Developmental and Comparative Immunology 35 (2011) 441–451
Contents lists available at ScienceDirect
Developmental and Comparative Immunology
journal homepage: www.elsevier.com/locate/dci
Acute phase response in Chinese soft-shelled turtle (Trionyx sinensis) with
Aeromonas hydrophila infection
Xiuxia Zhou a,b , Lu Wang a,b , Hong Feng a,b , Qionglin Guo a,∗ , Heping Dai a
a
b
Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan 430072, China
Graduate School of Chinese Academy of Sciences, Beijing 100039, China
a r t i c l e
i n f o
Article history:
Received 21 September 2010
Received in revised form
12 November 2010
Accepted 14 November 2010
Available online 30 November 2010
Keywords:
Innate immunity
Acute phase response (APR)
Acute phase protein (APP)
Serum amyloid A (SAA)
Bacterial infection
Chinese soft-shelled turtle (Trionyx sinensis)
a b s t r a c t
Chinese soft-shelled turtle (Trionyx sinensis) is an important culture reptile. However, little is known
about its acute phase response (APR) caused by bacteria. Serum amyloid A (SAA) is a major acute phase
protein (APP). In this study, a turtle SAA homologue was identiﬁed and described in reptiles. The fulllength cDNA of turtle SAA was 554 bp and contained a 381 bp open reading frame (ORF) coding for a
protein of 127 aa. Similar to other known SAA genes, the turtle SAA gene contained three exons and
two introns. The promoter region of turtle SAA gene contained the consensus binding sites for nuclear
factor (NF)-␬B and c-Rel. The turtle SAA amino acid sequence shared the highest identity to avian SAA
sequences. Meantime, we present the ﬁrst systematic study with expression levels of ﬁve genes encoding
APPs in immune response caused by Aeromonas hydrophila infection. After infection, turtle SAA mRNA was
induced in liver at 8 h, then increased more than 1200-fold at 2 d; in spleen and kidney, the SAA mRNAs
were also induced during 8 h–7 d, but the level was far lower than that in the liver. The complement
3 (C3), ﬁbrinogen-gamma chain (Fb-G) and cathepsin L (CathL) mRNAs were increased in liver at 2 d,
whereas the albumin (ALB) mRNA was signiﬁcantly decreased during 8 h–7 d. Our studies suggest that
the APR in turtle with A. hydrophila infection is similar to that in mammals, and SAA is a major indicator
of bacterial infection, especially at early stage, in reptiles. Additionally, the different expression patterns
of ﬁve APP genes observed in present studies could provide clues for understanding the innate immune
mechanisms in the APR of reptiles.
Crown Copyright © 2010 Published by Elsevier Ltd. All rights reserved.
1. Introduction
Innate immunity plays an important role in early defense mechanisms and serves to initiate the acquired immune response. The
acute phase response (APR) is a complicated and systemic earlydefense system activated by tissue injury, infection, surgical trauma
and inﬂammation (Cray et al., 2009; Uhlar and Whitehead, 1999),
which results in a remarkable change in the concentrations of many
plasma proteins, known as acute phase proteins (APPs) (Gabay and
Kushner, 1999). In its broadest context, the APR is involved in many
changes at least the hepatic, neuroendocrine, hematopoietic, and
immune system. It is not clear whether all species have an acute
phase response. Although the APR is non-speciﬁc, it serves as a
core part of the innate immunity involving physical and molecular barriers and responses (Cray et al., 2009). Upon infection
and inﬂammation or tissue damage and stress, APR is induced by
pro-inﬂammatory signals, such as IL-1, IL-6 and TNF-␣, which are
∗ Corresponding author. Tel.: +86 027 68780003; fax: +86 027 68780123.
E-mail address: [email protected] (Q. Guo).
generated by activated cells including monocytes, macrophages,
ﬁbroblasts, and T cells, then APR rapidly evokes the changes of APPs
(Bayne and Gerwick, 2001).
The great majority of APPs are synthesized in hepatocytes,
also in extra-hepatic sites such as the brain and leukocytes
(Bayne and Gerwick, 2001). It responses quickly and becomes a
complicated but precise regulation network. APPs play an important role in a variety of the defense-related activities such as
killing infectious microbes, repairing tissue damage and restoring healthy (homeostatic) state (Murata et al., 2004). The APPs
(more than 200) have been grouped according to the extent to
which their concentration changes (major, moderate, and minor),
and the direction of changes (positive and negative) during APR
(Steel and Whitehead, 1994), or according to function (Gabay
and Kushner, 1999). Major APPs (concentrations may increase
10- to 100-fold, or up to 1000-fold) include serum amyloid A
(SAA), C-reactive protein (CRP), and haptoglobin (HP); moderate
APPs (concentrations may increase 2- to 10-fold) include complement components, and ﬁbrinogen (Fb); minor APPs (concentrations
only a slight increase) include cathepsin L (CathL); negative APPs
(concentrations decline) include albumin (ALB), pre-albumin and
0145-305X/$ – see front matter. Crown Copyright © 2010 Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.dci.2010.11.011
442
X. Zhou et al. / Developmental and Comparative Immunology 35 (2011) 441–451
transferin (Cray et al., 2009; reviewed in Ref. [Bayne and Gerwick,
2001]).
SAA is a major APP in mammals. The studies showed that the
gene encoding SAA from trout acts as an effective gene of innate
immunity which is known to be regulated by the Toll-like receptor
(TLR) signaling cascade. It has also been discussed that SAA may
even constitute an endogenous TLR4 ligand (reviewed in Ref. Rebl
et al., 2009). SAA homologs have been identiﬁed in all vertebrates
investigated and are highly conserved (Uhlar and Whitehead,
1999). In recent years, SAA homologs have also been identiﬁed and
characterized from some ﬁsh, such as arctic char (Salvelinus alpinus), common carp (Cyprinus carpio), Atlantic salmon (Salmo salar),
zebraﬁsh (Danio rerio) and rainbow trout (Oncorhynchus mykiss)
(Jensen et al., 1997; Fujiki et al., 2000; Jorgensen et al., 2000; Lin
et al., 2007; Rebl et al., 2009). Although a tuatara (Sphenodon punctatus) SAA homologue in reptile (AAM46103) has been cloned, no
report for reptilian APP has been found yet. Moreover, the wider
application of APPs in veterinary medicine was not reported until
the early 1990s.
