Efficiencies and kinetics of infection in different cell types/lines by African and Asian strains of Zika virus (ZIKV)

Abstract

After recent outbreaks, Zika virus (ZIKV) was linked to severe neurological diseases including Guillain-Barré syndrome in adults and microcephaly in newborns. The severities of pathological manifestations have been associated with different ZIKV strains. To better understand the tropism of ZIKV, we infected 10 human and 4 nonhuman cell lines (types) with 2 African (IbH30656 and MR766) and 2 Asian (PRVABC59 and H/FP/2013) ZIKV strains. Cell susceptibility to ZIKV infection was determined by examining viral titers, synthesis of viral proteins, and replication of positive and negative strands of viral genome. Among nonhuman cell lines, only Vero cells were efficiently infected by ZIKV. Among human cell lines, all were permissive to ZIKV infection. However, 293T and HeLa cells showed differential susceptibility towards to African strains. In 293T cells, NS1 protein was expressed at high level by African strains but was almost not expressed by Asian strains though there was no obvious difference in viral genome replication, suggesting that the differential susceptibility might be controlled at the stage of viral protein translation. This study provides comprehensive results of the permissiveness of different cell types to both African and Asian ZIKV strains, which might help clarify their different pathogenesis.

Introduction

Zika virus (ZIKV) was first isolated in Uganda in 1947 ^1,2. ZIKV outbreaks have recently been described in Yap Island (2007) ³, French Polynesia (2013) ⁴ and Brazil (2015) ⁵, reaching many countries in the Americas by 2016 ^6–14. ZIKV is an arbovirus, belonging to the Flaviviridae family. It is transmitted by Aedes ssp, mainly Aedes aegypti and Aedes albopictus ^15,16. Initially, ZIKV infection was only associated with mild clinical symptoms such as cutaneous rash, slight fever and arthralgia ¹⁷. However, it has recently been associated with neurological disorders such as Guillain-Barré syndrome in adults ^18–23 and microcephaly in newborns ^24–26. In February, 2016, the World Health Organization (WHO) declared a Public Health Emergency of International concern due to ZIKV infection ²⁷, which ended in November 2016. As of March, 2017, 84 countries and territories had reported the presence of local ZIKV infection ²⁸.

The ZIKV genome is composed of a 5′-capped untranslated region (UTR), one open reading frame (ORF), and a 3’-UTR ²⁹. The ORF encodes a 3,419 amino acids (aa) long polypeptide, which is processed into three structural proteins including capsid (C), precursor of membrane (prM) and envelope (E), and seven nonstructural proteins, NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5 ²⁹. Numerous cell surface molecules have been described to mediate ZIKV attachment and entry, including Dendritic cell-specific intercellular adhesion molecule-3-grabbing non-integrin (DC-SIGN), AXL, Tyro3, and T-cell immunoglobin and mucin domain (TIM) ³⁰. However, it remains unclear which receptor is essential for ZIKV attachment and entry ^31–34.

ZIKV lineage is divided into African (West and East) and Asian strains. The first strain isolated in French Polynesia (H/PF/2013) was highly similar to the ones isolated in Cambodia in 2010 ³⁵ and Yap Island in 2007 ^36,37. Additionally, those isolated in Brazil in 2015 were very similar to the one from French Polynesia and were also classified as Asian strains³⁸. Although the strains isolated in the Americas belong to the same lineage as the ones isolated in the Pacific Islands, the clinical consequences of ZIKV infection were completely different. During the French Polynesia outbreak, there was an increase in cases of Guillain-Barré syndrome in adults, while during the Brazilian outbreak, there was an inexplicable increase of microcephaly in newborns ^23,24,38. Following ZIKV spread to other countries in the Americas, cases of microcephaly associated with ZIKV infection were also described ^39–42 but not in a profuse number as in Brazil, indicating that other co-factors might play a role in the pathogenesis of ZIKV infection.

