Echiuran Hox genes provide new insights into the correspondence between Hox subcluster organization and collinearity pattern

Abstract

In many bilaterians, Hox genes are generally clustered along the chromosomes and expressed in spatial and temporal order. In vertebrates, the expression of Hox genes follows a whole-cluster spatio-temporal collinearity (WSTC) pattern, whereas in some invertebrates the expression of Hox genes exhibits a subcluster-level spatio-temporal collinearity pattern. In bilaterians, the diversity of collinearity patterns and the cause of collinearity differences in Hox gene expression remain poorly understood. Here, we investigate genomic organization and expression pattern of Hox genes in the echiuran worm Urechis unicinctus (Annelida, Echiura). Urechis unicinctus has a split cluster with four subclusters divided by non-Hox genes: first subcluster (Hox1 and Hox2), second subcluster (Hox3), third subcluster (Hox4, Hox5, Lox5, Antp and Lox4), fourth subcluster (Lox2 and Post2). The expression of U. unicinctus Hox genes shows a subcluster-based whole-cluster spatio-temporal collinearity (S-WSTC) pattern: the anterior-most genes in each subcluster are activated in a spatially and temporally colinear manner (reminiscent of WSTC), with the subsequent genes in each subcluster then being very similar to their respective anterior-most subcluster gene. Combining genomic organization and expression profiles of Hox genes in different invertebrate lineages, we propose that the spatio-temporal collinearity of invertebrate Hox is subcluster-based.

1. Introduction

Hox genes encode a family of transcription factors with a highly conserved 60 amino acid homeodomain. Hox genes are the master regulatory genes in anterior–posterior axial patterning in many animals [1–3]. They were first described in Drosophila melanogaster [4,5] and later identified in most bilaterians [6,7].

Hox genes are clustered in the genomes of many bilaterians. Based on the variety of Hox cluster organization, Denis Duboule categorized the Hox clusters of bilaterians into four types: organized, disorganized, split and atomized clusters [8]. The organized clusters, represented by vertebrates, are well organized, with genes tightly arranged and devoid of unrelated sequences. The disorganized clusters (e.g. the sea urchin cluster [9]) are much larger and may contain non-Hox genes. The split clusters, represented by Drosophila [10], are broken into several parts with a significant gap. The atomized clusters (e.g. the urochordate Oikopleura dioica cluster [11]) represent the no-cluster situation where the genes are scattered throughout the genome. According to Duboule's model [8], the ancestral cluster is a disorganized cluster and subsequently evolves in bi-directions: consolidation and fragmentation. In vertebrates, the Hox cluster may undergo the consolidation (eliminating the unrelated genes and reducing the intergenic distance) to evolve into the organized cluster. As to the fragmentation direction, owing to losing the regulatory constraints that keep genes together, the ancestral cluster may be split into two parts or more and finally evolves into the atomized cluster.

Based on molecular phylogenies, the Hox cluster can be subdivided into four groups: anterior group, paralogue group 3 (PG3), central group and posterior group [12,13]. The gene composition of these groups differs across three major evolutionary lineages of Bilateria: Deuterostomia, Ecdysozoa and Lophotrochozoa [12–17]. The anterior group contains two genes corresponding to Lab, Pb in ecdysozoans and Hox1, Hox2 in lophotrochozoans and deuterostomes [18]. PG3 corresponds to zen in ecdysozoans and Hox3 in lophotrochozoans and deuterostomes [19,20]. The central group generally varies from five genes in ecdysozoans (Dfd-AbdA) and deuterostomes (Hox4-Hox8) to six genes (Hox4-Lox2) in lophotrochozoans [21,22]. The posterior group contains one gene (AbdB) in ecdysozoans, two genes (Post2 and Post1) in lophotrochozoans and six to seven genes (Hox9-Hox15) in deuterostomes (despite some variation in nomenclature) [23–25].

