Polyphasic Study of the Spatial Distribution of Microorganisms in ...

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Dec. 1999, p. 5464–5473 0099-2240/99/$04.00⫹0 Copyright © 1999, American Society for Microbiology. All Rights Reserved.

Vol. 65, No. 12

Polyphasic Study of the Spatial Distribution of Microorganisms in Mexican Pozol, a Fermented Maize Dough, Demonstrates the Need for Cultivation-Independent Methods To Investigate Traditional Fermentations ´ DE ´ RIC AMPE,1* NABIL FRE

OMAR,1 CLAIRE MOIZAN,1 CARMEN WACHER,2 JEAN-PIERRE GUYOT1

BEN

AND

Laboratoire de Biotechnologie Microbienne Tropicale, Institut de Recherche pour le De´veloppement, F-34032 Montpellier cedex 1, France,1 and Departamento de Alimentos y Biotecnologia, Facultad de Quimica, Universidad Nacional Autonoma de Me´xico, 04510 Me´xico D.F., Mexico2 Received 12 April 1999/Accepted 9 September 1999

The distribution of microorganisms in pozol balls, a fermented maize dough, was investigated by a polyphasic approach in which we used both culture-dependent and culture-independent methods, including microbial enumeration, fermentation product analysis, quantification of microbial taxa with 16S rRNA-targeted oligonucleotide probes, determination of microbial fingerprints by denaturing gradient gel electrophoresis (DGGE), and 16S ribosomal DNA gene sequencing. Our results demonstrate that DGGE fingerprinting and rRNA quantification should allow workers to precisely and rapidly characterize the microbial assemblage in a spontaneous lactic acid fermented food. Lactic acid bacteria (LAB) accounted for 90 to 97% of the total active microflora; no streptococci were isolated, although members of the genus Streptococcus accounted for 25 to 50% of the microflora. Lactobacillus plantarum and Lactobacillus fermentum, together with members of the genera Leuconostoc and Weissella, were the other dominant organisms. The overall activity was more important at the periphery of a ball, where eucaryotes, enterobacteria, and bacterial exopolysacharide producers developed. Our results also showed that the metabolism of heterofermentative LAB was influenced in situ by the distribution of the LAB in the pozol ball, whereas homolactic fermentation was controlled primarily by sugar limitation. We propose that starch is first degraded by amylases from LAB and that the resulting sugars, together with the lactate produced, allow a secondary flora to develop in the presence of oxygen. Our results strongly suggest that cultivation-independent methods should be used to study traditional fermented foods. However, for a large number of biotopes in which most of the microorganisms are unknown and “unculturable” (soil, rumen, wastewater treatment plants, etc.), a cultivation-dependent approach biases our view of microbial diversity (3, 28). In fact, it is thought that fermented foods contain mainly culturable organisms. Do culture-dependent methods bias our view of the microbial assemblages responsible for fermentation of these foods? To try to address this question, we studied the microflora of pozol by a polyphasic approach in which we used culture-dependent and culture-independent methods. Furthermore, as pozol dough is shaped into balls, it is thought that there are gradients between the inside and the periphery, as observed for bacterial aggregates in water treatment reactors (32). Thus, the spatial distribution of the microflora was investigated in this study. We believe that more information about the microflora present and active during pozol fermentation should help improve the microbiological quality and safety of this food and that the approach which we propose may contribute to new developments in the field of food microbial ecology.

Pozol is a traditional fermented maize dough prepared by Indians and mestizos in southeastern Mexico (Yucatan, Quintana Roo, Campeche, Tabasco, Chiapas, and Oaxaca states) and Guatemala (53). Cobs of white maize are shelled, and the kernels are cooked in the presence of lime and washed to remove the pericarps. The grains are then coarsely ground, shaped into balls, wrapped in banana leaves, and allowed to ferment at ambient temperature for 2 to 7 or more days. The resulting fermented dough is suspended in water and drunk daily as a refreshing beverage. A wide variety of microorganisms have already been isolated from this spontaneous fermentation; these microorganisms include fungi, yeasts, lactic acid bacteria (LAB), and non-lactic acid bacteria (non-LAB) (15, 43, 53, 57). However, little is known about the ecophysiological importance of these organisms. In particular, despite the presence of LAB, enterobacteria have reportedly been found in pozol from various areas (10, 56, 57). In order to study the ecology of this fermented food, it is important to know the structure of the microbial community in it and to identify the physiologically active organisms. Classically, these questions are addressed by enumerating members of some microbial groups in fermented foods by using various culture media, followed by identification of some dominant microorganisms by taxonomic and/or phylogenetic methods (9, 22, 27, 30, 45).

MATERIALS AND METHODS Pozol sampling. Freshly ground pozol shaped into 250-g balls and wrapped in banana leaves was bought at Atasta market, Villahermosa, Tabasco, Mexico, and incubated at 30°C for 5 days. Three concentric samples (periphery, intermediate, and center of the ball) that weighed approximately 50 g were then aseptically diluted 10-fold in sterile 0.9% NaCl and mixed in a Waring blender. Each sample was immediately fractionated and used for further analyses. Subsamples which were used for microbial enumeration, direct microscopic observation, and pH determination were processed immediately. For high-pressure liquid chromatography (HPLC) analyses, 0.2 ml of 2 N H2SO4 was added to each 1.3-ml sub-

* Corresponding author. Mailing address: IRD (ex. Orstom) LBMT, 911 Avenue Agropolis, BP 5045, F-34032 Montpellier cedex 1, France. Phone: 33 4 67 41 62 78. Fax: 33 4 67 41 62 83. E-mail: Frederic.Ampe @mpl.ird.fr. 5464

VOL. 65, 1999

MICROBIOLOGY OF POZOL

5465

TABLE 1. 16S rRNA-targeted oligonucleotide probes and PCR primers used in this study Probe or primer

Univ1390 Eub338R Euc502R Entero Lacb0722

Universal Eubacteria Eucarya Enterobacteria LAB

Strc493

Streptococcus, Lactococcus, some Leuconostoc spp. Lactococcus Lactobacillus, Enterococcus, Pediococcus, Leuconostoc, Weissella Leuconostoc NAd NA NA NA

212RLa Lab158 LU2 27f 907r 338fe 518r

Positiona

Oligonucleotide Probe Database designationb

1,407–1,390 355–338 516–502 1,432–1,418 746–722

S-*-Univ-1390-a-A-18 S-D-Bact-0338-a-A-18 S-D-Euca-0502-a-A-16 not available S-*-Lab-0722-a-A-25

44 54 52 44 54

60 5 5 41 49

511–493

S-*-Strc-0493-a-A-19

50

18

CTT TGA GTG ATG CAA TTG CAT C GGT ATT AGC AYC TGT TTC CA

233–212 177–158

S-S-L.lac-0212-a-A-22 S-G-Lab-0158-a-A-20

46 45

47 58

GAT AGA CCG ACT ATT

242–220 46–27 926–907 357–338 534–518

Not available NA NA NA NA

46 NA NA NA NA

42 31 31 31 40

Target taxon(a)

Sequence (5⬘-3⬘)

GAC GCT ACC CTT CCA CAC GTT

GGG GCC AGA TTG CCG T AGC

CCA GTT TCA CCT ACC

CGG TCC CTT CAR CTA

TGT CGT GCC CCC CAC

GTA AGG CTC ACT ATG

CAA AGT C GAG TTC

CGT CCC TTT CTG G

TCT TGA ATT ACG GCG

CTA TCM CMT GGA GCT

GGT TGG TTR GGC GCT

GAC CTC AGT AGC GG

GCC G AG TT AG

Wash temp (°C)c

Reference

a

E. coli numbering. See reference 4. Wash temperature in 1⫻ SSC–1% SDS. d NA, not applicable. e A GC clamp was attached to the 5⬘ end of primer 338f to obtain primer gc338f (GC clamp, 5⬘CGCCCGCCGCGCGCGGCGGGCGGGGCGGGGGCACGGG GGG) (40). b c

