Lactobacillus casei metabolic potential to utilize citrate as an energy source in ripening cheese: a bioinformatics approach

Authors


  • Present addresses
    I. Díaz-Muñiz, United States Department of Agriculture, Agricultural Research Services, Food Science Research Unit, North Carolina State University, 322 Schaub Hall, Campus Box 7624, Raleigh, North Carolina 27695-7624, USA.

  • E.G. Dudley, University of Maryland-School of Medicine, 685 W. Baltimore St., Room 447, Baltimore, MD 21201, USA.

J.L. Steele, 1605 Linden Drive, University of Wisconsin-Madison, Madison, WI, 53706, USA. E-mail: jlsteele@wisc.edu

Abstract

Aims:  To identify potential pathways for citrate catabolism by Lactobacillus casei under conditions similar to ripening cheese.

Methods and Results:  A putative citric acid cycle (PCAC) for Lact. casei was generated utilizing the genome sequence, and metabolic flux analyses. Although it was possible to construct a unique PCAC for Lact. casei, its full functionality was unknown. Therefore, the Lact. casei PCAC was evaluated utilizing end-product analyses of citric acid catabolism during growth in modified chemically defined media (mCDM), and Cheddar cheese extract (CCE). Results suggest that under energy source excess and limitation in mCDM this micro-organism produces mainly l-lactic acid and acetic acid, respectively. Both organic acids were produced in CCE. Additional end products include d-lactic acid, acetoin, formic acid, ethanol, and diacetyl. Production of succinic acid, malic acid, and butanendiol was not observed.

Conclusions:  Under conditions similar to those present in ripening cheese, citric acid is converted to acetic acid, l/d-lactic acid, acetoin, diacetyl, ethanol, and formic acid. The PCAC suggests that conversion of the citric acid-derived pyruvic acid into acetic acid, instead of lactic acid, may yield two ATPs per molecule of citric acid. Functionality of the PCAC reductive route was not observed.

Significance and Impact of the Study:  This research describes a unique PCAC for Lact. casei. Additionally, it describes the citric acid catabolism end product by this nonstarter lactic acid bacteria during growth, and under conditions similar to those present in ripening cheese. It provides insights on pathways preferably utilized to derive energy in the presence of limiting carbohydrates by this micro-organism.

Introduction

Lactic acid bacteria (LAB) are utilized in cheese manufacture due to their ability to primarily metabolize lactose into lactic acid. Starter lactic acid bacteria (SLAB) are added to milk at the beginning of cheese manufacture. Nonstarter lactic acid bacteria (NSLAB) are present in the cheese matrix as a result of contamination during the manufacturing procedure. Numbers of NSLAB begin to increase in the cheese matrix after the first two weeks postmanufacture. Therefore, when NSLAB cell numbers increase, most of the residual lactose in cheese has been utilized by SLAB. Alternate potential energy sources for NSLAB in cheese are nucleic acids derived from the autolysis of SLAB (Thomas 1986), amino acids (Kristoffersen 1956; Kieronczyk et al. 2001), sugars liberated from glycoproteins and glycolipids present in milk (Williams and Banks 1997; Fox et al. 1998; Williams et al. 2000), and citrate (Campbell and Gunsalus 1944; Fryer 1970; Fryer et al. 1970; Branen and Keenan 1977; Starrenburg and Hugenholtz 1991). To further examine the role of these potential energy sources for NSLAB in ripening cheese, it is necessary to explore the metabolic potential of NSLAB. This will not only advance the knowledge of NSLAB physiology, but also our understanding of the role of these micro-organisms in cheese flavour development (Reiter et al. 1967; Law et al. 1976).

Citrate catabolism and uptake by LAB, more specifically Lactococcus lactis biovar diacetylactis and Leuconostoc mesenteroides is generally considered as a secondary metabolic energy-generating process (Hugenholtz et al. 1993; Marty-Teysset et al. 1996a; Bandell et al. 1998). The citrate permeases from both micro-organisms are secondary transporters, which transport divalent citrate in symport with a proton, and efficiently exchange divalent citrate with monovalent d-lactate (Marty-Teysset et al. 1995; Bandell et al. 1998). Translocation of a net negative charge into the cell during transport results in the generation of a membrane potential (Marty-Teysset et al. 1995). It has been reported that intracellular conversion of the citrate-derived divalent oxaloacetic acid to monovalent pyruvic acid consumes a cytoplasmic proton (Hugenholtz et al. 1993). Therefore, citrate transport and conversion into lactate generates both membrane potential and a pH gradient or proton motive force. The Leuc. mesenteroides citrate permease requires glucose as a co-metabolite to generate proton motive force (Marty-Teysset et al. 1996b). An equivalent citrate transport system in NSLAB, more specifically lactobacilli, has not been described to date. However, an open reading frame with homology to citrate transporters has been identified upstream to the putative cit operon in Lact. casei ATCC334 (http://maple.lsd.ornl.gov/microbial/lcas/02feb04/ (February 2004).

