key: cord-0861330-x92ed83i
authors: Liu, Yuan; Lear, Travis; Larsen, Mads; Lin, Bo; Cao, Qing; Alfaras, Irene; Kennerdell, Jason; Salminen, Laura; Camarco, Daniel; Lockwood, Karina; Ma, Jing; Liu, Jie; Tan, Jay; Myerburg, Michael; Chen, Yanwen; Croix, Claudette St; Sekine, Yusuke; Evankovich, John; Finkel, Toren; Chen, Bill
title: Modulation of lysosomal function as a therapeutic approach for coronaviral infections
date: 2021-04-23
journal: Res Sq
DOI: 10.21203/rs.3.rs-419305/v1
sha: 0cf62c9f5e8560740392c800c5f93dab5ee4687d
doc_id: 861330
cord_uid: x92ed83i

The endo-lysosomal pathway plays an important role in pathogen clearance and both bacteria and viruses have evolved complex mechanisms to evade this host system. Here, we describe a novel aspect of coronaviral infection, whereby the master transcriptional regulator of lysosome biogenesis – TFEB – is targeted for proteasomal-mediated degradation upon viral infection. Through mass spectrometry analysis and an unbiased siRNA screen, we identify that TFEB protein stability is coordinately regulated by the E3 ubiquitin ligase subunit DCAF7 and the PAK2 kinase. In particular, viral infection triggers marked PAK2 activation, which in turn, phosphorylates and primes TFEB for ubiquitin-mediated protein degradation. Deletion of either DCAF7 or PAK2 blocks viral-mediated TFEB degradation and protects against viral-induced cytopathic effects. We further derive a series of small molecules that interfere with the DCAF7-TFEB interaction. These agents inhibit viral-triggered TFEB degradation and demonstrate broad anti-viral activities including attenuating in vivo SARS-CoV-2 infection. Together, these results delineate a viral-triggered pathway that disables the endogenous cellular system that maintains lysosomal function and suggest that small molecule inhibitors of the E3 ubiquitin ligase DCAF7 represent a novel class of endo-lysosomal, host-directed, anti-viral therapies.

Autophagy plays a key role in the cellular response to microbial threats including a direct role for pathogen elimination through the endo-lysosomal pathway 1 . Master transcriptional control of autophagy rests with the Transcription factor EB (TFEB), which regulates genes involved in all stages of autophagy 2, 3 . Various stressors stimulate the nuclear translocation of TFEB where it can activate >400 different gene products containing the CLEAR (coordinated lysosomal expression and regulation) regulatory motif 4 . These gene products work to increase lysosomal number and function, endocytic function, lysosomal exocytosis, and maintain lysosomal acidi cation 5 . TFEB's role in host defense is evolutionarily conserved, as it is known to play a prominent role in the response to bacterial pathogens in organisms such as C. elegans 6 . In many cases, pathogens have evolved sophisticated mechanisms to disrupt or evade the host's autophagic machinery [7] [8] [9] [10] . Here, we report that coronaviruses trigger TFEB protein degradation through E3 ligase-dependent ubiquitination. A small molecule that speci cally blocks virally-induced TFEB degradation is shown to have potent anti-viral activity.

