Zika virus disrupts steroidogenesis and impairs spermatogenesis by stalling the translation of CYP17A1 mRNA

Yang, Wei; Li, Hanyu; Wang, Shanshan; Huang, Rui; Zhang, Yifei; Guo, Moujian; Huang, Li; Li, Shihua; Yang, Ruirui; Zhao, Dingran; Xiong, Yuxin; Liu, Yifei; Huang, Mengjing; Hui, Lixia; Xiao, Wei; Wu, Ying

doi:10.1038/s41467-025-62044-x

Download PDF

Article
Open access
Published: 22 July 2025

Zika virus disrupts steroidogenesis and impairs spermatogenesis by stalling the translation of CYP17A1 mRNA

Nature Communications volume 16, Article number: 6756 (2025) Cite this article

3463 Accesses
16 Altmetric
Metrics details

Subjects

Abstract

ZIKV infection is associated with testicular damage and abnormal spermatogenesis. However, the molecular mechanisms underlying these pathogenic processes remain unclear. Here, we demonstrate that ZIKV disrupts Leydig cells’ ability to produce testosterone, leading to decreased sperm counts and motility. Specifically, the non-structural protein NS2A of ZIKV downregulates testosterone production by directly binding to mRNA of CYP17A1, a key enzyme in testosterone synthesis, thereby inhibiting its translation. Notably, the sole membrane-traversing segment and its flanking loops of NS2A are crucial for this interaction with CYP17A1 mRNA. Scanning mutagenesis studies within this sequence identified amino acid residues critical for NS2A binding and the suppression of CYP17A1 mRNA translation. Testicular inoculation of adeno-associated virus (AAV) delivering ZIKV-NS2A or its mutant showed that ZIKV-NS2A alone is sufficient to affect steroidogenesis and spermatogenesis in vivo. Moreover, a mutant virus generated by reverse genetics, containing a single amino acid mutation that abolishes NS2A’s binding to CYP17A1 mRNA, exhibited significantly lower inhibition of steroidogenesis and spermatogenesis compared to the wild-type virus in mouse models. These findings enhance our understanding of how ZIKV impacts male reproductive health and provide crucial insights for future preventive and therapeutic strategies.

Structural basis of Zika virus NS1 multimerization and human antibody recognition

Article Open access 25 April 2024

Single-cell RNA sequencing reveals the fragility of male spermatogenic cells to Zika virus-induced complement activation

Article Open access 29 April 2023

Antiviral activity of ouabain against a Brazilian Zika virus strain

Article Open access 23 July 2022

Introduction

Zika virus (ZIKV) is a mosquito-borne flavivirus that can cause severe neurological symptoms, including microcephaly in newborns and Guillain-Barre syndrome in adults^1,2. In addition, ZIKV infection adversely affects male reproductive health, observed both in ZIKV-infected patients and mouse models^3,4,5. The male reproductive system comprises the testes, epididymis, prostate gland, and seminal vesicles. The testes, primarily responsible for spermatogenesis and steroidogenesis, have been identified as a direct target of ZIKV. Indeed, some male patients with ZIKV exhibited symptoms of the reproductive system, such as oligospermia, hematospermia, and prostatitis^6,7. Moreover, testosterone levels in ZIKV-infected men also showed lower values⁸. Nevertheless, the mechanism by which ZIKV infection leads to impaired testosterone and sperm production remains elusive.

In the testicular tissue, various cells, including Germ cells (GCs), Sertoli cells (SCs), Leydig cells (LCs), and peritubular-myoid cells (MCs), are susceptible to ZIKV infection^4,9,10,11. Given that LCs serve as the primary site for the synthesis and secretion of male sex hormones, particularly testosterone, ZIKV infection of LCs may potentially be a crucial factor in reducing testosterone levels. Testosterone plays a pivotal role in sustaining spermatogenesis and fertility^12,13. Spermatogenesis occurs in the seminiferous tubules of the testes. The seminiferous tubules are composed of two major cell types: SCs and GCs. GCs are reproductive cells that develop into sperm, while SCs provide the nutritional and structural support to GCs, facilitating their maturation. Within the seminiferous tubules, only SCs have receptors for testosterone. GCs do not express androgen receptors (AR)^14,15. Consequently, testosterone produced by LCs diffuses into the seminiferous tubules and regulates the expression of proteins involved in spermatogenesis through its action on SCs^16,17, thus maintaining normal sperm production. Sperm then mature and are stored in the epididymis, where high testosterone levels are vital for supporting sperm maturation¹⁸. Thus, maintaining normal testosterone levels is crucial for the entire spermatogenic process. For instance, withdrawal of testosterone in the testis leads to a substantial decline in sperm production^19,20,21. Spermatogenesis is also regulated by the hypothalamic-pituitary-gonadal (HPG) axis. ZIKV infection in male suckling mice has been shown to cause long-term disruptions to the HPG axis²² and, ultimately, fertility. However, ZIKV-induced functional abnormalities in HPG axis have not been observed in adult mice or human patients. Thus, our hypothesis posits a strong correlation between ZIKV infection of LCs and the resultant decline in testosterone levels and sperm counts.

Testosterone production in LCs starts with cholesterol transfer into mitochondria, mediated by the steroidogenic acute regulatory (STAR) protein²³. Cholesterol is then converted to pregnenolone by the cholesterol side-chain cleavage enzyme (CYP11A1) in the mitochondria²⁴. Pregnenolone is then transported to the endoplasmic reticulum (ER) and converted sequentially to progesterone by 3-beta-hydroxysteroid dehydrogenase/Δ5-Δ4 isomerase (3β-HSD), to 17-hydroxyprogesterone, and then to androstenedione by 17α-hydroxylase/C17–20 lyase (CYP17A1), and finally to testosterone by 17-beta-hydroxysteroid dehydrogenase (17β-HSD)^25,26. The synthesis of androgens, particularly testosterone, is a crucial aspect of testicular metabolism and is influenced by various factors²⁷. ZIKV-induced low testosterone levels reveal metabolic abnormalities in ZIKV-infected mouse testes. Notably, virus-induced metabolic changes are often critical for viral survival and host pathogenesis²⁸. For example, ZIKV infection can lead to abnormalities in NAD⁺ metabolism in the brains of fetal mice, ultimately resulting in microcephaly²⁹. To investigate the molecular mechanisms of ZIKV-induced testosterone reduction, we employed unbiased metabolomics and quantitative proteomics analyses in the testes of both ZIKV-infected and Mock-infected mice, aiming to allow for a multi-dimensional understanding of how ZIKV affects testicular metabolism, facilitating the identification of critical pathways and molecular targets potentially involved in the disruption of testosterone production and steroidogenesis. The proteomics profiling showed a significant downregulation of CYP17A1, a finding supported by the metabolomics analysis. CYP17A1, a membrane-bound dual-function monooxygenase, plays a critical role in the synthesis of many human steroid hormones³⁰. Functional deficits in CYP17A1 can disrupt hormone synthesis, including glucocorticoids and sex hormones, leading to underdeveloped male physical characteristics and reduced fertility^31,32,33. Clinically, CYP17A1 inhibitors, like abiraterone acetate, are utilized to lower testosterone levels in patients with advanced prostate cancer^34,35,36.

The ZIKV genome encodes a polyprotein that is cleaved into three structural proteins (C, prM/M, E) and seven nonstructural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5). The structural proteins form viral particles. The nonstructural proteins play critical roles in various stages of the virus life cycle, including replication, assembly, and evasion of host immune responses. NS1 is involved in the formation of the viral replication complex and immune evasion³⁷. NS2A is crucial for the formation of replication vesicles, coordinating membrane rearrangements for RNA synthesis, and assembling viral particles³⁸. It also suppresses interferon signaling, enhancing immune evasion. NS2B serves as a cofactor for NS3, forming a protease complex that facilitates the cleavage of the polyprotein precursor³⁹. NS3 functions as both a protease and helicase, facilitating protein cleavage and RNA replication³⁹. NS4A and NS4B collaborate to regulate the formation of replication vesicles and inhibit the host interferon response³⁷. NS5, as an RNA-dependent RNA polymerase and methyltransferase, is responsible for viral RNA replication and stabilization, while also suppressing host immune responses⁴⁰.

Here, we show that ZIKV infection impairs CYP17A1 mRNA translation via its non-structural protein NS2A (ZIKV-NS2A) binding, reducing CYP17A1 expression and consequently testosterone production. Further experiments confirmed the direct impact of ZIKV-NS2A expression on reduced testosterone and abnormal sperm production in mice. Our results reveal a previously unidentified pathogenic mechanism and potential interventions for ZIKV-related symptoms.

Results

Metabolomics profiling reveals disruption in steroidogenesis in the testes of ZIKV-infected mice

The testes are integral components of the male reproductive system, responsible primarily for spermatogenesis and hormone synthesis. ZIKV infection leads to disrupted testicular structure⁵, leakage of blood-testis barrier⁴¹, diminished testosterone secretion, and culminates⁴ in male infertility in mice, strongly suggesting a disturbance on steroid homeostasis. In order to investigate the impact of ZIKV infection on steroidogenesis in testes, we performed untargeted metabolomics analyses on mouse testes from Mock- and ZIKV-infected A129 mice (IFNα/β receptor-deficient), a well-established model for studying ZIKV pathogenesis⁵. The infected mice displayed evident viremia two days post-ZIKV infection, followed by gradual clearance of the virus (Fig. 1a). The viral load in testes increased with the duration of infection (Fig. 1b), which is consistent with previous reports⁴¹. Principal component analysis (PCA) indicated well-defined clustering of individual samples within each group, signifying low variation among replicates, and revealed distinct separation of metabolite profiles between the Mock- and ZIKV-infected groups at 10 dpi (Fig. 1c), suggesting significant metabolic changes. OPLS-DA further confirmed the distinct separation between the two groups (Supplementary Fig. 1a), with Q2 and R2 values showing no overfitting. The metabolomics analysis identified a total of 693 metabolites, of which 335 were differentially represented and mainly enriched in 35 metabolic pathways. Notably, steroid hormone biosynthesis is shown as one of the top 10 enriched pathways among the altered metabolites (Fig. 1d).

**Fig. 1: Metabolomic profiling reveals disruption in steroidogenesis in the testes of ZIKV-infected mice.**

To gain a deep insight into identified metabolites involved in the steroid hormone biosynthesis, we examined the differential metabolites (FDR < 0.05) and identified several prominently upregulated and downregulated metabolites (Fig. 1e). Of note, a substantial decrease in testosterone and androstenedione levels, accompanied by elevated levels of their common precursor, progesterone, was observed. All three metabolites are implicated in testosterone production in Leydig cells (Fig. 1f), indicating a potential blockage in testosterone biosynthesis at intersections downstream of progesterone due to ZIKV infection.

We next performed enzyme-linked immunosorbent assays (ELISA) to quantify testosterone and progesterone levels in both mouse serum and testes. The results revealed a significant decrease in testosterone levels in both serum and testes following ZIKV infection (Fig. 1g, h), with an accumulation of progesterone (Fig. 1i, j), aligning with the observed metabolic profiles. Additionally, we observed a significant reduction in sperm counts and motility in mice after ZIKV infection (Fig. 1k, l). Histological analysis with hematoxylin and eosin (H&E) staining of the epididymis revealed both a decrease in sperm count and structural abnormalities, such as constriction of the epididymal lumen and thickening of inter-luminal tissue (Fig. 1m). Microscopic analysis of sperms further confirmed these observations (Fig. 1n). Together, these data indicate that ZIKV infection contributes to dysfunction in steroidogenesis and spermatogenesis in mouse testes, by hindering testosterone biosynthesis at reactions downstream from progesterone.

