Collagens are the predominant proteins in the extracellular matrix of animals. They play diverse roles in cell adhesion, differentiation, cell migration as well as tissue regeneration. Currently, there are at least 29 different collagens known (Söderhäll et al, 2007
; Kadler et al, 2007
). Generally, collagen fibrils are composed of three peptide alpha chains forming either heterotrimers or homotrimers. Depending on their functions and domain homology, collagens can be grouped into seven different types (Kadler et al, 2007
). Collagen type I, II, III, V and XI as well as recently discovered XXIV and XXVII are fibril-forming collagens. Historically, the fibrillar collagens are further divided into major (I, II and III) and minor (V, XI, XXIV and XXVII) fibrillar collagens based on their relative abundance in the collagen fibrils. For collagen type V, three different alpha chains have been characterized in human: COL5A1, COL5A2 and COL5A3. Likewise, collagen type XI also has three alpha chains, COL11A1, COL11A2 and COL11A3. However, COL11A3 is actually COL2A1 (alpha1 chain, type II), one of the major fibril-forming collagens. Additionally, another alpha chain has been identified in rat, Col5a4, is expressed in Schwann cells and shows a high similarity to Col5a3 (Chernousov et al, 2000
). However, Col5a4 has not been reported for other species.
Minor fibril-forming collagens are composed of several domains, including the C-propeptide (NC1 or CPP), a major helix (COL1) and minor helix (COL2) separated by a short sequence (NC2), and the N-terminal end (NC3). The NC1 or CPP domain is well conserved among fibril-forming collagens, whereas NC3 domain varies in terms of size, primary structure and subdomains (Ricard-Blum, 2005). In the minor fibrillar collagens, NC3 contains either a cysteine-rich repeat domain (CRR) feature with 10 cysteines in the sequence (Col5a2), or a thrombospondin N-terminal-like (TSPN) domain adjacent to a variable region (VR) (Col5a1, Col5a3, Col11a1 and Col11a2). The structural difference between Col5a2 and the other minor fibrillar collagens is in agreement with recent phylogenetic analysis (Zhang et al, 2007
). The VR, historically, is defined as the region of low homology among different collagen alpha chains. For both collagens type V and XI, the amino acid sequence of the VRs are rich in charged residues under physiological conditions. The VR in Col11a1 contains a heparan sulfate binding site (Warner et al, 2006
). In human, mouse, rat and other species, Col11a1 and Col11a2 also exhibit extensive alternative splicing (AS) in the VR (Li et al, 1995
; Lui et al, 1996
; Oxford et al, 1995
; Tsumaki & Kimura, 1995
). Although there can be many different combinations of exons in the VR, the most abundant forms of Col11a1 detected are the spliceforms that include exons 6A-7-8, exons 6B-7, and exon 7 alone between the constitutively expressed exons 5 and 9 (Morris et al, 2000
; Oxford et al, 1995
); in the case of Col11a2, exon 6, exon 7 and exon 8 can be either included or excluded independently (Lui et al, 1996
; Tsumaki and Kimura, 1995
The importance of collagens has been showed by their profound influence in connective tissue formation and function. In human, mutations in collagen type V genes cause Ehlers-Danlos syndrome, resulting in hypermobility of joints (Richards et al, 1998
; Schwarze et al, 2000
); whereas mutations in collagen type XI can cause Marshall’s or Stickler syndrome, both resulting in altered facial appearance, eye abnormalities, joint problems and hearing loss (Annunen et al, 1999
; Chen et al, 2005
; McGuirt et al, 1999
; Melkoniemi et al, 2000
; Pihlajamaa et al, 1998
; Sirko-Osadsa et al, 1998
). However, the pathogenic mechanisms for these syndromes are still not completely understood.
