Macrophage Tropism of HIV-1 Depends on Efficient Cellular dNTP Utilization by Reverse Transcriptase*

Retroviruses utilize cellular dNTPs to perform proviral DNA synthesis in infected host cells. Unlike oncoretroviruses, which replicate in dividing cells, lentiviruses, such as human immunodeficiency virus type 1 (HIV-1) and simian immunodeficiency virus, are capable of efficiently replicating in non-dividing cells (terminally differentiated macrophages) as well as dividing cells (i.e. activated CD4+ T cells). In general, non-dividing cells are likely to have low cellular dNTP content compared with dividing cells. Here, by employing a novel assay for cellular dNTP content, we determined the dNTP concentrations in two HIV-1 target cells, macrophages and activated CD4+ T cells. We found that human macrophages contained 130-250-fold lower dNTP concentrations than activated human CD4+ T cells. Biochemical analysis revealed that, unlike oncoretroviral reverse transcriptases (RTs), lentiviral RTs efficiently synthesize DNA even in the presence of the low dNTP concentrations equivalent to those found in macrophages. In keeping with this observation, HIV-1 vectors containing mutant HIV-1 RTs, which kinetically mimic oncoretroviral RTs, failed to transduce human macrophages despite retaining normal infectivity for activated CD4+ T cells and other dividing cells. These results suggest that the ability of HIV-1 to infect macrophages, which is essential to establishing the early pathogenesis of HIV-1 infection, depends, at least in part, on enzymatic adaptation of HIV-1 RT to efficiently catalyze DNA synthesis in limited cellular dNTP substrate environments.

Cellular dNTP levels serve as a biomarker for the replicative capacity of mammalian cells (for review, see Ref. 1). Consistent with this function, numerous studies have reported that dNTP contents in cancer cells and transformed cell lines are elevated 3-10-fold compared with normal cells and that dividing cells have higher dNTP contents than non-dividing cells (1)(2)(3)(4)(5)(6). Cellular dNTP levels also fluctuate during the cell cycle and upon DNA damage, when replication and repair DNA synthesis lev-els are altered (6 -10). In addition, many viruses utilize cellular dNTPs to carry out their replicative cycle. These viruses include lentiviruses, such as the human and simian immunodeficiency viruses (HIV-1 1 and SIV), which are unique among retroviruses in their ability to infect both non-dividing cells, such as terminally differentiated macrophages (11), and dividing cells, such as activated CD4ϩ T cells. Changes in cell tropism between macrophages and T cells that occur during the course of lentiviral infection are closely related to immunological and clinical manifestations of lentivirus infection (i.e. CD4ϩ T cell decrease and AIDS). In contrast to lentiviruses, oncoretroviruses productively replicate only in dividing cells (11,12). Therefore, it is intriguing to consider how lentiviral DNA polymerases may have evolved to replicate efficiently in quiescent cells (i.e. terminally differentiated/non-dividing macrophages) that presumably contain low levels of dNTPs.
In this study, we established a sensitive dNTP assay to measure dNTP concentrations in primary HIV-1-susceptible host cells. Biochemical analysis revealed that HIV-1 reverse transcriptase (RT) efficiently performs processive DNA synthesis under reaction conditions that contained dNTP concentrations equivalent to those found in macrophages. In contrast, oncoretroviral RTs were unable to efficiently polymerize under these conditions. Finally, we designed pseudotyped HIV-1 virus vectors that contained either wild-type RT or RTs with mutations that affected dNTP binding affinity. An analysis of reporter gene expression in vector-infected target cells revealed that the mutant constructs could transduce only actively dividing cells (cell lines, primary activated T cells), whereas the wild-type vectors could transduce both dividing and non-dividing (macrophage) host cells. Collectively, these data provide evidence that the unique ability of HIV-1 to infect non-dividing cells depends, at least in part, on the evolutionary adaptation of its reverse transcriptase to function under conditions of limiting dNTP availability.

Single Nucleotide Incorporation Assay
Four different 19-mer DNA templates containing sequence variations (N) at the 5Ј end nucleotide (5Ј-NTGGCGCCCGAACAGGGAC-3Ј) were individually annealed to an 18-mer DNA primer (5Ј-GTCCCTGT-TCGGGCGCCA-3Ј), which was 32 P-labeled at its 5Ј end (template: primer, 4:1). The nucleotide at the 5Ј end of the primer determines the dNTP to be measured. The primer (160 pmol in 80 l) was labeled using 40 units of T4 polynucleotide kinase (New England Biolabs) with 80 Ci of [␣-32 P]ATP (Amersham Biosciences) for 30 min at 37°C. An additional 40 units of T4 polynucleotide kinase was added for an additional 30-min period of labeling. After heat inactivation (95°C, 10 min), the labeled primer was split into four separate tubes annealed with each of the four 19-mer templates (160 pmol) with the addition of 10ϫ STE buffer (100 mM NaCl and 5 mM EDTA for final concentration) in a final volume of 50 l. The template/primer mixture was incubated for 10 min at 95°C, 5 min at 55°C, and 5 min at 22°C and stored on ice until used.
