Screening and selection of exopolysaccharide ... - Wiley Online Library

Journal of Applied Microbiology 2003, 95, 1200–1206

doi:10.1046/j.1365-2672.2003.02122.x

Screening and selection of exopolysaccharide-producing strains of Lactobacillus delbrueckii subsp. bulgaricus A.D. Welman1,2,*, I.S. Maddox1 and R.H. Archer2 1

Institute of Technology and Engineering, Massey University, Palmerston North, New Zealand and 2Fronterra Cooperative Group Ltd, Palmerston North, New Zealand

2003/0309: received 13 April 2003, revised 16 June 2003 and accepted 17 June 2003

ABSTRACT A . D . W E L M A N , I . S . M A D D O X A N D R . H . A R C H E R . 2003.

Aims: The selection of exopolysaccharide (EPS)-producing strains of Lactobacillus delbrueckii subsp. bulgaricus. Methods and Results: Improved EPS-overproducing strains of L. delbrueckii subsp. bulgaricus were derived by chemical mutagenesis and selection. Initial screening of the chemically induced mutant pool relied primarily on the selection of strains with raised levels of lactic acid and reduced biomass formation. Supporting selection criteria used were ropiness and colonial mucoidy. Final screening of candidate strains undertaken in a semi-defined medium in batch culture, resulted in the selection of a mutant with a 35% improvement in specific EPS yield relative to the parent strain. Conclusions: Initial selection of mutants of L. delbrueckii subsp. bulgaricus on the basis of enhanced formation of lactate and reduced biomass formation, coupled with a ropy or mucoid phenotype, proved to be a satisfactory means of isolating strains with the potential for a higher level of specific EPS production than the parent strain. Significance and Impact of the Study: The assay protocol allowed for the selection of an EPS-overproducing strain of L. delbrueckii subsp. bulgaricus. Such strains are useful for the purposes of metabolic studies related to EPS-production. Keywords: exopolysaccharides, lactate, Lactobacillus delbrueckii subsp. bulgaricus, metabolism, screening.

INTRODUCTION Exopolysaccharides (EPSs) produced by lactic acid bacteria (LAB) possess the possibility of replacing stabilizers and thickeners currently produced commercially by nonfoodgrade bacteria. Over the past number of years, various studies have concentrated upon understanding the biochemistry and genetics of EPS production in LAB, so that rational strategies can be developed for the improvement of EPS yield and the design of tailor-made polysaccharides. Metabolic engineering strategies that target increasing metabolic flux to EPS should include EPS formation * Present address: A.D. Welman, Fonterra Cooperative Group Ltd, Palmerston North, New Zealand. Correspondence to: A.D. Welman, Fonterra Research Centre, Fonterra Cooperative Group Ltd, Private Bag 11029, Palmerston North, New Zealand (e-mail: [email protected]).

pathways (de Vos 1996). It has been suggested that a potential controlling factor in EPS biosynthesis is the availability of sugar nucleotides which are necessary for the construction of the polymers (Boels et al. 2001). In the present investigation, a stable, chemically induced mutant of Lactobacillus delbrueckii subsp. bulgaricus NCFB 2483 was isolated which had an enhanced specific yield of EPS, relative to the parent strain. This mutant was generated for future studies in order to compare the anabolic processes of sugar-nucleotide synthesis between the two strains. Primary screening was based upon the rationale, that any enhancements in the EPS-synthesizing pathway would be needed to be supported by a strain which produced sufficient ATP for the synthesis of sugar nucleotides, and the ability to adequately control the redox balance of the cell. The strains obtained for further screening on the basis of enhancement of EPS yield, would exhibit a global enhancement in glycolysis, and demonstrate a mucoidy and/or ropiness. ª 2003 The Society for Applied Microbiology

