The black pigment of Porphyromonas gingivalis is composed of the mu-oxo bishaem complex of Fe(Ill) protoporphyrin IX (mu-oxo oligomer, dimeric haem), namely [Fe(III)PPIX](2)O.P. gingivalis W50 and Rgp (Arg-gingipain)- and Kgp (Lys-gingipain)deficient mutants K1A, D7, E8 and W501 [Aduse-Opoku, Davies, Gallagher, Hashim, Evans, Rangarajan, Slaney and Curtis (2000) Microbiology 146, 1933-1940] were grown on horse blood/agar for 14 days and examined for the production of mu-oxo bishaem. mu-oxo Bishaem was detected by UV-visible, Mossbauer and Raman spectroscopies in wild-type W50 and in the blackpigmented RgpA- and RgpB-deficient mutants (W501 and D7 respectively), whereas no haem species were detected in the straw-coloured colonies of Kgp-deficient strain K1A. The dark brown pigment of the double RgpA/RgpB knockout mutant (E8) was not composed of mu-oxo bishaem, but of a high-spin monomeric Fe(III) protoporphyrin IX species (possibly a haem-albumin complex). In vitro incubation of oxyhaemoglobin with cells of the W50 strain and the RgpA- and RgpB-deficient mutants (W501 and D7) resulted in the formation of mu-oxo bishaem via methaemoglobin as an intermediate. Although the Kgp-deficient strain K I A converted oxyhaemoglobin into methaemoglobin, this was not further degraded into mu-oxo bishaem. The double RgpA/RgpB knockout was also not capable of producing mu-oxo bishaem from oxyhaemoglobin, but instead generated a haemoglobin haemichrome. Inhibition of Arg-X protease activity of W50, W501, D7 and K1A with leupeptin, under conditions where Lys-X protease activity was unaffected, prevented the production of mu-oxo bishaem from oxyhaemoglobin, but resulted in the formation of a haemoglobin haemichrome. These results show that one or both of RgpA and RgpB gingipains, in addition to the lysine-specific gingipain, is necessary for the production of mu-oxo bishaem from haemoglobin by whole cells of P. gingivalis.