Microcephaly protein ANKLE2 promotes Zika virus replication

Abstract

Orthoflaviviruses are positive-sense single-stranded RNA viruses that hijack host proteins to promote their own replication. Zika virus (ZIKV) is infamous among orthoflaviviruses for its association with severe congenital birth defects, notably microcephaly. We previously mapped ZIKV-host protein interactions and identified the interaction between ZIKV non-structural protein 4A (NS4A) and host microcephaly protein ankyrin repeat and LEM domain-containing 2 (ANKLE2). Using a fruit fly model, we showed that NS4A induced microcephaly in an ANKLE2-dependent manner. Here, we explore the role of ANKLE2 in ZIKV replication to understand the biological significance of the interaction from a viral perspective. We observe that ANKLE2 localization is drastically shifted to sites of NS4A accumulation during infection and that knockout of ANKLE2 reduces ZIKV replication in multiple human cell lines. This decrease in virus replication is coupled with a moderate increase in innate immune activation. Using microscopy, we observe dysregulated formation of virus-induced endoplasmic reticulum rearrangements in ANKLE2 knockout cells. Knockdown of the ANKLE2 ortholog in Aedes aegypti cells also decreases virus replication, suggesting ANKLE2 is a beneficial replication factor across hosts. Finally, we show that NS4A from four other orthoflaviviruses physically interacts with ANKLE2 and is also beneficial to their replication. Thus, ANKLE2 likely promotes orthoflavivirus replication by regulating membrane rearrangements that serve to accelerate viral genome replication and protect viral dsRNA from immune detection. Taken together with our previous results, our findings indicate that ZIKV and other orthoflaviviruses hijack ANKLE2 for a conserved role in replication, and this drives unique pathogenesis for ZIKV since ANKLE2 has essential roles in developing tissues.IMPORTANCEZIKV is a major concern due to its association with birth defects, including microcephaly. We previously identified a physical interaction between ZIKV NS4A and host microcephaly protein ANKLE2. Mutations in ANKLE2 cause congenital microcephaly, and NS4A induces microcephaly in an ANKLE2-dependent manner. Here, we establish the role of ANKLE2 in ZIKV replication. Depletion of ANKLE2 from cells significantly reduces ZIKV replication and disrupts virus-induced membrane rearrangements. ANKLE2's ability to promote ZIKV replication is conserved in mosquito cells and for other related mosquito-borne orthoflaviviruses. Our data point to an overall model in which ANKLE2 regulates virus-induced membrane rearrangements to accelerate orthoflavivirus replication and avoid immune detection. However, ANKLE2's unique role in ZIKV NS4A-induced microcephaly is a consequence of ZIKV infection of important developing tissues in which ANKLE2 has essential roles.

Keywords: ANKLE2; NS4A; Zika virus; microcephaly; orthoflavivirus; virus-host interaction.

Fig 1

ANKLE2 co-localizes with ZIKV proteins during infection. (A–C) Immunofluorescence confocal microscopy of HEK293T cells expressing either ANKLE2-, GFP-, or ANKLE1-FLAG and infected with designated ZIKV strain at a multiplicity of infection (MOI) of 5 for 48 hours. Nucleus observed via Hoechst 33342 staining. (D) Pearson’s correlation was used to quantify degree of co-localization between FLAG and NS4A, n = 16–24 cells per condition. ****P < 0.0001, one-way analysis of variance with Šidák multiple comparison test. Gray circles represent the values of individual cells. (E) Microscopy of HEK293T cells infected with ZIKV MR766 at an MOI of 5 for 48 hours and evaluated for ZIKV E and endogenous ANKLE2, n = 27 cells. All scale bars = 5 µm.

