Restriction of HIV-1 Requires the N-Terminal Region of MxB as a Capsid-Binding Motif but Not as a Nuclear Localization Signal

ABSTRACT

The interferon alpha (IFN-α)-inducible restriction factor MxB blocks HIV-1 infection after reverse transcription but prior to integration. Fate-of-capsid experiments have correlated the ability of MxB to block HIV-1 infection with stabilization of viral cores during infection. We previously demonstrated that HIV-1 restriction by MxB requires capsid binding and oligomerization. Deletion and gain-of-function experiments have mapped the HIV-1 restriction ability of MxB to its N-terminal 25 amino acids. This report reveals that the N-terminal 25 amino acids of MxB exhibit two separate functions: (i) the ability of MxB to bind to HIV-1 capsid and (ii) the nuclear localization signal of MxB, which is important for the ability of MxB to shuttle into the nucleus. To understand whether MxB restriction of HIV-1 requires capsid binding and/or nuclear localization, we genetically separated these two functions and evaluated their contributions to restriction. Our experiments demonstrated that the ¹¹RRR¹³ motif is important for the ability of MxB to bind capsid and to restrict HIV-1 infection. These experiments suggested that capsid binding is necessary for the ability of MxB to block HIV-1 infection. Separately from the capsid binding function of MxB, we found that residues ²⁰KY²¹ regulate the ability of the N-terminal 25 amino acids of MxB to function as a nuclear localization signal; however, the ability of the N-terminal 25 amino acids to function as a nuclear localization signal was not required for restriction.

IMPORTANCE MxB/Mx2 blocks HIV-1 infection in cells from the immune system. MxB blocks infection by preventing the uncoating process of HIV-1. The ability of MxB to block HIV-1 infection requires that MxB binds to the HIV-1 core by using its N-terminal domain. The present study shows that MxB uses residues ¹¹RRR¹³ to bind to the HIV-1 core during infection and that these residues are required for the ability of MxB to block HIV-1 infection. We also found that residues ²⁰KY²¹ constitute a nuclear localization signal that is not required for the ability of MxB to block HIV-1 infection.

INTRODUCTION

Myxovirus resistance proteins represent a family of interferon (IFN)-inducible factors with a wide range of antiviral activities (1–3). The myxovirus B (MxB) gene was originally cloned from a human glioblastoma cell line treated with interferon alpha (IFN-α) (4, 5). MxB, as well as the related protein MxA, belongs to the dynamin-like family of proteins, which have diverse functions ranging from vesicle transport to antiviral activity (1, 6–11). The most studied dynamin-like protein that exhibits antiviral activity is MxA (1, 2). Contrary to MxB, the antiviral role of MxA has been extensively studied for viruses, including the influenza (1, 12–15), tick-borne Thogoto (16), African swine fever (17), hepatitis B (18), and La Crosse (19, 20) viruses. The antiviral activity of the long form of MxB was recently described (9, 21–23); these investigations led to the discovery that the IFN-α-inducible protein MxB blocks HIV-1 infection.

Genetic evidence has suggested that HIV-1 capsid is the determinant for the ability of MxB to block HIV-1 infection (9, 22, 23). In agreement with these findings, we have recently demonstrated that MxB binds to the HIV-1 capsid and have correlated the ability of MxB to block HIV-1 infection with inhibition of uncoating (24). We also showed that the ability of MxB to block infection requires a capsid binding domain and an oligomerization domain provided by the 90 N-terminal and 143 C-terminal amino acids of MxB, respectively (24). In addition, the work of others and our work showed that the 90 N-terminal amino acids of MxB are important for its ability to bind capsid and restrict HIV-1 infection (24–26).

MxB contains a previously described putative nuclear localization signal on its N-terminal 25 amino acids (4). Deletion of the N-terminal 25 amino acids annihilates the ability of MxB to block HIV-1 infection and to bind to the HIV-1 core (23, 24, 27). To understand the contributions of capsid binding and nuclear localization to the ability of MxB to block HIV-1 infection, we generated a series of MxB variants in the N-terminal 25 residues of MxB. The different variants were tested for capsid binding, nuclear localization, and restriction. These results genetically mapped the capsid-binding ability of MxB to the protein motif ¹¹RRR¹³. In addition, we mapped the nuclear localization signal of the N-terminal 25 amino acids to residue K20. Overall, our results showed that HIV-1 restriction by MxB requires the capsid binding motif ¹¹RRR¹³ but not the nuclear localization signal provided by the N-terminal 25 amino acids of the protein.

