COMPOSITE PREDICTIONS OF SECONDARY STRUCTURES OF LAC REPRESSOR

The secondary structure of the lac repressor protein proposed by Chou et al. has been modified to include the recent revisions in sequence. In addition to the Chou and Fasman method, five other methods were used; they include those of (1) Lim, (2) Ptitsyn and Finkelstein, (3) Burgess et al., (4) Bunting et al., and (5) Wu and Kabat. Any two individual methods gave results differing sharply from one another. Three or more methods were in agreement for 91, 39, and 126 residues in helix, in β, and in combined coil plus turn conformations, respectively; there were such agreements for a total of 256 of the 360 residues. Agreements in the amino‐terminal third of the molecule were found for 68% of the residues, whereas in the remainder of the molecule only 53% of the residues showed such agreements. Only two helix‐breaking and two β‐breaking tripeptides were inconsistent with the composite predictions by three or more methods. The large number of disagreements among the results for different methods indicates that only very limited information is provided by each method and that the basis on which they operate is not clear. There is no a priori reason for a composite prediction to be more reliable than any individual prediction, and such a procedure does not permit the determination of an unambiguous secondary structure. Since these methods were applied to lac repressor before any three‐dimensional crystallographic structure was known, the methods may ultimately be evaluated should such a structure become available. Copyright © 1979 John Wiley & Sons, Inc.