Bald sea urchin disease shifts the surface microbiome on purple sea urchins in an aquarium

Abstract Bald sea urchin disease (BSUD) is most likely a bacterial infection that occurs in a wide range of sea urchin species and causes the loss of surface appendages. The disease has a variety of additional symptoms, which may be the result of the many bacteria that are associated with BSUD. Previous studies have investigated causative agents of BSUD, however, there are few reports on the surface microbiome associated with the infection. Here, we report changes to the surface microbiome on purple sea urchins in a closed marine aquarium that contracted and then recovered from BSUD in addition to the microbiome of healthy sea urchins in a separate aquarium. 16S rRNA gene sequencing shows that microhabitats of different aquaria are characterized by different microbial compositions, and that diseased, recovered, and healthy sea urchins have distinct microbial compositions, which indicates that there is a correlation between microbial shifts and recovery from disease.


Introduction
Bald sea urchin disease (BSUD) is a bacterial infection that impacts many species of sea urchins and was first described in the r ed sea urc hin, Mesocentrotus franciscanus (Johnson 1971 ).Spine loss is the k e y c har acteristic of the disease, hence the gener al name of BSUD.Since the original c har acterization, ther e has been a wide range of descriptions of BSUD that has mostly been based on the pr esence of discr ete lesions on the surface of sea urchins and the loss of a ppenda ges within those lesions that may include test erosion and body wall perforation (Table 1 ).Howe v er, other symptoms include general spine loss in the presence or absence of surface lesions and disruption of the peristomial membrane.Sea urc hins ar e known to r ecov er completel y fr om BSUD or may be at risk of death depending on the se v erity of the infection (Lafferty et al. 2004 ).Declines in echinoid populations as a result of disease can have significant consequences in marine systems because they may result in phase shifts in ecosystems depending on the extent of the echinoid population reduction from a mass die-off and whether the infected species is a k e ystone member of an ecosystem (r e vie wed in Smith et al. 2022 ).For example, in the 1980s the near disa ppear ance of a k e ystone herbi vore, the longspined black sea urchin, Diadema antillarum , due to disease of unknown etiology resulted in a massive ecological phase shift in the Caribbean Sea from coral cover to uncontrolled algal growth on the reefs (Sammarco 1980, Vega Thurber et al. 2012, Smith et al. 2022 ).A repeat mass mortality event of D. antillarum beginning in 2022 in the Caribbean Sea shows general spine loss prior to death and the pathogen has been identified as a scuticociliate (Hewson et al. 2023 ).Ec hinoid diseases ar e also associated with substandard aquacultur e pr actices including suboptimal water temper atur e, eutr ophication, poor food quality, and injuries from handling and sorting (Tajima et al. 1998, Brink et al. 2019, Chi et al. 2022 ).Sea urchin disease results in economic losses for mariculture facilities and financial impacts on the mariculture industry (Wang et al. 2013a ).Consequently, the causes of diseases in echinoids and the possibilities of their pr e v ention ar e important to understand for both marine ecosystem structure and for successful mariculture.
Man y r eports of BSUD outbr eaks hav e focused on identifying the underlying causative agent and have indicated that the pathogen is unlikely to be fungal, blue-green algal, or viral, but is likely bacterial (Maes and Jangoux 1984a, Wang et al. 2013b, Bauer and Young 2000 ).A wide variety of bacteria are associated with BSUD, and many can reproduce the disease (Table 2 ).Given this v ariety, whic h has been correlated with a wide range of oceanic locations, and given the plethora of potentially pathogenic bacteria in marine systems, there may be a vast number of causative agents of BSUD.This variety may be the basis for the range of disease symptoms that are evident on the surface of sea urchins (Table 1 ).
Although the presence and characteristics of BSUD for sea urchins in the ocean or in aquaculture facilities have been reported (Table 1 ) including the associated bacterial pathogens (Table 2 ), there are no studies that describe the onset and r ecov ery of Table 1.Infections on the surface of sea urchins show a wide range of symptoms.

Disease
Symptoms Reference BSUD Spine loss without lesions This paper, Clemente et al. ( 2014 ), Brink et al. ( 2019 ) Black mouth disease or no name given Spine loss without lesions Chi et al. ( 2022 ), Tajima et al. ( 1988 ) Blackish peristomial membrane

BSUD
Lesions on the body surface showing green, pink, red, and purple discoloration Lafferty et al. ( 2004 ), Girard et al. ( 2011 ), Brink et al. ( 2019 ) Lesions of various colors Maes and Jangoux ( 1984a ,b ), Gilles and Pearse ( 1986 ), Roberts-Regan et al. ( 1988 ), Becker et al. ( 2008 ) BSUD in closed aquaria in the absence of experimentally induced infection.Her e, we r eport on purple sea urc hins, Strong ylocentrotus purpuratus , that were shipped from Southern California to Washington DC, and subsequently acquired BSUD after being housed in a recirculating marine aquarium.Symptoms included the loss of all primary spines, no discrete lesions as described in pr e vious reports (Clemente et al. 2014, Brink et al. 2019 ), and the animals r ecov er ed fr om the disease, r egr e w their primary spines, and r eturned to what a ppear ed to be a healthy state.Because housing was in a closed aquarium system, this controlled environment provided an opportunity to investigate the changes in the surface micr obiome comm unity associated with BSUD.Samples were col-lected from sea urchin surfaces during the disease, after recovery, and from healthy animals housed in a separate aquarium, plus aquarium seawater and used for sequencing the 16S rRNA gene amplicons.Results indicated that the surface microbiomes on the three groups of sea urchins had significantly different compositions suggesting a microbial shift during disease recovery, and a difference in the microbial community on sea urchins from different shipments housed in two different aquaria.We speculate that the changes in the local environment of the ocean vs. a closed aquarium system, and perhaps our standard antibiotic treatments undertaken for all newly shipped sea urchins , ma y have contributed to an altered community state of the micro-biomes .T he microbial shifts may have led to the onset of disease in the sea urchins with subsequent shifts associated with recovery.

