The Anopheles gambiae mosquito is the principle vector of the malaria parasite Plasmodium falciparum in sub-Saharan Africa. Current estimates suggest that nearly half of the global population is at risk of malaria and there are annually approximately 250 million cases resulting in a million deaths [1]. In addition, An. gambiae vectors o'nyong-nyong virus (ONNV), a single-stranded (+) RNA virus (Togaviridae; Alphavirus) [2-4]. Symptoms of ONNV infection in humans include rash, fever and polyarthritis often resulting in high morbidity rates during epidemics [5,6].

Although most arthropod-borne viruses (arboviruses) are vectored by culicine mosquitoes, ONNV is primarily vectored by the anopheline mosquitoes An. gambiae and An. funestus [7]. In spite of the unusual vector specificity, ONNV shares a common host cell entry mechanism with many other arboviruses. Endocytosis and subsequent fusion with the host's membrane in the endosome are exploited by ONNV to infect host cells [8]. Alphaviruses, including ONNV, Sindbis virus, and Chikungunya virus have class II fusion proteins such as E glycoproteins that mediate membrane fusion between virus and host cells during virus entry [8,9]. Class II E glycoproteins mainly consist of beta sheet-folded domains while class I E proteins are α-helices [10,11]. Since membrane fusion is one of the protein maturation processes mediated by molecular chaperones, such as the HSP70 family, it is possible that HSP70 may enhance or suppress maturation of viral proteins [12-14].

Members of the HSP70 family contain three conserved domains: an ATPase domain at the N-terminus, a peptide binding domain, and a GP-rich region at the C-terminus that contains an EEVD motif [15-17]. HSP70, a molecular chaperone, changes its conformation in an ATP dependent manner to mediate proper target protein folding, degradation and translocation [18,19]. The carboxy-terminal EEVD motif is a unique feature of cytosolic heat shock proteins that is recognized by chaperone cofactors to initiate chaperone activity [20-22]. The heat shock cognate protein 70 (HSC70) is a constitutively expressed member of the HSP70 family and functions as a molecular chaperone under normal cellular conditions. However, the expression of the HSC70 gene may be increased in response to environmental and physiological stress [19].

The An. gambiae HSC70B is an ortholog of Drosophila melanogaster Hsc70-4 [23]. cDNA microarray studies demonstrated that HSC70B is upregulated during ONNV infection in adult An. gambiae, suggesting an important role during virus infection [24]. The functional importance of HSC70B upregulation in ONNV-infected female An. gambiae was further elucidated by RNAi gene silencing of the hsc70B transcript [23]. Reduction of the hsc70B transcript by RNAi silencing enhanced ONNV replication in vivo. Likewise, enhanced ONNV replication in HSC70B-knockdown mosquitoes suggests that HSC70 proteins play an important role in arbovirus suppression and maintaining homeostasis during infection [23].

To further elucidate the transcriptional regulation of the hsc70B locus in response to viral infection, we characterized the 5' upstream region of the hsc70B coding sequence ex vivo using cell culture and luciferase reporter systems. Herein, we report the identification of a regulatory region essential for hsc70B transcription. Furthermore, the kinetic properties of hsc70B transcription during arbovirus infections were examined with ONNV (Togaviridae; Alphavirus), West Nile virus (Flaviviridae; Flavivirus) and La Crosse virus (Bunyaviridae; Orthobunyavirus). The results showed that the hsc70B promoter region was responsive to all three arboviruses. Induction of hsc70B transcription by three taxonomically different arboviruses suggests that the HSC70B protein may be expressed to cope with cellular stress imposed during infection. The biological implications of these data are discussed.

Figure 1 Nucleotide sequence of the hsc70B promoter region. Putative binding sites for transcription factors are underlined. The binding sites were evaluated in silico by both the TFSEARCH and AliBaba2.1 program. Transcription factors predicted by both programs (more ...)

Figure 2 Deletion Analysis of the hsc70B promoter. (A) The solid black line represents the full length of the promoter where positions -2599 and +1 denote the 5' end of the hsc70B promoter and the putative starting codon ATG, respectively. (B) The bars on the (more ...)

Figure 3 Induction of the hsc70B promoter in transfected Vero cells by ONNV infection. The constructs containing the 2.6 kb and 2.2 kb-long 5' upstream regions were evaluated for hsc70B promoter activity during ONNV infection (MOI = 0.001). The luciferase activity (more ...)

