Structural characterization of the protein cce_0567 from Cyanothece 51142, a metalloprotein associated with nitrogen fixation in the DUF683 family

Abstract

The genomes of many cyanobacteria contain the sequence for a small protein with a common “Domain of Unknown Function” grouped into the DUF683 protein family. While the biological function of DUF683 is still not known, their genomic location within nitrogen fixation clusters suggests that DUF683 proteins may play a role in the process. The diurnal cyanobacterium Cyanothece sp. PCC 51142 contains a gene for a protein that falls into the DUF683 family, cce_0567 (78 aa, 9.0 kDa). In an effort to elucidate the biochemical role DUF683 proteins may play in nitrogen fixation, we have determined the first crystal structure for a protein in this family, cce_0567, to 1.84 Å resolution. Cce_0567 crystallized in space group P2₁ with two protein molecules and one Ni²⁺ cation per asymmetric unit. The protein is composed of two α-helices, residues P11 to G41 (α1) and L49–E74 (α2), with the second α-helix containing a short 3₁₀-helix (Y46–N48). A four-residue linker (L42–D45) between the helices allows them to form an anti-parallel bundle and cross over each other towards their termini. In solution it is likely that two molecules of cce_0567 form a rod-like dimer by the stacking interactions of ~1/2 of the protein. Histidine-36 is highly conserved in all known DUF683 proteins and the N2 nitrogen of the H36 side chain of each molecule in the dimer is coordinated with Ni²⁺ in the crystal structure. The divalent cation Ni²⁺ was titrated into ¹⁵N-labeled cce_0567 and chemical shift perturbations were observed only in the ¹H–¹⁵N HSQC spectra for residues at, or near, the site of Ni²⁺ binding observed in the crystal structure. There was no evidence for an increase in the size of cce_0567 upon binding Ni²⁺, even in large molar excess of Ni²⁺, indicating that a metal was not required for dimer formation. Circular dichroism spectroscopy indicated that cce_0567 was extremely robust, with a melting temperature of ~62 °C that was reversible.

1. Introduction

Cyanobacteria, the progenitors of chloroplasts in plants and algae [1, 2], radically altered the chemistry of our planet starting about 2.3 billion years ago through the release of a photosynthesis byproduct, oxygen [3, 4]. Some cyanobacteria can also still fix nitrogen [5] with nitrogenase, a multimeric enzyme complex [6]. One of the byproducts of the process, hydrogen, is currently of much interest as a clean source of energy [7]. However, nitrogen fixation is biochemically intolerant of the presence of oxygen because nitrogenase is irreversibly inhibited by molecular oxygen [8–10]. To perform the two mutually exclusive functions, photosynthesis and nitrogen fixation, some cyanobacteria have temporally separated them into daytime and nighttime activities, respectively. Indeed, cyanobacteria are the simplest known organisms that display circadian rhythms [11–13]. To better understand the details of how cyanobacteria control the diurnal processes, the genome of the marine diazotrophic cyanobacterium, Cyanothece ATCC 51142, was recently sequenced and deposited into the GenBank database (accession nos. CP000806–CP000811) [14]. Present is a gene, cce_0567, that falls into a family of proteins with a similar “Domain of Unknown Function”, DUF683, that is conserved among various prokaryotic species. Excluding Cyanothece, 54 DUF683 proteins are currently found in GenBank with similarity searches and the conserved domain is also listed 39 times in the Pfam database (PF05082) [15]. Genes that fall into the DUF683 family are typically under 100 amino acid residues in size, are always observed in the nitrogen fixation gene cluster of the genome, and appear to be most prevalent in cyanobacteria. In Cyanothece, the cce_0567 gene is observed between NifW, a gene that codes for a protein associated with nitrogen fixation whose precise function is uncertain, and cce_0566, a gene that falls into the DUF269 family of genes found exclusively in nitrogen fixation operons. The appearance of the cce_0567 gene in nitrogen fixation operons suggests that it may play a role in the process, a role that is still undefined. To obtain clues as to the biochemical function of proteins in the DUF683 family of conserved proteins [16, 17] we have determined the first crystal structure of a protein in the DUF683 family, cce_0567, to a resolution of 1.8 Å. Because the solved protein structure contained a nickel cation, the metal binding properties of cce_0567 towards the divalent cations Ni²⁺, Zn²⁺, and Co²⁺ were assessed by NMR spectroscopy using previously determined backbone NMR assignments [18]. The thermal stability of cce_0567 was assessed by circular dichroism spectroscopy and verified by NMR spectroscopy. On the basis of our crystal structure for cce_0567, a recently determined NMR structure of another member of the DUF683 protein family (Q60C73_METCA from Methylococcus capsulatus (2JS5)), and sequence homology of all proteins in the DUF683 family, the biochemically significant regions and potential biological role of proteins in the DUF683 family are speculated upon.