The template/primer complexes (T/Ps) were extended using RT proteins in a standard dNTP assay reaction. Each assay reaction (20 l) contained 0.2 pmol of T/P (10 nM, primer concentration), 2 l of appropriate dNTPs (Amersham Biosciences) or extracted cellular dNTPs, 25 mM Tris-HCl, pH 8.0, 100 mM KCl, 2 mM dithiothreitol, 5 mM MgCl 2 , 5 M (dT) 20 , and 0.1 mg/ml bovine serum albumin (New England Biolabs). Reactions were initiated by adding excess RT proteins (60 nM) in relation to dNTP concentration (0.2-6.4 nM or 4 -128 fmol in 20 l) and incubated at 37°C for 5 min. Reactions were terminated with 10 l of 40 mM EDTA, 99% formamide. Reaction products were immediately denatured by incubating at 95°C for 5 min, and 4 l of each 30 l of final reaction mixture was quantitated by phosphorimaging analysis (PerkinElmer Life Sciences) of 14% acrylamide-urea denaturing gels (SequaGel, National Diagnostics; model S2 sequencing gel electrophoresis apparatus, Labrepco). Note that in our purification protocol, ϳ30 -60% of the purified RT protein was active as determined by our pre-steady-state kinetic assay using a similar 18-mer/40-mer primer/ RNA template (17,18). The reaction with less RT (10 nM), which is still a higher concentration than the highest dNTP concentration used (6.4 nM), gave a similar incorporation profile under the assay condition.

Standard Curves for Single Nucleotide Incorporation
The gels obtained from the incorporation assay (see Fig. 1A) were subjected to phosphorimaging analysis (see above). The percent of primer extension in each reaction was calculated by determining the ratios of extended versus total (extended and unextended) primers. Each signal for the extended products was normalized using the background signal in the control reactions incubated without RT (see Fig.  1C, lane C). The calculated percentage primer extension was plotted from triplicate reactions as a function of the dNTP quantity used, generating standard curves for all four dNTPs. The percentage primer extension determined in triplicate reactions containing extracted cellular dNTPs was then extrapolated to the standard curves to determine the dNTP contents present in the cellular samples.

Primer Extension by RT Proteins
The primer extension assay was modified from a previously described misincorporation assay (18). Briefly, an RNA T/P was prepared by annealing a 38-mer RNA (5Ј-AAGCUUGGCUGCAGAAUAUUGC-UAGCGGGAAUUCGGCGCG-3Ј, Dharmacon Research) to the 17-mer A primer (5Ј-CGCGCCGAATTCCCGCT-3Ј; template:primer, 2.5:1, Invitrogen) 32 P-labeled at the 5Ј end by T4 polynucleotide kinase. Assay mixtures (20 l) contained 10 nM T/P, the RT protein concentrations specified in the legends to Fig. 2, A and B, and 4 dNTPs at concentrations indicated in the figure legends under the condition described in the dNTP assay above. Reactions were incubated at 37°C for 5 min and terminated for analysis as described in the dNTP assay. Concentrations of RTs (i.e. 1 and 4ϫ) and dNTPs (i.e. 4 and 0.04 M) used in each primer extension experiment were described in each figure legend. This reaction condition allows multiple rounds of primer extension. We also performed the single round primer extension. In this reaction, RTs were preincubated with the 32 P-labeled 17-mer/38-mer RNA template (10 nM), and the reaction was initiated by adding a mixture of dNTP (0.04 M) and a molar excess of cold T/P (1 M) for 3 min at 37°C. For the trap control, RTs were preincubated with a mixture of the 32 P-labeled T/P and the cold T/P, and the reaction was initiated by adding dNTP (0.04 M). We also employed a 120-nucleotide-long RNA template annealed to a 32 P-labeled 20-mer primer (19). This T/P was also extended by differ-ent concentrations of RTs and dNTP for 10 min at 37°C (see Fig. 1D legend).

Steady-state Kinetic Analysis of dNTP Incorporation
The steady-state kinetic assay protocol described by Boosalis et al. (20) was modified to determine DNA polymerization efficiencies of DNA polymerases with the 18-mer/19-mer template/primers used in the dNTP assay (see above). Reaction conditions were the same as those described in the dNTP assay except for DNA polymerase concentrations (0.5-2 nM). For the analysis of dNTP incorporation kinetics, four different templates (5Ј template nucleotides T, C, G, or A for dATP, dGTP, dCTP, or dTTP incorporation, respectively) annealed to the 32 P-labeled 18-mer primer were used (see above). The amounts of RTs and incubation times were adjusted to yield an extension that was 30 -60% of the total labeled primer (0.8 pmol) at the highest dNTP concentrations, and the reactions were repeated with eight decreasing concentrations of each dNTP. For HIV-1 RT, 5, 10, 20, 40, 80, 160, 320, and 720 nM dNTPs were used, and for MuLV RT, 1, 2, 4, 8, 16, 32, 64, and 128 M were used. We also measured the steady-state kinetic values using the processive reverse transcription with the 38-mer RNA annealed to the 32 P-labeled 17-mer. In this analysis, the different concentrations of all four dNTP mixtures were used. Products (single nucleotide-incorporated primer or all extended products) were resolved in 14% polyacrylamide-urea gel and quantitated by phosphorimaging analysis using the OptiQuant software (PerkinElmer Life Sciences). The k cat and K m values were determined from the Michaelis-Menten equation.

