The Pinholin of Lambdoid Phage 21: Control of Lysis by Membrane Depolarization

Abstract

The phage 21 holin, S²¹, forms small membrane holes that depolarize the membrane and is designated as a pinholin, as opposed to large-hole-forming holins, like S^λ. Pinholins require secreted SAR endolysins, a pairing that may represent an intermediate in the evolution of canonical holin-endolysin systems.

For most phages, the termination of each infection cycle is the strictly programmed and regulated lysis of the host, brought about by two phage-encoded proteins (28). One of these, the endolysin, is capable of degrading the cell wall, while the second, the holin, is a small membrane protein which controls endolysin function. During the assembly of progeny virions, holin molecules accumulate in the cytoplasmic membrane without damaging the host. Then, at a time dictated by their primary structure, holins trigger to disrupt the cytoplasmic membrane. For many phages, like λ and T4, this event releases to the periplasm an endolysin that has accumulated fully folded and enzymatically active in the cytosol. By contrast, phages P1 and 21 encode endolysins that are exported by the host sec system and accumulate in the periplasm as enzymatically inactive proteins tethered to the membrane by an N-terminal SAR (signal anchor-release) domain (25, 26). These SAR endolysins become enzymatically active when their SAR domains exit the membrane to generate the mature, soluble form in the periplasm. This process occurs spontaneously at a low rate but is greatly accelerated when the cytoplasmic membrane is deenergized. Thus, for phages encoding SAR endolysins, holins need only to depolarize the membrane in order to fulfill their role in controlling the timing of lysis. The formation of large membrane lesions like those resulting from S^λ triggering (22) would not be necessary. This raises the possibility that holins serving SAR endolysins may not function with canonical, soluble endolysins to effect host lysis.

R²¹ expression allows holin-independent lysis by phage λ.

To further characterize the phage 21 holin and endolysin, which are the products of the genes S²¹68 and R²¹, respectively (14), we replaced the lysis genes of λ with the lysis genes from phage 21 by homologous recombination between λΔSR and plasmids carrying either S²¹68R²¹ or S²¹68_amR²¹ (Table 1). The recombinant λ S²¹68R²¹ formed plaques of uniform size (Fig. 1A) that were slightly smaller than those produced by λ on the same host (not shown). In liquid culture, the lysis of synchronously induced λS²¹68R²¹ lysogens exhibited a saltatory character, indicative of the synchronous triggering of the S²¹68 holin which, in the absence of endolysin function, results in a cessation of growth (Fig. 2A). With nonsuppressor hosts, the behavior of λS²¹68_amR²¹ was different with respect to both phenotypes. First, the plaques formed by λ S²¹68_amR²¹ were small and showed a considerable size variation; this heterogeneity persisted when phage from large and small plaques were replated (Fig. 1B to D). Thus, like phage P1 but unlike λ and T4, the S²¹ holin gene is nonessential for plaque formation (7, 8, 10, 27). Second, for induced λS²¹68_amR²¹ lysogens, lysis in liquid culture is less saltatory, requiring 30 to 40 min for completion as assessed by monitoring the decrease in culture A₅₅₀ (Fig. 2A).

TABLE 1.

