The sulfate-binding site structure of the human eosinophil cationic protein as revealed by a new crystal form

Journal of Structural Biology 179 (2012) 1–9

Contents lists available at SciVerse ScienceDirect

Journal of Structural Biology journal homepage: www.elsevier.com/locate/yjsbi

The sulfate-binding site structure of the human eosinophil cationic protein as revealed by a new crystal form Ester Boix a,⇑, David Pulido a, Mohammed Moussaoui a, M. Victòria Nogués a, Silvia Russi b a b

Department de Bioquímica i Biologia Molecular, Facultat de Biociències, Universitat Autònoma de Barcelona, Spain European Molecular Biology Laboratory, 6 Rue Jules Howoritz, 38042 Grenoble, France

a r t i c l e

i n f o

Article history: Received 24 November 2011 Received in revised form 25 April 2012 Accepted 26 April 2012 Available online 9 May 2012 Keywords: Eosinophil RNase Immunity Cytotoxicity Sulfate Glycosaminoglycans Crystal form

a b s t r a c t The human eosinophil cationic protein (ECP), also known as RNase 3, is an eosinophil secretion protein that is involved in innate immunity and displays antipathogen and proinflammatory activities. ECP has a high binding affinity for heterosaccharides, such as bacterial lipopolysaccharides and heparan sulfate found in the glycocalix of eukaryotic cells. We have crystallized ECP in complex with sulfate anions in a new monoclinic crystal form. In this form, the active site groove is exposed, providing an alternative model for ligand binding studies. By exploring the protein–sulfate complex, we have defined the sulfate binding site architecture. Three main sites (S1–S3) are located in the protein active site; S1 and S2 overlap with the phosphate binding sites involved in RNase nucleotide recognition. A new site (S3) that is unique to ECP is one of the key anchoring points for sulfated ligands. Arg 1 and Arg 7 in S3, together with Arg 34 and Arg 36 in S1, form the main basic clusters that assist in the recognition of ligand anionic groups. The location of additional sulfate bound molecules, some of which contribute to the crystal packing, may mimic the binding to extended anionic polymers. In conclusion, the structural data define a binding pattern for the recognition of sulfated molecules that can modulate the role of ECP in innate immunity. The results reveal the structural basis for the high affinity of ECP for glycosaminoglycans and can assist in structure-based drug design of inhibitors of the protein cytotoxicity to host tissues during inflammation. Ó 2012 Elsevier Inc. All rights reserved.

1. Introduction The human eosinophil cationic protein (ECP) is an eosinophil secretion protein with antipathogen activities and is involved in the host immune defense response (Boix et al., 2008; Venge et al., 1999). Mature ECP has a MW ranging from 15 to 21 kDa due to several glycosylation grades (Eriksson et al., 2007; Rubin et al., 2009). The protein is highly basic (pI  11), and it targets a wide range of pathogens, including helminths, protozoa and bacteria (Boix et al., 2012), suggesting a rather nonspecific mechanism of action. Indeed, structural and functional studies have identified protein regions that are involved in a membrane destabilizing mechanism (Carreras et al., 2003; Sanchez et al., 2011; Torrent et al., 2007, 2009a,b). ECP (ID:P12724; EC:3.1.27.⁄), also known as RNase 3, was first identified as an eosinophil cytotoxic protein and ascribed to the Abbreviations: ECP, eosinophil cationic protein; EDN, eosinophil derived neurotoxin; GAGs, glycosaminoglycans; LPS, lipopolysaccharides; pn, protein interaction site for phosphate anions; Sn, protein interaction site for sulfate anions. ⇑ Corresponding author. Address: Departament de Bioquímica i Biologia Molecular, Facultat de Biociències, Universitat Autònoma de Barcelona, 08193 Cerdanyola del Vallès, Spain. Fax: +34 93 5811264. E-mail address: [email protected] (E. Boix). 1047-8477/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jsb.2012.04.023

RNase A superfamily by sequence homology (Gleich et al., 1986; Rosenberg et al., 1989). The first ECP crystal structure (Boix et al., 1999a) confirmed that the protein overall topology followed the RNase A superfamily fold, and the structure of ECP in complex with the 20 50 ADP nucleotide (Mohan et al., 2002) revealed the main structural traits of the catalytic site. RNase A binds RNA via several subsites, which recognize the phosphate, base and ribose units and are designated as pn, Bn and Rn, respectively (Cuchillo et al., 2011; Nogues et al., 1998; Pares et al., 1991; Raines, 1998). p1 being the main site where the phosphodiester bond is cleaved. Following an equivalent nomenclature, the corresponding RNA binding sites for the distinct members of the RNase A family were assigned. The protein reduced RNase catalytic efficiency (Boix et al., 1999b; Sorrentino and Libonati, 1994) and the distinct polynucleotide cleavage pattern were related to impaired interactions at substrate binding secondary sites (Boix et al., 1999a,b). On the other hand, ECP has a high affinity for sulfated heterosaccharides. In fact, ECP was first reported to bind heparin when it was originally purified from eosinophils (Olsson and Venge, 1974; Venge et al., 1999), and key interacting residues were recently located at the active site cleft by molecular docking prediction (Torrent et al., 2011a) and NMR spectroscopy (García-Mayoral et al., 2010). Binding

