The Autophagy-Initiating Protein Kinase ULK1 Phosphorylates Human Cytomegalovirus Tegument Protein pp28 and Regulates Efficient Virus Release

So far, investigations on the role of autophagy during HCMV infection have mainly focused on a few autophagy-related proteins or viral factors (11–17, 34, 35). In order to gain a more complete picture concerning the impact of individual autophagy-related proteins on HCMV replication, we decided to perform an RNA interference (RNAi)-based screening (Fig. 1). For this, primary human foreskin fibroblasts (HFFs) were transfected with either a nontargeting control or one of the 48 individual siRNA mixtures targeting various autophagy-related genes. To avoid interference between transfection and infection, cells were infected at 48 h posttransfection (hpt). Infection was performed with HCMV strain AD169 at a multiplicity of infection (MOI) of 0.1. At 96 h postinfection (hpi), viral genome equivalents in the supernatant were quantitated by TaqMan-based quantitative PCR (qPCR), and the cells from one replicate per sample were lysed and subjected to Western blot analyses (Fig. 1A and C). Thirty-five of the tested siRNAs, including siRNAs targeting the ULK1 complex constituents RB1CC1, ULK1, ATG13, and ULK2, effectively impaired the release of viral genomes into the supernatant, whereas 13 did not have a statistically significant effect compared to the nontargeting control (Fig. 1A). The reduction of viral genomes ranged between approximately 26% for CTSS/CTSD and 66% for CXCR4/UVRAG. Surprisingly, none of the siRNAs tested in this setting enhanced the release of viral genomes into the supernatant. To exclude that the transfection of the nontargeting control had any effect on viral replication and thereby affected the interpretation of the results, we also quantitated viral genomes in the supernatant of cells transfected with increasing amounts of the nontargeting control (6.25 to 100 nM) and compared them to those in mock-transfected samples but observed no difference (Fig. 1B). Regarding the expression levels of different viral marker proteins, we found that only a few siRNAs negatively affected viral gene expression (Fig. 1C). This was clearly the case for siRNAs targeting MTOR, UVRAG, or WDR45 (WIPI4). None of the other siRNAs showed a noticeable effect on viral protein expression, implying that the cause of the reduced virus release shown in Fig. 1A is an impairment of late steps of viral replication, such as tegumentation, secondary envelopment, or viral particle release.

We were particularly interested in the role of ULK1 in HCMV infection as it acts as a key initiator of autophagy. Furthermore, ULK1 is the only protein kinase of the core autophagy machinery and may thus serve as a promising drug target. Knockdown of ULK1 resulted in a significant reduction of virus release of approximately 49% (Fig. 1A). Interestingly, the transfection of an siRNA targeting ATG13 also resulted in the depletion of ULK1 and a significant reduction of virus release (Fig. 1A and C, compare lanes 6 and 42). This is consistent with previous reports indicating that ATG13 as part of the ULK1 protein complex affects the stability of ULK1 (22, 25). Taken together, these data indicate that the depletion of many autophagy-related proteins, including ULK1, negatively affects HCMV replication, suggesting that autophagy-related factors may be required for efficient viral replication.

To gain further insight into the regulation of proteins of the ULK1 complex during HCMV infection, HFFs were infected with HCMV strain AD169 or TB40/E (MOI of 1), harvested throughout the course of infection, and subjected to Western blot analyses (Fig. 2). The viral proteins IE1 (immediate early 1), pUL44, and pp28 (pUL99) were stained as marker proteins for the different phases of the HCMV replication cycle (immediate early, early, and late). The three major components of the ULK1 complex, namely, ULK1, RB1CC1 (FIP200), and ATG13, were clearly upregulated in both AD169- and TB40/E-infected cells (Fig. 2A and B). The same applied to the catalytic subunits of AMPK as well as mTOR, both regulators of ULK1 activity (21, 22, 25–28). Epidermal growth factor receptor (EGFR), as an additional control, displayed the opposite trend, as it was downregulated.

Since phosphorylation is a major mechanism affecting the activity of ULK1, we next investigated the phosphorylation status of ULK1 in AD169-infected cells using a panel of phospho-specific antibodies (29). Interestingly, the upregulation of ULK1 was not limited to total protein levels but could also be observed concerning the phosphorylation of ULK1 at four distinct serines (S317, S556, S638, and S758) (Fig. 2C). This was most evident at late times of infection, correlating with a distinct mobility shift of ULK1 upon detection with anti-ULK1 antibody R600 (Fig. 2C, top panel, lanes 7 to 10). Surprisingly, since we observed enhanced phosphorylation signals at both S317 and S758, AMPK-dependent activating as well as mTOR-dependent inactivating phosphorylation sites of ULK1 were affected equally (29). These findings in conjunction with previous results suggesting that autophagy is counteracted by HCMV led us to the assumption that ULK1 may play a role during HCMV infection beyond its function in the initiation of autophagy (11–13).

