key: cord-0274896-upxhdvcv
authors: Gupta, Ravi K.; Mlcochova, Petra
title: Cell cycle independent role of cyclin D3 in host restriction of SARS-CoV-2 infection
date: 2022-05-08
journal: bioRxiv
DOI: 10.1101/2022.05.07.491022
sha: e31be4635eed0eec3f0c3a4da36b085adaa598a5
doc_id: 274896
cord_uid: upxhdvcv

The COVID-19 pandemic caused by Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) presents a great threat to human health. The interplay between the virus and host plays a crucial role in successful virus replication and transmission. Understanding host-virus interactions is essential for development of new COVID-19 treatment strategies. Here we show that SARS-CoV-2 infection triggers redistribution of cyclin D1 and cyclin D3 from the nucleus to the cytoplasm, followed by its proteasomal degradation. No changes to other cyclins or cyclin dependent kinases were observed. Further, cyclin D depletion was independent from SARS-CoV-2 mediated cell cycle arrest in early S phase or S/G2/M phase. Cyclin D3 knockdown by small interfering RNA specifically enhanced progeny virus titres in supernatants. Finally, cyclin D3 co-immunoprecipitated with SARS-CoV-2 Envelope and Membrane proteins. We propose that cyclin D3 inhibits virion assembly and is depleted during SARS-CoV-2 infection to restore efficient assembly and release of newly produced virions.

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the causative agent for 26 the global Covid-19 pandemic. To date, SARS-CoV-2 has infected over 265 millions of people 27 with death toll of more than 5 million people 1 . While strategies of counteracting SARS-CoV-28 2 infection through vaccination have been partially successful, there is still a need for effective 29 antiviral drugs given the emergence of vaccine escape variants such as Omicron. 30

Coronaviruses, including SARS-CoV-2, like other viruses are intracellular pathogens 31 exploiting the host cell machinery to their own advantage. The identification of cellular 32 mechanisms and host cell targets required for SARS-CoV-2 life cycle will provide us with new 33 knowledge that could be used to interfere with viral replication and therefore presents an 34 alternative approach to block viral infection. 35

Cyclins and cyclin dependent kinases (CDKs) are the major regulators of cell cycle 36 progression. Many viruses, including coronaviruses, adopt a strategy of manipulating cell cycle 37 progression through cyclin-CDKs complexes 2-5 to facilitate viral replication. Several SARS-38

CoV-1 proteins have been shown to reduce cyclin D and cyclin E and A expression that is 39 connected to cell cycle arrest 6-8 . For example, SARS-CoV-1 N protein directly interacts with 40 cyclin D to prolong the S phase 6 that ensure enough supply of nucleotides for viral replication. 41

Nsp13 protein both in SARS-CoV-1 and Infectious bronchitis virus (IBV) interacts with DNA 42 polymerase subunit to induce DNA damage and cell cycle arrest 9 . It is believed that the virus 43 infection associated cell cycle arrest increases essential DNA repair processes and replication 44 proteins that are required by virus replication. 45 to transduce cells. Cells were then infected with a replication competent strain of SARS-CoV-125 2 for 24h, fixed and stained for nucleocapsid. Flow cytometry analysis was used to visualise 126 G1, early S and S/G2/M phase. The FUCCI system cannot demonstrate G0 phase as it is 127 defined as a cell population void of any fluorescent protein but cannot be differentiated from 128 an untransduced cell population. Cells were gated on infected (expressing SARS-CoV-2 129 nucleocapsid) or uninfected cells with Cdt1/Geminin expression determined in both gated 130 populations and compared (Fig. 4A-D) . Indeed, SARS-CoV-2 infection mediated cycle arrest 131 in S/G2/M phase in VERO AT2 cells (Fig. 4C ), confirming previously published data 10 . 132

Interestingly, SARS-CoV-2 mediated cell cycle arrest in A549 AT2 cells was identified 133 specifically in early S phase (Fig. 4D ). This cell cycle arrest was caused by productive SARS- Furthermore, cell cycle kinetics in virus exposed but uninfected cells were similar to unexposed 139 uninfected cells (Fig. S6 F,G) . 140

These observations support the hypothesis that a productive SARS-CoV-2 infection is 141 responsible for cell cycle arrest and that the arrest is not the result of by-stander effects. 142

However, the data suggest that the specific phase in which cells are arrested appears to be cell-143 type dependent. 144

To dissect whether the increase in viral titer is a direct consequence of cylin D depletion or 147 aftermath of cell cycle arrest caused by the absence of this cyclin we investigated cell cycle 148 arrest during SARS-CoV-2 infection and linked it to cyclin D expression/cellular localization. 149