2

E. Boix et al. / Journal of Structural Biology 179 (2012) 1–9

to heparan sulfate may facilitate adhesion to eukaryote extracellular matrix and promote the RNase cytotoxic activity (Chao and Raines, 2011; Chao et al., 2010; Fan et al., 2007). Because heparin and heparan sulfate derivatives modulate key biological cell paths, the pharmaceutical industry is applying a great effort to unravel the underlying structural determinants for protein binding (Boix et al., 2011; Capila and Linhardt, 2002; Imberty et al., 2007). Target recognition is primarily driven by electrostatic interactions that depend on the particular arrangement of protein basic clusters (Capila and Linhardt, 2002; Cardin and Weintraub, 1989; Fromm et al., 1995). These interactions can favor binding to sulfated saccharides and other anionic polymers, such as nucleic acids. Not surprisingly, the Cardin and Weintraub motif for heparin recognition shows a preference for sulfate anion binding and secondarily for phosphate groups (Boix et al., 2011). Indeed, ECP represents a nice illustrative example of protein affinity for both sulfated and phosphorylated polymers. We report here the structure of ECP bound to sulfate anions, which provides the opportunity to analyze the putative binding sites for sulfated ligand molecules of biological significance. We also compare the new monoclinic crystal form to the previously reported hexagonal and tetragonal crystal forms (Boix et al., 1999a; Mallorqui-Fernandez et al., 2000) and to the NMR solution structure (Laurents et al., 2009). 2. Materials and methods 2.1. Protein expression and purification Recombinant ECP was expressed and purified in Escherichia coli BL21(DE3) cells as described previously (Boix, 2001). Briefly, a synthetic gene for human ECP was cloned into the pET11c expression vector, and the protein was purified from inclusion bodies. The protein was purified by cationic exchange FPLC on a Resource-S column followed by reverse phase chromatography on a Vydac C4 column. 2.2. Protein crystallization Crystallization conditions were screened at the High Throughput Crystallization facility of the EMBL Grenoble Outstation, starting from a protein sample of 14 mg/ml in 20 mM Na Cacodylate, pH 5. A single crystal appeared after two months of incubation at 20 °C in 0.2 M lithium sulfate monohydrate, 0.1 M Tris hydrochloride, pH 8.5, and 15% PEG 4000, which corresponds to condition number 17 from the Crystal Screen Lite, Hampton Research. A cryoprotectant was prepared by supplementing the reservoir crystallization solution with 20% PEG 400. Crystallization conditions were also set to reproduce and optimize the previously reported ECP tetragonal crystal form (Mallorqui-Fernandez et al., 2000) using the hanging drop methodology in a 1 lL:1 lL protein/precipitant mixture. The precipitant concentration was slightly reduced to increase the crystal size. A sample of 15 mg/ml of protein in 20 mM Na citrate, pH 5.2 was incubated at 16 °C in 0.1 M Na citrate, pH 5.2, 6% Jeffamine M600, 8 mM KH2PO4, 2 mM K2HPO4, and 10 mM FeCl3. Crystals appeared after 5 days and were soaked in a cryoprotectant by supplementing the crystallization buffer with 30% 2-methyl-2, 4-pentanediol (MPD). 2.3. Data processing and structure refinement Diffraction data for the new ECP crystal form were collected at 100 K at the BM14 beamline station at the European Synchrotron Radiation Facility (ESRF), Grenoble. One hundred fifty images were collected at a 1° oscillation range. Diffraction data for the ECP

tetragonal crystal form were collected at 100 K at the X13 beamline station at the Deutsches Elektronen Synchrotron (DESY)-EMBL, Hamburg. Two hundred images were collected at a 0.5° oscillation range. Data processing was performed with the XDS software, and scaling of integrated reflections was performed with SCALA. The ECP tetragonal crystal form (1DYT (Mallorqui-Fernandez et al., 2000)) was used as a starting model to solve both structures. Molecular replacement was performed with PHASER, and refinement was achieved with REFMAC5 (Murshudov et al., 1997), as implemented in the CCP4 suite package (CCP4i, 1994). The model was improved by alternate cycles of refinement and manual rebuilding with COOT (Emsley and Cowtan, 2004). Water molecules were incorporated with COOT at several steps of refinement when electron density peaks of either (Fo–Fc) or (2Fo–Fc) maps were higher than 3 or 1.5r, respectively, and were located within a hydrogen bond distance from appropriate atoms. A final refinement cycle with TLS Motion Determination (Painter and Merritt, 2006) was performed, and the stereochemistry of the final structure was validated with PROCHECK (Laskowski et al., 1993).

3. Results and discussion 3.1. A new crystal form A new crystal form for ECP was identified by screening multiple crystallization conditions using the High Throughput Crystallization facility of the EMBL Grenoble Outstation (https://htxlab.embl.fr). The new crystal form belongs to the space group C2 and contains two molecules in the asymmetric unit. Data to 1.7 Å resolution were collected, and the final refined parameters are summarized in Table 1. The protein molecules are tightly packed in the crystal with a solvent content of 24%. A total of fourteen sulfate anions were fit unambiguously to the (Fo–Fc) electron density map. A sigma-A weighted map was also built to confirm the anion positions (Fig. S1). Five sulfate anions are associated with molecule A and nine with molecule B (Fig. 2 and Table S1). Three sulfate anions overlap in both molecules. Sulfate binding sites were classified and named S1 to S11, according to their relative position. The final average B factors are below 15 Å2 for the two protein molecules in the asymmetric unit, where only segment 89–93 in loop L7 shows higher B factors (Fig. 1B). Few side chains show only partial density (Arg 28, Trp 35, Arg 36, Gln 58, Arg 73, Phe 76, Asn 92, Arg 101 and Arg 105), and residues Ile 13, Asn 53, Ser 59, Asn 69, His 82, Ile 86 and Arg 105 were built with alternate conformations. Comparison of molecules A and B in the asymmetric unit reveals no major differences (RMSD between both main chains is 0.56 Å2). The side chain of Tyr 122 has a different orientation in both molecules, which can be attributed to its position at the interface between the two asymmetric unit molecules. The interface between molecules A and B is mainly driven by interactions between loops L2 and L8 and the external side of a-helices 1–4 (Fig. 2 and Table S2). Two sulfate anions, A306 (S6) and B310 (S10), connect both molecules at each end of the interface. Segment 115–122, which showed high B factors in previous structures, is fairly rigid in our structure, as it is involved in the interface between molecules A and B. The intermolecular packing interactions between symmetry related molecules are detailed in Table S3. In particular, exposed Arg residues in loop regions are the critical elements connecting symmetry molecules, often through sulfate molecules, providing tight packing and contributing to the reduced mobility of some of the exposed loops, such as L5 and L8 (Fig. 1B). A close side to side comparison was performed with the other two available crystal forms, PDB ID 1QMT (Boix et al., 1999a) and

E. Boix et al. / Journal of Structural Biology 179 (2012) 1–9 Table 1 Data collection, processing and structure refinement statistics.

