Combination of Multiplex PCR and PCR-Denaturing Gradient Gel Electrophoresis for Monitoring Common Sourdough-Associated Lactobacillus Species

Abstract

A combination of denaturing gradient gel electrophoresis (DGGE) and a previously described multiplex PCR approach was employed to detect sourdough lactobacilli. Primers specific for certain groups of Lactobacillus spp. were used to amplify fragments, which were analyzed by DGGE. DGGE profiles obtained from Lactobacillus type strains acted as standards to analyze lactobacilli from four regional Abruzzo (central Italy) sourdoughs.

PCR-based methods are frequently applied for the purpose of bacterial typing (13) or monitoring changes in microbial populations (2). Although a polyphasic methodology, based on both culture-dependent and culture-independent methods, is the preferred approach to determine the presence of metabolically active microflora of fermented foods (1), culture-independent methods are useful tools to rapidly follow microbial communities that are significant for the development of the specific characteristics of the final product, e.g., the population dynamics of various lactobacilli during sourdough fermentation (15).

Among culture-independent methods, developed to circumvent the limitations of conventional microbial community analysis (17), denaturing gradient gel electrophoresis (DGGE) analysis is one of the most widely applied genetic methods to quickly and economically study microbial diversity (10, 12). DGGE represents a method to obtain species information through analysis of PCR products of the same length but with different sequences and is based on differential mobility in an acrylamide gel matrix of increasing denaturant (urea and formamide) concentration. Generally, total bacterial DNA from the habitat of interest is extracted and a region with a hypervariable nucleotide base sequence is amplified by PCR and subjected to DGGE.

This work aimed to develop a single-step identification methodology for the differential detection of 16 Lactobacillus species that are generally associated with the sourdough ecosystem by combining DGGE with a previously described multiplex PCR method (15). The applicability of the system was verified with four sourdoughs from artisan bakeries of the Abruzzo region of central Italy.

Strains used in this study are listed in Table 1. Sourdough lactic acid bacteria were propagated in sourdough bacterial broth (8) at 30°C for 24 h, while all other strains were propagated as indicated by the respective culture collections for 24 h.

TABLE 1.

Lactic acid bacterial strains used in this study

Strain	Source	Aim
Lactobacillus alimentarius LMG 9187T	Culture collectiona	DGGE development
L. alimentarius DSM 20181	Culture collection	System validation
Lactobacillus brevis ATCC 14869T	Culture collection	DGGE development
L. brevis 5Z	Sourdough, Italy	System validation
Lactobacillus farciminis DSM 20184T	Culture collection	DGGE development
L. farciminis DSM 20180	Culture collection	System validation
Lactobacillus fermentum ATCC 14931T	Culture collection	DGGE development
L. fermentum DSM 20391	Culture collection	System validation
Lactobacillus fructivorans DSM 20203T	Culture collection	DGGE development
L. fructivorans DSM 20607	Culture collection	System validation
Lactobacillus frumenti DSM 13145T	Culture collection	DGGE development
Lactobacillus hilgardii DSM 20176T	Culture collection	DGGE development
L. hilgardii DSM 20051	Culture collection	System validation
Lactobacillus mindensis DSM 14500T	Culture collection	DGGE development
Lactobacillus panis DSM 6035T	Culture collection	DGGE development
Lactobacillus paralimentarius DSM 13238T	Culture collection	DGGE development
L. paralimentarius CQ5	Sourdough, Italy	System validation
Lactobacillus paraplantarum DSM 10667T	Culture collection	DGGE development
L. paraplantarum DSM 10641	Culture collection	System validation
Lactobacillus pentosus ATCC 8041T	Culture collection	DGGE development
L. pentosus DSM 20199	Culture collection	System validation
Lactobacillus plantarum ATCC 14917T	Culture collection	DGGE development
L. plantarum 20	Sourdough, Italy	System validation
Lactobacillus pontis DSM 8475T	Culture collection	DGGE development
L. pontis DSM 8476	Culture collection	System validation
Lactobacillus rossiae DSM 15814T	Culture collection	DGGE development
Lactobacillus sanfranciscensis DSM 20451T	Culture collection	DGGE development
L. sanfranciscensis 7A	Sourdough, Italy	System validation
Lactococcus lactis subsp. lactis DSM 20481T	Culture collection	Specificity check
Weissella confusa DSM 20196T	Culture collection	Specificity check

^aAmerican Type Culture Collection (ATCC), Manassas, Va.; Deutsche sammlung von Mikroorganismen und Zellkulturen (DSM), Braunschweig, Germany; and Laboratorium voor Microbiologie, University of Ghent, (LMG), Ghent, Belgium.

