The zebrafish progranulin gene family and antisense transcripts
- 9.6k Downloads
Progranulin is an epithelial tissue growth factor (also known as proepithelin, acrogranin and PC-cell-derived growth factor) that has been implicated in development, wound healing and in the progression of many cancers. The single mammalian progranulin gene encodes a glycoprotein precursor consisting of seven and one half tandemly repeated non-identical copies of the cystine-rich granulin motif. A genome-wide duplication event hypothesized to have occurred at the base of the teleost radiation predicts that mammalian progranulin may be represented by two co-orthologues in zebrafish.
The cDNAs encoding two zebrafish granulin precursors, progranulins-A and -B, were characterized and found to contain 10 and 9 copies of the granulin motif respectively. The cDNAs and genes encoding the two forms of granulin, progranulins-1 and -2, were also cloned and sequenced. Both latter peptides were found to be encoded by precursors with a simplified architecture consisting of one and one half copies of the granulin motif. A cDNA encoding a chimeric progranulin which likely arises through the mechanism of trans-splicing between grn1 and grn2 was also characterized. A non-coding RNA gene with antisense complementarity to both grn1 and grn2 was identified which may have functional implications with respect to gene dosage, as well as in restricting the formation of the chimeric form of progranulin. Chromosomal localization of the four progranulin (grn) genes reveals syntenic conservation for grna only, suggesting that it is the true orthologue of mammalian grn. RT-PCR and whole-mount in situ hybridization analysis of zebrafish grns during development reveals that combined expression of grna and grnb, but not grn1 and grn2, recapitulate many of the expression patterns observed for the murine counterpart. This includes maternal deposition, widespread central nervous system distribution and specific localization within the epithelial compartments of various organs.
In support of the duplication-degeneration-complementation model of duplicate gene retention, partitioning of expression between grna and grnb was observed in the intermediate cell mass and yolk syncytial layer, respectively. Taken together these expression patterns suggest that the function of an ancestral grn gene has been devolved upon four paralogues in zebrafish.
KeywordsAntisense Transcript Head Kidney Antisense Transcription Yolk Syncytial Layer Pronephric Duct
List of abbreviations
apical ectodermal ridge
central nervous system
enveloping monolayer of cells
hematopoietic stem cell
intermediate cell mass
lateral plate mesoderm
million years ago
single sequence length polymorphism
yolk syncitial layer.
In mammals, a single progranulin gene, also known as proepithelin, acrogranin and PC-cell-derived growth factor (PCDGF), encodes a glycoprotein precursor exhibiting pleiotropic tissue growth factor activity (reviewed in [1, 2, 3, 4]). Progranulin is secreted in an intact form [5, 6, 7, 8], or undergoes proteolysis leading to the release of its constituent peptides, the granulins [9, 10, 11]. Individual granulins have an approximate molecular weight of 6 kDa, and are structurally defined by the presence of 12 cysteines arranged in a characteristic motif: X2–3C X5–6C X5CC X8CC X6CC X5CC X4C X5–6C X2 . Comparison of the biosynthetic origin of granulin peptides in various mammals reveals that all are commonly derived from a precursor composed of one amino-terminal half followed by seven non-identical copies of the granulin motif.
A role of progranulin in mammalian embryogenesis has been suggested. The exogenous addition of recombinant progranulin to eight-cell stage mouse embryos grown in culture ex vivo accelerates the onset of cavitation, stimulates the rate of blastocoel expansion, and leads to an increase in the number of trophectoderm cells compared to controls . Conversely, the use of a progranulin function-blocking antibody arrests growth and prohibits embryo implantation [13, 14]. These results are consistent with the growth-promoting activity of progranulin upon epithelial cells in vitro.
Despite these advances, the evolutionary history and phylogenetic distribution of the progranulin gene outside the mammalian radiation remain largely unexplored. In order to shed light on this issue, and to establish a model for studying the functional contribution of progranulin to vertebrate development, we undertook the characterization of the biosynthetic origins of progranulins in the zebrafish. The widely documented evidence in favor of a pan-genomic duplication event at the base of the teleost radiation , commonly referred to as 3R, predicts that the single mammalian progranulin gene will likely be represented by two zebrafish co-orthologues [16, 17]. Results reported here demonstrate that, in zebrafish, progranulins arise as members of an extended gene family represented by two distinct architectures, in excess of that predicted by 3R.
Comparative chromosomal mapping of the various progranulin genes (grns) was performed to assist in the establishment of an orthologous relationship to their mammalian counterpart and to provide a point of reference to discuss the evolutionary origins of the distinct progranulin architectures. In support of the duplication-degeneration-complementation (DDC) model  gene expression analysis of the zebrafish progranulins reveals spatio-temporal divergence among the different family members, possibly reflecting extensive functional devolution of an ancestral form. Also, the occurrence of natural antisense transcription to some grns suggests that gene dosage may have influenced the retention of extra grn paralogues in zebrafish.
Evidence for a progranulin multigene family in teleosts
Zebrafish progranulins are represented by two distinct architectures
Progranulin-1 and progranulin-2
Hybrid grn RNA
Antisense progranulin1-2 gene
During the cloning of grn1, genomic sequences were used to perform BLAST searches for sequences deposited at NCBI . An EST was detected (GenBank accession AW777232) whose sequence was an exact match to a portion of the grn1 gene, but in the reverse complement orientation. The full characterization of this EST, designated ASgrn1-2, revealed that it is spliced and shares exonic complementarity to exons 2 and 3, in addition to flanking intronic sequences, of both grn1 and grn2 genes (Figure 3 and Additional File 6). This observation suggested that grn2 is located upstream of grn1 in a head-to-tail organization, thus providing strong support for trans-splicing between the grn1 and grn2 primary transcripts as a mechanism for the genesis of the hybrid progranulin RNA (Figure 3).
In addition to its partial complementarity to grn1 and grn2, the last exon of the ASgrn1-2 gene (corresponding to nucleotides 1015 to 1989 of the cDNA) shares a high degree of sequence conservation with the tzf transposon , a subclass belonging to the Tc1/mariner superfamily of class II DNA mobile elements . However, this mobile element is in the reverse complement orientation within the antisense transcript, and has undergone extensive mutations resulting from nucleotide insertions, deletions and point mutations (Additional File 7). Thus, ASgrn1-2 never possessed the ability to encode a translatable transposase protein, nor does it have a clearly predictable ORF in view of the presence of several termination codons in all three possible reading frames. For these reasons, this naturally occurring antisense transcript is considered to belong to the category of non-coding RNA genes.
