ABSTRACT
Corals are colonized by symbiotic microorganisms that exert a profound influence on the animal’s health. One noted symbiont is a single-celled alga (from the family Symbiodiniaceae), which provides the coral with most of its carbon. During thermal stress, the algae’s photosystems are impaired, resulting in a toxic accumulation of reactive oxygen species (ROS) that cause cellular damage to both the host and symbiont. As a protective mechanism the coral host and algal symbiont disassociate; this process is known as bleaching. Our goal was to construct a probiotic comprised of host-associated bacteria able to neutralize free radicals such as ROS. Using the coral model, the anemone Exaiptasia diaphana, and pure bacterial cultures isolated from the model animal, we identified six strains with high free radical scavenging ability belonging to the families Alteromonadaceae, Rhodobacteraceae, Flavobacteriaceae, and Micrococcaceae. In parallel, we established a “negative” probiotic consisting of genetically related strains with poor free radical scavenging capacities. From their whole genome sequences, we explore genes of interest that may contribute to the potential beneficial roles of these putative probiotic members, which may help facilitate the therapeutic application of a bacterial probiotic. Probiotics is one of several interventions currently being developed with the aim of augmenting climate resilience in corals and increasing the likelihood of coral reef persistence into the future.
IMPORTANCE Coral bleaching is tightly linked to the production of reactive oxygen species (ROS), whereby ROS accumulates to a toxic level in host-harboring algae cells leading to coral-algal dysbiosis. Interventions targeting toxic ROS accumulation, such as the application of exogenous antioxidants, have shown promise for maintaining the coral-algal partnership. With the feasibility of administering antioxidants directly to corals low, we have applied bioengineering strategies to develop a probiotic to neutralize toxic ROS during a thermal stress event. This probiotic can then be tested with corals or a coral model to assess its efficacy in improving coral resistance to environmental stress.
INTRODUCTION
Coral reefs are among the most biologically and economically productive ecosystems on Earth [1, 2]. Though they make up less than 0.1% of the ocean floor [3], coral reefs support fisheries, tourism, pharmaceuticals and coastal development with a global value of $8.9 trillion “international $”/year [4]. Corals and other reef organisms have been dying, largely due to anthropogenic influences such as climate change[5, 6], which has led to an increased frequency, intensity and duration of summer heat waves that cause coral bleaching [7, 8].
The coral holobiont [the sum of the coral animal and its symbiotic partners, including intracellular algae, endolithic algae, fungi, protozoans, bacteria, archaea and viruses; 9] is an ecosystem engineer. By secreting a calcium carbonate skeleton, the reef structure rises from the ocean floor, forming the literal foundation of the coral reef ecosystem. The success of corals to survive and build up reefs over thousands of years [10] is tightly linked to their obligate yet fragile symbioses with endosymbiotic dinoflagellates of the family Symbiodiniaceae [11].
Intracellular Symbiodiniaceae translocate photosynthetically fixed carbon to the coral host [12] in exchange for inorganic nitrogen, phosphorus and carbon and location in a high light environment with protection from herbivory [13, 14]. During periods of intense thermal stress, the relationship between the coral host and their Symbiodiniaceae breaks down, resulting in a separation of the partners and a state of dysbiosis. This phenomenon, ‘coral bleaching’, is devastating to the host and detrimental to the reef system. Debilitating effects of bleaching on the coral include reduced skeletal growth and reproductive activity, a lowered capacity to shed sediments, and an inability to resist invasion of competing species and diseases. Severe and prolonged bleaching can cause partial to total colony death, resulting in diminished reef growth, the transformation of reef-building communities to alternate, non-reef building community types, bioerosion and ultimately the disappearance of reef structures [11].
Although there are several hypotheses detailing the mechanisms driving bleaching [see 15, 16- 18], a common theme is the overproduction and toxic accumulation of reactive oxygen species (ROS) from the algal symbiont. Excess ROS are generated by a number pathways including heat damage to both photosynthetic and mitochondrial membranes [19, 20], and are shown to play a central role in injury to both symbiotic partners and to inter-partner communication of a stress response [15]. Once generated, ROS causes damage to many cell components including photosystem II (PSII) reaction centers in the Symbiodiniaceae, specifically at the D1 and D2 proteins, [see review in 21]. Exposure to elevated temperatures [22] can result in photoinhibition of photosynthesis in Symbiodiniaceae. Once damaged, Symbiodiniaceae are no longer able to maintain their role in the symbiotic relationship with corals and separate from the host tissue via in situ degradation, exocytosis, host cell detachment, host cell apoptosis or host cell necrosis [15].
