Major loss of coralline algal diversity in response to ocean acidification

Calcified coralline algae are ecologically important in rocky habitats in the marine photic zone worldwide and there is growing concern that ocean acidification will severely impact them. Laboratory studies of these algae in simulated ocean acidification conditions have revealed wide variability in growth, photosynthesis and calcification responses, making it difficult to assess their future biodiversity, abundance and contribution to ecosystem function. Here, we apply molecular systematic tools to assess the impact of natural gradients in seawater carbonate chemistry on the biodiversity of coralline algae in the Mediterranean and the NW Pacific, link this to their evolutionary history and evaluate their potential future biodiversity and abundance. We found a decrease in the taxonomic diversity of coralline algae with increasing acidification with more than half of the species lost in high pCO2 conditions. Sporolithales is the oldest order (Lower Cretaceous) and diversified when ocean chemistry favoured low Mg calcite deposition; it is less diverse today and was the most sensitive to ocean acidification. Corallinales were also reduced in cover and diversity but several species survived at high pCO2; it is the most recent order of coralline algae and originated when ocean chemistry favoured aragonite and high Mg calcite deposition. The sharp decline in cover and thickness of coralline algal carbonate deposits at high pCO2 highlighted their lower fitness in response to ocean acidification. Reductions in CO2 emissions are needed to limit the risk of losing coralline algal diversity.


| INTRODUC TI ON
When atmospheric CO 2 levels rise increasing amounts of this gas dissolve in seawater causing ocean acidification. This acidification can cause seawater carbonate saturation to fall below levels suitable for the biogenic construction of calcareous reefs (Albright et al., 2018), as confirmed by the fossil record from before and after ocean acidification events (Hönisch et al., 2012). For example, volcanic activity caused a quadrupling of atmospheric CO 2 levels 201 million years ago (Mya) which acidified the ocean and triggered the extinction of around 80% of all living species on Earth. After ocean acidification events, weathering of rocks on land slowly increased the carbonate saturation state of the ocean, allowing calcareous organisms to diversify over timescales of millions of years.
Volcanic CO 2 release 55.6 Mya is the closest geological analog to anthropogenic ocean acidification: Although it was slower than the present day rate of ocean acidification it led to major declines in the diversity and abundance of marine calcified organisms (Haynes & Hönisch, 2020).
Coralline algae are the only group of seaweeds that deposit calcite within their cell walls (Hurd et al., 2014), and thanks to these mineral cell walls they have an extensive fossil record (Aguirre et al., 2000). As calcium carbonate 'biofactories', they act as ecosystem engineers (Ballesteros, 2006;Nelson, 2009;Peña et al., 2021;Riosmena-Rodríguez et al., 2017) from the intertidal to deep water (down to 265 m) and from the tropics to the poles (Amado-Filho et al., 2012;Littler et al., 1985). They stimulate the settlement and metamorphosis of many invertebrates, including commercially important species such as lobsters, scallops, sea urchins and abalone (Huggett et al., 2006;Nelson, 2009). Loss of coralline algae simplifies coastal ecosystems (Harvey et al., 2021;Kroeker et al., 2013) and has a negative impact on ocean health and ecosystems services ). An understanding of which coralline algae will be able to survive the current ocean acidification event is lacking, but needed given their critical role in coastal ecosystems.
Laboratory and field studies have shown that coralline algal recruitment, growth, skeletal strength and survival are generally negatively affected by increased CO 2 (McCoy & Kamenos, 2015;Smith et al., 2020). Observations along natural gradients of increasing CO 2 around shallow water volcanic seep systems show that many coralline algae are vulnerable to ocean acidification but a lack of taxonomic information in these observations is a key knowledge gap (Agostini et al., 2018;Fabricius et al., 2015;Kamenos et al., 2016;Martin et al., 2008;Porzio et al., 2011). Some coralline algae are resilient to ocean acidification which seems to be partly down to the conditions that they live in. For example, intertidal coralline algae can divert energy to fight ocean acidification conditions (Bradassi et al., 2013) and the intertidal often has rapid changes in water carbonate chemistry (Wootton et al., 2008). The rate of acidification is also important, as some coralline algae can tolerate gradual but not rapid change (Kamenos et al., 2013). Coralline algae are mostly long-lived organisms (Halfar et al., 2000(Halfar et al., , 2011; but some thin coralline algal crusts can grow and reproduce quickly and build resilience to ocean acidification conditions within a few generations . Our key question is 'Does ocean acidification change the diversity of coralline algal communities?' and some clues lay in their fossil record. The Sporolithales, Hapalidiales, Corallinales and the Corallinapetrales are fully calcified orders of coralline algae (Jeong et al., 2021;Le Gall et al., 2010;Nelson et al., 2015;Peña et al., 2020). The Sporolithales is the oldest of these and first appeared ca. 137 Mya, in the Lower Cretaceous (Peña et al., 2020). In the Cretaceous, surface seawater carbonate saturation levels were high and marine life with calcareous shells and skeletons proliferated. This is when the Hapalidiales originated (ca. 116 Mya) followed by the Corallinales (ca. 112 Mya). A meteor strike that killed most dinosaurs 66 Mya caused ocean acidification which killed an estimated 67% of coralline algal species (Aguirre et al., 2000;Henehan et al., 2019). After this mass extinction event Sporolithales diversity remained low, whereas the Hapalidiales diversified (Aguirre et al., 2000) and then, during an increase in tropical coral reefs worldwide (ca. 28-12 Mya), the Corallinales became highly diverse with ca. 600 species alive today (Gabrielson et al., 2018;Guiry & Guiry, 2021;Peña et al., 2019;Rösler et al., 2016).
Research into the effects of ocean acidification on coralline algae has so far relied on the use of morphological characteristics for their identification, likely underestimating the impacts of ocean acidification on their diversity. Molecular systematics show that coralline algae have high levels of cryptic diversity globally (Gabrielson et al., 2018;Pardo et al., 2014;Pezzolesi et al., 2019) and are morphologically very variable and so are difficult to identify without DNA-based methods (Carro et al., 2014;Peña et al., 2021;Sissini et al., 2014).
Here, we used molecular systematic tools to evaluate the impact of ocean acidification on the biodiversity of coralline algae along gradients of increasing seawater CO 2 in the Pacific and Mediterranean basins to assess the capacity of this important algal group to resist or adapt to changing ocean conditions.

