We report convergent structural, biochemical and cell biological observations supporting the conclusion that HIV’s Nef protein interacts with human β-catenin. A sequence near the C-terminus of Nef is similar to the central β-catenin binding motif in diverse β-catenin ligands. Isolated peptide structures exhibiting this sequence and its variations in Nef have a high degree of 3D structural compatibility with, and indeed a biophysical preference for, the molecular surface pocket on β-catenin, in which known β-catenin ligands bind (shown by unconstrained computational molecular docking). The distribution of the β-catenin binding motif in Nef across diverse HIV-1, SIV and HIV-2 strains suggests conservation of the interaction site. The interaction is specifically detected in vitro and in HEK293 cells. Finally, the interaction is functionally significant in HEK293 cells: the endogenous Wnt signaling pathway in this human cell line is affected by Nef and this influence disappears when Nef positions 186 and 191, which mirror key residues in known β-catenin ligands, are mutated.
The 3D structural and informatics results strongly suggest that the interaction is relevant for HIV/AIDS, as the motif appears across HIV strains, suggesting functional conservation in the virus. However, the HEK293 cell line in which we demonstrated the cellular activity is not directly relevant to HIV/AIDS. Nevertheless, endogenous human β-catenin was bound in these cells. This fact, along with the structural results, increases the probability that Nef may play a role in AIDS pathogenesis.
From a cell biology point of view, the probability that Nef interacts with β-catenin leads to intriguing speculation that β-catenin-mediated functions in T-cells and macrophages are involved in HIV pathophysiology. Nef is already known to be associated with T-cell chemotaxis, which is the driving force of T-cell extravasation to infected tissues [42
]. Interestingly, β-catenin is involved directly in the physical process of T-cell extravasation: stabilized β-catenin protein in T-cells directly targets matrix metalloproteinase (MMP) promoters through tandem TCF sites and MMP expression augments T-cell transmigration [29
]. In addition, β-catenin stabilization in T-cells enhanced survival of CD4+/CD25+ T-regulatory cells (Treg) while CD4+/CD25- T-non-regulatory cells (Tnon-reg) became anergic and proliferated poorly in response to CD3 antibody stimulus [46
]. A mirroring pattern of immune activation appears to occur during HIV/SIV infection [47
]. Finally, recent studies show that β-catenin signaling programs dendritic cells in the gut into a tolerogenic state, limiting the inflammatory response there [49
]. The gut is the principal site where HIV-1 replicates [50
If it is physiologically relevant, the β-catenin interaction site is predicted by to be highly prevalent in HIV subtype B, SIV and HIV-2, while it is nearly absent in the other HIV subtypes. Although differences in anti-retroviral treatment confound analysis, there appear to be phenotypic differences between B and non-B HIV subtypes [56
]. There are clear phenotypic differences between HIV-1 and HIV-2, as well as between SIV and HIV. Our findings raise the possibility that Nef’s interaction with β-catenin may contribute to phenotypic differences observed between B and non-B subtypes in HIV-1, and to the phenotypic differences between HIV-1, HIV-2 and SIV. If this were true, the exact contributions and specific phenotypes in question would require further investigation.
How might Nef influence β-catenin signaling in vivo
? At first approximation, our results suggest that Nef can compete for the same site occupied by TCF/LEF on β-catenin, thereby inhibiting TCF-based transcription (). However, Nef is primarily (although not exclusively) localized to the cytoplasm [58
], while the TCF- β-catenin complex is active in the nucleus of cells [24
]. This raises the possibility, even though we have observed functional inhibition of TCF-based transcription, that Nef could affect β-catenin stabilization in the cytoplasm, compete with E-cadherin at cell-cell junctions [60
] or even compete with ICAT [61
] in the nucleus resulting in increased
TCF-based transcription in T-cells or macrophages (which may be fundamentally physiologically different from the HEK293 cells used in this study). Fractionation studies performed in our lab in HEK293 cells (Figure S2
) show that in this particular cell line, Nef localization is limited to the cytoplasm with only traces in the nucleus. Nevertheless, the predominant in vivo
functional effect, if any, of Nef on T-cells and macrophages infected with HIV could diverge significantly from our observations in HEK293 cells.
