Defining molecular and domain boundaries in the bacteriophage phi29 DNA packaging motor.

Production of ϕ29 Particles Fibered proheads with 174-base pRNA were produced in B. subtilis SpoOA12( sup − ) cells infected with the mutant sus 16(300)- sus 14(1241) that is defective for the DNA packaging ATPase gp16 as described previously ( Morais et al., 2005 ). These proheads were treated with RNAse A (1mg/ml) to digest pRNA, purified by sucrose density gradient centrifugation, reconstituted with in vitro transcribed 71- or 120-base pRNA, and repurified on a sucrose density gradient. Proheads with 174-base pRNA were incubated with gp16 to produce the prohead-ATPase complex. Proheads without connectors were observed in RNaseA-treated proheads which were reconstituted with in vitro transcribed 120-base pRNA and stored at −70°C for 3 months.

Electron Microscopy

Prohead particles were flash-frozen on holey grids in liquid ethane. Images were recorded at 39,000 × magnification with a CM200 FEG microscope, with electron dose levels of approximately 20 e − /Å 2 . All micrographs were digitized at 3.59 Å pixel −1 using either a Zeiss SCAI scanner or a Nikon Super CoolScan 9000 scanner. For all reconstructions, individual particle images were boxed, floated, and preprocessed to normalize mean intensities and variances and to remove linear background gradients. For each data set, reference projections of a previously published prohead reconstruction ( Morais et al., 2005 ) were used to initially classify particles for three-dimensional reconstruction. The resulting model was used to recalculate reference projections for better particle classification. Several cycles of iterative particle classification and reconstruction were performed until convergence had been reached. Structure factor phases were modified as indicated by the parameters of the contrast transfer function. All steps of the reconstruction process, including determination of the contrast transfer function parameters, were performed with the program EMAN ( Ludtke et al., 1999 ). Five-fold symmetry was assumed in all the reconstructions. The number of particles incorporated into each reconstruction, as well as the final resolution of each reconstruction, is summarized in Table 1 . The resolutions of these reconstructions were determined by the Fourier shell correlation method using a correlation coefficient of 0.5 between independent half data sets as the cut-off criterion. Difference maps between the reconstructions were calculated after adjusting radial and density scale factors using RobEM ( http://cryoem.ucsd.edu/programs.shtm ). In order to eliminate artifacts that might arise as a result of calculating difference maps between reconstructions of differing resolutions, the same low-pass filter with a Gaussian fall-off was applied to each of the two reconstructions in a given difference map prior to scaling and subsequent map subtraction. The cut-off value of this low-pass filter was chosen to match the poorer resolution of the two maps being compared.