Molecular Immunology 66 (2015) 299–309

Contents lists available at ScienceDirect

Molecular Immunology journal homepage: www.elsevier.com/locate/molimm

Evolution of the complement system C3 gene in Antarctic teleosts Daniela Melillo a,1 , Sonia Varriale b,1 , Stefano Giacomelli b , Lenina Natale a , Luca Bargelloni c , Umberto Oreste b , Maria Rosaria Pinto a , Maria Rosaria Coscia b,∗ a b c

Department of Biology and Evolution of Marine Organisms, Stazione Zoologica Anton Dohrn, Villa Comunale, 80121 Napoli (SZN), Italy Institute of Protein Biochemistry, CNR, Via Pietro Castellino 111, 80131 Napoli, Italy Department of Comparative Biomedicine and Food Science, University of Padua, Via Ugo Bassi 58/B, 35131 Padova, Italy

a r t i c l e

i n f o

Article history: Received 21 November 2014 Received in revised form 16 March 2015 Accepted 18 March 2015 Keywords: Antarctic teleost Complement system component C3 Anaphylatoxin Positive evolution Cold adaptation Molecular flexibility

a b s t r a c t Notothenioidei are typical Antarctic teleosts evolved to adapt to the very low temperatures of the Antarctic seas. Aim of the present paper is to investigate sequence and structure of C3, the third component of the complement system of the notothenioid Trematomus bernacchii and Chionodraco hamatus. We determined the complete nucleotide sequence of two C3 isoforms of T. bernacchii and a single C3 isoform of C. hamatus. These sequences were aligned against other homologous teleost sequences to check for the presence of diversifying selection. Evidence for positive selection was observed in the evolutionary lineage of Antarctic teleost C3 sequences, especially in that of C. hamatus, the most recently diverged species. Adaptive selection affected numerous amino acid positions including three residues located in the anaphylatoxin domain. In an attempt to evaluate the link between sequence variants and specific structural features, we constructed molecular models of Antarctic teleost C3s, of their proteolytic fragments C3b and C3a, and of the corresponding molecules of the phylogenetically related temperate species Epinephelus coioides, using human crystallographic structures as templates. Subsequently, we compared dynamic features of these models by molecular dynamics simulations and found that the Antarctic C3s models show higher flexibility, which likely allows for more pronounced movements of both the TED domain in C3b and the carboxyl-terminal region of C3a. As such dynamic features are associated to positively selected sites, it appears that Antarctic teleost C3 molecules positively evolved toward an increased flexibility, to cope with low kinetic energy levels of the Antarctic marine environment. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction The complement system (CS) participates in host protection acting as a first line defensive molecule, promoting local inflammatory reactions, and coordinating the adaptive immune response (Carroll and Isenman, 2012; Ricklin et al., 2010). It consists of over 30 secreted or membrane-bound proteins and can be activated through three different pathways (classical, alternative, and lectin). CS activation pathways converge in the formation of the C3 convertase, which cleaves the third component C3. The

Abbreviations: ␣ NT, ␣ N-terminal tag; ANATO, anaphylatoxin domain; C3, third complement system component; C345C, C3, C4, C5 domain; CS, complement system; CUB, complement C1r/C1s, Uegf, Bmp1; LINCS, linear constraint solver; LNK, linker region; MD, molecular dynamics; MG, ␣2 macroglobulin domain; RACE, rapid amplification of cDNA ends; RMSD, root mean square deviation; RMSF, root mean square fluctuation; SAR, short anchor region; TED, thioester containing domain. ∗ Corresponding author. Tel.: +39 0816132556; fax: +39 0816132277. E-mail address: [email protected] (M.R. Coscia). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.molimm.2015.03.247 0161-5890/© 2015 Elsevier Ltd. All rights reserved.

proteolytic cleavage of C3 generates a small bioactive fragment, the anaphylatoxin C3a, which mediates many biological activities, and a large fragment, C3b, which undergoes an extensive conformational change resulting in the exposure of the thioester group that, in turn, interacts with bacterial matrices (Gros et al., 2008; Law and Dodds, 1997). C3b proteolytic break down, initiated by Factor I, produces multiple fragments; among them iC3b and its portion C3d, are the ligands of complement receptor 2 (CR2) on the B cells surface. Ligand–receptor binding mediates the adaptive immune response (Kalli et al., 1991). The C3 molecular structure has been determined at atomistic level (Janssen et al., 2005) allowing the comprehension of the complex molecular functions exerted by this molecule. A single primary transcript encodes the C3 pro-protein that undergoes relevant posttranslational modifications consisting in the glycosylation at two asparagine residues (Hirani et al., 1986), the loss of the signal peptide, the cleavage in two polypeptide chains (␤ and ␣) with the loss of four arginine residues, the formation of 13 disulfide bridges (Dolmer and Sottrup-Jensen, 1993; Huber et al., 1980), and the formation of the internal thioester bond (Tack et al., 1980).

300

D. Melillo et al. / Molecular Immunology 66 (2015) 299–309

The CS, very ancient from an evolutionary viewpoint, reached the highest level of complexity in vertebrates. The evolution of C3 genes during the vertebrate radiation has been investigated and a positive selection pressure has been proposed to occur early during teleost evolutionary history (Meng et al., 2012). The CS has been investigated in many teleost species. In contrast to mammals, it functions also at low temperature and several serum components exhibit titers drastically higher than in mammals (Boshra et al., 2006). Moreover, C3 is present in different isoforms, irrespectively of the teleost-specific whole genome duplication: two isoforms have been found in Oryzias latipes (Kuroda et al., 2000) and Dicentrarchus labrax (Mauri et al., 2011), eight in Danio rerio (Forn-Cuni et al., 2014) and Oncorhynchus mykiss (Sunyer et al., 1996), five in Sparus aurata (Sunyer et al., 1997) and eight in the tetraploid Cyprinus carpio (Nakao et al., 2000). All the isoforms originated by an event of gene duplication, occurred at the basis of teleost divergence giving raise to two C3 paralogs (Forn-Cuni et al., 2014). C3 isoforms differ in physico-chemical parameters such as electrophoretic mobility and glycosylation patterns, as well as in the sequence of functional sites (Zarkadis et al., 2001); these modifications have been suggested to confer the ability to bind diverse complement-activating surfaces (Sunyer et al., 1998). In this context, cold-adapted teleost species can be considered attractive models for evolutionary studies. Antarctic teleosts, living isolated in a constantly cold environment, belong mainly to the Notothenioidei suborder, which comprises eight families and 120 species (Eastman, 2000). The most recently diverged family, the Channichthyidae, radiated during the most pronounced cooling down period in the late Miocene (Near et al., 2012). In Antarctic teleosts, adaptive evolution to the cold has been proposed for different molecules (Bargelloni et al., 1998; Chen et al., 1997; Detrich et al., 2000; Rizzello et al., 2013), including immune molecules (Coscia et al., 2010, 2011; Varriale et al., 2012). Recently, advances in Antarctic gene analysis were made possible through the availability of sequenced cDNA libraries from the Antarctic notothenioid teleosts Dissostichus mawsoni (Nototheniidae) (Chen et al., 2008), Harpagifer antarcticus (Harpagiferidae) (Thorne et al., 2010), Chionodraco hamatus (Channichthyidae) (Coppe et al., 2013), Chaenocephalus aceratus (Channichthyidae), Notothenia coriiceps and Pleurogramma antarcticum (Nototheniidae) (Shin et al., 2012). In the present paper, we report the identification and sequencing of C3 isoforms of the notothenioid species Trematomus bernacchii and C. hamatus, which have been selected as model species to study Antarctic teleost immune molecules. We analyzed Antarctic C3 genes in comparison with other teleost orthologous genes and found evidence for adaptive evolution. The dynamic properties of Antarctic teleosts, temperate teleosts, and human C3b and C3a molecular models were also analyzed.

