The rational design of the “super aspirin” Vioxx.

The molecular design of the nonsteroidal anti-inflammatory drug (NSAID) vioxx (Fig. 1 is the chemical structure of vioxx) is an elegant example of rational drug design despite the withdrawal of vioxx from the market over safety concerns. The available NSAID’s, like aspirin or ibuprofen, inhibit both cyclooxygenase (COX) enzymes COX-1 and COX-2 whereas vioxx was designed to specifically inhibit COX-2. The precise inhibition of COX-2 is advantageous because COX-2 mediates the synthesis of “bad prostaglandins” responsible for pain and inflammation, while COX-1 mediates the synthesis of “good prostaglandins” responsible for protecting the stomach lining. Therefore creating selective NSAIDs, like vioxx, that specifically inhibit COX-2 promotes pain relief but doesn’t pose a significant risk for adverse side effects such as peptic ulcers.
As I mentioned in a previous post, rational drug design is guided by the molecular structure of the enzyme active site. The creation of vioxx is a great example of drug design directed by the target enzyme active site, COX-2, as well as the off-target enzyme active site from COX-1. More specifically, vioxx assembly was guided by the structural differences between these two enzymes producing a specific inhibitor for COX-2. The major structural difference between these two enzymes is the substitution of an active site isoleucine in COX-1 with a valine in COX-2. The smaller valine side chain in COX-2 permits access to a small hydrophobic pocket in the enzyme which is blocked by the larger isoleucine side chain in COX-1 (Fig. 2 shows a space filling model of vioxx superpositioned into the active sites of COX-1 in magenta and COX-2 in grey highlighting the important difference between the two enzymes that confers vioxx's preferential binding to COX-2. Notice the clash between the vioxx phenyl ring and the COX-1 isoleucine side chain). Vioxx was designed to bind to this additional pocket enhancing its selectivity for COX-2.

The insulin mimicking effects of vanadyl sulfate.

Vanadyl sulfate (Fig. 1 is the chemical structure of vanadyl sulfate) is a compound consumed by athletes and bodybuilders to enhance muscular strength and development. The biochemical function of vanadyl sulfate in the human body is not completely understood, yet a great deal of attention has been paid to this compound due to its ability to mimic insulin. Although all of the molecular details concerning vanadyl sulfate as an insulin mimic are unclear some aspects have been elucidated as described below.
After ingestion of vanadyl sulfate, vanadyl is further oxidized into vanadate (Fig. 2 is the chemical structure of vanadate) which competitively inhibits the protein tyrosine phosphatase 1B (PTP1B). PTP1B negatively regulates insulin function by dephosphorylating and inactivating the insulin receptor (Fig. 3 schematically shows how vanadate inhibition of PTP1B stimulates the function of insulin leading to glucose influx into the cell via the glucose transporter GLUT4). During the PTP1B catalyzed dephosphorylation reaction the catalytic cysteine residue’s sulfur atom nucleophilically attacks the phosphate group from the insulin receptor’s phosphotyrosine residue leading to dephosphorylation (Fig. 4 shows the dephosphorylation reaction catalyzed by PTP1B). The transition state during this reaction is a pentacovalent intermediate (Fig. 5 depicts the pentacovalent transition state intermediate) that is structurally mimicked by vanadate (Fig 6 illustrates vanadate bound in the active site of PTP1B near the catalytic cysteine residue). This makes vanadate an excellent competitive inhibitor of PTP1B preventing the insulin receptor’s phosphotyrosine binding and dephosphorylation as enzyme’s like PTP1B preferably bind the transition state in order to lower the free energy of activation promoting catalysis. Consequently vanadate inhibition of PTP1B enhances insulin activity increasing the conversion of glucose into energy (ATP) for greater athletic performance.

The best part of waking UP... is 1,3,7-trimethylxanthine in your CUP.