Viruses induce proteasomal-mediated degradation of TFEB Recent observations have demonstrated that coronaviruses, including SARS-CoV-2, induce a critical deacidi cation of lysosomes, as part of their pathogenesis 10 . We were intrigued by this observation, since normally, lysosomal pH is kept in a narrow physiological range by the feedback, homeostatic activity of TFEB and related transcription factors 11 . To begin to understand how coronaviruses are able to modulate lysosomal function, we inoculated human airway cells with two human coronavirus strains, OC43 and 229E; strains associated with the common cold 12 . Surprisingly, we observed a marked dose and time dependent reduction of TFEB protein level after viral infection ( Fig. 1A -B, S1A-D). This fall in protein levels occurred without a concomitant reduction in TFEB transcription ( Fig. 1C-D) . Viral infection decreased the cytoplasmic and nuclear pools of TFEB protein, revealing a nuclear-compartment loss of TFEB protein and thus presumed impairment of its transcriptional activity ( Fig. S1E-F) . These results suggest that viral infection likely triggers the post-translational degradation of TFEB, which may impair the cell's ability to maintain lysosomal function. In support of this, we observed that following viral infection, there was an increase in K48-linked TFEB poly-ubiquitin, the canonical linkage associated with proteasome degradation (Fig. 1E) . To begin to understand whether this observation might be a point for future therapeutic intervention, we performed an unbiased, high throughput chemical screen using a library of FDA-approved molecules (approximately 1100 compounds). We used an EGFP-tagged TFEB chimeric protein, and speci cally focused on nuclear TFEB levels in the setting of concurrent viral infection, as the nuclear pool of TFEB is responsible for its biological activity. From this screen, we noted that the top hits, namely molecules that could maintain nuclear TFEB levels in the setting of coronaviral infection, all represented chemical inhibitors of the ubiquitin-proteasomal system (UPS) (Fig. 1F ).

Together, these results suggest that coronaviral infection leads to reduced TFEB protein levels, likely mediated through the ubiquitin proteasomal pathway.

The E3 ligase DCAF7 mediates TFEB degradation E3 ubiquitin ligases catalyze substrate ubiquitination, and we therefore sought to identify the E3 ubiquitin ligase regulating TFEB degradation. TFEB-EGFP pulldown (PD) and MS/MS analysis detected only one E3 ligase protein, DCAF7, present in the TFEB-EGFP PD samples (total 449 proteins detected) (Fig. S2A ).

DCAF7 is a substrate receptor for the Cul4-based CRL4 ubiquitin E3 ligase complex, and aside from a few studies, including a study validating its nuclear localization, the mechanistic function of DCAF7 remains unknown [13] [14] [15] . Ectopic DCAF7 expression decreased TFEB protein levels in a dose-dependent manner ( Fig. S2B) , and signi cantly increased TFEB poly-ubiquitination (Fig. S2C) . Furthermore, RNAi knockdown of DCAF7 led to an accumulation of TFEB protein in the nucleus ( Fig. 2A) . The bioinformatic tool, BDM-PUB, predicted a key ubiquitin acceptor within TFEB (Lys-232), which when mutated increased TFEB protein stability and reduced DCAF7-mediated ubiquitination ( Fig. S2D-F) 16 . CRISPR-based gene editing of DCAF7 revealed that in DCAF7 knockout (KO) cells, TFEB exhibited a signi cantly prolonged protein half-life (Fig 2B) , as well as increased nuclear localization and reduced poly-ubiquitination, (Fig. S2G , S3A).

Prior to viral infection, DCAF7 KO cells demonstrated an approximate 2-fold increase in the level of TFEB protein levels compared to WT control cells (Fig. 2C ). This supports a role for DCAF7 control of TFEB levels under basal conditions. Following coronavirus challenge, TFEB protein declined in WT cells but was preserved in DCAF7 KO cells (Fig. 2C, S3B) . Interestingly, DCAF7 KO cells were also resistant to the cytopathic effects of OC43 infection (Fig. 2D ) and demonstrated an increase in viability after infection (Fig. 2E, S3C) . Preservation of TFEB also correlated with a signi cant decrease in viral load (Fig. S3B, D) .

These observations appear relevant to other viral pathogen families, as H1N1 in uenza infection of airway epithelial cells also led to a dose-dependent reduction in TFEB protein which was blunted in DCAF7 KO cells (Fig. S2E ). To validate the delity of our KO cells, we re-expressed DCAF7. Reconstitution of DCAF7 in DCAF7 KO cells abrogated the protective KO phenotype, demonstrating comparable viral infection as WT cells (Fig. 2F, S4A -B).

