Deficient uracil base excision repair leads to persistent dUMP in HIV proviruses during infection of monocytes and macrophages

Abstract

Non-dividing cells of the myeloid lineage such as monocytes and macrophages are target cells of HIV that have low dNTP pool concentrations and elevated levels of dUTP, which leads to frequent incorporation of dUMP opposite to A during reverse transcription ("uracilation"). One factor determining the fate of dUMP in proviral DNA is the host cell uracil base excision repair (UBER) system. Here we explore the relative UBER capacity of monocytes (MC) and monocyte-derived macrophages (MDM) and the fate of integrated uracilated viruses in both cell types to understand the implications of viral dUMP on HIV diversification and infectivity. We find that the kinetics for MC infection is compatible with their lifetime in vivo and their near absence of hUNG2 activity is consistent with the retention of viral dUMP at high levels at least until differentiation into macrophages, where UBER becomes possible. Overexpression of human uracil DNA glycosylase in MDM prior to infection resulted in rapid removal of dUMP from HIV cDNA and near complete depletion of dUMP-containing viral copies. This finding establishes that the low hUNG2 expression level in these cells limits UBER but that hUNG2 is restrictive against uracilated viruses. In contrast, overexpression of hUNG2 after viral integration did not accelerate the excision of uracils, suggesting that they may poorly accessible in the context of chromatin. We found that viral DNA molecules with incorporated dUMP contained unique (+) strand transversion mutations that were not observed when dUMP was absent (G→T, T→A, T→G, A→C). These observations and other considerations suggest that dUMP introduces errors predominantly during (-) strand synthesis when the template is RNA. Overall, the likelihood of producing a functional virus from in vitro infection of MC is about 50-fold and 300-fold reduced as compared to MDM and activated T cells. The results implicate viral dUMP incorporation in MC and MDM as a potential viral diversification and restriction pathway during human HIV infection.

Fig 1. dUTP and dTTP pool measurements from methanol extracts of MDM, MC and comparison with Hap1 dividing cells.

A single nucleotide extension assay (SNE) was used to measure dUTP and dTTP levels in MDM, MC and the Hap1 dividing cell line (S1 Fig). The specific measurement of dUTP in a mixed pool of TTP + dUTP is accomplished by the in vitro conversion of dUTP to dUMP + PP_i using dUTPase prior to the SNE assay. (a) dUTP and dTTP pool measurements from methanol extracts of MDM, MC and comparison with Hap1 dividing cells. The RTase control (dTTP only) establishes that RTase is required for converting the n bp substrate into n + 1 and that dUTPase does not inhibit extension in the absence of dUTP (see methods). The dUTPase control establishes that RTase can fully extend the probe by a single nucleotide in the presence of a dNTP mixture with dUTP replacing dTTP. The addition of dUTPase completely abolishes extension under these conditions. The ratio of the intensities of the SNE extension bands to the original primer band gives the fractional extension arising from [dTTP + dUTP] ie. (n+1 band)/{(n + 1 band) + (primer band)}. When dUTPase is added before the reaction, the extension can only arise from dTTP. Thus, the level of dUTP associated extension is determined by the difference in extension between the +/- dUTPase reactions. (b) Combined dTTP + dUTP pool measurements for MDM and MC from two donors (D1 and D2) and the Hap1 dividing cell line. Error bars are standard errors from triplicate measurements. (c) dUTP/dTTP ratio for the various cell types. ND, not detected. Error bars are standard errors from triplicate measurements. (d) Determination of the relative activity of HIV RTase for incorporation of dTTP or dUTP opposite to adenine. The solid line is a theoretical fit to the entire time course, while the dashed lines are initial rate linear fits, which are statistically indistinguishable. The concentration of the template DNA was 50 nM and concentration of each dNTP was 100 nM, which is the estimated concentration of these dNTPs in MDM or MC.

Fig 2. Determination of Uracil Base Excitation Repair (UBER) capacity of infected and uninfected MCs, MDMs and comparison with the Hap1 dividing cell line.

MC or MDM were infected with replication competent CCR5 tropic HIV-1^BAL virus at MOI = 5. Extracts from MC (a, c, e, g, i) and MDM (b, d, f, h, j) using uninfected (-) or infected cells (+) at 3 days post-infection were processed for western blotting. Blots were performed using 10 μg of total cell protein, except for Lig IIIα, where 20 μg was used. Due to the low levels of hUNG present in non-dividing cells, native extracts (10 μg) were prepared for measurement of UNG enzymatic activity using a sensitive real-time fluorescence assay (see methods). The specific measurement of UNG activity was established by the addition of UNG inhibitor UGI (4 μM), which completely inhibited the measured activity. Panels a to j show western blots or activity assays for detection of (a, b) uracil DNA glycosylase (UNG); (c, d) AP endonuclease 1, (APE1); (e, f) DNA polymerase β, (polβ); (g, h) ligase III (Lig); and (i, j). APOBECA3A (A3A) and APOBECA3G (A3G); tubulin (Tub) loading control was used in each blot.

Fig 3. Characteristics of HIV infection in MC and macrophages infected before differentiation (MDM^EI) and after differentiation (MDM^LI).

