Dengue virus replication enhances labile zinc pools by modulation of ZIP8

Abstract Zinc‐dependent viral proteins rely on intracellular zinc homeostasis for successful completion of infectious life‐cycle. Here, we report that the intracellular labile zinc levels were elevated at early stages of dengue virus (DENV) infection in hepatic cells and this increase in free zinc was abolished in cells infected with UV‐inactivated virus or with a DENV replication inhibitor implicating a role for zinc homeostasis in viral RNA replication. This change in free zinc was mediated by zinc transporter, ZIP8, as siRNA‐mediated knockdown of ZIP8 resulted in abrogation of increase in free zinc levels leading to significant reduction in DENV titers suggesting a crucial role for ZIP8 in early stages of DENV replication. Furthermore, elevated free zinc levels correlated with high copy numbers of dengue genome in peripheral blood leukocytes obtained from dengue patients compared to healthy controls suggesting a critical role for zinc homeostasis in dengue infection. Take Aways Dengue virus utilises cellular zinc homeostasis during replication of its RNA. ZIP8 upregulates free zinc levels during dengue virus replication. Enhanced viremia associates with elevated intracellular free zinc in dengue.


| INTRODUCTION
Zinc is an essential micronutrient and an important structural constituent of various proteins such as transcription factors, enzymes, growth factors, cytokines and receptors involved in cellular signalling cascade (Fukada, Yamasaki, Nishida, Murakami, & Hirano, 2011). It is the second most abundant transition metal in cellular organism after iron and is an integral part of about 10% of all human proteins (Andreini, Bertini, & Cavallaro, 2011). The total cellular zinc content is in the range of hundreds of micromolar which is mainly constituted by protein-bound form of zinc. The labile (free) zinc pool in cytosol, which is more accessible, is of significant interest as its exchange between the cytosol and subcellular compartments by zinc transporters regulates physiological functions (Frederickson, Koh, & Bush, 2005). This exchange is mediated majorly through two protein families, Zrt, Irt-like proteins (ZIP) and zinc transporters (ZnT). ZIPs are responsible for importing zinc inside the cytosol from extracellular space or from the intracellular compartments while ZnTs transport Aleksha Panwar and Jigme Wangchuk contributed equally to this work. zinc out of the cytosol into intracellular organelles or into the extracellular space (Kambe, Hashimoto, & Fujimoto, 2014;Maret, 2017). Zinc homeostasis is modulated by external stimuli and redistribution of zinc into tissue spaces as part of acute-phase response may also limit or enhance the availability of zinc which is required for the survival and propagation of the invading pathogen depending on its tissue tropism.
Although the cellular level of free zinc is of several orders of magnitude lesser compared to the total zinc or the protein-bound form (Maret, 2013), they represent the zinc pool which is readily available and involved in cellular metabolism.
We further confirmed this in Huh-7 cells by infecting cells with DENV at 3 MOI and 1 hr after virus adsoprtion, cells were washed and serum-free medium containing DMSO or 0.5 μM TPEN was added and cultured for 24 hr. Viral titers in the supernatant were measured by plaque assay at 24 hpi. Consistent with our previous findings, we observed that zinc depletion by TPEN led to a significant reduction in DENV titers indicating the importance of zinc in DENV life-cycle ( Figure 1a). Similarly, we performed zinc supplementation experiments, where cells were infected as above and serum-free medium containing 100 μM of ZnSO4 was added after virus adsorption and cultured for 24 hr and virus titers were measured by plaque assay at 24 hpi. As observed earlier in epithelial cells (Kar et al., 2019;Khan et al., 2020), supplementing media with excess zinc did not have any effect suggesting that unlike other viruses which are susceptible to inhibition by zinc (Suara & Crowe, 2004;Yuasa et al., 2006;Zhang et al., 1991), DENV infection is not perturbed by excess zinc ( Figure 1b). Since DENV replication was sensitive to perturbation in zinc homeostasis and labile zinc pool acts as an indicator of change in zinc homeostasis, we next sought to determine whether DENV infection leads to modulation of zinc homeostasis. Labile zinc levels were visualised by confocal microscopy using zinc fluorophore, fluozin-3 a.m. (FLZ-3 a.m.), in Huh-7 cells infected with DENV at 8 and 16 hpi.
We observed a two-fold increase in labile zinc levels in DENV infection at 16 hpi (Figure 1c(i),(ii)). The labile zinc levels at 24 hr were comparable with the mock-infected samples further suggesting that the transient nature of this induction is most likely to meet the demand of viral RNA replication (Figure 1c(i),(ii). The increase in labile zinc levels were found to be MOI-dependent as we observed $1.5-fold and $3-fold increase at 0.3 and 3 MOI, respectively, which coincided with exponential phase of virus replication as estimated by qRT-PCR and plaque assay respectively (Figure S1a-c). Further, we verified the above observations in primary human hepatocytes which were infected with DENV at 5 MOI and stained with FLZ-3 a.m. at 16 hpi. Similar to Huh-7 cells, primary hepatocytes showed elevated free zinc levels upon DENV infection (Figure 1d(i),(ii)). We observed $80-90% infection under these conditions in these primary cells which was similar to what we observe with 3 MOI in Huh-7 cells ( Figure S1d). These data suggest that DENV infection modulates zinc homeostasis to meet the demand for zinc in viral replication by inducing elevation in cytosolic free zinc levels in cells of hepatic origin. To verify this, we infected cells with UV-inactivated DENV and measured labile zinc levels at 16 hpi. We found that UVinactivated virus did not lead to increase in zinc levels as efficiently as an actively replicating virus (Figure 1e(i), (ii)). We next used fluoxetine hydrochloride (FLX), which we have shown earlier to inhibit DENV replication (Medigeshi, Kumar, Dhamija, Agrawal, & Kar, 2016). We observed no change in zinc levels when cells were treated with FLX post-infection ( Figure 1f). To further validate this observation, we used BHK-21 cells stably expressing a DENV-2 replicon ( Figure S1e) (Boonyasuppayakorn, Reichert, Manzano, Nagarajan, & Padmanabhan, 2014). Mock BHK-21 cells and cells stably expressing DENV-2 replicon were stained with FLZ-3 a.m.