Minor informatics improvements enabled our detection of the Nef-β-catenin interaction. Searching for new ligands to β-catenin with a motif, rather than using standard sequence-sequence alignment methods enabled us to identify new ligands. In 1993, Shugars et al. identified four Nef-defining sequences, referred to as “blocks”. The blocks were aligned with host proteins in order to shed light on the function of Nef. The C terminal block, overlapping with the β-catenin binding motif (), did not align with host proteins and its function remained unclear [62
]. The use of a motif in contrast to sequence-sequence
alignment is therefore more productive and this aligns with work done by others [63
]. In addition, 3D structural information is encapsulated within our motif, which further increases the sensitivity of the motif.
Structurally, the match between Nef and β-catenin is very strong, with the local sequence of Nef at the location of the β-catenin motif easily adopting, and actually preferring, a structural conformation that fills the key Asp and Phe/Tyr restricted pockets on β-catenin (). In addition, although the major hotspot of the interaction was determined to be this small segment of Nef at 186-191, Nef binds a variety of armadillo repeat proteins as a whole domain [10
]. Accordingly, there may be other distributed contact points on the Nef and concave β-catenin surfaces that contribute to the interaction. Indeed, a nuclear receptor [71
] is known to bind to β-catenin as a whole domain within the concave surface of its armadillo repeats. In order to visualize whether such an interaction could be consistent with our findings, we built a theoretical model of the whole Nef domain bound to β-catenin via the segment at 186-191. Our model shows that the C-terminal tail of Nef after position 185 must detach from the core Nef domain and unfold in order to assume the extended conformation predicted by our studies and make the key β-catenin interactions (). The remaining core domain almost perfectly fills the volume of the concave surface of β-catenin. A highly flexible loop in Nef that is important for its association with adaptor proteins (marked with an arrow in ) does not clash with β-catenin in our model. reveals that, if Nef were to bind to β-catenin in this manner, it would not interfere with Nef dimerization or with other Nef binding sites such as that for the SH3 domain from Fyn.
Model of Nef in situ on β-catenin.
The Wnt pathway is an intrinsic molecular mechanism to limit HIV replication in PBMC’s (peripheral blood mononuclear cells) and in astrocytes [72
]. It was shown that HIV replication is repressed when TCF-4 transcription factor binds the HIV long terminal repeat (LTR)[73
]. At least four such TCF-4 binding sites have been identified in the LTR. The -143 site has garnered attention since it (i) has 100 % homology to the TCF-4 core (5′-(A/T)(A/T)CAAAG-3′), (ii) it is present in approximately one-third of 500 HIV LTR sequences from the the Los Alamos gene bank, (iii) it has the highest affinity for TCF-4 (iv) SMAR1, a nuclear matrix binding protein, was shown to complex with β-catenin/TCF-4 at the -143 site to facilitate transcriptional repression from HIV promoters[74
]}. In addition, it was shown that expression of TCF-4 in human astrocytic cells decreased the basal and Tat-mediated transcription of the HIV-1 LTR [77
]. TCF-4/β-catenin repression of basal LTR activity likely prevents Tat from reaching a threshold level which would allow it to tether on the TAR region of the LTR in association with a positive elongation complex (pTEFb) to accelerate the rate and efficiency of HIV transcription[75
TCF-4 and β-catenin regulate the expression of other transcription factors relevant to HIV transcription. β-catenin and TCF-4 inhibit C/EBP β/δ tethering on the HIV LTR, suggesting that β-catenin and TCF-4 cooperate in this repression. TCF-4, independent of β-catenin, also negatively regulates NFκB tethering on the LTR [78
]. In astrocytes, NFκB suppression is mediated by TCF-4 without the involvement of β-catenin while in other cells, β-catenin suppresses NFκB activity [79
]. TCF-4 suppression of NFκB may be mediated by direct interaction between TCF-4 and NFκB or an indirect effect on upstream regulators of NFκB. Both C/EBP and NFκB are inducers of HIV promoter activity and thus β-catenin/TCF-4 inhibition of these inducers could also contribute to the overall mechanism by which the Wnt/β-catenin pathway represses HIV transcription and replication.
These data correlate with our reporter assay results in which Nef inhibits transcription from a luciferase reporter with multiple TCF binding sites. Our results suggest that Nef prevents TCF-4 from repressing HIV replication via its action on β-catenin. Extrapolating our findings from HEK293 cells to immune cells in vivo
, it is possible, though speculative, that the Wnt pathway may mediate the lowered viral load and lack of progression associated with Nef deletion [1
We have detected a previously unrecognized interaction between the HIV-1 protein Nef and human β-catenin, part of the Wnt signaling pathway. This finding potentially implicates β-catenin and the Wnt signaling pathway in T-cell transmigration defects and immune activation phenomena observed during the development of AIDS from HIV infection.