Discovery of Novel Small-Molecule Inhibitors of LIM Domain Kinase for Inhibiting HIV-1

ABSTRACT A dynamic actin cytoskeleton is necessary for viral entry, intracellular migration, and virion release. For HIV-1 infection, during entry, the virus triggers early actin activity by hijacking chemokine coreceptor signaling, which activates a host dependency factor, cofilin, and its kinase, the LIM domain kinase (LIMK). Although knockdown of human LIM domain kinase 1 (LIMK1) with short hairpin RNA (shRNA) inhibits HIV infection, no specific small-molecule inhibitor of LIMK has been available. Here, we describe the design and discovery of novel classes of small-molecule inhibitors of LIMK for inhibiting HIV infection. We identified R10015 as a lead compound that blocks LIMK activity by binding to the ATP-binding pocket. R10015 specifically blocks viral DNA synthesis, nuclear migration, and virion release. In addition, R10015 inhibits multiple viruses, including Zaire ebolavirus (EBOV), Rift Valley fever virus (RVFV), Venezuelan equine encephalitis virus (VEEV), and herpes simplex virus 1 (HSV-1), suggesting that LIMK inhibitors could be developed as a new class of broad-spectrum antiviral drugs. IMPORTANCE The actin cytoskeleton is a structure that gives the cell shape and the ability to migrate. Viruses frequently rely on actin dynamics for entry and intracellular migration. In cells, actin dynamics are regulated by kinases, such as the LIM domain kinase (LIMK), which regulates actin activity through phosphorylation of cofilin, an actin-depolymerizing factor. Recent studies have found that LIMK/cofilin are targeted by viruses such as HIV-1 for propelling viral intracellular migration. Although inhibiting LIMK1 expression blocks HIV-1 infection, no highly specific LIMK inhibitor is available. This study describes the design, medicinal synthesis, and discovery of small-molecule LIMK inhibitors for blocking HIV-1 and several other viruses and emphasizes the feasibility of developing LIMK inhibitors as broad-spectrum antiviral drugs.

Structurally, these lead LIMK inhibitors represent three chemotypes (41). Scaffold A is based on the 4-yl-piperidine-or piperazine-benzimidazole derivative; scaffold B is derived from phenyl benzimidazole analogs, and scaffold C is based on the bis-aryl urea moiety. All three scaffolds share a substituted or unsubstituted 7H-pyrrolo[2,3-d]pyrimidine moiety (Table 1). We also performed a molecular docking study using R10015 (Fig. 3A) and the   1 3A). Our docking study suggests that all the compounds from these three scaffolds are type I ATP-competitive kinase inhibitors. In addition to the hinge interactions, the terminal aromatic moiety of these LIMK inhibitors is bound to a pocket under the P loop with strong hydrophobic interactions (Fig. 3A). An additional H-bonding interaction(s), which could contribute to the high affinity of these LIMK inhibitors, might also be present in this binding motif, depending on the functional substitution(s). For example, for R10015, there is also a potential H-bonding interaction between the ester carbonyl group and the backbone amide of residue Gly351 ( Fig. 3A and B). Molecular and cellular characterization of the LIMK inhibitor R10015. We selected R10015 from the potential LIMK inhibitors for further detailed mechanistic studies. The synthesis, purification, and biochemical characterization of R10015 are described in Materials and Methods below and shown in Fig. 3. We performed a cell-free assay to measure R10015 inhibition of the LIM domain kinase using purified human LIMK1 and its substrate, cofilin. The biochemical half-maximal inhibition concentration (IC 50 ) of R10015 against purified LIMK1 was determined to be approximately 38 nM (Fig. 3C). We also performed profiling of R10015 against a panel of 62 kinases ( Table 2). The results demonstrated that R10015 has good selectivity, with off-target inhibition of only LRRK2 and p70S6K (Ն90% inhibition at 1 M R10015) and moderate inhibition of protein kinase A (PKA) (ϳ76%), ROCK2 (ϳ70%), and FLT3 (ϳ68%).
