Abstract
Improving the annotation of sugar catabolism-related genes requires functional characterization. Our objective was to identify the genes necessary for lactose utilization by Lactobacillus gasseri ATCC 33323 (NCK334). The mechanism of lactose transport in many lactobacilli is a lactose/galactose-specific permease, yet no orthologue was found in NCK334. Characterization of an EI knockout strain [EI (enzyme I) is required for phosphotransferase system transporter (PTS) function] demonstrated that L. gasseri requires PTS(s) to utilize lactose. In order to determine which PTS(s) were necessary for lactose utilization, we compared transcript expression profiles in response to lactose for the 15 complete PTSs identified in the NCK334 genome. PTS 6CB (LGAS_343) and PTS 8C (LGAS_497) were induced in the presence of lactose 107- and 53-fold, respectively. However, L. gasseri ATCC 33323 PTS 6CB, PTS 8C had a growth rate similar to that of the wild-type on semisynthetic deMan, Rogosa, Sharpe (MRS) medium with lactose. Expression profiles of L. gasseri ATCC 33323 PTS 6CB, PTS 8C in response to lactose identified PTS 9BC (LGAS_501) as 373-fold induced, whereas PTS 9BC was not induced in NCK334. Elimination of growth on lactose required the inactivation of both PTS 6CB and PTS 9BC. Among the six candidate phospho-β-galactosidase genes present in the NCK334 genome, LGAS_344 was found to be induced 156-fold in the presence of lactose. In conclusion, we have determined that: (1) NCK334 uses a PTS to import lactose; (2) PTS 6CB and PTS 8C gene expression is strongly induced by lactose; and (3) elimination of PTS 6CB and PTS 9BC is required to prevent growth on lactose.
-
Edited by: D. A. Mills
Introduction
Lactic acid bacteria (LAB) are a group of related organisms that produce lactic acid through carbohydrate fermentation. The probiotic potential of this bacterial group has been shown to be due to the ability of specific strains to survive passage through the human gastrointestinal tract (GIT) and exert benefits upon the general health and wellness of the host (Liévin-Le Moal & Servin, 2006). Probiotics have been defined as live micro-organisms that, when administered in adequate amounts, confer a health benefit to the host (Reid et al., 2003). Some of these benefits include a positive influence on the normal microbiota present in the GIT, the competitive exclusion of pathogens, and the stimulation or adjustment of mucosal immunity (Ouwehand et al., 2002). Amongst LAB, Lactobacillus gasseri is considered to be indigenous to the human GIT and have a favourable influence on human health (Azcarate-Peril et al., 2008). Recently, the L. gasseri ATCC 33323 genome was published and revealed that carbohydrate utilization resembles that of Lactobacillus johnsonii, yet the specific function of many of these genes is unknown (Azcarate-Peril et al., 2008).
Lactose metabolism, the primary sugar metabolism of LAB, has been well characterized (Wood & Warner, 2003). Lactose utilization occurs through two different pathways in this bacterial group. The first pathway, which is more common in characterized lactobacilli, employs a carbohydrate-specific permease (Barrangou et al., 2003; Leong-Morgenthaler et al., 1991). Once imported, an intracellular β-galactosidase hydrolyses the disaccharide lactose into galactose and glucose (Premi et al., 1972). Galactose can then be further metabolized by the Leloir pathway (Kandler, 1983), while glucose can enter glycolysis.
The second pathway of lactose utilization occurs through the use of lactose-specific phosphoenolpyruvate-dependent phosphotransferase system transporters (PTSs). PTS-dependent utilization of lactose appears to be common in many LAB, including Lactococcus lactis subsp. lactis (McKay et al., 1970; de Vos et al., 1990) and Streptococcus mutans (Honeyman & Curtiss, 1992). The only lactobacillus with a characterized lactose PTS is Lactobacillus zeae (Alpert & Chassy, 1990; Chassy et al., 1976; Chassy & Thompson, 1983; Gosalbes et al., 1999). However, lactose-specific PTS genes have recently been indentified using bioinformatics in Lactobacillus sakei (Nyquist et al., 2011) and Lactobacillus iners (Macklaim et al., 2011), suggesting that lactose-specific PTSs are not rare amongst lactobacilli.
