Effect of Different Nuclear Localization Signals on the Subcellular Localization and Anti-HIV-1 Function of the MxB Protein

Interferon exerts its antiviral activity by stimulating the expression of antiviral proteins. These interferon stimulate genes (ISGs) often target a group of viruses with unique molecular mechanisms. One such ISG is myxovirus resistance B (MxB) that has been reported to inhibit human immunodeficiency virus type 1 (HIV-1) by targeting viral capsid and impairing nuclear import of viral DNA. The antiviral specificity of MxB is determined by its N-terminal 25 amino acids sequence which has the nuclear localization activity, therefore functions as a nuclear localization signal (NLS). In this study, we report that the bipartite NLS, but not the classic NLS, the PY-NLS, nor the arginine-rich NLS, when used to replace the N-terminal sequence of MxB, drastically suppress HIV-1 gene expression and virus production, thus creates a new anti-HIV-1 mechanism. MxB preserves its anti-HIV-1 activity when its N-terminal sequence is replaced by the arginine-rich NLS. Interestingly, the arginine-rich NLS allows MxB to inhibit HIV-1 CA mutants that are otherwise resistant to wild type MxB, which suggests sequence specific targeting of viral capsid. Together, these data implicate that it is not the nuclear import function itself, but rather the sequence and the mechanism of action of the NLS which define the antiviral property of MxB.


INTRODUCTION
Interferons are produced in response to viral infections, and provide one main mechanism of host innate antiviral defense (Bonjardim, 2005;Randall and Goodbourn, 2008). Interferons operate by inducing the expression of hundreds of genes, collectively called interferon-stimulated genes (ISGs), many of which have direct antiviral activities. The myxovirus resistance (Mx) genes are typical ISGs, and were discovered for their protection of mice from lethal infection by influenza virus (Staeheli et al., 1986;Pavlovic et al., 1992).
In addition to its interaction with HIV-1 capsid core in concert with the GTPase and the stalk domains, the NTD of MxB also has the nuclear localization function Goujon et al., 2015;Kane et al., 2018). This NTD, when attached to the heterologous proteins, can direct their localization to the nucleus (Kane et al., 2018). However, MxB is mainly seen on the nuclear envelope and within the cytoplasm, which likely results from the action of a potential nuclear export signal (NES) in MxB, or strong interaction of MxB with nuclear pore complex (NPC) components (King et al., 2004;Busnadiego et al., 2014;Goujon et al., 2014;Dicks et al., 2018;Kane et al., 2018;Xu et al., 2020). In any event, the nuclear envelope location does not appear to be crucial for MxB inhibition of HIV-1 infection, since certain MxB mutants, such as the MxB 1-25 + NLS mutant and MX2(N91)-ARFAPTIN 2 mutant, lost localization to the nuclear envelope, yet still exhibited strong anti-HIV-1 activity Kane et al., 2018). It is likely that localization to the nuclear envelope may be required for certain cellular functions of MxB.
The NTD NLS of MxB does not fit in the definition of classic NLS, the bipartite NLS, nor the PY-NLS. Nonetheless, its nuclear localization function does depend on importin-β, but exhibits clear difference in the requirement of nucleoporins as compared to the classic NLS of SV40 large T antigen (Kane et al., 2018). The pressures for human MxB to evolve and maintain such an NLS are unknown. It is worth noting that the MxB proteins of primates all keep this NLS in their NTDs (Busnadiego et al., 2014). To further understand this conservation of NLS in MxB proteins, we changed the MxB NLS to other well-characterized NLS sequences and examine which heterologous NLS is able to at least partially rescue the known functions of MxB including inhibition of HIV-1. We found that the bipartite NLS is able to restore MxB localization to the nuclear envelope, whereas the arginine-rich NLS enables MxB to inhibit HIV-1, albeit exhibiting different restriction profiles toward HIV-1 capsid mutants.