2. Materials and methods

2.1. Cloning, expression, and purification

The cloning, expression, and purification protocol for ¹⁵N-labeled DUF683 has previously been reported [18]. The procedure for preparing selenomethionine-substituted DUF683 was similar using a protocol that inhibited the methionine biosynthesis pathway [19,20]. Basically, after growing the cells at 310 K to mid-log phase (OD_600nm~0.8) in M9 minimal medium supplemented with 34 µg/mL kanamycin (RPI Corporation, Prospect, IL),120 µg/mL MgSO₄, 11 µg/mL CaCl₂, 10 ng/mL Fe₂Cl₃, 50 µg/mL NaCl, and 4 mg/mL glucose, the temperature was lowered to 298 K. At this point lysine (0.1 mg/mL), phenylalanine (0.1 mg/mL), threonine (0.1 mg/mL), isoleucine (0.05 mg/mL), valine (0.05 mg/mL), and selenomethionine (SeMet) (Acrös Organics, Geel, Belgium) (0.06 mg/mL) were added followed by the induction of protein expression ~15 min later with isopropyl β-d-1-thiogalactopyranoside (0.026 mg/mL). Approximately 6 h later the cells were harvested by mild centrifugation and then frozen at 193 K. From this point forward the protocol was identical to the previously described method with the protein exchanged into the buffer (500 mM NaCl, 20 mM Tris, 1.0 mM dithiothreitol, pH 7.2) used for the crystallization, NMR, and circular dichroism experiments. Although NMR spectroscopy and other evidence indicated that cce_0567 was primarily in the metal-free form at this stage (>90%), protein was treated with 2 mM EDTA and then dialyzed extensively to remove the EDTA for metal titration studies.

2.2. Nuclear magnetic resonance spectroscopy

One microliter aliquots of divalent metal cation were titrated into a 250 µL NMR sample of ¹⁵N-labeled cce_0567 (~0.4 mM) using 0.1 M stock solution of NiCl₂, CoCl₂, and ZnCl₂ (Hampton Reseach, Aliso Viejo, CA). Following mild agitation, two-dimensional ¹H–¹⁵N HSQC spectra were collected on a Varian Inova-750 spectrometer equipped with a triple resonance probe and pulse field gradients. Spectra were recorded after every addition at a metal:protein molar ratio of 0.5, 1.0, 2.0, and 4.0 to 1.0. To test the efficiency of EDTA in removing the metal from the protein, a solution of cce_0567 and Ni²⁺ (2:1 metal:protein molar ratio) was treated with excess EDTA and an ¹H–¹⁵N HSQC spectrum collected. To assay the effect of heating, a 0.4 mM sample of ¹⁵N-labeled cce_0567 (metal-free) was heated to 80 °C for 30 min and an ¹H–¹⁵N HSQC spectrum collected after cooling back to room temperature. All NMR data was processed using Felix2007 (Felix NMR, Inc., San Diego, CA) software with ¹H and ¹⁵N chemical shifts referenced to DSS (DSS = 0 ppm) using indirect methods [21]. The assigned ¹H, ¹³C, and ¹⁵N chemical shifts for cce_0567 have been deposited in the BioMagResBank database (www.bmrb.wisc.edu) under the accession number BMRB-15539 [18].

