Cytidine monophospho-N-acetylneuraminic acid hydroxylase (CMAH) mutations associated with the domestic cat AB blood group

Background The cat has one common blood group with two major serotypes, blood type A that is dominant to type B. A rare type AB may also be allelic and is suspected to be recessive to A and dominant to B. Cat blood type antigens are defined, N-glycolylneuraminic acid (NeuGc) is associated with type A and N-acetylneuraminic acid (NeuAc) with type B. The enzyme cytidine monophospho-N-acetylneuraminic acid hydroxylase (CMAH) determines the sugar bound to the red cell by converting NeuAc to NeuGc. Thus, mutations in CMAH may cause the A and B blood types. Results Genomic sequence of CMAH from eight cats and the cDNA of four cats representing all blood types were analyzed to identify causative mutations. DNA variants consistent with the blood types were genotyped in over 200 cats. Five SNPs and an indel formed haplotypes that were consistent with each blood type. Conclusion Mutations in type B cats likely disrupt the gene function of CMAH, leading to a predominance of NeuAc. Type AB concordant variants were not identified, however, cDNA species suggest an alternative allele that activates a downstream start site, leading to a CMAH protein that would be altered at the 5' region. The cat AB blood group system is proposed to be designated by three alleles, A > aab > b. The A and b CMAH alleles described herein can distinguish type A and type B cats without blood sample collections. CMAH represents the first blood group gene identified outside of non-human primates and humans.


Background
Only one major blood group system has been identified for the domestic cat, whereas a plethora of systems are recognized in dogs, humans, horses and other species [1]. Except in primates, no gene mutations have been identified that cause the many animal blood groups. The major blood group system of the domestic cat was identified in the early 1900's [2,3] and was later found to contain two types as determined by agglutination by naturally occurring alloantibodies [4,5]. The two major blood types are called type A and type B [6].
The common types are allelic, with type A dominant to type B. A rare type AB, which shows agglutination with both anti-A and anti-B, has not been as clearly defined, but is suspected to be allelic to types A and B [6,7]. Several studies have shown familial segregation for cat blood types A and B. However, families segregating for type AB are limited and the parentage of the matings was not confirmed. Therefore, the inheritance of the third type AB allele was not clearly resolved [6,7], although it appears to be recessive to type A and dominant to type B [8].
However, some cat breeds have a high frequency of type B individuals. The Siamese, Tonkinese and Burmese breeds are nearly fixed for type A, whereas the prevalence of type B individuals of British Shorthairs, Birmans, Devon and Cornish Rex can range between 25 -50% [14]. Although type B cats are rare in random bred populations, admixtures with cat breeds are increasing the frequency of type B cats in the random bred population. For example, type B cats are more common in the random bred populations of California [15] and Australia than in Europe and the rest of the United States [13]. Thus, veterinary clinics and hospitals are finding it increasingly necessary to maintain both type A and type B cats for blood transfusions.
The ability to determine feline blood types is especially important because of the phenomena of neonatal isoerythrolysis [16][17][18] and transfusion reactions [19][20][21][22] between previously non-sensitized donors. Similar to the unrelated ABO system of humans, cats often possess alloantibodies against their opposite blood type.
Blood type B cats possess strong agglutinins and hemolysins to type A red cells.
Type A cats possess alloantibodies that are less strongly reactive for type B red cells. Thus, a blood transfusion of a type B cat with type A blood can cause a severe reaction. Milder reactions occur if type B blood is transfused into a type A cat. Interestingly, blood type AB cats are universal recipients since they lack alloantibodies against either blood type. The presence of strong, naturally occurring alloantibodies in the type B cat against the A antigens also leads to neonatal isoerythrolysis for type A kittens born from a type B queen. The prevention of transfusion reactions and neonatal isoerythrolysis requires close monitoring of blood types of donor and recipient, and toms and queens [17,[23][24][25][26].
The cause and the likely enzymatic mechanism for the cat blood groups is known. The A and B blood types of cats are caused by differences in the neuraminic acid residues present on a ceramide dihexose backbone on the   (Table 1), which included the cats comprising the two purebred pedigrees (Figures 2, 3). The 18 bp indel was deleted in the type A homozygous cats, but inserted in the type B cat. All type B cats were homozygous for the identified variants. All type A cats that were obligate carriers of type B were heterozygous, suggesting 2 haplotypes controlling the cat blood types. No cats that were type A were homozygous for the type B haplotype. In addition, these haplotypes segregated concordantly with the blood types in the two breed pedigrees (Figures 2,   3). Only one coding region SNP, the silent G1266A mutation in exon 11, was inconsistent with the blood types in the expanded sample pool. Approximately 20 individuals had genotypes that were discordant with the blood types (data not shown).
Flanking sequences were insufficient to genotype the A324C SNP in additional cats.
The remaining six silent coding region mutations were not present in all the cats with the same blood type or were unique to an individual cat [Additional file 1]. GenBank accessions represent the sequences of the identified DNA variants consistent with type B cats [GenBank: EF127686].