Probiotics are preparations of viable microorganisms that are introduced to alter a microbial community in a way that is beneficial to the host. Microbiome engineering through the addition of probiotics has been postulated as a key strategy to manipulate host phenotypes and ecosystem functioning for coral reefs [23-28]. The differences in the bacterial species composition of healthy and thermally stressed corals [29-34] and the coral model Exaiptasia diaphana [35-37] suggest a role for microbiome engineering in cnidarian health. A disruption to the bacterial community of Pocillopora damicornis with antibiotic treatment diminished the resilience of the holobiont during thermal stress, whereas intact microbial communities conferred resilience to thermal stress and increased the rate of recovery after bleaching events to the coral holobiont [38]. The relative stability of coral-associated bacterial communities have also been linked to coral heat tolerance; the bacterial community of heat sensitive Acropora hyacinthus corals shifted when transplanted to thermal stress conditions, whereas heat- tolerant A. hyacinthus corals harbored a stable bacterial community [39].
In recent years, researchers have begun to explore the use of probiotics in corals and the model organism for corals, E. diaphana. To inhibit the progression of white pox disease in E. diaphana, caused by the pathogen Serratia marcescens, an Alphaproteobacteria cocktail containing several Marinobacter spp. isolates was applied [40]. These strains were able to inhibit both biofilm formation and swarming in S. marcescens, which halted disease progression in E. diaphana. The probiotic was deemed effective as anemones exposed to both the cocktail and pathogen survived after seven days, while anemones in the S. marcescens control experiment died. A bacterial consortium native to the coral Mussismilia harttii was selected to degrade water-soluble oil fractions[41]. This bioremediation strategy reduced the negative impacts of oil on M. harttii health and accelerated the degradation of petroleum hydrocarbons [41]. Coral microbiomes have also been manipulated to mitigate the effects of thermal stress. This manipulation of the coral-associated microbiome was facilitated through addition of a consortium of native or seawater derived bacteria to the surface of P. damicornis [42]. The results from this study suggest the consortium was able to partially mitigate coral bleaching and provides promising initial results in the field of coral probiotics.
The goal of this research was to identify bacterial strains suitable for use in a probiotic to mitigate the effects of thermal stress in E. diaphana. Given the potential role of ROS in the bleaching process, our focus was to select diverse E. diaphana-sourced bacterial strains with antioxidant properties while avoiding potential pathogens. Antioxidant properties were measured using the stable free radical 2,2-diphenyl-1-picrylhydrazyl (DPPH), which is reduced in the presence of an antioxidant molecule, undergoing a color change from a violet to a colorless solution.
RESULTS
Diversity of culturable bacteria associated with E. diaphana
A total of 842 isolates were obtained from four genotypes of GBR-sourced E. diaphana, with no significant differences in bacterial colony forming units (CFUs) between the four genotypes, regardless of growth medium, with (mean±SE) 5.9-10.3 x 103 cells per anemone on Reasoner’s 2A agar (R2A) and 6.3- 10.4 x103 cells per anemone on marine agar (MA) (p>0.05). Partial 16S rRNA gene sequences (∼1000 bp) were used to identify the closest matches from the NCBI database using BLASTn. In total there were 109 species in 64 genera, 27 families and six phyla (Fig. 1). The most abundant genera were Alteromonas, Labrenzia, and Ruegeria (Table 1). Gram-positive bacteria comprised 23 species, including Microbacterium (31 isolates) and Micrococcus (28 isolates). Eight genera were found to be associated with all four genotypes (Table 1); these eight genera made up 59.4% of all E. diaphana-associated bacterial isolates.
