Occurrence of Homologs of the Escherichia coli lytB Gene in Gram-Negative Bacterial Species

Abstract

The Escherichia coli LytB protein regulates the activity of guanosine 3′,5′-bispyrophosphate synthetase I (RelA). A Southern blot analysis of chromosomal DNA with the E. coli lytB gene as a probe revealed the presence of lytB homologs in all of the gram-negative bacterial species examined but not in gram-positive species. The lytB homologs from Enterobacter aerogenes and Pseudomonas fluorescens complemented the E. coli lytB44 mutant allele.

Amino acid deprivation of Escherichia coli activates guanosine 3′,5′-bispyrophosphate (ppGpp) synthetase I (RelA), the ribosome-associated enzyme encoded by the relA gene (see reference 5 for a review). The resulting accumulation of ppGpp is believed to mediate the inhibition of a diverse array of metabolic processes comprising the phenomenon known as the stringent response. The stringent response is phenotypically suppressed by treatment of amino acid-deprived bacteria with certain ribosome inhibitors, e.g., chloramphenicol, which inhibit the activation of RelA. Phospholipid synthesis is one process which is inhibited during the stringent response. Heath et al. (8) have presented evidence indicating that this is due to the inhibitory action of ppGpp on sn-glycerol-3-phosphate acyltransferase. Cell wall peptidoglycan synthesis and lysis induced by treatment with β-lactam antibiotics are also inhibited during the stringent response, and this accounts for the well-known penicillin tolerance of amino acid-deprived relA⁺ E. coli (16). Peptidoglycan metabolism has been shown to be obligately coupled to phospholipid synthesis in both growing and amino acid-deprived bacteria (6, 14–16). Therefore, the inhibition of peptidoglycan synthesis and the induction of penicillin tolerance during the stringent response are consequences of the inhibition of phospholipid synthesis by ppGpp.

Mutations in the lytB gene of E. coli result in a temperature-sensitive lysis-defective phenotype and temperature-dependent penicillin tolerance (17). The lytB mutants are as sensitive as their parent strain to the lysis-inducing activities of β-lactam antibiotics at the permissive temperature. On the other hand, although growth of the mutants is inhibited by β-lactam antibiotics at the restrictive temperature, no lysis occurs. The penicillin-tolerant phenotypes of the mutants at the restrictive temperature are suppressed by inhibitors of RelA activation such as chloramphenicol or by introduction of a mutation in the relA gene (11). Furthermore, the lytB mutants accumulate ppGpp at the restrictive temperature. Therefore, the thermoinactivation of the mutant LytB protein apparently causes the activation of RelA, and the resulting accumulation of ppGpp is responsible for the observed penicillin tolerance. The function of LytB is presently unknown, but it would appear that its direct or indirect interaction with RelA is necessary to maintain RelA in an inactive form during normal growth. The lytB gene has been identified as a previously described open reading frame, originally designated orf316 (4), occurring in the E. coli ileS-lsp operon (7). The ileS-lsp operon consists of (i) a gene, recently designated ribF, which encodes riboflavin flavokinase (3); (ii) ileS (isoleucyl tRNA synthetase); (iii) lsp (prolipoprotein signal peptidase); (iv) an open reading frame designated orf149 which is proposed to encode a 17-kDa PFKB homolog; and (v) lytB. There are no obvious functional relationships among the identified gene products encoded by this operon. The in vivo transcription and mRNA 5′ end mapping experiments of Miller et al. (13) suggest that the transcription of the iles-lsp operon may be regulated in a complex fashion. The operon may be expressed from three promoters, one preceding the ribF gene, a second within ribF, and a third preceding orf149. Interestingly, the operon structure is conserved in at least two other species, Enterobacter aerogenes (10) and Pseudomonas fluorescens (9). However, recent reports of lytB homologs from other bacteria indicate that E. aerogenes and P. fluorescens are so far the only other examples of species in which there is a lytB association with the ileS-lsp operon. Table 1 summarizes the sequenced lytB homologs in the GenBank and TIGR (The Institute for Genomic Research) databases. Note that the sequence of the E. aerogenes lytB homolog (cited in reference 10) apparently has not been reported. The objectives of this study were to screen genomic DNA from a collection of bacterial species for the occurrence of lytB homologs and to determine whether the lytB homologs from E. aerogenes and P. fluorescens could complement the E. coli lytB mutation.

