Special - October 2010
Short description: The genome projects have yielded the sequences of many genes of unknown function. Insight into the mechanisms of action of the products of these genes can come from collaboration between structural biology and follow-up functional analyses..
The number of genes that have been annotated in recent years has left us with a wealth of insight into the gene products encoded in any given genome, but the next step is to understand the function of these proteins. As illustrated by a number of recent papers covering a variety of processes and organisms, sometimes those insights can come from solving the structure of the protein, but a complement to these efforts is collaborative follow-up analyses that elucidate the biochemical function or specific binding properties of a given protein.
The poly(A)-binding proteins (PABPs) are involved in the post-transcriptional regulation of eukaryotic transcripts and fall into two classes. Structures of type I PABPs, alone and in complex with RNA, have been determined previously and involve the engagement of two RNA-binding RRM (RNA recognition motif) domains. Sheets and colleagues (PSI CESG) have now determined the solution structure of XlePABP2, an embryonically expressed type II PABP, revealing a homodimer formed by the interaction of two β-strands 1 . In the homodimer, the RNA-binding site, as predicted from related RRMs, is occluded, but by follow-up RNA-binding experiments, computational modeling and NMR analyses, the authors find that poly(A) RNA shifts the conformation of the homodimer, promoting monomer formation upon RNA binding. These analyses illustrate unexpected conformational changes that might regulate the function of a protein with a common and much-studied protein fold. In another collaboration targeting RNA-binding proteins, Brow and colleagues examined the RRM1-3 domains of Prp24 (a factor that affects U4:U6 snRNA pairing during spliceosome assembly) 2 . RRMs 1 and 2 interact, thus occluding their canonical RNA-binding interfaces, leaving RRM3 as a likely key interaction site. Binding and interaction assays, including chemical-shift analyses, lead to a specific model for the roles of the different Prp24 RRMs.
Having access to high-resolution structures can facilitate modeling to glean further information from new functional studies, as demonstrated in recent real-time probing analyses of bacterial transcription. Two key kinetic intermediates leading to initiation-competent RNA polymerase complex formation during bacterial transcription have previously been observed biochemically. Using real-time probing analyses, Saecker and colleagues (PSI CESG) were able to first define exposed regions in one of these intermediates, but also were able to combine these data with previous structural analyses of the polymerase to constrain and model what one such intermediate might look like structurally 3 . Thus, by combining functional analysis — probing — with a published structure related to transcription initiation, the authors were able to suggest what an important intermediate for initiation, previously derived as existing only by kinetic analysis, might actually look like.
Meanwhile, Tong and colleagues (PSI NESG) have investigated the still mysterious details of salicylic acid (SA) signaling in plants 4 . SA is critical for signaling, locally and away from infection sites, in the plant immunity system. Structures of tobacco SA binding protein 2 (SABP2), alone and in complex with SA to 2.1 Å resolution, have been defined. SABP2 belongs to the α/β hydrolase enzyme superfamily, with Ser81, His238, and Asp210 as the catalytic triad, and SA observed in the active site. Enzymatic studies demonstrate that SABP2 indeed functions as an esterase targeting methyl salycilate (MeSA, a biologically inactive SA derivative). Through structural and functional analyses, the authors hypothesize that the biological role of SABP2 may be to convert MeSA from its storage form to the active SA form as part of the wound healing response in plants. Further insights into plant signaling, based on examining function in the light of form, have come from the PSI CESG. Abscisic acid (ABA) is a small molecule involved in the plant stress response. The synthetic growth inhibitor pyrabactin acts as an ABA agonist and functional analyses with this compound allowed the authors to overcome complicated genetic redundancy in the system and help provide evidence for the identity of the elusive ABA receptor 5 . The protein PYR1 turns out to be an ABA receptor and inhibits the phosphatase PPC2C. This shows how a small-molecule ligand can be used to fish out other pathway components genetically.
Another striking example of the discovery of biochemical function driven by structural analysis is the case of Bacillus subtilis polyamine N-acetyltransferase (PaiA), which has been implicated in the negative regulation of sporulation. The 1.9 Å crystal structure determined by PSI NESG revealed that PaiA belongs to the N-acetyltransferase superfamily 6 , although PaiA had previously been suggested to have a DNA-binding domain. The structure further revealed a binding pocket that could act on a linear charged compound and, indeed, PaiA can act as an acetyltransferase with spermidine as substrate, as confirmed in both biochemical and cell-biological assays.