2.3. Circular dichroism spectroscopy

Circular dichroism data were obtained on an Aviv Model 62DS spectropolarimeter calibrated with an aqueous solution of ammonium d-(+)camphorsulfonate. Measurements on cce_0567were made in the same buffer used for the NMR studies (500 mM NaCl, 20 mM TrisHCl, 1 mM DTT, pH 7.2). A thermal denaturation curve for cce_0567 was obtained on a 30 µM solution in a quartz cell of 0.1 cm path length by recorded the ellipticity at 220 nm in 2.5 °C intervals from 10 to 80 °C. Using the same sample, a far-UV wavelength spectrum between 200 and 250 nm was recorded at 25, 80, and again at 25 °C. Spectra were the result of averaging three consecutive scans with a bandwidth of 1.0 nm and a time constant of 1.0 s. The wavelength spectra were processed by subtracting a blank spectrum, correcting the base-line, and then noise reduction.

2.4. Crystallization, structure determination and refinement

Vapor-diffusion crystallization trials using hanging drops were set up on SeMet-labeled cce_0567 at room temperature (~295 K) using screens from Hampton Research (Aliso Viejo, CA). Crystals began appearing 24–48 h later under one condition that was then optimized by adjusting the protein concentration. Crystals were harvested 3–4 days after mixing 2 µL of protein (~2 mg/mL) with 2 µL of reservoir buffer containing 30% (w/v) PEG 1500. Protein crystals of similar morphology were also grown under the same conditions with precipitant containing 1 mM NiCl₂ (Hampton Research). Crystals were directly mounted in nylon CryoLoops (Hampton Research), flash-frozen in liquid nitrogen, stored under liquid nitrogen, and shipped to the National Synchrotron Light Source (NSLS) at Brookhaven National Laboratory for X-ray data collection.

2.5. X-ray data collection and processing

All X-ray diffraction data were collected at the X29A beamline with an ADSC Q315 CCD detector. Data was initially collected on crystals grown in the absence of 1.0 mM NiCl₂ in the precipitant solution which contained 30% (w/v) PEG 1500 that served well as a cryoprotectant when cryocooling the crystals in a nitrogen stream (100 K) during data collection. A three-wavelength SAD data set was collected on these first crystals to a resolution of 2.6 Å, however, it was not possible to generate an interpretable electron density map from the data suggesting that perhaps there was too much disorder in the crystal lattice. In an attempt to remove this disorder, crystallization screens were repeated in the same precipitant in the presence of 12 different metal additives (Hampton Research, Aliso Viejo, CA). In the presence of these additives, crystals were only obtained in the presence of 1 mM NiCl₂ and they indexed to the same space group (P2₁) as the original crystals. Such crystals diffracted to a resolution of 1.84 Å and yielded excellent electron density maps. X-ray fluorescence spectroscopy on the single crystals grown in the presence and absence of NiCl₂ conclusively identified the presence of nickel in the latter, but not in the former, crystals.

A three-wavelength MAD data set at 1.84 Å resolution was obtained on the crystals grown in the presence of NiCl₂. The images were integrated and scaled with HKL2000 [22]. Selenium and nickel heavy atom sites were determined using the SHELX program suite [23–25] and HKL2MAP [26]. The peak wavelength data of the MAD data set, which contained useful anomalous signal out to 1.84 Å resolution, was used together with the program SOLVE [27] to produce an electron density map at 1.84 Å resolution. The majority of the structure was built automatically into the resolution map using RESOLVE [27–29] with the remainder built manually using coot [30] and refined with refmac5 [31] to 2.3 Å resolution. A final check on the stereochemical quality of the final model was assessed using the program MolProbity [32] and PROCHECK [16] and any conflicts addressed. MolProbity analysis indicated that the overall geometry of the final model ranked in the 45th percentile (MolProbity score of 2.22) where the 100th percentile is best among structures of comparable resolution. The clash score for “all-atoms” was 15.12 corresponding to a 45th percentile ranking for structures of comparable resolution. PROCHECK analysis showed that 98% of the psi/phi pairs were in most favored region and the remainder in additionally allowed regions. Overall, MolProbity and PROCHECK assessment indicated that the final model was a quality representation of the structure of cce_0567. The data collection and structure refinement statistics are given in Table 1 and the coordinates have been submitted to the Protein Data Bank (PDB ID 3CSX).

