- Open Access
Hox gene expression during postlarval development of the polychaete Alitta virens
EvoDevovolume 4, Article number: 13 (2013)
Hox genes are the family of transcription factors that play a key role in the patterning of the anterior-posterior axis of all bilaterian animals. These genes display clustered organization and colinear expression. Expression boundaries of individual Hox genes usually correspond with morphological boundaries of the body. Previously, we studied Hox gene expression during larval development of the polychaete Alitta virens (formerly Nereis virens) and discovered that Hox genes are expressed in nereid larva according to the spatial colinearity principle. Adult Alitta virens consist of multiple morphologically similar segments, which are formed sequentially in the growth zone. Since the worm grows for most of its life, postlarval segments constantly change their position along the anterior-posterior axis.
We studied the expression dynamics of the Hox cluster during postlarval development of the nereid Alitta virens and found that 8 out of 11 Hox genes are transcribed as wide gene-specific gradients in the ventral nerve cord, ectoderm, and mesoderm. The expression domains constantly shift in accordance with the changing proportions of the growing worm, so expression domains of most Hox genes do not have stable anterior or/and posterior boundaries.
In the course of our study, we revealed long antisense RNA (asRNA) for some Hox genes. Expression patterns of two of these genes were analyzed using whole-mount in-situ hybridization. This is the first discovery of antisense RNA for Hox genes in Lophotrochozoa.
Hox gene expression in juvenile A. virens differs significantly from Hox gene expression patterns both in A. virens larva and in other Bilateria.
We suppose that the postlarval function of the Hox genes in this polychaete is to establish and maintain positional coordinates in a constantly growing body, as opposed to creating morphological difference between segments.
The amazing diversity of body plans of bilateral animals is a result of the structural and regulatory evolution of genes directly connected to morphogenetic processes. Among these, Hox genes, which play a crucial role in regionalization of the anterior-posterior (AP) axis in bilateral animals, are particularly interesting [1, 2]. These genes display clustered organization and colinear expression, and are highly conserved.
Current knowledge of the function of Hox genes, and even of their expression patterns, is nonuniform across different clades of Bilateria. While in Deuterostomia and Ecdysozoa Hox gene functions are well studied, at least for vertebrates and arthropods, research in Lophotrochozoa is still at the initial stage. There are only a few studies of this animal group that describe expression patterns of all Hox genes in the cluster [3–5]. However, the Lophotrochozoa group includes an unsurpassed amount of diverse body plans and is very promising for the study of mechanisms of morphogenetic evolution.
One of the major phyla among Lophotrochozoa is Annelida. Polychaeta is a basal class of Annelida . Many species in this group have indirect development that includes stages of an unsegmented trochophore larva (a trait shared with many other Lophotrochozoa phyla) and a segmented larva, nectochaete. In the postlarval stage, polychaetes generate segments through a subterminal growth zone (GZ).
The diversity of polychaetes is manifest in various segment morphologies. There are species with morphologically similar segments, like representatives of the Nereididae family, as well as heteronomously segmented species that have different segments grouped into tagmata. To date, the expression of Hox genes has been studied in two heteronomously segmented polychaetes, Chaetopterus and Capitella. As in most bilateral animals, the expression boundaries of individual Hox genes in these species correspond with morphological boundaries of the body [5, 7]. These expression patterns are consistent with the possible role of Hox genes in establishing morphological identity along the AP axis.
Our model object is a nereid polychaete Alitta (Nereis) virens. This is an errant homonomously segmented worm. The ontogenesis of this polychaete includes a lecithotrophic trochophore, a nectochaete with three seta-bearing segments and a multi-segmented worm that grows throughout most of its life. This body plan is likely to be basal among Polychaeta . Previously, we completed a study of larval Hox gene expression in nereids Alitta virens and Platynereis dumerilii. The genomes of these species contain the whole complement of Hox genes specific to Lophotrochozoa: Hox1 (PG1), Hox2(PG2), Hox3(PG3), Hox4(PG4), Hox5(PG5), Lox5(PG6–8), Hox7(PG6–8), Lox4(PG6-8), Lox2(PG6-8), Post2(PG9+) and Post1(PG9+) . The Hox genes in a segmented larva seem to define its body plan according to the principle of spatial colinearity, as they do in most bilaterian animals.
During postlarval development, A. virens continues to form new segments in the GZ for almost its whole life. Alitta’s body does not have any apparent morphological boundaries. This raises a question about the role of the Hox genes in such an animal.
Apart from morphological heteronomy of the segments, most polychaetes display primary segmental heteronomy, which is based on differences between larval and postlarval segments. This is characteristic of both heteronomously and homonomously segmented polychaetes. Many researchers point out significant dissimilarities in formation, structure and ability to regenerate larval and postlarval segments [9–11]. In particular, the larval segments form almost simultaneously by splitting the single somatic plate into metameres; they do not produce reproductive products, and do not have metanephridia (only protonephridia are present) [4, 10, 12]. Postlarval segments, on the other hand, are formed sequentially in the posterior GZ, have metanephridia, and can produce reproductive products. There are some groups of polychaetes that use only one type of segmentation. For example, Polygordiidae lack a segmented larva, and a juvenile worm is formed right after trochophore metamorphosis. Dinophilidae (Archiannelida), on the other hand, have only ‘larval’ segments. Sometimes within a single family (such as Nereididae), there are species that have all stages of development from trochophore to an adult segmented worm (Platynereis dumerilii), as well as species with direct development (Neanthes arenaceodentata) [10, 13, 14]. Finally, different individuals of the same species can have direct development as well as development through a larval stage; moreover, the larvae can be of different types (planktonic and benthic) [15, 16].
The evolutionary and ontogenetic flexibility of polychaetes suggests that stages of their development are independent ‘modules’, controlled by different morphogenetic programs. In this case, we can expect that Hox gene expression patterns in larval and postlarval development are significantly different.
In this study, we describe detailed Hox gene expression patterns using whole-mount in-situ hybridization (WMISH) in polychaete Alitta virens during postlarval stages. We address three main questions. First, are Hox genes involved in patterning of postlarval segments? Second, are their expression patterns during postlarval development consistent with a conserved Hox gene function to convey morphological segment identity? Third, does A. virens have considerable differences between larval and postlarval Hox gene expression?
