Author’s Accepted Manuscript

A squalene synthase protein degradation method forimproved sesquiterpene production inSaccharomyces cerevisiae

Bingyin Peng, Manuel R. Plan, PanagiotisChrysanthopoulos, Mark P. Hodson, Lars K.Nielsen, Claudia E. Vickers

PII: S1096-7176(16)30242-7DOI: http://dx.doi.org/10.1016/j.ymben.2016.12.003Reference: YMBEN1191

To appear in: Metabolic Engineering

Received date: 21 April 2016Revised date: 17 November 2016Accepted date: 7 December 2016

Cite this article as: Bingyin Peng, Manuel R. Plan, Panagiotis Chrysanthopoulos,Mark P. Hodson, Lars K. Nielsen and Claudia E. Vickers, A squalene synthaseprotein degradation method for improved sesquiterpene production inSaccharomyces cerevisiae, Metabolic Engineering,http://dx.doi.org/10.1016/j.ymben.2016.12.003

This is a PDF file of an unedited manuscript that has been accepted forpublication. As a service to our customers we are providing this early version ofthe manuscript. The manuscript will undergo copyediting, typesetting, andreview of the resulting galley proof before it is published in its final citable form.Please note that during the production process errors may be discovered whichcould affect the content, and all legal disclaimers that apply to the journal pertain.

www.elsevier.com/locate/ymben

1

A squalene synthase protein degradation method for improved

sesquiterpene production in Saccharomyces cerevisiae

Bingyin Peng1, Manuel R. Plan

1,2, Panagiotis Chrysanthopoulos

1,2, Mark P. Hodson

1,2, Lars K.

Nielsen1, Claudia E. Vickers

1*

1Australian Institute for Bioengineering and Nanotechnology (AIBN), the University of Queensland,

St. Lucia, QLD 4072, Australia

2Metabolomics Australia (Queensland Node), the University of Queensland, St. Lucia, QLD 4072,

Australia.

*Corresponding author. c.vickers@uq.edu.au

Abstract

Sesquiterpenes are C15 isoprenoids with utility as fragrances, flavours, pharmaceuticals, and potential

biofuels. Microbial fermentation is being examined as a competitive approach for bulk production of

these compounds. Competition for carbon allocation between synthesis of endogenous sterols and

production of the introduced sesquiterpene limits yields. Achieving balance between endogenous

sterols and heterologous sesquiterpenes is therefore required to achieve economical yields. In the

current study, the yeast Saccharomyces cerevisiae was used to produce the acyclic sesquiterpene

alcohol, trans-nerolidol. Nerolidol production was first improved by enhancing the upstream

mevalonate pathway for the synthesis of the precursor farnesyl pyrophosphate (FPP). However,

excess FPP was partially directed towards squalene by squalene synthase (Erg9p), resulting in

squalene accumulation to 1 % biomass; moreover, the specific growth rate declined. In order to re-

direct carbon away from sterol production and towards the desired heterologous sesquiterpene, a

novel protein destabilisation approach was developed for Erg9p. It was shown that Erg9p is located

on endoplasmic reticulum and lipid droplets through a C-terminal ER-targeted transmembrane peptide.

2

A PEST (rich in Pro, Glu/Asp, Ser, and Thr) sequence-dependent endoplasmic reticulum-associated

protein degradation (ERAD) mechanism was established to decrease cellular levels of Erg9p without

relying on inducers, repressors or specific repressing conditions. This improved nerolidol titre by 86 %

to ~100 mg L-1

. In this strain, squalene levels were similar to the wild-type control strain, and

downstream ergosterol levels were slightly decreased relative to the control, indicating redirection of

carbon away from sterols and towards sesquiterpene production. There was no negative effect on cell

growth under these conditions. Protein degradation is an efficient mechanism to control carbon

allocation at flux-competing nodes in metabolic engineering applications. This study demonstrates

that an engineered ERAD mechanism can be used to balance flux competition between the

endogenous sterol pathway and an introduced bio-product pathways at the FPP node. The approach of

protein degradation in general might be more widely applied to improve metabolic engineering

outcomes.

Keywords: sesquiterpene, trans-nerolidol, metabolic engineering, Saccharomyces cerevisiae,

endoplasmic reticulum-associated protein degradation (ERAD)

1. Introduction

Sesquiterpenes are 15-carbon isoprenoid compounds with a wide variety of industrial applications,

including as fragrances and flavours; and as active molecules in pharmaceuticals, traditional herbs,

and personal care products. If cost-competitive production can be achieved, they may also be useful as

alternative fuels due to their high energy content. It is commonly uneconomical to produce bulk

amounts of sesquiterpenes from natural (usually plant) sources or via semi/total chemical synthesis.

Consequently, microbial fermentation from sugars using microbial cell factories is being pursued as a

competitive approach for their bulk production (Ajikumar et al., 2010; Scalcinati et al., 2012b; Shiba

et al., 2007; Vickers et al., 2015a).

Sesquiterpenes are synthesized biologically from the common precursor farnesyl pyrophosphate (FPP)

by specific sesquiterpene synthases. FPP is produced by condensation of the universal isoprenoid

3

backbone molecules, isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP;

Fig. 1). In S. cerevisiae, IPP and DMAPP are synthesized via the mevalonate (MVA) pathway from

acetyl-CoA, with consumption of NADPH and ATP. The MVA pathway is naturally tightly regulated

at transcriptional and post-transcriptional levels (Goldstein and Brown, 1990; Kuzuyama, 2002;

Miziorko, 2011), limiting the overall metabolic flux for sesquiterpene production. In order to increase

availability of FPP in the presence of a heterologous sesquiterpene synthase, the entire MVA pathway

as well as FPP synthase are typically overexpressed. This strategy is used both in yeast and in E. coli,

the latter of which uses the alternative methylerythritol phosphate pathway for native isoprenoid

production (Bongers et al., 2015; Ignea et al., 2011; Tsuruta et al., 2009; Yoon et al., 2009).

Fig. 1: The engineered mevalonate pathway and relevant downstream pathways, including

endogenous genes: ERG10, acetoacetyl-CoA thiolase; ERG13, 3-hydroxy-3-methylglutaryl-

CoA (HMG-CoA) synthase; tHMG1, truncated HMG-CoA reductase; ERG12, mevalonate

kinase; ERG8, phosphomevalonate kinase; MVD1, mevalonate pyrophosphate decarboxylase;

IDI1, isopentenyl diphosphate:dimethylallyl diphosphate isomerase; ERG20, farnesyl

Acetyl-CoA

Mevalonate

IPP DMAPP

FPP

Squalene

Ergosterol

Trans-nerolidol

“Upper”

mevalonate

pathway

“Lower”

mevalonate

pathway

ERG10 / EfmvaE

ERG13 / EfmvaS

tHMG1 / EfmvaE

ERG12

ERG8

MVD1

IDI1

ERG20

ERG20

AcNES1ERG9

Vitamin E/K

Teprenone

α-Sinensal

Chemical

synthesis

ERG1

4

pyrophosphate synthetase; ERG9, squalene synthase; ERG1, squalene epoxidase; and

exogenous genes: EfmvaS, Enterococcus faecalis HMG-CoA synthase; EfmvaE, Enterococcus

faecalis acetoacetyl-CoA thiolase/HMG-CoA reductase; AcNES1, Actinidia chinensis tran-

nerolidol synthase.

In addition to the low capacity for prenyl phosphate (IPP/DMAPP/FPP) synthesis, non-productive

consumption of FPP for sterol synthesis is a limitation for heterogeneous sesquiterpene production.

Sterol synthesis from FPP is initiated by the branch-point enzyme squalene synthase, which

condenses two molecules of FPP to squalene. Yeast strains with an enhanced MVA pathway exhibit

dramatically elevated levels of squalene and sterols (Liu et al., 2013; Polakowski et al., 1998). It is not

feasible to knock out squalene synthase (erg9) to decrease competition for production of novel

sesquiterpenes, because sterols are essential cell components; erg9 dysfunction causes ergosterol

auxotrophy. Conditional down-regulation of ERG9 at the transcriptional level has therefore been used

to constrain squalene/sterol synthesis and increase heterogeneous sesquiterpene production. This has

been achieved by replacing the native ERG9 promoter on the genome with various repressible

promoters: MET3 (repressed upon addition of methionine to the medium), CRT3 (repressed on high-

concentration copper) or HXT1 (induced by high glucose and reduced under low glucose;

transcription decreases after the exponential phase of batch cultivation) (Asadollahi et al., 2008;

Paddon et al., 2013; Scalcinati et al., 2012a; Scalcinati et al., 2012b; Xie et al., 2014). However, these

strategies require strict fermentation conditions and/or addition of repressors to achieve regulation,

conditions that can increase the complexity and/or cost in industrial processes.

Protein concentration is dynamically regulated in the cell by modulating the rates of transcription,

RNA degradation, translation and/or protein degradation (Babiskin and Smolke, 2011; Jungbluth et al.,

2010; Raina and Crews, 2010; Sekar et al., 2016). In metabolic engineering, specific protein levels are

typically controlled through transcription (at the promoter level); control mechanisms at other levels,

especially at the protein level, have not been well exploited in yeast. Protein degradation is a natural

mechansim widely used in many cellular/metabolic regulation processes, including cell cycling

(Sandoval et al., 2013) and sterol homeostasis (Christiano et al., 2014; Foresti et al., 2013; Meusser et

al., 2005). In S. cerevisiae, the squalene-consuming enzyme squalene epoxidase (Erg1p; Fig. 1) has a

5

short half-life (60-80 min) due to protein degradation, whereas squalene synthase (Erg9p) is more

stable, with a half life of 10 hours (Christiano et al., 2014; Foresti et al., 2013; Ruckenstuhl et al.,

2007). The short half-life of Erg1p provides a mechanism for the cell to rapidly shut down squalene

consumption when high flux is not required for downstream synthesis (Foresti et al., 2013).

Conversely, the stability of squalene synthase, presumably coupled with its ability to compete

effectively with introduced sesquiterpene synthases, results in squalene accumulation in the absence

of downstream consumption or when upstream flux is high (Liu et al., 2013; Polakowski et al., 1998).

The stability of squalene synthase also means that the squalene synthase node is less sensitive to

down-regulation using transcriptional approaches (Yuan and Ching, 2015), because a decrease in

cellular squalene synthase levels after transcritpional inhibition depends on subsequent dilution of the

Erg9p protein through cell division (Sekar et al., 2016). As a novel metabolic engineering strategy,

protein destabilization might be able to effect a rapid reduction in absolute enzyme concentration

(Jungbluth et al., 2010), thereby quickly diminishing competition at specific nodes.