Caffeine, 1,3,7-trimethylxanthine, (Fig. 1 is the chemical structure of caffeine) is the active ingredient in coffee that stimulates the central nervous system (CNS) impeding drowsiness and restoring alertness. For the athlete, however, it can be used to boost exercise capacity by several mechanisms described below.
Caffeine’s positive affect on exercise capacity comes from its ability to increase the cellular concentration of the molecule cyclic AMP (cAMP). This is achieved by caffeine competitively inhibiting the enzyme that degrades cAMP, the cAMP phosphodiesterase (PDE) (Fig. 2 shows the reaction catalyzed by PDE). PDE converts cAMP into AMP by hydrolyzing the cyclic phosphodiester bond by nucleophilic attack from a zinc coordinated water molecule (Fig. 3 highlights the zinc atom, purple sphere, coordinated to the hydrolyzing water molecule, red sphere, just before nucleophilic attack of the phosphodiester bond). Caffeine is capable of competitively inhibiting PDE because of the structural similarity between caffeine and cAMP’s adenine ring (Fig 4 is the chemical structure of cAMP, compare its adenine ring structure to caffeine’s structure in Fig. 1). The caffeine molecule binds the PDE active site preventing cAMP binding and catalysis (Fig. 3 shows the cAMP molecule in purple with a superpositioned caffeine molecule in magenta highlighting their similar ring structures). Increased cAMP levels activate the protein kinase a (PKA) which subsequently activates specific enzymes involved in lipolysis and glycolysis producing energy from fatty acids and glucose, respectively. These processes increase the capacity to exercise because of the increased energy (ATP) supplied to working muscles.

The molecular based diet

Human health is enhanced by physical exercise (see previous post) and an appropriate diet. The ideal contemporary diet includes adequate intake of low-calorie foods full of essential vitamins and minerals. These essential vitamins and minerals perform an important role in specific biochemical reactions and pathways. Moreover, a low calorie regimen enhances wellbeing by lowering blood sugar, triglycerides, and cholesterol levels as well as reducing body fat. Unfortunately, the modern human diet is calorically dense contributing to numerous ailments. However, a diet supplemented with novel compounds that reverse the harmful effects of high-caloric diets could help alleviate this emerging epidemic. The recent discovery of the molecule resveratrol supports this proposal (Fig. 1 is the chemical structure of the polyphenol resveratrol). Resveratrol is found in the skin of red grapes and red wine and promotes wellbeing by mimicking caloric restriction. Some of reseveratrol’s health benefits include anti-cancer and anti-diabetic effects in humans in addition to life span extension in budding yeast. Resveratrol mimics caloric restriction by binding and activating the protein deacetylase SIRT1. SIRT1 is an energy-sensing molecule normally activated by an increase in NAD+ concentration which can be brought about by low caloric intake. NAD+ is the required cofactor in the SIRT1 catalyzed deacetylation reaction (Fig. 2 shows the proposed SIRT1 reaction scheme). Upon activation by resveratrol or caloric restriction, SIRT1 deacetylates many targets including DNA packaging histones. SIRT1 deacetylates lysine residues at the N-terminus of histone molecules. Histone deacetylation reinstates the positive charge on these lysine residues strengthening histone interaction with the negatively charged phosphate backbone of DNA. This interaction encourages high-affinity binding between histones and DNA (Fig. 3a illustrates lysines positive charge and Fig. 3b reveals the proximity of two lysine residues from the histone tail to the phosphate groups from DNA). The increased DNA binding condenses the DNA structure and prevents transcription (Fig. 4 depicts how deacetylation produces histone conformational changes consequently silencing gene transcription). For example, SIRT1 activation turns off genes that inhibit gluconeogenesis leading to the production of glucose for energy production during prolonged periods of caloric restriction. In addition, SIRT1 also indirectly activates the AMP-activated protein kinase (see previous post) by deacetylating and turning on LKB1, which phosphorylates and activates AMPK. Triggering AMPK produces energy (ATP) by stimulating glycolysis, fatty acid oxidation, and inhibiting ATP consuming pathways. This energy production is vital during times of depleted food intake and contributes to the healthier metabolic profile characteristically obtained from caloric limitation. Interestingly, AMPK activation has also been shown to activate SIRT1 in a positive-feedback loop. This intimate relationship intensifies SIRT1 and AMPK function creating a synergistic effect and a more robust response to metabolic stress. This partnership explains why low caloric diets rich in beneficial compounds, like resveratrol, combined with plenty of exercise are so advantageous to human health.
SIRT1 represents an attractive pharmaceutical target. Still, the molecular details regarding SIRT1-resveratrol interaction have not been uncovered despite several years of work by many groups including work done by my colleagues and I at Columbia University. Interestingly residues 183-225 of SIRT1 are essential for resveratrol interaction. Due to the dearth of structural data and for some fun, I generated a molecular model of the resveratrol-binding region by threading SIRT1 residues 183-225 onto an acceptable protein fold. Then I performed resveratrol docking experiments with this newly generated SIRT1 molecular model. Intriguingly, resveratrol docks nicely into a large hydrophobic pocket within the molecular model of SIRT1 possibly representing the resveratrol-SIRT1 binary complex (Fig. 5 depicting the putative resveratrol binding pocket on the SIRT1 molecular model surface colored in grey). Structural information like this could be a starting point for rational drug design (see post from 7-26-10) to enhance resveratrol’s beneficial properties and potentially generate treatment for a number of metabolic disorders. Finally, I believe further investigation will provide greater insight into the intricate biology of energy balance and impart a deeper understanding of how a molecular-based diet and rigorous exercise can mitigate human disease.

