Skip to main page content

• HOME
• Current Issue
• Archive
• Alerts
• About ASM
• Contact us
• Tech Support
• Journals.ASM.org

This Article Abstract Full Text (PDF) Supplemental material Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Bilhère, E. Articles by Lonvaud-Funel, A. Search for Related Content PubMed PubMed Citation Articles by Bilhère, E. Articles by Lonvaud-Funel, A. Agricola Articles by Bilhère, E. Articles by Lonvaud-Funel, A.

 Previous Article  |  Next Article 

Applied and Environmental Microbiology, March 2009, p. 1291-1300, Vol. 75, No. 5
0099-2240/09/$08.00+0     doi:10.1128/AEM.02563-08
Copyright © 2009, American Society for Microbiology. All Rights Reserved.

Multilocus Sequence Typing of Oenococcus oeni: Detection of Two Subpopulations Shaped by Intergenic Recombination ^,

Eric Bilhère, Patrick M. Lucas,^* Olivier Claisse, and Aline Lonvaud-Funel

UMR 1219, Université de Bordeaux, INRA, ISVV, Talence 33405, France

Received 10 November 2008/ Accepted 21 December 2008

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Oenococcus oeni is the acidophilic lactic acid bacterial species most frequently associated with malolactic fermentation of wine. Since the description of the species (formerly Leuconostoc oenos), characterization of indigenous strains and industrially produced cultures by diverse typing methods has led to divergent conclusions concerning the genetic diversity of strains. In the present study, a multilocus sequence typing (MLST) scheme based on the analysis of eight housekeeping genes was developed and tested on a collection of 43 strains of diverse origins. The eight targeted loci were successfully amplified and sequenced for all isolates. Only three to 11 different alleles were detected for these genes. The average nucleotide diversity also was rather limited (0.0011 to 0.0370). Despite this limited allelic diversity, the combination of alleles of each strain disclosed 34 different sequence types, which denoted a significant genotypic diversity. A phylogenetic analysis of the concatenated sequences showed that all strains form two well distinct groups of 28 and 15 strains. Interestingly, the same groups were defined by pulsed-field gel electrophoresis, although this method targets different genetic variations. A minimum spanning tree analysis disclosed very few and small clonal complexes. In agreement, statistical analyses of MLST data suggest that recombination events were important during O. oeni evolution and contributed to the wide dissemination of alleles among strains. Taken together, our results showed that MLST is more efficient than pulsed-field gel electrophoresis for typing O. oeni strains, and they provided a picture of the O. oeni population that explains some conflicting results previously obtained.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Oenococcus oeni forms part of the flora of indigenous lactic acid bacteria (LAB) present on grapes and in wine, where it generally becomes the unique LAB species following alcoholic fermentation due to a remarkable tolerance to high-ethanol and low-pH conditions. One or several indigenous O. oeni strains may conduct malolactic fermentation, which contributes to the deacidification, flavor modification, and microbial stability of wine (24). However, the start and beneficial impact of malolactic fermentation are better controlled by inoculating the wine with industrially produced cultures of O. oeni in place of indigenous strains which are unable to develop in certain types of wine.

Since the first description of the species by Garvie in 1967 (12), O. oeni, formerly Leuconostoc oenos, was studied not only for characterizing O. oeni isolates from various origins and eventually for identifying candidates for commercial purposes but also in order to decrypt the evolution and population structure of this singular species. Yang and Woese deduced from the analysis of 16S rRNA genes of LAB that O. oeni is a tachytelic (fast-evolving) species (47). Although this conclusion led to controversy (4, 34), it was well supported by a phylogeny derived from the genome sequences of 12 LAB (27, 33). O. oeni lacks the ubiquitous mismatch repair system genes mutS mutL, which normally contribute to lowering the rates of spontaneous mutations and recombination (28). The absence of these genes was recently correlated to the hypermutability of the species and thus provides a rational basis for the tachytelic status of O. oeni (29). At the infraspecies level, 16S, 23S, and16S-23S spacer sequences proved to be highly conserved among O. oeni strains, suggesting that the species is genetically homogeneous (23, 31, 32, 48). This was supported by the levels of DNA-DNA homology and by similarities of the genetic maps of diverse strains (7, 50). In addition, differentiation of O. oeni strains can only be achieved by using methods targeting minor genotypic differences: random amplification of polymorphic DNA (RAPD) (36, 48, 49), DNA fingerprinting (46), amplified fragment length polymorphism (3), or pulsed-field gel electrophoresis (PFGE) patterns of low-frequency restricted genomes (14, 20-22, 38, 45). However, the concept of genetic homogeneity was repeatedly challenged by the detection of groups of strains—often two major groups—distinguished on the basis of molecular and/or phenotypic/metabolic characteristics, leading authors to suggest that O. oeni might be divided into two subspecies (35, 45).

Analysis of O. oeni strains by multilocus sequence typing (MLST) has provided a new picture of the diversity and population structure of the species (5). MLST is a strategy based on the sequence polymorphism of a set of genes, usually 7 to 10, which has the advantages of being robust (based on genetic data) and electronically portable, to generate data that can be used not only for strain differentiation but also for evolutionary and population studies (25). Although it was originally developed for pathogenic bacteria, MLST became the gold standard for studying lineages and population structures of all kinds of microorganisms (26). The only MLST analysis of O. oeni reported to date has examined variations at five genetic loci among 18 strains (5). The authors of that study detected a high allelic diversity and evidences that recombination events contributed to the dissemination of alleles. They concluded that the O. oeni population is panmictic (no line of clonal descent is easily discernible) and suggested that frequent recombination events and horizontal gene transfers (HGTs) occurred between strains. This hypothesis was well supported by the comparative analysis of LAB genomes, which argued in favor of a substantial number of gene losses and acquisitions during the evolution of LAB (27, 28). In addition, HGTs may be particularly favored in O. oeni that lacks the mutS mutL genes (29). However, the hypothesis of a panmictic population contrasted with both the genetic homogeneity and the existence of subspecies suggested by other typing methods.