Fig 2

Knockout of ANKLE2 reduces ZIKV replication in vitro across multiple human cell lines. (A) Control Huh7 cells (H-ncg) and clonal ANKLE2 KO Huh7 cells (H1 and H2) were generated with CRISPR and validated by immunofluorescence confocal microscopy of ANKLE2. (B) Western blot of ANKLE2 and related proteins PPP2R1A and VRK1 in control and ANKLE2 KO cells. (C) Western blot of viral proteins following ZIKV PLCal infection of control and ANKLE2 KO cells. (D–G) Virus titer was determined by plaque assay following infection of control and ANKLE2 KO cells. All statistical comparisons are made to negative control (H-ncg) cells. (H) Western blot of control (SK-ncg) and ANKLE2 KO (SK-g1) SK-N-SH cells. (I) Virus titers measured by plaque assay following infection of control and ANKLE2 KO cells. (J) Western blot of control (J-ncg) and clonal ANKLE2 knockout (J1 and J2) JEG-3 cells following infection with ZIKV PLCal at an MOI of 0.1. (K) Virus titers measured by plaque assay following infection of control and ANKLE2 KO cells. (L) Western blot of JEG-3 cells transduced with mCherry (mCh) or ANKLE2-mCherry-3×FLAG (A2) lentivirus to overexpress or restore ANKLE2 expression. (M) Confocal microscopy of A2 rescue cells. (N) Virus titers measured by plaque assay from rescue cells infected with ZIKV PLCal at an MOI of 0.1 for 72 hours. Student’s two-tailed t-test: n.s., not significant; **P < 0.01, ***P < 0.001, ****P < 0.0001. Bars represent average values + standard deviations of three technical replicates.

Fig 3

Knockout of ANKLE2 reduces ZIKV genome replication and enhances innate immune response activation. (A) RT-qPCR to assess ZIKV in ANKLE2 KO Huh7, JEG-3, and SK-N-SH cells infected with ZIKV PLCal at an MOI of 0.1 for 48 hours. Student’s unpaired t-test compared to corresponding normalized ncg control (dashed line). (B–D) Entry and internalization assay was performed after incubation with ZIKV PRVABC59 at an MOI of 2. RNA or virus titers were evaluated by RT-qPCR (C) or plaque assay (D). (E–H) RT-qPCR to assess innate immune responses in cells infected with ZIKV PRVABC59 at an MOI of 5 for 24 hours. Relative expression shown compared to the same cell line under mock infection conditions. Student’s unpaired t-test: n.s., not significant; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Bars represent average values + standard deviations of three biological replicates.

Fig 4

ZIKV membrane rearrangements are dysregulated in ANKLE2 knockout Huh7 cells. (A) TEM images of cells that were either mock-infected (top) or infected with ZIKV PRVABC59 at an MOI of 1 for 48 hours (bottom). Stereotypical orthoflavivirus replication organelles were consistently observed in infected H-ncg cells, including virus-induced replication vesicles (Ve) and convoluted membranes (CM). The number of cells with observed CM and Ve are noted in the top left corners. Other abbreviations: ER, endoplasmic reticulum, M, mitochondrion, Nuc, nucleus. (B–D) H-ncg or H2 cells were transfected with a plasmid expressing ZIKV NS2B-3-4A-2K-4B for 48 hours. Induced membrane rearrangements were evaluated by confocal microscopy (C) or quantified using a semi-automated imaging pipeline (D). Data represent four biological replicates. Black lines indicate median NS4A aggregate size; dashed line indicates limit of detection set to ≥4 pixels of signal. Student’s unpaired t-test: ***P < 0.001.

Fig 5

Silencing of ANKLE2 ortholog in mosquito Aag2 cells reduces ZIKV replication. (A) Mosquito Aag2 cells were transfected with either GFP- or ANKLE2 (LOC5576059, UniProtID A0A6I8U7J4) ortholog-targeting dsRNA for 48 hours prior to RT-qPCR or ZIKV infection. (B) ANKLE2 ortholog expression was measured by RT-qPCR. Eight technical replicates across two biological experiments. (C–E) Virus titers were measured by plaque assay for GFP and ANKLE2 ortholog knockdown cells for the noted ZIKV strain and MOI. Four knockdown/infection technical replicates were performed for each condition. All statistical comparisons are made to the corresponding GFP condition. Student’s unpaired t-test: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig 6

The NS4A-ANKLE2 interaction is conserved across mosquito-borne orthoflaviviruses, but the impact of ANKLE2 in virus replication varies. (A) NS4A total amino acid biochemical similarity using EMBOSS Needle (46). (B) HEK293T cells were co-transfected with ANKLE2-3×FLAG and NS4A-2×Strep from the corresponding orthoflavivirus. FLAG affinity purification (AP) and Western blotting were performed to determine physical interaction between proteins. (C–H) Huh7 or SK-N-SH ANKLE2 KO cells were infected with indicated viruses at an MOI of 0.1 for 72 hours. Virus titers were measured using plaque assay. All values shown are means ± standard deviations of three technical replicates. All statistical tests compared to corresponding ncgRNA condition. Student’s t-test: n.s., not significant. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. ncg, negative control CRISPR gRNA control.