ZIKV infection decreases CYP17A1 at the protein level

To elucidate the molecular mechanisms by which ZIKV influences testosterone biosynthesis, we conducted proteomic profiling on mouse testes infected with ZIKV. The results revealed that, after 6 days of infection, 400 proteins displayed significant differences between ZIKV-infected and Mock-infected groups [ZIKV (6 dpi) vs Mock]. Following 10 days of infection, 647 proteins exhibited significant differences compared against the Mock-infected group [ZIKV (10 dpi) vs Mock]. A total of 239 proteins were jointly enriched in both infection groups (Fig. 2a). Through gene-annotation-independent analysis of the differentially expressed proteins, we identified a series of proteins associated with cellular metabolism and reproductive development (Fig. 2b). Hierarchical clustering analysis of these proteins revealed a significant downregulation of the key enzyme CYP17A1, integral to testosterone biosynthesis, in both ZIKV (6 dpi) and ZIKV (10 dpi) groups (Fig. 2c). To examine whether ZIKV infection leads to a downregulation of CYP17A1 expression, we confirmed a decrease in the protein abundance of CYP17A1 in the testes following ZIKV infection by immunofluorescence (Fig. 2d) and Western blotting (Fig. 2e), which was consistent with the results obtained from mass spectrometry analysis, despite the differences in fold change at 6 dpi and 10 dpi. However, there was no change in the mRNA levels of CYP17A1 (Fig. 2f). Other proteins related to testosterone synthesis, such as CYP11A1, 3β-HSD, and STAR, showed no significant differences at both the protein and mRNA levels (Fig. 2e, g and Supplementary Fig. 1b–d). Additionally, we generated a stable A549 cell line expressing CYP17A1 (CYP17A1-A549 cells), where CYP17A1 is strongly expressed in ZIKV-uninfected cells (short white arrows), but its expression is nearly absent in ZIKV-infected cells (long white arrows), demonstrating the loss of stable transfected CYP17A1 expression in ZIKV-infected cells (Fig. 2h).

Leydig cells are the only testicular cell type that can express CYP17A1 and synthesize testosterone⁴². We isolated and cultured primary mouse Leydig cells (mLCs) in vitro to assess the impact of ZIKV infection. The endogenous marker protein CYP11A1 exhibited well expression in mLCs (Supplementary Fig. 1e). Additionally, we analyzed the levels of testosterone and progesterone in the culture supernatant of mLCs at different time points and observed that the secretion levels of these hormones peaked at 24 h after isolation and subsequently declined over time (Supplementary Fig. 1f, g). We found that ZIKV efficiently infected mLCs without significantly compromising their viability, even at high viral infection doses (Supplementary Fig. 1h–k). ZIKV infection in mLCs led to a time- and dose-dependent reduction in CYP17A1 protein expression (Fig. 2i, j), with no significant changes in mRNA expression levels (Fig. 2k). Other testosterone synthesis-related proteins such as STAR, CYP11A1, and 3β-HSD showed no significant differences in both protein and mRNA levels (Fig. 2i, j and Supplementary Fig. 1l–n). We also performed ELISA assays to quantify testosterone and progesterone levels in the supernatant of ZIKV-infected mLCs. The results revealed a significant decrease in testosterone levels, with an accumulation of progesterone (Supplementary Fig. 1o, p). In summary, these findings demonstrate that ZIKV infection downregulates the protein expression levels of endogenous CYP17A1 in both cells and testes.

ZIKV-NS2A inhibits the production of CYP17A1 protein

To further investigate the mechanism behind the reduction of CYP17A1 production due to ZIKV infection, we individually co-transfected 10 ZIKV-encoded proteins with human CYP17A1 (hCYP17A1) or mouse CYP17A1 (mCYP17A1) into HEK293T cells. NS2A, but not other viral proteins, disrupted hCYP17A1/mCYP17A1 expression (Fig. 3a, b) and this reduction in CYP17A1 levels was not attributable to cell death, as NS2A showed no significant cytotoxicity in HEK293T cells under the conditions tested (Supplementary Fig. 2a). Similarly, NS2A exhibited no substantial cytotoxic effects on A549 or mLCs (Supplementary Fig. 2b, c). The expression of hCYP17A1/mCYP17A1 was inhibited by NS2A in a dose-dependent manner (Fig. 3c, d). We also verified this phenomenon in CYP17A1-A549 cells (Fig. 3e), which was in line with immunofluorescence results indicating the loss of CYP17A1 expression in NS2A-positive cells (Fig. 3f). Furthermore, NS2A decreased the endogenous CYP17A1 protein in mLCs in a dose-dependent manner (Fig. 3g), with no impact on CYP17A1 mRNA levels (Fig. 3h). Notably, NS2A had no effect on the protein and mRNA levels of other testosterone synthesis-related proteins (Fig. 3g and Supplementary Fig. 2d–f). To assess specificity, we evaluated the impact of different flavivirus NS2As on CYP17A1 expression, including West Nile virus (WNV), Japanese Encephalitis virus (JEV), and Dengue virus (DENV). Only ZIKV-NS2A showed the most pronounced effectiveness on CYP17A1 expression (Supplementary Fig. 2g). Additionally, we selected the endoplasmic reticulum protein Calnexin for validation and found that neither ZIKV infection nor ZIKV-NS2A overexpression had any impact on Calnexin expression levels (Supplementary Fig. 2h–k). In summary, our results demonstrate that ZIKV-NS2A induces a downregulation of CYP17A1 expression.

ZIKV-NS2A sequestrates CYP17A1 mRNA and reduces its translation

To investigate if the decrease in CYP17A1 protein levels induced by ZIKV-NS2A is attributed to specific protein degradation pathways, we co-transfected CYP17A1 with ZIKV-NS2A in the presence of distinct pathway inhibitors. However, the inhibitors, including the autophagy inhibitor bafilomycin A1 (Baf-A1) and Chloroquine (CQ), the proteasome inhibitor MG-132, and the caspase inhibitor Z-VAD-FMK, failed to prevent the ZIKV-NS2A-induced reduction in CYP17A1 levels (Supplementary Fig. 3a, b). Additionally, we conducted protein stability assays in HEK293T cells expressing CYP17A1 with or without ZIKV-NS2A. These cells were treated with cycloheximide (CHX), a translation inhibitor, for up to 3 or 6 h. The results reinforced that ZIKV-NS2A had no effect on CYP17A1 at the post-translational level (Fig. 4a and Supplementary Fig. 3c), strongly indicating that ZIKV-NS2A may exert its inhibitory function during the translation of CYP17A1 mRNA, given ZIKV-NS2A had no effect on CYP17A1 mRNA levels. To further explore the impact of ZIKV-NS2A on the translation of CYP17A1 mRNA, we transfected ZIKV-NS2A into HEK293T cells, followed by transfection with in vitro synthesized CYP17A1 mRNA. The results showed a downregulation of CYP17A1 expression in the presence of ZIKV-NS2A, further supporting its suppressive effect on CYP17A1 mRNA translation (Fig. 4b).

**Fig. 4: ZIKV-NS2A interacts with CYP17A1 mRNA and reduces its translation.**

Previously, it has been reported that ZIKV-NS2A can bind a 3’-UTR stem-loop in the viral genome and present viral RNA to the structural proteins for virion assembly³⁸. This led us to hypothesize that ZIKV-NS2A might also bind to CYP17A1 mRNA, thereby affecting its translation. To validate our hypothesis, we investigated the potential binding of ZIKV-NS2A to CYP17A1 mRNA using an RNA immunoprecipitation (RIP) assay. However, one major challenge in studying flavivirus NS2A is the lack of specific and reliable antibodies against this protein. Therefore, we used a hemagglutinin (HA)-tagged NS2A ZIKV (HA-NS2A ZIKV) in the context of an infectious cDNA clone of Cambodian strain FSS13025³⁸. We performed RIP assays by HA, viral E protein, or cellular protein claudin-2 antibodies to precipitate cell lysates of HA-NS2A ZIKV-infected CYP17A1-A549 cells (Supplementary Fig. 3d). Quantitative RT-PCR was used to measure the amount of CYP17A1 mRNA co-precipitated with HA-NS2A, viral E or claudin-2, and GAPDH mRNA was quantified as a control from the precipitated samples. The results showed approximately a 2.5-fold enrichment of CYP17A1 mRNA in the HA-NS2A-precipitated samples, while no enrichment was observed for cellular GAPDH mRNA. Viral E or host claudin-2 protein did not enrich CYP17A1 or GAPDH mRNA (Fig. 4c), suggesting a specific interaction between ZIKV-NS2A and CYP17A1 mRNA in ZIKV-infected cells. The same phenomenon was observed in HEK293T cells co-transfected with plasmids encoding CYP17A1-HA and NS2A-Flag (Supplementary Fig. 3e). Additionally, we examined the affinity between ZIKV-NS2A and CYP17A1 mRNA through microscale thermophoresis technology (MST) assay⁴³, indicating that ZIKV-NS2A binds to CYP17A1 mRNA (K_d = 58.2 nM) (Fig. 4d, e). To test whether ZIKV-NS2A has a direct interaction with CYP17A1 mRNA, we preformed the RNA-protein pulldown experiment, indicating the CYP17A1 mRNA-biotin efficiently pulled down GST-NS2A but not GST (Fig. 4f, g). In contrast, a biotinylated RNA control (an iron response element RNA) failed to pull down GST or GST-NS2A (Fig. 4g). The result demonstrated that ZIKV-NS2A could directly bind to the CYP17A1 mRNA. Furthermore, an RNA competition assay was performed to examine the specificity of GST-NS2A and CYP17A1 mRNA interaction. Increasing amounts of unbiotinylated CYP17A1 mRNA competed away the GST-NS2A/CYP17A1 mRNA interaction, whereas yeast tRNAs did not (Fig. 4h), confirming the specificity of the interaction between GST-NS2A and CYP17A1 mRNA. Additionally, the electrophoretic mobility shift assay (EMSA) results further supported that only the NS2A-GST, but not the GST, formed a stable complex with CYP17A1 mRNA and retarded its migration (Fig. 4i). If cytoplasmic CYP17A1 mRNA levels are constant in the presence of ZIKV-NS2A, yet less CYP17A1 protein is produced, this implies reduced translation. This, in turn, should be reflected in less efficient loading of ribosomes onto CYP17A1 mRNA. To determine whether this is the case, we used a published procedure⁴⁴ to measure the ribosome association of CYP17A1 mRNA in the presence or absence of ZIKV-NS2A. The results indicated that the relative level of ribosome-associated CYP17A1 mRNA was significantly decreased in the presence of ZIKV-NS2A, but not ZIKV-E (Fig. 4j). As controls, we also measured the ribosome loading of cellular GAPDH and Calnexin mRNAs (Fig. 4j). These results confirm that the reduced ribosome association observed with CYP17A1 mRNA in the presence of ZIKV-NS2A is specific to CYP17A1 mRNA and not due to a global reduction in ribosome loading onto all mRNAs. Furthermore, we observed that the addition of GST-NS2A decreased the synthesis of CYP17A1 protein in a dose-dependent manner by in vitro translation assays using rabbit reticulocyte lysates (Fig. 4k), indicating that ZIKV-NS2A inhibits translation of CYP17A1 mRNA. Collectively, these results suggest that ZIKV-NS2A interacts with CYP17A1 mRNA and reduces its translation

The 1200-1400 nucleotides of CYP17A1 mRNA is responsible for binding to ZIKV-NS2A

To identify the specific binding region between ZIKV-NS2A and CYP17A1 mRNA, we performed a photo-activatable, ribonucleoside-enhanced crosslinking and immunoprecipitation (PAR-CLIP) experiment. The hCYP17A1 mRNA-containing photoreactive 4-thiouridine (4SU) was prepared through in vitro transcription in the presence of 4-S-UTP. Human CYP17A1 mRNA was incubated with GST-NS2A in vitro, forming RNA-protein complex following UV-irradiation. The 4-thiouridine in the hCYP17A1 mRNA was cross-linked to ZIKV-NS2A under UV light, resulting in U to C conversions during reverse transcription of hCYP17A1 mRNA (Fig. 5a). Subsequently, the hCYP17A1 mRNA was then converted into a cDNA library and analyzed by deep sequencing. A summary of the T to C mutations in the hCYP17A1 mRNA when complexed with ZIKV-NS2A is shown in (Fig. 5b). The highest mutation sites mainly occurred within the 200–300 and 1200–1400 nucleotides of hCYP17A1 mRNA (Fig. 5b).