Results obtained in our laboratory suggest that growth of the NSLAB, Lact. casei ATCC 334, is enhanced in the presence of citrate in modified chemically defined media (mCDM)-containing limiting carbohydrate. Moreover, citrate catabolism by Lact. casei in the presence of excess carbohydrate is not initiated until the culture reaches stationary phase. Therefore, this research evaluated potential pathways for citrate catabolism by Lact. casei ATCC334. It identified catabolic end products in the presence of limiting and excess carbohydrate in mCDM and in our cheese model system, Cheddar cheese extract (CCE). Putative metabolic pathways for citrate catabolism by Lact. casei were identified using the genome sequence, and metabolic flux analyses.

Materials and methods

Genome sequence analysis

The Lact. casei ATCC334 finished genome sequence was accessed at http://maple.lsd.ornl.gov/microbial/lcas/02feb04/ (February 2004). The database was screened to identify putative gene products involved in citrate catabolism. The putative genes were grouped under three categories: (i) citric acid cycle (CAC)-related genes; (ii) pyruvate catabolism-related genes; (iii) and other relevant genes (Table 1).

Table 1. Lactobacillus casei ATCC334 putative genes potentially involved in citrate catabolism
Gene I. D. from genome sequenceFunctional descriptions of gene-coded productsAbbreviations for putative gene-coded products
  1. *Putative genes encoding for enzymes potentially involved in citrate catabolism, which showed E-values lower than E-5, lack a gene I.D. number. Therefore, gene I.D. numbers composed of GS X were assigned, were X represents a number between 1 and 3.

Citric acid cycle
 475Pyruvate carboxylasePyc
 1166Succinate dehydrogenaseSdh
 1341Possible fumarate reductaseFdr
 1936Fumarate hydratase (fumarase)Fum
 2347Citrate lyase gamma subunitCit
 2348Citrate lyase beta subunitCit
 2349Citrate lyase alpha subunitCit
 2373Phosphoenolpyruvate carboxikinasePck
Pyruvate metabolism
 69Malic enzyme, NAD binding domainMle1
 90Alcohol dehydrogenase/acetaldehyde dehydrogenaseAdh
 245Phosphate acetyltransferasePta
 465Pyruvate dehydrogenase complex, alpha subunitPdh1
 466Pyruvate dehydrogenase complex, beta subunitPdh2
 467Pyruvate dehydrogenase complex, dihydrolipoamide acetyltransferasePdh3
 468Pyruvate dehydrogenase complex, dihydrolipoamide dehydrogenasePdh4
 470Malate/lactate dehydrogenaseLdh1
 516Pyruvate kinasePyk
 557Pyruvate–formate lyase formate C-acetyltransferasePfl
 558Pyruvate–formate lyase-activating enzymePfl
 604AcylphosphataseAcyp
 730Acetate kinaseAck
 738Alanine dehydrogenaseAladh
 839Malate/lactate dehydrogenaseLdh2
 1033Hydroxyisocaproate dehydrogenaseHic1
 1047Acetate kinaseAck2
 1242Pyruvate oxidasePox1
 1599Malic enzymeMle2
 1777Glutathione reductase (dihydrolipoamide dehydrogenase)Glur
 1793l-2-hydroxyisocaproate dehydrogenaseL-Hic
 1987Phosphoenolpyruvate synthase/pyruvate phosphate dikinasePps
 2005D-lactate ferricytochrome c oxidoreductaseD-Lfcco
 2117Hydroxyisocaproate dehydrogenaseHic2
 2131Pyruvate oxidasePox3
 2160Acetyl-CoA carboxylase-Biotin carboxyl carrier proteinAcc
 2163Acetyl-CoA carboxylase beta subunitAcc
 2164Acetyl-CoA carboxylase alpha subunitAcc
 2344Oxaloacetate decarboxylase, sodium ion pump subunitOad
 2351Putative pyruvate carboxylase/oxaloacetate decarboxylase, alpha subunitOad
Pyruvate metabolism
 2360Acetolactate synthaseAls
 2361Alpha-acetolactate decarboxylaseAld
 2400Acetyl-CoA acetyltransferaseAcat
 2404Pyruvate oxidasePox2
 2689Dihydrolipoamide dehydrogenasePdh2
 2691Pyruvate dehydrogenase, beta subunitPdh2
 2955Malate/lactate dehydrogenaseLdh3
Other relevant genes
 59Acetoacetate decarboxylase (putative)Aadc
 772NADH dehydrogenaseNdh
 1062NADH oxidaseNox
 1602Aspartate-ammonia lyaseAspal
 2654Aspartate aminotransferaseAspat
 GS1*Formate dehydrogenase (cytochrome)Fdh
 GS2*Acetoin racemaseAr
 GS3*(S, S)-butanediol dehydrogenaseBdh

Metabolic flux analysis

A catalogue of reactions potentially catalysed by the products of the selected putative genes was created. This catalogue was utilized as an input file for the metabolic flux analysis programme, Metatool (Pfeiffer et al. 1999; Schuster et al. 2000) (http://pinguin.biologie.uni-jena.de/bioinformatik/networks/metatool/metatool.html; December 2004). Potential citrate catabolism pathways identified by Metatool were utilized to create a customized putative citric acid cycle (PCAC) map for Lact. casei ATCC334 (Fig. 1).