Chinese soft-shelled turtle (Trionyx sinensis) is commercially
cultured in Southeast Asia, especially in China, Japan and Taiwan
area, for its nutritious and medical values. However, infectious diseases caused by bacteria and viruses have caused severe losses
to the turtle culture industry. When screening a subtracted cDNA
library from Chinese soft-shelled turtle experimentally infected
with Aeromonas hydrophila by suppressive subtractive hybridization (SSH), we isolated some APP cDNA fragments (ESTs) (Zhou
et al., 2008). Herein, we further report the molecular characteristics of turtle SAA, and ﬁve APP expression proﬁles during bacterial
infection by real-time quantitative PCR (RQ-PCR). These present
studies will help us to better understand the evolution of the SAA
molecule, and its innate immune mechanisms in the anti-bacterial
response of reptiles.
2. Materials and methods
2.1. Chinese soft-shelled turtles and bacterial infection
Chinese soft-shelled turtles (T. sinensis), 50–110 g in body
weight, were purchased from a local turtle culture farm, and
kept in clean tanks for 7 days. These turtles showed no clinical signs or laboratory evidence of Aeromonas or other infections.
Thirty turtles were intraperitoneally injected with freshly prepared
A. hydrophila (a Gram-negative bacterium, Strain T4, 1.0 × 108
CFU/50 g body weight), a main pathogen originally isolated from
clinical diseased soft-shelled turtles, and then introduced into
a clean tank. Five were injected with sterile water as control.
Speciﬁc bacterial infection was conﬁrmed by HE stain analysis
of the liver, kidney and spleen. The isolated bacteria from the
liver, kidney and intestine were conﬁrmed to be A. hydrophila
by culture, Gram-staining and PCR assays of four virulence genes
(aerolysin, metalloprotase, hemolysin, and ser-protease genes)
(data not shown).
concentrations were determined by a spectrophotometer. Then the
RNA was stored at −20 ◦ C.
2.3. Five soft-shelled turtle APP cDNA fragments and rapid
ampliﬁcation of cDNA end (RACE)
Five soft-shelled turtle APP cDNA fragments (SAA, FF281765;
Fb-G, FF281766; C3, FF281762; CathL, FF281777; ALB, FF281769)
were initially isolated from an SSH cDNA library constructed with
the mixed liver, spleen, and kidney tissue of A. hydrophila infected
turtles (Zhou et al., 2008). The RNA of the mixed tissues was used
as template to amplify the cDNA fragments of the turtle SAA.
Primers used for cDNA and DNA cloning were shown in Table 1.
5 RACE was performed using SMART RACE cDNA Ampliﬁcation
kit (Clontech) according to the manufacturer’s instructions. Genespeciﬁc primers of SAA-3F, SAA-R1 and SAA-R2 were designed
based on the turtle SAA cDNA fragments. Brieﬂy, the primers, UPM
and primer SAA-R1 or SAA-R2 were, respectively, used for 5 RACE
under the conditions of 94 ◦ C denaturation for 3 min, running 30
cycles of 94 ◦ C 30 s; 56 ◦ C 30 s; 72 ◦ C 1 min, and 72 ◦ C elongation for
7 min.
For 3 RACE, the cDNA template was transcribed by AMV Reverse
Transcriptase (TaKaRa) with Oligo dT-adaptor primer (Table 1). PCR
was performed with the primers of 3 adaptor (Table 1) and SAA-3F
under the conditions of 94 ◦ C denaturation for 2 min, running 30
cycles of 94 ◦ C 30 s; 58 ◦ C 30 s; 72 ◦ C 1 min, and 72 ◦ C elongation for
7 min.
2.4. Cloning genomic sequence and promoter region
Genomic DNA was puriﬁed from the turtle liver by the phenol
chloroform method (Sambrook and Russell, 2001). The primer SAAGF was designed according to the 5 -untranslated region (UTR) of
the full-length cDNA of SAA. 25 ng of genomic DNA was used for
the genomic PCR with an Ex Taq HS (TaKaRa) using SAA-GF and
SAA-GR (Table 1). PCR was performed with an initial denaturation
step of 5 min at 95 ◦ C, and then 35 cycles were run as follows: 94 ◦ C
30 s; 52 ◦ C 30 s; 72 ◦ C 3 min, and 72 ◦ C elongation for 10 min.
To obtain the 5 ﬂanking region, genome walking approach was
used by constructing genomic libraries with a Universal Genome
WalkerTM Kit (Clontech). Each of the 2.5 ␮g genomic DNA was completely digested with DraI, EcoRV, PvuII or StuI. Then four pools
of adaptor-ligated DNA fragments were constructed. The primers
SAA-P1 and SAA-P2 (Table 1) were designed according to the 5 end
of SAA cDNA. Together with the adaptor primers AP1 and AP2, two
rounds of PCR were performed for the ampliﬁcation of 5 ﬂanking
region. The cycling protocol included a two-step method for longdistance PCR. The primary PCR was performed with a hot start at
94 ◦ C for 2 min; 6 cycles of 94 ◦ C 30 s, 72 ◦ C 3 min; and 30 cycles of
94 ◦ C 30 s, 67 ◦ C 3 min, followed by 67 ◦ C for 10 min. The secondary
PCR was carried out with 1 ␮L of the ﬁrst round PCR mixture under
the conditions of 20 cycles of 94 ◦ C 25 s and 67 ◦ C 3 min, followed
by 67 ◦ C for 10 min.
2.5. TA cloning, sequencing and database analysis
2.2. Sampling and RNA extraction
Thirty infected turtles were, respectively, euthanized at 8 h, 24 h
(1 d), 2 d, 4 d, and 7 d after infection (5 infected turtles as a group,
n = 5). Various tissues of the infected turtles at each time point
above and control turtles were collected and washed with DEPCtreated saline, respectively. These samples were frozen in liquid
nitrogen.
Total RNA was extracted from these frozen tissues using Trizol
reagent (Invitrogen) according to the manufacturer’s instructions.
The integrity was ensured by analysis on a 1.5% agarose gel and
PCR products were separated by agarose gel electrophoresis,
puriﬁed using a Gel Extraction kit (OMEGA), and then ligated into
pMD18-T vectors (TaKaRa) and transformed into competent E. coli
DH5␣ cells. Positive colonies were screened by PCR and at least two
recombinant plasmids were sequenced by dideoxy chain termination using an automatic DNA sequencer (ABI Applied Biosystems
Model 3730).
