Nucleotide polymorphisms in the canine Noggin gene and their distribution among dog (Canis lupus familiaris) breeds.

Noggin (NOG) is an important regulator for the signaling of bone morphogenetic proteins. In this study, we sequenced the complete coding sequence of the canine NOG gene and characterized the nucleotide polymorphisms. The sequence length varied from 717 to 729 bp, depending on the number of a 6-bp tandem repeat unit (GGCGCG), an insertion that has not been observed in other mammalian NOG genes investigated to date. It results in extensions of (Gly-Ala)3-5 in the putative NOG protein. To survey the distribution of these tandem repeat polymorphisms, we analyzed 126 individuals in seven dog breeds. We identified only three alleles: (GGCGCG)3, (GGCGCG)4, and (GGCGCG)5. Although the allele frequencies were remarkably different among the breeds, the three alleles were present in all seven of the breeds and did not show any deviation from Hardy-Weinberg equilibrium.

Introduction

Noggin (NOG) is a member of the bone morphogenetic protein (BMP) family and a major extracellular antagonist of the BMP. NOG regulates dorsal induction and joint formation in embryonic development by blocking BMP signaling (Smith and Harland 1992; Zimmerman et al. 1996; Brunet et al. 1998; McMahon et al. 1998; Canalis et al. 2003). NOG also controls bone regeneration and homeostasis in adults (Canalis et al. 2003; Canalis 2009). The coding sequences of mammalian NOG genes are contained in a single exon and are highly conserved among species (Valenzuela et al. 1995). Nucleotide and amino acid sequence similarities are more than 90% among humans, mice, and other mammals (Valenzuela et al. 1995). In humans, several point mutations in the NOG gene have been linked to congenital skeletal malformations, for example, proximal symphalangism or multiple synostoses syndrome (Gong et al. 1999; Marcelino et al. 2001; Hirshoren et al. 2008). Therefore, the NOG gene is a potential candidate for involvement in congenital skeletal malformations in domesticated dogs. Partial DNA sequences at this locus have been defined by the Canine Genome Project (NW_876332; Canis familiaris chromosome 9 genomic contig). To the best of our knowledge, no nucleotide polymorphisms have been described in the canine NOG gene. In the present study, we first determined the DNA sequence of the complete coding sequence of the canine NOG gene, to scan for all the possible nucleotide polymorphisms encoded at this locus. We subsequently surveyed the allelic distribution of the variable number of tandem repeats (VNTRs) that were identified in the canine NOG gene.

Materials and Methods

Tissue and nail samples from non-sibling dogs were collected at veterinary hospitals and dog-grooming shops in Japan; 127 samples were obtained from 24 long-haired Miniature Dachshunds, 25 long-haired Chihuahuas, 23 Miniature Poodles, 19 Shih Tzus, 13 Papillons, 12 Malteses, 10 Yorkshire Terriers, and 1 Beagle. Genomic DNA was extracted from tissue or nails by a standard phenol–chloroform extraction method (Sambrook et al. 1989) or using the QuickGene DNA tissue kit (Fuji Film, Tokyo, Japan).

Amplification and Sequencing of Canine NOG Coding Sequence

Four animals from three breeds were selected for DNA sequence determination of the complete coding sequence for the canine NOG gene (one Beagle, two Miniature Dachshunds, and one Chihuahua). Primers for polymerase chain reaction (PCR) and sequencing were designed from dog genome draft sequences (chromosome 9, NW_876332); rat (NM_012990), mouse (NM_008711), and human (NM_005450) NOG gene sequences (Table 1). The primer sets of CanNOG-F1 and CanNOG-R1 produce about 1.1 kbp DNA products, whereas the primer sets of CanNOG-F2 and CanNOG-R2 produce about 700 bp. The two products overlapped with one another in the canine NOG coding sequence.

Table 1

Primers for canine NOG gene

Primer	Sequence (5′ to 3′)	Position on NW_876332a
CanNOG-F1	TGGTGATGGAGCTGAAAGTG	13273923–13273942
CanNOG-R1	ACCACAGCCACATCTGTAACTTC	13274986–13275008
CanNOG-F2	ACCCGGACCCTATCTTTGAC	13274685–13274704
CanNOG-R2	TTTCTGGCTACAGAGACCTAGCT	13275418–13275440
CanNOG-S1b	GAAGTTACAGATGTGGCTGTGGT	Reverse complement of R1
CanNOG-S2b	GTCAAAGATAGGGTCCGGGT	Reverse complement of F2