In recent years, much attention has been paid to the relationship between the genomic arrangement and expression pattern of Hox genes. Collinearity is the correspondence between the order of Hox genes in the cluster and their spatio-temporal expression patterns. Spatial collinearity refers to the correspondence between the gene order in the cluster and the expression position along the anterior–posterior axis [26,27]. Temporal collinearity refers to the phenomenon that the anterior genes (in the 3′ part) in the cluster are activated earlier than the posterior genes (in the 5′ part) [28–31]. Temporal collinearity is often considered the main constraining force on Hox cluster organization [10,32–34], and is mainly found in vertebrates with organized Hox clusters [35,36]. In other animal lineages whose Hox clusters are split or atomized, temporal collinearity is lost [11,37–39]. However, there are some exceptions. For example, 10 Hox genes of the annelid Capitella teleta retain temporal collinearity (except Post1) [7,38–41].

Recently, Wang and his colleagues proposed a new model of subcluster-level temporal collinearity (STC) in the mollusc Patinopecten yessoensis [42]. In this model, the Hox cluster is divided into several subclusters according to their expression patterns, and genes in the same subcluster display temporal collinearity. Then, the similar expression patterns were also identified by reanalysing the published data in various bilaterians including the mollusc Crassostrea gigas, the annelid Alitta virens (formerly Nereis virens), the arthropod Litopenaeus vannamei and the urochordate Ciona intestinalis [42–46]. In this model, Wang et al. considered that the STC probably only concerns the integrity of subclusters rather than that of the whole Hox cluster. This pattern is different from whole-cluster temporal collinearity (WTC) in which all the Hox genes are expressed orderly in vertebrates. Then, a new question has arisen: why do these species exhibit this collinearity pattern?

In this study, we report the genomic organization and expression pattern of Hox genes in the echiuran worm Urechis unicinctus, an important commercial species of Annelida, Echiura. Furthermore, we performed a comprehensive analysis of the genomic organization and expression profiles of Hox genes in different species to understand the relationship between Hox subcluster organization and spatio-temporal collinearity.

2. Material and methods

(a) . Animals and sampling

Sexually mature U. unicinctus individuals with a mean fresh mass of 32.8 ± 7.4 g were collected from the coastal intertidal area in Yantai, China. Male and female worms at a proportion of 1 : 10 were dissected to obtain nephridia (gonoduct) with mature sperm and oocytes. Artificial fertilization was performed by mixing sperm and oocytes in filtered sea water (FSW) and then the fertilized eggs were raised in FSW at 17 ± 0.1°C, pH 7.9 ± 0.06 and salinity 30‰. The embryos underwent cleavage and gastrulation within 24 h and then hatched into the free-swimming trochophore larvae. The larvae were fed with single-cell algae (Chlorella vulgaris, Isochrysis galbana, Chaetoceros muelleri) at 500 cells individual⁻¹ daily and the feeding amount increased by 500 cells individual⁻¹ d⁻¹ (18 ± 0.3°C, pH 7.9 ± 0.13 and salinity 30‰). During development, the hyposphere of the trochophore elongated, and produced larval segments in segmentation larva. Then, the larval segments disappeared, and finally the larva metamorphosed into worm-shaped larva which burrowed in the sediment.

The embryos and larvae from different developmental stages were gathered, including two to eight cell embryo (3 h post fertilization, hpf), multicellular embryo (MC, 6 hpf), blastula (BL, 9 hpf), gastrula (GA, 16 hpf), early trochophore (ET, 30 hpf), mid-trochophore 1 (MT1, 2 days post fertilization, dpf), mid-trochophore 2 (MT2, 12 dpf), late-trochophore (LT, 25 dpf), early segmentation larva (ES, 30 dpf), segmentation larva (SL, 35 dpf) and worm-shaped larva (WL, 45 dpf). The parts of these samples were immediately frozen in liquid nitrogen and then stored at −80°C for total RNA extraction. Meanwhile, others were fixed in 4% paraformaldehyde for 16 h at 4°C, and dehydrated with serial methanol (25%, 50%, 75% and 100%) in phosphate-buffered saline (PBS, pH 7.4) and finally stored in 100% methanol at −30°C for whole mount in situ hybridization. All the larvae were anaesthetized with MgSO₄ saturated solution for 10 min prior to fixation.

