The Basic Domain of HIV-Tat Transactivating Protein Is Essential for Its Targeting to Lipid Rafts and Regulating Fibroblast Growth Factor-2 Signaling in Podocytes Isolated from Children with HIV-1–Associated Nephropathy

Abstract

Podocyte injury has a critical role in the pathogenesis of HIV-associated nephropathy (HIVAN). The HIV-1 transactivator of transcription (Tat), combined with fibroblast growth factor-2 (FGF-2), can induce the dedifferentiation and proliferation of cultured human podocytes. Cellular internalization of Tat requires interactions with heparan sulfate proteoglycans and cholesterol-enriched lipid rafts (LRs). However, the specific distribution of Tat in human podocytes and its ability to associate with LRs have not been documented. Here, we found that Tat is preferentially recruited to LRs in podocytes isolated from children with HIVAN. Furthermore, we identified arginines in the basic domain (RKKRRQRRR) of Tat as essential for (1) targeting Tat to LRs, (2) Tat-mediated increases in the expression of Rho-A and matrix metalloproteinase-9 in LRs, and (3) Tat-mediated enhancement of FGF-2 signaling in human podocytes and HIV-transgenic mouse kidneys and the exacerbation of renal lesions in these mice. Tat carrying alanine substitutions in the basic domain (AKKAAQAAA) remained localized in the cytosol and did not associate with LRs or enhance FGF-2 signaling in cultured podocytes. These results show the specific association of Tat with LRs in podocytes isolated from children with HIVAN, confirm Tat as a regulator of FGF-2 signaling in LRs, and identify the key domain of Tat responsible for promoting these effects and aggravating renal injury in HIV-transgenic mice. Moreover, these results provide a molecular framework for developing novel therapies to improve the clinical outcome of children with HIVAN.

HIV-associated nephropathy (HIVAN) is a progressive renal disease seen in HIV-infected patients from African ancestry. One feature characteristic of HIVAN is the collapse of glomerular capillaries associated with the dedifferentiation and proliferation of podocytes.¹ These changes affect the glomerular filtration barrier² and cause heavy proteinuria and renal failure. Previous studies suggest that podocytes can be productively infected in vivo³ and that cells derived from the parietal epithelium^4,5 may also be affected. In addition, circulating viral proteins and cytokines released by HIV-infected cells can induce renal injury as well.¹ In this regard, the HIV-1 transactivator of transcription (Tat) protein has received significant attention, because it is a transcription factor that enhances viral replication; it can also be released and taken up by noninfected cells. In this manner, Tat can affect the function of podocytes acting through at least two different mechanisms.^6–8 A concentration of Tat within the range detected in sera of HIV-infected patients^9,10 can cause glomerular permeability changes in vivo¹¹ and induce the proliferation of cultured podocytes by stimulating the release of fibroblast growth factor-2 (FGF-2).¹² Indeed, high plasma, kidney, and urinary levels of FGF-2 are detected in children with HIVAN,^13,14 and FGF-2 is accumulated in the kidney of HIV-Tg₂₆ mice with renal disease.^15–17 Overall, these studies suggest that both Tat and FGF-2 may play a role in the pathogenesis of childhood HIVAN.

Extracellular Tat and FGF-2 specifically bind and interact with negatively charged cell-surface heparan sulfate proteoglycans (HSPGs).^6,18–20 Tat entry into cells is facilitated by an endocytic pathway that originates in lipid rafts (LRs) and involves caveolar endocytosis²¹ or LR-dependent macropinocytosis.²² LRs are specialized membrane domains enriched in certain lipids, cholesterol, and proteins that serve as docking sites for signaling proteins, including G protein–coupled receptors.^23–25 They are resistant to low-temperature solubilization by nonionic detergents, and this property allows for their separation by differential flotation after density-gradient centrifugation in low-density fractions that are called detergent-resistant membranes (DRMs).²³ Actin binding proteins bind to polyphosphoinositides located in LRs, and these proteins link the actin cytoskeleton with signaling molecules that are enriched in the LRs.²³ In fact, activation of G protein–coupled receptors by HIV-Tat in brain endothelial cells results in the stimulation of small guanosine 5′‑triphosphatase of the Rho-family, which in turn, activate Rho-kinase and the myosin binding subunit of the myosin light chain (MLC).^26,27 Moreover, activation of the Rho-A pathway in podocytes can cause chronic renal failure in transgenic mice.²⁸ Therefore, we carried out this study to determine whether Tat, alone or combined with FGF-2, can induce cytoskeletal changes in cultured podocytes acting through an LR-mediated Rho-signaling mechanism and to define the Tat binding motif regulating this process in podocytes harvested from children with HIVAN.

Results

Extracellular Tat Increases the Ability of FGF-2 to Induce Rho-A/Phospho–Extracellular Signal-Regulated Kinase Signaling in Cultured Primary Podocytes Harvested from Children with HIVAN

To determine how extracellular Tat and FGF-2 modulate Rho-A signaling in children with HIVAN, we cultured primary podocytes from the urine of these patients (Figure 1, Supplemental Figure 1). We found that Tat, alone or combined with FGF-2, induced the expression of Rho-A, phospho–extracellular signal-regulated kinase (pERK), and phospho-MLC₂ (pMLC₂) in these cells (Figure 2). These changes were partially inhibited by the exoenzyme C3-transferase from Clostridium botulinum, which inhibits Rho-proteins by ADP-ribosylation in the effector binding domain of the guanosine 5′‑triphosphatase (Figure 2), and did not involve other HIV-1 genes, because the cells were not infected. Similar results were reproduced in the cell lines generated from the primary podocytes (Figure 3A), validating their clinical relevance for the purpose of this study.