Figure 1.

 Putative citric acid cycle for Lactobacillus casei ATCC334: enzyme names are abbreviated as shown in Table 1. The A/drc abbreviation represents acetoin/diacetyl reductase. The coding gene for this enzyme was not identified in the Lact. casei genome sequence. Potential end products are succinate, carbon dioxide, l-Alanine, l/d-lactate, diacetyl, ethanol, and 2,3-butanediol.

Bacterial strain

The bacterial strain, Lact. casei ATCC334 was obtained from the American Type Culture Collection (Rockville, MD, USA). This strain is a cheese isolate, and the proposed neotype of the Lact. casei cluster (Dicks et al. 1996; Dellaglio et al. 2002). This culture was maintained at −80°C in lactobacilli MRS broth (Difco Laboratories, Detroit, MI, USA) and 10% glycerol. Lactobacillus casei was transferred to MRS broth from frozen stocks. Incubation was performed in a water bath at 37°C for 16–18 h. Cultures were transferred twice in MRS broth prior to inoculation into the experimental media.

mCDM and growth conditions

mCDM was prepared as described by Christensen and Steele (2003) with the following modifications: (i) all individual amino acids were substituted by 10·0 g of Bacto Casitone (Becton Dickinson and Company, Sparks, Maryland, USA), and (ii) it was supplemented with 2·5 mg pyridoxamine dihydrochloride (Sigma-Aldrich Co., St Louis, MO, USA), 1 mg of resazurin sodium salt (Sigma-Aldrich Co.) per litre. Substitution of individual amino acids by Bacto casitone was necessary to observe Lact. casei ATCC334 growth. The mCDM pH was adjusted to 5·2 ± 0·1 with hydrochloric acid, if required. The carbohydrate content was modified as indicated in the text. Batch cultures were grown under a nitrogen headspace in a 2-l fermentor vessel (New Brunswick Scientific, Co., Edison, NJ, USA) at 37°C. The agitation rate was kept at 40 rev min−1. pH was monitored and maintained at 5·2 ± 0·1 with 0·1 mol l−1 sodium hydroxide using a VirTis Omni Culture Fermentor (The VirTis Co., Inc., Gardiner, NY, USA) connected to a Chemtrix Type 45AR pH Controller (Whatman Lab Sales, Inc., Hillsboro, OR, USA), a MasterFlex Speed Controller (Barnant Co., Barrington, IL, USA), and a MasterFlex Pump Model No. 701400 (Cole-Parmer Instruments Co., Chicago, IL, USA).

Growth was initiated with a 0·1% washed inoculum. Bacterial cells were washed twice using Dulbecco's phosphate-buffered saline (Invitrogen Corporation, Carlsbad, CA, USA) at 5930 g for 10 min at 25°C using an induction drive centrifuge (Beckman Coulter) prior to their transfer to the experimental media. All the experiments conducted in mCDM were performed in duplicate.

CCE and growth conditions

CCE was obtained from 2·5-month-old Cheddar cheese. A Cheddar cheese block was cut into 2 1/2′′ cubes, shredded using a commercial food processor (AB Hallde Maskiner, Kista, Sweden), divided into approximately 1 kg sub-samples, and placed in plastic bags. Sub-samples were frozen for 2 h at −20°C and subsequently held at −80°C until lyophilization was initiated. Sub-samples were lyophilized for 48 h at room temperature followed by a cycle of 24 h at 50°C (Unitrap Model, Virtis, Gardiner, NY, USA). The three freeze-dried subsamples of 500 g each were ground to a fine powder using a commercial food processor and stored in sealed plastic bags at 5°C until extracted. One lyophilized-powdered sub-sample of 500 g was extracted with 2 l of cold sterile distilled water in a commercial blender (Waring Product Division New Hartford, CT, USA) for 10 min at low speed. The extract was cooled in an ice bath for 10 min, and the liquid/solid mixture percolated through cheesecloth to remove solids. The liquid (2 l) was re-used to extract the two remaining lyophilized and powdered sub-samples to achieve the desired ratio of four parts of water to three parts of freeze-dried cheese. Three parts of freeze-dried cheese are equivalent to six parts of cheese. The desired ratio was based on the moisture content of the original cheese, which was 40%. Subsequently, the extracts were aliquoted into centrifuge tubes (Nalgene Nunc International, Rochester, NY, USA), cooled in an ice bath for 10 min, and centrifuged for 10 min at 9000 g at 5°C (Beckman Instruments Inc., Arlington Heights, IL, USA). Supernatants or the CCE sub-samples were transferred into 2-l plastic bottles and stored at −20°C for 16–18 h. CCE were thawed in a 37°C water bath (Precision Scientific, Chicago, IL, USA), and centrifuged for 20 min at 22 000 g at 5°C to remove fine solids. The supernatants were subsequently filtered through a 0·7-μm TCLP filter (Fisher, Fair Lawn, NJ, USA), and a 0·2-μm filter unit (Nalgene Nunc International, Rochester, NY, USA). The filtered CCE were mixed and re-filtered into a sterile 2·5 l bottle with a 0·2 μm bottle top filter to ensure sterility of the media. The filter-sterilized CCE pool was stored at −20°C until needed.

