Crystal Structure of the Light-Driven Chloride Pump Halorhodopsin from Natronomonas pharaonis

The light-driven chloride pump halorhodopsin from Natronomonas pharaonis (phR) crystallised into the monoclinic space group C2, with a phR trimer per the asymmetric unit. Diffraction data at 2.0-Å resolution showed that the carotenoid bacterioruberin binds to crevices between adjacent protein subunits in the trimeric assembly. Besides seven transmembrane helices (A to G) that characterise archaeal rhodopsins, the phR protomer possesses an amphipathic α-helix (A′) at the N-terminus. This helix, together with a long loop between helices B and C, forms a hydrophobic cap that covers the extracellular surface and prevents a rapid ion exchange between the active centre and the extracellular medium. The retinal bound to Lys256 in helix G takes on an all-trans configuration with the Schiff base being hydrogen-bonded to a water molecule. The Schiff base also interacts with Asp252 and a chloride ion, the latter being fixed by two polar groups (Thr126 and Ser130) in helix C. In the anion uptake pathway, four ionisable residues (Arg123, Glu234, Arg176 and His100) and seven water molecules are aligned to form a long hydrogen-bonding network. Conversely, the cytoplasmic half is filled mostly by hydrophobic residues, forming a large energetic barrier against the transport of anion. The height of this barrier would be lowered substantially if the cytoplasmic half functions as a proton/HCl antiporter. Interestingly, there is a long cavity extending from the main-chain carbonyl of Lys256 to Thr71 in helix B. This cavity, which is commonly seen in halobacterial light-driven proton pumps, is one possible pathway that is utilised for a water-mediated proton transfer from the cytoplasmic medium to the anion, which is relocated to the cytoplasmic channel during the photocycle.     

Retinal-Protein Interactions in Halorhodopsin from Natronomonas pharaonis: Binding and Retinal Thermal Isomerization Catalysis

Halorhodopsin from Natronomonas pharaonis (NpHR) is a member of the retinal protein group and serves as a light-driven chloride pump in which chloride ions are transported through the membrane following light absorption by the retinal chromophore. In this study, we examined two main issues: (1) factors controlling the binding of the retinal chromophore to the NpHR opsin and (2) the ability of the NpHR opsin to catalyze the thermal isomerization of retinal isomers. We have revealed that the reconstitution process of pharaonis HR (NpHR) pigment from its apoprotein and all-trans retinal depends on the pH, and the process has a pK_a of 5.8 ± 0.1. It was proposed that this pK_a is associated with the pK_a of the lysine residue that binds the retinal chromophore (Lys256). The pigment formation is regulated by the concentration of sodium chloride, and the maximum yield was observed at 3.7 M NaCl. The low yield of pigment in a lower concentration of NaCl (< 3 M) may be due to an altered conformation adopted by the apomembrane, which is not capable of forming the pigment. Unexpectedly and unlike the apomembrane of bacteriorhodopsin, NpHR opsin produces pigments with 11-cis retinal and 9-cis retinal owing to the thermal isomerization of these retinal isomers to all-trans retinal. The isomerization rate depends on the pH, and it is faster at a higher pH. The pK_a value of the isomerization process is similar to the pK_a of the binding process of these retinals, which suggests that Lys256 is also involved in the isomerization process. The isomerization is independent of the sodium chloride concentration. However, in the absence of sodium chloride, the apoprotein adopts such a conformation, which does not prevent the isomerization of retinal, but it prevents a covalent bond formation with the lysine residue. The rate and the thermodynamic parameter analysis of the retinal isomerization by NpHR apoprotein led to the conclusion that the apomembrane catalyzes the isomerization via a triplet mechanism.     

Spectral characteristics of the photocycle of channelrhodopsin-2 and its implication for channel function

