Structural Studies of HHARI/UbcH7∼Ub Reveal Unique E2∼Ub Conformational Restriction by RBR RING1

Abstract

SummaryRING-between-RING (RBR) E3s contain RING1 domains that are structurally similar yet mechanistically distinct from canonical RING domains. Both types of E3 bind E2∼ubiquitin (E2∼Ub) via their RINGs but canonical RING E3s promote closed E2∼Ub conformations required for direct Ub transfer from the E2 to substrate, while RBR RING1s promote open E2∼Ub to favor Ub transfer to the E3 active site. This different RING/E2∼Ub conformation determines its direct target, which for canonical RING E3s is typically a substrate or substrate-linked Ub, but is the E3 active-site cysteine in the case of RBR-type E3s. Here we show that a short extension of HHARI RING1, namely Zn2+-loop II, not present in any RING E3s, acts as a steric wedge to disrupt closed E2∼Ub, providing a structural explanation for the distinctive RING1-dependent conformational restriction mechanism utilized by RBR E3s.

Synergistic recruitment of UbcH7~Ub and phosphorylated Ubl domain triggers parkin activation

“…We interpreted the chemical shift changes within the tether region to result from both direct E2 binding and an altering of the tether position to accommodate the E2 enzyme. The open arrangement of the E2‐Ub bound to R0RBR parkin more closely resembles the interaction of UbcH5b‐Ub with HOIP (Lechtenberg et al , 2016) than either of the structures for UbcH7‐Ub with HHARI (Dove et al , 2017; Yuan et al , 2017; Fig EV4). Ub binding is governed predominantly by contacts from β1‐L1‐β2 (K6, L8, K11, I13‐T14), the linker following helix α1 (Q31‐D32) and C‐terminus (L73, R74) to an R0RBR surface including β1 (P333, P335) and the C‐terminus of the IBR domain (E370) and adjacent tether (V380, F381, S384, T386), RING1 helix H1 (N273) and the straightened RING1 helix H3 (R314, Y318).…”

“…In the complex, helix H1 (R6, K9) and loop L4 (E60, F63‐K64) in UbcH7 contact the R0RBR loop L1 (T240, T242) and helix H1 (L266, T270, R271, D274) respectively in RING1 (Fig 2C). As with other E2/RBR E3 ligase complexes (Lechtenberg et al , 2016; Dove et al , 2017; Yuan et al , 2017), the L7 loop (N94, K96, P97, A98) in UbcH7 sits adjacent to loop L2 (V290‐G292) in RING1, but has additional contacts with Y391‐D394 just prior to the short helix in the tether region. We noted that some of the largest chemical shift changes were in this region of the tether (Y391, R392, D394), which undergoes multiple rearrangements in crystal structures of parkin (Kumar et al , 2017).…”

“…These observations show the pUbl domain is no longer bound at the IBR/RING1 interface and, consistent with previous NMR, sedimentation velocity and computational experiments, indicates the pUbl domain adopts a range of bound/free conformations in the pParkin:pUb state (Fig 1D; Caulfield et al , 2014; Aguirre et al , 2017). In this scenario, release of the pUbl domain exposes the predicted E2~Ub binding site, based on other RING/E2 complexes (Lechtenberg et al , 2016; Dove et al , 2017; Yuan et al , 2017). …”

“…In particular, RBR E3 ligases incorporate a hybrid ubiquitination mechanism (Wenzel et al , 2011) whereby an E2 conjugating enzyme is recruited to the RING1 domain (similar to RING E3 ligases) and ubiquitin (Ub) is transferred from the E2~Ub conjugate to a catalytic cysteine in the RING2(Rcat) domain (similar to HECT E3 mechanisms) prior to labelling of a substrate lysine. RBR E3 ligases and RING E3 ligases have RING domains that are structurally similar and are expected to recruit E2 enzymes in a similar fashion (Budhidarmo et al , 2012), as recently shown in crystal structures of the RBR E3 ligases HHARI with UbcH7‐Ub (Dove et al , 2017; Yuan et al , 2017) and HOIP with UbcH5b‐Ub (Lechtenberg et al , 2016). However, a distinguishing feature of the HHARI and HOIP RBR E3 ligases is their ability to recognize an extended (“open”) form of the E2~Ub conjugate similar to that used by HECT E3 enzymes.…”

“…Current crystal structures of parkin show the proposed E2 binding site on the RING1 domain is > 50 Å from the catalytic site (C431) in the RING2(Rcat) domain suggesting a significant conformational rearrangement is needed (Riley et al , 2013; Trempe et al , 2013; Wauer & Komander, 2013; Kumar et al , 2015, 2017; Sauvé et al , 2015; Wauer et al , 2015). A similar dilemma arises from recent structures of HHARI in complex with UbcH7‐Ub that show the ubiquitin molecule is 47–53 Å from the catalytic site (Dove et al , 2017; Yuan et al , 2017). Alternatively, structures of HOIP:E2‐Ub and parkin:pUb complexes assembled from domain‐swapping or symmetry‐related molecules raise the possibility of co‐operation between multiple E3 ligase molecules to promote ubiquitin transfer (Lechtenberg et al , 2016; Kumar et al , 2017).…”