Proteins labeled with 14C-amino acids after infection of Escherichia coli B by T4 phage were examined by electrophoresis in the presence of sodium dodecyl sulfate. Four regA mutants (regA1, regA8, regA11, and regA15) failed to make a protein having a molecular weight of about 12,000, whereas mutant regA9 did make such a protein; regA15 produced a new, apparently smaller protein that was presumably a nonsense fragment, whereas regA11 produced a new, apparently larger protein. We conclude that the 12,000-dalton protein was the product of the regA gene. The molecular weight assignment rested primarily on our finding that the regA protein had the same mobility as the T4 gene 33 protein, which we identified by electrophoresis of whole-cell extracts of E. coli B infected with a gene 33 mutant, amE1120. Synthesis of wild-type regA protein occurred from about 3 to 11 min after infection at 37 degrees C in the DNA+ state and extended to about 20 min in the DNA- state. However, synthesis of the altered regA proteins of regA9, regA11, and regA15 occurred at a higher rate and for a much longer period in both the DNA+ and DNA- states; thus, the regA gene is autogenously regulated. At 30 degrees C, both regA9 and regA11 exhibited partial regA function by eventually shutting off the synthesis of many T4 early proteins; the specificity of this shutoff differed between these two mutants. We also obtained evidence that the regA protein is not Stevens's "polypeptide 3." As a technical point, we found that, when quantitating acid-precipitable radioactivity in protein samples containing sodium dodecyl sulfate, it was necessary to use 15 to 20% trichloroacetic acid; use of 5% acid, e.g., resulted in loss of over half of the labeled protein.