2. Materials and methods 2.1. Biological material Fishing activity was performed in Tethys bay (Ross Sea) at 74◦ 41 S, 164◦ 26 E, out of the specially protected areas, during the XXI Italian Antarctic Expedition (2005–2006). The activity permit, released by Italian National Program for Antarctic Research (PNRA), was in agreement with the “Protocol on environmental protection to the Antarctic Treaty” Annex V. Four fish specimens T. bernacchii, family Nototheniidae, and four C. hamatus, family Channichthyidae, both belonging to the suborder Notothenioidei, were collected by use of gill nets or traps and kept in aquaria with running, aerated seawater. The animals were euthanized with anesthetic overdose by immersion for 10 min in

seawater containing 0.1% 2-phenoxyethanol. After collection, tissues were immediately frozen in liquid nitrogen and stored at −80 ◦ C until used. The study was carried out in strict accordance with European and Italian legislation for the care and use of animals for scientific purposes (Directive 86/609/EEC; Decreto Legislativo n. 116/1992). 2.2. RNA isolation and cDNA synthesis Total RNA was prepared from the liver of either a T. bernacchii or a C. hamatus specimen using the SV Total RNA isolation system kit (Promega), according to the manufacturer’s instruction. The quality of the RNA was checked by gel electrophoresis on a 1% agarose gel. Oligo(dT)- and random hexamers-primed single-stranded cDNA was synthesized from liver RNA with the Superscript III First-Strand Synthesis System for RT-PCR (Invitrogen) according to the manufacturer’s instruction. 2.3. T. bernacchii C3-1 gene sequencing Sequencing strategy and primers used are illustrated in Supplementary Fig. S1 and Table S1. The first cDNA fragment of T. bernacchii C3 was obtained by PCR-amplification using a couple of heterologous primers (A forward and A reverse, Table S1) designed on C3 sequences identified in D. mawsoni adult liver library (Chen et al., 2008). A single band (1076 bp) was detected. This product, cloned and verified by sequencing, contained two different sequences, the fragments A and A1 (Fig. S1), named TbC3-1 and TbC3-2, respectively. TbC31 sequence was elongated toward the 5 - and 3 -end using pairs of primers consisting of a specific primer designed in the TbC31 sequenced fragment A and heterologous primers designed in D. mawsoni C3 sequences (B forward and B reverse, and C forward and C reverse, Table S1). This strategy allowed to obtain the fragments B and C, of 2482 and 1809 bp, respectively (Fig. S1). To obtain the full-length sequence, 3 - and 5 -RACE amplifications were performed using specific oligonucleotides designed within the C and B fragments, respectively (primers D and E in Table S1). These amplifications produced the fragments D (730 bp) and E (1243 bp) (Fig. S1). To exclude the possibility that the assembly of A, B, C and D fragments generated a chimeric sequence, a further step of 3 -RACE was carried out using a specific oligonucleotide designed in B fragment (primer F, Table S1). The fragment obtained (F, in Fig. S1), consisting of 2130 bp, agreed completely with the sequence previously obtained by the assembly of B, A, C and D fragments. All the heterologous primers were designed in regions of D. mawsoni C3 transcriptomic sequences exhibiting high sequence identity with other teleost fish C3 sequences. All the PCR and RACE amplified products were cloned into the pCRII-TOPO vector (Invitrogen). Both strands were sequenced using vector- or gene-specific primers on an Automated Capillary Electrophoresis Sequencer 3730 DNA Analyzer (Applied Biosystems). cDNA sequence analysis was routinely checked by running a basic BLAST search of the GenBank database. TbC3-1 gene sequence has been deposited in GenBank under the accession number KF15762. 2.4. T. bernacchii C3-2 and C. hamatus C3 gene sequencing The walking process carried out to determine the nucleotide sequences of the additional T. bernacchii C3 isoform, TbC3-2, and of C. hamatus C3, ChC3, is represented schematically in Fig. S1. Briefly, TbC3-2 fragments A1, B1, and C1 were amplified by PCR using the couple of primers A, B1 and C (Table S1). The 3 - and 5 -end fragments D1 and E1 (Fig. S1) were obtained by the RACE technology using the primers D and E1 (Table S1).

D. Melillo et al. / Molecular Immunology 66 (2015) 299–309

For C. hamatus, a first cDNA fragment of 2485 bp (Fragment G) was obtained by PCR-amplification of random hexamers-primed single-stranded cDNA, using the heterologous primers G forward and reverse (Table S1). To determine sequences at the 3 - and 5 ends of ChC3 mRNA, rapid amplification of cDNA ends (RACE) was carried out using the 3 - and 5 -RACE system (Invitrogen) with the primers H and I (Table S1), respectively. The presence of long overlapping portions in the sequenced fragments allowed the precise reconstruction of the entire TbC32 and ChC3 sequences, minimizing the occurrence of chimeric sequences. The vector used for the amplifications and the sequencing methods were as described for TbC3-1 gene. The obtained gene sequences have been registered in GenBank with the following accession numbers: TbC3-2, KF157663; ChC3, KF157664. 2.5. Computational tools The location of the signal peptide cleavage site in C3 amino acid sequences was predicted using the SignalP 4.1 server at http://www.cbs.dtu.dk/services/SignalP/ (Petersen et al., 2011). NetNGlyc 1.0 server predicted N-glycosylation sites by evaluating the sequence context of NXS/T sequons (http://www.cbs.dtu. dk/services/NetNGlyc/). A threshold of glycosylation potential of 0.5 was used to define the high propensity of NXS/T sequons to be glycosylated. The pI values and the net charges of amino acid sequences were calculated using the ProtParam tool at http:// expasy.org/tools/protparam.html (Gasteiger et al., 2005). Nucleotide sequence identity was calculated by ClustalW2 at http://www.ebi.ac.uk/Tools/clustalw2/index.html (Larkin et al., 2007). Amino acid sequences were aligned using MAFFT and L-INS-I option (online version available at http://mafft.cbrc.jp/alignment/ server/) (Katoh and Standley, 2013) and analyzed by the WebLogo tool (http://weblogo.berkeley.edu/) (Crooks et al., 2004). The amino acid alignment was then transferred into Translator X at http://translatorx.co.uk (Abascal et al., 2010) to guide nucleotide sequence alignment and to obtain codon-based aligned sequences. Sequences with insufficient sequence information were manually discarded. Aligned coding sequences were used to reconstruct a guide tree for further analysis, using PHYML 3.0 (Guindon et al., 2010). Four categories of rates in a gamma distribution as well as fraction of invariant sites were empirically optimized. A general time reversible model was used for nucleotide evolution. The obtained tree topology was employed as user tree option in the analysis of branch- and site-specific codon evolution, which was implemented in the Data Monkey web server (http://www. datamonkey.org/). Random Effects Likelihood (REL) (Kosakovsky Pond et al., 2011), a branch-specific test for positive selection was used to assess the presence of significantly divergent branches in the gene tree. Mixed Effects Model of Evolution (MEME) test (Murrell et al., 2012) was used to assess the presence of codons/positions under positive selection. Likewise, Fast Unconstrained Bayesian Approximation for inferring selection (FUBAR) (http://www.hyphy.org) (Murrell et al., 2013) and REL (Kosakovsky Pond and Frost, 2005) were used to identify positively as well as negatively selected codon positions. 2.6. Homology modeling T. bernacchii, C. hamatus, and Epinephelus coioides C3, C3b and C3a, and C. hamatus C3d models were built by homology modeling using Swiss-Model Workspace (Arnold et al., 2006; Bordoli et al., 2009). C3 models were constructed using as template the human C3 structure 2A73 (Janssen et al., 2006), refined at 3.00 A˚ (mean identity 43%); C3a models were built on the crystal structure of the human C3a anaphylatoxin 4HW5 (Bajic et al., 2013),

