Identification of White Spot Syndrome Virus Latency-Related Genes in Specific-Pathogen-Free Shrimps by Use of a Microarray

Abstract

To investigate whether specific-pathogen-free (SPF) shrimps are asymptomatic carriers of white spot syndrome virus (WSSV), we used a WSSV-specific DNA microarray to measure WSSV gene expression in SPF and WSSV-infected shrimps. Three WSSV genes were found to be relatively highly expressed in SPF shrimps. Reverse transcription-PCR using nested primers as well as real-time detection confirmed that these genes have no detectable counterparts in GenBank; structural analysis of the putative proteins revealed helix-loop-helix and leucine zipper motifs. Viral sequences could be PCR amplified from genomic DNA of SPF shrimp, further supporting the suggestion that these shrimps are asymptomatic carriers.

White spot syndrome virus (WSSV) is a major pathogen in shrimp that causes high mortality and huge economic losses in shrimp aquaculture. The WSS virion is a nonoccluded ellipsoid- or bacilliform-shaped enveloped particle about 275 nm in length and 120 nm in width. Its circular double-stranded DNA consists of 300 kbp covering approximately 185 open reading frames (ORFs) (13). Though the genome of WSSV has recently been sequenced (13, 16), little is known about the molecular mechanisms underlying the WSSV life cycle and its modes of infectivity. Uninfected shrimp populations are an important control for the study of WSSV infection, and they are important for use in the shrimp industry in general. However, given the highly infectious nature of WSSV, such populations have historically been difficult to establish. Recently, specific-pathogen-free (SPF) shrimps thought to lack WSSV have been commercialized (BIOTEC, Bangkok, Thailand). These shrimps have been grown for 6 generations in a controlled environment without any disease outbreak. Routine diagnosis performed at BIOTEC by using an IQ2000 WSSV detection kit (Farming IntelliGene Technology Corporation, Taipei, Taiwan) (8) also verified that these shrimps are indeed WSSV negative. However, others have observed symptoms of WSSV infection in normal shrimps that are thought to result from environmental stress rather than viral contamination, raising the possibility that these shrimps contain the virus in a dormant state (1, 9, 11). In this study, we sought to determine whether the commercial SPF shrimps, which represent a valuable control condition, harbor WSSV at molecularly detectable levels and whether we could identify candidate viral genes that facilitate latency. To this end, we employed DNA microarray technology to screen for viral genes expressed in WSSV-infected and SPF shrimps and to compare the viral transcriptional profiles of these shrimps in order to identify specific transcripts that may be involved in WSSV latency and pathogenesis.

We developed WSSV DNA arrays by using highly purified WSSV particles obtained from WSSV-infected shrimps, Penaeus monodon (giant tiger shrimp), purchased from Malaysia. Shrimps were homogenized and subjected to several rounds of differential centrifugation to remove cellular debris before application for 30 to 60% discontinuous sucrose gradient centrifugation (10). Viral particles were found primarily at the interface between 40 and 50% sucrose and were confirmed by WSSV-specific monoclonal antibody immunogold electron microscopy (7) and PCR using WSSV-specific primers (data not shown). The purified viral genomic DNA obtained was restricted with AluI (NEB), generating about 744 fragments, and ligated to dephosphorylated pBluescript KS-II vector (Stratagene). Approximately 3,000 DNA fragments ranging in size from 200 bp to 2 kb, expected to cover the entire WSSV genome, were selected from the WSSV plasmid library for PCR amplification and were spotted on poly-l-lysine slides.

Different groups of total RNA samples, approximately 15 g each, were extracted from the pool of WSSV-infected shrimps used in the extraction of viral genome by using Trizol reagent (Invitrogen) according to the manufacturer's protocol. Different groups of total RNA samples were also extracted from four pools of commercialized SPF shrimps, with each group containing RNA from five shrimps. The four pools of shrimps originated from different ponds and had been certified to be free of WSSV by use of an IQ2000 WSSV detection kit prior to delivery in liquid nitrogen. After treatment with DNase I, the RNA samples were stored in aliquots at −80°C. We further verified the status of SPF (all four pools) and infected shrimps as being WSSV negative and positive, respectively, by conventional one-step PCR using structural genes VP15, VP26, and VP24 (13). The extraction of nucleic acids from infected and SPF shrimps was carried out individually with utmost care to eliminate any cross contamination.