Data collection Space group Unit cell a, b, c (Å)

a = b = c (°) Resolution (Å) No. of reflections (measured/ unique) Rmergea (%) I/rI Completeness (%) Wilson B factor (Å2) Matthews coeff. (Å3/Da) Solvent content (%) Refinement Resolution range (Å) Rcryst/Rfreed,e (%) No. of protein atoms No. of water molecules No. of bound anions R.m.s. deviation from ideal geometry Bond lengths (Å) Bond angles (deg) B-factors (Å2) Protein atoms (mol A/mol B) All Main chain Side chain Anion atoms Water molecules PDB code

ECP–sulfate

ECP–citrate

C2

P4322

92.26, 51.29, 55.39 90, 111.28, 90 1.69 94301/26684

62.51, 62.51, 174.86 90, 90, 90 1.70 266250/38834

7.0 (45)b 11.4 (2.3)b 98.6 (90.8)b 23.6 1.63 24.17

6.0 (50)c 18.2 (2.9)c 99.2 (98.1)c 22.7 2.72 47.4

44–1.69 16.2/22.3 2536 371 14

62–1.70 19/21.9 2555 310 5

0.021 1.9

0.021 1.95

14.59/14.79 12.07/11.93 16.82/17.29 36.01 30.83 4A2O

13.56/15.82 11.69/13.89 15.30/17.64 20.54 30.54 4A2Y

P P P P Rmerge ¼ hkl j¼1 to NjIhklIhkl ðjÞj= hkl j¼1 to N Ihkl (j), where N is the redundancy of the data. b The outermost shell is 1.78–1.69 Å. c The outermost shell is 1.75–1.7 Å. P P d Rcryst ¼ hkl jjF obs j  jF calc jj= hkl jF obs j, where the Fobs and Fcalc are the observed and calculated structure factor amplitudes of reflection hkl. e Rfree = is equal to Rcryst for a randomly selected 5% subset of reflections that were not used in refinement. a

1DYT (Mallorqui-Fernandez et al., 2000), and with the threedimensional NMR structure (2KB5 (Laurents et al., 2009)). The first crystal structure reported was solved at 2.4 Å (Boix et al., 1999a). The protein crystallized in the hexagonal P63 space group with only one molecule in the asymmetric unit. The crystallization buffer was equilibrated at pH 8.5, as in the presently reported monoclinic crystal form. The basic pH correlates in both structures with a relative position of the catalytic His 128 in the ‘‘active’’ or A conformation, in comparison with the ‘‘non-active’’ or B conformation reported for the tetragonal crystal form, solved at pH 5 (Mallorqui-Fernandez et al., 2000). The active and non-active conformations of the catalytic His 128 also follow an equivalent pH dependent behavior in the counterpart residue in other RNase A homolog structures (Berisio et al., 1999; Boix et al., 1999a; Swaminathan et al., 2002). Alternate orientations are also found for His 64 and His 82, two residues that showed a high mobility in the NMR structure. In our structure, His 82 exists in two conformations and forms a hydrogen bond with either Asp 84 or Thr 42. In the other two crystal forms, His 82 exists in a unique conformation, with interactions with either Thr 42 or Asp 84. Interestingly, His 82 was also reported to adopt alternate conformations in the EDN high resolution structure (0.98 Å) and was also suggested to be dependent on the crystallization buffer pH (Swaminathan et al., 2002). The present monoclinic crystal form structure (PDB ID: 4A2O) was compared with the tetragonal form using the current solved

3

structure at 1.7 Å (PDB ID: 4A2Y) as a reference (see Table 1). The tetragonal crystal form reveals a distinct intermolecular interface between molecules A and B in the asymmetric unit. The active site groove remains buried at the interface, conformed by loops L1 and L3, and a citrate molecule is bound at Arg 34, one of the main exposed basic clusters for anionic ligand binding (Fig. S2 and Table S4). On the contrary, in the monoclinic crystal form, the main sites for potential ligand binding are freely exposed. Moreover, the tetragonal form is not suitable for nucleotide analog binding analysis because the catalytic His 128 is involved in crystal packing interactions in a ‘‘non-active’’ conformation. We also compared the ECP crystal structures with the NMR structure. The NMR structure, 2KB5 (Laurents et al., 2009), further defined the most mobile regions of the protein in solution. Conformational diversity was observed for loop segments 17–22, 65–68 and 92–95, where segment 92–95 (L7) showed the largest RMSD between NMR models. A particularity unique to the NMR structure that was not observed in any of the three crystal forms is the position of Tyr 33. In the NMR structure, Tyr 33 is oriented in the opposite direction than in the crystal structure and is hydrogen bonded to Lys 38. This interaction would remove Lys 38 from its catalytically favorable orientation. Interestingly, nitration of Tyr 33, which has been described to occur in vivo (Ulrich et al., 2008), would provide a negative charge that may enhance the interaction with the e amino group of Lys 38. Holding Lys 38 in an inactive conformation was suggested (Laurents et al., 2009) to be one of the reasons for ECP reduced catalytic activity. Finally, we compared the four structures together. We observed the most variability in loop L7 (89–95). Loop 88–95 is partially disordered, mostly from residues 90–92. This loop shows higher B factors in our structure (Fig. 1B), which is also observed in other RNase family homologs (Leonidas et al., 1999; Berisio et al., 1999; Swaminathan et al., 2002). In addition, in the present monoclinic crystal form structure, residues Ile 93–Asn 95 adopt a ‘‘ag turn’’ secondary structure, which has also been reported for the high-resolution EDN crystal structure (0.98 Å, 1GQV) (Swaminathan et al., 2002). On the contrary, this ‘‘ag turn’’ is not observed in the other two ECP crystal structures or in the NMR structure. In fact, in the tetragonal crystal form structure, the loop is involved at the interface between the two molecules in the asymmetric unit and presents lower mobility. Moreover, in the present structure, L7 loop is bound to two sulfate anions (S8 and S11). Comparison of all four structures also reveals variability at Trp 35, His 64, His 82, Ile 86 and His 128 side chain orientations. His 128 is hydrogen bonded to Asp 130 only in the NMR and tetragonal crystal form structures. An equivalent hydrogen bond is reported for the RNase A structure and may account for the optimum orientation for catalysis at the active site, although the direct contribution of the corresponding His 119–Asp 121 ‘‘dyad’’ in the RNase A catalytic mechanism is unclear (Raines, 1998). Comparison of the ECP structures also provides controversial data, as His 128–Asp 130 is concomitant to the B inactive form in the tetragonal crystal form structure but can also be formed in the A active form in the NMR structure. In the present monoclinic crystal structure, His 128 is in the ‘‘active’’ conformation, although the side chain is partially disordered in molecule B. Another residue reported previously in alternate conformations is Lys 38, which in our structure is bound to a sulfate anion. Arg 34 also adopts two alternate conformations in the free NMR structure, and, upon sulfate binding, is fixed in a single orientation. Comparison of the different crystal forms with the NMR structure is also helpful to predict and analyze the potential conformational variability of the protein. For example, hydrogen bond interactions that were deduced to participate in the overall stability of the protein by molecular dynamics simulations