301

refined at 2.25 A˚ (mean identity 37%); C3b models were built using human C3b structure 2I07 (Janssen et al., 2006) as template (mean identity 44%); C3d model was built using human C3d structure 1C3D (identity 46%) (Nagar et al., 1998). All models were minimized for 100 ps and validated using WHAT IF package at http:// swift.cmbi.ru.nl/servers/html/ (Vriend, 1990). 2.7. Molecular dynamics simulations All the computer simulations reported in this study were performed using GROMACS 4.5.4 package (Van Der Spoel et al., 2005) and the GROMOS96 force field. The starting models were immersed in boxes containing simple-point-charge water molecules. The dimensions of the box used for C3b models sim˚ those of the box used for C3a ulations were 50.1 × 70.1 × 70.1 A; ˚ No artifacts arising from the models were 9.15 × 8.69 × 16.86 A. contacts between images in the periodic boundary conditions were evident. The ionization state was set to mimic a neutral pH environment. The overall charge of the system was neutralized by adding the appropriate number of ions 7 A˚ further from the protein surface. The energy of the systems was minimized by a 20 ps MD at 300 K. The overall system was then minimized without restraints, before the productive run. All bond lengths were constrained by Linear Constraint Solver (LINCS) algorithm (Berendsen et al., 1981). Newton’s equations of motion were integrated with a time step of 2 fs and atomic coordinates were saved for analyses every 0.5 ps. A cut-off of 14 A˚ was used for the treatment of both electrostatic and Lennard–Jones interactions. All systems were simulated in the isobaric–isothermal ensemble at 300 K, using periodic boundary conditions in the three coordinate directions. The pressure and temperature were controlled using the Berendsen algorithm at 1 bar with a coupling constant tp = 1 ps at 300 K. All the trajectories were estimated to assess the quality of the simulations using GROMACS routines and in-house programs. The stability of the simulations was monitored by following a number of parameters such as root mean square deviation (RMSD) from the starting structure, radius of gyration, accessible surface area, number of hydrogen bonds. The flexibility of each C␣ atom was checked by determining the root mean square fluctuation (RMSF) with respect to the mean coordinates. The simulations showed stable trajectories and reached a plateau after a few nanoseconds. The electrostatic energy of each model was calculated by APBS version 0.5.1 (Baker et al., 2001). The simulation time of TbC3-1b model was 14 ns, that of E. coioides was 10 ns; TbC3-1a and E. coioides C3a simulations were carried out for 10 ns. The models deduced by the crystallographic data of human C3b and C3a required a shorter time to reach the equilibrium, and the simulations were stopped after 5 ns. Each simulation was repeated for a shorter time (4 ns) to ascertain that no different equilibrium state was reached. 3. Results and discussion 3.1. Analysis of Antarctic teleost C3 sequences Using the sequencing strategy described in Section 2 and summarized in Supplementary Fig. S1, we determined the complete nucleotide sequences of transcripts encoding two different isoforms of T. bernacchii C3, TbC3-1 and TbC3-2, and one of C. hamatus C3, ChC3. The coding region of TbC3-1 and TbC3-2 consisted of 4971 nt, whereas that of ChC3 included 4980 nt. The respective 3 UTRs were 131, 138 and 140 nt long, and all contained the less common polyadenylation signal ATTAAA. The full-length nucleotide sequences of TbC3-1, TbC3-2 and ChC3 are reported in the

302

D. Melillo et al. / Molecular Immunology 66 (2015) 299–309

Supplementary Figs. S2, S3, and S4, respectively. TbC3-1 and TbC32 shared 94.7% of nucleotide identity; ChC3 was more similar to TbC3-2 (93.3% identity) than to TbC3-1 (92.4% identity). To evaluate the multiplicity of the C3 isoforms in T. bernacchii, we sequenced 21 out of many clones obtained by the 3 RACE reaction, carried out using total RNA extracted from a single specimen. We recovered 10 different nucleotide sequences encoding the C3 C-terminal region. The allelic fragments possibly belong to at least five distinct genes sharing a high degree of nucleotide identity (91.2–98.5%) in the determined nucleotide region (sequences not shown). This finding is not unusual because multiple C3 isoforms have been reported in many teleost species (Nakao et al., 2006) and seem to arise by lineage-specific duplications (Meng et al., 2012). We recognized in the deduced amino acid sequence of TbC31 pro-protein a 23 amino acid residues long leader peptide, the ␤ chain (between residues 24 and 658) followed by the tetraarginine motif, site of post-translational processing, and the ␣ chain consisting of 994 amino acid residues (Supplementary Fig. S2). A search for conserved protein domains revealed that all the Antarctic C3 pro-proteins sequences reproduced the complete intricate arrangement of the mammalian C3 pro-proteins (Fredslund et al., 2006; Janssen and Gros, 2006; Janssen et al., 2005; Tack et al., 1980). Eight domains (MG1-MG8) were of ␣2-macroglobulin type; MG1MG5 belonged to the ␤ chain, MG7 and MG8 were in the ␣ chain whereas MG6 was shaped by residues of both the ␣ and ␤ chain. The linker region (LNK), the anaphylatoxin domain (ANATO) and the ␣ N-terminal tag (␣ NT) were inserted in the MG6 sequence. Two domains were present between MG7 and MG8: the CUB domain (Bork and Beckmann, 1993) and the thioester containing domain (TED) (Law and Dodds, 1997). The C3C4C5 (C345C) domain (Bramham et al., 2005) linked the C-terminal end of MG8 through the short anchor region (SAR) including an intra-chain disulfide bridge. The theoretical pI value of the Antarctic species C3 ranged between 5.85 and 5.92, which are values lower than human C3 (6.02) and non-Antarctic teleost C3 (mean 6.53); the net charge of Antarctic C3 was on average −23, lower than human (−18) and other teleost (−8.2 in mean) C3s (Table S2). By analyzing the alignment of TbC3-1, TbC3-2 and ChC3 with human C3, the cysteine residues were found in the same positions as the disulfide bridges localized in human. However, the two cysteines that form one of the two bridges of the human MG8 domain were absent in the Antarctic sequences (Fig. 1). The lack of this disulfide bond probably increases the MG8 domain flexibility that is crucial in the C3 to C3b conformational change, implying a ␤–␣–␤/␤–␣–␣ transition (Janssen and Gros, 2007). On the other hand, two additional cysteine residues were present in the MG1 domain of the Antarctic teleosts. On the basis of the molecular models built for the Antarctic species C3, these cysteine residues are possibly involved in the formation of a unique intra-␤ chain disulfide bond. Localization of the disulfide bridges in human and Antarctic species C3 is reported in Table S3. In Antarctic species C3, the N-linked glycosylation sites differ from those determined in human C3 (Crispin et al., 2004; Hirani et al., 1986; Janssen and Gros, 2006) located in MG1 (␤ chain) and at the beginning of the first CUB segment (␣ chain). In human C3, it has been suggested that the asparagine residue of the CUB domain (Asn939 in Fig. 1) is involved in the correct folding of the molecule and, after cleavage of C3b by factor I, it shifts from buried to exposed to the solvent (Crispin et al., 2004). In Antarctic C3 sequences at the position corresponding to human Asn939 residue there was a gap (Fig. 1). Four Asn-Xaa-Ser/Thr sequons were detected in TbC3-1 and ChC3, and six in TbC3-2, however high glycosylation propensity scores were found only for two sites of TbC3-1, for three of ChC3 and for four of TbC3-2. The putative glycosylated domains were ANATO (TbC3-1 and TbC3-2), MG7 (ChC3), MG8 (TbC3-1, TbC3-2

and ChC3), and the second CUB segment (TbC3-2, ChC3) (Table S4). It is to be noted that while the glycosylation sites in MG7, CUBII and MG8 are shared by many teleost species, no glycosylation sites have been detected in ANATO of non-Antarctic teleost species. The teleost glycosylation site in CUB II could substitute that present in mammalian CUB I. TbC3-2 and ChC3 lack the catalytic histidine, crucial for the cleavage of the thioester bond. In fact, the cysteinyl sulfhydryl and the glutaminyl ␥-carbonyl of the motif CGEQ, present in the TED, form a thioester bond, which closes a thiolactone ring, essential for the binding to hydroxyl or amino groups present on bacterial walls (Isaac and Isenman, 1992). In humans this reaction is catalyzed by a histidine residue, which participates in a nucleophilic reaction forming an intermediate with an intramolecular acyl-imidazole bond (Law and Dodds, 1997). It should be noticed that C4, the other human CS thioester containing protein component, presents two isotypes, C4A and C4B, characterized by the absence or presence of the catalytic histidine, respectively (Dodds et al., 1996). The lack of the catalytic histidine is not unusual in the teleost C3 molecule. In fact, it is absent in the trout C3-3 and C3-4 isoforms (Zarkadis et al., 2001), in three out of five C. carpio C3 isoforms (Nakao et al., 2000) and in O. latipes C3-2 (Kuroda et al., 2000). The catalytic histidine is replaced by various amino acid residues thus suggesting different molecular mechanisms of interaction of the C3b fragments with the bacterial wall. In summary, it emerges from the sequence analysis of Antarctic C3s that although their domain architecture reproduces that of mammals, there are significant amino acid differences affecting several physico-chemical parameters. The lower pI values probably account for the need to increase the solubility at low temperature of a molecule occurring at a rather high concentration in the teleost serum. The lack of one disulfide bridge in MG8 could favor the large conformational transition of the active C3b fragment; the presence of a larger number of N-glycosylation sites could increase the Antarctic C3 hydrophilic character and, in turn, its solubility.

