Characterization of a Galactokinase-Positive Recombinant Strain of Streptococcus thermophilus

Abstract

The lactic acid bacterium Streptococcus thermophilus is widely used by the dairy industry for its ability to transform lactose, the primary sugar found in milk, into lactic acid. Unlike the phylogenetically related species Streptococcus salivarius, S. thermophilus is unable to metabolize and grow on galactose and thus releases substantial amounts of this hexose into the external medium during growth on lactose. This metabolic property may result from the inability of S. thermophilus to synthesize galactokinase, an enzyme of the Leloir pathway that phosphorylates intracellular galactose to generate galactose-1-phosphate. In this work, we report the complementation of Gal⁻ strain S. thermophilus SMQ-301 with S. salivarius galK, the gene that codes for galactokinase, and the characterization of recombinant strain SMQ-301K01. The recombinant strain, which was obtained by transformation of strain SMQ-301 with pTRKL2TK, a plasmid bearing S. salivarius galK, grew on galactose with a generation time of 55 min, which was almost double the generation time on lactose. Data confirmed that (i) the ability of SMQ-301K01 to grow on galactose resulted from the expression of S. salivarius galK and (ii) transcription of the plasmid-borne galK gene did not require GalR, a transcriptional regulator of the gal and lac operons, and did not interfere with the transcription of these operons. Unexpectedly, recombinant strain SMQ-301K01 still expelled galactose during growth on lactose, but only when the amount of the disaccharide in the medium exceeded 0.05%. Thus, unlike S. salivarius, the ability to metabolize galactose was not sufficient for S. thermophilus to simultaneously metabolize the glucose and galactose moieties of lactose. Nevertheless, during growth in milk and under time-temperature conditions that simulated those used to produce mozzarella cheese, the recombinant Gal⁺ strain grew and produced acid more rapidly than the Gal⁻ wild-type strain.

Streptococcus thermophilus is widely used in yoghurt fermentation and cheese making. Specific S. thermophilus strains may also have the potential to improve human health as probiotics. For instance, they may interfere with the adhesion of indigenous cariogenic bacteria to teeth (9) and may prolong the useful life of indwelling voice prostheses by preventing colonization by Candida spp. (8).

The ability of S. thermophilus to rapidly take up sugars from the environment is a prerequisite for these applications. Unlike several other lactic acid bacteria, S. thermophilus only uses a few sugars, with most strains showing a marked preference for saccharose and lactose, while glucose and fructose are more slowly fermented, if they are fermented at all (15, 19). S. thermophilus takes up lactose via the membrane protein LacS, a member of the glycoside-pentoside-hexuronide:cation symporter family (29). Inside the cell, the disaccharide is hydrolyzed into glucose and galactose by the enzyme β-galactosidase. While all strains metabolize glucose via the Embden-Meyerhof-Parnas pathway, most are unable to metabolize galactose (23), which is expelled into the external medium (19). The galactose expulsion phenomenon is closely associated with S. thermophilus LacS, which is able to catalyze two modes of transport: a Δp-driven lactose uptake in symport with protons and a lactose-galactose exchange (16).

Accumulation of galactose in dairy products may lead to several undesirable effects, notably, browning of mozzarella cheese (21), production of CO₂ by nonstarter bacteria (36), and growth of spoilage and/or pathogenic microorganisms (37). Excess galactose in dairy products may also adversely affect human health, particularly in individuals with galactosemia (25).

The inability of S. thermophilus to metabolize galactose does not result from a lack of appropriate genetic information, as the bacterium possesses the genes that code for the enzymes of the Leloir pathway, namely, galactokinase (galK), galactose-1-phosphate uridylyltransferase (galT), and UDP-glucose 4-epimerase (galE) (11, 40, 42). These chromosomal genes make up the gal operon, which is immediately followed by galM, which codes for a galactose mutarotase, and the lac operon, which comprises the genes that code for LacS (lacS) and β-galactosidase (lacZ). The inability of S. thermophilus to metabolize galactose may result from poor expression of the gal genes because of a deficient promoter (42). This, however, is not consistent with the finding that S. thermophilus cells contain significant amounts of GalT and GalE activities (10, 30, 35, 40). Moreover, a 3.5- to 3.7-kb polycistronic galKTE mRNA transcript has been detected in S. thermophilus A147 and SMQ-301, indicating that the gal promoter is functional (30, 40).

