Demet Kekilli

Fingerprinting redox and ligand states in haemprotein crystal structures using resonance Raman spectroscopy

It is crucial to assign the correct redox and ligand states to crystal structures of proteins with an active redox centre to gain valid functional information and prevent the misinterpretation of structures. Single-crystal spectroscopies, particularly when applied in situ at macromolecular crystallography beamlines, allow spectroscopic investigations of redox and ligand states and the identification of reaction intermediates in protein crystals during the collection of structural data. Single-crystal resonance Raman spectroscopy was carried out in combination with macromolecular crystallography on Swiss Light Source beamline X10SA using cytochrome c from Alcaligenes xylosoxidans. This allowed the fingerprinting and validation of different redox and ligand states, identification of vibrational modes and identification of intermediates together with monitoring of radiation-induced changes. This combined approach provides a powerful tool to obtain complementary data and correctly assign the true oxidation and ligand state(s) in redox-protein crystals.

Photoreduction and validation of haem-ligand intermediate states in protein crystals by in situ single-crystal spectroscopy and diffraction.

Integrated structural biology can yield powerful synergies and maximize the biological information gained. Two examples are described of combining X-ray crystallography with single-crystal resonance Raman and UV–visible spectroscopies to study the functions of haem proteins.

Conformational control of the binding of diatomic gases to cytochrome c

The cytochromes c′ (CYTcp) are found in denitrifying, methanotrophic and photosynthetic bacteria. These proteins are able to form stable adducts with CO and NO but not with O2. The binding of NO to CYTcp currently provides the best structural model for the NO activation mechanism of soluble guanylate cyclase. Ligand binding in CYTcps has been shown to be highly dependent on residues in both the proximal and distal heme pockets. Group 1 CYTcps typically have a phenylalanine residue positioned close to the distal face of heme, while for group 2, this residue is typically leucine. We have structurally, spectroscopically and kinetically characterised the CYTcp from Shewanella frigidimarina (SFCP), a protein that has a distal phenylalanine residue and a lysine in the proximal pocket in place of the more common arginine. Each monomer of the SFCP dimer folds as a 4-alpha-helical bundle in a similar manner to CYTcps previously characterised. SFCP exhibits biphasic binding kinetics for both NO and CO as a result of the high level of steric hindrance from the aromatic side chain of residue Phe 16. The binding of distal ligands is thus controlled by the conformation of the phenylalanine ring. Only a proximal 5-coordinate NO adduct, confirmed by structural data, is observed with no detectable hexacoordinate distal NO adduct.

Active-site protein dynamics and solvent accessibility in native Achromobacter cycloclastes copper nitrite reductase

Multiple structures obtained from one crystal of copper nitrite reductase at elevated cryogenic temperature, together with molecular-dynamics simulations, reveal catalyically important protein and solvent dynamics at the active site.

Hydrogen bonding of the dissociated histidine ligand is not required for formation of a proximal NO adduct in cytochrome c

Cytochromes c’, that occur in methanotrophic, denitrifying and photosynthetic bacteria, form unusual proximal penta-coordinate NO complexes via a hexa-coordinate distal NO intermediate. Their NO binding properties are similar to those of the eukaryotic NO sensor, soluble guanylate cyclase, for which they provide a valuable structural model. Previous studies suggested that hydrogen bonding between the displaced proximal histidine (His120) ligand (following its dissociation from heme due to trans effects from the distally bound NO) and a conserved aspartate residue (Asp121) could play a key role in allowing proximal NO binding to occur. We have characterized three variants of Alcaligenes xylosoxidans cytochrome c’ (AXCP) where Asp121 has been replaced by Ala, Ile and Gln, respectively. In all variants, hydrogen bonding between residue 121 and His120 is abolished yet 5-coordinate proximal NO species are still formed. Our data therefore demonstrate that the His120–Asp121 bond is not essential for proximal NO binding although it likely provides an energy minimum for the displaced His ligand. All variants have altered proximal pocket structure relative to native AXCP.

Distinguishing Nitro vs Nitrito Coordination in Cytochrome c′ Using Vibrational Spectroscopy and Density Functional Theory

