Nonribosomal peptides (NRP) such as the antibiotic tyrocidine have D-amino acids, introduced by epimerase (E) domains embedded within modules of the enzymatic assembly lines. We predict that the peptide bond-forming condensation (C) domains immediately downstream of E domains are D-specific for the peptidyl donor and L-specific for the aminoacyl acceptor ((D)C(L)). To validate this prediction and establish that the C(5) domain of tyrocidine synthetase is indeed (D)C(L), the apoT (thiolation) forms of module 4 (TycB(3) AT(4)E) and module 5 (TycC(1) C(5)AT(5)) were expressed. T(5) was posttranslationally primed with CoASH to introduce the HS-pantetheinyl group and autoaminoacylated with radiolabeled L-Asn* or L-Asp*. Alternate donor substrates were introduced by priming apo AT(4)E with synthetically prepared tetrapeptidyl-CoA's differing in the chirality of Phe-4, D-Phe-L-Pro-L-Phe-L-Phe-CoA, and D-Phe-L-Pro-L-Phe-D-Phe-CoA. The tetrapeptidyl-S-T(4) and L-Asp-S-T(5) were studied for peptide bond formation and chain translocation by C(5) to yield pentapeptidyl-S-T(5), whose chirality (D-L-L-D-L- vs D-L-L-L-L-) was assayed by thioester cleavage and chiral chromatography of the released pentapeptides. Only the D-Phe-4 pentapeptidyl-S-T(5) was generated, implying that only D-L-L-D-S-T(4) was utilized, proving C(5) is indeed a (D)C(L) catalyst. Furthermore, a mutant with an inactive E domain transferred tetrapeptide only when loaded with D-Phe-4 tetrapeptidyl donor, not L-Phe-4, confirming that in the wild-type assembly line C(5) only transfers D-L-L-L-tetrapeptidyl-S-T(4) after in situ epimerization by the E domain. These results contrast the observation that C(5) can make both L-Phe-L-Asn and D-Phe-L-Asn when assayed with Phe as the donor substrate. Hence, utilizing an aminoacyl-S-T(4) versus the natural peptidyl-S-T(4) donor produced misleading information regarding the specificity of the condensation domain.