a Canis familiaris chromosome 9 genomic contig, whole genome shotgun sequence

bPrimers used only for sequencing

The PCR was performed according to the manufacturer’s instructions for PrimeStar HS DNA polymerase with GC buffer (Takara Biotechnology, Shiga, Japan). The program included first denaturation at 98°C for 2 min; 35 cycles of denaturation at 98°C for 20 s, annealing at 55°C for 5 s, and extension at 72°C for 60 s; and final extension at 72°C for 3 min. The amplified products of each reaction were cloned into the plasmid pCR-Blunt II-TOPO using the Zero Blunt TOPO PCR Cloning kit (Invitrogen, Carlsbad, USA). At least five clones were isolated from each individual and sequenced using the Big Dye Terminator version 3.1 Cycle Sequencing Kit and ABI 3100 sequencer (Applied Biosystems, Foster City, USA) with M13 universal primers and additional sequencing primers (Table 1).

Detection of the VNTR Polymorphisms in the NOG Coding Sequence

In order to survey the distribution of the VNTR alleles in dog breeds, we analyzed 126 animals from seven breeds (Table 2). The 300 bp DNA products of the canine NOG gene, including the 6-bp tandem repeats (GGCGCG), were PCR-amplified using primer sets of CanNOG-F2 and CanNOG-R1 and then purified with the High Pure PCR Product Purification Kit (Roche Applied Science, Mannheim, Germany). Purified DNAs were sequenced directly with the primer CanNOG-F2 as described above. Heterozygous individuals that had a different number of tandem repeats were genotyped from their electropherograms.

Table 2

Genotype and allele frequency of (GGCGCG)n VNTR in canine NOG gene among seven dog breeds

Breed (number of animals sampled)	Genotype						Allele frequency
	3	4	5	3/4	3/5	4/5	3	4	5
Miniature Dachshund, long hair (24)	1	7	0	7	7	2	0.333	0.479	0.188
Chihuahua, long hair (25)	3	0	6	5	7	4	0.360	0.180	0.460
Miniature poodle (23)	2	2	2	5	6	6	0.326	0.326	0.348
Shih Tzu (19)	9	0	2	3	4	1	0.658	0.105	0.237
Papillon (13)	5	0	0	2	4	2	0.615	0.154	0.231
Maltese (12)	2	1	0	0	6	3	0.417	0.208	0.375
Yorkshire Terrier (10)	8	0	0	1	1	0	0.900	0.050	0.050

Results and Discussion

In four domestic dogs, we sequenced the entire coding region of the NOG gene, which included an unidentified region of 91 nucleotides in NW_876332. These nucleotide sequences were entered into the DDBJ, EMBL, and GenBank nucleotide databases with accession numbers AB544074–AB544077. After multiple alignment of canine sequences with human NOG (NM_005450) and mouse Nog (NM_008711), we found that the coding sequence of the canine NOG gene is longer than that of humans and mice. The coding sequence of the human and mouse genes is consistently 699 bp in length. The coding sequence in the four dogs showed three polymorphisms, of 717, 723, and 729 bp, caused by inserted sequences of 18, 24, and 30 bp, respectively. These inserted sequences in the canine NOG gene were constructed by different numbers of the 6-bp tandem repeats (GGCGCG) unit (Fig. 1). These GGCGCG tandem repeats immediately preceded the inserted sequences. Thus, these sequence-length polymorphisms are the result of the duplicated insertion of this repeat unit. The insertions of tandem repeats were aligned with the reading frame of the NOG gene and found to encode a Gly–Ala repeat (Fig. 1). These (Gly–Ala)3–5 insertions are located in the polyglycine loop which is present only in mammalian Noggin (Fig. 2). The polyglycine loop is the most flexible part of the mammalian Noggin protein, as a complex with the BMP ligand (Groppe et al. 2002). Excluding these inserted-sequence regions, the coding sequence of the canine NOG gene displayed 96.7% sequence similarity with the nucleotides of human NOG and 94.0% with mouse Nog. No other polymorphic sites were found among the four canine NOG sequences, except the above-mentioned inserted sequences.

