Comparison of complete nuclear receptor sets from the human, Caenorhabditis elegans and Drosophila genomes

Abstract

Background: The availability of complete genome sequences enables all the members of a gene family to be identified without limitations imposed by temporal, spatial or quantitative aspects of mRNA expression. Using the nearly completed human genome sequence, we combined in silico and experimental approaches to define the complete human nuclear receptor (NR) set. This information was used to carry out a comparative genomic study of the NR superfamily.

Results: Our analysis of the human genome identified two novel NR sequences. Both these contained stop codons within the coding regions, indicating that both are pseudogenes. One (HNF4 gamma-related) contained no introns and expressed no detectable mRNA, whereas the other (FXR-related) produced mRNA at relatively high levels in testis. If translated, the latter is predicted to encode a short, non-functional protein. Our analysis indicates that there are fewer than 50 functional human NRs, dramatically fewer than in Caenorhabditis elegans and about twice as many as in Drosophila. Using the complete human NR set we made comparisons with the NR sets of C. elegans and Drosophila. Searches for the >200 NRs unique to C. elegans revealed no human homologs. The comparative analysis also revealed a Drosophila member of NR subfamily NR3, confirming an ancient metazoan origin for this subfamily.

Conclusions: This work provides the basis for new insights into the evolution and functional relationships of NR superfamily members.

Figure 1

Amino acid alignments of the novel NR sequences FXR-r and HNF4γ-r. (a) FXR-r amino acid alignment with FXR (NR1H4). The nucleotide sequences from complete and ordered clone AL358372 contained eight fragments with amino acid sequence similarity with FXR (from 5'-3' relative to FXR-r, nucleotide positions 13144-13480; 15053-15199; 21541-21672; 21781-21879; 22015-22115; 23681-23795; 24486-23795; 27041-27262). The positions of the intron gaps within FXR-r and FXR are boxed. Positions of termination codons within FXR-r are indicated by an asterisk. (b) HNF4γ-r amino acid alignment with HNF4γ (NR2A2). The complete and ordered clones AL449103 and AL138997 contained contiguous HNF4γ-r sequence. Seven frameshifts (indicated by asterisks) were required to preserve the reading frame of HNF4γ-r relative to HNF4γ. Intron locations in the HNF4γ gene are boxed.

Figure 2

Relationships within the completed sets of NR superfamily members from humans, C. elegans and Drosophila. A neighbor-joining tree of NR DNA-binding domain sequences was generated using the paupsearch feature of the GCG 10.1 program package (1,000 bootstrap replicates, midpoint rooting); analysis methods are described in detail in Sluder et al. [5]. Significant bootstrap support values are indicated by slashes on the appropriate branches: (/) 50-79%; (//) 80-94%; (///) ≥ 95%; branches also supported by parsimony analysis are marked by dots; subfamilies are boxed. All human (Hs) and Drosophila (Dm) NRs are included, except for the two human NRs that lack a canonical DBD (DAX-1 (NR0B1) and SHP (NR0B2)). The C. elegans (Ce) sequences include all members of the six major metazoan NR subfamilies as well as selected representatives of the major groupings of divergent C. elegans NRs evident in a larger tree containing all the nematode sequences.

Figure 3

Amino acid alignment of Drosophila CG7404 with the sequences of human ERRα, ERRβ, and ERRγ. DmCG7404 exhibits similarity to the human ERRs in both DNA- and ligand-binding domain sequences. Block boxes indicate residues identical in at least two of the four sequences; gray boxes indicate similar residues. Expression of CG7407 as mRNA and the splicing pattern has been confirmed by the recovery and sequencing of cDNA (K.K. et al., unpublished data).