**Fig. 5: The 1200–1400 nucleotides of CYP17A1 mRNA are responsible for binding to ZIKV-NS2A.**

To investigate the binding of these two regions to ZIKV-NS2A, we conducted a competition assay using in vitro transcribed hCYP17A1 mRNA segments of 200–300 nt and 1200–1400 nt. The full-length hCYP17A1 mRNA served as a positive control. The result indicated 1200–1400 nt segment, but not 200–300 nt segment, competed away the ZIKV-NS2A and hCYP17A1 mRNA interaction (Fig. 5c), indicating that 1200–1400 nt segment binds to ZIKV-NS2A. Next, we aimed to explore whether mCYP17A1 mRNA also interacts with ZIKV-NS2A and whether the structural characteristics of the region on mCYP17A1 mRNA that binds to ZIKV-NS2A are similar to the segment 1200–1400nt of hCYP17A1 mRNA. To address this, we predicted the secondary structures of hCYP17A1 and mCYP17A1 mRNA, revealing that the ZIKV-NS2A binding region, 1200–1400 nt segment on hCYP17A1 mRNA, contains a large number of stem-loop structures, and both hCYP17A1 and mCYP17A1 mRNA form similar secondary structures in this region (Supplementary Fig. 4a, b). Further experiments revealed that the 1200–1400nt segment of mCYP17A1 mRNA also competed away the ZIKV-NS2A and mCYP17A1 mRNA interaction (Fig. 5d), indicating that mCYP17A1 mRNA can bind ZIKV-NS2A, and its segment 1200–1400 nt is responsible for binding to ZIKV-NS2A. Furthermore, we generated several truncated forms of hCYP17A1 mRNA: Δ1 involved the removal of 200–300 nt, Δ2 involved the removal of 1200–1400 nt, and Δ3 involved the removal of both 200–300 nt and 1200–1400 nt (Fig. 5e). These constructs were then individually co-transfected with ZIKV-NS2A into HEK293T cells, indicating the deletion of the 1200–1400 nt segment abolished the inhibitory effect of ZIKV-NS2A on CYP17A1 expression (Fig. 5f). The findings further support the ZIKV-NS2A binds to CYP17A1 mRNA and the 1200–1400 nt of CYP17A1 mRNA is responsible for binding to ZIKV-NS2A.

Mutations at residues 67E/101P abolish the binding between ZIKV-NS2A and CYP17A1 mRNA and the inhibition of CYP17A1 translation

ZIKV-NS2A was reported to be a transmembrane protein with a transmembrane structure anchored on the endoplasmic reticulum, closely associated with the assembly of viral particles on the ER³⁸ (Fig. 6a). Based on the topology of ZIKV-NS2A, we generated various truncated forms of ZIKV-NS2A (Fig. 6b), to pinpoint critical amino acid residues responsible for its binding and subsequent suppression of CYP17A1 mRNA translation. The truncated NS2A protein 1-56aa (NS2A_1-56) had no effect on the expression levels of hCYP17A1 or mCYP17A1. Whereas the truncated NS2A protein 1-103aa (NS2A_1-103), 1-123aa (NS2A_1-123), 1-147aa (NS2A_1-147), 1-196aa (NS2A_1-196), continued to exhibit downregulation of hCYP17A1 or mCYP17A1 expression (Fig. 6c, d). We tentatively concluded that the residues on ZIKV-NS2A responsible for affecting CYP17A1 expression lie between amino acids 56 and 103, which includes a membrane-traversing segment and two flanking loops. We introduced mutations in the amino acid residues within the 56-103 region that could potentially affect NS2A’s RNA binding ability (Fig. 6e). These mutations involved hydrophilic or positively charged residues, including glutamic acid (67E/103E), aspartic acid (73D), histidine (76H), lysine (84 K), and arginine (86 R/96 R/102 R). Additionally, we included the proline (P101) due to its unique properties: the inability to form main-chain hydrogen bonds and the ability to adopt either the cis or trans peptide bond conformation, which can influence protein conformational changes and folding processes^45,46. We found that mutations of residue E67 or P101 on ZIKV-NS2A minimally affected the expression levels of hCYP17A1 or mCYP17A1 (Fig. 6f, g). Subsequent mutations at positions 84, 86, 102, or 103 on ZIKV-NS2A revealed partial compensation for the inhibition of hCYP17A1 or mCYP17A1 (Fig. 6f, g). Additionally, we validated these findings through RIP and RNA pulldown assays, wherein we demonstrated that the mutant NS2A^E67A / NS2A^P101A failed to interact with hCYP17A1 mRNA, unlike the mutant NS2A^D73A (Fig. 6h, i). These mutations (NS2A^K84A, NS2A^R86A, NS2A^R102A, and NS2A^E103A) weakened the interaction between NS2A and hCYP17A1 mRNA compared to the NS2A^WT (Fig. 6h). Moreover, NS2A^E67A/NS2A^P101A was unable to disrupt translation of hCYP17A1 mRNA, as observed by an in vitro translation assay (Fig. 6j, k). Collectively, these data indicate that the sole transmembrane segment of NS2A and its flanking loops are crucial for binding CYP17A1 mRNA and inhibiting its translation. Mutations at amino acid residues 67E or 101 P in this region abolish NS2A’s ability to bind CYP17A1 mRNA and inhibit its translation.

ZIKV^E67A mutant virus exhibits significantly lower inhibition on steroidogenesis and spermatogenesis than ZIKV^WT

To assess whether ZIKV-NS2A can impede steroidogenesis and spermatogenesis during ZIKV infection, we introduced the E67Aor P101A mutation into a ZIKV infectious clone based on Cambodian strain FSS13025 and attempted to rescue ZIKV^E67A or ZIKV^P101A mutant viruses from Vero cells. The original infectious clone construct, which produces wild-type viruses (ZIKV^WT), was used as a control. We confirmed the full length of the genome sequence of the mutant viruses (Supplementary Fig. 5a, b). ZIKV^E67A exhibited similar plaque morphologies to the ZIKV^WT, and growth kinetic analysis in BHK21 cells showed no significant difference between ZIKV^E67A and ZIKV^WT, suggesting that the E67A mutation in NS2A did not affect viral replication in vitro (Supplementary Fig. 5c, d), consistent with previous reports⁴⁷. However, the P101A mutant genomic RNAs produced neither E-positive cells nor infectious viruses in the transfected cells, as indicated by immunofluorescent assay (Supplementary Fig. 5e), plaque assay (Supplementary Fig. 5f), and qRT-PCR (Supplementary Fig. 5g), whereas the WT RNA did. Incubation of naive Vero cells with culture supernatants from the P101A-transfected cells did not generate any viral E-positive cells (Supplementary Fig. 5h). Subsequently, we intraperitoneally infected A129 mice with ZIKV^E67A or ZIKV^WT at the same dose for 6 days. ZIKV^E67A demonstrated the same infection efficiency as ZIKV^WT in vivo (Fig. 7a). We collected mouse serum and testes for testosterone and progesterone measurements. In contrast to ZIKV^WT, ZIKV^E67A failed to affect testosterone and progesterone levels in both mouse serum and testes (Fig. 7b–e). Similarly, ZIKV^E67A resulted in a less significant reduction in the protein levels of CYP17A1 in the testes, without affecting its mRNA production, compared to ZIKV^WT (Fig. 7f, g). ZIKV^E67A did not influence the protein and mRNA levels of other key enzymes involved in testosterone synthesis (Fig. 7f and Supplementary Fig. 5i–k). This phenomenon was also observed through immunofluorescence (Fig. 7h). Additionally, we noted diminished total and motile sperm counts in fluid collected from the cauda epididymis of ZIKV^E67A -infected mice, whereas ZIKV^E67A did not inhibit total and motile sperm production (Fig. 7i, j). Co-immunofluorescence staining with anti-DDX4, a marker for cells at various stages of spermatogenesis, revealed that ZIKV^WT significantly reduced the number of germ cells, while ZIKV^E67A had no inhibitory effect on germ cells (Fig. 7k). This trend was further supported by H&E staining of testes, showing that ZIKV^WT decreased the number of germ cells in seminiferous tubules, while testicular architecture remained intact in ZIKV^E67A-infected mice (Fig. 7l). Next, we determined whether ZIKV-NS2A alone is sufficient to affect steroidogenesis and spermatogenesis in vivo. Single mutations of ZIKV-NS2A at positions 101, 102, or 103 each attenuated the inhibitory effect of NS2A on mCYP17A1. Previous studies have also reported that the NS2A mutants (R102A) and (E103A) significantly attenuate viral replication and block virion assembly, respectively, indicating critical roles of these residues in NS2A fuction⁴⁷. Therefore, we speculated that a combined mutation of residues 101-103 (NS2A^PRE-AAA) would result in a more significant loss of NS2A’s inhibitory effect on mCYP17A1 compared to single mutations. As expected, NS2A^PRE-AAA exhibited a significantly reduced inhibitory effect on the expression of mCYP17A1 (Supplementary Fig. 6a). We then used adeno-associated virus (AAV) to deliver ZIKV-NS2A^WT, ZIKV-NS2A^PRE-AAA, or the control EGFP into the mouse testes through in situ injection, and observed abnormal steroidogenesis and spermatogenesis after a 5-day AAV-NS2A^WT infection compared with AAV-NS2A^PRE-AAA (Supplementary Fig. 6). In conclusion, our data demonstrate that NS2A can suppress sperm production by inhibiting testosterone synthesis. Additionally, the residue 101P on NS2A that regulates its binding to CYP17A1 mRNA is also critical for viral propagation.

Discussion

To date, ZIKV is the only flavivirus that can get across the BTB and establish long-term infection in testis⁴¹. The persistence of ZIKV particles in male semen for an extended duration after viremia clearance, along with its sexual transmission from male to female and vertical transmission exacerbate the concern of fetal infection in resulting pregnancies³. Moreover, related reproductive symptoms, including hematospermia and oligospermia, have been reported in male ZIKV patients, implying its harmful effects on spermatozoa production and fertility. The testis functions as the primary organ for spermatogenesis, which is tightly regulated by hormones, particularly testosterone. Considering the testis is a potential reservoir for ZIKV and major steroidogenesis organ, there exists a theoretical basis for investigating the impact of ZIKV infection on testosterone production. Indeed, early clinical observations in male patients infected with ZIKV exhibited notably diminished testosterone levels⁸, consistent with the observations both in ZIKV-infected IFNα/β receptor-deficient mice and immunocompetent macaques, suggesting the decrease in testosterone was independent of ZIKV-induced interferon activation, and some other mechanisms are involved. However, the absence of IFN signaling in A129 mice presents limitations for studying viral pathogenesis and antiviral immunity. For instance, IFN-α/β-dependent restrictions on viral tropism or effects on B and T cell priming will not be evident in this model⁴⁸. Although ZIKV infection in immune-deficient mice develop pronounced genital pathological changes, such as substantial inflammation in the testes and testicular atrophy, severe symptoms such as orchitis are rarely reported in humans⁴⁹, implying virus direct infection rather than the immune-mediated damage play more important roles in the decreased sex hormone production and oligospermia manifested in ZIKV-infected humans. However, the specific mechanisms by which ZIKV infection decreases testosterone secretion and impairs sperm production remain largely unknown.

