Human COA3 Is an Oligomeric Highly Flexible Protein in Solution

Oct 28, 2016 - We found that the protein is oligomeric in aqueous solution, as shown by ...... of some flexible (not fully formed) helical structures ...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/biochemistry

Human COA3 Is an Oligomeric Highly Flexible Protein in Solution José L. Neira,*,†,‡ Sergio Martínez-Rodríguez,†,∇ José G. Hernández-Cifre,§ Ana Cámara-Artigas,∥ Paula Clemente,⊥,#,○ Susana Peralta,⊥,#,● Miguel Á ngel Fernández-Moreno,⊥,# Rafael Garesse,⊥,# José García de la Torre,§ and Bruno Rizzuti*,@ †

Instituto de Biología Molecular y Celular, Universidad Miguel Hernández, Elche, Alicante, Spain Biocomputation and Complex Systems Physics Institute, Zaragoza, Spain § Department of Physical Chemistry, University of Murcia, Murcia, Spain ∥ Department of Chemistry and Physics, University of Almería, Agrifood Campus of International Excellence (ceiA3), Almería, Spain ⊥ Departamento de Bioquímica-Instituto de Investigaciones Biomédicas “Alberto Sols”, Universidad Autónoma de Madrid-Consejo Superior de Investigaciones Científicas and Centro de Investigación Biomédica en Red de Enfermedades Raras (CIBERER), Madrid, Spain # Instituto de Investigación Sanitaria, Hospital 12 de Octubre (i+12), Madrid, Spain @ CNR-NANOTEC, Licryl-UOS Cosenza and CEMIF.Cal, Department of Physics, University of Calabria, 87036 Rende, Italy ‡

S Supporting Information *

ABSTRACT: The assembly of the protein complex of cytochrome c oxidase (COX), which participates in the mitochondrial respiratory chain, requires a large number of accessory proteins (the so-called assembly factors). Human COX assembly factor 3 (hCOA3), also known as MITRAC12 or coiled-coil domain-containing protein 56 (CCDC56), interacts with the first subunit protein of COX to form its catalytic core and promotes its assemblage with the other units. Therefore, hCOA3 is involved in COX biogenesis in humans and can be exploited as a drug target in patients with mitochondrial dysfunctions. However, to be considered a molecular target, its structure and conformational stability must first be elucidated. We have embarked on the description of such features by using spectroscopic and hydrodynamic techniques, in aqueous solution and in the presence of detergents, together with computational methods. Our results show that hCOA3 is an oligomeric protein, forming aggregates of different molecular masses in aqueous solution. Moreover, on the basis of fluorescence and circular dichroism results, the protein has (i) its unique tryptophan partially shielded from solvent and (ii) a relatively high percentage of secondary structure. However, this structure is highly flexible and does not involve hydrogen bonding. Experiments in the presence of detergents suggest a slightly higher content of nonrigid helical structure. Theoretical results, based on studies of the primary structure of the protein, further support the idea that hCOA3 is a disordered protein. We suggest that the flexibility of hCOA3 is crucial for its interaction with other proteins to favor mitochondrial protein translocation and assembly of proteins involved in the respiratory chain. itochondria are the “powerhouses” of cells, generating the bulk of cellular ATP. Cellular respiration occurs within them, involving oxygen consumption and ATP release. This process is known as the “electron transport chain” and involves five protein complexes: four enzymatic respiratory ones and the ATP synthase. Electrons delivered from NADH and succinate go through the electron transport chain to O2, which is finally reduced to H2O. The cytochrome c oxidase

M

© XXXX American Chemical Society

(COX) or complex IV, the fourth enzyme of the electron transport chain, catalyzes the oxidation of cytochrome c, transferring its electrons to O2. The mammalian complex IV is formed by 13 protein subunits: three of them (COX1 to Received: June 24, 2016 Revised: October 26, 2016 Published: October 28, 2016 A

DOI: 10.1021/acs.biochem.6b00644 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

of several types of detergents. We found that the protein is oligomeric in aqueous solution, as shown by both hydrodynamic techniques. Far-UV CD spectroscopy also suggests oligomeric species in the presence of detergents. Thermal and chemical denaturations suggest that the presence of secondary structure in aqueous solution, as monitored by far-UV CD and fluorescence quenching, is labile and nonrigid. In the presence of detergents, there is a slight increase in helical content, but this structure is not rigid. Combined with an in silico analysis, our experimental results suggest that hCOA3 is a disordered protein, with some regions populating different types of secondary structure or, alternatively, forming hydrophobic clusters.

