Suppression of the ptsH Mutation in Escherichia coli and Salmonella typhimurium by a DNA Fragment from Lactobacillus casei

Abstract

A DNA fragment from Lactobacillus casei that restores growth to Escherichia coli and Salmonella typhimurium ptsH mutants on glucose and other substrates of the phosphoenolpyruvate:carbohydrate phosphotransferase system (PTS) has been isolated. These mutants lack the HPr protein, a general component of the PTS. Sequencing of the cloned fragment revealed the absence of ptsH homologues. Instead, the complementation ability was located in a 120-bp fragment that contained a sequence homologue to the binding site of the Cra regulator from enteric bacteria. Experiments indicated that the reversion of the ptsH phenotype was due to a titration of the Cra protein, which allowed the constitutive expression of the fructose operon.

In bacteria, a large number of sugars and hexitols are transported by the phosphoenolpyruvate (PEP)-dependent phosphotransferase system (PTS). The PTS consists of different proteins which catalyze the phosphorylation of sugars by PEP (for reviews, see references 14 and 16). The first member of the phosphorylation cascade, enzyme I (EI), is a soluble protein which phosphorylates the second component, HPr, on a catalytic histidine residue (P-His-HPr). P-His-HPr can transfer its phosphoryl group to a number of sugar-specific EIIs which catalyze the transport and concomitant phosphorylation of their substrates. In gram-positive bacteria, P-His-HPr is involved not only in sugar transport, but it has also been demonstrated that it can transfer its phosphate to other proteins, including enzymes such as the glycerol kinase of Enterococcus and the transcriptional antiterminators LicT, SacT, and SacY of Bacillus subtilis (5). In each case, phosphorylation stimulates the activity of these proteins. HPr of gram-positive bacteria can also be phosphorylated on a serine residue at the expense of ATP (phosphorylation catalyzed by a metabolically activated HPr kinase). The P-Ser-HPr protein plays a role in the control of sugar transport by regulating the processes of inducer exclusion (inhibition of sugar permeases) and inducer expulsion (activation of intracellular sugar-phosphate phosphatases) in gram-positive organisms (5). P-Ser-HPr is also involved in catabolite repression, acting as a corepressor with the transcriptional regulator CcpA.

We are interested in the study of the regulation of sugar metabolism of the lactic acid bacterium Lactobacillus casei, a species normally used in industry as a starter for dairy products. In this context, the gene encoding a global regulator of carbon metabolism in this organism, ccpA, was isolated and its involvement in the regulation of the expression of the lactose utilization system, the lactose PTS, was studied (9, 15). Due to the key regulatory role of HPr, we tried to isolate the gene encoding this protein, ptsH, from L. casei by complementation of Escherichia coli ptsH mutants. We based this approach on the fact that this cloning strategy was used to isolate the ptsHI operon from Bacillus stearothermophilus (11) and on the fact that the ptsHI genes from Streptococcus mutans and Staphylococcus carnosus were shown to complement E. coli pts mutants (1, 10). We obtained some clones from an L. casei genomic library that shared a common DNA region and which restored the growth on PTS sugars of an E. coli ptsH mutant. However, sequencing of the complementing DNA fragment revealed the absence of a ptsH homologue. In this report, we describe the characterization of the process of complementation in these clones.

Cloning of a DNA fragment from L. casei that restored PTS activity in an E. coli ptsH mutant.

E. coli TP2880 (Table 1) was transformed with an L. casei genomic library in the vector pJDC9. Transformants were plated on MacConkey agar plates with 0.5% glucose. A total of six colonies that were red were selected (the red color indicates that glucose was being fermented). All the complemented transformants contained plasmids with overlapping inserts. One of these plasmids, called pINC7, was chosen for the study. Transformants of TP2880 bearing pINC7 regained the ability to use a number of PTS substrates as the sole carbon source, such as glucose, mannose, mannitol, and N-acetylglucosamine, on minimal medium. PTS activity could also be detected in these transformants by measuring phosphorylation of [U-¹⁴C]methyl-α-d-glucopyranoside by PEP in cell extracts (19). PTS activity in TP2880/pINC7 was 4.19 nmol of methyl-α-d-glucopyranoside phosphorylated/min per mg of protein, while it was 0.53 nmol/min per mg of protein in the control TP2880/pJDC9. Subcloning of the 5.3-kb insert of pINC7 allowed the isolation of pINC72, which contained a Sau3AI/partial-EcoRI 1.7-kb fragment, which maintained the complementation ability (Fig. 1A).

