Polymorphisms within the APOBR gene are highly associated with milk levels of prognostic ketosis biomarkers in dairy cows
Jens Tetens, Claas Heuer, Iris Heyer, Matthias S. Klein, Wolfram Gronwald, Wolfgang Junge, Peter J. Oefner, Georg Thaller and Nina Krattenmacher Physiol. Genomics 47:129-137, 2015. First published 10 February 2015; doi:10.1152/physiolgenomics.00126.2014 You might find this additional info useful... This article cites 32 articles, 8 of which can be accessed free at: /content/47/4/129.full.html#ref-list-1 Updated information and services including high resolution figures, can be found at: /content/47/4/129.full.html Additional material and information about Physiological Genomics can be found at: http://www.the-aps.org/publications/pg
This information is current as of May 24, 2015. Downloaded from on May 24, 2015
Physiological Genomics publishes results of a wide variety of studies from human and from informative model systems with techniques linking genes and pathways to physiology, from prokaryotes to eukaryotes. It is published quarterly in January, April, July, and October by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright © 2015 by the American Physiological Society. ISSN: 1094-8341, ESSN: 1531-2267. Visit our website at http://www.the-aps.org/.
Physiol Genomics 47: 129 –137, 2015. First published February 10, 2015; doi:10.1152/physiolgenomics.00126.2014.
Polymorphisms within the APOBR gene are highly associated with milk levels of prognostic ketosis biomarkers in dairy cows Jens Tetens,1 Claas Heuer,1 Iris Heyer,1 Matthias S. Klein,2 Wolfram Gronwald,2 Wolfgang Junge,1 Peter J. Oefner,2 Georg Thaller,1 and Nina Krattenmacher1 1
Institute of Animal Breeding and Husbandry, Christian Albrecht University, Kiel, Germany; and 2Institute of Functional Genomics, University of Regensburg, Regensburg, Germany Submitted 2 December 2014; accepted in final form 4 February 2015
ketosis; metabolomics; GWAS; glycerophosphocholine; phosphocholine; APOBR OVER THE YEARS, MILK PRODUCTION per cow has increased considerably. But this increase has been accompanied by a greater prevalence of health problems (26). Predominant diseases in highly selected dairy cows (e.g., mastitis, metritis, and metabolic disorders) are most prevalent at the beginning of lactation and have been related to the extent and duration of the postpartum energy deficit (9, 17, 22). Following parturition, the energy requirements for milk production increase tremendously and, initially, are not met by daily feed consumption (5). Essentially all dairy cows experience a period of negative energy balance, which is compensated for by increased body fat mobilization (4, 22). Excessive mobilization due to a severe energy deficit impairs the cow’s immune function (18) and fertility (34) and leads to metabolic stress (9). As
Address for reprint requests and other correspondence: J. Tetens, Inst. of Animal Breeding and Husbandry, Christian Albrecht Univ., Kiel, Germany (e-mail: [email protected]).
a consequence, cows experiencing a prolonged period of negative energy balance are more prone to metabolic problems (9, 14). One of the most common metabolic disorders in dairy cattle is ketosis or acetonemia, which is characterized by an increased production of the ketone bodies beta-hydroxybutyrate, acetone, and acetoacetate (1). These arise from incomplete hepatic oxidation of nonesterified fatty acids (NEFAs) mobilized from depot fat, as the acetyl-CoA resulting from -oxidation of NEFAs cannot completely enter the tricarboxylic acid cycle, because oxaloacetate is redirected to gluconeogenesis (2). Synthesis from noncarbohydrate precursors is the main pathway to ensure glucose supply for lactose synthesis in the mammary gland (28), because enteral glucose availability is low due to ruminal fermentation of carbohydrates (3). Milk levels of ketone bodies are established biomarkers for subclinical ketosis (12, 13), which was estimated to occur at a prevalence of 21.8% across European countries (30). Recently, NMR metabolomic analysis revealed that the milk glycerophosphocholine (GPC)-to-phosphocholine (PC) ratio in early lactation and the GPC level in midlactation were prognostic biomarkers for the risk of ketosis in dairy cattle (20). In that study, healthy animals had significantly higher levels of milk GPC and lower levels of milk PC than animals suffering from ketosis. The authors hypothesized that the raised GPC/PC ratios might be due to higher rates of blood phosphatidylcholine (PtC) breakdown enabling the respective animals to utilize more blood PtC as a fatty acid source for milk lipid synthesis, in turn reducing lipomobilization (20). Thus, selecting animals with high GPC/PC or GPC values might help in breeding animals that cope better with the energy deficit in early lactation and, therefore, are less prone to ketosis (20). A genetic determination can be assumed, because significant correlations between breeding values for fat-protein ratio in milk and PC, GPC, and GPC/PC levels were found (20). In addition, Buitenhuis et al. (2013) (8) stated that GPC in milk has a heritable component. However, comprehensive recording of these milk metabolites is currently too expensive and not available in practice, constricting the implementation in conventional breeding systems. Recent developments of genomeassisted selection schemes might, nonetheless, provide opportunities to select for metabolic stability based on genetic markers. So far, no genetic mechanisms that might be involved in the control of metabolic stability have been identified. In this study, genome-wide association studies (GWAS) for GPC, PC, and the GPC/PC ratio were performed on the phenotypic data from Klein et al. (2012) (20) with the aim of identifying
1094-8341/15 Copyright © 2015 the American Physiological Society
129
Downloaded from on May 24, 2015
Tetens J, Heuer C, Heyer I, Klein MS, Gronwald W, Junge W, Oefner PJ, Thaller G, Krattenmacher N. Polymorphisms within the APOBR gene are highly associated with milk levels of prognostic ketosis biomarkers in dairy cows. Physiol Genomics 47: 129 –137, 2015. First published February 10, 2015; doi:10.1152/physiolgenomics.00126.2014.—Essentially all highyielding dairy cows experience a negative energy balance during early lactation leading to increased lipomobilization, which is a normal physiological response. However, a severe energy deficit may lead to high levels of ketone bodies and, subsequently, to subclinical or clinical ketosis. It has previously been reported that the ratio of glycerophosphocholine to phosphocholine in milk is a prognostic biomarker for the risk of ketosis in dairy cattle. It was hypothesized that this ratio reflects the ability to break down blood phosphatidylcholine as a fatty acid resource. In the current study, 248 animals from a previous study were genotyped with Illumina BovineSNP50 BeadChip, and genome-wide association studies were carried out for the milk levels of phosphocholine, glycerophosphocholine, and the ratio of both metabolites. It was demonstrated that the latter two traits are heritable with h2 ⫽ 0.43 and h2 ⫽ 0.34, respectively. A major quantitative trait locus was identified on cattle chromosome 25. The APOBR gene, coding for the apolipoprotein B receptor, is located within this region and was analyzed as a candidate gene. The analysis revealed highly significant associations of polymorphisms within the gene with glycerophosphocholine as well as the metabolite ratio. These findings support the hypothesis that differences in the ability to take up blood phosphatidylcholine from low-density lipoproteins play an important role in early lactation metabolic stability of dairy cows and indicate APOBR to contain a causative variant.
130
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
Table 1. Descriptive statistics of milk metabolite concentrations of cows in first and second lactation 1st Lactation (nCows ⫽ 238)
2nd Lactation (nCows ⫽ 25)
Trait
Unit
nSpecimens
Mean
SD
Min.
Max.
nSpecimens
Mean
SD
Min.
Max.
GPC PC GPC/PC
mmol/l mmol/l ratio
1,136 908 887
0.74 0.49 1.61
0.23 0.18 0.86
0.15 0.14 0.36
1.52 1.15 7.32
102 90 84
0.75 0.42 1.90
0.22 0.15 0.86
0.29 0.16 0.68
1.50 0.82 5.49
GPC, glycerophosphocholine; PC, phosphocholine.
quantitative trait loci (QTL) affecting energy metabolism in early lactation. A genome wide-significant QTL on cattle chromosome 25 (BTA25) identified the APOBR gene, which encodes the apolipoprotein B receptor, as a plausible candidate. Further analyses indicated that genetic variation within APOBR was likely to underlie the QTL. MATERIAL AND METHODS
Table 2. PCR and sequencing primers used in comparative resequencing of the bovine APOBR gene Primer
APOBR_Ex1 APOBR_Ex2A APOBR_Ex2B APOBR_Ex2C APOBR_Ex2D APOBR_Ex2E APOBR_Ex3 APOBR_Ex4
Location
exon exon exon exon exon exon exon exon
1 2 2 2 2 2/3 3 3/4
Forward (5=¡3=)
Reverse (5=¡3=)
Product Size, bp
Annealing Temp., °C
GGCCGAGGTGTAAATACAAGACC CTGCAGATCCTGTCCTCCTCTTT GAGGCTGGTCAAGATAGGAGGAG GGAGAGTGGAAGAAGAAGCCAAA AGACCACAGCAGAGCAACTTGAA AAGGGAGAGAGGAAGCTCACAGA AGTGCTGGACAGGATGCTCAC CTCTGTCTCGAGGAGCCGAGTAT
AGGAAAGGGCCTAGATGAGAGGT CTTCCCCTCAGGCTTCTTGTG ACCAGCCTTGACCTCTGAGTCTT CTTCCTCCTGGGTTTCTCTGGT CCCCTTCTCTTCTCTCCCTCTG GAATACTCGGCTCCTCGAGACA TCCAAAGTGACAGAGGGAGTCAG CATCTGGTTCTGTCTGCTCCAC
584 702 732 897 751 738 649 575
62 62 62 62 62 62 62 62
Locations are given with respect to transcript variant 1 (ENSBTAT00000037341). Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
Samples, metabolite analysis, and genotyping. Milk samples were collected repeatedly from Holstein-Friesian cows at the Karkendamm dairy research farm in Northern Germany. During the first 5 wk of lactation, specimens were obtained weekly. Thereafter, specimens were collected in midlactation (month 6) and at the end of lactation. Some cows left the herd early, and, thus, time series were not complete for all animals. Sample preparation and NMR measurements were performed at the Institute of Functional Genomics, University of Regensburg, Germany, according to established protocols (15, 19). A detailed description of the experimental settings is given by Klein et al. (2012) (20). The original dataset contained information on cows in lactation one to seven, but cows with lactation numbers ⬎2 were not genotyped. Therefore, the phenotypic dataset was reduced to 248 genotyped cows with 1,266 test-day records collected between days 3 and 351 in milk. Most milk samples (⬎ 90%) were collected from first-lactation cows. All samples were obtained between October 2008 and June 2010. Table 1 provides a more detailed overview of the phenotypic data used in this study. Cows were genotyped using the Illumina BovineSNP50 BeadChip (Illumina, San Diego, CA) comprising 54,001 single nucleotide polymorphisms (SNPs). Based on GC-score filtering of the genotyping results, with a threshold of 0.15 applied as recommended by the manufacturer, all animals had ⬍10% missing genotypes. SNPs without a chromosomal assignment with respect to genome build UMD 3.1 were excluded from the investigation. To increase the power of the GWAS, we used minimac (16) to impute sporadically missing SNP genotypes (⬃1% of genotypes) after prephasing the data with MaCH (23). Thereafter, SNPs with a minor allele frequency ⬍5% were removed, resulting in a final set of 41,409 SNPs. Statistical analyses. We carried out the GWAS by fitting the following linear mixed model to the data for one SNP at a time: y ⫽ Xb ⫹ Wv ⫹ Zaua ⫹ Zpup ⫹ e, where y is a vector of all phenotypic observations across individuals for GPC, PC, or GPC/
PC, respectively. W is the design matrix for the fixed allele substitution effects (v) allocating the SNP genotypes (coded as 0 for the heterozygous and ⫺1 or 1 for the alternative homozygote genotypes, respectively) to the individuals. The marker effect was omitted from the model for the estimation of the variance components of each trait. X is the design matrix for the fixed effects. The respective vector b includes the intercept and systematic effects on the milk metabolites, e.g., the fixed effects of lactation number and test day (119 levels for GPC, 97 levels for PC and GPC/PC, respectively) and, to account for differences between lactation stages, a linear regression on lactation day. Design matrices Za for the additive genetic effects (ua) and Zp for the permanent environmental effects (up) relate the elements of ua and up to their corresponding element in y. The additive genetic effects ua are assumed to follow a normal distribution with mean zero and variance A2a , where A represents a numerator relationship matrix obtained from pedigree information consisting of 4,577 animals in total including 184 unique sires of phenotyped cows. The permanent environmental effect up ⬃ N (0, I2p) was included because the dataset contained repeated observations within cows. I is an identity matrix with an order equal to the number of cows with records (248 for GPC, 247 for PC and GPC/PC, respectively). The residuals were assumed to be independently and normally distributed. Significance of the allele substitution effect was tested by a Wald 2-test with H0: v ⫽ 0 and HA: v ⫽ 0. Genome- and chromosomewide significance thresholds at the 5% level were derived by a Bonferroni correction. Linear mixed models for GWAS and pedigree processing were done using the pedigreemm (32) and lme4 package (4a) within the R statistical environment (27). Gibbs sampling was used to obtain estimates and standard deviations of the variance components and ratios from their marginal posterior distributions using 20,000 iterations and a burn-in period of 10,000. Additionally, bivariate models were fitted to estimate genetic correlations between the traits. Candidate gene analyses. Within 1 Mbp to each side of genomewide significantly associated markers, the current version of the cattle genome (UMD3.1, annotation release 103) was inspected for plausible candidate genes related to energy and especially phospholipid metabolism. The APOBR gene coding for the apolipoprotein B receptor and located on BTA25 within the most significantly associated QTL region for GPC and GPC/PC was selected for further analysis. We determined gene structure by aligning the
131
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
Table 3. Estimates and SD of variance components and genetic parameters for the milk levels of GPC, PC, and GPC/PC, respectively Model Without Any Marker Effect a2 ⫾ SD
Trait
2 pe ⫾ SD
e2 ⫾ SD
Variance Explained by QTL3
Model Including Top SNP r ⫾ SD2
h2 ⫾ SD1
a2 ⫾ SD
2 pe ⫾ SD
e2 ⫾ SD
2 QTL /a2
2 QTL /p2
0.33
0.17
0.24
0.12
GPC 0.018 ⫾ 0.004 0.004 ⫾ 0.003 0.020 ⫾ 0.001 0.43 ⫾ 0.08 0.52 ⫾ 0.03 0.012 ⫾ 0.003 0.003 ⫾ 0.002 0.020 ⫾ 0.001 PC 0.001 ⫾ 0.001 0.006 ⫾ 0.001 0.015 ⫾ 0.001 0.07 ⫾ 0.04 0.33 ⫾ 0.04 GPC/PC 0.215 ⫾ 0.071 0.121 ⫾ 0.001 0.301 ⫾ 0.017 0.34⫾.010 0.53 ⫾ 0.04 0.164 ⫾ 0.048 0.095 ⫾ 0.046 0.300 ⫾ 0.017 2 a
2 pe
2 e
2
Given are the additive genetic ( ), permanent environmental ( ) and residual ( ) variance components as well as heritabilities (h ) and repeatabilities (r) for each trait. For GPC and GPC/PC, variance components are also given as obtained from a model including the most significant marker within APOBR. The amount of additive genetic and phenotypic (2p) variance explained by the QTL was calculated from the reduction of a2 caused by the consideration of the effect 2 of the top marker in the model. Variance components were rounded to 3 and ratios to 2 decimal places, respectively. 1Heritability: a2/(a2 ⫹ pe ⫹ e2). 2 2 2 2 Repeatability: (a2 ⫹ pe )/(a2 ⫹ pe ⫹ e2). 3QTL was calculated as the difference between the additive genetic variance components and phenotypic variance, respectively, from the 2 models.