The present study was undertaken to develop a new MLST scheme targeting eight housekeeping genes and to compare its discriminatory potential with PFGE, which is generally recognized as the most efficient typing method for O. oeni strains. Comparing the two methods was also interesting since they target different genetic variations: MLST reveals punctual mutations in a few genes, whereas PFGE is more sensitive to large-scale genomic rearrangements due to the presence of genomic islands, insertion sequences or mobile elements. The discriminatory ability of both methods was compared by testing a collection of 43 O. oeni strains, and the obtained data were also used to revisit the population structure of O. oeni.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Bacterial strains and growth conditions.
O. oeni strains used in the present study are described in Table 1. Oenococcus kitaharae DSM 17330 was obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH. All strains were grown under anaerobic conditions at 25°C in MRS broth supplemented with 1% D-L malic acid.

View this table: [in this window] [in a new window]

TABLE 1. Origins and typing data of O. oeni strains analyzed in this study

PFGE.
PFGE was performed as described previously (13). Briefly, bacteria cells grown up to the mid-exponential phase were collected by centrifugation, washed with TE buffer (10 mM Tris, 1 mM EDTA [pH 7.5]), resuspended in T100E buffer (10 mM Tris, 100 mM EDTA [pH 7.5]), and embedded in pieces of certified megabase agarose (Bio-Rad). The cell wall and membranes were weakened by incubating agarose plugs at 37°C in T100E buffer containing 10 mg of lysozyme ml^–1 for 3 h, followed by 16 h in T100E buffer supplemented with 2 mg of pronase ml^–1 plus 15 mg of N-lauroylsarcosine ml^–1. After several washes in TE buffer and water, the agarose-embedded DNA was digested by incubating plugs for 16 h at 25°C in 120-µl reactions containing 30 U of NotI endonuclease and the buffer recommended by the manufacturer (New England Biolabs). The separation of digested DNA fragments was performed by electrophoresis in a 1% PFGE certified agarose gel (Bio-Rad) immersed in 0.5x TBE buffer (45 mM Tris-OH [pH 8.0], 45 mM boric acid, 1 mM EDTA) using a CHEF-DRIII apparatus (Bio-Rad) with pulse times of 1 to 25 s, at 6 V cm^–1, for 22 h and at 15°C. After electrophoresis, the gel was stained for 30 min with 0.5 µg of ethidium bromide ml^–1 and photographed under UV light. The obtained DNA fingerprint patterns were analyzed with Bionumerics 5.1 software (Applied Maths). A dendrogram representing the relatedness of bacterial strains was generated by the unweighted pair group method using arithmetic means (UPGMA) with the Dice coefficient of similarity and a tolerance limit of 2.3%.

DNA extraction, amplification, and sequencing.
The genomic DNA of bacteria was extracted by using a Wizard genomic DNA purification kit according to the manufacturer's instructions (Promega) with minor modifications, including a 1-h pretreatment at 37°C of bacteria cells in EDTA (50 mM, pH 8.0) containing 10 mg of lysozyme ml^–1. DNA preparations were resuspended in sterile water to 10 ng µl^–1 and stored at –20°C. The primers used for PCR amplifications and sequencing reactions are listed in Table 2. PCR amplifications were performed in 25-µl reactions containing a custom-made PCR master mix (MP-Biomedicals), 10 ng of template genomic DNA, and 10 pmol of each primer. The PCR program was 95°C for 2 min; followed by 30 cycles of 95°C for 30 s, 58°C for 30 s, and 72°C for 30 s; followed in turn by a final extension of 10 min at 72°C. PCR products were purified by using magnetic beads (Agencourt Ampure; Beckman Coulter), and their purity was controlled by analyzing aliquots on 1% agarose gels. The purified PCR products were sequenced on both DNA strands. Sequencing reactions were performed by using a BigDye Terminator v1.1 cycle sequencing kit (Applied Biosystems). The obtained products were purified and sequenced at the Genotyping and Sequencing Facility of Bordeaux.

View this table: [in this window] [in a new window]

TABLE 2. Genes and primers used for MLST

MLST locus selection.
MLST analyses traditionally focus on allelic diversity of housekeeping genes, which match admitted criteria, including their presence as a single copy in all strains, their conserved sequence, and their wide distribution across the chromosome. The genes gyrB, ddl, pgm, and recP were selected according to a previous MLST analysis of O. oeni (5). The other genes were searched in previous MLST studies of gram-positive bacteria (http://www.MLST.net), and g6pd, purK, dnaE, and rpoB were selected on the basis of their positions on the chromosome and the numbers of polymorphic sites detected by comparing gene sequences from the O. oeni PSU-1 (CP000411) and ATCC BAA-1163 (AAUV00000000) available genomes.

MLST data treatment and bioinformatic analyses.
Analysis, editing, and comparison of the 688 chromatograms and sequences obtained for the eight genes of the MLST scheme and the 43 bacterial strains were performed by using Bionumerics 5.1. Each distinct gene sequence was assigned an allele number, and each unique combination of eight allele numbers was assigned a sequence type (ST). The same ST was used for several strains when they share the same allelic profile.

Descriptive analyses of the genetic variability at MLST loci such as the determination of the mean G+C content, the number of polymorphic sites, and the nucleotide diversity were performed using DnaSP 4.5 (37). The same software was also used to calculate the d_N/d_S ratio (where d_N is the number of nonsynonymous substitutions per nonsynonymous site and d_S is the number of synonymous substitutions per synonymous site) and to perform the Tajima's D neutrality test (43).

The phylogenetic trees were constructed by the neighbor-joining method with a Kimura two-parameter distance model using MEGA 4 (44) or by the maximum-likelihood method with the HKY 85 model using Tree-Puzzle 5.2 (39). Bootstrap values were obtained after 1,000 replicates. Congruencies between maximum-likelihood trees were statistically assessed by performing the Shimodaira-Hasegawa test implemented in the CONSEL software (42).

Clonal complexes of strains were investigated by a minimum spanning tree analysis, based on number of mutations between concatenated sequences, using Bionumerics 5.1. Strain relationships were also analyzed using the eBURST program, which focuses on number of variable loci between allelic profiles (11).