The evolutionary biology of exercise and man

One of the most fascinating wonders in science is how exercise positively affects the human body, from lowering cholesterol levels to increasing lean body mass. Exercise has an extremely advantageous influence on our health principally due to the evolutionary history of man. Early humans (~100,000 years ago) were hunter-gatherers and used a great deal of energy pursuing their food. Finite hunting resources produced extended periods of time between feeding putting selective pressure on early human genomes to become accustomed to a physically active environment with periods of food shortage. This lead to the thrifty regulation of energy use where physical activity for hunting took precedence over many other energy consuming phenomenon. Moreover, the moment energy was consumed it would be used to recharge the energy deficit caused by physical activity with the remainder immediately stored for the imminent food shortage. As a result, naturally selected genes expressed protein molecules that quickly manufactured energy to recharge the expended energy from the hunt and adeptly store the rest of the captured energy for the approaching fasting period.
Unfortunately, due to the slow rate of evolution, modern human has not changed much genetically relative to early humans yet the modern human environment has changed significantly by becoming a lot more inactive. Consequently the modern human genome, which has been profoundly shaped by early human's environment to be energetically frugal, is poorly acclimated for our current sedentary environment laden with copious amounts food. A lack of activity contributes to caloric surplus, yet more importantly, genes triggered by physical exertion are not expressed creating pathological conditions.
The AMP-activated protein kinase (AMPK) is a perfect example of a thrifty gene dysregulated by physical inactivity that contributes to disease. AMPK functions as the central regulator of energy homeostasis. During energy expenditure/exercise a rising AMP:ATP ratio, signifying low energy, activates AMPK, which shuts off less important energy-consuming processes and stimulates essential energy-producing processes (Fig. 1 shows the regulatory functions of AMPK throughout the human body - green arrows indicate stimulated processes and red bars indicate inhibited processes) recharging the energy state and promoting energy storage. Dysregulation of AMPK’s parsimonious function by inactivity can lead to a host of health problems including obesity, diabetes, and cancer. For example, when AMPK is turned on by physical exertion it causes an increase in fatty acid oxidation producing ATP and inhibits the less important ATP consuming fatty acid biosynthetic pathway. Both of these AMPK functions reduce overall fatty acid concentration reducing the propensity for obesity. Moreover, AMPK stimulates glucose transporter translocation to the cell membrane promoting profound glucose influx into the cell. Initially imported glucose is converted to ATP to compensate the energy debt and the rest of the glucose is stored as glycogen preparing for food shortage. Both of these functions reduce the amount of circulating glucose in the blood inhibiting the progression of diabetes. Furthermore, active AMPK inhibits energetically costly cell division by functioning as an energy checkpoint in the cell division cycle. Improved regulation of cell division by AMPK is antithetical to the unrestrained division of cancer cells and proliferation of cancer. All of these AMPK functions highlight the absolute necessity for modern humans to live a physically active lifestyle in order to prevent many of the ailments plaguing us today.
The molecular details regarding AMPK have increased greatly in recent years due to X-ray crystallographic studies of AMPK, including work done by my colleagues and I at Columbia University. Most AMPKs are heterotrimeric enzymes, consisting of one catalytic subunit and two regulatory subunits. The protein kinase domain, the auto-inhibitory domain, and the regulatory sequence (RS) are located in the catalytic alpha subunit. The regulatory gamma subunit binds AMP or ATP controlling AMPK activity. AMPK functions as a molecular switch changing conformation depending on cellular energy status. During low energy AMP concentrations are high therefore AMP predominantly binds the gamma subunit. The gamma subunit interacts with the RS and pulls the auto-inhibitory domain away from the kinase domain turning on the kinase function. After activated AMPK recharges the energy status and ATP concentration increases, ATP predominantly binds the gamma subunit which releases the RS. The proximity of super-positioned ATP to the RS in our AMPK crystal structure (Fig. 2 highlights the distance of ~7 Angstrom from the RS to the gamma-subunit bound ATP) suggests ATP’s greater negative charge triggers release of the RS. Subsequently the RS interacts with the alpha subunit where the auto-inhibitory domain can impede kinase activity (Fig. 3 represents a model of AMPK function dependent on energy status). Our structure reveals that part of the regulatory sequence is sequestered by the gamma subunit indicating that our structure represents AMPK in its activated conformation (Fig. 4 is a ribbon diagram representation of the crystal structure of AMPK in the active conformation bound to AMP).
As their are no pharmaceuticals directly activating AMPK, this structural information could be used to rationally design a drug that activates AMPK and fights diseases such as diabetes. AMPK binds two molecules of ATP where the negatively charged beta and gamma phosphate groups from ATP putatively trigger inhibition. Our crystal structure with two superpositioned ATP molecules reveals the two gamma phosphate groups from each ATP are ~ 5 Å apart (Fig. 5 illustrates the proximity of the two gamma phosphate groups from the two superpositioned ATP molecules, drawn in all-bonds, onto AMPK's molecular surface colored grey). One possibility for a potential lead compound could be two ATP-like molecules ligated together via the gamma phosphate moiety with the beta and gamma phosphates negative charge mitigated to support AMPK activation. The lead compound should have a higher affinity for AMPK than ATP therefore activating AMPK.

