High-Resolution Structure of a Virally Encoded DNA-Translocating Conduit and the Mechanism of DNA Penetration

Abstract

Although ϕX174 DNA pilot protein H is monomeric during procapsid assembly, it forms an oligomeric tube on the host cell surface. Reminiscent of a double-stranded DNA phage tail in form and function, the H tube transports the single-stranded ϕX174 genome across the Escherichia coli cell wall. The 2.4-Å resolution H-tube crystal structure suggests functional and energetic mechanisms that may be common features of DNA transport through virally encoded conduits.

INTRODUCTION

Upon host cell recognition, bacteriophages must propel their genomes across a large barrier, the cell wall. For double-stranded DNA (dsDNA) phages, emblematic tail structures often perform this function. Myovirus T4 employs a contractile sheath surrounding a tail tube. Contraction ultimately leads the tube spanning both of the Escherichia coli membranes and the intervening periplasmic space (1). Podovirus T7 lengthens its pre-existing short tail tube until it spans this barrier (2). These tubes become the genome's conduit into the cytoplasm. In contrast, siphoviruses (e.g., phage λ), filamentous, and tailless icosahedral bacteriophages co-opt host cell channels for genome transport (3–5). Until recently, ϕX174 DNA delivery could not be interpreted within these well-defined paradigms (6).

The strictly icosahedral ϕX174 virion does not use host cell channels to transport its genome (7). Thus, the penetration mechanism has been mysterious since the first Parisian plaque formed during the 1920s, Les Années Folles. In the more sobering 1970s, Jazwinski and colleagues recognized the critical involvement of protein H, a minor capsid protein, calling it the DNA pilot protein (8). They speculated that “protein H facilitates the transit of nucleic acid, a polyelectrolyte, through the membrane lipid bilayers, by the creation of a pore.” Thirty-five years later, their hypothesis was shown to be exceptionally insightful (6).

The structure of the ϕX174 H tube, a virally encoded DNA conduit.

The atomic structure of the DNA pilot protein's central domain was determined to 2.4-Å resolution (Fig. 1A). Ten α-helical monomers form a 170-Å-long, coiled coil, helical barrel long enough to span the periplasmic space or a membrane adhesion site (9). The 22-Å internal diameter can accommodate two un-base-paired single-stranded DNA (ssDNA) strands (10, 11). The helices run in parallel but kink at residues Y193, A194, and Q195. This divides the tube into two domains. Domain A is an 11/3 coiled-coil structure (11 residues per 3 helical turns), whereas domain B is a 7/2 coiled-coil structure. To the best of our knowledge, this is the first report of a viral DNA translocation tube determined at atomic resolution.

FIG 1.

Structure of the ϕX174 protein H DNA-translocating conduit (H tube), ϕX174 assembly, and DNA delivery. (A) The X-ray structure of the H tube, which contains 10 monomers. (B) The ϕX174 assembly pathway. Protein H is monomeric through the assembly of the 12S* particle. Protein B facilitates protein H incorporation. As the protein H structure during this stage of assembly is unknown, it is depicted as a red spot. The release of protein B as the procapsid is filled with DNA may allow H-H protein interactions within the capsid. However, this is not known. (C) The inward-facing and angled amide and guanidinium side chains that line the length of the tube suggest a one-way passage. (D) Conservation of structure, the inward-facing and angled teeth of the giant sea lamprey. (E) Coupling of DNA synthesis and genome delivery. If host cell DNA synthesis is inhibited, the entire genome is not transported into the cell. Restoration of the cell's DNA-synthesizing capabilities leads to full genome transport. After DNA delivery, protein H is associated with the inner membrane (IM). The structure of protein H and its oligomeric state at this stage of infection are unknown; therefore, it is depicted as a red spot. LPS, lipopolysaccharide; OM, outer membrane.

Demonstrating biological significance.

