Comparison of the CD14 sequences of 106 Canadian Holsteins and 46 Jersey cows with published sequences (GenBank Nos. NW_001495367 and D84509) revealed a total of five SNPs including one in the 5' untranslated region (UTR) (g.C1291T, numbering is according to GenBank No. EU148609), two in the coding regions (g.A1908G and g.A2318G) and two in the 3' UTR (g.A2601G and g.G2621T) (Table 1). Four of the SNPs are transitional mutations while SNP 2621 involves the transversion of guanine to thymine. SNP 1908 is responsible for a non-synonymous codon change in amino acid 175 of the protein, from Asn (aac) to Asp (gac), while SNP 2318 results in a synonymous codon change without a change in amino acid 311 (Pro, cca vs ccg) of the protein. The coding region SNPs characterizes three gene alleles A (A₁₉₀₈A₂₃₁₈) (GenBank No. EU148609), A₁ (A₁₉₀₈G₂₃₁₈) (GenBank No. EU148610) and B (G₁₉₀₈G₂₃₁₈) (GenBank No. EU148611) and two deduced protein variants A (Asn175) (A₁₉₀₈A₂₃₁₈, GenBank No. ABV68569 or A₁₉₀₈G₂₃₁₈, GenBank No. ABV68570) and B (Asp175) (G₁₉₀₈G₂₃₁₈, GenBank No. ABV68571). Protein variant A is more common in the breeds analyzed with a frequency of 88.6% in Holsteins and fixed in Jerseys (Table 1). Within protein variant A, gene allele A₁ occurred at a very high frequency (80.2%) in Jerseys. The non-coding SNPs occurred at about equal magnitudes in Holsteins while the frequencies of T₁₂₉₁, G₂₆₀₁ and T₂₆₂₁ were above 80% in Jerseys.

Comparison of deduced protein sequences with reported CD14 protein sequences for cattle (UniProt AAD32215 and UniProt NP_776433) revealed a further amino acid difference, 209Ser→Thr. Further comparisons revealed that amino acid 175Asn is conserved in cattle, buffalo (UniProtKB ABE68724), goat (UniProtKB ABE68725) and sheep (UniProtKB NP_001070677) while Asp is present at this position in cattle variant B.

Considering the five SNPs identified and the genotype information of all individuals sequenced, the program PHASE V2.2.1 determined a total of four potential haplotype combinations (CAAAG-C₁₂₉₁A₁₉₀₈A₂₃₁₈A₂₆₀₁G₂₆₂₁, CGAAG, TGGGT, TAGGT) in the analyzed breeds. TGGGT was absent in Jersey while CGAAG was absent in both populations. The gene alleles A1 and B were observed from the sequencing data to be associated with SNPs T₁₂₉₁, G₂₆₀₁ and T₂₆₂₁ while allele A was associated with C₁₂₉₁, A₂₆₀₁ and G₂₆₂₁ therefore giving rise to three actual haplotypes (TAGGT, TGGGT and CAAAG) in the analyzed populations (Table 1). The frequency of the haplotype associated with allele A (CAAAG) was highest (65.8%) in Holsteins while the haplotype associated with allele A₁ (TAGGT) was highest in Jerseys (80.2%) (Table 1).

Whole blood from healthy Holstein cows (animals showing no outward symptoms of infection and farm record indicating milk somatic cell counts below 200,000 cells/ml) with different CD14 genotypes were incubated with fluorescein isothiocyanate (FITC)-labeled mouse anti-human CD14 antibody to determine the effects of genotypes on the expression of CD14 antigens on the surfaces of monocytes and neutrophils. The results are presented in Figure 1 and Table 2. In Figure 1, a higher percentage of gated monocyte cells were found in the LogFITC region labeled M2 (log10² and above) which indicates a higher fluorescence intensity coming from cells with the most CD14 antigens on their surfaces and termed the high expression region. On the other hand, more neutrophils were in the M1 or low expression region (Log10¹ to 10²). The mean channel fluorescence (MCF) intensities (for all cows) observed for the gated regions were M1 = 32.04 (range 16.67 – 65.18), M2 = 288.97 (149.85 – 566.47) and M3 = 201.44 (73.57 – 437.56) for monocytes and M1 = 22.76 (16.20 – 35.52), M2 = 267.33 (148.26 – 536.87) and M3 = 34.76 (19.49–60.83) for neutrophils. As presented, the M3 (M1 + M2) MCF intensity for monocyte was higher than for neutrophils. SNP A1908G that changes amino acid 175Asn to Asp, and thus protein A to B, is significantly (P < 0.01) associated with a higher number of neutrophils in the M2 or higher expression zone. 7.26% of gated neutrophils from cows of genotype 1908AG were found in M2 as compared to 4.36% from cows of genotype 1908AA (Table 2). In the M1 zone, the percentage of neutrophils from cows of both genotypes was the same. For monocytes, a higher number of total cells stained was observed for 1291CT (P < 0.05) and the other genotypes (2318AG, 2601AG and 2621GT) in perfect linkage disequilibrium with C1291T but this difference disappeared when Scheffe's adjustments was applied to means. A similar result was recorded for haplotypes.