To date, there is little information about the early expression patterns of minor fibrillar collagens during early embryogenesis in vertebrate model organisms. Minor fibrillar collagens are expressed in a variety of different tissues in mouse, chicken, and rat. Col5a1 expression is observed in the aorta, heart, branchial arches, developing mesenchyme and neuroepithelium; later in development, Col5a1 expression is limited to bones, vertebral column and cornea as well as the tendons and ligaments (Imamura et al, 2000
; Roulet et al, 2007
). The location of expression of Col5a3 has been shown as a subset of that of Col5a1 (Imamura et al, 2000
). Col11a1 has been detected in limbs, vertebrae, mandibular bones, otic vesicle, and the atrioventricular valve in the heart, neuroepithelium of the brain, liver, kidney, lung, muscle and intestine (Lui et al, 1995
; Oxford et al, 1995
; Yoshioka et al, 1995
), mesothelial layer of the cornea, and perioptic mesenchyme within the sclera (Sugimoto et al, 1998
). Relative to Col11a1, Col11a2 has more restricted cartilage-related expression in developing limbs and axial skeleton (Sugimoto et al, 1998
). To further investigate the mechanisms of pathogenesis due to mutations in minor fibrillar collagens, we have characterized their spatial and temporal expression patterns during early development in zebrafish (Danio rerio
). Recently, work by Baas and colleagues demonstrated the importance of one of the minor fibrillar collagens located on chr19 during embryonic development (Baas et al. 2009
). Col11a2 expression pattern in embryonic structures was also described by Yokoi and colleagues (Yokoi et al. 2009
). Results from our study are compared to the results of a recent studied carried out by Hoffman and colleagues (Hoffman et al. 2010) in which they characterized early expression patterns of type V and XI collagens.
1.1 Characterization of zebrafish minor fibrillar collagens and proteins
To identify zebrafish minor fibrillar collagens (Col5a1, Col5a3, Col11a1 and Col11a2), the NC1 or CPP domain of human minor fibrillar collagens were used to search against the GenBank and Ensembl databases using the Blastp search algorithm (WU Blast2.0 default settings). Loci identified as encoding both the NC1 domain and the amino terminal thrombospondin-like domain were further analyzed for predicted exons. Predicted exons were identified in the database and compared to other vertebrate species. To confirm predicted exons, PCR primers were designed, synthesized, and used to amplify the predicted variable regions by RT-PCR. Identity and exon composition were verified by sequencing DNA contained within excised bands. Accession numbers and the primer information are listed in . Our analysis focused on the alternative splicing in the VR.
Accession numbers and primers for Col5a1, Col5a3, Col11a1, Col11a2.
Translation of the VRs of the minor fibrillar collagens were then compared among zebrafish, chimpanzee, mouse, rat and chicken. The level of conservation at the amino acid level is highlighted in , focusing on each exon in the VR. Although the VR demonstrates low homology among species, there are specific conserved patterns of residues that have been identified. Exon 6 (E6) in Col5a1 and E6A, E8 in Col11a1 as well as E8 in Col11a2 are rich in tyrosine and acidic residues aspartic acid and glutamic acid; whereas E6 in Col5a3 and E6B in Col11a1 include clusters of arginines and lysines. Tyrosine of Col5a1 is known to be extensively sulfonated in the extracellular matrix, and successive arginines and lysines have been shown to serve as a binding motif for proteoglycans (Erdman et al, 2002
; Yamaguchi et al, 2005
). However, none of the tyrosine-rich or arginine/lysine-rich domains are thoroughly understood with respect to specific function in the minor collagens. A gene encoding a minor fibrillar collagen most closely related to Col11a1 (designated Col11a1b by Hoffman et al., 2010) was identified on chromosome 2. No indication of alternative splicing was observed for the VR of Col11a1b in our study.
Figure 1 Sequence and domain structure alignment of minor fibrillar collagens in zebrafish (D. rerio), human (H. sapiens), mouse (M. musculus), rat (R. norvegicus), chimpanzee (P. troglodytes) and chicken (G. gallus). Protein sequence alignment was generated by (more ...)
1.2 Expression of Col5a1, Col5a3, Col11a1 and Col11a2 during zebrafish development Temporal expression patterns
We used RT-PCR to determine the temporal expression patterns of Col5a1, Col5a3, Col11a1 and Col11a2. The transcription of all minor fibrillar collagen genes was first apparent at approximately 4–6 hpf, with the exception of Col11a2, in which maternal transcripts were detected (Supplement Fig. 1
). Both Col11a1 and Col11a2 exhibited similar trends, demonstrating additional spliceforms at later developmental stages (). Sequence analysis of each band generated by PCR confirmed alternative splicing for these transcripts, which are likely to be developmentally regulated. In addition, we observed a previously undescribed exon within VR of Col11a2 in addition to exon 6, 7 or 8, which was between exon 6 and exon 7 in the genomic sequence. This exon (designated exon 6′) was approximately 1200bp in length. Surprisingly, our experiments also demonstrated alternative splicing for Col5a3 on chr3 within the VR (), which contained an additional 447bp exon. Alternative splicing within the VR of Col5a3 has not been reported for other species. This observation is in agreement with Hoffman and colleagues (Hoffman et al, 2010). The alternative splicing is summarized in . Sequence data for each of the exons within the VRs is shown in .
