Selective conservation of symbiont cell-surface glycans across generations in a vertically transmitting coral

19 Coral resilience under climate change depends on the stability of coral–Symbiodiniaceae 20 symbioses. While verJcally transmi[ng corals inherit symbionts directly from parental colonies, 21 the extent to which symbiont cellular traits are conserved across life stages remains unclear. 22 Here, we examined cell-surface glycan profiles of Symbiodiniaceae in parental colonies and eggs 23 of the coral Mon$pora capitata. Glycan signatures were structured by symbiont genus and 24 differed between life stages, with mannose/glucose- and galactose-containing glycoproteins as 25 primary drivers of variaJon. Despite life-stage differences, parent–offspring comparisons 26 revealed significant conservaJon of glycan profiles, indicaJng intergeneraJonal transmission of 27 symbiont cellular traits that differed between Cladocopium and Durusdinium and were driven 28 by disJnct glycan classes. These results suggest that verJcal transmission preserves key 29 recogniJon -relevant glycans while allowing flexibility in other symbionts’ surface traits, 30 providing a mechanisJc basis for symbiosis stability. 31 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint 32 Introduc%on 33 Reef-building corals form an obligate endosymbiosis with dinoflagellates of the family 34 Symbiodiniaceae, a partnership that underpins the producJvity and persistence of coral reef 35 ecosystems (1). Symbiodiniaceae provide photosyntheJcally derived carbon to the host, 36 supporJng metabolism, calcificaJon, and reef growth, while receiving inorganic nutrients and a 37 protected intracellular niche. This metabolic coupling represents one of the most ecologically 38 successful symbioses in the ocean and has persisted for hundreds of millions of years, shaping 39 the evoluJon and biogeography of coral reefs (2,3). 40 Despite this long evoluJonary history, the coral –Symbiodiniaceae symbiosis is highly sensiJve to 41 environmental change. Elevated sea surface temperatures destabilize this associaJon, causing 42 bleaching through impaired photosynthesis, nutriJonal decoupling (4), and reacJve oxygen –43 mediated symbiont loss (5,6). Recurrent marine heatwaves have increased the frequency and 44 severity of mass bleaching, driving coral mortality, declines in coral cover, and community shijs. 45 As a result, a substanJal proporJon of reefs ha ve already been degraded, with further losses 46 projected under conJnued warming (7). 47 Considering Symbiodiniaceae are indispensable to coral fitness, the mode by which corals 48 acquire their symbionts is a key determinant of coral biology. Symbiont transmission occurs 49 along a conJnuum between horizontal acquisiJon from the environment and verJcal 50 inheritance from the parent colony, with important consequences for host–symbiont specificity, 51 flexibility, and resilience (8). Horizontally transmi[ng corals acquire symbionts from the 52 surrounding seawater, enabling uptake of locally adapted partners with reduced fidelity. In 53 contrast, verJcally transmi[ng corals provision symbionts via the eggs, ensuring early symbiosis 54 and high intergeneraJonal fidelity (9), though potenJally limiJng rapid adjustment to 55 environmental change. These contrasJng transmission strategies are therefore expected to 56 influence not only the composiJon of symbiont communiJes but also holobiont performance 57 under stress (10). 58 The reef-building coral Mon$pora capitata is a dominant species in Kāneʻohe Bay, Hawaiʻi, and a 59 key model for coral–Symbiodiniaceae symbiosis under environmental stress. It is a broadcast 60 spawner with verJcal transmission, releasing egg –sperm bundles that develop into larvae 61 containing maternally derived symbionts (10,11). This strategy establishes symbiosis early and 62 links parental and offspring communiJes, yet variaJon across colonies and environments 63 indicates both inherited and environmental influences on symbiont dynamics. 64 Across its range, M. capitata associates primarily with symbionts from the genera Cladocopium 65 and Durusdinium, which differ in physiological performance and stress tolerance (12,13). 66 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint Cladocopium is generally considered more thermally sensiJve but can support higher rates of 67 photosynthesis and growth under stable environmental condiJons (2,12,14,15). In contrast, 68 Durusdinium is widely recognized for its enhanced thermal tolerance and resistance to 69 bleaching, although this resilience is ojen associated with trade-offs in host growth or 70 metabolic efficiency (2,12,14–16). As a result, the relaJve abundance of these symbionts can 71 influence coral performance across environmental gradients, parJcularly under elevated 72 temperature regimes (14,15), providing a framework to invesJgate how symbiont traits are 73 maintained across life stages. 74 At the cellular level, the establishment and maintenance of the coral–Symbiodiniaceae 75 symbiosis is mediated by molecular recogniJon processes ojen described as a “lock -and-key” 76 mechanism, in which symbiont cell-surface glycans interact with host lecJns to enable selecJve 77 recogniJon, uptake, and retenJon of compaJble symbioJc partners (1,17). Experimental 78 evidence has shown that lecJn –glycan interacJons influence symbiont acquisiJon, with certain 79 carbohydrate moJfs promoJng or inhibiJng symbiont colonizaJon, thereby contribuJng to 80 host–symbiont specificity (14,18–28). Most of our current understanding of these recogniJon 81 mechanisms comes from horizontally transmi[ng systems, where symbiont uptake from the 82 environment requires acJve discriminaJon among diverse potenJal partners and lecJn –glycan 83 interacJons mediate partner selecJon at the host –cell interface. How these molecular 84 recogniJon processes operate in verJcally transmi[ng corals remains poorly understood. 85 Because symbionts are inherited directly from the parent colony via symbiont-provisioned eggs, 86 the need for de novo partner acquisiJon is assumed to be reduced. However, recogniJon 87 mechanisms may sJll be essenJal for regulaJng symbiont retenJon, proliferaJon , and quality 88 control across life stages. In this context, lecJn–glycan interacJons may funcJon not only in 89 iniJal symbiont recogniJon but also in maintaining symbioJc fidelity and mediaJng 90 intergeneraJonal conJnuity of host –symbiont associaJons. 91 In M. capitata from Kāneʻohe Bay, cell-surface glycan profiles of Cladocopium and Durusdinium 92 differed between genera and were significantly altered by temperature and oxidaJve stress, 93 suggesJng that deviaJons from baseline glycan signatures may contribute to the breakdown of 94 specific partnerships and the plasJcity of symbiosis under stress (14). 95 Here, we invesJgate how symbiont cell-surface glycan profiles are maintained and structured 96 across life stages in a verJcally transmi[ng coral. Using M. capitata as a model system, we test 97 whether lecJn -binding signatures of Symbiodiniaceae are conserved between parental colonies 98 and their offspring, and whether this signature varies between Cladocopium and Durusdinium, 99 providing insight into the role of lecJn –glycan mechanisms in maintaining symbiosis across 100 generaJons. 101 102 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint

103 Coral collec%on, Symbiodiniaceae isola%on, and qPCR 104 M. capitata colonies in Kāneʻohe Bay, Oʻahu, were selected based on visual health and colony 105 size to maximize the likelihood of spawning. Colonies were tagged in situ, and thumb-sized 106 fragments (N = 3-5) were collected from each colony for symbiont isolation and molecular 107 analyses. 108 To characterize the algal symbiont community, fragments were sampled using a 5 mm diameter 109 dermal curette, immediately snap-frozen in liquid nitrogen, and stored at −80°C until 110 processing. Genomic DNA was extracted using the Zymo Quick-DNA kit (Fisher Scientific, cat. 111 no. 50-444-148) following the manufacturer's protocol. Quantitative PCR (qPCR) assays 112 targeting clade-specific actin loci were used to quantify the relative abundance 113 of Cladocopium and Durusdinium in corals (12,29). All qPCR reactions were performed on a 114 StepOnePlus Real-Time PCR System (Applied Biosystems). Reactions were carried out in 10 µL 115 volumes containing 5 µL TaqMan Genotyping Master Mix and 1 µL genomic DNA template. 116 The M. capitata assay contained 1 µM forward primer (PaxC-F, 5ʹ-117 GTGCAGGTGAGATTGAGTCTTATAACA-3ʹ), 1.5 µM reverse primer (PaxC-R, 5ʹ-118 CGGTTGAGCTTCGCTAAACAG-3ʹ), and 2 µM TaqMan probe (PaxC-Probe, 5ʹ-FAM-119 CAGTTCTTCCAACAATG-MGB-3ʹ). The multiplexed Cladocopium/Durusdinium assay contained 1 120 µM Cladocopium forward primer (CActF, 5ʹ-CCAGGTGCGATGTCGATATTC-3ʹ), 1.5 121 µM Cladocopium reverse primer (CActR, 5ʹ-TGGTCATTCGCTCACCAATG-3ʹ), and 2 122 µM Cladocopium probe (CActProbe, 5ʹ-VIC-AGGATCTCTATGCCAACG-MGB-3ʹ), together with 1 123 µM Durusdinium forward primer (DActF, 5ʹ-GGCATGGGGTAAGCACTTCTT-3ʹ), 1.5 124 µM Durusdinium reverse primer (DActR, 5ʹ-GATCCTTGAACTAGCCTTGGAAAC-3ʹ), and 2 125 µM Durusdinium probe (DActProbe, 5ʹ-6FAM-CAAGAACGATACCGCC-MGB-3ʹ). All samples were 126 run in triplicate for each assay over 40 cycles. Thermal cycling conditions consisted of an initial 127 incubation at 50°C for 2 min and 95°C for 10 min, followed by 40 cycles of denaturation at 95°C 128 for 10 s and annealing/extension at 60°C for 1 min. Cycle threshold (CT) values were 129 determined using the StepOnePlus software with a baseline interval of cycles 6–23 and a 130 fluorescence threshold of Rn = 0.01. Positive amplification was scored only when technical 131 replicates yielded CT values < 40 and no amplification was detected in no-template controls. CT 132 values were corrected for differences in fluorescence intensity among the three reporter dyes 133 associated with the TaqMan probes. Symbiont-to-host (S/H) cell ratios for 134 Cladocopium and Durusdinium were calculated as 𝑆/𝐻 = 2!𝐶𝑇"#$% −𝐶𝑇$'()*#+% ,, and values 135 were normalized for differences in probe fluorescence intensity and target locus gene copy 136 number. Total S/H ratios were calculated as the sum of Cladocopium and Durusdinium S/H 137 ratios (12,29). 138 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint In total, 16 M. capitata colonies were selected, including 6 dominated by Cladocopium 139 (genotypes ID 162, 176, 177, 190, 214, 236) and 10 dominated by Durusdinium (169, 185, 197, 140 203, 208, 228, 231, 243, 249, 672). Colonies were used for downstream analyses if their 141 symbiont community was composed of > 99.5% of a single genus; composition data can be 142 found in Supplementary Table 1. 143 Adult samples preparation 144 Algal symbionts were freshly isolated from a subset of 11 parental colonies (Cladocopium 145 genotypes 162, 177, 190, 214, 236, and Durusdinium genotypes 185, 203, 208, 228, 231, 243) 146 by airbrushing tissue in 0.2 μm filtered seawater (FSW). Logistical constraints prevented us from 147 sampling all parental colonies. The resulting homogenate was dissociated using a sterile 22-148 gauge needle and syringe to obtain single cells, centrifuged at 1,000 × g for 5 min to pellet 149 symbiont cells and resuspended in FSW. This wash step was repeated five times to remove 150 residual host material. Symbiont cell pellets were fixed in 2% paraformaldehyde in 1x PBS and 151 stored in the dark at 4°C until further processing. 152 Spawning and egg samples prepara%on 153 Tagged M. capitata colonies were transported to the Hawaiʻi InsJtute of Marine Biology (HIMB) 154 in Kāneʻohe Bay three days prior to the new moon in May and June 2025 and maintained in 155 ambient flow-through seawater tanks under natural light condiJons. Gamete bundles were 156 collected from individual spawning coral colonies using a net on May 28 and June 26, 2025. 157 Buoyant gamete bundles were gathered in individual 50mL conical tubes and were gently 158 agitated to dissociate bundles and release eggs and sperm. Eggs were repeatedly washed with 159 FSW to remove residual sperm and debris, then aliquoted into 1.5 mL microcentrifuge tubes. 160 To isolate algal symbionts from the eggs, samples were vortexed horizontally at maximum speed 161 for 5 min and centrifuged at 1,000 × g for 5 min. The supernatant was discarded, and the 162 symbiont pellet was resuspended in FSW using a sterile 22-gauge needle and syringe. This wash 163 cycle was repeated five Jmes to discard eggs debris. Symbiont cell pellets from replicate 164 samples were pooled, fixed in 2% paraformaldehyde in PBS, and stored in the dark at 4°C unJl 165 further processing. 166 Lec%n staining and flow cytometry of Symbiodiniaceae cell-surface glycans 167 Fluorescent lecJns were used to characterize glycans present on the cell-surface of symbioJc 168 algae. The following lecJns , purchased from Thermo Fisher ScienJfic, were used: ConA 169 (concanavalin A; specific for D-mannose and D-glucose; cat. no. C21401), LTL (Lotus 170 tetragonolobus lecJn; specific for L -fucose; cat. no. L32480), PNA (peanut aggluJnin; specific for 171 D-galactose; cat. no. L21409), WGA (wheat germ aggluJnin; specific for N -acetylglucosamine 172 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint and N-acetylneuraminic acid; cat. no. W11261), PHAL (phytohemaggluJnin -L from Phaseolus 173 vulgaris; specific for N-acetylglucosamine β(1–2) mannopyranosyl residues; cat. no. L11270), 174 and GSIB4 (isolecJn B4 from Griffonia simplicifolia; specific for N-acetyl-D-galactosamine and α-175 D-galactosyl residues; cat. no. I21411). LecJns were conjugated with either Alexa Fluor 488 176 (ConA, PNA, WGA, PHAL, GSIB4) or FITC (LTL). LecJn -binding profiles of Symbiodiniaceae cell-177 surface glycans were quanJfied for symbiont s isolated from the eggs of 6 Cladocopium-178 dominated and 10 Durusdinium-dominated colonies, and from adult Jssue of 5 Cladocopium-179 dominated and 6 Durusdinium-dominated parental colonies. 180 Algal cell density was determined from quadruplicate counts using a Countess 3 FL Automated 181 Cell Counter (Thermo Fisher ScienJfic). For each sample, 2.5 × 10⁵ fixed symbiont cells were 182 incubated separately with each fluorescent lecJn at a final concentraJon of 1 mg ml⁻¹ for 1 hour 183 at room temperature in the dark. Three stained replicates were prepared for each biological 184 replicate. Following incubaJon, cells were washed with 1× PBS by centrifugaJon at 3,000 × g for 185 5 min and resuspended in a final volume of 500 µl PBS (14). 186 Samples were processed on a CytoFlex S flow cytometer (Beckman-Coulter) for at least 90 s at a 187 flow rate of 20 µl min-1 and data were analyzed using FloJo sojware (v10.10.0, BD BioSciences). 188 Gain parameters were established before running experimental samples using unstained and 189 stained algal cells. Algal cell populaJons were first gated by chlorophyll autofluorescence and 190 Alexa Fluor 488 fluorescence (on the SYBR Green channel) based on excitaJon at 488 nm and 191 detected by the PerCP channel (690/50 bandpass filter) and FITC channel (525/40 bandpass 192 filter), respecJvely. The median fluorescence intensity (MFI) of the lecJn -binding height signal 193 (SYBR Green) was quanJfied using the gated algal populaJon (Supplementary Fig. 1). 194

