|Home | About | Journals | Submit | Contact Us | Français|
The crystal structures of SPO0140 and Sbal_2486 were determined using the semiautomated high-throughput pipeline of the Joint Center for Structural Genomics (JCSG) as part of the NIGMS Protein Structure Initiative (PSI). The structures revealed a conserved core with domain duplication and a superficial similarity of the C-terminal domain to pleckstrin homology-like folds. The conservation of the domain interface indicates a potential binding site that is likely to involve a nucleotide-based ligand, with genome-context and gene-fusion analyses additionally supporting a role for this family in signal transduction, possibly during oxidative stress.
To extend the structural coverage of proteins with uncharacterized biological function, we targeted Pfam protein family Pfam06938 (DUF1285), for which we determined the structures of two representative members. The SPO0140 gene of Silicibacter pomeroyi DSS-3, a marine α-proteobacterium, encodes a protein with a molecular weight of 21.2 kDa (residues 1–193) and a calculated isoelectric point of 5.7. The Sbal_2486 gene of Shewanella baltica OS155, a psychrotropic marine γ-proteobacterium, encodes a protein with a molecular weight of 17.7 kDa (residues 1–157) and a calculated isoelectric point of 4.8.
The clones for SPO0140 and Sbal_2486 were generated using the Polymerase Incomplete Primer Extension (PIPE) cloning method (Klock et al., 2008 ). The gene encoding SPO0140 (GenBank YP_165412, gi:56695065, UniProt Q5LWU5) was amplified by polymerase chain reaction (PCR) from Silicibacter pomeroyi DSS-3 genomic DNA using PfuTurbo DNA polymerase (Stratagene) and I-PIPE (Insert) primers (forward primer, 5′-ctgtacttccagggcATGAGCGGACAAAAGCCTGTGAAACC-3′; reverse primer, 5′-aattaagtcgcgttaGCCGCGCTCCAGCTCCTCGACCGTCATC-3′; target sequence in upper case) that included sequences for the predicted 5′ and 3′ ends. The expression vector pSpeedET, which encodes an amino-terminal tobacco etch virus (TEV) protease-cleavable expression and purification tag (MGSDKIHHHHHHENLYFQ/G), was PCR-amplified with V-PIPE (Vector) primers (forward primer, 5′-taacgcgacttaattaactcgtttaaacggtctccagc-3′; reverse primer, 5′-gccctggaagtacaggttttcgtgatgatgatgatgatg-3′). The V-PIPE and I-PIPE PCR products were mixed to anneal the amplified DNA fragments together. Escherichia coli GeneHogs (Invitrogen) competent cells were transformed with the V-PIPE/I-PIPE mixture and dispensed onto selective LB–agar plates. The cloning junctions were confirmed by DNA sequencing. Expression was performed in a selenomethionine-containing medium at 310 K with suppression of normal methionine synthesis. At the end of fermentation, lysozyme was added to the culture to a final concentration of 250 µg ml−1 and the cells were harvested and frozen. After one freeze–thaw cycle, the cells were homogenized in lysis buffer [50 mM HEPES pH 8.0, 50 mM NaCl, 10 mM imidazole, 1 mM tris(2-carboxyethyl)phosphine–HCl (TCEP)] and the lysate was clarified by centrifugation at 32 500g for 30 min. The soluble fraction was passed over nickel-chelating resin (GE Healthcare) pre-equilibrated with lysis buffer, the resin was washed with wash buffer [50 mM HEPES pH 8.0, 300 mM NaCl, 40 mM imidazole, 10%(v/v) glycerol, 1 mM TCEP] and the protein was eluted with elution buffer [20 mM HEPES pH 8.0, 300 mM imidazole, 10%(v/v) glycerol, 1 mM TCEP]. The eluate was buffer-exchanged with TEV buffer (20 mM HEPES pH 8.0, 200 mM NaCl, 40 mM imidazole, 1 mM TCEP) using a PD-10 column (GE Healthcare) and incubated with 1 mg TEV protease per 15 mg of eluted protein. The protease-treated eluate was run over nickel-chelating resin (GE Healthcare) pre-equilibrated with HEPES crystallization buffer (20 mM HEPES pH 8.0, 200 mM NaCl, 40 mM imidazole, 1 mM TCEP) and the resin was washed with the same buffer. The flowthrough and wash fractions were combined and concentrated to 13.9 mg ml−1 by centrifugal ultrafiltration (Millipore) for crystallization trials. SPO0140 was crystallized by mixing 200 nl protein solution with 200 nl crystallization solution and using a 50 µl reservoir using the nanodroplet vapor-diffusion method (Santarsiero et al., 2002 ) with standard Joint Center for Structural Genomics (JCSG; http://www.jcsg.org) crystallization protocols (Lesley et al., 2002 ). The crystallization reagent consisted of 20%(v/v) ethanol and 100 mM Tris pH 8.5. Glycerol was added to a final concentration of 17%(v/v) as a cryoprotectant. A rod-shaped crystal of approximate dimensions 0.08 × 0.02 × 0.02 mm was harvested after 29 d at 277 K for data collection. Initial screening for diffraction was carried out using the Stanford Automated Mounting (SAM) system (Cohen et al., 2002 ) at the Stanford Synchrotron Radiation Lightsource (SSRL, Menlo Park, California, USA). The diffraction data were indexed in the tetragonal space group P43212.