Sequences were analyzed based on nucleotide and protein databases using the BLASTN and BLASTX program
(http://www.ncbi.nlm.nih.gov/BLAST/). The protein and its
X. Zhou et al. / Developmental and Comparative Immunology 35 (2011) 441–451
443
Table 1
Primers used for cDNA and DNA cloning (F forward and R reverse primers).
Primers
Sequence (5 –3 )
Application
SAA-3F
SAA-R1
SAA-R2
SAA-GF
SAA-GR
SAA-P1
SAA-P2
Oligo dT Adaptor
3 adapter
SMART II A Oligo
5 CDS
UPM
AP1
AP2
␤-Actin-F
␤-Actin-R
TGGCAGAGGAGCAGAAGAC
ATGTCACCCTGCCAACCTTC
TGTCATCGTAAGCACGCCAC
GCAGTACAGCAGTGCCTAAATAAG
CCTCCATTTCTACCCCAGGCATT
CAGTTCTGGGCACTCACACACAGTAC
CACTTATTTAGGCACTGCTGTACTGC
GGCCACGCGTCGACTAGTAC(T)17
GGCCACGCGTCGACTAGTAC
AAGCAGTGGTATCAACGCAGAGTACGCGGG
(T)25VN
CTAATACGACTCACTATAGGGC
GTAATACGACTCACTATAGGGC
ACTATAGGGCACGCGTGGT
GTGATGGTGGGAATGGGTC
ATGGCTGGGGTGTTGAAGGT
3 RACE-PCR
5 RACE ﬁrst round PCR and RT-PCR
5 RACE second round PCR and genomic PCR
Genomic PCR
Genomic PCR
Genomic walking ﬁrst round PCR
Genomic walking second round PCR
3 RACE cDNA synthesis
3 RACE PCR
5 RACE cDNA synthesis
5 RACE cDNA synthesis
5 RACE-PCR
Genomic walking ﬁrst round PCR
Genomic walking second round PCR
RT-PCR control
RT-PCR control
topology prediction were performed using software at the
ExPASy Molecular Biology Server (http://expasy.pku.edu.cn).
The protein family signature was identiﬁed by InterPro
(http://www.ebi.ac.uk/interpro/). Multiple sequence alignment was carried out using the CLUSTALW 1.81 program
(http://clustalw.genome.jp/) and the sequence identities were
calculated using GeneDoc (http://www.psc.edu/biomed/genedoc).
A phylogenetic tree was constructed using the neighbor-joining
(NJ) method in the Mega3.1 software package (Kumar et al., 2004).
2.6. Tissue distribution of ﬁve soft-shelled turtle APP mRNAs
The RNAs (10 ␮g) extracted from various tissues of the control
turtles were respectively treated with RNase-free DNase I (TaKaRa).
The ﬁrst strand of cDNA was synthesized using AMV Reverse Transcriptase and oligo (dT)18 (TaKaRa). The total amount of cDNA was
calibrated on the basis of the ampliﬁcation of turtle ␤-actin. The
cDNA was properly diluted and used as a template in PCR reactions. The RT-PCR primers of ﬁve soft-shelled turtle APP fragments
were listed in Table 2. The PCR condition was: initial denaturation
at 94 ◦ C for 2 min, 30 cycles of 94 ◦ C 30 s, 58-60 ◦ C 30 s, 72 ◦ C 30 s,
followed by 72 ◦ C for 7 min. The PCR products were electrophoresed
on a 1.5% agarose gel stained with ethidium bromide.
2.7. Quantiﬁcation of ﬁve APP mRNAs in infected turtles
Plasmids containing APP cDNAs were, sequenced to make sure
the PCR ampliﬁcations were correct. Then standard curves were
constructed by using these tenfold serial diluted plasmids, respectively.
Primers listed in Table 2 were also designed for quantitative
analysis of APP mRNAs in infected turtles. RQ-PCR was performed
with Chromo 4TM Continuous Fluorescence Detector from MJ
Research using SYBR Green Realtime PCR Master Mix (TOYOBO).
Each tissue sample assay was performed in triplicate. Each PCR
ampliﬁcation was carried out following the conditions: 2 min at
94 ◦ C, followed by 42 cycles consisting of 94 ◦ C 10 s, 58–60 ◦ C 15 s,
72 ◦ C 20 s, and a ﬁnally 72 ◦ C 10 min. The reactions performed without DNA sample were used as negative control. A standard curve
was constructed by using tenfold serial diluted plasmids containing
APP cDNAs. RQ-PCR reactions of the standard curves were always
included in all runs in order to relate quantitative data from run
to run. Concentration of cDNA in each sample was calculated from
the standard curve. Melting curve analysis of ampliﬁcation products was performed at the end of each PCR reaction to conﬁrm
that only one PCR product was ampliﬁed and detected. The ␤-actin
gene was used as internal standard in all RQ-PCR experiments. The
data obtained from the RQ-PCR analysis were subjected to one-way
analysis of variance (one-way ANOVA) using SPSS 13.0 software.
3. Results
3.1. Sequence analysis of soft-shelled turtle SAA cDNA
The full-length cDNA of soft-shelled turtle SAA was 554 bp (GenBank accession no. HQ186287) and contained a 381 bp open reading
frame (ORF) coding for a protein of 127 amino acids (aa), and
had a 116 bp of 5 -UTR and a 109 bp of 3 -UTR including a 31 bp
poly (A). The polyadenylation signal AATAAA was 14 bp upstream
of poly (A) (Fig. 1). The Kozak sequence -(A/G)NNATG-, recognized by ribosomes as the translational start site, was present
within the 5 UTR sequence of turtle SAA. The deduced turtle
SAA amino acid sequence contained an N-terminal signal peptide
(1–18 aa) (Fig. 1). The isolated turtle SAA gene was 3122 bp (GenBank accession no. HQ186288) and contained a 1265 bp promoter
Table 2
Primers used for RT-PCR and RQ-PCR (F forward and R reverse primers).
