Mentor: Dimtrios Morikis
Bioengineering 175-Senior Design
4/8/2010

Abstract
Project Objectives
Background
I: Rationale of Project
II: Business Opportunity Prospects
Prior Review
Functional and Performance Specifications
Block Diagram of Problem
Evolution of Final Design
Detailed Description Final Solution
Material Selection
Method of Prototyping Discussion
Performance Testing Protocol Discussion
Performance Testing Results Discussion
Financial Consideration of Design
Conclusion
Future Work
Statement of Social Impact
Appendix
I: List of Abbreviations
II: Project Budget
III: List of Equipment and Facilities
IV: Team Job Responsibilities
V: Detailed Design Drawings
VI: Testing Results
VII: References
2

Many biological processes occur due to shifts in pKa’s. Processes including, catalysis, binding, and intramolecular and intermolecular associations, transpire because protein ioniziablity changes due to changes in environmental conditions. Embedded in a proteins structure reside ionizable amino acids, which help to define the overall charge or electrostatic potential of the protein. These charge characteristics can be elucidated through the calculations of ionizable amino acid titration properties, most important being pKa’s. pKa’s are effected by changes in pH and ionic strength of the environment, and are important because pKa will help to show where the protein is most stable. The analysis presented here will give insight to the generations of a program suite that has the capability of evaluating a wide range of protein characteristics, including pKa’s titrations curves, and stability curves. This suite is unique, when compared to similar packages, in the fact it calls to the Adaptive Possion Boltzmann Solver
(APBS) for accurate free energy calculations, and ultimately more accurate pKa values. Results show correlations of pKa values to already established softwares, with a minimal standard deviation. This program suite has a higher degree of efficacy and accuracy; furthermore this program software will help to further understand protein intrinsic properties that allow it to function and carry out many biological processes that are vital for life.
3

The purpose of our project will be to write a script using the programming language R, which will take any pdb file and calculate the desired thermodynamic properties along with electrostatic potentials, and the incorporation of other programs to create a local and portable convenience package for any to use. This convenience package will be divided into two phases corresponding to winter and spring quarters. The winter phase will consist of first learning the programming language R with the help of graduate student Chris Kieslich and Dr. Thomas Girke, who will be holding an introduction workshop the end of January. We will be working on the incorporation of APBS, which will allow us to calculate the intrinsic pK a
values. These pK a
values will be calculated much like they were in BIEN 135, by the use and understanding of the thermodynamic cycle. Upon the completion of the winter phase, we will continue with algorithms that will calculate different thermodynamic properties. At this time, we will be able to take the “divide and conquer” approach and each group member will be responsible for a particular algorithm. The Spring Phase will cover the statistical approximation for the calculation of apparent pK a
values using the intrinsic values. We will approach the algorithm by either the method of clustering or the Monte Carlo method. To further continue in the convenience of the total package, we will add scripts to print out titration curves for each and all ionizable amino acids; to calculate protein stabilities and, in the case of complexes, binding free energies. The scripts will then
4

be optimized to reduce their size, also to increase speed and efficiency, and to create a pleasant and effective computer-user interface
Proteins are essential for carrying out many biological processes. Proteins carry out their biological due to their functionality. Protein structure and ultimately functionality is influenced in large part by the environment they reside in. Understanding this protein structure-function paradigm is paramount in designing new technologies and advancing fields such as pharmaceuticals and bioengineering. Embedded in a protein structure are ionizable amino acids.
These ionizable groups help to define the overall intrinsic charge characteristics and delineate the overall stability of the proteins. Ionizable amino acids include: Aspartic acid (D), Glutamic acid (E),
Arginine (R), Lysine (K), Histidine (H), Cysteine (C), when it is not involved in a disulfide bond, and
Tyrosine (Y) at high pH’s. Intrinsic to these ionizable amino acids are the corresponding pK a
. The pK a
for an individual amino acid is defined as the pH at which there is a 50% probability of ionization. This
Figure 1 -Depicts electrostatic potential
(isopotential contour) red represents the negative, and the blue represent the positive ionization is described as the extraction and release of proton, which has been termed as the proton transfer phenomena (ref). This proton transfer phenomena is highly dependent on pH,
5

