Part of our antiviral strategy against HIV-I was based on the use of a defective interfering HIV-1 particle which encodes an antiviral gene product such as a multitarget-ribozyme. In addition, the goal was that these particles were able to target HIV-1 infected cells that express the HIV-1 Env protein. The hypothesis was that this could be achieved by placing the HIV-1 receptor, CD4, and one of the co-receptors, such as CXCR4, into the envelope of these defective HIV-1 particles. We have generated HIV-like particles which carry CD4 or several different chimeric versions of CD4 with different anchoring domains in their envelope. These particles showed a novel and unexpected cell binding activity which was dependent on the presence of the CD4 ectodomain but independent of the presence of the HIV-1 Env protein on the target cell. Binding of CD4(+) particles to cells was very efficient and similar in strength as the binding of CD 4 to HIV-1 Env. Polyclonal antibodies to CD4 as well as soluble CD4 did not totally abolish cell binding under conditions which totally blocked CD4-Env binding. CD4 had to be on present in the viral envelope to be able to bind to cells. CD4(+) cells did not bind the virus particles. The results indicate that a different microenvironment for CD4 may exist in the envelope of virus particles which exposes new binding activities that are hidden in cellular membranes. The CD4 particles could also be immunoprecipitated using a monoclonal antibody to CDCR4 when the particles were generated in human HeLa or HEK 293 cells, but not when they were generated in COS or NIH3T3 cells. Particles carrying both, CD4 and CSCR4 were tested for binding and transduction of cells expressing HIV-1 Env or the HIV-2 Env proteins. Control transduction included pseudotype HIV-1 like particles carrying the vesicular stomatitis virus (VSV) G protein. Until now no specific infection of cells expressing Env protein has been detected at a level of detectability of approximately 1/1000th of the transduction efficiency through the VSV G protein. It is possible that the unique cell binding activity which was observed and which does not require Env protein may have inhibited membrane fusion and infection. To be able to achieve a targeted infection, most likely requires an elimination of this new cell binding activity, which probably resides within the membrane proximal region of the CD4 ectodomain. Our results also indicate that the insertion of foreign proteins into viral membranes, including even a seven transmembrane protein like CSCR4, is not as restricted as previously assumed. This opens the possibility of using enveloped viruses such as our HIV-1 like particle to study cellular membrane proteins and potentially transfer them to other cells by membrane fusion.