Lactobacillus casei growth was monitored in CCE, which naturally contained 0·58 ± 0·01 mmol l−1 of lactose, 1·40 ± 0·09 mmol l−1 of galactose, and 2·3 ± 0·13 mmol l−1 of citrate. This CCE was supplemented with 12·7 mmol l−1 of diammonium citrate to obtain a final concentration of 15 mmol l−1 of citrate prior to inoculation. CCE pH was re-adjusted to 5·1 ± 0·1 with 0·1 mol l−1 of sodium hydroxide, following supplementation of diammonium citrate. CCE was prereduced in 40-ml GC vials, fitted with teflon septa and screw caps (I-Chem Brand Products, Rockwood, TN, USA), by incubation in an anaerobic jar for 48 h prior to inoculation. Growth was initiated with a 0·1% washed inocula. Bacterial cells were washed twice using Dulbecco's phosphate-buffered saline (Invitrogen Corporation, Carlsbad, CA, USA) at 5930 g for 10 min at 25°C prior to their transfer to the experimental media. Cultures were incubated at 37°C in anaerobic jars. All the experiments performed in CCE were done in triplicate.

Organic acids and galactose detection

l- and d-lactic acids, l-malic acid, succinic acid, formic acid, acetic acid, and ethanol production was monitored using the enzymatic kits available from R-Biopharm, Inc. (Marshall, MI, USA). Similarly, carbohydrate (lactose and galactose), and citrate disappearance was monitored using the enzymatic detection kits available from R-Biopharm, Inc. Samples for detection assays were collected at the beginning of the lag phase, early logarithmic phase, late logarithmic phase, and stationary phase. Collected samples were spun at 16 000 g for 5 min. Supernatants were stored at −20°C until organic acids, ethanol, and carbohydrate concentrations were determined.

Gas chromatography

Samples (2·0 ml) were collected at the beginning of the lag phase, early logarithmic phase, late logarithmic phase, and stationary phase in sterile 8-ml vials fitted with teflon septa and screw caps (I-Chem Brand Products). Samples were stored at −20°C until volatile profiles were analysed using solid-phase microextraction (SPME) combined with gas chromatography-mass spectrometry (GC-MS). For analyte equilibration purposes, samples were retained at 30°C for 30 min. A manual SPME holder fitted with an 85 μm carboxen/PDMS-coated fibre (Supelco Inc, Bellefonte, PA, USA) was used for volatile absorption. The fibre was placed in the static, equilibrated sample headspace for 10 min. Volatiles were analysed with a gas chromatograph (6890N; Agilent Technologies Inc., Palo Alto, CA, USA) with a mass selective detector (5973N; Agilent Technologies Inc.) fitted with a capillary column (Rtx-Wax, 30 m, 0·25 mm i.d., 0·5 μm stationary phase, Restek Inc., Bellefonte, PA, USA). Chromatography conditions were: inlet at 280°C using helium carrier gas at 0·8 ml min−1. The fibre was held in the injection port over the complete GC programme cycle to ensure complete desorption. The GC temperature programme included an initial temperature of 35°C for 3 min ramping at 20°C min−1 until a final temperature of 250°C with 4 min hold. The mass scan range was m/z 21–400 at 3·81 scans s−1. The ionization energy was 70 eV. The quantitation of volatiles was done using integration values using standard curves.