Adult Alitta virens were collected near the Kartesh Marine Biological Station of the Zoological Institute (RAS), at the White Sea, Chupa Inlet. Mature animals were caught with a hand net at the water surface during their spawning period (June and July). Artificial fertilization and cultivation of the embryos were carried out at 10.5°C . A culture of postlarval animals was kept in the Laboratory of Experimental Embryology (Peterhof, Russia) under the following conditions: temperature −18°C, salinity −23‰, artificial seawater (Red Sea salt). The size of nechtochaetes is about 0.8 mm; 4 to 6 segment worms, 1 mm; 10 to 12 segment worms, 2 mm; 15 to 20 segment worms, 3 to 4 mm; worms with more than 20 segments, up to 6 mm.
Cloning of A. virens Hox genes
The cloning of A. virens Hox genes was described previously [4, 18, 19]. Gene fragments for probe synthesis were received by 3'RACE. Gene fragments, except Nvi-Hox3, were inserted into pGEM®-T Easy Vector (Promega). Nvi-Hox3 was inserted into pBluescript II SK+ (Fermentas). The vector sequence allows sense and antisense probes to be obtained from different promoters (T7 and Sp6). Riboprobes were generated from fragments of the following lengths: 548bp for Nvi-Hox1, 580bp for Nvi-Hox2, 550bp for Nvi-Hox3, 453bp for Nvi-Hox4, 1010bp for Nvi-Hox5, 573bp for Nvi-Lox5, 522bp for Nvi-Hox7, 302bp for Nvi-Lox4, 498bp for Nvi-Lox2 and 380bp for Nvi-Post2.
Whole-mount in-situ hybridization (WMISH)
Whole-mount in-situ hybridization was performed for Alitta as described previously  with the following modifications. Hybridization was carried out at 65°C, and washings from probes at 67°C. Collagenase treatment (collagenase (Sigma) 100 γ/ml, 2.5 mM dithiothreitol (DTT); 1 mM CaCl2) was performed for 5 to 10 min, proteinase K (Sigma) treatment was performed for 10 to 20 min (10 γ/ml). Washings from the probes were performed as follows: 100% Hybe 2 × 60 min, 80% prehybe/20% PTw 2 × 20 min, 50% prehybe/50% PTw 4 × 30 min, 20% prehybe/80% PTw 2 × 20 min, 100% PTw 2 × 20 min at 67°C. Washings from antibodies were carried out for 10 × 20 min in PTw on the shaker. The detailed protocol is available on request. Between 8 and 20 worms were used for each stage. BM-purple (Roche) was used as a chromogenic substrate to localize the hybridized probe. The time of incubation in substrate was 12 h for sense transcripts and 24 to 48 h for antisense transcripts and sense transcript of Nvi-Lox4. The worms were mounted in clove oil before microscopic analysis. We used the following negative control: animals taken through the entire in-situ protocol, but with no probe (see Additional file 1). The results were imaged on a DMRXA microscope (Leica) with a Leica DC500 digital camera with Nomarski optics. Optical sections were assembled with Helicon Focus software. Brightness, contrast, and color values were corrected in all images using Adobe Photoshop CS5 image processing software.
Brief description of Alitta virens postlarval development
The main stages of Alitta virens larval development were described previously . The postlarval growth begins with origination of the fourth segment anlage (this is the first postlarval segment) from posterior GZ (Figure 1). Postlarval segments form sequentially, one by one, in the GZ. When a worm has 6 to 8 segments, its first chaete-bearing segment (the first larval segment) changes significantly. Its parapodia are considerably reduced, they lose their chaetae, their cirri grow and transform into a second pair of peristomial cirri. Then, the first larval segment becomes part of the peristome of the adult head. After this stage, the second larval segment becomes the first body segment of a juvenile worm (Figure 1). Formation of new segments continues for most of Alitta’s life and stops when the worm has about 200 segments. The morphology of the juvenile worm is shown in Figure 2.
Nvi-Hox gene expression patterns
To bring descriptive data for expression patterns into a system, we divided ten Hox genes into groups based on their expression behavior during larval development . As in larval development , Nvi-Post1 is not detectable in broad ectodermal domains of juvenile worms. Nvi-Post1 expression was observed in the chaetal sacs of developing chaetae (data not shown). Apparently, its expression is not related to the AP patterning, and we will not discuss it here. In general, most Nvi-Hox genes are expressed in parapodia; this calls for a separate study and is not covered in this article.
Nvi-Hox1, Nvi-Hox4, Nvi-Hox5, Nvi-Lox5, Nvi-Post2
The first group includes Nvi-Hox1, Nvi-Hox4, Nvi-Hox5, Nvi-Lox5,and Nvi-Post2. These genes have colinear expression patterns during larval development.
At the end of the larval development at the late nectochaete stage, expression domain of this gene has a sharp boundary in the first larval segment (Figure 3Aa). Nvi-Hox1 is detected in the ganglia of the ventral nerve cord (VNC) and parapodial ectoderm of all three larval segments with the highest level in the second segment and the lowest level in the third segment. In addition, Nvi-Hox1 expression marks the bases of peristomial and pygidial cirri. After metamorphosis, the expression expands to include the fourth and the fifth segments (first and second postlarval segments) (Figure 3Ab). Expression in pygidial cirri is significantly reduced. In juvenile worms with six to ten segments, two new Nvi-Hox1 expression domains appear: in the digestive tube at the foregut-midgut boundary (Additional file 2) and in the lateral part of the ganglia of the new segments. The expression patterns in anterior and posterior ganglia are different. In anterior segments, expression is detected in most of the ganglia but in posterior segments, Nvi-Hox1 expression appears as two spots in the lateral part of each ganglion (Figure 3Ac-Ae, Additional file 2). Despite significant changes in the first segment associated with the reorganization of the peristome, the anterior border of Nvi-Hox1 expression in this segment is persistent (Figure 3Ab-Ae; Figure 3A,E). Nvi-Hox1 expression during later postlarval stages forms a gradient with two maxima: in the first body segment (second larval segment) and in the ganglia of posterior third of the body. Nvi-Hox1 expression in the medial segments is weak. The most posterior and youngest segments and the GZ are Nvi-Hox1-negative (Figure 3Ab-Ae).
During larval development, Nvi-Hox4 expression is associated with development of the second larval segment. At the nectochaete stage, expression extends to include the third segment, but is not detected in the pygidium (Figure 3Ba).