Figure 1.

Podocytes cultured from the urine of children with HIVAN express WT-1, synaptopodin, and nestin. A–D show light microscopy images of (A and B) primary podocytes cultured from the urine of two children with HIVAN and (C and D) two cell lines derived from the primary podocytes. (E and G) Immunohistochemistry staining identified the nuclear localization of the WT-1 antigen in cultured podocytes (red). (F and H) No specific WT-1 staining was detected in cells incubated with nonimmune IgG. (I and K) By immunofluorescence microsocopy, synaptopodin was detected in red in the cytoplasm of cultured podocytes, whereas cell nuclei were counterstained in blue with diamino-2 phenylindol. (J and L) No specific synaptopodin staining was detected in podocytes incubated with irrelevant isotypic Igs. (M and O) Nestin was detected in red in the cytoplasm of podocytes, showing both a partial filamentous and homogenous pattern. Cell nuclei were counterstained in blue with diamino-2 phenylindol. (N and P) Nestin was not detected in podocytes incubated with irrelevant isotypic Igs. Original magnification, ×200 in A and E–P; ×400, B.

Figure 2.

Extracellular Tat increases the ability of FGF-2 to induce Rho-A/pERK signaling in primary cultured podocytes isolated from the urine of children with HIVAN. Primary podocytes were starved overnight in serum-free media and treated with Tat (100 ng/ml) alone or combined with FGF-2 (50 ng/ml). The Rho-A inhibitor C3-transferase (20 ng/ml) was added 4 hours before stimulation. The cells were treated for 5 minutes as described above and then harvested to assess the phosphorylation of Rho-A, pERK, and pMLC as described in Concise Methods. The graph shows mean±SEM corresponding to three different experiments. Results were expressed in OD units expressed as a ratio of the total activity. Values significantly different from the corresponding control group are marked: *P<0.05; **P<0.01.

Figure 3.

Extracellular Tat increases the ability of FGF-2 to induce Rho-A/pERK signaling in a podocyte cell line generated from a child with HIVAN. (A) Cultured podocytes were starved overnight in serum-free media and treated with Tat (100 ng/ml) alone or combined with FGF-2 (50 ng/ml). The Rho-A inhibitor C3-transferase (20 ng/ml) was added 4 hours before stimulation. The cells were treated for 5 minutes as described above and then harvested to assess the phosphorylation of Rho-A, pERK, and pMLC as described in Concise Methods. The graph shows mean±SEM corresponding to three different experiments. Results were expressed in arbitrary OD units as a ratio of the total activity. Values significantly different from the corresponding control group are marked: *P<0.05; **P<0.01. (B) Cultured podocytes were treated as in A, and 20 minutes after treatment, F-actin fibers were visualized in cells by staining with 2 µg/ml Alexa Fluor 488–labeled phalloidin. Cell nuclei were stained with Hoechst 33342. Scale bar, 10 µm.

C3-Transferase Inhibits Tat+FGF-2–Induced Stress Fiber Formation and Rho-Activation in Cultured Podocytes

Follow-up experiments were done to determine how Tat+FGF-2 modulate the formation of stress fibers in the presence and absence of C3-transferase. As shown in Figure 3B, incubation of podocytes with Tat+FGF-2 resulted in a remarkable increase in stress fibers. These structures appeared as thick actin cables running along the length of the cell. In contrast, control cells or Tat+FGF-2–treated cells incubated with C3-transferase showed a more rounded shape, exhibited F-actin in the periphery of the cells, and formed fewer stress fibers on Tat+FGF-2 stimulation. The activation of pMLC₂ was also inhibited by the C3-transferase (Figure 3A). These results suggest that Rho-A activation mediates the cytoskeletal changes induced by Tat+FGF-2 in cultured podocytes.

Tat Preferentially Associates with LRs in Cultured Human Podocytes

To investigate the intracellular distribution of Tat in cultured podocytes, we generated recombinant adenoviral vectors (rAds) carrying a cDNA fragment encoding the full-length Tat protein and the FLAG peptide sequence. As discussed in Concise Methods and Supplemental Figure 2, the cDNA fragment encoding Tat was derived from cultured renal tubular epithelial cells infected with HIV-1.²⁹ This Tat variant, named Tat-HIVAN, was conserved in the N terminus but not the polymorphic C-terminal region downstream of the basic domain, and it was missing the RGD motif in the C-terminal region because of a frameshift deletion (Supplemental Figure 2). To validate the proper function of the Tat-HIVAN, we confirmed its ability to translocate to the nucleus in cultured podocytes and transactivate the HIV long terminal repeat in GHOST (GFP–human osteosarcoma) cells carrying a green fluorescent protein (GFP) reporter system³⁰ (Supplemental Figures 3 and 4). Subsequently, using flotation sucrose gradients, we isolated LRs from cultured podocytes transduced with either rAd-Tat-FLAG or rAd-GFP control vectors. Using both FLAG and Tat antibodies, we found that a significant proportion of Tat was preferentially localized in fraction 4 of the DRMs, whereas additional Tat was detected in detergent-soluble fractions (non-DRMs) (Figure 4A). Caveolin-1, a protein known to associate with LRs, was almost exclusively detected in DRMs fraction 4, confirming that an efficient isolation of LRs was achieved (Figure 4). In contrast, GFP was not detected in DRMs (Figure 4B), but it was located in the bottom of the gradient, which is composed of non-DRMs that contain detergent-soluble cytoplasmic and nuclear components. Similar results were also found in cultured primary podocytes harvested from children with HIVAN (Figure 4C).