The gene encoding Sbal_2486 (GenBank YP_001050848.1, gi:126174699, UniProt A3D5G6) was amplified from Shewanella baltica OS155 genomic DNA. Using the PIPE method (Klock et al., 2008 ), the initial clone was generated by using I-PIPE (Insert) primers (forward primer, 5′-ctgtacttccagggcATGGAAAAGATGACTGACAGTATTCAAC-3′; reverse primer, 5′-aattaagtcgcgttaCTGCTCATTTAGATCAGATAAATTG-3′; target sequence in upper case) that included sequences for the predicted 5′ and 3′ ends. Cloning, expression and purification were performed as described above for SPO0140. Crystals obtained from the full-length construct were not suitable for data collection. Bioinformatic predictions suggested that a 12-residue N-terminal truncation might produce better diffracting crystals than the full-length (157 residues) wild-type protein. Cloning attempts were initiated to generate nested truncations in steps of four residues around this prediction for a truncated sequence. Additionally, C-terminal truncations (NB these were not part of the prediction to improve crystallization) were attempted. Truncation clones were successfully generated for the construct boundaries 9–157, 17–157, 21–157, 1–153, 1–149, 1–145 and 1–141. These constructs were screened in parallel for solubility, crystallization and diffraction (Table 1 ). No clone was obtained for the initially proposed truncation (13–157) construct. The other three N-terminal truncations produced soluble protein that led to harvestable crystals that were of higher quality and diffracted better than the crystals of the full-length protein. The only C-terminal truncation that produced soluble protein was that for residues 1–153. This construct also led to harvestable crystals, which again were of higher quality and diffracted better than the full length construct. The improvement was less than that observed for the N-terminal truncation constructs. By making several truncation constructs and empirically testing all of them, we found a construct that was better suited to crystallization and structure determination. A crystal of the 9–157 construct was used for structure solution. The primers used to generate the 9–157 truncation clone by PIPE mutagenesis were I-PIPE (Insert) forward primer 5′-ctgtacttccagggcCAACACACACTCAAACAATTCGCCGCCG-3′ and reverse primer 5′-gccctggaagtacaggttttcgtgatgatgatgatgatg-3′ (Klock et al., 2008 ). Purified Sbal_2486 was concentrated to 20 mg ml−1 by centrifugal ultrafiltration (Millipore) for crystallization trials. Sbal_2486 was crystallized by mixing 200 nl protein solution and 200 nl crystallization solution and using a 50 µl reservoir volume using the nanodroplet vapor-diffusion method with standard JCSG crystallization protocols. The crystallization reagent consisted of 20%(v/v) 2-propanol, 20%(w/v) PEG 4000 and 0.1 M sodium citrate pH 5.6. Ethylene glycol (1,2-ethanediol) was added to a final concentration of 10%(v/v) as a cryoprotectant. A rhombohedral crystal of approximate size 0.1 × 0.1 × 0.1 mm was harvested after 20 d at 277 K for data collection. Initial screening for diffraction was carried out using the SAM system and an X-ray microsource (Miller & Deacon, 2007 ) installed at SSRL. The diffraction data were indexed in the orthorhombic space group P212121.