Nitrite coordination to heme cofactors is a key step in the anaerobic production of the signaling molecule nitric oxide (NO). An ambidentate ligand, nitrite has the potential to coordinate via the N- (nitro) or O- (nitrito) atoms in a manner that can direct its reactivity. Distinguishing nitro vs nitrito coordination, along with the influence of the surrounding protein, is therefore of particular interest. In this study, we probed Fe(III) heme-nitrite coordination in Alcaligenes xylosoxidans cytochrome c (AXCP), an NO carrier that excludes anions in its native state but that readily binds nitrite (Kd ∼ 0.5 mM) following a distal Leu16 → Gly mutation to remove distal steric constraints. Room-temperature resonance Raman spectra (407 nm excitation) identify ν(Fe-NO2), δ(ONO), and νs(NO2) nitrite ligand vibrations in solution. Illumination with 351 nm UV light results in photoconversion to {FeNO}6 and {FeNO}7 states, enabling FTIR measurements to distinguish νs(NO2) and νas(NO2) vibrations from differential spectra. Density functional theory calculations highlight the connections between heme environment, nitrite coordination mode, and vibrational properties and confirm that nitrite binds to L16G AXCP exclusively through the N atom. Efforts to obtain the nitrite complex crystal structure were hampered by photochemistry in the X-ray beam. Although low dose crystal structures could be modeled with a mixed nitrite (nitro)/H2O distal population, their photosensitivity and partial occupancy underscores the value of the vibrational approach. Overall, this study sheds light on steric determinants of heme-nitrite binding and provides vibrational benchmarks for future studies of heme protein nitrite reactions.

MSOX crystallography and simulations to capture redox enzyme catalysis

Michael Alexander Hough1, Demet Kekilli1, Sam Horrell1, Kakali Sen1, Chin Yong2, Thomas W.K. Keal2, Svetlana V Antonyuk3, Robert R. Eady3, S Samar Hasnain3, Richard W. Strange1 1School Of Biological Sciences, Colchester, United Kingdom, 2Scientific Computing Department, STFC Daresbury Laboratory, Warrington, United Kingdom, 3Institute of Integrative Biology, University of Liverpool, Liverpool, United Kingdom E-mail: mahough@essex.ac.uk

Single Crystal Serial Crystallography to Capture Redox Enzyme Catalysis and Dynamics

Relating individual protein crystal structures to enzyme mechanisms remains a challenging goal for structural biology. The mechanisms of radiation damage to macromolecular crystals have become increasingly well-characterised [1] and attention is paid to minimising the deleterious effects. Alternatively, X-rays may be used to drive enzymes to particular redox states or intermediates. Serial crystallography using multiple crystals has recently been reported in both SR and XFEL experiments [2,3]. I will describe our approach to exploit rapid, shutterless X-ray detector technology on synchrotron MX beamlines to perform low-dose serial crystallography on a single Cu nitrite reductase crystal, allowing 10-50 consecutive X-ray structures at high resolution to be collected, all sampled from the same crystal volume. This serial crystallography approach captures the gradual conversion of the substrate bound at the catalytic type 2 Cu centre, from nitrite to the product, nitric oxide, following reduction of the electron transfer type 1 Cu centre by X-ray generated solvated electrons. Significant, well defined structural rearrangements in the active site are evident in the series as the enzyme moves through its catalytic cycle, which is a vital step in the global denitrification process. We propose that such a serial crystallography approach is widely applicable for studying any redox or electron-driven enzyme reactions in a single protein crystal. It can provide a ‘catalytic reaction movie’ highlighting structural changes that occur during enzyme catalysis. Anticipated developments in the automation of data analysis and modelling are likely to allow seamless and near-real-time analysis of such data on-site at synchrotron crystallographic beamlines. We describe such serial crystallographic experiments conducted at 100K [4] at the elevated cryogenic temperatures of 180-200K, exploiting previously characterised changes in solvent viscosity and dynamics [5]. Finally, we extend our approach to crystals at room temperature, allowing a more complete protein conformational response to active site structural changes to be observed. References [1] E. F. Garman & M. Weik (2015) J. Synchrotron Rad. 22, 195-200. [2] e.g. M. Suga, et al. (2015) Nature 517, 99-103. [3] e.g. C. Gati, G. et. al. (2014) IUCrJ 1, 87-94. [4] M. Weik & J.-P. Colletier (2010) Acta Cryst. Section D. 66, 437-446. [5] S. Horrell, S. V. et al. (2016) IUCrJ (submitted).

Mechanisms of ligand discrimination in cytochromec

The gas-binding heme protein cytochrome c’ discriminates between nitric oxide (NO) and carbon monoxide (CO) while excluding the binding of molecular oxygen. In the absence of gaseous ligands, the heme Fe is 5 coordinate with a proximal histidine ligand and a vacant distal coordination site. CO binds at the distal face to form a 6-coordinate (6c) adduct, while NO forms a stable 5-coordinate (5c) proximal adduct, involving displacement of the proximal histidine, via 6c distal and transient dinitrosyl intermediates. The 6to 5coordinate conversion of NO ligation in cytochrome c’ has recently been confirmed to be highly relevant to the mechanism of activation of soluble guanylate cyclase. Using site directed mutagenesis, biophysical characterisation and correlated crystallography with single crystal resonance Raman spectroscopy we have investigated the mechanisms by which this remarkable example of ligand specificity is controlled and regulated.