<%BANNER%>

Structure-Function Studies of the Carnitine/Choline Acyltransferase Family


PAGE 1

STRUCTURE-FUNCTION STUDIES OF THE CARNITINE/CHOLINE ACYLTRANSFERASE FAMILY By BRENDA DAWN PEDERSEN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Brenda Dawn Pedersen

PAGE 3

This dissertation is dedicated to my family and friends, who have supported and believed in me along the way.

PAGE 4

ACKNOWLEDGMENTS I would first like to thank my mentor, Dr. Donghai Wu. I thank him for his encouragement, for his patience in answering my questions, and for always pushing me to go farther. Without his support this dissertation would not have been possible. I would also like to thank all of my committee members: Dr. Sloan, Dr. Schulman, Dr. Prokai and Dr. Oh. Their suggestions for the direction of this project, their time and their guidance are greatly appreciated. My gratitude and thanks go to the other members of my lab who have made this research possible. I thank Yunrong Gu and Wei Lian who, very patiently, shared their technical expertise with me on numerous occasions. I sincerely appreciate all of their contributions to this project. I also thank fellow graduate student and senior lab gremlin, Dr. Tom Kukar, for his assistance and support. Ice cream is a small price to pay for your friendship. In addition, I would like to acknowledge my fellow graduate students. I hope that the friendships that have been made along the way will continue for many years to come. I will always be thankful for the continued support and enthusiasm of my parents, Erik and Lynnae Pedersen, throughout my academic career. I thank them for believing in me and constantly encouraging me. Finally, I thank my fianc, Eric Moore. I will always be grateful for his unwavering love and support for all of my endeavors. iv

PAGE 5

TABLE OF CONTENTS Page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT.........................................................................................................................x CHAPTER 1 LITERATURE REVIEW.............................................................................................1 Fatty Acid Metabolism.................................................................................................2 The Carnitine/Choline Acyltransferase Family............................................................9 Carnitine Palmitoyltransferase..............................................................................9 Carnitine Palmitoyltransferase I...................................................................12 Carnitine Palmitoyltransferase II.................................................................14 Carnitine Octanoyltransferase.............................................................................15 Carnitine Acetyltransferase.................................................................................16 Choline Acetyltransferase...................................................................................17 Regulation of L-Carnitine Acyltransferases...............................................................18 L-Carnitine/Choline Acyltransferase Disorders.........................................................19 L-Carnitine/Choline Acyltransferase Deficiencies..............................................20 L-Carnitine Acetyltransferase Increases..............................................................23 Structure-Function Studies.........................................................................................24 Specific Aims..............................................................................................................28 2 MATERIALS AND METHODS...............................................................................30 Determine Structure of Carnitine Acetyltransferase with L-Carnitine.......................31 Expression...........................................................................................................31 Extraction............................................................................................................32 Purification..........................................................................................................33 Crystallization......................................................................................................34 Mutate Specific Residues of CAT..............................................................................34 Evaluate Enzymatic Activity......................................................................................36 Evaluate Biological Activity.......................................................................................38 Structural Determination of Choline Acetyltransferase.............................................38 v

PAGE 6

3 STRUCTURE DETERMINATION OF CARNITINE ACETYLTRANSFERASE WITH L-CARNITINE................................................................................................40 Introduction.................................................................................................................40 Materials and Methods...............................................................................................43 Results and Discussion...............................................................................................44 Structure Determination and Overall Structure of the Enzyme Complex...........44 Comparison of CAT-L-Carnitine with Free CAT...............................................44 L-Carnitine Binding Site.....................................................................................46 Comparison of CAT-L-Carnitine with Mouse CAT...........................................50 Conclusions.................................................................................................................51 4 FUNCTIONAL AND BIOLOGICAL CHARACTERIZATION OF ESSENTIAL RESIDUES OF CARNITINE ACETYLTRANSFERASE........................................52 Introduction.................................................................................................................52 Materials and Methods...............................................................................................55 Results and Discussion...............................................................................................56 Functional Characterization of L-Carnitine Binding Residues...........................56 Biological Characterization of L-Carnitine Binding Residues............................61 Conclusions.................................................................................................................67 5 STRUCTURE OF CHOLINE ACETYLTRANSFERASE........................................68 Introduction.................................................................................................................68 Materials and Methods...............................................................................................72 Results and Discussion...............................................................................................72 Structure Determination and Overall Structure of the Enzyme...........................72 Comparison of CAT-L-Carnitine with Free CAT...............................................76 Active Site...........................................................................................................78 Choline Binding Site....................................................................................79 Coenzyme A Binding Site............................................................................80 Disease Associated Mutations.............................................................................83 Temperature Sensitive Mutations........................................................................86 Conclusion..................................................................................................................87 6 CONCLUSIONS........................................................................................................88 LIST OF REFERENCES...................................................................................................91 BIOGRAPHICAL SKETCH...........................................................................................101 vi

PAGE 7

LIST OF TABLES Table page 4-1 Kinetic parameters of recombinant wild type and mutant forms of carnitine acetyltransferase.......................................................................................................57 4-2 Growth of mutant yeast cells with glucose and ethanol media................................65 4-3 Enzymatic activity and growth with glucose and ethanol of Glu326 mutants...........67 5-1 Reported mutations of choline acetytlransferase.*..................................................70 5-2 Choline substrate binding sites for rat choline acetyltransferase (rChAT), human peroxisomal L-carnitine acetyltransferase (hCAT), and mouse L-carnitine acetyltransferase (mCAT)........................................................................................82 5-3 Coenzyme A substrate binding sites for rat choline acetyltransferase (rChAT), human peroxisomal L-carnitine acetyltransferase (hCAT), and mouse L-carnitine acetyltransferase (mCAT).....................................................................82 vii

PAGE 8

LIST OF FIGURES Figure page 1-1 Chemical structure of a triglyceride...........................................................................3 1-2 The release of stored fatty acids.................................................................................4 1-3 General scheme of the L-carnitine acyltransferase reaction......................................6 1-4 Long-chain fatty acid transport across the mitochondrial membrane........................7 1-5 Long-chain and branched fatty acid metabolism in the peroxisome..........................8 1-6 Plasmalemmal L-carnitine transporter, L-carnitine palmitoyltransferase I, and L-carnitine/acyl-L-carnitine translocase......................................................................14 1-7 Structure of malonyl-CoA........................................................................................18 2-1 The pQE-9 expression vector...................................................................................31 2-2 Interaction of histidine residues of 6xHis tag and the Ni-NTA agarose..................33 2-3 Overview of the QuickChange site-directed mutagenesis method..........................35 2-4 Schematic representation of the CAT assay.............................................................37 3-1 Overall structure of human peroxisomal carnitine acetyltransferase.......................43 3-2 Structural similarity and superimposition of CAT and CAT-L-carnitine................45 3-3 Ser433 positional change in free and L-carnitine bound form of CAT.....................46 3-4 Electron density map of the carnitine binding site of L-carnitine acetyltransferase.......................................................................................................47 3-5 The active site of L-carnitine acetyltransferase-L-Carnitine....................................48 3-6 The interactions between L-carnitine and L-carnitine acetyltransferase..................49 4-1 The two pathways for the metabolism of fatty acids................................................53 4-2 Sequence alignment of carnitine and choline acyltransferases................................56 viii

PAGE 9

4-3 The active site of carnitine acetyltransferase with Tyr431........................................58 4-5 Human carnitine acetyltransferase complements mutant S. cerevisiae....................64 4-5 The active site of carnitine acetyltransferase...........................................................65 4-7 The reaction mechanism of carnitine acetyltransferase...........................................66 5-1 Schematic of choline acetyltransferase reaction mechanism...................................69 5-2 Photograph of recombinant rat choline acetyltransferase crystals...........................73 5-3 X-ray diffraction pattern of rat choline acetyltransferase crystal.............................74 5-4 Overall Structure of Rat Choline Acetyltransferase.................................................75 5-5 Sequence alignment of choline acetyltransferases...................................................77 5-6 The conformation arrangement of the active site histidine......................................79 5-7 The modeled binding site of choline and CoA.........................................................81 5-8 Structural comparison of rat choline acetyltransferase (rChAT) and human peroxisomal carnitine acetyltransferase (hCAT) and mutations in human ChAT that cause myasthenic syndrome associated with episodic apnea (CMS-EA).........85 ix

PAGE 10

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy STRUCTURE-FUNCTION STUDIES OF THE CARNITINE/CHOLINE ACYLTRANSFERASE FAMILY By Brenda Dawn Pedersen August 2004 Chair: Donghai Wu Major Department: Medicinal Chemistry To further understand the process of fatty acid metabolism, the enzymes of the carnitine acyltransferase family were studied. These enzymes play essential roles in fatty acid metabolism by facilitating the transfer of activated fatty acids across intracellular membranes. To understand the mechanism of these enzymes, the structure of human peroxisomal L-carnitine acetyltransferase (CAT) in complex with its substrate, L-carnitine, was determined. The structure revealed several residues involved in substrate binding that were further investigated by an enzymatic and biological assay. The structure of CAT with L-carnitine is very similar to the structure of the free enzyme. Both are monomeric proteins with two equally sized domains that form a central active site tunnel. The most significant difference between these two structures is with residue Ser433. The chi angle of this amino acid changes from 58 to to position the hydroxyl group of the residue closer to the carboxylate group of L-carnitine, possibly to help in stabilization during the transition state. x

PAGE 11

The structure also reveals several residues (His322, Tyr431, Thr444, Arg497, Phe545, and Val548) and two water molecules (W621 and W679) involved with binding L-carnitine. Four of these residues (Tyr431, Thr444, Arg497 and Phe545) were further investigated. These studies showed that Tyr431 and Thr444 interact with the carboxylate group of L-carnitine through hydrogen bonds, that Arg497 interacts with the carboxylate group through electrostatic interactions, and that Phe545 interacts with the quaternary amine group through pi-cation interactions. Another member of the L-carnitine/choline acyltransferase family, choline acetyltransferase (ChAT), was investigated. This enzyme catalyzes the synthesis of acetycholine (a neurotransmitter that functions in memory, learning and muscle contraction). The structure of rat ChAT (rChAT) was determined, and the substrates were modeled into the active site. The structure of rChAT shares several structural features with CAT. The rChAT is also a monomeric protein, and consists of two domains that form a central active site tunnel. Besides identifying several active site residues, the structure also provides a molecular basis for the disease, congenital myasthenic syndrome. xi

PAGE 12

CHAPTER 1 LITERATURE REVIEW Obesity is a serious health concern in the United States and in other developed countries. Obesity is defined by the body mass index, which is a measure of weight in relation to height. A person with a body mass index greater than 30 is clinically defined as obese, while a person with a body mass greater than 25 is defined as overweight. According to the NIH, based on the US Census Bureau Census 2000, nearly two thirds of United States adults are overweight, including one third that are obese [1]. Surprisingly less than half of United States adults maintain a healthy weight. The prevalence of overweight and obesity in adults has steadily increased over the past 20 years among both genders, all ages, all racial and ethnic groups, and all educational levels [1]. According to the CDC, in 2002 every state reported an obesity prevalence of at least 15-19%, while 29 states reported a prevalence of 20-24%, and 3 states reported an obesity prevalence over 25%. The obesity prevalence has huge implications for public health and society. Several diseases are associated with overweight and obesity. Specifically, overweight and obesity are risk factors for diabetes, heart disease, stroke, hypertension, gallbladder disease, osteoarthritis, sleep apnea and other breathing problems, and some forms of cancer [1]. Obesity is also associated with high blood cholesterol, complications with pregnancy, menstrual irregularities, hirsutism, stress incontinence, psychological disorders and an increased risk of complications during surgery [1]. In addition, due to the rise in excess weight and obesity, health care costs have increased dramatically. The annual medical 1

PAGE 13

2 spending due to excess weight and obesity is approximately 93 billion or 9% of health care expenditures, which can be broken down into costs associated with heart disease, type 2 diabetes, osteoarthritis, hypertension, gallbladder disease, cancer, and lost productivity at work [1]. In addition to these health care costs, Americans spend $33 billion dollars per year on weight-loss products and services. Many people blame obesity on a sedentary lifestyle or poor food choices (such as over-eating, too many carbohydrates, too much saturated fat). However, research has shown that there may be a genetic influence on weight disorders [2]. One of the National Health Objectives is to reduce the prevalence of obesity among adults to less than 15%. Perhaps the most effective way to combat the problem is through further understanding of the mechanism of fatty acid metabolism [1]. Fatty Acid Metabolism Fatty acids are involved in several cellular processes. They regulate transcription [3, 4], serve as building blocks for components of cell membranes [3], serve as precursors to cell-signaling molecules [5-7], and provide a significant source of energy for heart and skeletal muscle [8, 9]. The primary sources of fatty acids for energy production are diet, synthesis, and release from cellular stores. Dietary fatty acids account for at least 30% of caloric intake of humans in industrialized countries [1]. Before these fatty acids can be oxidized for energy, they must be absorbed, transported to the appropriate organ, taken up into the cell, and then transferred to the mitochondria, where oxidation occurs. First, fatty acids from dietary sources are digested by emulsification. In this process, large aggregates of insoluble fat particles are broken down physically (by bile salts from the gall bladder), and are held in suspension. The fat particles are then hydrolyzed to monoglycerides and free fatty acids

PAGE 14

3 by pancreatic lipases. These lipids are absorbed from the gut by diffusing into the epithelial cells lining the intestinal surface. There they are transported to the endoplasmic reticulum, to be reconverted to triglycerides. The newly formed triglycerides are packaged with cholesterol, lipoproteins, and other lipids to form chylomicrons. Chylomicrons move through the lymphatic system to the bloodstream. Lipoprotein lipase disassembles the chylomicron, and releases the fatty acid for uptake into cells. In muscle cells, the fatty acid will be oxidized for energy; but in adipose tissue, the fatty acid will be re-esterified for storage as triglycerides (Figure 1-1). H2COOCH3HCOOCH3H2COOCH3 Triglyceride Figure 1-1. Chemical structure of a triglyceride Stored triglycerides (Figure 1-1) are released in response to energy demands [10, 11]. As shown in Figure 1-2, the hormone adrenalin is secreted in response to low blood glucose levels and binds to the -adrenergic receptor (a G-protein coupled cell-surface receptor of adipocytes), which activates adenylate cyclase [10]. The resulting increase in cyclic AMP levels stimulates cyclic AMP-dependent protein kinase to phosphorylate and activate hormone-sensitive lipase. This enzyme sequentially hydrolyzes ester linkages of triglycerides to diglyceride and monoglyceride, resulting in the release of three moles of

PAGE 15

4 fatty acid and one mole of glycerol [10]. The free fatty acid is released from the adipocyte into the blood, where it will bind to serum albumin and travel to tissues to serve as fuel. PKA Hormone sensitive lipase TAG DAGMAG glycerol MAG lipase Adrenalin -adrenergic receptor G-protein Adenylate cyclase + ATP cAMP + Hormone sensitive lipase ATPADP P fatty acidfatty acidfatty acid Figure 1-2. Representation of the release of stored fatty acids. PKA, protein kinase; TAG, triacylglycerol; DAG, diacylglycerol; MAG, monoacylglycerol. The process of fatty acid uptake into cells is quite controversial, with experimental evidence pointing to both rapid diffusion and a protein-mediated process [12]. Several investigators have shown that fatty acids traverse cell membranes very quickly suggesting that a protein is not necessary for uptake. These studies show that the molecule uses a flip-flop mechanism to diffuse across the membrane [13, 14]. However,

PAGE 16

5 at least three different membrane associated proteins have been identified as putatively involved in trafficking fatty acids across cell membranes: the plasma membrane fatty acid binding protein (FATPPM); the fatty acid transporter protein (FATP); and the fatty acid translocase, which is homologous to the human leukocyte differentiation antigen CD36, (FAT/CD36) [12, 15]. It appears that fatty acid uptake increases with expression and over expression of these proteins [12]. Some researchers still contest that the presence of a protein mechanism does not negate the fact that a proportion of fatty acids also diffuse across the membrane [16]. Conversely it is thought that these proteins could function as fatty acid acceptors, while the actual movement of fatty acids through the membrane occurs by diffusion. Nonetheless, it is also possible that these two mechanisms co-exist as separate routes, and each contributes to the overall rate of fatty acid uptake [14]. Once fatty acids are inside the cell, they must first be activated via ligation with coenzyme A (CoA), forming a fatty acyl-CoA ester, before they are oxidized for energy production [17]. Several fatty acyl-CoA synthetases that catalyze this ATP-dependent process have been identified, and are grouped by their chain length specificity. Short-chain acyl-CoA synthetases prefer fatty acids with 2 and 3 carbon atoms as a substrate; while medium-chain acyl-CoA synthetases act on fatty acids with 4 to 12 carbon atoms; and long-chain acyl-CoA synthetases catalyze the reaction with fatty acids with 10 to 20 carbon atoms [17]. The activity of these enzymes has been observed in the mitochondrial matrix (short and medium chain) and the outer mitochondrial membrane (long chain) [17]. Long-, mediumand short-chain acyl-CoAs are oxidized in the mitochondrial matrix. Since acyl-CoAs are impermeable to membranes, they are transported into the

PAGE 17

6 matrix by an enzyme system consisting of carnitine palmitoyltransferase (CPT) I and II and an L-carnitine/acyl-L-carnitine transporter [8, 18-20]. The CPT I and II are members of the L-carnitine/choline acyltransferase enzyme family. These enzymes catalyze the reversible transfer of acyl groups between L-carnitine/choline and CoA to facilitate the transfer of fatty acids across intracellular membranes (Figure 1-3) [18]. +CoA SOR +CoA SHON+OHCH3CH3CH3O ON+OCH3CH3CH3ORO L-CarnitineAcyl-CoAAcyl-L-CarnitineCoA-Figure 1-3. General scheme of the L-carnitine acyltransferase reaction. R represents the carbon chain that consists of 1-21 carbon atoms. This reaction commits the activated fatty acid to oxidation in the mitochondrial matrix (Figure 1-4) [8]. The CPT I is an integral membrane protein with two transmembrane domains in the outer mitochondrial membrane, and its catalytic domain facing the cytosol [18]. The formed acyl-L-carnitine is then transported across the mitochondrial membrane by the L-carnitine/acyl-L-carnitine transporter. The transporter is an integral membrane protein that uses a ping-pong mechanism to exchange L-carnitine for acyl-L-carnitine [8]. Once inside the mitochondrial matrix, CPT II catalyzes the transfer of the acyl group from acyl-L-carnitine to CoA to form acyl-CoA, which can then enter the -oxidation spiral (Figure 1-4). The CPT II is anchored to the inner membrane as a peripheral membrane protein. The -oxidation occurs through the sequential removal of 2 carbon units at the -carbon position of the fatty acyl-CoA molecule. With each round of oxidation, 1 mole

PAGE 18

7 Outer MembraneInner MembraneAc y l CoA CPT I Ac y l L-Carnitine Ac y l L-Carnitine CPT II Ac y l CoA CACT -oxidation The Citric Acid C y cle Figure 1-4. Long-chain fatty acid transport across the mitochondrial membrane. CoA, coenzyme A; CPT I, L-carnitine palmitoyltransferase I; CACT, L-carnitine/acyl-L-carnitine transporter; CPT II, L-carnitine palmitoyltransferase II. of NADH, 1 mole of FADH2, and 1 mole of acetyl-CoA are produced. The acetyl-CoA then enters the citric acid cycle, where it is further oxidized to CO2; and 3 moles of NADH, 1 mole of FADH2, and 1 mole of ATP are generated. The NADH and FADH2 formed from -oxidation and the citric acid cycle can proceed to the respiratory pathway for the production of ATP. In contrast, very long-chain acyl-CoAs (with 24 or more carbon atoms) and branched fatty acyl-CoAs (such as pristanic acid) are poor substrates for the mitochondrial system, and are therefore oxidized in peroxisomes (Figure 1-5) [21, 22]. Transport of these fatty acids across the peroxisomal membrane appears to differ from transport of acyl-CoAs across the mitochondrial membrane in that the carnitine system is

PAGE 19

8 not necessary [21]. It is suspected that a transporter protein is involved; however the mechanism of this arrangement is unknown. Once inside the peroxisome, the very long-chain and branched acyl-CoAs are oxidized using the same mechanism that occurs in the mitochondria: dehydrogenation, hydration, dehydrogenation again, and thiolytic cleavage [21, 23]. However, there are differences between the two systems. The acetyl-CoA that is produced from -oxidation in the peroxisomes cannot be degraded into CO2 and water, because there is no functional Krebs cycle in the peroxisome. In addition fatty acids are not fully oxidized to acetyl-CoA units in the peroxisome: they are only chain-shortened [23]. For complete oxidation, the chain-shortened acyl-CoAs are transferred to the mitochondria. Peroxisome Very Long-Chain Fatty Acids Branched Fatty Acids Fatty Acid -oxidation Acetyl-CoA + Fatty Acid CAT Acetyl-L-Carnitine COT Acyl-L-Carnitine Figure 1-5. Long-chain and branched fatty acid metabolism in the peroxisome. CAT, L-carnitine acetyltransferase; COT, L-carnitine octanoyltransferase. However, these molecules cannot freely diffuse out of the peroxisome or into the mitochondria: they must use the carnitine system. A peroxisomal carnitine octanoyltransferase (COT) (another member of the L-carnitine/choline acyltransferase

PAGE 20

9 family) transfers medium-chain acyl groups (with 6 to 22 carbon atoms) from acyl-CoA to L-carnitine to form acyl-L-carnitine. Another member of the L-carnitine/choline acyltransferase family, carnitine acetyltransferase (CAT), catalyzes the same reaction with short-acyl groups (2 to 6 carbon atoms) [18, 22]. The acyl-L-carnitine can then diffuse across the peroxisomal and mitochondrial membranes. After the acyl-L-carnitine is in the mitochondrial matrix, CPT II transfers medium-chain acyl groups from L-carnitine back to CoA to form acyl-CoA; while a mitochondrial form of CAT catalyzes the reverse reaction, and transfers short-acyl groups from L-carnitine back to CoA to form acyl-CoA. Acyl-CoAs with 2 carbon atoms will enter the citric acid cycle directly while longer acyl-CoAs will be further -oxidized before they enter the citric acid cycle. The Carnitine/Choline Acyltransferase Family As discussed, the enzymes of the carnitine acyltransferase family play vital roles in fatty acid oxidation by facilitating in the transfer of acyl groups across intracellular membranes. The molecular genetics of these enzymes is known and is presented here. Carnitine Palmitoyltransferase As previously discussed, activated long-chain fatty acids are transported across the mitochondrial membrane by an enzyme system consisting of CPT I, CPT II, and an L-carnitine/acyl-L-carnitine translocase. This enzymatic arrangement was first conceptualized in the 1960s by Fritz and Yue [24] and Bremer [25]. Since then, the general concept of the system has undergone several refinements. In 1955, it was shown that L-carnitine is an active component of muscle extracts, and that it stimulates long-chain fatty acid oxidation in liver perfusions and skeletal and heart muscle fractions. Then it was shown that, in the presence of L-carnitine, CoA also stimulates fatty acid oxidation [26]. Further studies showed that L-carnitine was

PAGE 21

10 acetylated, and fatty acid esters of L-carnitine were isolated. It was thought that L-carnitine might react with long-chain fatty acids to form acyl-L-carnitine derivatives, which could then be converted to the corresponding acyl-CoA compounds [24]. Experimental evidence demonstrated that a transferase in heart muscle catalzyed the transfer of palmitoyl from palmitoyl-L-carnitine to CoA, to form palmitoyl-CoA [24]. It was then thought that L-carnitine may enhance fatty acid oxidation via acyl-L-carnitine formation, and that palmitoyl-L-carnitine may function as a carrier through membranes [24]. The presence of CPT on both sides of the membrane was essential to this hypothesis. Experimental data confirmed this hypothesis, establishing that formation of palmitoyl-L-carnitine and de-acetylation of palmitoyl-L-carnitine could occur on either side of the mitochondria [25]. Further functional studies with isolated mitochondria demonstrated that the membrane contains two pools of CPT: an external (overt) (CPT I) and an internal (latent) (CPT II) [26]. After the identification of CPT, several experiments were performed to characterize the system. Inhibition studies with the enzyme determined that CPT I is inhibited by malonyl-CoA (reversibly), tetradecylglycidyl-CoA and etomoxir-CoA (irreverisbly), and 2-bromopalmitoyl-CoA (essentially irreversibly) [27]. Interestingly, it was found that detergent-solubilized CPT is insensitive to all inhibitors, while it is still enzymatically active. This form of CPT is also stable, and migrates on sodium dodecylsulphate (SDS) gels as a single protein with a molecular weight of 68-70 kDa [27]. Based on these findings, a model of the CPT system was developed. In this model, CPT I and II are the same protein. The CPT I is located in the outer membrane, and is associated with a regulatory protein to which the inhibitors bind and cause allosteric inactivation of the

PAGE 22

11 enzyme. Detergents can then disrupt the interaction between the catalytic and regulatory components of CPT I, which allows the co-purification of CPT I and II as single entities [27]. Several experiments provided additional support for this model. In particular, Bergstrom and Reitz [28] found that when extracted from the mitochondrial membrane, CPT I and II had very similar properties (such as the Km value toward acyl-CoA, L-carnitine, acyl-L-carnitine, and CoASH; and the ratio of Vforward/Vreverse). Additionally, antibodies raised against CPT I inhibited CPT II and vice versa, suggesting immunoidentity of the two enzymes [28]. However, several experiments also raised questions about the validity of the model. The existence of a regulatory component came into question. Previous studies had shown that 2-bromopalmitoyl-CoA inhibited CPT I, and formed a ternary complex at the active site of the enzyme. Thereafter it was demonstrated that all of the inhibitors of CPT I (malonyl-CoA, tetradecylglycidyl-CoA, etomoxir-CoA, and 2-bromopalmitoyl-CoA) interacted at a common locus, suggesting that the other inhibitors must also act at the active site and not at a regulatory component [27]. The theory that CPT I and II are the same protein was also in question. Radioactive tetradecylglycidyl-CoA was found to label a single protein in rat liver mitochondria with a molecular weight of 90-94 kDa, while the labeled protein in rat skeletal muscle was only 86 kDa. Interestingly, CPT from both tissues migrated to the same position on SDS gels (70 kDa), and reacted equally well with an antibody raised against the liver form of the enzyme [27]. An alternative model to explain these observations was proposed. In this model CPT I and II are different proteins. Inhibitors bind to the catalytic center of CPT I, and do

PAGE 23

12 not affect the activity of CPT II due to the lack of necessary binding sites [27, 29]. In support of this model, irradiation and inactivation experiments concluded there were two catalytic entities of different sizes: 83 and 70 kDa [27]. Furthermore, a Western Blot from the mitochondrial membrane (containing CPT I and II) and a mild detergent (Tween-20) extracted membrane (rich in CPT I, devoid of CPT II) demonstrated a strong signal (with an antibody raised toward CPT II) in the normal membrane, and a lack of signal in the detergent extracted membrane [27]. Another facet of the model is that CPT I is detergent labile, while CPT II is not [27, 29]. A mild detergent (such as Tween 20) separated the mitochondrial membrane extract into pellet and supernatant fractions; both express CPT activity. The pellet fraction was inhibited by etomoxir-CoA, while the supernatant was insensitive to the inhibitor. Upon further treatments with stronger detergents (such as octylglucoside and Triton X-100) enzyme activity decreased. This indicates that CPT II was released from the membrane with mild detergent, while CPT I was more firmly anchored in its membrane environment. However, CPT I lost its catalytic activity upon membrane disruption, suggesting that CPT I requires a critical membrane component [27, 29]. It can now be concluded that CPT I and II are different proteins that act in concert with an L-carnitine/acyl-L-carnitine translocase to transport long-chain fatty acids across the mitochondrial membrane. Of the two enzymes, CPT I is larger and detergent labile. However, many factors of the model are still unresolved. Carnitine Palmitoyltransferase I Three forms of CPT I have been identified: a muscle (M-CPT I), a liver (L-CPT I), and a putative brain (CPT I-C) isoform. The gene encoding the muscle form of CPT I has been mapped to chromosome 22, position 22q13.31-q13.32, and encodes a protein that

