Introduction

Lactic acid bacteria (LAB) are a group of Gram-positive bacteria occurring on grapes and in wine during malolactic fermentation (MLF). The latter process, carried out mainly by Oenococcus oeni and Lactobacillus plantarum, involves the conversion of l-malic acid to l-lactic acid and CO2. This reaction contributes to wine acidity reduction, increased microbial stability and changes in the sensory attributes of wine (Bartowsky 2005; Lerm et al. 2010). LAB are generally fastidious microorganisms with multiple amino acid requirements, thus requiring the proteolytic system necessary to degrade proteins (present in the environment in which they occur) into smaller peptides and amino acids crucial for their growth (Juillard et al. 1995). In the dairy industry, for example, the proteolytic activity of bacterial species also favours the development of texture and flavour of fermented milk products (Lynch et al. 1999; El-Soda et al. 2000; Hynes et al. 2003).

While the proteolytic system of Lactococcus lactis, a model microorganism used as the starter culture in the dairy industry, has been extensively investigated (Tan et al. 1993; Tynkkynen et al. 1993; Hagting et al. 1994; Foucaud et al. 1995; Kunji et al. 1996), the proteolytic enzymes of bacterial species of wine origin remain poorly understood. Nevertheless, the production of exocellular proteinases by O. oeni strains has been detected (Rollan et al. 1993; Manca de Nadra et al. 1997, 1999; Remize et al. 2005), albeit Davis et al. (1988) observed no protease production by several wine LAB strains of Oenococcus, Lactobacillus and Pediococcus. More recently, Strahinic et al. (2010) demonstrated the presence of a protease enzyme-encoding gene (prtP) in a natural isolate of L. plantarum originating from home-made cheese. This gene encodes an enzyme that participates in protein degradation to form small peptides and amino acids. The resulting peptides are subsequently translocated into the cell via the specific transport system where they are acted upon by a variety of peptidases (Kunji et al. 1996).

Following the primary and secondary proteolysis, some peptidases are able to release sulphur-containing amino acids from peptide degradation (Kunji et al. 1996). In the dairy industry, the catabolism of sulphur amino acids by microbial enzymes plays a key role in the development of a typical cheese flavour (Dias and Weimer 1999). The LAB originating from cheese are also known to metabolize methionine into methanethiol, a volatile sulphur compound contributing to cheese flavour (Weimer et al. 1999). In wine, Pripis-Nicolau et al. (2004) showed that the LAB isolated from wine could degrade methionine to form volatile sulphur compounds. At a molecular level, Knoll et al. (2011) recently reported on the cloning and characterization of a cystathionine lyase gene from two Oenococcus oeni strains originating from wine. The latter authors also observed that O. oeni seemed to play a minor role in the production of volatile sulphur compounds during winemaking. This is the extent of our knowledge regarding the enzymes acting on sulphur-containing amino acids and, in addition, the distribution of genes encoding a variety of peptidases in wine-associated LAB strains has still not been thoroughly investigated.

The aim of this study was, therefore, to investigate the presence of genes encoding enzymes involved in the proteolytic and amino acid catabolic pathways in lactobacilli and pediococci of oenological origin. Reference strains of non-wine origin that are representative of different LAB species were also included in the study. The genes tested for included those encoding peptide transporter (DtpT), aminopeptidases (PepC, PepN, PepM), endopeptidase (PepO), tripeptidase (PepT), proline peptidase (PepX, PepI), as well as enzymes involved in the catabolism of sulphur-containing amino acids.

Materials and methods

Bacterial strains and growth conditions

The bacterial strains used in this work were previously isolated from grape and wine samples obtained from five different commercial wineries situated in the Western Cape region, South Africa, during the 2001 and 2002 harvest seasons (Krieling 2003). Taxonomic identification of these isolates was based on the results of PCR analysis with species-specific primers and 16S rDNA sequence analysis. The type and reference strains included in this work are presented in Table 1. All of the strains were cultured at 30 °C in MRS medium (BIOLAB Diagnostics, Wadeville, South Africa).

Table 1 Reference or type strains included in this work
Full size table

Primer design

To design the primers for PCR assays, we aligned nucleotide sequences of the target genes from several LAB species belonging to four LAB genera: Lactobacillus, Leuconostoc, Pediococcus and Oenococcus. Two conserved domains of each gene were then selected to design the primers for PCR amplifications. All PCR primers used in this study (Table 2) were synthesized by Integrated DNA Technologies. (Coralville, IA).

