Enterovirus Infection Induces Massive Recruitment of All Isoforms of Small Cellular Arf GTPases to the Replication Organelles

Abstract

Enterovirus replication requires the cellular protein GBF1, a guanine nucleotide exchange factor for small Arf GTPases. When activated, Arfs associate with membranes, where they regulate numerous steps of membrane homeostasis. The requirement for GBF1 implies that Arfs are important for replication, but which of the different Arfs function(s) during replication remains poorly understood. Here, we established cell lines expressing each of the human Arfs fused to a fluorescent tag and investigated their behavior during enterovirus infection. Arf1 was the first to be recruited to the replication organelles, where it strongly colocalized with the viral antigen 2B and mature virions but not double-stranded RNA. By the end of the infectious cycle, Arf3, Arf4, Arf5, and Arf6 were also concentrated on the replication organelles. Once on the replication membranes, all Arfs except Arf3 were no longer sensitive to inhibition of GBF1, suggesting that in infected cells they do not actively cycle between GTP- and GDP-bound states. Only the depletion of Arf1, but not other class 1 and 2 Arfs, significantly increased the sensitivity of replication to GBF1 inhibition. Surprisingly, depletion of Arf6, a class 3 Arf, normally implicated in plasma membrane events, also increased the sensitivity to GBF1 inhibition. Together, our results suggest that GBF1-dependent Arf1 activation directly supports the development and/or functioning of the replication complexes and that Arf6 plays a previously unappreciated role in viral replication. Our data reveal a complex pattern of Arf activation in enterovirus-infected cells that may contribute to the resilience of viral replication in different cellular environments.IMPORTANCE Enteroviruses include many known and emerging pathogens, such as poliovirus, enteroviruses 71 and D68, and others. However, licensed vaccines are available only against poliovirus and enterovirus 71, and specific anti-enterovirus therapeutics are lacking. Enterovirus infection induces the massive remodeling of intracellular membranes and the development of specialized domains harboring viral replication complexes, replication organelles. Here, we investigated the roles of small Arf GTPases during enterovirus infection. Arfs control distinct steps in intracellular membrane traffic, and one of the Arf-activating proteins, GBF1, is a cellular factor required for enterovirus replication. We found that all Arfs expressed in human cells, including Arf6, normally associated with the plasma membrane, are recruited to the replication organelles and that Arf1 appears to be the most important Arf for enterovirus replication. These results document the rewiring of the cellular membrane pathways in infected cells and may provide new ways of controlling enterovirus infections.

Keywords: GBF1; GTPases Arf; RNA replication; enteroviruses; membrane metabolism; picornavirus; replication organelles.

FIG 1

Characterization of cell lines expressing individual Arf-Venus fusions. (A) HeLa cells stably expressing corresponding Arf-Venus fusions were analyzed with antibodies against cis- and trans-Golgi markers ERGIC53 and GM130, respectively. Plasma membrane was stained with wheat germ agglutinin (WGA). Nuclear DNA was stained by Hoechst 33342 (blue). Fluorescently tagged Arfs demonstrated the expected behavior. (B) Expression of fluorescently tagged Arfs does not interfere with poliovirus replication. The original HeLa cell line and its derivatives stably expressing individual Arf-Venus fusions were mock infected or infected with poliovirus (50 PFU/cell), the cells were collected at 6 h p.i. in radioimmunoprecipitation assay (RIPA) buffer, and the total cell lysates were analyzed by Western blotting with an antibody against viral antigen 2C and anti-GFP antibodies that recognize the Arf-Venus fusions. Actin is shown as a loading control. (C) Arf-Venus fusions are not significantly affected by poliovirus infection. Cell lines expressing corresponding Arf-Venus fusions were infected with poliovirus (PV; 50 PFU/cell) or mock infected (M), the cells were collected at 6 h p.i. in RIPA buffer, and the total cell lysates were analyzed by Western blotting with an antibody against viral antigen 2C and anti-GFP antibodies that recognize the Arf-Venus fusions. Actin is shown as a loading control.

FIG 2

Recruitment of Arfs to the replication organelles in poliovirus-infected cells. HeLa cell lines expressing individual Arf-Venus fusions were mock infected or infected with poliovirus type I Mahoney at 50 PFU/cell and fixed at the indicated times postinfection. To maximally preserve the pattern of Arf recruitment, no other manipulations with cells were performed. For quantifications, several random fields with no fewer than 100 total cells were counted. The cellular phenotypes were defined as Arf1 to -4, mostly at the Golgi membrane (Arf6 was at the plasma membrane), Arfs mostly distributed at the cytoplasmic dots, and Arfs mostly recruited to the perinuclear rings of replication organelles (RO).

