Host IP3R channels are dispensable for rotavirus Ca2+ signaling but critical for intercellular Ca2+ waves that prime uninfected cells for rapid virus spread

ABSTRACT Rotavirus is a leading cause of viral gastroenteritis. A hallmark of rotavirus infection is increased cytosolic Ca2+ caused by nonstructural protein 4 (NSP4). NSP4 is a viral ion channel that releases endoplasmic reticulum (ER) Ca2+, and the increased Ca2+ signaling is critical for rotavirus replication. In addition to NSP4, host inositol 1,4,5-trisphosphate receptor (IP3R) ER Ca2+ channels may contribute to rotavirus-induced Ca2+ signaling and by extension, virus replication. Thus, we set out to determine the role of IP3R Ca2+ signaling during rotavirus infection using CRISPR/Cas9 IP3R-knockout of MA104 cells stably expressing the GCaMP6s Ca2+ indicator (MA104-GCaMP6s-IP3R-KO). Live Ca2+ imaging showed that IP3R-KO did not reduce Ca2+ signaling in infected cells but eliminated rotavirus-induced intercellular Ca2+ waves (ICWs) and, therefore, the increased Ca2+ signaling in surrounding, uninfected cells. MA104-GCaMP6s-IP3R-TKO cells showed similar rotavirus susceptibility, single-cycle replication, and viral protein expression as parental MA104-GCaMP6s cells. However, MA104-GCaMP6s-IP3R-TKO cells exhibited significantly smaller rotavirus plaques, decreased multi-round replication kinetics, and delayed virus spread, suggesting that rotavirus-induced ICW Ca2+ signaling stimulates virus replication and spread. Inhibition of ICWs by blocking the purinergic receptor P2RY1 (P2Y1), which mediates the ICW Ca2+ signals, also decreased rotavirus plaque size. Conversely, exogenous expression of P2Y1 in LLC-MK2-GCaMP6s cells, which natively lack P2Y1 and rotavirus ICWs, rescued the generation of rotavirus-induced ICWs and enabled plaque formation. In conclusion, this study shows that NSP4 Ca2+ signals fully support rotavirus replication in individual cells; however, IP3R is critical for rotavirus-induced ICWs and virus spread by priming Ca2+-dependent pathways in surrounding cells. IMPORTANCE Many viruses exploit host Ca2+ signaling to facilitate their replication; however, little is known about how Ca2+ signals from different host and viral channels contribute to the overall dysregulation of Ca2+ signaling or promote virus replication. Using cells lacking IP3R, a host ER Ca2+ channel, we delineated intracellular Ca2+ signals within virus-infected cells and intercellular Ca2+ waves (ICWs), which increased Ca2+ signaling in neighboring, uninfected cells. In infected cells, IP3R was dispensable for rotavirus-induced Ca2+ signaling and replication, suggesting the rotavirus NSP4 viroporin supplies these signals. However, IP3R-mediated ICWs increase rotavirus replication kinetics and spread, indicating that the Ca2+ signals from the ICWs may prime nearby uninfected cells to better support virus replication upon eventual infection. This “pre-emptive priming” of uninfected cells by exploiting host intercellular pathways in the vicinity of virus-infected cells represents a novel mechanism for viral reprogramming of the host to gain a replication advantage.

C alcium (Ca 2+ ) signaling is a cornerstone of cellular communication and is critical for maintaining homeostasis and responding to damage or infection (1).As such, it is unsurprising that many viruses have evolved strategies to exploit these pathways to facilitate virus replication and spread (2).However, the highly interconnected nature of Ca 2+ signaling has made it challenging to understand how viral and host proteins interact to orchestrate the patterns of Ca 2+ dysregulation that are observed during infection.One key strategy for virus-induced Ca 2+ dysregulation is the expression of viral ion channels, or viroporins, which can conduct Ca 2+ across host cell membranes (3).The rotavirus nonstructural protein 4 (NSP4) is among the most well-characterized Ca 2+ conducting viroporins, and rotavirus has become a leading model system to study how viruses exploit Ca 2+ signaling to promote their replication (4).Yet, the interplay between rotavirus-and host-induced Ca 2+ signaling pathways remains incompletely characterized.
NSP4 is a multifunctional endoplasmic reticulum (ER), transmembrane glycoprotein.Within infected cells, NSP4 functions both as a viroporin to dysregulate Ca 2+ signaling and traffics to viroplasm-associated membranes to serve as an intracellular receptor for VP6 on immature double-layered particles, facilitating the assembly of the rotavirus outer capsid proteins VP4 and VP7 (5)(6)(7).Numerous studies show that NSP4 elevates cytosolic Ca 2+ levels during infection by increasing the Ca 2+ permeability of the ER through the formation of a Ca 2+ permeable viral ion channel (8)(9)(10)(11)(12).The NSP4-mediated release of ER Ca 2+ , in turn, activates a host process known as store-operated calcium entry (SOCE) to further elevate cytosolic Ca 2+ levels (13).This is critical for multiple steps in rotavirus replication, including the activation of the autophagy pathway and assembly of the rotavirus outer capsid protein, VP7 (13,14).Yet, NSP4 may play both a direct and an indirect role in ER Ca 2+ release, through interaction with host channels.The two main eukaryotic ER Ca 2+ release channels are the ryanodine receptor (RyR) and the inositol-1,4,5-trisphosphate receptor (IP 3 R), and while RyR expression is generally restricted to excitable cells, IP 3 R is widely expressed in most cell types (15).Mammals have three IP 3 R genes, each with multiple splice variants that provide nuance to the regulation of IP 3 R Ca 2+ release and, therefore, the ability to shape Ca 2+ signals (15,16).The IP 3 R channel is activated by IP3, and channel opening is regulated by cytosolic Ca 2+ in a biphasic manner, such that low cytosolic Ca 2+ levels potentiates ER Ca 2+ release and increasing cytosolic Ca 2+ levels inhibit ER Ca 2+ release (17).Thus, even though the NSP4 viroporin directly alters ER and cytosolic Ca 2+ , NSP4 Ca 2+ release may also potentiate IP 3 R activity, increasing the Ca 2+ dysregulation of host cells.
In addition to NSP4 and IP 3 R crosstalk within infected cells, we recently discovered that rotavirus infection triggers intercellular Ca 2+ waves (ICWs) through the release of ADP from infected cells and activation of P2Y1 purinergic receptors on neighboring cells.Activation of P2Y1 receptors results in an IP 3 R-mediated ER Ca 2+ signal, and since many ICWs are produced during infection, this significantly increases Ca 2+ signaling in neighboring, uninfected cells as well (18).Thus, the ER Ca 2+ store is a critical source of Ca 2+ signals in both rotavirus-infected and the nearby uninfected cells, and IP 3 R has the potential to significantly impact the overall landscape of rotavirus-induced Ca 2+ signaling dysregulation.
The goal of this study is to examine the contribution of IP 3 R to rotavirus Ca 2+ dysregulation and by extension, its role in rotavirus infection, replication, and spread.While HEK293 and HeLa cells with genetic knockout of all three IP 3 R genes (e.g., IP 3 R1, IP 3 R2, and IP 3 R3) have been established previously, these cell lines are suboptimal for studying rotavirus replication, and it is not known whether they produce the robust rotavirus-induced ICW signals that were previously characterized using MA104 cells and human intestinal organoids (16,19).Thus, we generated an IP 3 R triple knockout in MA104 cells, a vervet monkey (Chlorocebus pygerythrus) kidney cell line commonly used to study rotavirus (18,20,21).We found that IP 3 R was not necessary for aberrant Ca 2+ signaling in infected cells but was necessary for Ca 2+ dysregulation in neighboring, uninfected cells.In parallel, IP 3 R was dispensable for single-round rotavirus infection and replication, but lack of IP 3 R, and therefore lack of ICW signaling, strongly reduced rotavirus spread.This study provides new insights into mechanisms exploited by viruses not only to reprogram infected cells but also to pre-program surrounding cells for subsequent rounds of infection.