to visualise labile zinc levels. (Figure 1g). Similar to virus infection in Huh-7 cells, we observed a significant increase in labile zinc levels in cells expressing the DENV replicon suggesting that active replication of DENV leads to increase in labile zinc pools in the cytosol.
We next sought to identify the mechanism of increase in labile zinc levels in dengue infection. Free zinc levels in the cytoplasm may increase due to enhanced cellular uptake of zinc or due to redistribution of zinc from intracellular organelles to cytosol or by the release of zinc from metallothioneins which act as zinc stores (Kimura & Kambe, 2016). Huh-7 cells were infected with DENV and at 16 hpi, total cell extracts were subjected to inductively coupled-plasma mass spectrometry (ICP-MS) for the detection of total zinc (Zn) content.
The amount of zinc ions detected by ICP-MS in parts per billion (ppb) was normalised to the total protein content of the cells to account for any difference in cell numbers. We did not observe any change in the total zinc content in DENV infection conditions ( Figure 2a). Flavivirus replication utilises cytosolic membranous compartments derived from the ER and Golgi (Mackenzie, 2005). Therefore, we were interested to examine whether this free zinc pool is localised to flavivirus replication compartments. However, localization of free zinc with viral proteins and viral replication compartment was not possible as permeabilization of cells led to loss of FLZ-3 a.m. signal. Therefore, we transfected plasmid constructs coding for mCherry-tagged ER or Golgi-resident proteins in Huh-7 cells followed by infection with DENV after 24 hr and visualised free zinc levels by confocal microscopy at 16 hpi. We consistently observed higher levels of free zinc in infection conditions which concentrated in the perinuclear region as We speculated that the increase in cytosolic zinc induced by dengue virus infection could be regulated by zinc transporters. To test F I G U R E 1 DENV infection leads to increase in cytosolic free zinc levels. Huh-7 cells were infected with DENV at 3 MOI and after viral adsorption, cells were cultured in either 0.5 μM TPEN (a) or 100 μM ZnSO 4 (b) and viral titers in supernatants was estimated by plaque assay at 24 hpi. (n = 9 for both [a] and [b]) (c) Huh-7 cells were infected with DENV and free zinc levels were determined by immunofluorescence assay at indicated time points. (i) Representative images of cells stained with FLZ-3 a.m., (ii) Sum grey intensity of FLZ-3 stain was evaluated using cellSens software. Data represent relative change in sum grey intensity of FLZ-3 a.m. stain in cells (n = 95-115) counted from eight different fields. (d) Primary hepatocytes were infected with 5 MOI of DENV and free zinc levels were assessed by immunofluorescence assay at 16 hpi. (i) Representative images of cells stained with FLZ-3 a.m., (ii) Data represent relative change in sum grey intensity of FLZ-3 a.m. stain in cells (n = 15-20) counted from seven different fields. (e) Huh-7 cells were either mock-infected or with 3 MOI of DENV or UV-inactivated virus. Cells were stained with FLZ-3 staining at 16 hpi. (i) Representative images (upper panel) and DENV replication was assessed by dsRNA staining using antibody against dsRNA (lower panel), (ii) Data represent relative change in sum grey intensity of FLZ-3 a.m. stain in cells (n = 80-100) counted from seven different fields. (f) Huh-7 cells were infected with DENV and fluoxetine hydrochloride (FLX) was added at a concentration of 4 μM after virus adsorption. Data represent fold change in sum grey intensity of FLZ-3 a.m. stain in cells (n = 80-100) counted from eight different fields. (g) Free zinc levels were determined in BHK-21 cells stably expressing DENV-2 replicon and mock BHK-21 cells using FLZ-3 a.m. stain. Representative images are shown. All the data presented here are from two or three independent experiments. Data are presented as M ± SD. Scale bar is 10 μM. ns, not significant. **p < .01; ***p < .001  (King et al., 1999;Schmid, Diamond, & Harris, 2014;Valero et al., 2014). Based on our in vitro data, we next probed whether blood cells from dengue patients show an increase in labile zinc levels. We recruited 21 paediatric dengue patients who were in the viremic phase (within 7 days of fever) into the study to collect blood samples and analyse for free zinc levels in peripheral blood cells by flow cytometry. The clinical features of these patients have been reported recently as part of another study (Khan et al., 2021). As per WHO classification guidelines, 8 patients had mild dengue, 7 had dengue with warning signs and 6 were with severe dengue. Total RNA was isolated from whole blood and processed for estimation of viral RNA copy numbers by qRT-PCR as described previously (Kar et al., 2017;Khan et al., 2021;Singla et al., 2016). We also