Sea urchin husbandry
Purple sea ur chins, S. purpuratus , w ere collected from the nearshore water in San Diego CA, transferred to the Southern California Sea Urc hin Compan y (Cor ona del Mar, CA) and placed in holding tanks for 2 weeks in the open seawater system at the K erc khoff Marine Labor atory (California Institute of Tec hnology).Appar entl y healthy sea urchins ( n = 40) that survived collection w ere pur chased and shipped overnight to George Washington University in Washington DC.Sea urchins were housed in aquarium B (125 gallon) with recirculating ASW (Premium Marine Salt, OmegaSea), salinity of 32-35 ppt, 13-14 • C, and outfitted with both physical and biofilters, a UV light housing, and a protein skimmer.The central aquarium pump (Pond-Mag 9.5, Pondmaster) that was positioned in the aquarium sump, circulated 950 gallons/hour through the system.The water quality was maintained with w eekly seaw ater changes of 5 gallons that also served in solid waste r emov al.
Aquarium A (125 gallons) in our laboratory, held 40 healthy sea urchins that had been housed in that aquarium for 9 months, and for which no diseases had been observed.The salinity, temperature, filtering, and UV light housing were the same as for aquarium B, and the central pump circulated 1057 gallons/hour through the system (Marine DC pump DCT-4000, Jecod).All animals were fed weekly with rehydrated brown seaweed, Saccharina angustata (Kjellman) (WEL-PAC).

Treatment of sea urchins with penicillin and streptomycin
Based on our standard protocol, upon arrival all shipments of sea ur chins w ere treated by immersion for 1-2 hours at 14 • C in a tray (8 l) of fr eshl y pr epar ed ASW with 12 mg/l penicillin and 50 mg/l stre ptom ycin sulfate (pen/stre p).After treatment, different sea urchin shipments were placed in different aquaria.

Sample collection from sea urchin surfaces and from aquarium seawater
Samples were collected from randomly selected healthy sea urchins ( n = 4, H1-H4) in aquarium A and diseased sea urchins ( n = 4, D1-D4) in aquarium B, plus seawater samples from aquarium A ( n = 2, WH1 and WH2) and aquarium B ( n = 2, WD1 and WD2).Subsequent samples were collected from randomly selected r ecov er ed sea urc hins ( n = 4, R1-R4) in aquarium B as w ell as seaw ater samples ( n = 2, WR1 and WR2).The cellular material collected from the surface of sea urchins was retained on nylon filters (0.22 μm, 47 mm diameter; GVS Filter Technology).Filters were held on a filter-holder assembly composed of a 300-ml funnel (Fisherbrand) that was spring clamped (Millipore, item XX1004703) to a fritted glass support base (Millipore) with a rubber stopper to insert into the top of a 1-l sidearm flask connected to the building vacuum system.Each sea urchin was placed in the funnel or was held over the vacuum filtration apparatus and 500 ml of seawater from its respective aquarium was pour ed slowl y ov er all surfaces of the sea urc hin, suc h that the material washed from the animal plus cells in the seawater were collected on the filter.The filters were inserted into individual 50 ml falcon tubes and stored at −80 • C for later processing.After each collection, the filter apparatus and the funnel were rinsed with deionized water before collecting the sample from the next sea ur chin.Seaw ater (500 ml) fr om eac h aquarium was filter ed similarly to collect samples that served as controls.A sample of 500 ml of fr eshl y mixed Omega ASW contr ol was filter ed in the same manner.Samples from diseased sea urchins were collected on day 140 after arrival and housing aquarium B at a time when all sea urchins were exhibiting the disease symptom of primary spine loss.Samples from recovered sea urchins were collected 300 days after arrival and housing in aquarium B, at a time when all sea urchins had regenerated their primary spines and were exhibiting normal behavior.Samples collected from healthy sea urchins in aquarium A were collected on the same day as when the diseased sea urchin samples were collected.

Genomic DNA isolation from samples collected on filters
The genomic DN A (gDN A) isolation from nylon filters was carried out according to Turner et al. ( 2014 ) with modifications as tested with cultures of E. coli ( Supplementary Data File 1 ).Each filter was placed in a sterile plastic petri dish (10 cm diameter), cov er ed with 1 ml cetyltrimethylammonium bromide [CTAB; 2% CTAB, 100 mM Tris base (pH 7.4), 1.4 M NaCl, 1% pol yvin ylpyrr olidone, and 20 mM EDTA] and incubated for 10 minutes at 65 • C with constant shaking.The filters were removed from heat and a cell scr a per was used to r emov e all material from the filter, which was tr ansferr ed to a 1.5-ml tube.Chloroform:isoamyl alcohol (24:1; 1 ml) was added to the CTAB solution, mixed by inversion, and centrifuged at 20,800 × g for 2 minutes at r oom temper atur e .T he aqueous layer was tr ansferr ed to a new tube, NaCl was added to 2.5 M, follo w ed b y the addition of an equal volume of 100% isopropanol and mixed by in version.T he samples were chilled at −80 • C for 15 minutes and spun at 20,800 × g for 15 minutes at 4 • C. The supernatant was discarded, then the pellet and the inside of the tube were washed with 150 μl of 70% ethanol, followed by a wash with 150 μl of 90% ethanol and air drying.gDNA pellets wer e r esuspended in 10 μl Tris-EDTA buffer [TE; 10 mM Tris base (pH 7.4), 1 mM EDTA] and the concentr ation was e v aluated on a spectr ophotometer (NanoDr op 2000c , T hermoFisher).T he gDNA size and le v el of degr adation was e v aluated with a 0.75% a gar ose gel with Tris-acetate-EDTA buffer (TAE; 40 mM Tris; 20 mM acetic acid, 1 mM EDTA) plus 1% ethidium bromide and imaged with a UV imaging system (Kodak Molecular Imaging, Kodak Gel Logic 1500 Imaging System).

Pol ymer ase chain reactions
The 16S rRNA gene was amplified from the isolated gDN A b y PCR to test for the presence of prokaryotic DNA in the samples (27F, 5 AGA GTT TGA TCC TGG CTC AG; and 1492R, 5 ACG GTT ACC TTG TTA CGA CTT) that resulted in approximately a 1.5-kb amplicon (Weisburg et al. 1991 ).In total, two different concentrations of gDNA template (2 and 0.2 ng μl − 1 ) were used in PCR (Bio-Rad T100 Thermal Cycler) and the final reaction volume of 20 μl contained 1X Primestar Buffer, 200 μM dNTPs, 0.3 μM each primer, and 0.5 units PrimeSTAR GXL DNA pol ymer ase (Takar a).The PCR pr ogr am was 98 • C for 1 minute, follo w ed b y 30 c ycles of 98 • C for 10 seconds, 60 • C for 15 seconds, and 68 • C for 15 seconds, with a final extension of 68 • C for 1 minute and a 4 • C hold.The amplicons wer e anal yzed on a 0.8% a gar ose gel with TAE buffer plus ethidium bromide and imaged on an UV system as described abo ve .