To further determine the specificity of R10015 in human CD4 T cells, we treated CEM-SS T cells with R10015 for a time course. Similar to the cell-free assay, we used LIMK-mediated cofilin phosphorylation as the readout. Following R10015 treatment, the cells were lysed, and cofilin phosphorylation was measured by Western blotting. As shown in Fig. 3D, we detected drastic inhibition of cofilin serine 3 phosphorylation, confirming that R10015 blocks the LIMK activity in cells. The decrease of phosphocofilin (p-cofilin) did not result from possible inhibition of cofilin protein synthesis. When total protein levels were quantified, R10015 did not cause a decrease in the total amount of cofilin (Fig. 3D). The total protein levels of LIMK itself were also not decreased (Fig. 3D). To further confirm that the inhibition of cofilin phosphorylation occurs directly through blocking LIM kinase, we examined upstream signaling molecules that regulate LIMK activity. We found that R10015 did not inhibit the threonine 508/505 phosphorylation of LIMK1/2 itself by the upstream kinase PAK2, whose activity is regulated by upstream Rho family GTPases, such as Rac, cdc42, and Rho (11,42) (Fig.  3D). The phosphorylation of PAK2 was also not affected by R10015 (Fig. 3D). These results strongly suggest that the inhibition of cofilin serine 3 phosphorylation by R10015 came from the direct inhibition of the activity of LIMK; the LIMK upstream kinases and GTPases are not the targets of R10015. These cell-based studies confirmed the high selectivity of R10015 in blocking the LIM kinase activity in human CD4 T cells. The Western blot quantification of R10015 on cofilin phosphorylation (Fig. 3D) represents the drug effect on cells as a whole population. We further quantified cofilin phosphorylation in individual T cells by intracellular staining and flow cytometry. Cells  a Profiling was carried out at Reaction Biology Corporation following a protocol described on its website. b The activity data are averages of two determinations (DMSO control, 100% activity). The ATP concentration used was 10 M; staurosporine was used as the control for assay validation.
were treated with R10015 for a time course, permeabilized, and stained with a rabbit anti-human p-cofilin antibody, followed by staining with Alexa Fluor 647-labeled chicken anti-rabbit antibodies. Stained cells were analyzed by flow cytometry. As shown in Fig. 3E, in cells not treated with R10015 (0 h), approximately 80% of the cells stained relatively high on p-cofilin. Following R10015 treatment for 1 h, 80% of the cells stained low on p-cofilin. These results confirm that R10015 inhibits cofilin phosphorylation in a majority of T cells. We further tested R10015 for inhibition of stromal cell-derived factor 1 (SDF-1)mediated chemotaxis and actin dynamics, which are known to be regulated by LIMK (43). Human Jurkat T cells were treated with different doses of R10015 and then placed into the upper chamber of a transwell plate. SDF-1 was added to the lower chamber. The migration of the cells to the lower chamber was enumerated. As shown in Fig. 3F, we observed a dose-dependent inhibition of Jurkat T cell chemotaxis, and the IC 50 was around 10 M, which is higher than the IC 50 (38 nM) observed in the in vitro LIMK assay (Fig. 3C). This difference in R10015 potency between the in vitro assay and that in live cells likely resulted from nonoptimal properties of R10015 for intracellular delivery, which may require further medicinal chemistry optimization. To further confirm that the inhibition of SDF-1-mediated chemotaxis is directly related to R10015-mediated inhibition of actin dynamics, we measured actin polymerization following SDF-1 and R10015 treatment. For measuring actin polymerization, human blood resting CD4 T cells were treated with R10015 and SDF-1 for a time course. The cells were permeabilized, stained with fluorescein isothiocyanate (FITC)-phalloidin for F-actin, and analyzed by flow cytometry. Consistently, pretreatment of resting CD4 T cells with R10015 markedly depressed SDF-1-mediated actin polymerization ( Fig. 3G and H), confirming that R10015 blocks LIMK-regulated actin dynamics. Confocal microscopy observation of R10015-treated T cells revealed that the drug dramatically blocked actin polymerization and actin capping following SDF-1 stimulation ( Fig. 3I and J).