In our microarray study, RNAs from infected and uninfected shrimps was firstly amplified by generating cDNA templates by using oligo(dT) 17, followed by in vitro transcription using a Megascript T7 polymerase kit (Ambion) as previously described (2, 12). Probes generated from 4 μg of the amplified RNA (aRNA) by using reverse transcription with a random primer were then labeled with Cy3 (Perkin-Elmer). Replicate single-channel hybridizations to the microarrays were performed to detect expression of WSSV genes in the infected and SPF shrimps, notwithstanding the various copy numbers of virus. Labeling of both aRNAs with a second fluorescent dye, Cy5, followed by microarray hybridization, was performed to assess possible dye bias effects and data reproducibility. Hybridized slides were then scanned on a GenePix 4000B array scanner and analyzed with GenePix Pro array analysis software (Axon Instruments, Foster City, Calif.). Signal intensities of transcripts binding microarray probes were background subtracted and used as approximate measures of absolute expression. Two array replicates with each fluorescent dye were performed. The RNA samples used in each replicate were independent of the other, and genes manifesting signal intensities that were four times that of background were considered to be expressed.

In general, the majority of the transcripts showed high signal intensities when hybridized with infected sample, indicating considerable expression during infection, while expression was seen to a much lesser extent in the SPF sample. Nevertheless, hybridization with the SPF sample revealed exceptionally high signal intensities from some elements on the array, indicating that these shrimps had been carriers of the virus and were actively expressing viral genes. Thirty clones that showed significantly high expression from each hybridization were selected. Of these, clones showing consistently high signal intensity across the four independent hybridizations were sequenced. A BLAST search revealed a few WSSV ORFs; however, only three were represented at least twice on the array. They were identified as WSSV ORFs 151, 427, and 366 (GenBank accession no. NC_003225) (16). These proteins showed no obvious counterpart in GenBank. Interestingly, these genes are highly expressed in SPF shrimps relative to other transcripts such as the VP15, VP26, and VP24 genes, while the reverse was observed in WSSV-infected shrimps, as represented in Fig. 1. The designated values for the clones in Fig. 1 are arbitrary and do not reflect the true viral copy numbers and expression levels of transcripts in both infected and SPF shrimps but merely depict a difference in distribution patterns of WSSV transcripts.

FIG. 1.

Schematic representation of the expression profiles and arrangement of WSSV ORFs 151, 366, and 427 in relation to structural genes and other clones in the WSSV genome in WSSV-infected shrimps (a) and SPF or noninfected shrimps (b).

Reverse transcription-PCR (RT-PCR) and SYBR Green real-time RT-PCR were then carried out to confirm the presence of these genes in SPF shrimps (Fig. 2 and 3). We used two-step PCR to amplify the three genes by using cDNA reverse transcribed from SPF shrimps. Primers amplifying full-length WSSV ORFs 151 (4.3 kb), 427 (1.87 kb), and 366 (252 bp) were used in the first step of the PCR in the presence of a higher magnesium chloride concentration (Table 1). Nested PCR was then performed using inner primers generating 510-, 210-, and 900-bp products of WSSV ORFs 151, 366, and 427, respectively (Fig. 2a). The amplicons were then gel purified (Qiagen) and cloned into pGEM T-vector prior to sequencing, which confirmed that the amplicons corresponded to the three genes. Shrimp β-actin gene (339 bp) was included as a positive control. We also detected the presence of these transcripts in the three other groups of SPF shrimps by using similar RT-PCR conditions (data not shown). Extending our results to the DNA of all four groups of SPF shrimps, we found that SPF shrimps contained genes that transcribed these mRNAs (Fig. 2b). RNA and DNA from WSSV-infected shrimps had also been included as positive controls (Fig. 2b). All amplified products were subsequently sequenced to verify the findings.