5' UTR mRNA analysis
The cloned 5' UTR sequences detected more than two species of cDNA sequences in three of the four cats analyzed by mRNA. All of the clones from the homozygous A cat and 50% of the clones from the heterozygous A cat had sequences that extended to approximately -121 bp, which included an in-frame ATG at -102 bp ( Figure 4).
Approximately 50% of the clones from the type B cat also extended to -121 bp, which included the ATG at -99 bp and the 18 bp insertion at -53 bp. Unexpectedly, the remaining type B cDNA clones were completely missing exon 1. In these particular clones, intron 1 sequence appeared prior to exon 2 sequence, with novel ATGs 16 bp and 59 bp upstream of the splice site for the initial serine codon for exon 2 (Figures 1   & 4). Either of these intronic ATGs caused a shift in the reading frame of the protein.
The two cDNA forms found in the Type B cat were also identified in the heterozygous A and AB cats, but at a lower frequency.

Blood type AB cats
Approximately 8,650 samples of cat blood have been typed by hemagglutination assay at the collaborating service laboratory (data not shown). Approximately 133 cats were type AB, suggesting a frequency of approximately 1.5% in the European cat population. However, a majority of the type AB cats represented the Ragdoll breed (N=104) and European shorthairs (N = 18). Approximately 50% of cDNA sequences from the type AB cat were missing exon 1 and had intron 1 sequence upstream of exon 2, similar to the type B cats. Blood type AB cats were never homozygous for the haplotype of the B blood group; they were either homozygous or heterozygous A.
Unexpectedly, two type AB cats of the Sphynx breed were homozygous for the 18 bp insertion, but heterozygous for the four other mutations in the haplotypes (Table 1).