(b) . Identification of Hox genes

Hox genes and non-Hox genes were identified in the genome and transcriptomes [47,48] of U. unicinctus using BLAST searches. The homology-based cloning of Post1 in U. unicinctus was also attempted using degenerate PCR amplification with larval cDNA as template. The degenerate primers, Post1-F: 5′-GCMCCHDCVACMGTBACRYT-3′ and Post1-R: 5′-ARYGTVACKGTBGHDGGKGC-3′ were designed according to the sequences of Post1 from Ca. teleta, P. yessoensis, Cr. gigas and Terebratalia transversa (electronic supplementary material, table S1). Multiple sequence alignment was conducted using DNAMAN. Phylogenetic analysis was performed using neighbour-joining methods with 1000 bootstrap replicates in MEGAX.

(c) . RNA extraction and quantitative PCR

Total RNA was extracted using thiocyanate–phenol–chloroform method and digested with RNase-free DNase I (TaKaRa, Dalian, China) to remove genomic DNA. The quantity and quality of total RNA were detected with a NanoPhotometer N60 (Implen GmbH, Munich, Germany) and 1.2% agarose gel electrophoresis. Total RNA was reverse transcribed into cDNA using the PrimerScript^TM RT reagent Kit (TaKaRa, Dalian, China) according to the manufacturer's instructions. qPCR analysis was performed to determine the expression level of Hox mRNA with the gene-specific primers (electronic supplementary material, table S2) designed using Primer 3. qPCR was performed using the Roche LightCycler 480 Real-Time PCR System (Roche, Basel, Switzerland) with SYBR Premix Ex Taq^TM (TaKaRa, Dalian, China). The relative expression levels were normalized to the reference gene ATPase, and expression ratios were calculated using the 2^−△△Ct method. All data were presented as mean ± standard error from three samples with three parallel repetitions. All qPCR analysis was validated in compliance with ‘the MIQE guidelines'. The heat map and bar charts were created using GraphPad Prism 8.

(d) . Whole-mount in situ hybridization

Linear templates for probe synthesis were generated by PCR using the specific primers with the T7 or Sp6 promoter sequence flanking their 5′ ends (electronic supplementary material, table S3). Digoxigenin-labelled riboprobes were transcribed in both sense and antisense directions using the DIG RNA Labelling Kit SP6/T7 (Roche, Basel, Switzerland). Embryos and larvae at the different developmental stages were rehydrated for 5 min each time in serial methanol solutions (75%, 50%, 25%) and PBT (PBS, 0.1% Tween 20). Then these samples were digested with proteinase K (2 µg ml⁻¹ in PBT) for 20 min at 37°C. These samples were fixed in 4% paraformaldehyde for 30 min. Prehybridization was conducted in a hybridization buffer (50% formamide, 5x SSC, 1.5% blocking reagent, 1 ng µl⁻¹ yeast total RNA, 0.1% Tween 20) without the probe for 5 h at 60°C, and then hybridization was performed for 16 h at 60°C by using a probe concentration of 1 ng µl⁻¹ in the hybridization buffer. The excess probe was removed by increasingly stringent washes of SSC to 0.2x SSC. After incubation in blocking buffer for 1 h at room temperature, the samples were exposed to anti-DIG-AP antibody (Roche, Basel, Switzerland) with a 1 : 2500 dilution for 16 h at 4°C, and then were stained in AP buffer (100 mM Tris, 100 mM NaCl, 5 mM MgCl₂·6H₂O, 0.5% Tween 20, pH 9.5) with the addition of NBT/BCIP at room temperature. Finally, they were dehydrated with serial ethanol (70%, 80%, 95% and 100%), transparentized with xylene and mounted with neutral balsam. The samples were observed and photographed using a Nikon E80i microscope (Nikon, Tokyo, Japan).

(e) . Statistical analysis

All data were presented as mean ± s.e. (n = 3). Significant differences between means were determined using one-way analysis of variance (ANOVA) followed by Duncan's multiple-range test with the SPSS statistical software (version 22, SPSS Inc., Chicago, IL, USA). The significance level was set at p < 0.05.

