PSEUDOMONAS OLEOVORANS

Diss. ETH No. 12987

DUAL (C,N) NUTRIENT LIMITED GROWTH OF

PSEUDOMONAS OLEOVORANS

A dissertation submitted to the

SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZÜRICH

for the degree of

Doctor of Natural Sciences

presented by

Manfred Stephan Zinn

Dipl. Natw. ETH

born June 28, 1967

in Bern, Switzerland

citizen of Lützelflüh, BE

accepted on the recommendation of

Prof. Dr. B. Witholt, examiner

PD Dr. H. Brandl, co-examiner

PD Dr. Th. Egli, co-examiner

Zürich 1998

3

TABLE OF CONTENTS

Summary 4

Zusammenfassung 6

Chapter 1 General introduction 9

Chapter 2 Dual nutrient limited growth: an overview 23

Chapter 3 Dynamics of growth and accumulation of

poly(3-hydroxyalkanoate) (PHA) in Pseudomonas

oleovorans cultivated in continuous culture under

transient feed conditions: rapid identification of dual

nutrient (C,N) limited growth zones

49

Chapter 4 Intracellular degradation of poly(3-hydroxyalkanoate)

in Pseudomonas oleovorans

79

Chapter 5 A journey to the unknown: triple (C,N,P) limited

growth of Pseudomonas oleovorans in chemostat

97

Chapter 6 Towards understanding of dual nutrient limited growth

of Pseudomonas oleovorans

115

Chapter 7 General conclusions 139

References 145

Curriculum Vitae 158

5

SUMMARY

Physiological studies of bacteria are usually carried out in liquid cultures. These cultures are

then often exposed to cultivation conditions where one specific nutrient is limiting growth.

Therefore synthetic media are used to adjust the nutrient composition such that only one

nutrient limits growth and all others are in excess. However, it has been shown that microbial

growth need not be limited by one particular nutrient only (as formulated in the concept of

the “law of the minimum” by Justus von Liebig 1840) but that several nutrients may limit

growth at the same time. Two basic theories have been at the basis of numerous relevant

experiments: the kinetic aspect focuses on the microbial growth rate and its restriction by

nutrients, whereas the stoichiometric aspect aims to describe the relationship between the

biomass and its production on growth limiting nutrient concentrations (Chapter 2).

A suitable organism to study the stoichiometric aspect and the consequences of

simultaneous multiple nutrient limited growth is Pseudomonas oleovorans. This Gram

negative bacterium is especially of interest as it intracellularly accumulates medium chain

length (C6 to C12) poly[(R)-3-hydroxyalkanoates] (mclPHAs) when grown on mcl alkanes,

alkanols, or alkanoic acids and is exposed to growth limiting nutrients such as N, P, Mg, or S.

MclPHA can be chemically extracted from the cells at high purity. This isolate has the

properties of a flexible plastic but is in addition biodegradable and may therefore replace

petroleum based plastic in future.

We investigated whether mclPHA can be produced in continuous cultures under dual

carbon-nitrogen limited growth conditions with octanoic acid and ammonium as sole carbon

and nitrogen sources, respectively (Chapter 3). In a chemostat culture three distinct and

stable growth regimes could be detected when the carbon to nitrogen ratio in the feed

medium (Cf/Nf) was stepwisely increased by the increase of the octanoate concentration: a

carbon limitation for Cf/Nf < 6.4 g g-1, a dual (C,N) limitation for 6.4 g g-1< Cf/Nf < 9.6 g g-1,

and a nitrogen limitation for Cf/Nf > 9.6 g g-1. The accurate determination of the boundaries

of the dual (C,N) limited growth regime is a rather tedious process because many data points

are needed. An alternate method was developed where the Cf/Nf ratio was continuously

changed by either increase or decrease of the carbon or nitrogen source, respectively. It was

General Introduction 6

found that the dual (C,N) limited growth regime was shifted towards higher Cf/Nf ratios in

the case where Cf was increased over time. In contrast, a shift towards lower Cf/Nf ratios was

observed when Cf was reduced or Nf was increased. These observations were interpreted as

time delays and could be corrected to some extent mathematically. The remaining deviation

was related to a biological delay, presumably caused by the need to adapt to ever changing

growth conditions (e.g. change from carbon excess to carbon limitation) and could not be

corrected mathematically.

Since the best results were obtained with a medium gradient where the carbon supply was

continuously reduced thereby reducing the Cf/Nf ratio. This process raised questions about

the PHA degradation.

In batch experiments where the cells were initially starved for carbon or carbon and

nitrogen simultaneously, P. oleovorans degraded intracellular mclPHA exponentially at a

high rate. The PHA C8 and C6 monomers were degraded at identical degradation rates,

whereas the C10 monomer remained almost constant. As a result, the polymer composition

changed continuously over time. This indicated that the C10 monomer is an artefact caused

by the methanolysis procedure of freeze dried cells. Further, it was found that the degradation

of PHA occurred, although protein synthesis was blocked by rifampicin. This observation

indicated that intracellular PHA was continuously and simultaneously accumulated and

degraded in P. oleovorans.

Chemostat experiments were performed to assess whether P. oleovorans was able to

accumulate PHA to a larger amount under triple nutrient limited growth conditions. In a

series of chemostat experiments it was shown for the first time that a triple limitation can be

established when carbon, nitrogen, and phosphorus are limiting growth simultaneously

(Chapter 5). Under these stringent growth conditions PHA was accumulated in large amounts

but did not exceed the amount accumulated under single phosphorus limitation. The triple

limited growth regime was found to be predictable for a given growth rate and could be

expressed as ranges of Cf/Nf and Cf/Pf ratios of the feed medium.

A black-box model was developed to predict the dual (C,N) limited growth regime for the

whole growth rate range of P. oleovorans (Chapter 6). Thus, an optimal PHA production

under dual (C,N) limited growth could be postulated for a Cf/Nf = 12.4 g g-1 and a dilution

rate of 0.21 h-1.

Summary 7

The results obtained in this work indicate that multiple nutrient limited growth can be

established in continuous cultures. P. oleovorans was found to be a microorganism that

quickly adapts to environmental changes and is able to grow and synthesize PHA under

stringent growth conditions such as triple (C,N,P) limitation.

8

ZUSAMMENFASSUNG

Physiologische Untersuchungen von Bakterien werden im allgemeinen mit Flüssigkulturen

durchgeführt. Die Wachstumsbedingungen, bei denen spezifisch ein Nährstoff wachstums-

limitierend ist und alle anderen im Überschuss sind, werden durch die Anwendung von

synthetischen Medien bewerkstelligt. Es wurde jedoch bereits gezeigt, dass das mikrobielle

Wachstum nicht unbedingt nur von einem Nährstoff limitiert sein muss (wie es Justus von

Liebig im “Gesetz des Minimums” 1840 formulierte), sondern dass mehrere Nährstoffe das

Wachstum gleichzeitig limitieren können. Zwei Theorien stellten die Grundlage für

zahlreiche massgebende Experimente dar: der kinetische Aspekt untersucht den Einfluss von

limitierenden Nährstoffen auf die mikrobielle Wachstumsrate, der stöchiometrische Aspekt

hingegen hat zum Ziel, das Verhältnis zwischen limitierendem Nährstoff und der

Biomassenproduktion zu beschreiben (Kapitel 2).

Pseudomonas oleovorans ist ein passender Organismus, um den stöchiometrischen

Aspekt und die Auswirkungen des vielfach limitierten Wachstums zu studieren. Dieses

Gram negative Bakterium ist von speziellem Interesse, denn es akkumuliert intrazellulär

Poly[(R)-3-hydroxyalkanoat] aus mittellangen (C6 bis C12) Monomeren (mclPHAs), wenn

es auf mcl Alkanen, Alkanolen oder Alkansäuren gezüchtet wird und einer gleichzeitigen

Nährstofflimitation von N, P, Mg oder S ausgesetzt ist. MclPHA kann mit hohem

Reinheitsgrad chemisch aus den Zellen extrahiert werden. Dieses Extrakt besitzt die

Eigenschaften eines flexiblen Kunststoffes, ist aber zusätzlich biologisch abbaubar und

könnte deshalb den aus Erdöl gewonnenen Kunststoff zukünftig ersetzen.

Wir untersuchten, ob P. oleovorans mclPHA in kontinuierlicher Kultur unter simultaner

C- und N-Limitation, mit Oktansäure als einziger C- und Ammonium als einziger N-Quelle,

mclPHA produzieren kann (Kapitel 3). In einer Chemostatkultur konnten drei zu

unterscheidende und stabile Wachstumsbereiche detektiert werden, wenn das Kohlenstoff zu

Stickstoff-Verhältnis im Mediumzufluss (Cf/Nf) schrittweise durch die Zunahme der Oktan-

säurekonzentration erhöht wurde: eine Kohlenstofflimitation für Cf/Nf < 6.4 g g-1, eine

(C,N)-Doppellimitation für 6.4 g g-1 < Cf/Nf < 9.6 g g-1 und eine Stickstofflimitation für Cf/Nf

> 9.6 g g-1. Die genaue Bestimmung der Grenzen der (C,N)-Doppellimitation ist wegen der

Zusammenfassung 9

vielen dazu benötigten Datenpunkte ein umständliches Verfahren. Wir entwickelten eine

alternative Methode, bei der das Cf/Nf Verhältnis während des Experimentes kontinuierlich

durch die Erhöhung oder Erniedrigung der C- oder N-Quelle verändert wurde. Bei

kontinuierlicher Erhöhung von Cf konnte eine Verschiebung des doppelt-limitierten

Wachstumsbereiches zu erhöhten Cf/Nf Verhältnissen festgestellt werden. Im Gegensatz

dazu, wurde bei einer kontinuierlichen Erhöhung von Nf oder einer Verringerung von Cf eine

Verschiebung der (C,N)-Doppellimitation zu kleineren Cf/Nf Verhältnissen gemessen. Diese

Beobachtungen wurden als Zeitverzögerungen interpretiert und konnten bis zu einem

gewissen Masse mathematisch korrigiert werden. Die verbleibenden Abweichungen wurden

einer zusätzlichen biologischen Verzögerung zugeordnet, deren Ursache wahrscheinlich in

der nötigen Anpassung der Zellen an die andauernde Veränderung der Wachstumsbe-

dingungen (z.B. Wechsel von C-Überschuss zu C-Limitation) zu suchen wäre.

Die beste Annäherung an die Chemostat Daten wurden mit einem Mediumsgradienten

erzielt, bei dem die C-Zufuhr kontinuierlich reduziert wurde (Cf/Nf abnehmend), was

allerdings Fragen bezüglich des PHA-Abbaues aufwarf. Satzkulturen von P. oleovorans , die

zu Beginn C-oder (C,N)-limitiert waren, bauten intrazelluläres mclPHA exponentiell mit

hoher Rate ab (Kapitel 4). Die C8 und C6 Monomere des PHAs wurden mit gleicher Rate

abgebaut. Im Gegensatz dazu blieb die Menge von C10 Monomeren praktisch konstant. Die

Polymerzusammensetzung veränderte sich kontinuierlich, was darauf hinwies, dass die C10

Monomere ein Artefakt des Methanolyseprozesses von gefriergetrockneten Zellen waren. Im

weiteren wurde festgestellt, dass sich der Abbau von PHA trotz Blockierung der

Proteinsynthese durch Rifampicin fortsetzte. Diese Beobachtungen weisen darauf hin, dass P.

oleovorans das intrazelluläre PHA gleichzeitig ein- und abbaut.

Es wurden Chemostatexperimente durchgeführt, um zu untersuchen, ob P. oleovorans

fähig ist, grössere Mengen an PHA unter dreifacher Nährstofflimitation zu synthetisieren.

Dabei konnte zum ersten Mal eine Dreifachlimitation eingestellt werden, bei der die Nähr-

stoffe C, N und P das Wachstum gleichzeitig limitierten (Kapitel 5). Unter diesen

Wachstumsbedingungen wurde PHA in grossen Mengen eingelagert; der PHA-Gehalt unter

P-Limitation konnte dabei jedoch nicht übertroffen werden. Die dreifach limitierte

Wachstumszone konnte für die Wachstumsrate µ = 0.2 h-1 und für die dazu einzustellenden

Cf/Nf und Cf/Pf Verhältnisse bestimmt werden.

Aus den gewonnenen Erkentnissen wurde ein “black-box” Modell entwickelt, das die

(C,N)-Doppellimitation für den ganzen Wachstumsgeschwindigkeitsbereich von P. oleovo-

rans voraussagen kann (Kapitel 6). Dies führte zu einer theoretisch optimalen PHA-Pro-

duktion unter (C,N)-Doppellimitation mit Cf/Nf = 12.4 g g-1 und µ = 0.21 h-1.

Die Resultate dieser Arbeit zeigen, dass Vielfachlimitationen in kontinuierlichen

Kulturen erstellt werden können. Es konnte der Nachweis gebracht werden, dass P.

oleovorans ein Mikroorganismus ist, der sich schnell den Umweltbedingungen anpasst und

fähig ist, unter bestimmten Wachstumsbedingungen, wie der (C,N,P)-Dreifachlimitation, zu

wachsen und PHA zu synthetisieren.

Zusammenfassung 11

12

CHAPTER 1

GENERAL INTRODUCTION

Keywords: Pseudomonas oleovorans, bioplastic, poly(3-hydroxyalkanoate), metabolism,

PHA granules, waste, scope of thesis.

Chapter 1 13

Introduction

Within the last 50 years petrochemical plastics have joined the ranks of our most applied

materials. Their success was caused by their versatility, outstanding technical properties, and

relatively low price. Today’s applications are nearly universal: automobiles, home

appliances, computer equipment, construction, sport and leisure equipment, packages and

even medical applications are areas, where plastics clearly dominate. The enormous need for

this material can be expressed in a few numbers. In 1996 Switzerland produced 133’000 t

plastics, imported 1’552’000 t, and 911’000 t were exported (Schweiz, 1997). About

690’000 t were disposed of in landfills or incinerated, and only 58’000 t were recycled. There

are no plans to increase the recycling of plastics due to high costs (BUWAL, 1991;

Tellenbach, 1998).

Various studies have shown that these plastics are environmentally neutral, which means

they are not biologically degraded. Since landfills will be forbidden in Switzerland by the

year 2000 (TVA, 1990), communities have to burn their waste in large incinerators.

Although a very high percentage of the production energy can be regained by this method,

the former fixed carbon is released into the environment as carbon dioxide, which is known

to contribute to the green-house effect with the associated climate changes. The addition of

carbon dioxide to the global carbon cycle from incineration of petro-plastics could be

stabilized or even reduced when renewable resources are used for the production of plastics.

First materials that were able to replace petro-plastics were found in the mid-seventies,

although the motivation for the evaluation of alternates to plastic was different. At that time

the crude oil prices increased due to a predicted end of the oil reserves and an artificially

induced shortage by the oil producing countries. Since then many biopolymers have been

checked for their possible industrial applications and their biodegradability; cellulose, starch,

blends of those with synthetic polymers, poly(lactic acid), polyester-amide, and

poly(3-hydroxyalkanoate) (PHA). The latter were of particular interest since they can be

produced biotechnologically at a high purity (100% stereospecificity, since all chiral carbon

atoms are in D(-) configuration), which increases their biodegradability and biocompatibility


significantly (Brandl et al., 1990). Thus, the PHAs seem to have an enormous potential to be

applied in medical applications, such as blood vessel replacement, wound dressing, bone

replacement and plates. Studies showed that even anticancer drugs can be released at a well

defined rate when they are integrated in PHB microspheres (Abe et al., 1992). The

implantation of such drug dispensers could help patients that need daily medication.

In 1987 the global market for biodegradable plastics was estimated to be around 1.5 * 106

metric tons by the year 2000 (Lindsay, 1992). To date, however, the price of the only

commercial PHA, BIOPOLTM produced by Monsanto, USA, is still too expensive with 15

US$ kg-1 (Lee, 1996b) to become a bulk plastic, since the chemically synthesized

polyethylene, polystyrene, and polypropylene cost between 0.62 and 0.96 US $ kg-1.

Recently, studies were performed that aim to transfer the production of bioplastics to crop

plants using genetical engineering (Nawrath et al., 1995; Poirier et al., 1992a; Poirier et al.,

1992b; van der Leij and Witholt, 1995). This change of the production host would reduce the

price of the bioplastic close to that of starch (Lee, 1996b), because the cultivation costs for

plants are very low.

Occurrence of PHA

A first type of PHA, namely PHB, was discovered in Bacillus megaterium in 1926 by the

French scientist Lemoigne (1926). He found that the polymer consisted of 3-hydroxy butyric

acid monomers that were linked through an ester bond between the 3-hydroxyl group and the

carboxylic group of the next monomer (Fig. 1.1). Since the discovery of PHB more than 90

genera of microorganisms have been found, not only aerobic but also anaerobic heterotrophic

ones, that are able to accumulate PHB or other PHAs intracellularly (Steinbüchel, 1991a). It

is even suggested that the occurrence of PHA is in general an environmental marker for

unbalanced growth, since PHA is generally produced when the carbon source is in excess

(Findlay et al., 1990; Herron et al., 1978; White and Ringelberg, 1997). However, there are

also reports that organisms accumulate PHA under carbon limiting growth conditions

(Dunlop and Robards, 1973; Matin et al., 1979; Tal and Okon, 1985).

Polyhydroxybutyrate

15 Chapter 1

The best studied PHA is PHB. The polymer chain has a compact right-handed helix stiucture

and is optically active due to the fact that all hydroxybutyrate monomers are in R

configuration (Comibe1t and Marchessault, 1972). It has a molecular weight up to 1.4 * 106

Da (Doi, 1990), and a melting point of about 180° C. These prope1ties seemed to be sufficient

for an industi·ial production for the British company Imperical Chemical Industiy (ICI)

(Holmes, 1985). However, PHB has been found to be more brittle and stiffer than

polypropylene, which has reduced the field of possible applications. These prope1ties were

improved, when a PHA copolymer was fonned with 3-hydroxyvalerate (3HV) and

3-hydroxybutyrate. The result was a softer plastic with a lower melting point (102 - 157°C),

that was more flexible when the 3HV content was increased (Bloembergen et al., 1986). This

copolymer, which has the ti·ade name BIOPOL Thi' is now produced with an Alcaligenes

eutrophus sh'ain by the American company Monsanto. There are also plans to produce PHA

in large amounts using crop plants like rapeseed, soybean and sunflower (Poirier et al. 1992a;

1995). This could decrease the price of PHB to as low as 0.5 US$ kg-1 .

H ..... )i ' °"'c/~c OH

H

' . II R H 0

n

FIG. 1.1: Chemical stiucture of poly(3-hydroxybutyrate) (PHB) and medium chain length poly(3-hydroxyalkanoate) (mclPHA). All monomers have one chiral center(*) in R position. Poly(3-hydroxybutyrate): R = CH3, poly(3-hydroxyvalerate): R = C2Hs, mclPHA: R = C3H1 - C11H2s, n = 100 - 30,000 monomers.


A special class of PHB with a lower molecular weight (cPHB; Mr < 14,000 Da) was found in

Bacillus subtilis, and Azotobacter vinelandii (Reusch and Sadoff, 1988), in association with

polyphosphate and calcium ions. An interesting observation was that competent (ready to

exchange genetical information) Escherichia coli cells have a similar, complex structure in

their membranes that may serve as a voltage gated calcium channel (Huang and Reusch,

1995). Recent investigations revealed that cPHB is not only found in Eubacteria, but also in

eucaryotic cells, e.g. yeasts, peanut, spinach, sheep (intestine), cat muscles and many more

(Reusch, 1995). It can be assumed that PHB is an omnipresent polymer in nature.

Structure of mclPHA

PHAs with medium chain length of the monomers (mclPHAs, C6 - C14) were first found in

Pseudomonas oleovorans (de Smet et al., 1983). When P. oleovorans was grown on n-octane

as sole carbon source under growth limiting conditions by ammonium, the polymer consisted

of 3-hydroxyoctanoate and 3-hydroxyhexanoate monomers (de Smet et al. 1983). More

organisms could be found that are able to produce mclPHA. Huisman et al. (1989) concluded

that all fluorescent pseudomonads belonging to the rRNA homology group I are able to

synthesize such mclPHAs.

To date, P. oleovorans is the best investigated mclPHA producer. It is able to grow on a

wide variety of different substrates that can be incorporated into PHA, ranging from

n-alkanoic acids (Tab. 1.1), n-alkanals, to n-alkanes with a carbon backbone with a length of

6 to 14 carbon atoms. P. oleovorans is also able to produce PHA and to grow on unsaturated

alkenes (Lageveen et al., 1988), on branched alkanoates (Fritzsche et al., 1990a; Scholz et al.,

1994), on halogenated end groups such as brominated (Kim et al., 1992), fluorinated (Hori et

al., 1994; Kim et al., 1996) and chlorinated molecules (Doi and Abe, 1990).

TAB. 1.1: PHA polymer composition of P. oleovorans grown on different alkanoic acids in 50 ml 0.5*E2 cultures (adapted from (Huisman et al., 1992).

Composition (% of total 3-hydroxyfatty acids)

Substrate PHA C6 C7 C8 C9 C10 C11 C12

Chapter 1 17

[% cdw]

Butyrate 0.6 33 67

Valerate 0.7 35 65

Hexanoate 3.3 95 5

Heptanoate 2.3 100

Octanoate 8.7 8 91 1

Nonanoate 9.1 35 65

Decanoate 12.5 8 75 17

Undecanoate 9.8 28 59 13

Dodecanoate 6.6 6 57 32 5

Tridecanoate 5.4 32 48 5 14

Tetradecanoate 10.6 7 59 30 4 C6: 3-hydroxyhexanoate, C7: 3-hydroxyheptanoate, C8: 3-hydroxyoctanoate, C9: 3-hydroxynonanoate, C10: 3-hydroxydecanoate, C11: 3-hydroxyundecanoate, C12: 3-hydroxydodecanoate.

There are reports that even aromatic groups can be incorporated when they have a

sidegroup of at least five carbon atoms (Curley et al., 1996; Hazer et al., 1996). Probably a

reason for the large variation of the PHA compounds is the fact that P. oleovorans can only

produce PHA with substrates that are related to the growth substrate (Lageveen et al., 1988;

Williams and Peoples, 1996).

Along with the change of the length of the side chain of the monomers, the properties of

the bioplastic, e.g. melting point, glass transition temperature, and cristallinity

(stiffness/flexibility), varies considerably (Eggink et al., 1995; Preusting et al., 1990).

Accordingly, the molecular weight of the polymer depends on the carbon source. When P.

oleovorans was grown on n-alkanes from C6 to C10, the average molecular weight for PHA

varied between 1.8 and 3.3 * 105 Da (Preusting et al., 1990). The mclPHAs that contain

unsaturated 3-hydroxyfatty acids can be crosslinked with each other resulting in a

biodegradable rubber (de Koning et al., 1994; de Koning and Witholt, 1996; Molitoris et al.,

1996).

Synthesis of PHB


The formation of PHA precursors occurs in the cytosol, whereas the polymerization step is

on the surface of the PHA granule. The synthesis of PHB (Fig. 1.2) and mclPHA (Fig. 1.3)

differs with respect to the complexity.

The PHB synthesis occurs in a 3 step reaction starting with two acetyl CoA that are

derived from the tricarboxylic acid cycle (TCA). The enzyme ß-ketothiolase condenses the

acetyl CoA to one acetoacetyl CoA molecule which is then reduced at the 3 position by the

enzyme acetoacetyl CoA reductase and the cofactor NADPH + H+. This reaction occurs

stereospecifically which means all resulting 3-hydroxybutyryl CoA are in the R

configuration at position 3. In the final step the monomers are attached to the PHB polymer

by the polymerase through an ester bond.

The first key step is generally regulated by the intracellular concentration of AcCoA.

Under normal growth conditions that means that without a limitation by a nutrient, the pool

of free CoASH is enhanced. However, when the growth is nutrient limited the CoASH level

is reduced and the PHB synthesis is favored.

2O

SCoA

O O

CoASHSCoA

OH O

SCoAPHB

CoASHNADPH+H+ NAD+

1 2 3

FIG. 1.2: PHB synthesis occurs in 3 steps. 1: ß-ketothiolase, 2: acetoacetyl CoA reductase, 3: PHB polymerase.

An increased PHB content of Azotobacter beijerinckii was found under oxygen limitation

and was interpreted by Senior and Dawes (Senior et al., 1972) as an alternate electron sink for

oxygen into which reducing power would be channeled.

Synthesis of mclPHA

The metabolic pathway leading to mclPHA is more complex than that for PHB synthesis.

The substrate, e.g. a fatty acid, is channeled through the ß-oxidation cycle. Its intermediates,

2-trans-enoyl-CoA, S-3-OH-Acyl-CoA, and 3-ketoacyl-CoA, are candidates to be the

precursors for the R-3-OH-Acyl-CoA (van der Leij and Witholt, 1995). R-3-OH-Acyl-CoA

Chapter 1 19

is then condensed to mclPHA by the PHA polymerase through an ester bond. Interestingly, P.

oleovorans and P. aeruginosa have two polymerases, PhaC1 and PhaC2 (Huisman et al.,

1991; Steinbüchel et al., 1992). They have a base pair identity of 53.6% (Huisman, 1991).

The polymerase PhaC2 has a higher affinity for 3-hydroxyhexanoate monomers when the

gene is expressed in a polymerase lacking mutant of P. putida grown on octanoic acid

(Huisman et al., 1992). When the environmental growth conditions become limiting due to a

low carbon supply, mclPHAs are degraded by the PHA depolymerase delivering organic

carbon/energy for cell energy maintenance (Steinbüchel, 1991b) through the tricarboxylic

cycle. The PHA depolymerase is encoded by the gene phaZ and is located between phaC1

and phaC2 on the same open reading frame (Huisman et al., 1991).

P. oleovorans harboring the OCT plasmid (Schwartz and McCoy, 1973; van Beilen et al.,

1994) is able to grow on n-alkanes, n-alkanols, and n-alkanals consisting of 6 to 12 carbon

atoms whilst following the alkane oxidation pathway. However, in contradiction to P. putida

KT2442, P. oleovorans does not accumulate PHA when grown on fructose, glucose, and

glycerol and produces only small amounts of mclPHA under carbon limited growth

conditions. Huijberts et al. (1992) showed that PHA produced by P. putida grown on glucose

consists of saturated and unsaturated monomers and moreover that the PHA precursors made

on medium chain length fatty acids can be elongated by a C2 unit, presumably caused by a

reversed action of the enzyme 3-ketothiolase (Huijberts et al., 1995). Inhibition experiments

with acrylic acid (ß-oxidation) and cerulenin (de novo fatty acid synthesis) revealed that P.

putida seems to be able to synthesize PHA on de novo fatty acid synthesis, too (Eggink et al.,

1992; Huijberts et al., 1995).

@ Alkane-Alkanol--Alkanai--.

Acyl-CoA

0

R~ scoA B-Oxidation 2-trans-enoyl-CoA Cycle

0 0

• AA R SCoA

\ @ ~H 0 ~K•;•t<AA @ I N\_ ~r

\ R SCoA I \ S-3-0H-Acyl-CoA I @ \ I I

0

R~O ~ Alkanoate ~ ~

R ACP

Acyl-ACP

0

R~ACP 2-trans-enoyl-ACP

o ~I ASCoA

Acetyl-CoA de novo Fatty Acid

Synthesis

\ TCA

i 3-Ketoacyl-ACP

RMACP R-3-0H-Acyl-ACP

Biomass

a Alkane oxidation pathway i PHA depolymerase ' tCD /1

~ OHO ~/ @ f'~ ~ RM SCoA _4 ___.(D....._i ~•. lR~~H

b Acyl-CoA synthetase k 3-Keto-acyl reductase c Acyl-CoA dehydrogenase I 3-Ketothiolase d Enoyl-CoA hydratase m 3-Keto-acyl-ACP reductase e Enoyl-CoA hydratase n 3Hydroxy-acyl-ACP dehydratase f 3-Hydroxy-acyl-CoA epimerase o Enoyl-ACP reductase

R-3-0H-Acyl-CoA mclPHA

g 3-Keto-acyl-CoA reductase p Fatty acid thioesterase h PHA polymerase

FIG. 1.3: Pseudomonads of the rRNA homology group 1 are able to accumulate mclPHA through de novo fatty acid synthesis and subsequent B-oxidation. Only P. oleovorans harboring the OCT plasmid is able to degrade n-alkanes through the alkane oxidation pathway (a). Dashed ~mows represent hypothetical steps.

21 Chapter 1

PHAs are stored in granules

PHB and many other PHAs are collected in light refracting granules inside the cell. This site

specific accumulation is advantageous since the polymers are separated from the cell lumen

and consequently do not affect the osmotic pressure of the cell extensively (Anderson and

Dawes, 1990). The number of granules per cell seem to be strain specific, e.g. Alcaligenes

eutrophus has between 8 and 12 with variable sizes (Ballard et al., 1987). Although many

electron microscopy studies have been described, the formation and the basic structure of the

granules is not yet clear. Numerous models of the granule structure were published (Ballard,

Holmes and Senior, 1987; de Koning and Maxwell, 1993; Ellar et al., 1968; Fuller et al.,

1992; Horowitz et al., 1993; Lauzier et al., 1992; Lundgren et al., 1964; Mayer and Hoppert,

1997; Mayer et al., 1996; Mccool et al., 1996; Merrick et al., 1965; Stuart et al., 1995;

Wieczorek et al., 1996). A general model presented by Steinbüchel et al. (1995) is accepted

as a basic structure of granules (Fig. 1.4). According to this model the PHA polymers are

enwrapped by a lipid (0.5% w/w) and mainly by a protein (mono-)layer (2% w/w) (Lundgren

et al., 1964). There is no evidence that proteins or lipids are integrated in the core of the

granules (Dunlop and Robards, 1973).

The granule associated proteins can be compiled in four classes:

Class 1: PHA polymerases

Class 2: PHA depolymerases

Class 3: Phasins (granule associated proteins that have a stabilizing function)

Class 4: Proteins with unidentified function

The PHA polymerases seem to differ for each strain, although the central regions were highly

homologous between 21% and 87.3% for 22 different bacteria (Steinbüchel et al., 1992). It

can be assumed that the PHA content is dependent on the polymerase content of the cells,

since this is the case for P. oleovorans (Kraak et al., 1997). It is assumed that the PHA

polymerase is located on granule surfaces and is responsible for the formation of a granule.


To date, it is not known how the intracellular degradation of PHA occurs. First studies by

MeITick and Doudoroff (1964) indicated that the depolymerase needs to be activated by an

unidentified activator in order to degrade PHB granules. Isolated granules of Bacillus

megaterium were degraded by a soluble enzyme fraction of polymer depleted cells of

Rhodospirillum rubrum . The fraction could be separated into an activator and depolymerase

fraction. No PHB degradation occmTed when the activator was applied alone. They found

that trypsin could replace the activator, but, had no PHB degrading activity. Their conclusion

was that the activator and trypsin may remove some protective substance from the granule

surface, thus multiplying the free sites of attachment for the depolymerase. Whether this

mechanism is common for all granules is not known.

Phasins/proteins

PHA depolymerase

Phospholipid monolayer

FIG. 1.4: Scheme of a PHA granule. The core consists of PHA polymers which is covered by a lipid monolayer with integrated proteins. The lipid monolayer points with the hydrophobic side to the core. The integrated proteins consist of PHA polymerase, PHA depolymerase, str11ctural proteins (phasins ), and proteins of unknown function.

The third class of granule associated proteins, the phasins, Inight function as temporaiy

protectors of the granules from the depolymerase, since it is known that phasins can inhibit

agglutination of isolated granules in P. oleovorans (Foster et al. , 1994). It is even speculated

23 Chapter 1

that the phasins have a protective function to reduce the passive attachment of cytosolic

proteins (Steinbüchel et al., 1995).

Production of PHA

PHA can be produced in microorganisms with various culture methods, such as batch,

fed-batch, and chemostat cultures. The batch method is probably most often applied in

research, since it is very convenient to do growth studies and screening of potential PHA

accumulating organisms. Generally, the medium is designed in such a way that a nutrient, e.g.

nitrogen, limits the growth of the biomass and the carbon source is in excess. Depending on

the organism and on the growth medium the experiments are performed within one to two

days. During that time the organisms go through a sequence of growth stages, such as lag

phase, exponential growth phase, stationary phase, and finally the death phase (Monod, 1949;

Pirt, 1975). Concomitantly, the cells perceive a continuous change of their environment due

to the ever changing nutrient concentration caused by the nutrient consumption. This leads to

a significant problem when batches are used as a fast screening method to detect PHA

containing cells, since cells that are starved for carbon start to degrade PHA again (Anderson

et al., 1990). Thus, data that are produced with batch cultures may vary considerably

depending on harvest time.

The fed-batch culture is basically a batch culture that is continuously supplemented with

nutrients after it enters the late exponential phase (Pirt, 1975). Today, PHA is produced in

such a fed-batch mode on a large scale: BIOPOL™, which is a copolymer consisting of

poly(3-hydroxybutyrate) and poly(3-hydroxyvalerate). It is produced by the American

company MONSANTO with a modified strain of Alcaligenes eutrophus. The production

process consists of two main phases: In a first phase, the cells are cultured in a minimal

medium which contains the essential growth nutrients, glucose, and only low amounts of

phosphate, supporting only cell growth to a certain biomass concentration. Consequently,

only minor PHA accumulation occurs (Byrom, 1987) at that time. In a second phase, where

all phosphorus is consumed by the cells, PHA accumulation takes place. The production is

enhanced through the continuous addition of glucose and propionic acid to the culture at well

defined rates. After 48 hours of feed the copolymer BIOPOL™ consists of about 80%


poly(3-hydroxybutyrate) and 20% poly(3-hydroxyvalerate). The process is stopped when the

PHA content of the cells has reached a desired level, somewhat between 70% and 80% of the

cell dry weight. The fed-batch process is the culture method of choice since propionic acid is

toxic to the cells and can inhibit cell growth at high concentrations. The advantage of

fed-batch cultures in general are the high cell densities (Preusting et al., 1993b) that can be

obtained which reduce the costs of PHA purification significantly. A disadvantage of the

process is that the cells do not grow at a constant growth rate when the feed rate and feed

concentration are kept constant. The reason for this is that the added nutrients are consumed

by an ever increasing cell concentration in the same time unit. This can lead to unexpected

losses in PHA production (Suzuki et al., 1986a; 1986b).

The third method to produce PHA biotechnologically, the chemostat, is the most

controlled cultivation method. In such a system the growth medium is continuously added to

a culture under sterile conditions. The liquid volume of the culture is kept constant with an

overflow device or a scale controlled harvest pump. According to the theory of Monod

(Monod, 1942) the specific growth rate of the culture can be set by the volume of the culture

and the rate of volume changes. This unique feature is appropriate to determine the influence

of well defined growth conditions on PHA accumulation (Egli et al., 1986; Hazenberg, 1997;

Preusting et al., 1993a). However, this method is not yet used in the PHA production on a

large scale, although a high productivity of PHA can be obtained when appropriate growth

conditions are selected (Lee, 1996a; Lee, 1996b). The potential of the chemostat can be

increased further when the two chemostats are connected in sequence. Hazenberg and

Witholt (1997) reported that the mclPHA productivity in P. oleovorans could be increased

significantly and that the PHA content could be increased to more than 60% which is the

highest value reported to date.

In all methods mentioned here the medium composition is an important factor for the

successful production of PHAs. In addition, it is also a significant cost factor in a large scale

PHA production. Thus, it is necessary to design media such that only the amount of nutrients

needed for an optimal cell and PHA production are used. In other words, growth should be

limited by several nutrients at the same time. To cover this subject in more detail, the

following chapter gives an overview of multiple nutrient limited growth experiments

described in the literature.

25 Chapter 1

Aim and scope of the thesis

Microorganisms are generally cultured in media where a single nutrient limits the growth and

determines the productivity for many microbial processes. Recently, the existence of well

defined dual nutrient limited growth regimes has been proved during the continuous

cultivation in bioreactors (Egli, 1991). However, there is little knowledge about the extension

of the dual nutrient limited growth regime for various macro-nutrients, such as carbon,

nitrogen or phosphorus. There are indications that under dual nutrient limited growth

conditions a special physiological state of the cell can be achieved where higher

productivities of enzymes and metabolites were reported. This special physiology cannot be

achieved by traditional culture methods.

We want to demonstrate in this thesis the existence of multiple nutrient limited growth for

Pseudomonas oleovorans, a bioplastic (PHA) producing bacteria, and to investigate the

physiological consequences of such growth conditions. Especially of interest is whether the

amount of accumulated bioplastic can be increased through dual limited growth.

