The following primary antibodies were used: rabbit anti-LC3 (Novus Biologicals, Littleton, CO), mouse anti-LC3 (Nanotools, Teningen, Germany), mouse anti-p62 (Abcam, Cambridge, UK), mouse anti-tubulin (Sigma-Aldrich, St. Louis, MO), rabbit anti-ATG13 (Sigma-Aldrich), mouse anti-HSBP1 (Sigma-Aldrich, clone 2C3), rabbit anti-ATG16L1 (MBL, Woburn, MA), mouse anti-actin (Merck, Darmstadt, Germany), mouse anti-GFP (Clontech, Shiga, Japan), rabbit anti-ULK1 (Santa Cruz, Dallas, TX), rabbit anti-FIP200 (Bethyl Laboratories, Montgomery, TX), mouse anti-enterovirus (DAKO, Glostrup, Denmark), mouse anti-dsRNA (English & Scientific Consulting Bt., Budapest, Hungary), rabbit anti-capsid (EMCV, a kind gift from Ann Palmenberg, University of Wisconsin, Madison, WI), mouse anti-VP1 (EMCV, a kind gift from Hanchun Yang, China Agricultural University, Beijing, China), mouse anti-4G2 (Dengue virus (DENV) and Zika virus (ZIKV), Merk Millipore, Billerica, MA) and rabbit anti-E1 (Chikungunya virus (CHIKV), from Jolanda Smit, University Medical Center Groningen, The Netherlands), mouse anti-NP (influenza A virus (IAV), BioRad, Hercules, CA). The secondary antibodies were from Thermo Fisher Scientific (Waltham, MA) and were AlexaFluor488-conjugated goat anti-mouse or chicken anti-rabbit; AlexaFluor568-conjugated goat anti-mouse or donkey anti-rabbit; AlexaFluor680-conjugated goat anti-mouse or goat anti-rabbit; and AlexaFluor800-conjugated goat anti-mouse secondary antibodies were used for the visualization of the primary antibodies, and they were all from Thermo Fisher Scientific (Waltham, MA). Hoechst33342 was from Sigma Aldrich while bafilomycin A₁ (BafA1) was from BioAustralis (Smithfield NSW, Australia).

To induce autophagy, cells were washed two times with Earle’s balanced salt solution (EBSS, Sigma-Aldrich) and then incubated in the same medium for 2 h.

Virus stocks of EMCV, EMCV-Zn, Rluc-EMCV, CVB3, RLuc-CVB3, EV71 (kind gifts from Frank van Kuppeveld, University of Utrecht, Netherlands), A/Puerto Rico/8/1934 (H1N1) IAV (kind gift from Anke Huckriede, University Medical Center Groningen, The Netherlands), DENV-2 strain 16681, ZIKV (clinical isolate from Surinam) and CHIKV (La Reunion OPY1) (all a kind gift from Jolanda Smit, University Medical Center Groningen, The Netherlands) were generated and propagated as described previously (Mauthe et al., 2016; Bhide et al., 2019; Diosa-Toro et al., 2019; Troost et al., 2020).

Virus infections for EMCV, EMCV-Zn wt, Rluc-EMCV, CVB3, RLuc-CVB3 were performed at a multiplicity of infection (moi) of 0.25 (or 1 for the co-immunoprecipitation experiments) and virus inoculums were left onto the cells for 6 h. For DENV, CHIKV, ZIKV and IAV, cells were infected at a moi of 0.8 for 26 h, at a moi of 10 for 10 h, at a moi of 0.8 for 26 h and at a moi of 0.5 for 6 h, respectively.

GFP-HSBP1 was generated by cloning HSBP1 cDNA (Source BioScience, Nottingham, UK, cat# IRQMp5018C039D) into the pEGFP-C1 vector as EcoRI/SacII fragment. Construct correctness was confirmed by DNA sequencing and protein expression by western blot (WB).

