A microtiter plate-based assay method to determine fat quality

Mark Bertrama

Corrine Manschot-Lawrenceb

Eckhard Flöterb

Uwe T. Bornscheuera

a Greifswald University,Institute of Biochemistry,Department of Biotechnology& Enzyme Catalysis,Greifswald, Germany

b Unilever Research,Vlaardingen, The Netherlands

A microtiter plate-based assay method todetermine fat quality

A microtiter plate (MTP) method was developed to determine the quality of fats that areused in large-scale processing using lipase catalysis. Two assay formats were fol-lowed: In the first approach, the fats were interesterified with p-nitrophenol laurateusing a lipase from Thermomyces lanuginosa; in the second approach, pH indicatorswere added to the fat samples containing lipase. A blind study using 29 fats showedthat the MTP method using p-nitrophenol as pH indicator allowed a rapid and reliableassignment of bad fats and an acceptable differentiation between fats of moderate andgood quality.

Keywords: Assay, fat quality, immobilized lipase, lipase stability, microtiter plate.

1 Introduction

Fats and oils are highly important nutrients for humansand therefore play a major role in agriculture and in theindustry processing them. The vast majority of the ap-proximately 128 million metric tons of fats and oils pro-duced worldwide per year is indeed used in the foodsector and an estimate of 10% is used in oleochemistry.Especially vegetable oils and fats are frequently used asstarting material for food products. In order to adapt theproperties of fats and oils for best human nutrition (i.e. toproduce margarine), physical blending, chemical meth-ods (hydrogenation, interesterification) and enzymaticmethods are used. The latter method – in the majority ofapplications relying on the use of lipases – has gainedconsiderable importance in the past few years [1], forseveral reasons: Lipases exhibit unique specificitiesunknown to most chemical catalysts (i.e. specific for fattyacid chain length, degree of saturation, 1,3-regiospecifi-city), the enzymes became substantially cheaper and arenow available in bulk quantities in immobilized form, andseveral lipases have been tailor-designed for lipid mod-ification. The recent change in legislation to control thelevel of trans fatty acids in triglyceride-derived productsfor human consumption further boosted the demand formild reaction methods. Indeed, a broad range of productsis currently made using lipase catalysis on industrialscale. This includes for instance the interesterification oftriglycerides to produce margarine, the synthesis ofdiglyceride-rich cooking and frying oil, and the synthesisof structured triglycerides [2].

The fat and oils from the oil palm plant (Elaeis guineensis)are widely used for the production of margarines. Whilethe palm kernel oil mainly contains lauric acid, the majorfatty acids of palm oil are palmitic acid. The lipase-cata-lyzed interesterification of both types of fats leads to amargarine precursor. The quality and composition of aparticular fat intended for use in a given product isimportant in the development and manufacture of a high-quality product. Consequently, it is very important toknow whether a lipase is stable and active during thelarge-scale processing of natural oils and fats. Currently,the process stability of the biocatalyst is determined in asmall-scale interesterification process over several days.However, it would be much more desirable to have a rapidand at the same time reliable method at hand to quicklydecide whether a certain fat batch has a negative influ-ence on lipase stability. The aim of this project wastherefore the development of a microtiter plate (MTP)-based assay method to quickly assess the lipase stabilityin the presence of fats.

2 Materials and methods

2.1 Materials

The lipase from Thermomyces lanuginosa (TLL, liquidformulation) was from Novozymes (Bagsvaerd, Denmark).All fat samples were provided and classified with respectto their quality (good, moderate, bad) by Unilever (Vlaar-dingen, The Netherlands). All other chemicals and sol-vents were from common commercial suppliers and ofthe highest purity available. All experiments were per-formed at least in duplicate. MTP (PS-Microplate, black,clear bottom) were from Greiner Bio-One GmbH (Fri-ckenhausen, Germany).

Correspondence: Uwe T. Bornscheuer, Greifswald University,Institute of Biochemistry, Department of Biotechnology & En-zyme Catalysis, Felix-Hausdorff-Str. 4, D-17487 Greifswald,Germany. Phone: 149–3834–864367, Fax: 149 3834 8680066,e-mail: [email protected]