At the onset of postlarval growth, strong Nvi-Hox4 expression is detectable in the ganglia and ectoderm of each nascent segment (Figure 3Bb). In small worms (6 or 7 segments), Nvi-Hox4 expression takes on a posterior-anterior gradient (data not shown). The interior border of Nvi-Hox4 expression is retained in the second larval segment. After the first larval segment has merged with the peristome, the second larval segment becomes the first adult body segment. Nvi-Hox4 expression persists in this area during all postlarval stages (Figure 3Bb-Be). In animals larger than ten segments, the VNC expression gradient changes its shape. Now Nvi-Hox4 is expressed intensively in several anterior ganglia and in the most posterior ones, but staining is much weaker in central segments. In the ectoderm Nvi-Hox4 expression is strong in the most posterior newly formed segments and decreases gradually toward the anterior end so it is no longer detectable in the middle part of the body. Both gradients (in the VNC and the ectoderm) ‘stretch’ but keep their shapes at all analyzed stages (Figure 3Bc-Be). The anterior border of Nvi-Hox4 expression in VNC is stable and located in the first body segment (Figure 3Bb-Be, Figure 4B,F).
During larval development, expression of this gene begins later than Nvi-Hox1 or Nvi-Hox4 and is limited to the neuroectoderm of the third segment (Figure 3Ca). Since the beginning of postlarval growth, the Nvi-Hox5 domain expands to include the ectoderm of this segment and the fourth segment surface and the developing neuromere (Figure 3Cb). In juvenile Alitta virens, Nvi-Hox5 is expressed in each postlarval segment that already has parapodia. Several youngest, most posterior segments are Nvi-Hox5-negative. Moreover, Nvi-Hox5 expression is not detectable in the pygidium and the GZ at all analyzed stages (Figure 3Cb-Ce). The expression pattern of this gene takes on a broad bow-shaped gradient. The anterior border of Nvi-Hox5 expression is retained in the second body segment (third larval segment) (Figure 3Cc-Ce; Figure 4C,G). The anterior boundary of the expression gradient in the ectoderm is located two or three segments posterior to that of the VNC. The Nvi-Hox5 expression pattern does not change significantly during postlarval growth (Figure 3Cc-Ce).
This gene displays early expression in larva during the third segment development . At the late larval stages, expression persists in the third segment and expands to include the posterior GZ (Figure 3Da). After metamorphosis, the nascent postlarval segment displays strong Nvi-Lox5 expression, which gradually decreases toward the anterior end. Staining in the third larval segment is slightly downregulated but does not disappear (Figure 3Db). In worms larger than ten segments, Nvi-Lox5 forms the expression gradient in VNC similar to Nvi-Hox4. Nvi-Lox5 expression has two peaks: one in several anterior ganglia and one in posterior segments. The transcription level of Nvi-Lox5 is weak in the central part of Alitta’s body (Figure 3Dc-De). Expression in the second body segment is weak but persists at all analyzed postlarval stages (Figure 3Db-De; Figure 4 E,H).
At the nectochaete stage, Nvi-Post2 expression is restricted to the pygidium and anal cirri (Figure 3Ga). At the onset of postlarval segmentation, expression of this gene expands to the GZ and new postlarval segments (Figure 3Gb). As the segment moves toward the anterior end, Nvi-Post2 expression weakens and gradually disappears (Figure 3Gb-Ge). Nvi-Post2 displays a short posterior-to-anterior expression gradient in the VNC, segmental ectoderm, mesoderm, and distal gut. The number of Nvi-Post2-positive segments increases with the growth of the worm, but their percentage of the total number remains virtually the same (Figure 3Gb-Ge).
Nvi-Hox7, Nvi-Lox4, Nvi-Lox2
Expression of these genes starts at the later stages of larval development. Their expression zones overlap in prospective posterior GZ . They do not take part in larval morphogenesis.
In the late nectochaete and four-segment juvenile worm, Nvi-Hox7 expression is restricted to the posterior GZ (Figure 3Ea). Expression is initiated in postlarval segments after the formation of the fourth segment (Figure 3Eb). In older worms, theNvi-Hox7 expression pattern takes on a posterior-to-anterior gradient in the VNC, segment ectoderm, and parapodia (Figure 3Ec-Ee). The expression gradient is proportional to the length of the growing worm body. (Figure 3 Ec-Ee). The anterior expression boundary is in the third segment (Figure 3Eb) in four-segment worms, and moves backwards as the worm continues to grow. In animals with 15to 30 segments, it is located in the fifth to seventh segment (Figure 3Ec-Ee), but since the staining weakens towards the anterior end, we cannot determine the exact position of the anterior boundary.
At all stages, detected Nvi-Lox4 expression is very weak, which may be because of the small size of the RNA probe (302 b). Nvi-Lox4 is first detected at the late nectochaete stage in a small number of cells in the anterior part of the pygidium (Figure 5A). In juvenile worms, the expression level increases with age. The Nvi-Lox4 transcript is spread diffusely in VNC and does not display a sharp metameric pattern (Figure 5B,C). Expression is not detected in the ectoderm within the GZ and in the pygidium (Figure 5D). Low-level expression is detected in the mesoderm of the nascent segments (Figure 5E). Nvi-Lox4 displays a posterior-to-anterior expression gradient (Figure 5C,D). The anterior border of Nvi-Lox4 expression is loose and apparently not stable, like Nvi-Hox7.
This expression is very weak in the nectochaete (Figure 3Fa), but intensifies significantly with the onset of postlarval segmentation (Figure 3Fb-Fe). The Nvi-Lox2 transcript is detected in the posterior GZ, neural ganglia, and ectoderm of posterior postlarval segments and in the pygidium (Figure 3Fb). Nvi-Lox2 also displays a posterior-to-anterior expression gradient, like that for Nvi-Hox7, Nvi-Lox4, and Nvi-Post2. Expression in the VNC lasts longer than in the segment ectoderm (Figure 3Fc-Fe). The anterior border of the Nvi-Lox2 expression domain is unstable and constantly shifts toward the posterior end. The anterior expression boundary is in the third segment (Figure 3Fb) in four-segment worms, and moves backwards as the worm continues to grow. In animals with 15 to 30 segments, it is located in the sixth to eighth segment (Figure 3Fc-Fe), but since the staining weakens towards the anterior end, we cannot determine the exact position of the anterior boundary.
