Isolation of two C-reactive protein homologues from cod (Gadus morhua L.) serum

Fish & Shellfish Immunology (2009) 26, 210e219

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/fsi

Isolation of two C-reactive protein homologues from cod (Gadus morhua L.) serum Berglind Gisladottir a, Sigridur Gudmundsdottir a, Linton Brown b, Zophonias O. Jonsson c, Bergljot Magnadottir a,* a

Institute for Experimental Pathology, University of Iceland, Keldur v. Vesturlandsveg, 112 Reykjavik, Iceland Institute of Aquaculture, University of Stirling, Stirling FK9 4LA, Scotland, UK c Institute of Biology, University of Iceland, Askja, 101 Reykjavik, Iceland b

Received 11 January 2008; revised 18 March 2008; accepted 20 March 2008 Available online 1 April 2008

KEYWORDS Cod; Gadus morhua L; Pentraxin; CRP; SAP; Acute phase proteins

Abstract Pentraxins are important molecules in innate defence and play a role in the acute phase response of both mammals and fish. Isolation of cod pentraxins by affinity chromatography using phosphorylcholine agarose revealed two pentraxin-like proteins, referred to as PI and PII proteins. These varied in their overall charge, pentameric and subunit molecular size, glycosylation and N-terminal amino acid sequences. The PI protein was homologous with the CRP-like pentraxin previously described in cod whereas the PII protein was a new CRP homologue, which was characterized by substantial individual heterogeneity with regard to subunit size and relative density. The results indicate considerable genetic variations in the cod pentraxins. ª 2008 Published by Elsevier Ltd.

Introduction The pentraxins, serum amyloid P (SAP) and C-reactive protein (CRP), are pattern recognition proteins found in both invertebrates and vertebrates. Pentraxins belong to the innate immune system and are key factors in acute phase response to infection, injury or trauma. They can activate the classical complement pathway as well as take part in other immunological and biological processes [1]. SAP and CRP are the protypical acute phase proteins in mammals, their serum levels increasing drastically during

* Corresponding author. Tel.: þ354 5855100; fax: þ354 5674714. E-mail address: [email protected] (B. Magnadottir). 1050-4648/$ - see front matter ª 2008 Published by Elsevier Ltd. doi:10.1016/j.fsi.2008.03.015

an acute phase response [2]. Decreasing, as well as increasing serum levels have been reported during acute phase response of different fish species [3e8]. Pentraxins are pentameric, planar molecules. The five subunits are commonly non-covalently linked, often glycosylated and with a recognition- and effector-face; dimeric forms (decamers) and larger aggregates have also been described [9e11]. They show characteristic calcium dependent ligand specificity, CRP binds primarily to phosphorylcholine, a pneumococcal C-polysaccharide, and SAP shows specificity for phosphoethanolamine, agarose, zymosan and other glycans as well as DNA and chromatin [9]. The ligand specificity is generally accepted as being the defining characteristic of the pentraxin type. The two pentraxins, SAP and CRP, are both present in mammalian serum and

Isolation of C-reactive protein homologues from cod human SAP and CRP show about 70% amino acid sequence similarity when conservative substitutions are taken into account [9]. SAP- and/or CRP-like proteins have been described in the serum of several fish species, for example, plaice (Pleuronectes platessa L.) [3], rainbow trout (Oncorhynchus mykiss) [12,13], halibut (Hippoglossus hippoglossus) [14] and channel catfish (Ictalurus punctatus) [15]. Pentraxin homologue of cod has been described by Lund and Olafsen [14]. According to their data a CRP-like pentraxin was present in cod serum. This was essentially based on the binding specificity for phosphorylcholine (PC). During recent studies at our laboratory on the role of pentraxin in the acute phase response of cod, additional data was obtained and a previously noted heterogeneity of the cod pentraxins [16] was confirmed, which is the subject of this paper.

Materials and methods

211

Ion exchange chromatography MonoQ anion ion exchange column (HR5/5, Amersham Pharmacia Biotech) was used. The column was equilibrated in 20 mM TriseHCl, pH 7.5 start buffer. The crude pentraxin isolated above was equilibrated in the start buffer using a PD buffer exchange column (Amersham Pharmacia Biotech) and then loaded onto the MonoQ column and eluted with start buffer followed by a salt gradient of 0e1 M NaCl in the start buffer.

Size-exclusion chromatography Superose 12 (HR10/30, Amersham Pharmacia Biotech) was used and the elution buffer was 100 mM TriseHCl, 150 mM NaCl, pH 8.0. A size calibration graph was plotted using a gel filtration calibration kit from GE Healthcare UK Limited, UK, and the molecular weight of the pentraxins extrapolated from this.

Fish and serum SDS-PAGE analysis All fish came from the Marine Institute’s Experimental Station, Stadur, Grindavik, Iceland, originating from wild gamete fertilized and hatched at the station and maintained at 7  C under optimal cultural conditions [17]. For the isolation of pentraxin from serum, cod weighing 600e700 g was used. Blood, 2e10 ml, was collected from the caudal vessel of anaesthetized (40e50 mg tricaine methanesulphonate l1) fish and allowed to coagulate overnight at 4  C. Serum was then collected, pooled from several fish, and stored at 20  C until use. To examine the individual heterogeneity of pentraxins, serum from 12 fish, weighing about 100 g and originating from the same hatch, were used. To study the possible change with time serum collected at intervals over a period of 18 months from one individual cod as well as serum collected from six individuals at the beginning and the end of the 18 month period was used. These samples came from a previous experiment on cod [18].