Bacterial probiotic selection
Of the original 842 isolates, 709 were screened for their ability to scavenge free radicals. Those isolates were divided into three categories, positive (144), weakly positive (121), and negative (444). There was no clear pattern of free radical scavenging (FRS) capacity at the family level (Fig. 2A) with strain specific responses by species. Ninety-eight strains representing eight families and 18 genera were quantitatively assessed for FRS (Fig. 2B). From these isolates, probiotic members were selected by choosing E. diaphana-associated bacteria species with separate strains displaying a high (“positive”) or low (“negative”) FRS ability (Fig. 2C-G; Table 2). Of the 12 selected probiotic members, seven were catalase positive and five were catalase negative (Table 2). In each probiotic set (i.e. high or low FRS strains), none of the selected isolates showed antagonistic activity against one other as evidenced by the absence of any zone of inhibition and growth from each combination of isolates on a plate.
Comparative genomics
As part of the characterization of the 12 isolates (six positive and six conspecific negative FRS isolates), a draft genome sequence was determined for each isolate. A summary of the data and metrics for the draft genome sequences is presented in Table S1. The diversity of the six pairs of isolates is indicated by the %G+C range (35% to 72%) and genome size (2.4 Mb to 6.8 Mb). Each isolate pair was classified as the same species except the Micrococcus stains. Isolates MMSF00068 (high FRS strain) and MMSF00107 (low FRS strain) are classified as Micrococcus luteus by both 16S rRNA gene and genome-based methods, but MMSF00107 is classified as Micrococcus yunnanensis by NCBI. These isolates have the smallest genomes among the isolates (2.43 Mb 2.48 Mb, respectively) and the highest G+C content of 72.8% and 72.4%, respectively. The classification of isolate MMSF00107 is uncertain, particularly given that the genome sequence for the type strain for M. yunnanensis is not yet available.
Direct pairwise comparison of the genome sequences between the pairs of isolates revealed a wide range of genome variation, ranging from ∼190,000 single nucleotide polymorphisms (SNP) differences between the Alteromonas oceani strains, to fewer than five SNPs between the genomes of Labrenzia aggregata (Table S2). The genome sequences of Winogradskyella poriferorum isolates MMSF00910 and MMSF00046 are nearly identical with fewer than ten pairwise core SNP differences; but there were some accessory genome differences.
Genes of interest
The annotated genome sequences of each selected probiotic member were searched for key genes of interest (Table S3-S4). Dimethylsulfoniopropionate (DMSP) cleavage to dimethylsulfide (DMS) was identified by presence of one or more of the DMSP lyase genes; dddP, dddD, dddL, dddW, and dddQ. Only L. aggregata strains contained DMSP lyase genes (dddP and dddL) in their whole genome sequences (WGS). DMSP biosynthesis was identified by the presence of dsyB, which is the only described gene for an enzyme in the DMSP biosynthesis pathway. The presence of dsyB in both L. aggregata isolates suggests that they have the capacity to produce DMSP. cobP was used as an indicator gene for the presence of the dynamic vitamin B12 pathway, which contains 27 genes (Table S4). Again, only the L. aggregata isolates contained cobP. Catalase positive strains were identified by the presence of katG; all positive and negative FRS strains contained katG except the Micrococcus spp. strains, in which katA and katE were detected.
16S rRNA gene copy number
The 16S rRNA gene copy numbers of the 12 draft genomes were estimated using a read depth approach (Table S1). The copy numbers were close between each other within pairs of isolates, in which the pair of isolates MMSF00257 and MMSF00958 contained the most copies (5.15 and 4.79, respectively) and the isolates MMSF00046 and MMSF00910 contained the fewest copies (1.03 and 0.77, respectively).
DISCUSSION
The 842 E. diaphana bacterial isolates reported here and based on 16S rRNA gene-based taxonomic classification comprise 109 species from 64 genera and six phyla. Using metabarcoding, studies of microbiomes associated with E. diaphana have revealed a similar diversity at the phylum level for E. diaphana sourced from the GBR [35], Hawaii [strain H2; 43], Pacific and Caribbean [55], Atlantic [strain CC7; 45] and Red Sea [36], as well as stony corals (see a review by Blackall, Wilson [63]). Thus, our culture collection of E. diaphana bacterial isolates suitably represents the diversity of the E. diaphana-associated microbiome. Though previous studies have used only a narrow bacterial species diversity to develop a probiotic [40, 42], our culture collection is sufficiently diverse to better select potential probiotics.
