Scaffold Development

Grow Taller 4 Idiots

How to Grow Taller

Get Instant Access

Proteins provide structural integrity and molecular recognition, as well as catalytic processing of most chemical conversions within biological systems. It has become clear with the advent of genomics and proteomics how complex protein interactions are within living systems. The number of expressed proteins in humans has recently been

Figure 8 Crystal structures of CheY from four complexes with different proteins to show variation in loop conformation (green ribbon, top left): (a) IAOO, (b) 1F4V, (c) 1KMI, (d) 1CHN. (Rendered using PyMol.)

Figure 8 Crystal structures of CheY from four complexes with different proteins to show variation in loop conformation (green ribbon, top left): (a) IAOO, (b) 1F4V, (c) 1KMI, (d) 1CHN. (Rendered using PyMol.)

Figure 9 (a) Ribbon diagram of HIV protease with inhibitor (MVT-101, sticks; 4HVP, PDB) bound. The two b-hairpin flaps (on top) have closed down to complete the binding site for the inhibitor, MVT-101; this change is seen with essentially every inhibitor bound to HIV protease. (b) Crystal structure of calmodulin (ribbon diagram, 3CLN, PDB). Note the long blue helix connecting two calcium-binding domains. (c) Calmodulin-peptide complex; notice the dramatic change in the structure of the major helix to allow calcium-binding sites to enfold helical peptide. Calcium ions are shown in red. (Rendered using PyMol.)

Figure 9 (a) Ribbon diagram of HIV protease with inhibitor (MVT-101, sticks; 4HVP, PDB) bound. The two b-hairpin flaps (on top) have closed down to complete the binding site for the inhibitor, MVT-101; this change is seen with essentially every inhibitor bound to HIV protease. (b) Crystal structure of calmodulin (ribbon diagram, 3CLN, PDB). Note the long blue helix connecting two calcium-binding domains. (c) Calmodulin-peptide complex; notice the dramatic change in the structure of the major helix to allow calcium-binding sites to enfold helical peptide. Calcium ions are shown in red. (Rendered using PyMol.)

estimated to be approximately 25 0 00.100 In Escherichia coli, fruit fly, and yeast,101,102 the number of expressed proteins is less, around 6000-15 000, but each protein, on average, has been estimated to interact with up to six other proteins.103 In addition, protein expression and protein-protein interactions were dynamic when changes were monitored during the cell cycle in yeast.104 Despite their dynamic nature, proteins and protein-protein interactions have assisted in targeting particular residues for drug design and have provided scaffolds, or the basis for scaffolds, on which drugs can be designed.

4.02.5.1 Recognition 'Hot Spots'

Side-chain recognition is dominant in peptide-receptor complexes. For instance, aromatic residues have a rigid arrangement and large surface area, causing a great deal of potential free energy and a very low entropic cost that results from binding.105 Some side chains within an interface play a more significant role than others in the energetics of binding and determination of the relative orientation of the two proteins. In the human growth hormone/receptor complex, eight of the 31 side chains involved in the interface accounted for approximately 85% of the binding energy, providing the genesis of the 'hot spot' theory106 as a basis for inhibitor design and drug discovery. The recognition of the hormone somatostatin by its GPCRs further emphasizes the importance of side chains in peptide recognition, as many of the amide bonds can be reduced,107 the direction of the peptide backbone can be reversed, and even the whole peptide backbone can be replaced by a saccharide with recognition retained at the receptor.108'109

Charged groups, particularly the planar guanidinium of Arg and the carboxyl groups of Glu and Asp, are also often essential recognition 'hot spots' in peptide messages, providing a geometrical array of hydrogen-bond donors and/or acceptors. Asn and Gln also have planar groups with distinctive hydrogen-bond geometries. These observations are also valid within the interfaces of protein-protein interactions,53 as formation of salt bridges across the intermolecular interface is highly favorable.110,111 The studies support those of Marshall et al. derived from SAR studies on peptide ligands; tryptophan is most highly enriched at almost fourfold at protein interfaces (Trp having the largest planar surface), followed by Arg, Tyr, Ile, Asp, and His, respectively, with 50% enrichment for His.53