COX3) are encoded by the mitochondrial DNA and form the core of the protein complex, whereas the remaining 10 are encoded in the nuclear genome.1,2 COX assembly occurs in a complex linear fashion,3−5 with several cofactors necessary in some proteins (required for electron transport) and different subunits being incorporated in order by the help of the so-called assembly factors. These proteins are required at every step of complex formation, from translation of the individual subunits of COX to assembly of the holoenzyme to addition of the prosthetic groups present in a few of the protein subunits.4,5 The biogenesis of COX must be tightly modulated to prevent the accumulation of pro-oxidant assembly intermediates; for instance, there is a negative feedback regulatory system in Saccharomyces cerevisae, which regulates the translation of COX1 in membrane-bound ribosomes and the accumulation of complex IV assembly intermediates.5,6 COX biogenesis is important because defects in enzyme assembly are a cause of mitochondrial disorders in humans.7−9 In fact, mitochondrial disorders are thought to be the result of dysfunction of the respiratory chain:10 most of the COX deficiencies in patients affected by mitochondrial disorders are caused by mutations in COX assembly factors.11 Human COX assembly factor 3 (hCOA3), also known as MITRAC12 or coiled-coil domain-containing protein 56 (CCDC56), is a 106-residue (11.7 kDa) polypeptide that is essential for the formation of the catalytic core of COX. We and other laboratories have recently shown that hCOA3 and its 87-residue variant in Drosophila melanogaster are mitochondrial transmembrane proteins essential for COX biogenesis in either species;12−14 in fact, hCOA3 is an inner mitochondrial membrane protein.14 The sequence of the human protein also is highly similar with that of yeast COX assembly factor 3.11,12,15 From a functional point of view, hCOA3 interacts with both proteins involved in the early steps of the biogenesis of complex IV and the translocation machinery located at the inner membrane of the mitochondria;14 in fact, the lack of hCOA3 leads to defective complex IV.12 Therefore, hCOA3 could be considered a new pharmaceutical target, and its structure and conformational stability first must be known in depth to design appropriate and specific drugs. The primary structure of hCOA3 does not include any disulfide bridge or free cysteine residues and encompasses a unique tryptophan and three tyrosines as intrinsically fluorescent residues.12 The high abundance of charged amino acids in the sequence of hCOA3 (27%, including 13 Asp/Glu acidic residues and 16 Lys/Arg basic ones) suggests that it may possess large unstructured regions. In this respect, we suspected that hCOA3 may behave as an intrinsically disordered protein (IDP); that is, it might belong to a class of proteins lacking stable secondary or tertiary structures.16,17 IDPs exist as an ensemble of rapidly interconverting structures, which fold into a well-defined three-dimensional structure only in the presence of their binding partners or their specific ligands,17 in some cases. Because of their high flexibility, IDPs may act as hubs in interaction networks performing several functions in the cell,16,17 and they have been recognized as potential drug targets.18 In this work, we have used several biophysical methods, namely, fluorescence and circular dichroism (CD) spectroscopies, together with analytical ultracentrifugation (AUC) and dynamic light scattering (DLS) as hydrodynamic techniques, to describe the stability and structure of hCOA3 in aqueous solution. We have also performed experiments in the presence