In Chapter 2, after this general introduction, a literature overview gives an introduction to

dual nutrient limited growth and its stoichiometric and kinetic aspect in particular. It was

found that the systematic approach using chemostats gave the best description of the dual

nutrient limited growth regime. However, the determination of dual nutrient limited growth

in a chemostat is rather a tedious process, since for each measurement a steady-state of the

culture is needed. Chapter 3 describes the development of an automated system where the

medium of a continuous culture of P. oleovorans was continuously changed by the increase

or decrease of the carbon (octanoate) or the nitrogen (ammonium) concentration,

respectively. Dual (carbon,nitrogen) limited growth could be determined by the on-line

analysis of the residual carbon and nitrogen concentration in an efficient way. Further, a

mathematical model was developed to correct for a wash-in effect of the medium gradient in

the culture broth. Interestingly, the best results compared to a traditional chemostat

experiment was obtained by a medium gradient experiment, where the carbon source in the

feed medium was continuously decreased over time. During this experiment the PHA content


of P. oleovorans was reduced continuously. In order to estimate how fast such PHA

degradation may occur, a PHA degradation study with P. oleovorans was performed

(Chapter 4). Besides the total PHA content, the monomeric composition was followed over

the degradation process during starvation for nitrogen and dual (carbon,nitrogen). In addition,

it was investigated whether the blockade of the protein synthesis can inhibit the degradation

of PHA. Chapter 5 documents a journey to the unknown: triple nutrient limitation of P.

oleovorans by the three macro-elements carbon, nitrogen, and phosphorus. For the first time

triple nutrient limited growth could be documented for at least 6 steady-state cultures. Our

findings rose the question whether the growth conditions for multiple nutrient limitation

could be predicted for the whole growth rate range. This lead to the design of a black-box

model described in Chapter 6 which gives an extended stoichiometric comprehension of the

dual (C,N) limited growth regime of P. oleovorans. Chapter 7 is the conclusion of the thesis

and gives some prospect how the dual limited growth regime could be used for the industrial

production of PHA with P. oleovorans.

27

CHAPTER 2

DUAL NUTRIENT LIMITED GROWTH: AN OVERVIEW

Manfred Zinn, Thomas Egli, and Bernard Witholt

Overview 28

Keywords: Dual nutrient limited growth regime, kinetics, stoichiometry, limitation, batch,

fed-batch, chemostat, model, theory.

Chapter 2 29

SUMMARY

Dual nutrient limited growth, the control of the cell growth rate (kinetic aspect) or the

restriction of the amount of biomass (stoichiometric aspect) by two nutrients at the same time,

is a relatively unknown ability of the microorganisms and consequently still not mentioned in

the text books to date. Thus, it is the aim of this survey to close the gap and give an

introduction to this special topic with respect to basic models and published experimental

data.

It was found that nutrient limited or controlled growth was investigated and observed in

many different systems e.g. biofilm on glass beads and sand particles, batch, fed-batch, and

chemostat cultures. The best results were obtained in chemostat cultures because the growth

conditions can be set specifically and the experiments are highly reproducible.

Generally, distinct dual nutrient limited growth was observed when the microorganism of

interest a) had a variation of the basic cellular composition (e.g. RNA, Protein), b) was able

to accumulate a storage compound (e.g. polyphosphate, lipid, polyglucan,

poly(3-hydroxybutyrate)), c) changed the physiological efficiency (e.g. biomass/ATP

consumed and biomass/CO2 produced), or d) excreted metabolic intermediates into the

culture broth under different nutrient limitations. Consequently, stoichiometric models

dealing with biomass yield coefficients have been developed that help significantly to

estimate the growth conditions in a chemostat to obtain dual nutrient limited growth.

The general problem of the kinetic experiments is the accurate detection of the growth

controlling nutrients in the culture broth. This is not a trivial task because the cells may

consume residual nutrients very rapidly, e.g. while sampling from a batch culture. Moreover,

the detection of the nutrients to a very low concentration (µg l-1 range) represents still a major

problem. Nevertheless, most theoretical models of dual limited growth deal with the kinetic

aspect although the experiments are difficult to carry out since the residual concentrations of

the nutrients in the culture broth have to be determined most accurately.

The presented overview indicates that the two aspects, kinetic and stoichiometric, were

treated as being independent of each other.

Overview 30

INTRODUCTION

The basic tenet of biological nutrition, the “law of the minimum”, was formulated by the

German chemist Justus von Liebig in 1840. He stated that always a particular nutrient limits

the growth of biomass that can be produced in a biological system (von Liebig, 1840). This

concept has influenced the thinking of microbiologists for many years. Consequently,

synthetic media for the cultivation of microorganisms are usually designed such that one

specific nutrient limits the amount of biomass that can be formed; the biomass that can

theoretically be produced in such a system is a function of the initial concentration and the

biomass yield coefficient for this particular nutrient (Fig. 2.1a).

A different approach focuses on the kinetic aspect of biomass growth. The work of

Blackman (1905) and Monod (1942) opened the way for the description of growth using

kinetic models. In these models it is assumed that the concentration of a substrate and the

affinity of the cell towards this substrate determine the rate of increase of biomass (Fig. 2.1b).

Monod

Blackmanµ

ss01 (s0n = const.)

Xs1s2

s2 s1

Limitation by s1 Limitation by s2

X

a b

FIG. 2.1: Text book theory of nutrient limitation Panel a:Stoichiometric growth limitation. The amount of biomass that can be

produced in a system is limited by only one nutrient (law of the minimum, von Liebig, 1840). When the concentration of the limiting nutrient is increased, the biomass increases until a second nutrient becomes limiting, and consequently the biomass does not increase any further.

Panel b: Kinetic growth limitation. The growth rate of the biomass in a system is dependent on the actual concentration of the “limiting” nutrient. Various

Chapter 2 31

models, e.g. Blackman (1905) and Monod (1942), describe the growth rate as function of the actual nutrient concentration s.

X biomass concentration; S01 input concentration of the nutrient 1; s1, s2 residual concentrations of nutrients 1 and 2; µ specific growth rate; s nutrient concentration in the culture broth.

More sophisticated culture techniques, such as the invention of the chemostat (Monod,

1950; Novick and Szilard, 1950), allowed to investigate the stoichiometry (Herbert, 1961)

and kinetics of cell growth (Koch, 1982; Powell, 1967; Senn et al., 1994) in a more controlled

way. Quantitative studies revealed that the specific growth rate and the nature of the limiting

nutrient have decisive influences on the cell composition (Herbert, 1976; Herbert et al.,

1971), whereas kinetics focused on the affinity of cellular uptake systems towards nutrients

and the observed specific growth rate (Button, 1985; Button, 1991; Roels, 1983).

However, in contrast to what is written in many text books, the law of the minimum is a

rather simplistic description of biological growth conditions. There are many reports

showing that the growth of the biomass and the biomass concentration in a system are not

limited by one nutrient only, but rather by two or more nutrients simultaneously (dual or

multiple nutrient limitation). In this chapter a brief survey is given on studies where dual

nutrient limitation has been observed.

NOMENCLATURE

As we concentrate here on dual nutrient limited growth, the subsequent definitions and their

synonyms are described explicitly for the case where two or more nutrients interact. In

general, nutrients can be grouped according to their physiological function.

Homologous nutrients are nutrients that fulfill the same physiological function during

growth (Harder and Dijkhuizen, 1976). For instance, for a heterotrophic microorganism two

organic carbon compounds can be homologous, because they fulfill the same physiological

function. Other authors used the synonyms “perfectly substitutable substrates” (León and

Overview 32

Tumpson, 1975; Ramakrishna et al., 1997), or “mixed substrates” (Egli, 1995; Harder and

Dijkhuizen, 1976; Harder and Dijkhuizen, 1982; Narang et al., 1997).

Heterologous nutrients are used to satisfy different physiological requirements (Egli,

1995). For instance, a nitrogen (ammonium) and a carbon source (glucose) are heterologous

nutrients because one cannot be replaced by the other (at least when they contain no C or N,

respectively). The synonyms also used are “noninteractive“ (Bader, 1978; Bader et al., 1975),

“complementary nutrients” (Baltzis and Fredrickson, 1988; Straight and Ramkrishna, 1994),

or “essential substrates” (Tilman, 1980).

According to Baltzis and Frederickson (1988), two additional (functional) combinations

of nutrients can be distinguished/defined when a nutrient can fulfill two functions at the same

time. For instance, one substrate can be partially homologous to a second substrate, whereas

the second substrate is entirely homologous to the first substrate (for instance glucose and

alanine). This corresponds to the synonymous expression of “partially substitutable and

entirely substitutable” substrates by Baltzis and Frederickson (1988).

Additionally, combinations of substrates can be partially homologous and partially

heterologous (for instance, glucosamine and methylamine for a methylotrophic organism). A

corresponding synonym is “partially substitutable and partially complementary”.

Bader and coworkers (1978) divided the nutrients according to their physiological

function into “noninteractive” (heterologous) and “interactive” (homologous, partially

homologous and heterologous, and partially and entirely homologous) nutrients.

In the following we will use the terms homologous, heterologous, partially homologous

and entirely homologous, and partially homologous and partially heterologous for nutrient

pairs (Fig. 2.2).

MODELS OF DUAL NUTRIENT LIMITATION

Based on the history of the description of growth limitation, the kinetic and the

stoichiometric aspects led to distinct models to describe “dual nutrient limited growth”. The

two aspects cannot be linked in practice, because kinetics deal with the growth rate as a

Chapter 2 33

function of the nutrient concentration in the culture broth (“growth rate limitation”), whereas

the stoichiometric aspect specifies the relationship between biomass concentration and the

therefore consumed nutrients (“biomass amount limitation”). Thus, we differentiate these

two aspects in this chapter, too: For the dual (kinetic) limitation the expression dual nutrient

controlled growth will be used, whereas the term dual limitation will be continuously used

for the stoichiometric aspect.

Models of the kinetic aspect

Generally, dual nutrient controlled growth can be defined by the single condition that the

growth rate increases with the concentration of either substrate at a constant concentration of

the other substrate.

The first model of dual nutrient control appeared only in the seventies (Megee et al., 1972).

Their kinetic approach was based on the observation that in mixed cultures the nutrients are

often used up completely (Veldkamp and Jannasch, 1972). Bader and coworkers (Bader,

1978; Bader et al., 1975) proposed that even a monoculture may show dual nutrient limited

growth depending on the kind of nutrients available for growth. For heterologous nutrients

the growth rate of the organism is limited by only one nutrient at a time. Thus, the growth

rates were proposed to follow the subsequent equations:

for , and [2.1]

for , [2.2]

where µ is the specific growth rate, µmax the maximum specific growth rate, sn the residual

substrate concentration of substrates 1 (s1) and 2 (s2), and Ks1 and Ks2 are the Monod

constants for substrates s1 and s2, respectively.

Interactive nutrients are those that affect the growth rate of the organism simultaneously.

This is observed when both nutrients are present in less than saturating concentrations. Most

kinetic models and experiments that dealt with this combination of nutrients (Bae and

Rittmann, 1996a; Lee et al., 1984; Mankad and Bungay, 1988; Mankad and Nauman, 1992;

Narang et al., 1997) have focused on batch growth. An often applied approach to describe the

Overview 34

resulting growth rate is that of Megee et al. (1972), who extended the Monod equation with

the second limiting nutrient:

[2.3]

The critical factor that determines whether a dual nutrient controlled growth occurs is

according to the model the residual concentration of the nutrients s1 and s2. The extension of

the Monod equation by an identical term seems not to be correct, when more than two

nutrients are growth limiting. This means that the specific growth rate µ would be reduced

the more nutrients were considered analogous to the equation 2.2, which is not likely to take

place. To date, there is no experimental proof of the model of Megee (1972).

A very interesting report by Rutgers et al. (1990) analysed experimental data of chemostat

cultures of Klebsiella pneumoniae in which the C/N ratio in the feed was changed. They

based their model on the metabolic control theory of Kacser and Burns (1973). In this theory

the flux coefficient of a substrate s is defined by the equation:

, [2.4]

where is the flux control coefficient for the specific growth rate (µ) by the effector

“substrate” (s, residual substrate concentration), and where ss denotes steady state. Rutgers et

al. found that positive flux control coefficients for the nutrients determine the condition for

dual nutrient controlled growth. At the maximum growth rate all substrates are present at

saturating concentrations and control coefficients become zero, which means that the cell has

no flexibility for dual limited growth. In addition, they showed that in continuous culture

under steady-state conditions the residual substrate concentration is a function of the growth

rate and the relative concentrations of the substrates. This approach is very interesting

because it opens a new kinetic alternative to the Monod model. However, the accurate

Chapter 2 35

determination of the steady-state (residual) substrate remains an experimental problem and

therefore, not yet experimentally verified.

Bell et al. (1980), Bley and Babel (1992), and most recently Egli and coworkers

[Lendenmann, 1998 #1553; Lendenmann, 1996 #1564] (only for homologous nutrients)

have further extended these kinetic models for the description of chemostat cultures.

The model of Lendenmann and Egli (1998) is able to describe the residual concentrations

of homologous carbon sources under carbon limitation. The principle equation (eq. 2.5)

clearly indicates that the specific growth rate of the cell is proposed to be the sum of the

growth rates on individual sugars:

, [2.5]

where µmax is the highest of the individual µmax,i values, Ki is the Monod saturation constant

of sugar i, and si is the steady-state substrate concentration of sugar i during mixed substrate

utilization. The equation accurately described the growth rate of a sequence of chemostats of

E. coli ML30 grown at two dilution rates (D = 0.3 and 0.6 h-1) based only on the residual

concentration of up to six sugars. This approach has definitely an advantage over the

extended Monod equations, since it is not restricted to a system of two homologous nutrients.

c 0 ;; 'i:: -:::l z

() 'i:: Q) - So2 E u 0 Cl> ~ I/) .~ ns 0 -"'

Homologous Heterologous

08

µ = canst. µ = canst. µ = canst.

~ FIG. 2.2: Dual nutrient limited growth: nomenclature and theo1y with respect to physiology, kinetics, and stoichiometry

(modified from Bader, 1978; Baltzis and Frederickson, 1988; Tilman, 1980). Top: The filled area represents the nutritional need of a cell, whereas the circular areas represent the nuti·ients 1 and 2. Middle and bottom: Lines represent isolines of the same specific growth rate µ (kinetic aspect) or the same biomass concenti·ation X (stoichiometi·ic aspect).

Chapter 2 37 s: actual nutrient concentration in the culture broth, s0: total nutrient concentration that can be metabolized theoretically.

Overview 38

Several authors have developed cybernetic models which focus on cell metabolism

(Alexander and Ramkrishna, 1991; Baloo and Ramkrishna, 1991a; Baloo and Ramkrishna,

1991b; Straight and Ramkrishna, 1994). This mathematical approach requires a complete

description of enzyme activities and synthesis, and intermediate nutrient concentrations,

which is difficult to obtain experimentally. As a result, these models cannot be verified in

detail due to the insufficient quality of the experimental data.

Models of the stoichiometric aspect

Dual (stoichiometric) limited growth can be defined as the condition for which the biomass

yield changes with the input concentration of either substrate at a constant concentration of

the other substrate.

Studies with chemostat cultures have provided insight into physiological changes within

the dual limited growth regime (Al-Awadhi et al., 1990; Egli, 1995; Egli and Quayle, 1986;

Grätzer-Lampart et al., 1986; Minkevich et al., 1988).

A stoichiometric approach to predict dual nutrient limited growth for chemostat cultures has

been developed by Egli and Quayle (1986). (A similar suggestion was made by Thingstad,

1987). They found that the borders of the limited growth regime at a particular growth rate

can be predicted from the cellular composition (or the substrate growth yield factors) using

two simple equations, given below for the case of carbon and nitrogen as limiting

(heterologous) nutrients:

[2.6]

and

[2.7]

39 Chapter 2

where Cr and Nr are the medium feed concentrations of carbon and nitrogen, respectively,

and D and D are the respective growth yields based on carbon and nitrogen measured

under either C only (C lim) or N only (N lim) limited growth conditions. Equation 2 .6

calculates the lower boundaiy (Fig. 2.3, (Cti'Nr)iower) between carbon and dual (C,N)

limitation, whereas the equation 2 .7 gives the corresponding value for the bounda1y between

dual (C,N) limitation and the nitrogen limited growth regime ((Cti'Nr)upper).

Measurements of chemostat cultures of Hyphomicrobium ZV620 showed that the

predicted and the measured dual (C,N) limited growth regimes were identical

(Gratzer-Lampaii et al. 1986).

µ=D

Nitrogen limitation

Carbon limitation

FIG. 2.3: The simultaneous limitation of carbon (C) and nitrogen (N) is dependent on the medium composition (nutrient ratio, Cti'Nr) and the dilution rate (D) of the chemostat (stoichiometi·ic aspect) (adapted from Egli, 1991). The borders of the dual (C,N) limited growth regime can be predicted based on the equations 2 .6 and 2 . 7 . ( Cti'N r)1ower: boundaiy between C and dual ( C,N) limited growth regime; (Cti'Nr)upper: boundaiy between dual (C,N) and N lirnited gi-owth regime.

Overview 40

Moreover, Egli (1991) predicted that the dual nutrient limited growth regime is a function of

the specific growth rate and the Cf/Nf ratio. Egli called this growth regime the dual nutrient

limited zone, indicating that this growth regime can be represented graphically in a D vs

Cf/Nf diagram (Fig. 2.3). The zone has a “banana” shape: the dual nutrient limited growth

regime is broader at lower growth rates, becomes more narrow at higher growth rates and is

not detectable at the maximum growth rate. The explanation of the author is that the growth

yields do not differ greatly between C and N limitation at high growth rates, since the cells do

not have the plasticity to adapt to the nutrient limitation like it is the case under low growth

rates. The location and the shape of the “banana” are determined by the kind of substrate

limitation and by the substrates themselves. For instance, the zone for a dual (C,N) limitation

with methane is not as broad as with formate. In addition, the zone is shifted to higher Cf/Nf

ratios in the case of formate. The explanation for this observation is that methane is less

oxidized than formate, so that less methane than formate is used to produce a given amount

of energy (Fig. 2.4).

µ = D

Cf/Nf

Met

hane

Gly

cero

l

Citr

ate

Form

ate

relativeµmax

FIG. 2.4: The redox state of the carbon substrate determines the shape and position of the dual (C,N) limited growth zone (adapted and changed from Egli, 1991).

Chapter 2 41

The two aspects (kinetic and stoichiometric) can be graphically displayed (Fig. 2.2). The

diagrams represent the nutrition and the corresponding effect on the specific growth rate

(kinetic aspect) and on the amount of biomass that can be produced (stoichiometric aspect).

Note that the stoichiometric diagrams look different for each growth rate, as the influence of

the cell maintenance and the viability are a function of the growth rate.

Below, we discuss the literature on multiple limited growth with respect to the four

defined classes described in nomenclature.

Overview 42

EXPERIMENTAL DATA

Homologous nutrients

As mentioned above, homologous nutrients can replace each other completely. Differences

between such substrates relate to the maximum growth rate, the uptake rate, the growth

yields and inhibition effects associated with individual substrates.

Comprehensive surveys on the literature on homologous nutrients have been published by

Harder (Harder and Dijkhuizen, 1976; Harder and Dijkhuizen, 1983), Babel et al. (1993), and

recently Egli (1995). Thus, only a few representative cases are mentioned below.

Most reports on homologous nutrients deal with carbon sources (“mixed substrates”, Egli,

1995), although many other cases could be imagined, e.g. serine and methionine as only

sulfur sources. A typical and interesting observation for experiments with carbonaceous

nutrients is that they may be consumed in batch cultures in two time separated growth phases,

a common phenomenon described in literature as diauxic growth. For instance,

methylotrophic yeasts (Hansenula polymorpha and Candida boidinii ) show a diauxic

growth pattern when grown in batch culture on a mixture of glucose and methanol (Sahm and

Wagner, 1973; Sakai et al., 1987). However, when presented as mixtures, e.g.

glycerol/methanol (Wanner and Egli, 1990) or xylose/methanol (Brinkmann and Babel, 1992;

Volfová et al., 1988), both carbon sources were consumed simultaneously.

Kinetic aspect of homologous nutrients

Diauxic growth was generally observed when both nutrients were greatly in excess

(laboratory conditions, substrates in g l-1 range). When the initial concentration of the

nutrients in batch cultures are significantly lower (mg l-1 range), a different growth pattern

was found. This could be shown specifically for the case of E. coli growing with glucose and

galactose, where the initial concentration was reduced to 5 mg l-1 or lower (Lendenmann,

1994; Senn, 1989). Both sugars were consumed simultaneously.

A good example of the kinetic aspect for the description of homologous nutrient limitation

is certainly the work of Lendenmann et al. (1996). They showed that Escherichia coli used 6

different sugars at the same time under carbon limited growth conditions in the chemostat.

Chapter 2 43

Interestingly, the residual concentrations of all 6 sugars were reduced compared to the

residual concentration of the corresponding single sugar systems.

TAB. 2.1: Multiple substrate systems reported in literature.

Nutrition Organism Conditions Substrates Remarks Reference

Homologous Acinetobacter calcoaceticus

Chemostat Ac, Xyl Additive growth yields

(van Schie et al., 1993)

Homologous Candida boidinii Batch Xyl, MeOH (Volfová, Korínek and Kyslíková, 1988)

Homologous Clostridium thermohydrosulfu-ricum

Batch, chemostat Glc, Cel In chemostat cultures with D < 0.2 h-1 both sugars are used up completely

(Slaff and Humphrey, 1986)

Homologous Escherichia coli Chemostat Glc, Gal, Rib, Fru, Mal, Ara

Additive growth yields

(Lendenmann et al., 1996)

Homologous E. coli Chemostat Gal, Glc Feed concentrations < 5 mg l-1

(Lendenmann, 1994; Senn, 1989)

Homologous Hansenula polymorpha

Batch Gly, MeOH (Wanner and Egli, 1990)

Homologous H. polymorpha Batch MeOH, Gly Preculture was grown on glycerol only

(Wanner and Egli, 1990)

Homologous H. polymorpha Batch Xyl, MeOH (Brinkmann and Babel, 1992)

Homologous H. polymorpha, Kloeckera sp. 2201 (=C. boidinii)

Chemostat MeOH, Glc MeOH is used at higher D than growth on MeOH

(Egli et al., 1986; Egli et al., 1982a)

only

Table 2.1 continued


Homologous Kloeckera sp. 2201 (=C. boidinii)

Chemostat MeOH, Glc Additive growth yields

(Egli et al., 1982b)

Homologous Kloeckera sp. 2201 (=C. boidinii)

Chemostat MeOH, Glc Residual MeOH concentration is lower under DL compared to MeOH only limitation

(Egli et al., 1983)

Homologous Pseudomonas oxaliticus

Chemostat Ac, Ox No Ox utilised at D > 0.4 h-1

(Harder and Dijkhuizen, 1982)

Homologous Saccharomyces cerevisiae

Chemostat Glc, EtOH Additive growth yields

(de Jong-Gubbels et al., 1995)

Entirely homologous and partially homologous

mouse-mouse Hybridoma

Fed-batch Glc, Gln Slightly increased IgG production and end biomass

(Ljunggren and Häggström, 1994)


Hybridoma Fed-batch Glc, Gln Reduced lactate production due to dual control of Gln and Glc

(Yoshida et al., 1994)


Streptococcus mutans, Streptococcus milleri (mixed)

Chemostat Arg, Glc Mixed culture in stable proportions

(Rogers et al., 1987)

Heterologous Acinetobacter johnsonii

Biofilm Phe, NO3- Radioactive labelled

phenol as substrate, detection of labelled

(Hoyle et al., 1995)

CO2

Table 2.1 continued


Heterologous Aspergillus foetidus Fed-batch N, P Citric acid pro-duction is enhanced

(Chen, 1993)

Heterologous Aspergillus niger Fed-batch N, P Final citric acid concentration was increased to 95 g l-1

(Dawson et al., 1989)

Heterologous Bacillus sp. NCIB 12522

Chemostat MeOH, N DL between C/N = 4.8 and 8 g g-1 at D = 0.362 h-1

(Al-Awadhi et al., 1990)

Heterologous Brevibacterium glutamicum

Chemostat Arg, P Increased stability of Arg auxotrophic mutant in chemostat culture (only 0.1% revertants compared to 40% under Arg only lim.), ornithine production increased

(Choi et al., 1996)

Heterologous Candida valida Chemostat EtOH, N Discontinuous medium gradient with change of N concentration in the feed medium

(Minkevich et al., 1988)

Heterologous Enterobacter Chemostat (Glc,N); (Glc,P); Physiology of the (Cooney and Wang,

aerogenes (=Klebsiella pneumoniae)

(Glc,P) cell is dependent on nutrient combination and growth rate

1976b; Cooney et al., 1976a)

Heterologous E. coli Chemostat Glc, N Problems with the detection of the residual nutrient conc., kin. model

(Lee et al., 1984)

Table 2.1 continued


Heterologous Gracilaria (sea algae)

in situ cage techniques

N, P Natural habitat is limited first by P and second by N, G. stores N intracell.

(Lapointe, 1985; Lapointe and Ryther, 1979)

Heterologous H. polymorpha, Kloeckera sp. 201 (=C. boidinii)

Continuous culture with medium shift

MeOH, Glc, N Shift from C limi-tation to N limitation went through dual (C,N) limited transition zone

(Egli, 1982)

Heterologous H. polymorpha Chemostat C, N

(C = 87.7% Glc and 12.2% MeOH, w/w)

Dual limited growth between C/N=12 and 31 g g-1 at D = 0.1 h-1

(Egli and Quayle, 1986)

Heterologous Hyphomicrobium ZV620

Chemostat MeOH, N Increase of intracellular PHB content under dual limitation

(Grätzer-Lampart et al., 1986)

Heterologous Hyphomicrobium X Chemostat MeOH, N (Duchars and Attwood, 1989)

Heterologous Klebsiella pneumoniae

Chemostat Glc, N Experimental data used for flux theory

(Rutgers et al., 1990)

Heterologous Kluyveromyces fragilis

Chemostat Glc, N Increased heat production

(Cooney et al., 1996)

Heterologous Pseudomonas putida

Chemostat O, Ac NAD+/NADH is O dependent, whereas

(Bae and Rittmann, 1996a; Bae and

ATP/ADP*Pi is Ac influenced; model

Rittmann, 1996b)

Table 2.1 continued


Heterologous Saccharomyces cerevisiae, Candida utilis

Chemostat Glc, O (Weusthuis et al., 1994)

Heterologous S. cerevisiae Chemostat Glc, N Separation of anabolic and catabolic substrate consumption

(Larsson et al., 1993)

Heterologous S. mutans, Actinomyces viscosus (mixed)

Chemostat O, Glc Double Monod model

(van der Hoeven and Gottschal, 1989)

Heterologous Thalassiosira weissflogii (diatom)

Batch N, Si Cell cycle stopped in G1-phase, only addition of both nutrients enabled cell division

(Vaulot et al., 1987)

Heterologous Thiobacillus A2 Chemostat Thiosulfate, Glc No repression of RuBP carboxylase

(Smith et al., 1980)

Heterologous Thiobacillus A2 Chemostat Thiosulfate, Ac 30% increased cell dry weight

(Gottschal and Kuenen, 1980)

Heterologous Isolate from deep subsurface

Glass-bead-packed columns

O, intracellular metabolites of quinoline degradation

Application of dual limitation kinetics on intracellular metabolites

(Malmstead et al., 1995)

Abbreviations: Ac: acetate, Ara: arabinose, Arg: arginine, C: carbon, Cel: cellobiose, EtOH: ethanol, Fru: fructose, Gal: galactose, Gln: glutamine, Glc: glucose, Gly: glycerol, K: potassium, Mal: maltose, MeOH: methanol, N: ammonium, O: oxygen, Ox: oxalacetate, P: phosphorus, Phe: phenol, Rib: ribulose, Si: silicone.

Chapter 2 53

In a simpler system with only two homologous sugars, galactose and glucose, the same

authors could show that the residual concentration reflected the mixture of the sugars in the

medium and were homologous (perfectly substitutable) from a dilution rate of D = 0.2 h-1 up

to at least D = 0.45 h-1 (Lendenmann, 1994).

Stoichiometric aspect of homologous nutrients

The stoichiometric aspect is demonstrated in the publication of Egli et al. (1986) dealing with

the methylotrophic yeast Hansenula polymorpha. They cultured the yeast in chemostat with

different methanol/ glucose mixtures at dilution rates between 0.025 and 0.51 h-1. They were

able to show that H. polymorpha utilized both carbon sources simultaneously at low dilution

rates, whereas at higher growth rates only glucose was consumed. Interestingly, methanol

was even used as a second C source at higher dilution rates than when H. polymorpha was

grown on methanol only.

Van Schie et al. (1993) observed that the yields for Acetobacter calcoaceticus were

increased significantly in chemostat cultures, when acetate and xylose were supplied under

carbon limiting growth conditions. The acetate concentration of the feed was fixed to 30 mM,

whereas the xylose concentration was discontinuously increased. As a result the biomass

increased up to a xylose concentration of 25 mM and remained constant after that.

Heterologous nutrients

There is a wealth of reports on dual limitation with heterologous nutrients, either based on

kinetic or stoichiometric limitation. The variety of studies is large, ranging from the study of

biofilms (Hoyle et al., 1995; Malmstead et al., 1995; Rittmann and Dovantzis, 1983), mixed

cultures (van der Hoeven and Gottschal, 1989), and even studies in the natural habitat

(Lapointe, 1985; Lapointe and Ryther, 1979). Due to the large number of reports, we restrict

this discussion to well defined systems only, emphasising that the chemostat is certainly the

best experimental tool to reproducibly study dual nutrient limited growth.

Kinetic aspect of heterologous nutrient limitation

Overview 54

The reports investigating the kinetic aspect suffer heavily under the difficulty to measure the

residual substrate concentrations accurately enough. This difficulty consisted of two

experimental problems. First, the resolution of the detection system was in most cases not

well enough and could barely cover the concentration for the Monod saturation constant Ks.

(µg l-1 range, e.g. Lee et al., 1984). Second, the sampling of the culture had to be performed in

a quick and efficient way. Investigations of Senn et al. (1994) could clearly show that in a

sample of an E. coli chemostat (D = 0.88 h-1) the residual glucose concentration decreased

very fast (t1/2 < 0.5 min) when the cells were not separated from the culture broth

immediately.

Probably the best study on the kinetic aspect of heterologous nutrients to date was

presented by Rutgers et al. (1990). They investigated the dual (glucose, ammonium) limited

growth of Klebsiella pneumoniae under chemostat conditions (D = 0.2, 0.4, and 0.6 h-1).

They used low substrate feed concentrations to be able to measure the residual nutrient

concentration most accurately. Based on their results a flux model was developed that

included flux control coefficients (see theory of dual limited growth). Thus, they were able to

show that the control of growth gradually shifted from glucose to ammonia within the dual

(C,N) limited zone and did not change in a step function (Fig. 2.5).

Stoichiometric aspect of heterologous nutrient limitation

The chemostat is generally accepted as a very accurate tool to study the stoichiometric

relationship between nutrient input into a system and the produced biomass. The results

gained by this method revealed that the dual nutrient limited growth regime must be detected

for microorganisms that show a variation of the biomass yields under different heterologous

nutrient limitations (e.g. YX/C under C and N limitation, see also section models). This

variation may be caused by one of the following physiological changes:

• Variation of the basic cellular composition (e.g. RNA, Protein)

• Accumulation of a storage compound (e.g. polyphosphate, lipids, polyglucans,

polyhydroxybutyrate)

55 Chapter 2

• Change of physiological efficiency (e.g. biomass/ ATP consumed and biomass/C0 2

produced)

• Excretion of intermediate metabolic intennediates

In the following samples of such observations are presented.

A first stoichiometi·ic approach was explored by Cooney and coworkers (Cooney and

Wang, 1976b; Cooney et al., 1976a) with Enterobacter aerogenes. They showed that this

organism is able to grow under simultaneous limitations of (Gluc,N), (Gluc,P), or (N,P).

Depending on the kind of limitation, they could detect changes in the cell composition, such

as polysaccharide content, protein to ribonucleic acid ratio, and metabolite excretion. All

parameters were also influenced by the dilution rate.

100000

10000 ~

~ 1000 3 c .Q 100 -ro ..... -c 10 Q) () c • 0 () 1 E ::I c @ 0 E E 10000 ro ..... 0 1000 Q) (/) 0 () 100 ::I

C)

10

1

0.1 0.01 0.1 1 10 100

Molar ratio of glucose to ammonium in the feed medium

Overview 56

FIG. 2.5: Influence of the molar ratio of glucose to ammonilllll in the feed medilllll on the steady-state concentrations of glucose (II) and ammonium (0 ) in chemostat cultures of Klebsiella pneumoniae. The shaded area indicates dual (C,N) contrnlled growth regime. Panel a: dilution rate ~ 0.2 h-1

. Panel b: dilution rate ~ 0.4 h-1. Adapted from

Rutgers et al. 1990.

Gratzer-Lampaii et al. (1986) investigated the physiology of Hyphomicrobium ZV620 in

chemostat experiments. They increased the methanol to ammonium ratio in the feed medium

discontinuously at a constant dilution rate of 0.054 h-1. They observed a dual (C,N) limited

growth regime for 7.1 g g-1 < Ct/Nr < 12.6 g g-1 (Fig. 2 .6) and showed that the ammonium

assimilating enzymes glutamate dehydrogenase, glutamine synthetase, and glutamate

synthase changed specific activities within the dual lirnited growth regime. The NADP+

glutamate dehydrogenase was repressed under dual (C,N) and under N only limitation.

Duchars and Attwood (1989) confnmed, as previously found by Gratzer-Lampaii and

coworkers, that Hyphomicrobium X accumulates polyhydroxybutyrate (PHB) under dual

(C,N) and N lirnitation. Within the dual (C,N) lirnited growth regime the PHB content

showed a linear dependence on the Ct/Nrratio.

57 Chapter 2

~0. 1 8 2.7 C> ~ / 4.2 ";"

'E0.15 \ .2? ~2.1

:::J ® ~ 0 I

C> § 0.12 \ I 3.0~

i 1.5 E \ ~ E 0.09 \ 6 cu 1.8 ~ ~ ~ 0.06 o 0.9 I "O ·w "O Q)

0.3 -~ 0.03 0.6Ck:'. Ck:: - 1 OO;;o .... ;:' -3: .... 3:

~24 3: 3: >. >. 50 >. 80

..... ..... 30 ..... "O ~

"O "O ~ 0 ~ 40 • ~ 0

:::- 20 0 ::...... 60 -- .... c .... .... c Q) c 30 20 c Q)

c 16 Q) Q) .... .... .... 40 c 0 c c 0 (.) 0 20 O-c- 0 (.)

ijj 12 (.)

10 (.) c c aJ 20 .Qi

C> 0 10 I .... 0 .0 o--o a.. 0 ..... ..... .....

:!:: 8 cu 0 0 0 a.. z u 3 6 9 12 15 18 21 24

CtfNt [g g-1 I

FIG. 2.6: Growth of Hyphomicrobium ZV620 with methanol and ammonium in the chemostat at a constant dilution rate of D = 0.054 h·1, as a function of the carbon to nitrogen ratio in the feed medium (Cfi'Nc). Panel a: D1y weight of biomass produced (D); Residual concentrations of methanol (0 ) and ammonium (• ). Panel b: Cellular content (percentage diy weight) of carbon (• ), nitrogen (Cl), protein (0 ) and poly-B-hydi·oxyalkanoate (PHB) (II). The shaded area indicates the dual (C,N) limited growth regime. Adapted from Gratzer et al. , 1986.

Egli and Quayle (1986) demonstrated that the methylotrophic yeast Hansenula

polymorpha is able to grow under multiple limitations on a inixture of methanol-glucose and

ammonium. They found in chemostat experiments at a dilution rate of 0.1 h-1 that the cells

were C liinited below C/N = 12 g g·1. Both C sources were used up completely, whereas for

CIN ratios between 12 and 31 g g·1 the cells were dual (C,N) liinited. Within this growth

regime the cells showed different cellular and enzyinic composition, such as the repression of

the synthesis of methanol assiinilating and dissiinilating enzymes. Interestingly, highest

Overview 58

activity of the glutamate hydrogenase was found under dual (C,N) limited growth

(considerably higher than under C limitation, about 2 x higher than under N-limitation). This

is the only report where a transition from homologous (methanol and glucose) to

heterologous (glucose and ammonium) nutrient limitation was investigated.

A great versatility to adapt to unfortunate growth conditions is a trait of the bacterium

Thiobacillus A2. This organism is very adaptable because of its ability to grow

chemolithoautotrophically on thiosulfate and CO2, heterotrophically on glucose, or

mixotrophically with dual limitation by thiosulfate and glucose (Smith et al., 1980). Under

dual limited growth in a chemostat at a dilution rate of 0.08 h-1, the cells were able to use

glucose and thiosulfate as energy sources, and CO2 and glucose as carbon sources,

respectively. Under glucose only limitation, CO2 fixation via the Calvin cycle was not

detectable due to the repression of the ribulose bisphosphate (RuBP) carboxylase. For dual

(glucose,thiosulfate) limited growth this repression was relaxed. As a consequence,

enhanced activity of the (RuBP) was measured and carbon was obtained via both the Calvin

cycle and the assimilation of glucose. A second group, represented by Gottschal and Kuenen

(1980), investigated the mixotrophic growth of Thiobacillus A2 on acetate and thiosulfate.