Intrigued by the finding that both activating as well as inactivating phosphorylation of ULK1 were affected during HCMV infection, we became interested in further elucidating the role of the upstream regulators AMPK and mTOR. One approach was the treatment of AD169-infected HFFs (MOI of 0.5) with different inhibitors of AMPK and mTOR followed by Western blot analyses (Fig. 3). Compound C (10 μM) was used to block AMPK activity, while torin 1 (100 nM) and rapamycin (50 nM) were used as mTOR inhibitors. AG490 (10 μM), an EGFR inhibitor, served as a negative control. Inhibitors were added at either 72 or 96 hpi to circumvent well-known negative effects on HCMV replication upon inhibition of AMPK or mTOR at earlier time points of infection (Fig. 3A) (36–39). At 120 hpi, cells were lysed and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting (Fig. 3B). The addition of the AMPK inhibitor compound C reduced the phosphorylation of ULK1 at the AMPK-specific phosphorylation site S317 by 20 and 70% after treatment for 24 and 48 h (addition at 96 and 72 hpi), respectively (Fig. 3B, lanes 3 and 9). Viral gene expression was hardly affected, although the levels of pp28 appeared to be slightly increased. Total levels of ULK1 protein were reduced after treatment with compound C by 40 and 20%, indicative of either reduced ULK1 upregulation or decreased stability of the protein. Interestingly, the additions of the two mTOR inhibitors torin 1 and rapamycin differed concerning their effects on the phosphorylation of ULK1 at S758 (Fig. 3B, lanes 4, 5, 10, and 11). While rapamycin had no effect, torin 1 efficiently reduced the phosphorylation of ULK1 at S758 by 50 and 60% after addition at 96 and 72 hpi, respectively. Viral gene expression of IE proteins or pUL44 was not affected, but in contrast to compound C, the addition of both inhibitors slightly decreased pp28 levels. Regarding total protein levels of ULK1, torin 1 reduced ULK1 levels by 50 and 40%, whereas rapamycin had no prominent effect.

Based on the finding that solely compound C and torin 1 affected the phosphorylation of ULK1 in this setting, both inhibitors were analyzed concerning their effect on HCMV replication. To investigate this, we infected HFFs with AD169 (MOI of 1) and added compound C (10 μM), torin 1 (100 nM), or AG490 (10 μM) at 48 hpi (Fig. 4A). Cell culture supernatants including inhibitors were renewed at 72 and 96 hpi to ensure the integrity of the inhibitors throughout the experiment. At 96 and 120 hpi, supernatants were harvested, and virus release was quantitated via either the determination of viral genomes by qPCR or titration of infectious viral particles (determined as IE1 protein-forming units [IEU] per milliliter) (Fig. 4B and C). The addition of compound C impaired the release of viral genomes into the supernatant more than 10-fold at both investigated time points, and the numbers of infectious particles were significantly reduced as well. In contrast, the effect of torin 1 was less pronounced and reduced the levels of viral genomes and infectious particles to a much lesser extent. Considering that the addition of compound C inhibits ULK1, while torin 1 leads to activation, we speculated that ULK1 possibly acts as a proviral factor during HCMV replication.

To further analyze the function of ULK1 during HCMV replication, we decided to manipulate the activity of ULK1 directly by using the ULK1 kinase inhibitor SBI-0206965 (40). First, we wanted to confirm the inhibitory activity of SBI-0206965 by both in vitro kinase and cell culture experiments (Fig. 5). For the in vitro kinase assay, HEK293T cells were cotransfected with plasmids for myc-tagged ATG13 (serving as a substrate) in combination with either hemagglutinin (HA)-tagged ULK1 or the kinase-dead mutant HA-ULK1(K46I) (Fig. 5A). After lysis at 2 days posttransfection (dpt), small fractions of the cell lysates were analyzed by Western blotting (input control), and the remainder was subjected to immunoprecipitation (IP) with HA- or myc-specific antibodies. Again, small fractions of the immunoprecipitates were analyzed by Western blotting (precipitation control). The remaining samples were utilized in an in vitro kinase reaction in the presence or absence of SBI-0206965. Subsequent detection of phosphorylated proteins was accomplished via SDS-PAGE, blotting onto a membrane, and autoradiography (in vitro kinase reaction). The coexpression of kinase-active HA-ULK1 and myc-ATG13 clearly resulted in the presence of a second, higher-migrating ATG13 band (myc-ATG13[-P]) in the input control, which was not detectable upon the coexpression of the kinase-dead mutant HA-ULK1(K46I) (Fig. 5A, lanes 1 and 2). The autoradiograph obtained after the in vitro kinase reaction showed distinct phosphorylation signals for HA-ULK1 (autophosphorylation) and myc-ATG13 (Fig. 5A, top panel, lane 2). In line with the results of Egan et al. demonstrating that SBI-0206965 is a potent inhibitor of ULK1 kinase activity, the presence of SBI-0206965 considerably decreased the autophosphorylation of HA-ULK1 as well as the phosphorylation of myc-ATG13 (Fig. 5A, top panel, lane 3) (40).