CRISPR/Cas9 development and transduction
MA104 cells lacking IP 3 R expression (MA104-GCaMP6s-IP 3 R-TKO) were generated by lentivirus transduction to introduce Cas9 and small-guide RNAs (gRNAs) to IP 3 R1, IP 3 R2, and IP 3 R3 (Table 1).The lentivirus construct was designed by our lab and manufactured by VectorBuilder (Chicago, IL).MA104-GCaMP6s cells were transduced with MOI 10 in complete DMEM supplemented with 10 µg/mL polybrene (22).At 72 h post-transduc tion, cells were passaged in the presence of 40 µg/mL blasticidin.After 2 weeks of selection, cells were dilution cloned and resulting clones were screened for lack of agonist-induced IP 3 R Ca 2+ responses, as described below.

Sequencing of IP 3 R triple-knockout
Genomic DNA was extracted using a PureLink gDNA mini kit (Invitrogen, USA).PCR amplification was performed using KOD Hotstart polymerase kit (EMD Millipore) and primers flanking the gRNA target sites (Table 2).PCR products were cloned in TOPO-TA vectors (Invitrogen, USA), and a minimum of six bacterial colonies for each IP 3 R gene were sequenced using M13 forward and reverse primers (Azenta, NJ, USA).To determine the distribution of mutant alleles, sequences were mapped to the genomic DNA using SnapGene.

LLC-MK2-GCaMP6s P2Y1 knock-in generation
LLC-MK2-GCaMP6s P2Y1 knock in lines were generated by lentiviral transduction of the human P2Y1 receptor.The P2Y1 cDNA clone was purchased from GenScript (Piscataway, NJ, USA) and subcloned into pLVX-IRES-Neo lentivirus vector (Takara Bio) and packaged by the BCM Vector Development Core.LLC-MK2-GCaMP6s cells were transduced with MOI 10 in 10 µg/mL polybrene and at 72 h post-transduction selected using 500 µg/mL G418 (22).The presence of functional P2Y1 was verified by Ca 2+ imaging for a Ca 2+ response to 25 nM ADP.

Viral infectivity assay
MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO monolayers grown in 96-well plates were inoculated with twofold serial dilutions of SA11-mRuby.The inoculum was removed 1 h post-infection, and then the cells were rinsed twice with PBS and cultured in FBS-free DMEM for ~16 h.Monolayers were fixed in ice cold methanol for 20 min at 4°C, washed three times with PBS, and immunostained for 2 h using a rabbit anti-rotavirus antisera (strain Alabama) at 1:1,000 in PBS, followed by incubation for 1 h with goat anti-rabbit IgG:AlexaFluor555 at 1:1,000 in PBS, and quantified by fluorescence microscopy.

Single-cycle and multi-cycle rotavirus yield assay
MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO monolayers were inoculated with SA11-mRuby for 1 h.We used an MOI of 10 for single-cycle replication assays and 25 PFU/well for multi-cycle replication assays.The 25 PFU infection dose (MOI = 5.9 × 10 −5 ) was chosen empirically, based on plaque assays in parental MA104-GCaMP6s cells, as a low virus input that would maximize the number of uninfected cells but result in total monolayer destruction by 4 days post infection.For single-cycle assays, cells were maintained in DMEM without trypsin, which prevents rotavirus spread due to lack of VP4 spike protein cleavage, and harvested at 24 hpi.For multi-cycle assays, cells were maintained in DMEM with 1 µg/mL Worthington's trypsin and harvested at 24, 48, 72, and 96 hpi.Virus titration was performed by plaque assay (see below).

Plaque assay
Monolayers of MA104-GCaMP6s or LLC-MK2-GCaMP6s cells, or their modified derivatives, were inoculated with 10-fold serial dilutions of virus samples in FBS-free DMEM.After the inoculum was removed, 3 mL of overlay (1.2% Avicel in FBS-free DMEM supplemented with 0.1 mg/mL DEAE dextran and 1 µg/mL Worthington's trypsin) was added to each well, and plates were incubated for 3-4 days.After removing the overlay, plates were washed and stained with crystal violet or imaged by fluorescence microscopy prior to crystal violet staining.