PAGE 24

13 consists of 772 amino acids [30]. This isoform is the primary form in skeletal muscle, heart, adipose tissue, and testis [31, 32]. The liver form of CPT I is the predominant isoform in the liver, kidney, lung, spleen, intestine, pancreas, ovary, and brain [31, 32]. The gene encoding L-CPT I has been mapped to chromosome 11, position 11q13.1-q13.5, and encodes a protein that consists of 773 amino acids [30]. The muscle and liver isoforms are very similar: in fact, they have a 63% amino acid sequence identity [33]. The brain form of CPT I was only recently identified [34]. The gene that encodes CPT I-C has been mapped to chromosome 19, position 19q13.33, and encodes a protein that is similar to M-CPT I and L-CPT I, with an amino acid sequence identity of 53 and 55%, respectively [34]. This isoform is expressed at high levels in the brain, particularly in regions that have been shown to be involved with feeding behavior and appetite control (such as the hippocampus, paraventricular arcuate, and the superchiasmatic nuclei) [34]. However, no catalytic activity was observed with acyl-CoA esters that are substrates for the muscle and liver isoforms. This suggests that CPT I-C may participate in highly specialized lipid metabolism of the brain, and explains why the brain may require the expression of an additional member of the CPT I family [34]. The structural features of the three forms of CPT I are similar. As previously discussed, CPT I is an integral membrane protein localized to the outer membrane of the mitochondria. Topographical studies have shown that CPT I contains two transmembrane domains (amino acid residues 46-76 and 104-126), with both Nand C-terminals (amino acid residues 1-128 and 128-772/773, respectively) facing the cytosol and the linker region between the two domains (amino acid residues 77-103) jutting into the intermembrane space (Figure 1-6) [35]. The N-terminal of CPT I contains a

PAGE 25

14 mitochondrial targeting signal peptide (amino acid residues 123-147) for import into the mitochondria. CPT I is the first committed step to oxidation in the mitochondria, and therefore is a highly regulated process, which will be discussed later in this chapter. N C PM L carnitine transporter MOM N C C N L carnitine palmitoyltransferase L carnitine/acyl L carnitine translocase MIM Figure 1-6. Topographical representation of the plasmalemmal L-carnitine transporter, L-carnitine palmitoyltransferase I, and L-carnitine/acyl-L-carnitine translocase. Membranes are shown in blue, pm, plasma membrane; mom, mitochondrial outer membrane; mim, mitochondrial inner membrane. Regions spanning the membrane are shown by purple barrels. The amino-terminus (N) and carboxy-terminus (C) of each protein are shown. Adapted from information reviewed in [18]. Carnitine Palmitoyltransferase II The genes that encode the human and rat form of CPT II have been identified and mapped to chromosome 1, position 1p32, and translate to a 658 amino acid protein [36, 37]. CPT II is a mitochondrial matrix protein, and is associated with the inner mitochondrial membrane. The first 25 amino acid residues are part of a mitochondrial

PAGE 26

15 targeting signal, and are removed during import into the mitochondria [18]. In contrast to CPT I, only one form of CPT II has been identified, and it shares 50% amino acid sequence homology with L-CPT I in the major portion of the sequences, except the N-terminus [32]. It appears that CPT II is regulated by the homeostasis of the cell. In conditions of starvation and diabetes, when CPT II activity is critical for generating fuel sources, the mRNA levels (and consequently the activity of CPT II) appear to increase [18]. The CPT II gene promoter does not contain a fat activated response element, but it does respond to specific fibrate induction through peroxisome proliferator activated receptor [18]. Carnitine Octanoyltransferase Evidence for the existence of a third carnitine acyltransferase that prefers acyl groups composed of 6-22 carbon atoms emerged in 1971. This suggestion was based on studies that examined the substrate specificity profiles of muscle mitochondria of pigeon breast; and mitochondria of liver, heart, and testis of rat [38, 39]. Subsequent substrate profiles of mitochondria from beef heart showed that the enzyme preferred octanoyl-carnitine [40]. Obviously this was surprising, since the previously identified carnitine acyltransferases prefer substrates with acyl groups of 2-6 (CAT) and 8-18 (CPT) carbon atoms. The gene that encodes COT has been mapped to chromosome 7, position 7q21.1, and encodes a protein composed of 612 amino acids [18, 30]. Significant COT activity has been observed in the peroxisomes of the liver, intestine, and ovary [41]. The enzyme contains a peroxisomal targeting signal at the C-terminus; however, the signal differs in the human (Thr-His-Leu), rat (Ala-His-Leu), and bovine (Pro-His-Leu) forms [18].

PAGE 27

16 Interestingly, like CPT II, COT levels increase after induction by peroxisomal proliferators. This suggests a parallel mechanism between the two enzymes [18]. It was also found that diets heavy in fish oils induce COT expression in rats [18]. Since fish oils are composed of very long-chain acyl groups, these fatty acids must be shortened through -oxidation in the peroxisome before they can be transferred to the mitochondria for energy production, therefore demonstrating a need for COT activity. Carnitine Acetyltransferase The gene that encodes CAT has been mapped to chromosome 9, position 9q34.1, and displays an alternative splicing pattern that results in a mitochondrial and peroxisomal form of the enzyme [42, 43]. The mitochondrial form of CAT is composed of 626 amino acid residues, with a mitochondrial targeting signal at the N-terminus (residues 1-28) [18]. The peroxisomal form of CAT starts at the second codon, which is downstream from the first mitochondrial start codon, and is composed of 605 amino acid residues with a peroxisomal targeting signal at the C-terminus [18]. CAT activity has also been observed in the endoplasmic reticulum [18]. It is not known if this form of the enzyme is analogous to the mitochondrial or peroxisomal form, or if it is a separate entity. Early acyl group specificity studies with CAT isolated from pigeon breast muscle showed that the enzyme prefers substrates with 2-4 carbon atoms [44]. Further studies showed that acyl groups with 12 or more carbon atoms inhibit the enzyme by competing with acetyl-CoA [44]. However, an endogenous inhibitor of CAT activity has not been identified.

PAGE 28

17 Choline Acetyltransferase Choline Acetyltransferase (ChAT) is the fourth member of the L-carnitine/choline acyltransferase enzyme family. This enzyme catalyzes a similar reaction to that of CAT, except it transfers an acetyl group from CoA to choline to form the neurotransmitter acetylcholine (ACh). ACh is implicated in both the central and peripheral nervous systems. In the central nervous system, ACh is associated with sleep, memory, and learning. In fact, ACh appears to be the primary neurotransmitter involved with memory and learning. In the peripheral nervous system, ACh is involved in smooth muscle contraction, and helps to regulate heart rate. The gene that encodes ChAT has been mapped to position 10q11.2 [45, 46]. Human ChAT cDNA was cloned, and appears to encode two forms of the enzyme of differing lengths, with 748 and 630 amino acids. The form with 748 amino acids contains a N-terminal extension of 118 amino acids compared to the other form. It is not clear if this form is actually synthesized in vivo [47]. The enzyme ChAT is located in the nerve terminals of cholinergic nerves in the central and peripheral nervous systems. In the central nervous system, ChAT-immunoreactive neurones are detected in the medial septal nucleus; the nucleus of the diagonal band of Broca; the basal nucleus of Meynert; the caudate nucleus; the putamen; the nucleus accumbens; the pedunculopontine tegmental nucleus; the laterodorsal tegmental nucleus; the medial habenular nucleus; the parabigeminal nucleus; some cranial nerve nuclei (including oculomotor, trochlear, trigeminal, abducent, facial, prepositus, vagus, accessory, and hypoglossal); and the anterior horn of the spinal cord [47].

PAGE 29

18 The rate-limiting step of ACh synthesis is the uptake of choline into the cholinergic neuron. However, ChAT is also regulated by several factors and chemicals (such as cholinergic differentiating factor/leukemia inhibitory factor, ciliary neurotrophic factor, camp, and retinoids) [47]. Regulation of L-Carnitine Acyltransferases As previously described, CPT I is regulated by the reversible inhibitor, malonyl-CoA (Figure 1-7). Malonyl-CoA is an intermediate in fatty acid synthesis, produced during the first step in a reaction catalyzed by acetyl-CoA carboxylase (ACC). This is the rate-limiting step in fatty acid synthesis, and is regulated by adenosine monophosphate (AMP) activated protein kinase (AMPK) [48]. AMPK functions as a fuel sensor to monitor the status of the cell. Once activated, AMPK inactivates ACC by phosphorylation, resulting in decreased levels of malonyl-CoA and increased CPT I activity [48]. Interestingly, AMPK activity increases after exercise to promote fatty acid oxidation in muscle cells, to generate ATP and glucose uptake to replenish glycogen stores [48]. A decrease in AMPK activity has been observed in the hypothalmus of rats as a result of refeeding after a fast, glucose and insulin administration, or the injection of leptin [49]. Thus it appears that CPT I activity depends on the metabolic state of the organism. CoA SOOO Figure 1-7. Structure of malonyl-CoA, an inhibitor of L-carnitine palmitoyltransferase I Recently, it was shown [50] that the hormone leptin modulates fatty acid metabolism through AMPK. The 16 kDa hormone is synthesized by white adipocyte

PAGE 30

19 tissue in proportion to adipose cell size and number, and in response to stimuli [51]. Leptin plays a central role in regulating food intake, energy expenditure, and neuroendocrine function. Leptin crosses the blood-brain barrier to reach its receptors (cytokine type 1), which signal through the janus kinase signal transducer and activator of transcription (JAK-STAT) pathway [51]. The activation of these receptors leads to the decrease of food intake, which is mediated by several downstream mechanisms. Recently, it was shown that leptin activates AMPK by phosphorylation of the alpha 2 catalytic subunit of AMPK, and therefore stimulates fatty acid oxidation and glucose uptake [50]. In addition, leptin also suppresses the activity of ACC to further promote fatty acid oxidation [50]. However, fasting and sympathetic activity have been shown to modulate leptin expression and serum leptin [52]. In particular, suppression of fatty acid oxidation in vivo or in isolated adipocytes increases leptin expression, indicating that metabolites of fatty acid oxidation may regulate leptin expression [52]. The activity of CPT I could affect leptin expression. A recent experiment showed that inhibition of fatty acid oxidation via inhibition of CPT I increased leptin expression in fasting rats [52]. The mechanism of the increased expression is unclear, and needs further work to sort out. However, it is thought that CPT I is a key modulator of leptin expression in fasting rats, perhaps through a complex feedback mechanism in which suppression of fatty acid synthesis (due to inhibition of fatty acid oxidation and increase in adipocyte content) causes a decrease in related metabolites (such as malonyl-CoA) that may suppress leptin expression [52]. L-Carnitine/Choline Acyltransferase Disorders The enzymes of the L-carnitine/choline acyltransferase family play vital roles in fatty acid oxidation. These metabolic roles are so important, that a deficiency in CPT

PAGE 31

20 and/or CAT can cause serious health conditions. In fact, CPT deficiencies are among the most common causes of inherited fatty acid oxidation disorders. Interestingly, increases in CAT activity have been shown to have beneficial effects and prevent age-related neurodegeneration. L-Carnitine/Choline Acyltransferase Deficiencies The first case of CPT I deficiency was described in 1981, with a patient who presented fasting induced non-ketotic hypoglycemia and displayed an absence of fatty acid metabolites in the blood and urine [53]. Since then, approximately 30 patients have been reported worldwide. Typically the symptoms, hypoketotic hypoglycemia (usually associated with fasting or a viral illness), and hepatomegaly (with or without acute liver failure), with subsequent hypoglycemic attacks, begin from birth to 18 months of age [32]. The outcome of the disease is variable. A persistent neurological deficit is common, but not constant. Recurrent episodes are frequent up to five years of age, but are successfully treated with symptomatic therapy [32]. CPT II deficiency is more common than CPT I deficiency and can be classified according to age of onset, with adult, infant and neonatal forms. The adult form was first described in 1973, and now more than 150 patients have been reported [32]. The symptoms begin between 6 and 20 years of age, but the age of onset can also be over 50 years or as early as 4 years [32]. This disease is characterized by muscle pain and stiffness, which can be complicated by acute renal failure (as a result of myoglobinuria) and respiratory insufficiency (secondary to respiratory muscle involvement) [32, 54]. These symptoms are triggered by exercise, fasting, high-fat intake, exposure to cold, mild infection, fever, and/or emotional stress and can be treated by restriction of long-chain fat

PAGE 32

21 intake along with medium-chain triglyceride supplementation, and frequent meals with extra carbohydrate intake before and during exercise [32, 54]. The infantile form of CPT II deficiency has been reported in 10 patients and is characterized by recurrent attacks of acute liver failure with hypoketotic hypoglycemia, that can result in coma and seizures, and transient hepatomegaly [32]. In approximately half of these cases the heart muscle was also symptomatic and displayed dilated and hypertrophic cardiomyopathy or arrythmias and conduction disorders [32]. Fasting or illness also triggers these attacks. The treatment is the same as with the adult form of the disorder, but the prognosis is not as good. Of the 10 reported patients, 6 died within the first year of life due to heart beat disorders, while the other 4 patients died from Reye syndrome [32]. The neonatal form of CPT II appears to be even more severe than the infant form. To date, 13 patients have been reported and all died shortly after birth, except for one patient who recovered after resuscitation [32]. These patients had dramatic malformations including dysmorphic features, cystic renal dysplasia, and neuronal migration defects in addition to hypoketotic hypoglycemia and cardiomyopathy [32, 54]. Over 25 mutations have been identified in the gene encoding CPT II in the adult, infantile, and neonatal forms of CPT II deficiency [32]. Two of these mutations, Ser113Leu (allele frequency of 60% in European patients) and Pro50His (allele frequency 6.5%), are common in the adult form of the disease [32]. Most of the mutations occur in amino acid sequences that are highly conserved among the carnitine acyltransferase enzyme family. However, it is difficult to understand the enzymatic role of these residues and thus the molecular basis of the disease without a three-dimensional structure.

PAGE 33

22 As previously discussed, CAT deficiencies also have serious medical consequences. A decrease in CAT activity has been observed and presents clinical symptoms of progressive ataxic encephalopathy [55]. This disease is characterized by psychomotor and growth retardation and is associated with weakness, ataxia, ophthalmoplegia, and increased blood pyruvate levels [55]. In the affected patient the COT activity was slightly decreased in the brain (approximately 60%), CPT activity was slightly increased in the liver and kidney, and CAT activity was significantly decreased in all tissues examined [55]. It was suggested that the CAT defect was coupled to a deficiency in acetyl-CoA metabolism in the mitochondria of the brain and that the CAT defect affected oxidation [55]. Deficiencies in ChAT have also been observed. Mutations in the gene encoding ChAT have been shown to cause congenital myasthenic syndrome (CMS) [56]. CMS is a genetic disorder that presents in infancy or childhood and is characterized by generalized weakness and fatigability of voluntary muscles including those controlling movement, eye movement, swallowing, and breathing. These symptoms are a result of disorders at the neuromuscular junction, including defects in the synthesis and release of ACh, in the gene encoding acetylcholinesterase, and the ACh receptor. Recently ChAT has also been implicated with CMS. Sequencing of the gene encoding ChAT in 5 patients with CMS revealed 10 recessive mutations [56]. Three of these mutations resulted in reduced expression of the enzyme while kinetic studies with mutant forms of the enzyme showed decreased catalytic activities with eight mutations, and a complete loss of catalytic activity with one of the mutations [56]. The molecular basis for these mutations is unclear.

PAGE 34

23 L-Carnitine Acetyltransferase Increases Increases in CAT activity have also been observed and appear to protect against age-associated neurodegeneration. A hallmark of aging is mitochondrial decay which is characterized by high generation of oxidants, a decline in the activity of electron transport complexes, and a decrease in the fatty acid composition of cardiolipin (an essential phospholipid for normal function of the mitochondria) [57]. Recently, it was shown that acetyl-L-carnitine reverses age-associated mitochondrial decay in rats. Specifically, feeding rats acetyl-L-carnitine decreased mortality, improved nerve regeneration, protected neurons from the toxicity of mitochondrial uncouplers or inhibitors, increased cardiolipin contents, and increased the activity of cytochrome c oxidase [57]. Feeding rats a combination of acetyl-L-carnitine and lipoic acid had an even more beneficial effect. This treatment reversed age-related mitochondria decline, improved hepatocellular ascorbate levels, increased metabolism, and decreased the oxidative stress [58]. In addition, feeding rats acetyl-L-carnitine and lipoic acid also affected cognitive function. These rats showed an improved performance on memory function tasks, a decrease in brain mitochondrial structural decay, and a decrease in oxidative damage in the brain [59]. Further studies showed that acetyl-L-carnitine may protect CAT from interacting with aldehydes from lipid peroxidation while lipoic acid may prevent proteosomes from oxidative modification and play a role in reactivating CAT by increasing its substrate binding affinity [60]. However, the molecular basis of acetyl-L-carnitine and lipoic acid treatment is unclear. To further understand the (a) mechanism of fatty acid oxidation and how it is regulated (b) carnitine/choline acyltransferase disorders and (c) the benefits of acetyl-L-carnitine and lipoic acid it is of the utmost importance to study this enzyme family.

PAGE 35

24 Structure-Function Studies Enzymatic studies of the carnitine acyltransferase family began in the late 1960s with CAT. Understanding of this enzyme family has significantly increased with the advent of biological tools such as sequencing, site-directed mutagenesis, and crystallography. These studies have helped to elucidate the catalytic activity of these enzymes and to identify specific residues involved in the mechanism of catalysis, with the ultimate goal of understanding endogenous regulation and developing exogenous regulators. Early kinetic studies with CAT (isolated from pigeon breast muscle) determined that the enzyme follows a random ordered, rapid equilibrium mechanism that involves the formation of an intermediate [61]. Further studies with CAT and bromoacetyl-L-carnitine showed that there are two steps to the reaction, reversible binding and irreversible alkylation [62]. A combination of electrophoretic studies, chemical modification, and substrate analogues implicated histidine as the active site residue that is alkylated during catalysis [62]. Further inactivation studies confirmed that a histidine residue is crucial to activity and that it serves as a general acid/base catalyst [63]. In this role, the histidine residue removes a proton from L-carnitine or CoA (depending on the direction of the reaction) to promote the formation of an intermediate [18]. Recently, the x-ray crystal structures of human peroxisomal CAT as well as the free and substrate bound forms of mouse CAT have been determined [64, 65]. The overall structure reveals that CAT is a monomeric protein consisting of 17 helices and 14 strands that form two equally sized domains [64, 65]. The two domains are tightly associated with each other and form a central, active site tunnel. The active site histidine,

PAGE 36

25 His322, is located in the middle of the active site tunnel, suggesting that L-carnitine and acetyl-CoA bind to opposite sides of the tunnel [64]. Several structure function experiments have also been aimed at determining the structural basis of malonyl-CoA sensitivity of CPT I. Since this regulation affects fat metabolism, it is conceptualized as a potential drug target for type II diabetes. Interestingly, the muscle and liver isoforms of CPT I have different malonyl-CoA sensitivities, an IC50 of 0.03 M and 3 M, respectively [66]. Therefore, the primary question is what is the structural basis of the malonyl-CoA sensitivity difference between these two isoforms. In 1997, Zhu et al., developed a high level expression system for CPT I in the yeast, Pichia pastoris [67, 68]. This organism is devoid of endogenous CPT activity and therefore, is an excellent expression system. From this system, several structure function studies, including deletion mutations, site-directed mutagenesis, and chimeric protein studies, have been carried out to characterize malonyl-CoA binding. The deletion mutation studies were designed to characterize the N-terminal of CPT I. The first 124 amino acid residues are conserved in M-CPT I and L-CPT I, but not in CPT II (which is not responsive to malonyl-CoA). Therefore, it was predicted that this region is the malonyl-CoA sensitive site. Several deletions of this area were constructed to elucidate the amino acid residues involved in malonyl-CoA sensitivity. In a particular study the residues 1-18, 1-35, 1-52, 1-73, 1-83 and 1-129 in L-CPT I were deleted and the mutant forms of the enzyme were expressed, purified, and the kinetic parameters were determined and compared to the wild type form of L-CPT I [33]. Deleting the first eighteen residues of L-CPT I completely abolished malonyl-CoA inhibition and high

PAGE 37

26 affinity malonyl-CoA binding [33]. However, to achieve insensitivity to malonyl-CoA and loss of high affinity malonyl-CoA binding in M-CPT I, the residues 1-28 and 1-39 must be deleted [69]. This suggests that the first 18 amino acid residues play a different role in the muscle and liver forms of CPT I. Another interesting deletion mutation study demonstrated that, the removal of amino acid residues 3-18 and 1-80 of M-CPT I decreases malonyl-CoA sensitivity and that the deletion of residues 19-30 increases malonyl-CoA sensitivity [70]. This was the first demonstration of positive and negative determinants of malonyl-CoA sensitivity. To evaluate specific residues within this region site-directed mutagenesis studies were performed. Substitution of glutamate at position 3 by alanine or glutamine results in a complete loss of malonyl-CoA sensitivitiy and high and low affinity binding while other kinetic parameters remain intact [71, 72]. Furthermore, an aspartate at position 3 can only partially substitute for the glutamate. This suggests that a negative charge at that position is essential and that the longer side chain of glutamate is essential for optimal malonyl-CoA sensitivity [72]. In addition, substitution of serine at position 24 or glutamine at position 30 by alanine resulted in a 10-20 fold decrease in IC50 and Ki values for malonyl-CoA [72]. These two residues may affect malonyl-CoA sensitivity through interactions with glutamate 3 [72]. The involvement of the N-terminus in malonyl-CoA sensitivity was further evaluated through the use of chimeric proteins from combinations of L-CPT I and M-CPT I. These chimeric proteins were constructed from 3 segments of CPT I; (a) the cytosolic N-terminus and transmembrane domain one, (b) the amino acid residues that link the transmembrane domains together and (c) transmembrane domain two and the

PAGE 38

27 cytosolic C-terminus [73]. The mutant forms of the enzyme were expressed and purified and the kinetic parameters were determined. Replacing the first segment and the first and second segments of L-CPT I with M-CPT I altered the Km values for both substrates, L-carnitine and palmitoyl-CoA, and increased the malonyl-CoA sensitivity. However, M-CPT I was much less susceptible to influence by the alternate segments [73]. This suggests that transmembrane domains interact with each other and these interactions are important for determining the kinetic parameters for the catalytic domain of L-CPT I [73]. Despite the high sequence identity between the first two segments there is an interesting amino acid difference in transmembrane domain one that might account for the difference in malonyl-CoA sensitivity. An arginine residue is located at position fifty two in M-CPT I whereas in L-CPT I this is a threonine residue. The arginine residue might make the transmembrane domain shorter in M-CPT I than in L-CPT I and thus extend the N-terminal domain to confer a different malonyl-CoA sensitivity [73]. The charged residue may also restrict the mobility of the transmembrane helix and result in a more rigid structure of M-CPT I that would also influence malonyl-CoA sensitivity [73]. An interesting observation with the pig form of L-CPT I helped to further the chimera studies. This form of L-CPT I consists of 772 amino acids and shares an 86% amino acid sequence identity with the rat form of L-CPT I and a 62% sequence identity with the rat form of M-CPT I. Interestingly, the pig, human, and rat form of L-CPT I have similar kinetic properties with respect to Km values for L-carnitine and acyl-CoA, but the pig form of L-CPT I demonstrates a higher sensitivity to malonyl-CoA, more like the human and rat forms of M-CPT I [74]. These characteristics demonstrate that the pig

PAGE 39

28 form of L-CPT I behaves like a natural chimera between the L-CPT I and M-CPT I isoforms. To further understand the difference in malonyl-CoA sensitivity between M-CPT I and L-CPT I, chimeras between the pig and rat forms of L-CPT I were constructed. These studies demonstrated that the Km value for the substrates and the malonyl-CoA sensitivity depends on the C-terminal domain [75]. In addition, replacing different regions of the C-terminus of rat L-CPT I, 128-220, 220-590 or 590-772, with pig L-CPT I showed that the C-terminus behaves as a single unit and the degree of sensitivity is determined by the structure of this domain [75]. Additional deletion studies showed that the first 18 residues determine the malonyl-CoA sensitivity; this is most likely due to interactions with the C-terminal [75]. It was then concluded that the difference in malonyl-CoA sensitivity between pig and rat L-CPT I is due to the difference in the interaction of the first 18 residues with the C-terminal of these forms [75]. This was the first study that described how interactions between the Nand C-terminals could affect malonyl-CoA sensitivity in pig and rat L-CPT I. There are several unresolved issues with the CPT I system. These studies demonstrate the need for a three dimensional structure to resolve these issues. Specific Aims The specific aims of this project are to further understand this enzyme family at the protein level. The first specific aim is to identify and characterize amino acid residues that are involved with substrate recognition and binding. Determining the structure of CAT binding to the substrate L-carnitine will identify specific substrate binding residues These residues will then be functionally and biologically characterized by an enzymatic assay and a novel genetic assay.

PAGE 40

29 Determining the structure of ChAT will also identify specific residues involved in catalysis and substrate recognition and binding. These two enzymes will most likely share several structural features. However, there are differences in L-carnitine and choline that may be reflected in the enzyme structures. In addition, the information gleaned from these structure studies can be used to determine the molecular basis of congenital myasthenic syndrome (CMS).

PAGE 41

CHAPTER 2 MATERIALS AND METHODS This chapter describes the general methods used to express, purify, crystallize and determine the structure of ChAT and CAT in complex with L-carnitine, and to characterize essential enzymatic amino acid residues of CAT. The results of these studies are discussed in the following chapters. The first goal of this project was to identify residues of CAT that interact with the substrate, L-carnitine, and then characterize these residues through an enzymatic and biological assay. The identification and characterization of these residues will provide further insight into the reaction mechanism of CAT and will help to advance understanding of other members of the L-carnitine/choline acyltransferase family. The general experimental design to accomplish these goals was as follows. Determine the structure of CAT with the substrate L-carnitine Mutate specific residues of CAT using site-directed mutagenesis Express and purify the mutant forms of CAT Determine kinetic parameters of the mutant forms Evaluate biological activity of mutant forms using a genetic selection system The second specific aim of this project was to determine the structure of ChAT. The structure of ChAT can not only be used to gain further structural and mechanistic understating of the L-carnitine/choline family, but also to understand the molecular basis of the disease, congenital myasthenic syndrome (CMS). 30

PAGE 42

31 Determine Structure of Carnitine Acetyltransferase with L-Carnitine To determine the x-ray crystal structure of CAT with the substrate L-carnitine, the enzyme was expressed, purified, and co-crystallized with L-carnitine. The crystals were then x-rayed and the structure was determined from the diffraction data. Expression Recombinant human peroxisomal CAT was expressed in Eschericia coli B834 cells. The coding region of CAT was amplified by polymerase chain reaction (PCR) and cloned into the pQE-9 (Qiagen) expression vector (Figure 2-1). This expression vector allows for high-level expression of proteins with a six-histidine tag on the N-terminus. Interestingly, the histidine tag does not affect enzyme activity and actually improves expression levels. The vector also contains an ampicillin selection marker and the T5 Figure 2-1. Diagram of the pQE-9 expression vector. PT5: T5 promoter, lacO: lac operon, RBS: ribosome-binding site, ATG: start codon, 6xHis: 6xHis tag sequence, MCS: multiple cloning site with restriction sites (BamH, SalI, PstI, HindII), Stop codons: stop codons in all three reading frames, Col E1: Col E1 origin of replication, Ampicillin: ampicillin resistance gene. Source: Qiagen.