The ergogenic effects of creatine monohydrate

Ergogenic compounds improve physical performance in many different ways. They may enhance physiological capacity, augment the recovery rate from training and competition, or remove psychological barriers. Thus far the biochemical underpinnings of many ergogenic aids are tremendously misunderstood. This creates an environment where many ergogenic compounds possess unproven qualities potentially leading to misuse.
Creatine is somewhat of an exception to that rule. It has been heavily examined by research scientists yet some of its functional details are still unclear. What is known about creatine is that after human consumption during times of energy surplus, creatine will be converted to creatine phosphate (Fig. 1 is the chemical structure of creatine phosphate) by the enzyme creatine kinase. Creatine phosphate functions by donating its phosphate group to ADP replenishing ATP concentration and extending the possible duration of muscular contraction (Fig. 2 illustrates the ATP regeneration cycle utilizing the phosphate group from creatine phosphate [CP]). For that reason, athletes consume a copious amount of creatine, which is converted to creatine phosphate and enhances muscular contraction and strength.
In addition to creatine phosphate’s well established impact as a phosphate donor, creatine may confer additional ergogenic effects during energy-depleted periods by putatively modulating the energy-sensing AMP-activated protein kinase (AMPK)(see post from 9-6-10). Creatine, signifying low energy, might activate AMPK. Creatine activation of AMPK would enhance physical performance by encouraging cellular influx of glucose increasing ATP production which facilitates muscle contraction and hypertrophy.
The molecular features of creatine/creatine phosphate’s interaction with AMPK are unknown. Interestingly, our AMPK crystal structure reveals a large positively charged surface near the ligand binding sight (Fig. 3 shows the AMPK electrostatic surface potential with blue and red representing electropositive [+4 kbT] and electronegative [- 4 kbT] regions, respectively. The putative creatine phosphate binding site is highlighted by a red asterisk)., suggesting a possible interaction with the negatively charged creatine phosphate and a possible mode of regulation similar to ATP.
In conclusion, greater molecular insight into creatine's function during high and low energy conditions will maximize benefit by revealing the most efficacious prescription for creatine consumption. For instance, given that creatine purportedly functions during high and low energy states, perhaps creatine consumption should occur before, during, and immediately after exercise.