In 2003, channelrhodopsin-2 (ChR2) from Chlamydomonas reinhardtii was discovered to be a light-gated cation channel, and since that time the channel became an excellent tool to control by light neuronal cells in culture as well as in living animals with high temporal and spatial resolution in a noninvasive manner. However, little is known about the spectral properties and their relation to the channel function. We have expressed ChR2 in the yeast Pichia pastoris and purified the protein. Flash-photolysis data were combined with patch-clamp studies to elucidate the photocycle. The protein absorbs maximally at ∼ 480 nm before light excitation and shows flash-induced absorbance changes with at least two different photointermediates. Four relaxation processes can be extracted from the time course that we have analysed in a linear model for the photocycle leading to the kinetic intermediates P₁ to P₄. A short-lived photointermediate at 400 nm, suggesting a deprotonation of the retinal Schiff base, is followed by a red-shifted (520 nm) species with a millisecond lifetime. The first three kinetic intermediates in the photocycle, P₁ to P₃, are described mainly by the red-shifted 520-nm species. The 400-nm species contributes to a smaller extent to P₁ and P₂. The fourth one, P₄, is spectroscopically almost identical with the ground state and lasts into the seconds time region. We compared the spectroscopic data to current measurements under whole-cell patch-clamp conditions on HEK 293 cells. The lifetimes of the spectroscopically and electrophysiologically determined intermediates are in excellent agreement. The intermediates P₂ and P₃ (absorbing at 520 nm) are identified as the cation permeating states of the channel. Under stationary light, a modulation of the photocurrent by green light (540 nm) was observed. We conclude that the red-shifted spectral species represents the open channel state, and the thermal relaxation of this intermediate, the transition from P₃ to P₄, is coupled to channel closing.     

Functional characterization of sensory rhodopsin II from Halobacterium salinarum expressed in Escherichia coli

Sensory rhodopsin II (SRII) from Halobacterium salinarum is heterologously expressed in Escherichia coli with a yield of 3-4 mg of purified SRII per liter cell culture. UV/Vis absorption spectroscopy display bands characteristic for native SRII. The resonance Raman spectrum provides evidence for a strongly hydrogen-bonded Schiff base like in mammalian rhodopsin but unlike to the homologous pSRII from Natronobacterium pharaonis. Laser flash spectroscopy indicates that SRII in detergent as well as after reconstitution into polar lipids shows its typical photochemical properties with prolonged photocycle kinetics. The first functional heterologous expression of SRII from H. salinarum provides the basis for studies with its cognate transducer HtrII to investigate the molecular processes involved in phototransduction as well as in chemotransduction.     

Crystal structure of the eukaryotic light-driven proton-pumping rhodopsin, Acetabularia rhodopsin II, from marine alga

Acetabularia rhodopsin (AR) is a rhodopsin from the marine plant Acetabularia acetabulum. The opsin-encoding gene from A. acetabulum, ARII, was cloned and found to be novel but homologous to that reported previously. ARII is a light-driven proton pump, as demonstrated by the existence of a photo-induced current through Xenopus oocytes expressing ARII. The photochemical reaction of ARII prepared by cell-free protein synthesis was similar to that of bacteriorhodopsin (BR), except for the lack of light-dark adaptation and the different proton release and uptake sequence. The crystal structure determined at 3.2 Å resolution is the first structure of a eukaryotic member of the microbial rhodopsin family. The structure of ARII is similar to that of BR. From the cytoplasmic side to the extracellular side of the proton transfer pathway in ARII, Asp92, a Schiff base, Asp207, Asp81, Arg78, Glu199, and Ser189 are arranged in positions similar to those of the corresponding residues directly involved in proton transfer by BR. The side-chain carboxyl group of Asp92 appears to interact with the sulfhydryl group of Cys218, which is unique to ARII and corresponds to Leu223 of BR and to Asp217 of Anabaena sensory rhodopsin. The orientation of the Arg78 side chain is opposite to the corresponding Arg82 of BR. The putative absence of water molecules around Glu199 and Arg78 may disrupt the formation of the low-barrier hydrogen bond at Glu199, resulting in the “late proton release”.     

Expression of the halobacterial transducer protein HtrII from Natronomonas pharaonis in Escherichia coli

Archaeal phototaxis is mediated by sensory rhodopsins which form complexes with their cognate transducers. Whereas the receptors sensory rhodopsin I and sensory rhodopsin II (SRII) have been expressed in Escherichia coli (E. coli) only shortened fragments of HtrII from Natronomonas pharaonis (NpHtrII) are available. Here we describe the heterologous expression of full length NpHtrII which was achieved in yields of up to 0.9 mg per litre cell culture. Gel filtration analysis reveals the tendency of the transducer to form dimers and higher-order oligomers which was also observed when complexed to NpSRII. A circular dichroism (CD) spectrum of NpHtrII is comparable to those obtained for the E. coli chemoreceptors indicating a similar folding with predominantly alpha-helical structure. NpHtrII dissociates from the NpSRII/HtrII complex with an apparent K(D) of about 0.6 microM. Photocycle kinetics of the complex is comparable to that obtained for NpSRII in complex with a truncated transducer with slight differences in the M-decay. The data indicate that the heterologously expressed NpHtrII adopt a native like structure, providing the means for elucidating transmembrane signal transduction and activation of microbial signalling cascades.     