| Study sites and carbonate chemistry
We used natural gradients in seawater carbonate chemistry at seabed CO 2 seeps off the volcanic coasts of Italy (Vulcano Island, Mediterranean Sea, 34°19′N, 139°12′E) and Japan (Shikine Island, North-Western Pacific, 38°25′N, 14°57′E). Physicochemical surveys have established these locations as natural analogues for the future effects of ocean acidification, as long as care is taken to avoid confounding factors (Agostini et al., 2015;Boatta et al., 2013).
Carbonate chemistry parameters for both Shikine and Vulcano were calculated using the CO 2 SYS software (Pierrot et al., 2006). Measured pH, total alkalinity, temperature and salinity were used as the input variables, alongside the disassociation constants from Mehrbach et al. (1973), as adjusted by Dickson and Millero (1987), KSO 4 using Dickson (1990), and total borate concentrations (Table 1).

| Sampling and data collection
In the Mediterranean, intertidal and subtidal bedrock at four CO 2 levels were surveyed and sampled in May 2014 (Table 1) and (iv) 'Very high CO 2 ' (mean pCO 2 1012 ± 139 [SD] μatm).
Intertidal bedrock was surveyed and sampled using 25 × 25 cm quadrats (n = 5 per site) thrown haphazardly. Coralline algal cover (%, from 0-absence-to 100%) was recorded using a 5 × 5 cm grid in the quadrats to assist these in situ estimates. Specimens were TA B L E 1 Carbonate chemistry of reference and elevated pCO 2 sites in the NW Pacific (Shikine Island) and Mediterranean (Vulcano Island). The pH T , salinity and total alkalinity (AT) were measured, others were calculated using the CO 2 SYS program. Values presented as means, with standard deviation below pCO 2 (μatm)