4

E. Boix et al. / Journal of Structural Biology 179 (2012) 1–9

A

B

L3

L1 α3

α1

L8

L7

β7

α2

L2 α4

L4 β1

L5

β4 β3

β6 L6

Fig. 1. (A) Sequence alignment of ECP, EDN and RNase A using the ClustalW server. Secondary structure elements are numbered and depicted according to the ECP structure (4A2O). Cys residues involved in disulfide bonds are labeled in green. The figure was drawn using the Espript software (Gouet et al., 1999). (B) Schematic ribbon representation of the ECP three-dimensional structure (4A2O). Protein molecule A and the main S1–S3 sulfate anions are shown together with the disulfide bonds and colored according to the residue and anion average B factors. Secondary structure elements are indicated. The picture was drawn using PyMOL (www.pymol.org).

(Sanjeev and Vishveshwara, 2004) can adopt alternate conformations. Sanjeev and Vishveshwara predicted that the interactions that fix the N-terminal a-helix to the core of the protein at the end of loop L8 (112–122) are the main contributors to protein stability. Arg 117 is one of the key residues that anchors the N-terminus to the C-terminus loop. In the monoclinic form, the hydrogen bonds between residues 115 and 117 hinder Arg 117 from forming a hydrogen bond with the main chain of residues 3 and 5, as is observed in the other two crystal forms. Moreover, the bond between Asp 112 and Arg 7 is lost in the monoclinic crystal form structure, where Arg 7 is involved in sulfate binding. Another critical hydrogen bond that fixes the N-terminus is between Asp 118 and Thr 6, which is found in all the structures. The authors defined two ‘‘open’’ and ‘‘closed’’ conformations of the protein, which depend on the potential interactions between the N- and C-termini. These two conformations could modulate ligand accessibility at the active site (Sanjeev and Vishveshwara, 2004). Another emblematic interaction is the bond between Asp 130 and Asn 65, which is observed in the new structure, but not in the other crystal structures due to an alternate orientation of the 62–70 loop. In the new structure, Asp 130 simultaneously interacts

with both His 64 and His 128, as in the NMR structure. Variability in the side chain conformation of His 64 also offers diverse partner connections. His 64 is unbound in the new structure but participates in connecting loop L5 loop (62–70) to the last loop by a hydrogen bond with Thr 131 in the other two crystal forms. Finally, we further compared the previously reported tetragonal crystal form (Mallorqui-Fernandez et al., 2000) with our structure of the tetragonal crystal form (PDB ID: 4A2Y) at slightly higher resolution (1.7 Å) and refined to considerably much lower R factors (Table 1). Our structure has a total of three bound citrate molecules (Figs. S1, S2 and Table S4), where Cit 303 was not reported in the previous structure. In addition, while Cit 301 is found in equivalent positions and is very well defined in our structure, the second citrate molecule (Cit 302) shows high mobility, and the electron density contour suggests several alternate conformations, which are difficult to model properly. 3.2. Exploring the ECP sulfate binding mode The structure of ECP crystallized in the presence of lithium sulfate provides an excellent model to study the protein sulfate

E. Boix et al. / Journal of Structural Biology 179 (2012) 1–9

5

Fig. 2. Schematic representation of the two molecules in the asymmetric unit (4A2O). Residues involved in the A and B interface are colored in red. The side chains of the hydrogen bonded residues at the interface are shown. Sulfate anions are depicted as spheres and labeled according to the nomenclature in Table S1. The figure was drawn using PyMOL (www.pymol.org).

binding mode. A total of fourteen sulfate molecules are bound to the two protein molecules in the asymmetric unit. The positions of three of these anions overlap, and binding sites S1–S11 can be defined (Fig. 2). S1, S2 and S3 are located in the protein active site groove, whereas the other sulfate anions are at the protein surface (Fig. S3) and partly contribute to crystal packing interactions (Table S3). Individual average B-factors range from 15 to 55 Å2.

Most anions involved in the crystal packing show slightly lower B-factors. Table S1 details the contact residues for each assigned anion. In particular, residues Arg 34, Arg 36 and Asn 39 in S1; Gln 14, Lys 38 and His 128 in S2; and Arg 1 and Arg 7 in S3 interact with the sulfate anions in the protein active site groove (Fig. 3). A detailed analysis of the ligand environment revealed that two sulfates are located at the equivalent phosphate binding sites. In

Fig. 3. Schematic picture showing all the protein residue side chains at van der Waals distance from the sulfate anions and the hydrogen bond interactions with sulfates S1 (A), S2 (B) and S3 (C). Hydrogen bond distances (Å) are indicated. The figure was drawn using PyMOL (www.pymol.org).

