Translation initiation factor 2 (IF2) is involved in the early steps of bacterial protein synthesis. It promotes the stabilization of the initiator tRNA on the 30S initiation complex (IC) and triggers GTP hydrolysis upon ribosomal subunit joining. While the structure of an archaeal homologue (a/eIF5B) is known, there are significant sequence and functional differences in eubacterial IF2, while the trimeric eukaryotic IF2 is completely unrelated. Here, the crystal structure of the apo IF2 protein core from Thermus thermophilus has been determined by MAD phasing and the structures of GTP and GDP complexes were also obtained. The IF2–GTP complex was trapped by soaking with GTP in the cryoprotectant. The structures revealed conformational changes of the protein upon nucleotide binding, in particular in the P-loop region, which extend to the functionally relevant switch II region. The latter carries a catalytically important and conserved histidine residue which is observed in different conformations in the GTP and GDP complexes. Overall, this work provides the first crystal structure of a eubacterial IF2 and suggests that activation of GTP hydrolysis may occur by a conformational repositioning of the histidine residue.

The recently identified plant Bcl-2-associated athanogene (BAG) family plays an extensive role in plant programmed cell death (PCD) processes ranging from growth and development to stress responses and even cell death. In the Arabidopsis thaliana BAG (AtBAG) protein family, four members (AtBAG1–4) have a domain organization similar to that of mammalian BAG proteins. Here, crystal structures of the BAG domains (BDs) of AtBAG1–4 have been determined; they have high homology and adopt a structure comprising three short parallel α-helices, similar to some mammalian BAG proteins. The crystal structure of a complex of the AtBAG1 ubiquitin-like domain and BAG domain (UBD) with the Hsc70 nucleotide-binding domain (NBD) was also determined. The binding of the AtBAG1 BD to the Hsc70 NBD induces conformational change of the Hsc70 NBD to the open state and reduces the affinity of the NBD for ADP. In vivo studies showed that bag2-1 mutant plants are larger than wild-type plants when growing under normal conditions, indicating that the AtBAG proteins might regulate plant PCD and confer tolerance to stresses in plants. These structural and functional analyses indicate that the AtBAG proteins function as nucleotide-exchange factors for Hsp70/Hsc70 in A. thaliana and that the mechanism of regulation of chaperone-mediated protein folding is conserved in plants.

Dynamic behavior of proteins is critical to their function. X-ray crystallography, a powerful yet mostly static technique, faces inherent challenges in acquiring dynamic information despite decades of effort. Dynamic `structural changes' are often indirectly inferred from `structural differences' by comparing related static structures. In contrast, the direct observation of dynamic structural changes requires the initiation of a biochemical reaction or process in a crystal. Both the direct and the indirect approaches share a common challenge in analysis: how to interpret the structural heterogeneity intrinsic to all dynamic processes. This paper presents a real-space approach to this challenge, in which a suite of analytical methods and tools to identify and refine the mixed structural species present in multiple crystallographic data sets have been developed. These methods have been applied to representative scenarios in dynamic crystallography, and reveal structural information that is otherwise difficult to interpret or inaccessible using conventional methods.

The βγ-crystallin superfamily includes highly diverse proteins belonging to all of the kingdoms of life. Based on structural topology, these proteins are considered to be evolutionarily related to the long-lived βγ-crystallins that constitute the vertebrate eye lens. This study reports the crystallographic structure at 0.99 Å resolution of the two-domain βγ-crystallin (geodin) from the sponge Geodia cydonium. This is the most ancient member of the βγ-crystallin superfamily in metazoans. The X-ray structure shows that the geodin domains adopt the typical βγ-crystallin fold with a paired Greek-key motif, thus confirming the hypothesis that the crystallin-type scaffold used in the evolution of bacteria and moulds was recruited very early in metazoans. As a significant new structural feature, the sponge protein possesses a unique interdomain interface made up by pairing between the second motif of the first domain and the first motif of the second domain. The atomic resolution also allowed a detailed analysis of the calcium-binding site of the protein.