It has been previously reported that the Cyclin D degradation is sufficient to cause cell cycle 150 arrest in G1 phase 16,17 . We confirmed cell cycle arrest in G1 phase in A549 AT2 after D-cyclin 151 depletion (Fig. S7A,B) . Further VERO AT2 cells showed similar G1 arrest after individual D1 152 and D3 cyclin depletion but as well when both D1 and D3 cyclins were depleted together 153 (Fig.S7C) . Importantly, no increase in early S nor S/G2/M phase have been observed after 154 cyclin D depletion in uninfected cells. On the contrary, a decrease in these phases has been 155 identified in concordance with more cells arresting in G1 phase and not progressing through 156 the cell cycle (Fig.S7 ). As we have already shown that SARS-CoV-2 infection increases 157 percentage of cells in early S (in A549 AT2 cells) and S/G2/M (in VERO AT2) specific cell cycle phases (Fig.4) , we investigated cell cycle progression in infected cells were D-cyclins 159 and cyclin A had been depleted (Fig. 5, S8 ). A549 AT2 cells have been depleted for D and A 160 cyclins (Fig. 5A ) and infected with Delta variant. The percentage of infected cells was 161 determined 24h later and a small increase (3 fold) compare to NT control was detected in cells 162 depleted for cyclin D3 (Fig.5B,C) . The A549 AT2 cell population in early S increased without 163 any changes in S/G2/M phase following infection in both non-targeted (NT) control cells and 164 cells depleted for D and A cyclins (Fig.5D ). This data were confirmed using an Alpha SARS-165

CoV-2 variant (Fig. S8A,B) . Furthermore, cell cycle in virus exposed but uninfected cells 166 Cyclin D3 has been previously implicated in the restriction of influenza A virus through 187 impairment of virus assembly 18 . Cyclin D3 has been shown to interact with IAV protein M2, 188 a ion channel that promotes viral replication 18,19 . Interestingly, SARS-CoV-2 E protein has 189 been suggested to be an ion channel 20,21 . In the light of our data showing cyclin D3 depletion 190 increased SARS-CoV-2 viral titer, we investigated potential implication of cyclin D3 in SARS-protein was investigated (Fig. 6A ). HA-tagged cyclin D3 was co-expressed together with tagged E and nsp9 protein in 293T cells. Nsp9 was chosen as a control on the basis of its diverse 194 cellular localization both in the nucleus and cytoplasm 22 . Of note, 293T cells showed 195 undetectable endogenous expression of cyclin D3 (Fig. S9A ). We showed that SARS-CoV-2 196

Envelope (E) coimmunoprecipitated with HA-tagged cyclin D3 using anti-HA antibody while 197 SARS-CoV-2 nsp9 protein did not (Fig. 6A ) suggestive of specific binding to E protein. leading to S is retention in the ER and Golgi, preventing syncytia formation 27 . It is possible 214 that this interaction between structural proteins and spike retention allows S to target the virion 215 assembly sites. We hypothesised that if cyclin D3 is interacting with E and M it might impact 216 their function in Spike processing/trafficking, and we can use it as a read-out by assessing the 217 syncytia formation. Firstly, a split GFP system 28 was used to confirm that E and M or 218 combination of both (E/M) impact Spike-mediated syncytia formation. Indeed, both structural 219 proteins when coexpressed with S significantly decreased GFP+ve area (cell-cell fusion) 220 ( Fig.6C,E) . Secondly, 293T GFP11 cells were transfected with full length S (WT), and/or with 221 E, M, and cyclin D3. 24h post-transfection cells were seeded at a 1:1 ratio with Vero-GFP10 222 and cell to cell fusion was measured 18h later to determine a proportion of green area to total 223 phase area ( Fig.6D-G) . Interestingly, cyclin D3 had no effect on reduction of syncytia when 224 expressed together with S or S + E or S + M (Fig.6F ). However, it increased syncytia formation in combination with S+M+E suggestive of compromising E and M impact on Spike 226 processing/trafficking towards the cell surface (Fig.6D,G) . 227

These data together demonstrate that cyclin D3 associates with proteins important for SARS-228

CoV-2 assembly and impairs their optimal function.

