A structure-based mechanism for HEXIM displacement from 7SK

Productive transcriptional elongation of many cellular and viral mRNAs requires transcriptional factors to extract pTEFb from the 7SK snRNP by modulating the 11 association between the HEXIM protein and the 7SK snRNA. Here we report the 12 structure of the HEXIM arginine rich motif in complex with the apical stemloop-1 of 13 7SK (7SK-SL1 apical ) and detail how the HIV transcriptional regulator Tat from various 14 subtypes overcome the structural constraints required to displace HEXIM. While the 15 majority of interactions between 7SK and HEXIM and Tat are similar, critical 16 differences exist that guide function. First, the conformational plasticity of 7SK 17 enables the formation of three different base pair configurations at a critical 18 remodeling site, which allows for the modulation required for HEXIM binding and its 19 subsequent displacement by Tat. Furthermore, the specific sequence variations 20 observed in various Tat subtypes all converge on remodeling 7SK at this region. 21 Second, we show that HEXIM primes its own displacement by causing specific local 22 destabilization upon binding — a feature that is then exploited by Tat to bind 7SK 23 more efficiently. Overall, our study details the molecular environment presented by HEXIM and uncovers a destabilization-driven displacement strategy that increases the 25 conformational sampling of 7SK-snRNP, which may allow diverse transcriptional 26 factors to competitively regulate pTEFb.


Preformed configurations of ASM1 and ASM2 provide a common mode of interaction
the G74 and G70 bases of ASM1 and ASM2 to the H and the Hδ protons confirm that 152 consecutive arginines R156 and R155 interact in a ladder-like configuration with the 153 tandem preformed motifs ASM1 and ASM2, respectively ( Fig. 3a and Supplementary Fig.   154 4c). Such NOEs are also observed in both the Tat Fin and the Tat G -bound complexes, 155 confirming the similar placement of the C-terminal R57 and R56 into the tandem ASM1 156 and ASM2, respectively ( Fig. 3a and Supplementary Fig. 5a, b, d, e). Taken together, the 157 structures reveal a common mode of interaction between the non-varying C-terminal 158 arginines and the tandem ASMs. Rearrangement of pseudo-ASM3 allows for HEXIM N-terminal interactions. 161 In the free 7SK-SL1 apical , pseudo-ASM3 and ASM4 adopt a pseudo-symmetrical 162 architecture where the two motifs are spatially opposed. Upon HEXIM N-ARM binding, the 163 pseudo-ASM3 maintains its U40:A43-U66 triple base interaction although the base of the 164 sandwich, A39, rearranges from a reverse Hoogsteen interaction with U68 into a cis 165 Hoogsteen/sugar interaction, giving rise to an alternate pseudo configuration. (Fig. 3b,f). 166 This is evidenced both by NOEs from the U68 imino proton to the A39 amino protons and 167 NOEs from the U68 H2′ and H3′ protons to the A39 H8 proton ( Supplementary Fig. 6b). 168 This frees up the U68 imino proton to engage the backbone carbonyl of K152 while 169 simultaneously bringing the N1 proton acceptor of A39 into the major groove to hydrogen 170 bond with the side chain Hε protons of K151 (Fig. 3b). Thus, both K151 and 152 can enter 171 deep into the major groove by remodeling the pseudo-ASM3. 172 The amino side chain of K151 is within hydrogen-bonding distance of the A39 N1  Fig. 4d, e). Taken together, these data show that despite the lack of 192 arginines, the lysine-rich N-terminus of HEXIM N-ARM is able to be accommodated by 7SK: 193 the Watson-Crick face of A39 turns from the minor into the major groove to interact with an additional arginine docks into the preformed ASM4. While the mechanism of 202 remodeling pseudo-ASM3 is conserved upon binding of both Tat Fin and Tat G ARMs ( Fig.   203 3c, e, g, h; Supplementary Fig. 6), both the drivers of the conformational switch and the 204 engagement of the ASM4 vary depending on differences in amino acid sequences.