EXPERIMENTAL PROCEDURES Materials. Ultrapure urea was from ICN Biomedicals Inc. Concentrations of urea were calculated from the refractive index of the corresponding solution.19 Trizma acid and base, DNase I from bovine pancreas, chloramphenicol, tetracycline, NaCl, nickel resin, polyethylenimine (PEI), 3-[(3cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), sodium dodecyl sulfate (SDS), Nonidet P-40 (NP40), and 8-anilinonaphthalene-1-sulfonic acid (ANS) were from Sigma-Aldrich. Isopropyl β-D-thiogalactopyranoside (IPTG) and ampicillin were from Apollo. The β-ME was from BioRad. Triton X-100 was from Calbiochem. Standard suppliers were used for all other chemicals. Water was deionized and purified on a Millipore system. Cloning of hCOA3. Primers Ccdc56_5 (CATATGGCGTCTTCGGGAGCTGGTGAC) and Ccdc56_3 (CTCGAGGGACCCTGACGCCCTTGCCAGAGC), including NdeI and XhoI restriction sites, were used for polymerase chain reaction (PCR) amplification using a pRSET-B vector containing hCOA3.12 The PCR fragment obtained was purified from an agarose gel using QIAquick (Qiagen) and subcloned using a StrataClone PCR cloning kit (Stratagene). The isolated subcloning plasmid was purified using the QIAprep Spin miniprep kit (Qiagen) and then digested using NdeI and XhoI (Fermentas). The digested fragment was purified from an agarose gel using QIAquick (Qiagen) and then ligated into the pET22b+ plasmid (Novagen) cut with the same enzymes to create plasmid pET22b_CCDC56. Once the fragment had been cloned, it was sequenced using the dye dideoxynucleotide sequencing method in an ABI 377 DNA sequencer (Applied Biosystems). This construction allows the expression of hCOA3 fused to a C-terminal His6 tag in BL21 (DE3) RIL C+ cells (Agilent). We also tried Escherichia coli strains BL21 (DE3) and BL21 (DE3) pLys (Novagen), C41 (DE3) and C43 (DE3) (Lucigen), and BL21 (DE3) Star (Invitrogen, Life Technologies); however, none of them yielded a visible expression of hCOA3 in denaturing SDS gels, after staining. hCOA3 was obtained by growing chemically transformed BL21 (DE3) RIL C+ cells in 250 mL flasks containing ampicillin (100 μg/mL), chloramphenicol (50 μg/mL), and tetracycline (100 μg/mL). The flasks were left overnight at 37 °C. The next morning 1 L flasks, with the corresponding amounts of antibiotics, were inoculated with the culture that had been grown overnight. The cells cultured in the 1 L flasks were grown at 37 °C. When the absorbance at 600 nm reached 0.6−0.8 units, protein expression was induced by addition of IPTG at a final concentration of 700 μM, and the temperature was decreased to 15 °C. The cells in the 1 L flasks were grown overnight. B

DOI: 10.1021/acs.biochem.6b00644 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

did not affect the spectroscopic properties of the protein, as expected.21,22 Fluorescence. Spectra were recorded at 25 °C on a Cary Varian spectrofluorimeter (Agilent), by using a Peltier temperature controller. A 1 cm path length quartz cell (Hellma) was used. The final buffer concentrations were 10 mM in all cases. Excitation was carried out at 280 and 295 nm. Excitation and emission slits were 5 nm. Emission was measured from 300 to 400 nm, unless stated otherwise. The experiments were prepared the day before and left overnight at 5 °C for sample equilibration. Experiments at concentrations ranging from 2.6 to 13.2 μM (in protomer units) showed no changes in the maximal wavelength, but there were small changes in the normalized intensity. These results suggest the presence of concentrationdependent equilibria. Intrinsic Fluorescence. Urea denaturations at pH 7.0 (phosphate buffer), followed by either fluorescence or CD, were carried out by dilution of the proper amount of an 8 M urea stock solution. Urea denaturations were reversible. Appropriate blank corrections were made in all spectra. Protein concentrations used were 2 and 3 μM (in protomer units); both urea denaturations did show the same pattern (see Results). The pH of the samples was measured after completion of experiments with an ultrathin Aldrich electrode in a Radiometer (Copenhagen, Denmark) pH-meter. Salts and acids used (from pH 2.0 to 12.0) have been described previously.24 For the pH experiments, the protein concentration was 10 μM (in protomer units). Experiments in the presence of detergents (CHAPS, NP-40, SDS, and Triton X-100) were carried out with the same experimental set described above, except that the protein concentration was 2.6 μM (in protomer units) and the detergent concentrations used were either below or above their respective critical micelle concentrations (CMCs). For SDS, we assumed that the CMC was 8 mM;25,26 for CHAPS, the CMC was assumed to be 6 mM.27,28 Triton X-100 and NP40 were not used in the fluorescence experiments because of the strong fluorescence emission shown. Experiments were carried out at pH 8.0 (20 mM Tris) with 500 mM NaCl at 25 °C. Control experiments with 2.6 and 13.2 μM protein (in protomer units) in CHAPS (at 8 mM) and SDS (at 10 mM) were also carried out to determine whether different protein concentrations affected the shape and the intensity of the spectra (after normalization). The spectra had the same maximal wavelengths at the two concentrations, but the normalized intensity slightly changed. Furthermore, as the shape of both spectra and the maximal wavelengths were the same as those of the experiments performed in aqueous solution, we suggest that the presence of the detergent did not alter the oligomerization state of the protein (see Results). Thermal Denaturations. Experiments were performed at heating rates of 60 °C/h, from 25 to 90 °C, with an average time (sampling time) of 1 s. Thermal scans were collected at 315, 330, and 350 nm after excitation at 280 and 295 nm. Thermal denaturations were reversible. Thermal experiments in the presence of both concentrations of detergents (either above or below their CMCs) were performed with the same experimental set and at the same protein concentration (2.6 μM in protomer units). Thermal denaturations were reversible.