dNTP Extraction from Cells
The protocol for dNTP extraction from cells was similar to that described previously (21). Briefly, cell pellets (5 ϫ 10 4 -2 ϫ 10 6 cells) were washed twice with 1ϫ Dulbecco's phosphate-buffered saline (Mediatech) and resuspended in 100 l of ice-cold 60% methanol. Samples were vortexed vigorously to lyse the cells and then heated at 95°C for 3 min prior to centrifugation at 12,000 ϫ g for 30 s. The supernatants were collected and completely vacuum-dried using a SpeedVac (Savant) with medium heat. The dried pellets were subsequently resuspended in dNTP buffer (50 mM Tris-HCl, pH 8.0, and 10 mM MgCl 2 ; 100 l for 1 ϫ 10 6 HeLa cells and 10 l for 1 ϫ 10 6 primary cells), and usually 1-2 l of the extracted dNTP samples were used for each 20-l single nucleotide incorporation reaction (see above). The proper dilutions of the dNTP samples were prepared for the assay to make the primer extension values lie within the linear ranges of the dNTP incorporation (2-32% primer extension). The extracted dNTP samples were stored at Ϫ70°C until used. Several different volumes of the extracted dNTP samples were also used to confirm the linearity of the primer extension. In addition, the dNTP samples were prepared from different cell numbers depending on the recovery efficiency of the primary cells from each blood sample (see below). However, the dNTP content of each cell type was normalized by pmol/1 ϫ 10 6 (see Fig. 1F).

Isolation and Culture of Human T Cells and Monocyte-derived Macrophages
Human macrophages and CD4ϩ T cells were isolated from human buffy coats (Blood Research Institute) as described (22). Peripheral blood mononuclear cells were harvested from Ficoll density gradients (Lymphoprep, Axis-Shield PoC AS, Oslo, Norway), and monocytes were then purified using immunomagnetic selection with anti-CD14 antibody-conjugated magnetic beads following the manufacturer's recommendations (Miltenyi Biotech). CD4ϩ lymphocytes were isolated from monocyte-depleted buffy coats by immunomagnetic selection using anti-CD4 antibody-conjugated magnetic beads. The purified human monocytes were incubated in 10-cm dishes in RPMI 1640 medium (Mediatech) containing 20% human AB serum (Sigma) for 4 days in the presence of 5 ng/ml human recombinant GM-CSF (R&D Systems) and then incubated for an additional 3 days in the absence of GM-CSF to allow differentiation into macrophages. The purified resting T cells were maintained in RPMI 1640 medium supplemented with 10% autologous human serum, although activated T cells were propagated in RPMI 1640 medium supplemented with 10% autologous human serum and 5 ng/ml phytohemagglutinin (Sigma) for 24 h and then incubated in RPMI 1640 medium with 10% human serum and 5 ng/ml human recombinant interleukin 2 (Sigma) for 3 days. These cell preparations were to prepare for dNTP extracts and also for analysis of cell volumes using confocal microscopy.

Pseudotyped Virus Cloning and Preparation
pHCMV-VSVG envelope vector (23) and pD3-GFP transfer vector, which encodes the HIV-1 NL4-3 genome with the eGFP gene in place of HIV-1 envelope, were used for preparing pseudotyped HIV-1. The V148I and Q151N mutations were cloned into pD3-GFP (pNL4-3-based) vector using the overlapping PCR with proper primers that were designed based on the pNL4-3 RT sequence. The created mutations were confirmed by sequencing with the 3305 primer (5Ј-GCACTATAGGCT-GTACTGTCC-3Ј), and correct SwaI relegation was confirmed with D3-GFP SEQ SwaI F (5Ј-CAGGCCATATCACCTAG-3Ј) and D3-GFP SEQ SwaI R (5Ј-TCTAACTGGTACCATAAC-3Ј) primers. 293FT cells (Invitrogen) were grown to ϳ90% confluency with Dulbecco's modified Eagle's medium, 10% fetal bovine serum, and 500 g/ml G418 sulfate in T-175 flasks. The cells were transfected with 5 g of envelope vector and 25 g of transfer vector using Lipofectamine 2000 (Invitrogen) following the manufacturer's recommendations (in the absence of geneticin). Cells were split in half into two new T-175 flasks 16 h posttransfection. Supernatant was collected 48 and 72 h post-transfection and stored at 4°C. Virus was harvested by ultracentrifugation at 22,000 rpm for 2 h at 4°C with the Beckman SW 28 rotor. Virus was titrated by infection of HeLa cells 2 in the presence of polybrene, and the infection was monitored by fixation and flow cytometry analysis for GFP expression 48 h post-infection. p24 levels were also measured using the HIV-1 p24 antigen enzyme-linked immunosorbent assay test system (Beckman).