Current studies on ZIKV proteins have partially answered questions about viral evolution, virulence, tissue tropism, immune evasion, and pathogenicity, with a specific focus on the mechanisms whereby ZIKV induces neurodevelopmental diseases^50,51,52,53. Nevertheless, there is too limited knowledge of the molecular-level mechanisms underlying male reproductive consequences of virus infection. Here, from an unbiased and systematic screening of ZIKV-encoded proteins, we identified ZIKV-NS2A as a causal factor for ZIKV-induced low testosterone levels. Mechanistically, ZIKV-NS2A sequestrates CYP17A1 mRNA and inhibits its translation by diminishing the recruitment of ribosomes, leading to a decrease in testosterone synthesis, which in turn suppresses spermatogenesis. Our study, along with others⁵⁴, revealed that ZIKV infection does not significantly impact the viability of Leydig cells—even at high viral infection doses (Supplementary Fig. 1g–j). As Leydig cells are the exclusive site of CYP17A1 expression in the testes, this finding indicates that the reduced testosterone levels observed during ZIKV infection are not a result of Leydig cell death. This underscores a direct correlation between the decreased testosterone levels resulting from ZIKV infection and the inhibition of CYP17A1 mRNA translation by ZIKV-NS2A. Additionally, we performed a comparative analysis of the impact of different flavivirus NS2As on CYP17A1 expression, revealing that only ZIKV-NS2A exhibited a notably inhibitory effect on CYP17A1 expression. We postulated that this outcome is likely attributed to high sequence diversity and distinct membrane topology observed across various flavivirus NS2As^47,55,56. Indeed, a series of studies by Shi et al. have unveiled that, in stark contrast to the topologies of NS2As in other flavivirus, ZIKV-NS2A has a single transmembrane segment traversing the ER membrane, with the remaining six pTMSes peripherally associated with the ER membrane, forming three cytoplasmic loop regions^47,56,57. Through truncation experiments, we determined that the amino acid sequence spanning residues 56 to 103 predominantly inhibits the expression of CYP17A1 by NS2A. Notably, scanning mutagenesis studies within this sequence identified amino acid residues 67E and 101P as pivotal for facilitating NS2A binding and suppressing CYP17A1 mRNA translation. Subsequently, we sought to determine if the incapability of NS2A^E67A or NS2A^P101A to block the translation of CYP17A1 could also be observed in the context of an authentic viral infection. To investigate this, we generated mutant viruses through reverse genetics based on the infectious clone of the ZIKV strain FSS13025. Interestingly, ZIKV^E67A infection, compared to ZIKV^WT, resulted in the loss of its capacity to suppress testosterone synthesis and spermatogenesis. We also revealed that amino acid residues 84K, 86R, 102R, and 103E exhibited partial inhibition of CYP17A1 expression by NS2A. However, these mutants can reduce viral production and abolish virion assembly⁴⁷. Functionally, these charged or hydrophilic amino acid residues could potentially influence the conformation and subcellular localization of the NS2A protein within the endoplasmic reticulum. Consequently, we hypothesize that the loss of CYP17A1 expression inhibition by the E67A mutant results from mutations altering the topology and localization of NS2A. However, the precise mechanism by which ZIKV-NS2A influences CYP17A1 remains unclear, and other contributing factors may be involved. It is not yet understood how NS2A 56-103, which compose the pTMS3 domain flanked by two loops oriented towards the ER lumen and the cytosol, interacts with CYP17A1 mRNA while NS2A is recruited with other viral proteins into the ER membrane-associated viral replication complexes. In this study, NS2A was observed to be inherently unstable compared to other flavivirus NS2A proteins (Supplementary Fig. 2g). Our recent study has reported that NS2A undergoes ubiquitination and proteolytic degradation processes⁵⁸ that may facilitate the release of NS2A fragments and their subsequent exposure to the cytosolic CYP17A1 mRNA. Additionally, the impact of mutations at positions E67, K84, R86, P101, R102, or E103 in NS2A 56-103 on CYP17A1 mRNA binding and translation remains unknown. Further research is needed to clarify these interactions by analyzing the dynamic structure of the NS2A with CYP17A1 mRNA during viral replication.

On the other hand, the RNA secondary structures, such as hairpins and stem loops, play a pivotal role in influencing gene expression, translation, as well as protein recognition and binding^59,60,61,62. Integration of PAR-CLIP analysis with predicted secondary structure analysis of hCYP17A1 mRNA disclosed that the region spanning nucleotides 1200–1400, consisting of numerous stem-loops, is responsible for binding to ZIKV-NS2A. Similar secondary structures are also predicted on mCYP17A1 mRNA in the same region. Through in vitro binding assays, we confirmed the binding of 1200–1400 nt segment to ZIKV-NS2A, and the removal of this segment annulled the inhibitory effect of ZIKV-NS2A on CYP17A1 expression. Interestingly, ZIKV-NS2A binds to a stem-loop structure in the 3′ UTR of the ZIKV genome, serving as an “RNA recruitment signal” for ZIKV assembly³⁸. Based on this information, we propose that the presence of these structures may be the primary reason for the specific binding of ZIKV-NS2A to CYP17A1 mRNA.

RNA-binding proteins (RBPs) are commonly defined as proteins that bind RNA through one or multiple RNA-binding domains (RBDs) and change the fate or function of the bound RNAs^63,64,65. Flavivirus NS2As, including KUNV-NS2A, ZIKV-NS2A, and DENV-NS2A, have been shown to bind RNA molecules of various species, such as the viral genomic RNA^38,57,66 and dsRNA⁶⁷. Additionally, our study also identified binding to a host mRNA by ZIKV-NS2A. This prompts an investigation into whether these flaviviruses NS2As, especially ZIKV-NS2A, function as RNA-binding proteins. Future work should explore if these interactions are mediated by consensus motifs, represented as linear sequence motifs, or RNA structural motifs, or the structural context in which linear motifs are embedded. This could be achieved through a systematic combination of advanced biochemical methods, like PAR-CLIP, enabling the identification of RNAs bound by RBPs and the corresponding protein binding sites, and computational analysis based on an expanded dataset of RNA secondary structures⁶². Taken together, our study reveals a unique molecular mechanism by which ZIKV inhibits steroidogenesis and spermatogenesis, offering potential therapeutic targets for mitigating ZIKV-induced male productive system injury and low testosterone secretion. These findings expand our comprehension of virus-host cell interactions in viral pathogenesis and would represent an exciting avenue for future studies.

Methods

Ethics statement

All the animal experiments were performed according to the Regulations of Hubei Province Laboratory Animal Management and approved by Wuhan University Animal Experiment Ethics Committee under project licence WP20230062. All animal experiments were conducted under isoflurane anesthesia to minimize animal suffering. Studies with ZIKV were conducted under biosafety level 2 and Animal Biosafety Level 3 (BSL3) containment.

Cells and viruses

A549 cells (Adeno-carcinomic human alveolar basal epithelial cells), HEK293T cells (Human embryonic kidney cells), Vero cells (African green monkey kidney epithelial cells) and BHK-21 cells (Baby hamster kidney) were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS), 1% penicillin-streptomycin (P/S) at 37 °C in an atmosphere containing 5% CO₂. Mosquito C6/36 cells (Aedes albopictus mosquito cells) were grown at 28 °C in RPMI 1640 medium containing 10% FBS, 1% P/S. The CYP17A1-A549 cells were maintained in DMEM supplemented with 10% FBS 1% P/S. Primary mouse Leydig cells (mLCs) were isolated from C57BL/6 mice. ZIKV strain ZIKA-SMGC-1 (GenBank accession number: KX266255) was kindly provided by Dr. George F. Gao (Institute of Microbiology, Chinese Academy of Sciences, Beijing, China). HA-NS2A ZIKV (HA-tagged NS2A ZIKV using an infectious cDNA clone of Cambodian strain FSS13025) was kindly provided by Dr. Gong Cheng (Tsinghua University-Peking University Joint Center for Life Sciences, School of Medicine, Tsinghua University, Beijing, China). Viruses were propagated in C6/36 cells, and viral titers were determined by crystal violet-based plaque assay.

Mouse experiments

A129 mice (IFNα/β receptor-deficient) were obtained from the Institute of Laboratory Animal Science, Chinese Academy of Medical Sciences, and Peking Union Medical College. C57BL/6 mice were purchased from the Hubei Provincial Center for Disease Control and Prevention (Wuhan, China). All mice were housed on a 12 light /12 dark cycle, at ~18–23 °C with 40%–60% humidity. For ZIKV infection, 6–8-week-old male A129 mice were intraperitoneally infected with 1 × 10⁶ plaque-forming units (PFU) of ZIKV. Mice administered 1 × PBS served as a control. Mice were euthanized at the indicated time points post-infection. Whole blood was collected for qRT-PCR (n = 5 per group) and hormone measurements (n = 5 per group), while testes were harvested for metabolomics analysis (n = 8 per group), proteomics analysis (n = 3 per group) and other assays, such as qRT-PCR (n = 5 per group), hormone measurements (n = 5 per group), immunofluorescence analysis (n = 3 per group), and Western blotting analysis (n = 3 per group). For AAV infection, 10 μL (equivalent to 1 × 10⁹ vg) of AAV8-NS2A^WT, AAV8-NS2A^PRE-AAA, or AAV8-GFP was injected into testes. Mice administered 1 × PBS served as a control. At the indicated time points post-infection, the mice were euthanized, and both whole blood and testes were collected for qRT-PCR (n = 5 per group), hormone measurements (n = 5 per group), immunofluorescence analysis (n = 3 per group), and Western blotting analysis (n = 3 per group).

Plaque assay

BHK-21 cells were seeded in 6-well plates and grown to 80%–90% confluence. Viral samples were ten-fold serially diluted six times in DMEM and added to each well, followed by incubation at 37 °C for 2–3 h. The viral solution was then replaced with 2 mL of 1.2% methylcellulose, and the plates were incubated for 5 days and then fixed and stained using formaldehyde-crystal violet, and the visible plaques were counted and viral titers (PFU/mL) were calculated as described previously⁶⁸.

Antibodies

Rabbit monoclonal anti-CYP17A1 (Cat. no. 94004, 1:1000), rabbit monoclonal anti-CYP11A1(Cat. no. 14217, 1:1000), rabbit monoclonal anti-β-actin (Cat. no. 4970, 1:1000), mouse monoclonal normal IgG (Cat. no. 5415, 1:200), rabbit monoclonal anti-Cleaved Caspase-3 (Cat. no. 9664, 1:1000), and rabbit monoclonal anti-LC3 (Cat. no. 12741, 1:1000) were purchased from Cell Signaling Technology. Mouse monoclonal anti-GAPDH (Cat. no. G9295, 1:5000), mouse monoclonal anti-FLAG M2 (Cat. no. F1804, 1:1000), mouse monoclonal anti-HA (Cat. no. H9658, 1:5000), mouse monoclonal anti-GST (Cat. no. G1160, 1:5000), goat anti-mouse IgG conjugated with HRP (Cat. no. A5278, 1:5000), and goat anti-rabbit IgG conjugated with HRP (Cat. no. AP307P, 1:5000) were purchased from Sigma-Aldrich. Rabbit polyclonal anti-ZIKV-E (Cat. no. GTX133314, 1:5000) and mouse anti-ZIKV antibody 4G2 (Cat. no. GTX57154, 1:1000) were purchased from GeneTex. Rabbit polyclonal anti-GFP (Cat. no. 50430-2-AP, 1:1000), rabbit polyclonal anti-p62 (Cat. no. 18420-1-AP, 1:10000), and rabbit polyclonal anti-Calnexin (Cat. no. 10427-2-AP, 1:20000) were purchased from Proteintech. Rabbit polyclonal anti-STAR (Cat. no. ab96637, 1:1000) and rabbit polyclonal anti-DDX4 (Cat. no. ab13840, 1:1000) were purchased from Abcam. Alexa Fluor 488 goat anti-mouse IgG antibody (Cat. no. 406406, 1:200) and Alexa Fluor 488 goat anti-mouse IgG antibody (Cat. no. 405319, 1:200) were purchased from BioLegend. Mouse monoclonal anti-3β-HSD (Cat. no. sc-515120, 1:1000) was purchased from Santa Cruz Biotech. Fluorescein-Conjuated Affinipure Goat anti-human IgG (Cat. no. ZF-0308, 1:200) was purchased from ZSGB-BIO. Anti-ZIKV antibody (FITC-Z6, 1:1000) was Kindly gifted by Dr. George F. Gao.

Plasmid construction

A mammalian expression vector pcDNA3.1 (Thermo Fisher Scientific) with a cytomegalovirus (CMV) promoter was used to express various viral proteins (C, prM, E, NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5) in HEK293T cells. To construct the pcDNA3.1-ZIKV-C/prM/E/NS1/NS2B/NS3/NS4A/NS4B/NS5-FLAG plasmid, the genes encoding these ZIKV proteins (strain FSS13025) were synthesized and fused with a C-terminal FLAG tag sequence, followed by cloning into the pcDNA3.1 vector. For constructs pcDNA3.1-SPG-C16-NS2A-FLAG, a DNA fragment containing the Gaussia luciferase signal peptide (SPG), the last 16 amino acids of NS1 protein (C16) to ensure the correct processing and membrane topology of NS2A, and NS2A tagged with a C-terminal FLAG tag sequence was synthesized and cloned into pcDNA3.1 vector. DNA fragments that encode different truncated NS2A segments were amplified from the above pcDNA3.1-SPG-C16-NS2A-FLAG vector and cloned into the pcDNA3.1, resulting in the pcDNA3.1-SPG-C16-ctNS2A-FLAG plasmids (ctNS2A represents C-terminal truncated NS2A). Similarly, pcDNA3.1-SPG-C16-mutNS2A-FLAG plasmids (mutNS2A indicating NS2A mutants) were constructed using the same method. human and mouse CYP17A1(hCYP17A1/mCYP17A1) genes were fused with a C-terminal human influenza hemagglutinin (HA) tag and cloned into the pcAGGS vector under a CMV promoter, resulting in the pcAGGS-hCYP17A1/mCYP17A1-HA plasmid.

pGEX-4T-1 (GE Healthcare Life Sciences) was used as the vector for expressing all N-terminally glutathione S-transferase (GST)-tagged NS2A fusion proteins. The DNA fragments containing various NS2A mutants were cloned into pGEX-4T-1 through restriction enzyme sites BamHI and XhoI. To generate an A549 cell line constitutively expressing CYP17A1, the CYP17A1 gene was constructed by overlap PCR and inserted into the pWPI vector via PmeI and SpeI restriction sites and co-transfected into HEK293T cells with pMD2.G and pSPAX2 plasmids to produce lentivirus.

