sign in sign up
<<
Home , Lactobacillus pontis

1'i I. ..

_." ~ .

INTERNATIONALJOURNAL OF Food Microbiology

ELSEVIER International Journal of Food Microbiology 65 (2001) 45-54 www.elsevier.nl/locate/ijfoodmicro

l. Dynarnics of the microbial community responsible for traditional sour cassava studied by denaturing gradient gel electrophoresis and quantitative rRNA hybridization

FrédBribAmpe a, * , Audrey Sirvent a, Nadine Zakhia b*c I Laboratoire de Biotechnologie Microbienne Tropicale, Institut de Recherche pour le Développement, 911 Avenue Agropolis, BP 5045, F-34032 Montpellier cedex I, France CIRAD, Programme Agroalirnentaire, TA 40/ 16, 73 rue JF Breton, 34398 Montpellier cedex 5, France E CIAT, Rural Agroenterprise Development Project, AA 6713, Cali, Colombia Received 20 August 2000; accepted 15 November 2000 ..

Abstract

The microbial community developing during the spontaneous fermentation of sour cassava starch was investigated by cultivation-independent methods. Denaturing gradient gel electrophoresis (DGGE) of partially amplified 16s rDNA followed by sequencing of the most intense bands showed that the dominant organisms were all lactic (LAB), mainly close relatives of Bifdobacteriunt minintuin, Lactococcus lactis, Streptococcus sp., Enterococcus saccharolyticus and plantarum. Close relatives of Lb. panis, Leuconostoc ïneseiiteroides and Ln. citreuni were also found. A complementary analysis using hybridization of 16s rRNA with phylogenetic probes was necessary to detect the presence of the recently discovered Lb. iizaiiilzotivoraiis. Although it represented up to 13% of the total bacteria of sour cassava starch, this species could not be detected by DGGE as the PCR product migrated to the same position as Lc. lactis. In addition, it was shown that a strong pH decrease in the time course of fermentation was most probably responsible for the competitive selection of acid-resistant LAB vs. both homo and heterofermentative acid-sensitive LAB. O 200 1 Elsevier Science B.V. All rights reserved.

Keywords: Cassava; Sour starch; ; Denaturing gradient gel electrophoresis; RNA quantification; Oligonucleotide probe; Lb. manihotiuorans.

1. Introduction tation is widely used for root preservation and the a'I elaboration of well-appreciated cassava fermented Ckxwi is a Very imPoltant Staple crop for the dishes and foods. Sour cassava starch is typically diet Of mally tropical Countries. SmalbSCde fermen- processed in Southem America (Colombia, Brazil) and traditionally used for the preparation of cheese- breads. The natural lactic fermentation of wet-ex- * Corresponding author. LBMRPM, CNRS-INRA BP 27, Chemin de Borde Rouge, 3 1326 Castanet-Tolosan Cedex, France. tracted cassava starch, along with sun-drying, were Fax: +33-4-561-285061. proved to be the key steps for conferring sour cas- E-mail address: [email protected] (F. Ampe). sava starch some specific functional properties, such I l 0168-1605/01/$ - see front matter O 2001 Elsevier Science B.V. All rights reserved. I D PII SOl68-1605(00)00502-X Il bonds I ,/ ' -_I_- - -_ - - ____ , Documenhire R 46 (2001) F. Ampe et al. /Internotional Journal of Food Microbiology 65 45-54 as expansion during dough baking (Westby and formed in the same conditions. Numerical results are Cereda, 1994; Zakhia et al., 1996). the means of the data obtained for these five fermen-

