A Protein Antibiotic in the Phage Qβ Virion: Diversity in Lysis Targets

See allHide authors and affiliations

Science  22 Jun 2001:
Vol. 292, Issue 5525, pp. 2326-2329
DOI: 10.1126/science.1058289


A2, a capsid protein of RNA phage Qβ, is also responsible for host lysis. A2 blocked synthesis of murein precursors in vivo by inhibiting MurA, the catalyst of the committed step of murein biosynthesis. An A2-resistance mutation mapped to an exposed surface near the substrate-binding cleft of MurA. Moreover, purified Qβ virions inhibited wild-type MurA, but not the mutant MurA, in vitro. Thus, the two small phages characterized for their lysis strategy, Qβ and the small DNA phage φX174, effect host lysis by targeting different enzymes in the multistep, universally conserved pathway of cell wall biosynthesis.

Double-stranded DNA phages encode at least two and as many as five proteins, including a muralytic enzyme, to effect a precisely scheduled host lysis. At a genetically programmed time, one protein, the holin, acts to permeabilize the membrane, allowing the muralytic enzyme to degrade the cell wall. Other proteins serve as negative regulators of the holin or as agents to destabilize the outer membrane (1). In contrast, theMicroviridae, Leviviridae, andAlloleviridae, which are lytic phages with small, single-stranded nucleic acid genomes, have only a single gene required for lysis and do not produce a muralytic activity (2). TheMicrovirus [single-stranded DNA (ssDNA)] φX174 has only 10 genes and produces a single lysis protein, E, which is encoded by a 91-codon reading frame embedded in the +1 register within the essential morphogene D. Similarly, the Levivirus (ssRNA) MS2 has only four genes; the lysis gene L, overlapping the coat and replicase genes in the +1 register, encodes a 75–amino acid membrane protein. In contrast, there is no separate lysis gene in theAllolevivirus (ssRNA) Qβ. Instead, synthesis of the maturation or A2 protein, a single-copy virion protein responsible for absorption to the sex pilus and protection of the virion RNA against external ribonuclease, is also necessary and sufficient for lysis (Fig. 1A). Because of the absence of a muralytic activity, the mode of action of these single-gene lysis systems has been mysterious and controversial (2, 3). Recently, we demonstrated that E is lytic for the same reason that the fungal cell wall antibiotic, mureidomycin, is lytic; E causes lysis by acting as a specific inhibitor of MraY, a membrane-embedded enzyme, conserved throughout eubacteria, that catalyzes the synthesis of the first lipid-linked intermediate in peptidoglycan synthesis (4, 5) (Fig. 1B). This finding led us to wonder whether other simple phage lysis systems also encode proteins that function as cell wall antibiotics and, if so, at which step they act. To address these questions, we investigated the lytic mechanism of A2 using a combined genetic and biochemical approach.

Figure 1

(A) The genome of the ssRNA phage Qβ. The coat protein is the major virion structural protein; A1 is a product of read-through of the leaky UGA stop codon of the coat protein and is a minor component of the virion. Replicase is the viral-encoded subunit of the RNA-dependent RNA polymerase. A2, or maturation protein, is present in one copy per virion; it is required for adsorption to the sex pilus of the host and is also responsible for cell lysis (28). Bar, 1 kb. (B) Pathway for murein biosynthesis [adapted from Nanninga (29)]. The committed step, the addition of the pyruvyl moiety to UDP-GlcNAc, is catalyzed by the product of the murA gene, UDP-N-acetylglucosamine enol-pyruvyl transferase. The externalization of the undecaprenol-pyrophospho-MurNAc–GlcNAc disaccharide pentapeptide is catalyzed by an unknown flippase activity, indicated by a question mark. The pathway by which the disaccharide pentapeptide is covalently linked into the murein is catalyzed by the high molecular weight penicillin binding proteins (PBPs), although the molecular details of the pathway are uncertain. The undecaprenol pyrophosphate generated by the PBPs is recycled back to undecaprenol phosphate and then flipped to the cytoplasmic face, again by an unknown flippase activity (30). The undecaprenyl moiety is represented by a wavy line embedded in the cytoplasmic membrane (CM). The enzyme inhibited by the φX174 E protein, MraY, is indicated by the blunt arrow.

In E-mediated lysis, incorporation of the specific peptidoglycan precursor [3H]diaminopimelate (DAP) into murein is blocked long before lysis. Labeling cells expressing a clonedA2 gene gives similar results, with incorporation stopping at least 20 min before lysis is detectable (Fig. 2). However, unlike E, A2 also causes a subsequent degradation of newly synthesized peptidoglycan. Also unlike E, which by blocking MraY causes uridine 5′-diphosphate–N-acetylmuramic acid (UDP-MurNAc)–pentapeptide, the last cytoplasmic precursor in the peptidoglycan synthesis pathway, to accumulate (5), no DAP-containing precursors accumulate in A2-inhibited cells (Table 1). Thus, MraY is not the target of A2, which instead must block either the step where DAP is added to the oligopeptide or one of the five steps preceding it; the enzymes catalyzing these steps are encoded by murA,murB, murC, murD, murE, and murI (Fig. 1B).