Despite efficient transcription of the gal genes, most S. thermophilus strains do not synthesize substantial amounts of galactokinase, the first enzyme of the Leloir pathway (19, 20, 35, 40). This may account, at least in part, for the inability of S. thermophilus to grow on galactose (6, 19, 29). This hypothesis is reinforced by the observation that spontaneous S. thermophilus Gal⁺ mutants possess higher galactokinase activity than the wild-type strains (6, 19, 35). It is noteworthy that these mutants are unstable and lose their Gal⁺ phenotype under normal growth conditions in milk (35).

Advances in molecular biology have made possible the metabolic engineering of microorganisms, which likely constitutes the best approach to generating stable Gal⁺ S. thermophilus strains. However, an in-depth understanding of the targeted metabolic pathway, at the biochemical and genetic levels, is needed to successfully apply this strategy. We previously showed that the galK, galT, and galE gene products of a Gal⁻ S. thermophilus strain are more than 95% identical to the homologues of Streptococcus salivarius, a phylogenetically related Gal⁺ species (40). The gal promoters of both species are virtually identical and direct the transcription of the gal genes into a galKTE polycistronic mRNA. The two species have similar levels of GalT and GalE, but S. thermophilus has much less galactokinase activity than does S. salivarius. Sequence analysis revealed that the ribosome binding site of S. thermophilus galK differs from that of S. salivarius by two nucleotides, suggesting that the S. thermophilus galK gene is poorly translated. In this work, we report the complementation of a Gal⁻ strain of S. thermophilus with S. salivarius galK, which bears its own ribosome-binding site, as well as the characterization of the recombinant strain.

MATERIALS AND METHODS

Strains and growth conditions.

The strains and plasmids used in this study are listed in Table 1. S. salivarius was grown at 37°C, and S. thermophilus was grown at 42°C in M17 broth (Difco Laboratories, Detroit, Mich.) supplemented with sugars as described in the text. For DNA isolation, S. salivarius was grown in a medium containing 10 g of tryptone and 5 g of yeast extract (Difco), 2.5 g of NaCl, 2.5 g of disodium phosphate, and 5 g of glucose per liter. Growth was monitored by following the optical density at 660 nm (OD₆₆₀). For determination of sugar concentrations in the medium of growing cells, samples (0.25 to 0.5 ml) were taken at intervals, heated at 100°C for 10 min, and centrifuged to remove cells. The supernatants were stored at −20°C until used for the sugar assays.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant genotype or characteristicsa	Source or reference
Strains
Escherichia coli XL1 Blue	recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac [F′ proAB lacIqZΔM15 Tn10 (Tetr)]	Stratagene
S. salivarius ATCC 25975	Wild type; Lac+ Glu+ Gal+	17
S. thermophilus
SMQ-301	Wild type; Lac+ Glu− Gal−	38
SMQ-301K01	S. thermophilus SMQ-301 bearing plasmid pTRKL2TK; Lac+ Glu− Gal+	This work
Plasmids
pTRKL2	Cloning vector; Emr	27
pTRKL2T	Cloning vector; Emr	This work
pTRKL2TK	Contains S. salivarius galK, including the ribosome binding site and the gal promoter	This work

^aEm^r, erythromycin resistance.

DNA purification and manipulations.

Chromosomal DNA was isolated from S. salivarius as described previously (14). Unless otherwise mentioned, DNA manipulations were performed in accordance with standard procedures (3). Escherichia coli XL-1 Blue cells were made competent and transformed with plasmid DNA by electroporation (32). S. thermophilus was transformed by electroporation as previously described (7).

Construction of a galactose-positive recombinant strain of S. thermophilus.

S. salivarius galK (GenBank accession no. AF389474) was PCR amplified with forward primer galKR2-S (5′-AAACAGTCGACATATTGATTTTCCTTGT-3′), which contained an engineered SalI site (underlined), and reverse primer galKT1R-B (5′-CAGCCATAAGATCTACTCCTTTCTCATT-3′), which contained an engineered BglII site (underlined). Primer galKR2-S covered positions −77 to −50 relative to the adenine of the ATG initiation codon of S. salivarius galK, and primer galKT1R-B covered positions 1,166 to 1,193. The PCR was performed with a DNA Thermal Cycler 480 (Perkin-Elmer) in a total volume of 25 μl containing 20 mM Tris-HCl (pH 8.8), 10 mM (NH₄)₂SO₄, 10 mM KCl, 2 mM MgSO₄, 0.1% Triton X-100, 1 μg of DNA, primers at 0.2 μM, and the four deoxynucleotide triphosphates at 200 μM (each). The reaction was carried out for 25 cycles in the presence of 0.5 U of Vent DNA polymerase (NEB) with the following temperature-time profile: 94°C for 1 min, 52°C for 1 min, and 72°C for 2 min. The amplified DNA fragment (1,270 bp) comprised the entire galK gene with its own ribosome binding site, as well as the promoter region of the gal operon, but not the binding site of GalR, a transcriptional regulator of the gal operon (2, 42). After digestion with SalI and BglII, the amplicon was cloned into low-copy-number shuttle plasmid pTRKL2T to yield plasmid pTRKL2TK, which was propagated in E. coli XL-1 Blue. Plasmid pTRKL2T was obtained by cloning the transcription terminators of the Omegon kanamycin from pCIV2 (26) at the EcoRV site of pTRKL2 (27). The nucleotide sequence of the DNA fragment cloned into pTRKL2TK was confirmed by sequencing both DNA strands by the dideoxy chain termination method of Sanger et al. (33). DNA sequencing was carried out by the DNA sequencing service at the Université Laval with an ABI Prism 3100 apparatus.