Cell pellets were resuspended in 50 mL of buffer A [20 mM Tris (pH 8.0), 0.5 M NaCl, 1 mM β-ME, 0.1% Triton X-100, 5 mM imidazole, 10 μg/mL DNase I, and 20 mM MgSO4] supplemented with a tablet of Sigma Protease Cocktail EDTAfree. After being incubated with gentle agitation at 4 °C for 10 min, cells were disrupted by sonication (Branson, 750 W), with 10 cycles of 45 s at 55% of maximal power output and an interval of 15 s between the cycles. All the sonication steps and the interval waits were carried out in ice. The lysates were clarified by centrifugation at 24000 rpm for 40 min at 4 °C in a Beckman JSI30 centrifuge. The clarified lysate from this first centrifugation did not contain hCOA3, which was present in the precipitate. Thus, the precipitate was treated with buffer A supplemented with 8 M urea. The resuspended sample was treated with another 10 cycles of sonication in ice, and the sample was clarified by centrifugation at 20000 rpm for 30 min at 4 °C. The hCOA3 was in the supernatant and was purified by immobilized affinity chromatography (IMAC). The supernatant was added to 5 mL of His-Select HF nickel affinity gel (Sigma-Aldrich) previously equilibrated in buffer A supplemented with 8 M urea. The mixture was incubated for 20 min at 4 °C, and after that time, the lysate was separated from the resin by gravity. On-column refolding was carried out during column wash with 20 mL of buffer B [20 mM Tris (pH 8.0), 0.5 M NaCl, 1 mM β-ME, and 20 mM imidazole]; the protein was then eluted by gravity from the column with buffer C [20 mM Tris (pH 8.0), 0.5 M NaCl, 1 mM β-ME, and 500 mM imidazole]. The eluted hCOA3 was extensively dialyzed against buffer C, in the absence of imidazole and β-ME. The final yield of protein was 1.5−2.5 mg/L of culture, and the protein was 85−90% pure as judged by SDS gels. We attempted to repurify the protein recovered from IMAC by using gel filtration chromatography in a Superdex 16/600, 75 pg column performed in an AKTA Basic system (GE Healthcare) by monitoring the absorbance at 280 nm; nevertheless, the protein was bound to the column, eluting at volumes larger than the bed volume (∼125 mL). Further attempts with a Heparin column [5 mL HiTrap Heparin column (GE Healthcare)] were also unsuccessful, leading to protein precipitation during dialysis. The eluted protein from IMAC showed absorbance at 260 nm, suggesting that hCOA3 was probably contaminated with traces of DNA, even though DNase was used during purification. We believe that this contamination was due to electrostatic interactions between the highly charged hCOA3 (theoretical pI of 9.22) and the nucleic acid. The total protein concentration, Pc (in milligrams per milliliter) was determined by using the expression20 Pc = 1.55A280 − 0.76A260, where A280 and A260 are the absorbances of the dialyzed protein solution at 280 and 260 nm, respectively. However, it is important to note that the presence of DNA traces did not affect the spectroscopic signals in the far-UV region or fluorescence, because DNA is spectroscopically silent in these techniques.21,22 We also tried to eliminate the DNA traces by using different concentrations of PEI [ranging from 0.2 to 1% (v/v)],23 but most of the hCOA3 coprecipitated with DNA. After PEI precipitation, the supernatant was concentrated by using Amicon centrifugal devices (molecular weight cutoff of 3500 Da, Millipore), and fluorescence and far-UV CD spectra were acquired with the concentrated protein stock solution. Both types of spectra were similar to those acquired in the absence of PEI (i.e., in the presence of small amounts of DNA); therefore, these findings indicate that such DNA traces C