Experimentally, information on 'hot spots' has been obtained by systematically mutating side chains within the interface to alanine and determining the changes in binding affinity. Bogan and Thorn112 collected a database of 2325 alanine mutants for which the change in free energy of binding upon mutation to alanine had been measured.300 Analysis of the database by Bogan and Thorn112 generated several observations; amino acid side chains in hot spots are located near the center of protein-protein interfaces, are generally solvent-inaccessible, and are self-complementary across protein-protein interfaces, i.e., they align and pack against one another. Out of 31 contact residues involved in the interaction of growth hormone with its receptor, for example, two tryptophan residues of the receptor accounted for over 75% of the free energy of binding, as determined by mutation to alanine.113 Polar residues have also been localized at hot spots in protein-protein recognition.53

4.02.5.2 Protein Engineering

Protein engineering involves designing novel three-dimensional protein scaffolds onto which amino acid side chains can be introduced to obtain functionality for molecular recognition and catalysis. Applications could include therapeutics, biomaterials, biosensors, industrial catalysts, nano materials, etc.114-118 Despite the exciting applications, protein engineering is difficult, due to the inability to predict the three-dimensional protein structure from amino acid sequence alone. Within the conformational space a protein fold can adopt, there are often small energy differences between several different low-energy protein folds. Many low-energy alternative folds can bury hydrophobic surface in a compact structure and satisfy most internal hydrogen-bonding groups. An accurate scoring potential and adequate sampling of configurational space for each candidate fold is needed to approximate the entropy of the hydrated system and to help distinguish between alternative folds. When there are no constraints on a protein's function and interactions within a given fold are maximized, particular folds have been stabilized by optimization of the amino acid sequences.119 Nevertheless, optimization utilizing a particular scaffold does not guarantee the sequence will fold to the desired state. In fact, an alternative fold, perhaps a novel fold, may be stabilized to a greater extent than the original fold, causing the alternative fold to form.

Many examples of protein engineering utilize prefolded subunits as scaffolds, simply transferring and/or evolving binding sites or scrambling protein domains to generate new functionality.120-127 This cut-and-paste approach assumes that the scaffolds and domains are sufficiently stable and that they will retain the same three-dimensional structures, despite any perturbation in amino acid sequences involved. The Hellinga group has focused on bacterial periplasmic binding proteins as functional scaffolds for grafting desired binding sites128 with considerable success. Metal-binding sites,129 such as those for zinc,130 have been appended131,132 and eliminated133 from proteins. Others have used covalent modification of existing proteins to generate enzyme activity,134,135 or modified existing enzymes to cause a predicted change in function.136,137

Several recent examples of successful cut-and-paste applications are given below to illustrate the power of this approach. These, and many other examples from the literature, illustrate the creative way in which novel proteins with designed properties can be generated by utilization of pre-existing protein building blocks.

a. Conversion of protein scaffold to enzyme. Dwyer etal.138 designed a novel triose phosphate isomerase active site within the bacterial ribose-binding protein. By knowing the geometry of amino acids residues required for catalytic activity in triose phosphate isomerases, an appropriate geometry was found in the ribose-binding protein that could accommodate the active site. Kaplan and DeGrado139 have reported phenol oxidase activity in a de novo designed four-helix bundle containing two iron-binding sites (Figure 10a and 10b).

b. Change of specificity of bacterial receptors as biosensors. Looger etal.140 attached a fluorophore to the ribose-binding protein to signal ribose binding. The binding site was then mutated to selectively recognize TNT and other small ligands instead of similar ligand analogs shown in Figure 10c. Similar results were reported for the nerve gas agent soman.141

c. Ligand activation of protein splicing. Some bacteria have the capability of splicing two protein segments together through excision of a small intervening cysteine protease or intein.142 By fusing a domain of the receptor responsible for thyroid hormone recognition to an intein, Skretas and Wood143 generated an expression system that allows an inactive protein to be expressed within a cell, and then spliced together to generate an activate protein by the addition of thyroid hormone.

