BBAPAP-39353; No. of pages: 11; 4C: 4, 5, 7 Biochimica et Biophysica Acta xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbapap

5

F

O

4

Consuelo Latorre Fortes-Dias a,⁎, Paula Ladeira Ortolani a, Carlos Alexandre H. Fernandes b, Kelli Roberta Lobo a, Lutiana Amaral de Melo a, Márcia Helena Borges a, Wallance Moreira Pazin c, Mário de Oliveira Neto b, Roberto Morato Fernandez b, Marcos Roberto M. Fontes b,⁎

6 7 8

a

9

a r t i c l e

10 11 12 13 14

Article history: Received 11 February 2014 Received in revised form 28 April 2014 Accepted 1 May 2014 Available online xxxx

15 16 17 18 19 20 21

Keywords: Phospholipase A2 inhibitor PLI Phospholipase A2 PLA2 Crotalus Quaternary structure

i n f o

a b s t r a c t

Several snake species possess endogenous phospholipase A2 inhibitors (sbPLIs) in their blood plasma, the primary role of which is protection against an eventual presence of toxic phospholipase A2 (PLA2) from their venom glands in the circulation. These inhibitors have an oligomeric structure of three to six subunits and have been categorized into three classes (α, β and γ) based on their structural features. SbγPLIs have been further subdivided into two subclasses according to their hetero or homomeric nature, respectively. Despite the considerable number of sbγPLIs described, their structures and mechanisms of action are still not fully understood. In the present study, we focused on the native structure of CNF, a homomeric sbγPLI from Crotalus durissus terrificus, the South American rattlesnake. Based on the results of different biochemical and biophysical experiments, we concluded that, while the native inhibitor occurs as a mixture of oligomers, tetrameric arrangement appears to be the predominant quaternary structure. The inhibitory activity of CNF is most likely associated with this oligomeric conformation. In addition, we suggest that the CNF tetramer has a spherical shape and that tyrosinyl residues could play an important role in the oligomerization. The carbohydrate moiety, which is present in most sbγPLIs, is not essential for the inhibitory activity, oligomerization or complex formation of the CNF with the target PLA2. A minor component, comprising no more than 16% of the sample, was identified in the CNF preparations. The amino-terminal sequence of that component is similar to the B subunits of the heteromeric sbγPLIs; however, the role played by such molecule in the functionality of the CNF, if any, remains to be determined. © 2014 Published by Elsevier B.V.

E

T

R

E

c

Diretoria de Pesquisa e Desenvolvimento, Fundação Ezequiel Dias (FUNED), Rua Conde Pereira Carneiro 80, CEP 30510-010, Belo Horizonte, MG, Brazil Departamento de Física e Biofísica, Instituto de Biociências, Universidade Estadual Paulista, UNESP, Botucatu, SP, Brazil Departamento de Física e Matemática, Faculdade de Filosofia Ciências e Letras, USP, Ribeirão Preto, SP, Brazil

C

b

R O

3Q1

P

2

Insights on the structure of native CNF, an endogenous phospholipase A2 inhibitor from Crotalus durissus terrificus, the South American rattlesnake

D

1

38

R

42 40 39

N C O

41

1. Introduction

44 45

Snake venoms are amongst the richest sources of phospholipases A2 (PLA2, EC 3.1.1.4), a widespread superfamily of structurally related enzymes that hydrolyze glycerophospholipids in lysophospholipids and free fatty acids. Based on molecular and biochemical criteria, secretory PLA2 (sPLA2) has been classified into 15 different groups [1]. Snake venom PLA2 belongs to groups I and II, which also encompasses pancreatic and inflammatory sPLA2. The sPLA2s in snake venoms may exert their deleterious actions as monomeric or multimeric toxins with at least one catalytically active subunit. A well-known example of a multimeric sPLA2 is the crotoxin (Ctx), a β-neurotoxin [2], which is the major toxic component of the venom of the South American rattlesnake, Crotalus durissus terrificus.

48 49 50 51 52 53 54 55

U

43

46 47

Q6 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

⁎ Corresponding authors. E-mail addresses: [email protected] (C.L. Fortes-Dias), [email protected] (M.R.M. Fontes).

Ctx has a heterodimeric structure formed by an acidic, non-enzymatic subunit (CA) and a basic, enzymatically active counterpart (CB) [3] that are strongly bound by non-covalent interactions. CA acts as a chaperone, preventing the non-specific adsorption of CB to membrane structures other than its physiological target, thereby enhancing the pharmacological potency and, consequently, the lethal effect of CB [4]. The structural and functional activities of Ctx have been recently reviewed by [5]. Aiming, primarily, at a physiological protection against the eventual presence of venom gland contents in their circulating blood, several snake species have been provided with PLA2 inhibitors, generally referred to as sbPLIs (snake blood phospholipase A2 inhibitors). During the last two decades, a growing number of sbPLIs has been described for several venomous and non-venomous snake species [6]. These inhibitors have an oligomeric structure of, at least, three subunits and, based on known mammalian protein domains, they have been categorized into three structural classes (α, β and γ). Members of these three classes can be concomitantly present in a single snake species.

http://dx.doi.org/10.1016/j.bbapap.2014.05.001 1570-9639/© 2014 Published by Elsevier B.V.

Please cite this article as: C.L. Fortes-Dias, et al., Insights on the structure of native CNF, an endogenous phospholipase A2 inhibitor from Crotalus durissus terrificus, the South American..., Biochim. Biophys. Acta (2014), http://dx.doi.org/10.1016/j.bbapap.2014.05.001

56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73

C.L. Fortes-Dias et al. / Biochimica et Biophysica Acta xxx (2014) xxx–xxx

128

Heparinized blood plasma from C. d. terrificus snakes was obtained from the Serpentarium of Fundação Ezequiel Dias, depending on the availability of specimens. The procedure followed the protocol approved by the Committee for Ethics in Animal Use of the Fundação Ezequiel Dias (CEUA FUNED 022/2012). Native CNF was purified in two steps, i.e., ion exchange in DEAE-Sephacel [14] followed by a hydrophobic interaction chromatography on a Hitrap Phenyl FF (GE HealthCare) [7], starting with plasma from different C. d. terrificus specimens. All the final preparations consistently displayed the expected 24 kDa and 20 kDa protein bands on SDS-PAGE [15], under variable ratios. These bands correspond to the glycosylated and the non-glycosylated monomers, respectively, in native CNF. The non-glycosylated monomer is always present at much lower proportion than the glycosylated monomer. Crotoxin (Ctx) was purified from whole C. d. terrificus snake venom by classical gel filtration on Sephadex G75 [16].

129

2.2. Gel filtration chromatography

130 131

Gel filtration chromatography was performed in an Akta Purifier 10 (GE HealthCare) with Protein-Pak 300 SW (Waters, Millipore Co.), Superose 6 HR10/30, Superose 12 HR10/30 and Superdex 200 HR10/ 30 (GE HealthCare) columns, using 0.1 M sodium phosphate pH 7.0 as elution buffer. Each column was previously calibrated with the following

101 102 103 104 105 106 107 108 109 110

115 116 117 118 119 120 121 122 123 124 125 126 127

132 133 134

C

99 100

E

97 98

R

95 96

R

93 94

O

91 92

C

89 90

N

87 88

U

85 86

2.3. Cross-linking assays

147

F

114

83 84

137 138 139 140 141 142 143 144 145 146

BS3 (Bis[sulfosuccinimidyl]suberate, Pierce) or glutaraldeyde (25% EM grade, Merck) was added in increasing concentrations (1 to 10 mM) to a fixed amount of CNF (native or enzymatically deglycosylated) in 20 mM HEPES pH 7.5, achieving a final concentration of 6.7 mg/ml of the inhibitor. A control sample was prepared using the same concentration of CNF but omitting the BS3. In addition, a control experiment was performed with 6.7 mg/ml of bovine serum albumin (B2518 from Sigma Co.) in the presence of increasing concentrations of BS3 (1 to 10 mM), in order to check for non-specific reactions of the cross-linker. The mixtures were incubated in the dark for 3 h at room temperature before subjecting them to the analysis by SDS-PAGE on 8–25% gradient Phast® gels (GE HealthCare) under non-reducing conditions. The CNF molecular mass (in kDa) was calculated by interpolation, using the calibration curves obtained with protein markers and expressed as mean ± SD of two independent experiments per cross-linker (n = 4). The cross-linking experiment was repeated by incubating increasing concentrations of native CNF with 10 mM of BS3. The final concentrations of CNF were 6.7, 16.7 and 33.3 mg/ml. A control sample was prepared with 6.7 mg/ml of CNF but omitting the BS3. The cross-linked samples were analyzed by SDS-PAGE, as before.

148 Q8

2.4. Dynamic light scattering

168

Dynamic light scattering (DLS) measurements with native CNF (2.5 mg/ml in 10 mM ammonium formate pH 6.5) were carried out at 4 °C (277 K), 10 °C (283 K) and 18 °C (291 K) on a DynaPro Titan equipment (Wyatt Technology). The average values of one hundred measurements were analyzed with the Dynamics DynaPro version 6.10 software.