Figure 2 Temporal expression and splicing patterns of minor fibrillar collagens. RT-PCR was performed using RNA extracted from wild type embryos at 4, 6, 10, 16, 24, 36, 48, 72hpf. (A). Multiple bands were observed when cDNA was amplified within the region between (more ...) Spatial expression patterns
The spatial expression patterns of the minor fibrillar collagens were then examined by in situ hybridization. Antisense riboprobes were synthesized based on the mRNA sequences in the VR for each gene. Because Col5a1 does not show alternative splicing within the VR, the antisense riboprobe was designed to be complimentary to exons 6-7-8. However, as Col5a3, Col11a1a and Col11a2 undergo alternative splicing within the VR, the riboprobes were designed from the earliest or most prevalent spliceform. For Col11a1a, the probe was designed to hybridize to exons 6a-7-8, which is the major product demonstrated by RT-PCR (). Detection of Col5a3 and Col11a1b was accomplished using probes representing all exons within the variable region.
Generally, Col5a1 was expressed in developing somites, neural crest, and head mesenchyme at early timepoints; at later stages, it was expressed in the developing cranial cartilage, pharyngeal arches and vertebrae. Col5a3 was detected in the notochord at early timepoints in development, and later in mesenchymal cells of the eyes and lens. Col11a1a and Col11a2 demonstrated similar expression patterns, including notochord, otic vesicle, and developing cranial cartilages. Interestingly, Col11a1b on chromosome 2 shared a very similar expression pattern with Col5a1.
Comparative expression at 16 hpf
At 16 hpf, Col5a1 was expressed predominantly in the cranial neural crest and in the region of developing somites (). Col5a3 was expressed in the notochord (), as was Col11a1a on chromosome 24 (Fig. D, I) and Col11a2 (). Col11a1b on chromosome 2 was detected near the otic placode and hindbrain () as was Col11a1a on chromosome 24 (). Col11a2 was expressed around the developing notochord (). Controls showed no detectable signal.
Figure 3 Spatial expression patterns of minor fibrillar collagen genes in zebrafish at 16 hpf by whole mount in situ hybridization. The expression patterns of Col5a1 (A, F, K), Col5a3 (B, G), Col11a1b (C, H, L), Col11a1a (D, I, M) and Col11a2 (E, J) are shown (more ...) Comparative expression at 24 hpf
At 24 hpf, Col5a1 was detected in the locations observed at 16 hpf (), and also in agreement with expression patterns observed in mouse (Roulet et al, 2007
). Col5a3 was expressed in the notochord with a stronger signal toward the tip of the tail. Col11a1b on chromosome 2 seemed to have a similar pattern as Col5a1 at this time point although weaker. Both Col11a1a on chromosome 24 and Col11a2 were detected in the notochord. The previously observed expression of Col11a1a on chromosome 24 in otic placode and hindbrain was not detected at 24 hpf, indicating transient expression in the hindbrain and otic placode during early development. Marshall’s and Stickler syndrome patients commonly have hearing problems, and this transient expression may shed light on the mechanism of how mutations in Col11a1 are associated with these symptoms. Col11a2 was detected faintly in the region of the developing eye ( right-hand column). Controls showed no detectable signal.
Figure 4 Spatial expression patterns of minor fibrillar collagen genes in zebrafish at 24 hpf by whole mount in situ hybridization. The expression patterns of Col5a1, Col5a3, Col11a1b, Col11a1a, and Col11a2 are shown in both dorsal (A) and lateral (B) views. Magnified (more ...) Comparative expression at 48 and 72 hpf
At 48 and 72 hpf, the expression of Col5a1 was detected in the region of developing somites, perioptic mesenchyme cells in the eyes, hindbrain and otic vesicle (). Col5a3 was expressed at 48 hpf at the posterior end of the notochord, the lens and mesenchyme cells of the eyes (). In addition, Col5a3 also was expressed in the developing anterior pericardium (, second column). Col11a1b on chromosome 2 showed a similar expression pattern to Col5a1 in the region of the developing somites ( first and third columns). Expression of both Col11a1a on chromosome 24 and Col11a2 was detected in the notochord, in the region of the otic vesicles, and in the developing ethmoid plate and trabeculae. ( and ). Col11a1a was expressed in polar cartilage, whereas Col11a2 exhibited strong expression within the parachordal cartilage.
Figure 5 Spatial expression patterns of minor fibrillar collagen genes in zebrafish at 48 hpf by whole mount in situ hybridization. The expression patterns of Col5a1, Col5a3, Col11a1b, Col11a1a, and Col11a2 are shown in both dorsal (A) and lateral (B) views. Expression (more ...)