Since the most intriguing fact related to ZIKV infection is the association of some viral strains with neurological diseases, several animal models have been developed to address these clinical findings. However, when high doses of virus were used in the experiments, both lineages caused neurological diseases ^43–46. The direct comparison of African and Asian strains is complicated by the fact that African strains have been passaged for hundreds of time since they were isolated 70 years ago, which could have introduced modifications in the virus genome. This might explain the new pathological manifestations in the animal models, which were not previously observed in humans at the time of the initial virus isolation. Up to now, there has been no systematic evaluation of ZIKV strains isolated from different outbreaks. In this study, we systematically examined the permissiveness of 4 ZIKV strains in 14 cell lines/types, of which 10 are human and 4 are non-human. The results provide important insights into the tropism of the ZIKV strains in different cell types.

Results

Permissiveness of ZIKV infection among human cell lines

Most of the human cell lines were efficiently infected by ZIKV, producing viral titers at the range of 6–7.5 log₁₀ PFU/mL (Fig. 1A). THP-1 cells have previously been reported as non-permissive to ZIKV infection ^30,47,48. Before infection, these cells were treated with phorbol 12-myristate 13-acetate (PMA) for 72 h to induce differentiation. Unexpectedly, we observed viral replication in THP-1 cells producing viral titers up to 5 log₁₀ PFU/mL, which was one of the lowest among all the human cells tested (Fig. 1A). Human hepatocellular carcinoma cells Huh7 produced the highest viral titers (> 7 log₁₀ PFU/mL) at 72 hpi for all the ZIKV strains examined (Fig. 1A). The most interesting results were observed with human embryonic kidney cells 293T. A few studies have shown that HEK293 and 293T cells cannot be infected by ZIKV while others have shown that they can be infected but with low efficiencies ^30,47,48. Our results showed that 293T cells were permissive to ZIKV infection but the replication efficiencies between African and Asian ZIKV strains differed in these cells. The Asian strains generated 4.0–4.5 log₁₀ PFU/mL viral titers while the African strains produced 5.5–6.0 log₁₀ PFU/mL viral titers (Fig. 1A). We repeated the experiments with another source of 293T cells and obtained similar results (results not shown). These results indicate that 293T cells are permissive to African strains. However, the infection efficiencies of the Asian strains are lower than those of African strains, and are more comparable to or lower than in THP-1 cells. Infection efficiencies of Asian strains were also slightly lower than African strains in Huh7 cells (6.0–6.5 vs 7.0–7.5 log₁₀ PFU/mL) and human cervical adenocarcinoma HeLa cells (5.0–5.5 vs 5.5–6.0 log₁₀ PFU/mL) (Fig. 1A). On the other hand, infection efficiencies of African strains were slightly lower than Asian strains in astroglial cells SVG p12, human lung carcinoma cells A549 and human renal carcinoma cells Caki (Fig. 1A).

Figure 1. Kinetics of production of infectious virions in different cell types/lines infected by African and Asian ZIKV strains.

All cells except 293T and THP-1 were infected with ZIKV at a MOI of 1.0 for 1 h with gentle shaking every 15 min. The cultures were then washed three times with PBS, replaced with new medium and cultured for 3, 24, 48, 72 h. Cells were then collected for RNA and protein examination, and supernatants were collected for viral titration. 293T cells were infected with the same procedures for 2 h except without shaking. THP-1 cells were treated with 0.125 μg/mL of phorbol 12-myristate 13-acetate (PMA) for 72 h, then infected the same procedures as 293T cells. A. Kinetics of production of virions in different human cell types/lines infected by IbH30656, MR766, PRVABC59, or H/FP/2013 ZIKV strains. Supernatants were collected at indicated time-points and plaque-assay was performed using Vero cells. B. Kinetics of production of virions by IbH30656, MR766, PRVABC59, or H/FP/2013 ZIKV strains in different nonhuman cell types/lines. Viral titers were examined as described in A.

Vero is the only non-human cell line used in this study that can be efficiently infected by ZIKV

Among the four types of non-human cells (African green monkey kidney cells Vero, primary rat mesenchymal stem cells MM, mouse embryonic fibroblast cells MEF and murine macrophage cells Raw 264.7) infected with ZIKV, only Vero cells produced high viral titers, reaching up to 7 log₁₀ PFU/mL at 48 hpi (Fig. 1B). Vero cells are defective in interferon (IFN) signaling ⁴⁹. Although the influence of IFN signaling on ZIKV infection remains unclear, numerous mouse studies showed that only animals that were deficient in IFN-α/β responses could be efficiently infected ^50,51. MM, MEF and Raw cells were not permissive to Asian strains. Raw cells were also not permissive to African strains but MEF and MM cells supported low levels of replication of African strains, producing viral tiers of 4 and 3 log₁₀ PFU/mL at 48 hpi, respectively, indicating limited viral replication (Fig. 1B).