SCREENING FOR EXOPOLYSACCHARIDES

Chemical mutagenesis of LAB has been undertaken in a number of instances using N-methyl-N¢-nitro-N-nitrosoguanidine (NTG). NTG has been used to induce an increase in the formation of diacetyl and acetaldehyde in Streptococcus diacetylactis (Kuila et al. 1971). Similarly, an increase in acetoin and diacetyl has been reported in Strep. lactis subsp. diacetylactis (Piatkiewicz et al. 1981). Genetic variants of Lactobacillus casei subsp. alactosus which produced increased quantities of flavour compounds and lactic acid in soya milk, have been obtained (Miyamoto et al. 1983). Mutants were generated which were defective in the structural gene of malolactic enzyme, catalysing the conversion of L-malate to L-lactate and carbon dioxide in Streptococcus lactis (Renault and Heslot 1987), and stable galactose-fermenting strains of Streptococcus thermophilus were obtained by Benateya et al. (1990). b-Galactosidase-overproducing strains of Lact. delbrueckii subsp. bulgaricus, Strep. thermophilus, Bifidobacterium breve and Bif. longum have also been generated using NTG (Ibrahim and O’Sullivan 2000). Few reported cases exist, however, of attempts to induce EPS overproduction in LAB using chemical mutagenesis. This paper describes a rational strategy to select for a raised specific EPS yield in a chemically mutagenized population of Lact. delbrueckii subsp. bulgaricus.

MATERIALS AND METHODS Media The medium used was according to Kimmel and Roberts (1998) which was modified by replacement of the glucose component with lactose and consisted of lactose, 20 g l)1, yeast nitrogen base (without amino acids and ammonia), 5 g l)1; bacto casitone, 20 g l)1; sorbitan monoleate (Tween 80), 1 g l)1; dipotassium phosphate (K2HPO4), 2 g l)1; magnesium sulphate (MgSO4Æ7H2O), 0Æ1 g l)1; manganese sulphate (MnSO4Æ4H2O), 0Æ05 g l)1; ammonium citrate, 2 g l)1; sodium acetate, 5 g l)1. In all instances, the medium was prepared in separate, double-strength aliquots of lactose and the remainder of the nutrients as described above. After steam-sterilization, the aliquots were allowed to cool, and pooled, achieving the desired concentration of medium constituents. Where solid medium was prepared, 15 g l)1 of agar was used. Bacterial strains The strain of Lact. delbrueckii subsp. bulgaricus NCFB 2483 (NCIMB 702483) was obtained from the National Collection of Industrial and Marine Bacteria (Aberdeen, Scotland) and preserved as a cell bank in 1Æ0-ml aliquots at )80C. Mutants selected for batch culture testing were preserved in the same fashion.

1201

Chemical mutation The method used was a modification of that described by Monnet and Corrieu (1998). A mid-log phase culture of Lact. delbrueckii subsp. bulgaricus NCFB 2483, washed and suspended in 0Æ1 M McIlvaine’s buffer (sodium citrate) (pH 6Æ2), was treated with NTG and incubated at 37C for 60 min. A two to three log reduction in cell number was desired. The mutated cells were subsequently washed three times in 0Æ1 M McIlvaine’s Buffer, and resuspended in the modified medium of Kimmel and Roberts (see above) for a further 3 h period of incubation at 37C to resuscitate the cells prior to storage in 4-ml aliquots at )80C. The individual mutants were isolated for colony picking by serial dilution. Culture conditions For the culture of the parent (NCFB 2483) and mutant strains in 96-well format microtitre plates, individual colonies were inoculated into wells containing 250 ll medium, and the plates incubated under slow agitation (100 rev min)1) for a period of 24 h under anaerobic conditions. This was achieved by incubating the microtitre plates in sealed plastic bags, inflated with nitrogen gas. Batch-culture testing was undertaken in 250 ml volumes in shake bottles (Schott Duran, Mainz, Germany), inoculated from the working cell banks, as described above. The cultures were incubated in an orbital incubator shaker (Gallenkamp INR-200, Gallenkamp, Leicester, UK) at 100 rev min)1 which was sufficient to keep the culture well-mixed, at a temperature of 37C for a period of 24 h. Sample aliquots (15 ml) were withdrawn aseptically at 4 h intervals for the determination of dry cell weight and sugar conversion to extracellular metabolites. Incubation of cultures on solid medium was undertaken at 37C under anaerobic conditions for a period of 48 h. In instances where analyses could not be immediately undertaken, samples were stored in presterilized bottles at )20C. Screening techniques and analyses Growth was monitored by absorbance measurement at 650 nm (Pharmacia Biotech Ultrospec spectrophotometer, Amersham Pharmacia, Uppsala, Sweden). Biomass was determined from a standard curve relating absorbance at 650 nm to washed dry cell weights. Plates for colony picking were obtained by serial dilution. Colonies were picked from plates with a spread of 90–100 colonies. Replica-plating was achieved using a sterilizable 96-pin replicator, custom manufactured to transfer droplets of cell-suspension medium from the microtitre plates to

ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 1200–1206, doi:10.1046/j.1365-2672.2003.02122.x

1202 A . D . W E L M A N ET AL.

15-cm-diameter agar plates and to additional microtitre plates for the testing of the presence of acetoin. Acidifiability tests were undertaken according to a modification of the method described by Venus et al. (1991), using an OPTImax tunable Microplate reader (SOFTmax Pro software, Molecular Devices, Sunnyvale, CA, USA). After measurement of the absorbance at 650 nm after 24 h growth in the microtitre plate wells, 5 ll of a mixture of 0Æ5% aqueous bromocresol purple and 0Æ5% aqueous bromocresol green solutions was added to each well. After agitation, the absorbance was measured at 600 nm. The acidification was expressed as a ratio between the absorbance at 600 nm relative to the absorbance at 650 nm. The presence of acetoin, was tested using the Voges– Proskauer reaction according to the method of O’Meara (1931) and as described by Eddy (1961). O’Meara’s reagent (5 ll) was added to 250 ll cultures grown for 24 h in microtitre plates. Mucoidy of colonies was determined by visual appearance, and the ropiness of colonies and liquid broth were determined by testing with a sterile toothpick. For the measurement of extracellular polysaccharides, aliquots (100 ll) of fermentation broth (including bacterial cells) were subjected to a twofold precipitation with chilled ethanol (2Æ9 ml distilled H2O and 7 ml 99Æ7% ethanol) for 24-h periods at 4C. The precipitate was obtained by centrifugation (35 850 g, 40 min, 4C) and resuspended in 1Æ0 ml distilled water. The total sugar concentration in the resuspended sample was measured according to the method of Dubois et al. (1956), using dextran as the standard. Lactose depletion, and galactose and lactate formation in shake-bottle cultures was measured by HPLC (Waters Alliance 2690, coupled to a Waters 2410 Differential Refractometer, and Waters 2487 UV spectrophotometer, Waters, Milford, MA, USA). The compounds were separated using an Aminex HPX-87H, 300 · 7Æ8 mm column (Biorad, Richmond, CA, USA), according to the method as described in Ross and Chapital (1987). The column temperature was maintained at 40C, and the flow rate at 0Æ6 ml min)1. Broth samples were clarified by centrifugation at 16 000 g for 15 min prior to dilution in distilled and filtered water (MilliQ, Millipore, Billerica, MA, USA), prior to analysis. External standards were prepared for lactose, galactose and lactic acid. All analyses were undertaken in duplicate.