180 DOI 10.1002/ejlt.200600193 Eur. J. Lipid Sci. Technol. 109 (2007) 180–185

© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com

Sp

ecia

lTo

pic

Eur. J. Lipid Sci. Technol. 109 (2007) 180–185 Lipase assay for fat quality 181

2.2 MTP preparation and reaction conditions

All investigations of lipase inactivation in oil were per-formed in MTP scale. To avoid solidification of the fatsduring the measurement in the fluorimeter (max. analysistemperature 45 7C), the small spaces between the wellsof the MTP were filled with glycerol serving as insulator.This modification allowed the removal of the plate fromthe thermoshaker for up to 10 min without any formationof solids. The fats were melted in a water bath at 75 7Cand then distributed into the wells of the MTP. A typicalreaction solution contained 200 mL fat and 50 mL enzymesolution (sodium phosphate buffer, 50 mM, pH 7.5). TheseMTP were incubated on a thermoshaker at 70 7C and350 rpm for a given time. To determine the lipase activityand stability, the MTP was transferred to the fluorimeter(FLUOstar Optima; BMG Labtech GmbH, Offenburg,Germany) and the measurement was performed beforereturning the MTP again to the 70 7C incubation. To initiatethe measurement in the fluorimeter, 50 mL substrate [i.e.p-nitrophenyl laurate (pNPL) in DMSO/buffer – see belowfor details – or a pH indicator] was added automatically,before the first measurement using the built-in fluorimeterpump was started.

2.3 Protocols for lipase stability test in MTPscale

2.3.1 p-Nitrophenol ester assay

For the direct measurement of lipase activity, pNPL or p-nitrophenol palmitate (pNPP) were used. Stock solutions(30 mM in DMSO) were diluted to appropriate final con-centrations (1–3 mM) using sodium phosphate buffer(50 mM, pH 7.5). The released chromophor p-nitrophenol(pNP) was quantified at 410 nm.

2.3.2 pH indicator assay

Alternativately, pNP or bromothymolblue (BTB) served aspH indicators. 1 mM pNP in sodium phosphate buffer(5 mM, pH 7.2) was added automatically to the fat(200 mL) and enzyme solution (50 mL). Incubation andmeasurement were performed as described above for thepNP assay. In case of BTB, a pH decrease leads to a colorshift (blue to yellow) and a decrease in absorbance at620 nm. Of fat, 150 mL was mixed with 100 mL BTB (15%,vol/vol) dissolved in sodium phosphate buffer (10 mM,pH 7.5) in the wells of an MTP. Finally, 50 mL enzyme buf-fer solution was added. Incubation and measurementwere performed as described above for the pNP assay.

2.3.3 pH measurement

Of fat, 100 mL was dissolved in 1 mL ethanol and filled upwith distilled water to 10 mL. The pH value of this solutionwas determined in triplicate using a pH meter (Hannainstruments, Kehl am Rhein, Germany). The pH value of asolution of ethanol with distilled water served as blank.

2.3.4 Thin-layer chromatography

Samples from pure and enzyme-modified fats were ana-lyzed. Reaction samples from the lipase incubationexperiments were withdrawn and extracted according toFolch et al. [3]. The samples were mixed with 800 mLchloroform/methanol (2 : 1, vol/vol) and 400 mL distilledwater. After centrifugation for 3 min at 13,000 rpm, theorganic layer was removed, spotted on a thin-layer chro-matography (TLC) plate and developed with n-hexane/diethylether/acetic acid (70 : 30 : 1, vol/vol). For visuali-zation, the TLC plates were sprayed with Cerium-IV-reagent (24.7 mM Cerium-IV-sulfate, 13.6 mM molybda-tophosphoric acid, 3% vol/vol concentrated sulfuric acid)and heated to visualize the monoacylglycerides, 1,2- and1,3-diacylglycerides, free fatty acids (FFA) and triacyl-glycerides.

2.3.5 Gas chromatographic analysis

The compositions of the different fats were qualitativelyanalyzed using high-temperature gas chromatography(Hewlett-Packard 5890 Series II Plus; Hewlett-Packard,Böblingen, Germany) equipped with an Optima-17-TGcolumn (25 m 6 0.32 mm; Macherey-Nagel, Düren, Ger-many). Analysis was carried out with a temperature pro-gram from 150 to 340 7C at 20 7C/min and 340 to 360 7Cat 3 7C/min. Injector and detector temperatures werechosen as 350 and 370 7C, respectively. This temperatureprogram allowed separation of FFA, partial glycerides andtriacylglycerides.

3 Results and discussion

3.1 Establishment of suitable measurementconditions

The aim of this project was the establishment of a meth-odology to assay fat quality in MTP by spectro-photometric or fluorimetric measurement. Unfortunately,the currently available devices for this measurement allowa maximum working temperature of 45 7C. In contrast, thefat mixture (palm oil and palm kernel oil) to be investigatedhad a melting point of 65 7C. Initial experiments showed


182 M. Bertram et al. Eur. J. Lipid Sci. Technol. 109 (2007) 180–185

that the fats solidified rapidly when not thermostatedabove 65 7C, making a measurement impossible. Thismajor problem was solved by filling the small space be-tween the wells of the MTP with glycerol serving as insu-lator. This modification allowed the removal of the platefrom the thermoshaker for up to 10 min without any soli-dification, even for fats with a melting point of 65 7C.