169 170

2.5. Small angle X-ray scattering

174

Small angle X-ray scattering (SAXS) data for CNF was collected on the D02A-SAXS2 beam line at the Brazilian Synchroton Light Laboratory (LNLS). CNF samples were prepared in 20 mM ammonium formate pH 6.5 at the concentrations of 1 and 5 mg/ml (41.7 μM and 208.4 μM, respectively). The samples were centrifuged for 30 min at 23,500 ×g at 4 °C to remove potential aggregates. The radiation wavelength was set to 1.48 Å and a 165 nm MarCCD detector was used to record the scattering patterns. The sample-to-detector distance was set to 1000 mm to obtain the range of the scattering vector q from 0.013 to 0.33 Å−1, where q is the magnitude of the q-vector defined by q = 4π sinθ/λ (2θ is the scattering angle). Two successive frames, each of 300 s duration, were recorded for each sample to monitor radiation damage and beam stability. After buffer scattering subtraction, protein SAXS patterns were integrated using Fit2D software [17]. The radius of gyration, Rg, was computed from the Guinier equation [18] and by indirect Fourier transform method using the Gnom package [19]. The distance distribution p(r) also was calculated using the Gnom, and the maximum diameter, Dmax, was obtained. The molecular mass estimation was calculated using the SAXS Mow web tool [20].

175 176

O

2.1. Purification of CNF and crotoxin

81 82

R O

113

80

P

2. Materials and methods

78 79

135 136

D

112

76 77 Q7

reference proteins acquired from Sigma Co: aprotinin (A1153), bovine serum albumin (Sigma A7517), cytochrome C (C3131), concanavalin A (C2010), ovalbumin (A7642), and RNase (R6513). Catalase, ferritin and thyreoglobulin from an old kit (Pharmacia Fine Chemicals, 170441-01) complemented the range of reference molecular masses. Firstly, each protein was loaded individually to guide their unequivocal identification in the elution profiles of the protein mixture. Calibration curves were constructed by the linear least squares regression method using the Graph Prism® 6.0 for Mac OS X (GraphPad software Inc., La Jolla, California). Horseradish peroxidase (HRPO P6782, Sigma Co.) and Ctx (prepared in our laboratory) were run as controls of a glycosylated and a non-glycosylated protein, respectively.

T

111

The αPLIs have a C-type lectin domain, whereas the distinguishing feature in the βPLIs is the occurrence of leucine-rich repeats, known as LRR. The members of these three classes can be concomitantly PLIs, which in turn, are composed of two structural units of highly conserved tandem repeats of half cysteines known as three-finger motifs [7]. Based on the current understanding, the last class comprises the highest number of sbPLIs that are, typically, oligomers of glycosylated and nonglycosylated subunits. Based on the identity of their subunits, Lizano et al. [8] proposed a sub-classification into heteromeric (subclass I) or homomeric (subclass II) sbMembers of these three classes can be concomitantly PLIs. A gamma sbγPLI from C. d. terrificus snakes, named CNF (for an acronym of Crotalus neutralizing factor) has been extensively studied by the authors of this paper, leading to important conclusions on its mechanism of action. CNF is able to displace CA from the native Ctx complex and to bind tightly to CB, thus forming a stable CNF.CB complex. CNF.CB is devoid of any PLA2 activity and may be considered as reminiscent of the interaction of Ctx with its target receptor at the pre-synaptic neuromuscular junctions [9,10]. Native CNF is an oligomer of glycosylated and non-glycosylated single subunits of 24 kDa and 20 kDa [9] and has been assigned to subclass II of sbγPLIs [8]. However, fundamental questions remain to be answered such as the number of subunits in the oligomer, the role of glycosylation in the inhibitor functionality and its homomeric character. These issues are not exclusive to CNF but can be extended to most sbγPLIs. Although oligo/polymeric structures have been assigned to all of these inhibitors, the number of forming subunits, in most cases, remains undetermined. Regarding glycosylation, the ability to bind or to inhibit phospholipases A2 was solely demonstrated for homologous recombinants with no carbohydrates from Python and Protobothrops inhibitors [11,12]. In addition, the fact that a second subtype of gamma inhibitor was isolated from a cDNA library from Protobothrops flavoviridis snake liver [13] made the homomeric character of the subclass II sbγPLIs questionable. The translated protein is similar to the subunit B of heteromeric sbγPLIs. In the present study, we focused on the native structure of CNF, the sbγPLI from C. d. terrificus snakes, in an effort to increase our understanding of oligomerization, role of glycosylation and subunit composition.

74 75

E

2

Please cite this article as: C.L. Fortes-Dias, et al., Insights on the structure of native CNF, an endogenous phospholipase A2 inhibitor from Crotalus durissus terrificus, the South American..., Biochim. Biophys. Acta (2014), http://dx.doi.org/10.1016/j.bbapap.2014.05.001

149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167

171 172 173

177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193

C.L. Fortes-Dias et al. / Biochimica et Biophysica Acta xxx (2014) xxx–xxx

216

2.7. Deglycosylation of CNF

217

229 230

Deglycosylation was performed by incubating 250 μg of CNF with 5000 U of PNGase F (704S, Bio-Rad Laboratories, Inc.) in 250 μl of the accompanying enzyme buffer. A control without enzyme was run under the same conditions. The deglycosylation was confirmed by a decrease in the intensity of the 24 kDa band (glycosylated monomer) and the concomitant enhancement of the 20 kDa band (non-glycosylated monomer), on SDS-PAGE. The reaction was let to proceed until sufficient for our purposes and, in some cases, it did not reach completion. The enzymatic reaction was interrupted by immediate application to a reversed-phase chromatography, without PNGase removal. Deglycosylated CNF was assayed for PLA2 inhibition, analyzed by reversed-phase chromatography, submitted to chemical cross-linking and incubated with Ctx to investigate any complex formation. Native CNF was run in parallel in every experiment.

231

2.8. PLA2 inhibition assay

232

240

Forty-five microliters of native or deglycosylated CNF at increasing concentrations (19.95 μg/ml to 3.16 mg/ml) was incubated with 5 μl of C. d. terrificus venom solution at 0.5 mg/ml for 30 min at 37 °C and PLA2 activity was assayed by the egg yolk clearing method [22]. The data were normalized as percentages of the positive (venom only) and negative (PBS with no venom) controls and fitted by nonlinear four-parameter logistic regression by the least squares method, using the Graph Prism 6.0 for Mac OS X (GraphPad software Inc., La Jolla, California).

241

2.9. Reversed-phase chromatography of native and deglycosylated CNF

242 243

The chromatography was performed in a μRPC C2/C18 ST 4.6/100 (Amersham Biosciences, Sweden) using 0.1% TFA in H2O (A) and 0.1% TFA in CH3CN (B) as mobile phases, according to the published protocols used in the purification of heteromeric sbγPLIs [7,23,24]. A control run was performed by loading PNGase at the same concentration used in the deglycosylation step (item 2.7). In every case, the elution gradient varied from 25% to 70% of B in 72 ml under 0.5 ml/min flow rate. The eluted samples were analyzed by SDS-PAGE, mass spectrometry and amino acid sequencing. The relative component percentages were calculated by integration of the areas under the corresponding peaks. When applicable, samples were blotted to a nitrocellulose membrane

219 220 221 222 223 224 225 226 227 228

233 234 235 236 237 238 239

244 245 246 247 248 249 250 251 252

Matrix-assisted laser desorption ionization time-of-flight mass spectrometry was performed on an AutoFlex III MALDI-TOF-TOF instrument (Bruker Daltonics, Bremen, Germany) controlled by the FlexControl 3.0 software. The samples were prepared by mixing 0.5 ml of the protein with 0.5 ml sinapinic acid matrix solution directly on the target plate and dried at room temperature. The average masses were recorded in positive linear mode with external calibration using the Protein Calibration Standard kit (Bruker Daltonics, Bremen, Germany), containing [M + H]+: Trypsinogen 23982; Protein A: 44613 and [M + H]2 +: Protein A: 22307. MS data were acquired in the 14000–100,000 m/z range and the spectra were subsequently processed with FlexAnalysis 3.3 (Bruker Daltonics, Bremen, Germany). Amino acid sequencing was performed by Edman degradation using a Shimadzu PPSQ-21A automated protein sequencer.

256 257

2.11. Complex formation analysis

270

Native or deglycosylated CNF were incubated with Ctx (1:1 molar ratio) for 1 h at 37 °C in 50 mM Tris–HCl pH 7.4 containing 50 mM NaCl and 20 mM CaCl2 (buffer A). Native CNF, deglycosylated CNF and Ctx preparations were run in parallel. The samples were loaded into a Protein Pak SW300 column (Waters, Millipore Co.) and eluted with buffer A under a flow rate of 0.5 ml/min. The resulting peaks were analyzed by reversed-phase chromatography on a μRPC C2/C18 ST 4.6/100 column (Amersham Biosciences, Sweden). Mobile phases were 0.1% TFA in H2O (A) and 0.1% TFA in CH3CN (B). Elution was performed with a gradient of 25 to 75% of B in 72 ml, under a flow rate of 0.5 ml/min.