The zebrafish co-orthologues of mammalian progranulin
Zebrafish progranulin-A and progranulin-B are 48.6% identical over their aligned sequences, and each is similarly related to human progranulin with 44.8% and 42.9% identity, respectively. As expected, sequence conservation is seen primarily within the aligned granulin domains of the zebrafish precursors (Figure 4). However, the deduced granulins within the zebrafish grnA and grnB precursors cannot be aligned strictly on the basis of the mammalian nomenclature (i.e. in the order granulin-G, -F, -B, -A, -C, -D, -E) [5, 26, 27, 28]. Indeed, progranulin-A contains five granulin peptides (domains 4, 6, 7, 8, 9) whose sequences bear close sequence similarity with one-another and with that of human granulin-A. In addition, zebrafish progranulin-A encodes a deduced granulin structure (domain 2) displaying the modified cysteine motif found in human granulin G, which is characterized by the absence of cysteine residues 4 and 7 (Figure 4).
Northern blot analysis
The size of the cloned sense and antisense zebrafish progranulin transcripts was assessed by northern blot analysis (Additional File 8). With the exception of the putative ASgrna (see below), the observed transcript sizes were in agreement with those predicted from cloned sequences. However, many pre-mRNA transcripts of higher than expected molecular weight were identified for several family members. Only grnb demonstrated the presence of splice variants of unknown composition.
Chromosomal mapping of zebrafish grns
Assessment of zebrafish grn gene expression by RT-PCR
Zebrafish grn gene expression in adult tissues
Developmental expression of Zebrafish grn genes
Assessment of zebrafish grn gene expression by whole mount in situ hybridization
In order to evaluate the relationship between sense and antisense transcription, and to gain insights into the potential contributions of the four grn paralogues to development, the spatio-temporal expression of zebrafish grns was monitored by whole-mount in situ hybridization. For all stages examined, sense and antisense riboprobes for sonic hedgehog were used as experimental controls since the tissue expression of this gene is discrete and well documented . No non-specific hybridization signal for sonic hedgehog was detected (data not shown). Unless stated otherwise, the respective sense riboprobes to grna and grnb did not give rise to detectable signals.
Grna and grnb
The sense riboprobe corresponding to grna but not grnb detects staining in the brain, intestine and pronephros at this stage (Figure 9, panel A and B, b), suggesting either non-specific hybridization or the presence of an antisense transcript. This observation prompted a search for sequences deposited at GenBank corresponding to parts of the grna cDNA sequence, but in the reverse complement orientation. Three unidirectionally cloned ESTs (accessions CD585878, CD585963 and CD596001) were all revealed upon sequencing to correspond to a 914 nucleotide long cDNA sharing perfect complementarity to nucleotides 2701–3614 of the grna cDNA (Additional File 11, panel A). This sequence corresponds to the 3'UTR region of the grna gene and is not bisected by an intron at the genomic level. This precludes the conclusion that this candidate antisense transcript is not an artifact of cloning. To confirm its directionality, cDNA was synthesized for subsequent PCR amplification using a primer located within the 3'UTR exon of grna that shared complementarity to ASgrna (sense relative to grna) or using another primer that was located downstream of the cloned ASgrna sequence and within a known intron for grna (Additional File 11, panel A). This RT-PCR strategy suggests that the ASgrna deduced from cloned EST sequences may represent a splice variant, and that antisense transcription extends further in the 3'direction (Additional File 11, panel B).
Grn1 and grn2
Zebrafish progranulins: simplified molecular forms and orthologues of the mammalian gene
The granulin peptide family was originally discovered as a component of the granule fraction of mammalian phagocytic leukocytes. A series of related cysteine-rich peptides (designated granulins A, B, C and D) were purified from extracts of human neutrophils . The definition of the structure of human progranulin as a glycoprotein bearing multiple copies of the granulin motif made it apparent that granulin peptides are generated through proteolytic cleavage of this precursor within the phagolysosomal compartment of the neutrophil. This explains the roughly equimolar ratios observed for members of the granulin family that are co-packaged within this subcellular compartment. The hematopoietic tissues (spleen and head kidney) of the carp (Cyprinus carpio) were also shown to be abundant sources of three granulin-like peptides (granulins-1, -2 and -3) . However, the non-stoichiometric ratios observed for carp granulins-1,-2 and -3 suggested that the granulin gene family in this teleost species may expand beyond the prototypic single grn gene found in mammals. Specifically, carp spleen contains mainly granulin-1, whereas granulins-1, -2 and -3 were found in extracts of the head kidney (Belcourt et al. 1993). To simplify the identification of teleost granulin gene family members, the zebrafish (Danio rerio) was chosen based on its usefulness as a model of vertebrate development and disease.
The presence of extra gene paralogues in teleost fish has been extensively documented, often [15, 31, 32] but not invariably supporting [33, 34] the hypothesis that the actinopterygian (ray-finned) lineage underwent an additional round of genome duplication (3R) subsequent to diverging from the sarcopterygian (lobe-finned) lineage approximately 450 mya. Specific examples include the Hox clusters [16, 35], the annexins , the claudins  and the Nodal-related genes squint and cyclops . Further, comparative chromosomal mapping studies have shown that duplicated zebrafish genes often reside on distinct chromosomes that exhibit extensive blocs of conserved synteny with their mammalian counterpart [39, 40]. Using this approach, it has been estimated that approximately 20% of human genes may be represented by two co-orthologues in the genome of zebrafish .
As predicted by 3R, we demonstrate that grns are members of an extended gene family in zebrafish. We have identified two deduced precursors, progranulins A and B, harbouring 10 and 9 granulin peptide repeats respectively that bear close structural and sequence relationship to human progranulin (Figure 4). Grna was localized to a region of LG3 known to show syntenic correspondence to where human grn is found on human chromosome 17 (Figure 5). Grnb was localized to LG 24 rather than being positioned on LG12, predicted to bear synteny with LG3 . Despite this apparent discrepancy, a co-orthologous relationship between grna and grnb relative to mammalian grn is supported by their sequence conservation as well as their extensive overlapping expression patterns observed during development.
Two smaller grn genes, grn1 and grn2, each encoding one full and one amino-terminal half copies only of the granulin motif, were also characterized (Figure 2). Interestingly, a common ancestry of the smaller zebrafish grn genes with grna and grnb, and thus mammalian grn, is implied by conservation of the strict exonic organization that these genes display. However, it is unlikely that grn1 and grn2 arose from the postulated whole genome duplication event corresponding to 3R since they were found to be localized in tandem on LG19 (Figure 5). It is notable that both grn1 and grn2 are linked to a Hox cluster and dlx gene paralogues, similar to that observed for grna. This suggests that a smaller grn may have originated coincidentally with the duplication of a Hox-bearing chromosome in a primordial species. Evidence suggests that this putative structure would then have been retained within the teleosts but lost within the sarcopterygian line of evolution leading to mammals.