Table 2 PCR primers designed in this study, annealing temperatures and amplicon sizes
Full size table

Colony PCR assays

PCR amplifications were performed in a Biometra Thermocycler machine (Biometra® GmbH, Göttingen, Germany). The 25-μl PCR mixture comprised a template DNA, 0.6 μM of each primer, 1.5 mM MgCl2, 250 μM of dNTP mix, 1× PCR buffer and 1.25 units of Supertherm DNA polymerase (Southern Cross Biotechnology, Cape Town, South Africa). The reaction conditions were 94 °C for 5 min, followed by 35 cycles of 94 °C for 1 min, 49–54 °C for 45 s and 72 °C for 1 min (specific annealing temperatures are indicated in Table 2), with a final extension step at 72 °C for 10 min. PCR fragments were checked by gel electrophoresis in 1 % w/v agarose gels containing ethidium bromide (0.5 μg ml-1) and documented with Alpha Imager software (Alpha Innotech Corporation, San Leandro, CA).

Sequence verification

To verify whether the PCR-generated amplicons corresponded to the target genes, the L. plantarum IWBT B349 strain was chosen for sequencing. DNA fragments from this strain were amplified with the same primers in a 50-μl reaction containing TaKaRa Ex Taq™ polymerase (Southern Cross Biotechnology) instead of Supertherm polymerase. Amplicons were cleaned with the QIAquick® PCR Purification kit (Qiagen, Southern Cross Biotechnology) and subsequently sequenced using the corresponding forward and reverse primers. The GenBank database was searched for homologous DNA sequences.

Phylogenetic analysis

In an attempt to study the phylogenetic relationship between the LAB strains tested in this study and closely related taxa, we randomly selected five strains from which the pepC and pepM genes were amplified and sequenced. The obtained nucleotide gene sequence data were assembled and compared to the database sequences. The phylogenetic trees were inferred using the neighbour-joining method (Saitou and Nei 1987). The evolutionary distances were computed using Kimura 2-parameter method (Kimura 1980) from MEGA 4.1 software (Kumar et al. 2008) and are given in the units of the number of base substitutions per site. Bootstrapping analysis was performed to evaluate the reliability of topologies of the constructed phylogenetic trees using 1000 bootstrap replications (Felsenstein 1985).

Results

PCR detection of genes

A total of 104 strains of lactobacilli and pediococci were tested in this study. Using a PCR detection approach with gene-specific primers, we screened these strains for the presence of genes encoding peptidases and enzymes involved in the catabolism of sulphur-containing amino acids. PCR assays generated amplicons of the expected sizes (Table 2), and the different enzyme-encoding genes detected by the PCR assays are presented in Table 3. Among the lactobacilli screened, more than 80 % possessed all of the genes tested for. It is also worth noting that the pepC and pepM genes were present in all of the Lactobacillus species tested and that although L. plantarum strains seemed to possess different combinations of the tested genes, other Lactobacillus species, including Lactobacillus casei, L. curvatus, L. delbrueckii and L. reuteri, possessed all of the genes tested for. Among the L. fermentum strains tested, only the reference strain (DSM 20052) did not possess the pepO and pepX genes. All ten strains of pediococci possessed the dtpT, pepC, pepI, pepN and pepM genes. In contrast, the pepX and metB/metC genes were not detected in any of the Pediococcus strains tested.

Table 3 Results of PCR detection of genes with enzyme-specific primers
Full size table

Sequence verification

PCR amplifications of the IWBT B349 strain of L. plantarum were verified by sequencing using the same sets of primers used for the PCR amplifications. The newly determined sequence data were first assembled and aligned, and then a comparative analysis of the DNA sequences was performed. A homology search was performed by comparing these sequences to other DNA sequences published in GenBank database. The BLAST results confirmed that the sequenced fragments corresponded to the genes tested. The highest homology of nucleotide sequences was recorded with three L. plantarum strains (JDM1, WCFS1 and ATCC 14917) published in the NCBI databank (Table 4). The comparative analysis of sequences revealed that IWBT B349 showed a higher sequence identity to the JDM1 strain than WCFS1 and ATCC 14917.

Table 4 Highest identities found between DNA sequences of the Lactobacillus plantarum IWBT B349 isolate and other L. plantarum strains published in the NCBI databasea
Full size table

Phylogenetic analysis

The phylogenetic relationship between bacterial strains was also studied by constructing neighbour-joining trees based on nucleotide sequences of the pepC and pepM genes as examples. Five strains belonging to L. plantarum, Lactobacillus paracasei, Lactobacillus paraplantarum, Pediococcus parvulus and Pediococcus pentosaceus were randomly selected, from which the pepC and pepM genes were PCR-amplified and subsequently sequenced. Based on the phylogenies inferred with nucleotide sequences of the two selected genes (Fig. 1), there was a similar trend of clustering between bacterial species—a conspicuous feature observed for both phylogenies derived from pepC and pepM genes. For example, L. plantarum IWBT B349 clustered with other L. plantarum strains (WCFS1 and JDM1) whose nucleotide gene sequences were retrieved from GenBank database. Similarly, P. pentosaceus LMG 13561 also clustered with ATCC 25745 strain of the same species.