Immunofluorescence
MA104 cell monolayers infected with SA11-mRuby (MOI 1) were fixed at 9 hpi, washed and incubated overnight at 4°C with rabbit antisera against NSP4 (anti-NSP4-aa114-135) and guinea pig antisera against NSP5 (gift from Dr. Mary Estes at Baylor College of Medicine) (25,26).Monolayers were washed and incubated for 2 h at room temperature with fluorescent-conjugated secondary antibodies [anti-Rabbit DyLight 488; anti-guinea pig DyLight 549 (Rockland)] diluted 1:2,000 in PBS.Monolayers were imaged using a 63× objective on a Zeiss LSM980 confocal microscope with Airyscan.Mander's coefficients were determined by calculating the ratio signal intensity for NSP4 colocalized with that of NSP5 to the overall intensity of the NSP4 signal in a given image following background subtraction.Analyses used three images per group for a total of 46 parental MA104-GCaMP6s and 63 MA104-GCaMP6s-IP 3 R-TKO cells.

Detection of ADP release from cells
Confluent MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO cells were infected with SA11-mRuby (MOI 10) and then maintained in 300 µL FluoroBrite-Plus with 20 µM ARL67156, an ecto-ATPase inhibitor.At ~5 hpi, the supernatant was removed and used to treat MA104-GCaMP6s cell monolayers to evoke a Ca 2+ response, which was monitored by live imaging of GCaMP6s fluorescence and quantitated by measuring the maximum increase in GCaMP6s fluorescence after addition of the supernatant.The specificity of the Ca 2+ response for P2Y1 activation was determined by pretreating cells with 10 µM BPTU, a P2Y1-selective blocker.

Quantitative PCR
LLC-MK2-GCaMP6s and LLC-MK2-GCaMP6s + P2Y1 cells were grown to confluency in a 24-well plate.Total RNA was extracted from cell monolayers using TRIzol (Ambion) according to the manufacturer's instructions.Three hundred nanograms of RNA was used to generate cDNA with SensiFAST synthesis reagents (Bioline).Quantitative PCR was performed with Fast SYBR green (Applied Biosystems) with a QuantStudio 3 thermocy cler.Purinergic receptor genes were normalized to 18 s, and expression relative to the lowest expressed gene (P2Y12) was calculated using the ΔΔC t method.Primer sequences have been published previously and can be found in Table 3 (27).

SA11-mRuby detection in plaques
Rotavirus SA11-mRuby plaques were visualized by fluorescence microscopy using a Nikon TiE inverted microscope as described above.Images were taken from a 6-well tissue culture plate with a 10× Plan Fluor objective (NA 0.30).Whole-well images were obtained by taking a 15 mm × 15 mm stitch with blending and 10% overlap.

Statistical analyses
Statistics were completed using GraphPad Prism (version 8.4.3).Unless stated, all experiments were performed in biological triplicate.The threshold for Ca 2+ spikes is a 5% increase in GCaMP fluorescence, as previously defined for this system (21).We performed column statistics to determine the normality of the data sets.We used unpaired Student's t test for data sets with a parametric distribution or a Mann-Whitney test for data sets with a nonparametric distribution.We applied a one-way ANOVA with proper correction when applicable for comparing sample groups to control groups.Differences were determined statistically significant if the P value was <0.05.For all figures, P value notations are as follows: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