Fig. 1

The coding region of the canine NOG gene and polymorphisms detected in this study. a Structure of the coding region of the canine NOG gene. The solid upper line indicates the partial sequence of canine chromosome 9 (NW_876332), and the broken line represents the undeciphered region in NW_876332. The start codon of canine NOG was located in the unidentified region, −91 bp from position 13274612 on NW_876332. The white blocks of the diagram represent the canine NOG coding sequence, and the shaded block indicates the region of the inserted sequence. b The nucleotide and amino acid sequences around the canine-specific insertion. The repeating units are ruled off by vertical lines. The numbers at the bottom indicate the number of repetitions. Repetition of GGCGCG was translated into the amino acids Gly–Ala

Fig. 2

Comparison of partial amino acid sequences of vertebrate Noggin. Asterisks indicate the conserved amino acids among the six species. The gray box encloses the polyglycine loop (Gly–Gly–Gly–Gly–Gly–Ala–Ala) insertion, which is present only in mammalian Noggin (Groppe et al. 2002). The highlighted letters indicate extension of the Gly–Ala repeat in the canine Noggin. The Gly–Ala repeat is connected to the polyglycine loop. Citation of sequences: human NM_005450, mouse NM_008711, pig NM_001143691, dog AF544077, chicken NM_204123, and Xenopus NM_00185644

We sequenced part of the NOG gene, including the tandem repeat (GGCGCG)n, in 126 dogs from seven breeds, and identified three alleles, (GGCGCG)3, (GGCGCG)4, and (GGCGCG)5, among these individuals (Table 2). The genotypic and allelic frequencies of these VNTR alleles did not deviate from Hardy–Weinberg equilibrium among the seven breeds. Although the three alleles of the VNTR were present in all the dog breeds analyzed in this study, distributions varied widely among the breeds. (GGCGCG)4 had the highest frequency in the Miniature Dachshund, whereas (GGCGCG)3 was extremely high in the Shih Tzu, Papillon, and Yorkshire Terrier (Table 2). Since the three NOG VNTR alleles were distributed widely among various dog breeds, these sequence-length polymorphisms arose long before the modern dog breeds were established several hundred years ago.

Because we collected DNA samples from visibly healthy dogs, each VNTR allele in the canine NOG gene did not appear to be linked to fatal outcomes and severe pathological phenotypes in young dogs. Considering the important role of NOG in bone regeneration and homeostasis in adults (Canalis 2009), however, the three canine NOG variations theoretically could be associated with susceptibility to bone or joint disorders in the lifetime of the domestic dog. Nevertheless, the polymorphisms are most likely silent with respect to function given that (1) the additional residues encoded by the repeats were inserted in the most flexible portion of the Noggin protein, (2) the segment is well removed from the ligand-binding interface, (3) the loop is expanded in other mammals with no apparent loss of function, and (4) the inserted residues are small and functionally inert in the context of a disordered loop (Supplementary Fig. 1). In contrast, polymorphisms resulting in insertion of tracts of polyglutamine in transcription factors can have a toxic gain-of-function effect in neurodegenerative disorders, triggering conformational changes that disrupt gene regulation (Riley and Orr, 2006) or culminate in formation of amyloid-like fibrils (Robertson et al. 2011). Indeed, insertion of glutamine repeats in Runx-2 protein, a transcription factor that regulates osteoblast differentiation, can have dramatic effects on canine skeletal morphogenesis (Fondon and Garner 2004). Introduction of amide-containing side-chains, with two functional groups capable of hydrogen bonding, can enhance interactions within multiprotein transcriptional complexes, leading to upregulation of target genes involved in skeletogenesis. Therefore, although polymorphisms within domestic dog breeds can have pronounced effects on development, the structural context and nature of the side-chain of the inserted residues appears to dictate whether the effect is functionally silent, accumulating over evolutionarily large time spans, or pronounced, capable of generating rapid diversity of form.

Below is the link to the electronic supplementary material.

Supplementary Fig. 1. Crystal structure of human Noggin in complex with BMP-7. (A) Cartoon representation of the homodimeric Noggin and BMP-7 proteins. Monomers are differentially colored for clarity. The terminal residues of the model that flank the disordered polyglycine loop (Pro88, Gly96) are indicated. The perspective is approximately perpendicular to the two-fold symmetry axis, with the complex tilted slightly behind the plane of the image and to the right (cf. Fig. 2, Groppe et al. 2002). (B) Zoomed view of the site of insertion of the encoded tandem repeat residues, the disordered polyglycine loop. The 2Fo-Fc electron density map (gray mesh), contoured at 1.5 sigma, is shown for one of the Noggin monomers, as well as a line representation of the side-chains. Note the diminished and fragmented nature of the map for the backbone and side-chains of residues preceding the disordered loop, which lacked detectable density for model building, hence was omitted. The graphic model was rendered from coordinates of the Noggin:BMP-7 complex (Protein Data Bank accession code 1M4U) with the MacPyMOL Molecular Graphics System, Schrödinger, LLC. The 2Fo-Fc map was created with the Fast Fourier Transform program (FFT) in the CCP4i Suite (TIFF 571 kb)