Results

Lactobacillus casei PCAC

Putative genes potentially involved in citrate catabolism were identified (Table 1). The PCAC map generated from the steady-state reactions identified by Metatool (Pfeiffer et al. 1999; Schuster et al. 2000) shows a completely balanced cycle (Fig. 1). The putative map includes the reductive tricarboxylic acid cycle route, which is involved in the conversion of citric acid into succinic acid; the pyruvate oxidative route, and the nonoxidative cleavage of pyruvate route, which requires pyruvate dehydrogenase, and pyruvate-formate lyase, respectively, and may produce acetic acid and/or ethanol from citric acid; and the pyruvate condensation route, which theoretically converts two molecules of the citric acid-derived pyruvic acid into one molecule of either acetoin, 2,3-butanediol or diacetyl via α-acetolactate. These four pathways have traditionally been included in the citrate catabolism map for LAB. However, the putative fumarase and malic enzyme encoding genes are unique to the Lact. casei ATCC334 PCAC map. These putative genes encode proteins theoretically linking the potential conversion of the citric acid-derived fumaric acid into pyruvic acid. The possible conversion of the citrate-derived fumaric acid into pyruvic acid closes the cycle.

Detection of end products

The detection of end products was performed from cultures growing in the presence of either excess (50 mmol l−1) or limiting (2·5 mmol l−1) carbohydrate in mCDM and in CCE. Citrate utilization in the presence of both limiting and excess carbohydrate in mCDM was initiated after 70% of the initial carbohydrate concentration was consumed. Acetic acid was the main product of limiting galactose catabolism even in the absence of citric acid (Fig. 2a and Table 2). Lactobacillus casei mainly converted 14·9 ± 2·9 mmol l−1 of citric acid into 28·6 ± 3·9 mmol l−1 of acetic acid and 3·68 ± 0·41 mmol l−1 of d-lactic acid when the initial galactose concentration was limiting (2·30 ± 0·23 mmol l−1) (Fig. 3a and Table 2). In the presence of excess carbohydrate (49·9 ± 0·61 mmol l−1) and 14·1 ± 1·0 mmol l−1 of citric acid, the main catabolic product was l-lactic acid (68·6 ± 1·1 mmol l−1) (Fig. 4a and Table 2). However, 17·4 ± 3·0 mmol l−1 of acetic acid and 8·48 ± 1·5 mmol l−1 of d-lactic acid were produced (Fig. 4a and Table 2).

Figure 2.

 Limiting galactose catabolism end products formed by Lactobacillus casei: Lact. casei ATCC334 was grown in modified chemically defined media supplemented with 2·31 ± 0·29 mmol l−1 of galactose at pH 5·1. (a) Conversion of galactose into acetate, L-lactate, and D-lactate is depicted (inline image, L-lactate; bsl00011, acetate; □, D-lactate; bsl00001, galactosc; bsl00066, growth). (b) Formation of diacetyl (bsl00001), and formate (bsl00036) from catabolism of limiting galactose.

Table 2.   End products of citrate and galactose catabolism by Lactobacillus casei
Substrates/end productsLimiting galactose (mmol l−1 OD−1)Limiting galactose and citrate (mmol l−1 OD−1)Excess galactose and citrate (mmol l−1 OD−1)
Galactose5·01 ± 0·32·31 ± 0·227·2 ± 0·6
Citrate14·3 ± 0·67·88 ± 0·9
Acetic acid17·4 ± 2·427·5 ± 3·911·9 ± 2·1
l-lactic acid1·13 ± 0·30·55 ± 0·337·3 ± 1·1
d-lactic acid1·69 ± 0·13·54 ± 0·44·61 ± 4·5
Formic acid0·16 ± 0·10·29 ± 0·1
Diacetyl0·13 ± 0·10·07 ± 0·10·11 ± 0·2
Acetoin0·57 ± 0·13·37 ± 0·8
Figure 3.

 Limiting galactose and citrate catabolism end products formed by Lactobacillus casei: Lact. casei ATCC334 was grown in modified chemically defined media supplemented with 2·31 ± 0·23 mmol l−1 of galactose and 14·9 ± 1·5 mmol l−1 of diammonium citrate at pH 5·1 ± 0·2. (a) Conversion of galactose and diammonium citrate into acetate and D-lactate (inline image, L-lactate; bsl00011, acetate; □, D-lactate; bsl00001, galactose; –bsl00066–, growth; –□–, citrate). (b) Formation of diacetyl (bsl00001), formate (bsl00036), and acetoin (□) from catabolism of limiting galactose and citrate is also depicted.

Figure 4.

 Excess galactose and citrate catabolism end products formed by Lactobacillus casei: Lact. casei ATCC334 was grown in modified chemically defined media supplemented with 49·9 ± 0·63 mmol l−1 of galactose and 14·1 ± 1·0 mmol l−1 of diammonium citrate at pH 5·1 ± 0·2. (a) Conversion of galactose and diammonium citrate into L-lactate, acetate, and D-lactate is depicted (inline image, L-lactate; bsl00011, acetate; □, D-lactate; bsl00001, galactose; –bsl00066–, growth; –□–, citrate). (b) Formation of diacetyl (bsl00001) and acetoin (□) from catabolism of excess galactose and citrate.