6

E. Boix et al. / Journal of Structural Biology 179 (2012) 1–9

particular, sulfates S1 and S2 occupy positions near the phosphate binding sites that were previously identified in the ECP-20 50 ADP complex (Fig. 4A). Comparison with the EDN sulfate complex structure (Leonidas et al., 2001; Mosimann et al., 1996) also confirms the conserved position of both sulfates (Fig. 4B). EDN residues involved in site p-1 were confirmed by site-directed mutagenesis (Sikriwal et al., 2007) and crystal complexes with nucleotide analogs (Baker et al., 2006; Leonidas et al., 2001). However, significant differences are indeed observed between both eosinophil RNases. We suggest that Arg 34, which is unique to ECP, would strengthen the interactions at site p-1. On the contrary, interacting residues in p1 (Leonidas et al., 2001; Sikriwal et al., 2009) may be more favored in EDN and would account for its relative higher catalytic efficiency (Sorrentino and Libonati, 1994). Interestingly, no sulfate anion corresponding to the S3 site defined in ECP was reported in the EDN crystal complexes. 3.3. ECP unique distribution of basic clusters We have also compared the sulfate anion positions in the ECP structure to the corresponding RNase multimeric phosphate binding subsites (Nogues et al., 1998) using the tetranucleotide–RNase A complex structure (Fontecilla-Camps et al., 1994) as a reference (Fig. 5A). In detail, Arg 34 and Arg 36, which bind to sulfate S1, would be analogous to Arg 85 in RNase A, defined as the p-1 site (Fisher et al., 1998); residues Gln 14, Lys 38 and His 128 in the S2 sulfate binding site correspond to the main p1 RNase A site, and Arg 1 and Arg 7, which bind to sulfate S3, are located in the vicinity of a putative p3 site. Therefore, the new monoclinic crystal form with sulfate bound molecules allows us to confirm three defined phosphate binding sites: p-1 (S1), p1 (S2) and p3 (S3) sites along the catalytic groove, suggesting that impaired interaction at the p0 and p2 sites could explain the lower catalytic activity of ECP (Boix et al., 1999b). A second scenario should also be considered in which the sulfate site architecture is specifically optimized to suit the binding of sulfated glycosaminoglycans (GAGs), such as heparin. The relative location of bound sulfate anions has been used as a reference for predicting the binding mode of GAGs (Boix et al., 2011; Lortat-Jacob et al., 2002; Raghuraman et al., 2006). GAGs can mediate ECP immunomodulatory properties and cell surface attachment. Heparin binds with high affinity to ECP through interactions at its active site (Torrent et al., 2011a) (García-Mayoral et al., 2010). Residues Arg 1, Arg 7 and Arg 34 were predicted by molecular docking to provide the main driving force for heparin probe binding (Torrent et al., 2011a), and NMR spectroscopy data confirmed interactions between the a1 and b1 N-terminus secondary structure elements and the IdoA(2S)–GlcNS(6S) disaccharide (García-Mayoral et al., 2010). Residues 33–36 in L3 loop were also identified to be involved in heparin binding by site-directed mutagenesis (Fan et al., 2008), and a recent study on ECP antiparasitic activity identified Arg 34 as the key basic surface exposed residue (Singh and Batra, 2011). Interestingly, additional contribution of Arg 1 and Arg 7 is observed by NMR for an extended heparin probe (García-Mayoral et al., 2010, unpublished results). To further analyze the potential protein binding sites, we compared the location of the heparin tetrasaccharide docked to ECP (Boix et al., 2011) with the position of the sulfate bound molecules in the present crystal structure (Fig. 5B). Molecular modeling reveals how the IdoA–GlcNS–IdoA–GlcNS tetrasaccharide fits in the ECP active site groove and overlaps with the three consecutive sulfate molecules S1–S3. In particular, S1 and S2 are found in the environment of the second sugar unit. The GlcNS 20 sulfate is within hydrogen bonding distance from the S1 sulfate, and the S3 sulfate anion is located at the level of the 20 sulfate of the fourth sugar unit. Indeed, the corresponding distance between sites S1 and S3 fits the

Fig. 4. (A) Superposition of the 20 50 ADP nucleotide analog from the ECP crystal complex (1H1H) (Mohan et al., 2002) onto the ECP–sulfate anion complex structure (4A2O). The main chains of both ECP structures were superposed by least square fit with a mean deviation of 0.9 Å. The ECP-20 50 ADP complex is colored in green, and the ECP–sulfate complex is in cyan. 20 50 ADP is colored according to atom type, and sulfate anions are shown in cyan. (B) Superposition of the EDN–sulfate complex structure (1HI2) (Leonidas et al., 2001) onto the ECP–sulfate complex structure (4A2O). EDN and ECP protein main chains were superposed by least square fit with a 1.5 Å mean deviation. EDN and ECP protein main chains are colored in green and cyan, respectively. EDN sulfates are colored according to atom type, and ECP sulfates are shown in cyan. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

structural requirements reported for heparin binding proteins (Capila and Linhardt, 2002). Based on these data, we can conclude that S1 and S3 would be the main ECP interacting sites for GAG binding. Finally, the particular distribution of exposed basic clusters may also contribute to the interaction between ECP and anionic heterosaccharides at the bacteria outer envelope. ECP is an antimicrobial RNase (Boix and Nogues, 2007) with high bactericidal activity and a particularly high binding affinity for the lipopolysaccharides (LPS) exposed at the outer membrane of Gram-negative bacteria (Boix et al., 2008; Torrent et al., 2008, 2010). In vivo studies using bacteria strains with progressively truncated LPSs confirmed that the protein interaction between ECP and the LPS sugar core mediates bacterial agglutination and triggers cell death (Pulido et al., 2012). Docking simulations of the LPS NAG4P–NAG1P disaccharide unit onto ECP identified residues Arg 1, Gln 14, Lys 38 and Arg 34 to be within hydrogen bonding distance to the selected target (Torrent et al., 2011a), and further modeling studies using the whole LPS molecule (1FI1) also predicted that the protein N-terminus was involved in anchoring the phosphate groups (Pulido et al., 2012). Complementary synthetic peptide screening studies also revealed that the protein region involved in LPS binding was located at the N-terminus (Torrent et al., 2011b).