In order to analyze whether SBI-0206965 is cell permeable and inhibits the phosphorylation of ATG13 in analogous cell culture experiments, we cotransfected HEK293T cells with plasmids for HA-ULK1, the kinase-dead mutant HA-ULK1(K46I), and myc-ATG13. One hour prior to lysis, SBI-0206965 or the solvent control dimethyl sulfoxide (DMSO) was added directly to the cell culture supernatant (Fig. 5B). Staining with a phospho-specific antibody that detects ULK1-specific ATG13 phosphorylation at S318 revealed that the coexpression of catalytically active HA-ULK1 and myc-ATG13 resulted in a substantially increased phosphorylation of ATG13 that was not present in samples transfected with the kinase-dead mutant HA-ULK1(K46I) (Fig. 5B, third panel, lanes 4 and 7) (41). The addition of SBI-0206965 at two different concentrations led to a considerable decrease in the phosphorylation signal at S318 by up to 50% in proportion to the concentration of the inhibitor (Fig. 5B, lanes 5 and 6). This experiment demonstrates that the inhibitory effect of SBI-0206965 is evident not only in vitro but also in cell culture experiments.

We already observed that RNAi-mediated depletion of ULK1 impairs virus release into the cell culture supernatant (Fig. 1A). As a next step, we were interested in the role of ULK1 kinase activity in HCMV replication. Consequently, virus release assays were performed in the absence or presence of either SBI-0206965 or ULK-101, a novel and potent ULK1 inhibitor (Fig. 6A to C) (42). To investigate whether ULK1 kinase activity is important for HCMV replication at the immediate early, early, or late stage of infection, HFFs were infected with AD169 (MOI of 0.1), and the respective inhibitors were added at 1.5, 24, or 48 hpi at different concentrations as indicated (Fig. 6A). At 96 hpi, viral genomes were quantitated in cell culture supernatants by TaqMan-based qPCR. The addition of both inhibitors resulted in the reduced release of viral genomes into the supernatant irrespective of the time point of addition. SBI-0206965 impaired virus release starting at a concentration of 10 μM, whereas ULK-101 proved to be considerably more potent, with significant impairment starting at a concentration of 0.63 μM (Fig. 6A). At a concentration of 20 μM, SBI-0206965 decreased the number of viral genomes in the supernatant by 58 to 93%, while ULK-101 impaired the release of viral genomes by 65 to 89% at a concentration of 1.25 μM. The impact of ULK1 kinase inhibition on the release of viral genomes was less pronounced after the addition of the inhibitors at 48 hpi than after an earlier addition at 1.5 or 24 hpi. However, the fact that the addition of ULK1 inhibitors at 48 hpi was still able to induce a significant impairment of virus release implies that ULK1 kinase activity plays an important proviral role in late events of the viral replication cycle. As a next step, we aimed to analyze whether the reduced release of viral genomes that we observed in the absence of ULK1 kinase activity is limited to a specific HCMV strain, MOI, or time point of analysis. Thus, HFFs were infected with either AD169 or TB40/E (MOI of 0.1 or 1), followed by the addition of SBI-0206965 (20 μM) or ULK-101 (1.25 μM) at 48 hpi. Viral genomes in the cell culture supernatants were quantitated by TaqMan-based qPCR at 96 to 168 hpi (Fig. 6B and C). In AD169-infected cells, the release of viral genomes was reduced by 72 to 91% (MOI of 0.1) and 65 to 83% (MOI of 1) upon inhibition of ULK1 kinase activity (Fig. 6B). Similarly, in TB40/E-infected cells, release was reduced by 61 to 80% (MOI of 0.1) and 36 to 87% (MOI of 1) (Fig. 6C). In conclusion, irrespective of the viral strain, MOI, or time point of analysis, the release of viral genomes into the supernatant was significantly reduced in the absence of ULK1 kinase activity, with one exception. The addition of ULK1 kinase inhibitors did not show a significant impairment after infection with TB40/E at an MOI of 0.1 and analysis at 96 hpi; however, the total number of released viral genomes ranged between 695 and 2,321 on average and was therefore very low (Fig. 6C). To answer the question of whether inhibition of virus release by SBI-0206965 or ULK-101 is due to negative effects on viral gene expression, we investigated the expression pattern of different viral marker proteins throughout the replication cycle in the presence of ULK1 kinase inhibitors (Fig. 6D). Expression of the immediate early and early proteins IE1 and pUL44 was not affected by the addition of ULK1 kinase inhibitors up to 96 hpi. Surprisingly, we observed a modest increase in steady-state levels of the late proteins′ major capsid protein (MCP) and pp28 (pUL99) at 96 hpi in the presence of ULK1 kinase inhibitors. Most strikingly, however, ULK-101 elicited a distinct mobility shift of the phosphoprotein pp28 (Fig. 6D, lanes 14, 15, 19, and 20). In particular, at 96 hpi, a slightly slower-migrating form of pp28 merged with a faster-migrating form in the presence of ULK-101 (Fig. 6D, lane 20). This suggests that phosphorylation of pp28 depends on ULK1 kinase activity. Taken together, our data demonstrate that two chemically distinct ULK1 kinase inhibitors interfere with the release of viral particles without an obvious negative effect on viral gene expression, indicating that ULK1 kinase activity is required for late events of the viral replication cycle.