271 272

2.12. Circular dichroism spectroscopy

281

Circular dichroism (CD) measurements were carried out in a J-815 spectropolarimeter (Jasco Inc.) equipped with a Peltier thermocontroller within a spectral range of 195–260 nm. The experiments were performed at 20 °C using 0.4 mg/ml (16.7 μM) of CNF in 10 mM ammonium formate pH 6.5 with 0.5 mm optical length, 100 nm/min scanning speed, 1 nm bandwidth and step resolution of 0.5 nm. Twenty spectra were collected, averaged and corrected for the buffer baseline contribution. The data were normalized to mean residue ellipticity Φ (deg·cm2·dmol−1·residue−1) and curve deconvolution was conducted using the Contin algorithm [26] available in the Dichroweb server [27].

282

2.13. Protein modeling and molecular dynamic simulations

292

The monomer U of the crystallographic structure of urokinase plasminogen activator from Homo sapiens (PDB ID: 2FD6) was selected as the best template for initial in silico CNF structural model, according to the data obtained from based-threading method program HHPred [28] (score: 212.63; e-value: 1.7e−30; identity: 31%) available at the MaxPlanck Institute for Developmental Biology server. The initial in silico model of the CNF structure was generated by the program MODELLER v.9.10 [29] using the selected template. Next, this initial model was submitted to molecular dynamic (MD) simulations using the GROMACS (Groningen Machine for Chemical Simulation) v.4.5.3 software [30] in the presence of explicit water molecules. Protonation states of charged groups were set according to pH 7.0. Counter ions were added to neutralize the system and the GROMOS 96 53a6 force field [31] was chosen for performing the MD simulation. The minimum distance between any atom of the protein and the box wall was 1.0 nm. An energy minimization (EM) using a steepest descent algorithm was performed to generate the starting configuration of the system. After this step, 200 ps of MD simulation with position restraints applied to the protein (PRMD)

293

D

218

255

E

212 213

T

210 211

C

208 209

E

206 207

R

204 205

R

202 203

N C O

200 201

U

199

2.10. Mass spectrometry and amino acid sequencing

F

214 215

Steady-state fluorescence measurements of the emission intensity and anisotropy of CNF were performed in a Hitachi F-700 spectrofluorimeter using a 2 mm optical path cuvette and a bandwidth of 2 nm. CNF samples were prepared in 20 mM ammonium formate pH 6.5 at concentrations of 1.7 mg/ml (72 μM), 2.5 mg/ml (100 μM) and 3.5 mg/ml (140 μM). Due to the absence of tryptophan residues and the presence of tyrosine residues in the CNF primary structure, the fluorescence emission spectra of tyrosines were measured from 290 nm to 500 nm (1200 nm/s) using a 280 nm excitation wavelength. Fluorescence anisotropy was obtained in the maximum intensity region (330 to 350 nm). The steady-state fluorescence anisotropy was measured using a set of polarizer filters. The steady-state anisotropy r is defined as (IVV − G.IVH) / (IVV + 2.G.IVH) where IVH is the intensity obtained from the vertically polarized excitation and horizontally polarized emission and IVV is the intensity obtained from the vertically polarized excitation and emission. The G factor accounts for the ratio between the sensitivities of detection system for vertically and horizontally polarized light [21]. This method provides information on molecular size and shape and local viscosities of a fluorophore's environment, as well as offering insight into changes in molecular sizes of polymers, proteins and other macromolecules.

O

195 196 197 198

and revealed for glycoprotein using periodic acid and Schiff's reagent 253 (PAS) [25]. 254

R O

2.6. Fluorescence spectroscopy

P

194

3

Please cite this article as: C.L. Fortes-Dias, et al., Insights on the structure of native CNF, an endogenous phospholipase A2 inhibitor from Crotalus durissus terrificus, the South American..., Biochim. Biophys. Acta (2014), http://dx.doi.org/10.1016/j.bbapap.2014.05.001

258 259 260 261 262 263 264 265 266 267 268 269

273 274 275 276 277 278 279 280

283 284 285 286 287 288 289 290 291

294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310

4

Table 1 Snake blood phospholipase A2 inhibitors from the gamma class (sbγPLIs) for which native molecular masses were estimated by gel filtration chromatography.

Viperidae

t1:5 t1:6

Name

Estimated molecular mass (kDa)

Gel filtration column

Composition and molecular mass of the subunits (kDa)

Reference

Bothrops jararacussu

BjussuMIP

160

Sephacryl S200

[43]

Ccrrophidion godmani (Formerly Bothrops godmani) Godman's viper Crotalus durissus terrificus South Amencan rattlesnake

CgMIP-I

110

Superdex 200

CNF

160

Superdex75

CICS

130

Superdex 200

PLI

100

Superose 12

LsPLI

100

Superose 12

Homopolymer 23.5 kDa Homopolymer 25 kDa Homopolymer 24 kDa Heteropolymer 25 kDa/23 kDa Heteropolymer (2:1) 25 kDa/20 kDa Heterotrlmer 25 kDa/20 kDa (2:1) Heterotrlmer 31 kDa/25 kDa (2:1) Heteropolymer 30 kDa/25 kDa Heteropolymer 30 kDa/25 kDa Heteropolymer 25 kDa/25 kDa Heteropolymer 30 kDa/29 kDa Homopolymer 6 × 23.5 kDa

t1:7 t1:8 t1:9

Elapidae

t1:10 t1:11 t1:12 t1:13

Colubridae

t1:14

Pythonidae

PLI

90

Sephadex G-200

NAI

110

Superose 12

NSI

110

Superose 12

EC PLI

120

Superdex 200

EQPL1

130

Sephadex G200

PIP

140

Superdex 200

P

t1:15

Gloydius brcvicaudus (Formerly Agktstrodon blomhoffi siniticus) Chinese mamushi Laticauda semifasciata Sea krait Naja naja kaouthia Thailand cobra Notechis ater Tiger snake Notcchis scutatus Common tiger snake Elaphe climacophora Japanese rat snake Elaphe quadrivirgata Japanese striated snake Python reticulatus

3. Results

317

3.1. Oligomerization of CNF: insights in its quaternary structure

318

The number of forming subunits remains undetermined for most polymeric sbγPLIs with the exception of the inhibitors from Laticauda semifasciata [24] and Naja naja kaouthia [23] elapid snakes, which have been identified as trimers (Table 1). In order to determine the number of subunits in native CNF, we combined several biochemical and biophysical approaches, namely, chemical cross-linking, gel filtration chromatography, fluorescence spectroscopy, dynamic light and small angle X-ray scatterings (DLS and SAXS, respectively). In the cross-linking experiments, we employed two different chemical agents, BS3 (Fig. 1A) and glutaraldeyde (data not shown), obtaining identical results. In the experiment using BSA instead of CNF, no extra

326 327 328

C

E

R

R

O

324 325

C

322 323

[24] [23] [12] [46] [48] [49] [41]

band was observed on SDS-PAGE after incubation of the control protein with up to 10 mM of BS3, hence discarding any non-specific reaction due to the cross-linker. SDS-PAGE analysis of the cross-linked samples of CNF revealed up to four protein bands with average molecular masses compatible with monomers, dimers, trimers and tetramers, in addition to the non-glycosylated monomer at much lower proportion. The molecular mass of the highest protein band in the gel was 94.7 ± 7.7 kDa, which is compatible with a tetramer. This result suggests that tetramers are the highest aggregation state identified in the crosslinked inhibitor. Faint bands just below the main oligomer bands identified on the right side of the gels may be observed (Fig. 1A). These bands may possibly correspond to oligomers formed by glycosylated/ non-glycosylated or non-glycosylated/non-glycosylated molecules but our data do not support any conclusion about the composition of those extra bands. The cross-linking with BS3 was repeated with different concentrations of native CNF (Fig. 2). Even after a five-fold increase in the inhibitor concentration, tetramers were confirmed as the highest aggregation state of CNF on SDS-PAGE. In the gel filtration chromatography, we used four different columns including those described in the literature for other sbγPLIs (Table 1). Each column was calibrated with reference protein samples loaded

N

321

[7,40]

U

319 320

[10]

D

316

[9]

T

314 315

was executed in order to relax the system gently. Then, 50 ns of unrestrained MD simulation were calculated to evaluate the stability of the structure. All the MD simulations were carried out in a periodic truncated dodecahedron box under constant temperature (298 K) and pressure (1.0 bar) maintained by coupling to an external heat and an isotropic.