3.2. Analysis of Antarctic teleost C3a sequences Particular attention was devoted to the ANATO domain contained in the C3a fragment generated by the C3 activation. The percentage of identity between human C3a and each Antarctic sequence ranged between 33.3 and 35.9. A WebLogo representation of the alignment of the Antarctic sequences with other notothenioid species sequences, deduced by single contigs of C. aceratus, H. antarcticus, N. coriiceps and P. antarcticum transcriptomes, is shown in Supplementary Fig. S5A. In the same figure, the alignments of 10 non-Antarctic teleosts (Supplementary Fig. S5B) and 10 mammalian species (Supplementary Fig. S5C) are also shown. Only 12 residues were shared by all examined sequences, including the six cysteines involved in the three canonical disulfide bonds, the Cterminal motif LAR, and the residues Gly27 (26 in mammals), Arg40 (39 in mammals) and Phe54 (53 in mammals). Amino acid residues exclusively shared by the Antarctic teleost sequences were Asn14, Val41, and Val70. The alignment of the selected C3a sequences is reported in Supplementary Fig. S6. A peculiar feature of several Antarctic teleost was the presence of a N-glycosylation site at the amino-terminal region. The theoretical pI values of TbC3-1a, TbC3-2a and ChC3a were 5.23, 4.86 and 5.60, respectively, similar to those of other teleost C3a, but lower than human (9.69) and other mammalian C3a (ranging between 8.26 and 9.84); the ionic character of teleost C3a anaphylatoxins (net charge ranging between −2 and −4) is opposite to that of mammalian ones (net charge ranging between +2 and +10). In particular, in the last 12 C-terminal residues, teleost C3a

D. Melillo et al. / Molecular Immunology 66 (2015) 299–309

303

Fig. 1. Clustal X alignment of Antarctic and human C3 pro-protein amino acid sequences. Numbers on the right refer to the amino acid positions. Structural domains (shaded in gray) and defined regions are indicated at the top of the amino acid sequence (MG1-8: ␣2 macroglobulin domain 1–8; ANATO: anaphylatoxin domain; LNK: linker region; ␣ NT: amino-terminal region of the ␣ chain; CUB: C1r/C1s, Uegf, Bmp1 domain; TED: thioester containing domain; C345C: C3C4C5 domain). The thioester motif and the catalytic histidine, contained in the TED domain, are in bold. Cysteine residues marked by the same number in bold, form disulfide bridges. The sequons NXS/T that represent potential glycosylation sites are underlined.

contains three negative charged residues absent in the mammalian C3a (Supplementary Fig. S5B). Antarctic teleost sequences seem to be quite different from nonAntarctic ones because they lack one negatively charged amino acid within the last 12 C-terminal residues (Supplementary Fig. S5A). The different charge distribution at the C-terminus in the teleost and mammalian sequences raises questions about the mechanism of interaction of teleost C3a with the specific C3a receptor (C3aR). The positive charges of mammalian C3a C-termini have been suggested to interact with negatively charged amino acid residues, localized at the basis of the second extracellular loop of the specific C3aR (Chao et al., 1999; Wetsel et al., 2000). The remarkable differences found between human and teleost sequences in the second extracellular loop of the C3aR (Boshra et al., 2005) could account

for a different ligand–receptor interaction mechanism. The inability of teleost C3a to promote chemotaxis could be associated with the failure to bind to the specific C3aR as in the mammalian model. 3.3. Phylogenetic analysis We constructed a data set of full-length teleost C3-like sequences, from 16 different teleost species. A C3 gene tree, reconstructed using maximum likelihood method, is presented in Fig. 2. A branch-specific test for positive selection (REL), which uses an empirical Bayesian approach, was used to assess the presence of significantly divergent branches in the gene tree. This analysis detected a pervasive action of positive selection on teleost C3, with 17 branches in the gene tree (Supplementary Table S5)

304

D. Melillo et al. / Molecular Immunology 66 (2015) 299–309

Fig. 2. Maximum likelihood tree based on codon-alignment of teleosts C3. Internal branches where episodic positive selection was inferred are indicated with numbers. For terminal branches undergoing positive selection, sequence names are in bold.

showing significant evidence for a set of positions with the ratio between non-synonymous and synonymous substitutions greater than 1. With regard to Antarctic teleosts, episodes of positive selection were inferred in the branch of C. hamatus C3, T. bernacchii

C3-1, and in the lineage leading to C. hamatus C3 and T. bernacchii C3-2. Three different approaches were used to assess the presence of codons/positions under positive selection. Using MEME, FUBAR,

Fig. 3. Ribbon representation of the structural domain arrangement of TbC3-1 (A), TbC3-1b (B), and TbC3-1a (C) models. Each domain is marked by a different color. Cysteine residues involved in TbC3-1 disulfide bonds are in yellow; the asparagine residues in the context NXT are in red. In (C), the disulfide bridges are in yellow. Abbreviations are as follows: MG1-8: ␣2 macroglobulin domain 1–8; ANA: anaphylatoxin domain; CUB: C1r/C1s, Uegf, Bmp1 domain; TED: thioester containing domain; C345C: C3C4C5 domain; SAR: short anchor region; LNK: linker region. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

D. Melillo et al. / Molecular Immunology 66 (2015) 299–309

and REL tests, the number of positively selected residues was found to be different in TbC3-1, TbC3-2 and ChC3 sequences, being 23, 28, 37, respectively, and indicating that the sequence of the most recently diverged species, C. hamatus, underwent a higher selective pressure. The highest percent of selected residues was found in MG4, ANA and TED domains (Supplementary Table S6). In particular six positively selected residues have been found in the TED of TbC3-1, 10 in that of TbC3-2 and 15 in that of ChC3. The positively selected residues of TbC3-1, TbC3-2 and ChC3 (listed in Supplementary Tables S7–S9) were compared to the corresponding residues of the E. coioides C3 sequence. The latter was chosen for comparison because it represented one of the closest species to the notothenioid group in the phylogenetic tree, without evidence of divergent selection as the case of Anarhichas minor and Gasterosteus aculeatus. Furthermore we analyzed in more details the different character of the positively selected residues of the TED. In all the Antarctic molecules we found an increase of the positive charges: in fact positive charges were introduced or negative charges were removed. We calculated the net charge variation due to the positive selection and found the values of +2.5, +4.5 and +4.5 for the TED of TbC3-1, TbC3-2 and ChC3, respectively. Negatively selected codons were identified across the entire tree using REL and FUBAR approaches. Both methods revealed a large number of positions under purifying selection: 719 (FUBAR) and 749 (REL) out of 1870 analyzed codons using a conservative threshold (p < 0.99). The distribution of negatively selected positions in the different C3 structural regions is reported in Supplementary Table S10. They are most frequent in MG7 and MG4 domains. These protein regions are fundamental for properly maintaining the MG ring and the ␣ chain assembly, respectively. The regions containing a lower percentage of sites under purifying selection were located in the unstructured regions, such as the ␣ NT, the leader peptide and the LNK. Overall, the phylogenetic analysis showed putative evidence for positive selection especially in the most recently diverged species C.

Fig. 4. Thioester site activation. Residues involved in the thioester site activation in TbC3-1 (A) and TbC3-2 (B) are indicated. Numbering is in agreement with Supplementary Figs. S2 and S3.