Standard SDS-PAGE analysis of reduced and non-reduced samples was carried out using the Mini-PROTEAN II system from Bio-Rad Laboratories (USA) according to the manufacturer’s instructions. The separation gel was 12% or 14% and the stacking gel 4.5% acrylamide. Reduced samples were mixed with an equal volume of 0.125 M Tris, pH 6.8 sample buffer containing 0.5% 2-mercaptoethanol and 2% SDS and heated for 2e5 min at 100  C. Non-reduced samples were mixed with the same buffer, omitting the mercaptoethanol and the heating. Non-reduced samples were also analysed in homogeneous 4% SDS-PAGE acrylamide gel strengthened with 0.6% agarose according to a method described by Avatlion and Mor [21]. Native-PAGE analysis in a 12% gel, omitting SDS from all solutions, was also carried out on non-reduced samples. Proteins were stained with Silver Stain Plus (Bio-Rad) or Phast Gel Blue R (SigmaeAldrich, USA). Prestained molecular weight markers from Fermentas Life Sciences, Germany, or Invitrogen, Germany, were used for estimating molecular weight.

Affinity chromatography Western blotting The isolation of pentraxins by affinity chromatography, using phosphorylcholine (PC)-agarose and ethylenediaminetetraacetic acid (EDTA) elution was based on previously published protocols [6,12e14,19,20]. PC-agarose from Pierce Chemicals (USA) was used. Ten milliters of the PC-agarose was equilibrated with several changes of excess 50 mM TriseHCl buffer, pH 8.0, containing 150 mM NaCl (TBS) and 2 mM CaCl2, (TBS-Ca). The gel was then mixed with 13 ml of TBS-Ca buffer and 13 ml of pooled cod serum and gently rotated overnight at 4  C. The supernatant was discarded and the gel washed once with the TBS-Ca buffer. The gel was packed onto a chromatographic column (XK16, Amersham Pharmacia Biotech, ¨ KTA FPLC system from AmerSweden), connected to the A sham Pharmacia Biotech and washed in TBS-Ca buffer until the absorbance of the eluate was back to zero background. Bound protein (crude pentraxins) was then eluted with 10 mM EDTA in TBS buffer.

Following standard SDS-PAGE separation, proteins were transferred to a nitrocellulose (NC) membrane (ImmobilonP transfer membrane, Millipore, UK), using a semi-dry MilliBlot Graphite Electroblotter (Millipore) according to the manufacturer’s instructions. Transfer was for 1 h at room temperature. Residual sites of the NC membrane were blocked with 0.1% semi-skimmed milk powder in 0.1 M Tris buffered saline, pH 7.8, containing 0.1% Tween20, for 1 h at room temperature or overnight at 4  C. Incubations with the primary and the secondary peroxidase or alkaline phosphatase conjugated antibodies, diluted in the blocking solution, were for 1 h at room temperature with extensive washing between each step. When using peroxidase conjugated secondary antibody, the blots were developed using the ECL system from Amersham Pharmacia Biotech. When using alkaline phosphatase conjugated secondary antibody the blots were developed in a substrate buffer of 100 mg ml1

212 NBT (nitro blue tetrazolium), 60 mg ml1 BCIP (5-bromo4-chloro-3-indonylphosphate) and 4 mM MgCl2 in 0.1 M ethanolamineeHCl buffer, pH 9.6.

B. Gisladottir et al. produced against crude pentraxin preparation isolated by affinity chromatography, (b) anti-PI produced against peak I from the MonoQ separation and (c) anti-PII produced against peak II from the MonoQ separation.

Electron microscopy A drop of the protein solution, 900 mg ml1, was placed onto a 200 mesh Formvar-coated copper grid and left for 2 min. Excess fluid was removed by touching the edge with filter paper. A drop of 1% phosphotungstic acid, pH 6.5, was added and left for 2 min before excess fluid was removed as before. This was allowed to dry and then viewed using an FEI Tecnai Spirit G2 Bio Twin transmission electron microscope (TEM) (FEI UK Ltd.).

Ligand specificity Approximately 10 mg of agarose (BDH Chemicals, Ltd., UK), zymosan (insoluble preparation of yeast, Saccharomyces cerevisiae, prepared at our laboratory according to Llanos et al. [22]) or PC-agarose were washed extensively in TBS-Ca buffer and then incubated with about 20 mg protein in 100 ml TBS-Ca buffer or buffer only for 1 h at room temperature with gentle shaking. One microlitre of the supernatant was dotted on a pair of nitrocellulose membranes, dried and then blocked in 0.1% semi-skimmed milk powder as described in ‘‘Western Blotting’’. One membrane was incubated with anti-pentraxin antibody and the other with anti-PII antibody (see ‘‘Antibodies’’), this was followed by incubation with alkaline phosphatase conjugated antibody and the colour developed using the NBT/BCIP substrate buffer as described in ‘‘Western Blotting’’.

