Two samples of A. veratra were collected from Port Cartwright, Queensland, Australia, in February 2016 and maintained at the marine laboratory at the Queensland University of Technology. In this facility, they were housed in 50 L aquarium glass tanks under controlled conditions that reflected their natural environment as detailed in [25] for details. The first A. veratra individual was aerially exposed for three hours before RNA extraction, while the second sample was kept immersed in the tank as a control. In this study, the air exposed sample was named (Air) and the control sample, named (Water) (Figure 2).

Individual sea anemones (air and water samples) were snap frozen in liquid nitrogen and stored at −80˚C until RNA extraction. The individuals were homogenised in liquid nitrogen and total RNA was extracted from the whole organisms using a Trizol/chloroform RNA extraction protocol [26] [27] . RNA quality and integrity were tested using a bioanalyzer 2100 RNA nano chip following the protocol of [28] . Library preparation was undertaken using an Illumina True-Seq stranded mRNA sample preparation kit (Illumina) and following the manufacturer’s instructions. Sequencing was performed on the Illumina NextSeq 500 platform using 150 bp paired end reads. The non-biological sequences as well as low quality reads (Q < 20) were removed using Trimmomatic [29] [30] . High-quality sequence reads with thresholds of Q > 20 were used for assembly and other downstream analyses. The raw RNA sequence reads were deposited to GenBank® at the National centre of biotechnology information (NCBI), the accession numbers are available in the supplementary file: Table S1.

The two sets of clean reads were assembled using the Trinity v2.0.6 short read de novo assembler using default settings and the stranded tag [31] . CD-Hit v.4.6.1 [32] [33] was used to remove redundant and chimeric sequences from all assemblies [28] . In addition, the core eukaryotic genes mapping approach (CEGMA v.2.5) [34] was used to assess the assembly completeness by determining the percentage of full-length sequences corresponding to 248 highly conserved eukaryotic proteins [26] .

Transcriptome annotation was conducted using the Trinotate pipeline V3.0 following method detailed as per [13] for details. Gene Ontology (GO) terms were assigned to contigs that received BLAST hits and had functional annotation information. The distribution of GO terms across Molecular Function (MF), Biological Process (BP) and Cellular Component (CC) categories were visualised in WEGO [35] as per [36] .

BOWTIE2 software v.2.2.5 was used to map reads back to the reference assembly [37] for both species in which differential gene expression analysis was undertaken. Transcript abundance was estimated using RSEM v1.2.19 [38] . Fragments per kilobase of transcript per millions (FPKM) was calculated for all transcripts. Differential gene expression (DGEs) analysis was performed using the EdgeR Bioconductor package [39] in the trinity pipeline following the scripts at (https://github.com/trinityrnaseq/trinityrnaseq/wiki/Trinity-Differential-Expression). The gene list with statistically significant differential expression based on a false discovery rate (FDR) of <0.001 and log fold change (FC) ≥ 2 were then clustered using hierarchical clustering to produce a heatmap at (https://github.com/trinityrnaseq/trinityrnaseq/wiki/Trinity-Differential-Expression). The list of differentially expressed genes was filtered to find if any of the mucin candidates were in this list.

Gene set enrichment analysis was conducted on the list of DGEs using trinity scripts at (https://github.com/trinityrnaseq/trinityrnaseq/wiki/Trinity-Differential-Expression). The results were filtered to determine if specific ontologies were over or under-represented in comparison to the overall assembly based on FDR of <0.01.

A total of 69,513,111 clean reads were assembled into 118,019 contigs. The assembly was highly complete and contained a high proportion of full-length transcripts (98.4%). More statistics are available in the supplementary file: Table S2.

Overall, 47,513 contigs returned significant BLASTx hits with a stringency of 1E × 10⁻⁵, and gene ontology terms were assigned to 35,574 contigs. The distribution of GO terms showed that the highest number of GO terms was assigned to BP, then MF and finally the least terms were assigned to CC. In the BP category, cellular process (20,728 gene) (74.5%), metabolic process (14,716 genes) (52.9%) and biological regulation (12,237 genes) (44%) were the most assigned GO terms. Binding (18,438 genes) (66.3%) and catalytic activity (11,099 genes) (39.9%) were the top assigned GO terms under the MF category. The greatest number of GO terms assigned under the CC category were; cell (22,942 genes) (82.4%), cell part (22,940 genes) (82.4%) and organelle (15,978 genes) (57.4%). WEGO plots was presented in the supplementary: Figure S1.

At the beginning of the experiment, the A. veratra oral disc and tentacles were completely closed. Then, during the first hour, the oral disc opened slightly while the tentacles were still drawn, and there was some mucus production. During the second hour, there was a clear increase of mucus covering the oral disc. During the third hour, mucus production increased again.