16S rRNA amplicon sequencing
The gDNA samples were processed and sequenced using the Zy-moBIOMICS targeted sequencing service at ZymoResearch (Irvine CA).Targeted sequencing of the bacterial 16S rRNA gene was performed using the Quick -16S NGS Library Prep Kit (ZymoResearch) with custom-designed, proprietary primers to amplify the V3 and V4 region of the 16S rRNA gene .T he libr ary pr epar ation was completed by quantitativ e r eal-time PCR (qPCR) and quantification of the PCR products using qPCR fluorescence readings.Library fragment sizes were selected and optimized with the Select-a-Size DNA Clean & Concentr ator (ZymoResearc h) and quantified with Ta pestation (Agilent Tec hnologies) and Qubit (ThermoFisher Scientific).The positiv e contr ol sample used for libr ary pr epar ation was the ZymoBIOMICS Microbial Community DN A Standar d (Zy-moResearch).In addition to the ASW sample collected by DNA extr action fr om fr eshl y made Omega sea water (see abo ve), Zy-moResear ch emplo y ed a blank during libr ary pr epar ation, and both of these samples served as negative controls .T he completed library was sequenced using a V3 r ea gent kit with 600 cycles on Illumina MiSeq, which was calibrated by a 10% spike-in of PhiX DNA.T he ra w sequence reads were uploaded to the Sequence Read Arc hiv e database at NCBI under the BioProject ID PR-JNA851819.

Amplicon sequence analysis
D AD A2 (Callahan et al. 2016 ) was used to filter and trim sequences, infer amplicon sequence variants (ASVs) and r emov e sequencing errors and chimeric sequences.Taxonomy assignment was performed using SILVA release 138.1 (Quast et al. 2013 ).Phyloseq pac ka ge (v ersion 1.42.0)(McMur die et al. 2013 ) w as used to calculate alpha and beta diversity.Alpha diversity was estimated using the Observed Species, Chao1 (Chao 1984 ), and abundancebased cov er a ge estimator (ACE) (Chao and Lee 1992 ) indices.Beta diversity was analyzed using Bray-Curtis distances (Bray and Curtis 1957 ) and visualized with nonmetric multidimensional scaling (nMDS).
Statistical significance of differences among groups for alpha diversity indices was performed using one-way ANOVA ( P ≤ .05)and Tuk e y test in R. To e v aluate statistical significance among groups for beta diversity, a permutational multivariate analysis of variance (PERMANOVA, P ≤ .05)using distance matrices was performed with the adonis2 function (Permutations = 999) with the v egan pac ka ge (v ersion 2.6.4)(Oksanen et al. 2015 ).Taxonomic groups that had significant differences in abundance among differ ent gr oups wer e identified by Linear Discriminant Anal ysis Effect Size analysis (LEfSe) (Segata et al. 2011 ) using the micro-biomeMarker pac ka ge (v ersion 1.3.2) (Cao 2022 ).The Upset plot was generated using UpSetR (version 1.4.0)(Gehlenborg 2019 ) and ComplexUpset (version 1.3.3)(Lex et al. 2014, Krassowski 2020 ).Rar efaction curv es wer e gener ated using the Micr obiotaPr ocess pac ka ge (v ersion 1.6.6)(Xu and Yu 2022 ).The afor ementioned anal yses wer e performed using R version 4.1.1(R Core Team 2021 ).The R code with complete pipeline utilized can be found in the GitHub repository https:// github.com/chloeshaw8/ Bald-seaurc hin-disease _ pr oject .

Amplicon sequence analysis by Zymo Research
The sequenced reads were also evaluated by the 16S rRNA Amplicon Sequencing Data Inter pr etation service at ZymoResearch using their bioinformatics pipeline and analysis ( Supplementary Data File 3 ).Briefly, the D AD A2 pipeline was used to infer ASVs fr om r aw r eads, to r emov e err ors and c himeric sequences, and taxonomic assignment was performed using the internal curated ZymoResearch dataset.

Progression of BSUD shows a complete loss of primary spines follo w ed b y reco very and spine regrowth
Sea urchins shipped across the country generally arrive stressed and spawning, and the survival of different shipments of sea urchins can range from as much as 80% to as poor as no survivors.Consequently, since 2012 we have emplo y ed a standar d protocol for sea urchin care to improve survival.Sea urchins ( n = 40 per shipment) r eceiv ed fr om California wer e tr eated upon arriv al (day 0) by immersion in pen/strep for at least an hour at 15 • C, which was adapted from standard culturing conditions for embryos and larvae (Leahy 1987, Adams et al. 2019, Schuh et al. 2020 ).Because the pen/strep treatment resulted in higher survival after shipping, it was incor por ated for all sea urc hins since implementing this treatment.Although the polychete, Flabesymbios commensalis , has been noted on sea urchins in some shipments (unpublished observ ations), none wer e e vident on the animals in aquarium B, and no other ectoparasites were present (Becker et al. 2007, Virwani et al. 2021 ).On day 91 after arriv al, m ultiple sea urchins in aquarium B sho w ed unusual behavior of drooping spines (Fig. 1 A), although when disturbed by sound or touc h, they quic kl y r eoriented their spines to the normal position of pointing dir ectl y out from the spheroid body (Video 1; Supplementary Data File 2 , Video 1 Legend).On days 93-108, the sea ur chins w er e constantl y mo ving their spines , pointing them in differ ent dir ections r ather than the expected orientation of perpendicular from the body surface, which was another unusual behavior.By day 108, many sea urchins in aquarium B showed spine loss (Fig. 1 B), and the animals with the greatest spine loss were placed in floating plastic boxes to minimize their interaction with other sea ur chins.Ho we v er, isolation in boxes did not reduce or block the spread of the disease to all animals in the aquarium.In an effort to save the animals fr om pr ogr ession to death (whic h was assumed at the time), the w eekly seaw ater change w as incr eased fr om 5 to 10 gallons .Sea water chemistry for pH, ammonia, nitrate , nitrite , copper, and phosphorus for aquarium B and aquarium A, which housed healthy sea ur chins, w ere normal and deemed not to be the basis for the diseased sea urchins in aquarium B. On day 113, a 20-gallon seawater change was carried out and the sea urchins were treated with a second immersion in pen/strep for 2 hours at 15 • C. On day 116, the UV light bulbs for both aquaria A and B were changed to ensure that they were working optimally to sterilize the microbes in the seawater.By day 132, all sea urchins in aquarium B were diseased and had lost all primary spines, although the secondary spines (Fig. 1 C and D, y ello w arro w), pedicellariae, and tube feet remained intact.On day 132, sea urchins in aquarium B were treated a third time by pen/strep immersion, with the exception of nine animals that held tightly to the aquarium walls and could not be r emov ed without significant injury.The sea ur chins sho wing BSUD symptoms were difficult to grasp because their surfaces wer e abnormall y slimy, whic h was consistent with infection and tissue disintegration (Fig. 1 C and D).Throughout the infection, some sea urchins held tightly to the walls of the aquarium yet also sho w ed the opposite behavior of failing to hold kelp secur el y during feeding.Notably, all animals with BSUD fed throughout the disease and none sho w ed discrete lesions as described in other reports of spotting disease and some reports of BSUD (Table 1 ).From Figure 1.BSUD pr ogr ession and r ecov ery in the pur ple sea urc hin.(A) An earl y symptom of BSUD is dr ooping spines suggesting that the sea urc hin may be sleeping as has been observed on rare occasions for healthy sea urchins.Alternatively, the surface infection that would impact the muscles associated with the base of the spines and the tubercles may have altered spine movement (see Video 1; Supplementary Data File 2 , Video 1 Legend).(B) A sea urchin infected with BSUD that has lost many of its primary spines.(C) A sea urchin with BSUD that has lost all of its primary spines and shows the red tubercles to which primary spines are normally attached.Shorter and smaller secondary spines remain attached.(D) A magnified image of the surface of a sea urchin with BSUD that has lost all primary spines .T he secondary spines (y ello w arro w), pedicellariae that are too small to see in this image, and tube feet remain on the animal surface .T he beginnings of newly growing spines (white arrow) are present on some tubercles.(E) A r ecov ering sea urchin with newly regrowing primary spines that are short, pointed, and light purple.(F) A healthy sea urchin after full recovery from BSUD with the c har acteristic of rigid primary spines oriented perpendicular to the animal body.days 154 to 246, all sea urchins began to r ecov er fr om the disease as evidenced by regrowing their primary spines (Fig. 1 D, white arr ow, and E) and r eturned to what a ppear ed to be a healthy state (Fig. 1 F).