Although the H tube's length and internal diameter were consistent with genome transport across a Gram-negative cell wall, these properties appeared to be inconsistent with three previous observations. (i) Neither an external tail-like feature nor an internal core had been observed. (ii) The fully assembled tube was likely too long to fit into the capsid (12). (iii) During procapsid assembly (Fig. 1B), protein H monomers are incorporated into pentameric assembly intermediates (13). Thus, tube formation occurs after procapsid assembly. These hypotheses, after procapsid assembly tube formation and subsequent tube function in DNA transport through the cell wall, were tested biochemically in vitro, genetically in vivo, and tomographically in situ.

In vitro testing.

Amino acid substitutions that inhibited tube-forming helix-helix interactions were introduced into the protein. Circular-dichroism spectra indicated that these substitutions did not alter the α-helical fold. However, by size exclusion chromatography, the mutant proteins migrated as monomers vis-à-vis the oligomeric wild-type control. Moreover, trypsin sites, normally buried in the wild-type structure, were accessible in the mutant proteins. These data indicated that the mutant proteins cannot form tubes.

In vivo testing.

As particle assembly involves H protein monomers, inhibition of tube formation should not affect in vivo morphogenesis. In both protein and DNA contents, virus-like particles assembled with the mutant H proteins were indistinguishable from the wild type. However, the resulting particles lacked infectivity, as measured in standard plaque assays.

In situ testing.

Cryoelectron microscopic tomography with minicells was used to visualize the tube performing its hypothesized function. Upon cell contact, fully formed tubes emerged from capsids. After DNA delivery, the conduits appeared to dissociate within the cell wall, which most likely happens to preserve the cell's membrane potential. Thus, the presence of this newly discovered phage tail is ephemeral.

H-tube formation—temporal questions.

Twelve pentameric intermediates, each of which contains only one copy of protein H, are organized into the procapsid by the external scaffolding protein (Fig. 1B). Although mechanistic details remain obscure, internal scaffolding protein B facilitates protein H incorporation, perhaps keeping it monomeric during assembly (14). As protein B is found in the procapsid, it may prevent tube formation. During genome packaging, DNA binding protein J displaces protein B (12). This may allow H monomers to associate before the entire genome volumetrically constrains further protein H movement. Indeed, the noise created by packaged DNA was too large to allow protein H detection in virions (6). Thus, to visualize partially formed tubes within capsids, both DNA and protein B need to be eliminated. Fortuitously, a recently characterized internal scaffolding protein B mutant produces B-less, DNA-less procapsids (15).

Evolving a one-way conduit.

The tube's inner surface is lined with amino acids containing amide and guanidinium side chains (Fig. 1C). The primary amino acid is glutamine, but asparagine and arginine residues are also represented. The long side chains are oriented toward the tube's center and point toward the N termini of the constituent polypeptides. Analogous to the mouth of a giant see lamprey (Fig. 1D), with inward-pointing teeth angled to funnel prey down its throat, the orientation of the internal H-tube side chains suggests that the DNA is transported from the C to the N terminus and cannot move backward.

Ten or twelve.

During assembly, 12 pentameric intermediates (12S* particles) assemble into procapsids. As each 12S* particle contains one protein H molecule, virions likely contain 12 protein H molecules. Yet, the crystal structure is a 10-mer. This implies that two protein H molecules are excluded from the tube or functional tubes can be formed with various numbers of subunits. The crystallized H tube was assembled in vitro, not in the context of a DNA-crowded, volumetrically constrained capsid. Thus, the number of monomers required for in vivo tube formation and function is uncertain. However, virally encoded dsDNA-translocating vertices in tailed phages exhibit 12-fold symmetry, although, like the H tube, filamentous phage capsids, which also transport two antiparallel strands of ssDNA, exhibit 10-fold symmetry (11).

Energy.