DOI: 10.1021/acs.biochem.6b00644 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry Fluorescence Quenching. Quenching of tryptophan and tyrosines with either iodide or acrylamide was examined under different solution conditions. The ionic strength was kept constant during KI quenching by addition of KCl; also, Na2S2O3 was added (concentration of 0.1 M) to prevent the formation of I3−. The presence of KCl did not alter the hCOA3 structure as suggested by the absence of changes in the shape and ellipticity of the CD spectra in the presence of 0 and 1 M KCl (data not shown). For quenching with KI, the data were fitted to the equation29 F0/F = 1 + Ksv[X], where Ksv is the Stern−Volmer constant for collisional quenching, F0 is the fluorescence when KI is not present, and F is that at any KI concentration. The range of KI concentrations explored was 0− 0.7 M. We measured the quenching of fluorescent residues at pH 4.0, 7.0, and 10.0. Experiments were also carried out in the presence of 5 M urea at pH 7.0. For acrylamide, the experimental spectroscopic parameters were the same as in KI. However, the Stern−Volmer equation was modified to include an exponential term to account for dynamic quenching:29 F0/F = (1 + Ksv[X])eν[X], where ν is the dynamic quenching constant (component). This equation is identical to the one given above, when ν = 0. ANS Binding. The excitation wavelength was 380 nm, and emission fluorescence was measured from 400 to 600 nm during the pH denaturation. ANS stock solutions were prepared in water and diluted to yield a final concentration of 100 μM. Signals from blank solutions were subtracted from the corresponding spectra. Circular Dichroism. CD spectra were recorded on a Jasco J815 (Japan) or Jasco J810 spectropolarimeter fitted with a thermostated cell holder and interfaced with a Peltier unit. The instrument was periodically calibrated with (+)-10-camphorsulfonic acid. The molar ellipticity was calculated as described previously.24 Protein concentrations were 10 μM (in protomer units) in a 0.1 cm path length cell for pH experiments and 8 μM (in protomer units) for chemical denaturations. Far-UV Spectra. Isothermal wavelength spectra of hCOA3 at different pH values or urea concentrations were acquired at a scan speed of 50 nm/min with a response time of 4 s and averaged over six scans at 25 °C, with 10 mM buffer. Spectra were corrected by subtracting the baseline in all cases. The chemical and pH denaturations were repeated at least twice with different samples. The samples were prepared the day before and left overnight at 5 °C to allow for equilibration. For steady-state experiments in the presence of detergents, two protein concentrations were explored: 9.9 and 20 μM (in protomer units). Experiments were carried out at pH 8.0 (20 mM Tris) with 500 mM NaCl at 25 °C. The concentrations used for the detergents were 4 and 10 mM for SDS, 2 and 8 mM for CHAPS, 0.1 and 0.5 mM for NP-40, and 0.1 and 1 mM for Triton X-100. Control experiments with 10 and 16.2 μM protein (in protomer units) in CHAPS (at 8 mM) and SDS (at 10 mM) were also carried out to determine whether different protein concentrations affected the shape and/or intensity of the spectra after normalization. The spectra had the same shape, but the normalized intensity slightly changed, suggesting the presence of self-association equilibria even in the presence of detergents. Thermal Denaturations. The experiments were performed at heating rates of 60 °C/h and a response time of 8 s. Thermal scans were collected following the ellipticity at 222 nm from 25 to 80 °C. Solution conditions were the same as those in the

steady-state experiments. No drifting of the spectropolarimeter was observed. Thermal denaturations were reversible. For experiments in the presence of detergents, 9.9 μM protein (in protomer units) was used. The employed detergent concentrations were the lowest among those indicated above. Thermal denaturations were always reversible. Analytical Ultracentrifugation. Experiments were conducted as described previously.30,31 Briefly, they were performed in a Beckman Coulter Optima XL-I analytical ultracentrifuge (Beckman-Coulter, Palo Alto, CA) equipped with UV−visible absorbance as well as interference optics detection systems, using an An50Ti eight-hole rotor and 12 mm path length charcoal-filled Epon double-sector centerpieces. The experiments were performed at 20 °C in 50 mM Tris (pH 7.9), with 20 μM protein (in protomer units). The laser delay was adjusted prior to the runs to obtain high-quality interference fringes. Light at 675 nm was used in the interference optics mode. Sedimentation velocity runs were carried out at a rotor speed of 40000 rpm using 400 μL samples. A series of 900 scans, without time intervals, were acquired for each sample. A least-squares boundary modeling of the sedimentation velocity data was used to calculate sedimentation coefficient distributions with the size distribution c(s) method,32 implemented in SEDFIT version 13.0b. The sedimentation coefficient in water, s20,w, solvent density (ρ = 1.0089 g/mL), and viscosity (η0 = 1.002 cP) at 20 °C were estimated using SEDNTERP.33 The partial specific volume of the protein, V̅ , was 0.7298 mL/g. Dynamic Light Scattering. Experiments were performed with a Zetasizer Nano ZS instrument (Malvern Instruments Ltd., Worcestershire, U.K., 4 mV He−Ne laser, λ0 = 633 nm, and Θ = 173°) using a thermostated 12 μL quartz sample cuvette at 20 °C, as described previously.30 Briefly, protein samples were prepared in 50 mM Hepes buffer (pH 7.0). The sample concentration was 20 μM (in protomer units). All the solutions were filtered immediately before measurement, and protein samples were centrifuged for 30 min at 14000 rpm to remove any aggregates and dust. Data were analyzed using the manufacturer’s software: the hydrodynamic radius, Rs, and molecular mass, M, were determined from the Stokes−Einstein equation, assuming a spherical shape for the protein. Analysis of the pH Denaturation Curves. The pH denaturation curves were analyzed assuming that both protonated and deprotonated species contributed to the monitored spectral feature. The physical property, X, being observed, either ellipticity or fluorescence intensity, depends upon pH according to X=