Previous studies have shown that the natural sour tations. '8 cassava starch fermentation was mainly due to the action of lactic acid bacteria (LAB), and that fermen- 2.2. Scanning electron inicroscopy tation temperature and duration as well as the com- position of the microflora influenced the expansion properties of the final cassava sour starch (Brabet, Sour cassava starch samples were fixed with glu- 1994; Figueroa et al., 1995). During these works, taraldehyde and dehydrated with an graded some LAB strains involved in the natural sour cas- series as described by Giraud et al. (1994). Ethanol sava starch fermentation were isolated, identified and was then removed by application of CO, at the pitical point. The samples were coated with plat- characterized using classical microbiological " tech- niques (Figueroa et al., 1995; Morlon-Guyot et al., inum and SEM observation was performed with a 1998; ben Omar et al., 2000). JEOL JSM-6300F microscope (University of Mont- However, recent studies on the fermentation of pellier UMII, France). pozol, a traditional maize-fermented food, demon- strated that classical microbiological approach failed 2.3. Analysis of pH, sugars and ferrnentation prod- to adequately describe the microflora of fermented ucts foods and recommended that a polyphasic approach should be used for a better description of the micro- Ten grams of starch was diluted fivefold in sterile bial communities of these environments (Ampe et . 0.2 ml of 2 N H,S04 was added to 1.3 ml of al., 1999b). Therefore, this paper aims at describing this suspension and the mixture was centrifuged 10 the dynamics of the microbial community during min at 10,000 X g. The concentrations of soluble traditional cassava starch fermentation using culture- starch, sugars, ethanol and organic in the su- independent methods. pernatants were assayed by high-pressure liquid chromatography using an Aminex HPX87H column (BioRad, Richmond CA). Running conditions were: ml 2. Materials and methods mobile phase, H,S04 6 mmol 1-I; flow rate, 0.8 min-' ; temperature, 65°C. Detection was performed using a refractometer (PU 4026 Philips, Heindoven). 2.1. Sour cassava starch The pH was measured directly in the suspension. All results were the means of five determinations. The cassava Mbra 383 variety usually cultivated in the Cauca Department, Colombia (around 1100 m above sea level), was processed into sour starch 2.4. DNA isolation according to the traditional small-scale processing in this region. Twelve kilograms of the wet-extracted Total DNA was extracted from fermented starch starch (45% to 50% of water content) was put in a at different fermentation times by a method previ- l-m-height PVC tube and covered with a 1- to 2-cm ously tested for high starch-containing foods (Ampe layer of running water. The PVC tubes were intro- et al., 1999b). One gram of starch was resuspended duced in the traditional fermentation tank and the in 10 ml 0.9% NaCl and homogenized for 30 s at starch was allowed to naturally ferment for 30 days maximal speed with an Ultraturrax T2.5 (Janke and at ambient temperature (around 25°C) in liters tanks. Kunkel, IKA' Labortechnik). Two tubes (1.5 ml After fermentation, the starch was laid on black each) of this suspension were then centrifuged at polyethylene sheets and sun-dried until a final 10% 7000 X g for 10 min. Five hundred microliters of water content was reached. Sour cassava starch was lysosyme (20 pg pl-!) in TES buffer (50 mM Tris mM sampled at t = 1, 6, 16 and 30 days of fermentation pH 8-1 EDTA-8.56% (w/v) saccharose) for further analysis. Five were per- and 10 pl of mutanolysin (1 U pl-') were added to 41 F. Aiiipe et al. /Interiintiorial Journal of Food Microbiology 65 (2001) 45-54

