Dystroglycan (DG) is a cell surface adhesion complex, originally isolated from rabbit skeletal muscle, representing the pivotal element of a multimeric complex defined as dystrophin-glycoprotein complex (DGC). In mammals DAG1 possesses an uncomplicated exon-intron-exon structure, and its transcription and translation generates a precursor protein that is post-translationally cleaved into two noncovalently associated subunits: the highly glycosylated extracellular α-DG and the transmembrane β-DG 1. The DG subunits are believed to establish a molecular bridge linking the extracellular matrix to the cytoskeleton 2. In skeletal muscle and in a wide variety of tissues α-DG binds extracellular matrix molecules, such as laminins, agrins and perlecan, and interacts non covalently with β-DG, that binds dystrophin via its cytoplasmic tail 3. Several cDNA sequences, which in most cases correspond to a highly conserved protein product 895 aa long, have been reported in different organisms such as human, mouse, dog, amphibia and fish DGs. The degree of sequence identity among mammals is remarkably high (> 90%), while the recently identified cDNA sequences of X. laevis and D. rerio (zebrafish) confirm that a very high degree of similarity is found also in lower vertebrate species 45.

DG is believed to have an increasingly important role in human health, being involved in pathological processes ranging from cancer progression to infective diseases 6. In particular, in human skeletal muscle DG, as well as several proteins belonging to the DGC (like dystrophin and sarcoglycans), is involved in severe forms of muscular diseases 7. On the other hand, until now there are no reports about muscular diseases directly generated by DAG1 mutations (primary dystroglycanopathies), not surprisingly since the DG knockout experiment in mice causes an early arrest of the embryonic development (at day 6.5), due to the disruption of the Reichert's membrane 8. However, in particular muscular diseases, known as congenital dystrophies (Muscle-Eye-Brain disease, MEB; Fukuyama Congenital Muscular Dystrophy, FCMD; Walker-Warburg Syndrome, WWS), mutations in different genes encoding for glycosyltransferases are regarded to generate an abnormal glycosylation of α-DG 91011. This alteration of the glycosylation pattern of α-DG compromises its binding to extracellular matrix molecules and it is thought to be the reason for the progressive muscle fibre degeneration; this kind of human congenital disorders have been defined as "secondary dystroglycanopathies" 12.

In the last years, a large body of knowledge originated from comparative biochemical and physiological studies about dystroglycan and the dystrophin glycoprotein complex in D. rerio 5131415, which showed that DG indeed plays a crucial role for adult skeletal muscle stability 5. With the aim of carrying out an expanded genetic and biochemical comparative analysis, we examined DAG1 from several fish species; besides D. labrax (sea bass) and D. rerio (zebrafish), we also analysed pufferfish characterized by compact genomes (T. nigroviridis and T. rubripes), other teleosts such as O. latipes (medaka) and G. aculeatus (stickleback), and Antarctic species (T. bernacchii and C. hamatus). So far it was generically assumed that all vertebrate species would share only one copy of DAG1, even if a whole genome duplication (WGD) event, involving a large number of genes, has been described in Actinopterygii 16.

Although a DG gene duplication event has not been identified in D. rerio 5, our computer mining of genomic data available for pufferfish indicates that two different DG sequences are present. Accordingly, via the analysis of DNAs (and cDNAs) from T. nigroviridis, we identified two functional paralogous DG sequences, hereinafter defined as DAG1a and DAG1b. Moreover, for the first time we have cloned and sequenced DAG1 in sea bass (D. labrax), showing that it contains an additional mini-intronic sequence of about 150 bp, which is properly spliced out upon transcription.

