Methods to refine macromolecular structures in cases of severe diffraction anisotropy

Specific Xray diffraction patterns of membrane proteins caused by secondary structure collinearity

Diffraction anisotropy is a phenomenon that impacts more specifically membrane proteins, compared to soluble ones, but the reasons for this discrepancy remained unclear. Often, it is referred to a difference in resolution limits between highest and lowest diffraction limits as a signature for anisotropy. We show in this article that there is no single correlation between anisotropy and difference in resolution limits, with notably a substantial number of structures displaying various anisotropy with no difference in resolution limits. We further investigated diffraction intensity profiles, and observed a peak centred on 4.9 Å resolution more predominant in membrane proteins. Since this peak is in the region corresponding to secondary structures, we investigated the influence of secondary structure ratio. We showed that secondary structure content has little influence on this profile, while secondary structure collinearity in membrane proteins correlate with a stronger peak. Finally, we could further show that the presence of this peak is linked to higher diffraction anisotropy. These results bring to light a specific diffraction of membrane protein crystals, which calls for a specific handling by crystallographic software. It also brings an explanation for investigators struggling with their anisotropic data.

Structures of GapR reveal a central channel which could accommodate B-DNA

GapR is a nucleoid-associated protein required for the cell cycle of Caulobacter cresentus. We have determined new crystal structures of GapR to high resolution. As in a recently published structure, a GapR monomer folds into one long N-terminal α helix and two shorter α helices, and assembles into a tetrameric ring with a closed, positively charged, central channel. In contrast to the conclusions drawn from the published structures, we observe that the central channel of the tetramer presented here could freely accommodate B-DNA. Mutation of six conserved lysine residues lining the cavity and electrophoretic mobility gel shift experiments confirmed their role in DNA binding and the channel as the site of DNA binding. Although present in our crystals, DNA could not be observed in the electron density maps, suggesting that DNA binding is non-specific, which could be important for tetramer-ring translocation along the chromosome. In conjunction with previous GapR structures we propose a model for DNA binding and translocation that explains key published observations on GapR and its biological functions.

X-ray diffraction reveals the intrinsic difference in the physical properties of membrane and soluble proteins

Membrane proteins are distinguished from soluble proteins by their insertion into biological membranes. This insertion is achieved via a noticeable arrangement of hydrophobic amino acids that are exposed at the surface of the protein, and renders the interaction with the aliphatic tails of lipids more energetically favorable. This important difference between these two categories of proteins is the source of the need for a specific handling of membrane proteins, which transpired in the creation of new tools for their recombinant expression, purification and even crystallization. Following this line, we show here that crystals of membrane proteins display systematically higher diffraction anisotropy than those of soluble proteins. This phenomenon dramatically hampers structure solution and refinement, and has a strong impact on the quality of electron-density maps. A farther search for origins of this phenomenon showed that the type of crystallization, and thus the crystal packing, has no impact on anisotropy, nor does the nature or function of the membrane protein. Membrane proteins fully embedded within the membrane display equal anisotropy compared to the ones with extra membranous domains or fusions with soluble proteins. Overall, these results overturn common beliefs and call for a specific handling of their diffraction data.

Crystal Structure of the Marburg Virus VP35 Oligomerization Domain

Marburg virus (MARV) is a highly pathogenic filovirus that is classified in a genus distinct from that of Ebola virus (EBOV) (genera Marburgvirus and Ebolavirus, respectively). Both viruses produce a multifunctional protein termed VP35, which acts as a polymerase cofactor, a viral protein chaperone, and an antagonist of the innate immune response. VP35 contains a central oligomerization domain with a predicted coiled-coil motif. This domain has been shown to be essential for RNA polymerase function. Here we present crystal structures of the MARV VP35 oligomerization domain. These structures and accompanying biophysical characterization suggest that MARV VP35 is a trimer. In contrast, EBOV VP35 is likely a tetramer in solution. Differences in the oligomeric state of this protein may explain mechanistic differences in replication and immune evasion observed for MARV and EBOV.

IMPORTANCE

Marburg virus can cause severe disease, with up to 90% human lethality. Its genome is concise, only producing seven proteins. One of the proteins, VP35, is essential for replication of the viral genome and for evasion of host immune responses. VP35 oligomerizes (self-assembles) in order to function, yet the structure by which it assembles has not been visualized. Here we present two crystal structures of this oligomerization domain. In both structures, three copies of VP35 twist about each other to form a coiled coil. This trimeric assembly is in contrast to tetrameric predictions for VP35 of Ebola virus and to known structures of homologous proteins in the measles, mumps, and Nipah viruses. Distinct oligomeric states of the Marburg and Ebola virus VP35 proteins may explain differences between them in polymerase function and immune evasion. These findings may provide a more accurate understanding of the mechanisms governing VP35's functions and inform the design of therapeutics.