PAGE 43

32 promoter, which is tightly regulated by the lac operon. The vector containing the CAT coding region was then transformed into E. coli cells. These cells were grown at 30C to an optical density of 0.5-0.7 at 600 nm. Expression was induced by adding isopropyl-b-D-thiogalacto-pyranoside (IPTG) to a final concentration of 1 mM. Cells were then grown for an additional 2 hours. The bacteria was harvested by centrifugation and stored at C until further use. Extraction Recombinant CAT is soluble and was extracted by cell lysis using the French Press. To purify the protein, the E. coli extract, prepared in 20 mM sodium phosphate buffer pH 7.6, was added to Ni-nitriloacetate-agarose (Ni-NTA) (Qiagen), which was equilibrated with the extraction buffer. After a 1 hour incubation the mixture was transferred to a column and the buffer was allowed to flow through, then collected. The imidazole ring of the histidine residues of the histidine tag binds to the nickel ions immobilized by the NTA groups of the matrix (Figure 2-2). Thus the enzyme is retained on the column. The column was then washed batch wise with buffer containing increasing concentrations of NaCl from 0 to 2 M. Free imidazole can also bind to the nickel ions and disrupt the binding of the histidine residues. This mechanism was employed to elute the enzyme, using an imidazole gradient from 5 to 150 mM, was employed to elute the enzyme. Next, the enzyme was applied onto a MonoQ column (Pharmacia) that was equilibrated with 20 mM sodium phosphate buffer pH 7.6. The enzyme was eluted with a linear salt gradient from 0 to 1 M and then concentrated/dialyzed using Ultra-free Tangential membranes (Millipore). The purified

PAGE 44

33 enzyme was either used immediately or stored at C in 20 mM sodium phosphate buffer pH 7.6 containing 40% glycerol. Figure 2-2. Interaction of histidine residues of 6xHis tag and the Ni-NTA agarose. Source: Qiagen Purification Purity of the purified enzyme was analyzed by SDS-polyacrylamide gel electrophoresis while the concentration of the enzyme was determined by the BIO-RAD Bradford Protein Assay with bovine albumin as the standard. Protein expression was further characterized by Western Blot. In this procedure, the enzyme was electrophoretically transferred to an Immobilon-P nylon membrane and incubated for 1 hour in blocking buffer consisting of 10 mM Tris-HCl buffer pH 7.4, 150 mM sodium chloride, 5% nonfat dry milk, and 0.2% Nonidet P-40. After this incubation the membrane was incubated overnight with an affinity purified anti-CAT antibody diluted in fresh blocking buffer. The next day the membrane was washed three times with a solution containing 10 mM Tris-HCl buffer pH 7.4, 150 mM sodium chloride, 0.25% sodium-7-deoxycholate and 0.1% SDS and then washed three more times with a solution

PAGE 45

34 containing 10 mM Tris-HCl buffer pH 7.4, and 150 mM sodium chloride. The enzyme was visualized after incubation with a goat anti-rabbit antiserum coupled to alkaline phosphatase and 5-bromo-4-chloro-3-indolyl phosphatase linked with p-nitroblue tetrazolium chloride as the color reagent. Crystallization Initial crystallization experiments were carried out using the hanging-drop vapor-diffusion method in a 24-well tissue culture Linbro plate (MP Biomedicals Inc.) at three different temperatures, 4, 16, and 22C. A sparse-matrix crystallization screen based on the original report by Jancarik & Kim (1991) [76] was used and promising conditions were further optimized with respect to pH and precipitants. Each drop of CAT-L-carnitine was formed by mixing 3 uL of 20 mg/mL CAT, 3 uL of 1 mM L-carnitine, and 6 uL of reservoir solution that was 50 mM Bis-Tris buffer pH 6.2, 100 mM sodium chloride, and 12-15% PEG 8000. Mutate Specific Residues of CAT Specific residues of CAT thought to be involved in substrate binding were investigated using the QuickChange (Stratagene) site-directed mutagenesis procedure. This method (Figure 2-3) has proven to be very effective, with an efficiency of 80%. First, oligonucleotide primers containing the desired mutation are incubated with a double-stranded DNA vector including the gene that encodes CAT and PfuTurbo DNA polymerase. During temperature cycling, the polymerase extends the primers generating a mutated plasmid containing staggered nicks. After temperature cycling, the plasmid is treated with Dpn I endonuclease. This endonuclease is specific for methylated and hemimethylated DNA and will digest the parental DNA template since DNA isolated

PAGE 46

35 from almost all E. coli strands is dam methylated and thus subject to Dpn I digestion. Conversely, newly synthesized DNA will not have been methylated and is therefore not Figure 2-3. Overview of the QuickChange site-directed mutagenesis method. Step 1, gene in plasmid with target site for mutation. Step 2, denature the plasmid and anneal the oligonucleotide primers containing the desired mutation. Use PfuTurbo DNA polymerase to extend and incorporate the mutagenic primers resulting in nicked circular strands. Step 3, Digest the methylated, non-mutated parental DNA template with Dpn I. Step 4, transform the circular, nicked dsDNA into XL-1 blue supercompetent cells. After transformation, the XL-1 blue supercompetent cells repair the nicks in the mutated plasmid. Source: Stratagene.

PAGE 47

36 susceptible. The nicked vector DNA containing the mutation is then transformed into XL1-Blue supercompetent cells, which will repair the nicks in the vector. The colonies are plated and grown on LB media. Individual colonies are screened to ensure correct insertion of the mutation by checking for a restriction site that has been introduced by the primers. The high fidelity of PfuTurbo DNA polymerase and the low number of thermo cycles contribute to the high mutation efficiency and low potential for random mutations. However, to ensure that random mutations have not been introduced, the correct sequence was verified by DNA sequencing (UF DNA Sequencing Core) before further experiments are completed. The mutant forms of the enzyme were expressed and purified as previously described. In addition, expression levels of the mutant forms of the enzyme were compared to the wild type form by SDS-PAGE and Western Blot, as previously described. Evaluate Enzymatic Activity The kinetic parameters were then determined using a continuous, spectrophotometric assay. This is an indirect measure of the forward direction of the reaction. In the forward direction the substrates, acetyl-CoA and L-carnitine, yield the products acetyl-L-carnitine and CoASH. The release of CoASH is monitored spectrophotometrically with the reagent 5,5-dithiobis-2-nitrobenzoic acid (DTNB). DTNB reacts with the free sulfhydryl group of CoASH to form the yellow colored 5-thio-2-nitrobenzoate (NTB), which absorbs strongly at 412 nm (Figure 2-4). Activity was measured as an increase in absorbance at 412 nm over time. Kinetic values were used to interpret the data from the mutant and wild type forms of the enzyme. The initial velocity of the enzymes with different concentrations of

PAGE 48

37 substrates (L-carnitine and acetyl-CoA) was determined. This data was fitted to the Michaelis-Menton equation using nonlinear regression with the program Prism (Graphpad). The values of Vmax, Km, kcat and kcat/Km were calculated to determine how the COOHO2NO2NCOOHSS + O2NCOOHSCoAS +COOHO2NSH SHCoA DTNBNTB Figure 2-4. Schematic representation of the CAT assay. In this reaction DTNB reacts with the free sulfhydryl group of CoA-SH to form the chromogenic NTB. amino acid mutations effect the enzyme. The Vmax is the theoretical maximum velocity of the enzyme reaction expressed as the product formed per unit of time. Km is the concentration of the substrate at half the maximum velocity and is expressed as a concentration. The turnover number, or the kcat value, is the maximum velocity (Vmax) divided by the concentration of the enzyme and gives information about the efficiency of the enzyme. The kcat/Km is a measure of the catalytic efficiency. By comparing these values for the mutant and wild type forms of the enzyme we will be able to understand how these residues interact with the substrates and determine their role in catalytic function.

PAGE 49

38 Evaluate Biological Activity The yeast strain used in this study is derived from the strain FY23. In this strain, the CIT2 and CAT2 gene were deleted [77]. Yeast were either grown on YPD media, containing 1% (w/v) yeast extract, 2% (w/v) peptone, and 2% (w/v) glucose, or YND media, containing 0.67% (w/v) yeast nitrogen base without amino acids and 2% (w/v) glucose. Media containing ethanol (YNEC) as the carbon source was made with 0.67% (w/v) yeast nitrogen base without amino acids and 3% (v/v) ethanol. This media was also supplemented with L-carnitine at a concentration of 10 mg/L (YNEC). Plasmid containing wild type and mutated CAT was introduced into the mutant yeast strain by the lithium acetate method. Briefly, yeast cells from an overnight culture of yeast, grown in YPD media, were harvested and washed in buffer containing 10 mM Tris, 0.1 mM lithium acetate, and 1 mM EDTA. Next, 50% PEG 4000, 1 M lithium acetate, water, and plasmid DNA were added to the cells, in the described order, and the mixture was vortexed. After a 25-minute incubation in a 42C water bath, the cells were centrifuged and the supernatant was removed. The remaining cells were resuspended in water, plated onto YND agar plates and grown in a 30C incubator. The yeast cells that were plated onto YND agar plates were allowed to grow for 2-3 days before being replica plated onto YNEC and YND agar plates. Growth on YNEC and YND plates was monitored for several days. Structural Determination of Choline Acetyltransferase To determine the structure of ChAT, the enzyme was expressed, purified, and crystallized following the same general procedure as described previously with CAT and L-carnitine. Briefly, recombinant rat ChAT was expressed in E. coli B834 cells and

PAGE 50

39 extracted by cell lysis with the French Press. The extracted enzyme was applied to a column containing Ni-NTA agarose (Qiagen) and washed batch wise with buffer containing increasing concentrations of NaCl from 0 to 2 M and eluted by an imidazole gradient from 5 to 150 mM. The enzyme was then applied to a MonoQ column (Pharmacia) and eluted with a NaCl gradient from 0 to 1 M and then concentrated/dialyzed using Ultra-free Tangential membranes (Millipore). The purified enzyme was either used immediately or stored at -80C in 20 mM sodium phosphate buffer pH 7.6 containing 40% glycerol. The purity of the enzyme was evaluated by SDS-polyacrylamide gel electrophoresis and Western Blot while the concentration of the enzyme was determined by the BIO-RAD Bradford Protein Assay with bovine albumin as the standard. Purified ChAT was crystallized using the hanging-drop vapor-diffusion method in a 24-well tissue culture Linbro plate (MP Biomedicals, Inc.) by mixing 3 uL of 10 mg/mL ChAT with 3 uL of the reservoir solution, containing 50 mM MES buffer pH 6.0, 100 mM sodium chloride, and 8-10% PEG 8000 at 4C.

PAGE 51

CHAPTER 3 STRUCTURE DETERMINATION OF CARNITINE ACETYLTRANSFERASE WITH L-CARNITINE The goal of the first specific aim was to identify and characterize essential amino acid residues of human CAT. These objectives were met by determining the structure of CAT with its substrate L-carnitine, using an enzymatic assay to evaluate the catalytic role of specific residues, and by developing a genetic selection system to evaluate the biological role of these residues. This chapter discusses structure determination of the enzyme complex and residues involved in catalysis. Introduction L-carnitine is a highly polar zwitterionic quaternary amine carboxylic acid present in some prokaryotes and all eukaryotes. In prokaryotic cells L-carnitine serves either as a nutrient (such as a carbon and nitrogen source) or as an osmolyte. However, in eukaryotic cells, L-carnitine serves exclusively as a carrier of acyl moieties through various subcellular compartments [22]. L-carnitine helps to remove toxic accumulations of poorly metabolized fatty acids in the mitochondria but, more significantly, the carnitine system plays an essential physiological role in fatty acid -oxidation and in the maintenance of acyl-CoA pools [78, 79]. L-carnitine and its esters are transported across the inner mitochondrial membrane to facilitate the entrance of long chain fatty acids into the mitochondria for -oxidation [19, 78]. The physiological actions of L-carnitine are mediated via a sodium dependent high affinity carnitine/acylcarnitine plasma membrane transporter, a carnitine/acylcarnitine 40

PAGE 52

41 translocase located on the inner mitochondrial membrane, as well as a group of L-carnitine utilizing enzymes, the carnitine acyltransferases [18, 22, 78]. As previously discussed, carnitine acyltransferases (CAT, COT and CPTs) facilitate the reversible transfer of acyl groups between CoA and L-carnitine, and fall into three main classes which differ in their acyl-chain length selectivity, sub-cellular compartmentalization and physiological function [19, 37, 42]. CAT has a substrate preference for short-chain acyl-CoAs and is found in the mitochondrial matrix, peroxisome and the endoplasmic reticulum. COT is predominantly localized in peroxisomes and has a substrate preference for medium length acyl-CoAs. CPTs are found both in the outer mitochondrial membrane (CPT I) and the mitochondrial matrix (CPT II) with a substrate preference for long-chain acyl-CoAs. Decreases in CAT activity (particularly in the brain) have been described and present clinical symptoms of progressive ataxic encephalopathy. These symptoms are most likely due to a decrease in acetyl-CoA available for the citric acid cycle in the mitochondria [55]. In addition, recent experiments have suggested that CAT also plays a role in age-associated neurodegeneration. It was found that feeding rats acetyl-L-carnitine and lipoic acid increases CAT activity, improves performance on memory function tasks, reduces brain mitochondrial structural decay, and reduces oxidative damage in the brain [58-60]. One explanation is that over time CAT becomes inactivated by aldehydes formed by lipid peroxidation and that acetyl-L-carnitine has a protective effect while lipoic acid reduces lipid peroxidation and reactivates CAT [60]. In spite of its importance, fundamental questions remain about the structure and function of the carnitine acyltransferases. To further understand the molecular basis of

PAGE 53

42 the benefits of acetyl-L-carnitine and lipoic acid on CAT, carnitine acetyltransferase disorders and amino acid residues involved in catalysis, it is important to know the structure and mechanism of action of this enzyme. Previous chemical modification, inactivation, site-directed mutagenesis, and three-dimensional structure studies have helped to identify residues of CAT that interact with the substrates (L-carnitine and CoA) and explain the mechanism of this enzyme. As previously discussed, early studies with CAT isolated from pigeon breast muscle demonstrated that the enzyme follows a random-order equilibrium reaction mechanism that includes the formation of an intermediate ternary complex [61]. Further studies revealed that the reaction product of CAT and radiolabeled bromoacetyl-L-carnitine is the alkylation of a histidine residue. The alkylation specifically occurs at the N-3 position and suggests that the active site of CAT contains a histidine residue that is directly involved in catalysis [62]. The histidine residue is believed to act as a general base, removing a proton from L-carnitine or CoA (depending on the direction of transfer) to promote the formation of a tetrahedral intermediate [18]. The crystal structure of human peroxisomal and mouse CAT have shown that the enzyme is composed of 17 helices and 14 strands, that form two equal sized domains [64, 65]. These two domains (Figure 3-1) are tightly associated with each other and form a central active site tunnel with His322 located in the middle of the tunnel, suggesting that L-carnitine and acetyl-CoA bind to opposite sides of the tunnel [64]. Several amino acid residues have been implicated in the active site and putative binding residues of L-carnitine and CoA (L-carnitine: His322, Glu326, Tyr431, Arg443, Ser531, Thr532; CoA:

PAGE 54

43 Ser533, Leu142, Lys398, Ser407 and Asp487) [64, 65]. To verify that these residues do indeed interact with L-carnitine and to identify additional L-carnitine binding residues, the structure of CAT with its substrate, L-carnitine, was determined. Figure 3-1. Overall structure of human peroxisomal carnitine acetyltransferase [64]. Ribbon diagram representation depicts domains I (blue) and II (orange). The N and C termini are indicated. Materials and Methods The experimental procedures involved in this project were described in Chapter 2. Briefly, recombinant human peroxisomal CAT was expressed in bacterial cells with a histidine tag and then purified using a column containing Ni-NTA agarose. The purified enzyme was then co-crystallized with its substrate (L-carnitine) and the three-dimensional structure was determined.

PAGE 55

44 Results and Discussion Structure Determination and Overall Structure of the Enzyme Complex X-ray data of the CAT-L-carnitine crystal were collected at the Cornell High Energy Synchrotron Source (CHESS). Data was collected using a crystal-to-detector distance of 100 mm and a 1 oscillation angle per image. A total of 110 images were collected from a single crystal at 100 K. Before flash freezing, the crystal was cryo-protected in the crystallization reservoir solution using 15% PEG 8000 and 40% glycerol. Data were indexed, integrated and scaled using the HKL2000 suite programs DENZO and SCALEPACK [80]. Data were then phased using CNS program version 1.1 [81]. The structure was further refined using standard protocols within SHELXL97 [82] with additional interactive modeling using the program O [83]. The crystal of the CAT-L-carnitine complex was isomorphous with the free CAT crystals and belonged to the space group P212121 with unit cell dimensions of a = 137.6, b = 84.7, c = 57.3 The structure was refined to a 1.8 resolution. The final model of the complex structure consisted of 590 amino acid residues (9-599), 465 water molecules and the substrate L-carnitine. Comparison of CAT-L-Carnitine with Free CAT The crystal structure of the complex is almost identical to that of the free form of CAT with a backbone root-mean-square deviation of 0.41 (Figure 3-2) [64]. The overall structure of the free form of CAT reveals that the enzyme is a monomeric protein composed of two equally sized domains that are tightly associated with one another to form a central solvent accessible tunnel, which constitutes the active site of the enzyme. Examination of the complex structure shows that there are subtle changes in electron

PAGE 56

45 density for the side chains of some active site residues in the L-carnitine binding site, but no major structural changes occur on substrate binding. Figure 3-2. Structural similarity and superimposition of CAT and CAT-L-carnitine. Coil diagram shows the superimposition of the overall structure of CAT (gray) and CAT-L-carnitine (green), the active site residues, His322, Tyr431, Thr444, Arg497, Phe545, Val548, and the water molecules, W621 and W679, with L-carnitine in the active site. Domain I and II and the Nand C-termini are indicated. Comparisons of the structures of the CAT-L-carnitine complex and free CAT crystal structures show that there are slight changes in the conformation of different surface regions and small displacement of active site residues. However, the only significant conformational change occurs with Ser433. Upon L-carnitine binding the chi angles of Ser433 change from 58 to positioning the hydroxyl group 3.03 from one of the carboxylic oxygen atoms of L-carnitine (Figure 3-3). The orientation of Ser433 and L-carnitine in CAT suggests that this residue may also contribute to transition state stabilization.

PAGE 57

46 His322 L-Carnitine Ser433 Ser433-Free Figure 3-3. Representation of Ser433 positional change in free and L-carnitine bound form of CAT. The chi angle of Ser433 rotates from 58 to -31 upon substrate binding to interact with the carboxylic oxygen of L-carnitine and/or contribute to transition state stabilization. L-Carnitine Binding Site The binding site for L-carnitine in CAT is located at one entrance of the active site tunnel, as was predicted by the previous structure of the free enzyme [64]. The binding surface of L-carnitine is formed by helix 4, strand 1, strand 7, and strand 8 from domain I and strand 9 to 10, helix 13 to 14, and strand 11 to 14 from domain II. Six residues, His322, Tyr431, Thr444, Arg497, Phe545, Val548, and two tightly associated water molecules (H2O621 and H20679), in particular, constitute a distinct pocket to specifically recognize and interact with L-carnitine (Figure 3-4). Among these active site residues, His322 is completely conserved in all L-carnitine/choline acyltransferases and is believed to be the catalytic residue that serves as a general base to catalyze fatty acyl transfer [18, 64, 65]. L-carnitine binds to CAT with its -hydroxyl moiety aligned directly with His322 to form a hydrogen bond with the N-3 atom of the imidazole ring (2.76 ) (Figure 3-5). This interaction supports the catalytic role of His322, which has been proposed to deprotonate the -hydroxyl group of L-carnitine or the thiol group of CoA to facilitate nucleophilic attack on the ester linkage

PAGE 58

47 Figure 3-4. Electron density map of the carnitine binding site of L-carnitine acetyltransferase. Residues His322, Tyr431, Thr444, Arg497, Phe545, and Val548 and water molecules W621 and W679 are shown with L-carnitine. of the acyl group [18, 64, 65]. Additionally, the alignment of L-carnitine and His322 is stereospecific and explains why D-carnitine is the inactive enantiomer, since the -hydroxyl of D-carnitine would point in the opposite direction of His322 and would not be accessible for acyl group transfer. Besides His322 a water molecule (H2O679) is within hydrogen bonding distance with the -hydroxyl moiety of L-carnitine. This water also interacts with the side chain hydroxyl group of Ser533, which is the last residue in a Ser-Thr-Ser motif that is conserved among all carnitine/choline acyltransferases. Ser533 is believed to play an essential role in transition state stabilization during catalysis and may also be involved in L-carnitine binding through its interaction with H2O679 [64, 84, 85].

PAGE 59

48 Figure 3-5. Close up figure of the active site of L-carnitine acetyltransferase-L-Carnitine. This three-dimensional representation depicts the interaction between His322 and L-carnitine. Hydrogen bond interactions seem to be the primary force for the recognition of L-carnitine since a hydrogen bond network is formed between the carboxylate oxygens (O1 and O2) of L-carnitine and a number of active site residues (Figure 3-6) [65, 85]. Specifically, Thr444 and an ordered active site water molecule, H2O621, interact with one oxygen (O2) of the carboxylate group. The water molecule is further coordinated by three residues in the active site, Trp81, Tyr86, and Glu326. The other carboxylic oxygen appears to primarily interact with Tyr431 through a strong hydrogen bond (2.53 ). In addition to the numerous hydrogen bond interactions with L-carnitine there is also an electrostatic interaction between the negatively charged carboxylate group and the positively charged guanidinium group of Arg497. Furthermore, the side chain NH of Arg497 forms hydrogen

PAGE 60

49 bonds with the hydroxyl groups of Tyr431 and Thr444, which may further influence L-carnitine binding. Figure 3-6. Two-dimensional representation of the interactions observed between L-carnitine and L-carnitine acetyltransferase. Dashed lines denote hydrogen bonds and distances in Hydrophobic interactions are shown as arcs and radial spokes. In contrast to the negatively charged carboxylate group of L-carnitine, the positively charged quaternary nitrogen does not appear to have electrostatic interactions with negatively charged residues in the active site. However, the aromatic side chain of Phe545 is close to the methyl carbon of the quaternary amine head group of L-carnitine

PAGE 61

50 (3.6 ). Thus, Phe545 appears to be the primary residue responsible for interacting with the quaternary nitrogen of L-carnitine through pi-cation interactions (Figure 3-4 and 3-6) [86, 87]. The de-localized electron clouds on the side chain benzene ring can interact with and stabilize the positively charged quaternary amine head group of L-carnitine. Val548 may also contribute to the binding of the trimethyl amino head group of L-carnitine through van der Waals interactions. Comparison of CAT-L-Carnitine with Mouse CAT Recently the crystal structure of mouse CAT in complex with L-carnitine was determined to 1.9 [65]. Based on sequence homology, human CAT and mouse CAT share 90% identity at the amino acid level. Interestingly, the two proteins crystallize in different space groups, human CAT-L-carnitine crystallized in the orthorhombic space group P212121 with one monomer in the asymmetric unit whereas mouse CAT-L-carnitine crystallized in the monoclinic space group C2 which contains two molecules in the crystallographic asymmetric unit [65, 85]. This may reflect a difference in the unit cell packing arrangement of the proteins themselves, the different crystallization conditions [65, 88] or both. Despite this difference, the structures are very similar with a root-mean-square deviation of 0.6 over 581 C atoms. Furthermore, there is a slight variation in atomic distances between L-carnitine and the active site residues in human and mouse CAT. This difference is within experimental error but may be attributable to subtle structural differences between the two proteins. Overall, the conformation of L-carnitine and the amino acid residues that form its binding site are almost identical between these two enzymes.

PAGE 62

51 Conclusions The structure of the complex of CAT and L-carnitine has been determined and reveals several key residues that recognize and interact with the substrate L-carnitine. This structure also supports the proposed mechanism that His322 plays an essential role in catalysis by abstracting the proton from the -hydroxyl of L-carnitine or the thiol of CoA depending on the direction of the reaction [64, 65]. Additionally, active site residues involved in binding L-carnitine and catalysis were identified. Specifically, Tyr431, Thr444, and Arg497 form a binding pocket for the carboxylate group while Phe545 and Val548 interact with the quaternary amine head group of L-carnitine. This study also provides important information about L-carnitine binding and the catalytic mechanism of other carnitine acyltransferases since His322, Tyr431, Thr444, and Phe545 are highly conserved among the carnitine acyltransferase family.

PAGE 63

CHAPTER 4 FUNCTIONAL AND BIOLOGICAL CHARACTERIZATION OF ESSENTIAL RESIDUES OF CARNITINE ACETYLTRANSFERASE Introduction The structure of CAT in complex with L-carnitine indicated several residues involved in catalysis, His322, Tyr431, Thr444, Arg497, Phe545, and Val548. To confirm these residues are essential enzymatic residues and determine their role in catalysis, they were functionally and biologically characterized. To functionally characterize these residues, they were replaced with amino acid residues of varying charge, side-chain length, aromaticity and polarity. The kinetic properties of the mutant forms of the enzymes were determined by an enzymatic assay, as described in Chapter 2. To biologically characterize active site residues a novel genetic system was developed in the yeast, Saccharomyces cerevisiae. In this system an endogenous CAT is disrupted rendering the cells unable to grow on media with short-chain carbon groups, such as ethanol, as the sole energy source. However, human CAT complements these mutant cells and enables the cells to grow on ethanol media by providing a pathway for short-chain carbon groups to enter the mitochondria for oxidation. Degradation of fatty acids in yeast differs from mammals. In yeast, long-chain fatty acid oxidation occurs exclusively in the peroxisomes, where they are oxidized to acetyl-CoA [89, 90]. For further metabolism the acetyl-CoA must be transported from the peroxisome to the mitochondria either by the peroxisomal glyoxylate cycle or the carnitine shuttle (Figure 4-1) [91]. In the five steps of the glyoxylate cycle, acetyl-CoA 52

PAGE 64

53 forms succinate which is then transferred to the mitochondria where it enters the citric acid cycle [91]. In the first step of the glyoxylate cycle, acetyl-CoA is combined with oxaloacetate to form citrate by the action of the enzyme peroxisomal citrate synthase. The enzyme aconitrase catalyzes the conversion of citrate to isocitrate, which is then hydrolyzed by isocitrate lysase 1 to glyoxylate, which in turn forms malate by the enzyme malate synthase, and succinate, which is transported to the mitochondria. Once in the mitochondria the succinate will enter the citric acid cycle and release malate, which will start the process of gluconeogenesis. Fatty acyl-CoA Acetyl-CoA + L-Carnitine Acetyl-L-CarnitineCarnitine Shuttle -oxidation CAT Glyoxylate Cycle Oxaloacetate CitrateIsocitrate GlyoxylateMalate CIT2 A CO MLS ICL2 Succinate Acetyl-L-Carnitine CAT L-Carnitine + Acetyl-CoA Succinate TCA Cycle TCA Cycle Malate PeroxisomeMitochondria CO2 CO2 Figure 4-1. Schematic representation of the two pathways for the metabolism of fatty acids. Long-chain fatty acids are oxidized in the peroxisome by -oxidation to acetyl-CoA. The acetyl-CoA can use either the carnitine shuttle or glyoxylate cycle for further metabolism, as described in the text. CAT, carnitine acetyltransferase; CIT2, peroxisomal citrate synthetase, ACO, aconitase, ICL, isocitrate lysase; MLS, malate synthase.