Formation of formic acid, diacetyl, and acetoin was detected in trace amounts. In the presence of limiting galactose and absence of citric acid, 0·080 ±0·01 mmol l−1 of formic acid, and 0·141 ± 0·01 mmol l−1 of diacetyl were produced (Fig. 2b and Table 2). Acetoin production was not detected under these conditions. In the presence of limiting galactose and 14·9 ± 2·9 mmol l−1 of citric acid, 0·290 ± 0·02 mmol l−1 of formic acid, 0·121 ± 0·01 mmol l−1 of diacetyl, and 0·59 ± 0·06 mmol l−1 of acetoin were produced (Fig. 3b and Table 2). In the presence of excess galactose and citric acid, 0·211 ± 0·03 mmol l−1 of diacetyl and 6·21 ± 0·82 mmol l−1 of acetoin were produced (Fig. 4b and Table 2). Formic acid was not produced in the presence of excess carbohydrate. Production of ethanol, butanediol, malic acid, and succinic acid was not detected under any of the conditions studied in mCDM (data not shown).

End products detected from cultures growing in CCE containing 0·580 ± 0·01 mmol l−1 of lactose, 1·40 ± 0·09 mmol l−1 of galactose, and 14·7 ± 0·13 mmol l−1 of citrate include all of the end products detected in mCDM and ethanol (Fig. 5). Lactobacillus casei produced 2·67 ± 0·06 mmol l−1 of acetic acid while lactose catabolism occurred (Fig. 5a). Subsequently, Lact. casei ATCC334 mainly produced acetic acid (22·5 ± 1·2 mmol l−1), l-lactic acid (14·2 ± 0·09 mmol l−1), and d-lactic acid (1·75 ± 0·05 mmol l−1) (Fig. 5a). Trace concentrations of diacetyl (0·130 ±0·01 mmol l−1), formic acid (0·070 ± 0·03 mmol l−1), acetoin (1·03 ± 0·07 mmol l−1), and ethanol (0·101 ± 0·02 mmol l−1) were also detected (Fig. 5b). Production of succinic acid, malic acid, and butanediol was not observed in CCE (data not shown).

Figure 5.

 Citrate catabolism end products formed by Lactobacillus casei in Cheddar cheese extract (CCE): Lact. casei was grown in CCE containing 0·58 ± 0·01 mmol l−1 of lactose, 1·40 ± 0·09 mmol l−1 of galactose, and 14·7 ± 0·13 mmol l−1 of citrate. (a) Conversion of carbohydrates and citrate into acetate, and L-/D-lactate (bsl00003, acetate; inline image, L-lactate; □, D-lactate; bsl00018, galactose; bsl00001, citrate; bsl00008, lactose; bsl00066 CFU/ ml−1). (b) Formation of diacetyl (bsl00001), formate (bsl00036), acetoin (□), and ethanol (bsl00003) is also depicted.

Discussion

Traditionally, the citrate catabolic pathways are depicted with the citrate derived-pyruvate being at the centre of a multibranch map, and there is no connection between the different pathways. Identification of the putative genes potentially involved in citrate catabolism by Lact. casei using the genome sequence and Metatool (Pfeiffer et al. 1999; Schuster et al. 2000) allowed for the design of an unbiased PCAC map. Conversion of the traditionally multibranch citric acid catabolism map into a cycle was possible due to the presence of the putative fumarase and malic enzyme coding genes in the Lact. casei ATCC334 genome. It was the objective of this research to identify end products of citrate catabolism by Lact. casei, and the potential pathways utilized from the PCAC map.

The influence of ecological factors present in ripening cheese in citric acid catabolism by Lact. casei ATCC334 has been previously studied in our laboratory. Results obtained suggest that this micro-organism is able to utilize citric acid in the absence of other more readily metabolized energy sources, especially lactose. Moreover, under carbohydrate limitation in mCDM, growth of this micro-organism is enhanced concomitantly with citric acid utilization. Therefore, galactose and citric acid catabolism end products in the presence of limiting and excess carbohydrate were identified in this research.

The data shown herein suggest that under energy limitation, the main catabolic product is acetic acid. However, in the presence of an excess of energy sources, the main catabolic product is l-lactic acid. It has been reported that l-Ldh, which is the main enzyme catalysing the conversion of pyruvic acid into l-lactic acid in LAB, is fructose diphosphate dependent (Wolin 1964; Thomas 1976). Under carbohydrate limitation, intracellular levels of fructose diphosphate are reduced and thus l-Ldh activity is likely to be hindered. Under this condition the intracellular pyruvic acid pool would be converted into acetic acid instead of l-lactic acid. Formation of acetic acid instead of l-lactic acid from the citric acid and galactose-derived pyruvic acid may yield from two to four ATPs per molecule, respectively.