Analysis of progranulin expression by RT-PCR
The DDC model predicts that an important driving force behind the retention of duplicated genes is the devolution of an ancestral function onto the resultant pair through quantitative and qualitative changes in gene expression which, when combined, may reflect the sum of the ancestral expression pattern . Thus, it was of interest to determine the extent of expression partitioning and overlap between the two grn gene classes during development by RT-PCR and to compare these patterns to those known for the mammalian counterpart.
As an initial survey of the differential expression patterns of the zebrafish grns, we conducted semi-quantitative RT-PCR analyses using adult tissues and staged embryos. Similar to the well documented widespread expression pattern of human [26, 27, 42], rat , mouse and guinea pig grns in several tissues and cell lines of epithelial, mesenchymal, and hematopoietic origin, both grna and grnb were observed to be expressed ubiquitously in several adult zebrafish organs. In contrast, grn1 and grn2 exhibit a more restricted pattern of expression. It was previously shown that carp granulin-1 and granulin-2 peptides are differently distributed in the spleen and head kidneys of the carp . Interestingly, RT-PCR experiments demonstrate that the expression of the homologous structures in zebrafish were similarly uncoupled at the level of mRNA, (Figure 6, panel C), reflected in a lack of detectable zebrafish grn2 expression in adult spleen. Also, grn1 appears to be the predominant form in the heart, while the eyes express higher levels of grn2. Whether the widespread pattern of expression for zebrafish grns is due in part to leukocyte entrapment in some organs, cells known to express granulins in carp  and goldfish  cannot be determined using this experimental approach.
Other notable differences in the expression of the paralogous and orthologous pairs of genes were observed. First, maternal transcripts for grna and grnb are readily detectable, with grnb showing higher abundance than grna until early somitogenesis when the two genes become expressed at similar levels (Figure 7, panel A). In contrast, although very low levels of grn1 mRNA are also detected in the newly-fertilized egg, the combined expression of grn1 and grn2 only becomes detectable by 30 hpf, at a stage when most organogenesis is well advanced (Figure 7, panel A). Thus, the absence of grn1 and grn2 expression prior to the onset of zygotic expression argues that these genes are functionally dispensable in early embryogenesis.
Analysis of progranulin expression by whole mount in situ hybridization
In order to determine possible roles for the grns during development, their spatio-temporal distribution was examined by whole mount in situ hybridization. Overall, grna and grnb expression patterns share conserved features with their murine orthologue in the early embryo. Transcripts for both zebrafish co-orthologues are maternally deposited and remain ubiquitously expressed subsequent to the onset of zygotic transcription (mid-blastula transition – 3 hpf). Similarly, in a pattern that reflects the replacement of maternal mRNAs with zygotically expressed transcripts , murine grn mRNA levels fall rapidly after egg fertilization, reaching negligible levels as early as the 2-cell stage, but rise again to detectable levels by the eight-cell stage  Notably, this precedes the morula stage and subsequent blastocyst stage when the epithelium is first formed. It is interesting to note that during zebrafish epiboly, which comprises the morphogenetic movements of the blastoderm towards the vegetal pole (dome to tailbud stage – 4.5 hpf to 10 hpf), grna and grnb are still ubiquitous but more intense in the outer enveloping monolayer of cells (EVL), which ultimately will give rise to an epithelium covering the blastoderm. Expression of these grns in the EVL and later in the skin ectoderm is reminiscent of the more elevated levels of expression for grn in the apical surface of the mouse blastocyst epithelium, the trophectoderm, relative to the inner cell mass population .
Although several regions display increased expression of grna and grnb during brain segmentation (Figure 8, panel A and B, c–f), regional specificity is more apparent at 24 hpf. Distinct or non-overlapping patterns observed include grna expression within the tectum and a more expansive grnb expression pattern encompassing the midbrain-hindbrain boundary, tegmentum and telencephalon (Figure 8, panel B, g). Despite the expression of both grna and grnb in the epithelial lining of the eyes and lens, transient expression within the retina and tectum is noticeable for grna only, suggesting that this orthologue may affect the development of the retinotectal projections. Similar functions may also be implied for mammalian grn, given its expression within the retinal glia during murine development . Interestingly, murine grn expression is abundant throughout the central and peripheral nervous system, similar to that observed for grnb at later developmental time points within the zebrafish CNS (Figure 8 and 9). The generally unrestricted gene expression pattern suggests a role for progranulin in cell survival or proliferation or as a competence factor. Indeed, mammalian grn has been demonstrated to be a potent glial cell mitogen in vitro  and is consistently up-regulated in malignant human gliomas [47, 48]. Grn2 and ASgrn1-2 but not grn1 demonstrate similar unrestricted expression within the zebrafish brain (Figure 11d,h). Taken together these expression patterns suggest that CNS development in the zebrafish may involve a functional interplay between the various molecular forms of granulin.
The only known functional and physiological contribution that grn gene expression is known to make during neural maturation is its involvement in the sexual differentiation of the male rat brain. It has been shown that male sexual behaviour is associated with steroid-dependent grn expression in the male neonatal hypothalamus [49, 50]. Similar distinctions in grn gene expression were not determined in the present study given that sexual differentiation of zebrafish gonads occurs much later and spans roughly 21–28 dpf .
The expression of the zebrafish grn gene family members also displays overlapping and distinctive patterns with respect to the endoderm and tissues derived from this germ layer. At 24 hpf both grna and grnb can be found within the pharyngeal and foregut endoderm, whereas only the latter is located within the YSL (Figure 8, panels A and B, g). Both transcripts maintain a degree of dispersed endodermal expression until 120 hpf where grna is located within the epithelial lining of the stomach and anterior intestine while grnb can be found within the YSL, pancreas and proctodeum (Figure 8 and 9).
Unlike the expression of grna and grnb, the paralogues grn1, grn2 and hybrid granulin are highly abundant and for the most part restricted to pharyngeal and visceral endodermal derivatives from 72 hpf onward. Although grn1 and grn2 demonstrate some endodermal tissue-specific expression, these transcripts often co-localize (Figure 11a,b,d,e). The restricted expression of hybrid granulin within the proctodeum is particularly striking (Figure 11g). Consistent with the manner in which the hybrid transcript is formed, both grn1 and grn2 are likewise expressed in the proctodeum. However, expression of these two transcripts overlaps within a large portion of the intestine and stomach where no hybrid grn transcript is found. This suggests that wherever the hybrid grn is observed, the generation of this chimeric peptide must be a highly regulated process and that a specific function is implied.