As a control, transfections were also performed in the absence of the pHCMV-VSVG vector, and this virus was titered on HeLa cells revealing no detectable GFP expression. In addition, controls were performed by pretreating HeLa cells for 2 h prior to and during infection with 0.5 mM cycloheximide, which is a general translation inhibitor. Cells were fixed and analyzed by flow cytometry 48 h post-infection. Results showed that very little protein was carried over with the pseudotyped virus preparations, as evidenced by Ͻ1% GFP-positive cells with plus cycloheximide infections compared with Ͼ20% GFP-positive cells in minus cycloheximide infections (data not shown).

Infection of Primary CD4ϩ T Cells and Monocyte-derived Macrophage
T cells and macrophage were purified essentially as described above.
T Cell Infections-After 5 days of T cell stimulation with phytohemagglutinin and interleukin 2 (see above), 2 ϫ 10 5 stimulated T cells were infected in the presence of polybrene with a multiplicity of infection (MOI) of 100 based on HeLa cell titering. For QN virus infections, an MOI of 65 was used because of trouble generating high titer QN virus. These infections reflect the addition of viral stocks containing 200 -300 ng of p24. Infected cells were incubated at 37°C and 5% CO 2 in a 100-l total volume. Two hours post-infection, 100 l of RPMI 1640 medium with 10% heat-inactivated human AB serum and 4 ng/l interleukin 2 was added, and infections were incubated for an additional 46 h. Cells were visualized and photographed by fluorescent microscopy to detect GFP fluorescence by infected cells. The percent of infection was determined after cells were prepared for flow analysis by fixation with 0.5% formaldehyde.
Macrophage Infections-After the initial magnetic bead purification, 2 ϫ 10 6 CD14ϩ cells/well were plated into 6-well plates and cultured in RPMI 1640 medium with 2% heat-inactivated human AB serum and 5 ng/ml GM-CSF for 48 h. This was followed by a 48-h incubation in RPMI 1640 medium with 20% heat-inactivated human AB serum and 5 ng/ml GM-CSF. Cells were then incubated for an additional 8 days in the absence of GM-CSF, with fresh medium added after the initial 4 days. Macrophages were infected with pseudotyped virus at an MOI of 40 based on HeLa cell titering and assuming 1 ϫ 10 6 cells/well in the presence of polybrene. These infections reflect the addition of viral stocks containing 80 -130 ng of p24. Initial infections were in 250-l volumes for the first 2 h at 37°C and 5% CO 2 followed by the addition of 2 ml of RPMI 1640 medium with 20% heat-inactivated human AB serum. Incubations were continued for up to 6 days, and cells were periodically examined for GFP fluorescence by fluorescent microscopy. Cells were fixed in 0.5% formaldehyde after disassociation from the plates by 30 min of incubation with 2 mM EDTA and gentle scraping. The percent of infection was determined by flow cytometry.

Determination of dNTP Levels in HIV-1 Target Cells, Human
Macrophages, and Resting or Activated CD4ϩ T Cells-First, we established a highly sensitive enzymatic dNTP assay that allowed us to measure the dNTP contents of primary and non-dividing cells such as macrophages. For the enzymatic measurement of each dNTP contained in a mixture of four different dNTPs, we designed a 32 P-labeled 18-mer primer that could anneal to any of four different 19-mer DNA templates (see "Experimental Procedures"). Because these dNTP assay T/Ps contain only one of the four possible nucleotides at the 5Ј protruding end of the template, DNA polymerases can incorporate only the complementary dNTP onto each of the four T/Ps. Using the specific T/Ps, we found that HIV-1 RT efficiently performed single nucleotide incorporation, even at very low dNTP concentrations, compared with KF of Escherichia coli DNA polymerase I (data not shown) and MuLV RT (see below). When 200 fmol of T/Ps were extended by HIV-1 RT (60 nM) in the presence of different amounts of the individual dNTPs for 5 min at 37°C, 19-mer extended products were generated. As shown in Fig. 1A, incorporation of dCTP and dTTP was detected at concentrations as low as 4 fmol, which extended ϳ2% of the primers. Similar profiles were seen with dGTP and dATP (data not shown). The percent of primer extension in each reaction was then calculated to generate standard curves for incorporation of each dNTP onto the appropriate template/ primer (Fig. 1B, dTTP and dGTP). No detectable primer extension was observed in control reactions lacking either HIV-1 RT or the required dNTP Fig. 1C, lane C). The single nucleotide primer extension reaction for each of the four dNTPs was found to be linear for dNTP quantities between 4 and 128 fmol (0.2-6.4 nM), as shown in the standard curve with dTTP and dGTP (Fig. 1B) as well as with dATP and dCTP (data not shown). The linear dNTP incorporation appeared to be affected at higher dNTP amounts in some cases (see 128 fmol dTTP in Fig. 1B) because the observed primer extension levels with 128 fmol of dNTPs (55-60%) were lower than expected (64% at 128 fmol). Therefore, the upper dNTP limit of the assay should be 128 fmol. This suggested that the dNTP content measurement with the dNTP samples extracted from cells should be performed within the linear dNTP ranges (4 -64 fmol of dNTP or 2-32% of primer extension). The minimum detection of the dNTP content in this HIV-1 RT-based assay (4 fmol) is 25-and 500-fold more sensitive than the KF (0.1 pmol) (21) and high pressure liquid chromatography-based assays (Ͼ2 pmol) (25-28), respectively, that have been used previously for dNTP content measurement.