Genomic DNAs were extracted as reported by De Los Reyes-Gavilán et al. (6) from 2-ml samples of overnight cultures.

The combination of multiplex PCR and PCR-DGGE developed in this study was performed as outlined in Table 2. Briefly, group-specific DNA fragments obtained following multiplex PCR were directly amplified (group 1) or amplified after band excision and purification (groups 2 to 4) (GFX PCR DNA and gel band purification kit [Amersham Biosciences]) using primer combinations in which one of the primers contained a GC clamp for DGGE purposes (18). For groups 3 and 4, the latter primer combinations were based on sequences within the multiplex-generated DNA fragment (nested set of primers) so as to generate a shorter PCR fragment in order to suit the DGGE, which allows separation of DNA fragments with a maximum length of 500 bp (16). All PCRs were carried out using the same conditions as those for Grouping-multiplex PCR (15). The high annealing temperature (65°C) of the PCR replaced the touchdown PCR approach normally associated with DGGE (11). The optimal range of band separation was determined by a perpendicular gel prepared as indicated by the Bio-Rad DCode apparatus (Bio-Rad) supplier. Parallel electrophoresis was run at 60°C for 10 min at 20 V, 14 h at 40 V, and 2 h at 140 V. Gels were stained with ethidium bromide for 15 min and rinsed for 30 min in distilled H₂O.

TABLE 2.

Groups of Lactobacillus species, oligonucleotide primer sequences and locations, and DGGE conditions

Lactobacillus group and organism	Multiplex PCR primersa	DGGE primersb	DGGE conditionsc
1 L. paraplantarum L. pentosus L. plantarum	Gru1′"F, Gru2"F, Gru3F, Gru4F, LrosF, LrosF, MulISRR	60g1F (5′-GGCATTCGTCGTGGGATTGAAGAAGC-3′ [40-65, AY424357])/GC-60g1R (5′-CCATGTATTGTGATAAGTAGCCGCGG-3′ [282-307, AY424357 reversed])	9% (wt/vol) polyacrylamide gel in 1× TAE, denaturing gradients (FU) 41-46%
2 L. alimentarius L. farciminis L. mindensis L. paralimentarius	Gru1′"F, Gru2"F, Gru3F, Gru4F, LrosF, LrosF, MulISRR	Gru2"F/GC-MulISRR	9% (wt/vol) polyacrylamide gel in 1× TAE, denaturing gradients (FU) 41-46%
3 L. brevis L. fructivorans L. hilgardii L. sanfranciscensis	Gru1′"F, Gru2"F, Gru3F, Gru4F, LrosF, LrosF, MulISRR	Nest3F (5′-GGGGTTGTAGGACTGAACATTTG-3′ [512-534, AJ616221])/GC-MulISRR	9% (wt/vol) polyacrylamide gel in 1× TAE, denaturing gradients (FU) 41-46%
4 L. fermentum L. frumenti L. panis L. pontis	Gru1′"F, Gru2"F, Gru3F, Gru4F, LrosF, LrosF, MulISRR	Nest4F (5′-CGATTCCCTGAGTAGCGGC-3′ [438-456, AJ616012])/GC-MulISRR	6% (wt/vol) polyacrylamide gel in 1× TAE, denaturing gradients (FU) 53-58%

^aThe primers written in bold represent primers specific for the group indicated in that row. See reference 15 for primer sequences.

^bPrimer sequences are given in parentheses, and nucleotide positions and GenBank accession numbers are given in brackets.

^cTAE, 40 mM Tris, 20 mM acetic acid, and 1 mM EDTA (pH 8.0); FU, 100% denaturant contained 40% (vol/vol) formamide and 7 M urea.

Four standard ladders, one for each group, were developed, representing 15 Lactobacillus species in total (Fig. 1). Although methods for qualitative analysis of specific populations of complex ecosystems based on genus-specific PCR and DGGE have been developed (7, 14), the commonly used universal 16S rRNA gene-derived primers can readily amplify yeast, fungal, or plant DNAs copurified with bacterial DNA (9). In order to avoid misleading DGGE profiles, we approached this problem by using primers specific for lactobacilli based on parts of the hsp60 gene for group 1 species and on parts of the 23S rRNA gene for species belonging to groups 2 to 4.