These genes are expressed in an intensive and dynamic manner during early larval development , but at the nectochaete stage their expression domains narrow considerably.
In the nectochaete, Nvi-Hox2 is detected in small patches of cells in all larval segments (Figure 6A). A new expression domain appears in the mesodermal part of the GZ after metamorphosis (Figure 6B,C,E). Weak diffuse expression of Nvi-Hox2 is detected in nascent segments in small juvenile worms (Figure 6B). Older worms do not have this expression domain. At the late juvenile stages, Nvi-Hox2 is expressed in the mesoderm of the GZ and in small patches of 3 to 5 cells at the midline within each segment (Figure 6C,F,G). In worms of 8 to 10 segments and larger worms, Nvi-Hox2 is also detected in the pharynx (Figure 6E).
Since the early nectochaete stage, Nvi-Hox3 expression is present in the GZ. At all analyzed juvenile stages, strong, ring-like expression is detected in ectodermal cells of the GZ (Figure 7A-E).
We found antisense transcripts for some Nvi-Hox genes (Nvi-Hox1, Nvi-Hox4, Nvi-Hox5, Nvi-Hox7). Expression of Nvi-antiHox5 and Nvi-antiHox7 was analyzed in detail using WMISH.
After the beginning of postlarval segmentation, Nvi-antiHox5 is detectable in the GZ and the posterior segments. Sense and antisense transcripts have adjacent and even slightly overlapping expression domains. The Nvi-antiHox5 expression domain is more posterior than sense Hox5 (Figure 8A,D). In a juvenile worm, antiHox5 is detected in the GZ and the most posterior segments, where sense Nvi-Hox5 is absent (Figure 8B,C,E,F). During postlarval development, Nvi-antiHox5 expression does not have a stable anterior border. Its expression domain constantly shifts toward the posterior end.
After the beginning of postlarval growth, this transcript is detected in VNC of the third and fourth segments anterior in relation to the sense transcript (Figure 9A,D). The anterior border of Nvi-antiHox7 expression is always closer to the head than the sense transcript. Unlike the Nvi-Hox7 sense transcript, Nvi-antiHox7 is absent from nascent segments and the GZ (Figure 9B,E). In older worms, sense and antisense Nvi-Hox7 domains seem to be broadly overlapping, but particular expression patterns in the same ganglia seem to be different (Figure 9C,F). These RNAs are probably expressed in different neurons.
Summary of Hox gene expression
According to our results, most Nvi-Hox genes (Nvi-Hox1, Nvi-Hox4, Nvi-Hox5, Nvi-Lox5, Nvi-Hox7, Nvi-Lox4, Nvi-Lox2, and Nvi-Post2) form expression gradients in the central nervous system and ectoderm of A. virens. These gradients overlap significantly, but their span and shape are unique for each gene (Figures 3,5,10).
We do not have any data on physical linkage of Alitta Hox genes. However, among Lophotrochozoa, the cluster organization of Hox genes was shown for Capitella sp. I. We presume that the genomic order of the Hox genes of A. virens is similar to that of Capitella sp. I.
The spatial colinearity is clearly present for Nvi-Hox1, Nvi-Hox2, Nvi-Hox4, Nvi-Hox5,and Nvi-Lox5 genes, which have well-defined anterior boundaries (Figures 4). For Nvi-Hox1, Nvi-Hox4, Nvi-Hox5,and Nvi-Lox5, these boundaries were established during larval development (Figure 4). Anterior boundaries of other Hox genes (Nvi-Hox7, Nvi-Lox4, Nvi-Lox2, and Nvi-Post2) are not so clear, but the spatial organization of their expression domains also shows some colinearity. The Nvi-Hox7 expression gradient spreads further towards the anterior end of the body than the gradients of Nvi-Lox2 and Nvi-Post2. Nvi-Post2 displays the shortest posterior-to-anterior gradient (Figure 3Ec-Ee,Fc-Fe,Gc-Ge). According to our results, the Nvi-Lox4 anterior boundary lies posterior to the anterior boundary of Nvi-Lox2 expression. This seems to violate the spatial colinearity principle. However, the detected expression of Nvi-Lox4 is very weak; therefore, we cannot ascertain the true localization of its anterior boundary. The only gene that does not display any colinear expression is Nvi-Hox3, which is expressed in the GZ of the worm from the nectochaete stage.
Hox gene expression in annelids
In most bilateral animals, Hox genes regionalize the AP axis of the body. A discrete distribution of Hox proteins divides the early embryo into separate domains by differential regulation of target genes. This leads to establishment of expression domains correlated with morphological boundaries of the body regions [21–23]. Hox gene expression in previously studied annelids is consistent with this general principle . At the moment there are only a few studies concerning expression of Hox genes in annelids. Among these are studies on the expression of some Hox genes in larval development of Chaetopterus sp. and Platynereis dumerilii, development of the leech Helobdella sp., and larval and postlarval development of the polychaete Capitella sp.I.
Among all studied polychaetes, Chaetopterus has the most morphologically complex larva. It consists of three tagmata (A, B, and C), and there are morphological differences not only between segments in different tagmata, but also within a single tagma (В). At the late larval stages, posterior boundaries of СН-Нох1 and СН-Нох2 expression coincide with the boundary between tagmata A and B, and posterior boundary of СН-Нох5 expression coincides with the border between morphologically diverse segments within tagma B (Figure 10) .
The Capitella sp. I body consists of a thoracic region, which includes nine larval segments, and an abdominal region, which includes four additional larval and all of the postlarval segments. There is no great morphological difference between thoracic and abdominal segments. During development, expression boundaries of some Hox genes are stabilized in the region between thoracic and abdominal tagmata (Figure 10) .
Even leeches, being highly specialized annelids, retain the axial specification of the nervous system by means of the Hox cluster. The leech Helobdella sp. consists of 32 segments: 4 anterior R1-R4, 21 central М1-М21, and 7 caudal С1-С7. The body nervous system of the leech is patterned by Hox genes according to the principle of spatial colinearity. Anterior expression boundaries of He-Hox7 (Hox1), He-Lox6 (Hox4), He-Lox20 (Hox5),and He-Lox5 genes correspond to nervous system structures in four sequential segments of the rostral region, while the posterior boundary of He-Lox2 expression correlates with the anterior border of the caudal ganglion (Figure 10) .