These reactions contained excess HIV-1 RT in relation to substrates, allowing depletion of provided dNTPs within the linear range, as shown in Fig. 1. This explanation is evidenced by the following: 1) the assay was performed at 10 and 20 min, and no additional primer extension compared with the 5 min standard reaction was observed (data not shown); and 2) the assay showed a 1:1 correlation between the amount of extended primer and the amount of dNTPs included in the reaction, particularly at the low dNTP amount (e.g. adding 4 fmol of dCTP resulted in 2% primer extension). dNTP samples obtained from cells contained all four dNTPs as well as structurally related rNTPs, which may be present in cells with concentrations 200 -1000-fold higher than the dNTPs (1). Thus, it is conceivable that an enzymatic dNTP assay might be influenced by the presence of such a vast excess of rNTPs. This is unlikely in the case of HIV-1 RT, which has been shown to incorporate rNTPs very inefficiently (i.e. only when their concentration is at least 0.5 mM) (29). Nonetheless, we tested our assay by adding rNTPs at a concentration 10,000fold higher than the highest dNTP concentration in the linear range of the assay, and rNTP incorporation was not detected (Fig. 1C). A second potential concern with using a primer extension reaction based on the HIV-1 RT is that this enzyme is known to be error-prone (i.e. capable of incorporating "incorrect" dNTPs) (30,31). We therefore tested whether incorrect dNTPs could be incorporated under the assay conditions used. For this test, the dGTP-specific template/primer pair was extended in the presence of a mixture of dATP, dCTP, and dTTP (0.5 M each), the three incorrect dNTPs; this incorrect dNTP concentration is ϳ100-fold higher than the highest correct dNTP concentrations that showed linear incorporation under these assay conditions (6.4 nM dGTP). As shown in Fig. 1C, no incorporation of incorrect dNTPs was observed even at this artificially high concentration of incorrect dNTPs. In fact, HIV-1 RT needs ϳ200 M incorrect dNTPs to carry out misincorporation under similar reaction conditions (32, 33). Therefore, incorrect dNTPs and rNTPs present in cellular dNTP samples should not interfere with the determination of individual dNTP levels under the assay conditions established.
We used the HIV-1 RT-based assay to examine dNTP levels in the major primary target cells for HIV-1 (i.e. human monocyte-derived macrophages and activated or resting CD4ϩ T cells). For this analysis, dNTPs were extracted from these primary cells (see "Experimental Procedures"), and the extracted dNTP samples with the proper dilutions were used as the dNTP source in single nucleotide extension assays with each of the four T/Ps. As shown in Fig. 1, D and F, human macrophages were found to contain ϳ4and ϳ20-fold less dNTPs than resting and activated T cells, respectively. The dNTP content of T cells determined in this assay was found to be similar to that previously determined by the KF based assay FIG. 1. Primary cell dNTP content determination by the HIV-1 RT-based dNTP assay. A, dCTP and dTTP incorporation by HIV-1 RT (60 nM) was measured using the dCTP-specific template/primer pair (200 fmol) in the presence of different amounts of dCTP. B, standard curve for the incorporation of dTTP and dGTP onto the dTTP-and dGTP-specific template/primer pairs. The percent of primer extension in each reaction was plotted after background normalization (see "Experimental Procedures"). Each data point was calculated from three independent reactions; error bars denote the standard deviation from the mean. C, incorporation of incorrect dNTPs and rNTPs. The G-specific template/primer pair was incubated with HIV-1 RT (60 nM) and 50 M rNTP or a mixture of three incorrect dNTPs (ϪdGTP), which corresponds to a mixture that contained 0.5 M each dTTP, dCTP, and dATP. D, dNTP samples were isolated from human activated (A/T) or resting (R/T) CD4ϩ T cells or CD14ϩ macrophages (M). dNTP samples contained in 1 (T cells) to 2 (macrophages) ϫ 10 5 cells were used for each assay reaction using dGTP-or dTTP-specific template/primer pairs. E, effect of other chemicals contained in the extracted dNTP samples. The dNTP assay was performed with either pure dNTP or mixtures of pure and extracted dNTPs. The amounts of the pure dNTPs used in the reactions are shown below the gels, and 1 (T cells) or 2 (macrophages) l of the cellular dNTP sample was used. The percent of primer extension in each reaction is also shown. Two different amounts of the T cell dNTP sample, 1 and 2ϫ, were used. The results for all template/primer pairs with the extracted dNTP samples are summarized in F. dNTP values shown represent mean values calculated from three independent experiments performed in triplicate; standard deviations are shown for each value. All reactions were performed as described under "Experimental Procedures," and products were analyzed by electrophoresis through 14% denaturing polyacrylamide gels followed by phosphorimaging analysis (see "Experimental Procedures").  (21). We also confirmed the dNTP content of human resting CD4ϩ T cells using the KF-based assay previously established, but this assay required 10 -20-fold more cells per assay because of the higher assay sensitivity limit (0.1 pmol).