gDN A isola ted from the surface of sea urchins includes bacterial DNA
Because the symptoms of BSUD were restricted to the external surface of the sea ur chins, w e reasoned that the surface microbiome was involved in disease onset.All sea urchins were fed the same diet, ther efor e, we also reasoned that the gut microbiome was not involved in BSUD.Furthermore, because a limited number of sea urchins can be deliv er ed per shipment, we opted to save as many animals as possible and to investigate the surface micr obes r ather than sacrificing animals for tissue dissection.This also r emov ed the pr oblem of contamination of tissue samples with microbes released from the gut during dissection.When e v aluating the gDNA (larger than 10 kb), more was isolated from the surface of the sea urchins than from the seawater samples ( Supplementary Data File 2 ; Table S1 , Supporting Information ) although it was likely that the gDNA was a mixtur e fr om both sea urchin cells and surface fauna ( Supplementary Data File 2 ; Figure S1A , Supporting Information ).To test for the presence of bacterial DNA in the samples prior to sequencing, we e v aluated the gDNA by PCR amplification of the 16S rRNA gene using our primers (see the section "Pol ymer ase c hain r eactions").Amplification of the 16S rRNA gene using the gDNA as the template (0.2 and 2.0 ng) resulted in the expected amplicons of 1.5 kb ( Supplementary Data File 2 ; Figure S1B -E , Supporting Information ).This indicated that bacterial gDN A w as present in all samples in sufficient quantities to support subsequent microbiome analysis.

Sufficient sampling depth is reached and unique ASVs are identified
Sequencing for all gDNA samples resulted in a total of 283,297 raw sequence reads following filtering and processing ( Supplementary Data File 2 ; Table S2 , Supporting Information ).A total of 41,217 ASVs were inferred and after taxonomic assignment, 31 phyla were identified ( Supplementary Data File 2 ; Table S3 , Supporting Information ).Rar efaction curv es r eac hed plateaus indicating sufficient sampling depth had been ac hie v ed for all samples ( Supplementary Data File 2; Figure S2, Supporting Information ).The sequenced reads were also analyzed by the ZymoResearch 16S Amplicon Sequencing Data Inter pr etation service using their own bioinformatics pipeline and analysis ( Supplementary Data File 3 ).We chose to compare the ZymoResearch pipeline results, which utilizes their own internal database, to our results .T he rarefaction curves for the two pipelines were similar ( Supplementary Data File 2 ; Figure S2 , Supporting Information ; Supplementary Data File 3 ; Figure S1 and Table S1 , Supporting Information ), ho w e v er, the filtering and processing of the reads resulted in the identification of 25 phyla for the ZymoResearch pipeline ( Supplementary Data File 3 ; Table S2 , Supporting Information ).

The microbiomes identified for each group are distinct
T hr ee distinct microbiome compositions were identified for the three groups of sea ur c hins The surface microbiomes sampled for the diseased, r ecov er ed, and healthy sea urchin groups each had many unique ASVs (Fig. 2 ) e v en though there were nearly 5000 shared ASVs .T he reco vered and healthy gr oups shar ed the most ASVs, whereas the diseased and healthy groups had the fewest number of shared ASVs .T he diversity of the microbiomes was evaluated for alpha and beta div ersity using se v er al metrics (Fig. 3 ).The surface microbiome samples of the diseased sea urchins had decreased alpha div ersity compar ed to the surface micr obiome samples fr om the healthy and r ecov er ed sea urc hins based on r esults fr om all metrics (Fig. 3 A-C).Analysis by Chao1 and ACE identified significant differences in the alpha diversity of the surface microbiomes for the diseased compared to the recovered groups, but not between the diseased and healthy groups (ANOVA, P < .05).The microbiome samples from the recovered group had the greatest alpha diversity based on all indices, ho w ever, the samples sho w ed no significant differ ences compar ed to the micr obiome samples fr om healthy group.Although the ZymoResearch pipeline did not identify any significant differences in the alpha diversity among the gr oups, r esults wer e similar to our analysis .T he Chao1 and Shannon metrics sho w ed that the microbiomes on the sea urchins in the diseased group had lo w er alpha diversity based on compared to the r ecov er ed and healthy groups ( Supplementary Data File 3 ; Figure S2 , Supporting Information ).Beta diversity based on Bray-Curtis distances r e v ealed clustering of the samples from within eac h gr oup and sho w ed minimal ov erla p of clusters (Fig. 3 D).This indicated that each microbiome sample within a group had a similar microbial composition and that the different groups (diseased, r ecov er ed, and healthy) had significantly different microbial compositions (PERMANOVA, P < .05).A nearly identical result was obtained from the ZymoResearch pipeline that sho w ed significantl y differ ent micr obial compositions among the gr oups ( Supplementary Data File 3 ; Figure S3 , Supporting Information ).