Overall 5686 differentially expressed sequences were found in A. veratra across water and air treatments. The 5685 sequences included 3243 genes significantly up-regulated in the water treatment, while the remaining 2443 genes were up-regulated in the air treatment. The expression level across the two samples based on (FC ≥ 2 and FDR of ≤0.001) is shown in a hierarchically clustered heat map (Figure 3). Contigs with up-regulation are indicated in yellow, the contigs with down-regulation indicated, in purple. The contigs expression pattern similarity across the samples is indicated by the three generated sub-cultures, as shown in Figure 4.

Overall, mucin5B-like, mucin4-like and mucin-like genes were differentially expressed. Specifically, the mucin5B-like and mucin4-like were differentially expressed under the aerial exposure treatment, while the mucin-like was differentially expressed in the water treatment.

Functional enrichment analysis was performed on the differentially expressed contigs, and it revealed a number of over-represented GO terms in the air and water treatments. The enrichment results based on FDR < 0.01 identified 42 statistically over-represented and 29 under-represented genes in the water sample; while in the air sample, 28 were over-represented and 7 were under-represented genes. The highest number of over-represented genes in the water and air samples was under the BP category, followed by CC in the water treatment and MF in the air treatment sections, with the lowest being the CC in the air treatment option, and MF in the water sample (Figure 5). The set of over-represented genes in the air treatment is available in the supplementary file: Table S3. Specifically, the majority of the gene ontology terms under the BP in the air treatment belonged to the defence and immune response functions. Additionally, genes belonging to “binding”, “transport”, and “developments” were over-represented in response to air exposure. On other hand, a group of extracellular genes constituted the majority of over-represented genes under the CC category in the water treatment, available in supplementary file: Table S5. However, the under-represented genes in the air treatment included those which referred to cellular activities or to the receptor pathway and activity are shown in supplementary file: Table S4. This result showed a reduction in the gene activity inside the cell when the animal body was exposed to environmental stress. In contrast, the majority of under-represented genes in the water treatment were rich in receptor activity, signal transduction and transporter genes are presented in supplementary file: Table S6. Among the over represented genes from the air treatment, the extracellular region ontology was enriched by mucin5B-like, while the calcium ion binding ontology was enriched by the mucin4-like.

The identified mucin5B-like and mucin4-like genes were the only responsive mucins to the three hours of air exposure, indicating that they may play an important role when the animal is out of the water and confronting conditions of stress such as air exposure. From this data we hypothesise that mucin5B-like and mucin4-like are the mucins that respond to this type of stress in the sea anemone A. veratra. In a previous study, a change in mucin5B-like expression was observed during development of the sea anemone species N. vectensis [24] . This expression data indicates that mucin5B-like may have different roles across ecologically divergent sea anemone species such as A. veratra and N. vectensis, or that it undertakes multiple roles in different metabolic and developmental processes. This indicates that mucin5B-like and mucin4-like may have a role in the production of mucus and protection of this sea anemone during aerial exposure.

The function of the mucin-like gene that was differently expressed in the water treatment is not clearly understood, but it may have a protective under submerged conditions. For example, in fish natural mucus secretions occur to safeguard the skin against pathogen invasion [40] . The expression of the mucin-like in the water sample could also be interpreted as mechanism to increase mucus production for a similar role as that observed in fish.

Most of the over-represented genes under the aerial treatment were related to defence and immune response functions. This indicated that the response to air exposure in A. veratra induces stress related genes, which has been observed in a number of invertebrate species under environmental stress [41] . An earlier study on N. vectensis identified three different groups of genes related to stress-response in sea anemones, including genes belonging to wound healing, genes belonging to immune response and genes belonging to chemical stresses [42] . [43] examined genome wide patterns of gene expression in Acropora hyacinthus under thermal stress and found genes belonging to these classes to induced in thermally sensitive individuals. Overall, this data indicates that aerial exposure in A. veratra is stressful, and that the induction of stress response genes may serve a protective role during this time.

The gene expression data from A. veratra is indicative of the cellular stress response (CSR), a universal cellular defence mechanism, which is initiated in response to changes in the extracellular environment [44] . Depending on the duration and severity of stress, cells either reinstate the process of cellular homeostasis to the previous state or take on a distorted state in the new and changed environment [45] . The response includes control of the cell cycle, DNA and chromatin stabilization and repair, removal of damaged proteins, protein chaperoning and repair, as well as various metabolism functions [44] . Consequently, the induction of stress response genes seen in A. veratra may be associated with the CSR and allow cells of this animal to re-establish normal function when returned to water.