fluorescence was corrected by subtracJng the MFI of unstained cells from lecJn -195 stained samples. 196 Sta%s%cal analysis 197 All staJsJcal analyses were conducted in R (v4.1.2). Coral colonies were classified according to 198 their dominant symbiont genus based on qPCR esJmates of Cladocopium and Durusdinium 199 relaJve abundance following Cunning et al. 2013 (29). 200 LecJn -binding profiles of symbiont cell-surface glycans were quanJfied as median fluorescence 201 intensity (MFI) for each of the six lecJns (ConA, LTL, PNA, WGA, PHAL, GSIB4) and unstained 202 controls. LecJn MFI values were corrected by subtracJng the chlorophyll autofluorescence 203 signal measured for each sample from the corresponding unstained negaJve control, thereby 204 accounJng for background fluorescence and isolaJng the lecJn -specific binding signal. 205 MulJvariate differences in lecJn -binding profiles between symbiont genera and sample origin 206 (eggs vs. parents) were assessed using permutaJonal mulJvariate analysis of variance 207 (PERMANOVA; 999 permutaJons) and homogeneity of mulJvariate dispersion ( β-dispersion) 208 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint implemented in the vegan package (30). We used parJal least squares discriminant analysis 209 (PLS-DA) to evaluate the ability of each glycan’s z-scored MFI to disJnguish symbiont types in 210 eggs and parents using the mixOmics package (31). 211 To test the effects of lecJn type and symbiont genus on lecJn -binding profiles of symbionts 212 isolated from eggs and parental coral colonies, we fi†ed a generalized linear model with a 213 Student’s t error distribuJon using the glmmTMB package (32). The model included lecJn type, 214 symbiont genus, and host genotype nested within symbiont genus as fixed effects (MFI ~ lecJn 215 × symbiont + symbiont/genotype). Model diagnosJcs were performed using the DHARMa 216 package (33). Significance of fixed effects was assessed using analysis of variance (ANOVA), and 217 post hoc pairwise comparisons were conducted using Tukey-adjusted contrasts implemented in 218 the emmeans package (34). 219 AssociaJons between lecJn -binding profiles of symbionts isolated from eggs and parental 220 colonies were examined using pairwise correlaJon analyses. CorrelaJon matrices were 221 constructed from z-scaled MFI values across all six lecJns, and similarity between matrices was 222 assessed using Mantel tests (999 permutaJons) implemented in the vegan package (30). 223 Comparisons were performed between symbiont genera from eggs and between eggs and 224 parents within each symbiont genus. 225 We compared Mahalanobis distances between matched parent–offspring pairs (POP) and 226 unmatched comparisons (other) within each symbiont genus to assess whether lecJn -binding 227 profiles of symbiont cell-surface glycans were conserved between parental colonies and their 228 offspring. Distances were analyzed using a generalized linear model including symbiont genus, 229 comparison type (POP or other) and their interacJon (distance ~ symbiont × comparison type). 230 Significance was assessed using analysis of variance (ANOVA), and post hoc pairwise 231 comparisons were conducted using Tukey-adjusted contrasts implemented in the emmeans 232 package (34). To assess transmission of each glycan, we used z-scaled MFI data as input for 233 parent-offspring linear regressions separately for each lecJn and symbiont type (egg ~ parent), 234 and an ANCOVA to assess differences in slope between symbiont types for each lecJn. 235 236

237 Symbiont genera and life stage differences in lec%n-binding profiles 238 MulJvariate analysis revealed clear separaJon in lecJn -binding profiles of symbiont cell-surface 239 glycans between symbiont genera at both life history stages. Principal component analysis 240 showed disJnct clustering of Cladocopium and Durusdinium in eggs (PERMANOVA, P = 0.001; 241 Fig. 1A) and parents (PERMANOVA, P = 0.001; Fig. 1B), indicaJng genus -specific lecJn -binding 242 signatures. The lecJn ConA contributed strongly to the separaJon between symbiont genera in 243 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint both eggs and adults. WGA and PNA also strongly separated symbiont genera in eggs, but not in 244 parental corals, suggesJng a more complex glycan landscape in gametes than adults. PC2 245 captured substanJal addiJonal variaJon in eggs and within Cladocopium in adults. MulJvariate 246 dispersion was not different between symbiont genera in eggs (β-dispersion, P > 0.05; Fig. 1A) 247 but was significantly higher in Cladocopium-hosJng parents ( β-dispersion, P = 0.043; Fig. 1B). 248 Differences between symbionts isolated from eggs and parental colonies were detected in both 249 genera (PERMANOVA, P = 0.001; Fig. 1C, D), indicaJng that variability in lecJn -binding profiles 250 also differed based on symbiont community. In Cladocopium, the separaJon was less 251 pronounced, and confidence ellipses demonstrated some overlap (Fig. 1C), but in Durusdinium, 252 egg- and parent-derived symbionts formed discrete clusters (Fig. 1D). Symbionts isolated from 253 eggs had higher mulJvariate dispersion than symbionts isolated from parents in both genera (β-254 dispersion, P < 0.001; Fig. 1C, D). ParJal least squares discriminant analysis (PLS -DA) showed 255 that the lecJns PNA and WGA were disproporJonately important for disJnguishing symbiont 256 communiJ es in eggs, while LTL and PHAL were more important in adults (Fig. 1E). The lecJns 257 ConA and GSIB4 were approximately equally important for disJnguishing the symbiont 258 community in eggs and parents. 259 Analysis of proporJonal median fluorescent intensity (MFI) of lecJn -binding signal composiJon 260 revealed that both symbiont genera from eggs and parental colonies were dominated by ConA 261 and WGA signals, with relaJvely minor contribuJons from other lecJns (Supplementary Fig. 2). 262 In Cladocopium from eggs, genotype-level profiles were relaJvely consistent, with ConA 263 contribuJng the majority of total signal across all genotypes (Supplementary Fig. 2C). 