Man y ph yla show shifts in abundance as the sea ur c hins r ecover fr om BSUD
To e v aluate the taxa in the micr obiomes fr om the thr ee gr oups of sea urchins, the phyla with an av er a ge r elativ e abundance of > 0.1% across all samples were selected for comparisons among the thr ee gr oups ( Supplementary Data File 2; Tables S3  and S4, Supporting Information ).Results sho w ed that all micr obiomes wer e similarl y dominated by Pr oteobacteria and Bacteroidota (Fig. 4 ).The most abundant phyla sho w ed minor differences among the three groups (Fig. 4 A), and these differences wer e e vident for eac h of the samples from the sea urchins within eac h gr oup (Fig. 4 B).A few phyla had different abundances among the gr oups, suc h as Desulfobacter ota, Spir oc haetota, and Firmicutes that sho w ed reduced abundances in the microbiome samples of the diseased group compared to samples from the recovered and healthy group microbiomes .T he ZymoResearch pipeline similarly sho w ed that all microbiomes w ere dominated b y Proteobacteria and Bacteroidota, and that there were a few phyla that differed in abundance among the groups, including Spirochaetota, Firmicutes, and Verrucomicrobia ( Supplementary Data File 3; Figure S4 and Table S2, Supporting Information ).Because these c hanges wer e e vident at the le v el of phylum, this indicated significant changes to the microbiomes .Furthermore , differences in the abundances of certain phyla suggested that the microbiomes on the diseased sea urchins were altered and that changes to the micr obiomes occurr ed as the sea urc hins r ecov er ed fr om the BSUD infection.

T he micr obiomes on the thr ee gr oups of sea ur c hins show differences in microbial composition
Taxa identified in the microbiomes were also evaluated at the genus le v el and those with an av er a ge r elativ e abundance of > 1% for at least one sample across all groups were selected and e v aluated for their r elativ e abundance per sample (Fig. 5 ; Supplementary Data File 2; Table S5, Supporting Information ), which included the seawater samples for comparison ( Supplementary Data File 2; Table S6, Supporting Information ).Major differences in the compositions of the microbial genera were evident among the three groups (Fig. 5 A) and differ ences wer e also e vident for all samples collected within the groups of sea urchins (Fig. 5 B).Psychromonas and Vibrio had similar abundances in all three groups, ho w ever, the majority of the most abundant genera had abundances that differed among the thr ee gr oups (see the GitHub r epository for the complete table).For example, the genera Colwellia , Leucothrix , and a genus fr om the famil y Erwiniaceae wer e highl y abundant in the diseased group microbiomes compared to the recovered and healthy gr oup micr obiomes.Alternativ el y, Desulfotalea and a genus of the family Marinifilaceae were lo w er in abundance in the diseased group microbiome compared to the recovered and healthy gr oup micr obiomes.Similarl y, the ZymoResearc h pipeline sho w ed that Psychromonas and Vibrio had similar abundances among all thr ee gr oups, but that man y other gener a differ ed in abundance, such as Colwellia , Leucothrix , and Erwinia , which were elevated in the diseased gr oup micr obiomes ( Supplementary Data File 3; Figure S5 and Table S3, Supporting Information ).These results suggested distinct microbial compositions of the microbiomes for the three groups of sea urchins, which was in agreement with significant differences identified for beta diversity (Fig. 3 D).Figur e 3. T he surface microbiomes of sea urchins with BSUD have decreased alpha diversity compared to both recovered and healthy sea urchins.Alpha diversity was analyzed by (A) the Observed Species index, (B) the Chao1 index, and (C) the ACE index.Groups include microbiomes collected from the sea urchins and microbes in the aquarium sea water.T he box plots show the av er a ge and quartile values for each group.The Chao1 index and the ACE index show significant differences between the diseased and r ecov er ed gr oups (ANOVA, P < .05).Error bars in the Chao1 index indicate the result as an estimate of diversity.(D) The microbial compositions of the surface microbiome are different among the three groups of sea urchins.Bray-Curtis distances estimates of beta diversity shows distinct differences for the bacterial composition and membership of ASVs in the microbiomes among the three groups of sea urchins.Ellipses show the 95% confidence intervals for the samples collected from sea urchins in each group.

Major shifts to the microbial composition are associated with r ecover y fr om BSUD
Because the microbiome samples collected for the diseased and r ecov er ed gr oups of sea urc hins wer e obtained at different times from the same animals housed in aquarium B, changes in the taxa on their surface microbiomes could be compared directly to identify taxa associated with BSUD.Notably, the surface microbiomes of the diseased group had ele v ated abundances of Colwellia , Erwiniaceae , Leucothrix , Lutibacter , and NS10 marine group, compared to the microbiomes on the sea urchins after recovery (Fig. 5 ).Alternativ el y, the r ecov er ed gr oup had ele v ated abundances of Alteromonadaceae , Bacteroida, Desulfotalea , Gammaproteobacteria, and Marinifilaceae compared to the diseased group samples .T he ZymoResearc h anal ysis identified man y similar gener a as ele v ated in the micr obiomes of diseased sea urc hins, including Erwinia , Leucothrix , Cowellia , and Lutibacter.Furthermore, many similar taxa were also identified as ele v ated in the recov-er ed gr oup including Desulfotalea and Alteromonas ( Supplementary Data File 3 ; Figures S5 , S6 , and Table S3 , Supporting Information ).In contrast to our pipeline, the ZymoResearch pipeline assigned matches to species for many taxa, which identified species within gener a that wer e either ele v ated or r educed in the diseased gr oup micr obiomes ( Supplementary Data File 3 ; Figur es S7 and S8 , Supporting Information ).For example, species Colwellia meonggei , Leucothrix mucor , and Erwinia rhapontici wer e ele v ated in the diseased group microbiomes ( Supplementary Data File 3 ; Figure S7A -C , Supporting Information ).These taxonomic differences suggested that as the sea urc hins r ecov er ed fr om BSUD, the species membership on their surface microbiomes underwent major shifts in composition.