Figure 6 Spatial expression patterns of minor fibrillar collagen genes in zebrafish at 72 hpf by whole mount in situ hybridization. The expression patterns of Col5a1, Col5a3, Col11a1b, Col11a1a, and Col11a2 are shown in dorsal (A), lateral (B and C), and ventral (more ...)
At 72 hpf, Col5a1 had a similar expression patterns as that observed for 48 hpf (). Additionally, Col5a1 was detected strongly in the parachordal cartilage, ceratohyal, trabeculae, and madibular cartilage, the outlining of the cranial cavity (). The expression of Col5a3 was detectable in the notochord, in the region of the retinal pigmented epithelium, and lens (). Col11a1b on chromosome 2 demonstrated similar expression patterns as those of Col5a1 (, column 1 and 3). Both Col11a1a on chromosome 24 and Col11a2 were expressed in the notochord, ceratohyal, palatoquadrate and Meckel’s cartilage in addition to those structures that were detected at 48 hpf ( column 4 and 5). Controls showed no significant signal at all time points.
Our study has identified orthologs of minor fibrillar collagens in zebrafish, characterizing and contrasting the expression patterns. Col11a1a on chromosome 24 and Col11a1b on chromosome 2 as well as Col5a1 have been extensively characterized during early development in zebrafish by our study. For Col5a1, the strong expression in the neural crest, head mesenchyme and in the region of developing somites may indicate a broad spectrum of functions. Sequence alignment and synteny analysis suggest that the gene on chromosome 2 is most closely related to Col11a1. Our results show that this gene does not undergo alternative splicing within the VR, and exhibits a similar expression pattern to that of Col5a1. Interestingly, we have identified transient expression of Col11a1a on chromosome 24 around 16 hpf, which includes hindbrain and otic placode. This finding has not been reported in other species. With regard to the collagenopathies of Col11a1 in humans, the transient expression in hindbrain and otic placode may contribute to the clinical symptoms that have been described in Marshall’s and Stickler syndromes. The developmental events associated with transient expression require further investigation.
Baas and colleagues reported the identification and characterization of collagen XI cDNA using zebrafish genomic sequence reported to be the ortholog of human COL11A2 (GenBank accession number AL672176 and XM685088). Within the coding sequence, they identified characteristic features of all minor fibrillar collagens, a 60 amino acid variable region, and a carboxyl terminal domain with the length, number and location of cysteine residues within the NC1 domain that is more consistent with the identity as an ortholog of human COL11A1. (Baas et al 2009
.) Overall amino acid sequence identity however, indicates that this minor fibrillar collagen alpha chain is more likely to be an ortholog of human COL11A2. Sequence comparison of the six amino acids between cysteines 3 and 4 of the amino propeptide domain was also cited as supporting the identification of this minor fibrillar collagen gene as Col11a1 rather than the previous identification as Col11a2. Both COL11A1 and COL11A2 play critical roles in vertebrate craniofacial development and mutations in either of these genes can result in Stickler syndrome. Therefore, it is possible that the in situ
hybridization data and characterization of the morphants carried out by Baas and colleagues may actually reflect the essential role that Col11a2 plays in craniofacial development and morphogenesis in the zebrafish rather than Col11a1, as their title indicates. It is possible that all minor fibrillar collagens belonging to the V/XI collagen family share common function as well as serving unique roles in tissue and species-specific manners.
Additional findings from our study include the detection of exon 6b of Col11a1a at 72 hpf. While not reported in previously related work (Baas et al., 2009
, Hoffman et al., 2010), sequence corresponding to exon 6b was previously submitted to GenBank as a result of expression profiling and comparative genomics analysis (Dickmeis et al., 2004
). Characterization of Col11a1b on chr2 in our study agrees with findings by Hoffman and colleagues (Hoffman et al., 2010). A short form of Col11a2 was reported by Hoffman and colleagues; here we report several spliceforms, including a form with an extensive alternatively spliced exon of approximately 1.2 kb (termed exon 6′ in our current study). While expression of this exon has not been previously reported, sequence corresponding to this region was included in the entry from Howden (Howden, P. Direct Submission, Submitted (12-JAN-2009) Wellcome Trust Sanger Institute, Hinxton, Cambridgeshire, CB10 1SA, UK). Results from our laboratory agree with those reported by Hoffman and colleagues for Col5a1, and to a large extent for Col5a3 with only minor sequence differences.