Bacterial Quorum-Sensing Signal Arrests Phytoplankton Cell Division and Impacts Virus-Induced Mortality

Bacteria and phytoplankton form close associations in the ocean that are driven by the exchange of chemical compounds. The bacterial signal 2-heptyl-4-quinolone (HHQ) slows phytoplankton growth; however, the mechanism responsible remains unknown.

are complex, often being species specific (4) or temporally ephemeral (5), and can span the spectrum from antagonistic to beneficial (6,7). Increasingly, it is clear that these intricate interkingdom interactions are facilitated by excreted chemical compounds that mediate a suite of processes such as nutrient transfer, primary production, and shifts in community composition. Linking chemical compound identity with a mechanism of action and ecological consequences will strengthen our understanding of how these fundamental and multifaceted interactions govern marine ecosystem function.
First discovered in marine systems 4 decades ago (8), quorum sensing (QS) is a form of microbial cell-cell communication through which marine bacteria use diffusible chemical signals to facilitate coordinated and cooperative biogeochemically important behaviors (9). Recent work finds that alkylquinolone-based QS signals can modulate interspecies behavior, suggesting that these molecules may influence cellular communication at the interkingdom level (10). In particular, the alkylquinolone QS signal 2-heptyl-4-quinolone (HHQ) functions as a messenger molecule able to modulate bacterial virulence behavior, facilitating the emergence of the pathogen Pseudomonas aeruginosa within polymicrobial communities (11,12). Trafficking of hydrophobic alkylquinolones, including HHQ, is aided by the release of outer membrane vesicles containing micromolar concentrations of alkylquinolones that are produced by P. aeruginosa and serve as signal delivery vehicles to neighboring recipient cells (13). Purified outer membrane vesicles isolated from P. aeruginosa have also been shown to possess significant antimicrobial activity, inhibiting the growth of adjacent Gram-positive bacteria (13). Additionally, HHQ has also been implicated in antagonizing fungal biofilm formation (12); downregulating eukaryotic host immune responses via the suppression of a key transcription factor, NF-κB (10); and activating receptors found to play a role in innate immune signaling in airway epithelia (14). These findings support the influence of alkylquinolones in mediating hostmicrobe interactions.
Recently, HHQ was isolated from marine gammaproteobacteria (Pseudomonas sp. and Pseudoalteromonas sp.) and was observed to cause significant shifts in both natural phytoplankton and microbial communities (15), including species-specific static phytoplankton growth (no growth or mortality) at nanomolar concentrations (16). Static growth in phytoplankton has been observed previously, in relation to both bacterial exudates (17,18) and nutrient stress (19)(20)(21)(22). However, the underlying molecular mechanism(s) by which HHQ influences phytoplankton fitness and the outcomes of ecological interactions remains unknown. For example, host physiology has been demonstrated to be an integral factor in the success of viral infection of phytoplankton, with infection success and burst size being influenced by host conditions (23). However, the role that HHQ plays in mediating microbial interactions beyond phytoplankton growth alterations has yet to be investigated.
To better understand how HHQ alters molecular function and ecological interactions in marine microbes, ultrastructural observations and diagnostic biochemical assays were integrated with transcriptomic and proteomic studies to link the persistent but reversible physiological impact of nanomolar concentrations of HHQ on a model marine phytoplankton, Emiliania huxleyi. Furthermore, we examined if HHQ could disrupt virus-induced mortality in E. huxleyi, thereby ascribing a new role for bacterial quorum-sensing signals. E. huxleyi plays a central role in mediating ocean carbon (24) and sulfur (25) cycling; thus, the results presented here emphasize the importance of considering the ecological consequences of chemically mediated bacterium-phytoplankton interactions on global primary production and biogeochemical cycles.

RESULTS AND DISCUSSION
Response to HHQ exposure. Following 96 h of exposure to 100 ng ml 21 of HHQ, batch cultures of axenic E. huxleyi (CCMP2090) exhibited cellular stasis (no cell division or mortality) concomitant with a significant increase in forward scatter, red fluorescence, and side scatter, proxies for cell size, chlorophyll content, and cell granularity, respectively (P value of ,0.01 for all comparisons by repeated-measures analysis of variance [ANOVAR]) (Fig. 1). The photosynthetic efficiency (F v /F m ) did not change in response to long-term HHQ exposure (ANOVAR). Additionally, after only 24 h of HHQ exposure, phytoplankton cells were observed to have enlarged chloroplasts with distended thylakoid membranes containing numerous intraorganelle vesicles, abundant cytoplasmic vesicles/vacuoles, homogeneous nucleus staining lacking defined euchromatin/heterochromatin regions with disintegrated nuclear envelops, and osmium-rich puncta within and adjacent to the chloroplasts, likely indicating enhanced lipid storage (see Fig. S1 in the supplemental material). To examine if the physiological effects induced by HHQ exposure were reversible, 96-h-exposed HHQ cultures were diluted roughly ;80-fold with f/2 medium without silica (26) to a final concentration of 1.25 ng ml 21 HHQ, a concentration demonstrated not to influence E. huxleyi growth. Cells previously exposed to HHQ showed recovery mirroring paired vehicle control cultures (ANOVAR) (Fig. S2). Taken together, HHQ-treated E. huxleyi cells appear to mirror previous studies in which cellular arrest has been observed in phytoplankton in response to bacterially derived chemical exposure (17,18,(27)(28)(29) as well as nutrient limitation (20)(21)(22). In order to elucidate if the observed cellular stasis is mechanistically similar to those observed previously in the literature, we conducted cell cycle, transcriptomic, and proteomic analyses of HHQ-exposed E. huxleyi.