A. virens and P. dumerilii have larvae with morphologically similar segments. The anterior expression boundaries of Hox genes coincide with the segments’ borders (Nvi-Hox1, Nvi-Hox2, Nvi-Hox4, Nvi-Hox5, Nvi-Lox5). The posterior boundaries are located between the segmented area and the pygidium. The Nvi-Post2 gene marks the pygidium territory (Figure 10) .
During most of its postlarval life, A. virens continues to form morphologically identical segments, which are not divided into tagmata. In this case, most Hox gene expression domains do not possess the stable posterior boundaries that lie in the nascent segments or in the GZ. The expression domains of A. virens Hox genes cover most of the body and overlap significantly. Comparison of the Hox gene expression in annelids with different body plans confirms that the presence of posterior expression boundaries correlates with the presence of different body tagmata, as previously shown for other bilaterian animals, for example, Arthropoda .
Larval and postlarval developmental programs
As mentioned, there are significant differences in formation of larval and postlarval segments of the polychaetes. These differences indicate the primary heteronomy of polychaete segments. Our data on Hox gene expression during ontogenesis of A. virens reveals the differences in molecular mechanisms of patterning of larval and postlarval segments.
First, a certain subset of Hox genes is expressed in each larval segment. Meanwhile, the postlarval segments express all Hox genes but not at the same time during postlarval growth. Each nascent segment expresses all Hox genes except Nvi-Hox1, Nvi-Hox3, and Nvi-Hox5. As new segments continue to form, the older segments move towards the head and begin to express Nvi-Hox1 and Nvi-Hox5, while the expression of other genes is gradually downregulated or upregulated according to the shape of gene expression gradients (Figure 11). Interestingly, neural expression of ParaHox cluster genes in A. virens is governed by the same principle; each ParaHox gene is expressed at different periodsin all postlarval (but not the larval) neuromeres .
Second, three Hox genes, Nvi-Hox7, Nvi-Lox4 and Nvi-Lox2, are not expressed during larval segment formation . They are switched on in each postlarval segment on its formation in the GZ (Nvi-Hox7, Nvi-Lox2) or at the beginning of growth (Nvi-Lox4).
Taken together, the existing morphological data and the differential character of Hox gene expression in larval and postlarval development of A. virens support the idea of separate morphogenetic developmental programs in the segmented larva and adult worm.
Comparing Hox gene expression in different polychaetes (Chaetopterus, Capitella, Platynereis, Alitta), one can notice the fundamental similarity of Hox gene expression in their larval development. In all four cases (Figure 10), Hox genes are activated early in wide spatial domains of the larval body. The Hox gene expression corresponds to the location of primordial structures organized along the main body axis, and the expression anterior boundaries are colinear. These features link the program of polychaete larval body organization to the programs utilizing the Hox genes during the embryonic development of animals in other evolutionary clades, such as Deuterostomia and Ecdysozoa.
Apart from A. virens, Hox gene expression in only one other postlarval polychaete has been studied to date, Capitella sp. I. Unfortunately, it is difficult to compare these two worms. All the segments of A. virens formed from the subterminal GZ are considered to be postlarval. In contrast, several segments of Capitella that are referred to as larval ones are formed sequentially from the posterior GZ . The Hox gene expression in Capitella after metamorphosis is considerably different from that in Alitta. First, in postlarval segments only CapI-Lox4, CapI-Lox2, and CapI-Post2 genes are active. Second, Hox gene expression in a juvenile worm is maintained only within the neural system ganglia. Third, expression boundaries in the juvenile worm are stable and correspond to the morphological tagma boundaries.
Nevertheless, a comparison of Hox ortholog expression in late Capitella larva, when it has already formed several segments from the posterior GZ, and juvenile A. virens reveals fundamental similarities in many gene patterns. Hox1, Hox4, and Lox5 genes have anterior as well as posterior expression domains in both species (Figure 11). In both cases, the expression has a gradient shape (Figure 3) . We can speculate that this stage in Capitella development is a separate phase that can be compared to the postlarval development of A.virens.
There are also significant differences of Hox gene expression between larval and juvenile stages of Capitella development. In juvenile stages of Capitella, almost all expression domains of Hox genes become limited to the VNC and discrete posterior expression boundaries appear .
Our data and results for Capitella suggest that there are actual differences between larval and postlarval developmental programs. The question arises: which mode of segmentation, larval or postlarval, is more ancestral? Several authors [12, 27] believe that larval segmentation (that is, the simultaneous metamerization of the somatic plate that creates larval segments) is a basic type, while sequential formation of postlarval segments from the GZ is an evolutionary novel segmentation type. Anderson, on the other hand, believes that postlarval segmentation is a primitive type [14, 27].
During A. virens nectochaete development, only the ‘ancient’ Hox genes are expressed. Indeed, the minimal Hox gene complement of Urbilateria, the last common ancestor of bilateral animals, consisted of at least seven Hox genes. It contained five anterior genes (PG1-5), at least one gene from a central group (PG6/8), and at least one posterior gene (PG9+) . This is exactly the set of Hox paralogs that is expressed during development of A. virens larva . At the onset of postlarval growth, the genes of the central group, namely Hox7, Lox4, and Lox2, are activated. This may support the notion that postlarval segmentation is an evolutionary novelty. However, there is a certain contradiction with data for larval Hox gene expression in Capitella. All CapI-Hox genes are expressed during larval segment formation, including CapI-Hox7, CapI-Lox4, and CapI-Lox2.
The existing evidence is not sufficient to draw final conclusions about larval and postlarval developmental programs in the polychaetes. This emphasizes the need for Hox gene expression studies in other polychaete families, since this will probably elucidate the evolution of postlarval morphogenesis in this class.
Possible Hox gene function in A. virens postlarval development
The nereid A. virens produces morphologically similar segments for almost its whole life. According to our data, most of A. virens Hox genes are expressed in the segment ectoderm and in the neural system of each postlarval segment. Hox gene transcripts are distributed as complex gene-specific gradients. As the worm grows, the position of each segment along the body AP axis changes, and this is accompanied by a change in Hox gene expression profile (Figure 11). We assume that the system of Hox gene transcriptional gradients is necessary not for specification of segment morphology, but to create and maintain the positional information for each segment. Within the constantly growing polychaete body, this regulatory system allows one to assign each metamere a unique ‘Hox code’, which gradually changes as the segment moves with respect to the terminal structures, that is, the head and pygidium (Figure 11). In this case, the ‘Hox code’ serves for positional, rather than morphological, specification.