All plasmids were propagated in E. coli DH5α and verified by DNA sequencing (Tsingke Biotechnology Co., Ltd). All restriction enzymes were purchased from New England Biolabs. Primers were synthesized by Genecreate Biotech.

Adenovirus-associated virus (AAV) production

GFP, HA-NS2A^WT, or HA-NS2A^PRE-AAA were cloned into AAV vector (pAAV-CMV), and then transfected into HEK293T cells with pAAV2/8-RC and pHelper plasmid (gifted from Dr. Yuchen Xia) for 3 days to produce AAVs. AAVs-containing supernatants were harvested and filtered with 0.22 μm filter, then concentrated with Amicon Ultra-15 centrifugal filter unit (Sigma-Aldrich). The concentration of purified AAVs was determined by qRT-PCR as described previously in ref. ⁶⁹ and finally AAVs at the concentration of 1 × 10¹¹ vg/mL were obtained.

Isolation and culture of primary mouse Leydig cells (mLCs) and cell viability assay

The procedure for isolation of mouse Leydig cells was refined from a procedure described previously in ref. ⁷⁰. Briefly, 6-week-old male mice were anesthetized with CO₂ and then killed by decapitation. Mice were soaked in 75% alcohol for 5 min, and bilateral testicles were removed through abdominal asepsis. The white membrane was removed, and the testicular tissue of the mice was kept intact. Testes obtained from these mice were dispersed by treatment with 0.5 mg/mL combined type II and IV collagenase (Sigma-Aldrich, C2-BIOC and C4-BIOC) at room temperature for 30 min under gentle oscillation. This suspension was then filtered through 100 μm copper meshes and centrifuged with centrifugation at 1000 × g for 5 min. The cells were then cultured in a medium DMEM/F-12 (Gibco) supplemented with 10% FBS, 1% P/S at 37 °C with 5% CO₂. After 2 h of incubation, spermatogenic cells and cell debris were removed by washing twice with phosphate-buffered saline (1 × PBS). Primary mLCs that adhere to culture dishes were used in this study. The purity of these Leydig cells was assessed by staining for the cholesterol side-chain cleavage enzyme (CYP11A1) (a marker of Leydig cells). The viability of the primary mLCs were measured using the Alamar Blue (resazurin) fluorescent dye (Sigma-Aldrich).

Enzyme-linked immunosorbent assay (ELISA)

A129 male mice were intraperitoneally challenged with 10⁶ PFU of ZIKV. Mice administered 1 × PBS served as a control. At different time points, the mouse testes were collected by carefully removing the fascia and large visible blood vessels from the surface of each testis to minimize blood contamination. The testes were then rinsed three times with chilled 1 × PBS, followed by homogenization. Blood samples were collected via retro-orbital bleeding and centrifuged to separate the serum. The concentrations of testosterone and pregnenolone in the serum and testis homogenates were measured using mouse testosterone ELISA Kits and mouse pregnenolone ELISA Kits (Beyotime) following the manufacturer’s instructions.

H&E analyses and immunofluorescence

Testes from ZIKV-infected and Mock-infected mice were collected at different time points, fixed in 4% paraformaldehyde (PFA) solution overnight, then dehydrated and paraffin-embedded. For immunofluorescence, 6 μm testis sections were stained with specific antibodies, including CYP17A1, CYP11A1, ZIKV-E, or DDX4, followed by fluorophore-labeled secondary antibodies. For H&E staining, 5 μm testis sections were used. Images were then captured with a Zeiss LSM510 laser scanning confocal microscope.

Confocal microscopy

The cells (Vero cells, CYP17A1-A549 cells, or mLCs) were fixed in 4% PFA at room temperature for 20 min. After three washes with 1 × PBS, the cells were permeabilized using wash buffer (1 × PBS with 0.1% Triton X-100) for 5 min, and subsequently blocked with 1 × PBS containing 5% bovine serum albumin (BSA) for 1 h. The cells were then incubated overnight at 4 °C with the primary antibody, followed by incubation with secondary antibodies conjugated to Alexa Fluor 488 and Alexa Fluor 649. Finally, the cells were stained with 4′,6-diamidino-2-phenylindole (DAPI) solution for 5 min and subjected to analysis using a confocal laser scanning microscope (Fluo View FV1000; Olympus, Tokyo, Japan).

Real-time quantitative RT-PCR (qRT-PCR)

Total RNA was extracted with TRIzol reagent (Invitrogen), and the concentration was detected by NanoDrop (Thermo Fisher Scientific). Quantitative RT-PCR was performed with iTaq Universal SYBR Green Supermix (Bio-Rad Laboratories) on a Bio-Rad CFX96 Real-Time PCR System. A probe-based assay was performed to quantify viral RNA copy number by qPCR amplification of ZIKV-E gene. A standard curve was generated using serially diluted expression plasmids containing the coding sequence of ZIKV-E. mRNA expression levels of the indicated genes were normalized to the expression levels of GAPDH as an endogenous reference control.

The following primers were used:

Human CYP17A1 forward: 5′-GGACCTAGTCCCCTGGTTGA-3′

Human CYP17A1 reverse: 5′-CACTCCGGAATTTCTCCTTGT-3′

Human GAPDH forward: 5′-ATGACATCAAGAAGGTGGTG-3′

Human GAPDH reverse: 5′-CATACCAGGAAATGAGCTTG-3′

Mouse CYP17A1 forward: 5′-TTCGCCTGGGTACCACAACTGC-3′

Mouse CYP17A1 reverse: 5′-TAGAGTCACCATCTGGGGCCGA-3′

Mouse CYP11A1 forward: 5′-GGGGCAACAAGCTGCCCTTCAA-3′

Mouse CYP11A1 reverse: 5′-TGCAGGGTCATGGAGGTCGTGT-3′

Mouse GAPDH forward: 5′-TGTGTCCGTCGTGGATCTGA-3′

Mouse GAPDH reverse: 5′-TTGCTGTTGAAGTCGCAGGAG-3′

Mouse STAR forward: 5’-AAAGCCAGCAGGAGAACGGGGA-3’

Mouse STAR reverse: 5′-GCCTCCATGCGGTCCACAAGTT-3′

Mouse 3β-HSD forward: 5’-GCGGCTGCTGCACAGGAATA-3′

Mouse 3β-HSD reverse: 5′-GACGCATGCCTGCTTCGTGA-3′

ZIKV-E forward: 5′-TGAYAAGCARTCAGACAC-3′

ZIKV-E reverse: 5′-TCACCARRCTCCCTTTGC-3′

Western blotting analysis

The total amount of sample protein is determined by BCA Protein Assay Kit. The samples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to 0.45 μm pore size polyvinylidene difluoride (PVDF) membranes (Sigma-Aldrich) as described previously⁴¹.

Microscale thermophoresis technology (MST)

MST-experiments were performed using Monolith NT.115 (Blue/Red) instrument (NanoTemper Technologies GmbH, Germany). The cell lysates of HEK293T cells expressing GFP-NS2A were used as a source of fluorescently labeled NS2A. HEK293T cells were transfected with GFP-fused NS2A and lysed 24 h after transfection. To evaluate binding of CYP17A1 mRNA to NS2A, cell lysates were diluted with MST buffer (10 mM Na-phosphate buffer, pH 7.4, 1 mM MgCl₂, 3 mM KCl, 150 mM NaCl, 0.05% Tween-20, and RNase Inhibitor) to provide optimal level of fluorescence. The lysates of GFP-transfected HEK293T cells were used as negative control. Cell lysates were incubated with titration series of CYP17A1 mRNA (0–100 mM). Samples were loaded into Premium Coated Capillaries, and MST measurements were performed using 20% MST power and 40% LED power at 25 °C. Laser-on and -off times were 30 and 5 s respectively. The MO Affinity Analysis v.1.6.1 software (NanoTemper Technologies) was used to fit the data and to determine the Kd values.

Electrophoretic mobility shift assay (EMSA)

EMSA-experiments were performed to evaluate the binding affinity of GST-NS2A to CYP17A1 mRNA. In brief, 20 ng biotinylated CYP17A1 mRNA was incubated with various concentration of recombinant GST-NS2A proteins in 50 μL binding buffer (50 mM Tris-HCl, pH 7.5,10 mM 2-mercaptoethanol, 2 mM MgCl₂, 50 mM KCl) for 10 min at 24 °C, and the protein-RNA complex migrated in 0.4% agarose gel without sodium dodecyl sulfate (SDS) and was transferred to a BrightStar-Plus positively charged nylon membrane (Thermo Fisher Scientific) according to the instructions of the NorthernMax-Gly kit (Ambion). After ultraviolet cross-linking treatment (200 mJ/cm² of 25 nm UV light), the nylon membrane was incubated with Streptavidin-HRP Conjugate and detected by chemiluminescence analysis.

Purification of GST recombinant protein

GST-NS2A^WT, GST-NS2A^E67A, GST-NS2A^D73A, and GST-NS2A^P101A were expressed in E. coli strain BL21 (DE3) at 18 °C overnight in the presence of 0.5 mM IPTG. The cells were pelleted down by centrifugation at 9000 × g for 10 min at 4 °C and resuspended in a lysis buffer [pH 7.5, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄, 2.7 mM KCl, 400 mM NaCl, 1 mM EDTA, 10 U/mL Benzonase nuclease (Sigma-Aldrich), and EDTA-free protease inhibitor cocktail (Roche)], followed by ultrasonication at 4 °C. The lysates were clarified by centrifugation at 20,000 × g for 10 min. The supernatants were further centrifuged at 39,000 × g for 1.5 h in an SW40Ti rotor (Beckman Coulter) at 4 °C to pellet down the membrane fractions. Afterward, the pellet was resuspended in a solubilization buffer (pH 7.5, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄, 2.7 mM KCl, 1 M NaCl, 1 mM EDTA, 1.5% DDM) with agitation for 2 h at 4 °C. After centrifugation at 39,000 × g for 30 min to remove undissolved debris, the proteins in the supernatant were purified by Glutathione Beads (Smart-Lifesciences) according to the manufacturer’s protocol, and the purified proteins were quantified by BCA protein assay Kit.

RNA-protein pulldown assay

CYP17A1 cDNAs and truncated constructs were transcribed in vitro using the HiScribe T7 ARCA mRNA Kit (New England Biolabs). Biotinylated CYP17A1 mRNA (CYP17A1 mRNA-biotin) was synthesized by Pierce™ RNA 3’End Desthiobiotinylation Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. The biotinylated RNAs were purified using the Monarch RNA Cleanup Kits (New England Biolabs). Briefly, the RNA samples were transferred into an RNase-free tube and mixed with two volumes of RNA Cleanup Buffer. Subsequently, one volume of ethanol (≥95%) was added, and the mixture was thoroughly mixed by pipetting up and down or flicking the tube (do not vortex). The sample was loaded onto an RNA cleanup column placed in a collection tube, and the cap was closed. The column was spun for 1 min, and the flow-through was discarded. The column was re-inserted into the collection tube, and 500 μL of RNA Cleanup Wash Buffer was added. After a 1-min spin, the flow-through was discarded, and the wash step was repeated. The column was then transferred to an RNase-free tube. Finally, the purified RNAs were eluted in nuclease-free water and stored at −80 °C. RNA-protein pulldown assay was performed using the Pierce magnetic RNA-protein pulldown Kit (Thermo Fisher Scientific) according to the manufacturer’s guideline.