Evidence for S-phase arrest. The DNA content of E. huxleyi cells following HHQ exposure was tracked for 96 h via flow cytometry, and cells treated with HHQ ceased the typical diurnal cell cycle progression within 24 h of HHQ addition (Fig. 2). Over 96 h, the proportions of HHQ-exposed cells in both G 1 and G 2 phases were shown to steadily decrease, whereas the proportion of cells in S phase significantly increased (P . 0.01 by ANOVAR) (Fig. 2c through e). Additionally, HHQ-exposed cells found in G 1 phase FIG 1 Exposure to HHQ halts cell division and alters cell morphology. (a through d) E. huxleyi cultures (n = 3) were exposed to HHQ or the vehicle control (DMSO) at the 0-h time point (T 0 ) and monitored by flow cytometry for cell abundance (a), red fluorescence in relative fluorescence units (RFU) (695/50 nm) (a proxy for chlorophyll a intensity) (b), forward scatter (a proxy for cell size) (c), and side scatter (a proxy for cell granularity) (d) over 96 h. Means 6 standard deviations are shown. In 100-ng ml 21 HHQ-exposed cells, all parameters measured were significantly different from those for the vehicle control (P , 0.05 by repeated-measures analysis of variance). Note that in panel a, data for HHQ-treated cells at 1 ng ml 21 sit directly beneath data for the vehicle control (DMSO). (e and f) Transmission electron microscopy micrographs of E. huxleyi cells exposed to the vehicle control (DMSO) (e) or 100 ng ml 21 HHQ (f) for 24 h. Subcellular structures include the chloroplast (c), lipid droplet (l), mitochondria (m), nucleus (n), and vacuoles (black arrowheads). demonstrated a significantly higher DNA content per cell than their paired vehicle controls (P value of ,0.05 by Welch's approximate t test) (Fig. 2f). These results suggest that HHQ-exposed cells are attempting to duplicate their genome but are stalling in early S phase.
Whole-cell transcriptomic and proteomic analyses were performed on E. huxleyi cells exposed to HHQ concentrations of 1 ng ml 21 (low), 10 ng ml 21 (medium), and 100 ng ml 21 (high), with samples taken at 24 h (transcripts) and 72 h (transcripts and proteins) (Fig. S3). E. huxleyi cultures demonstrated unique transcriptomic and proteomic profiles in response to each HHQ concentration, with the greatest numbers of differentially expressed genes being found in higher-HHQ treatments, compared to the dimethyl sulfoxide (DMSO) vehicle control (Table 1; Fig. S3a and b). After 24 h of HHQ exposure, 39.8% of transcripts in high-HHQ samples were differentially expressed relative to the DMSO vehicle controls (q value of ,0.05 by a Wald test) (Table 1). Similarly, after 72 h of exposure, replicate high-HHQ samples continued to appear distinct from the DMSO vehicle control samples ( Fig. S3a and b), with 37.6% of transcripts (q value of ,0.05 by a Wald test) and 15.9% of proteins (q value of ,0.05 by Welch's approximate t test) significantly changing in relative abundance and abundance, respectively (Table 1). When examined together, a total of 665 genes and corresponding proteins were found to significantly change in abundance at 72 h under high-HHQ treatment relative to the vehicle control ( Fig. 3) (see Supplemental Data File 1 at https://doi.org/10.6084/m9.figshare.14414285.v1). In general, processes associated with DNA replication and repair, aerobic respiration, and protein catabolism yielded higher relative transcript and protein abundances under high-HHQ treatment, while  21 HHQ for 96 h and stained with propidium iodide were compared via Welch's approximate t test (P , 0.01). As DNA replication occurs only in S phase, the increase in the mean fluorescence for HHQ-treated cells that fall within the G 1 gate suggests that these cells are currently in S phase but stall early in the process of DNA synthesis and are unable to synthesize enough additional DNA to fall within the Sphase region. photosynthetic components/processes were detected at lower relative transcript and protein abundances ( Fig. 3a) (see Supplemental Data File 1 at https://doi.org/10 .6084/m9.figshare.14414285.v1). Far fewer genes and proteins were found to be differentially expressed in the low-and medium-HHQ treatments (Table 1), which is likely related to the observed recovery of cell growth in these treatments (Fig. S3c). The growth of cells exposed to low HHQ concentrations was nearly identical to that of the DMSO control throughout the experiment. By 72 h, no genes or proteins in the low-HHQ treatment were differentially expressed compared to the control. While the medium-HHQ treatment demonstrated some growth inhibition over the first 24 h, by 72 h, the population had largely recovered (Fig. S3c). For cell populations exposed to medium HHQ concentrations at 24 h, transcripts related to cell cycle progression, cytoskeletal regulation, and mitosis demonstrated increased relative abundances compared to the control (see Supplemental Data File 1 at https:// doi.org/10.6084/m9.figshare.14414285.v1). By 72 h, there was no clear trend observed in functions related to differentially expressed transcripts in the medium-HHQ samples. For the purposes of elucidating the molecular target(s) of HHQ, we therefore focused our efforts on the analysis of the high-HHQ transcriptomic and proteomic data. Indeed, at the physiological level, the response of E. huxleyi to HHQ parallels phosphorus (P) limitation in phytoplankton (i.e., S/G 2 -phase arrest, decreased growth rate, and increased chlorophyll content, forward scatter, and side scatter) (19)(20)(21)30). However, the canonical response in P-limited cells of the upregulation of both alkaline phosphatase and phosphodiesterases (31)(32)(33) was not observed in cells exposed to HHQ, nor did we see a significant induction of acid phosphatases, pyrophosphatase, phosphorus transporters, or ATP-sulfurylase enzymes known to be induced following P limitation in HHQexposed cells, indicating the lack of phosphorus stress (see Supplemental Data File 1 at https://doi.org/10.6084/m9.figshare.14414285.v1). Therefore, while the patterns of cell cycle arrest are similar between HHQ-treated E. huxleyi and nutrient limitation, the underlying mechanisms are distinct.