Amino acid sequence analysis A total of six subunits, two from the PI protein and four from the PII protein, were subjected to sequence analysis. For Nterminal amino acid sequencing, using Edman’s degradation method, samples were excised from PVDF membrane following SDS-PAGE and Western blotting and protein staining using red ponceau S or amido black. This analysis was kindly carried out by J. d’Alayer at the Institut Pasteur, Paris, France. For mass spectrometry based protein identification (proteomics), using a Waters QTOF2 mass spectrometer, samples were excised from acrylamide gel following SDSPAGE and Phast Gel Blue staining. This analysis was kindly carried out by K. Bailey, School of Biomedical Sciences, Queens Medical Centre, Nottingham, UK.

Results Affinity isolation and ion exchange chromatography Thirteen milliters of cod serum, pooled from several individuals, following affinity chromatography, yielded about 1 mg of crude pentraxin or approximately 80 mg ml1 serum. Separation on an anion exchange column gave two peaks, I and II (Fig. 1). Peak I was eluted in a single 0.5 ml fraction with 0.48 M NaCl, peak II had a broader elution profile and was eluted in three 0.5 ml fractions with 0.54e0.64 M NaCl. The proportional protein yield of peaks I and II was about 1:1. These will be referred to as PI and PII proteins below. Fractions were analysed in SDS-PAGE followed by silver staining. The result is shown as an insert in Fig. 1. The PI protein was composed of two subunits whereas the PII protein was composed of several (here 4) subunits, the number of these varied from one serum pool to another (not shown).

Size estimation of the PI and PII proteins Various methods were used to estimate the molecular size of the PI and PII proteins. Some degree of discrepancy was observed between these methods reflecting the variable behaviour of proteins, reduced or non-reduced, in different media [24]. In aqueous phase (size-exclusion chromatography, results not shown) both proteins were eluted as a single, symmetrical peak. The PI protein was eluted with 11.54 ml and the PII with 12.46 ml. The corresponding molecular size

Deglycosylation A kit from New England BioLabs (USA) containing the enzyme PNGase, which removes N-linked oligosaccharides from proteins under reducing or non-reducing conditions, was used following the manufacturer’s instructions. The protein sample was treated with the enzyme both before and after reduction in reducing buffer containing mercaptoethanol (from the kit).

Antibodies Three polyclonal antibodies were produced in mice according to Overkamp et al. [23]: (a) anti-pentraxin antibody

Figure 1 Ion exchange chromatography (MonoQ anion exchanger) of crude cod pentraxins. Two distinct types of C-reactive proteins were eluted. The PI protein (peak I) was eluted with 0.48 M NaCl and the PII protein (peak II) with 0.54e0.64 M NaCl. The insert shows the SDS-PAGE/silver staining analysis of the fractions from peak I (PI protein, 1 fraction) and peak II (PII protein, three fractions) under reducing conditions.

Isolation of C-reactive protein homologues from cod extrapolated from the standard graph was 158 and 90 kDa, respectively. Using standard SDS-PAGE the two subunits of the PI protein were 28 and 31 kDa under reducing conditions (Fig. 2b) and under non-reducing conditions 49 and 54 kDa subunits were seen as well as two larger proteins, approximately 200 and 240 kDa (Fig. 2a). This indicated the presence of some inter-subunit disulfide bonds. When examined in 4% homogeneous SDS-PAGE gel 110 and 125 kDa proteins were seen and in native-PAGE and 118 and 158 kDa proteins were observed (results not shown). Using standard SDS-PAGE, the PII protein separated into several protein subunits, 22e29 kDa under reducing conditions (Fig. 2b) and 20e26 kDa under non-reducing conditions (Fig. 2a), indicating the absence of inter-subunit disulfide bonds. The slightly larger size observed under reducing conditions probably reflects the effects of reduced intra-subunit disulphide bonds changing the protein shape and movement through the gel. No proteins larger than 70 kDa were seen in 4% homogeneous SDS-PAGE whereas in native-PAGE a major 82 kDa protein and a minor 158 kDa protein were seen. All the PI and PII subunits seen in Fig. 2 were also detected in Western blotting using anti-pentraxin antibody.

Electron microscopy Results from the EM analysis of the PII protein showed a pentameric-like form, which was 8.18 nm and 7.96 nm across (Fig. 3). Analysis of the PI protein was unsuccessful.

213

Figure 3 Electron micrograph of the PII protein showing pentameric structure at an original magnification of 300.000.

Antibody production The mouse antibody produced against crude pentraxin, isolated by affinity chromatography (anti-pentraxin antibody), reacted in Western blotting with both the PI and PII proteins (Fig. 4). The mouse antibodies produced against the PI protein or against the PII protein were monospecific and did not cross react. Ample quantity was obtained of the anti-pentraxin and anti-PII antibody but only a limited amount of the anti-PI antibody, which also showed relatively weak activity.

Ligand specificity The Phosphorylcholine-agarose removed both the PI and PII proteins from the solution, agarose alone reduced the protein density slightly, in particular of the PI protein, and zymosan had no effect on either the PI or the PII

Figure 2 Standard SDS-PAGE analysis in 12% gel and silver staining of the PI and PII proteins under (a) non-reducing and (b) reducing conditions. The analysis revealed different degrees of heterogeneity and different covalent bondings of the PI and PII protein subunits.

Figure 4 Western blotting analysis of the isolated PI and PII proteins under reducing conditions with the three antibodies generated: (a) anti-pentraxin antibody, (b) anti-PII protein antibody and (c) anti-PI protein antibody. The anti-pentraxin antibody detected both proteins whereas the anti-PI and anti-PII antibodies were monospecific for the relevant protein. The activity of the anti-PI antibody was relatively poor and required prolonged incubation.

