Is every component of a virus absolutely essential for its infection and replication in a host cell? Or can you just have parts of it to cause infection?
Unless you have a specific virus in mind, I can speak generally about viruses. It seems that the viral genome is reduced to just essential genes. Many viruses segregate their genome into periods of expression: early, mid and late. The early proteins are so-called because they encode for proteins that help their entry into cells (what is called infection). Middle and late genes tend to help the virus co-opt cellular machinery to aid in viral replication. Late genes are then useful for final stages of viral assembly and ultimtely their exit from the cell (lysis) and the newly made viral particles can go on and infect more cells. When viewed in this manner, it follows that removing even one coding region from the viral genome would inhibit the overall infection/reproduction/assembly process. In fact, if you remove the gene encoding for a viral envelope protein, purified virus added to cells in culture are able to infect cells, but not produce complete and functional viral particles. In this sense, they are infection-competent, but reproduction-incompetent.
HIV has nine genes. According to the paper cited below, three of these genes (vpu, vpr and nef) are not required for replication in cultured cells.
(PBMC = peripheral blood mononuclear cell)
As might be expected, none of these genes encodes a component of the virus - they are regulatory genes involved in co-ordinating the infective cycle of the virus. They are however implicated in HIV-associated disease symptoms. According to the Los Alamos National Laboratory HIV Sequence Compendium 2008:
vpu is an integral membrane protein which promotes extracellular release of viral particles and degrades CD4 in the ER.
vpr promotes nuclear localization of the preintegration complex, inhibits cell division, and arrests infected cells at G2/M
nef downregulates CD4 and class I MHC