On the function of anabolic steroids

Anabolic steroids are synthetic derivatives of testosterone that display androgenic and anabolic effects in a tissue dependent manner. The primary anabolic effect stems from increased protein synthesis and decreased protein catabolism within muscle tissue accelerating muscle growth and strength. The major androgenic properties also come from protein synthesis regulation yet primarily within the gonads and brain. This regulation leads to the development of characteristics such as facial hair and increased aggressiveness. Furthermore, anabolic steroids can also mimic the function of the female hormone estrogen promoting some female secondary characteristics such as breast development.
The hydrophobic anabolic steroid molecule (Fig. 1 depicts the hydrophobic nature of the anabolic steroid oxymetholone) is permeable to cell membranes. This permeability facilitates anabolic steroid penetration across the cell membrane where it binds to the androgen receptor located in the cytoplasm of that cell. Anabolic steroid binding (Fig. 2 depicts the androgen receptor ligand binding domain in ribbon diagram with red alpha helices, blue beta strands, and green loops bound to testosterone colored white emphasized by yellow arrow) changes the androgen receptor conformation causing the dissociation of bound heat shock proteins, homodimerization, and exposure of its nuclear localization signal. All of this causes the anabolic steroid-receptor complex to diffuse into the nucleus. In the nucleus, the androgen receptor regulates gene expression by binding to specific DNA sequences known as promoters and interacting with other proteins in the nucleus that either activate or inactivate gene transcription (Fig. 3 demonstrates the functional scheme for anabolic steroids on the androgen receptor). Activation of transcription results in increased synthesis of messenger RNA which is translated by ribosomes into certain proteins. Two target genes activated by the androgen receptor are the muscle protein actin which is essential for muscle growth, and the insulin-like growth factor I (IGF-1) which stimulates skeletal muscle growth.
Because anabolic steroids possess desirable anabolic properties along with unwanted androgenic and estrogen-like qualities, designing anabolic steroids with greater anabolism and fewer androgenic and estrogenic properties is of great interest to the medical community. In order to augment these properties, rational drug design (see post 7-26-10) has been used to modify the anabolic steroid’s chemical structure enhancing the anabolic-steroid androgen-receptor interaction and minimizing anabolic steroid binding to the estrogen receptor. The design strategy for the anabolic steroid tetrahydrogestrinone (THG) lucidly demonstrates the structure-activity relationship where even minor modifications in the structure have a great impact on the receptor interaction. THG was chemically modified with two diethyl groups added to the 13th and 17th carbon positions within the steroid molecule (Fig. 4 is the chemical structure of THG highlightling the two key diethyl groups with asterisks). These modifications establish greater van der Waals interactions with a distance of ~4 angstroms between THG and the androgen receptor (Fig. 5 shows the androgen receptor bound to testosterone in cyan with a superpositioned THG in magenta highlighting with red asterisks the additional van der Waal contacts between THG and the androgen receptor) and generate steric clashes between THG and the estrogen receptor with a distance of less than 1 angstrom abrogating THG binding (Fig 6 shows the estrogen receptor bound to estrone in cyan with a superpositioned THG molecule in magenta highlighting with red asterisks the steric clashes between THG and the estrogen receptor). Up to this point a diverse group of anabolic steroids have been synthesized with similarly optimized characteristics. However, no anabolic steroid has eliminated all androgenic and estrogen-like effects because of the inseparable nature of anabolic and androgenic function combined with the inability to completely eliminate estrogen receptor interaction and activation.

The ambiguous anabolism of androstenedione

Androstenedione is the substrate in the final biosynthetic step producing testosterone as well as certain estrogens like estrone (Fig. 1 depicts the biosynthetic steps converting androstenedione to testosterone and estrone). In theory, androstenedione consumption boosts testosterone levels because an equilibrated enzymatic system experiencing an increased substrate concentration counteracts this change by rapidly generating product and restoring equilibrium. The conversion of androstenedione to testosterone is catalyzed by the enzyme 17β-hydroxysteroid dehydrogenase (17βHSD). Enzymes, like 17βHSD, catalyze the conversion of the substrate (androstenedione) to the product (testosterone) by increasing the rate of the chemical reaction (lowering the free energy of activation or delta G++) while maintaining the reaction’s equilibrium (reduction in overall free energy or delta Gr) (Fig. 2 illustrates the reaction coordinate of androstenedione to testosterone). For example, if the consumption of 10 androstenedione molecules led to an equilibrated ratio of 9 testosterone molecules to 1 androstenedione molecule then ingesting more substrate would perturb equilibrium and generate more product. For instance, the intake of 100 molecules of androstenedione would produce 90 more molecules of testosterone at equilibrium.
Although androstenedione intake should hypothetically increase testosterone levels and lead to greater muscle mass, different studies show varying results. Some studies show ingesting androstenedione increases testosterone levels whereas other data reveal no significant influence on testosterone level at all. More importantly, no investigation has ever shown an increase in muscle size or strength. Furthermore, as previously mentioned androstenedione converts to estrone and certain studies have indicated increased estrone levels could present adverse side effects especially for men. The end result clearly suggests that androstendione doesn’t promote muscle size or strength and may be harmful to your health.