8
Phosphocholine (PC) 8
− log10(Pobserved)
− log10(P)
6
4
2
0
6
4
λ = 1.16
2
0 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
0
X
2
10
6
8
Glycerophosphocholine (GPC)
8
8
6
− log10(Pobserved)
− log10(P)
4
− log10(Pexpected)
Cattle Chromosome
4
2
0
6
4
λ = 1.21
2
0 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
X
0
2
4
6
8
− log10(Pexpected)
Cattle Chromosome
10
GPC to PC ratio 8
6
− log10(Pobserved)
− log10(P)
8
4
2
0
6
4
λ = 1.19
2
0 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
X
Cattle Chromosome
0
2
4
6
8
− log10(Pexpected)
Fig. 1. Results of the genome-wide association study for glycerophosphocholine (GPC), phosphocholine (PC), and GPC/PC. Left: Manhattan plots of the negative decadic logarithm of the raw P values. The black line indicates the genome-wide significance threshold at a Bonferroni-corrected 5% level; the gray lines indicate the chromosome-wide significance threshold. Right: Q-Q plots comparing the observed and expected distribution of the P values along with the genomic inflation factor indicating a sufficient correction for population structure. Genome- and chromosome-wide significantly associated single nucleotide polymorphisms (SNPs) are highlighted in the Q-Q plot in red and blue, respectively. Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
a final extension step at 72°C for 10 min. After purification of the PCR products with thermosensitive alkaline phosphatase (FastAP; Fermentas, St. Leon-Rot, Germany) and exonuclease I (Fermentas), sequencing was carried out with one of the correspondent primers using an ABI 3130xl Genetic Analyzer and BigDye Terminator v3.1 Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, CA). Sequence analysis was done using the software Sequencher 4.9 (Gene Codes, Ann Arbor, MI); genotypes were directly called from the chromatograms. The polymorphisms within the APOBR gene were tested for trait association with the GWAS model described above. Furthermore, the most significantly associated SNP for GPC and GPC/PC, respectively, were included in the model used to estimate the variance components.
available transcripts of two different splice variants (Ensembl transcript IDs ENSBTAT00000037341 and ENSBTAT00000065319; GenBank Acc. no. XM_002698084, see Fig. 2) to the current version of the cattle genome sequence (UMD3.1, GenBank Acc. no. AC_000182.1). All exons and adjacent intronic regions were then resequenced in 237 animals from the dataset used for the GWAS as described above. For sequence analysis, PCR primers were designed with primer3 (21, 31) (Table 2). The amplification was performed in a 12 l reaction volume containing 20 ng of genomic DNA, 0.2 M of each primer, 200 M dNTPs, and 5 U of Taq polymerase (Invitek, Berlin, Germany) in the reaction buffer supplied by the manufacturer. Cycling conditions were as follows: initial denaturation at 94°C for 5 min, 35 cycles each of 94°C for 30 s, 62°C for 60 s, 72°C for 90 s, and
132
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
The amount of variance explained by the QTL was determined from the reduction of additive genetic variance and heritability, respectively (Table 3). RESULTS AND DISCUSSION
Genetic parameters. The estimated variance components and ratios for all traits analyzed are summarized in Table 3. The highest heritability of 0.43 was estimated for GPC, which is in line with the results of Buitenhuis et al. (2013) (8), who estimated a heritability of 0.48 for milk levels of GPC in dairy cattle. The repeatability of GPC/PC was almost identical to that of GPC, but its heritability was lower (0.34). The heritability of the PC level was close to zero (h2 ⫽ 0.07). In summary, both the milk level of GPC and the GPC/PC ratio are highly
heritable traits, while there seems to be no significant additive genetic effect for the level of PC. These findings indicate that genetic variation in GPC/PC is mainly explained by variation in GPC levels. However, a genetic correlation of ⫺0.59 ⫾ 0.1 (phenotypic correlation ⫽ ⫺0.12) between GPC and PC was estimated, which might point to a joint genetic control of both traits. This should though be interpreted with care, because such analysis requires the simultaneous estimation of several variance components in a small dataset. GWAS. The hypothesis that genetic variation of the GPC/PC ratio is due to GPC variation is supported by the results of the GWAS, which revealed three genome-wide significant (Padjusted ⱕ 0.05) association signals on BTA7, 23, and 25 for GPC, while for PC only a single genome-wide significantly
Table 4. Results of the genome-wide association study PC Positionb
1 3 3 5 7 7 7 9 9 9 9 9 9 10 10 10 13 13 14 14 15 15 15 19 19 22 23 23 23 23 23 25 25 25 25 25 25 25 25 25 25 25 25 25 29
41,732,937 23,737,990 37,897,553 33,374,776 89,026,917 90,239,425 90,753,869 39,187,022 39,779,363 39,816,061 41,209,760 45,121,053 48,722,389 51,827,797 51,850,642 52,011,775 79,730,805 80,848,032 57,728,143 59,398,170 3,669,413 31,947,879 65,255,763 31,638,453 42,289,012 19,588,936 5,524,345 5,598,043 7,398,527 7,874,236 14,394,214 24,228,117 25,547,267 25,740,992 26,310,196 26,468,456 26,982,725 27,501,732 27,545,745 29,693,401 30,039,582 30,062,438 30,927,675 31,491,172 40,695,267
SNP
IDc
ARS-BFGL-NGS-118356 BTA-66968-no-rs BTA-88540-no-rs BTB-01226625 ARS-BFGL-NGS-116915 Hapmap58847-rs29023123 Hapmap51054-BTA-80144 ARS-BFGL-NGS-112742 ARS-BFGL-NGS-75693 Hapmap45687-BTA-83338 ARS-BFGL-NGS-111730 ARS-BFGL-NGS-61170 ARS-BFGL-NGS-27097 ARS-BFGL-NGS-82941 ARS-BFGL-NGS-24765 Hapmap43018-BTA-69257 Hapmap47255-BTA-34035 ARS-BFGL-NGS-7071 Hapmap50325-BTA-35116 ARS-BFGL-NGS-112566 ARS-BFGL-BAC-18253 ARS-BFGL-NGS-111015 BTB-00610177 Hapmap49181-BTA-45080 ARS-BFGL-NGS-111199 BTA-10297-rs29015855 ARS-BFGL-NGS-13215 Hapmap41168-BTA-117434 ARS-BFGL-NGS-61010 ARS-BFGL-NGS-115177 ARS-BFGL-NGS-59308 Hapmap43502-BTA-18199 ARS-BFGL-NGS-21527 Hapmap40645-BTA-110440 ARS-BFGL-NGS-22050 ARS-BFGL-BAC-42500 ARS-BFGL-NGS-100347 ARS-BFGL-NGS-37360 ARS-BFGL-NGS-3731 Hapmap43775-BTA-115377 ARS-BFGL-NGS-56926 ARS-BFGL-BAC-37178 ARS-BFGL-BAC-47171 Hapmap26309-BTC-057772 ARS-BFGL-NGS-86495
Allelesd
MAFe
A/G A/G G/C G/A A/G A/G A/G A/G A/G G/C A/G A/G G/A G/A A/G A/G G/A G/A G/A A/G A/G A/G C/A A/G C/A G/A C/A A/G A/C G/A C/A G/A A/G G/A G/A A/G A/G G/A A/G A/G C/A A/G A/G G/A G/A
0.42 0.17 0.15 0.22 0.37 0.24 0.26 0.21 0.40 0.40 0.33 0.20 0.35 0.19 0.19 0.08 0.28 0.33 0.35 0.27 0.07 0.32 0.06 0.36 0.37 0.23 0.47 0.45 0.12 0.45 0.06 0.30 0.42 0.44 0.28 0.06 0.50 0.49 0.36 0.47 0.11 0.10 0.36 0.41 0.26
P
Valuef
GPC ⫾
SEg
1.85E-05
⫺0.04 ⫾ 0.01
6.85E-07
ⴚ0.07 ⫾ 0.01
P Value
 ⫾ SE
1.89E-05
⫺0.09 ⫾ 0.02
6.88E-06 1.86E-05 3.25E-07
9.44E-06 2.26E-06 1.05E-05
2.05E-05
⫺0.05 ⫾ 0.01
1.24E-05
0.04 ⫾ 0.01
3.39E-05 4.08E-05
⫺0.04 ⫾ 0.01 0.04 ⫾ 0.01
GPC/PC
3.06E-05
⫺0.07 ⫾ 0.02 ⫺0.07 ⫾ 0.02 ⴚ0.08 ⫾ 0.02
⫺0.08 ⫾ 0.02 ⫺0.09 ⫾ 0.02 ⫺0.12 ⫾ 0.03 ⫺0.06 ⫾ 0.02
3.72E-05 7.78E-07
⫺0.06 ⫾ 0.01 0.07 ⫾ 0.01
9.17E-07 1.84E-05 3.72E-08 3.26E-06 3.24E-06 8.64E-08 4.83E-05 4.54E-09 2.87E-08 1.74E-05 4.90E-05 8.35E-06 3.44E-05
ⴚ0.07 ⫾ 0.01 ⫺0.13 ⫾ 0.03 0.08 ⫾ 0.02 0.07 ⫾ 0.01 ⫺0.07 ⫾ 0.01 0.09 ⫾ 0.02 ⫺0.13 ⫾ 0.03 ⴚ0.08 ⫾ 0.01 ⴚ0.08 ⫾ 0.01 0.07 ⫾ 0.02 0.06 ⫾ 0.01 0.10 ⫾ 0.02 0.10 ⫾ 0.02
5.78E-06
⫺0.07 ⫾ 0.01
P Value
 ⫾ SE
6.60E-07 1.21E-05
0.44 ⫾ 0.09 ⫺0.30 ⫾ 0.07
1.96E-07 2.91E-06 2.91E-06 1.01E-05 8.82E-08 2.98E-05
ⴚ0.36 ⫾ 0.07 ⫺0.30 ⫾ 0.06 ⫺0.30 ⫾ 0.06 ⫺0.29 ⫾ 0.07 ⴚ0.40 ⫾ 0.08 ⫺0.26 ⫾ 0.06
2.42E-05 1.03E-05
⫺0.27 ⫾ 0.06 ⫺0.29 ⫾ 0.07
2.32E-05 1.89E-05 2.05E-05
⫺0.48 ⫾ 0.11 ⫺0.27 ⫾ 0.06 ⫺0.50 ⫾ 0.12
3.54E-05
⫺0.28 ⫾ 0.07
3.86E-05
⫺0.39 ⫾ 0.09
2.77E-06 3.74E-05 2.31E-05 2.14E-05
0.30 ⫾ 0.06 0.25 ⫾ 0.06 ⫺0.26 ⫾ 0.06 0.29 ⫾ 0.07
6.14E-09 1.62E-07
ⴚ0.33 ⫾ 0.06 ⴚ0.30 ⫾ 0.06
1.72E-06
⫺0.28⫾.06
2.32E-05
0.30 ⫾ 0.07
Single nucleotide polymorphism (SNP) markers from the 50k SNP chip significantly associated with milk levels of GPC and PC as well as GPC/PC, respectively. aCattle chromosome. bGenomic position according to the bovine genome build UMD3.1. cIllumina SNP identifier. dThe minor allele is given first. e Minor allele frequency. fGiven is the raw P value, boldfaced letters indicate genome-wide significance at the Bonferroni-corrected 5% level; all other listed results represent chromosome-wide significant associations. gGiven is the additive effect of the minor allele together with its SE. Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
BTAa
133
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
Table 5. Polymorphisms identified within the APOBR gene and their association with milk levels of GPC and the ratio of GPC and PC, respectively GPC Location
a
CDS
b
AA
p.Thr98Ala p.Arg122Trp p.His163_Glu164insAsp p.Trp178Arg syn. syn. p.Glu220Gln syn. syn. p.Glu263Gly p.Thr310Ser syn. syn. syn. p.Gly452Glu syn. p.Glu518Lys p.Ser531Pro p.Glu549Arg p.Gly565Asp syn. p.Pro606(594)Leu p.Ala621(609)Thr syn. p.Gly647(635)Arg p.Val681(669)Met syn. p.Asn704(692)Asp p.Arg751(739)Gly syn. syn. p.Gly866(815)Ala
p.Glu967(916)Gly p.Leu970(919)Ile syn. syn.