| EXPERIMENTAL PROCEDURES
List of reagents used in the study is provided in Table S1. Screening and enrolment of patients were exactly as described previously (Kar et al., 2017;Khan et al., 2021;Singla et al., 2016). Written informed consent for the study was taken from parents/guardians to collect blood samples at the time of admission. Children aged between 4-14 years with symptoms suggestive of dengue were screened using a

| Detection of free zinc levels by confocal microscopy
FluoZin-3 a.m. (FLZ-3 a.m.) (Invitrogen) was used to determine the intracellular free zinc levels. Medium was removed and cells were washed twice with 1X PBS followed by incubation with FLZ-3 a.m. at a concentration of 10 μM. Cells were incubated for 20 min at 37 C.
After incubation, cells were washed three times with 1X PBS and stained with 4 0 ,6-diamidino-2-phenylindole (DAPI) at a dilution of 1:10,000 in PBS for 5 min. Cells were fixed using 3% paraformaldehyde (PFA) for 10 mins followed by washing with 1X PBS three times and mounted using Prolong Gold antifade reagent (Invitrogen). Images were captured using FV3000 confocal microscope (Olympus). Fluorescence images with FLZ-3 a.m. were captured using an excitation wavelength of 488 nm. The parameters for detection and capturing images were digitally controlled to keep same settings throughout the experiments. For quantitative analysis, sum grey intensity per cell was calculated using CellSens software (Olympus) and plotted as relative change.
For zinc supplementation experiment, 100 μM of ZnSO 4 was added in serum-free media after viral adsorption and viral titers were measured at 24 hpi. To study the effect of fluoxetine hydrochloride on DENVinduced free zinc levels, Huh-7 cells were infected with DENV and fluoxetine hydrochloride was added after viral adsorption at a concentration of 4 μM in 2% DMEM. Cells were incubated for 16 hpi and processed for FLZ-3 a.m. staining. Coverslips were mounted on glass slides, sealed and images were captured at 100X using FV3000 confocal microscope (Olympus).