In contrast to the observed grn distribution in zebrafish, murine grn is not detected within developing endodermal derivatives, with the exception of adult deep crypt enterocytes [42, 46]. This suggests that mammalian dependence on grn expression may be developmentally restricted to endodermal-epithelial transitions in the gut or subsequent maintenance of this organ. Alternatively, it is possible that endodermal expression of zebrafish grns reflects a species-specific requirement.
Of particular interest in regards to functional equivalence between species and functional separation between duplicated co-orthologues, is the almost complete partitioning of grn hematopoietic expression onto grna and to a much lesser extent grn2. In zebrafish, primitive hematopoiesis occurs within the anterior ICM from where nucleated erythroblasts originate and myeloid cells that can be seen circulating at 24 hpf . Hematopoietic stem cells (HSCs) are then believed to populate the dorsal aorta and yolk sac which represent the zebrafish equivalent of the mammalian aorta-gonad-mesonephros, the tissue presumed to be responsible for later definitive erythropoiesis [53, 54]. The posterior ICM, located within the ventral tail and positive for molecular markers of all three hematopoietic lineages, may represent a secondary zebrafish HSC population or region required for HSC maturation . Unlike the murine model, zebrafish definitive hematopoiesis undergoes a migratory transition from the dorsal aorta/ventral tail to the kidney (roughly 96 hpf), without the involvement of liver or bone marrow.
In accordance with expression patterns mentioned previously, grna only acquires distinctive tissue-specificity at 24 hpf, as is the case for its prevalence within the caudal ICM (Figure 8, panel A, g–j), restricting its involvement to definitive hematopoietic waves. Grna can be found at low levels within the dorsal aorta at 48 hpf (data not shown) and is highly expressed within the caudal-ventral tail region throughout all stages post-24 hpf (Figure 8 and 9), suggesting its involvement in multiple hematopoietic lineages. Significantly, sustained grna expression in this hematopoietic organ is coupled with the appearance of grna-expressing leukocytes dispersed throughout the animal with levels that peak at approximately 72 hpf (Figure 8, panel A, j–k). Thus grna expression in presumed granulocytes all over the body of the animal may suggest its involvement in the innate immune response of the host. Grn2 expression is also found within peripheral leukocytes, but in a sporadic pattern that is distinct from that observed for grna. Whether these differences reflect leukocyte sub-populations or activation states for these cells has not been addressed. Commensurate with the transition of ICM to kidney as the major site of hematopoiesis, grna can be found within the pronephric ducts (Figure 9) along with grn1 and grn2, which are also present in the head kidney (Figure 11, panel B and E).
Mammalian grn exhibits a similar expression pattern, particularly in neutrophils. Furthermore, murine grn can modulate the inflammatory response during wound healing, acting as both a chemokinetic factor and inhibitor of neutrophil degranulation and respiratory burst [56, 57]. It remains to be determined whether grna supports a similar role in zebrafish.
During the initial degenerate primer amplification of cDNAs encoding grn1 and 2, a third cDNA was cloned and identified as sharing strict identity with portions of both grn1 and 2. Interestingly, this granulin-hybrid showed 100% identity with exons 1 and 2 of grn-1, and with exons 3, 4 and 5 of grn-2, suggesting that this hybrid cDNA may represent a splicing of granulin-1 and 2 primary transcripts (Figure 3). Chimeric transcripts usually result from one of the following mechanisms: chromosomal translocations, transcription of neighboring genes as a single transcription unit or alternative splicing in trans. In all cases, joining of exons is predicted to occur through the recognition of canonical splice acceptor and donor sites.
A hybrid granulin structure has been previously reported through cloning of cDNA sequences in the rat . Specifically, a structural splice variant of progranulin cDNA was retrieved and predicted to encode a granulin domain consisting of the amino-terminal domain of granulin-C fused to the carboxyl-terminal domain of granulin-D, consistent with the removal of an exon from the larger primary transcript [22, 28]. The zebrafish hybrid grn described here likely originates through a mechanism other than alternative splicing from a larger primary transcript since the grn2 gene is located 5' to the grn1 gene (Figure 3). This topology was confirmed through the cloning and structural analysis of the partially complementary ASgrn1-2 gene. Also, we found no evidence for the presence of additional grn1-like genomic sequences located upstream of the grn2 gene, or elsewhere in the genome by Southern analysis (data not shown). In particular, no equivalent of carp grn3 was found in zebrafish. These observations suggest that the presence of hybrid grn in zebrafish likely occurs through a splicing reaction in trans between grn1 and grn2 pre-mRNAs, similar to the mechanism originally documented in trypanosomatids . Although rare, scrambled or intergenic RNA molecules consisting of exons originating from distinct genes through a trans-splicing reaction have also been documented in vertebrates. For instance, acyl-CoA:cholesterol acyltransferase-1 (ACAT-1) and the CYP3A family of P450 cytochrome genes produce hybrid mRNA variants in humans [59, 60]. Trans-splicing of the voltage-gated sodium channel in response to nerve growth factor stimulation, further suggests that this alternative mode of splicing can be a regulated process . Indeed, there is evidence suggesting that splicing in trans may be facilitated through the recognition of regulatory elements within transcript sequences called splicing enhancers that require binding of SR proteins for activity . Whether grn1 or grn2 harbour a necessary enhancer sequence that could explain the directionality of hybrid grn is currently not known. We believe the hybrid granulin represents the first example of trans-splicing with regards to the modification of a growth factor gene product.
Although the majority of deposited zebrafish EST library sequencing confirmed the existence of grn1 and grn2, one particular sequence (AW777232) corresponded to the exact reverse complement to both grn1 and grn2 within the same exon/intron spanning region (exons 2–3 and intervening intron), and was named ASgrn1-2 accordingly. In addition, ASgrn1-2 harbours sequences for an extensively mutated tzf transposon (Tc1/mariner superfamily) in its last exon, but in the reverse complement orientation (Figure 3 and Additional File 7). Despite its polyadenylation, the lack of an ORF classifies ASgrn1-2 as a non-coding RNA. To our knowledge, ASgrn1-2 represents the first example of a single spliced transcript with antisense complementarity to two tandemly organized paralogous protein-coding genes.