U2OS (a kind gift from Ger Strous), HeLa (a kind gift from Peter van der Sluijs), U2OS cells stably expressing either GFP (GFP U2OS) or GFP-HSBP1 (GFP-HSBP1 U2OS), hsbp1^-/- U2OS (HSBP1KO), atg7^-/- U2OS (ATG7KO) (Janssen et al., 2018), atg13^-/- U2OS (ATG13KO) cells and HeLa RFP-GFP-LC3 cells (a kind gift of Tamotsu Yoshimori) were cultured in Dulbecco’s Modified Eagle Medium (DMEM, Life Technology, Carlsbad, CA) supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin and 10% fetal calf serum (FCS), at 37°C in 5% CO₂ humidified atmosphere. The culture medium of HeLa RFP-GFP-LC3, GFP U2OS and GFP-HSBP1 U2OS was supplemented with 0.6 µg/ml G418 (Thermo Fisher Scientific). The HSBP1KO and ATG13KO cells were generated as previously described (Janssen et al., 2018). In brief, for the generation of HSBP1KO and ATG13KO cells using the CRISPR/Cas9 system, guides targeting exon 1 (CGTCCCTTACCACCGAGGTGAGG) and exon 2 (GGATATTTCTCCCAATGATCTGG) of HSBP1 and exon 3 (TTTGCTTCATGTGTAACCTCTGG and AGTCGGGAGGTCCATGTGTGTGG) of ATG13, respectively, were designed using optimized CRISPR design (http://crispr.mit.edu/). Guides were cloned into pX458 plasmid (Addgene #48138) allowing expression of guide RNAs and Cas9 along with GFP. U2OS cells were transfected for 48 h and subsequently clonally sorted based on GFP expression using a SH800S Cell Sorter (Sony biotechnology, San Jose, CA). Clones were then sequenced and protein expression was assessed by WB to verify the deletion of HSBP1 and ATG13. Characterization of the ATG13KO cells can be found in Figure S2C .

To generate stable GFP or GFP-HSBP1 U2OS cells, U2OS cells were transfected with the pEGFP or the pEGFP-HSBP1 plasmid, respectively and then selected in a medium containing 0.6 µg/ml G418 for 10 days, resulting in a stable GFP U2OS or GFP-HSBP1 U2OS bulk population.

U2OS cells either stably or transiently expressing EGFP and EGFP-HSBP1 and grown in a 10 cm dish, were treated with EBSS for 2 h, EMCV infected for 6 h or left untreated before being subjected to lysis on ice in the following buffer: 20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM MgCl₂, 5 mM DTT, 0.5% Tween, Complete protease inhibitor (Roche, Basel, Switzerland), 1 mM PMSF. Co-immunoprecipitations were performed using the GFP‐trap beads (Chromotek, Planegg, Germany). Beads were incubated with the lysates for 2 h at 4°C and washed with the washing buffer (20 mM Tris-HCl, pH 7.4, 250 mM NaCl, 2 mM MgCl₂, 5 mM DTT, 0.5% Tween-20). Proteins were eluted by boiling the beats in Laemmli loading buffer (65.8 mM Tris-HCl, pH 6.8, 26.3% glycerol, 2.1% SDS, 0.01% bromophenol blue) (Laemmli, 1970) and then examined by WB.

Cells grown in 6-well or 24-well plates were washed with PBS and harvested in 100 µl of lysis buffer (20 mM Tris-HCl, pH 7.6, 130 mM NaCl, 1% Triton-X100, Complete protease inhibitor). The lysates were incubated on ice for 30 min, vortexed and centrifuged at 14,000 g for 10 min at 4°C. Supernatants were finally collected and mixed with the Laemmli loading buffer. Alternatively, cells were directly lysed in the Laemmli loading buffer and sonicated for 1 min. Equal protein amounts were separated by SDS-PAGE and after WB, proteins were detected on PDVF membranes (Merck) using specific antibodies and the Odyssey Imaging System (LI-COR Biosciences, Lincoln, NE). Densitometric values of the bands were quantified on WB images at non-saturating exposures using the ImageJ software (Schneider et al., 2012), and normalized against the loading control. For the detection of the small HSBP1 protein, we followed an optimized protocol previously described (Visweshwaran et al., 2018). In brief, after running, the gels were put into a renaturation buffer (20% glycerol, 50 mM Tris–HCl, pH 7.4) for 30 min at room temperature and after transfer, PVDF membranes were incubated in PBS containing 0.4% paraformaldehyde for 30 min to cross-link the proteins before proceeding with the detection.