Bacterial strains, bacteriophages, and plasmids used in this study

Strain, phage, or plasmid	Genotype and relevant features	Source or reference
Strains
MC4100	E. coli K-12 F−araD139 Δ(argF-lac)U169 rpsL150 relA1 flbB5301 deoC1 ptsF25 rbsR	18
MDS12	MG1655 with 12 deletions, totaling 376,180 nucleotides, including cryptic prophages	11
MDS12 tonA::Tn10	tonA::Tn10 transductant of MDS12	This study
Phages and prophages
λ Δ(SR)	λ stf::cat::tfa cI857 Δ(SR)	19
λ SamR+	λcI857 Sam7 (amber in position 56 of S)	Laboratory stock
λ S+Ram	λcI857 Ram54am60 (ambers in positions 26 and 73 of R)	Laboratory stock
λ S2168R21	λ-21 hybrid phage carrying S68RRzRz1 of phage 21 under λPR′	This study
λ S2168R21am	λ-21 hybrid phage carrying S68Ram,amRzRz1 of phage 21 under λ PR′; the R21 gene carries am codons in positions 39 and 42	This study
λ S2168amR21	λ-21 hybrid phage carrying S68amRRzRz1 of phage 21 under λ PR′; the S2168 gene carries an am codon in position 46	This study
Plasmids
pRE	Vector with pBR322 origin, carrying λ late promoter PR′	Supplemental information of Park et al. (14)
pS105	pBR322 origin, PR′ promoter, and S105RRzRz1 from λ; in the S105 gene, the first codon of S is converted to a CTG; this allele produces only the S105 holin gene product	19
pTP2	S105RRzRz1 of pS105 replaced with S68RRzRz1 of phage 21; in the S2168 gene, the first codon of S21 is converted to a CTG; this allele thus produces only the S2168 gene product	Supplemental information of Park et al. (14)
pTP3	pTP2 with amber codon at position 46 of S2168	Supplemental information of Park et al. (14)
pTP4	pTP2 with amber codons at positions 39 and 42 of R21	Supplemental information of Park et al. (14)
pS105Ram	pS105 with ochre and amber codons at positions 7 and 9, respectively, of R	Laboratory stock
pJFLyz	pJF118 tac vector carrying the P1 lyz SAR endolysin gene	Supplementary information of Xu et al. (26)
pRλ	pJF118 tac vector carrying the λ R endolysin gene, analogous to pJFLyz	Supplementary information of Xu et al. (26)
pJFT4E	pJF118 tac vector carrying the Τ4 e endolysin gene	M. Xu, unpublished data
pTGS	TorA TAT leader sequence fused to SsrA-tagged GFP in pBAD33	4

FIG. 1.

The absence of S²¹68 holin contributes to heterogeneity of plaque morphology. MDS12 tonA::Tn10 was used as a host for plating the indicated λ 21 hybrid phages. (A) λS²¹68R²¹; (B) λS²¹68_amR²¹; (C and D) replatings of the small and large plaques from panel B, respectively.

FIG. 2.

The S²¹68 holin triggers but does not allow release of cytoplasmic endolysins. (A) S²¹68 supports abrupt lysis with the SAR endolysin R²¹. Lysogens of MDS12 tonA::Tn10 growing logarithmically in LB medium at 30°C were thermally induced at an A₅₅₀ of 0.2 by aerating at 42°C for 15 min and at 37°C thereafter. Prophage symbols: ▪, λS²¹68⁺R²¹; ○, λS²¹68⁺R²¹_am; •, λS²¹68_amR²¹⁺; □, λS²¹68_am R^21am. (B) S²¹68 and S^λ are not functionally equivalent. Lysis-defective lysogens of MC4100 carrying a plasmid with the indicated alleles of holin-endolysin gene pairs were grown and induced as for panel A. The lysis genes on the plasmids are under the transcriptional control of the λ P_R′ promoter, which is activated by the Q protein supplied from the induction of the lysogen, as previously described (14). Symbols: ▪, λΔ(SR) and pTP2 (S²¹68R²¹); ○, λΔ(SR) and pS105 (S^λR^λ); ▾, λS_am7R⁺ and pTP4 (S²¹68R^21am), with the arrow indicating the time of addition of CHCl₃; □, λS⁺R_am and pTP3 (S²¹68_amR²¹). (C and D) Lysogens of MDS12 tonA::Tn10 bearing the indicated prophage and plasmid, the latter carrying an endolysin gene under tac promoter control, were grown and thermally induced as for panel A, except that isopropyl-β-d-thiogalactopyranoside was also added at time zero to induce the plasmid. (C) S²¹68 does not facilitate lysis by T4 E. •, λΔ(SR) and pJFT4E (e⁺); □, λS²¹68⁺R^21am and pJFT4E. CHCl₃ was added (arrow) to both cultures (• and □) at 65 min after induction. (D) S²¹68 allows abrupt lysis by the SAR endolysin P1 lyz. •, λS²¹68R^21am and pJFLyz (lyz⁺); ○, λS²¹68R^21am and pR^λ; ▪, Δ(SR) and pJFLyz; □, λΔ(SR) and pR^λ. To verify the expression of the endolysin gene, CHCl₃ was added (arrow) to two cultures (○ and □) at 75 min after induction.

S²¹ and S^λ are not functionally equivalent.

