Replacement of P1 of soybean mosaic virus with P1 of clover yellow vein virus has no impact on virus viability and host specificity.

Potyvirus genomes are expressed as polyproteins that are autocatalytically cleaved to produce 10 to 12 multifunctional proteins, among which P1 is the most variable. It has long been hypothesized that P1 plays role(s) in host adaptation and host specificity. We tested this hypothesis using two phylogenetically distinct potyviruses: soybean mosaic virus (SMV), with a narrow host range, and clover yellow vein virus (ClYVV), with a broader host range. When the full-length P1 cistron of SMV-N was replaced with P1 from ClYVV-No.30, the chimera systemically infected only SMV-N-permissive hosts. Hence, there were no changes in the host range or host specificity of the chimeric viruses. Despite sharing only 20.3% amino acid sequence identity, predicted molecular models of P1 proteins from SMV-N and ClYVV-No.30 showed analogous topologies. These observations suggest that P1 of ClYVV-No.30 can functionally replace P1 of SMV-N. However, the P1 proteins of these two potyviruses are not determinants of host specificity and host range.

Characterizing protein conformations by correlation analysis of coarse-grained contact matrices.

We have developed a method to capture the essential conformational dynamics of folded biopolymers using statistical analysis of coarse-grained segment-segment contacts. Previously, the residue-residue contact analysis of simulation trajectories was successfully applied to the detection of conformational switching motions in biomolecular complexes. However, the application to large protein systems (larger than 1000 amino acid residues) is challenging using the description of residue contacts. Also, the residue-based method cannot be used to compare proteins with different sequences. To expand the scope of the method, we have tested several coarse-graining schemes that group a collection of consecutive residues into a segment. The definition of these segments may be derived from structural and sequence information, while the interaction strength of the coarse-grained segment-segment contacts is a function of the residue-residue contacts. We then perform covariance calculations on these coarse-grained contact matrices. We monitored how well the principal components of the contact matrices is preserved using various rendering functions. The new method was demonstrated to assist the reduction of the degrees of freedom for describing the conformation space, and it potentially allows for the analysis of a system that is approximately tenfold larger compared with the corresponding residue contact-based method. This method can also render a family of similar proteins into the same conformational space, and thus can be used to compare the structures of proteins with different sequences.