Deep-learning map segmentation for protein X-ray crystallographic structure determination.

When solving a structure of a protein from single-wavelength anomalous diffraction X-ray data, the initial phases obtained by phasing from an anomalously scattering substructure usually need to be improved by an iterated electron-density modification. In this manuscript, the use of convolutional neural networks (CNNs) for segmentation of the initial experimental phasing electron-density maps is proposed. The results reported demonstrate that a CNN with U-net architecture, trained on several thousands of electron-density maps generated mainly using X-ray data from the Protein Data Bank in a supervised learning, can improve current density-modification methods.

Multivariate estimation of substructure amplitudes for a single-wavelength anomalous diffraction experiment.

To determine a substructure from single-wavelength anomalous diffraction (SAD) data using Patterson or direct methods, the substructure-factor amplitude (|Fa|) is first estimated. Currently, the absolute value of the Bijvoet difference is widely used as an estimate of |Fa| values for SAD data. Here, an equation is derived from multivariate statistics and tested that takes into account the correlation between the observed positive (F+) and negative (F-) Friedel pairs and Fa along with measurement errors in the observed data. The multivariate estimation of |Fa| has been implemented in a new program, Afro. Results on over 180 test cases show that Afro provides a higher correlation to the final substructure-factor amplitudes (calculated from the refined, final substructures) than the Bijvoet differences and improves the robustness of direct-methods substructure detection.

CCP4 Cloud for structure determination and project management in macromolecular crystallography.

Nowadays, progress in the determination of three-dimensional macromolecular structures from diffraction images is achieved partly at the cost of increasing data volumes. This is due to the deployment of modern high-speed, high-resolution detectors, the increased complexity and variety of crystallographic software, the use of extensive databases and high-performance computing. This limits what can be accomplished with personal, offline, computing equipment in terms of both productivity and maintainability. There is also an issue of long-term data maintenance and availability of structure-solution projects as the links between experimental observations and the final results deposited in the PDB. In this article, CCP4 Cloud, a new front-end of the CCP4 software suite, is presented which mitigates these effects by providing an online, cloud-based environment for crystallographic computation. CCP4 Cloud was developed for the efficient delivery of computing power, database services and seamless integration with web resources. It provides a rich graphical user interface that allows project sharing and long-term storage for structure-solution projects, and can be linked to data-producing facilities. The system is distributed with the CCP4 software suite version 7.1 and higher, and an online publicly available instance of CCP4 Cloud is provided by CCP4.

A new MR-SAD algorithm for the automatic building of protein models from low-resolution X-ray data and a poor starting model.

Determining macromolecular structures from X-ray data with resolution worse than 3 Šremains a challenge. Even if a related starting model is available, its incompleteness or its bias together with a low observation-to-parameter ratio can render the process unsuccessful or very time-consuming. Yet, many biologically important macromolecules, especially large macromolecular assemblies, membrane proteins and receptors, tend to provide crystals that diffract to low resolution. A new algorithm to tackle this problem is presented that uses a multivariate function to simultaneously exploit information from both an initial partial model and low-resolution single-wavelength anomalous diffraction data. The new approach has been used for six challenging structure determinations, including the crystal structures of membrane proteins and macromolecular complexes that have evaded experts using other methods, and large structures from a 3.0 Šresolution F1-ATPase data set and a 4.5 Šresolution SecYEG-SecA complex data set. All of the models were automatically built by the method to Rfree values of between 28.9 and 39.9% and were free from the initial model bias.

CCP4i2: the new graphical user interface to the CCP4 program suite.

The CCP4 (Collaborative Computational Project, Number 4) software suite for macromolecular structure determination by X-ray crystallography groups brings together many programs and libraries that, by means of well established conventions, interoperate effectively without adhering to strict design guidelines. Because of this inherent flexibility, users are often presented with diverse, even divergent, choices for solving every type of problem. Recently, CCP4 introduced CCP4i2, a modern graphical interface designed to help structural biologists to navigate the process of structure determination, with an emphasis on pipelining and the streamlined presentation of results. In addition, CCP4i2 provides a framework for writing structure-solution scripts that can be built up incrementally to create increasingly automatic procedures.

Substructure determination using phase-retrieval techniques.

Thus far, the application of phase-retrieval methods in crystallography has mainly been aimed at variants of charge flipping or structure-factor flipping. In this work, the relaxed averaged alternating reflections (RAAR) algorithm is applied to determine anomalously scattering substructures from single-wavelength anomalous diffraction (SAD) data of macromolecules. The algorithm has been implemented in a new program, PRASA, and has been shown to significantly outperform charge flipping in determining anomalously scattering substructures on a test sample of 169 SAD data sets with resolutions up to 3.88 Å.