In phytoplankton, cellular arrest is often accompanied by the induction of autocatalytic or programmed cell death (PCD) responses such as increased reactive oxygen production or caspase-like activity (34), and previous findings in mammalian cells indicate that HHQ has the ability to activate PCD pathways (35). However, no evidence of PCD/ apoptosis was observed in HHQ-exposed E. huxleyi cells using a series of diagnostic fluorescence assays (i.e., membrane permeabilization, caspase activity, and reactive oxygen species [ROS] and nitrous oxide [NO] production) (Fig. S4). Additionally, no transcripts or proteins associated with PCD increased in abundance with exposure to HHQ (see Supplemental Data File 1 at https://doi.org/10.6084/m9.figshare.14414285.v1). The lack of PCD induction in HHQ-exposed cells may stem from E. huxleyi's arrest in early S phase (Fig. 2d), as cellular arrest during S phase does not induce apoptotic pathways but rather curtails DNA replication, thereby dramatically extending the cell cycle (36). The transcriptomic profile of HHQ-exposed cells demonstrates an increased relative abundance of canonical transcripts facilitating the G 1 /S transition, including cell division control protein 6 (CDC6), origin recognition complex subunit 1 (ORC1), and cyclins A, B, E, and K (Fig. 3b). Moreover, significant increases in relative transcript abundances of DNA replication fork machinery (i.e., DNA polymerases a, « , and d ; DNA primase; replication protein A; topoisomerases [TOPO]; the minichromosomal maintenance complex; proliferating cell nuclear antigen; and replication factor C) (Fig. 3c) 72 h after HHQ exposure suggest an intent to replicate DNA, a hallmark of S phase (37). Yet despite this observed induction of DNA replication machinery, DNA synthesis was severely curtailed following HHQ exposure (Fig. 2), suggesting that HHQ exposure interferes with the ability of E. huxleyi cells to correctly complete the DNA replication process. Disruption of DNA replication induces DNA damage response (DDR) pathways, activating effector kinases such as Chk1 and Chk2 necessary for halting DNA synthesis and the induction of cell cycle arrest to allow time for repair (38). We observed transcripts for Chk1 and Chk2 to be differentially expressed under HHQ treatment (Fig. 3b). Furthermore, a significant decrease in relative histone transcript and protein abundances, a hallmark of DNA synthesis disruption, was observed (see Supplemental Data File 1 at https://doi.org/10.6084/m9.figshare.14414285.v1) following HHQ exposure. As DNA replication and histone production are coupled, cells experiencing DNA replication stress will remove histone transcripts (39).
Possible protein targets of HHQ. During S phase, a cell must tightly regulate the availability of nucleotides to ensure faithful DNA replication (40). Therefore, S-phase cells rely on de novo nucleotide synthesis pathways to produce enough materials for complete genome replication (41). Several transcripts and proteins involved in de novo purine (amidophosphoribosyltransferase, trifunctional purine biosynthetic protein adenosine 3, phosphoribosylformylglycinamidine synthase, bifunctional purine biosynthesis protein, adenylosuccinate synthase, IMP dehydrogenase, and GMP synthase) and pyrimidine (carbamoyl phosphate synthase II, aspartate carbamoyltransferase, and CTP synthases) nucleotide synthesis increased in abundance with HHQ exposure (see Supplemental Data File 1 at https://doi.org/10.6084/m9.figshare.14414285.v1). Increased nucleotide synthesis may indicate the need to produce the necessary materials to replenish nucleotide pools during replication. However, only partial replication of the E. huxleyi genome following HHQ exposure was observed (Fig. 2), suggesting that HHQ may disrupt nucleotide production, thereby limiting nucleotide availability.
Select alkylquinolones are known to inhibit a key rate-limiting enzyme directly involved in pyrimidine synthesis, dihydroorotate dehydrogenase (DHODH) (42). DHODH inhibition in eukaryotes may induce intra-S-phase arrest due to severely diminished cellular nucleotide pools that can disrupt DNA replication, stall replication forks, and increase the frequency of genomic DNA lesions, including strand breaks, during S phase (43,44). Indeed, after 46 h of HHQ exposure, a significant increase in DNA strand breaks was observed in culture (P = 0.032 by Welch's approximate t test) (Fig. 4a) and was not observed when HHQ was directly exposed to genomic E. huxleyi or lambda DNA (Fig. S5). This indicates that DNA strand breaks are not caused directly by HHQ but are caused indirectly through other mechanisms. It has been previously observed that following the induction of DNA damage during S phase, cells will enter intra-S-phase arrest that drastically decreases the rate of DNA replication to allow the DDR to resolve any DNA lesions (36). With the exception of preliminary work in Chlamydomonas reinhardtii and dinoflagellates, the DDR has not been well characterized in phytoplankton (45,46). Of the 57 mammalian DDR protein homologs in the E. huxleyi genome (E value of #10 220 ), 41 were significantly differentially expressed (at the transcript and/or protein level), of which 37 increased in relative abundance at 72 h under high-HHQ exposure (Fig. 3c), indicating that the cell is attempting to repair DNA lesions. However, DNA damage induced by the inhibition of DHODH is known to activate apoptotic pathways through the hyperactivation of the DDR (47). No apoptotic pathway activation was observed with HHQ exposure, suggesting that the DDR itself may also be impacted by HHQ.