Plasmid DNA
DNA sequence encoding wild type human MxB (GenBank accession number NM_002463.1) was cloned into either pEGFP-N1 vector (Clontech) with green fluorescence protein (GFP) sequence attached to the C-terminus or pQCXIP retroviral vector (Clontech) with a Flag-tag at the C-terminus. The MxB (1-25) mutation was generated by deleting sequences from Met1 to Glu25. The different NLS-MxB (1-25) chimeras were generated by replacing the first 25 amino acids with corresponding heterogeneous NLS sequences, the different N1-NLS constructs were generated by inserting the corresponding NLS sequences into pEGFP-N1 vector, and primers used are listed in Table 1. The HIV-1 proviral DNA clone NL4-3 was obtained from the NIH AIDS Reagent Program. HIV-1 DNA mutants NL4-3/CA-P90T, NL4-3/CA-E187V, and NL4-3/CA-P207S were generated using site-directed mutagenesis PCR method (Clontech). Tat-HA was cloned into pcDNA3.1 expression vector (Invitrogen) with a C-terminal HA tag. Luciferase reporter constructs CMV-Luc, SV40-Luc, PFV-LTR-Luc, and HIV-LTR-Luc were generated by inserting the promoter sequences into pGL3-basic luciferase reporter vector (Promega), which lacks promoter and enhancer sequences, expression of luciferase activity depends on insertion of a functional promoter at the upstream of luciferase gene. The importin-β siRNA (ID s7918) was purchased from Ambion. All constructs were confirmed by sequencing.

Luciferase Reporter Assay
Transfected or infected cells were collected and lysed in cell culture lysis buffer (Promega). Luciferase activity was measured using the luciferase assay system (Promega). All experiments were performed in triplicate and repeated at least three times.

Virus and Infection
HIV-1 particles were produced by transfection of HEK293T cells with proviral DNA constructs NL4-3, NL4-3/CA-P90T, NL4-3/CA-E187V, or NL4-3/CA-P207S. The culture medium was changed at 6 h post-transfection. Viruses in the supernatants were harvested at 48 h post-transfection, and clarified by centrifugation at 3,000 rpm for 10 min at 4 • C. The amount of virus was determined by measuring HIV-1 reverse transcriptase activity. For infection assays, equal amounts of viruses were used to infect SupT1 cell lines for 4 h at 37 • C in presence of 5 µg/ml polybrene, after which virus inoculate were washed off. At 48 h post-infection, the number of infected SupT1 cells was determined by flow cytometry after staining with FITCconjugated anti-p24 antibody. The amounts of infectious HIV-1 viruses in the culture supernatants were determined by infecting the TZM-bl indicator cells and measuring luciferase activity. All infection experiments were performed in triplicate, and repeated at least three times.
RT-qPCR to Quantify HIV-1 RNA Analysis of HIV-1 RNA by RT-qPCR was performed as follows. Briefly, total cellular RNA was extracted using Trizol (Invitrogen), followed by DNase (Invitrogen) treatment to remove DNA. Equal amounts of DNase-treated RNA were subjected to reverse transcription for cDNA synthesis using random hexamers (Invitrogen, catalog 8080127) and the MuMLV reverse transcriptase (Invitrogen, catalog number 28025013). Reverse transcription products were quantified by qPCR using Fast SYBR Green Master mix (Invitrogen, catalog 4385612) according to the manufacturer's instructions. qPCR reactions were performed in triplicate. β-actin mRNA also quantified in each sample to normalize HIV-1 RNA levels. The sequences of primers are as follows: 5 -GAC GCT CTC GCA CCC ATC TC-3 and 5 -CTG AAG CGC GCA CGG CAA-3 to quantify HIV-1 full-length RNA, and 5 -GAG CGG TTC CGC TGC CCT GAG GCA CTC-3 and 5 -GGG CAG TGA TCT CCT TCT GCA TCC TG-3 to quantify β -actin mRNA.

Statistical Analysis
Statistical difference between the two groups was analyzed using the Student's t-test with GraphPad Prism version 8.0. Differences were considered statistically significant when the p-value was <0.05. In the figures, p-values are indicated as follows: * for p < 0.05, * * for p < 0.001, * * * for p < 0.0001, ns for not significant.