Next, we tested the nonspecific inhibition of the HIV-1 RT activity by other chemicals co-extracted with cellular dNTPs under the dNTP assay condition. For this test, we performed the dNTP assay with either pure dNTP (64 or 8 fmol of dTTP for the macrophage samples and 16 fmol of dGTP for the T cell samples; Fig. 1E) or mixtures of pure dNTP and the extracted dNTP samples used in Fig. 1D. As shown in Fig. 1E, even with 2 l of the macrophage extract sample, which contains 16.8 fmol of dTTP (Fig. 1F), the primer extension with 64 fmol of dTTP was not reduced. Rather, compared with the reaction with only pure dTTP (30.1%), a slight increase (ϳ7%) was observed in the reaction with this mixture. This increase likely resulted from the additional incorporation of dTTP contained in the macrophages sample (16.8 fmol), which should have given ϳ8% primer extension. More obviously, as shown in Fig.  1E, the reaction with the mixture of 8 fmol of pure TTP (theoretically giving 4% primer extension) and 2 l of the macrophage sample (16.8 fmol of TTP giving 8.4% primer extension) showed 12.4% primer extension, whereas the reaction with only 8 fmol of pure dTTP showed the expected 4.3% primer extension. The 8.1% primer extension increase in the reaction with this mixture likely resulted from the incorporation of the 16.8 fmol of dTTP (Fig. 1F) contained in the 2 l of the macrophage sample. As shown in Fig. 1E, the additive primer extension was also observed in the reactions with the mixtures of pure dGTP (16 fmol, 7.6% primer extension) and the resting T cell samples containing 28 (1ϫ and 1 l) or 56 (2ϫ and 2 l) fmol of dGTP (giving additional 14.2 and 27.4% primer extensions, respectively) as well as activated T cells (data not shown). In addition, the expected primer extension increases were seen when the pure dNTPs (Fig. 1E) were added into the reactions with only extracted samples (Fig. 1, D and F). For example, the reaction with a mixture of 16 fmol of pure dGTP, which gave 7.6% primer extension, and the resting T cell sample (1ϫ and 1 l) containing 28 fmol dGTP (theoretically giving 14% primer extension (Fig. 1F) or actually giving 13.1% primer extension (Fig. 1D)) showed 21.8% primer extension (Fig. 1E). This indicates that the 16 fmol of pure dGTP, which produces 8% primer extension, contributed to 7.8 (theoretical) to 8.7% (actual calculated) primer extension increases. Therefore, this result supports the notion that the incorporation of the pure dGTP was not significantly affected by the addition of the extracted sample (2 l). Similar cases can be observed with the 2ϫ T cell samples as well as with the macrophage samples, when the data shown in Fig. 1F (with only extracted dNTP sample) and Fig. 1E (with mixtures of pure and extracted dNTPs) are compared. We also performed the poly(rA)/oligo(dT)-based RT assay measuring the [ 3 H]TTP incorporation in the presence or absence of 2 l of cellular dNTPs samples, and in this experiment also we did not observe any reduction of the [ 3 H]TTP incorporation by the cellular dNTP samples (macrophages, T cells, and HeLa cells; data not shown). Therefore, the data shown in Fig. 1E, together with the observations from the [ 3 H]TTP incorporation RT assay, clearly demonstrate that HIV-1 RT activity is either not affected or is only slightly affected by the inhibitory effects of other chemicals contained in the extracted dNTP samples. In fact, the protocol employed in this assay has been established previously for similar types of enzymatic dNTP assays (21). We also used several different volumes of the extracted dNTP samples for the dNTP measurements, and with these varied amounts of the samples, we observed the expected linear changes in the percent of primer extension. These data allow us to use 1) the percentages of the primer extension with the extracted cellular dNTP samples and 2) the dNTP standard curves shown in Fig. 1B for determining the dNTP content in each extracted sample.
Because there is no compartmentalization of cellular dNTPs between the nucleus and cytoplasm (1), the average cell volumes have been used to calculate the cellular dNTP concentrations. The dNTP concentrations in these three human primary cells were then calculated by taking into account the volumes of these cell types. The cell volumes of human macrophages and resting and activated CD4ϩ T cells were calculated using confocal microscopy and found to be 2660, 186, and 320 m 3 , respectively, which are similar to previously published estimates (34 -36). Therefore, the calculated dNTP concentrations in human macrophages are ϳ130 -250-fold lower (ϳ0.026 M) than in resting and activated CD4ϩ T cells (3.4 and 6.2 M, respectively; Fig. 1D).