| Transient transfections
Huh-7 cells were seeded at 60,000 cells per well in a 24-well plate on glass coverslips. 24 hr later, cells were washed with 1X PBS, and transfected with 800 ng of plasmids expressing mCherry-tagged endoplasmic reticulum (Addgene-55,041) and Golgi (Addgene-55,052) markers in serum-free media using Lipofectamine2000 (Invitrogen) as per the manufacturers' instructions. After 6 hr incubation period, the transfection mixtures were removed and replaced with fresh growth medium. After 24 hr post transfection, cells were infected with DENV at 3 MOI. At 16 hpi, cells were stained with FLZ-3 a.m. and processed for confocal microscopy as described in the previous sections.

| Viral dsRNA staining
At 16 hpi, cells were washed with cold PBS and fixed in ice-cold methanol at À20 C for ≥20 min. Cells were washed twice with PBS followed by incubation in 0.2% BSA in IMF buffer (20 mM HEPES pH 7.5, 0.1% Triton X-100, 150 mM NaCl, 5 mM EDTA, 0.02% sodium azide) for 1 hr at room temperature (RT). Cells were then incu- Signals were detected using a gel documentation system (Azure biosystems C400).

| Ultraviolet (UV)-inactivation of DENV
DENV inactivation was carried out using a UV crosslinker (UVP-CL1000S). DENV stock was diluted in a ratio of 1:100 in 1 ml of MEM media and placed in a 35 mm diameter petri dish. DENV stock was treated with a dose of UV light at 9999 Â 10 2 μJ/cm 2 . The supernatant was exposed three times for 10 min with an interval of 3 min.
After the inactivation process, virus stock was aliquoted and stored at À80 C, later titered and used for the experiment. UV-inactivated virus showed no replication as observed by viral dsRNA staining.

| siRNA knockdown
Smartpool siRNAs targeting the human ZIP1 and ZIP8 genes or nontargeting controls (NTC) were purchased from Dharmacon. Transfections were carried out as described previously . All transfections were performed as per manufacturer's instructions.
Briefly, 10 nM concentrations of NTC, ZIP1 and ZIP8 siRNAs were mixed with Opti-MEM (Life Technologies) and 1 μl of Lipofectamine RNAiMax to a total volume of 100 μl in a 24-well plate. Cells were trypsinized and volume made up so as to contain 30,000 cells in 400 μl antibiotic-free medium. After 20 min incubation of the transfection complex, cell suspension was added into each well. Knockdown efficiency was determined by qRT-PCR at 48 hr post transfection.

| Quantitative real time PCR (qRT-PCR)
Huh-7 cells were infected with DENV and at indicated time points, cells were collected in TRIzoL reagent (Takara) and RNA was isolated using manufacturer's instructions. cDNA synthesis was performed using PrimeScript RT reagent kit with gDNA eraser (Takara). 100 ng of cDNA was used to determine genes expression using DyNAmo flash SYBR green quantitative PCR reagent (Thermo Scientific). Reaction conditions used were as follows: (95 C-7 min; 95 C-10 s followed by 60 C for 30 s). GAPDH primer was used as housekeeping control. For DENV RNA detection by reverse-transcription polymerase chain reaction (RT-PCR), total RNA was extracted from cells at the indicated time points using RNAiso Plus (TaKaRa), and 200 ng of RNA was used in multiplex TaqMan one-step RT-PCR with DENV primers, DENV probe and human GAPDH primer-probe mix (Applied Biosystems). At indicated time points, supernatant was collected for estimating viral titers by plaque assay and cells were harvested for positive and negative strand detection PCR as described previously (Kar et al., 2017).
Expression levels of GAPDH was used to calculate fold change and normalisation. Data were analysed using the ΔΔ C T method, where C T is threshold cycle.

| Data analysis
Data were analysed and graphs were prepared using Prism 7 (Version 7.0e) software (GraphPad Software Inc.). All the graphs represent results from two or more independent experiments; values are presented as M ± SD. Statistical significance was estimated by Mann-Whitney test. The data were corrected for multiple comparisons using Bonferroni-Dunn method wherever applicable.

ACKNOWLEDGMENTS
We thank all the members of the CCV lab for their support and critical inputs. We thank all the patients for providing their samples for the study. This work was supported by DBT/Wellcome Trust India Alliance Fellowship [grant number IA/S/14/1/501291] awarded to GRM.
The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

CONFLICT OF INTEREST
The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT
All the data presented in the manuscript is available as part of this manuscript and as supplementary information.