sample in a tube, the tubes were centrifuged for 10 min at 10,000 ⫻ g, and the supernatants were analyzed as described below. Subsamples which were used for RNA and DNA extraction were frozen at ⫺20°C until they were analyzed. SEM. Undiluted samples from the periphery and the center of a pozol ball were fixed with glutaraldehyde and dehydrated with a graded ethanol series as described by Giraud et al. (21). Ethanol was then removed by adding CO2 at the critical point. The samples were coated with gold, and scanning electron microscopy (SEM) was performed with a JEOL model JSM-6300F microscope at Universite´ de Montpellier II. Enumeration and isolation of microorganisms. Serial dilutions of homogenized pozol samples in 0.9% NaCl were used for microbial enumeration with the following media: MRS-glucose medium (Difco) for LAB (14); MRS-starch medium containing 2% soluble starch instead of glucose for amylolytic LAB (19); potato dextrose agar (PDA) (Merck) containing 14 mg of tartaric acid per liter, 50 mg of chloramphenicol per liter, and 50 mg of rose bengal per liter for yeasts and molds; plate count agar (PCA) (Difco) to estimate the number of total aerobic mesophilic bacteria (the number of non-LAB was estimated by determining the number of colonies with diameters larger than 1 mm [56]); MFT medium (5 g of tryptone per liter, 2.5 g of yeast extract per liter, 1 g of dextrose per liter, 15 g of agar per liter) to estimate the size of the total microflora (27); and LFB medium (0.5 g of Trypticase per liter, 1 g of yeast extract per liter, 0.5 g of cysteine-HCl per liter, 0.1 g of sodium acetate per liter, 5 mg of resazurin per liter, 20 ml of a mineral solution per liter, 1 ml of a trace element solution [59] per liter, 5 g of lactate per liter) for lactate-fermenting anaerobes (9). LFB medium was prepared and inoculated by using strictly anaerobic techniques. We also used violet red bile agar medium (Difco) for enterobacteria (56); M17 medium (Difco) for Streptococcus and Lactococcus strains; ENC medium (azidedextrose agar; Sanofi Diagnostics Pasteur, Marnes-la-Coquette, France) for Enterococcus strains; and LEU medium (10 g of tryptone per liter, 5 g of yeast extract per liter, 100 g of sucrose per liter, 1 g of sodium citrate per liter, 5 g of glucose per liter, 2.5 g of gelatin per liter, 15 g of agar per liter) for Leuconostoc exopolysaccharide (EPS)-producing strains (19). For lactate-fermenting anaerobes, three tubes of LFB medium were inoculated anaerobically for each dilution, and most-probable-number values were obtained by using McCrady’s tables (35). All other counts were obtained by the plate count method; 0.1-ml portions of appropriate dilutions were directly inoculated in triplicate onto solid media when PDA, PCA, violet red bile agar, M17 medium, ENC medium, and LEU medium were used, whereas MRS-glucose medium, MRS-starch medium, and MFT medium cultures were prepared by using an overlay consisting of 10 ml of medium containing a 0.1-ml dilution. Counts were obtained after 48 h and 5 days of incubation at 30°C. The results given below are means and standard deviations based on three determinations. The square roots of total colonies (25, 27) were randomly picked from MRSglucose, MRS-starch, LEU, M17, and PDA media. The strains recovered were purified further by streaking them onto the same media, and liquid cultures were stored by using 20% glycerol at ⫺80°C. Strains isolated from MRS-glucose, MRS-starch, LEU, and M17 media were examined by performing Gram stain, catalase, sporulation, and motility tests. Gram-positive, non-spore-forming, cata-

lase-negative strains were considered LAB. Strains isolated from PDA were tested for amylase production as described by Nuraida et al. (43). Analysis of sugars and fermentation products. The concentrations of soluble starch, sugars, ethanol, and organic acids were determined by HPLC by using an Aminex HPX87H column (Bio-Rad, Richmond, Calif.). The following conditions were used: mobile phase, H2SO4 (6 mmol 䡠 liter⫺1); flow rate, 0.8 ml 䡠 min⫺1; and temperature, 65°C. A refractometer (model PU 4026; Philips, Heindoven, The Netherlands) was used for detection. The retention times were 4.75 min for soluble starch, 5.65 min for maltose, 6.9 min for glucose, 9.7 min for lactate, 10.3 min for formate, 11.2 min for acetate, and 16.3 min for ethanol. Preparation of rRNA standards from pure cultures. Several rRNA standards were prepared by extracting RNA from laboratory cultures of the following strains: Escherichia coli JM109, Lactococcus lactis subsp. lactis ATCC 11454T, Lactobacillus plantarum DSM20174T, Leuconostoc mesenteroides ATCC 10832, and Saccharomyces cerevisiae FL100. All of the strains were grown in appropriate rich media, and total RNA was extracted from exponentially grown cells as previously described (6). Isolation of RNA from pozol. Total RNA was extracted from pozol by using a previously described method adapted to samples with high starch contents and optimized for the study of pozol (6). Hybridization probes. The oligonucleotide probes used are shown in Table 1. The temperatures used for the stringent washes are also shown. The specificity of the probes was checked by using the CHECK_PROBE command of a recent release of the Ribosomal Database Project (RDP) (37, 46a). Synthetic HPLCpurified oligonucleotides (Eurogentec, Seraing, Belgium) were 3⬘ end labeled with digoxigenin by following the instructions of the manufacturer (Boehringer Mannheim). rRNA quantitative hybridization. RNA blotting was performed as described by Stahl et al. (50). RNA was denatured by adding 3 volumes of 2% glutaraldehyde immediately before dilution to a concentration of 0.5 to 20 ␮g 䡠 ml⫺1 in water containing 1 ␮g of poly(A) (Sigma) per ml (50). Samples were applied in a total volume of 100 ␮l to Hybond N⫹ nylon membrane filters (Amersham, Les Ulis, France) by using a slot blot device (model PR648; Hoeffer Scientific Instruments, San Francisco, Calif.) under a slight vacuum. The membranes were air dried and baked for 30 min at 80°C before hybridization. The baked membranes were prehybridized in 20 ml of hybridization buffer [0.9 mol of NaCl per liter, 50 mmol of NaPO4 per liter, 5 mmol of EDTA per liter, 10⫻ Denhardt solution (48), 0.5 mg of poly(A) per ml (pH 7)] for 2 h at 40°C. Hybridization was performed overnight at 40°C with 10 ml of hybridization buffer containing the labeled probe. The filters were washed twice in 100 ml of 1⫻ SSC–1% sodium dodecyl sulfate (SDS) (1⫻ SSC is 0.15 mol of NaCl per liter plus 0.015 mol of sodium citrate per liter) at the temperatures indicated in Table 1 for 30 min. Chemiluminescence was then detected as recommended by the manufacturer (Boehringer Mannheim). Bound probe was quantified by densitometry relative to reference standards after autoradiography. The hybridization controls used are listed above. The RNA contents of controls were estimated by hybridization with universal probe S-*-Univ-1390-a-A-18 (60) before the controls were used as internal standards; E. coli RNA (Boehringer Mannheim) was the absolute stan-

5466

AMPE ET AL.

