at SciVerse ScienceDirect

Fish & Shellfish Immunology 35 (2013) 429e437

Contents lists available

Fish & Shellfish Immunology

journal homepage: www.elsevier .com/locate / fs i

A heat shock protein 90 b isoform involved in immune response tobacteria challenge and heat shock from Miichthys miiuy

Tao Wei a, Yunhang Gao b, Rixin Wang a,*, Tianjun Xu a,*

a Laboratory for Marine Living Resources and Molecular Engineering, College of Marine Science, Zhejiang Ocean University, 105 Wenhua Road, Zhoushan316000, PR ChinabCollege of Animal Science and Veterinary Medicine, Jilin Agriculture University, Changchun 130118, PR China

a r t i c l e i n f o

Article history:Received 21 January 2013Received in revised form8 April 2013Accepted 29 April 2013Available online 17 May 2013

Keywords:Heat shock protein 90 (HSP90)Miichthys miiuy (Miiuy croaker)Gene expressionWestern blotEvolutionary analysis

* Corresponding authors. Tel.: þ86 580 2550826.E-mail addresses: wangrixin1123@126.com (R.

(T. Xu).

1050-4648/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.fsi.2013.04.045

a b s t r a c t

Heat shock protein 90 (HSP90) is highly conserved molecular chaperone that plays a critical role incellular stress response. In this study, we reported the identification and functional analysis of a heatshock protein 90 gene from miiuy croaker (designated Mimi-HSP90). Mimi-HSP90 contained fiveconserved HSP90 protein family signatures and shared 89.6%e99.5% similarity with other known HSP90b isoform. Homology analysis and structure comparison further indicated that Mimi-HSP90 should be bisoformmember of the HSP90 family. The molecular evolutionary analysis showed that HSP90 was underan overall strong purifying select pressure among fish species. Mimi-HSP90 gene was constitutivelyexpressed in ten examined tissues, and the expression level of liver was higher than in other tissues. Theexpression level of Mimi-HSP90 gene under bacterial infection and heat shock were analyzed by real-time quantitative RT-PCR, resulted in significant changes in liver, spleen, and kidney tissues. The puri-fied recombinant pET-HSP90 proteinwas used to produce the polyclonal antibody in mice. The specificityof the antibody was determined by Western blot analysis. All results suggested that Mimi-HSP90 wasinvolved in thermal stress and immune response in miiuy croaker.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Heat shock proteins (HSPs) are ubiquitous and highly conservedstress proteins, and exist widely from bacteria to mammals andplants. They play key roles in response to potential stress conditions[1,2]. According to their apparent molecular mass, HSPs can begrouped into several families: HSP110, HSP90, HSP70, HSP60 andlow molecular weight HSP [3,4]. Among all HSP family, HSP90 is ahighly conserved and the most abundant cellular protein,accounting for 1%e2% of cellular proteins in most tissues undernon-stress conditions [5]. HSP90 participates in some cellar pro-cess, such as cell signaling, signal transduction, protein folding, andprotein degradation under normal metabolism and stressful con-ditions [6e9]. As a molecular chaperone, HSP90 plays a key role infolding newly synthesized and refolding denatured proteins understress environment [10], and it is also involved in the immuneresponse. There are two cytoplasmic isoforms of HSP90, namelyHSP90 a (inducible type) and HSP90 b (constitutive type) in cell,which are the consequences of gene duplication in 500million year

Wang), tianjunxu@163.com

All rights reserved.

ago [11e14]. HSP90 b lacks the glutamine-rich sequence (QTQDQ)at the N-terminus and is larger than HSP90 a, while they exhibitdifferent molecular chaperone functions [15]. The studies of HSP90have demonstrated that expression of HSP90 is inducible and canbe regulated by all kind of environmental stresses, such as heatshock [16,17], bacterial challenge [18], and heavy metals [19].

The aquatic environment is the more complex system. Aquaticanimals usually face a variety of environmental stress, such asthermal shock, bacteria, virus, and oxygen levels [20]. Therefore,the aquatic animalsmust develop an effective and helpful system toadapt to bad environments. A mass of evidence have demonstratedthat HSP is responsible for protecting animals from harm. HSP90 isa member of the HSP family and has received some attention inresearch. To our knowledge, several HSP90s have been reportedand the expression level of HSP90 gene is shown in scallop and fish,such as bay scallop [21], carp [22], senegalese sole [23], andrainbow trout [24]. However, the report about the response ofHSP90 gene at functional level in teleost remains deficient.

Miiuy croaker, Miichthys miiuy, is an economically importantfish as it is extensively distributed from the western Japan Sea tothe East China Sea. After it flourishes for several years, diseases ofcultured miiuy croaker have occurred frequently. Meantime hightemperature also restrains fish healthy developing. These factors

Table 1Primers sequences used in this study and their application.

Name Sequence (50e30) PCR objective

Mimi-HSP90-outer-F TCTGACCAATGACTGGGAGG 30RACEMimi-HSP90-inner-F GTGCCCTGCTCTTCATTCMimi-HSP90-RT-F ATGACCAAAGCCGACCTG Real time RT PCRMimi-HSP90-RT-R GTGAAAGAACCTCCAGCAMimi-HSP90-Actin-F GTGATGAAGCCCAGAGCA Real time RT PCRMimi-HSP90-Actin-R CGACCAGAGGCATACAGGMimi-HSP90-F TAGGATCCATGCCTGAAGAAATGCACC Recombinant

expressionMimi-HSP90-R GTCGAATTCAATCAACTTCTTCCATGCGT

T. Wei et al. / Fish & Shellfish Immunology 35 (2013) 429e437430

limit the profitability and development of miiuy croaker [25], andresult in a great loss of miiuy croaker in aquaculture. In order toelucidate the molecular immune mechanisms in miiuy croaker, aseries of immune response and evolution mechanism studies of theimmune-related genes have been carried out and reported [26e34]. However, the stress response of miiuy croaker HSP90 againstbacterial infection and heat shock is largely unknown. It is impor-tant to understand the functions and expression characteristics ofHSP90 under environmental stress. In this study, for the first time,we report the isolation of the HSP90 cDNA of miiuy croaker(designated Mimi-HSP90). The gene expression analysis is used toclarify the role of HSP90 in response to Vibrio anguillarum infectionand heat shock. Meanwhile, HSP90 protein expression situationsare analyzed from miiuy croaker. To understand possible evolu-tionary process of the HSP90, an evolutionary study of HSP90among fish species is discussed.