During recent years, the biological role of dystroglycan (DG) in higher vertebrates has been in part elucidated. The DG adhesion complex, composed of two subunits (α and β), is a pivotal member of a large transmembraneous group of glycoproteins associated with the cytoskeleton representing, together with integrins, the major molecular bridge involved in the formation and stabilization of contacts at the cell/extracellular matrix interface during embryogenesis and in a wide variety of adult tissues 817. In mice, the concerted action of DG and laminin is believed to trigger the initial phase of embryogenesis, when the first contacts between cells and basement membranes are established. In fact, DAG1 knockout mice exhibit gross developmental abnormalities beginning around 6.5 days of gestation, while in contrast heterozygous mice are viable and fertile 8. However, the role of DG during embryogenesis remains controversial. Although no mutations have been identified so far in human populations, thus confirming the DG crucial primary role during peri-implantation in mammals, knock-out experiments in zebrafish showed that early development remained unaffected by the absence of DG while a severe dystrophic phenotype emerged during adulthood 5.

The comparison of DAG1 among different vertebrate species, including several fish species and even antarctic ones, which typically underwent the evolutionary process of cold-adaptation, could be useful to understand how the selection pressure influenced the actual organization of DAG1 in fish and the whole genome duplication process. In fact, several lines of evidence suggest that a whole-genome duplication (WGD) event occurred within the teleost lineage after separation from the tetrapod lineage, and that only a subset of duplicates have been retained in modern teleost genomes 16. The analysis of genomic sequences obtained from zebrafish and pufferfish provided further evidence for WGD during the evolution of ray-finned fish (Actinopterygii) 1620. It was estimated that WGD should have taken place about 350 Myr ago, after the separation of ray-finned and lobe-finned fish, but before the origination of teleost fish 21. While several duplicated genes were subsequently lost, many others were maintained during evolution. Preserved genes might have underwent small changes and adopted slightly different functions and this might have further protected the gene from being lost 2223. These assumptions are of primary importance when searching for possible orthologous versions of mammalian genes in fish genomes 2425.

The major piece of data collected so far on the structure and function of DAG1 in zebrafish is the work published by Parsons and colleagues, which shows that the inactivation of the DG gene by antisense morpholino oligonucleotides causes severe muscular dystrophy in the adult stage 5. Genome analysis reveals that only one copy of DAG1 is present in D. rerio, displaying the typical uncomplicated exon/intron mammalian structure 26. On the other hand, the analysis of available genomic sequence drafts from T. rubripes, T. nigroviridis, O. latipes and G. aculeatus, reveals the presence of two ORFs encoding DG, that we here name as DAG1a and DAG1b, based on their alignment scores with respect to other mammalian DGs and in particular to human DG (see Fig. 1, Table 1 and 2).

Surprisingly, the gene copy that we propose to define DAG1a, displays a novel intronic sequence at the level of the region corresponding to the second exon. The intron is very short in size: 137 bp in T. rubripes and T. nigroviridis, 116 bp in G. aculeatus and D. labrax and only 86 bp in O. latipes (see Table 1) in close similarity with the shortest sizes of introns already identified in other species 27. The gain of this "mini-intron" did not produce any frameshift affecting the resulting protein sequence, as also demonstrated by our Western blot results (see below). Accordingly, experiments performed with specific primer pairs designed for both DAG1a and DAG1b, reveals that in pufferfish both the DAG1 copies are transcribed and therefore likely to be functional and expressed (Fig. 2). This result was somehow anticipated by the high conservation of both paralogous DAG1 sequences and by the absence of nonsense mutations or any other major genetic alteration that would imply a drift towards a pseudogene status. In fact, pseudogenes are known to constantly drift until they are either deleted or become unrecognizable 28. However, further analysis will be needed to investigate in detail such intron gaining event 29.

As already reported for several other genes, it is likely that DAG1 underwent duplication as part of the whole genome duplication (WGD) event that took place during the Actinopterygii speciation process 1625 (black arrow in Fig. 4) and subsequently a sporadic gain of a mini-intronic sequence took place either before the separation between Ostariophysi and Acanthopterygii (green arrow in Fig. 4) or afterwards (red arrow).