d. Flexible loops in zinc fingers. Another example is the use of the zinc finger domain as a scaffold for protein mimicry. These small proteins associated with DNA regulation also function in protein recognition.144 Sharpe et al.145 demonstrated that approximately 70% of the residues in a small zinc finger could be mutated to alanine without disruption of the fold. A binding site for the co-repressor CtBP2 was grafted onto a flexible loop of a plant homeodomain motif, Mi2b, and yielded the expected change of function.146

e. Surface residues of cystine knot proteins. Venoms from snakes and invertebrates often contain small proteins (cystine knots) with multiple disulfide bridges.147'148 These can be used as scaffolds by grafting critical side chains from one partner in a protein-protein interaction to the appropriate residues on the cystine knot to mimic the recognition surface and inhibit the protein-protein interaction.149

Considerable effort has been expended to design proteins from first principles to test the functional level of our understanding of protein motifs, stability, and protein-protein interactions.150'151 One of the first serious efforts was that of the Richardsons152 to design a novel b-barrel protein, betabellin. They enlisted the synthetic expertise of Erickson to prepare the designed protein, and iterative cycles of design, synthesis, and disappointment ensued until the 15th iteration in design153 gave the desired structure. A major difficulty with designing proteins with b-sheet proteins is their tendency to aggregate.154 Examples of several other successful miniproteins that have been engineered

Figure 10 Examples of cut-and-paste applications. (a) Schematic diagram of four-helix bundle with two iron-binding sites (adapted from 1JM0). (b) Model of active site with substrate interaction with diiron catalytic site (adapted from 1JM0). (c) Conversion of ribose-binding protein (C) into a biosensor for TNT that can discriminate TNB, 2, 4-DNT, and 2,6-DNT; a biosensor that can discriminate L-lactate from D-lactate or pyruvate; and a biosensor that can discriminate serotonin from tryptamine and tryptophan (Rendered using Pymol.).

Figure 10 Examples of cut-and-paste applications. (a) Schematic diagram of four-helix bundle with two iron-binding sites (adapted from 1JM0). (b) Model of active site with substrate interaction with diiron catalytic site (adapted from 1JM0). (c) Conversion of ribose-binding protein (C) into a biosensor for TNT that can discriminate TNB, 2, 4-DNT, and 2,6-DNT; a biosensor that can discriminate L-lactate from D-lactate or pyruvate; and a biosensor that can discriminate serotonin from tryptamine and tryptophan (Rendered using Pymol.).

Pymol Interaction
Figure 11 Examples of miniproteins that have been designed (optimized) to have the folds shown. Top - predominately helical: (a) Mayo et a/.,158 (b) Wells et a/.,164 (c) Andersen et a/.163 Bottom - predominately sheet, (d) Imperiali et a/.,133-165 (e) Serrano eta/.166-167 (Rendered using PyMol.)

are shown in Figure 11. Miniproteins are discrete autonomously folding proteins less than 50 amino acids155 and can be comprised of a-helices, helical bundles, b-sheet, b-hairpins, etc.156-163 (These references are meant to be illustrative, and many other examples are becoming common in the literature.)

Oligomerization of designed units to explore the basis of protein-protein interactions has also been an active field. Helix bundles have been the goal of many efforts in de novo protein design.168 Hecht and co-workers have used binary patterning of polar and nonpolar amino acids169 to guide folding in accord with the lattice model studies of Dill and coworkers.170 Helix bundles have also been designed by DeGrado and co-workers.171-177 Recently, the DeGrado group has generated catalytic activity by adding a diiron catalytic center.139 Hodges,178-181 Kim,182-185 and Keating186-188 have thoroughly explored the physical basis of coiled-coil interactions (Figure 12). Further, DeGrado189'190 and Kim182'183 have successfully designed coiled-coil systems.

The Imperiali group has focused on transforming a zinc finger structure into a structure that assumes the same bba fold without the need to bind zinc.133 Recently, the bba fold was subjected to selection conditions favoring oligomerization.159 The recently solved crystal structure revealed a homotetramer with each monomer containing a bba secondary structure.191 Computational design on the same system by the Keating group has led to a hetero-tetramer (Figure 13).192 This work provides a practical demonstration of the use of positive and negative strategies to prevent homotetramerization while stabilizing heterotetramerization. Bolon etal.193 also discuss the results of the use of positive strategies to stabilize a particular fold, combined with negative strategies to destabilize alternative folds.

Was this article helpful?

0 0

Post a comment