The standardized index of association (I^S_A) was calculated to determine the degree of linkage disequilibrium between alleles using START 2 (19). The split decomposition method was used to assess the degree of tree-like structure for alleles of each locus or for concatenated sequences using SPLITSTREE 4.1 (17). A compatibility matrix of all of the informative sites identified in MLST loci was generated by using the RETICULATE program (18). Recombination events were searched between sequences of single and concatenated loci using seven algorithms (RDP, Geneconv, BootScan, Maximum ², 3Seq, Chimaera, and Sister Scanning) implemented in the RDP 3.27 software (30). Only recombination events detected by at least three methods and involving parental sequences present in the MLST data set were considered.

Nucleotide sequence accession numbers.
Nucleotide sequences of MLST loci were deposited in GenBank under accession numbers FJ392687 to FJ392689 (ddl), FJ392690 to FJ392698 (dnaE), FJ403333 to FJ403343 (g6pd), FJ403344 to FJ403350 (pgm), FJ403351 to FJ403361 (purK), FJ403362 to FJ403370 (recP), FJ403371 to FJ403373 (rpoB), and FJ413033 to FJ413040 (gyrB).

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
PFGE analysis of 43 O. oeni strains.
In order to compare the discriminatory ability of PFGE and MLST, 43 strains of O. oeni expected to be as diverse as possible were selected on the basis of their isolation date (1972 to 2003), origin (Australia, France, Greece, Italy, Spain, and the United States), and type of product (red wine, white wine, champagne, fortified wine, grape fruit, and cider) (Table 1). Nine strains were commercial starters from six different companies. One strain of Oenococcus kitaharae, the only other species of the genus Oenococcus identified to date (10), was included in the collection for comparison with O. oeni.

All strains were analyzed by PFGE using NotI for macrodigestion of genomic DNAs. The 43 O. oeni displayed 30 well-discriminated DNA patterns composed of 6 to 11 fragments of 24 to 300 kb in size and 1 or 2 fragments greater than 300 kb (Fig. 1). The similarities between DNA patterns did not always reflect the proximity of the strain isolation sites. The DNA pattern of the O. kitaharae strain was strikingly different since it comprised only small fragments not exceeding 170 kb. A UPGMA tree was constructed from PFGE data and rooted with O. kitaharae (Fig. 1). In this tree the 43 O. oeni strains were clearly subdivided into two clusters, named A and B, each comprising 28 and 15 strains of PFGE patterns sharing more than 67.5 and 64.8% similarity, respectively. Interestingly, eight of the nine commercial strains were grouped together in cluster A, along with all strains isolated from fortified wines (Pineau, Banyuls, and Floc) and two strains from champagne. In addition, at least 11 of the 15 strains of cluster B were isolated before 1993 (isolation date is missing for four strains, Table 1), whereas cluster A contained only three strains obtained before this date.

View larger version (96K): [in this window] [in a new window]

FIG. 1. UPGMA tree based on NotI-PFGE macrorestriction patterns of 43 O. oeni isolates.

MLST scheme and genetic diversity at eight loci.
Eight pairs of primers bordering DNA regions of 597 to 714 bp of eight housekeeping genes of O. oeni were designed according to the genome sequences of O. oeni strains PSU-1 (33) and ATCC BAA1163 (NZ_AAUV00000000). Four genes were used in a previous analysis of O. oeni strains: gyrB, ddl, pgm, and recP (5), while the four other genes were previously included in MLST schemes of diverse bacteria species: g6pd, dnaE, purK, and rpoB (Table 2).

The eight genes were successfully amplified and sequenced for all 43 O. oeni strains but not for O. kitaharae, which therefore was not further considered in the MLST analysis. Nucleotide sequences of 496 bp (ddl) to 665 bp (dnaE) were determined (Table 3). Their mean G+C content varied from 36.6% (ddl) to 46.0% (recP), while it is 38% in the whole O. oeni genome (33). The level of nucleotide variation differed greatly among genes (Table 3). Only two polymorphic sites were detected in ddl compared to 48 sites in rpoB. Similarly, the nucleotide diversity (the average number of nucleotide differences per site from two randomly selected sequences) was very low at two loci (ddl, 0.0011; pgm, 0.0014) but high in recP and rpoB (0.0204 and 0.0370, respectively). Taken together, rpoB and recP accounted for more than half of all polymorphic sites (81 of the 156 sites). Despite the high diversity of these two genes the data demonstrated that the overall nucleotide diversity of the 43 strains at the eight loci was low. The d_N/d_S ratios were calculated to estimate the level of selection applied to each gene (the value obtained for ddl was unreliable since it was based on only two polymorphic sites; Table 3). All values did not exceed 0.5, indicating that there was a strong selective pressure against amino acid changes, as typically observed for housekeeping genes. Besides nucleotide variations, we identified a one-base deletion at position 381 in the recP sequence of O. oeni IOEB-SARCO 422 and a 860-bp transposon inserted at position 456 in the purK sequences of strains IOEB-SARCO 422 and 444. Since these modifications disrupted or dramatically changed the open reading frames, there is no doubt that the encoded proteins were not functional. The genetic equilibrium of alleles was analyzed by using Tajima's D neutrality test (43). D values obtained for gyrB, g6pd, pgm, ddl, dnaE, purK, and recP did not deviate significantly from zero, supporting a neutral selection of the alleles of these genes (Table 3). In contrast, the positive D value (3.5) measured for rpoB denoted an important balancing selection, which may explain why this gene presents the highest number of polymorphic sites (n = 48), along with the lowest number of alleles (n = 3).