They found that acetate has a repressing effect on the thiosulfate-oxidising capacity in batch

cultures. In chemostat cultures at a dilution rate of 0.05 h-1 a dual limited growth regime was

observed. They observed a tight enzyme regulation which was balanced by the induction by

thiosulfate and repression by acetate. Thus, growth on mixtures of acetate and thiosulfate

revealed up to 30% more cell dry weight than predicted based on the growth yields of

comparable amounts of thiosulfate or acetate under separate single limitation.

Larsson et al. (1993) cultured S. cerevisiae under carbon, nitrogen, or simultaneous

nitrogen and carbon limitation, in a chemostat. The carbon source (glucose) was kept

constant. The nitrogen source (ammonium) and the dilution rate were varied at the same time.

They found that the critical dilution rate was dependent on the type of limitation: on glucose

limited growth the critical dilution rate was 0.2 h-1, whereas under ammonium limitation it

was 0.1 h-1. They concluded that the nitrogen concentration has an influence on the specific

oxygen consumption rate (qO2): A decrease of qO2 was observed when the nitrogen

concentration was decreased. At a dilution rate of 0.07 h-1 the specific oxygen consumption

rate was more than doubled under dual (C,N) limitation compared to C only limited growth.

Chapter 2 59

Their conclusion was that S. cerevisiae is able to separate anabolic biomass formation from

catabolic energy substrate consumption. The same research group compared the results of S.

cerevisiae with those of Kluyveromyces fragilis (Cooney et al., 1996) obtained under similar

growth conditions. K. fragilis used glucose mainly in a respirative way in contrast to S.

cerevisiae. The equipment they used (Marison and von Stockar, 1987) enabled a detailed

calorimetric investigation of chemostat cultures. For dual (C,N) limited growth they

observed increased consumption of glucose and a concomitant increase in heat production.

They concluded that the anabolic formation of biomass was uncoupled from the substrate

consumption. This overflow metabolism might be observed for other microorganisms, too.

Dual nutrient limited growth is strongly dependent on the physiology of the particular

microbial species. Weusthuis et al. (1994) showed that Saccharomyces cerevisiae and

Candida utilis can grow under simultaneous limitations on glucose and oxygen in chemostat

(D = 0.1 h-1). Depending on the extent of the oxygen supply, the cells grew respiratorily with

simultaneous alcoholic fermentation. With maltose as C substrate only, S. cerevisiae

produced ethanol as an overflow metabolite. Under the same growth condition C. utilis grew

respiratorily only (no alcohol production), although the fermentative enzymes were present

at high levels. The amount of converted maltose correlated with the oxygen feed, indicating

that C. utilis growth can be dual (maltose,O2) limited.

A very successful experimental application of dual substrate limited growth was reported

by Choi et al. (1996). They produced L-ornithine by an arginine auxotrophic mutant of

Brevibacterium ketoglutamicum in a chemostat at a dilution rate of 0.1 h-1. This mutant was

not stable in a culture where only arginine limited growth. After 90 hours, the productivity of

L-ornithine synthesis decreased due to the fact that only 40% of the cells were arginine

auxotrophs. The revertants grew considerably faster than the production strain and were

therefore able to outcompete the original strain. As a stabilising effect, a second limitation by

phosphorus was imposed over the arginine limitation. The chemostat culture could be kept

stable over 250 hours with a phosphate/arginine ratio of 2.01 g g-1 in the feed medium. Thus,

the revertant fraction remained below 0.1% of total cells. The authors concluded that this

method is also applicable to the large scale production of amino acids. Certainly this idea

seems to be applicable for the stabilisation of cultures of genetically engineered organisms.

Overview 60

Entirely homologous and partially homologous nutrients

There are only a few publications regarding growth on this combination of substrates.

However, one can speculate that this growth type is not very rare and can be found with all

auxotrophic organisms. Interestingly, all reports that could be found to date explained their

findings with the kinetic aspect of dual nutrient controlled growth.

Heterotrophic and auxotrophic organisms are found in the dental plaque of humans. Two

representatives of this complex community are Streptococcus mutans and Streptococcus

milleri. Rogers et al. (1987) determined the growth of both organisms in chemostat cultures

at D = 0.1 h-1 and determined a strict arginine auxotrophy only for S. milleri. Based on a

kinetic aspect, the authors predicted a stable mixed culture of S. mutans and S. milleri under

conditions where glucose and arginine are limiting at the same time. They determined

experimentally that an arginine concentration in the feed medium lower than 50 µM and a

glucose concentration of 10 mM resulted in a stable community. Disturbances of the nutrient

concentrations caused by pulses of arginine, arginine and glucose, and glucose resulted in a

change of the community composition. After consumption of such substrate pulses, the

mixed culture went back to the composition observed prior to the substrate pulses. Thus, it

can be assumed that mixed cultures are only stable in continuous culture when they are

limited by multiple nutrients at the same time. Consequently, the chance that one organism

can outgrow the other one seems to be low.

Dual controlled limited growth can have a beneficial effect on eucaryotic cultures, too.

Ljunggren and Häggström (1994) reported that overflow metabolism was important in

fed-batch cultures of mouse-mouse Hybridoma cells. On the basis of the kinetic aspect, the

authors postulated a beneficial effect on antibody production under nutrient limited growth.

They showed that in fed-batch cultures, where only glucose limited growth and all other

nutrients were in excess, lactate formation was significantly reduced, whereas at the same

time, the glutamine consumption and the ammonium production were increased. When

glutamine was the sole limiting substrate and all other nutrients were in excess, the

ammonium and alanine formation decreased and concomitantly, the glucose consumption

and lactate formation decreased. A dual (glucose,glutamine) limitation led to a negligible

formation of lactate, alanine, ammonium, and a higher antibody production (total amount of

Chapter 2 61

antibody produced: glucose lim.: 40.2 mg; glutamine lim.: 64.3 mg; glucose and glutamine

lim.: 92.2 mg). The authors explained their findings by the reduced uptake of glucose and

glutamine, which had an influence on the pyruvate pool and on restriction of the flux through

glutaminase and lactate dehydrogenase. Thus, glutamine entered the tricarboxylic acid cycle

through the glutamate dehydrogenase pathway, which releases more energy than the

transamination route. This could be confirmed by the slightly increased cell densities and

total antibody production. The authors noted that the productivity could be enhanced even

further when an exponential instead of a linear feed was applied. This had the advantage that

the cells grew at a constant growth rate rather than a steadily decreasing µ.

The constant growth rate is certainly a challenging requirement for dual limited fed-batch

cultures. Based on the stoichiometric aspect, the specific growth rate has a significant

influence on the physiological activity, and as a consequence the productivity of

side-products is influenced, too. At the worst, a decreasing growth rate, which occurs in

fed-batch cultures with linear feed, may result in changing nutrient limitations, in general an

energy limitation (compare Fig. 2.3).

Partially homologous and partially heterologous nutrients

Unfortunately, no example for this special growth condition could be found in the literature.

We are sure that this growth type exists, however, it is difficult to detect and has therefore

apparently not yet been cited in the literature.

An example for such a system would be a diauxotrophic bacterium that needs two amino

acids to grow. Thus, both amino acids are essential and cannot be replaced by one another.

However, under nitrogen limiting growth conditions, both nutrients can additionally be used

as nitrogen sources.

63

CHAPTER 3

DYNAMICS OF GROWTH AND ACCUMULATION OF

POLY(3-HYDROXYALKANOATE) (PHA) IN

PSEUDOMONAS OLEOVORANS CULTIVATED IN

CONTINUOUS CULTURE UNDER TRANSIENT FEED

CONDITIONS: RAPID IDENTIFICATION OF DUAL

NUTRIENT (C,N) LIMITED GROWTH ZONES

Manfred Zinn, Roland Durner, Hanspeter Zinn, Thomas Egli, and Bernard Witholt

Dynamics of growth 64

Keywords: Pseudomonas oleovorans, chemostat, continuous culture, dual nutrient limited

growth, medium feed gradient, medium chain length poly(3-hydroxyalkanoate), growth

inhibition, octanoic acid.

65

SUMMARY

Pseudomonas oleovorans can grow in continuous culture simultaneously limited by

ammonium and octanoate, because of its ability to store carbon intracellularly as

poly(3-hydroxyalkanoate) (PHA). The boundaries for limitation of growth by the two

nutrients carbon and nitrogen were determined from analysis of steady-state chemostat

cultures (dilution rate = 0.3 h-1) by changing the carbon to nitrogen ratio in the feed medium

(Cf/Nf) stepwise (discontinuous medium gradient). At Cf/Nf ≤ 6.4 g g-1 growth was purely C

limited, when Cf/Nf was 9.5 g g-1 it was purely N limited. When Cf/Nf was between 6.4 and

9.5 g g-1, growth was dual (C,N) nutrient limited. To investigate whether this dual nutrient

limited zone is also observed under transient growth conditions a dynamic approach was

taken, where in the experiments the Cf/Nf ratio was changed continuously through a convex

increase of Cf, a concave increase of Nf, or a linear decrease of Cf (gradients 1, 2, and 3). In

these experiments the dual (C,N) limited growth regime was between 7.2 and 11.0 g g-1 for

gradient 1, 4.3 and 6.9 g g-1 for gradient 2, and 5.1 and 8.9 g g-1 for gradient 3. A

mathematical equation was developed that compensates a time delay of the gradient that is

caused by the wash-in effect of the medium feed. Taking this into account the boundaries

observed in dynamic experiments coincided considerably better with those obtained under

steady-state conditions. The best correction could be done for gradient 3, where correction

shifted the dual limited growth regime to C/N ratios between 6.5 and 9.6 g g-1. This gradient

was done over a longer period and was consequently less steep than gradients 1 and 2. It was

concluded that the physiological adaptation to the feed medium gradient occurred in different

ways depending on the design of gradient (nutrient increase or decrease and physiological

function of the nutrient).

INTRODUCTION

Biotechnological processes are becoming increasingly important in the production of fine

chemicals. Therefore, there is an interest that cultivation of cells takes place in defined state


that ensures reproducible and highest productivity. The composition of growth media is a

very important factor for costs and productivity. In a recent cost analysis for bulk chemical

processes it was reported that 70% of the total production costs can be attributed to the media

(Hepner, 1996). Hence, the strategy must be to make best use of medium components

because large amounts of unused nutrients are costly. Generally, in such designed media, one

nutrient is limiting cell growth and all others are supplied in excess. Usually, the choice of

nutrient limitation depends on the product to be produced and selectively applying suitable

nutrient limitation and forcing the microbial cell into a physiological state particularly

suitable for production (Herbert, 1976). Recently, it has been shown that simultaneous

limitation for two nutrients can change metabolic activities (Duchars and Attwood, 1989;

Egli and Quayle, 1986; Hueting and Tempest, 1979; Larsson et al., 1993) and that such

conditions even enhance productivities or activities (Chen, 1993; Egli and Quayle, 1986;

Hoyle et al., 1995; Lucca et al., 1991).

A good model organism to study simultaneous carbon-nitrogen limited growth is the soil

bacterium Pseudomonas oleovorans. This microorganism accumulates the storage polymer

poly-3(R)-hydroxyalkanoate (PHA) under nitrogen, phosphorus, magnesium, or sulfur

nutrient limitation, when cultivated with alkanes, alkanols or alkanoic acids of medium chain

length (Lageveen et al., 1988). This PHA is a biodegradable polyester that may be used as an

alternate material to the traditional plastics (for a survey see Lee, 1996 or Van der Leij and

Witholt, 1995).

Traditionally, to define dual nutrient limited growth regimes, steady-state chemostat

cultures are analysed that are run at a given dilution rate, with a medium feed containing

carbon and nitrogen at distinct Cf/Nf ratios between 2 to 20 g g-1 (Durner, 1998)

(discontinuous method). At low Cf/Nf ratios, carbon will be limiting, while at high Cf/Nf

ratios, nitrogen will be the limiting nutrient. Between these extremes, a Cf/Nf range can

generally be identified in which both C and N are limiting (Egli, 1991).

Although this approach allows identification of dual limitation growth regimes, it is quite

time consuming, particularly at low dilution rates. We have therefore developed a more rapid

alternative, using a feed strategy based on continuous gradients in the C/N ratio in the feed

medium (Babel et al., 1983; Domínguez et al., 1992; Müller et al., 1985; Pagni et al., 1995).

Using this approach transitions between pure carbon and pure nitrogen limitation, or vice

67

versa were imposed on a continuous culture, and the behaviour of the culture was observed.

Here we describe the influence of different nutrient gradients of the medium and their effects

on cell physiology and PHA production.


MATERIAL AND METHODS

Microorganism

Pseudomonas oleovorans (ATCC 29347) was used throughout all experiments. For the

production of frozen stock cultures an exponential culture with medium X was mixed with

30% glycerol (1:1 v/v) and stored in 1 ml portions at -80°C.

Culture media

Precultures

The minimal growth medium X for precultures contained E2-salts, MT micro elements

(Lageveen et al., 1988), and citric acid (in grams per liter): NaNH4HPO4*4H2O, 3.5;

K2HPO4*3H2O, 7.5; KH2PO4, 3.7; citric acid monohydrate, 2.0. After sterilization the

following nutrients were added to 1 l medium: 1 ml of 1 M MgSO4*7H2O, 1 ml of 0.01 M

FeSO4*7H2O in 1 N HCl and 1 ml of MT stock solution that contained (in grams per liter 1 N

HCl): MnCl2*4H2O, 1.98; CoSO4* 7H2O, 2.81; CaCl2*2H2O, 1.47; CuCl2*2H2O, 0.17;

ZnSO4*7H2O, 0.29. Medium X is carbon limited.

Continuous cultivation

For continuous cultivation we used medium XCC, which contained (in grams per liter):

(NH4)2SO4, 0.71; KH2PO4, 1.0; and different concentrations of Na-octanoate. After

sterilization we added per liter: 1 ml 1M MgSO4*7H2O; 1 ml of 0.01 M FeSO4*7H2O in 1 N

HCl and 1 ml of MT stock solution.

Differently nutrient limited media (limited by nitrogen, carbon or both) were obtained by

varying either the octanoate or the ammonium concentration of the medium (Durner, 1998).

The media XCC1 and XCC2 refer to medium XCC in vessel V1 and V2, respectively (Fig.

3.1).

Culture conditions

69

Pseudomonas oleovorans was cultivated in a 3 l continuously stirred tank bioreactor

(Wubbolts et al., 1996) with a working volume of 1.5 l. The aeration was adjusted to 0.6 vvm.

The temperature was set to 30°C and pH was held constant at 7.0 ± 0.02 by automatic

addition of either 4 N H2SO4 or 2 N NaOH.


Steady-state and transient continuous culture experiments

The chemostat (steady-state) approach to determine dual nutrient limited growth was

compared with the results obtained by transient experiments. For the chemostat study only

the medium of vessel VI (Fig. 3. I) was used to feed the culture. After each steady-state

measurement the vessel VI was exchanged by a new one that contained medium of a

different CtlNr ratio. The CtlNr ratio during this experiment was therefore changed stepwise

(Fig. 3.2).

Pump 2

V1 ~ S1

Medium XCC2 Medium XCC1 ((C/N)0 = 20) ((C/N)0 = 1)

Feed gradient maker

Base

S2

Waste

N H4 + Analyser

0

on-line GC (Na-octanoate)

FIG. 3.1: Bioreactor set-up to produce medium feed gradients. Chemostat conditions were first established by feeding the culture only from vessel VI that contained medium XCCI with the initial C/N. After steady-state growth was achieved, a gradient feed was initiated by switching on pump 2. The medium gradient depended on the flow rates of pumps I and 2, and the nutrient concentrations of the media in vessels VI and V2. The actual medium gradient was determined by analysis of samples taken from the sample po1i SI. The residual carbon and nitrogen concentrations in the bioreactor were measured on-line by a gas chromatograph (Na-octanoate) and an enzymatic analyzer (ammonium). Biomass and its composition were determined off-line.

71

S1: Sample port to determine the C and N concentration in the feed medium; S2: Sample port of culture broth; V1, V2: Medium vessels 1 and 2; R1: Bioreactor with the working volume VR.


For transient experiments, steady-state conditions were established first, after which a

C/N gradient was established in the medium feed by addition of XCC2 with a different C/N

ratio from vessel V2 into vessel V1 (Fig. 3.1). Because XCC1 and XCC2 differed only with

respect to the carbon or nitrogen concentration, it was possible to obtain a continuous

increase or decrease in the C/N ratio of the bioreactor feed medium (Fig. 3.2).

On-line analysis of octanoic acid

Cell-free sampling

A cross-flow-filter (Bioengineering, Wald, Switzerland) was used to produce a cell-free

sample flow for on-line analysis. Teflon membranes (Tech-Sep) with a pore size of 0.2 µm

were used.

On-line GC

The concentration of octanoate in the culture broth was measured with an on-line gas

chromatograph (Biospectra, Schlieren, Switzerland) equipped with a flame ionisation

detector and a fused silica capillary column (Permabond-FFAP-0.35, 25 m x 0.32 mm ID,

Machery-Nagel, Germany). Every 20 minutes a cell-free sample from the cross flow filter

was diluted 1:2 with internal standard (nonanoic acid 1.82 g l-1 in H2O/ethanol 1:1) with

controlled syringes and 1 µl was injected into the gaschromatograph (HP 5890) through an

on-top injector. The temperature of the injector port and detector were 240 and 260°C,

respectively. The temperature program was set as follows: isothermal at 160°C for 2 minutes;

an increase at a rate of 3°C min-1 to 169°C; heating to 200°C and maintaining that

temperature for 2.5 minutes. The total analysis time per sample was 9 minutes. Hydrogen

served as carrier gas. The retention times for octanoic acid and nonanoic acid were 3.08 min

and 4.23 min, respectively. For signal processing the software Spectracontrol™ was used

and the data were saved in Excel files. Every 50th to 54th sample was used for the

recalibration of the system with 0.8 g octanoate l-1. The relative error of the measurements

was less than 4%.

73

On-line ammonium analyzer

The residual ammonium concentration in the culture was measured by an on-line enzymatic

analyzer (Biospectra, Schlieren, Switzerland) . A cell free sample was mixed with a reaction

mixture that resulted in the final concentration:

u ~ @ z

c r--0 l,g/N}r-_r- I Ct :.;::::; <ti ...... -c Q) Nt = constant () c --0 Time u

Steady-state Gradient feed

@ (C/N)t I C-limited l

--- -u Nt = constant ~ © - -z -c N-limited

0 :.;::::; <ti ...... -c Q)

Ct= constant () c 0 u - - ........

N-limited

Nt = constant t = 0 t=t Time

Co ' No Ct ' Nt

FIG. 3.2: Nutrient feeding strategies that are typically used to investigate dual (C,N) nun·ient limited growth. The panels show the C and N concentrations prior to entering the reactor and the shaded area indicates trnnsitionaiy (non-steady-state) conditions of the culture.


Panel a: Discontinuous gradient. Nf is constant and Cf is increased stepwise. After the cell density reaches a steady-state samples of the culture were analysed and the carbon concentration is increased and maintained until a new steady-state is reached.

Panel b: Continuous gradient. Nf is constant, and Cf is increased continuously (convex increase of Cf/Nf ratio). The parameters that determine the shape of the gradient are the initial volume, [C] and [N] in vessels V1 and V2, and the flow rates of pumps 1 and 2.

Panel c: As for b, with Cf constant, and Nf decreasing continuously (concave decrease of Cf/Nf ratio).

Panel d: As for b, with Cf decreasing linearly (linear decrease of Cf/Nf ratio). Glutamate-DH, 6.17 ml l-1 (7400 U); 2-oxoglutarate, 2.56 g l-1; ADP, 0.34 g l-1; NADH,

0.135 g l-1; triethanolamine-HCl, 29.07 g l-1. The NADH decrease was measured at 340 nm

and the calculated concentration was saved as an Excel file. When the concentration of

ammonium exceeded the valid measurement range, a new sample was taken automatically

and diluted in order to ensure a correct measurement. The time between two samples was

20 minutes. Recalibration of the analyzer was done manually every 24 hours.

Off-line determinations

Sampling

Samples of the culture and of the medium feed (S1 and S2, Fig. 3.1) were put on ice

immediately. Typically the sample volume of the culture broth was 6% of VR and the culture

volume was re-established after 12 minutes. The sampling intervals varied from experiment

to experiment and can be taken from the time scale plotted for each transient experiment.

Optical density

Optical density was measured with a spectrophotometer (Pharmacia, Uppsala, Sweden) at a

wavelength of 450 nm against 10 mM MgSO4. For OD 450 nm > 0.3, samples were diluted

with 10 mM MgSO4. Cell dry mass concentration was estimated using a factor of 0.29 mg

ml-1 per OD 450 = 1.

Cell dry weight

Polycarbonate filters (Nuclepore, 47 mm diameter and 0.2 µm pore size) were initially dried

before use at 80°C for one day and tared after cooling down to room temperature in a

75

desiccator. About 5 mg cell dry mass based on the OD450 measurements was filtered

through a tared filter. Filters were washed with the same volume of 10 mM MgSO4, dried to

constant weight at 80°C for at least 3 days, and weighed after one additional day in a

desiccator.

Elemental composition

The carbon, hydrogen, nitrogen and sulfur content of 2 mg freeze dried cell samples were

measured with an elemental analyzer (EAGER 2000, Carlo Erba, Milan, Italy). Sulfanylamid

was used as a standard for calibration.


PHA-measurement

PHA content of 5 mg freeze dried cell samples was determined through methanolysis

according to Lageveen (1988).

Nitrogen concentration of the feed medium

Nitrogen was measured according to the method of Scheiner (1976).

Protein content

The protein content of cells was determined by the modified Biuret method of Koch and

Putnam (1971). 10 mg freeze dried cells were diluted in 1 ml distilled water in a 2 ml

Eppendorf tube, 0.5 ml 3 M NaOH was added and the samples were boiled in a water bath for

5 minutes. Subsequent to alkaline hydrolysis the samples were cooled on ice for 5 min, 0.5

ml CuSO4 (25 g l-1 in water) was added and the tubes were vortexed for half a minute. After 5

minutes, the tubes were centrifuged at 18,800 g in an Eppendorf centrifuge for 5 min. The

supernatant was collected and absorbence at 555 nm was measured against a standard

without protein. Protein standards were prepared from bovine serum albumin (Boehringer

Mannheim, Germany) for the range of 0-10 mg l-1 protein.

Chemicals

All chemicals were purchased from Merck (Darmstadt, Germany) and were >99% pure.

THEORETICAL ASPECTS

The discontinuous gradient approach to determine dual nutrient limitation zones (Fig. 3.2a)

is time consuming. We therefore explored an alternative technique that provides equivalent

data in a shorter time. Instead of changing the Cf/Nf ratio discontinuously, we produced

different nutrient gradients in the feed medium with the apparatus shown in Figure 3.1. Thus,

77

linear, convex, and concave, increasing and decreasing nutrient gradients at different rates

were applied (Figs. 3.2b, c, and d).

Consequently, the culture is not anymore at steady-state and the real specific growth rate

(µ), which depends on nutrient availability and determines the increase or decrease of the

biomass during the experiment, may be higher or lower than the hydraulic dilution rate (D).


Increase of the medium feed concentration

The selection of a nutrient gradient has to be done carefully. One must consider that the

specific growth rate µ cannot exceed maximum growth rate µmax. Consequently, the gradient

should be designed such that the cells are able to adapt to the medium gradient (µ< µmax).

[3.1]

The growth rate µ is dependent on the increase of the limiting nutrient in the bioreactor dsR/dt,

on the corresponding maximum yield of this nutrient YX/S, and on the dilution rate D of the

bioreactor.

According to Pagni et al. (1995) the cells adapt to moderate changes in the medium feed

immediately. We chose gradients that were sufficiently small to fulfill the above

requirement.

Decrease of the medium feed concentration

The situation is different for negative medium gradients. A rapid decrease of growth limiting

nutrients could largely increase the endogenous metabolism (consumption of own mass in

the absence of an energy source) and consequently, the physiology of the culture would not

represent that of a continuously growing culture.

This effect can be avoided by decreasing the concentration of feed medium components at

a rate which is considerably lower than the wash-out rate. The cells will grow normally, at a

specific growth rate µ which is lower than the hydraulic dilution rate D.

Specific uptake of nutrients

During a medium gradient with increasing concentration of the growth limiting nutrient,

more of this nutrient becomes available to the cell and consequently more of this nutrient is

taken up at a higher rate as long as the nutrient is limiting growth. When the growth limitation

of the gradient nutrient is released and a second nutrient becomes limiting instead, the

biomass concentration should stay constant and consequently the specific uptake then has to

remain constant, too.

79

The situation is different when the concentration of the initially non-limiting nutrient is

decreased over time. The availability of this nutrient is continuously reduced and

consequently the specific uptake rate will become smaller as soon as the nutrient becomes

growth limiting.

The specific uptake rate (qS) for a nutrient is calculated as follows (Powell, 1967):

[3.2]

where dS/dt is the consumption of the nutrient S over a time interval, dX/dt is the biomass

produced over the same time interval, and µ is the actual specific growth rate of the culture

(under steady-state conditions is µ = D).

With respect to the herein described medium gradients, the specific uptake rates of carbon

(qC) and nitrogen (qN) are of particular interest.

RESULTS

Initially we determined inhibitory concentrations of various nutrient sources to establish

appropriate feed strategies.

Carbon source

Na-octanoate was the sole carbon and energy source in our experiments. Complete growth

inhibition by octanoate was reported for P. oleovorans with concentrations higher than 50

mM (Brandl et al., 1988). In order to determine the tolerance of P. oleovorans towards lower

concentrations of octanoate, a series of batch experiments was carried out according to

Owens and Legan (1987) with a modified medium X, where citrate was replaced by different

starting concentrations of octanoate. The inoculum was taken from 2 precultures cultivated

on medium X that were grown to late exponential phase.


Three different growth regimes were observed (Fig. 3.3): I) With increasing

concentrations of octanoate up to 3 mM (0.29 gC l-1) the specific growth rate increased

hyperbolically. From the experimental data the Ks value for octanoate was estimated to be

lower than 0.5 mM for growth in batch culture. II) Between 3 and 10 mM initial octanoate

concentrations the growth rate reached a plateau at a µmax of 0.48 h-1. III) In batch cultures

with initial octanoate concentrations between 10 and 20 mM the µmax decreased from 0.48 to

0.37 h-1. Based on this information, the octanoate feed concentration was always kept within

a range of 1.56 and 31.25 mM, corresponding to 0.15 gC l-1

and 3 gC l-1

, respectively. In a

subsequent experiment using the chemostat method (Owens and Legan, 1987) we found Ks

to be 0.05 mM (Monod kinetics). Recently, a Ks value of 0.1 mM octanoate was published

for another PHA producing organism, Pseudomonas putida (Carnicero et al., 1997).

0

0.1

0.2

0.3

0.4

0.5

0 5 10 15 20

Octanoate [mM]

Max

imum

spe

cific

gro

wth

rate

[h-1

]

FIG. 3.3: Growth of Pseudomonas oleovorans in shake flask cultures with different initial octanoate concentrations. The cultures were grown in 250 ml of a modified medium X with octanoate instead of citrate.

Nitrogen source

We chose a range of 0.1 - 2.0 g N l-1. After testing for effects of octanoate on cell growth, we

selected a range of Cf/Nf between 1 and 20 g g-1.

81

Discontinuous medium gradient experiments

Since we wished to develop a more rapid procedure to determine dual nutrient limited growth

regimes with continuous medium gradients, we first caITied out discontinuous gradient

experiments to serve as a comparative benchmark.

Pseudomonas oleovorans was cultured in a chemostat (Fig. 3.1) with medium XCC at a

dilution rate of 0.3 h-1. The feed medium contained a constant nitrogen concentration and the

octanoate concentrntion was increased stepwise after each steady-state measurement,

resulting in a discontinuous gradient (Fig. 3 .2a).

C-limitation N-limitation l~~~~~~~~I

C: 0.60 2.25 N: 0.15 0.15 c

Q) O> e

0.30------------------- 8 7 0.25 ~

c -g 0.20 ro § ~ 0.15 .0 O> ............... ~ 0.10

~ 0.05 "O

6 -5 E c 4 0

l.()

3 ~ 0

2 0 1

~ 0.00 0 0::: 0 4 5 6 7 8 9 10 11 12 13 14 15 16

Carbon to nitrogen ratio [g g-1]

FIG_ 3-4: Determination of dual (C,N) nutrient limited growth in a discontinuous gradient. Pseudomonas oleovorans was cultured in a chemostat on medium XCC at a dilution rate of 0.3 h-1. The carbon to nitrogen ratio in the feed was increased stepwise after steady-state was reached (see Figure 3.2a). The nitrogen concentration in the feed was constant at 0.15 gN r 1

, while [C] increased from 0.6 to 2.25 gC r1

. The culture was exposed to excess nitrngen below Ct/Nr= 6.4 g g-1, while carbon was in excess above CtlNr = 9 .5 g g-1. Both C and N were limiting between these C/N ratios, which define the shaded dual (C,N) limited zone. Carbon (O); nitrogen (X); biomass (optical density) (0 ).


The culture was carbon limited at low Cf/Nf values and consequently the biomass increased

as more C was introduced (Fig. 3.4). At a Cf/Nf = 6.4 g g-1 ammonia had decreased to a

concentration below the detection limit of 0.2 mg N l-1. A dual (C,N) limited zone could be

detected between feed Cf/Nf ratios 6.4 and 9.5 g g-1. Within this zone the optical density

increased linearly as more carbon was added to the feed medium and no unutilized carbon

was detected in the culture. When the Cf/Nf ratio exceeded 9.5 g g-1 the culture became

clearly limited by nitrogen only. As unutilized octanoate rose with increasing Cf/Nf ratios the

concentration of biomass decreased slightly. This effect was also observed by Durner (1998)

in similar chemostat experiments at different growth rates.

The time needed to collect steady-state data for each C/N ratio data set was one to two

days; based on 18 data sets, the total time required for this experiment was more than 3 weeks

without including any time for restarts of the chemostat due to wall growth.

83

Continuous medium gradient experiments

Instead of changing the medium step-wise, the C/N ratio in the feed medium was changed

continuously, using the approach described in Figures 3.1 and 3.2. The effect of different C

or N gradients on substrate utilization and cellular composition of Pseudomonas oleovorans

was studied. From a number of different medium gradients performed, three were selected

and presented here. For a better comparison between the experiments the data of the medium

gradients are presented in the same way as the chemostat data (C/N ratio of the feed medium

as reference) and do not represent the time course (compare time scale and the small icon in

the first panel of the figures).

Gradient 1: Convex increase of Cf/Nf over 66 hours with increasing carbon

concentration in the medium feed

Chemostat conditions were initially established at a dilution rate of 0.3 h-1, using a medium of

Cf/Nf = 0.9 g g-1 (N = 0.15 g l-1 and C = 0.135 g l-1). After steady-state was attained, the

octanoate concentration in the feed medium in vessel V1 was increased continuously from

0.135 to 2.1 g l-1, at a constant N concentration, resulting in a convex C/N feed gradient from

0.9 to 14 g g-1, within 20 volume changes (Figs. 3.2b and 3.5).

With carbon as the limiting nutrient, the biomass increased with the C concentration in the

feed (Fig. 3.5a). Concomitantly, more nitrogen was utilised, until at a Cf/Nf of 7.2 g g-1 it also

became limiting. The biomass increased further between Cf/Nf ratios of 7.2 and 11 g g-1 due

to intracellular accumulation of PHA (Fig. 3.5b). When the Cf/Nf ratio exceeded 11 g g-1,

residual carbon was detected in the supernatant, indicating that growth of the culture was

now limited by nitrogen only . As a result, the biomass leveled off and remained constant at

1.6 g l-1.


C-l imitation N-l imitation I I

c::: c 0.135 2.25 Q) O> N 0.15 0.151 .8 0 0.30 ...... ~

:-.:

®~ .,....

1.6 ..L c::: 0.25 "'C 1.4 ~ c::: -ro ~ 0.20 1.2 ..c::: c::: ~ O> o-

0.15 1.0 'Q) .0 O>

0.8 ~ ...... .......... ro ~ () 0.10 0.6 "'C ro 0.4 Q) ::I 0.05 "'C 0.2 u ' ii) Q) 0 0 0:: en 40 65 en en ro 35 en

E 60 ro 0 30 E :0 55 .Q - 25 .0 0 50 -0 ~ 20

~ <( 15 45

c::: I 10 40 Q) a.. fl -0 5 35 ...... a..

en 0 ©

0 en 55 13 ro E 54 c::: en

.Q 53 12 «> en

.0 O> ro ro 52 11 e ~

51 "'C ·-- 10 ~.o 0 - 50 - - ro 0 49 9 c:::-

Q) 0 ~ 48 8

O>-o-c::: 47 ...... 0 0 7 :-.: ~ .0 46 * z ~

...... ro o- 0 u 0 1 2 3 4 5 6 7 8 9101112131415


I I I I 0 10 20 30 40 50 60 64

Time [h]

FIG.3.5: Determination of dual (C,N) limited growth in a continuous C/N gradient, with a constant N concentration. P. oleovorans was grown continuously at D = 0.3 h-1,

[N] = 0.15 g r 1' while [C] was increased continuously from 0.135 to 2.1 g r 1 (see

Figme 3 .2b ).

85 Chapter 3

Panel a: Cell growth indicated as total biomass (0 ) and rest biomass (PHA-free biomass)(• ). Residual nitrogen (ammonium) ~) and carbon (octanoic acid) (0 ) were measured on-line; nitrogen was in excess below C/N = 7.2 g g-1

, while carbon was in excess above C/N = 11.0 g g-1

, and these boundaries define the shaded dual limited zone. Panel b: Protein (II) and PHA (a ) as a% (w/w) of total biomass. Panel c: Elemental composition as a % (w/w) of total biomass; carbon (<>), nitrngen ~), hydrogen (•).

The PHA content within the dual-limited growth zone increased linearly from 9 to 33% as

the C/N ratio of the feed increased (Fig. 3.5b). Consequently, the composition of the biomass

also changed: protein decreased from 65% to 45% of cell diy weight, whereas the total

carbon content increased from 47.5% to 52.5% of the cell diy weight (Fig. 3.5c). These

changes were not observed when the non-PHA cell biomass was taken as the basis for the

calculation of the cellular protein and carbon contents (data not shown).

Gradient 2: Concave decrease of the CtlNe ratio over 43 hours with increasing nitrogen


To verify whether the same dual nuti·ient limited regime boundaries as those observed in

gradient 1 are obtained when the Cr/Ne ratio is decreased instead of increased, the same

experiment was caITied out, but now with a convex niti·ogen concenti·ation gradient at a

constant carbon concenti·ation in the feed. In this experiment the effective C/N ratio in the

feed decreased from 19.8 to 1.3 g g·1 (Figs. 3.2c and 3.6). As the niti·ogen concentration

increased the cell diy weight also increased and as a result more carbon was utilised. The

maximum biomass concentration was reached at a Cr/Ne ratio of 6.9 g g-1, when the cells

entered the dual nuti·ient limited growth regime. As the N supply increased further, the PHA

content decreased from 26% to 3% of cell diy weight. At the same time, the protein content

rose from 52% to about 61 % of cell diy mass. At Cr/Ne ratios lower than 4.3 g g-1, residual

niti·ogen became detectable and biomass concentration remained constant at about 2.3 g r1•

Thus, in this reversed gradient experiment a dual limitation zone was observed between Cr/Ne - 1 = 4.3 and 6.9 gg .


C-limitation N-limitation I I

c 2.0 2.0 c 1.6 N 2.0 0.13.0 Q) O>

~ ~ e 1.4 ......

:!: 2.5 ..L c 1.2 ~ "O 2.0 -c 1.0 £: ro ~ O> c '7 0.8 1.5

.Q) o- ~ .0 O> ..... .......... 0.6 ~ ro 1.0 () "O

ro 0.4 Q)

::I 0.2 0.5 0 "O (/) 0 0 Q)

a::: 40 70 7ii' (/)

7ii' 35 65 ro (/) E ro 30 60 0 E :.0 0 25 .!le._ -ll - - - - 55 -:.0 v --"' - '6;

0 - 20 50 ~ 0 - ::..... ~ 15 I 45 c

[6] [6] .Q)

<( 10 40 -6 0 I ..... a.. 5 35 a..