Mechanistic study of R10015 for blocking HIV infection. Previous shRNA/siRNA LIMK knockdown studies have demonstrated that LIMK is involved in viral entry, DNA synthesis, nuclear migration, and viral budding (11,14). We first quantified R10015 inhibition of the early steps of HIV infection of CD4 T cells, in which cells were exposed to R10015 only briefly during viral infection; R10015 was removed from the infection culture following infection. As shown in Fig. 4, Rev-CEM-GFP-Luc cells were pretreated with different doses of R10015 for 1 h, infected with HIV for 3 h, washed to remove the input virus and R10015, and then cultured in the absence of R10015 for 48 h. HIV-dependent luciferase expression was quantified. For inhibition of HIV, R10015 showed an IC 50 of 14.9 M (Fig. 4A), which is close to the IC 50 (10 M) for inhibition of SDF-1-mediated chemotaxis (Fig. 3F). The drug had no detectable cytotoxicity for this short period of treatment (4 h) at all the concentrations tested (up to 200 M) ( Fig. 4B and C). In addition, R10015 minimally inhibited the early steps of vesicular stomatitis virus (VSV) G-pseudotyped HIV, even at 100 M ( Fig. 4A and E), confirming that the inhibition of HIV did not result from nonspecific cytotoxicity and is indeed specific to viral processes related to HIV gp120-mediated fusion and entry (44); HIV gp120mediated infection of human CD4 T cells has been known to require cortical actin dynamics (8). On the other hand, VSV G-mediated endocytosis bypasses cortical actin (45) and thus is less susceptible to R10015. This differential susceptibility of gp120versus VSV G-pseudotyped virus to R10015 is in agreement with the hypothesis that R10015 inhibits HIV through direct blockage of the LIMK-regulated actin dynamics.
To further elucidate the anti-HIV mechanism, we pretreated human CEM-SS T cells with R10015 for 1 h, and the effects of R10015 on the surface expression of the HIV receptors, both CD4 and CXCR4, were examined. This brief treatment did not alter the surface expression of CD4 or CXCR4; prolonged treatment (4 h), however, slightly decreased the surface density of CD4 ( Fig. 5A) (11). We further examined the effect of R10015 on viral entry using two HIV entry assays, the Vpr-␤-lactamase-and Nef-luciferase-based assays (46, 47). R10015 did not significantly inhibit viral entry, as measured by infection with HIV-1(Vpr-␤-lactamase) (46) (Fig. 5B) or HIV-1(Nefluciferase) virion particles (reference 47 and data not shown). We then followed viral postentry steps, such as viral DNA synthesis and nuclear migration. For these stepwise mappings, we used a single-cycle virus, HIV-1(Env), which is pseudotyped with HIV-1 gp160 (45). CEM-SS T cells were pretreated with R10015 for 1 h and then infected with HIV-1(Env) for 2 h in the presence of R10015. Following infection, the cells were washed to remove HIV-1 and R10015 and cultured for 24 to 48 h. The late products of HIV reverse transcription were quantified by real-time PCR. As shown in Fig. 5C, R10015 inhibited viral late DNA synthesis at all time points. We also measured HIV nuclear   migration using the synthesis of HIV circular DNA containing 2 copies of the viral long terminal repeat (2-LTR circles) as a surrogate. R10015 also inhibited viral nuclear migration, as measured by viral 2-LTR circles (Fig. 5D).
Recently, LIMKs have been suggested to be involved in retroviral budding and release (14); siRNA knockdown of LIMK1 prevents mature virions from leaving the plasma membrane. We further tested the effects of R10015 on the late stage of HIV infection. CEM-SS T cells were infected with HIV-1(Env) for 2 h, washed, incubated for 12 h, and then treated with R10015. Virion release from 12 to 72 h postinfection was quantified by measuring p24 release into the supernatant. As shown in Fig. 5E, R10015 also inhibited p24 release when applied at a later stage of HIV infection, confirming previous findings on LIMK in HIV budding and release (14). As a complementary approach, we also treated a chronically HIV-infected T cell line, ACH-2, with R10015, and observed similar inhibition of virion release (Fig. 5F). Furthermore, R10015 also inhibited virion release from HIV DNA-transfected HEK293 cells (Fig. 5G); the use of DNA transfection to assemble virion particles abrogates the requirements for viral early steps, such as entry, reverse transcription, and integration. Thus, the inhibition of virion release by R10015 is associated with direct blocking of viral late steps rather than early steps. Together, our results demonstrate that R10015 inhibited viral DNA synthesis, nuclear migration, and virion release, phenotypes consistent with those observed in shRNA/siRNA LIMK1 knockdown cells (11,14).
