An UBR-box from a non-N-recognin: Crystal Structure of the UBR domain of UBR6

The degradation of many short-lived proteins in eukaryotic cells is carried out by the Ubiquitin Proteasome System. The N-end rule pathway links the half-life of proteins to the identity of its N-terminal residue, also called N-degron. Destabilizing N-degrons, are recognized by E3 ubiquitin ligases termed N-recognins. N-degrons are grouped into type 1, composed of basic residues, and type 2, composed of bulky hydrophobic residues. In mammals, four N-recognins mediate the N-end rule pathway: UBR1, UBR2, UBR4 and UBR5. These proteins share a ~70-residue zinc finger-like motif termed the Ubiquitin Recognin (UBR) box, responsible for their specificity. The mammalian genome encodes at least three more UBR-box proteins: UBR3, UBR6/FBXO11 and UBR7. However, these UBRs cannot recognize any type of N-degrons. Our lab reported the crystal structures of the UBR boxes from the human UBR1 and UBR2, rationalizing the empirical rules for the classification of type 1 N-degrons. Despite the valuable information obtained from those structures there is not a clear explanation for the no recognition of N-degrons by other UBR-box proteins. Here we report the crystal structure of the UBR-box domain from UBR6 also known as FBXO11. UBR6 is a F-box protein of the SKP1-Cullin1-F-box (SCF) ubiquitin ligase complex and does not recognize any type of N-degrons. We crystallized a 77-residue fragment of the UBR-box of UBR6 and determined its structure at 1.7 Å resolution. Unexpectedly, this domain adopts an open conformation compared to UBR1-box, without any N-degron binding pockets. Its zinc-binding residues are conserved as in the N-recognins, but they are arranged in different zinc-binding motifs. Molecules form dimmers stabilized by zinc ions. The crystal had 4 molecules per asymmetric unit and space group P212121. For phasing we used Zn-SAD. With this structure we hope to obtain clues that explain the absence of N-degron recognition in some members of the UBR family.

Paip1 and Paip2: conformational modulators of PABP

Translation is regulated in cells. Key among the different stages of translation, translation initiation is controlled by regulatory interactions of eIFs and also translation regulatory proteins interacting with eIFs and PABP. PABP-interacting protein (Paip) 1 and 2 have been identified as regulatory proteins affecting the rate of translation initiation through their interactions with PABP. These interactions are mediated by the PABP interacting Motifs (PAM) 1 and 2 of Paips. Paip1 enhances the rate of translation while Paip2 suppresses it. To further understand their mechanisms of actions, here, we have studied the interactions of PABP RRMs with the PAM1 regions of Paip1 and 2. Compared to Paip1, affinity measurements using ITC show that Paip2 binds to PABP with higher affinity. While Paip2 dissociates poly(A) from PABP RRMs, Paip1 can not compete with poly(A) binding to PABP. Our binding studies show that Paip1 increases eIF4G-binding to PABP RRM1-2, supporting its stimulatory role in translation. Upon binding, Paip1 and Paip2 affect the conformation of the PABP RRMs differently. Compared to RRMs-Paip1, the complex of RRMs-Paip2 has a compact conformation. Our data suggest that Paip1 and Paip2 regulate PABP function through modulating its conformation, stimulating its incorporation in the pre-initiation complex, or sequestering it from translational machinery.

Structural studies of Legionella effectors

Legionella pneumophila is a gram-negative bacterium that causes Legionnaires' disease. It uses a Dot/Icm type IV secretion system to inject effector proteins into the host cell to manipulate host processes. Currently, about 300 Icm/Dot dependent effectors of L.pneumophila have been identified. Lpg1496 is an effector protein, which contains a conserved domain from the SidE family. To date, the middle domain and the conserved SidE domain have been crystallized and the structure solved at a resolution of 1.15Å and 2.3Å, respectively. A structural homology search using the middle domain suggested a similarity to phosphoribosylaminoimidazolesuccinocarboxamide (SAICAR) synthase, an ATPase involved in purine nucleotide synthesis. We performed 1H-15N HSQC NMR titrations to show that this domain binds ATP, ADP and AMP, with the highest binding affinity for ADP. A structural homology search using the SidE domain showed a similarity to cyclic nucleotide phosphodiesterases. To further elucidate the function of lpg1496, other fragments have been cloned, expressed, and subjected to crystallization trials. Currently, we have successfully crystallized the N-terminal domain, with crystals diffracting to <2.0Å. Obtaining the crystal structure of lpg1496 and revealing its function will not only lead to a better understanding of the virulence of L. pneumophila, but also contribute to the development of novel therapeutic treatments of Legionnaires' disease.