Minor Allele MAFc A T T C G C C G T ins C C A C T A G T T G A A C A G C A G T A T A A A A G G C G G A A A C G G C T G
0.063 0.057 0.057 0.232 0.233 0.233 0.223 0.011 0.235 0.222 0.236 0.237 0.403 0.228 0.235 0.057 0.057 0.013 0.057 0.236 0.011 0.227 0.239 0.213 0.057 0.226 0.011 0.226 0.011 0.016 0.226 0.011 0.213 0.011 0.225 0.011 0.013 0.226 0.226 0.293 0.057 0.226 0.226 0.224 0.225 0.230 0.230 0.246 0.234
GPC/PC
Genomic Positiond
rs/ss Number
P Value
 ⫾ SE
P Value
 ⫾ SE
26,314,750 26,314,795 26,315,111 26,315,136 26,315,159 26,315,323 26,315,325 26,315,577 26,315,649 26,315,776 26,315,817 26,315,906 26,315,914 26,315,943 26,315,972 26,316,068 26,316,073 26,316,213 26,316,260 26,316,476 26,316,500 26,316,640 26,316,734 26,316,837 26,316,976 26,316,931 26,316,979 26,317,028 26,317,102 26,317,146 26,317,223 26,317,224 26,317,326 26,317,337 26,317,395 26,317,536 26,317,709 26,317,751 26,317,882 26,318,027 26,318,043 26,318,079 26,318,276 26,318,284 26,318,331 26,318,367 26,318,458 26,318,562 26,318,580
rs133727743 rs208487368 rs209746216 rs42071229 rs42071228 rs42071227 rs42071226 rs379736164 rs42071225 rs385299093 rs42071224 rs207724169 rs135372626 rs209397879 rs42071223 rs380807821 rs211666640 ss1424629033 rs208523209 rs42071222 rs470845672 rs42071221 rs42071220 rs211378099 rs110409129 rs210148195 rs378139306 rs42071219 rs477688696 rs447746766 rs42071218 rs379443921 rs210845083 rs382514340 rs42071217 rs379029473 rs42071216 rs42071215 rs42071214 rs42071213 rs209511923 rs42071212 rs42071211 rs42071210 rs42071209 rs42071208 rs42071207 rs42071206 rs42071205
1.65E-05 9.36E-05 9.35E-05 8.40E-16 3.23E-15 1.62E-13 1.16E-16 7.10E-03 1.05E-14 1.03E-16 1.14E-14 7.27E-15 3.83E-05 2.53E-17 1.81E-14 9.35E-05 9.35E-05 3.75E-01 9.35E-05 1.14E-14 7.10E-03 2.29E-18 1.53E-14 2.41E-18 9.35E-05 5.46E-18 7.10E-03 5.46E-18 7.10E-03 5.00E-01 5.46E-18 7.10E-03 2.42E-18 7.10E-03 2.85E-17 7.10E-03 4.73E-01 5.46E-18 5.46E-18 1.72E-06 9.35E-05 5.46E-18 3.75E-18 2.64E-18 2.16E-18 1.68E-17 1.52E-16 4.00E-16 8.84E-17
0.13 ⫾ 0.03 0.12 ⫾ 0.03 0.12 ⫾ 0.03 ⴚ0.13 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 ⴚ0.12 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 0.21 ⫾ 0.08 ⴚ0.12 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 ⴚ0.12 ⫾ 0.02 ⴚ0.12 ⫾ 0.02 ⫺0.06 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 ⴚ0.12 ⫾ 0.02 0.12 ⫾ 0.03 0.12 ⫾ 0.03 0.06 ⫾ 0.07 0.12 ⫾ 0.03 ⴚ0.12 ⫾ 0.02 0.21 ⫾ 0.08 ⴚ0.13 ⫾ 0.02 ⴚ0.12 ⫾ 0.02 ⴚ0.14 ⫾ 0.02 0.12 ⫾ 0.03 ⴚ0.13 ⫾ 0.02 0.21 ⫾ 0.08 ⴚ0.13 ⫾ 0.02 0.21 ⫾ 0.08 ⫺0.04 ⫾ 0.06 ⴚ0.13 ⫾ 0.02 0.21 ⫾ 0.08 ⴚ0.14 ⫾ 0.02 0.21 ⫾ 0.08 ⴚ0.13 ⫾ 0.02 0.21 ⫾ 0.08 ⫺0.04 ⫾ 0.06 ⴚ0.13 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 ⫺0.08 ⫾ 0.02 0.12 ⫾ 0.03 ⴚ0.13 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 ⴚ0.14 ⫾ 0.02 ⴚ0.14 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 ⴚ0.12 ⫾ 0.02 ⴚ0.12 ⫾ 0.01 ⴚ0.13 ⫾ 0.02
1.07E-05 4.31E-04 4.31E-04 1.57E-09 3.95E-09 4.39E-09 5.12E-12 4.31E-04 6.65E-09 8.36E-12 3.96E-09 3.83E-09 8.12E-04 4.50E-12 4.70E-09 4.31E-04 4.31E-04 1.94E-01 4.31E-04 3.96E-09 4.31E-04 3.15E-12 1.48E-09 1.04E-12 4.31E-04 5.21E-12 4.31E-04 5.21E-12 4.31E-04 9.41E-01 5.21E-12 4.31E-04 1.04E-12 4.31E-04 1.23E-11 4.31E-04 8.51E-01 5.21E-12 5.21E-12 3.64E-04 4.31E-04 5.21E-12 2.97E-12 1.16E-12 3.34E-12 4.98E-12 1.06E-11 1.16E-12 1.68E-12
0.54 ⫾ 0.12 0.44 ⫾ 0.13 0.44 ⫾ 0.13 ⴚ0.41 ⫾ 0.07 ⴚ0.41 ⫾ 0.07 ⴚ0.4 ⫾ 0.07 ⴚ0.46 ⫾ 0.07 1.07 ⫾ 0.31 ⴚ0.39 ⫾ 0.07 ⴚ0.46 ⫾ 0.07 ⴚ0.4 ⫾ 0.07 ⴚ0.4 ⫾ 0.07 ⫺0.21 ⫾ 0.06 ⴚ0.46 ⫾ 0.07 ⴚ0.39 ⫾ 0.07 0.44 ⫾ 0.13 0.44 ⫾ 0.13 0.37 ⫾ 0.28 0.44 ⫾ 0.13 ⴚ0.4 ⫾ 0.07 1.08 ⫾ 0.31 ⴚ0.46 ⫾ 0.07 ⴚ0.4 ⫾ 0.07 ⴚ0.49 ⫾ 0.07 0.44 ⫾ 0.13 ⴚ0.46 ⫾ 0.07 1.08 ⫾ 0.31 ⴚ0.46 ⫾ 0.07 1.08 ⫾ 0.31 ⫺0.02 ⫾ 0.24 ⴚ0.46 ⫾ 0.07 1.08 ⫾ 0.31 ⴚ0.49 ⫾ 0.07 1.08 ⫾ 0.31 ⴚ0.45 ⫾ 0.07 1.08 ⫾ 0.31 ⫺0.05 ⫾ 0.24 ⴚ0.46 ⫾ 0.07 ⴚ0.46 ⫾ 0.07 ⫺0.24 ⫾ 0.07 0.44 ⫾ 0.13 ⴚ0.46 ⫾ 0.07 ⴚ0.46 ⫾ 0.07 ⴚ0.47 ⫾ 0.07 ⴚ0.46 ⫾ 0.07 ⴚ0.45 ⫾ 0.07 ⴚ0.44 ⫾ 0.06 ⴚ0.45 ⫾ 0.06 ⴚ0.46 ⫾ 0.07
e
f
a Given is the location with respect to transcript ENSBTAT00000037341; exon 2 of this transcript is represented by 3 adjacent exons in the alternative transcript ENSBTAT00000065319; see Fig. 2 for details. bCDS ⫽ coding sequence, AA ⫽ amino acid sequence; nucleotide positions are with respect to transcript ENSBTAT00000037341, positions within the alternative transcript ENSBTAT00000065319 are given in parentheses; as both transcript have the same reading frame, amino acid changes are denoted equivalently. cMinor allele frequency. dGenomic position according to the bovine genome build UMD3.1. eGiven is the raw P value, boldfaced letters indicate genome-wide significance at the Bonferroni-corrected 5% level. fGiven is the additive effect of the minor allele together with its SE.
associated SNP was detected on BTA3 (Fig. 1, Table 4). The association signal on BTA25 was the most significant for GPC and equivalently the most prominent for the ratio trait (Fig. 1, Table 4). The association signal for PC on BTA3 was likewise present for the GPC/PC ratio, but only represented by a single SNP. Furthermore, there was a notable genome-wide significant association for GPC/PC on BTA9 that was neither present for GPC nor for PC (Fig. 1, Table 4). A total of 10 SNPs reached genome-wide significance; four out of these SNPs were located on BTA25, two of which were
concomitantly associated with GPC levels and the GPC/PC ratio (Table 4); these markers are in linkage disequilibrium with D= ⫽ 0.65 and r2 ⫽ 0.38. The lowest nominal P values for GPC (P ⫽ 4.54E-09) and GPC/PC (P ⫽ 6.14E-09) were obtained for the SNP marker ARS-BFGL-NGS-100347 located at 26,982,725 bp on BTA25. The association signals for GPC and GPC/PC on BTA25 were essentially the same (Figs. 1 and 2), while there was no signal for PC. The direction of the allele substitution effects estimated for the most significantly associated markers on BTA25 was in line with the hypothesis, that
Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
Exon 1 (5=-UTR) c.1-196G⬎A c.1-151C⬎T Intron 1 c.57 ⫹ 109C⬎T c.57 ⫹ 134T⬎C c.57 ⫹ 157A⬎G c.58-20T⬎C c.58-18T⬎C Exon 2 c.292A⬎G c.364A⬎T c.487_488insATG c.532T⬎C c.621T⬎C c.629A⬎G c.658G⬎C c.687C⬎T c.783G⬎A c.788A⬎G c.928A⬎T c.975C⬎T c.1191A⬎G c.1215G⬎A c.1355G⬎A c.1449G⬎C c.1552G⬎A c.1591T⬎C c.1646A⬎G c.1694G⬎A c.1743A⬎G c.1817(1781)C⬎T c.1861(1825)G⬎A c.1938(1902)C⬎T c.1939(1903)G⬎A c.2041(2005)G⬎A c.2052(2016)G⬎A c.2110(2074)A⬎G c.2251(2215)C⬎G c.2424(2388)G⬎A c.2466(2430)C⬎T c.2597(2444)G⬎C Intron 2 c.2730(2577)⫹12A⬎G c.2730(2577)⫹28G⬎A c.2731(2578)-28G⬎A Exon 3 c.2900(2747)A⬎G c.2908(2755)C⬎A c.2955(2802)G⬎C c.2991(2838)G⬎A Intron 3 c.2994(2841)-6T⬎C Exon 4 (3=-UTR) c.3063(2910)*29G⬎T c.3063(2910)*99G⬎A
b
134
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
receptor (6, 7), but it interacts with apolipoprotein B, a main constituent of LDL. Furthermore, the uptake and conversion of LDL-associated PtC have been shown at least for hepatocytes (24). The inspection of the genomic regions adjacent to associated SNPs on other chromosomes revealed no equally obvious candidates. However, on BTA23, ⬃800 kb downstream of the SNP Hapmap41168-BTA-117434 (Table 4), the bovine HMGCLL1 gene is located, which encodes the 3-hydroxymethyl-3-methylglutaryl-CoA (HMG-CoA) lyase-like protein 1. HMG-CoA lyase catalyzes the cleavage of HMGCoA into acetoacetate and acetyl-CoA (29), an essential step in ketogenesis. HMGCLL1 codes for an extramitochondrial form of this enzyme (25). This is an obvious link to ketosis, although a possible connection with PtC or GPC metabolism remains to be elucidated. Furthermore, the ITPR3 gene, coding for the inositol 1,4,5-triphosphate receptor type 3, is located ⬃100 kb upstream of the genome-wide significantly associated SNP ARS-BFGL-NGS-115177 on BTA23 (Table 4). This gene is also not directly linked to phospholipid metabolism but is strongly linked to ketosis, as it is known to regulate hepatic gluconeogenesis in fasting and diabetes (31). Analysis of the APOBR gene. A total of 49 polymorphisms were identified within the APOBR gene (Table 5), which are almost exclusively SNPs. The only exception was a 3 bp insertion within the open reading frame leading to the insertion of an aspartic acid residue after position 163 of the protein (Table 5). With respect to transcript ENSBTAT00000037341
A
20
GPC−SNP50 GPC/PC−SNP50 GPC−APOBR GPC/PC−APOBR
− log10(P)
15
10
5
0 15
20
25
30
35
Position BTA25 [Mb] 20
B
GPC GPC/PC
− log10(P)
15 10 5 0 APOBR−1 (ENSBTAT00000037341)
5’UTR
I
APOBR−2 (ENSBTAT00000065319)
5’UTR
I
26314000
26315000
II II
26316000
III
26317000
IV
26318000
III
IV
3’UTR
V
VI
3’UTR
26319000
Position BTA25 [bp]
Fig. 2. Detailed view of the association results for GPC and GPC/PC within the QTL region on BTA25 and within the APOBR gene. Depicted are the negative decadic logarithms of the P values for both traits within the QTL region illustrating the matching association signals. Markers from the 50k SNP chip are depicted as dots, intragenic polymorphisms as triangles (A). Intragenic association results are shown with respect to 2 different splice variants (transcripts ENSBTAT00000037341 and ENSBTAT00000065319, see Table 4). Yellow asterisks indicate nonsynonymous mutations. The black lines in both A and B indicate the genome-wide significance threshold at a Bonferroni-corrected 5% level; the gray line indicates the chromosome-wide significance threshold. Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
variation in GPC levels determines variation in GPC/PC (Table 4). A QTL for milk levels of GPC has previously been reported on BTA25 (8), but that study did not analyze PC. Identification of candidate genes. Klein et al. (2012) (20) hypothesized that the individual variability of the milk GPC/PC ratio might reflect different abilities to break down blood PtC as a fatty acid source for milk fat synthesis, thus reducing the amount of body fat mobilization. This hypothesis might be extended in so far as the increased PtC utilization reduces the need for de novo fatty acid synthesis in the mammary gland and, thus, the recruitment of NADPH from the pentose phosphate cycle. This in turn might lead to an increased C3 return from the mammary pentose phosphate cycle in favor of hepatic gluconeogenesis. It has been reported that this mechanism might contribute to the allocation of glucogenic carbon in dairy cattle (2). Based on the above hypothesis, we screened the close vicinity of genome-wide significantly associated SNPs for possible candidate genes involved especially in phospholipid metabolism, but also in energy metabolism in general. There are no genes annotated within these intervals coding for phospholipases directly involved in PtC breakdown. An alternative mechanism involved in blood PtC breakdown as a fatty acid source in the mammary gland might be related to cellular low-density lipoprotein (LDL) uptake. Approximately 600 kb upstream of the most significant association signal on BTA25, the gene encoding for the apolipoprotein B-receptor (APOBR) is located. This receptor has been described as a macrophage
135
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
more, it cannot be excluded that the actual causative variant is a nearby regulatory mutation not captured by the current experiment. In their previous work, Klein et al. (20) demonstrated the predictive ability of the analyzed metabolites with respect to ketosis at a phenotypic level. They determined thresholds of 2.5 for the GPC/PC ratio in lactation weeks 1– 4 and 1.2 mmol/l for GPC in lactation month six, respectively. Cows exceeding these limits were found to be less prone to ketosis. Compared with the mean phenotypic values of different QTL genotypes (Fig. 3), the thresholds are rather high. Although predominantly animals homozygous for the favorable QTL allele fall above the limits, the QTL genotype cannot be used to classify the animals with respect to ketosis liability. This result has to be expected, because the QTL only explains a part of the additive genetic variance, which in turn only explains half of the phenotypic variance of metabolite concentrations. The metabolites are, however, highly informative, reflecting both the genetic determination and the actual environment. This has also been found in a study on the prediction of individual ketosis risk by means of artificial neural networks using the very same dataset (11). The determination of the causative variant for the current and further identified QTL will enhance our knowledge about the mechanisms underlying metabolic stability in early lactation and help to determine combinations of metabolites, genetic markers, and possibly other phenotypes that are more predictive with respect to individual disease risk.