Resonance Raman and FTIR spectroscopic characterization of the closed and open states of channelrhodopsin-1

Channelrhodopsin-1 from Chlamydomonas augustae (CaChR1) is a light-activated cation channel, which is a promising optogenetic tool. We show by resonance Raman spectroscopy and retinal extraction followed by high pressure liquid chromatography (HPLC) that the isomeric ratio of all-trans to 13-cis of solubilized channelrhodopsin-1 is with 70:30 identical to channelrhodopsin-2 from Chlamydomonas reinhardtii (CrChR2). Critical frequency shifts in the retinal vibrations are identified in the Raman spectrum upon transition to the open (conductive P₂³⁸⁰) state. Fourier transform infrared spectroscopy (FTIR) spectra indicate different structures of the open states in the two channelrhodopsins as reflected by the amide I bands and the protonation pattern of acidic amino acids.     

Role of Arg-72 of pharaonis Phoborhodopsin (sensory rhodopsin II) on its photochemistry

Pharaonis phoborhodopsin (ppR, or pharaonis sensory rhodopsin II, NpsRII) is a sensor for the negative phototaxis of Natronomonas (Natronobacterium) pharaonis. Arginine 72 of ppR corresponds to Arg-82 of bacteriorhodopsin, which is a highly conserved residue among microbial rhodopsins. Using various Arg-72 ppR mutants, we obtained the following results: 1). Arg-72(ppR) together possibly with Asp-193 influenced the pK(a) of the counterion of the protonated Schiff base. 2). The M-rise became approximately four times faster than the wild-type. 3). Illumination causes proton uptake and release, and the pH profiles of the sequence of these two proton movements were different between R72A mutant and the wild-type; it is inferred that Arg-72 connects the proton transfer events occurring at both the Schiff base and an extracellular proton-releasing residue (Asp-193). 4). The M-decays of Arg-72 mutants were faster ( approximately 8-27 folds at pH 8 depending on mutants) than the wild-type, implying that the guanidinium prevents the proton transfer from the extracellular space to the deprotonated Schiff base. 5), The proton-pumping activities were decreased for mutants having increased M-decay rates, but the extent of the decrease was smaller than expected. The role of Arg-72 of ppR on the photochemistry was discussed.     

Crystal Structure of HypA, a Nickel-Binding Metallochaperone for [NiFe] Hydrogenase Maturation

HypA is one of the auxiliary proteins involved in the maturation of [NiFe] hydrogenases. By an unknown mechanism, HypA functions as a metallochaperone in the insertion of the Ni atom into hydrogenases. We have determined the crystal structures of HypA from Thermococcus kodakaraensis KOD1 in both monomeric and dimeric states. The structure of the HypA monomer consists of Ni- and Zn-binding domains. The relative arrangement of the two metal-binding domains has been shown to be associated with local conformations of the conserved Ni-binding motif, suggesting a communication between the Ni- and Zn-binding sites. The HypA dimer has been shown to be stabilized by unexpected domain swapping through archaea-specific linker helices. In addition, the hexameric structure of HypA is formed in the crystal packing. Several hydrogen bonds and hydrophobic interactions stabilize the hexamer interface. These findings suggest the functional diversity of HypA proteins.     

Predicted bacteriorhodopsin from Exiguobacterium sibiricum is a functional proton pump

The predicted Exigobacterium sibiricum bacterirhodopsin gene was amplified from an ancient Siberian permafrost sample. The protein bacteriorhodopsin from Exiguobacterium sibiricum (ESR) encoded by this gene was expressed in Escherichia coli membrane. ESR bound all-trans-retinal and displayed an absorbance maximum at 534 nm without dark adaptation. The ESR photocycle is characterized by fast formation of an M intermediate and the presence of a significant amount of an O intermediate. Proteoliposomes with ESR incorporated transport protons in an outward direction leading to medium acidification. Proton uptake at the cytoplasmic surface of these organelles precedes proton release and coincides with M decay/O rise of the ESR.     