The Cullin based E3 ligase complex typically targets phosphorylated substrates [17] [18] [19] , and TFEB translocation is known to be regulated by phosphorylation status 5 . Taken together, we hypothesized that viral infection might alter the phosphorylation of TFEB, marking the protein for ubiquitination. We used a siRNA library targeting all 600 kinases and high content imaging to identify regulators of TFEB nuclear localization. This unbiased approach uncovered p21-activated kinase 2 (PAK2) as a key determinant of We next explored the mechanism whereby PAK2 modulates TFEB stability. Through protein binding assays we noted that PAK2 and TFEB could directly interact, and we de ned a critical region within TFEB's basic helix-loop-helix domain required for PAK2 binding ( Fig. S5B-D) . We sought to uncover the key substrate phosphorylation sites, phospho-degrons, within TFEB that may serve as E3 ligase engagement motifs 20 . We hypothesized that proteasomal inhibition would speci cally enrich for TFEB phosphorylation sites that might act as phospho-degrons, identi able through phospho-proteomics ( Fig.   S6A ). We identi ed several TFEB phosphorylation sites signi cantly enriched with proteasomal inhibition (MG132), making them candidate phospho-degrons (Supplemental Table 1 ). Of these candidates, serine-138 and serine-142 t the pattern of a classical E3 ligase phospho-degron motif: S/T XXX S/T 21 (Fig.   3H ). Mutation of each of these candidate serine residues to alanine resulted in reduced DCAF7-TFEB interaction (Fig. 3I) . The double serine to alanine mutant (S138A/S142A), hereafter termed the TFEB phospho-mutant, further weakened this interaction when compared to mutation of each residue alone, and was resistant to DCAF7-mediated ubiquitination (Fig. S6B) . We next explored the functional importance of this mechanism. The TFEB phospho-mutant exhibited enhanced nuclear localization and resistance to OC43 mediated TFEB phosphorylation (Fig. S6C-D) . Further, the TFEB phospho-mutant was resistant to degradation following OC43 infection and its persistent expression decreased viral load (Fig.  3J , S6E-F). Together, this suggests PAK2-mediated TFEB phosphorylation generates a critical molecular motif for substrate engagement by the DCAF7-E3 ligase complex (Fig. 3K) Small molecule DCAF7 inhibitors modulate in vitro and in vivo TFEB activity From these data, we hypothesized that small molecule inhibition of PAK2 or DCAF7 could prevent TFEB protein degradation, leading to preserved endo-lysosomal activity and the potential to improve the cell's anti-viral response. To date, it has been di cult to generate PAK2 or PAK family inhibitors without toxicity 22, 23 . As such, we focused on novel small molecules targeting the E3 ubiquitin ligase DCAF7. DCAF7 harbors a conserved WD repeat domain within its C-terminus which is a key substrate binding domain 24, 25 . We hypothesized that small molecule inhibition of the WD repeat domain would disrupt DCAF7's interaction with TFEB. We constructed a DCAF7 homology model using the Nurf55 WD domain crystal structure (2XYI.pdb) 26 and screened 3 million compounds (ChemDiv, INC) as potential ligands in silico (Fig. S7A) . The top score-ranking molecules were selected and further evaluated in vitro. The initial top hit, termed BC1753, dose-dependently increased TFEB protein, and nuclear localization, and increased autophagosome formation and lysosomal activity (Fig. S7B-E) . We executed multiple rounds of hit-tolead and subsequent lead optimization and arrived at two closely related candidate DCAF7 inhibitors, BC18813 and BC18630 (Fig. S7F) . Direct target engagement was con rmed through a thermal shift assay (Fig. S8A ).