Circular permutation of streptavidin was carried out in order to investigate the role of a main-chain amide in stabilizing the high-affinity complex of the protein and biotin. Mutant proteins CP49/48 and CP50/49 were constructed to place new N-termini at residues 49 and 50 in a flexible loop involved in stabilizing the biotin complex. Crystal structures of the two mutants show that half of each loop closes over the binding site, as observed in wild-type streptavidin, while the other half adopts the open conformation found in the unliganded state. The structures are consistent with kinetic and thermodynamic data and indicate that the loop plays a role in enthalpic stabilization of the bound state via the Asn49 amide–biotin hydrogen bond. In wild-type streptavidin, the entropic penalties of immobilizing a flexible portion of the protein to enhance binding are kept to a manageable level by using a contiguous loop of medium length (six residues) which is already constrained by its anchorage to strands of the β-barrel protein. A molecular-dynamics simulation for CP50/49 shows that cleavage of the binding loop results in increased structural fluctuations for Ser45 and that these fluctuations destabilize the streptavidin–biotin complex.

A series of bovine insulin samples were obtained as 14 polycrystalline precipitates at room temperature in the pH range 5.0–7.6. High-resolution powder X-ray diffraction data were collected to reveal the T₆ hexameric insulin form. Sample homogeneity and reproducibility were verified by additional synchrotron measurements using an area detector. Pawley analyses of the powder patterns displayed pH- and radiation-induced anisotropic lattice modifications. The pronounced anisotropic lattice variations observed for T₆ insulin were exploited in a 14-data-set Rietveld refinement to obtain an average crystal structure over the pH range investigated. Only the protein atoms of the known structure with PDB code 2a3g were employed in our starting model. A novel approach for refining protein structures using powder diffraction data is presented. In this approach, each amino acid is represented by a flexible rigid body (FRB). The FRB model requires a significantly smaller number of refinable parameters and restraints than a fully free-atom refinement. A total of 1542 stereochemical restraints were imposed in order to refine the positions of 800 protein atoms, two Zn atoms and 44 water molecules in the asymmetric unit using experimental data in the resolution range 18.2–2.7 Å for all profiles.

Ribonuclease from Bacillus intermedius (binase) is a small basic protein with antitumour activity. The three-dimensional structure of the binase mutant form Glu43Ala/Phe81Ala was determined at 1.98 Å resolution and its functional properties, such as the kinetic parameters characterizing the hydrolysis of polyinosinic acid and cytotoxicity towards Kasumi-1 cells, were investigated. In all crystal structures of binase studied previously the characteristic dimer is present, with the active site of one subunit being blocked owing to interactions within the dimer. In contrast to this, the new mutant form is not dimeric in the crystal. The catalytic efficiency of the mutant form is increased 1.7-fold and its cytotoxic properties are enhanced compared with the wild-type enzyme.

The structure of the large ribosomal subunit from the halophilic archaeon Haloarcula marismortui (Hma) is the only crystal structure of an archaeal ribosomal particle that has been determined to date. However, the first model of the Hma 50S ribosomal subunit contained some gaps: the structures of functionally important mobile lateral protuberances were not visualized. Subsequently, some parts of the P (L12) stalk base were visualized at 3.0 Å resolution [Kavran & Steitz (2007), J. Mol. Biol.371, 1047–1059]: the RNA-binding domain of r-protein P0 (L10), the C-terminal domain of L11 and helices 43 and 44 of the 23 S rRNA. Here, the 2.4 Å resolution electron-density map of the Hma 50S ribosomal subunit was revisited and approximately two-thirds of the P0 protein, residues 1–58 of the N-terminal domains of two P1 protein molecules, residues 130–156 of L11, the full-length r-protein LX, nucleotides 2137–2149 and 2226–2237 of the 23S rRNA helix H76 forming the L1 stalk, nucleotides 2339–2343 of the 23S rRNA (contacting L5 protein) and loops 29–34 and 108–128 of protein L5 could be visualized. Thus, this paper provides a supplemented version of the Hma 50S ribosomal subunit model.