During HIV transmission, early viruses utilize CCR5 and mainly infect memory CD4 T cells and macrophages, whereas late viruses also utilize CXCR4 (48)(49)(50). We tested the ability of R10015 to inhibit R5 and X4 latent infection of blood memory and resting CD4 T cells (8,51). In our viral latent infection system, total resting CD4 T cells or the CD45RO ϩ RA Ϫ memory CD4 T cells were purified from the peripheral blood, cultured overnight without stimulation, and then treated with R10015 for 1 h. Resting T cells were infected with HIV-1(NL4-3) for 2 h, washed, cultured for 5 days in the absence of R10015, and then activated with CD3/CD28 beads to stimulate viral replication. Viral p24 release was measured. As shown in Fig. 6, we observed that R10015 blocked X4 HIV-1 latent infection of resting CD4 T cells (Fig. 6A); we also observed R10015 dosedependent inhibition of R5 HIV latent infection of resting memory CD4 T cells (Fig. 6C). The inhibition of HIV latent infection of resting T cells did not result from possible drug cytotoxicity, because the short drug treatment did not inhibit T cell function, as judged by the surface expression of CD25 or CD69 following T cell activation ( Fig. 6B and D). These results confirmed that the LIMK inhibitor is effective in blocking both X4 and R5 viral infection of primary target cells. We further tested the ability of R10015 to inhibit infection by primary HIV isolates. As shown in Fig. 6E, HIV 92/BR/018 (Brazil) and HIV 93UG070 (Uganda) were used to infect peripheral blood mononuclear cells (PBMC) in the presence of R10015. We observed inhibition of HIV replication by R10015.
Inhibition of multiple viruses by R10015. Given that the requirement for actin dynamics is shared among viruses, we tested R10015 for its ability to inhibit several other viruses, including Zaire ebolavirus (EBOV), Rift Valley fever virus (RVFV), Venezuelan equine encephalitis virus (VEEV), and herpes simplex virus 1 (HSV-1). Ebola virus is a negative-sense, single-stranded RNA virus of the family Filoviridae. EBOV acquires an envelope from the host cell membrane during virion budding. The cellular entry of EBOV has been suggested to occur through a macropinocytosis-like mechanism that involves localized polymerization of the actin filaments (52). To test the possible effects of R10015, we pretreated human HFF-1 cells with R10015 for 2 h and then infected them with EBOV (Zaire) for 48 h in the continuous presence of R10015. To quantify viral replication, the cells were fixed and stained for the EBOV glycoprotein with an Alexa 488-labeled antibody and then analyzed by confocal imaging. We observed dosedependent inhibition of EBOV ( Fig. 7A and B). This inhibition is likely a combined effect of R10015 on viral early and late processes, such as intracellular migration and virion release.
We further tested the effects of R10015 on another negative-sense, single-stranded RNA virus, RVFV, of the bunyavirus family. RVFV is a cytoplasmic virus, and the cellular entry of RVFV is predicted to occur via a class II fusion mechanism that becomes activated by low pH following endocytosis of the virus (53). Given that R10015 is less effective in inhibiting endocytotic entry mediated by VSV G (Fig. 4C), we treated cells with R10015 for the entire course of viral replication to block both early and late steps, such as viral assembly. We used a luciferase-tagged virus, RVFV-Luc (MP12), to measure   we also treated cells with R10015 for the entire course of viral replication. We used a luciferase-tagged virus, VEEV-Luc (TC-83), and the attenuated VEEV (TC-83) (biosafety level 2 [BSL-2]) and observed dosage-dependent inhibition in both luciferase and plaque assays with an IC 50 of approximately 5 M (Fig. 7D). We confirmed these results with the infectious VEEV (TrD) strain (BSL3), and observed 2 log units of inhibition of VEEV (TrD) with R10015 (Fig. 7D).
Recently, studies have suggested that LIMK is also involved in HSV-1 infection of neurons (18). During viral entry, HSV-1 triggers biphasic remodeling of the actin cytoskeleton through phosphorylation and dephosphorylation of cofilin, which is mediated through the LIMK-cofilin signaling pathway (18). To confirm the involvement of LIMK in the early infection steps of HSV-1 infection, we pretreated Vero cells with R10015 for 2 h and then infected them with HSV-1 for 2 h. Following infection, both HSV-1 and R10015 were washed away, and the cells were cultured in the absence of the drug. As shown in Fig. 7E, we observed inhibition of HSV-1 by R10015, even with this brief treatment, in agreement with a previous report that LIMK is involved in HSV-1 early infection steps. Together, our results demonstrate the broad antiviral activity of R10015.