Xa + Xb10n(pH − pKa) 1 + 10n(pH − pKa)

(1)

where Xa is the value observed for the acidic species, Xb is that at high pH, pKa is the apparent midpoint of the titrating group, and n is the Hill coefficient,34 which was in the range of 0.9−1.1 in all the curves. The apparent pKa reported from intrinsic or ANS fluorescence and CD was obtained from three different measurements for each technique, performed on different samples. Fitting by nonlinear least-squares analysis to eq 1 was performed by using Kaleidagraph (Abelbeck software) working on a personal computer. D

DOI: 10.1021/acs.biochem.6b00644 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry



RESULTS Sequence Analysis Indicates That hCOA3 Contains Intrinsically Disordered Regions. Our initial studies of the primary structure of hCOA3 suggest that the protein has a single transmembrane domain (from residue 58 to 78) and a coiled-coil domain (from residue 79 to 104).12 The overall amino acid composition identifies hCOA3 as a folded protein in a charge−hydropathy plot.35,36 On the other hand, as shown in Figure 1A, both charged and hydrophobic residues are

(including the transmembrane domain) is ordered, whereas the N-terminal and C-terminal regions are predicted to be disordered. The use of Fold-Index43 suggests that only the first 40 residues were disordered, while the rest of the protein (the transmembrane domain and the coiled-coil regions) is ordered. Furthermore, s2D44 yielded high percentages of helical structure for the regions of residues 30−50 and 80−95 (which is the predicted coiled-coil region12), and the rest of the sequence appeared to be disordered. hCOA3 Is an Oligomeric Protein in Aqueous Solution. We tried to characterize the self-association state of hCOA3 with four hydrodynamic techniques. Nuclear magnetic resonance diffusion ordered spectroscopy did not yield reliable results, because the observed peaks were very broad (data not shown). Likewise, we could not carry out size exclusion chromatography experiments because, as discovered during the purification protocol (see Experimental Procedures), hCOA3 was bound to the gel filtration column. Therefore, we carried out hydrodynamic experiments with only AUC and DLS techniques. In the AUC measurements in the interference mode, we could detect one main peak with an s20,w of 1.8 S. However, there were also a secondary peak around an s20,w of 3 S and very broad peaks with an average of s20,w of 8 S (Figure 2A). Fitting

Figure 1. Sequence-based disorder predictors of hCOA3. (A) Protein sequence and distribution of charged residues. Acidic and basic amino acids are indicated with red triangles and blue circles, respectively. (B) Probability of disorder propensity as a function of residue index, as predicted by RONN,39 IUpred,40 DISOclust,41 and PrDOS.42 The cutoff between ordered and disordered conformations is 0.5 in all cases.

unevenly distributed along the protein sequence, with the transmembrane domain lacking the former. In fact, both regions preceding and following the transmembrane domain are separately classified as natively unfolded in the same charge−hydropathy plot.36 In addition, hCOA3 contains a large amount of Ala residues (14.2%, approximately twice the frequency found for proteins37), which typically tend to increase the RS and to modulate the backbone sensitivity to local structural perturbations in proteins.38 These results suggest that hCOA3 could behave as an IDP under a range of conditions, showing a high flexibility with a low degree of order in large regions of its structure. To test this hypothesis, we submitted the protein sequence to different disorder prediction servers. As shown in Figure 1B, all of them (RONN, DISOclust, IUPred, and PrDOS) suggest39−42 that only the region from residue 50 to 80

Figure 2. Hydrodynamic measurements in hCOA3. (A) Interference mode of the AUC experiment. The rate of the frictional factor of the protein, f, to that of a spherical protein with the same number of residues, f/f 0, was fixed to 1.5, and the root-mean-square deviation obtained was 0.0075. (B) DLS experiments shown as a percentage of volume peak.