3. Results

(a) . Genomic organization of Hox genes in Urechis unicinctus

Ten Hox genes were identified in the genome and transcriptomes of U. unicinctus, which were orthologous to Hox1, Hox2, Hox3, Hox4, Hox5, Lox5, Antp, Lox4, Lox2 and Post2 in other lophotrochozoans, and the Post1 existing in many lophotrochozoans was not obtained by BLAST and PCR in U. unicinctus. The 10 Hox genes were located on two scaffolds (figure 1a). Scaffold A spanned 2.2 Mb and contained eight Hox genes (Hox1, Hox2, Hox3, Hox4, Hox5, Lox5, Antp and Lox4) which occupied a genomic region of 758 kb. Scaffold B spanned 1.3 Mb and included Lox2 and Post2 in a genomic region of 42 kb. The distance is more than 500 kb from Hox-coding areas to the end of the scaffold, suggesting the gap between Hox1-Lox4 and Lox2-Post2 is more than 1 Mb. Non-Hox genes were found among the Hox genes, such as between Hox2 and Hox3, Hox3 and Hox4 in Scaffold A, as well as on the flank of the two scaffolds (figure 1a; electronic supplementary material, table S4). Generally, the distances between adjacent Hox genes were over 30 kb. Particularly, the distance between Hox2 and Hox3 was 330 kb due to the existence of some non-Hox genes. Moreover, the Hox genes were transcribed in the same orientation in respective scaffold. Taken together, U. unicinctus might have the same property of gene order and transcriptional orientation with the representative lophotrochozoans (figure 1b). Phylogenetic analysis indicated that Hox genes in U. unicinctus were clustered into four groups: anterior group, Hox1 and Hox2; PG3, Hox3; central group, Hox4, Hox5, Lox5, Antp, Lox4 and Lox2; posterior group, Post2, which were consistent to those of the representative lophotrochozoans (electronic supplementary material, figure S1, table S1 and figure 1c).

Figure 1.

The Hox cluster of U. unicinctus. (a) Genomic organization of Hox genes in U. unicinctus. The 10 Hox genes of U. unicinctus are located on two scaffolds. The genomic regions containing Hox genes are represented in scale. Black lines depict the two scaffolds. Each colour box indicates a Hox orthologous gene and vertical lines represent non-Hox genes. Below each box are the direction of transcription (the horizontal arrows) and the transcription units (exon–intron composition). (b) Schematic diagram of Hox clusters in the representative lophotrochozoans, with the putative ancestral state of Hox clusters in the lophotrochozoan ancestor on the top. U. unicinctus, Urechis unicinctus; Ca. teleta, Capitella teleta; E. fetida, Eisenia fetida; P. yessoensis, Patinopecten yessoensis; Cr. gigas, Crassostrea gigas; T. transversa, Terebratalia transversa. Each Hox orthologous gene is represented by a specific colour. (c) Phylogenetic analysis of Hox genes from five lophotrochozoans. Tree topology is constructed with neighbour-joining methods using an alignment of the 60 amino acids of the homeodomain. Numbers above branches are supporting percentages of 1000 bootstrap replicates from neighbour-joining tree. Uu, U. unicinctus; Ct, Ca. teleta; Py, P. yessoensis; Cg, Cr. gigas; Tt, T. transversa.

(b) . Temporal expression of Hox genes in Urechis unicinctus

We performed qPCR assays to investigate temporal expression of Hox genes in different developmental stages of U. unicinctus (figure 2, electronic supplementary material, figure S2). The results showed that U. unicinctus Hox genes demonstrated significant temporal differential expression during development. All the Hox genes were not expressed in two to eight cell embryos. The anterior Hox gene Hox1 was first activated in multicellular embryos, then upregulated from mid-trochophore 2 to segmentation larva. The highest level of Hox1 expression was exhibited in segmentation larva, and finally it was downregulated in worm-shaped larva. The expression of Hox2 began also in multicellular embryos and then decreased in blastula and gastrula, increased in the hatched larvae, reached its peak in segmentation larva, and finally decreased in worm-shaped larva. Activation of Hox3 occurred first in mid-trochophore 1, then the highest expression level was apparent in segmentation larva, and finally decreased in worm-shaped larva. The central Hox genes Hox4, Hox5, Lox5, Antp and Lox4 exhibited a similar temporal expression pattern, which were initiated in mid-trochophore 2 and exhibited the highest expression level in segmentation larva, and then downregulated in worm-shaped larva. However, Lox2 and Post2 were not activated until the ES larva stage, and the highest expression levels were demonstrated in segmentation larva and then decreased in worm-shaped larva (figure 2; electronic supplementary material, figure S2).