The existence of ASgrn1-2 has potential implications in aspects of grn1 and grn2 function. Classically, antisense transcripts often function as inhibitors of the expression of their associated gene, through repression of transcription or promotion of mRNA degradation. There is the possibility that ASgrn1-2 is involved in the formation of hybrid grn wherein this complementary transcript may provide a scaffold to sequester grn1 and grn2 mRNA transcripts within the same intracellular locale, preventing or alternatively facilitating the trans-splicing reaction.
Regarding the RT-PCR data, despite evidence of clear tissue-specific and temporal ASgrn1-2 expression in adult tissues and various developmental stages respectively, this transcript shows no clear reciprocal relationship to grn1 or grn2 (Figures 6 and 7). These expression patterns were confirmed at all stages examined using sense riboprobes for grn1, grn2, and verified for specificity by using the corresponding sense riboprobe to hybrid grn as negative control (Figures 10 and 11, and data not shown). At 5 dpf, ASgrn1-2 is expressed in the presumed hyoid ossification center, stomach and rostral intestine, as well as swim bladder (Figure 11i–n). Although grn1 and grn2 are found within most of these tissues (particularly grn1), specific expression of ASgrn1-2 in the hyoid region undergoing osteogenesis may suggest the required down-regulation of the expression of its counterpart genes during bone development. Therefore a reciprocal relationship may exist between these genes under strict spatio-temporal regulation. Although ASgrn1-2 may mediate a degradation independent mode of gene regulation, such as alternative splicing, it clearly does not perform as a universal negative regulator of grn1 and grn2 expression.
At least one type of transcription modulation may be associated with ASgrn1-2 based on whole mount in situ hybridization analysis. There is a clear reciprocal relationship between hybrid grn expression and absence of ASgrn1-2 transcription. At 120 hpf, hybrid grn is restricted to the distal intestine and despite the expression of its substrate transcripts within several other locales, no other grn1/grn2 rich setting is devoid of ASgrn1-2 (Figure 11). This pattern further suggests that ASgrn1-2 expression prevents formation of the hybrid grn RNA.
The existence of ASgrn1-2 suggested that an equivalent entity may exist for one or both co-orthologues, grna and grnb. Indeed, the sense probe for grna in whole mount in situ hybridization analysis produced a consistent and reproducible signal within the intestine and pronephric ducts at 5 dpf (Figure 9, panel A, b). Several unidirectionally cloned cDNAs (accession numbers (CD585878, CD585963, and CD596001) were found to correspond to the reverse complement sequence of grna within the 3'UTR. Northern blot analysis using sense grna demonstrated a putative 4 kb transcript and directional cDNA synthesis followed by RT-PCR indicated this putative antisense transcript extended within the known grna intronic sequence. However, repeated attempts to clone the full-length transcript by RACE were unsuccessful. Nevertheless, these observations provide evidence for the existence of a second naturally occurring grn complementary transcript, namely ASgrna.
The granulin motif – phylogenetic and functional implications
In contrast to the conserved protease/granulin gene architecture found in plants, members of the animal kingdom have expanded their granulin repertoire, not via genomic or segmental duplication, but rather through tandem multiplication of the granulin ORF. For instance the Xenopus leavis and the early chordate Ciona intestinalis granulin genes encode of five and six tandemly repeated near identical granulin motifs, respectively. The variance in the number of granulin intragenic regions demonstrates a degree of plasticity during grn gene expansion. Intragenic multiplication is unlikely to have occurred as a single ancestral event; rather grn gene expansion has taken place independently in various species to varying extents. This mechanism of conserved domain repetition is not unique to grn genes. Indeed, the same type of genetic expansion is likely responsible for the repetition of immunoglobulin, EGF and lectin domains in numerous proteins.
The presence of single grn genes in protostomes is not surprising since the origins of these species predate the estimated genomic duplications within the vertebrate radiation (1R and 2R, Figure 12); these large-scale events appear not to have given rise to grn gene expansion. Specifically, all members of the sarcopterygian lineage (derived from lobe-finned fish) harbour a single granulin gene of varying motif number, indicating that intragenic multiplication remains the preferred and tolerated means of granulin expansion for most vertebrates, including mammals. The same intolerance to gene duplication has not encompassed the actinopterygians (ray-finned fish), including Danio rerio and Takifugu rubripes. These species have undoubtedly expanded their granulin gene repertoire through tandem intragenic expansion as well as genome duplication (3R, Figure 12), to yield the co-orthologues grna and grnb. Interestingly, the existence of grn1 and grn2 does not necessarily conform to these methods of gene expansion, suggesting that a specific and unidentified means of gene expansion, possibly involving ASgrn1-2, may exist.
Although the existence of two co-orthologues of mammalian progranulin in zebrafish is likely a result of genome-wide duplication, similar genetic events have occurred within chordates prior to divergence of the ray-finned (teleost) and lobe-finned (mammalia) radiations. Indeed two rounds of genome-wide duplication are believed to have occurred . More precisely, the first genome duplication probably occurred in a common ancestor of all agnathans and gnathostomes after its divergence from cephalochordates, ~594 mya (million years ago). The second round is presumed to have occurred ~488 mya, within the lineage leading to jawed vertebrates after the jawless line diverged, presumably before the split between cartilaginous and bony fish. Despite this, all mammals studied thus far have retained only a single copy of the progranulin gene, whereas two rounds of genome duplication would theoretically create four progranulin genes. It is therefore interesting to consider the biological rationale behind retention of grna and grnb following the teleost genome duplication, an event not permitted within other vertebrates, in conjunction with the appearance of two extra paralogues, grn1 and grn2. Regulation by gene dosage through complementary transcription may have allowed for the retention of the smaller paralogues, while putative antisense transcription to grna may be necessary for precisely regulating the spatio-temporal activity of this growth factor.
Overall, the expression patterns of zebrafish progranulins faithfully replicate those observed for the mouse counterpart in a similar context [42, 46]. Importantly, this indicates that the use of zebrafish will enable modeling of the contributions of progranulin activity to vertebrate development through investigating both grna and grnb. These studies will be uncomplicated by the presence of grn1 and grn2, whose expression patterns largely do not overlap with the co-orthologues. Overall, the expression patterns for the grns indicate that these growth factors may subserve multiple functions in vivo that are consistent with the known role of their mammalian counterpart in cell growth, motility and survival.