Crystal structures of mammalian glutamine synthetases illustrate substrate-induced conformational changes and provide opportunities for drug and herbicide design

Glutamine synthetase (GS) catalyzes the ligation of glutamate and ammonia to form glutamine, with concomitant hydrolysis of ATP. In mammals, the activity eliminates cytotoxic ammonia, at the same time converting neurotoxic glutamate to harmless glutamine; there are a number of links between changes in GS activity and neurodegenerative disorders, such as Alzheimer's disease. In plants, because of its importance in the assimilation and re-assimilation of ammonia, the enzyme is a target of some herbicides. GS is also a central component of bacterial nitrogen metabolism and a potential drug target. Previous studies had investigated the structures of bacterial and plant GSs. In the present publication, we report the first structures of mammalian GSs. The apo form of the canine enzyme was solved by molecular replacement and refined at a resolution of 3 Å. Two structures of human glutamine synthetase represent complexes with: a) phosphate, ADP, and manganese, and b) a phosphorylated form of the inhibitor methionine sulfoximine, ADP and manganese; these structures were refined to resolutions of 2.05 Å and 2.6 Å, respectively. Loop movements near the active site generate more closed forms of the eukaryotic enzymes when substrates are bound; the largest changes are associated with the binding of the nucleotide. Comparisons with earlier structures provide a basis for the design of drugs that are specifically directed at either human or bacterial enzymes. The site of binding the amino acid substrate is highly conserved in bacterial and eukaryotic GSs, whereas the nucleotide binding site varies to a much larger degree. Thus, the latter site offers the best target for specific drug design. Differences between mammalian and plant enzymes are much more subtle, suggesting that herbicides targeting GS must be designed with caution.     

Crystal structure of the stationary phase survival protein SurE with metal ion and AMP

The stationary phase survival protein SurE is a metal ion-dependent phosphatase distributed among eubacteria, archaea, and eukaryotes. In Escherichia coli, SurE has activities as nucleotidase and exopolyphosphatase, and is thought to be involved in stress response. However, its physiological role and reaction mechanism are unclear. We report here the crystal structures of the tetramer of SurE from Thermus thermophilus HB8 (TtSurE) both alone and crystallized with Mn(2+) and substrate AMP. In the presence of Mn(2+) and AMP, differences between the protomers were observed in the active site and in the loop located near the active site; AMP-bound active sites with the loops in a novel open conformation were found in the two protomers, and AMP-free active sites with the loops in a conventional closed conformation were found in the other two protomers. The two loops in the open conformation are entwined with each other, and this entwining is suggested to be required for enzymatic activity by site-directed mutagenesis. TtSurE exists as an equilibrium mixture of dimer and tetramer in solution. The loop-entwined structure indicates that SurE acts as a tetramer. The structural features and the absence of negative cooperativity imply the half-of-the-sites reactivity mechanism resulting from a pre-existing tendency toward structural asymmetry.     

Crystal Structures of Wild-type and Mutant Methicillin-resistant Staphylococcus aureus Dihydrofolate Reductase Reveal an Alternate Conformation of NADPH That May Be Linked to Trimethoprim Resistance

Both hospital- and community-acquired Staphylococcus aureus infections have become major health concerns in terms of morbidity, suffering and cost. Trimethoprim-sulfamethoxazole (TMP-SMZ) is an alternative treatment for methicillin-resistant S. aureus (MRSA) infections. However, TMP-resistant strains have arisen with point mutations in dihydrofolate reductase (DHFR), the target for TMP. A single point mutation, F98Y, has been shown biochemically to confer the majority of this resistance to TMP. Using a structure-based approach, we have designed a series of novel propargyl-linked DHFR inhibitors that are active against several trimethoprim-resistant enzymes. We screened this series against wild-type and mutant (F98Y) S. aureus DHFR and found that several are active against both enzymes and specifically that the meta-biphenyl class of these inhibitors is the most potent. In order to understand the structural basis of this potency, we determined eight high-resolution crystal structures: four each of the wild-type and mutant DHFR enzymes bound to various propargyl-linked DHFR inhibitors. In addition to explaining the structure-activity relationships, several of the structures reveal a novel conformation for the cofactor, NADPH. In this new conformation that is predominantly associated with the mutant enzyme, the nicotinamide ring is displaced from its conserved location and three water molecules complete a network of hydrogen bonds between the nicotinamide ring and the protein. In this new position, NADPH has reduced interactions with the inhibitor. An equilibrium between the two conformations of NADPH, implied by their occupancies in the eight crystal structures, is influenced both by the ligand and the F98Y mutation. The mutation induced equilibrium between two NADPH-binding conformations may contribute to decrease TMP binding and thus may be responsible for TMP resistance.     