The PTS functions by the transfer of a phosphate group from phosphoenolpyruvate to the incoming sugar through a series of sequential steps that involve the different components of the PTS. The PTS consists of cytoplasmic components, which lack sugar specificity, and membrane-associated enzymes, which are specific for a few sugars, at most. The cytoplasmic components are enzyme I (EI) and a histidine-phosphorylatable protein (HPr). The membranous component of the PTS system, enzyme II (EII), is made up of three to four subunits: IIA, IIB, IIC and sometimes IID. The PTS system phosphorylates carbohydrates as they enter the cell (Reizer & Saier, 1997). The resulting lactose phosphate is hydrolysed by phospho-β-galactosidase, which results in glucose and galactose 6-phosphate (Kandler, 1983). The galactose 6-phosphate can be further metabolized by the tagatose pathway (Kandler, 1983).
Phospho-β-galactosidase and β-galactosidase are members of the glycoside hydrolase family 1 (GH1). In lactobacilli, phospho-β-galactosidase and β-galactosidase are distinct enzymes with minimal overlapping activity (Honda et al., 2007). While the structures of the active sites are highly conserved in this family (Schulte & Hengstenberg, 2000), only eight active-site residues are responsible for the substrate specificity (Hill & Reilly, 2008), making it difficult to predict the specific activity of an uncharacterized GH1 protein. Consequently, the correct annotation of phospho-β-galactosidase and β-galactosidase ORFs in lactobacillus genomes is difficult, leading to general annotations which include both activities. For example, six ORFs are annotated as ‘β-glucosidase/6-phospho-β-glucosidase/β-galactosidase’ in the publicly available annotation of L. gasseri ATCC 33323 (NCK334).
The currently available annotation of the NCK334 genome describes numerous genes potentially involved in the uptake and metabolism of lactose, yet the specific functions of these genes remain unknown. While NCK334 can utilize lactose, the majority of L. gasseri strains tested are unable to utilize lactose (Azcarate-Peril et al., 2008). Our objective was to characterize lactose utilization by NCK334. Bioinformatic analysis failed to identify a lactose/galactose-specific permease in the NCK334 genome. The involvement of a PTS in lactose utilization was investigated.
Methods
Bacterial strains, plasmids and growth conditions.
The bacterial strains and plasmids used in this study are listed in Table 1. NCK334 was grown at 37 °C, with deMan, Rogosa, Sharpe (MRS) broth (Difco) or on MRS supplemented with 1.5 % agar (Fisher). Plates were grown anaerobically in a Coy anaerobic chamber with a gas composition of 90 % nitrogen, 5 % hydrogen and 5 % carbon dioxide. When necessary, 2.5 µg erythromycin ml–1 and 5 µg chloramphenicol ml–1 (both Fisher BioReagents) were added to the MRS agar plates. For RT-PCR studies, NCK334 was grown at 37 °C in 10 ml semisynthetic MRS (sMRS) medium supplemented with 1 % carbohydrate (w/v) as described elsewhere, except that bromocresol purple was not used (Barrangou et al., 2006).
Escherichia coli cells were grown at 37 °C aerobically with shaking in Luria–Bertani (LB) broth (Difco) or on LB supplemented with 1.5 % agar (Fisher BioReagents). Antibiotics were added when appropriate at the following concentrations: 40 µg kanamycin ml–1 (Teknova), 150 µg erythromycin ml–1 (Fisher BioReagents) and 15 µg chloramphenicol ml–1 (Fisher BioReagents).
Nucleic acid isolation, manipulation and transformation.
Genomic DNA was isolated from NCK334 using the Microbial DNA Isolation kit (MO BIO) according to the manufacturer’s protocol. E. coli plasmid DNA was isolated using the QIAprep Spin Miniprep kit (Qiagen).