RESULTS
Previous studies have shown that rotavirus NSP4's viroporin function is a critical mechanism for the dysregulation of cellular Ca 2+ homeostasis during infection (10,21).Yet, NSP4 crosstalk with host IP 3 R channels could contribute to the overall Ca 2+ signaling in infected cells and the recently discovered intercellular Ca 2+ waves (ICWs) induced by rotavirus are generated by IP 3 R Ca 2+ release upon activation of P2Y1 receptors (15)(16)(17)(18).Thus, we investigated the role of IP 3 R in the rotavirus-induced Ca 2+ signaling landscape using IP 3 R-null cell lines generated by genomic editing.First, we obtained IP 3 R tripleknockout (IP 3 R-TKO) HEK293 cells, which express no functional IP 3 R isoforms and tested whether rotavirus infection elevates Ca 2+ signaling similar to MA104 cells (16).Using long-term, time-lapse Ca 2+ imaging, we found that rotavirus-infected MA104-GCaMP6s cells exhibited a dynamic increase in Ca 2+ signaling in both the RV-infected and neighboring uninfected cells, consistent with our previous studies (18,21).To dissect these Ca 2+ signals further, we examined the imaging data and graphs of the GCaMP6s signal from both rotavirus-infected and neighboring uninfected cells (Fig. 1A and B).
Representative images show both intracellular Ca 2+ signals observed in rotavirus-infected cells (at 354 min and 360 min) and ICWs that propagate from infected cells to surround ing uninfected cells (396 min) (Fig. 1A; Movie S1).Representative traces of GCaMP6s fluorescence show mock-infected cells exhibited no strong Ca 2+ signaling events (Fig.  1B, black line).In contrast, rotavirus-infected cells (Fig. 1B, red line) exhibited a strong increase in discrete Ca 2+ signaling events, and the GCaMP6s trace from an uninfected neighboring cell (NB3-third neighbor cell) shows some coordinated Ca 2+ signals, which represent the ICWs (Fig. 1B, blue; Movie S1).Together, these data showed the presence of two distinct types of Ca 2+ signals: (i) intracellular Ca 2+ signals within rotavirus-infected cells that do not propagate to neighboring cells and (ii) ICWs that initiate from rotavirusinfected cells and propagate to surrounding uninfected cells.Furthermore, we noted that in most infected cells the onset of the intracellular Ca 2+ signals occurred prior to the onset of ICWs (Fig. 1B, magenta brackets).However, as shown in the inset graph in Fig. 1B, even after the onset of ICWs, the rotavirus-infected cell continues to exhibit intracellular signals that are independent of the ICWs (Fig. 1B, inset).This increase in Ca 2+ signaling results in significantly more Ca 2+ spikes in rotavirus-infected than in mock-infec ted cells (Fig. 1C).While IP 3 R is expected to play a critical role in the P2Y1-mediated ICWs, the complicated intracellular Ca 2+ signaling phenotype raised the question of the extent to which IP 3 R-mediated Ca 2+ release contributes to the intracellular Ca 2+ signals observed in rotavirus-infected cells and prompted us to examine rotavirus-induced Ca 2+ signaling in IP 3 R null cells.
To test the role of IP 3 R in dysregulation of Ca 2+ signaling during rotavirus infection, we examined Ca 2+ signaling in SA11-mRuby-infected parental HEK293-GCaMP6s and HEK293-GCaMP6s-IP 3 R-TKO cells (Fig. 1D through F).Representative images (Fig. 1D) and Ca 2+ signaling traces (Fig. 1E) show that both parental HEK293-GCaMP6s and HEK293-GCaMP6s-IP 3 R-TKO cells exhibit robust Ca 2+ signaling during rotavirus infection (red lines) that is much greater than in mock-infected cells (black lines).While there were numerous intracellular Ca 2+ signals in both cell lines, only a few, small ICWs were observed in parental HEK293-GCaMP6s cells, as illustrated in Fig. 1D and the GCaMP6s trace of a uninfected neighboring +3 cell (Fig. 1E, top blue line), and no ICWs in the HEK293-GCaMP6s-IP 3 R-TKO cells (Fig. 1E, bottom blue line).The lack of ICWs in IP 3 R-TKO cells was expected because P2Y1 is a Gq-coupled GPCR that evokes an IP 3 R-mediated release of ER Ca 2+ upon activation.Quantitation of Ca 2+ signaling in parental HEK293-GCaMP6s and HEK293-GCaMP6s-IP 3 R-TKO cells shows both cell types had a significant increase in overall spikes per cell compared to mock cells (Fig. 1F).While we observed a small, but statistically significant, increase in overall Ca 2+ signal ing in HEK293-GCaMP6s-IP 3 R-TKO cells compared to parental HEK293-GCaMP6s cells (Fig. 1F), the critical observation is that there was no decrease in rotavirus-induced Ca 2+ signaling in HEK293-GCaMP6s-IP 3 R-TKO cells.Together, these data show rotavirus infection produces robust Ca 2+ signaling in both MA104 and HEK293 cells, and this is not decreased in the absence of IP 3 R.However, we found that HEK293 cells were suboptimal for examining rotavirus-induced Ca 2+ signaling because the ICWs produced are smaller than those from MA104 cells, which potentially complicates quantitation of infected cell versus neighboring cell Ca 2+ signals.Furthermore, MA104 cells are a well-established cell line for studies on rotavirus replication and spread.Thus, for further investigation into the role of IP 3 R-mediated signaling during rotavirus infection, we aimed to establish an MA104-GCaMP6s-IP 3 R-TKO cell line.

Generation MA104-GCaMP6s cells lacking IP 3 R expression
To generate an MA104-GCaMP6s cell line lacking IP 3 R expression, we generated a CRISPR-Cas9 lentivirus construct that expresses individual gRNAs targeted to each IP 3 R gene (Table 1).Our workflow to generate MA104-GCaMP6s-IP 3 R-TKO cells is depicted in Fig. 2A.For screening and validation, we used two different GPCR agonists to activate IP 3 R-mediated Ca 2+ responses: 50 µM ADP to activate P2Y purinergic receptors and 0.75 µM AC55541 to activate the Protease-activated Receptor 2 (PAR2).We identified eight clones that exhibited little to no response to these two agonists (Fig. 2B) and chose clone F3 for full characterization, which will be designated MA104-GCaMP6s-IP 3 R-TKO cells for the remainder of this paper.
To functionally validate the loss of IP 3 R Ca 2+ signals in MA104-GCaMP6s-IP 3 R-TKO cells, we performed live imaging to measure Ca 2+ responses to ADP and PAR2 agonist AC55541 (Fig. 2C and D).For both agonists, MA104-GCaMP6s-IP 3 R-TKO cells exhibited no increase in Ca 2+ and a small, but reproducible, decrease in GCaMP6s fluorescence (Fig. 2C and D, red traces; Movie S3).To validate that these cells maintain an ER Ca 2+ store, we treated cells with thapsigargin to inhibit SERCA pumps.We found thapsigargin induced an increase in GCaMP6s fluorescence for both parental and IP 3 R-TKO cells, and peak signal from IP 3 R-TKO cells was significantly greater than that of parental MA104-GCaMP6s cells (Fig. 2E).This indicates that IP 3 R-TKO cells have a higher ER Ca 2+ load than that of parental MA104-GCaMP6s cells, which would be expected in the absence of basal IP 3 R activity.Finally, sequence analyses of MA104-GCaMP6s-IP 3 R-TKO cells identified the mutations at the gRNA sites for each IP 3 R gene.To assess the distribution of mutations, we amplified the gRNA sites for each gene, PCR-cloned amplicons, and sequenced six plasmids for each IP 3 R gene.We found a deletion mutation for IP 3 R1, three different insertion/deletion mutations for IP 3 R2, and a single allele insertion for IP 3 R3 (Fig. 2F), but no wild-type sequences were isolated, consistent with the lack of agonist-induced Ca 2+ responses from these cells.Together, these data show that we have successfully generated an MA104-GCaMP6s cell line lacking IP 3 R expression to better examine the role of IP 3 R-mediated ER Ca 2+ release during rotavirus infection.