MATERIALS AND METHODS

Binding of MxB variants to in vitro-assembled HIV-1 CA-NC complexes.

293T cells were transfected with plasmids expressing wild-type (WT) or mutant MxB proteins. Forty-eight hours after transfection, cell lysates were prepared as follows. Previously washed cells were resuspended in hypotonic lysis buffer (10 mM Tris [pH 7.4], 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM dithiothreitol [DTT]). The cell suspension was frozen and thawed and incubated on ice for 10 min. Afterwards the lysate was centrifuged at maximum speed in a refrigerated Eppendorf microcentrifuge (∼14,000 × g) for 5 min. The supernatant was supplemented with 1/10 volume of 10× phosphate-buffered saline (PBS) and then used in the binding assay. In some cases, samples containing the MxB variants were diluted with extracts prepared in parallel from untransfected cells. To test binding, 5 μl of capsid-nucleocapsid (CA-NC) complex particles assembled in vitro was incubated with 200 μl of cell lysate at room temperature for 1 h (28, 29). A fraction of this mixture was stored (input). The mixture was spun through a sucrose cushion (70% sucrose, 1× PBS, 0.5 mM DTT) at 100,000 × g in an SW55 rotor (Beckman) for 1 h at 4°C. After centrifugation, the supernatant was carefully removed, and the pellet was resuspended in 1× SDS-PAGE loading buffer (pellet). The level of MxB proteins was determined by Western blotting using anti-FLAG antibodies. The level of HIV-1 CA-NC protein in the pellet was assessed by Western blotting using anti-p24 CA antibodies.

Creation of cells stably expressing wild-type and mutant MxB proteins.

Retroviral vectors encoding wild-type or mutant human MxB proteins were created using the LPCX or pRetroX-tet-one vectors (Clontech). The experiments throughout this article were performed using the LPCX vector, with the exception of the experiment shown in Fig. 6, where the pRetroX-tet-one vector was used. The MxB proteins contained a FLAG epitope tag at the C terminus. Recombinant viruses were produced in 293T cells by cotransfecting the pRetroX-tet-one plasmids with the pVPack-GP and pVPack-VSV-G packaging plasmids (Stratagene). The pVPack-VSV-G plasmid encodes the vesicular stomatitis virus (VSV) G envelope glycoprotein, allowing efficient entry into a wide range of vertebrate cells. Cf2Th canine thymocytes were transduced and selected in puromycin (Sigma).

FIG 6.

The ¹¹RRR¹³ motif but not ²⁰KY²¹ is required for the ability of MxB to block HIV-1 infection. (A) Wild-type and mutant MxB proteins were stably expressed in Cf2Th cells using the doxycycline-inducible pRetroX-tet-one vector. Protein induction was achieved using 100 ng/ml of doxycycline. Expression levels in whole-cell extracts were measured by Western blotting using anti-FLAG antibodies. Protein load was assayed by Western blotting using anti-glyceraldehyde-3-phosphate dehydrogenase (anti-GAPDH) antibodies. (B) Cf2Th cells stably expressing wild-type and mutant MxB proteins were challenged with increasing amounts of HIV-1-GFP. Forty-eight hours postinfection, the percentage of GFP-positive cells was measured by flow cytometry. As a control, Cf2Th cells stably transduced with the empty vector pRetroX were challenged with HIV-1. Similar results were obtained in three independent experiments, and a representative experiment is shown.

Infection with viruses expressing GFP.

Recombinant HIV-1 expressing green fluorescent protein (GFP) was prepared as described previously (30). Recombinant viruses were pseudotyped with the VSV-G glycoprotein. For infections, 24-well plates were seeded with 3 × 10⁴ Cf2Th cells stably expressing the different MxB variants and incubated at 37°C with virus for 24 h. Cells were washed and returned to culture for 48 h. Subsequently, GFP-positive cells were analyzed using a flow cytometer (Becton Dickinson). Luciferase-expressing HIV-1 (HIV-1-Luc) viruses pseudotyped by VSV-G were prepared by cotransfecting NL4.3 Luc env⁻ with the VSV-G envelope.

Indirect immunofluorescence microscopy.