The CD14 sequence analyzed in this study is made up of 2630 bps with a 1213 bp promoter (Figure 2), two exons and one intron (GenBank No. EU148609). With the use of bioinformatics tools and published information in the literature, we were able to identify putative transcription factor binding sites (TFBSs) (with 100% match against searched data bases) on the bovine CD14 promoter. The putative TFBSs described here are, in particular, those already demonstrated to control the expression of the gene in human and rat. The putative motifs are shown in Figure 2 (boxed) and include amongst others 9 PU.1, 7AP-1, 5 SP1, 4 C/EBP, 3 c-Myb and 2 AP-2 sites.

We also searched for conserved regulatory motifs in the core promoters of the CD14 genes of different species (cattle, human, mouse, rat and pig) through a query in the NSITEM data base http://linux1.softberry.com/berry.phtml?topic=nsitem&group=programs&subgroup=promoter. The results indicated a total of 63 motifs of 50 regulatory elements (REs) that were conserved in the analyzed breeds with zero to 3 bp mismatches (Figure 3). Nine of the motifs representing 6 REs including C/EBP and Spi-1/PU.1 were conserved in all the species while 4 motifs of 3 REs including AP1 and C/EBP-alpha were conserved in 4 species including cattle (Figure 3). Furthermore, 60.30% (38) of the motifs were conserved between cattle and at least two other species. Sequence alignment of the same region revealed a perfect conservation, both in nucleotide number and orientation of the TATA box in cattle, human, mouse and rat (data not shown). The TATA box of pig differed from the others by only one bp mismatch.

Further, we checked the degree of conservation between the CD14 proteins of cattle with those of different species. Protein sequences compared were those deduced in this work, variants A (ABV68569) and B (ABV68571), other published bovine sequences UniProt: BAA21517, AAD32215 and NP_776433, protein sequences of buffalo (ABE68724), goat (ABE68725), sheep (NP_001070677), pig (AAY98033), mouse (CAA32166), rat (NP_068512) and human variant 1 (NP_001035110) and 2 (NP_000582). The analysis revealed extensive conservations in the amino acid composition and structure. For the bovine proteins, the sequence of ABV68569 is the same as BAA21517 while AAD32215 and NP_776433 differed from ABV68569 by having amino acid 209 changed from Ser to Thr. This indicates the presence of a further CD14 protein variant in cattle here named C. Furthermore, the amino acids of the CD14 proteins of buffalo, sheep and goat shared high conservation rates of 97.05%, 95.17%, and 87.40% respectively, with the bovine ABV68569, followed by pig (76.94%) (Figure 4). The rate of amino acid conservation of bovine A variant was less with the human (72.39%), mouse (61.66%) and rat (60.59%) proteins. While bovine, buffalo, goat, sheep and pig proteins are made up of 373 amino acids, the human protein has two more, and the rat and mouse have one more and seven less animo acids, respectively. The signal peptide, made up of the first 20 amino acids, was highly conserved (only one amino acid difference, 14Ser to Pro) between the bovine and buffalo/goat/sheep proteins. The difference between cattle and the pig was 4 amino acids while being highly diverged with human, mouse and rat (7 to 12 amino acid differences). These relationships were further represented phylogenetically (Figure 4). As depicted in Figure 4, three main groups were evident; the rat and mouse proteins in group one (bootstrap value 100), the human proteins in another group (bootstrap value 100) while members of the Artiodactyla order (bovine, buffalo, goat, sheep) and pig formed a group of their own (bootstrap value 99). In the third group, the pig formed an outcrop of its own, while a closer relationship was visible between cattle and buffalo on the one hand (bootstrap value 72) and, goat and sheep on the other hand (bootstrap value 78).

Genomic DNA was extracted from the blood of 106 Canadian Holstein cows kept at the Howard Webster Centre-Macdonald Teaching Farm, McGill University and the milks of 46 Jersey cows enrolled in the Quebec Dairy Production Centre of Expertise program http://www.valacta.com using Nucleospin Blood Mini Kit (Macherey-Nagel Inc. Easton, PA) as described by the manufacturer. In the case of milk DNA isolation, the manufacturer's protocol was slightly modified. Milk samples were initially centrifuged at 13000 rpm for 15 min at 4°C to remove excess fat before proceeding with the manufacturer's protocol.

Four primer pairs were designed with Invitrogen's OligoPerfect™ software (Invitrogen, Canada Inc., Burlington, ON, Canada) based on GenBank No. NW_001495367 and used to amplify overlapping regions covering the whole Bos taurus CD14 gene (Table 3). Invitrogen synthesized the primers.

PCR reactions with all primer pairs were each carried out in a total volume of 45 μL containing 50 ng DNA, 0.25 mM dNTPs, 2.0 to 2.5 mM MgCl₂ (Table 3), 10 μM of each primer, 2 units Tag DNA polymerase (Fermentas Life Sciences, Burlington, ON, Canada) and 1× Taq buffer. The cycling conditions, with PTC-100™ thermal cycler (MJ Research, Inc., Watertown, MA, USA) included an initial denaturation for 2 min at 94°C followed by 30 cycles comprising 30 sec at 94°C, 30 sec at 60°C, 50 sec at 72°C, and a final elongation for 5 min at 72°C. Both directions of amplified PCR products were sequenced by McGill University/Genome Quebec Innovation Centre using the big dye termination technique and an ABI 3700 sequencer.

