The trifunctional trp-1 gene from Neurospora crassa was cloned by complementation of a phosphoribosylanthranilate isomerase-deficient mutant of E. coli. A 2.7-kb DNA sequence containing trp-1 was determined. Homology of the deduced trp-1 polypeptide sequence to the corresponding E. coli proteins is striking; the order of functional domains within trp-1 is NH2-glutamine amidotransferase-indoleglycerolphosphate synthase-phosphoribosylanthranilate isomerase-COOH (NH2-trpG-trpC-trpF-COOH). Whereas trpF complementing activity can be detected in E. coli, trpC activity is absent. It is likely that translation of trp-1 does not proceed from the proper start site in E. coli; the carboxy terminal portion of the trp-1 polypeptide may be the only portion synthesized. Fusion of a bacterial amino terminus and ribosome binding site to the trp-1 coding region results in expression of trpC as well as trpF activity in E. coli. The locations of several startpoints for trp-1 mRNA synthesis were determined by the S1 nuclease mapping technique. DNA immediately 5' to the trp-1 transcription initiation region does not possess a sequence resembling the canonical TATAAA of eukaryotes.