2. Materials and methods

2.1. Animals, immune challenge and heat shock stress

Miiuy croakers (mean weight 810 g) were collected fromZhoushan Fisheries Research Institute (Zhejiang, China) and main-tained in aerated seawater at 20 �C for two weeks before processing.Ten tissues (kidney, liver, spleen, fin, brain, heart, intestine, gill,muscle and eye) from miiuy croaker were removed and keptat �80 �C until use. Challenge of miiuy croaker with V. anguillarumwas performed as described by Xu et al. [35]. Miiuy croaker wereanaesthetized by immersion in MS222 and injected intraperitoneallywith 1 ml bacterial suspension. The infected fish were killed at 6 h,12 h, 24 h, 36 h, 48 h, and 72 h after injection. For heat shock treat-ment experiment, ten miiuy croaker fish were put into aeratedaquaria at different temperature (25 �C, 28 �C, 31 �C, 34 �C, 37 �C,respectively) for 2 h and thenmoved back to natural environment for30min. At 37 �C, experimental fishwould die rapidly.Meantime, non-heat shock fish were kept in the 20 �C as a control group. Finally,tissues (liver, spleen, and kidney) of experiment under bacterialchallenge and heat shock stress and control fish were removed andkept at �80 �C until use.

2.2. RNA isolation, cDNA synthesis

Total RNA was extracted from various tissues of miiuy croakerusing RNAiso Reagent (TaKaRa) according to the manufacturer’sinstructions. Then it was resuspended in DEPC-treated water.Quality of the RNA was assessed by electrophoresis on 1.0%agarose gel. Finally, the cDNA was synthesized using QuantscriptRT kit (TIANGEN) according to the manufacturer’s instructions.Then the cDNAwas used for PCR reactions in gene expression andcloning.

2.3. Cloning the full-length of the Mimi-HSP90

One EST (GW669537), similar to other fish species HSP90 gene,was acquired from the spleen cDNA library of miiuy croaker by ESTanalysis in our laboratory [35]. To obtain the complete cDNAsequence of miiuy croaker, Two specific sense primers, HSP90-outer-F 50-TCTGACCAATGACTGGGAGG-30, HSP90-inner-F 50-GTGCCCTGCTCTTCATTC-30 (Table 1), were designed based on thesequence of EST to clone the 30 end of Mimi-HSP90 cDNA sequence.RACE-PCR was performed using a Smart RACE cDNA amplificationkit (Clontech) according to the manufacturer’s instructions. Thereaction condition of PCR were follows: predenaturalization at95 �C for 5 min; 35 cycles of 30 s at 95 �C, 30 s at 60 �C, 2 min at72 �C, and a final step of 10min at 72 �C. The obtained PCR products

were purified by a PCR purification kit (TIANGEN). The purifiedfragments were ligated into PMD-19T vectors (Takara) and clonedto TOP10 cell according to the standard protocol. Positive cloneswere screened via PCR with M13� primers. At least three cloneswere sequenced from both strands on an ABI3730XL AutomatedSequencer with M13 primer.

2.4. Sequence analysis of Mimi-HSP90

The searches for nucleotide and amino acid sequence similar-ities were conductedwith BLAST programs at the NCBI (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Multiple alignment of Mimi-HSP90 wasperformed with the Clustal WMultiple Alignment program (http://www.ebi.ac.uk./clustalw/). The sequence identities between Mimi-HSP90 and vertebrates were calculated by the software MegAlign(DNAStar, USA). The SignalP3.0 was used for predicting Mimi-HSP90 signal peptide (http://www.cbs.dtu.dk/services/Signalp/).Additional the domain structure of Mimi-HSP90 was predicted bySMART (http://smart.embl-heidelberg.de/).

2.5. Expression of Mimi-HSP90 protein and purification of therecombinant proteins

The open reading frame (ORF) of Mimi-HSP90 was amplifiedwith specific primers (Table 1) and cloned into the pMD-18T vector.The recombination plasmid DNA containing the Mimi-HSP90 ORFwas digestedwith BamH I and EcoR I enzymes, and then ligated intothe pET-28a vector (digested with the same enzymes). The formingrecombinant plasmid pET-HSP90 was transformed into Escherichiacoli BL21 cells, and correct insertion was determined by enzymedigestion and sequencing. Positive clone with correct insertion wascultured overnight in 5 mL LB medium containing kanamycin at37 �C. Overnight culture was diluted at (1:20) and incubated at37 �C with shaking at 180 rpm until the cell density (OD600) ach-ieved 0.6e0.7. In order to obtain over-expressed recombinantprotein, a different time point’s induction with 1 mM isopropyl-b-thiogalactopyranoside (IPTG) and an IPTG induction using differentconcentrations (0.05 mM, 0.1 mM, 0.3 mM, 0.5 mM and 0.7 mM) at37 �C were both performed. After the IPTG induction, proteinsamples were analyzed by 12% SDS-PAGE. The recombinant proteinwas purified according to the instruction manual supplied withHis-Bind purification kit.