Differential expression of viral protein in human cells infected by African or Asian ZIKV strains

To confirm the results of viral titers (Fig. 1), we examined the expression of ZIKV proteins. ZIKV NS1 protein was highly expressed in 293T cells infected by African strains IbH30656 and MR766 but was almost undetectable in 293T cells infected by Asian strains PRVABC59 and H/FP/2013 at 48 and 72 hpi (Fig. 2A). These findings were consistent with the results of viral titers (Fig. 1A). Since the permissiveness of 293T cells to ZIKV infection is unclear in the literature, the differential infection efficiencies of 293T cells by different ZIKV strains might explain the reported contradictory findings ^30,47,48.

Figure 2. Kinetics of viral protein expression in different cell types/lines infected by African and Asian ZIKV strains.

Different cell types/lines were infected with ZIKV strains as described in Fig. 1. A. Kinetics of viral protein expression in different human cell types/lines infected by IbH30656, MR766, PRVABC59, and H/FP/2013 ZIKV strains. Cells were lysed at indicated time points (3, 24, 48, and 72 hpi). Non-structural (NS1) and/or structural (envelope) viral proteins were detected by Western Blotting. B. Kinetics of viral protein expression in different nonhuman cell types/lines infected by IbH30656, MR766, PRVABC59, and H/FP/2013 ZIKV strains. Viral proteins were examined as described in A.

Results of viral titers indicated that HeLa and Huh7 cells supported viral replication of African strains slightly better than that of Asian strains (Fig. 1A). The expression levels of NS1 protein were stronger in HeLa cells infected with African strains than those infected with Asian strains (Fig. 2A). In contrast, while viral replication in astroglial, A549 and Caki cells were slightly more robust for Asian strains than African strains (Fig. 1A), there was no consistent variations of NS1 protein expression between the two lineages (Fig. 2A). While A549 cells can be infected by all ZIKV strains, most of the cells had detached from the plate at 72 hpi, which explained the low β-tubulin protein level detected (Fig. 2A) and the cytopathic effect observed in these cells (results not shown). For THP1 cells, we were unable to detect the expression of NS1 protein though we detected the expression of ZIKV envelope protein, confirming the results that THP-1 cells were not efficiently infected by ZIKV (Fig. 1A). NS1 protein was detected in human mesenchymal stem cells (MSC), fibroblast and primary human umbilical vein endothelial cells (HUVEC) infected by all ZIKV strains (Fig. 2A).

ZIKV protein expression in nonhuman cell lines

Vero cells infected with African or Asian ZIKV strains were the only type of nonhuman cells that had NS1 protein expression. Vero and A549 cells (Fig. 2A) were the only cells in this study that had detectable NS1 protein at as early as 24 hpi (Fig. 2B). MM, MEF or Raw cells did not have any expression of ZIKV NS1 or envelope proteins (Fig. 2B).

Differential viral genome replication in human cells infected by African or Asian ZIKV strains

We examined the levels of positive and negative viral genomes in cells infected by different ZIKV strains. Interestingly, HeLa cells had different replication kinetics between African and Asian ZIKV strains albeit we had only observed minor differences in viral titers and protein expression levels between these strains. The levels of positive and negative viral genomes in HeLa cells infected with IbH30656 and MR766 increased more than 1×10⁴ fold at 72 hpi compared to those at 3 hpi (Fig. 3A). In contrast, the levels of positive viral genome in HeLa cells infected with PRVABC59 and H/FP/2013 had less than 1×10² fold change while those of negative viral genome had almost no change at 72 hpi compared to those at 3 hpi (Fig. 3A). The differential viral genome levels were also observed in astroglial and 293T cells though they were less dramatic with less than 1×10² fold differences between African and Asian strains (Fig. 3A). High levels of positive and negative viral genomes were detected in other human cells including MSC, fibroblasts, A549, Caki, HUVEC and Huh7 cells infected by all ZIKV strains, reaching at 1×10³ fold change at 72 hpi compared those at 3 hpi without much difference among the strains. THP-1 cells exhibited less than 1×10² fold changes among all the ZIKV strains tested, indicating a low level of viral genome replication, which was consistent with the results of viral titers (Fig. 1A).