Cells were fixed with 3.7% paraformaldehyde or 100% methanol, washed and blocked with blocking buffer (PBS, 1% bovine serum albumin, 0.1% saponin). Primary and secondary antibodies were diluted in the blocking buffer and incubated for 1 h at room temperature. Nuclei were stained with Hoechst33342 during the incubation with the secondary antibody for automated image acquisition. Alternatively, nuclei were stained with DAPI (Thermo Fisher Scientific) for confocal microscopy. Fluorescent microscopy images were collected with a DeltaVision RT fluorescence microscope (Applied Precision, Issaquah, WA) equipped with a CoolSNAP HQ camera (Photometrix, Kew, Australia). Images were generated by collecting a stack of 6 to 16 images with focal planes 0.30 μm apart, and subsequently deconvolved using the SoftWoRx software (Applied Precision). Quantification of puncta number was performed using the Icy software (http://icy.bioimageanalysis.org) using spot detector plugin or the ImageJ software. For automatic acquisition, fluorescence images were automatically acquired using a TissueFAXS (TissueGnostics, Vienna, Austria), which is based on a high-end fully motorized Zeiss AxioObserver Z1 microscope with a Zeiss- LD “Plan-Neofluar” 20x/0,4 Corr Dry objective (Zeiss, Oberkochen, Germany). The following filters were used: DAPI for the imaging of the nuclei, GFP for the acquisition of the GFP-HSBP1 and LC3 signals, and TexasRed for the imaging of the p62, EMCV and CVB3 capsid signal. The GFP and TexasRed filter were used for the GFP-RFP-LC3 tandem analysis. The acquired images were analyzed using the TissueQuest fluorescence analysis software (TissueGnostics GmbH, Vienna, Austria) to determine the cell count (based on the nuclei staining), the percentage of infected cells (based on the signal intensity in the cells), the mean signal intensity of infected cells and the translocation of GFP-HSBP1 into the nucleus (based on the signal intensities of GFP in the cytosol versus the nucleus). LC3 and p62 puncta were automatically quantified using the icy bioimage analysis software (De Chaumont et al., 2012).

U2OS cells were transfected for 48 h with 20 nM of either control siRNA or siRNA targeting HSBP1 (SMARTpool from Dharmacon, Lafayette, CO) using 0.1 µl, 0.5 µl or 2 µl of Lipofectamine RNAiMAX (Thermo Fisher Scientific) for 96-, 24- or 6-wells plate cultures, respectively, according to the manufacturer’s protocol. For the GFP and GFP-HSBP1 transfection, U2OS cells were seeded in 10 cm dishes (for co-immunoprecipitation, Co-IP) or in 6-well plates (for making stable cell lines), followed by a transfection procedure with Fugene (Promega, Madison, WI), according to the manufacturer’s protocol.

Cells grown in 96-well plates were washed with PBS and incubated with 50 µl of Lysis buffer (Thermo Fisher Scientific) at room temperature for 15 min, before storing the cell lysates at -20°C. 25 µl aliquots of thawed cell lysates were then used to measure renilla luciferase expression using the Renilla luciferase flash assay kit (Thermo Fisher Scientific). Alternatively, renilla luciferase activity was measured in the following reaction buffer: 45 mM EDTA, 30 mM sodium pyrophosphate, 1.425 M NaCl, 10 µM coelenterazine h (Promega) (Baker and Boyce, 2014).

Enzymatic activities were measured using a GloMax^®-Multi Detection System (Promega) and the following program: 25 µl substrate; 2 s delay; 10 s measuring. Background luminescence was subtracted from each value and the results were normalized towards cells transfected with control siRNA.

