Ex vivo promoter analysis of antiviral heat shock cognate 70B gene in Anopheles gambiae

Background The Anopheles gambiae heat shock cognate gene (hsc70B) encodes a constitutively expressed protein in the hsp70 family and it functions as a molecular chaperone for protein folding. However, the expression of hsc70B can be further induced by certain stimuli such as heat shock and infection. We previously demonstrated that the An. gambiae hsc70B is induced during o'nyong-nyong virus (ONNV) infection and subsequently suppresses ONNV replication in the mosquito. To further characterize the inducibility of hsc70B by ONNV infection in An. gambiae, we cloned a 2.6-kb region immediately 5' upstream of the starting codon of hsc70B into a luciferase reporter vector (pGL3-Basic), and studied its promoter activity in transfected Vero cells during infection with o'nyong-nyong, West Nile and La Crosse viruses. Results Serial deletion analysis of the hsc70B upstream sequence revealed that the putative promoter is likely located in a region 1615–2150 bp upstream of the hsc70B starting codon. Sequence analysis of this region revealed transcriptional regulatory elements for heat shock element-binding protein (HSE-bind), nuclear factor κB (NF-κB), dorsal (Dl) and fushi-tarazu (Ftz). Arbovirus infection, regardless of virus type, significantly increased the hsc70B promoter activity in transfected Vero cells. Conclusion Our results further validate the transcriptional activation of hsc70B during arbovirus infection and support the role of specific putative regulatory elements. Induction by three taxonomically distinct arboviruses suggests that the HSC70B protein may be expressed to cope with cellular stress imposed during infection.


Introduction
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'nyongnyong virus (ONNV), a single-stranded (+) RNA virus (Togaviridae; Alphavirus) [2][3][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][13][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][16][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][21][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 HSC70Bknockdown 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.

Sequence analysis of the 5' upstream of hsc70B
Transcription factor binding elements along the 5'upstream sequence of the hsc70B gene (2559 bp) were analyzed in silico. The binding sites identified by both the TFSEARCH and AliBaba2.1 programs are shown in Figure  1. In addition to core promoter sequences (e.g., TATA and CAT boxes), putative binding sites for heat shock proteins such as HSE-bind and heat shock transcription factor (HSF) were also identified. Putative binding sites for NF-κB, Dl, c-AMP response element binding protein (CREB), signal transducers and activators of transcription protein (STAT), and fushi-tarazu (Ftz) factors were also identified.

Deletion analysis of the hsc70B promoter
To identify the critical elements required for transcription, various deletions of the 5' upstream region of the hsc70B locus were generated and ligated into the pGL3-Basic vector. The promoter activities of the different deletion constructs were compared to that of the full-length construct (2.6 kb). The full length promoter pGL3-2.6k contains 2599 bp of the 5' upstream region (-2599 to -1); +1 denotes the first base of the starting codon ( Figure 2A). Deletions of 449 bp, 975 bp, 1649 bp and 2267 bp from the 5' end of the full length promoter produced pGL3-2.2k, pGL3-1.6k, pGL3-0.9k and pGL3-0.3k, respectively. The promoter activity of these deletion constructs was measured by firefly luciferase expression and normalized by the Renilla luciferase expression. Both pGL3-2.6k and pGL3-2.2k constructs had luciferase expression levels 5fold higher than that of the pGL3-Basic control. The luciferase expression levels of pGL3-0.3k, pGL3-0.9k and pGL3-1.6k did not differ from that of the control ( Figure  2B). These data suggest that elements critical for the transcription of hsc70B reside in the 526 bp region between 2.2 and 1.6 kb upstream of the starting codon.

Effect of ONNV infection on the pGL3-2.6k and pGL3-2.2k hsc70B promoter plasmid constructs
To determine if differences between the promoter activities of the pGL3-2.6k and pGL3-2.2k constructs occurred during arbovirus infection, the constructs were initially evaluated in the context of ONNV infection. Transfected with either the pGL3-2.6k or pGL3-2.2k plasmids, Vero cells were subsequently infected with ONNV (MOI = 0.001). The cells were harvested after cytopathic effects (CPE) were confirmed at 60 hpi. ONNV infection significantly increased the hsc70B promoter activity (Figure 3). The luciferase activity of both pGL3-2.6k and pGL3-2.2k constructs in ONNV-infected Vero cells was ~2-fold higher than uninfected Vero cells.

Effect of arbovirus infection on the hsc70B promoter activity
Based on the previous results, the pGL3-2.2k construct was used to assay the effect of arbovirus infection on the hsc70B promoter. A time course experiment with ONNV (MOI = 0.001) in Vero cells transfected with the pGL3-2.2k construct demonstrated increases in hsc70B pro-moter activities at 48 and 72 hpi. However, at earlier time points the hsc70B promoter activity was comparable to that of the uninfected control ( Figure 4A). This enhanced hsc70B promoter activity in ONNV-infected cells appeared to occur with increasing ONNV titers at 48 and 72 hpi. The titers were 1.5 × 10 2 , 3 × 10 5 , 1.4 × 10 8 , and 1.1 × 10 8 plaque forming units (pfu)/mL at 1, 24, 48 and 72 hpi, respectively ( Figure 4A). Furthermore, CPE in ONNVinfected Vero cells became evident at 48 hpi, corresponding with the elevated viral titers at later time points ( Figure  4B). The hsc70B promoter activity in ONNV infected Vero cells was 1.4 and 1.6-fold higher at 48 hpi and 72 hpi, respectively, than that in uninfected Vero cells ( Figure  4A).
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 Nucleotide sequence of the hsc70B promoter region hsc70B promoter activity as determined by the luciferase assay ( Figure 5).