214

B. Gisladottir et al.

protein (Fig. 5). The monospecificity of the anti-PII antibody was again demonstrated (Fig. 5b).

Glycosylation of cod pentraxins The PNGase removed about 2 kDa from the larger subunit of the PI protein following reduction making it the same size as the smaller subunit but the enzyme did not remove carbohydrate from the non-reduced PI protein (Fig. 6). The PNGase removed carbohydrate from the largest subunit of the PII protein both before and after reduction. Although in the case of the PII protein it was not possible to ascertain the amount of carbohydrate removed it must be at least 2 kDa.

Sequence analysis of the P I and PII proteins The following protein subunits (see insert in Table 1) were subjected to N-terminal analysis: PI subunit 2 and PII subunit 4 and the following to mass spectrometric analysis: PI subunit 1 and PII subunits 1e3. Both types of analysis demonstrated a significant similarity to pentraxins from other species (Fig. 7, Table 1). The N-terminal sequence of the PI subunit 2 was identical to the N-terminal sequence of cod CRP homologue, published by Lund and Olafsen [14] but also showed relatively high identity to SAP from, for example, wolfish (Anarhichas lupus) and halibut. Lower identity was generally shown to CRP of other species. (Fig. 7). Mass spectrometric analysis of the PI subunit 1 gave four sequences, three of which showed significant identity to pentraxins (SAP) from other species. One of these (QDLSGK) matched a proportion of the N-terminal sequence of the PI subunit 2 (Table 1). The N-terminal sequence of the PII subunit 4 showed significant similarity to a CRP-like protein isolated from rainbow trout, published by Murata et al. [13] and to CRP described in several other species (Fig. 7). Mass spectrometric analysis of the PII subunits 1e3 gave 6, 4 and 3

Figure 6 Western blotting analysis of the PI and PII proteins following deglycosylation using PNGase (PNG), both before (NR) and after (R) reduction. Deglycosylation of the PI protein was only achieved under reduced conditions whereas N-linked carbohydrate could be removed from the PII protein both before and after reduction.

sequences, respectively (Table 1). With the exception of two sequences from the PII subunit 1, all the sequences showed significant identity to CRP from other species. One sequence, VFPEETANSFVELTPAK, from each of the three PII subunits (PII-1, -2 and -3) matched a proportion of the N-terminal sequence of the PII subunit 4 (Table 1). Two sequences were shared by the PI subunit 1 and the PII subunits 1. One of these, ALFSLATR, also showed significant identity to SAP from salmonids whereas the other, FQAVTATTGR, showed no significant match to pentraxins (Table 1).

Heterogeneity of cod pentraxin Twelve sera from individual cod of a similar size and of the same origin were tested in Western blotting using the monospecific anti-PII antibody (Fig. 8a). With regard to the number of subunits, their size and relative density most of the subunit patterns were different from one individual to another. Sera from six individuals from another group of cod retained the same subunit pattern at the beginning and end of a period of 18 months (Fig. 8b) and when one of these (number 1) was tested at intervals during this period the same subunit pattern was seen throughout although the density varied (Fig. 8c). When these sera were tested with the weak anti-PI antibody there were indications that such heterogeneity was not associated with the PI protein (results not shown).

Discussion

Figure 5 Dot blotting of the PI and PII proteins following incubation with different ligands: agarose, zymosan and PCagarose (PC) tested with (a) anti-pentraxin antibody and (b) anti-PII antibody. The ligand specificity was primarily for phosphorylcholine although the PI protein showed weak affinity with the agarose alone.

In the present study, two types of calcium dependent phosphorylcholin (PC) binding proteins, referred to as PI and PII proteins, were isolated from cod serum. The total quantity of pentraxin in serum, 80 mg ml1, was comparable to that of other fish species [6,8]. The PC agarose used for the affinity isolation would be expected to retain both PC (CRP) and agarose (SAP) binding proteins. Further examination of the ligand specificity

Subunits from proteins PI and PII

Type of analysis

Results

Blast analysis: examples of % identity (Swiss-Prot and NCBI)

PI subunit 1

MS

TPGNVLNWR

78% identity to SAP from Atlantic salmon [5] and 67% identity to SAP from rainbow trout [5] 75% identity to SAP from Atlantic salmon and rainbow trout [5] No significant match to pentraxin 100% identity to CRP from cod and SAP from Atlantic salmon, wolffish [14] and rainbow trout [12]

ALFSLATR* FQAVTATTGR** QDLSGK PI subunit 2

N-terminal

IPQDLSGKMLTFPKE

100% identity to CRP from cod [14] 73% identity to SAP from wolffish [14] and 57% identity to SAP from human [27]

PII subunit 1

MS

LGQDPDNFLGDFQAK*** VFPEETANSFVELTPAK*

67% identity to pentraxin from carp [26] and zebrafish (CAM46930) 71% identity to CRP from rainbow trout [13] and 65% identity to CRP from plaice [25] No significant match to pentraxin No significant match to pentraxin 53% identity to SAP from Atlantic salmon and rainbow trout [5] 75% identity to SAP from Atlantic salmon and rainbow trout [5]