264 Durusdinium from eggs exhibited greater variability among genotypes, parJcularly in the 265 relaJve contribuJon of WGA and secondary lecJns (Supplementary Fig. 2D), suggesJng 266 increased heterogeneity in lecJn -binding profiles within this genus. 267 Lec%n-specific differences in symbiont cell-surface glycan binding 268 To idenJfy lecJn -specific differences underlying mulJvariate pa†erns, we compared median 269 fluorescence intensity (MFI) across symbiont genera using generalized linear models. In eggs 270 and parental colonies, lecJn type, symbiont genus, and their interacJon , all significantly 271 influenced MFI (GLM, P < 0.0001 for all terms), indicaJng that glycan -binding differences are 272 lecJn -dependent and vary between symbiont genera (Fig. 2). Host genotype nested within 273 symbiont genus also contributed significantly to variaJon in MFI (GLM, P < 0.001). Consistent 274 with these effects, Durusdinium exhibited higher MFI than Cladocopium for several lecJns ( Fig. 275 2, Supplementary Fig. 2), with the strongest differences observed in eggs in ConA and WGA (P < 276 0.0001), as well as GSIB4 (P = 0.003) and PNA (P = 0.012), and in parents in ConA (P = 0.012), 277 GSIB4 (P = 0.045) and PHAL (P = 0.042). 278 Correla%on structure of lec%n-binding profiles across symbiont genera and life stages 279 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint To characterize relaJonships between lecJn -binding profiles of symbiont cell-surface glycans, 280 we constructed correlaJon matrices for each symbiont genus and coral life stage (Fig. 3). In 281 Cladocopium, correlaJons were exclusively posiJve in eggs and parents, although some lecJn 282 pairs showed weak or no associaJon. The overall correlaJon structure was significantly different 283 between life history stages in Cladocopium (Mantel r = 0.861, P = 0.001; Fig. 3A). In contrast, 284 Durusdinium exhibited more variable correlaJon pa†erns, with both posiJve and negaJve 285 associaJons observed in egg - and parent-derived symbionts. Generally weaker correlaJons 286 within PNA and ConA in both eggs and parents with Durusdinium contributed to a lower 287 correlaJon between sample types, although the overall correlaJon structure was sJll 288 significantly different between life history stages (Mantel r = 0.505, P = 0.044; Fig. 3B). 289 CorrelaJon matrices were significantly different between symbiont genera in samples from 290 parents (Mantel r = 0.726, P = 0.003) and eggs (Mantel r = 0.777, P = 0.019). 291 Host genotype-specific similarity in symbionts lec%n-binding profiles 292 To assess whether lecJn -binding profiles of symbiont cell-surface glycans were conserved 293 between parental colonies and their offspring, we compared Mahalanobis distances between 294 matched parent–offspring pairs (POP) and unmatched comparisons (other) within each 295 symbiont genus. Distances were analyzed using a generalized linear model including symbiont 296 genus, comparison type (POP or other), and their interacJon. Distance was significantly 297 influenced by the interacJon between symbiont genus and comparison type ( χ² = 23.26, df = 1, 298 P < 0.001), but the main effect of symbiont and comparison type alone were not significant 299 (Symbiont χ² = 0.07, df = 1, P = 0.795; Group χ² = 1.61, df = 1, P = 0.205). Post hoc comparisons 300 revealed that distances were significantly smaller in POP than in unrelated pairs in Cladocopium 301 (P = 0.003) and significantly higher than in unrelated pairs in Durusdinium (P < 0.0001; Fig. 4A). 302 The parent-offspring regressions showed a significant parent–symbiont relaJonship in the 303 lecJns ConA ( P = 0.023) and PNA (P = 0.028; Fig. 4B), while there was no significant effect in the 304 other lecJns ( P > 0.05). For the significant lecJns, R2 of the scaled parent-offspring regression 305 were ConA (R2 = 0.415) and PNA (R2 = 0.027), indicaJng substanJal explained variance. We 306 observed near-zero or negaJve slopes for other lecJns, which could reflect biological variaJon 307 in glycomes across life history traits or may be representaJve of noise in a relaJvely small 308 dataset. The slope of the parent–offspring regression was significantly different between 309 symbiont genera in ConA (P = 0.038) and PNA (P = 0.032), but not other glycans (Fig. 4B). We 310 observed several negaJve correlaJons between parent and egg MFIs in several combinaJons, 311 but none were significant. 312 313

314 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint The ability of coral reefs to persist under climate change depends in part on the composiJon 315 and flexibility of corals’ symbioJc partnerships , which represent the energeJc foundaJon of the 316 ecosystem. AssociaJons between Symbiodiniaceae and their cnidarian hosts rely on molecular 317 specificity, which underpins the establishment and maintenance of the endosymbiosis. In 318 verJcally transmi[ng corals, symbionts are directly provisioned to offspring via eggs and are 319 therefore expected to promote the conJnuity of symbiont communiJes across generaJons, 320 increasing symbiont fidelity by coupling fitness between both partners (35). This method of 321 symbiont transmission is hypothesized to arise from more mutually beneficial partnerships (36) 322 and may reduce the likelihood of incompaJble or marginal associaJons (37). However, it 323 remains unclear whether the symbionts’ cellular traits are conserved during transmission and 324 across host life stages, and whether they contribute to the stability of verJcally transmi†ed 325 symbioses. 326 Here, we examined cell-surface glycan profiles of Cladocopium and Durusdinium in parental 327 colonies and eggs in the verJcally transmi[ng coral M. capitata. The Symbiodiniaceae ITS2 328 sequences associated with eggs largely mirror those of their parental colonies in Cladocopium- 329 and Durusdinium-dominated genotypes (10), indicaJng stable transmission at the community 330 level. Our results showed that glycan signatures were structured by symbiont genus and differed 331 between life stages, with mannose/glucose and galactose emerging as key contributors to this 332 variaJon. 333 Egg-derived symbionts exhibited greater variability in lectin-binding profiles compared to those 334 from parental colonies (Fig. 1C, D). This pattern was evident in both symbiont genera, but was 335 more pronounced in Durusdinium (Fig. 1D). Increased variability in eggs suggests that glycan 336 profiles may be less constrained at early life stages, potentially reflecting heterogeneity in 337 symbiont populations prior to stabilization within the adult host environment. In contrast, the 338 reduced variability observed in parental colonies may indicate more consistent or regulated 339 glycan expression in established symbioses. This shift from higher variability in eggs to more 340 uniform profiles in adults is consistent with increasing constraint in host–symbiont interactions 341 over time (8) and provides a potential framework for interpreting the roles of specific glycans 342 across life stages. 