Establishing and maintaining of positional information may be necessary for regeneration. Nereididae can regenerate posterior body parts. After a part of the body is lost, it is necessary to change the scale of the positional coordinates quickly in accordance with the new boundaries of the body. Since genes in the Hox cluster are often coordinated by common regulatory elements and a common feedback system, one can expect that the expression pattern of the whole complex would be easily reorganized during positional failure.
Planarian worms (Turbellaria; Platyhelminthes) also have gradient Hox gene expression. In planarian Dugesia japonica, Plox4-Dj (PG5; Hox5), Plox5-Dj (PG6-8; Lox5), and Dj-Abd-Ba (PG9+, Post2) transcripts are distributed as gradients along the AP axis of the adult worm [28, 29]. The key aspect is that in this case Hox gene expression does not correspond to any morphological structures, has no clear boundaries and is proportional to the length of the worm body. It was also experimentally shown that Dj-Abd-Ba gene quickly restores expression pattern in head and tail regenerates, long before the new head and tail are formed, and this pattern is consistent with the new planarian body proportions .
Persistent 5' HoxC gene expression was surprisingly discovered in the spinal cord of an adult newt Pleurodeles waltl. RT-PCR data showed that PwHoxc13, PwHoxc12, and PwHoxc10 expression is upregulated during tail regeneration. The authors of this study suggest that Hox genes in an adult newt act as carriers for the positional memory necessary to achieve effective regeneration .
It is noteworthy that in newts and polychaetes the definitive Hox gene expression is associated with the neural system. Expression patterns of various Hox genes in the A. virens VNC ganglia are somewhat different (Figure 3). Hox gene function in A. virens is probably associated with specification of particular neuron types within VNC ganglia. Studies of another polychaete, Platynereis dumerilii, support this hypothesis. There is evidence that Pdu-Hox genes are expressed in different parts of the neuromeres . Moreover, in the leech Helobdella expression patterns of different Нох genes in the ganglia are also different . Colinear Hox gene expression was found during nervous system development in arthropods, vertebrates, ascidians, and annelids [5, 32, 33]. In some animals, the nervous system is the only structure that shows spatial and temporal colinearity of Hox genes expression [3, 24, 34]. Probably, this role in nervous system regionalization is the most conserved function of the Hox cluster.
Hox genes cooptions in A. virens
Some Alitta Hox genes are likely to take part in patterning of particular structures rather than working in an orchestrated way together with other genes of the cluster. For example, Nvi-Hox1 and Nvi-Hox2 genes are expressed in the pharynx and at the esophagus-midgut boundary. Interestingly, their orthologs in other studied polychaetes, such as Capitella and Chaetopterus, are also expressed in the pharynx [5, 7]. In addition, Hox orthologs in deuterostomes are involved in establishing the posterior boundary of pharyngeal entoderm [35, 36]. This cooption is likely to date back to the early stages of evolution, probably to the origins of all Eubilateria.
Hox1 orthologs work in peristomial and pygidial cirri in A. virens and P. dumerilii. Since Hox1 expression is found in the notopodia and neuropodia of these polychaetes, one can assume that it is expressed in cirri as in serially homologous structures and thus is not purely a cooption. In general, most Nvi-Hox genes are expressed in parapodia, and this calls for a separate study. It is noteworthy, however, that this expression changes in accordance with general gradient pattern in the neural system and segment ectoderm of the worm.
The aforementioned Nvi-Post1 gene was identified in parapodial chaetal sacs not only in Alitta and Platynereis, but even in Capitella[5, 20], which is located far from Nereididae in the phylogenetic tree [37, 38]. This suggests that this gene fell out of the common regionalization program early on.
In contrast to other Lophotrochozoa, Hox genes in Alitta have a rather limited range of cooptions. For example, Nvi-Hox7 functions only in the neural system and in the segment ectoderm of postlarval segments, while its Capitella ortholog CapI-Antp has additional expression domains in the brain and pharynx . In P. dumerilii, the Lox2 gene works in the coelomic epithelium; this sets it apart from other Pdu-Hox genes and the Nvi-Lox2 gene, which exhibits strong expression in ectodermal tissues .
In general, studied Hox gene cooptions in different species within the Lophotrochozoa clade support an idea that these genes are easily involved in new morphogenetic programs. They often play a role in the morphogenesis of structures, specific for large taxa within this animal group. For example, they pattern the shell gland in gastropods and the brachial crown in cephalopods [3, 39].
Nvi-Hox genes antisense transcripts
Our results represent the first discovery of Hox gene antisense transcripts in an animal from the Lophotrochozoa clade. These are long antisense RNA, complementary to the sense transcripts.
Hox cluster regulation by noncoding RNA was reported for mammals and arthropods [40, 41]. Over 200 noncoding RNAs of different sizes and directions are transcribed from four human Hox clusters . Our finding suggests that this is a universal regulatory mechanism in bilateral animals.
The Nvi-antiHox5 transcript is expressed in a pattern almost complementary to the sense transcript (Figure 8). This resembles anti-Ubx expression in centipedes [41, 42]. Anti-Ubx is functional in embryogenesis and exhibits a complex expression pattern that is mutually exclusive with the sense transcript. There have been no functional studies so far, but this early and specific nature of expression suggests a possible regulatory role of anti-Ubx.
The Nvi-antiHox7 anterior expression boundary lies ahead of Nvi-Hox7 (Figure 9). Both transcripts are detected in VNC ganglia and in the segment ectoderm, but their expression patterns are different within the same ganglia (Figure 9C,F). The mutually exclusive territories of sense and antisense transcript distribution suggest that one may be controlled by another.
So far, we have identified antisense transcripts for most A. virens Hox genes. Some of them are cloned (data are being prepared for publication). It should be noted that Hox genes long antisense RNAs in polychaetes are located in the cytoplasm. We do not know their function or biogenesis, but existing examples allow us to suppose that antisense RNAs take part in epigenetic tuning of the Hox cluster. They can participate in transcriptional and translational repression of their targets, protein-encoding RNAs. Nvi-antiHox5 and Nvi-antiHox7 expression patterns suggest that they may be controlling the anterior (Nvi-Hox7) or posterior (Nvi-Hox5) expression boundaries of their sense counterparts. The worm is constantly growing, and its Hox genes change levels of expression in a coordinated manner in accordance with the body proportions. In this situation, turning the genes on and off quickly at a translational level may be very advantageous.