View this table: [in this window] [in a new window]

TABLE 3. Genetic variability at O. oeni loci

Comparison of MLST and PFGE data.
The number of alleles per locus was rather low in view of the number of strains that were analyzed: only three alleles for ddl and rpoB and a maximum of 11 alleles for dnaE and g6pd (Table 3). Nevertheless, when the eight alleles of each strain were compiled to determine the STs, 34 different STs were obtained from the 43 O. oeni isolates, which denotes a significant allelic diversity (Table 1). A total of 28 STs were assigned to single strains, 3 STs were assigned to two strains (ST-11, -16, and -28), and 3 STs were assigned to three strains (ST-14, -15, and -31). No feature, such as the origin or isolation date, was associated to strains sharing the same ST (Table 1). It is possible that some of these bacteria (isolates with ST-14, -15, -28, and -31) were different isolates of the same strain given that they also share identical PFGE patterns. In only one instance could PFGE differentiate two strains that were not distinguished by MLST (ST-11). Conversely, MLST proved to be more accurate than PFGE for differentiating many strains: ST-3 and -4, ST-8 and -18, ST-16 and -17, ST-20 and -21, ST-25 and -26, and ST-28 and -29 (Fig. 1 and Table 1).

Phylogeny based on MLST data.
The phylogeny of the 43 O. oeni isolates was analyzed by constructing a neighbor-joining tree from the 4,581-bp concatenated sequence of the eight loci (Fig. 2). The tree revealed two major phylogroups, named A and B, strongly supported by bootstrap values, and two descents were also detected: A1, A2, B1, and B2. The phylogroup A included all strains of ST-1 to ST-22 and in B were strains with ST-23 to ST-34. Interestingly, there was a perfect correlation between these two phylogroups and the two PFGE clusters (Fig. 1). To determine whether one of the eight genes used in the concatenated sequence influenced this tree topology, this was compared to the topologies of the eight trees constructed independently from each gene. Six trees based on gyrB, g6pd, ddl, dnaE, purK, and rpoB showed the same topology supporting the two phylogroups, while only two trees derived from pgm and recP showed different topologies (see Fig. S1 in the supplemental material). Therefore, the distribution of O. oeni strains in two distinct groups did not result from the allelic diversity of a single gene but more likely from a general tendency of whole genomes. A statistical comparison of tree topologies performed by the Shimodaira-Hasegawa's test (41) revealed a significant lack of congruence in many pairwise comparisons and particularly for the concatenated sequence (see Table S1 in the supplemental material), indicating that independent evolutionary mechanisms affected the MLST genes and that the phylogeny of O. oeni strains cannot be inferred accurately from only one or a few genes.

View larger version (17K): [in this window] [in a new window]

FIG. 2. Neighbor-joining phylogenetic tree constructed from 34 concatenated nucleotide sequences of eight loci. Bootstrap values above 80% are indicated. Two major phylogroups, designated A and B, and their respective descents, A1, A2, B1, and B2, are mentioned.

Population structure and strain relationships.
To investigate the mechanisms at the origin of the allelic diversity of O. oeni strains, the population structure was analyzed by the minimum spanning tree method (40) with an algorithm improved to take into account the number of mutations between allelic profiles (Bionumerics 5.1). This approach revealed the descents A1, A2, B1, and B2 described above and also seven possible clonal complexes (CC-1 to CC-7), defined as groups of strains that diverge by three or fewer mutations (Fig. 3). The seven CCs contained 2 to 8 strains, for a total of 26 strains (19 STs). No specific feature was detected between strains found in a same CC, except for CC-3 which contained two strains isolated from Italy (ST-21 and ST-22). Besides the seven CCs, 17 strains (15 STs) were displayed as singletons showing 4 to 12 differences with the closest ST. Relationships between O. oeni strains were also analyzed by using the eBURST algorithm (11), which focuses on allelic profiles and identifies CCs by linking single- or double-locus variants. eBURST revealed a picture similar to that obtained by the minimum spanning tree method: a similar ratio of linked and single STs (20/14) and similar CCs (data not shown). The small sizes of CCs and the large number of single STs suggested that during the evolution of O. oeni the expansion of clonal groups by accumulation of punctual mutations is limited, while strains with remote genotypes appear frequently, most likely as a result of genetic recombination.

View larger version (17K): [in this window] [in a new window]

FIG. 3. Minimum spanning tree analysis of 43 O. oeni strains based on the concatenated sequences of eight loci. Each circle corresponds to an ST. Different shadings were used for phylogenetic groups A1 (black), A2 (dark gray), B1 (clear), and B2 (light gray). Circle sizes denote the number of strains sharing the same ST (1, 2, or 3). The number of mutations between STs is indicated. STs that belong to the same clonal complex (CC) are shown as circles grouped in a gray area.

Recombination in O. oeni.
Evidences for recombination were investigated by determining the standardized index of association (I^S_A) which estimates the level of linkage disequilibrium between alleles and varies from 0 to 1 (from recombinant to clonal distribution) (15). I^S_A calculated from the 34 O. oeni STs was low but significantly positive (0.197, P < 0.001), thus confirming that recombination played a key role in the distribution of alleles, although a low allelic linkage was still detectable. This trace of linkage disequilibrium detected in the whole O. oeni population was in fact the result of specific allelic contents in distinct subpopulations of O. oeni, given that I^S_A values calculated from the 22 and 12 STs of subpopulations A and B were not significantly different from zero (0.027 and 0.026; P > 0.001).

The split decomposition method (17) was used to examine the impact of recombination in the eight loci and in the concatenated sequence. Split graphs of each locus showed treelike structures, except for purK and recP, where some networks were detected, indicating that most of the genes were not significantly affected by intragenic recombination (see Fig. S2in the supplemental material). In contrast, the split graph of the concatenated sequence had a "rectangular" network shape where the two subpopulations A and B and their descents were clearly disconnected (Fig. 4), which implies that intergenic recombination events were important during O. oeni evolution and for the emergence of lineages.

View larger version (22K): [in this window] [in a new window]

FIG. 4. Split graph deduced from the concatenated sequences of the eight loci for the 34 STs. Circles indicate the positions of the groups A1, A2, B1, and B2.