305

hamatus, whereas structurally relevant regions seemed to be under strong evolutionary constraints. 3.4. Antarctic teleost C3, C3b, C3a and C3d models To link amino acid residues putatively under positive selection in each domain fold with protein function, we built, by Homology Modeling, models of the Antarctic and E. coioides C3, C3b and C3a. E. coioides was selected for comparative analysis because it belongs to the Perciformes order, the same as the Antarctic species. Furthermore, as indicated by evolutionary analyses, E. coioides C3 does not show evidence for divergent selection. The Homology Modeling method, although less informative than experimental biophysical methods, is appropriate to study molecules exhibiting sequence homology with available template structures. A representation of the TbC3-1, TbC3-1b and TbC3-1a models is reported in Fig. 3A, B, and C, respectively.

Fig. 5. ChC3d model. (A) Ribbon representation of the ChC3d. The positively selected residues are shown: in blue those increasing the positive net charge and in red that introducing a negative charge; the residues of the thioester site are in yellow. The ChC3d surface electrostatic potential is shown in (B) (convex side) and (C) (concave side). The blue and red color intensity indicates the positive and negative charges, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

306

D. Melillo et al. / Molecular Immunology 66 (2015) 299–309

Fig. 6. Potential movement of the TbC3-1b TED domain. The movement of the TED domain with respect to the CUB is shown in the left. Arrows indicate the amplitude of the thioester site movement. The interactions between the TED and MG1 domains limiting the extension of the TED movement are shown in the right.

The amino acid residues encoded by positively selected codons were identified in the structure and the 62.5% of them were found exposed to the solvent; in many cases polar residues have been introduced to replace hydrophobic ones. The presence of such residues may indicate that under evolutionary pressure the superficial hydrophilicity of Antarctic C3 increased to render more soluble the molecule at low temperature. In an attempt to investigate the differences resulting from the absence of the catalytic histidine, we directed our attention to the thioester site of TbC3-2. In the model deduced by the X-ray diffraction datasets of human C3 crystal, the distance between the pyrrolic ring of the catalytic histidine residue and the ␥-carbonyl of the glutamyl residue of the thioester site, which allows the formation of an ˚ intermediate by an intramolecular acyl-imidazole bond, is 7.90 A. In TbC3-1, the distance between Gln␣346 and the catalytic His␣459 was 5.92 A˚ (Fig. 4A). In TbC3-2, lacking the catalytic histidine, we suggest that the catalytic role could be exerted by the Tyr␣402, whose phenolic oxygen is 6.43 A˚ distant from the ␥-carbonyl of Gln␣346 (Fig. 4B). Analysis of other teleost C3 sequences revealed that in four other sequences, lacking the catalytic histidine, a tyrosine residue was localized at the same position as Tyr␣402, whereas in all the sequences having the catalytic histidine, the tyrosine residue was never found at that position. Our observations suggest that the C3 molecules lacking the catalytic histidine utilize the ionizable oxygen of a tyrosine residue to generate the catalytic intermediate. This hypothesis is in agreement with previously reported results, obtained by site-directed mutagenesis of the catalytic histidine of human C4B, indicating that, among several amino acids tested, tyrosine was the residue allowing the most efficient covalent binding of the thioester to glycerol (Ren et al., 1995). TbC3-1b model differs extensively from that of TbC3-1 (Fig. 3B). It is well known that C3 proteolysis, mediated by the C3 convertase, produces a dramatic conformational change, which essentially concerns the TED that rotates and translates; in parallel, CUB loses its ␤-sheet structure and stretches in ␣-helix and several ␤-turns

(Ajees et al., 2006). Consequently, MG3, MG8, and ␣ NT are destabilized. As a result of this rearrangement, the chemical reactive groups of the thioester bond become exposed (Gros et al., 2008). In this context, the high flexibility of the MG8 domain allows radical changes in structure (Janssen and Gros, 2007). A ribbon representation of TbC3-1b model is reported in Fig. 3B. The C3d fragment resulting from C3b proteolysis consists essentially of the TED; the surface of its molecular structure presents two opposite sides, one concave and one convex. A close relationship has been suggested to exist between this structural organization and the dual functional role exerted by this fragment. In fact, while the C3d concave side interacts with the CR2 located on the B cells surface (Isenman et al., 2010; Mohan et al., 2015), triggering the adaptive immune response, its convex side, carrying the thioester site, induces the covalent binding to the pathogen surface. Analysis of localization of positively selected residues indicated that the majority of those present in Antarctic C3d were crowded in the close proximity of the thioester site, with their side chains sticking out. ChC3d structure, containing the largest number of positively selected sites, is shown in Fig. 5A. We found that the positive selection introduced positive charges in C3d. The analysis of the surface electrostatic potential of the model revealed that these charges are positioned on the convex surface of the ChC3d containing the thioester site (Fig. 5B), being the concave side essentially negatively charged (Fig. 5C); in this side, we found the single negative charge introduced by the positive selection. The distribution of the selected charged residues on Antarctic C3d suggests that a number of residues underwent evolutionary pressure to modify the electrostatic potential of the molecular surface. The influence of the evolution on the C3d electric charges has been already described by Kieslich and Morikis (2012). A possible explanation might be that Antarctic fish C3b needs to cope with specific surface components of psychrophilic bacteria whose exopolysaccharides are more abundant and show high content of uronic acid and other organic and inorganic acidic molecules, which contribute a negative charge to the overall polymer (Nichols et al., 2005; Poli et al., 2010).

D. Melillo et al. / Molecular Immunology 66 (2015) 299–309

Molecular models of TbC3-1a and E. coioides C3a were constructed using the human C3a crystallographic structure (4HW5) as template. The structure consisted in a bundle comprising four helices connected by three disulfide bridges (Fig. 3C). In mammals, the C3a disulfide bridges are unusually labile and suggest in vivo disulfide interchange (Chang et al., 2008). Mammalian C3a amphipathic C-terminal region is thought to regulate the biological functions of the anaphylatoxin being the site of interaction with the C3a receptor (Chao et al., 1999; Ember et al., 1998; Klos et al., 2013). The last C-terminal residue, arginine, is particularly interesting because its cleavage by the carboxypeptidase B generates C3a desArg, which is functionally inactive; however, structural differences between human C3a and C3a desArg have not been evidenced (Bajic et al., 2013). In addition, we found that the putative glycosylation site identified on the Antarctic C3a molecular structures was accessible to the solvent and far from the C-terminus involved in mammals in the interaction with the C3a receptor. 3.5. Dynamic properties of C3b models The dynamic properties of the TED of TbC3-1b were analyzed in comparison with those of the corresponding domains of E. coioides and human C3b by performing MD simulations. The resulting RMSD plots are shown in Supplementary Figs. S7–S9. The physical parameters of each structure, calculated at the equilibrium, are shown as average values in Supplementary Table S11. The secondary structures were maintained during the simulation time (Supplementary Table S12). By the analysis of the MD trajectory at the equilibrium, we identified a potential movement of TED with respect to the CUB domain (Fig. 6), more pronounced in Antarctic C3bs. In fact, the C␣ atoms shifted within the two most diverging states on average by 9.41 A˚ in the TbC3-1b, and 3.32 A˚ and 2.52 A˚ in the E. coioides and human molecules, respectively. The extension of the mobility varied in the different C␣ atoms of the TED domain; the largest mobility was found to be associated to C␣ atoms of the thioester site, differing by 10.59 A˚ in the TbC3-1b. It could be speculated that the high motion of the C3b thioester site enables sensing the bacterial surface and helps in finding the best orientation to perform the covalent binding. In contrast, the less mobile residues ˚ were Glu␣325 and Lys␣326 oscillating by only 3.80 and 3.75 A, respectively. It seems likely that these residues, involved in polar bonds with the residues Lys␤101 and Asp␤102 located in the MG1 domain, function as a brake, thus modulating the TED movement (Fig. 6). In addition to the polar interactions, hydrophobic contacts link the Phe␣372 in TED to two residues in the ␤ chain, Tyr␤108 in MG1 and Tyr␤648 in LNK, contributing to limit the TED movement (Fig. 6B). Similar contacts, not found in the human C3b, suggest the parallel increase of the TED movement and its regulation during evolution of Antarctic C3s.