Re-refinement of the spliceosomal U4 snRNP core-domain structure

The core domain of small nuclear ribonucleoprotein (snRNP), comprised of a ring of seven paralogous proteins bound around a single-stranded RNA sequence, functions as the assembly nucleus in the maturation of U1, U2, U4 and U5 spliceosomal snRNPs. The structure of the human U4 snRNP core domain was initially solved at 3.6 Å resolution by experimental phasing using data with tetartohedral twinning. Molecular replacement from this model followed by density modification using untwinned data recently led to a structure of the minimal U1 snRNP at 3.3 Å resolution. With the latter structure providing a search model for molecular replacement, the U4 core-domain structure has now been re-refined. The U4 Sm site-sequence AAUUUUU has been shown to bind to the seven Sm proteins SmF-SmE-SmG-SmD3-SmB-SmD1-SmD2 in an identical manner as the U1 Sm-site sequence AAUUUGU, except in SmD1 where the bound U replaces G. The progression from the initial to the re-refined structure exemplifies a tortuous route to accuracy: where well diffracting crystals of complex assemblies are initially unavailable, the early model errors are rectified by exploiting preliminary interpretations in further experiments involving homologous structures. New insights are obtained from the more accurate model.

Nucleotide binding interactions modulate dNTP selectivity and facilitate 8-oxo-dGTP incorporation by DNA polymerase lambda

8-Oxo-7,8,-dihydro-2'-deoxyguanosine triphosphate (8-oxo-dGTP) is a major product of oxidative damage in the nucleotide pool. It is capable of mispairing with adenosine (dA), resulting in futile, mutagenic cycles of base excision repair. Therefore, it is critical that DNA polymerases discriminate against 8-oxo-dGTP at the insertion step. Because of its roles in oxidative DNA damage repair and non-homologous end joining, DNA polymerase lambda (Pol λ) may frequently encounter 8-oxo-dGTP. Here, we have studied the mechanisms of 8-oxo-dGMP incorporation and discrimination by Pol λ. We have solved high resolution crystal structures showing how Pol λ accommodates 8-oxo-dGTP in its active site. The structures indicate that when mispaired with dA, the oxidized nucleotide assumes the mutagenic syn-conformation, and is stabilized by multiple interactions. Steady-state kinetics reveal that two residues lining the dNTP binding pocket, Ala(510) and Asn(513), play differential roles in dNTP selectivity. Specifically, Ala(510) and Asn(513) facilitate incorporation of 8-oxo-dGMP opposite dA and dC, respectively. These residues also modulate the balance between purine and pyrimidine incorporation. Our results shed light on the mechanisms controlling 8-oxo-dGMP incorporation in Pol λ and on the importance of interactions with the incoming dNTP to determine selectivity in family X DNA polymerases.

Structure determination of Murine Norovirus NS6 proteases with C-terminal extensions designed to probe protease-substrate interactions

Noroviruses are positive-sense single-stranded RNA viruses. They encode an NS6 protease that cleaves a viral polyprotein at specific sites to produce mature viral proteins. In an earlier study we obtained crystals of murine norovirus (MNV) NS6 protease in which crystal contacts were mediated by specific insertion of the C-terminus of one protein (which contains residues P5-P1 of the NS6-7 cleavage junction) into the peptide binding site of an adjacent molecule, forming an adventitious protease-product complex. We sought to reproduce this crystal form to investigate protease–substrate complexes by extending the C-terminus of NS6 construct to include residues on the C-terminal (P′) side of the cleavage junction. We report the crystallization and crystal structure determination of inactive mutants of murine norovirus NS6 protease with C-terminal extensions of one, two and four residues from the N-terminus of the adjacent NS7 protein (NS6 1′, NS6 2′, NS6 4′). We also determined the structure of a chimeric extended NS6 protease in which the P4-P4′ sequence of the NS6-7 cleavage site was replaced with the corresponding sequence from the NS2-3 cleavage junction (NS6 4′ 2|3).The constructs NS6 1′ and NS6 2′ yielded crystals that diffracted anisotropically. We found that, although the uncorrected data could be phased by molecular replacement, refinement of the structures stalled unless the data were ellipsoidally truncated and corrected with anisotropic B-factors. These corrections significantly improved phasing by molecular replacement and subsequent refinement.The refined structures of all four extended NS6 proteases are very similar in structure to the mature MNV NS6—and in one case reveal additional details of a surface loop. Although the packing arrangement observed showed some similarities to those observed in the adventitious protease-product crystals reported previously, in no case were specific protease–substrate interactions observed.