We next designed experiments to determine if S²¹68 and R²¹ could complement the lysis defect of phages λS_amR⁺ and λS⁺R_am, respectively. Previously, we had reported that, when expressed from the pUC18 derivative pTZ18R, the S²¹ gene appeared to be the functional equivalent of S^λ (2). However, the lysis of the culture was not complete even an hour after its onset, despite the fact that the S²¹ protein was produced at supraphysiological concentrations from the very-high-copy-number plasmid. For this reason, we repeated these experiments with various alleles of S²¹68 and R²¹ transactivated from the λ late promoter on a medium-copy-number plasmid, in trans to lysis-defective prophages. This system was shown in other studies to support lysis with approximately normal timing (1, 6). As can be seen in Fig. 2B, expression of R²¹ from the plasmid complemented the lysis defect of an induced λS⁺R_am lysogen, with lysis beginning 55 min after induction and completed within 10 min. In contrast, expression of S²¹68 did not complement an induced λS_amR⁺ lysogen, despite the fact that the S²¹ holin triggered, as can be seen from the halt in cell growth at approximately 15 min after induction. Moreover, the addition of CHCl₃ resulted in immediate lysis, indicating the presence of a pool of cytoplasmic R endolysin. Similarly, unlike S^λ, S²¹68 was unable to promote the release of E, the cytosolic endolysin from phage T4 (Fig. 2C). However, coexpression of lyz, encoding the SAR endolysin from phage P1, and S²¹68 resulted in saltatory and rapid lysis of the host, a characteristic of holin-triggered lysis (Fig. 2D). This S²¹68-facilitated lysis was easily distinguished from the delayed and gradual lysis that occurs when lyz is induced in the absence of a holin (Fig. 2D) (26). Thus, the phage 21 holin facilitates lysis only when paired with SAR endolysins. We interpret this to mean that when S²¹68 triggers, it eliminates the proton motive force, causing release and activation of the membrane-tethered inactive SAR endolysin, but does not form holes in the membrane large enough to allow passage of a cytoplasmic endolysin.

Macromolecules easily pass through S^λ but not S²¹ holes.

In order to demonstrate that S^λ but not S²¹ allows the nonspecific movement of macromolecules across the inner membrane, the genes for either holin were expressed in cells producing the fluorescent periplasmic marker TorA-GFP-SsrA (4). The latter protein has the leader peptide and the first 8 amino acids of TorA fused to the N terminus of a green fluorescent protein (GFP) variant, allowing the Tat-specific secretion of the chimera to the periplasm. The SsrA sequence at its C terminus promotes the degradation by the ATP-dependent proteases ClpAP and ClpXP of any of the chimeric protein that escapes export and remains in the cytoplasm. When examined by fluorescence microscopy, a thin ring of fluorescence at the periphery of cells expressing the torA-gfp-ssrA gene is observed (Fig. 3A), indicative of the periplasmic localization of the TorA-GFP-SsrA protein. The induction and triggering of S^λ in such cells result in a uniform, diffuse fluorescence throughout the cytoplasm, indicating that the chimeric GFP has reentered the cytoplasm through the S^λ holes (Fig. 3B). Lack of degradation of the fluorescent chimera by ClpAP and ClpXP is due to the rapid depletion of ATP subsequent to the formation of S^λ holes in the inner membrane. By contrast, the induction and triggering of S²¹68 in cells with periplasmic TorA-GFP-SsrA did not cause its redistribution to the cytoplasm (Fig. 3C).

FIG. 3.

Assessing the passage of a periplasmic marker through membrane lesions generated by S105 and S²¹68. MDS12 tonA::Tn10 λΔ(SR) lysogens bearing pTGS (torA-gfp-ssrA) and either pRE (vector) (A), pS105 (S105) (B), or pTP2 (S²¹68) (C) were grown in the presence of 0.2% arabinose for 100 min to induce Tor-GFP-SsrA fusion and then thermally induced. After 1 h, cells were collected by centrifugation, washed, and immediately examined under a Zeiss Axioplan 2 imaging fluorescence microscope.

Implications for the evolution of holin-endolysin systems.