The formation of trace amounts of formic acid in the presence of citric acid and/or limiting galactose suggest the possibility of an active pyruvate–formate lyase complex in this micro-organism. Formation of formic acid was uniquely observed in cultures growing in limited galactose. The lactococcal pyruvate–formate lyase is strongly inhibited by either d-glyceraldehyde-3-phosphate or dihydroxyacetone phosphate (Takashi et al. 1982). The observations gathered herein suggest that if indeed an active pyruvate–formate lyase is present in Lact. casei, its regulatory mechanism would not be significantly different from previously described pyruvate–formate lyases.

Genes encoding for some of the pyruvate–formate lyase subunits could not be identified by sequence homology in the Lact. casei genome sequence (Table 1). The pyruvate–formate lyase system in Escherichia coli is composed of six proteins: the pyruvate–formate lyase dimmer, an activating enzyme, a deactivating enzyme, a flavodoxin, an NADPH:flavodoxin oxidoreductase, and a pyruvate:flavodoxin oxidoreductase (Knappe and Sawers 1990). Sequence homology analysis suggest that the Lact. casei genome encodes for the pyruvate–formate lyase (formate C-acetyltransferase) and the activating enzyme. Additionally, this analysis suggests that the Lact. casei genome contains multiple genes with some degree of homology to oxidoreductases. Therefore, identification of putative genes encoding for a specific oxidoreductase with a significant degree of certainty is difficult. Although, indirect evidence for the presence of this complex in Lact. casei was obtained, further studies are required to confirm the presence of such a complex in this micro-organism.

The concentration of formate detected was significantly lower than the expected value for the amount of substrate initially present. This result suggests that enzymes competing with pyruvate–formate lyase for pyruvate may be present in this organism. Other enzymes, which could be competing with pyruvate–formate lyase for pyruvate, are pyruvate dehydrogenase, and pyruvate oxidase.

Pyruvate oxidase catalyses the oxidative decarboxylation of pyruvate in the presence of phosphate and oxygen, yielding acetyl phosphate, carbon dioxide, and hydrogen peroxide. In an independent study in our laboratory, it has been observed that Lact. casei reduces the media redox potential during the lag phase of growth to at least −150 mV. Therefore, growth occurs in either the absence of oxygen or in the presence of limiting oxygen. These observations suggest that participation of the Lact. casei pyruvate oxidase in the metabolism of pyruvate is unlikely. Therefore, the only alternate pathway remaining is through the activity of the pyruvate dehydrogenase complex.

The pyruvate dehydrogenase complex from Lact. lactis biovar diacetylactis reaches maximal activity in aerobic and lactose-limited cultures and is virtually absent in anaerobic cultures (Snoep et al. 1992). However, it has been reported that under anaerobic conditions cells of Enterococcus faecalis have significant pyruvate dehydrogenase activity at low pH values (Snoep et al. 1993). Therefore, it is still possible that this enzymatic complex is mediating conversion of pyruvate to acetyl-CoA in Lact. casei under the conditions studied.

d-lactic acid formation was observed either concomitanly with citrate catabolism or catabolism of limiting galactose. Although, the activities of both l-Ldh and d-Ldh are mostly fructose diphosphate dependent, some dehydrogenases are fructose diphosphate independent. Multiple putative genes encoding for both l- and d-lactate dehydrogenases were identified in the Lact. casei genome sequence. Additionally, a total of three putative genes potentially encoding for hydroxyisocaproate dehydrogenases were identified. Hydroxyisocaproate dehydrogenases are capable of reducing pyruvic acid and dehydrogenate lactic acid (Schütte et al. 1984). Therefore, d-lactic acid could have been formed by a fructose diphosphate-independent d-lactate dehydrogenase or by a hydroxyisocaproate dehydrogenase.

Gas chromatography results revealed production of trace amounts of diacetyl and acetoin in the presence of both limiting or excess galactose and citric acid (Figs 3b and 4b). In the presence of limiting carbohydrate alone, no acetoin formation was observed. These observations confirm previous reports regarding an increased yield of α-acetolactate-derived end products as the intracellular pyruvic acid concentration increases. Formation of the diacetyl and acetoin precursor, α-acetolactate, requires the condensation of two molecules of pyruvic acid (Juni 1951). Therefore, competitive pathways for the metabolism of pyruvic acid may be favoured over formation of α-acetolactate. Conversion of the acetolactate-derived diacetyl into acetoin has the potential to regenerate one additional NAD+. Lack of butanediol formation was unexpected. However, theoretically the conversion of butanediol to either diacetyl or acetoin would generate NADH instead of NAD+. Generation of NADH may have been necessary to balance the redox potential.