Inoculated bacteria can be acquired by developing coral larvae [25] and probiotic mixes applied to cnidarians have been shown to be effective in inhibiting disease progression [40] and increasing resistance to the negative effects of oil [41] and heat exposure [42]. The E. diaphana probiotics generated in this study were assembled from 12 isolated bacterial strains selected based on their FRS ability since free radical production, specifically ROS, is relevant in coral bleaching. The broader culture collection contains bacteria with a wide range of FRS capacity.
The consistent and frequent reporting of our probiotic bacteria genera in E. diaphana and coral studies (see Table 2) suggests these bacteria may have key functions in cnidarian holobionts. Among these potential functions are the production of antioxidants such as DMSP, the breakdown of DMSP to other antioxidants (dimethyl sulfide (DMS), acrylate, dimethyl sulfoxide (DMSO), and methane sulfinic acid (MSA) [64]. L. aggregata has been reported to produce DMSP in the absence of any methylated sulfur compounds with dsyB identified as the first DMSP biosynthesis gene in any organism [65]. dsyB was found in the whole genome sequences of both high and low FRS L. aggregata strains (Table S3). Many E. diaphana-sourced bacterial species, specifically relatives of our selected probiotic members, are implicated in the degradation of DMSP to DMS (Alteromonas spp., [66]; Labrenzia spp., [67]). dddP codes for the enzyme responsible for cleaving DMSP to DMS and acrylate and was used (from the Prokka annotation) as an indicator of a DMSP degradation genotype. Only the L. aggregata isolates were found to be able to degrade DMSP (Table S3).
Carotenoids are among the strongest antioxidants and are highly reactive against both reactive oxygen species and free radicals [68-72]. Carotenoids are lipid-soluble pigments, and in bacteria they give an orange-yellow hue to colonies. Two of the five selected probiotic genera produce orange/yellow colonies (Winogradskyella, Micrococcus), and there is evidence of carotenoid production by marine Flavobacteriaceae [72, 73] and Micrococcus strains [74]. A marine Flavobacteriaceae (strain GF1) was found to produce the potent antioxidant carotenoid zeaxanthin that protected Symbiodiniaceae from thermal and light stress [75].
Vitamin B12 is a cofactor involved in the production of the amino acid methionine, which is needed to synthesize every protein and in diverse metabolic pathways including generation of the antioxidants glutathione and DMSP [76]. Vitamin B12 is synthesized by many heterotrophic bacteria [77]. Genomic evidence suggests that Symbiodiniaceae may have lost the capacity to synthesize vitamin B12 due to changes in their metabolic enzymes during their evolution [78], agreeing with other work that free-living Symbiodiniaceae depend on bacterial symbionts to gain access to this important cofactor [79]. The genes involved in the biosynthesis of vitamin B12 have been found in coral-associated bacteria, specifically L. aggregata cultured from the Caribbean coral, Orbicella faveolata [80].From the annotated draft genome sequence, using the cobP gene [77] as the genotypic indicator, both L. aggregata isolates are capable of Vitamin B12 biosynthesis. The Marinobacter salsuginis isolates lack this gene but have other detectable vitamin B12 synthesis genes (Table S4).
Bacteria have developed highly specific mechanisms to protect themselves against oxidative stress with enzymes such as catalase/peroxidase and superoxide dismutase (SOD), small proteins like thioredoxin and glutaredoxin, and molecules such as glutathione (GSH) in combination with glutathione peroxidase and glutathione reductase [see review 81]. It has been suggested that increasing the in hospite concentration of catalase in the coral holobiont by the application of a probiotic with catalase-positive organisms, could possibly minimize the impact of thermal stress by neutralizing hydrogen peroxide [27]. While this is a justifiable hypothesis worth investigating further, to our knowledge, no studies have shown that catalase producing bacteria can reduce the concentration of ROS in hospite. Here we tested all probiotic candidates for catalase production using a standard hydrogen peroxide assay [82]. Catalase participates in cellular antioxidant defense by decomposing hydrogen peroxide; this decomposition is enzymatically driven by two catalases (yielding H2O and O2): hydroperoxidase I (HPI),which is present during aerobic growth and transcriptionally controlled at different levels, and hydroperoxidase II (HPII), which is induced during stationary phase [83]. Most phenotypically determined catalase positive strains (MMSF00249, L. aggregata; MMSF00257, A. macleodii; MMSF00964, M. salsuginis) had homologs for the catalase-peroxidase gene, katG; the Micrococcus spp. isolates did not (Table S3). However, katG homologs were found in all strains that were phenotypically catalase negative, indicating that these genes may not be active during culture. Given the inconsistency between the catalase and DPPH results (in many cases high FRS isolates were catalase negative while the low FRS strain was catalase positive) the catalase results were not used as a primary factor in selecting members of the positive and negative probiotic.