Reverse Transcription Activity of Retroviral RTs at dNTP Concentrations Found in T Cells and Macrophages-Lentiviruses differ from oncoretroviruses in their ability to infect non-dividing cells such as macrophages (11,12). HIV-1 infection of terminally differentiated, tissue-resident cells of monocytic lineage may be especially important during viral trans- mission and early replication events (37,38). In contrast, like other RNA tumor viruses, MuLV (11) and AMV (39,40) infection requires cell proliferation. However, AMV infection of chicken embryo-derived macrophages has been described only when these cells were actively dividing as confirmed by [H 3 ]thymidine incorporation and mitotic index (39,40). Because the calculated dNTP concentration present in primary macrophages is very low (ϳ0.03 M), we performed a series of experiments employing various substrates and reaction condi-tions to examine whether lentiviral (HIV-1 and SIV) and oncoretroviral (MuLV and AMV) RTs were able to execute processive DNA synthesis under reaction conditions that contained dNTP concentrations present in macrophages (0.04 M) or T cells (4 M) (Fig. 2).
First, a 17-mer primer annealed to a 38-mer RNA template was employed for monitoring reverse transcription. The RT enzymes being compared were then added at two different quantities (4 and 1ϫ); as expected, these different levels of input enzyme resulted in an ϳ4-fold difference in the amount of fully extended primer (Fig. 2, A and B, lane F) at the dNTP concentration present in T cells (4 M). However, when the primer extension was performed in the presence of 0.04 M of each dNTP (i.e. the concentration present in macrophages), the two lentiviral RTs still maintained efficient DNA synthesis capabilities ( Fig. 2A), whereas the two oncoretroviral RTs showed a dramatic reduction in their ability to conduct multiple rounds of DNA synthesis (Fig. 2B). Note that slightly more reverse transcription activity was added for reactions with MuLV RT to demonstrate that even with more DNA polymerase activity, MuLV RT did not extend the primer efficiently at 0.04 M dNTP. In addition, reactions were performed with HIV-1 and MuLV RTs for longer incubation times with 0.04 M dNTPs, and we observed further extension of the primer, implying that all substrate dNTPs were not depleted under these reaction conditions (data not shown). Therefore, the low levels of primer extension seen for MuLV RT at 0.04 M dNTP concentrations are not because of dNTP depletion.
Second, we examined the effect of dNTP concentration on the DNA polymerization of HIV-1 and MuLV RTs under the condition allowing a single round of primer extension. This analysis employed the same reaction condition as described in the legend for Fig. 2, A and B (multiple round primer extension), except using a trap that is a 100-fold molar excess of cold T/P. In this reaction, HIV-1 and MuLV RTs (4ϫ concentration) were preincubated with the 32 P-labeled 17-mer primer annealed to the 38-mer RNA template, and then the RT reaction was initiated by adding a mixture of 0.04 M dNTP and the cold T/P trap. As shown in Fig. 2C, under this single round extension condition with 0.04 M dNTP, less primer extension was observed in the reaction with HIV-1 RT than with the multiple round primer extension by HIV-1 RT at 0.04 M (Fig. 2A). We also confirmed the single round primer extension by initiating the reaction containing RTs preincubated with both the cold trap and the 32 Plabeled T/P by adding 0.04 M dNTPs. In this trap control, no primer extension was observed (data not shown), indicating that the data shown in Fig. 2C was generated under the single round primer extension condition. Indeed, as shown in Fig. 2C, HIV-1 RT was still able to perform the single round of the primer extension very efficiently at 0.04 M dNTP compared with MuLV RT, which did not incorporate more than one nucleotide. This confirms the dNTP utilization discrepancy between HIV-1 RT and MuLV RT at low dNTP concentrations.
Third, we also used the multiple round primer extension assay using a 120-nucleotide-long HIV-1 RNA template annealed to a 32 P-labeled 20-mer primer (19). As shown in Fig.  2D, even with a 4-fold greater RT activity (8ϫ; see 4 M dNTP reactions), the two oncoretroviral RTs (M, MuLV RT; A, AMV RT) showed significant low primer extension at 0.04 M dNTP, compared with the two lentiviral RTs (H, HIV-1 RT; S, SIV RT). This result is similar to those described in Fig. 1, B and C, using the 38-mer RNA template.