FIG 3

Recruitment of Arfs to the enterovirus but not cardiovirus replication organelles. HeLa cell lines expressing individual Arf-Venus fusions were mock infected or infected with 50 PFU/cell of coxsackievirus B3 (an enterovirus) (A) or encephalomyocarditis virus (a cardiovirus) (B) and fixed at 6 and 8 h p.i. The cells were counterstained for dsRNA, an intermediate product of replication of RNA viruses, to confirm the infection.

FIG 4

Arf1 fusions with red and green fluorescent proteins recapitulate the same Arf1 behavior. HeLa cells coexpressing Arf1-Venus (green) and Arf1-FRP (red) fusions were mock infected or infected with poliovirus type I Mahoney at 50 PFU/cell, and the cells were fixed at the indicated times postinfection. Nuclear staining was performed with a cell-permeable dye, Hoechst 33342, in live cells 30 min before fixation. To maximally preserve the pattern of Arf recruitment, no other manipulations with cells were performed.

FIG 5

Arf1 is the first to be recruited to the replication organelles. (A) HeLa cells coexpressing pairs of Arf1-FRP (red) with other Arf-Venus (green) fusions were infected with poliovirus type I Mahoney at 50 PFU/cell and fixed at 3 h p.i. To maximally preserve the pattern of Arf recruitment, no other manipulations with cells were performed. (B) HeLa cells coexpressing pairs of Arf1-FRP (red) with other Arf-Venus (green) fusions were infected with poliovirus type I Mahoney at 50 PFU/cell and fixed at 4 h p.i. The cells were stained with antibody A12 that recognizes only mature poliovirus capsids (35), and early in infection denotes the localization of the active replication organelles, since the assembly of poliovirus virions is intimately linked to active RNA replication (52). Nuclear DNA was stained by Hoechst 33342 (blue).

FIG 6

Poliovirus antigens differentially associate with Arf-enriched membranes. (A) Scheme of the processing of poliovirus polyprotein. (B) Mock-infected HeLa cells expressing Arf1-Venus fusion. (C) HeLa cells expressing Arf1-Venus fusion were infected with poliovirus type I Mahoney at 50 PFU/cell, fixed at 4 h p.i., and stained with antibodies recognizing dsRNA, mature poliovirus capsid (A12) (35), or the antigens in the viral nonstructural protein 2B, 3A, 2C, or 3D. (D) HeLa cells expressing Arf1-Venus fusion were infected with 50 PFU/cell of a poliovirus type I Mahoney mutant with FLAG-Y inserted in the nonstructural protein 3A (30). The cells were stained with anti-FLAG and anti-2C antibodies simultaneously. Nuclear DNA was stained by Hoechst 33342 (blue).

FIG 7

Only Arf3 undergoes active cycling through GTP- and GDP-bound forms on the replication organelles. (A) Scheme of the experiment with the addition of brefeldin A (BFA), an inhibitor of GBF1, to the infected cells upon the formation of well-developed replication organelles. (B) Scheme of Arf cycling through membrane-bound GTP- and cytoplasmic GDP-bound forms. The guanidine nucleotide activating factors (GEFs) may be inhibited by BFA, while the GTPase-activating proteins (GAP) are still functional. (C) HeLa cells expressing individual Arf-Venus fusions were mock infected or infected with poliovirus type I Mahoney at 50 PFU/well, incubated for 5 h, and then incubated for 1 h in the presence of 1 μg/ml of BFA and fixed. No other manipulations with cells were performed to maximally preserve the pattern of Arf recruitment.

FIG 8

Arf1 significantly increases the sensitivity of polio replication to GBF1 inhibition. (A) Specificity and efficacy of the Arf isoform-targeting siRNAs. HeLa cell lines expressing individual Arf-Venus fusions were transfected with isoform-specific siRNAs (e.g., 1A corresponds to anti-Arf1 siRNA) or scrambled siRNA control (C). The cells were lysed on the third day after siRNA transfection and analyzed by Western blotting with anti-GFP antibodies that recognize Arf-Venus fusions. The samples reflecting the effect of a specific siRNA on a cognate cell line (e.g., cells expressing Arf1-Venus treated with anti-Arf1 or control siRNAs) are highlighted. Actin is shown as a loading control. The siRNAs shown are those taken for further replication assay. (B) HeLa cells were treated with Arf isoform-specific siRNAs (or scrambled siRNA; siC) and, on the third day after siRNA transfection, were transfected with a poliovirus replicon RNA coding for a Renilla luciferase gene. (Left) The cells were incubated with the indicated concentrations of BFA, and the Renilla signal was monitored in live cells every hour for 18 h after replicon RNA transfection (kinetic curves). (Middle) The total replication signal was calculated as the area under the curve, and normalization was performed for each sample for the replication signal in the absence of the inhibitor. (Right) A parallel sample of siRNA-transfected cells was used for cell viability assay. P values are indicated: ****, P < 0.0001; ***, P < 0.001. RLU, relative light units.