Figure 2

A2 expression blocks murein synthesis before lysis. ET505 cells carrying either pA2, with the A2 gene under tacPO control, or the isogenic lacZ plasmid, pJFlacZK (25), were labeled and induced as described (5). Briefly, cells were grown in minimal M9 glucose media in 250-ml culture flasks at 37°C to an A550 of 0.3. A portion of each culture was transferred with constant aeration to a prewarmed 50-ml flask containing [3H]DAP at a final activity of 5 μCi/ml. After a 10-min prelabeling period, both labeled and unlabeled cultures were induced with 1 mM IPTG. Culture growth was monitored asA550 in the unlabeled culture, and [3H]DAP incorporation into cell wall was monitored in the labeled culture as radioactivity insoluble in boiling 4% SDS, as described (5).

Table 1

Cell wall and precursor labeling in vivo.

View this table:

To discriminate between these candidates, we took a genetic approach. After induction of a culture of cells carrying a plasmid-borneA2 gene, spontaneous survivor colonies were isolated at a frequency of 10−6. Two of 90 survivor colonies were found to be Qβr, as judged by cross-streaking and plating tests, and were designated rat1and rat2 (resistance toA-two) (6). rat1 cells accumulate plaque-forming virions at a rate indistinguishable from that of the parental cells up until the normal time of lysis and continue to accumulate virions beyond that time (Fig. 3A); thus, the defect in ratcells is only in phage-mediated lysis and is not due to a gross defect in A2 expression. To determine whether the spontaneous rat mutations mapped to one of the sixmur genes implicated by the labeling data, we took advantage of the dispersal of these genes in three clusters: murA at 71.8 min; murB and murI at 89.7 min; andmurC, murD, and murE in the cluster of murein synthesis and cell-division genes at 2.1 min (7). When the Rat mutants were transduced to kanamycin resistance (Kanr) with P1 lysates prepared from donor strains with Tn10kan markers at 1.8, 72.5, and 90.4 min, Qβ-sensitive transductants were obtained only in the case of the 72.5-min insertion, with ∼20% linkage, consistent only withrat1 being allelic to murA (8). Sequence analysis of the murA gene from rat1revealed a single missense change, Leu138→Gln; the identical change was found in rat2, whereas no change was found in the parental sequence (9). Multicopy clones ofmurA delay lysis after infection with Qβ (Fig. 3B). The basal level of expression of murA+ results in a considerable delay of lysis, whereas induction of themurA+ clone abolishes macroscopic lysis entirely. In contrast, even the basal level of themurArat allele abolishes lysis. Similar results were obtained for inductions of a plasmid clone ofA2 in which the phage gene is under the control of the tac promoter; lysis begins about 20 min after induction in cells carrying only the chromosomal copy ofmurA, but is not observed with a multicopy clone ofmurA or if a single copy of murArat1 is present (10). These data indicate that the A2protein effects lysis by titrating MurA, which catalyzes the committed step of murein biosynthesis. Consistent with this finding, the residue altered in rat1 is exposed at the surface near the substrate-binding cleft of MurA (11) and is thus available for direct contact with the A2 protein without requiring conformational changes in MurA.

Figure 3

(A) Host lysis, but not virion production, is compromised in the rat mutant.murA+ ormurArat1 cells were infected with Qβ at time t = 0, and both culture mass and the total intracellular and extracellular production of virions were determined at the indicated times (31). Circles:A 550; squares: total Qβ titer. Open symbols: parental; solid symbols: rat1. (B) Multicopy clones of murA and murArat1 can block Qβ lysis. Male murA+ cells carrying the plasmid vector pZE12-luc (24) or its derivatives pZE12-murA or pZE12-rat1, with murA + ormurArat1 cloned under the control of the hybrid pL-lac promoter (9), were infected at an MOI of 5 at t = 0. IPTG (final concentration of 1 mM) was added when the culture reached anA 550 of 0.2 to induce the expression of the cloned murA locus, where indicated. Triangles: vector. Solid and open circles: induced and uninduced pZE12-murA, respectively. Solid and open squares: induced and uninduced pZE12-rat1, respectively.