The yellow fluorescent protein phiYFPv (λ_em^max≃ 537 nm) with improved folding has been developed from the spectrally identical wild-type phiYFP found in the marine jellyfish Phialidium. The latter fluorescent protein is one of only two known cases of naturally occurring proteins that exhibit emission spectra in the yellow–orange range (535–555 nm). Here, the crystal structure of phiYFPv has been determined at 2.05 Å resolution. The `yellow' chromophore formed from the sequence triad Thr65-Tyr66-Gly67 adopts the bicyclic structure typical of fluorophores emitting in the green spectral range. It was demonstrated that perfect antiparallel π-stacking of chromophore Tyr66 and the proximal Tyr203, as well as Val205, facing the chromophore phenolic ring are chiefly responsible for the observed yellow emission of phiYFPv at 537 nm. Structure-based site-directed mutagenesis has been used to identify the key functional residues in the chromophore environment. The obtained results have been utilized to improve the properties of phiYFPv and its homologous monomeric biomarker tagYFP.

Starch, a polymer of glucose, is the major source of calories in the human diet. It has numerous industrial uses, including as a raw material for the production of first-generation bioethanol. Several classes of enzymes take part in starch biosynthesis, of which starch synthases (SSs) carry out chain elongation of both amylose and amylopectin. Plants have five classes of SS, each with different roles. The products of the reaction of SS are well known, but details of the reaction mechanism remain obscure and even less is known of how different SSs select different substrates for elongation, how they compete with each other and how their activities are regulated. Here, the first crystal structure of a soluble starch synthase is presented: that of starch synthase I (SSI) from barley refined to 2.7 Å resolution. The structure captures an open conformation of the enzyme with a surface-bound maltooligosaccharide and a disulfide bridge that precludes formation of the active site. The maltooligosaccharide-binding site is involved in substrate recognition, while the disulfide bridge is reflective of redox regulation of SSI. Activity measurements on several SSI mutants supporting these roles are also presented.

The members of the Dicistroviridae family are non-enveloped positive-sense single-stranded RNA (+ssRNA) viruses pathogenic to beneficial arthropods as well as insect pests of medical importance. Triatoma virus (TrV), a member of this family, infects several species of triatomine insects (popularly named kissing bugs), which are vectors for human trypanosomiasis, more commonly known as Chagas disease. The potential use of dicistroviruses as biological control agents has drawn considerable attention in the past decade, and several viruses of this family have been identified, with their targets covering honey bees, aphids and field crickets, among others. Here, the crystal structure of the TrV capsid at 2.5 Å resolution is reported, showing that as expected it is very similar to that of Cricket paralysis virus (CrPV). Nevertheless, a number of distinguishing structural features support the introduction of a new genus (Triatovirus; type species TrV) under the Dicistroviridae family. The most striking differences are the absence of icosahedrally ordered VP4 within the infectious particle and the presence of prominent projections that surround the fivefold axis. Furthermore, the structure identifies a second putative autoproteolytic DDF motif in protein VP3, in addition to the conserved one in VP1 which is believed to be responsible for VP0 cleavage during capsid maturation. The potential meaning of these new findings is discussed.

VLD (vive la difference) is a novel ab initio phasing approach that is able to drive random phases to the correct values. It has been applied to small, medium and protein structures provided that the data resolution was atomic. It has never been used for non-ab initio cases in which some phase information is available but the data resolution is usually very far from 1 Å. In this paper, the potential of VLD is tested for the first time for a classical non-ab initio problem: molecular replacement. Good preliminary experimental results encouraged the construction of a pipeline for leading partial molecular-replacement models with errors to refined solutions in a fully automated way. The pipeline moduli and their interaction are described, together with applications to a wide set of test cases.

