Dept. of Biological Sciences, Univ. of Pittsburgh, Pittsburgh PA 15260, USA
Chaperonins, including E. coli GroEL and other members of the Cpn60 family are a ubiquitous class of proteins that assist in protein folding in vivo by binding to partially folded proteins. This binding prevents aggregation of non-folded proteins and may unfold "off pathway" (misfolded) intermediates and direct them back into the correct folding pathway. Bacteriophage HK97 is a lambdoid virus that attacks E. coli. The major capsid protein in its icosahedral head (gp 5) is a natural substrate of GroEL. Here we describe properties of the complex and crystals of it that are suitable for structure determination by X-ray diffraction.
Cocrystals of a GroEL-HK97 gp5 complex have been obtained. The crystals can be "flash frozen" in liquid propane and remain stable in the X-ray beam (several hours at CHESS, up to eight weeks on our RU200). The space group is P2 and the unit cell parameters are a=138.0, b= 270.9, c=155.4 Å, and beta=101.4. A partial data set was obtained during a recent feasibility study at CHESS; meaningful data were recorded to at least 3.6Å resolution (the mean I/sigma ratio for all measured data in the 3.6Å shell was approximately 2 and some data were observed to 3.2Å). The Matthews coefficient is 3.39; the asymmetric unit contains one HK97-GroEL complex. This is the smallest asymmetric unit that could contain a complete substrate-GroEL complex. Preliminary self rotation functions calculated with X-PLOR show a clear peak for the 7-fold rotational symmetry expected for GroEL.
The crystallographic analyses complement in vitro studies of the interaction between GroEL and gp5 which suggest that binding of gp5 induces a change in the structure of GroEL. As with many proteins, unfolded gp5 makes a relatively stable complex with GroEL, and we have shown a stoichiometry of one gp5 subunit per GroEL 14-mer. This complex runs more slowly than free GroEL in non-denaturing electrophoresis, causing a band shift. Our data clearly indicate that this change in mobility is due primarily to a change in the charge of the complex upon binding gp5. Surprisingly, the experiments show that the band shift is not simply the result of the additive effect of the charges on the GroEL molecule and those on gp5: two "halves" of gp5, produced by tryptic cleavage after Lys167, each bind individually to GroEL but do not cause band shifts, and in any case the direction of the band shift is opposite to what would be predicted by a simple addition of charges. We have tested a large number of proteins, including duplication and deletion derivatives of gp5 and other unrelated proteins, and find that protein subunits smaller than about 31 kD do not cause a band shift upon binding to GroEL but that proteins larger than this do. Our current picture of this process is that when GroEL is presented with a protein bigger than about 31 kD, it undergoes a conformational adjustment to accommodate that protein. This adjustment is revealed experimentally as a change in the net charge of the complex, most likely due to a change in the number of bound counterions.