PAGE 65

54 With the carnitine shuttle, the acetyl group of acetyl-CoA is transferred to L-carnitine and the resulting acetyl-L-carnitine is transferred across intracellular membranes to the mitochondria. Once in the mitochondria the acetyl group is transferred from L-carnitine to a free CoA where it can then enter the citric acid cycle [91]. CAT catalyzes the reversible transfer of L-carnitine to acetyl-CoA since yeast does not contain endogenous COT or CPT activity. Three genes encoding CAT in S. cerevisiae have been identified, CAT2, YAT1 and YAT2. The first identified CAT gene, CAT2, is translated to CAT, which is localized to the peroxisomes and inner mitochondrial membrane and is responsible for more than 99% of total CAT activity in galactose grown cells [92]. The second gene, YAT1, codes for a CAT located in the outer surface of the mitochondria that contributes to 5% of the total CAT activity in acetate and ethanol grown cells [93]. The most recently identified gene, YAT2, codes for a cytosolic CAT that is responsible for 50% of total CAT activity in ethanol grown cells [77]. Disruption of any one of the three genes that code for CAT (CAT2, YAT1 and YAT2) allows for growth similar to wild type cells on all carbon sources [77]. However, in yeast cells devoid of the glyoxylate cycle (by disruption of the CIT2 gene and disruption of any one of the genes encoding CAT) the cells cannot grow on media with fatty acids, such as oleic acid, or non-fermentable carbon sources, such as ethanol or acetate, as the sole carbon source. However, these cells grow as normal with glucose as the carbon source [77]. This indicates that in the absence of the glyoxylate cycle, all three CATs are essential for growth. We hypothesized that human CAT would complement the

PAGE 66

55 mutant yeast strain and allow growth on media containing ethanol as the sole carbon source. To develop the genetic system, we obtained a strain of S. cerevisiae in which CIT2, the gene that encodes the first enzyme of the glyoxylate cycle, and one of the CAT genes, CAT2, was disrupted, thus blocking both the glyoxylate cycle and the carnitine shuttle. These cells could not grow on media with ethanol as the sole carbon source since they were not capable of utilizing ethanol as an energy source. We hypothesized that human CAT would complement the mutant yeast strain and that we could use this strain to evaluate the biological function of CAT and confirm the role of amino acid residues thought to be involved in substrate recognition/binding and catalysis. Materials and Methods The general procedures used in these experiments were described in Chapter 2. Briefly, recombinant wild type and mutant human peroxisomal CAT was expressed in E. coli B834 cells with a histidine tag and then purified using a column containing Ni-NTA-agarose. The kinetic parameters of the purified enzyme were determined using a spectrophotometric assay that measures the forward direction of the reaction. The biological activity of the wild type and mutant forms of the enzyme were determined using an assay in the yeast. Plasmids containing wild type and mutant forms of CAT were introduced to a mutant strain of yeast. These yeast cells were able to grow with media containing glucose as the carbon source, but were unable to grow with non-fermentable carbon sources, such as ethanol. Yeast cells expressing wild type and mutant forms of CAT were screened for growth on media containing ethanol or glucose as the sole carbon source.

PAGE 67

56 Results and Discussion Functional Characterization of L-Carnitine Binding Residues The structure of the binary complex of CAT and L-carnitine reveals several important residues responsible for the recognition and/or binding of L-carnitine. These residues are conserved among all carnitine acyltransferases with the exception of Phe545 and Val548 that become valine and methionine respectively in carnitine octanoyltransferases (Figure 4-2). In order to determine their contribution to L-carnitine binding, four residues, Tyr431, Thr444, Arg497, and Phe545, have been investigated using site-directed mutagenesis. Figure 4-2. Sequence alignment of carnitine and choline acyltransferases. The amino acid sequences of human and mouse hpCAT, human and rat carnitine octanoyltransferase, human and rat carnitine palmitoyltransferase 1 liver isoform, human and rat carnitine palmitoyltransferase 1 muscle isoform and human and rat choline acetyltransferase were aligned using ClustalW [94]. The residue numbers bordering each segment are indicated. The residues in bold typeface, Tyr431, Thr444, Arg497 and Phe545, were mutated in this study using site directed mutagenesis. Substitution at these positions creates mutant forms of CAT with altered steady-state kinetic constants. Generally, there is a large increase in Km values for L-carnitine, with minor changes in Km values for acetyl-CoA compared to L-carnitine (Table 4-1). In almost all cases the catalytic efficiency of L-carnitine with the mutant forms compared to

PAGE 68

57 the wild type enzyme was decreased. A decrease in turnover rates was also observed, although in two cases, Tyr431Ala and Arg497Gln, the kcat value increased (Table 4-1). Replacement of Tyr431 with phenylalanine produced the most defective mutant enzyme, which increased the L-carnitine Km value by 326 fold and decreased the kcat value by 49 fold compared to wild type CAT (Table 4-1). This mutation also results in the most dramatic decrease in catalytic efficiency, 2.2 x 101 compared to 3.5 x 105 for wild type (Table 4-1). The substitution of tyrosine with phenylalanine removes a phenolic hydroxyl group, which is involved in a strong hydrogen bond with one of the oxygen atoms of the carboxylic group of L-carnitine (2.53 ). Replacing Tyr431 with alanine also increased the L-carnitine Km constant by 100 fold, but this increase was three fold lower than Tyr431Phe (Table 4-1). The decreased enzymatic activity and increased Km and lowered catalytic efficiency towards L-carnitine for both of the Tyr431 mutants demonstrates the importance of this residue and the requirement of a hydroxyl group for high affinity binding. Table 4-1. Kinetic parameters of recombinant wild type and mutant forms of carnitine acetyltransferase. Enzyme Vmax L-Carnitine (mol/min*mg) Km L-Carnitine (mM) Km ACoA (M) kcat (sec-1) kcat/Km (M-1sec-1) Wild-Type 33.1 1.4 0.11 0.02 42.3 2.7 38.5 3.5 x 105 Tyr431Ala 38.0 3.2 11.1 1.4 87.1 1.6 44.2 4.0 x 103 Tyr431Phe 0.67 0.07 35.9 6.4 497.9 83.5 0.78 2.2 x 101 Thr444Ala 5.0 0.3 7.5 0.9 188.3 18.7 5.76 7.7 x 102 Arg497Gln 81.5 4.9 25.6 3.0 120.2 15.7 94.8 3.7 x 103 Phe545Ala 2.9 0.1 2.0 0.2 177.9 15.2 8.57 4.3 x 103 Phe545Tyr 26.2 0.8 0.15 0.01 63.7 6.1 30.5 2.0 x 105

PAGE 69

58 It is interesting that substitution of Phe, rather than Ala, at this position produced a more defective mutant, since Phe is a more conservative mutation. This may be explained by a couple of factors. First, charge repulsion most likely occurs between the aromatic side chain and the negatively charged carboxylate group of L-carnitine, reducing substrate affinity. Furthermore, this substitution may produce unfavorable interactions between the Phe residue and adjacent amino acids, which could distort the geometry of the L-carnitine binding pocket. This may explain the high Km as well as extremely low specific activity of this mutant. Alternatively, the finding that insertion of Ala at this position yields a less drastic change in kinetic parameters can perhaps be explained by the fact that substitution of Tyr by a much smaller Ala residue may allow a water molecule to fit into the vacant space to form hydrogen bond interactions with L-carnitine (Figure 4-3). Tyr431 L-Carnitine His322 A His322 L-Carnitine Ala431 B His322 L-Carnitine Phe431 C Figure 4-3. Three-dimensional representation of the active site with Tyr431. A, Wild type with His322, L-carnitine and Tyr431; B, Mutant form of the enzyme with His322, L-carnitine and Ala modeled into position 431; C, Mutant form of the enzyme with His322, L-carnitine and Phe modeled into position 431. Another residue that forms putative interactions with L-carnitine was also investigated. Thr444 is 2.82 from the other carboxylic oxygen of L-carnitine and can form a hydrogen bond with the side chain hydroxyl group. Substitution of Thr444 with an

PAGE 70

59 alanine increases the L-carnitine Km value by 68 fold and decreases the kcat value by 6 fold (Table 4-1). In addition, Ala444 CAT is not as effective as a catalyst compared to the wild type enzyme since the catalyst efficiency of Ala444 CAT is more than 450 fold less than wild type (Table 4-1). The substitution of threonine by alanine removes the side chain hydroxyl group and presumably the hydrogen bond between Thr444 and L-carnitine, resulting in a large increase in the L-carnitine Km value and a corresponding decrease in the catalytic efficiency utilizing L-carnitine (Table 4-1). The above mutations and kinetic results clearly establish the importance of hydrogen bonds among Tyr431, Thr444, and the carboxylic group of L-carnitine. Next the role of Arg497 was examined. This residue was proposed to influence L-carnitine binding through both electrostatic interactions and its participation in a hydrogen-bonding network with Tyr431 and Thr444. In addition, Arg497 is conserved among all carnitine acyltransferases [64], but the analogous position in all known choline acetyltransferases is an asparagine (Figure 4-2) [95]. Since choline does not have a carboxylate group as L-carnitine does, it was hypothesized that this arginine residue is partially responsible for the ability of carnitine acyltransferases to distinguish L-carnitine from choline [95]. In fact, substitution of this residue with an asparagine residue in the bovine form of COT increased the L-carnitine Km value ~1650 fold and enabled the mutant form of the enzyme to use choline as a substrate [95]. Substitution of Arg497 with glutamine in CAT increases the L-carnitine Km value by 232 fold and the kcat value by 2.5 fold (Table 4-1). By replacing the positively charged Arg497 with an uncharged residue the binding pocket for the carboxylate group of L-carnitine is most likely disrupted, which may account for a decreased apparent affinity

PAGE 71

60 towards L-carnitine and a decreased catalytic efficiency of 95 fold compared with the wild type CAT (Table 4-1). These results are consistent with the previous study conducted with bovine COT [95] and the structural analysis reported in this work suggesting a critical role for the electrostatic interactions in binding and recognition of L-carnitine. Finally, the interaction of the positively charged quaternary amine head group of L-carnitine in the active site of CAT was examined. Phe545 is hypothesized to be the residue primarily responsible for the binding of the quaternary amine of L-carnitine. The structure of the binary complex shows that the aromatic side chain of Phe545 is 3.6 from the methyl carbon of the quaternary nitrogen of L-carnitine. Substitution of Phe545 with alanine increased the L-carnitine Km value by ~18 fold and decreased the turnover rate by 4.5 (Table 4-1). On the other hand, substitution with tyrosine caused only minor changes in the kinetic parameters compared to wild type. These results demonstrate the importance of an aromatic ring for L-carnitine interactions, since removal of it led to a more drastic change than replacement with a phenolic side chain. It is interesting to note that the phenylalanine at this position is conserved in all carnitine acyltransferases except for COT. Mutations may have occurred in COT to compensate for the loss of an aromatic group at this position. Furthermore, these results are consistent with the hypothesis that the pi-electrons of the aromatic group of Phe545 interact with the positively charged nitrogen of L-carnitine through pi-cation interactions. This type of interaction is also observed with acetylcholinesterase, an enzyme associated with the cholinergic nerve terminals that hydrolyzes acetylcholine to terminate neurotransmission. The crystal structure of acetylcholinesterase and photoaffinity labeling revealed that the aromatic ring

PAGE 72

61 of Phe330 as well as Trp84 interacts with the quaternary amine group of acetylcholine in the active site [96]. Based on the crystal structure of CAT, the residues that were investigated using site-directed mutagenesis appear to be primarily involved in binding L-carnitine. The kinetic results from this study presented here further support this conclusion. For the most part, the substitutions at these positions produce only minor effects on the Km values for acetyl-CoA (1.5-11 fold) compared to the dramatic affects on L-carnitine (1.5-330 fold) (Table 4-1). The largest change in Km value for acetyl-CoA occurred with the Tyr431 Phe substitution, which also had the largest Km value for L-carnitine. This mutant also showed the lowest specific activity and turnover number compared to any other substitution. Thus it seems likely that this mutation causes a significant perturbation in the active site leading to decreased activity and ligand affinity. Overall, the substitutions at these positions in the active site of CAT reveal the key role these residues play in high affinity binding of L-carnitine and catalysis. Biological Characterization of L-Carnitine Binding Residues Functional characterization studies have confirmed four residues to be intimately involved in L-carnitine recognition and binding. Next, the biological importance of these residues was evaluated with a novel genetic selection system. The genetic system is based on the observation that there are only two pathways for peroxisomally produced acetyl-CoA to be further metabolized: the glyoxylate cycle and the carnitine shuttle. The first step of the glyoxylate cycle is the formation of citrate from acetyl-CoA and oxaloacetate, catalyzed by the enzyme CIT2 [91]. Disruption of the CIT2 gene blocks the glyoxylate cycle and forces the cell to employ the carnitine shuttle for further metabolism of acetyl-CoA. The carnitine shuttle utilizes carnitine

PAGE 73

62 acetyltransferase activity to transport acetyl-CoA formed in the peroxisome to the mitochondria where it can enter the citric acid cycle for energy production. CAT also facilitates in the transfer of cytosolic acetyl-CoA formed by the catablolism of ethanol and/or acetate to the mitochondria. As previously discussed, disruption of any one of the three genes that code for CAT blocks the carnitine shuttle [77]. Yeast cells devoid of the glyoxylate cycle (by disruption of the CIT2 gene) and the carnitine shuttle (by disruption of the CAT2 gene) cannot grow on media with fatty acids, such as oleic acid, or non-fermentable carbon sources, such as ethanol or acetate, as the sole carbon source. However, they grow as normal with glucose as the carbon source [77]. This indicates that in the absence of the glyoxylate cycle, all three CATs are essential for growth. Therefore, it was hypothesized that human CAT would complement the mutant yeast strain and allow growth on media containing ethanol as the sole carbon source. To test this hypothesis mutant yeast cells with the CIT2 and CAT2 genes disrupted were screened on ethanol plates to determine their viability. The mutant yeast cells grew on YND media, but when replica plated onto media with ethanol as the sole carbon source there were no viable cells, as expected. However, when human CAT was introduced into the cells the mutant yeast cells were able to grow on both YND and YNEC media (Figure 4-4). This indicates that human CAT was able to provide a mechanism for the transport of cytosolic acetyl-CoA to the mictochondria for energy production and therefore allow growth on non-fermentable carbon sources such as ethanol. This system suggests that the mutant yeast phenotype can function as a genetic screen for the verification of residues of CAT important for the biological function of the enzyme.

PAGE 74

63 To test the feasibility of the genetic system and verify that human CAT does complement the mutant yeast strain, a mutant form of CAT was introduced to the cells. In this form of CAT the active site histidine, His322, was substituted with alanine, serine and glutamine, rendering the enzyme inactive. The resulting yeast cells were able to grow on YND media; however, no viable cells were observed on the ethanol media, as shown in Figure 4-5. Since the mutant forms of CAT were inactive, cytosolic acetyl-CoA was unable to cross the mitochondrial membrane and could not be used as an energy source. Since ethanol was the only carbon source available the mutant yeast cells could not survive without metabolizing it for energy. These results show that the mutant form of CAT could not complement the yeast strain. This provides further evidence that human CAT complements the mutant yeast strain and verifies that this genetic system can be used to evaluate the biological activity of essential enzymatic residues of human CAT. Next, the biological activity of the L-carnitine binding residues was evaluated. Changing these residues were all shown to decrease the enzymes effectiveness by displaying altered catalytic efficiencies. However the distorted kinetic values for each mutation were different. To further understand how these residues affect L-carnitine binding and to what degree, they were characterized biologically using the genetic system in the yeast. The mutant forms of the enzyme were sub-cloned into the mutant yeast strain and screened for growth on ethanol and glucose plates. All yeast cells grew on glucose plates indicating the cells were viable (Table 4-2). However, yeast cells with Thr444Ala, Phe545Ala or Phe545Tyr CAT grew on ethanol media while cells with the substitution of

PAGE 75

64 Tyr431 by either alanine or phenylalanine did not grow with ethanol. Kinetic analysis of Tyr431 illustrates how essential hydrogen bonding is to L-carnitine binding by producing YNEC YND Mutant hCAT His322Ala Mutant hCAT hCAT hCAT His322Ala Figure 4-5. Human carnitine acetyltransferase complements mutant S. cerevisiae. Shows a replica plate of yeast media with glucose as carbon source (YND) and yeast media with ethanol as carbon source (YNEC). Growth of the mutant yeast strain with endogenous CAT that has been disrupted is shown as the Mutant section of the YND and YNEC plates. The hCAT section of the YND and YNEC plates shows growth of the mutant yeast strain with human CAT introduced. The His322Ala section of the YND and YNEC plates shows growth of the mutant yeast strain with mutant human CAT, containing the substitution of the active site histidine with alanine, introduced. the most defective mutants (Table 4-1). The genetic screen furthers this point by showing that the enzyme is not active enough to provide a mechanism for the transport of acetyl-CoA to the mitochondria for energy production, thus inhibiting growth of the cells. Although Thr444 and Phe545 are also essential to L-carnitine binding, mutation at these positions resulted in less defective mutants (Table 4-1). The Km value towards L-carnitine for the Thr444Ala, Phe545Ala, and Phe545Tyr forms of CAT were increased 68, 18, and 1.3-fold, respectively while the acetyl-CoA Km values for Thr444Ala, Phe545Ala, and Phe545Tyr CAT were closer to the wild type form, only increasing 4.4, 4.2, and

PAGE 76

65 1.5-fold, respectively. However, despite these altered kinetic parameters the mutant forms of CAT were able to complement the mutant yeast and allow growth on ethanol media (Table 4-2). Table 4-2. Growth of mutant yeast cells with glucose and ethanol media. Glucose Ethanol Tyr431Ala Yes No Tyr431Phe Yes No Thr444Ala Yes Yes Phe545Ala Yes Yes Phe545Tyr Yes Yes Characterization of Glu326: Glu326 is located in close proximity to the active site histidine and coordinates with a water molecule that is involved with binding the carboxylate group of L-carnitine (Figure 4-5). Due to its close proximity to His322, it is thought that this negatively charged residue helps to stabilize the protonated histidine during catalysis (Figure 4-6). Figure 4-5. Three-dimensional representation of the active site of CAT. Shows L-carnitine, His322 and Glu326. Green and yellow dashes represent hydrogen bonding potential.

PAGE 77

66 His 322NHN SCoA R O N+ O O O H N+ O O O R O N+ O O O R SCoA O His 322NHNH+ His 322NHN HSCoA Glu326O O Glu326O O Glu326O O Figure 4-6. Schematic of the reaction mechanism of CAT. The nitrogen of His322 deprotonates L-carnitine, which facilitates the nucleophilic attack on acyl-CoA, forming a tetrahedral intermediate. The intermediate falls apart releasing coenzyme A and acyl-L-carnitine. It appears as though the carboxylate group of Glu326 stabilizes the quaternary nitrogen of histidine during the transition state. To evaluate this residue, Glu326 was replaced by amino acids that would alter the physical properties of this residue: specifically, serine, asparagine, histidine, lysine, glutamine, alanine, and aspartate. The mutant forms of the enzyme were functionally and biologically characterized using the enzymatic assay and genetic selection system. As shown in Table 4-3, the only mutant form of the enzyme enzymatically active was Asp326 CAT. Replacement of glutamate by an aspartate retains the negative charge, but decreases the side chain length by one carbon atom. The other amino acid substitutions eliminated the negative charge. As theorized by the three-dimensional structure and reaction mechanism, these results provide additional support that the negative charge is necessary during the transition state for stabilizing the protonated histidine. To determine the biological role of this residue the mutant forms of the enzyme were introduced to the mutant yeast strain and screened for growth on ethanol and glucose media. Interestingly, the mutant yeast strain was able to grow with the Asp326 form of CAT, but not with any of the other substitutions (Table 4-3). Apparently this

PAGE 78

67 mutant form of CAT was the only form able to complement the mutant yeast strain. This means that the Glu326Asp CAT was active enough to provide energy to sustain growth while the other mutant forms of CAT were not. These results agree with the functional characterization studies and further support the role of this residue. Table 4-3. Enzymatic activity and growth with glucose and ethanol of Glu326 mutants. Growth Activity Glucose Ethanol Glu326Gln Yes No None Glu326Asp Yes Yes Yes Glu326Asn Yes No None Glu326His Yes No None Glu326Lys Yes No None Conclusions The structure of the complex of CAT and L-carnitine reveals several key residues of CAT that recognize and interact with the substrate L-carnitine. To further evaluate the role of these residues, they were functionally and biologically characterized. Substitution of four active site residues produced CAT mutants with altered kinetic properties that support their functional role in binding L-carnitine and catalysis. Specifically, Tyr431, Thr444, and Arg497 form a binding pocket for the carboxylate group while Phe545 interacts with the quaternary amine head group of L-carnitine. This study also provides important information about L-carnitine binding and the catalytic mechanism of other carnitine acyltransferases since His322, Tyr431, Thr444, Arg497, and Phe545 are highly conserved among the carnitine acyltransferase family.

PAGE 79

CHAPTER 5 STRUCTURE OF CHOLINE ACETYLTRANSFERASE The goal of the second specific aim is to further understand the L-carnitine/choline acyltransferase family by solving the structure of ChAT. Given that ChAT and CAT have a high sequence homology (~ 40%) and the active site residues of CAT are shared with ChAT, it can be presumed that these two enzymes will share several structural features. The x-ray crystal structure of ChAT can confirm these estimations and can also be used to understand several disease causing mutations. Introduction As previously discussed, ChAT is the fourth member of the L-carnitine/choline acyltransferase family and catalyzes the synthesis of the neurotransmitter acetylcholine (ACh). ACh was the first neurotransmitter to be reported [97] and has since been shown to play a pivotal role in fundamental brain processes such as learning, memory, and sleep [98-100]. ACh functions in the cholinergic neurons of the peripheral and central nervous systems. In the peripheral nervous system ACh stimulates muscle contraction and in the central nervous system ACh facilitates learning and short-term memory formation. In 1943 ChAT was identified as the enzyme responsible for the synthesis of ACh [101]. Similar to other members of the L-carnitine/choline acyltransferase family, ChAT catalyzes the reversible transfer of an acetyl group from CoA to choline (Figure 5-1). Although ChAT has been known as the enzyme responsible for the synthesis of ACh for more than 60 years, its mechanism of action is still unclear [47]. Comparably to studies with CAT, several inactivation and modification studies of ChAT have suggested 68

PAGE 80

69 several histidine and arginine residues that are critical for the enzyme to function (Table 5-1) [102-108]. N+OHCH3CH3CH3 +CoA SOCH3 N+OCH3CH3CH3CH3O +CoA SH Figure 5-1. Schematic of ChAT reaction mechanism. Site-directed mutagenesis studies in Drosophila ChAT have lead to the proposal that a highly conserved histidine residue, histidine 426, (histidine 334 in the rat form of ChAT) is essential for activity and may serve as an acid/base catalyst [109]. Additional mutagenesis studies in rat ChAT have also suggested that another highly conserved residue, arginine 452, interacts with acetyl-CoA [63]. Interestingly, an equivalent mutation in human ChAT, arginine 560 to histidine, has also been shown to significantly reduce the affinity for both substrates, choline and acetyl-CoA [56]. The ChAT gene appears to have multiple splice variants that result in different mRNA species. Three non-coding exons, R, N and M, have been identified upstream of the coding exon 4 in rat, mouse and human. However, the difference between these species is that the human form of exon M contains a sequence that can be translated to amino acids [47]. At least seven, five and four different mRNA species have been isolated from the mouse, rat and human, respectively [47]. These mRNA species are formed from different combinations of the non-coding exons. Since the exons are non-coding in the mouse and rat, a single form of ChAT protein is generated in these species. In contrast, the human M exon can be translated to amino acids and thus generates two forms of ChAT in humans, small and large.

PAGE 81

70 Table 5-1. Reported mutations of choline acetytlransferase.* Animal species Mutation Corresponding residue in rChAT Effect Site-Directed Mutations Drosophila His393Leu His301 Inactive His436Leu His334 Inactive His426Gln His334 Inactive Rat Arg452Ala Increase Km for CoA (12-fold) and ACh (8-fold) and increase kcat (7-fold) Arg452Gln Increase Km for CoA (12-fold) and ACh (8-fold) and increase kcat (7-fold) Arg452Glu Increase Km for CoA (54-fold) and ACh (60-fold) and increase kcat (15-fold) Arg453Glu Increase Km for CoA (7-fold) and ACh (12-fold) and increase kcat (5-fold) Arg452Gln/ Arg453Glu Increase Km for CoA (170-fold) and ACh (70-fold) and increase kcat (17-fold) Disease Cauing Mutations Human Glu441Lys Glu331 Low exression, no catalytic activity Arg560His Arg452 Increase in Km toward acetyl-CoA and choline Temperature Sensitive Mutations Drosophila Met403Lys Leu318 Decrease activity *Adapted from information reviewed in [56, 63, 109, 110]. The small form of ChAT in humans is the primary form observed and is composed of 630 amino acids with a molecular weight of 67 68 kDa. The observed molecular weight is less than the calculated molecular weight, most likely due to post-translational modification or proteolysis during enzyme purification [47]. The large form of ChAT is

PAGE 82

71 composed of 748 amino acids and contains a 118 amino acid N-terminal extension. It is not clear whether this form is actually synthesized in vivo [47]. ChAT is localized in the cholinergic nerve terminals (Figure 1-6) in two forms, soluble (80-90% of total enzyme activity) and membrane bound forms (10-20%) [47, 111]. The significance of the membrane association is unknown; however, phosphorylation appears to be required for membrane association [112]. ChAT is thought to exist as an extremely low abundance and monomeric globular protein in nature, with its mechanism of action still unknown. Decreases in ChAT activity are an indicator of cholinergic neuron damage and have been observed with patients who have neurodegenerative disorders, such as Alzheimers disease, Huntingtons disease and amyotrophic lateral sclerosis. Decreases in ChAT activity have also been correlated with schizophrenia, Rett syndrome and sudden infant death syndrome [47]. Recently, direct sequencing of the ChAT gene from five patients, exhibiting congenital myasthenic syndrome with episodic apnea (CMS-EA), revealed 10 recessive mutations, which result in severe muscle weakness and respiratory insufficiency. Kinetic analysis of nine of these mutants (the tenth was a frameshift null mutation), demonstrated that one mutant (Glu441Lys; residue 332 in rChAT) lacked catalytic activity while the others (Leu210Pro, Pro211Ala, Ile305Thr, Arg420Cys, Arg482Gly, Ser498Leu, Val506Leu, and Arg560His; residues 101, 102, 196, 311, 373, 388, 397, and 452 in rChAT, respectively) had significantly impaired catalytic efficiencies [56]. Understanding the basis of these mutations, however, has been hindered since no three dimensional structure was available.

