key: cord-0430703-l1xfzxpr
authors: Costa, Vincenzo A.; Bellwood, David R.; Mifsud, Jonathon C.O.; Geoghegan, Jemma L.; Holmes, Edward C.; Harvey, Erin
title: Limited Cross-Species Virus Transmission in a Spatially Restricted Coral Reef Fish Community
date: 2022-05-18
journal: bioRxiv
DOI: 10.1101/2022.05.17.492384
sha: 5e5b39dbd385166a478b3876cbce6477100bc94c
doc_id: 430703
cord_uid: l1xfzxpr

The Great Barrier Reef (GBR) – the largest coral reef ecosystem in the world – supports over 1200 fish species with some of the highest population densities and diversities seen in vertebrates, offering a high potential for virus transmission among species. As such, the GBR represents an exceptional natural ecosystem to determine the impact of host community diversity on virus evolution and emergence. In recent decades the GBR has also experienced significant threats of extinction, making it one of the most vulnerable ecosystems on the planet. However, our understanding of virus diversity and connectivity in tropical reef fishes remains poor. Here, we employed metatranscriptomic sequencing to reveal the viromes of 61 reef fish species. This identified a total of 132 viruses, 38 of which were vertebrate-associated and therefore likely infecting the fish, including a novel isolate of Santee-cooper ranavirus (Iridoviridae). Notably, we found little evidence for virus transmission between fish species living within a very restricted geographical space – a 100 m2 coral reef ecosystem – suggesting that there might be important host genetic barriers to successful cross-species transmission despite regular exposure. We also identified differences in virome composition between reef fish families, such that cryptobenthic reef fishes – characterized by small body sizes and short life-spans – exhibited greater virome richness compared to large reef fishes. This study suggests that there are important barriers to cross-species transmission, and that successful emergence in a reef fish community likely requires active host adaptation, even among closely related host species.

We discovered a basal group of four parvoviruses that fell within the subfamily Parvovirinae 281 (Parvoviridae). The genome of Ecsensius stictus parvovirus was a single contig of 4286 nt, 282 containing two main ORFs. The left ORF was 1374 nt and encoded the conserved non-283 structural protein NS1 that has DNA helicase and ATPase function [65] . The right ORF 284 encoded the structural VP1 protein (2571 bp). I. ornatus parvovirus exhibited a genome of 285 4284 nt and contained three main ORFs, one encoding NS1 (1878 nt) and the other two, 286 encoding putative structural proteins (Figure 4 ). 287 We also identified the partial genome of a novel icthamaparvovirus in L. lupus and a novel 288 circovirus in Abudefduf bengalensis that clustered within a distinct clade of fish circoviruses. reptiles [52, 53, 54, 55] . ENV was the closest relative across all proteins identified 302 (Supplementary Table 2 ). 304 Despite the large number of viruses identified, we only found evidence for one cross-species 305 transmission within our GBR ecosystem. This involved astroviruses found in five different fish species that exhibited high levels of amino acid sequence similarity and phylogenetic 307 clustering ( Figure 3) . Specifically, phylogenetic comparisons of ORF1b (RdRp) revealed two 308 viral species: goby astrovirus 1, identified in I. goldmanni, and goby astrovirus 2, identified in 309 I. nigroocellatus, I. decoratus, Asterropteryx semipunctatus, and Cabillus tongarevae ( Figure   310 4). These two viruses exhibit 82.5% amino acid sequence similarity across the virus 311 genome, while the four sequences of goby astrovirus 2 had 96.8% amino acid similarity 312 ( Figure 3 ). 313 We also detected related viruses (i.e., those from the same virus family) in several different 314 fish species, including the Hantaviridae, Rhabdoviridae, Paramyxoviridae, and 315 Picornaviridae ( Figure 6 ). However, most of these were highly divergent and likely reflect 316 common ancestry rather than direct host-jumping in the reef ecosystem. For example, out of 317 the eight picornaviruses identified in this study, the closest relatives were Asterropteryx 318 spinosa picornavirus and A. semipunctatus picornavirus that shared only 44% amino acid 319 similarity. 320 Also of note was the identification of viral species not directly infecting the fish themselves, 321 but rather associated with the local environment, diet or microbiome (i.e., non-fish) that were 322 transmitted between reef fish assemblages. These were quenyaviruses (95% amino acid 323 similarity between Pomacentrus moluccensis and A. bengalensis), flavi-like viruses (91.2% 324 between I. goldmanni and I. ornatus), narnaviruses (97.7% between I. goldmanni, I. 325 nigroocellatus and I. rigillus) and totiviruses (93.5% between Amblygobius buanesis and 326 Amblygobius rainfordi) (Supplementary figs 1, 2 and 4; Figure 6 ). That there were more 327 instances of cross-species transmission of non-fish viruses compared to those viruses that 328 actively replicate in fish suggests that the latter group are subject to strong host barriers, 329 even among closely related species.

