ANP32 proteins from ticks and vertebrates are key host factors for replication of Bourbon virus across species

Abstract

Bourbon virus (BRBV) is a tick-borne virus in the genus Thogotovirus in the Orthomyxoviridae family. BRBV was initially identified as the presumptive causative agent of a fatal human infection in 2014 and has since been identified in ticks in the Midwest, Northeast, and Southern United States, with occasional spillovers into humans. However, little is known about how virus-host interactions impact their large host range. Here, we show that BRBV polymerase activity in human cells is completely dependent on cellular ANP32 proteins. BRBV polymerase activity was completely lost in cells lacking ANP32A and ANP32B, resulting in failed infections. BRBV polymerase activity was restored in the presence of ANP32 proteins from diverse hosts. Dhori virus and Thogoto virus, other related Thogotovirus members, retained high activity in the absence of ANP32 proteins, showing reduced dependence on these host factors. Interaction studies revealed that the BRBV polymerase trimer binds human ANP32A or ANP32B. Genetic analysis revealed that tick vectors for BRBV encode a single ANP32 locus corresponding to ANP32A. Tick ANP32A produces multiple protein variants through alternative splicing and start-site selection, all of which enhance polymerase activity for Thogotoviruses. Unexpectedly, the BRBV polymerase was highly sensitive to changes at the N-terminus of ANP32, while it was insensitive to changes in the body of ANP32 that restrict the activity of influenza virus polymerases. Thus, ANP32A is a deeply conserved pro-viral cofactor, and Thogotoviruses show remarkable plasticity utilizing ANP32 homologs from different hosts separated by almost 1 billion years of evolution.IMPORTANCEViral polymerases rely on cellular cofactors to support efficient transcription of viral genes and replication of the viral genome. The RNA-dependent RNA polymerase of influenza virus, an orthomyxovirus, requires the cellular ANP32A or ANP32B proteins for genome replication. However, little is known about whether ANP32 proteins are required by other orthomyxovirus family members, like the tick-borne thogotoviruses. We show that thogotoviruses use ANP32 proteins from diverse hosts to enhance polymerase activity, including that encoded by the single ANP32A gene found in ticks. However, thogotovirus polymerase showed varying levels of dependence on ANP32 proteins, with some polymerases functioning at near full activity even in the absence of ANP32 proteins. Thus, ANP32 proteins are deeply conserved viral cofactors, with each virus displaying distinct patterns of ANP32 usage and requirements for function.

Keywords: Bourbon virus; anp32; host range; influenza virus; orthomyxovirus; thogotovirus; tick; virus:host interactions.

Fig 1

Human ANP32A and ANP32B are required host factors for thogotovirus polymerase activity and replication. (A–F) ANP32A and ANP32B increase polymerase activity for diverse orthomyxoviruses. Polymerase activity assays were performed in WT or knockout cells by expressing the polymerase proteins, nucleoprotein, and a viral firefly luciferase reporter for the indicated virus. Polymerase activity was normalized to an internal Renilla luciferase control and reported as the ratio of firefly/Renilla luciferase. FLUCV, influenza C virus. The polymerase subunit PB2 was expressed with a FLAG epitope tag. PB2 and the loading control β-actin were detected by western blotting. (G and H) Multicycle replication of BRBV (multiplicity of infection [MOI] = 0.1) or DHOV (MOI = 1) in 293T and knockout cells. Titers determined by focus-forming assay. LOD, limit of detection. (A–F) Data are mean of n = 3 ± sd and significance was assessed by a one-way analysis of variance (ANOVA) with a post hoc Dunnett’s multiple comparisons test against WT 293T cells. (G and H) Data are mean of n = 6 ± sd with comparisons to DKO cells made by a two-way ANOVA with a Tukey’s post hoc test. **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001; ns, not significant.