Table 1.

Summary of the diffraction data collection and refinement statistics for cce_0567¹

X-ray source	X29
Detector	ADSC Q315 CCD
X-ray wavelength (Å)	0.9791
Temperature (K)	100
Data set	SAD
Space group	P21 (P 1 21 1)
Unit-cell parameters
a (Å)	29.44
b (Å)	38.38
c (Å)	62.64
α = γ	90°
β	96.70°
Resolution range (Å)	30–1.84 (1.92–1.84)
Molecules per AU	2
Total number of reflections	269,646
Total number of unique reflections	11,230 (625)
Completeness (%)	92 (52)
Rmerge (%)a	6.9 (26.1)
Mean I/σ(I)	36.9 (2.6)
Wilson B (Å2)	25.12
Refinement
R (%)	21.6
Rfree (%)	29.4
Mean B-value (Å2)	28.14
Δbonds (Å)	0.015
Δangles (°)	1.50
MolProbity Ramachandran analysis
Most favored	98%
Additionally allowed	2%
MolProbity
Clash score, all atoms	15.1 (46th)
MolProbity score	2.22 (45th)

¹Values in parenthesis are the statistics for the highest resolution shell (1.92–1.84 Å).

^aR_merge = ∑_h∑_j|I_h,j−<I_h>|/∑_h∑_j, where I_h,j is the intensity of the jth observation of unique reflection h.

3. Results and discussion

3.1. Crystal structure of cce_0567

The asymmetric unit of the crystal contains two molecules of cce_0567 plus one nickel ion. Interpretable electron density was absent at the termini of both molecules suggesting the ends of the protein were disordered. Ten fewer residues could be mapped in molecule A (A15–E75)—then in molecule B (T5–E75), and therefore, the latter molecule will be discussed when referring to a single cce_0567 molecule. Aside from more missing electron density in one molecule, the structures of both molecules in the asymmetric unit are similar with a backbone RSMD of 0.96 Å and an all atoms RMSD of 1.79 Å (A15–E75) as determined using Superpose [33]. Indeed, the structure of both molecules are similar enough that previous NMR data, while suggesting cce_0567 was oligomeric in solution, generated only one set of amide cross peaks in the ¹H–¹⁵N HSQC spectrum indicating that each subunit was in a magnetically similar environment [18]. Each individual cce_0567 molecule is composed of two α-helices, residues P11 to G41 (α1) and L49–E74 (α2), with the second α-helix containing a short 3₁₀-helix (Y46–N48). Four residues (L42–D45) connect the two helices and allow them to form an anti-parallel bundle. Fig. 1A shows that the two-helix bundle adopts an “X”-type shape as they cross over towards their termini.

Fig. 1.

(A) Two orientations of the likely cce_0567 dimer in solution that differ by ~90° about the horizontal axis. Each subunit is colored separately and the side chains of the H36 residue from each protein and the nickel (orange) highlighted. (B) The solvent accessible electrostatic surface for the cce_0567 dimer shown in two orientations. The upper orientation is similar to the orientation shown on the top of (A) with the lower orientation differing by ~180° about the horizontal axis.

Previous NMR data suggested that cce_0567 was oligomeric [18]. However, in the asymmetric unit of the crystal there clearly was little intermolecular contact between the two molecules as reflected in a calculated buried surface area of only 860 Ų between them. A more likely interaction between two molecules of cce_0567 in solution is as shown in Fig. 1A. Such a dimer was observed upon expanding the symmetry related molecules of the asymmetric unit. The buried surface area due to the formation of such a dimer is 1860 Ų, a value larger than the value observed due to dimer formation in the asymmetric unit. Strong evidence in support of such a dimer is the recently determined solution structure for another protein that falls into the DUF683 family, Q60C73_METCA, from the methane-oxidizing bacterium M. capsulatus (PDB ID 2JS5). The crystal structure of the cce_0567 dimer shown in Fig. 1A (chain B, T10–A75) superimposes over the solution structure of the Q60C73_METCA dimer (model 1, chain B, M1–H66) with a backbone RMSD of 1.88 Å and an all-atoms RMSD of 2.32 Å (as determined using Superpose).