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant characteristicsa	Source or reference
Strains
E. coli TP2880	ptsH Tn10	A. Danchin
E. coli TP2811	xyl argH1 ΔlacX74 aroB ilvA Δ(ptsHI crr)	13
E. coli HK881	ptsH fruR fruB	H. L. Kornberg
S. typhimurium SB2226	trpB223 ptsH38	4
S. typhimurium SB1690	trpB223 ptsI34	4
Plasmids
pUC19	Cloning vector; Apr	Pharmacia
pJDC9	Cloning vector; Emr	2
pACYC177	Cloning vector; Apr Kmr	NEBb
pINC7	pJDC9 with 5.3-kb insert from L. casei	This study
pINC72	pJDC9 with 1.7-kb Sau3AI/partial-EcoRI fragment from pINC7	This study
pFRU	pUC19 with 120-bp SphI-PvuI fragment from pINC72	This study
pFRUm	pFRU with CG in Cra site changed to TA	This study
pBCP30	pACYC177-containing cra gene from E. coli	7
pBCP35	pACYC177-containing cra gene from S. typhimurium	7
pBCP381	E. coli ptsH under the control of trc promoter	21

^aAp^r, ampicillin resistance; Em^r, erythromycin resistance; Km^r, kanamycin resistance. 

^bNEB, New England BioLabs. 

FIG. 1.

(A) Complementation of the fermentation ability of E. coli TP2880 for PTS sugars by different DNA fragments from L. casei. The solid black arrow is the ORF coding for a PriA homologue. Symbols: +, complementation; −, no complementation. (B) Comparison of the Cra binding site present in the fragment cloned in pFRU with established Cra sites in E. coli. The underlined base is the only deviation from the consensus sequence. The base changes in pFRUm are indicated over the pFRU sequence. The sequences of the first Cra site in the fru operon (Eco fruB O1) and the (Eco pps) Cra site in the phosphoenolpyruvate synthase gene are shown. In the consensus sequence, R is A or G, S is C or G, W is A or T, H is A or T or C, and N is any nucleotide.

The DNA insert in pINC72 did not contain any gene related to the PTS.

The 1.7-kb insert carried by pINC72 was sequenced. Surprisingly, no ptsH gene or other genes related to the PTS could be found. Instead, a 1,458-bp open reading frame (ORF) was found that could encode a protein that showed homology to the PriA protein of E. coli and Haemophilus influenzae (43.7 and 36.5% identity in the 222-amino-acid sequence of the C-terminal part, respectively). This finding was confusing, as the PriA protein is involved in DNA replication (12), and no function related to those of the PTS had been reported for it. Strain TP2880 was described as a ptsH mutant obtained by Tn10 transposon insertion. In order to test the stability of this insertion, Southern hybridization experiments with a Tn10-derived probe, as well as phage P1 transduction of the tetracycline marker to other E. coli strains, were performed. Data revealed that Tn10 was not inserted in ptsH but 40 kb upstream of it (data not shown). In order to investigate the TP2880 strain, it was transformed with plasmid pBCP381 (21), which contains the E. coli ptsH gene under the control of the inducible trc promoter and growth on minimal medium plus glucose was restored. Furthermore, amplification of the TP2880 ptsH gene by PCR and sequencing showed the presence of a G-to-C change that produced a substitution of the alanine in position 65 of HPr for a proline. These data indicated that TP2880 was a true ptsH mutant. Moreover, pINC72 plasmid could also complement a Salmonella typhimurium ptsH mutant, strain SB2226. This complementation needed a functional EI protein, as no complementation was observed in E. coli TP2811 (ptsHI) or S. typhimurium SB1690 (ptsI) with pINC72.

The gene present in pINC72 was not required for complementation of the ptsH mutation.

The putative priA gene present in pINC72 was inactivated by inserting a chloramphenicol resistance cassette (Fig. 1A). This new construct could still complement strain TP2880. The same situation was observed when deletions were introduced in the 5′ or 3′ end of the gene (Fig. 1A). These data unambiguously showed that the priA homologue was not needed for complementation. The complementation capacity was located in a DNA fragment spanning approximately positions 300 to 960 in its sequence (Fig. 1A). No other ORF could be found in this area.

Titration of the Cra repressor by an L. casei DNA fragment.

It has been reported that revertants of the ptsH mutation that regained the ability to use PTS sugars can be isolated from enteric bacteria (3, 7). These revertants have secondary site mutations in the gene coding for the Cra regulator (formerly named FruR) (20). cra mutants are constitutive in the expression of the fru operon that contains the genes coding for the EIICB^Fru PTS transport element, fructose-1-phosphate kinase and the diphosphoryl transfer protein (DTP). The latter protein has two domains, an N-terminal EIIA^Fru and a C-terminal domain called FPr and homologous to HPr. The FPr domain, when constitutively expressed (i.e., in cra mutants), can replace HPr and allows the cells to grow on PTS substrates (6). If the fru operon were deregulated in E. coli TP2880 or S. typhimurium SB2226 after transformation with pINC72, these cells should be able to constitutively transport fructose when present at low concentrations, due to the high affinity of EIICB^Fru for fructose. Figure 2 shows the uptake of 0.25 mM [U-¹⁴C]fructose by SB2226 bearing pINC72 or pJDC9 and grown under inducing (fructose) and noninducing (glycerol) conditions. Fructose transport in SB2226/pJDC9 was induced by fructose to a level 12-fold higher than in SB2226/pJDC9 grown on glycerol. In SB2226 containing the pINC72 plasmid, transport was high and at the level of the fructose-induced SB2226/pJDC9, irrespective of the carbon source used. Basically, identical results were obtained when E. coli TP2800 was used in similar assays. In this E. coli strain, the induction by fructose was only sixfold higher but fructose transport of TP2880/pINC72 grown on glycerol was threefold higher than in TP2880/pJDC9 grown in the same medium (data not shown). These data indicated that the presence of pINC72 or a subclone of it covering positions 300 to 960 of the L. casei DNA insert led to a constitutive expression of the fru operon and hence to a production of the DTP protein. This was supported by the fact that pINC72 did not complement E. coli HK881, a double mutant (ptsH fruB) defective in both HPr and DTP. According to this, pINC72 transformants behaved like cra mutants.