A total of 3 × 10⁴ HeLa cells per well were plated in complete medium (Dulbecco's minimal essential medium [DMEM] plus 10% [vol/vol] fetal bovine serum [FBS] plus penicillin-streptomycin) for 18 h on 12-mm-diameter cover glasses inside 24-well plates. Each well was transfected with 300 ng of the corresponding DNA in serum-free DMEM for 6 h. Subsequently, the medium was replaced by complete medium, and expression was monitored 18 h posttransfection. In the case of transfection with FLAG-tagged constructs, immunofluorescence staining was performed with rabbit anti-FLAG antibody and Cy3-conujugated anti-rabbit antibody, as shown previously (31). FLAG-tagged proteins were stained using the rabbit anti-FLAG antibody (polyclonal, affinity-purified; Sigma-Aldrich, F7425) at a dilution of 1:250, and Cy3-conjugated anti-rabbit antibodies (affinity-purified IgG; Jackson ImmunoResearch, 711-165-152), diluted 1:1,000. Before mounting, cells were stained with DAPI (4′,6-diamidino-2-phenylindole) as previously described (31). When indicated, leptomycin B (Sigma) was used at 20 to 40 ng/ml for 6 h.

Luciferase assays.

Infectivity of HIV-1-Luc was measured by luciferase (Promega) activity 48 h postinfection according to the manufacturer's instructions (Victor, PerkinElmer). Lysates were normalized by protein content by using a Bio-Rad protein assay.

Cell fractionation of nucleus cytoplasm.

Cells were lysed with hypotonic buffer (10 mM Tris [pH 8], 10 mM KCl, 1.5 mM MgCl₂, 1 mM DTT) containing 0.1% Triton X-100, which does not destroy nuclei, and incubated on ice for 15 min. The lysate was centrifuged at 5,000 rpm for 5 min at 4°C, and the supernatant as cytoplasmic fraction was collected. The nuclear pellet was washed with 1 ml hypotonic buffer without Triton X-100 twice and centrifuged at 13,000 rpm for 5 min at 4°C. The DNA was isolated from nuclei using the QIAamp DNA microkit (Qiagen).

Real-time PCR to detect HIV-1 late reverse transcripts and the formation of 2-LTR circles.

Total cellular DNA from infected cells was isolated using the QIAamp DNA microkit (Qiagen) at 7, 24, and 48 h postinfection (hpi). Before infection, viruses were treated for 30 min at 37°C with 1,000 U of DNase I (Roche). As a control of the experiment, we performed infection in the presence of 5 μM nevirapine. We measured late reverse transcripts using the following primers and probe: forward primer, MH531 (TGTGTGCCCGTCTGTTGTGT); reverse primer, MH532 (GAGTCCTGCGTCGAGAGAGC); and probe, LRT-P (6-carboxyfluorescein [FAM]–CAGTGGCGCCCGAACAGGGA–6-carboxytetramethylrhodamine [TAMRA]). Two-long-terminal-repeat (2-LTR)-containing circles were detected with primers MH535/536 and probe MH603, and a standard curve based on the pUC2LTR plasmid, which contains the HIV-1 2-LTR junction, was used (32). β-Actin amplification was used for normalization using the following primers and probes for actin: forward primer 50-AACACCCCAGCCATGTACGT, reverse primer 50-CGGTGAGGATCTTCATGAGGTAGT, and probe FAM-CCAGCCAGGTCCAGACGCAGGA-BlackBerry quencher (BBQ). Reaction mixtures contained 1× Fast Start Universal Probe master mix, 2× Rox (Roche), 300 nM forward primer, 300 nM reverse primer, 100 nM probe primer, and template DNA. Denaturation steps (95°C for 15 min) with 40 cycles of amplification were carried out (95°C for 15 s, 58°C for 30 s, and 72°C for 30s).

RESULTS

The N-terminal 25 amino acids of MxB are important for capsid binding and nuclear localization.

Melen and colleagues suggested that the N-terminal domain of MxB contains a putative nuclear localization signal (4) (Fig. 1). The ability of MxB to restrict HIV-1 requires an intact N-terminal 25 amino acids (23, 24, 27). Interestingly, the ability of MxB to bind capsid also maps to the N-terminal residues of MxB (24). These observations suggested that the N-terminal residues of MxB might contribute to two separate functions, capsid binding and nuclear localization. To understand the role of these two functions in the ability of MxB to block HIV-1, we generated a series of MxB variants to test for capsid binding, nuclear localization, and HIV-1 restriction (Fig. 1 and Table 1).

FIG 1.

The N-terminal region of MxB. The underlined residues constitute a putative tripartite nuclear localization signal.

TABLE 1.