Figure 4.

HIV-1 Tat protein is associated with LRs in human podocytes isolated from the urine of children with HIVAN. LRs were isolated from a cultured podocyte cell line (P2) infected with adenovirus carrying (A) FLAG-tagged Tat (Ad-Tat-FLAG) or (B) GFP (Ad-GFP) or (C) primary podocytes infected with Ad-Tat-FLAG by sucrose gradient–based flotation assay. Twelve fractions of the gradient were run by SDS-PAGE to show the distribution of DRM and non-DRM fractions. FLAG-tagged Tat, GFP, and the LR marker Caveolin-1 (Cav-1) were analyzed by Western blotting.

Arginines in the Tat Basic Domain Sequences Mediate LR Association in Podocytes

The basic domain of Tat, (RKKRRQRRR)_49–57 (also called protein transduction domain), is known to mediate membrane crossing for protein transduction. In addition, arginines of the basic domain have been reported to specifically interact with negatively charged lipids in the plasma membrane.³¹ Therefore, we explored whether this domain is necessary and sufficient to mediate Tat association with LR in cultured podocytes. As shown in Figure 4A, we observed that two different wild-type (WT) –Tat variants, carrying or missing the RGD motifs, were localized in the LR domains and that the presence or absence of the RGD motif did not affect this localization. In contrast, we found that six alanine substitutions (AKKAAQAAA)_49–57 carried by a mutant Tat, called Tat–basic domain mutant 1 (Tat-BDM1-HIVAN), almost completely abolished Tat association with LRs in cultured podocytes (Figure 5). Interestingly, partial substitutions of three arginines (_55–57) did not affect the recruitment of Tat (Tat-BDM2-HIVAN) to the LRs (Figure 5B). Alternatively, another Tat mutant (Tat-BDM3-HIVAN) carrying partial substitutions of the other three arginines (_49,52,53) showed a poor association with LRs (Figure 5B). Indeed, the OD ratio of DRM bands over non-DRM bands for Tat-BDM3-HIVAN was 5-fold lower than the ratio of Tat-BDM2-HIVAN or Tat-WT-HIVAN (data not shown). These data suggest that positions of arginines present in the Tat basic domain are important for its association with LRs.

Figure 5.

Arginines located in the basic domain are essential for targeting Tat to the LRs in cultured podocytes. (A) Several mutations were introduced in the basic domain of the WT Tat isolated from a child with HIVAN (Tat-HIVAN) to generate different Tat BDMs named for the purpose of this study: Tat-BDM1, Tat-BDM2, Tat-BDM3, and Tat-Y47A (mutation introduced in the putative cholesterol recognition consensus sequence). The Tat-101 control gene derived from the HIV-1 cDNA clone pCV1 (Supplemental Figure 2) carries the RDG motif, which is missing in Tat-HIVAN and all the other Tat BDMs. All these sequences were aligned using the Clustal Omega multiple sequence alignment program. Residues aligned identical are marked with an asterisk. Mutated residues are underlined. (B) Podocytes were transiently transfected with Tat-101/pCV1, Tat-HIVAN, Tat-HIVAN BDMs, or Tat-HIVAN-Y47A mutant. Cells were lysed with Triton X-100 and fractionated using a sucrose gradient flotation assay. FLAG-tagged Tat was detected by Western blot using the anti-FLAG antibody. (C) Western blot analyses showing changes in pERK, Rho-A, pMLC, and MMP-9 expression in isolated LRs and whole-cell lysates (WCLs) extracted from cultured podocytes transfected with WT Tat and Tat-mutated in BDM1 and the empty vector. GTP, guanosine triphosphate.

Because cholesterol is a major component of LRs, we also explored the potential role of the consensus cholesterol recognition/interaction amino acid consensus sequence (CRAC). The CRAC sequence is defined as (L/V–(X)_1–5–Y–(X)_1–5–R/K) and used by several proteins for binding cholesterol in the LRs, including Caveolin-1, HIV gp41, and human cytomegalovirus UL37 exon 1 protein.^32,33 We found that arginines_49,52,53 in the basic domain could form part of a potential CRAC motif (LGISYGRKKRR). Thus, because tyrosine in the CRAC motif is an essential residue for cholesterol binding,^32,33 we generated a mutant Tat with Y47A (Tat-Y47A) and tested its association with LRs. We found that the Y47A mutation did not affect the association of Tat with the LRs (Figure 5B). These findings indicate that the CRAC sequence alone is not sufficient to warrant cholesterol binding and that other structural factors are needed for this process.³²

Tat Basic Binding Domain Modulates Rho-A Signaling and the Expression of Matrix Metalloproteinase-9 in LRs

Rho-A signaling and matrix metalloproteinases (MMPs) modulate the migration of HIV-infected cells²⁶ and facilitate the release of FGF-2 in patients with Kaposi's sarcoma.^34,35 Children with HIVAN excrete high urinary levels of MMPs and FGF-2, which is in correlation with the progression of their renal disease.^14,15 Therefore, we explored the effect of Tat-HIVAN on Rho-A signaling and MMP-9 expression. We found that Rho-A and MMP-9 were localized in the LRs of the podocytes and that the basic domain of Tat increased Rho-A phosphorylation and MMP-9 expression in LRs (Figure 5C).

