Identification of genetic variants in differentially expressed sequences in cattle of different metabolic type – potential genetic markers of nutrient utilization

Two cattle breeds serve as a model to identify genes and genetic variants, respectively, that are potentially associated with nutrient transformation: Holsteins bred for high milk production mainly, and Charolais bred for high body weight with outstanding muscular growth. The major differences between Charolais and Holstein regarding many general body parameters originate from differences in pathways and deposition of nutrients. In an initial experiment, expressed sequence tags (ESTs) differentially displayed between both cattle breeds were isolated by mRNA differential display in liver and intestine. Of the in total identified 277 ESTs, 79 showing the most prominent differences, were screened for single nucleotide polymorphisms (SNPs). Thirty four SNPs were detected in 30 ESTs In a direct sequencing approach based on the comparative sequencing of the corresponding amplicons generated by PCR from genomic DNA pools of 20 animals each of both cattle breeds,. Eighteen of these SNPs showed breed specific distribution of allelic variants. Occurrence of ESTs with a breed specific SNP distribution and localisation of the respective ESTs to chromosome regions known to be affecting carcass and growth traits in cattle suggest a trait association of the respective SNPs. The polymorphic nature of the SNP markers suggests that they will be useful for evaluating whether variation in these genomic regions influences nutrient pathways in cattle.

Schlüsselwörter: Einzelbasenpaarpolymorphismen, different exprimierte Sequenzen, Stoffwechseltypen, Rind Introduction Identifying the genetic and physiological background of variation in nutrient turnover of organisms is a well recognised prerequisite for optimised nutrition in relation to health and performance and also for efficient breeding strategies in livestock (HESKETH et al., 1998).Cattle breeds being primarily selected for either milk or meat production and strikingly differing in their respective phenotype represent a suitable model for study the genetic and physiological reasons of variances in nutritional turnover (KUEHN et al., 2002).The causative reasons why ruminants transform feed components preferentially in body tissues (accretion type) or milk (secretion type) are to a great extent unknown.So far systematic characterisation of metabolic types was limited to a descriptive level only.During the past 5 years many experiments have identified a high number of different quantitative trait loci (QTL) regions in cattle affecting milk performance, growth and meat quality (e. g., STONE et al., 1999;CASAS et., 2001CASAS et., , 2003CASAS et., , 2004;;KIM et al., 2003;KHATKAR et al., 2004).QTL positions and highly significant QTL effects repeatedly confirmed in independent studies emphasize the potential value of mapped QTL in selection for growth and milk performance using marker-assisted selection programs.However, efficient utilisation of mapped QTL in breeding programs requires a higher mapping resolution or the ultimate positional cloning of the corresponding genes or genetic variants, but also a better understanding of the complex physiological process at the cellular level.To obtain a better understanding of the underlying causative interactions a suitable animal model for the genetic and physiological investigations intended, an F 2 -resource family has been designed by crossing Charolais bulls and German Holstein cows.These cattle breeds representing the accretion and the secretion metabolic type, respectively, are the background for a maximum phenotypic variation introduced into a large number of traits including lactation, growth, carcass composition, and meat quality, as well as physiological characteristics (KUEHN et al., 2002).By the application of QTL mapping within this F 2 -design genomic localisation and characterisation of genetic variation of complex traits like nutrient transformation for accretion and secretion will be studied.One of the major efforts to identify the causal gene or gene variant itself is the identification of coding sequences or transcript units (COLLINS, 1995), especially those that are localized in the QTL region of interest and that show trait-associated expression.We hypothesize that allelic variation in ESTs differentially expressed between the two metabolic types may be associated with variation in nutrient turnover in cattle.Testing this hypothesis requires genetic markers for identifying allelic variants at each gene locus.In initial experiments, expressed sequence tags (ESTs) differentially displayed between Charolais and Holstein were identified, isolated by mRNA differential display (DDRT-PCR) and studied for their trait-associated expression by real-time RT-PCR (SCHWERIN et al., 1999;DORROCH et al., 2001).This paper describes the identification of 34 single nucleotide polymorphism (SNP) markers of a set of 30 of differentially expressed ESTs and characterisation of their distribution among each 20 cows of the Charolais and German Holstein breed.According to their localisation in QTL regions based on established integrated marker/gene-maps and their traitassociated expression, candidate genes were suggested.
Materials and Methods SNP screening Amplicons were generated using genomic DNA pools of each 20 bulls of both cattle breeds and amplified by PCR using EST-specific primers.A combined Charolais and Holstein DNA pool was established and used as an intermediate control.Amplicons were comparatively sequenced by Taq cycle sequencing with a model 310C sequencer (Perkin Elmer/Applied Biosystems, Weiterstadt, Germany).Potential SNPs identified according to the occurrence of two nucleotide peaks at the same position in the sequence profile were confirmed by sequencing of DNA samples of individual animals.Figure 1 demonstrates exemplarily the sequencing approach used.Primer sequences, PCR conditions, accession number of polymorphic sequences, length of overlapping fragments, sequence similarity percentage, potential homologous gene, and chromosome assignements are reported within Tables 1 and 2.