FIG. 2.

[U-¹⁴C]fructose uptake in S. typhimurium SB2226 transformed with pJDC9 or pINC72 and grown with 0.5% fructose or glycerol. The assay was performed with cells resuspended in a solution of 50 mM sodium phosphate (pH 7.4) and 5 mM MgCl₂ at an optical density at 550 nm of 0.3 and 37°C. At time zero, [U-¹⁴C]fructose (0.3 mCi/mmol) was added to a final concentration of 0.25 mM. At different times, 1-ml aliquots were withdrawn and rapidly filtered through 0.45-μm-pore-size filters and washed. The radioactivity in the filters was determined by liquid scintillation counting.

These facts raised the hypothesis that a sequence present in pINC72 could be titrating the Cra regulator by directly binding to it. A detailed analysis of the sequence from positions 300 to 960 in the insert carried by pINC72 revealed the presence of a 12-bp sequence homologous to the DNA binding site of Cra (Fig. 1B) (18). This sequence had only one deviation from the consensus sequence and also conserved the GC pair in its 5′ end known to be important for Cra recognition (17). This sequence was located in a 120-bp SphI-PvuI fragment that was cloned in pUC19, giving pFRU. Transformants of strains TP2880 or SB2226 bearing pFRU were able to grow on minimal medium plus glucose. When the central CG pair of the putative Cra binding site was mutated to TA (plasmid pFRUm [Fig. 1B]), the complementation capacity was lost. These results suggested that a 14-bp sequence present in the reading frame of the L. casei priA homologue could be titrating out the Cra repressor, thus allowing a constitutive expression of the fru operon. This was confirmed by transformation of TP2880 and SB2226 carrying pFRU with pBCP30 or pBCP35. These last two plasmids overexpress the Cra protein of E. coli and S. typhimurium, respectively (7). Transformants of TP2880 or SB2226 carrying pFRU and pBCP30 or pBCP35 were not able to use glucose as the sole carbon source, while these plasmids had no effect on fructose utilization (Table 2).

TABLE 2.

Growth on minimal medium of strain TP2880 or SB2226 transformed with different plasmids^a

Plasmid(s)	Growthb on MM with:
	Glucose	Fructose
pUC19	−	+
pFRU	+	+
pFRUm	−	+
pFRU, pACYC177c	+	+
pFRU, pBCP30	−	+
pFRU, pBCP35	−	+

^aE. coli TP2880 (ptsH) and S. typhimurium SB2226 (ptsH) were transformed with different plasmids and streaked on M9 minimal medium (MM) agar plates plus 0.2% glucose or fructose. pFRU carries a 120-bp SphI-PvuI fragment containing a Cra binding site. In pFRUm, the central CG pair of the Cra site was changed to an TA by site-directed mutagenesis. pBCP30 and pBCP35 are pACYC177 derivatives which overexpress Cra from E. coli and S. typhimurium, respectively. 

^bSymbols +, growth; −, no growth. 

^cThe bla gene from pACYC177 was deleted in order to make this plasmid compatible with pFRU, a pUC19 derivative. 

The Cra protein of enteric bacteria is involved in the regulation of carbon fluxes by activating glycolytic genes in the presence of sugars and gluconeogenic genes in the presence of gluconeogenic substrates (17). This protein is displaced from its binding site by micromolar amounts of fructose-1-phosphate (the product of fructose uptake by the fructose PTS) or millimolar amounts of fructose-1,6-bisphosphate (18). cra mutants are unable to grow on gluconeogenic substrates due to a lack of activation of genes encoding enzymes involved in their metabolism by Cra (3, 8). pINC72 transformants grew normally on lactate, alanine, or acetate. This indicated that, although titration of Cra was leading to a deregulation of the fru operon, apparently there was still enough Cra to activate genes like pps, encoding PEP synthase, which allows the cells to grow on lactate. The effects of Cra titration on different promoters would largely depend on the affinity of Cra for the various DNA binding sites and its mode of action. This is supported by the finding that the chromosomal copy of cra was sufficient to activate the pps promoter in multiple copies, while multiple copies of the fru operon required multiple copies of cra to be repressed (8).

This work represents an example of how initially unexpected results can be analyzed in-depth to yield valuable conclusions on the fine regulatory balance controlling gene expression. It also stresses the need for in-depth knowledge of the genetics and regulation of certain processes related to target genes in order to understand certain unexplained results often found in cloning experiments by complementation in E. coli and S. typhimurium.

Nucleotide sequence accession number.

The nucleotide sequence reported in this study has been deposited in the EMBL/GenBank data bank under accession no. AJ006018.