A master regulator of the DDR involved in chromatin remodeling, nucleolar structure, and genome stability is poly(ADP-ribose) polymerase (PARP) (48). PARP binds to sites of DNA damage and stalls replication forks, producing negatively charged ADPribose polymer scaffolds that attract repair proteins (49). PARP homologs in E. huxleyi were found to be increased in both relative transcript abundance and protein abundance under HHQ treatment (Fig. 3c). Under high levels of DNA damage or if repair mechanisms are compromised, PARP can become overactivated and deplete cellular NAD 1 and ATP pools, thereby initiating apoptotic pathways (50). However, no apoptotic activity was observed in E. huxleyi cells following HHQ exposure (Fig. S4), indicating that HHQ may inhibit PARP activity. Indeed, HHQ was found to significantly inhibit human PARP activity (P = 0.0002 by Welch's approximate t test) (Fig. 4b), while a closely related alkylquinolone, 2-heptyl-3-hydroxy-4(1H)-quinolone (PQS), did not possess PARP-inhibitory activity, nor did it impact E. huxleyi growth (Fig. 4b and Fig. S6a). Interestingly, the genomes of phytoplankton species unaffected by HHQ (16) did not reveal the presence of any PARP homologs, further implicating PARPs in the response of phytoplankton to HHQ.
Inhibition of PARP activity in the presence of DNA damage drastically reduces the effectiveness of the DDR and is known to induce cellular arrest in the S phase (51). Together, our observations of prolonged S-phase arrest (Fig. 2), the upregulation of the DDR in HHQ-exposed cultures (Fig. 3c), the conserved nature of the mammalian and E. huxleyi PARP catalytic sites (Fig. S6b through d), and the chemical structural similarities of HHQ to known inhibitors of both PARP and DHODH with core benzimidazole moieties (52) collectively suggest that HHQ may function simultaneously to inhibit both PARP and DHODH activity in E. huxleyi. Additional experiments using E. huxleyi enzymes are needed to fully characterize whether PARP and DHODH are molecular targets of HHQ.
HHQ impacts on energy production. To facilitate DNA synthesis and repair, the cell requires large ATP pools (53). In HHQ-exposed cells, the increased relative transcript abundance of enzymes in the tricarboxylic acid (TCA) cycle (i.e., isocitrate dehydrogenase, a-ketoglutarate dehydrogenase, succinate dehydrogenase, fumarase, and malate dehydrogenase) (see Supplemental Data File 1 at https://doi.org/10.6084/m9 .figshare.14414285.v1) may signal the overproduction of reducing equivalents for ATP production via oxidative phosphorylation. Additionally, the increase in the relative transcript abundance of metabolic efficiency controllers, sirtuin-like deacetylases (54), observed following HHQ treatment (see Supplemental Data File 1 at https://doi.org/10 .6084/m9.figshare.14414285.v1) may be a direct result of PARP inhibition. Sirtuins compete with PARPs for NAD 1 , and the expression of deacetylases is dependent on NAD 1 availability (55). PARP inhibition is known to drastically increase cellular NAD 1 pools, thereby promoting sirtuin expression and activity (56). Increased sirtuin activity in HHQ-exposed cells may also explain the increase in the relative transcript abundance of manganese superoxide dismutase (Mn-SOD) (see Supplemental Data File 1 at https://doi.org/10.6084/m9.figshare.14414285.v1), an antioxidant enzyme that protects the cell from ROS-induced damage, as sirtuins are known to induce the production of FIG 4 Exposure to HHQ leads to cellular DNA damage and inhibition of human PARP. (a) Cultures (n = 4) of E. huxleyi were exposed to 100 ng ml 21 HHQ or the vehicle control (DMSO) for 46 h before pigments were removed and cells were stained using an in vivo TUNEL assay to detect the presence of DNA ends, a proxy for DNA breaks. (b) Inhibition of the human PARP-1 enzyme by 50 mM HHQ and 2-heptyl-3-hydroxy-4(1H)-quinolone (PQS). Percent PARP inhibition was measured using the PARP universal colorimetric assay kit (R&D Systems). The absorbance values for quadruplicate wells containing HHQ or PQS were compared to those of the vehicle control, and this ratio was subtracted from 100% to determine PARP inhibition. Points represent individual replicates. Asterisks indicate a significant difference between the treatment and the vehicle control (P value of ,0.05 by Welch's approximate t test).
Mn-SOD proteins (57). Finally, increased relative transcript abundance of the tryptophan-mediated de novo NAD 1 synthesis pathway was also observed, potentially in an attempt to increase NAD 1 availability (see Supplemental Data File 1 at https://doi.org/ 10.6084/m9.figshare.14414285.v1). Taken together, these results suggest that HHQ exposure promotes increased energy production in E. huxleyi, which can fuel various cellular biosynthesis and repair pathways while staving off the induction of PCD.