dard used. The amounts of microorganisms are expressed below as fractions of the total rRNA in samples (i.e., RNA indices). The lower limit for detecting a unique small-subunit (SSU) rRNA in the 2 ␮g of nucleic acid spotted onto a membrane was approximately 5 ng of SSU-like rRNA. Preparation of DNA from pure cultures. Two-milliliter samples of overnight cultures were centrifuged at 7,000 ⫻ g for 10 min. The cell pellets were resuspended in 200-␮l portions of a solution of lysozyme (20 ␮g 䡠 ␮l⫺1) in TES buffer (50 mmol of Tris [pH 8] per liter, 1 mmol of EDTA per liter, 8.56% [wt/vol] saccharose) containing 10 ␮l of a mutanolysin solution (1 U 䡠 ␮l⫺1) and were incubated for 1 h at 37°C. Fifty microliters of 20% SDS was then added to each preparation to lyse the cells. Total DNA was then purified by repeated extraction with phenol-chloroform-isoamyl alcohol (25:24:1), followed by final extraction with chloroform-isoamyl alcohol (24:1). The DNA was precipitated with isopropanol, washed with 70% ethanol, and vacuum dried. The resulting pellets were resuspended in 200-␮l portions of a solution containing 50 ␮g of RNase per ml and incubated for 15 min at 37°C. The quality of the extracts was checked on 1% agarose gels, and DNA was quantified by spectrophotometry. DNA isolation from pozol. Ten milliliters of pozol resuspended in 0.9% NaCl (10-fold dilution) was homogenized for 30 s at the maximum speed with an Ultraturrax T25 apparatus (Janke & Kunkel, IKA Labortechnik). Two 1.5-ml tubes containing the resulting suspension were then centrifuged at 7,000 ⫻ g for 10 min, and 500 ␮l of a lysozyme solution (20 ␮g 䡠 ␮l⫺1) in TES buffer, and 10 ␮l of a mutanolysin solution (1 U 䡠 ␮l⫺1) were added to each pellet. Samples were vortexed for 1 min and incubated for 1 h at 37°C. Fifty microliters of a proteinase K solution (10 mg 䡠 ml⫺1) was added to each tube, and the tubes were incubated for 50 min at 50°C and then for 10 min at 65°C. Then 300 ␮l of warm (65°C) buffer (0.2 M NaCl, O.1 M Tris-Hcl [pH 8], 2% SDS) was added to each tube, and the tubes were incubated for 10 min at 65°C. Three hundred microliters of 5 M NaCl was added to each tube, and the tubes were gently mixed for 30 s, incubated at 4°C for 10 min, and centrifuged at 7,000 ⫻ g and 4°C for 10 min. Each supernatant was divided and placed into two tubes, and the DNA in each tube was precipitated with 780 ␮l of isopropanol by incubating the tube at ⫺20°C for 30 min. Pellets were recovered by centrifugation at 12,000 ⫻ g and 4°C for 15 min, washed with 1 ml of 70% ethanol, vacuum dried, and resuspended in 100 ␮l of water. The contents of the tubes representing the same sample were mixed, 700 ␮l of water was added, and then 800 ␮l of phenol (pH 8) was added. The contents of the tubes were mixed for 3 min and centrifuged at 12,000 ⫻ g and room temperature for 10 min. The aqueous phase was extracted once with phenol and two or three times with phenol-chloroform-isoamyl alcohol (25:24:1) (pH 8) before the final extraction with chloroform-isoamyl alcohol (24:1). The aqueous phase was then precipitated with isopropanol, and the pellets were washed with 70% ethanol, vacuum dried, and resuspended in 200 ␮l water. The quality of the extracts was routinely checked by using 1% agarose-0.5⫻ TBE gels. PCR-DGGE analysis. Purified DNA was amplified with primers gc338f and 518r spanning the V3 region of the 16S ribosomal DNA (rDNA) (Table 1) (44). Amplification was performed with a Perkin-Elmer model 9400 thermal cycler. Each mixture (final volume, 25 ␮l) contained 1 ␮l of template DNA, each primer at a concentration of 0.25 ␮M, each deoxynucleoside triphosphate at a concentration of 200 ␮M, 1.5 mM MgCl2, 2.5 ␮l of 10⫻ PCR buffer, and 1.25 U of Taq polymerase (Eurogentec). Template DNA was denatured for 5 min at 94°C. To increase the specificity of amplification and to reduce the formation of spurious by-products, a “touchdown” PCR was performed (40). The initial annealing temperature used was 10°C above the expected annealing temperature (65°C), and the temperature was decreased by 1°C every second cycle until the touchdown temperature, 55°C, was reached; then 10 additional cycles were carried out at 55°C. Primer extension was carried out at 72°C for 3 min. The tubes were then incubated for 10 min at 72°C (final extension). Aliquots (2 ␮l) of the amplification products were analyzed first by electrophoresis in agarose gels. The PCR products were then analyzed by denaturing gradient gel electrophoresis (DGGE) by using a Bio-Rad DCode apparatus and the procedure first described by Muyzer et al. (40). Samples were applied to 8% (wt/vol) polyacrylamide gels in 1⫻ TAE. The optimal denaturation gradient was determined by electrophoresing a perpendicular gel (40). Optimal separation was achieved with a 30 to 60% urea-formamide denaturing gradient (100% corresponded to 7 M urea and 40% [vol/vol] formamide) (data not shown). All parallel electrophoresis experiments were then performed at 60°C by using gels containing a 30 to 60% urea-formamide gradient increasing in the direction of electrophoresis. The gels were electrophoresed for 10 min at 20 V and for 3 h at 200 V, stained with ethidium bromide for 10 to 15 min, and rinsed for 20 to 30 min in distilled water. Sequencing of DGGE fragments and 16S rDNA from pure strains. DGGE fragments were cut out with a sterile scalpel. The DNA of each fragment was eluted in 20 ␮l of sterile water overnight at 4°C. One microliter of the eluted DNA from each DGGE band was reamplified by using the conditions described above. The success of this procedure was checked by electrophoresing 3-␮l portions of the PCR products in DGGE gels as described above; pozol amplified DNA was used as the control. PCR products which yielded single bands which comigrated with an original band were then purified and sequenced. For pure strains, 1 ␮l of total DNA was amplified with primers 27f and 907r (31) by using conditions described elsewhere (38). Sequences of the gene fragments were determined by the dideoxy chain termination method with an ABI PRISM dye terminator kit (Perkin-Elmer). Sequence products were loaded into and analyzed with a model 373 DNA se-

APPL. ENVIRON. MICROBIOL. quencer (Applied Biosystems). DGGE fragments were sequenced by using primer gc338f, and pure strain 16S rDNA were sequenced with primer 27f. Analysis of the sequence data. To determine the closest known relatives of the partial 16S rDNA sequences obtained, searches of public data libraries (RDP and GenBank) were performed by using the BLAST and RDP programs (37). The CHECK_CHIMERA command of the RDP facilities was used to try to detect chimeric sequences. Nucleotide sequence accession numbers. The GenBank accession numbers for the partial 16S rDNA sequences of strains MRS-III06, MRS-II22, MRS-I08, MRS-II19, MRS-I09, and LEU-I02 reported in this paper are AF138777 through AF138782. The accession numbers for the sequences of DGGE bands 1 to 6 are AF138783 through AF138788.

RESULTS SEM. Phase-contrast and SEM observations revealed that morphologically the organisms present at the periphery of a single ball (masa) of pozol were very diverse, whereas the diversity was somewhat lower in the center of the ball (Fig. 1). Bacteria of various shapes and sizes, as well as cylindrical yeasts (Geotrichum-like organisms) and filamentous fungi, were present at the periphery of the ball. Starch granules, which represented the larger fraction of the substratum, appeared to be unevenly degraded, suggesting that microenvironments were important. Enumeration and isolation of microorganisms. The total microflora and specific groups of organisms were enumerated by using 10 different culture media (Table 2). PCA without an overlay was used to estimate the total microflora concentration because MFT medium, the medium used by Hounhouigan et al. (27), gave counts that were lower than the counts obtained with specific media, such as MRS-glucose medium. The concentration of the total microflora was high (9 to 10 log CFU 䡠 g [dry weight]⫺1), and it was four to five times higher at the periphery of the ball than it was inside the ball. The LAB were an important part of the flora, as illustrated by the high counts on MRS-glucose medium. Among the LAB, the EPS producers were a major group at the periphery (the counts were similar to the counts obtained with MRS-glucose medium), but the counts were lower in the center of the ball. Starch degraders apparently also were an important part of the LAB. The counts on M17 medium were also high and close to the counts obtained with MRS-glucose medium. By contrast, no Enterococcus strain was detected in either sample (the counts obtained with ENC medium were always less than 4 log CFU 䡠 g [dry weight]⫺1). The counts of lactate-fermenting strict anaerobes were also less than 4 log CFU 䡠 g [dry weight]⫺1. At the periphery of the ball, an important population of yeasts and filamentous fungi was present, and the numbers of these organisms decreased significantly to undetectable levels at the center. The same observation was made for enterobacteria, which represented up to 1% of the total flora at the periphery but less than 0.001% in the center. Rapid estimates of the catalase activities revealed that a high proportion of the colonies detected on M17 medium were not LAB colonies, which made estimation of Streptococcus and Lactococcus counts with this medium impossible. A few catalase-positive colonies were also detected on MRS-starch medium plates. The square roots of total colonies on MRS-glucose medium, MRS-starch medium, M17 medium, LEU medium, and PDA plates containing 30 to 300 colonies were randomly picked and restreaked onto the same media until pure strains were obtained. The bacteria were examined by performing Gram stain, catalase production, sporulation, and motility tests. The shapes and associations of the resulting 136 LAB (gram positive, cata-

VOL. 65, 1999


5467

FIG. 1. SEM micrographs of a 5-day fermented pozol sample. (A through C) Periphery of the ball. (D through F) Center of the ball. (A) Geotrichum-like yeasts and bacteria. (B) Filamentous fungi. (C) Consortium of bacteria on a degraded starch granule. (D and E) Bacteria on cell debris. (F) Homogeneous rods on a starch granule.

lase negative, non-spore-forming, nonmotile organisms) were variable (rods and lenticular cells occurred alone, in pairs, or in chains). Twenty-five percent of the strains isolated with M17 medium were catalase positive and were not used for further studies. Yeasts and fungi were tested to determine whether

they produced amylase and an aroma. None of the strains isolated appeared to produce extracellular amylase, but 70% of the strains produced a fruity aroma like that of fresh pozol. All of the strains isolated were also used in comparative DGGE analyses (see below).

