Pc is an often fatal opportunistic pathogen. The study of Pc has been hampered by our inability to grow this organism; thus our ability to understand and manipulate Pc is limited to organisms isolated directly from animal models. Cross-reactive epitopes on Pc isolated from various mammalian hosts have been demonstrated but appear to be rare. Active immunization is generally not considered for the prevention of classic opportunistic pathogens. However, we have shown that immunization with whole Pc can protect mice from Pc pneumonia (PCP) when they are subsequently immunosuppressed and challenged with Pc. To determine how the known host species-specific antigenic variation of Pc would affect the immune response, mice were immunized with either mouse or ferret-derived Pc, with subsequent analysis of the immune response and ability of the mice to resist infection after immunosuppression and challenge. Measurement of Pc specific antibody response by ELISA showed that the antibody response was highly restricted to the strain of Pc used to immunize. For example, the ELISA OD to mouse Pc glycoprotein A after immunization with mouse Pc, ferret Pc, or normal ferret lung homogenate was.609;.115; and.122, respectively. The same results were observed using a crude homogenate of Pc as solid phase antigens in the ELISA. Western blot analysis confirmed the poor induction of antibodies to mouse Pc by immunization with ferret Pc. Mice immunized with mouse-derived Pc were completely protected from PCP when challenged by intratracheal inoculation with mouse Pc. Consistent with the results observed in our antibody assays, immunization with ferret Pc had no protective effect after a similar challenge. These results show that: 1) Pc is host species restricted; 2) immunization with one type of Pc does not induce significant cross-reactive immunity; and 3) any possible immunotherapy for PCP in humans must take into consideration these biologically significant antigenic differences in Pc of animal origin.