Figure 3. Kinetics of replication of positive and negative strands of viral genomes in different cell types/lines infected by African and Asian ZIKV strains.

Different cell types/lines were infected with ZIKV strains as described in Fig. 1. A. Kinetics of replication of positive and negative strands of viral genomes in different human cell types/lines infected by IbH30656, MR766, PRVABC59, and H/FP/2013 ZIKV strains. Cells were lysed at indicated time points. Both positive- and negative-strand viral genome were measured by RT-qPCR. B. Kinetics of replication of positive and negative strands of viral genomes in different nonhuman cell types/lines infected by IbH30656, MR766, PRVABC59, and H/FP/2013 ZIKV strains. Levels of positive and negative strands of viral genomes were examined as described in A.

Of interest, we found that the levels of negative viral genome lagged behind those of the positive viral genome in 293T and HUVEC cells infected by African strain IbH30656, in MSC, fibroblast, A549, Caki and HeLa cells infected by Asian strain PRVABC59, and in MSC, fibroblast, HUVEC, Caki and HeLa cells infected by Asian strain H/FP/2013 (Fig. 3A). However, these differences did not affect the levels of positive viral genome.

ZIKV genome replication in nonhuman cell lines

Vero cells had the highest fold increase of positive and negative viral genomes compared to other non-human cells reaching around 1×10³ fold at 48 hpi compared those at 3 hpi for all ZIKV strains (Fig. 3B). No evidence of ZIKV genome replication was detected in MEF cells (Fig. 3B), confirming that ZIKV did not replicate in these cells. Raw cells had increases of positive and negative viral genomes for up to 1×10² fold change at 72 hpi compared to those at 3 hpi. However, since we failed to detect any infectious virions (Fig. 1B) and viral protein (Fig. 2B), we concluded that Raw cells were not permissive to ZIKV infection. Only slight increases of viral genomes (<1×10² fold) were detected in MM cells infected by both African strains and H/FP/3013 Asian strain, and no evidence of replication of viral genome was detected with Asian strain PRVABC59 (Fig. 3B). Since we only detected low levels of viral titers and no viral protein in MM cells (Fig. 1B and2B), we concluded that ZIKV infection was inefficient in these cells.

Discussion

In this study, we examined the susceptibility of 10 human and 4 nonhuman cell types/lines to infection of different ZIKV strains. We found that all the 10 human cell types/lines are permissive to ZIKV infection. Of these cells, Huh7 has the highest infection efficiencies, followed by astroglial, fibroblast and Caki cells. A549, MSC, 293T, HeLa and HUVEC cells can also be infected but with lower efficiencies. The infection efficiency in THP-1 cells is the lowest among all the human cell types tested. It is worth noted that infection of MSC cells has not been reported before. Our results showed that ZIKV can efficiently infect precursor/stem cells, and hence assess hematopoietic cells. Whether 293T and THP-1 cells can be infected by ZIKV is controversial ^30,47,48. We showed that both 293T and THP-1 cells could be infected by ZIKV though the efficiencies of THP-1 cells were low. Overall, our results have shown the variations of tropisms of different cells to ZIKV infection.

Among the nonhuman cells, Vero is the only cell type that can be efficiently infected. Low infection efficiencies were detected with MM and MEF cells while no infection was detected for Raw cells. Since Vero cells are defective in IFN signaling, it is possible that the susceptibility of nonhuman cells is determined by IFN signaling. Numerous animal models have been developed for ZIKV infection ^50–53, and neurological diseases were only observed in animals defective in IFN response, suggesting the importance of IFN signaling in controlling ZIKV infection and ZIKV-related diseases. Recently, it was shown that fetal demise and growth restriction after ZIKV infection was caused by type I interferon response ⁵⁴. Hence, IFN signaling might have a more complex influence in ZIKV-associated diseases. On the other hand, most human cell types examined in this study support ZIKV replication and have full functional IFN signaling, suggesting that IFN signaling alone is insufficient to control ZIKV infection in human cells. It is possible that another innate immune response pathway also regulates ZIKV infection, and that this pathway is evaded by ZIKV in human but not nonhuman cells.