Comparisons of viral alpha and beta diversity between reef fish families 331 We next compared vertebrate virome composition between reef fish families, as well as 332 between cryptobenthic reef fishes and large reef fishes (i.e. that differ in size). In our data set, cryptobenthic reef fish families included the Gobiidae, Apogonidae, Blenniidae, Three statistical measures were used to assess alpha diversity: viral abundance (i.e., 338 standardised number of viral reads), observed viral richness (i.e., the number of viruses) and 339 Shannon diversity. Notably, we found an association between fish size and observed viral 340 richness, with cryptobenthic reef fishes harbouring more viruses than large reef fishes (χ 2 341 = 2.795, df = 1, P = 0.028). In particular, the Tripterygiidae exhibited greater observed viral 342 richness than all other reef fish families (χ 2 = 16.678, df = 15, P = 0.007). However, we 343 found no association between reef fish family and viral abundance (P = 0.153), Shannon 344 diversity (p = 0.901) or beta diversity (R 2 = 0.334, P = 0.064). Likewise, we identified no 345 significant differences in viral abundance (P = 0.271), Shannon diversity (P = 0.142) nor beta 346 diversity (R 2 = 0.048, P = 0.121) between cryptobenthic reef fishes and large reef fishes.

As a form of internal control, we repeated our analyses of viral abundance and diversity on 348 the non-fish viruses identified here. This analysis revealed no significant differences in viral 349 abundance between reef fish families (P = 0.994) nor between cryptobenthic reef fishes and 350 larger reef fishes (P = 0.355). Similarly, we found no significant difference in observed viral 351 richness between fish families (P = 0.733), although cryptobenthic reef fishes exhibited 352 higher observed non-vertebrate viral richness than large reef fish families (χ 2 = 10.805, 353 df = 1, P = 0.016). We observed no difference in Shannon diversity between fish families (P = 354 0.453), as well as between cryptobenthic reef fishes and larger reef fishes (P = 0.070).

Finally, we found no association between beta diversity and reef fish families (R 2 = 0.279, P 356 = 0.126) nor between cryptobenthic reef fishes and large reef fishes (R 2 = 0.335, P = 0.058).

The GBR supports over 1200 species of fish and is the largest coral reef ecosystem in the suggests that there may be important host genetic barriers to virus switching among the reef 376 fishes sampled herein. This is supported by the observation that non-fish viruses -that are 377 not impacted by aspects of host genetics -were characterized by higher levels of cross-378 species virus transmission. Although currently of unknown nature, these barriers are likely to 379 be subtle and may reflect nuances in host cell receptor binding [56] . For instance, although 380 the betacoronavirus RaTG13 sampled from Rhinolophus affinis bats is closely related (~96% 381 sequence similarity) to SARS-CoV-2, it is unable to bind to the human ACE2 receptor [73] .

Another notable result was that we detected differences in vertebrate virome composition 383 between reef fish families, with cryptobenthic reef fishes harbouring more viruses than large 384 reef fishes (Figure 7 ). Due to their small body size, cryptobenthic reef fishes exhibit 385 significantly higher rates of metabolism compared to large reef fishes, resulting in high energy demands with a low tolerance for starvation [10] . Given their extremely short 

Guangdong catfish astro like virus AVM87132

Roe deer astrovirus QHW11790

Avian astrovirus

Bleniella astrovirus Chicken astrovirus NP 620617

Wenling japanese topeshark astrovirus AVM87192

Feline astrovirus

Feline astrovirus 2 YP 008519301.1 Goby astrovirus 3

Chinese pond turtle astro like virus AVM87480

Goby astrovirus 2 Istigobius decoratus Dongtou red stingray astrovirus AVM87143

Dongbei arctic lamprey astrovirus 1 AVM87496

Zhejiang chinese fire belly newt astrovirus AVM87170

Hainan black spectacled toad astrovirus 2 AVM87164

Wenling rattails astrovirus 1 AVM87150

Yellow striped sandperch astrovirus AVM87184

Western African lungfish astro like virus AVM87136

Bovine astrovirus

Beihai mudskipper astro like virus AVM87121

Goby astrovirus 2 Cabillus tongarevae Wenling plagiopsetta astrovirus AVM87176

Longspine snipefish astrovirus AVM87174

Wenling righteye flounders astrovirus AVM87607

Wuhan astro like virus AVM87125

Goby astrovirus 2 Asterropteryx semipunctatus Wenling gobies fish astro like virus AVM87598

Murray Darling rainbowfish astrovirus

Zhejiang gunthers frog astrovirus AVM87611.1 Wenling pterygotrigla hemisticta astrovirus AVM87158

Porcine astrovirus ADO30543