Discussion
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 consist-ing 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 Deletion Analysis of the hsc70B promoter The bars on the left represent the lengths of the 5' upstream region that were generated by PCR. The bars on the right represent relative firefly luciferase activities (mean ± SD) that were normalized by the Renilla luciferase activity. The relative luciferase activity indicates the promoter activity of the 5' upstream deletion constructs of hsc70B. The promoter activities of constructs less than 2.2 kb were significantly lower than the 2.6 kb full length construct. 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  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.

Induction of the hsc70B promoter in transfected Vero cells by ONNV infection
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.

Conclusion
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.

Construction of An. gambiae hsc70B promoter-luciferase reporter gene
The 2599 bp 5' region upstream of the putative starting codon of the hsc70B gene was amplified from BAC clone 132E18 http://www.ensembl.org by a PCR method using Phusion High-Fidelity DNA polymerase (NEB, MA). The primers used were as follows: AngaHsc_F1, 5'-CCCGAGCTCGATGGTCACAAATGTTTCACAGG-3' and AngaHsc_R, 5'-CCGCTCGAGCTGCGAACACG-CAACACAC-3' with a SacI or an XhoI recognition site (underlined) incorporated at the 5' end of the primers, respectively. The PCR conditions were as follows: 98°C for 30 sec, followed by 30 cycles of denaturation at 98°C for 10 sec, annealing at 68°C for 30 sec and extension at 72°C for 80 sec, a final extension at 72°C was performed for 10 min. The amplified DNA fragment was doubledigested with SacI and XhoI and subcloned into the promoterless pGL3-Basic vector (Promega) predigested with SacI and XhoI to construct pGL3-2.6k. Serial deletions of the 5'-flanking region of the hsc70B gene were also prepared from pGL3-2.6k using a PCR method with the primers listed in table 1.

Analysis of 5' upstream sequence of hsc70B
Putative binding sites for transcription factors in the 5' upstream region of hsc70B were predicted in silico using the TFSEARCH [36,37] and AliBaba2.1 [38] programs.

Transfection and luciferase activity assay of the hsc70B promoter activity in Vero cells
Transfection experiments were performed in 24-well plates using the Lipofectamine reagent according to the manufacturer's instructions (Invitrogen, CA). Briefly, Vero cells (ATCC: CCL-81) were seeded and incubated at 37°C with 5% CO 2 in Dulbecco's Modified Eagle Medium (DMEM) for 24 h prior to transfection at a density of 0.5 × 10 5 cells/well. When the cells reached ~80% confluency, the culture media was removed and 200 μl of fresh DMEM without antibiotics or fetal bovine serum (FBS) was added. The cells were then co-transfected with 400 ng of pGL3 firefly (Photinus pyralis) luciferase constructs containing varying lengths of the hsc70B upstream region (e.g., pGL3-2.6k, pGL3-2.2k, pGL3-1.6k, pGL3-0.9k, pGL3-0.3k, or pGL3-Basic) and 0.05 ng of a pRL-cmv Renilla reniformis luciferase construct. The pRL-cmv construct was used as an internal control, in which the Renilla luciferase expression is driven by the cytomegalovirus promoter (cmv). Because the pGL3-Basic is a promoterless reporter plasmid containing only the coding sequence of firefly luciferase it served as a background control. At 3 h  Hours-post infection Relative luciferase activity uninfected WNV post transfection, the transfection mixture was replaced with a complete medium including 100 U/mL Penicillin-Streptomycin, and 10% FBS. Cells were harvested at predefined time points post transfection. The luciferase activities were measured by the Dual-Luciferase Reporter Assay System (Promega, WI) according to manufacturer's instructions. Quantification of the luminescent signals was performed using a Synergy HT microplate reader (BioTek, USA). In order to account for heterogeneous transfection efficiencies and cell viabilities among different samples, the firefly luminescence values were normalized as a ratio of the Renilla luminescence values. A minimum of three biological replicates were included for the time course experiments with ONNV. For time course experiments with WNV and LACV, the mean values and standard deviations were calculated from four biological replicates out of six replicates. The largest and the smallest values from these replicates were excluded from the analysis.

Viruses
The SG650 strain of ONNV has previously been described [23]. The WNV isolate (LA-  was isolated by BDB from the brain tissue of a blue jay (Cyanocitta cristata) found in New Iberia, LA during 2005. A cloacal swab from the bird tested positive for WNV by the Rapid Analyte Measurement Platform (RAMP, Adapco, Inc.). Subsequent nucleic acid amplification and sequencing of the PreM-Envelope region of the isolate confirmed the RAMP identification (GenBank Accession Number DQ646699). The virus was isolated in Vero cells and had not been further passaged. The LACV (78-V-13193) was obtained from the World Reference Center for Arboviruses at the University of Texas Medical Branch, Galveston, TX. The virus had been passed once in suckling mouse brain and twice in Vero cells.

Virus infection
To determine the effect of viral infection on the promoter activity of hsc70B, Vero cells cotransfected with pGL3-2.6k or pGL3-2.2k and pRL-cmv were infected with ONNV, WNV or LACV 12 h post transfection. For ONNV, confluent monolayers of Vero cells were infected at an MOI (multiplicity of infection) of 0.001. The infected cells were harvested at predetermined time points (e.g., 1, 24, 48 and 72 h post-infection) during time course experiments. Otherwise, the cells were harvested at 60 h post infection when CPE were evident. For the WNV and LACV time course experiments, confluent monolayers of Vero cells were infected at an MOI of 0.01 and the infected cells were harvested at 1, 24, 36 and 48 hpi. Viral titers were determined by a standard plaque assay in Vero cells [39].