In the current study, trans-nerolidol was used to investigate a series of engineering strategies for

improving sesquiterpene production in yeast. Trans-nerolidol is an important fragrance and flavour

(Berger, 2007). Nerolidol also exhibits antiulcer (Klopell et al., 2007), antileishmanial (Arruda et al.,

2005) and insect repelling (Di Campli et al., 2012) activities. It is the main component of the essential

oil and the biologically active component from fragrant rosewood (Dalbergia odorifera; Jiang Xiang

in Chinese) heart wood; D. odorifera essential oil is used in Traditional Chinese Medicine for

dispersing blood stasis in organs/tissues, and applied for treating coronary heart disease and angina

pectoris (Li et al., 2014). As an intermediate, nerolidol is used to synthesize vitamins E and K1

(Nissen et al., 1982), α-sinensal (a sweet orange aroma flavour, as in orange marmalade) (Ashurst,

2012; Berger, 2007), and teprenone (geranylgeranylacetone; a drug used for the treatment of gastric

ulcers) (Chen and Hu, 2004). In this work, core MVA pathway flux was first improved by over-

expression of native and heterologous pathway genes, leading to increased nerolidol production.

Nerolidol production was then specifically improved by engineering a protein destabilization

mechanism to reduce the FPP-consuming squalene synthase protein Erg9p, facilitating re-direction of

carbon from squalene to nerolidol.

6

2. Materials and Methods

2.1 Plasmid and Strain construction

Plasmids used in this work are listed in Table 1 and strains are listed in Table 2. Primers used

polymerase chain reaction (PCR) and PCR performed in this work are listed in Supplementary

Materials: Table S1. Plasmid construction processes are listed in Supplementary Materials: Table

S2. The genes encoding trans-nerolidol synthase (AcNES1; GenBank-KX064240) from Actinidia

chinensis (Green et al., 2012), and 3-hydroxy-3-methylglutaryl-Coenzyme A (HMG-CoA) synthase

(EfmvaS; GenBank-KX064238) and bi-functional acetoacetyl-CoA thiolase /HMG-CoA reductase

(EfmvaE; GenBank-KX064239) from E. faecalis (Wilding et al., 2000) were synthesized by

GenScript (Piscataway, NJ, USA) with codon optimization according to the codon bias of S.

cerevisiae (Supplementary Materials: Table S3).

Table 1: Plasmids used in this work

Table 2: S. cerevisiae strains used in this work

Plasmid Features Reference

pILGFP89S Yeast integration plasmid; PTEF1-yEGFP (Peng et al., 2015)

pILGFPJ1 Yeast integration plasmid; PTEF1-yEGFP-ERG9Sig

This work

pILGFPU0 Yeast integration plasmid; PTEF1-yEGFP-ERG9Sig

-

CLN2PEST

This work

pCEV-G1-Km E.coli/S. cerevisiae shuttle plasmid; 2μ, PPGK1-TCYC1 (Vickers et al.,

2013)

pRS423 E.coli/S. cerevisiae shuttle plasmid; 2μ, HIS3 (Christianson et al.,

1992)

pRS424 E.coli/S. cerevisiae shuttle plasmid; 2μ, TRP1 (Christianson et al.,

1992)

pRS425 E.coli/S. cerevisiae shuttle plasmid; 2μ, LEU2 (Christianson et al.,

1992)

pPMVAu5 pRS423: PSSA1-ERG10-TNAT1-PRPL4A-ERG13-TEFM1-

PRPL15A-tHMG1-TEBS1

This work

pPMVAu8 pRS423: PRPL4A-EfmvaS-TEFM1-PRPL15A-EfmvaE -TEBS1 This work

pPMVAd3 pRS424: PRPL8B-ERG12-TNAT5-PSSB1-ERG8-TIDP1-PRPL3-

MVD1-TIDP1-PYEF3-IDI1-TRPL15A

This work

pJT1 pRS425: PTEF2-ERG20-TRPL3-PTEF1-AcNES1-TRPL41B This work

pJT1R pRS425: PTEF2-ERG20-TRPL3 This work

7

Strain Genotype Resource/reference

CEN.PK2-1C MATa ura3-52 trp1-289 leu2-3,112 his3Δ 1 (Entian and Kötter,

1998)

CEN.PK113-5D MATa ura3-52 (Entian and Kötter,

1998)

CEN.PK113-7D MATa (Entian and Kötter,

1998)

ILHA series strains:

oURA3 CEN.PK2-1C derivative; URA3 This work

Oref oURA3 derivative; [pRS423] [pRS424] [pRS425] This work

N1 oURA3 derivative; [pRS423] [pRS424] [pJT1] This work

N3 oURA3 derivative; [pPMVAu5] [pPMVAd3] [pJT1] This work

N6 ILHA oURA3 derivative; [pPMVAu8] [pPMVAd3]

[pJT1]

This work

O6 ILHA oURA3 derivative; [pPMVAu8] [pPMVAd3]

[pJT1R]

This work

GH4 CEN.PK113-5D derivative; ura3(1,704 )::KlURA3 (Peng et al., 2015)

G89S CEN.PK113-5D derivative; ura3(1, 704 )::KlURA3-

PTEF1-yEGFP

(Peng et al., 2015)

G89 CEN.PK113-5D derivative; ura3(1, 704)::KlURA3-

PTEF1-yEGFP-CLN2PEST

(Peng et al., 2015)

GJ1 CEN.PK113-5D derivative; ura3(1, 704)::KlURA3-

PTEF1-yEGFP-ERG9Sig

This work

GJ13GH GJ1 derivative; [pCEV-G1-Km(PPGK1-SEC63-mRFP1-

TCYC1)]

This work

GU0 CEN.PK113-5D derivative; ura3(1, 704)::KlURA3-

PTEF1-yEGFP-ERG9Sig

-CLN2PEST

This work

oH5 oURA3 derivative; ERG9(1333, 1335)::yEGFP-

CLN2PEST

-TURA3-loxP-KlURA3-loxP

This work

oH5C oH5 derivative; [pRS423] [pRS424] [pRS425] This work

oH6C oURA3 derivative; ERG9(1333, 1335)::yEGFP-TURA3-

loxP-KlURA3-loxP [pRS423] [pRS424] [pRS425]

This work

N6D oH5 derivative; [pPMVAu8] [pPMVAd3] [pJT1] This work

Nerolidol-producing strains were constructed using S. cerevisiae CEN.PK2-1C (Table 2). The URA3

gene was amplified from CEN.PK113-7D genomic DNA (Supplementary Materials: Table S1, #18)

and was transformed to replace the dysfunctional ura3 gene in CEN.PK2-1c. To modify the Erg9p

carboxyl-terminal (C-terminal) sequence, ERG9C-terminal

-yEGFP-CLN2PEST

-TURA3-LoxP-KlURA3-LoxP-

TERG9 and ERG9C-terminal

-yEGFP- TURA3-LoxP-KlURA3-LoxP- TERG9 were constructed through over-lap

extension PCR (Supplementary Materials: Table S1, #19 & 20). The constructs included green

fluorescence protein (GFP) with/without the G1 cyclin degradation domain (GLN2PEST

), a selectable

marker gene (URA3 from Kluyveromyces lactis), and two genomic integration sites (ERG9C-terminal

and

TERG9) targeted to the C-terminal and terminator of ERG9. The resulting fragments were transformed

into CEN.PK2-1c to generate two strains, ILHA oH5 (with CLN2PEST

) and ILHA oH6 (without

8

CLN2PEST

). Furthering strains were generated by co-transforming with the plasmids as indicated in

Table 2. For nerolidol-producing strains, biological replication was achieved by storing at least three

individual transformants from transformation plates in 20% glycerol stock at -80 C, as described

previously (Bruschi et al., 2012).

The strains GJ1 and GU0 expressing yEGFP-ERG9Sig

(GFP fused with Erg9p C-terminal 34 residues)

and yEGFP-ERG9ER

-CLN2PEST

respectively, were constructed by transforming SwaI-digested

plasmids (pILGFPJ1 and pILGFPU0 respectively) into CEN.PK113-5D. The strain GJ13GH was

generated by co-transforming the BamHI-digested pCEV-G1-Km and PCR fragments SEC63 and the

red fluorescent protein (RFP) mRFP1 (http://parts.igem.org/Part:BBa_J61031) for the in-vivo plasmid

assembly.

2.2 Two-phase flask cultivation and sampling

Nerolidol-producing strains were evaluated through two-phase flask cultivation (Vickers et al., 2015b).

Synthetic minimal (SM) medium (containing 6.7 g L-1

yeast nitrogen base, YNB, Sigma-Aldrich

#Y0626; pH 6.0) with 20 g L-1

glucose as the carbon source was used (SM-glucose). 100 mM 2-(N-

morpholino) ethanesulfonic acid (MES, Sigma-Aldrich#M8250) was used to buffer medium pH and

the initial pH was adjusted to 6.0 by adding ammonium hydroxide. Strains were recovered from

glycerol stocks by streaking on SM-glucose agar plates and pre-cultured in MES-buffered SM-

glucose medium to exponential phase (cell density OD600 between 1 and 4). Pre-cultured cells were

collected by centrifugation and resuspend in fresh media before initiating two-phase cultivations.

Two-phase flask cultivation was initiated by inoculating pre-cultured cells to OD600 = 0.2 in 25 ml

MES-buffered SM-glucose medium in 250 ml flasks with solvent-resistant screw-caps (Vickers et al.,

2015b); 2 ml dodecane was added to extract nerolidol. Flask cultivation was performed at 30 C and

200 rpm. For all cultivations, 1 ml culture was sampled at 0 hour; over the first 10 hr, 5 x 0.4 ml

samples were taken to measure cell density; and at 72 hour, dodecane and culture were sampled for

metabolite analysis (see below). For RNA extraction, 2 ml exponential-phase culture (OD600 = 1 to 1.5)

was sampled and cells were collected and stored at -80 C. The cells from the final cultures (t = 72

9

hour) were washed twice with distilled water and aliquots were stored at -20 C for analysis of

squalene and ergosterol.

2.3 Confocal microscopy

Yeast GJ13GH were grown in YPD broth (20 g L-1

peptone, 10 g L-1

yeast extract and 20 g L-1

glucose) with 500 μg mL-1

G418 to exponential phase (OD600 = 1.0) and fixed in paraformaldehyde

solution as described previously (Mazumder et al., 2013). Nucleic acid was dyed with 4',6-diamidino-

2-phenylindole (DAPI). Yeast cells were mounted on a poly-lysine bed on a glass slide. Fluorescent

(DAPI, GFP and RFP) and transparent (T PMT) channels were photographed using an LSM 710

confocal microscope (Zeiss) at the Queensland node of the Australian National Fabrication Facility.