Rotavirus-induced Ca 2+ signaling in the absence of IP 3 R
We previously found that rotavirus infection results in activation of highly dynamic Ca 2+ signaling throughout infection (21).Now by generating MA104-GCaMP6s-IP 3 R-TKO cells, we set out to determine how the lack of IP 3 R affects rotavirus-induced Ca 2+ signaling.We performed long-term, time-lapse Ca 2+ imaging of both parental MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO cells infected with SA11-mRuby rotavirus, and representative images are shown in Fig. 3A.As previously characterized, rotavirus-infec ted parental MA104-GCaMP6s cells exhibited an increase in intracellular Ca 2+ signals, as well as production of intercellular Ca 2+ waves (ICWs) (Fig. 3A, top; Movie S4).In MA104-GCaMP6s-IP 3 R-TKO cells, we still observed robust intracellular Ca 2+ signaling in rotavirus-infected cells, but no ICWs were produced (Fig. 3A, bottom; Movie S4).Next, we compared Ca 2+ signaling phenotypes between parental MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO cells, with representative traces shown in Fig. 3B.First, neither MA104-GCaMP6s nor MA104-GCaMP6s-IP 3 R-TKO mock inoculated cells exhibited any strong Ca 2+ signaling events (Fig. 3B, black lines).Next, we compared the Ca 2+ signal ing phenotype of both infected (RV) and an uninfected neighboring cell (NB3) for the MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO cells (Fig. 1B).We found that the rotavirus-infected cells for both cell lines exhibit a very similar pattern of dynamic Ca 2+ signaling that progressively increased over the course of infection (Fig. 1B, red and dark red lines).As shown in Fig. 1A, parental MA104-GCaMP6s cells generated ICWs, as indicated by the coordinated Ca 2+ signals in both the infected (RV, red line) and neighboring (NB3, blue line) cells (Fig. 1B; Movie S4).In contrast, MA104-GCaMP6s-IP 3 R-TKO cells lacked ICWs, as indicated by the absence of Ca 2+ signaling events in the neighboring uninfected cells (Fig. 1B, NB3, blue line; Movie S4).Next, we quan titated the number of Ca 2+ spikes in rotavirus-infected and uninfected neighboring cells three and five cells away (NB3 and NB5, respectively) to capture ICWs (Fig. 3C).In parental MA104-GCaMP6s cells, we found a significant increase in Ca 2+ signaling in both infected and neighboring cells, consistent with our previous studies (21).In MA104-GCaMP6s-IP 3 R-TKO cells, there was a significant increase in Ca 2+ signaling in rotavirus-infected cells but no increase in the neighboring cells, consistent with the lack of ICWs (Fig. 3C).Interestingly, in rotavirus-infected MA104-GCaMP6s-IP 3 R-TKO cells, we found no significant decrease in overall Ca 2+ signaling compared to parental MA104-GCaMP6s cells (Fig. 3C).Since MA104-GCaMP6s-IP 3 R-TKO cells did not produce ICWs, we wanted to determine whether the cells were still releasing ADP during infection.To test this, parental MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO cells were mock or rotavirus-infected and incubated in the presence of ARL67156, an ecto-nucleotidase inhibitor, to stabilize extracellular purines.At 5 hpi, a time point when ICWs have begun, media was removed from mock and infected cells and added to parental MA104-GCaMP6s cells.We found that media from rotavirus-infected cells induced a significantly greater Ca 2+ response than that from mock-infected cells, and the Ca 2+ response was blocked by the P2Y1 inhibitor BPTU (Fig. 3D).Together, this indicated that the loss of IP 3 R does not reduce the rotavirus-induced release of ADP during infection and that the IP 3 Rindependent arms of the P2Y1 signaling pathway are likely intact.These data in MA104 cells echo that of our initial studies in HEK293-GCaMP6s-IP 3 R-TKO cells, in that rotavirus still dramatically increases Ca 2+ signaling within infected cells in the absence of IP 3 R (Fig. 1E).Yet, loss of IP 3 R resulted in the abrogation of rotavirus-induced ICWs and significantly reduced the increase in Ca 2+ signaling in surrounding uninfected cells, which ultimately changes the overall Ca 2+ signaling landscape during infection.

Role of IP 3 R in rotavirus infection and replication
Numerous studies have confirmed that increased cytosolic Ca 2+ is critical for rotavi rus replication (13,14,28).Thus, we next used parental MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO cells to assess whether IP 3 R signaling plays a role in rotavirus infectivity and/or replication.First, we compared SA11-mRuby plaque phenotypes between parental MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO cells, discovering a major difference in plaque morphology (Fig. 4A).MA104-GCaMP6s cell plaques fully cleared the monolayer and had clean margins (Fig. 4A, Top) in contrast to MA104-GCaMP6s-IP 3 R-TKO cells, which exhibited a similar number of plaques (i.e., no difference in measured titer); however, the plaques were turbid with most of the monolayer remaining intact (Fig. 4A, Bottom).Quantification of plaque diameters showed rotavi rus formed significantly smaller plaques on MA104-GCaMP6s-IP 3 R-TKO cells than on parental MA104-GCaMP6s cells (Fig. 4A).Due to this substantial difference in plaque formation, we next determined whether loss of IP 3 R affected rotavirus infectivity or single-cycle virus yield.We first determined infectivity of SA11-mRuby for both parental MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO cells using a fluorescent focus assay (FFA) and found no differences in the observed titer, indicating that rotavirus infects both cell lines with the same efficiency (Fig. 4B).This is consistent with both our live imaging studies (Fig. 3) and plaque assay (Fig. 4A), in which we observed a similar number of infected cells and plaques, respectively.Next, we tested whether the MA104-GCaMP6s-IP 3 R-TKO cells efficiently supported rotavirus replication.We compared single-cycle virus yield between parental MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO cells.Cells were infected with MOI 10 and maintained in the absence of trypsin to limit the infection to a single cycle.We found that virus yields at time points from 2 to 10 HPI were similar between parental MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO cells, with no statistical differences observed (Fig. 4C).Thus, the small plaque phenotype on the MA104-GCaMP6s-IP 3 R-TKO cells cannot be explained by an inability of these cells to support rotavirus replication.We evaluated rotavirus protein expression by western blot and found similar levels and kinetics of viral protein synthesis for both structural proteins and NSP4 in both cell lines (Fig. 4D).Finally, using immunofluorescence staining of rotavirus-infected cells, we found no difference in viroplasm formation or localization of NSP4 to viroplasms (Fig. 4E), and this was confirmed by determining no significant difference in the Manders' coefficient for each cell (MA104: 0.047 ± 0.018; MA104-IP 3 R-TKO: 0.034 ± 0.013, P = 0.7).Together, these data indicate the loss of IP 3 R did not significantly affect the ability of rotavirus to infect, produce proteins, or assemble new progeny virus during the initial round of replication and, therefore, suggests the defect in plaque formation occurs during rotavirus spread through multiple rounds of replication.