In a completely different field and a different organism, soluble N-ethylmaleimide-sensitive factor attachment protein gamma (γ-SNAP) is involved in membrane trafficking. The zebrafish γ-SNAP crystal structure at 2.6 Å resolution 7 , combined with further analyses by the PSI CESG, indicates that γ-SNAP shares similar preferred directions of bending and twisting of its twisted sheet motif to Sec17, leading to possible insights into disassembly of the membrane fusion complex. Another example of insights from collaborative functional investigations involves a soluble Rieske-type ferrodoxin. Although such ferrodoxins have been observedare known from in bacterial membrane complexes, they have not previously been observed as soluble in a eukaryotic system. The 2.07 Å crystal structure of a mouse Rieske ferrodoxin 8 (determined by PSI CESG) defines a β-barrel followed by a smaller [2Fe-2S]-binding region. This structure resembles those of bacterial dioxygenases, and subsequent studies find that it can even accept electrons from a bacterial oxidoreductase. Although not amenable so far to structural analysis, the human homolog behaves functionally like its mouse counterpart.
In a number of studies by Burley and colleagues at the PSI NYSGXRC, the activities of unannotated or misannotated proteins have been defined through collaboration with an NIH-funded program focused on functional annotation of enolases and amidohydrolases. Structural analysis of an uncharacterized enolase from Oceanobacillus iheyensis, along with functional analysis leading to the structure of a catalytic mutant, have defined galactarate as its substrate 9 . In another study, two new carboxypeptidases have been identified through cloning, expression and analysis of Caulobacter crescentus and Global Ocean Survey environmental genome sequences. In subsequent functional studies, a dipeptide library was used to define activity and specificity, and the crystal structure of one of these carboxypeptidases from the Sargasso Sea sequencing project was determined to 2.33 Å in complex with L-methionine 10 . Other amidohydrolase family members have also been defined. Another C. crescentus protein and the sequence Sgx9359b, derived from the Sargasso Sea project, were previously annotated as prolidases, but structural analyses in collaboration with functional analyses using combinatorial libraries of dipeptides now define them as amidohydrolases 11 .
Providing insight into a completely new function, the first enzyme capable of deaminating 8-oxoguanine (8-oxoG) to uric acid has been identified from Pseudomonas aeruginosa 12 . As 8-oxoG is a common DNA lesion, this discovery could have an impact on the study of DNA damage and repair pathways. The 2.2 Å crystal structure revealed that this enzyme is a member of the amidohydrolase superfamily, with a single zinc ion in the active site. Computational docking of potential intermediates identified conserved residues that might be involved in 8-oxoG binding to the active site and cross-species sequence comparisons have now identified other potential 8-oxoG deaminases.
In another example of the use of computational docking to help elucidate function, Thermotoga maritima Tm0936, known to belong to the amidohydrolase family and for which structures had emerged from the PSI NYSXGRC and the JCSG, was docked with high-energy intermediates of metabolites to try to predict its function 13 . The top-scoring list contained mostly adenine analogs, and likely hits suggested deaminase activity, which was confirmed experimentally. The crystal structure with the product of S-adenosylhomocysteine deamination was derived and resembled the expected product structure predicted from in silico docking analysis of intermediates.
Montelione and colleagues (PSI NESG) have pursued an alternative experimental approach to identify likely ligands using high-throughput NMR screening methods (FAST-NMR) to overcome the lack of immediate functional insight from sequence or structural homology 14 . An unannotated protein, SAV1430 from Staphylococcus aureus, was used as the test case. A one-dimensional NMR experiment was used as an easy first-pass screen to narrow down potential ligands and this was followed by two-dimensional 1H-15N heteronuclear single quantum correlation NMR on positive hits. For SAV1430, 21 ligand hits, all similar to O-phospho-L-tyrosine (pTyr) were uncovered. Structural bioinformatic analysis suggested that the binding site resembled an SH2 domain bound to pTyr, thus giving insight into both the function and the homology relationships of a protein for which the structure had begun to narrow down potential ligands, but where the details of binding specificity had been unclear.
Altogether, these examples illustrate a variety of processes examined through collaborations that give insight into the biological functions of a range of proteins.