Tat Basic Domain Modulates FGF-2–Induced pERK, Rho-A, and pMLC₂ in Cultured Podocytes

Because FGF-2 is accumulated in the kidney of children with HIVAN,^{13,14,16,17,36} we explored how the Tat basic domain modulated FGF-2 signaling in podocytes transfected with Tat-WT-HIVAN or Tat-BDM1-HIVAN. As shown in Figure 6, we found that Tat-WT-HIVAN, alone or combined with FGF-2, induced a more significant activation of pERK, Rho-A, and pMLC₂ compared with Tat-BDM1-HIVAN. These changes are consistent with the results obtained in cultured primary podocytes using an extracellular Tat variant that carries the RGD motif (Figure 2). In contrast, Tat-BDM1-HIVAN was unable to increase the activity of FGF-2 (Figure 6). Overall, these findings confirm that the Tat basic domain plays a critical role modulating FGF-2 signaling in cultured podocytes and that the RGD motif is not essential for this process.

Figure 6.

The basic domain of Tat is essential for enhancing the FGF-2–induced activation of pERK, Rho-A, and pMLC2 and increasing the expression of MMP-9 in cultured podocytes. (A) Western blot analyses showing representative results in podocytes transiently transfected Tat-WT (WT) or Tat-BDM1 (BDM1) for 44 hours. Transfected cells were subsequently treated with FGF-2 (50 ng/ml) for 5 minutes and processed for the signaling studies. (B) The graph shows mean±SEM corresponding to five samples per group. Results are expressed in arbitrary OD units as a ratio of the total activity or normalized for β-actin for the MMP-9 expression. Values significantly different from the corresponding control group are marked: *P<0.05; **P<0.01.

Tat Basic Domain Modulates the Expression of pERK, Rho-A, pMLC₂, and MMP-9 in the Kidneys of Young HIV-Tg₂₆ Mice

To determine the role of the Tat basic domain in vivo, we used an adenoviral gene transferring technique developed in our laboratory³⁷ to express Tat in the kidney of young mice (Supplemental Figures 5 and 6). As shown in Figure 7, both WT and HIV-Tg₂₆ mice infected with Ad-Tat-WT-HIVAN showed a significant upregulation of renal pERK, Rho-A, pMLC, and MMP-9 compared with mice infected with Ad-Tat-BDM1-HIVAN. These changes were more remarkable in HIV-Tg₂₆ mice (Figure 7) and also associated with an upregulated expression of HIV-1 genes and the development of more severe renal lesions compared with HIV-Tg₂₆ mice injected with rAd-Tat BDM1-HIVAN or rAd-LacZ vectors (Supplemental Figures 5 and 6).

Figure 7.

The basic domain of Tat is important for inducing the activation of pERK, Rho-A, and pMLC2 and increasing the expression of MMP-9 in the kidney of WT and HIV-Tg₂₆ young mice. Newborn FVB/N WT and HIV-Tg₂₆ mice (n=3 per group) were infected with adenoviral vectors carrying Tat-WT (Ad-Tat-WT) or Tat-BDM1 (Ad-Tat-BDM1). Mice were euthanized after 7 days, and their kidneys were processed for the signaling studies as described in Concise Methods. (A) The Western blots show representative results corresponding to three mice in each group. (B) The graphs summarize the results corresponding to all the kidney samples per group. Results were expressed in arbitrary OD units as a ratio of the total activity or normalized for β-actin for the MMP-9 expression. Values significantly different from the corresponding control group are marked: *P<0.05; **P<0.01.

Circulating FGF-2 Induces the Expression of pERK, Rho-A, pMLC, and MMP-9 in the Kidney of Adult HIV-Tg₂₆ Mice

Because children with HIVAN show high plasma and urinary levels of FGF-2,^13,14 we used Ad-FGF-2 vectors carrying a secreted form of FGF-2 to determine whether circulating FGF-2 can induce similar renal signaling changes in adult WT and HIV-Tg₂₆ mice. In this experimental adult mouse model,^38,39 only the hepatocytes are infected with the adenoviral vectors, and FGF-2 released into the circulation is trapped in the kidney bound to HSPGs.^{13,15,16,36,40} As shown in Figure 8, FGF-2 increased the renal expression of pERK, Rho-A, pMLC₂, and MMP-9 in both WT and HIV-Tg₂₆ mice. These changes, however, were more remarkable in HIV-Tg₂₆ mice (Figure 8) and associated with the development of more severe renal lesions and albuminuria compared with all other groups (Figure 9, Supplemental Figure 7), Moreover, FGF-2, alone or combined with Tat, induced the proliferation and survival of cultured podocytes harvested from children with HIVAN acting through similar signaling pathways (Figure 10).

Figure 8.

Circulating FGF-2 induces the expression of pERK, Rho-A, pMLC2, and MMP-9 in the kidneys of WT and HIV-Tg₂₆ adult mice. Adult FVB/N WT and HIV-Tg₂₆ mice (n=3 per group) were infected with adenoviral vectors carrying the Lac-Z gene (rAd-Lac-Z) or a secreted form of human FGF-2 (rAd-FGF-2). Mice were euthanized after 28 days, and their kidneys were processed for the signaling studies as described in Concise Methods. (A) The Western blots show representative results corresponding to one experiment. (B) The graphs summarize the results corresponding to six kidney samples per group. Results were expressed in arbitrary OD units as a ratio of the total activity or normalized for β-actin for the MMP-9 expression. Values significantly different from the corresponding control group are marked: *P<0.05; **P<0.01.

Figure 9.