E. Boix et al. / Journal of Structural Biology 179 (2012) 1–9

7

Arg 121 at S6 and Arg 66 at S7 (Fig. S3), which may assist in binding to longer anionic polymers. Some of these residues, such as Arg 34, Arg 36, Arg 75, Arg 77, Arg 104 and Arg 121 have been identified to contribute to the antipathogenic activities of the protein (Carreras et al., 2003; Singh and Batra, 2011). To fully illustrate the role of ECP basic clusters, we should also take into account the potential posttranslational modifications that would modulate the function of the protein function in vivo, such as N-glycosylation (Eriksson et al., 2007) and, in particular, Tyr-33 nitration (Ulrich et al., 2008), which may hinder anionic ligand binding at the S1 site. 4. Conclusions The new ECP crystal form in complex with sulfate anions provides direct evidence of the main protein sulfate binding sites. We have defined three sulfate interacting sites at the RNase catalytic groove, namely S1–S3, that would define the main region involved in sulfated heterosaccharide binding. In particular, site S3 is unique to ECP in the RNase A superfamily. Additional basic residues at the protein surface may assist in anchoring longer anionic polymers. The data show how the eosinophil RNase contains a specific sulfate binding site architecture, which could explain its reduced catalytic activity and enhanced affinity for relevant sulfated compounds that would mediate the protein’s role in host defense. 5. Accession numbers The atomic coordinates and structure factors have been deposited in the European Protein Data Bank (PDBe) (http:// www.pdbe.org/) (PDB ID: 4A2O and 4A2Y for ECP-sulfate and ECP-citrate anion complexes, respectively). Acknowledgments Fig. 5. Superposition of the tetranucleotide from the RNase A complex (1RCN) (Fontecilla-Camps et al., 1994) (A) and the heparin tetrasaccharide docked onto the ECP structure (Boix et al., 2011) and (B) onto the ECP–sulfate complex structure (4A2O). The protein electrostatic surface is shown with the ECP sulfate anions (S1, S2 and S3) in ball and sticks. ECP and RNase A main chain atoms were superposed by least square fit with a 6.2 Å mean deviation. The RNase A nucleotide is shown in brown, and the heparin tetramer, composed of N-sulfated glucosamine (GlcNS) and sulfated iduronic acid (IdoA(2S)) alternate units, is shown in brown and colored according to atom type. The corresponding phosphate and sulfate binding sites are indicated. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Finding the structural determinants for LPS recognition is one of priorities of the pharmaceutical industry (Brandenburg et al., 2011; Cohen, 2002), and work is in progress to develop peptide analogs as LPS binders that can work both as antibiotics and sepsis inhibitors (Brandenburg et al., 2010; Frecer et al., 2004; Gutsmann et al., 2010). The particularly high affinity of ECP to anionic biomolecular targets, such as heparin and LPS, which is not shared with other RNase A family counterparts, relies mainly on its high cationicity. Inspection of ECP Arg residues distributed at the protein surface reveals a unique pattern (Fig. 1A) acquired during the divergence of both eosinophil RNases from a common ancestor in a relative short evolution period (Rosenberg et al., 1995; Zhang et al., 1998). During this divergent path, ECP increased its number of Arg residues to a total of 19, 15 of which are unique to ECP. Interestingly, 10 of these Arg residues are involved in sulfate binding (Table S1 and Fig. S3). In addition to the Arg residues involved in the S1–S3 binding sites (Fig. 3), ECP also contains Arg 77 and Arg 104 at S4, Arg 75 at S5,

We acknowledge the synchrotron staff at BM14, ESRF, Grenoble and at X13, DESY, EMBL, Hamburg. We also want to thank Dr. Demetres Leonidas and Dr. Daniel Fernández for their help during data collection. We acknowledge the European Community-Research Infrastructure Action PCUBE, FP7 Program. This research was funded by the European Community0 s Seventh Framework Programme (FP7/2007-2013) under Grant Agreement No. 227764 (P-CUBE). The work was supported by the Ministerio de Educación y Cultura (grant number BFU2009-09371) and co-financed by FEDER funds and the Generalitat de Catalunya (2009 SGR 795). D.P. is a recipient of a UAB predoctoral fellowship. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jsb.2012.04.023. References Baker, M.D., Holloway, D.E., Swaminathan, G.J., Acharya, K.R., 2006. Crystal structures of eosinophil-derived neurotoxin (EDN) in complex with the inhibitors 50 -ATP, Ap3A, Ap4A, and Ap5A. Biochemistry 45, 416–426. Berisio, R., Lamzin, V.S., Sica, F., Wilson, K.S., Zagari, A., Mazzarella, L., 1999. Protein titration in the crystal state. J. Mol. Biol. 292, 845–854. Boix, E., 2001. Eosinophil cationic protein. Method Enzymol. 341, 287–305. Boix, E., Nogues, M.V., 2007. Mammalian antimicrobial proteins and peptides: overview on the RNase A superfamily members involved in innate host defence. Mol. Biosyst. 3, 317–335. Boix, E., Torrent, M., Sánchez, D., Nogués, M.V., 2008. The antipathogen activities of eosinophil cationic protein. Current Pharm. Biotechnol. 9, 141–152.