2.6. Production of polyclonal antibodies against the recombinantHSP90

Five male BALB/c mice (4-week-old) were used for immunizingwith recombinant protein HSP90 mixed with equal volumeFreund’s complete adjuvant. Subsequently, Mice were boosterimmunized with the mixture of antigen and Freund’s incompleteadjuvant. Themice anti-HSP90 serumwas collected after the fourthimmunization. It was clarified by overnight incubation at 4 �C, andcentrifuged at 1600 rpm for 10 min. Finally the supernatant was

Fig. 1. Nucleotide and deduced amino acid sequences of Mimi-HSP90 from miiuy croaker. Heat shock protein 90 family signature motifs and consensus sequence MEEVD arehighlighted as shaded regions. ATP binding domain is underlined, conserved “GxxGxG” motif is italicized. The nucleotides and amino acids are numbered along the left and rightmargin.

T. Wei et al. / Fish & Shellfish Immunology 35 (2013) 429e437 431

stored at �20 �C. Simultaneously, the negative control group wasimmunized with PBS using the same procedures.

2.7. Western blot analysis

After the electrophoresis, the gel was transferred onto thePVDF membranes, and it was blocked by incubation in TBSTcontaining 5% skimmed milk for 4 h. Meanwhile, polyclonal

antibodies from mice were adsorbed with E. coil BL21 for 6 h. Themembrane was hybridized with HSP90 polyclonal antibodies(1:5000) for 2 h, followed by washes 15 min for 3 times with PBST.Finally, it was blocked in TBS for 10 min. The goat-anti-mouse IgGantibody was used as the secondary antibody dissolved in TBST.The membrane was incubated with the second antibody for 1 h atroom temperature. It was developed by NBT/BCIP and stopped bystrips with distilled water.

Fig. 2. Alignment of amino acid sequences of Mimi-HSP90. HSP90 family signature sequences and the consensus sequence MEEVD are shaded with gray background. The aminoacids are numbered along the right margin.

T. Wei et al. / Fish & Shellfish Immunology 35 (2013) 429e437432

Table 2HSP90 amino acid identity determined by the DNAStar.

Species Gene isoform Accession number Miichthys miiuy

Identity(%)

Larimichthys crocea HSP90 beta AEP68104 99.5Lates calcarifer HSP90 beta AEH27541 97.4Paralichthys olivaceus HSP90 beta ABG56394 96.9Oreochromis niloticus HSP90 beta XP_003446339 95.8Kryptolebias marmoratus HSP90 beta AEM65180 95.7Danio rerio HSP90 beta AF068772 94.1Acipenser ruthenus HSP90 beta AFA25806 92.5Salmo salar HSP90 beta AF135117 91.7Paralichthys olivaceus HSP90 alpha ABG56393 85.2Danio rerio HSP90 alpha BC075757 81.7Astyanax mexicanus HSP90 alpha AY222612 81.4

Fig. 3. Phylogenetic tree of HSP90 gene from miiuy croaker and other species wereconstructed using the Bayesian inference method. The lineage of miiuy croaker ismarked by the red circle. The species names and the GenBank accession numbers areas follows: Miichthys miiuy HSP90 JQ929760; Lutjanus sanguineus HSP90 HM59273;Scophthalmus maximus HSP90 ABG56394; Lates calcarifer HSP90 AEH27541; Para-lichthys olivaceus HSP90 ABG56394; Solea senegalensis HSP90 BAF92790; Kryptolebiasmarmoratus HSP90 AEM65180; Oreochromis niloticus HSP90 XP_003446339; Onco-rhynchus mykiss HSP90 AB196458; Salmo salar HSP90 AF135117; Gallus gallus HSP90CAA49704; Homo sapiens HSP90 NP_031381; Bos taurus HSP90 NP_001073105; Xen-opus laevis HSP90 AY785160; Dicentrarchus labrax HSP90 AY395632; Takifugu obscurusHSP90 EU853673. (For interpretation of the references to colour in this figure legend,the reader is referred to the web version of this article.)

T. Wei et al. / Fish & Shellfish Immunology 35 (2013) 429e437 433

2.8. Quantitative analysis of Mimi-HSP90 mRNA expression

The primers Mimi-HSP90-RT-F and Mimi-HSP90-RT-R (Table 1)were used for amplifyingMimi-HSP90 gene fragment (Table 1). TheRT-PCR of b-actin expression was carried out with b-actin-F andb-actin-R (Table 1). The real-time quantitative PCR assay was car-ried out on a 7300 real time PCR system (Applied Biosystems, USA)using a RealMasterMix kit (TIANGEN). The reaction carried outwithout the template was used as blank control. The PCR wascarried out in a total volume of 20 ml, including 1 ml cDNA sample,9 ml SYBR Green Real-time PCR master mixtures, 1 ml of each primerand 8 ml ddH2O. The amplification protocol was carried out at 10 s at95 �C, followed by 45 cycles consisting of 5 s at 95 �C and 34 s at60 �C. Dissociation curve analysis of amplification products wasperformed after each PCR reaction to determine target specificity.Fluorescent real-time PCR data were analyzed with 7300 SystemSDS Software v 1.3.0 (Applied Biosystems, USA) after the PCR run.The base line was set automatically by the software. The 2-DDCT

method was used for analyzing the relative expression level ofMimi-HSP90 (Livak and Schmittgen, 2001). Comparisons werepresented by one-way analysis of variance (ANOVA) with softwareSPSS 13.0 between groups, followed by a Duncan test to determinedifferences among the treatments. Differences were accepted sta-tistically significant if the p value was less than 0.05.