3.2 Measurement of lipase activity

All fat samples used in this study were already classifiedby Unilever (Vlaardingen, The Netherlands) with respectto their quality as good, moderate, and bad fats based ontheir influence on lipase stability determined in large-scaletests. Still, the use of GC-MS or HPLC analysis did notprovide any indication of the components present in thesesamples responsible for the observed differences inlipase stability (data not shown).

First, the lipase activity in the presence of the fats ofdifferent quality was directly measured by addition ofpNPL or pNPP. Hydrolysis of these esters releases thechromophore pNP, which can be easily quantified at410 nm in the fluorimeter. Fat samples were incubatedin an MTP on a thermoshaker at 70 7C together with thelipase (TLL) and the pNP ester for a certain time inter-val; then, the amount of pNP released was measured inthe fluorimeter (at 45 7C) and the MTP was again incu-bated at 70 7C. Fig. 1 shows the change in absorbanceover time for pNPL for six fat samples. Is it obvious thata clear distinction between the bad fats (nos. 69, 199)from the moderate (nos. 134, 181) and good fats

(nos. 61, 84) is possible using this method. However,the use of pNPL is preferable, as significantly largerabsorption values and stronger differences between thebad and the other fats were observed compared to theuse of pNPP (data not shown). This observation is pos-sibly related to the fatty acid chain length specificity ofTLL [4].

3.3 pH indicator assay

Fig. 1 shows that even before the addition of pNPL to thefats, a significant difference in initial absorbance can bemeasured between the bad and the moderate/good fats(Fig. 1; 0 min incubation time). Thus, we decided to di-rectly use pH indicators to estimate the fat quality by theamount of fatty acid released from the fat by lipase cata-lysis. High enzyme activity leads to the formation of largeramounts of FFA and therefore a stronger decrease in thepH value. pNP was already used to estimate the enan-tioselectivity of lipases [5], and it is also useful because itspKa value of 7.14 is close to the pH optimum of mostlipases. The results of this experiment are shown in Fig. 2.Autohydrolysis (A) was rather low and resulted in largerabsorbance values, while the lipase activity led to astronger decrease in absorbance. This can also be fol-lowed over the entire incubation period. The insert inFig. 2 shows the differences between the values meas-ured in the presence of TLL minus the autohydrolysis ofeach sample. Again, a differentiation between bad (whitecolumns) and good or moderate fats (grey and black col-umns) is possible.

Fig. 1. Time course of the formation of pNP during incubation of lipase from Thermomyces lanugi-nosa (TLL) in the presence of fats of different quality using pNPL as substrate. The columns indicatebad (white; nos. 69, 199), moderate (grey; nos. 134, 181) and good (black; nos. 61, 84) fats. Theerror bars indicate the mean values of ten experiments.



Fig. 2. Time course of the change in absorbance using pNP as pH indicator. A, Autohydrolysis; TLL, lipase-catalyzed hydrolysis. The insert shows the difference between the lipase-catalyzed reaction and the auto-hydrolysis (A – TLL) at 1140 min. The shading and numbering of columns is identical to Fig. 1. The error barsindicate the mean values of six experiments.

Fig. 3. Time course of the change in absorbance using BTB as pH indicator. A, Autohydrolysis; TLL, lipase-catalyzed hydrolysis. The insert shows the difference between the lipase-catalyzed reaction and the auto-hydrolysis (A – TLL). The shading and numbering of columns is identical to Fig. 1. The error bars indicate themean values of six experiments.


184 M. Bertram et al. Eur. J. Lipid Sci. Technol. 109 (2007) 180–185

In a similar manner, BTB was used as pH indicator. Varia-tion of the pH between 6.0 and 7.4 leads to a color shiftfrom yellow to blue. Additionally, the absorbance of thealkaline form of BTB can be measured at 620 nm [6, 7].Again, a similar profile as found for pNP was observed(Fig. 3); however, the data deviation is less pronounced ascan be seen from the blank experiments. Interestingly, themeasurable differences between the bad fats on the onehand and the good/moderate fats on the other hand canalready be highly reproducibly observed before the addi-tion of lipase. This would be in accordance with theknown observation that large amounts of FFA can inacti-vate lipases [8].

As the pH of the fats appeared to be a characteristic cri-terion to judge the quality of the fats and hence the lipasestability, the pH value was measured. For this, 100 mL fatwas solved in ethanol and then distilled water was added,before the pH of the aqueous phase was determined witha conventional pH meter. However, no clear correlationbetween the pH value and the fat quality was found (datanot shown).