Figure 2.

Quantitative expression of Hox genes in U. unicinctus developmental stages. Two to eight cell, two to eight cell embryo; MC, multicellular embryo; BL, blastula; GA, gastrula; ET, early trochophore; MT1, mid-trochophore 1; MT2, mid-trochophore 2; LT, late-trochophore; ES, early segmentation larva; SL, segmentation larva; WL, worm-shaped larva.

(c) . Spatial expression of Hox genes in Urechis unicinctus

To analyse spatial expression of Hox genes in U. unicinctus, we performed whole mount in situ hybridization from embryos to segmentation larvae. Specificity of the probes was confirmed by conducting controls with sense probes. No signal was detected in these controls (data not shown). The spatial expression was analysed by combing the histological observation of the U. unicinctus larvae (electronic supplementary material, figure S3).

The strong positive signal of the anterior Hox gene Hox1 was first detected in multicellular embryos (electronic supplementary material, figure S4A1), which was distributed throughout the embryo. Then the positive signal was weak from blastula to gastrula (electronic supplementary material, figure S4B1-C1). Until early trochophore, the positive signal became strong (electronic supplementary material, figure S4D1), and mainly located in the ventral ectoderm of the hyposphere (electronic supplementary material, figure S4D1-G1). In the ES larva and segmentation larva, the strong positive signals of Hox1 were presented in the dorsal and ventral ectoderm of entire hyposphere from the first segment to the 11th segment (figure 3a1–b1). Like Hox1, the second anterior Hox gene Hox2 was also expressed firstly in multicellular embryos and presented a similar expression pattern to Hox1 during the embryonic and larval development (electronic supplementary material, figure S4A2-G2 and figure 3a2–b2). Hox3 positive signal was observed for the first time in mid-trochophore 1, which was mainly located in the ventral epidermis of the hyposphere (electronic supplementary material, figure S4E3) and maintained in mid-trochophore 2 and late-trochophore (electronic supplementary material, figure S4F3–G3). In segmentation larva, the positive signal of Hox3 was exhibited in the dorsal and ventral ectoderm of larval abdominal segments from the second segment to the 11th segment (figure 3b3). The positive signals of the central Hox genes Hox4, Hox5, Lox5, Antp and Lox4 were initially detected in mid-trochophore 2 (electronic supplementary material, figure S4F4-F8), and then were obviously concentrated in the ventral ectoderm of the hyposphere in late-trochophore (electronic supplementary material, figure S4G4-G8). It was worth noting that Hox genes exhibited gene-specific expression patterns in the segmental ectoderm of ES larva (figure 3a4–a8) and segmentation larva (figure 3b4–b8). The anterior-most segment of Hox4 and Hox5 expression was the third segment, and the posterior-most segment was the 11th segment for Hox4 and the sixth segment for Hox5; Lox5 was expressed from the fourth segment to the 11th segment, while Antp and Lox4 were expressed from the sixth segment to the 11th segment. The expression of Lox2 and Post2 was not observed until ES larva (figure 3a9–a10, electronic supplementary material, figure S4), which was showed in the dorsal and ventral ectoderm of the posterior hyposphere. In segmentation larva, Lox2 was continuously expressed from the seventh segment to the 11th segment (figure 3b9), while Post2 was expressed from the eighth segment to the 11th segment (figure 3b10).

Figure 3.

Spatial expression of Hox genes in U. unicinctus from ES larvae and segmentation larvae. (a0–b0) Morphology of U. unicinctus larvae. (a1–a10) ES larvae (ventral views); (b1–b10) segmentation larvae (lateral views). Scale bars: 100 µm. The genomic organization of the Hox genes is shown on the top.