Crystal structures of the luciferase and green fluorescent protein from Renilla reniformis

Due to its ability to emit light, the luciferase from Renilla reniformis (RLuc) is widely employed in molecular biology as a reporter gene in cell culture experiments and small animal imaging. To accomplish this bioluminescence, the 37-kDa enzyme catalyzes the degradation of its substrate coelenterazine in the presence of molecular oxygen, resulting in the product coelenteramide, carbon dioxide, and the desired photon of light. We successfully crystallized a stabilized variant of this important protein (RLuc8) and herein present the first structures for any coelenterazine-using luciferase. These structures are based on high-resolution data measured to 1.4 Å and demonstrate a classic α/β-hydrolase fold. We also present data of a coelenteramide-bound luciferase and reason that this structure represents a secondary conformational form following shift of the product out of the primary active site. During the course of this work, the structure of the luciferase's accessory green fluorescent protein (RrGFP) was also determined and shown to be highly similar to that of Aequorea victoria GFP.     

Snapshots of RecA protein involving movement of the C-domain and different conformations of the DNA-binding loops: crystallographic and comparative analysis of 11 structures of Mycobacterium smegmatis RecA

Mycobacterium smegmatis RecA and its nucleotide complexes crystallize in three different, but closely related, forms characterized by specific ranges of unit cell dimensions. The six crystals reported here and five reported earlier, all grown under the same or very similar conditions, belong to these three forms, all in space group P6(1). They include one obtained by reducing relative humidity around the crystal. In all crystals, RecA monomers form filaments around a 6(1) screw axis. Thus, the c-dimension of the crystal corresponds to the pitch of the RecA filament. As reported for Escherichia coli RecA, the variation in the pitch among the three forms correlates well with the motion of the C-terminal domain of the RecA monomers with respect to the main domain. The domain motion is compatible with formation of inactive as well as active RecA filaments involving monomers with a fully ordered C domain. It does not appear to influence the movement upon nucleotide-binding of the switch residue, which is believed to provide the trigger for transmitting the effect of nucleotide binding to the DNA-binding region. Interestingly, partial dehydration of the crystal results in the movement of the residue similar to that caused by nucleotide binding. The ordering of the DNA-binding loops, which present ensembles of conformations, is also unaffected by domain motion. The conformation of loop L2 appears to depend upon nucleotide binding, presumably on account of the movement of the switch residue that forms part of the loop. The conformations of loops L1 and L2 are correlated and have implications for intermolecular communications within the RecA filament. The structures resulting from different orientations of the C domain and different conformations of the DNA-binding loops appear to represent snapshots of the RecA at different phases of activity, and provide insights into the mechanism of action of RecA.     

Association of pharaonis phoborhodopsin with its cognate transducer decreases the photo-dependent reactivity by water-soluble reagents of azide and hydroxylamine

pharaonis phoborhodopsin (ppR; also pharaonis sensory rhodopsin II, psRII) is a receptor of the negative phototaxis of Natronobacterium pharaonis. In halobacterial membrane, ppR forms a complex with its transducer pHtrII, and this complex transmits the light signal to the sensory system in the cytoplasm. In the present work, the truncated transducer, t-Htr, was used which interacts with ppR [Sudo et al. (2001) Photochem. Photobiol. 74, 489-494]. Two water-soluble reagents, hydroxylamine and azide, reacted both with the transducer-free ppR and with the complex ppR/t-Htr (the complex between ppR and its truncated transducer). In the dark, the bleaching rates caused by hydroxylamine were not significantly changed between transducer-free ppR and ppR/t-Htr, or that of the free ppR was a little slower. Illumination accelerated the bleach rates, which is consistent with our previous conclusion that the reaction occurs selectively at the M-intermediate, but the rate of the complex was about 7.4-fold slower than that of the transducer-free ppR. Azide accelerated the M-decay, and its reaction rate of ppR/t-Htr was about 4.6-fold slower than free ppR. These findings suggest that the transducer binding decreases the water accessibility around the chromophore at the M-intermediate. Its implication is discussed.     