Both BC18813 and BC18630 decreased TFEB ubiquitination and signi cantly reduced the measured interaction between DCAF7 and TFEB in nucleus of cells (Fig. 4A, S8B-D) . Consistent with a role for DCAF7 in basal TFEB regulation, treatment with either DCAF7 inhibitor signi cantly increased nuclear levels of TFEB in a time and dose-dependent manner (Fig. S9A-H) in a wide variety of cell types (Fig.  S10A-D) . DCAF7 inhibition also resulted in a signi cant dose and time-dependent increase in multiple key TFEB-transcription targets (lysosomal proteases, autophagic transport, autophagy receptors) (Fig. S11A-D) . To further validate our DCAF7 inhibitors, we treated DCAF7 KO cells with our compounds, and observed no further increase in nuclear TFEB level (Fig. S12A-B) . Similarly, while DCAF7 inhibitor treatment increased lysosomal number (Lysotracker) and activity (Magic Red) in WT cells, DCAF7 KO cells were insensitive to these agents (Fig. 4B , S12C-H). To assess the speci city of these compounds, we also tested their effects on known protein substrates of other closely related E3 ligases to DCAF7; DCAF7 inhibitors selectivity increased nuclear TFEB protein levels, without altering other substrates (Fig. S13A ). In addition, both compounds exhibited minimal cellular toxicity at concentrations below 10 µM ( Fig.  S13B-C) . We found that BC18630 displayed a superior pharmacokinetic pro le in rodents compared to BC18813, and that BC18630 appeared to increase hepatic endo-lysosomal capacity in mice, as measured by dextran cascade blue uptake [27] [28] [29] (Fig. S14A-C) . Thus, these small molecule inhibitors of DCAF7 can function to increase basal TFEB levels and lysosomal activity in vitro and in vivo.

As our DCAF7 inhibitors demonstrated the capacity to modulate basal TFEB levels and lysosomal activity, we next sought to explore the utility of these compounds in the setting of viral infection. DCAF7 inhibitors prevented virally-triggered TFEB degradation and dose-dependently reduced OC43 coronavirus infection (Fig.4C, S15A) . Consistent with previous observations 10 , coronavirus infection led to a signi cant increase in lysosomal pH (Fig. 4D) . However, both BC18813 and BC18630 inhibited this virally-induced alkalization (Fig. 4D) . These agents appeared to be relatively potent as we estimated that the reduction of HCoV-OC43 NP expression occurred at an IC50 of roughly 0.3 µM for both inhibitors (Fig.  S15B, C) . Further, both DCAF7 inhibitors strongly decreased the proportion of OC43 positive cells, (Fig.  S15D, E) . DCAF7 inhibitors also protected against 229E induced infection and cytopathic effects (Fig.  S15F-J) . Further, BC18813 demonstrated e cacy in maintaining TFEB protein level in the setting of infection with the structurally unrelated in uenza virus, H1N1 (Fig. S15L) .

Given the e cacy of this strategy to treat several relatively common coronaviruses, we next assessed the effect of modulating TFEB in the setting of live SARS-CoV-2 infection. We utilized the BSL3 facilities at the IIT Research Institute to test DCAF7 inhibitors in SARS-CoV-2 infection assays with Calu-3 cells. Both compounds signi cantly reduced SARS-CoV-2 viral RNA load and cellular infection as measured by SARS-CoV-2 nucleoprotein immuno uorescence (Fig. 4E-G, S16A-C) . To further characterize the utility of our DCAF7 inhibitors, we next assessed an in vivo model of SARS-CoV-2 infection. We chose to study BC18630, given its favorable pharmacokinetic pro le ( Fig. 14A-C) . We employed a Syrian hamster model of SARS-CoV-2 infection and treated with vehicle or two doses of BC18630 (oral BID at 20mg/kg and 50 mg/kg for 5 days) (Fig. 4H) . We observed that BC18630 dose-dependently reduced lung viral titer (PFU) at all time points (Fig. 4I) . The higher dose strategy produced a >90% viral load reduction at day 4 and 6. Immunohistochemical staining of lung slices for SARS-CoV-2 N-protein (NP) at day 4 post infection revealed signi cantly reduced epithelial NP-positive signal with BC18630 treatment (Fig. 4J-K) . BC18630 also reduced the observed level of lung injury, with the higher dose particularly reducing in ammatory in ltrates (Fig. 4L) . In summary, BC18630 appears to decrease viral titer, reduce viral protein abundance, and prevent adverse pathophysiological changes in the lungs of hamsters infected with SARS-CoV-2. These data demonstrate the utility of pharmacological-based methods to modulate TFEB levels as an anti-viral strategy.