2.4 GFP fluorescence and degradation assay

To monitor GFP fluorescence during the early exponential phase (OD600 = 0.4 to 0.8), cells were

cultivated aerobically in 20 ml SM-glucose medium in a 100 ml flask. To monitor GFP fluorescence

over the entire batch cultivation, cells were cultivated aerobically in MES-buffered 20 ml SM-glucose

medium in 100 ml flasks. Samples were collected at indicated time points and GFP fluorescence in

single cells was analysed immediately after sampling using a flow cytometer (BD AccuriTM

C6; BD

Biosciences, USA). GFP fluorescence was excited by a 488 nm laser and monitored through a 530/20

nm band-pass filter; 10,000 events were counted per sample. The GFP fluorescence signal was

corrected according to the values of forward scatter and side scatter (Peng et al., 2015). GFP

fluorescence level was expressed as the percentage relative to the background auto-fluorescence from

a GFP-negative reference strain GH4 (empty vector integration strain). For the GFP degradation assay,

cycloheximide was added into the culture to a final concentration of 50 mg L-1

.

2.5 Quantitation of mRNA level

Total RNA was extracted using a yeast RiboPure™ RNA Purification Kit (Ambion #AM1926). After

DNase treatment, 0.1 or 0.2 μg total RNA was used for first-strand cDNA synthesis in a 10 μl reaction

10

using ProtoScript® II Reverse Transcriptase (NEB #M0368). The diluted cDNA was used as the

template for quantitative real-time (QRT) PCR (primers are listed in Supplementary Materials:

Table S1, # 23 to # 38). KAPA SYBR® FAST qPCR Kit (KaPa Biosystems#KP-KK4601) and

CFX96 Touch™ Real-Time PCR Detection System (BIO-RAD) were used in QRT-PCR. Yeast

genomic or plasmid DNA was used for preparing the standard curve. Ct values were analysed using

CFX Manager software (Bio-Rad Laboratories, QLD, Australia). The mRNA levels (N QRT-PCR

copies pg-1

total RNA) were calculated by referring to the standard curve equations.

2.6 Extraction of ergosterol and squalene

In order to extract ergosterol and squalene, methanolic pyrogallol saponification (Adams and Parks,

1968) was performed. The cell pellets were resuspended in 400 μl alkaline methanol (2 mol L-1

potassium hydroxide) and 100 μl internal standard solution (400 μmol L-1

pyrene and 10 g L-1

pyrogallol in methanol). Saponification was performed at 80 C in 2.2 mL screw-capped tubes for 1

hour with mixing for 30 second every 15 minutes. Distilled water (500 μl) was added to dilute

methanol saponification mixture and lipids were extracted with 1 ml hexane by regular vortex mixing

over a period of 30 minutes. The upper 750 μl of the hexane phase was desiccated using a vacuum-

vent centrifuge, and the pellet was resuspended in 150 μl methanol:ethanol (1:1) solution containing 2

g L-1

pyrogallol for HPLC analysis. Standards of ergosterol and squalene were prepared with 200

μmol L-1

pyrene as internal standard, and pyrogallol was used as an anti-oxidant.

2.7 Metabolite analysis

Ergosterol, squalene and pyrene were analysed through reverse-phase chromatography using an

Agilent 1200 HPLC system and a Gemini C18 column (150 x 4.6mm, 3µm, 110Å, Phenomenex PN:

00F-4439-E0) with a guard column (SecurityGuard Gemini C18, Phenomenex PN: AJO-7597).

Analytes were eluted isocratically at 1 mL min-1

with methanol at 35 oC. Analytes were monitored

using a diode array detector (Agilent Technologies) at the following UV wavelengths: 200 nm (for

squalene), 274 nm (for ergosterol) and 334 nm (for pyrene).

11

For trans-nerolidol analysis, dodecane samples were diluted 40-fold using hexane spiked with a β-

pinene internal standard. Diluted samples were analysed with gas chromatography-mass spectrometry

using an Agilent 7890A GC system at Metabolomics Australia (Queensland Node). Samples (3 μL)

were injected in split mode (ratio 1:30) at 250 C; analytes were separated on a FactorFour VF-5ms

GC column (30m*0.25um*0.25um+10m EZ-guard; Agilent Technologies, PN: CP9013) with 1 mL

min-1

helium flow and the following over temperature gradient program: 50 C for 2 min, +15 C min-

1 to 200 C, 200 C for 3 min, +40 C min

-1 to 325 C, 325 C for 0 min. Analytes were ionized and

fragmented using electron impact (EI) ionization source and detected by mass spectrometry using a

mass selective detector (MSD; Agilent Technologies#5975C).

A novel HPLC method was developed to analyse geraniol, linalool, trans,trans-farnesol, trans-

nerolidol, geranylgeraniol, and geranyllinalool. Dodecane samples from fermentations were diluted in

a 40-fold volume of ethanol. These samples (20μl) were analysed using an Agilent 1200 HPLC

system with a Zorbax Extend C18 column (4.6 x 150 mm, 3.5 µm, Agilent PN: 763953-902) and a

guard column (SecurityGuard Gemini C18, Phenomenex PN: AJO-7597). Analytes were eluted at 35

oC at 0.9 ml min

-1 using a mixture of solvent A (high purity water,18.2 kΩ) and solvent B (45%

acetonitrile, 45% methanol and 10% water), with a linear gradient of 5-100% solvent B from 0-24

min, then 100% from 24-30 min, and finally 5% from 30.1-35 min. Analytes of interest were

monitored using a diode array detector (Agilent DAD SL, G1315C) at 202 nm wavelength. Spectral

scans were also performed on each of the compounds from 190-400 nm in steps of 2 nm to confirm

their identity and purity.

The standards of geraniol (98% purity; Sigma-Aldrich # 163333), linalool (97% purity; Sigma-

Aldrich # L2602), trans,trans-farnesol (96% purity; Sigma-Aldrich # 277541), trans-nerolidol (93.7%

purity; Sigma-Aldrich #04610590), geranylgeraniol (85% purity; Sigma-Aldrich # G3278), and

geranyllinalool (95% purity; Sigma-Aldrich # 48809), were used to prepare the standard curve for

quantification.

12

3. Results and Discussion

To optimize heterologous sesquiterpene production in yeast, it is necessary to increase metabolic flux

flowing through the MVA pathway, and to redirect flux from competing native isoprenoid production

towards the desired end product (Asadollahi et al., 2008; Paddon et al., 2013; Tsuruta et al., 2009;

Vickers et al., 2015a; Yoon et al., 2009). In metabolic engineering, transcriptional regulation is

commonly employed to control enzyme levels (and consequently, metabolic flux) in desired pathways.

Transcriptional outputs can be tuned by choosing promoters with different strength and/or expression

patterns, or dynamically regulated by applying synthetic molecular circuits (Peng et al., 2015;

Williams et al., 2015a; Williams et al., 2015b; Xie et al., 2014). Here, we examined a novel approach

to modulate metabolic flux through destabilisation of a competing enzyme at a key node. In order to

provide an industrially-relevant comparison between engineered strains intended for cheap industrial

processes (Hahn-Hägerdal et al., 2005), all strains were evaluated on synthetic minimal mineral

medium (without any amino acid supplementation) with 20 g L-1

glucose as the carbon source. Genes

for flux modification were expressed from plasmids, with the exception of the Erg9p tagging, which

was done by editing the chromosomal Erg9 gene.

3.1 Increasing trans-nerolidol production by enhancing the MVA pathway

The production of sesquiterpenes in yeast is primarily limited by the weak synthetic capacity of the

native MVA pathway (Vickers et al., 2015a). Initial strain engineering therefore focussed on

increasing flux through the MVA pathway by overexpression of pathway genes (Fig. 1; Fig. 2a). FPP

synthase (ERG20) was first introduced to improve flux to FPP, in conjunction with nerolidol synthase

(AcNES1) from kiwifruit (A. chinensis) to catalyse nerolidol production. Strain N1, harbouring the

nerolidol synthesis plasmid pJT1 (Fig. 2a; containing ERG20 under the control of the TEF2 promoter

and the codon-opitmized AcNES1 under the control of TEF1 promoter) successfully produced

nerolidol, but at low titres of only ~5 mg L-1

(Fig. 2b). No nerolidol was detected from the reference

strain (Fig. 2b; strain Oref, harbouring empty plasmid vectors).

13

Fig. 2: Characterisation of nerolidol-producing strains: (a) schematic of over-expression

cassettes on plasmids pJT1, pPMVAu5 (yeast upper MVA pathway genes; yMVAupper

),

pPMVAu8 (E. faecalis upper MVA pathway genes; EfMVAupper

), and pPMVAd3 (yeast lower

MVA pathway genes; yMVAlower

) and the endoplasmic reticulum-associated protein

degradation model for squalene synthase Erg9p destabilization: yEGFP, yeast enhanced green

florescence protein; Cln2pPEST, PEST sequence (rich in Pro, Glu/Asp, Ser, and Thr) from the

0

10

20

30

N6 O6 N6D

FarnesolGeranylgeraniol

pPMVAd3

pPMVAu8

pPMVAu5

pJT1ERG20 AcNES1

TRPL41BTRPL3PTEF2 PTEF1

EfmvaS EfmvaE

TEBS1TEFM1PRPL4A PRPL15A

ERG12 ERG8 MVD1

TPRM9TIDP1TNAT5PRPL8B PSSB1 PRPL3IDI1

PYEF3TRPL15A

ERG10 ERG13 tHMG1

TEBS1TEFM1TNAT1PSSA1 PRPL4A PRPL15A

0

50

100

150

200

Ne

roli

do

l (m

g L

-1)

Sqalene

synthase

destabilization

Endoplasmic

reticulum

Erg9p yEGFPCln2p

PEST

Cytosol

Membrane Protein

degradation

C-terminalN-terminal

0

0.1

0.2

0.3

0.4

Oref N1 N3 N6 O6 N6Dμ

ma

x (

ho

ur-

1)

0

5

10

15

20

25

0 20 40 60 80

OD

600

Time (hour)

Oref N1N3 N6O6 N6D

a b

c

d

e

Belo

w

dete

ctio

n

s

s

s

n

sn

s

0

2

4

6

8

10

12

14

16

Lo

g2(m

RN

A l

ev

el)

OrefN3N6

N.D

.N

.D.