Materials and methods
Tissue extraction and granulin peptide purification
Carp (Cyprinus carpio) were purchased live at a local fish market (Waldman Plus, Montreal, QUE). Peptides from fish spleens were extracted using C18 Sep-Pak cartridges (Waters Canada Ltd. Mississauga, ONT) and separated using reversed-phase high-performance liquid chromatography (RP-HPLC) on a C18 Bondapak column (Waters) as previously described [19, 68]. Column fractions were screened for cysteine content by amino acid analysis and granulin-1/2 immunoreactivity by radioimmunoassay . Fractions positive for both criteria were further purified by RP-HPLC using solvents containing 0.13% (v/v) heptafluorobutyric acid as counter-ion, and subsequently purified to homogeneity using the original solvent system containing 0.1% (v/v) trifluoroacetic acid. Molecular weight of purified peptides was determined using a Voyager matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometer (Perceptive Biosystems, Framingham, MA) located at the Sheldon Biotechnology Centre of McGill University.
Microsequencing of carp granulin-A
The putative carp granulin-A was alkylated according to a previously published protocol . Approximately 20 μg of the alkylated carp granulin-A was digested with sequencing grade chymotrypsin (Roche Diagnostics Canada, Laval, QUE) according to the manufacturer's instructions, and the resulting fragments separated by RP-HPLC on a C18 Bondapak column. Amino-terminal sequence analysis of carp granulin-A and its chymotryptic fragments was undertaken using a Procise sequencer (Applied Biosystems, Foster City, CA) located at the Sheldon Biotechnology Centre of McGill University.
Wild type zebrafish were purchased from Scientific Hatcheries (Huntington Beach, CA) and maintained on a 14 h/10 h light/dark cycle at 28.5°C in a laboratory aquarium (Allantown Aquaneering, Allantown, NJ). Fish were fed twice daily, and bred as described elsewhere . Embryos for developmental studies were collected from tanks and staged according to conventional criteria  and by hours post-fertilization (hpf).
Library screening and cloning of zebrafish progranulins
The zebrafish grn1 cDNA was cloned using a PCR strategy (Additional File 1). The carp granulin-1 amino acid sequence was used to design degenerate forward DF1 (5'-GTI ATY CAY TGY GAY GC-3') and reverse DR1 (5'-CAR CAR TGR ATI CCR TC-3') and DR2 (5'-TCR CAR TGR TAI CCR TG-3') primers for use in the polymerase chain reaction (IUPAC codes are used to refer to the bases in primer sequences). The template for the initial amplification reaction was a 5'-STRETCH plus cDNA library cloned in lambda gt10 vector (Clontech BD Biosciences, Mississauga, ONT). cDNA for this library was prepared from 1-month-old zebrafish using a combination of oligo-dT and random priming. 0.25 μl of library (approximately 108 pfu/ml) was used in a final reaction volume of 100 μl for each new amplification attempt. The annealing temperature was determined empirically in order to maximize yield of product. PCR amplifications were performed with Taq DNA polymerase, unless specified otherwise, using a Hybaid thermal cycler from Bio/Can Scientific Inc. (Etobicoke, ONT). Amplified products were isolated by agarose gel electrophoresis, purified with the QIAquick Gel Extraction Kit (Qiagen Inc. Mississauga, ONT) and sequenced after cloning into TOPO pCR2.1 (Invitrogen, Carlsbad, CA). An initial reaction using the DF1 and DR1 primer pair yielded several products. 5 μl of this reaction was subjected to re-amplification using DF1 primer in combination with the nested (anchored) DR2 primer, which revealed a product of 126-bp encoding a partial sequence for granulin-1 (Additional File 1, step 1). New grn1 primers F126 (5'-ACTGTGTGTCCAGACGG-3') and R215 (5'-CCATCCCTGCAACACTG-3') were then designed based on this sequence and were used, respectively, in combination with flanking gt10 primers in order to obtain the 5'- and 3'-untranslated region (UTR) cDNA sequences (Additional File 1, steps 2 and 3). Finally, the entire ORF was amplified with Pwo DNA polymerase (Roche Diagnostics), using forward F1 (5'-ATGTTCCCAGTGTTGATG-3') and reverse R (STOP) (5'-GCTTACAACTCCAACCCG-3') primers (Additional File 1, Step 4). This PCR was performed in a final volume of 100 μl, containing 0.25 μl of library, 50 mM KCl, 10 mM Tris-HCl, (pH 8.8), 1.5 mM MgCl2, 0.1% Triton X-100, 0.2 mM concentration of each dNTP, 0.5 unit of Pwo DNA polymerase, and 100 pmol of each primer. An initial denaturation step was carried out at 94°C for 3 min. Annealing temperatures of 54°C, 56°C and 58°C were used sequentially for 10 cycles each. Typical denaturation, annealing, and amplification reactions were carried out at 94°C for 30 sec, 54°C for 1 min, and 68°C for 1 min, respectively. A final extension step of 10 min at 72°C was carried out after adding 0.25 unit Taq DNA polymerase. An amplification product specific for grn1 was sequenced on both strands. The 5'-UTR, and a portion of the 3'-UTR for grn1, were amplified using grn1-specific primers in conjunction with a lambda gt10 primer. Distinct cDNAs encoding progranulin-2 and a chimeric progranulin were uncovered through this approach. Each transcript was confirmed through sequencing of independent amplification reactions using template cDNA derived from either adult organs or embryos of mixed stages. Following a strategy similar to that used for the isolation of zebrafish grn1 cDNA, primers based on the purified carp granulin-A peptide sequence were designed to clone partial cDNAs for zebrafish grna and grnb, respectively (not shown). BLAST searches using the cloned sequences retrieved two unique ESTs at NCBI sharing an exact match with grna and grnb, respectively (accession numbers AW174591 and AW184435). These respective ESTs were purchased from RZPD GmbH (Heidelberg, Germany; clone ID: UCDMp574E2318Q2 and UCDMp574I0223Q2) and sequenced on both strands to create a final assembly of the full-length cDNAs encoding zebrafish progranulin-A and progranulin-B. In addition to our cloning strategy, the rapid amplification of cDNA ends (RACE) was performed with the GeneRacer kit (Invitrogen, Burlington, ONT) using total RNA isolated from adult zebrafish intestine. For grn1 and grn2 transcripts, a reverse primer that corresponded to nucleotides 195–215 based on the cloned ORF of both transcripts (5'-CCATCCCTGCAACACTGACCC-3'), was used to perform the 5' RACE, while a forward primer corresponding to nucleotides 1–22 of each transcript (5'-ATGTTCCCAGTGTTGATGTTAC-3') was used to perform the RACE in the 3' direction. Similarly a 5' UTR sequence for grna was obtained using the map reverse primer (see below). Repeated RACE attempts in both directions for ASgrna were unsuccessful.