FQAVTASATR** AVADFHMASLPLNK PLTLLGQEQDSYGGGDFLK ALFSLATR* PII subunit 2

MS

LGQDPDNFLGDFQAK*** VFPEETANSFVELTPAK* EVLLFAYR DELNVWR

PII subunit 3

MS

LGQDPDNFLGDFQAK*** VFPEETANSFVELTPAK* DELNVWR

PII subunit 4

N-terminal

GRSLVFPEETANSFVELFPAKELSL*

Isolation of C-reactive protein homologues from cod

Table 1 The amino acid sequence analysis of subunits from protein PI and PII as shown in Fig. 7 and examples of percentage identity (Blast analysis, GenBank accession number shown when relevant)

67% identity to pentraxin from carp [26] and zebrafish (CAM46930) 71% identity to CRP from rainbow trout [13] and 65% identity to CRP from plaice [25] 75% identity to pentraxin from zebrafish (CAM46925) 100% identity to pentraxin from carp [26] and zebrafish (CAM46924) 67% identity to pentraxin from carp [26] and zebrafish (CAM46930) 71% identity to CRP from rainbow trout [13] and 65% identity to CRP from plaice [25] 100% identity to pentraxin from carp [26] and zebrafish (CAM46924) 60% identity to CRP from rainbow trout [13] and 56% identity to CRP from plaice [25]

Same symbols show identical or similar sequences. MS: Mass spectrometric analysis.

215

216

B. Gisladottir et al.

a

Identity PI protein

I

P

Q

D

L

S

G

K

M

L

T

F

P

Cod CRP (14)

(I)

P

Q

D

L

S

G

K

M

L

T

F

Wolffish SAP (14)

A

P

Q

D

L

S

G

K

M

F

I

F

H

T

D

L

S

G

K

V

F

V

P

I

D

L

L

G

K

V

F

Q

T

D

M

S

R

K

A

Human SAP (27) Halibut SAP (14)

(E)

Human CRP (28)

K

E

P

K

?E

100

P

Q

E

73

F

P

R

E

57

V

F

S

K

E

53

F

V

F

P

K

E

50

Salmon SAP (14)

A

H

Q

D

L

S

G

K

V

F

V

I

P

M

A

47

Rainbow trout SAP (12)

T

G

Q

D

L

S

G

K

V

F

V

S

P

M

T

47

Plaice SAP (25)

Z

P

I

D

K

E

47

P

R

Q

Dogfish SAP (29) Syrian hamster SAP (30)

D

L

M

G

K

V

F

V

D

G

F

P

G

K

S

L

I

F

Q

T

D

L

T

G

K

V

F

V

F

D

F

Dogfish CRP (29)

A

T

A

G

L

A

G

K

A

L

44 43 42

Frog CRP (31)

A

Q

E

D

L

V

G

N

V

F

L

F

P

K

P

40

Mouse SAP (32)

C

Q

T

D

L

K

R

K

V

F

V

F

P

R

E

40

Rat SAP (33)

A

Q

T

D

L

N

Q

K

V

F

V

F

P

R

E

40

b

Identity

PII protein (CRP)

G

Rainb. trout CRP (13)

R

Plaice CRP (25)

I

R

S

L

V

F

P

E

E

T

R

S

L

V

K

T

L

V

A

N

S

F

V

E

L

F

P

F

P

M

E

F

Q

S

E

A

K

T

D

N

S

Y

V

E

L

V

S

N

N

S

F

V

E

L

I

E

L

S

L

P

Q

K

P

M

K

P

L

N

L

56

60

Syr. hamster CRP (34)

K

T

A

F

V

F

P

K

E

S

A

N

S

Y

V

S

L

E

A

Q

S

K

K

T

L

40

Rabbit CRP (35)

K

K

A

F

V

F

P

K

E

S

D

N

S

Y

V

S

L

N

A

Q

L

T

K

P

L

36

Human CRP (28)

R

K

A

F

V

F

P

K

E

S

D

T

S

Y

V

S

L

K

A

P

L

T

K

P

L

32

Mouse CRP (36)

K

K

A

F

V

F

P

K

E

S

D

T

S

Y

V

S

L

E

A

E

S

K

K

P

L

32

Rat CRP (37)

K

Q

A

F

V

F

P

G

V

S

A

T

A

Y

V

S

L

E

A

E

S

K

K

P

L

28

Frog CRP (31)

G

N

V

F

L

F

P

K

P

S

V

T

T

Y

A

I

L

K

P

E

V

E

K

P

L

28

Limulus CRP (38)

I

T

S

K

V

K

F

P

P

S

S

P

S

F

P

R

L

V

M

V

G

T

L

P

D

24

Figure 7 Comparison of N-terminal amino acid sequences of PI and PII (subunits 2 and 4, respectively, see insert in Table 1) with known pentraxin sequences [12e14,27e38]. (a) N-terminal amino acid sequence of PI compared to known CRP and SAP from other species. (b) Nterminal amino acid sequence of PII compared to known CRP from other species. Residues identical to PI or PII are greyish within boxes. The percentage values indicate identity between the PI or PII and other proteins. Uncertain or unidentified amino acids are indicated by ( ) or ?.