N.D

.N

.D.

N.D

.

Belo

w

LO

Q

Belo

w d

ete

ctio

n

Fa

rne

so

l,

ge

ran

ylg

era

nio

l(m

g L

-1)

f

14

G1 cyclin Cln2p, which leads protein degradation; (b) nerolidol titres at 72 hour, quantified

through GC method (except that in O6, nerolidol was analysed through HPLC method and

below the limit of quantification); (c) maximum specific growth rates (μmax) during the

exponential phase; (d) growth curves over the batch cultivation; (e) titres of farnesol and

geranylgeraniol at 72 hour, quantified through HPLC method; (f) mRNA levels during the

exponential phase (OD600 = 1.0 to 1.5) for strains Oref, N3 and N6: the difference of mRNA

levels of reference genes ACT1 and PDA1 is not significant (two-tailed t-test p > 0.1). Mean

values ± standard deviations are shown (N ≥ 3). Two-tailed Welch's t-test was used for

statistical analysis: n (not significant), p > 0.1; s (significant), p < 0.05. N.D., not detected.

Augmentation of the MVA pathway was tested next. Comparative studies in E. coli have examined

relative efficiency of MVA pathway genes from different species; in this system, the pathway is

commonly divided into expression modules on either side of the core intermediate mevalonate,

resulting in an “upper” part and a “lower” part (Fig. 1). The upper MVA pathway, especially the

reaction catalysed by HMG-CoA reductase, is regarded as the main flux bottleneck (Paddon and

Keasling, 2014). A truncated version of HMG-CoA reductase (tHMG1) lacking the membrane-

binding domain is used to avoid degradation-mediated feedback regulation at that node (Polakowski

et al., 1998; Shimada et al., 1998). The E. coli studies have demonstrated that the “upper” module

genes from E. faecalis are superior to S. cerevisiae genes, whereas “lower” module genes from S.

cerevisiae and Streptococcus pneumoniae operate at similar efficiency, and are more efficient than the

genes from E. faecalis, Staphylococcus aureus and Streptococcus pyogenes genes (Tsuruta et al., 2009;

Yoon et al., 2009). Several modules were examined to test if augmentation of the “upper” MVA

module in yeast is more efficient using genes from E. faecalis as opposed to the equivalent yeast

genes.

A yeast “upper” module plasmid, pPMVAu5 (including yeast genes ERG10, ERG13 and tHMG1) and

an E. faecalis upper MVA module plasmid, pPMVAu8 (including the codon-optimized E. faecalis

genes EfmvaS and EfmvaE, the latter being a bi-functional acetoacetyl-CoA thiolase/HMG-CoA

reductase) were first constructed (Fig. 2a). These modules were coupled with a “lower” MVA module

15

plasmid, pPMVAd3 (yeast ERG12, ERG8, MVD1 and IDI1) and the nerolidol synthesis plasmid pJT1

(Fig. 2a).

Augmentation of the core MVA pathway by over-expression of the native yeast genes increased

nerolidol titre by 3.5-fold, to ~17.5 mg L-1

(Fig. 2b, strain N3 compared to strain N1). Heterologous

expression of the E. faecalis “upper” MVA module increased nerolidol titre by a further ~3-fold, to 56

mg L-1

(Fig. 2b, strain N6 compared to strain N3). This demonstrates that in yeast, as in E. coli

(Yoon et al., 2009), the E. faecalis genes are more effective than the yeast genes. However, both

MVA pathway-supplemented strains exhibited decreased maximum specific growth rate (μmax)

compared to strain N1 (Fig. 2c-d; p<0.01). Compared to strain N1 (μmax = 0.3), strain N3 grew ~30 %

more slowly (0.22 h-1

), and strain N6 grew at about half the rate (0.14 h-1

). These data confirm that

supplementation of the MVA pathway improves accumulation of end products; however, they also

demonstrate that pathway over-expression has a negative effect on growth rate.

Native terpene synthases have not been found in S. cerevisiae; despite this, yeast can produce terpene

alcohols (notably farnesol and some nerolidol), and production increases dramatically when FPP

pools are increased by either over-expression of a truncated HMG-CoA reductase and/or deletion of

squalene synthase (Carrau et al., 2005; Grabowska et al., 1998; Karst and Lacroute, 1977; McNeil,

2008; McNeil et al., 2012; Zhuang, 2013). Proposed mechanisms for native terpenoid production

include activity of FPP synthase, which can catalyse solvolysis of GPP to produce geraniol and FPP to

produce nerolidol and farnesol (Poulter and Rilling, 1976; Saito and Rilling, 1979); non-specific

phosphatase/pyrophosphatase activity (Chambon et al., 1990); and non-enzymatic hydrolysis (McNeil,

2008). Non-enzymatic hydrolysis has been shown to be a major mechanism, and adjusting the

medium pH to titrate farnesol:nerolidol ratios can yield nerolidol titres of up to 50 mg/L (McNeil,

2008; McNeil et al., 2012). In squalene synthase knock-out strains with enhanced uptake of

exogenous sterols and tHMGR over-expression, farnesol can accumulate up to 140 mg L-1

in 10-day

cultivation on 20 g L-1

glucose (Zhuang, 2013). Over-expression of a strawberry (Fragaria ×

ananassa) NES1 (FaNES1) effectively converted this farnesol to nerolidol (Zhuang, 2013). In some

cases, native conversion of accumulated FPP to farnesol can outcompete a heterologously-expressed

terpene synthase (Takahashi et al., 2007).

16

We therefore examined terpene alcohol production in the engineered strains (Supplementary

Materials: Fig. S1). AcNES1 is known to have additional linalool/geranyllinalool synthase activities

(Green et al., 2012); however, geraniol, linalool, farnesol, and geranyllinalool, were all below the

limit of dectection in strain N6 (Supplementary Materials: Fig. S1). Conversely, a geranylgeraniol

peak was detected (Supplementary Materials: Fig. S1). To examine this phenomenon more closely,

a control strain overexpressing the MVA pathway genes and FPP synthase (plasmids pPMVAu8,

pPMVAd3, and pJT1R) but without the AcNES1 gene, was made (strain O6). Strain O6 produced

significantly lower geranylgeraniol (~0.7 mg L-1

), compared to strain N6 (~9 g L-1

; Fig. 2e). This

indicates that geranylgeraniol can be produced in the absence of the nerolidol synthase, but that the

nerolidol synthase also has geranylgeraniol synthase activity. This is surprising considering that

exactly the same protein sequence was used as in the previous study, where linalool and

geranyllinalool, but not geranylgeraniol, were reported (Green et al., 2012). We presume that

differences in in vitro conditions used previously, compared to our in vivo experiment, are responsible

for this discrepancy.

No trans-nerolidol was observed in strain O6, confirming that trans-nerolidol production is the result

of AcNES1 activity. Instead, trans, trans-farnesol was detected in strain O6 (Supplementary

Materials: Fig. S1) with a titre of ~11.7 mg L-1

(Fig. 2e). This is consistent with previous reports

describing production of farnesol in strains over-expressing the mevalonate pathway with or without a

sesquiterpene synthase (Muramatsu et al., 2009; Scalcinati et al., 2012a; Song, 2003). This activity is

attributed to the presense of endogeneous phosphatase(s) (Farhi et al., 2011; Scalcinati et al., 2012a)

The improved performance of the E. faecalis genes (strain N6) might be due to either higher catalytic

efficiency of the enzymes encoded by EfmvaS and EfmvaE relative to the cognate yeast genes

(ERG10/ERG13 and tHMG1, respectively), higher transcription/translation rate, or higher

protein/mRNA stability. To examine relative transcript levels, the mRNA levels for genes involved in

nerolidol production were analysed during exponential phase using QRT-PCR (Fig. 2f). Data are

presented in log scale to facilitate direct comparisons. Transcripts for control genes actin (ACT1) and

pyruvate dehydrogenase (PDA1) were similar (two-tailed t-test p > 0.1) in all strains in the

exponential phase. The fold-changes (compared to the reference strain Oref) for the over-expressed

17

lower MVA pathway genes were similar in strains N3 (over-expressing yeast upper and lower MVA

modules) and N6 (over-expressing E. faecalis upper and yeast lower MVA modules) - as expected,

considering the same expression plasmids was used in both strains.

For the “upper” MVA pathway genes, the transcript levels in strain N3 were increased by 10-fold for

ERG10 and 5-fold for ERG13 and HMG1 relative to in the reference strain Oref. However, the

observed transcript levels for the E. faecalis genes in strain N6 were significantly higher than those of

the over-expressed yeast upper MVA module genes in strain N3: transcription levels of EfmvaS and

EfmvaE in strain N6 were respectively 17-fold and 24-fold higher than the levels of genes ERG13 and

HMG1 in strain N3. Although the constructs were controlled by the same promoter/terminator

systems for both the yeast and E. faecalis modules (Fig. 2a), differences in mRNA levels may be due

to stability of different transcripts (especially considering that the E. faecalis genes are not native

genes while the yeast genes are native), or to idiosyncratic characteristics of the ORFs/constructs and

their compatibility with specific promoters and terminators (Carquet et al., 2015; Yamanishi et al.,

2013). Regardless of the mechanism, these data demonstrate that expression of the yeast genes is

limited at the RNA level relative to the E. faecalis genes. This presumably results in improved

pathway flux thus contributing to the improved nerolidol titre in the E. faecalis module strain, and

also confirms that the upper part of the pathway is rate-limiting in this system.

Despite our findings here, it has been previously been shown that the yeast HMG-CoA synthase

(Erg13p) has higher catalytic efficiency than E. faecalis enzyme (EfmvaS) (Middleton, 1972; Skaff et

al., 2012; Vannice et al., 2013). This suggests that the improved catalytic efficiency is not able to

overcome the penalty conferred by decreased transcriptional efficiency. The kinetic parameters for

yeast HMG-CoA reductase are not available.

3.2 Minimising competition for FPP by destabilising squalene synthase

Controlling carbon flux around FPP, the flux competing node, is required to improve accumulation of

the desired sesquiterpenes. For metabolic engineering applications, transcription/RNA-level

engineering has been used to regulate the flux-competing nodes, e.g., selecting a weaker promoter

18

(Scalcinati et al., 2012a) or applying RNA interference (Williams et al., 2015a). To direct more FPP

toward sesquiterpene production, transcription of the native squalene synthase gene (ERG9) is

commonly decreased by replacing the native promoter with a weaker promoter or a promoter that can

be specifically down-regulated (Asadollahi et al., 2008; Paddon et al., 2013; Scalcinati et al., 2012a;

Scalcinati et al., 2012b; Xie et al., 2014). Here we investigated a novel alternative strategy to regulate

the level of Erg9p: a protein degradation mechanism to decrease protein half-life.