Cloning of the zebrafish grn1 gene
A zebrafish genomic library constructed in P1 artificial chromosome (PAC)  and represented on filters at high-density (RZPD GmbH) was screened for the presence of the grn1 gene using standard procedures. The cDNA bearing the grn1 ORF was labeled with [α-32P] dCTP by random priming using the Oligolabeling kit (Amersham Biosciences, Baie d'Urfe, QUE) for use as probe, and purified using a Sephadex G-50 column (Amersham Biosciences). Kodak X-OMAT AR film was used for autoradiography (Fisher Scientific Ltd, Whitby, ONT). Three positive clones (706K2254Q, BUSMP706K14116Q2, 706F20133Q2) were detected by autoradiography, and the first two were confirmed to carry at least part of the grn1 gene by PCR, using the F1 (5'-ATGTTCCCAGTGTTGATG-3') and R215 (5'-CCATCCCTGCAACACTG-3') primer pair which does not discriminate between grn1 and grn2, and sequencing. DNA from a positive clone (706K2254Q) was purified with the Plasmid Midi Kit (Qiagen). 1.5 μg of this DNA was subjected to restriction digest with EcoRI to generate fragments suitable for cloning into pBluescript II KS (Stratagene, La Jolla, CA), and was followed by transformation in TOP 10F' electrocompetent cells (Invitrogen). Screening of colonies transferred onto nitrocellulose membranes (Xymotech, Montreal, QUE), employing the same probe used for the original library screening, was performed in the following prehybridization and hybridization conditions: 2 × SSC, 0.5% SDS, 0.05% Na Pyrophosphate at 65°C. Membranes were washed twice in 1 × SSC, 0.1% SDS, 0.05% Na Pyrophosphate at 60°C for 15 min, followed by two washes in 0.1% SSC, 0.1% SDS, 0.05% Na Pyrophosphate at 60°C for 10 min. Plasmid DNA from a positive clone was purified using the high pure plasmid isolation kit (Roche Diagnostics, Laval, QUE). An insert of ~9-kb was fully sequenced and revealed the presence of the promoter region and approximately half of the grn1 gene. The remaining gene sequence was found in a ~6-kb insert clone isolated by re-screening the colony lifts with 32P-labeled reverse R(STOP) oligonucleotide (5'-GCTTACAACTCCAACCCG-3') as probe. A PCR was performed using primers flanking this EcoRI site, and sequenced to confirm that the isolated 9 kb and 6 kb clones represent continuous sequences within the original PAC clone.
Retrieval of antisense transcripts from NCBI
While in the process of analyzing cloned grn1 genomic sequences (data not shown) through BLAST searches for corresponding sequences at GenBank, an EST harbouring sequences corresponding to unspliced grn1, but in the reverse complement orientation, was noticed. This clone (accession number AW777232) was purchased (RZPD, clone ID: DKFZp717B091Q2) and further analyzed through sequencing. A putative transcript exhibiting perfect complementarity to the 3'UTR region of zebrafish grna (ASgrna) was deduced from sequencing four unidirectionally cloned ESTs deposited at GenBank (CD585878, CD585963, CD596001 and CD588938) that originated from an oligo-dT-primed cDNA synthesis from adult kidney marrow RNA (Song et al. 2004; kindly provided by Dr. Chen, Shanghai Institute of Biological Science). Unidirectional cDNA synthesis using total RNA derived 5 day-old larvae was synthesized using a sense primer relative to the 3'UTR exon (ASgrna 2) or to a known intron (ASgrna 3) of grna (Additional File 10 and 11). These primers were then used in conjunction with the following primer (ASgrna 1) in subsequent RT-PCR to confirm antisense transcription to grna.
Chromosomal assignment and syntenic analysis
Zebrafish grns were mapped using the LN54 Radiation Hybrid Panel as previously described . Primers for each gene are noted in Additional File 9. Each PCR reaction was carried in a final volume of 20 μl containing 100 ng "hybrid DNA", 500 mM KCl, 100 mM Tris-HCl (pH 8.3), 15 mM MgCl2, 0.2 mM each dNTP, 1 unit Taq DNA polymerase, and 5 pmol of each oligo. Denaturation, annealing and amplification were performed at 94°C for 30 sec, 55°C (grn1 and grn2) or 60°C (grna and grnb) for 30 sec, and 72°C for 30 sec, respectively, followed by an extension step of 7 min at 72°C. To determine syntenic relationships between zebrafish and human genomes, mapped zebrafish genes flanking a given zebrafish grn gene were identified using the consolidated zebrafish maps available from ZFIN  and data from LocusLink .
Gene expression profiling by RT-PCR
Total RNA from various adult tissues and developmental stages was isolated using Trizol LS reagent (Gibco BRL, Burlington, ONT), treated with DNaseI and used in first strand synthesis using the Revert Aid H-synthesis kit (MBI Fermentas Inc. Burlington, ONT). PCR conditions used for each family member consisted of an initial denaturation at 94°C for 2 min, followed by 30–40 cycles at 94°C for 45 sec, gene-specific annealing temperature (Additional File 10) for 1 min and extension at 72°C for 1 min, with a final single cycle extension at 72°C for 7 min. To discriminate between grn1, grn2 and hybrid grn, an (NH4)2SO4 buffer with 1 mM MgCl2 (MBI Fermentas) was used for RT-PCR in conjunction with cloned template controls (pBluescript, Clonetech) for all three forms using the same reaction parameters. Two independent reverse primers for grn1 and grn2 yielded products of expected size and hybrid grn was produced using grn1 forward with either of the grn2 reverse primers. All PCR products were resolved on 2% agarose gels, ethidium bromide stained and visualized on Polaroid 667 Film. The authenticity of all PCR products was confirmed by sequencing after cloning into TOPO/pCR2.1.
Northern blot analysis
Full-length mRNA transcript size was assessed for each progranulin family member by Northern analysis of poly-A enriched mRNA (Micro Poly (A) Purist Small Scale Purification Kit; Ambion) derived from whole adult or 5 dpf animals . Hybridization (Ultrahyb; Ambion, Austin, TX) was carried out using non-radioactive biotin-labeled cRNA probes (Psoralen-Biotin Non-iosotopic labeling Kit; Ambion) and detected with Brightstar BioDetect Nonisotopic Detection Kit (Ambion) according to manufacturers instructions. Band size was determined using pre-labeled biotin markers (Ambion).