The Power SYBR^® Green Cells-to-CT™” kit (Thermo Fisher Scientific) was used according to manufacturer’s protocol to isolate RNA, reverse transcribe it and synthesize cDNA. Quantitative PCR was performed in a CFX connect Thermocycler (Bio-Rad, Hercules, CA) using the following specific primers (Mauthe et al., 2016): HSBP1 (TATCGCGGACCTCATGACAC and TAGCAACCTTCAACTCTTTTGCG), ULK1 (TGGGCAAGTTCGAGTTCTCC and CTCCAAATCGTGCTTCTCGC), ULK2 (TGGAGACCTCGCAGATTATTTGC and ACACTCTGATCGTGTCTTCACT), ATG101 (TCCTCCAGCTTCCGAGTCCA and CCACGTAACCAGGGAGGAAC), FIP200 (CTCAAACCAGGTGAGGGTGCTTCA and TGTTTTGTGCCTTTTTGGCTTGACA), ATG13 (TCCAGACAGTTCGTGTTGGG and CTCAAATTGCCTGGTAGACATGA), GAPDH (GGGAACGCATTGACTGTTTT and CTCGGGCTTCTCAAAGTCAC).

The mRNA expression levels were first normalized against the expression of GAPDH, before comparing gene expression levels in HSPB1-depleted cells with those in the control cells.

Statistical significance was evaluated using two-tailed heteroscedastic t-testing before calculating the p-values. Individual data points from each independent experiment (the number of the independent experiments is indicated in each figure legend) were used to determine significances.

We have found previously that FIP200 and ATG13 restrict picornavirus replication independently of their roles as components of the ULK kinase complex (Mauthe et al., 2016). We furthermore identified two potential binding partners that were shared by ATG13 and FIP200, i.e., HSBP1 and cell cycle progression 1 (CCPG1) (Mauthe et al., 2016). Since CCPG1 has recently been characterized as an ER-phagy receptor (Smith et al., 2018), we focused our attention to HSBP1 because no connection to autophagy or virus replication has been previously reported. To verify that HSBP1 is indeed interacting with FIP200 and ATG13, we performed Co-IP experiments in U2OS cells ectopically expressing GFP-HSBP1 (GFP-HSBP1 U2OS cells) using the GFP-Trap^® resin ( Figure 1A and Figure S1A ). We could confirm that GFP-HSBP1 specifically interacts with all the subunits of the tested ULK kinase complex, i.e. FIP200, ATG13 and ULK1 ( Figure S1A ). Next, we explored how HSBP1 binds to the ULK kinase complex also with Co-IP experiments. When ATG13 was knocked down in GFP-HSBP1 U2OS cells, the interaction between HSBP1 and ULK1, but not with FIP200 was abolished ( Figure 1A ). ULK1 knockdown did not influence the interaction of both FIP200 and ATG13 with HSBP1. In contrast, FIP200 depletion eliminated the binding between HSBP1 and ULK1 or ATG13 ( Figure 1A ). Altogether, these Co-IP experiments revealed that HSBP1 binds to the ULK kinase complex via FIP200.

The ULK kinase complex is essential for autophagy initiation (Licheva et al., 2021). Therefore we tested whether HSBP1 depletion alters the autophagic response to nutrient deprivation ( Figures 1 , 2 and Figures S1 , S2 ). First, we performed a classical autophagic flux assay (Klionsky et al., 2021) in which we treated control and HSBP1-depleted cells ( Figures S1B , S4 ) with EBSS for 2 h, to induce autophagy in the presence or absence of BafA1, a lysosomal inhibitor ( Figures 1B, C ). As a readout, we measured the conversion of non-lipidated microtubule associated protein 1 light chain 3 (LC3/MAP1LC3)-I into lipidated, autophagosomal membrane-associated LC3-II ( Figure 1B ) and the levels of sequestosome 1 (p62/SQSTM1) ( Figure 1C ) by WB, and examined the RFP-GFP-LC3 fluorescent reporter by fluorescence microscopy ( Figure S1C ), which are all assays that allow to assess autophagy induction and progression (Klionsky et al., 2021). The WB analyses revealed that BafA1 treatment increased the LC3-II and p62 levels in control and HSBP1-depleted cells to a similar extend ( Figures 1B, C ). Moreover, the number of formed autolysosmes was also equal in these cells ( Figure S1C ). Together, these data clearly show that autophagic progression is not blocked in the absence of HSBP1. Nonetheless, we detected a reduced autophagosome formation rate under starvation conditions, i.e., lower ratio of LC3-II/LC3-I, in HSBP1-depleted cells in comparison to the control ( Figure 1B ). To confirm the reduced autophagosome formation rate, we treated control and HSBP1-depleted cells with EBSS for 2 h and quantified endogenous ATG13 puncta by immunofluorescence microscopy (IF), which represent autophagosome formation sites (Karanasios et al., 2013). Knockdown of HSBP1 caused a significant reduction in the number of ATG13 puncta per cell ( Figure 1D ). We also generated a HSBP1 knockout cell line in U2OS cells, HSBP1KO ( Figures S2A , S4 ), and repeated the autophagic flux assay to corroborate the effects that HSBP1 depletion had on autophagy ( Figures 2A, B and Figure S2B ). Analogously to the result observed in HSBP1-depleted cells, we detected an impairment in autophagosome formation reflected by a lower LC3-II/LC3-I ratio ( Figure 2A ). We furthermore observed a significantly reduced number of p62 puncta per cell in HSBP1KO under control and starvation conditions ( Figure 2B ), whereas LC3 puncta were reduced under some conditions, but not significantly ( Figure S2B ). p62 forms discrete puncta when autophagy is induced and therefore measuring the amount of those is widely used as a read-out for autophagy induction (Orhon and Reggiori, 2017). Altogether, these results show that depletion of HSBP1, a novel interactor of the ULK kinase complex, impairs nutrient starvation-induced early autophagy events.