Phenotypes of the MxB variants in this study

MxB variant	MxB		Preferential subcellular localization of MxB(1–25)-SAMHD1(15–626) variantsc
	% of HIV-1 restrictiona	Relative binding to HIV-1 CA-NC tubesb
WT	100	1	Cytosolic + nuclear
P9A	NDd	0.95 ± 0.08	Cytosolic + nuclear
Y10A	ND	0.89 ± 0.11	Cytosolic + nuclear
R11A	3 ± 0.39	0.00 ± 0.00	Cytosolic + nuclear
R12A	11 ± 3.37	0.14 ± 0.07	Cytosolic + nuclear
R13A	1 ± 0.36	0.23 ± 0.10	Cytosolic + nuclear
R11A R12A	5 ± 0.45	0.00 ± 0.00	ND
R11A R12A R13A	2 ± 0.30	0.00 ± 0.00	ND
R19A	91 ± 0.39	1.04 ± 0.01	Cytosolic + nuclear
K20A	100 ± 5.16	1.05 ± 0.26	Cytosolic
K20D	ND	ND	Cytosolic
K20E	ND	ND	Cytosolic
R19A K20A	76 ± 9.01	0.39 ± 0.29	Cytosolic
Y21A	ND	ND	Cytosolic
K23A	ND	1.18 ± 0.11	Cytosolic + nuclear
K24A	ND	1.13 ± 0.05	Cytosolic + nuclear

^aRestriction was measured by infecting cells stably expressing the indicated MxB variant with HIV-1-GFP. After 48 h, the percentage of GFP-positive cells (infected cells) was determined by flow cytometry. Restriction is expressed as the percentage of HIV-1 restriction by the MxB variants relative to HIV-1 restriction by wild-type MxB. The mean ± SD from three independent experiments is shown.

^bBinding to in vitro-assembled HIV-1 CA-NC complexes was determined for each MxB variant as described in Materials and Methods. The binding is expressed as the bound fraction relative to input normalized to the binding of wild-type MxB. The mean ± SD from three experiments is shown.

^cSubcellular localization of the different MxB(1–25)-SAMHD1(15–626) variants was measured by analyzing the localization of the variants in 200 cells (Fig. 4C). Every cell was characterized for nuclear or nuclear plus cytosolic localization.

^dND, not determined.

The ability of MxB to bind capsid maps to residues ¹¹RRR¹³.

We previously demonstrated that the variant MxB-Δ(1–20) loses the ability to block HIV-1 and to bind capsid (24). To find determinants for the ability of MxB to bind capsid, we tested the ability of variants in the first 25 amino acids of MxB to bind in vitro-assembled HIV-1 capsid-nucleocapsid (CA-NC) complexes (Fig. 2 and Table 1). Our investigations revealed that the sequence ¹¹RRR¹³ is important for the ability of MxB to interact with capsid (Fig. 2). On the contrary, changing residues R19 or K20 did not affect the ability of MxB to bind to in vitro-assembled HIV-1 CA-NC complexes. These results showed that the motif ¹¹RRR¹³ contributes to the ability of MxB to bind in vitro-assembled HIV-1 CA-NC complexes.

FIG 2.

An intact ¹¹RRR¹³ motif is required for the ability of MxB to bind to in vitro-assembled HIV-1 CA-NC complexes. The ability of MxB variants to bind in vitro-assembled HIV-1 CA-NC complexes was measured. 293T cells were transfected with plasmids expressing wild-type or mutant FLAG-tagged MxB proteins. Twenty-four hours posttransfection, cells were lysed. Subsequently, the lysates were incubated at room temperature for 1 h with in vitro-assembled HIV-1 CA-NC complexes. The mixtures were applied to a 70% (wt/vol) sucrose cushion and centrifuged. “Input” represents the lysates analyzed by Western blotting before being applied to the 70% cushion. The input mixtures were Western blotted using anti-FLAG antibodies. The pellet from the 70% sucrose cushion (“Bound”) was analyzed by Western blotting using anti-FLAG or anti-p24 antibodies. Similar results were obtained in three independent experiments, and the standard deviation (SD) for the bound fraction relative to input, and normalized to binding of wild-type MxB is shown.

Residues ²⁰KY²¹ control the ability of the N-terminal domain of MxB to function as a nuclear localization signal.