pTRKL2TK was then used to transform Gal⁻ strain S. thermophilus SMQ-301. After electroporation, cells were spread on galactose-supplemented M17 plates containing 5 μg of erythromycin per ml, which allowed the isolation of the recombinant strain SMQ-301K01. The identity of the recombinant S. thermophilus strain was confirmed by optical microscopy, plasmid profiling, gram staining, phage typing, and determination of the sugar fermentation pattern with API50CH strips (bioMérieux). Lastly, to ascertain that the gal locus was intact in the chromosome of SMQ-301K01, PCR experiments were conducted with primers galR875 (5′-TAGTATATCCGATTTCATCAGCGAT-3′) and galT1800R (5′-CGTTAAAGTAGGCATAAGGCGAG-3′), which hybridized with galR and galT, respectively, the genes flanking galK on the chromosome.

RNA isolation and Northern blot analysis.

Cells were grown in 25 ml of M17 medium supplemented with 0.5% lactose or galactose to an OD₆₆₀ of 0.4, 0.6, or 0.8. Cells were collected by centrifugation and suspended in 600 μl of the RLT buffer provided with the RNeasy kit used for RNA isolation (Qiagen), supplemented with 6 μl of 14.3 M 2-mercaptoethanol. Glass beads (0.35 g, 150 to 212 μm in diameter) were added, and the cells were broken by vortexing the suspension three times for 1 min with a Vortex set at maximum. Between treatments, the cell suspensions were chilled on ice for 1 min. After removal of the glass beads and cellular debris by centrifugation (2 min at 13,000 × g), the RNA was isolated by the procedure recommended by the manufacturer of the RNeasy kit and stored at −80°C. The RNA was fractionated on a 0.8% agarose gel as described by Ausubel et al. (3), transferred to a positively charged nylon membrane with a Turboblotter apparatus (Schleicher & Schuell), and fixed by UV cross-linking for 3 min. PCR products labeled by random priming with [α-³²P]ATP (Pharmacia) were used as hybridization probes. These included galK, galT, and two plasmid lacZ-specific probes. Prior to radiolabeling, the probes were purified as described previously (12).

Enzyme assays.

Bacteria were grown in 500 ml of M17 broth containing 0.5% lactose or galactose to an OD₆₆₀ of 0.35. Cell extracts were prepared by grinding with alumina as described previously (39). The cell fraction obtained after centrifugation at 20,000 × g for 20 min was used for β-galactosidase assays. Galactokinase activities were determined with the supernatant obtained after centrifugation of the cell fraction mentioned above at 150,000 × g for 90 min. β-Galactosidase activity was assayed with o-nitrophenyl-β-d-galactopyranoside as the substrate (18). Galactokinase activity was assayed by measuring the oxidation of NADH in a coupled assay with pyruvate kinase and lactate dehydrogenase from rabbit muscle as described previously (44).

Milk acidification activity.

Strains SMQ-301 and SMQ-301K01 were grown overnight at 35°C in Elliker broth (Difco). The cultures were mixed with glycerol (20% [vol/vol]) and frozen at −80°C. Reconstituted 12% (wt/vol) low-heat skim milk (Agropur) was pasteurized at 85°C for 45 min and then inoculated at 2% (wt/vol) with frozen cells. The cultures were incubated at 35°C until the pH reached about 4.6. Whole milk used for acidification kinetics was pasteurized at 65°C for 30 min and inoculated at 2% with the cultures obtained in the reconstituted milk. The acidification kinetics, performed in triplicate, were recorded for 171 min with a Cinac system (version 3; Ysebaert). The time-temperature profile simulated that used to produce mozzarella cheese. Plate counts with M17-lactose (Difco) were done for samples withdrawn at time zero and 171 min.

Sugar and protein assays.

