Scientists have caught a rare view of an unusual kinase poised to phosphorylate its substrate. Led by David Komander, MRC Laboratory of Molecular Biology in Cambridge, England, U.K., the researchers reported the crystal structure of the kinase PINK1 bound to ubiquitin in the October 30 Nature.

“We captured a pre-catalytic state, a snapshot of PINK1 holding ubiquitin before the phosphate group jumps on,” said first author Alexander Schubert.

In the Pink.

Top: PINK1 bound to its ubiquitin substrate (red) and stabilized by an antibody fragment (grey). Bottom: Key PINK1 features (right) include a stabilizing C-terminus region (CTR), a catalytic N-terminal lobe with three loops, a.k.a. insertions (yellow), and two phosphorylated serines (red) that are essential for catalysis. [Courtesy of Schubert et al., Nature 2017.]

Operating as a master regulator of mitochondrial quality control, PINK1 triggers the disposal of dysfunctional mitochondria (McWilliams and Muqit, 2017). Researchers know little about how the enzyme works, and why it fails when mutated. Assuming a fluid structure, the protein has been notoriously difficult to crystallize. Though recombinant versions might offer the best chance of getting enough human protein to study, they have turned out to be inactive. 

Instead, Komander’s group turned to the body louse Pediculus humanus corporis, which makes a constitutively active homolog that can be mass produced in E. coli (Woodroof et al., 2011). They mixed this recombinant PhPINK1 with a ubiquitin containing two point mutations, T66V and L67N. Komander’s group found that this Ub TVLN undergoes a conformational twist of its C-terminal that extends its phosphate-accepting Serine 65 outward. Ub TVLN binds PINK1 more tightly than WT ubiquitin and is phosphorylated 50 times more effectively. Komander contends that even WT ubiquitin occasionally assumes this conformation and that it might be the true physiological substrate for PINK1 (Gladkova et al., 2017). Finally, to keep PhPINK1 and Ub TVLN together, Schubert used a small fragment of an antibody, a.k.a. nanobody, made against PhPINK1 chemically cross-linked to Ub TVLN.

The PhPINK1-Ub TVLN-nanobody complex formed crystals suitable for x-radiography. A 3.1 Å structure revealed many features of PhPINK1 that are unique for a kinase (see image above). The CTR sports a large hydrophobic core that likely stabilizes PhPINK1. In addition, its N-lobe, a kinase domain normally composed of β-strands, includes three unique loops—“insertions” in PINK1 parlance—not found in other kinases. Insertion 3, the most conserved among PINK1 from different species, binds ubiquitin.

At a finer level, the authors also pinpointed key residues involved in substrate binding and catalysis. Tyrosine 198 forms part of a pocket that embraces a Ub TVLN hairpin (Ala46-Gly47) and helps hold ubiquitin in place by making hydrophobic contacts with Ub TVLN residues Ile44 and Val70, and a hydrogen bond with Gly47 (image below). In addition, a kinase activation segment in PhPINK1’s C-terminus binds Ub TVLN’s extended Ser65 loop via three polar bonds. Activation segments typically undergo phosphorylation as a prerequisite for  binding the substrate's target phosphorylation site. The researchers suggest that the flexibility of the Ser65 loop and the hinging of PhPINK1’s N- and C-lobes enable catalysis involving a nearby ATP-binding pocket. Interestingly, several Parkinsonism mutations that impact PINK1 activity cluster around this pocket.

The structure predicts that autophosphorylation contributes to organization of the N-lobe. PhosphoSer202 helps insertion 3 adopt its functional conformation (image above), and when mutated, disrupts both ubiquitin binding and kinase activity, highlighting its key role in PINK1 function.

Might wild-type ubiquitin interact with PhPINK1 in the same way? Possibly. The researchers proposed that after binding to the N-lobe, ubiquitin retracts its C-terminal domain to expose Ser65. Parkin, another PINK1 substrate, has a ubiquitin-like phosphorylation site and might morph in the same fashion, the authors suggested.

Close-up. The N-lobe (insertion 3, β3-αB loop, and Gly-rich loop), and the activation segment interact with ubiquitin (red) at the ubiquitin-binding site (left). A close-up (right) reveals Y198 forming a hydrogen bond (dotted line) with ubiquitin. Two PD mutations map to this region (light red). [Courtesy of Schubert et al., Nature.]

The structure pinpoints the locations for dozens of PINK1 PD mutations, shedding light on potential disease-causing mechanisms (see image below). Most of them affected protein folding, with the activation segment surfacing as a hot spot for structural mutations. The researchers reported that two mutations in insertion 3 of PhPINK1, G281D and P268L, eliminated and reduced, respectively, kinase activity toward Ub TVLN or ubiquitin. Also, many other mutations known to affect catalysis mapped to the ATP binding pocket.