GPC by genotype across lactation stages
A GPC [mmol/L]
1.4 1.2 1.0 0.8 0.6 0.4 0.2 Week 1
Week 2
GPC/PC [Molar ratio]
B
Week 3
Week 4
Week 5
Month 6
GPC/PC by genotype across lactation stages Genotype c.2955G>C
6
G/G G/C C/C
4 2 0 Week 1
Week 2
Week 3
Week4
Week5
Fig. 3. Milk GPC levels (A) and ratios of GPC to PC (B) by genotype across different lactation stages. The phenotypic values are expressed with respect to the genotype of the polymorphism within the APOBR gene most significantly associated with GPC. For lactation month 6, only GPC values were available. The dashed lines indicate the thresholds determined by Klein et al. (19). Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
(Table 5, Fig. 2), 36 of the 49 polymorphisms were located within the open reading frame, and 22 of these represent nonsynonymous mutations. A total of 29 polymorphisms were found to be genome-wide significantly associated at the 5%Bonferroni level with both GPC and the GPC/PC ratio. Given the small dataset, the obtained P values were exceedingly small. For GPC, a minimum of 2.16E-18 was found for a synonymous G⬎C transversion in exon 3 at position 2955 of the coding sequence, while the best results for GPC/PC (P ⫽ 1.04E-12) were obtained for two nonsynonymous A⬎G transitions causing the amino acid substitutions Glu518Lys and Val681Met in exon 2, respectively (Table 5, Fig. 2). The top SNPs explain 33 and 23% of the additive genetic variance for GPC and the GPC/PC ratio, corresponding to 17 and 12% of explained phenotypic variance, respectively. Overall, the genome-wide significantly associated polymorphisms were scattered across the entire gene. They all showed a similar minor allele frequency of ⬃0.23 and almost identical absolute effect sizes of ⬃0.13 for GPC and ⬃0.4 to 0.5 for GPC/PC, respectively (Table 5) and probably represent the haplotype containing the causative mutation. Consequently, the pairwise linkage disequilibrium (LD) between the significantly associated markers averaged to an r2 value of 0.96. Thus, it is not possible to determine an actual causative variant by a purely genetic approach. Also, tools to predict possibly deleterious variants are of limited informativeness for prioritization, as the variant underlying the analyzed QTL is segregating at high frequencies and might only be of functional impact in high-yielding animals, thus possibly being neutral under natural selection conditions. Further-
136
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
Conclusions It has previously been reported that the ratio of milk GPC and PC is a prognostic biomarker for the risk of ketosis in dairy cattle. It was hypothesized that this might be due to a different ability of blood PtC breakdown as a source for mammary fatty acid synthesis. Here, using the same set of animals, we found that the milk level of GPC and the GPC/PC ratio are highly heritable traits. A major QTL on BTA25 significantly affects milk GPC and GPC/PC, but not the level of PC. Within the QTL region, the APOBR gene coding for the apolipoprotein B receptor is located. The analysis of this gene revealed highly significant associations with GPC and the GPC/PC ratio, suggesting APOBR harbors the causative variant underlying this QTL. These findings support the hypothesis that differences in the ability to take up blood PtC from LDL play a role in the metabolic stability of dairy cows during early lactation. However, the identification of the causative mutation was not possible by purely genetic approaches; further studies will be necessary. ACKNOWLEDGMENTS
GRANTS This study was funded by the German Federal Ministry of Education and Research in context of the Fugato-plus MeGA-M framework and by the Bavarian Genome Network (BayGene), Munich, Germany. The genotyping of the cows was financially supported by the German Federal Ministry of Education and Research (project FUGATO-plus GENOTRACK, grant no. 0315134A) and the KMSH (Kompetenzzentrum Milch - Schleswig-Holstein, Kiel, Germany). DISCLOSURES No conflicts of interest, financial or otherwise, are declared by the author(s). AUTHOR CONTRIBUTIONS Author contributions: J.T., G.T., and N.K. conception and design of research; J.T., I.H., M.S.K., W.G., and P.J.O. performed experiments; J.T. and C.H. analyzed data; J.T. and N.K. interpreted results of experiments; J.T. prepared figures; J.T. drafted manuscript; J.T., M.S.K., W.G., W.J., P.J.O., G.T., and N.K. edited and revised manuscript; J.T., C.H., I.H., M.S.K., W.G., W.J., P.J.O., G.T., and N.K. approved final version of manuscript. REFERENCES 1. Andersson L. Concentrations of blood and milk ketone bodies, blood isopropanol and plasma glucose in dairy cows in relation to the degree of hyperketonaemia and clinical signs. Zentralblatt Veterinärmedizin A 31: 683–693, 1984. 2. Aschenbach JR, Kristensen NB, Donkin SS, Hammon HM, Penner GB. Gluconeogenesis in dairy cows. The secret of making sweet milk from sour dough. IUBMB Life 62: 869 –877, 2010. 3. Baird GD, Lomax MA, Symonds HW, Shaw SR. Net hepatic and splanchnic metabolism of lactate, pyruvate and propionate in dairy cows in vivo in relation to lactation and nutrient supply. Biochem J 186: 47–57, 1980. 4. Banos G, Coffey MP, Brotherstone S. Modeling daily energy balance of dairy cows in the first three lactations. J Dairy Sci 88: 2226 –2237, 2005. 4a.Bates D, Maechler M, Bolker B, Walker S. lme4: Linear mixed-effects models using Eigen and S4. http://CRAN.R-project.org/package⫽lme4, 2014.
Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
The authors thank the staff at the Karkendamm research farm for providing milk and plasma specimens and NORD-OST GENETIC GmbH & Co. KG (Verden/Aller, Germany) for kind permission to include their animals in this study. Gabi Ottzen is acknowledged for expert technical assistance. Current address for M. S. Klein: Faculty of Kinesiology, Univ. of Calgary, Calgary, Alberta, Canada.
5. Bell AW. Regulation of organic nutrient metabolism during transition from late pregnancy to early lactation. J Anim Sci 73: 2804 –2819, 1995. 6. Bermudez B, Lopez S, Varela LM, Ortega A, Pacheco YM, Moreda W, Moreno-Luna R, Abia R, Muriana Francisco JG. Triglyceride-rich lipoprotein regulates APOB48 receptor gene expression in human THP-1 monocytes and macrophages. J Nutr 142: 227–232, 2012. 7. Brown ML, Ramprasad MP, Umeda PK, Tanaka A, Kobayashi Y, Watanabe T, Shimoyamada H, Kuo W, Li R, Song R, Bradley WA, Gianturco SH. A macrophage receptor for apolipoprotein B48. Cloning, expression, and atherosclerosis. Proc Natl Acad Sci USA 97: 7488 –7493, 2000. 8. Buitenhuis AJ, Sundekilde UK, Poulsen NA, Bertram HC, Larsen LB, Sorensen P. Estimation of genetic parameters and detection of quantitative trait loci for metabolites in Danish Holstein milk. J Dairy Sci 96: 3285–3295, 2013. 9. Collard BL, Boettcher PJ, Dekkers JCM, Petitclerc D, Schaeffer LR. Relationships between energy balance and health traits of dairy cattle in early lactation. J Dairy Sci 83: 2683–2690, 2000. 11. Ehret A, Hochstuhl D, Krattenmacher N, Tetens J, Klein MS, Gronwald W, Thaller G. Short communication: Use of genomic and metabolic information as well as milk performance records for prediction of subclinical ketosis risk via artificial neural networks. J Dairy Sci 98: 322–329, 2015. 12. Enjalbert F, Nicot MC, Bayourthe C, Moncoulon R. Ketone bodies in milk and blood of dairy cows. Relationship between concentrations and utilization for detection of subclinical ketosis. J Dairy Sci 84: 583–589, 2001. 13. Geishauser T, Leslie K, Tenhag J, Bashiri A. Evaluation of eight cow-side ketone tests in milk for detection of subclinical ketosis in dairy cows. J Dairy Sci 83: 296 –299, 2000. 14. Goff JP, Horst RL. Physiological changes at parturition and their relationship to metabolic disorders. J Dairy Sci 80: 1260 –1268, 1997. 15. Gronwald W, Klein MS, Kaspar H, Fagerer SR, Nürnberger N, Dettmer K, Bertsch T, Oefner PJ. Urinary metabolite quantification employing 2D NMR spectroscopy. Analyt Chem 80: 9288 –9297, 2008. 16. Howie B, Fuchsberger C, Stephens M, Marchini J, Abecasis GR. Fast and accurate genotype imputation in genome-wide association studies through pre-phasing. Nat Genet 44: 955–956, 2012. 17. Ingvartsen KL, Andersen JB. Integration of metabolism and intake regulation. A review focusing on periparturient animals. J Dairy Sci 83: 1573–1597, 2000. 18. Kehrli, ME Jr, Nonnecke BJ, Roth JA. Alterations in bovine neutrophil function during the periparturient period. Am J Vet Res 50: 207–214, 1989. 19. Klein MS, Almstetter MF, Schlamberger G, Nürnberger N, Dettmer K, Oefner PJ, Meyer HHD, Wiedemann S, Gronwald W. Nuclear magnetic resonance and mass spectrometry-based milk metabolomics in dairy cows during early and late lactation. J Dairy Sci 93: 1539 –1550, 2010. 20. Klein MS, Buttchereit N, Miemczyk SP, Immervoll AK, Louis C, Wiedemann S, Junge W, Thaller G, Oefner PJ, Gronwald W. NMR metabolomic analysis of dairy cows reveals milk glycerophosphocholine to phosphocholine ratio as prognostic biomarker for risk of ketosis. J Proteome Res 11: 1373–1381, 2012. 21. Koressaar T, Remm M. Enhancements and modifications of primer design program Primer3. Bioinformatics (Oxford) 23: 1289 –1291, 2007. 22. Leblanc S. Monitoring metabolic health of dairy cattle in the transition period. J Reprod Dev 56: S29 –S35, 2010. 23. Li Y, Willer CJ, Ding J, Scheet P, Abecasis GR. MaCH: using sequence and genotype data to estimate haplotypes and unobserved genotypes. Genet Epidemiol 34, 816 –834, 2010. 24. Minahk C, Kim KW, Nelson R, Trigatti B, Lehner R, Vance DE. Conversion of low density lipoprotein-associated phosphatidylcholine to triacylglycerol by primary hepatocytes. J Biol Chem 283: 6449 –6458, 2008. 25. Montgomery C, Pei Z, Watkins PA, Miziorko HM. Identification and characterization of an extramitochondrial human 3-hydroxy-3-methylglutaryl-CoA lyase. J Biol Chem 287: 33227–33236, 2012. 26. Pryce JE, Veerkamp RF, Thompson R, Hill WG, Simm G. Genetic aspects of common health disorders and measures of fertility in Holstein Friesian dairy cattle. Anim Sci 65: 353–360, 1997. 27. R Core Team. R: a Language and Environment for Statistical Computing. Vienna, Austria: http://www.R-project.org/, 2014. 28. Reynolds CK, Huntington GB, Tyrrell HF, Reynolds PJ. Net portaldrained visceral and hepatic metabolism of glucose, L-lactate, and nitrog-
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS enous compounds in lactating holstein cows. J Dairy Sci 71: 1803–1812, 1988. 29. Stegink LD, Coon MJ. Stereospecificity and other properties of highly purified beta-hydroxy-beta-methylglutaryl coenzyme A cleavage enzyme from bovine liver. J Biol Chem 243: 5272–5279, 1968. 30. Suthar VS, Canelas-Raposo J, Deniz A, Heuwieser W. Prevalence of subclinical ketosis and relationships with postpartum diseases in European dairy cows. J Dairy Sci 96: 2925–2938, 2013. 31. Untergasser A, Cutcutache I, Koressaar T, Ye J, Faircloth BC, Remm M, Rozen SG. Primer3–new capabilities and interfaces. Nucleic Acids Res 40: e115, 2012.
137
32. Vazquez AI, Bates DM, Rosa GJM, Gianola D, Weigel KA. Technical note: an R package for fitting generalized linear mixed models in animal breeding. J Anim Sci 88: 497–504, 2010. 33. Wang Y, Li G, Goode J, Paz JC, Ouyang K, Screaton R, Fischer WH, Chen J, Tabas I, Montminy M. Inositol-1,4,5-trisphosphate receptor regulates hepatic gluconeogenesis in fasting and diabetes. Nature 485: 128 –132, 2012. 34. Wathes DC, Fenwick M, Cheng Z, Bourne N, Llewellyn S, Morris DG, Kenny D, Murphy J, Fitzpatrick R. Influence of negative energy balance on cyclicity and fertility in the high producing dairy cow. Theriogenology 68: S232–S241, 2007.
