How are signalling molecules organized into different pathways within the same cell? In Drosophila, the inaD gene encodes a protein consisting of five PDZ domains which serves as a scaffold to assemble different components of the phototransduction cascade, including the principal light-activated ion channels, the effector phospholipase C-beta and protein kinase C. Null inaD mutants have a dramatically reorganized subcellular distribution of signalling molecules, and a total loss of transduction complexes. Also, mutants defective in a single PDZ domain produce signalling complexes that lack the target protein and display corresponding defects in their physiology. A picture emerges of a highly organized unit of signalling, a 'transduclisome', with PDZ domains functioning as key elements in the organization of transduction complexes in vivo.

Classical studies show that for many proteins, the information required for specifying the tertiary structure is contained in the amino acid sequence. Here, we attempt to define the sequence rules for specifying a protein fold by computationally creating artificial protein sequences using only statistical information encoded in a multiple sequence alignment and no tertiary structure information. Experimental testing of libraries of artificial WW domain sequences shows that a simple statistical energy function capturing coevolution between amino acid residues is necessary and sufficient to specify sequences that fold into native structures. The artificial proteins show thermodynamic stabilities similar to natural WW domains, and structure determination of one artificial protein shows excellent agreement with the WW fold at atomic resolution. The relative simplicity of the information used for creating sequences suggests a marked reduction to the potential complexity of the protein-folding problem.

Statistical analyses of protein families reveal networks of coevolving amino acids that functionally link distantly positioned functional surfaces. Such linkages suggest a concept for engineering allosteric control into proteins: The intramolecular networks of two proteins could be joined across their surface sites such that the activity of one protein might control the activity of the other. We tested this idea by creating PAS-DHFR, a designed chimeric protein that connects a light-sensing signaling domain from a plant member of the Per/Arnt/Sim (PAS) family of proteins with Escherichia coli dihydrofolate reductase (DHFR). With no optimization, PAS-DHFR exhibited light-dependent catalytic activity that depended on the site of connection and on known signaling mechanisms in both proteins. PAS-DHFR serves as a proof of concept for engineering regulatory activities into proteins through interface design at conserved allosteric sites.Proteins typically adopt well-packed three-dimensional structures in which amino acids are engaged in a dense network of contacts (1,2). This emphasizes the energetic importance of local interactions, but protein function also depends on nonlocal, long-range communication between amino acids. For example, information transmission between distant functional surfaces on signaling proteins (3), the distributed dynamics of amino acids involved in enzyme catalysis (4-6), and allosteric regulation in various proteins (7) all represent manifestations of nonlocal interactions between residues. To the extent that these features contribute to defining biological properties of protein lineages, we expect that the underlying mechanisms represent conserved rather than idiosyncratic features in protein families.On the basis of this conjecture, methods such as statistical coupling analysis (SCA) quantitatively examine the long-term correlated evolution of amino acids in a protein family -the statistical signature of functional constraints arising from conserved communication between positions (8,9). This approach has identified sparse but physically connected networks of coevolving amino acids in the core of proteins (8)(9)(10)(11)(12). The connectivity of these networks is remarkable, given that a small fraction of total residues are involved and that no tertiary structural information is used in their identification. Empirical observation in several protein families shows that these networks connect the main functional site with distantly † To whom correspondence should be addressed. rama.ranganathan@utsouthwestern.edu. * These authors contributed equally to this work. The finding that certain surface sites might be statistical "hotspots" for functional interaction with active sites suggests an idea for engineering new regulatory mechanisms into proteins. What if two proteins were joined at surface sites such that their statistically correlated networks were juxtaposed and could form functional interactions (Fig. 1A)? If the connection sites are functionally linked to their respective activ...