Ankyrin repeats as a dimerization module.

Legionella pneumophila is a pathogen, causing severe pneumonia in humans called Legionnaires' disease. AnkC (LegA12) is a poorly characterized 495-residue effector protein conserved in multiple Legionella species. Here, we report the crystal structure of a C-terminally truncated AnkC (2-384) at 3.2 Å resolution. The structure shows seven ankyrin repeats (ARs) with unique structural features. AnkC forms a dimer along the outer surface of loops between ARs. The dimer exists both in the crystal form and in solution, as shown by analytical ultracentrifugation. This is the first example of ARs as a dimerization module as opposed to solely a protein interaction domain. In addition, a novel α-helix insert between AR3-AR4 is positioned across the surface opposite the ankyrin groove. Sequence conservation suggests that the ankyrin groove of AnkC is a functional site that interacts with binding targets. This ankyrin domain structure is an important step towards a functional characterization of AnkC.

The dynamic complex of cytochrome c6 and cytochrome f studied with paramagnetic NMR spectroscopy.

The rapid transfer of electrons in the photosynthetic redox chain is achieved by the formation of short-lived complexes of cytochrome b6f with the electron transfer proteins plastocyanin and cytochrome c6. A balance must exist between fast intermolecular electron transfer and rapid dissociation, which requires the formation of a complex that has limited specificity. The interaction of the soluble fragment of cytochrome f and cytochrome c6 from the cyanobacterium Nostoc sp. PCC 7119 was studied using NMR spectroscopy and X-ray diffraction. The crystal structures of wild type, M58H and M58C cytochrome c6 were determined. The M58C variant is an excellent low potential mimic of the wild type protein and was used in chemical shift perturbation and paramagnetic relaxation NMR experiments to characterize the complex with cytochrome f. The interaction is highly dynamic and can be described as a pure encounter complex, with no dominant stereospecific complex. Ensemble docking calculations and Monte-Carlo simulations suggest a model in which charge-charge interactions pre-orient cytochrome c6 with its haem edge toward cytochrome f to form an ensemble of orientations with extensive contacts between the hydrophobic patches on both cytochromes, bringing the two haem groups sufficiently close to allow for rapid electron transfer. This model of complex formation allows for a gradual increase and decrease of the hydrophobic interactions during association and dissociation, thus avoiding a high transition state barrier that would slow down the dissociation process.

Automatic protein structure solution from weak X-ray data.

Determining new protein structures from X-ray diffraction data at low resolution or with a weak anomalous signal is a difficult and often an impossible task. Here we propose a multivariate algorithm that simultaneously combines the structure determination steps. In tests on over 140 real data sets from the protein data bank, we show that this combined approach can automatically build models where current algorithms fail, including an anisotropically diffracting 3.88 Å RNA polymerase II data set. The method seamlessly automates the process, is ideal for non-specialists and provides a mathematical framework for successfully combining various sources of information in image processing.

Recent advances in the CRANK software suite for experimental phasing.

For its first release in 2004, CRANK was shown to effectively detect and phase anomalous scatterers from single-wavelength anomalous diffraction data. Since then, CRANK has been significantly improved and many more structures can be built automatically with single- or multiple-wavelength anomalous diffraction or single isomorphous replacement with anomalous scattering data. Here, the new algorithms that have been developed that have led to these substantial improvements are discussed and CRANK's performance on over 100 real data sets is shown. The latest version of CRANK is freely available for download at http://www.bfsc.leidenuniv.nl/software/crank/ and from CCP4 (http://www.ccp4.ac.uk/).

Reduction of density-modification bias by ß correction.

Density modification often suffers from an overestimation of phase quality, as seen by escalated figures of merit. A new cross-validation-based method to address this estimation bias by applying a bias-correction parameter 'ß' to maximum-likelihood phase-combination functions is proposed. In tests on over 100 single-wavelength anomalous diffraction data sets, the method is shown to produce much more reliable figures of merit and improved electron-density maps. Furthermore, significantly better results are obtained in automated model building iterated with phased refinement using the more accurate phase probability parameters from density modification.

REFMAC5 for the refinement of macromolecular crystal structures.