We have found that ZIKV African strains in general have higher infection efficiencies than Asian strains. It has been speculated that the more severe neurological diseases associated with the recent ZIKV outbreaks (which were not observed in the past) could be due to the differences in the strain virulence. In this context, higher replication efficiencies of the African strains in human cells might offer some explanations. In this study, we have also observed faster cytopathic effect of African strains and larger plaque formation for Asian strains during virus production. Similar results were observed in another study, which described that an African strain had higher infection and dissemination rates in Aedes aegypti mosquitoes than an Asian strain ⁵⁵. Additionally, infection by African strain resulted in fatal disease in Ifnar1^−/− mice while infection by an Asian strain had a lower mortality rate ⁵⁵. Interestingly, a recent isolated Asian strain had a guanine-adenine nucleotide mutation in the membrane protein resulting in smaller plaque sizes and less virulence than other Asian strains ⁵⁶. Another study showed that African and Asian strains had similar replication rates in Vero cells but African strain had a faster replication rate in C6/36 cells and low infection rate in Aedes aegypti mosquitoes. Although both strains could infect chick embryos, infection by the African strain resulted in a higher mortality rate ⁵⁷. Thus, virus virulence, infection efficiency and host immune response could all determine the outcomes of ZIKV-associated diseases.

Our results of ZIKV protein expression, and replication of the positive and negative strands viral genomes in general support those of viral titers. However, there are inconsistencies among some cell types. While we observed that African strains had higher viral titers and higher levels of viral genome replication than Asian strains had, it was still surprising to observe that 293T cells infected by Asian strains had almost no NS1 protein expression despite detection of high levels of NS1 protein expression in cells infected by the African strains. This might suggest that viral protein translation could be an important block for the Asian strains. Of interest, HeLa cells infected by Asian strains had much lower levels of viral genome replication than those infected by African strains. Similar results were also observed for astroglial, and Huh7 cells.

Some studies focused on in vitro experiments to understand the basic biology of virus infection and explore what cell types might be susceptible and permissive to ZIKV infection. The susceptibilities of human and non-human cell lines to PRVABC59 and 976 Uganda ZIKV strains have been evaluated ⁴⁸. 293T cells were efficiently infected by both strains with viral genomes detected at 6.5 to 7.5 log₁₀ copies/mL; however, the protein expression was detected in 1% and 20% of the cells for PRVABC59 and 976 Uganda strains, respectively ⁴⁸. In our study, we have shown a drastic difference in viral titers and virus protein expression in 293T cells infected with African or Asian strains, indicating that 293T cells are more permissive to African than Asian strains (Fig. 1A and2A). Together, these results indicate higher susceptibilities of 293T and HEK293 to African than Asian ZIKV strains. In agreement with these results, Hamel et al (2015) were not able to detect viral envelope protein at 24 hpi in HEK293 cells infected with PF-25013–18 strain isolated in French Polynesia ³⁰. Similar to our results, another study also showed that A549 cells were susceptible to ZIKV infection ³⁰.

Chan et al. (2016) reported efficient infection by both Asian and African strains in HeLa cells with viral genomes detected at 8 to 10 log₁₀ copies/mL but drastic differences in protein expression (1% vs 20% positive cells for PRVABC59 and 976 Uganda strains, respectively) ⁴⁸. Our results showed that there were only minor differences in viral titers in HeLa cells infected by African or Asian strains (Fig. 1A). However, African strains IbH30656 and MR766 expressed higher levels of NS1 protein than those of Asian strains at 72 hpi (Fig. 2A). Surprisingly, the efficiencies of viral genome replication were more than 100 times higher for African than Asian strains at 72 hpi with the Asian strains at almost undetectable levels (Fig. 3A). It was unclear why such differences in viral protein expression and viral genome replication did not translate into viral titers.

