Differential expression analysis of the developmental CEL-Seq dataset

2.3.5 Differential expression analysis of the developmental CEL-Seq dataset

The developmental CEL-Seq dataset (Anavy et al., 2014) was used to investigate whether the shift in cell behavior (i.e. increase in proliferation and choanocyte chamber formation) could be detected between stages with and without choanocytes during metamorphosis. Pairwise differential expression analysis was conducted on the 23-24 h stage and chamber stage samples using edgeR (McCarthy et al., 2012; Robinson et al., 2010) on Galaxy Queensland (Genomics Virtual Laboratory 4.0) (Afgan et al., 2015). Default values were used for most parameters: non-differential contig count quantile threshold was set to 0.3, prior degrees of freedom was set to 10, using an FDR control method (Benjamini and Hochberg, 1995) with a p-value threshold for FDR filtering for family-wise error rate control of 0.05. Principal component analysis (PCA) was conducted on R. The Gene Ontology (GO) category enrichment analysis was conducted using BiNGO (Maere et al., 2005), and the result of this analysis was summarized and visualized using REVIGO (Supek et al., 2011). 2.4 Results 2.4.1 Changes in ciliation patterns during metamorphosis

One of the distinct morphological features of choanocytes is the apical flagellum or cilium (Figure 2.1). To visualize ciliated cells and to constrain the timing of choanocyte chamber formation during metamorphosis, fixed larvae and postlarvae were labeled with an anti-acetylated tubulin antibody, which recognizes ciliary microtubules, and DAPI, which binds to DNA and thus labels nuclei (Figure 2.2). Ciliation occurs in columnar epithelial cells and flask cells in the competent larvae, consistent with previous observations using transmission electron microscopy (TEM) (Figure 2.2A; Leys and Degnan, 2002). In the early phase of metamorphosis, the cilia in most if not all cells appear to be partially resorbed, such that the length of the cilia appear much shorter than in the larvae (Figure 2.2B), as previously observed by TEM (Leys and Degnan, 2002). These cells with resorbed cilia are enriched towards the center of the postlarva (Figure 2.2C). By 24 hpr, cells containing resorbed cilia have spread out to the edge of the postlarval body (Figure 2.2D). By 48 hpr, the first choanocyte chambers are visible while there are still many cells containing resorbed cilia not associated with forming chambers (Figure 2.2E). These newly formed choanocyte chambers

are generally smaller and more uniform in size compared to juvenile and adult chambers. At 72 hpr, much larger chambers are detected, with adjacent chambers appearing to be interconnected (Figure 2.2F). There are fewer individual cells with resorbed cilia present at 72 hpr compared to earlier stages. 2.4.2 Multiple larval cell types including larval epithelial cells contribute to postlarval and juvenile choanocyte chambers As previous studies have shown that ciliated larval epithelial cells generate a range of juvenile cell types including choanocytes (Leys and Degnan, 2002; Nakanishi et al., 2014), we sought to determine the specific contribution of larval epithelial cells to juvenile choanocyte chambers in A. queenslandica. Larval epithelial cells were labeled with CM-DiI and were traced through metamorphosis to investigate the clonality of postlarval choanocyte chambers. Only the epithelial layer has CM-DiI labeled cells (Figure 2.3A), with cilia clearly labeled by CM-DiI as well (Figure 2.4B; Nakanishi et al., 2014). CM-DiI does not label all larval epithelial cell types, such as the epithelial globular (spherulous) cells, which in Figure 2.3B are labeled by phallacidin. As such, the CM-DiI pulse-chase treatment used here follows only a subset of cells in the epithelial layer through metamorphosis, namely the ciliated columnar epithelial cells and the flask cells (Nakanishi et al., 2014).

By 24 hpr, CM-DiI labeled cells are no longer localized in the outer region of the postlarvae. The most abundant CM-DiI-labeled cell type is an archeocyte-like cell (Figure 2.3C, D), which has an amoeboid shape and a large nucleolus, often possessing putative apoptotic nuclear fragments as observed in previous studies (Nakanishi et al., 2014). Indeed by 12 hpr there is no evidence of CM-DiI labeled larval ciliated epithelial cells; thus, it appears that most ciliated epithelial cells had rapidly transformed into these archeocyte-like cells within a few hours of the initiation of metamorphosis. Only a subset of cells in early postlarvae are labeled with CM-DiI (Figure 2.3C, D) and these all have an archeocyte-like morphology. Based on the size of nuclei and large nucleolus, many unlabeled cells are likely to be archeocyte-like. Some of these are presumably derived directly from larval archeocytes and not a transdifferentiation event (Nakanishi et al., 2014).