Figure 8 Western blotting of different cod sera with anti-PII antibody. (a) sera from 12 individuals, (b) sera from six individuals sampled at an interval of 18 months (e.g. 1a and 1b, respectively) and (c) sera from one individual sampled ten times over a period of 18 months.

showed that both proteins were PC specific, the PI protein showed weak binding specificity for agarose but neither protein showed specificity for zymosan. By definition both proteins should therefore be classified as CRP homologues [9]. The PI protein was basically equivalent to the cod PCbinding protein (Gm-PCBP, CRP homologue) described by Lund and Olafsen [14] although additional characteristics were observed. The PII protein, on the other hand, was a new identity not described before in cod and was notable for its heterogeneity. Ion exchange chromatography of the two proteins demonstrated a higher and a more heterogeneous overall negative charge of the PII protein compared to the PI protein. The two proteins also varied with respect to the total and subunit size and composition. The PI protein was about 158 kDa, composed of two types of subunits, one of which was glycosylated and some inter-subunit covalent bonds were indicated. The PII protein was about 90 kDa, composed of several subunits, one of which was glycosylated and inter-subunit bonds were absent. It has been suggested that pentraxins (e.g. SAP) require Ca2þ and a stabilizing ligand like albumin to form a stable pentamer [10]. Both factors are present in serum but would be absent from the purified PI and PII protein preparations.

Isolation of C-reactive protein homologues from cod This could explain the difficulties encountered in determining the exact molecular size and the appearance of both truncated and complexed pentamers in the different SDSPAGE analyses. Some authors have verified the pentameric structure of pentraxin-like proteins using electron microscopic analysis [13,39]. A pentameric structure was demonstrated for the PII protein using EM analysis but a pentameric structure of the PI protein remains to be verified. The largest subunit of both the PI and the PII protein (subunit 1) was glycosylated and the degree of glycosylation appeared similar for both proteins, corresponding approximately to a single carbohydrate moiety of 2 kDa. The presence of inter-subunit covalent bonds between the PI subunits, absent from the PII protein, probably made its carbohydrate moiety less accessible to enzyme digestion under non-reducing conditions. Heterogeneity with respect to the glycosylation of the pentraxin subunits has been described in Indian carp (Labeo rohita) by Mandal et al. [40] who found that the degree and type of glycosylation was influenced by environmental factors. The present samples come from cultured cod kept under similar and controlled conditions; hence, variations of this nature would not be expected. Examination of the influence of environmental factors on the glycosylation of cod pentraxins could be an interesting project. The PI protein (subunit PI-2) and the PII protein (subunit PII-4) had distinct N-terminal amino acid sequences. Both showed significant percentage identity to N-terminal sequences of pentraxins from other species. The PI protein showed closer N-terminal sequence identity with SAP than with CRP [12,14,25] whereas the N-terminal sequence of PII protein showed a closer identity with CRP of other fish species [13,25]. The mass spectrometric analysis further demonstrated that the two subunit of the PI protein had a common N-terminal sequence and similarly the four subunits of the PII protein had an identical N-terminal sequence. Some sequence identity was observed between the PI and the PII protein as is commonly seen between pentraxins from different sources [9]. Interestingly the shared sequences were associated with the largest, glycosylated subunit 1 of each protein. The differences observed between the PI and PII proteins were reflected in the distinct antigenicity of the two proteins demonstrated by the lack of cross reactivity between the anti-PI and PII antibodies. In general, when species possess both SAP- and CRP-like pentraxins, these are distinguished by their ligand specificity as well as by their N-terminal amino acid sequence [25,31]. In instances, when more than one PC-binding protein was identified, only one showed characteristic pentraxin features and sequence homology [13,41]. This paper describes, apparently for the first time, two types of PC-binding pentraxins, with distinct SAP- or CRP-like N-terminal sequences. In view of these results it should perhaps be considered whether the ligand specificity alone is a valid criterion for distinguishing between pentraxin types. Remarkable heterogeneity was a feature of the PII protein. When isolated from pooled sera up to six subunits of varying size and density were observed. Notable heterogeneity was also observed between individual cod, each carrying a characteristic pattern of one to four subunits.

217 Examination of serum collected from the same individual over a period of 18 months suggested that subunit pattern did not change with time although the overall or proportional density of each subunit might vary. Such individual heterogeneity did not appear to be associated with the PI protein. These results indicate a large amount of allelic variation of cod pentraxins and that the PII protein variants might be encoded at multiple loci in the cod genome. This heterogeneity may be an important factor in the innate immune defence and acute phase response of cod. Heterogeneity of immunologically important molecules, like the isoforms of complement proteins C3 and B/C2 factor, has been described in different fish species [42e44]. In the case of the C3 of rainbow trout these isoforms have been shown to have a different binding specificity for complement activators [45]. Recent work on the complement system of cod failed to demonstrate such heterogeneity of C3 in cod [46,47]. It could therefore be speculated whether the heterogeneity of the pentraxins compensates for the lack of C3 isoforms in cod. The diversity of the pattern recognition molecule (pentraxin) might therefore make up for the lack of diversity of its effector protein (C3). Examination of the binding specificity and pattern recognition properties of the different subunits/variants of the PI and in particular of the PII protein would be of considerable interest in this context. Work is now in progress at our laboratory to determine the genetic source of the PII protein heterogeneity. The possible role of pentraxins the PI or PII protein or both, in the acute phase response of cod is also being investigated.