Primers
Sequence (5 –3 )
Accession no. of EST
SAA-QF
SAA-QR
C3-QF
C3-QR
CL-QF
CL-QR
Fb-G-QF
Fb-G-QR
ALB-QF
ALB-QR
␤-Actin-QF
␤-Actin-QR
TGTGTGTGAGTGCCCAGAAC
TGTCATCGTAAGCACGCCAC
CCAGGAGCTGTCAAGGTCTATGA
GGCAAATATCCCCGTGGCAGA
GGTCCAGTCTCTGTGGCTATTG
CCATCTTCGTCTGCTCCCTGA
TCACGCTGCTAACCTCAATGGC
CATGGAATACCACCGAGAACGC
GGATTGTATGCAC GA AA GGGTAG
GCAGGTTTGTCATCATTGTCCA
GAGACCCGACAGACTACCT
AGGATGATGAAGCAGCAGT
FF281765
96
FF281762
115
FF281777
149
FF281766
124
FF281769
147
EO727174
156
Products (bp)
444
X. Zhou et al. / Developmental and Comparative Immunology 35 (2011) 441–451
Fig. 1. Genomic sequence and deduced amino acid sequence of turtle SAA gene. Exons and predicted amino acid sequences are shown in upper case, whereas introns are
shown in lower case. Intron splice sites (gt. . .ag) are indicated in italics and bolded. The start codon (ATG) is boxed and the stop codon (TAA) is marked by an asterisk. The
predicted signal peptide sequence is underlined. The polyadenylation signal AATAAA is in bold.
sequence in the 5 ﬂanking region. Typical intron splice motifs were
present at the 5 (GT) and 3 (AG) ends of each intron (Fig. 1). The
turtle SAA gene had three exons and two introns, which is similar to other known SAAs genomic organization (Fig. 2). Exon 1
encodes the 5 -UTR and the N-terminal 29 aa residues of the SAA
protein. Exon 2 encodes the residual 45 aa and exon 3 encodes
47 aa.
After the deﬁnition of the transcriptional start site of turtle SAA gene, the 1207 bp 5 ﬂanking region was identiﬁed by
genomic walking (Fig. 3). TATA box was found 24–29 bp upstream
X. Zhou et al. / Developmental and Comparative Immunology 35 (2011) 441–451
445
Fig. 2. Genomic structure and organization of the turtle SAA gene compared with other species. The three exons are indicated by rectangles and two introns by lines. The
sizes of exons are indicated in the rectangles, and the sizes of introns are indicated above the lines.
Fig. 3. Promoter region of turtle SAA gene. Putative binding sites for TATA box, Oct-1, C/EBP ␤, NF-␬B, ARP-1, HSF and HLF are underlined. The c-Rel sites are bolded and
underlined.
of the start codon so that the transcriptional start site was veriﬁed. Computational analysis of the promoter sequence revealed
some putative binding sites for several important transcription
factors, including octamer-binding protein 1 (Oct-1, −62 to −76),
CCAAT/enhancer binding protein ␤ (C/EBP ␤, −90 to −99), nuclear
factor (NF)-␬B (−171 to −180), three c-Rel (−227 to −236, −799
to −808, −1051 to −1060), two apolipoprotein regulatory protein 1 (ARP-1, −690 to −705, −989 to −1004), heat shock factor
(HSF, −799 to −813), hepatic leukemia factor (HLF, −1100 to
−1109).
3.2. Similarity comparison and phylogenetic analysis of
soft-shelled turtle SAA
The deduced amino acid sequence of turtle SAA protein is
remarkably conserved (Table 3). It shared the highest identity
(74%) to duck SAA sequence, with 72–74% identity to avian SAA
sequences, and 62–69% identity to ﬁsh and mammalian SAA
sequences, respectively. Similarity between the protein sequences
of the only two reptile SAA molecules, soft-shelled turtle and
tuatara SAA, is 72%.
446
X. Zhou et al. / Developmental and Comparative Immunology 35 (2011) 441–451
Table 3
Amino acid identity comparison of the turtle SAA protein with other known SAA proteins.
Species
Common name
Proteins
Accession no.
Identity (%)
Anas platyrhynchos
Monodelphis domestica
Sphenodon punctatus
Anser anser domesticus
Taeniopygia guttata
Ornithorhynchus anatinus
Oryctolagus cuniculus
Canis familiaris
Homo sapiens
Homo sapiens
Felis catus
Equus caballus
Macaca mulatta
Acinonyx jubatus
Stichopus japonicus
Branchiostoma belcheri
Oncorhynchus mykiss
Holothuria glaberrima
Duck
Opossum
Tuatara
Goose
Poephila guttata
Plalypust
Rabbit
Dog
Human
Human
Cat
Horse
Rhesus monkey
Cheetah
Stichopus japonicus
Japanese lancelet
Rainbow trout
Sea cucumber
SAA
SAA
SAA
SAA precursor
SAA A
SAA
SAA
SAA precursor
SAA1
SAA2
SAA
SAA
SAA1
SAA
SAA
SAA
SAA
SAA
P02740
XP 001379259
AAM46103
AAV33332
XP 002198615
XP 001508384
NP 001075771
P19708
AAA64799
NP 110381
NP 001035288
XT 001505005
XP 001086724
BAG06986
ABX55S30
BAB97379
AM422447
AAG24633
74
72
72
73
72
69
64
65
62
62
62
65
66
62
66
63
62
62
Multiple sequence alignments were carried out using the
CLUSTALW 1.81 program (Fig. 4). The 18-aa signal peptide is shown
together with the 109-aa mature protein. The hydrophobic Nterminal portion of the molecule has been shown to be a major
determinant for amyloid formation (Westermark et al., 1992) and
C-terminal portion is the proposed neutrophil and GAG binding
region. Secondary structure predictions indicated that the turtle
SAA molecule is likely to contain two regions of ␣-helix and two ␤
strands (Fig. 4).
Phylogenetic and molecular evolutionary analysis was conducted using Mega 3.1. The phylogenetic NJ-tree, containing two
reptilian SAA sequences, was presented based on the homology of
their amino acid sequences of SAAs. As shown in Fig. 5, turtle and
tuatara SAAs clustered together with bird SAAs, while sequences of
mammals formed a separate cluster. The branches of ﬁsh and invertebrates SAA molecules were also included. The different degrees
of divergence among reptilian, mammalian, avian, ﬁsh and invertebrates SAAs may reﬂect their phylogenetic difference.
3.3. Tissue expression of ﬁve soft-shelled turtle APP mRNAs
The turtle SAA mRNA distribution was examined using semiquantitative RT-PCR in control turtles (Fig. 6A). Normalized with ␤actin, no SAA transcript was detected in liver, spleen, kidney, heart,
intestine and blood tissue of control turtles (Fig. 6A: Left ﬁgure). A
signiﬁcant induction of SAA was detected in liver after A. hydrophila
infection, while a weak expression was also detected in the kidney
and spleen (Fig. 6A: Right ﬁgure).