The data on expression patterns of A. virens and Capitella sp.I Hox genes support the idea of different morphogenetic programs underlying larval and postlarval development of the polychaetes. The larval program seems to be more conservative among the polychaetes. The principles of Hox gene expression during larval segmentation are similar to those during the body plan formation of other bilaterians. The postlarval programs display more differences between species and seem to be more evolutionary flexible.
The gradient expression of the Nvi-Hox genes in the postlarval period may be responsible for establishing and maintaining the positional values along the AP axis of multi-segmented, constantly growing polychaete body. The role of A. virens Hox cluster in positional specification can be studied in regeneration experiments.
The system of long antisense RNAs is a possible clue to the question of ‘sliding’ Hox gene expression boundaries in growing worms. This has not yet been described in existing model systems (planarian, mouse, drosophila), and we aim to investigate this phenomenon thoroughly.
Reverse transcriptase polymerase chain reaction
Ventral nerve cord
Whole-mount in-situ hybridization.
Carroll SB: Homeotic genes and the evolution of arthropods and chordates. Nature. 1995, 376: 479-485. 10.1038/376479a0.
Wray GA: Transcriptional regulation and the evolution of development. Int J Dev Biol. 2003, 47: 675-684.
Lee PN, Callaerts P, de Couet HG, Martindale MQ: Cephalopod Hox genes and the origin of morphological novelties. Nature. 2003, 424: 1061-1065. 10.1038/nature01872.
Kulakova M, Bakalenko N, Novikova E, Cook CE, Eliseeva E, Steinmetz PRH, Kostyuchenko RP, Dondua A, Arendt D, Akam M, Andreeva T: Hox gene expression in larval development of the polychaetes Nereis virens and Platynereis dumerilii (Annelida, Lophotrochozoa). Dev Genes Evol. 2007, 217: 39-54. 10.1007/s00427-006-0119-y.
Fröbius AC, Matus DQ, Seaver EC: Genomic organization and expression demonstrate spatial and temporal Hox gene colinearity in the Lophotrochozoan Capitella sp. I. PLoS One. 2008, 3 (12): e4004-10.1371/journal.pone.0004004.
Brusca RC, Brusca GJ: Invertebrates. 2002, Massachusetts: Sinauer
Irvine SQ, Martindale MQ: Expression patterns of anterior Hox Genes in the polychaete Chaetopterus: correlation with morphological boundaries. DevBiol. 2000, 217: 333-351.
Tessmar-Raible K, Arendt D: Emerging systems: between vertebrates and arthropods, the Lophotrochozoa. CurrOpin Genet Dev. 2003, 13: 331-340. 10.1016/S0959-437X(03)00086-8.
Wilson EB: The cell lineage of Nereis. J Morph. 1892, 6: 361-480. 10.1002/jmor.1050060301.
Ivanoff PP: Die Entwicklung der Larvalsegmentebei den Anneliden. Z Morph Ökol Tiere. 1928, 10: 62-161. 10.1007/BF00419279.
Beklemishev WN: Principles of Comparative Anatomy in Invertebrates. 1969, Moscow: Nauka
Fischer AHL, Henrich T, Arendt D: The normal development of Platynereis dumerilii (Nereididae, Annelida). Frontiers in Zoology. 2010, 7: 31-10.1186/1742-9994-7-31.
Anderson TD: Embryology and Phylogeny in Annelids and Arthropods. 1973, Oxford: Pergamon Press
Winchell CJ, Valencia JE, Jacobs DK: Confocal analysis of nervous system architecture in direct-developing juveniles of Neanthes arenaceodentata (Annelida, Nereididae). Frontiers in Zoology. 2010, 7: 17-10.1186/1742-9994-7-17.
Gibson GD, Gibson JF: Heterochrony and the evolution of poecilogony: generating larval diversity. Evol. 2004, 58 (12): 2704-2717.
Kesäniemi JE, Geuverink E, Knott KE: Polymorphism in developmental mode and its effect on population genetic structure of a spionid polychaete, Pygospio elegans. Integr Comp Biol. 2010, 52 (1): 181-196.
Dondua AK: Effect of actinomycin D and sibiromycin on the embryonic and larval development of Nereis virens (Sars.). Ontogenez. 1975, 6 (5): 475-484.
Andreeva TF, Kuk C, Korchagina NM, Akam M, Dondya AK: Cloning and analysis of structural organization of Hox genes in the Polychaete Nereis virens. Ontogenez. 2001, 32 (3): 225-233.
de Rosa R, Grenier JK, Andreeva T, Cook CE, Adoutte A, Akam M, Carroll SB, Balavoine G: Hox genes in brachiopods and priapulids and protostome evolution. Nature. 1999, 399: 772-776. 10.1038/21631.
Kulakova MA, Kostyuchenko RP, Andreeva TF, Dondua AK: The abdominal-B-like gene expression during larval development of Nereis virens (Polychaeta). Mech Dev. 2002, 115 (1–2): 177-179.
Averof M, Akam M: HOM/Hox genes of Artemia: implications for the origin of insect and crustacean body plans. Curr Biol. 1993, 3: 73-78. 10.1016/0960-9822(93)90158-K.
Hughes CL, Kaufman TC: Hox genes and the evolution of the arthropod body plan. Evol Dev. 2002, 4 (6): 459-499. 10.1046/j.1525-142X.2002.02034.x.
Burke AC, Nowicki JL: Hox genes and axial specification in vertebrates. Amer Zool. 2001, 41: 687-697. 10.1668/0003-1569(2001)041[0687:HGAASI]2.0.CO;2.
Kourakis MJ, Master VA, Lokhorst DK, Nardelli-Haefliger D, Wedeen CJ, Martindale MQ, Shankland M: Conserved anterior boundaries of Hox gene expression in the central nervous system of the leech Helobdella. Dev Biol. 1997, 190: 284-300. 10.1006/dbio.1997.8689.
Kulakova MA, Cook CE, Andreeva TF: ParaHox gene expression in larval and postlarval development of the polychaete Nereis virens (Annelida, Lophotrochozoa). BMC Dev Biol. 2008, 8: 61-10.1186/1471-213X-8-61.