each pellet. Samples were vortexed for 1 min and The PCR products were then analyzed by DGGE incubated for 1 h at 37°C. Fifty microliters of pro- using the Bio-Rad DCode apparatus following the teinase K (10 mg ml-’) was added and the tubes procedure first described by Muyzer et al. (1993). were incubated for 50 min at 50°C then for 10 min at Samples were applied onto 8% (w/v) polyacryl- 65°C. Three hundred microliters of warm (65°C) amide gels in 1X TAE with a denaturing gradient buffer (0.2 M NaCl, 0.1 M Tris-HC1 pH8, 2% SDS) ranging from 25% to 50% UF (100% corresponded was added and the tubes were incubated for 10 min to 7 M urea and 40% [v/v] formamide). The gels at 65°C. Three hundred microliters of 5 M NaCl was were run for 10 min at 20 V and 3 h at 200 V, added and the tubes were gently mixed for 30 s, stained with ethidium bromide for 10 to 15 min, then incubated at 4°C for 10 min, and centrifuged at rinsed for 20 to 30 min with distilled water and 7000 X g, 4°C for 10 min. The supernatant was documented with the GelDoc system (BioRad, Rich- divided into two tubes, and precipitated with 780 pl mond,’CA). of isopropanol by incubation at -20°C for 30 min. A DGGE standard was prepared as follows: (1) The pellets were recovered ‘by centrifugation at total DNA was extracted from exponentially grown X t 12,000 g and 4°C for 15 min, washed with 1 ml Lactobacillus plantaruin LMG18053, Lb. cellobio- 70% ethanol, vacuum-dried and resuspended in 100 sus ATCC9846, Lb. yaracasei I 2030,. Leuconostoc p1 of water. The tubes corresponding to the same mesenteroides ATCC10832, Ln. dextranicurn , sample were then mixed and 700. pl of water was INRA18G, Pediococcus pentosaceus ATCC43200, added. Eight hundred microliters of phenol pH 8 was Weissella paraniesenteroides ATCC333313, Lacto- added. Tubes were mixed for 3 min and centrifuged coccus lactis ATCC1 1454 and Streptococcus sali- at 12,000 X g at room temperature for 10 min. The varius spp. therinopldus CNCM10303. These were aqueous phase was extracted once more with phenol chosen among 40 strains as best spanning the whole and two to three times with pheno1:chloroform:iso- gradient of the DGGE gels prepared here (data not amyl alcohol (25:24: 1) pH 8 before a final extraction shown). (2) PCR products were generated from each with ch1oroform:isoamyl alcohol (24: 1). The aqueous of these DNA preparations. (3) A mixture of these phase was then precipitated with isopropanol, and PCR products was made with final concentrations of: the pellets were washed with 70% ethanol, vacuum- 10 ng pl-’ of PCR products from Lb. plantarum, dried and resuspended in 200 pl water. The extract Lc. lactis and L. paracasei, and 2 ng pl-’ of PCR quality was routinely checked using 1% agarose-TBE products of DNA from the other strains. Ten micro- 0.5 X gels. liters ‘of this standard was routinely used in all DGGE gels. 2.5. PCR-DGGE analysis 2.6. Sequences of DGGE jkaginents The purified DNA was amplified with primers specific to the bacteria or eukarya domains. The DGGE fragments were cut out with a sterile bacterial community DNA was amplified using scalpel. The DNA of each fragment was eluted in primers gc338f (S’CGCCCGCCGCGCGCG- 2O-pl sterile water overnight at 4°C. One microliter y GCGGGCGGGGCGGGGGCACGGGGGGACTCC- of the eluted DNA of each DGGE band was reampli- TACGGGAGGCAGCAG) and 5 18r (5’ATTAC- fied using the same conditions as above. The success CGCGGCTGCTGG) spanning the V3 region of the of this operation was checked by running 3 pl of the a 16s rDNA (0vreas et al., 1997) as previously PCR products in DGGE gels as described above with described (Ampe et al., 1999b). The eukaryotic sour starch-amplified DNA as control. The PCR community DNA was amplified using primers gc- products which yielded a single band co-migrating Euk1427f (5’TCTGTGATGCCCTTAGATGT- with the original band were then purified and se- TCTGGG) and Euk1616r (SGCCGTGTGTACA- quenced. , AAGGGCAGGG) spanning the 1427-1637 region Sequences of these gene fragments were deter- of the 18s rDNA, as described by van Hannen et al. mined by the dideoxy chain-termination method with (1999). the ABI PRISM dye terminator kit (Perkin Elmer). 48 F. Ampe et al./International Journal of Food Microbiology 65 (2001)45-54