Recent indirect immunofluorescence studies have shown that MxB exhibits a punctate pattern that extends from the cytosol to the perinuclear region of the cell (24–27, 33). To test whether human MxB shuttles between the cytosol and the nucleus, we treated human cells expressing MxB with the drug leptomycin B, which blocks nuclear export by interfering with the CRM1-mediated mechanism (34–37). Immunofluorescence experiments revealed that treatment of cells expressing MxB with the nuclear export inhibitor leptomycin B accumulated the protein in the nuclear compartment (Fig. 3A and B). As a control, we treated cells expressing MxA with leptomycin B and showed that MxA is not accumulated in the nuclear compartment. Although the steady-state localization of MxB is cytosolic, these experiments suggested that the protein is shuttling in and out of the nucleus. By using leptomycin B, we also tested the shuttling ability of the protein chimera MxB(1–90)-MxA(43–662), which is known to block HIV-1 infection (24, 25). Interestingly, the protein chimera MxB(1–90)-MxA(43–662) gained the ability to shuttle in and out of the nucleus compared to MxA.

FIG 3.

MxB but not MxA shuttles in and out of the nucleus. (A) Human HeLa cells expressing MxB, MxA, or MxB(1–90)-MxA(43–662) were treated with 20 ng/ml of leptomycin B or dimethyl sulfoxide (DMSO) (solvent) for 6 h. Treated cells were fixed and stained using anti-FLAG antibodies (red) as described in Materials and Methods. Cellular nuclei were stained by using DAPI (blue). Representative images are shown. (B) Cellular localization was determined visually for 200 randomly selected cells using a 40× objective. The results of three independent experiments with standard deviations are shown. (C) Cf2Th cells stably expressing MxB or containing the empty vector were challenged with increasing amounts of HIV-1-GFP in the presence of leptomycin B or DMSO. Similar results were obtained in three independent experiments, and a representative experiment is shown.

These experiments demonstrated that leptomycin B affects the cellular localization of MxB; however, it did not affect the ability of MxB to block HIV-1 infection (Fig. 3C).

To understand the contribution of the N-terminal 25 amino acids of MxB to nuclear import, we fused the N-terminal amino acids 1 to 10, 1 to 15, and 1 to 25 of MxB to residues 15 to 626 of SAMHD1 [SAMHD1(15–626)] (Fig. 4A). The protein SAMHD1(15–626) does not contain a nuclear localization signal (SAMHD1-ΔNLS) and localizes to the cytosol of the cell (Fig. 4B and C), as previously shown (31, 38, 39). As shown in Fig. 4B and C, the fusion construct MxB(1–25)-SAMHD1(15–626)-FLAG localized to the nuclear compartment, suggesting that the first 25 amino acids of MxB contain the necessary information to allow the transport of the protein SAMHD1(15–626) to the nucleus. On the contrary, proteins MxB(1–10)-SAMHD1(15–626) and MxB(1–15)-SAMHD1(15–626) did not localize to the nuclear compartment compared to MxB(1–25)-SAMHD1(15–626)-FLAG (Fig. 4B and C). As expected, full-length SAMHD1 localized to the nuclear compartment (31, 38–40). Fusion of the N-terminal 1 to 10, 1 to 15, and 1 to 25 amino acids of MxB to the green fluorescent protein showed similar results (data not shown). These results showed that the N-terminal 25 amino acids are a functional nuclear localization signal.

FIG 4.

Residues ²⁰KY²¹ regulate the ability of the N-terminal region of MxB to function as a nuclear localization signal. (A) The different constructs used to analyze the ability of the N-terminal 25 amino acids to work as nuclear localization signal are shown. MxB residues 1 to 10, 1 to 15, and 1 to 25 were fused to SAMHD1 residues 15 to 626. The fragment SAMHD1(15–626) does not contain the nuclear localization signal of SAMHD1 and is also referred to as SAMHD1-ΔNLS, which preferentially localizes to the cytoplasm. (B) Human HeLa cells expressing the indicated variants were fixed and stained using anti-FLAG antibodies (red) as described in Materials and Methods. Cellular nuclei were stained by using DAPI (blue). Representative images are shown. (C) Cellular localization was determined visually for 200 randomly selected cells using a 40× objective. Each cell was examined for cytosolic, nuclear, or nuclear plus cytosolic localization. The results from three independent experiments with standard deviations are shown. Similar results were obtained in three independent experiments, and a representative experiment is shown.

Next we decided to explore the contribution of the different residues of the N-terminal 25 amino acids of MxB to mediate nuclear import (Fig. 4B and C and Table 1). Interestingly, we found that disruption of K20 or Y21 affected the ability of the N-terminal 25 amino acids of MxB to promote nuclear import (Fig. 4B and C and Table 1). In summary, variants K20A, K20D, K20E, R19A K20A, and Y21A disrupted the ability of the N-terminal 25 amino acids of MxB to promote nuclear import (Fig. 4B and C and Table 1). In contrast, disruption of residues P9, Y10, R11, R12, R13, R19, K23, and K24 did not change the ability of N-terminal 25 amino acids of MxB to promote nuclear import (Fig. 4B and C and Table 1). Overall, our results suggested that residues ²⁰KY²¹ control the ability of the N-terminal 25 amino acids of MxB to function as a nuclear localization signal.