To further test the feasibility of developing LIMK inhibitors for in vivo use, we used five different LIMK inhibitors to treat mice daily for 7 days (Fig. 7F). R10904, R10906, R10907, or R7826 was delivered via oral gavage at 20 mg/kg of body weight. Phosphate-buffered saline (PBS) was used as a control. R10015 (dissolved in dimethyl sulfoxide [DMSO]) was delivered via intraperitoneal (i.p.) injection at 10 mg/kg. DMSO was used as a control. The mice were weighed daily and monitored for morbidity and mortality, including lethargy and ruffled fur. None of the LIMK inhibitors displayed any indication of toxicity. In fact, mice treated with R10904, R10906, and R7826 gained weight over the course of a week compared to the PBS control. These results suggest the possibility of short-term use of LIMK inhibitors to block viral infections.

DISCUSSION
In this article, we described the design, medicinal synthesis, and discovery of LIMK inhibitors for blocking HIV-1, EBOV, and other viruses. Our study is the first proof-ofconcept study for developing LIMK inhibitors to block viruses. Our results emphasize the feasibility of using LIMK inhibitors as anti-HIV and broad-spectrum antiviral drugs. The rational design and development of LIMK inhibitors as antiviral drugs are largely based on recent studies on the role of LIMK/cofilin in HIV-1 infection (8,11,14). It has been shown that HIV requires LIMK1/cofilin activity for entry, intracellular migration, cell-cell transmission, and virion budding (8,11,14). Given that the requirement for actin dynamics is shared among viruses, these LIMK inhibitors are capable of inhibiting multiple viruses.
As a proof-of-concept study, our current lead compounds, particularly R10015, have low molecular weight (Ͻ450) and are amenable to further medicinal chemistry optimization to achieve higher solubility, cell potency, and appropriate drug metabolism and pharmacokinetic properties (DMPK). When profiled against a panel of 62 kinases in vitro (Table 2) To further ensure that R10015 blocks HIV by inhibiting LIMK, we performed a series of experiments demonstrating that (i) R10015 strongly blocked cofilin phosphorylation, whereas the other upstream kinases were largely unaffected, demonstrating its good selectivity (Fig. 3D); (ii) R10015 effectively blocked actin polymerization and T cell chemotaxis ( Fig. 3F and G), as these are the expected properties of LIMK inhibitors; and (iii) most importantly, R10015 selectively blocked wild-type HIV infection, but not VSV G-pseudotyped HIV infection, when used early during viral entry (Fig. 4). This strongly suggests that R10015-mediated viral inhibition does not result from nonspecific cytotoxicity but is related to the inhibition of specific processes, such as actin dynamics, that are involved in HIV entry; it has been well documented that HIV gp120-mediated fusion and entry require cortical actin activities (8), while VSV G-mediated endocytotic entry is less dependent on actin dynamics (45). Furthermore, our detailed stepwise molecular mapping confirmed that R10015-mediated inhibition of viral nuclear entry (Fig. 5D) is similar to that observed in shRNA LIMK knockdown cells (11).