The direct oxygen sensor DosP is a multidomain protein that contains a gas-sensing haem domain and an EAL effector domain. EAL domains are omnipresent signal transduction domains in bacteria. Many EAL domains are active phosphodiesterases and are involved in breakdown of the ubiquitous bacterial second messenger cyclic di-GMP. Despite a great deal of information on the functional and structural aspects of active and inactive EAL domains, little is known about the structural basis of their regulation by their associated sensory domains. Here, two crystal structures of the Escherichia coli DosP EAL domain derived from cubic and monoclinic crystal forms that were obtained under tartrate and PEG conditions, respectively, are described. Both of the structures display the typical TIM (triosephosphate isomerase) barrel fold with one antiparallel β-strand. However, unlike other EAL structures, access to the active site in DosP EAL is sterically restricted by the presence of a short helical stretch (Ser637-Ala-Leu-His640) in loop L3 between strand β3 and helix α3. This element, together with an unordered fragment, replaces the short α-helix (named α5 in Tbd1265 EAL) that is found in other EAL-domain structures. Since DosP EAL is an active c-di-GMP phosphodiesterase, the observed inactive conformation is suggested to be of functional relevance for the regulation mechanism of DosP.

The vault particle, with a molecular weight of about 10 MDa, is the largest ribonucleoprotein that has been described. The X-ray structure of intact rat vault has been solved at a resolution of 3.5 Å [Tanaka et al. (2009), Science, 323, 384–388], showing an overall barrel-shaped architecture organized into two identical moieties, each consisting of 39 copies of the major vault protein (MVP). The model deposited in the PDB includes 39 MVP copies (half a vault) in the crystal asymmetric unit. A 2.1 Å resolution structure of the seven N-terminal repeats (R1–7) of MVP has also been determined [Querol-Audíet al. (2009), EMBO J.28, 3450–3457], revealing important discrepancies with respect to the MVP models for repeats R1 and R2. Here, the re-refinement of the vault structure by incorporating the high-resolution information available for the R1–7 domains, using the deformable elastic network (DEN) approach and maintaining strict 39-fold noncrystallographic symmetry is reported. The new refinement indicates that at the resolution presently available the MVP shell can be described well as only one independent subunit organized with perfect D39 molecular symmetry. This refinement reveals that significant rearrangements occur in the N-terminus of MVP during the closing of the two vault halves and that the 39-fold symmetry breaks in the cap region. These results reflect the highly dynamic nature of the vault structure and represent a necessary step towards a better understanding of the biology and regulation of this particle.

The toolbox for computational protein crystallography is full of easy-to-use applications for the routine solution and refinement of periodic diffraction data sets and protein structures. There is a gap in the available software when it comes to aperiodic crystallographic data. Current protein crystallography software cannot handle modulated data, and small-molecule software for aperiodic crystallography cannot work with protein structures. To adapt software for modulated protein data requires training data to test and debug the changed software. Thus, a comprehensive training data set consisting of atomic positions with associated modulation functions and the modulated structure factors packaged as both a three-dimensional supercell and as a modulated structure in (3+1)D superspace has been created. The (3+1)D data were imported into Jana2006; this is the first time that this has been performed for protein data.

The crystal structure of a 75 kDa central fragment of GBS104, a tip pilin from the 2063V/R strain of Streptococcus agalactiae (group B streptococcus; GBS), is reported. In addition, a homology model of the remaining two domains of GBS104 was built and a model of full-length GBS104 was generated by combining the homology model (the N1 and N4 domains) and the crystal structure of the 75 kDa fragment (the N2 and N3 domains). This rod-shaped GBS104 model is constructed of three IgG-like domains (the N1, N2 and N4 domains) and one vWFA-like domain (the N3 domain). The N1 and N2 domains of GBS104 are assembled with distinct and remote segments contributed by the N- and C-termini. The metal-binding site in the N3 domain of GBS104 is in the closed/low-affinity conformation. Interestingly, this domain hosts two long arms that project away from the metal-binding site. Using site-directed mutagenesis, two cysteine residues that lock the N3 domain of GBS104 into the open/high-affinity conformation were introduced. Both wild-type and disulfide-locked recombinant proteins were tested for binding to extracellular matrix proteins such as collagen, fibronectin, fibrinogen and laminin, and an increase in fibronectin binding affinity was identified for the disulfide-locked N3 domain, suggesting that induced conformational changes may play a possible role in receptor binding.