Due to the finding that inhibition of ULK1 kinase activity by ULK-101 changes the migration pattern of pp28 in SDS-PAGE gels, we asked the question of whether pp28 acts as a substrate for ULK1. To address this, an in vitro kinase assay was performed. After cotransfection of either HA-ULK1 or the kinase-dead mutant HA-ULK1(K46I) and green fluorescent protein (GFP)-tagged pp28 or GFP as a control, cells were lysed and subjected to immunoprecipitation with the indicated HA- or GFP-specific antibodies (Fig. 7). As evident in the input and precipitation controls, all proteins were expressed and immunoprecipitated to similar extents (Fig. 7A). In the in vitro kinase reaction, a clear phosphorylation signal was detectable for GFP-pp28 only after the coexpression of HA-ULK1, possibly representing direct phosphorylation (Fig. 7B, lane 3). Distinctively, in the absence of coexpressed HA-ULK1 or upon the coexpression of the kinase-dead mutant HA-ULK1(K46I), this phosphorylation signal was not observed (Fig. 7B, lanes 2 and 4). Furthermore, the coexpression of HA-ULK1 and GFP demonstrated that the GFP tag did not serve as a phosphorylation target (Fig. 7B, lane 5).

To further characterize the role of ULK1 in HCMV infection, we investigated whether ULK1 kinase activity affects the subcellular localization of pp28, pp65 (pUL83), MCP, and GM130 at late time points of infection (Fig. 8). For this, we infected HFFs with AD169 (MOI of 1) and added DMSO or ULK-101 (1.25 μM) at 48 hpi to exclude any potential immediate early or early effects of ULK1 inhibition. At 96 hpi, cells were fixed; stained for pp28, pp65, MCP, or GM130; and analyzed via fluorescence microscopy. In infected and DMSO-treated samples, pp28 localized primarily next to indentations of the crescent-shaped nuclei, consistent with the described accumulation of pp28 in cytoplasmic virion assembly complexes (cVACs) (Fig. 8A, panels 1 and 3) (43). The tegument protein pp65 localized to cytoplasmic speckles, MCP was detected in nuclear viral replication compartments and cVACs, and the cis-Golgi marker GM130 localized to cVACs (Fig. 8A, panels 5, 7, 9, 11, 13, and 15). In contrast, in the presence of the ULK1 kinase inhibitor ULK-101, pp28 and pp65 localized at large circular structures in the cytoplasm (Fig. 8C, panels 21, 23, 25, and 27, marked by white arrowheads). These structures were also readily visible in bright-field images but not in 4′,6-diamidino-2-phenylindole (DAPI) and MCP stainings, indicating that these structures do not represent accumulations of viral capsids (Fig. 8C, panels 21 to 32). The same applies to the cellular protein GM130, which did not colocalize with the cytoplasmic structures, indicating that cis-Golgi membranes are not involved in their formation (Fig. 8C, panels 33 and 35). To investigate whether the inhibition of ULK1 kinase activity itself results in the formation of these cytoplasmic structures, we analyzed mock-infected HFFs treated with either DMSO or ULK-101 (1.25 μM) (Fig. 8B and D). We found none of these structures in ULK-101-treated cells resembling the ones detected after infection, indicating that their formation is confined to infected cells (Fig. 8C and D, compare, e.g., panels 22 and 24 and panels 38 and 40). The substantial change in the localization pattern of pp28 and pp65 in the presence of ULK-101 implies that ULK1 kinase activity is important for the proper subcellular localization of both proteins at late time points of infection.