Huh7 cells were efficient infected by PRVABC59 and 976 Uganda strains ZIKV strains with viral genomes detected at 9 and 10 log₁₀ copies/mL, respectively ⁴⁸. Viral protein expression was detected in 60% and 80% of cells infected by PRVABC59 and 976 Uganda, respectively ⁴⁸. In our study, Huh7 cells had the highest viral titers among all types of cells tested (> 7 log₁₀ PFU/mL). Chan et al. (2016) also described that Vero cells were efficiently infected by Asian and African strains, with 50% and 80% of cells expressing viral protein, respectively ⁴⁸. On the other hand, our results for THP-1 showed low infection rate of THP-1 cells as evidenced by the detection of low viral titers (Fig. 1A), and low levels of viral protein (Fig. 2A) and viral genome replication for all the strains examined (Fig. 3A). Similar results in THP-1 cells were also observed by Chan et al. (2016) ⁴⁸. A study by Hou et al. (2017) also showed that THP-1 and MEF were not permissive to ZIKV infection while Vero and MRC-5 cells had up to 70% of infected cells, and A549 cells had 10–30% infected cells ⁴⁷.

There is no explanation for the differential susceptibilities of different cell types to different ZIKV strains. The Golgi protease furin is essential for the cleavage of the pr/M junction ⁵⁸ and ZIKV infection. Since all the cell lines used in our study have furin, it should not determine the susceptibility of these cells to ZIKV infection. On the other hand, the presence and absence of viral receptors could determine ZIKV infection. Although AXL has been described as the main receptor for ZIKV infection, some studies have shown that AXL is not essential ^{31–34,59,60}. There is no correlation of the degree of susceptibilities of cells to infection with the levels of furin or receptors for attachment and entry. In our study, we have shown that there is a general delay in the infection kinetics for Asian strains when compared to the African ones with most of the cell lines examined.

The current burning question concerning ZIKV infection is why some strains are associated with microcephaly in newborns, some with Guillain-Barré syndrome in adults, and still others without any association with clinical symptom. It is possible that, co-factors such as environmental, genetics or co-infections with other viruses might play a role in the clinical consequences of ZIKV infection. Further studies with multiple cell types and different virus strains are essential to help unveil the mechanisms involved in the different pathologies observed after ZIKV infection.

Materials and Methods

Antibodies

Peptides of ZIKV structural envelope protein (NDTGHETDENRAK) and non-structural NS1 protein (EAWRDRYKYHPDSPRR) were designed based on Asian ZIKV strain H/PF/2013 (Genbank# KJ776791.2), synthesized and primary antibodies against these proteins were generated by Chempeptide Limited, Shanghai, China. As previously described both antibodies can recognize Asian and African strains⁶¹. Antibody to β-tubulin was purchased from Sigma-Aldrich (St. Louis, MO). Horseradish peroxidase (HRP)-conjugated goat anti-mouse and goat anti-rabbit antibodies were obtained from Santa Cruz Biotechnology (Dallas, TX).

Cell culture

Primary rat mesenchymal stem cells (MM), mouse embryonic fibroblast cells (MEF), murine macrophage cells (Raw 264.7), African green monkey kidney cells (Vero), human renal carcinoma cells (Caki), human cervical adenocarcinoma cells (HeLa), human hepatocellular carcinoma cells (Huh7), human astroglia (SVG p12, ATCC^®CRL-8621™), human lung fibroblasts, and human embryonic kidney cells (293T) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Hyclone, GE Healthcare, Chicago, IL) supplemented with 10% fetal bovine serum (FBS) (Sigma, Saint Louis, MO), 1 × penicillin-streptomycin (P/S) solution (Genesee Scientific, San Diego, CA). Human mesenchymal stem cells (MSC) from adipose tissue were cultured in MSC medium complemented with 5% FBS, 1% of MSC growth supplement and 1% of P/S solution (ScienCell, Carlsbad, CA). Human lung carcinoma cells (A549) and human monocytic cells (THP-1) were cultured with RPMI-1640 (Hyclone, GE Healthcare, Chicago, IL) supplemented with 10% FBS, and 1 × penicillin-streptomycin solution. Primary human umbilical vein endothelial cells (HUVEC) were cultured in VascuLife VEGF complete media (Lifeline cell technology, Frederick, MD). All cells were cultivated at 37°C with 5% CO₂.