Dozens of blood group systems have been recognized by the International
Society of Blood Transfusion and 39 are genetically characterized in humans [34,35]. Extensive polymorphism is present in some blood groups of humans and some agricultural species, allowing blood groups for use in identification and parentage testing. In addition to pedigree testing, blood typing has been of crucial importance in preventing transfusion reactions and neonatal isoerythrolysis in a variety of species.
The cat AB blood group system is both unique and typical of blood groups of other species. It is unique because it appears to be the sole major blood group system. Since it has only two common alleles and the vast majority of cats have the same blood type A, the cat blood group system is also somewhat unique in not being useful for parentage and identification testing. It is typical of other animal blood groups in that cats with one blood type naturally produce alloantibodies to the opposing blood type [17,26]. Therefore, like other blood group systems, such as ABO of humans, the AB system of cats is important in transfusion reactions and neonatal isoerythrolysis [17,26].
The cat AB blood group is coded by a unique gene, which is not listed among the 39 blood type genes that have been identified in humans. This study suggests that the cat AB blood group system resulted from mutation(s) in CMAH that prevented the conversion of NeuAc to NeuGc, hence type B to type A. Several mutations, including 2 SNPs upstream of the start, an indel in the exon 1 5' UTR and three missense mutations in the coding region were concordant with the type A and type B blood types of cats representing 18 breeds from the USA and Europe. The six mutations were homozygous in all type B cats and heterozygous in all type A cats that were obligate carriers for type B. No type A cats were homozygous for these six variants.
The function of the identified mutations needs to be confirmed by protein studies, however, the functions appear simple in the type A and type B cats (see Figure  4). We hypothesize that CMAH may be extended at the amino terminus of the protein as compared to rodents, which would be unique to the cat. For both type A and type B cats, cDNA sequence extended approximately -121 bp upstream of the exon 1 start site. Within this sequence, a potential start codon is located -102 Approximately half of the mRNA sequences from the type B and type AB cats had intron 1 spliced to exon 2. Variants in the 5' upstream region may interact with transcription initiation in the mRNA, leading to the intron 1 sequence that is upstream to exon 2 for approximately 50% of the cDNA species in the type B.
Notably, the C-369T mutation disrupts a GATA transcription factor site in the type B cat, which is recognized for erythropoeitic transcription in other species. An additional GATA site is located in intron 1 and is present in both the type A and type B cats [38]. Two ATG sequences are present in this second cDNA variant with intron 1 sequence, however, in frame stop sites occur within a few codons of both sites. The first in frame ATG on the b allele would occur at the G139A SNP, which would truncate the amino terminus of the protein by 47 amino acids from the traditional start and by an additional 28 amino acids if the actual start site in the cat is the newly identified ATG at -102 bp. A similar situation would occur in type AB cats with the Ab haplotype. The first ATG in frame in a blood type AB cat with an AA haplotype is at M87. Each protein should be highly compromised and likely non-functional.
One of the most intriguing findings of this study concerned the cats that tested positive for both A and B forms of sialic acids. Within the analyzed regions, the blood type AB positive cat could not be differentiated by genomic sequence from a blood type A positive cat. However, no type AB positive cats (N=26) were homozygous for the type B haplotype. Approximately 50% of the type AB cats have the homozygous A haplotype and 50% have the heterozygous A haplotype. The blood type AB cat in the mRNA analysis had the homozygous A haplotype. Some clones had the type A cDNA form, while some had the intron 1 spliced to exon 2, as found in type B cats.
However, the coding region SNPs were homozygous for the A haplotype. Since both sialic acids are present on the red cell antigens in the type AB cats [7], CMAH must convert some NeuAc to NeuGc, which is different from a blood type A cat carrying a B allele. Although previous studies have not been definitive on the inheritance of Type AB, the present studies indicate that AB is allelic to A and B in cats, and that the unidentified mutation responsible for the AB positive blood type resides on a type A allelic background, altering splicing similarly to the type B cats. Thus, we propose a renaming of the cat AB blood group system to contain three alleles represented as A > a ab > b. The allele A would encode the wild-type and fully active enzyme, a ab the enzyme that is partially active, and b the inactive enzyme. Possible genotypes/phenotypes would be AA (Type A); Aa ab (Type A); Ab (Type A); a ab b (Type AB), and a ab a ab (Type AB), and bb (Type B).
Griot-Wenk et al. (1996) have presented strong evidence that type AB of cats is allelic, but the possibility of cis-acting alleles due to unequal recombination or linked loci was not disregarded [6]. Auer and Bell (1981)