Figure 4 hsc70B promoter activity (mean ± SD) time course experiments during ONNV infection. (A) The hsc70B promoter activity, as measured by luciferase activity, was significantly higher at 48 h (P < 0.01, t = 8.53, df = 4) and 72 hpi (P < (more ...)

To determine if the observed transcriptional activation of hsc70B was virus specific, two taxonomically distinct arboviruses were chosen for additional time course experiments. Vero cells transfected with pGL3-2.2k constructs were infected with WNV or LACV at MOI = 0.01. The infected cells were harvested at 1, 24, 36 and 48 hpi. Infection with either virus also significantly increased the hsc70B promoter activity as determined by the luciferase assay (Figure 5).

Figure 5 Increased hsc70B promoter activities (mean ± SD) in Vero cells during West Nile virus (A) and La Crosse virus (B) infection. The hsc70B promoter activity, as measured by luciferase activity, is higher in infected cells at 36 and 48 hpi.

Repression of ONNV replication by the HSC70B protein was previously shown in An. gambiae [23,24]. Of particular interest in this result is the transcriptional regulation of hsc70B expression in response to ONNV infection in An. gambiae. To map and characterize the promoter activity of hsc70B, the upstream region up to 2599 bp from the putative starting codon of hsc70B was subjected to a luciferase reporter assay. Initially, the 2599 bp upstream sequence of the hsc70B showed a promoter activity (Figure 2). Subsequent deletion analysis of this region revealed that the regulatory elements critical for hsc70B transcription reside between 2150 ~ 1615 bp upstream of the hsc70B starting codon (Figure 2). Deletion of this 535 bp region abolished the promoter activity of hsc70B. This regulatory region contains several binding sites for transcription factors such as HSE-bind, CRE, NF-κB, dorsal, and Ftz (Figure 1). HSE is a binding site for heat shock transcription factors that are activated in response to environmental and physical stresses such as heat shock and microbial infection [25,26]. In hsc70B, there is one putative HSE consisting of a block of three repeats of a 5-bp sequence, 5'-nGAAn-3'. Although the number of HSE blocks can vary among different HSPs, the 5-bp HSE repeat is highly conserved in the regulatory region of various heat shock proteins such as hsp70, hsp83, and hsp27 in Drosophila [27]. The second and third repeat in the HSE block of An. gambiae hsc70B has a tail-to-tail (5'nTTCnnGAAn3') arrangement with 6-bp gaps between them (Figure 1). In Drosophila HSPs, there are 5 or less gaps, if any, between the 5-bp repeats [27]. It will thus be interesting to learn how the additional gap in An. gambiae hsc70B regulates hsc70B expression.

CRE is a response element for phosphorylated CREB (c-AMP response element-binding protein) which regulates transcription of genes. CREB is involved in human hsp90 gene expression which is constitutively expressed [28]. Thus, CRE may be a key element to induce basic transcription of An. gambiae hsc70B gene as it is also a constitutively expressed member of HSPs. NF-κB is a transcription factor which responds to stresses including viral infection [29]. Transcription of NF-κB was shown to be increased by downregulation of HSC70B protein in rat pancreatic acinar AR42J cells [30]. Ftz is a transcription factor that was originally isolated in Drosophila. It has many orthologs in various species and is involved in fushi tarazu gene expression which functions in embryonic segmentation in Drosophila and sex determination in zebrafish [31,32]. Further biochemical and molecular characterization using electrophoretic mobility shift assays (EMSA) and DNase I protection assay should elucidate key elements that transcriptionally regulate An. gambiae hsc70B expression in response to ONNV infection. These assays will further improve our understanding of transcriptional regulation of hsc70B, and facilitate the identification of transcriptional factors and co-factors in the signal transduction pathway of hsc70B expression.

ONNV was used to infect Vero cells to examine the effects on hsc70B promoter activity. The different lengths, 2150 bp and 2599 bp, of the 5' upstream sequences were tested because these two constructs contain the regulatory sequence for the basic transcription of hsc70B. Both 2150 bp and 2599 bp upstream genomic fragments responded to ONNV infection and the promoter activities of both constructs increased during ONNV infection (Figure 3). When Vero cells were transfected with either pGL3-2.6k or pGL3-2.2k reporter plasmid, the promoter activity in the reporter plasmids was about 2-fold higher in infected cells than the uninfected control (Figure 3). This suggests that induction of An. gambiae hsc70B gene, leading to expression of the HSC70B protein, results from virus infection. Therefore, it is reasonable to speculate that cellular signals are transduced to the regulatory region of the hsc70B locus in An. gambiae.