Circulating FGF-2 precipitated the development of HIVAN in adult HIV-Tg₂₆ mice. WT and HIV-Tg₂₆ male adult mice without preexisting renal disease were infected with adenoviral viral vectors carrying a secreted form of human recombinant FGF-2 or the Lac-Z gene (n=5 per group) as described in Concise Methods. Renal sections were harvested 28 days after the adenoviral infection. (B and D) Both WT and HIV-Tg₂₆ mice infected with rAd-FGF-2 showed a significant recruitment of glomerular and tubular epithelial cells expressing proliferating cell nuclear antigen (PCNA), a marker of DNA synthesis, replication, and intrinsic repair activity when compared to mice (A and C) injected with rAd-LacZ vectors. (H) HIV-Tg₂₆ mice infected with rAd-FGF-2 showed a decreased number of WT-1⁺ cells compared with all other groups (E–G) (*P<0.01; ANOVA). (D and H) These changes were associated with the presence of FSGS lesions, tubular casts, and microcysts. The bar graphs show the immunohistochemistry staining scores corresponding to each group. Results are expressed as percent changes in PCNA or WT-1⁺ cells relative to the control group (WT mice infected with rAd-Lac-Z). Values significantly different from the corresponding control group are marked: *P<0.05; **P<0.01. Original magnification, ×200 in A–D; ×400 in E–H.

Figure 10.

FGF-2, alone or combined with Tat, increased the proliferation and survival of cultured podocytes. (A) Proliferation assay. Podocytes were seeded at a density of 5×10⁴ cells/well and cultured in DMEM media supplemented with 1% FBS and penicillin/streptomycin for 4 days. HIV-Tat, FGF-2, or both combined were added daily at a concentration of 50 ng/ml each. On day 4, all cells were trypsinized and counted as described in Concise Methods. Values significantly different from controls are marked: *P<0.05. (B) The MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) survival assay. Podocytes were seeded at a density of 6×10⁴ cells per well and starved overnight on DMEM serum-free media containing antibiotics. Subsequently, all cells were treated with FGF-2+Tat at a concentration of 10 ng/ml each (black bars) in the presence or absence of the corresponding inhibitors. The kinase inhibitor PD98059 (5 µM), the Rho-associated protein kinase inhibitor Y27632 (10 µM), and the Rho-A inhibitor C3-transferase (20 ng/ml) were added 2 hours before the FGF-2+Tat treatment. On day 3, 10 µl MTT solution (5 mg/ml) was added to each well, incubated for 3 hours at 37°C, and then, treated with 100 µl MTT solvent (4 mM HCl and 0.1% Triton X-100 in isopropronal) as described in Concise Methods. Results were recorded in MTT absorbance units and expressed as a percent of control values (open bars) considering six independent readings. Values significantly different from controls are marked: **P<0.01.

Discussion

This study shows that Tat is preferentially recruited to LRs in podocytes harvested from children with HIVAN and induces cytoskeletal changes through the stimulation of the Rho-A/pMLC pathways. In addition, we found that alanine substitution of the six arginines in the Tat basic binding domain prevented the association of Tat with LRs, impaired its ability to enhance FGF-2 signaling or MMP-9 expression in cultured podocytes, and failed to induce severe renal disease in HIV-Tg₂₆ mice. Taken together, these findings provide compelling evidence to support the notion that the Tat basic domain is essential for the recruitment of Tat to LRs and the regulation of FGF-2 signaling in podocytes isolated from children with HIVAN.

The Tat protein is a powerful transcriptional factor encoded by two exons. The first exon encodes the activation domain, which interacts with cyclin T1, and the basic domain (amino acids 49–57), which is required for the nuclear localization of Tat,^41,42 HIV-1 transcription,⁷ and many other functions.^7,41,42 The second exon encodes the RGD motif (C-terminal amino acids 73–86), which enhances the angiogenic activity of Tat acting through integrin receptors.^43,44 However, the Tat protein derived from a child with HIVAN used in this study had an incomplete RGD sequence, and our results show that this sequence is not essential for the association of Tat with LRs or the regulation of FGF-2 signaling. Indeed, we found that Tat variants carrying the RGD motif were also recruited to LRs and induced FGF-2 signaling. Thus, our findings should be relevant for children infected with viruses carrying different Tat variants.

In contrast to the RGD motif, the Tat basic binding domain contains a cluster of basic residues (RKKRRQRRR) that are known to carry proteins and DNA molecules across the cell membranes and affect the activity of extracellular Tat peptides in human podocytes.^11,12 LRs were also reported to play a critical role modulating Tat activity in other cell types^26,45; however, these studies did not explore the specific localization of Tat in LRs. Moreover, to date, the interactions between Tat and LRs in podocytes are unclear, and the residues responsible for the association of Tat with LRs remain undefined. Here, we used purified LRs to show that three arginine residues in the basic domain are essential for the stable association of Tat with LRs and the regulation of FGF-2 signaling in cultured podocytes harvested from children with HIVAN. It is possible that Tat may be recruited to LRs microdomains by binding to anionic lipids that are enriched in the LRs.³¹ However, the basic residues are not equally important in membrane insertion or binding to polyphosphoinositides, because basic residues 49–51 but not arginines_55–57 are critical for this process.³¹ Similarly, we found that three alanine substitutions in arginines_49,52,53 but not arginines_55–57 diminished Tat association with LRs by over 80%. These results indicate that lack of positive charges in the Tat-BDM1-HIVAN per se cannot fully explain its defective association with LRs. In that regard, there may be conformation-specific interactions between the basic domain and the LRs proteins to allow this stable association to occur.