3.2.1 The C-terminal of Erg9p is a signal peptide targeting the endoplasmic reticulum and lipid

droplets

Human squalene synthase was shown to be bound to the endoplasmic reticulum (ER) membrane

through C-terminal residues (Thompson et al., 1998). Consistent with this, the TMHMM Server

online tool (Krogh et al., 2001) predicted that Erg9p has a C-terminal transmembrane domain with the

N-terminal residues lying in the non-cytosol side of the membrane and C-terminal residues located

across the membrane and in the cell cytosol (Fig. 3a). To test if the C-terminal residues could direct

localisation to the ER membrane, the C-terminal 34 residues of Erg9p were attached to the C-terminal

of yeast enhanced green fluorescent protein (yEGFP), and the fusion was expressed in yeast.

Fluorescent microscopy showed that GFP was located in the perinuclear/peripheral organelles, a

region recognised as the ER (Feldheim et al., 1992). Consistent with this, the GFP flouorescence in

this region overlapped with red fluorescence from RFP-tagged Sec63p (Fig. 3b) , a protein known to

localise to the ER (Feldheim et al., 1992). In addition to the ER, GFP signal was also observed in

optically bright regions connecting to the ER (Fig. 3b; differential interference contrast imaging).

These regions have been reported as lipid droplets (LDs) (Bouchez et al., 2015; Wolinski et al., 2011).

Consistent with this localisation pattern, several other enzymes from ergosterol synthetic pathway,

including Erg1p, Erg7p, and Erg27p, are also co-located in ER and LDs (Goodman, 2009; Leber et al.,

1998). As observed previously (Llopis et al., 1998; Mateus and Avery, 2000), un-tagged GFP is

located in both cytosol and nucleus (Supplementary Materials: Fig. S2). These data indicate that

the C-terminal transmembrane domain of Erg9p function as an signal peptide targeting to ER and LDs.

19

Fig.3: Modification of the C-terminal of squalene synthase (Erg9p): (a) Erg9p transmembrane

domain prediction via TMHMM Server v. 2.0 (http://www.cbs.dtu.dk/services/TMHMM/); (b)

imaging of GJ13GH cells expressing GFP labelled with the C-terminal 34 residues of Erg9p

(ERG9Sig

) and Sec63p-mRFP1; (c) GFP fluorescence levels in the strains G89S (expressing

yEGFP), G89 (expressing yEGFP- CLN2PEST

), GJ1(expressing yEGFP-ERG9Sig

) and GU0

(expressing yEGFP-ERG9Sig

-CLN2PEST

); (d) GFP degradation analysis after adding

cycloheximide; (e) GFP fluorescence levels in strains with Erg9p C-terminal labelled with

yEGFP- Cln2pPEST

(oH5C) or yEGFP (oH6C) over the whole batch cultivation. Mean values ±

standard deviations are shown (N = 3).

0

1000

2000

3000

4000

5000

Without

CLN2PEST

With

CLN2PEST

yEGFPyEGFP-

ERG9Sig

Th

e r

ela

tive

GF

P f

luo

res

ce

nc

e

(%A

uto

-flu

ore

sc

en

ce

)0

20

40

60

80

100

0 15 30 45 60

Time (min)

yEGFP

yEGFP-ERG9Sig

yEGFP-CLN2PEST

yEGFP-ERG9Sig -

CLN2PEST

0

5

10

15

20

25

0

100

200

300

400

0 20 40 60 80

OD

60

0

Time (Hour)

GFP (ERG9-yEGFP- CLN2PEST)

GFP (ERG9-yEGFP)

OD600 (ERG9-yEGFP- CLN2PEST)

OD600 (ERG9-yEGFP)

a b c

d e

Th

e r

ela

tive

GF

P f

luo

res

ce

nc

e

(%A

uto

-flu

ore

sc

en

ce

)

Th

e G

FP

flu

ore

sc

en

ce

re

sid

ue

(%

)

0

0.2

0.4

0.6

0.8

1

0 100 200 300 400

Pro

bab

ilit

y

Residue

InsideOutsideTransmembrane

20

3.2.2 Engineering CLN2PEST sequence-dependent degradation of ER-associated protein with the

Erg9p signal peptide

In order to reasonably minimize Erg9p level, we exploited a post-transcriptional regulation strategy—

increasing protein degradation rate of Erg9p. Cytosolic proteins can be degraded through either N-

terminal or C-terminal destabilization peptides (Jungbluth et al., 2010; Varshavsky, 1996). ER-

associated protein degradation (ERAD) is an important cellular pathway that targets misfolded ER

proteins for degradation. The pathway includes mechanisms for targeting protein substrates,

retrotranslocating the substrate to cytosol, ubiquitination at specific residues, and finally degradation

via proteasomes (Lemus and Goder, 2014; Meusser et al., 2005). The Erg9p transmembrane peptide

has histidine and arginine residues at the carboxyl end of the transmembrane, indicating that this end

of the peptide is located in the cytosol (the “positive-inside rule”) (Goder et al., 2004; Krogh et al.,

2001). C-terminal ER-associated degradation was therefore identified as a potential degradation

strategy. Several cytosol-side sequences rich in Pro, Glu (and Asp), Ser, and Thr amino acids (known

as PEST seqeunces) have been shown to be responsible for the degradation of ER-located enzymes

(Arteaga et al., 2006; Kato et al., 2006). To exploit this process for engineering Erg9p degradation, we

labeled the cytosolic end of the Erg9p transmembrane domain with the G1 cyclin PEST sequence

(CLN2PEST

). The yeast G1 cyclin protein is very unstable, a characteristic that is partially determined

by C-terminal PEST motif; and the G1 cyclin PEST sequence is sufficient to destabilise heterologous

cytosolic proteins (Mateus and Avery, 2000; Salama et al., 1994).

We first tested the PEST sequence-dependent ERAD model using ER-localized yEGFP. yEGFP

constructs with or without the ERG9 ER signal peptide (ERG9Sig

) and/or CLN2PEST

were expressed

under the control of the TEF1 promoter. ERG9Sig

decreased GFP fluorescence by 33% (Fig. 3c), but

did not affect GFP half-life: both the yEGFP construct strain and the yEGFP-ERG9Sig

construct strain

demonstrated very stable GFP activity following addition of cycloheximide to prevent translation,

with half-lives well in excess of the 60 min that the experiment was performed over (Fig. 3d). The

CLN2PEST

fusion resulted in a significant decrease in GFP fluorescence in both the control yEGFP

construct strain (10-fold) and in the yEGFP-ERG9Sig

construct strain (8-fold) (Fig. 3c). Presence of

the CLN2PEST

decreased the GFP fluorescence half-life to just 10 min for the yEGFP construct strain

21

and to ~25 min for the yEGFP-ERG9Sig

construct strain (Fig. 3d). We presume that the GFP

fluorescence behaves as a proxy for decreased protein half-life, indicating that the CLN2PEST

sequence

can successfully destabilise the ER membrane-integrating proteins when fused to the cytosolic C-

terminal end, resulting in rapid protein degradation.

To test the effect of a C-terminal CLN2PEST

peptide tag on Erg9p, both yEGFP and yEGFP-CLN2PEST

were inserted at the 3′ end of the chromosomal ERG9 gene (Fig. 2a). GFP fluorescence was used as a

proxy for the relative abundance of the resulting fusion proteins (Chong et al., 2012; Newman et al.,

2006). Interestingly, the Erg9p-yEGFP strain exhibited a dramatic decrease in GFP fluorescence level

in the post-exponential growth phase (Fig. 3e). This is in contrast to previous studies, which showed

that β-galactosidase activity driven by the ERG9 promoter was stable throughout batch cultivation

(Scalcinati et al., 2012a). In support of this, transcription of ERG9 decreased less than 15% relative to

exponential phase transcription at 36 hours in the nerolidol-producing strains N6 (ERG9) and N6D

(ERG9-yEGFP-CLN2PEST

); in contrast, transcription of other tested genes decreased significantly

more (ACT1, ERG1 and ERG2 by 50 %; and PDA1 by 30%; Supplementary Materials: Fig. S3).

These results support natural regulation of squalene synthase at the post-transcriptional level

(Devarenne et al., 2002; Otero et al., 2010). In addition, β-galactosidase is cytosolic, whereas the

Erg9p-yEGFP protein is ER-associated. ER-associated proteins can be degraded through an

autophagy process (Mochida et al., 2015; Schuck et al., 2014) which is induced during the diauxic

shift (Galdieri et al., 2010). These two considerations may account for the post-transcriptional

regulation on Erg9p observed in the current study. Supporting this idea, the fluorescence pattern we

observed is consistent with dynamic squalene accumulation observed in an MVA pathway-enhanced

strain, which showed a decrease in squalene content in the ethanol growth phase relative to in the

exponental growth phase (Mantzouridou and Tsimidou, 2010). These data suggest that squalene

synthase is down-regulated as cell growth slows down during the diauxic shift.

The presence of the CLN2PEST

tag (strain oH5C; ERG9-yEGFP-CLN2PEST

) resulted in an ~4-fold

decrease in GFP fluorescence in the exponential phase, and an ~2-fold decrease in the ethanol phase,

compared to the control (strain oH6C; ERG9-yEGFP; Fig. 3e). Fluorescence levels still decreased

over the diauxic shift, but by a relatively small amount compared to the sharp decrease seen in the

22

Erg9p-yEGFP strain. This suggests that the destabilisation approach successfully decreased fusion

protein accumulation throughout the batch cultivation, and particularly during exponential phase. The

modifications did not influence the cell growth (strain Oref in Fig. 2c; strains oH5C and OH6C in Fig.

3e), suggesting that a decrease in Erg9p levels during the exponential phase can be well tolerated by

the cell. This implies that squalene synthase levels might be naturally above the level required for

normal cell growth during batch culture.

3.2.3 Destabilisation of Erg9p increases nerolidol production

To test for improved availability of FPP for nerolidol synthesis, plasmids harbouring the E. faecalis

upper MVA pathway module, the yeast lower MVA pathway module, and ERG20 / NES1 were

introduced into the Erg9p-destabilized background strain oH5, to produce strain N6D. Strain N6D

exhibited an increase in nerolidol titre of 86 % compared to strain N6, to 105 mg L-1

(Fig. 2b). In

contrast with strain N6, a farnesol peak was detected in strain N6D (Supplementary Materials: Fig.