of the first peak with a fixed value of the rate of the frictional coefficient of the protein with that of a sphere to 1.5 yields a molecular mass of 15.6 kDa (which is close to that theoretically calculated from the sequence, 12.8 kDa). However, the presence of self-associated species, with estimated M values of 40 and 200 kDa, is clear (Figure 2A). These results suggest that hCOA3 does not form a single self-associated species; instead, E

DOI: 10.1021/acs.biochem.6b00644 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

Figure 3. Spectroscopic characterization of hCOA3 in aqueous solution. (A) Fluorescence intensity at 330 nm after excitation at 280 nm (●, left axis) and mean molar ellipticity at 222 nm (○, right axis) as the pH was modified. In fluorescence, similar curves were obtained after excitation at 295 nm at any other emission wavelengths. Experiments were performed at 25 °C. (B) Thermal denaturations followed by fluorescence at 330 nm (excitation at 280 nm), at several pH values (the units on the y-axis are arbitrary). (C) ANS fluorescence at 480 nm as the pH was modified. The inset shows the far-UV CD of hCOA3 at pH 7.0. Experiments were conducted at 25 °C. (D) Thermal denaturations followed by changes in ellipticity at 222 nm (the units on the y-axis are arbitrary).

hCOA3 is 15.6 Å. In addition, the RS of a folded spherical protein can also be approximated by the equation46 RS = (4.75 ± 1.11)N0.29, where N is the number of residues; for a spherical, monomeric hCOA3, this expression yields a value of 19 ± 4 Å. A more accurate formula, specifically designed for IDPs,47 that takes into account the fraction of Pro residues and the absolute net charge of the protein sequence, yields an RS of 25 Å for hCOA3, which is larger than the values given above. The statistical and maximal uncertainties on this latter estimate are ∼2 and ∼4 Å,47 respectively, and therefore, even this value is not compatible with the experimental ones we obtained. Then, these theoretical results clearly suggest that the shape of hCOA3 was not spherical, nor was it a monomer in solution. hCOCA3 Acquires a Nonrigid Structure in a Narrow pH Range in Aqueous Solution. To elucidate which kind of structure (if any) is present in hCOA3, and whether it is stable over a wide pH range, we could use several biophysical techniques: ANS fluorescence, intrinsic fluorescence, and farUV CD. All those techniques provide complementary information about different structural features of hCOA3. We used ANS fluorescence to monitor the burial of solventexposed hydrophobic patches.48 Intrinsic fluorescence was used to monitor changes in the tertiary structure of the protein around its sole tryptophan (Trp46) and tyrosines (Tyr72, Tyr74, and Tyr77). Finally, we carried out far-UV CD experiments to monitor changes in the secondary structure. Steady-State Fluorescence and Thermal Denaturation. The fluorescence spectrum of hCOA3 at pH 7.0 and 20 °C had a maximum at 341 nm, suggesting that (i) protein fluorescence was dominated by the unique tryptophan, Trp46, and (ii) this

there are several oligomerization equilibria with different reaction orders. The DLS experiments (in volume) also showed several peaks corresponding to a hydrodynamic radius of 44 ± 6 Å (yielding an M of 114 ± 65 kDa, assuming a spherical shape for hCOA3) (Figure 2B), 325 ± 250 Å, and 2600 ± 700 Å. The overall polydispersity index is 0.47 (68.5% of polydispersity), which indicates that the sample is, in fact, polydisperse. However, the 94.6% of the sample in the volume distribution graph corresponds to the first peak with an only 13.9% polydispersity; this result suggests that this first peak is monodisperse, but with a very specific self-association. In summary, all the peaks observed in DLS suggest the presence of self-associated protein species. These results further support the presence of several oligomerization equilibria with different reaction orders. Although we cannot obtain any measurement of the order of the oligomerization or the molecular weight of the species involved, we also recorded far-UV CD spectra at two different protein concentrations: 10 and 16.2 μM (in protomer units). The results (Figure 1 of the Supporting Information) suggest that, although both spectra were similarly shaped, there were changes in intensity, and therefore, the protein had a tendency to aggregate. We can further elaborate on the expected theoretical value of the RS for a protein of the size of hCOA3. The RS value for an ideal spherical molecule can be theoretically calculated by considering that the unsolvated molecular volume, MV̅ /NA, is that of a sphere:45 RS = (3MV̅ /4NAπ)1/3, where NA is Avogadro’s number. The calculated RS for a spherical, monomeric protein with M and V̅ corresponding to those of F