Different aquaria show distinct microbial compositions
Microbiome samples collected from recovered and healthy sea ur chins w ere both from noninfected sea urchins but were housed Figur e 4. T he surface microbiomes are dominated by Proteobacteria and Bacteroidota.The phyla shown have an av er a ge r elativ e abundance of > 0.1% for at least one sample across the groups ( Supplementary Data File 2; Table S3, Supporting Information ).The taxa are shown as (A) the average r elativ e abundance in the microbiomes for each group, and (B) the relative abundance for samples collected from each sea urchin in the three groups as well as the water samples.Samples from sea urchins are indicated in (B) as diseased (D), recovered (R), and healthy (H), and the numbers (1-4) correlate with the four sea urchins within each group.Water samples are indicated with a (W).
Figur e 5. T he most abundant genera in the microbiomes differ among the three groups of sea urchins .T he genera shown have an average relative abundance of > 1% for at least one sample ( Supplementary Data File 2; Table S5, Supporting Information ).The results show (A) the av er a ge r elativ e abundance of microbial taxa from all sea urchins in each sample group and (B) the relative abundance of taxa for samples collected from each sea urchin in each of the three sample groups as well as the water samples.Taxa are listed to the right, and ASVs that could not be assigned a genus are indicated as the next matching taxonomic level.Samples from sea urchins are indicated in (B) as diseased (D), recovered (R), and healthy (H), and the numbers (1-4) correlate with the four sea urchins within each group.Water samples are indicated with a (W). in differ ent aquaria.Ther efor e, a comparison among the microbial compositions for these two gr oups wer e used to identify differences attributed to different shipments of sea urchins and to housing in separate aquarium en vironments .T he genera in the microbiomes of the recovered sea urchins in aquarium B had ele v ated Alteromonadaceae , Gamma pr oteobacteria, Ruegeria , and Spongiibacteraceae compared to the microbiomes of the healthy sea urchins in aquarium A (Fig. 5 ).In contrast, the healthy sea urchin microbiomes had elevated Bacteroidetes BD2-2, Desulfotalea , Leucothrix, Marinifilaceae , and Roseimarinus compared to the microbiomes from the recovered group.Results from the ZymoResearch pipeline identified many taxonomic differences between the re-cov er ed and healthy group microbiomes, but the results differed from the taxa identified by our pipeline ( Supplementary Data File 3; Figures S5, S6, and Table S3, Supporting Information ).For example, taxa ele v ated in the r ecov er ed gr oup micr obiomes wer e Alteromonas , Sulfurimonas , and Cobetia , and taxa ele v ated in the healthy group microbiomes were Parvularcula , Pseudoalteromonas , and Neiella (Supplementary Data File 3; Figure S6, Supporting Information ).Nonetheless, both e v aluations of the taxonomic differences indicated distinct microbial compositions of sea urchins fr om differ ent shipments housed in differ ent aquaria.

The major differences in microbial composition based on the most abundant taxa are demonstrated for the three groups of sea ur c hins
In addition to e v aluating the most abundant taxa, those that sho w ed significant differences among samples were evaluated using LEfSe ( Supplementary Data File 2; Table S7, Supporting Information ).The most abundant taxa in the microbiomes from the three groups of sea ur chins w ere generally different (Fig. 6 ).The differ entiall y abundant taxa in the microbiomes of the diseased gr oup r elativ e to the other two gr oups wer e also some of the most abundant taxa.Ther efor e, these taxa may be k e y to the composition of the microbiomes on sea urchins with BSUD and included ASVs of the genera Colwellia , Leucothrix , Roseimarinus , and the families Spongiibactereaceae and Rhodobacteraceae .There were also differ entiall y abundant taxa that were not the taxa of highest abundance , which ma y still be k e y to the composition of the micr obiomes.Notabl y, families fr om whic h m ultiple ASVs wer e identified as differ entiall y abundant in the diseased gr oup micr obiomes included Saprospiraceae , of which two ASVs were within the genus Aureispira .Furthermor e, fiv e ASVs were identified in the family Roseobacteraceae , of which two were within the genus Haloc ynthiibacter , and there w ere three ASVs in the family Flavobacteriaceae .Results from LEfSe in the ZymoResearch pipeline identified many similar differentially abundant taxa in the diseased gr oup micr obiomes, including C. meonggei , Lutibacter agarilyticus , L. mucor , and a genus of the family Saprospiraceae , among others ( Supplementary Data File 3; Tables S4 and S5, Supporting Information ).Man y taxa wer e also differ entiall y abundant in the r ecov er ed and healthy gr oup micr obiomes (Fig. 6 ).Notabl y, the r ecov er ed gr oup micr obiomes had m ultiple ASVs of the gener a Clostridia and Desulfovibrio , whereas ASVs of the genus Desulfotalea , Marinifilaceae , and Pseudahrensia wer e differ entiall y abundant in the microbiomes of the healthy gr oup.The r esults of the LEfSe analysis in the ZymoResearch pipeline ov erla pped with the results of our pipeline and included Desulfovibrio and Coxiella as differ entiall y abundant in the r ecov er ed gr oup micr obiomes, and Desulfotalea , Shewanella , and Draconicbacterium in the healthy gr oup micr obiomes ( Supplementary Data File 3; Tables S4 and S5, Supporting Information ).These differ entiall y abundant taxa identified by LEfSe for both pipelines in the microbiomes for the recovered and healthy groups defined the differences between the two shipments of sea urchins that were housed in different aquaria.The differences among the microbiomes of the three groups were in a gr eement with the beta div ersity r esults (Fig. 3 ) and identified the set of k e y taxa underlying the differences in microbial composition.Ov er all, differ ential abundances of taxa sho w ed that as sea urc hins tr ansitioned fr om infection to r ecov ery, their micr obiomes underwent major shifts in bacterial composition.This was in a gr eement with the e v aluation of taxa that wer e of the gr eatest abundance .Furthermore , differential abundances were also evi-dent for sea urchins from different shipments housed in different aquaria.