Uridine at position 34 of bacterial transfer RNAs is commonly modified to uridine-5-oxyacetic acid (cmo⁵U) to increase the decoding capacity. The protein CmoA is involved in the formation of cmo⁵U and was annotated as an S-adenosyl-L-methionine-dependent (SAM-dependent) methyltransferase on the basis of its sequence homology to other SAM-containing enzymes. However, both the crystal structure of Escherichia coli CmoA at 1.73 Å resolution and mass spectrometry demonstrate that it contains a novel cofactor, S-adenosyl-S-carboxymethyl-L-homocysteine (SCM-SAH), in which the donor methyl group is substituted by a carboxymethyl group. The carboxyl moiety forms a salt-bridge interaction with Arg199 that is conserved in a large group of CmoA-related proteins but is not conserved in other SAM-containing enzymes. This raises the possibility that a number of enzymes that have previously been annotated as SAM-dependent are in fact SCM-SAH-dependent. Indeed, inspection of electron density for one such enzyme with known X-ray structure, PDB entry 1im8, suggests that the active site contains SCM-SAH and not SAM.

Lyme disease is a tick-borne infection caused by the transmission of Borrelia burgdorferi from infected Ixodes ticks to a mammalian host during the blood meal. Previous studies have shown that the expression of B. burgdorferi surface-localized lipoproteins, which include BBA64, is up-regulated during the process of tick feeding. Although the exact function of BBA64 is not known, this lipoprotein is critical for the transmission of the spirochete from the tick salivary glands to the mammalian organism after a tick bite. Since the mechanism of development of the disease and the functions of the surface lipoproteins associated with borreliosis are still poorly understood, the crystal structure of the B. burgdorferi outer surface lipoprotein BBA64 was solved at 2.4 Å resolution in order to obtain a better insight into the pathogenesis of B. burgdorferi and to promote the discovery of novel potential preventive drugs against Lyme disease. In this study, the crystal structure of BBA64 was also compared with that of the paralogous protein CspA (also referred to as BbCRASP-1, CRASP-1 or BBA68). CspA is the complement regulator-acquiring surface protein-1 of B. burgdorferi; its structure is known, but its function apparently differs from that of BBA64. It is demonstrated that unlike the homologous CspA, BBA64 does not form a homodimer. Their differences in function could be explained by divergence in their amino-acid sequences, electrostatic surface potentials and overall tertiary structures. The C-terminal part of BBA64 has a different conformation to that of CspA; the conformation of this region is essential for the proper function of CspA.

2-Haloacid dehalogenases (2-HADs) catalyse the hydrolytic dehalogenation of 2-haloalkanoic acids, cleaving the carbon–halide bond at the C^α-atom position and releasing a halogen atom. These enzymes are of interest for their potential use in bioremediation and in the synthesis of industrial chemicals. Here, the crystal structure of 2-HAD from Pseudomonas syringae pv. tomato DC3000 (ps-2-HAD) at 1.98 Å resolution solved using the single-wavelength anomalous dispersion method is reported. The ps-2-HAD molecule consists of two structurally distinct domains: the core domain and the subdomain. Enzymatic activity analysis of ps-2-HAD revealed its capacity to catalyse the dehalogenation of both L- and D-substrates; however, the structure of ps-2-HAD is completely different from that of DehI, which is the only DL-2-HAD enzyme that has been structurally characterized, but shows similar overall folding to L-HADs. Single mutations of four amino-acid residues at the putative active site showed that they are related to its enzymatic activity, yet three of them are nonconserved among HADs. These observations imply that ps-2-HAD has a novel active site and a unique catalytic behaviour compared with other HADs. This study provides a structural basis and biochemical evidence for further elucidation of the catalytic mechanism of 2-HAD.

