Bacteriophages (phages) are viruses that infect bacteria. The phages that are described by this dissertation encompass 2 classes, double-stranded DNA phages and single-stranded RNA phages. While both of these phages infect similar bacteria, they have adopted different mechanisms to lyse, or destroy, the cell in order to release phage progeny. dsDNA phages have large genomes (> 20 kb) and use multiple lysis proteins (holin, endolysin, and spanin complex) to lyse the cell. ssRNA phages, on the other hand, have small genomes (< 6 kb) and only encode one lysis protein. The two X-ray crystallography projects outlined here deal with the phage proteins involved in these lysis mechanisms. The project described in the first study deals with the holin (T) and the antiholin (RI) of the ds-DNA phage T4, the major players of the lysis inhibition (LIN) phenomenon. Crystal structures of the holin and of the holin-antiholin complex are presented. The structures provide new molecular level insights into the phenomenon of LIN in bacteriophage T4 and the T-even phages in general. The second investigation describes ongoing efforts at structural characterization of A2, the maturation protein of the ssRNA bacteriophage Qbeta that inhibits E. coli MurA. In addition, the structure of Bacillus subtilis MurA, which is not recognized by A2, is presented. The crystal structure of B. subtilis MurA, the first structure of MurA from a Gram-positive organism, allows for a direct comparison of Gram-positive and Gram-negative homologs and for identification of any significant structural differences. The more flexible catalytic loop of B. subtilis MurA protrudes farther out compared to the loop of E. coli MurA and creates enough hindrance to prevent A2 from establishing secure contact points.

The N-terminal hydrophobic domain of the phage P1 endolysin Lyz was found to facilitate the export of Lyz in a sec-dependent fashion, explaining the ability of Lyz to cause lysis of E.coli in the absence of the P1 holin. The N-terminal domain of Lyz is demonstrated to be both necessary and sufficient not only for export to the membrane but also for release into the periplasm of this endolysin. We propose that this unusual N-terminal domain functions as a "signal arrest- release" (SAR) sequence, which first directs the endolysin to the periplasm in membrane-tethered form and then allows it to be released as a soluble active enzyme in the periplasm. To understand why release from the membrane is required for the physiological expression of the lytic activity of Lyz, we examined the role of its seven cysteine residues in the biogenesis of the active endolysin. The inactive, membrane-tethered and the active, soluble forms of Lyz differ in their pattern of intramolecular disulfide bonding. We conclude that the release of Lyz from the membrane leads to an intramolecular thiol-disulfide bond isomerization causing a dramatic conformational change in the Lyz protein. As a result, an active site cleft that is missing in nascent Lyz is generated in the mature form of the endolysin. Examination of the protein sequences of related bacteriophage endolysins suggests that the presence of an SAR sequence is not unique to Lyz. Studies on holin and antiholin indicated that P1 encodes two holins, LydA and LydC. The antiholin LydB inhibits LydA by binding to it directly on the membrane. All above results demonstrate a new paradigm for control of phage lysis, which is, upon depolarization of the membrane by holin function at a programmed time, endolysin is released from the bilayer leading to the immediate lysis of the host.