RESULTS Chemical mutagenesis A mutant pool generated by treatment of a stock of Lact. delbrueckii subsp. bulgaricus NCFB 2483 with NTG was used for assessment of traits associated with raised EPS formation. A 10)7 dilution derived from a serial dilution of

the mutant mixture yielded ca 90 colonies per agar plate, after incubation at 37C for 24 h under anaerobic conditions. These colonies were toothpicked into wells of medium in microtitre plates, and cross-inoculated onto replica agar plates. Screening tests for EPS, mucoidy, ropiness and acetoin production The discriminatory value of the methods to test for mucoidy, and ropiness of the bacterial colonies or broth cultures was limited, as any differences in EPS production (or alternatively structure) were relatively small. The methods served to confirm these characteristics in strains selected on the basis of biomass and acid production. Neither the NCFB 2483 parent strain nor mutant strains demonstrated acetoin production when tested with O’Meara’s reagent. Analysis of biomass and acidification in microtitre plates For comparative purposes, acidification and absorbance analyses were undertaken on the Lact. delbrueckii subsp. bulgaricus NCFB 2483 parent strain. A 250 ml culture of Lact. delbrueckii subsp. bulgaricus NCFB 2483 was prepared from the stock culture in modified Kimmel and Roberts medium, as described in the Materials and methods section. Colonies (64) cultured on solid agar medium from a dilution of the culture harvested at 16 h were toothpicked into 250 ll volumes of medium in the microtitre wells, and incubated under anaerobic conditions and gentle agitation at 37C for 24 h. A distribution of the acidifying ability and absorbance values of the NCFB 2483 strain is depicted in Fig. 1. Whilst a small number of colonies of L. delbrueckii subsp. bulgaricus NCFB 2483 had a low acidifying ability over the 24 h time period (i.e. high A600/A650 values) (Fig. 1), the mean value and standard deviation derived from the 64 colonies was 3Æ2 ± 1Æ2 at a 95% level of confidence. In the same manner, 1200 individual mutants were screened within a relatively short time-interval. Strains were sought which formed less biomass than the parent strain, but which produced more acid per cell mass. As the A600 value diminishes with increasing acid production, mutants with an A600/A650 ratio which was significantly lower than that of the parent strain (3Æ2 ± 1Æ2) were selected for further screening. Typical results of a spread of A650 values and A600/A650 ratios for mutants from individual colonies are depicted in Fig. 2. From the 1200 mutants screened, nine were selected on the basis of their lower biomass and higher acidifying ability relative to the parent strain, mucoid appearance of colonies and ropiness in broth. Upon reculturing of these strains, seven retained their viability.



10·0

9·0

9·0

8·0

8·0

7·0

7·0

6·0

6·0

5·0

5·0

4·0

4·0

3·0

3·0

2·0

2·0

1·0

1·0

0·0

A600 /A650

10·0

A650

Fig. 1 Distribution of absorbance readings related to cell concentration and acidifying ability of Lactobacillus delbrueckii subsp. bulgaricus NCFB 2483 (A600/A650) cultured for 24 h in microtitre plate wells. The absorbance corresponding to cell concentration (j) was measured at 650 nm. Acidifying ability (n) was quantified by the ratio of A600 of the broth with bromocresol purple and bromocresol green added, to the A650 measurement

1203

0·0 0

10

20

30 40 Well number

50

60

6·0

6·0

5·5 5·0

5·0

4·5 4·0

4·0

3·0

3·0

2·5 2·0

2·0

A600 /A650

3·5 A650

Fig. 2 Distribution of absorbance readings related to cell concentration and acidifying ability of Lactobacillus delbrueckii subsp. bulgaricus mutants cultured for 24 h in microtitre plate wells. The absorbance corresponding to cell concentration (j) was measured at 650 nm. Acidifying ability (n) was quantified by the ratio of A600 of the broth with bromocresol purple and bromocresol green added, to the A650 measurement. The mean ± S.D. of the A600/A650 ratio for the parent strain is shown for comparison