| Molecular identification of coralline algae
Specimens were rinsed in freshwater, air-dried and preserved in ziplock bags with silica gel. A fragment of each specimen was cleaned under a stereomicroscope, then a clean part was ground into powder for DNA extraction. Genomic DNA was extracted using a NucleoSpin® 96 Tissue kit (Macherey-Nagel, GmbH and Co. KG) following the manufacturer's protocol. The psbA locus was amplified using primer pairs: psbA-F1/psbA-R2 and psbA-F1/psbA-600R (Yoon et al., 2002), and eventually the primer pair psbA21-350F/ psbA22-350R generated for coralline algae (Anglés d'Auriac et al., 2019). The mitochondrial COI-5P fragment was PCR-amplified for some Mediterranean specimens using the primer pair Gaz-F1/ Gaz-R1 (Saunders, 2005). The thermal profile for psbA and COI-5P amplifications and PCR reactions followed Peña et al. (2015). PCR products were purified and sequenced by Genoscope ( for 10 million generations, and tree sampling every 1000 generations. Our species delimitation was contrasted with interspecific divergence-uncorrected p-distances-usually applied for psbA and COI-5P sequences generated for coralline red algae (e.g. Hind et al., 2016;Peña et al., 2015Peña et al., , 2021Pezzolesi et al., 2017Pezzolesi et al., , 2019. For species identification, sequences were compared with publicly available sequences of coralline algae from GenBank.

| Carbonate biomass and complexity of biogenic habitat
Coralline algae collected from quadrats (n = 6) at each NW Pacific site (excluding the 'Very high CO 2 ' area as it lacked corallines) were dried at 60°C, then they were weighed and decalcified using HCl.
After decalcification, samples were rinsed with distilled water, dried again and re-weighed. The CaCO 3 content in each site (as g 0.1 m −2 ) was calculated based on the weight difference before and after decalcification, the estimated area occupied by each specimen within the quadrat (625 cm 2 ) and adjusted with the percentage cover estimated for coralline algae. A further set of samples of a coralline algal species that was common at the low tide mark was collected at low water from different levels of seawater pCO 2 and examined using micro-computed tomography (micro-CT scan, Skyscan 1172 system). Tomography was used to illustrate changes in thickness and

| Statistical analyses
Differences in coralline algae cover recorded at increasing levels of seawater CO 2 were assessed using one-way ANOVA tests, or by pCO 2 ) and both of these locations had low mean aragonite seawater saturation levels with periods of aragonite undersaturation (Table 1).

| Fall in coralline algal diversity at increasing levels of CO 2
We generated 223 sequences (  Although coralline algal diversity declined with increasing ocean acidification, 15 of the coralline taxa found in 'Reference' sites (nine in the NW Pacific, six in the Mediterranean) were also found living at elevated CO 2 sites ( Figure 1; Table S2). In particular, ten species were able to survive in the 'High CO 2 ' areas although none were found at the 'Very high CO 2 ' sites. Except for two taxa of the order Hapalidiales (Melobesioideae sp.1 and Phymatolithon sp.5), the remaining species found across the CO 2 gradient belonged to the order Corallinales. By contrast, seven species collected at 'Increased' and 'High CO 2 'sites were not recorded in the 'Reference' sites; all these species corresponded to the order Corallinales, and four of them belonged to the same genus (Lithophyllum).

| Coralline algal cover, complexity of biogenic habitat and carbonate biomass
In both regions, the cover of coralline algae-expressed in %-decreased significantly as CO 2 levels rose (Figure 3; Table S3).
At community scale, in both regions, as coralline algae declined, non-calcified macroalgae were more abundant among which we can list the reds Grateloupia elata (Okamura) S. Kawaguchi & H.W.
Wang and crustose Peyssonnelia spp. and Hildenbrandia spp., the browns Cystoseira spp. sensu lato, Padina pavonica (Linnaeus) Thivy and Taonia atomaria (Woodward) J. Agardh, and the green Caulerpa spp., as well as large diatom colonies in the 'Very high CO 2 ' NW Pacific site (Figure 4). In addition to the reduction in coralline cover, the complexity of biogenic habitat created by coralline algae decreased with increasing CO 2 . Intertidal samples of Phymatolithon sp.