Previous studies have demonstrated that many viruses have evolved mechanisms to circumvent or hijack the endo-lysosomal pathway 1, 30, 31 . Our data demonstrates that viral infection induces the proteasomal mediated degradation of TFEB. We observed that PAK2 activation is a critical priming step for TFEB degradation, as this kinase is activated by viral infection, generating a phospho-degron required for subsequent DCAF7 recognition (see model in Fig S17) . Other groups have independently characterized the PAK kinase family, including PAK2, as associated with viral infectivity, including recent studies with SARS-CoV-2 32,33 . Intriguingly, either PAK2 deletion or genetic/pharmacological DCAF7 inhibition ablated viral-induced TFEB degradation, thereby maintaining lysosomal tness and improving the host response. While our study has concentrated on TFEB during infection, the e cacy of DCAF7 inhibitors to modulate TFEB activity under basal conditions suggests potential application of this approach for other, noninfectious conditions associated with impaired auto-lysosomal activity. Indeed, pharmacological activation of TFEB represents an attractive target for a growing number of diseases 5,34-36 . DCAF7 inhibitors showed e cacy against SARS-CoV-2 in cell and hamster-based infection models. With the rapid spread of SARS-CoV-2, and the high burden of acute respiratory failure, there is an urgent need for additional off-the-shelf therapies. Several small molecules have been investigated but each has drawbacks. For instance, remdesivir showed only modest clinical effects, that were most pronounced in severely ill patients 37, 38 . In contrast to viral-centered interventions, modulators of the DCAF7-TFEB-endolysosomal axis function to augment intrinsic cellular defense mechanisms. One appeal of this strategy is that it should hypothetically remain e cacious against future novel viral strains that might emerge. Uninvestigated in this study is the immunological consequences of TFEB augmentation. Increased lysosomal tness among the sentinel epithelial compartment may also enhance antigen presentation generating a more potent immune-mediated clearance response. Taken together, our observations demonstrate a novel mechanism of TFEB regulation and reveal a host-centric, endo-lysosomal focused strategy to limit viral infections. Viral infection activates PAK2 kinase to prime TFEB for degradation. A. High content screening of kinase regulators of TFEB-EGFP nuclear localization. BEAS-2B cells stably expressing TFEB-EGFP were screened with an RNAi library targeting all human kinases prior to quanti cation of TFEB nuclear to cytosolic ratio.

The kinase PAK2 was identi ed as a top regulator of TFEB nuclear levels and con rmed by immunoblotting of PAK2 siRNA treatment in BEAS-2B cells (inset). B. Representative images of TFEB-EGFP localization following control or PAK2 RNAi treatment. PAK2 siRNA increased nuclear TFEB levels.

Boxed area is shown at higher magni cation (Zoom). Scale bar = 50 µm. C. Immunoblot analysis demonstrating PAK2 activation (Serine-20 phosphorylation) in BEAS-2B following OC43 infection (48h), 

Autophagy during viral infection -a double-edged sword

A Gene Network Regulating Lysosomal Biogenesis and Function

TFEB links autophagy to lysosomal biogenesis

A lysosome-to-nucleus signalling mechanism senses and regulates the lysosome via mTOR and TFEB

TFEB at a glance

Innate host defense requires TFEB-mediated transcription of cytoprotective and antimicrobial genes

The Legionella effector RavZ inhibits host autophagy through irreversible Atg8 deconjugation

Mycobacterium tuberculosis is protected from NADPH oxidase and LC3-associated phagocytosis by the LCP protein CpsA

A Bacterial Effector Reveals the V-ATPase-ATG16L1 Axis that Initiates Xenophagy

β-Coronaviruses Use Lysosomes for Egress Instead of the Biosynthetic Secretory Pathway

The common cold

The adaptor protein DCAF7 mediates the interaction of the adenovirus E1A oncoprotein with the protein kinases DYRK1A and HIPK2

Wdr68 Mediates Dorsal and Ventral Patterning Events for Craniofacial Development

DYRK1A regulates Hap1-Dcaf7/WDR68 binding with implication for delayed growth in Down syndrome

Computational Prediction of Ubiquitination Proteins Using Evolutionary Pro les and Functional Domain Annotation

A combinatorial F box protein directed pathway controls TRAF adaptor stability to regulate in ammation