312 313

[44]

E

311

F

t1:4

Snake species

O

t1:3

R O

t1:1 t1:2 Q2

C.L. Fortes-Dias et al. / Biochimica et Biophysica Acta xxx (2014) xxx–xxx

Fig. 1. Non-reducing SDS-PAGE on 8–25% gradient Phast® gels (GE HealthCare) of native (A) or deglycosylated (B) CNF after cross-linking with BS3 at the concentrations marked on top. Legend: (n) glycosylated CNF monomer; (ng-) nonglycosylated or enzymatically deglycosylated CNF monomer. Dimers, trimers and tetramers are identified by the respective number of subunits times (n), followed by (g-) if non-glycosylated/deglycosylated. Molecular mass markers (LMW protein mixture, GE HealthCare) are indicated on the left side. The gels were stained with Coomassie Blue.

Please cite this article as: C.L. Fortes-Dias, et al., Insights on the structure of native CNF, an endogenous phospholipase A2 inhibitor from Crotalus durissus terrificus, the South American..., Biochim. Biophys. Acta (2014), http://dx.doi.org/10.1016/j.bbapap.2014.05.001

329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349

C.L. Fortes-Dias et al. / Biochimica et Biophysica Acta xxx (2014) xxx–xxx

5

O

F

Fig. 3. Estimated molecular masses of CNF by gel filtration chromatography on different columns: ◆ Superose 12 HR10/30; ♦ Superdex 200 HR10/30; ● Protein-Pak 300 SW, ■ Superose 6 HR10/30. Horseradish peroxidase (HRPO, glycosylated protein) and crotoxin (Ctx, non-glycosylated protein) were used as controls.

350

spherical particles [32], thus indicating that CNF is of spherical shape. In fact, this idea is corroborated by the Rh/Rg ratio (0.789) that is very close to that obtained for perfect sphere particles (0.775) [33]. Finally, the estimation of molecular mass from the SAXS data using SAXS MoW web tool indicates a molecule with 118 kDa, which is also compatible with a predominantly tetrameric structure of CNF. Dynamic light scattering (DLS) measurements showed that CNF presents a monodisperse unimodal size distribution (N 99% of mass in samples measured with b 15% of polidispersivity) and an average hydrodynamic radius (Rh) of 4.4 ± 0.1 nm. This value corresponds to an average molecular mass of 107.7 ± 6.5 kDa, which is roughly compatible with a tetramer (Table 3). This data supports the results obtained by cross-linking assays and SAXS, where we observed that, despite limited aggregation, the tetramer appears to be the prevailing quaternary structure of CNF. Fluorescence anisotropy experiments revealed a direct relationship between the increase in CNF concentration and the resulting higher anisotropy values (Fig. 5A): 0.092 for 1.7 mg/ml, 0.105 for 2.5 mg/ml and 0.118 for 3.5 mg/ml of CNF. This finding reveals a direct relationship between molecular size and CNF concentration, suggesting the oligomerization of CNF in solution. Furthermore, fluorescence emission intensity of tyrosines (Fig. 5B) shows an inverse relationship with CNF concentration, whereby the fluorescent signal decreases with the increase of CNF concentration. This reduction of fluorescent intensity can be explained by the oligomerization. In the oligomers, the Tyr fluorescence signal could be reduced by an increase in the non-radioactive processes or by Föster resonance energy homotransfer to other Tyr residues.

378 379

3.2. Secondary and tertiary CNF structures

406

R O

Fig. 2. Non-reducing SDS-PAGE on 8–25% gradient Phast® gels (GE HealthCare) of native CNF after cross-linking with 10 mM of BS3. Lane 1 (control): 6.7 mg/ml of CNF incubated under the same conditions with no BS3. CNF concentrations in the cross-linking reactions were 6.7 mg/ml (lane 2), 16.7 mg/ml (lane 3) and 33.3 mg/ml (lane 4). Legend: (n) CNF monomer, (ng-) non-glycosylated CNF monomer, (2n) CNF dimer, (3n) CNF trimer, (4n) CNF tetramer. The gel was stained with Coomassie Blue.

t2:1 t2:2 t2:3 Q3

Table 2 Parameters of the calibration curves obtained for the reference proteins on the different gel filtration columns. The regression curves were obtained by linear regression using the least square method.

365 366 367 368 369 370 371 372 373 374 375

t2:4

D

E

T

C

363 364

E

361 362

R

359 360

R

357 358

N C O

355 356

Gel filtration column

Matrix

Reference proteins

t2:5 t2:6 t2:7 t2:8 t2:9 t2:10 t2:11 t2:12 t2:13

Protein Pak SW 300

Silica

Superose 6 HR 10/30

Cross-linked agarose

Superose 12 HR 10/30

Cross-linked agarose

Superdex 200 HR10/30

Dextran-grafted agarose

382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405

Circular dichroism (CD) measurements were used to investigate the 407 secondary structure of CNF (Fig. 6A). CD spectra showed a minimum 408 and a maximum in ellipticity (ϕ) at 217 nm and 197 nm, respectively, 409

U

353 354

P

376 377

individually and as a mixture and the regression curves obtained in both cases were fully compatible as indicated by comparable slopes (Table 2). The goodness of fit of the experimental data to linear regression curves was also ascertained by correlation coefficients (R2) consistently very close to 1.0 (Table 2). Taking into account the entire set of columns, average molecular masses of 18.2 ± 2.5 kDa and 42.0 ± 2.9 kDa were estimated for the control proteins (Ctx and HRPO) with variation coefficients of 13.7% and 6.9%, respectively. In contrast, the estimated molecular masses of native CNF varied within a wider range depending on the column: 86.0 kDa (Superose 6), 114.4 kDa (Protein Pak SW300), 128.3 kDa (Superdex 200) and 179.6 kDa (Superose 12). The average molecular mass calculated for native CNF was 127.1 ± 39.2 kDa with a variation coefficient of 30.8%. The estimated molecular masses of CNF, Ctx and HRPO are depicted in Fig. 3. In the case of CNF, the gel filtration approach alone was considered inappropriate for our purpose. Therefore, the data obtained here were analyzed after association with other biochemical and biophysical methods. Small angle X-ray scattering (SAXS) experiments presented similar scattering curves for CNF at 1 and 5 mg/ml and subsequent experiments were performed at 5 mg/ml. However, due to beam stop shadow as well as the observation of some sample aggregation, the twenty initial points (q values from 0.01 to 0.04 Å−1) were disconsidered in the analysis of the scattering curve. The Guinier plot gave a radius of gyration (Rg) of 36.48 ± 0.38 Å that is well correlated with the Rg obtained by the p(r) analysis (34.75 ± 0.03 Å) (Fig. 4). As two minima could be observed in the CNF scattering curve, this indicated that the p(r) function has a maximum at approximately Dmax/2. These characteristics are typical of SAXS data from

351 352

380 381

Mixture Individual Mixture Individual Mixture Individual Mixture Individual

Regression parameters Slope (−1)

R2

2.47 2.40 2.28 2.14 2.89 2.91 5.08 5.08

0.9687 0.9709 0.9926 0.9689 0.9660 0.9667 0.9596 0.9616

± ± ± ± ± ± ± ±

0.22 0.19 0.11 0.16 0.22 0.22 0.60 0.59

Please cite this article as: C.L. Fortes-Dias, et al., Insights on the structure of native CNF, an endogenous phospholipase A2 inhibitor from Crotalus durissus terrificus, the South American..., Biochim. Biophys. Acta (2014), http://dx.doi.org/10.1016/j.bbapap.2014.05.001

C.L. Fortes-Dias et al. / Biochimica et Biophysica Acta xxx (2014) xxx–xxx

O

F

6

Q19 Fig. 4. Small angle X-ray scattering curves of CNF (A). Experimental scattering curve (open squares with error bars) and fitted scattering curve (straight line) using GNOM [19]. The inset

Temperature

Hydrodynamic

Estimated molecular

Polidispersivity

Mass

t3:5

°C

Radius (nm)

Mass (kDa)

%

%

t3:6 t3:7 t3:8

4 10 18

4.3 4.4 4.5

101 108 114

6 14.4 16.2

99.5 99.8 99.7

421 422 423 424 425 426

436

Subunit homogeneity in a standard preparation of native CNF was investigated by reversed-phase chromatography, under conditions used by other authors for heteromeric sbγPLIs. Two main peaks, ‘A’ and ‘B’, corresponding to 72% and 25% of the loaded sample based on the absorbance at 280 nm, respectively, were obtained (Fig. 7). On SDS-PAGE, peak ‘A’ displayed a single band with an apparent molecular mass of 24 kDa (Fig. 8). This band was positive for carbohydrates (data not shown) and displayed an m/z ratio of 22,201 by mass spectrometry (Fig. 8). Higher m/z ratios compatible with dimers, trimers and tetramers were also observed. The amino-terminal sequence determined for peak ‘A’ (RSCDFCHNIGKDCDGYEEEC) was identical to that previously published for the CNF monomer [9,10], demonstrating that this peak comprises the glycosylated subunit routinely observed as the major component in the preparations. Similar sequences have been obtained for subunits A of heteromeric sbγPLIs. The peak ‘B’ was poorly resolved