5468

AMPE ET AL.

APPL. ENVIRON. MICROBIOL.

TABLE 2. Distribution of microorganisms in a pozol ball as estimated by plate enumeration or most-probable-number counting with different culture media No. of cells (log CFU 䡠 g [dry wt]⫺1) at theb: Mediuma

Organisms targeted

PCA (all colonies) PCA (large colonies) MFT MRS-glucose M17 MRS-starch LEU VRBGc ENC LFB PDA PDA

Total flora Non-LAB Total flora LAB Streptococcus, Lactococcus Amylolytic LAB Leuconostoc Enterobacteria Enterococcus Lactate-fermenting strict anaerobes Yeasts Fungi

Periphery

Intermediate position

Center

9.79 (8.78) 8.47 (7.18) 8.87 (8.09) 8.82 (7.75) 8.63 (7.77) 8.69 (8.28) 8.81 (8.05) 7.80 (7.28) ⬍4 ⬍4 7.82 (7.14) 6.85 (6.78)

9.15 (8.02) 5.73 (4.78) 7.89 (7.24) 8.99 (7.72) 7.78 (7.01) 8.71 (8.18) 6.45 (6.02) 5.91 (5.01) ⬍4 ⬍4 6.08 (5.24) 5.34 (5.00)

9.14 (8.31) 5.72 (4.75) 7.87 (7.15) 8.89 (7.49) 8.32 (7.32) 8.72 (8.18) 5.12 (5.18) 3.88 (3.85) ⬍4 ⬍4 ⬍4 ⬍4

a

For an explanation of the media see the text. The data are means based on three replicates for a single pozol ball. The values in parentheses are standard deviations. c VRBG, violet red bile agar. b

pH, sugars, and fermentation products. Low pH values were found, and the acidity increased towards the inside of the pozol ball (from pH 3.87 to 3.63) (Table 3). High levels of soluble starch were found, but only low concentrations of mono- or disaccharides were detected in the samples; only glucose and maltose were found (there was no fructose or sucrose), and significant concentrations were present mainly at the periphery of the ball. Conversely, high levels of fermentation products were observed. Lactate was the major fermentation product, and the lactate concentrations ranged from 158 to 175 ␮mol 䡠 g (dry weight)⫺1 (periphery to center); high concentrations of acetate, ethanol, and formate were also detected. Higher concentrations of acetate were found at the periphery, whereas the ethanol concentrations were higher in the center of the ball. Finally, no butyrate, propionate, or other volatile fatty acid was detected. Quantification of 16S rRNA. Total RNA was extracted directly from the same pozol samples and was quantified by using a strategy previously optimized for the study of starchy foods, including pozol (6, 7). Twofold more RNA was recovered at the periphery than at the center. rRNA indices for microbial taxa were then determined by hybridization with 16S-rRNA targeted oligonucleotide probes (Table 4). The data show that

TABLE 3. pH values and sugar, ethanol, and organic acid contents of pozol Amt at thea:

Substrate or product

Periphery


Center

Soluble starch Lactate Acetate Formate Ethanol Glucose Maltose

25.3 (0.6) 158.4 (10.6) 30.4 (4.5) 22.3 (5.2) 22.5 (9.5) 7.7 (0.6) 1.4 (0.1)

24.1 (0.3) 164.8 (4.0) 15.6 (4.5) 29.9 (3.9) 30.2 (5.4) 1.8 (0.7) 0.2

20.2 (0.6) 174.7 (9.4) 12.2 (3.0) 7.5 (0.9) 36.6 (10.3) 1.8 (0.2) 0.0

a For all of the compounds except soluble starch the data are expressed in micromoles per gram (dry weight). The data for soluble starch are expressed in milligrams per gram (dry weight). The data are means based on three replicates, and the values in parentheses are standard deviations. The pH values at the periphery, the intermediate position, and the center were 3.87, 3.71, and 3.63, respectively.

the sum of the domain signals (eubacteria plus eucarya) was within the 100% ⫾ 25% range with respect to the signal obtained with the universal probe. Although this result was expected, it provided a necessary control for the experimental validity of the technique used (7, 33). The rRNA indices obtained with the Lacb0722 probe targeting all LAB showed that the LAB accounted for the vast majority of the active microorganisms in pozol, especially inside the ball, where they accounted for more than 94% of the active microflora (Table 4). Group- and genus-specific probes were then used to determine the microbes in the LAB assemblage. Quantification with probe LU2 showed that members of the genus Leuconostoc accounted for almost 7% of the active flora at the periphery and 3 to 4% inside the ball. Conversely, almost no members of the genus Lactococcus (probe 212RLa) were detected at the periphery, but members of this genus accounted for around 1.5% of the flora inside the ball. Two other probes that targeted subgroups of LAB were also used. Probe Lab158 targets the genera Lactobacillus, Enterococcus, Pediococcus, Leuconostoc, and Weissella (6, 58), and probe Strc493 targets the genera Streptococcus and Lactococcus plus some members of the genera Leuconostoc and Weissella (18). The rRNA indices obtained with probe Lab158 showed that the former group accounted for around 60% of the total active flora. As (i) Enterococcus strains were not found by the plate counts method, (ii) members of the genus Leuconostoc accounted for less than 7% of the active population, and (iii) no member of the genus Pediococcus (perfectly round cocci that occur alone or in tetrads) was observed in either sample, members of the genera Lactobacillus and Weissella probably accounted for more than 50% of the total active flora. Members of the genera targeted by probe Strc493 accounted for around 25 and 50% of the total active flora at the periphery and in the center, respectively. As the rRNA indices obtained for members of the genera Lactococcus and Leuconostoc were relatively low, the high values probably meant that the proportion of members of the genus Streptococcus in pozol was high. In addition, the sum of the signals obtained with the two LAB subgroup probes (Lab158 and Strc493) minus the signal obtained for Leuconostoc spp. (probe LU2) was within the 100% ⫾ 25% range with respect to the signal obtained with probe Lacb0722 targeting all LAB, which validated the numerical data. Two other non-LAB microbial groups were quantified. Eu-

VOL. 65, 1999


5469

TABLE 4. Percentages of the total SSU rRNA in pozol samples as determined with individual probes % of total SSU rRNA at thea: Probe(s)

Univ1390 Eub338 Euc502 Eub338 ⫹ Euc502 Entero Lacb0722 Strc493 Lab158 LU2 212RLa Lab158 ⫹ Strc493 ⫺ LU2

Target taxon(a)

Universal Eubacteria Eucarya Control for universal probe Enterobacteria LAB Streptococcus, Lactococcus, some Leuconostoc spp. Lactobacillus, Enterococcus, Pediococcus, Leuconostoc, Weisella Leuconostoc Lactococcus Control for all LAB probes

Periphery


Center

100 88.9 6.09 95.0 0.85 88.6

100 75.4 4.65 80.1 NDb 97.4

100 98.2 2.25 100.5 ND 94.3

26.9

26.8

54.2

68.1 6.82 0.27 89.2

65.20 3.84 0.74 88.2

59.5 4.08 1.47 99.6

a The data are expressed as percentages of the signal obtained with universal probe Univ1390 and are means based on three replicates. The amounts of total RNA recovered at the periphery, the intermediate position, and the center were 29, 17.7, and 13.3 ␮g/g of pozol, respectively. b ND, not detected.