Next, we analyzed ULK-101-treated cells by transmission electron microscopy (TEM) (Fig. 9 and 10). This was done to further characterize the pp28/pp65-positive cytoplasmic accumulations and to investigate whether the altered localization of pp28 leads to an impairment of secondary envelopment as described previously for HCMV mutants containing a manipulated UL99 open reading frame (18). HFFs were infected with AD169 or TB40/E (MOI of 1) followed by the addition of either the solvent control DMSO or ULK-101 (1.25 μM) at 48 hpi. At 96 hpi, samples were prepared by high-pressure freezing, freeze substitution, Epon embedding, and ultrathin sectioning. In both DMSO- and ULK-101-treated cells, typical kidney-shaped nuclei and cVACs (delineated by dashed lines) in nuclear indentations could be observed (Fig. 9A and E and Fig. 10A and E). The cVAC structures and compositions were characteristic of late stages of lytic HCMV infection and contained numerous viral particles (capsids and dense bodies), as typically observed in HCMV infection (Fig. 9B, C, F, and G and Fig. 10B, C, E, and F, marked by arrowheads). Most importantly, in AD169 and TB40/E infection, TEM of ULK-101-treated cells revealed no obvious defect in secondary envelopment, defined as an increased ratio of naked or budding capsids compared to control cells, as the majority of viral particles were enveloped. This suggests that the phosphorylation of pp28 by ULK1 is not essential for secondary envelopment. However, it is noteworthy that in cells infected with either of the virus strains, we observed at least a 2-fold reduction (∼2.7-fold for AD169 and ∼3.6-fold for TB40/E) of viral particles in the cVACs of ULK-101-treated cells (Table 1). Furthermore, TEM confirmed the presence of prominent cytoplasmic structures in the presence of ULK-101 that appeared as circular, electron-dense structures with diameters of between 1 and 7 μm often found in close proximity to the cVACs (Fig. 9H and I and Fig. 10G and H). These structures were clearly distinguishable from accumulations observed in DMSO-treated cells due to their different size, localization, and texture (Fig. 9D). The relatively high electron density and the absence of an enclosing membrane together with the results of immunofluorescence staining (Fig. 8) led us to the assumption that these structures are most likely extraordinarily large protein accumulations containing at least the tegument proteins pp28 and pp65. Higher magnifications of the large cytoplasmic structures in AD169-infected cells treated with ULK-101 revealed that they were often directly associated with vesicles that sometimes contained viral particles or intraluminal vesicles (Fig. 9H and I, marked as V) and membranes (Fig. 9H and I, marked by asterisks). However, the large structures were never completely enclosed by these membranes. In TB40/E-infected cells treated with ULK-101, these accumulations were generally surrounded by numerous DNA-containing, tegumented but nonenveloped capsids (Fig. 10G and H, marked by gray arrowheads). Taken together, our data suggest that ULK1 kinase activity is important to prevent the deregulated accumulation of pp28 and pp65 in the cytoplasm.

The expression plasmid encoding HA-ULK1 was generated by amplification of the ULK1 coding sequence from pRK5_myc-ULK1 (a gift from Do-Hyung Kim; Addgene plasmid 31961) (22) utilizing 5′-GCATGAATTCATGTACCCATACGATGTTCCAGATTACGCTGAGCCCGGCCGCGGCGGCACAG-3′ (containing the HA sequence) and 5′-CGTATCTAGATTAGGCACAGATGCCAGTCAGC-3′ as primers and subsequent ligation into pcDNA3 (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) with EcoRI and XbaI. An expression plasmid encoding the kinase-dead mutant HA-ULK1(K46I) was generated by site-directed mutagenesis with 5′-GATTTGGAGGTCGCCGTCATCTGCATTAACAAGAAGA-3′ and 5′-TCTTCTTGTTAATGCAGATGACGGCGACCTCCAAATC-3′ as primers and the above-mentioned plasmid pcDNA3_HA-ULK1 as a template. pRK5_myc-ATG13 was a gift from Do-Hyung Kim (Addgene plasmid 31965) (22). The expression plasmid encoding GFP-pp28 was generated by amplification of the pp28 (pUL99) coding sequence from pcDNA3_pp28 utilizing 5′-GCATAAGCTTTGGGTGCCGAACTCTGCAAACG-3′ and 5′-GCATGGATCCTTAAAAGGGCAAGGAGGCGGC-3′ as primers and subsequent ligation into pEGFP-C1 (Clontech/TaKaRa Bio, Saint-Germain-en-Laye, France) with HindIII and BamHI. The template vector pcDNA3_pp28 was generated by amplification of the pp28 (UL99) coding sequence utilizing 5′-GCATGAATTCATGGGTGCCGAACTCTGC-3′ and 5′-CGTATCTAGATTAAAAGGGCAAGGAGGCGGC-3′ as primers and subsequent ligation into pcDNA3 (Invitrogen, Thermo Fisher Scientific) with EcoRI and XbaI.