Members of the selected probiotic consortium (e.g., Marinobacter spp. and Winogradskyella spp.) have some described roles in coral/E. diaphana health. For example, coral-bleaching events are often followed by disease outbreaks [84-86] such as white pox [87]. When inoculated in the presence of the white pox pathogen, S. marcescens, Marinobacter strains were able to inhibit the progression of the infection in E. diaphana [40]. In a bioremediation study on coral and oil, Winogradskyella sp. showed a significant decrease in abundance during oil treatment, which was correlated with a decrease in coral holobiont health measured via maximum quantum yield [41]. The presence of this microbiome member could play a role in protecting the Symbiodiniaceae photochemical ability during periods of stress.
A critical characteristic in the selection of probiotic members is the maintenance and proliferation of the inoculated bacteria in the host over time [49] and their potential transmission to the next generation. There is evidence that corals release bacteria with their offspring such as Alteromonas [46], Flavobacteriaceae [46], Rhodobacteraceae [88], and Marinobacter [56]. While the majority of broadcast spawning corals do not transfer their bacterial symbionts with their gametes (vertical transfer) [89], the brooding coral Porites astreoides transmits bacteria vertically to planulae with two bacterial taxa (Roseobacter clade- associated bacteria and Marinobacter spp.) consistently and stably associating with juvenile P. astreoides [56]. In addition to the potential antioxidant properties of Marinobacter [90], others like Roseobacter spp. might be beneficial in facilitating larval settlement. If adult corals stably associate with inoculated probiotic candidates like Marinobacter, Alteromonas, and Winogradskyella, they may be passed on to offspring and thus have a long-term positive impact on these individuals.
Interactions within the microbiota associated with marine holobionts are undoubtedly complex. Results presented in this manuscript show that pure cultured bacteria from E. diaphana can scavenge free radicals, albeit at a strain-specific rate. This suggests that the selection and inoculation of these high FRS strains could be beneficial to the host under high oxidative stress conditions, such as those that contribute to coral bleaching. Conspecific pairs of six bacteria provide an opportunity to determine the genetic basis for measured phenotypic differences between the pairs. An essential element of this future work will be to investigate the stability of the phenotypic differences observed and this stability may be reflected in the nature of the genetic differences between the pairs of strains. Where isolate pairs are distantly related, it is unlikely that the genetic basis for the phenotypic difference will be identified using comparative genomics.
The probiotic members were chosen from a highly diverse pool of E. diaphana-sourced bacterial isolates. While the selected probiotic members are phylogenetically diverse, potentially promising probiotic bacteria in the culture collection were omitted based on our selection criteria. For example, Ruegeria spp., which were excluded based on the absence of a high FRS strain, have the ability to breakdown DMSP and can participate in denitrification [80]. Muricauda isolates had high FRS abilities, but they did not grow consistently in the selected medium and therefore were excluded from the consortium. Like Ruegeria, Muricauda also has genes for denitrification [80], can oxidize DMS to produce DMSO [67], and produces potent carotenoids [91] that can mitigate thermal and light stresses in Symbiodiniaceae cultures [75]. Muricauda will be involved in future probiotic evaluations.
At present there is no biological treatment that can minimize coral bleaching in the field. Management priorities for coral reefs must move beyond documenting their declines and toward investigating potential approaches for mitigating coral bleaching, such as the application of coral probiotics. We believe that the application of a coral probiotic specifically tailored to address coral bleaching by neutralizing ROS could provide hope for the future of coral reefs. We also understand that this form of intervention may not work alone, but could benefit by pairing with other strategies such as enhancing coral resistance and resilience using other assisted evolution approaches such as assisted gene flow, hybridization and experimental evolution of the algal symbionts [24, 26, 92-98].Climate warming will continue even with the most drastic reductions in greenhouse gas emissions, thus, additional interventions such as coral probiotics present an alternative that could lead to relief from coral bleaching in real time.