3.4 Two-phase phenomenon

During the analysis of the fat samples in the MTP, it wasobserved that under certain conditions no formation of afull two-phase system took place, but the fat layer wasmore droplet-like on top of the aqueous phase, as illus-trated in Fig. 4. Upon incubation for 10–20 min, this drop-let ‘spread’ over the entire radius of the MTP well(Fig. 4B), but this phenomenon was only observed for theknown bad fats. This phenomenon might explain the dif-ferences found in the pNP ester assay of the bad fats

compared to the better-quality fats. Further analysis ofthis observation revealed that this effect only occured ifthe volume of the aqueous phase was 50 mL and whenphosphate or citrate buffer (at 50 mM concentration andpH �7.5) was used. An explanation for this phenomenonis not easy, but we assume that the presence of emulsi-fying substances in good and moderate fats facilitates thespreading of the fat phase, whereas in bad fats this isprevented. Muralidhar et al. described a similar phenom-enon based on the influence of monoacylglycerides andphospholipids [9]. However, no signficant differences inthe concentration of these components were found dur-ing fat analysis by TLC or GC (data not shown).

Fig. 4. Illustration of the two-phase phenomenon: Thetop picture shows the top view of two wells of an MTP.Below this, the front view of the same MTP wells is shownschematically. (A) refers to the droplet formationobserved for moderate and good fats, (B) indicates thetwo-phase system found for bad fats.

Fig. 5. Results of the blind study of 29 fats ofinitially unknown quality. The fat samples wereincubated with pNP as pH indicator for 1 h andthe change in absorbance was monitored. Aclear distinction of bad fats (white columns)from samples of higher quality (moderate, grey;good, black columns) is visible.



3.5 Application of the assay in a blind study

In order to verify the predictive manner of the MTP meth-od, 29 fat samples of disclosed quality (known to Unileverbut not to the Greifswald research group) where sub-jected to the assay method in a blind study. For this,200 mL of each fat was mixed with 50 mL pNP (1 mM) dis-solved in sodium phosphate buffer (pH 7.5, 50 mM) inwells of an MTP. The plate was incubated for 1 h and theabsorption was measured at 410 nm. Fig. 5 shows thatsignificant differences between the 29 fat samples can bedetermined and that the bad fats can be clearly identified.In most cases, the higher-quality (moderate and good)fats could also be correctly assigned.

In summary, this method allows a highly reproducibleidentification of fats of poor quality within a short analysistime using an MTP format. It therefore enables a rapiddecision whether a certain fat batch can be used safely inlipase-catalyzed processing of fats.

Acknowledgments

Financial support by the Eureka project TSIN2006 “Natu-ral Spread” is gratefully acknowledged. We are especiallygrateful to Dr. Hilda ten Brink and Dr. Rob Diks from Uni-lever, Vlaardingen, The Netherlands, for useful discus-sions and the provision of fat samples.

References

[1] A. Balksten: Enzyme technology in the oils and fats industry.Lipid Technol. 2006, 18, 154–157.

[2] U. T. Bornscheuer, M. Adamczak, M. M. Soumanou: Lipase-catalyzed synthesis of modified lipids. In: Lipids as Constitu-ents of Functional Foods. Ed. F. D. Gunstone, P. J. Barnes,Bridgwater (UK) 2002, pp. 149–182.

[3] J. Folch, M. Lees, G. H. S. Stanley: A simple method for theisolation and purification of total lipids from animal tissues. JBiol Chem. 1957, 226, 497–509.

[4] M. Berger, M. P. Schneider: Lipases in organic solvents: Thefatty acid chain length profile. Biotechnol Lett. 1991, 13, 641–645.

[5] L. E. Janes, A. C. Löwendahl, R. J. Kazlauskas: Quantitativescreening of hydrolase libraries using pH indicators: Identify-ing active and enantioselective hydrolases. Chem Eur J.1998, 4, 2324–2331.

[6] G. T. John, E. Heinzle: Quantitative screening method forhydrolases in microplates using pH indicators: Determinationof kinetic parameters by dynamic pH monitoring. BiotechBioeng. 2001, 78, 620–627.

[7] F. X. Malcata, H. R. Reyes, H. S. Garcia, C. G. J. Hill, C. H.Amundson: Kinetics and mechanisms of reactions catalyzedby immobilized lipases. Enzyme Microb Technol. 1992, 14,426–446.

[8] F. Moris-Varas, A. Shah, J. Aikens, N. P. Nadkarni, J. D. Roz-zell, D. C. Demirjian: Visualization of enzyme-catalyzed reac-tions using pH indicators: Rapid screening of hydrolaselibraries and estimation of the enantioselectivity. Bioorg MedChem. 1999, 7, 2183–2188.

[9] R. V. Muralidhar, R. R. Chirumamilla, R. Marchant, V. N.Ramachandran, O. P. Ward, P. Nigam: Understanding lipasestereoselectivity. World J Microbiol Biotechnol. 2002, 18, 81–97.

[Received: September 3, 2006; accepted: November 28, 2006]


Documents

A microtiter plate-based assay method to determine fat quality