Virus preparation

ZIKV African strains MR766 (Uganda, 1947) and IbH30656 (Nigeria, 1968), and Asian strains H/PF/2013 (French Polynesia, 2013) and PRVABC59 (Puerto Rico, 2015) were grown in Vero cells. To prepare ZIKV stocks, Vero cells were inoculated with ZIKV at 0.01 or 0.1 multiplicity of infection (MOI), and the supernatants were harvested at 48–96 hpi depending on the ZIKV strain. The titers of ZIKV stocks were determined by plaque assay with Vero cells. All the experiments used the same virus stocks for all the cell lines.

Virus infection

All cell lines were infected at 1 MOI 24 h after seeding. All cell lines except 293T and THP-1 were infected for 1 h with gentle shaking every 15 min. 293T cells were infected without shaking for 2 h. THP-1 cells were first treated with 0.125 μg/mL of phorbol 12-myristate 13-acetate (PMA) for 3 days, then infected with ZIKV for 2 h without shaking. After infection, cells were washed 3 times with PBS to remove unbound viruses, and the respective culture medium was added to each cell line.

Western-blot analysis

Total protein preparations extracted from cells infected with ZIKV for specific time-points were separated in sodium dodecyl sulfate-polyacrylamide gels, transferred to nitrocellulose membranes, and detected with antibodies. Specific signals were revealed with chemiluminescence substrates and recorded with a UVP MultiSpectral Imaging System (UVP LLC, Upland, CA).

Reverse transcription quantitative real-time PCR (RT-qPCR)

The expression levels of viral genes were analyzed by RT-qPCR. Briefly, total RNAs from ZIKV-infected cells were prepared with TRI reagent as recommended by the manufacturer (Sigma, Saint Louis, MO). The RNA was treated with RNase-free DNase (Thermo Fisher Scientific, Inc, Waltham, WA) and reverse transcribed using a Maxima Reverse Transcriptase system (Thermo Fisher Scientific). Forward primers were used to transcribe cDNA from (−) strand RNA, while reverse primers were used to transcribe cDNA from (+) strand RNA. All primers are indicated in table 1, including housekeeping genes. For each sample, a control without reverse transcriptase was included in parallel. Quantitative PCR (qPCR) analysis was performed on an Biorad CFX Connect Real-time System using Sso Advanced Universal SYBR Green Supermix (#172–5271, Biorad, Hercules, CA). The relative expression levels of target genes were normalized to the expression levels of the internal control genes yielding cycle threshold (2^−ΔΔCT) values. As an internal reference, α-tubulin was used for human cell lines while β-actin was used for murine and monkey cell lines (Table 1). All reactions were run in triplicates. Relative expression of normalized target genes at time-points at 24, 48 and 72hpi were normalized to 3hpi, which is the first time-point to have a detectable cycle threshold. This might reflect background of the assay and/or inoculum that remained attached to the cells after post-infection washing steps.

Table 1.

Primers used in this study

Primer target	Primer sequence (5’−3’)
IbH30656	Forward: CCA TAC GGC CAA CAA AGA GT Reverse: CAG CTC CTT CCA TAG CCA AG
MR766	Forward: AGA TTG AAG GGC GTG TCA Reverse: TCC ATC TGT CCC TGC ATA
H/PF/2013 and PRVABC59	Forward: TGG GTT GAT GTT GTC TTG GA Reverse: ATC TTA CCT CCG CCA TGT TG
Human α-tubulin	Forward: AGA TCA TTG ACC TCG TGT TGG A Reverse: ACC AGT TCC CCC ACC AAA G
Murine β-actin	Forward: GCA GGA GTA CGA TGA GTC CG Reverse: ACG CAG CTC AGT AAC AGT CC
African green monkey β-actin	Forward: GGC CAG GTC ATC ACC ATT Reverse: ATG TCC ACG TCA CAC TTC ATG

Plaque assay

Plaque assay was performed to determine the viral titer for the ZIKV stock. Supernatants collected at different time-points were analyzed to determine the kinetics of virion production. Briefly, 10-fold dilutions of virus inoculum were applied to Vero cells at 37 °C for 1 h with gentle shaking every 15 min to facilitate virus attachment to the cells. The cells were washed with PBS, and a mix of nutriment solution with agarose (Lonza Walkersville, MD) was added. The cells were maintained at 37 °C for 5 days, and then fixed with 10% formaldehyde (Sigma, Saint Louis, MO). After removal of the agarose layer, the cells were stained with 0.1% of crystal violet in water to reveal the plaques.