FIG. 1.

DGGE ladders. The Lactobacillus organisms and strains are as follows: lane 1, L. plantarum ATCC 14917^T; lane 2, L. pentosus ATCC 8041^T; lane 3, L. paraplantarum DSM 10667^T; lane 4, L. alimentarius LMG 9187^T; lane 5, L. paralimentarius DSM 13238^T; lane 6, L. farciminis DSM 20184^T; lane 7, L. mindensis DSM 14500^T; lane 8, L. sanfranciscensis DSM 20451^T; lane 9, L. fructivorans DSM 20203^T; lane 10, L. hilgardii DSM 20176^T; lane 11, L. brevis ATCC 14869^T; lane 12, L. fermentum ATCC 14931^T; lane 13, L. pontis DSM 8475^T; lane 14, L. panis DSM 6035^T; and lane 15, L. frumenti DSM 13145^T.

In order to validate the above-described system and verify the standard ladders, DGGE fragments from lactobacilli, found at the same position of the type strains, were excised from gel and eluted in 20 μl sterile double-distilled H₂O overnight at 4°C. One microliter of the eluted DNA of each DGGE band was amplified as reported above. This operation was repeated once or twice until a single band was obtained on gel, and then the single bands were amplified with primers that did not contain the GC clamp for their subsequent insertion in plasmid vectors for direct sequencing (MWG Biotech AG).

Since a 180-bp band and a 280-bp band, characteristic of groups 2 and 3, respectively, were obtained with Weissella confusa DSM 20196^T and Lactococcus lactis subsp. lactis DSM 20481^T, respectively, when the strains were subjected to Grouping-multiplex PCR (15), these two strains were subjected to DNA amplification with primers containing the GC clamp and loaded on DGGE gel together with the respective group ladder. Neither strain generated an amplification product, thus excluding false-positive results.

Under the DGGE conditions used in this study, all Lactobacillus strains tested could be distinguished on the basis of their unique band migration distances. The system was subsequently tested on total DNA samples isolated from dough environments.

Four doughs were inoculated with a mixture of Lactobacillus strains that belonged to the same multiplex group, thus representing the most complex situation in terms of species discrimination, since all strains are identified by an identically sized PCR product. In addition, two reference doughs, named R (without starter, not fermented) and FR (without starter, fermented at 30°C for 24 h), were prepared and treated as previously described by Settanni et al. (15) for analysis by the multiplex PCR/PCR-DGGE combined system. Total DNA from doughs was extracted by means of two different protocols following the methods reported by Settanni et al. (15) and Ampe et al. (1). Isolated DNAs were first subjected to Grouping-multiplex PCR, and then the bands characteristic of doughs 1 to 4 were excised from the agarose gel and, following amplification with primers specific for DGGE analysis (Table 2), separated by DGGE. Each species showed the band produced when the pure culture was processed to develop the DGGE ladders. As expected, no amplification products from doughs R and FR were obtained with Grouping-multiplex PCR.

The system was also applied to analyze four artisan wheat sourdoughs (A to D) collected in the Abruzzo region (central Italy) at the end of fermentation (6 h at about 25°C). Analysis showed a prevalence of Lactobacillus plantarum and Lactobacillus sanfranciscensis (results not shown), common species in Italian sourdoughs (5), after DNA extraction using both methods. Furthermore, doughs A and C also contained Lactobacillus rossiae (formerly Lactobacillus rossii [3]) and L. paralimentarius (results not shown), respectively, but these species were detected only by use of DNA extracted with the protocol reported by Ampe et al. (1), probably since the detection level of this method is more sensitive than that of the method reported by Settanni et al. (15). These results were in agreement with those obtained with the cultivation method performed as previously described by Corsetti et al. (4), and they showed isolation of L. plantarum and L. sanfranciscensis from the 10⁻⁸ plate count. In contrast, L. rossiae and L. paralimentarius were not found; they were isolated from 10⁻⁷ and 10⁻⁶ plate counts, respectively (data not shown).

The multiplex PCR/PCR-DGGE combined system developed in this study is highly specific for the sourdough Lactobacillus species considered; it is based on differently sized PCR products generated from the various groups of Lactobacillus spp., combined with different nucleotide sequences among species of each group. Hence, it allows the reduction of species to follow to just those belonging to each group.