0 0 55 13 7ii' 54

(/) 53 12 ro c E 52 0 11 Q) ~ .Q o- O> (/)

51 xx_ 0 (/) .0 x- ~ - 10 ..... ro - 50 ~E 0 - -~ 49 _o-o-o ::IE- ~ 9 £: .Q

::..... 48 - .0 8 c_

c __£_. Q) 0 0 47 _._..--A - ..- O> .0 46 7 e ~ ..... :!: __, ro 0 0 0 z

0 2 4 6 8 10 12 14 16 18 20 Carbon to nitrogen ratio [g g-1]

I I I 53 40 30 20 10 0

Time [h]

FIG. 3.6: Determination of dual (C,N) limited growth in a continuous C/N gradient with a constant C concentration. P. oleovorans was grown continuously at D = 0.3 h-1 in

Chapter 3 87

a medium which contained [C] = 2.0 g l-1, while [N] was increased continuously from 0.1 to 2 g l-1 (see Figure 3.2c). The symbols are as defined for Fig. 3.5. Nitrogen was in excess below 4.3 g g-1 while carbon was in excess above C/N = 6.9 g g-1, and these boundaries define the shaded dual limited zone.

Gradient 3: Linear decrease of the Cf/Nf ratio over 88 hours with decreasing carbon


To test the effect of a decreasing rather than an increasing C concentration (as illustrated in

Figure 3.2d), we carried out a gradient experiment in which the nitrogen feed was kept

constant at 125 mg N l-1 and the carbon feed concentration was decreased over a period of 88

hours, resulting in a Cf/Nf ratio that changed linearly from 18.5 to 2 g g-1 (Figs. 3.2d and 3.7).

After an initial steady-state was reached, the total biomass was constant at 1.45 g l-1. The

carbon source was present in excess and only nitrogen limited growth. After the feed gradient

was initiated, the residual carbon decreased according to the decreasing carbon concentration

in the medium feed until it became limiting at Cf/Nf = 8.9 g g-1, at which point the PHA

content of the cells started to decrease. The dual nutrient limited zone ended at a lower

boundary of 5.1 g g-1, when N began to accumulate in the culture broth. Simultaneously, the

protein content increased from 51% to 63% of cell dry weight in the dual limited zone.

Interestingly, the PHA content of the cells decreased to less than 0.2% (w/w) of total cell

dry mass, as the gradient passed through the C, N dual limited zone, and no PHA could be

detected at a Cf/Nf ratio below 5.1 g g-1, when only carbon limited growth (compare Figure

3.5).

DISCUSSION

The dual nutrient limited growth regimes determined in the continuous gradient experiments

differed from those found with the discontinuous gradient (chemostat). In general, the zones

found with continuous gradients were narrower and shifted in the direction of the Cf/Nf

gradient (e.g. decreasing Cf/Nf ratios resulted in a shift of the dual nutrient limited growth

regime towards lower Cf/Nf ratios). This was apparently due to experimental and

physiological effects.


C-limitation N-limitation I I

c 0.14 1.75 N 0.125 0.125

c 0.30 1.8 Q) O> ® 1.6 ~ 0 0.25 ...... ......

0 ..L :!: 1.4 c ~ "'C 0.20 1.2 -c £: ro ~ 1.0 .Q> c ~ 0.15 Q) o- 0.8 ~ .0 O> ...... ..........

0.10 0.6 ~ ro () "'C

ro 0.05 0.4 Q)

:J 0.2 0 "'C ·en 0 0 Q)

@ 0:: 40 70 ~ (J) • (J)

'7ii 35 - ........ 65 ro (J) ..... E ro 30

~tr - 4 60 0 E :0

.Q 25 .-- • 6 55 -.0 20 {{ -.-•. 50 0 - ~ 0 I £.....

~ 15 45 c 10 ti 40 .Q)

<( -0 I 5 I 35 ...... a.. a..

0 30 0 1 2 3 4 5 6 7 8 91011121314


80 70 60 50 40 30 24 ll_J

0 Time [h]

FIG. 3.7: Determination of dual (C,N) limited growth in a reversed continuous C/N gradient with a constant Nand a decreasing C concentration. P. oleovorans was grown continuously at D = 0.3 h-t in a medium which contained [N] = 0.125 g r t, while [C] was decreased continuously from 1.75 to 0.14 g r t (see Figure 3.2d). Symbols are used as defined in Figure 3.5. Nitrngen was in excess below C/N 5 .1 g g-t, while carbon was in excess above C/N = 8.9 g g-t, and these boundaries define the shaded dual limited zone.

Chapter 3 89

Experimental effects

Generally, the maximum biomass concentration that can be attained in a continuous culture

in steady-state depends on the concentration of the growth limiting nutrient(s) in the medium

feed. The situation is more complex under transient conditions when the continuous culture

is exposed to a changing nutrient concentration in the feed medium.

Thus, the state of a gradient culture with a distinct nutrient concentration in the feed

medium cannot be compared with a steady-state culture with the same feed concentration,

because the history of the gradient has an influence on the behaviour of the culture

(previously added volume units are more or less concentrated, wash-in and wash-out effect).

This effect is well known from nutrient shift experiments (Egli, 1982) and the magnitude of

the effect depends on the volume exchange rate of the bioreactor and the slope of the nutrient

gradient concentration. For instance, a slow nutrient gradient at a low dilution rate will show

less physical delay than a faster gradient at the same dilution rate.

When the medium gradient in the feed and the dilution rate are known accurately, this

physical delay can be corrected mathematically.

Correction of the physical delay

The measured Cf/Nf ratios were used to create a polynomial equation of 3rd degree (Cf/Nf(t) =

α1 + α2t + α3t2 + α4t3). The coefficients of this polynomial were used to calculate the

corrected C/N ratio for the time t (Zinn HP, personal communication) (see also Appendix ii):

[3.3]

where: (C/N)corr(t) is the corrected carbon to nitrogen ratio at time t, (C/N)0 the carbon to

nitrogen ratio of the initial steady-state medium, and n the polynomial coefficients of the

gradient. In our experiments (C/N)0 was always equal to α1.


A summary of the C/N values at the dual nutrient limitation boundaries with and without

correction is given in Table 3.1. It can be seen that the mathematical correction considerably

helped to approach steady-state boundaries. The best correction was achieved for the linear

gradient 3 (linear decrease of the C source over time).

Based on mathematical simulations, we concluded that linear gradients are best for such

experiments. First, they are easier to perform (see Appendix ii). Second, they have a higher

accuracy, because all measured feed concentrations correlate linearly to the time. Third, the

mathematical correction is more accurate, because fewer parameters are involved

(α3 = α4 = 0 in eq. 3.3).

Physiological effects

In addition to the experimental effects, a physiological effect also influences the observed

C/N ratios. In contrast to the experimental wash-in effect, physiological effects cannot be

quantified a priori and consequently no mathematical correction is possible. We found that

there are several factors that can affect the outcome of a gradient experiment:

1. Number of generations

Unquestionably, an important factor is how many generations elapse during the experiment;

the faster a gradient is performed over the same Cf/Nf range, the fewer generations are

involved, and consequently, the greater the lag between the imposed nutrient concentration

change and the resulting cellular response. The lowest generation number was calculated for

gradient 2, which had only 19 generations and showed the greatest deviation from true

steady-state chemostat data (gradient 1: 28.5 and gradient 3: 38 generations). According to

table 3.1 the gradient 3 with the highest generation number, showed the smallest deviation to

the boundaries of the chemostat culture after the correction of the physical delay.

2. Driving force of the physiological adaptation

A second reason for a deviation from the chemostat data is related to the nature of the

physiological adaptation elicited by the environmental pressure. The changes of the growth

conditions over the performed medium gradients are considerable. Thus, changes in the

Chapter 3 91

cellular management of the intracellular carbon and nitrogen pool must proceed efficiently in

order to satisfy the needs for cell and energy maintenance (e.g. PHA accumulation or

degradation under N and C shortage, respectively).

Depending on the kind of the gradient, the cells may react in a different way.

In gradient 1 the total biomass increased, due to the increase of the carbon supply, so that

ultimately all of the fed nitrogen was consumed. A further increase of the carbon source

resulted in an accumulation of PHA, and simultaneously, the cells were constrained to

optimize their nitrogen utilzation under the existing nitrogen limitation.

Gradient 2 was similar to gradient 1: the biomass increased over the experiment, due to the

increased nitrogen supply under initial nitrogen limited conditions. As the nitrogen input

increased the nitrogen limitation was relaxed and carbon limitation ensued. This challenge

could initially be simply satisfied by reducing the accumulation of PHA. Moreover, the cells

had a lower PHA content than the first and third gradient under N only limitation. This

observation can be explained by the reduced PHA content of P. oleovorans at higher growth

rates (Durner, 1998; Preusting et al., 1991).

TAB. 3.1: Summary of the dual nutrient boundaries obtained by chemostat and transient

experiments of Pseudomonas oleovorans at D = 0.3 h-1.

Chemostat Gradient 1 Gradient 2 Gradient 3

Shape of C/N gradient

stepwise increase

convex increase concave decrease

linear decrease

Change of C/N ratio by

increase of C increase of C increase of N decrease of C

Duration > 500 h 66 h 43 h 88 h

Generations 216 28.6 18.6 38.1

Experimentala

Lower boundary [gC/gN]

6.4

7.2 (+12.5 %)c

4.3 (-32.8 %)c

5.1 (-20.3 %)c

Upper boundary [gC/gN]

9.5

11.0 (+15.8 %)c

6.9 (-27.4 %)c

8.9 (-6.3 %)c


Mathematical correctionb

Lower boundary [gC/gN]

not necessary

6.7 (+4.7 %)c

5.2 (-18.8 %)c

6.5 (+1.6%)c

Upper boundary [gC/gN]

not necessary 10.1 (+6.3 %)c

8.1 (-14.7 %)c

9.6 (+1.1 %)c

a Experimental boundaries. The dual nutrient limited growth regime was determined by the measurement of the residual concentrations of nitrogen and carbon. Dual nutrient limitation was defined as a nitrogen concentration below 0.2 mg l-1 and a carbon concentration below 8 mg l-1. As a reference the C/N ratio in the feed medium was used.

b Corrected boundaries. The results obtained in the transient experiments were corrected with equation 3.3 (see also Appendix).

c The numbers in parentheses represent the deviation of the experimental or corrected C/N value from the steady-state value.

Chapter 3 93

Gradient 3 differed completely from gradients 1 and 2. The biomass decreased over the

time course due to the decreasing C concentration. The driving force to adapt to the medium

gradient and to degrade PHA is probably much more prominent than in gradient 2, because

the culture is kept in a famine state throughout the whole gradient. PHA degradation was so

effective that less than 0.2% PHA could be detected under C only limitation.

3. Uptake of the limiting nutrients

In the here presented experiments the cells were always limited by at least one nutrient. This

was not the case when the slope of gradient 1 was doubled (data not shown), unlimited

growth occurred (N and C source in excess), and the cells were not able to adapt to the

changing medium gradient. The analysis of the data revealed that the theoretical rules

described in section theoretical aspects were not satisfied sufficiently. The gradient was

performed too rapidly, so that the cells were unable to take up nutrients at a sufficiently high

rate. Moreover, the biomass should have grown faster than µmax in order to adapt to the

gradient (see eq. 3.1).

Physiological changes of P. oleovorans during the medium gradients

Changes of the cellular composition

The C, N, and H content of the cells is strongly dependent on the Cf/Nf ratio under dual

nutrient limited growth conditions. Under purely C or purely N limitation the cellular

composition was almost constant. The main reason for the differences in composition is

obviously the accumulation of PHA. PHA produced on octanoate consists up to 67% of

carbon (w/w), therefore the C content of the cells is higher than under C limitation. The

change of the PHA content per (C/N)corr was essentially constant within the dual limited

growth zone. Gradient 1 produced 5.6 % PHA per (C/N)corr unit, gradient 2 degraded 6.25 %

PHA, and gradient 3 degraded 6.4 % PHA per (C/N)corr unit. The different ratios of PHA

production/consumption to (C/N)corr might be caused by the additional utilization of PHA as

internal carbon source. As a consequence, gradient 2 and 3 show a steeper slope than gradient

1. The degradation process in P. oleovorans is not fully understood yet. Depending on how

PHA degradation is regulated, the cell might react rapidly (activation of existing


depolymerase), or might need more time (de novo synthesis). As the data of chapter 4

indicate, PHA is accumulated and degraded simultaneously in P. oleovorans, as is the case

also for PHB in Alcaligenes eutrophus (Doi et al. , 1990).

en 80 en ro

E 75 0 :0

70 ro :::J 65 "'C .ii) Q) ...... 60 -0 55 ~ c 50 ·m -0 ...... 0

a_

c C-limitation c N-limitation

Gradient 1

Gradient 2

Gradient 3

FIG. 3.8: Protein content of the PHA-free (residual) biomass for medium gradients 1, 2, and 3 under only carbon or nitrogen limiting growth conditions. The ve1i ical en or bars represent the standard deviation of the average protein content of the carbon or nitrogen limited growth zone for each medium gradient.

The protein content of the cells decreased with increasing (CIN)corr ratio (e.g. in gradient 1

from 65% to 45% of total cell diy weight) because protein was diluted as a fraction of the

total biomass due to the PHA accumulation. However, when taking into account only the

residual biomass (biomass without PHA), the protein content was almost identical for cells

grown under N or C limitation in gradient 2 as expected (no stringent N limitation since N

was continuously increased) (Fig. 3.8). Interestingly, the cells exposed to gradient 1 showed

a higher protein content under C limitation than under N limitation, whereas the opposite was

the case in gradient 3. This is likely to be due to the fact that the cells of gradient 1

experienced an increase of the carbon source, so that there always was enough nitrogen

available to produce protein under C limited conditions, whereas under nitrogen limiting

conditions the cells had to reduce their nitrogen needs. The situation was completely

95 Chapter 3

different in gradient 3. The cells experienced a carbon and simultaneously an energy

limitation. Thus, all approaches to satisfying their energy needs, such as reduction of the

protein content, are exploited by these cultures.

--------------.....----- 0.08 0.4

0.06 0.3 •••

o e ~ 0.2 0 O o 0.04 oa:o

0.1 @ 0.02

~ 0 0 I

~ 0 10 20 30 40 50 60 70 .,....

,........, .,.... I

~

b> O> ....... -------.....----.-------- 0.08 O>

.......... 0.4 • • ~ • • 0 "" <l> • 0 0.06 .::£

~ 0.3 ~ ~ ... ~ ~ 0.2 0 0 0 0 0 0.04 ~

c c 0 0.1 ..c

0.02 <l> ~ ro 0 .,._ ___ .,._ __ ---1......_--+...._....._ ....... ___......._...or 0

O> 0 ~

:t: () () 0 1 0 20 30 40 50 c ~ ()

·5 ....... -------.......---........ ---- 0.08 ~ ~ 0.4 • • ~

Cf) • • • • 0.06 Cf) 0.3 •• •••e o o 04 0.2 0 0 0 0 0 • 00 . 0 0 0 00 0 •• 0.1 0.02

0 ...... """-1 ............... --............... ......-.i ........ --............... +-~ 0 0 10 20 30 40 50 60 70 80

Time [h]

FIG. 3.9: Time comse of the specific uptake rate of carbon (• ) and nitrogen (0 ) for the medium gradients. Panel a: Gradient 1, Panel b: Gradient 2, Panel c: Gradient 3.

Changes of the specific uptake rates of nitrogen and carbon


The calculation of the specific uptake rates (eq. 3.2) for the medium gradients 1, 2, and 3

revealed a complex picture of the physiological adaptation (Fig. 3.9) and may also explain

the above mentioned changes of the cellular composition. In the first medium gradient qC

increased and qN decreased over the time course. The increase of qC can be explained by the

enhanced consumption of C due to PHA accumulation. Since the cellular mass increased for

the same reason, qN had to decrease accordingly. In Figure 3.4 it can be seen that the PHA

content and the total biomass still increased although the cells were purely N limited. This

effect can be considered as a delay of the cellular adaptation.

The second medium gradient, where the Cf/Nf ratio was decreased by an increase of Nf,

the situation is more complex. Under N limitation the cells consumed nitrogen at almost a

constant rate, whereas qC decreased simultaneously. This would indicate that the cells

already started to degrade PHA under N limiting growth conditions, which in fact can be

estimated based on the PHA data (Fig. 3.5b). Under dual (C,N) and C only limitation qN

increased further and qC remained almost constant. The increase of qN is again a function of

the cellular PHA content under dual (C,N) limited growth conditions. However, under C

limitation one would expect to see a constant qN, since the cellular composition should be

constant under N limitation. However, the data indicate that the cell accumulated nitrogen or

excreted nitrogenous compounds the more nitrogen was available for the cell. Finally, it

appears reasonable that the specific carbon uptake remained constant since carbon was then

the new and only limiting nutrient.

In the third medium gradient with a continuous decrease of Cf, qC and qN showed an

interesting pattern. Initially, qN remained constant, since the cells were nitrogen limited. At

the same time the uptake of carbon constantly decreased probably due to a simultaneous

accumulation and degradation of PHA (see above). Under dual (C,N) limited growth qN

increased because the cells degraded PHA now almost completely and thus the N uptake per

g biomass was larger. At the same time qC approached a plateau of 0.23 g g-1 h-1 for the same

reason. Surprisingly, qC and qN decreased under C only limitation further. This effect could

be related to the wash-out effect discussed previously.

Chapter 3 97

OUTLOOK

Our experiments demonstrate the utility of medium gradient experiments to investigate the

cell physiology and the adaptation of an organism to constantly changing growth conditions.

Thus, it could be shown that medium gradients with small changes of the nutrient feed

composition may be used to demonstrate the ability of microorganisms to grow multiple

nutrient limited. Although the boundaries of the dual nutrient limited growth regime differed

from the steady-state cultures, a mathematical correction term could be used to approach the

steady-state boundaries. This is particular useful since the time to carry out a medium

gradient experiment is significantly shorter than a series of chemostat experiments.

Moreover, the medium gradients could give also useful information how the feed technique

of a fed-batch culture could be optimized further.

ACKNOWLEDGEMENTS We thank Biospectra, Schlieren, Switzerland, for help and advice with on-line analytics.

NOMENCLATURE

α1, α2, α3, α4 [g g-1] Polynomial coefficients for the description of the C/N

gradient in the feed medium (Cf/Nf(t)= α1 + α2t + α3t2 + α4t3)

C [g l-1] Carbon concentration in the culture supernatant

Cf [g l-1] Carbon concentration in the feed medium

(C/N)0 [g g-1] Carbon to nitrogen ratio in feed medium before a gradient

was initiated

Cf/Nf(t) [g g-1] Carbon to nitrogen ratio in the feed medium of the bioreactor

at time t measured at the sample port S1 (Fig. 3.1)

(C/N)corr(t) [g g-1] Carbon to nitrogen ratio that is corrected for the wash-in

effect

D [h-1] Dilution rate (D = F1/VR)


F1 [l h-1] Flow of medium XCC1 into the bioreactor (F1 is equal to the

outflow)

F2 [l h-1] Flow of the medium XCC2 into vessel V1

µ [h-1] Specific growth rate of a culture

N [g l-1] Nitrogen concentration in the culture supernatant

Nf [g l-1] Nitrogen concentration in the feed medium

qC, qN [g g-1 h-1] Specific uptake rates for carbon and nitrogen

V1, V2 [l] Volume of the medium vessels V1 and V2

VR [l] Working volume of the bioreactor

YX/C [g g-1] Yield of biomass on used carbon

YX/N [g g-1] Yield of biomass on used nitrogen

Chapter 3 99

APPENDIX A

i) Medium gradient maker

Two medium bottles were connected in series with the bioreactor. Pumps enabled the

transfer of medium XCC2 from vessel V2 into the mixing vessel V1 (at constant rate F2) and

from there into the bioreactor at rate F1 (Fig. 3.1). Depending on the flow rates, gradients

with different shapes were possible: concave (F2 > 2F1), linear (F2 = 2F1), and convex (F2 <

2F1). The slope of these gradients was determined by the volume in the mixing vessel V1, the

nutrient concentrations involved, and the flow rate F2.

The volumes needed for such a transient can be calculated based on the equation:

[A1]

and [A2]

In all cases the medium concentration in vessel V2 determines the end concentration of the

medium gradient.

Typical values for a gradient experiment (Fig. 3.5) were V1 = 20 l, V2 = 10 l, F1 = 0.45 l h-1,

F2 = 0.15 l h-1, and Vr = 1.5 l.

We have measured the medium gradient over the whole experiments to check for errors in the

pumping rates. Thus, errors caused by the medium preparation were considered for the

polynomial calculation of the carbon to nitrogen ratio in the feed medium Cf/Nf(t).

ii) Mathematical correction of the wash-in effect

The wash-in effect of the gradient must be considered: The fed medium is diluted by the

liquid in the bioreactor. Because of this fact, changes of the nutrient and biomass

concentrations in the bioreactor will be delayed in comparison to the steady-state conditions

with the same C/N ratio in the feed medium. This delay can be expressed with a corrected


C/N ratio as a new reference. This approach can best be understood by considering the same

set-up in operation in the same way, however this time without biological activity. This is so

because the wash-in effect cannot be measured in a system with biological activity. Below, a

mathematical expression is derived to correct for the difference in (C/N)corr and Cf/Nf.

The mixture Cf/Nf(t) enters the bioreactor from vessel V1 at a constant rate F1. The culture

volume VR in the bioreactor is regulated with an overflow device. The hydraulic dilution rate

D is therefore constant, too.

[A3]

Data for the effective (C/N) ratio in the feed as a function of time were fit by a polynomial

equation of 3rd: ((C/N)(t) = α1 + α 2t + α 3t2 + α 4t3). The coefficients of this polynomial were

used to calculate the effective C/N ratio in the bioreactor as a function of time t.

The correction of the C/N ratio caused by the wash-in or wash-out effects is described by

the following inhomogeneous differential equation (Zinn HP, personal communication):

[A4]

thus giving the general solution:

[A5]

The following equation represents the solution for medium gradient (C/N)corr(t) whose feed

gradient is expressed as a polynomial of 3rd degree:

[A6]

Chapter 3 101

where: (C/N)corr(t) is the effective carbon to nitrogen ratio at time t in the bioreactor, (C/N)0

the initial effective carbon to nitrogen ratio in the bioreactor, and n the polynomial

coefficients of the gradient. In our case (C/N)0 was always equal to α1.

To demonstrate the dilution effect the 3 medium gradients are displayed in Figure A1 with

and without correction.

10 20 30 40 50 02468101214161820

02468101214161820

0 10 20 30 40 50 60Time [h]

010203040506070800

ba c

(C/N

) cor

r [g

g-1 ]

Cf/N

f [g

g-1 ]

FIG. A1: The medium gradient feed subject to a wash-in effect, which results in a time delay of the gradient. The dashed line represents the C/N ratio of the medium feed Cf/Nf, whereas the solid line represents the gradient (C/N)corr corrected for the time delay.

Panel a: Medium gradient 1 (convex increase) Panel b: Medium gradient 2 (concave decrease) Panel c: Medium gradient 3 (linear decrease)

102

CHAPTER 4

INTRACELLULAR DEGRADATION OF POLY(3-HYDROXYALKANOATE) IN

PSEUDOMONAS OLEOVORANS

Manfred Zinn, Thomas Egli, and Bernard Witholt Keywords: Starvation, batch, PHA, degradation rate, monomer, rifampicin, depolymerase.

Chapter 4 103

SUMMARY

Pseudomonas oleovorans was grown in 2 chemostats (D = 0.29h-1 and 0.32 h-1) with

octanoate as the only source of carbon where either nitrogen (N) or both carbon and nitrogen

(C,N) limited the growth. Under the selected growth conditions the cells accumulated the

storage compound poly(3-hydroxyalkanoate) (PHA). Cells of steady-state cultures were

harvested for subsequent degradation experiments in 1 l shake flasks which were

supplemented with different additives.

The degradation rate was significantly increased by the addition of nitrogen (ammonia).

Thus, carbon starved cultures degraded PHA at about twice the rate of dual carbon and

nitrogen limited cultures. Generally, it could be shown that the depolymerization of the PHA

monomers 3-hydroxyhexanoate and 3-hydroxyoctanoate occurs at equal rates, whereas the

monomer 3-hydroxydecanoate is not degraded at all. This strongly suggests that

3-hydroxydecanoate is not a part of PHA but is a constituent of the biomass.

Cell cultures that were blocked in their protein synthesis at the mRNA level, degraded

PHA at equally low rates whether nitrogen and carbon or carbon only were limiting. This

indicates that depolymerase is always present and active and it suggests that PHA is

continuously synthesised and depolymerised in continuous cultures.

INTRODUCTION

Microorganisms have developed special techniques to get over short term nutrient

restrictions in their ecosystem. Dawes (1985) postulated the most important points for

successful survival of vegetative bacteria, amongst which the ability to store and reutilize

carbonaceous materials as carbon and energy sources is of importance. The energy storage

compounds can be divided into 4 classes: polyphosphate, carbohydrates (polyglucans,

glycogen), cyanophycin/phycocyanin, and lipids amongst them poly(3-hydroxyalkanoate)s

(PHAs).

Intracellular Degradation of PHA 104

The latter are widespread among microorganisms and mainly found as

poly-ß-hydroxybutyrate (PHB) (Anderson and Dawes, 1990). However, pseudomonads of

the rRNA group I do accumulate a PHA polymer of variable composition which depends on

the carbon source on which they are growing (Huisman et al., 1991). For instance, when

Pseudomonas oleovorans is cultured in a batch with octanoate as the sole carbon source and

the single nitrogen source ammonium sulfate is present in limiting amounts, PHA is

produced with an average composition of 91 mol% 3-hydroxyoctanoate, 8 mol%

3-hydroxyhexanoate, and 1 mol% 3-hydroxydecanoate (Huisman et al. 1989). The polymer

chain has an average molecular weight of 0.16 * 106 g mol-1 (Gross et al., 1989; Preusting et

al., 1990) and consists on average of more than 1100 monomers. These polymers are

deposited intracellularly in the form of granules (de Smet et al., 1983). Granules have a size

of approxymately 0.2 - 0.6 µm in diameter and their surface is covered by proteins and a lipid

layer (Steinbüchel et al., 1995; van der Leij and Witholt, 1995). It has been shown (Huisman

et al., 1991; Steinbüchel et al., 1992) that the granule proteins consist of two PHA

polymerases C1 and C2, a PHA depolymerase, and structural proteins (phasins).

Huisman and coworkers (1991) sequenced the pha operon and determined the nucleotide

sequence based amino acid sequence of the PHA depolymerase and found a serine box

(Gly-X-Ser-X-Gly), which is common for esterases. However, the intracellular P.

oleovorans PHA depolymerase differs from the extracellular depolymerases produced by

other organisms (Jendrossek et al., 1996). The depolymerase of P. oleovorans has not yet

been detected extracellularly, which is consistent with the observation that P. oleovorans is

not able to grow on PHA supplied in the growth medium as the only carbon source.

The specific PHA degradation rate catalysed by the depolymerase has been measured in

vitro only recently and is 16.1 mg PHA h-1 µg-1 depolymerase (Stuart et al., 1996). There are

no reports on the PHA degradation rate in vivo. Based on the degradation kinetics found for

isolated PHA and PHB granules, the degradation rate is thought to approximate a zero order

reaction (dPHA/dt = -k) (Foster et al., 1994; Merrick and Doudoroff, 1964).

Here we report that in vivo PHA degradation depends on the nature of the starvation

conditions (C or simultaneous C,N). Degradation, which occurs to some extent in the

absence of protein synthesis, is enhanced significantly when protein synthesis is enabled by

Chapter 4 105

the addition of nitrogen. Cells that were blocked in de novo protein synthesis degraded PHA,

indicating that starved cells do not need to synthesise new PHA depolymerases. This implies

that the depolymerases are present during the PHA accumulation phase, suggesting that there

might be simultaneous synthesis and degradation of PHA, as also described for Alcaligenes

eutrophus by Doi and coworkers (Doi et al., 1990).



Microorganism and medium

Pseudomonas oleovorans GPo1 (ATCC 29347) was used in all experiments. The cells were

precultured in 500 ml shake flasks in 150 ml of medium E (Lageveen et al., 1988)

supplemented with trace elements in a shaker incubator at 200 rpm and 30°C. One liter of

medium E supplemented with trace elements consisted of: 3.5 g NaNH4HPO4*4H2O; 7.5 g

K2HPO4*3H2O; 3.7 g KH2PO4; 2.0 g citric acid monohydrate. After heat sterilization the

trace elements were added: 1 ml of 246.5 g l-1 MgSO4*7H2O in 1 l distilled H2O, 1 ml of 2.87

M FeSO4*7H2O in 1 l 1 N HCl and 1 ml of MT stock solution that contained (per liter 1 M

HCl): 1.98 g MnCl2*4H2O, 2.81 g CoSO4*7H2O, 1.47 g CaCl2*2H2O, 0.17 g CuCl2*2H2O,

0.29 g ZnSO4*7H2O.

Generally, when the culture entered the late exponential phase, 100 ml of this culture were

inoculated into a lab-scale bioreactor containing 1.4 l medium XC, which consisted of (per

liter): 2.6 g Na-octanoate, 0.71 g (NH4)2SO4, 1.0 g KH2PO4, 0.01 g

ethylenediamine-tetraacetic acid disodium dihydrate (EDTA disodium) and the pH was

adjusted with 10 M NaOH to 7.2. After autoclaving, the trace elements were added as

indicated above for medium E.

In medium XC, Na-octanoate and (NH4)2SO4 were the sole sources of carbon and nitrogen,

respectively, and the initial medium carbon to nitrogen ratio was 10 g g-1

. Antifoam

polypropylene glycol 2000 was added in variable amounts only when the culture was

changed from batch to continuous feed mode.

The antibiotic rifampicin was used as a mRNA synthesis inhibitor to block the synthesis

of protein. In preliminary studies 50 µg of rifampicin (Sigma, USA) per ml of culture proved

sufficient to stop cell growth of P. oleovorans GPo1. Rifampicin was diluted in DMSO and

was always prepared freshly.

Continuous cultures

Chapter 4 107

All continuous cultures were performed in a 3 l bioreactor (Wubbolts et al., 1996) with a

working volume of 1.5 l. Sterile medium XC was fed continuously by a peristaltic pump

(Watson Marlow 502S) and the volume was kept constant with an overflow device. The

dilution rate of the starter culture in the first series of experiments (B, N) was 0.29 h-1,

whereas in the second series (P, Q, R, T) it was 0.32 h-1. The cultures were aerated at 0.6 vvm

with filter-sterilized air and stirring at 1700 rpm. The temperature of the culture broth was

regulated to 30°C with heating and cooling fingers. The pH was held at 7.0 ± 0.02 by the

controlled addition of 4 N H2SO4 and 2 N NaOH.

Steady-states of the continuous cultures with respect to biomass concentration were

attained after 3 - 5 volume changes (Chen et al., 1995). This was confirmed with optical

density measurements.

Degradation experiments

Once steady-state of the continuous culture was established, medium feed was stopped and

aliquots of the culture were transferred from the bioreactor to 1 l shake flasks with a

peristaltic pump with sterile tubing. Broth was added to each shake culture to an end-volume

of 250 ml. The shake flasks were supplemented with different nutrients or inhibitors as listed

in Table 4.1.

TAB. 4.1: Conditions for PHA degradation in batch experiments of Pseudomonas oleovorans subsequent to cultivation in a chemostat.

Experimenta Dilution rate [h-1]

Limitation in chemostat

Starvation in batch

Additions

B 0.29 Dual (C,N) C, N -b N 0.29 Dual (C,N) C Nc

P 0.32 N C Nd Q 0.32 N C, N NaCle R 0.32 N C N , rifampicinf T 0.32 N C, N rifampicing

a Two different chemostat cultures were used to generate cell material for the PHA degradation experiments. For each experiment 250 ml of culture broth were transferred by a peristaltic pump with sterile tubing into 1 l shake flasks. Additions were made as indicated.


b Culture B (control) was kept in the bioreactor and no addition was made. c Culture N: 3.3 ml 1 M (NH4)2SO4 was added to a final concentration of 0.4 g N l-1. d Culture P: 3.3 ml 1 M (NH4)2SO4 in 0.9% NaCl was added to a final concentration of 0.4 g N l-1. e Culture Q: 3.3 ml 0.9% NaCl was added. f Culture R: 3.3 ml 1 M (NH4)2SO4 and rifampicin in 0.9% NaCl were added to final concentrations of 0.4 g N

l-1 and 50 mg rifampicin l-1. g Culture T: 3.3 ml rifampicin in 0.9% NaCl was added to a final concentration of 50 mg rifampicin l-1.

The cultures were incubated on a rotary shaker at 200 rpm and 30°C. As a function of time,

samples of 10 ml were withdrawn and put on ice immediately. Two milliliters were used for

the measurement of the optical density and the remaining 8 ml were centrifuged at 8,400g

and 4°C. The supernatant was collected and stored at -20°C for further analyses. The pellet

was resuspended, split into two 2 ml Eppendorf tubes, and the cells were washed with 1 ml

distilled water. They were resuspended in 1 ml distilled water and frozen at -20°. The cell

suspensions were freeze dried and used for the measurement of PHA and protein content.

Off-line determinations

Optical density

Cell growth was determined by measuring the optical density at 450 nm (OD450) against 10

mM MgSO4 with a spectrophotometer (Pharmacia, Uppsala, Sweden). Previous studies

showed that the optical density and biomass (cell dry weight) correlate up to an absorption of

0.3 under carbon limitation. If necessary, samples were diluted with 10 mM MgSO4. Optical

density was used for the determination of the specific growth rate, because cell dry weight

measurements were not reliable with small sample volumes.

PHA-measurement

The PHA content of 5 mg of freeze dried cells was determined with a GCMS MD 800

(Fisons Instruments Ltd., United Kingdom) following the methanolysis procedure described

by Lageveen et al. (1988).

Degradation rate and half life time of PHA

The degradation of PHA was calculated as follows:

Chapter 4 109

, [4.1]

where r(PHA) is the degradation rate of PHA, PHA1 and PHA2 are the PHA contents of the

cells at time t1 and t2, respectively. The half life time (t1/2) of PHA was calculated based on

the following equation:

. [4.2]

The same equations were used to determine the degradation rate of the monomeric PHA

compounds, such as 3-hydroxyhexanoate, 3-hydroxyoctanoate, and 3-hydroxydecanoate.


Elemental composition of the biomass

The nitrogen, carbon, hydrogen, and sulfur composition of 2 mg freeze-dried cells was

measured with a Carlo Erba elemental analyzer (EAGER 2000, Carlo Erba, Italy).

Nitrogen determination

The ammonium concentration in the growth medium and supernatants was measured by the

indophenol method described by Scheiner (1976).

Octanoate concentration

To determine the octanoate concentration in medium and supernatants, the GC method

described by Rothen (1997) was used.

RESULTS

Poly(3-hydroxyalkanoate) (PHA) is accumulated by Pseudomonas oleovorans under growth

conditions where nutrients other than carbon (N, P, Mg, or S) are limiting growth when

cultivated with medium chain length alkanes, alkanoic acids or alkanols (Lageveen et al.,

1988). If the restricting growth condition is released by the addition of the limiting nutrient,

PHA is degraded again and used as a carbon/energy source (Steinbüchel and Valentin, 1995).

Degradation of PHA

Culturing Pseudomonas oleovorans in the chemostat with medium XC resulted in the

accumulation of large amounts of PHA. A series of experiments (Tables 4.1 to 4.3) was

carried out in which the intracellular PHA degradation in such cells was investigated.

Chapter 4 111

The dilution rate (D) was 0.29 h-1

for experiments B and N, and 0.32 h-1

for experiments P,

Q, R, and T. This slight difference in growth rates resulted in changes of the residual

concentration of octanoate in the medium and in the PHA content of the cells. For the slower

growth rate the residual nitrogen concentration was 0.15 mg N l-1

, no residual octanoate was

detected (C below detection limit of 5 mg C l-1), and the PHA content was 27% of the cell dry

weight under these dual limited growth conditions. The slightly higher growth rate (D = 0.32

h-1) resulted in nitrogen limitation: residual N = 0.17 mg l-1

and residual carbon = 63 mg C l-1

.

The PHA content was somewhat lower at 24% of the cell dry weight.

These chemostat-grown cells were withdrawn form the bioreactor and were subjected to a

set of environmental conditions under which accumulated PHA was degraded (see Table 4.1).

Two groups of experiments were done: one group aimed at determining whether PHA is

degraded in cells cultivated under dual nutrient limitation and whether the degradation is

influenced by the addition of nitrogen, e.g. when shifting the cells to conditions limited by

carbon only. The second set of experiments was done to study the influence of a blockage of

protein synthesis on PHA depolymerisation. Rifampicin, which blocks the synthesis of

protein at the transcription level (mRNA synthesis), was used as the antibiotic.