Just say NO to muscle growth!

The signaling molecule nitric oxide (NO), made up of one nitrogen atom and one oxygen atom, has become a popular compound in the weightlifter’s arsenal for increased muscle growth. NO is biosynthesized from the amino acid L-arginine and oxygen by the enzyme nitric oxide synthase (NOS), therefore supplementation with the amino acid L-arginine could increase NO production. (Fig 1 shows the reaction mechanism of NOS). NO has many roles but the one athlete’s desire is increased blood flow caused by vasodilation. Vasodilation occurs when NO binds guanlyl cyclase’s heme moiety triggering conformational changes within the heme and protein that generate a fully active guanlyl cyclase (Fig. 2 illustrates the guanlyl cyclase heme group, highlighted by the green arrow, bound to NO, depicted by the blue arrow, with the heme’s axial ligand histidine, emphasized by the red arrow, just before the histidine residue dissociates from the heme initiating further conformational change of the heme group and protein molecule). The activated guanlyl cyclase catalyzes the conversion of GTP to cyclic GMP. Cyclic GMP then activates protein kinase G (PKG), which subsequently leads to myosin light chain dephosphorylation, smooth muscle relaxation, and vasodilation within the arterial wall (Fig. 3 depicts NO’s role during arterial vasodilation) generating increased blood flow. Greater blood flow brings more essential nutrients and oxygen to laboring muscles stimulating recovery and growth of the muscle tissue.
Nitric oxide is a highly reactive and transient signaling molecule. For NO to function non-locally in the human body it is transported via protein molecules. Intriguingly during physical exertion, as vasodilation increases blood flow to working muscles it increases muscle concentration of red blood cells and hemoglobin, the protein abundantly found in red blood cells. This hemoglobin may be post-translationally modified with NO attached to the sulfur atom from the amino acid cysteine (Fig. 4 shows the electron density map around the NO modified cysteine residue in hemoglobin), essentially increasing the local arterial concentration of NO after it dissociates from the hemoglobin molecule proliferating vasodilation. Furthermore, NO modified hemoglobin has an altered affinity for oxygen and carbon dioxide producing a more efficient exchange of oxygen for carbon dioxide in hypoxic working muscles. This NO modification generates additional vasodilation and superior oxygen replenishment promoting all of the benefits of increased blood flow for working muscles and the avid weightlifter.

Rational drug design

Drugs play an integral role in the battle against the many diseases afflicting humans. In order to fight this battle drugs are designed to enhance their disease-fighting properties. There are two major types of drug design. The first is referred to as structure-based drug design and the second is ligand-based drug design. Structure-based drug design relies on knowledge of the three-dimensional structure of the protein molecule target usually obtained by X-ray crystallography which enhances the ability to create novel drugs that combat disease. The detailed knowledge of the protein structure serves as a blueprint for the design of a lead compound. The lead compound is designed atom by atom optimizing both shape and charge complementarity with the active site of the protein target to enhance their interaction and suppress protein function (Fig. 1 illustrates the structural similarity between the lead compound and the enzyme’s active site). After the lead compound has been synthesized, scientists then use X-ray crystallography to analyze the structure of the protein target bound to the lead compound (Fig. 2 demonstrates the binary complex of a lead compound bound to the enzyme’s active site). The binary complex demonstrates how the compound binds the active site of the target protein. Using this structural information, redesigned lead compounds are then synthesized and further refined and analyzed in an iterative process until a sufficiently potent compound has been designed and optimized. Ligand-based drug design depends on information from all ligands that bind to the protein target. These ligands are utilized to generate a molecular framework describing all the critical elements responsible for ligand interaction with the protein target (Fig. 3 demonstrates all the ligand features essential for protein target interaction, such as the H-bond acceptor region). Incorporation of these elements into one ligand should augment protein-ligand interaction. Additionally, a protein target model is generated based on the composite ligand. This protein model is used to design additional ligands with features that improve interaction with this model. The classic target for rational drug design is the protein molecule functioning as an enzyme. Enzymes catalyze biochemical reactions by lowering the energy barrier from substrate to product. Occasionally a malfunctioning enzyme can cause disease. Rational drug design aims to create an extremely selective compound that will only bind to the active site of this malfunctioning enzyme, thereby preventing the defective enzyme’s function and ultimately halting progression of the disease.