To confirm the genetic analysis, we compared the sugar nucleotide pool from A2-inhibited cells, prepared by gel-filtration and ion-exchange chromatography (12), to that of control cells. Upon ion-exchange chromatography, ∼70% of the N-acetyl sugar from the A2-inhibited culture was recovered in a single peak that was barely detectable in identically prepared extracts from a control culture. The elution position of this peak was identical to that of authentic UDP–N-acetylglucosamine (UDP-GlcNAc) and eluted at a much lower salt concentration than the more acidic UDP-MurNAc derivatives involved in cell wall biosynthesis. The material in this peak was subjected to mild acid hydrolysis, and the sugar composition of the hydrosylate was analyzed by paper chromatography. A single reducing sugar was found that comigrated with authentic GlcNAc and stained purple with the aminosugar-specific Elson-Morgan reagent, as expected for an N-acetylated aminosugar. The hydrolysate was also examined by thin-layer chromatography, which revealed the presence of single ultraviolet-absorbing constituent comigrating with uridine 5′-monophosphate in both neutral and acidic solvent systems (10). Based on an extinction coefficient of 10,000 for uracil, the sugar nucleotide that accumulates afterA2 induction had anN-acetylsugar/uracil ratio of 0.86. Taken with the genetic and physiological data, the finding that UDP-GlcNAc accumulates provides conclusive evidence that MurA is inhibited duringA2 -mediated lysis in vivo. Estimates based on the absorbance at 262 nm (A 262) of the peak fractions indicated that the concentration of this UDP-GlcNAc pool was ∼100 μM in the control cells and about 600 μM in the inhibited cells.

MurA activity can be detected in vitro as UDP-GlcNAc–dependent release of inorganic phosphate (Pi) from phosphoenolpyruvate (PEP) (13). Efforts to obtain purified A2 protein were unsuccessful owing to the insolubility of the overproduced protein. Unexpectedly, however, inhibition of MurA could be demonstrated with Qβ virions purified in a CsCl gradient. Addition of purified virions abolished MurA activity in extracts prepared from cells expressingmurA + from a plasmid (Fig. 4) (14). In extracts prepared from cells expressing murArat1 from a plasmid and murA + from the chromosome, less than 20% inhibition was observed, reflecting the presence of A2-sensitive MurA molecules from the chromosomal locus. We conclude that the resistance of the mutant MurA to A2inhibition provides the cell with resistance to A2-induced lysis.

Figure 4

Purified Qβ virions inhibit MurA, but not MurARat1. MurA activity was assayed as UDP-GlcNAc–dependent Pi release from PEP in crude, cell-free extracts prepared from cells expressing multicopymurA+ or murArat1 , either with buffer or purified Qβ virions added (14). The fractional inhibition of the activity in the extracts from the multicopy murArat1 cells presumably reflects the small proportion of wild-type MurA enzyme encoded by the chromosomalmurA locus in these cells.

The mode of inhibition by A2 remains to be determined. The simplest notion is that the virion-associated A2 binds to MurA and that the rat mutation destabilizes the complex. On the basis of the established turnover number for MurA (13), each 50-μl reaction contained about 2.5 × 1011 MurA enzyme molecules. Inhibition was quantitative with the addition of 3 × 1012 virions, but was not observed with 10-fold fewer virions (15), which indicates that the dissociation constant of the virion-MurA complex would be in the 10 nM range. This tight binding would result in a titration of MurA activity as virions accumulate. Kozak and Nathans (16) showed that the maturation protein of RNA viruses accompanies the RNA into the cell, whereas the rest of the virion is discarded in the medium. Thus, the A2 protein, like fungal cell-wall antibiotics, can come from outside the cell to inhibit the cytoplasmic synthesis of murein precursors. However, based on the fraction of A2-sensitive MurA activity observed above (Fig. 4), we estimate that cells with a single copy ofmurA contain about 1500 MurA molecules; this ensures that even at a high multiplicity of infection (MOI), where 30 or more virions can infect a single host (17), substantial inhibition of cell wall synthesis by the virion-associated A2 does not occur until virion production is advanced well beyond the eclipse period.

We have recently isolated Qβ por (plateson rat) mutants that overcome themurArat plating block (18). Potentially, these mutants may help distinguish which regions of A2 are involved in MurA inhibition. Moreover, the ease of isolating these mutants raises the possibility that allele-specific inhibitors can be obtained, if more rat alleles are isolated.

Now two single-gene lysis systems have been analyzed in detail; one, φX174 E, inhibits MraY, which catalyzes the formation of the first lipid-linked murein precursor, and the second, QβA2 , inhibits MurA, which catalyzes the committed step in the murein biosynthesis pathway. We propose that this is a general strategy, and that, in the absence of a muralytic enzyme, the only way for phage to compromise the murein sacculus sufficiently to effect host lysis is to interfere with peptidoglycan synthesis during growth. We predict that an analogous mode of action will be found for the remaining unresolved single-gene lysis system known for male-specific coliphages: the MS2 L gene. Preliminary results in this laboratory indicate that the target of L is a different gene in the same pathway (19). Small-genome phages thus appear to accomplish host lysis by elaborating polypeptides that inhibit murein synthesis at different steps. This raises the attractive possibility that DNA-encoded oligopeptide antibiotics, subject to facile genetic manipulation, might be designed on the basis of these lysis proteins. Moreover, given that all three prototype small-genome phages attack three different steps of the murein synthesis pathway, it also suggests that a search for new classes of small, lytic bacteriophages, not only of Escherichia colibut also for other bacteria, is in order.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: ryland{at}


View Abstract

Navigate This Article