4. Discussion

(a) . Urechis unicinctus Hox cluster and subclusters

Our results revealed that U. unicinctus Hox cluster contained 10 Hox genes which were located on two scaffolds. If we only summarize Hox-coding areas, the total size of the cluster will be at least 800 kb, which is much larger than vertebrate Hox clusters (approx. 100 kb) [49] but smaller than that of the mollusc P. yessoensis (approx. 1.82 Mb) [42] (electronic supplementary material, figure S5), implying U. unicinctus Hox cluster may be more compact than that of the mollusc P. yessoensis. According to Duboule's model [8], the Hox cluster of U. unicinctus should belong to the split type because of a significant gap (more than 1 Mb) between Hox1-Lox4 and Lox2-Post2, while P. yessoensis Hox cluster is a disorganized one as these Hox genes are in a single scaffold [42].

Our results revealed that U. unicinctus Hox genes were grouped into four clades in the phylogeny: anterior, PG3, central and posterior group (figure 1c), which are consistent with other lophotrochozoans [13,17,39,40,42] and fit well with the hypothesis of the ancestral Hox cluster of the last common ancestor of bilaterians, Urbilateria [12,50–52]. In addition, the U. unicinctus Hox cluster lacks Post1 compared to the representation lophotrochozoans, such as annelids [40,44], molluscs [17,42,45] and brachiopods [19,39]. Expression analysis of brachiopods [39,53] and annelids [40,44,54] indicated that the Post1 was mainly expressed in the chaetae sacs. Urechis unicinctus only has anal chaetae rings around the posterior trunk [55]. It appears that loss of the Post1 may result from the degeneration of the anatomical arrangement of the chaetae sacs. Therefore, we propose that the Post1 of U. unicinctus undergoes secondary loss in the process of evolution.

Furthermore, our genomic information revealed that some non-Hox genes are distributed in the Hox cluster (figure 1a). These genes subdivided the Hox cluster into four subclusters: S1 (Hox1 and Hox2), S2 (Hox3), S3 (Hox4, Hox5, Lox5, Antp and Lox4) and S4 (Lox2 and Post2). Molecular phylogenetic analysis indicated that Hox cluster can be subdivided into four groups in lophotrochozoans [13,17,40]. The four subclusters subdivided by non-Hox genes in U. unicinctus were almost consistent with the results of molecular phylogenies (figure 1c) except Lox2 which belonged to central group.

(b) . The Hox expression of Urechis unicinctus exhibits a subcluster-based whole-cluster spatio-temporal collinearity pattern in the larval ectoderm

The expression analysis of U. unicinctus Hox genes revealed that all Hox genes were not expressed in two to eight cell embryos, which is consistent with the results of the annelid Ca. teleta [40] and the mollusc P. yessoensis [42]. However, another annelid, Platynereis dumerilii, displayed maternal expression pattern [56]. It remains unclear whether the maternal expression is lost in these three species or whether this maternal expression is an exception. In addition, we did not detect the expression of the sense probes, which is similar to the results of Ca. teleta [40] and P. yessoensis [42], suggesting that antisense transcripts of Hox are absent in these lophotrochozoans. Nevertheless, these results are challenged by a recent report of antisense Hox transcripts in two annelids A. virens and P. dumerilii [57]. Given this situation, more studies are required to address the antisense Hox transcripts in different species, which may provide insights into the ancestral feature of lophotrochozoans.

After analysing gene expression at multiple developmental stages, we concluded that the Hox gene expression of U. unicinctus displayed a subcluster-based whole-cluster temporal collinearity during the embryonic and larval development. The first subcluster (S1) genes Hox1 and Hox2 were activated firstly in the multicellular embryos, and then second subcluster (S2) gene Hox3 in the mid-trochophore 1, the third subcluster (S3) genes Hox4, Hox5, Lox5, Antp and Lox4 in the mid-trochophore 2, finally the fourth subcluster (S4) genes Lox2 and Post2 in the ES larva (figures 2, 3 and electronic supplementary material, figure S4). In summary, Hox genes expression in the different subclusters exhibited temporal collinearity, while Hox genes expression in the same subcluster exhibited the co-activation expression pattern.