PAGE 83

72 Recently, the crystal structures of human and mouse carnitine acetyltransferase have been reported [64, 65]. These enzymes show significant amino acid sequence homology (~ 40% identity) with ChAT, and therefore can be used for the molecular replacement solution of the structure of the rat form of ChAT (rChAT). The structure, together with the modeled substrate (choline and acetyl-CoA) complexes has given rational to understanding the mechanisms of catalysis of the enzyme and understanding to disease causing mutations of ChAT. Materials and Methods The experimental procedures involved in this project were described in Chapter 2. Briefly, recombinant rat ChAT was expressed in bacterial cells with a histidine tag and then purified using a column with Ni-NTA-agarose. The purified enzyme was then crystallized and the three-dimensional structure was determined. Results and Discussion Structure Determination and Overall Structure of the Enzyme RChAT was over-expressed in Escherichia coli, purified by affinity chromatography and crystallized (Figure 5-2). A single crystal was cryoprotected by a 60 second soak in a reservoir solution containing 15% PEG 8000 and 40% glycerol. The crystal was then mounted in a thin fiber nylon loop (Hampton Research) and flash frozen prior to data collection at 100 K. The diffraction data were collected at the Cornell High Energy Synchrotron Source (CHESS) using the ADSC Quantum 4 CCD detector system. Data were collected using a 0.2 mm collimator, detector-to-crystal distance of 150 mm, 1 oscillation angle with an exposure time of 120 seconds for each image. Reflection intensities were indexed and evaluated with DENZO and scaled using SCALEPACK [80].

PAGE 84

73 Figure 5-2. Photograph of recombinant rat choline acetyltransferase crystals. Approximate dimensions are 0.4 x 0.2 x 0.2 mm. A total of 70o of data were collected (70 images) from a single crystal. Figure 5-3 shows a typical 1.0 oscillation diffraction image. The rChAT crystal was shown to be orthorhombic P212121 with unit cell parameters a = 139.0, b = 77.7, c = 59.7 The structure was refined to 1.55 resolution [113]. The structure of ChAT was solved by molecular replacement procedures [114], with the coordinates of human peroxisomal CAT (with the side-chains of all non-glycine residues truncated to alanines) as a search model [64] using the CNS program package [81]. A sequence alignment between rChAT and CAT showed there was ~ 40% amino acid identity indicating the overall structural fold of the two enzymes would be similar. The structure consists of ~40 % helix (20 helices: H1-H20) and 15% strand (16 strands: S1-S16) arranged into two domains, I (residues 89-400) and II (residues 20-88 and 401-616; Figure 4.4). These domains are interconnected and create a tunnel, approximately 16 in diameter that passes through the center of the enzyme, which forms the active site. Domain I is composed of a six-stranded mixed -sheet (S1, S8, S7, S6, S4, and S5) that is interconnected on both flanks by eight -helices (Figure 5-4). Domain II also

PAGE 85

74 Figure 5-3. X-ray diffraction pattern of rat choline acetyltransferase crystal. Data were collected on the F1 beam line ( = 0.9504 ) with ADSC Quantum 4 CCD detector system at the Cornell High Energy Synchrotron Source. This data was collected with a crystal-to-detector distance of 150 mm with an exposure time of 120 seconds/image and an oscillation range of 1.0 at 100 K. The black concentric rings show the demarcation of the 5.0, 3.5, 2.0 resolution diffraction shells. The highest observed diffraction extended beyond 1.55 resolution. consists of six-mixed -sheets (S10, S16, S15, S13, S11, and S12) that are connected with 11 -helices (seven from the C-terminal and four from the N-terminal). Helices H1 and H2 generate an anti-parallel bundle that elongates to an extended structure composed of helices H3 and H4, which crosses over to domain I to form the front entrance (choline) of the tunnel. In domain I residues 322-335 generate anti-parallel -sheets (S7 and S8) with the catalytic His334 located at the end of strand S8 positioned at the center of the active site tunnel (Figure 5-4). This turn (residues 333-358) extends to form helix H12

PAGE 86

75 S16 H3 H1 H2H16H17C-termH20H15 H14S12H18H19 S11S13S14S15S10H4H12S3S1S8H6S2H9 H10 S6 S7 H5 S4 H7 H8S5 S9 His334CoA CholineH11N-termI IIH13S16 Figure 5-4. Overall Structure of Rat Choline Acetyltransferase. Ribbon diagram representation depicting -helices H1-H20 (red) and -strands S1-S16 (blue). Also shown in ball and stick representation the active site His334 and docked substrates choline and coenzyme A. that lines one side of the active site tunnel. The long helix H13 is arranged in front of the -sheet of domain I and connects to -strand S10 that links domain I to domain II. Strand S10 leads into helix H14 that is parallel with the -sheet of domain II. The -sheet of domain II facilitates the flipping out of strand S14 (between S13 and S15), which forms an anti-parallel interaction with the strand S1 of domain I. Five helices, H15-H20, are between strands S11 and S16, generate a supporting scaffold that holds domain II -sheet open face towards the active site of the enzyme. The -strand S16 connects with the

PAGE 87

76 C-terminal helix H20 via a loop that terminates domain II. The two domains of rChAT adopt a similar topology with each other in the absence of any recognizable amino acid sequence homology (~ 11% identity) between them. The overall surface area of the molecule is 25,000 2. Out of the 644 amino acids, 89 (mostly hydrophobic residues) are buried at the interface between the two domains (~4000 2). In domain I residues Ala109, Ile111, and Glu347 have hydrogen bond interactions with domain II amino acids Thr555, Met556, Phe559, and Cys561. The carbonyl oxygens of Glu347 hydrogen bond with the side chain of Thr555 (3.1, 2.8 ) and the main-chain nitrogen of Met556 (2.8 ). Both residues, Ala109 and Ile111, have hydrogen bond interactions with Cys561 (3.0 ) and Phe559 (2.8 ) main chain atoms. These interactions are predominantly main-chain and therefore there is no evolutionary pressure to conserve them between species (Figure 5-5). Comparison of CAT-L-Carnitine with Free CAT The amino acid sequence of rChAT is highly homologous with ChATs from other species and has 95, 86, and 85% identity with mouse, human, and pig respectively. In addition, the amino acid sequence of rChAT has ~ 40% identity compared to human and mouse CAT [64, 65] and shares a common structure with a RMS deviation of 1.30 and 1.25 respectively, for approximately 570 C atoms. The rChAT structure has 2 additional -strands (S9 and S10) and an additional helix (H14) compared to the structures of human and mouse forms of CAT. Most of the main-chain structural variation (3.0 to 6.0 ) occurs between residues 209-216 and 233-240 of rChAT compared to human and mouse CAT.

PAGE 88

77 Figure 5-5. Sequence alignment of choline acetyltransferases. Shows the amino acid sequence alignment of rat, human, mouse, pig, D. melanogaster, and C. elegans ChAT. The amino acid numbering and secondary structural elements

PAGE 89

78 are based on the structure of rat ChAT. Strands S1-S16 are represented by blue arrows while helices H1-H20 are depicted by red spirals. Identical and conserved residues are indicated by red and yellow blocks, respectively. Residues mutated using site-directed mutagenesis are indicated by purple triangles [63, 95, 109]. Residues mutated in patients with congenital myasthenic syndrome with episodic apnea (CMS) are shown by green triangles [56]. A blue star indicates the residue, Arg312, that was mutated by site-directed mutagenesis and shown to be mutated in patients with CMS. Interestingly, E2p, an enzyme that catalyzes a similar reaction to that of rChAT, in that it transfers an acetyl group from acetyl-CoA to an organic substrate, shares an 8% sequence identity with rChAT. When E2p is structurally aligned with domain I of rChAT it has an RMS deviation of 1.87 for 117 C atoms [115]. Active Site The catalytic center of rChAT is a 16 long tunnel, formed between the interface of domains I and II. Chemical modification, inactivation, and mutagenesis studies have identified the catalytic histidine of rChAT as His334, which is located at the center of this tunnel between -strand S8 and helix H8 in domain I (Figure 5-4) [18, 116]. This observation is consistent with sequence homology comparison with the known catalytic histidine residues of other acyl group transfer enzymes including His322 in human CAT [64], His343 in mouse CAT [65], His610 in E2p [115], and His195 in chloramphenicol transferase [117]. In rChAT, His334 assumes an unusual conformational arrangement (1 = -146; 2 = -49) to form a hydrogen bond (3.0 ) interaction between the imidazole nitrogen N1 and its main chain carbonyl oxygen. In the crystal structure of human and mouse CAT, E2p, and chloramphenicol transferase, the catalytic histidine also shows a similar conformation [64, 65, 115, 117]. Interestingly, this conformation of His334 allows the N3

PAGE 90

79 nitrogen to point towards the hydroxyl group of choline to extract a proton, as shown in Figure 5-6 with His322 of human CAT and L-carnitine. His322 L-Carnitine L-Carnitine Figure 5-6. Three-dimensional representation of the conformation arrangement of the active site histidine, His322 in human CAT. Dashed line (yellow) represents hydrogen bond interaction between imidazole nitrogen and backbone carbonyl group (green) represents hydrogen bond interaction between N3 of His322 and hydroxyl group of L-carnitine. These interactions are common with mouse CAT, E2p, chloramphenicol transferase and ChAT. A highly conserved glutamate, Glu338, interacts with His334 and may be important for catalytic activity. The carbonyl oxygen O1 of Glu338 is hydrogen bonded (3.7 ) with the N3 of the active site His334 and the O2 of Glu338 hydrogen bonds (2.8 ) with Arg458. Most likely these interactions contribute to potentiate the catalytic activity of His334 and also stabilize the binding of choline. Choline Binding Site The docking of choline into the active site of rChAT indicates five residues (Tyr95, His334, Ser548, Tyr562, and Val565) that may be involved in its binding (Figure 5-7, Table 5-2). The hydroxyl group of choline interacts with His334 and forms a hydrogen bond (3.0 ). This hydrogen bond formation supports the catalytic role of His334. Tyr95 is also

PAGE 91

80 within hydrogen bonding distance to the hydroxyl group of choline (2.0 ). The quaternary amine head group of choline is positioned close to the side chain of Tyr562, separated by 4.6 and has hydrophobic interactions with the side chain residues Ser548 and Val565 (3.0 and 3.2 respectively) (Figure 5-7, Table 5-2). Interestingly, in other highly conserved members of the L-carnitine/choline acyltransferase family, CAT, the quaternary amine of L-carnitine appears to be positioned close to a phenylalanine residue, Phe545 and Phe566 respectively. However, in the case of rChAT the quaternary amine of choline interacts with a tyrosine residue (Tyr562) [64, 65]. In CAT, the highly conserved residues within carnitine acyltransferases, Thr444 and Arg497 (Val459 and Asn514 in rChAT) play critical roles by interacting with the carboxylate group of L-carnitine [65, 85]. Since choline does not have a carboxylate group at this position, these residues are not conserved in rChAT. All members of the L-carnitine/choline acyltransferase family contain a conserved Ser-Thr-Ser (STS) motif near the carboxy terminal: Ser548, Thr549, and Ser550 in rChAT [84]. The first serine of this motif, Ser548, is positioned close to the quaternary amine group of choline and may play a critical role in transition state stabilization during catalysis. Coenzyme A Binding Site The CoA binding site was modeled on the opposite side of His334 from the choline binding site. The strands S11 and S12 of domain II are separated from each other and create a space that can accommodate the pantothenic arm of CoA, which extends towards the active site. Highly conserved residues in L-carnitine/choline acyltransferases surround the CoA binding site at the C-terminal end of the enzyme (Table 5-3).

PAGE 92

81 L y s413L y s417CoAH14H15S11H12S8 S1 S14Asp424Glu447Gln551 Glu338T y r95His334CholineT y r446Ser548Tyr562Val565I II Figure 5-7. Close up view of the modeled binding site of choline and CoA. The C trace of domain I (blue) and domain II (red). RChAT residues putatively involved in binding shown in ball and stick. Red dash lines indicate possible hydrogen bond interactions Residues Lys413 and Lys417 may interact with the 3-phosphate group of CoA and a clustered group of residues (Asp424, Glu447, and Gln551) possibly interact with the pantothenic arm of CoA (Figure 5-7, Table 5-3). The carboxyl oxygens of Asp424 are hydrogen bonded with Gln551 (3.0 ) and Glu447 (2.6 ). This would imply at least some effect on the protonation state of at least one of the acidic residues in the active site.

PAGE 93

82 Table 5-2. Choline substrate binding sites for rat choline acetyltransferase (rChAT), human peroxisomal L-carnitine acetyltransferase (hCAT), and mouse L-carnitine acetyltransferase (mCAT). Atom rChAT hCAT** mCAT*** O3 His334 (3.02) His322 (2.76) His434 (2.68) N(CH3)3 Ser548 (3.04) Ser533 (3.61) Ser552 (3.62) Tyr562 (4.6) Phe545 (3.6) Phe545 (3.96) Val565 (3.16) Val548 (3.83) Val569 (3.86) O1 Tyr431 (2.55) Tyr452 (2.59) O2 Thr444 (2.82) Thr465 (2.63) *Residue (Distance ()) ** Interactions taken from the crystal structure of hCAT, PDB number, 1S5O [85]. ***Interactions taken from the crystal structure of mCAT, PDB number, 1NDF [65]. Table 5-3. Coenzyme A substrate binding sites for rat choline acetyltransferase (rChAT), human peroxisomal L-carnitine acetyltransferase (hCAT), and mouse L-carnitine acetyltransferase (mCAT). Atom rChAT hCAT** mCAT*** P3(O9) Lys413 (3.23) Lys398 (3.27) Lys419 (3.64) Lys417 (2.72) Lys402 (2.67) Lys423 (3.58) *Residue (Distance ()) ** Interactions taken from the crystal structure of hCAT, PDB number, 1S5O [85]. ***Interactions taken from the crystal structure of mCAT, PDB number, 1NDF [65]. Previous chemical modification, inactivation, and site-directed mutagenesis studies have suggested that Arg452 of rChAT interacts with the 3-phosphate group of CoA to stabilize CoA binding [63, 105]. Although substitution of this residue by an alanine resulted in a 7-12 fold increase in Km for both CoA and acetyl-CoA as well as an increase in the kcat value [63], based on observations from the structure of rChAT it appears that this residue is too far from the 3-phosphate group of CoA (~ 11 ) to form an interaction. Perhaps the mutation of Arg452 with alanine disrupts the overall structure of the active site or the stability of the enzyme to cause such a decreased affinity towards the substrates.

PAGE 94

83 Disease Associated Mutations Congenital myasthenic syndrome (CMS) is an inherited disease that affects the neuromuscular junction and causes muscle weakness from birth or early childhood. The neuromuscular junction is the junction between a terminal motor neuron and a muscle fiber. Nerve impulses travel down the motor neuron to cause the muscle fiber to contract. Once the action potential reaches the neuron terminal ACh, synthesized by ChAT and stored in synaptic vesicles, is released into the synaptic cleft where it binds to its receptors to elicit neurotransmission. The action of ACh is terminated by acetylcholinesterase, which is the target of several Alzheimers disease therapeutics and chemical warfare. CMS is classified into three categories based on the site of the mutation. Pre-synaptic defects, which account for 8% of cases, are related to decreased production or release of ACh. Synaptic basal lamina associated defects are responsible for 16% of cases and are due to endplate acetylcholineterase deficiency that leads to excessive muscle stimulation and ultimately damages the muscle. The third category accounts for the largest proportion of cases and is a result of deficiency or mechanical changes of the ACh receptor. Interestingly, the cause of the defect determines the symptoms. With the pre-synaptic form of the disorder patients experience dyspnea (shortness of breath) and bulbar (muscles of the face, jaw, pharynx, larynx and tongue) weakness. The distinguishing feature of this form of CMS is that the dyspnea and bulbar weakness is sudden and can lead to apnea or temporary cessation of breathing. This symptom is precipitated by infection, fever or excitement. The only treatment of this form of CMS is

PAGE 95

84 acetylcholinesterase inhibitors. However, recent studies on the molecular basis of the disease have revealed mutations of the ChAT encoding gene. An analysis of the ChAT encoding gene from five patients with CMS-EA revealed recessive mutations, one (523insCC) null mutation, three mutations (Ile305Thr, Arg420Cys, and Glu441Lys) that reduce ChAT expression, one mutant (Glu441Lys) devoid of catalytic activity and eight mutants (Leu210Pro, Pro211Ala, Ile305Thr, Arg420Cys, Ser498Leu, Val506Leu, and Arg560His) that have decreased catalytic efficiencies towards CoA (Table 5-1, Figures 5-5 & 5-7) [56]. Further mutational analysis of the ChAT encoding gene from three patients with CMS-EA from two Turkish families shows an additional mutation (Ile336Thr) and studies from five patients from three independent families shows three further mutations (Val194Leu, Arg548Stop, and Ser694Cys) [118, 119]. Figure 5-8 shows that most of these residues are located in the periphery of the structure, away from the active site tunnel and perhaps are important for stability or maintaining the conformation of the active site. The most severe mutation reported, Glu441Lys, drastically reduces ChAT expression in COS cells (29% compared to wild type) and the mutant form of the enzyme is devoid of catalytic activity [56]. This conserved residue corresponds to Glu333 in rChAT and is located in strand S8, adjacent to the catalytic histidine (Figure 5-5). Although this residue points in the opposite direction of the active site and appears to not be involved directly in the catalytic mechanism of the enzyme, mutation to a lysine could introduce interactions with a nearby arginine, Arg250, (4.0 ). These interactions could disrupt the active site or the orientation of the active site histidine and thus eliminate catalytic activity.

PAGE 96

85 hpCAT CoA CarnitiI II His334 274-279 rCHA Choline/ N C 233-241 209-216 Figure 5-8. Structural comparison of rat choline acetyltransferase (rChAT) and human peroxisomal carnitine acetyltransferase (hCAT) and mutations in human ChAT that cause myasthenic syndrome associated with episodic apnea (CMS-EA). Coil diagram shows the superimposition of rChAT (red) onto hCAT (blue). Arrows show structural variation (C RMS deviation > 2.0 ), residues 209-216 and 233-241, between rChAT and hCAT. Shown in ball and stick His334, choline/L-carnitine, and coenzyme A. Residues mutated in patients with CMS-EA are shown by green spheres [56]. Residues mutated using site directed mutagenesis are indicated by purple spheres [63, 95, 109]. A blue sphere indicates the residue, Arg312, that was mutated by site directed mutagenesis and shown to be mutated in patients with CMS-EA. (Refer to text, Table 5-1, and Figure 5-5). Another interesting mutation is Arg560His. This mutation causes a decreased affinity towards both acetyl-CoA and choline [56]. Arg560 is absolutely conserved and corresponds to Arg452 in rChAT (Figure 5-5). As previously discussed, chemical

PAGE 97

86 modification, inactivation and site-directed mutagenesis studies have shown that this residue interacts with the 3-phosphate group of CoA [63, 105]. However, based on the structure, Arg452 does not interact with either substrate. Therefore, it is quite perplexing that a mutation of this residue can impair catalytic activity to such a degree that it causes a disease state. Based on the structure, it appears that a histidine located at position 452, could interact with Gln164 (~2.5 ), which could disrupt the conformation of the active site histidine. A histidine residue at position 452 could also interact with Glu338 (~4.0 ), which, as previously discussed, forms a hydrogen bond interaction with the active site histidine. The other mutations, Leu102Pro, Pro103Ala, Ile197Thr, Arg312Cys, Arg374Gly, Ser390Leu, and Val398Leu, have been shown to cause a decrease in the affinity for acetyl-CoA and therefore decrease catalytic efficiency when compared to wild type [56]. Temperature Sensitive Mutations The structure of rChAT can also be used to understand the temperature sensitive mutations in Drosophila that were generated to determine the role of the neurotransmitter acetylcholine in development and mediating behavioral function in the brain [120]. These mutations (Chats1 and Chats2) show a conditional temperature sensitive phenotype in which at increased temperature, 30C, ChAT activity decreases. Drosophila expressing either mutation become paralyzed and subsequently die when incubated at 30C [120]. At the higher temperature it was shown that the mutant phenotype appears more rapidly with the Chats2 mutant, indicating that the Chats2 mutation is more severe than the Chats1 [120]. Recently these mutations have been identified as Met403Lys for Chats1, which corresponds to Leu318 in rChAT, and Arg397His for Chats2, which corresponds to Arg312 in rChAT (Figure 5-5) [121]. Arg312 is absolutely conserved within all known ChATs while

PAGE 98

87 Leu318 is not conserved. In fact, this residue is a leucine residue in rat, human, mouse and pig, an isoleucine in C. elegans, and a methionine in the Drosophila ChAT. Both of these residues are close to the active site histidine. The replacement of Leu318 by lysine appears to affect the overall stability of the enzyme since it appears that this mutation does not introduce deleterious interactions with surrounding residues. In contrast, substitution of Arg312 by histidine may change the choline binding site. It appears that a histidine residue located at position 312 may interact with Glu333, which is located adjacent to the active site histidine. As previously discussed, this residue faces the opposite direction of the active site. However, it may be critical in formation of the choline binding pocket, such that an unfavorable interaction will disrupt choline binding. Conclusion The structure of rChAT has been determined to a 1.55 resolution and exhibits conserved structural features found in human and mouse CAT [64, 65]. RChAT is a monomeric protein containing two domains that interconnect to form an active site tunnel. Within the active site the catalytic residue His334 is located at the center of the tunnel, suggesting a common catalytic mechanism for the entire L-carnitine/choline acyltransferase family. The substrates, choline and acetyl-CoA, have been modeled in the active site based on the structure of mouse CAT, and reveal several residues that may be involved in substrate recognition and ligand interactions and provide insight into the molecular basis of disease causing mutations.

PAGE 99

CHAPTER 6 CONCLUSIONS The carnitine/choline acyltransferase family plays an important role in fat metabolism by facilitating the transfer of acyl groups across intracellular membranes. These enzymes catalyze the reversible transfer of acyl groups between CoA and L-carnitine. There are three known classes of carnitine acyltransferases that differ in their acyl group specificity, subcellular location, tissue distribution and physiological function. Of these carnitine acetyltransferase (CAT) is of particular interest. CAT has a substrate preference for short chain (2-6 carbon atoms) acyl groups and is found in the mitochondrial matrix, the endoplasmic reticulum and peroxisomes. Chemical modification and site-directed mutagenesis experiments have identified an active site histidine that acts as a general base to catalyze the reaction. However, in spite of its importance, fundamental questions remain about the structure of CAT. The structure of CAT with its substrate L-carnitine was determined and reveals several residues involved in catalysis. The structure supports the putative mechanism of the enzyme and identified the active site histidine as His322. His322 plays an essential role in catalysis by abstracting a proton from the -hydroxyl group of L-carnitine or the thiol group of CoA depending on the direction of the reaction. In addition, specific L-carnitine binding residues were identified. It was shown that three residues, Tyr431, Thr444 and Arg497, interact with the carboxylate group of L-carnitine while two residues, Phe545 and Val548, interact with the quaternary amine 88

PAGE 100

89 group. These residues are highly conserved within the carnitine/choline acyltransferase family. To confirm their role in catalysis these residues were functionally and biologically characterized. Functional characterization was carried out by a spectrophotometric enzymatic assay. A novel genetic assay with Saccharomyces cerevisiae as the model organism was developed to biologically characterize these residues. In this system an endogenous yeast CAT gene is disrupted rendering the cells incapable of using short chain carbon groups as an energy source and thus unable to grow on ethanol media. Human CAT complements the mutant yeast strain and allows growth on short chain carbon sources while an inactive form of CAT is unable to complement the yeast strain. Substitution of the four residues (Tyr431, Thr444, Arg497 and Phe545) produced mutant forms of the enzyme with altered kinetic parameters. The mutant forms of the enzyme were also unable to complement the mutant yeast strain. Taken together these studies support the role of these residues in binding L-carnitine and catalysis. In addition to these studies, the role of Glu326 was evaluated using the enzymatic and biological assays. This residue is located in close proximity to the active site histidine and is thought to stabilize the protonated histidine during catalysis. Replacement of Glu326 by other amino acids, except by aspartate, produced mutant forms of the enzyme that were catalytically and biologically inactive. This provides additional support that the negative charge is necessary during the transition state for stabilization. The structure of the fourth member of the carnitine/choline acyltransferase family, choline acetyltransferase (ChAT), was also determined. It was found that rat ChAT and human CAT share several structural features; both enzymes are composed of two

PAGE 101

90 domains that form an active site tunnel. Similar to CAT, His334 is located in the center of the tunnel, suggesting a common mechanism for the entire carnitine/choline acyltransferase family. The substrates, choline and acetyl-CoA, were modeled in the active site and reveal several residues that may be involved in substrate recognition: Tyr95, Ser548, Tyr563, Val565 and Lys413, Lys417, Asp424, Glu447, Gln551. The structure of ChAT has also provided insight to the molecular basis of the disease congenital myasthenic syndrome (CMS). Previous studies have revealed that mutations, which reduce ChAT expression, eliminate catalytic activity and decrease catalytic efficiencies towards CoA from five patients with CMS. Based on the structure of rChAT, the position of these residues could be determined and the possible detrimental interactions by the mutations could be ascertained. In conclusion, these studies have helped to further understand this enzyme family. The information gleaned from structural studies of CAT and ChAT have identified several key residues which, could be carried over to other members of the enzyme family, COT or the CPTs. CPT I plays an essential role in fatty acid metabolism, but is also highly regulated by an endogenous substrate analogue. Understanding of the molecular basis of this relationship is key to developing an exogenous modulator of CPT I activity.