For SPO0140, multiple-wavelength anomalous diffraction (MAD) data were collected on beamline BL11-1 at SSRL at wavelengths corresponding to the inflection (λ3-2re3), high-energy remote (λ1-2re3) and peak (λ2-2re3) of a selenium MAD experiment. The data sets were collected at 100 K with a MAR Mosaic 325 mm CCD detector (Rayonix) using the Blu-Ice (McPhillips et al., 2002 ) data-collection environment. The MAD data were integrated and reduced using MOSFLM (Leslie, 1992 ) and were scaled using the program SCALA (Collaborative Computational Project, Number 4, 1994 ). Selenium-substructure solution and phasing were performed with SHELXD (Sheldrick, 2008 ) and autoSHARP (Bricogne et al., 2003 ) with a mean figure of merit of 0.49 for 12 selenium sites (NB there are five unique Se sites per chain, but SeMet131 adopts two different conformations resulting in two partial occupancy sites). Automatic model building was performed with ARP/wARP (Cohen et al., 2004 ). Model completion and refinement were performed with Coot (Emsley & Cowtan, 2004 ) and REFMAC5.2 (Winn et al., 2003 ) using the remote (λ1-2re3) data set. The refinement included experimental phase restraints in the form of Hendrickson–Lattman coefficients from SHARP, NCS restraints (positional weight 0.5 and thermal weight 2.0) and TLS refinement with one TLS group per chain. Data-reduction and refinement statistics for SPO0140 are summarized in Table 2 .
For Sbal_2486, MAD data were collected on beamline 8.2.2 at the ALS at wavelengths corresponding to the inflection (λ2-2ra9), low-energy remote (λ3-2ra9) and peak (λ1-2ra9) of a selenium MAD experiment. The inflection and remote data were collected first using an interleaved protocol with a 10° wedge size over a total sweep of 100° and the peak data were then collected in a 130° sweep. The data sets were collected at 100 K with a Quantum 315 CCD detector (ADSC). The MAD data were integrated and reduced using MOSFLM and were scaled with the program SCALA. Selenium-substructure solution and phasing were performed with SHELXD and autoSHARP with a mean figure of merit of 0.48 for a single selenium site. Automatic model building was performed with ARP/wARP. Model completion and refinement were performed with Coot (Emsley & Cowtan, 2004 ) and REFMAC5.2 (Murshudov et al., 1999 ) using the peak (λ1-2ra9) data set. The refinement included experimental phase restraints in the form of Hendrickson–Lattman coefficients from SHARP and restrained anisotropic ADP refinement. Data-reduction and refinement statistics for Sbal_2486 are summarized in Table 3 .
The quality of the crystal structures was analyzed using the JCSG Quality Control server (http://smb.slac.stanford.edu/jcsg/QC). This server processes the coordinates and data through a variety of validation tools including AutoDepInputTool (Yang et al., 2004 ), MolProbity (Davis et al., 2007 ), WHAT IF v.5.0 (Vriend, 1990 ), RESOLVE (Terwilliger, 2003 ) and MOLEMAN2 (Kleywegt, 2000 ), as well as several in-house scripts, and summarizes the output. Protein quaternary-structure analysis used the PISA server (Krissinel & Henrick, 2007 ). Fig. 1 (b) was adapted from an analysis using PDBsum (Laskowski et al., 2005 ) and all others were prepared with PyMOL (DeLano Scientific). Atomic coordinates and experimental structure factors for SPO0140 at 2.5 Å resolution and Sbal_2486 at 1.4 Å resolution have been deposited in the PDB and are accessible under codes 2re3 and 2ra9, respectively.
The crystal structure of SPO0140 (Fig. 1 ) was determined to 2.5 Å resolution using the MAD method. Data collection, model and refinement statistics are summarized in Table 2 . The final model included two protomers (residues 10–192 for chain A, residues 10–193 for chain B), one glycerol molecule and 186 water molecules in the asymmetric unit. No electron density was observed for residues Gly0 (which remained at the N-terminus after cleavage of the purification tag), SeMet1–Pro9 in chains A and B and Gly193 in chain A. Side-chain atoms of Lys108, Thr137 and Glu139 in chain A and Ser10, Lys108, Gln133, Thr137 and Glu139 in chain B had poorly defined electron density and were omitted from the model. The Matthews coefficient (V M; Matthews, 1968 ) was 3.0 Å3 Da−1 and the estimated solvent content was 59.0%. The Ramachandran plot produced by MolProbity (Davis et al., 2007 ) showed that 97.0% of the residues were in favored regions, with no outliers.
