Phylogenetic and evolutionary analysis of the PLUNC gene family

Abstract

The PLUNC family of human proteins are candidate host defense proteins expressed in the upper airways. The family subdivides into short (SPLUNC) and long (LPLUNC) proteins, which contain domains predicted to be structurally similar to one or both of the domains of bactericidal/permeability-increasing protein (BPI), respectively. In this article we use analysis of the human, mouse, and rat genomes and other sequence data to examine the relationships between the PLUNC family proteins from humans and other species, and between these proteins and members of the BPI family. We show that PLUNC family clusters exist in the mouse and rat, with the most significant diversification in the locus occurring for the short PLUNC family proteins. Clear orthologous relationships are established for the majority of the proteins, and ambiguities are identified. Completion of the prediction of the LPLUNC4 proteins reveals that these proteins contain approximately a 150-residue insertion encoded by an additional exon. This insertion, which is predicted to be largely unstructured, replaces the structure homologous to the 40s hairpin of BPI. We show that the exon encoding this region is anomalously variable in size across the LPLUNC proteins, suggesting that this region is key to functional specificity. We further show that the mouse and human PLUNC family orthologs are evolving rapidly, which supports the hypothesis that these proteins are involved in host defense. Intriguingly, this rapid evolution between the human and mouse sequences is replaced by intense purifying selection in a large portion of the N-terminal domain of LPLUNC4. Our data provide a basis for future functional studies of this novel protein family.

Figure 1.

Unrooted phylogenetic tree of BPI and PLUNC family members. The PLUNC family branch of the tree is colored black, and the BPI branch is in green. Branches with bootstrap values <85% are colored red; the detailed branching order is uncertain in these regions. Thickened branches delineate groups of orthologous proteins, as defined in the text. See Materials and Methods for details of the regions of sequence used to construct the tree.

Figure 2.

The organization of the mouse, rat, and human PLUNC family gene loci. The stippled and black boxes represent “long” (LP) and “short” PLUNC family proteins (SP). Due to the less complete nature of the rat genomic sequences in the htgs database (as of December 22, 2002) the absolute size of the rat locus may be underestimated. The updated human locus is also presented showing the position of LPLUNC6, XM and BASE and the partial sequence of the ortholog of LPLUNC5 (shown as the white box), which may represent a pseudogene. The sequence of rLPLUNC5 is a near complete prediction. The gene designated XM represents a predicted gene (XM_206487), which has definite PLUNC family characteristics but which we are presently unable to assign to either the Long or Short PLUNC family proteins due to the lack of expression information and the lack of a definitive prediction in GENSCAN (Burge and Karlin 1997). Genes contained within brackets are those for which there is no expression information in any form. Immediately 5′ of the mouse SPLUNC4 gene is a viral gag gene (gag), which due to its position may block transcription of the otherwise completely predicted gene. The size bar represents 50 kb. During our study the cloning and sequencing of human LPLUNC2 and LPLUNC6 was reported (Mulero et al. 2002). In this article the authors designated these proteins as BPI-like proteins (BPIL) 1 and 3, respectively. The position of the human BASE gene (Egland et al. 2003) is also shown adjacent to SPLUNC3.

Figure 3.

Conservation of exon sizes in BPI and PLUNC family proteins. The number of coding nucleotides in the coding exons of BPI and PLUNC family proteins (exons 2–8 in one domain proteins and exons 2–16 in two domain proteins) for (A) hBPI, hLBP, hPLTP, and hCETP; (B) hLPLUNC1, hLPLUNC2, hLPLUNC3, hLPLUNC4, mLPLUNC5, and hLPLUNC6; (C) hSPLUNC1, hSPLUNC2, hSPLUNC3, mPSP, rSMGB, mSPLUNC4, mSPLUNC5. For LPLUNC4, the large additional exon 3b is shown positioned between exons 3 and 4. For CETP, the two exons that correspond to exon 5 of BPI have been shown with a combined size of 159 nucleotides.

Figure 4.

Sequence comparison of mouse and human LPLUNC proteins. (A) The 3D structure of human BPI colored to show the region encoded by exon 2 (purple) and that encoded by exon 3 (white), the remainder of the N-terminal domain (green), and the C-terminal domain (gold). (B) Locally averaged human/mouse sequence identity traces overlaid for SPLUNC1, SPLUNC3, LPLUNC1, LPLUNC2, LPLUNC3, LPLUNC4, and LPLUNC6. The trace for LPLUNC4 is in red; traces for the other proteins are in black. The average at each sequence position was calculated as a Gaussian weighted average of neighboring positions. The standard deviation of the Gaussian function used was 15 residues. The colored bars mark the regions corresponding to the regions colored in (A) mapped to the residue numbers of LPLUNC4. The thin bar shows the position of the GL-rich region in the sequence of LPLUNC4. The two horizontal lines mark the 16th to 83rd percentile ranges of amino acid identity observed in domain-containing protein regions in a survey of the human and mouse genomes (Waterston et al. 2002). Individual traces are available as Supplemental Material.