Glucose concentrations were measured by a peroxidase-glucose oxidase assay (Sigma). Galactose was determined by a peroxidase-galactose oxidase assay (4). Lactose was assayed by measuring the concentration of glucose or galactose in samples before and after hydrolysis with β-galactosidase for 1 h at 37°C in 233 mM citrate buffer (pH 6.6) containing 60 mM MgSO₄ and 0.05 U of β-galactosidase (Worthington) per μl. Protein concentrations were measured by the method of Peterson (28) with bovine serum albumin as the standard.

RESULTS

Growth properties.

Complementation of Gal⁻ strain S. thermophilus SMQ-301 with the S. salivarius galactokinase-encoding gene galK enabled the recombinant strain SMQ-301K01 to grow on galactose (Fig. 1). The generation time of strain SMQ-301K01 on 0.2% galactose was approximately 55 min, which was roughly double the generation time on lactose (Table 2). The generation time of SMQ-301K01 on galactose was also about twice that of S. salivarius, which also uses the Leloir pathway to metabolize galactose (40). S. thermophilus SMQ-301 (Gal⁻) and SMQ-301K01 (Gal⁺), as well as S. salivarius ATCC 25975 (Gal⁺), grew at virtually the same rate on lactose. It is worth noting that even after repeated culturing in lactose-containing medium (approximately 50 generations), strain SMQ-301K01 retained its ability to grow on galactose.

FIG. 1.

Growth on galactose. One colony of either S. salivarius or S. thermophilus SMQ-301K01 was used to inoculate 10 ml of M17 medium supplemented with 0.2% galactose, and one colony of S. thermophilus SMQ-301 was used to inoculate 10 ml of M17 medium containing 0.2% lactose. The cultures were incubated overnight at 37°C for S. salivarius and at 42°C for S. thermophilus. These cultures (100 μl) were then used to inoculate 10 ml of fresh medium containing 0.2% galactose. Open circles, growth of S. salivarius; solid circles, growth of S. thermophilus SMQ-301K01; open squares, growth of S. thermophilus SMQ-301.

TABLE 2.

Generation times

Sugar (0.2%)	Mean generation timea (min) ± SE
	SMQ-301	SMQ-301K01	S. salivarius
Lactose	25 ± 0	26.5 ± 2.8	27.3 ± 0.8
Galactose	No growth	54.4 ± 4.9	29.8 ± 1.3

^aGrowth was monitored by following the OD₆₆₀. Generation times were calculated for cultures in the exponential growth phase by plotting the logarithm of the OD₆₆₀ as a function of time. Values represent the means of three separate experiments.

Growth of strain SMQ-301K01 on galactose resulted from complementation with S. salivarius galK.

To demonstrate that the Gal⁺ phenotype of SMQ-301K01 resulted from complementation with S. salivarius galK and not from a secondary unknown chromosomal mutation, pTRKL2TK was cured from SMQ-301K01 by repeated culturing in the presence of lactose without erythromycin for 15 days. Cells were then plated on M17 supplemented with lactose. Five colonies were picked at random, and the cells were tested for the ability to grow on galactose. Cells from the five colonies had lost plasmid pTRKL2TK, were unable to grow on galactose, and were sensitive to erythromycin. They were, however, able to grow on lactose at a rate identical to that of strain SMQ-301 (data not shown).

To verify that plasmid pTRKL2TK was not integrated into the chromosome at the gal locus, we carried out PCR experiments with DNA from strains SMQ-301 and SMQ-301K01 with primers that hybridized with galR and galT, which flank the chromosomal S. thermophilus galK gene. A single 1.2-kb amplicon, which was the size expected for the amplified DNA fragment without an insertion, was obtained with both strains (data not shown), indicating that pTRKL2TK was not integrated into the chromosome at the gal locus.

mRNA analysis of the plasmid-borne S. salivarius galK gene.

We carried out Northern blot analyses to determine whether the S. salivarius galK gene on pTRKL2TK was transcribed in S. thermophilus SMQ-301K01. Previous studies demonstrated that transcription of the S. thermophilus gal operon generates an abundant 3.7-kb galKTE mRNA transcript and a barely detectable 10-kb transcript that encompasses the gal operon, galM, and the lac operon (40). With an S. salivarius galK-specific probe, we detected in galactose-grown (not shown) and lactose-grown (Fig. 2A) cells of strain SMQ-301K01 the expected 3.7- and 10-kb transcripts. We also detected a short 1.2-kb transcript, which corresponded to the expected size of the galK mRNA resulting from the transcription of S. salivarius galK borne by plasmid pTRKL2TK. Surprisingly, an additional 3-kb mRNA transcript was detected. This transcript was not detected with a galT-specific probe, indicating that it was not directed from the chromosomal gal promoter (Fig. 2B). Northern analyses performed with probes that specifically hybridized with either the 5′ or the 3′ end of the plasmid lacZ gene interrupted by S. salivarius galK revealed that the 3-kb transcript resulted from readthrough by the RNA polymerase beyond the Omegon kanamycin terminators of pTRKL2TK (data not shown).