T he char acteristics of BSUD are highl y v ariable
The BSUD outbreak in aquarium B sho w ed consistent symptoms for e v ery sea urc hin, including unusual primary spine mov ement and positioning before their loss over the entire surface of the animals, in addition to abnormal tube foot behavior and damage to the epidermal tissue.Disease r ecov ery was evident based on regrowth of the primary spines for all animals.Because the symptoms were observed more than 90 days after the arrival of the sea urchins, the basis for the disease onset is not understood but ma y ha v e been a combination of se v er al possible factors.At least some of the microbes that arrived with the sea urchins housed in aquarium B may have been involved in the disease onset, in addition to unknown effects of the pen/strep treatments on the microbiomes, a shift in the surface microbiome that occurs when sea urc hins ar e mov ed fr om the ocean to a closed aquarium (Wessel et al. 2022 ), and perhaps the microbial environment of aquarium B in which they were housed.The recovery from BSUD infers that the skeletogenic cells, which are involved in the regeneration pr ocess, ar e either not affected by BSUD, or that these cells migr ate fr om noninfected r egions of the body including the subdermal test onto the spine tubercles to function in rebuilding the primary spines (Heatfield 1970, 1971, Heatfield and Travis 1975, Märkel and Röser 1983, Dubois and Ameye 2001, Politi et al. 2004, Reinardy et al. 2015, Emerson et al. 2017 ).BSUD symptoms reported here differ somewhat from most other reports of BSUD, spotting disease, and other similar diseases in echinoids (Table 1 ).The majority of reports describe spine loss from discrete lesions as the most common c har acteristic of the disease (e.g .Maes and J angoux 1984b ), ho w e v er, ther e ar e a fe w r eports of BSUD showing general spine loss in the absence of lesions (Clemente et al. 2014, Brink et al. 2019 , this paper).The wide range of BSUD symptoms suggests significant variation in the pathogenicity of the infection, which infers that many different microbes and/or opportunistic bacteria may underlie this variable pathology and disease se v erity.The micr obial composition of BSUD r eported her e may not have been as pathogenic as in other reports given that the sea urc hins r ecov er ed fr om the disease in our facility, wher eas BSUD has been documented to cause mass mortalities in marine environments (Pearse et al. 1977, Boudouresque et al. 1980, Azzolina et al. 1985 ). Variation in the bacterial composition of diseased sea urchins based on sequence data is not unusual and has been reported pr e viousl y in whic h sea urc hins fr om differ ent locations in the ocean show distinct bacterial compositions associated with discrete lesions (Becker et al. 2008 ).This may be a basis for the differences in the microbiomes reported here for the two shipments of animals housed in different aquaria.This concept is further supported by the taxa identified in the diseased microbiome for the sea urchins in aquarium B, which differ greatly from the taxa identified in other reports of BSUD (Gilles and P ear ce 1986, Becker et al. 2008, Brink et al. 2019 ).Our results suggest that the bacterial composition associated with BSUD can include a variety of bacterial species that can differ based on the location of the sea urchins, suggesting that many combinations of bacteria may cause symptoms consistent with BSUD.Ov er all, the v ariation in the symptoms described in reports of BSUD suggest that differences in the microbiomes and/or the types of opportunistic bacteria may impact the le v el of symptoms and se v erity of the Figure 6.Man y micr obial taxa ar e significantl y differ entiall y abundant among the gr oups.A heatma p shows the r elativ e abundance of the taxa in each sample within the three groups including the water samples.Identified taxa are ASVs and the lo w est kno wn taxonomic classification is listed.Taxonomic names that are colored have a notably reduced abundance or are completely absent in the other two group microbiomes .T he taxa displa yed ha v e an LDA scor e of > 3.1 based on identification by LEfSe.All taxa with an LDA scor e of > 3.0 along with their LDA scor e and P -v alues ar e available in Supplementary Data File 2; Table S7 (Supporting Information ).The threshold was increased for better visualization of the data.
disease.Based on results presented here, we suggest that the observed symptoms, including spine loss and surface discoloration ar e gener al indicators of man y differ ent sea urc hin diseases that fall within the category of BSUD.

Many factors may contribute to an altered microbial community
Our findings for S. purpuratus indicate that the microbiome on diseased sea urchins is distinct from that on nondiseased sea urc hins, and that r ecov ery fr om disease is associated with significant changes in the bacterial composition of the surface micr obiome.Furthermor e, the similarities between the microbiome compositions on the sea urchins compared to that in the aquarium seawater are maintained over time as the sea urchins recover ed fr om BSUD.Because both aquaria are equipped with UV lamp housings that function to eliminate live microbes from the seawater, this suggests that the animals release bacteria from their surfaces that impact the aquarium micr obiome r ather than the r e v erse.Based on our practices of sea urchin care in which each shipment of sea urchins is housed in different aquaria, the composition of the microbiome of the sea urchins that arrive in a given shipment may determine or gr eatl y influence the microbial composition of the aquarium.Clearly, this variation among sea urchins in different aquaria is maintained to some extent e v en though the micr obiome under goes major changes when sea urc hins ar e tr ansferr ed fr om their natur al envir onment to an aquarium.For example, less than 10% of ASVs identified in the microbiome of spines are retained on the variegated sea urchin, Lytec hinus v ariegatus , when they ar e mov ed fr om Bisca yne Ba y, Florida to a labor atory envir onment (Wessel et al. 2022 ).Although the pen/str ep tr eatments that all sea urc hins r eceiv e upon arriv al may also alter the surface microbiome, this treatment improves survival after shipping and has been reported to impr ov e surviv al of a lesion syndrome disease similar to BSUD in Strongylocentrotus intermedius (Wang et al. 2013a ).It is noteworthy that the animals housed in aquarium B that acquired BSUD have been the only shipment to our laboratory in which all sea urchins contracted a disease .T his ma y infer that these particular sea ur chins w ere different in some way, possibly in the composition of the microbiome of one or more of the animals, that resulted in all animals contracting BSUD.