Whole-mount mRNA in situ hybridization
In situ hybridization for progranulin family gene expression was carried out essentially as previously described . Briefly, digoxigenin-labeled RNA probes for each full-length cDNA, with the exception of ASgrn-1/2 which corresponded to exons 1–3 only, were hybridized at 70°C using various developmental stages from cleavage to larval. In some cases, polyvinyl alcohol was added to the staining solution in order to minimize the occurrence of background, especially when the reaction was required to proceed for several days . Stained whole-mount and sectioned embryos were mounted in glycerol and visualized under a Leica MZFLIII stereomicroscope (Richmond Hill, ONT). Pictures were taken with a Leica DC350F camera and processed with Adobe Photoshop 7.0 software.
Sequence Accession Numbers
GenBank accession numbers of all zebrafish proranulin genes and antisense transcripts described in this paper are as fellows: grn1, AF273479; grn2, AF273480; hybrid grn, AF273481; ASgrn1-2, AY289607; grna, AF375477, ASgrna, AY826190; grnb, AY289606 .
This work was supported by an operating grant (MOP-53105) from the Canadian Institute of Health Research. We are indebted to Drs. Marc Ekker and Marie-Andrée Akimenko (University of Ottawa) for their hospitality and for sharing their technical expertise with the whole mount in situ hybridisation technique during the initial phases of this project, as well as for performing the linkage analyses on the LN54 mapping panel. We also thank Ms. Jo-Ann Bader (Molecular Oncology Group, Royal Victoria Hospital, Montreal) for providing assistance with sectioning and staining tissue sections.
- 5.Baba T, Hoff HB, Nemoto H, Lee H, Orth J, Arai Y, Gerton GL: Acrogranin, an acrosomal cysteine-rich glycoprotein, is the precursor of the growth-modulating peptides, granulins, and epithelins, and is expressed in somatic as well as male germ cells. Mol Reprod Dev. 1993, 34 (3): 233-243. 10.1002/mrd.1080340302.PubMedCrossRefGoogle Scholar
- 8.Xu SQ, Tang D, Chamberlain S, Pronk G, Masiarz FR, Kaur S, Prisco M, Zanocco-Marani T, Baserga R: The granulin/epithelin precursor abrogates the requirement for the insulin-like growth factor 1 receptor for growth in vitro. J Biol Chem. 1998, 273 (32): 20078-20083. 10.1074/jbc.273.32.20078.PubMedCrossRefGoogle Scholar
- 14.Qin J, Diaz-Cueto L, Schwarze JE, Takahashi Y, Imai M, Isuzugawa K, Yamamoto S, Chang KT, Gerton GL, Imakawa K: Effects of Progranulin on Blastocyst Hatching and Subsequent Adhesion and Outgrowth in the Mouse. Biol Reprod. 2005Google Scholar
- 17.Postlethwait JH, Yan YL, Gates MA, Horne S, Amores A, Brownlie A, Donovan A, Egan ES, Force A, Gong Z, Goutel C, Fritz A, Kelsh R, Knapik E, Liao E, Paw B, Ransom D, Singer A, Thomson M, Abduljabbar TS, Yelick P, Beier D, Joly JS, Larhammar D, Rosa F, Westerfield M, Zon LI, Johnson SL, Talbot WS: Vertebrate genome evolution and the zebrafish gene map. Nat Genet. 1998, 18 (4): 345-349. 10.1038/ng0498-345.PubMedCrossRefGoogle Scholar
- 25.Lam WL, Lee TS, Gilbert W: Active transposition in zebrafish. Proc Natl Acad Sci USA. 1996, , 93: 10870-10875. 10.1073/pnas.93.20.10870.Google Scholar
- 29.Hukriede NA, Joly L, Tsang M, Miles J, Tellis P, Epstein JA, Barbazuk WB, Li FN, Paw B, Postlethwait JH, Hudson TJ, Zon LI, McPherson JD, Chevrette M, Dawid IB, Johnson SL, Ekker M: Radiation hybrid mapping of the zebrafish genome. Proc Natl Acad Sci U S A. 1999, 96 (17): 9745-9750. 10.1073/pnas.96.17.9745.PubMedPubMedCentralCrossRefGoogle Scholar
- 31.Jaillon O, Aury JM, Brunet F, Petit JL, Stange-Thomann N, Mauceli E, Bouneau L, Fischer C, Ozouf-Costaz C, Bernot A, Nicaud S, Jaffe D, Fisher S, Lutfalla G, Dossat C, Segurens B, Dasilva C, Salanoubat M, Levy M, Boudet N, Castellano S, Anthouard V, Jubin C, Castelli V, Katinka M, Vacherie B, Biemont C, Skalli Z, Cattolico L, Poulain J, De Berardinis V, Cruaud C, Duprat S, Brottier P, Coutanceau JP, Gouzy J, Parra G, Lardier G, Chapple C, McKernan KJ, McEwan P, Bosak S, Kellis M, Volff JN, Guigo R, Zody MC, Mesirov J, Lindblad-Toh K, Birren B, Nusbaum C, Kahn D, Robinson-Rechavi M, Laudet V, Schachter V, Quetier F, Saurin W, Scarpelli C, Wincker P, Lander ES, Weissenbach J, Roest Crollius H: Genome duplication in the teleost fish Tetraodon nigroviridis reveals the early vertebrate proto-karyotype. Nature. 2004, 431 (7011): 946-957. 10.1038/nature03025.PubMedCrossRefGoogle Scholar
- 57.Zhu J, Nathan C, Jin W, Sim D, Ashcroft GS, Wahl SM, Lacomis L, Erdjument-Bromage H, Tempst P, Wright CD, Ding A: Conversion of proepithelin to epithelins: roles of SLPI and elastase in host defense and wound repair. Cell. 2002, 111 (6): 867-878. 10.1016/S0092-8674(02)01141-8.PubMedCrossRefGoogle Scholar
- 59.Li BL, Li XL, Duan ZJ, Lee O, Lin S, Ma ZM, Chang CC, Yang XY, Park JP, Mohandas TK: Human acyl-CoA:cholesterol acyltransferase-1 (ACAT-1) gene organization and evidence that the 4.3-kilobase ACAT-1 mRNA is produced from two different chromosomes. J Biol Chem. 1999, 274: 11060-11071. 10.1074/jbc.274.16.11060.PubMedCrossRefGoogle Scholar
- 63.The Pfam database of protein families and HMMs. [http://pfam.wustl.edu/index.html]
- 64.European Bioinformatics Institute -Interpro. [http://www.ebi.ac.uk/interpro/]
- 72.ZFIN. [http://zfin.org/ZFIN/]
- 73.LocusLink. [http://www.ncbi.nlm.nih.gov/LocusLink/]
- 74.Sambrook J, Fritsch EF, Maniatis T: Molecular cloning - A laboratory manual. Second edition. 1989, Cold Spring Harbor Laboratory PressGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.