Downloaded from on May 24, 2015
Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Jens Tetens, Claas Heuer, Iris Heyer, Matthias S. Klein, Wolfram Gronwald, Wolfgang Junge, Peter J. Oefner, Georg Thaller and Nina Krattenmacher Physiol. Genomics 47:129-137, 2015. First published 10 February 2015; doi:10.1152/physiolgenomics.00126.2014 You might find this additional info useful... This article cites 32 articles, 8 of which can be accessed free at: /content/47/4/129.full.html#ref-list-1 Updated information and services including high resolution figures, can be found at: /content/47/4/129.full.html Additional material and information about Physiological Genomics can be found at: http://www.the-aps.org/publications/pg
This information is current as of May 24, 2015. Downloaded from on May 24, 2015
Physiological Genomics publishes results of a wide variety of studies from human and from informative model systems with techniques linking genes and pathways to physiology, from prokaryotes to eukaryotes. It is published quarterly in January, April, July, and October by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright © 2015 by the American Physiological Society. ISSN: 1094-8341, ESSN: 1531-2267. Visit our website at http://www.the-aps.org/.
Physiol Genomics 47: 129 –137, 2015. First published February 10, 2015; doi:10.1152/physiolgenomics.00126.2014.
Polymorphisms within the APOBR gene are highly associated with milk levels of prognostic ketosis biomarkers in dairy cows Jens Tetens,1 Claas Heuer,1 Iris Heyer,1 Matthias S. Klein,2 Wolfram Gronwald,2 Wolfgang Junge,1 Peter J. Oefner,2 Georg Thaller,1 and Nina Krattenmacher1 1
Institute of Animal Breeding and Husbandry, Christian Albrecht University, Kiel, Germany; and 2Institute of Functional Genomics, University of Regensburg, Regensburg, Germany Submitted 2 December 2014; accepted in final form 4 February 2015
ketosis; metabolomics; GWAS; glycerophosphocholine; phosphocholine; APOBR OVER THE YEARS, MILK PRODUCTION per cow has increased considerably. But this increase has been accompanied by a greater prevalence of health problems (26). Predominant diseases in highly selected dairy cows (e.g., mastitis, metritis, and metabolic disorders) are most prevalent at the beginning of lactation and have been related to the extent and duration of the postpartum energy deficit (9, 17, 22). Following parturition, the energy requirements for milk production increase tremendously and, initially, are not met by daily feed consumption (5). Essentially all dairy cows experience a period of negative energy balance, which is compensated for by increased body fat mobilization (4, 22). Excessive mobilization due to a severe energy deficit impairs the cow’s immune function (18) and fertility (34) and leads to metabolic stress (9). As
Address for reprint requests and other correspondence: J. Tetens, Inst. of Animal Breeding and Husbandry, Christian Albrecht Univ., Kiel, Germany (e-mail: [email protected]).
a consequence, cows experiencing a prolonged period of negative energy balance are more prone to metabolic problems (9, 14). One of the most common metabolic disorders in dairy cattle is ketosis or acetonemia, which is characterized by an increased production of the ketone bodies beta-hydroxybutyrate, acetone, and acetoacetate (1). These arise from incomplete hepatic oxidation of nonesterified fatty acids (NEFAs) mobilized from depot fat, as the acetyl-CoA resulting from -oxidation of NEFAs cannot completely enter the tricarboxylic acid cycle, because oxaloacetate is redirected to gluconeogenesis (2). Synthesis from noncarbohydrate precursors is the main pathway to ensure glucose supply for lactose synthesis in the mammary gland (28), because enteral glucose availability is low due to ruminal fermentation of carbohydrates (3). Milk levels of ketone bodies are established biomarkers for subclinical ketosis (12, 13), which was estimated to occur at a prevalence of 21.8% across European countries (30). Recently, NMR metabolomic analysis revealed that the milk glycerophosphocholine (GPC)-to-phosphocholine (PC) ratio in early lactation and the GPC level in midlactation were prognostic biomarkers for the risk of ketosis in dairy cattle (20). In that study, healthy animals had significantly higher levels of milk GPC and lower levels of milk PC than animals suffering from ketosis. The authors hypothesized that the raised GPC/PC ratios might be due to higher rates of blood phosphatidylcholine (PtC) breakdown enabling the respective animals to utilize more blood PtC as a fatty acid source for milk lipid synthesis, in turn reducing lipomobilization (20). Thus, selecting animals with high GPC/PC or GPC values might help in breeding animals that cope better with the energy deficit in early lactation and, therefore, are less prone to ketosis (20). A genetic determination can be assumed, because significant correlations between breeding values for fat-protein ratio in milk and PC, GPC, and GPC/PC levels were found (20). In addition, Buitenhuis et al. (2013) (8) stated that GPC in milk has a heritable component. However, comprehensive recording of these milk metabolites is currently too expensive and not available in practice, constricting the implementation in conventional breeding systems. Recent developments of genomeassisted selection schemes might, nonetheless, provide opportunities to select for metabolic stability based on genetic markers. So far, no genetic mechanisms that might be involved in the control of metabolic stability have been identified. In this study, genome-wide association studies (GWAS) for GPC, PC, and the GPC/PC ratio were performed on the phenotypic data from Klein et al. (2012) (20) with the aim of identifying
1094-8341/15 Copyright © 2015 the American Physiological Society
129
Downloaded from on May 24, 2015
Tetens J, Heuer C, Heyer I, Klein MS, Gronwald W, Junge W, Oefner PJ, Thaller G, Krattenmacher N. Polymorphisms within the APOBR gene are highly associated with milk levels of prognostic ketosis biomarkers in dairy cows. Physiol Genomics 47: 129 –137, 2015. First published February 10, 2015; doi:10.1152/physiolgenomics.00126.2014.—Essentially all highyielding dairy cows experience a negative energy balance during early lactation leading to increased lipomobilization, which is a normal physiological response. However, a severe energy deficit may lead to high levels of ketone bodies and, subsequently, to subclinical or clinical ketosis. It has previously been reported that the ratio of glycerophosphocholine to phosphocholine in milk is a prognostic biomarker for the risk of ketosis in dairy cattle. It was hypothesized that this ratio reflects the ability to break down blood phosphatidylcholine as a fatty acid resource. In the current study, 248 animals from a previous study were genotyped with Illumina BovineSNP50 BeadChip, and genome-wide association studies were carried out for the milk levels of phosphocholine, glycerophosphocholine, and the ratio of both metabolites. It was demonstrated that the latter two traits are heritable with h2 ⫽ 0.43 and h2 ⫽ 0.34, respectively. A major quantitative trait locus was identified on cattle chromosome 25. The APOBR gene, coding for the apolipoprotein B receptor, is located within this region and was analyzed as a candidate gene. The analysis revealed highly significant associations of polymorphisms within the gene with glycerophosphocholine as well as the metabolite ratio. These findings support the hypothesis that differences in the ability to take up blood phosphatidylcholine from low-density lipoproteins play an important role in early lactation metabolic stability of dairy cows and indicate APOBR to contain a causative variant.
130
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
Table 1. Descriptive statistics of milk metabolite concentrations of cows in first and second lactation 1st Lactation (nCows ⫽ 238)
2nd Lactation (nCows ⫽ 25)
Trait
Unit
nSpecimens
Mean
SD
Min.
Max.
nSpecimens
Mean
SD
Min.
Max.
GPC PC GPC/PC
mmol/l mmol/l ratio
1,136 908 887
0.74 0.49 1.61
0.23 0.18 0.86
0.15 0.14 0.36
1.52 1.15 7.32
102 90 84
0.75 0.42 1.90
0.22 0.15 0.86
0.29 0.16 0.68
1.50 0.82 5.49
GPC, glycerophosphocholine; PC, phosphocholine.
quantitative trait loci (QTL) affecting energy metabolism in early lactation. A genome wide-significant QTL on cattle chromosome 25 (BTA25) identified the APOBR gene, which encodes the apolipoprotein B receptor, as a plausible candidate. Further analyses indicated that genetic variation within APOBR was likely to underlie the QTL. MATERIAL AND METHODS
Table 2. PCR and sequencing primers used in comparative resequencing of the bovine APOBR gene Primer
APOBR_Ex1 APOBR_Ex2A APOBR_Ex2B APOBR_Ex2C APOBR_Ex2D APOBR_Ex2E APOBR_Ex3 APOBR_Ex4
Location
exon exon exon exon exon exon exon exon
1 2 2 2 2 2/3 3 3/4
Forward (5=¡3=)
Reverse (5=¡3=)
Product Size, bp
Annealing Temp., °C
GGCCGAGGTGTAAATACAAGACC CTGCAGATCCTGTCCTCCTCTTT GAGGCTGGTCAAGATAGGAGGAG GGAGAGTGGAAGAAGAAGCCAAA AGACCACAGCAGAGCAACTTGAA AAGGGAGAGAGGAAGCTCACAGA AGTGCTGGACAGGATGCTCAC CTCTGTCTCGAGGAGCCGAGTAT
AGGAAAGGGCCTAGATGAGAGGT CTTCCCCTCAGGCTTCTTGTG ACCAGCCTTGACCTCTGAGTCTT CTTCCTCCTGGGTTTCTCTGGT CCCCTTCTCTTCTCTCCCTCTG GAATACTCGGCTCCTCGAGACA TCCAAAGTGACAGAGGGAGTCAG CATCTGGTTCTGTCTGCTCCAC
584 702 732 897 751 738 649 575
62 62 62 62 62 62 62 62
Locations are given with respect to transcript variant 1 (ENSBTAT00000037341). Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
Samples, metabolite analysis, and genotyping. Milk samples were collected repeatedly from Holstein-Friesian cows at the Karkendamm dairy research farm in Northern Germany. During the first 5 wk of lactation, specimens were obtained weekly. Thereafter, specimens were collected in midlactation (month 6) and at the end of lactation. Some cows left the herd early, and, thus, time series were not complete for all animals. Sample preparation and NMR measurements were performed at the Institute of Functional Genomics, University of Regensburg, Germany, according to established protocols (15, 19). A detailed description of the experimental settings is given by Klein et al. (2012) (20). The original dataset contained information on cows in lactation one to seven, but cows with lactation numbers ⬎2 were not genotyped. Therefore, the phenotypic dataset was reduced to 248 genotyped cows with 1,266 test-day records collected between days 3 and 351 in milk. Most milk samples (⬎ 90%) were collected from first-lactation cows. All samples were obtained between October 2008 and June 2010. Table 1 provides a more detailed overview of the phenotypic data used in this study. Cows were genotyped using the Illumina BovineSNP50 BeadChip (Illumina, San Diego, CA) comprising 54,001 single nucleotide polymorphisms (SNPs). Based on GC-score filtering of the genotyping results, with a threshold of 0.15 applied as recommended by the manufacturer, all animals had ⬍10% missing genotypes. SNPs without a chromosomal assignment with respect to genome build UMD 3.1 were excluded from the investigation. To increase the power of the GWAS, we used minimac (16) to impute sporadically missing SNP genotypes (⬃1% of genotypes) after prephasing the data with MaCH (23). Thereafter, SNPs with a minor allele frequency ⬍5% were removed, resulting in a final set of 41,409 SNPs. Statistical analyses. We carried out the GWAS by fitting the following linear mixed model to the data for one SNP at a time: y ⫽ Xb ⫹ Wv ⫹ Zaua ⫹ Zpup ⫹ e, where y is a vector of all phenotypic observations across individuals for GPC, PC, or GPC/
PC, respectively. W is the design matrix for the fixed allele substitution effects (v) allocating the SNP genotypes (coded as 0 for the heterozygous and ⫺1 or 1 for the alternative homozygote genotypes, respectively) to the individuals. The marker effect was omitted from the model for the estimation of the variance components of each trait. X is the design matrix for the fixed effects. The respective vector b includes the intercept and systematic effects on the milk metabolites, e.g., the fixed effects of lactation number and test day (119 levels for GPC, 97 levels for PC and GPC/PC, respectively) and, to account for differences between lactation stages, a linear regression on lactation day. Design matrices Za for the additive genetic effects (ua) and Zp for the permanent environmental effects (up) relate the elements of ua and up to their corresponding element in y. The additive genetic effects ua are assumed to follow a normal distribution with mean zero and variance A2a , where A represents a numerator relationship matrix obtained from pedigree information consisting of 4,577 animals in total including 184 unique sires of phenotyped cows. The permanent environmental effect up ⬃ N (0, I2p) was included because the dataset contained repeated observations within cows. I is an identity matrix with an order equal to the number of cows with records (248 for GPC, 247 for PC and GPC/PC, respectively). The residuals were assumed to be independently and normally distributed. Significance of the allele substitution effect was tested by a Wald 2-test with H0: v ⫽ 0 and HA: v ⫽ 0. Genome- and chromosomewide significance thresholds at the 5% level were derived by a Bonferroni correction. Linear mixed models for GWAS and pedigree processing were done using the pedigreemm (32) and lme4 package (4a) within the R statistical environment (27). Gibbs sampling was used to obtain estimates and standard deviations of the variance components and ratios from their marginal posterior distributions using 20,000 iterations and a burn-in period of 10,000. Additionally, bivariate models were fitted to estimate genetic correlations between the traits. Candidate gene analyses. Within 1 Mbp to each side of genomewide significantly associated markers, the current version of the cattle genome (UMD3.1, annotation release 103) was inspected for plausible candidate genes related to energy and especially phospholipid metabolism. The APOBR gene coding for the apolipoprotein B receptor and located on BTA25 within the most significantly associated QTL region for GPC and GPC/PC was selected for further analysis. We determined gene structure by aligning the
131
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
Table 3. Estimates and SD of variance components and genetic parameters for the milk levels of GPC, PC, and GPC/PC, respectively Model Without Any Marker Effect a2 ⫾ SD
Trait
2 pe ⫾ SD
e2 ⫾ SD
Variance Explained by QTL3
Model Including Top SNP r ⫾ SD2
h2 ⫾ SD1
a2 ⫾ SD
2 pe ⫾ SD
e2 ⫾ SD
2 QTL /a2
2 QTL /p2
0.33
0.17
0.24
0.12
GPC 0.018 ⫾ 0.004 0.004 ⫾ 0.003 0.020 ⫾ 0.001 0.43 ⫾ 0.08 0.52 ⫾ 0.03 0.012 ⫾ 0.003 0.003 ⫾ 0.002 0.020 ⫾ 0.001 PC 0.001 ⫾ 0.001 0.006 ⫾ 0.001 0.015 ⫾ 0.001 0.07 ⫾ 0.04 0.33 ⫾ 0.04 GPC/PC 0.215 ⫾ 0.071 0.121 ⫾ 0.001 0.301 ⫾ 0.017 0.34⫾.010 0.53 ⫾ 0.04 0.164 ⫾ 0.048 0.095 ⫾ 0.046 0.300 ⫾ 0.017 2 a
2 pe
2 e
2
Given are the additive genetic ( ), permanent environmental ( ) and residual ( ) variance components as well as heritabilities (h ) and repeatabilities (r) for each trait. For GPC and GPC/PC, variance components are also given as obtained from a model including the most significant marker within APOBR. The amount of additive genetic and phenotypic (2p) variance explained by the QTL was calculated from the reduction of a2 caused by the consideration of the effect 2 of the top marker in the model. Variance components were rounded to 3 and ratios to 2 decimal places, respectively. 1Heritability: a2/(a2 ⫹ pe ⫹ e2). 2 2 2 2 Repeatability: (a2 ⫹ pe )/(a2 ⫹ pe ⫹ e2). 3QTL was calculated as the difference between the additive genetic variance components and phenotypic variance, respectively, from the 2 models.