Solving the phase problem.

In order to solve the structure of a protein molecule using X-ray crystallography the intensity and phase, also known as the Structure Factor (Fp), of each reflection must be determined. Unfortunately, only the intensity information can be directly measured in a crystallographic experiment while the phase information must be indirectly calculated. Typically the phase information is acquired in a "divide and conquer” approach where the substructure of “anomalous atoms” within the protein molecule is deciphered leading to the estimated protein reflection phases. In the standard approach today, the “anomalous atom” Selenium replaces the sulfur atom in methionine generating selenomethionine. The selenomethionine is incorporated into the protein molecule by normal protein translation and crystallized. At a specific X-ray wavelength, the crystallized protein modified with “anomalous atoms” alters the reflections relative to the unmodified protein reflections due to the absorption and delayed emission of X-rays by the “anomalous atoms” thus providing the information necessary to elucidate the “anomalous atom” substructure. The traditional approach for solving the “anomalous atom” substructure is calculating a Fourier transformation using only the anomalous intensities (really it’s the amplitudes which are essentially the square root of the intensities) calculated from the simple equation fa = fpa-fp where the measured intensities from the protein atoms (fp) are subtracted from the measured intensities of the protein plus anomalous atoms (fpa). This produces a "difference Patterson map" specifically containing the interatomic vectors between all “anomalous atoms”. Due to crystallographic symmetry, some interatomic vectors are located on a particular section of the Patterson map known as the Harker section. Patterson maps use the coordinate system U,V,W calculated from the simple relationship U=X1-X2, V=Y1-Y2, and W=Z1-Z2 where, for example, X1 is the x coordinate of any atom and X2 is a symmetrically related x coordinate with similar relationships for both Y and Z. These simple equations produce Harker sections and "anomalous atom" coordinates due to crystallographic symmetry. For example, a 2-1 screw axis symmetry operation along the Y axis converts Y1 to Y2+1/2 producing the Harker section Y = ½ from the relationship V=Y1-(Y2+1/2)=1/2. Subsequently peaks on the Y = 1/2 Harker section are given the U coordinate where that same 2-1 screw axis converts X1 to a negative X2. Therefore U=X1-(-X2) equaling 2X. Therefore, measuring the U peak position on this Harker section and dividing it by 2 will give you the X coordinate of the "anomalous atom". When this is done for all three coordinates X, Y, and Z and for every "anomalous atom", the positions are converted to the anomalous Structure Factor (Fa). From this information one can calculate the estimated protein Structure Factors (Fp) by using the simple equation Fp = Fpa-Fa in combination with a Harker circle diagram. The original estimation produces significant phase ambiguity as there are two possible solutions to this simple equation Fp = Fpa-Fa (Fig. 1 shows the two possible solutions for the protein phase, highlighted by red asterisks, on the Harker circles). However, this phase ambiguity is rectified by precise “real space” modifications of the protein molecules electron density map. The density modification procedure is based on the immutable physical characteristics of protein molecules. Finally, the protein Structure Factors (Fp) are plugged into the Fourier transformation producing the protein molecule's initial electron density map. Subsequently, a model of the protein molecule is built into the original electron density map and refined by iterative cycles of model building and refinement. Protein model alterations continue until the Structure factor amplitudes calculated from the model (Fc) converge with the Structure factor amplitudes observed from the diffraction experiment (Fo). Once the phase refinement is done, the real excitement begins. The crystallographer interprets the complex structural features that influence protein molecule function.

Why we need protein crystals.