The sequencing products were loaded and analyzed Promega (Charbonnières, France), E. coli RNA, on a 373 DNA sequencer (Applied Biosystems). DIG-labelling kit from Roche Diagnostics (Meylan, DGGE fragments were sequenced using primer France). Other molecular biology products were from gc338f. Sigma (Saint-Quentin Fallavier, France). All other To determine the closest known relatives of the chemicals were from Prolabo (Lyon, France). partial 16s rDNA sequences obtained, searches were performed in public data libraries (RDP and Gen- bank) with the FASTA, BLAST and RDP programs 3. Results (Maidak et al., 1999). The CHECK-CHIMERA command of the RDP facilities was used to try to detect chimeric sequences. 3.1. Scanning electron iizicroscopy The Genbank accession numbers for the 16s rDNA partial sequences of DGGE bands were Phase contrast and scanning electron microscopy AF192510 through AF192519. (Fig. 1) indicated that (i) several different morpho- types were involved in the traditional sour cassava * 2.7. RNA extraction and quantitative hybridization starch fermentation including long and short rods, with phylogenetic probes coccoid cells, as well as some typical bifidobacteria The total RNA was extracted from sour cassava (Fig. la-d); (ii) there was a large increase in cell , starch and pure strains using a method previously number in the time course of fermentation; (iii) a optimized for high starch-containing food samples small fraction of starch was degraded during the (Ampe et al., 1998). The recovered RNAs were used fermentation, as suggested from some alveoles de- for quantification with probes S-S-Lbma-0207-a-A- tected on the surface of starch granules after 6 days 20 (Ampe, 2000) and S- * -Lab-0722-a-A-25 (Sghir of fermentation (Fig. lb). Fig. If shows the layered et al., 1998) specifically targeting Lb. inanihotivo- raw starch structure on some degraded granules as runs and all LAB species, respectively. Synthetic the starch degradation increases; (iv) a network of HPLC purified oligonucleotides (Eurogentec, Bel- fibers, most probably corresponding to bacterial gium) were 3'-end labeled with digoxigenin follow- , appeared to be synthesized during ing the instructions of the manufacturer (Boehringer the second half of the fermentation process (Fig. le). Mannheim). RNA blotting and hybridization was performed as described before (Stahl et al., 1988; 3.2. pH, sugars and fermentation metabolites Ampe et al., 1999a). The bound probe was quan- titated by densitometry in relation to reference As. seen from Table 1, the pH rapidly decreased standards after autoradiography. The control RNA during the fermentation time course and reached a content was estimated by hybridization with the uni- final value equal to the pK, of lactic acid (3.5). versal probe *-Univ-1390-a-A-18 (Zheng et al., s- Lactate was the most important fermentation metabo- 1996) prior to use as internal standards, with E. coli lite found in the samples, although significant RNA (Boehringer Mannheim) as absolute reference. amounts of were also measured. Other fer- Abundances of Lb. manihotiuorans were expressed mentation metabolites such as butyrate or propionate, Y as the fraction of the total LAB rRNA in the sample as well as free sugars, were not detected. (namely RNA indexes). The lower limit for detecting

a unique SSU rRNA in the 1 pg of nucleic acid l spotted on the membrane was approximately 3 ng of 3.3. DGGE fingerprinting of bacterial and eukary- SSU-like rRNAs. Results were given as the mean of otic communities three determinations. Two separate extractions of total DNA from each 2.8. Chemicals and reagents sour cassava starch sample were performed. One HPLC-pure primers were from Eurogentec (Sera- microliter of lo-' dilutions and undiluted total DNAs ing, Belgium), Tnq polymerase and dNTPs from were subjected to amplification, and the equal-size F. Aiizpe et al. / Iiiteniatioiial Jolinial of Food Microbiology 65 (2001)45-54 49

Fig. 1. Scanning electron micrographs of sour cassava starch samples taken after 1 (a), 6 (b and c), 15 (d and e) and 30 (f) days of fermentation. The 30-day-fermented starch was sampled after sun-drying and represents the end product.

16s rDNA PCR products were analyzed by DGGE. The fingerprints obtained for the sour cassava starch Repeated extractions as well as dilutions of a given sampled during the fermentation time course showed sample gave similar fingerprints (data not shown). 11 visible bandS (Fig. 2). Bands #1, 4, 6, 8 and 10 were the most intense ones. The same bands were r' Table 1 observed during the time course of fermentation, but pH and fermentation metabolites produced during the sour cassava the relative intensity of several bands (e.g. band #6) starch fermentation. Results are the means of five determinations varied significantly reflecting the changes in the 1 3. standard deviation bacterial assemblage. Time pH Lactate Acetate g-'1 Individual bands observed in the DGGE profiles (days) (pm01 (pmol g-') were then excised from acrylamide gels (Fig. 2) and 1 4.75 f O. 11 6.57 3. 1.40 O reamplified with primers gc338f and 518r. Before 6 3.80f0.05 15.5f6.06 5.63f7.39 15 3.78rk0.11 14.7f 13.4 4.53f5.06 sequencing, each PCR-reamplified DGGE band was 30 3.47f0.05 61.7& 17.0 4.97f2.34 run on a denaturing gradient gel to confirm that the expected product was obtained. All sequences re- 50 F. Ampe et nl./Internationnl Journal of Food Microbiology 65 (2001) 45-54