Seaver EC, Thamm K, Hill SD:Growth patterns during segmentation in the two polychaete annelids Capitella sp. I and Hydroides elegans: comparisons at distinct life history stages.Evol Dev. 2005, 7 (4): 312-326. 10.1111/j.1525-142X.2005.05037.x.
Giangrande A, Gambi MC: Metamerism and life‒style within polychaetes: morpho-functional aspects and evolutionary implications. Ital J of Zool. 1998, 65 (1): 39-50. 10.1080/11250009809386725.
Orii H, Kato K, Umesono Y, Sakurau T, Agata K, Watanabe K:The planarianHOM/HOXHomeobox genes (Plox) expressed along the anteroposterior axis.Dev Biol. 1999, 210: 456-468. 10.1006/dbio.1999.9275.
Nogi T, Watanabe K: Position-specific and non-colinear expression of the planarian posterior (Abdominal-B-like) gene. Dev Growth Differ. 2001, 43: 177-184. 10.1046/j.1440-169X.2001.00564.x.
Nicolas S, Papillon D, Perez Y, Caubit X, Le Parco Y: The spatial restrictions of 5'Hox C genes expression are maintained in adult newt spinal cord. Biol Cell. 2003, 95: 589-594.
Pfeifer K, Dorresteijn AWC, Fröbius AC: Activation of Hox genes during caudal regeneration of the polychaete annelid Platynereis dumerilii. Dev Genes Evol. 2012, 222: 165-179. 10.1007/s00427-012-0402-z.
Ikuta T, Yoshida N, Satoh N: Ciona intestinalis Hox gene cluster: its dispersed structure and residual colinear expression in development. PNAS. 2004, 101 (42): 15118-15123. 10.1073/pnas.0401389101.
Tara A, Christof N, Robb K: Hox genes and segmentation. Annu Rev Cell DevBiol. 2009, 25: 431-456. 10.1146/annurev.cellbio.042308.113423.
Hinman VF, O’Brien EK, Richards GS, Degnan BM: Expression of anterior Hox genes during larval development of the gastropod Haliotis asinina. Evol Dev. 2003, 5: 508-521. 10.1046/j.1525-142X.2003.03056.x.
Schubert M, Yu JK, Holland ND, Escriva H, Laudet V, Holland LZ: Retinoic acid signaling acts via Hox1 to establish the posterior limit of the pharynx in the chordate amphioxus. Dev. 2005, 132: 61-73.
Aronowicz J, Lowe CJ:Hoxgene expression in the hemichordateSaccoglossus kowalevskiiand the evolution of deuterostome nervous systems.Integr Comp Biol. 2006, 46 (6): 890-901. 10.1093/icb/icl045.
Struck TH, Schult N, Kusen T, Hickman E, Bleidorn C, McHugh D, Halanych KM: Annelid phylogeny and the status of Sipuncula and Echiura. BMC Evol Biol. 2007, 7: 57-10.1186/1471-2148-7-57.
Zrzavý J, Říha P, Piálek L, Janouškovec J: Phylogeny of Annelida (Lophotrochozoa): total-evidence analysis of morphology and six genes. BMC Evol Biol. 2009, 9: 189-10.1186/1471-2148-9-189.
Samadi L, Steiner G: Expression of Hox genes during the larval development of the snail, Gibbulavaria (L.)-further evidence of non-colinearity in molluscs. Dev Genes Evol. 2010, 220 (5): 161-172.
Rinn JL, Kertesz M, Wang JK, Sharon L, Squazzo SL, Xu X, Brugmann SA, Goodnough H, Helms JA, Farnham PJ, Segal E, Chang HY: Functional demarcation of active and silent chromatin domains in human HOX loci by non-coding RNAs. Cell. 2007, 129 (7): 1311-1323. 10.1016/j.cell.2007.05.022.
Brena C, Chipman AD, Minelli A, Akam M: Expression of trunk Hox genes in the centipede Strigamia maritima: sense and anti-sense transcripts. Evol Dev. 2006, 8 (3): 252-265. 10.1111/j.1525-142X.2006.00096.x.
Janssen R, Budd GE: Gene expression suggests conserved aspects of Hox gene regulation in arthropods and provides additional support for monophyletic Myriapoda. Evo Devo. 2010, 1: 4-
We are grateful to Tanya Andreeva, who laid the foundation for this research; to Michael Akam and Charles Cook, who helped to clone Nvi-Hox-genes fragments and who supported us. We thank the staff of the Kartesh White Sea Biological Station (Zoological Institute, Russian Academy of Sciences) for help in collecting and maintaining the A. virens.
We thank Olga B. Lavrova and Victor Starunov for their help in maintaining the Alitta virens culture in the Laboratory of Experimental Embryology. We are grateful to the Chromas center for providing the opportunity to use the Leica microscope. The research was supported by RFBR grants: 06-04-49654-a and 09-04-01322-a.
The authors declare that they have no competing interests.
NIB, ELN, and MAK performed the experiments and data analysis. AYN participated in material collection, maintenance of the worm culture and editing the manuscript. ELN, NIB, and MAK conceived the study and drafted the manuscript. All authors have read and approved the final manuscript.
Nadezhda I Bakalenko, Elena L Novikova contributed equally to this work.
Electronic supplementary material
Additional file 1: Negative control with no probe. All these worms were processed in the same WMISH. (A,B,C) The worms with different numbers of segments.(D,E,F) The posterior ends of 25 to 35 segment worms in higher magnification. The anterior end is to the right on all panels. All views are ventral. (A) The small worms do not display any background. (B-E) Weak background is detected in the gut (arrowheads), epithelial glands (arrows), parapodia, pygidial glands (red arrows) and at the ventral surface of the heads of 15 to 35 segment animals. (F) Some worms display strong background in the parapodia and pygidial glands. We have never seen any background in the nervous system or mesoderm. WMISH, whole-mount in-situ hybridization. (PDF 91 KB)
Additional file 2: Some details of Nvi-Hox1 expression. (A) Nvi-Hox1 expression in the postlarval worm (overview). Images on panels (B) and (C) indicate the different expression patterns of Nvi-Hox1 in the anterior (C) and posterior (B) ganglia of the VNC. (D) Nvi-Hox1 expression in the peristomial cirri. (E) Nvi-Hox1 expression in the pharynx and foregut-midgut boundary. VNC, ventral nerve cord. (PDF 61 KB)
Authors’ original submitted files for images
Below are the links to the authors’ original submitted files for images.