At its inception, molecular biology aspired to explain biological processes within the context of physical laws. Phage systems have contributed significantly to this endeavor. Although it is counterintuitive, mature virions, with the stability to withstand hostile extracellular environments, represent the particle's highest energy state (1, 16). Energy enters the system during DNA packaging, which requires ATP hydrolysis. This potential energy is related to the internal capsid pressure generated by a highly compacted, volumetrically constrained genome (16). For dsDNA phages, energy propels the genome through tail structures that act as DNA-translocating conduits (1). In this regard, ϕX174 differs only in the timing of tail assembly. The tail tube extensions observed in T7-like viruses and the protein-rich membrane tubes formed by the PRD1-like phages represent other paradigm variations (2, 17).

The H-tube structure demonstrates that the entire inner passage is lined with amide and guanidinium side chains, which are known to interact with purines (18). Although other high-resolution DNA conduit structures have yet to be determined, their sequences contain many glutamine, arginine, and asparagine residues. There is enough homology between the ϕX174 H protein and T7 gp16, a likely component of the tail tube extension in T7, to generate a reasonable alignment in which the positioning of the glutamine residues is extremely similar (6).

Although the energetic contribution of the DNA-interacting side chains is still unknown, reasonable speculation is warranted. As exemplified by the catabolic oxidation of carbon and the electron transport systems, potential energy performs work more efficiently if released in small, usable increments. The amide side chains may act like the air gauges that regulate compressed-gas-driven machines. These gauges prevent energy from overwhelming the system. In this analogy, the compressed gas represents the volumetrically constrained genome and its potential energy, whereas the gas cylinder represents the capsid. The H tube is the conduit through which the energy is released. The amide side chains perform the function of a regulatory gauge. If too much air pressure (energy) is released at one time, it can rupture the conduit or blow the gasket that connects the tubing to the cylinder. Vis-à-vis the phage, the gasket constitutes the H-tube–capsid interactions. By interacting with the DNA, the amide side chains modulate or slow down transport, releasing potential energy in small increments. Thus, the amide side chains provide a frictional force countering the one driving the genome into the cell. This seems counterintuitive, resulting in genomes that do not completely transit the tube. However, the results of previous experiments (19) suggest that the first stage of viral DNA replication, ssDNA→dsDNA, may be linked to genome transport, suggesting that energy released during DNA replication could provide an additional energy source (see below).

Immediately upon infection, host cell DNA replication enzymes use the positive-ssDNA genome as a template for negative-strand DNA synthesis, converting the ssDNA genome into a dsDNA molecule (Fig. 1E). The entire genome need not be internalized to begin this process. Mano and colleagues used temperature-sensitive host DNA replication proteins to inhibit penetration. The process begins at the restrictive temperature, but the entire genome is not transported (19). Upon shifting infections to the permissive temperature, which restores the host cell's DNA synthesis capability, the rest of the viral genome could enter the cell. Thus, the potential energy stored within the packaged capsid is likely necessary but not sufficient to fully transport the genome into the cell.

The glutamine side chains on the inside of the H tubes may have a unique role in transporting the ssDNA through a narrow passage. The ϕX174 genome has some secondary structure (20). While the tube's inner diameter can accommodate two antiparallel strands of ssDNA (10), it is far too narrow to allow the passage of dsDNA and hairpin loops. Prior to entering the tube, H bonds between paired purines and pyrimidines would have to be broken, which would require energy. However, this barrier could be lowered if the broken purine-to-pyrimidine H bonds were replaced with hydrogen bonds between the purines and amide side chains. These same interactions may suppress secondary structure that could clog the tubing during transport, another gauge-like function.

Although we answered the 9-decade-old “tailless” DNA delivery riddle, the solution poses additional puzzles. When and how does the tube form (see H-tube formation—temporal questions)? After DNA delivery, the tube dissipates but protein H remains associated with the inner membrane. Thus, the undefined connection with the capsid is broken. Perhaps the C and N termini, which are not contained in the structure, mediate these functions. We endeavor to solve these puzzles in less time than it took to answer the original riddle.