1·5 1·0

1·0

0·5 0·0 0

Batch fermentation testing Six candidate strains were selected from the initial screening of 1200 mutants for further evaluation. The strains were cultured in 250 ml medium for a period of 24 h, and analysed for EPS and lactate production. Biomass levels were all significantly lower, and specific lactate yields of all the mutant strains significantly higher than that of the parent (NCFB 2483) (Fig. 3). Strains A2483M, B2483M, E2483M and F2483M exhibited significantly higher specific EPS yields (Yp/x) after 24 h incubation, than that of the parent strain (Yp/x ¼ 0Æ20 ± 0Æ01) (Fig. 3). The E2483M strain was selected for further evaluation on the basis of having the highest Yp/x value for EPS (0Æ26 ± 0Æ003) amongst the selection of mutants tested. The performance of the E2483M mutant

100

200

300 Well number

400

500

(Fig. 4b) was compared directly with the NCFB 2483 parent strain (Fig. 4a) over the time course of 24 h batch fermentations, undertaken in triplicate. Specific yields of EPS relative to biomass of the parent and E2483M strains at 24 h, were similar to those achieved in the preliminary batch-screening (a Yp/x value of 0Æ20 g g)1 for the parent strain vs 0Æ27 g g)1 for the E2483M strain, representing a 35% improvement in specific EPS yield). The E2483M strain thus proved to be suitable for further investigation into the metabolism of EPS production. DISCUSSION The microtitre plate-based system with automatic analysis allowed a relatively rapid means of selecting mutant strains


1·0

20·0

0·9

18·0

0·8

16·0

0·7

14·0

0·6

12·0

0·5

10·0

0·4

8·0

0·3

6·0

0·2

4·0

0·1

2·0

0·0

Yp/x (g lactate g–1 dcw)

Dry cell weight (g l–1); Yp/x [g EPS (dextran equivalents) g–1 dcw]

1204 A . D . W E L M A N ET AL.

0·0 NCFB2483 A2483M

B2483M

C2483M

D2483M

E2483M

F2483M

Strain

1·0

20·0

0·8 15·0 0·6 10·0 0·4 5·0

0·2

0·0 0

4

8

12 Time (h)

16

20

Lactose (g l–1); galactose (g l–1); lactate (g l–1)

(b) 25·0

EPS (g l–1 dextran equivalents); dry cell weight (g l–1)

1·2

0·0 24

0·6 0·5

20·0

0·4 15·0 0·3 10·0 0·2 5·0

0·1

0·0 0

4

8

12 Time (h)

16

20

of Lact. delbrueckii subsp. bulgaricus NCFB 2483 which had a higher acidifying ability than the parent strain. The system was similar to a method described by Venus et al. (1991),

0·0 24

EPS (g l–1 dextran equivalents); dry cell weight (g l–1)

Lactose (g l–1); galactose (g l–1); lactate (g l–1)

(a) 25·0

Fig. 3 Biomass (dry cell weight, ), specific lactate yield (u) and specific EPS yield ( ) of mutants derived from Lactobacillus delbrueckii subsp. bulgaricus NCFB 2483. All values represent the mean of three fermentations. Error bars represent standard deviations

Fig. 4 (a) Batch fermentation profile of Lactobacillus delbrueckii subsp. bulgaricus NCFB 2483 (parent strain). Residual lactose ()); biomass (dry cell weight, n); galactose (s); lactate (u); EPS (*). (b) Batch fermentation profile of L. delbrueckii subsp. bulgaricus mutant strain E2483M. Residual lactose ()); biomass (dry cell weight, n); galactose (s); lactate (u); EPS (*) (mean values). Error bars represent standard deviations

differing in respect of the medium used. Similar small-scale fermentations in microtitre plates have been used to measure growth and lactate production by Lactobacillus LB-WT