PAGE 102

LIST OF REFERENCES 1. U.S. Department of Health and Human Services Centers for Disease Control and Prevention National Center for Chronic Disease Prevention and Health Promotion Division of Nutrition and Physical Activity Overweight & Obesity. http://www.cdc.gov/nccdphp/dnpa/obesity/trend (Last accessed: April 25, 2004) 2. Damcott CM, Sack P, Shuldiner AR. (2003) The genetics of obesity. Endocrinol Metab Clin North Am, 32, 761-86. 3. Amri EZ, Ailhaud G, Grimaldi PA. (1994) Fatty acids as signal transducing molecules: involvement in the differentiation of preadipose to adipose cells. J Lipid Res, 35, 930-7. 4. Sellmayer A, Danesch U, Weber PC. (1996) Effects of different polyunsaturated fatty acids on growth-related early gene expression and cell growth. Lipids, 31 Suppl, S37-40. 5. Uauy R, Mena P, Valenzuela A. (1999) Essential fatty acids as determinants of lipid requirements in infants, children and adults. Eur J Clin Nutr, 53 Suppl 1, S66-77. 6. Dutta-Roy AK, Kahn NN, Sinha AK. (1991) Interaction of receptors for prostaglandin E1/prostacyclin and insulin in human erythrocytes and platelets. Life Sci, 49, 1129-39. 7. Dutta-Roy AK. (1994) Insulin mediated processes in platelets, erythrocytes and monocytes/macrophages: effects of essential fatty acid metabolism. Prostaglandins Leukot Essent Fatty Acids, 51, 385-99. 8. Kerner J, Hoppel C. (2000) Fatty acid import into mitochondria. Biochim Biophys Acta, 1486, 1-17. 9. Felig P, Wahren J. (1975) Fuel homeostasis in exercise. N Engl J Med, 293, 1078-84. 10. Raclot T. (2003) Selective mobilization of fatty acids from adipose tissue triacylglycerols. Prog Lipid Res, 42, 257-88. 11. Kunau WH, Dommes V, Schulz H. (1995) beta-oxidation of fatty acids in mitochondria, peroxisomes, and bacteria: a century of continued progress. Prog Lipid Res, 34, 267-342. 91

PAGE 103

92 12. Pownall HJ, Hamilton JA. (2003) Energy translocation across cell membranes and membrane models. Acta Physiol Scand, 178, 357-65. 13. Hamilton JA, Guo W, Kamp F. (2002) Mechanism of cellular uptake of long-chain fatty acids: Do we need cellular proteins? Mol Cell Biochem, 239, 17-23. 14. Glatz JF, Luiken JJ, van Bilsen M, van der Vusse GJ. (2002) Cellular lipid binding proteins as facilitators and regulators of lipid metabolism. Mol Cell Biochem, 239, 3-7. 15. van der Vusse GJ, van Bilsen M, Glatz JF, Hasselbaink DM, Luiken JJ. (2002) Critical steps in cellular fatty acid uptake and utilization. Mol Cell Biochem, 239, 9-15. 16. Bonen A, Luiken JJ, Glatz JF. (2002) Regulation of fatty acid transport and membrane transporters in health and disease. Mol Cell Biochem, 239, 181-92. 17. Eaton S, Bartlett K, Pourfarzam M. (1996) Mammalian mitochondrial beta-oxidation. Biochem J, 320 ( Pt 2), 345-57. 18. Ramsay RR, Gandour RD, van der Leij FR. (2001) Molecular enzymology of carnitine transfer and transport. Biochim Biophys Acta, 1546, 21-43. 19. Ramsay RR. (2000) The carnitine acyltransferases: modulators of acyl-CoA-dependent reactions. Biochem Soc Trans, 28, 182-6. 20. Zammit VA. (1999) Carnitine acyltransferases: functional significance of subcellular distribution and membrane topology. Prog Lipid Res, 38, 199-224. 21. Hettema EH, Tabak HF. (2000) Transport of fatty acids and metabolites across the peroxisomal membrane. Biochim Biophys Acta, 1486, 18-27. 22. Ramsay RR, Arduini A. (1993) The carnitine acyltransferases and their role in modulating acyl-CoA pools. Arch Biochem Biophys, 302, 307-14. 23. Wanders RJ, Vreken P, Ferdinandusse S, Jansen GA, Waterham HR, van Roermund CW, Van Grunsven EG. (2001) Peroxisomal fatty acid alphaand beta-oxidation in humans: enzymology, peroxisomal metabolite transporters and peroxisomal diseases. Biochem Soc Trans, 29, 250-67. 24. Fritz IB, Yue KT. (1963) Long-Chain Carnitine Acyltransferase and the Role of Acylcarnitine Derivatives in the Catalytic Increase of Fatty Acid Oxidation Induced by Carnitine. J Lipid Res, 58, 279-88. 25. Bremer J. (1963) Carnitine in Intermediary Metabolism. The Biosynthesis of Palmitylcarnitine by Cell Subfractions. J Biol Chem, 238, 2774-9. 26. Bremer J. (1983) Carnitine--metabolism and functions. Physiol Rev, 63, 1420-80.

PAGE 104

93 27. McGarry JD, Sen A, Esser V, Woeltje KF, Weis B, Foster DW. (1991) New insights into the mitochondrial carnitine palmitoyltransferase enzyme system. Biochimie, 73, 77-84. 28. Bergstrom JD, Reitz RC. (1980) Studies on carnitine palmitoyl transferase: the similar nature of CPTi (inner form) and CPTo (outer form). Arch Biochem Biophys, 204, 71-9. 29. McGarry JD. (1995) The mitochondrial carnitine palmitoyltransferase system: its broadening role in fuel homoeostasis and new insights into its molecular features. Biochem Soc Trans, 23, 321-4. 30. van der Leij FR, Huijkman NC, Boomsma C, Kuipers JR, Bartelds B. (2000) Genomics of the human carnitine acyltransferase genes. Mol Genet Metab, 71, 139-53. 31. Esser V, Brown NF, Cowan AT, Foster DW, McGarry JD. (1996) Expression of a cDNA isolated from rat brown adipose tissue and heart identifies the product as the muscle isoform of carnitine palmitoyltransferase I (M-CPT I). M-CPT I is the predominant CPT I isoform expressed in both white (epididymal) and brown adipocytes. J Biol Chem, 271, 6972-7. 32. Bonnefont JP, Demaugre F, Prip-Buus C, Saudubray JM, Brivet M, Abadi N, Thuillier L. (1999) Carnitine palmitoyltransferase deficiencies. Mol Genet Metab, 68, 424-40. 33. Shi J, Zhu H, Arvidson DN, Cregg JM, Woldegiorgis G. (1998) Deletion of the conserved first 18 N-terminal amino acid residues in rat liver carnitine palmitoyltransferase I abolishes malonyl-CoA sensitivity and binding. Biochemistry, 37, 11033-8. 34. Price N, van der Leij F, Jackson V, Corstorphine C, Thomson R, Sorensen A, Zammit V. (2002) A novel brain-expressed protein related to carnitine palmitoyltransferase I. Genomics, 80, 433-42. 35. Fraser F, Corstorphine CG, Zammit VA. (1997) Topology of carnitine palmitoyltransferase I in the mitochondrial outer membrane. Biochem J, 323 ( Pt 3), 711-8. 36. Woeltje KF, Esser V, Weis BC, Sen A, Cox WF, McPhaul MJ, Slaughter CA, Foster DW, McGarry JD. (1990) Cloning, sequencing, and expression of a cDNA encoding rat liver mitochondrial carnitine palmitoyltransferase II. J Biol Chem, 265, 10720-5. 37. Finocchiaro G, Taroni F, Rocchi M, Martin AL, Colombo I, Tarelli GT, DiDonato S. (1991) cDNA cloning, sequence analysis, and chromosomal localization of the gene for human carnitine palmitoyltransferase. Proc Natl Acad Sci U S A, 88, 661-5.

PAGE 105

94 38. Solberg HE. (1971) Carnitine octanoyltransferase. Evidence for a new enzyme in mitochondria. FEBS Lett, 12, 134-136. 39. Solberg HE. (1972) Different carnitine acyltransferases in calf liver. Biochim Biophys Acta, 280, 422-33. 40. Kopec B, Fritz IB. (1971) Properties of a purified carnitine palmitoyltransferase, and evidence for the existence of other carnitine acyltransferases. Can J Biochem, 49, 941-8. 41. Markwell MA, Tolbert NE, Bieber LL. (1976) Comparison of the carnitine acyltransferase activites from rat liver peroxisomes and microsomes. Arch Biochem Biophys, 176, 497-88. 42. Corti O, Finocchiaro G, Rossi E, Zuffardi O, DiDonato S. (1994) Molecular cloning of cDNAs encoding human carnitine acetyltransferase and mapping of the corresponding gene to chromosome 9q34.1. Genomics, 23, 94-9. 43. Corti O, DiDonato S, Finocchiaro G. (1994) Divergent sequences in the 5' region of cDNA suggest alternative splicing as a mechanism for the generation of carnitine acetyltransferases with different subcellular localizations. Biochem J, 303 ( Pt 1), 37-41. 44. Chase JF. (1967) The substrate specificity of carnitine acetyltransferase. Biochem J, 104, 510-8. 45. Cohen-Haguenauer O, Brice A, Berrard S, Nguyen VC, Mallet J, Frezal J. (1990) Localization of the choline acetyltransferase (CHAT) gene to human chromosome 10. Genomics, 6, 374-8. 46. Strauss WL, Kemper RR, Jayakar P, Kong CF, Hersh LB, Hilt DC, Rabin M. (1991) Human choline acetyltransferase gene maps to region 10q11-q22.2 by in situ hybridization. Genomics, 9, 396-8. 47. Oda Y. (1999) Choline acetyltransferase: the structure, distribution and pathologic changes in the central nervous system. Pathol Int, 49, 921-37. 48. Ruderman NB, Park H, Kaushik VK, Dean D, Constant S, Prentki M, Saha AK. (2003) AMPK as a metabolic switch in rat muscle, liver and adipose tissue after exercise. Acta Physiol Scand, 178, 435-42. 49. Ruderman NB, Saha AK, Kraegen EW. (2003) Minireview: malonyl CoA, AMP-activated protein kinase, and adiposity. Endocrinology, 144, 5166-71. 50. Minokoshi Y, Kim YB, Peroni OD, Fryer LG, Muller C, Carling D, Kahn BB. (2002) Leptin stimulates fatty-acid oxidation by activating AMP-activated protein kinase. Nature, 415, 339-43.

PAGE 106

95 51. Hofbauer KG. (2002) Molecular pathways to obesity. Int J Obes Relat Metab Disord, 26 Suppl 2, S18-27. 52. Wolkowicz P, Grenett H, Belousova M, Urthaler F. (2004) Activation of leptin expression by an inhibitor of carnitine palmitoyltransferase-1. Int J Obes Relat Metab Disord, 28, 649-51. 53. Rinaldo P, Matern D, Bennett MJ. (2002) Fatty acid oxidation disorders. Annu Rev Physiol, 64, 477-502. 54. Taggart RT, Smail D, Apolito C, Vladutiu GD. (1999) Novel mutations associated with carnitine palmitoyltransferase II deficiency. Hum Mutat, 13, 210-20. 55. DiDonato S, Rimoldi M, Moise A, Bertagnoglio B, Uziel G. (1979) Fatal ataxic encephalopathy and carnitine acetyltransferase deficiency: a functional defect of pyruvate oxidation? Neurology, 29, 1578-83. 56. Ohno K, Tsujino A, Brengman JM, Harper CM, Bajzer Z, Udd B, Beyring R, Robb S, Kirkham FJ, Engel AG. (2001) Choline acetyltransferase mutations cause myasthenic syndrome associated with episodic apnea in humans. Proc Natl Acad Sci U S A, 98, 2017-22. 57. Liu J, Atamna H, Kuratsune H, Ames BN. (2002) Delaying brain mitochondrial decay and aging with mitochondrial antioxidants and metabolites. Ann N Y Acad Sci, 959, 133-66. 58. Hagen TM, Liu J, Lykkesfeldt J, Wehr CM, Ingersoll RT, Vinarsky V, Bartholomew JC, Ames BN. (2002) Feeding acetyl-L-carnitine and lipoic acid to old rats significantly improves metabolic function while decreasing oxidative stress. Proc Natl Acad Sci U S A, 99, 1870-5. 59. Liu J, Head E, Gharib AM, Yuan W, Ingersoll RT, Hagen TM, Cotman CW, Ames BN. (2002) Memory loss in old rats is associated with brain mitochondrial decay and RNA/DNA oxidation: partial reversal by feeding acetyl-L-carnitine and/or R-alpha -lipoic acid. Proc Natl Acad Sci U S A, 99, 2356-61. 60. Liu J, Killilea DW, Ames BN. (2002) Age-associated mitochondrial oxidative decay: improvement of carnitine acetyltransferase substrate-binding affinity and activity in brain by feeding old rats acetyl-Lcarnitine and/or R-alpha -lipoic acid. Proc Natl Acad Sci U S A, 99, 1876-81. 61. Chase JF, Tubbs PK. (1966) Some kinetic studies on the mechanism of action of carnitine acetyltransferase. Biochem J, 99, 32-40. 62. Chase JF, Tubbs PK. (1970) Specific alkylation of a histidine residue in carnitine acetyltransferase by bromoacetyl-L-carnitine. Biochem J, 116, 713-20.

PAGE 107

96 63. Wu D, Hersh LB. (1995) Identification of an active site arginine in rat choline acetyltransferase by alanine scanning mutagenesis. J Biol Chem, 270, 29111-6. 64. Wu D, Govindasamy L, Lian W, Gu Y, Kukar T, Agbandje-McKenna M, McKenna R. (2003) Structure of human carnitine acetyltransferase. Molecular basis for fatty acyl transfer. J Biol Chem, 278, 13159-65. 65. Jogl G, Tong L. (2003) Crystal structure of carnitine acetyltransferase and implications for the catalytic mechanism and fatty acid transport. Cell, 112, 113-22. 66. McGarry JD, Brown NF. (2000) Reconstitution of purified, active and malonyl-CoA-sensitive rat liver carnitine palmitoyltransferase I: relationship between membrane environment and malonyl-CoA sensitivity. Biochem J, 349, 179-87. 67. Zhu H, Shi J, Cregg JM, Woldegiorgis G. (1997) Reconstitution of highly expressed human heart muscle carnitine palmitoyltransferase I. Biochem Biophys Res Commun, 239, 498-502. 68. Zhu H, Shi J, de Vries Y, Arvidson DN, Cregg JM, Woldegiorgis G. (1997) Functional studies of yeast-expressed human heart muscle carnitine palmitoyltransferase I. Arch Biochem Biophys, 347, 53-61. 69. Shi J, Zhu H, Arvidson DN, Woldegiorgis G. (2000) The first 28 N-terminal amino acid residues of human heart muscle carnitine palmitoyltransferase I are essential for malonyl CoA sensitivity and high-affinity binding. Biochemistry, 39, 712-7. 70. Jackson VN, Zammit VA, Price NT. (2000) Identification of positive and negative determinants of malonyl-CoA sensitivity and carnitine affinity within the amino termini of rat liverand muscle-type carnitine palmitoyltransferase I. J Biol Chem, 275, 38410-6. 71. Shi J, Zhu H, Arvidson DN, Woldegiorgis G. (1999) A single amino acid change (substitution of glutamate 3 with alanine) in the N-terminal region of rat liver carnitine palmitoyltransferase I abolishes malonyl-CoA inhibition and high affinity binding. J Biol Chem, 274, 9421-6. 72. Jackson VN, Price NT, Zammit VA. (2001) Specificity of the interactions between Glu-3, Ser-24, and Gln-30 within the N-terminal segment of rat liver mitochondrial overt carnitine palmitoyltransferase (L-CPT I) in determining the malonyl-CoA sensitivity of the enzyme. Biochemistry, 40, 14629-34. 73. Jackson VN, Cameron JM, Fraser F, Zammit VA, Price NT. (2000) Use of six chimeric proteins to investigate the role of intramolecular interactions in determining the kinetics of carnitine palmitoyltransferase I isoforms. J Biol Chem, 275, 19560-6.

PAGE 108

97 74. Nicot C, Hegardt FG, Woldegiorgis G, Haro D, Marrero PF. (2001) Pig liver carnitine palmitoyltransferase I, with low Km for carnitine and high sensitivity to malonyl-CoA inhibition, is a natural chimera of rat liver and muscle enzymes. Biochemistry, 40, 2260-6. 75. Nicot C, Relat J, Woldegiorgis G, Haro D, Marrero PF. (2002) Pig liver carnitine palmitoyltransferase. Chimera studies show that both the Nand C-terminal regions of the enzyme are important for the unusual high malonyl-CoA sensitivity. J Biol Chem, 277, 10044-9. 76. Jancarik J, Kim SH. (1991) Sparse matrix sampling: a screening method for crystallization of proteins. J Appl Crystallogr, 24, 409-411. 77. Swiegers JH, Dippenaar N, Pretorius IS, Bauer FF. (2001) Carnitine-dependent metabolic activities in Saccharomyces cerevisiae: three carnitine acetyltransferases are essential in a carnitine-dependent strain. Yeast, 18, 585-95. 78. Bieber LL. (1988) Carnitine. Annu Rev Biochem, 57, 261-83. 79. Breningstall GN. (1990) Carnitine deficiency syndromes. Pediatr Neurol, 6, 75-81. 80. Otwinowski ZaM, W. (1997) Methods in Enzymology: Macromolecular Crystallography. Academic, San Diego. 81. Brunger AT, Adams PD, Clore GM, DeLano WL, Gros P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J, Nilges M, Pannu NS, Read RJ, Rice LM, Simonson T, Warren GL. (1998) Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr D Biol Crystallogr, 54 (Pt 5), 905-21. 82. Sheldrick GM. (1998) Direct Methods for Solving Macromolecular Structures. Kluwer Academic, Dordrecht. 83. Jones TA, Zou JY, Cowan SW, Kjeldgaard. (1991) Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr A, 47 (Pt 2), 110-9. 84. Cronin CN. (1997) The conserved serine-threonine-serine motif of the carnitine acyltransferases is involved in carnitine binding and transition-state stabilization: a site-directed mutagenesis study. Biochem Biophys Res Commun, 238, 784-9. 85. Govindasamy L, Kukar T, Lian W, Pedersen B, Gu Y, Agbandje-McKenna M, Jin S, McKenna R, Wu D. (2004) Structural and mutational characterization of L-carnitine binding to human carnitine acetyltransferase. J Struct Biol, 146, 416-424. 86. Mecozzi S, West AP, Jr., Dougherty DA. (1996) Cation-pi interactions in aromatics of biological and medicinal interest: electrostatic potential surfaces as a useful qualitative guide. Proc Natl Acad Sci U S A, 93, 10566-71.

PAGE 109

98 87. Dougherty DA. (1996) Cation-pi interactions in chemistry and biology: a new view of benzene, Phe, Tyr, and Trp. Science, 271, 163-8. 88. Lian W, Govindasamy L, Gu Y, Kukar T, Agbandje-McKenna M, McKenna R, Wu D. (2002) Crystallization and preliminary X-ray crystallographic studies on recombinant human carnitine acetyltransferase. Acta Crystallogr D Biol Crystallogr, 58, 1193-4. 89. Kawamoto S, Nozaki C, Tanaka A, Fukui S. (1978) Fatty acid beta-oxidation system in microbodies of n-alkane-grown Candida tropicalis. Eur J Biochem, 83, 609-13. 90. Kunau WH, Buhne S, de la Garza M, Kionka C, Mateblowski M, Schultz-Borchard U, Thieringer R. (1988) Comparative enzymology of beta-oxidation. Biochem Soc Trans, 16, 418-20. 91. van Roermund CW, Elgersma Y, Singh N, Wanders RJ, Tabak HF. (1995) The membrane of peroxisomes in Saccharomyces cerevisiae is impermeable to NAD(H) and acetyl-CoA under in vivo conditions. Embo J, 14, 3480-6. 92. Kispal G, Sumegi B, Dietmeier K, Bock I, Gajdos G, Tomcsanyi T, Sandor A. (1993) Cloning and sequencing of a cDNA encoding Saccharomyces cerevisiae carnitine acetyltransferase. Use of the cDNA in gene disruption studies. J Biol Chem, 268, 1824-9. 93. Schmalix W, Bandlow W. (1993) The ethanol-inducible YAT1 gene from yeast encodes a presumptive mitochondrial outer carnitine acetyltransferase. J Biol Chem, 268, 27428-39. 94. Thompson JD, Higgins DG, Gibson TJ. (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res, 22, 4673-80. 95. Cronin CN. (1997) cDNA cloning, recombinant expression, and site-directed mutagenesis of bovine liver carnitine octanoyltransferase--Arg505 binds the carboxylate group of carnitine. Eur J Biochem, 247, 1029-37. 96. Harel M, Schalk I, Ehret-Sabatier L, Bouet F, Goeldner M, Hirth C, Axelsen PH, Silman I, Sussman JL. (1993) Quaternary ligand binding to aromatic residues in the active-site gorge of acetylcholinesterase. Proc Natl Acad Sci U S A, 90, 9031-5. 97. Loewi O. (1921) Pflugers Arch, 189, 239-242. 98. Aigner TG, Mishkin M. (1986) The effects of physostigmine and scopolamine on recognition memory in monkeys. Behav Neural Biol, 45, 81-7.

PAGE 110

99 99. Bartus RT, Dean RL, 3rd, Beer B, Lippa AS. (1982) The cholinergic hypothesis of geriatric memory dysfunction. Science, 217, 408-14. 100. Karczmar AG. (1976) Central actions of acetylcholine, cholinomimetics and related drugs. Raven Press, New York. 101. Nachmansohn D, Machado AL. (1943) The formation of acetylcholine. A new enzyme choline acetylase. J Neurophysiol, 6, 397-403. 102. Carbini L, Rodriguez G, Hersh LB. (1990) Kinetic and inactivation studies of recombinant Drosophila choline acetyltransferase. Brain Res Bull, 24, 119-24. 103. Hersh LB, Nair RV, Smith DJ. (1979) The reaction of choline acetyltransferase with sulfhydryl reagents. Methoxycarbonyl-CoA disulfide as an active site-directed reagent. J Biol Chem, 254, 11988-92. 104. Malthe-Sorenssen D. (1976) Choline acetyltransferase--evidence for acetyl transfer by a histidine residue. J Neurochem, 27, 873-81. 105. Mautner HG, Pakula AA, Merrill RE. (1981) Evidence for presence of an arginine residue in the coenzyme A binding site of choline acetyltransferase. Proc Natl Acad Sci U S A, 78, 7449-52. 106. Roskoski R, Jr. (1974) Choline acetyltransferase. Inhibition by thiol reagents. J Biol Chem, 249, 2156-9. 107. Roskoski R, Jr. (1974) Choline acetyltransferase and acetylcholinesterase: evidence for essential histidine residues. Biochemistry, 13, 5141-4. 108. Roskoski R, Jr., Lim CT, Roskoski LM. (1975) Human brain and placental choline acetyltransferase: purification and properties. Biochemistry, 14, 5105-10. 109. Carbini LA, Hersh LB. (1993) Functional analysis of conserved histidines in choline acetyltransferase by site-directed mutagenesis. J Neurochem, 61, 247-53. 110. Engel AG, Ohno K, Shen XM, Sine SM. (2003) Congenital myasthenic syndromes: multiple molecular targets at the neuromuscular junction. Ann N Y Acad Sci, 998, 138-60. 111. Smith CP, Carroll PT. (1980) A comparison of solubilized and membrane bound forms of choline-O-acetyltransferase (EC 2.3.1.6) in mouse brain nerve endings. Brain Res, 185, 363-71. 112. Dobransky T, Davis WL, Rylett RJ. (2001) Functional characterization of phosphorylation of 69-kDa human choline acetyltransferase at serine 440 by protein kinase C. J Biol Chem, 276, 22244-50.

PAGE 111

100 113. Lian W, Gu Y, Pedersen B, Kukar T, Govindasamy L, Agbandje-McKenna M, Jin S, McKenna R, Wu D. (2004) Crystallization and Preliminary X-ray Crystallographic Studies on Recombinant Rat Choline Acetyltransferase. Acta Crystallogr D Biol Crystallogr, 60, 374-375. 114. Rossmann MG. (1990) The molecular replacement method. Acta Crystallogr A, 46 ( Pt 2), 73-82. 115. Mattevi A, Obmolova G, Kalk KH, Teplyakov A, Hol WG. (1993) Crystallographic analysis of substrate binding and catalysis in dihydrolipoyl transacetylase (E2p). Biochemistry, 32, 3887-901. 116. McGarry JD, Brown NF. (1997) The mitochondrial carnitine palmitoyltransferase system. From concept to molecular analysis. Eur J Biochem, 244, 1-14. 117. Leslie AG, Moody PC, Shaw WV. (1988) Structure of chloramphenicol acetyltransferase at 1.75-A resolution. Proc Natl Acad Sci U S A, 85, 4133-7. 118. Maselli RA, Chen D, Mo D, Bowe C, Fenton G, Wollmann RL. (2003) Choline acetyltransferase mutations in myasthenic syndrome due to deficient acetylcholine resynthesis. Muscle Nerve, 27, 180-7. 119. Schmidt C, Abicht A, Krampfl K, Voss W, Stucka R, Mildner G, Petrova S, Schara U, Mortier W, Bufler J, Huebner A, Lochmuller H. (2003) Congenital myasthenic syndrome due to a novel missense mutation in the gene encoding choline acetyltransferase. Neuromuscul Disord, 13, 245-51. 120. Greenspan RJ, Finn JA, Jr., Hall JC. (1980) Acetylcholinesterase mutants in Drosophila and their effects on the structure and function of the central nervous system. J Comp Neurol, 189, 741-74. 121. Wang W, Kitamoto T, Salvaterra PM. (1999) Drosophila choline acetyltransferase temperature-sensitive mutants. Neurochem Res, 24, 1081-7.

PAGE 112

BIOGRAPHICAL SKETCH Brenda Dawn Pedersen was born and raised in Sarasota, FL, as the only child of parents Erik and Lynnae Pedersen. She graduated from Sarasota High School, and attended Florida Institute of Technology in Melbourne, FL, where she received a BS with honor in biochemistry. After graduation she enrolled in graduate school at the University of Florida in Gainesville, FL, to pursue a doctoral degree in medicinal chemistry. After graduation, she plans to continue her training with a postdoctoral fellowship. 101


Permanent Link: http://ufdc.ufl.edu/UFE0005701/00001

Material Information

Title: Structure-Function Studies of the Carnitine/Choline Acyltransferase Family
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0005701:00001

Permanent Link: http://ufdc.ufl.edu/UFE0005701/00001

Material Information

Title: Structure-Function Studies of the Carnitine/Choline Acyltransferase Family
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0005701:00001


This item has the following downloads:


Full Text












STRUCTURE-FUNCTION STUDIES OF THE CARNITINE/CHOLINE
ACYLTRANSFERASE FAMILY
















By

BRENDA DAWN PEDERSEN


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2004

































Copyright 2004

by

Brenda Dawn Pedersen

































This dissertation is dedicated to my family and friends, who have supported and believed
in me along the way.















ACKNOWLEDGMENTS

I would first like to thank my mentor, Dr. Donghai Wu. I thank him for his

encouragement, for his patience in answering my questions, and for always pushing me

to go farther. Without his support this dissertation would not have been possible. I would

also like to thank all of my committee members: Dr. Sloan, Dr. Schulman, Dr. Prokai

and Dr. Oh. Their suggestions for the direction of this project, their time and their

guidance are greatly appreciated.

My gratitude and thanks go to the other members of my lab who have made this

research possible. I thank Yunrong Gu and Wei Lian who, very patiently, shared their

technical expertise with me on numerous occasions. I sincerely appreciate all of their

contributions to this project. I also thank fellow graduate student and senior lab gremlin,

Dr. Tom Kukar, for his assistance and support. Ice cream is a small price to pay for your

friendship. In addition, I would like to acknowledge my fellow graduate students. I hope

that the friendships that have been made along the way will continue for many years to

come.

I will always be thankful for the continued support and enthusiasm of my parents,

Erik and Lynnae Pedersen, throughout my academic career. I thank them for believing in

me and constantly encouraging me. Finally, I thank my fiance, Eric Moore. I will always

be grateful for his unwavering love and support for all of my endeavors.
