DNA manipulations were performed according to standard procedures. Restriction enzymes and T4 ligase were obtained from Invitrogen. When necessary, DNA fragments were isolated from agarose gels using the Zymoclean Gel DNA Recovery kit (Zymo Research). PCRs were performed according to standard procedures using EconoTaq polymerase (Lucigen). PCR primers were designed using Clone Manager 9 (Sci-Ed Software) and purchased from Integrated DNA Technologies, Inc. For cloning, restriction enzyme sites were added at the 5′ end of the primers. When necessary, PCR products were purified using the DNA Clean and Concentrator kit (Zymo Research).
Electrocompetent NCK334 cells were prepared using 3.5× sucrose MgCl2 electroporation buffer as described elsewhere (Luchansky et al., 1991). To perform the electroporation, the Electroporator 2510 (Eppendorf) was used at a setting of 2.5 kV.
RNA isolation and analysis.
sMRS (Barrangou et al., 2003) was used to analyse PTS gene expression in response to carbohydrates. The carbohydrates added to the medium were glucose (Fisher), fructose (Sigma-Aldrich) or lactose (Fisher). A 0.1 % overnight culture was transferred six times before isolation of RNA, as described elsewhere (Barrangou et al., 2006). The final transfer of L. gasseri was grown to OD595 0.6 in order to obtain mid-exponential phase growth of the bacterial cells. A 1.5 ml aliquot of culture was collected by centrifugation at 10 000 g at room temperature. RNA for RT-PCR was isolated from the L. gasseri strains using the Microbial RNA Isolation kit (MO BIO) according to the manufacturer’s protocol. Typically, 100 ng RNA µl−1 was treated with TURBO DNA-free according to the supplier’s instructions (Ambion) in a 50 µl reaction volume.
Two-step RT-PCR was performed to carry out the relative quantification as described previously (Francl et al., 2010). Primers were designed for the IIC subunit of the 15 complete PTSs which were reported previously (Francl et al., 2010). The primers for the seven potential β-galactosidase genes in L. gasseri were determined using Clone Manager 9 and are shown in Table 2. Primers used in the RT-PCR experiments were synthesized by Invitrogen. Relative fold changes were reported by using a phosphofructokinase (PFK) gene in L. gasseri that was previously shown in L. plantarum WCFS1 to exhibit qualities of an acceptable internal standard (Marco et al., 2007). The ΔΔCt method was used to calculate the relative fold change of the PTS and potential β-galactosidase genes, as previously described (Francl et al., 2010). Relative fold change values were compared with either glucose or fructose. Reported relative fold changes are the mean of three independent experiments.
F, forward; R, reverse.
PTS 6CB and PTS 8C gene inactivation.
The inactivation of PTS 6CB (LGAS_343) and PTS 8C (LGAS_497) was carried out by the two-plasmid lactococcal integration strategy (Law et al., 1995) adapted for gene replacements in lactobacilli (Bruno-Bárcena et al., 2005). The non-replicative vectors pMJM-2 and pMJM-3 were designed to disrupt the gene function of PTS 6 IIBC and PTS 8 IIC, respectively. The primers AF_PTS6HindAB and AF_PTS6XhoAB (Table 3) were used to amplify a 715 bp internal region of PTS 6CB from L. gasseri (amplicon AB). The primers AF_PTS6XhoBC and AF_PTS6XbaBC (Table 3) were used to amplify a 716 bp internal region of PTS 6CB from L. gasseri (amplicon BC). These fragments were cloned via the HindIII/XbaI sites (Table 3, underlined) into pORI28 and resulted in pMJM-2. The pMJM-3 construction followed the same strategy as that for pMJM-2, except that PTS 8 primers were used (Table 3). Plasmids pMJM-2 and pMJM-3 were introduced into L. gasseri containing pTRK669 (MJM79) by electroporation. Subsequent steps to facilitate the gene replacement were carried out using protocols described elsewhere (Bruno-Bárcena et al., 2005). Suspected gene replacements were confirmed by PCR and subsequent DNA sequencing using standard procedures. MJM76 is L. gasseri ATCC 33323 PTS 6CB and MJM77 is L. gasseri ATCC 33323 PTS 8C.