(Patnaik et al. 2002). The variation in acidification levels of Lact. delbrueckii subsp. bulgaricus NCFB 2483 (Fig. 1) can be partially explained by the influence of biomass, the tendency of the cells to form pellets in the microtitre wells, and natural variation amongst the population of cells. It was thus important to establish the range of variation in acidification of the parent strain in order for significant differences in mutant strains to be identified. The medium used was effective for purposes of incorporation of dyes. The tests for mucoidy and ropiness were limited in their ability to discriminate smaller differences in EPS production, and were hence used as a confirmatory test. The ability to use an indicator dye in a medium suitable for EPS production, without excessive background interference, represented a useful means by which minor changes in lactate formation could be measured. O’Meara’s reagent was used to detect if any carbon in the mutants was diverted towards acetoin, diacetyl or 2,3-butanediol. It has been proposed that mutants which are defective in lactate dehydrogenase (LDH) would allow the production of increased amounts of alternative catabolites in LAB (Montville et al. 1988). These cell products would however most likely be limited to those arising from the breakdown of pyruvate e.g. diacetyl and acetoin. Kuila and Ranganathan (1978) induced mutations in Streptococcus lactis subsp. diacetylactis with raised levels of diacetyl and acetaldehyde. LDH-deficient mutants of Streptococcus mutans were generated which produced elevated levels of acetoin (Hillman et al. 1987). Whilst this strategy is useful for the diversion of carbon flux to catabolic products, a rerouting of carbon away from glycolysis towards the production of anabolic products such as EPS, would be hampered by the associated reduction in the formation of ATP. The ATP is necessary for the formation of the sugar-nucleotide precursors of EPS. In the present study, it was proposed that selection of mutants which produced an excess supply of ATP via glycolysis while retaining their ability to control the redox potential via the pyruvate–lactate reaction, would be more suited to producing EPS. Some support for this approach was to be found in the findings of Bouzar et al. (1996) who determined that variants of Lact. delbrueckii subsp. bulgaricus CNRZ 1187 which formed more EPS, produced slightly more lactate than the poorer EPS-producers, over 24 h of fermentation. A lower biomass than the parent strain and hence less diversion of sugar nucleotides for cell-wall synthesis was seen to be an additional desirable characteristic. As expected, none of the mutants selected in this study produced products other than lactate from the pyruvate branch-point, as the cells would not need to meet any ATP shortfall via a heterofermentative route (Stephanopoulos et al. 1998). The second tier of fermentation-screening confirmed the effectiveness of the microplate screen in the selection of

1205

mutants with raised acid production and reduced flux of carbon to biomass formation (Fig. 3). The mutant E2483M proved to have stable characteristics in terms of enhanced EPS production (Fig. 4b). Mutants of L. delbrueckii subsp. bulgaricus NCFB 2483 with a presumptively enhanced EPS production relative to the parent strain were selected in the first instance on the basis of enhanced glycolytic activity and the formation of lactate, and reduced biomass. A second tier of fermentationscreening of mutants with these characteristics, yielded four mutants with a higher level of specific EPS production than the parent strain. Final fermentation testing of a selected mutant confirmed a stable trait of enhanced specific EPS production, suitable for further study of EPS metabolism. The screening procedure described above, proved to be an effective strategy for isolating chemically induced mutants with a raised carbon flux to EPS, for the purposes of metabolic studies of EPS-overproduction. ACKNOWLEDGEMENTS This study was supported in part by the New Zealand Foundation for Research, Science and Technology through the Technology for Business Growth programme. REFERENCES Benateya, A., Braquart, P. and Linden, G. (1990) Galactose-fermenting mutants of Streptococcus thermophilus. Canadian Journal of Microbiology 37, 136–140. Boels, I.C., van Kranenburg, R., Hugenholtz, J., Kleerebezem, M. and de Vos, W.M. (2001) Sugar catabolism and its impact on the biosynthesis and engineering of exopolysaccharide production in lactic acid bacteria. International Dairy Journal 11, 723–732. Bouzar, F., Cerning, J. and Desmazeaud, M. (1996) Exopolysaccharide production in milk by Lactobacillus delbrueckii ssp. bulgaricus CNRZ 1187 and by two colonial variants. Journal of Dairy Science 79, 205–211. Dubois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A. and Smith, F. (1956) Colorimetric method for determination of sugars and related substances. Analytical Chemistry 28, 350–356. Eddy, B.P. (1961) The Voges-Proskauer reaction and its significance: a review. Journal of Applied Bacteriology 24, 27–41. Hillman, J.D., Andrews, S.A. and Dzuback, A.L. (1987) Acetoin production by wild-type strains and a lactate dehydrogenasedeficient mutant of Streptococcus mutans. Infection and Immunity 55, 1399–1402. Ibrahim, S.A. and O’Sullivan, D.J. (2000) Use of chemical mutagenesis for the isolation of food grade b-galactosidase overproducing mutants of bifidobacteria, lactobacilli, and Streptococcus thermophilus. Journal of Dairy Science 83, 923–930. Kimmel, S.A. and Roberts, R.F. (1998) Development of a growth medium suitable for exopolysaccharide production by Lactobacillus delbrueckii ssp. bulgaricus RR. International Journal of Food Microbiology 40, 87–92.