8
Phosphocholine (PC) 8
− log10(Pobserved)
− log10(P)
6
4
2
0
6
4
λ = 1.16
2
0 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
0
X
2
10
6
8
Glycerophosphocholine (GPC)
8
8
6
− log10(Pobserved)
− log10(P)
4
− log10(Pexpected)
Cattle Chromosome
4
2
0
6
4
λ = 1.21
2
0 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
X
0
2
4
6
8
− log10(Pexpected)
Cattle Chromosome
10
GPC to PC ratio 8
6
− log10(Pobserved)
− log10(P)
8
4
2
0
6
4
λ = 1.19
2
0 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
X
Cattle Chromosome
0
2
4
6
8
− log10(Pexpected)
Fig. 1. Results of the genome-wide association study for glycerophosphocholine (GPC), phosphocholine (PC), and GPC/PC. Left: Manhattan plots of the negative decadic logarithm of the raw P values. The black line indicates the genome-wide significance threshold at a Bonferroni-corrected 5% level; the gray lines indicate the chromosome-wide significance threshold. Right: Q-Q plots comparing the observed and expected distribution of the P values along with the genomic inflation factor indicating a sufficient correction for population structure. Genome- and chromosome-wide significantly associated single nucleotide polymorphisms (SNPs) are highlighted in the Q-Q plot in red and blue, respectively. Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
a final extension step at 72°C for 10 min. After purification of the PCR products with thermosensitive alkaline phosphatase (FastAP; Fermentas, St. Leon-Rot, Germany) and exonuclease I (Fermentas), sequencing was carried out with one of the correspondent primers using an ABI 3130xl Genetic Analyzer and BigDye Terminator v3.1 Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, CA). Sequence analysis was done using the software Sequencher 4.9 (Gene Codes, Ann Arbor, MI); genotypes were directly called from the chromatograms. The polymorphisms within the APOBR gene were tested for trait association with the GWAS model described above. Furthermore, the most significantly associated SNP for GPC and GPC/PC, respectively, were included in the model used to estimate the variance components.
available transcripts of two different splice variants (Ensembl transcript IDs ENSBTAT00000037341 and ENSBTAT00000065319; GenBank Acc. no. XM_002698084, see Fig. 2) to the current version of the cattle genome sequence (UMD3.1, GenBank Acc. no. AC_000182.1). All exons and adjacent intronic regions were then resequenced in 237 animals from the dataset used for the GWAS as described above. For sequence analysis, PCR primers were designed with primer3 (21, 31) (Table 2). The amplification was performed in a 12 l reaction volume containing 20 ng of genomic DNA, 0.2 M of each primer, 200 M dNTPs, and 5 U of Taq polymerase (Invitek, Berlin, Germany) in the reaction buffer supplied by the manufacturer. Cycling conditions were as follows: initial denaturation at 94°C for 5 min, 35 cycles each of 94°C for 30 s, 62°C for 60 s, 72°C for 90 s, and
132
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
The amount of variance explained by the QTL was determined from the reduction of additive genetic variance and heritability, respectively (Table 3). RESULTS AND DISCUSSION
Genetic parameters. The estimated variance components and ratios for all traits analyzed are summarized in Table 3. The highest heritability of 0.43 was estimated for GPC, which is in line with the results of Buitenhuis et al. (2013) (8), who estimated a heritability of 0.48 for milk levels of GPC in dairy cattle. The repeatability of GPC/PC was almost identical to that of GPC, but its heritability was lower (0.34). The heritability of the PC level was close to zero (h2 ⫽ 0.07). In summary, both the milk level of GPC and the GPC/PC ratio are highly
heritable traits, while there seems to be no significant additive genetic effect for the level of PC. These findings indicate that genetic variation in GPC/PC is mainly explained by variation in GPC levels. However, a genetic correlation of ⫺0.59 ⫾ 0.1 (phenotypic correlation ⫽ ⫺0.12) between GPC and PC was estimated, which might point to a joint genetic control of both traits. This should though be interpreted with care, because such analysis requires the simultaneous estimation of several variance components in a small dataset. GWAS. The hypothesis that genetic variation of the GPC/PC ratio is due to GPC variation is supported by the results of the GWAS, which revealed three genome-wide significant (Padjusted ⱕ 0.05) association signals on BTA7, 23, and 25 for GPC, while for PC only a single genome-wide significantly
Table 4. Results of the genome-wide association study PC Positionb
1 3 3 5 7 7 7 9 9 9 9 9 9 10 10 10 13 13 14 14 15 15 15 19 19 22 23 23 23 23 23 25 25 25 25 25 25 25 25 25 25 25 25 25 29
41,732,937 23,737,990 37,897,553 33,374,776 89,026,917 90,239,425 90,753,869 39,187,022 39,779,363 39,816,061 41,209,760 45,121,053 48,722,389 51,827,797 51,850,642 52,011,775 79,730,805 80,848,032 57,728,143 59,398,170 3,669,413 31,947,879 65,255,763 31,638,453 42,289,012 19,588,936 5,524,345 5,598,043 7,398,527 7,874,236 14,394,214 24,228,117 25,547,267 25,740,992 26,310,196 26,468,456 26,982,725 27,501,732 27,545,745 29,693,401 30,039,582 30,062,438 30,927,675 31,491,172 40,695,267
SNP
IDc
ARS-BFGL-NGS-118356 BTA-66968-no-rs BTA-88540-no-rs BTB-01226625 ARS-BFGL-NGS-116915 Hapmap58847-rs29023123 Hapmap51054-BTA-80144 ARS-BFGL-NGS-112742 ARS-BFGL-NGS-75693 Hapmap45687-BTA-83338 ARS-BFGL-NGS-111730 ARS-BFGL-NGS-61170 ARS-BFGL-NGS-27097 ARS-BFGL-NGS-82941 ARS-BFGL-NGS-24765 Hapmap43018-BTA-69257 Hapmap47255-BTA-34035 ARS-BFGL-NGS-7071 Hapmap50325-BTA-35116 ARS-BFGL-NGS-112566 ARS-BFGL-BAC-18253 ARS-BFGL-NGS-111015 BTB-00610177 Hapmap49181-BTA-45080 ARS-BFGL-NGS-111199 BTA-10297-rs29015855 ARS-BFGL-NGS-13215 Hapmap41168-BTA-117434 ARS-BFGL-NGS-61010 ARS-BFGL-NGS-115177 ARS-BFGL-NGS-59308 Hapmap43502-BTA-18199 ARS-BFGL-NGS-21527 Hapmap40645-BTA-110440 ARS-BFGL-NGS-22050 ARS-BFGL-BAC-42500 ARS-BFGL-NGS-100347 ARS-BFGL-NGS-37360 ARS-BFGL-NGS-3731 Hapmap43775-BTA-115377 ARS-BFGL-NGS-56926 ARS-BFGL-BAC-37178 ARS-BFGL-BAC-47171 Hapmap26309-BTC-057772 ARS-BFGL-NGS-86495
Allelesd
MAFe
A/G A/G G/C G/A A/G A/G A/G A/G A/G G/C A/G A/G G/A G/A A/G A/G G/A G/A G/A A/G A/G A/G C/A A/G C/A G/A C/A A/G A/C G/A C/A G/A A/G G/A G/A A/G A/G G/A A/G A/G C/A A/G A/G G/A G/A
0.42 0.17 0.15 0.22 0.37 0.24 0.26 0.21 0.40 0.40 0.33 0.20 0.35 0.19 0.19 0.08 0.28 0.33 0.35 0.27 0.07 0.32 0.06 0.36 0.37 0.23 0.47 0.45 0.12 0.45 0.06 0.30 0.42 0.44 0.28 0.06 0.50 0.49 0.36 0.47 0.11 0.10 0.36 0.41 0.26
P
Valuef
GPC ⫾
SEg
1.85E-05
⫺0.04 ⫾ 0.01
6.85E-07
ⴚ0.07 ⫾ 0.01
P Value
 ⫾ SE
1.89E-05
⫺0.09 ⫾ 0.02
6.88E-06 1.86E-05 3.25E-07
9.44E-06 2.26E-06 1.05E-05
2.05E-05
⫺0.05 ⫾ 0.01
1.24E-05
0.04 ⫾ 0.01
3.39E-05 4.08E-05
⫺0.04 ⫾ 0.01 0.04 ⫾ 0.01
GPC/PC
3.06E-05
⫺0.07 ⫾ 0.02 ⫺0.07 ⫾ 0.02 ⴚ0.08 ⫾ 0.02
⫺0.08 ⫾ 0.02 ⫺0.09 ⫾ 0.02 ⫺0.12 ⫾ 0.03 ⫺0.06 ⫾ 0.02
3.72E-05 7.78E-07
⫺0.06 ⫾ 0.01 0.07 ⫾ 0.01
9.17E-07 1.84E-05 3.72E-08 3.26E-06 3.24E-06 8.64E-08 4.83E-05 4.54E-09 2.87E-08 1.74E-05 4.90E-05 8.35E-06 3.44E-05
ⴚ0.07 ⫾ 0.01 ⫺0.13 ⫾ 0.03 0.08 ⫾ 0.02 0.07 ⫾ 0.01 ⫺0.07 ⫾ 0.01 0.09 ⫾ 0.02 ⫺0.13 ⫾ 0.03 ⴚ0.08 ⫾ 0.01 ⴚ0.08 ⫾ 0.01 0.07 ⫾ 0.02 0.06 ⫾ 0.01 0.10 ⫾ 0.02 0.10 ⫾ 0.02
5.78E-06
⫺0.07 ⫾ 0.01
P Value
 ⫾ SE
6.60E-07 1.21E-05
0.44 ⫾ 0.09 ⫺0.30 ⫾ 0.07
1.96E-07 2.91E-06 2.91E-06 1.01E-05 8.82E-08 2.98E-05
ⴚ0.36 ⫾ 0.07 ⫺0.30 ⫾ 0.06 ⫺0.30 ⫾ 0.06 ⫺0.29 ⫾ 0.07 ⴚ0.40 ⫾ 0.08 ⫺0.26 ⫾ 0.06
2.42E-05 1.03E-05
⫺0.27 ⫾ 0.06 ⫺0.29 ⫾ 0.07
2.32E-05 1.89E-05 2.05E-05
⫺0.48 ⫾ 0.11 ⫺0.27 ⫾ 0.06 ⫺0.50 ⫾ 0.12
3.54E-05
⫺0.28 ⫾ 0.07
3.86E-05
⫺0.39 ⫾ 0.09
2.77E-06 3.74E-05 2.31E-05 2.14E-05
0.30 ⫾ 0.06 0.25 ⫾ 0.06 ⫺0.26 ⫾ 0.06 0.29 ⫾ 0.07
6.14E-09 1.62E-07
ⴚ0.33 ⫾ 0.06 ⴚ0.30 ⫾ 0.06
1.72E-06
⫺0.28⫾.06
2.32E-05
0.30 ⫾ 0.07
Single nucleotide polymorphism (SNP) markers from the 50k SNP chip significantly associated with milk levels of GPC and PC as well as GPC/PC, respectively. aCattle chromosome. bGenomic position according to the bovine genome build UMD3.1. cIllumina SNP identifier. dThe minor allele is given first. e Minor allele frequency. fGiven is the raw P value, boldfaced letters indicate genome-wide significance at the Bonferroni-corrected 5% level; all other listed results represent chromosome-wide significant associations. gGiven is the additive effect of the minor allele together with its SE. Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
BTAa
133
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
Table 5. Polymorphisms identified within the APOBR gene and their association with milk levels of GPC and the ratio of GPC and PC, respectively GPC Location
a
CDS
b
AA
p.Thr98Ala p.Arg122Trp p.His163_Glu164insAsp p.Trp178Arg syn. syn. p.Glu220Gln syn. syn. p.Glu263Gly p.Thr310Ser syn. syn. syn. p.Gly452Glu syn. p.Glu518Lys p.Ser531Pro p.Glu549Arg p.Gly565Asp syn. p.Pro606(594)Leu p.Ala621(609)Thr syn. p.Gly647(635)Arg p.Val681(669)Met syn. p.Asn704(692)Asp p.Arg751(739)Gly syn. syn. p.Gly866(815)Ala
p.Glu967(916)Gly p.Leu970(919)Ile syn. syn.