As APPs are mainly synthesized in liver during APR, the other
four APP (C3, CathL, Fb-G and ALB) mRNAs were tested only in the
liver. Three individuals were tested for each APP mRNA. As shown in
Fig. 6B, they were all constitutively expressed in the liver of control
turtles.
3.4. Real-time quantiﬁcation of ﬁve APP mRNAs in infected turtles
The ampliﬁcation speciﬁcities for APPs and ␤-actin were determined by analyzing the melting curves. Only one peak presented in
the melting curves for two genes above, indicating that the ampliﬁcations were speciﬁc. The inducible expression of turtle SAA in
various tissues at different time point after A. hydrophila infection
was shown in Fig. 7. After infection, in liver, turtle SAA mRNA was
induced at 8 h, and increased more than1200-fold at 2 d (p < 0.001),
then about 170-fold at 7 d; in spleen, the SAA mRNA was induced
during 8 h–7 d and increased 6-fold (p < 0.001) at 1 d; in kidney, the
SAA mRNA was induced at 8 h, increased 3.5-fold at 2 d (p < 0.001).
In liver, the Fb-G mRNA increased about 5-fold at 1 d (p < 0.001),
and 20-fold at 2 d (p < 0.001); the C3 mRNA was slightly decreased
during 8 h–1 d, and up-regulated 5.8-fold at 2 d (p < 0.001), then
decreased; the CathL mRNA increased 2.6-fold at 2 d (p < 0.05),
while the ALB mRNA was signiﬁcantly decreased during 8 h–7 d
(p < 0.001) (Fig. 8).
4. Discussion
4.1. Turtle SAA homologue is highly conserved, and clusters
together with bird SAAs in phylogenetic tree
In the present work, we described for the ﬁrst time the molecular
characterization of the SAA in reptiles. Our results strongly support
that this sequence obtained from Chinese soft-shelled turtle is a
mammalian SAA homologue. Similar to other known SAAs (Fig. 4),
turtle SAA contained a signal sequence of 18 amino acids. The neutrophil and GAG binding region were well conserved in turtle SAA.
An octapeptide insertion was lack in turtle SAA. While the aa insertion is present in horse and cattle SAA. Human and mouse SAA4
also contain the aa insertion which is a characteristic of constitutive
SAAs (Lin et al., 2007). The length of turtle SAA protein (127 aa) is
similar to those of trout, zebraﬁsh and pufferﬁsh Tetraodon (121 aa)
and carp (123 aa) (Rebl et al., 2009). Almost all of the results in this
study suggest that the turtle SAA may be a functionally conserved
protein.
The ﬁrst reptilian SAA gene was identiﬁed and it shared the same
genomic organization with other known SAA genes. It contains a
1265 bp promoter region, in which several putative binding sites for
TATA box, NF-␬B, C/EBP ␤, c-Rel (3 sites), oct-1, ARP-1 (2 sites), HLF
and HSF are present, suggesting that the transcription of turtle SAA
may be regulated by multiple transcription factors in inﬂammatory
response. Among these factors, NF-␬B is a major transcription factor that regulates the genes responsible for both the innate and the
adaptive immune response (Thapa et al., 2008); c-Rel, a member of
NF-␬B family, is predominantly expressed in immune cells. It regulates gene transcription in T cells and is implicated in leukocyte
trafﬁcking, is particularly related to Th1-mediated inﬂammation
(Bunting et al., 2007). Three binding sites for c-Rel locate within
turtle SAA gene promoter region, which suggests that c-Rel plays
an important role in controlling turtle SAA gene transcription in
inﬂammatory responses. In early studies, the combined effect of
mammalian SAA promoters proved that NF-␬B and C/EBP-like are
essential for inducing transcriptional activation of SAA (Li and Liao,
1992; Ray et al., 1995).
X. Zhou et al. / Developmental and Comparative Immunology 35 (2011) 441–451
447
Fig. 4. Multiple alignment of the amino acid sequences of turtle SAA with other known SAAs. Black shaded sequence indicates positions that have a fully conserved residue,
gray shaded sequence indicates conserved amino acid substitutions, light gray shaded sequence indicates semi-conserved amino acid substitutions, and dashes indicate gaps.
Signal peptide is indicated upon the amino acid residues. N-terminal hydrophobic region is underlined. The predicted tertiary structure (␣-helix and ␤-sheet) is boxed, based
on the structure of human SAA (Uhlar and Whitehead, 1999). GenBank accession numbers for these SAA protein sequences used are listed in Table 3.
4.2. SAA is a major indicator of bacterial infection, especially at
the early stage, in reptiles
It is generally accepted that APPs are inductors of a proinﬂammatory reaction and fever, their over-expression can lead to
an anti-inﬂammatory response. Thus, APPs are used today as potential biological markers for monitoring animal welfare and health
status (Ceciliani et al., 2002; Petersen et al., 2004; Pallarés et al.,
2008). SAA are considered as main APP in mammals. Major APPs
are often observed to increase markedly with the ﬁrst 48 h after the
triggering event and have a rapid decline due to their short half-life
(reviewed in Ref. Cray et al., 2009). The aim of this study was mainly
to determine the immune changes (APR) caused by bacterial infec-
tion, and to highlight the valuable APPs in diagnosing reptiles with
inﬂammation. As shown in Fig. 7, after A. hydrophila infection a
remarkable APR is evoked. Our studies revealed that SAA mRNA
was signiﬁcantly up-regulated in turtle liver during 8 h–1 d and
reached 1200-fold at 2 d. This is agreement with that in carp skin
(1600-fold at 36 h) infected by Ichthyophthirius multiﬁliis (Gonzales
et al., 2007), which emphasizes the importance of the ﬁsh skin as
one of the main sites of SAA production during the APR. Meantime,
SAA mRNAs were also up-regulated in the spleen and kidney. Presumably, the expression may be due to leukocytes. This suggests a
critical protective role of the APR in reptiles.
It is reported that the SAA levels had 100% sensitivity, while neutrophil counts had much lower sensitivity and speciﬁcity (30–70%).