The following monoclonal antibodies were used for Western blotting, immunoprecipitation (IP), and indirect immunofluorescence: anti-β-actin (AC-15) (catalog number A1978; Sigma-Aldrich, Taufkirchen, Germany), anti-EGFR (D38B1) (catalog number 4267; Cell Signaling Technology, Frankfurt am Main, Germany), anti-GFP (clones 7.1 and 13.1, a mixture of two monoclonal antibodies) (catalog number 11814460001; Roche Diagnostics, Mannheim, Germany), anti-GM130 (clone 35) (catalog number 610822; BD Biosciences, San Jose, CA, USA), anti-HA (HA-7) (catalog number H9658; Sigma-Aldrich), anti-IE1 (p63-27) (48), anti-MCP (28-4) (49), anti-myc (MYC 1-9E10.2; ATCC, LGC Standards, Wesel, Germany), anti-pp28 (41-18) (50), anti-pp65 (65-33) (kindly provided by W. J. Britt, Birmingham, AL, USA), anti-pUL44 (BS510) (kindly provided by B. Plachter, Mainz, Germany), and anti-ULK1 (D9D7) (catalog number 6439; Cell Signaling Technology). The following polyclonal antibodies were used for Western blotting: anti-AMPKα (catalog number 2532; Cell Signaling Technology), anti-ATG13 (catalog number SAB4200100; Sigma-Aldrich), anti-IE86 (51), anti-mTOR (catalog number 2972; Cell Signaling Technology), anti-RB1CC1 (catalog number 17250-1-AP; Proteintech Europe, Manchester, United Kingdom), and anti-ULK1 (R600) (catalog number 4773; Cell Signaling Technology). The following phospho-specific antibodies were used for Western blotting: polyclonal anti-ATG13-P (S318) (catalog number PAB19948; Abnova, Taipei City, Taiwan), monoclonal anti-ULK1-P (S317) (D2B6Y) (catalog number 12753; Cell Signaling Technology), monoclonal anti-ULK1-P (S556, equivalent to S555 of mouse Ulk1) (D1H4) (catalog number 5869; Cell Signaling Technology), polyclonal anti-ULK1-P (S638) (catalog number 12097; Cell Signaling Technology), and polyclonal anti-ULK1-P (S758, equivalent to S757 of mouse Ulk1) (catalog number 6888; Cell Signaling Technology). As secondary antibodies for Western blotting, horseradish peroxidase (HRP)-conjugated goat anti-mouse or -rabbit antibodies (Dianova, Hamburg, Germany) were used. As a secondary antibody for indirect immunofluorescence, a goat anti-mouse antibody conjugated with Alexa Fluor 488 (Invitrogen, Thermo Fisher Scientific) was used.

Primary human foreskin fibroblasts (HFFs) were prepared from human foreskin tissue and maintained in minimum essential medium (MEM; Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 7% fetal bovine serum (FBS; Sigma-Aldrich, Taufkirchen, Germany) and 350 mg/liter l-glutamine (Sigma-Aldrich) or GlutaMAX (Gibco, Thermo Fisher Scientific). Infection experiments were performed with the HCMV strains AD169 and TB40/E, which were titrated by quantitating IE1 protein-forming units (IEU) as described previously (52). For standard infection, 3 × 10⁵ or 1.2 × 10⁵ HFFs were seeded into 6- or 12-well dishes 1 day prior to inoculation with 1 or 0.5 ml of the viral supernatant. At 1.5 h postinfection (hpi), fresh growth medium was added. Subsequently, infected cells and the resulting viral supernatant were used for analyses.

Human embryonic kidney HEK293T cells were maintained in Dulbecco’s modified Eagle medium (DMEM; Gibco, Thermo Fisher Scientific) supplemented with 10% FBS and 350 mg/liter l-glutamine. Transfection of HEK293T cells was performed utilizing calcium phosphate-DNA precipitates. For standard transfection, 7 × 10⁵ HEK293T cells were seeded into 6-well dishes 1 day prior to transfection with 1 μg plasmid DNA. At 16 h posttransfection (hpt), cells were washed twice with phosphate-buffered saline (PBS) and provided with fresh growth medium. At 48 hpt, cells were used for subsequent analyses.

SBI-0206965 was purchased from either BioVision (Milpitas, CA, USA) (catalog number 9580) or Sigma-Aldrich (Taufkirchen, Germany) (catalog number SML1540), and ULK-101 was obtained from Selleck Chemicals (Munich, Germany) (catalog number S8793). Compound C (dorsomorphin) (catalog number 171260), torin 1 (mTOR inhibitor XI) (catalog number 475991), rapamycin (sirolimus) (catalog number 553210), and AG490 (tyrphostin B42) (catalog number 658401) were purchased from Sigma-Aldrich (Calbiochem, Taufkirchen, Germany). All compounds were dissolved in dimethyl sulfoxide (DMSO), aliquoted, and stored at −20°C or −80°C.

For Western blotting, whole-cell extracts were generated by resuspending cell pellets in PBS, mixing with the Laemmli-based buffer Roti Load 1 (Carl Roth, Karlsruhe, Germany), boiling at 95°C for 10 min, and subsequent sonication. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with 6 to 12.5% polyacrylamide gels and blotted onto nitrocellulose (GE Healthcare, Chicago, IL, USA) or polyvinylidene difluoride (PVDF) (Bio-Rad, Hercules, CA, USA) membranes. Chemiluminescence was detected using an ECL Western blotting detection kit (GE Healthcare) according to the manufacturer’s instructions and digitized with Fusion FX (Vilber Lourmat, Collégien, France) or a Fujifilm LAS-1000 luminescent image analyzer (Fujifilm Europe, Düsseldorf, Germany). Signal intensities were quantitated utilizing Image Lab 6.0, and images were edited in Adobe Photoshop Elements 2018.