Some of the results fit with a recently determined crystal structure of part of another PINK1 homolog from the red flour beetle, Tribolium castaneum. Researchers led by co-senior author Miratul Muqit at the University of Dundee, Scotland, also identified Ser202 (Ser205 in the beetle protein) as important for ubiquitin binding, recognized the CTR as an important structural domain, and gained insights into the mechanistic underpinnings of Parkinsonism-causing mutations (Kumar et al., 2017). However, their structure looked very different in the catalytic N-lobe, noted Komander in an email to Alzforum, probably because it lacked insertion 3, which was deleted to enhance stability.—Marina Chicurel

Comments

  1. Among the proteins implicated in the pathogenesis of Parkinson’s disease, the mitochondrial kinase PINK1 is undoubtedly one of the most eclectic, and over the past decade it has been clearly demonstrated that its neuroprotective activity unravels through a multitude of distinct mechanisms and pathways, within and outside mitochondria. These include the pro-survival phosphorylation of mitochondrial and anti-apoptotic proteins (such as TRAP1, NdufA10 and Bcl-xL), as well as a pivotal role in the regulation of mitophagy, exerted through two parallel mechanisms: on the one hand, the recruitment of Beclin1 to the mitochondria-associated membranes to form autophagosomes, and on the other hand, phosphorylation and recruitment of both parkin and ubiquitin onto the surface of damaged mitochondria to “process” them for mitophagic digestion. Yet the structure of PINK1 has remained elusive for more than a decade, hampering a deeper knowledge of PINK1 interaction with its substrates and, most importantly, of the different impact of distinct PD-related mutations on the various PINK1 functions.

    Now, the structure of Pediculus humanus corporis (Ph)PINK1 in complex with its substrate ubiquitin has been finally revealed in this study led by David Komander. This is a very important work for a number of reasons. Firstly, the authors demonstrated that (Ph)PINK1 does not phosphorylate ubiquitin in its common conformation, but specifically binds to a recently described but less represented “C-terminally retracted” conformation of ubiquitin. Secondly, they revealed some unique peculiarities of PINK1, such as the structure of the C-terminus domain, a region with no homologies with other proteins that was long known to modulate PINK1 catalytic activity, and the presence of three unique insertions in the kinase N-lobe, of which insertion-3 seems to be the most relevant to enable kinase activity (of note, this was also reported in the recent study on the structure of Tribolium castaneum PINK1, by Kumar and collaborators). Finally, Komander’s group was able to map most of the PINK1 mutations causative of PD on the (Ph)PINK1 structure, and showed that distinct mutations had variable effects on either protein folding, substrate binding, autophosphorylation, or complete kinase activity. This finding makes sense of previous observations noting a different impact of various PINK1 mutations on the protein’s many functions. For instance, we reported that PINK1W437X, a deletion mutant which entirely lacks the C-terminus domain and shows increased autophosphorylation activity, had impaired ability to bind Beclin1 and to enhance autophagy, while this was not observed for PINK1G309D, a mutant with defective kinase activity but unaltered ability to bind Beclin1 (Michiorri et al., 2010). In this light, the definition of the structure of PINK1 (albeit not yet that of the human protein) opens novel intriguing perspectives to better characterize its many interactions, and to establish whether some of these interactions may not necessarily implicate phosphorylation of the substrate, as seems to be the case for Beclin1. 

    References:

    . The Parkinson-associated protein PINK1 interacts with Beclin1 and promotes autophagy. Cell Death Differ. 2010 Jun;17(6):962-74. PubMed.

Make a Comment

To make a comment you must login or register.

References

Paper Citations

  1. . PINK1 and Parkin: emerging themes in mitochondrial homeostasis. Curr Opin Cell Biol. 2017 Apr;45:83-91. Epub 2017 Apr 22 PubMed.
  2. . Discovery of catalytically active orthologues of the Parkinson's disease kinase PINK1: analysis of substrate specificity and impact of mutations. Open Biol. 2011 Nov;1(3):110012. PubMed.
  3. . An invisible ubiquitin conformation is required for efficient phosphorylation by PINK1. EMBO J. 2017 Nov 13; PubMed.
  4. . Structure of PINK1 and mechanisms of Parkinson's disease associated mutations. Elife. 2017 Oct 5;6 PubMed.

Further Reading

Primary Papers

  1. . Structure of PINK1 in complex with its substrate ubiquitin. Nature. Published online 30 October 2017.