Effect of a nitrogen pulse on PHA degradation in cells growing under simultaneous

nitrogen and carbon limitation (cultures B, N)

Figure 4.1 shows the reaction of P. oleovorans when the cells were starved for either carbon

and nitrogen simultaneously or for carbon only over a time period of 5 hours. The specific

growth rate for culture B was -0.07 h-1

and for culture N -0.06 h-1

, respectively. The decrease

of the optical density was caused by the depolymerization of intracellular PHA and to a lesser

degree by cell lysis (Dawes, 1985), as the optical refraction of the cells is enhanced with

higher PHA content (R. Durner, personal communication). The intracellular

depolymerization of PHA occurred exponentially, with a degradation rate which was

strongly influenced by the addition of nitrogen. Thus, the PHA degradation rate for the dual

limited culture (B) was 0.14 h-1

, whereas the rate for the culture limited by carbon only (to

which nitrogen had been added) was 0.31 h-1

(Table 4.2).


3.4 30.0

3.2 25.0

- 3 20.0 E 0 c: 2.8 E c 15.0 c: U') 2.6 "It' c c U')

0 2.4 "It' - c c: 10.0 0 ...J 2.2 - <C

<C 2 ::c ::c a.. a.. 1.8 -c: ...J

1.6 5.0 1.4 4.1

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Time [h]

FIG. 4.1: Effect of a nitrogen-(ammonimn) pulse on PHA degradation in a P. oleovorans culture simultaneously starved for nitrogen and carbon. The feed of a chemostat culture (D = 0.29 h·1, C/N = 10 g g·1

) was stopped and the culture was separated into two part s: Culture B (open symbols) remained (N,C) limited (no nutrient pulse), and culture N (filled symbols) was supplemented with an ammonia pulse (remained starved for C), respectively. Time-course of the optical density: culture B (D), culture N C• ), and the PHA content in the cell (% of cell chy weight): culture B (0 ), culture N (• ).

The data in Figure 4.2a demonstrntes that the composition of the extracted intracellular·

alkanoic acids was not constant during the degradation process. Initially, cells from

experiments B and N contained PHA with the same monomeric composition found for cells

of the original chemostat culture: 87 mol% 3-hydroxyoctanoate (C8), 11 mol%

3-hydroxyhexanoate (C6) and 2 mol% 3-hych·oxydecanoate (Cl O) . After five hours of PHA

degradation the 3-hych·oxy fatty acid monomer composition of the cells had changed: the

fraction of Cl O had increased, whereas the molar fraction of C6 and C8 monomers was

almost constant. Figure 4.2b clear·ly demonstrates that this is due to the fact that the Cl O

monomer was not degraded, while C6 and C8 monomers disappeared with the same kinetics

Chapter 4 113

seen for the PHA. In both experiments (Tab. 4.2) the degradation rates of C6 and C8 were

equal (culture B: r(C8) = r(C6) = 0.16 h-1, culture N: r(C8) = r(C6) = 0.3 h-1).

TAB. 4.2: Intracellular degradation of PHA in P. oleovorans GPo1 under different

conditions of nutrient starvation.

Batcha r(C8) [h-1]b

r(C6) [h-1]b

r(C10) [h-1]b

r(PHA)early [h-1]c

r(PHA)late

[h-1]d

t1/2 [h]e

Number of data pointsf

B 0.165 0.163 -0.006 0.14 - 4.85 10 N 0.305 0.304 0.014 0.31 - 2.48 13 P 0.432 0.592 nd 0.49 0.37 1.41 12 Q 0.267 0.359 nd 0.27 1.56 2.56 5 R 0.066 0.143 nd 0.16 0.03 4.33 7 T 0.151 0.253 nd 0.16 0.03 4.33 7

a Growth conditions as described in Table 4.1 b The intracellular PHA composition is given by the product of the PHA content (% CDW) and the monomeric

composition in the isolated PHA (% mol). Based on these values the degradation rate was calculated specifically for each monomer with equation 4.1.

c Degradation rate of the intracellular PHA for the first (early) degradation phase (hours 0 to 4). The degradation rate was calculated with equation 4.1.

a Degradation rate of the intracellular PHA for the second (late) degradation phase (hours 4 till 8). The degradation rate was calculated as described in c.

e The half life time of the intracellular PHA in the “early phase” was calculated with equation 4.2. f The degradation of the intracellular PHA was exponential only in the beginning of the experiment. For

subsequent calculations only these data points were used. nd.: not determined.

The relative elemental compositions of the total biomass and of the residual biomass are

given in Figure 4.3. For the biomass, the weight percentage of carbon (C) and hydrogen (H)

decreased, whereas the weight percentage of nitrogen (N) content increased in both cultures.

The mass/mass ratios of N/C increased over 5 hours by 27% for culture N and by 18.8% for

culture B, whereas the ratio of N/H increased by 27.7% (N) and by 19.8% (B), respectively.

In contrast, these ratios remained nearly constant for the PHA-free biomass (with a tendency

of a slight increase of the N/C ratio in the C starved culture) with an elemental composition of

40 - 41% carbon, 11 - 12% nitrogen, and 6% hydrogen (Tab. 4.3).


® 16 90 ...... ~ 14 85 0

0 E 12 ......

80 ~ ...... 0 a.. 0 Cl)

E 10 75 E ...... 0 a.. c: 8 Cl) 0 70 E :!!: 0 I c 6 65 c:

"""" 0 () :!!: - I

I 4 60 co (Q () () C10

2 55

0 50 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Time [h]

3.5 @ -- 3 c: Cl) -c: 2.5

a.. 0 Cl) () 2 E <C o::c 1.5 c: a.. C6 0 "'

:!!: g 1 ~~-o - .()._ 0-0 -·-c: ~ .......... .... -o - .0.. ...J I/I 0.5 ~.--. _ - - - -0 0 c.. 0 -a E - -0 -() -0.5

D~- -------1 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Time [h]

FIG. 4.2: Monomer composition of PHA during intrncellular PHA degradation in starved P. oleovorans cells. The PHA of the cultures of Figure 4.1 was analysed. Panel a: Composition of the isolated PHA. Panel b: Cellular content of the PHA monomers. Culture B: (0 ), culture N: (• ). C6: 3-hydroxyhexanoate; C8: 3-hydroxy-octanoate; Cl O: 3-hydroxydecanoate.

@ 0.35

-.... b> Cl 0.30 ...... 0 :;:; ~ 0.25 c 0 .c ... ~ 0.20 0 -c Q) Cl 0.15 0 ...

;1: z

-'";' Cl Cl ...... 0 :;:; ~

0.10

2

1.8

1.6

c Q) Cl 1.4 2

11 5

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Time [h]

6 6 6 - -0 -ts - - - ~ -6- - - -~~----~----~-----0 0

0

"O >. .c 1.2 -r-• 0 -; 1 Cl 0 ...

;1: z 0.8

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Time [h]

Chapter 4

FIG. 4.3: Elemental composition of total and PHA-free biomass in aP. oleovorans culture dming PHA degradation. The cultures of Figure 4. 1 were analysed and elemental compositions were expressed as a mass/mass ratio of the nitrogen to carbon content (Panel a), and the nitrngen to hydrogen content (Panel b) . ( • ) carbon starved total biomass (culture N); ( 6 ) carbon starved PHA-free biomass (culture N); C• ) simultaneously nitrngen and carbon starved, total


biomass (culture B); ( ) simultaneously nitrogen and carbon starved, PHA-free biomass (culture B).

Chapter 4 117

TAB. 4.3: Average and standard deviation of carbon, hydrogen, and nitrogen composition of residual (PHA-free) biomass under PHA degrading conditions. a

B 40.2 ± 1.37 6.1 ± 0.37 11.0 ± 0.50 8 N 41.3 ± 1.37 6.1 ± 0.25 11.7 ± 0.29 12

P 43.5 ± 2.80 6.5 ± 0.43 12.1 ± 0.79 8 Q 43.7 ± 1.79 6.4 ± 0.34 11.6 ± 0.41 8 R 43.4 ± 1.00 6.4 ± 0.16 12.4 ± 0.28 12 T 44.1 ± 1.10 6.5 ± 0.19 12.2 ± 0.38 10

a The elemental composition of residual biomass was calculated on the basis of the measured PHA content and the elemental composition of lyophilized total biomass.

b Degradation conditions: N, P: carbon starved (N added); B, Q: simultaneously carbon and nitrogen starved; T: simultaneously carbon and nitrogen starved, mRNA synthesis blocked with rifampicin; R: carbon starved (N added), mRNA synthesis blocked with rifampicin.

c PHA-free biomass

Effect of a nitrogen pulse on the degradation of PHA in initially nitrogen limited

cultures (cultures P, Q, R, T)

To determine whether the degradation of PHA depends on the induction of the synthesis of

PHA-depolymerase or whether it occurs simultaneously with PHA synthesis due to the

presence of constitutively produced depolymerase, as is the case for Alcaligenes eutrophus

(Doi et al., 1992; Doi et al. 1990), the synthesis of protein was blocked at the mRNA level

with rifampicin prior to degradation tests (cultures R and T). As a control, cultures with no

antibiotic were followed with and without the addition of nitrogen (cultures Q and P,

respectively). The time-courses of the optical density and the PHA content of the cells are

shown in Figure 4.4. Two time phases could be clearly distinguished:

The early phase (t = 0 to 4 h) included the change of all cultures from nitrogen only (0.61

mg C l-1 in the supernatant) to dual (C,N) starvation (no carbon detected at t = 0.62 h). All

cultures had the same initial PHA content of 24% of the cell dry weight. An exponential

decrease of the polymer content was observed for all cultures until hour 4, ranging from

20.3% to 3.2% for culture P, from 22.5% to 10.3% for culture Q, from 22.7% to 16.7% for

culture R, and from 23.2% to 13.1% for culture T.

Batchb Carbon (% of Rc)

Hydrogen (% of Rc)

Nitrogen (% of Rc)

Number of data points


@ 1.8

1.7

e- 1.6 c: c ~ 1.5 c 0 -.5 1.4

1.3

1.2 0 1 2 3 4 5 6 7

Time [h]

® 4

3 -...... 3: -c 2 a; () • .:a =I=

--c: ns o- 1 () .s ct 0 ~ ct 0 -::c .5 c.. ~ -1 -

~ ~

0 \

. _.__ • \

\

-2 °' \0

-3 0 1 2 3 4 5 6 7

Time [h]

FIG. 4.4: Degradation of intracellular PHA in P. oleovorans.

5.00

4.00

3.32 8

54.6 30.0 20.0 10.0

5.0

1.0

0.05 8

E c: c U') "It' c 0

-3: c c: () Q) -- ns c: -0 0 () --ct 0

::c ct c.. ::c c.. ~ 0 -

The cells were grown in a chemostat (D = 0.32 h-1) under nitrogen limitation with

a residual Na-octanoate concentration of 63 mg r1. The feed was stopped and 250

ml aliquots of the culture were trnnsfe1Ted to four 1 1 shake flasks. Nitrogen and/or rifampicin were added to the isolated cultures resulting in 4 cultures: Culture P (• ): C starved; culture Q (0 ): C, N starved; culture R (II): C starved

Chapter 4 119

and protein synthesis blocked by rifampicin; culture T ( ): C, N starved and protein synthesis blocked by rifampicin. Panel a: Time-course of the optical density. Panel b: Time-course of the cellular PHA content.

The fastest degradation rate with r(PHA) = 0.49 h-1 was observed for culture P which was

only carbon limited. The dual (C, N) limited cells (experiment Q) degraded PHA at r(PHA) =

0.27 h-1, whereas the rifampicin treated cultures (R, T) showed identical degradation rates of

r(PHA) = 0.16 h-1. This degradation was obviously not enhanced by a nitrogen pulse as was

the case for culture P.

During the late phase (t = 4 till 8 h) the rifampicin treated cultures (R, T) no longer

degraded PHA. Their PHA content remained at a high level of 16.7% for culture R and

13.1% for culture T, respectively. The carbon only starved cells (culture P), however,

continued to degrade the remaining PHA (3.2% w/w) at a slightly slower rate of r(PHA) =

0.37 h-1. A surprising effect could be observed for the dual (C, N) starved culture. The PHA

content at t = 4 h (10% PHA) was degraded at a considerably higher rate than in the early

phase. A PHA degradation rate of r(PHA) = 1.5 h-1 was calculated.

The elemental composition of the PHA-free biomass remained constant over 8 hours and

did not differ much between the four cultures (Tab. 4.3). On average the PHA-free cell mass

consisted of 43.7% carbon, 6.5% hydrogen, and 12.1% nitrogen.

DISCUSSION

This report shows that the degradation of PHA in P. oleovorans occurred exponentially when

harvested chemostat cultures were starved for carbon and nitrogen (Figs. 4.1 and 4.4). Two

time phases of the PHA degradation could be distinguished, an early phase for the first 4-5

hours and a late phase between 4 and 8 hours.

Observations in the early phase of PHA degradation

The cells degraded PHA immediately when the carbon source in the medium was consumed.

Rifampicin reduced the rate of PHA degradation but did not eliminate it completely. The fact

that there was depolymerization while protein synthesis was inhibited, indicates that


depolymerase was present and active during continuous growth, before the transfer of these

chemostat cultures to flasks in experiments R or T. Thus, there was simultaneous

accumulation and degradation of PHA in the chemostat for a net rate constant of PHA

formation of 0.32 h-1. Since we measured a degradation rate of the order of 0.16 h-1, the rate

of PHA synthesis during continuous growth was probably in the order of 0.4 - 0.45 h-1. A

similar conclusion, based on a different experimental approach, was reported by Doi et al.

(1990). They showed that the accumulation and degradation of PHB in Alcaligenes

eutrophus occurred simultaneously. In contrast to these findings, Kim et al. (1997) found that

P. oleovorans grown in a batch culture with sequential feeding of nonanoic acid (NA) and

10-undecenoic acid (UA) did not synthesise and degrade PHA simultaneously, since the

NA-PHA content in the culture remained constant after addition of UA.

The rate of PHA degradation in cultures without added antibiotic was almost doubled by

the addition of ammonium (Tab. 4.2). Thus, cells starved only for carbon were able to

degrade PHA efficiently. Either the depolymerase is activated by added nitrogen, or more

depolymerase is synthesized in the presence of added nitrogen. The latter option is supported

by the fact that inhibition of protein synthesis reduced the rate of PHA degradation to that

found in the absence of nitrogen. Support for this notion comes from Foster (1996) who

showed that the addition of depolymerase to isolated granules of P. oleovorans increased the

degradation rate from 0.93 mg PHA h-1 to 1.3 mg PHA h-1.

Degradation of PHA monomers

3-Hydroxyhexanoate and 3-hydroxyoctanoate were released at equal rates (Tab. 4.2). In

contrast, 3-hydroxydecanoate was clearly not degraded (Fig. 4.2b). This either means the

esterase does not have the possibility to split off PHA-C10 monomers, or these monomers are

not a component of the PHA, but rather of the PHA-free biomass. The latter assumption is

supported by the data of Durner (1998). They found that the C10 monomer is detected in the

PHA assay when lyophilized cells were boiled with sulfuric acid in the presence of methanol

and chloroform, but not in the PHA assay of chloroform extracted PHA. In the first

publication about the PHA of Pseudomonas oleovorans (de Smet et al., 1983), C10

monomers were also detected only as cellular and not PHA as components. The C10

Chapter 4 121

monomers detected in subsequent studies (Huisman et al., 1991; Lageveen et al., 1988) were

probably found because whole cells were used for the PHA assay in these experiments.

Foster et al. (1996) measured the monomeric composition of PHA granules isolated from

P. oleovorans. They found 92.5 mol% 3-hydroxyoctanoate and 7.5 mol%

3-hydroxyhexanoate, but no 3-hydroxydecanoate. However, they reasoned their observation

was related to a special feed strategy (intermittent feeding of octanoic acid).

Elemental cell composition

The elemental constituents of the total biomass changed due to the degradation of PHA (Fig.

4.3). The weight percentage of nitrogen increased, whereas carbon and hydrogen decreased

over the time-course. However, considering the residual (PHA-free) biomass the elemental

composition remained constant. The nitrogen sufficient cultures showed a slightly higher

average nitrogen content in the residual biomass compared to the nitrogen deficient cultures

(B, Q). However, this difference is within the range of errors.

Observations in the late PHA degradation phase

An interesting phenomenon could be seen in the second time phase (4 - 8 hours). The carbon

only starved culture continued to degrade PHA at a slightly lower rate (shift from r(PHA) =

0.49 h-1 to 0.37 h-1), whereas the dual (C,N) starved culture showed an increase in

degradation. This second degradation phase was not observed in the corresponding culture B,

perhaps due to the shorter analysis time of less than 5 hours. An explanation for the faster

degradation might be that after lysis of cells nitrogen containing nutrients became available

again for other cells. Such cryptic growth (growth on cell debris) was also observed when

Pseudomonas putida KT2442 was starved for phosphorus for 5 days (Eberl et al., 1996).

The antibiotic treated cultures (R, T) ceased PHA degradation altogether after 3.5 hours

and the PHA content remained high during the late PHA degradation phase. Why the

degradation stopped is not understood. Either an additional protein set (phasins or other

unknown helper proteins) is needed for a complete degradation or, more likely, the

intracellular depolymerase was simply inactivated after 3.5 hours and no new depolymerase


was synthesised. Hence, it would be interesting to determine whether the regulation of the

degradation occurs at the transcriptional level, post-synthesis, or via enzyme activity

regulation (phosphorylation/ dephosphorylation).

Rates of PHA degradation under various medium conditions

Within the first 4 - 5 hours the PHA decay rate constants varied between 0.14 h-1 (B) and 0.27

h-1 (Q) for dual (C, N) limited and between 0.31 h-1 (N) and 0.49 h-1 (P) for carbon only

limited cultures. The observed difference in the degradation rates under similar starvation

conditions could be due to one or more of the following conditions:

a) The starting conditions were not identical. Both chemostat cultures, the sources for the

experiments B, N and P, Q, R, and T, respectively, grew at different growth rates (D =

0.29 h-1 and 0.32 h-1) and under different nutrient limitations (dual (C,N) and N only).

Thus, cultures Q, P, R, and T were starved dually (C,N) only after 0.61 hours. This time

might have been sufficient for the cells to better prepare for nutrient starvation.

b) Smaller granules may be degraded faster (surface to volume ratio). An influence of the

substrate particle size has been described in the shrinking particle model of

Blancarte-Zurita and Branion (1986) that has been set up for the sulfur leaching bacterium

Thiobacillus ferrooxidans; particles with a lower volume to surface ratio are degraded

faster. Since nothing is known about the amount, activity, and distribution of PHA

polymerase on or near PHA granules in P. oleovorans, the influence of these variables on

the PHA degradation rate cannot be assessed. Possibly small variations in PHA content

and granule size might enhance or reduce granule degradation.

c) The degradation of the PHA might be slowed down by the C10 content of the cells, as the

faster degrading cells contain only traces of C10 (Table 4.2). Thus, there might be

inhibition of the depolymerase by cellular C10 monomer containing components. It would

be interesting to determine whether PHA granules produced during growth on

3-hydroxydecanoate are degraded at a much slower rate.

What is the advantage of PHA accumulation for P. oleovorans

Chapter 4 123

There are several reasons why PHA accumulation is advantageous for P. oleovorans. First,

the simultaneous accumulation and degradation described in this report provides an effective

mechanism to handle fluctuations of the carbon concentration in the environment. Second,

mobilisation of stored carbon on sudden carbon starvation occurs rapidly and might help the

cell to prepare for a long term starvation, since it is known that carbon starvation induces a

cascade of genes in fluorescent pseudomonads (Givskov et al., 1994a; Givskov et al., 1994b;

van Overbeek et al., 1995). Thereafter, the cells are less sensitive towards a wide array of

adverse conditions, such as pH, temperature, and osmotic stress. Third, a large PHA content

can be advantageous for survival during long-term starvation. Matin et al. (1979) found that

the survival of a Spirillum sp. after 30 hours of carbon starvation was linearly related to the

initial PHB contents. To date, no such experiment has been reported for fluorescent

pseudomonads: it will be interesting to determine if similar results are found for

pseudomonads.

124

CHAPTER 5

A JOURNEY TO THE UNKNOWN: TRIPLE (C,N,P)

LIMITATION OF PSEUDOMONAS OLEOVORANS IN

CHEMOSTAT


Keywords: Chemostat, multiple nutrient limitation, PHA, protein content, growth yield,

model.

125 Chapter 5

SUMMARY

Pseudomonas oleovorans is able to accumulate a biodegradable polyester,

poly(3-hydroxyalkanoate) (PHA). P. oleovorans was grown in a series of 6 chemostats at a

dilution rate of 0.2 h-1 to investigate whether this organism is able to grow under triple (C,N,P)

limited growth conditions and simultaneously accumulate PHA to higher amounts. By the

variation of the carbon (C, octanoate), nitrogen (N, ammonium), and phosphorus (P,

di-potassium hydrogen phosphate) concentration of the feed medium resulted in three groups

of nutrient limited cultures: firstly, single limitations by C, N, or P, secondly, dual (C,N),

dual (C,P), or dual (N,P) limitation, and thirdly, triple (C,N,P) limitation, here for the first

time reported.

Analysis of the data revealed that the triple limited growth regime can be described more

accurately, based on the idea that it is a superposition of dual (C,N; C,P) limited growth

regimes. Thus, triple limited growth should be found within the range of the following

nutrient ratios of the medium feed: 6.6 g g-1 < Cf/Nf < 11.8 g g-1 and 31 g g-1 < Cf/Pf < 78 g g-1,

which was in fact the case for all 6 triple limited chemostat cultures. All triple limited

steady-state cultures had a higher PHA content than cultures grown under dual (C,N) or dual

(C,P) limited growth conditions. However, triple limited growth did not lead to the maximum

PHA content of N only limitation (45.1% of cell dry weight) and was comparable to P only

limited growth. The average monomeric composition of PHA in triple limited cells was 85 ±

1.5 mol% 3-hydroxyoctanoate and 15 ± 1.5 mol% 3-hydroxyhexanoate. The PHA monomer

3-hydroxdecanoate was only found for a few N limited cultures.

INTRODUCTION

In addition to abiotic growth requirements, microorganisms require nutrients in sufficient

amounts. To cultivate microorganisms in an artificial environment, microbiologists have

designed synthetic media which satisfy all the needs for maximum biomass productivity. In

general, a single nutrient is selected that limits the maximum biomass that can be produced

Triple (C,N,P) limited growth 126

using such a synthetic medium. However, mathematical modelling and chemostat

experiments, it has become clear that dual limited growth can occur in pure cultures when

several key nutrients in the feed medium are supplied in a specific mass/mass ratio (Egli,

1991; Egli, 1995; Egli and Quayle, 1986; Minkevich et al., 1988).

Recently, it has been shown that Pseudomonas oleovorans is able to grow in continuous

cultivation under simultaneous limitation of nitrogen (ammonium) and carbon (octanoate)

(Chapter 3 and Durner, 1998). This observation has been made for steady-state chemostat

cultures, where the C/N ratio in the feed medium (Cf/Nf) was changed by the increase of the

C concentration. The experimental results showed that P. oleovorans did not change from

carbon to nitrogen limitation in a step function. Instead, there is an intermediate regime,

where both nutrients are consumed completely (schematically shown in Figure 5.1a). This

was named the dual (C,N) limited growth regime. Under dual (C,N) limited growth

conditions poly(3-hydroxyalkanoate) (PHA) was accumulated as an intracellular carbon

storage compound and the amount formed showed a linear relationship to the Cf/Nf ratio in

the growth medium. Under sole carbon or sole nitrogen limited growth conditions the PHA

content remained constant. As a consequence the yields for the biomass on used carbon

(YX/C), on used nitrogen (YX/N), and the ratio of both parameters (YX/N/YX/C) changed only in

the dual limited growth regime (Fig. 5.1b).

One can wonder whether more than two nutrients can limit growth simultaneously. Since

multiple limited growth is thought to be a general phenomenon in nature, Egli (1991)

postulated that triple limited growth might be observed in chemostat cultures. Because dual

limited growth had been shown for the nutrient pairs nitrogen and carbon (Duchars and

Attwood, 1989; Egli, 1982; Rutgers et al., 1990), nitrogen and phosphorus (Cooney and

Wang, 1976b), and carbon and phosphorus (Lucca et al., 1991), the potential nutrients to

enable such a particular triple limited growth regime for P. oleovorans are the three elements

carbon, nitrogen, and phosphorus (P). In addition, Lageveen et al. (1988) showed that P.

oleovorans is able to accumulate PHA to more than 30% of the cell dry weight under P

limitation in batch cultures with octane as only C source.

According to Egli (1991) the triple limited growth regime can be understood as a

superposition of dual (C,N) and dual (C,P) limited growth regimes. It is known that specific

growth rate is an important determinant of the position of the dual (C,N) limited growth

127 Chapter 5

regime. For instance, the faster P. oleovorans was grown under N limitation in a chemostat,

the less PHA was accumulated (Dmner, 1998) . Because the PHA content achieved was lower,

the dual (C,N) limited regime was resti·icted to a smaller range of Ct!Nr units. The opposite

was the case when the cells were grown at a slower growth rate. Moreover, more of the

carbon source was spent for maintenance pmposes. Consequently, when the cells were

cultured more slowly, the dual limited growth regime was shifted towards higher Ct!Nrratios.

The same effects should be observed for the dual (C,P) limited growth.

x z u

~ .,.... I O>

.2? z x >-

C limitation Dual (C,N) N limitation limitation

y X/N/Y X/c>Cf"Nt y X/N/Y x1c- Cf"Nt y X/N/Y X/c<Cf"Nt

® Yx/N

YX/c ,,.--

/ /

/

/ /

/ // c

------

~ 0 u -0

<( I a..

~

N I O>

N .2?

(.)

<5 - - - - _,,.,,,,, >'!"" ......

x c x ------ z >- YX/N/YX/C ...... x ...... >-- - - -

FIG. 5.1: Dual (C,N) limited growth in a chemostat culture of Pseudomonas oleovorans as a function of the carbon to niti·ogen ratio in the feed medium (Ct!Nr). Panel a: With increasing carbon feed concenti·ation the biomass (X) increases until nih'ogen (N) is the sole limiting nuh'ient. The poly(3-hydroxyalkanoate) (PHA) content of the cells starts to increase when the nih'ogen source becomes limiting and remains constant after residual carbon (C) becomes detectable in the supernatant.


Panel b: The biomass yields on carbon (YX/C) and on nitrogen (YX/N) are affected by the accumulation of poly(3-hydroxyalkanoate) (PHA). The ratio of the yields YX/N/YX/C approximates the Cf/Nf ratio under the dual (C,N) limited growth regime only.

129 Chapter 5

The supe1position of the dual (C,N) over the dual (C,P) limited growth conditions would

ideally result in triple limited growth regimes. How these growth conditions have to be

understood is seen in Figure 5.2. The pyramid depicted represents the triple limited growth

regime such that points inside the pyramid and on its surface are limited by C, N, and P

simultaneously. Triple limited growth conditions are clearly determined by the Cr/Nr, the

Cr/Pr ratio in the feed medium, and the dilution rate (D). It can be seen that the medium

composition that results in triple limited growth at a low growth rate does not necessarily

cause triple limited growth at high growth rates. Consequently, we chose to use a moderate

constant dilution rate of 0.2 h-1 for our experimental journey to the triple limited growth

regime which is described in this chapter.

Dilution rate

C1· 'll1itat· I On

Dua1 (C N . , ) hrnitat·

I On N1· irnitat· e--- ion - - - -.. -

FIG. 5.2: Triple (C,N,P) limited growth of Pseudomonas oleovorans as a function of the dilution rate, the carbon to nitrogen (Cr/Nr), and the carbon to phosphorns (Cr!Pr)


ratio of the feed medium. The pyramid represents assumed triple (C,N,P) limited growth. The cross-cut through the pyramid which is depicted with dashed lines represents the conditions of a constant dilution rate.

131 Chapter 5


Microorganism

The Gram-negative bacterium Pseudomonas oleovorans (ATCC 29347) was used in all

experiments. For the production of frozen stock cultures an exponential culture with medium

X was mixed with 30% glycerol (1:1 v/v) and stored in 1 ml portions at -80°C. The cells were

revived in medium X.

Growth conditions

Pseudomonas oleovorans was cultivated in a 3 l stirred tank reactor (Wubbolts et al., 1996).

The working volume of the bioreactor was weight-controlled to 1.5 l by a scale-controlled

waste pump. The feed pump ran continuously at a constant rate of 0.3 l h-1. The aeration was

adjusted to 0.6 vvm. The temperature was set to 30°C and the pH was held at 7.0 ± 0.02 by

automatic addition of either 4 N H2SO4 or 2 N NaOH.

Culture medium

For precultures the minimal growth medium X was used. This mineral medium contained

(per liter): 3.5 g NaNH4HPO4*4H2O; 7.5 g K2HPO4*3H2O; 3.7 g KH2PO4; 2.0 g citric acid

monohydrate. After heat sterilization and cooling to room temperature the following

minerals were added: 1 ml of 1 M MgSO4*7H4O, 1 ml of 0.01 M FeSO4*7H2O in 1N HCl,

and 1 ml of mineral trace element (MT) stock solution that contained (per liter 1N HCl): 1.98

g MnCl2*4H2O; 2.81 g CoSO4*7H2O; 1.47 g CaCl2*2H2O; 0.17 g CuCl2*2H2O; 0.29 g

ZnSO4*7H2O. Medium X is carbon limited up to 2.5 g l-1 citric acid.

Medium for the continuous cultivation

The basic medium consisted of the following components (per liter): 2.13 g morpholine

propane sulfonic acid (MOPS) (used as pH-buffer), 0.01 g ethylenediamine tetra acetic acid

(EDTA), 1 ml of 1 M MgSO4*7H2O, 2 ml of 0.01 M FeSO4*7H2O in 1 N HCl and 1 ml of

MT stock solution. The carbon source (octanoate), the nitrogen source (ammonium sulfate),


and the phosphorus source (K2HPO4) varied in concentration as described later in the

individual experiments. The medium for 50 l was prepared in a 20 l bottle, where pH was

adjusted with 10N NaOH and 4N H2SO4 to 7.1. Thereafter the medium was filter-sterilized

into a sterile 50 l tank and filled with sterile water up to 50 l.

During the experiment the medium composition was changed by the addition of

concentrated sterile carbon (0.5 M octanoate), nitrogen (0.5 M ammonium sulfate), or

phosphorus (1 M K2HPO4) stock solution to achieve triple limited growth. For each

steady-state measurement a sample of the medium was taken and analysed off-line. The

residual concentrations of C, N, and P in the supernatant were then used to calculate the

potential in biomass gain when the residual concentration would be converted into biomass

completely.

The potential biomass gain was calculated as follows:

[5.1]

[5.2]

[5.3]

where ∆XC, ∆XN, and ∆XP are the potential biomass gains that can be theoretically produced

with the residual nutrient concentrations of the supernatant, such as carbon (C), nitrogen (N),

and phosphorus (P). The applied biomass yields were as follows: YX/C = 0.7 g g-1, YX/N = 6.6

g g-1, and YX/P = 30 g g-1. The largest potential biomass obtained in equations 5.1 to 5.3 was

named ∆Xmax and then used for further calculations. Finally, the necessary nutrient

concentrations that would be needed to generate ∆Xmax can be calculated as follows:

[5.4]

[5.5]

[5.6]

133 Chapter 5

where ∆Cf, ∆Nf, and ∆Pf are the concentrations of carbon, nitrogen, and phosphorus,

respectively that had to be increased in the feed medium by that specific amount with the

addition of the respective stock solution. In order to avoid an unknown limitation by another

nutrient the medium was modified such that the biomass did not exceed 2.5 g l-1.

Biomass determination via optical density

The optical density of the culture broth was measured with a spectrophotometer (Pharmacia,

Uppsala, Sweden) at a wavelength of 450 nm against 10 mM MgSO4. For OD 450 nm > 0.3,

samples were diluted with 10 mM MgSO4. The cell dry mass was estimated using a

correlation factor of 0.29 mg ml-1 per OD 450 of 1.

Biomass determination by cell dry weight measurements

Polycarbonate filters (Nuclepore, 47 mm and 0.2 µm) were dried at 80°C for two days and

tared after cooling down in a desiccator to room temperature. Based on the OD450

measurements cell dry mass (5 mg) was filtered by vacuum through a tared filter. Filters were

washed with the same volume of 10 mM MgSO4, dried at 80°C for at least 3 days, and

weighed after one additional day in a desiccator (identical results as with 105°C dried filters).

PHA-measurement

PHA content of freeze dried cell samples (5 mg) was determined according to the method

described by Lageveen et al. (1988).

Carbon concentration

The Na-octanoate concentration of the medium and the supernatant was measured by gas

chromatography (HP5890, Hewlett Packard, Camas, WA, USA) equipped with a carbowax

column (CW20M, 25 m x 0.25 mm, Marchery-Nagel, Germany). The same temperature

program was used as described in chapter 3.


Nitrogen concentration

Nitrogen was measured according to the method of Scheiner (1976). The detection range

was between 0.2 and 2 mg N l-1 and the detection limit was below 0.15 mg l-1.

Phosphorus concentration

The phosphorus concentration in the medium and in the supernatant was determined

spectrophotometrically according to the method published by Chen et al. (1956). The range

of the assay was between 0.2 and 2 mg P l-1 and the detection limit was below 0.1 mg l-1.

Chemicals

Chemicals from Merck (Darmstadt, FRG) were used throughout the experiment and were

>99% pure.

135 Chapter 5

RESULTS

A series of 6 chemostat experiments at a dilution rate of 0.2 h-1 were used to achieve triple

limited growth (Tab. 5.1). To obtain other growth conditions the medium feed composition

was changed through the preparation of a new medium tank, or through the addition of sterile

stock solutions of carbon, nitrogen or phosphorus to the medium. The amount of nutrients

that had to be added were calculated based on the equations 5.1 to 5.6, consequently, the

desired triple limited growth regimes were achieved on a try and error basis.

Nutrient limitations

A culture was considered to be carbon limited when the inflowing carbon was consumed to a

residual carbon concentration that was below the detection limit (< 8 mg C l-1). Nitrogen and

phosphorus limitation occurred as residual concentrations of nitrogen in the culture fell

below 0.6 mg N l-1 and 2 mg P l-1, respectively.

Thus, based on the residual nutrient concentrations, the measured steady-states could be

divided into three different classes of growth regimes: first, 10 single limitations (C, N, and

P), second, 8 dual limitations ((C,N), (C,P), and (N,P)), and third, 6 triple limitations (C,

N,P).

To obtain an overview on the nutrient ratios in the feed medium, the steady-states are

presented in a two-dimensional plot with Cf/Pf on the X-axis and Cf/Nf on the Y-axis (Fig.

5.3). Theoretically this configuration represents a cross-cut through the pyramid (triple

(C,N,P) limited growth regime) shown in Figure 5.2 at a constant dilution rate of 0.2 h-1. In

fact, all triple limited steady-states were found within a Cf/Nf range of 8.25 and 11.46 g g-1

and a Cf/Pf range of 40.78 to 64.97 g g-1. Interestingly, both dual (N,P) limited steady-states

were located in this triple limited area, however, they were clearly not carbon limited.

PHA content under different nutrient limitations

Figure 5.3b shows a graphical overview of the PHA content of the cells as a function of the

nutrient ratios in the feed medium. In order to present the PHA contents they were divided

into classes: 1.: 0-10%, 2.:10-20%, 3.: 20-30%, 4.: 30-40%, and 5.: >40% PHA of the cell dry


mass (Fig. 5.3b). Thus, it can be seen that the PHA content of P. oleovorans was dependent

on the composition of the medium. The cellular amount of PHA was highest (Tab. 5.1) with a

maximum of 45.1% under nitrogen only limited conditions, and lowest with a PHA content

of 4.2% under C only limitation.

16 15 • 14 • 13 • 12 • • 11 x ++• •+ + .,.... 10 + I • O> 9 .2? 8 + -z 7 )K - )K u 6 5 4 3 2 @ 1 0

15 ~ 14

13 @) 12

·~~ @) ~ 11 @)

~ 10 ~ @) b> 9 .2? 8 ~ - @) z 7 - 6 @)

~ u 5 4 3 0

2 @) @ 1 0

@)

0 0 20 40 60 80 100 120 140

Ct !Pt [g g-1]

FIG. 5.3a: Determined nutrient limitations of P. oleovorans grown in chemostats (D = 0.2 h-1) as a function of the carbon to nitrogen (Cfi'Nc) and carbon to phosphoms (CtlPc) ratio in the feed medium. The following symbols indicate specific types of limitations: carbon only(• ), phosphoms only (6 ), nitrogen only (II), dual

137 Chapter 5

carbon-nitrogen ( ), dual carbon-phosphorus ( ), dual nitrogen-phosphorus ( ), and triple carbon-nitrogen-phosphorus limitation ( ).

5.3b: The poly(3-hydroxyalkanoate) (PHA) content of P. oleovorans is dependent on the carbon to nitrogen (Cf/Nf) and on the carbon to phosphorus (Cf/Pf) ratio in the feed medium. The growth conditions of the culture are plotted with respect to the Cf/Pf and the Cf/Nf ratio. The corresponding PHA content of the cells are grouped into subclasses. The symbols represent the weight of PHA per cell dry mass: 1 circle: 0-10 % PHA, 2 circles: 10-20 % PHA, 3 circles 20-30 % PHA, 4 circles: 30-40 % PHA, and 5 circles >40 % PHA (w/w).