Interesting findings of this study are the localization of MMP-9 within LRs in cultured podocytes and the ability of Tat-WT-HIVAN to induce the expression of MMP-9 in this location. MMP-9 belongs to a family of zinc binding endopeptidases that degrade extracellular matrix proteins, including HSPG and collagen.⁴⁶ Previous studies have shown that MMP-9 was associated with LRs in several cancer cell lines, where they modulate angiogenesis and cell migration.^47–49 In children with HIVAN, the urinary levels of MMP-9 are increased in correlation with the progression of their renal disease.^14,16 Therefore, because MMP-9 facilitates the release of FGF-2,³⁴ our findings may provide an alternative mechanism by which Tat can regulate the activity of FGF-2 in podocytes.

Also, Rho is a small guanosine triphosphate binding protein that plays a central role regulating the dynamic organization of contractile actin–myosin filaments and the formation of stress fibers in mammalian cells.⁵⁰ When bound to guanosine triphosphate, Rho-proteins activate the Rho-kinase and other downstream effector proteins, including the phosphorylation of the myosin binding subunit of MLC.⁵⁰ All these factors are important to maintain the cytoskeletal structure of podocytes and the integrity of the glomerular basement membrane.² In line with this notion, our data suggest that Tat, combined with FGF-2, can induce the crosslinking of F-actin and the formation of stress fibers in human podocytes through activation of Rho-A and pMLC signaling pathways. In addition, we found that Tat increased the FGF-2–induced phosphorylation of pERK, a pathway leading to cell proliferation. Podocytes are terminally differentiated cells and have limited in vivo ability to undergo nuclear division in response to FGF-2.⁵¹ Thus, differentiated podocytes that are forced to re-enter the cell cycle under the influence of Tat and FGF-2 may be unable to divide and could detach or undergo apoptosis. Considering that FGF-2 is accumulated in the circulation, glomeruli, and urine of children with HIVAN,^{13,14,16,17,36} it is tempting to speculate that the podocytes carrying the genetic risk variants that predispose to HIVAN^52,53 may be more sensitive to this pathogenic mechanism. In support of this notion, we found that both Tat and FGF-2 can precipitate the development of HIVAN in HIV-Tg₂₆ mice. However, HIVAN is a complex renal disease, and other viral proteins like Nef, which inhibits Rho-A activation in conditionally immortalized cultured murine podocytes,⁵⁴ also play a key role in this disease. Nonetheless, children may be more sensitive to the renal accumulation of FGF-2, because their kidneys are growing, and they have high expressions of HSPG, MMPs, FGF receptors, and binding proteins and higher plasma and tissue levels of FGF-2 relative to all other HIV-1 proteins.^{13,14,16,17,36} They also develop more severe immunosuppression and higher viral loads, leading to additional secretion of Tat and inflammatory cytokines that release more FGF-2.^6,34 Moreover, our data did not rule out the possibility that Tat, acting through the release of systemic cytokines that upregulate the expression of HIV-1 genes through NF-κB activation, may play an additional role in HIVAN as well. Both Tat and FGF-2 are cleared from the circulation and stored bound to HSPG,^6,18 and they can reach high tissue concentrations in organs with a high blood flow and HSPG content.^15,17,40

In conclusion, we found that Tat is preferentially recruited to LRs in cultured podocytes from children with HIVAN. This event is mediated by the Tat basic domain, because mutations of six arginines in this region abolished Tat association with LRs and impaired its ability to induce Rho-A activation, MMP-9 expression, and FGF-2 signaling both in vitro and in vivo. These findings may provide a molecular framework to define the role of LRs in the pathogenesis of childhood HIVAN and/or identify new therapeutic targets to improve the outcome of children with other HIV renal diseases.

Concise Methods

Collection of Human Samples

The collection of human samples was carried out in accordance with the principles of the Declaration of Helsinki. This study was approved by the Institutional Review Board of Children’s National Medical Center, and a waiver of Documentation of Informed Consent and Health Insurance Portability and Accountability Act Authorization was obtained to allow for anonymous data and specimen collection after a verbal agreement was obtained from the patients or their parents.

Construction of Tat Expression Vectors and Adenoviruses

The cDNA fragment encoding Tat was derived from renal epithelial cells harvested from the urine of a child with HIVAN and infected with PBMCs isolated from the same child as previously described.²⁹ Infected renal epithelial cells were then lysed using TRIzol (Invitrogen) to isolate total mRNA. The Tat gene was amplified from the subsequently synthesized cDNA with the high-fidelity DNA polymerase Pfu (Invitrogen) using PCR primers flanking the open reading frame of the full-length Tat.⁵⁵ A cDNA fragment encoding the full-length Tat protein was cloned into the pCXN2-FLAG vector and used to generate E1-deleted recombinant adenoviruses as previously described.⁵⁶ Both Tat-FLAG and GFP adenoviruses were purified, desalted, and titrated as described before.⁵⁷ To generate adenoviruses that express secreted Tat protein, the signal peptide for secretion was amplified by PCR from the sp-FGF4:FGF-1_1–154–pMEXneo vector⁵⁸ using forward 5′-ATACTCGAGATGGCGGGGCCCGGGACGGC-3′ and reverse 5′- GCGAAGCTTGGGCGCCAGCAAGGCCAGCAG-3′ primers. The PCR product of this 78-bp signal peptide was then ligated with the PCR product of the full-length Tat-FLAG (WT or BDM1) from the pCXN2 vector using forward 5′-ACTAAGCTTGACTACAAGGACGACGATGA-3′ and reverse 5′-TACGGATCCCTAACTAGCTAATCGAATCG -3′ primers. The resulting recombinant Tat-WT-FLAG or Tat-BDM1-FLAG gene with the FGF4 signal peptide sequence was cloned into the pVQAd CMV K-NpA shuttle plasmid (provided by ViraQuest, Inc., North Liberty, IA) to generate Tat adenoviruses (rAd-Tat-WT* and rAd-Tat-BDM1*) as described previously.⁵⁹