Last, we examined the effect of the dNTP concentration on the DNA synthesis kinetics of HIV-1 and MuLV RTs, comparing the steady-state kinetic parameters of these two RTs measured by a single nucleotide incorporation assay. This experiment was performed on the 18-mer/19-mer template/primer used for the dNTP assay but with more limiting RT concentrations than those used in Fig. 1A. Fig. 2E demonstrates the qualitative difference between HIV-1 and MuLV RT in the single nucleotide incorporation with varying dNTP concentrations. Initially, RT activity showing ϳ60% extension at 50 M dTTP under the standard reaction condition described in Fig. 2A was determined, and then the same reactions were repeated at three different lower dTTP concentrations (5, 0.5, and 0.05 M). Indeed, as shown in Fig. 2E, at lower dNTP concentrations, HIV-1 RT was able to polymerize single nucleotide incorporation much more efficiently than MuLV RT. Next, the steady-state kinetic parameters K m and k cat were calculated under similar reaction conditions but with lower RT activity than was used in Fig. 2E. As seen in Fig. 2F, the k cat values for the two enzymes with the four dNTPs did not vary greatly, whereas there was a large increase in the K m values for the MuLV RT as compared with HIV-1 RT. Furthermore, in processive DNA synthesis assays similar to Fig. 2, A and B, we also found that the K m values of MuLV and AMV RTs were 23and 20-fold higher than that of HIV-1 RT, respectively, with similar k cat values between these RTs (data not shown). Therefore, these K m differences confirm that at lower dNTP concentrations, HIV-1 RT is able to polymerize DNA more efficiently than MuLV RT.
The data obtained from the various assays shown in Fig. 2 clearly suggest that the lentiviral RTs have evolved to efficiently catalyze DNA polymerization even at very low dNTP concentrations (such as those found in primary macrophages). In contrast, the oncoretroviral RTs may have adapted to efficiently catalyze DNA synthesis only at the much higher dNTP concentrations (i.e. 4 M) that are present in actively dividing cells.
HIV-1 Replication in Human Primary Target Cells with Different dNTP Concentrations-To further examine the relationship between cellular dNTP content and retroviral replication, we took advantage of two mutated derivatives of the HIV-1 RT, V148I and Q151N, that had recently been shown by pre-steadystate kinetic analysis to have severely reduced dNTP binding affinity (k d , 25.6-and 234.2-fold respectively) compared with wild-type HIV-1 RT, even though these mutants possess wildtype levels of catalysis (k pol ) (14,17). As seen in Fig. 3, we examined the dNTP concentration dependence of processive DNA polymerization by these mutant RTs as compared with wild-type RT. Each of the RTs exhibited similar levels of activity in reactions that contained high dNTP concentrations (i.e. 5 M in Fig. 3), which is supported by their identical k pol values. The wild-type HIV-1 RT was also able to efficiently polymerize DNA even at the low dNTP concentrations found in macrophages (Fig. 3A). However, similar to the two oncoretroviral RTs tested above (Fig. 2B), the two mutant HIV-1 RTs were significantly diminished in their ability to conduct processive DNA synthesis at the low dNTP concentrations found in macrophages. The dNTP binding to the active site of the mutant RTs likely becomes a rate-limiting step during the steady-state primer extension reaction at low dNTP concentrations. Overall DNA polymerization capability at the dNTP concentrations found in macrophages is greatest for wild-type RT, followed by V148I RT and then Q151N RT, which displayed the most significant decrease in dNTP binding affinity (14,17). Our kinetic analysis demonstrated that, like these HIV-1 dNTP binding mutants, MuLV RT also has a 40 -75-fold higher k d than wild-type HIV-1 (data not shown). Basically, these HIV-1 RT dNTP-binding mutants kinetically mimic MuLV (Fig. 2B).
Next, to examine the effect of RT dNTP binding affinity on lentiviral replication in primary cells, we generated pseudotyped HIV-1 vectors that contained an eGFP reporter cassette and that were isogenic except for the previously defined RT mutations (i.e. wild type, V148I, and Q151N). The pseudotyped particles were titered on HeLa cells, and then equivalent MOIs were used to infect primary human T cells and macrophages. As seen in Fig. 4, A and D, all three vectors had similar abilities to transduce activated human T cells and transformed cell lines (i.e. HeLa cells) containing high cellular dNTP concentrations, which is consistent with the efficient polymerization capability of wildtype RTs and these mutant HIV-1 RTs at high concentrations (Fig. 2C) and their similar k cat values (Fig. 2D). In contrast, the vectors containing the V148I and Q151N RT mutants were greatly diminished in their ability to transduce primary macrophages even with higher MOI infections (Fig. 4, B-D). Therefore, this reflects the enzymatic inability of these mutant RTs to synthesize DNA at low dNTP concentrations and suggests that the dNTP binding affinity of lentiviral RTs is a key contributor to viral replication in macrophages.
In summary, using both biochemical and virological analyses, we were able to investigate the relationship between cellular dNTP pools, RT binding affinity for dNTPs, and viral replication in non-dividing host primary cells that contain limiting dNTP concentrations. Our results strongly suggest that the enzymatic adaptation of the HIV-1 RT DNA polymerase active site to catalyze DNA synthesis in cellular environments that contain low levels of dNTPs is a major contributor to the unique ability of these viruses to efficiently replicate in macrophages. Therefore, this study provides evidence that RT plays an essential role in establishing and maintaining the macrophage tropism of HIV-1 during the course of viral infection, which presents unique virological and pathological characteristics that differentiate lentiviruses from oncoretroviruses.