The 2150-bp 5' upstream sequence was used to further investigate the effects of ONNV infection on the hsc70B promoter activity at different time points after infection. The promoter activity of hsc70B was significantly higher in infected cells at later time points (48 and 72 hpi) than earlier points (1 and 24 hpi) (Figure 4A). The elevated hsc70B promoter activity corresponded with increasing viral titers at 48 and 72 hpi because plaque assays of the cell culture media showed higher ONNV titers at these later time points (Figure 4A). These plaque assay data were further evaluated by observing CPE in ONNV-infected Vero cells. CPE became apparent at 48 and 72 hpi in Vero cells while uninfected control cells did not show cell lysis (Figure 4B). The appearance of CPE in ONNV-infected Vero cells corresponded to higher ONNV titers at 48 and 72 hpi. It can be thus inferred that induction of hsc70B transcription may be triggered in response to cellular stresses burdened by rapidly replicating viruses. In cells at immediate or early infection stages, hsc70B expression may not be activated.

The inducibility of the hsc70B promoter was also examined using two additional arboviruses, WNV (Flaviviridae) and LACV (Bunyaviridae). Like ONNV, both WNV and LACV were also able to upregulate the transcription activity of hsc70B during infection (Figure 5). Due to more rapid kinetics of replication, both WNV and LACV caused the hsc70B promoter activity to rise earlier than ONNV. For example, WNV-infected Vero cells started to show transcriptional induction as early as 24 hpi. Transcriptional activation of hsc70B by three different arboviruses suggests that upregulation of hsc70B expression indeed results from cellular stresses caused by virus infection in host cells. In addition, activation of the hsc70B promoter by virus infection was recently shown in shrimp (Penaeus monodon) [33]. Using a luciferase reporter in Sf21 cells, Chuang et al. (2007) demonstrated 5.5-fold induction of the shrimp hsc70B promoter when the Sf21 cells were infected with Autographa californica multiple nuclear polyhedrosis virus (AcMNPV; MOI = 0.1). Therefore, it appears that induction of hsc70B expression may be a general cellular response of host cells to virus infection.

We previously reported that the transcriptional activation of hsc70B in ONNV-infected An. gambiae renders the mosquito an ability to repress ONNV replication [23,24]. These in vivo findings and our current ex vivo characterization of the hsc70B regulatory region unequivocally indicate that the induction of HSC70B may be a mosquito innate immune response against virus infection. To support this hypothesis, mosquito cells (e.g., C6/36 cells from Ae. albopictus) do not show any CPE during arbovirus infection while mammalian cells including Vero cells display prominent CPE and subsequent cell lysis due to overreplication of viruses. Evolutionally, mosquitoes may have acquired the ability to maintain viral titers below a certain threshold, below which mosquitoes may serve as arboviral vectors without pathogenesis from viral infections. Interestingly, a potent antiviral drug, prostaglandin A, showed antiviral effects against Sendai or Sindbis virus through induction of HSP70 proteins in AGMK cells (African green monkey kidney) or Vero cells, respectively [34,35]. Therefore, comparative studies on HSP expression in response to viral infection between mosquito and mammalian cells will provide a deeper insight into innate immune responses to viral infection between mosquito vectors and mammalian hosts.

The authors declare that they have no competing interests.

SK performed the experiments, analyzed the data, and drafted the manuscript. CS contributed to the cloning of the hsc70B locus and reviewed the manuscript. BDB conducted cell culture and viral plaque assays and reviewed the manuscript. FHC initiated the project and provided materials and a critical review of the manuscript. YSH provided overall direction and conducted experimental design, data analysis and wrote the manuscript. All authors read and approved the final manuscript.

The authors are grateful to M. Kern of University of Notre Dame for providing BAC clones containing the hsc70B locus. We appreciate Dr. D. Wesson of Tulane University for providing reagents and cell culture facilities and M. Rider of Tulane University for his critical review of the manuscript. This work was supported by the Louisiana Board of Regents Fund (LEQSF(2005-08)-RD-A-35) and Tulane Research Enhancement Fund to YSH.