Barry Olafson, founder and president of Molecular Engineering Corporation, and Professor Fred Lee, creator of the influential molecular simulation software POLARIS, for teaching me everything about molecular simulations. Newburgh, Executive Officer of the Protein Society, Professor Steve Frautschi, Professor Henry Lester, Professor Sela Mager, Professor Steve Mayo, Dr. Jen Sun and Lavonne Martin for getting me out of the lab and encouraging me to work out in the gym.
First, the effect of explicit backbone motion on the selection of amino acids in protein design was assessed in the core of the streptococcal protein G~1 domain (G~1).
Chapter 1 Introduction
In addition, some "other" residues may be misinterpreted or poorly defined in the crystallographic structures. Sequence ribbon and proposed three-dimensional structure of disulfide-labeled Felix (reprinted from Hecht et al., 1990). At the heart of the design approach is a "design cycle" in which theory and experiment alternate.
This is followed by experimental construction and analysis of the properties of the engineered protein. On the other hand, it was observed experimentally that the backbone of the target structure could move to facilitate potentially devastating mutations (Baldwin et al., 1993; Lim et al., 1994). Analysis and prediction of the packing of a-helices versus a b-sheet in the tertiary structure of globular proteins.
Chapter 2
Ribbon diagram of G,Bl showing the positions of the 10 core residues examined in this study. The van der Waals radii of the atoms in the simulation were scaled by either 1.0 or 0.9. Global optimum sequences for each of the backbone variants were found using the Dead End Elimination (DEE) theorem (Desmet et al Goldstein, 1994).
Y3F and V391 are likely the result of the hydrophobic surface burial term in the scoring function. The wild-type G,BI sequence and position numbers are shown at the top of the table. 34; Wild-type Gfil sequence and position numbers are shown at the top of the table.
Vol is the fraction of core-side chain volume relative to the Gfil sequence.: T,;, IS the melting temperature measured by eo: NMR is a qualitative indicator of the degree of chemical shift dispersion in the 10 IH NMR spectra; NO mdlcates a propeny that has not been determined. 34; The GpI wild-type sequence and position numbers are shown at the top of the table. The GPI wild-type sequence and position numbers are shown at the top of the table.
Far-UV CD spectra of G,Bl and the most disordered of the mutants of the tJ.1l series. Two crystal structures of the J31 immunoglobulin-binding domain of streptococcal G protein and comparison by NMR.
Chapter 3
The algorithm we developed uses supersecondary structure parameters to determine the orientation between secondary structural elements relative to the target protein fold. Given a certain number of residues for each secondary structural element, a single optimal configuration is found. When choosing a set of motifs to demonstrate the generality of the algorithm, we chose motifs /3/3a, /3a/3 and aa, as their combination can be used to approximate the most possible backbone bend.
The combination of these three motifs includes all possible classes of binding ring (Donate et al., 1996) and all possible secondary structural elements (a-helix, parallel. ~-sheet, and anti-parallel /3-sheet) . Our design methodology consists of an atomistic backbone design algorithm that quantitatively considers specific interactions between secondary structural elements. The backbone design algorithm screens all possible geometric orientations of secondary structural elements and finds the optimal placement of these elements given a target protein fold.
Given a specified number of residues for each secondary structural element, this procedure yields a family of optimal configurations for each motif. The parameters are shown in Figure 1. The cr parameter describes the rotation angle around the helix axis, and the Ie parameter describes the translational distance along the helix axis. The parameter d describes the distance between the centers of the two strands or helices, and the parameter 1 describes the angle between them.
The Q parameter describes the angle between the strand axis and the helix axis after projection onto the sheet. The parameter h describes the distance between the center of the helix a and the plane of the sheet ~.
N· helix axis
There are no local variables, $i and 1jIi are the number of residues and the backbone $/1jI dihedral angles for the ith axis. These atoms were then mapped to target locations via standard Euler matrices, where the matrix elements consist of combinations of supersecondary structure parameters (see Table 2a,b,c). Structures that clash (with positive van der Waals energy) and structures where none of the corresponding loops provided closure were rejected.
The cost function value was then determined for each of the remaining structures based on nucleating C~-C~ interactions and the total number of hydrogen bonds. The number of hydrogen bonds between the ~ strands was determined as described in Figure 2; each hydrogen bond was counted as -l. The best backbone structures obtained for each of the three motifs are shown in Figure 3.