Fig. 1B shows an estimation of the electrostatic surface potential of the protein. There appears to be defined regions of alternating positive and negative charge. The biological importance of the electrostatic surface potential is unknown. However, sequence alignment of all 48 currently identified proteins in the DUF683 family using ClustalW2 [34], shown in Fig. 2, indicates that the charged residues are highly conserved at least in the N-terminal α-helix (α1). Such conservation of charged residues suggests that they may have a biological property.

Fig. 2.

ClustalW2 sequence alignment and color coding of cce_0567, Q60C73_METCA, and 46 other proteins listed by Pmaf in the DUF683 protein family (PF05082). Red = small and hydrodrophobic but not aromatic (A, V, F, P, M, I, L, W), blue = acidic (D,E); magenta = basic (R, K); green = hydrophilic (S, T, Y, H, C, N, G, Q).

3.2. Metal binding site

Crystals of SeMet-labeled cce_0567 that diffracted to 2.6 Å were original grown in the absence of divalent cation additives. However, it was not possible to generate an interpretable electron density map due to excessive disorder in the crystals. On the other hand, crystals of SeMet-labeled cce_0567 grown in the presence of NiCl₂ diffracted to 1.84 Å resolution and these crystals generated excellent electron density maps. Nickel was unambiguously identified in the crystals grown in the presence of NiCl₂ by X-ray fluorescence spectroscopy (data not shown) and the electron density observed for the metal ion in the electron density maps corroborated well for a Ni²⁺ ion. There was no evidence for nickel by X-ray fluorescence spectroscopy in the original crystals grown in the absence of NiCl₂. Evidently, the nickel cation was necessary to provide additional order to the crystal lattice.

Fig. 3 illustrates the electron density at 1.5 sigma about the metal center and highlights the metal coordination to the protein side chains and water. The nickel is ligated to the N2 nitrogen atom of H36 from two cce_0567 molecules with a typical, average Ni–N bond length of 2.09±0.02 Å [35]. In addition to coordinating two nitrogen atoms, four additional “bulges” of electron density are observed around the nickel to form an overall pseudooctahedral geometry [35]. Two of the “bulges” are approximately in the plane of the two Ni–N bonds with the other two above and below this plane. Nickel (II) can exist in either the paramagnetic high-spin octahedral electronic configuration or the diamagnetic low-spin form. The latter form can adopt either a five-coordinate trigonal bipyramidal geometry or a four-coordinate square-planar geometry [36]. Consequently, the nickel appears to be in the paramagnetic high-spin state in the cce_0567 crystal structure, a configuration that differs from its low-spin state in solution as determined by NMR spectroscopy (to be discussed).

Fig. 3.

The electron density surrounding the Ni²⁺ ion in the crystal structure of cce_0567 illustrated at 1.5 σ. The N2 ring nitrogen from the histidine-36 side chain of two molecules coordinate the Ni²⁺ with water occupying the other sites around the metal. In the diagram on the right, H36′ has been rotated ~90° about the z-axis into the plane of the paper.