343 Mannose/glucose (ConA) and galactose (PNA) terminal glycans have been central to studies of 344 Symbiodiniaceae cell–surface features and host–symbiont recogniJon. In our study, ConA 345 exhibited the strongest MFI profile across samples (Fig. 1, 2; Supplementary Fig. 2), consistent 346 with previous reports (14,20,24). This pa†ern likely reflects the abundance of mannose- and 347 glucose-containing glycans on the Symbiodiniaceae surface (11,14,18,19), supported by the 348 cellulose-rich dinoflagellate cell wall (21) and the enrichment of terminal mannose residues on 349 surface glycoproteins (20). FuncJonally , mannose-rich glycans have been directly linked to 350 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint symbiosis establishment and maintenance. Host-derived mannose-binding lecJns localized 351 around symbionts within the gastrodermis of Acropora millepora (38) and Pocillopora 352 damicornis (39), and showed substanJal variability in binding domains, suggesJng a role in fine -353 scale recogniJon of compaJble partners (38). Experimental saturaJon of Exaiptasia diaphana 354 lecJn receptors with D-mannose reduced colonizaJon success by homologous symbionts (24), 355 further supporJng a role in recogniJon rather than a purely structural funcJon in dinoflagellate 356 cell walls. 357 D-galactose–containing glycans similarly contribute to symbioJc interacJons. Galactose 358 residues are present on the surface of mulJple Symbiodiniaceae species (14,22,24,25) and have 359 been implicated in symbiosis establishment in A. tenuis (40) larvae and anemone model (24). 360 Two D-galactose-binding lecJn s, SLL-2 and CeCL, induced the transiJon of Symbiodiniaceae 361 from a moJle, flagellated state to a non -moJle coccoid form characterisJc of the symbioJc 362 state, and localized around symbiont cells within host Jssues (19,40–42). Together, these 363 findings highlight both mannose/glucose- and galactose-containing glycans as key mediators of 364 host–symbiont integraJon and emphasize the relaJve importance of galactose in gametes, 365 where PNA is a stronger marker of symbiont variaJon and idenJty than in parents (Fig. 1, 4). 366 MulJvariate differences between Cladocopium and Durusdinium were supported by pairwise 367 correlaJons in lecJn -binding intensiJes, with ConA and PNA – key drivers of genus-level 368 separaJon – showing coordinated posiJve pa†erns in egg -derived symbionts (Fig. 3). In 369 parental colonies, ConA and PNA correlaJons diverged between genera, posiJve in 370 Cladocopium-dominated corals and negaJve in Durusdinium-dominated corals, although these 371 relaJonships were not significant (Fig. 3). Previous work on M. capitata from Kāne‘ohe Bay 372 showed that ConA binding profiles shij under thermal stress and are linked to oxidaJve stress, 373 with similar trends observed for PNA (14). Considering oxidaJve stress can alter glycan 374 composiJon through redox -mediated modificaJon of glycoproteins (43), these observaJons 375 suggest that symbiont cell-surface glycans may integrate physiological state with recogniJon 376 processes. 377 Within this framework, our results provide evidence that the transmission of specific cell-378 surface glycans is conserved across life stages in Cladocopium and Durusdinium. The significant 379 difference in Mahalanobis distances between matched parent–offspring pairs relaJve to 380 unrelated comparisons (Fig. 4A) suggests that verJcally transmi†ed Cladocopium symbionts 381 retain genotype-specific glycan signatures, consistent with parental effects mediated through 382 direct transmission of symbiont traits. However, this conservaJon is not uniform across the 383 glycome. LecJn -level analyses revealed that only ConA and PNA exhibited significant parent–384 offspring correlaJons (Fig. 4B), and these were genus-specific, indicaJng that intergeneraJonal 385 similarity is driven by disJnct glycan classes in each lineage. The high similarity in Cladocopium-386 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint dominated parents and eggs is likely driven by the strong heritability of mannose/glucose in this 387 genus, which is also the most abundant glycan, weighJng this analysis. 388 Notably, these same lecJns were primary contributors to genus-level differences in eggs and 389 parents (Fig. 1), linking divergence between symbiont genera with parent–offspring conJnuity. 390 This suggests that some glycan features, primarily mannose/glucose (ConA specificity), that 391 disJnguish Cladocopium and Durusdinium, are also those most consistently maintained across 392 life stages, and may therefore play a funcJonal role in maintaining host–symbiont compaJbility. 393 Conversely, galactose and N-acetylglucosamine (PNA and WGA specificity, respecJvely) were 394 disproporJonately important for disJnguishing symbionts in gametes (Fig. 1A), but these 395 glycans are orthogonal to the axis of separaJon in adults (Fig. 1B), suggesJng differenJal 396 provisioning and funcJon in early life histories . This type of variaJon across life history stages 397 matches moderate heritability differences between symbiont genera, suggesJng some degree 398 of strategic provisioning that differs based on the type of symbiont and corresponds to the 399 overall breakdown in correlaJon structure in the glycome of the two corals. 400 All pairwise correlaJons in Cladocopium adults and eggs were posiJve, although some were not 401 significant, indicaJng that differenJaJon between individuals within this populaJon is driven by 402 variaJon in the intensity of the lecJn -probe signal and presumably the relaJve abundance of 403 the glycans on cell surfaces (Fig. 2A). Conversely, in Durusdinium, negaJve correlaJons suggest 404 some tradeoffs in relaJve abundance (Fig. 2B), which may be indicaJve of stronger 405 composiJonal variaJon marks in Durusdinium, especially in offspring where mulJple strong 406 negaJve correlaJons emerge that are not found in adults or in Cladocopium. 407 Together, our findings support a model in which verJcal transmission preserves recogniJon -408 relevant glycan features while allowing flexibility in other components of the symbiont surface. 409 Rather than broad inheritance of the glycome, specific glycan classes—likely those involved in 410 host–lecJn interacJons —appear to be selecJvely maintained across generaJons. Future work 411 should extend these analyses to larvae and juveniles to determine whether glycan traits are 412 maintained at these life stages, and to test whether shijs in the ConA:PNA raJo—reflecJng 413 mannose/glucose vs. galactose glycans—mediate parental effects and the preferenJal 414 associaJon of Durusdinium under stress through glyco-redox–linked mechanisms. This selecJve 415 retenJon may represent a mechanism by which corals ensure symbiont fidelity while retaining 416 physiological plasJcity under changing environmental condiJons. 417 Author Contribu%ons 418 GT and CD conceived the study and analyzed data. GT, KH and CD secured funding. GT, NLF, ACV, 419 SLR, IAA, EM, KS and CD performed the experiments. GT wrote the manuscript. All authors 420 collected data and revised the manuscript. 421 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint

422 Hawaiian culture fosters a reciprocal connecJon between people and coral reefs, including 423 those in Kāneʻohe Bay, O‘ahu, Hawai‘i where this research was conducted. We hope to honor 424 this relaJonship by recognizing its foundaJonal importance in our work. We are grateful to the 425 Coral Resilience Lab for field and laboratory support. All collecJons were made under Hawaiʻi 426 Department of Land and Natural Resources permits to HIMB (SAP 2026-16). We thank the 427 Walder-Christensen Charitable Fund for supporJng this work. This is SOEST contribuJon #xx and 428 HIMB contribuJon #xx. 429 430 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint 431 Fig. 1. | Lec+n-binding profiles dis+nguish symbiont genera and host life stages. 432 A. Principal component analysis (PCA) of lec5n-binding intensi5es in Cladocopium (orange, N = 18) 433 and Durusdinium (blue, N = 30) symbionts isolated from coral eggs. B. PCA of lec5n-binding 434 intensi5es in Cladocopium and Durusdinium symbionts isolated from parental colonies (N = 15 435 and N = 18, respec5vely). C. PCA of lec5n-binding intensi5es in Cladocopium symbionts isolated 436 from coral eggs (circles) and parental colonies (triangles) pairs (N = 15). D. PCA of lec5n-binding 437 intensi5es in Durusdinium symbionts isolated from coral eggs and parental colony pairs (N = 18). In 438 A-D, vectors represent the contribu5on of individual lec5ns to the ordina5on. Ellipses represent the 439 95% confidence intervals around group centroids. Group separa5ons were sta5s5cally significant 440 (PERMANOVA P = 0.001). Asterisks denote adjusted P values (*P< 0.05, ** P< 0.01, *** P< 0.001). E. 441 ParJal least squares discriminant analysis (PLS -DA) of each lec5n-binding intensity profile for 442 symbionts in eggs and parents. Lec5ns: ConA (concanavalin A, specific for D-mannose and D-443 glucose), LTL (Lotus tetragonolobus lec5n, specific for L-fucose), PNA (Arachis hypogaea lec5n, 444 specific for D-galactose), WGA (wheat germ agglu5nin, specific for N-acetylglucosamine and N-445 acetylneuraminic acid), PHA-L (phytohemagglu5nin-L from Phaseolus vulgaris, specific for N-446 acetylglucosamine β(1-2) mannopyranosyl) and GS-IB4 (isolec5n from Griffonia simplicifolia, specific 447 for N-acetyl-D-galactosamine and a-D-galactosyl residues). 448 449 PNA LTL PHAL WGA GSIB4 ConA Eggs P =0.001*** β> 0.05 -2 0 2 -2.5 0.0 2.5 5.0 PC1 (57%) PC2 (28%) Cladocopium Durusdinium A ConA LTL PNA WGA PHAL GSIB4 Parents P =0.001*** β= 0.043* -3 0 3 -4 -2 0 2 PC1 (47%) PC2 (35%) B PNA LTL PHAL WGA GSIB4 ConA C l ad ocopi um P =0.001*** β= 0.005** -2 0 2 -4 0 4 PC1 (65%) PC2 (31%) Egg Parent C PNA LTL PHAL WGA GSIB4 ConA D urusd i ni um P =0.001*** β= 0.001***-5.0 -2.5 0.0 2.5 5.0 -4 -2 0 2 PC1 (55%) PC2 (39%) Egg Parent D More Important for Eggs More Important for Adults 0.0 0.1 0.2 0.3 0.4 0.0 0.1 0.2 0.3 0.4 Adult Importance Egg Importance ConA GSIB4 LTL PHAL PNA WGA E .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint 450 Fig. 2. | Lec%n-binding intensity differs between genera in symbionts isolated from coral eggs 451 and parental colonies. 452 A. Median fluorescence intensity (MFI) raw data of lecJn -binding measured in symbiont cells 453 from the genera Cladocopium (orange, N = 18) and Durusdinium (blue, N = 30) isolated from 454 coral eggs. B. Median fluorescence intensity (MFI) raw data of lecJn -binding measured in 455 symbiont cells from the genera Cladocopium (N = 15) and Durusdinium (N = 18) isolated from 456 parental coral colonies. Points represent genotype-level mean values. Bars indicate the mean 457 across genotypes, and error bars represent the standard error (SE). Asterisks denote adjusted P 458 values (*P< 0.05, ** P< 0.01, *** P< 0.001). 459 460 **** ** * **** lectin*symbiont<0.001 lectin*symbiont*genotype<0.001 0e+00 1e+06 2e+06 3e+06 4e+06 5e+06 ConA GSIB4 LTL PHAL PNA WGA Lectin Eggs MFI (raw values) Cladocopium Durusdinium A * * * lectin*symbiont<0.001 lectin*symbiont*genotype<0.001 0.0e+00 5.0e+05 1.0e+06 1.5e+06 2.0e+06 ConA GSIB4 LTL PHAL PNA WGA Lectin Parents MFI (raw values) B .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint 461 Fig. 3. | Pairwise Pearson correla%ons between lec%n-binding intensi%es differ between 462 symbiont genera and host life stages. 463 Heatmaps show pairwise Pearson correlaJons in lecJn -binding intensiJes measured in A. 464 Cladocopium and B. Durusdinium cells isolated from coral eggs (upper triangles) and parental 465 colonies (bo†om triangles). Color scale represents the strength and direcJon of correlaJons 466 (purple: posiJve, green: negaJve ). Similarity between lecJn correlaJon matrices was assessed 467 using Mantel tests based on 999 permutaJons ( PC-eggs vs C-parents= 0.001; PD-eggs vs D-parents= 0.044; 468 PC-eggs vs D-eggs= 0.003; PC-parents vs D-parents= 0.019). Asterisks within cells indicate significant 469 correlaJon coefficients between lecJn -binding intensiJes (* P< 0.05, ** P< 0.01, *** P< 0.001). 470 *** ****** *** ****** ****** * *** *** *** *** ** *** ****** *** *** *** PNA GSIB4LTL PHAL ConA WGA WGA ConA PHAL LTL GSIB4 PNA Cladocopium Upper triangle: eggs | Lower triangle: parents A *** *** *** *** * *** *** ** *** ***** ** * ****** *** PNA GSIB4LTL PHAL ConA WGA WGA ConA PHAL LTL GSIB4 PNA Durusdinium Upper triangle: eggs | Lower triangle: parents B Pearson r -1.0 -0.5 0.0 0.5 1.0 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint 471 Fig. 4 | Lec%n-binding profiles are conserved between symbiont genera%ons. 472 A. Mahalanobis distance distribuJons comparing lecJn -binding profiles for symbiont cells 473 associated with coral eggs. Distances were grouped by symbiont (Cladocopium and 474 Durusdinium) and parental origin (parent-offspring populaJon, POP; other populaJons, Other). 475 Significance was assessed using Tukey-adjusted contrasts. B. Pairwise correlaJons between 476 scaled lecJn -binding intensiJes measured in symbionts isolated from coral eggs and their 477 corresponding parental colonies, for Cladocopium (orange) and Durusdinium (blue). Each panel 478 shows one lecJn (ConA, GSIB4, LTL, PHAL, PNA, WGA). The do†ed line represents the 1:1 479 idenJty line , and the solid line represents a linear regression fit with 95% confidence interval 480 shading. Pearson correlaJon coefficients and significance are indicated. Asterisks denote 481 adjusted P values (*P< 0.05, ** P< 0.01, *** P< 0.001). 482 483 ** **** Cladocopium Durusdinium POP Other POP Other 2 4 6 Parental Origin Mahalanobis Distance A 0.024* 0.538 0.325 0.331 0.846 0.239 0.611 0.3 0.341 0.029* 0.722 0.328 ConA GSIB4 LTL PHAL PNA WGA CladocopiumDurusdinium -3 -2 -1 0 1 2 -3 -2 -1 0 1 2 -3 -2 -1 0 1 2 -3 -2 -1 0 1 2 -3 -2 -1 0 1 2 -3 -2 -1 0 1 2 -2 -1 0 1 2 -2 -1 0 1 2 Scaled Egg Intensity Scaled Parent Intensity B .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint

484 1. Davy SK, Allemand D, Weis VM. Cell Biology of Cnidarian-Dinoflagellate Symbiosis. Vol. 485 76. 2012;76(2):229–61. doi:10.1128/MMBR.05014-11 486 2. LaJeunesse TC, Parkinson JE, Gabrielson PW, Jeong HJ, Reimer JD, Voolstra CR, et al. 487 SystemaJc Revision of Symbiodiniaceae Highlights the AnJquity and Diversity of Coral 488 Endosymbionts. Current Biology. 2018;28(16):2570-2580.e6. 489 doi:10.1016/j.cub.2018.07.008 PubMed PMID: 30100341. 490 3. Vaga CF, Qua†rini AM, Galvão de Lossio e Seiblitz I, Huang D, Quek ZBR, Stolarski J, et al. 491 A global coral phylogeny reveals resilience and vulnerability through deep Jme. Nature. 492 2025;1–6. 493 4. Rädecker N, Pogoreutz C, Gegner HM, Cárdenas A, Roth F, Bougoure J, et al. Heat stress 494 destabilizes symbioJc nutrient cycling in corals. Proceedings of the naJonal academy of 495 sciences. 2021;118(5):e2022653118. 496 5. Weis VM. Cellular mechanisms of Cnidarian bleaching : stress causes the collapse of 497 symbiosis. 2008;3059–66. doi:10.1242/jeb.009597 498 6. Helgoe J, Davy SK, Weis VM, Rodriguez-Lane†y M. Triggers, cascades, and endpoints: 499 connecJng the dots of coral bleaching mechanisms. Biological Reviews. 2024;99(3):715 –500 52. 501 7. Eakin CM, Heron SF, Connolly SR, Devo†a DA, Liu G, Geiger EF, et al. Severe and 502 widespread coral reef damage during the 2014-2017 Global Coral Bleaching Event. Nat 503 Commun. 2026;17(1):1318. 504 8. Baker AC. Flexibility and specificity in coral-algal symbiosis: diversity, ecology, and 505 biogeography of Symbiodinium. Annu Rev Ecol Evol Syst. 2003;34(1):661–89. 506 9. Baird AH, Guest JR, Willis BL. SystemaJc and Biogeographical Pa†erns in the 507 ReproducJve Biology of ScleracJnian Corals. Annu Rev Ecol Evol Syst. 2009;40(1):551 –71. 508 doi:10.1146/annurev.ecolsys.110308.120220 509 10. Padilla-Gamiño JL, Pochon X, Bird C, Concepcion GT, Gates RD. From parent to gamete: 510 verJcal transmission of Symbiodinium (Dinophyceae) ITS2 sequence assemblages in the 511 reef building coral MonJpora capitata. PLoS One. 2012;7(6):e38440. 512 11. de Souza MR, Caruso C, Ruiz-Jones L, Drury C, Gates R, Toonen RJ. Community 513 composiJon of coral -associated Symbiodiniaceae differs across fine-scale environmental 514 gradients in Kāne ‘ohe Bay. R Soc Open Sci. 2022;9(9). 515 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint 12. Cunning R, Ritson-Williams R, Gates RD. Pa†erns of bleaching and recovery of MonJpora 516 capitata in Kāne ‘ohe Bay, Hawai ‘i, USA. Mar Ecol Prog Ser. 2016;551:131–9. 517 13. Drury C, Bean NK, Harris CI, Hancock JR, Huckeba J, Roach TNF, et al. IntrapopulaJon 518 adapJve variance supports thermal tolerance in a reef -building coral. Commun Biol. 519 2022;5(1):486. 520 14. Tortorelli G, Rosset SL, Sullivan CES, Woo S, Johnston EC, Walker NS, et al. Heat-induced 521 stress modulates cell surface glycans and membrane lipids of coral symbionts. ISME J. 522 2025;19(1):wraf073. 523 15. Dilworth J, Caruso C, Kahkejian VA, Baker AC, Drury C. Host genotype and stable 524 differences in algal symbiont communiJes explain pa†erns of thermal stress response of 525 MonJpora capitata following thermal pre -exposure and across mulJple bleaching events. 526 Coral Reefs. 2021;40(1):151–63. 527 16. Silverstein RN, Cunning R, Baker AC. Tenacious D: Symbiodinium in clade D remain in reef 528 corals at both high and low temperature extremes despite impairment. Journal of 529 Experimental Biology. 2017;220(7):1192–6. 530 17. Rosset SL, Oakley CA, Ferrier-Pagès C, Sugge† DJ, Weis VM, Davy SK. The molecular 531 language of the cnidarian–dinoflagellate symbiosis. Trends Microbiol. 2021;29(4):320–33. 532 18. Wood-charlson EM, Weis VM. The diversity of C-type lecJns in the genome of a basal 533 metazoan , Nematostella vectensis. Vol. 33. 2009;33:881–9. 534 doi:10.1016/j.dci.2009.01.008 535 19. Jimbo M, Yanohara T, Koike K, Koike K. The D -galactose-binding lecJn of the octocoral 536 Sinularia lochmodes : characterizaJon and possible relaJonship to the symbioJc 537 dinoflagellates. Vol. 125. 2000;125:227–36. 538 20. Lin K lin, Wang J terng, Fang L shing. ParJcipaJon of Glycoproteins on Zooxanthellal Cell 539 Walls in the Establishment of a SymbioJc RelaJonship with the Sea Anemone, Aiptasia 540 pulchella. Vol. 39. 2000;39(3):172–8. 541 21. Markell et al 1992. Macromolecules associated with the cell walls of symbioJc 542 dinoflagellates. 1992. 543 22. Logan DDK, Laflamme AC, Weis VM, Davy SK. FLOW-CYTOMETRIC CHARACTERIZATION OF 544 THE CELL-SURFACE GLYCANS OF SYMBIOTIC DINOFLAGELLATES ( SYMBIODINIUM SPP .) 1. 545 Vol. 533. 2010;533:525–33. doi:10.1111/j.1529-8817.2010.00819.x 546 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint 23. Wood-charlson EM, Hollingsworth LL, Krupp DA, Weis VM. LecJn / glycan interacJons 547 play a role in recogniJon in a coral / dinoflagellate symbiosis. Vol. 8. 2006;8(July):1985–548 93. doi:10.1111/j.1462-5822.2006.00765.x 549 24. Tortorelli G, Rautengarten C, Bacic A, Segal G, Ebert B, Davy SK, et al. Cell surface 550 carbohydrates of symbioJc dinoflagellates and their role in the establishment of 551 cnidarian–dinoflagellate symbiosis. ISME J. 2022;16(1):190–9. 552 25. Trevor R. Tivey, John Evere† Parkinson, Paige E. Mandelare, Donovon A. Adpressa, 553 Wenjing Peng, Xue Dong, Yehia Mechref, Virginia M. Weis SL. N-linked surface glycan 554 biosynthesis, composiJon, inhibiJon, and funcJon in cnidarian - dinoflagellate symbiosis. 555 2020. 556 26. Maruyama S, Mandelare-Ruiz PE, McCauley M, Peng W, Cho BG, Wang J, et al. Heat stress 557 of algal partner hinders colonizaJon success and alters the algal cell surface glycome in a 558 cnidarian-algal symbiosis. Microbiol Spectr. 2022;10(3):e01567-22. 559 27. Bay LK, Cumbo VR, Abrego D, Kool JT, Ainsworth TD, Willis BL. InfecJon Dynamics Vary 560 between Symbiodinium Types and Cell Surface Treatments during Establishment of 561 Endosymbiosis with Coral Larvae. 2011;356–74. doi:10.3390/d3030356 562 28. Jimbo M, Yamashita H, Zooxanthellae ÁÁ. Effects of lecJn in the scleracJnian coral 563 CtenacJs echinata on symbioJc zooxanthellae. 2010;355 –63. doi:10.1007/s12562-009-564 0204-z 565 29. Cunning R, Baker AC. Excess algal symbionts increase the suscepJbility of reef corals to 566 bleaching. Nat Clim Chang. 2013;3(3):259–62. 567 30. Oksanen J, Blanchet FG, Kindt R, Legendre P , Minchin PR, O’hara RB, et al. Community 568 ecology package. R package version. 2013;2(0):321–6. 569 31. Rohart F, GauJer B, Singh A, Lê Cao KA. mixOmics: An R package for ‘omics feature 570 selecJon and mulJple data integraJon. PLoS Comput Biol. 2017;13(11):e1005752. 571 32. Brooks ME, Kristensen K, Van Benthem KJ, Magnusson A, Berg CW, Nielsen A, et al. 572 glmmTMB balances speed and flexibility among packages for zero-inflated generalized 573 linear mixed modeling. 2017. 574 33. HarJg F. DHARMa: residual diagnosJcs for hierarchical (mulJ -level/mixed) regression 575 models. CRAN: contributed packages. 2016. 576 34. Lenth R. emmeans: EsJmated Marginal Means, aka Least -Squares Means_. R package 577 version 18 5. 2023. 578 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint 35. Douglas AE. Host benefit and the evoluJon of specializaJon in symbiosis. Heredity 579 (Edinb). 1998;81(6):599–603. 580 36. Fabina NS, Putnam HM, Franklin EC, Stat M, Gates RD. Transmission mode predicts 581 specificity and interacJon pa†erns in coral -Symbiodinium networks. 2012. 