Increased cellular demand for ATP would necessitate the induction of glycolytic enzymes. However, following HHQ treatment, there was a significant decrease in the relative transcript abundance of hexokinase (see Supplemental Data File 1 at https:// doi.org/10.6084/m9.figshare.14414285.v1), the first step in glycolysis, consistent with previous work demonstrating that alkylquinolones suppress the induction of this glycolytic enzyme through direct targeting of transcription factor hypoxia-inducible factor 1 (HIF-1) protein degradation via proteasomal pathways (58). Furthermore, we observed a shift to the Entner-Doudoroff glycolytic pathway in HHQ-treated cells (see Supplemental Data File 1 at https://doi.org/10.6084/m9.figshare.14414285.v1), which can conserve amino acid resources due to a low protein demand in comparison to other pathways (59). Moreover, we observed increases in relative transcript abundances leading to the production of aspartate (i.e., the TCA cycle, the aspartate-arginosuccinate shunt, glutamic oxaloacetic transaminase [GOT], and C 4 -like photosynthesis) in parallel with decreases in transcripts for aspartate utilization pathways, with the exception of nucleotide synthesis (see Supplemental Data File 1 at https://doi.org/10.6084/ m9.figshare.14414285.v1). Aspartate is known to rescue cells from S-phase arrest by fueling de novo nucleotide synthesis (60).
HHQ impacts on photosynthesis and redox. HHQ-induced cell cycle arrest in E. huxleyi did not significantly alter the photosynthetic energy conversion efficiency; however, the majority of light-harvesting complexes and transcripts of the Calvin cycle decreased in relative abundance under HHQ exposure (Fig. 3a). These findings parallel those described previously for the diatom Phaeodactylum tricornutum undergoing chemically mediated cell cycle arrest (61). In plants, the coordinated downregulation of transcripts involved in photosynthesis, electron transport, and the Calvin cycle is thought to allow for the reallocation of resources toward defense against bacterial and viral pathogens (62). However, a decrease in transcript abundance does not always correlate with a loss of protein function, as photosynthetic proteins have a long functional half-life in the cell, with the exception of ferredoxin (Fd) and ferredoxin NADP 1 oxidoreductase (FNR), both of which are involved in maintaining the cellular redox state following pathogen infection (62). Together, both ferredoxin and the isofunctional flavodoxin (Fld) participate in electron shuttling, preventing electron misrouting that can lead to ROS accumulation and restoring chloroplast redox homeostasis under environmental stress (63). Indeed, the genes and proteins with the most significant differential expression levels under HHQ exposure in E. huxleyi were Fd (58-fold increase in transcript and 3fold increase in protein abundances), FNR (85-fold increase in transcript abundance), and Fld (38-fold increase in transcript and 186-fold increase in protein abundances) (Fig. 3a) (see Supplemental Data File 1 at https://doi.org/10.6084/m9.figshare.14414285.v1), which may explain the observed lack of ROS production (Fig. S4). Additional reduction systems, including FAD/NAD(P)-binding oxidoreductase, ferredoxin nitrite reductase (Fd-NR), and glutathione reductase (GR), in HHQ-treated E. huxleyi cells were also significantly induced, which could ameliorate NADPH buildup (see Supplemental Data File 1 at https://doi.org/ 10.6084/m9.figshare.14414285.v1). Moreover, increased relative expression of vitamin B 6 (VitB 6 ) transcripts following HHQ treatment could protect against oxidative stress in chloroplasts (64), while increased relative expression levels of transcripts encoding proline oxidase (POX), pyrroline-5-carboxylate reductase (P5CR), and manganese superoxide dismutase (Mn-SOD) could explain the lack of mitochondrial ROS toxicity (Fig. S4) (see Supplemental Data File 1 at https://doi.org/10.6084/m9.figshare.14414285.v1). Together, these results suggest that HHQ-exposed E. huxleyi uniformly decreased the relative abundance of photosynthetic gene transcripts in support of a coordinated induction of defense responses aimed at maintaining cellular redox homeostasis without debilitating photosynthetic capacity.