For indirect immunofluorescence, HFFs were cultivated at a density of 2.5 × 10⁴ cells/well in μ-slide 8-well dishes (ibidi, Gräfelfing, Germany) with ibiTreat surface modification to ensure reliable adhesion of the cells. Cells were washed once with PBS before fixation with either 4% (wt/vol) paraformaldehyde (PFA) (dissolved in PBS) for 10 min at room temperature or methanol for 15 min at −20°C. Subsequently, cells were washed three times with PBS and, in the case of PFA fixation, permeabilized with 0.2% (vol/vol) Triton X-100 (dissolved in PBS) for 20 min at 4°C and additionally washed five times with PBS. Next, cells were incubated with PBS containing 2 mg/ml gamma globulins from human blood (Sigma-Aldrich) for 30 min at 37°C to block unspecific binding of the antibodies. Next, cells were first incubated with the respective primary antibodies diluted in 1% (vol/vol) FBS (diluted in PBS) for 1.5 h at 37°C and then incubated with secondary goat anti-mouse antibody conjugated with Alexa Fluor 488 (Invitrogen, Thermo Fisher Scientific) diluted 1:400 in 1% (vol/vol) FBS (diluted in PBS) for 45 min at 37°C after three washing steps with PBS. Following another three washing steps with PBS, cells were mounted with Vectashield antifade mounting medium (Vector Laboratories, Burlingame, CA, USA) containing 4′,6-diamidino-2-phenylindole (DAPI) and analyzed with a Zeiss Axio Observer.Z1 instrument with Apotome.2 using a 40× objective. Fluorescent dyes were excited at a wavelength of 475 or 385 nm, and emission was visualized utilizing appropriate filters excluding a signal overlap. Images were edited in ZEN 2.3 lite.

One day prior to transfection, 5 × 10⁶ HEK293T cells were seeded into 100-mm dishes. A total of 10 to 20 μg plasmid DNA per construct coding for either the kinase or the substrate protein was cotransfected utilizing calcium phosphate-DNA precipitates. At 2 days posttransfection (dpt), cells were lysed on ice for 20 min with IP buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 5 mM EDTA, 0.5% [vol/vol] Nonidet P-40 substitute) freshly supplemented with protease inhibitors (1 mM phenylmethylsulfonyl fluoride [PMSF], 2 μg/ml aprotinin, 2 μg/ml leupeptin, 2 μg/ml pepstatin A). Cellular debris was removed by centrifugation at 20,000 × g for 10 min. One-twentieth of the lysate was mixed with the Laemmli-based buffer Roti Load 1 (Carl Roth) and boiled at 95°C for 10 min and served as the input control. The remaining lysate was mixed with protein A-Sepharose beads (Sigma-Aldrich) coupled with the respective antibodies and incubated for 2 h at 4°C in a rotating mixer. Beads were washed twice with IP buffer including protease inhibitors and twice with kinase buffer (20 mM HEPES [pH 7.4], 0.4 mM EDTA, 1 mM EGTA, 5 mM MgCl₂, 50 μM dithiothreitol [DTT], 100 μM Na₃VO₄) including protease inhibitors. Two-sevenths of the protein-loaded beads were mixed with Roti Load 1 and boiled at 95°C for 10 min and served as the precipitation control. If appropriate, beads were preincubated with SBI-0206965 or DMSO for 30 min at 4°C in a rotating mixer. The kinase reaction was performed by mixing the remaining beads with kinase reaction buffer (kinase buffer supplemented with 10 μM ATP, 3 μCi [γ-³³P]ATP per sample, and SBI-0206965 or DMSO, if appropriate) followed by incubation at 30°C for 30 min in a microtube shaker. The reaction was stopped by mixing the samples with Roti Load 1 and boiling at 95°C for 10 min. SDS-PAGE and Western blotting were performed as described above. Proteins labeled with [γ-³³P]ATP during the kinase reaction were visualized by applying the Western blot membrane onto a photostimulable phosphor plate (Fujifilm) and subsequent scanning (in vitro kinase reaction). Images were edited in Aida software v4.22 and Adobe Photoshop Elements 2018.