343: PTS, cellobiose-specific, EIICB; 497: PTS, galacitol-specific, EIIC; 501: PTS, cellobiose-specific. Restriction sites are underlined.
L. gasseri ATCC 33323 PTS 6CB, PTS 8C was created by making a PTS 8C gene replacement in MJM76. Initially, pTRK669 was transformed into MJM76 followed by the steps for generating the PTS 8C gene replacement described above. One candidate was then selected and designated L. gasseri ATCC 33323 PTS 6CB, PTS 8C (MJM78).
PTS 9BC gene knockouts.
The inactivation of PTS 9BC (LGAS_501) was performed by targeted insertion of an erythromycin-resistant, non-replicative vector pMJM-10 by homologous recombination using an established method (Russell & Klaenhammer, 2001). The primers MM_PTS9Bam and MM_PTS9Nco (Table 3) were used to amplify an 868 bp internal region of PTS 9BC from NCK334. The amplicon was cloned via the BamHI/NcoI sites (Table 3, underlined) into pORI28. The plasmid pMJM-10 was introduced into MJM79 by electroporation. Subsequent steps to facilitate the gene knockout were carried out using protocols described elsewhere (Russell & Klaenhammer, 2001). The inactivation of PTS 9BC in MJM76, MJM77 and MJM78 followed a similar procedure. Suspected PTS 9BC knockouts were confirmed by PCR and subsequent DNA sequencing using standard procedures. Candidate knockouts were selected: L. gasseri ATCC 33323 PTS 9BC (MJM116), L. gasseri ATCC 33323 PTS 6CB, PTS 9BC (MJM117), L. gasseri ATCC 33323 PTS 8C, PTS 9BC (MJM118) and L. gasseri ATCC 33323 PTS 6CB, PTS 8C, PTS 9BC (MJM119).
Carbohydrate utilization analysis.
Carbohydrate utilization assays and kinetic growth curves were performed with the following strains: NCK334, MJM76, MJM77 and MJM78. Strains were analysed for their ability to utilize carbohydrates with the API 50 CH carbohydrate utilization assay (bioMérieux) according to the manufacturer’s protocol. The API 50 CH assay does not include the ability to discriminate strains with β-galactosidase and/or phospho-β-galactosidase activity. Additional kinetic growth assays were done with the following strains: MJM116, MJM117, MJM118 and MJM119. For the kinetic growth curves, the strains were grown at 37 °C in sMRS supplemented with 1 % carbohydrate (w/v) as described elsewhere, except that bromocresol purple was not used (Barrangou et al., 2006). Cell growth was performed in 250 µl sMRS covered with 50 µl mineral oil in a Bioscreen 100-well honeycomb plate. Cell growth was monitored by measuring OD600 using the Bioscreen C Automated Microbiology Growth Curve Analysis System (Growth Curves USA). The plate reader was operated in discontinuous mode, with absorbance readings performed at 30 min intervals, and preceded by 30 s shaking intervals at maximum speed. Controls consisted of inoculated medium lacking carbohydrate, and uninoculated medium was used as a blank.
Results
Bioinformatic analysis
Bioinformatic analysis was used to identify genes potentially involved in lactose utilization in NCK334. Earlier studies have identified a lactose/galactose-specific permease in the closely related Lactobacillus acidophilus NCFM (Barrangou et al., 2003). The identification of bacteria that contain a homologue to the lactose/galactose-specific permease (LBA1463) was determined using concise protein blast, available on the blast database (). Concise protein blast identifies protein clusters which are reciprocal best blast hits (Klimke et al., 2009). We defined homologues as being members of the same protein cluster. Cluster CLS1225721, of which LBA1463 is a member, contains several proteins from the order Lactobacillales, yet none from NCK334. Additionally, cluster CLS1225721 only has members from the order Lactobacillales and genus Bifidobacterium.