Consequences of HHQ-induced cellular stasis. Given that viral replication requires the hijacking of host replication machinery and HHQ exposure inhibited DNA replication and repair in E. huxleyi, the impact of HHQ on host-virus dynamics was investigated. When E. huxleyi cells were exposed to HHQ and Emiliania huxleyi virus (EhV) strain 207 concurrently, virus-induced cellular death was significantly reduced (P value of ,0.0001 by ANOVAR) (Fig. 5a). This outcome was observed regardless of whether viruses were added simultaneously with HHQ (Fig. 5a) or 72 h after HHQ treatment (Fig. 5b). However, if HHQ addition was delayed even by 24 h, virus-induced mortality occurred in E. huxleyi (Fig. 5c). These results indicate the possibility that HHQ exposure early in viral infection critically impacts the effectiveness of the virus. There are numerous mechanisms by which HHQ may inhibit virus-induced mortality of E. huxleyi (Fig. 6). For example, HHQ may impact the entry of the virus into the cell. Significant morphological restructuring occurred following 24 h of HHQ exposure, which may prevent viral recognition, attachment, and/or endocytosis. Previous work has demonstrated that within 24 h of E. huxleyi viral infection, the virus requires the induction of host DNA replication machinery (65). Thus, HHQ may either inhibit the virus' ability to manipulate DNA replication or acquire necessary nucleotides for transcription, thereby stalling infection success. HHQ may also stall the induction of ROS production, which has been demonstrated previously to be necessary for successful E. huxleyi viral infection (66). In the transcriptomic and proteomic data presented here, significant upregulation of a variety of antioxidants, including Fd, FNR, Fld, Fd-NR, GR, Mn-SOD, POX, and VitB 6 , may counteract virus-induced remodeling of the host antioxidant network essential for viral replication. Likewise, the expression and activation of caspase and metacaspase proteases during infection are critical for enabling virus-induced lysis in E. huxleyi (67). However, these proteases were not upregulated and did not show activity in HHQ-exposed cells (Fig. S4) (see Supplemental Data File 1 at https://doi.org/10.6084/ m9.figshare.14414285.v1), further suggesting that HHQ exposure may disrupt these critical processes in viral infection. Attenuation of viral mortality would theoretically permit increased survival of phytoplankton and allow bacteria to continue to take advantage of coordinated nutrient exchange, common between bacteria and phytoplankton (68). Thus, the impacts of HHQ exposure on phytoplankton may have ecological consequences beyond shifts in algal physiology, to impacts on large-scale biogeochemical cycles.
Summary. Our laboratory findings demonstrate that a quorum-sensing signal produced by a marine bacterium significantly, but reversibly, leads to DNA lesions and cell cycle arrest in a eukaryotic phytoplankter, which can influence interkingdom virus-host interactions. In the eastern tropical South Pacific, .1-ng liter 21 surface concentrations of HHQ were found to correlate with areas of enhanced phytoplankton biomass (Fig. S7). These low concentrations of HHQ in bulk seawater are not surprising, as Nacyl homoserine lactones (69), vitamins (70), and other highly labile, trafficked compounds fundamental for growth and signaling are often found in low concentrations in bulk seawater. We anticipate that the primary abiotic sink for HHQ will be photooxidation by sunlight; however, the rate of photooxidation in seawater will strongly depend on a myriad of factors (e.g., depth, day length, and dissolved organic matter [DOM] concentrations, etc.). Previous work showed that the half-life of quinoline, the parent compound of HHQ, varied between 8 and 53 days using solar simulations (71). Furthermore, these measured bulk concentrations likely do not represent the effective concentration that a marine microbial cell would experience in the phycosphere (72).
Previous work has demonstrated that HHQ can significantly alter natural microbial community composition and growth rates (15), and here, we find that detectable in situ HHQ concentrations correlated with enhanced phytoplankton biomass. Together, these findings suggest that alkylquinolone signaling may play a significant role in structuring complex microbial communities, ultimately influencing primary production and biogeochemical cycles. In addition, our findings highlight the functional duality of bacterial cues that serve as diffusive messengers used as a communication tool in microbial communities but also as chemical mediators of marine microbial interactions.

MATERIALS AND METHODS
General cultivation conditions. For all experiments, axenic Emiliania huxleyi (CCMP2090, non-lith forming) (from the National Center for Marine Algae and Microbiota, East Boothbay, ME) was grown in natural seawater-based f/2 medium without silica (26). Cultures were maintained on a 14-h/10-h light (80 6 5 mmol photons m 22 s 21 )/dark cycle at 18°C, with a salinity of 35. These conditions are referred to here as general culturing conditions. Strain purity was confirmed using f/2 MM and f/2 MB purity test broths and visually confirmed by epifluorescence microscopy (73). Cultures (20 ml) were transferred weekly to maintain exponentially growing cultures.
Phytoplankton cells were enumerated by a hemocytometer or using a flow cytometer (Guava; Millipore). Via the flow cytometer, cell abundance was determined by using species-specific settings, including their forward scatter, side scatter, and red fluorescence (695/50-nm) emission characteristics for evaluating chlorophyll intensity. All samples were run at 0.24 ml s 21 for 3 min, either live or fixed with glutaraldehyde (0.5% final concentration). A correction factor was applied to fixed cell abundances to account for cell loss due to preservation.
Growth experiments. The HHQ concentration resulting in 50% growth inhibition (IC 50 ) was determined using triplicate, 2-or 20-ml cultures of E. huxleyi (;100,000 cells ml 21 ) exposed to HHQ (between 0.25 and 512 ng ml 21 ), PQS (0.5 to 530 mg ml 21 ), or the vehicle control (0.1% DMSO) for 72 h. Growth rates were calculated using an exponential growth equation and were plotted against the HHQ concentration to determine the IC 50 at 72 h postexposure as described previously (16). Concentrations of DMSO below 0.5% (vol/vol) have no impact on axenic E. huxleyi growth. DMSO was used as the solvent vehicle for HHQ and PQS.
To examine the impacts of HHQ, triplicate flasks of 30-ml cultures of E. huxleyi (;50,000 cells ml 21 ) were exposed to either 1 or 100 ng ml 21 HHQ or a vehicle (0.1% DMSO) control. The experiment mixture was sampled daily for 96 h to monitor E. huxleyi abundance, forward scatter, side scatter, red fluorescence (695/50 nm), and photosynthetic efficiency (F v /F m ). F v /F m was measured using a fluorescence induction and relaxation (FIRe) system (Satlantic). Samples were dark adapted for 30 min, and photosystem II kinetics were measured from the average of 10 iterations of an 80-ms single-turnover event and 1,000 ms of weak modulated light.