S1); the titre was ~23.5 mg L-1

(Fig. 2e). We presume that this is due to accumulation of intracellular

FPP due to a reduction in squalene sythase activity. Similarly to strain N6, geranylgeraniol was also

produced in strain N6D, at a titre of ~12 mg L-1

.

The MVA pathway genes and FPP synthase gene (pPMVAu8, pPMVAd3, and pJT1R) were

introduced to the Erg9p-destabilized strain. The transformants grew on SM-glucose agar, but

exhibited a growth deficiency characterised by instability and very slow growth rates in broth (data

not shown). We presume this is due to over-accumulation of FPP,which can cause cellular toxicity

(Dahl et al., 2013).

The fold-increase achieved using Erg9p destabilization is comparable to the fold-increase observed

for other terpene end-products when using the low-glucose-repressed HXT1 promoter to control

ERG9 (Scalcinati et al., 2012a; Xie et al., 2014). The advantage for protein destablization strategy is

that its regulation does not depend on having a low glucose concentration or addition of a repressor –

as the HXT1/MET3/CRT3 promoters require. This is beneficial in an industrial process. Also, no

difference in growth rate between strain N6 and strain N6D was observed (Fig. 2c-d: N6D vs. N6).

23

This suggets that destabilisation of Erg9p (and presumably concommitent increase in FPP availability)

was successfully achieved, and can increase sesquitepene levels without further inhibiting growth in

the presence of an over-expressed MVA pathway.

3.4 Erg9p destabilisation redirects carbon from sterol synthesis towards heterologous

sesquiterpene production

Squalene accumulates at up to 1 % cell dry weight in strains with an enhanced MVA pathway (Liu et

al., 2013; Polakowski et al., 1998). Successful redirection of carbon towards heterologous

sesquiterpenes at the FPP node should result in decreased squalene accumulation coincident with

increased sesquiterpene titres. We examined both squalene and downstream ergosterol accumulation

(Fig. 4) to determine the effects of different engineering steps on sterol biosynthesis.

Fig. 4: Intracellular squalene (a) and ergosterol (b) contents in nerolidol-producing yeast strains

at 72 h. Mean values ± standard deviations are shown (N = 3). Two-tailed Welch's t-test was

used for statistical analysis: n (not significant), relative to Oref, p > 0.1; s (significant), relative

to Oref, p < 0.05.

In the strain N3, harbouring the over-expressed yeast MVA pathway and exhibiting relativey low

levels of nerolidol (Fig. 2b), the squalene and ergosterol levels did not change significantly relative to

the reference strain Oref (Fig. 4). This suggests that pathway flux was not strongly increased

compared to the wild type, and that any excess pathway flux overflowed to nerolidol or other sterol

0

2

4

6

8

10

12

14

Oref N3 N6 N6D

0

2

4

6

8

10

12

14

Oref N3 N6 N6DSq

ua

len

e (

mg

g-1

bio

ma

ss

)

Erg

oste

rol

(mg

g-1

bio

mass)

nn

s

p=0.03

n n

s

p=0.04

a b

24

intermediates. Heterologous expression of the the E. faecalis upper MVA module (strain N6)

produced ~3-fold more nerolidol (Fig. 2b) and ~10-fold more squalene (Fig. 4a), consistent with

previous studies showing accumulation of squalene in pathway-engineered strains with high flux (Liu

et al., 2013; Polakowski et al., 1998). Ergosterol levels were not significantly affected in this strain

(Fig. 4b).

Destabilisation of squalene synthase (strain N6D, harbouring the E. faecalis module and the

destabilised Erg9) caused a significant decrease in squalene relative to strain N6, to levels similar to

the reference strain Oref (Fig. 4a). Moreover, ergosterol pools were also affected, with levels

decreasing by 35% compared strain N6 and 43% compared to strain Oref. This is consistent with

decreased ergosterol levels observed in strains with transcriptional down-regulation of squalene

synthase using the methionine-repressible MET3 promoter or the low-glucose-repressible HXT1

promoter (Asadollahi et al., 2010; Scalcinati et al., 2012a). Raw mass balance calculations indicated

the decreased squalene/ergosterol contents could contribute to ~30 mg L-1

increase in sesquiterpene

(nerolidol + farnesol) titre. Here, ~70 mg L

-1 improvement in in N6D relative to N6 was observed,

suggesting that other sterol intermediates may also be re-directed to nerolidol production in the

destabilised strain. These data indicate that destabilization of Erg9p can redirect carbon at the FPP

node away from squalene/sterol production and towards alternative sesquiterpene production in the

presence of an enhanced MVA pathway.

4. Summary and Conclusion

In this study, we first improved MVA pathway flux by over-expression/heterologous expression of

pathway genes. As in E. coli (Tsuruta et al., 2009; Yoon et al., 2009), the heterologous E. faecalis

genes in the upstream module of the MVA pathway provided better pathway flux than the native yeast

genes (Fig. 2b). We further demonstrated that improved transcript stability is likely a contributing

factor to the higher efficiency of the E. faecalis genes (Fig. 2f). Further improvement of trans-

nerolidol production was achieved by regulating the flux-competing enzyme squalene synthase

(Erg9p) via a novel engineered ERAD mechanism. We demonstrated that the C-terminal region of

Erg9p behaved as a signal peptide partially targeting to ER membrane; then showed that fusing a

25

PEST sequence to the cytosolic ER-integration end of a protein can drive an ERAD response,

significantly shortening protein half-life (Fig. 3). This protein destabilizing strategy was then applied

to reduce squalene synthase level; this improved the trans-nerolidol production and also resulted in

some farnesol production by redirecting carbon from sterol synthesis and into sesquiterpene synthesis

(Fig. 2b & Fig. 4). In contrast to previous transcriptionally-mediated approaches to regulate squalene

synthase (Asadollahi et al., 2008; Paddon et al., 2013; Scalcinati et al., 2012a; Scalcinati et al., 2012b;

Xie et al., 2014), the destabilization strategy does not require either addition of repressors which may

increase the cost of the bioprocess (in the case of repressible promoters), or appropriately repressing

fermentation conditions (in the case of glucose-responsive promoter). Destabilisation of squalene

synthase does not inhibit the maximum specific growth rate. Protein degradation is a novel

mechanism to effectively control carbon allocation at flux-competing nodes in metabolic engineering

applications. In future studies, titration of squalene synthase levels by modifying expression levels

might improve carbon allocation balance, further improving heterologous sesquiterpene levels.

Balancing the upstream pathway over-expression might also avoid growth inhibition in pathway over-

expression strains.

Authors’ contributions

BP, LKN and CEV designed the experiments; BP carried out the experiments; MP developed HPLC

method and performed the HPLC analysis; PC developed the GC-MS method and performed the GC-

MS analysis; MPH contributed to development of analytics methods; BP, LKN and CEV drafted and

revised the manuscript; all authors contributed to the result analysis and the discussion of the research.

All authors read and approved the final manuscript.

Acknowledgements

This work was performed in part at the Queensland node of the Australian National Fabrication

Facility, a company established under the National Collaborative Research Infrastructure Strategy to

26

provide nano- and micro-fabrication facilities for Australia’s researchers. BP was supported by a

University of Queensland International Postgraduate Research Scholarship. CEV was supported by

Queensland Government Smart Futures and Accelerate Fellowships. We thank Prof. Dr.-Ing. Lars

Blank and Dr.-Ing. Birgitta Ebert (RWTH-Aachen) for useful discussions.

References

Adams, B. G., Parks, L. W., 1968. Isolation from yeast of a metabolically active water-soluble form of ergosterol. Journal of lipid research. 9, 8-11.

Ajikumar, P. K., Xiao, W. H., Tyo, K. E., Wang, Y., Simeon, F., Leonard, E., Mucha, O., Phon, T. H., Pfeifer, B., Stephanopoulos, G., 2010. Isoprenoid pathway optimization for Taxol precursor overproduction in Escherichia coli. Science. 330, 70-4.

Arruda, D. C., D'Alexandri, F. L., Katzin, A. M., Uliana, S. R., 2005. Antileishmanial activity of the terpene nerolidol. Antimicrobial agents and chemotherapy. 49, 1679-87.

Arteaga, M. F., Wang, L., Ravid, T., Hochstrasser, M., Canessa, C. M., 2006. An amphipathic helix targets serum and glucocorticoid-induced kinase 1 to the endoplasmic reticulum-associated ubiquitin-conjugation machinery. Proceedings of the National Academy of Sciences of the United States of America. 103, 11178-83.

Asadollahi, M. A., Maury, J., Moller, K., Nielsen, K. F., Schalk, M., Clark, A., Nielsen, J., 2008. Production of plant sesquiterpenes in Saccharomyces cerevisiae: effect of ERG9 repression on sesquiterpene biosynthesis. Biotechnology and bioengineering. 99, 666-77.

Asadollahi, M. A., Maury, J., Schalk, M., Clark, A., Nielsen, J., 2010. Enhancement of farnesyl diphosphate pool as direct precursor of sesquiterpenes through metabolic engineering of the mevalonate pathway in Saccharomyces cerevisiae. Biotechnology and bioengineering. 106, 86-96.

Ashurst, P. R., 2012. Food flavorings. Springer Science & Business Media. Babiskin, A. H., Smolke, C. D., 2011. A synthetic library of RNA control modules for predictable tuning

of gene expression in yeast. Molecular systems biology. 7, 471. Berger, R. G., 2007. Flavours and fragrances: chemistry, bioprocessing and sustainability. Springer

Science & Business Media. Bongers, M., Chrysanthopoulos, P. K., Behrendorff, J. B. Y. H., Hodson, M. P., Vickers, C. E., Nielsen, L.

K., 2015. Systems analysis of methylerythritol-phosphate pathway flux in E. coli: insights into the role of oxidative stress and the validity of lycopene as an isoprenoid reporter metabolite. Microbial cell factories. 14, 1-16.

Bouchez, I., Pouteaux, M., Canonge, M., Genet, M., Chardot, T., Guillot, A., Froissard, M., 2015. Regulation of lipid droplet dynamics in Saccharomyces cerevisiae depends on the Rab7-like Ypt7p, HOPS complex and V1-ATPase. Biol Open. 4, 764-75.

Bruschi, M., Boyes, S. J., Sugiarto, H., Nielsen, L. K., Vickers, C. E., 2012. A transferable sucrose utilization approach for non-sucrose-utilizing Escherichia coli strains. Biotechnology advances. 30, 1001-1010.