Conclusions
The identified mutations in feline CMAH are strongly correlated with cat blood types and may represent the first causative mutations for a blood group in a nonprimate species. Genomic and cDNA sequence analyses suggest that the cat exon 1 translation initiates upstream from the putative mouse start codon, generating a protein with 28 additional amino acids at the amino terminus in the cat. Disruption of the reading frame by an 18 bp insertion in exon 1 or an altered splice likely causes enzymatic failure and the production of red cells with NeuGC, which would be the type b allele. Altered splicing and transcription sites may influence the rare, type AB blood group, however, specific mutations were not identified. The present work identifying the causative gene and concordant haplotypes for type A and type B alleles provides a powerful tool to identify bb and b allele carriers. Until the protein alterations are deduced, genotyping of all or several mutations in the type B haplotype may be warranted to clearly identify b carriers. Genetic testing can now help prevent neonatal isoerythrolysis in cat breeding by detecting b carrier cats and preventing the breeding of type B queens bred to type A toms that carry b.

Sampling
To maintain clarity within the discussions, cats that are obligate carriers of type B are presented as type Ab, whereas the third, less defined blood group, will be presented in the traditional manner as type AB. Whole blood anti-coagulated with EDTA or buccal swab samples from non-related, as well as related cats with A, B and AB blood types were obtained from owners in the USA. European cat DNA samples, which were received as part of a blood typing service, were also provided by the collaborating service laboratory (Oy Triniini Company, Helsinki, Finland). The complete sample set (N = 221) included 107 type A, 79 type B and 35 type AB cats representing 18 breeds ( Table 1). The sample set included; i) 153 purebred cats representing 18 breeds, ii) two purebred pedigrees segregating for the A and B blood types developed for the British Shorthair (N = 31) and Birman (N = 13) cat breeds (Figures 2, 3), iii) twelve obligate carrier F1 hybrid cats produced by crossing a type B Persian with two type A Oriental Shorthairs,thus producing type Ab, and iv) two random bred domestic shorthairs (

CMAH mRNA analyses
Total RNA was extracted from whole blood samples from a cat of each blood type, including a type A Oriental Shorthair, a type B Persian, a type AB Siamese and an obligate carrier of type B (type Ab) that was a hybrid between a Persian and an Oriental Shorthair) ( Table 1). RNA was extracted using the Versagene™ RNA Purification Kit (Gentra System, Minneapolis, MN) or the PAXgene™ Blood RNA System (Qiagen). Three pools of cDNA template were synthesized by reverse transcription of 1 µg of mRNA with random primers, a gene specific primer, CMAH_GSP_ 5'-GAGGTACGTGATCTGCACTT-3', and oligo dT primers using SuperScript III (Invitrogen, Carlsbad CA). PCR amplification of CMAH from the random primer pool of cDNA was performed using nested primers listed in Table 2. To obtain the 5' and 3' portions of the gene, RNA ligase-mediated rapid amplification of 5' and 3' cDNA ends (RLM-RACE) were performed using the GeneRacer™ Kit (Invitrogen). The RACE primers were also presented in Table   2. The 5' RACE used the cDNA pool generated by the CMAH specific primer listed above and the 3' RACE used the cDNA pool generated from the oligo dT primers. The PCR conditions included an initial denaturation for 2 min, 5 cycles at 95°C for 30 sec and 72°C for 1 min, 5 cycles at 95°C for 30 sec and 70°C for 1 min, 27 cycles at 95°C for 30 sec, 68°C for 30 sec and 72°C for 1 min, followed by a 10 min extension at 72°C. All PCR products were directly sequenced as previously described except that the 5'RACE-PCR products that were cloned using the TOPO TA Cloning Kit for Sequencing (Invitrogen) prior to sequencing.
Seven to twenty-four 5' RACE cDNA clones from each of the four cats were selected and sequenced.

Mutation genotyping
Six SNPs identified by genomic sequence analyses (Table 1) were screened across the entire sample set by pyrosequencing using an M13 biotinylated primer (5' AGCGCATAACAATTTCACAGG 3') ( Table 2)       The traditional ATG start at bp 1 would be the first in-frame start site.
Approximately 50% of the mRNA from the blood type AB cat also had mRNA species as depicted in figure 5c but not the G139A mutation as found on the b allele.