Podocyte Cultures

Primary podocytes were isolated from clean-catch urine as previously described.^29,60 Proliferating podocytes were obtained consistently from the urine of two children with HIVAN. Primary colonies that showed typical podocyte morphology (Figure 1, A–E), were further characterized by immunohistochemistry or immunofluorescence with the podocyte markers Wilms' tumor 1 (WT-1), synaptopodin, and nestin (Figure 1, F–P). Cells were fixed in 4% paraformaldehyde permeabilized with 0.1% Triton X-100 (Sigma-Aldrich) and stained using well established methods with specific antibodies against WT-1 (Dako), synaptopodin (Maine Biotechnology, Inc.), and nestin (Chemicon Int., Inc.). Subsequently, selected colonies were transduced with adenoviral vectors carrying DNA sequences encoding the SV-40 large T antigen (provided by Janice Chou, National Institutes of Health, Bethesda, MD)⁶¹ and human telomerase (rAd-TERT; Applied Biologic Materials, Inc.). Colonies of transformed podocytes were characterized again by RT-PCR using specific primers for the podocyte markers WT-1, synaptopodin, podocalyxin, nestin, podocin, and nephrin as described in previous studies^60,62 and shown in Supplemental Figure 1 and Supplemental Table 1. Podocyte colonies expressing all these markers, with the exception of podocin (which was not detected in any of these colonies), were selected and tested again by immunohistochemistry with WT-1, synaptopodin, and nestin antibodies (Figure 1, G, K, and O). Two podocyte colonies, named for the purpose of this study as P-2 and P-3 (Supplemental Figure 1), were expanded and cultured in DMEM supplemented with 10% FBS and 1% antibiotic–antimycotic (Invitrogen), which contains 100 units/ml penicillin, 100 µg/ml streptomycin, and 0.25 µg/ml Fungizone (amphotericin B). Western blots were done to confirm the expression of WT-1, synaptopodin, and nestin using antibodies from Dako (6F-H2), Santa Cruz Biotechnology (P-19), and Chemicon (MAb5326), respectively. Control sections were stained omitting the first antibody and using nonimmune IgG. All podocytes clones were screened and tested for the presence of HIV-1 genes by PCR as previously described.²⁹

Mutagenesis

Site-specific mutations were introduced to the Tat gene using the QuikChange II XL Site-Directed Mutagenesis Kit (Agilent Technologies, Santa Clara, CA) following the manufacturer’s manuals. To generate mutations in the basic domain RKKRRQRRR, forward primer 5′-GGCAGGAAGAAGCGGAGACAGGCAGCAGCAGCTCCTCAAGACAGTCAGAC-3′ and reverse primer 5′-GTCTGACTGTCTTGAGGAGCTGCTGCTGCCTGTCTCCGCTTCTTCCTGCC-3′ were first used to mutate arginines_55–57 and generate a mutant Tat (RKKRRQAAA)_49–57. Forward primer 5′-TTAGGCATCTCCTATGGCGCGAAGAAGGCGGCACAGGCAGCAGCAGCTCC-3′ and reverse primer 5′-GGAGCTGCTGCTGCCTGTGCCGCCTTCTTCGCGCCATAGGAGATGCCTAA-3′ were then used to produce mutant Tat with (AKKAAQAAA)_49–57. Forward 5′-CTTAGGCATCTCCTATGGCGCGAAGAAGGCGGCACAGCGACGAAGAGCTCCT-3′ and reverse 5′-AGGAGCTCTTCGTCGCTGTGCCGCCTTCTTCGCGCCATAGGAGATGCCTAAG-3′ were used to generate mutant Tat with (AKKAAQRRR)_49–57. Forward primer 5′-AGGCTTAGGCATCTCCGCTGGCAGGAAGAAGCGG-3′ and reverse primer 5′-CCGCTTCTTCCTGCCAGCGGAGATGCCTAAGCCT-3′ were used to introduce the Y47A mutation in the potential CRAC sequence, LGISYGRKKRR.

Extracellular Treatments, Adenovirus Infections, and Transient Transfection

Podocytes were exposed for 5 minutes to the control buffer, 100 ng/ml HIV-1 Tat protein (catalog number 2222; National Institutes of Health AIDS Reagent Program),⁶³ or 50 ng/ml recombinant human FGF-2 (R&D Systems) alone or combined with HIV-Tat and harvested for the signaling or immunofluorescence studies. Both Tat and FGF-2 preparations were screened for endotoxin contamination (<0.1 ng/μg protein). As indicated, C3-transferase toxin (20 ng/ml) was added 4 hours before stimulation to block Rho-A activity. Alternatively, podocytes were infected with 2–4×10⁹ particles/ml rAd-Tat or rAd-GFP vectors (Quantum Biotechnologies, Inc.) for 24–30 hours and then harvested for LRs isolation. For transfection, podocytes were transiently transfected with pCXN2-Tat-FLAG, pcDNA3.1-TAT-1–101-FLAG⁶⁴ (Addgene), which was originally derived from the HIV-1 cDNA clone pCV1,⁶⁵ or the corresponding Tat mutant or control vectors using Lipofectamine 2000 (Invitrogen) and harvested 36–48 hours later for LR isolation as described before.³³ Cell proliferation and survival assays were done as previously described.³⁶