To detect genes that could have been involved in recombination events, the phylogenetic compatibility of mutations found in the eight loci was tested by using the algorithm RETICULATE (18). Consistent with the split trees obtained for each locus, very few incompatibilities were detected among mutations located within a same locus, which supports the idea that intragenic recombination was rather low (Fig. 5). In contrast, some incompatibilities were noticed between mutations of gyrB, pgm, g6pd, dnaE, and purK loci and, surprisingly, all mutations in recP were incompatible with almost all mutations of other loci. Therefore, it is very likely that HGT events were responsible for the distribution of recP alleles. Such events were searched among the MLST data set using the seven algorithms implemented in the RDP3 package (30). RDP3 did not detect any trace of intragenic recombination event but disclosed four possible intergenic events, among which three involved recP and one purK (Table 4). Interestingly, each event aroused from one parental sequence of subpopulation A and another one of subpopulation B. In addition, all of the strains of descents A2, B1, and B2 could originate from one of the events involving recP. In agreement with this possibility, a phylogenetic tree constructed by the neighbor-joining method with the concatenated sequences of all loci except recP showed that removing recP makes undetectable the descents A1, A2, B1, and B2 (data not shown).

View larger version (32K): [in this window] [in a new window]

FIG. 5. Compatibility matrix of the 137 informative sites of the eight loci. Highly incompatible sites are indicated by black squares.

View this table: [in this window] [in a new window]

TABLE 4. Detection of possible recombination events among O. oeni strainsa

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
MLST, a valuable method for typing O. oeni strains.
PFGE is generally recognized as the most efficient method for typing O. oeni strains. It is widely used to characterize O. oeni isolates from laboratory collections or commercial strains used by winemakers or malolactic starter producers (13). The main limitation of PFGE is that DNA patterns which are produced may vary with experimental conditions, which limits possibilities of data exchange and comparison. Considering the growing interest for the discrimination of O. oeni strains, there is a need for a reliable and more portable typing method. In a previous study, De Las Rivas et al. (5) showed that MLST might fulfill these requirements. Here we have improved the MLST strategy of the authors by including four new loci in the analysis. The new method was tested on a collection of 43 O. oeni strains and compared to PFGE. Its discriminatory power was slightly better than PFGE: in the set of 43 strains, MLST distinguished 34 STs (24 STs when only genes of the former MLST scheme were considered) instead of 30 DNA patterns of PFGE. According to the method of Hunter (16), this corresponds to discriminatory powers of 0.987 and 0.980 for MLST and PFGE, respectively, where 0.987 represents a probability of 98.7% that two randomly selected strains are discriminated. The new MLST scheme is well adapted for industrial applications since it discriminates all nine commercial strains that were tested, and it produces data within two to three working days, whereas more than 2 weeks are required to produce PFGE patterns due to the slow growth of O. oeni. The evidence supports the use of MLST rather than PFGE for typing O. oeni strains. The method can prove particularly convenient for differentiating O. oeni isolates during the process of selection of new commercial strains. However, when considering industrial applications, it is noteworthy that analyzing a strain or a sample of wine by MLST is more expensive and requires more specialization than PFGE. We suggest an MLST scheme including only seven of the eight genes analyzed in this work since removing the ddl gene does not decrease its discriminatory power. Similarly, the rpoB gene is not very informative (only three alleles); however, the gene sequence does allow segregation of a strain into one of the two subpopulations A or B (discussed below).

Two subpopulations in the species O. oeni.
A phylogenetic tree constructed from the concatenated sequences of the eight loci showed that the 43 O. oeni strains form two distinct phylogroups of 28 and 15 strains that we have designated subpopulations A and B. The detection of two subpopulations was not the result of an evolutionary distortion due to one or a few genes given that it was supported by the topologies of six independent trees constructed from gyrB, g6pd, ddl, dnaE, purK, and rpoB sequences. It denotes more likely a global evolutionary tendency of bacterial genomes. In agreement with this hypothesis, I^s_A values calculated from the whole O. oeni population (0.197) and from each subpopulation (0.026 and 0.027) support the existence of two groups of strains carrying their own allelic contents. PFGE analysis of the 43 strains also provided a strong support to this possibility since it disclosed two groups of strains that are perfectly correlated to the two subpopulations A and B. It is noteworthy that PFGE and MLST analyses target different variations of the genome: large-scale modifications due to genomic rearrangements or insertions or deletion of mobile DNA elements and point mutations in a few genes, respectively. Therefore, strains that belong to the same subpopulation share not only similarities of allelic content but also similarities in genome organization. Interestingly, eight of the nine commercial strains-which are expected to have optimal abilities for performing the malolactic fermentation of wine—were grouped together in subpopulation A. In future studies, it will be useful to examine possible relationships between phenotypic and genotypic traits of strains since this could help to select industrial strains and also to decide if the nontaxonomic but informative concept of subspecies can be used instead of subpopulation.

The detection of two subpopulations differs with a previous MLST analysis of 18 O. oeni strains that revealed only one group (5). However, the number of loci analyzed in this former study was probably insufficient to disclose accurately phylogenetic groups, given that the same 18 strains formed two distinct groups by ribotyping analysis (5). Although only two subpopulations were identified in the set of 43 strains analyzed here, we cannot exclude the possibility that more would have been detected by analyzing a larger collection of strains. However, previous studies often reported two groups of strains. The most recent is the typing of 67 isolates from Germany that were classified in two major groups based on PFGE patterns (22). Similarly, two groups of strains were detected in other studies based on PGFE analysis (45), RAPD analysis (49), ribotyping (5, 49), and metabolic characterization (35). Unfortunately, the data reported in these works were produced by different techniques or even by PFGE analyses, which are difficult to compare. Therefore, it is not yet possible to precisely determine whether the species O. oeni contains several subpopulations or only two that were repeatedly detected in independent works. However, in favor of the latter possibility, it is important to note that the 43 strains that we analyzed were collected from many different sources.