307

core of the molecule. All the molecular structures shared the same core consisting of three helices, H1, H2, and H3 (Fig. 7A), stabilized by three disulfide bonds; a large loop connected the second and third helix. The total helix content was found to be different among the structures of Antarctic, non-Antarctic and mammalian models (Supplementary Table S13). In particular, the H3 length consisted of 16, 19, and 26 residues in T. bernacchii, E. coioides, and human, respectively. A fourth short helix (H4) was found only in TbC3-1a; it was located in the C-terminal region and included seven amino acid residues that, in human C3a anaphylatoxin, have been supposed to interact with C3aR. In the human C3a, the presence of Arg65 interacting with i-3 Glu62 allows the helix H3 to extend up to the Leu73 residue. Indeed, in the T. bernacchii sequence the loss of the Arg65 broke the helix; accordingly, the Ala61 adopted an elbow role permitting the unison movement of the terminal helix. Other relevant differences concern the superficial electrostatic potential. It is noteworthy that TbC3-1a superficial electrostatic potential (3.02 × 104 kJ) is higher than the entire TbC3-1 molecule (1.74 × 104 kJ) and TbC3-1b, (1.52 × 104 kJ). The distribution of positive and negative charges on the surface of T. bernacchii, E. coioides and Homo sapiens C3a is evident in Fig. 7B. We tested also the flexibility of the C␣ of each residue by calculating the RMSF of the C␣ atoms at the equilibrium. RMSF is an

3.6. C3a models dynamics We performed MD simulations of T. bernacchii and E. coioides C3a models, and of the crystallographic structure of human C3a. The resulting RMSD plots (Supplementary Figs. S10–S12) showed that all the structures reached the equilibrium. The physical parameters of each structure at the equilibrium are shown as average values in Supplementary Table S13. Starting from the four helices structures, all the examined models after MD simulations became less compact because of the unfolding of the third helix, which changed in a loop. However, the collected data indicated that the shape of TbC3-1a was different from that of non-Antarctic species (Fig. 7). In fact, TbC3-1a structure presented the C-terminal region oriented outwards, whereas the corresponding C3a region of E. coioides was back-faced toward the

Fig. 7. Analysis of the C3a models. (A) The models of C3a of T. bernacchii C3-1 (top), E. coioides (center) and H. sapiens (bottom), as mean of eight structures at the equilibrium after MD simulations, are shown. The helices are numbered starting from the N-terminus. (B) Charge distribution on the C3a surface of T. bernacchii C3-1 (top), E. coioides (center) and H. sapiens (bottom) is shown. The blue and red color intensity indicates the positive and negative charges, respectively. In (C), the amplitude of the swinging movement of the C-terminal regions of TbC3-1a (top), E. coioides (center) and human C3a (bottom) is compared. For each species, the two most divergent models with respect to the average are superimposed, and the calculated distances between C␣ atoms of the Arg77 are indicated by the double pointed arrow. The position of the Ala 61, which assumes in T. bernacchii an elbow function, is indicated by an arrow. It should be noted that the C-terminal regions have different orientations: outward in T. bernacchii and human, and inward in E. coioides. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

308

D. Melillo et al. / Molecular Immunology 66 (2015) 299–309

index that indicates the intensity of fluctuation of each residue with respect to the average structure. The RMSF profiles of T. bernacchii, E. coioides and H. sapiens, were calculated using a 5 ns window and are reported in Supplementary Fig. S13. The plots clearly indicate a higher flexibility of the C-terminal region in TbC3-1a. By analyzing the MD trajectory at the equilibrium we identified a possible swinging movement of the C-terminal tail. We calculated the swinging amplitude of the 16 residues long Cterminal region, which resulted in 12.55 A˚ in TbC3-1a, but 4.47 A˚ in E. coioides and only 0.91 A˚ in human C3a (Fig. 7C). The large extension of the movement of the C-terminal region, together with its outward orientation with respect to the core of the peptide, and its helical structure that preserves the putative functional site, support the opinion that the Antarctic molecule gained an increased functionality. In fact, the ability of the C3a carboxylterminal sequence to assume different orientations could certainly facilitate the encounter with its specific receptor on the cell membrane. 4. Conclusions Our study on Antarctic teleost C3s further improved the comprehension of the evolution of the CS. In particular, comparative analysis of C3 amino acid sequences, in addition to the identification of the most important and canonical molecular signatures in Antarctic species, confirmed the differences between teleost and mammalian C3 sequences in terms of both disulfide bonds and putative N-glycosylation sites. Moreover, this analysis emphasized the lower pI of teleost C3a and suggested that in teleost C3 sequences lacking the catalytic histidine, a tyrosine residue could play a similar role. By combining sequential, structural and dynamic data we hypothesize that the structural features of Antarctic teleost C3, especially those at functional sites, have been modified under evolutionary pressure. Such evidence could be explained by the need to increase the solubility at the extremely low temperature of the Antarctic seas, and to enhance the molecular flexibility in an environment characterized by low kinetic energy levels. High flexibility, in turn, could facilitate the conformational changes particularly crucial for C3 and its proteolytic fragments, C3a and C3b. More interestingly, we demonstrated in Antarctic models an amplified swinging movement of the TED domain in C3b and an extended mobility of the C-terminal end of C3a. Our results added weight to the idea that the cold adaptive evolution modified genes to produce more flexible molecules. Recent data obtained by MD simulations of psychrophilic microorganism enzymes, support this hypothesis (Giordano et al., 2013; Mohamad Ali et al., 2013; Sigtryggsdottir et al., 2014). The present work has approached the topic of the cold adaptive evolution by analyzing a very large molecule, whose flexibility is crucial to exert its complex functions and found a strong relationship between positively selected sites and higher flexibility of functional sites. Funding This work has been supported by the Italian National Program for Antarctic Research (PNRA), grant PEA2009/A1.12. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. 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.molimm.2015. 03.247