2.9. Evolutionary analysis

The nucleotide sequence of HSP90 was aligned with otherspecies based on the Clustal W alignment. The phylogenetic treewas reconstructed using Bayesian inference in MrBayes version 3.1[36]. For Bayesian inference, we evaluated the optimal model asGTR þ I þ G by Bayesian Information Criterion (BIC) using jMo-deltest [37]. Four chains were run simultaneously and two inde-pendent runs were used to assess the convergence of posteriorprobability distributions. Each run was conducted with 5 milliongenerations and sampled every 1000 generations, and then builtthe consensus tree after discarding the first 25% of trees as burn-in.In order to investigate the evolution pressure on the HSP90 gene,evolutionary analysis was performed with PAML 4.5 program suite[38]. The hypothesis of positive selection was tested using site-specific model in the CODEML program. The random-sites models[39] were employed in assuming some heterogeneous sites withdifferent the ratio of nonsynonymous and synonymous sub-stitutions (u) parameters without a priori knowledge of which class(positive, neutral, or purifying selection) a given codon belonged to.In order to account for whether positive selection had been oper-ating on any codon sites, we estimated parameters under sixdifferent codon substitution models (M0, M1a, M2a, M3, M7, M8,model) [39]. The likelihood ratio tests (LRT) were performed tocompare the corresponding models with and without selection(M0 vs M3, M1a vs M2a, M7 vs M8). When the alternative modelsM2a and M8 suggested the presence of sites with u > 1, this wouldbe thought as evidence of positive selection [39]. Posterior proba-bilities were calculated by Empirical Bayes in the models M2a andM8 [40]. In all cases, twice the difference of log-likelihood values(2OLnL) was calculated following a c2 distribution with degreesof freedom equal to the difference in parameter numbers estimatedin the compared models [39].

3. Results

3.1. Characteristics of Mimi-HSP90 gene

In this study, the cDNA sequence of Mimi-HSP90 was depositedin GenBank under accession no. JQ929760. The full length of miiuy

croaker HSP90 cDNA was 2736 bp, consisting of a 50termainal un-translated region of 91 bp, a 30 terminal un-translated region of467 bp with a canonical polyadenylation signal sequence ATTAAA.The open reading frame (ORF) was 2178 bp that could encode a 725amino acids protein with a predicted molecular mass of 83.3 kDaand theoretical isoelectric point of 4.83 (Fig. 1). Five conservedHSP90 family signatures were identified in the deduced amino acidsequence of Mimi-HSP90: signature 1 NKEIFLRELISNASDALDKIR,signature 2 LGTIAKSGT, signature 3 IGQFGVGFYSAYLVAE, signature4 IKLYVRRVFI, and signature 5 GVVDSEDLPLNISRE (Figs. 1 and 2).The glutamine-rich sequence (QTQDQ) was not found at the N-terminus of Mimi-HSP90. However, the conserved motif “GxxGxG”and C-terminal-conservedMEEVDmotif were determined inMimi-HSP90. Meanwhile, SMART program analysis revealed the typicalhistidine kinase-like ATPase domain, which was located fromposition 34e188. It was ubiquitous in all HSP90 family members(Fig. 1).

3.2. Bioinformatics analysis of Mimi-HSP90

The deduced amino acid sequence of Mimi-HSP90 was comparedwith the sequence of previous other species HSP90. It displayed high

Table 3Site-specific model analyses for the HSP90 gene of fish species.

Model np Log-likelihood Model comparison Parameter estimates Positivelyselected sites

M0 (one ration) 31 �11062.766 K ¼ 1.675, u ¼ 0.023 NoneM1a (nearly Neutral) 32 �10942.335 K ¼ 1.744, p0 ¼ 0.971, p1 ¼ 0.029,

u0 ¼ 0.016, u1 ¼ 1.000NA

M2a (Positive Selection) 34 �10942.335 M2 vs M1, 2DlnL ¼ 0.0, df ¼ 2, p ¼ 1.000 K ¼ 1.744, p0 ¼ 0.971, p1 ¼ 0.029,p2 ¼ 0.000, u0 ¼ 0.016, u1 ¼ 1.000,u2 ¼ 89.763

Not allowed

M3 (discrete) 35 �10860.207 M3 vs M0, 2DlnL ¼ 405.1, df ¼ 4, p ¼ 0.000 K ¼ 1.664, p0 ¼ 0.767, p1 ¼ 0.202,p2 ¼ 0.031, u0 ¼ 0.002, u1 ¼ 0.067, u2 ¼ 0.429

Not allowed

M7 (beta) 32 �10872.706 K ¼ 1.700, p ¼ 0.129, q ¼ 3.389 NAM8 (beta and omega) 34 �10864.381 M8 vs M7, 2DlnL ¼ 16.65, df ¼ 2, p ¼ 2.43E-4 K ¼ 1.700, p0 ¼ 1.000, p1 ¼ 0.000,

p ¼ 0.129, q ¼ 3.389, u ¼ 3.872238 K (0.568)

Fig. 4. Expression of HSP90 gene in various tissues (brain (B), eye (E), gill (G), heart (H),intestine (I), liver (L), fin (F), muscle (M), spleen (S), and kidney (K)) of miiuy croaker(uninfected). HSP90 gene mRNA levels were expressed as a ratio relative to beta-actinlevels in the same samples after real-time PCR. Deviation bars represent the standarderrors of three experiments at each time point. Values with the same superscript arenot significantly different (P > 0.05).