Mutability and evolution of O. oeni genome.
The absence of mutS and mutL genes in O. oeni was correlated to the hypermutability of its genome (29). However, the nucleotide diversity measured at the eight loci of the 43 strains analyzed here was rather low and not as important as expected for a hypermutable genome. Only 3.4% of variable sites were detected in the 4,581-bp concatenated sequence. This is in the same range as values obtained from other LAB: 1 to 7.7% in Lactobacillus plantarum (6), 0 to 2.67% in Pediococcus parvulus (2), and 1.4 to 7.8% in Lactobacillus casei (1). The corresponding allelic diversity reaches only 7.6 alleles per locus, whereas higher values are usually measured. For instance, an analysis of six housekeeping genes in 40 L. casei strains showed a mean value of 13.8 alleles per locus (1). Therefore, it is possible that the hypermutable status concerns only some genes of O. oeni. In the species Mycobacterium tuberculosis that also lacks mut genes (9), housekeeping genes are extremely well conserved, whereas a significant mutation rate was measured for genes involved in DNA repair, replication, and recombination (8). The authors of that study have suggested that the lack of fidelity in genome maintenance would be the starting point of evolution in stressful conditions, such as antibiotic resistance or adaptation to a particular niche. A similar situation could be responsible for the adaptation of O. oeni to the wine environment. Consistent with the low genetic diversity at most MLST loci, the pictures of the O. oeni population structure obtained by eBURST and the minimum spanning tree methods revealed only a few small clonal complexes and a large number of single STs. During the evolution of O. oeni the emergence of clonal descents by accumulation of punctual mutations was limited, while the impact of recombination events was probably much more important and produced many strains with remote genotypes.

Intergenic recombination shapes the O. oeni population.
A critical role of recombination in O. oeni evolution was pointed out in a former MLST analysis (5). These authors suggested that the impact of recombination was so important in this species that it could be an example of panmictic population, i.e., a population where clonal descents are hardly detectable. This conclusion contrasted with other results obtained by DNA-DNA hybridizations, 16S-23S ISR sequences, ribotyping, or RAPD patterns (7, 23, 48, 49) that suggested a genetically homogeneous O. oeni population and a clonal mode of evolution. According to our results, the apparent genetic homogeneity and clonality were most likely due to the limited genetic diversity in O. oeni and the presence of two subpopulations. Although these two subpopulations exist, our data confirmed the importance of recombination in O. oeni evolution. As described above, the two subpopulations represent strains sharing genomic and allelic similarities, but their alleles are widely disseminated and close to linkage equilibrium (I^s_A = 0.026 and 0.027 in subpopulations A and B, respectively). A split graph representation of the concatenated sequence of the eight loci has clearly shown that the two subpopulations and their descents originated from intergenic recombination events. According to analyses made with the RDP3 package, it is possible that the descents A2, B1, and B2 result from events that occurred between strains of each subpopulation involving DNA regions that include the recP gene. To better understand how recombination has modeled the O. oeni genome, it will be interesting to compare the full genomes of O. oeni PSU1 (33) and ATCC BAA 1163 (unfinished genome) that belong to each subpopulations.

 
E.B. was the recipient of a doctoral fellowship from the French Ministry for Higher Education and Research.

We thank C. Miot-Sertier for technical assistance and the Genotyping and Sequencing facility of Bordeaux for performing the sequencing reactions (grants from the Conseil Régional d'Aquitaine [20030304002FA and 20040305003FA] and from the European Union [FEDER 2003227]).

 
* Corresponding author. Mailing address: Université Bordeaux 2, Faculté D'Oenologie, ISVV, 210 Chemin de Leysotte, CS 50008, 33882 Villenave d'Ornon, France. Phone: 33 5 57575833. Fax: 33 5 57575813. E-mail: patrick.lucas{at}u-bordeaux2.fr

Published ahead of print on 29 December 2008.