References Abascal, F., Zardoya, R., Telford, M.J., 2010. TranslatorX: multiple alignment of nucleotide sequences guided by amino acid translations. Nucleic Acids Res. 38, W7–W13. Ajees, A.A., Gunasekaran, K., Volanakis, J.E., Narayana, S.V.L., Kotwal, G.J., Murthy, H.M.K., 2006. The structure of complement C3b provides insights into complement activation and regulation. Nature 444, 221–225. Arnold, K., Bordoli, L., Kopp, J., Schwede, T., 2006. The SWISS-MODEL workspace: a web-based environment for protein structure homology modelling. Bioinformatics 22, 195–201. Bajic, G., Yatime, L., Klos, A., Andersen, G.R., 2013. Human C3a and C3a desArg anaphylatoxins have conserved structures, in contrast to C5a and C5a desArg. Protein Sci. 22, 204–212. Baker, N.A., Sept, D., Joseph, S., Holst, M.J., McCammon, J.A., 2001. Electrostatics of nanosystems: application to microtubules and the ribosome. Proc. Natl. Acad. Sci. U. S. A. 98, 10037–10041. Bargelloni, L., Marcato, S., Patarnello, T., 1998. Antarctic fish hemoglobins: evidence for adaptive evolution at subzero temperature. Proc. Natl. Acad. Sci. U. S. A. 95, 8670–8675. Berendsen, H.J.C., Postma, J.P.M., van Gusteren, W.F., Hermans, J., 1981. Interaction models for water in relation to protein hydration. In: Pullman, B. (Ed.), Intermolecular Forces. Reidel Publishing Company, Dordrecht, The Netherlands, pp. 331–342. Bordoli, L., Kiefer, F., Arnold, K., Benkert, P., Battey, J., Schwede, T., 2009. Protein structure homology modeling using SWISS-MODEL workspace. Nat. Protoc. 4, 1–13. Bork, P., Beckmann, G., 1993. The cub domain – a widespread module in developmentally-regulated proteins. J. Mol. Biol. 231, 539–545. Boshra, H., Li, J., Sunyer, J.O., 2006. Recent advances on the complement system of teleost fish. Fish Shellfish Immunol. 20, 239–262. Boshra, H., Wang, T., Hove-Madsen, L., Hansen, J., Li, J., Matlapudi, A., Secombes, C.J., Tort, L., Sunyer, J.O., 2005. Characterization of a C3a receptor in rainbow trout and Xenopus: the first identification of C3a receptors in nonmammalian species. J. Immunol. 175, 2427–2437. Bramham, J., Thai, C.T., Soares, D.C., Uhrin, D., Ogata, R.T., Barlow, P.N., 2005. Functional insights from the structure of the multifunctional C345C domain of C5 of complement. J. Biol. Chem. 280, 10636–10645. Carroll, M.C., Isenman, D.E., 2012. Regulation of humoral immunity by complement. Immunity 37, 199–207. Chang, J.Y., Lin, C.C.J., Salamanca, S., Pangburn, M.K., Wetsel, R.A., 2008. Denaturation and unfolding of human anaphylatoxin C3a: an unusually low covalent stability of its native disulfide bonds. Arch. Biochem. Biophys. 480, 104–110. Chao, T.H., Ember, J.A., Ember, J.A., Wang, M.Y., Bayon, Y., Hugli, T.E., Ye, R.D., 1999. Role of the second extracellular loop of human C3a receptor in agonist binding and receptor function. J. Biol. Chem. 274, 9721–9728. Chen, L.B., DeVries, A.L., Cheng, C.H.C., 1997. Evolution of antifreeze glycoprotein gene from a trypsinogen gene in Antarctic notothenioid fish. Proc. Natl. Acad. Sci. U. S. A. 94, 3811–3816. Chen, Z., Cheng, C.H., Zhang, J., Cao, L., Chen, L., Zhou, L., Jin, Y., Ye, H., Deng, C., Dai, Z., Xu, Q., Hu, P., Sun, S., Shen, Y., 2008. Transcriptomic and genomic evolution under constant cold in Antarctic notothenioid fish. Proc. Natl. Acad. Sci. U. S. A. 105, 12944–12949. Coppe, A., Agostini, C., Marino, I.A., Zane, L., Bargelloni, L., Bortoluzzi, S., Patarnello, T., 2013. Genome evolution in the cold: Antarctic icefish muscle transcriptome reveals selective duplications increasing mitochondrial function. Genome Biol. Evol. 5, 45–60. Coscia, M.R., Varriale, S., De Santi, C., Giacomelli, S., Oreste, U., 2010. Evolution of the Antarctic teleost immunoglobulin heavy chain gene. Mol. Phylogenet. Evol. 55, 226–233. Coscia, M.R., Varriale, S., Giacomelli, S., Oreste, U., 2011. Antarctic teleost immunoglobulins: more extreme, more interesting. Fish Shellfish Immunol. 31, 688–696. Crispin, M.D., Ritchie, G.E., Critchley, A.J., Morgan, B.P., Wilson, I.A., Dwek, R.A., Sim, R.B., Rudd, P.M., 2004. Monoglucosylated glycans in the secreted human complement component C3: implications for protein biosynthesis and structure. FEBS Lett. 566, 270–274. Crooks, G.E., Hon, G., Chandonia, J.M., Brenner, S.E., 2004. WebLogo: a sequence logo generator. Genome Res. 14, 1188–1190. Detrich, H.W., Parker, S.K., Williams, R.C., Nogales, E., Downing, K.H., 2000. Cold adaptation of microtubule assembly and dynamics – structural interpretation of primary sequence changes present in the alpha- and beta-tubulins of antarctic fishes. J. Biol. Chem. 275, 37038–37047. Dodds, A.W., Ren, X.D., Willis, A.C., Law, S.K., 1996. The reaction mechanism of the internal thioester in the human complement component C4. Nature 379, 177–179. Dolmer, K., Sottrup-Jensen, L., 1993. Disulfide bridges in human complement component C3b. FEBS Lett. 315, 85–90. Eastman, J.T., 2000. Antarctic notothenioid fishes as subjects for research in evolutionary biology. Antarct. Sci. 12, 276–287. Ember, J.A., Jagels, M.A., Hugli, T.E., 1998. Characterization of complement anaphylatoxins and their biological responses. In: Volanakis, J.E., Frank, M.M. (Eds.), The Human Complement System in Health and Disease. Marcel Dekker, New York. Forn-Cuni, G., Reis, E.S., Dios, S., Posada, D., Lambris, J.D., Figueras, A., Novoa, B., 2014. The evolution and appearance of C3 duplications in fish originate an exclusive teleost C3 gene form with anti-inflammatory activity. PLOS ONE 9, e99673.

D. Melillo et al. / Molecular Immunology 66 (2015) 299–309 Fredslund, F., Jenner, L., Husted, U.B., Nyborg, J., Andersen, G.R., Sottrup-Jensen, L., 2006. The structure of bovine complement component 3 reveals the basis for thioester function. J. Mol. Biol. 361, 115–127. Gasteiger, E., Hoogland, C., Gattiker, A., Duvaud, S., Wilkins, M.R., Appel, R.D., Bairoch, A., 2005. Protein identification and analysis tools on the ExPASy server. In: Walker, J.M. (Ed.), The Proteomics Protocols Handbook. Humana Press, pp. 571–607. Giordano, D., Coppola, D., Russo, R., Tinajero-Trejo, M., di Prisco, G., Lauro, F., Ascenzi, P., Verde, C., 2013. The globins of cold-adapted Pseudoalteromonas haloplanktis TAC125: from the structure to the physiological functions. Adv. Microb. Physiol. 63, 329–389. Gros, P., Milder, F.J., Janssen, B.J., 2008. Complement driven by conformational changes. Nat. Rev. Immunol. 8, 48–58. Guindon, S., Dufayard, J.F., Lefort, V., Anisimova, M., Hordijk, W., Gascuel, O., 2010. New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst. Biol. 59, 307–321. Hirani, S., Lambris, J.D., Muller-Eberhard, H.J., 1986. Structural analysis of the asparagine-linked oligosaccharides of human complement component C3. Biochem. J. 233, 613–616. Huber, R., Scholze, H., Paques, E.P., Deisenhofer, J., 1980. Crystal structure analysis and molecular model of human C3a anaphylatoxin. Hoppe-Seyler’s Z. Physiol. Chem. 361, 1389–1399. Isaac, L., Isenman, D.E., 1992. Structural requirements for thioester bond formation in human complement component C3. Reassessment of the role of thioester bond integrity on the conformation of C3. J. Biol. Chem. 267, 10062–10069. Isenman, D.E., Leung, E., Mackay, J.D., Bagby, S., van den Elsen, J.M., 2010. Mutational analyses reveal that the staphylococcal immune evasion molecule Sbi and complement receptor 2 (CR2) share overlapping contact residues on C3d: implications for the controversy regarding the CR2/C3d cocrystal structure. J. Immunol. 184, 1946–1955. Janssen, B.J., Christodoulidou, A., McCarthy, A., Lambris, J.D., Gros, P., 2006. Structure of C3b reveals conformational changes that underlie complement activity. Nature 444, 213–216. Janssen, B.J., Gros, P., 2006. Conformational complexity of complement component C3. Adv. Exp. Med. Biol. 586, 291–312. Janssen, B.J.C., Gros, P., 2007. Structural insights into the central complement component C3. Mol. Immunol. 44, 3–10. Janssen, B.J.C., Huizinga, E.G., Raaijmakers, H.C.A., Roos, A., Daha, M.R., NilssonEkdahl, K., Nilsson, B., Gros, P., 2005. Structures of complement component C3 provide insights into the function and evolution of immunity. Nature 437, 505–511. Kalli, K.R., Ahearn, J.M., Fearon, D.T., 1991. Interaction of iC3b with recombinant isotypic and chimeric forms of CR2. J. Immunol. 147, 590–594. Katoh, K., Standley, D.M., 2013. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol. Biol. Evol. 30, 772–780. Kieslich, C.A., Morikis, D., 2012. The two sides of complement C3d: evolution of electrostatics in a link between innate and adaptive immunity. PLoS Comput. Biol. 8, e1002840. Klos, A., Wende, E., Wareham, K.J., Monk, P.N., 2013. International Union of Pharmacology. LXXXVII. Complement peptide C5a, C4a, and C3a receptors. Pharmacol. Rev. 65, 500–543. Kosakovsky Pond, S.L., Frost, S.D., 2005. Not so different after all: a comparison of methods for detecting amino acid sites under selection. Mol. Biol. Evol. 22, 1208–1222. Kosakovsky Pond, S.L., Murrell, B., Fourment, M., Frost, S.D., Delport, W., Scheffler, K., 2011. A random effects branch-site model for detecting episodic diversifying selection. Mol. Biol. Evol. 28, 3033–3043. Kuroda, N., Naruse, K., Shima, A., Nonaka, M., Sasaki, M., 2000. Molecular cloning and linkage analysis of complement C3 and C4 genes of the Japanese medaka fish. Immunogenetics 51, 117–128. Larkin, M.A., Blackshields, G., Brown, N.P., Chenna, R., McGettigan, P.A., McWilliam, H., Valentin, F., Wallace, I.M., Wilm, A., Lopez, R., Thompson, J.D., Gibson, T.J., Higgins, D.G., 2007. Clustal W and clustal X version 2.0. Bioinformatics 23, 2947–2948. Law, S.K., Dodds, A.W., 1997. The internal thioester and the covalent binding properties of the complement proteins C3 and C4. Protein Sci. 6, 263–274. Mauri, I., Roher, N., MacKenzie, S., Romero, A., Manchado, M., Balasch, J.C., Bejar, J., Alvarez, M.C., Tort, L., 2011. Molecular cloning and characterization of European seabass (Dicentrarchus labrax) and Gilthead seabream (Sparus aurata) complement component C3. Fish Shellfish Immunol. 30, 1310–1322. Meng, F., Sun, Y., Liu, X., Wang, J., Xu, T., Wang, R., 2012. Analysis of C3 suggests three periods of positive selection events and different evolutionary patterns between fish and mammals. PLoS ONE 7, e37489. Mohamad Ali, M.S., Mohd Fuzi, S.F., Ganasen, M., Abdul Rahman, R.N., Basri, M., Salleh, A.B., 2013. Structural adaptation of cold-active RTX lipase from Pseudomonas sp. strain AMS8 revealed via homology and molecular dynamics simulation approaches. Biomed. Res. Int. 2013, 925373.