caryotes (yeasts and fungi) accounted for 6% of the active population at the periphery, and the rRNA index was lower in the center. Finally, although enterobacteria accounted for less than 1% of the active population, they were detected at the periphery but not in the center of the ball. Community fingerprinting. Total DNA was extracted from each pozol sample three times independently. One-microliter portions of 10⫺2 and 10⫺1 dilutions and undiluted total DNA were used in amplification reactions, and the equal-size 16S rDNA PCR products were analyzed by DGGE. Repeated extractions, as well as different dilutions, of an individual sample resulted in similar fingerprints (data not shown). The fingerprints obtained with the three concentric pozol samples contained nine sharp bands and a few fainter bands (Fig. 2). The three different profiles contained the same intense bands (bands 5 through 7, 9, and 10), but the relative intensities of several bands (e.g., band 7) varied between the samples. Band 4 was present at the periphery and not in the center. Conversely, band 2 was detected in the center and not at the periphery. We then examined the melting behavior of the PCR products obtained with DNA extracted from pure strains isolated from pozol (see above) used as templates; we used the same conditions that we used for the pozol DNA analysis. The great majority of the strains tested produced a single DGGE band with a melting position identical to that of one of the bands identified in the pozol DNA fingerprint (Table 5). Eighty-two strains comigrated with band 10, 25 strains comigrated with band 9, and 17 strains comigrated with band 7. Seven strains did not comigrate with any of the bands identified, and five strains produced patterns with more than one band. None of the strains tested corresponded to band 1, 2, 3, 4, 5, 6, 8, or 11, and only three bands could be attributed to strains that were isolated. This was particularly striking in the case of band 6; no strain that was isolated corresponded to this band even though it was the most intense band in the three pozol DGGE profiles (Fig. 2). Strains corresponding to band 7 were isolated only at the periphery of the ball, and most of these strains were isolated with LEU medium. Conversely, strains corresponding to band 9 were found throughout the ball when MRS-glucose medium was used, at the periphery when MRS-starch and M17 media were used, and never when LEU medium was used. Finally, strains corresponding to band 10 were found throughout the pozol ball when MRS-glucose, MRS-starch, and LEU

media were used. The latter organisms were by far the most numerous organisms (60% of the strains isolated). Although primer gc338f, which was used here for DGGE analysis, is normally specific for eubacteria and therefore eu-

FIG. 2. DGGE analysis of PCR-amplified 16S rDNA fragments from pozol bacterial communities. DNA was derived from three concentric fractions obtained from the same 5-day fermented ball of pozol. Lane A, center; lane B, intermediate; lane C, periphery. The bands are discussed in the text, and the positions of the bands are indicated on the left and on the right.

5470

AMPE ET AL.

APPL. ENVIRON. MICROBIOL. TABLE 5. Correspondence between strains isolated from pozol and DGGE bands

Isolation medium

MRS-glucose MRS-starch LEU M17 Total for all media Total for all media and all samples

Location

Periphery Intermediate Center Periphery Intermediate Center Periphery Intermediate Periphery Periphery Intermediate Center

No. of strains that comigrated with the following pozol DGGE bands and positionsa: Band 10

Band 9

Band 7

13 25 14 3 15 2 5 5

4 4 10 2

2

21 45 16

5 11 4 10

82

25

Bands 9 and 10

Position a

Position b

Position c

4 1

12 3

1

14 3

5

1

17

5

1

2 2

4 4

2

4

Total

19 33 24 5 16 2 17 9 11 52 58 26 136

a

None of the strains tested produced bands that comigrated with pozol bands 1 through 6, 8, and 11. Positions a, b, and c did not correspond to pozol DGGE bands. Position a was between bands 6 and 7; position b was between bands 4 and 5; and position c was also between bands 4 and 5.

caryal rDNA sequences should not be amplified, we tried to amplify DNA extracted from yeasts and molds isolated from the periphery of the pozol ball by using the conditions that we used for pozol total DNA. Only 2 of the 18 eucaryote strains tested gave amplification products, and these products were the expected size. When these PCR products were subjected to DGGE analysis, they had melting behaviors different from those of pozol bands 1 to 11. The bands in the DGGE profiles of pozol total DNA were excised from the acrylamide gel and reamplified with primers gc338f and 518r (Fig. 2 shows the original gel from which the bands were excised). Before sequencing, each PCR-amplified DGGE band was electrophoresed on a denaturing gradient gel to confirm its position relative to its position in the original pozol sample gel. In addition, the 16S rDNA genes of two representative strains corresponding to bands 7, 9, and 10 were partially sequenced. All of the sequences obtained corresponded to portions of 16S rDNA genes. Band 1 was identified as an Acetobacter sp. sequence (100% homology with Acetobacter aceti); band 2 corresponded to the sequence of a member of the Bacillus subtilis group (⬎99% homology with an environmental clone from a benzene-mineralizing consortium and Exiguobacterium aurantiacum); and band 3 corresponded to the sequence of members of the Lactobacillus casei subgroup (⬎99% homology with L. casei, Lactobacillus paracasei, and Lactobacillus zeae). Band 4 could not be affiliated with members of a known genus, although the closest relatives found were members of the highG⫹C-content gram-positive bacteria. Band 5 exhibited 90% homology with known Zea mays chloroplast 16S rDNA sequences. The band 6 sequence was identical (⬎99% homology) to the sequence reported for members of the Streptococcus bovis subgroup (including S. bovis, Streptococcus equinus, Streptococcus intestinalis, and Streptococcus macedonius). Partial 16S rDNA sequences (length around 700 bp starting at position 338) of six strains corresponding to bands 7, 9, and 10 resulted in good identification of the strains. Strains LEU-I02 and MRS-I09 corresponding to band 7 were identified as members of the Weissella-Leuconostoc group. MRS-I09 exhibited 97.7% homology with Weissella paramesenteroides, and LEUI02 exhibited 97.7% homology with Leuconostoc mesenteroides. Strains MRS-I08 and MRS-II19, corresponding to band 9, were close relatives of Lactobacillus fermentum (97.5 and 97.0% homology, respectively). Strains MRS-III06 and MRS-

II22, corresponding to band 10, were close relatives of the Lactobacillus plantarum-Lactobacillus pentosus group (97.7 and 97.1% homology, respectively). None of the sequences determined was found to have a chimeric nature. We were not able to purify bands 8 and 11. DISCUSSION The spatial distribution of the microflora in a pozol ball was investigated by using polyphasic microbial ecology. We hoped that this approach would minimize the biases inherent in both culture-dependent and culture-independent methods. Quantification of the main taxa with 16S rRNA-targeted probes and the community fingerprinting by DGGE analysis independently highlighted the limitations of using culture media to study the ecology of fermented foods. rRNA quantification with 16S rRNA-targeted oligonucleotide probes clearly demonstrated that LAB were the dominant microorganisms in all of the samples, accounting for between 89 and 97% of total rRNA. However, although enumeration with MRS-glucose medium revealed the importance of the LAB, the proportions of LAB obtained with this medium were much smaller, especially at the periphery, where the counts on this medium represented only 10% of the total mesophilic counts. If we considered only the results obtained with PCA, it appeared that non-LAB (the large colonies) accounted for around 5% of total flora, a result which is in agreement with rRNA quantification results (percentage of non-LAB ⫽ percentage obtained with Univ1390 ⫺ percentage obtained with Eu502 ⫺ percentage obtained with Lacb0722 ⫽ 100% ⫺ 6.09% ⫺ 88.6% ⫽ 5.3%). Besides, the very high concentrations of lactate observed (158 to 175 ␮mol 䡠 g [dry weight]⫺1), which were similar to the concentrations observed for other typical lactic acid fermentations, such as cassava retting (110 to 220 ␮mol 䡠 g [dry weight]⫺1) (9), also indicate that the LAB were the dominant organisms. Furthermore, although DGGE and rRNA analyses clearly demonstrated that members of the genus Streptococcus (and probably strains closely related to S. bovis) accounted for between 25 and 50% of the total active population in pozol, no Streptococcus strain was isolated with the culture media used, including MRS medium. It should also be noted that in none of the previous studies of the microbiology of pozol has a Streptococcus strain been found. Our DGGE analysis also revealed that a Lactobacillus strain (a