8

E. Boix et al. / Journal of Structural Biology 179 (2012) 1–9

Boix, E., Torrent, M., Nogués, M., Salazar, V., 2011. Searching for heparin binding patterns. In: Piyathilake, D., Liang, R. (Eds.), Heparin: Properties, Uses and Side Effects Nova Sciences Publishers, Inc. Boix, E., Leonidas, D.D., Nikolovski, Z., Nogues, M.V., Cuchillo, C.M., Acharya, K.R., 1999a. Crystal structure of eosinophil cationic protein at 2.4 A resolution. Biochemistry 38, 16794–16801. Boix, E., Nikolovski, Z., Moiseyev, G.P., Rosenberg, H.F., Cuchillo, C.M., Nogues, M.V., 1999b. Kinetic and product distribution analysis of human eosinophil cationic protein indicates a subsite arrangement that favors exonuclease-type activity. J. Biol. Chem. 274, 15605–15614. Boix, E., Salazar, V.A., Torrent, M., Pulido, D., Nogués, M.V., Moussaoui, M., 2012. Structural determinants of the eosinophil cationic protein antimicrobial activity. Biol. Chem. 393. Brandenburg, K., Andra, J., Garidel, P., Gutsmann, T., 2011. Peptide-based treatment of sepsis. Appl. Microbiol. Biotechnol. 90, 799–808. Brandenburg, K., Garidel, P., Fukuoka, S., Howe, J., Koch, M.H., Gutsmann, T., Andra, J., 2010. Molecular basis for endotoxin neutralization by amphipathic peptides derived from the alpha-helical cationic core-region of NK-lysin. Biophys. Chem. 150, 80–87. Capila, I., Linhardt, R.J., 2002. Heparin–protein interactions. Angew. Chem. Int. Ed. Engl. 41, 391–412. Cardin, A.D., Weintraub, H.J., 1989. Molecular modeling of protein– glycosaminoglycan interactions. Arteriosclerosis 9, 21–32. Carreras, E., Boix, E., Rosenberg, H.F., Cuchillo, C.M., Nogues, M.V., 2003. Both aromatic and cationic residues contribute to the membrane-lytic and bactericidal activity of eosinophil cationic protein. Biochemistry 42, 6636– 6644. CCP4i, 1994. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D Biol. Crystallogr. 50, 760–763. Cohen, J., 2002. The immunopathogenesis of sepsis. Nature 420, 885–891. Cuchillo, C.M., Nogues, M.V., Raines, R.T., 2011. Bovine pancreatic ribonuclease: fifty years of the first enzymatic reaction mechanism. Biochemistry 50, 7835–7841. Chao, T.Y., Raines, R.T., 2011. Mechanism of Ribonuclease A endocytosis analogies to cell-penetrating peptides. Biochemistry 50, 8374–8382. Chao, T.Y., Lavis, L.D., Raines, R.T., 2010. Cellular uptake of ribonuclease A relies on anionic glycans. Biochemistry 49, 10666–10673. Emsley, P., Cowtan, K., 2004. Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132. Eriksson, J., Woschnagg, C., Fernvik, E., Venge, P., 2007. A SELDI–TOF MS study of the genetic and post-translational molecular heterogeneity of eosinophil cationic protein. J. Leukocyte Biol. 82, 1491–1500. Fan, T.C., Chang, H.T., Chen, I.W., Wang, H.Y., Chang, M.D., 2007. A heparan sulfatefacilitated and raft-dependent macropinocytosis of eosinophil cationic protein. Traffic 8, 1778–1795. Fan, T.C., Fang, S.L., Hwang, C.S., Hsu, C.Y., Lu, X.A., Hung, S.C., Lin, S.C., Chang, D.T., 2008. Characterization of molecular interactions between eosinophil cationic protein and heparin. J. Biol. Chem. 283, 25468–25474. Fisher, B.M., Grilley, J.E., Raines, R.T., 1998. A new remote subsite in ribonuclease A. J. Biol. Chem. 273, 34134–34138. Fontecilla-Camps, J.C., de Llorens, R., le Du, M.H., Cuchillo, C.M., 1994. Crystal structure of ribonuclease A.d(ApTpApApG) complex. Direct evidence for extended substrate recognition. J. Biol. Chem. 269, 21526–21531. Frecer, V., Ho, B., Ding, J.L., 2004. De novo design of potent antimicrobial peptides. Antimicrob. Agents Chemother. 48, 3349–3357. Fromm, J.R., Hileman, R.E., Cadwell, E.E., Weiler, J.M., Linhardt, R.J., 1995. Differences in the interaction of heparin with arginine and lysine and the importance of these basic amino acids in the binding of heparin to acidic fibroblast growth factor. Arch. Biochem. Biophys. 323, 279–287. García-Mayoral, M.F.M., de la Torre, B.G., Andreu, D., Boix, E., Nogués, M.V., Rico, M., Laurents, D.V., Bruix, M., 2010. NMR structural determinants of eosinophil cationic protein binding to membrane and heparin mimetics. Biophys. J. 98, 2702–2711. Gleich, G.J., Loegering, D.A., Bell, M.P., Checkel, J.L., Ackerman, S.J., McKean, D.J., 1986. Biochemical and functional similarities between human eosinophilderived neurotoxin and eosinophil cationic protein: homology with ribonuclease. Proc. Natl. Acad. Sci. USA 83, 3146–3150. Gouet, P., Courcelle, E., Stuart, D.I., Metoz, F., 1999. ESPript: analysis of multiple sequence alignments in PostScript. Bioinformatics 15, 305–308. Gutsmann, T., Razquin-Olazaran, I., Kowalski, I., Kaconis, Y., Howe, J., Bartels, R., Hornef, M., Schurholz, T., Rossle, M., Sanchez-Gomez, S., Moriyon, I., Martinez de Tejada, G., Brandenburg, K., 2010. New antiseptic peptides to protect against endotoxin-mediated shock. Antimicrob. Agents Ch. 54, 3817–3824. Imberty, A., Lortat-Jacob, H., Perez, S., 2007. Structural view of glycosaminoglycan– protein interactions. Carbohyd. Res. 342, 430–439. Laskowski, R.A., MacArthur, M.W., Moss, D.S., Thornton, J.M., 1993. PROCHECK – a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr. 26, 283–291. Laurents, D.V., Bruix, M., Jimenez, M.A., Santoro, J., Boix, E., Moussaoui, M., Nogues, M.V., Rico, M., 2009. The (1)H, (13)C, (15)N resonance assignment, solution structure, and residue level stability of eosinophil cationic protein/RNase 3 determined by NMR spectroscopy. Biopolymers 91, 1018–1028. Leonidas, D.D., Shapiro, R., Allen, S.C., Subbarao, G.V., Veluraja, K., Acharya, K.R., 1999. Refined crystal structures of native human angiogenin and two active site variants: implications for the unique functional properties of an enzyme involved in neovascularisation during tumour growth. J. Mol. Biol. 285, 1209– 1233.