In D. labrax (sea bass), the result of our homologous cloning strategy for DG fishing was a gene fragment of ≈ 2000 bp (data not shown), including a sequence corresponding to a 116 bp mini-intron which, based on the alignment score, can be assigned to the family of DAG1a sequences (Table 3). The expression of DG was preliminary tested by Western blot using a monoclonal antibody directed versus the C-terminal tail of the β-DG subunit, since this region is highly conserved in all the vertebrates 30. In fact, positive signals of 43 kDa were detected in all the samples analysed, including antarctic species 131415. Up to now, any attempt at homologous cloning of DG sequences from antarctic species exploiting the same primers employed for D. labrax DG were unsuccessful. Therefore, further experiments employing new designed degenerate primers will be required in order to clone the DG sequences from antarctic species.

The secondary structure of the α-DG N-terminal region of T. rubripes, predicted from the gene sequence (both DAG1a and DAG1b) exploiting SSpro software http://www.igb.uci.edu/tools/scratch3132 (data not shown), suggests a significant similarity with the α-helical and β-strand elements detected in the crystal structure of mouse α-DG N-terminal domain that was recently solved 33. This region is composed by two autonomous domains: an Ig-like one, and the second one resembling ribosomal RNA-binding protein S6. Moreover, additional predictions performed using NetOGlyc software, confirm the presence of a mucin-like domain also in the central region of teleost α-DGs. It is noteworthy that the β-DG binding epitope, spanning the amino acid positions 550–565 of the C-terminal domain of α-DG, is highly conserved also in fish 34. In contrast, its counterpart, the putative α-DG binding epitope spanning the amino acidic positions 691–719 in β-DG, displays a much lower degree of identity with the mammalian one, with few exceptions, such as the conservation of Phe692 and Phe718, which have been shown to play a crucial role in the α/β subunits interface formation (Fig. 1) 35. Another region highly conserved is the C-terminal domain of β-DG which contains the dystrophin binding site 173036.

Our analysis clearly shows that the WGD event that took place in Actinopterygii involved also DAG1. During evolution, WGD events are expected to have had a high impact on speciation. To fully understand this impact means to unravel all the genetic and molecular details underlying the speciation process, and the knowledge of which genes were retained in duplicate and how the duplication modified their evolutionary fitness is crucial to that aim. Generally, the functional consequences of WGD in fish have been mitigated both by partial gene loss and acquisition of new useful functions 37. Indeed, in some cases the presence of two functional copies of an important gene like DG, could have represented an improvement of their fitness. The morpholino oligos-driven disruption of DG in zebrafish results in the emergence of a severe dystrophic phenotype in adults that could not be alleviated or compensated by a paralogue isoform of DG 5.

The importance of fish model systems for the study of Duchenne muscular dystrophy and other human muscular diseases is clearly emerging, as highlighted by the work carried out in Kunkel's lab 38. Due to the important role played by DG in human congenital muscular disorders, comparative genetic and biochemical analyses could be particularly relevant in the race for fully elucidating its function or misfunction in severe diseases, eventually leading to innovative therapeutical strategies related to DG. For example, the DG's high affinity towards the proteoglycan agrin has been one of the factors leading to the design of miniaturized agrin, rescuing the dy/dy dystrophic phenotype in mice 3940.

A comprehensive understanding of the biological implications of DAG1 duplication in some teleost fish species may have unexpected repercussions on the view of "secondary dystroglycanopathies", since recently paralogue isoforms of glycosyltransferases thought to act specifically on DG (LARGE and POMT among others) were also identified and characterized 4142. It is intriguing to hypothesize that in the future "evolution-inspired" gene therapy approaches, implying the introduction (or reintroduction) of a second DAG1 copy or isoform, will be used to alleviate the symptoms of dystrophy in human skeletal muscle.

EP did all the experimental work including the collection, storage and homogenization of the fish tissue samples and also gave a major contribution to the writing of the manuscript and assembly of the final version of figures and tables. DC carried out some of the PCR experiments dealing with the identification of two expressed sequence in cDNA and readapted the fish pictures. MO carried out extensive sequence alignment analysis. RT carried out the Western blot analysis. AG and BG contributed to edit the manuscript. AB conceived the project, supervised the experiments and wrote the manuscript together with EP. All authors read and approved the final manuscript.