Search

Human c-myc is critical for cell homeostasis and growth but is a potent oncogenic factor if improperly regulated. The c-myc far-upstream element (FUSE) melts into single-stranded DNA upon active transcription, and the noncoding strand FUSE recruits an activator [the FUSE-binding protein (FBP)] and a repressor [the FBP-interacting repressor (FIR)] to fine-tune c-myc transcription in a real-time manner. Despite detailed biological experiments describing this unique mode of transcriptional regulation, quantitative measurements of the physical constants regulating the protein-DNA interactions remain lacking. Here, we first demonstrate that the two FUSE strands adopt different conformations upon melting, with the noncoding strand DNA in an extended, linear form. FBP binds to the linear noncoding FUSE with a dissociation constant in the nanomolar range. FIR binds to FUSE more weakly, having its modest dissociation constants in the low micromolar range. FIR is monomeric under near-physiological conditions but upon binding of FUSE dimerizes into a 2:1 FIR(2)-FUSE complex mediated by the RRMs. In the tripartite interaction, our analysis suggests a stepwise addition of FIR onto an activating FBP-FUSE complex to form a quaternary FIR(2)-FBP-FUSE inhibitory complex. Our quantitative characterization enhances understanding of DNA strand preference and the mechanism of the stepwise complex formation in the FUSE-FBP-FIR regulatory system.

Bioactive peptides have multiple conformations in solution but adopt well-defined conformations at lipid surfaces and in interactions with receptors. We have used side chain lactam cross-links to stabilize secondary structures in the following peptide models of a conserved N-terminal domain of apolipoprotein E (cross-link periodicity in parentheses): I, H2N-GQTLSEQVQEELLSSQVTQELRAG-COOH (none); III, [sequence; see text] (i to i + 3); IV,[sequence; see text] (i to i + 4); IVa, [sequence, see text] (i to i + 4) (lactams above the sequence, potential salt bridges below the sequence). We previously demonstrated [Luo et al. (1994) Biochemistry 33, 12367-12377; Braddock et al. (1996) Biochemistry 35, 13975-13984] that peptide III, containing lactam cross-links between the i and i + 3 side chains, enhances specific binding of LDL via a receptor other than the LDL-receptor. Peptide III in solution consists of two short alpha helices connected by a non alpha helical segment. Here we examine the hypothesis that the domain modeled by peptide III is one antipode of a conformational switch. To model another antipode of the switch, we introduced two strategic modifications into peptide III to examine structure-function relationships in this domain: (1) the spacing of the lactam cross-links was changed (i to i + 4 in peptides IV and IVa) and (2) peptides IV and IVa contain the two alternative sequences at a site of a possible end-capping interaction in peptide III. The structure of peptide IV, determined by 2D-NMR, is alpha helical across its entire length. Despite the remarkable degree of structural order, peptide IV is biologically inactive. In contrast, peptides III and possibly IVa contain a central interruption of the alpha helix, which appears necessary for biological activity. These and other studies support the hypothesis that this domain is a conformational switch which, to the extent that it models apolipoprotein E itself, may modulate interactions between apo E and its various receptors.

In this paper, we test the hypothesis that peptide models of a highly conserved domain of apolipoprotein E (amino acids 41-60 in human apo E) modulate the binding and internalization of LDL to cell surface receptors in a conformationally specific manner. Three peptides were compared: peptide I containing the natural sequence of amino acids 41-60 of human apo E; peptide III containing side-chain lactam cross-links designed to enhance alpha-helical structure; and peptide II containing cross-links designed to prevent formation of alpha-helices. Peptide III was shown previously to consist of two short alpha-helical domains linked by a turn and to have more alpha-helical content than peptide I, while peptide II was shown to have less helical content than either peptide III or I(Luo et al., 1994). Peptide III induced a 30-fold increase in the specific binding of 125I-LDL to normal human skin fibroblasts and a 60-fold increase in the binding to fibroblasts lacking the LDL-R. This same peptide also restored the binding to normal fibroblasts of 125I-LDL from a patient with familial defective apolipoprotein B, the R3500-->Q mutation. Analysis of binding indicated an increase in the apparent number of binding sites, with little effect on the affinity of 125I-LDL for the cell surface. Heparinase treatment of the cells did not abrogate this effect, suggesting that the increased binding is not mediated by cell surface glycans. LDL internalization but not degradation was also increased by peptide III. Similar but smaller effects were also induced by peptide I. Peptide II was much less active than peptide I or III. Thus, the order of biological activity was the same as the order of alpha-helical content, i.e., peptide III > peptide I > peptide II. These results suggest a hitherto unknown biological function for a highly conserved domain of apolipoprotein E, and this bioactivity was shown by peptide models to be specific to the alpha-helical conformation.

Apolipoprotein E plays a critical role in plasma lipoprotein clearance. A peptide model of a highly conserved domain of this protein has been shown to increase low-density lipoprotein binding to fibroblast cell surface receptors. To distinguish between two potential structures--one essentially alpha-helical and nonamphiphilic, the other an amphiphilic pi-helix--synthetic side-chain lactam constraints have been incorporated into model peptides in order to restrict conformational flexibility favoring either the alpha- or pi-helix. Here we provide CD and 1H NMR data suggesting that the more biologically active, putatively alpha-helical peptide indeed contains two alpha-helical domains separated by a central bend. Whereas previous studies (Osapay & Taylor, 1992; Felix et al., 1988) indicated stabilization of alpha-helices by cross-links between the i and i + 4 residues, the current paper demonstrates that cross-links between the i and i + 3 residues also stabilize the helix. Indeed, the stabilization afforded by these cross-links is approximately 1 kcal/mol, similar to that reported for peptides cross-linked between the i and i + 4 residues, and derives exclusively from a loss of entropy of the unfolded state. The presence of the alpha-helical structure appears to correlate well with biological activity. This study provides initial insight into the bioactive structure of this domain of apo E and suggests strategies as to how peptides can be conformationally constrained to enhance their stability and biological function.