Previously, we determined that a functional PTS was required for lactose utilization by NCK334. Briefly, an EI knockout strain (MJM75) was evaluated for its ability to grow on lactose using two different growth assays (Francl et al., 2010). PTSs require a functional EI to import carbohydrates (Cvitkovitch et al., 1995). First, an API 50 CH assay revealed that NCK334 could utilize lactose, while the EI knockout strain could not. Second, growth curves using sMRS supplemented with 1 % lactose confirmed the API 50 CH results (Francl et al., 2010). From these results, we concluded that PTS(s) are required for the utilization of lactose by NCK334.
NCK334 has 15 complete PTSs (Francl et al., 2010). A complete PTS was defined as having the IIA, IIB and IIC components present in the enzyme II of the PTS, while an incomplete PTS lacked one or more of these three subunits (Barabote & Saier, 2005). Of the 15 complete PTSs in NCK334, none has been annotated as lactose-specific. However, bioinformatic analysis was used to identify several possible candidates for lactose transport in NCK334, including PTS 5, PTS 6, PTS 8, PTS 9 and PTS 15. PTS 6 (LGAS_342–343; Fig. 1a) and PTS 9 (LGAS_500–501; Fig. 1b) are members of the same protein cluster as the characterized lactose-specific PTS in L. zeae ATCC 393 (Gosalbes et al., 1999). PTS 6 and PTS 9 also have a potential phospho-β-galactosidase in their respective operons (LGAS_0344 and LGAS_0502, respectively). PTSs in the same protein cluster as PTS 8 (LGAS_495–497; Fig. 1b) have not been characterized. However, members of the tagatose pathway, which is required for utilization of lactose 6-phosphate, are in the same operon as PTS 8 (galactose-6-phosphate isomerase A and B; LGAS_492–493; Fig. 1b), suggesting that PTS 8 may be involved in lactose utilization. PTS 5 (LGAS_192, 194–195) and PTS 15 (LGAS_1669) also have putative phospho-β-galactosidases in their respective operons (LGAS_185, 190, 196 and LGAS_1668, respectively). Since we cannot eliminate the possibility of other PTSs being involved in lactose import, we included all 15 complete PTSs in our transcript analysis.
PTS 6 (LGAS_341–347) (a), PTS 8 and PTS 9 regions (LGAS_491–502) (b) drawn to scale. The regions are flanked by ORFs encoded on the opposite strand. PTS 6, PTS 8 and PTS 9 ORFs are represented by black arrows. Annotated gene functions listed are from the publicly available annotation of NCK334 on NCBI.
The potential β-galactosidase genes in NCK334 were selected for further study due to the importance of β-galactosidases in the utilization of lactose (Kandler, 1983). Seven genes were identified as being potential β-galactosidase genes and their transcript expression profiles were analysed.
Transcript expression profiles
The genes necessary for the utilization of a carbohydrate show increased expression in the sole presence of the specific carbohydrate (Barrangou et al., 2006; Duong et al., 2006). Therefore, RT-PCR was used to study the transcript expression profiles of the 15 complete PTSs in NCK334 in response to lactose when compared with fructose as the calibrator. As shown in Fig. 2a, PTS 6 was induced 107±20-fold and PTS 8 was induced 53±32-fold in the presence of lactose. PTS 3 also showed a slight level of induction at ninefold. All other PTSs tested were induced less than fivefold in the presence of lactose.
(a) Relative fold changes of all PTSs in NCK334. (b) Relative fold changes of the potential phospho-β-galactosidase ORFs in NCK334. Results are the mean of three independent experiments; error bars, sd.