Interestingly solving the structure of a protein molecule using X-ray diffraction requires the protein molecule to be in a crystalline lattice (Fig. 1 is a representation of a crystal lattice). The main reason for this is quite simple. The structural information acquired is significantly amplified from diffracting X-rays off of millions of identically arranged protein molecules in a crystalline lattice. In order for you to get a better grasp of this fact we’ll have to explore a few of the finer points of X-ray diffraction from crystals. So here we go, X-rays are photons that behave like waves (and particles but I don’t want to even think about quantum mechanics). When X-rays are “in phase” the wave peaks and valleys line up (Fig. 2 demonstrates how the top two waves are in phase and when added produce the bottom wave with higher peaks and valleys). This is important because in an X-ray diffraction experiment when diffracted X-rays line up “in phase” (Fig. 3 demonstrates two incoming X-rays “in phase” that bounce off two electron dense planes "in phase" resulting in diffraction) they augment one another leading to a diffracted wave that has larger peaks and valleys. This diffracted X-ray can now be measured due to the amplification of its peaks and valleys from the crystalline lattice. The diffracted X-rays contain information necessary to solve the structure of the protein molecule because the diffracted X-rays produce a diffraction pattern (see Fig. 4 depicts a standard diffraction pattern) that is intimately connected to the structure of the protein molecule within the crystal. The diffraction pattern and a protein molecule are somewhat analogous to the fingerprint and a finger as the structure of both the protein and finger produce a unique diffraction pattern or fingerprint, respectively. The features of the diffraction pattern are mathematically described by Bragg’s law (see Fig. 5 – the mathematical equation for Bragg’s law) which illustrates the geometric conditions within the crystal lattice necessary for X-ray waves to be “in phase” during the diffraction experiment. The diffraction pattern from a typical X-ray diffraction experiment is essentially a manifestation of Bragg’s law.
As I mentioned in my previous post, diffraction data provides the intensity information required to solve the protein's structure but it lacks the phase information also necessary to solve the protein structure. In my next post I will describe how the protein crystallographer gathers this phase information and solves what is known as “the phase problem.”

Why we use X-rays to visualize protein molecules.

In order to visualize a very small object, such as a cell, one magnifies the object using a light microscope. When using a light microscope, light bounces off the object and the scattered light’s intensity and phase are reassembled and magnified by the objective lens (Fig. 1 is a diagram of a light microscope). However, the light microscope can only be used to visualize objects no smaller than the wavelength of light being used which is ~500 nanometers (nm). Therefore, in order to visualize individual atoms within a protein molecule, X-rays are the correct light source as their wavelength of ~0.1 nm matches the typical distance between two atoms within a protein molecule. Unfortunately, while X-rays are the correct wavelength, their diffracted intensity and phase are unable to be reassembled by an objective lense as X-rays pass right through most material. Therefore, we must substitute the objective lense with X-ray crystallography. Here we can directly measure the diffracted intensity of each reflection but must indirectly extrapolate each reflections phase. This phase and intensity information is used in a mathematical calculation, known as the Fourier Transformation, to generate an electron density map of the protein molecule that can be readily visualized on a molecular graphics terminal and used to build a molecular model of the protein.

Structural Biology is Cool

My name is Michael Rudolph. I am a scientist who investigates how the three-dimensional shape of a protein molecule affects how the protein functions. In this blog, I will write about this seemingly peculiar fact. I'll write about how we study these extremely small protein molecules using a technique called X-ray crystallography and also how we benefit from protein structure information. For example, how protein structural knowledge facilitates the drug design process. Furthermore, I will write about structural biological phenomenon related to exercise and human performance as a consequence of my research interests and my background as an athlete (I played college football at Hofstra University).
In a nutshell, protein function comes from protein form. Subsequently, when a protein molecule has an altered shape it may perform differently, potentially leading to disease or improved health. Scientists, like me, primarily study the form/function relationship of protein molecules with the hope of understanding and abrogating disease. For example, when hemoglobin (figure above is the structure of hemoglobin - red arrows point to oxygen binding sites), the protein that binds and transports the oxygen we breathe, changes shape due to a "mutation" it may loose the ability to bind oxygen. This inability to bind oxygen may lead to disease as well as premature death. Therefore, if we can figure out why the deformed protein malfunctions this may lead to a cure. Theoretically there are many ways to manage poorly functioning proteins and understanding the shape of a protein molecule plays a massive part in this task. On the other hand, a different hemoglobin "mutation" could improve oxygen binding, imparting advantage in endurance events such as the Tour de France. Perhaps the great cyclist Lance Armstrong has a strong-oxygen binding form of hemoglobin enhancing his ability to ride so effectively under low oxygen conditions (ie. high altitude).
So, this blog will focus on the molecular structure of proteins and how their shape affects their function either positively or negatively particularly related to human performance and how small molecules (ie. drug design and ergogenic compounds) affect the protein molecules involved. I hope you enjoy this expedition with me. I am sure you will because science is one of the coolest ways to interpret our world especially at the molecular level.