days 1 6 15 30 Std

Fig. 2. DGGE analysis of PCR-amplified 16s ribosomal DNA fragments from sour cassava starch bacterial communities. DNA was derived from the fermentation time course of a single experiment, but repeated experiments gave similar profiles. The positions and numbering of the bands sequenced (Table 2) and discussed in the text are indicated. A DGGE standard (Std) derived from PCR products of LAB pure strains was included in all the gels run and the corresponding identity of the bands is indicated on the right side of the standard. trieved corresponded to portions of 16s rDNA genes A similar work was performed to study the eu- (Table 2), and except for the band #2 corresponding karyotic community of sour cassava starch. The to the cassava chloroplast DNA, $11 identified se- DGGE fingerprints revealed identical profiles for all quences were clearly homologues of lactic acid bac- tested samples. Two main bands were visible, but the teria 16s rDNA. The closest relatives identified were profiles of fermented cassava were identical to that sequences of the bacteria included in Table 2. In of unfermented cassava (data not shown); therefore, addition, bands #4 and 10 Co-migrated with the the observed bands obviously corresponded to cas- amplified DNA from Lb. plantarum and Lc. laetis sava nuclear DNA, and no DNA from or fungi reference strains, respectively, an observation in good was therefore found in the sour cassava starch sam- agreement with the sequencing results. No sequence ples. could be retrieved from band #3. The CHECK- CHIMERA function of the RDP facilities was used 3.4. Quantification of Lb. inanihotivorans rRNA to detect possible chimeric sequences but none of the retrieved sequences was found to exhibit a chimeric Previous works report on the isolation of LAB nature (Maidak et al., 1999). strains from different traditional cassava starch fer- F. Anipe et al. / International Journal of Food Microbiology 65 (2001) 45-54 51