T. Wei et al. / Fish & Shellfish Immunology 35 (2013) 429e437434

similarity to HSP of Larimichthys crocea (99.5% to b isoform), Latescalcarifer (97.4% to b isoform), Paralichthys olivaceus (96.9% to b iso-form, 85.2% to a isoform), Danio rerio (94.1% to b isoform, 81.7% to aisoform), and Astyanax mexicanus (81.4 to a isoform) (Table 2). Thephylogenic analysis using Bayesian inference method placed Mimi-HSP90 inside the main cluster of teleostean sequence. Then theyformed a larger branch with frog and bird (Fig. 3). The relationshipsdisplayed in the phylogenic tree were generally in agreement withthe concept of traditional taxonomy. As for its unique characteristics,some tests for selection were used for determining the evolutionarypattern of Mimi-HSP90. Under the one-ratio model, uwas estimatedto be 0.023 for all branches, which was significantly smaller than 1(Table 3). This result suggested that an overall strong purifying se-lection existed in HSP90 gene. Model M3 seemed to be better fit tothe data than the M0 model, and it revealed variable selectionpressure across the HSP90 sequences. Model M2a was not signifi-cantly better than the null hypothesis model M1a, but the alternativehypothesis M8 was better than the M7 (Table 3). The site-specificanalysis in PAML identified a positively selected site (P ¼ 56.8%) inthe HSP90 group under M8 model. Posterior probabilities for siteclasses had been calculated by empirical Bayes. SignalP softwareanalysis found no signal peptide in Mimi-HSP90 sequence. The sec-ondary structure indicated that a helix accounted for 50.21%; beta-sheet, 15.31%; beta-turn, 5.66%; and random coil, 28.83%. Mean-while, the DNAStar softwarewas used to predict the hydrophilicity ofMimi-HSP90. The result revealed that the amount of hydrophilicresidueswasmore than hydrophobic residues. In addition, theMimi-HSP90 might locate in cytoplasm (26.1%), nucleus (8.7%), endo-plasmic reticulum (8.7%), vesicles of secretory system (21.7%), mito-chondria (13%), plasma membrane (8.7%), cytoskeleton (4.3%),vacuole (4.3%), and golgiosome (4.3%).

3.3. Quantitative analysis of Mimi-HSP90 gene expression

The mRNA transcripts of Mimi-HSP90 could be determined inten normal tissues by quantitative real-time RT-PCR, including liver,gill, spleen, brain, fin, eye, intestine, kidney, heart, and muscle. Butthe expression level of them was distinctly different (Fig. 4). Thehighest expression of Mimi-HSP90 was observed in liver tissue(p < 0.05), moderately in fin, brain, spleen, gill, muscle (p < 0.05),and expressed weakly in intestine, kidney, heart, eye.

In heat shock stress experiment, the expression level of Mimi-HSP90 was shown at different temperature points (Fig. 5). In liver,the expression of Mimi-HSP90 gene was lowest level at 28 �C(p < 0.05). Moreover, the expression level at 31 �C, 34 �C and 37 �Cwas less than control experiment (p < 0.05). In spleen, the expres-sion level of Mimi-HSP90 gene was the lowest at 28 �C, but theexpressionpeakwas at 31 �C (p< 0.05).While the expression level ofMimi-HSP90 gene was maximum level in kidney at 37 �C (p < 0.05).

The temporal expression levels of Mimi-HSP90 in liver, spleen,kidney with the pathogenic V. anguillarum were shown (Fig. 6). Inliver, the expression level decreased from infection starting time to6 h, and then fluctuated from 6 h to 72 h (p< 0.05). While in spleentissue, the expression level decreased at firstly, then increased topeak at 24 h (p < 0.05), but decreased again the low level in 36 h,then increased to peak level. In kidney, the expression level ofMimi-HSP90 decreased dramatically after challenge, and thenfluctuated from 6 h to 72 h (p < 0.05).

3.4. Protein expression, purification and western blot

The complete ORF of Mimi-HSP90 was amplified and cloned intothe pET-28a expression vector. It was used to acquire expressionlevel of recombinant protein in E.coil BL21. Protein expression levelwas induced by adding different incubation time and concentrationof IPTG. The idea expression condition was 0.7 mmol/L IPTG incu-bation for 6 h (Fig. 7). After the protein was purified by Ni-NTAaffinity chromatography, a signal specific band of 95 kDa wasshowed by SDS-PAGE. The polyclonal antibody produced from theimmunized mice was used for detecting the recombinant HSP90protein. The results indicated a specific band emerged in the puri-fied recombinant HSP90 protein, and there was not specific band inthe induced BL-21 with pET-28a (Fig. 8). At the same time, thespecific band was not detected in the negative control, suggestingthat the polyclonal antibody only bound specifically with HSP90protein.

Fig. 5. Expression analysis of Mimi-HSP90 by relative quantitative real-time PCR in liver, spleen, and kidney under heat shock stress. Expression of b-actin was used as an internalcontrol for real-time PCR. Deviation bars represent the standard errors of three experiments at each temperature point. Significant differences between heat shock group andcontrol group were indicated by an asterisk (P < 0.05).