Given the critical role of LIMK in regulating cell migration, chemotaxis, and T cell activation, long-term suppression of LIMK may inhibit immune responses. Nevertheless, the possibility of developing future clinical LIMK inhibitors is suggested by genetic studies in humans and mice. In the human genome, the LIMK1 gene is located in a region on human chromosome 7 (7q11.23) and is haplodeleted, along with 24 other genes, in adults living with Williams syndrome (WS), a rare neurodevelopmental genetic disorder associated with mild mental retardation (58)(59)(60). The deletion of elastin (ELN) has been explicitly linked to most of the cardiovascular problems in WS (61). However, the neurodevelopmental genotype-phenotype correlation is still uncertain and may be related to the hemizygosity of WBSCR11, CYLN2, GTF2I, NCF1, and perhaps LIMK1 (60). Specifically, LIMK1 knockout in mice has been linked to alterations in spine morphology and synaptic function. The knockout mice also showed heightened locomotor activity and altered fear responses and spatial learning (62). Nevertheless, LIMK1-null mice and human adults with WS do not have the severe multiple developmental disorders normally seen in other developmental genetic diseases, suggesting that blocking LIMK1 is not fatal and that short-term or localized LIMK inhibition is likely tolerable in adults. Therefore, LIMK is considered a valuable target for treating various human diseases, such as metastatic cancer, Alzheimer's disease, and drug addiction (63)(64)(65). These previous studies are also consistent with our results showing that the LIMK inhibitors are generally nontoxic in our cell-based assays and in mice (Fig. 7F), demonstrating the feasibility of developing highly specific LIMK inhibitors for short-term use to inhibit HIV and other viruses. Given the lack of an effective HIV vaccine, these novel inhibitors may prove to be valuable alternatives for preventing HIV sexual transmission; these LIMK inhibitors inhibit viral reverse transcription, nuclear migration, and release and are expected to be effective against a broad spectrum of HIV strains because of the highly conserved nature of viral dependency on actin dynamics for infection. In addition, these LIMK inhibitors have anti-inflammatory properties and can inhibit the migration of infected immune cells for HIV cell-cell transmission (Fig. 3F) (14). Thus, these LIMK inhibitors are ideal candidates, as potential topical microbicides, for preexposure prophylaxis that may complement the current antiretroviral drugs. Finally, our results also emphasize the feasibility of developing LIMK inhibitors as a new class of broad-spectrum antiviral drugs for urgent treatment of exposure to viral agents.

MATERIALS AND METHODS
Approvals from IRB and IACUC. Peripheral blood was drawn from HIV-negative donors. All protocols involving human subjects were reviewed and approved by the George Mason University (GMU) institutional review board (IRB). Mouse experiments were carried out in animal BSL-2 facilities in accordance with the National Research Council's Guide for the Care and Use of Laboratory Animals and under GMU IACUC protocol number 0211.
Synthesis of R10015. Commercially available reagents and anhydrous solvents were used without further purification unless otherwise specified. Thin-layer chromatography (TLC) analyses were performed with precoated silica gel 60 F254 plates. Mass spectra were recorded by liquid chromatography (LC)-mass spectrometry (MS) with a Finnigan LCQ Advantage Max spectrometer from Thermo Electron. Flash chromatography was performed on prepacked silica gel columns (230 to 400 mesh; 40 to 63 m) by CombiFlash with ethyl acetate (EtOAc)-hexane or methanol (MeOH)-dichloromethane (DCM) as the eluent. Preparatory high-performance liquid chromatography (HPLC) was performed on a SunFire C 18 OBD column, 10 m (30 by 250 mm), with CH 3 CN plus 50% MeOH-H 2 O plus 0.1% trifluoroacetic acid (TFA) as the eluent to purify the targeted compounds. Analytic HPLC was performed on an Agilent Technologies 1200 series HPLC with CH 3 CN (solvent B)-H 2 O plus 0.9% CH 3 CN plus 0.1% TFA (solvent A) as the eluent; the targeted products were detected by UV in the detection range of 215 to 310 nm. Nuclear magnetic resonance (NMR) spectra were recorded with a Bruker 400-MHz spectrometer at ambient temperature, with the residual solvent peaks as internal standards. The line positions of multiplets are given in parts per million (␦), and the coupling constants (J) are given in hertz. The high-resolution mass spectrum (HRMS) (electrospray ionization) experiments were performed with a Thermo Finnigan Orbitrap mass analyzer. Data were acquired in the positive ion mode at a resolving power of 100,000 at m/z 400. Calibration was performed with an external calibration mixture immediately prior to analysis.
General synthetic procedures. The scheme and synthetic procedures described below are for the inhibitor R10015. The syntheses of the other LIMK inhibitors listed in Table 1  Stirring was continued at room temperature overnight, after which LC-MS indicated a complete reaction. The solvent was removed in vacuo to residue, which was suspended in EtOAc. The suspension was washed with brine and saturated NaHCO 3 , dried over anhydrous Na 2 SO 4 , and evaporated under reduced pressure to give a mixture of crude amide products 3 and 4. Without further purification, the mixture was suspended in acetic acid and heated at 70°C for 4 h for ring closure to give the Boc-protected 4-yl-piperidinobenzimidazole, which was purified by flash chromatography. The Boc protection was then removed with 30% TFA in DCM to yield 5 as an oil. Finally, a mixture of 5 and 4,5-dichloro-7H-pyrrolo [2,3d]pyrimidine in a small amount of isopropanol was heated at 130°C under microwave conditions for 3 h to produce the LIMK inhibitor R10015, which was purified by reverse-phase HPLC to a purity of Ͼ95% based on analytical HPLC analysis (UV detection at 254 nm).