Supplemental material for this article may be found at http://aem.asm.org/.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1
1. Cai, H., B. T. Rodrguez, W. Zhang, J. R. Broadbent, and J. L. Steele. 2007. Genotypic and phenotypic characterization of Lactobacillus casei strains isolated from different ecological niches suggests frequent recombination and niche specificity. Microbiology 153:2655-2665.[Abstract/Free Full Text]
2
2. Calmin, G., F. Lefort, and L. Belbahri. 2008. Multi-loci sequence typing (MLST) for two lacto-acid bacteria (LAB) species: Pediococcus parvulus and P. damnosus. Mol. Biotechnol. 40:170-179.[CrossRef][Medline]
3
3. Cappello, M. S., D. Stefani, F. Grieco, A. Logrieco, and G. Zapparoli. 2008. Genotyping by amplified fragment length polymorphism and malate metabolism performances of indigenous Oenococcus oeni strains isolated from Primitivo wine. Int. J. Food Microbiol. 127:241-245.[CrossRef][Medline]
4
4. Chelo, I. M., L. Zé-Zé, and R. Tenreiro. 2007. Congruence of evolutionary relationships inside the Leuconostoc-Oenococcus-Weissella clade assessed by phylogenetic analysis of the 16S rRNA gene, dnaA, gyrB, rpoC, and dnaK. Int. J. Syst. Evol. Biol. 57:276-286.[Abstract/Free Full Text]
5
5. De Las Rivas, B., A. Marcobal, and R. Muñoz. 2004. Allelic diversity and population structure in Oenococcus oeni as determined from sequence analysis of housekeeping genes. Appl. Environ. Microbiol. 70:7210-7219.[Abstract/Free Full Text]
6
6. De Las Rivas, B., A. Marcobal, and R. Muñoz. 2006. Development of a multilocus sequence typing method for analysis of Lactobacillus plantarum strains. Microbiology 152:85-93.[Abstract/Free Full Text]
7
7. Dicks, L. T. M., H. J. J. Van Vuuren, and F. Dellaglio. 1990. Taxonomy of Leuconostoc species, particularly Leuconostoc oenos, as revealed by numerical analysis of total soluble cell protein pattern, DNA base composition and DNA-DNA hybridization. Int. J. Syst. Bacteriol. 40:83-91.[Abstract/Free Full Text]
8
8. Dos Vultos, T., O. Mestre, J. Rauzier, M. Golec, N. Rastogi, V. Rasolofo, T. Tonjum, C. Sola, I. Matic, and B. Gicquel. 2008. Evolution and diversity of clonal bacteria: the paradigm of Mycobacterium tuberculosis. PLoS ONE 3:e1538.[CrossRef][Medline]
9
9. Eisen, J. A., and P. C. Hanawalt. 1999. A phylogenomic study of DNA repair genes, proteins, and processes. Mutat. Res. 435:171-213.[Medline]
10
10. Endo, A., and S. Okada. 2006. Oenococcus kitaharae sp. nov., a non-acidophilic and non-malolactic-fermenting oenococcus isolated from a composting distilled shochu residue. Int. J. Syst. Evol. Microbiol. 56:2345-2348.[Abstract/Free Full Text]
11
11. Feil, E. J., B. C. Li, D. M. Aanensen, W. P. Hanage, and B. G. Spratt. 2004. eBURST: inferring patterns of evolutionary descent among clusters of related bacterial genotypes from multilocus sequence typing data. J. Bacteriol. 186:1518-1530.[Abstract/Free Full Text]
12
12. Garvie, E. I. 1967. Leuconostoc oenos sp. nov. J. Gen. Microbiol. 48:431-438.[Abstract/Free Full Text]
13
13. Gindreau, E., A. Joyeux, G. de Revel, O. Claisse, and A. Lonvaud-Funel. 1997. Evaluation of the settling of malolactic starters within the indigenous microflora of wines. J. Int. Sci. Vigne Vin. 31:197-202.
14
14. Guerrini, S. A. Bastianini, G. Blaiotta, L. Granchi, G. Moschetti, S. Coppola, P. Romano, and M. Vincenzini. 2003. Phenotypic and genotypic characterization of Oenococcus oeni strains isolated from Italian wines. Int. J. Food Microbiol. 83:1-14.[CrossRef][Medline]
15
15. Haubold, B., and R. R. Hudson. 2000. LIAN 3.0: detecting linkage disequilibrium in multilocus data: linkage analysis. Bioinformatics 16:847-848.[Abstract/Free Full Text]
16
16. Hunter, P. R. 1990. Reproducibility and indices of discriminatory power of microbial typing methods. J. Clin. Microbiol. 28:1903-1905.[Abstract/Free Full Text]
17
17. Huson, D. H., and D. Bryant. 2006. Application of phylogenetic networks in evolutionary studies. Mol. Biol. Evol. 23:254-267.[Abstract/Free Full Text]
18
18. Jakobsen, I. B., and S. Easteal. 1996. A program for calculating and displaying compatibility matrices as an aid in determining reticulate evolution in molecular sequences. Comput. Appl. Biosci. 12:291-295.[Abstract/Free Full Text]
19
19. Jolley, K. A., E. J. Feil, M. S. Chan, and M. C. Maiden. 2001. Sequence type analysis and recombinational tests (START). Bioinformatics 17:1230-1231.[Abstract/Free Full Text]
20
20. Kelly, W. J., C. M. Huang, and R. V. Asmundson. 1993. Comparison of Leuconostoc oenos strains by pulsed-field gel electrophoresis. Appl. Environ. Microbiol. 59:3969-3972.[Abstract/Free Full Text]
21
21. Lamoureux, M., H. Prévost, J. F. Cavin, and C. Diviès. 1993. Recognition of Leuconostoc oenos strains by the use of DNA restriction profiles. Appl. Microbiol. Biotechnol. 39:547-552.[CrossRef][Medline]
22
22. Larisika, M., H. Claus, and H. König. 2008. Pulsed-field gel electrophoresis for the discrimination of Oenococcus oeni isolates from different wine-growing regions in Germany. Int. J. Food Microbiol. 123:171-176.[CrossRef][Medline]
23
23. Le Jeune, C., and A. Lonvaud-Funel. 1997. Sequence of DNA 16S/23S spacer region of Leuconostoc oenos (Oenococcus oeni): application to strain differentiation. Res. Microbiol. 148:79-86.[Medline]
24
24. Lonvaud-Funel, A. 1999. Lactic acid bacteria in the quality improvement and depreciation of wine. Antonie van Leeuwenhoek 76:317-331.[CrossRef][Medline]
25
25. Maiden, M. C., J. A. Bygraves, E. Feil, G. Morelli, J. E. Russell, R. Urwin, Q. Zhang, J. Zhou, K. Zurth, D. A. Caugant, I. M. Feavers, M. Achtman, and B. G. Spratt. 1998. Multilocus sequence typing: a portable approach to the identification of clones within populations of pathogenic microorganisms. Proc. Natl. Acad. Sci. USA 95:3140-3145.[Abstract/Free Full Text]
26
26. Maiden, M. C. 2006. Multilocus sequence typing of bacteria. Annu. Rev. Microbiol. 60:561-588.[CrossRef][Medline]
27