Role of IP 3 R and ICWs in rotavirus spread
To further investigate why rotavirus did not plaque efficiently on MA104-GCaMP6s-IP 3 R-TKO cells, we used fluorescent microscopy to examine the spread of SA11-mRuby in the area around plaques for both parental MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO cells.We performed plaque assays with SA11-mRuby, and, on day 4 post-infection, we replaced the overlay with PBS and imaged plaques by brightfield and fluorescence microscopy to visualize mRuby expression in the plaques.In parental MA104 cells, there was full clearing of the plaque and strong mRuby expression in a wide margin around the plaque (Fig. 5A, top).In contrast, plaques on MA104-GCaMP6s-IP 3 R-TKO cells did not clear, and the area of mRubypositive cells was substantially smaller (Fig. 5A, bottom).
Next, we examined whether the smaller, turbid plaques formed on MA104-GCaMP6s-IP 3 R-TKO cells were the result of a decrease in virus replication kinetics in multi-step replication.We infected parental MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO cells with 25 PFU to ensure well-isolated infected cells and cultured in the presence of trypsin to visualize SA11-mRuby spread and measure the kinetics of virus replication over 96 h post-infection.Using live microscopy, we tracked SA11-mRuby spread throughout the monolayer in both parental MA104-GCaMP6s and MA104-GCaMP6s-IP 3 R-TKO cells (Fig. 5B).In parental MA104-GCaMP6s cells, there was more rapid spread with complete destruction of the monolayer by 48-72 hpi (Fig. 5B, top).In contrast, virus spread in MA104-GCaMP6s-IP 3 R-TKO cells was substantially delayed, especially between 24 and 48 hpi (Fig. 5B, bottom).Next, we determined virus yield from 24 to 96 hpi and found parental MA104-GCaMP6s cells supported faster virus replication than MA104-GCaMP6s-IP 3 R-TKO cells, with peak titers reached in parental cells by 48 hpi (Fig. 5C, blue line).In contrast, virus replication was significantly lower in MA104-GCaMP6s-IP 3 R-TKO cells from 48 to 72 hpi and did not peak until 72-96 hpi (Fig. 5C, red line).
Furthermore, we examined rotavirus spread using live, time-lapse microscopy to measure the rate of mRuby expression in the absence or presence of trypsin (Fig. 5D).In the absence of trypsin, there were single infected cells and no spread of mRuby to neighboring cells (data not shown).In the presence of trypsin, rotavirus spread rapidly from the initial infected cells to many of the surrounding cells in parental MA104-GCaMP6s cells, but in MA104-GCaMP6s-IP 3 R-TKO cells, this was much slower and was restricted primarily to adjacent cells (Fig. 5D; Movie S5).Together, these data show that in infected cells, the host IP 3 R ER Ca 2+ channel was largely dispensable for rotavirus-induced Ca 2+ signaling; however, IP 3 R-mediated Ca 2+ signaling was critical for rotavirus spread.
Lastly, we wanted to test if loss of IP 3 R contributed to a defect in virus spread of a human rotavirus strain.Similar to our results with SA11-mRuby, the human rotavirus strain Ito formed plaques that fully cleared the MA104-GCaMP6s monolayer and had clean margins (Fig. 5E, left).In contrast, while Ito-infected MA104-GCaMP6s-IP 3 R-TKO cells exhibited a similar number of plaques (i.e., no difference in measured titer), the plaques were substantially smaller with most of the monolayer remaining intact (Fig. 5E, right).Next, we determined virus yield in a multi-round replication assay and found parental MA104-GCaMP6s cells supported faster virus replication than MA104-GCaMP6s-IP 3 R-TKO cells, with peak titers in parental cells reached by 48 hpi (Fig. 5E, orange line).In contrast, virus replication was significantly slower in MA104-GCaMP6s-IP 3 R-TKO cells, and peak titers were not reached until 96 hpi (Fig. 5E, teal line).
(Continued on next page) All experiments were performed with a minimum of three biological repeats of at least three technical