This paper describes various components of the macromolecular crystallographic refinement program REFMAC5, which is distributed as part of the CCP4 suite. REFMAC5 utilizes different likelihood functions depending on the diffraction data employed (amplitudes or intensities), the presence of twinning and the availability of SAD/SIRAS experimental diffraction data. To ensure chemical and structural integrity of the refined model, REFMAC5 offers several classes of restraints and choices of model parameterization. Reliable models at resolutions at least as low as 4 Šcan be achieved thanks to low-resolution refinement tools such as secondary-structure restraints, restraints to known homologous structures, automatic global and local NCS restraints, `jelly-body' restraints and the use of novel long-range restraints on atomic displacement parameters (ADPs) based on the Kullback-Leibler divergence. REFMAC5 additionally offers TLS parameterization and, when high-resolution data are available, fast refinement of anisotropic ADPs. Refinement in the presence of twinning is performed in a fully automated fashion. REFMAC5 is a flexible and highly optimized refinement package that is ideally suited for refinement across the entire resolution spectrum encountered in macromolecular crystallography.

Multivariate phase combination improves automated crystallographic model building.

Density modification is a standard technique in macromolecular crystallography that can significantly improve an initial electron-density map. To obtain optimal results, the initial and density-modified map are combined. Current methods assume that these two maps are independent and propagate the initial map information and its accuracy indirectly through previously determined coefficients. A multivariate equation has been derived that no longer assumes independence between the initial and density-modified map, considers the observed diffraction data directly and refines the errors that can occur in a single-wavelength anomalous diffraction experiment. The equation has been implemented and tested on over 100 real data sets. The results are dramatic: the method provides significantly improved maps over the current state of the art and leads to many more structures being built automatically.

A multivariate likelihood SIRAS function for phasing and model refinement.

A likelihood function based on the multivariate probability distribution of all observed structure-factor amplitudes from a single isomorphous replacement with anomalous scattering experiment has been derived and implemented for use in substructure refinement and phasing as well as macromolecular model refinement. Efficient calculation of a multidimensional integration required for function evaluation has been achieved by approximations based on the function's properties. The use of the function in both phasing and protein model building with iterative refinement was essential for successful automated model building in the test cases presented.

Extending the resolution and phase-quality limits in automated model building with iterative refinement.

Previously, the direct use of prior phase information from a single-wavelength anomalous diffraction (SAD) experiment with a multivariate likelihood function applied to automated model building with iterative refinement has been proposed [Skubák et al. (2004), Acta Cryst. D60, 2196-2201]. In this approach, the anomalous information from the experimental data is used in refinement to derive phase information in a maximum-likelihood formalism and provided a more theoretically valid way of incorporating prior phase information compared with current approaches. In the present work, the SAD multivariate likelihood function that directly uses prior phase information is tested against currently used functions on many different SAD data sets which exhibit a wide range of resolution limits and anomalous signal. The results clearly show the importance of the more theoretically valid utilization of prior phase information: the SAD function extends the resolution and phase-quality limits needed for successful automated model building with iterative refinement. Indeed, the multivariate likelihood function reduces the overfitting in the refinement procedure and performs consistently better than the current refinement targets in terms of the quality of the models obtained and the number of residues built.

Direct incorporation of experimental phase information in model refinement.

The incorporation of prior phase information into a maximum-likelihood formalism has been shown to strengthen model refinement. However, the currently available likelihood refinement target using prior phase information has shortcomings; the 'phased' refinement target considers experimental phase information indirectly and statically in the form of Hendrickson-Lattman coefficients. Furthermore, the current refinement target implicitly assumes that the prior phase information is independent of the calculated model structure factor. This paper describes the derivation of a multivariate likelihood function that overcomes these shortcomings and directly incorporates experimental phase information from a single-wavelength anomalous diffraction (SAD) experiment. This function, which simultaneously refines heavy-atom and model parameters, has been implemented in the refinement program REFMAC5. The SAD function used in conjunction with the automated model-building procedures of ARP/wARP leads to a successful solution when current likelihood functions fail in a test case shown.

Electronic structure of the nickel(II) complex of the Schiff base of (S)-N-(2-benzoylphenyl)-1-benzylprolinamide and glycine.

The experimental charge density of the Ni(II) complex of the Schiff base of (S)-N-(2-benzoylphenyl)-1-benzylprolinamide and glycine was derived from high-resolution single-crystal X-ray diffraction data (lambda = 0.5604 A) at low temperature (100 K) with synchrotron radiation at the beamline F1 using a CCD area detector. The central Ni atom is pseudo-square-planar coordinated by three N atoms [1.9414 (3), 1.8559 (3) and 1.8533 (3) A] and by one O atom [1.8620 (4) A]. The N(1) atom is 0.359 A above the plane defined by the atoms Ni(1), N(2) and N(3). The d-orbital population analysis reveals an oxidation state for the Ni atom of +2 with the configuration d(8) and a hole mainly in the d(x(2)-y(2)) orbital, located in the plane of the four ligating atoms. The prochiral reaction centre was examined by topological analysis.