VOL. 65, 1999

close relative of L. casei) was present; this strain was not one of the strains isolated. Such a discrepancy between the results of the cultivation-dependent and cultivation-independent approaches can clearly be attributed to the fact that elective culture media, such as MRS-glucose, were used. MRS-glucose medium was first developed for dairy LAB and has been (and still is) extensively used to study the ecology of lactic acid fermented plant material (8, 9, 23, 24, 27, 30, 43, 45), but our results indicate that it may be not suitable for enumerating LAB in vegetal environments. In particular, this medium is probably not suitable for growing the Streptococcus strains in pozol. This statement concerning MRS-glucose medium also applies to the other media tested since all of the dominant LAB were not isolated with any of them. The bias introduced by cultivation of microorganisms has already been demonstrated for environments whose microorganisms are unculturable so far. Our results demonstrate that ecological studies of well-known and culturable organisms may suffer from the same bias. This of course raises the question of which one culture medium (or several culture media) should be used to study the ecology of fermented foods. A rational approach may be to test the media from which microorganisms are to be isolated by performing a DGGE analysis of the overall growth in various enrichment cultures and to compare the profiles obtained with the total-community fingerprint (by using a strategy resembling the strategy recently proposed by Jackson et al. [29]). An appropriate combination of media should then reflect the diversity found in the original sample. In any case, if an isolation strategy is going to be the sole approach used to study the ecology of a fermented food, it is obvious that a great number of strains (certainly more than 136 strains) have to be isolated and characterized; this would make the study labor-intensive and time-consuming, especially if the number of samples is high. However, despite the advantages of the alternative cultivation-independent methods, these methods also have inherent biases (3, 51, 55). Biases may be introduced by (i) selective extraction of nucleic acids, (ii) selective amplification of 16S rDNA, and (iii) comigration of bands of different sequences in a DGGE analysis. First, we have shown previously that the total-RNA extraction procedure is not selective for a group of organisms (6). The results obtained here (for example, for enterobacteria the rRNA quantification and plate counts results are in good agreement) confirmed that extraction of nucleic acids was not selective for a group of organisms (in particular, gram-positive or gram-negative bacteria). Second, PCR-DGGE analysis yielded results which agreed well with the results of rRNA quantification (that did not include an amplification step), particularly as far as the dominant organisms were concerned. In addition, the same gradients of populations were observed when we compared DGGE band intensities and plate counts (e.g., band 7, which was much more intense at the periphery, corresponded to organisms isolated mainly at the periphery). This suggests that the PCR step prior to DGGE analysis did not distort the community fingerprints. However, our study of the microbial assemblage was limited to bacteria, and a detailed study of yeasts and fungi may still be necessary. Third, it has been shown that different species may yield PCR products which comigrate in DGGE gels (39). Here, strains LEU-I02 and MRS-I09, although they belong to different species, had the same melting position. Both strains belong to the same group (the Leuconostoc-Weissella group), but a more detailed study of the microbial assemblage may be needed and could be performed by investigating other regions of the 16S rDNA gene (26). In addition to members of the genus Streptococcus, the lac-


5471

tobacilli were the other dominant microorganisms in pozol (they accounted for around 50% of the active population). DGGE analysis combined with strain isolation and identification by 16S rDNA sequencing revealed that the following two species were important: L. plantarum and L. fermentum. The former organism has been found in a great variety of vegetal fermented products, where its ability to maintain a pH gradient in the presence of high organic acid concentrations and its low growth-limiting internal pH enable it to terminate the plant fermentations (9, 36). L. plantarum and L. pentosus accounted for more than one-third of 92 LAB isolated from pozol by Escalante et al. (15), which confirmed that these organisms play a role in pozol fermentation. In our study, these species accounted for 60% of the LAB isolated, although this proportion probably overestimates the contribution of these organisms, as previously discussed. So far, L. fermentum has received less interest, although it seems that this heterofermentative species is found in a large number of fermentations of starchy products (2, 9, 22, 23, 27). The other LAB in pozol were lactococci and members of the Leuconostoc-Weissella group. The latter organisms generally develop in the early stages of vegetal fermentations, where they contribute to acidification. This is probably why they accounted for only a minor fraction of the total LAB in the 5-day fermented product studied here. The role of lactococci in the early stage of pozol fermentation remains to be documented, but several workers have reported that Lactococcus strains, including Lactococcus lactis and Lactococcus raffinolactis, are present in this food (15, 43). The other objective of this work was to study the spatial distribution of microorganisms in pozol, and both molecular and classical approaches demonstrated that there are gradients of microbial populations and activities in this food. At the periphery of a pozol ball, where oxygen should not have been limiting, the overall population and metabolic activity were greater than in the center, as demonstrated by both plate count and rRNA quantification methods. In the outer part of the ball, yeasts, fungi, EPS producers (including members of the genus Leuconostoc), and enterobacteria, as well as other nonLAB, such as members of the genus Acetobacter, developed, whereas these organisms were poorly represented inside the ball. However, the very similar DGGE profiles, at least with respect to the most intense bands, of the different samples strongly suggest that the same species were the dominant organisms in all samples, although they were present in different relative proportions, and that the apparently higher diversity observed at the periphery as revealed by SEM or microbial counting probably was due to minor groups of organisms in terms of cell number and overall activity, except for yeasts and fungi. A recent study in which multiple quantitative reverse transcription-PCR and temperature gradient gel electrophoresis were used with soil samples showed that visible bands in gradient electrophoresis fingerprints may account for 50% of the total microflora (16), while the other half includes minor species, each of which accounts for less than 1%. In the case of pozol, this means that the species present only at the periphery are not major organisms. rRNA quantification and enumeration of EPS producers both demonstrated that members of the genus Leuconostoc were predominant organisms at the periphery of the ball. The presence of a higher number of Leuconostoc cells at the periphery may be attributed to the slightly higher pH. In time course fermentations of vegetal products, members of the genus Leuconostoc generally develop before more acid-tolerant lactobacilli develop; for example, this occurs in fermentations

5472

AMPE ET AL.

of cassava, sauerkraut, and cucumber (9, 13, 46). As the pH at the periphery of the ball was higher, these organisms were probably not totally outcompeted by lactobacilli, as was the case in the center of the pozol ball. Another explanation may be related to the ability to produce EPS. This characteristic, although not specific to the genus Leuconostoc alone, is found in many Leuconostoc strains (11) and in a great number of the strains isolated in this study. It may give an ecological advantage to cells confronted by dehydration, but it should also increase the moisture retention in the food, as found in low-fat mozzarella (34). Another advantage of EPS production may be that EPS anchors organisms to solid substrates, such as starch granules or pectocellulosic debris of maize cells. We have not measured in situ production of EPS, but formation of ropy slime on the surfaces of pozol balls in the case of longer fermentations has been reported to be common, although undesirable to consumers, in Chiapas (10). Both homofermentative and heterofermentative LAB were present and active in pozol. The activity of the latter organisms (mainly L. fermentum and Weissella and Leuconostoc spp.) was revealed by the large amounts of acetate and ethanol present (even though members of the genus Acetobacter and yeasts may also produce these compounds). The double-inverted gradient of the concentrations of these products was probably generated by the oxygen gradients; in the absence of oxygen (in the center), heterofermentative LAB produced lactate and ethanol, whereas in the presence of oxygen, these organisms regenerated some of the cofactors through the action of an NADH oxidase, which allowed them to obtain energy through the production of acetate (1). This phenomenon has been observed previously under laboratory conditions, and our data provide evidence that it is significant in the environment. As the acetate/lactate ratio was shown to play an important role in the taste of nondairy fermented products, strict control of aeration may allow workers to control this ratio (1). Homofermentative LAB produce formate when the sugar source is limiting, either in the presence or in the absence of oxygen (52); therefore, the activity of these organisms in pozol was revealed by the presence of significant amounts of formate together with very low concentrations of free sugars throughout the ball. The low concentrations of free sugars also raise the question of the sugar and carbon sources responsible for the high levels of microbial growth observed. Starch is the main sugar source in maize, but is it an important substrate? Several observations support this hypothesis. First, large amounts of soluble starch are available; this substrate is more readily degraded than raw starch. Second, SEM showed that starch granules are degraded, even though the microbes responsible for the degradation remain to be determined. Third, a great number of amylolytic LAB were present, as shown by the MRS-starch medium counts. Thus, a large portion of the LAB isolated from pozol seems to be amylolytic (the counts obtained on MRSstarch medium were similar to the counts obtained on MRSglucose medium). In the past few years, amylolytic LAB belonging to different species have been isolated from nondairy products. These organisms include L. plantarum (20) and S. bovis (12). Recent work (2) has shown that L. fermentum strains isolated from Beninese maize sourdough also exhibit high levels of amylolytic activity, and several of the L. fermentum strains isolated in this study were also found to efficiently degrade soluble starch (data not shown). By contrast, none of the yeast and fungi tested were able to hydrolyze this polysaccharide. All of these results strongly suggest that amylolytic LAB play an important role in making the carbon in starch available. However, direct evidence of this amylolytic activity