We have previously suggested that HIV-1 restriction by MxB requires a capsid binding motif and an oligomerization motif (24, 41). To further test this hypothesis, we tested the ability of MxB(1–25)-SAMHD1(15–626)-FLAG to bind in vitro-assembled HIV-1 CA-NC complexes and restrict HIV-1 infection. As shown in Fig. 5, MxB(1–25)-SAMHD1(15–626)-FLAG binds capsid and restricts HIV-1 infection. On the contrary, MxB(1–25)-R11A-SAMHD1(15–626)-FLAG does not bind capsid or restrict HIV-1 infection. These experiments reinforce the notion that HIV-1 restriction by MxB requires capsid binding and an oligomerization motif.

FIG 5.

MxB(1–25)-SAMHD1(15–626) binds to capsid and restricts HIV-1 infection. The ability of MxB(1–25)-SAMHD1(15–626)-FLAG and MxB(1–25)-R11A-SAMHD1(15–626)-FLAG to bind in vitro-assembled HIV-1 CA-NC complexes was measured as described in Materials and Methods (A). Cf2Th cells stably expressing MxB(1–25)-SAMHD1(15–626) and MxB(1–25)-R11A-SAMHD1(15–626) (B) were challenged with increasing amounts of HIV-1-GFP. Forty-eight hours postinfection, the percentage of GFP-positive cells was measured by flow cytometry (C). As a control, Cf2Th cells stably transduced with the empty vector LPCX were challenged with HIV-1. Similar results were obtained in three independent experiments, and a representative experiment is shown.

Changes on the ¹¹RRR¹³ motif disrupted the ability of MxB to block HIV-1 infection.

To understand the contribution of the N-terminal 25 amino acids of MxB to HIV-1 restriction, we tested the abilities of the different MxB variants to block HIV-1 infection. To this end, we stably expressed the different MxB variants in canine Cf2Th cells (Fig. 6A) and challenged them with increasing amounts of HIV-1 expressing the green fluorescent protein (HIV-1-GFP) as a reporter of infection. As shown in Fig. 6B, the ability of MxB to block HIV-1 infection requires an intact ¹¹RRR¹³ motif. Interestingly, changes on the ¹¹RRR¹³ motif also affected the ability of MxB to interact with in vitro-assembled HIV-1 CA-NC complexes (Fig. 2 and Table 1). These experiments narrowed down the interaction of MxB with HIV-1 capsid to the ¹¹RRR¹³ motif. On the contrary, disruption of K20 did not affect the ability of MxB to block HIV-1 infection, suggesting that the ability of the N-terminal 25 amino acids to function as a nuclear localization signal is not necessary for restriction.

Cellular localization and shuttling ability of the different MxB variants.

To evaluate the effect of N-terminal changes to the cellular distribution of MxB, we studied the localization of the different MxB variants by indirect immunofluorescence microscopy in HeLa cells. As shown Fig. 7A and B, MxB variants R11A, R11A R12A, R11A R12A R13A, and R19A K20A exhibit a more diffuse localization than wild-type MxB, which showed a punctate localization that extends from the cytosol to the perinuclear structures of the cell (Fig. 7A and B). All of the other variants studied here showed a similar localization to wild-type MxB (Fig. 7A and B). We tested the abilities of the different MxB variants to shuttle in and out of the nucleus by using leptomycin B. As shown in Fig. 7B, all of the studied MxB variants shuttle in and out of the nucleus, suggesting that MxB contains a separate nuclear localization signal from ²⁰KY²¹. Similar results were observed in the canine cells Cf2Th (data not shown).

FIG 7.

Cellular localization of the different MxB variants. (A) Human HeLa cells expressing the indicated MxB variant were fixed and stained using anti-FLAG antibodies (red) as described in Materials and Methods. Cellular nuclei were stained using DAPI (blue). Representative images are shown. (B) Cellular localization was determined visually for 200 randomly selected cells using a 40× objective. Each cell was examined for cytosolic or cytosolic plus nuclear localization. The results from three independent experiments with standard deviations are shown.