Priming uninfected cells via P2Y1-mediated ICWs
As the loss of IP 3 R did not affect rotavirus infectivity or replication in the single-cycle replication studies, the above data indicated the defect in rotavirus spread was related to the lack of increased Ca 2+ signaling in neighboring cells.Increased Ca 2+ signaling in rotavirus-infected cells has been shown to be critical for robust replication (13,29,30), so these observations led us to hypothesize that the increased Ca 2+ signaling in neighbor ing uninfected cells could promote rotavirus replication/spread by priming them for infection.Thus, we set out to test this hypothesis by examining how the manipulation of ICWs affected rotavirus spread.
We previously showed P2Y1 receptor blockers significantly inhibit rotavirus-induced ICWs (18).Thus, we next tested whether blocking ICWs with BPTU, a P2Y1 selective blocker, would also reduce rotavirus spread.We performed a plaque assay using parental MA104 cells in the presence of 1 µM BPTU, or DMSO vehicle control, and found BPTU treatment resulted in significantly smaller plaques (Fig. 6).Thus, reducing ICWs by blocking P2Y1 causes a similar reduction in the ability of rotavirus to form plaques as the loss of IP 3 R; however, P2Y1 blockers, or even P2Y1 knockout cells, did not fully abrogate ICWs as strongly as MA104-GCaMP6s-IP 3 R-TKO cells (18).Therefore, to further test the role of ICWs in rotavirus replication, we examined rotavirus spread in LLC-MK2-GCaMP6s cells, a rhesus monkey kidney cell line we used previously to study Ca 2+ signaling by Recoviruses (Rhesus enteric caliciviruses), for which we used rotavirus as a positive control virus for increased Ca 2+ signaling (23).We observed that while rotavirus-infected LLC-MK2-GCaMP6s cells exhibited increased Ca 2+ signaling, there were no ICWs and, therefore, no increase in Ca 2+ signaling in neighboring uninfected cells (Fig. 7A).Furthermore, treatment of LLC-MK2-GCaMP6s cells with ADP did not evoke a Ca 2+ response (Fig. 7B through D), indicating a lack of functional P2Y1 signaling.So, we reasoned that if ICWs are a critical factor in rotavirus spread, exogenous expression of P2Y1 in LLC-MK2-GCaMP6s cells might rescue the rotavirus-induced ICW phenotype and increase rotavirus spread.We used lentivirus transduction to generate a stable P2Y1 knock-in LLC-MK2-GCaMP6s cell line (LLC-MK2-GCaMP6s + P2Y1), and the presence of an ADP-stimulated Ca 2+ signal confirmed P2Y1 was functional in these cells (Fig. 7C and D; Movie S6).We confirmed a significant increase in P2Y1 expression in the LLC-MK2-GCaMP6s + P2Y1 cells by qRT-PCR and found that the P2Y1 knock-in did not affect the expression of any other P2Y purinergic receptor (Fig. 7E).
We next investigated whether exogenous expression of P2Y1 was able to rescue rotavirus-induced ICWs and if this affected rotavirus spread.We performed live imaging of rotavirus-infected LLC-MK2-GCaMP6s + P2Y1 cells and found these cells exhibited both higher basal Ca 2+ signaling and ICWs that originate from rotavirus-infected cells (Fig. 8A; Movie S7), resulting in greater Ca 2+ signaling in the neighboring, uninfected cells than mock-inoculated cells (Fig. 8B).Next, we used plaque assays to assess rotavirus spread in LLC-MK2-GCaMP6s cells.In parental LLC-MK2-GCaMP6s cells, rotavirus formed few plaques, and those visible were turbid (Fig. 8C, top); however, LLC-MK2-GCaMP6s + P2Y1 cells showed more plaques with greater clearing (Fig. 8C, bottom).We used fluorescent microscopy to examine spread of SA11-mRuby in the area around plaques for both cell lines.We saw strong mRuby expression for both cell types, but LLC-MK2-GCaMP6s + P2Y1 had significantly more infected cells surrounding the plaques regions with cell monolayer clearing (Fig. 8D and E).Thus, introducing P2Y1 expression in LLC-MK2-GCaMP6s cells provides a gain of function to produce ICWs in response to rotavirus infection, and this results in better plaque formation and more efficient virus spread.Together these data support our model that the Ca 2+ signals from P2Y1-mediated ICWs, which require ER Ca 2+ release by IP 3 R, primes neighboring cells to promote more robust rotavirus replication and spread in vitro.