FIG. 2.

Northern blot analysis. (A) S. thermophilus SMQ-301K01 was grown at 42°C in M17 medium supplemented with 0.5% lactose until the OD₆₆₀ reached 0.4. Total RNA (5 μg) was hybridized with a galK-specific probe. The 10-kb transcript is barely detectable in S. thermophilus (40) and, although apparent in the autoradiogram, is not discernible in the figure. (B) Total RNA (5 μg) was hybridized with a galT-specific probe. Lanes: 1, RNA isolated from S. thermophilus SMQ-301K01 grown at 42°C in M17 medium supplemented with 0.5% lactose and harvested at mid-log phase; 2, same as lane 1 but with RNA from S. thermophilus SMQ-301; 3, same as lane 1 but with cells grown on 0.5% galactose. The estimated sizes of the transcripts are shown on the left, and the position of the 23S rRNA is shown on the right.

Enzyme activities.

Galactokinase and β-galactosidase activities were determined in S. thermophilus SMQ-301K01 after growth on lactose and galactose and in S. thermophilus SMQ-301 after growth on lactose. For comparison, the enzyme activities were also determined in lactose- and galactose-grown cells of S. salivarius (Table 3). Galactokinase activity was twofold higher in strain SMQ-301K01 than in strain SMQ-301 after growth on lactose, β-galactosidase activities were virtually the same in lactose-grown cells of SMQ-301 and SMQ-301K01, and growth of S. thermophilus SMQ-301K01 and S. salivarius on galactose led to higher levels of galactokinase and β-galactosidase activities than growth on lactose. These results confirmed that S. salivarius galK was efficiently translated in S. thermophilus and suggested that expression of this extrachromosomal gene did not interfere with expression of the chromosomal lac operon.

TABLE 3.

β-Galactosidase and galactokinase activities

Bacterium and enzyme	Mean sp acta ± SE after growth on:
	Lactose	Galactose
SMQ-301
β-Galactosidase	1,268 ± 223
Galactokinase	103 ± 48
SMQ-301K01
β-Galactosidase	1,021 ± 376	2,506 ± 250
Galactokinase	233 ± 50	567 ± 74
S. salivarius
β-Galactosidase	490 ± 36	678 ± 93
Galactokinase	1,174 ± 189	1,841 ± 241

^aSpecific activities are expressed as nanomoles per milligram of protein per minute. The values were obtained from measurements carried out in triplicate on samples from three cultures. Cells were grown in M17 medium supplemented with 0.5% sugar and harvested at the mid-exponential phase of growth.

Transcription of the gal and lac operons as a function of the growth phase.

We reported previously that gal mRNA is hardly detectable in exponentially growing Gal⁻ S. thermophilus cells but is abundant at the end of the exponential phase of growth on lactose (40). We looked at whether this phenomenon persists in a Gal⁺ strain and, if so, whether it still occurs when cells are grown on galactose. Transcription of the gal operon as a function of the growth phase was determined with a specific galT probe with SMQ-301 cells grown on lactose and with SMQ-301K01 cells grown on lactose and galactose (Fig. 3A). As previously observed, transcription of the gal genes in lactose-grown cells of strain SMQ-301 was enhanced when the cells reached the end of the exponential phase of growth (OD₆₆₀ = 0.8). Similar results were obtained with SMQ-301K01 cells grown on lactose, while the levels of mRNA remained nearly constant in galactose-grown cells. Transcription of the lac operon was also studied with a lacS-specific probe with cells grown on lactose and harvested at ODs of 0.4, 0.6, and 0.8 (Fig. 3B). The levels of lac mRNA, a 5-kb transcript, remained virtually the same in both strains. This result confirmed that the extrachromosomal galK gene did not interfere with transcription of the chromosomal lac operon.

FIG. 3.

Northern analysis of galT and lacS in S. thermophilus SMQ-301 and SMQ-301K01. (A) Total RNA (5 μg) was hybridized with a galT-specific probe. RNA was isolated from cells grown in M17 medium containing 0.5% lactose or galactose and harvested when the OD₆₆₀ reached 0.4 (lanes A), 0.6 (lanes B), and 0.8 (lanes C). 301lac, RNA isolated from S. thermophilus SMQ-301 grown on lactose; K01lac, RNA isolated from S. thermophilus SMQ-301K01 grown on lactose; K01gal, RNA isolated from S. thermophilus SMQ-301K01 grown on galactose. (B) Same as panel A with a lacS-specific probe. the estimated sizes of the transcripts are shown on the left, and the positions of the 16S and 23S rRNAs are shown on the right.