Conversion of citric acid into the products reported herein suggests that Lact. casei ATCC334 has a functional citrate lyase and the capability to convert oxaloacetic acid into pyruvic acid via either pyruvate carboxylase, oxaloacetate decarboxylase (Oad), and/or pyruvate carboxykinase/pyruvate kinase. Citrate lyase activity in Lact. casei ATCC334 was previously described (Dudley and Steele 2000). Genes putatively encoding citrate lyase in Lact. casei ATCC334 are found adjacent to genes for a putative membrane-bound Oad. Previous studies done in our laboratory suggest that both Lact. casei ATCC334 and Lactobacillus zeae ATCC393 posses an active Oad (data not shown). DNA sequence analysis suggested that Oad from both of these strains is a three-subunit protein complex, which comprises of a dicarboxylic acid decarboxylase (OadA), a membrane-spanning carboxybiotin-carrier protein decarboxylase (OadB), and a biotin-binding protein (OadC) (Dudley and Steele 2000). This analysis suggests that OadABC belongs to a family of membrane-bound decarboxylases found in fermentative bacteria (Dimroth and Schink 1998) that use the energy released during decarboxylation to generate a membrane chemical gradient. This observation may explain why citrate enhances the growth of Lact. zeae ATCC 393 (Branen and Keenan 1977), and suggests that a similar phenotype may occur with Lact. casei ATCC 334. A putative gene encoding for a cytoplasmic oxaloacetate decarboxylase, homologous to the Lact. lactis enzyme was not identified in the Lact. casei genome sequence. Although, pyruvate carboxylase has been characterized in Lact. lactis subsp. lactis C2 and identified among other lactococci (Wang et al. 2000), it is still unknown whether pyc is a functional gene in Lact. casei.

The results presented herein are in agreement with previous reports for Lact. lactis and other lactobacilli. Starrenburg and Hugenholtz (1991) reported that Lact. lactis biovar diacetylactis homofermentatively converts pyruvic acid to lactic acid in the presence of high lactose concentrations. Additionally, this group reported that mixed-acid fermentation with formic acid, ethanol, and acetic acid as products is observed under conditions of lactose limitation in continuous culture. Ethanol production by growing cells of Lact. casei was not observed in the presence of either limiting or excess galactose in mCDM. However, trace amounts of ethanol were produced by Lact. casei when growing in CCE. Putative genes encoding for alcohol dehydrogenase and acetaldehyde dehydrogenase were identified by sequence homology in the Lact. casei genome. This result may suggest that although the machinery required for the production of ethanol is present in Lact. casei, this metabolic route is not favoured. Formation of ethanol from pyruvic acid may provide the cells with additional ways to regenerate NADH. However, regeneration of NADH by this route would reduce the quantities of ATP that could be formed via substrate level phosphorylation from the conversion of pyruvic acid to acetic acid.

The lack of production of succinic acid by Lact. casei under all the conditions studied confirms observations previously reported. Kaneuchi et al. (1988) reported that a number of Lactobacillus strains produced succinic acid in MRS broth to various extents. However, under the conditions studied by Kaneuchi et al., ten Lact. casei strains did not produce succinic acid. Similarly, Dudley and Steele (2004) reported that although Lactobacillus plantarum strains are able to synthesize succinic acid from citric acid by the reductive TCA cycle, Lact. casei, Lact. zeae, and Lactobacillus rhamnosus lack one or more enzymatic activities present in this pathway, and do not produce succinic acid. Nonetheless, putative genes encoding enzymes involved in a reductive TCA cycle were identified in the Lact. casei ATCC334 genome sequence. Further studies to identify the causes for the lack of succinic acid production in this micro-organism are needed.

This is the first report of metabolic end products by Lact. casei in CCE. Although, CCE represents a more complex environment, it is a more accurate model for the conditions present in ripening Cheddar cheese. Both lactose and galactose were present in limiting concentrations. Although, it naturally contained 2·31 ± 0·13 mmol l−1 of citrate, supplementation with 12·7 ± 0·24 mmol l−1 of diammonium citrate was performed to obtain a final concentration of 15 ± 0·65 mmol l−1 of citric acid. An increased final concentration allowed a more extended analysis of the end products of citric acid catabolism in this particular media. Lactose was the first substrate utilized, and was mainly converted to acetic acid. Galactose and citric acid catabolisms were observed immediately following lactose utilization. Similar to the results obtained in mCDM, the main catabolic end products during galactose and citric acid metabolism were acetic acid, and l-/d-lactic acids.

Acknowledgements

Ilenys Díaz-Muñiz was funded by the Advanced Opportunity Fellowship from the Graduate School at the University of Wisconsin-Madison, and the Robert D. Watkins Minority Graduate Fellowship from the American Society for Microbiology (ASM), Washington, DC. This project was funded by Dairy Management, Inc. through the Center for Dairy Research and the College of Agricultural and Life Sciences at the University of Wisconsin-Madison.

Ancillary