Fig 2

Human ANP32A or ANP32B is sufficient for thogotovirus polymerase activity and replication. (A–F) Polymerase activity was measured in WT 293T cells, DKO cells, or DKO cells stably complemented with human ANP32A or ANP32B. Polymerase activity assays were performed by expressing the polymerase proteins, nucleoprotein, and a viral firefly luciferase reporter for the indicated virus. Polymerase activity was normalized to an internal Renilla luciferase control. (G and H) Human ANP32A and ANP32B enhance BRBV polymerase activity in a dose-dependent manner. Polymerase activity was measured in DKO cells as in (A) in the presence of increasing amounts of expression vector for (G) human ANP32A or (H) human ANP32B. PB2-FLAG, ANP32-FLAG, and the loading control β-actin were detected by western blotting. (I) Multicycle replication of BRBV in 293T cells, DKO cells, or DKO cells stably complemented with human ANP32A or human ANP32B (MOI = 0.01). Titers were determined by focus-forming assay. (A–H) Data are mean of n = 3 ± sd. Significance was assessed by a one-way ANOVA with a post hoc Dunnett’s multiple comparisons test against WT 293T cells (A–F) or empty vector controls (G and H). (I) Data are mean of n = 6 ± sd with comparisons to DKO cells made by a two-way ANOVA with a Tukey’s post hoc test. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001; ns, not significant.

Fig 3

BRBV polymerase binds ANP32 and requires its C-terminal low complexity acidic region. (A) Schematic diagram of full-length human ANP32B and three C-terminal truncations at amino acid 160 (T160), 190 (T190), and 220 (T220). Domains and boundaries are indicated. LRR = leucinerich repeat; LCAR = low complexity acidic region. T160, T190, and T220 = terminal amino acids in C-terminal truncations. (B) Human ANP32A (hu32A) interacts with the BRBV polymerase trimer. Human ANP32A-FLAG was expressed in cells with the BRBV polymerase or versions lacking PB2 or PA. Human ANP32A-FLAG was immunoprecipitated from whole-cell lysate and probed for huANP32A or V5-tagged PB2 and PA. Input samples were blotted as an expression control. β-actin served as a loading control. (C) The LCAR of human ANP32B (hu32B) is important for interaction with the BRBV polymerase. Polymerase proteins were expressed in DKO cells in the presence of human ANP32B-FLAG, the indicated truncations, or a negative control. Interaction with the polymerase was tested by immunoprecipitating huANP32B-FLAG from whole-cell lysate and probing for huANP32B and co-precipitating PB2 and PA that contained a C-terminal V5 tag. Input samples were blotted as an expression control. β-actin served as a loading control. (D) BRBV polymerase activity assays were performed in DKO cells expressing WT or truncated huANP32B, or an empty vector control. Polymerase activity was normalized to an internal Renilla luciferase control. Data are mean of n = 3 ± sd. Significance was assessed by a one-way ANOVA with a post hoc Dunnett’s multiple comparisons test against cells expressing WT huANP32B. ****P ≤ 0.0001.

Fig 4

Ticks encode a single ancestral locus expressing ANP32A variants that support BRBV polymerase activity. (A) Syntenic gene blocks were identified for ANP32 genes, revealing a single ancestral locus present in the lone star (Aam) ticks and brown dog (Rsa) ticks. Comparisons were made to humans (Hsa, Homo sapiens), chickens (Gga), swine (Ssc), and cows (Bta). (B) De novo transcriptome assembly identifies multiple splice variants of tick ANP32. The genomic locus for Rsa and Aam ticks is indicated, as well as accession numbers for previously cataloged transcript variants. Exons are shown as boxes with non-coding regions in pink, open reading frames in red, and a small upstream open reading frame (uORF) in gray. The conserved start site present in vertebrate ANP32A is indicated. (C) BRBV polymerase activity assays were performed in DKO cells expressing the viral polymerase, nucleoprotein, and a viral reporter as well as the indicated ANP32A or ANP32B from humans (hu), cows (co), chickens (ch), lone star ticks (Aam), or brown dog ticks (Rsa). Naming of tick variants corresponds to those diagrammed in (B). Activity was normalized to an internal Renilla luciferase control. (D) Polymerase activity assays were performed for DHOV and THOV as in (C). (E) Activity of the FLUAV polymerase in cells expressing diverse ANP32 proteins was measured as described for (C). For (C)–(E), data are mean of n = 3 ± sd. Significance was assessed by a one-way ANOVA with a post hoc Dunnett’s multiple comparisons test against the empty vector. **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001; ns, not significant. (F) Expression of ANP32 proteins was detected by western blot. Tubulin was probed as a loading control.