Quantification of ribosome loading on to CYP17A1 transcripts

Ribosome-loaded RNA was isolated essentially as previously described⁴⁴. Briefly, CYP17A1 was co-transfected into HEK293T cells with ZIKV-NS2A, ZIKV-E, or empty vector as control. At 48 h post-transfection, cells were then treated with 100 μg/mL of cycloheximide for 10 min at 37 °C, washed in PBS and lysed in ribosome lysis buffer (10 mM Tris-HCl pH 7.4, 5 mM MgCl₂, 100 mM KCl, 1% Triton X-100, protease inhibitor, 2 mM DTT, 100 μg/mL cycloheximide and RNase inhibitor). One-tenth of the lysate was then used for RNA isolation for the “input” sample. The rest of the lysates were sheared by passage through a 26-gauge needle before clarification by centrifugation at 1300 × g for 10 min. The supernatants were brought up to 10 mL in lysis buffer and overlaid on a 30% sucrose cushion in lysis buffer, and centrifuged at 164,000 × g for 2 h at 4 °C. RNAs were then extracted from the ribosomal pellet using TRIzol. RNA isolated from the input and ribosome loaded fractions then underwent reverse transcription followed by qRT-PCR.

Photoactivatable-ribonucleoside-enhanced crosslinking and immunoprecipitation (PAR-CLIP)

PAR-CLIP-experiments were performed according to the protocol as described previously^57,71. Briefly, 4-thiouridine labeled CYP17A1 mRNA was prepared by adding 4-Thio-UTP (TriLink BioTechnologies) using the HiScribe T7 ARCA mRNA Kit (New England Biolabs). Recombinant GST-NS2A protein was expressed in E. coli strain BL21 (DE3). Ten μg of 4-thiouridine-labeled CYP17A1 mRNA was incubated with 5 μg recombinant GST-NS2A and then irradiated on ice with 150 mJ/cm² at UV light at 365 nm. An aliquot of the same reaction without UV irradiation was used as the uncross-linked control. The RNA-protein mixtures were immunoprecipitated using BeyoMag Anti-GST Magnetic Beads (Beyotime). The beads were washed and resuspended in proteinase K digestion buffer (100 mM Tris, pH 7.5, 50 mM NaCl, 10 mM EDTA, and 1% SDS) supplemented with proteinase K at a final concentration of 2 mg/mL. Subsequently, the RNA in the reaction was purified using the Monarch RNA Cleanup Kits (New England Biolabs) and resuspended in nuclease-free water. The extracted RNA was reverse transcribed, amplified, and the PCR product was separated on 10% agarose gels, followed by purification using E.Z.N.A. Gel Extraction Kit (Omega BioTek). The purified PCR product was turned into a cDNA library using a TransNGS Tn5 Plasmid DNA Library Prep Kit (TransGen Biotech) and subjected to sequencing using the Illumina MiSeq platform. The indexed sequencing reads of both forward and reverse strands were proceeded with the following bioinformatic pipelines: first, an initial quality filtering was conducted with Fastp to remove low quality reads. Then, the remaining reads underwent adapter trimming with TrimGalore 0.6.10 automaticly (https://github.com/FelixKrueger/TrimGalore/). Subsequently, a paired-end alignment was conducted using the Bowtie2(2.5.2), followed by sorting using SAMtools v1.4 with command line parameters sort, index, and mpileup. Finally, mutation calling for variant positions was performed using varscan (https://github.com/dkoboldt/varscan/)(--min-var-freq0--min-avg-qual10), and the results were visualized using the R4.2.3 with ggplot2 (3.4.4).

Untargeted metabolism analysis

Male A129 mice were infected with ZIKV to form the ZIKV group, while mice administered 1 × PBS served as the Mock group. At 10 days post-infection, all mice were euthanized, and testes were collected for metabolomic analysis (n = 8 per group). Briefly, 20 mg of the testis sample was thawed on ice and homogenized in 400 μL of ice-cold methanol: water (70%, V/V). The homogenized mixtures were incubated on ice for 15 min and centrifuged at 4 °C, 12,000 g for 10 min. The supernatant (300 μL) was collected, incubated at −20 °C for 30 min, and then centrifuged again at 4 °C, 12,000 × g for 3 min. The final supernatant was collected for LC-MS/MS analysis. In brief, the sample extracts were analyzed using a LC-MS/MS system (Waters Acquity SM-FTN-H; Applied Biosystems, TripleTOF 6600 + ) with a Phenomenex Kinetex C18 column (2.6 µm, 2.1 mm × 100 mm). Column temperature, flow rate, and injection volume were set 40 °C, 0.3 mL/min, and 2 μL, respectively. Mobile phase was composed of water containing 0.1% formic acid (A) and acetonitrile containing 0.1% formic acid (B). The gradient program initiated from 5% B increased to 95% B in 11.0 min, and held for 5 min, and then decreased 5% B for re-equilibrium. For the quality control (QC) of metabolomic analysis, we pipette 10 μL of each sample to pool a QC sample. When running sample sets on column, one QC sample was injected after 8 samples in the sequence.

Data acquisition was performed using the information-dependent acquisition (IDA) mode, implemented with Analyst TF 1.7.1 software (Sciex). The source operation parameters were as follows: ion source gas 1 and gas 2 were set at 55 psi, curtain gas at 35 psi, and source temperature at 500 °C for positive ion mode and 450 °C for negative ion mode. The declustering potential was configured to 80 V for positive ion mode and −80 V for negative ion mode, while the ion spray voltage was set to 5500 V for positive and −4500 V for negative ion modes.

Data processing and quantification were carried out using the Analytics module of SCIEX OS 3.0 software (AB SCIEX, USA). For metabolite identification, MS/MS spectra were matched to the SCIEX Metabolite_HRMS/MS library (version 1.0), with additional annotations obtained from public databases including HMDB　and MetDNA2. The mass tolerance values applied for both MS1 and MS2 ions during metabolite identification were 5ppm. MS-DIAL software also used for metabolite annotation with library ESI (+/−)-MS/MS from authentic standards. The mass tolerance value used in MS-DIAL for MS1 was 0.01 Da and for MS2 was 0.025 Da. Minimum peak height was 1000 amplitude, and mass slice width was 0.1 Da. Peaks with a detection rate below 80% in any sample group were excluded from further analysis.

Finally, statistical analysis was carried out by the R program, with both univariate analysis (Student’s t test) and multivariate analysis (principal component analysis, PCA, and orthogonal partial least-squares discriminant analysis, OPLS-DA). The PCA was employed to visualize the clustering of metabolite profiles and to assess the intrinsic variation between ZIKV-infected and Mock-infected groups. The OPLS-DA model was evaluated with the relevant R2 and Q2, which assess the model’s goodness of fit and predictive ability, respectively. VIP (Variable Importance in Projection) values were extracted from the OPLS-DA results. Only metabolites with VIP > 1 and P value < 0.05 were considered statistically significant. Metabolites with significant differences were subjected to KEGG pathway enrichment analysis using MetaboAnalyst v.4.0, with metabolic pathways with P value < 0.05 deemed significant. Volcano plots were used to filter metabolites of interest which based on log2(FC) and -log10(P-value) of metabolites.

iTRAQ-LC-MS/MS analysis

The iTRAQ experiment was performed in triplicates, with each replicate containing three testicular samples from a Mock-infected mouse, a ZIKV-infected mouse (6 dpi), and a ZIKV-infected mouse (10 dpi). In each replicate, the Mock group was compared with both the 6 dpi and 10 dpi ZIKV-infected groups. Total proteins were extracted from the testicular samples. The testes samples were homogenized in lysis buffer (2.5% SDS, 100 mM Tris-HCl, pH 8.5). The lysates were boiled for 15 min and then sonicated in an ice-water bath to denature the proteins, shear the DNA, and enhance cell disruption. After centrifugation at 4 °C, 12,000 × g for 30 min, the protein in the supernatant was precipitated with four volumes of cold acetone at 4 °C overnight. After centrifugation, the protein pellets were washed twice in cold acetone, air-dried, and dissolved in a buffer consisting of 8 M urea and 100 mM Tris-HCl, pH 8.0. The protein samples were reduced with 10 mM DL-dithiothreitol reducing agent at 37 °C for 1 h and alkylated with 40 mM iodoacetamide at room temperature for 1 h in the dark. Protein concentrations were determined by the Bradford method. Then, the urea concentrations in the samples were reduced to <2 M with 100 mM Tris-HCl (pH 8.0). The proteins were digested overnight at 37 °C with trypsin at an enzyme/protein ratio of 1:50 (w/w). The digestion was stopped by adding TFA until the pH reached 6.0. The supernatant of the digest was subjected to peptide purification using a Sep-Pak C18 desalting column, and the eluate was vacuum dried and stored at −20 °C until use. For iTRAQ labeling, 100 μg of each trypsin-digested peptide sample was reconstituted in iTRAQ reagent buffer, and separately labeled with different channels of 4plex-iTRAQ labeling kit: Mock-infected mouse was labeled with iTRAQ-114, ZIKV-infected mouse for 6 days (6 dpi) with iTRAQ-116, ZIKV-infected mouse for 10 days (10 dpi) with iTRAQ-117. The labeled samples in each replicate were mixed together, desalted using a C18 column, and vacuum dried. Each replicate mixture was fractionated by high pH reverse chromatography using a C18 5 μm 4.6 × 150 mm column (TechMate) on an Agilent 1260 HPLC operated at flow rate of 0.2 mL/min. Buffer A (10 mM ammonium formate, pH 10.0) and Buffer B (10 mM ammonium formate, 80% ACN, pH 10.0) were used to establish the separation gradient. For each replicate, 60 fractions were collected and merged into 15 fractions. Each fraction was vacuum dried and stored at −80 °C.

The LC-MS/MS analysis was performed on a TripleTOF 5600 mass spectrometer (Sciex), as described previously in ref. ⁷². All mass spectrometry analyses were conducted by SpecAlly Life Technology Co., Ltd, Wuhan, China. In brief, samples were reconstituted in 0.1% formic acid and injected by an autosampler onto a trapping column (C18, 5 µm, 5 × 0.3 mm) for desalting, followed by injection onto a C18 analytical column (75 μm × 150 mm, 3 μm particle size, 100 Å pore size, Eksigent/Sciex). Samples were run using a 100 min gradient from 5% to 80% solvent B (solvent A: 3% DMSO, 97% H2O, 0.1% formic acid; solvent B: 3% DMSO, 97% ACN, 0.1% formic acid), with a flow rate of 300 nL/min. The mass spectrometer was operated in information-dependent acquisition mode with an MS survey range of 350–1500 amu, followed by MS/MS scans of the 40 most intense ions. Dynamic exclusion was set for a period of 18 s and a tolerance of 120 cps. The MS/MS spectra were acquired in high sensitivity mode with the rolling collision energy and iTRAQ reagent collision energy adjustment on.

MS raw data were analyzed with ProteinPilot V4.5 using the Paragon database search algorithm (Sciex). Spectra files were searched against the mouse protein sequence database downloaded from UniProt using the following parameters: Sample Type was iTRAQ 4plex (peptide labeled); Cys alkylation was set to iodoacetamide; digestion was set to trypsin; quantitate was set to bias correction; background correction was enabled for specific processing; search effort was set to Rapid ID. Mass tolerance parameters were automatically set and optimized by ProteinPilot software (for the SCIEX TripleTOF 5600 data, the precursor mass tolerance was set to 0.05 Da and the fragment mass tolerance was set to 0.1 Da). A strict cutoff for an unused protein score ≥ 1.3 was used as the filter criterion for protein identification, which corresponds to less than 1% FDR on protein level (as indicated by ProteinPilot software). Protein quantification matrix was calculated by averaging the protein ratios in 3 replicates of Protein Summary table. P values of T test were calculated using log2-transformed ratios from 3 replicates. Significantly differentially expressed proteins (DEPs) were determined by the P value from Student’s t-test and fold change thresholds. Functional annotations were performed using Gene Ontology (GO), the Kyoto Encyclopedia of Genes and Genomes (KEGG) databases. A Fisher’s Exact Test was performed to reveal annotations and pathways. Hierarchical clustering in the heatmap was performed on log2-transformed and normalized intensities. The clusterProfiler package (v3.14.3) was employed for functional enrichment analysis.