Minor Allele MAFc A T T C G C C G T ins C C A C T A G T T G A A C A G C A G T A T A A A A G G C G G A A A C G G C T G
0.063 0.057 0.057 0.232 0.233 0.233 0.223 0.011 0.235 0.222 0.236 0.237 0.403 0.228 0.235 0.057 0.057 0.013 0.057 0.236 0.011 0.227 0.239 0.213 0.057 0.226 0.011 0.226 0.011 0.016 0.226 0.011 0.213 0.011 0.225 0.011 0.013 0.226 0.226 0.293 0.057 0.226 0.226 0.224 0.225 0.230 0.230 0.246 0.234
GPC/PC
Genomic Positiond
rs/ss Number
P Value
 ⫾ SE
P Value
 ⫾ SE
26,314,750 26,314,795 26,315,111 26,315,136 26,315,159 26,315,323 26,315,325 26,315,577 26,315,649 26,315,776 26,315,817 26,315,906 26,315,914 26,315,943 26,315,972 26,316,068 26,316,073 26,316,213 26,316,260 26,316,476 26,316,500 26,316,640 26,316,734 26,316,837 26,316,976 26,316,931 26,316,979 26,317,028 26,317,102 26,317,146 26,317,223 26,317,224 26,317,326 26,317,337 26,317,395 26,317,536 26,317,709 26,317,751 26,317,882 26,318,027 26,318,043 26,318,079 26,318,276 26,318,284 26,318,331 26,318,367 26,318,458 26,318,562 26,318,580
rs133727743 rs208487368 rs209746216 rs42071229 rs42071228 rs42071227 rs42071226 rs379736164 rs42071225 rs385299093 rs42071224 rs207724169 rs135372626 rs209397879 rs42071223 rs380807821 rs211666640 ss1424629033 rs208523209 rs42071222 rs470845672 rs42071221 rs42071220 rs211378099 rs110409129 rs210148195 rs378139306 rs42071219 rs477688696 rs447746766 rs42071218 rs379443921 rs210845083 rs382514340 rs42071217 rs379029473 rs42071216 rs42071215 rs42071214 rs42071213 rs209511923 rs42071212 rs42071211 rs42071210 rs42071209 rs42071208 rs42071207 rs42071206 rs42071205
1.65E-05 9.36E-05 9.35E-05 8.40E-16 3.23E-15 1.62E-13 1.16E-16 7.10E-03 1.05E-14 1.03E-16 1.14E-14 7.27E-15 3.83E-05 2.53E-17 1.81E-14 9.35E-05 9.35E-05 3.75E-01 9.35E-05 1.14E-14 7.10E-03 2.29E-18 1.53E-14 2.41E-18 9.35E-05 5.46E-18 7.10E-03 5.46E-18 7.10E-03 5.00E-01 5.46E-18 7.10E-03 2.42E-18 7.10E-03 2.85E-17 7.10E-03 4.73E-01 5.46E-18 5.46E-18 1.72E-06 9.35E-05 5.46E-18 3.75E-18 2.64E-18 2.16E-18 1.68E-17 1.52E-16 4.00E-16 8.84E-17
0.13 ⫾ 0.03 0.12 ⫾ 0.03 0.12 ⫾ 0.03 ⴚ0.13 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 ⴚ0.12 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 0.21 ⫾ 0.08 ⴚ0.12 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 ⴚ0.12 ⫾ 0.02 ⴚ0.12 ⫾ 0.02 ⫺0.06 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 ⴚ0.12 ⫾ 0.02 0.12 ⫾ 0.03 0.12 ⫾ 0.03 0.06 ⫾ 0.07 0.12 ⫾ 0.03 ⴚ0.12 ⫾ 0.02 0.21 ⫾ 0.08 ⴚ0.13 ⫾ 0.02 ⴚ0.12 ⫾ 0.02 ⴚ0.14 ⫾ 0.02 0.12 ⫾ 0.03 ⴚ0.13 ⫾ 0.02 0.21 ⫾ 0.08 ⴚ0.13 ⫾ 0.02 0.21 ⫾ 0.08 ⫺0.04 ⫾ 0.06 ⴚ0.13 ⫾ 0.02 0.21 ⫾ 0.08 ⴚ0.14 ⫾ 0.02 0.21 ⫾ 0.08 ⴚ0.13 ⫾ 0.02 0.21 ⫾ 0.08 ⫺0.04 ⫾ 0.06 ⴚ0.13 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 ⫺0.08 ⫾ 0.02 0.12 ⫾ 0.03 ⴚ0.13 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 ⴚ0.14 ⫾ 0.02 ⴚ0.14 ⫾ 0.02 ⴚ0.13 ⫾ 0.02 ⴚ0.12 ⫾ 0.02 ⴚ0.12 ⫾ 0.01 ⴚ0.13 ⫾ 0.02
1.07E-05 4.31E-04 4.31E-04 1.57E-09 3.95E-09 4.39E-09 5.12E-12 4.31E-04 6.65E-09 8.36E-12 3.96E-09 3.83E-09 8.12E-04 4.50E-12 4.70E-09 4.31E-04 4.31E-04 1.94E-01 4.31E-04 3.96E-09 4.31E-04 3.15E-12 1.48E-09 1.04E-12 4.31E-04 5.21E-12 4.31E-04 5.21E-12 4.31E-04 9.41E-01 5.21E-12 4.31E-04 1.04E-12 4.31E-04 1.23E-11 4.31E-04 8.51E-01 5.21E-12 5.21E-12 3.64E-04 4.31E-04 5.21E-12 2.97E-12 1.16E-12 3.34E-12 4.98E-12 1.06E-11 1.16E-12 1.68E-12
0.54 ⫾ 0.12 0.44 ⫾ 0.13 0.44 ⫾ 0.13 ⴚ0.41 ⫾ 0.07 ⴚ0.41 ⫾ 0.07 ⴚ0.4 ⫾ 0.07 ⴚ0.46 ⫾ 0.07 1.07 ⫾ 0.31 ⴚ0.39 ⫾ 0.07 ⴚ0.46 ⫾ 0.07 ⴚ0.4 ⫾ 0.07 ⴚ0.4 ⫾ 0.07 ⫺0.21 ⫾ 0.06 ⴚ0.46 ⫾ 0.07 ⴚ0.39 ⫾ 0.07 0.44 ⫾ 0.13 0.44 ⫾ 0.13 0.37 ⫾ 0.28 0.44 ⫾ 0.13 ⴚ0.4 ⫾ 0.07 1.08 ⫾ 0.31 ⴚ0.46 ⫾ 0.07 ⴚ0.4 ⫾ 0.07 ⴚ0.49 ⫾ 0.07 0.44 ⫾ 0.13 ⴚ0.46 ⫾ 0.07 1.08 ⫾ 0.31 ⴚ0.46 ⫾ 0.07 1.08 ⫾ 0.31 ⫺0.02 ⫾ 0.24 ⴚ0.46 ⫾ 0.07 1.08 ⫾ 0.31 ⴚ0.49 ⫾ 0.07 1.08 ⫾ 0.31 ⴚ0.45 ⫾ 0.07 1.08 ⫾ 0.31 ⫺0.05 ⫾ 0.24 ⴚ0.46 ⫾ 0.07 ⴚ0.46 ⫾ 0.07 ⫺0.24 ⫾ 0.07 0.44 ⫾ 0.13 ⴚ0.46 ⫾ 0.07 ⴚ0.46 ⫾ 0.07 ⴚ0.47 ⫾ 0.07 ⴚ0.46 ⫾ 0.07 ⴚ0.45 ⫾ 0.07 ⴚ0.44 ⫾ 0.06 ⴚ0.45 ⫾ 0.06 ⴚ0.46 ⫾ 0.07
e
f
a Given is the location with respect to transcript ENSBTAT00000037341; exon 2 of this transcript is represented by 3 adjacent exons in the alternative transcript ENSBTAT00000065319; see Fig. 2 for details. bCDS ⫽ coding sequence, AA ⫽ amino acid sequence; nucleotide positions are with respect to transcript ENSBTAT00000037341, positions within the alternative transcript ENSBTAT00000065319 are given in parentheses; as both transcript have the same reading frame, amino acid changes are denoted equivalently. cMinor allele frequency. dGenomic position according to the bovine genome build UMD3.1. eGiven is the raw P value, boldfaced letters indicate genome-wide significance at the Bonferroni-corrected 5% level. fGiven is the additive effect of the minor allele together with its SE.
associated SNP was detected on BTA3 (Fig. 1, Table 4). The association signal on BTA25 was the most significant for GPC and equivalently the most prominent for the ratio trait (Fig. 1, Table 4). The association signal for PC on BTA3 was likewise present for the GPC/PC ratio, but only represented by a single SNP. Furthermore, there was a notable genome-wide significant association for GPC/PC on BTA9 that was neither present for GPC nor for PC (Fig. 1, Table 4). A total of 10 SNPs reached genome-wide significance; four out of these SNPs were located on BTA25, two of which were
concomitantly associated with GPC levels and the GPC/PC ratio (Table 4); these markers are in linkage disequilibrium with D= ⫽ 0.65 and r2 ⫽ 0.38. The lowest nominal P values for GPC (P ⫽ 4.54E-09) and GPC/PC (P ⫽ 6.14E-09) were obtained for the SNP marker ARS-BFGL-NGS-100347 located at 26,982,725 bp on BTA25. The association signals for GPC and GPC/PC on BTA25 were essentially the same (Figs. 1 and 2), while there was no signal for PC. The direction of the allele substitution effects estimated for the most significantly associated markers on BTA25 was in line with the hypothesis, that
Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
Exon 1 (5=-UTR) c.1-196G⬎A c.1-151C⬎T Intron 1 c.57 ⫹ 109C⬎T c.57 ⫹ 134T⬎C c.57 ⫹ 157A⬎G c.58-20T⬎C c.58-18T⬎C Exon 2 c.292A⬎G c.364A⬎T c.487_488insATG c.532T⬎C c.621T⬎C c.629A⬎G c.658G⬎C c.687C⬎T c.783G⬎A c.788A⬎G c.928A⬎T c.975C⬎T c.1191A⬎G c.1215G⬎A c.1355G⬎A c.1449G⬎C c.1552G⬎A c.1591T⬎C c.1646A⬎G c.1694G⬎A c.1743A⬎G c.1817(1781)C⬎T c.1861(1825)G⬎A c.1938(1902)C⬎T c.1939(1903)G⬎A c.2041(2005)G⬎A c.2052(2016)G⬎A c.2110(2074)A⬎G c.2251(2215)C⬎G c.2424(2388)G⬎A c.2466(2430)C⬎T c.2597(2444)G⬎C Intron 2 c.2730(2577)⫹12A⬎G c.2730(2577)⫹28G⬎A c.2731(2578)-28G⬎A Exon 3 c.2900(2747)A⬎G c.2908(2755)C⬎A c.2955(2802)G⬎C c.2991(2838)G⬎A Intron 3 c.2994(2841)-6T⬎C Exon 4 (3=-UTR) c.3063(2910)*29G⬎T c.3063(2910)*99G⬎A
b
134
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
receptor (6, 7), but it interacts with apolipoprotein B, a main constituent of LDL. Furthermore, the uptake and conversion of LDL-associated PtC have been shown at least for hepatocytes (24). The inspection of the genomic regions adjacent to associated SNPs on other chromosomes revealed no equally obvious candidates. However, on BTA23, ⬃800 kb downstream of the SNP Hapmap41168-BTA-117434 (Table 4), the bovine HMGCLL1 gene is located, which encodes the 3-hydroxymethyl-3-methylglutaryl-CoA (HMG-CoA) lyase-like protein 1. HMG-CoA lyase catalyzes the cleavage of HMGCoA into acetoacetate and acetyl-CoA (29), an essential step in ketogenesis. HMGCLL1 codes for an extramitochondrial form of this enzyme (25). This is an obvious link to ketosis, although a possible connection with PtC or GPC metabolism remains to be elucidated. Furthermore, the ITPR3 gene, coding for the inositol 1,4,5-triphosphate receptor type 3, is located ⬃100 kb upstream of the genome-wide significantly associated SNP ARS-BFGL-NGS-115177 on BTA23 (Table 4). This gene is also not directly linked to phospholipid metabolism but is strongly linked to ketosis, as it is known to regulate hepatic gluconeogenesis in fasting and diabetes (31). Analysis of the APOBR gene. A total of 49 polymorphisms were identified within the APOBR gene (Table 5), which are almost exclusively SNPs. The only exception was a 3 bp insertion within the open reading frame leading to the insertion of an aspartic acid residue after position 163 of the protein (Table 5). With respect to transcript ENSBTAT00000037341
A
20
GPC−SNP50 GPC/PC−SNP50 GPC−APOBR GPC/PC−APOBR
− log10(P)
15
10
5
0 15
20
25
30
35
Position BTA25 [Mb] 20
B
GPC GPC/PC
− log10(P)
15 10 5 0 APOBR−1 (ENSBTAT00000037341)
5’UTR
I
APOBR−2 (ENSBTAT00000065319)
5’UTR
I
26314000
26315000
II II
26316000
III
26317000
IV
26318000
III
IV
3’UTR
V
VI
3’UTR
26319000
Position BTA25 [bp]
Fig. 2. Detailed view of the association results for GPC and GPC/PC within the QTL region on BTA25 and within the APOBR gene. Depicted are the negative decadic logarithms of the P values for both traits within the QTL region illustrating the matching association signals. Markers from the 50k SNP chip are depicted as dots, intragenic polymorphisms as triangles (A). Intragenic association results are shown with respect to 2 different splice variants (transcripts ENSBTAT00000037341 and ENSBTAT00000065319, see Table 4). Yellow asterisks indicate nonsynonymous mutations. The black lines in both A and B indicate the genome-wide significance threshold at a Bonferroni-corrected 5% level; the gray line indicates the chromosome-wide significance threshold. Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
variation in GPC levels determines variation in GPC/PC (Table 4). A QTL for milk levels of GPC has previously been reported on BTA25 (8), but that study did not analyze PC. Identification of candidate genes. Klein et al. (2012) (20) hypothesized that the individual variability of the milk GPC/PC ratio might reflect different abilities to break down blood PtC as a fatty acid source for milk fat synthesis, thus reducing the amount of body fat mobilization. This hypothesis might be extended in so far as the increased PtC utilization reduces the need for de novo fatty acid synthesis in the mammary gland and, thus, the recruitment of NADPH from the pentose phosphate cycle. This in turn might lead to an increased C3 return from the mammary pentose phosphate cycle in favor of hepatic gluconeogenesis. It has been reported that this mechanism might contribute to the allocation of glucogenic carbon in dairy cattle (2). Based on the above hypothesis, we screened the close vicinity of genome-wide significantly associated SNPs for possible candidate genes involved especially in phospholipid metabolism, but also in energy metabolism in general. There are no genes annotated within these intervals coding for phospholipases directly involved in PtC breakdown. An alternative mechanism involved in blood PtC breakdown as a fatty acid source in the mammary gland might be related to cellular low-density lipoprotein (LDL) uptake. Approximately 600 kb upstream of the most significant association signal on BTA25, the gene encoding for the apolipoprotein B-receptor (APOBR) is located. This receptor has been described as a macrophage
135
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
more, it cannot be excluded that the actual causative variant is a nearby regulatory mutation not captured by the current experiment. In their previous work, Klein et al. (20) demonstrated the predictive ability of the analyzed metabolites with respect to ketosis at a phenotypic level. They determined thresholds of 2.5 for the GPC/PC ratio in lactation weeks 1– 4 and 1.2 mmol/l for GPC in lactation month six, respectively. Cows exceeding these limits were found to be less prone to ketosis. Compared with the mean phenotypic values of different QTL genotypes (Fig. 3), the thresholds are rather high. Although predominantly animals homozygous for the favorable QTL allele fall above the limits, the QTL genotype cannot be used to classify the animals with respect to ketosis liability. This result has to be expected, because the QTL only explains a part of the additive genetic variance, which in turn only explains half of the phenotypic variance of metabolite concentrations. The metabolites are, however, highly informative, reflecting both the genetic determination and the actual environment. This has also been found in a study on the prediction of individual ketosis risk by means of artificial neural networks using the very same dataset (11). The determination of the causative variant for the current and further identified QTL will enhance our knowledge about the mechanisms underlying metabolic stability in early lactation and help to determine combinations of metabolites, genetic markers, and possibly other phenotypes that are more predictive with respect to individual disease risk.