Acknowledgements The authors would like to thank Matthias Oddsteinsson, Agnar Steinarsson and other staff at the Marine Institute’s Experimental Station, Stadur, Grindavik, Iceland for providing the fish and assisting with the blood collection. Eggert Gunnarsson and Sonja Vilhjalmsdottir are thanked for the antibody production. The project was supported by the Icelandic Research Council, the Research Fund of the University of Iceland and AVS R&D Fund of the Ministry of Fisheries in Iceland.

References [1] Bayne CJ, Gerwick L. The acute phase response and innate immunity of fish. Dev Comp Immunol 2001;25:725e43. [2] Ballou S, Kushner I. C-reactive protein and the acute phase response. Adv Intern Med 1992;37:313e34. [3] White A, Fletcher TC, Pepys MB, Baldo BA. The effect of inflammatory agents on C-reactive protein and serum amyloid P-component levels in plaice (Pleuronectes platessa L.) serum. Comp Biochem Physiol C 1981;69:325e9. [4] Szalai AJ, Norcum MT, Bly JE, Clem LW. Isolation of an acutephase phosphorylcholine-reactive pentraxin from channel catfish (Ictalurus punctatus). Comp Biochem Physiol 1992; 102B:535e43. [5] Jensen LE, Hiney MP, Shields DC, Uhlar CM, Lindsay AJ, Whitehead AS. Acute phase proteins in salmonids, evolutionary analysis and acute phase response. J Immunol 1997;158: 384e92.

218 [6] Lund V, Olafsen JA. Changes in serum concentration of a serum amyloid P-like pentraxin in Atlantic salmon, Salmo salar L.during infection and inflammation. Dev Comp Immunol 1999;23:61e70. [7] Kodama H, Matsuoka Y, Tanaka Y, Liu Y, Iwasaki T, Watarai S. Changes of C-reactive protein levels in rainbow trout (Oncorhynchus mykiss) sera after exposure to anti-ectoparasitic chemicals used in aquaculture. Fish Shellfish Immunol 2004; 16:589e97. [8] Liu Y, Iwasaki T, Watarai S, Kodama H. Effect of turpentine oil on C-reactive protein (CRP) production in rainbow trout (Oncorhynchus mykiss). Fish Shellfish Immunol 2004;17:203e10. [9] Tennent GA, Pepys MB. Glycobiology of the pentraxins. Biochem Soc Trans 1994;22:74e9. [10] Hutchinson WL, Hohenester E, Pepys MB. Human serum amyloid P component is a single uncomplexed pentamer in whole serum. Mol Med 2000;6:482e93. [11] Mortensen RF. C-reactive protein, inflammation and innate immunity. Immunol Res 2001;24:163e76. [12] Murata M, Onuma M, Kodama H. isolation and characterization of rainbow trout (Oncorhynchus mykiss) serum amyloid P component (SAP). J Vet Med Sci 1994;56:661e5. [13] Murata M, Kodama H, Onuma M. Characterization of rainbow trout C-polysaccharide binding proteins. J Vet Med Sci 1995; 57:419e25. [14] Lund V, Olafsen JA. A comparative study of pentraxin-like proteins in different fish species. Dev Comp Immunol 1998;22: 185e94. [15] Szalai AJ, Bly JE, Clem LW. Changes in serum concentration of channel catfish (Ictalurus punctatus Rafinesque) phosphorylcholine-reactive protein (PRP) in response to inflammatory agents, low temperature-shock and infection by the fungus Saprolegnia sp. Fish Shellfish Immunol 1994;4:323e36. [16] Magnado ´ttir B. Innate immunity of fish (overview). Fish Shellfish Immunol 2006;20:137e51. [17] Bjornsson B, Steinasson A. The food-unlimited growth rate of Atlantic cod (Gadus morhua). Can J Fish Aquat Sci 2002;59: 494e502. [18] Magnadottir B, Jonsdottir H, Helgason S, Bjo ¨rnsson B, Solem ST, Pilstro ¨m L. Immune parameters of immunized cod (Gadus morhua L.). Fish Shellfish Immunol 2001;11:75e89. [19] Fletcher TC, White A, Youngson A, Pusztai A, Baldo BA. Characterization of C-reactive protein from the eggs of the marine teleost, Cyclopterus lumpus L. Biochim Biophys Acta 1981;671:44e9. [20] Cartwright JR, Tharia HA, Burns I, Shrive AK, Hoole D, Greenhouth TJ. Isolation and characterisation of pentraxinlike serum proteins from the common carp Cyprinus carpio. Dev Comp Immunol 2004;28:113e25. [21] Avatlion RR, Mor A. Monomeric IgM is transferred from mother to egg in tilapius. Israeli J Aquacult 1992;44:93e8. [22] Llanos RJ, Miceli DC, Valz-Gianinet JN, Whitacre CM. A simple procedure for isolation of Bufo arenarum C sub (3) complement fraction and preparation of antiserum. Comp Biochem Physiol 1999;124C:1e5. [23] Overkamp D, Mohammed-Ali S, Cartledge C, Landon J. Production of polyclonal antibodies in ascitic fluid of mice: technique and applications. J Immunoassay 1988;9:51e68. [24] Tennent GA, Butler PJG, Hutton T, Woolfitt AR, Harvey DJ, Rademacher TW, et al. Molecular characterization of Limulus polyphemus C-reactive protein. I. Subunit composition. Eur J Biochem 1993;214:91e7. [25] Pepys MB, de Beer FC, Milstein CP, March JF, Feinstein A, Butress A, et al. C-reactive protein and serum amyloid P component in the plaice (Pleuronectes platessa L.), a marine teleost, are homologous with their human counterparts. Biochim Biophys Acta 1982;704:123e33.