RNA immunoprecipitation (RIP) assay

The cells were harvested and subjected to RNA-binding protein immunoprecipitation (RIP) with the RIP Kit (Genecreate) using specific antibodies or IgG isotype control, according to the manufacturer’s instructions. The precipitated RNA was reverse transcribed into cDNA and subjected to qRT-PCR analysis.

Generation of ZIKV mutants

The infectious clone of ZIKV FSS13025 was kindly provided by Dr. Pei-Yong Shi. All mutants (E67A and P101A) were introduced into infectious clone of FSS13205 individually by using a Mut Express II Fast Mutagenesis Kit V2 (Vazyme Biotech). Briefly, the plasmid is amplified by reverse PCR using a pair of primers (P101A-Forward: 5′-AATTGGACAGCCCGTGAGAGCATGCTGCTGGC-3′, P101A-Reverse: 5′-TCACGGGCTGTCCAATTAGCTCTGAAAATGAAA-3′, E67A-Forward: 5′-CTTCGCGGCAATGAACACTGGAGGAGATGTTGC, E67A-Reverse: 5′-TGTTCATTGCCGCGAAGGTGGCACCCATCAAA-3′), followed by digestion of the amplification product using Dpn1, and cyclization recombination reaction of the amplification product using Exnase II. The recombinant product is transformed into DH5α competent cells and the correct clone is selected. The substitution sites were confirmed by DNA sequencing. The infectious clone plasmids were linearized with restriction enzyme ClaI (engineered immediately downstream of the 3′ end of the viral genomic sequence) digestion and purified using the Gel Extraction Kit (Omega). In vitro transcribed genome-length RNA was prepared using HiScribe T7 ARCA mRNA Kit (New England Biolabs) and purified using Monarch RNA Cleanup Kit (New England Biolabs). The RNA was then transfected into Vero cells using Lipofectamine 3000 reagent (Thermo Fisher Scientific) to generate the virus. Six hours later, culture fluids were replaced with fresh medium to remove extracellular input RNAs. After incubation at 37 °C for 24 h, cells were cultured in DMEM with 2% FBS at 30 °C for another 4 days. At various time points, the culture fluids from viral RNA-transfected cells were collected and filtered with 0.22 μm filter, then stored at −80 °C. Extracellular viral titers were determined by plaque assay on BHK-21 cells. Further, viral antigen expression in ZIKV RNA-transfected Vero cells was detected with Confocal microscopy. The culture fluids were used to extract extracellular viral RNA using the QIAamp viral RNA minikit (Qiagen). The total cellular RNAs were extracted by RNeasy minikit (Qiagen). Quantitative RT-PCR was performed according to the protocol described previously in ref. ⁴⁷.

Statistical analyses

All analyses were performed using GraphPad Prism software (Version 8.0). Measurement data are presented as the mean ± s.d. Statistical differences of normally distributed data were determined using one-way analysis of variance (ANOVA) or with two-way ANOVA statistical tests, followed by Tukey’s post hoc test. P values less than 0.05 were considered significant. ns not significant. Quantitation of western blotting was measured by Image J software (1.52p/java 1.8.0 172).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The full-genome sequencing data for ZIKV^WT, ZIKV^E67A, and ZIKV^P101A generated in this study have been deposited to the GenBank database under accession number: ZIKV^WT (PQ849442), ZIKV^E67A (PQ849443), and ZIKV^P101A (PQ849444). The mass spectrometry proteomics data have been submitted to the ProteomeXchange Consortium with the dataset identifier PXD055433, with differentially expressed proteins detailed in Supplementary Data 1. Metabolomics data have been deposited in the MetaboLights database under the accession number MTBLS12549 [https://www.ebi.ac.uk/metabolights/reviewerb80bfff7-4d2c-4882-a1f0-17859a1ed9d4], and the output results are provided in Supplementary Data 2. Source data are provided with this paper.

References

Petersen, L. R., Jamieson, D. J., Powers, A. M. & Honein, M. A. Zika virus. N. Engl. J. Med. 374, 1552–1563 (2016).
Article CAS PubMed Google Scholar
Pierson, T. C. & Diamond, M. S. The emergence of Zika virus and its new clinical syndromes. Nature 560, 573–581 (2018).
Article CAS PubMed Google Scholar
D’Ortenzio, E., Matheron, S. & Yazdanpanah, Y. Evidence of sexual transmission of Zika virus. N. Engl. J. Med. 374, 2195–2198 (2016).
Article PubMed Google Scholar
Govero, J. et al. Zika virus infection damages the testes in mice. Nature 540, 438–422 (2016).
Article CAS PubMed PubMed Central Google Scholar
Ma, W. et al. Zika virus causes testis damage and leads to male infertility in mice. Cell 168, 542–542 (2017).
Article CAS PubMed Google Scholar
Kurscheidt, F. A. et al. Persistence and clinical relevance of Zika virus in the male genital tract. Nat. Rev. Urol. 16, 211–230 (2019).
Article PubMed Google Scholar
Calvet, G. A. et al. Study on the persistence of Zika virus (ZIKV) in body fluids of patients with ZIKV infection in Brazil. BMC Infect. Dis. 18, 49 (2018).
Joguet, G. et al. Effect of acute Zika virus infection on sperm and virus clearance in body fluids: a prospective observational study. Lancet Infect. Dis. 17, 1200–1208 (2017).
Article PubMed Google Scholar
Ma, W. et al. Zika virus causes testis damage and leads to male infertility in mice. Cell 167, 1511–1524 e1510 (2016).
Article CAS PubMed Google Scholar
Matusali, G. et al. Zika virus infects human testicular tissue and germ cells. J. Clin. Investig. 128, 4697–4710 (2018).
Article PubMed PubMed Central Google Scholar
Mlera, L. & Bloom, M. E. Differential Zika virus infection of testicular cell lines. Viruses 11, 42 (2019).
Smith, L. B. & Walker, W. H. The regulation of spermatogenesis by androgens. Semin. Cell Dev. Biol. 30, 2–13 (2014).
Article CAS PubMed Google Scholar
Wang, R.-S., Yeh, S., Tzeng, C.-R. & Chang, C. Androgen receptor roles in spermatogenesis and fertility: lessons from testicular cell-specific androgen receptor knockout mice. Endocr. Rev. 30, 119–132 (2009).
Article PubMed PubMed Central Google Scholar
Walker, W. H. Non-classical actions of testosterone and spermatogenesis. Philos. Trans. R. Soc. B Biol. Sci. 365, 1557–1569 (2010).
Article CAS Google Scholar
Sar, M., Lubahn, D. B., French, F. S. & Wilson, E. M. Immunohistochemical localization of the androgen receptor in rat and human tissues. Endocrinology 127, 3180–3186 (1990).
Article CAS PubMed Google Scholar
Walker, W. H. Testosterone signaling and the regulation of spermatogenesis. Spermatogenesis 1, 116–120 (2011).
Article PubMed PubMed Central Google Scholar
Griswold, M. D. The central role of Sertoli cells in spermatogenesis. Semin. Cell Dev. Biol. 9, 411–416 (1998).
Article CAS PubMed Google Scholar
Robaire, B. & Hamzeh, M. Androgen action in the epididymis. J. Androl. 32, 592–599 (2011).
Article CAS PubMed Google Scholar
Odonnell, L., McLachlan, R. I., Wreford, N. G., deKretser, D. M. & Robertson, D. M. Testosterone withdrawal promotes stage-specific detachment of round spermatids from the rat seminiferous epithelium. Biol. Reprod. 55, 895–901 (1996).
Article CAS PubMed Google Scholar
Wong, C. H. et al. Regulation of ectoplasmic specialization dynamics in the seminiferous epithelium by focal adhesion-associated proteins in testosterone-suppressed rat testes. Endocrinology 146, 1192–1204 (2005).
Article CAS PubMed Google Scholar
Zirkin, B. R., Santulli, R., Awoniyi, C. A. & Ewing, L. L. Maintenance of advanced spermatogenic cells in the adult rat testis: quantitative relationship to testosterone concentration within the testis. Endocrinology 124, 3043–3049 (1989).
Article CAS PubMed Google Scholar
Liu, L.-B. et al. Zika virus infection leads to hormone deficiencies of the hypothalamic-pituitary-gonadal axis and diminished fertility in mice. J. Virol. 97, e0100623 (2023).
Clark, B. J. & Stocco, D. M. Steroidogenic acute regulatory protein: the StAR still shines brightly. Mol. Cell. Endocrinol. 134, 1–8 (1997).
Article CAS PubMed Google Scholar
Payne, A. H. & Hales, D. B. Overview of steroidogenic enzymes in the pathway from cholesterol to active steroid hormones. Endocr. Rev. 25, 947–970 (2004).
Article CAS PubMed Google Scholar
BurgosTrinidad, M. et al. Repression of cAMP-induced expression of the mouse P450 17 alpha-hydroxylase/C17-20 lyase gene (Cyp17) by androgens. Mol. Endocrinol. 11, 87–96 (1997).
CAS PubMed Google Scholar
Payne, A. H. & Youngblood, G. L. Regulation of expression of steroidogenic enzymes in Leydig cells. Biol. Reprod. 52, 217–225 (1995).
Article CAS PubMed Google Scholar
Li, L. et al. Hormone regulation in testicular development and function. Int. J. Mol. Sci. 25, 5805 (2024).
Olive, A. J. & Sassetti, C. M. Metabolic crosstalk between host and pathogen: sensing, adapting and competing. Nat. Rev. Microbiol. 14, 221–234 (2016).
Article CAS PubMed Google Scholar
Pang, H. et al. Aberrant NAD⁺ metabolism underlies Zika virus-induced microcephaly. Nat. Metab. 3, 1109–1124 (2021).
Article CAS PubMed Google Scholar
Miller, W. L. & Auchus, R. J. The molecular biology, biochemistry, and physiology of human steroidogenesis and its disorders. Endocr. Rev. 32, 81–151 (2011).
Article CAS PubMed Google Scholar
Van Den Akker, E. L. T. et al. Differential inhibition of 17α-hydroxylase and 17,20-lyase activities by three novel missense CYP17 mutations identified in patients with P450c17 deficiency. J. Clin. Endocrinol. Metab. 87, 5714–5721 (2002).
Article PubMed Google Scholar
Kardelen, A. D. et al. A rare cause of congenital adrenal hyperplasia: clinical and genetic findings and follow-up characteristics of six patients with 17-hydroxylase deficiency including two novel mutations. J. Clin. Res. Pediatr. Endocrinol. 10, 206–215 (2018).
Article PubMed PubMed Central Google Scholar
Ma, L. Z. et al. Combined 17 alpha-hydroxylase/17,20-lyase deficiency with short stature: case study. Gynecol. Endocrinol. 32, 264–266 (2016).
Article CAS PubMed Google Scholar
Danila, D. C. et al. TMPRSS2-ERG status in circulating tumor cells as a predictive biomarker of sensitivity in castration-resistant prostate cancer patients treated with abiraterone acetate. Eur. Urol. 60, 897–904 (2011).
Article CAS PubMed PubMed Central Google Scholar
Grist, E. & Attard, G. The development of abiraterone acetate for castration-resistant prostate cancer. Urol. Oncol. Semin. Orig. Investig. 33, 289–294 (2015).
CAS Google Scholar
Thibault, C. & Massard, C. New therapies in metastatic castration resistant prostate cancer. Bull. Cancer 102, 501–508 (2015).
Article PubMed Google Scholar
Shi, Y. & Gao, G. F. Structural biology of the Zika virus. Trends Biochem. Sci. 42, 443–456 (2017).
Article CAS PubMed Google Scholar
Zhang, X. et al. Zika virus NS2A-mediated virion assembly. Mbio 10, e02375–19 (2019).
Zhang, Z. et al. Crystal structure of unlinked NS2B-NS3 protease from Zika virus. Science 354, 1597–1600 (2016).
Article CAS PubMed Google Scholar
Ren, W. et al. Zika virus NS5 protein inhibits type I interferon signaling via CRL3 E3 ubiquitin ligase-mediated degradation of STAT2. Proc. Natl Acad. Sci. USA 121, e2403235121 (2024).
Article CAS PubMed PubMed Central Google Scholar
Hui, L. et al. Matrix metalloproteinase 9 facilitates Zika virus invasion of the testis by modulating the integrity of the blood-testis barrier. PLos Pathog. 16, e1008509 (2020).
Zirkin, B. R. & Papadopoulos, V. Leydig cells: formation, function, and regulation. Biol. Reprod. 99, 101–111 (2018).
Article PubMed PubMed Central Google Scholar
Jerabek-Willemsen, M. et al. MicroScale thermophoresis: Interaction analysis and beyond. J. Mol. Struct. 1077, 101–113 (2014).
Article CAS Google Scholar
Subtelny, A. O., Eichhorn, S. W., Chen, G. R., Sive, H. & Bartel, D. P. Poly(A)-tail profiling reveals an embryonic switch in translational control. Nature 508, 66 (2014).
Article CAS PubMed PubMed Central Google Scholar
Macarthur, M. W. & Thornton, J. M. Influence of proline residues on protein conformation. J. Mol. Biol. 218, 397–412 (1991).
Article CAS PubMed Google Scholar
Zhou, X. et al. Mutations linked to neurological disease enhance self-association of low-complexity protein sequences. Science 377, eabn5582 (2022).
Article CAS PubMed PubMed Central Google Scholar
Zhang, X. et al. Genetic and biochemical characterizations of Zika virus NS2A protein. Emerg. Microbes. Infect. 8, 585–602 (2019).
Article CAS PubMed PubMed Central Google Scholar
Lazear, H. M. et al. A mouse model of Zika virus pathogenesis. Cell Host Microbe 19, 720–730 (2016).
Article CAS PubMed PubMed Central Google Scholar
Miner, J. J. & Diamond, M. S. Zika virus pathogenesis and tissue tropism. Cell Host Microbe 21, 134–142 (2017).
Article CAS PubMed PubMed Central Google Scholar
Zeng, J. et al. The Zika virus capsid disrupts corticogenesis by suppressing dicer activity and miRNA biogenesis. Cell Stem Cell 27, 618–623.e9 (2020).
Article CAS PubMed PubMed Central Google Scholar
Zhou, J. et al. Zika virus degrades the omega-3 fatty acid transporter Mfsd2a in brain microvascular endothelial cells and impairs lipid homeostasis. Sci. Adv. 5, eaax7142 (2019).
Article CAS PubMed PubMed Central Google Scholar
Yoon, K. J. et al. Zika-virus-encoded NS2A disrupts mammalian cortical neurogenesis by degrading adherens junction proteins. Cell Stem Cell 21, 349–358 e346 (2017).
Article CAS PubMed PubMed Central Google Scholar
Guo, M., Hui, L., Nie, Y., Tefsen, B. & Wu, Y. ZIKV viral proteins and their roles in virus-host interactions. Sci. Chin. Life Sci. 64, 709–719 (2021).
Article CAS Google Scholar
de La Vega, M.-A. et al. Zika-induced male infertility in mice is potentially reversible and preventable by deoxyribonucleic acid immunization. J. Infect. Dis. 219, 365–374 (2019).
Article PubMed Google Scholar
Vossmann, S., Wieseler, J., Kerber, R. & Kuemmerer, B. M. A basic cluster in the N terminus of yellow fever virus NS2A contributes to infectious particle production. J. Virol. 89, 4951–4965 (2015).
Article CAS PubMed PubMed Central Google Scholar
Xie, X., Gayen, S., Kang, C., Yuan, Z. & Shi, P.-Y. Membrane topology and function of dengue virus NS2A protein. J. Virol. 87, 4609–4622 (2013).
Article CAS PubMed PubMed Central Google Scholar
Xie, X. et al. Dengue NS2A protein orchestrates virus assembly. Cell Host Microbe 26, 606–602.e8 (2019).
Article CAS PubMed Google Scholar
Zhang, L. et al. AMFR-mediated flavivirus NS2A ubiquitination subverts ER-phagy to augment viral pathogenicity. Nat. Commun. 15, 9578 (2024).
Article CAS PubMed PubMed Central Google Scholar
Garneau, N. L., Wilusz, J. & Wilusz, C. J. The highways and byways of mRNA decay. Nat. Rev. Mol. Cell Biol. 8, 113–126 (2007).
Article CAS PubMed Google Scholar
Warf, M. B. & Berglund, J. A. Role of RNA structure in regulating pre-mRNA splicing. Trends Biochem. Sci. 35, 169–178 (2010).
Article CAS PubMed Google Scholar
Kozak, M. Regulation of translation via mRNA structure in prokaryotes and eukaryotes. Gene 361, 13–37 (2005).
Article CAS PubMed Google Scholar
Wan, Y., Kertesz, M., Spitale, R. C., Segal, E. & Chang, H. Y. Understanding the transcriptome through RNA structure. Nat. Rev. Genet. 12, 641–655 (2011).
Article CAS PubMed Google Scholar
Ramanathan, M., Porter, D. F. & Khavari, P. A. Methods to study RNA-protein interactions. Nat. Methods 16, 225–234 (2019).
Article CAS PubMed PubMed Central Google Scholar
Castello, A., Fischer, B., Hentze, M. W. & Preiss, T. RNA-binding proteins in Mendelian disease. Trends Genet. 29, 318–327 (2013).
Article CAS PubMed Google Scholar
Hentze, M. W., Castello, A., Schwarzl, T. & Preiss, T. A brave new world of RNA-binding proteins. Nat. Rev. Mol. Cell Biol. 19, 327–341 (2018).
Article CAS PubMed Google Scholar
Mackenzie, J. M., Khromykh, A. A., Jones, M. K. & Westaway, E. G. Subcellular localization and some biochemical properties of the flavivirus Kunjin nonstructural proteins NS2A and NS4A. Virology 245, 203–215 (1998).
Article CAS PubMed Google Scholar
Qiu, Y. et al. Flavivirus induces and antagonizes antiviral RNA interference in both mammals and mosquitoes. Sci. Adv. 6, eaax7989 (2020).
Nie, Y. et al. Rearrangement of actin cytoskeleton by Zika virus infection facilitates blood-testis barrier hyperpermeability. Virol. Sin. 36, 692–705 (2021).
Article CAS PubMed PubMed Central Google Scholar
Teng, Y. et al. Novel function of SART1 in HNF4α transcriptional regulation contributes to its antiviral role during HBV infection. J. Hepatol. 75, 1072–1082 (2021).
Article CAS PubMed Google Scholar
Sun, J., Zhong, L., Zhu, Y. & Liu, G. Research on the isolation of mouse Leydig cells using differential digestion with a low concentration of collagenase. J. Reprod. Dev. 57, 433–436 (2011).
Article CAS PubMed Google Scholar
Hafner, M. et al. Transcriptome-wide identification of RNA-binding protein and microRNA target sites by PAR-CLIP. Cell 141, 129–141 (2010).
Article CAS PubMed PubMed Central Google Scholar
Yang, N. et al. Identification and characterization of proteins that are differentially expressed in adipose tissue of olanzapine-induced insulin resistance rat by iTRAQ quantitative proteomics. J. Proteomics 212, 103570 (2020).

