How do viral packaging motors drive double-stranded DNA into the capsid?

    Packaging of dsDNA is opposed by large forces, arising from the DNA-DNA electrostatic repulsions, elastic deformations, and the large loss of DNA conformational entropy upon confinement (although this is offset by the favorable entropy change of solvent release). As a consequence, viral packaging motors are among the strongest of all known biological motors.

   It is generally accepted that the chemical cycle (ATP binding, hydrolysis and product release) drives a cycle of conformational changes in the motor proteins. It is widely supposed that this cycle moves the DNA forward into the capsid using some sort of lever-like protein motions, coupled to a protein-DNA grip-and-release cycle. In this conventional view, DNA is a passive substrate, moved through the motor just as a rope moves through the hands of someone pulling a bucket up out of a well, or through the hands of an athlete climbing a rope.



   I proposed that DNA is not simply a passive substrate, driven into the capsid by lever-like protein motions. Instead, it is an active component of the force-generating mechanism.

   In its original form, this scrunchworm hypothesis suggested that the proteins alternatively dehydrate and rehydrate the DNA, driving it through a cycle of transitions between the A-DNA and B-DNA conformations. A-DNA is 23% shorter than B-DNA, and the model proposed that the transitions between these two forms are coupled to a protein-DNA grip-release cycle, rectifying the DNA motion and driving it forward into the capsid.

   The scrunchworm model offered an explanation for the large forces generated during DNA packaging (~50 pN), and it made a series of experimentally testable predictions.

     We carried out a series of computer simulations to test the scrunchworm model. In the first of these, we found that the phi29 connector induces DNA scrunching (red in panel "a"). Panel "b" shows A-DNA in blue at the left, and the scrunched conformations of four different DNA sequences in the phi29 portal; these are even shorter (more "scrunched") than A-DNA. While the results support the basic idea of the scrunchworm model, they suggest that the mechanism involves more than B- to A-DNA transitions (publication),

   Subsequent simulations on the portal from phage T4 led to a surprise: in contrast with the scrunched DNA conformation seen in phi29, DNA in the T4 portal is stretched. And when we extended the simulations to phage P22, the situation got even murkier: when the P22 portal is in the "PC" conformation, the DNA stretches, but the DNA is essentially B-form in when the P22 portal has the "MV" conformation. The figure below summarizes these results.

   The differences between the DNA conformation in the P22-PC and P22-MV portal conformations is important, because they demonstrate that changes in the conformation of the portal protein (presumably driven by the biochemical cycle) can drive DNA conformational changes, as required by the scrunchworm model. But these results disprove the original proposal involving transitions between the A- and B-forms of DNA.

   They also raise the question of exactly how the protein conformation affects the DNA conformation.

   This question was answered when my colleague Kim Sharp examined the complexes in the figure above using Delphi, a computer program for determining the electrostatic potential and interaction energies in biomolecular complexes (figure below). The DNA phosphate groups are attracted to regions of high positive electrostatic potential (blue), leading to DNA scrunching in the phi29 portal. And they are repelled from regions of high negative potential (red), producing the stretched DNA conformation seen in the T4 and P22-PC portals.

   This suggest an outline for how DNA translocation can be driven by coupling cyclic changes in the electrostatic potential with cyclic motions of two protein-DNA grips. The figure below illustrates the case with DNA stretching.

   The motor proteins are shown schematically in green. (A) The cycle begins with B-DNA (orange) in the channel of the complex of motor proteins. The lower protein-DNA grip is closed at the bottom of the channel, and the upper grip is open (pairs of green triangles). (B) A conformational change in the proteins brings a negatively charged protein domain (pink) close to the DNA, driving the nearby DNA into an extended form (red).  Since the tail of the DNA is held fixed, the upper end of the DNA molecule is pushed forward into the capsid (arrow). (C) The upper grip closes to capture the DNA's advance. (D) The lower grip opens. (E) A second conformational change in the protein pulls the negatively charged domain away from the DNA. The DNA returns to the B-form. Since the upper part of the DNA is gripped by the protein, the return to the B-form pulls the tail of the DNA upward. (arrow). (F) The lower grip closes to prevent DNA backsliding. (G) The upper grip opens, returning the protein to its original conformation. The DNA has advanced upward from its original position, which is marked by the dashed orange lines.

   This work has been published in Biophysical Journal (PMID 31103227).

   Whether further experiments, simulations and theoretical analyses support the scrunchworm model, disprove it, or require that it be modified, they will provide information that will contribute to the ultimate understanding of how these motors work.