437 438

T

C

E

R

419 420

R

417 418

O

415 416

3.3. Subunits homogeneity

C

413 414

indicating predominance of beta-sheets. In fact, spectral deconvolution indicates that CNF consist of 44.4% of beta-sheets and 50.7% of loops, in addition to disordered elements. In order to gain further insight into tertiary structural features of CNF, an in silico three-dimensional model was constructed using threading modeling technique further enhanced by 50 ns molecular dynamic (MD) simulation (Fig. 6B). The proposed model presents a high stereochemical adequacy as expressed by a high percentage of residues (97.7%) in the allowed regions of the Ramachandran plot as well as an overall good quality (Z-score equal to −4.2). The analysis of CNF in silico structure after MD simulation confirmed that beta-sheets and loop/disordered elements form this structure, in line with the previously shown CD data. The arrangement of these secondary structure elements forms two three-finger domains, well ordered in the tertiary structure of CNF, which are typical of the sbγPLIs. Fluorescent emission intensity of tyrosines (Fig. 5B) showed an inverse relationship with CNF concentration, suggesting the CNF

N

411 412

U

410

427 428

P

t3:4

oligomerization towards a region comprising tyrosine residues (Section 3.1). Inspection of the solvent accessible area (SAS) of the tyrosinyl residues by the AREAIMOL software [34] available in the CCP4 program suite [35] showed that six of those residues have good accessible area to solvent, in particular those in positions 16, 113, 132 and 166 (Table 4). The tyrosinyl residues were located in the in silico structural model of CNF (Fig. 6C). These findings further confirm that a region that comprises some of those tyrosines may form the interface between the CNF monomers in the tetramer.

D

Table 3 Dynamic light scattering for native CNF at 2.5 mg/ml in 20 mM ammonium formate pH 6.5. The results are expressed by the mean values of 100 measurements.

E

t3:1 t3:2 t3:3

R O

shows the Guinier plot. (B) Normalized distance distribution function for CNF.

Fig. 5. Steady-state fluorescence spectroscopy measurements of Tyr residues from CNF. (A) Anisotropy r (B) emission intensity I in different concentrations of CNF as indicated on the left side of the curves.

Please cite this article as: C.L. Fortes-Dias, et al., Insights on the structure of native CNF, an endogenous phospholipase A2 inhibitor from Crotalus durissus terrificus, the South American..., Biochim. Biophys. Acta (2014), http://dx.doi.org/10.1016/j.bbapap.2014.05.001

429 430 431 432 433 434 435

439 440 441 442 443 444 Q9 445 446 447 448 449 450 451

C.L. Fortes-Dias et al. / Biochimica et Biophysica Acta xxx (2014) xxx–xxx t4:1 t4:2 t4:3 t4:4 Q4

ASA (Å2)

t4:6 t4:7 t4:8 t4:9 t4:10 t4:11 t4:12 t4:13

Tyr16 Tyr68 Tyr97 Tyr98 Tyr113 Tyr132 Tyr141 Tyr166

131.7 2.2 42.2 39.3 75.8 127.2 27.8 207.7

F

under our experimental conditions, even after rechromatography attempts under isocratic gradient elution. Therefore, peak ‘B’ was separated into ‘B1’ and ‘B2’ portions (Fig. 7) and analyzed as such. ‘B1’

U

N C O

R

R

E

C

T

E

D

P

R O

454

Residue

O

452 453

accounted for about 67% of peak ‘B’ and 16% of the CNF sample. ‘B2’ corresponded to about 33% of peak B and 8% of the entire CNF sample. On SDS-PAGE, peak ‘B’ was composed of two bands of 24 and 20 kDa, respectively, with faint carbohydrate positivity in the first one (data not shown). Mass spectrometry confirmed the presence of two components under peak ‘B’ with m/z ratios of 22,252 and 20,173, respectively (Fig. 8). While amino acid sequencing attempts with ‘B2’ resulted in an unresolved mixture of residues, a main amino-terminal sequence IECEVCMKPGDRCNGSMMT was obtained for ‘B1’. This sequence is comparable to the amino-terminal of the second subunit of heteromeric sbγPLIs (Table 5). Therefore, in addition to the 24 kDa monomer, our CNF preparation contained about 16% of a second component with estimated molecular mass of 20 kDa in SDS-PAGE and a m/z ratio of 20,173 in the mass spectrum (peak ‘B’ in Fig. 8). The small peak ‘X’ between A and B in the native CNF (Fig. 7) was submitted to amino acid sequencing. The first five amino-terminal

Table 4 Accessible solvent area (ASA) of tyrosinyl residues (Å2) from in silico CNF structural model calculated with AREAIMOL (Saff & Kuijlaars [34]) software.

t4:5

7

Q20 Fig. 6. Secondary and tertiary structure of CNF. (A) Circular dichroism (CD) spectra for CNF; (B) cartoon representation of in silico CNF structural model obtained after 50 ns molecular dynamic (MD) simulation. The inset shows the root mean square deviation (r.m.s.d.) of CNF model during the MD simulation. Generated by GROMACS [31]. The stabilization of the structure occurred in approximately 10 ns; (C) localization of tyrosine residues (showed in sticks) at in silico CNF structural model. The tyrosines with higher solvent accessible areas (see Table 3) are represented in yellow and the remaining in cyan.

Please cite this article as: C.L. Fortes-Dias, et al., Insights on the structure of native CNF, an endogenous phospholipase A2 inhibitor from Crotalus durissus terrificus, the South American..., Biochim. Biophys. Acta (2014), http://dx.doi.org/10.1016/j.bbapap.2014.05.001

455 456 457 458 459 460 Q10 Q11 461 462 463 464 465 466 467 Q12 468 469 470

C.L. Fortes-Dias et al. / Biochimica et Biophysica Acta xxx (2014) xxx–xxx

O

F

8

3.4. Role of carbohydrates

477

The role played by the carbohydrate moiety in the functionality of CNF was evaluated after enzymatic deglycosylation of the native protein. Although the PLA2 inhibiting activity decreased after carbohydrate 480 removal, as observed by a displacement of the inhibition curve to the 481 right, the deglycosylated molecule retained its functionality (Fig. 9). 482 Moreover, the deglycosylated CNF was able to form a stable complex 483 with CB, as previously demonstrated for the native inhibitor (repeated 484 here as a control). After pre-incubation of the native or deglycosylated 485 CNF with Ctx, two protein peaks were separated by gel filtration chro486 matography (Fig. 10). In both cases, RPC analysis showed that the first 487 peaks contained the native or the deglycosylated inhibitor as well as 488 CB, whereas the unbound CB and CA from Ctx were eluted under the 489 second peak (Fig. 11). As observed before with other matrices (Fortes490 Q13 Dias et al., 1994), gel filtration chromatography failed to separate the 491 CNF from the CNF/CB complex. They are eluted together in the first 492 peak (P1 or P′1), independently of the presence of the carbohydrate 493 moiety. In the RPC of P′1, an extra peak “y” was eluted at volumes 494 between “A” and “a”. This peak is exclusively observed after incubation 495 of deglycosylated CNF with Ctx, but its nature and composition will 496 deserve further investigation. 497 In order to evaluate the interference of the carbohydrate moiety on 498 the oligomerization of CNF, the deglycosylated molecule was submitted 499 to chemical cross-linking. The carbohydrate removal did not impair the 500 formation of monomers, dimers, trimers and tetramers (Fig. 1B). This 501 suggests that oligomerization is independent on the presence of 502 carbohydrate, and occurs, similarly, in the native and deglycosylated 503 inhibitor. 504 Reversed-phase fractionation of deglycosylated CNF gave an elution 505 profile comparable to that obtained for the native molecule, with the ex506 ception of the displacement of the entire profile towards higher reten507 tion times (Fig. 7). In the elution profile of deglycosylated CNF, peak ‘a’ 508 corresponded to 67% and peak ‘b’ to 25% of the loaded sample, respec509 tively. Minor components contributed to the absorbing material by 510 about 8%. On SDS-PAGE, peak ‘a’ produced a single 20 kDa band with a 511 m/z ratio of 20,008 (Fig. 8). Additional m/z ratios of 40,158, 60,332 512 and 80,521 were also observed, thus confirming the oligomerization 513 of deglycosylated CNF previously observed after cross-linking. The first

U

N

C

O

R

R

E

C

T

478 479

P

476

473 474

D

475

residues (RSCDF) were the same of the glycosylated CNF monomer eluted ender peak ‘A’. Moreover, the time of elution of peak ‘x’ is the same of peak ‘a’ in the deglycosylated CNF, as it will be presented in the next section. Therefore, we concluded that peak ‘X’ corresponds to the non-glycosylated monomer in the native preparation of CNF.

E

471 472

R O

Fig. 7. Reversed phase (μRPC C2/C18 ST 4.6/100) of CNF before (continuous line) and after (dotted line) enzymatic deglycosylation. Absorbance (mAU) was read at 280 nm. Inset: Same chromatogram with absorbance readings (mAU) at 214 nm. Mobile phases: 0.1% TFA in H2O (A) and 0.1% TFA in CH3CN (B). Protein peaks are identified by capital letters (before deglycosylation) or small letters (after deglycosylation), respectively.