The siRNAs used in this study were a mixture of four individual siRNAs targeting the same gene (SMARTPool) that were further chemically modified to reduce off-target effects (ON-TARGETplus siRNA). They were purchased from Horizon Discovery (Cambridge, UK) and are as follows: nontargeting control pool (D-001810-10), AMBRA1 (L-029987-01), ATG2A (L-026591-02), ATG2B (L-016822-02), ATG3 (L-015375-00), ATG4A (L-005789-00), ATG4B (L-005786-00), ATG4C (L-005788-00), ATG4D (L-005790-00), ATG5 (L-004374-00), ATG7 (L-020112-00), ATG9A (L-014294-01), ATG9B (L-018082-02), ATG10 (L-019426-01), ATG12 (L-010212-00), ATG13 (L-020765-01), ATG14 (L-020438-01), ATG16L1 (L-021033-01), ATG16L2 (L-026687-01), ATG101 (L-017816-01), BECN1 (L-010552-00), CTSB (L-004266-00), CTSD (L-003649-00), CTSS (L-005844-00), CXCR4 (L-005139-00), GABARAP (L-012368-00), GABARAPL1 (L-014715-00), GABARAPL2 (L-006853-00), MAP1LC3A (L-013579-00), MAP1LC3B (L-012846-00), MAP1LC3C (L-032399-01), MTOR (L-003008-00), PIK3C3 (L-005250-00), PIK3CG (L-005274-00), PIK3R4 (L-005025-00), PRKAA1 (L-005027-00), RB1CC1 (L-021117-00), SQSTM1 (L-010230-00), TMEM74 (L-015923-02), TNFSF10 (L-011524-00), ULK1 (L-005049-00), ULK2 (L-005396-00), UVRAG (L-015465-00), VMP1 (L-015899-01), WDR45 (L-019758-01), WDR45B (L-017119-01), WIPI1 (L-018205-01), WIPI2 (L-020521-01), and ZFYVE1 (L-013086-01). One day after seeding 5 × 10⁴ HFFs into 24-well dishes, the cells were transfected in triplicates with 12.5 pmol of the respective siRNA mixtures (corresponding to a final concentration of 25 nM) utilizing Lipofectamine 2000 transfection reagent (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA). As a control, the same amount of the ON-TARGETplus nontargeting control pool was used. At 48 hpt, HFFs were infected with the HCMV laboratory strain AD169 at a multiplicity of infection (MOI) of 0.1 and subjected to a virus release assay. At 96 hpi, the cells of one well of each triplicate were additionally lysed and used for Western blot analyses. Statistical analyses were performed with VassarStats (http://www.vassarstats.net/).

One day prior to infection, 1.2 × 10⁵ HFFs were seeded in triplicates into 12-well dishes. Before infection with either of the HCMV strains AD169 and TB40/E (MOI of 0.1 or 1), cells were washed once with growth medium. After inoculation for 1.5 h, cells were washed twice with growth medium, and 1 ml growth medium was added. At 96 to 168 hpi, the entire supernatant was harvested and centrifuged at 200 × g for 5 min. Infectious particles (determined as IEU per milliliter) in the supernatant were quantitated as described previously (52). For qPCR, 20 μl of the supernatant per sample was mixed with 80 μl of a proteinase K solution (1.65 U/ml proteinase K, 50 mM KCl, 15 mM Tris, 2.5 mM MgCl₂, 0.5% [vol/vol] Tween 20 [pH 8.0]) and incubated at 56°C for 1 h and subsequently at 95°C for 5 min. Next, 5 μl of the proteinase K-digested samples was mixed with 2× SsoAdvanced universal probes supermix (Bio-Rad, Hercules, CA, USA), primers (250 nM each) (5′-AAGCGGCCTCTGATAACCAAG-3′ and 5′-GAGCAGACTCTCAGAGGATCGG-3′), and a probe (150 nM) (5′-CATGCAGATCTCCTCAATGCGGCG-3′ [labeled with 6-carboxyfluorescein {FAM} and 6-carboxytetramethylrhodamine {TAMRA}]). qPCR was performed in an AriaMx real-time PCR system (Agilent, Santa Clara, CA, USA) and analyzed with AriaMx software v1.5. Thermal cycling conditions consisted of an initial denaturing step at 95°C (3 min) and 40 amplification cycles at 95°C (10 s) and 60°C (30 s). Alternatively, 5 μl of the proteinase K-digested samples was mixed with 2× TaqMan PCR master mix (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA), primers (375 nM), and a probe (200 nM), and qPCR was performed in an Applied Biosystems 7500 real-time PCR system (Applied Biosystems, Thermo Fisher Scientific). Thermal cycling conditions consisted of two initial steps at 50°C (2 min) and 95°C (10 min) and 40 amplification cycles at 95°C (15 s) and 60°C (60 s). In this case, analysis was performed using 7500 system SDS software v1.4. As a reference for quantitation of genome equivalents, serial 10-fold dilutions (10⁷ to 10²) of a plasmid containing the IE1 coding sequence were analyzed in parallel. Statistical analyses were performed with VassarStats (http://www.vassarstats.net/).

HFFs were seeded on sapphire discs (3 mm in diameter; Engineering Office M. Wohlwend GmbH, Sennwald, Switzerland) in μ-slide 8-well dishes (ibidi) at a density of 2.5 × 10⁴ cells/well. One day later, cells were infected with HCMV strain AD169 or TB40/E (MOI of 1), and DMSO or ULK-101 (1.25 μM) was added at 48 hpi. At 96 hpi, EM sample preparation was conducted as described previously (53, 54). In brief, infected HFFs were immobilized by high-pressure freezing (Compact 01; Engineering Office M. Wohlwend GmbH) and subjected to freeze substitution in a substitution medium consisting of acetone with 0.2% osmium tetroxide, 0.1% uranyl acetate, and 5% water (55). After washing and stepwise embedding in epoxy resin (Sigma-Aldrich), ultrathin sections of a 70-nm thickness were prepared, mounted on Formvar-coated single-slot grids (Plano, Wetzlar, Germany), and examined with a JEM-1400 transmission electron microscope (JEOL, Tokyo, Japan) at an accelerating voltage of 120 kV.