MATERIALS AND METHODS
Isolation of bacterial isolates
Great Barrier Reef (GBR) origin E. diaphana were maintained in the laboratory at 26°C [99] and used to isolate probiotic candidates. Sixteen individuals from each of four E. diaphana genotypes (AIMS1-4) were collected using sterile disposable pipets and gently transferred to filter-sterilized (0.2 µm) reverse osmosis (RO) water reconstituted Red Sea Salt(tm) (Red Sea; RSS) at ∼34 parts per thousand (ppt) salinity (fRSS) and placed in the dark for 30 min. This was done to remove some of the influence of the external seawater on the bacterial community. After 30 min, each anemone was transferred to a sterile glass homogenizer with 1 mL of fRSS. Each homogenate was used to prepare serial dilutions from 10− 1 to 10−4. From each dilution, 50 µL was spread plated onto three replicate plates each of MA (Difco(tm) Marine Agar 2216) and R2A (CM0906, Oxoid) supplemented with 40 g L-1 RSS and incubated at 26°C. After one week, CFU counts were completed. Individual bacterial isolates were sub-cultured to purification from plates with <100 CFUs onto the initial isolation medium. All purified bacterial isolates were resuspended in 40% glycerol, aliquoted into 1.2 mL cryotubes and stored at −80°C.
Identification of Exaiptasia-sourced isolates
Colony PCR with the universal bacterial primers 27f (5’ – AGA GTT TGA TCM TGG CTC AG – 3’) and 1492r (5’ – TAC GGY TAC CTT GTT ACG ACT T – 3’) [100] was used to generate 16S rRNA gene amplicons from each isolate. Briefly, cells from each pure culture were suspended in 20 µL Milli-Q water and denatured at 95°C for 10 min. The suspension was then centrifuged at 2,000 x g at 4°C for two minutes and the supernatant was used as the DNA template for PCR amplification. The PCR was performed with 20 µL Mango Mix(tm) (Bioline), 0.25 µM of each primer and 2 µL of DNA template in a final volume of 40 µL. The thermal cycling protocol was as follows: 95°C for 5 min; 35 cycles of 95°C for 1 min, 50 °C for 1 min, and 72°C for 1 min; and a final extension of 10 min at 72°C. Amplicons were purified and Sanger sequenced on an ABI sequencing instrument by Macrogen Inc. (Seoul, South Korea) or by the Australian Genome Research Facility (AGRF) using the 1492r primer. Trimmed high quality read data from each isolate was used for presumptive identification by querying the 16S rRNA gene sequences via the Basic Local Alignment Search Tool (BLASTn). For some isolates the near-complete 16S rRNA gene sequence was determined by sequencing with additional primers (27f, 357f (5′-CCT ACG GGA GGC AGC AG-3′, [101]), 926f (CCG TCA ATT CMT TTR AGT TT, [102]), 519r (5’-GWA TTA CCG CGG CKG CTG-3’, [101]), 926r (5’–AAA CTR AAA MGA ATT GAC GG–3’,[102]), and 1492r). The six reads for each isolate were aligned using Geneious Prime 2019.1.2 (https://www.geneious.com) via the Geneious global alignment default settings with automatic determination of read direction. From this alignment, a consensus sequence for the 16S rRNA gene was constructed based on the frequency of a base and its quality (from chromatogram data) in each alignment column. The consensus sequence length for each of the six isolate pairs varied from 1352 to 1495 nucleotides. GenBank accession numbers for sequences are shown in Table 1.
Qualitative free radical scavenging assay
DPPH is a stable free radical that is purple in its oxidized state but becomes white-yellow when reduced by antioxidants, and has been used to identify FRS marine bacteria [103, 104]. To qualitatively assess E. diaphana-associated bacteria isolates for FRS ability, a sterile Whatman #1 filter paper was gently pressed against fresh (2-4 days old) colonies from a streak plate. Plates (with filter paper) were then incubated overnight at 26°C. The following day, filter papers were removed with forceps, allowed to dry in a fume hood for 30 min, and 500 µL of a 0.2 mM DPPH (Cat# D9132, Sigma-Aldrich) solution in methanol was applied with a pipette over individual colonies. As a positive control, a few drops of 0.1% (w/v) L-ascorbic acid (Cat# A7631, Sigma-Aldrich) were placed on a separate filter. The response of each isolate to DPPH was recorded within 3 min of DPPH application; a positive response was recorded when a white-yellow halo appeared around individual colonies within 1 min, a weak positive response was assigned to strains that had a halo form between 1 and 3 min after DPPH application, and a negative response was listed for strains that failed to form a halo (Fig. 3). Approximately 700 isolates were screened using the qualitative DPPH assay.