mentations in Colombia, the majority of which be- Table 3 the longed to the species Lb. plantaruin, Lb. maniho- Importance of species Lb. manihotiuoruiis in sour cassava starch as determined by quantitative dot blot measurements of 16s tivorans and Ln. mesenteroides (Figueroa et al., rRNA 1995; ben Omar et al., 2000). The sour cassava Fermentation % Lb. inaiiihotivoram starch samples used in our study showed bands with time (days) rRNA sequences corresponding to Lb. plantarunz and Ln. 1 1.6k 1.4 mesenteroides, but none of the sequences retrieved 6 13.3 f4.6 corresponded to Lb. manihotiuorans. At the same 15 11.8 f8.6 time, the PCR-DGGE product obtained with DNA 30 ND from a pure culture of Lb. manihotivorans co- ~ The probes used (S-S-Lbma-0207-a-A-20 and S- * -Lab-0722-a-A- migrated with that of Lc. lactis in our experimental 25) were DIG-labelled for hybridization and the RNA controls conditions. As band #4-identified by sequencing used for standardization were from the Lb. maniliotiuorans LMG to be from Lc. Lactis-was very intense, the pres- 18010T and Lb. phiitariirii LMG 18053 strains. Results are ence of Lb. manihotivorans in the analysed samples presented as the relative percentage in comparison with the total LAB and are the means of data obtained from five separate Y could not be dismissed by this experiment only. Therefore, the total RNA was extracted from the fermentationsf standard deviation. ND: not detected. sour cassava starch samples and the relative impor- tance of Lb. nzanihotiuorans was quantified using a phylogenetic probe targeting the 16s rRNA of this species (Ampe, 2000). As (i) all identified organisms results were given as percentages of the total LAB were lactic acid bacteria and (ii) a bias might be (Table 3). Lb. inanilzotivorans was initially present introduced by the presence of cassava RNA, the at a relatively low percentage, but the importance of this species increased during the fermentation until 13% of total LAB after 6-15 days of fermentation; this proportion then decreased to undetected levels at the end of the process. Results were given as the Table 2 means of quantifications obtained with five separate Identity of bands retrieved from DGGE analysis of the sour cassava starch bacterial community fermentations using the same cassava cultivar. Lb. manihotiuomns rRNA was found in all these fermen- Band Closest relative % of Accession #" identity number tations even though the different sour cassava starch samples showed a high variability as reflected by the 1 B$ miriirnum 98% AF1925 10 2 chloroplast from Euphorbiaceae' 94% AF192511 high standard deviations (Table 3). 4 Lc. lactis 93% AF192512 5 Lb. punis and Lb. potitis 94% AF 1925 13 6 Streptococcus sp. 92% AFl925 14 4. Discussion 7 Ln. meseriteroides 97% AF1925 15 8 Ec. saccharolyticus 97% AF1925 16 9 Ln. citreurii 95% AF1925 17 4.1. Succession of microbial populations during the k 10 Lb. plantaruni 93% AF192518 sour cassava starch fernientation 11 Lb. plaiitarim 96% AF 192519 The molecular tools used in this study for follow- "Bands were extracted from the DGGE gel shown in Fig. 2. ing the microbial dynamics of traditional sour cas- 3 bPercentage of identical nucleotides between the sequences retrieved from the DGGE gel and the closest relative found in sava starch fermentation confirmed that this process GenBank or RDP. Comparison was made using partial 16s rDNA was ruled by lactic acid bacteria as previously sug- sequences only (around 180 bases, corresponding to the region gested by many authors (Brabet, 1994; Figueroa et sequenced). al., 1995; ben Omar et al., 2000). Besides, the 'The public banks do not have the sequence for cassava sequences of several partial 16s rDNA genes re- chloroplast 16s rDNA (last verification: August 2000). The com- parison could only be done with other members of the Euphor- trieved from DGGE fingerprints corresponded to the biaceae. dominant LAB species identified in traditional fer- 52 (2001) F. Ainpe et al. /Internatiorial Journal of Food Microbiology 65 45-54 mentations of several starchy foods all over the and Enterococcus species, with intense correspond- world, such as Lb. pontis and Lb. panis for the sour ing bands in the DGGE profiles. Interestingly, the dough in Europe (Vogel et al., 1994), Ln. mesen- intensity of band #1 corresponding to Bifidobac- teroides, Lb. plantarum, Lc. lactis for cassava terizim increased during the fermentation, likely due retting in the Congo (Brauman et al., 1996) or to the progressive setting of anaerobic conditions as Bifidobacterium minimum, Ln. mesenteroides, Ente- first suggested by Cereda (1973). On the contrary, rococcus saccharolyticus, Lb. plantarum for the po- the amount of Lc. lactis increased during the first 6 zol in Mexico (Ampe et al., 1999b). This suggests days of fermentation and then decreased probably that some species might be typically associated with due to the pH decrease and to the sensitivity of this the fermentation of starchy foods. Surprisingly, Sifi- species to low pH values. The same was observed dobacterium and Enterococcus were until recently for Lb. manihotiuorans, another acid-sensitive known to be mainly associated with the gastrointes- species (Guyot et al., 2000), which represented 13% tinal tracts but not to participate in the spontaneous of the total flora after 6 days of fermentation but was lactic acid fermentations of foods. The present study not detected anymore at the end of the fermentation. along with other observations on the Mexican pozo1 Therefore, even though the traditional fermentation (Ampe et al., 1999b) open new orientations for involves a succession of microbial populations as studying the ecology of these organisms. On the suggested by Figueroa et al. (19951, this succession other hand, as several sequences varied significantly seems to be ruled by the resistance vs. sensitivity of' from the closest reference 16s rDNA sequence, the towards the very acidic conditions sour cassava starch might be a reservoir for unknown (pH down to 3.5) reached within 2 weeks of fermen- bacterial species as already illustrated by the isola- tation and not by the type (homo vs. hetero) of tion of the new species Lb. manihotivorans so far fermentative (both Lc. Zactis and Lb. not found in other natural environments (Morlon- manihotivomns are strictly homofermentative LAB). Guyot et al., 1998). This is in agreement with Brabet (1994) who re- Figueroa et al. (1995) have suggested that the ported that 74-85% of the isolated microorganisms traditional sour cassava starch fermentation was ruled throughout three different sour cassava starch fer- by the succession of two microbial groups: heterofer- mentations in Colombia were homofermentative mentative LAB, mainly Leuconostoc species, would LAB. start the acidification followed by the development Cereda (1973) suggested that nonlactic acid bacte- of homofermentative LAB dominated by Lb. plan- ria-in particular clostridia-may also participate in tarum. However, this hypothesis was based on a the fermentation of cassava starch. In our study, limited number of isolated strains and unavailable neither the DGGE profiles nor the HPLC measure- numerical data. In addition, studying the ecology of ments evidenced the presence of these microorgan- microorganisms by cultivation-dependent methods isms, strongly suggesting that they play no signifi- was shown to be strongly biased, even for fermented cant role in this process. foods (Ampe et al., 1999b). The molecular tools In addition, the presence of and fungi was used in the present study confirmed the importance not detected using either SEM and PCR-DGGE with of the species Lb. plantarum in the traditional fer- eukaryotic primers. This result is in agreement with mentation of sour cassava starch, but also demon- Martinez and Zapata (1981) who found neither yeasts strated that this species was present throughout the nor fungi using classical culture-dependent methods process and not only in the later stages as previously after 8 days of fermentation, and with Brabet (1994) ' thought. By contrast, the very low intensities of the who only found very low counts of yeasts and fungi bands corresponding to the two Leuconostoc species using culture media. The absence of fungi is most suggests that this genus does not represent a domi- probably due to the anaerobiosis during fermenta- nant group during the fermentation. In addition, the tion. However, the absence of yeasts remains to be molecular tools helped in identifying a number of explained as these were found at the end of cassava organisms previously not mentioned among the sour retting, another fermentation type performed in the cassava sour starch flora, such as Bifidobacteriiim Congo, with similar pH and anaerobiosis conditions F. Aiiipe et al. / Interiiatioiial Joliriin1 of Food Microbiology 65 (2001)45-54 53