Utilization of galactose by cells growing on lactose.

During growth on lactose, Gal⁻ strain S. thermophilus SMQ-301 releases galactose into the medium (40). We thus looked at whether the Gal⁺ phenotype conferred by S. salivarius galK allows strain SMQ-301K01 to grow on lactose without expelling galactose. SMQ-301 and SMQ-301K01 expelled similar amounts of galactose during growth on lactose (Fig. 4). However, while strain SMQ-301 was able to utilize galactose only when lactose was almost totally consumed (less than 0.1 mg/ml), strain SMQ-301K01 started to utilize galactose when the medium still contained lactose (∼0.5 mg/ml).

FIG. 4.

Growth of Gal⁻ strain S. thermophilus SMQ-301 (A) and Gal⁺ strain SMQ-301K01 (B) on lactose. Cells were grown at 42°C in M17 medium containing approximately 0.2% lactose. Symbols: diamonds, growth; grey and black circles, concentrations of lactose and galactose, respectively.

This difference prompted us to look at whether galactose utilization in the presence of lactose by strain SMQ-301K01 could be improved by varying the growth conditions. Cells were first grown on either galactose or lactose and then on lactose. Similar levels of galactose utilization were observed under both conditions (data not shown). As the utilization of accumulated galactose started when the culture approached the stationary phase, we speculated that the cells, at this point in their growth, had reached a physiological status that enabled them to metabolize galactose in the presence of lactose. To test this hypothesis, SMQ-301K01 was first cultured in a medium supplemented with 0.2% lactose. Cells were withdrawn when they started metabolizing galactose and used to inoculate a lactose-containing medium. The cells still released galactose during growth and resumed galactose metabolism only when the lactose concentration in the medium decreased to approximately 0.5 mg/ml (data not shown). Since the pH of the medium decreased during growth, we hypothesized that the onset of galactose utilization was pH dependent. We therefore adjusted the pH of a lactose-containing medium to 6.1, which corresponded to the pH at which galactose utilization started, and inoculated this medium with lactose-grown cells. The cells still released galactose during growth and began to utilize galactose only when the lactose concentration reached 0.5 mg/ml (data not shown).

This lactose concentration threshold was further confirmed in three other series of experiments. First, SMQ-301K01 cells were repeatedly cultured in a medium supplemented with 0.2% lactose. The cells were transferred into fresh medium only after the lactose had been completely consumed (Fig. 5). Every addition of lactose resulted in a period of galactose accumulation, followed by disappearance of the galactose when the lactose concentration fell below 0.05%. Second, SMQ-301K01 was grown in culture medium containing 0.1, 0.2, or 0.5% lactose and the residual lactose was measured when galactose utilization started. Again, galactose consumption began only when the lactose concentration in the medium reached 0.05% (data not shown). Lastly, the growth of strain SMQ-301K01 in a medium in which the lactose concentration was maintained at 0.05% did not result in galactose accumulation (data not shown).

FIG. 5.

Repeated culturing of S. thermophilus SMQ-301K01 on lactose. A tube containing 10 ml of M17 medium supplemented with 0.2% lactose was inoculated with 100 μl of an overnight culture of S. thermophilus SMQ-301K01 grown in M17 medium containing 0.2% lactose and incubated at 42°C. When the lactose in the medium was almost totally depleted, 3 ml of this culture was used to inoculate another tube containing 7 ml of fresh M17 medium that was supplemented with lactose at a final concentration of approximately 0.2%. This process was repeated four times. Symbols: diamonds, growth; grey and black circles, concentrations of lactose and galactose, respectively. The arrows indicate the amount of lactose in each culture immediately after transfer of 3 ml of the previous culture.

Similar experiments were conducted with S. thermophilus SMQ-301 (Gal⁻) and S. salivarius (Gal⁺). Strain SMQ-301 behaved like strain SMQ-301K01, except that galactose utilization occurred only when the lactose was almost totally consumed, while S. salivarius never expelled galactose (data not shown).

These results clearly indicate that (i) increasing galactokinase activity in S. thermophilus was not sufficient to prevent galactose expulsion during growth on lactose and (ii) the amount of extracellular lactose determined the ability of S. thermophilus Gal⁺ cells to concomitantly metabolize the glucose and galactose moieties of lactose.