Sequences were processed with Chromas, version 1.45 http://www.technelysium.com.au/chromas14x.html) and comparison with other published sequences was done with the multiple sequence alignment program with hierarchical clustering, Multalign http://bioinfo.genopole-toulouse.prd.fr/multalin/multalin.html. CD14 protein sequences of different species (cattle, buffalo, goat, sheep, pig, man, mouse and rat) were aligned or processed with MEGA3.1 software 31 and phylogenetic relationships also constructed with the same software.

The promoter region was analyzed for the presence of putative transcription factor-binding sites using the combined search query against the TRANSFAC database with a maximum allowable string mismatch of 10% [32; http://www.cbil.upenn.edu/cgi-bin/tess/tess]. The combined search query option was used to take advantage of the full power of combined string and weight matrix searching, pre-filtering of factors, significance p-values, and new information in new databases. Particular attention was paid to binding sites already proven to be of significance in regulating the CD14 gene of other species and other common binding sites in mouse and human.

Since regions of conserved non-coding sequences between closely related or divergent species are likely to have common functional roles, we searched the region, about 500 bps of the core promoters of cattle (this study or GenBank No. EU148609), human (GenBank No. U00699), mouse (GenBank No. X13987), rat (GenBank No. AF087944) and pig (GenBank No. DQ079063) for described conserved regulatory element (RE) motifs against the NSITEM data base http://linux1.softberry.com/berry.phtml?topic=nsitem&group=programs&subgroup=promoter. These species were chosen because of the availability of complete or partial promoter sequence information.

Flow cytometry was used to study the effects of identified CD14 SNPs on the expression of CD14 on the surfaces of neutrophils and monocytes in healthy cows. Blood was collected from the caudal vein of 64 Holstein cows with known CD14 genotypes by venipuncture into vacutainer tubes coated with heparin anticoagulant (BD Biosciences, Franklin Lakes, NJ, USA). After collection, samples were stored on ice and analyzed within four hours. One hundred microlitre of heparinized whole blood was placed in a 12 × 75 mm flow cytometric (FCM) tube and incubated with 10 μL of fluorescein isothiocyanate (FITC)-labeled mouse anti human CD14 antibody (ABD Serotec Inc., Raleigh, NC, USA). This was mixed (Barnstead Thermolyne, Dubuque, IOWA, USA) thoroughly and incubated at room temperature on an orbitron rotator (Boekel Ind. Inc., PA, USA) for 30 minutes. Lysis and fixation of erythrocytes was done by adding 2 mL of lysing solution (PHAGOTEST^® Kit, Orpegen Pharma, Heidelberg, Germany) to the mixture. This was mixed gently, incubated for 20 minutes on an orbitron rotator at room temperature and centrifuged at 250 g for 5 minutes at 4°C. The supernatant was aspirated leaving approximately 400 μL of cells in the FCM tube. This was washed with 3 mL of Dulbecco's phosphate buffered saline (DPBS) pH 7.2 (Life Technologies) by centrifuging at 250 g for 5 minutes at 4°C. The supernatant was aspirated as described above and the cells resuspended in 1 mL of DPBS and analyzed by flow cytometry (Becton Dickinson Immunocytochemistry Systems, San José, CA, USA) within 30 minutes. Excitation of samples was at 488 nm; with FITC fluorescence measured at 525 nm ± 10 nm. Acquisition was stopped when 20,000 gated events were collected in the fluorescence cell count histogram. Gating of monocytes and polymorphonuclear leukocytes was based on forward scatter and side scatter dot plots by encircling the populations with amorphous regions. All parameters were recorded with logarithmic amplifications. List mode flow cytometric data from 20,000 events were stored and processed with the Windows Multiple Document Interface for flow cytometry (WinMDI) software version 2.8 (Joseph Trotter, The Sripps Research Institute, http://facs.scripps.edu/software.html) on a personal computer.

The viability of neutrophils and monocytes in whole blood was determined by propidium iodide (PI) exclusion (50 μg/mL, final concentration) using flow cytometry after cells were incubated for 10 minutes in the dark at room temperature. The cells showed 99% viability.

Allele frequencies were estimated with GENEPOP program 33 while haplotypes and their frequencies were determined with the program PHASE V2.1.1 3435. PHASE implements a Bayesian method of haplotype reconstruction based on genealogies reconstructed from coalescent theory under a Markov Chain Monte Carlo framework and has been shown to outperform other strategies such as the maximum likelihood expectation maximization algorithm in most cases 35.

Flow cytometric data were analyzed as a one-way ANOVA using the MIXED procedure SAS 36. Treatment means were separated using the least square means option of SAS. Differences between treatment means were tested using Scheffe's Multiple Comparison test and statistical significance was declared at P < 0.05.

Statistical model used: Y_ij = μ + genotype_i + e_ij