Fig. 8. Non-reducing SDS-PAGE on 8–25% gradient Phast® gels (GE HealthCare) (top) and mass spectrometry (bottom) analysis of native and enzymatically deglycosylated CNF after fractionation by reversed-phase chromatography (Fig. 7). Molecular mass markers (LMW protein mixture, GE HealthCare) are marked on the right side of the gels. Sample identification is the same adopted in the reversed-phase chromatograms (Fig. 7) except that B1/B2 were loaded together as B. m/z ratios N25,000, possibly corresponding to oligomers, are indicated by *.

amino-terminal residues of peak ‘a’ (RSCDFCHNIG KDCDGYEEEC) were identical to those determined for peak ‘A’ in native CNF. Therefore, peak ‘a’ corresponds to the non-glycosylated monomer that is usually observed in CNF preparations on SDS-PAGE. While the ratio of the glycosylated (24 kDa) and non-glycosylated (20 kDa) monomers varies according to the preparation, the latter is always present at much lower amounts. Attempts to resolve the amino acid mixture under peak ‘b’ were unfruitful; however, it is possible that its contents are similar to those present under peak ‘B’ of native CNF. After SDS-PAGE (Fig. 8), carbohydrates were not detected in either ‘a’ or ‘b’ samples (data not shown). A control run was performed with PNGase, at the same amounts used for deglycosylation of CNF. No extra peak was noted that could interfere in the chromatogram of deglycosylated CNF.

Please cite this article as: C.L. Fortes-Dias, et al., Insights on the structure of native CNF, an endogenous phospholipase A2 inhibitor from Crotalus durissus terrificus, the South American..., Biochim. Biophys. Acta (2014), http://dx.doi.org/10.1016/j.bbapap.2014.05.001

514 515 516 517 518 519 520 521 522 523 524 525 526 527

C.L. Fortes-Dias et al. / Biochimica et Biophysica Acta xxx (2014) xxx–xxx t5:1 t5:2 t5:3

9

Table 5 Partial alignment of the secondary component of CNF (identified as C. d. terrificus) with the B subunits/subtypes from sbγPLIS described so far.

O

F

Fig. 10. Protein-Pak SW300 chromatogram of native (continuous line) and enzymatically deglycosylated (dotted line) CNF, after pre-incubation with crotoxin. Absorbance (mAU) was read at 280 nm. The peaks obtained (P1 and P2 for native CNF and P1′ and P2′ for deglycosylated CNF) were analyzed by reversed phase chromatography (Fig. 11).

542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557

R O

C

540 541

E

538 539

R

536 537

R

534 535

N C O

532 533

U

530 531

P

The determination of the quaternary structure of native sbγPLIs is a crucial step for a better understanding of the functionality of those complex molecules with high biotechnological potential. Although the molecular masses of the individual subunits have been estimated for the great majority of sbγPLIs by classical gel filtration (Table 1), the number of forming subunits in the native molecules remained to be determined. In a first attempt to better understand its quaternary structure, cross-linked CNF was analyzed by SDS-PAGE. Protein bands with masses compatible with oligomers composed of one to four subunits were obtained. The highest molecular mass (94.7 ± 7.7 kDa) is compatible with the presence of tetramers as the highest quaternary structures in the mixture. This value is lower than those previously estimated by gel filtration chromatography of native CNF: 160 kDa on Superdex G75 [9] and 130 kDa on Superdex 200 column [10]. Based on their data, authors of these studies suggested that CNF oligomer consist of six to eight monomers. In order to clarify the discrepancy between the molecular masses previously obtained for CNF (six to eight monomers) and to establish the number of forming subunits in other sbγPLIs, we performed the gel filtration using four different columns (all displaying optimum separation ranges, adequate for our purpose). Superdex G75 was performed but it is not reported here, since its optimal separation range (3 to 70 kDa) is far below the molecular mass of about 95 kDa estimated for CNF by cross-linking. Although we obtained adequate calibration curves for every column employed (Table 2) and took every precaution to minimize sample and temperature variations during the gel filtration experiments, the molecular mass estimated for CNF varied from 86 to about 180 kDa. However, it has long been known that some glycoproteins do not conform to the

D

529

expected globular behavior in gel filtration [36,37]. This may be due to the higher hydration of carbohydrate chains, as compared to polypeptide chains, as well as far from globular conformations and/or interactions with the column matrix [38,39]. Since we used the same CNF preparation in our experiments and the predominant CNF tetramer was later shown to be globular, the most plausible explanation for the observed variability in the molecular mass estimations of the inhibitor could be attributed to differences in the composition and average pore sizes of the column matrices (Table 6). In the particular case of CNF, the estimate by the silica-based matrix gave more compatible results with the other methods employed herein. The HRPO used as control

E

4. Discussion

T

528

Fig. 9. PLA2 inhibition curves for native (●) and deglycosylated (○) CNF fitted as variable slope (four-parameter logistic) curves. [CNF] concentration is expressed in μg/ml. C. d. terrificus snake venom was used as the PLA2 source. The IC50 values for native and deglycosylated CNF were 505 μg/ml and 726 μg/ml with 95% confidence intervals of (411–620) and (637.9–825.4), respectively. Bars represent standard deviations of triplicates.

Fig. 11. Reversed-phase (μRPC C2/C18 ST 4.6/100) chromatograms of the peaks obtained by gel filtration chromatography of native and enzymatically deglycosylated CNF pre-incubated with crotoxin (Fig. 10). The samples are identified on the right side of the chromatograms. Control profiles: Native and fully deglycosylated CNF as in Fig. 7; crotoxin and its forming subunits, CA and CB. CB is present as a mixture of three main isoforms in the crotoxin preparation used in the experiments.

Please cite this article as: C.L. Fortes-Dias, et al., Insights on the structure of native CNF, an endogenous phospholipase A2 inhibitor from Crotalus durissus terrificus, the South American..., Biochim. Biophys. Acta (2014), http://dx.doi.org/10.1016/j.bbapap.2014.05.001

558 559 560 561 562 563 564 565 566 567 568

Table 6 Summary of the estimated molecular masses of CNF using biochemical and biophysical techniques. The standard deviations were calculated whenever technically possible. Abbreviations: unglycosylated monomer (ng-); glycosylated molecules: monomer (n), dimer (2n), trimer (3n) and tetramer (4n). Estimated molecular mass (kDa)

583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614

671

We acknowledge Miss Ana do Carmo Valentim for excellent technical help and Dr. Amando Siuiti Ito for fruitful discussions concerning fluorescence spectroscopy. We are thankful to the Fundação de Amparo à Pesquisa de Minas Gerais (FAPEMIG), Fundação de Amparo à Pesquisa de São Paulo (FAPESP), INCTTox (CNPq/FAPESP) and CAPES (Toxinology 063/2011) for financial support.

672 673

O

F

Acknowledgements

T

C

581 582

E

579 580

R

577 578

R

575 576

showed an estimated molecular mass close to the accepted 40 kDa [36] in every column. Hence, the carbohydrate moiety did not interfere in the gel filtration behavior of HRPO. Thus, it is also possible that observed variability of CNF mass could be due to the eventual formation of CNF aggregates. The SAXS results suggest that their occurrence and this possibility cannot be discarded in the gel filtration samples. Anyway, gel filtration chromatography does not seem to be an appropriate technique to estimate the molecular mass of CNF, and probably of other sbγPLIs. By combining SAXS and DLS techniques, we showed that the CNF tetramer is a globular protein and that tyrosines may play an important role in oligomerization. A closer analysis of the in silico CNF threedimensional model indicated the presence of, at least, four tyrosines with high accessible solvent areas (Fig. 6, Table 4). In the spectroscopy fluorescence, an increase of fluorescence intensity at higher concentration of CNF could be expected, due to a decrease in non-radiative processes of deexcitation. However, the opposite effect was observed (Fig. 5B), and the decrease of tyrosine fluorescence emission can be attributed to an increase in the homotransfer process due to the oligomerization of CNF. The homomeric character of some sbγPLIs, and CNF in particular, has been previously questioned [24,40]. Although two subunits of different molecular masses (25 and 23 kDa) have been detected once in a preparation of CICS, a homologous inhibitor to CNF, both had identical aminotermini sequences [10]. Later, this observation was explained by the simultaneous presence of glycosylated and non-glycosylated subunits with electrophoretic migration compatible with the observed masses [15]. In our work, additional components, such as peak ‘B’ observed herein, have not been previously detected during the purification of functional CNF. One possible reason is the co-migration of peak ‘B’ and the non-glycosylated CNF monomer under SDS-PAGE. On the other hand, the amino terminal sequence of peak ‘B’ was previously detected in a preparation of CICS. The component comprised less than 10% of the total protein sample and was considered a contaminant peptide at that time [10]. It is worth noting that two homomeric sbγPLIs were shown to bind to PLA2 [11,41,42] Functional homomeric inhibitors have been also purified by affinity chromatography with immobilized PLA2 [43–45]. Thus, the importance of that second component in the functionality of CNF and possibly in other homomeric sbγPLIs, if any, remains unclear. In attempting to establish any role played by this secondary component, the unfavorable conditions of the reversed-phase chromatography needed for the separation of sbγPLIs subunits pose the main challenge. Such a step invariably leads to inactivation, thus hindering any functional assay of the isolated components. Even in the absence of any functional assay of the isolated subunits A and B, two well-resolved peaks with area ratios of 2:1, respectively,

