Expansion of the Bactericidal/Permeability Increasing-like (BPI-like) protein locus in cattle

Abstract

Background: Cattle and other ruminants have evolved the ability to derive most of their metabolic energy requirement from otherwise indigestible plant matter through a symbiotic relationship with plant fibre degrading microbes within a specialised fermentation chamber, the rumen. The genetic changes underlying the evolution of the ruminant lifestyle are poorly understood. The BPI-like locus encodes several putative innate immune proteins, expressed predominantly in the oral cavity and airways, which are structurally related to Bactericidal/Permeability Increasing protein (BPI). We have previously reported the expression of variant BPI-like proteins in cattle (Biochim Biophys Acta 2002, 1579, 92-100). Characterisation of the BPI-like locus in cattle would lead to a better understanding of the role of the BPI-like proteins in cattle physiology

Results: We have sequenced and characterised a 722 kbp segment of BTA13 containing the bovine BPI-like protein locus. Nine of the 13 contiguous BPI-like genes in the locus in cattle are orthologous to genes in the human and mouse locus, and are thought to play a role in host defence. Phylogenetic analysis indicates the remaining four genes, which we have named BSP30A, BSP30B, BSP30C and BSP30D, appear to have arisen in cattle through a series of duplications. The transcripts of the four BSP30 genes are most abundant in tissues associated with the oral cavity and airways. BSP30C transcripts are also found in the abomasum. This, as well as the ratios of non-synonymous to synonymous differences between pairs of the BSP30 genes, is consistent with at least BSP30C having acquired a distinct function from the other BSP30 proteins and from its paralog in human and mouse, parotid secretory protein (PSP).

Conclusion: The BPI-like locus in mammals appears to have evolved rapidly through multiple gene duplication events, and is thus a hot spot for genome evolution. It is possible that BSP30 gene duplication is a characteristic feature of ruminants and that the BSP30 proteins contribute to an aspect of ruminant-specific physiology.

Figure 1

The BPI-like protein locus in cattle, human and mouse. The position on the chromosome is indicated by nucleotide number (in small font) of the fully assembled human (hg17) and mouse (mm6) chromosomes, obtained from the UCSC Goldenpath genome browser [48]. The cattle coordinates are that of the 759 kbp assembled contig. The shaded and solid bars above the cattle contig indicate the positions of the sequenced BACs. The arrowed end of each gene indicates the direction of the ORF. Intact genes are shaded while pseudogenes are not filled.

Figure 2

Tissue-specific expression of BPI-like genes: Northern blots were loaded with RNA from the following bovine tissues: (1) liver, (2) brain, (3) spleen, (4) skin, (5) abomasum (6) heart, (7) endometrium, (8) bladder, (9) lymph node, (10) testes, (11) kidney, (12) small intestine, (13) white blood cells, (14) sublingual salivary gland, (15) submandibular salivary gland, (16) parotid salivary gland, (17) buccal salivary gland, (18) tonsil, (19) pharyngeal lymph node, (20) nasal mucosa, (21) cheek mucosa, (22) trachea, (23) tongue, (24) soft palate, (25) lung. The blots were probed with full length cDNA encoding SPLUNC3 (top panels), BSP30C (middle panels) and BPIL1 (bottom panels).

Figure 3

Abundance of transcripts for BPI-like genes in different tissues: Reverse transcriptase-PCR analyses were performed using RNA extracted from the following bovine tissues: (1) parotid salivary gland, (2) submandibular salivary gland, (3) buccal salivary gland, (4) tracheal lining, (5) olfactory tract mucosa, (6) nasal mucosa, (7) vomeronasal organ. The PCR was performed using primer pairs specific for bovine BASE and BPIL3, as indicated. The primer sequences are listed in the Methods. The expected product sizes are 596 bp (BPIL3), 454 bp (BASE), and 428 bp (BSP30D).

Figure 4

Phylogenetic analysis of the BPI-like genes: A consensus Maximum-Likelihood tree was generated using the PHYML program after aligning the nucleotide sequences from the N-terminal domain of the two-domain and the full length sequences of the single-domain BPI-like proteins. The values associated with the branch nodes indicate the level of support derived from bootstrap analysis of 1000 replicates. The bar indicates 10 nucleotide substitutions per 100 sequence positions. The sequences used were from human (HS), mouse (MM) or cattle (BT) (see Table 1, Additional file 1 and Additional file 2).

Figure 5

Predicted protein fold for BSP30A: A secondary structure view is shown of the fold predicted for BSP30A by the PHYRE protein threading program, produced as described in the Methods. The amino acids identified by Codeml as being under evolutionary pressure for divergence between the BSP30 and PSP genes, listed in Table 6, are indicated. These amino acids are viewed as a wire representation.

Figure 6

Dot-plot pairwise alignment of the BPI-like locus between species: The genomic nucleotide sequence in mouse and human were aligned using the OWEN program as described in the Methods. The top panels show the cattle – human alignment while the bottom panels show the cattle – mouse alignment. The left hand panels show the alignment across the full locus, while the right hand panels show the area of the locus containing the cattle duplications in more detail.

Figure 7

Mapping of the duplications onto the cattle BPI-like locus: The position of the genomic duplications is shown as shaded boxes above the line. The position of the intact genes is indicated by shaded arrows while the pseudogene is indicated by an unshaded arrow. The stippled bars with arrows indicate the position and direction of the L1Bt LINE element while the shaded vertical bars with arrows indicate the position and direction of the RTEBt1 LINE element.