At physiological levels of expression, the holin of phage 21 is lethal and can mediate host lysis when coexpressed with cognate and noncognate SAR endolysins but not with the cytoplasmic endolysins (Fig. 2). We interpret this to mean that the S²¹ holin makes holes too small to allow the passage of proteins the size of phage endolysins (∼15 kDa) from the cytoplasm to the periplasm. We propose that holins of this type be called “pinholins” to emphasize their small hole size. We suggest that the S²¹/R²¹ gene pair, encoding a pinholin and a SAR endolysin, may represent an intermediate stage in the evolution of holin-endolysin systems. The minimum requirement for an effective phage lysis system, other than the muralytic activity itself, is a delay in lysis after the onset of late gene expression, to allow for assembly of progeny virions (28). Originally, phages may have had no dedicated lysis system at all but simply relied on the fact that redirection of the host macromolecular metabolism towards phage replication and assembly would eventually cause cellular disintegration because of a failure in the functions required for maintenance of the envelope. The most primitive dedicated lysis system could have consisted of a SAR endolysin alone. This mode would provide a lysis delay because of the gradual release and activation of the membrane-tethered endolysins. In addition, due to their sensitivity to membrane depolarization, the SAR endolysins would provide a sentinel function (20) to effect immediate lysis in the event of any condition which disrupted the integrity of the membrane, including superinfection, which, in the case of myophage or siphophage, results in a temporary depolarization of the cytoplasmic membrane concomitant with DNA injection (12, 13).

However, a lysis system employing a SAR endolysin alone would be inherently inferior to canonical holin-endolysin systems for two reasons. First, because canonical holins function with cytoplasmic endolysins, the muralytic activity elaborated during the infection cycle can be produced in great excess. Not only does this mean that once the holin triggers, host lysis occurs in a matter of seconds, reducing the dwell time in the dead, nonproductive host to a minimum, but also it means that lysis timing is completely dependent on the holin. Secondly, it has been shown that most missense changes in holin proteins alter the timing of lysis, unpredictably advancing or retarding the instant of triggering (5, 9, 15-17, 24). This malleability would provide a distinct evolutionary advantage, since maintaining fitness under different environmental conditions requires the ability to tune the timing of lysis; for example, increased and decreased host cell densities favor shorter and longer infection cycles, respectively (3, 21, 23). By contrast, SAR endolysins would offer few mutational paths to advance or retard the timing of lysis. A small number of mutations affecting active site residues would meaningfully change the k_cat and, similarly, only mutations in the N-terminal SAR domain would be expected to alter the kinetics of membrane release. This combination of malleability and uniformity in the canonical holins would make selection of a new, fitter holin allele, with altered lysis timing, much more rapid when the selective environment changed to the advantage of a shortened or lengthened vegetative cycle.

These advantages are partially replicated in the phage 21 system, with a SAR endolysin and a pinholin, which exhibits a more saltatory lysis profile than the SAR endolysin alone (Fig. 2A). This presumably derives from the quantitative activation of the SAR endolysin at the time of the pinholin triggering, rather than relying on its gradual spontaneous activation (Fig. 2). However, the pinholin has a restricted tuning range, since the SAR endolysin itself will cause lysis at some point after induction, irrespective of the pinholin allele. Moreover, the level of muralytic activity has to be much lower to avoid inappropriately early lysis; in fact, the specific activity of the classic T4 gpe endolysin is >10³-fold higher than that of P1 Lyz (25). Presumably, further evolutionary optimization would involve, first, alterations in the holin that would allow it to form protein-sized membrane lesions and, second, loss of the N-terminal SAR domain from the endolysin.

About 25% of phages possess SAR endolysins, as judged by manual inspection of endolysin genes in the currently available phage genomes (I.-N. Wang and R. Young, unpublished). However, the holin of phage P1, which is paired with the SAR endolysin Lyz, is a canonical holin that can complement defects in λ S (M. Xu, D. K. Struck, and R. Young, unpublished), and so there is no way, a priori, to determine how many of the SAR endolysins are served by pinholins. The canonical holins thus have a selective advantage not only for fitness, in terms of the mechanistic advantages of holin function, but also because they can function with either cytoplasmic endolysins or SAR endolysins, whereas the pinholin genes can function only with SAR endolysins. It will be interesting to see whether the S²¹ pinholin gene can be mutated to a larger hole size, allowing passage of a fully folded cytoplasmic endolysin like R^λ, and thus attain the universal functionality of a canonical holin.