DOI: 10.1021/acs.biochem.6b00644 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

pH, the fluorescent residues are more solvent-shielded than they are under the other conditions tested. This tendency is not as clearly seen in the presence of acrylamide, because at pH 10.0, the Ksv was smaller than at physiological pH; we explain this effect as being due to the contribution of ν. In conclusion, although no sigmoidal thermal transition was observed at any pH, it seems clear that at physiological pH the fluorescence residues had a more restricted solvent accessibility. ANS Binding. At low pH, the ANS fluorescence intensity at 480 nm was very large and decreased as the pH was increased (Figure 3C), indicating that hCOA3 had solvent-exposed hydrophobic regions at low pH. The intensity at 480 nm showed a sigmoidal behavior, with a pKa of 4.4 ± 0.2. This value was smaller than that obtained from intrinsic fluorescence, suggesting that the solvent-exposed hydrophobic patches are buried before acquisition of tertiary structure around the fluorescent residues occurs. Circular Dichroism. We recorded the far-UV spectrum of the protein at pH 7.0 and 20 °C. The far-UV CD spectrum had a shape resembling that of α-helices (Figure 3C, inset), but the maxima at 208 and 222 nm were not very intense. Deconvolution of the spectrum by using the online Dichroweb server49,50 yields percentages of 8% for α-helix, 44% for β-sheet, and 48% for random coil, but the root-mean-square deviation of the experimental and fitted spectra was very poor. These results suggest that the main secondary structure component observed is indicative of the presence of a β-sheet, in disagreement with the values of some of the theoretical predictions (see above), based on the primary structure. Furthermore, the deconvoluted values do not agree with the shape of the far-UV CD spectrum (Figure 3C, inset), where the presence of an α-helix structure could be guessed. The reasons of such a discrepancy are not clear, but a possible explanation is that the minimum at 222 nm (and, thus, the expected high content of helical structure) could be due to the presence of aromatic residues, which also absorb at this wavelength.51−53 We also carried out pH denaturation experiments. As the pH was increased, the ellipticity at 222 nm increased (in absolute value) in a sigmoidal-like fashion, with a pKa of 6.0 ± 0.1 [Figure 3A (○, right axis)], and thus, at low pH values, the ellipticity at this wavelength was smaller than at the physiological wavelength. The pKa value is different from that observed by intrinsic and ANS fluorescence, suggesting that the acquisition of secondary structure (monitored by far-UV CD) occurred at a later stage than both (i) the formation of tertiary structure around fluorescent residues and (ii) the burial of solvent-exposed hydrophobic amino acids. Thermal denaturations followed by CD, at the same pH values as in the fluorescence experiments, did not show a clear sigmoidal behavior (Figure 3D). Attempts to fit these curves to the Gibbs−Helmholtz equation30 led to unrealistic values for the thermal denaturation midpoints. Chemical Denaturations of hCOA3 in Aqueous Solution. Because thermal denaturations did not lead to a proper measurement of hCOA3 stability, we carried out urea denaturations followed by CD (at 10 μM in protomer units) and fluorescence (at 2 and 3 μM in protomer units). Denaturations were conducted at pH 7.0 (Tris buffer), where the spectroscopic properties of the protein seemed to reach a plateau (Figure 3A). In the CD curves, as the concentration of urea was increased, the minimum at 222 nm disappeared, indicating the disruption of the protein secondary structure. The denaturation curves

residue was slightly buried (for a fully solvent-exposed Trp residue, the maximum should be 350 nm19,29). The intensity at 330 nm (as at any of the other explored wavelengths) showed a bell shape as the pH varied, with two transitions [Figure 3A (●)]. The first occurred with a pKa of 5.4 ± 0.1, whereas the second happened at basic pH and was probably due to the titration of the phenol group of some of the three tyrosines; this transition was not yet complete at pH 12.0, and therefore, we could not obtain its midpoint. Thermal denaturations at several pH values (3.1, 5.4, 7.4, and 12.1) were performed (Figure 3B). We did not observe any sigmoidal behavior at any pH; only at pH 7.4 could a very flat transition be guessed. Trying to fit this curve led to unreliable results, with apparent thermal denaturation midpoints below 0 °C. All transitions were reversible, except that at pH 3.1, where scattering at high temperatures was observed (Figure 3B). Quenching Experiments. From the results given above, it seems that there are several pH intervals where the structure of hCOA3 could be different: the one at pH