To measure recovery, after 96 h of HHQ exposure, triplicate 2-ml aliquots of an HHQ-exposed culture were transferred into 198 ml of fresh medium, effectively diluting HHQ to 1 ng ml 21 . The same dilution was made with the vehicle control treatment, and the experiment mixture was sampled daily for E. huxleyi growth rate, forward scatter, side scatter, and red fluorescence (695/50 nm).
To investigate viral infection dynamics, triplicate 50-ml cultures were prepared for the following treatments: E. huxleyi (;40,000 cells ml 21 ) plus the vehicle control (0.1% DMSO), E. huxleyi plus EhV 207 (3.2 Â 10 6 EhV particles ml 21 ), E. huxleyi plus HHQ (100 ng ml 21 ), and E. huxleyi plus HHQ and EhV 207. Once inside the cell, the virus hijacks host replication machinery to produce additional viral particles. This replication is dependent on functional de novo nucleotide synthesis enzymes, such as dihydroorotate dehydrogenase (DHODH), to provide the cell with sufficient nucleotide materials. Likewise, functional DNA repair, often mediated by poly(ADP-ribose) polymerase (PARP), is necessary to ensure that replication can continue. Successful viral replication then generates intracellular reactive oxygen species (ROS) and nitric oxide (NO) signaling, which in turn activates caspase proteases, allowing the release of replicated viral particles via programmed cell death (PCD)-induced cell lysis. (b) In HHQ-exposed phytoplankton cells, virus-induced mortality was not observed, but the mechanism by which HHQ impacts viral cycling remains unclear. HHQ may directly inhibit (shown as red lines) the activity of DHODH and PARP, which would prevent the production of viral particles via the collapse of DNA replication machinery. HHQ may also indirectly impact parts of the virus cycle (shown as dotted lines) by changing host physiology to disrupt recognition, nucleotide production, ROS production, caspase activation, or PCD.
The multiplicity of infection (MOI) was 80, to ensure successful viral infection potential. Samples were taken daily to monitor E. huxleyi abundance.
For all growth experiments, excluding the IC 50 calculation, significant differences between treatments were determined by comparing abundances over time using ANOVAR, followed by Dunnett's multiple-comparison test (74). All data were tested to ensure that they passed the assumptions for normality and sphericity prior to running the ANOVAR.
Physiological assays. Propidium iodide (PI) was used to quantitatively discriminate cell cycle stages in HHQ-exposed phytoplankton cultures over 122 h. Three replicate 2-liter cultures at ;33,000 cells ml 21 were dosed with either 100 ng ml 21 HHQ or the vehicle control (0.002% DMSO). Fixed cells were enumerated every 24 h via flow cytometry. Every 2 h, approximately 10 6 cells were subsampled, pelleted, and washed twice via centrifugation at 3,214 Â g for 15 min at 18°C. The dry cell pellets were resuspended in 1 ml of ice-cold liquid chromatography-mass spectrometry (LC-MS)-grade methanol, transferred to microcentrifuge tubes, and stored at 280°C. For reading, methanol-fixed cells were centrifuged at 16,000 Â g for 10 min at 4°C, methanol was removed, and pellets were resuspended in 1 ml of 1Â Dulbecco's phosphate-buffered saline (DPBS) before repelleting by centrifugation at 16,000 Â g for 10 min at 4°C. The pellet was resuspended in 0.5 ml of FxCycle PI/RNase solution (Thermo Fisher), incubated for 30 min in the dark, and then measured via flow cytometry (583/26-nm emission).
Diagnostic fluorescent dye assays were used to measure indicators of cell stress and programmed cell death (PCD) following HHQ treatment. Intercellular reactive oxygen species (ROS) and nitric oxide (NO) production, mitotoxicity, cytotoxicity, and caspase protease levels and activity in whole-cell lysates were measured in E. huxleyi (starting cell concentration of ;100,000 cells ml 21 ) following HHQ treatment (70 ng ml 21 or 100 ng ml 21 ) at various time points up to 72 h postexposure. See Text S1 in the supplemental material for detailed protocols.
E. huxleyi DNA integrity was examined using a modified protocol for the Click-iT terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) Alexa Fluor 488 imaging assay kit (Thermo Fisher). Four replicate E. huxleyi cultures (;250,000 cells ml 21 ) were assayed according to the manufacturer's protocol and sampled after 46 h of HHQ exposure, with tagged cells being enumerated via flow cytometry (512/18-nm emission). See Text S1 in the supplemental material for detailed protocols.
Transmission electron microscopy (TEM). Replicate 20-ml cultures of exponentially growing E. huxleyi cells (;100,000 cells ml 21 ) were exposed to either 100 ng ml 21 HHQ or the vehicle control (0.2% DMSO) for 24 h. Samples were concentrated by filtration on a 0.45-mm polycarbonate filter, transitioned out of f/2 medium via three sequential washes with 10 ml of 0.2 M sodium cacodylate buffer (pH 7.2), and then fixed in 2% glutaraldehyde in 0.2 M sodium cacodylate buffer (pH 7.2). Samples were postfixed in 2.0% osmium tetroxide for 1 h at room temperature and rinsed in double-distilled water (ddH 2 O) prior to en bloc staining with 2% uranyl acetate. After dehydration through a graded ethanol series, the cells were infiltrated and embedded in Embed-812 (Electron Microscopy Sciences). Thin sections were stained with uranyl acetate and lead citrate and examined with a JEOL 1010 electron microscope fitted with a Hamamatsu digital camera and AMT Advantage NanoSprint500 software.