to those prevailing in sour cassava ,starch fermenta- cassava fermentation. Appl. Environ. Microbiol. 66, 2224- tion (Brauman et al., 1996). 2226. Ampe, F., ben Omar, N., Guyot, J.-P., 1998. Recovery of total microbial RNA from lactic acid fermented foods with a high 4.2. Coinbination of DGGE and RNA quantifcatioii starch content. Lett. Appl. Microbiol. 27,' 270-274. with phylogenetic psobes Ampe, F., ben Omar, N., Guyot, J.-P., 1999a. Culture-independent quantification of physiologically active microbial groups in This work represents the first attempt to study Mexican pozol, a lactic acid fermented dough, using rRNA- cassava starch fermentation with molecular tools. It targeted oligonucleotide probes. J. Appl. Microbiol. 87, 131- 140. clearly demonstrated that DGGE fingerprinting fol- Ampe, F., ben Omar, N., Moizan, C., Wacher, C., Guyot, J.-P., lowed by 16s rDNA sequencing can successfully 1999b. Polyphasic study of the spatial distribution of microor- help in the faster identification of the bacterial species ganisms in Mexican pozol, a maize-fermented dough, demon- involved in cassava fermentation. This advantage strates the need for cultivation-independent methods to investi- would thus allow to avoid biased and time-consum- gate traditional fermentations. Appl. Environ. Microbiol. 65, 5464-5473. ing culture-dependent methods and open new per- ben Omar, N., Ampe, F., Raimbault, M., Guyot, J.-P., Tailliez, P., spectives for studying other cassava fermentation 2000. Molecular diversity of lactic acid bacteria from cassava @ types, such as those performed in Western and Cen- sour starch (Colombia). Syst. Appl. Microbiol. 23, 285-291. tral Africa. However, it also pointed out one of the Brabet, C., 1994. Etude des mécanismes physico-chimiques et limits of DGGE analysis already reported by Murray biologiques responsables du pouvoir de panification de 3 l'amidon fermenté de manioc. PhD thesis dissertation. Univer- et al. (1996): several DNA fragments sharing differ- sit6 de Montpellier II, France. ent sequences-and potentially corresponding to dif- Brauman, A., Kéléké, S., Malonga, M., Miambi, E., Ampe, F., ferent species-can exhibit the same melting be- 1996. Microbiological characterization of cassava retting a haviour on a DGGE gel. In our study, the PCR traditional lactic acid fermentation for foo-foo (cassava flour) products obtained from the amplification of DNA production. Appl. Environ. Microbiol. 62, 2854-2858. Cereda, M.P., 1973. Alguns aspectos sobre a fermentação da from Lc. lactis and Lb. nzaiiihotiuol-ans could not be fecula de mandioca. PhD thesis dissertation, Botucatu, Brasil. distinguished by the sole DGGE analysis. The impor- Figueroa, C., Davila, A.M., Pourquié, J., 1995. Lactic acid bacte- tance of the latter species could only be demon- ria of the sour cassava starch fermentation. Lett. Appl. Micro- strated through the use of rRNA quantification with bid. 21, 126-130. S-S-Lbma-0207-a-A-20 probe. This result empha- Giraud, E., Champailler, A., Raimbault, M., 1994. Degradation of raw starch by a wild amylolytic strain of Lactobacillus plan- sizes the need for combining several molecular taruriz. Appl. Environ. Microbiol. 60, 43 19-4323. methods to avoid the biases inherent to each one. Guyot, J.P., Calderon, M., Morlon-Guyot, J., 2000. Effect of pH Besides, it confirms the importance of pursuing ef- control on lactic acid fermentation of starch by Lactobacillus forts for isolating microbial species from their natu- i~iaiiiliofi~~ormzsLMG 18010T. J. Appl. Microbiol. 88, 176- ral environments. 182. Maidak, B.L., Cole, J.R., Parker Jr., C.T., Garrity, G.M., Larsen, N., Li, B., Lilburn, T.G., McCaughey, M.J., Olsen, G.J., Overbeek, R., Pramanik, S., Schmidt, T.M., Tiedje, J.M., Acknowledgements Woese, C.R., 1999. A new version of the RDP (Ribosomal Database Project). Nucl. Acids Res. 27, 171-173. The authors thank F. Alarcon (CIAT, Colombia) Martinez, A., Zapata, L.E., 1981. Aislamiento y caracterizacion Y dela microflora presente durante la fermentacion del almidon and N. ben Omar (IRD) for technical assistance in de yuca. Technologia 174, 23-30. . sampling and analysis and M. Grill, Université Morlon-Guyot, J., Guyot, J.P., Pot, B., Jacobe de Haut, I., Raim- Montpellier II, for assistance with SEM. bault, M., 1998. Lactobacillus nianihotiuorans sp. nov., a new 1 starch hydrolyzing lactic acid bacterium isolated from cassava sour starch fermentation. Int. J. Syst. Bacteriol. 48, 1101-1 109. Murray, A.E., Hollibaugh, J.T., Orrego, C., 1996. Phylogenetic References compositions of bacterioplankton from two California estuar- ies compared by denaturing gradient gel electrophoresis. Appl. Ampe, F., 2000. Design and evaluation of a Lacfobacilliis mani- Environ. Microbiol. 62, 2676-2680. hotiuoraris species-specific ribosomal RNA-targeted hy- Muyzer, G., De Wall, E.C., Uitterlinden, A.G., 1993. Profiling of the bridization probe and its application to study of sour complex microbial populations by denaturing gradient gel 54 F. Ainpe et ~l./ International Journal of Food Microbiology 65 (2001)45-54

