Membrane topography of the hydrophobic anchor sequence of poliovirus 3A and 3AB proteins and the functional effect of 3A/3AB membrane association upon RNA replication

Abstract

Replication of poliovirus RNA takes place on the cytoplasmic surface of membranous vesicles that form after infection of the host cell. It is generally accepted that RNA polymerase 3D(pol) interacts with membranes in a complex with viral protein 3AB, which binds to membranes by means of a hydrophobic anchor sequence that is located near the C-terminus of the 3A domain. In this study, we used fluorescence and fluorescence quenching methods to define the topography of the anchor sequence in the context of 3A and 3AB proteins inserted in model membranes. Mutants with a single tryptophan near the center of the anchor sequence but lacking Trp elsewhere in 3A/3AB were constructed which, after the emergence of suppressor mutations, replicated well in HeLa cells. When a peptide containing the mutant anchor sequence was incorporated in model membrane vesicles, measurements of Trp depth within the lipid bilayer indicated formation of a transmembrane topography. However, rather than the 22-residue length predicted from hydrophobicity considerations, the transmembrane segment had an effective length of 16 residues, such that Gln64 likely formed the N-terminal boundary. Analogous experiments using full-length proteins bound to preformed model membrane vesicles showed that the anchor sequence formed a mixture of transmembrane and nontransmembrane topographies in the 3A protein but adopted only the nontransmembrane configuration in the context of 3AB protein. Studies of the function of 3A/3AB inserted into model membrane vesicles showed that membrane-bound 3AB is highly efficient in stimulating the activity of 3D(pol) in vitro while membrane-bound 3A totally lacks this activity. Moreover, in vitro uridylylation reactions showed that membrane-bound 3AB is not a substrate for 3D(pol), but free VPg released by cleavage of 3AB with proteinase 3CD(pro) could be uridylylated.

Figure 1

Genomic organization of poliovirus, processing of the polyprotein and amino acid sequences of protein 3AB. (A) The single-stranded RNA genome is covalently linked to the viral-encoded protein VPg at the 5′ end of the non-translated region (5′NTR). The linkage of the 5′ terminal UMP to the hydroxyl group of tyrosine in VPg is shown on the left. The 5′NTR consists of two cis-acting domains. The cloverleaf is involved in genome replication, and the internal ribosomal entry site (IRES), controls translation of a 247 kDa polyprotein (open box). The polyprotein is processed by the virus encoded proteinases 2A^pro and 3C^pro/3CD^pro into structural proteins (P1) and nonstructural (P2 and P3), the latter specifying the proteins involved in replication. Triangles indicate cleavages by 3C^pro/3CD^pro and circles cleavage by 2A^pro. Filled symbols indicate fast cleavages and open symbols indicate slow cleavages. The open diamond marks the VP0 capsid maturation cleavage. The 3′NTR contains a structured heteropolymeric region and is polyadenylylated. The open arrow indicates one of two suppressor mutations resulting from the W42F/F69W mutations in 3AB (see Fig. 1B). (B). Presentation of 3AB and its cleavage product 3A. Shown is the amino acid sequence of 3AB with the Q/G cleavage site between 3A and 3B(VPg). The hydrophobic anchor domain of 3A, starting at amino acid 59, is indicated in a box, which is enlarged below. The shaded area represents the TM segment as defined in this report. Downward or upward arrows indicate two separate sets of mutations. Filled arrows depict mutations engineered into the protein, open arrows the suppressor mutations that rapidly emerged during the first cycle of virus replication. Anchor peptide-1 and anchor peptide-2 delineate the peptides used for membrane binding studies.

Figure 2

One step growth curves of wt poliovirus and a 3A mutant (W42F/F69W/A70V). Growth curves and plaque assays were carried out as described in Experimental Procedures. Upper panel: growth at 37 °C. Lower panel: growth at 39 °C. Note that this 3A mutant also has a suppressor mutation (I47V) within the first hydrophobic domain of the 2B-coding sequence (Fig. 1A). The Y-axis shows pfu/ml in exponential units (base 10).