The Bipartite NLS Supports MxB Localization to the Nuclear Envelope
We first asked which type of NLS, when used to replace the first 25 amino acids sequence of MxB, can support MxB localization to the nuclear envelope. We started with the wellcharacterized classic NLS of SV40 large T antigen (SV40T-NLS), the bipartite NLS of nucleoplasmin (NP) (BI-NP-NLS), the PY-NLS of hnRNPA1 (M9-NLS) (Lange et al., 2007). We also tested the arginine-rich non-classic NLS sequences of HIV-1 Tat and Rev proteins (Tat-NLS and Rev-NLS), given the presence of the arginine cluster in the NTD of MxB ( Figure 1A). The green fluorescence protein (GFP) sequence was attached to the C-terminus of MxB to facilitate detection by confocal microscopy. Similar to what was shown for HA-tagged MxB (Kane et al., 2013), MxB-GFP was mainly seen at the nuclear envelope ( Figure 1B). Deletion of the N-terminal 25 amino acids led to diffused cytoplasmic distribution ( Figure 1B). None of the SV40T-NLS, M9-NLS, Tat-NLS, or Rev-NLS, when used to replace the 25-amino acid NTD of MxB, restored localization of MxB to the nuclear envelope. In contrast, the BI-NP-NLS rendered complete nuclear localization of the engineered MxB mutant (Figure 1B). To test whether this observation is specific to NP-NLS or is a general feature of bipartite NLS, we further examined the bipartite NLS sequences of retinoblastoma (RB) protein and Ras-related C3 botulinum toxin substrate 3 (RAC3) protein (Cherezova et al., 2011). The results showed that both BI-RB-NLS and BI-RAC3-NLS enabled MxB localization to the nuclear envelope concurrent with dispersed distribution within the cytoplasm (Figure 1B). Therefore, the bipartite NLS, not the other NLS sequences tested herein, is able to position MxB to the nuclear envelope, with notable differences in the overall subcellular localization of the engineered MxB variants between the bipartite NLS sequences tested.

Nuclear Localization of BI-NP-NLS-MxB
Depends on Importin-β The bipartite NLS functions by binding the importin-α/β complex (Pawlowski et al., 2010). The nuclear localization activity of MxB NTD, when attached to reporter protein GFP-LacZ, is also dependent on importin-β (Kane et al., 2018). We thus knocked down importin-β with siRNA and examined how the subcellular localization of MxB and its mutants BI-NP-NLS-MxB (1−25), BI-RB-NLS-MxB (1−25), and BI-RAC3-NLS-MxB (1−25) were affected. In agreement with the study by Kane et al. (2018), depletion of importin-β did not affect localization of MxB to the nuclear envelope (Figure 2). Among the three MxB mutants tested, the BI-NP-NLS-MxB (1−25) mutant changed its location from the nucleus to the cytoplasm and formed aggregates, whereas the BI-RB-NLS-MxB (1-25) and BI-RAC3-NLS-MxB (1-25) mutants still showed strong association with nuclear envelope, and moderately increased nuclear presence in importin-β knockdown cells (Figure 2). Therefore, BI-NP-NLSmediated localization to the nucleus is sensitive to importin-β knockdown, as opposed to the NTD sequence of MxB, BI-RB-NLS, or BI-RAC3-NLS whose activity in directing MxB localization to the nuclear envelope is not affected.

The Bipartite NLS-Bearing MxB Dramatically Suppresses HIV-1 Gene Expression
MxB itself does not affect viral gene expression and virus production when both MxB and HIV-1 DNA are transfected into HEK293T cells (Liu et al., 2013). We posited that changing the NTD to different NLS sequences might not endow MxB with the ability of modulating HIV-1 gene expression. Indeed, the levels of HIV-1 Gag expression were not affected by MxB and its variants containing the NLS of SV40T, M9, Tat or Rev in HEK293T cells that were transfected with HIV-1 DNA and the above MxB DNA clones (Figure 3A), nor the levels of HIV-1 RNA and the amounts of HIV-1 particles were affected (Figures 3B-D). Unexpectedly, all three bipartite NLS transformed MxB into a strong suppressor of HIV-1 Gag expression ( Figure 3A). This inhibition occurred at the viral RNA level, since a dramatic reduction in the fulllength HIV-1 RNA, was detected with RT-PCR ( Figure 3B). As a result, BI-NLS-bearing MxB led to more than 100-fold decrease in the production of HIV-1 particles (Figures 3C,D). We further examined the subcellular localization of MxB-Flag and its mutants by immunofluorescence staining, and observed nuclear localization of BI-NP-NLS, BI-RB-NLS, and BI-RAC3-NLS-bearing MxB ( Figure 3E).
HIV-1 RNA synthesis is initiated from the viral LTR promoter, and is dramatically enhanced by viral Tat protein. We thus examined the effect of MxB and its mutants on the transcription from HIV-1 LTR and the transactivation activity of viral Tat, by transfecting HEK293T cells with the HIV-1 LTR-Luc reporter DNA, MxB or its mutants, with or without Tat DNA. The results showed that the BI-NLS-bearing MxB inhibited luciferase expression from HIV-1 LTR by more than 10-fold, whereas MxB and other mutants did not exhibit any effect (Figure 4A, left panel). The Tat protein increased luciferase expression from HIV-1 LTR by 250-fold. The magnitude of stimulation by Tat was not affected by MxB, nor its mutants, including bipartite NLS-bearing MxB mutants (Figure 4A, middle panel). But the overall gene FIGURE 2 | Nuclear localization of BI-NP-NLS-MxB (1-25) depends on importin-β. MxB or MxB mutants bearing a C-terminal GFP tag were expressed in HeLa cells that were transfected with control siRNA (siCtrl) or siRNA targeting importin-β (siImportin-β). Endogenous importin-β was immuno-stained with anti-importin-β antibodies. The subcellular localization of MxB and importin-β was visualized by confocal laser scanning microscopy. Scale bars represent 10 µm. expression from HIV-1 LTR in the presence of Tat was reduced by 10-fold by BI-NLS-MxB proteins. To demonstrate that BI-NLS needs to act together with the MxB sequence to inhibit HIV-1 LTR promoter, we attached BI-NLS to GFP and measured the effect of NLS-GFP fusion proteins on HIV-1 LTR transcription. None of the NLS-GFP proteins affected the activity of HIV-1 LTR ( Figure 4A, right panel).