Despite truly impressive achievements in the global battle against HIV there remains a need for new drugs directed against novel targets, and the viral capsid protein (CA) may represent one such target. Intense structural characterization of CA over the last two decades has provided unprecedented insight into the structure and assembly of this key viral protein. Furthermore, several inhibitor-binding sites that elicit antiviral activity have been reported on CA, two of which are located on its N-terminal domain (CA_NTD). In this work, the binding of a novel capsid-assembly inhibitor that targets a unique inhibitory site on CA_NTD is reported. Moreover, whereas cocrystallization of CA_NTD in complex with ligands has proven to be challenging in the past, the use of this inhibitor as a tool compound is shown to vastly facilitate ternary cocrystallizations with CA_NTD. This improvement in crystallization is likely to be achieved through the formation of a compound-mediated homodimer, the intrinsic symmetry of which greatly increases the prospect of generating a crystal lattice. While protein engineering has been used in the literature to support a link between the inherent symmetry of a macromolecule and its propensity to crystallize, to our knowledge this work represents the first use of a synthetic ligand for this purpose.

The aspartic protease BACE2 is responsible for the shedding of the transmembrane protein Tmem27 from the surface of pancreatic β-cells, which leads to inactivation of the β-cell proliferating activity of Tmem27. This role of BACE2 in the control of β-cell maintenance suggests BACE2 as a drug target for diabetes. Inhibition of BACE2 has recently been shown to lead to improved control of glucose homeostasis and to increased insulin levels in insulin-resistant mice. BACE2 has 52% sequence identity to the well studied Alzheimer's disease target enzyme β-secretase (BACE1). High-resolution BACE2 structures would contribute significantly to the investigation of this enzyme as either a drug target or anti-target. Surface mutagenesis, BACE2-binding antibody Fab fragments, single-domain camelid antibody V_HH fragments (Xaperones) and Fyn-kinase-derived SH3 domains (Fynomers) were used as crystallization helpers to obtain the first high-resolution structures of BACE2. Eight crystal structures in six different packing environments define an ensemble of low-energy conformations available to the enzyme. Here, the different strategies used for raising and selecting BACE2 binders for cocrystallization are described and the crystallization success, crystal quality and the time and resources needed to obtain suitable crystals are compared.

Staphylococci cause a wide range of diseases in humans and animals, and the proteins of the multiple antibiotic-resistance repressor (MarR) family in staphylococci function as regulators of protein expression and confer resistance to multiple antibiotics. Diverse mechanisms such as biofilm formation, drug transport, drug modification etc. are associated with this resistance. In this study, crystal structures of the Staphylococcus aureus MarR homologue SAR2349 and its complex with salicylate and the aminoglycoside antibiotic kanamycin have been determined. The structure of SAR2349 shows for the first time that a MarR protein can interact directly with different classes of ligands simultaneously and highlights the importance and versatility of regulatory systems in bacterial antibiotic resistance. The three-dimensional structures of TcaR from S. epidermidis in complexes with chloramphenicol and with the aminoglycoside antibiotic streptomycin were also investigated. The crystal structures of the TcaR and SAR2349 complexes illustrate a general antibiotic-regulated resistance mechanism that may extend to other MarR proteins. To reveal the regulatory mechanism of the MarR proteins, the protein structures of this family were further compared and three possible mechanisms of regulation are proposed. These results are of general interest because they reveal a remarkably broad spectrum of ligand-binding modes of the multifunctional MarR proteins. This finding provides further understanding of antimicrobial resistance mechanisms in pathogens and strategies to develop new therapies against pathogens.

SspCA, a novel `extremo-α-carbonic anhydrase' isolated from the thermophilic bacterium Sulfurihydrogenibium yellowstonense YO3AOP1, is an efficient catalyst for the hydration of CO₂ and presents exceptional thermostability. Indeed, SspCA retains a high catalytic activity even after being heated to 343–373 K for several hours. Here, the crystallographic structure of this α-carbonic anhydrase (α-CA) is reported and the factors responsible for its function at high temperature are elucidated. In particular, the study suggests that increased structural compactness, together with an increased number of charged residues on the protein surface and a greater number of ionic networks, seem to be the key factors involved in the higher thermostability of this enzyme with respect to its mesophilic homologues. These findings are of extreme importance, since they provide a structural basis for the understanding of the mechanisms responsible for thermal stability in the α-CA family for the first time. The data obtained offer a tool that can be exploited to engineer α-CAs in order to obtain enzymes with enhanced thermostability for use in the harsh conditions of the CO₂ capture and sequestration processes.