Two scenario could explain the negative impact that HSBP1 knockdown has on the nutrient-induced autophagic response. The first is that HSBP1 is a positive regulator for the ULK kinase complex. The second is that HSBP1 could be important to stabilize this complex, a possibility evoked by the fact that HSBP1 has been shown to be crucial in the assembly and stabilization of the WASH complex (Visweshwaran et al., 2018). This second scenario was tested by knocking down HSBP1 and examine the levels of the ULK kinase complex subunits by WB. Indeed, HSBP1 depletion led to a significant decrease of ATG13, FIP200 and ULK1 levels ( Figure 2C ). This effect was specific to these proteins since the expression levels of other ATG proteins, i.e. ATG16L1, was not influenced by HSBP1 knockdown. Since the decrease in protein expression was not due to a reduced mRNA expression ( Figure S2D ), this result shows that HSBP1 is important to stabilize the ULK kinase complex.

Since HSBP1 binds to FIP200, we next examined whether HSBP1 also functions together with the ATG13-FIP200 subcomplex in controlling picornavirus infection. We repeated the Co-IP experiments using GFP-HSPB1 as a bait and measured the binding to ATG13 after either autophagy induction or EMCV infection ( Figure 3A ). EMCV is a member of the picornavirus family that we have previously employed to characterize the anti-picornaviral role of the FIP200-ATG13 subcomplex (Mauthe et al., 2016). The binding of HSBP1 to ATG13 remained unchanged upon autophagy induction ( Figure S3A ) showing that nutrient starvation and subsequent activation of the ULK kinase complex activity is not regulated through dynamic HSBP1 binding. In contrast, when the cells were exposed to EMCV, HSBP1 association with ATG13 was significantly reduced ( Figure 3A ). This result and the fact that HSBP1 binds the ULK complex via FIP200 ( Figure 1A ), suggests that HSBP1 is also involved and possibly regulates the ATG13-FIP200 subcomplex function in controlling the replication of EMCV and by extension, of other picornaviruses. To test whether HSBP1 influences EMCV replication, we infected control and HSBP1-depleted U2OS cells with a wild-type EMCV strain and quantified both viral capsid expression and the percentage of virus infected cells. In parallel, the same cells were also infected with a luciferase-expressing EMCV strain to assess virus replication by measuring luciferase activity (Mauthe et al., 2016). We found that HSBP1 knockdown reduced EMCV replication to 40-60% in comparison to control cells ( Figure 3B ). We also infected the HSBP1KO cells and observed that EMCV replication was significantly reduced, confirming the results obtained with the siRNA ( Figure 3C ). Conversely, GFP-HSBP1 overexpression led to an increase in EMCV replication, revealing that HSBP1 is a novel host factor that positively regulates EMCV propagation ( Figure 3C ). Finally, we also tested CVB3 replication in HSBP1-depleted cells to determine whether what we observed is EMCV specific or not. Measurement of luciferase expression and quantification of virus-positive cells in HeLa and U2OS cells upon HSBP1 knockdown ( Figure 3D ) and in HSBP1KO cells ( Figure S3B ) revealed HSBP1 depletion also reduced CVB3 replication. These data established HSBP1 as a novel host factor that is required for optimal EMCV and CVB3 replication.

