The objective of this study was to examine the potential of vaginal and rectal mucosal routes for feline immunodeficiency virus (FIV) uptake and infection, as a model of mucosal HIV infection, and to determine the fate of virus at these mucosal sites following transmission of infection. SPF cats were exposed to FIV isolates (PET, GL-8, T637), administered as either cell-associated or cell-free inocula, via the rectum or vagina. Establishment of infection was confirmed by isolation of infectious FIV from peripheral blood mononuclear cells (PBMC), and by presence of FIV proviral DNA in PBMC using a nested polymerase chain reaction. Fate of virus in tissue taken at necropsy from cats infected for 6-48 weeks was assessed by localizing FIV core and envelope proteins, p24 and gp41, using a biotin-streptavidin linked immunoperoxidase (IP) technique. Cells susceptible to infection were identified by an in situ hybridization technique for FIV viral DNA and RNA. Cell-free, as well as cell-associated, virus was infectious across intact vaginal and rectal mucosal surfaces. Transmission was most successful using cell-associated inocula, and via the rectal route. Cells infected with FIV were detected by IP staining in the colon of 6/9 rectally challenged cats and 1/5 vaginally challenged cats. Virus was predominantly localized within the epithelium at the base of the colonic crypts associated with lymphoid aggregates (follicle associated epithelium; FAE), and within the lymphoid follicle itself. Occasional infected cells were also noted within the lamina propria. The distribution of FIV DNA positive cells in the colon was similar to that for FIV antigen whilst FIV RNA positive cells were found more extensively, including within the lamina propria and lymphoid follicle. FIV infected cells were not detected within the vagina, or colonic and ileac lymph nodes. Similar patterns of infected cells were seen in all of the positive cats, indicating that colonic tissues remain persistently actively infected with FIV. We conclude that the FIV/cat model of rectal and vaginal mucosal infection should prove useful for characterizing the mechanism by which HIV infects mucosal surfaces and as a challenge system for the design of vaccines effective at preventing HIV infection via rectal and vaginal routes.