Structure and Mechanism of Ubiquitin Hydrolases (DUBs)

Humans have approximately 100 deubiquitinases (DUBs), which belong to five distinct groups: USPs (ubiquitin-specific proteases), UCHs (ubiquitin C-terminal hydrolases), MJDs (Machado-Joseph domain proteins), OTUs (ovarian tumor associated proteases), and JAMMs (JAB1/MPN/MOV34 domain proteins). The USPs, UCHs, MJDs, and OTUs are all cysteine proteases while the JAMM family DUBs are zinc-dependent metalloproteases. Many of these enzymes are linked to pathological conditions such as neurological disorders and cancer, but the underlying molecular mechanism is elusive. Our lab is engaged in dissecting the mechanism of these enzymes, specifically, how their catalytic activities are regulated. The enzymes we are studying are known to be activated when they associate with larger macromolecular complexes. Examples include UCHL5 (also known as UCH37), a proteasome-associated UCH DUB, and AMSH, an ESCRT (endosomal sorting complexes required for transport)-associated JAMM family DUB.

UCHL5 is activated for polyubiquitin chain processing when it binds to the 19S regulatory subunit of the proteasome via its interaction with a 19S member, Rpn13. We are seeking structural insight into this activation process. Our approach here is to crystallize UCHL5 in apo form and bound to its protein cofactors such as Rpn13 to visualize structural transformations in the active site of the enzyme. These crystallographic studies are supported by biophysical and mutational analysis that we hope will give us a complete picture of proteasome-mediated activation of UCHL5. Our studies in this direction have broader implication as to how the proteasome functions as a protein degradation machinery.

AMSH, associated molecule with the SH3 domain of STAM, is a JAMM domain-containing metalloprotease whose catalytic activity is stimulated when recruited to the ESCRT complexes (ESCRT-0 and ESCRT-III), which are essential for endocytosis. Using X-ray crystallography supported by biophysical analysis and site-directed mutagenesis, we are trying to elucidate the mechanism of its recruitment to and subsequent activation by the ESCRT complexes.

Identification and Purification of DUB complexes

Schematic illustration of affinity-based isolation DUB complexes and identification of their subunits.
Figure 1: Schematic illustration of affinity-based isolation DUB complexes and identification of their subunits.

Growing evidence seems to indicate that many DUBs bind to other proteins, forming macromolecular complexes that are capable of deubiquitinating specific targets. This is reflected in their primary structure; most of the human DUBs have domains for protein-protein interaction in addition to the catalytic domain. These complexes likely represent the true biologically active forms of these DUBs, so information about these complexes is essential in advancing our understanding of the biological pathways in which DUBs play a role. Using reactive ubiquitin-based affinity probes, we are trying to develop new methods for identifying DUB macromolecular complexes. An example of this is shown below.

Ubiquitin is attached to a metal chelator (MC) at its C terminus and is HA (hemagglutinin epitope)-tagged at its N terminus. This probe (HA-Ub-MC) can be synthesized by intein-mediated semisynthesis. This reagent is applied to cellular lysate from mammalian cell culture. The probe is expected to bind to metallo DUB complexes present in the cells (e.g., BRISC complex, a multi-subunit Lys63-specific DUB complex with a JAMM domain as its catalytic center). The probe-bound complexes can be affinity purified on anti-HA agarose beads and analyzed by mass spectrometry in order to identify members of the complex that make the active holoenzyme. Studies from this proteomics-based approach are likely to yield novel information about active DUB complexes and will pave the way for identification of the biological function of these enzymes.

Protein Degradation and Neurological Disorder and cancer

Crystal structure of UCHL1 bound to an inhibitor (cyan).
Figure 2: Crystal structure of UCHL1 bound to an inhibitor (cyan).

Ubiquitination is deregulated in a number of diseases, including neurological disorders such as Parkinson's and Alzheimer's diseases (PD and AD). For example, UCHL1 is mutated in PD (the I93M mutation). Our lab is trying to gather insight into the molecular nature of the effect of this mutation. Among other properties that may have changed, such as catalytic efficiency, the I93M mutation appears to have destabilized the protein, exposing parts of the hydrophobic patch that may be associated with increased propensity to aggregate in vivo. Our studies are directed toward understanding the change in properties, physical and/or biochemical, of the mutant to establish a clear basis of UCHL1's link with PD.

Normally an abundant neuronal protein, UCHL1 is also found highly expressed in certain cancers, such as pancreatic and colorectal cancers. Recent studies have shown that UCHL1 could behave as an oncogene and may drive tumor growth. We are actively trying to develop small-molecule inhibitors of the enzyme that we hope will be used one day as a therapeutic for UCHL1-mediated cancers. Our efforts in this direction involve structure-based rational design and library screening approaches for identifying small-molecule modulators of this and other disease-associated DUBs. The picture below illustrates a crystallographic study from our lab of a peptide-fluoromethylketone inhibitor bound in the active site of UCHL1. We are now trying to design and synthesize potent analogs of this inhibitor for testing in mice models.

Structural Studies to understand regulation of metabolic enzymes

Substrate binding to cPAH.
Figure 3: Substrate binding to cPAH.

Our lab is also actively engaged in understanding how metabolic enzymes are regulated. The specific system we are studying in this project is phenylalanine hydroxylase (PAH), a non-heme iron enzyme that catalyzes phenylalanine oxidation to tyrosine. Human PAH features a regulatory domain where binding of the substrate leads to allosteric activation of the enzyme. Phenylketonuria (PKU) arises when dysfunctional PAH in humans impairs metabolism of phenylalanine, leading to accumulation of abnormal levels of phenylalanine in the blood. Working with the Abu-Omar group in our department, we are trying to address the effect of PKU mutations on the structure, stability and catalytic function of this key metabolic enzyme. We are now extending this study to understand the structure and function PAH from pathogenic organisms such as Leishmania major.

We recently discovered an additional substrate binding site, distinct from the active site, on a single-domain, monomeric bacterial enzyme, cPAH (PAH from Chromobacterium violaceum). We call this the distal site because it is located on a part of the enzyme 15 Å from its active site. This raises an intriguing possibility that the additional binding sites may be used to regulate enzymatic activity even in single-domain, monomeric enzymes. We are currently trying to determine if the distal site contributes to catalysis or regulation thereof.