The crystal structure of Sbal_2486 was determined to 1.4 Å resolution using the MAD method. Data collection, model and refinement statistics are summarized in Table 3 . The final model included a monomer of 127 residues, seven ethylene glycol molecules, one sodium ion and 231 water molecules in the asymmetric unit. No electron density was observed for residues Gly0 (which remained at the N-terminus after cleavage of the purification tag), Gln9–Cys28 and Glu156–Gln157. The Matthews coefficient (V M; Matthews, 1968 ) was 2.6 Å3 Da−1 and the estimated solvent content was 52.5%. The Ramachandran plot produced by MolProbity showed that 97.6% of the residues were in favored regions, with no outliers.
SPO0140 is an α/β protein comprising two domains (Fig. 1 ). The N-terminal domain (residues 10–93) consists of a β-meander–α-helix–β-meander core (residues 24–92), with the two β-meanders hydrogen bonding along the first and sixth strands to form a twisted mixed six-stranded β-sheet. Two N-terminal helices (H1 and H2) pack against the sheet and complete this domain. The same β3αβ3 unit (β1–3, H3 and β4–6; residues 36–88) is encountered again in the C-terminal domain (β10–12, H5 and β13–15; residues 129–83), with an additional three-stranded meander (β7–9; residues 95–122) packing perpendicularly against the second meander to form a β-sandwich (Fig. 1 a). Both repeats share the same overall fold and topology and their structural comparison is considered to be significant by different alignment methods. Both FATCAT (Ye & Godzik, 2003 ) and DALI (Holm et al., 2008 ) show a statistically significant similarity, with the FATCAT alignment yielding a Cα r.m.s.d. of 3.0 Å over 51 residues (Figs. 2 a and 2 b) and a sequence identity of 8% (DALI Z score 4.0, r.m.s.d. of 3.6 Å over 48 residues, 15% sequence identity). The main difference involves the orientation of the β3 meanders with respect to each other, resulting in different hydrogen-bonding patterns (i.e. the two meanders are connected in the N-terminal domain but form two separate sheets in the C-terminal domain).
Three smaller β3–α-helix (β3α) repeats (residues 36–67, 129–164 and 167–192) were also identified in the structure and can be generated from the first repeat by an approximate 90° anticlockwise rotation along the helix axis between the first and second motif and a clockwise 90° rotation between the second and third (Fig. 2 c). Although the C-terminal helix is both shorter and differently oriented with respect to the meander when compared with the first two β3α repeats (Fig. 2 d), this unit may constitute a minimal supersecondary structure motif for this protein.
The two domains of Sbal_2486 (residues 29–84 and 85–155) show close structural similarity to SPO0140 (Cα r.m.s.d. of 2.8 Å over 127 residues with a sequence identity of 25%). Structural elements missing from Sbal_2486 involve SPO0140 helices H1 and H2 and part of strand β1, as well as the C-terminal β-meander (strands β13–β15) and helix H6 (Fig. 3 a). Thus, although the β3αβ3 repeat described for SPO0140 is present in the N-terminal domain of Sbal_2486, the loss of the C-terminal meander results in a truncated version of this repeat in the second domain. The hypothesis of the fold having originated via duplication of the β3αβ3 repeat is supported by homolog sequence analysis, which shows the N-terminal domain to be more strongly conserved than the C-terminal domain. Helix H1 and the β7–β9 meander can then be viewed as additions to the core repeat that help to stabilize each domain. Alternatively, the fold can also be viewed as consisting of a repetition of β3 units followed either by an α-helix or a turn. The β3-turn repeats (strands β4–β6 and β7–β9) are conserved in both structures, with the conservation of the β7–β9 meander arguing in favor of this possibility. In the case of the β3α motifs, three repeats are encountered in longer homologs, such as SPO0140, and two are found in shorter versions, such as Sbal_2486.