Furthermore, the spatial collinearity of Hox gene expression was well presented in U. unicinctus epidermal ectoderm. In ES larva and segmentation larva, the anterior expression boundaries of adjacent Hox genes were staggered backward by about one segment, except for Hox2, Hox5 and Antp (figure 3). The orders of Hox gene expression in the segments of U. unicinctus larvae from anterior to posterior were consistent with the position of each Hox gene within the cluster, indicating the whole-cluster spatial collinearity. In addition, a striking finding is that spatial collinearity of Hox gene expression exhibit subcluster-based characteristics: genes in same subcluster have similar expression regions. The S1 genes Hox1 and Hox2 have the same anterior and posterior boundaries. Two S3 genes Hox4 and Hox5 have the same anterior boundary at the third segment, while the other two S3 genes Antp and Lox4 share the same anterior and posterior boundaries. Together, expression of Hox genes follows the spatial collinearity among the whole cluster and exhibits similar features in the subcluster-level.

Based on the data described above, we propose that the expression pattern of U. unicinctus Hox genes exhibits a subcluster-based whole-cluster spatio-temporal collinearity (S-WSTC) pattern in the larval ectoderm: the anterior-most genes in each subcluster are activated in a spatially and temporally colinear manner, with the subsequent genes in each subcluster then being very similar to their respective anterior-most subcluster gene.

(c) . Correspondence between Hox subcluster organization and spatio-temporal collinearity pattern

A direct link between the organized arrangement of Hox cluster and spatio-temporal collinearity pattern in vertebrates has been proposed [58]. In vertebrates, the Hox clusters are well organized and compact with a low density of extraneous sequences. Hox genes are tightly collinearly arranged in the clusters and are expressed in a whole-cluster spatio-temporal collinearity (WSTC) pattern in the mesoderm: each Hox gene is expressed in spatial and temporal order [36] (figure 4a).

Figure 4.

Different spatio-temporal collinearity patterns in three different species. (a) The whole-cluster spatio-temporal collinearity (WSTC) pattern is recorded in the mesoderm of vertebrates which possess the organized Hox clusters [36]. (b) Hox cluster of U. unicinctus should be a split type and it can be divided into four subclusters, and the Hox expression exhibits a subcluster-based whole-cluster spatio-temporal collinearity (S-WSTC) pattern in the ectoderm. (c) The Hox expression of P. yessoensis with disorganized cluster contains four subclusters exhibits a subcluster-level spatio-temporal collinearity (SSTC) pattern in the ectoderm [42]. Red diagonal arrows indicate gene activation order. Black vertical arrows indicate co-activation of leading genes of four subclusters in P. yessoensis. The expression regions of Hox genes are represented by different colours.

Notably, Hox genes of invertebrates possess distinctly different cluster organization and collinearity patterns compared with vertebrates. The mollusc P. yessoensis has a disorganized Hox cluster, with four subclusters including S1 (Hox1, Hox2 and Hox3), S2 (Hox4 and Hox5), S3 (Lox5, Antp, Lox4 and Lox2) and S4 (Post2 and Post1) [42]. The leading gene in each Hox subcluster is expressed simultaneously, followed by sequential expression of their subcluster followers. Their spatial expression of four leading genes also follows spatial collinearity in the ectoderm. This pattern was called subcluster-level spatio-temporal collinearity (SSTC; figure 4c). Likewise, the Hox genes of A. virens and P. dumerilii display the similar SSTC pattern: the first gene in each Hox subcluster Hox1, Hox4, Lox5 and Post2 is expressed simultaneously, and the expression of each gene in the same subcluster follows temporal collinearity; Meanwhile, Hox1, Hox4, Lox5 and Post2 are expressed in spatial collinearity in the ectoderm [44].