DISCUSSION
A hallmark of rotavirus infection, and several other viruses, is an elevation in cytosolic Ca 2+ and decrease in ER Ca 2+ stores, which facilitates virus replication and contributes to pathogenesis through a variety of downstream pathways.Host cells also rely on Ca 2+ signaling pathways to maintain homeostasis.The IP 3 R Ca 2+ channel is an important Ca 2+ signaling relay hub converting extracellular signals (e.g., ADP) into intracellular signals in which Ca 2+ itself is the second messenger that controls a myriad of cellular pathways via Ca 2+ -regulated proteins (31).While the NSP4 viroporin initiates the dysregulation of Ca 2+ homeostasis during rotavirus infection, the role of IP 3 R in generating rotavirusinduced Ca 2+ signals, particularly in virus-infected cells, had not been characterized.By developing a MA104-GCaMP6s-IP 3 R-TKO cell line, we uncovered the following: (i) IP 3 R is not required for elevated Ca 2+ signaling observed in rotavirus-infected cells; (ii) IP 3 R ER Ca 2+ release is critical for P2Y1-mediated ICWs that increase Ca 2+ signaling in neighbor ing, uninfected cells; and (iii) while IP 3 R Ca 2+ signaling is not necessary for rotavirus replication, the rotavirus-induced ICWs, which require IP 3 R, increase the kinetics of rotavirus replication and spread in vitro.Together, these findings indicate that increased Ca 2+ signals in neighboring cells, caused by ICWs, primes these cells to better support rotavirus replication.This process of "pre-emptively priming" uninfected cells within a viral niche represents a novel mechanistic paradigm by which viruses exploit intercellular host responses to promote their replication.
We recently showed that rotavirus-induced increases in cytosolic Ca 2+ occur through a massive increase in discrete Ca 2+ signaling events, and a substantial part of this signaling comes from the release of ER Ca 2+ (21).By examining IP 3 R-null cells (both HEK293 and MA104), we have been able to differentiate two distinct types of Ca 2+ signals that are induced during rotavirus infection: (i) the intracellular Ca 2+ signals that occur within rotavirus-infected cells, which were IP 3 R-independent, and (ii) the multicellular ICWs that propagate from infected to neighboring, uninfected cells, which are IP 3 R-dependent.By dissecting out these distinct Ca 2+ signals, we can gain new mechanistic insights into how these signals support rotavirus replication and spread.
The intracellular Ca 2+ signals are IP 3 R-independent as knockout of IP 3 R did not reduce the number of Ca 2+ signals observed.Since NSP4 alone is sufficient to increase Ca 2+ signaling by release of ER Ca 2+ , we propose that the Ca 2+ signals observed in the IP 3 R-TKO cells are ER Ca 2+ release events by the NSP4 viroporin (28,32); however, some of these Ca 2+ signals could be Ca 2+ entry via SOCE channels (e.g., Orai) activated by NSP4-medi ated decrease in ER Ca 2+ levels (29).The elevation in Ca 2+ signaling, in the absence of IP 3 R, was sufficient to support rotavirus replication though the kinetics of multi-step virus replication was impaired (discussed below).The fact that rotavirus encodes an intrinsic capability to drive this degree of Ca 2+ dysregulation further highlights the critical role Ca 2+ signaling by NSP4, and perhaps other viroporins play in virus replication (3,10,13).This raises the question of what cellular pathways are activated by these intracellular Ca 2+ signals.NSP4 viroporin Ca 2+ signals are critical for rotavirus replication by initiating the cellular autophagy pathway via activation of the Ca 2+ -dependent kinase CaMKKβ (13).Furthermore, these dynamic ER Ca 2+ signals increase mitochondrial metabolism and suppress the activation of apoptosis, which would also benefit rotavirus replication (21).Yet, unraveling the complexity of virus-induced Ca 2+ signaling remains a burgeoning field, so we are certain to discover other Ca 2+ -regulated pathways involved in virus replication.
While IP 3 R Ca 2+ signaling was dispensable for rotavirus replication, MA104-GCaMP6s-IP 3 R-TKO cells revealed that rotavirus-induced ICWs significantly increased rotavirus spread and the rate of virus replication.The importance of this paracrine ADP/P2Y1generated ICW signaling pathway was further supported by rescue of rotavirus plaque formation and increased spread in LLC-MK2-GCaMP6s cells when rotavirus ICW generation was reconstituted by exogenous P2Y1.Furthermore, because MA104-GCaMP6s-IP 3 R-TKO cells still released ADP, we can infer that the loss of IP 3 R Ca 2+ signaling, and not other downstream effects of P2Y1 activation, causes the defect in rotavirus spread.Finally, although rotavirus replication in MA104-GCaMP6s-IP 3 R-TKO cells was slower, the final yields for both SA11-mRuby and Ito were similar by 96 hpi to those from parental MA104-GCaMP6s cells, which indicates that IP 3 R Ca 2+ signaling increased the rate of virus replication primarily via the increase in Ca 2+ signaling in neighboring, uninfected cells as a result of the P2Y1-mediated ICWs.
Based on these data, we propose the rotavirus-induced ICWs activate pro-viral Ca 2+ -regulated pathways in neighboring cells, ultimately priming these cells for more rapid and/or increased replication.This raises the question of which Ca 2+ -regulated pathways are activated by ICWs in neighboring cells and how they increase rotavi rus replication.One likely pathway is the activation of autophagy because ICWs, like NSP4-mediated Ca 2+ signals, are generated by the release of ER Ca 2+ and IP 3 R Ca 2+ release can activate CaMKKβ and AMPK phosphorylation and upregulate autophagy (33)(34)(35).In rotavirus-infected cells, early autophagosome membranes are usurped to traffic NSP4 and VP7 out of the ER and form a membrane compartment associated with viroplasms (i.e., virus replication complexes) which are the site of final rotavirus assembly (13).Thus, ICWs could activate the early, biosynthetic stages of the autophagy pathway, making more of these membranes available to be utilized upon rotavirus infection.
Rotavirus-induced Ca 2+ signaling is also associated with a broad dysregulation in the actin cytoskeleton, though this has primarily been studied in infected or NSP4expressing cells and the potential role for ICWs has not been investigated (36).Neverthe less, ICWs are an important mode of multicellular epithelial signaling that directs cell extrusion from monolayers in response to cell damage (37)(38)(39).The resulting IP 3 R-medi ated ER Ca 2+ release triggers a global, but transient, actin reorganization process termed Ca 2+ -mediated Actin Reset (CaAR) (40).The CaAR response can drive significant changes in host gene transcription by releasing transcription factors otherwise sequestered in the cytosol (40), and this may help identify cellular pathways activated by ICWs.Importantly, the CaAR responses studied thus far have come from singular signals/damage events, but during rotavirus infection, neighboring cells are stimulated by hours of ICWs, providing ample time and numbers of signals to drive substantial changes.Thus, it will likely require a detailed, multi-omics approach to identify ICW-response pathways and elucidate which of those support rotavirus spread.
In summary, we have uncovered a dichotomous role for IP 3 R in the overall Ca 2+ signaling landscape during rotavirus infection and in rotavirus replication and spread.Within infected cells, rotavirus (presumedly via NSP4) generates sufficient Ca 2+ signaling to support its replication without IP 3 R, making this host channel dispensable.In contrast, rotavirus spread was increased by the presence of ICWs, indicating that the increased Ca 2+ signaling in neighboring cells essentially primes them for future rotavirus infection.This implies that viral take-over of the infection niche goes beyond rotavirus-infected cells to include nearby uninfected cells, which undergo "preemptive reprogramming" by repeated ICWs prior to becoming infected.Furthermore, ICWs triggered by rotavirus infection may be a common host response to many different virus infections, and, if so, this P2Y1-mediated Ca 2+ signaling pathway would have a broader importance in virus replication and pathogenesis.Finally, purinergic signaling is just one of the myriad of other intercellular signaling molecules/pathways that could be exploited by viruses to prime or otherwise reprogram uninfected cells within the infection niche.Identification of analogous virus-induced intercellular signaling pathways may uncover new mechanisms by which viruses, or other microbes, exploit host responses to benefit their replication and spread.

FIG 5 (
FIG 5 (Continued) replicates, except for panel E, which used two technical replicates.All data are shown as the mean ± SD.Scale bars are 50 µm in B and D. **P < 0.01 by t-test.

FIG 6
FIG 6 Blocking P2Y1-mediated ICWs reduces rotavirus plaque size.(A) Representative image of RV plaques formed on parental MA104 cell monolayers treated with DMSO (Top) or 1 µM BPTU (bottom).(B) Measurement of plaque diameter for DMSO and 1 µM BPTU treated cells.Experiments were performed with three biological repeats of two technical replicates each.Data are shown as the mean ± SD. ***P < 0.001 by Mann-Whitney t test.

TABLE 3
Quantitative PCR primer sequences