B. Gisladottir et al. [26] Fujiki K, Bayne CJ, Shin DH, Nakao M, Yano T. Molecular cloning of carp (Cyprinus carpio) C-type lectin and pentraxin by use of suppression subtractive hybridisation. Fish Shellfish Immunol 2001;11:275e9. [27] Mantzouranis EC, Dowton SB, Whitehead AS, Edge MD, Bruns GA, Colten HR. Human serum amyloid P component, cDNA isolation, complete sequence of pre-serum amyloid P component, and localization of the gene to chromosome 1. J Biol Chem 1985;260:7752e6. [28] Woo P, Korenberg JR, Whitehead AS. Characterization of genomic and complementary DNA sequence of human C- reactive protein, and comparison with the complementary DNA sequence of serum amyloid P component. J Biol Chem 1985; 260:13384e8. [29] Robey FA, Tanaka T, Liu TY. Isolation and characterization of two major serum proteins from the dogfish, Mustelus canis, Creactive protein and amyloid P component. J Biol Chem 1983; 258:3889e94. [30] Rudnick CM, Dowton SB. Serum amyloid P (female protein) of the Syrian hamster. Gene structure and expression. J Biol Chem 1993;268:21760e9. [31] Lin L, Liu TY. Isolation and characterization of C-reactive protein (CRP) cDNA and genomic DNA from Xenopus laevis. A species representing an intermediate stage in CRP evolution. J Biol Chem 1993;268:6809e15. [32] Ishikawa N, Shigemoto K, Maruyama N. The complete nucleotide and deduced amino acid sequence of mouse serum amyloid P component. Nucleic Acids Res 1987;15:7186. [33] Dowton SB, McGrew SD. Rat serum amyloid P component. Analysis of cDNA sequence and gene expression. Biochem J 1990;270(2):553e6. [34] Dowton SB, Holden SN. C-reactive protein (CRP) of the syrian hamster. Biochemistry 1991;30:9531e8. [35] Wang CM, Nguyen NY, Yonaha K, Robey F, Liu TY. Primary structure of rabbit C-reactive protein. J Biol Chem 1982;257:13610e5. [36] Whitehead AS, Zahedi K, Rits M, Mortensen RF, Lelias JM. Mouse C-reactive protein. Generation of cDNA clones, structural analysis, and induction of mRNA during inflammation. Biochem J 1990;266:283e90. [37] Rassouli M, Sambasivam H, Azadi P, Dell A, Morris HR, Nagpurkar A, et al. Derivation of the amino acid sequence of rat C-reactive protein from cDNA cloning with additional studies on the nature of its dimeric component. J Biol Chem 1992;267:2947e54. [38] Nguyen NY, Suzuki A, Boykins RA, Liu TY. The amino acid sequence of Limulus C-reactive protein. Evidence of polymorphism. J Biol Chem 1986;261:10456e65. [39] Sinha S, Mandal C. Microheterogeneity of C-Reactive protein in the sera of fish Labeo rohita induced by metal pollutants. Biochem Biophys Res Commun 1996;226:681e7. [40] Mandal C, Sinha S, Mandal C. Lectin like properties and differential sugar binding characteristics of C-reactive protein purified from sera of normal and pollutant induced Labeo rohita. Glycoconj J 1999;16:741e50. [41] Lund V, Olafsen JA. Atypical phosphorylcholine-reactive protein from Atlantic salmon, Salmo salar L. Comp Biochem Biophys B 1998;119:471e7. [42] Sunyer JO, Lambris JD. Evolution and diversity of the complement system of poikilothermic vertebrates. Immunol Rev 1998;166:39e57. [43] Nakao M, Matsumoto M, Nakazawa M, Fujiki K, Yano T. Diversity of complement factor B/C2 in the common carp (Cyprinus carpio): three isotypes of B/C2-A expressed in different tissues. Dev Comp Immunol 2002;26:533e41. [44] Lovoll M, Fischer U, Mathisen GS, Bogwald J, Ototake M, Dalmo RA. The C3 subtypes are differentially regulated after immunostimulation in rainbow trout, but head kidney

Isolation of C-reactive protein homologues from cod macrophages do not contribute to C3 transcription. Vet Immunol Immunopathol 2007;117:284e95. [45] Sunyer JO, Zarkadis IK, Sahu A, Lambris JD. Multiple forms of complement C3 in trout that differ in binding to complement activators. Proc Natl Acad Sci U S A 1996;93:8451e6. [46] Lange S, Dodds AW, Magnadottir B. Isolation and characterisation of complement component C3 from Atlantic

219 cod (Gadus morhua L.) and Atlantic Halibut (Hippoglossus hippoglossus L.). Fish Shellfish Immunol 2004;16:227e39. [47] Lange S, Dodds AW, Gudmundsdottir S, Bambir SH, Magnadottir B. The ontogenic transcription of complement component C3 and apolipoprotein A-I tRNA in Atlantic cod (Gadus morhua L.) e a role in development and homeostasis? Dev Comp Immunol 2005;29:1065e77.