Protein stitchery: design of a protein for selective binding to a specific DNA sequence.

Abstract

We present a general strategy for dsgning proteins to recognize DNA sequences and illustrate this with an example based on the "Y-shaped scissors grip" model for leucine-zipper gene-regulatory proteins. The designed protein is formed from two copies, in tandem, of the basic (DNA binding) region of v-Jun. These copies are coupled through a tripeptide to yield a "dimer" expected to recognize the sequence TCATCGATGA (the v-Jun-v-Jun homodimer recognizes ATGACTCAT). We synthesized the protein and oligonucleotides containing the proposed binding sites and used gel-retardation assays and DNase I footprinting to esablh that the dimer binds specifically to the DNA sequence TCATC-GATGA but does not bind to the wild-type DNA-sequences, nor to oligonucleotides in which the recognition half-site is modifled by single-base changes. These results also provide strong support for the Y-shaped scissors grip model for binding of leucine-zipper proteins.We propose a general strategy for designing proteins to recognize specific DNA-binding sites: this strategy is to select segments of proteins, each of which recognizes particular DNA segments and to stitch these segments together via a short peptide with a cystine crosslink in a way compatible with each peptide being able to bind to its own DNA segment. This technique creates a protein that recognizes the composite site.As a starting point we consider the gene-regulatory leucinezipper proteins. They are characterized by two structural motifs (1-3): (i) the leucine zipper, which is responsible for dimerization, and (ii) the basic region for DNA binding (4-7). The basic regions of unbound leucine-zipper dimers are unfolded but fold into the a-helix conformation upon binding to the specific site (8-10). The most plausible model for the conformation of leucine-zipper protein is the "Y-shaped scissors grip" model (1, 2), in which the basic region of each monomer interacts with DNA on either side of the dyad axis of the binding site. Thus, for yeast transcriptional activator GCN4 each arm recognizes the half-site AGTA (11,12). DESIGN Our design strategy assumes this Y-shaped scissors grip model (Fig. la). We design proteins by crosslinking (stitching together) various binding arms so as to be consistent with the orientation of the recognition helix in each half-site. Here we build upon the results of Kim and coworkers (5, 6), who showed that the leucine zipper of GCN4 can be replaced with linkers (Gly-Gly-Cys) at the C terminus of the DNA binding segment, which upon oxidation dimerize and bind to the same site (ATGACTCAT) as GCN4. As a model system to explore the design of additional DNA-binding proteins, we have chosen the v-Jun leucine-zipper dimer (Fig. la), which also binds to the site ATGACTCAT as a homodimer with itself or as a heterodimer with Fos (4, 13-16), another member of this DNA-binding protein family. We will reverse the sequence relationship of the a-helix to the target nucleotide of the binding arms by adding the Gly-Gly-Cys linker to the N terminus (rather than ...

The Role of Dimerization in Noise Reduction of Simple Genetic Networks

“…The alternative pathway is motivated by the behavior of leucine zipper proteins (Park et al, 1996;Berger et al, 1998). In this pathway, a protein monomer first forms a complex with the DNA by binding to the operator site as a monomer.…”

Factors Governing the Sequence-Selective DNA Binding of Geometrically Constrained Peptide Dimers

“…Similarly, preferential binding to the CRE sequence over the AP1 sequence was observed when a host−guest inclusion complex of β-cyclodextrin and the adamantyl group was used as a non-covalent dimerization module for the basic region peptide of GCN4 . It has been shown that an N-terminal disulfide dimer of the basic region of v-Jun binds to a palindromic DNA sequence in which an arrangement of each half site is reversed in terms of the polarity …”

“…The basic region peptides derived from the basic leucine zipper proteins (bZIP) afford ideal systems for studying the sequence-specific DNA binding of short peptide dimers since the disulfide-bonded dimer of the basic region peptides is sufficient for the sequence-specific DNA binding. , It is now possible to design peptide dimers that possess DNA binding specificities different from those of the native bZIP by appropriately arranging two basic region peptides with an artificial dimerization domain such as a bulky metal complex, − a disulfide linkage, − a host−guest inclusion complex of β-cyclodextrin and the adamantyl group, − enantiomeric bridged biphenyl derivatives, , and a lysine residue . With a bulky metal complex as a dimerization domain for the basic region peptide of GCN4, the dimeric peptide preferentially binds the CRE (5‘-ATGACGTCAT-3‘) sequence over the non-palindromic AP1 (5‘-ATGACTCAT-3‘) sequence. − Because the native GCN4 dimer prefers the AP1 sequence, switching the dimerization module from the leucine-zipper coiled coil into the metal complex changes the sequence selectivity of the basic region dimer of GCN4.…”

Cooperative Oligomerization Enhances Sequence-Selective DNA Binding by a Short Peptide

“…The protein−protein interaction, namely, the dimerization, is mediated through a coiled-coil structure of the leucine zipper domain, while the region rich in basic amino acid residues (the basic region) located N-terminal to the leucine zipper domain directly contacts the DNA major groove. The finding by Kim and co-workers that disulfide-bonded dimers of the basic region specifically bind the DNA sequence as observed for the native bZIP further encourages the use of the basic region as a DNA binding unit for the design of novel DNA binding peptides. , By extending this method, several disulfide dimers and trimers of basic region peptides that target novel nonpalindromic DNA sequences have been reported. − Also sequence-specific DNA binding of covalent basic region dimers using metal complexes − or bridged biphenyl derivatives , as the covalent dimerization domain has been reported. These studies demonstrated that the basic region by itself can be used as the DNA binding domain.…”