Fig 5

BRBV polymerase activity is sensitive to changes at the N-terminus of ANP32. (A) Alignment of the N-terminus of human ANP32A (hu32A), ANP32B (hu32B), and lone star tick (Aam) ANP32A (Aam 32). Mutants used below are highlighted in red. (B–D) Changes to the N-terminus of ANP32 disrupt its ability to support BRBV polymerase. (B–G) Polymerase activity assays for the indicated polymerase were performed in DKO cells expressing huANP32A with a C-terminal FLAG tag (C-FLAG), an N-terminal FLAG tag (N-FLAG), a C-terminal FLAG tag with an N-terminal deletion (∆ME) or mutation (E2A), or an empty vector control. Polymerase lacking the PB2 subunit (∆PB2) served as a negative control. (H–I) Polymerase activity assays were performed with the indicated polymerase in DKO cells expressing huANP32B with a C-terminal FLAG tag (C-FLAG), an N-terminal FLAG tag (N-FLAG), a C-terminal FLAG tag with an N-terminal deletion (∆MD) or mutation (D2A), or an empty vector control. Polymerase lacking the PB2 subunit (∆PB2) served as a negative control. For all, polymerase activity was normalized to an internal Renilla luciferase control. Data are mean of n = 3 ± sd. Significance was assessed by a one-way ANOVA with a post hoc Dunnett’s multiple comparisons test against huANP32 C-FLAG. ****P ≤ 0.0001; ns, not significant. (J) Representative blot of ANP32 proteins and variants confirming equivalent expression. Tubulin was probed as a loading control.

Fig 6

The impact of adaptive variants in chicken ANP32A and ANP32B differs between BRBV and FLUAV polymerases. (A–D) Polymerase activity assays were performed in DKO expressing the indicated viral polymerase, nucleoprotein, and a viral reporter. Cells were complemented with human (huANP32), chicken (chANP32), chANP32A lacking residues 175-207 (ch∆33), or mutant huANP32B N129I/D130N proteins. Activity was normalized to an internal Renilla luciferase control. Data are mean of n = 3 ± sd. Significance was assessed by a one-way ANOVA with a post hoc Dunnett’s multiple comparisons. ANP32A variants were compared against huANP32A, and ANP32B variants were compared against huANP32B. *P ≤ 0.05; ***P ≤ 0.001; ns, not significant. (E) Representative blot confirming expression of ANP32 variants. Tubulin was detected as a loading control.

Fig 7

Amino acid variations in the N-terminus and central regions of ANP32E preclude its use by BRBV or FLUAV, respectively. (A) Diagram of chimeric proteins between human ANP32A and ANP32E. Fusions were made after residues 120 and 170. (B) BRBV polymerase, nucleoprotein, and a viral reporter were expressed in DKO cells with the indicated human ANP32A, ANP32E, or chimeric clone. Activity was normalized to an internal Renilla luciferase control. (C) FLUAV polymerase activity was measured as in (B). (D) Representative blot showing expression of ANP32 proteins and chimeras. Tubulin was detected as a loading control. Data are mean of n = 3 ± sd. Significance was assessed by a one-way ANOVA with a post hoc Dunnett’s multiple comparisons test against human ANP32A or ANP32E. ***P ≤ 0.001; ****P ≤ 0.0001; ns, not significant.

Fig 8

ANP32 usage is a deeply conserved feature of Orthomyxoviridae. Schematic of the relationship between different species of ANP32 proteins and Orthomyxoviridae family members, highlighting the promiscuous usage by Thogotoviruses.