The Search for MxB Orthologs Carrying Bipartite NLS
The drastic inhibition of HIV-1 gene expression by the MxB mutants carrying bipartite NLS prompted us to search for MxB orthologs that may carry bipartite NLS at their N-terminal regions. We aligned the N-terminal sequences of MxB proteins from 65 species, which are available in the NCBI database, and found bipartite-like NLS in MxB proteins from three species, Ailuropoda melanoleuca, Callithrix jacchus, and Tupaia chinensis ( Figure 5A). We then changed the first 25 amino acids sequence of human MxB to each of the corresponding MxB sequences from these three species. When these MxB mutants were co-transfected into HEK293T cells with HIV-1 DNA, no effect on HIV-1 Gag/p24 expression was detected by Western blotting, whereas the biparpite NLS-bearing MxB mutants drastically decreased HIV-1 Gag/p24 levels ( Figure 5B). Thus, unlike the bipartite NLS, the N-terminal sequences of MxB from species Ailuropoda melanoleuca, Callithrix acchus, and Tupaia chinensis do not transform human MxB into a strong inhibitor of HIV-1 gene expression.

DISCUSSION
In this study, we observed that the bipartite NLS, when being used to replace the N-terminal sequence of MxB, is able to maintain the nuclear envelope localization of MxB. This function is specific for bipartite NLS, since none of the other types of NLS tested in this study is able to restore MxB localization to the nuclear envelope. This unique activity of bipartite NLS may benefit from its ability of directly interacting with importin-β, as opposed to the NLS of SV40 T which needs to interact with importin-α before engaging importin-β, whereas the M9 NLS acts by binding to transportin 1. Alternatively, the bipartite NLS may interact with resident proteins in the nuclear pore complex or on the nuclear envelope so as to anchor MxB to the nuclear envelope.
The bipartite NLS-MxB mutants potently suppress HIV-1 gene expression, at least partly through inhibiting HIV-1 LTR promoter. It is possible that this inhibitory function results from the localization of bipartite NLS-MxB to the nucleus where these MxB mutants either modulate the binding of key transcription factors to HIV-1 LTR promoter DNA or directly interact with specific promoter sequences. Our data showed that this inhibition is specific for HIV-1 LTR, as these bipartite NLS-MxB mutants FIGURE 6 | MxB and its mutants bearing the arginine-rich NLS differentially inhibit wild type HIV-1 and viral mutants having CA mutations. SupT1 cells were transduced with retroviral particles that express MxB, Tat-NLS-MxB (1-25), or Rev-NLS-MxB (1-25), followed by infection with HIV-1 and its mutants CA-P90T, CA-E187V, or CA-P207S. MxB and its mutants contain a FLAG tag at the C-terminus. (A) Expression of MxB and its mutants was detected by anti-FLAG immunostaining. The percentages of FLAG-positive cells were scored by flow cytometry, and mean fluorescence intensity (MFI) values of FLAG staining were calculated and presented in the bar graphs. Results shown are the average of three independent experiments. *p < 0.05, **p < 0.001, ***p < 0.0001, ns, not significant. (B) HIV-1 infected cells were detected by immunostaining viral p24. The percentages of viral p24 positive cells in the Flag-positive cell population were scored by flow cytometry. Results shown are the average of three independent experiments. Fold of inhibition of HIV-1 infection by MxB or its mutants were calculated with reference to the infection of control cells by wild type HIV-1 or viral mutants. *p < 0.05, **p < 0.001, ***p < 0.0001. ns, not significant.
do not affect the gene expression of the other promoters tested (Figures 4B,D). One possible scenario is that bipartite NLS-MxB has blocked the function of a transcription factor that binds to HIV-1 LTR but not other promoters.