O

573 574

19.4 25.1 47.7 69.8 94.7 107.7 118.0

C

571 572

114.4 128.3 86.0 179.6

N

569 570

Gel filtration chromatography ProteinPAK SW300 Superdex 200 Superose 6 Superose 12 Chemical cross-linking/SDS-PAGE ngn 2n 3n 4n Dynamic light scattering (DLS) Small angle X-ray scattering (SAXS)

U

t6:7 t6:8 t6:9 t6:10 t6:11 t6:12 t6:13 t6:14 t6:15 t6:16 t6:17 t6:18 t6:19

R O

t6:6

615 616

P

Analytical method

were consistently demonstrated for the heteromeric sbγPLIs [7,12,23, 24,46]. This ratio is considerably higher compared to the 2:0.3 observed herein for the CNF monomer and component 2, respectively. A structural, rather than functional, role was previously suggested for the B subunit of heteromeric inhibitors from elapid snakes [46]. In the hypothetical model envisaged by these authors, the A subunit would interact with the PLA2, whereas the B subunit would maintain the correct threedimensional structure of A subunit(s). Recently, a cDNA encoding a secondary component was identified in the homomeric sbγPLIs from two species of Protobothrops snakes. The new transcribed molecules were considered a novel subtype of sbγPLIs belonging to a distinct family from that of the previously identified single monomer [13,47]. The contribution of that second subtype of inhibitor to the functionality of the Protobothrops sbγPLI is, however, presently unclear. At least one subunit in sbγPLIs is consistently glycosylated in snake species belonging to Elapidae, Colubridae and Pythonidae families. Homologues of the homomeric sbγPLI from Python reticulatus (PIP) are capable of inhibiting a broad spectrum of toxic svPLA2 as well as PLA2-induced mouse edema [41]. In the heteromeric inhibitors from Notechis ater (NAI) and L. semifasciata (LsPLI) glycosylation is not essential for their PLA2 inhibiting activity, since deglycosylated molecules were as active as native ones [12,46]. According to our data, the carbohydrate moiety of CNF is not essential, but increases the inhibition capacity of CNF in vitro. Since the native inhibitor is a mixture of glycosylated and non-glycosylated molecules, the role played by the latter, although present in lesser amounts, in the functionality of CNF requires further study. Putative glycosylation sites have been identified in the encoding cDNAs for the second subtype of sbγPLI from Protobothrops snakes mentioned before [13,47]. In our work, no carbohydrate was detected in peak ‘B’, which could be due to insufficient sensitivity of the techniques employed. Deglycosylation does not impair the oligomerization of CNF, abolish its functionality or impair the formation of a stable CNF-CB complex. Two other homomeric inhibitors from Python and Protobothrops snakes were shown to retain their ability to bind or to inhibit phospholipases A2, when homologous recombinants with no carbohydrate were tested [11,12]. This property seems advantageous for future, larger scale, production of sbγPLI for biotechnological purposes. As reported before (Fortes-Dias et al., 1994), the separation of CNF alone from the CNF/CB complex, was not achieved by gel filtration chromatography. We are presently investigating the mechanism of complex formation between CNF and CB, mainly concerning the stoichiometry of the reaction, and we hope to obtain a plausible explanation to that behavior. In the present study, using a combination of biochemical and biophysical techniques, we demonstrated that CNF in solution occurs as a mixture of glycosylated and deglycosylated oligomers. Tetramers appear to be the predominant oligomerization state of this sbγPLI with a molecular mass of about 100 kDa. The oligomerization may occur through tyrosine-rich regions. While carbohydrate moiety is not essential, it contributes favorably to PLA2 inhibition. Furthermore, we identified a second component in CNF, which comprises no more than 16% of the purified protein. This component is similar to the B subunits present in the heteromeric sbγPLIs. Whether this minor component plays any role in the structure or functionality of CNF remains to be determined.

D

t6:1 t6:2 t6:3 t6:4 t6:5 Q5

C.L. Fortes-Dias et al. / Biochimica et Biophysica Acta xxx (2014) xxx–xxx

E

10

Please cite this article as: C.L. Fortes-Dias, et al., Insights on the structure of native CNF, an endogenous phospholipase A2 inhibitor from Crotalus durissus terrificus, the South American..., Biochim. Biophys. Acta (2014), http://dx.doi.org/10.1016/j.bbapap.2014.05.001

617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 Q14 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 Q15 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670

674 Q16 Q17 675 676 677 Q18

C.L. Fortes-Dias et al. / Biochimica et Biophysica Acta xxx (2014) xxx–xxx

C

E

R

R

N C O

U

817

F

O

R O

P

[1] J.E. Burke, E.A. Dennis, Phospholipase A2 biochemistry, Cardiovasc. Drugs Ther. 23 (2009) 49–59. [2] O. Vital-Brazil, B.J. Excell, Action of crotoxin and contractin from the venom of Crotalus durissus terrificus (South American rattlesnake) on the frog neuromuscular junction, J. Physiol. 212 (1970) 34–35. [3] G. Faure, C. Bon, Crotoxin, a phospholipase A2 neurotoxin from the South American rattlesnake Crotalus durissus terrificus: purification of several isoforms and comparison of their molecular structure and of their biological activity, Biochemistry 27 (1988) 730–738. [4] C. Bon, C. Bouchier, V. Choumet, G. Faure, M.S. Jiang, M.P. Lambezat, F. Radvanyi, B. Saliou, Crotoxin, half-century of investigations on a phospholipase A2 neurotoxin, Acta Physiol. Pharmacol. Latinoam. 39 (1989) 439–448. [5] S.C. Sampaio, S. Hyslop, M.R. Fontes, J. Prado-Franceschi, V.O. Zambelli, A.J. Magro, P. Brigatte, V.P. Gutierrez, Y. Cury, Crotoxin: novel activities for a classic betaneurotoxin, Toxicon 55 (2010) 1045–1060. [6] C.L. Fortes-Dias, M.I. Estevão-Costa, Snakes as a source of phospholipase A2 inhibitors with biotechnological potential, in: M.E. Lima (Ed.), Animal Toxins: State of the Art Perspectives in Health and Biotechnology, Belo Horizonte, 2009, pp. 607–620. [7] N. Ohkura, H. Okuhara, S. Inoue, K. Ikeda, K. Hayashi, Purification and characterization of three distinct types of phospholipase A2 inhibitors from the blood plasma of the Chinese mamushi, Agkistrodon blomhoffii siniticus, Biochem. J. 325 (1997) 527–531. [8] S. Lizano, G. Domont, J. Perales, Natural phospholipase A2 myotoxin inhibitor proteins from snakes, mammals and plants, Toxicon 42 (2003) 963–977. [9] C.L. Fortes-Dias, Y. Lin, J. Ewell, C.R. Diniz, T.-H. Liu, A phospholipase A2 inhibitor from the plasma of the South American rattlesnake (Crotalus durissus terrificus), protein structure, genomic structure and mechanism of action, J. Biol. Chem. 269 (1994) 15646–15651. [10] J. Perales, C. Villela, G.B. Domont, V. Choumet, B. Saliou, M. Moussatché, C. Bon, G. Faure, Molecular structure and mechanism of action of the crotoxin inhibitor from Crotalus durissus terrificus serum, Eur. J. Biochem. 227 (1995) 19–26. [11] I. Nobuhisa, T. Chiwata, Y. Fukumaki, S. Hattori, Y. Shimohigashi, M. Ohno, Structural elements of Trimeresurus flavoviridis serum inhibitors for recognition of its venom phospholipase A2 isozymes, FEBS Lett. 429 (1998) 385–389. [12] P.G. Hains, K.L. Sung, A. Tseng, K.W. Broady, Functional characteristics of a phospholipase A(2) inhibitor from Notechis ater serum, J. Biol. Chem. 275 (2000) 983–991. [13] S. So, T. Chijiwa, N. Ikeda, I. Nobuhisa, N. Oda-Ueda, S. Hattori, M. Ohno, Identification of the B subtype of gamma-phospholipase A2 inhibitor from Protobothrops flavoviridis serum and molecular evolution of snake serum phospholipase A2 inhibitors, J. Mol. Evol. 6 (2008) 298–307. [14] C.L. Fortes-Dias, B.C.B. Fonseca, E. Kochva, C.R. Diniz, Purification and properties of an antivenom factor from the plasma of the South American rattlesnake (Crotalus durissus terrificus), Toxicon 29 (1991) 997–1008. [15] R.M.M. Santos, L.C. Oliveira, M.I. Estevão-Costa, M.E. de Lima, M.M. Santoro, C.L. Fortes-Dias, Inhibition of crotoxin binding to synaptosomes by a receptor-like protein from Crotalus durissus terrificus (the South American rattlesnake), Biochim. Biophys. Acta 1717 (2005) 27–33. [16] C. Seki, J.C. Vidal, A. Barrio, Purification of gyroxin from a South American rattlesnake (Crotalus durissus terrificus) venom, Toxicon 18 (1980) 235–247. [17] A. Hammersley, K. Brown, W. Burmeister, L. Claustre, A. Gonzalez, S. McSweeney, E. Mitchell, J. Moy, S. Svensson, A. Thompson, Calibration and application of an X-ray image intensifier/charge-coupled device detector for monochromatic macromolecular crystallography, J. Synchrotron Radiat. 4 (1997) 67–77. [18] A. Guinier, G. Fornet, Small Angle Scattering of X-rays, first ed. Wiley, London, 1995. [19] D.I. Svergun, Determination of the regularization parameter in indirect-transform methods using perceptual criteria, J. Appl. Crystallogr. 25 (1992) 495–503. [20] H. Fischer, M.O. Neto, H.B. Napolitano, I. Polikarpov, A.F. Craievich, Determination of the molecular weight of proteins in solution can be determined from a single SAXS measurement on a relative scale, J. Appl. Crystallogr. 43 (2010) 101–109. [21] S. Joshi, D.D. Pant, Steady-state and time-resolved fluorescence spectroscopy of quinine sulphate dication bound to sodium dodecylsulfate micelles: fluorescent complex formation, J. Lumin. 145 (2014) 224–231. [22] E. Habermann, K.L. Hardt, A sensitive and specific plate test for the quantification of phospholipases, Anal. Biochem. 50 (1972) 163–173. [23] N. Ohkura, S. Inoue, K. Ikeda, K. Hayashi, Isolation and characterization of a phospholipase A2 inhibitor from the blood plasma of the Thailand cobra Naja naja kaouthia, Biochem. Biophys. Res. Commun. 200 (1994) 784–788.