Quantitative free radical scavenging assay
To quantitatively assess the FRS ability, select isolates were grown in R2A broth (see Table S5 for composition) made by suspending 43.12 g in 1 L of MilliQ water, dissolving the medium completely, and sterilization by autoclaving at 121°C for 15 min. Fifty mL aliquots of autoclaved medium were distributed into sterile 250 mL Erlenmeyer flasks and each flask was inoculated with an isolate colony grown on agar (R2A or MA). Cultures were grown with shaking (150 rpm; Ratek orbital incubator) at 37°C for 48 h. A minimum of three replicate cultures were grown per isolate. After 48 h, the optical density of each culture was measured at 600 nm (OD600, CLARIOstar PLUS, BMG Labtech), and the cultures (including negative medium controls) were centrifuged at 3000 x g at 4°C for 30 min (Allegra X- 12R) to pellet the bacterial cells. The cell-free supernatants (CFSs) were collected, frozen at - 80°C, freeze dried (Alpha 1-4 LDplus, Martin Christ), and stored under inert gas in a dark, dry environment until analysis. Antioxidants were extracted from the CFSs by resuspending at 50 mg mL-1 in 100% methanol, sonicating (Branson 2510) for 5 min, then centrifuging at 3000 x g for 5 min at 4°C. Quantitative DPPH assays were run by creating a 1:1 solution of 0.2 mM DPPH in methanol and CFS extract to a final volume of 1 mL, vortexing, and reaction in the dark for 30 min at room temperature. Samples were then vortexed briefly, and three 300 µL replicates of each sample were transferred into a well of a 96 well plate. FRS was measured by determining absorbance at 517 nm (Enspire 2300 plate reader, Perkin Elmer). Decolourization of DPPH was determined by measuring the decrease in absorbance at 517 nm, and the FRS activity was calculated according to the formula, % DPPH scavenging activity = (Control – Sample) / Control ×100, where, Control is the absorbance of the DPPH control (1:1 0.2 mM DPPH:methanol), and Sample is the absorbance of CFS extract in DPPH. All samples were measured against a 100% methanol blank. Positive controls consisting of 0.01 - 0.001% (w/v) L-ascorbic acid were run on each 96-well plate. FRS activity ranged from 0-90%.
Catalase assay
The pelleted cells from above were resuspended in 2 mL fRSS and 500 µL hydrogen peroxide giving a final concentration of 16 mM. If bubbles appeared, the organism was considered catalase positive. If there were no bubbles, the organism was classified as catalase negative.
Inhibition testing
Each paired set of high and low FRS strains were inoculated crosswise along the middle of MA plates to test for antagonism. Plates were kept at 26 °C and monitored daily for up to 7 days for antagonistic activity by documenting the presence or absence of both inoculated isolates and if there was a zone of inhibition between them.
Phylogenetic analysis
All partial 16S rRNA gene sequences (842) were aligned with reference sequences (72) of closely related organisms using Geneious Prime 2019.1.2 (https://www.geneious.com). This alignment was used to construct a neighbor-joining phylogenetic tree using the Jukes-Cantor method. Maximum-likelihood dendrograms were generated with bootstrap values of 1000.
Whole genome sequence analysis
Positive FRS strains along with conspecific negative FRS strains were selected for genome sequencing; in total, six pairs of isolates were sequenced. Genomic DNA was isolated from a single colony using a JANUS Chemagic Workstation and Chemagic Viral DNA/RNA kit (PerkinElmer). Libraries were prepared with the Nextera XT DNA sample preparation kit (Illumina). Readsets were produced using the Illumina sequencing platform (Instrument: Illumina NextSeq 500, 150 base, paired-end) and the whole genome shotgun (WGS) method. Read depth coverage was approximately 100 times assuming a genome size of 4 M bases.
