Genetic sequences of hormone response elements share similarity with predicted alpha helices within DNA binding domains of steroid receptor proteins: A basis for site-specific recognition

Abstract

The 150 amino acid sequence comprising the DNA binding region of rat glucocorticoid receptor protein, RGRDBR, was compared to amino acid sequences of members of the superfamily of eukaryotic DNA regulatory proteins. Maximal similarity fell within the 86 amino acid sequence of RGRDBR reported to contain both DNA binding and transcription regulating properties and within the reported DNA binding regions of those proteins to which it was compared. Chou-Fasman secondary structure predictions within these DNA binding domains revealed a conserved alpha helix-beta turn-alpha helix motif. The 450 nucleotide sequence comprising the complementary DNA (cDNA) of amino acids making up RGRDBR was compared to a nucleotide sequence (−312 to −38) from mouse mammary tumor virus 5′ long terminal repeat, MMTV5LTR, known to contain glucocorticoid response elements (GREs). The maximally similar subsequence was found within the coding region for predicted alpha helix B of RGRDBR (nucleotides 1376 to 1412) and within a reported GRE of MMTV5LTR (nucleotides −199 to −131). This MMTV5LTR GRE sequence contains an imperfect palindrome of TGTTCT which is the specific recognition motif for DNA binding by both glucocorticoid and progesterone receptors. Since there are multiple coding possibilities for the majority of the 20 known amino acids, the exceptions being methionine and tryptophan which have a single codon, to thoroughly investigate the extent of genetic information conserved between RGRDBR and GRE, we converted this MMTV5LTR GRE nucleotide subsequence (−199 to −131) to amino acids in all three reading frames reading rightward and leftward in both strands. This procedure revealed all coding possibilities within the MMTV5LTR nucleotide subsequence, as well as the location of the codon sites. A comparison of these MMTV5LTR amino acid coding possibilities to RGRDBR predicted helix B amino acids revealed highly conserved genetic information localized within the GRE half-sites, predominantly in the right half-site containing the TGTTCT sequence. In the absence of atomic coordinates for eukaryotic DNA regulatory proteins, a computer model of a eukaryotic/procaryotic hybrid protein was created with RGRDBR predicted helix B replacing helix of F of E. coli cAMP-dependent regulatory protein (CRP) for which coordinates from X-ray crystallography were available. This hybrid protein was docked onto MMTV5LTR at the region of maximal similarity to helix B. Our computer model shows that the side chains of amino acids within RGRDBR helix B are oriented toward, and appear to be capable of interacting with, nucleotides on both strands of their respective codons within a functional GRE. Calculations of H-bonding in this model indicate that amino acids of helix B are forming H-bonds with nucleotides of their cognate codon/anti-codon sites within the major grooves of the GRE half-sites.