Leonidas, D.D., Boix, E., Prill, R., Suzuki, M., Turton, R., Minson, K., Swaminathan, G.J., Youle, R.J., Acharya, K.R., 2001. Mapping the ribonucleolytic active site of eosinophil-derived neurotoxin (EDN). High resolution crystal structures of EDN complexes with adenylic nucleotide inhibitors. J. Biol. Chem. 276, 15009– 15017. Lortat-Jacob, H., Grosdidier, A., Imberty, A., 2002. Structural diversity of heparan sulfate binding domains in chemokines. Proc. Natl. Acad. Sci. USA 99, 1229– 1234. Mallorqui-Fernandez, G., Pous, J., Peracaula, R., Aymami, J., Maeda, T., Tada, H., Yamada, H., Seno, M., de Llorens, R., Gomis-Ruth, F.X., Coll, M., 2000. Threedimensional crystal structure of human eosinophil cationic protein (RNase 3) at 1.75 A resolution. J. Mol. Biol. 300, 1297–1307. Mohan, C.G., Boix, E., Evans, H.R., Nikolovski, Z., Nogues, M.V., Cuchillo, C.M., Acharya, K.R., 2002. The crystal structure of eosinophil cationic protein in complex with 20 ,50 -ADP at 2.0 A resolution reveals the details of the ribonucleolytic active site. Biochemistry 41, 12100–12106. Mosimann, S.C., Newton, D.L., Youle, R.J., James, M.N., 1996. X-ray crystallographic structure of recombinant eosinophil-derived neurotoxin at 1.83 A resolution. J. Mol. Biol. 260, 540–552. Murshudov, G.N., Vagin, A.A., Dodson, E.J., 1997. Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. D Biol. Crystallogr. 53, 240–255. Nogues, M.V., Moussaoui, M., Boix, E., Vilanova, M., Ribo, M., Cuchillo, C.M., 1998. The contribution of noncatalytic phosphate-binding subsites to the mechanism of bovine pancreatic ribonuclease A. Cell Mol. Life Sci. 54, 766–774. Olsson, I., Venge, P., 1974. Cationic proteins of human granulocytes. II. Separation of the cationic proteins of the granules of leukemic myeloid cells. Blood 44, 235– 246. Painter, J., Merritt, E.A., 2006. Optimal description of a protein structure in terms of multiple groups undergoing TLS motion. Acta Crystallogr. D Biol. Crystallogr. 62, 439–450. Pares, X., Nogues, M.V., de Llorens, R., Cuchillo, C.M., 1991. Structure and function of ribonuclease A binding subsites. Essays Biochem. 26, 89–103. Pulido, D., Moussaoui, M., Andreu, D., Nogués, M., M, T., E, B., 2012. Antimicrobial action and cell agglutination by eosinophil cationic protein is modulated by the cell wall lipopolysaccharide structure. Antimicrob. Agents Chemother. 56, 2378–2385. Raghuraman, A., Mosier, P.D., Desai, U.R., 2006. Finding a needle in a haystack: development of a combinatorial virtual screening approach for identifying high specificity heparin/heparan sulfate sequence(s). J. Med. Chem. 49, 3553–3562. Raines, R.T., 1998. Ribonuclease A. Chem. Rev. 98, 1045–1066. Rosenberg, H.F., Ackerman, S.J., Tenen, D.G., 1989. Human eosinophil cationic protein. Molecular cloning of a cytotoxin and helminthotoxin with ribonuclease activity. J. Exp. Med. 170, 163–176. Rosenberg, H.F., Dyer, K.D., Tiffany, H.L., Gonzalez, M., 1995. Rapid evolution of a unique family of primate ribonuclease genes. Nat. Genet. 10, 219–223. Rubin, J., Zagai, U., Blom, K., Trulson, A., Engstrom, A., Venge, P., 2009. The coding ECP 434(G > C) gene polymorphism determines the cytotoxicity of ECP but has minor effects on fibroblast-mediated gel contraction and no effect on RNase activity. J. Immunol. 183, 445–451. Sanchez, D., Moussaoui, M., Carreras, E., Torrent, M., Nogues, V., Boix, E., 2011. Mapping the eosinophil cationic protein antimicrobial activity by chemical and enzymatic cleavage. Biochimie 93, 331–338. Sanjeev, B.S., Vishveshwara, S., 2004. Conformational transitions in eosinophil cationic protein: a molecular dynamics study in aqueous environment. J. Biomol. Struct. Dyn. 22, 171–182. Sikriwal, D., Seth, D., Batra, J.K., 2009. Role of catalytic and non-catalytic subsite residues in ribonuclease activity of human eosinophil-derived neurotoxin. Biol. Chem. 390, 225–234. Sikriwal, D., Seth, D., Dey, P., Batra, J.K., 2007. Human eosinophil-derived neurotoxin: involvement of a putative non-catalytic phosphate-binding subsite in its catalysis. Mol. Cell Biochem. 303, 175–181. Singh, A., Batra, J.K., 2011. Role of unique basic residues in cytotoxic, antibacterial and antiparasitic activities of human eosinophil cationic protein. Biol. Chem. 392, 337–346. Sorrentino, S., Libonati, M., 1994. Human pancreatic-type and nonpancreatic-type ribonucleases: a direct side-by-side comparison of their catalytic properties. Arch. Biochem. Biophys. 312, 340–348. Swaminathan, G.J., Holloway, D.E., Veluraja, K., Acharya, K.R., 2002. Atomic resolution (0.98 A) structure of eosinophil-derived neurotoxin. Biochemistry 41, 3341–3352. Torrent, M., Nogues, M.V., Boix, E., 2011a. Eosinophil cationic protein (ECP) can bind heparin and other glycosaminoglycans through its RNase active site. J. Mol. Recognit. 24, 90–100. Torrent, M., Navarro, S., Moussaoui, M., Nogues, M.V., Boix, E., 2008. Eosinophil cationic protein high-affinity binding to bacteria-wall lipopolysaccharides and peptidoglycans. Biochemistry 47, 3544–3555. Torrent, M., de la Torre, B.G., Nogues, V.M., Andreu, D., Boix, E., 2009a. Bactericidal and membrane disruption activities of the eosinophil cationic protein are largely retained in an N-terminal fragment. Biochem. J. 421, 425–434. Torrent, M., Sánchez, D., Buzón, V., Nogués, M.V., Cladera, J., Boix, E., 2009b. Comparison of the membrane interaction mechanism of two antimicrobial RNases: RNase 3/ECP and RNase 7. BBA 1788, 1116–1125. Torrent, M., Badia, M., Moussaoui, M., Sanchez, D., Nogues, M.V., Boix, E., 2010. Comparison of human RNase 3 and RNase 7 bactericidal action at the Gramnegative and Gram-positive bacterial cell wall. FEBS J. 277, 1713–1725.

E. Boix et al. / Journal of Structural Biology 179 (2012) 1–9 Torrent, M., Pulido, D., De la Torre, B.G., Garcia de la Torre, J., Nogues, M.V., Bruix, M., Andreu, D., Boix, E., 2011b. Refining the Eosinophil cationic protein antibacterial pharmacophore by rational structure minimization. J. Med. Chem. 54, 7. Torrent, M., Cuyas, E., Carreras, E., Navarro, S., Lopez, O., de la Maza, A., Nogues, M.V., Reshetnyak, Y.K., Boix, E., 2007. Topography studies on the membrane interaction mechanism of the eosinophil cationic protein. Biochemistry 46, 720–733. Ulrich, M., Petre, A., Youhnovski, N., Promm, F., Schirle, M., Schumm, M., Pero, R.S., Doyle, A., Checkel, J., Kita, H., Thiyagarajan, N., Acharya, K.R., SchmidGrendelmeier, P., Simon, H.U., Schwarz, H., Tsutsui, M., Shimokawa, H., Bellon,

9

G., Lee, J.J., Przybylski, M., Doring, G., 2008. Post-translational tyrosine nitration of eosinophil granule toxins mediated by eosinophil peroxidase. J. Biol. Chem. 283, 28629–28640. Venge, P., Bystrom, J., Carlson, M., Hakansson, L., Karawacjzyk, M., Peterson, C., Seveus, L., Trulson, A., 1999. Eosinophil cationic protein (ECP): molecular and biological properties and the use of ECP as a marker of eosinophil activation in disease. Clin. Exp. Allergy 29, 1172–1186. Zhang, J., Rosenberg, H.F., Nei, M., 1998. Positive Darwinian selection after gene duplication in primate ribonuclease genes. Proc. Natl. Acad. Sci. USA 95, 3708– 3713.