APPL. ENVIRON. MICROBIOL.

in situ is still needed, and so far our efforts to measure in situ amylase activity in pozol suspensions have been unsuccessful (data not shown). Previous work (43) has shown that yeasts isolated from pozol are capable of using lactate as a sole carbon source. This consumption of lactate explains why the lactate concentration was lower (and the pH was higher) at the periphery, where high numbers of yeasts and fungi were found by both plate count and rRNA quantification methods, whereas the absolute LAB activity was higher than in the center. During cassava fermentation, clostridia play an important role (9). In the case of pozol, both the low numbers of lactate-fermenting bacteria and the absence of typical volatile fatty acids, such as propionate and butyrate, demonstrate that strict anaerobes, such as clostridia, do not play a key role in the overall process. Both the plate count and rRNA quantification methods showed that enterobacteria accounted for up to 1% of the active flora at the periphery of a pozol ball, but these organisms were hardly detected in the center. The presence of a large population of enterobacteria despite the high lactate concentration may have been due to the presence of microenvironments or to the resistance of some strains, as demonstrated by Wacher-Rodarte (57); this important problem should be addressed since the control pathogens in traditional foods are a major concern (17). Conclusion. In this work we demonstrated that the ecology of fermented foods cannot be effectively studied by cultivationdependent methods alone, as some of the dominating taxa, although culturable, are not recovered by this traditional approach. Just as polyphasic taxonomy is required for classification of microorganisms (54), polyphasic ecology must be used to study the microbiology of traditional fermented foods. A polyphasic approach is also applicable to other fermented foods, such as cheese, and should allow us to better understand the dynamic changes during traditional fermentation, as well as the variability in similar processes; the ultimate objective is to control the process and the quality of the final food product. ACKNOWLEDGMENTS N. ben Omar was supported by a Grant del Plan Proprio from the University of Granada. This study was supported in part by the Bureau des Ressources Ge´ne´tiques. We thank A. Brauman and E. Miambi for helpful discussions, D. Centurion Hidalgo and J. Espinosa Moreno for pozol samples, and M. Maurin, Universite´ Montpellier II, for assistance with SEM. REFERENCES 1. Adler-Nissen, J., and A. L. Demain. 1994. Aeration-controlled formation of acetic acid in heterolactic fermentations. J. Ind. Microbiol. 13:335–343. 2. Agati, V., J.-P. Guyot, J. Morlon-Guyot, P. Talamond, and J. Hounhouigan. 1998. Isolation and characterization of new amylolytic strains of Lactobacillus fermentum from fermented maize doughs (mawe` and ogi) from Benin. J. Appl. Microbiol. 85:512–520. 3. Akkermans, A.DL., M. S. Mirza, H. J. M. Harmsen, H. J. Blok, P. R. Herron, A. Sessitsch, and W. M. Akkermans. 1994. Molecular ecology of microbes: a review of promises, pitfalls, and true progress. FEMS Microbiol. Rev. 15: 185–194. 4. Alm, E. W., D. B. Oerther, N. Larsen, D. A. Stahl, and L. Raskin. 1996. The oligonucleotide probe database. Appl. Environ. Microbiol. 62:299–306. 5. Amann, R. I., B. J. Binder, R. J. Olson, R. J. Chisholm, R. Devereux, and D. A. Stahl. 1990. Combination of 16S rRNA-targeted probes with flow cytometry for analyzing mixed microbial populations. Appl. Environ. Microbiol. 56:1919–1925. 6. Ampe, F., N. ben Omar, and J.-P. Guyot. 1998. Recovery of total microbial RNA from lactic acid fermented foods with a high starch content. Lett. Appl. Microbiol. 27:270–274. 7. Ampe, F., N. ben Omar, and J.-P. Guyot. 1999. Culture-independent quantification of physiologically active microbial groups in Mexican pozol, a lactic acid fermented dough, using rRNA-targeted oligonucleotide probes. J. Appl. Microbiol. 87:131–140.

VOL. 65, 1999 8. Bouton, Y., P. Guyot, and R. Grappin. 1998. Preliminary characterization of microflora of Comte´ cheese. J. Appl. Microbiol. 85:123–131. 9. Brauman, A., S. Ke´le´ke´, M. Malonga, E. Miambi, and F. Ampe. 1996. Microbiological characterization of cassava retting, a traditional lactic acid fermentation for foo-foo (cassava flour) producion. Appl. Environ. Microbiol. 62:2854–2858. 10. Can ˜ as, A., E. Barzana, J. D. Owens, and C. Wacher. 1993. La elaboracion de pozol en los altos de Chiapas. Ciencia (Mexico City) 44:219–229. 11. Cerning, J. 1990. Exocellular polysaccharides produced by lactic acid bacteria. FEMS Microbiol. Rev. 87:113–130. 12. Cotta, M. A., and T. R. Whitehead. 1993. Regulation and cloning of the gene encoding amylase activity of the ruminal bacterium Streptococcus bovis. Appl. Environ. Microbiol. 59:189–196. 13. Daeschel, M. A., R. E. Anderson, and H. P. Fleming. 1987. Microbial ecology of fermenting plant materials. FEMS Microbiol. Rev. 46:357–367. 14. de Man, J., M. Rogosa, and M. E. Sharpe. 1960. A medium for the cultivation of lactobacilli. J. Appl. Bacteriol. 23:130–135. 15. Escalante, A., T. Flores, J. Morlon-Guyot, A. Farre´s, and C. Wacher-Rodarte. 1999. Unpublished data. 16. Felske, A., A. D. L. Akkermans, and W. M. De Vos. 1998. Quantification of 16S rRNAs in complex bacterial communities by multiple competitive reverse transcription-PCR in temperature gradient gel electrophoresis fingerprints. Appl. Environ. Microbiol. 64:4581–4587. 17. Food and Agriculture Organization. 1995. Workshop on fermentation as a household technology to improve food safety. Food and Agriculture Organization, World Health Organization, Pretoria, South Africa. 18. Franks, A. H., H. J. M. Harmsen, G. C. Raands, G. J. Jansen, F. Schut, and G. W. Welling. 1998. Variations of bacterial populations in human feces measured by in situ hybridization with group-specific 16S rRNA-targeted oligonucleotide probes. Appl. Environ. Microbiol. 64:3336–3345. 19. Garvie, E. I. 1960. The genus Leuconostoc and its nomenclature. J. Dairy Res. 27:283–292. 20. Giraud, E., A. Brauman, S. Ke´le´ke´, B. Lelong, and M. Raimbault. 1991. Isolation and physiological study of an amylolytic strain of Lactobacillus plantarum. Appl. Microbiol. Biotechnol. 36:379–383. 21. Giraud, E., A. Champailler, and M. Raimbault. 1994. Degradation of raw starch by a wild amylolytic strain of Lactobacillus plantarum. Appl. Environ. Microbiol. 60:4319–4323. 22. Halm, M., A. Lillie, A. K. Sørensen, and M. Jakobsen. 1993. Microbiological and aromatic characteristics of fermented maize doughs for kenkey production in Ghana. Int. J. Food Microbiol. 19:135–143. 23. Hamad, S. H., G. Bo ¨cker, R. F. Vogel, and W. P. Hammes. 1992. Microbiological and chemical analysis of fermented sorghum dough for Kisra production. Appl. Microbiol. Biotechnol. 37:728–731. 24. Hamad, S. H., M. C. Dieng, M. A. Ehrmann, and R. F. Vogel. 1997. Characterization of the bacterial flora of Sudanese sorghum flour and sorghum sourdough. J. Appl. Microbiol. 83:764–770. 25. Harrigan, W. F., and M. EM. McCance. 1976. Laboratory methods in food and dairy microbiology. Academic Press, London, United Kingdom. 26. Heuer, H., M. Krsek, P. Baker, K. Smalla, and E. M. Wellington. 1997. Analysis of actinomycete communities by specific amplification of genes encoding 16S rRNA and gel-electrophoretic separation in denaturing gradients. Appl. Environ. Microbiol. 63:3233–3241. 27. Hounhouigan, D. J., M. J. R. Nout, C. M. Nago, J. H. Houben, and F. M. Rombouts. 1993. Characterization and frequency distribution of species of lactic acid bacteria involved in the processing of mawe`, a fermented maize dough from Benin. Int. J. Food Microbiol. 18:279–287. 28. Hugenholtz, P., B. M. Goebbel, and N. R. Pace. 1998. Impact of cultureindependent studies on the emerging phylogenetic view of bacterial diversity. J. Bacteriol. 180:4765–4774. 29. Jackson, C. R., E. E. Roden, and P. F. Churchill. 1998. Changes in bacterial species composition in enrichment cultures with various dilutions of inoculum as monitored by denaturing gradient gel electrophoresis. Appl. Environ. Microbiol. 64:5046–5048. 30. Johansson, M.-L., A. Sanni, C. Lo ¨nner, and G. Molin. 1995. Phenotypically based taxonomy using API 50CH of lactobacilli from nigerian ogi, and the occurrence of starch fermenting strains. Int. J. Food. Microbiol. 25:159–168. 31. Lane, D. J. 1991. 16S/23S rRNA sequencing, p. 115–175. In E. Stackebrandt and M. D. Goodfellow (ed.), Nucleic acid techniques in bacterial systematics. Wiley, Chichester, United Kingdom. 32. Lens, P. N. L., D. De Beer, C. C. H. Cronenberg, F. P. Houwen, S. P. P. Ottengraf, and W. H. Verstrate. 1993. Heterogeneous distribution of microbial activity in methanogenic aggregates: pH and glucose microprofiles. Appl. Environ. Microbiol. 59:3803–3815. 33. Lin, C., L. Raskin, and D. A. Stahl. 1997. Microbial community structure in gastrointestinal tracts of domestic animals: comparative analyses using rRNA-targeted oligonucleotides probes. FEMS Microbiol. Ecol. 22:281–294. 34. Low, D., J. A. Ahlgren, D. Horne, D. J. McMahon, C. J. Oberg, and J. R. Broadbent. 1998. Role of Streptococcus thermophilus MR-1C capsular exopolysaccharide in cheese moisture retention. Appl. Environ. Microbiol. 64:2147–2151.