Statistical analysis
For all analyses the SAS/STAT package (SAS, 1999) was used.Distribution of allele frequencies between breeds and co-incidence of breed specific gene expression and allelic distribution were compared by χ 2 test using 2x2 contingence table.

Screening for single nucleotide polymorphisms in differently expressed sequences
In a previous study, ESTs differentially expressed between both cattle breeds were identified in liver and intestine by DDRT-PCR (DORROCH et al., 2001).Of the in total identified 277 ESTs 79 showing the most prominent differences, were screened for SNP using a direct sequencing approach based on comparative sequencing of PCR generated DNA fragments.Amplicons were generated using genomic DNA pools of each 20 unrelated bulls of both cattle breeds.A combined Charolais and Holstein DNA pool was established and used as an intermediate control.Potential SNPs identified according to the occurrence of two nucleotide peaks at the same position in the sequence profile were proved by sequencing of DNA samples of individual animals.Figure 1 demonstrates exemplarily the sequencing approach used.The use of this sequencing approach resulted in the identification of one or more SNPs in 30 ESTs of the 79 ESTs screened (Table 2).Of the identified 34 SNPs, allele distribution of 33 SNPs were comparatively analysed in each 20 cows of the Holstein and the Charolais breed, respectively.Nineteen SNPs showed significantly different distribution between both breeds.The assumed mutant variants (allele with overall lower frequency) were more frequent in the Charolais breed in 12 SNPs and in the Holstein breed in 7 SNPs, respectively.