RT-PCR was used to study the response of potential β-galactosidase genes to lactose. LGAS_344 was induced 156±60-fold in the presence of lactose (Fig. 2b). All other potential β-galactosidase genes tested were induced less than sixfold in the presence of lactose. LGAS_344 is located immediately downstream of PTS 6 (LGAS_342–343; Fig. 1a). LGAS_502, which is located immediately downstream of PTS 9 (LGAS_500–501; Fig. 1b) and five ORFs downstream of PTS 8 (LGAS_495–497; Fig. 1b), is not induced in the presence of lactose.
Inactivation of PTS 6CB and PTS 8C
To confirm the functions of PTS 6 and PTS 8 in lactose utilization, gene replacements of PTS 6CB and PTS 8C were created in L. gasseri, designated MJM76 and MJM77, respectively, and confirmed individually. The gene replacements resulted in frameshift deletions within the desired PTS gene. Subsequently, a double gene replacement strain in which both PTS 6CB and PTS 8C were inactivated was created (MJM78).
The three L. gasseri gene replacement strains and the wild-type were evaluated for their ability to grow on sMRS supplemented with 1 % lactose (Table 4). MJM76, MJM77, MJM78 and NCK334 had very similar ODmax (maximum OD600), growth rate, doubling time and lag time.
Values shown are the mean±sd of three independent replicates. ODmax, maximum OD600; μmax, maximum specific growth rate; td, doubling time; tlag, lag phase; nd, not determined.
Identification of an additional PTS
To determine which PTSs were responsible for lactose utilization following the inactivation of PTS 6CB and PTS 8C, transcript expression profiles in MJM76, MJM77 and MJM78 in response to lactose were evaluated. In addition to PTS 6 and PTS 8, MJM76, MJM77 and MJM78 showed induction of PTS 9 in response to lactose by 170±107-, 76±41- and 373±83-fold, respectively, using glucose as the calibrator. However, PTS 9 was not induced by lactose with NCK334. All other PTSs tested were induced less than 15-fold in the presence of lactose (data not shown).
To confirm the function of PTS 9 in lactose utilization, PTS 9BC knockout strains were created in NCK334, MJM76, MJM77 and MJM78, and are designated MJM116, MJM117, MJM118 and MJM119, respectively. These eight strains were grown on sMRS supplemented with 1 % lactose to determine their ability to utilize lactose (Table 4). The lag time and ODmax of MJM116 and MJM118 were indistinguishable from those of NCK334. Only MJM117 and MJM119 did not grow in the presence of lactose.
In order to confirm that MJM117 and MJM119 did not have a larger deficiency related to sugar catabolism, growth curves were performed with NCK334, MJM75, MJM117, MJM119 and MJM100 (Fig. 3). All five strains grew very well on glucose and were indistinguishable. MJM75, which lacks a functioning EI, was unable to grow on trehalose, sucrose and lactose. Previously, we demonstrated that growth on trehalose and sucrose requires a functioning PTS system (Francl et al., 2010). MJM100, which is a PTS 20BCA knockout strain, was unable to grow on sucrose but was otherwise indistinguishable from the wild-type strain. While MJM117 and MJM119 cannot ferment lactose, they were indistinguishable from the wild-type when grown on trehalose and sucrose, which require a functioning PTS 11ABC and PTS 20BCA, respectively, for growth. In addition, an API 50 CH assay with NCK334, MJM117 and MJM119 revealed that these three strains were indistinguishable apart from the inability of MJM117 and MJM119 to ferment lactose (data not shown).
Growth curves of NCK334, gene replacement strains and gene knockout strains. Strains were grown on sMRS+1 % carbohydrate: glucose (a), trehalose (b), sucrose (c) and lactose (d). Results are the mean of three independent experiments.
Discussion
In regard to sugar catabolism, the utilization of lactose is among the most studied in LAB (de Vos & Vaughan, 1994). Lactose permeases have been described in numerous LAB (Barrangou et al., 2003; Leong-Morgenthaler et al., 1991). However, a bioinformatic analysis did not identify a lactose permease in L. gasseri ATCC 33323 (NCK334). The creation of an EI knockout in L. gasseri ATCC 33323 and subsequent carbohydrate utilization analysis confirmed that lactose could not be utilized by the knockout strain, confirming that one or more PTSs are required for lactose utilization by NCK334 (Francl et al., 2010).