Illumina readsets for each isolate were assembled using Skesa [105] and the draft genome sequence annotated using Prokka [106]. No evidence of mixed colonies or sequence contamination was detected. A genome sequence based taxonomic classification for each isolate was determined using Kraken2 [107] with the Genome Taxonomy Database [GTDB; 108] as the curated genomic data source. Classification was primarily based on the genome sequence of related isolates (within the relevant species where possible), which were obtained from GenBank. In situations where genomes of taxonomically relevant individuals were available, a species level classification was possible. Where available, closed genome sequences from GenBank were used for comparative genomics analysis. Core genome comparisons were performed, as implemented in Nullarbor (https://github.com/tseemann/nullarbor), for each of the six pairs of isolates, with phylogenies inferred using core SNP differences. Genes of interest for DMSP synthesis and degradation, vitamin B12 synthesis, and catalase were identified from the annotated genome sequence (GFF format) produced by Prokka; specific genes were identified by both name and Refseq accession number.
16S rRNA gene copy number estimation
The 16S rRNA gene copy number of the 12 draft genomes was predicted by the 16Stimator pipeline [109]. Briefly, all the 12 genomic assemblies were submitted to the RAST server [110], and the positions of 16S rRNA and a set of single-copied housekeeping genes (Table S6) were extracted from the RAST annotations. The clean readsets were mapped back to the corresponding genomic assemblies by Bowtie 2 [111] to determine the read depth of each position. Finally, the 16S copy number of each isolate was calculated by dividing the median depth of 16S gene by the median depth of the single-copied housekeeping genes after the read depths were calibrated by the model parameters provided by 16Stimator.
Statistical analysis
CFU counts were analyzed in R [v3.6.2, 112] by first checking the assumptions of equal variance and homogeneity. An analysis of variance test was used to detect differences in the mean number of bacterial colonies from each anemone genotype by solid growth media (R2A or MA). A one-way analysis of variance [one-way ANOVA; 113] was used to determine if there were significant differences between FRS abilities of selected positive (high FRS), negative (low FRS), and media controls, and pairwise comparisons were performed using Tukey’s HSD [114, 115]. Each probiotic pair and media control was tested to determine if data met the assumptions of normality and homoscedasticity. If either assumption was violated, the non-parametic Kruskal-Wallis rank sum test [116] was used with a Dunn test [117] for multiple comparisons (p-values adjusted with the Benjamini-Hochberg method [118]) with the R package “FSA” [119].
Data availability
WGS raw reads are freely available in the Sequence Read Archive under BioProject PRJNA574193; the complete data set is listed in Table S1.
Supplemental Tables
ACKNOWLEDGEMENTS
This research was supported by the Australian Research Council Discovery Project grant DP160101468 (to MJHvO and LLB). MJHvO acknowledges Australian Research Council Laureate Fellowship FL180100036. We are grateful to Leon Hartman, Giulia Holland, and Shona Elliot- Kerr for their contributions in the preliminary culturing and screening of anemone-associated bacteria. Dr. Gayle Philip contributed with bacterial whole genome sequence analysis and Leon Hartman assisted with figure designs and reviewed the manuscript. Xavier Smith assisted with bacterial inhibition tests. Whole genome sequencing was organized by Dr. Glen Carter at the Peter Doherty Institute, Melbourne, Australia.
AMD, MvO and LLB conceived and designed the study. AMD performed the sampling and sample processing. AMD, DB, and HL completed bioinformatic analyses. AMD wrote the first draft. All authors edited and approved the final manuscript.
Footnotes
Importance statement updated.
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.
- 17.
- 18.
- 19.↵
- 20.↵
- 21.
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.
- 31.
- 32.
- 33.
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.
- 44.
- 45.
- 46.↵
- 47.
- 48.
- 49.↵
- 50.
- 51.
- 52.
- 53.
- 54.
- 55.↵
- 56.↵
- 57.
- 58.
- 59.
- 60.
- 61.
- 62.
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.
- 70.
- 71.
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.
- 82.↵
- 83.↵
- 84.↵
- 85.
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.
- 94.
- 95.
- 96.
- 97.
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
- 106.↵
- 107.↵
- 108.
- 109.↵
- 110.↵
- 111.↵
- 112.
- 113.
- 114.↵
- 115.↵
- 116.↵
- 117.↵
- 118.↵
- 119.↵