D

679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746

[24] N. Ohkura, Y. Kitahara, S. Inoue, K. Ikeda, K. Hayashi, Isolation and amino acid sequence of a phospholipase A2 inhibitor from the blood plasma of the sea krait, Laticauda semifasciata, J. Biochem. 125 (1999) 375–382. [25] G. Fairbanks, T.L. Steck, D.F. Wallach, Electrophoretic analysis of the major polypeptides of the human erythrocyte membrane, Biochemistry 10 (1971) 2606–2617. [26] N. Sreerama, R.W. Woody, Estimation of protein secondary structure from circular dichroism spectra: comparison of CONTIN, SELCON, and CDSSTR methods with an expanded reference set, Anal. Biochem. 287 (2000) 252–260. [27] L. Whitmore, B.A. Wallace, Protein secondary structure analysis from circular dichroism spectroscopy: methods and reference databases, Biopolymers 89 (2008) 392–400. [28] J. Söding, Protein homology detection by HMM–HMM comparison, Bioinformatics 21 (2005) 951–960. [29] M.A. Marti-Renom, A.C. Stuart, A. Fiser, A. Sali, Comparative protein structure modeling of genes and genomes, Annu. Rev. Biophys. Biomol. Struct. (2000) 291–325. [30] H.J.C. Berendsen, D. van der Spoel, R. van Drunen, GROMACS: a message-passing parallel molecular dynamics implementation, Comput. Phys. Commun. 91 (1995) 43–56. [31] C. Oostenbrink, T.A. Soares, N.F.A. van der Vegt, W.F. van Gunsteren, Validation of the 53A6 GROMOS force field, Eur. Biophys. J. 34 (2005) 273–284. [32] D.I. Svergun, M.H.J. Koch, Small-angle scatterings studies of biological macromolecules in solution, Rep. Prog. Phys. 66 (2003) 1735–1782. [33] E. Kokufuta, K. Ogawa, R. Doi, R. Kikuchi, R.S. Farinato, Geometrical characteristics of polyelectrolyte nanogel particles and their polyelectrolyte complexes studied by dynamic and static light scattering, J. Phys. Chem. B 111 (2007) 8634–8640. [34] E.B. Saff, A.B.J. Kuijlaars, Distributing many points on a sphere, Math. Intell. 19 (1997) 5–11. [35] M.D. Winn, C.C. Ballard, K.D. Cowtan, E.J. Dodson, P. Emsley, P.R. Evans, R.M. Keegan, E.B. Krissinel, A.G.W. Leslie, A. McCoy, S.J. McNicholas, G.N. Murshudov, N.S. Pannu, E.A. Potterton, H.R. Powell, R.J. Read, A. Vagin, K.S. Wilson, Overview of the CCP4 suite and current developments, Acta Crystallogr. D 67 (2011) 235–242. [36] P. Andrews, Estimation of the molecular weights of proteins by sephadex gelfiltration, Biochem. J. 91 (1964) 222–233. [37] P. Andrews, The gel-filtration behaviour of proteins related to their molecular weights over a wide range, Biochem. J. 96 (1965) 595–606. [38] S.-C. Lee, J.R. Whitaker, Are molecular weights of proteins determined by superose 12 column chromatography correct? J. Agric. Food Chem. 52 (2004) 4948–4952. [39] D.J. Winzor, Analytical exclusion chromatography, J. Biochem. Biophys. Methods 56 (2003) 15–52. [40] K. Okumura, S. Inoue, N. Ohkura, K. Ikeda, K. Hayashi, cDNA cloning of the two subunits of phospholipase A2 inhibitor PLI gamma from blood plasma of the Chinese mamushi, Agkistrodon blomhoffii siniticus, IUBMB Life 48 (1999a) (1999) 99–104. [41] M.-M. Thwin, P. Gopalakrishnakone, R.M. Kini, A. Armugam, K. Jeyaseelan, Recombinant antitoxic and antiinflammatory factor from the nonvenomous snake Python reticulatus: phospholipase A2 inhibition and venom neutralizing potential, Biochemistry 39 (2000) 9604–9611. [42] M.-M. Thwin, R.L. Satish, S.T. Chan, P. Gopalakrishnakone, Functional site of endogenous phospholipase A2 inhibitor from Python serum, Eur. J. Biochem. 269 (2002) 719–727. [43] I. Nobuhisa, S. Inamasu, M. Nakai, A. Tatsui, T. Mimori, T. Ogawa, Y. Shimohigashi, Y. Fukumaki, S. Hattori, H. Kihara, M. Ohno, Characterization and evolution of a gene encoding a Trimeresurus flavoviridis serum protein that inhibits basic phospholipase A2 isozymes in the snake's venom, Eur. J. Biochem. 249 (1997) 838–845. [44] S. Lizano, Y. Angulo, B. Lomonte, J.W. Fox, G. Lambeau, M. Lazdunski, J.M. Gutiérrez, Two phospholipase A2 inhibitors from the plasma of Cerrophidion (Bothrops) godmani which selectively inhibit two different group-II phospholipase A2 myotoxins from its own venom, Biochem. J. 346 (2000) 631–639. [45] C.Z. Oliveira, N.A. Santos-Filho, D.L. Menaldo, J. Boldrini-Franca, J.R. Giglio, L.A. Calderon, R.G. Stabeli, F.H. Rodrigues, L. Tasic, S.L. da Silva, A.M. Soares, Structural and functional characterization of a γ-type phospholipase A2 inhibitor from Bothrops jararacussu snake plasma, Curr. Top. Med. Chem. 11 (2011) 2509–2519. [46] P.G. Hains, K.W. Broady, Purification and inhibitory profile of phospholipase A2 inhibitors from Australian elapid sera, Biochem. J. 346 (2000) 139–146. [47] S. So, T. Murakami, N. Ikeda, T. Chijiwa, N. Oda-Ueda, T. Kuraishi, S. Hattori, M. Ohno, Identification and evolution of venom phospholipase A2 inhibitors from Protobothrops elegans serum, Biosci. Biotechnol. Biochem. 75 (2011) 480–488. [48] R. Shirai, M. Toriba, K. Hayashi, K. Ikeda, S. Inoue, Identification and characterization of phospholipase A2 inhibitors from the serum of the Japanese rat snake, Elaphe climacophora, Toxicon 53 (2009) 685–692. [49] K. Okumura, K. Masui, S. Inoue, K. Ikeda, K. Hayashi, Purification, characterization and cDNA cloning of a phospholipase A2 inhibitor from the serum of the nonvenomous snake Elaphe quadrivirgata, Biochem. J. 341 (1999) 165–171.

E

References

T

678

11

Please cite this article as: C.L. Fortes-Dias, et al., Insights on the structure of native CNF, an endogenous phospholipase A2 inhibitor from Crotalus durissus terrificus, the South American..., Biochim. Biophys. Acta (2014), http://dx.doi.org/10.1016/j.bbapap.2014.05.001

747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816