205
While Tat Fin has two major differences from Tat NL4-3 (K53 to R53 and spacer H54 to 206 Q54, respectively), it only differs in a single amino acid from HEXIM (R52 and K151, 207 respectively). Like Tat NL4-3 , R52 is responsible for remodeling pseudo-ASM3 ( shows that the K51 amino side chain is positioned to hydrogen-bond with the U63 ribose 217 ring in a stabilizing interaction (Fig. 3c). This is evidenced by NOEs between the K51 Hδ 218 protons with the U63 H5 and H1′ protons and the K51 Hε protons with the U63 2′ hydroxyl 219 proton ( Supplementary Fig. 5c). Furthermore, the N-terminal K50 exits near the apical 220 loop with NOEs observed between the K50 Hδ and Hε protons with the C38 and C37 H5 221 and H1′ protons position the amino side chain of K50 to the C38 phosphate backbone 222 ( Supplementary Fig. 5a). the C35, C36, and C37 H5 protons, placing H54 near ASM2 whereas the R55 (R154 in HEXIM) positioning this spacer residue near ASM1 (Fig. 3i, j; Supplementary Fig. 4, 5). This is in 229 contrast with the binding mode of Tat NL4-3 in which the intercalation of R53 into ASM4 230 drags both the Q54 and R55 spacer residues towards the apical ASMs.

231
The importance of the histidine H54 spacer is even more evident in the Tat G strain 232 where it represents the only difference between Tat NL4-3 . This single difference changes 233 the identity of the arginine that remodels pseudo-ASM3. In this ARM, the positioning of 234 H54 near ASM2 precludes R53 from reaching ASM4 to accomplish the inverse  Supplementary Fig. 6c). The destabilization of this region is also indicated by the line-broadening of K150, which interacts with U63 in the folded configuration ( Supplementary Fig. 4e).

253
The destabilization of 7SK-SL1 apical only by HEXIM is further evident when

278
The 7SK snRNP represents a central biomolecule that a wide range of 279 transcriptional factors need to interact with to access pTEFb to control transcriptional 280 elongation. In particular, pTEFb extraction by HIV Tat from this complex requires 281 manipulating the interaction between the 7SK snRNA and the HEXIM adapter protein.

282
In this study, we solved the structures of the RNA binding domains of HEXIM and Tat  Similarly, while ASM4 retains its preformed configuration found in the free state upon  The structures also provide insights into specific sequence variations that occur in 340 the highly conserved Tat ARM to displace HEXIM. When two arginines are available in 341 the N-terminal residues, both are involved in arginine sandwich interactions, providing 342 a 2-fold increase in affinity; however, either R52 (Tat NL4-3 ) or R53 (Tat G ) can act as the 343 remodeler. This can be explained by the presence of either a glutamine or histidine spacer, 344 respectively, which is the only amino acid difference between the two strains. As 345 glutamine (75%) and histidine (15%) make up the majority of the sequence variation in 346 this spacer, the structures show that these two spacer residues drive the differential 347 positioning of the arginine remodeler. In the Tat Fin strain, which has a histidine spacer, it 348 is the R52 that acts as a remodeler. In this case, the R53K substitution provides the 349 stabilizing interactions to reposition the single R52 arginine near pseudo-ASM3. 350 Furthermore, it is also interesting to compare the mode of binding of Tat Fin to additional ASM intercalation required for displacement. Thus, Tat has evolved specific 353 sequence variations that allow for the reconfiguration of pseudo-ASM3, even in cases 354 where there is only a single variation from HEXIM. Second, despite both ARMs having 355 lysines positioned near ASM4, only HEXIM leads to local destabilization. Our studies 356 therefore provide HEXIM as an example of a negative regulator that primes its own 357 displacement by locally destabilizing 7SK (Fig. 4c). Overall, these studies have broader 358 implications for 7SK-snRNP mediated regulation. Given that the destabilization-driven 359 displacement is a robust mechanism, it is possible that other yet to be identified cellular