Carquet, M., Pompon, D., Truan, G., 2015. Transcription interference and ORF nature strongly affect promoter strength in a reconstituted metabolic pathway. Frontiers in bioengineering and biotechnology. 3, 21.

27

Carrau, F. M., Medina, K., Boido, E., Farina, L., Gaggero, C., Dellacassa, E., Versini, G., Henschke, P. A., 2005. De novo synthesis of monoterpenes by Saccharomyces cerevisiae wine yeasts. FEMS Microbiology Letters. 243, 107-115.

Chambon, C., Ladeveze, V., Oulmouden, A., Servouse, M., Karst, F., 1990. Isolation and properties of yeast mutants affected in farnesyl diphosphate synthetase. Current Genetics. 18, 41-46.

Chen, H.-R., Hu, Z.-L., 2004. Improved Synthesis of Teprenone. Chinese Journal of Pharmaceuticals. 35, 449-450.

Chong, Y. T., Cox, M. J., Andrews, B., 2012. Proteome-Wide Screens in Saccharomyces cerevisiae Using the Yeast GFP Collection. In: Goryanin, I. I., Goryachev, B. A., Eds.), Advances in Systems Biology. Springer New York, New York, NY, pp. 169-178.

Christiano, R., Nagaraj, N., Frohlich, F., Walther, T. C., 2014. Global proteome turnover analyses of the Yeasts S. cerevisiae and S. pombe. Cell reports. 9, 1959-65.

Christianson, T. W., Sikorski, R. S., Dante, M., Shero, J. H., Hieter, P., 1992. Multifunctional yeast high-copy-number shuttle vectors. Gene. 110, 119-22.

Dahl, R. H., Zhang, F., Alonso-Gutierrez, J., Baidoo, E., Batth, T. S., Redding-Johanson, A. M., Petzold, C. J., Mukhopadhyay, A., Lee, T. S., Adams, P. D., Keasling, J. D., 2013. Engineering dynamic pathway regulation using stress-response promoters. Nature biotechnology. 31, 1039-46.

Devarenne, T. P., Ghosh, A., Chappell, J., 2002. Regulation of squalene synthase, a key enzyme of sterol biosynthesis, in tobacco. Plant physiology. 129, 1095-106.

Di Campli, E., Di Bartolomeo, S., Delli Pizzi, P., Di Giulio, M., Grande, R., Nostro, A., Cellini, L., 2012. Activity of tea tree oil and nerolidol alone or in combination against Pediculus capitis (head lice) and its eggs. Parasitology research. 111, 1985-92.

Entian, K.-D., Kötter, P., 1998. 23 yeast mutant and plasmid collections. Methods in microbiology. 26, 431-449.

Farhi, M., Marhevka, E., Masci, T., Marcos, E., Eyal, Y., Ovadis, M., Abeliovich, H., Vainstein, A., 2011. Harnessing yeast subcellular compartments for the production of plant terpenoids. Metabolic engineering. 13, 474-481.

Feldheim, D., Rothblatt, J., Schekman, R., 1992. Topology and functional domains of Sec63p, an endoplasmic reticulum membrane protein required for secretory protein translocation. Molecular and cellular biology. 12, 3288-96.

Foresti, O., Ruggiano, A., Hannibal-Bach, H. K., Ejsing, C. S., Carvalho, P., 2013. Sterol homeostasis requires regulated degradation of squalene monooxygenase by the ubiquitin ligase Doa10/Teb4. eLife. 2, e00953.

Galdieri, L., Mehrotra, S., Yu, S., Vancura, A., 2010. Transcriptional regulation in yeast during diauxic shift and stationary phase. Omics : a journal of integrative biology. 14, 629-38.

Goder, V., Junne, T., Spiess, M., 2004. Sec61p contributes to signal sequence orientation according to the positive-inside rule. Molecular biology of the cell. 15, 1470-8.

Goldstein, J. L., Brown, M. S., 1990. Regulation of the mevalonate pathway. Nature. 343, 425-30. Goodman, J. M., 2009. Demonstrated and inferred metabolism associated with cytosolic lipid

droplets. Journal of lipid research. 50, 2148-56. Grabowska, D., Karst, F., Szkopinska, A., 1998. Effect of squalene synthase gene disruption on

synthesis of polyprenols in Saccharomyces cerevisiae. FEBS letters. 434, 406-8. Green, S. A., Chen, X., Nieuwenhuizen, N. J., Matich, A. J., Wang, M. Y., Bunn, B. J., Yauk, Y. K.,

Atkinson, R. G., 2012. Identification, functional characterization, and regulation of the enzyme responsible for floral (E)-nerolidol biosynthesis in kiwifruit (Actinidia chinensis). Journal of experimental botany. 63, 1951-1967.

Hahn-Hägerdal, B., Karhumaa, K., Larsson, C. U., Gorwa-Grauslund, M., Görgens, J., van Zyl, W. H., 2005. Role of cultivation media in the development of yeast strains for large scale industrial use. Microbial cell factories. 4, 1-16.

Ignea, C., Cvetkovic, I., Loupassaki, S., Kefalas, P., Johnson, C. B., Kampranis, S. C., Makris, A. M., 2011. Improving yeast strains using recyclable integration cassettes, for the production of plant terpenoids. Microbial cell factories. 10, 4.

28

Jungbluth, M., Renicke, C., Taxis, C., 2010. Targeted protein depletion in Saccharomyces cerevisiae by activation of a bidirectional degron. BMC systems biology. 4, 176.

Karst, F., Lacroute, F., 1977. Ergosterol biosynthesis in Saccharomyces cerevisiae: mutants deficient in the early steps of the pathway. Mol Gen Genet. 154, 269-77.

Kato, H., Sakaki, K., Mihara, K., 2006. Ubiquitin-proteasome-dependent degradation of mammalian ER stearoyl-CoA desaturase. Journal of cell science. 119, 2342-53.

Klopell, F. C., Lemos, M., Sousa, J. P., Comunello, E., Maistro, E. L., Bastos, J. K., de Andrade, S. F., 2007. Nerolidol, an antiulcer constituent from the essential oil of Baccharis dracunculifolia DC (Asteraceae). Zeitschrift fur Naturforschung. C, Journal of biosciences. 62, 537-42.

Krogh, A., Larsson, B., von Heijne, G., Sonnhammer, E. L., 2001. Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. Journal of molecular biology. 305, 567-80.

Kuzuyama, T., 2002. Mevalonate and nonmevalonate pathways for the biosynthesis of isoprene units. Bioscience, biotechnology, and biochemistry. 66, 1619-27.

Leber, R., Landl, K., Zinser, E., Ahorn, H., Spok, A., Kohlwein, S. D., Turnowsky, F., Daum, G., 1998. Dual localization of squalene epoxidase, Erg1p, in yeast reflects a relationship between the endoplasmic reticulum and lipid particles. Molecular biology of the cell. 9, 375-86.

Lemus, L., Goder, V., 2014. Regulation of Endoplasmic Reticulum-Associated Protein Degradation (ERAD) by Ubiquitin. Cells. 3, 824-47.

Li, X., Wu, L., Liu, W., Jin, Y., Chen, Q., Wang, L., Fan, X., Li, Z., Cheng, Y., 2014. A network pharmacology study of Chinese medicine QiShenYiQi to reveal its underlying multi-compound, multi-target, multi-pathway mode of action. PloS one. 9, e95004.

Liu, J., Zhang, W., Du, G., Chen, J., Zhou, J., 2013. Overproduction of geraniol by enhanced precursor supply in Saccharomyces cerevisiae. Journal of biotechnology. 168, 446-51.

Llopis, J., McCaffery, J. M., Miyawaki, A., Farquhar, M. G., Tsien, R. Y., 1998. Measurement of cytosolic, mitochondrial, and Golgi pH in single living cells with green fluorescent proteins. Proceedings of the National Academy of Sciences of the United States of America. 95, 6803-8.

Mantzouridou, F., Tsimidou, M. Z., 2010. Observations on squalene accumulation in Saccharomyces cerevisiae due to the manipulation of HMG2 and ERG6. FEMS yeast research. 10, 699-707.

Mateus, C., Avery, S., 2000. Destabilized green fluorescent protein for monitoring dynamic changes in yeast gene expression with flow cytometry. Yeast (Chichester, England). 16, 1313 - 1323.

Mazumder, A., Pesudo, L. Q., McRee, S., Bathe, M., Samson, L. D., 2013. Genome-wide single-cell-level screen for protein abundance and localization changes in response to DNA damage in S. cerevisiae. Nucleic acids research. 41, 9310-24.

McNeil, C. V., Metabolic engineering approaches to biosynthesize terpenoids in Saccharomyces cerevisiae. Department of Chemistry; Department of Biochemistry and Cell Biology, Vol. PhD. Rice, Houston, Texas, 2008.

McNeil, C. V., Morlacchi, P., Baevich, A., Matsuda, S. P. T., Nerolidol, terpene, and terpene deriviative synthesis. Google Patents, 2012.

Meusser, B., Hirsch, C., Jarosch, E., Sommer, T., 2005. ERAD: the long road to destruction. Nature cell biology. 7, 766-72.

Middleton, B., 1972. The kinetic mechanism of 3-hydroxy-3-methylglutaryl-coenzyme A synthase from baker's yeast. The Biochemical journal. 126, 35-47.

Miziorko, H. M., 2011. Enzymes of the mevalonate pathway of isoprenoid biosynthesis. Archives of biochemistry and biophysics. 505, 131-43.

Mochida, K., Oikawa, Y., Kimura, Y., Kirisako, H., Hirano, H., Ohsumi, Y., Nakatogawa, H., 2015. Receptor-mediated selective autophagy degrades the endoplasmic reticulum and the nucleus. Nature. 522, 359-62.

Muramatsu, M., Ohto, C., Obata, S., Sakuradani, E., Shimizu, S., 2009. Alkaline pH enhances farnesol production by Saccharomyces cerevisiae. Journal of bioscience and bioengineering. 108, 52-5.

29

Newman, J. R., Ghaemmaghami, S., Ihmels, J., Breslow, D. K., Noble, M., DeRisi, J. L., Weissman, J. S., 2006. Single-cell proteomic analysis of S. cerevisiae reveals the architecture of biological noise. Nature. 441, 840-6.

Nissen, A., Kaibel, G., Woerz, O., Arnold, L., Braun, M., Rebafka, W., Preparation of higher unsaturated ketones. Google Patents, 1982.

Otero, J. M., Vongsangnak, W., Asadollahi, M. A., Olivares-Hernandes, R., Maury, J., Farinelli, L., Barlocher, L., Osteras, M., Schalk, M., Clark, A., Nielsen, J., 2010. Whole genome sequencing of Saccharomyces cerevisiae: from genotype to phenotype for improved metabolic engineering applications. BMC genomics. 11, 723.