J. Song, J. V. McGivern, K. W. Nichols, J. L. Markley & M. D. Sheets. Structural basis for RNA recognition by a type II poly(A)-binding protein.
Proc. Natl Acad. Sci. USA 105, 15317-15222 (2008). doi:10.1073/pnas.0801274105
E. Bae, N. J. Reiter, C. A. Bingman, S. S. Kwan, D. Lee et al. Structure and interactions of the first three RNA recognition motifs of splicing factor prp24.
J. Mol. Biol. 367, 1447-1458 (2007). doi:10.1016/j.jmb.2007.01.078
C. A. Davis, C. A. Bingman, R. Landick, M. T. Record, Jr & R. M. Saecker. Real-time footprinting of DNA in the first kinetically significant intermediate in open complex formation by Escherichia coli RNA polymerase.
Proc. Natl Acad. Sci. USA 104, 7833-7838 (2007). doi:10.1073/pnas.0609888104
F. Forouhar, Y. Yang, D. Kumar, Y. Chen, E. Fridman et al. Structural and biochemical studies identify tobacco SABP2 as a methylsalicylate esterase and implications for plant innate immunity.
Proc. Natl Acad. Sci. USA 102, 1773-1778 (2005). doi:10.1073/pnas.0409227102
S.-Y. Park, P. Fung, N. Nishimura, D. R. Jensen, H. Fujii et al. Abscisic acid inhibits type 2C protein phosphatases via the PYR/PYL family of START proteins.
Science 324, 1012-1013 (2009). doi:10.1126/science.1173041
F. Forouhar, I. Lee, J. Vujcic, S. Vujcic, J. Shen et al. Structural and functional evidence for Bacillus subtilis PaiA as a novel N1-spermidine/spermine acetyltransferase (SSAT).
J. Biol. Chem. 280, 40328-40336 (2005). doi:10.1074/jbc.M505332200
E. Bitto, C. A. Bingman, D. A. Kondrashov, J. G. McCoy, R. M. Bannen et al. Structure and dynamics of gamma-SNAP: insight into flexibility of proteins from the SNAP family.
Proteins 70, 93-104 (2008). doi:10.1002/prot.21468
E. J. Levin, N. L. Elsen, K. D. Seder, J. G. McCoy, B. G. Fox et al. X-ray structure of a soluble Rieske-type ferredoxin from Mus musculus.
Acta Crystallogr. D Biol. Crystallogr. 64, 933-934 (2008). doi:10.1107/S0907444908021653
J. F. Rakus, C. Kalyanaraman, A. A. Fedorov, E. V. Fedorov, F. P. Mills-Groninger et al. Computation-facilitated assignment of the function in the enolase superfamily: a regiochemically distinct galactarate dehydratase from Oceanobacillus iheyensis.
Biochemistry 48, 11546-11558 (2009). doi:10.1021/bi901731c
D. F. Xiang, C. Xu, D. Kumaran, A. C. Brown, J. M. Sauder et al. Functional annotation of two new carboxypeptidases from the amidohydrolase superfamily of enzymes.
Biochemistry 48, 4567-4576 (2009). doi:10.1021/bi900453u
D. F. Xiang, Y. Patskovsky, C. Xu, A. J. Meyer, J. M. Sauder et al. Functional identification of incorrectly annotated prolidases from the amidohydrolase superfamily of enzymes.
Biochemistry 48, 3730-3742 (2009). doi:10.1021/bi900111q
R. S. Hall, A. A. Fedorov, R. Marti-Arbona, E. V. Fedorov, P. Kolb et al. The hunt for 8-oxoguanine deaminase.
J. Am. Chem. Soc. 132, 1762-1763 (2010). doi:10.1021/ja909817d
J. C. Hermann, R. Marti-Arbona, A. A. Fedorov, E. Fedorov, S. C. Almo et al. Structure-based activity prediction for an enzyme of unknown function.
Nature 448, 775-779 (2007). doi:10.1038/nature05981
K. A. Mercier, M. Baran, V. Ramanathan, P. Revesz, R. Xiao et al. FAST-NMR – functional annotation screening technology using NMR spectroscopy.
J. Am. Chem. Soc. 128, 15292-15299 (2006). doi:10.1021/ja0651759