448
X. Zhou et al. / Developmental and Comparative Immunology 35 (2011) 441–451
Fig. 5. Phylogenetic analysis of the deduced SAA amino acid sequences. Neighbor joining tree was constructed with Mega3.1 program. The numbers on the branches represent
the conﬁdence level of 10,000 bootstrap replications. The bar indicates the substitution rate per residue. The GenBank accession numbers used are listed in Table 3.
In a cat with pancreatitis, SAA was increased with onset of symptoms, whereas the white blood cell (WBC) count was normal. With
rapid resolution of the symptoms in this cat, SAA returned to normal levels, whereas the WBC count was just beginning to increase
(Tamamoto et al., 2009). In mammals, SAA is synthesized extrahepatically by different cellular types like monocytes/macrophages
(Yamada et al., 2000). SAA has been demonstrated to result in the
chemotaxis of monocytes, T cells and polymorphonuclear leukocytes (Ceciliani et al., 2002; Petersen et al., 2004). In fact, the
up-regulation of turtle IL-8 gene (Zhou et al., 2009), including the
neutrophils increase observed in liver section of infected turtle (Figure not shown), suggest that the turtle SAA could be also involved in
the attraction of neutrophils to the site of inﬂammation. Undoubtedly, the changes of SAA level should have higher sensitivity than
that of neutrophil counts. In other words, the SAA could result
in neutrophil increase, and it may be more sensitive indicator of
inﬂammation. Moreover, in trout, SAA may constitute an endogenous TLR4 ligand (reviewed in Ref. Rebl et al., 2009). TLR activation
recruits several downstream factors regulating the expression of
immune relevant genes (such as pro-inﬂammatory cytokine genes).
The pattern recognition functions of SAA have also been identiﬁed with binding to a range of Gram-negative bacteria (Hari-Dass
et al., 2005). Thereby, SAA up-regulation observed in this study may
contribute to the establishment of an efﬁcient reptilian immune
Fig. 6. RT-PCR analysis of APPs in different tissues of soft-shelled turtle. (A) Expression and tissue distribution analysis of turtle SAA mRNA in various tissues of infected turtles
(Right ﬁgure) and control turtles (Left ﬁgure). (B) Constructively expressed C3, Fb-G, CathL and ALB in turtle liver. Different turtle individuals were represented numerically.
Expression of a gene encoding ␤-actin was used as the control. Marker: DL2000.
X. Zhou et al. / Developmental and Comparative Immunology 35 (2011) 441–451
449
Fig. 7. Expression analysis of SAA gene in the liver, spleen and kidney of infected and control turtles at different time points by RQ-PCR. ␤-actin was used as internal control.
Data are expressed as the mean + SD for triplicate samples in two experiments. The signiﬁcant differences to the control (p < 0.05 and p < 0.001) are denoted with ‘*’ and ‘**’,
respectively.
Fig. 8. Expression analysis of Fb-G, C3, CathL and ALB genes in the liver of infected and control turtles at different time points by RQ-PCR. ␤-actin was used as internal control.
Data are expressed as the mean + SD for triplicate samples in two experiments. The signiﬁcant differences to the control (p < 0.05 and p < 0.001) are denoted with ‘*’ and ‘**’,
respectively.
450
X. Zhou et al. / Developmental and Comparative Immunology 35 (2011) 441–451
response and TLR signaling pathway, and serve as a main APP and
biomarker of infection, especially at the early stage, in reptiles.
4.3. Other three APPs (C3, Fb-G and CathL) were involved in
anti-bacterial immune in a time-dependent manner
The complement system is one of the major effective arms of
immune responses in vertebrates. Activation of the complement
system results in cleavage of C3, followed by cleavage of C5, the
latter component forms a ring-shaped membrane attack complex
(MAC, C5b-9) (Favoreel et al., 2003). Activated complement elicits some potent biological activities, such as the promotion of
inﬂammation, the lysis of microorganisms or target cells and the
opsonization of pathogens or immune complexes with complement activation products (Mollnes et al., 2002; Walport, 2001).
Level of C3 (a moderate APP) can increase 2- to 10-fold in the
course of mammalian APR. In this study, turtle C3 mRNA was upregulated about 5.8-fold at 2 d in the infected liver, striking similar
to that in mammals (human and mouse), and zebraﬁsh in which
C3 mRNA also increased 4.3-fold after Aeromonas salmonicida (a
Gram-negative bacterium) infection (Lin et al., 2007). Our results
demonstrate that the turtle C3 may play an essential role in antiinfection as in other species. Whereas, in rainbow trout, within
just 10 min of the initiation of acute stressors, C3 was observed to
increase in plasma (reviewed in Ref. Bayne and Gerwick, 2001). The
short response time seen in trout was interpreted to mean that the
response was probably not due to the classical APR. It appears likely
that a pool of pre-synthesized C3 is held in store and released to the
plasma as part of a pre-acute phase response (Bayne and Gerwick,
2001).
Among the three up-regulated genes, Fb-G is noteworthy. Fb
can provide a substrate for ﬁbrin formation in homeostasis, and
a matrix for the migration of inﬂammatory-related cells in tissue repair. Fb speciﬁcally binds to CD11/CD18 integrins on the
cell surface of migrated phagocytes, leading the enhancement of
degranulation, phagocytosis, antibody-dependent cellular cytotoxicity and delay of apoptosis (Sitrin et al., 1998; Rubel et al., 2001).
Although Fb is a moderate APP (concentrations may increase 2- to
10-fold) in human, cow, goat, horse, mouse and rabbit, it has been
used as a reliable indicator of the presence of inﬂammation, bacterial infection or surgical trauma (Cray et al., 2009; reviewed in
Ref. Murata et al., 2004). However, in this study, Fb-G mRNA could
be increased 20-fold at 2 d in the liver of infected turtles, suggesting that turtle Fb-G may play a more important role than that of
other species involving in blood coagulation course, tissue repair,
and protect bacterial diffusion. Although the diagnostic accuracy of
plasma Fb in traumatic reticuloperitonitis in cattle was signiﬁcantly
lower than either SAA or Hp (Naziﬁ et al., 2009), our results showed
that in turtle, Fb-G may be a useful indicator for early infective
diagnosis, or may serve as a major APP.
CathL was also identiﬁed in infected turtles to increase at 2 d.