GPC by genotype across lactation stages
A GPC [mmol/L]
1.4 1.2 1.0 0.8 0.6 0.4 0.2 Week 1
Week 2
GPC/PC [Molar ratio]
B
Week 3
Week 4
Week 5
Month 6
GPC/PC by genotype across lactation stages Genotype c.2955G>C
6
G/G G/C C/C
4 2 0 Week 1
Week 2
Week 3
Week4
Week5
Fig. 3. Milk GPC levels (A) and ratios of GPC to PC (B) by genotype across different lactation stages. The phenotypic values are expressed with respect to the genotype of the polymorphism within the APOBR gene most significantly associated with GPC. For lactation month 6, only GPC values were available. The dashed lines indicate the thresholds determined by Klein et al. (19). Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
(Table 5, Fig. 2), 36 of the 49 polymorphisms were located within the open reading frame, and 22 of these represent nonsynonymous mutations. A total of 29 polymorphisms were found to be genome-wide significantly associated at the 5%Bonferroni level with both GPC and the GPC/PC ratio. Given the small dataset, the obtained P values were exceedingly small. For GPC, a minimum of 2.16E-18 was found for a synonymous G⬎C transversion in exon 3 at position 2955 of the coding sequence, while the best results for GPC/PC (P ⫽ 1.04E-12) were obtained for two nonsynonymous A⬎G transitions causing the amino acid substitutions Glu518Lys and Val681Met in exon 2, respectively (Table 5, Fig. 2). The top SNPs explain 33 and 23% of the additive genetic variance for GPC and the GPC/PC ratio, corresponding to 17 and 12% of explained phenotypic variance, respectively. Overall, the genome-wide significantly associated polymorphisms were scattered across the entire gene. They all showed a similar minor allele frequency of ⬃0.23 and almost identical absolute effect sizes of ⬃0.13 for GPC and ⬃0.4 to 0.5 for GPC/PC, respectively (Table 5) and probably represent the haplotype containing the causative mutation. Consequently, the pairwise linkage disequilibrium (LD) between the significantly associated markers averaged to an r2 value of 0.96. Thus, it is not possible to determine an actual causative variant by a purely genetic approach. Also, tools to predict possibly deleterious variants are of limited informativeness for prioritization, as the variant underlying the analyzed QTL is segregating at high frequencies and might only be of functional impact in high-yielding animals, thus possibly being neutral under natural selection conditions. Further-
136
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS
Conclusions It has previously been reported that the ratio of milk GPC and PC is a prognostic biomarker for the risk of ketosis in dairy cattle. It was hypothesized that this might be due to a different ability of blood PtC breakdown as a source for mammary fatty acid synthesis. Here, using the same set of animals, we found that the milk level of GPC and the GPC/PC ratio are highly heritable traits. A major QTL on BTA25 significantly affects milk GPC and GPC/PC, but not the level of PC. Within the QTL region, the APOBR gene coding for the apolipoprotein B receptor is located. The analysis of this gene revealed highly significant associations with GPC and the GPC/PC ratio, suggesting APOBR harbors the causative variant underlying this QTL. These findings support the hypothesis that differences in the ability to take up blood PtC from LDL play a role in the metabolic stability of dairy cows during early lactation. However, the identification of the causative mutation was not possible by purely genetic approaches; further studies will be necessary. ACKNOWLEDGMENTS
GRANTS This study was funded by the German Federal Ministry of Education and Research in context of the Fugato-plus MeGA-M framework and by the Bavarian Genome Network (BayGene), Munich, Germany. The genotyping of the cows was financially supported by the German Federal Ministry of Education and Research (project FUGATO-plus GENOTRACK, grant no. 0315134A) and the KMSH (Kompetenzzentrum Milch - Schleswig-Holstein, Kiel, Germany). DISCLOSURES No conflicts of interest, financial or otherwise, are declared by the author(s). AUTHOR CONTRIBUTIONS Author contributions: J.T., G.T., and N.K. conception and design of research; J.T., I.H., M.S.K., W.G., and P.J.O. performed experiments; J.T. and C.H. analyzed data; J.T. and N.K. interpreted results of experiments; J.T. prepared figures; J.T. drafted manuscript; J.T., M.S.K., W.G., W.J., P.J.O., G.T., and N.K. edited and revised manuscript; J.T., C.H., I.H., M.S.K., W.G., W.J., P.J.O., G.T., and N.K. approved final version of manuscript. REFERENCES 1. Andersson L. Concentrations of blood and milk ketone bodies, blood isopropanol and plasma glucose in dairy cows in relation to the degree of hyperketonaemia and clinical signs. Zentralblatt Veterinärmedizin A 31: 683–693, 1984. 2. Aschenbach JR, Kristensen NB, Donkin SS, Hammon HM, Penner GB. Gluconeogenesis in dairy cows. The secret of making sweet milk from sour dough. IUBMB Life 62: 869 –877, 2010. 3. Baird GD, Lomax MA, Symonds HW, Shaw SR. Net hepatic and splanchnic metabolism of lactate, pyruvate and propionate in dairy cows in vivo in relation to lactation and nutrient supply. Biochem J 186: 47–57, 1980. 4. Banos G, Coffey MP, Brotherstone S. Modeling daily energy balance of dairy cows in the first three lactations. J Dairy Sci 88: 2226 –2237, 2005. 4a.Bates D, Maechler M, Bolker B, Walker S. lme4: Linear mixed-effects models using Eigen and S4. http://CRAN.R-project.org/package⫽lme4, 2014.
Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
Downloaded from on May 24, 2015
The authors thank the staff at the Karkendamm research farm for providing milk and plasma specimens and NORD-OST GENETIC GmbH & Co. KG (Verden/Aller, Germany) for kind permission to include their animals in this study. Gabi Ottzen is acknowledged for expert technical assistance. Current address for M. S. Klein: Faculty of Kinesiology, Univ. of Calgary, Calgary, Alberta, Canada.
5. Bell AW. Regulation of organic nutrient metabolism during transition from late pregnancy to early lactation. J Anim Sci 73: 2804 –2819, 1995. 6. Bermudez B, Lopez S, Varela LM, Ortega A, Pacheco YM, Moreda W, Moreno-Luna R, Abia R, Muriana Francisco JG. Triglyceride-rich lipoprotein regulates APOB48 receptor gene expression in human THP-1 monocytes and macrophages. J Nutr 142: 227–232, 2012. 7. Brown ML, Ramprasad MP, Umeda PK, Tanaka A, Kobayashi Y, Watanabe T, Shimoyamada H, Kuo W, Li R, Song R, Bradley WA, Gianturco SH. A macrophage receptor for apolipoprotein B48. Cloning, expression, and atherosclerosis. Proc Natl Acad Sci USA 97: 7488 –7493, 2000. 8. Buitenhuis AJ, Sundekilde UK, Poulsen NA, Bertram HC, Larsen LB, Sorensen P. Estimation of genetic parameters and detection of quantitative trait loci for metabolites in Danish Holstein milk. J Dairy Sci 96: 3285–3295, 2013. 9. Collard BL, Boettcher PJ, Dekkers JCM, Petitclerc D, Schaeffer LR. Relationships between energy balance and health traits of dairy cattle in early lactation. J Dairy Sci 83: 2683–2690, 2000. 11. Ehret A, Hochstuhl D, Krattenmacher N, Tetens J, Klein MS, Gronwald W, Thaller G. Short communication: Use of genomic and metabolic information as well as milk performance records for prediction of subclinical ketosis risk via artificial neural networks. J Dairy Sci 98: 322–329, 2015. 12. Enjalbert F, Nicot MC, Bayourthe C, Moncoulon R. Ketone bodies in milk and blood of dairy cows. Relationship between concentrations and utilization for detection of subclinical ketosis. J Dairy Sci 84: 583–589, 2001. 13. Geishauser T, Leslie K, Tenhag J, Bashiri A. Evaluation of eight cow-side ketone tests in milk for detection of subclinical ketosis in dairy cows. J Dairy Sci 83: 296 –299, 2000. 14. Goff JP, Horst RL. Physiological changes at parturition and their relationship to metabolic disorders. J Dairy Sci 80: 1260 –1268, 1997. 15. Gronwald W, Klein MS, Kaspar H, Fagerer SR, Nürnberger N, Dettmer K, Bertsch T, Oefner PJ. Urinary metabolite quantification employing 2D NMR spectroscopy. Analyt Chem 80: 9288 –9297, 2008. 16. Howie B, Fuchsberger C, Stephens M, Marchini J, Abecasis GR. Fast and accurate genotype imputation in genome-wide association studies through pre-phasing. Nat Genet 44: 955–956, 2012. 17. Ingvartsen KL, Andersen JB. Integration of metabolism and intake regulation. A review focusing on periparturient animals. J Dairy Sci 83: 1573–1597, 2000. 18. Kehrli, ME Jr, Nonnecke BJ, Roth JA. Alterations in bovine neutrophil function during the periparturient period. Am J Vet Res 50: 207–214, 1989. 19. Klein MS, Almstetter MF, Schlamberger G, Nürnberger N, Dettmer K, Oefner PJ, Meyer HHD, Wiedemann S, Gronwald W. Nuclear magnetic resonance and mass spectrometry-based milk metabolomics in dairy cows during early and late lactation. J Dairy Sci 93: 1539 –1550, 2010. 20. Klein MS, Buttchereit N, Miemczyk SP, Immervoll AK, Louis C, Wiedemann S, Junge W, Thaller G, Oefner PJ, Gronwald W. NMR metabolomic analysis of dairy cows reveals milk glycerophosphocholine to phosphocholine ratio as prognostic biomarker for risk of ketosis. J Proteome Res 11: 1373–1381, 2012. 21. Koressaar T, Remm M. Enhancements and modifications of primer design program Primer3. Bioinformatics (Oxford) 23: 1289 –1291, 2007. 22. Leblanc S. Monitoring metabolic health of dairy cattle in the transition period. J Reprod Dev 56: S29 –S35, 2010. 23. Li Y, Willer CJ, Ding J, Scheet P, Abecasis GR. MaCH: using sequence and genotype data to estimate haplotypes and unobserved genotypes. Genet Epidemiol 34, 816 –834, 2010. 24. Minahk C, Kim KW, Nelson R, Trigatti B, Lehner R, Vance DE. Conversion of low density lipoprotein-associated phosphatidylcholine to triacylglycerol by primary hepatocytes. J Biol Chem 283: 6449 –6458, 2008. 25. Montgomery C, Pei Z, Watkins PA, Miziorko HM. Identification and characterization of an extramitochondrial human 3-hydroxy-3-methylglutaryl-CoA lyase. J Biol Chem 287: 33227–33236, 2012. 26. Pryce JE, Veerkamp RF, Thompson R, Hill WG, Simm G. Genetic aspects of common health disorders and measures of fertility in Holstein Friesian dairy cattle. Anim Sci 65: 353–360, 1997. 27. R Core Team. R: a Language and Environment for Statistical Computing. Vienna, Austria: http://www.R-project.org/, 2014. 28. Reynolds CK, Huntington GB, Tyrrell HF, Reynolds PJ. Net portaldrained visceral and hepatic metabolism of glucose, L-lactate, and nitrog-
APOBR GENE IS HIGHLY ASSOCIATED WITH KETOSIS BIOMARKERS enous compounds in lactating holstein cows. J Dairy Sci 71: 1803–1812, 1988. 29. Stegink LD, Coon MJ. Stereospecificity and other properties of highly purified beta-hydroxy-beta-methylglutaryl coenzyme A cleavage enzyme from bovine liver. J Biol Chem 243: 5272–5279, 1968. 30. Suthar VS, Canelas-Raposo J, Deniz A, Heuwieser W. Prevalence of subclinical ketosis and relationships with postpartum diseases in European dairy cows. J Dairy Sci 96: 2925–2938, 2013. 31. Untergasser A, Cutcutache I, Koressaar T, Ye J, Faircloth BC, Remm M, Rozen SG. Primer3–new capabilities and interfaces. Nucleic Acids Res 40: e115, 2012.
137
32. Vazquez AI, Bates DM, Rosa GJM, Gianola D, Weigel KA. Technical note: an R package for fitting generalized linear mixed models in animal breeding. J Anim Sci 88: 497–504, 2010. 33. Wang Y, Li G, Goode J, Paz JC, Ouyang K, Screaton R, Fischer WH, Chen J, Tabas I, Montminy M. Inositol-1,4,5-trisphosphate receptor regulates hepatic gluconeogenesis in fasting and diabetes. Nature 485: 128 –132, 2012. 34. Wathes DC, Fenwick M, Cheng Z, Bourne N, Llewellyn S, Morris DG, Kenny D, Murphy J, Fitzpatrick R. Influence of negative energy balance on cyclicity and fertility in the high producing dairy cow. Theriogenology 68: S232–S241, 2007.
Downloaded from on May 24, 2015
Physiol Genomics • doi:10.1152/physiolgenomics.00126.2014 • www.physiolgenomics.org