Dual-specificity phosphatases (DUSPs) play an important role in regulating cellular signalling pathways governing cell growth, differentiation and apoptosis. Human DUSP26 inhibits the apoptosis of cancer cells by dephosphorylating substrates such as p38 and p53. High-resolution crystal structures of the DUSP26 catalytic domain (DUSP26-C) and its C152S mutant [DUSP26-C (C152S)] have been determined at 1.67 and 2.20 Å resolution, respectively. The structure of DUSP26-C showed a novel type of domain-swapped dimer formed by extensive crossover of the C-terminal α7 helix. Taken together with the results of a phosphatase-activity assay, structural comparison with other DUSPs revealed that DUSP26-C adopts a catalytically inactive conformation of the protein tyrosine phosphate-binding loop which significantly deviates from that of canonical DUSP structures. In particular, a noticeable difference exists between DUSP26-C and the active forms of other DUSPs at the hinge region of a swapped C-terminal domain. Additionally, two significant gaps were identified between the catalytic core and its surrounding loops in DUSP26-C, which can be exploited as additional binding sites for allosteric enzyme regulation. The high-resolution structure of DUSP26-C may thus provide structural insights into the rational design of DUSP26-targeted anticancer drugs.

In the past decade many structures of nucleic acids have been determined, which have contributed to our understanding of their biological functions. However, crystals containing nucleic acids often diffract X-rays poorly. This makes electron-density interpretation difficult and requires a great deal of expertise in crystallography and knowledge of nucleic acid structure. Here, new programs called NAFIT and NABUILD for fitting and extending nucleic acid models are presented. These programs can be used as modules in the automated refinement system LAFIRE, as well as acting as independent programs. NAFIT performs sequential grouped fitting with empirical torsion-angle restraints and antibumping restraints including H atoms. NABUILD extends the model using a skeletonized map in a coarse-grained manner. It has been shown that NAFIT greatly improves electron-density fit and geometric quality and that iterative refinement with NABUILD significantly reduces the R_free factor.

X-ray crystal structures of the spermine⁴⁺ form of the Z-DNA duplex with the self-complementary d(CG)₃ sequence in complexes with Mn²⁺ and Zn²⁺ cations have been determined at the ultrahigh resolutions of 0.75 and 0.85 Å, respectively. Stereochemical restraints were only used for the sperminium cation (in both structures) and for nucleotides with dual conformation in the Zn²⁺ complex. The Mn²⁺ and Zn²⁺ cations at the major site, designated M²⁺(1), bind at the N7 position of G6 by direct coordination. The coordination geometry of this site was octahedral, with complete hydration shells. An additional Zn²⁺(2) cation was bis-coordinated in a tetrahedral fashion by the N7 atoms of G10 and G12 from a symmetry-related molecule. The coordination distances of Zn²⁺(1) and Zn²⁺(2) to the O6 atom of the guanine residues were 3.613 (6) and 3.258 (5) Å, respectively. Moreover, a chloride ion was also identified in the coordination sphere of Zn²⁺(2). Alternate conformations were observed in the Z-DNA–Zn²⁺ structure not only at internucleotide linkages but also at the terminal C3′—OH group of G12. The conformation of the sperminium chain in the Z-DNA–Mn²⁺ complex is similar to the spermine⁴⁺ conformation in analogous Z-DNA–Mg²⁺ structures. In the Z-DNA–Zn²⁺ complex the sperminium cation is disordered and partially invisible in electron-density maps. In the Z-DNA–Zn²⁺ complex the sperminium cation only interacts with the phosphate groups of the Z-DNA molecules, while in the Z-DNA–Mn²⁺ structure it forms hydrogen bonds to both the phosphate groups and DNA bases.

The two structures reported in the article by Chin et al. [(2013). Acta Cryst. D69, 352–366] have been further refined and corrected.

The article by Koval' et al. [(2013). Acta Cryst. D69, 213226] is corrected.