Here we show that SARS-CoV-2 infection depletes levels of cyclin D and suggest that this 232 depletion is independent from changes to cell cycle arrest in infected cells. Further, cyclin D3 233 seems to interfere with efficient SARS-CoV-2 assembly by interacting with Envelope and 234

Membrane SARS-CoV-2 proteins. Importantly, our data show that cyclin D3 associates with M as well, supporting our hypothesis 296 that cyclin D3 impairs SARS-CoV-2 assembly and spread. Further, E and M proteins have been implicated in Spike processing and trafficking 27 . It has been shown that Spike is retained 298 inside cells when expressed together with E and M probably to target S to proximity of 299 intracellular virus assembly sites. Our data show that S is retained in the cells in the presence 300 of M and E but its trafficking towards membrane and ability to form syncytia is partially 301 rescued when cyclin D3 is present. This supports the concept of cyclin D3 being restriction 302 factor impairing role of M and E in SARS-CoV-2 optimal assembly. 303

Our work provides important insight into mechanism through which cyclin D3 limits SARS-305

CoV-2 infection. In the light of immune evasion from vaccination, it is important that this 306 phenomenon was observed across different SARS-CoV-2 variants suggesting that this 307 mechanism provides a universal target for development of antivirals. Our data suggest that 308 cyclin D3 associates with SARS-CoV-2 E and M proteins, thereby interfering with efficient 309 assembly. SARS-CoV-2 has therefore evolved strategies to degrade cyclin D3 that require 310 further investigation, with the hope that it can be translated to therapeutics. proportion of green area to total phase area using ArrayScan high-content system 440 

COVID-19) Dashboard

Murine coronavirus replication induces cell cycle arrest in 448 G0/G1 phase

Murine coronavirus 450 nonstructural protein p28 arrests cell cycle in G0/G1 phase

Porcine epidemic diarrhea virus through p53-dependent pathway causes 453 cell cycle arrest in the G0/G1 phase

The M2 proton channels of influenza A and B viruses

SARS-CoV-2 envelope protein causes acute respiratory distress syndrome 499 (ARDS)-like pathological damages and constitutes an antiviral target

SARS-CoV-2 E protein is a potential ion channel that 502 can be inhibited by Gliclazide and Memantine

A systemic and molecular study of subcellular localization of SARS-505

CoV-2 proteins

The M, E, and N structural proteins of the severe acute respiratory 508 syndrome coronavirus are required for efficient assembly, trafficking, and release of 509 virus-like particles

Assembly and Entry of Severe Acute Respiratory Syndrome 511 Coronavirus 2 (SARS-CoV2): Evaluation Using Virus-Like Particles

SARS-CoV-2 viral budding and entry can be modeled using BSL-514 2 level virus-like particles

Structural and Functional Analysis of the D614G SARS-CoV

The SARS-CoV-2 envelope and membrane proteins modulate

VERO AT2 Cells were fixed 24h post-infection and stained for viral proteins and cyclins

Arrowheads highlight un-infected cells and cyclin D/A nuclear localization. Arrowheads: 601 Nuclear cyclin staining in uninfected cells

Uninfected (-) and SARS-CoV-2 infected cells were identified by 604 negative/positive nucleocapsid or Spike staining. Ratio between nuclear and cytoplasm 605 staining intensity of cyclins was measured using ImageJ and Harmony (PerkinElmer)

Student's t-tests; ns, non-significant

A549 AT2 cells were infected with WT, alpha (α), and delta (∆) SARS-CoV-2 variants

Cells were lysed 48h post-infection and viral, cyclin proteins expression was analysed by 610 western blot. N, nucleocapsid

Uninfected (-) and SARS-CoV-2 infected cells were identified by negative/positive 613 nucleocapsid or Spike staining. Ratio between nuclear and cytoplasm (N/C ratio) staining 614 intensity of cyclins was measured using ImageJ and Harmony (PerkinElmer)

Proteasome inhibition abolishes effect of SARS-CoV-2 infection on D-cyclins depletion 621 (A) A549 AT2 cells were infected with Delta SARS-CoV-2 variant

BZ, 1uM) was added to cells 18h post-infection. Cells were lysed 24h post-623 addition of inhibitor and cyclins and viral proteins were detected by western blot

Densitometry analysis of western blots for D-cyclins (normalized to actin) in A549 AT2 626 cells. Plots are average of 3 independent experiments. Bars indicate mean with SD

Statistical analysis was performed using ordinary two way ANOVA; ns, non-significant

A459 AT2 cells were infected with Delta SARS-CoV-2 variant

BZ, 1uM) was added to cells 8h post-infection. Cells were fixed and stained 631 24h post-addition of inhibitor

Quantification of D3 cyclin relocalization from nucleus after infection

SARS-CoV-2 infected cells were identified by negative/positive nucleocapsid staining

/C ratio) staining intensity of cyclins was 635 measured using ImageJ and Harmony (PerkinElmer)