(a) Infection of MC and MDM with HIV-1^BaL virus was performed using three conditions (MOI = 5). In one condition, MC were immediately infected and then cultured using nonadherent conditions without M-CSF for seven days to maintain a monocyte phenotype. MDM^EI refers to “early infection” of freshly isolated MC, which were subsequently differentiated into MDM under adherent conditions in the presence of M-CSF. MDM^LI refers to “late infection” where MDM were first fully differentiated before infection. In all cases, HIV fusion inhibitor drug, enfuvirtide, (0.2 μM final concentration) was added to culture medium after 24 hours of infection to prevent multiple rounds of infection. (b) Fraction of provirus in MC, MDM^EI and MDM^LI that contain uracil as determined by alu-gag Ex-qPCR. ND: not detected (c) Time course for appearance of early reverse transcription (ERT) products as determined by RT-qPCR using a primer set targeting the 5’LTR region of the HIV genome. (d) Time course for appearance of late reverse transcription (LRT) products as determined by RT-qPCR using a primer set targeting the LTR region of the HIV genome. (e) Time course for appearance of proviral DNA copies determined by alu-gag qPCR. Identical infections activated T cells were used as controls and are shown in panels d, e and f. Abbrev: ND, not detected. * not done. All data are averages obtained from three blood donors.

Fig 4. dUMP in HIV DNA during infection of MC and MDM arises predominantly from dUTP.

MC and MDM^LI were infected with HIV-1^Bal virus at an MOI of 5 (1 pg p24/cell) in the presence (+) and absence (-) of 5 mM deoxythymidine (dThyd). Copy number measurements of uracilated (red) and non-uracilated (blue) proviruses were made 7 dpi using the Alu-gag Ex-qPCR method. (a) MDM^LI and (b) MC infected in the absence and presence of dThyd. p-values from an unpaired Student’s t-test are shown.

Fig 5. Mutagenic effects of uracil incorporation on HIV proviral DNA and extracellular viral RNA in MDM and MC.

Sequences of single proviral clones and extracellular viral RNA from infected MDM and MC. MC and MDMs were infected with equivalent amount of virus (1 pg p24/cell). Seven days after infection total cellular DNA and extracellular viral RNA were extracted from cells and the culture supernatants, respectively. Single viral copies were amplified by limiting-dilution nested PCR and sequenced by the Sanger method. Sequences were aligned to our laboratory reference HIV-1^BaL sequence. (a) Mutational spectrum of HIV proviral DNA sequences from infected MC (left top), MDM (left middle) and extracellular viral RNA extracted from MDM (left bottom). (b) A population of viral cDNAs was sequenced before and after the removal of dUMP-containing sequences as indicated. Removal of dUMP-containing viral sequences was accomplished by treatment of the total DNA population with UNG prior to PCR amplification and clonal sequencing. HIV proviral DNA sequences where the dUMP fraction was subtracted contained almost exclusively transition mutations (left pie chart), while the total proviral DNA without subtraction showed the same transition mutations but also unique transversion mutations (right pie chart). (c) Transversion mutations are associated with loci that contain higher average A/T(U) content. In this analysis, the average A + T(U) content over a window of three flanking bases on both sides of each transversion or transition mutation was calculated. The loci with transversion mutations had a higher average content than the 29 loci that showed transition mutations. p-values from Wilcoxon rank-sum test indicate the statistical significance of the average A + T(U) content in the six base window surrounding the transition and transversion sites.

Fig 6. Normalized viral output from HIV proviruses in MC, MDM and activated T cells.

All cell types were infected with Bal virus at MOI of 5 (1 pg p24/cell) and seven days after infection culture supernatants and cellular DNA were processed. MC were infected immediately after isolation and then cultured for seven days without differentiation using nonadherent conditions in the absence of M-CSF. MDM^EI were infected as freshly isolated MC and then differentiated into MDM under adherent conditions in the presence of M-CSF. MDM^LI were infected after differentiation for seven days in the presence of M-CSF. T cell controls were infected and cultured in IL-2 medium for seven days. T20 entry inhibitor (10 μM) was used to prevent reinfection. (a) Viral protein 24 (p24) levels in culture supernatants were measured by ELISA and normalized to total proviral copies determined by alu-gag qPCR. (b) Extracellular viral RNA was measured using RT-PCR and normalized to total proviral copies determined by alu-gag qPCR. p-values from an unpaired Student’s t-test are shown. (c) Experimental scheme for determining infectivity of virus generated from MC, MDM^EI and MDM^LI producer cells. (d) Viral supernatants from MC, MDM^EI and MDM^LI and T cell producer cell cultures were collected at 7 dpi and MOLT-4/CCR5 target cells were infected using a 0.1 pg p24/target cell. Seven days after infection of target cells, cellular DNA was extracted and levels of proviral DNA were determined by qPCR. Relative infectivity is shown as LRT copies per target cell. p-values from an unpaired Student’s t-test are shown. (e) p24 levels were measured by the ELISA method. The data were obtained from one of the three donor samples analyzed at the proviral level in panel d.

Fig 7. Exogenous expression of hUNG2 depletes uracilated HIV DNA products.

Exogenous full-length hUNG2 was over expressed in MDM under control of a doxycycline (Dox) inducible promoter by lentiviral transduction (MOI = 5). The timelines above each panel indicate the experimental course and the points where hUNG2 expression was induced relative to infection with HIV^NL4-3. (a) hUNG2 expression was induced with doxycycline (1 μg/ml) one day before HIV^NL4-3 infection. Proviral copy numbers (left panel) and the fraction of HIV proviral DNA that contained dUMP were measured at indicated times (FracU, right panel). (b) Expression of hUNG2 was induced 7-days after infection with HIV^NL4-3 in order to allow HIV to fully integrate. Proviral copy numbers (left panel) and the fraction of HIV proviral DNA that contained dUMP were measured at indicated times (FracU, right panel). Uninduced (blue) and no lentiviral infection (green) controls were also performed.