5473

35. McCrady, M. H. 1918. Tables for rapid interpretation of fermentation tube results. Can. Public Health J. 9:201. 36. McDonald, L. C., H. P. Fleming, and H. M. Hassan. 1990. Acid tolerance of Leuconostoc mesenteroides and Lactobacillus plantarum. Appl. Environ. Microbiol. 56:2120–2124. 37. Maidak, B. L., J. R. Cole, C. T. Parker, Jr., G. M. Garrity, N. Larsen, B. Li, T. G. Lilburn, M. J. McCaughey, G. J. Olsen, R. Overbeek, S. Pramanik, T. M. Schmidt, J. M. Tiedje, and C. R. Woese. 1999. A new version of the RDP (Ribosomal Database Project). Nucleic Acids Res. 27:171–173. 38. Morlon-Guyot, J., J. P. Guyot, B. Pot, I. Jacobe de Haut, and M. Raimbault. 1998. Lactobacillus manihotivorans sp. nov., a new starch-hydrolyzing lactic acid bacterium isolated from cassava sour starch fermentation. Int. J. Syst. Bacteriol. 48:1101–1109. 39. Murray, A. E., J. T. Hollibaugh, and C. Orrego. 1996. Phylogenetic compositions of bacterioplankton from two California estuaries compared by denaturing gradient gel electrophoresis. Appl. Environ. Microbiol. 62:2676–2680. 40. Muyzer, G., E. C. De Wall, and A. G. Uitterlinden. 1993. Profiling of complex microbial populations by denaturing gradient gel electrophoresis of 16S ribosomal DNA fragments. Appl. Environ. Microbiol. 59:695–700. 41. Nierswicky-Bauer, S. 1994. In situ hybridization workshop. Microbial diversity summer course. Woods Hole Oceanographic Institution, Woods Hole, Mass. 42. Nissen, H., A. Holck, and R. H. Dainty. 1994. Identification of Carnobacterium spp. and Leuconostoc spp. in meat by genus-specific 16S rRNA probes. Lett. Appl. Microbiol. 19:165–168. 43. Nuraida, L., M. C. Wacher, and J. D. Owens. 1995. Microbiology of pozol, a Mexican fermented maize dough. World J. Microbiol. Biotechnol. 11:567– 571. 44. Øvreas, L., L. Forney, F. L. Daae, and V. Torsvik. 1997. Distribution of bacterioplankton in meromictic Lake Saelenvannet, as determined by denaturing gradient gel electrophoresis of PCR-amplified gene fragments coding for 16S rRNA. Appl. Environ. Microbiol. 63:3367–3373. 45. Paludan-Mu ¨ller, C., H. H. Huss, and L. Gram. 1998. Characterization of lactic acid bacteria isolated from a Thai low-salt fermented fish product and the role of garlic as substrate for fermentation. Int. J. Food Microbiol. 46:219–229. 46. Pederson, C. S., and M. N. Albury. 1961. The effect of pure culture inoculation on fermentation of cucumber. Food Technol. 15:351–354. 46a.Ribosomal Database Project. 9 July 1999, revision date. [Online.] Ribosomal Database Project. http://www.cme.msu.edu/rdp/. [28 September 1999, last date accessed.] 47. Salama, M., W. Sandine, and S. Giovannoni. 1991. Development and application of oligonucleotide probes for identification of Lactococcus lactis subsp. cremoris. Appl. Environ. Microbiol. 57:1313–1318. 48. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 49. Sghir, A., D. Antonopoulos, and R. I. Mackie. 1998. Design and evaluation of a Lactobacillus group-specific ribosomal RNA-targeted hybridization probe and its application to the study of intestinal microecology in pigs. Syst. Appl. Microbiol. 21:291–296. 50. Stahl, D. A., B. Flescher, H. R. Mansfield, and L. Montgomery. 1988. Use of phylogenetically based hybridization probes for studies of ruminal microbial ecology. Appl. Environ. Microbiol. 54:1079–1084. 51. Tanner, M. A., B. M. Goebbel, M. A. Dojka, and N. Pace. 1998. Specific ribosomal DNA sequences from diverse environmental settings correlate with experimental contaminants. Appl. Environ. Microbiol. 64:3110–3113. 52. Tseng, C.-P., and T. J. Montville. 1993. Metabolic regulation of end-product distribution in lactobacilli: causes and consequences. Biotechnol. Prog. 9:113–121. 53. Ulloa, M. 1974. Mycofloral succession in pozol from Tabasco, Mexico. Bol. Soc. Mex. Mycol. 8:17–48. 54. Vandamme, P., B. Pot, M. Gillis, K. Kersters, and J. Swings. 1996. Polyphasic taxonomy, a consensus approach to bacterial systematics. Microbiol. Rev. 60:407–438. 55. von Wintzingerode, F., U. B. Go ¨bel, and E. Stackebrandt. 1997. Determination of microbial diversity in environmental samples: pitfalls of PCR-based rRNA analysis. FEMS Microbiol. Rev. 21:213–229. 56. Wacher, C., A. Canas, P. E. Cook, E. Barzana, and J. D. Owens. 1993. Sources of microorganisms in pozol, a traditional Mexican fermented maize dough. World J. Microbiol. Biotechnol. 9:269–274. 57. Wacher-Rodarte, M. C. 1995. Estudios sobre la microbiologia del pozol. Ph.D. thesis. Universidad Nacional Autonoma de Mexico, Mexico D.F., Mexico. 58. Welling, G. G. 1998. Personal communication. 59. Widdel, F., and N. Pfennig. 1984. Dissimlatory sulfate or sulfur-reducing bacteria, p. 663–679. In N. R. Krieg and J. G. Holt (ed.), Bergey’s manual of systematic bacteriology, vol. 1. Williams & Wilkins Co., Baltimore, Md. 60. Zheng, D., E. W. Alm, D. A. Stahl, and L. Raskin. 1996. Characterization of universal small-subunit rRNA hybridization probes for quantitative molecular microbial ecology studies. Appl. Environ. Microbiol. 62:4504–4513.