XPB and XPD: defense against infection
We have shown that retroviral infection increases when XPB or XPD are mutant. This was shown for both the lentivirus HIV-1 and the distantly related gamma retrovirus Moloney Murine Leukemia Virus (MoMLV). Cells expressing wild type XPB protein show the least infection efficiency, while cells expressing the nucleotide excision repair mutant XPB(F99S) gene show the greatest infection efficiency. XPB mediated DNA repair capacity inversely correlated with retroviral infection efficiency. This suggests that XPB and XPD are part of an innate cellular host defense against retroviral infection.
XPB and XPD are DNA helicases with opposing polarity. They are both essential genes and part of the basal transcription complex TFIIH. Due to their essential nature, only point mutations of these genes may be studied. Mutations of XPB and XPD are extremely rare and are associated with the human diseases xeroderma pigmentosum, Cockayne Syndrome, and tricothiodystrophy. Ten proteins comprise TFIIH. The complex opens bubbles in DNA allowing access of RNA polymerase during transcription or additional repair proteins during nucleotide excision repair. Only DNA repair mutations of XPB and XPD affect retroviral infection, transcription associated mutations have no affect on infection. Additional proteins of the nucleotide excision repair pathway also did not affect retroviral infection.
Further analysis revealed that wild type XPB and XPD were associated with less retroviral cDNA accumulation during infection. It was unclear if this was an effect on cDNA synthesis by reverse transcriptase or an effect on degradation. XPB and XPD are nuclear resident proteins. XPB and XPD cell lines were arrested with aphidicolin and infected. Following 20 hours of infection during arrest, aphidicolin was removed allowing the cell cycle to resume. MoMLV cDNA is not able to enter the nucleus when cells are arrested. When MoMLV is trapped in the cytoplasm, there is no difference in the accumulation of cDNA between wild type or mutant cell lines. However, when the cell cycle resumes and MoMLV is able to enter the nucleus, the cDNA is more rapidly degraded in wild type cells compared to mutant cells. This data suggests that the TFIIH complex participates in the degradation of retroviral cDNA after it enters the nucleus.
Base Excision Repair: pirated by lentiviruses
All genes of the short patch oxidative base excision repair pathway
were identified in an siRNA screen for host factors affecting HIV-1 infection. HIV infection of cells with deletions of base excision repair genes compared to matched wild type cells confirmed that these DNA repair proteins are important for efficient infection. Deletion of two genes led to an even greater decrease of HIV infection.
While these proteins affect lentiviral infection, they do not affect infection by the gamma retrovirus Moloney murine leukemia virus. In contrast to the nucleotide excision repair proteins, the base excision repair proteins had no effect on the accumulation of retroviral cDNA. Instead the base excision repair pathway affects the integration efficiency of HIV.
Retroviral integration to the host DNA is not random. All retroviruses display a unique profile of preference for sites at both the chromatin level and the sequence level. The chromatin elements may include transcription units, promoters, and CpG islands. HIV favors integration to transcription units, but not promoters; while MoMLV favors integration to promoters. This level of preference is largely driven by host integration co-factors. In the case of HIV, the host co-factor is LEDGF/p75. This host protein is a transcription activator that directs HIV integration to transcription units. However, LEDGF has no effect on the subtle sequence preference at the HIV integration site.
Sequencing of HIV integration sites in base excision repair deletion cell lines showed that these host proteins have no effect on HIV integration near chromatin level elements, including transcription units, promoters, or CpG islands. However, the sequence preference at the HIV integration sites did change in the absence of base excision repair proteins. The original siRNA screen showed that only glycosylases recognizing oxidative base damage affect HIV infection; glycosylases that recognize any other types of base damage have no effect on HIV infection. Oxidative base damage is common and most frequently occurs on guanine (G). The HIV integration site sequence preference includes many G/C base pairs. Interestingly, MoMLV is not affected by base excision repair proteins and has no preference for G/C base pairs at its integration sites. The glycosylase OGG1 only recognizes oxidative damage of G. In the absence of OGG1, the HIV integration site preference for G/C base pairs is lost at several positions flanking the points of joining. When wild type cells are treated with the oxidative damage agent hydrogen peroxide, the HIV integration site sequence preference for G/C base pairs is extended. These data suggest that the base excision repair pathway influences HIV integration at the sequence level.