In late postlarvae (48-72 hpr), CM-DiI labeled choanocyte chambers can be observed (Figure 2.4). As no choanocyte-like cell types labeled with CM-DiI are detected in earlier

stages of metamorphosis, these choanocytes appear to have differentiated from archeocyte- like cells prevalent in earlier stages of metamorphosis (Figure 2.3C, D; Nakanishi et al., 2014). Moreover, while there are choanocyte chambers fully labeled with CM-DiI (Figure 2.4A), only a portion of choanocytes is CM-DiI labeled in the majority of both newly formed small chambers (Figure 2.4B) and larger more mature chambers (Figure 2.4C). Although the origin of these unlabeled cells cannot be determined, this indicates that choanocytes in a given chamber originate from more than one larval cell lineage, resulting in a non-clonal choanocyte chamber. Partially labeled choanocyte chambers indicate two possible scenarios: (1) initial chamber formation involving choanocyte precursors of multiple origins, some of which were CM-DiI labeled, or (2) recruitment of CM-DiI labeled choanocytes after initial chamber formation. The occasional presence of a single CM-DiI-labeled choanocyte not integrated into chambers (Figure 2.4D) is consistent with the latter scenario, with these cells being future recruits into existing chambers. Completely unlabeled choanocyte chambers are also abundant in postlarvae and juveniles (Figure 2.4D), consistent with non-labeled larval epithelial cells or non-epithelial larval lineages such as the larval archeocytes contributing to the postlarval/juvenile choanocyte population, as previously documented in (Nakanishi et al., 2014). 2.4.3 Clusters of choanocyte precursors appear by 30 hpr Cell division in choanocyte chambers was visualized by using an anti-phospho-histone H3 (PH3) antibody that labels mitotic cells (Goto et al., 1999), and anti-acetylated tubulin antibody to label cilia of choanocytes (Figure 2.5). To identify periods and regions of high cell proliferation during postlarval development, we then used sequential 6-h incubation windows of EdU during metamorphosis. A large proportion of cell proliferation detected by this method localized to regions where choanocyte chambers appear to be forming (Figures 2.6, 2.7). In the first 24 h of metamorphosis (24 hpr), cell proliferation appears to be rare based on the number of nuclei labeled with EdU (Figure 2.6A & B). Most of the EdU-labeled nuclei are large, consistent with archeocyte-like cells being the predominant proliferating cell type at this stage; there is no evidence of choanocytes or choanocyte chambers at 24 hpr. Between 24 and 30 hpr, there is a large increase in proliferating cells (Figure 2.6C & D), with many EdU-labeled cells clustered together in a sphere-like pattern similar to a primary chamber. Although ciliation within these clusters was not observed at this stage, these labeled nuclei are markedly smaller than those observed in the archeocyte-like cells and more similar

to choanocytes. Between 30 and 36 hpr, the proliferating cell clusters are forming a sphere (Figure 2.7A) with some ciliation occurring inside these clusters (Figure 2.7B). These localized clusters of EdU-labeled cells appear to be choanocyte chamber progenitors and suggest the clusters of proliferative cells observed at 30 hpr are early-forming choanocyte chambers. 2.4.4 Differential expression analysis of the developmental CEL-Seq dataset shows enrichment in proliferation- and putative choanocyte-related genes As choanocyte chambers start appearing from 30 hpr, differential expression analysis was conducted using the developmental CEL-Seq dataset (Hashimshony et al., 2012; Anavy et al., 2014; Levin et al., 2016), to investigate whether the emergence of choanocyte chambers could be transcriptionally detected. From the 17 developmental stages in this dataset, 23-24 h and chamber stage (between 30-48 hpr) were chosen for pairwise comparison, as these stages represent time points before and after the emergence of the first choanocyte chambers during metamorphosis. Figure 2.8 shows the PCA plot of the analysis, demonstrating a clear difference in transcriptomic profiles between the 23-24h and chamber stage. There were 1,410 genes differentially expressed (p-value <0.05), with 932 of those genes significantly upregulated in the chamber stage (for full list of genes differentially upregulated in the chamber stage with annotations, see Appendix 2.1). The Gene Ontology (GO) category enrichment analysis of these 932 genes (for full list see Appendix 2.2) demonstrated an enrichment of genes in line with the metamorphic processes and cellular changes observed in this study (Figure 2.9). Enrichment in GO terms related to nucleotide biosynthesis corresponds with the increase in proliferating cells from 30 hpr observed during metamorphosis (Figure 2.6). The anatomical structure development supercluster is comprised of GO terms that are related to cilia (e.g. GO:0005929 cilium, GO:0032421 stereocilium bundle, GO:0015631 tubulin binding, etc.), representing the emergence of choanocytes at this stage, supporting the observational studies (Figures 2.2, 2.6, 2.7).

2.4.5 Different rates of proliferation occur in juvenile choanocyte chambers

Metamorphosis is complete at about 72 hpr when a functional aquiferous system is established (Degnan et al., 2015). At this stage there are numerous choanocyte chambers of varying size and shape (Figure 2.10). In general, most newly forming or early-stage