ESTs with breed specific SNP distribution map within the vicinity of quantitative trait affecting growth and beef quality traits in cattle
Comparative analysis of chromosomal position of the 18 ESTs with SNPs showing breed-specific distribution (DORROCH et al., 2001;GOLDAMMER et al., 2002), and chromosomal localisation of quantitative trait loci affecting carcass and growth traits in cattle (STONE et al., 1999;CASAS et., 2001CASAS et., , 2003CASAS et., , 2004;;KIM et al., 2003; ) indicated that about 80 % of the polymorphic ESTs identified are located in chromosome regions known to be affecting carcass and growth traits in cattle, supporting the putative candidate gene character of the ESTs identified.
FASTA search of GenBank/EMBL database revealed that 10 ESTs showed similarity with BAC sequences, 3 ESTs did not show any similarity with database entries, whereas 4 of the polymorphic ESTs were similar to the previously described genes NACA (nascent-polypeptide-associated complex alpha polypeptide; fbn-l102), NDUFB8 (NADH-ubiquinone oxidoreductase ASHI subunit; fbn-l165), FUS/TLS (FUS/TLS gene product; fbn-i062), and UAP1 (UDP-N-acetylglucosamine pyrophosphorylase 1; fbn-l125). 1Position of the maximum of the test statistic along the chromosome In Table 3 chromosomal localisation of the loci NACA, NDUFB8, FUS/TLS, and UAP1 (DORROCH et al., 2001;GOLDAMMER et al., 2002) is accompanied with results from studies on detection of QTLs for carcass and growth traits in cattle (STONE et al., 1999;CASA et al., 2003CASA et al., , 2004)).The somatic cell panel mapping data represent the ESTs assignment to a chromosome and the radiation hybrid mapping data the loci of ESTs characterised by the neighbouring microsatellite markers in the RH 5000 second-generation cattle map (EVERTS VAN DER WIND et al., 2004).In addition, the position of the corresponding microsatellite markers in the bovine linkage map (IHARA et al., 2005) is given.
Discussion This paper presents data on identification and distribution of genetic variants of DNA sequences potentially associated with energy transformation in cattle, because of their trait-associated expression.Based on trait-associated mRNA pattern identified by DDRT-PCR (SCHWERIN et al., 1999;DORROCH et al., 2001) of 79 ESTs screened for potential trait-associated SNPs 30 ESTs were identified exhibiting one or more SNPs, of which 19 SNPs (18 ESTs) showed significantly different distribution between cows of Holstein and Charolais breed, respectively.Localisation of 15 of the 18 ESTs with SNPs showing breed-specific distribution (DORROCH et al., 2001;GOLDAMMER et al., 2002) in chromosomal regions known to harbour quantitative trait loci affecting carcass and growth traits in cattle (TAYLOR et al., 1998;STONE et al., 1999;REXROAD et al., 2001;CASAS et., 2003CASAS et., , 2004;;KIM et al., 2003;LI et al., 2004) support the putative candidate gene character of these ESTs.The identified SNPs represent potential informative genetic markers to test our hypothesis that allelic variation in ESTs differentially expressed between the two distinct metabolic cattle types may be associated with variation in nutrient turnover.The SNPs identified in this study provide a pool of genetic polymorphisms, which can be exploited in detailed investigations of a F 2 resource population between Charolais and Holstein (KUEHN et al., 2002) as well as in further commercial populations for their potential use as direct genetic markers.Holstein and Charolais cattle differ in process of growth and nutrition accretion.In comparison to Charolais the maximum of daily gain is earlier in Holstein cattle and the daily body gain drops stronger after reached maximum.From energetic feed evaluation systems (INRA, 1988;AfB, 1995), it can be concluded that the energetic requirements for the same body weight gain is significantly higher in German Holstein in comparison with Charolais.The major differences between Charolais and German Holstein regarding many general production parameters originate from differences in nutrient pathways and storage.Genetic differences in partitioning and utilisation of nutrients and nutrition-gene interactions seem to be main factors for differences in metabolic type of ruminants (BAUMAN and CURRIE, 1980;CRONJÉ, 2000).Growth, lactation, and metabolism are controlled by multiple hormones and factors acting in an endocrine and an autocrine manner (CANT et al., 1999;BREIER, et al., 2000).Both, endocrine and autocrine mechanisms, control the partitioning of absorbed nutrients between various body tissues and organs.According to sequence identity with known genes physiological function of four ESTs showing breed-specific distribution of alleles can be allocated: NACA, NDUFB8, FUS/TLS, and UAP1.NACA is acting as a transcriptional coactivator complexing with newly synthesized polypeptide chains and showing chaperone effects (WIEDMANN et al., 1994;HAMMERLE et al., 2003).NDUFB8 belongs to a family of eukaryotic NADH-ubiquinone oxidoreductase ASHI subunits proteins.Its main function is the transport of electrons from NADH to ubiquinone, which is accompanied by translocation of proteins from mitochondrial matrix to the intermembrane space (LOEFFEN et al., 1998).UAP1 catalyses the synthesis of N-acyl-alpha-Dglucosamine 1-phosphate into diphosphate UDP-N-acetyl-D-glucosamine in the presence of UTP (MIO et al., 1998).The FUS/TLS gene product is a member of the serine-arginine (SR) family of proteins, which is involved in constitutive and regulated RNA splicing.It interacts with the oncoprotein TLS and abrogates the influence of TLS on E1A pre-mRNA splicing (MEISSNER et al., 2003).Genetic variations in genes involved in energy metabolism are assumed considering the presumable physiological function of the ESTs exhibiting single nucleotide polymorphisms and their breed-specific distribution..In a further study based on the polymorphic nature of SNP markers, it will be proved whether variation in the genomic regions influences nutrient turn over in cattle.

Fig. 1 :
Fig. 1: DNA sequencing profiles of the expressed sequence tag fbn-l057 harbouring a G/A transition generated by Taq Cycle sequencing of A) DNA pools involving 20 unrelated Holstein and Charolais cattle and B) individuals with the alternative genotypes.Arrows indicate the variant nucleotide position.In the sequence profile, the green "A" peak under the black "G" peak in the Holstein pool (Figure 1A) represents a allelic frequency of 22 per cent proved by sequencing of the individual animals.Figure 1B shows the sequence profiles of the variant EST region of individuals of the two alternative homozygous (GG, AA) and the heterozygous genotypes (GA) (DNA-Sequenzierprofile der exprimierten Sequenz fbn-l057, die eine GA-Transition aufweist, nach Taq-Cycle-Sequenzierung von A) DNA-Pools aus 20 unverwandten Holstein-und Charolais-Rindern und B) Individuen mit den alternativen Genotypen.Pfeile markieren die variable Nukleotidposition.In den Sequenzprofilen repräsentiert der grüne "A"-Peak unter dem schwarzen "G"-Peak im Holstein-Pool (Abb.1A), verifiziert durch Sequenzierung der Einzeltiere, eine Allelfrequenz von 22 Prozent.Abb.1B zeigt die Sequenzprofile der variablen EST-Region von Individuen mit den alternativen homozygoten (GG, AA) und dem heterozygoten Genotyp (GA))