Milk acidification activity.

To verify whether the ability to metabolize galactose has an impact on the ability of the recombinant strain to ferment milk, we compared the acidification rates of S. thermophilus SMQ-301 and SMQ-301K01 with the Cinac milk fermentation system and a time-temperature profile that simulated the profile used for producing mozzarella cheese. The acidification rates indicated that recombinant strain SMQ-301K01 reduced the pH of milk more rapidly than did strain SMQ-301 (Fig. 6). Moreover, strain SMQ-301K01 produced a greater difference in pH between T = 0 min and T = 171 min (1.64 ± 0.05 versus 1.38 ± 0.05), a shorter lag time (25.2 ± 3.1 versus 29.4 ± 1.1 min), a shorter time required to reach the maximum rate of metabolism (50.7 ± 12.2 versus 62.7 ± 6.1 min), and a greater maximum rate of metabolism (maximum acidification rate, 0.02 versus 0.01 pH unit per min). Lastly, strain SMQ-301K01 grew faster (μ = 1.01 ± 0.16 h⁻¹) than strain SMQ301 (μ = 0.75 ± 0.03 h⁻¹) in milk.

FIG. 6.

Rates of milk acidification. The acidification kinetics were determined in triplicate with the Cinac system. The bars indicate the standard error. The temperature profile was the same as that used to produce mozzarella cheese. Symbols: inverted triangles, temperature; filled circles, rate of acidification by S. thermophilus SMQ-301; open circles, rate of acidification by S. thermophilus SMQ-301K01.

DISCUSSION

In a prior study (40), we showed that S. thermophilus SMQ-301 possesses a gal operon comprising the genes galK, galT, and galE. These genes code for the enzymes of the Leloir pathway, which catalyzes the first steps of galactose metabolism (13). In strain SMQ-301, the gal operon is efficiently transcribed into a polycistronic galKTE mRNA from a promoter located upstream from galK. Nonetheless, SMQ-301, like most S. thermophilus strains (19, 23), is unable to grow on galactose and expels galactose into the external medium when grown on lactose (40). It has been proposed that the inability to grow on galactose may result from the inability to translate the galK mRNA, depriving the cells of suitable levels of galactokinase, the enzyme that catalyzes the first reaction of the Leloir pathway. In the work reported here, we demonstrated that S. thermophilus SMQ-301 was able to grow on galactose when complemented with the S. salivarius galK gene cloned on a low-copy plasmid. Our results established that the inability of strain SMQ-301 to grow on galactose resulted from low galactokinase activity, a phenotypic defect that has previously been reported in several other S. thermophilus strains (6, 20, 35). As galactose metabolism also requires the galE and galT gene products, our results also confirmed that the S. thermophilus gal promoter was functional, as previously demonstrated by determinations of enzyme activities (10, 30, 35, 40) and transcriptional studies (40).

Previous studies conducted with spontaneous or chemically induced S. thermophilus Gal⁺ mutants revealed that these strains still excreted galactose during growth on lactose (6, 20, 35, 37). However, as the genotypes of these mutant strains were unknown, it was not possible to determine whether their phenotype resulted solely from an increase in galK expression. The presence of a functional extrachromosomal S. salivarius galK allele was the only difference between the SMQ-301 wild-type strain and the SMQ-301K01 recombinant strain. Despite its ability to metabolize and grow on galactose, SMQ-301K01 still excreted galactose during growth on lactose. These results unequivocally demonstrate that an increase in galactokinase activity is not sufficient to allow S. thermophilus to simultaneously metabolize the glucose and galactose moieties of lactose. This metabolic deficiency cannot be explained by a decrease in the transcription of the gal genes in lactose-grown cells since even galactose-adapted cells expelled galactose during growth on lactose.

Two biochemical events might be involved in the galactose expulsion phenomenon in S. thermophilus, namely, lactose transport by LacS and the conversion of β-d-galactose into α-d-galactose by galactose mutarotase (GalM). Lactose is transported into S. thermophilus by LacS, a transporter that catalyzes two modes of transport: Δp-driven lactose uptake in symport with protons and lactose-galactose exchange. The antiporter mode appears to be dominant in vivo (16). Within the cell, lactose is hydrolyzed by β-galactosidase into glucose and galactose. Since the presence of a functional galactokinase did not prevent lactose-galactose exchange, binding and expulsion of galactose by LacS are likely more rapid than galactose phosphorylation by galactokinase.