TABLE OF CONTENTS
Page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES ....................................................... ............ .............. .. vii

L IST O F FIG U R E S .............. ............................ ............. ........... ........... viii

A B ST R A C T .......... ..... ...................................................................................... x

CHAPTER

1 L ITER A TU R E R E V IE W .............................................................. .. .....................1

Fatty A cid M etabolism ................... ...... ................................................................ 2
The Carnitine/Choline Acyltransferase Family .....................................................9
Cam itine Palm itoyltransferase ........................................ ......................... 9
Cam itine Palm itoyltransferase I...................................... ............... 12
Cam itine Palm itoyltransferase II ...................................... ............... 14
Cam itine O ctanoyltransferase ........................................ ......... ............... 15
Cam itine A cetyltransferase ........................................ ........................... 16
Choline A cetyltransferase ............................................................................17
Regulation of L-Camitine Acyltransferases .............. ....... ........................18
L-Camitine/Choline Acyltransferase Disorders ............................... ...............19
L-Camitine/Choline Acyltransferase Deficiencies.................... ...................20
L-Cam itine Acetyltransferase Increases................................... ............... 23
Structure-F unction Stu dies .............................................................. .....................24
Specific A im s........................................................................................ 28

2 MATERIALS AND METHODS ........................................ ......................... 30

Determine Structure of Carnitine Acetyltransferase with L-Camitine.....................31
E x p re ssio n ..................................................................... .. 3 1
E x traction n ................................................................3 2
P u rifi catio n ................................................................3 3
Crystallization .......................... ......... ........ ...............34
M utate Specific R esidues of CA T ........................................ ......................... 34
Evaluate Enzym atic A activity ............................................. ............................. 36
Evaluate Biological Activity...................................................... 38
Structural Determination of Choline Acetyltransferase ..........................................38



v









3 STRUCTURE DETERMINATION OF CARNITINE ACETYLTRANSFERASE
W IT H L -C A R N IT IN E ..................................................................... .....................40

In tro du ctio n ...................................... ................................................ 4 0
M materials and M methods ....................................................................... ..................43
R esu lts an d D iscu ssion ......................................................... .......... ...... ......... .. 44
Structure Determination and Overall Structure of the Enzyme Complex ..........44
Comparison of CAT-L-Carnitine with Free CAT............... ....................44
L -C am itine B finding Site ...................... ......................... ............... ... 46
Comparison of CAT-L-Carnitine with Mouse CAT .......................................50
C o n c lu sio n s..................................................... ................ 5 1

4 FUNCTIONAL AND BIOLOGICAL CHARACTERIZATION OF ESSENTIAL
RESIDUES OF CARNITINE ACETYLTRANSFERASE ........................................52

In tro du ctio n ...................................... ................................................ 52
M materials and M methods ....................................................................... ..................55
R results and D discussion ............................. ................................ .. .. .... ......... .. 56
Functional Characterization of L-Carnitine Binding Residues ...........................56
Biological Characterization of L-Camitine Binding Residues............................61
C o n clu sio n s..................................................... ................ 6 7

5 STRUCTURE OF CHOLINE ACETYLTRANSFERASE....................................68

In tro du ctio n ...................................... ................................................ 6 8
M materials and M methods ....................................................................... ..................72
R results and D discussion ................. ....... ........ ............................ ........ ...... ....72
Structure Determination and Overall Structure of the Enzyme...........................72
Comparison of CAT-L-Carnitine with Free CAT............... ....................76
A ctiv e S ite ..................................................................................................... 7 8
Choline Binding Site ......................... ...... .............. .......... .. .......... ....79
Coenzyme A Binding Site......................... .................80
Disease Associated Mutations... .. ................ ........................83
Temperature Sensitive Mutations.................... ............................86
C o n clu sio n ...................... .. .. ......... .. .. ................................................. 8 7

6 CON CLU SION S .................................. .. .......... .. .............88

LIST OF REFEREN CES ................................................................... ............... 91

BIOGRAPHICAL SKETCH ................ ........................................... ............... 101















LIST OF TABLES


Table pge

4-1 Kinetic parameters of recombinant wild type and mutant forms of camitine
acetyltransferase. .......................................................................57

4-2 Growth of mutant yeast cells with glucose and ethanol media.............................65

4-3 Enzymatic activity and growth with glucose and ethanol of Glu326 mutants..........67

5-1 Reported mutations of choline acetytlransferase.* ................................................70

5-2 Choline substrate binding sites for rat choline acetyltransferase (rChAT), human
peroxisomal L-camitine acetyltransferase (hCAT), and mouse L-carnitine
acetyltransferase (m CA T). ............................................. .............................. 82

5-3 Coenzyme A substrate binding sites for rat choline acetyltransferase (rChAT),
human peroxisomal L-camitine acetyltransferase (hCAT), and mouse
L-carnitine acetyltransferase (mCAT). ........................................ ............... 82
















LIST OF FIGURES


Figure p

1-1 Chemical structure of a triglyceride ..................... ............... ...............3

1-2 The release of stored fatty acids.......................................... .......................... 4

1-3 General scheme of the L-carnitine acyltransferase reaction. ............. .................6

1-4 Long-chain fatty acid transport across the mitochondrial membrane....................7

1-5 Long-chain and branched fatty acid metabolism in the peroxisome.........................8

1-6 Plasmalemmal L-carnitine transporter, L-carnitine palmitoyltransferase I, and L-
carnitine/acyl-L-cam itine translocase. ........................................ ............... 14

1-7 Structure of m alonyl-CoA ......................................................... ............... 18

2-1 The pQ E -9 expression vector. ...................................................................... .......31

2-2 Interaction of histidine residues of 6xHis tag and the Ni-NTA agarose ................33

2-3 Overview of the QuickChange site-directed mutagenesis method. .........................35

2-4 Schematic representation of the CAT assay .........................................................37

3-1 Overall structure of human peroxisomal camitine acetyltransferase.....................43

3-2 Structural similarity and superimposition of CAT and CAT-L-carnitine ...............45

3-3 Ser433 positional change in free and L-carnitine bound form of CAT.. ...................46

3-4 Electron density map of the carnitine binding site of L-carnitine
acetyltransferase .................................... ............................... ........47

3-5 The active site of L-carnitine acetyltransferase-L-Carnitine.............................. 48

3-6 The interactions between L-carnitine and L-carnitine acetyltransferase..................49

4-1 The two pathways for the metabolism of fatty acids.................................... 53

4-2 Sequence alignment of carnitine and choline acyltransferases.. ............................56









4-3 The active site of carnitine acetyltransferase with Tyr431. ......... .....................58

4-5 Human carnitine acetyltransferase complements mutant S. cerevisiae ..................64

4-5 The active site of carnitine acetyltransferase.. ................................................... 65

4-7 The reaction mechanism of carnitine acetyltransferase.. ......................................66

5-1 Schematic of choline acetyltransferase reaction mechanism ................................69

5-2 Photograph of recombinant rat choline acetyltransferase crystals.........................73

5-3 X-ray diffraction pattern of rat choline acetyltransferase crystal...........................74

5-4 Overall Structure of Rat Choline Acetyltransferase .............................................75

5-5 Sequence alignment of choline acetyltransferases. .............................................77

5-6 The conformation arrangement of the active site histidine................................79

5-7 The modeled binding site of choline and CoA ......................................................81

5-8 Structural comparison of rat choline acetyltransferase (rChAT) and human
peroxisomal carnitine acetyltransferase (hCAT) and mutations in human ChAT
that cause myasthenic syndrome associated with episodic apnea (CMS-EA). .......85















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

STRUCTURE-FUNCTION STUDIES OF THE CARNITINE/CHOLINE
ACYLTRANSFERASE FAMILY

By

Brenda Dawn Pedersen

August 2004

Chair: Donghai Wu
Major Department: Medicinal Chemistry

To further understand the process of fatty acid metabolism, the enzymes of the

carnitine acyltransferase family were studied. These enzymes play essential roles in fatty

acid metabolism by facilitating the transfer of activated fatty acids across intracellular

membranes. To understand the mechanism of these enzymes, the structure of human

peroxisomal L-camitine acetyltransferase (CAT) in complex with its substrate,

L-camitine, was determined. The structure revealed several residues involved in substrate

binding that were further investigated by an enzymatic and biological assay.

The structure of CAT with L-carnitine is very similar to the structure of the free

enzyme. Both are monomeric proteins with two equally sized domains that form a central

active site tunnel. The most significant difference between these two structures is with

residue Ser433. The chi angle of this amino acid changes from 58 to -310 to position the

hydroxyl group of the residue closer to the carboxylate group of L-carnitine, possibly to

help in stabilization during the transition state.









The structure also reveals several residues (His322, Tyr431, Thr444, Arg497, Phe545,

and Val548) and two water molecules (W621 and W679) involved with binding L-carnitine.

Four of these residues (Tyr431, Thr444, Arg497 and Phe545) were further investigated. These

studies showed that Tyr431 and Thr444 interact with the carboxylate group of L-carnitine

through hydrogen bonds, that Arg497 interacts with the carboxylate group through

electrostatic interactions, and that Phe545 interacts with the quaternary amine group

through pi-cation interactions.

Another member of the L-carnitine/choline acyltransferase family, choline

acetyltransferase (ChAT), was investigated. This enzyme catalyzes the synthesis of

acetycholine (a neurotransmitter that functions in memory, learning and muscle

contraction). The structure of rat ChAT (rChAT) was determined, and the substrates were

modeled into the active site. The structure of rChAT shares several structural features

with CAT. The rChAT is also a monomeric protein, and consists of two domains that

form a central active site tunnel. Besides identifying several active site residues, the

structure also provides a molecular basis for the disease, congenital myasthenic

syndrome.














CHAPTER 1
LITERATURE REVIEW

Obesity is a serious health concern in the United States and in other developed

countries. Obesity is defined by the body mass index, which is a measure of weight in

relation to height. A person with a body mass index greater than 30 is clinically defined

as obese, while a person with a body mass greater than 25 is defined as overweight.

According to the NIH, based on the US Census Bureau Census 2000, nearly two thirds of

United States adults are overweight, including one third that are obese [1]. Surprisingly

less than half of United States adults maintain a healthy weight. The prevalence of

overweight and obesity in adults has steadily increased over the past 20 years among both

genders, all ages, all racial and ethnic groups, and all educational levels [1]. According to

the CDC, in 2002 every state reported an obesity prevalence of at least 15-19%, while 29

states reported a prevalence of 20-24%, and 3 states reported an obesity prevalence over

25%.

The obesity prevalence has huge implications for public health and society. Several

diseases are associated with overweight and obesity. Specifically, overweight and obesity

are risk factors for diabetes, heart disease, stroke, hypertension, gallbladder disease,

osteoarthritis, sleep apnea and other breathing problems, and some forms of cancer [1].

Obesity is also associated with high blood cholesterol, complications with pregnancy,

menstrual irregularities, hirsutism, stress incontinence, psychological disorders and an

increased risk of complications during surgery [1]. In addition, due to the rise in excess

weight and obesity, health care costs have increased dramatically. The annual medical









spending due to excess weight and obesity is approximately 93 billion or 9% of health

care expenditures, which can be broken down into costs associated with heart disease,

type 2 diabetes, osteoarthritis, hypertension, gallbladder disease, cancer, and lost

productivity at work [1]. In addition to these health care costs, Americans spend $33

billion dollars per year on weight-loss products and services.

Many people blame obesity on a sedentary lifestyle or poor food choices (such as

over-eating, too many carbohydrates, too much saturated fat). However, research has

shown that there may be a genetic influence on weight disorders [2]. One of the National

Health Objectives is to reduce the prevalence of obesity among adults to less than 15%.

Perhaps the most effective way to combat the problem is through further understanding

of the mechanism of fatty acid metabolism [1].

Fatty Acid Metabolism

Fatty acids are involved in several cellular processes. They regulate transcription

[3, 4], serve as building blocks for components of cell membranes [3], serve as precursors

to cell-signaling molecules [5-7], and provide a significant source of energy for heart and

skeletal muscle [8, 9]. The primary sources of fatty acids for energy production are diet,

synthesis, and release from cellular stores.

Dietary fatty acids account for at least 30% of caloric intake of humans in

industrialized countries [1]. Before these fatty acids can be oxidized for energy, they

must be absorbed, transported to the appropriate organ, taken up into the cell, and then

transferred to the mitochondria, where oxidation occurs. First, fatty acids from dietary

sources are digested by emulsification. In this process, large aggregates of insoluble fat

particles are broken down physically (by bile salts from the gall bladder), and are held in

suspension. The fat particles are then hydrolyzed to monoglycerides and free fatty acids









by pancreatic lipases. These lipids are absorbed from the gut by diffusing into the

epithelial cells lining the intestinal surface. There they are transported to the endoplasmic

reticulum, to be reconverted to triglycerides. The newly formed triglycerides are

packaged with cholesterol, lipoproteins, and other lipids to form chylomicrons.

Chylomicrons move through the lymphatic system to the bloodstream. Lipoprotein lipase

disassembles the chylomicron, and releases the fatty acid for uptake into cells. In muscle

cells, the fatty acid will be oxidized for energy; but in adipose tissue, the fatty acid will

be re-esterified for storage as triglycerides (Figure 1-1).








HC- vcH3
0


H-2C-_OCH3

Triglyceride


Figure 1-1. Chemical structure of a triglyceride

Stored triglycerides (Figure 1-1) are released in response to energy demands [10,

11]. As shown in Figure 1-2, the hormone adrenalin is secreted in response to low blood

glucose levels and binds to the P-adrenergic receptor (a G-protein coupled cell-surface

receptor of adipocytes), which activates adenylate cyclase [10]. The resulting increase in

cyclic AMP levels stimulates cyclic AMP-dependent protein kinase to phosphorylate and

activate hormone-sensitive lipase. This enzyme sequentially hydrolyzes ester linkages of

triglycerides to diglyceride and monoglyceride, resulting in the release of three moles of









fatty acid and one mole of glycerol [10]. The free fatty acid is released from the

adipocyte into the blood, where it will bind to serum albumin and travel to tissues to

serve as fuel.

Adrenalin

P-adrenergic G-protein Adenylate cyclase
receptor
ATP

...........................

cAMP

Hormone sensitive
lipase

ATP

4 .. PKA

ADP

Hormone sensitive
lipase

MAG lipase


glycerol MAG DAG TAG

fatty acid fatty acid fatty acid

Figure 1-2. Representation of the release of stored fatty acids. PKA, protein kinase; TAG,
triacylglycerol; DAG, diacylglycerol; MAG, monoacylglycerol.

The process of fatty acid uptake into cells is quite controversial, with

experimental evidence pointing to both rapid diffusion and a protein-mediated process

[12]. Several investigators have shown that fatty acids traverse cell membranes very

quickly suggesting that a protein is not necessary for uptake. These studies show that the

molecule uses a flip-flop mechanism to diffuse across the membrane [13, 14]. However,









at least three different membrane associated proteins have been identified as putatively

involved in trafficking fatty acids across cell membranes: the plasma membrane fatty acid

binding protein (FATPpM); the fatty acid transporter protein (FATP); and the fatty acid

translocase, which is homologous to the human leukocyte differentiation antigen CD36,

(FAT/CD36) [12, 15]. It appears that fatty acid uptake increases with expression and over

expression of these proteins [12]. Some researchers still contest that the presence of a

protein mechanism does not negate the fact that a proportion of fatty acids also diffuse

across the membrane [16]. Conversely it is thought that these proteins could function as

fatty acid acceptors, while the actual movement of fatty acids through the membrane

occurs by diffusion. Nonetheless, it is also possible that these two mechanisms co-exist as

separate routes, and each contributes to the overall rate of fatty acid uptake [14].

Once fatty acids are inside the cell, they must first be activated via ligation with

coenzyme A (CoA), forming a fatty acyl-CoA ester, before they are oxidized for energy

production [17]. Several fatty acyl-CoA synthetases that catalyze this ATP-dependent

process have been identified, and are grouped by their chain length specificity.

Short-chain acyl-CoA synthetases prefer fatty acids with 2 and 3 carbon atoms as a

substrate; while medium-chain acyl-CoA synthetases act on fatty acids with 4 to 12

carbon atoms; and long-chain acyl-CoA synthetases catalyze the reaction with fatty acids

with 10 to 20 carbon atoms [17]. The activity of these enzymes has been observed in the

mitochondrial matrix (short and medium chain) and the outer mitochondrial membrane

(long chain) [17].

Long-, medium- and short-chain acyl-CoAs are oxidized in the mitochondrial

matrix. Since acyl-CoAs are impermeable to membranes, they are transported into the









matrix by an enzyme system consisting of carnitine palmitoyltransferase (CPT) I and II

and an L-carnitine/acyl-L-camitine transporter [8, 18-20]. The CPT I and II are members

of the L-camitine/choline acyltransferase enzyme family. These enzymes catalyze the

reversible transfer of acyl groups between L-carnitine/choline and CoA to facilitate the

transfer of fatty acids across intracellular membranes (Figure 1-3) [18].


H3C H3C
H3 N CoA S R H3CN "Y 0 + CoA SH
H3N \ CoAH \
CH3 OH CH3 0 0

L-Carnitine Acyl-CoA R

Acyl-L-Carnitine CoA


Figure 1-3. General scheme of the L-carnitine acyltransferase reaction. R represents the
carbon chain that consists of 1-21 carbon atoms.

This reaction commits the activated fatty acid to oxidation in the mitochondrial

matrix (Figure 1-4) [8]. The CPT I is an integral membrane protein with two

transmembrane domains in the outer mitochondrial membrane, and its catalytic domain

facing the cytosol [18]. The formed acyl-L-carnitine is then transported across the

mitochondrial membrane by the L-carnitine/acyl-L-carnitine transporter. The transporter

is an integral membrane protein that uses a ping-pong mechanism to exchange

L-camitine for acyl-L-carnitine [8]. Once inside the mitochondrial matrix, CPT II

catalyzes the transfer of the acyl group from acyl-L-carnitine to CoA to form acyl-CoA,

which can then enter the P-oxidation spiral (Figure 1-4). The CPT II is anchored to the

inner membrane as a peripheral membrane protein.

The P-oxidation occurs through the sequential removal of 2 carbon units at the

P-carbon position of the fatty acyl-CoA molecule. With each round of oxidation, 1 mole









c -- CoA A L-Carnitine

Outer
CPT I Outer
Membrane


CACT Inner
Membrane
CPTII

Acy- CoA o Acy- L-Carnitine



3-oxidation The Citric Acid Cycle

Figure 1-4. Long-chain fatty acid transport across the mitochondrial membrane. CoA,
coenzyme A; CPT I, L-carnitine palmitoyltransferase I; CACT, L-
carnitine/acyl-L-camitine transporter; CPT II, L-camitine palmitoyltransferase
II.

of NADH, 1 mole of FADH2, and 1 mole of acetyl-CoA are produced. The acetyl-CoA

then enters the citric acid cycle, where it is further oxidized to C02; and 3 moles of

NADH, 1 mole of FADH2, and 1 mole of ATP are generated. The NADH and FADH2

formed from P-oxidation and the citric acid cycle can proceed to the respiratory pathway

for the production of ATP.

In contrast, very long-chain acyl-CoAs (with 24 or more carbon atoms) and

branched fatty acyl-CoAs (such as pristanic acid) are poor substrates for the

mitochondrial system, and are therefore oxidized in peroxisomes (Figure 1-5) [21, 22].

Transport of these fatty acids across the peroxisomal membrane appears to differ from

transport of acyl-CoAs across the mitochondrial membrane in that the carnitine system is









not necessary [21]. It is suspected that a transporter protein is involved; however the

mechanism of this arrangement is unknown.

Once inside the peroxisome, the very long-chain and branched acyl-CoAs are

P-oxidized using the same mechanism that occurs in the mitochondria: dehydrogenation,

hydration, dehydrogenation again, and thiolytic cleavage [21, 23]. However, there are

differences between the two systems. The acetyl-CoA that is produced from P-oxidation

in the peroxisomes cannot be degraded into CO2 and water, because there is no functional

Krebs cycle in the peroxisome. In addition fatty acids are not fully oxidized to

acetyl-CoA units in the peroxisome: they are only chain-shortened [23]. For complete

oxidation, the chain-shortened acyl-CoAs are transferred to the mitochondria.

Very Long-Chain Fatty Acids
Branched Fatty Acids

Peroxisome



S 3-oxidation
Fatty Acid Acetyl-CoA + Fatty Acid

CAT COT


Acetyl-L- Acyl-L-
Carnitine Camitine


Figure 1-5. Long-chain and branched fatty acid metabolism in the peroxisome. CAT, L-
carnitine acetyltransferase; COT, L-carnitine octanoyltransferase.

However, these molecules cannot freely diffuse out of the peroxisome or into the

mitochondria: they must use the carnitine system. A peroxisomal camitine

octanoyltransferase (COT) (another member of the L-camitine/choline acyltransferase









family) transfers medium-chain acyl groups (with 6 to 22 carbon atoms) from acyl-CoA

to L-carnitine to form acyl-L-camitine. Another member of the L-carnitine/choline

acyltransferase family, carnitine acetyltransferase (CAT), catalyzes the same reaction

with short-acyl groups (2 to 6 carbon atoms) [18, 22]. The acyl-L-carnitine can then

diffuse across the peroxisomal and mitochondrial membranes. After the acyl-L-carnitine

is in the mitochondrial matrix, CPT II transfers medium-chain acyl groups from L-

carnitine back to CoA to form acyl-CoA; while a mitochondrial form of CAT catalyzes

the reverse reaction, and transfers short-acyl groups from L-camitine back to CoA to

form acyl-CoA. Acyl-CoAs with 2 carbon atoms will enter the citric acid cycle directly

while longer acyl-CoAs will be further 3-oxidized before they enter the citric acid cycle.

The Carnitine/Choline Acyltransferase Family

As discussed, the enzymes of the carnitine acyltransferase family play vital roles in

fatty acid oxidation by facilitating in the transfer of acyl groups across intracellular

membranes. The molecular genetics of these enzymes is known and is presented here.

Carnitine Palmitoyltransferase

As previously discussed, activated long-chain fatty acids are transported across the

mitochondrial membrane by an enzyme system consisting of CPT I, CPT II, and an

L-carnitine/acyl-L-carnitine translocase. This enzymatic arrangement was first

conceptualized in the 1960s by Fritz and Yue [24] and Bremer [25]. Since then, the

general concept of the system has undergone several refinements.

In 1955, it was shown that L-camitine is an active component of muscle extracts,

and that it stimulates long-chain fatty acid oxidation in liver perfusions and skeletal and

heart muscle fractions. Then it was shown that, in the presence of L-camitine, CoA also

stimulates fatty acid oxidation [26]. Further studies showed that L-carnitine was









acetylated, and fatty acid esters of L-camitine were isolated. It was thought that

L-camitine might react with long-chain fatty acids to form acyl-L-camitine derivatives,

which could then be converted to the corresponding acyl-CoA compounds [24].

Experimental evidence demonstrated that a transferase in heart muscle catalzyed the

transfer of palmitoyl from palmitoyl-L-carnitine to CoA, to form palmitoyl-CoA [24]. It

was then thought that L-camitine may enhance fatty acid oxidation via acyl-L-camitine

formation, and that palmitoyl-L-camitine may function as a carrier through membranes

[24]. The presence of CPT on both sides of the membrane was essential to this

hypothesis. Experimental data confirmed this hypothesis, establishing that formation of

palmitoyl-L-carnitine and de-acetylation of palmitoyl-L-camitine could occur on either

side of the mitochondria [25]. Further functional studies with isolated mitochondria

demonstrated that the membrane contains two pools of CPT: an external (overt) (CPT I)

and an internal (latent) (CPT II) [26].

After the identification of CPT, several experiments were performed to characterize

the system. Inhibition studies with the enzyme determined that CPT I is inhibited by

malonyl-CoA (reversibly), tetradecylglycidyl-CoA and etomoxir-CoA (irreverisbly), and

2-bromopalmitoyl-CoA (essentially irreversibly) [27]. Interestingly, it was found that

detergent-solubilized CPT is insensitive to all inhibitors, while it is still enzymatically

active. This form of CPT is also stable, and migrates on sodium dodecylsulphate (SDS)

gels as a single protein with a molecular weight of 68-70 kDa [27]. Based on these

findings, a model of the CPT system was developed. In this model, CPT I and II are the

same protein. The CPT I is located in the outer membrane, and is associated with a

regulatory protein to which the inhibitors bind and cause allosteric inactivation of the









enzyme. Detergents can then disrupt the interaction between the catalytic and regulatory

components of CPT I, which allows the co-purification of CPT I and II as single entities

[27].

Several experiments provided additional support for this model. In particular,

Bergstrom and Reitz [28] found that when extracted from the mitochondrial membrane,

CPT I and II had very similar properties (such as the Km value toward acyl-CoA, L-

carnitine, acyl-L-carnitine, and CoASH; and the ratio of Vforward/Vreverse). Additionally,

antibodies raised against CPT I inhibited CPT II and vice versa, suggesting

immunoidentity of the two enzymes [28]. However, several experiments also raised

questions about the validity of the model.

The existence of a regulatory component came into question. Previous studies had

shown that 2-bromopalmitoyl-CoA inhibited CPT I, and formed a ternary complex at the

active site of the enzyme. Thereafter it was demonstrated that all of the inhibitors of CPT

I (malonyl-CoA, tetradecylglycidyl-CoA, etomoxir-CoA, and 2-bromopalmitoyl-CoA)

interacted at a common locus, suggesting that the other inhibitors must also act at the

active site and not at a regulatory component [27]. The theory that CPT I and II are the

same protein was also in question. Radioactive tetradecylglycidyl-CoA was found to

label a single protein in rat liver mitochondria with a molecular weight of 90-94 kDa,

while the labeled protein in rat skeletal muscle was only 86 kDa. Interestingly, CPT from

both tissues migrated to the same position on SDS gels (70 kDa), and reacted equally

well with an antibody raised against the liver form of the enzyme [27].

An alternative model to explain these observations was proposed. In this model

CPT I and II are different proteins. Inhibitors bind to the catalytic center of CPT I, and do









not affect the activity of CPT II due to the lack of necessary binding sites [27, 29]. In

support of this model, irradiation and inactivation experiments concluded there were two

catalytic entities of different sizes: 83 and 70 kDa [27]. Furthermore, a Western Blot from

the mitochondrial membrane (containing CPT I and II) and a mild detergent (Tween-20)

extracted membrane (rich in CPT I, devoid of CPT II) demonstrated a strong signal (with

an antibody raised toward CPT II) in the normal membrane, and a lack of signal in the

detergent extracted membrane [27].

Another facet of the model is that CPT I is detergent labile, while CPT II is not [27,

29]. A mild detergent (such as Tween 20) separated the mitochondrial membrane extract

into pellet and supernatant fractions; both express CPT activity. The pellet fraction was

inhibited by etomoxir-CoA, while the supernatant was insensitive to the inhibitor. Upon

further treatments with stronger detergents (such as octylglucoside and Triton X-100)

enzyme activity decreased. This indicates that CPT II was released from the membrane

with mild detergent, while CPT I was more firmly anchored in its membrane

environment. However, CPT I lost its catalytic activity upon membrane disruption,

suggesting that CPT I requires a critical membrane component [27, 29].

It can now be concluded that CPT I and II are different proteins that act in concert

with an L-camitine/acyl-L-camitine translocase to transport long-chain fatty acids across

the mitochondrial membrane. Of the two enzymes, CPT I is larger and detergent labile.

However, many factors of the model are still unresolved.

Carnitine Palmitoyltransferase I

Three forms of CPT I have been identified: a muscle (M-CPT I), a liver (L-CPT I),

and a putative brain (CPT I-C) isoform. The gene encoding the muscle form of CPT I has

been mapped to chromosome 22, position 22q13.31-q13.32, and encodes a protein that









consists of 772 amino acids [30]. This isoform is the primary form in skeletal muscle,

heart, adipose tissue, and testis [31, 32]. The liver form of CPT I is the predominant

isoform in the liver, kidney, lung, spleen, intestine, pancreas, ovary, and brain [31, 32].

The gene encoding L-CPT I has been mapped to chromosome 11, position 11q13.1-

q13.5, and encodes a protein that consists of 773 amino acids [30]. The muscle and liver

isoforms are very similar: in fact, they have a 63% amino acid sequence identity [33].