While nickel was necessary to solve the crystal structure of cce_0567 and is held in place by a histidine residue from each of the two molecules in the dimer, NMR studies indicate that nickel coordination is not necessary for dimer formation. Fig. 4 is an ¹H–¹⁵N HSQC spectrum of cce_0567 obtained in the absence and presence of nickel. As illustrated, only a small subset of resonances in the ¹H–¹⁵N HSQC spectrum are perturbed in the presence of nickel and these are at, or near, the nickel binding site or the turn between the two helices. Nickel associates with cce_0567 relatively tightly with a K_d of 10⁻⁶ M or less. Such a value is inferred from the observation of two subsets of cross peaks for residues about the metal-binding site in the presence of less than a 1:1 protein-to-metal molar ratios (data not shown) indicating that the equilibrium between the metal-free and metal-bound states are in slow exchange [37]. If nickel binding was responsible for dimer formation one would expect a larger subset of resonances to be perturbed in the ¹H–¹⁵N HSQC spectrum with the addition of nickel to the solution. However, instead of chemical shift perturbations to residues through-out the dimer interface, colored black for one cce_0567 monomer in the inset of Fig. 4, perturbations are only observed for residues in a small part of α1, near the nickel binding site, and part of the linker region (colored red in the Fig. 4 inset). Furthermore, dimer formation would dictate a doubling of the protein's molecular weight. Such an event would manifest an increase in the visible line shape of the cross peaks in the ¹H–¹⁵N HSQC spectrum (they would get broader with increasing molecular weight) and clearly this does not occur in Fig. 4. Dimer formation would also result in an increase in the estimated isotropic overall rotational correlation time (τ_c) for cce_0567, as inferred from ¹⁵N spin relaxation times [38]. There was no experimentally significant change in τ_c with the addition of nickel to cce_0567, measured at 17.1 ± 0.5 ns in the metal-free form [18]. Finally, our crystal structure for cce_0567 with nickel corroborates with the dimer structure determined by NMR-based methods for Q60C73_METCA deposited into the Protein Data Bank (2JS5) that did not contain a metal ligand. While the presence of Ni²⁺ would have been difficult, if not impossible, to identify using NMR methods alone, it is likely that the authors of the Q60C73_METCA structure analyzed their protein for metals using alternative methods.

Fig. 4.

Overlay of the ¹H–¹⁵N HSQC spectrum of ¹⁵N-labeled cce_0567 (blue) over the spectrum collected in the presence a ~1:1 molar ratio of cce_0567:NiCl₂ (red). The inset illustrates one molecule of cce_0567 with the residues in the dimer interface colored black and those residues with perturbed chemical shifts after the addition of Ni²⁺ colored red. Spectra were collected at a proton resonance frequency of 750 MHz, 25 °C, in the same buffer used for protein crystallization.

To establish that cce_0567 was initially expressed and purified in the metal-free form the metal chelator EDTA was added to the ¹⁵N-labeled NMR sample [39]. Ten fold molar excess EDTA had no effect on the ¹H–¹⁵N HSQC spectrum suggesting that there was no significant level of metal bound to cce_0567 initially and that the chemical shifts deposited in the BioMagRes Bank database (BMRB-15539) are for the metal-free form of the protein. Indeed, after the addition of nickel to the solution, the ¹H–¹⁵N HSQC spectrum could be converted back to the metal-free spectrum by the addition of EDTA (data not shown). Similar metal binding experiments were repeated with the divalent cations Zn²⁺ and Co²⁺. Similar effects were observed to the ¹H–¹⁵N HSQC spectrum with the addition of Zn²⁺ as observed with Ni²⁺, while Co²⁺ had greater perturbations to the ¹H–¹⁵N HSQC spectrum due to the paramagnetic effects of the ion [35, 40]. The one noticeable difference was that at molar ratios greater than 1:1, the protein was visibly observed to start precipitating out of solution in the presence of Zn²⁺ and Co²⁺ while excess Ni²⁺ had no visible effect on solubility.

3.3. Circular dichroism profile and thermal stability of metal-free cce_0567

Circular dichroism spectroscopy is a powerful tool to probe the conformation of proteins in solution [41, 42] because small changes in the backbone conformation can cause large changes in the CD spectrum [43]. Consequently, different elements of protein secondary structure each have distinct CD patterns and it is possible to estimate the secondary structure content for a protein from the deconvolution of its CD spectrum [42]. CD spectroscopy can also be used to follow the effect of many variables, such as pH, salt content, and temperature, on the structure of a protein.