Structural basis of DDB1-and-Cullin 4-associated Factor 1 (DCAF1) recognition by merlin/NF2 and its implication in tumorigenesis by CD44-mediated inhibition of merlin suppression of DCAF1 function

Activation of TNF-alpha/NF-kappaB axis enhances CRL4B(DCAF)(11) E3 ligase activity and regulates cell cycle progression in human osteosarcoma cells

Mechanisms and function of substrate recruitment by Fbox proteins

A partnership with the proteasome; the destructive nature of GSK3

Chemically Diverse Group I p21-Activated Kinase (PAK) Inhibitors Impart Acute Cardiovascular Toxicity with a Narrow Therapeutic Window

Precise design of highly isoform-selective p21-activated kinase 4 inhibitors: computational insights into the selectivity mechanism through molecular dynamics simulation and binding free energy calculation

The proin ammatory role of HECTD2 in innate immunity and experimental lung injury

Ubiquitin E3 ligase FIEL1 regulates brotic lung injury through SUMO-E3 ligase PIAS4

Chromatin-modifying complex component Nurf55/p55 associates with histones H3 and H4 and polycomb repressive complex 2 subunit Su(z)12 through partially overlapping binding sites

Abnormal transport along the lysosomal pathway in mucolipidosis, type IV disease

Sorting of an internalized plasma membrane lipid between recycling and degradative pathways in normal and Niemann-Pick, type A broblasts

Herpes Simplex Virus: The Hostile Guest That Takes Over Your Home

Autophagy and Viral Infection

Actionable Cytopathogenic Host Responses of Human Alveolar Type 2 Cells to SARS-CoV-2

An emerging role for p21-activated kinases (Paks) in viral infections

Lysosome biogenesis in health and disease

Key Roles of MiT Transcription Factors in Innate Immunity and In ammation

Transcription factor EB: an emerging drug target for neurodegenerative disorders

Compassionate remdesivir treatment of severe Covid-19 pneumonia in intensive care unit (ICU) and Non-ICU patients: Clinical outcome and differences in post-treatment hospitalisation status

Compassionate Use of Remdesivir for Patients with Severe Covid-19

TFEB protein densitometry was corrected to β-actin and normalized to WT-control. Data represent mean ±SEM (n=3). G. Quanti cation of viral infection through an in-cell ELISA assay of WT or PAK2 KO BEAS-2B cells infected with OC43 (48h). Data were normalized to the WT MOI=0.1 treatment levels and represent mean ±SEM (n=6). H. A canonical E3 ligase phospho-degron motif 21 S/T XXX S/T is present within TFEB at serine-138 and serine-142. I. Binding assay of TFEB serine-alanine mutants with DCAF7. In vitro synthesized V5-tagged TFEB mutants were incubated with immunoprecipitated DCAF7 prior to washing, elution, and immunoblot analysis

Schematic of PAK2-induced priming phosphorylation required for DCAF7-mediated poly-ubiquitination

0001; compared to vehicle or control or as indicated by two-way ANOVA with Tukey's multiple comparisons

After a 75 min inoculation, cells were washed and media was replaced with compound at the indicated concentration. Samples of the supernatant were taken 48 hours later to quantify viral RNA levels. Data in panel E represents mean ± SEM (n=6), and IC50 values were determined by sigmoidal nonlinear regression. Cells were xed and stained for SARS-CoV-2 nucleoprotein (NP) for quanti cation of viral signal (F-G). Data represent mean ± SEM (n=6), and IC50 values were determined by sigmoidal nonlinear regression

Six hours later, a second dose of BC18630 was administered and thereafter animals were treated BID for 5 days. Animals were sacri ced at days 2, 4, and 6 for analysis

Data represent mean ± SEM of technical replicates (n=3). J. Lung samples from 2 animals at day 4 post infection were stained for immunohistochemical detection of SARS-CoV-2 NP expression in tissues. K. Semi-quanti cation of IHC staining in panel (F)

0001; as indicated by one-way ANOVA with Dunnett's multiple comparisons

The authors declare the following competing interests: BBC, TF and YL are listed on a patent application based on the work described here. BBC, TF and YL are the founders and employees of Generian Pharmaceuticals and may own company stock.