FIG. 2.

Detection of WSSV ORF 151, 366, and 427 transcripts in SPF shrimps by nested PCR. (a and b) The up-regulated genes in the array of SPF shrimps were validated by nested PCR using total cDNA (a) and DNA (b) from SPF shrimps. WSSV-infected RNA and DNA were also included as a positive control. Lanes 1, WSSV ORF 151 (510 bp); lanes 2, WSSV ORF 427 (900 bp); lanes 3, WSSV ORF 366 (210 bp); lanes 4, β-actin transcript (339 bp). (c) Two other genes, VP24 (266 bp) (A) and WSSV ORF 249 (726 bp) (B) were detected via nested PCR in WSSV-infected DNA (lane 1), WSSV-infected RNA (lane 2), and SPF shrimp DNA (lane 3), but not in SPF shrimp RNA (lane 4). Results are representative of all four groups of commercial SPF shrimps.

FIG. 3.

Amplification profiles (a) and dissociation curves (b) of WSSV ORF 151 using total RNA from WSSV-infected shrimps and SPF shrimps and aRNA from SPF shrimps. The crossing points are 6.89, >11, and 3.99, respectively. The negative control is a SYBR Green RT-PCR mix without template. The thermal denaturation values of amplicons are indicated alongside the corresponding dissociation curves.

TABLE 1.

Primer sequences used in the two-step conventional PCR and SYBR Green real-time RT-PCR

Primer	WSSV ORF	Forward sequence	Reverse sequence
PCR outer primers	151	ATGGATTTTGAAGGAACTACCA	CTTCTTTGTTTTCTTTG
	366	ATGAGGAAAATGACCTCTATGA	TCAAGAAAGCGCGTGCTTTAG
	427	ATGGCATGGACCGTAATGGC	TTCCTTGATCTAGAGCT
PCR inner primers	151	GTGGTCACATCTGACAGTGGA	GCATAATGCAGTAGCGTCAACGGC
	366	GAGACGTCGCTCATCAAAGATGGGGAAG	GAAACCTGGACCATATTGAATACGGCCAG
	427	GAGCTGGCAAAGGAAACC	ACAGACAACAGAACCTCCTTC
Real-time RT-PCR primers	151	GGTCGAATCTGACAGGCGTCTTT	CTCCACAACTGCGTTACCTTCGTAT
	366	TGAGAGTGATGATGACGACGATGA	TGCTTTAGCCGGACGGTTAAGAA
	427	ATAATAATAGTGCCCACCAAATGTCGG	GTTGTGACTGGAGTGTTGATTGGGTA

Parallel cycling using other genes such as VP24 and WSSV ORF 249 proved that SPF shrimps harbor viral DNA at a molecularly detectable level. However, the transcripts of these genes were beyond detection level (Fig. 2c). The results presented here are representative of all four groups of SPF shrimps. In SYBR Green real-time RT-PCR, SPF shrimp aRNA, which had been prepared as previously described (2, 12), was used as the template for amplification. All three genes were detected by using this approach, indicating the presence of virus in SPF shrimps. Based on the obtained dissociation curves of WSSV ORF 151, more amplified product was observed in WSSV-infected samples than in aRNA of SPF shrimps despite the similar amounts of template used. Little or no amplification was seen when total RNA of SPF shrimps was used. Thus, the use of aRNA but not total RNA allowed the detection of low levels of viral sequence in SPF shrimps (Fig. 3). A negative control without template was included in the experiment.

We next analyzed protein structures in each of the WSSV sequences. Structural signatures of known regulatory proteins were identified (Fig. 4) by using the Network Protein Analysis program. The analysis revealed a Myc-type helix-loop-helix dimerization domain signature, a leucine zipper motif, an EF-hand Ca²⁺-binding domain, a homeobox domain, and Nt-DnaJ among these three genes. The presence of these motifs suggests that these viral proteins may modulate host and/or viral transcription via protein-DNA interaction, thereby potentially affecting viral pathogenesis. Leucine zipper and helix-loop-helix motifs have been reported to be involved in latency of herpes simplex virus (3, 6). These motifs may also play some roles in WSSV latency, and this requires further rigorous investigation.