CathL, a potent lysosomal cysteine protease primarily responsible for degradation and turnover of intracellular proteins, has been
implicated in a variety of physiological and pathological processes
including antigen presentation, pro-hormone activation, and CD4
+ cell selection (Villadangos et al., 1999; Honey et al., 2002). In
common condition, CathL stores in lysosome in an enzymogenshape, but during pathological injury (by microbes, inﬂammatory
factor and stress), a large amount of CathL can release into the
cytoplasm and intracellular tissue. Chicken CathL mRNA was upregulated in monocyte-derived macrophages (MDM) under avian
pathogenic E. coli infection (Lavric et al., 2008). Chinese white
shrimp (Fenneropenaeus chinensis) CathL was also up-regulated in
the hepatopancreas with white spot syndrome virus (WSSV) infection, and was proposed to play a role in shrimp innate immunity
(Ren et al., 2010). Thus, the results of the present study, taken
together with previous report, suggest that turtle CathL can be
involved in inﬂammatory course, and like other known species
CathL may serve as a minor APP in reptiles.
4.4. ALB is also a negative APP of bacterial infection in reptiles
As described above, with bacterial infection, the pathogen needs
to be neutralized, and the damage tissue needs to be cleared away,
the mRNA level of some APP genes will be different from that in
a healthy, homeostatic state. Consequently, the concentration of
some defense and ‘clean-up’ molecules in the blood plasma is likely
increased. To avoid increasing of the osmotic pressure and viscosity
of blood, increases in levels of some plasma molecules are accompanied by decreases in others during the APR (Bayne and Gerwick,
2001; Cray et al., 2009).
ALB, only synthesized in liver, is the most abundant protein in
human blood plasma. During the APR, the decrease in ALB synthesis
is postulated to allow for the unused pool of amino acids to instead
be used to generate positive APPs and other important mediators
of inﬂammation (Paltrinieri, 2008). Jensen et al. (1997) reported
that ALB mRNA levels declined late in the course of salmonids by A.
salmonicida infection. In this study, as shown in Fig. 8, ALB mRNA
was signiﬁcantly down-regulated in infected turtle liver during
8 h–7 d. In the APR, the plasma proteins that decrease by 25% or
more are called negative APP. Undoubtedly, ALB is also a negative
APP of bacterial infection in reptiles, although we do not know the
exact decreased mechanism.
A single APP should not be used exclusively to monitor a disease process. An assay of many APPs (including both positive and
negative APPs, as well as APPs that increase both rapidly and
slowly) has been used in both human and veterinary medicine
(Gruys et al., 2006). In this study, we present the ﬁrst systematic
study with expression levels of ﬁve genes encoding APPs caused by
A. hydrophila infection. Our results provide new insights into the
innate immunity of reptiles, suggesting that the APPs can serve as
potential applications in the diagnosis of reptilian infective disease
and/or homeostasis.
Acknowledgements
This work was supported by the National Natural Science Foundation of China (Grant no. 30070588 and 30871912)
and the National Basic Research Program of China (Grant no.
2009CB118704). We thank Prof. Xudong Xu and Zhan Yin (Institute
of Hydrobiology, Chinese Academy of Sciences) for their technical
assistance and manuscript correction.
References
Bayne, C.J., Gerwick, L., 2001. The acute phase response and innate immunity of ﬁsh.
Dev. Comp. Immunol. 25, 725–743.
Bunting, K., Rao, S., Hardy, K., Woltring, D., Denyer, G.S., Wang, J., Gerondakis, S.,
Shannon, M.F., 2007. Genome-wide analysis of gene expression in T cells to identify targets of the NF-␬B transcription factor c-Rel. J. Immunol. 178, 7097–7109.
Ceciliani, F., Giordano, A., Spagnolo, V., 2002. The systemic reaction during inﬂammation: the acute-phase proteins. Protein Pept. Lett. 9, 211–223.
Cray, C., Zaias, J., Altman, N.H., 2009. Acute phase response in animals: a review.
Comp. Med. 59, 517–526.
Favoreel, H.W., Gerlinde, R., de Walle, V., Nauwynck, H.J., Pensaert, M.B., 2003. Virus
complement evasion strategies. J. Gen. Virol. 84, 1–15.
Fujiki, K., Shin, D.-H., Nakao, M., Yano, T., 2000. Molecular cloning and expression
analysis of carp (Cyprinus carpio) interleukin-1␤ high afﬁnity immunoglobulin E
Fc receptor ␥ subunit and serum amyloid A. Fish Shellﬁsh Immunol. 10, 229–242.
Gabay, C., Kushner, I., 1999. Acute-phase proteins and other systemic responses to
inﬂammation. N. Engl. J. Med. 340, 448–454.
Gonzales, S.F., Buchmann, K., Nielsen, M.E., 2007. Ichthyophthirius multiﬁliis infection
induces massive up-regulation of serum amyloid A in carp (Cyprinus carpio). Vet.
Immunol. Immunopathol. 115, 172–178.
Gruys, E., Toussaint, M.J., Niewold, T.A., Koopmans, S.J., van Dijk, E., Meloen, R.H.,
2006. Monitoring health by values of acute phase protein. Acta Histochem. 108,
229–232.
X. Zhou et al. / Developmental and Comparative Immunology 35 (2011) 441–451
Hari-Dass, R., Shah, C., Meyer, D.J., Raynes, J.G., 2005. Serum amyloid A protein
binds to outer membrane protein A of gramnegative bacteria. J. Biol. Chem. 280,
18562–18567.
Honey, K., Nakagawa, T., Peters, C., Rudensky, A., 2002. Cathepsin L regulates CD4+
T cell selection independently of its effect on invariant chain: a role in the
generation of positively selecting peptide ligands. J. Exp. Med. 195, 1349–1358.
Jensen, L.E., Hiney, M.P., Shields, D.C., Uhlar, C.M., Lindsay, A.J., Whitehead, A.S.,
1997. Acute phase proteins in salmonids. Evolutionary analyses and acute phase
response. J. Immunol. 158, 384–392.
Jorgensen, J.B., Lunde, H., Jensen, L., Whitehead, A.S., Robertsen, B., 2000. Serum amyloid A transcription in Atlantic salmon (Salmo salar L.) hepatocytes is enhanced
by stimulation with macrophage factors, recombinant human IL-1 beta, IL-6 and
TNF alpha or bacterial lipopolysaccharide. Dev. Comp. Immunol. 24, 553–563.
Kumar, S., Tamura, K., Nei, M., 2004. MEGA3: integrated software for molecular evolutionary genetics analysis and sequence alignment. Brief Bioinform. 5, 150–163.