TABLE 1.

Reported homologs of the E. coli lytB gene

Bacterium	Accession no.a	% Nucleotide sequence identityb
P. fluorescens	M35366	66c
Haemophilus influenzae	U32781	68
B. anthracis	L48552	59d
Helicobacter pylori	HP0400	60
Synechocystis sp. strain PCC6803	U38915	62
Campylobacter jejuni	X89371	34

^aAll accession numbers are from GenBank except that of H. pylori, which is from The Institute for Genome Research. 

^bCompared to the E. coli lytB sequence. 

^cBased on a partial sequence (552 bases on 5′ end). 

^dBased on a partial sequence (370 bases of an open reading frame). 

A 1,050-bp fragment representing the entire E. coli lytB gene was amplified from plasmid pGM4 (12) by PCR. The sequences of the 5′ and 3′ primers used for this purpose were GATCCGGACTTGGAGGGAATTCATGCAATCCTGTTGGCC and AGGTAAACGCATGTTTTCTGCAAAAAATGCCGCTAACA, respectively. Both primers contained single point mutations introducing an EcoRI site at the 5′ end and a PstI site at the 3′ end. The resulting PCR product was cloned into plasmid pBDGAL-4 (Stratagene) cleaved with a combination of EcoRI and PstI to produce plasmid pXY10. The E. coli lytB probe used in Southern blot analyses was prepared from the 1,050-bp EcoRI-PstI fragment excised from pXY10. The probe was random prime labeled with digoxigenin-dUTP (Boehringer Mannheim).

All bacterial strains used in this study were from our laboratory collection. Minipreps of chromosomal DNA were prepared from bacteria by a method employing cetyltrimethylammonium bromide as described by Ausubel et al. (2). For Southern blot analysis, 3-μg samples of chromosomal DNA were digested with either EcoRI or EcoRV and separated by electrophoresis on 0.7% agarose gels. The DNA samples were depurinated, denatured, transferred to Hybond-N membranes (Amersham) with a Bio-Rad model 785 vacuum blotter, and then UV cross-linked to the membrane. Hybridization and detection of the probe were performed according to digoxigenin labeling kit protocols supplied by Boehringer Mannheim.

Figure 1 shows the results of an analysis of DNA samples from 14 gram-negative bacterial species. In Fig. 1A, high-stringency conditions (hybridization performed at 42°C followed by washes at 68°C) were used. The E. coli lytB probe hybridized equally well with EcoRI-generated DNA fragments of identical size (15 kbp) from E. coli and Shigella sonnei (Fig. 1A, lanes a and b, respectively). Slightly weaker hybridization occurred with Klebsiella pneumoniae, Salmonella typhimurium, Citrobacter freundii, and Morganella morganii (Fig. 1A, lanes c through f). These results indicate that lytB is highly conserved among these species. Discrete but weak hybridization signals were observed with DNA samples from (lane letters for Fig. 1 follow in parentheses): Proteus mirabilis (g), Serratia marcescens (h), Alkaligenes faecalis (j), Aeromonas hydrophila (m), and Chromobacterium violaceum (n). Even weaker hybridization signals occurred with Providentia rettgeri (i), Acinetobacter calcoaceticus (k), and Pseudomonas putida (l), although these results are not obvious in Fig. 1A. However, Fig. 1B shows discrete bands of 4.1, 9.4, and 8.4 kbp for P. rettgeri (i), A. calcoaceticus (k), and P. putida (l), respectively, when the stringency conditions were reduced (hybridization and washes performed at 30°C). The E. coli lytB probe was also used to locate a homolog on a 525.25-kbp CeuI fragment from the physical map of the Aeromonas salmonicida genome (18). Therefore, homologs of lytB were detected in all gram-negative bacterial species tested. On the other hand, DNA from the gram-positive bacteria Lactobacillus acidophilus, Staphylococcus aureus, Micrococcus luteus, Streptococcus lactis, and Bacillus anthracis yielded negative results under low-stringency conditions (hybridization and washes performed at 30°C). Our negative result with B. anthracis genomic DNA is notable because B. anthracis is so far the only gram-positive bacterium reported to possess a lytB homolog (1) (Table 1). This report is based on an incomplete sequence of a putative open reading frame (370 bp), and our result suggests that this open reading frame probably does not represent a lytB homolog. Sequences related to lytB were not detected in the recently completed genome sequences of Saccharomyces cerevisiae, Mycoplasma genitalium, Archaeoglobus fulgidus, and Bacillus subtilis.