First structures of an active bacterial tyrosinase reveal copper plasticity

Tyrosinase is a member of the type 3 copper enzyme family that is involved in the production of melanin in a wide range of organisms. The crystal structures of a tyrosinase from Bacillus megaterium were determined at a resolution of 2.0-2.3 Å. The enzyme crystallized as a dimer in the asymmetric unit and was shown to be active in crystal. The overall monomeric structure is similar to that of the monomer of the previously determined tyrosinase from Streptomyces castaneoglobisporus, but it does not contain an accessory Cu-binding “caddie” protein. Two Cu(II) ions, serving as the major cofactors within the active site, are coordinated by six conserved histidine residues. However, determination of structures under different conditions shows varying occupancies and positions of the copper ions. This apparent mobility in copper binding modes indicates that there is a pathway by which copper is accumulated or lost by the enzyme. Additionally, we suggest that residues R209 and V218, situated in a second shell of residues surrounding the active site, play a role in substrate binding orientation based on their flexibility and position. The determination of a structure with the inhibitor kojic acid, the first tyrosinase structure with a bound ligand, revealed additional residues involved in the positioning of substrates in the active site. Comparison of wild-type structures with the structure of the site-specific variant R209H, which possesses a higher monophenolase/diphenolase activity ratio, lends further support to a previously suggested mechanism by which monophenolic substrates dock mainly to CuA.     

Crystal structure of the Clostridium limosum C3 exoenzyme

C3-like toxins ADP-ribosylate and inactivate Rho GTPases. Seven C3-like ADP-ribosyltransferases produced by Clostridium botulinum, Clostridium limosum, Bacillus cereus and Staphylococcus aureus were identified and two representatives - C3bot from C. botulinum and C3stau2 from S. aureus - were crystallized. Here we present the 1.8 Å structure of C. limosum C3 transferase C3lim and compare it to the structures of other family members. In contrast to the structure of apo-C3bot, the canonical ADP-ribosylating turn turn motif is observed in a primed conformation, ready for NAD binding. This suggests an impact on the binding mode of NAD and on the transferase reaction. The crystal structure explains why auto-ADP-ribosylation of C3lim at Arg41 interferes with the ADP-ribosyltransferase activity of the toxin.     

The role of Glu196 in the environment around the substrate binding site of leucine aminopeptidase from Streptomyces griseus

To investigate the role of Glu196 of leucine aminopeptidase from Streptomyces griseus (SGAP) in SGAP activation by calcium and substrate specificity, we constructed E196X SGAP by saturation mutagenesis. Most mutations led to the abrogation of SGAP activation by calcium, and substitution with Lys led to a marked increase in activity toward Asp-p-nitroanilide (pNA) and a decrease in that toward Lys-pNA. A similar result was obtained from the investigation using non-calcium-activated enzyme from Streptomyces septatus (SSAP). These results indicate that Glu196 of SGAP is associated with the environment around the substrate binding site besides its role in SGAP activation by calcium.     

Structural Basis of Heme Binding in the Cu,Zn Superoxide Dismutase from Haemophilus ducreyi

The Cu,Zn superoxide dismutase from Haemophilus ducreyi is characterized by the unique ability to bind heme at its dimer interface. Here we report the high-resolution crystal structures of this protein in the heme-loaded (holo) and heme-free (apo) forms. Heme is asymmetrically bound between the two enzyme subunits, where heme iron is coordinated by two histidine residues, His64 and His 124, provided by the two subunits. Moreover, the binding of heme to the protein is ensured by stabilizing contacts between the prosthetic group and a limited number of other residues, most of which are not present in other bacterial enzyme variants. We show that the introduction of only three mutations at the dimer interface of the enzyme from Haemophilus parainfluenzae, a closely related bacterial species, is sufficient to induce heme-binding ability by this enzyme variant. Heme binding does not alter protein activity. Moreover, the binding of the prosthetic group does not induce any significant structural perturbation at the subunit level and requires only limited local structural rearrangements that widen the cleft at the dimer interface and cause a limited shift in the relative orientation between the subunits. The presence of a preformed heme-binding pocket and the significant solvent exposure of the cofactor to the solvent are compatible with the suggested protective role of the enzyme against heme toxicity or with its involvement in heme trafficking in the periplasmic space.