T. Wei et al. / Fish & Shellfish Immunology 35 (2013) 429e437 435

4. Discussion

In recent years, aquaculture industry has been greatly affectedby disease epidemics and temperature change. Themortality rate ofmarine animal has been associated with bacteria infection andtemperature stress. In order to survival, marine animal mustdevelop protective mechanisms in response to environment stress.As one member of the stress-related protein, HSP90 is involved inresistance to environment stress and plays key role in defendingcells against bacterial and viral infection. In this paper, the completecDNA sequence of HSP90 gene was cloned and characterized fromspleen of miiuy croaker. Five conserved HSP90 family signaturesand ATP binding domainwere found in theMimi-HSP90 amino acidsequence. They were essential role for functions of HSP90. Mean-while, a conserved “GxxGxG” motif was critical for binding ATP[41]. MEEVD motif was also found in the C-terminal of Mimi-HSP90, which was similar to other species HSP90 protein. It wasimportant for HSP90 to bind to tetratricopeptide repeat domain-containing proteins [42]. To date, there is only one type of HSP90

Fig. 6. Expression analysis of Mimi-HSP by relative quantitative real-time PCR in liver spleeused as an internal control for real-time PCR. Deviation bars represent the standard errors ofand control group were indicated by an asterisk (P < 0.05).

gene in invertebrates, except that Anopheles albimanus containstwo type HSP90 genes. But there exist two cytosolic isoforms ofHSP90 gene (HSP90 a and HSP90 b) in vertebrates, which aredifferent in the structure of glytamine-rich sequence (QTQDQ) atthe N-terminus [15]. In general, the QTQDQ do not exist the HSP90b isoform amino sequence. In this study, there was no QTQDQsequence at Mimi-HSP90 N-terminus. The observation suggestedthatMimi-HSP90was concernedwith vertebrate HSP90 b isoforms.Homology analysis demonstrated that Mimi-HSP90 shared 91.7%e99.5% with other known species HSP90s b isoform. Especially, thesimilarity of amino sequence came up to 99.5% between miiuycroaker and large yellow croaker HSP90 b isoform. The observationssuggested that Mimi-HSP90 was concluded to be a cytosolicmember of HSP90 b isoform family. In a word, the results ofstructure analysis and homology analysis supported that Mimi-HSP90 should belong to HSP90 b isoforms family. The randomsites models proved extreme variability in selective pressureamong sites. ModelM3 appeared to be better fit to the data than theM0 model, indicating that variable selection pressure was across

n, and kidney, during 72 h of induction with V. anguillarum. Expression of b-actin wasthree experiments at each time point. Significant differences between challenged group

Fig. 7. SDS-PAGE analysis result of expression product. (A) Lane 1: pET-HSP90 without induction; M: low protein molecular marker; 2:IPTG induction pET-28a; 3e9 IPTG inductionpET-HSP90 for 1 h, 2 h, 3 h, 4 h, 5 h, 6 h, 7 h. (B) IPTG induction pET-28a; M: low protein molecular marker; 2e6: IPTG concentration 0.7, 0.5, 0.3, 0.1, 0.05 mmol/L.

T. Wei et al. / Fish & Shellfish Immunology 35 (2013) 429e437436

the fish HSP90 sequences. The model M2a was not significantlybetter than the null hypothesis model M1a (p ¼ 1.00). The resultssuggested no reliable positive selection sites were found in positiveselection model. But the M8 was better than the null hypothesismodel M7, and it detected one positively selected amino acid(238 K) in the miiuy croaker with posterior probability of 56.8%.This site was likely to be under positive selection pressure. M8model was more fitted for smaller data sets by using Bays EmpiricalBayed method [40], and it should be consider the reliable result.Under the one-ratio model, u was estimated to be 0.023 for allbranches. The result illustrated that the HSP90was under an overallstrong purifying select pressure in these species.

Real-time PCR analysis indicated that Mimi-HSP90 gene wasconstitutively expressed at different level in ten examined tissues ofmiiuy croaker, which in general corresponded to situation thatconstitutive HSP90 gene expressed. The specific expression of theMimi-HSP90 in different tissues might be owing to tissue proteinsynthesis difference. Previous studies had reported that HSP90 wasinducible by a range of stressors [43,44]. To demonstrate this point,Challenge of miiuy croaker with the V. anguillarum infection wasconducted, and resulted in significant changes in the expression of

Fig. 8. Purification of HSP90 recombinant protein and specificity of HSP90 polyclonal antibodmolecular marker; 2: unpurified recombinant protein; 3: Purified recombinant protein. (B) M2: IPTG induced E. coil BL with pET-28a for 6 h; 3: Negative control.

HSP90 gene in liver, spleen, and kidney tissues. The result indicatedthat Mimi-HSP90 was inducible expressed and played a key role indefending bacterial infection. This was also the first time that suchthe response of HSP90 had been observed in the miiuy croaker,suggesting that Mimi-HSP90 was involved in immune responses tobacterial challenge. These data could be valuable to understand thesignificance of HSP90 to miiuy croaker immune defense. Further-more, we performed another experiment to determine the expres-sion levels of Mimi-HSP90 gene after heat shock stress. The resultshowed that there existed significant expression difference in theimmune tissues. It indicated that Mimi-HSP90 was inducible andconstitutive expressed in responded to heat shock stress, and couldplay key role in resisting in temperature elevation. Based on ourresults, we could prove that Mimi-HSP90 was involved in environ-mental stresses. To further study the function of Mimi-HSP90, theMimi-HSP90 was expressed in E. coli and purified by Ni-NTA affinitychromatography. SDS-PAGE analysis displayed that the size of HSP90was approximately 95 kDa in miiuy croaker. The size of the 95 kDawas larger than the deduced 83.3 kDa protein for molecular weight.The phenomenon also emerged in ab-Rbx1 protein [45], bZIP protein[46] and Ca-CREB protein [47]. It was reported that aggregation,

y was determined byWestern blot. (A) Lane 1: IPTG induction pET-28a. M: low protein: Low molecular marker; Lane 1: IPTG induced E. coil BL with pET-HSP90 for 6 h; Lane

T. Wei et al. / Fish & Shellfish Immunology 35 (2013) 429e437 437

asymmetrics structure or unsuitable electrophoretic conditionsmight have lead to the phenomenon [48]. In addition, Western blotanalysis indicated that the polyclonal antibody produced in micereacted to the purified HSP90, and not reacted pET-28a. This resultdemonstrated specificity of this antibody for HSP90 inmiiuy croaker.