Transcriptomic and proteomic analyses. A large-scale culturing experiment was performed with axenic E. huxleyi cells treated with either three concentrations of HHQ (1 ng ml 21 , 10 ng ml 21 , and 100 ng ml 21 ) or the vehicle control (0.002% DMSO) for 72 h. Following HHQ/DMSO exposure, 400-ml subsamples were taken from each quadruplicate 2-liter bottle at both 24 and 72 h for total RNA isolation, and an additional 1,200-ml subsample was taken at 72 h for total protein isolation. Total RNA and protein were isolated and quantified as described in Text S1 in the supplemental material.
For proteomic analysis, proteins were solubilized in urea, reduced, alkylated, and trypsin digested as described previously (81). The resulting peptide samples were desalted with a minicentrifugal C 18 column according to the manufacturer's instructions (Nest Group). Peptides were chromatographically separated (precolumn, 3-cm by 100-mm internal diameter [ID]; analytical column, 30-cm by 75-mm ID; resin, 3-mm C 18 -AQ) with a nanoAcquity ultraperformance liquid chromatography (UPLC) system (2 to 35% acetonitrile [ACN] and 0.1% [vol/vol] formic acid; 250 nl min 21 for 90 min) directly in line with a Fusion Lumos Orbitrap Tribrid mass spectrometer (Thermo Fisher Scientific) operated in the data-independent acquisition (DIA) mode according to methods described previously (82). To generate a peptide spectral library, 1 mg of a pooled sample containing equal parts from each peptide digest was analyzed with six gas-phase fractions covering m/z 400 to 1,000 in increments of 100 m/z (4 m/z staggered MS 2 windows and 2 m/z overlap). Each bioreplicate was then quantified in single DIA analyses (MS 1 , m/z 400 to 1,000; 8 m/z staggered MS 2 windows and 4 m/z overlap).
In order to generate absolute abundance measurements of detected proteins, raw MS data files were processed using msconvert (ProteoWizard) for demultiplexing and peak picking. EncyclopeDIA (V0.7.4) was used to (i) search the resulting fragmentation spectra against the UniProt E. huxleyi CCMP1516 protein and contaminant database (10.0-ppm precursor, fragment, and library tolerances), (ii) provide peptide-level area under the curve (AUC) data, and (iii) generate quantitative reports of identified peptides and proteins for each HHQ MS experiment (1% false discovery rate). Significant changes (q , 0.05) in protein abundances between HHQ treatments and the vehicle control were calculated as log 2 fold changes between treatments. Complete details of protein sample preparations, chromatographic separations, mass spectrometry detection, and quantification can be found in Text S1 in the supplemental material.
Proteomic data were matched to the transcriptomic data utilizing the corresponding NCBI accession numbers. As many of the genes and proteins were uncharacterized, potential homologs of known proteins of interest were identified by querying the amino acid sequences of selected human proteins against the translated E. huxleyi (CCMP2090) genome, utilizing a significance threshold of an E value of ,1 Â 10 220 . Combined data were visualized utilizing the ggplot2 and pheatmap packages in R.
PARP inhibition and homology modeling. To examine the impact of alkylquinolone exposure on mammalian PARP activity, an inhibition assay was performed using the PARP universal colorimetric assay kit (R&D Systems) according to the manufacturer's instructions. Human PARP enzyme (0.5 U) was exposed to 50 mM HHQ (n = 4), 50 mM PQS (n = 4), or the vehicle control (0.25% DMSO) (n = 4) for 15 min prior to the addition of PARP activity buffer. See Text S1 in the supplemental material for a detailed protocol.
The E. huxleyi sequence under GenBank accession number XP_005783504.1 was aligned to the Protein Data Bank (PDB) database to determine the closest structural homolog with a small molecular inhibitor, veliparib, in the active site that could lend insight into HHQ binding.
Detection of HHQ in environmental samples. Seawater samples were collected along a cruise track from Manta, Ecuador, to Tahiti from October to December 2013 (U.S. GEOTRACES EPZT GP16) as described previously (83). Briefly, seawater was collected at a 3-m depth by a tow fish and pumped at a flow rate of 250 ml min 21 through a 0.2-mm filter and a polytetrafluoroethylene column packed with 20 g of polystyrene resin (Bondesil ENV; Agilent). Each sample represents an integrated average of 400 to 600 liters of water across a wide region. Samples were frozen onboard at 220°C. Prior to analysis, thawed columns were rinsed with 500 ml of 18.2 MX-cm ultrahigh-purity water (qH 2 O) and eluted with 250 ml of LC-MS-grade methanol. The extracts were concentrated by rotary evaporation and brought up in a final volume of 6 ml of qH 2 O that was stored at 220°C. The organic extracts were separated by a high-pressure liquid chromatography system (Dionex Ultimate 3000) coupled to an Orbitrap Fusion MS instrument (Thermo Scientific). The specific methodology can be found in Text S1 in the supplemental material.
Data availability. Sequences from this study have been deposited in the Gene Expression Omnibus (GEO) and are accessible through GEO series accession number GSE131846. The raw mass spectrometry proteomics data and subsequent spectral libraries have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository under accession number PXD011560 (https://www.ebi.ac .uk/pride/archive/projects/PXD011560).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. TEXT S1, PDF file, 0.2 MB.