Statistical analysis was performed using ordinary two-way ANOVA

D and A-cyclins were depleted using siRNA in A549 AT2 cells. Cells were infected 18h 641 later with Delta (∆), Alpha (α) or wild type (WT) SARS-CoV-2 variants at MOI 0.001, 0.1, 0.1 642 respectively. Cells were washed 4h post-infection and new media were added

were depleted using siRNA in VERO AT2 cells. Cells were infected 645 18h later with Alpha (α) or wild type (WT) SARS-CoV-2 variants at MOI 0.1. Cells were 646 washed 4h post-infection and new media added

Representative example of western blot from cell lysates collected at 48h post-infection

D) Virus titres in cell culture supernatants were determined as TCID50 in VERO AT2 cells

Statistical analysis was performed using one-way ANOVA 651 with Dunnett's multiple comparisons test. NT, non-targeting control. ns, non-significant

* * p < 0.01. Bars indicate mean with SD

SARS-CoV-2 infection arrests cell cycle SARS-CoV-2 variants for additional 24h

Example of gating strategy for cell cycle analysis. The 659 population of cells exposed to SARS-CoV-2 was stained for nucleocapsid (N) protein and 660 gated on N+ve and N-ve population

See also Supplemental Figure S5

VERO AT2 or (D) A549 AT2 cells infected with WT SARS-CoV-2. Quantification of

Cdt1/Geminin +ve cells (early S phase) and Geminin +ve cells (S,G2,M 664 phase) cells. n = 5; one-way ANOVA with Dunnett's multiple comparisons test: ns, non-665 significant

* p < 0.0001. Bars indicate mean with SD

Quantification of cell cycle arrest after exposure of 667 cells to SARS-CoV-2 variants. α, alpha (MOI 0

n = 4; one-way ANOVA with Dunnett's multiple comparisons test: ns, non-significant

* p < 669 0.0001; * * * p < 0.001. Bars indicate mean with SD

SARS-CoV-2 mediated depletion of D-cyclins is cell cycle arrest independent 674 (A-D) A549 AT2 cells were depleted for D and A2 cyclins and 18h later infected with Delta 675 variant SARS-CoV-2 for 24h. Cells were fixed, stained for SARS-CoV-2 nucleocapsid and 676 analysed for infection and Fucci cell cycle sensor

A) A representative western blot from lysates of uninfected knock-down cells

Example of gating strategy for flow cytometry analysis

Percentage of infected cells in cells depleted for cyclins. n=3; Ordinary two-way ANOVA 680 with Sidak's multiple comparisons test: ns, non-significant

< 681 0.01; * p < 0.1. Bars indicate mean with SD

Flow cytometry analysis of early S and S/G2/M cell cycle phases comparing cyclin D1

NT (non-target siRNA). n = 3. Statistical analysis was performed 684 using two-sided unpaired Student's t-tests; ns, non-significant

* * p < 0.01; * p < 685 0.1. Bars indicate mean with SD

VERO AT2 cells were transduced with VSV-G pseudotyped Fucci containing lentiviral 687 particles and 18h later infected with Delta variant SARS-CoV-2. Cells were fixed and stained 688 for D-cyclins

Example of acquisition using automated microscopic platform. Cells are identified for 690 infection, expression of cyclin D3, and cell cycle

Quantification of D-cyclins re-localization from nucleus to cytoplasm and correlation 693 with cell cycle phases using ImageJ and Harmony (PerkinElmer). (F) Cyclin D3

At least 50-200 cells were analysed in each condition. Statistical analysis was performed 695 using two-sided unpaired Student's t-tests

A) 293T cells were contransfected with HA -cyclin D3 and Strep-tag-SARS-CoV-2 E or nsp9, 700 and control plasmid (EV). Immunoprecipitation was performed using anti-HA antibody. The 701 immunoprecipitates were blotted with anti-Strep

293T cells were contransfected with HA -cyclin D3 and Strep-tag-SARS-CoV-2 E, M, 703 both E and M or nsp9. Immunoprecipitation was performed using anti-HA antibody. The 704 immunoprecipitates were blotted with anti-Strep, anti-HA antibodies. WCL, whole cell lysate

293T GFP11 cells were transfected with Spike, and/or with Envelope, Membrane, and 706 cyclin D3. 24h post-transfection cells were seeded at a 1:1 ratio with Vero-GFP10 cells and 707 percentage of GFP+ve area (syncytia) were determined

D) Representative images of GFP+ syncytia

Quantification of cell-to-cell fusion showing percentage of the GFP+ve area to the 711 acquired total cell area. n = 5; one-way ANOVA with Dunnett's multiple comparisons test: ns

* * p < 0.01 . Bars indicate mean with SD

The authors declare that all data supporting the findings of this study are available within the 444 article and the Supplementary Information. 445