Obviously, the collinearity pattern from the U. unicinctus Hox cluster is different from the WSTC pattern in vertebrates [35,36] and the SSTC pattern in molluscs [42] (figure 4b). In the split Hox cluster of U. unicinctus, the collinearity pattern was subcluster-based: not only subcluster-based whole-cluster temporal collinearity pattern, but also subcluster-based whole-cluster spatial collinearity pattern in the larval ectoderm. The similar Hox subcluster classification and expression collinearity pattern have been revealed in Ca. teleta [40]. The Ca. teleta Hox cluster comprises four subclusters: S1 (Hox1, Hox2 and Hox3), S2 (Hox4 and Hox5), S3 (Lox5, Antp and Lox4) and S4 (Lox2 and Post2). Although there is a difference between the composition of the four subclusters in Ca. teleta and U. unicinctus, their expression patterns are alike. In Ca. teleta, S1 genes Hox1, Hox2 and Hox3 initiated simultaneously, then S2 genes Hox4 and Hox5 initiated earlier than S3 genes Lox5, Antp and Lox4, and finally S4 genes Lox2 and Post2 are activated. In addition, Hox1 and Hox3 have the same anterior and posterior boundaries, while Lox2 and Post2 share the same anterior and posterior boundaries. These results robustly demonstrate that the expression of Hox genes in Ca. teleta also exhibits the S-WSTC pattern.

Taken together, our findings demonstrate the correspondence between Hox subcluster organization and collinearity pattern in invertebrates. Some invertebrates show disorganized and split type Hox clusters which can be divided into subclusters. The spatio-temporal collinearity of Hox genes expression may be based on subcluster in these invertebrates. This conception is also supported by some invertebrates which exhibit partial collinearity. For example, in the mollusc Cr. gigas, the Hox cluster is divided into four subclusters and Hox genes display subcluster-level temporal collinearity [45]. In the brachiopod T. transversa, two subclusters (subcluster A composed of Hox3, Hox4 and Hox5; subcluster B composed of Lox5, Antp and Lox4) exhibit temporal collinearity, three Hox genes (Hox2, Hox3 and Hox4) present spatial collinearity in the developing larval mesoderm [39]. Based on previous research data and the above analysis, we propose the spatio-temporal collinearity of invertebrate Hox is subcluster-based.

5. Conclusion

The results of our study reveal the relationship between Hox subcluster organization and spatio-temporal collinearity in U. unicinctus. Genomic organization indicates that the Hox cluster of U. unicinctus has a split type with four subclusters: S1 (Hox1 and Hox2), S2 (Hox3), S3 (Hox4, Hox5, Lox5, Antp and Lox4) and S4 (Lox2 and Post2). The expression of U. unicinctus Hox genes follows a subcluster-based whole-cluster spatio-temporal collinearity (S-WSTC) pattern. Furthermore, comprehensive analysis of the genomic and expression information of Hox genes indicates that the spatio-temporal collinearity may be based on subcluster in invertebrate with the disorganized and split Hox clusters. This study provides new insights into the correspondence between Hox subcluster organization and collinearity pattern in invertebrates.

Ethics

The collection and handling of the U. unicinctus were performed in agreement with the Guidelines for Experimental Animals of the Ministry of Science and Technology (Beijing, China) and approved by the Institutional Animal Care and Use Committee of the Ocean University of China (OUC-IACUC).

Data accessibility

The datasets analysed for this study are available in the NCBI SRA repository (SRX2999418-SRX2999431, SRX4526070-SRX4526081). The genomic scaffolds including Hox genes are available from the GenBank database (accession numbers: OM221352 and OM221353).

The data are provided in the electronic supplementary material [59].

Authors' contributions

M.W.: data curation, formal analysis, investigation, methodology, software, visualization, writing—original draft; Z.Q.: data curation, investigation, software; D.K.: data curation, investigation, methodology; D.L.: data curation, software; Q.Z.: investigation; S.B.: investigation; Z.Z.: conceptualization, funding acquisition, project administration, supervision, writing—review and editing; Y.M.: funding acquisition, supervision, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

This work was supported by the National Natural Science Foundation of China (grant nos. 42176122 and 31572601), and Shandong province science outstanding Youth Fund (grant no. ZR2020YQ20). The funding bodies played no role in the design of the study and collection, analysis and interpretation of data and in writing the manuscript.