LR Isolation

LR fractions were isolated using the flotation assay as described previously.³³ In brief, podocytes were lysed on ice, homogenized, and sonicated using a Polytron tip sonicator. The cell lysate was then mixed with an equal volume of 80% sucrose dissolved in the isolation buffer and overlaid with discontinuous layers of 35% and 5% sucrose solutions at a volume ratio of 1:2:1 in a 4-ml ultracentrifuge tube (Beckman). The tubes were loaded on a Beckman SW60i rotor, and gradients were separated by centrifugation at 187,813×g (39,000 rpm) for 16 hours at 4°C. After centrifugation, 12 fractions (about 330 μl each) were sequentially collected from top to bottom, and 25 μl of each fraction were resolved by SDS-PAGE and analyzed by Western blot.

Western Blot Analyses

Cells were lysed using RIPA lysis buffer containing protease inhibitor and phosphatase inhibitor cocktail 2 (Sigma-Aldrich) and processed for Western blots as described before.⁶⁶ The following primary antibodies were used: phospo-p44/42 mitogen-activated protein kinase (Thr202/Tyr204), p44/42 mitogen-activated protein kinase (ERK1/2), phospho-MLC2 (Thr18/Ser 19), and total MLC2 from Cell Signaling Technology, anti–MMP-9 (Calbiochem), β-actin (AC-15; Sigma-Aldrich), α-tubulin (DM1A; Abcam), Caveolin 1 (BD Biosciences), FLAG (M2; Sigma-Aldrich), GFP (Santa Cruz Biotechnology), and Tat Rabbit pAb (catalog number 705; National Institutes of Health AIDS Reagent Program).⁶⁷ Protein bands were detected using Supersignal West Pico Chemiluminescent Substrate (Thermo Scientific) or an ECL Western Blot Detection Kit (GE Healthcare) following the manufacturer’s instructions.

Glutathione S-Transferase Pull-Down Assays

Rho-A activation was monitored by glutathione S-transferase (GST) pull-down using GST-Rhotekin recombinant protein bound to glutathione slurry resin using the protocol described before.⁶⁶ Total and phosphorylated Rho-A were assessed by Western blotting using Rho-A (67B9) rabbit mAb (Cell Signaling Technology). For total Rho-A, we used an equal amount of protein corresponding to the cell lysates obtained before they were mixed with the GST beads.

Immunofluorescence Staining

Podocytes were cultured on cell growth-promoting coverslips (Fisher) or coverslips coated with type I collagen for 24 hours. Cells were treated as indicated and then fixed and permeablized as described before.⁶⁶ After blocking with 1% BSA, cells were stained with 2 µg/ml Phalloidin–Alexa Fluor 488 (Invitrogen), and nuclei were stained with Hoechst 33342 (1:2000; Invitrogen) for 20 minutes at room temperature. After mounting, the confocal imaging was performed using an Olympus FV1000 confocal microscope, and a magnification of 60× was used for imaging.

Studies in WT and HIV-Tg₂₆ Mice

These experiments were approved by the Children’s Research Institute Animal Care and Use Committee. WT and HIV-Tg₂₆ FVB/N mice were housed in a pathogen-free environment on a 12:12-hour light/dark cycle in the animal facility at Children’s National Medical Center. All mice had free access to water and standard food and were treated in accordance with the National Institutes of Health guidelines for care and use of research animals. Both the generation of the HIV-Tg₂₆ mouse colony and the adenoviral gene-transferring technique to express foreign genes in newborn mouse kidneys have been described in detail.^37,68,69 Briefly, newborn FVB/N WT and HIV-Tg₂₆ mice (n=3–8 per group) were injected through the retro-orbital plexus with adenoviruses carrying the Escherichia coli LacZ gene (rAd-LacZ), Tat-WT-FLAG, or Tat-BDM1-FLAG (1×10⁸ pfu/pup). Tat mRNA expression was assessed by RT-PCR using the following primers: forward 5′-ATGGAGCCAGTAGATCCTAGAC-3′ and reverse 5′- CTAATCGAATCGATCTGTCTCTGC-3′. In other experiments, WT or HIV-Tg₂₆ adult male mice without preexisting renal disease were divided in two groups (n=3–5 mice per group) and injected through the retro-orbital vein plexus with adenoviral vectors carrying LacZ or a 700-bp cDNA sequence encoding a secreted form of human FGF-2 (rAd-FGF-2; 5×10⁸ pfu/mouse) as previously described.^38,69,70 Mice were euthanized 7 or 28 days after the adenoviral infection, and their kidneys were processed for signaling or renal histologic studies as described above. Renal injury was assessed in sections stained with period acid–Schiff by counting the percentage of glomeruli exhibiting segmental/global sclerosis and the percentage of tubular casts and microcysts. In addition, we counted the number of cells that stained positive for proliferating cell nuclear antigen and WT-1 and assessed the magnitude of albuminuria as previously described.³⁸

Statistical Analyses

If not specified otherwise, the data are expressed as means±SEMs. Multiple sets of data were compared by ANOVA with Newman–Keuls post hoc comparisons. The significant differences between the means of two groups were analyzed by unpaired t tests. Statistical analyses were performed using GraphPad Prism software (version 5.00; GraphPad Software, San Diego, CA). Values of P<0.05 were considered statistically significant.

Disclosures

None.