Fig. 5A contains the far-UV CD spectrum (solid line) for metal-free cce_0567 obtained in crystallization buffer at 25 °C. The dominant feature of the spectrum is the double minimumat 222 and 208–210 nm and maximum between 190 and 195 nm that are characteristic of a α-helical secondary structure [44]. Such an observation is expected given the crystal structure reported here for cce_0567. Fig. 5B assays the thermal stability of cce_0567 by monitoring the ellipticity at 220nmas a function of temperature between 10 and 80 °C. Avery gentle decrease in the ellipticity at 220 nmis observed from 10 to ~55 °C, upon which a steep decrease occurs up to ~75 °C at which point a plateau is reached. The inflection point for the transition is approximately 62 °C. Such a decrease in ellipticity with heating is typical of a phase transition from a folded to denatured state [45–48]. Indeed, the features of the CD spectrum of cce_0567 at 80 °C (dashed line) are characteristic of an unfolded protein [44]. However, after cooling the sample back to 25 °C (dotted line) the spectrum returns to essentially the identical pattern observed prior to heating the protein (solid line) indicating that denaturation is reversible. Such resilience to heating was confirmed by collecting ¹H–¹⁵N HSQC spectra for ¹⁵N-labeled cce_0567 before and after heating for 30 min at 80 °C. The spectrum after heating was essentially identical to the spectrum collected prior to heating (data not shown).

Fig. 5.

(A) Circular dichroism spectrum of cce_0567 (30 µM) at 25 °C (solid line), 80 °C (dashed line), and again at 25 °C (dotted line) in buffer containing 500 mM NaCl, 20 mM Tris, 1 mM DTT, pH 7.2. (B) CD thermal melt for cce_0567 (30 µM) collected at 220 nm in 2.5 °C intervals between 10 and 80 °C.

3.4. Biological context

Circumstantial evidence that cce_0567 plays a role in nitrogen fixation is the gene's location in the nitrogen fixation operon. Physical evidence that cce_0567 may play a biochemical role in nitrogen fixation comes from transcriptomics microarray data for Cyanothece sp. 51142 cultures grown as previously described [49] for 48 h in alternating 12-hour periods of light and dark [50]. The expression profile for cce_0567 oscillates in sync with the known, essential nif genes, suggesting that cce_0567 may play some biochemical role in the nitrogen fixation process. Indeed, in Cyanothece it was observed that all 34 genes in the nitrogen fixation transcriptional regulon exhibited strong co-regulation in their expression [50]. The function of 16 of these 34 genes, including cce_0567 (DUF683), has not yet been determined.

Many of the nif proteins involved in nitrogen fixation contain metals and the crystal structure and NMR data presented here shows that cce_0567 binds the divalent cation nickel. However, these nif proteins bind the metals Fe, Mo, and S [51] and not nickel. On the other hand, uptake hydrogenases associated with nitrogenases in cyanobacteria [52] often contain nickel, and therefore, it may be speculated that cce_0567 is somehow associated with these uptake hydrogenases. However, nickel binding to cce_0567 may simply be serependitious as the NMR metal titration studies presented here shows that cce_0567 also binds to the divalent cations Zn²⁺ and Co²⁺. Further studies are necessary to determine if nickel is the biologically significant ligand for cce_0567 in vivo.

While the biologically significant metal that cce_0567 binds is not known, ClustalW2 sequence alignment of all 48 currently identified proteins in the DUF683 family strongly suggest that metal binding is an important property of the protein members in this family. As shown by such a sequence alignment in Fig. 2, the histidine residue is conserved in 46 of the 48 DUF683 family members. Furthermore, residues near this conserved histidine, and especially those participating in the linker region between the two helices, are also highly conserved through the family members. It is therefore very likely that divalent metal binding mediated through the conserved histidine residue plays a role in the biological function of cce_0567 and all other protein in the DUF683 family. Further biochemical studies are necessary to define the precise biological role cce_0567 has in the nitrogen fixation process. Given that proteins in the DUF683 family are primarily observed in cyanobacteria and cyanobacteria are the only bacteria known to have circadian clocks, perhaps the role of cce_0567 and other proteins in the DUF683 family is related to the circadian control of nitrogen fixation.