electrophoresis of 16s ribosomal DNA fragments. Appl. Envi- community structure after mass lysis of filamentous cyanobac- ron. Microbiol. 59, 695-700. teria associated with viruses. Appl. Environ. Microbiol. 65, avreas, L., Fomey, L., Daae, F.L., Torsvik, V., 1997. Distribution 795-801. of bacterioplankton in meromictic lake Salenvannet, as deter- Vogel, R.F., Böcker, G., Stolz, P., Ehrmann, M., Fanta, D., mined by denaturing gradient electrophoresis of PCR-ampli- Ludwig, W., Pot, B., Kersters, K., Schleifer, K.-H., Hammes, fied gene fragments coding for 16s rRNA. Appl. Environ. W., 1994. Identification of lactobacilli from and Microbiol. 63, 3367-3373. description of Lactobacillus pontis sp. nov. Int. J. Syst. Sghir, A., Antonopoulos, D., Mackie, R.I., 1998. Design and Bacteriol. 44, 223-229. evaluation of a Lactobacillus group-specific ribosomal RNA- Westby, A., Cereda, M.P., 1994. Production of fermented cassava targetea hybridization probe and its application to the study of starch (polvilho azedo) in Brazil. Trop. Sci. 34, 203-210. intestinal microecology in pigs. Syst. Appl. Microbiol. 21, Zakhia, N., Dufour, D., Chuzel, G., Griffon, D., 1996. Review of 291-296. sour cassava starch production in rural Colombian areas. Trop. Stahl, D.A., Flescher, B., Mansfield, H.R., Montgomery, L., 1988. Sci. 36, 247-255. Use of phylogenetically based hybridization probes for studies Zheng, D., Alm, E.W., Stahl, D.A., Raskin, L., 1996. Characteri- of ruminal microbial ecology. Appl. Environ. Microbiol. 54, zation of universal small-subunit rRNA hybridization probes 1079- 1084. for quantitative molecular microbial ecology studies. Appl. van Hannen, EJ., Zwart, G., van Agterveld, P., Gons, H.J., Ebert, Environ. Microbiol. 62. 4504-4513. J., Laanbroek, H.J., 1999. Changes in bacterial and eukaryotic