Structural studies of a novel phosphatase-transpoter interaction

The Phosphatase of Regenerating Liver (PRL) family, a subgroup of the protein tyrosine phosphatases, comprises three highly oncogenic members, PRL-1, -2 and -3, implicated in the progression of numerous cancer types. PRL-2 is overexpressed in breast cancer and was shown to promote mammary tumour growth in mice, but its full role in these oncogenic effects remains elusive. Recently, PRL-2 was found to interact with cyclin M3 (CNNM3), a magnesium transporter. To characterize this novel interaction, I crystallized PRL-2 in complex with the cytosolic cystathionine-β-synthase (CBS) domain of CNNM3. The binding surface consists of an elongated loop from CBS that makes contact close to the catalytic site of PRL-2. Site-directed CBS mutants confirmed the loop residues important for binding. PRL-2 and CBS bind particularly tightly, as determined by isothermal titration calorimetry, and PRL-2 enzyme assays revealed that CBS binding reduces the phosphatase's catalytic activity in vitro. The novel role of oncogenic PRL-2 as modulator of intracellular magnesium levels may represent the link between its overexpression and its effects on tumour growth. Small-molecule inhibitors of the PRL-2/CNNM3 complex formation are a potentially valuable tool for exploring the physiological function of this new interaction and may be used as future drug leads for the treatment of breast cancer.

Structural and functional studies of Sacsin

Autosomal recessive spastic ataxia of Charlevoix-Saguenay (ARSACS) is an early-onset neurodegenerative disorder caused by mutations in the SACS gene. It was first described in the French-Canadian population in 1978 by JP Bouchard, but ARSACS cases have since been found worldwide. The SACS gene codes for Sacsin, a 520kDa protein, which localizes in the cytoplasm, close to the mitochondria. Sacsin is composed of an N-terminal ubiquitin-like domain, which binds the proteasome, followed by three Sacsin repeat regions (SRR), a DnaJ domain, and at the C-terminus a nucleotide-binding HEPN domain, which mediates dimerization. While over 120 mutations in Sacsin are known to cause disease, the cellular function of the protein remains unclear. We recently crystallized the ATPase-like fragment of the first SRR domain from human Sacsin. The crystals diffract to better than 2A resolution and the structure determination is in progress. Future studies will focus on larger fragments of Sacsin as well as biochemical studies to investigate the cellular function of the protein.

Deciphering the activation of the E3 ubiquitin ligase parkin

Parkin is an E3 ubiquitin ligase responsible for some autosomal recessive forms of Parkinson's disease. Even though parkin is a RING-type E3 ligase, it uses a hybrid RING/HECT mechanism for its activity. The crystal structures of full-length and the RING0-RING1-In-Between-RING-RING2 module of parkin reveal a conformation of parkin in which its E2 binding site is too far from its catalytic cysteine for the transfer of ubiquitin [1]. Many intramolecular interactions occur between the different RING domains, as well as with a repressor element, which, with RING0, are unique to parkin. Mutations of residues involved in those interactions lead to an increase of parkin activity. This suggests that parkin adopts an auto-inhibited state in basal conditions. Therefore, under stress-response conditions, parkin needs to undergo molecular rearrangements, modulated by post-translational modification and/or interactions with other proteins, to become active. The phosphorylation of serine 65 in the Ubl domain of parkin by Pink1, a kinase also found mutated in some Parkinson's patient, was shown to increase the activity of parkin. Recent publications have demonstrated that ubiquitin is also phosphorylated by Pink1 and, furthermore, that phosphorylated ubiquitin could activate parkin [2,3]. We have used different techniques of structural biology and protein-protein interactions to further characterize the interaction of phosphorylated ubiquitin with parkin. This work provides insight into the mechanism of activation of parkin and that causes Parkinson's disease.