TAB. 5.1: Utilization of C, N, and P by Pseudomonas oleovorans cultivated in a series of 6 chemostat experiments (D = 0.2 h-1) under different limiting growth conditions.

Limita-tion

Nutrients in feed Nutrients in chemostat

CDW [g l-1]

PHA [% of CDW]

YX/C [g g-1]

YX/N [g g-1]

YX/P [g g-1]

Cf [g l-1]

Nf [g l-1]

Pf [g l-1]

Cf/Nf [g g-1]

Cf/Pf [g g-1]

C [g l-1]

N [g l-1]

P [g l-1]

C, P 1.0030 0.8723 0.0284 1.15 35.33 <dl 0.6783 0.0010 1.10 8.9 1.10 5.67 40.18 C, P 1.3746 0.8698 0.0283 1.58 48.61 <dl 0.6866 0.0008 1.15 18.7 0.84 6.28 41.88 C, P 2.1337 0.8177 0.0276 2.61 77.20 <dl 0.6597 0.0010 1.11 19.8 0.52 7.03 41.59

P 2.5204 0.4370 0.0237 5.77 106.21 0.4680 0.2220 0.0004 1.99 33.6 0.97 9.26 85.40 N 2.7500 0.1741 0.2364 15.80 11.63 0.4420 <dl 0.1866 1.64 37.8 0.71 9.42 32.93 N 2.3846 0.1620 0.2359 14.72 10.11 0.2760 <dl 0.1713 1.62 45.1 0.77 10.00 25.08 C 0.6283 0.1781 0.2360 3.53 2.66 <dl 0.0817 0.2160 0.78 4.2 1.24 8.09 39.00

C, P 1.6244 0.2217 0.0337 7.33 48.20 <dl 0.0044 0.0006 1.89 17.7 1.16 8.70 57.10 C, P 1.6096 0.2448 0.0348 6.58 46.25 <dl 0.0011 0.0008 1.93 17.2 1.20 7.92 56.76 N 2.8475 0.2349 0.0384 12.12 74.25 0.0912 <dl 0.0124 2.00 10.2 0.73 8.52 77.07

C, N, P 2.4556 0.2494 0.0436 9.85 56.35 <dl <dl 0.0003 2.10 29.7 0.86 8.42 48.53 C, N, P 2.1337 0.2585 0.0487 8.25 43.82 <dl <dl 0.0003 2.23 32.1 1.05 8.63 46.09

N 2.4456 0.1782 0.2263 13.72 10.81 0.2333 0.0003 0.1870 2.18 42.8 0.99 12.25 55.41 P 3.1318 0.2168 0.0239 14.45 131.04 0.7854 0.0127 <dl 1.25 27.7 0.53 6.12 52.68 P 2.5613 0.2131 0.0286 12.02 89.55 0.2235 0.0149 0.0003 1.72 27.9 0.74 8.68 60.86 P 2.3952 0.2461 0.0303 9.73 79.05 0.4206 0.0471 0.0003 1.30 15.2 0.66 6.53 43.37 P 2.4764 0.1892 0.0298 13.09 83.10 0.0597 0.0035 0.0005 2.43 31.2 1.01 13.09 82.97

N, P 2.4735 0.2135 0.0523 11.59 47.29 0.0143 0.0003 <dl 2.45 32.2 1.00 11.49 47.05 C, N, P 2.4365 0.2234 0.0375 10.91 64.97 <dl 0.0004 <dl 2.50 35.7 1.03 11.21 66.77 C, N, P 2.3935 0.2193 0.0410 10.91 58.38 <dl 0.0006 0.0017 2.40 22.8 1.00 10.97 61.01

N, P 2.4402 0.2168 0.0462 11.26 52.82 0.0104 0.0004 0.0005 2.32 23.6 0.95 10.72 50.76 C, N, P 2.4365 0.2164 0.0545 11.26 44.71 <dl 0.0004 0.0009 2.35 27.3 0.96 10.88 43.84 C, N, P 2.4221 0.2114 0.0594 11.46 40.78 <dl 0.0004 0.0012 2.27 16.1 0.94 10.76 39.02

139 Chapter 5

C, N 2.3484 0.2069 0.0753 11.35 31.19 <dl 0.0004 0.0027 2.35 13.7 1.00 11.38 32.36 <dl: below detection limit


In general, it appeared that the accumulation of PHA was clearly favoured under N only

limiting growth conditions when a high residual P concentration (> 150 mg l-1) could be de-

tected. Interestingly, there was one N limited steady-state (Cf/Nf = 12.12 and Cf/Pf = 74.25 g

g-1) that had a PHA content of 10.2% with a residual P concentration of 12.4 mg l-1.

In contrast, the PHA content of P only limited cells appeared not to be influenced by the

residual concentration of N. It could be seen that the triple (C,N,P) limited cells didn’t lead

on average to a more efficient PHA accumulation than P limitation (Tab 5.2).

Monomeric composition of PHA

The analysis of the monomeric composition of PHA of cells grown under C, N, P only, and

triple limited growth revealed a general pattern of high 3-hydroxyoctanoate (>80 mol%) and

low 3-hydroxyhexanoate contents (< 16 mol%). Surprisingly, the monomer 3-hydroxyde-

canoate was only detected for two N limited and one C limited steady-states.

TAB. 5.2: PHA production of P. oleovorans under different nutrient limitations in a chemostat at a dilution rate of 0.2 h-1.

Relative composition of PHA in mol %

Limitationa Number of data points

PHA [% of CDW]

HHb HOb HDb

Carbon 1 4.2 9 84 6 Nitrogen 4 34.0 ± 16.14 13 ± 2.1 84 ± 4.6 3 ± 4.2

Phosphorus 5 26.1 ± 7.76 13 ± 0.0 87 ± 0.0 0 Triple (C,N,P) 6 27.3 ± 6.99 15 ± 1.5 85 ± 1.5 0

a Only the data points were considered that had a larger residual concentration of the non limiting nutrients than 10 mg l-1

b The PHA is composed of 3-hydroxyalkanoates: HH: 3-hydroxyhexanoate, HO: 3-hydroxyoctanoate, HD: 3-hydroxydecanoate.

DISCUSSION

141 Chapter 5

The experimental journey to triple limited growth went through 6 differently limited growth

regimes. These limitations were defined based on the residual nutrient concentrations in the

supernatant. Unfortunately, this approach focuses only on the very specific growth

conditions of the particular chemostat. As a consequence, a definition of the extent of the

growth regime does not seem to be possible with such an approach. Another approach will be

needed for a better understanding of the growth requirements required to achieve specific

nutrient limitations, e.g. triple (C,N,P) limited growth.

Determination of regimes of dual and triple nutrient limitations

The modified approach of Egli (1991) helps to estimate the growth conditions of the triple

limited growth regime with respect to the Cf/Nf and Cf/Pf ratio. According to the

stoichiometric approaches (Chapter 2) growth limiting nutrients are completely converted

into biomass under nutrient limitations. Thus, the following equations must be valid

simultaneously under triple limited growth conditions:

[5.7]

. [5.8]

In Figure 5.4 the experimental yield ratios YX/N/YX/C and YX/P/YX/C were plotted against the

nutrient ratios of the feed medium Cf/Nf and Cf/Pf, respectively. Theoretically all data points

of the plots that were on a line with the slope = 1 were dual (C,N) (Fig. 5.4a) or dual (C,P)

(Fig. 5.4b) limited.

The boundary between carbon and dual (C,N) or dual (C,P) limited growth was

determined graphically based on the fact that the yields are constant under C limitation at a

constant growth rate, as shown previously by Durner (1998) (see also Figure 5.1b).

To determine the transition point between dual (C,N) and N only limitation, the data

points which were N only limited were linearly correlated in Figure 5.4a and the junction

point of the resulting curve with the “dual limited curve” gave the transition point. The


procedure to determine the boundaiy between dual (C,P) and P only limitation was

perfonned with P only limited data points in Figure 5.4b. Thus, we found that dual (C,N)

limited growth can be expected for 6.6 g g-1 < Ct/Ne< 11.8 g g-1. The coITesponding dual (C,P)

limited regime would be found accordingly for 31 g g-1 < Ct/Pr < 78 g g-1.

Consequently, data points that were on both lines of Figures 5.4a and 5.4b should

represent triple (C,N,P) limited growth conditions according to the superposition principle

(Egli, 1991), which is in fact the case for all 6 triple limited steady-states.

14 13 @ x 12 11 10

~ 9 c 8 x z 7 x 6 >-

5 x 4 3 2 1 0

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Ctf Nt [g g-1]

120 @ • 100 • --

~ 80

c 60 a.. x >-40

20

0 0 20 40 60 80 100 120 140

Ctf Pt [g g-1]

143 Chapter 5

FIG. 5.4: Determination of dual (C,N), dual (C,P), and triple (C,N,P) limited growth regimes. A chemostat culture was dual nutrient limited when its yield ratio (YX/N/YX/C or YX/P/YX/C) was equal to the ratio of the corresponding nutrients in the feed medium (Cf/Nf or Cf/Pf) (Fig. 5.1). This is the case when in this figure the data points are on the continuous line. Data points that were located on both curves of Figs. a and b may theoretically be considered as triple (C,N,P) limited cultures. Five exceptions were detected: dual (C,N) lim. (1), dual (N,P) lim. (2, 3), and dual (C,P) lim. (4, 5) and were considered as potentially triple (C,N,P) limited steady-state cultures (see also Fig. 5.5)

Panel a: dual (C,N) limited growth. Panel b: dual (C,P) limited growth. The applied symbols are identical to those of Figure 5.3.

Surprisingly, not only were data points that represented triple limited growth on both lines

of Figure 5.4a and b, but also two dual (N,P), one dual (C,N), and two dual (C,P) limited

steady-states. One may now speculate that these cultures were physiologically triple (C,N,P)

limited although the third nutrient was clearly detectable in the supernatant. A further

indication that a physiological triple nutrient limitation might be the case is the fact that all

five steady-states were also within or very close to the predicted triple (C,N,P) limited

growth regime (Fig. 5.5). This means that our definition of nutrient limitation (C lim. < 8 mg

C l-1, N lim. < 0.6 mg N l-1, P lim. < 2 mg P l-1) might not be sufficient enough. An interesting

approach to solve this problem was previoulsy presented by Rutgers et al. (1990). They

investigated the dual (glucose,ammonium) limited growth of Klebsiella pneumoniae in

chemostat. They focused on flux of glucose and ammonium and showed that the control of

growth gradually shifted from glucose to ammonium within the dual (C,N) limited growth

regime and did not occur in a step function (see Fig. 2.5). The situation for triple (C,N,P)

limited growth is probably more complex since the nitrogen and/or phosphorus limitation

might additionally affect the uptake of octanoate.

Physiological flexibility of the cells towards nutrient limitations

The hypothesis that a limitation of nitrogen and phosphorus might decrease the uptake of

carbon certainly needs verification. However, one can speculate what kind of cellular

components are affected by a limitation of N and P. In Table 5.3 the distribution of total

nitrogen and total phosphorus are shown for an Escherichia coli cell (Neidhardt, 1987). The


data of E. coli may be used for P. oleovorans, because only minor changes in the cell

composition can be expected as both organisms are Gram negative.

The data in table 5.3 indicate that the RNA and DNA, and to a certain extent the lipid and

lipopolysaccharide content may be affected by nitrogen and phosphorus limitation in the first

place. It is known from literature that Pseudomonas putida KT2442 (Givskov et al., 1994b)

and Escherichia coli (Davis et al., 1986) have a more decreased viability under phosphorus

starvation than under carbon or nitrogen limitation. A way to handle the shortage of

phosphorus is to lower the RNA content of the cell. For instance, after one day under

phosphorus starvation P. putida KT2442 decreases the number of ribosomes per cell to a

tenth of its normal value (Eberl et al., 1996). This decrease of the ribosomes had also an

indirect influence on the protein synthesis. In addition, nitrogen limitation leads to a

decreased protein content (Herbert, 1961). These stringent limitations of the protein

synthesis may decrease the uptake of carbon. Recently, it was found that P. putida, a close

relative of P. oleovorans, has a specific constitutively expressed and energy dependent

uptake system for octanoate (Carnicero et al., 1997).

Prediction of nutrient limited growth regimes

Since the growth conditions resulting in triple limited growth regimes seem to be well

defined, one can wonder whether the other single and dual limited growth regimes can be

predicted:

Assuming that the boundaries of the triple limited growth would also be the boundaries of the

other limitations (superposition principle, Fig. 5.2), one can draw a scheme which presents

a pattern with 9 different growth regimes with respect to the Cf/Nf and Cf/Pf ratio (Fig. 5.5).

The borders of these regimes were to be found at Cf/Nf = 6.6 and 11.8 g g-1 and at Cf/Pf = 31

and 78 g g-1 (Fig. 5.5). However, the application of the superposition principle revealed that

not all zones shown in Figure 5.5 are defined sufficiently. In theory, it cannot be predicted

what kind of limitation would result in the superposition of N over dual (C,P) limited growth

(field A), N over P limitation (field B), and P over dual (C,N) limitation (field C). The

experimental data of this study suggest that the growth regimes in field A would represent

mainly N limitation and field C mainly P limitation. In field B only P limited steady-states

145 Chapter 5

were detected. However, it can be assumed that for increasing Cf/Nf ratios first a dual (N,P)

and then a N only limited growth regime has to be found.

TAB. 5.3: Calculated distribution of the total nitrogen and total phosphorus over the cell components of an E. coli B/r cell (data adapted from Neidhardt et. al., 1987)a.

Cell component Cellular content of the cell component [% of

cell dry weight]

Nitrogenb [% of total

nitrogen]

Phosphorusb [% of total

phosphorus]

Protein 55.0 67.9 - RNA 20.5 24.3 71.3 DNA 3.1 3.7 11.3 Lipids 9.1 0.9 14.5 Lipopolysachharides

3.4 0.4 2.9

Peptidoglycan 2.5 1.9 - Polyamines 0.4 0.9 -

a E. coli B/r was cultivated aerobically on a glucose minimal medium at 37°C with a mass doubling time of 40 minutes.

b Under these growth conditions E. coli B/r contains 14.4% nitrogen and 2.7% phosphate relative to total cell dry mass.


PHA content and composition of the cells

The PHA content of the cells (Fig. 5.3b) is a function of the composition of the nutrients in

the feed medium (Tab. 5.2). Interestingly, on the average the amount of accumulated PHA is

higher for N only (34% PHA w/w) than for P only (26.1 % PHA w/w) limited growth. The

average PHA content under triple (C,N,P) limited growth is in the order of the PHA content

of P only limited cultures (27.3% w/w). One can argue that the P limitation enables only

lower PHA contents and thus a larger PHA content under ti·iple limited growth conditions

seems not to be possible. However, a systematic study of the ti·iple limited growth regime

should give a better understanding of the above observations.

16

14

12 ~ .,.... 10 I O>

~ 8 -z ..._ -u 6

4

2

0

I • • I A 2 I B N limitation I 3~ ,/ I • ----~-~r~~;---r-~-------

Dual (C,N) I' + L Limitation I 1 Triple (C,N,P)f C

I + limitation I ____ L_~~---~---------1 ~54 I •

C limitation I Dual (C,P) I P limitation I limitation )If

Ix x I I I

0 20 40 60 80 100 120 140 Ct !Pt [g g-1]

FIG. 5.5: Comparison of predicted nuti·ient limited growth regimes of P. oleovorans with data of chemostat experiments. The growth conditions are indicated as functions of the carbon to phosphoms (Cc/Pr) and carbon to niti·ogen ratio (Ct!Nr) in the feed medium. Nine different growth regimes could be theoretically expected based on the principle of supe1position (Fig. 5.2). Six of these growth regimes are clearly detennined. However, for three cases (A, B, C) the growth regime cannot be predicted exactly. The numbers refer to potential physiologically ti·iple (C,N,P) limited steady-states (Fig. 5.2).

147 Chapter 5

The monomeric composition of the PHA is also influenced by the limitation. In the case of

P and triple nutrient limitation the monomer 3-hydroxydecanoate could not be detected at all.

In a previous study (Chapter 4) we showed that 3-hydroxydecanoate is a component of the

residual (PHA-free) biomass. An explanation for this observation has not yet been found.

148

CHAPTER 6

TOWARDS UNDERSTANDING OF DUAL NUTRIENT

LIMITED GROWTH OF PSEUDOMONAS OLEOVORANS:

A BLACK-BOX MODEL


Keywords: Black-box model, uptake, growth yield, growth rate, cell energy maintenance,

dual nutrient limited growth regime, chemostat, productivity.

Chapter 6 149

Black-Box Model 150

SUMMARY

Pseudomonas oleovorans accumulates poly(3-hydroxyalkanoate) (PHA) under ammonium

limited growth conditions, when octanoate, the sole carbon source, is in excess. In chemostat

cultures, P. oleovorans exhibits an intermediate growth regime between carbon (C) and

nitrogen (N) limited growth regimes, due to this ability to store carbon intracellularly. Within

this regime both nutrients (C, N) are used up to growth limiting amounts. This dual (C,N)

limited growth regime is dependent on the growth rate and the C/N ratio in the feed medium.

Based on experiments and published data, a black-box model has been set up, which is able

to predict under which conditions dual (C,N) limited growth with PHA production can be

expected. In the model the boundary between C only and dual (C,N) limited growth

((Cf/Nf)lower) is determined by the ratio of carbon and nitrogen uptake. The introduction of the

Pirt equation in the carbon uptake term resulted in an upward shift of (Cf/Nf)lower towards

higher C/N ratios for lower dilution rates. This is due to the fact that, as the portion of carbon

that is consumed for maintenance increases, the cells grow more slowly on a given amount of

carbon. The upper boundary is a function of the C storage (PHA) capabilities, which in turn

depends on the specific growth rate (µ). An equation with Moser kinetics of 2nd order gave a

good description of the PHA content. A conversion factor ß was introduced to express how

much carbon (C/N units) of the feed medium is needed to produce the predicted PHA content.

Thus, the black-box model was able to predict the dual (C,N) limited growth regime with

respect to the dilution rate and the Cf/Nf ratio. The model was used to calculate under which

conditions the highest specific PHA productivity and the most economical use of the carbon

source can be expected.

INTRODUCTION

Growth of microorganisms in nature and in artificial environments depends on an adequate

availability of nutrients. Efficient management of the available nutrients represents a first

adaptation to restricted growth conditions. Cell components, such as RNA, proteins,

Chapter 6 151

carbohydrates, and many others, are degraded and recycled. The energy source limits the

amount of biomass more strictly than other, also essential growth nutrients. Consequently,

several techniques have been developed by microorganisms to store energy when this source

is in excess.

For instance, Pseudomonas oleovorans and other fluorescent pseudomonads of the rRNA

homology group I accumulate a carbon and energy compound consisting of

poly(3-hydroxyalkanoates) (PHA) (Huisman et al., 1989). PHA is produced in P. oleovorans

when it is grown on medium chain length (mcl, C6-C14) alkanes, alkanols or alkanoic acids

under a nutrient limitation, such as nitrogen, phosphorus, sulfur, or magnesium (Lageveen,

Huisman et al., 1988). PHAs are getting more attention in the polymer industry because of

their unique properties that makes them into potential substitutes for the petrochemical

plastics, which have adverse effects on the environment (Byrom, 1987). At present, the

production costs of PHAs are still too high to permit their use as bulk plastics. However,

production costs can be significantly reduced by improving the bioprocess to maximize PHA

yield on substrates, PHA content, and volumetric productivity (de Koning et al., 1996; Lee,

1996b). In addition the medium costs should be minimized. A good approach is the

application of dual nutrient limited growth conditions, where the carbon and a nutrient source

are consumed completely. With respect to the production of PHA in P. oleovorans a good

carbon source would be a mcl alkanoic acid, e.g. octanoic acid, with ammonium as the only

nitrogen source. It is known from other chemostat experiments that the transition between

two nutrient limitations (e.g. from C to N limitation) is not a step function (Duchars and

Attwood, 1989; Egli and Quayle, 1986; Grätzer-Lampart et al., 1986; Minkevich et al., 1988),

but rather an intermediate growth regime where both nutrients limit growth simultaneously.

This regime can be graphically displayed as a zone with respect to the ratio of the two

nutrients in the feed medium. This zone depends on the specific growth rate of the cells (Egli,

1991; Minkevich et al., 1988), and can be predicted by the approach of Egli (Egli, 1991; Egli

and Quayle, 1986, and Chapter 2), which uses the biomass yields on the nutrients of interest

to calculate the boundaries of the dual limited growth regime (Fig. 6.1). This is possible

because the elemental cell composition differs significantly for each nutrient limitation

(Herbert, 1961; Herbert, 1976; Stouthamer, 1979).

Black-Box Model 152

A dual limited growth is likely to be found for P. oleovorans because of its ability to store

PHA intracellularly. The PHA content decreases as the growth rate of P. oleovorans

increases (Preusting et al., 1991), which will reduce the extent of the dual (C,N) limited

growth zone at higher growth rates. The opposite is expected for slower growth rates.

In this chapter we propose a black-box model to predict under which conditions P.

oleovorans has to be grown in order to attain dual (C,N) limited growth. Further, the highest

specific productivity and the most efficient usage of C source were determined.

1 0.9 0.8 0.7 Cir

~ ..... Nlr

I ..c 0.6 .......... Q) -ro ...... 0.5 c::: 0.4 0

:.;::::; 0.3 ::J

0 0.2 0.1 0

0 5 10 15 20 25 30 35

FIG. 6.1: Predicted dual (C, N) limited growth zone of Klebsiella aerogenes in chemostat cultures with different dilution rates and ratios of carbon (glycerol) to nitrngen (ammonium) in the feed medium (Cti'Nr) . K. aerogenes shows 3 different growth regimes: carbon limited at low Ctf'Nr ratios (Ch) , dual (C,N) limited growth regime (Dk), and sole nitrogen limitation at high Ctf'Nr ratios (Nk) separated by the lower Ctf'Nr boundaiy between Ck and Dk, or (Cti'Nr)iower, and upper Ctf'Nr boundaiy between Dk and Nk, or (Cti'Nr)upper· The data points were calculated by the approach of Egli (adapted from Egli, 1991).


Microorganism

Chapter 6 153

The strain P. oleovorans (ATCC 29347) was used for all experiments. Frozen stock cultures

were prepared from an exponential culture with medium X that was mixed with 30% glycerol

(1:1 v/v). Portions of 1 ml were stored at -80°C.

Culture media

Preculture

The minimal growth medium X for precultures contained E2-salts, MT micro elements

(Lageveen et al., 1988), and citric acid (per liter): 3.5 g NaNH4HPO4*4H2O; 7.5 g

K2HPO4*3H2O; 3.7 g KH2PO4; 2.0 g citric acid monohydrate. After sterilization we added

per liter 1 ml of 1 M MgSO4*7H2O, 1 ml of 0.01 M FeSO4*7H2O in 1 N HCl, and 1 ml of

MT stock solution that contained (in grams per liter 1 N HCl): 1.98 g MnCl2*4H2O; 2.81 g

CoSO4*7H2O; 1.47 g CaCl2*2H2O; 0.17 g CuCl2*2H2O; 0.29 g ZnSO4*7H2O. Medium X

is carbon-limited.

Batch culture

The medium X was modified for studies of batch cultures in the bioreactor, using 3.25 g l-1

Na-octanoate as sole carbon source and 2.24 g l-1 NaNH4HPO4*4H2O resulting in a C/N ratio

of 12.5 g g-1.

Continuous cultivation

Continuous cultivations were performed with medium XCC, which contained (in grams per

liter): 0.71 g (NH4)2SO4; 1.0 g KH2PO, and different concentrations of Na-octanoate. After

sterilization we added per liter: 1 ml 1 M MgSO4*7H2O, 1 ml of 0.01 M FeSO4*7H2O in 1 N

HCl, and 1 ml of MT stock solution.

Differently limited medium compositions (nitrogen, carbon or both), we obtained by

varying either the octanoate or the ammonium concentration of the medium.

Culture conditions

Black-Box Model 154

P. oleovorans was cultivated in a 3 l stirred tank bioreactor (Wubbolts et al., 1996) with a

working volume of 2.0 l. The aeration was adjusted to 0.6 vvm. The temperature was set to

30°C and pH was held constant at 7.0 ± 0.02 by automatic additions of either 4N H2SO4 or

2N NaOH.

In continuous cultures the volume of the culture was kept constant by an overflow device.

Generally, the bioreactor was inoculated with 25 ml of a preculture that grew to an OD450 =

3.5 overnight.

Analytical procedures

Optical density

The optical density was measured with a spectrophotometer (Pharmacia, Uppsala, Sweden)

at a wavelength of 450 nm against 10 mM MgSO4. For OD450 nm > 0.3, samples were

diluted with 10 mM MgSO4. Cell dry mass was estimated using a concentration factor of

0.29 mg ml-1 per OD 450 = 1 (Hazenberg, 1997).

Cell dry weight

Polycarbonate filters (Nuclepore, 47 mm and 0.2 µm) were dried at 80°C for two days and

tared after cooling down to room temperature in a desiccator. About 5 mg cell dry mass based

on the OD450 measurements was filtered through a tared filter. Filters were washed with the

same volume of 10 mM MgSO4, dried at 80°C for at least 3 days, and weighed after one

additional day in a desiccator.

PHA-measurement

PHA content of 5 mg freeze dried cell samples was determined by gas chromatography

according to the method described by Lageveen (1988).

Carbon concentration

To determine the octanoate concentration in the medium, the GC method described by

Rothen (1997) was used. The on-line measurement of octanoate was carried out as described

in chapter 3.

Chapter 6 155

Nitrogen concentration

Residual nitrogen concentration in the medium was measured spectrophotometrically (Dr.

Bruno Lange GmbH, Berlin, Germany). In the batch experiment the ammonium

concentration was measured on-line according to the method described in chapter three.

Chemicals

All chemicals were purchased from Merck (Darmstadt, Germany) and were >99% pure.

BLACK-BOX MODEL TO DETERMINE THE DUAL (C,N) LIMITED GROWTH

REGIME

In Figure 6.2 a black-box model is presented, which helps to define the growth conditions to

obtain dual (C,N) limited growth of P. oleovorans in a chemostat. Carbon (octanoate) and

nitrogen (ammonium) are fed to the culture at a defined ratio. Under dual (C,N) limited

growth conditions nitrogen and carbon are taken up to very low residual concentrations

(residual carbon, Cr < 8 mg l-1 and residual nitrogen, Nr < 0.2 mg l-1). Thus, the uptake ratio of

carbon to nitrogen (qC/qN) is very close to the feed ratio under dual (C,N) limited growth

conditions:

. [6.1]

The uptake ratio at a given growth rate (µ) can be calculated by the use of the yield

coefficients for carbon (YX/C) and nitrogen (YX/N), respectively:

[6.2]

Black-Box Model 156

The cell metabolizes the nuti·ients (black box), produces PHA, and residual (PHA-free)

biomass or rest biomass (XR). A ce1iain amount of carbon is "lost" due to cell energy

maintenance (turnover of cell materials, osmotic work to maintain concenh'ation gradients

(Pili, 1975)).

PHA

m

FIG. 6.2: Schematic representation of dual (C,N) limited growth of P. oleovorans. The ratio of the carbon (octanoate) to nih'ogen (ammonium) source in the feed medium (CtlNr) and the specific growth rate (~L) of P. oleovorans determine the amount of poly(3-hydroxyalkanoate) (PHA), the amount of rest biomass (XR), and the amount of carbon used for cell maintenance (m). When dual (C,N) limited growth conditions are established, the residual carbon (Cr) and niti·ogen (Nr) concenti·ations in the culture broth are ve1y low (Nr < 0.2 mg r 1

, Cr < 8 mg r 1).

According to the Put theo1y (Put, 1965; Pi1i , 1975), the experimentally detennined growth

yield changes with the growth rate due to the cell maintenance. This fact is considered in the

equation of Pi1i:

____ I. [6.3]

Chapter 6 157

where is the observed biomass yield on used carbon, is the true biomass yield

on carbon without the influence of maintenance carbon consumption, m is the maintenance

coefficient, and µ the specific growth rate (in steady-state cultures µ = dilution rate).

Thus, the dual limited growth regime can be predicted for chemostat cultures using a

modified method of Egli (1991) which is extended with the equation of Pirt (eq. 6.3). The

boundaries between carbon and dual (C,N) limitation, and between dual (C,N) and nitrogen

limitation were calculated in two steps.

(Cf/Nf)lower, the boundary between carbon and dual (C,N) limited growth regime

The integration of the energy maintenance equation of Pirt into the model of Egli (1991)

gives a continuous description of the lower boundary of dual (C,N) limited growth. Therefore,

the boundary can be predicted for a specific growth rate µ:

, [6.4]

where (Cf/Nf)lower denominates the boundary between C and dual (C,N) limitation,

and are the growth yields for nitrogen and carbon under C limitation, respectively.

(Cf/Nf)upper, the boundary between dual (C,N) and nitrogen limited growth

We assumed that the carbon used for cell maintenance and cell growth under dual (C,N)

limited growth equals that found under C only limited growth conditions. This is based on the

idea that the increase of the osmotic pressure due to PHA can be neglected (Dawes and

Senior, 1973) and consequently does not change the value of the cell maintenance.

Thus, the upper boundary of the dual (C,N) limited zone is calculated as the sum of

(Cf/Nf)lower and Cf/Nf units needed to produce the maximum PHA content at a given growth

rate. This idea is expressed by the following equation:

Black-Box Model 158

, [6.5]

where (Cf/Nf)upper stands for the boundary between dual (C,N) and N limited growth, ∆PHAµ

is the difference of the PHA content of the cells during carbon only and nitrogen only

limitation at the specific growth rate µ, and ß is the conversion factor that expresses how

much PHA is produced per Cf/Nf unit.

Description of the model parameters

Carbon to nitrogen ratio of the feed medium (Cf/Nf)

The concentration of the limiting substrate in the feed medium of a chemostat determines the

concentration of the total biomass that is present at steady-state (Monod, 1949). In order to

obtain a feed concentration independent expression, we used the Cf/Nf (g g-1) ratio as our

reference parameter. Thus, upscaling of the medium towards higher nutrient concentrations

is practicable.

Uptake ratio of carbon to nitrogen (qC/qN)

The uptake ratio of carbon and nitrogen is a key parameter for the model, because it describes

the physiological relation to the feed conditions (eqs. 6.1 and 6.2). To document the

relationship to the approach of Egli, we used for further calculations the ratio of the biomass

yields (YX/N/YX/C), which is equal to the uptake ratio of carbon and nitrogen (eq. 6.2) under

dual (C,N) limited growth.

Q) -ro .....

0

Cir

A- -- A

\

A-A Cir [)Ir Nlr c /

PHA

159 Chapter 6

Nlr

B-B Cir [)Ir Nlr x

FIG. 6.3: Model of dual (C,N) limited growth of P. oleovorans. The PHA content of the cells is a function of the Cff'Nr ratio and the dilution rate of the continuous culture. Since PHA is a carbon storage compound, the dual (C,N) limited growth regime is strongly dependent on the PHA content that can be accumulated inside the cell at a given dilution rate. The dual (C,N) limited growth regime includes fewer Cr!Nr units at high dilution rates (e.g. cut A-A) than under slow growth conditions (e.g . cut B-B), because the PHA content is smaller. In addition, the bounda1y between carbon and dual (C,N) limited growth regimes is shifted towards higher Cff'Nr for slower growth rates because the cells utilise a higher percentage of carbon for cell energy maintenance. Ch : carbon limited growth regime, Dh: dual (C,N) limited growth regime, Nh: nitrngen limited growth regime, (Cff'Nr)iower: bounda1y between carbon and dual (C,N) lirnitation, (Cff'Nr)upper: boundaiy between dual (C,N) and N limited growth, C: cai·bon concentration in the cell supernatant, N: nitrogen concentration in the cell supernatant, X: total biomass concentration, PHA: cellular content of PHA.

Black-Box Model 160

Specific growth rate (µ)

Generally, the specific growth rate (µ) of cells in a steady-state chemostat is equal to the

dilution rate (Monod, 1942). A culture growing at a certain µ has a generation time (G) that

can be calculated by the following equation:

[6.6]

The maximum value of the specific growth rate can be determined in a batch culture or by

a chemostat wash-out experiment. Here, a steady-state culture is exposed to an up-shift of the

dilution rate, and the wash-out of the culture is monitored by measurement of the optical

density. The maximum specific growth rate can then be calculated by equation 6.7:

. [6.7]

D is the dilution rate of the chemostat and X is the biomass, which is proportional to the

optical density of the culture.

Pirt parameters

The experimentally measured carbon yield is a function of the specific growth rate

(eq. 6.3) and according to Stouthamer (1979) this relationship is valid only for carbon-energy

limited cultures. The maintenance coefficient was determined by measurement of the

specific growth rate and the corresponding carbon yield in chemostat cultures. A double

reciprocal plot of against the experimental values of µ gives a linear correlation with

an intercept of and a slope of m (eq. 6.3 and Fig. 6.4).

In this model is assumed to be independent of the growth rate, since this

assumption is justified by the observations of Herbert (1961).

Chapter 6 161

PHA content and conversion to Cf/Nf units

Depending on µ, PHA in P. oleovorans is accumulated to different amounts under N only

limitation (de Smet et al., 1983; Preusting et al., 1991).

In order to describe the maximum PHA content of the cell as a function of the generation

time G, an equation analogous to Moser kinetics (Moser, 1958) was used:

, [6.8]

where PHA(G) is the PHA content for generation time G under nitrogen only limited growth

conditions, PHAmax is the maximum PHA content of the cell (48% of cell dry weight), G is

the generation time, Gmin is the generation time at maximum specific growth rate (µmax = 0.60

h-1, G = 1.155 h), n is the degree of the polynomial, and KG is the generation time at which

half of the maximum PHA content was measured minus Gmin (KG = 1.1 hn).

To reduce the number of model parameters the equation was transformed into an

expression with µ instead of G:

. [6.9]

PHA is not only accumulated under nitrogen limitation but also under carbon limitation.

Therefore, the difference between the PHA content under carbon and nitrogen limited growth

(∆PHAµ = PHAN lim - PHAC lim) reflects the actual PHA production in the dual (C,N) limited

zone.

On average the PHA of P. oleovorans consists of 67 % carbon, 10 % hydrogen, and 23 %

oxygen (w/w). Thus, ∆PHAµ may be used to calculate the width of the dual (C,N) limited

growth regime expressed in Cf/Nf units. This is done by the conversion factor ß (see also eq.

6.5) :

Black-Box Model 162

, [6.10]

where ∆PHA is the increase of the PHA content when the C/N ratio in the feed medium is

increased by the carbon concentration over a range of ∆(Cf/Nf) under dual (C,N) limited

growth conditions.

Highest specific and most efficient PHA production under dual (C/N) limited growth

conditions

The highest specific PHA productivity of P. oleovorans is the maximum PHA productivity

of the rest biomass (active biomass):

, [6.11]

where qP is the specific PHA productivity at growth rate µ ( D), PHA(µ) the cellular PHA

content, X the total biomass concentration and Xr the rest biomass concentration at growth

rate µ.

However, the highest specific productivity of PHA is not always the most economical one

with respect to the carbon substrate, especially if this is extraordinarily expensive. The most

efficient production of PHA therefore should be done at the lowest Cf/Nf ratio at which it is

possible to attain high PHA contents in the cell and a maximum conversion of all carbon

supplied (carbon limitation). Thus, this efficiency (E) has a maximum at boundary

(Cf/Nf)upper. The efficiency is calculated as follows :

. [6.12]

Chapter 6 163

RESULTS

In order to model the dual (C,N) limited growth zone (Fig. 6.3) of P. oleovorans, the essential

parameters were determined.

Specific growth rate (µ)

To determine the range of the specific growth rate, a batch culture of P. oleovorans was

grown in a bioreactor with 2 l modified medium X. Two growth phases could be

distinguished (Fig. 6.4): an exponential phase, from 0 to 5.5 h, followed by a linear growth

phase, where PHA accumulation occurred. The maximum growth rate in the exponential

(unlimited) growth phase of this batch was 0.58 h-1.

The maximum specific growth rate was verified by a wash-out experiment. P. oleovorans

was cultured in a chemostat at a dilution rate of 0.45 h-1 under octanoate limited conditions

(Cf/Nf=4.0 g g-1) in medium XCC. The dilution rate was then increased in one step to 0.9 h-1.

The wash-out of the culture was followed by measurement of the optical density (data not

shown), and a maximum specific growth rate of 0.63 h-1 was obtained. For further

calculations µmax = 0.60 h-1, the average of both experiments, was used. This value

corresponds to a generation time G = 1.16 h.