Fig. 1
figure 1

Phylogenetic tree showing the relationships between various lactic acid bacteria strains tested in this study (in bold) and closely related species (GenBank accession numbers in parentheses) based on partial pepC (a) and pepM (b) gene sequences. The tree was constructed by the neighbour-joining method with Kimura’s two-parameter correction model. Bootstrap values of >80 % are shown at branching points. The horizontal scale bar represents the number of base substitutions per site

Full size image

Discussion

The proteolytic system has been studied extensively in lactococci. This system consists of protein breakdown, peptide transport and hydrolysis (Kunji et al. 1996). The degradation of proteins is initiated by an extracellular cell-wall-associated proteinase (PrtP) enzyme that releases small peptides and amino acids essential for bacterial growth. The degradation products of the proteins are then transported across the cell membrane into the cell via the peptide transporters (e.g. DtpT, DtpP, Opp). Once inside the cell, these peptides are further hydrolyzed into amino acids by a variety of intracellular peptidases (Kunji et al. 1996; Christensen et al. 1999).

Strahinic et al. (2010) reported on the presence of prtP proteinase gene in a natural strain of L. plantarum isolated from home-made cheese. In our study, we screened different oenological species of lactobacilli and pediococci for the presence of genes encoding various peptidases. These peptidases can be classified into aminopeptidases (PepC, PepN, PepM, PepA), endopeptidases (PepO, PepF), dipeptidases (PepD, PepV), tripeptidase (PepT) and proline-specific peptidases (PepX, PepI, PepR, PepP, PepQ) (Liu et al. 2010). Among the bacterial strains tested in this study, it was shown that peptidase genes are present in the majority of different species. The presence of the dtpT gene encoding a di-/tripeptide transporter was also detected in all pediococci strains as well as in the majority of lactobacilli (>90 %).

In order to complement the amino acid auxotrophy, LAB require a range of proteolytic and peptidolytic enzymes to enable them to garner amino acids from the proteins and peptides. As such, wine also contains proteins which can ultimately be degraded by protein-utilizing LAB to produce smaller peptides and amino acids. Other studies have indicated that wine peptides act as carbon and nitrogen substrates for bacterial growth (Aredes Fernandez et al. 2004). The presence of peptidolytic enzyme-encoding genes in wine lactobacilli and pediococci strains tested in this study has implications on the genetic capability of these strains to release amino acids from peptides. This can also favour the development of wine flavour, as has been reported for dairy LAB (El-Soda et al. 2000; Hynes et al. 2003).

Following the breakdown of peptides by various peptidases, sulphur-containing amino acids can also be released (Kunji et al. 1996). These sulphurous amino acids have been shown to enhance flavour formation in various dairy products, such as cheese (Dias and Weimer 1999). In our study, we also tested the LAB strains for the presence of enzyme-encoding genes involved in the catabolism of sulphur-containing amino acids. Different Lactobacillus species displayed the presence of these genes (i.e. metK, gshR and metB/metC), albeit not all the strains tested possessed the genes. Some strains of pediococci only possessed the metK and gshR genes, whereas the metB/metC gene was not detected in any of the Pediococcus strains tested. These findings suggest that Pediococcus strains differ with regard to their amino acid utilization and that the presence of genes in some strains is rather strain-dependent. On the other hand, the genomes of O. oeni oenological strains PSU-1 (accession no. CP000411), ATCC BAA-1163 (accession no. AAUV00000000) and AWRI B429 (accession no. ACSE00000000) possess metK, gshR, metB and metC genes, suggesting that these strains have the genetic potential to metabolize sulphur-containing amino acids. This was also evidenced in a recent study (Knoll et al. 2011) in which two oenological strains of O. oeni were shown to contribute to the formation of volatile sulphur compounds, although the role played by these strains during vinification was insignificant.

Conclusion

To conclude, the results presented in this study provide an indication that Lactobacillus and Pediococcus strains, particularly those of wine origin, possess the genetic potential to degrade peptides and amino acids during winemaking. It should also be noted that certain genes were not present in some strains, suggesting that the presence of the tested enzyme-encoding genes is rather strain-dependent. Nevertheless, it remains to be further confirmed if the tested enzyme-encoding genes are active under oenological conditions and that they are not repressed by winemaking parameters such as pH, temperature and ethanol.