The structure of the ~~a motif resembles a zinc finger, and the structure of the ~a~ motif resembles a segment of a ~ barrel. The ~-strands and the a-helix are first built along the z-axis and then mapped according to x~Rx+f+Aii, where x runs through all atoms in the strand or helix. Once located, the backbone configuration of these atoms consisted of four terms: (1) the number of hydrogen bonds between paired strands; (2) the number of inter-segment CI3-CI3 pairs that are within 4.1 A to 6.6 A of each other;.
(3) the number of CI3-CI3 pairs between segments closer than 4.1 Å; The costs are set to a large positive number, if the Van der Waals interaction is positive or if no loops from the database can successfully close the ends, otherwise they are set to the negative sum of the number of nucleating CI3-CI3- pairs plus the number of hydrogen bonds.
4) rejects structures
5) rejects structures
Chapter 4
In this experimental paper, we evaluate our backbone design method by testing the thermal stability and structural properties of the designed peptides. It was found that small differences in the number and location of the hydrophobic residues can significantly change the thermodynamic behavior of the designed peptides. The automation of the second step has been successfully carried out by Dahiyat and Mayo (Dahiyat & Mayo, 1997).
To validate our backbone design algorithm, we decided to measure the thermal stability and structural properties of one representative family of optimal configurations using CD and 1D NMR. The thermal stability and structural properties of a representative configuration for the ~~a motif (bbal, see Figure 1 and Table 1a) were measured using CD and 1D NMR. This means that the sequence selection algorithm was required to select one of the hydrophobic residues for that position.
The number of residues in each secondary structural element can be determined by considering the geometry of the fold. This means that the relative orientation of the ~-strands and the a-helix determines the length of the secondary structural element. First, we assumed that the number of residues at the Nand · C ends matches the periodicity of the neighboring secondary structural elements.
Next, we optimized the number of residues for each of the ~-strands and the α-helix. Therefore, out of the 28 residues in the /3/3a fold, the first 12 residues will belong to the /3 strands (2B + N terminal residues), the next 2 residues will belong to the /3-a loop, and the last 14 residues will belong to the ex-helix (A + C terminal residues). b) Determine the balance between hydrophobic and hydrophilic residues.
2nd I)-strand
Chapter 5
764 Biophysical Journal, Volume 70 February 1996. the current version of the model. the intrinsic jumping frequency varies depending on the substrate, but is constant at all locations in the pore. These attraction and repulsion factors also include the possibility that the field changes with the ionic content of the channel. The integration of the transient response gives the total number of charges within the membrane.
This additional constraint in the simulations allowed us to determine the absolute values of the rate constants. Simulations showed leak currents of ∼1% of the saturated current in the absence of external GABA ( Fig. 4 C ). A Na+/GABA flux ratio between 1 and 2 was obtained (Fig. 4 D); simulated ratio of flux deJX!Dds to membrane potential, a point not tested experimentally for GATt.
In all the above simulations, the internal concentrations of the substrates are assumed to be zero. There are also differences in the electrical positions of the binding sites between these two carriers, accounting for the different sensitivities of the transport voltage. The model can also be used to interpret the variable stoichiometry of the 5 lit transporter.
Much of the flexibility of the model can be attributed to the minimal assumptions made about the transport structures. Ions required for electrogenic GABA transport by catfish retinal horizontal cells.
In chapter 2 I assessed the effect of explicit backbone movement on the selection of amino acids in the design of the core of the 13 1 domain of the streptococcal protein G. The stability and structural flexibility of seven of the redesigned proteins were determined experimentally and showed that core variants containing as many as 6 of 10 possible mutations retained native properties. This result shows that backbone flexibility can be explicitly combined with amino acid side chain selection and that the selection algorithm is sufficiently robust to tolerate perturbations as large as 15% of G131's native supersecondary structure parameter values.
In Chapter 3, I attempted to extend the range of computational protein design by developing a general, quantitative design method for computing de novo backbone models. The method had to calculate atomic resolution backbones compatible with the atomic sequence selection algorithm I was using and had to be applicable to all protein motifs. I chose three motifs to test our method (~~a, ~a~ and aa) as their combination can be used to approximate the folding of the backbone.
The best structure found for the ~~a motif resembles a zinc finger, and the best structure for ~a~. In Chapter 4, I evaluated the backbone design method by testing the thermal stability and structural properties of the designed peptides. I compared our five sequences with FSD-1, a thermally stable protein with a well-defined structure developed in our laboratory using the same sequence selection algorithm (Dahiyat & Mayo, 1997).
For a three-site model, simulated annealing provides parameters to fit steady-state measurements of flux coupling, transport-coupled currents, and charge movements for the GABA GATl transporter. The model also provides parameters that describe the available data for the rat 5-HT transporter and the rabbit Na + glucose transporter.