MxB blocks HIV-1 infection after reverse transcription but prior to formation of 2-LTR circles.

Our results demonstrated that MxB has the ability to shuttle into and out of the nuclear compartment. Previous investigations have shown that MxB blocks HIV-1 after reverse transcription but before integration (9, 22, 23, 27). To understand whether the block imposed by MxB to HIV-1 infection is before or after nuclear import, we measured the production of HIV-1 2-LTR circles during HIV-1 infection of cells stably expressing MxB. The generation of 2-LTR circles during HIV-1 infection is an indirect measure that the preintegration complex has migrated to the nuclear compartment (32). To this end, we challenged cells stably expressing MxB or containing the empty vector LPCX with HIV-1 expressing luciferase (HIV-1-Luc) as a reporter of infection and measured the HIV-1 2-LTR circles 24 h postinfection. As shown on Fig. 8A, 2-LTR circles were reduced or not generated in cells stably expressing MxB compared to those in cells containing the empty vector LPCX. As a control, we performed similar experiments in the presence of the reverse transcription inhibitor nevirapine. Because reverse transcription, measured as production of late reverse transcripts 7 h postinfection, was similar for both cells stably expressing MxB and those containing the empty vector LPCX, our results indicated that MxB blocks HIV-1 infection after reverse transcription but prior to nuclear import.

FIG 8.

MxB blocks HIV-1 infection after reverse transcription but prior to nuclear import. (A) Cf2Th cells stably expressing MxB or containing the empty vector LPCX were challenged with DNase-pretreated HIV-1 viruses containing luciferase as a reporter for infection (HIV-1-Luc) at a multiplicity of infection (MOI) of <0.5. Infection was determined by measuring luciferase relative light units (RLU) 48 h postinfection (left panel). In parallel, cells from similar infections were lysed at 7 and 24 h postinfection, and total DNA was extracted. The DNA samples collected at 7 h postinfection were used to determine the levels of late reverse transcripts by real-time PCR (middle panel). HIV-1 2-LTR circles were quantified by real-time PCR in DNA samples collected at 24 h postinfection (right panel). As a control, we performed similar infections in the presence of the reverse transcription inhibitor nevirapine (Nev). Similar results were obtained in three independent experiments, and the standard deviation is shown. (B) Similarly, Cf2Th cells stably expressing MxB or containing the empty vector LPCX were challenged with DNase-pretreated HIV-1-Luc viruses. Infection was determined by measuring luciferase RLU 48 h postinfection (left panel). In parallel, cells from similar infections were harvested 7, 24, and 48 h postinfection. Harvested cells were separated into cytoplasmic and nuclear fractions as described in Materials and Methods. Extracted DNA from cytoplasmic and nuclear fractions was used to determine the levels of HIV-1 2-LTR circles by real-time PCR as described in Materials and Methods. To control for the bona fide origin of nuclear extracts, we measured the levels of the actin DNA in extracts, which is only enriched in nuclear fractions. We also performed infections in the presence of the reverse transcription inhibitor nevirapine. Similar results were obtained in three independent experiments, and the standard deviation is shown. (C) Similar experiments were performed to measure the formation of HIV-1 2-LTR circles during infection of Cf2Th cells stably expressing MxB in the presence of leptomycin B. Similar results were obtained in three independent experiments, and the standard deviation is shown.

To gain a deeper understanding of the viral life cycle step at which MxB blocks HIV-1 infection, we measured the production of HIV-1 2-LTR circles in cells biochemically fractionated into nuclear and cytoplasmic fractions 7, 24, and 48 h postinfection (Fig. 8B). Measurement of HIV-1 2-LTR circles on the nuclear fraction revealed that cells stably expressing MxB blocked the production of 2-LTR circles, suggesting that MxB blocks HIV-1 infection after reverse transcription but prior to nuclear import. As a control to differentiate cytoplasmic from nuclear fractions, we measured the levels of the actin gene, which is only enriched in nuclear fractions (Fig. 8B). Because leptomycin B affected the subcellular localization of MxB, we tested the ability of leptomycin B to affect the formation of HIV-1 2-LTR circles during infection. As shown in Fig. 8C, treatment with leptomycin B did not affect HIV-1 infection and the formation of 2-LTR circles.