Paddon, C. J., Keasling, J. D., 2014. Semi-synthetic artemisinin: a model for the use of synthetic biology in pharmaceutical development. Nature reviews. Microbiology. 12, 355-67.

Paddon, C. J., Westfall, P. J., Pitera, D. J., Benjamin, K., Fisher, K., McPhee, D., Leavell, M. D., Tai, A., Main, A., Eng, D., Polichuk, D. R., Teoh, K. H., Reed, D. W., Treynor, T., Lenihan, J., Fleck, M., Bajad, S., Dang, G., Dengrove, D., Diola, D., Dorin, G., Ellens, K. W., Fickes, S., Galazzo, J., Gaucher, S. P., Geistlinger, T., Henry, R., Hepp, M., Horning, T., Iqbal, T., Jiang, H., Kizer, L., Lieu, B., Melis, D., Moss, N., Regentin, R., Secrest, S., Tsuruta, H., Vazquez, R., Westblade, L. F., Xu, L., Yu, M., Zhang, Y., Zhao, L., Lievense, J., Covello, P. S., Keasling, J. D., Reiling, K. K., Renninger, N. S., Newman, J. D., 2013. High-level semi-synthetic production of the potent antimalarial artemisinin. Nature. 496, 528-32.

Peng, B., Williams, T., Henry, M., Nielsen, L., Vickers, C., 2015. Controlling heterologous gene expression in yeast cell factories on different carbon substrates and across the diauxic shift: a comparison of yeast promoter activities. Microbial cell factories. 14, 91.

Polakowski, T., Stahl, U., Lang, C., 1998. Overexpression of a cytosolic hydroxymethylglutaryl-CoA reductase leads to squalene accumulation in yeast. Applied microbiology and biotechnology. 49, 66-71.

Poulter, C. D., Rilling, H. C., 1976. Prenyltransferase: the mechanism of the reaction. Biochemistry. 15, 1079-83.

Raina, K., Crews, C. M., 2010. Chemical inducers of targeted protein degradation. The Journal of biological chemistry. 285, 11057-60.

Ruckenstuhl, C., Lang, S., Poschenel, A., Eidenberger, A., Baral, P. K., Kohut, P., Hapala, I., Gruber, K., Turnowsky, F., 2007. Characterization of squalene epoxidase of Saccharomyces cerevisiae by applying terbinafine-sensitive variants. Antimicrobial agents and chemotherapy. 51, 275-84.

Saito, A., Rilling, H. C., 1979. Prenyltransferase. Product binding and reaction termination. The Journal of biological chemistry. 254, 8511-5.

Salama, S. R., Hendricks, K. B., Thorner, J., 1994. G1 cyclin degradation: the PEST motif of yeast Cln2 is necessary, but not sufficient, for rapid protein turnover. Molecular and cellular biology. 14, 7953-66.

Sandoval, P. C., Slentz, D. H., Pisitkun, T., Saeed, F., Hoffert, J. D., Knepper, M. A., 2013. Proteome-wide measurement of protein half-lives and translation rates in vasopressin-sensitive collecting duct cells. Journal of the American Society of Nephrology : JASN. 24, 1793-805.

Scalcinati, G., Knuf, C., Partow, S., Chen, Y., Maury, J., Schalk, M., Daviet, L., Nielsen, J., Siewers, V., 2012a. Dynamic control of gene expression in Saccharomyces cerevisiae engineered for the production of plant sesquitepene alpha-santalene in a fed-batch mode. Metabolic engineering. 14, 91-103.

Scalcinati, G., Partow, S., Siewers, V., Schalk, M., Daviet, L., Nielsen, J., 2012b. Combined metabolic engineering of precursor and co-factor supply to increase alpha-santalene production by Saccharomyces cerevisiae. Microbial cell factories. 11, 117.

Schuck, S., Gallagher, C. M., Walter, P., 2014. ER-phagy mediates selective degradation of endoplasmic reticulum independently of the core autophagy machinery. Journal of cell science. 127, 4078-88.

Sekar, K., Gentile, A. M., Bostick, J. W., Tyo, K. E., 2016. N-Terminal-Based Targeted, Inducible Protein Degradation in Escherichia coli. PloS one. 11, e0149746.

30

Shiba, Y., Paradise, E. M., Kirby, J., Ro, D. K., Keasling, J. D., 2007. Engineering of the pyruvate dehydrogenase bypass in Saccharomyces cerevisiae for high-level production of isoprenoids. Metabolic engineering. 9, 160-168.

Shimada, H., Kondo, K., Fraser, P. D., Miura, Y., Saito, T., Misawa, N., 1998. Increased carotenoid production by the food yeast Candida utilis through metabolic engineering of the isoprenoid pathway. Applied and environmental microbiology. 64, 2676-80.

Skaff, D. A., Ramyar, K. X., McWhorter, W. J., Barta, M. L., Geisbrecht, B. V., Miziorko, H. M., 2012. Biochemical and structural basis for inhibition of Enterococcus faecalis hydroxymethylglutaryl-CoA synthase, mvaS, by hymeglusin. Biochemistry. 51, 4713-22.

Song, L., 2003. Detection of farnesyl diphosphate accumulation in yeast ERG9 mutants. Analytical biochemistry. 317, 180-5.

Takahashi, S., Yeo, Y. S., Zhao, Y., O'Maille, P. E., Greenhagen, B. T., Noel, J. P., Coates, R. M., Chappell, J., 2007. Functional characterization of premnaspirodiene oxygenase, a cytochrome P450 catalyzing regio- and stereo-specific hydroxylations of diverse sesquiterpene substrates. The Journal of biological chemistry. 282, 31744-54.

Thompson, J. F., Danley, D. E., Mazzalupo, S., Milos, P. M., Lira, M. E., Harwood, H. J., Jr., 1998. Truncation of human squalene synthase yields active, crystallizable protein. Archives of biochemistry and biophysics. 350, 283-90.

Tsuruta, H., Paddon, C. J., Eng, D., Lenihan, J. R., Horning, T., Anthony, L. C., Regentin, R., Keasling, J. D., Renninger, N. S., Newman, J. D., 2009. High-level production of amorpha-4,11-diene, a precursor of the antimalarial agent artemisinin, in Escherichia coli. PloS one. 4, e4489.

Vannice, J., Skaff, D. A., Wyckoff, G. J., Miziorko, H. M., 2013. Expression in Haloferax volcanii of 3-Hydroxy-3-methylglutaryl-CoA synthase Facilitates Isolation and Characterization of the Active Form of a Key Enzyme Required for Polyisoprenoid Cell Membrane Biosynthesis in Halophilic Archaea. Journal of bacteriology.

Varshavsky, A., 1996. The N-end rule: functions, mysteries, uses. Proceedings of the National Academy of Sciences of the United States of America. 93, 12142-9.

Vickers, C. E., Behrendorff, J. B., Bongers, M., Brennan, T. C., Bruschi, M., Nielsen, L. K., 2015a. Production of Industrially Relevant Isoprenoid Compounds in Engineered Microbes. Microorganisms in Biorefineries. Springer, pp. 303-334.

Vickers, C. E., Bongers, M., Bydder, S. F., Chrysanthopoulos, P., Hodson, M. P., 2015b. Protocols for the production and analysis of isoprenoids in bacteria and yeast. Springer Protocols Handbooks. Springer, pp. 1-30.

Vickers, C. E., Bydder, S. F., Zhou, Y., Nielsen, L. K., 2013. Dual gene expression cassette vectors with antibiotic selection markers for engineering in Saccharomyces cerevisiae. Microbial cell factories. 12, 96.

Wilding, E. I., Brown, J. R., Bryant, A. P., Chalker, A. F., Holmes, D. J., Ingraham, K. A., Iordanescu, S., So, C. Y., Rosenberg, M., Gwynn, M. N., 2000. Identification, evolution, and essentiality of the mevalonate pathway for isopentenyl diphosphate biosynthesis in gram-positive cocci. Journal of bacteriology. 182, 4319-27.

Williams, T. C., Averesch, N. J., Winter, G., Plan, M. R., Vickers, C. E., Nielsen, L. K., Kromer, J. O., 2015a. Quorum-sensing linked RNA interference for dynamic metabolic pathway control in Saccharomyces cerevisiae. Metabolic engineering. 29, 124-34.

Williams, T. C., Espinosa, M. I., Nielsen, L. K., Vickers, C. E., 2015b. Dynamic regulation of gene expression using sucrose responsive promoters and RNA interference in Saccharomyces cerevisiae. Microbial cell factories. 14, 43.

Wolinski, H., Kolb, D., Hermann, S., Koning, R. I., Kohlwein, S. D., 2011. A role for seipin in lipid droplet dynamics and inheritance in yeast. Journal of cell science. 124, 3894-904.

Xie, W., Ye, L., Lv, X., Xu, H., Yu, H., 2014. Sequential control of biosynthetic pathways for balanced utilization of metabolic intermediates in Saccharomyces cerevisiae. Metabolic engineering. 28C, 8-18.

31

Yamanishi, M., Ito, Y., Kintaka, R., Imamura, C., Katahira, S., Ikeuchi, A., Moriya, H., Matsuyama, T., 2013. A genome-wide activity assessment of terminator regions in Saccharomyces cerevisiae provides a ''terminatome'' toolbox. ACS synthetic biology. 2, 337-47.

Yoon, S. H., Lee, S. H., Das, A., Ryu, H. K., Jang, H. J., Kim, J. Y., Oh, D. K., Keasling, J. D., Kim, S. W., 2009. Combinatorial expression of bacterial whole mevalonate pathway for the production of beta-carotene in E. coli. Journal of biotechnology. 140, 218-226.

Yuan, J., Ching, C. B., 2015. Dynamic control of ERG9 expression for improved amorpha-4,11-diene production in Saccharomyces cerevisiae. Microbial cell factories. 14, 38.

Zhuang, X., Engineering novel terpene production platforms in the yeasts Saccharomyces cerevisiae. Plant Physiology, Vol. PhD. University of Kentucky, 2013.

Highlights:

Squalene synthase was modulated by a protein degradation mechanism.

Squalene synthase C-terminal peptide is an ER-targeting transmembrane signal peptide.

A mechanism was engineered to degrade ER membrane-integrating protein.

Enterococcus mvaS and mvaE were used to improve sesquiterpene production in yeast.

Destabilizing squalene synthase redirects carbon flux to nerolidol production.