and thus any fluctuations of the protein’s environment can disrupt its overall charge characteristics and abolish the structure-function paradigm. pK a
’s are vital to the many biological processes of proteins. The processes include: catalysis, binding, inter- and intramolecular associations. A specific example of these would be the catalysis of chymotrypsinogen into chymotrypsin which depends on the pH of the small bintestine. A Histidine comes in proximity to an aspartic acid which shifts the pKa of Histidine down to allow Histidine to be charged at a lower pH. The charged Histidine leads to the hydrolysis for the catalysis of the reaction to occur. The intrinsic pK a
describes the ionization process of a specific ionizable group when all other ionizable groups are held constant (Morikis et al ). The intrinsic pK a
can be used to discern the apparent pK a
value of an ionizable group. The intrinsic pK a
can be extracted using the thermodynamic cycle. Other factors that affect pK a
values of ionizable amino acids include ionic strength, temperature, and dielectric constants of the medium. The thermodynamic cycle presented by Morikis et al allows for the calculation of pK a
values for amino acids residues free in solution and within the protein. Furthermore this cycle allows for the
Figure 2- Depicts a thermodynamic cycle, where arrow 1 repersents bound protonated, arrow 2 represents bound deprotonated, arrow 3 represents free deprotonated, arrow 4 represents free protonated calculation of pK a
’s values in both acidic and alkaline environments to help in discerning weather or not the protein is in a favorable environment. Using the thermodynamic cycle, directionality is defined by going with the arrows by adding and subtracted when going against it. Within the thermodynamic cycle there are four states that will be addressed. The states
6

correspond to the free amino acid charged and uncharged then the amino acid bound to the protein in a charge and uncharged state. Each state has a corresponding free energy value that will be calculated using the linearized Poisson-Boltzmann equation (ref).
Previous research has elucidated protein characteristics by first calculating free energies using the Poisson-Boltzmann equation, the Generalized Born models, and semi-empirical approaches (ref). These approaches have proved to computationally inefficient and notably less accurate than using a linearized Poisson-Boltzmann approach. There have already been established web based software to calculate protein titration characteristics using the approaches mentioned above, however they have proved to be inaccurate and inefficient because of empirical data and total calculation time(delete if no ref). Our method is unique because of the incorporation of the Adaptive Poisson-Boltzmann Solver (APBS), which uses the linearized
Poisson-Boltzmann equation to calculate free energies. APBS will then allow us to make pK a calculations, generate titration curves, and also stability curves. This approach has also been proven to increase efficiency and accuracy compared to the approaches previously mentioned.
Our approach will also be a local stand-alone software package without the need of an Internet connection and for any to use while remaining operating system independent
Prior Art Review
There have been programs made to in essence achieve the same goals, as our program will. However as stated previously our is unique in that it uses a linearized Poisson Boltzmann algorithm to calculate free energies, which allows for our software to be more accurate compared to the other software and also will be a computationally more efficient method. Another limitation of already established software are many are web interfaced, which puts a limitation to
7

the portability a limitation that will not be present in our program. Our software will also provide the user not only titration curves the entire protein, and single amino acids, but will also provide the user stability curves, which is a property that has been previously neglected by other software. Some available and open sourced software include Dr. Jens Nielsen pKD Residing
Protein pK a
Values software package developed at the Technical University of Denmark;
Karlsberg+ software, there method is based on the numerical solutions of the linearized Poisson
Boltzmann equation. This software is also capable of doing Monte-Carlo sampling of protein wprotonation patterns; H++: a server for estimating pK a
‘s and adding missing hydrogens to macromolecules. By Gordon JC, Myers JB, Folta T, Shoja V, Heath LS, Onufriev A; Nucleic
Acids Res. 33, W368-71 (2005
)
The design project will follow the functional specifications as desired by Dr. Morikis. Its main purpose will be to take any PDB file and generate titration curves and the statistical information needed for the analysis of protein stabilities. To do this, it will incorporate APBS and PBD2PQR in order to calculate free energies, which will be used to calculate the intrinsic pK a
values. The software package will then use these values to further calculate apparent pK a
values by taking the statistical approximations from the Monte Carlo method or Clustering for n titratable groups. The software package will generate the corresponding titration curves for the overall protein and will be used for the stability analysis. The package also needs to be user friendly, so it will include a graphical user interface that will make it easy for anybody to use. The calculations mentioned will take a considerable amount of time, so performance of the written scripts is an important issue. The scripts written will be optimized to best minimize the time needed in order to complete calculations.
8

Learn R
Use APBS
•VMD interface
•Stand Alone
Workout the analytic expressions for the calculations of intrinsic pK a values
Test each ionizable amino acid individually, using APBS and thermodynamic cycle with simple scripts
Work out statistical approximation to be used to calculate pK a values
Consolidate The scripts for each amino acid intrinsic pK a calculation into one general script and optimize
Devise a methd to reduce the number (and time) of required calculations
Optimize Calculation parameters
Use algorithm either
Monte Carlo or clustering
Write scripts in R to display protein tritration curves
Write scripts in R to determine thermodynamic stabiliity of protein
Design easy to use graphical interface for program suite
9

Calculate apparent pK a
Generation of titration curves for overall protein and its ionizable groups
Analysis of protein stability within any environment
Calculation of intrinsic pka for
1,2,3, and n ionizable group
Statistical analysis for n ionizable groups
Financial considerations for this project since all software software used are open source; there are no monetary financial needs. All computers to design software were all ready purchased by the group individuals prior to beginning of the project. Since this program will be also be open source, there are no financial considerations in the selling of the product.
10