No PTSs in L. gasseri are currently annotated as involved in lactose utilization. Among LAB, lactose-specific PTSs have been described in Lactococcus lactis subsp. lactis (McKay et al., 1970; de Vos et al., 1990), S. mutans (Honeyman & Curtiss, 1992) and L. zeae (Alpert & Chassy, 1990; Chassy et al., 1976; Chassy & Thompson, 1983; Gosalbes et al., 1999). Based on our bioinformatic analysis, we predicted that the lactose PTSs were PTS 5, PTS 6, PTS 8, PTS 9 and/or PTS 15. Transcript expression analysis identified PTS 6, PTS 8 and a phospho-β-galactosidase near PTS 6 as induced in the presence of lactose. PTS 5, PTS 9, PTS 15 and the other six potential phospho-β-galactosidases were not induced in the presence of lactose.
Gene replacement strains (MJM76, MJM77 and MJM78) were created to confirm the transcript expression profiles of NCK334 and determine the roles of PTS 6 and PTS 8 in lactose utilization. Based on the EI knockout results, PTSs are solely responsible for the utilization of lactose (Francl et al., 2010), although MJM78 was still able to grow on lactose.
We correctly hypothesized that an additional PTS is expressed in MJM78 that is capable of transporting lactose and which is not differentially expressed in NCK334. While it is unknown why PTS 9 is not induced by lactose in NCK334, growth on lactose was prevented in MJM117 and MJM119. Conversely, PTS 8 expression is induced by lactose yet it cannot import enough lactose to enable detectable growth.
It has been reported that lactose utilization by other strains of L. gasseri is not common (Azcarate-Peril et al., 2008). However, all three L. gasseri strains that we tested (ATCC 33323, ATCC 19992 and ADH) are able to utilize lactose (Francl et al., 2010). Additionally, analysis of newly available draft genomic DNA sequences for four additional L. gasseri strains (224-1, 202-4, MV-22 and JV-V03) has confirmed the presence of PTS 6, PTS 8 and PTS 9 in these strains as well. Similar to NCK334, these four L. gasseri strains also lack a lactose permease, yet the draft nature of their genomes prevents a definitive conclusion.
The differences among L. gasseri strains and the closely related L. acidophilus NCFM and L. johnsonii NCC 533 suggest that the species L. gasseri may have lost the lactose-specific permease early in speciation. Of particular interest is the apparent gain of lactose-specific PTSs in the L. gasseri strains, including NCK334, that are able to utilize lactose. These changes among L. gasseri strains could have occurred through horizontal gene transfer (HGT), a well-known process in lactobacilli for adaptation to specific niches (Wood & Warner, 2003). Furthermore, genes involved in sugar metabolism, including enzymes and components of PTSs, appear to have a higher rate of HGT than other genes in LAB (Makarova et al., 2006).
Our study demonstrates the importance of characterization of PTS function. Due to the limited number of PTSs that have been characterized, bioinformatic analysis can be insufficient in predicting PTS substrate specificity. The transcript expression profiles point to PTS 6 and PTS 8 as being lactose-specific PTSs, yet MJM78 indicated that an additional PTS contributed to lactose utilization. Subsequent transcript analysis revealed that PTS 9BC was induced by lactose in MJM78 but not in NCK334. Only when PTS 6CB and PTS 9BC are eliminated is growth on lactose prevented.
Acknowledgements
We acknowledge Rodolphe Barrangou and Tri Duong for insightful discussions and technical help. The authors would also like to acknowledge Julia Willett for her help in bioinformatic analysis. This project was supported by the USDA Cooperative State Research, Education and Extension Service, Hatch project number #ILLU-698-339. A. L. F. was supported by the Bill and Agnes Brown Fellowship.