DISCUSSION

Our previous observations suggested that the ability of MxB to block HIV-1 infection requires oligomerization and capsid binding (24). Deletion experiments showed that the N-terminal 25 amino acids of MxB are important for the ability of MxB to bind HIV-1 capsid (24). In agreement with the notion that the ability of MxB to bind HIV-1 capsid is important for restriction, deletion of the N-terminal ∼25 amino acids resulted in loss of restriction (23, 24, 27). Overall these results suggested that these 25 amino acids are important for capsid binding and restriction. This work tested the hypothesis that the N-terminal 25 amino acids of MxB contained determinants for the ability of MxB to bind capsid. Accordingly, this work showed that the motif ¹¹RRR¹³ is essential for the ability of MxB to bind to capsid. In agreement with these results, mutations in the ¹¹RRR¹³ motif resulted in loss of restriction. These results implied that the ¹¹RRR¹³ motif is likely to be interacting with the surface of the HIV-1 core. During the preparation of this article, Goujon et al. reported that disruption of residues ¹¹RRR¹³ affected the ability of MxB to block HIV-1 infection (42).

The fact that variants in the ¹¹RRR¹³ motif simultaneously lost capsid binding and restriction implies that capsid binding is necessary for restriction. This type of evidence was lacking in the field since most of the capsid mutants tested in the capsid binding assay only partially overcome restriction (24). Future experiments will identify capsid mutants that completely overcome the restriction imposed by MxB to HIV-1 infection; identification of such mutants will allow testing of capsid binding, which will provide additional evidence for capsid as the determinant for MxB restriction. Overall, these findings are in agreement with a mechanism in which MxB binds directly to the HIV-1 core and prevents the uncoating process of HIV-1 (24, 41, 43).

MxB has the ability to shuttle in and out of the nucleus. Experiments in which MxB-expressing cells were treated with leptomycin B, which blocks the CRM1-dependent nuclear-export pathway, showed accumulation of MxB in the nuclear compartment. Although the steady-state localization of MxB showed a punctate pattern that extends from the cytosol to the perinuclear region of the cell, these experiments suggested that MxB has the ability to shuttle in and out of the nucleus, which might be important for its ability to block HIV-1 infection.

The N-terminal 25 amino acids of MxB are a functional nuclear localization signal. Melen and colleagues, using N-terminal deletions of MxB, suggested that the N-terminal 25 amino acids of MxB constitute a nuclear localization signal (4). To functionally test the ability of the N-terminal 25 amino acids of MxB to function as a nuclear localization signal, we fused the N-terminal 25 amino acids to the fragment SAMHD1(15–626), which localizes to the cytoplasm of the cell. Remarkably, the fusion MxB(1–25)-SAMHD1(15–626) localized to the nuclear compartment, suggesting that the N-terminal 25 amino acids of MxB are a functional nuclear localization signal. Using this construct as a reporter for nuclear localization, we mapped the nuclear localization signal of MxB to ²⁰KY²¹. Because our studies did not encompass mutagenesis of the entire region, we could not conclude whether ²⁰KY²¹ alone or additional residues are involved in the nuclear localization signal function. However, this provided a valuable tool to understand whether the nuclear localization signal of MxB is required for restriction.

The nuclear localization signal of the N-terminal 25 amino acids of MxB is not required for the ability of MxB to block HIV-1 infection. The fact that the variant MxB-K20A blocks HIV-1 infection suggested that the nuclear localization signal of the N-terminal 25 amino acids is not required for restriction. However, these experiments do not discard the possibility that a second nuclear localization signal located in a different region of MxB is important for the ability of MxB to block HIV-1 infection. In agreement with this notion, variants that inhibit the nuclear import of MxB(1–25)-SAMHD1(15–626) did not abrogate shuttling when they were introduced in the full-length MxB. Future investigations will provide information on whether the shuttling of MxB into the nucleus is important for restriction.

MxB blocks HIV-1 infection after reverse transcription but prior to nuclear import. To understand whether MxB blocks HIV-1 infection before or after nuclear import, we decided to measure formation of 2-LTR circles in whole-cell extracts or in biochemically fractionated cells. In agreement with the hypothesis that MxB blocks HIV-1 infection before nuclear import, we found that MxB prevented the formation of HIV-1 2-LTR circles, which are an indirect measure of nuclear import (32).

Our investigations revealed that the N-terminal 25 amino acids of MxB exhibit two functions: (i) the ¹¹RRR¹³ motif is required for the ability of MxB to bind capsid and restrict HIV-1 infection, and (ii) the nuclear localization signal of the N-terminal 25 amino acids of MxB is not required for the ability of MxB to block HIV-1 infection. An important contribution to the MxB field is the fact that this work reports single mutations that lose capsid binding and restriction, reinforcing the notion that capsid binding is required for restriction.