Download references

Acknowledgements

This work was supported by the National Key Research and Development Plan of China (2021YFC2300200 to Ying Wu) and the National Natural Science Foundation of China (32122008, 82472277 to Ying Wu).

Author information

These authors contributed equally: Wei Yang, Hanyu Li.

Authors and Affiliations

State Key Laboratory of Virology and Biosafety and Hubei Province Key Laboratory of Allergy and Immunology, Institute of Medical Virology, TaiKang Medical School (School of Basic Medical Sciences), Wuhan University, Wuhan, China
Wei Yang, Hanyu Li, Shanshan Wang, Rui Huang, Yifei Zhang, Moujian Guo, Dingran Zhao, Yuxin Xiong, Yifei Liu, Mengjing Huang, Lixia Hui, Wei Xiao & Ying Wu
Research Center for Medicine and Structural Biology, TaiKang Medical School (School of Basic Medical Sciences), Wuhan University, Wuhan, China
Li Huang
CAS Key Laboratory of Pathogenic Microbiology and Immunology, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China
Shihua Li & Ruirui Yang

Authors

Wei Yang
View author publications
Search author on:PubMed Google Scholar
Hanyu Li
View author publications
Search author on:PubMed Google Scholar
Shanshan Wang
View author publications
Search author on:PubMed Google Scholar
Rui Huang
View author publications
Search author on:PubMed Google Scholar
Yifei Zhang
View author publications
Search author on:PubMed Google Scholar
Moujian Guo
View author publications
Search author on:PubMed Google Scholar
Li Huang
View author publications
Search author on:PubMed Google Scholar
Shihua Li
View author publications
Search author on:PubMed Google Scholar
Ruirui Yang
View author publications
Search author on:PubMed Google Scholar
Dingran Zhao
View author publications
Search author on:PubMed Google Scholar
Yuxin Xiong
View author publications
Search author on:PubMed Google Scholar
Yifei Liu
View author publications
Search author on:PubMed Google Scholar
Mengjing Huang
View author publications
Search author on:PubMed Google Scholar
Lixia Hui
View author publications
Search author on:PubMed Google Scholar
Wei Xiao
View author publications
Search author on:PubMed Google Scholar
Ying Wu
View author publications
Search author on:PubMed Google Scholar

Contributions

W.Y. and H.L. performed the experiments. S.W., R.H., Y.Z., M.G., R.Y., S.L., D.Z., Y. X., Y.L., M.H., and W.X. analyzed the results. Li.H. and Lx.H. contributed significantly to the metabolomics experiments and data analysis. Y.W. initiated and designed the project and directed the research. W.Y., H.L., and Y.W. wrote the manuscript. All authors discussed the results and approved the manuscript.

Corresponding author

Correspondence to Ying Wu.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Gilberto Domont, Philippe Despres, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Description Of Additional Supplementary File

Supplementary Data 1

Supplementary Data 2

Reporting summary

Transparent Peer Review file

Source data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Yang, W., Li, H., Wang, S. et al. Zika virus disrupts steroidogenesis and impairs spermatogenesis by stalling the translation of CYP17A1 mRNA. Nat Commun 16, 6756 (2025). https://doi.org/10.1038/s41467-025-62044-x

Download citation

Received: 19 August 2024
Accepted: 09 July 2025
Published: 22 July 2025
DOI: https://doi.org/10.1038/s41467-025-62044-x

Subjects

Abstract

Similar content being viewed by others

Introduction

Results

Metabolomics profiling reveals disruption in steroidogenesis in the testes of ZIKV-infected mice

ZIKV infection decreases CYP17A1 at the protein level

ZIKV-NS2A inhibits the production of CYP17A1 protein

ZIKV-NS2A sequestrates CYP17A1 mRNA and reduces its translation

The 1200-1400 nucleotides of CYP17A1 mRNA is responsible for binding to ZIKV-NS2A

Mutations at residues 67E/101P abolish the binding between ZIKV-NS2A and CYP17A1 mRNA and the inhibition of CYP17A1 translation

ZIKVE67A mutant virus exhibits significantly lower inhibition on steroidogenesis and spermatogenesis than ZIKVWT

Discussion

Methods

Ethics statement

Cells and viruses

Mouse experiments

Plaque assay

Antibodies

Plasmid construction

Adenovirus-associated virus (AAV) production

Isolation and culture of primary mouse Leydig cells (mLCs) and cell viability assay

Enzyme-linked immunosorbent assay (ELISA)

H&E analyses and immunofluorescence

Confocal microscopy

Real-time quantitative RT-PCR (qRT-PCR)

Western blotting analysis

Microscale thermophoresis technology (MST)

Electrophoretic mobility shift assay (EMSA)

Purification of GST recombinant protein

RNA-protein pulldown assay

Quantification of ribosome loading on to CYP17A1 transcripts

Photoactivatable-ribonucleoside-enhanced crosslinking and immunoprecipitation (PAR-CLIP)

Untargeted metabolism analysis

iTRAQ-LC-MS/MS analysis

RNA immunoprecipitation (RIP) assay

Generation of ZIKV mutants

Statistical analyses

Reporting summary

Data availability

References

Acknowledgements

Author information

Authors and Affiliations

Contributions

Corresponding author

Ethics declarations

Competing interests

Peer review

Peer review information

Additional information

Supplementary information

Source data

Rights and permissions

About this article

Cite this article

Share this article

Search

Quick links

ZIKV^E67A mutant virus exhibits significantly lower inhibition on steroidogenesis and spermatogenesis than ZIKV^WT