Black-Box Model 164

1.8 0.18

1.6 0.16 ~ - ......

C/) 1.4 x-x 0 0.14 I

Q) O> ..... )I( -)I( ')I(. ~

"C 1.2 ~/ 0.12 <( c ~ • rn L )I( • • I C/) ~ 1.0

')I( 0.10 a..

C/) "x I ro C/) ~

E C/) \ I ...... 0.8 0.08 ~ 0 ro I

:0 E )I( ~ - .Q 0.6 5K 0.06 c

c .0 \ /~ Q)

0 )I( O> .0 \ ~ 0 ..... 0.4 0.04 ..... ro /' :!: u z

0.2 -- - )I(

0.02 \.

0.0 ))I( 0.00 0 1 2 3 4 5 6 7 8 9 10 11 12 13

Time [h]

FIG. 6.4: Accumulation of PHA in a batch culture of P. oleovorans with an initial C/N ratio of 12.5 g g-1 in medium X with octanoate as C source. P. oleovorans sta1ted to accumulate PHA (• )when nitrogen ~) (ammonium) was exhausted. The total biomass (0 ) increased ftnther at a linear rate of 0.09 g r 1 h-1, whereas the rest biomass (XR) (• ) remained nearly constant. Residual carbon (O); residual nitrogen ~).

Pir t parameters (m, D The Put parameters were determined in a series of chemostat experiments under carbon

liinitation (Cr/Nr:'.S 4 g g-1) at different dilution rates (Fig. 6.5); the maintenance coefficient m

= 0.055 g g-1 h-1 and the hue biomass yielD = 1.59 g g-1.

1.4

1.2

1 ~ .,.... '0> 0.8 ~

Q 0.6 x < 0.4 .......

0.2

0 0 2

165

4 6 8

1/µ [h]

Slope = m

y(x)=0.055*x+0.629 RSQ=0.96

10 12

Chapter 6

14

FIG. 6.5: Detennination of the Pili parameters for growth of P. oleovorans on octanoate. The experimental carbon yield was measured in 6 independent steady-state experiments with limiting carbon concentrations (Cc/Nr < 4 g g·1

) in medium XCC.

The nitrogen yield under carbon limitation D The cell composition is not constant over the range of the growth rates. The cellular content

of nitrngen containing compounds, such as DNA, RNA, and proteins, is growth rate

dependent (Herbe1i, 1976), and Dis therefore a function of the dilution rate. As an

example, the N content of Klebsiella aerogenes was 12.8% and 14.2% during growth at

dilution rates of 0.1 and 0.85 h-1, respectively (Herbe1i, 1961 ). This variation is rather small,

and we therefore used a constant nitrogen yield of 7.8 g g·1 (Preusting et al., 1993a) as

detennined previously for P. oleovorans.

PHA content of the cells as a function of the growth rate (PHAµ.)

In the batch experiment (Fig. 6.4), a linear biomass increase could be detected after 5.5 h

which was caused by PHA accumulation at a constant rate of2.8% (w/w) PHA h.1 from6 to

Black-Box Model 166

9 h. If PHA accumulation is linear under these conditions the cellular PHA content should

coITelate with the generation time. To test this, the PHA data obtained in a single phase

system (excess octanoate, ammonium limiting; data shown in Fig. 6.6) were plotted against

the generation time (Fig. 6.6), and compared to data of Preusting et al. (1991). In both cases

the PHA content increased with increasing generation times. However, when G exceeded 5.3

hours (µ = 0.13 h-1) the PHA content remained constant at about 48% of cell diy mass. The

data were modelled according to equation 6.9, with n = 1 or 2. A best fit was obtained for n =

2, for G > 1.73 h or D < 0.4 h-1.

en 50 (/') ro E 45 ·------~-.~~~-o

>-~

"'C

(1.) () -0

~ 0 .......... ....... c 2 c 0 ()

~ I a..

40 35 30 25 20 15 10 5 0

0 2

'/ I

------- ------_,,,, --./

/ /

4 6 8 10 Generation time [h]

FIG. 6.6: PHA content of P. oleovorans as a function of the generation time in a chemostat under nitrngen limited growth conditions. The data were coITelated with Moser kinetics (n = 1: dashed line; n = 2 full line). Data for the PHA content of cells in octanoate lirnited ( • ) and of two-liquid-phase chemostats with n-octane as organic carbon source (0 ) (adapted from Preusting et al., 1991).

The PHA content under carbon only limited growth conditions varied strongly at 5% ± 4%

PHA per g cell diy weight over the whole growth rate range and there was no clear tendency

towards higher PHA contents during faster growth.

Chapter 6 167

∆PHAµ was therefore PHA(µ) reduced by 5% PHA per g cell dry weight.

PHA content of the cells and its conversion into Cf/Nf units

In a previous experiment (chapter 3, gradient 3) a continuous culture of P. oleovorans was

exposed to a medium feed gradient, where the Cf/Nf ratio was decreased continuously by

decreasing the carbon source feed. The cells passed from nitrogen only limitation through a

dual (C,N) to a carbon only limited growth regime. Within this growth regime the PHA

content of the cells decreased at a rate of ß = 6.4% PHA (Cf/Nf)-1. This relationship is

considered to be constant for all growth rates because PHA has no influence on the osmotic

pressure of the cell (Anderson and Dawes, 1990) and thus no influence on the cell

maintenance.

The dual (C,N) limited growth regime

The above mentioned variables are summarised in Table 6.1 and were used to calculate the

growth conditions for dual (C,N) limited growth in a chemostat (eqs. 6.4 and 6.5).

The model predicts a “banana” shaped dual limitation zone (Figs. 6.3 and 6.7). There is

good agreement with experimentally determined boundaries of the dual (C,N) limited growth

zone (see chapters 3 and 5) which are inserted in Figure 6.7a. At dilution rates higher than

0.45 h-1 ∆(Cf/Nf) is very small due to the relatively small difference of the PHA content under

carbon only and nitrogen only limitation.

A slightly different presentation (Fig. 6.7b), which uses the generation time instead of the

dilution rate, emphasizes that the dual (C,N) limited growth zone has a constant ∆(Cf/Nf)

width for generation times larger than 4 h, since ∆PHAµ remains constant.

TAB. 6.1: The following parameters were used to determine the growth conditions for dual (C,N) limited growth with a model.

Parameter Value/ Range Unit

µ 0 - 0.6 h-1 ∆PHAµ 0 - 0.43 g PHA g-1 cdw

ß 6.4 %PHA (Cf/Nf)-1

Black-Box Model 168

8 7.8 g g-1

1.59 g g-1

Ill 0.055 g g-1 h-1 KG 1.1 h2

0.6

0.5 @

Nlr 0.4 .,....

I

..c .......... 0.3 ::J...

0.2 Cir

0. 1

0.0

30 @ 25

,........, 20 ..c ..........

(9 Cir 15

10

5

0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

C/ Nt [g g-1]

FIG. 6.7: Modelling of a simultaneous nitrogen and carbon limited growth regime (Dfr) of P. oleovorans in chemostat cultures. Dual nitrogen (ammonium) and carbon ( octanoate) limited growth occurs within a transition zone between sole carbon

169 Chapter 6

(Ch') and sole nitrogen (Nfr) limited growth. (0 ) data points obtained in chemostat experiments (see chapter 3 and 5). Best PHA productivity (• ) , highest efficiency of PHA synthesis (II) Panel a: Dual (C,N) limited growth as a function of the specific growth rate ~L (equal to the dilution rate during steady-state growth in a chemostat) and the Cti'Nr ratio in the feed medium. Panel b: Dual (C,N) limited growth as a function of the generation time G and the Cti'Nr ratio in the feed medium of a steady-state culture. Ch': carbon limited growth, Dk: dual (C,N) limited growth, and Nfr: nitrogen limited growth regime.

Black-Box Model 170

Optimal PHA production states

According to the model, the highest specific productivity of P. oleovorans was 0.133 g PHA

(g XR)-1 h-1 at a dilution rate of 0.21 h-1 and at a Cf/Nf ratio of 12.4 g g-1 (Figs. 6.7 and 6.8).

The PHA content of the cell under these growth conditions is approximately 39%.

The efficiency of PHA synthesis (eq. 6.12) is highest with D = 0.18 h-1 and Cf/Nf = 13.2 g g-1.

The PHA content of the cell was calculated to be about 42% of the biomass and the specific

productivity was 0.129 g PHA (g XR)-1 h-1.

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.00 0.10 0.20 0.30 0.40 0.50 0.60Dilution rate [h-1]

q P [g

PH

A g

-1 X

r h-1

]

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

Effi

cien

cy o

f PH

A p

rodu

ctio

n[g

2 g-

2 ]

FIG. 6.8: Optimal PHA production conditions under dual (C,N) limitation. The highest specific PHA productivity qP (full line) and the most efficient PHA

production E (dashed line) were calculated for steady-state cultures of P. oleovorans growing at (Cf/Nf)upper boundary.

DISCUSSION

A black-box model was developed to describe the growth conditions that are needed to

establish dual (C,N) limited growth in chemostats. Dual (C,N) limited growth is especially

interesting with respect to the efficiency of the PHA production with P. oleovorans. The

Chapter 6 171

model calculates the Cf/Nf ratio needed to obtain dual (C,N) limited growth for a given

dilution rate. Since the model does not depend on absolute C or N concentrations, it can be

used at higher concentrations in the medium than used here.

The lower Cf/Nf boundary

The boundary between the C and dual (C,N) limited growth regime ((Cf/Nf)lower) is strongly

influenced by the cell maintenance (eq. 6.4). At high dilution rates the maintenance does not

show a large effect on the (Cf/Nf)lower ratio. However, as µ decreases the portion of consumed

carbon that is used for maintenance increases, resulting in an increase in (Cf/Nf)lower. For

instance, cells growing at a dilution rate of 0.01 h-1 are theoretically carbon limited up to a

Cf/Nf ratio of 55 g g-1.

In the model it is assumed that the nitrogen yield ( ) does not change with the

dilution rate. Thus, all data points on the boundary represent the same biomass concentration.

The upper Cf/Nf boundary

The width ∆(Cf/Nf) of the dual limited growth regime was assumed to be related only to the

achievable PHA content under dual (C,N) limited growth conditions. This is certainly the

case for dilution rates above D = 0.2 h-1 (G ≤ 3.5 h), because all carbon that is taken up in

addition to the rest biomass carbon, is accumulated as PHA, resulting in a linear relationship

between PHA content and generation time shorter than 3.5 h.

The model predicts a very small zone of dual (C,N) limitation for dilution rates larger than

D = 0.45 h-1. The reason for this is that 5% PHA are already accumulated under carbon

limitation and only 5.7% PHA is accumulated under N limitation. Hence, ∆PHAµ, the

difference between the PHA content at (Cf/Nf)upper and (Cf/Nf)lower is small.

At D < 0.2 h-1 the PHA content seems not to increase any further and remains at about

48% of the cell dry weight (Fig. 6.5). It can be speculated that a higher PHA content disturbs

cell growth. However, the carbon uptake rate is not necessarily affected by the cellular PHA

and consequently the cell has to get rid of the additional carbon and excretes metabolic

intermediates. Thus, the dual (C,N) limited growth regime could in principle be larger for

Black-Box Model 172

dilution rates below 0.2 h-1 than predicted by the model. Whether this can in fact be observed

remains to be determined in further experiments under slow growth conditions.

Refinements of the model

In this model we have not considered variations in the maintenance coefficient. This

parameter is sensitive to many environmental conditions, such as salt concentrations, pH,

temperature and other factors (McGrew and Mallette, 1962; Pirt, 1975). It may be useful to

extend the model to include these effects.

The maximum PHA content of cells in the model is 48% based on the data of Figure 6.6.

However, higher PHA contents have been reported, such as a PHA content of 66%, which

was found for P. oleovorans in a two stage chemostat culture with n-octane as carbon source

(Hazenberg, 1997), indicating that in these experiments ß and ∆(Cf/Nf) are larger. Such

changes in these experimental parameters can clearly be incorporated in the model.

Benefits of the model

The model can be used to establish optimal PHA production conditions based on a few

experiments. For P. oleovorans the highest productivity was found to be at D = 0.21 h-1 at

(Cf/Nf)upper = 12.4 g g-1. The same optimum (D = 0.2 h-1) was found experimentally for the

two-liquid phase cultures with n-octane and nitrogen limitation by Preusting (1991). In

contrast Ramsay et al. (1991) found a productivity of 0.074 g of PHA g-1 of cellular protein

h-1 in a chemostat at a dilution rate of 0.25 h-1, which was assumed to be the maximum

productivity. In carrying out the experiments at dilution rates between D = 0.1 h-1 and 0.5 h-1

at a constant Cf/Nf ratio of 12.7 g g-1, most of the data were collected outside of the dual

limitation zone, in the N only limited zone, rather than at (Cf/Nf)upper where higher PHA

productivities can be obtained.

The most efficient PHA production was calculated with our model to be at a D = 0.18 h-1

for a (Cf/Nf)upper = 13.2 g g-1. However, the PHA content of the cells increased by only 4% of

total biomass, compared to the conditions determined for maximum productivity. This

difference is not particularly interesting in the production with octanoate as carbon source.

Chapter 6 173

However, one can imagine that an expensive substrate is added to an inexpensive medium

when the culture is dual (C,N) limited. In this case, P. oleovorans would be cultured in a

chemostat on a cheap substrate, e.g. acetate, under conditions of (Cf/Nf)lower at the dilution

rate of best productivity (D = 0.21 h-1). After steady-state conditions are established, the

more expensive carbon substrate is added to the culture at a defined flow rate. Since the PHA

production would be initiated, the culture would additionally incorporate the new carbon

source. Thus, a most economic production could be achieved since the expensive carbon

source would only be used for PHA production (uncoupled PHA production). Primary tests

were made with octanoate and citrate by Durner (1998), where it could be shown in

chemostat experiments that an uncoupled PHA production exists within a very limited range

of Cf/Nf ratios.

A useful characteristic of dual limited growth over single compound limitations is that

growth inhibiting carbon sources are consumed completely, and steady-states of such dual

limited cultures are therefore stable, since octanoic acid is known to inhibit growth of P.

oleovorans (Brandl et al., 1988 and chapter 4). Consequently, a PHA production with

octanoate should be performed under dual (C,N) limited growth conditions.

Fed-batch cultivation represents an alternate PHA production method. Given that the

specific growth rate and the qC/qN ratio are important factors for efficient PHA production,

special feed techniques can be applied that enable a nutrient supply that maintains the culture

at the (Cf/Nf)upper boundary.

NOMENCLATURE

ß [% PHA (Cf/Nf)-1] Yield of PHA per C/N unit of the feed medium under dual (C,N)

limited growth conditions

Cf/Nf [g g-1] Carbon to nitrogen ratio in the feed medium

∆(Cf/Nf) [g g-1] Width of the dual (C,N) limited growth regime

(Cf/Nf)lower [g g-1] Boundary between carbon and dual (C,N) limited growth regime

with respect to the (Cf/Nf) ratio

Black-Box Model 174

(Cf/Nf)upper [g g-1] Boundary between dual (C,N) limited and nitrogen limited growth

regime with respect to the (Cf/Nf) ratio

E [g2 g-2] Efficiency

µ [h-1] Specific growth rate

µmax [h-1] Maximum specific growth rate

G [h] Generation time

Gmin [h] Minimum generation time

KG [h] Generation time at which half of the maximum PHA content was

measured minus the generation time of fastest growth

m [g g-1 h-1] Cell energy maintenance

∆PHAµ [g g-1] Difference of poly(3-hydroxyalkanoate) content between carbon

and nitrogen limited growth conditions at a given µ

qP [g g-1 h-1] Specific PHA productivity

[g g-1] Biomass yield on used carbon under carbon limitation

[g g-1] Biomass yield on used nitrogen under carbon limitation

[g g-1] True yield of biomass on used carbon under carbon limitation

(determined by the method of Pirt)

X [g l-1] Total biomass concentration

XR [g l-1] Rest biomass concentration, biomass without PHA

175

176

CHAPTER 7

GENERAL CONCLUSIONS

Keywords: Bioplastic, Pseudomonas oleovorans, chemostat, continuous culture, dual

nutrient limited growth regime, medium feed gradient, medium chain length

poly(3-hydroxyalkanoate), production, degradation.

Chapter 7 177

A bioplastic is needed for the future

In the past 20 years the poly(3-hydroxyalkanoate) (PHA) production in microorganisms

gained much interest in science and industry due to its particular properties, such as

biocompatibility, biodegradibility, and uniform stereospecifity (Lee, 1996). These qualities

gave PHA the reputation of being the plastic of the future (Hänggi, 1995; Page, 1995).

However, the market price of the only commercial PHA, BIOPOL

(poly(3-hydroxybutyrate-co-3-hydroxyvalerate)), remained too high at about $16 per kg

(Lee, 1996). Therefore, the production volume remained at a small scale of about 1,000 tons

per year (Poirier et al., 1995). As a consequence, PHA did not gain a large market share

because it could only be applied for special products, where the customer was willing to pay

a higher price for the “greener” product, e.g. shampoo bottles made of BIOPOL (SANARA,

Wella, Germany). For cheaper bulk-goods, such as packaging material or low cost products

(e.g. diapers), to date PHA cannot compete with the petroleum based plastics. Transfer of the

essential genes into crop plants like soybean, rapeseed, and potato, may well result in

production cost reduction (Poirier et al., 1995; Pool, 1989; van der Leij and Witholt, 1995).

The use of PHA in medical surgery is still at the stage of basic research. This is certainly

related to the fact that the plastic is of bacterial origin and was so far purified with harmful

chemicals like chloroform.

A further application is the use of PHA monomers as a source for stereospecific chemicals

(Holmes, 1988). The PHA polymer could be degraded enzymatically resulting in a high yield

of (R)-3-hydroxyalkanoic acid monomers. This appears to be of particular interest since there

are more than 90 different PHA monomers described in literature (Lee, 1996; Steinbüchel

and Valentin, 1995).

Production of PHA in Pseudomonas oleovorans

Pseudomonas olevorans is a potential strain for industrial scale PHA production. P.

oleovorans is able to grow and synthesize PHA on many different carbon sources ranging

from hydrophobic n-alkanes to hydrophilic n-alkanoic acids with medium chain lengths

(between 6 and 14 carbon atoms) (Abe et al., 1990; Brandl et al., 1988; Curley et al., 1996; de

Smet et al., 1983; Eggink et al., 1995; Fritzsche et al., 1990b; Fritzsche et al., 1990c; Hori et

General Conclusions 178

al., 1994; Kim et al., 1995; Kim et al., 1997; Lageveen et al., 1988; Preusting et al., 1991;

Preusting et al., 1993b; Witholt et al., 1994). High PHA contents were previously reported

for cultures under nutrient starvations like nitrogen, phosphorus, sulfur, magnesium and

oxygen (Gagnon et al., 1992; Gross et al., 1989; Lageveen et al., 1988). The cellular

integration of carbon in the form of PHA enables P. oleovorans and all other PHA

accumulating microorganisms to grow dual nutrient limited in continuous culture. Both

nutrients have to be heterologous (satisfying different physiological functions, Chapter 1),

preferentially a carbon source that enables biomass and PHA synthesis and a second nutrient

that triggers the PHA synthesis.

An exciting property of P. oleovorans is the ability to grow triple (C,N,P) limited in a

chemostat at a dilution rate of 0.2 h-1 (Chapter 5). This special growth regime has never been

described before. In this feasibility study only a few steady-states were examined and many

questions remained unanswered: Is there a higher PHA content achievable under triple

(C,N,P) limited growth conditions? What is the stability of the culture at the steady-states or

more specific: will the old and the new steady-states be the same after disturbances of the

system, e.g. after nutrient pulses?

In the same series of experiments (Chapter 5) it was shown that P. oleovorans is able to grow

dual (N,P) limited with carbon in excess. This observation was based only on two data points

which had a PHA content of 23.6% and 32.2% cell dry weight. It would be interesting to

assess the influence of the Nf/Pf ratio on the cellular PHA content and the monomeric PHA

composition in more detail.

Methods to investigate dual nutrient limitation

A literature survey (Chapter 2) revealed that dual nutrient limitation was described for many

microorganisms in a variety of growth systems. The application of continuous culture

techniques and the systematic approach by the modification of the nutrient mixture in the

medium gave the best results and showed that the physiology of the cell under dual nutrient

limited growth conditions may differ significantly from the well explored single limitations.

This clearly indicates that there is a strong need for experimental investigations of the dual

nutrient limited growth regimes although the ability and the feed conditions for growth under

Chapter 7 179

dual nutrient limitation can be predicted under certain circumstances (Chapter 6 and Egli,

1991). The best experimental tool to study growth under nutrient limitations is the chemostat.

It ensures a well defined environment for the cells and the data of steady-state cultures are

considered to be very consistent (Pirt, 1975). However, chemostat experiments are quite time

consuming because a steady-state has to be established first until an analysis can be made.

This is generally considered to be the case after 3 - 5 volume changes (Pirt, 1975).

We developed a method where the medium composition was changed continuously by the

increase or decrease of the carbon (octanoate) or nitrogen (ammonium) source. Hence, a

steady-state of the culture was never achieved nor intended. The boundaries of the dual

nutrient limited growth regime obtained by this method differed from a reference chemostat

considerably. This was primarily due to a wash-in effect causing a time delay of the medium

gradient (Appendix A of Chapter 3). The mathematical correction of the delay resulted in a

good approach to the chemostat data. Interestingly, the cell physiology depended on the way

the medium gradient was performed, e.g. the decrease of the carbon concentration resulted in

a very efficient degradation of the cellular PHA to less than 1% cell dry weight.

Possibilities to predict dual limited growth regimes

The dual limited growth regime can be predicted approximately for a specific growth rate

when the growth yields of the two nutrients are known under single limitations of each

nutrient (Chapter 2 and Egli and Quayle, 1986).

The growth yields of P. oleovorans differ significantly between carbon (octanoate) and

nitrogen (ammonium) limitation because accumulated PHA increases the weight of a cell. A

comparison of our own data with that from the literature revealed that the generation time of

the cells may be used to calculate the capacity of the cells to store carbon as PHA. In addition,

a steady carbon flux is essential for a living cell (cell energy maintenance). This is especially

recognizable at small growth rates as the portion of carbon used for cell maintenance

increases (Pirt, 1965; Pirt, 1975). These influences were used to design a black-box model

that was able to predict the Cf/Nf ratios needed for dual (C,N) limited growth (Chapter 6).

Although only a few parameters are needed, the model could demonstrate the following

points:


• The boundary between carbon and dual (C,N) limited growth is shifted towards higher

Cf/Nf ratios for slower growth rates (confirmed by experimental data of Durner, 1998).

• The width of the dual (C,N) limited growth regime correlates with the PHA content up to

generation times of 4 hours. For larger generation times the model predicts a constant

width of the dual (C,N) limited growth zone (no experimental data available).

• The optimal PHA productivity was determined to be at a dilution rate of D = 0.21 h-1 and

a Cf/Nf ratio of 11.8 g g-1 (optimal productivity for growth on octane was determined to be

at D = 0.2 h-1 (Preusting et al., 1993a)).

When the key parameters are known the black-box model can be easily extended for

carbon sources other than octanoate. However, a new model has to be designed to determine

the dual limited growth regime for other heterologous nutrients, e.g. phosphorus and nitrogen.

A more sophisticated model could be proposed when the intracellular metabolic fluxes were

better understood (Sauer et al., 1996), in other words when the “lid” of the black-box could

be removed.

MclPHA production in P. oleovorans under dual nutrient limited growth conditions

The application of dual nutrient limited growth to produce mclPHA is advantageous for

several reasons:

• High concentrations of the carbon substrate in the medium may inhibit bacterial growth.

This could be shown for the case of octanoic acid (Chapter 3). This problem can be solved

by means of dual nutrient limited growth with carbon as one limiting nutrient, reducing

the residual carbon concentration in the culture to a minimum.

• The medium components are used very efficiently and therefore the production costs can

be reduced. In addition, this is especially useful when the carbon substrate, e.g. octane,

obstructs the down-stream process of the PHA isolation (K. Jung, personal

communication). The possibility to grow P. oleovorans under dual (octane,ammonium)

limited growth conditions is under current investigations (K. Jung, personal

communication).

Chapter 7 181

• The production of biomass can be uncoupled from the PHA production. This could be

performed by the specific feed of two carbon sources, a PHA producing and a biomass

producing one. Durner (1998) made preliminary experiments with citrate (biomass) and

octanoate (PHA) and could show that both carbon sources were used simultaneously in

batch cultures. In a series of chemostat experiments the findings were different: octanoate

was preferred over citrate and consequently biomass and PHA production occurred on

octanoate alone. It would be interesting to repeat this approach with other carbon sources

like glucose and octanoate. P. oleovorans is not able to produce PHA on glucose, thus

only octanoate is used for PHA synthesis.

• The dual limited growth regime could be used to produce tailor made PHAs. P.

oleovorans might be the ideal production strain for such an application, since it is able to

polymerize more than 60 different 3-hydroxymonomers to PHA. First studies were

performed by Durner (1998) with octanoate and nonanoate as carbon sources. He

demonstrated that the monomeric PHA composition reflected the molar ratio of the

mixture of the two carbon sources in the feed medium.


However, a few disadvantages have to be considered, too:

• Accumulated PHA in P. oleovorans is degraded at a high rate when the cells are starved

for carbon (no external carbon source available, Chapter 4). This may be of importance

when the cells of a continuous culture are collected first in a storage container to enable an

efficient down-stream process. A down-stream process that is continuously performed

could avoid this loss of PHA. This raises the question whether mclPHA can be produced

with P. oleovorans with a high PHA content in a large bioreactor (> 1,000 l). Due to

inefficient mixing in such large bioreactors gradients of nutrients may be established.

Thus, it would be useful to know what kind of effects occur when P. oleovorans is

exposed to fluctuation of nutrient concentrations.

• The dual nutrient limited growth regime is highly dependent on the specific growth rate

and on the nutrition of the cell (Chapter 6). This means that the production process has to

be designed carefully in order to avoid a transition to a different growth regime (e.g. from

dual (C,N) limitation to carbon only limitation).

Increase of PHA content with new combinations of limiting nutrients

A new combination of heterologous nutrients might be of great interest, e.g. a triple

limitation by carbon, nitrogen and oxygen. Oxygen as limiting nutrient might affect the PHA

accumulation in P. oleovorans, because enhanced accumulation could be detected in batch

and chemostat cultures of Ralstana eutrophus H16 (Sonnleitner et al., 1979) under

simultaneous oxygen and nitrogen limitation.

Low oxygen concentrations alone might already trigger the PHA accumulation. It was

shown that Azotobacter beijerinckii (Senior et al., 1972) accumulated PHB in response to

oxygen limitation. It was concluded that PHB is not only a storage compound of energy and

carbon, but also an electron sink. Whether this might be the case for P. oleovorans is under

current investigation at our institute (Prieto, personal communication).

In order to improve PHA production further, genetic engineering could be used. Recently,

it has been shown that an increase in the polymerase expression results in higher PHA

content (Kraak et al., 1997). This occurs when additional copies of the polymerase C1 gene

Chapter 7 183

controlled by its native promotor were transferred into P. oleovorans. This recombinant

strain P. oleovorans GPo1[pGEc405] showed an increased PHA content of 64% (w/w). As

mentioned previously, in this case dual (C,N) limited growth could be applied to produce

tailor made PHA.

References 184

REFERENCES

Abe, C., Taima, Y., Nakamura, Y., and Doi, Y. 1990. New bacterial copolyester of 3-hydroxyalkanoates and 3-hydroxy-fluoroalkanoates produced by Pseudomonas oleovorans. Polym. Comm. 31: 404-406.

Abe, H., Doi, Y., and Yamamoto, Y. 1992. Controlled release of lastet, an anticancer drug, from poly(3-hydroxybutyrate) microspheres containing acylglycerols. Macromolec. Rep. A29: 229-235.

Al-Awadhi, N., Egli, T., Hamer, G., and Mason, C. A. 1990. The process utility of thermotolerant methylotrophic bacteria: 1. An evaluation in chemostat culture. Biotechnol. Bioeng. 36: 816-820.

Alexander, M. L., and Ramkrishna, D. 1991. Cybernetic modeling of iron-limited growth and siderophore production. Biotechnol. Bioeng. 38: 637-652.

Anderson, A. J., and Dawes, E. A. 1990. Occurence, metabolism, metabolic role, and industrial uses of bacterial polyhydroxyalkanoates. Microbiol. Rev. 54: 450-472.

Anderson, A. J., Haywood, G. W., and Dawes, E. A. 1990. Biosynthesis and composition of bacterial poly(hydroxyalkanoates). Int. J. Biol. Macromol. 12: 102-105.

Babel, W., Brinkmann, U., and Müller, R. H. 1993. The auxilary substrate concept - an approach for overcoming limits of microbial performance. Acta Biotechnol. 13: 211-242.

Babel, W., Müller, R. H., and Markuske, K. D. 1983. Improvement of growth yield of yeast on glucose to the maximum by using an additional energy source. Arch. Microbiol. 136: 203-208.

Bader, F. G. 1978. Analysis of double-substrate limited growth. Biotechnol. Bioeng. 20: 183-202.

Bader, F. G., Meyer, J. S., Fredrickson, A. G., and Tsuchiya, H. M. 1975. Comments on microbial growth rate. Biotechnol. Bioeng. 17: 279-283.

Bae, W., and Rittmann, B. E. 1996a. Responses of intracellular cofactors to single and dual substrate limitations. Biotechnol. Bioeng. 49: 690-699.

Bae, W., and Rittmann, B. E. 1996b. A structured model of dual-limitation kinetics. Biotechnol. Bioeng. 49: 683-689.

Ballard, D. G. H., Holmes, P. A., and Senior, P. J. 1987. Formation of polymers of ß-hydroxybutyric acid in bacterial cells and a comparison of the morphology of growth with the formation of polyethylene in the solid state. Recent Adv. Mech. Synth. Aspects Pol. 215: 293-314.

Baloo, S., and Ramkrishna, D. 1991a. Metabolic regulation in bacterial continuous cultures: II. Biotechnol. Bioeng. 38: 1353-1363.

Baloo, S., and Ramkrishna, D. 1991b. Metabolic regulation in bacterial continuous culutures: I. Biotechnol. Bioeng. 38: 1337-1352.

References 185

Baltzis, B. C., and Fredrickson, A. G. 1988. Limitation of growth rate by two complementary nutrients: some elementary but neglected considerations. Biotechnol. Bioeng. 31: 75-86.

Bell, W. H. 1980. Bacterial utilization of algal extracellular products. 1. The kinetic approach. Limnol. Oceanogr. 25: 1007-1020.

Blackman, F. F. 1905. Optima and limiting factors. Annals of Botany 19: 281-295.

Blancarte-Zurita, M. A., and Branion, R. M. R. 1986. Particle size effects in the microbiological leaching of sulfide concentrates by Thiobacillus ferrooxidans. Biotechnol. Bioeng. 28: 751-755.

Bley, T., and Babel, W. 1992. Calculating affinity constants of substrate mixtures in a chemostat. Acta Biotechnol. 12: 13-15.

Bloembergen, S., Holden, D. A., Hamer, G. K., Bluhm, T. L., and Marchessault, R. H. 1986. Studies of composition and crystallinity of bacterial poly(ß-hydroxybutyrate-co-ß-hydroxyvalerate). Macromolecules 19: 2865-2871.

Brandl, H., Gross, R. A., Lenz, R. W., and Fuller, R. C. 1988. Pseudomonas oleovorans as a source of poly(beta-hydroxyalkanoates) for potential applications as biodegradable polyesters. Appl. Environ. Microbiol. 54: 1977-1982.

Brandl, H., R.A., G., Lenz, R. W., and Fuller, R. C. 1990. Plasics from bacteria and for bacteria: poly(ß-hydroxyalkanoates) as natural, biocompatible, and biodegradable polyesters. Adv. Biochem. Eng./Biotechnol. 41: 77-93.

Brinkmann, U., and Babel, W. 1992. Simultaneous utilization of heterotrophic substrates by Hansenula polymorpha results in enhanced growth. Appl. Microbiol. Biotechnol. 37: 98-103.

Button, D. K. 1985. Kinetics of nutrient-limited transport and microbial growth. Microbiol. Rev. 49: 270-297.

Button, D. K. 1991. Biochemical basis for whole-cell uptake kinetics: Specific affinity, oligotrophic capacity, and the meaning of the Michaelis constant. Appl. Environ. Microbiol. 57: 2033-2038.

BUWAL 1991. Altstoffmärkte für Kunststoffabfälle in der Schweiz. Schriftenreihe Umwelt 147: Z1-Z5.

Byrom, D. 1987. Polymer synthesis by microorganisms: technology and economies. Trends Biotechnol. 5: 246-250.

Carnicero, D., Fernandez Valverde, M., Canedo, L. M., Schleissner, C., and Luengo, J. M. 1997. Octanoic acid uptake in Pseudomonas putida U. FEMS Microbiol. Lett. 149: 51-58.

Chen, H. C. 1993. Citric acid production by Aspergillus foetidus in batch and fed-batch cultures. Food Biotechnol. 7: 221-234.

Chen, P. S., Toribara, T. Y., and Warner, H. 1956. Microdetermination of phosphorus. Anal. Chem. 28: 1756-1758.

Chen, Q., Nijenhuis, A., Preusting, H., Dolfing, J., Janssen, D. B., and Witholt, B. 1995. Effects of octane on the fatty acid composition and transition temperature of

References 186

Pseudomonas oleovorans membrane lipids during growth in two-liquid phase continuous cultures. Enzyme Microb. Technol. 17: 647-652.

Choi, D. K., Ryn, W. S., Choi, C. Y., and Park, Y. H. 1996. Production of L-ornithine by arginine auxotrophic mutants of Brevibacterium ketoglutamicum in dual substrate-limited continuous culture. J. Ferment. Bioeng. 81: 216-219.

Cooney, C. L., and Wang, D. I. C. 1976b. Transient response of Enterobacter aerogenes under a dual nutrient limitation in a chemostat. Biotechnol. Bioeng. 18: 199-208.

Cooney, C. L., Wang, D. I. C., and Mateles, R. I. 1976a. Growth of Enterobacter aerogenes in a chemostat with double nutrient limitations. Appl. Environ. Microbiol. 31: 91-98.

Cooney, M. J., Marison, I. W., van Gulik, W. M., and von Stockar, U. 1996. Calorimetric and stoichiometric analysis of growth of Kluyveromyces fragilis in continuous culture: nitrogen limitation imposed upon carbon-limited growth. Appl. Microbiol. Biotechnol. 44: 589-591.

Cornibert, J., and Marchessault, R. H. 1972. Physical properties of poly-ß-hydroxybutyrate. IV. Conformational analysis and crystalline structure. J. Mol. Biol. 71: 735-756.

Curley, J. M., Hazer, B., Lenz, R. W., and Fuller, R. C. 1996. Production of poly(3-hydroxyalkanoates) containing aromatic substituents by Pseudomonas oleovorans. Macromolecules 29: 1762-1766.

Davis, B. D., Luger, S. M., and Tai, P. C. 1986. Role of ribosome degradation in the death of starved Escherichia coli cells. J. Bacteriol. 166: 439-445.

Dawes, E. A., and Senior, P. J. 1973. Advances in microbiology physiology. Adv. Microbiol. Physiol. 10: 135-266.

Dawson, M. W., Maddox, I. S., and Brooks, J. D. 1989. Evidence for nitrogen catabolite repression during citric acid production by Aspergillus niger under phosphate limited growth conditions. Biotechnol. Bioeng. 33: 1500-1504.

de Jong-Gubbels, P., Vanrollehem, P., Heijnen, S., van Dijken, J. P., and Pronk, J. T. 1995. Regulation of carbon metabolism in chemostat cultures of Saccharomyces cerevisiae. Yeast 11: 407-418.

de Koning, G., Kellerhals, M., Van Meurs, C., and Witholt, B. 1996. A process for the production of bacterial medium-chain-length Poly[(R)-3-hydroxyalkanoates]: Reviewing the status quo. In: G. Eggink, A. Steinbüchel, Y. Poirier and B. Witholt (eds.), 1996 International symposium on bacterial polyhydroxyalkanoates, Davos, pp. 137-141.

de Koning, G. J. M., and Maxwell, I. A. 1993. Biosynthesis of poyhydroxyalkanoates: an emulsion polymerization. J. Environm. Polym. Degrad. 1: 223.

de Koning, G. J. M., van Bilsen, H. H. M., Lemstra, P. J., Hazenberg, W., Witholt, B., Preusting, H., van der Galiën, J. G., Schirmer, A., and Jendrossek, D. 1994. A biodegradable rubber by crosslinking poly(hydroxyalkanoate) from Pseudomonas oleovorans. Polymer 35: 2090-2097.