FIG. 4.

Demonstration of protein signatures in WSSV ORFs 151, 366, and 427. Putative protein motifs were found in translated products of the three genes by NPS@ PROSCAN. The percentage of homology of each segment corresponding to the known protein structures is indicated alongside its domains on the diagrams. The protein sizes of the products of WSSV ORFs 151, 366, and 427 are 1,437, 624, and 84 aa, respectively.

WSSV is more closely related to baculovirus in structure and morphology than is herpes simplex virus (14). Latent baculovirus sequences have been reported to be present in cell lines established from fat body tissue of Mamestra brassicae insects (4). In addition, Hughes and colleagues reported that virus-free control insects died of an M. brassicae multiple nucleocapsid nucleopolyhedrovirus-like infection after they had been fed with M. brassicae fat body cells despite the failure to detect virus sequences in M. brassicae fat body cells. It is thus possible that the M. brassicae cells harbored the infectious baculovirus at very low levels, referred to as persistent infection (5). This phenomenon is very similar to our observation of WSSV dormancy in SPF shrimps. Therefore, it is probable that the WSSV genome resides in hosts either in a quiescent state or by remaining as a persistent infection. The conclusion of persistent infection is in agreement with the results of other reports describing the prevalence of WSSV in the ecosystem and its predominant infection in shrimp aquaculture, which could only be detected by using sensitive PCR techniques (11).

In conclusion, we have constructed a WSSV-specific DNA microarray for the study of WSSV biology. Our procedure, which provides a sensitive detection method (i.e., using amplified material), facilitated the detection of WSSV genes in SPF shrimps, suggesting a novel utility for a microarray-based approach in detecting very low levels of infection. The commercialized diagnostic test, which uses single-step nested PCR and a conventional one-step PCR technique, was unable to detect WSSV in the same SPF shrimps. This could be due to the low copy number of WSSV particles in SPF shrimps, a number that is beyond the detection limit of the kit, which requires the presence of at least 20 copies of pure target plasmid per reaction for detection. Further, the SPF shrimps used in this study were reared over 6 generations in a controlled environment without any disease outbreak, suggesting that these SPF shrimps were likely to harbor very low levels of WSSV genome. This is in contrast to the asymptomatic shrimps that were used in a previous study (8), as those shrimps were reportedly reared in an exposed setting and over a shorter period of time and thus were likely to harbor more viral sequences than the SPF shrimps. The conventional one-step PCR is clearly not sensitive enough to detect low copy numbers of the viral sequence, as indicated by the negative results we obtained from tests conducted initially upon receipt of the uninfected shrimps. Our microarray procedure, however, was able to detect the presence of three transcripts, and the expression of these transcripts was further confirmed by real-time RT-PCR on SPF shrimp aRNA and by two-step PCR (which included more cycles of amplification) on several independent samples of cDNA from SPF shrimps, as well as on DNA material from these shrimps. Taken together, these results suggest that WSSV genomes are present in SPF shrimps and that the three viral genes, which appear to be latent in the shrimp, may subsequently contribute to active transcription. However, substantial additional evidence is required to ascertain the role of these three viral transcripts.

Clinical symptoms such as ambiguous white spots found on the carapace as a result of bacterial infection (15), recovery from infection, or other causes can hinder diagnosis of WSSV infection or lead to a misdiagnosis. Our findings could assist in the formulation of a more sensitive diagnostic method that uses WSSV ORFs 151, 366, and 427 to detect WSSV in shrimps and other crustaceans. In addition, our results can lead to better understanding of the establishment of viral latency in asymptomatic carriers, shedding light on the molecular mechanisms in WSSV-induced mortality, thereby identifying ways of regulating the expression of these regulatory proteins to prevent outbreak.