Figure 3

Effect of lipid bilayer width upon Trp emission λmax for anchor peptide-1 and anchor peptide-2. Anchor peptide-1 (filled triangles) or anchor peptide-2 (open triangles) were reconstituted into model membrane vesicles composed of monounsaturated PCs with different acyl-chain lengths. Samples contained 2 μM peptide incorporated into 500 μM lipid dispersed in PBS at pH 7.0. The values shown are the average of six samples. The λmax values are generally reproducible to ±1 nm.

Figure 4

One step growth curves of wt poliovirus and two 3A mutants (T14K/W42F/V72W and P17K/W42F/V72W). Growth curves and plaque assays were carried out as described in Experimental Procedures. Upper panel: growth at 37 °C. Lower panel: growth at 39 °C.

Figure 5

Stimulation of the elongation activity of 3D^pol by 3A- or 3AB-proteoliposomes. Primer-dependent elongation reactions were performed in the presence of 3A-(filled triangle) or 3AB (open triangles)-proteoliposomes, and ³²[P]UMP incorporated into polymer was measured by a scintillation counter. Values were corrected for using the same amount of empty vesicles.

Figure 6

Uridylylation reactions using 3AB-proteoliposomes. A. VPgpUpU synthesis as a function of 3AB and 3CD^pro concentrations. Uridylylation reactions were carried out as described in Experimental Procedures on a cre(2C) template RNAs using proteoliposomes containing 3 ng, 30 ng, 300 ng or 3 μg of 3AB, as indicated on the figure. At each 3AB concentration proteolytically active 3CD^pro (cleavage-site mutant to prevent cleavage between 3C^pro and 3D^pol) (lanes 4–18) were included in the reactions, at concentrations 0.35 μg, 0.7 μg, 1.4 μg or 2.8 μg, from left to right in each set of experiments. The reaction shown on lane 3 contained 3AB (3 μg) and inactive proteinase (0.7 μg). As a control the uridylylation of synthetic VPg was measured in the presence of 0.7 μg of inactive 3CD^pro (lanes 2 and 20). B. VPgpUpU synthesis requires proteolytically active 3CD^pro. Reactions were carried out using 3 μg 3AB in proteoliposomes alone (lane 1) or with different concentrations of proteolytically active or inactive 3CD^pro for 1h (0.35 μg, 0.7 μg, 1.4 μg or 2.8 μg per reaction from left to right (lanes 2–9)). As a control synthetic VPg peptide (2 μg) was used as substrate in the uridylylation reaction with 0.7 μg of active 3CD^pro (lane 14). Uridylylation reactions were also carried out with pre-incubation of 3AB-liposomes with active 3CD^pro for 2 hr (lanes 10–13), or using purified 3AB in detergent (0.8% octylglucoside), (lanes 15–18), again with different concentrations of active 3CD^pro (0.35 μg, 0.7 μg, 1.4 μg or 2.8 μg per reaction, from left to right). Note the loss of activity at very high 3CD^pro concentrations, the reason for this is unclear, but we have also observed this in other experiments.

Figure 7

Proposed model for the structures of 3AB and 3A on membranes and initiation of poliovirus replication. The P3 precursor is processed by proteinase 3CD^pro to yield 3AB and 3CD^pro. 3AB interacts with membranes forming a non-TM configuration and its VPg domain binds 3D^pol to the membranes. Protein 3CD^pro cleaves off VPg releasing the 3A protein, which then assumes both a TM and a non-TM form. 3D^pol, complexed with VPg, then interacts with the 3C^pro domain of 3CD^pro. Through its RNA binding domain 3CD^pro binds the entire complex to the cre(2C) RNA structure where the uridylylation of VPg takes place yielding VPgpUpU. The 3A(B) interaction with 2B(C) is not indicated on the model because we do not yet know whether this interaction is required during the initiation of RNA replication.