1206 A . D . W E L M A N ET AL.

Kuila, R.K. and Ranganathan, B. (1978) Ultraviolet light-induced mutants of Streptococcus lactis subspecies diacetylactis with enhanced acid- or flavor- producing abilities. Journal of Dairy Science 61, 379–383. Kuila, R.K., Ranganathan, B., Dutta, S.M. and Laxminarayana, H. (1971) Induction of mutation in Streptococcus diacetylactis by nitrosoguanidine and ultraviolet radiation. Journal of Dairy Science 54, 331–335. Miyamoto, T., Reddy, N.S. and Nakae, T. (1983) Induction of mutation in Lactobacillus casei subsp. alactosus by nitrosoguanidine. Agricultural and Biological Chemistry 47, 2755–2759. Monnet, C. and Corrieu, G. (1998) Selection and properties of Lactobacillus mutants producing a-acetolactate. Journal of Dairy Science 81, 2096–2102. Montville, T.J., Koch, G.E. and Johnston, S.J. (1988) Partial biochemical characterization of tetrazolium red-reactive Lactobacillus plantarum mutants. Food Biotechnology 2, 81–91. O’Meara, R. (1931) A simple delicate and rapid method of detecting the formation of acetylmethylcarbinol bacteria fermenting carbohydrates. Journal of Pathology and Bacteriology 34, 401– 406. Patnaik, R., Louie, S., Gavrilovic, V., Perry, K., Stemmer, W.P.C., Ryan, C.M. and del Cardayre´, S. (2002) Genome shuffling of

Lactobacillus for improved acid tolerance. Nature Biotechnology 20, 707–712. Piatkiewicz, A., Libudzisz, Z. and Oberman, H. (1981) Selection and mutation of Streptococcus diacetylactis for butter production. In Advances in Biotechnology. Fuels, Chemicals, Foods, and Waste Treatment, vol. II. ed. Moo-Young, M. and Robinson, C.W. pp. 491–495. Toronto: Pergamon Press. Renault, P.P. and Heslot, H. (1987) Selection of Streptococcus lactis mutants defective in malolactic fermentation. Applied and Environmental Microbiology 53, 320–324. Ross, L.F. and Chapital, D.C. (1987) Simultaneous determination of carbohydrates and products of carbohydrate metabolism in fermentation mixtures by HPLC. Journal of Chromatographic Science 25, 112–117. Stephanopoulos, G.N., Aristidou, A.A. and Nielsen, J. (1998) Metabolic flux analysis. In Metabolic Engineering Principles and Methodologies ed. Stephanopoulos, G.N., Aristidou, A.A. and Nielsen, J. pp. 309–351. San Diego, CA: Academic Press. Venus, J., Idler, F. and Albrecht, C. (1991) New ways of selecting lactic acid bacteria for biotechnological processes. Applied Microbiology and Biotechnology 37, 240–243. de Vos, W.M. (1996) Metabolic engineering of sugar catabolism in lactic acid bacteria. Antonie van Leeuwenhoek 70, 223–242.