F I G U R E 2
Coralline algal species richness recorded at different levels of CO 2 in the NW Pacific (a) and the Mediterranean (b). Species richness recorded per CO 2 level is shown as well as the number of species found in subtidal, intertidal and both shore levels (overlap). '0' indicates the absence of coralline algae at the given site In the NW Pacific site, the coralline algal calcium carbonate per unit area showed a profound decline with increasing CO 2 ( Figure 6).
Subtidal rubble was the most impacted habitat with low amounts of coralline algal CaCO 3 at the 'Increased CO 2 ' and 'High CO 2 ' sites.

| DISCUSS ION
This is the first molecular systematic study of the effects of ocean acidification on coralline algae and shows that this approach is essential for an accurate understanding of how the diversity of coralline algae is affected by CO 2 emissions (Twist et al., 2020). Highly diverse coralline algal assemblages were simplified under elevated levels of CO 2 . Corallinales, the most recently evolved order of the coralline algae (Peña et al., 2020), was the most diverse of the orders we recorded and the most resilient to acidification, with 35% of the species present at high CO 2 . By contrast, only ca. 12% of the Hapalidiales were found in the high CO 2 sites, and the Sporolithales disappeared with increased CO 2 . We expected Sporolithales and Hapalidiales to be more resistant to ocean acidification since they first appeared in a period of Earth's history when seawater CO 2 levels were high (600-1500 ppm) whereas Corallinales originated in waters with relatively low CO 2 (400 ppm; Bergstrom et al., 2020;Hansen et al., 2013;Hönisch et al., 2012;Royer et al., 2004). A mesocosm study of six species (Bergstrom et al., 2020) has found that the Sporolithales and Hapalidiales species generally had a greater capacity for CO 2 use than the Corallinales species and related these differences with ocean pCO 2 conditions that prevailed when each group originated. An experimental study on Sporolithon has shown that recruits were more sensitive to elevated temperature/pCO 2 than adult plants (Page & Diaz-Pulido, 2020) which may explain the absence of any of these taxa at our increased CO 2 sites. Laboratory work showing the vulnerability of Sporolithon durum to ocean acidification and the resilience of Neogoniolithon (Comeau et al., 2018;Cornwall, Comeau, et al., 2017) and Chamberlainium sp. (Kim et al., 2020) confirm our field results.
A lack of accurate identification of the coralline species using molecular tools makes comparison of our results with previous studies problematic. At Mediterranean CO 2 seeps, Porzio et al. (2011) observed the replacement of thick coralline algal communities by thin crusts of Hydrolithon cruciatum-order Corallinales-at low pH sites. On settlement tiles at volcanic CO 2 seeps, the most pH tolerant taxa were found to be thin crusts of Lithoporella melobesioides, Dawsoniolithon conicum and Porolithon onkodes (order Corallinales) in Papua New Guinea (Fabricius et al., 2015), and Lithophyllum, Titanoderma (Corallinales) and Phymatolithon (Hapalidiales) in Italy (Kamenos et al., 2016). In addition, except for a few experimental studies (e.g. Bergstrom et al., 2020;Comeau et al., 2018;Cornwall, Comeau, et al., 2017) that encompassed several coralline taxa F I G U R E 4 Representative pictures of coralline algal decline at 'Reference', 'Increased CO 2 ' and 'High CO 2 ' levels in the NW Pacific and the Mediterranean. Japan intertidal (a-c) and subtidal (d-f); Italy intertidal (g-i) and subtidal Reference Increased CO 2 High CO 2 Very high CO 2 belonging to different orders, most of the literature is focused on the physiological responses of a single species, and it often pertains to the order Corallinales because of their ecological role as reef builders (e.g. Hydrolithon, Porolithon, Lithophyllum; Comeau et al., 2013;Martin et al., 2013;Scherner et al., 2016;Semesi et al., 2009). In response to ocean acidification, both acclimation and adaptation can be crucial to the survival and prevalence of organisms (Sunday et al., 2014). Despite long-term exposure to increased pCO 2 , no evidence of acclimatization of the reef coralline Lithophyllum kotschyanum has been shown (Comeau et al., 2019) whereas Cornwall et al. (2020) observed that Hydrolithon reinboldii ( to the presence of Peyssonnelia algal crusts. These observed declines in coralline algal thickness and cover are likely due to a combination of disruption to spore settlement and growth, increased energetic burden of skeletal maintenance, increased dissolution, and high sensitivity of the unpigmented algal tissues plus disruption (Bradassi et al., 2013;Diaz-Pulido et al., 2012;Kato et al., 2014;McCoy & Kamenos, 2015;Ordoñez et al., 2017;Porzio et al., 2018). It was conspicuous in the field, and well illustrated using tomography, that a thickly encrusting species (Phymatolithon sp.5) dominated the low intertidal fringe of our study region of Japan but showed severe reductions in cover and thickness with increased pCO 2 .
In the Mediterranean CO 2 vents of Ischia (Italy), Porzio et al. (2011) recorded a decrease in the diversity and abundance of coralline algae.
This loss of habitat complexity provided by coralline algae has many ecological implications. For instance, decreased size and thickness increases coralline algal susceptibility to grazing (Johnson & Carpenter, 2012;McCoy & Kamenos, 2018) and reduces their ability to compete for space with other seaweeds (Kroeker et al., 2013;Linares et al., 2015). The knock-on effects of reduced coralline algal diversity and dominance by species with a thin growth form can include: reduced reef accretion in tropical and temperate environments (Adey, 1978;Ballesteros, 2006;Fine et al., 2017;Goreau, 1963), reduced habitat provisioning for associated fauna and endolithic organisms (e.g. Tribollet & Payri, 2001), and alterations in epiphytic microbial biofilms (Huggett et al., 2018) (Schubert et al., 2019), predatory fishery (Fragkopoulou et al., 2021) and the importance of their ecosystems to socio-environmental and economic well-being (Moura et al., 2021), besides reducing CO 2 emissions, further effort should look to improve ocean health as a whole (Laffoley et al., 2020), fostering the creation of no-take marine protected areas with focus on these reef builders to enhance their resilience and survival (Sissini et al., 2020).
Here we also showed with molecular systematic tools that cor- in global biodiversity hotspots such as Japan (Tittensor et al., 2010) which remains poorly studied using this approach (except for Kato & Baba, 2019;Kato et al., 2011Kato et al., , 2013.

| CON CLUS ION
We show, for the first time, that ocean acidification can cause a decline in the biodiversity of coralline algae. We reveal an exceedingly high level of cryptic diversity in the Japanese coralline algal flora and show that the effects of ocean acidification on coralline algal diversity worldwide have previously been underestimated. It is now clear that identification using molecular systematics tools significantly advances insights into the responses of marine communities to global change. Shallow-water CO 2 seep systems in two widely separated biogeographic regions revealed consistent long-term, multigenerational assemblage shifts in the coralline algae. A decrease of coralline algal species diversity was accompanied by a major loss in seabed cover and we quantify, for the first time, the extent to which ocean acidification reduces algal carbonate accretion. The order Corallinales, in particular the genus Lithophyllum, was by far the most diverse group of taxa able to survive ocean acidification. The ability of coralline algal species to tolerate rising CO 2 levels is underpinned by what these species experience in terms of environmental variability today, as well as their evolutionary history. Ocean acidification reduces coralline algal habitat complexity and is projected to adversely affect ecosystem services, and so reductions in CO 2 emissions are needed to reduce risks to coastal ocean function.

ACK N OWLED G EM ENTS
VP acknowledges support from the postdoctoral programs Campus

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that supports the findings of this study are available in the