In summary, we were first to over-express and purify HSP90protein in the E. coli expressed system. The molecular evolutionaryanalysis showed that HSP90 was under an overall strong purifyingselect pressure among fish species. Studies of the expressionpattern of Mimi-HSP90 under bacterial challenge and heat shockstress was also described for the first time in miiuy croaker. Furtherstudies are needed to study the immune mechanism in miiuycroaker.

Acknowledgments

We thank two anonymous reviewers for valuable comments onan earlier version of the manuscript. This study was supported byNation Nature Science Foundation of China (31001120), ZhejiangProvincial Natural Science Foundation of China (Y3100013) andImportant Science and Technology Specific Projects of ZhejiangProvince (2011C14012).

References

[1] Craig EA, Schlesinger MJ. The heat shock response. Crit Rev Biochem Mol Boil1985;18:239e80.

[2] Lindquist S, Craig EA. The heat-shock proteins. Annu Rev Genet 1988;22:631e77.

[3] Georgopoulos C, Welch WJ. Role of the major heat shock proteins as molecularchaperones. Annu Rev Cell Dev Biol 1993;9:601e34.

[4] Basua N, Todgham AE, Ackerman PA, Bibeau MR, Nakano K, Schulte PM, et al.Heat shock protein genes and their functional significance in fish. Gene2002;295:173e83.

[5] Buchner J. Hsp90 and Co.-a holding for folding. Trends Biochem Sci 1999;24:136e41.

[6] Wiech H, Buchner J, Zimmermann R, Jakob U. Hsp90 chaperones proteinfolding in vitro. Nature 1992;358:169e70.

[7] Bohen SP, Kralli A, Yamamoto KR. Hold’em and fold’em: chaperones andsignal transduction. Science 1995;268:1303e4.

[8] Freeman BC, Morimoto RI. The human cytosolic molecular chaperones hsp90,hsp70 (hsc70) and hdj-1 have distinct roles in recognition of a non-nativeprotein and protein refolding. EMBO J 1996;15:2969e79.

[9] Nathan DF, Vos MH, Lindquist S. In vivo functions of the Saccharomyces cer-evisiae Hsp90 chaperone. Proc Natl Acad Sci USA 1997;94:12949e56.

[10] Fu D, Chen JH, Zhang Y, Yu ZN. Cloning and expression of a heat shock protein(HSP) 90 gene in the haemocytes of Crassostrea hongkongensis under osmoticstress and bacterial challenge. Fish Shellfish Immunol 2011;31:118e20.

[11] Lai BT, Chin NW, Stanek AE, Keh W, Lanks KW. Quantitation and intracellularlocalization of the 85K heat shock protein by using monoclonal and polyclonalantibodies. Mol Cell Biol 1984;4:2802e10.

[12] Gupta RS. Phylogenetic analysis of the 90 kD heat shock family of proteinsequences and an examination of the relationship among animals, plants, andfungi species. Mol Biol Evol 1995;12:1063e73.

[13] Csermely P, Schnaider T, Soti C, Prohaszka Z, Nardai G. The90-kDa molecularchaperone family: structure, function, and clinical applications.A Comprehensive Review Pharmacol Ther 1998;79:129e68.

[14] Sreedhar AS, Kalmar E, Csermely P, Shen YF. Hsp90 isoforms: functions,expression and clinical importance. FEBS Lett 2004;562:11e5.

[15] Lees-Miller SP, Anderson CW. The human double-stranded DNA-activatedprotein kinase phosphorylates the 90-kDa heat-shock protein, hsp90 alpha attwo NH2-terminal threonine residues. J Biol Chem 1989;264:17275e80.

[16] Farcy E, Serpentini A, Fiévet B, Lebel JM. Identification of cDNAs encodingHSP70 and HSP90 in the abalone Haliotis tuberculata: transcriptional induc-tion in response to thermal in hemocyte primary culture. Comp BiochemPhysiol PartB: Biochem Mol Biol 2007;146:540e50.

[17] Jiang S, Qiu L, Zhou F, Huang J, Guo Y, Yang K. Molecular cloning andexpression analysis of a heat shock protein (Hsp90) gene from black tigershrimp (Penaeus monodon). Mol Biol Rep 2009;36:127e34.

[18] Young RA. Stress proteins and immunology. Annu Rev Immunol 1990;8:401e20.[19] Ivanina AV, Taylor C, Sokolova IM. Effects of elevated temperature and cad-

mium exposure on stress protein response in eastern oysters Crassostreavirginica (Gmelin). Aquat Toxicol 2009;91:245e54.

[20] Li FJ, Luan W, Zhang CS, Zhang JQ, Wang B, Xie YS, et al. Cloning of cytoplasmicheat shock protein 90 (FcHSP90) from Fenneropenaeus chinensis and its

expression response to heat shock and hypoxia. Cell Stress and Chaperones2009;14:161e72.

[21] Gao Q, Zhao J, Song L, Qiu L, Yu Y, Zhang H, et al. Molecular cloning, char-acterization and expression of heat shock protein 90 gene in the haemo-cytes of bay scallop Argopecten irradians. Fish Shellfish Immunol 2008;24:379e85.

[22] Hermesz E, Ábrahám M, Nemcsók J. Identification of two hsp90 genes in carp.Comp Bionchem Physiol Part C: Toxicol Pharmacol 2001;129:397e407.

[23] Manchado M, Salas-Leiton E, Infante C, Ponce M, Asensio E, Crespo A, et al.Molecular characterization, gene expression and transcriptional regulation ofcytosolic HSP90 genes in the flatfish senegalese sole (Solea senegalensis Kaup).Gene 2008;416:77e84.