582 37. Weis VM, Reynolds WS, DeBoer MD, Krupp DA. Host-symbiont specificity during onset of 583 symbiosis between the dinoflagellates Symbiodinium spp. and planula larvae of the 584 scleracJnian coral Fungia scutaria. Coral Reefs. 2001;20(3):301 –8. 585 doi:10.1007/s003380100179 PubMed PMID: 572. 586 38. Kvennefors ECE, Leggat W, Hoegh-guldberg O, Degnan BM, Barnes AC. An ancient and 587 variable mannose-binding lecJn from the coral Acropora millepora binds both pathogens 588 and symbionts. 2008;1582–92. doi:10.1016/j.dci.2008.05.010 589 39. Vidal-dupiol J, Adjeroud M, Roger E, Foure L, Duval D, Mone Y , et al. Coral bleaching 590 under thermal stress : putaJve involvement of host / symbiont recogniJon mechanisms. 591 Vol. 16. 2009;16:1–16. doi:10.1186/1472-6793-9-14 592 40. Kuniya N, Jimbo M, Tanimoto F, Yamashita H. Possible involvement of TachylecJn - 2 - like 593 lecJn from Acropora tenuis in the process of Symbiodinium acquisiJon. Fisheries Science. 594 2015;473–83. doi:10.1007/s12562-015-0862-y 595 41. Koike K, Jimbo M, Sakai R, Kaeriyama M, Muramoto K, Ogata T, et al. Octocoral chemical 596 signaling selects and controls dinoflagellate symbionts. Biological BulleJn. 597 2004;207(2):80–6. doi:10.2307/1543582 PubMed PMID: 15501849. 598 42. Jimbo M, Koike K, Sakai R, Muramoto K, Kamiya H. Cloning and characterizaJon of a lecJn 599 from the octocoral Sinularia lochmodes. Vol. 330. 2005;330:157–62. 600 doi:10.1016/j.bbrc.2005.02.137 601 43. Taniguchi N, Kizuka Y , Takamatsu S, Miyoshi E, Gao C, Suzuki K, et al. Glyco-redox, a link 602 between oxidaJve stress and changes of glycans: lessons from research on glutathione, 603 reacJve oxygen and nitrogen species to glycobiology. Arch Biochem Biophys. 604 2016;595:72–80. 605 606 607 608 609 610 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint 611 612 613 614 615 616 617 618 619 620 621 622 623 624 Supplementary Fig. 1 | Flow cytometry ga%ng and distribu%on of Symbiodiniaceae cells. 625 RepresentaJve flow cytometry plots showing the idenJficaJon and quanJficaJon of 626 Symbiodiniaceae cells based on chlorophyll autofluorescence (top panel) and lecJn molecular 627 probe staining (bo†om panel). Right panel shows density plots of FL2-A (Chlorophyll-A 628 autofluorescence) versus FL1-A (SYBR Green fluorescence), with the gated populaJon 629 corresponding to Symbiodiniaceae cells for unstained cells (control) and stained cells (ConA). 630 Colors indicate event density from low (blue) to high (red). Lej panel shows histograms of FL1-A 631 fluorescence intensity (SYBR Green) for the gated populaJon, showing the distribuJon of lecJn 632 molecular probe signal. Axes are displayed on a log scale. 633 634 635 636 637 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint 638 Supplementary Fig. 2 | Median fluorescent intensity of Symbiodiniaceae cells. 639 A. Proportional binding profiles (total positive MFI signal) across lectins in Cladocopium (N = 640 18) and Durusdinium (N = 30) isolated from coral eggs. B. Proportional binding profiles across 641 lectins in Cladocopium (N = 15) and Durusdinium (N = 18) isolated from parental colonies. In C 642 and D, bar height represents the relative contribution of each lectin to the total binding signal per 643 symbiont genus. C. Proportional binding profiles across lectins per host genotype in 644 Cladocopium (N = 18) isolated from coral eggs. D. Proportional binding profiles across lectins 645 per host genotype in Durusdinium (N =30) isolated from coral eggs. In E and F, bar height 646 represents the relative contribution of each lectin to the total binding signal per genotype. 647 Lectins: ConA (concanavalin A, specific for D-mannose and D-glucose), LTL (Lotus 648 tetragonolobus lectin, specific for L-fucose), PNA (Arachis hypogaea lectin, specific for D-649 galactose), WGA (wheat germ agglutinin, specific for N-acetylglucosamine and N-650 acetylneuraminic acid), PHA-L (phytohemagglutinin-L from Phaseolus vulgaris, specific for N-651 acetylglucosamine β(1-2) mannopyranosyl) and GS-IB4 (isolectin from Griffonia simplicifolia, 652 specific for N-acetyl-D-galactosamine and a-D-galactosyl residues). 653 654 655 0% 25% 50% 75% 100% Cladocopium Durusdinium Symbiont Eggs total MFI A 0% 25% 50% 75% 100% Cladocopium Durusdinium Symbiont Parents total MFI B 0% 25% 50% 75% 100% 162 176 177 190 214 236 Cladocopium genotypes Eggs total MFI C 0% 25% 50% 75% 100% 169 185 197 203 208 228 231 243 249 672 Durusdinium genotypes Eggs total MFI D Lectin ConA GSIB4 LTL PHAL PNA WGA .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint Genotype N Replicates Percentage Cladocopium Dominant Genus 162 4 100.000% Cladocopium 176 3 100.000% Cladocopium 177 5 100.000% Cladocopium 190 4 100.000% Cladocopium 214 3 100.000% Cladocopium 236 3 0.002% Cladocopium 169 3 0.001% Durusdinium 185 4 0.268% Durusdinium 197 4 0.000% Durusdinium 203 3 0.087% Durusdinium 208 3 0.006% Durusdinium 228 3 0.000% Durusdinium 231 3 0.000% Durusdinium 243 3 0.000% Durusdinium 249 3 0.000% Durusdinium 672 3 0.000% Durusdinium 656 Supplementary Table 1 | Symbiont densiJes in M. capitata colonies. 657 658 659 660 661 662 663 664 665 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint PNA LTL PHAL WGA GSIB4 ConA Eggs P =0.001*** β> 0.05 -2 0 2 -2.5 0.0 2.5 5.0 PC1 (57%) PC2 (28%) Cladocopium Durusdinium A ConA LTL PNA WGA PHAL GSIB4 Parents P =0.001*** β= 0.043* -3 0 3 -4 -2 0 2 PC1 (47%) PC2 (35%) B PNA LTL PHAL WGA GSIB4 ConA C l ad ocopi um P =0.001*** β= 0.005** -2 0 2 -4 0 4 PC1 (65%) PC2 (31%) Egg Parent C PNA LTL PHAL WGA GSIB4 ConA D urusd i ni um P =0.001*** β= 0.001***-5.0 -2.5 0.0 2.5 5.0 -4 -2 0 2 PC1 (55%) PC2 (39%) Egg Parent D More Important for Eggs More Important for Adults 0.0 0.1 0.2 0.3 0.4 0.0 0.1 0.2 0.3 0.4 Adult Importance Egg Importance ConA GSIB4 LTL PHAL PNA WGA E .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint **** ** * **** lectin*symbiont<0.001 lectin*symbiont*genotype<0.001 0e+00 1e+06 2e+06 3e+06 4e+06 5e+06 ConA GSIB4 LTL PHAL PNA WGA Lectin Eggs MFI (raw values) Cladocopium Durusdinium A * * * lectin*symbiont<0.001 lectin*symbiont*genotype<0.001 0.0e+00 5.0e+05 1.0e+06 1.5e+06 2.0e+06 ConA GSIB4 LTL PHAL PNA WGA Lectin Parents MFI (raw values) B .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint *** ****** *** ****** ****** * *** *** *** *** ** *** ****** *** *** *** PNA GSIB4LTL PHAL ConA WGA WGA ConA PHAL LTL GSIB4 PNA Cladocopium Upper triangle: eggs | Lower triangle: parents A *** *** *** *** * *** *** ** *** ***** ** * ****** *** PNA GSIB4LTL PHAL ConA WGA WGA ConA PHAL LTL GSIB4 PNA Durusdinium Upper triangle: eggs | Lower triangle: parents B Pearson r -1.0 -0.5 0.0 0.5 1.0 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint ** **** Cladocopium Durusdinium POP Other POP Other 2 4 6 Parental Origin Mahalanobis Distance A 0.024* 0.538 0.325 0.331 0.846 0.239 0.611 0.3 0.341 0.029* 0.722 0.328 ConA GSIB4 LTL PHAL PNA WGA CladocopiumDurusdinium -3 -2 -1 0 1 2 -3 -2 -1 0 1 2 -3 -2 -1 0 1 2 -3 -2 -1 0 1 2 -3 -2 -1 0 1 2 -3 -2 -1 0 1 2 -2 -1 0 1 2 -2 -1 0 1 2 Scaled Egg Intensity Scaled Parent Intensity B .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted April 24, 2026. ; https://doi.org/10.64898/2026.04.21.719984doi: bioRxiv preprint