Methyl 2-(1-(5-chloro-7H-pyrrolo[2,3-d]pyrimidin-4-yl) piperidin-4-yl)-1H-benzo[d]imidazole-5carboxylate (R10015).
There was a 25% yield of R10015 in 4 steps after HPLC purification. 1  R10015 docking studies. The inhibitor R10015 was prepared for glide docking using LigPrep (Schrodinger, New York, NY). Chain A of PDB accession no. 3S95 was prepared using the protein preparation wizard in Maestro V 9.8 (Schrodinger, New York, NY) by removing water molecules and bound ligand and adding hydrogen atoms. The docking grid was generated around the original ligand with a box size of 18 by 18 by 18 Å. Docking was conducted without any constraint. The top-scored docking pose was merged with the protein for energy minimization using Prime (Schrodinger, New York, NY).
Isolation of PBMC and resting CD4 T cells from peripheral blood. All protocols involving human subjects were reviewed and approved by the George Mason University institutional review board. PBMC were purified from peripheral blood of HIV-negative donors by centrifugation in lymphocyte separation medium (Corning), and resting CD4 T cells were further purified by two rounds of negative selection, as previously described (8,51). Briefly, for the first-round depletion, we used monoclonal antibodies against human CD14; CD56; and HLA-DR, -DP, and -DQ (BD Biosciences). For the second-round depletion, we used monoclonal antibodies against human CD8, CD11b, and CD19 (BD Biosciences, San Jose, CA). Antibody-bound cells were depleted using Dynabeads Pan Mouse IgG (Invitrogen, Carlsbad, CA). For further negative selection of the memory and naive CD4 T cell subsets, monoclonal antibody against either CD45RA (0.02 l per million cells) or CD45RO (0.1 l per million cells) (BD Biosciences, San Jose, CA) was added during the second round of depletion. Purified cells were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Invitrogen, Carlsbad, CA), penicillin Signaling, Danvers, MA), with rabbit anti-cofilin antibodies (Cell Signaling, Danvers, MA), or with rabbit anti-PAK2 antibody (Cell Signaling, Danvers, MA). These antibodies were diluted in 2.5% milk-TBST and incubated with the blots overnight at 4°C. The blots were washed three times for 15 min each time, incubated with anti-rabbit horseradish peroxidase-conjugated secondary antibodies (KPL, Gaithersburg, MD; 1:5,000) for 1 h, and then developed with SuperSignal West Femto Maximum Sensitivity substrate (Thermo Fisher Scientific, Grand Island, NY). For loading control, the same blots were stripped and reprobed with antibodies against GAPDH (glyceraldehyde-3-phosphate dehydrogenase) (Abcam, Cambridge, MA; 1:1,000). Images were captured with a charge-coupled-device (CCD) camera (Protein Simple, San Jose, CA).
Intracellular p-cofilin staining and flow cytometry. One million cells were fixed, permeabilized with methanol, washed twice, and then stained with rabbit anti-human p-cofilin antibodies (Cell Signaling, Danvers, MA) for 60 min at room temperature. The cells were washed twice and stained with 10 g/ml Alexa Fluor 647 (Abcam, Cambridge, MA)-labeled chicken anti-rabbit antibodies (Invitrogen, Carlsbad, CA). The cells were washed twice and then analyzed on a FACSCalibur (BD Biosciences, San Jose, CA).
Surface staining of CD4 and CXCR4. Cells were stained with FITC-labeled monoclonal antibody against human CD4 (clone RPA-T4) or CXCR4 (clone 12G5) (BD Biosciences, San Jose, CA). The cells were stained on ice in PBS plus 0.1% BSA for 30 min, washed with cold PBS-0.5% BSA, and then analyzed on a FACSCalibur (BD Biosciences, San Jose, CA).