Searches with FATCAT (Ye & Godzik, 2003 ), DALI (Holm et al., 2008 ) or SSM (Krissinel & Henrick, 2004 ) revealed no significant hits for the N-terminal domain of SPO0140 or Sbal_2486. In all three methods, the closest structural neighbor of the SPO0140 C-terminal domain is the PA2021 protein (PDB code 1ywy; Y. C. Lin, G. Liu, Y. Shen, A. Yee, C. H. Arrowsmith & T. Szyperski, unpublished work) from Pseudomonas aeruginosa, with a Cα r.m.s.d. of 3.4 Å over 67 residues and a sequence identity of 12% (Fig. 3 b). The similarity extends over strands β7–β9 and the second β3α motif (strands β10–β12 and helix H5). SCOP classifies PA2021 as a pleckstrin-homology (PH) domain-like barrel. However, both this similarity and the similarity of the same region of SPO0140 to the canonical prokaryotic PH domain (PDB code 3hsa; Joint Center for Structural Genomics, unpublished work; Fig. 3 c) are not statistically significant in all of the algorithms employed (DALI Z score 0.4, r.m.s.d. of 4.6 Å over 42 residues, 10% sequence identity; SSM Z score 1.4, P score 0, r.m.s.d. of 3.2 Å over 53 residues, 4% sequence identity; Laskowski et al., 2005 ) and the overall topologies of the two domains are different, with the PH domain containing an additional N-terminal strand and a longer helix in a different orientation with respect to the DUF1285 β3-helix meander (β10–β12). These considerations led us to classify the SPO0140 and Sbal_2486 structures as a new fold, an assessment that is supported by a preliminary SCOP analysis (Alexey Murzin, personal communication).
SPO0140 crystallized with two monomers (A and B) in the asymmetric unit. The largest packing interface buries ~810 Å2 of solvent-accessible surface per monomer and involves mainly helix H1, strand β1 and the H3–β4 loop from the N-terminal domain and helix H4 and strands β14–β15 from the C-terminal domain. However, packing analysis using PISA (Krissinel & Henrick, 2007 ) suggests that this would not be a stable dimerization interface. The results of analytical size-exclusion chromatography (anSEC) support the assignment of a monomer as the quaternary state in solution for SPO0140.
Sbal_2486 crystallized with one monomer in the asymmetric unit and PISA again suggested a monomer as the most probable oligomeric form. The largest packing interface buries ~760 Å2 of solvent-accessible surface area but is not predicted to be stable for complex formation. However, the results of anSEC coupled with static light scattering (SLS) for Sbal_2486 indicate a dimeric form in solution under the test conditions. This suggests that either the observed crystal-packing interface is stable or the oligomerization state is dependent on the buffer conditions, forming a dimer in the anSEC buffer but disassociating in the crystallization solution prior to crystal formation.
Analysis of SPO0140 and Sbal_2486 using the CastP server (Binkowski et al., 2003 ) revealed that the largest cavity occurs at the interdomain interface. Surface-conservation analysis using ConSurf (Landau et al., 2005 ) showed this to be the largest and most highly conserved contiguous region in the DUF1285 family; it includes loops β1–β2, β6–β7 and strand β4 from the N-terminal domain and strands β10–β12 from the C-terminal domain. No strictly conserved residues are found in this region among DUF1285 homologs, leading us to propose that this region might act as a binding site, but not one that exhibits catalytic activity.
A search against a database of cognate binding sites using IsoCleft (Najmanovich et al., 2008 ), a graph-matching algorithm that searches for similarities in both geometry and chemical composition, identified shared features between the inter-domain cleft of SPO0140 and sugar, phosphate and purine-binding proteins [PDB codes 1pwh (Yamada et al., 2003 ), 1dqa (Istvan et al., 2000 ), 1dm3 (Modis & Wierenga, 2000 ), 1gpe (Wohlfahrt et al., 1999 ), 1v0j (Beis et al., 2005 ), 1hwy (Smith et al., 2001 ), 2vfs (Forneris et al., 2008 ) and 1q6p (Scapin et al., 2003 )]. Similar hits (adenosylcobalamin, heme, dideoxy sugars, NAD, thiamine diphosphate) were obtained for Sbal_2486 (Supplementary Table S11). These similarities, combined with the high conservation observed in this region of the DUF1285 structures and genome-context analysis (see below), indicate that a nucleotide-based ligand may bind along the interdomain interface.