309

Mohan, R.R., Gorham Jr., R.D., Morikis, D., 2015. A theoretical view of the C3d:CR2 binding controversy. Mol. Immunol. 64, 112–122. Murrell, B., Moola, S., Mabona, A., Weighill, T., Sheward, D., Kosakovsky Pond, S.L., Scheffler, K., 2013. FUBAR: a fast, unconstrained Bayesian approximation for inferring selection. Mol. Biol. Evol. 30, 1196–1205. Murrell, B., Wertheim, J.O., Moola, S., Weighill, T., Scheffler, K., Kosakovsky Pond, S.L., 2012. Detecting individual sites subject to episodic diversifying selection. PLoS Genet. 8, e1002764. Nagar, B., Jones, R.G., Diefenbach, R.J., Isenman, D.E., Rini, J.M., 1998. X-ray crystal structure of C3d: a C3 fragment and ligand for complement receptor 2. Science 280, 1277–1281. Nakao, M., Kato-Unoki, Y., Nakahara, M., Mutsuro, J., Somamoto, T., 2006. Diversified components of the bony fish complement system: more genes for robuster innate defense? Adv. Exp. Med. Biol. 586, 121–138. Nakao, M., Mutsuro, J., Obo, R., Fujiki, K., Nonaka, M., Yano, T., 2000. Molecular cloning and protein analysis of divergent forms of the complement component C3 from a bony fish, the common carp (Cyprinus carpio): presence of variants lacking the catalytic histidine. Eur. J. Immunol. 30, 858–866. Near, T.J., Dornburg, A., Kuhn, K.L., Eastman, J.T., Pennington, J.N., Patarnello, T., Zane, L., Fernandez, D.A., Jones, C.D., 2012. Ancient climate change, antifreeze, and the evolutionary diversification of Antarctic fishes. Proc. Natl. Acad. Sci. U. S. A. 109, 3434–3439. Nichols, C.A., Guezennec, J., Bowman, J.P., 2005. Bacterial exopolysaccharides from extreme marine environments with special consideration of the southern ocean, sea ice, and deep-sea hydrothermal vents: a review. Mar. Biotechnol. (NY) 7, 253–271. Petersen, T.N., Brunak, S., von Heijne, G., Nielsen, H., 2011. SignalP 4.0: discriminating signal peptides from transmembrane regions. Nat. Methods 8, 785–786. Poli, A., Anzelmo, G., Nicolaus, B., 2010. Bacterial exopolysaccharides from extreme marine habitats: production, characterization and biological activities. Mar. Drugs 8, 1779–1802. Ren, X.D., Dodds, A.W., Enghild, J.J., Chu, C.T., Law, S.K., 1995. The effect of residue 1106 on the thioester-mediated covalent binding reaction of human complement protein C4 and the monomeric rat alpha-macroglobulin alpha 1 I3. FEBS Lett. 368, 87–91. Ricklin, D., Hajishengallis, G., Yang, K., Lambris, J.D., 2010. Complement: a key system for immune surveillance and homeostasis. Nat. Immunol. 11, 785–797. Rizzello, A., Romano, A., Kottra, G., Acierno, R., Storelli, C., Verri, T., Daniel, H., Maffia, M., 2013. Protein cold adaptation strategy via a unique seven-amino acid domain in the icefish (Chionodraco hamatus) PEPT1 transporter. Proc. Natl. Acad. Sci. U. S. A. 110, 7068–7073. Shin, S.C., Kim, S.J., Lee, J.K., Ahn do, H., Kim, M.G., Lee, H., Lee, J., Kim, B.K., Park, H., 2012. Transcriptomics and comparative analysis of three antarctic notothenioid fishes. PLoS ONE 7, e43762. Sigtryggsdottir, A.R., Papaleo, E., Thorbjarnardottir, S.H., Kristjansson, M.M., 2014. Flexibility of cold- and heat-adapted subtilisin-like serine proteinases evaluated with fluorescence quenching and molecular dynamics. Biochim. Biophys. Acta 1844, 705–712. Sunyer, J.O., Tort, L., Lambris, J.D., 1997. Diversity of the third form of complement, C3, in fish: functional characterization of five forms of C3 in the diploid fish Sparus aurata. Biochem. J. 326, 877–881. Sunyer, J.O., Zarkadis, I.K., Lambris, J.D., 1998. Complement diversity: a mechanism for generating immune diversity? Immunol. Today 19, 519–523. Sunyer, J.O., Zarkadis, I.K., Sahu, A., Lambris, J.D., 1996. Multiple forms of complement C3 in trout that differ in binding to complement activators. Proc. Natl. Acad. Sci. U. S. A. 93, 8546–8551. Tack, B.F., Harrison, R.A., Janatova, J., Thomas, M.L., Prahl, J.W., 1980. Evidence for presence of an internal thiolester bond in 3rd component of humancomplement. Proc. Natl. Acad. Sci. U. S. A. 77, 5764–5768. Thorne, M.A., Burns, G., Fraser, K.P., Hillyard, G., Clark, M.S., 2010. Transcription profiling of acute temperature stress in the Antarctic plunderfish Harpagifer antarcticus. Mar. Genomics 3, 35–44. Van Der Spoel, D., Lindahl, E., Hess, B., Groenhof, G., Mark, A.E., Berendsen, H.J., 2005. GROMACS: fast, flexible, and free. J. Comput. Chem. 26, 1701–1718. Varriale, S., Ferraresso, S., Giacomelli, S., Coscia, M.R., Bargelloni, L., Oreste, U., 2012. Evolutionary analysis of Antarctic teleost Toll-like receptor 2. Fish Shellfish Immunol. 33, 1076–1085. Vriend, G., 1990. WHAT IF: a molecular modeling and drug design program. J. Mol. Graph. 8, 52–56, 29. Wetsel, R.A., Kildsgaard, J., Haviland, D.L., 2000. Complement anaphylatoxins (C3a, C4a, C5a) and their receptors (C3aR, C5aR/CD88) as therapeutic targets in inflammation. In: Lambris, J.D., Holers, V.M. (Eds.), Therapeutics Interventions in the Complement System. Human Press, New York, p. 113. Zarkadis, I.K., Sarrias, M.R., Sfyroera, G., Sunyer, J.O., Lambris, J.D., 2001. Cloning and structure of three rainbow trout C3 molecules: a plausible explanation for their functional diversity. Dev. Comp. Immunol. 25, 11–24.