Quantitative real-time PCR. HIV DNA was quantified using the Bio-Rad iQ5 real-time PCR detection system, utilizing the forward primer 5=LTR-U5, the reverse primer 3= gag, and the probe FAM-U5/gag, as previously described (8). The prequalified, full-length proviral plasmid pNL4-3 was used as the DNA standard. Viral 2-LTR circles were measured as described previously (8). For measuring 2-LTR circles, the DNA was amplified by real-time PCR with primers MH536 and MH535 and probe MH603 (8).
Chemotaxis assay. A half million Jurkat T cells were resuspended in 100 l RPMI 1640 medium and then added to the upper chamber of a 24-well transwell plate (Corning, Corning, NY). The lower chamber was filled with 600 l of medium premixed with SDF-1 (40 ng/ml). The plate was incubated at 37°C for 2 h, and then the upper chamber was removed and the cells in the lower chamber were counted. Where indicated, different concentrations of R10015 were added to the culture supernatant prior to the assay, along with a DMSO control. Viral entry assays. The BlaM-Vpr-based viral entry assay was performed as previously described (8,46). We also used a Nef-luciferase-based entry assay as described previously (47). Briefly, cells (1 ϫ 10 6 ) were treated with 100 M R10015 for 1 h, infected with 200 ng of Nef-luciferase-containing viruses at 37°C for 3 h with the same R10015 concentration, and then washed three times with medium. The cells were resuspended in 100 l of luciferase assay buffer (Promega, Madison, WI), and luciferase activity was measured in live cells using a GloMax-Multi detection system (Promega, Madison, WI). Viral release assay. Chronically HIV-infected ACH2 cells were obtained from the NIH AIDS Reagent Program. The cells were cultured in RPMI 1640 (Invitrogen) plus 10% FBS. The cells were washed 3 times prior to treatment with R10015. DMSO was used as a control. The culture supernatants were harvested on days 2 and 3 post-R10015 treatment and analyzed for HIV-1 p24 by ELISA. For quantification of virion release from DNA-transfected HEK293 cells, the cells were transfected with pHIV(NL4-3) using Lipofectamine 2000 (Thermo Fisher Scientific). After 5 h, the cells were washed and cultured in complete DMEM and then treated with different concentrations of R10015 for 48 h. DMSO was used as a control. The supernatants were harvested and analyzed by ELISA for HIV-1 p24.
Surface staining of CD25 and CD69. A half million resting CD4 T cells were cultured for 5 days and stimulated with anti-CD3/CD28 beads (4 beads per cell) for 24 h. The cells were stained with phycoerythrin (PE)-labeled monoclonal antibody against human CD25 (clone M-A251) or CD69 (clone FN50) (BD Biosciences, San Jose, CA). The cells were stained on ice in PBS plus 0.1% BSA for 30 min, washed with cold PBS-0.5% BSA, and then analyzed on a FACSCalibur (BD Biosciences, San Jose, CA).
Cytotoxicity assay. The cytotoxicity assay was performed using the MultiTox-Glo Multiplex cytotoxicity assay kit (Promega), as suggested by the kit manufacturer. Cells were cultured in a Cellstar white 96-well plate (Greiner) at a density of 15,000 cells/well (CEM T cells) or 5,000 cells/well (293T cells) and then treated with R10015 for various times at various drug concentrations. DMSO was used as the control. For toxicity control, 100 l of 1% saponin was added to cells for 3 h to induce cell lysis. Cytotoxicity is presented in relative luminescence units as measured by the GloMax Discover System plate reader (Promega).
Confocal microscopy. Cells were imaged using a Carl Zeiss Zen 780 laser scanning microscope with a 40ϫ/numerical aperture (NA) Oil DIC Plan-Apochromatic objective in the green fluorescence channel. The images were then processed and analyzed with Zen 780 software.

ACKNOWLEDGMENTS
We thank the George Mason University Student Health Center for blood donations, the NIH AIDS Research and Reference Reagent Program for reagents, Jennifer Guernsey A provisional patent application pertaining to the results presented in the article has been filed.
The content is solely our responsibility and does not necessarily represent the official views of the National Institutes of Health.
This work was funded by the 2010 NYCDC AIDS Ride organized by Marty Rosen and in part by Public Health Service grants 1R01MH102144 and 1R03AI110174 from NIAID and NIMH to Y.W. and R21EY021799 to Y.F. and by the Defense Threat Reduction Agency/Joint Science and Technology Office.