[24] Caraa JB, Aluru N, Moyanoa FJ, Vijayanb MM. Food-deprivation induces HSP70and HSP90 protein expression in larval gilthead sea bream and rainbowtrout.Comp Bilchem Physiol Part B: Biochem Mol Biol 2005;142:426e31.

[25] Xu TJ, Cheng YZ, Shi G, Wang RX. Molecular cloning, characterization, andexpression analysis of a disease-resistance related CC chemokine gene inmiiuy croaker (Miichthys miiuy). Aquaculture 2011;318:25e32.

[26] Xu TJ, Sun YN, Cheng YZ, Shi G, Wang RX. Genomic sequences comparison anddifferential expression of miiuy croaker MHC class I gene, after infection byVibrio anguillarum. Dev Comp Immunol 2011;35:483e9.

[27] Xu TJ, Sun YN, Shi G, Cheng YZ, Wang RX. Characterization of the majorhistocompatibility complex class II genes in Miiuy croaker. PLoS One 2011;6:e23832.

[28] Xu TJ, Sun YN, Shi G, Wang RX. Miiuy croaker hepcidin gene and comparativeanalyses reveal evidence for positive selection. PLoS ONE 2012;7:e35449.

[29] Sun YN, Xu TJ, Wang JX, Cheng YZ, Wang RX. Sequence and expressionanalysis of cathepsin S gene in the miiuy croaker Miichthys miiuy. Fish PhysiolBiochem 2011;37:761e5.

[30] Cheng YZ, Wang RX, Xu TJ. Molecular cloning, characterization and expressionanalysis of a miiuy croaker (Miichthys miiuy) CXC chemokine gene resemblingthe CXCL9/CXCL10/CXCL11. Fish Shellfish Immunol 2011;31:439e45.

[31] Cheng YZ, Wang RX, Sun YN, Xu TJ. Molecular characterization of miiuycroaker CC chemokine gene and its expression following Vibrio anguillaruminjection. Fish Shellfish Immunol 2011;31:148e54.

[32] Liu XZ, Shi G, Cui DL, Wang RX, Xu TJ. Molecular cloning and comprehensivecharacterization of cathepsin D in the Miiuy croaker Miichthys miiuy. FishShellfish Immunol 2012;32:464e8.

[33] Meng FX, Wang RX, Xu TJ, Sun YN, Cheng YZ, Shi G. An unexpected loss ofdomains in the conservative evolution ninth complement component in ahigher teleost, Miichthys miiuy. Fish Shellfish Immunol 2012;32:1171e8.

[34] Sun YY, Zhu ZH, Wang RX, Sun YN, Xu TJ. Miiuy croaker transferrin gene andevidence for positive selection events reveals different evolutionary patterns.PLoS One 2012;7:e43936.

[35] Xu TJ, Meng FX, Sun YN, Shi G, Wang RX. Identification of immune genes ofthe miiuy croaker (Miichthys miiuy) by sequencing and bioinformatic analysisof ESTs. Fish Shellfish Immunol 2010;29:1099e105.

[36] Huelsenbeck JP, Ronquist F. MRBAYES: bayesian inference of phylogeny.Bioinformatics 2011;17:754e5.

[37] Posada D. jModelTest: phylogenetic model averaging. Mol Biol Evol 2008;25:1253e6.

[38] Yang Z. PAML4: phylogenentic analysis by maximum likelihood. Mol Biol Evol2007;24:1586e91.

[39] Yang Z, Nielsen R, Goldman N, Pedersen AMK. Codon substitution models forheterogeneous selection pressure at amino acid sites. Genetics 2000;155:431e49.

[40] Yang Z, Wong WSW, Nielsen R. Bayes empirical bayes inference of amino acidsites under positive selection. Mol Biol Evol 2005;22:1107e18.

[41] Prodromou C, Roe SM, O’Brien R, Ladbury JE, Piper PW, Pearl LH. Identificationand structural characterization of the ATP/ADP-binding site in the Hsp90molecular chaperone. Cell 1997;90:65e75.

[42] Scheufler C, Brinker A, Bourenkov G, Pegoraro S, Moroder L, Bartunik H, et al.Structure of TPR domain-peptide complexes: critical elements in the assemblyof the Hsp70-Hsp90 multichaperone machine. Cell 2000;101:199e210.

[43] Palmisano AN, Winton JR, Dickhoff WW. Sequence features and phylogeneticanalysis of the stress protein hsp90 alpha in chinook salmon (Oncorhynchustshawytscha), a poikilothermic vertebrate. Biochem Biophys Res Commun1999;258:784e91.

[44] Wu CX, Zhao FY, Zhang Y, Zhu YJ, Ma MS, Mao HL, et al. Overexpression ofHsp90 from grass carp (Ctenopharyngodon idella) increases thermal protectionagainst heat stress. Fish Shellfish Immunol 2012;33:42e7.

[45] Wu LJ, Wu XZ, Wang L. Identification and functional characterization of anRbx1 in an invertebrate Haliotis diversicolor supertexta. Dev Comp Immunol2011;35:72e80.

[46] Niu X, Guiltinan MJ. DNA binding specificity of the wheat bZIP protein EmBP-1. Nucleic Acids Res 1994;22:4969e78.

[47] Zhu BJ, Wu XZ. Characterization and function of CREB homologue fromCrassostrea ariakensis stimulated by rickettsialike organism. Dev CompImmunol 2008;32:1572e81.

[48] Sharma N, Lopez DI, Nyborg JK. DNA binding and phosphorylation induceconformational alterations in the kinase inducible domain of CREB: implica-tions for the mechanism of CREB transcription function. J Biol Chem2007;282:19872e83.