Picornaviruses cause a so-called nucleocytoplasmic traffic disorder by modulating the nuclear pore complexes and thereby disrupting the regulated transport of material, i.e., proteins and mRNA, in and out of the nucleus (Lidsky et al., 2006; Lizcano-Perret and Michiels, 2021). Interestingly, we also observed a relocalization of GFP-HSBP1 from the cytoplasm into the nucleus upon EMCV infection by fluorescence microscopy ( Figure 4A ). This redistribution was already seen at 3 h post-infection, in agreement with the notion that the nucleocytoplasmic traffic disorder is an early event during EMCV infection (Lidsky et al., 2006), while the capsid protein expression was only detected at 4-6 h post-infection when more than 90% of the infected cells showed nuclear GFP-HSBP1 ( Figures S3C, D ). GFP-HSBP1 relocalization was strongly reduced when we inhibited virus replication by either treating cells with cycloheximide 1 h after exposure to EMCV or inoculating cells with inactivated EMCV, showing that HSBP1 redistribution is induced by an active EMCV infection (data not shown). Previous studies have demonstrated that EMCV-induced nucleocytoplasmic traffic disorder is caused by the viral Leader (L) protein, which alters the phosphorylation status and thereby the function of the nuclear pore complexes (Lidsky et al., 2006). We also found that GFP-HSBP1 translocation into the nucleus depends on L since infection with EMCV-Zn, an EMCV strain lacking L (Hato et al., 2007), did not lead to the same change ( Figures S3C, D ). The mechanism underlying nucleocytoplasmic traffic disorders differs between members of the cardioviruses (Picornaviridae) like EMCV and members of the enteroviruses (Picornaviridae) such as CVB3 and EV71 (Lizcano-Perret and Michiels, 2021). Thus, we tested whether the relocalization of HSBP1 is specific to EMCV or also occurs upon infection with CVB3 and EV71. To this aim, we infected GFP-HSBP1 U2OS cells with EV71 and CVB3 before imaging them. Cell infection with these viruses also caused a relocalization of HSBP1 from the cytoplasm into the nucleus, i.e. more than 90% of infected cells displayed nuclear GFP-HSBP1 ( Figures S2 , 3C, D ), showing that HSBP1 redistribution is triggered by all the tested picornaviruses. Although it remains unclear whether the presence of HSBP1 in the nucleus is beneficial for the virus or simply a consequence of the nucleocytoplasmic traffic disorder, it is clearly specific for picornaviruses since viruses from other families (e.g., IAV, ZIKV, DENV or CHIKV) or treatments triggering autophagy via endoplasmic reticulum stress, did not cause HSBP1 translocation into the nucleus ( Figures S3E, F , and data not shown).

Replication of EMCV and other picornaviruses is enhanced when ATG13 and FIP200 are depleted (Mauthe et al., 2016). Since HSBP1 depletion has an opposite effect, we examined the functional relationship between these three proteins. To this aim, we infected wild type, ATG7KO (Janssen et al., 2018) and ATG13KO ( Figure S2C ) cells after knocking down or not HSBP1, and then measured EMCV infection by IF ( Figure 4B ). In agreement with previous results (Mauthe et al., 2016), EMCV replication was more efficient in the ATG13KO than in wild type and ATG7KO cells. Interestingly, HSBP1 depletion reduced EMCV infection also in ATG13KO cells, when compared to control siRNA treated cells ( Figure 4B ). To confirm that this observation is not specific for EMCV infection, we repeated the same experiment using CVB3, obtaining an identical result ( Figure 4B ). Collectively, these data show that the effect of HSBP1 on picornavirus replication is not mediated through its function in autophagy since HSBP1 depletion decreased viral replication in ATG7KO cells. Moreover, these results indicate that the anti-picornaviral function of the FIP200-ATG13 subcomplex is connected to HSBP1, since after the depletion of HSBP1 in ATG13KO cells ( Figure 4B ), the effect caused by the loss of ATG13 and FIP200 is lost.