Lavric, M., Maughan, M.N., Bliss, T.W., Dohms, J.E., Bencina, D., Keeler Jr., C.L., Narat,
M., 2008. Gene expression modulation in chicken macrophages exposed to
Mycoplasma synoviae or Escherichia coli. Vet. Microbiol. 126, 111–121.
Li, X., Liao, W.S., 1992. Cooperative effects of C/EBP-like and NF kappa B-like binding
sites on rat serum amyloid A1 gene expression in liver cells. Nucleic Acids Res.
20, 4765–4772.
Lin, B., Chen, S., Cao, Z., Lin, Y., Mo, D., Zhang, H., Gu, J., Dong, M., Liu, Z., Xu, A., 2007.
Acute phase response in zebraﬁsh upon Aeromonas salmonicida and Staphyloccus
aureus infection: striking similarities and obvious differences with mammals.
Mol. Immunol. 44, 295–301.
Mollnes, T.E., Song, W.C., Lambris, J.D., 2002. Complement in inﬂammatory tissue
damage and disease. Trends Immunol. 23, 61–64.
Murata, H., Shimada, N., Yoshioka, M., 2004. Current research on acute phase protens
in veterinary diagnosis: an overview. Vet. J. 168, 28–40.
Naziﬁ, S., Ansari-Lari, M., Asadi-Fardaqi, J., Rezaei, M., 2009. The use of receiver
operating characteristic (ROC) analysis to assess the diagnostic value of serum
amyloid A, haptoglobin and ﬁbrinogen in traumatic reticuloperitonitis in cattle.
Vet. J. 182, 315–319.
Pallarés, F.J., Martinez-Subiela, S., Seva, J., Ramis, G., Fuentes, P., Bernabé, A., Munoz,
A., Cerón, J.J., 2008. Relationship between serum acute phase protein concentrations and lesions in ﬁnishing pigs. Vet. J. 177, 369–373.
Paltrinieri, S., 2008. The feline acute phase reaction. Vet. J. 177, 26–
35.
Petersen, H.H., Nielsen, J.P., Heegaard, P.M., 2004. Application of acute phase protein
measurements in veterinary clinical chemistry. Vet. Res. 35, 163–187.
Ray, A., Hannink, M., Ray, B.K., 1995. Concerted participation of NF kappa B and C/EBP
heteromer in lipopolysaccharide induction of serum amyloid A gene expression
in liver. J. Biol. Chem. 270, 7365–7374.
451
Rebl, A., Goldammer, T., Fischer, U., Köllner, B., Seyfert, H.M., 2009. Characterization of two key molecules of teleost innate immunity from rainbow trout
(Oncorhynchus mykiss): MyD88 and SAA. Vet. Immunol. Immunopathol. 131,
122–126.
Ren, Q., Zhang, X.W., Sun, Y.D., Sun, S.S., Zhou, J., Wang, Z.H., Zhao, X.F., Wang,
J.X., 2010. Two cysteine proteinases respond to bacterial and WSSV challenge
in Chinese white shrimp Fenneropenaeus chinensis. Fish Shellﬁsh Immunol. 29,
551–556.
Rubel, C., Fernandez, G.C., Dran, G., Bompadre, M.B., Isturiz, M.A., Palermo, M.S., 2001.
Fibrinogen promotes neutrophil activation and delays apoptosis. J. Immunol.
166, 2002–2010.
Sambrook, J., Russell, D., 2001. Molecular Cloning: A Laboratory Manual, 3rd ed. Cold
Spring Harbor Laboratory Press, New York.
Sitrin, R.G., Pan, P.M., Srikanth, S., Todd 3rd, R.F., 1998. Fibrinogen activates NF-kappa
B transcription factors in mononuclear phagocytes. J. Immunol. 161, 1462–1470.
Steel, D.M., Whitehead, A.S., 1994. The major acute phase reactants: C-reactive protein, serum amyloid P component and serum amyloid A protein. Immunol. Today
15, 81–88.
Tamamoto, T., Ohno, K., Ohmi, A., Seki, I., Tsujimoto, H., 2009. Time-course monitoring of serum amyloid A in a cat with pancreatitis. Vet. Clin. Pathol. 38, 83–86.
Thapa, D., Lee, J.S., Park, S.-Y., Bae, Y.-H., Bae, S.-K., Kwon, J.B., Kim, K.-J., Kwak, M.K., Park, Y.-J., Choi, H.G., Kim, J.-A., 2008. Clotrimazole ameliorates intestinal
inﬂammation and abnormal angiogenesis by inhibiting interleukin-8 expression
through a Nuclear Factor-␬B-dependent manner. J. Pharmacol. Exp. Ther. 327,
353–364.
Uhlar, C.M., Whitehead, A.S., 1999. Serum amyloid A, the major vertebrate acutephase reactant. Eur. J. Biochem. 165, 501–523.
Villadangos, J.A., Bryant, R.A., Deussing, J., Drissen, C., Lennon-Dumenil, A.M., Riese,
R.J., Saftig, P., Shi, G.P., Chapman, H.A., Peters, C., Ploeghl, H.L., 1999. Protease
involved in MHC class II antigen presentation. Immunol. Rev. 172, 109–120.
Walport, M.J., 2001. Complement. First of two parts. N Engl. J. Med. 344, 1058–1066.
Westermark, G.T., Engstrom, U., Westermark, P., 1992. The N-terminal segment of
protein AA determines its ﬁbrillogenic property. Biochem. Biophys. Res. Commun. 182, 27–33.
Yamada, T., Wada, A., Itoh, K., Igari, J., 2000. Serum amyloid A secretion from monocytic leukaemia cell line THP-1 and cultured human peripheral monocytes.
Scand. J. Immunol. 52, 7–12.
Zhou, X.X., Guo, Q.L., Dai, H.P., 2008. Identiﬁcation of differentially expressed
immune-relevant genes in Chinese soft-shelled turtle (Trionyx sinensis) infected
with Aeromonas hydrophila. Vet. Immunol. Immunopathol. 125, 82–91.
Zhou, X.X., Guo, Q.L., Dai, H.P., 2009. Molecular characterization and expression proﬁles in response to bacterial infection of Chinese soft-shelled turtle interleukin-8
(IL-8), the ﬁrst reptilian chemokine gene. Dev. Comp. Immunol. 33, 838–847.