The importance of nuclear envelope localization for MxB function is largely unclear. Residing at the nuclear envelope may be crucial for the cellular functions MxB plays (King et al., 2004), but clearly is dispensable for its anti-HIV-1 activity. This is supported by the results of the Tat-NLS-MxB (1-25) and Rev-NLS-MxB (1-25) which inhibit HIV-1 to a similar degree to that by wild type MxB, yet are mostly dispersed within the cytoplasm and not found on the nuclear envelope. Similarly, the engineered proteins, such as that bear protein dimerization motifs and the N-terminal sequence of MxB, do not reside at the nuclear envelope, but still effectively inhibit HIV-1 Dicks et al., 2016;Kane et al., 2018).
The N-terminal sequence of MxB possesses the activity of directing protein localization to the nucleus. This is supported by the results showing nuclear location of reporter protein LacZ-GFP when bearing the N-terminal sequence of MxB (Kane et al., 2018). And this activity is dependent on importin-β, since knockdown of importin-β prevents LacZ-GFP/N-MxB from accessing the nucleus (Kane et al., 2018). However, depletion of importin-β did not affect the localization of MxB to the nuclear envelope, suggesting that, although the N-terminal sequence of MxB has the nuclear localization activity, its requirement for MxB localization to the nuclear envelope is independent of importin-β .
Detailed mutagenesis studies showed that the short argininerich motif 11-RRR-13 in the N-terminal region is essential for MxB inhibition of HIV-1 (Goujon et al., 2015;Schulte et al., 2015). Coincidentally, the NLS sequences of both Tat and Rev contain arginine clusters, and both exhibit the anti-HIV-1 property when used to replace the N-terminal sequence of MxB. Both structural and biochemical analysis have supported the interaction of MxB N-terminal domain with HIV-1 capsid at the tri-hexamer interface Summers et al., 2019). It is possible that the Tat and Rev NLS enable MxB to bind HIV-1 capsid, but the precise interactions may differ from each other and also differ from that of MxB binding to capsid, since the exact sequences of Tat NLS, Rev NLS and the N-terminal sequence of MxB are different. As a result, one MxB-resistant HIV-1 CA mutation may not necessarily resist the inhibition by Tat-NLS-MxB (1-25) or Rev-NLS-MxB (1-25). Indeed, our data showed that the Rev-NLS-MxB (1-25) strongly inhibits HIV-1 CA mutants that are refractory to wild type MxB and the Tat-NLS-MxB (1-25). This also suggests that Rev-NLS-MxB (1-25) may interact with HIV-1 capsid in a unique fashion which is distinct from either wild type MxB or Tat- . Consistent with what observed previously, MxB moderately increased the infection of HIV-1 carrying the P90T mutation in viral capsid (Liu et al., 2013). Interestingly, the Tat-NLS-MxB (1-15) mutant also enhanced the infection of the CA-E187V viral mutant ( Figure 6B). It is possible that these capsid mutants have changed the profile of host factors interacting with viral capsid, thus reverse viral response to MxB.
In summary, our results reveal the drastically different effects of various NLS on the subcellular localization and anti-HIV-1 activity of MxB, when used to replace the N-terminal sequence of MxB. Our data further support the essential role of the arginine-rich sequence in targeting HIV-1 capsid structure and thus inhibiting HIV-1 infection, albeit that the exact arginine-rich sequence dictates the specific binding of MxB to viral capsid and hence the resistance profile of HIV-1 capsid.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

AUTHOR CONTRIBUTIONS
CL and WQ conceived the study and acquired the funding. CL, JT, and KC designed the experiments. KC, ZW, and QP performed the experiments. CL, WQ, KC, ZW, and QP analyzed the data and wrote the manuscript. All authors contributed to the article and approved the submitted version.