Biomark ers rev eal microbial associations with disease and health
Culturing marine bacteria poses challenges because few microbes in a sample will grow on marine media, thereby making it difficult to identify causativ e a gents of disease.Her e, we opted to c har acterize the microbiome associated with BSUD using high throughput sequencing to capture the breadth of microbes associated with the disease.Because bacteria were not cultured, we did not identify a causative agent through reinfection experiments and addressing Koch's P ostulates .Furthermore , the sequence data did not identify a k e y underlying pathogen of BSUD, but instead sho w ed that ther e ar e man y taxa with elevated abundance or taxa that are significantly differentially abundant in the microbiomes of the diseased sea urchins.Notable taxa that we identified include Colwellia , Leucothrix , Aureispira , Spongiibacter aceae , Roseobacter aceae , Rickettsiaceae , Rhodobacter aceae , and Flavobacteriaceae , which may be involved in BSUD progression based on their ele v ated r elativ e abundances in the microbiomes of the infected sea urchins .T he analysis of the same sequences b y the ZymoResear c h pipeline also identified man y of the same taxa as associated with BSUD, particularly Colwellia and Leucothrix , whic h str engthens the likelihood of their association with BSUD.Ho w e v er, incr eased abundances of certain taxa may not necessarily be pathogenic, and instead may be the outcome of an alter ed host-micr obe inter action (Faust et al. 2015 ).Alternativ el y, it is also possible that some taxa with low abundance that are also differ entiall y abundant in the diseased gr oup micr obiomes may be involved in BSUD progression, such as species of the genera Aureispira , Lutimonas , and Fusibacter (Fig. 6 ).Ho w e v er, our assumption here is that taxa involved in BSUD hav e ele v ated abundance.
Because ther e ar e man y taxa with incr eased r elativ e abundance in the infected sea urchin microbiome samples and there is no dominant taxon identified by either pipeline for these samples, it is feasible that a subset of these bacteria act collectiv el y to cause the BSUD symptoms that we observed.This is conceptually akin to microbial dysbiosis (Sweet et al. 2020 ).
The taxa identified here as associated with BSUD mostly differ from a previous report that analyzed the microbiome of dissected tissues of the collector sea urchin, Tripneustes gratilla , that showed a BSUD symptom of primary spine loss (Brink et al. 2019 ).Howe v er, ther e ar e major differ ences in the taxa that show incr eased abundance in the microbiomes of infected T. gratilla, which included Agarivorans , Arcobacter , Loktanella , and Leisingera , among others that were not identified for infected S. purpuratus .Alternativ el y, Leucothrix and Rhodobacteraceae are associated with diseased S. purpuratus and T. gratilla , which suggests that these taxa either contribute to the disease symptoms or are common opportunists.Furthermor e, a causativ e a gent underl ying BSUD in T. gratilla was not identified, leading to the similar conclusion that a combination of microbial agents underlie BSUD symptoms .Nonetheless , the differences between the taxa that we report here for S. purpuratus and the report for T. gratilla (Brink et al. 2020) infer that different compositions of microbes may cause similar disease symptoms, e v en in different sea urchin species.
This analysis of BSUD on S. purpuratus based on results from both pipelines demonstrated that the majority of the taxa as-sociated with the microbiomes on the diseased sea urchins ar e Gr am negativ e. Ho w e v er, it was unexpected that Vibrio had a similar r elativ e abundance acr oss all thr ee gr oups (Fig. 5 ; Supplementary Data File 3; Figure S7I, Supporting Information ).This result is inconsistent with multiple reports of direct associations between Vibrio and sea urchin disease (Gilles and Pearse 1986, Li et al. 2000, Wang et al. 2005, 2013a, Ho et al. 2016, Hira and Stensvåg 2022 ) and suggests that Vibrio may not be an essential pathogen that causes BSUD in this study.These findings further support the notion that variations in the microbial composition of the microbiome underlies variations in BSUD symptoms, inferring that BSUD infections may be fundamentally different in different species of sea urchins and also in different populations of the same species of sea urchins located in different regions of the oceans .T his notion is supported b y the differences betw een the microbiomes of the r ecov er ed and healthy sea urc hins, whic h are housed in different aquaria, and have distinct microbial compositions despite being the same species of sea urchin.This may infer that the local microbial marine environment plus the host interaction with the surface microbiome both act to shape the microbial composition on the echinoid surface.
Taxa that are reduced in abundance in the diseased group samples and are elevated in the recovered group samples may not be involved in causing BSUD, but instead may be indicators of a healthy microbiome on the nondiseased sea urchins.Some taxa that are reduced in abundance in the diseased group samples may be outcompeted by other taxa that contribute to the disease symptoms.Bacteroidetes BD-2, Desulfotalea , and Marinifilaceae may be the best indicators of a healthy microbiome because they ar e highl y abundant on both r ecov er ed and healthy sea urchins from both aquaria (Fig. 5 ).The ZymoResearch pipeline identified similar taxa as reduced in the diseased group microbiome compared to the recovered and healthy groups, such as Desulfotalea , Alteromonas , and Sulfurimonas (Supplementary Data File 3; Figure S5, Supporting Information).If these taxa are normally elevated in the surface microbiome when sea urc hins ar e in a nondiseased state, decreases in these taxa may be indicators that sea urchins ar e under going dysbiosis, whic h may be a preliminary condition that leads to disease.Monitoring the taxa of the surface microbiome in aquaria or in natural environments could be used for predicting disease onset.
Recov ery fr om infectious bacterial diseases in any animal infers the involvement of the immune system.Although we did not investigate the complexities of immune functions in the sea ur chins, w e suggest that in addition to responding to internal infections, the immune system also acts to regulate host-microbe interactions on the animal surface.For example, CRISPRCas9 gene knoc k out in sea urchins that pr e v ents the pr oduction of se v er al na phthoquinones, including ec hinoc hr ome A and se v er al spinoc hr omes, alters the microbiome of spines and decreases survival of the sea urchins (Yaguchi et al. 2020, Wessel et al. 2022 ).Na phthoquinone knoc k-outs also r esult in the failur e of some larv ae to surviv e in a micr obial envir onment (Wessel et al. 2020 ).The antimicr obial pr operties of ec hinoc hr ome A (Service and Wardlaw 1984, Lebedev et al. 2005, Coates et al. 2018 ) in red spherule cells of adult coelomocytes (Coates et al. 2018 ; r e vie wed in Smith et al. 2018 ) and in pigment cells in larvae (Ho et al. 2016, Buckley et al. 2017 ) is consistent with their immune functions and responses to bacterial contact, infections, and injuries (Johnson 1969, Heatfield and Travis 1975, Coffaro and Hinegardner 1977, Allen et al. 2022 ).T hus , the na phthoquinone knoc k-outs in adult sea urc hins infer the presence of a surface immune system and when it is altered, the surface microbiome changes and may result in dysbiosis .T his concept of host-microbe regulation in aquatic organisms is consistent with functions of mucosal innate immunity in aquatic vertebrates (Gomez et al. 2013, Varga et al. 2019 ).

Conclusions
Here, we describe sea urchins housed in a single aquarium that all contracted BSUD, which is characterized as a surface infection with the loss of all primary spines, in the absence of discrete lesions.Based on the results from two different analytical pipelines, we identify distinct microbial compositions of the surface microbiomes among diseased, recovered, and healthy sea urc hins, whic h show a correlation between r ecov ery fr om disease and changes to the surface microbiome of sea urchins.We find that the surface microbiome composition can differ among individual animals or among different populations .T his suggests that BSUD may be caused by a variety of combinations of bacteria based on both the local microbial environment and the interactions with the sea urchin surface tissues.Results are consistent with our inability, and that of others, to identify a pathogen or a group of microbes that are consistently associated with BSUD, and infers that many BSUD infections, or e v en eac h infection, may be unique.

Figur e 2 .
Figur e 2. T her e ar e man y ASVs identified in the micr obiomes for eac h group of sea urchins.Upset plot shows that the microbiomes for the diseased, r ecov er ed, and healthy groups have many ASVs that are unique.Each bar shows the ASV count and the group in which the ASVs were identified is denoted directly below each bar.Shared ASVs are indicated by the line connecting groups below the bar indicating number of ASVs counted.Of the total number of ASVs ( 40,800), 4,878 (12%) ASVs are shared among the three groups, and the diseased and healthy groups have the fewest number of shared ASVs (5.1%).

Table 2 .
Bacteria associated with BSUD and related surface diseases.