The DUF1285 homolog from Marinobacter sp. ELB17 includes a phosphoglycerate mutase domain (Pfam PF00300) as part of the gene preceding DUF1285. This domain is further annotated as belonging to the subgroup of SixA phosphohistidine phosphatases (IPR004449). In prokaryotes, transcriptional profiling has shown that expression of phosphoglycerate mutase is increased under conditions of oxidative stress (Nodop et al., 2008 ), while the SixA phosphohistidine phosphatase in E. coli has been implicated in signal transduction under conditions of anaerobic respiratory growth (Matsubara & Mizuno, 2000 ). However, as this is the only example of such a domain fusion encountered in the DUF1285 family, the scope for functional speculation is limited.
Several genes predicted (http://string.embl.de) to have functional associations with SPO0140 are located in the same genome neighborhood. Nudix hydrolases are observed in the genome neighborhood of the majority of DUF1285 homologs, including SPO0140. Nudix hydrolases are pyrophosphatases that control the cellular concentrations of a variety of nucleoside diphosphate derivatives, including nucleoside diphosphates and triphosphates and their oxidized forms, dinucleoside polyphosphates, NADH and other signaling compounds (Kraszewska, 2008 ). In plants, these enzymes have been implicated in oxidative signaling (Jambunathan & Mahalingam, 2006 ; Mahalingam et al., 2006 ), including the maintenance of cellular redox homeostasis (Ge et al., 2007 ) and resistance to exogenous reactive oxygen species (Tong et al., 2009 ). In bacteria, nucleotide-based second messengers are involved in a range of signaling functions (Pesavento & Hengge, 2009 ), including the oxidative stress response (Johnstone & Farr, 1991 ). Similarly, the DUF1285 family might carry out a signaling function related to oxidative stress, possibly through binding to a small nucleotide derivative.
The SPO0140 protein family (DUF1285, PF06938) is encountered mainly in proteobacteria and contains around 200 sequence homologs which vary between 150 and 250 residues in length. Availability of more DUF1285 sequences and structures might shed light on the evolutionary history of this intriguing protein family. The information presented here, in combination with further biochemical and biophysical studies, should yield valuable insights into the functional role of SPO0140 and Sbal_2486. Models for SPO0140 and Sbal_2486 homologs can be accessed at http://www1.jcsg.org/cgi-bin/models/get_mor.pl?key=2re3A and http://www1.jcsg.org/cgi-bin/models/get_mor.pl?key=2ra9A, respectively.
Additional information about SPO0140 and Sbal_2486 is available from TOPSAN (Krishna et al., 2010 ) at http://www.topsan.org/explore?PDBid=2re3 and http://www.topsan.org/explore?PDBid=2ra9, respectively.
The first structural representatives of the DUF1285 family revealed a novel fold consisting of repeated motifs. Sequence-conservation and genome-context analysis suggests a signaling role possibly involving binding to a small nucleotide derivative under conditions of oxidative stress.
PDB reference: SPO0140 from Silicibacter pomeroyi DSS-3, 2re3
PDB reference: Sbal_2486 from Shewanella baltica, 2ra9
We thank N. Grishin (University of Texas Southwestern Medical Center), L. Aravind (NCBI, Bethesda), A. Murzin (MRC, Cambridge) and A. Andreeva (MRC, Cambridge) for stimulating discussions. This work was supported by the National Institute of General Medical Sciences, Protein Structure Initiative grant No. U54 GM074898. RJN is a member of the FRSQ-funded Centre de Recherche Clinique Étienne-Le Bel. Portions of this research were carried out at the Stanford Synchrotron Radiation Lightsource (SSRL) and Advanced Light Source (ALS). The SSRL is a national user facility operated by Stanford University on behalf of the US Department of Energy, Office of Basic Energy Sciences. The SSRL Structural Molecular Biology Program is supported by the Department of Energy, Office of Biological and Environmental Research and by the National Institutes of Health (National Center for Research Resources, Biomedical Technology Program and the National Institute of General Medical Sciences). The ALS is supported by the Director, Office of Science, Office of Basic Energy Sciences, Materials Sciences Division of the US Department of Energy under Contract No. DE-AC03-76SF00098 at Lawrence Berkeley National Laboratory. Genomic DNA from Silicibacter pomeroyi DSS-3 (ATCC 700808D) was obtained from the American Type Culture Collection (ATCC). Shewanella baltica OS155 was a gift from Dr Jim K. Fredrickson, Pacific Northwest National Laboratory, Washington, USA. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of General Medical Sciences or the National Institutes of Health.
1Supplementary material has been deposited in the IUCr electronic archive (Reference: WD5122).