References 187

de Koning, G. J. M., and Witholt, B. 1996. A biodegradable rubber from bacteria, poly(hydroxyalkanoate) from Pseudomonas. Materials Science & Engineering C 4: 121-124.

de Smet, M. J., Eggink, G., Witholt, B., Kingma, J., and Wynberg, H. 1983. Characterization of intracellular inclusions formed by Pseudomonas oleovorans during growth on octane. J. Bacteriol. 154: 870-878.

Doi, Y. 1990. Microbial polyesters. VCH Publishers, New York.

Doi, Y., and Abe, C. 1990. Biosynthesis and characterization of a new bacterial copolyester of 3-hydroxyalkanoates and 3-hydroxy-chloroalkanoates. Macromolecules 23: 3705-3707.

Doi, Y., Kawaguchi, Y., Koyama, N., Nakamura, S., Hiramitsu, M., Yoshida, Y., and Kimura, H. 1992. Synthesis and degradation of polyhydroxyalkanoates in Alcaligenes eutrophus. FEMS Microb. Rev. 103: 103-108.

Doi, Y., Segawa, A., Kawaguchi, Y., and Kunioka, M. 1990. Cyclic nature of poly(3-hydroxyalkanoate) metabolism in Alcaligenes eutrophus. FEMS Microbiol. Lett. 67: 165-170.

Domínguez, H., Núñez, M. J., Chamy, R., and Lema, J. M. 1992. Determination of kinetic parameters of fermentation process by a continuous unsteady-state method: Application to the alcoholic fermentation of D-xylose by Pichia stipitis. Biotechnol. Bioeng. 41: 1129-1132.

Duchars, M. G., and Attwood, M. 1989. The influence of C:N ratio in the growth medium on the cellular composition and regulation of enzyme activity in Hyphomicrobium X. J. Gen. Microbiol. 135: 787-793.

Dunlop, W. F., and Robards, A. W. 1973. Ultrastructural study of poly-ß-hydroxybutyrate granules from Bacillus cereus. J. Bacteriol. 114: 1271-1280.

Durner, R. 1998. Feast and starvation: accumulation of bioplastic in Pseudomonas oleovorans. PhD thesis No. 12591, ETH Zürich, Zürich.

Eberl, L., Gisvskov, M., Sternberg, C., Moller, S., Christiansen, G., and Molin, S. 1996. Physiological responses of Pseudomonas putida KT2442 to phosphate starvation. Microbiol. 142: 155-163.

Eggink, G., de Waard, P., and Huijberts, G. N. M. 1992. The role of fatty acid biosynthesis and degradation in the supply of substrates for poly(3-hydroxyalkanoate) formation in Pseudomonas putida. FEMS Microb. Rev. 103: 159-164.

Eggink, G., Dewaard, P., and Huijberts, G. N. M. 1995. Formation of novel poly(hydroxyalkanoates) from long-chain fatty acids. Can. J. Microbiol. 41: 14-21.

Egli, T. 1982. Regulation of protein synthesis in methylotrophic yeasts: repression of methanol dissimilating enzymes by nitrogen limitation. Arch. Microbiol. 131: 95-101.

Egli, T. 1991. On multiple-nutrient-limited growth of microorganisms, with special reference to dual limitation by carbon and nitrogen substrates. Antonie van Leeuwenhoek 60: 225-234.

References 188

Egli, T., Bosshard, C., and Hamer, G. 1986. Simultaneous utilization of methanol-glucose mixtures by Hansenula polymorpha in chemostat: Influence of dilution rate and mixture composition on utilization in microorganisms. Biotechnol. Bioeng. 28: 1735-1741.

Egli, T., Käppeli, O., and Fiechter, A. 1982a. Mixed substrate growth of methylotrophic yeasts in chemostat culture: Influence of the dilution rate on the utilisation of a mixture of glucose and methanol. Arch. Microbiol. 131: 8-13.

Egli, T., Käppeli, O., and Fiechter, A. 1982b. Regulatory flexibility of methylotrophic yeasts in chemostat cultures: simultaneous assimilation of glucose and methanol at a fixed dilution rate. Arch. Microbiol. 131: 1-7.

Egli, T., Lindley, N. D., and Quayle, J. R. 1983. Regulation of enzyme synthesis and variation of residual methanol concentration during carbon-limited growth of Kloeckera sp. 2201 on mixtures of methanol and glucose. J. Gen. Microbiol. 129: 1269-1281.

Egli, T., and Quayle, J. R. 1986. Influence of the carbon : nitrogen ratio of the growth medium on the cellular composition and ability of the methylotrophic yeast Hansenula polymorpha to utilize mixed carbon sources. J. Gen. Microbiol. 132: 1779-1788.

Ellar, D., Lundgren, D. G., Okamura, K., and Marchessault, R. H. 1968. Morphology of poly-ß-hydroxybutyrate granules. J. Mol. Biol. 35: 489-502.

Findlay, R. H., Trexler, M. B., and White, D. C. 1990. Response of a benthic microbial community to biotic disturbance. Mar. Ecol. Prog. Ser. 62: 135-148.

Foster, L. J. R., Lenz, R. W., and Fuller, R. C. 1994. Quantitative determination of intracellular depolymerase activity in Pseudomonas oleovorans inclusions containing poly-3-hydroxyalkanoates with long alkyl substituents. FEMS Microbiol. Lett. 118: 279-282.

Foster, L. J. R., Stuart, E. S., Tehrani, A., Lenz, R. W., and Fuller, R. C. 1996. Intracellular depolymerase and polyhydroxyoctanoate granule integrity in Pseudomonas oleovorans. Int. J. Biol. Macromol. 19: 177-183.

Fritzsche, K., Lenz, R. W., and Fuller, R. C. 1990a. Bacterial polyesters containing branched poly(ß-hydroxyalkanoate) units. Int. J. Biol. Macromol. 12: 92-101.

Fritzsche, K., Lenz, R. W., and Fuller, R. C. 1990b. Production of unsaturated polyesters by Pseudomonas oleovorans. Int. J. Biol. Macromol. 12: 85-91.

Fritzsche, K., Lenz, R. W., and Fuller, R. C. 1990c. An unusual bacterial polyester with a phenyl pendant group. Makromol. Chem. 191: 1957-1965.

Fuller, R. C., O'Donnel, J. P., Saulnier, J., Redlinger, T. E., Foster, J., and Lenz, R. W. 1992. The supramolecular architecture of the polyhydroxyalkanoate inclusions in Pseudomonas oleovorans. FEMS Microbiol. Rev. 103: 279-288.

Gagnon, K. D., Lenz, R. W., Farris, R. J., and Fuller, R. C. 1992. Crystallization behavior and its influence on the mechanical properties of a thermoplastic elastomer produced by Pseudomonas oleovorans. Macromol. 25: 3723-3728.

References 189

Givskov, M., Eberl, L., and Molin, S. 1994a. Responses to nutrient starvation in Pseudomonas putida KT2442: Two-dimensional electrophoretic analysis of starvation- and stress-induced proteins. J. Bacteriol. 176: 4816-4824.

Givskov, M. L., Eberl, S., Møller, S., Poulsen, L. K., and Molin, S. 1994b. Responses to nutrient starvation in Pseudomonas putida KT2442 analysis of general cross protection, cell shape, and macromolecular content. J. Bacteriol. 176: 7-14.

Gottschal, J. C., and Kuenen, J. G. 1980. Mixotrophic growth of Thiobacillus A2 on acetate and thiosulfate as growth limiting substrates in the chemostat. Arch. Microbiol. 126: 33-42.

Grätzer-Lampart, S. D., Egli, T., and Hamer, G. 1986. Growth of Hyphomicrobium ZV620 in the chemostat: Regulation of NH4+-assimilating enzymes and cellular composition. J. Gen. Microbiol. 132: 3337-3347.

Gross, R. A., DeMello, C., Lenz, R. W., Brandl, H., and Fuller, R. C. 1989. Biosynthesis and characterization of poly(ß-hydroxyalkanoates) produced by Pseudomonas oleovorans. Macromolecules 22: 1106-1115.

Hänggi, U. J. 1995. Requirements on bacterial polyesters as future substitute for conventional plastics for consumer goods. FEMS Microbiol. Rev. 16: 213-220.

Harder, W., and Dijkhuizen, L. 1982. Strategies of mixed substrate utilization in microorganisms. Phil. Trans. R. Soc. Lond. B. 297: 459-480.

Harder, W., and Dijkhuizen, L. 1983. Physiological responses to nutrient limitation. Ann. Rev. Micribiol. 37: 1-23.

Hazenberg, W., and Witholt, B. 1997. Efficient production of medium-chain-length poly(3-hydroxyalkanoates) from octane by Pseudomonas oleovorans: economic considerations. Appl. Microbiol. Biotechnol. 48: 588-596.

Hazenberg, W. M. 1997. Production of poly(3-hydroxyalkanoates) by Pseudomonas oleovorans in two-liquid-phase media. Ph.D. Thesis, Federal Institute of Technology Zürich, Zürich, CH.

Hazer, B., Lenz, R. W., and Fuller, R. C. 1996. Bacterial production of poly-3-hydroxyalkanoates containing arylalkyl substituent groups. Polymer 37: 5951-5957.

Hepner, L. 1996. Cost analysis of fermentation processes. Chimia 50: 442-443.

Herbert, D. 1961. The chemical composition of micro-organisms as a function of their environment, pp. 391-416. In: M. C.G. and G. H. (eds.), Microbial reaction to the environment, Cambridge University Press, Cambridge, England.

Herbert, D. 1976. Stoichiometric aspects of microbial growth, pp. 1-30. In: A. C. R. Dean, D. C. Ellwood, C. G. T. Evans and J. Melling (eds.), Continuous culture 6: Applications and new fields, Ellis Horwood, Chichester.

Herbert, D., Phipps, D. J., and Strange, R. E. 1971. Chemical analysis of microbial cells. Methods in Microbiol. 5B: 209-344.

References 190

Herron, J. S., King, J. D., and White, D. C. 1978. Recovery of poly-ß-hydroxybutyrate from estuarine microflora. Appl. Environ. Microbiol. 35: 251-257.

Holmes, P. A. 1985. Applications of PHB - a microbially produced biodegradable thermoplastic. Phys. Technol. 16: 32-36.

Hori, K., Soga, K., and Doi, Y. 1994. Production of poly(3-hydroxyalkanoates-co-3-hydroxy-ω-fluoroalkanoates) by Pseudomonas oleovorans from 1-fluorononane and gluconate. Biotechnol. Lett. 16: 501-506.

Horowitz, D. M., Clauss, J., Hunter, B. K., and Sanders, J. K. M. 1993. Amorphous polymer granules. Nature 363: 23.

Hoyle, B. L., Scow, K. M., Fogg, G. E., and Darby, J. L. 1995. Effect of carbon: nitrogen ratio on kinetics of phenol biodegradation by Acinetobacter johnsonii in saturated sand. Biodegrad. 6: 283-293.

Huang, R. P., and Reusch, R. N. 1995. Genetic competence in Escherichia coli requires poly-ß-hydroxybutyrate calcium polyphosphate membrane complexes and certain divalent cations. J. Bacteriol. 177: 486-490.

Hueting, S., and Tempest, D. W. 1979. Influence of the glucose input concentration on the kinetics of metabolite production by Klebsiella aerogenes NCTC 418: growing in chemostat culture in potassium- or ammonium-limited environments. Arch. Microbiol. 123: 189-194.

Huijberts, G. N. M., Eggink, G., de Waard, P., Huisman, G. W., and Witholt, B. 1992. Pseudomonas putida KT2442 cultivated on glucose accumulates poly(3-hydroxyalkanoates) consisting of saturated and unsaturated monomers. Appl. Environ. Microbiol. 58: 536-544.

Huijberts, G. N. M., Rijk, T. C. d., Waard, P. d., and Eggink, G. 1995. 13C Nuclear magnetic resonance studies of Pseudomonas putida fatty acid metabolic routes involved in poly(3-hydroxyalkanoate) synthesis. J. Bacteriol. 176: 1661-1666.

Huisman, G. W. 1991. Poly(3-hydroxyalkanoates) from Pseudomonas putida: From DNA to plastic. PhD Thesis, Rijksuniversiteit Groningen, Groningen, the Netherlands.

Huisman, G. W., de Leeuw, O., Eggink, G., and Witholt, B. 1989. Synthesis of poly(3-hydroxyalkanoates) is a common feature of fluorescent Pseudomonads. Appl. Environ. Microbiol. 55: 1949-1954.

Huisman, G. W., Wonink, E., de Koning, G. J. M., Preusting, H., and Witholt, B. 1992. Synthesis of poly (3-hydroxyalkanoates) by mutant and recombinant Pseudomonas strains. Appl. Microbiol. Biotechnol. 38: 1-5.

Huisman, G. W., Wonink, E., Meima, R., Kazemier, B., Terpstra, P., and Witholt, B. 1991. Metabolism of poly(3-hydroxyalkanoates) (PHAs) by Pseudomonas oleovorans. J. Biol. Chem. 266: 2191-2198.

Jendrossek, D., Schirmer, A., and Schlegel, H. G. 1996. Biodegradation of polyhydroxyalkanoic acids. Appl. Microbiol. Biotechnol. 46: 451-463.

References 191

Kim, O., Gross, R. A., Hammar, W. J., and Newmark, R. A. 1996. Microbial synthesis of poly(beta-hydroxyalkanoates) containing fluorinated side-chain substituents. Macromolecules 29: 4572-4581.

Kim, Y. B., Lenz, R. W., and Fuller, R. C. 1992. Poly(β-hydroxyalkanoate) copolymers containing brominated repeating units produced by Pseudomonas oleovorans. Macromolecules 25: 1852-1857.

Kim, Y. B., Lenz, R. W., and Fuller, R. C. 1995. Poly-3-hydroxyalkanoates containing unsaturated repeating units produced by Pseudomonas oleovorans. J. Polym. Sci. A.-Polym. Chem. 33: 1367-1374.

Kim, Y. B., Rhee, Y.-H., Lenz, R. W., and Fuller, R. C. 1997. Poly(3-hydroxyalkanoate)s produced by Pseudomonas oleovorans grown by feeding nonanoic and 10-undecenoic acids in sequence. Polymer 29: 894-898.

Koch, A. L. 1982. Multistep kinetics: Choice of models for the growth of bacteria. J. Theor. Biol. 98: 401-417.

Koch, A. L., and Putnam, S. L. 1971. Sensitive Biuret method for determination of protein in an impure system such as whole bacteria. Anal. Biochem. 44: 239-245.

Kraak, M. N., Smits, T. H. M., Kessler, B., and Witholt, B. 1997. Polymerase C1 levels and poly(R-3-hydroxyalkanoate) synthesis in wild-type and recombinant Pseudomonas strains. J. Bacteriol. 179: 4985-4991.

Lageveen, R. G., Huisman, G. W., Preusting, H., Ketelaar, P., Eggink, G., and Witholt, B. 1988. Formation of polyesters by Pseudomonas oleovorans: effect of substrates on formation and composition of poly-(R)-3-hydroxyalkanoates and poly-(R)-3-hydroxyalkenoates. Appl. Environ. Microbiol. 54: 2924-2932.

Lapointe, B. E. 1985. Strategies for pulsed nutrient supply to Gracilaria cultures in the Florida Keys: Interactions between concentration and frequency of nutrient pulses. J. Exp. Mar. Biol. Ecol. 93: 211-222.

Lapointe, B. E., and Ryther, J. H. 1979. The effects of nitrogen and seawater flow rate on the growth and biochemical composition of Gracilaria foliifera v. angustissima in mass outdoor cultures. Bot. Mar. 22: 529-537.

Larsson, C., von Stockar, U., Marison, I., and Gustafsson, L. 1993. Growth and metabolism of Saccharomyces cerevisiae in chemostat cultures under carbon-, nitrogen-, or carbon- and nitrogen-limiting conditions. J. Bacteriol. 175: 4809-4816.

Lauzier, C., Marchessault, R. H., Smith, P., and Chanzy, H. 1992. Structural study of isolated poly(ß-hydroxybutyrate) granules. Polymer 33: 823-827.

Lee, A. L., Ataai, M. M., and Shuler, M. L. 1984. Double-substrate-limited growth of Escherichia coli. Biotechnol. Bioeng. 26: 1398-1401.

Lee, S. Y. 1996a. Bacterial polyhydroxyalkanoates. Biotechnol. Bioeng. 49: 1-14.

Lee, S. Y. 1996b. Plastic bacteria? Progress and prospects for polyhydroxyalkanoate production in bacteria. Trends Biotechnol. 14: 431-438.

References 192

Lemoigne, M. 1926. Produits de deshydration et de polymerisation de l'acide β-oxybutyric. Bull. Soc. Chem. Biol. 8: 770-782.

Lendenmann, U. 1994. Growth kinetics of Escherichia coli with mixtures of sugars. Ph.D. thesis No. 10658, Swiss Federal Institute of Technology Zürich, Zürich.

Lendenmann, U., Snozzi, M., and Egli, T. 1996. Kinetics of the simultaneous utilization of sugar mixtures by Escherichia coli in continuous culture. Appl. Environ. Microbiol. 62: 1493-1499.

León, J. A., and Tumpson, D. B. 1975. Competition between two species for two complementary or substitutable resources. J. Theor. Biol. 50: 185-201.

Lindsay, K. F. 1992. 'Truly biodegradable' resins are now truly commercial. Modern plastics 69: 62-64.

Ljunggren, J., and Häggström, L. 1994. Catabolic control of hybridoma cells by glucose and glutamine limited fed batch cultures. Biotechnol. Bioeng. 44: 808-818.

Lucca, M. E., Romero, M. E., Diaz-Ricci, J. C., Garro, O. A., and Callieri, D. A. S. 1991. Composition and morphology of Candida utilis grown in continuous culture with decreasing concentrations of phosphate. World J. Microbiol. Biotechnol. 7: 359-364.

Lundgren, D. G., Pfister, R. M., and Merrick, J. M. 1964. Structure of poly-ß-hydroxybutyric acid granules. J. Gen. Microbiol. 34: 441-446.

Malmstead, M. J., Brockman, F. J., Valocchi, A. J., and Rittmann, B. E. 1995. Modeling biofilm biodegradation requiring cosubstrates: the quinoline example. Wat. Sci. Tech. 31: 71-84.

Mankad, T., and Bungay, H. R. 1988. Model for microbial growth with more than one limiting nutrient. J. Biotechnol. 7: 161-166.

Mankad, T., and Nauman, E. B. 1992. Modeling of microbial growth under dual limitations. Chem. Eng. J. 48: B9-B11.

Marison, I. W., and von Stockar, U. 1987. A calorimetric investigation of the aerobic cultivation of Kluyveromyces on various substrates. Enzyme Microb. Technol. 9: 33.

Matin, A., Veldhuis, C., Stegeman, V., and Veenhuis, M. 1979. Selective advantage of a Spirillum sp. in a carbon-limited environment. Accumulation of poly-ß-hydroxybutyric acid and its role in starvation. J. Gen. Microbiol. 112: 349-355.

Mayer, F., and Hoppert, M. 1997. Determination of the thickness of the boundary layer surrounding bacterial PHA inclusion bodies, and implications for models describing the molecular architecture of this layer. J. Basic Microbiol. 37: 45-52.

Mayer, F., Madkour, M. H., Pieper-Fürst, U., Wieczorek, R., Gesell, M. L., and Steinbüchel, A. 1996. Electron microscopic observations on the macromolecular organization of the boundary layer of bacterial PHA inclusion bodies. J. Gen. Appl. Microbiol. Tokyo 42: 445-455.

References 193

Mccool, G. J., Fernandez, T., Li, N., and Cannon, M. C. 1996. Polyhydroxyalkanoate inclusion-body growth and proliferation in Bacillus megaterium. FEMS Microbiol. Lett. 138: 41-48.

McGrew, S. B., and Mallette, M. F. 1962. Energy of maintenance in Escherichia coli. J. Bacteriol. 88: 844-850.

Megee, R. D., Drake, J. F., Frederickson, A. G., and Tsuchiya, H. M. 1972. Studies in intermicrobial symbiosis, Saccharomyces cerevisiae and Lactobacillus casei. Can. J. Microbiol. 18: 1733-1742.

Merrick, J. M., and Doudoroff, M. 1964. Depolymerization of poly-ß-hydroxybutyrate by an intracellular enzyme system. J. Bacteriol. 88: 60-71.

Merrick, J. M., Lundgren, D. G., and Pfister, R. M. 1965. Morphological changes in poly-ß-hydroxybutyrate granules associated with decreased susceptibility to enzymatic hydrolysis. J. Bacteriol. 89: 234-239.

Minkevich, I. G., Krynitskaya, A. Y., and Eroshin, V. K. 1988. A double substrate limitation zone of continuous microbial growth, pp. 171-189. In: E. A. D. P. Kyslik, V. Krumphanzl, and M. Novak (eds.), Continuous culture, Academic Press, London.

Molitoris, H. P., Moss, S. T., de Koning, G. J. M., and Jendrossek, D. 1996. Scanning electron microscopy of polyhydroxyalkanoate degradation by bacteria. Appl. Microbiol. Biotechnol. 46: 570-579.

Monod, J. 1942. Recherches sur la croissance des cultures bactériennes. Hermann and Cie., Paris.

Monod, J. 1949. The growth of bacterial cultures. Ann. Rev. Microbiol. 3: 371-394.

Monod, J. 1950. La technique de culture continue; théorie et applications. Ann. Inst. Pasteur 79: 390-410.

Moser, H. 1958. The dynamics of bacterial populations maintained in the chemostat. Carnegie institution, Washington, D.C.,

Müller, R. H., Markuske, K. D., and W, B. 1985. Formate gradients as a means for detecting the maximum carbon conversion efficiency of heterotrophic substrates: Correlation between formate utilization and biomass increase. Biotechnol. Bioeng. 27: 1599-1602.

Narang, A., Konopka, A., and Ramkrishna, D. 1997. The dynamics of microbial growth mixtures of substrates in batch reactors. J. Theor. Biol. 184: 301-317.

Nawrath, C., Poirier, Y., and Somerville, C. 1995. Plant polymers for biodegradable plastics: cellulose, starch and polyhydroxyalkanoates. Mol. Breeding 1: 105-122.

Novick, A., and Szilard, L. 1950. Description of the chemostat. Science 112: 715-716.

Owens, J. D., and Legan, J. D. 1987. Determination of the Monod substrate saturation constant for microbial growth. FEMS Microbiol. Rev. 46: 419-432.

Page, W. J. 1995. Bacterial polyhydroxyalkanoates, natural biodegradable plastics with a great future. Can. J. Microbiol. 41: 1-3.

References 194

Pagni, M., Egger, L., and Aragno, M. 1995. The relationship between kinetics of substrate-limited transitions and steady-state growth in continuous cultures of Aquaspirillum autotrophicum limited by pyruvate. Antonie van Leeuwenhoek 68: 181-189.

Pirt, S. J. 1965. The maintenance energy of bacteria in growing cultures. Proc. Roy. Soc. B 163: 224-231.

Pirt, S. J. 1975. Principles of microbe and cell cultivation. Blackwell, London.

Poirier, Y., Dennis, D., Klomparens, K., Nawrath, C., and Somerville, C. 1992a. Perspectives on the production of polyhydroxyalkanoates in plants. FEMS Microbiol. Rev. 103: 237.

Poirier, Y., Dennis, D. E., Klomparens, K., and Somerville, C. 1992b. Polyhydroxybutyrate, a biodegradable thermoplastic, produced in transgenic plants. Science 256: 520-523.

Poirier, Y., Nawrath, C., and Somerville, C. 1995. Production of polyhydroxyalkanoates, a family of biodegradable plastics and elastomers, in bacteria and plants. Biotechnology 13: 142-150.

Pool, R. 1989. In search of the plastic potato. Science 245: 1187-1189.

Powell, E. O. 1967. The growth rate of microorganisms as a function of substrate concentration. In: C. T. G. Evans, R. E. Strange and D. W. Tempest (eds.), Microbial physiology and continuous culture (Proc. 3rd Intern. Symp.), Her Majesty's Stationery Office, London, pp. 34-56.

Preusting, H., Hazenberg, W., and Witholt, B. 1993a. Continuous production of poly(3-hydroxyalkanoates) by Pseudomonas oleovorans in a high-cell-density, two-liquid-phase chemostat. Enzyme Microb. Technol. 15: 311-316.

Preusting, H., Kingma, J., and Witholt, B. 1991. Physiology and polyester formation of Pseudomonas oleovorans in continuous two-liquid-phase cultures. Enzym. Microbiol. Technol. 13: 770-780.

Preusting, H., Nijenhuis, A., and Witholt, B. 1990. Physical characteristics of poly(3-hydroxyalkanoates) and poly(3-hydroxyalkenoates) produced by Pseudomonas oleovorans grown on aliphatic hydrocarbons. Macromolecules 23: 4220-4224.

Preusting, H., van Houten, R., Hoefs, A., Kool van Langenberghe, E., Favre-Bulle, O., and Witholt, B. 1993b. High cell density cultivation of Pseudomonas oleovorans: growth and production of poly (3-hydroxyalkanoates) in two-liquid phase batch and fed-batch systems. Biotechnol. Bioeng. 41: 550-556.

Ramakrishna, R., Ramkrishna, D., and Konopka, A. E. 1997. Microbial growth on substitutable substrates: characterizing the consumer-resource relationship. Biotechnol. Bioeng. 54: 77-90.

Ramsay, B. A., Saracovan, I., Ramsay, J. A., and Marchessault, R. H. 1991. Continuous production of long-side-chain poly-ß-hydroxyalkanoates by Pseudomonas oleovorans. Appl. Environ. Microbiol. 57: 625-629.

Reusch, R. N. 1995. Low molecular weight complexed poly(3-hydroxybutyrate): a dynamic and versatile molecule in vivo. Can. J. Microbiol. 41: 50-54.

References 195

Reusch, R. N., and Sadoff, H. L. 1988. Putative structure and functions of a poly-ß-hydroxybutyrate/calcium polyphosphate channel in bacterial plasma membranes. Proc. Natl. Acad. Sci. USA 85: 4176-4180.

Rittmann, B. E., and Dovantzis, K. 1983. Dual limitation of biofilm kinetics. Water Res. 17: 1727-1734.

Roels, J. A. 1983. Energetics and kinetics in biotechnology. Elsiever Biomedical, Amsterdam.

Rogers, A. H., Zilm, P. S., and Gully, N. J. 1987. Influence of arginine on the coexistence of Streptococcus mutans and S. milleri in glucose-limited mixed continuous culture. Microb. Ecol. 14: 193-202.

Rothen, S. A. R. 1997. Continuous bioconversion of octane to octanoic acid. PhD thesis No. 12082, ETH Zürich, Zürich.

Rutgers, M., Balk, P. A., and van Dam, K. 1990. Quantification of multiple-substrate controlled growth - simultaneous ammonium and glucose limitation in chemostat cultures of Klebsiella pneumoniae. Arch. Microbiol. 153: 478-484.

Sahm, H., and Wagner, F. 1973. Mikrobielle Verwertung von Methanol. Eigenschaften der Formaldehyddehydrogenase und Formiatdehydrogenase aus Candida boidinii. Arch. Mikrobiol. 90: 263-268.

Sakai, Y., Sawai, T., and Tani, Y. 1987. Isolation and characterization of a catabolite repression-intensive mutant of a methanol yeast Candida boidinii A5 producing alcohol oxidase in glucose-containing medium. Appl. Environ. Microbiol. 53: 1812-1818.

Sauer, U., Hatzimanikatis, V., Hohmann, H. P., Manneberg, M., van Loon, A. P. G. M., and Bailey, J. E. 1996. Physiology and metabolic fluxes of wild-type and ribolflavin-producing Bacillus subtilis. Appl. Environ. Microbiol. 62: 3687-3696.

Scheiner, D. 1976. Determination of ammonia and Kjeldahl nitrogen by indophenol method. Water Res. 10: 31-36.

Scholz, C., Wolk, S., Lenz, R. W., and Fuller, R. C. 1994. Growth and polyester production by Pseudomonas oleovorans on branched octanoic acid substrates. Macromolecules 27: 6358-6362.

Schwartz, R. D., and McCoy, C. J. 1973. Pseudomonas oleovorans hydroxylation-epoxidation system: additional strain improvements. Appl. Microbiol. 26: 217-218.

Schweiz, K.-V. 1997. Kunststoff-Statistik 1997. Bulletin des KVS 3: 1-18.

Senior, P. J., Beech, G. A., Ritchie, G. A. F., and Dawes, E. A. 1972. The role of oxygen limitation in the formation of poly-ß-hydroxybutyrate during batch and continuous culture of Azotobacter beijerinckii. Biochem. J. 128: 1193-1201.

Senn, H. 1989. Kinetik und Regulation des Zuckerabbaus von Escherichia coli bei tiefen Zuckerkonzentrationen. Ph.D. thesis No. 8831, Federal Institute of Technology Zürich, Zürich.

References 196

Senn, H., Lendenmann, U., Snozzi, M., Hamer, G., and Egli, T. 1994. The growth of Escherichia coli in glucose-limited chemostat cultures: A re-examination of the kinetics. Biochim. Biophys. Acta 1201: 424-436.

Slaff, G. F., and Humphrey, A. E. 1986. The growth of Clostridium thermohydrosulfuricum on multiple substrates. Chem. Eng. Commun. 45: 33-51.

Smith, A. L., Kelly, D. P., and Wood, A. P. 1980. Metabolism of Thiobacillus A2 grown under autotrophic, mixotrophic and heterotrophic conditions in chemostat culture. J. Gen. Microbiol. 121: 127-138.

Sonnleitner, B., Heinzle, E., Braunegg, G., and Lafferty, R. M. 1979. Formal kinetics of poly-ß-hydroxybutyric acid (PHB) production in Alcaligenes eutrophus H16 and Mycoplana rubra R14 with respect to the dissolved oxygen tension in ammonium-limited batch cultures. European J. Appl. Microbiol. Biotechnol. 7: 1-10.

Steinbüchel, A. 1991b. Polyhydroxyfettsäuren - thermoplastisch verformbare Polyester aus Bakterien. Nachr. Chem. Tech. Lab. 39: 1112-1123.

Steinbüchel, A., Aerts, K., Babel, W., Follner, C., Liebergesell, M., Madkour, M. H., Mayer, F., Pieperfürst, U., Pries, A., Valentin, H. E., and Wieczorek, R. 1995. Considerations on the structure and biochemistry of bacterial polyhydroxyalkanoic acid inclusions. Can. J. Microbiol. 41: 94-105.

Steinbüchel, A., Hustede, E., Liebergesell, M., Pieper, U., Timm, A., and Valentin, H. 1992. Molecular basis for biosynthesis and accumulation of polyhydroxyalkanoic acids in bacteria. FEMS Microbiol. Rev. 103: 217-230.

Steinbüchel, A., and Valentin, H. E. 1995. Diversity of bacterial polyhydroxyalkanoic acids. FEMS Microbiol. Lett. 128: 219-228.

Stouthamer, A. H. 1979. The search for correlation between theoretical and experimental growth yields. Int. Rev. Biochem. 21: 1-15.

Straight, J. V., and Ramkrishna, D. 1994. Cybernetic modeling and regulation of metabolic pathways. Growth on complementary nutrients. Biotechnol. Progr. 10: 574-587.

Stuart, E. S., Foster, L. J. R., Lenz, R. W., and Fuller, R. C. 1996. Intracellular depolymerase functionality and location in Pseudomonas oleovorans inclusions containing polyhydroxyoctanoate. Int. J. Biol. Macromol. 19: 171-176.

Stuart, E. S., Lenz, R. W., and Fuller, R. C. 1995. The ordered macromolecular surface of polyester inclusion bodies in Pseudomonas oleovorans. Can. J. Microbiol. 41: 84-93.

Suzuki, T., Yamane, T., and Shimizu, S. 1986a. Kinetics and effect of nitrogen source feeding on production of poly-ß-hydroxybutyric acid by fed-batch culture. Appl. Microbiol. Biotechnol. 24: 366-369.

Suzuki, T., Yamane, T., and Shimizu, S. 1986b. Mass production of poly-ß-hydroxybutyric acid by fed-batch culture with controlled carbon/nitrogen feeding. Appl. Microbiol. Biotech. 24: 370-374.

Tal, S., and Okon, Y. 1985. Production of the reserve material poly-ß-hydroxybutyrate and its function in Azospirillum brasilense Cd. J. Microbiol. 31: 608-613.

References 197

Tellenbach, M. 1998. Kunststoffabfälle - Wieviel Verwertung ist optimal?, Abfallwirtschaft - wie weiter, Tagung des KVS, Wildegg.

Thingstad, T. F. 1987. Utilization of N, P, and organic C by heterotrophic bacteria. I. Outline of a chemostat theory with a consistent concept of "maintenance" metabolism. Mar. Ecol. Prog. Ser. 21: 47-62.

Tilman, D. 1980. Resources: A graphical-mechanistic approach to competition and predation. Am. Nat. 116: 362-393.

van Beilen, J. B., Wubbolts, M. G., and Witholt, B. 1994. Genetics of alkane oxidation by Pseudomonas oleovorans. Biodegradation 5: 161-174.

van der Hoeven, H. S., and Gottschal, J. C. 1989. Growth of mixed cultures of Actinomyces viscosus and Streptococcus mutans under dual limitation of glucose and oxygen. FEMS Microbiol. Ecol. 62: 275-284.

van der Leij, F. R., and Witholt, B. 1995. Strategies for the sustainable production of new biodegradable polyesters in plants: a review. Can. J. Microbiol. 41: 222-238.

van Overbeek, L. S., Eberl, L., Molin, S., and van Elsas, J. D. 1995. Survival of, and induced stress resistance in, carbon-starved Pseudomonas fluorescens cells residing in soil. Appl. Environ. Biotechnol. 61: 4202-4208.

van Schie, B. J., Pronk, J. T., van Dijken, J. P., and Kuenen, J. G. 1993. Experimental and theoretical discrepancies in growth yields of Acinetobacter calcoaceticus: a correlation of published data. Appl. Microbiol. Biotechnol. 40: 270-272.

Vaulot, D., Olson, R. J., Merkel, S., and Chisholm, S. W. 1987. Cell-cycle response to nutrient starvation in two phytoplancton species, Thalassiosira weissflogii and Hymenonas carterae. Marine Biology (Berlin) 95: 625-630.

Veldkamp, H., and Jannasch, H. W. 1972. Mixed cultures with the chemostat. J. Appl. Chem. Biotechnol. 22: 105-123.

Volfová, O., Korínek, V., and Kyslíková, E. 1988. Alcohol oxidase in Candida boidinii on methanol-xylose mixtures. Biotechnol. Lett. 10: 643-648.

von Liebig, J. 1840. Organic chemistry in its application to agriculture and physiology. Taylor and Walton, London.

Wanner, U., and Egli, T. 1990. Dynamics of microbial growth and cell composition in batch culture. FEMS Microbiol. Rev. 75: 19-44.

Weusthuis, R. A., Visser, W., Pronk, J. T., Scheffers, W. A., and van Dijken, J. P. 1994. Effects of oxygen limitation on sugar metabolism in yeasts: A continuous-culture study of the Kluyver effect. Microbiology (Reading) 140: 703-715.

Wieczorek, R., Steinbüchel, A., and Schmidt, B. 1996. Occurrence of polyhydroxyalkanoic acid granule-associated proteins related to the Alcaligenes eutrophus H16 GA24 protein in other bacteria. FEMS Microbiol. Lett. 135: 23-30.

Williams, S. F., and Peoples, O. P. 1996. Biodegradable plastics from plants. Chemtech 26: 38-44.

References 198

Witholt, B., Eggink, G., and Huisman, G. W. 1994. Microbiological production of polyesters. US Patent, 5344769.

Wubbolts, M. G., Favre-Bulle, O., and Witholt, B. 1996. Biosynthesis of synthons in two-liquid-phase media. Biotechnol. Bioeng. 52: 301-308.

Yoshida, T., Izuishi, T., Gyusop, O., Inoue, T., Takagi, M., Fujiyama, K., and Seki, T. 1994. On-line monitoring and control of fed-batch culture of hybridoma with the aid of an expert system. Cytotechnology 14: 7.10.

199

CURRICULUM VITAE 28 June 1967 Born in Bern, Switzerland

1974-1978 Primary school, Sumiswald, Switzerland

1978-1981 Secondary school, Sumiswald, Switzerland

1981-1987 Gymnasium Burgdorf, Matura Typus C

1987-1992 Seasonal Flight Attendant Swissair

1988-1993 Study of Biotechnology, Swiss Federal Institute of Technology

(ETHZ): Diploma of Natural Sciences

1994-1998 Institute of Biotechnology, Swiss Federal Institute of Technology

(ETHZ) and Department of Microbiology, Swiss Federal Institute

of Environmental Science and Technology (EAWAG): Doctoral

Dissertation

1996 Marriage to Sandra Schärer

1997-1998 Center for Environmental Biotechnology, University of

Tennessee, USA: Research Associate

1998 Birth of Leonard S. Zinn

Documents

PSEUDOMONAS OLEOVORANS