However, galactose phosphorylation requires prior transformation of β-d-galactose into α-d-galactose by galactose mutarotase GalM (13). While the affinity constant of the S. thermophilus galactose mutarotase for its substrate is unknown, it is likely similar to the K_m values of 4 and 20 mM for the E. coli and Lactococcus lactis enzymes, respectively (5, 34). These values are much higher than the affinity constant of LacS (80 μM) estimated for intracellular galactose (43). Thus, one may infer that high levels of GalM are required to efficiently compete with LacS for the binding of β-d-galactose and its conversion into α-d-galactose. The S. thermophilus galactose mutarotase is coded for by galM, a gene located downstream from galE, the last gene of the gal operon (40, 42). Transcription of galM in the Gal⁺ species S. salivarius occurs via the synthesis of an abundant 10-kb mRNA initiated from the gal promoter and covering the gal operon, galM, and the lac operon (40). As S. salivarius does not expel galactose during growth on lactose (22, 40), it may be inferred that GalM is produced in sufficient amounts to rapidly transform β-d-galactose into α-d-galactose. This would allow rapid galactose phosphorylation, preventing the recapture of free galactose by LacS. In S. thermophilus, the transcripts bearing galM are produced at very low levels (40). Thus, the inability of Gal⁺ S. thermophilus to metabolize the galactose moiety of lactose may result from its inability to synthesize appropriate levels of galactose mutarotase to rapidly produce the galactokinase substrate α-d-galactose. This may also explain why strain SMQ-301K01 grew slower than S. salivarius on galactose.

In Streptococcus mutans, expression of the galactose genes is repressed by GalR, a transcriptional regulatory protein that belongs to the LacI/GalR family (1, 2, 24). Previous analyses of the S. thermophilus and S. salivarius gal gene clusters revealed the presence of a galR gene upstream from the gal operon and a GalR binding site in the gal promoter region, which encompasses nucleotides −61 to −94 relative to the adenine of the ATG initiation codon of S. salivarius galK (40, 42). The S. salivarius galK gene transferred into S. thermophilus was under the control of its own promoter. The cloned DNA fragment did not, however, contain the GalR binding site. Since the gene was transcribed in strain SMQ301-K01, we concluded that GalR is not required to assist the S. thermophilus RNA polymerase to initiate at the S. salivarius gal promoter. Since the GalR binding site of the gal promoters of S. salivarius 25975 and S. thermophilus SMQ-301 are almost identical and since the GalR proteins of the two species are more than 75% identical, our results suggest that the GalR proteins of the two species are repressors of the gal operon, as found in S. mutans (1, 2). Amazingly, inactivation of galR in S. thermophilus CNRZ 302 prevents expression of the gal operon, suggesting that, in CNRZ 302, GalR is a transcriptional activator (42). We have no explanation for this discrepancy, but differences between strains cannot be excluded since phenotypic and genotypic heterogeneity among S. thermophilus strains has been observed (15).

Northern experiments also revealed higher levels of gal mRNAs in the stationary phase of lactose-grown cells (Fig. 3). Conversely, in SMQ-301K01 cells grown on galactose, gal mRNA levels remained roughly the same throughout growth. We also observed that β-galactosidase and galactokinase activities were higher in galactose-grown cells than in lactose-grown cells. These results may be explained by the intracellular concentrations of the gal and lac operon inducers. In S. mutans, the intracellular inducer of the gal operon is free galactose (1). Since galactose is expelled into the medium when S. thermophilus is grown on lactose, one may hypothesize that repression of the gal and lac operons is maintained to some extent as long as lactose is present in the medium. This phenomenon would not occur during growth on galactose, resulting in nearly optimal transcription of these operons at every stage of growth. Since GalR also controls the lac operon (42), it is expected that the rate of transcription of the lac genes would also vary with the growth phase. Surprisingly, no discernible change in the level of lac mRNA was observed in cells harvested at different times during growth. Interestingly, a catabolite-responsive element sequence that is involved in catabolite repression in gram-positive bacteria (31) is present in the promoter region of both operons (40), and CcpA, a protein that binds to the catabolite-responsive element, has been shown to be involved in the regulation of the S. thermophilus lac operon (41). A different mode of regulation of the gal and lac operons by CcpA may thus account for the differences in the rates of transcription observed in lactose-grown cells.

Lastly, the potential of the recombinant strain as a starter culture was evaluated by measuring several parameters during its growth in milk. For all of the properties analyzed, the recombinant Gal⁺ strain performed better than the wild-type Gal⁻ strain. Although the S. thermophilus strain engineered for this study is not a food grade organism (the plasmid construct contains an antibiotic resistance gene), the data suggest that derivation of food grade equivalent strains may provide an advantage as starter cultures for manufacture of mozzarella cheese and other fermented dairy foods.