FIG. 1.

Southern blot analysis of the lytB gene. Three-microgram samples of genomic DNA from several bacterial species were digested with EcoRI and analyzed with an E. coli lytB probe under high-stringency conditions (hybridization was performed at 42°C, followed by stringency washes at 68°C) (A) The blot was then stripped and reprobed under lower-stringency conditions (hybridization and stringency washes were done at 30°C) (B) Lanes: a, E. coli; b, S. sonnei; c, K. pneumoniae; d, S. typhimurium; e, C. freundii; f, M. morganii; g, P. mirabilis; h, S. marcescens; i, P. rettgeri; j, A. faecalis; k, A. calcoaceticus; l, P. putida; m, A. hydrophila; n, C. violaceum. Markers on the left represent positions of molecular size standards (HindIII-digested phage λ DNA; 23.1, 9.4, 6.6, 4.4, 2.3, 2.0, 0.56, and 0.13 kb, from top to bottom).

Plasmids pBK1 (10) and pBROC128 (9) carry the cloned lytB homologs of E. aerogenes and P. fluorescens, respectively. These plasmids were transformed into strain WV7 [lytB44 Δ(srl-recA)306::Tn10] to determine whether the lytB homologs could complement the E. coli lytB mutation. For comparison, plasmid pGM21, which carries the E. coli ileS-lsp operon (12), was transformed into strain WV7. Complementation was assessed by determining the plating efficiencies of the bacteria as follows. Dilutions of stationary-phase cultures grown in Luria broth (Difco Laboratories) were plated onto nutrient agar (Difco Laboratories) plates. For each strain, one set of dilutions was incubated at 30°C and a second set was incubated at 42°C. Colonies were counted after 48 h of incubation, and plating efficiencies (numbers of colonies at 42°C/number of colonies at 30°C) were calculated. The plating efficiency for strain WV7 was 6.7 × 10⁻⁷. When plasmid pGM21 was introduced into WV7, the plating efficiency was 0.28, indicating that pGM21 complemented lytB44. Plasmids pBK1 and pBROC128 also complemented lytB44, giving rise to plating efficiencies of 0.32 and 0.78, respectively. Complementation of lytB44 was also characterized in M9 minimal medium. Figure 2 compares the growth of strain WV7 with that of derivatives of WV7 carrying plasmids pBK1, pBROC128, and pGM21. There were no significant differences in the growth rates of the four strains at 30°C (Fig. 2A). In contrast, the growth of WV7 ceased after about one doubling at 42°C (Fig. 2B). All three plasmids carrying the lytB homologs complemented growth of WV7 at 42°C. Plasmids pGM21 and pBK1 were equally effective in this regard, but they did not support maximum growth of WV7, whereas pBROC128 did. We are not sure why the P. fluorescens homolog was more effective than the E. coli lytB gene. This may be related to the fact that the multicopy expression of the E. coli lytB gene is growth inhibitory (19), and the E. aerogenes lytB homolog may have the same effect, in light of its high degree of sequence similarity to the E. coli gene. Furthermore, pBROC128 does not contain the entire P. fluorescens lytB gene (9), and it is possible that the deletion of the C-terminal end of LytB relieves the toxicity associated with multicopy expression. On the other hand, it is not clear why the two toxic homologs had no obvious inhibitory effects at 30°C. This matter is currently under investigation. In summary, these results indicate that the heterologous lytB homologs were functional in E. coli.

FIG. 2.

Growth of the E. coli lytB44 mutant strain, WV7 (○), and derivatives of strain WV7 carrying plasmids pGM21 (•), pBK1 (▵), and pBROC128 (▴) at 30°C (A) and at 42°C (B).