The fate of phage-infected bacteria is determined by the holin, a small membrane protein that triggers disruption of the membrane at a programmed time, allowing a lysozyme to attack the cell wall. S2168, the holin of phage 21, has two transmembrane domains (TMDs) with a predicted N-in, C-in topology. Surprisingly, TMD1 of S2168 was found to be dispensable for function, to behave as a SAR ("signal-anchor-release") domain in exiting the membrane to the periplasm, and to engage in homotypic interactions in the soluble phase. The departure of TMD1 from the bilayer coincides with the lethal triggering of the holin and is accelerated by membrane depolarization. Basic residues added at the N-terminus of S2168 prevent the escape of TMD1 to the periplasm and block hole formation by TMD2. Lysis thus depends on dynamic topology, in that removal of the inhibitory TMD1 from the bilayer frees TMD2 for programmed formation of lethal membrane lesions. Like the holin S of λ, the holin of lambdoid phage 21 (S21) controls lysis by forming holes in the membrane. However, unlike Sλ, these holes are small, serving only to depolarize the membrane facilitating the release and activation of the SAR endolysin, R21. We were able to demonstrate that, unlike Sλ, S2168 forms a “pinhole”, thus macromolecules easily pass through Sλ but not S21 holes. This result again supports our interpretation: when S21 triggers, it only needs to collapse the membrane potential, thus 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 pre-folded, active cytoplasmic endolysin. The lysis defective S2168 mutant alleles were isolated throughout the S21 gene. Although the majority of lysis defective mutations occurred in the codons for the TMD2 domain, two mutations were found in the codons for the TMD1. This result suggests that only the TMD2 domain of S2168 is likely to participate in actual hole formation. One can assume that two mutant alleles of TMD1 are involved in two different interactions: (a) TMD1-TMD1 intermolecular interaction, (b) TMD1-TMD2 intramolecular interaction. We showed that there is a specific TMD1-TMD2 interaction. In terms of TMD1-TMD2 interaction, the mutated residues of the two TMD1 mutants might prevent a departure of TMD1 from TMD2, resulting in the lysis defective phenotype. Hopefully, these findings deliver some hints about the mechanism of S2168 hole formation and further provoke more extensive work which is required to provide a definite answer to many questions regarding this matter.

The purpose of the work described in this dissertation is to better understand the role of Rz and Rz1 function with respect to phage lysis. We determined using both a genetic and biochemical approach that the Rz protein is an inner membrane protein containing a single N-terminal transmembrane domain (TMD) with an Nin/Cout topology. Consistent with previous work on Rz1, the Rz1 lipoprotein was found to be localized to the outer membrane (OM). Following localization, both Rz and Rz1 form homodimers in vivo due to intermolecular disulfide formation. Despite being localized to apposing membranes, the two proteins form a complex. A small number of phages encode a potential single protein equivalent of Rz-Rz1. This protein, termed a spanin, is predicted to tether the inner and outer membranes by a single polypeptide chain. Based on complementation, it was concluded that gp11 from the phage T1 is a functional equivalent of Rz-Rz1. Gp11, and by analogy the Rz-Rz1 two-component spanin complex, threads the meshwork of the PG layer. The presence of an Rz-Rz1 complex, which forms in the presence of peptidoglycan (PG), is supported by in vivo results. The soluble periplasmic domains of Rz and Rz1, which are dimeric and monomeric respectively, were purified. Circular dichroism analysis indicates that Rz is structured, with significant α-helical content, whereas Rz1, in which 10 out 39 residues are proline, is unstructured. Mixing the proteins results in the formation of a complex with significant new α-helical content. Negative-stain images reveal ~ 25 nm x ~ 4 nm rod-shaped structures. Holin independent activity of Rz and Rz1 is found to disrupt whole cells. Furthermore, time lapse microscopy of λ and λRzam lysis allows us to conclude that Rz and Rz1 are essential for lysis. These results suggest a model for Rz-Rz1 function which begins with Rz and Rz1 forming a complex through direct interaction prior to holin and endolysin function. Holin-mediated hole formation allows the endolysin to degrade PG which sterically hinders Rz-Rz1 activity. Removal of PG by endolysin degradation thus triggers Rz-Rz1 OM disruption via fusion of the inner and outer membranes.