27. Makarova, K., A. Slesarev, Y. Wolf, A. Sorokin, B. Mirkin, E. Koonin, A. Pavlov, N. Pavlova, V. Karamychev, N. Polouchine, V. Shakhova, I. Grigoriev, Y. Lou, D. Rohksar, S. Lucas, K. Huang, D. M. Goodstein, T. Hawkins, V. Plengvidhya, D. Welker, J. Hugues, Y. Goh, A. Benson, K. Baldwin, J. H. Lee, I. Diaz-Muniz, B. Dosti, V. Smeianov, W. Wechter, R. Barabote, G. Lorca, E. Altermann, R. Barrangou, B. Ganesan, Y. Xie, H. Rawsthorne, D. Tamir, C. Parker, F. Breidt, J. Broadbent, R. Hutkins, D. O'Sullivan, J. Steele, G. Unlu, M. Saier, T. Klaenhammer, P. Richardson, S. Kozyavkin, B. Weimer, and D. Mills. 2006. Comparative genomics of the lactic acid bacteria. Proc. Natl. Acad. Sci. USA 103:15611-15616.[Abstract/Free Full Text]
28
28. Makarova, K., and E. Koonin. 2007. Evolutionary genomics of lactic acid bacteria. J. Bacteriol. 189:1199-1208.[Free Full Text]
29
29. Marcobal, A. M., D. A. Sela, Y. I. Wolf, K. S. Makarova, and D. A. Mills. 2008. Role of hypermutability in the evolution of the genus Oenococcus. J. Bacteriol. 190:564-570.[Abstract/Free Full Text]
30
30. Martin, D. P., C. Williamson, and D. Posada. 2005. RDP2: recombination detection and analysis from sequence alignments. Bioinformatics 21:260-262.[Abstract/Free Full Text]
31
31. Martinez-Murcia, A. J., and M. D. Collins. 1990. A phylogenetic analysis of the genus Leuconostoc based on reverse transcriptase sequencing of 16S rRNA. FEMS Microbiol. Lett. 70:73-84.[CrossRef]
32
32. Martinez-Murcia, A. J., N. M. Harland, and M. D. Collins. 1993. Phylogenetic analysis of some Leuconostoc and related organisms as determined from large-subunit rRNA gene sequences, assessment of congruence of small and large-subunit rRNA derived trees. J. Appl. Bacteriol. 74:532-541.[Medline]
33
33. Mills, D. A., H. Rawsthorne, C. Parker, D. Tamir, and K. Makarova. 2005. Genomic analysis of Oenococcus oeni PSU-1 and its relevance to winemaking. FEMS Microbiol. Rev. 29:465-475.[CrossRef][Medline]
34
34. Morse, R., M. D. Collins, K. O'Hanlon, S. Wallbanks, and P. T. Richardson. 1996. Analysis of the β' subunit of DNA-dependent RNA polymerase does not support the hypothesis inferred from 16S rRNA analysis that Oenococcus oeni (formerly Leuconostoc oenos) is a tachytelic (fast-evolving) bacterium. Int. J. Syst. Bacteriol. 46:1004-1009.[Abstract/Free Full Text]
35
35. Peynaud, E., and S. Domercq. 1968. Etude de quatre cents souches de coques hétérolactiques isolées de vins. Ann. Inst. Pasteur 19:159-169.
36
36. Reguant, R., and C. Bordons. 2003. Typification of Oenococcus oeni strains by multiplex RAPD-PCR and study of population dynamics during malolactic fermentation. J. Appl. Microbiol. 95:344-353.[CrossRef][Medline]
37
37. Rozas, J., J. C. Sánchez-DelBarrio, X. Messeguer, and R. Rozas. 2003. DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics 19:2496-2497.[Abstract/Free Full Text]
38
38. Sato, H., F. Yanagida, T. Shinohara, M. Suzuki, K. Suzuki, and K. Yokotsuka. 2001. Intraspecific diversity of Oenococcus oeni isolated during red wine-making in Japan. FEMS Microbiol. Lett. 202:109-114.[CrossRef][Medline]
39
39. Schmidt, H. A., K. Strimmer, M. Vingron, and A. von Haeseler. 2002. TREE-PUZZLE: maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioinformatics 18:502-504.[Abstract/Free Full Text]
40
40. Schouls, L. M., H. G. van der Heide, L. Vauterin, P. Vauterin, and F. R. Mooi. 2004. Multiple-locus variable-number tandem repeat analysis of Dutch Bordetella pertussis strains reveals rapid genetic changes with clonal expansion during the late 1990s. J. Bacteriol. 186:5496-5505.[Abstract/Free Full Text]
41
41. Shimodaira, H., and M. Hasegawa. 1999. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Mol. Biol. Evol. 16:1114-1116.
42
42. Shimodaira, H., and M. Hasegawa. 2001. CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinformatics 17:1246-1247.[Abstract/Free Full Text]
43
43. Tajima, F. 1989. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123:585-595.[Abstract/Free Full Text]
44
44. Tamura, K., J. Dudley, M. Nei, and S. Kumar. 2007. MEGA4: molecular evolutionary genetics analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24:1596-1599.[Abstract/Free Full Text]
45
45. Tenreiro, R. M. A. Santos, H. Pavela, and G. Vleira. 1994. Inter-strain relationships among wine leuconostocs and their divergence from other Leuconostoc species, as revealed by low frequency restriction fragment analysis of genomic DNA. J. Appl. Bacteriol. 77:271-280.[Medline]
46
46. Viti, C., L. Giovannetti, L. Granchi, and S. Ventura. 1996. Species attribution and strain typing of Oenococcus oeni (formerly Leuconostoc oenos) with restriction endonuclease fingerprints. Res. Microbiol. 147:651-660.[Medline]
47
47. Yang, D., and C. R. Woese. 1989. Phylogenetic structure of the leuconostocs: an interesting case of a rapidly evolving organism. Syst. Appl. Microbiol. 12:145-149.
48
48. Zavaleta, A. I., A. J. Martínez-Murcia, and F. Rodríguez-Valera. 1996. 16S-23S DNA intergenic sequences indicate that Leuconostoc oenos is phylogenetically homogeneous. Microbiology 142:2105-2114.[Abstract/Free Full Text]
49
49. Zavaleta, A. I., A. J. Martínez-Murcia, and F. Rodríguez-Valera. 1997. Intraspecific genetic diversity of Oenococcus oeni as derived from DNA fingerprinting and sequence analyses. Appl. Environ. Microbiol. 63:1261-1267.[Abstract]
50
50. Zé-Zé, R., L. Tenreiro, and H. Paveia. 2000. The Oenococcus oeni genome: physical and genetic mapping of strain GM and comparison with the genome of a ‘divergent’ strain, PSU-1. Microbiology 146:3195-3204.[Abstract/Free Full Text]

---

Applied and Environmental Microbiology, March 2009, p. 1291-1300, Vol. 75, No. 5
0099-2240/09/$08.00+0     doi:10.1128/AEM.02563-08
Copyright © 2009, American Society for Microbiology. All Rights Reserved.

This Article Abstract Full Text (PDF) Supplemental material Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Bilhère, E. Articles by Lonvaud-Funel, A. Search for Related Content PubMed PubMed Citation Articles by Bilhère, E. Articles by Lonvaud-Funel, A. Agricola Articles by Bilhère, E. Articles by Lonvaud-Funel, A.