STRUCTURE DETERMINATION OF THE PHI-X174 CLOSED PROCAPSID

The structure of a procapsid of the single-stranded DNA bacteriophage phi X174 was determined to 3.5 Angstrom resolution. The crystal space group was I2(1)3 with a unit-cell length of 774 Angstrom. The unit cel l contained 16 icosahedral virus particles. each situated on a crystal lographic threefold axis. Thus. there are two independent one-thirds o f a particle per asymmetric unit, and a total of 40-fold non-crystallo graphic redundancy To aid in the interpretation of the packing arrange ment, crystals were prepared for thin sectioning and analyzed by elect ron microscopy. Oscillation X-ray diffraction data was collected on im age plates using synchrotron radiation and oscillation angles of eithe r 0.25 or 0.30 degrees. A low-resolution 6.5 Angstrom data set collect ed from a single frozen crystal was particularly helpful in the struct ure determination, because of its completeness and internal consistenc y. The initial particle orientations were determined using self-rotati on functions. while the initial position of one particle was determine d from a Patterson map. The structure was solved by molecular replacem ent real-space averaging using a model based on a cryo-electron micros copy reconstruction as a starting point for the phase determination. T he initial structure determination used the data between 20 and 13 Ang strom resolution, which was then extended one reciprocal lattice point at a time to 6.5 A resolution. At this paint. a 3.5 Angstrom resoluti on data set compiled from a number of crystals collected at 277 K was introduced. Phase extension and averaging continued to 3.5 Angstrom re solution after re-determining the particle positions and orientations. The amino acid sequences of most Of the D, F and G proteins and part of the B protein could be unambiguously built into the 3.5 Angstrom el ectron-density map. Partial crystallographic refinement yielded an R f actor of 31.6%, consistent with the relatively low resolution and lack of completeness of the data.