The brain form of CPT I was only recently identified [34]. The gene that encodes

CPT I-C has been mapped to chromosome 19, position 19q13.33, and encodes a protein

that is similar to M-CPT I and L-CPT I, with an amino acid sequence identity of 53 and

55%, respectively [34]. This isoform is expressed at high levels in the brain, particularly

in regions that have been shown to be involved with feeding behavior and appetite

control (such as the hippocampus, paraventricular arcuate, and the superchiasmatic

nuclei) [34]. However, no catalytic activity was observed with acyl-CoA esters that are

substrates for the muscle and liver isoforms. This suggests that CPT I-C may participate

in highly specialized lipid metabolism of the brain, and explains why the brain may

require the expression of an additional member of the CPT I family [34].

The structural features of the three forms of CPT I are similar. As previously discussed,

CPT I is an integral membrane protein localized to the outer membrane of the

mitochondria. Topographical studies have shown that CPT I contains two transmembrane

domains (amino acid residues 46-76 and 104-126), with both N- and C-terminals (amino

acid residues 1-128 and 128-772/773, respectively) facing the cytosol and the linker

region between the two domains (amino acid residues 77-103) jutting into the

intermembrane space (Figure 1-6) [35]. The N-terminal of CPT I contains a









mitochondrial targeting signal peptide (amino acid residues 123-147) for import into the

mitochondria. CPT I is the first committed step to oxidation in the mitochondria, and

therefore is a highly regulated process, which will be discussed later in this chapter.



PM



N C

L-carnitine N C
transporter

L-carnitine
palmitoyltransferase M


N


L-carnitine/acyl-L- MIM
carnitine translocase



Figure 1-6. Topographical representation of the plasmalemmal L-carnitine transporter, L-
carnitine palmitoyltransferase I, and L-carnitine/acyl-L-carnitine translocase.
Membranes are shown in blue, pm, plasma membrane; mom, mitochondrial
outer membrane; mim, mitochondrial inner membrane. Regions spanning the
membrane are shown by purple barrels. The amino-terminus (N) and carboxy-
terminus (C) of each protein are shown. Adapted from information reviewed
in [18].

Carnitine Palmitoyltransferase II

The genes that encode the human and rat form of CPT II have been identified and

mapped to chromosome 1, position lp32, and translate to a 658 amino acid protein [36,

37]. CPT II is a mitochondrial matrix protein, and is associated with the inner

mitochondrial membrane. The first 25 amino acid residues are part of a mitochondrial









targeting signal, and are removed during import into the mitochondria [18]. In contrast to

CPT I, only one form of CPT II has been identified, and it shares 50% amino acid

sequence homology with L-CPT I in the major portion of the sequences, except the N-

terminus [32].

It appears that CPT II is regulated by the homeostasis of the cell. In conditions of

starvation and diabetes, when CPT II activity is critical for generating fuel sources, the

mRNA levels (and consequently the activity of CPT II) appear to increase [18]. The CPT

II gene promoter does not contain a fat activated response element, but it does respond to

specific fibrate induction through peroxisome proliferator activated receptor ac [18].

Carnitine Octanoyltransferase

Evidence for the existence of a third carnitine acyltransferase that prefers acyl

groups composed of 6-22 carbon atoms emerged in 1971. This suggestion was based on

studies that examined the substrate specificity profiles of muscle mitochondria of pigeon

breast; and mitochondria of liver, heart, and testis of rat [38, 39]. Subsequent substrate

profiles of mitochondria from beef heart showed that the enzyme preferred

octanoyl-carnitine [40]. Obviously this was surprising, since the previously identified

carnitine acyltransferases prefer substrates with acyl groups of 2-6 (CAT) and 8-18 (CPT)

carbon atoms.

The gene that encodes COT has been mapped to chromosome 7, position 7q21.1,

and encodes a protein composed of 612 amino acids [18, 30]. Significant COT activity

has been observed in the peroxisomes of the liver, intestine, and ovary [41]. The enzyme

contains a peroxisomal targeting signal at the C-terminus; however, the signal differs in

the human (Thr-His-Leu), rat (Ala-His-Leu), and bovine (Pro-His-Leu) forms [18].









Interestingly, like CPT II, COT levels increase after induction by peroxisomal

proliferators. This suggests a parallel mechanism between the two enzymes [18]. It was

also found that diets heavy in fish oils induce COT expression in rats [18]. Since fish oils

are composed of very long-chain acyl groups, these fatty acids must be shortened through

P-oxidation in the peroxisome before they can be transferred to the mitochondria for

energy production, therefore demonstrating a need for COT activity.

Carnitine Acetyltransferase

The gene that encodes CAT has been mapped to chromosome 9, position 9q34.1,

and displays an alternative splicing pattern that results in a mitochondrial and

peroxisomal form of the enzyme [42, 43]. The mitochondrial form of CAT is composed

of 626 amino acid residues, with a mitochondrial targeting signal at the N-terminus

(residues 1-28) [18]. The peroxisomal form of CAT starts at the second codon, which is

downstream from the first mitochondrial start codon, and is composed of 605 amino acid

residues with a peroxisomal targeting signal at the C-terminus [18]. CAT activity has also

been observed in the endoplasmic reticulum [18]. It is not known if this form of the

enzyme is analogous to the mitochondrial or peroxisomal form, or if it is a separate

entity.

Early acyl group specificity studies with CAT isolated from pigeon breast muscle

showed that the enzyme prefers substrates with 2-4 carbon atoms [44]. Further studies

showed that acyl groups with 12 or more carbon atoms inhibit the enzyme by competing

with acetyl-CoA [44]. However, an endogenous inhibitor of CAT activity has not been

identified.









Choline Acetyltransferase

Choline Acetyltransferase (ChAT) is the fourth member of the L-carnitine/choline

acyltransferase enzyme family. This enzyme catalyzes a similar reaction to that of CAT,

except it transfers an acetyl group from CoA to choline to form the neurotransmitter

acetylcholine (ACh). ACh is implicated in both the central and peripheral nervous

systems. In the central nervous system, ACh is associated with sleep, memory, and

learning. In fact, ACh appears to be the primary neurotransmitter involved with memory

and learning. In the peripheral nervous system, ACh is involved in smooth muscle

contraction, and helps to regulate heart rate.

The gene that encodes ChAT has been mapped to position 10ql 1.2 [45, 46].

Human ChAT cDNA was cloned, and appears to encode two forms of the enzyme of

differing lengths, with 748 and 630 amino acids. The form with 748 amino acids contains

a N-terminal extension of 118 amino acids compared to the other form. It is not clear if

this form is actually synthesized in vivo [47].

The enzyme ChAT is located in the nerve terminals of cholinergic nerves in the

central and peripheral nervous systems. In the central nervous system,

ChAT-immunoreactive neurones are detected in the medial septal nucleus; the nucleus of

the diagonal band of Broca; the basal nucleus of Meynert; the caudate nucleus; the

putamen; the nucleus accumbens; the pedunculopontine tegmental nucleus; the

laterodorsal tegmental nucleus; the medial habenular nucleus; the parabigeminal nucleus;

some cranial nerve nuclei (including oculomotor, trochlear, trigeminal, abducent, facial,

prepositus, vagus, accessory, and hypoglossal); and the anterior horn of the spinal cord

[47].









The rate-limiting step of ACh synthesis is the uptake of choline into the cholinergic

neuron. However, ChAT is also regulated by several factors and chemicals (such as

cholinergic differentiating factor/leukemia inhibitory factor, ciliary neurotrophic factor,

camp, and retinoids) [47].

Regulation of L-Carnitine Acyltransferases

As previously described, CPT I is regulated by the reversible inhibitor, malonyl-

CoA (Figure 1-7). Malonyl-CoA is an intermediate in fatty acid synthesis, produced

during the first step in a reaction catalyzed by acetyl-CoA carboxylase (ACC). This is the

rate-limiting step in fatty acid synthesis, and is regulated by adenosine monophosphate

(AMP) activated protein kinase (AMPK) [48]. AMPK functions as a fuel sensor to

monitor the status of the cell. Once activated, AMPK inactivates ACC by

phosphorylation, resulting in decreased levels of malonyl-CoA and increased CPT I

activity [48]. Interestingly, AMPK activity increases after exercise to promote fatty acid

oxidation in muscle cells, to generate ATP and glucose uptake to replenish glycogen

stores [48]. A decrease in AMPK activity has been observed in the hypothalmus of rats as

a result of refeeding after a fast, glucose and insulin administration, or the injection of

leptin [49]. Thus it appears that CPT I activity depends on the metabolic state of the

organism.


CoA S o0


0 0

Figure 1-7. Structure of malonyl-CoA, an inhibitor of L-carnitine palmitoyltransferase I

Recently, it was shown [50] that the hormone leptin modulates fatty acid

metabolism through AMPK. The 16 kDa hormone is synthesized by white adipocyte









tissue in proportion to adipose cell size and number, and in response to stimuli [51].

Leptin plays a central role in regulating food intake, energy expenditure, and

neuroendocrine function. Leptin crosses the blood-brain barrier to reach its receptors

(cytokine type 1), which signal through the janus kinase signal transducer and activator of

transcription (JAK-STAT) pathway [51]. The activation of these receptors leads to the

decrease of food intake, which is mediated by several downstream mechanisms.

Recently, it was shown that leptin activates AMPK by phosphorylation of the alpha 2

catalytic subunit of AMPK, and therefore stimulates fatty acid oxidation and glucose

uptake [50]. In addition, leptin also suppresses the activity of ACC to further promote

fatty acid oxidation [50]. However, fasting and sympathetic activity have been shown to

modulate leptin expression and serum leptin [52].

In particular, suppression of fatty acid oxidation in vivo or in isolated adipocytes

increases leptin expression, indicating that metabolites of fatty acid oxidation may

regulate leptin expression [52]. The activity of CPT I could affect leptin expression. A

recent experiment showed that inhibition of fatty acid oxidation via inhibition of CPT I

increased leptin expression in fasting rats [52]. The mechanism of the increased

expression is unclear, and needs further work to sort out. However, it is thought that

CPT I is a key modulator of leptin expression in fasting rats, perhaps through a complex

feedback mechanism in which suppression of fatty acid synthesis (due to inhibition of

fatty acid oxidation and increase in adipocyte content) causes a decrease in related

metabolites (such as malonyl-CoA) that may suppress leptin expression [52].

L-Carnitine/Choline Acyltransferase Disorders

The enzymes of the L-carnitine/choline acyltransferase family play vital roles in

fatty acid oxidation. These metabolic roles are so important, that a deficiency in CPT









and/or CAT can cause serious health conditions. In fact, CPT deficiencies are among the

most common causes of inherited fatty acid oxidation disorders. Interestingly, increases

in CAT activity have been shown to have beneficial effects and prevent age-related

neurodegeneration.

L-Carnitine/Choline Acyltransferase Deficiencies

The first case of CPT I deficiency was described in 1981, with a patient who

presented fasting induced non-ketotic hypoglycemia and displayed an absence of fatty

acid metabolites in the blood and urine [53]. Since then, approximately 30 patients have

been reported worldwide. Typically the symptoms, hypoketotic hypoglycemia (usually

associated with fasting or a viral illness), and hepatomegaly (with or without acute liver

failure), with subsequent hypoglycemic attacks, begin from birth to 18 months of age

[32]. The outcome of the disease is variable. A persistent neurological deficit is common,

but not constant. Recurrent episodes are frequent up to five years of age, but are

successfully treated with symptomatic therapy [32].

CPT II deficiency is more common than CPT I deficiency and can be classified

according to age of onset, with adult, infant and neonatal forms. The adult form was first

described in 1973, and now more than 150 patients have been reported [32]. The

symptoms begin between 6 and 20 years of age, but the age of onset can also be over 50

years or as early as 4 years [32]. This disease is characterized by muscle pain and

stiffness, which can be complicated by acute renal failure (as a result of myoglobinuria)

and respiratory insufficiency (secondary to respiratory muscle involvement) [32, 54].

These symptoms are triggered by exercise, fasting, high-fat intake, exposure to cold, mild

infection, fever, and/or emotional stress and can be treated by restriction of long-chain fat









intake along with medium-chain triglyceride supplementation, and frequent meals with

extra carbohydrate intake before and during exercise [32, 54].

The infantile form of CPT II deficiency has been reported in 10 patients and is

characterized by recurrent attacks of acute liver failure with hypoketotic hypoglycemia,

that can result in coma and seizures, and transient hepatomegaly [32]. In approximately

half of these cases the heart muscle was also symptomatic and displayed dilated and

hypertrophic cardiomyopathy or arrythmias and conduction disorders [32]. Fasting or

illness also triggers these attacks. The treatment is the same as with the adult form of the

disorder, but the prognosis is not as good. Of the 10 reported patients, 6 died within the

first year of life due to heart beat disorders, while the other 4 patients died from Reye

syndrome [32].

The neonatal form of CPT II appears to be even more severe than the infant form.

To date, 13 patients have been reported and all died shortly after birth, except for one

patient who recovered after resuscitation [32]. These patients had dramatic malformations

including dysmorphic features, cystic renal dysplasia, and neuronal migration defects in

addition to hypoketotic hypoglycemia and cardiomyopathy [32, 54].

Over 25 mutations have been identified in the gene encoding CPT II in the adult,

infantile, and neonatal forms of CPT II deficiency [32]. Two of these mutations,

Ser113Leu (allele frequency of 60% in European patients) and Pro50His (allele frequency

6.5%), are common in the adult form of the disease [32]. Most of the mutations occur in

amino acid sequences that are highly conserved among the camitine acyltransferase

enzyme family. However, it is difficult to understand the enzymatic role of these residues

and thus the molecular basis of the disease without a three-dimensional structure.









As previously discussed, CAT deficiencies also have serious medical

consequences. A decrease in CAT activity has been observed and presents clinical

symptoms of progressive ataxic encephalopathy [55]. This disease is characterized by

psychomotor and growth retardation and is associated with weakness, ataxia,

ophthalmoplegia, and increased blood pyruvate levels [55]. In the affected patient the

COT activity was slightly decreased in the brain (approximately 60%), CPT activity was

slightly increased in the liver and kidney, and CAT activity was significantly decreased in

all tissues examined [55]. It was suggested that the CAT defect was coupled to a

deficiency in acetyl-CoA metabolism in the mitochondria of the brain and that the CAT

defect affected oxidation [55].

Deficiencies in ChAT have also been observed. Mutations in the gene encoding

ChAT have been shown to cause congenital myasthenic syndrome (CMS) [56]. CMS is a

genetic disorder that presents in infancy or childhood and is characterized by generalized

weakness and fatigability of voluntary muscles including those controlling movement,

eye movement, swallowing, and breathing. These symptoms are a result of disorders at

the neuromuscular junction, including defects in the synthesis and release of ACh, in the

gene encoding acetylcholinesterase, and the ACh receptor.

Recently ChAT has also been implicated with CMS. Sequencing of the gene

encoding ChAT in 5 patients with CMS revealed 10 recessive mutations [56]. Three of

these mutations resulted in reduced expression of the enzyme while kinetic studies with

mutant forms of the enzyme showed decreased catalytic activities with eight mutations,

and a complete loss of catalytic activity with one of the mutations [56]. The molecular

basis for these mutations is unclear.









L-Carnitine Acetyltransferase Increases

Increases in CAT activity have also been observed and appear to protect against

age-associated neurodegeneration. A hallmark of aging is mitochondrial decay which is

characterized by high generation of oxidants, a decline in the activity of electron transport

complexes, and a decrease in the fatty acid composition of cardiolipin (an essential

phospholipid for normal function of the mitochondria) [57]. Recently, it was shown that

acetyl-L-camitine reverses age-associated mitochondrial decay in rats. Specifically,

feeding rats acetyl-L-camitine decreased mortality, improved nerve regeneration,

protected neurons from the toxicity of mitochondrial uncouplers or inhibitors, increased

cardiolipin contents, and increased the activity of cytochrome c oxidase [57]. Feeding rats

a combination of acetyl-L-camitine and lipoic acid had an even more beneficial effect.

This treatment reversed age-related mitochondria decline, improved hepatocellular

ascorbate levels, increased metabolism, and decreased the oxidative stress [58]. In

addition, feeding rats acetyl-L-carnitine and lipoic acid also affected cognitive function.

These rats showed an improved performance on memory function tasks, a decrease in

brain mitochondrial structural decay, and a decrease in oxidative damage in the brain

[59]. Further studies showed that acetyl-L-camitine may protect CAT from interacting

with aldehydes from lipid peroxidation while lipoic acid may prevent proteosomes from

oxidative modification and play a role in reactivating CAT by increasing its substrate

binding affinity [60]. However, the molecular basis of acetyl-L-carnitine and lipoic acid

treatment is unclear.

To further understand the (a) mechanism of fatty acid oxidation and how it is

regulated (b) carnitine/choline acyltransferase disorders and (c) the benefits of acetyl-L-

carnitine and lipoic acid it is of the utmost importance to study this enzyme family.









Structure-Function Studies

Enzymatic studies of the carnitine acyltransferase family began in the late 1960s

with CAT. Understanding of this enzyme family has significantly increased with the

advent of biological tools such as sequencing, site-directed mutagenesis, and

crystallography. These studies have helped to elucidate the catalytic activity of these

enzymes and to identify specific residues involved in the mechanism of catalysis, with

the ultimate goal of understanding endogenous regulation and developing exogenous

regulators.

Early kinetic studies with CAT (isolated from pigeon breast muscle) determined

that the enzyme follows a random ordered, rapid equilibrium mechanism that involves

the formation of an intermediate [61]. Further studies with CAT and bromoacetyl-L-

carnitine showed that there are two steps to the reaction, reversible binding and

irreversible alkylation [62].

A combination of electrophoretic studies, chemical modification, and substrate

analogues implicated histidine as the active site residue that is alkylated during catalysis

[62]. Further inactivation studies confirmed that a histidine residue is crucial to activity

and that it serves as a general acid/base catalyst [63]. In this role, the histidine residue

removes a proton from L-carnitine or CoA (depending on the direction of the reaction) to

promote the formation of an intermediate [18].

Recently, the x-ray crystal structures of human peroxisomal CAT as well as the

free and substrate bound forms of mouse CAT have been determined [64, 65]. The

overall structure reveals that CAT is a monomeric protein consisting of 17 helices and 14

strands that form two equally sized domains [64, 65]. The two domains are tightly

associated with each other and form a central, active site tunnel. The active site histidine,









His322, is located in the middle of the active site tunnel, suggesting that L-camitine and

acetyl-CoA bind to opposite sides of the tunnel [64].

Several structure function experiments have also been aimed at determining the

structural basis of malonyl-CoA sensitivity of CPT I. Since this regulation affects fat

metabolism, it is conceptualized as a potential drug target for type II diabetes.

Interestingly, the muscle and liver isoforms of CPT I have different malonyl-CoA

sensitivities, an IC5o of 0.03 [LM and 3 [LM, respectively [66]. Therefore, the primary

question is what is the structural basis of the malonyl-CoA sensitivity difference between

these two isoforms.

In 1997, Zhu et al., developed a high level expression system for CPT I in the

yeast, Pichiapastoris [67, 68]. This organism is devoid of endogenous CPT activity and

therefore, is an excellent expression system. From this system, several structure function

studies, including deletion mutations, site-directed mutagenesis, and chimeric protein

studies, have been carried out to characterize malonyl-CoA binding.

The deletion mutation studies were designed to characterize the N-terminal of CPT

I. The first 124 amino acid residues are conserved in M-CPT I and L-CPT I, but not in

CPT II (which is not responsive to malonyl-CoA). Therefore, it was predicted that this

region is the malonyl-CoA sensitive site. Several deletions of this area were constructed

to elucidate the amino acid residues involved in malonyl-CoA sensitivity. In a particular

study the residues 1-18, 1-35, 1-52, 1-73, 1-83 and 1-129 in L-CPT I were deleted and the

mutant forms of the enzyme were expressed, purified, and the kinetic parameters were

determined and compared to the wild type form of L-CPT I [33]. Deleting the first

eighteen residues of L-CPT I completely abolished malonyl-CoA inhibition and high









affinity malonyl-CoA binding [33]. However, to achieve insensitivity to malonyl-CoA

and loss of high affinity malonyl-CoA binding in M-CPT I, the residues 1-28 and 1-39

must be deleted [69]. This suggests that the first 18 amino acid residues play a different

role in the muscle and liver forms of CPT I.

Another interesting deletion mutation study demonstrated that, the removal of

amino acid residues 3-18 and 1-80 of M-CPT I decreases malonyl-CoA sensitivity and

that the deletion of residues 19-30 increases malonyl-CoA sensitivity [70]. This was the

first demonstration of positive and negative determinants of malonyl-CoA sensitivity.

To evaluate specific residues within this region site-directed mutagenesis studies

were performed. Substitution of glutamate at position 3 by alanine or glutamine results in

a complete loss of malonyl-CoA sensitivity and high and low affinity binding while

other kinetic parameters remain intact [71, 72]. Furthermore, an aspartate at position 3

can only partially substitute for the glutamate. This suggests that a negative charge at that

position is essential and that the longer side chain of glutamate is essential for optimal

malonyl-CoA sensitivity [72]. In addition, substitution of serine at position 24 or

glutamine at position 30 by alanine resulted in a 10-20 fold decrease in IC5o and K, values

for malonyl-CoA [72]. These two residues may affect malonyl-CoA sensitivity through

interactions with glutamate 3 [72].

The involvement of the N-terminus in malonyl-CoA sensitivity was further

evaluated through the use of chimeric proteins from combinations of L-CPT I and M-

CPT I. These chimeric proteins were constructed from 3 segments of CPT I; (a) the

cytosolic N-terminus and transmembrane domain one, (b) the amino acid residues that

link the transmembrane domains together and (c) transmembrane domain two and the









cytosolic C-terminus [73]. The mutant forms of the enzyme were expressed and purified

and the kinetic parameters were determined. Replacing the first segment and the first and

second segments of L-CPT I with M-CPT I altered the Km values for both substrates, L-

carnitine and palmitoyl-CoA, and increased the malonyl-CoA sensitivity. However, M-

CPT I was much less susceptible to influence by the alternate segments [73]. This

suggests that transmembrane domains interact with each other and these interactions are

important for determining the kinetic parameters for the catalytic domain of L-CPT I

[73].

Despite the high sequence identity between the first two segments there is an

interesting amino acid difference in transmembrane domain one that might account for

the difference in malonyl-CoA sensitivity. An arginine residue is located at position fifty

two in M-CPT I whereas in L-CPT I this is a threonine residue. The arginine residue

might make the transmembrane domain shorter in M-CPT I than in L-CPT I and thus

extend the N-terminal domain to confer a different malonyl-CoA sensitivity [73]. The

charged residue may also restrict the mobility of the transmembrane helix and result in a

more rigid structure of M-CPT I that would also influence malonyl-CoA sensitivity [73].

An interesting observation with the pig form of L-CPT I helped to further the

chimera studies. This form of L-CPT I consists of 772 amino acids and shares an 86%

amino acid sequence identity with the rat form of L-CPT I and a 62% sequence identity

with the rat form of M-CPT I. Interestingly, the pig, human, and rat form of L-CPT I

have similar kinetic properties with respect to Km values for L-carnitine and acyl-CoA,

but the pig form of L-CPT I demonstrates a higher sensitivity to malonyl-CoA, more like

the human and rat forms of M-CPT I [74]. These characteristics demonstrate that the pig









form of L-CPT I behaves like a natural chimera between the L-CPT I and M-CPT I

isoforms.

To further understand the difference in malonyl-CoA sensitivity between M-CPT I

and L-CPT I, chimeras between the pig and rat forms of L-CPT I were constructed. These

studies demonstrated that the Km value for the substrates and the malonyl-CoA sensitivity

depends on the C-terminal domain [75]. In addition, replacing different regions of the C-

terminus of rat L-CPT I, 128-220, 220-590 or 590-772, with pig L-CPT I showed that the

C-terminus behaves as a single unit and the degree of sensitivity is determined by the

structure of this domain [75]. Additional deletion studies showed that the first 18 residues

determine the malonyl-CoA sensitivity; this is most likely due to interactions with the C-

terminal [75]. It was then concluded that the difference in malonyl-CoA sensitivity

between pig and rat L-CPT I is due to the difference in the interaction of the first 18

residues with the C-terminal of these forms [75].

This was the first study that described how interactions between the N- and C-

terminals could affect malonyl-CoA sensitivity in pig and rat L-CPT I. There are several

unresolved issues with the CPT I system. These studies demonstrate the need for a three

dimensional structure to resolve these issues.

Specific Aims

The specific aims of this project are to further understand this enzyme family at the

protein level. The first specific aim is to identify and characterize amino acid residues

that are involved with substrate recognition and binding. Determining the structure of

CAT binding to the substrate L-carnitine will identify specific substrate binding residues .

These residues will then be functionally and biologically characterized by an enzymatic

assay and a novel genetic assay.






29


Determining the structure of ChAT will also identify specific residues involved in

catalysis and substrate recognition and binding. These two enzymes will most likely

share several structural features. However, there are differences in L-camitine and

choline that may be reflected in the enzyme structures. In addition, the information

gleaned from these structure studies can be used to determine the molecular basis of

congenital myasthenic syndrome (CMS).














CHAPTER 2
MATERIALS AND METHODS

This chapter describes the general methods used to express, purify, crystallize and

determine the structure of ChAT and CAT in complex with L-carnitine, and to

characterize essential enzymatic amino acid residues of CAT. The results of these studies

are discussed in the following chapters.

The first goal of this project was to identify residues of CAT that interact with the

substrate, L-carnitine, and then characterize these residues through an enzymatic and

biological assay. The identification and characterization of these residues will provide

further insight into the reaction mechanism of CAT and will help to advance

understanding of other members of the L-camitine/choline acyltransferase family. The

general experimental design to accomplish these goals was as follows.

* Determine the structure of CAT with the substrate L-carnitine
* Mutate specific residues of CAT using site-directed mutagenesis
* Express and purify the mutant forms of CAT
* Determine kinetic parameters of the mutant forms
* Evaluate biological activity of mutant forms using a genetic selection system

The second specific aim of this project was to determine the structure of ChAT.

The structure of ChAT can not only be used to gain further structural and mechanistic

understating of the L-camitine/choline family, but also to understand the molecular basis

of the disease, congenital myasthenic syndrome (CMS).









Determine Structure of Carnitine Acetyltransferase with L-Carnitine

To determine the x-ray crystal structure of CAT with the substrate L-carnitine, the

enzyme was expressed, purified, and co-crystallized with L-camitine. The crystals were

then x-rayed and the structure was determined from the diffraction data.

Expression

Recombinant human peroxisomal CAT was expressed in Eschericia coli B834

cells. The coding region of CAT was amplified by polymerase chain reaction (PCR) and

cloned into the pQE-9 (Qiagen) expression vector (Figure 2-1). This expression vector

allows for high-level expression of proteins with a six-histidine tag on the N-terminus.

Interestingly, the histidine tag does not affect enzyme activity and actually improves

expression levels. The vector also contains an ampicillin selection marker and the T5


LZ_---

PT5s -IclO-lcO-EN u ATsiG ;S


pQE-9
3A kb


cdfli


Figure 2-1. Diagram of the pQE-9 expression vector. PT5: T5 promoter, lacO: lac
operon, RBS: ribosome-binding site, ATG: start codon, 6xHis: 6xHis tag
sequence, MCS: multiple cloning site with restriction sites (BamH, SalI, PstI,
HindII), Stop codons: stop codons in all three reading frames, Col El: Col El
origin of replication, Ampicillin: ampicillin resistance gene. Source: Qiagen.









promoter, which is tightly regulated by the lac operon. The vector containing the CAT

coding region was then transformed into E. coli cells. These cells were grown at 300C to

an optical density of 0.5-0.7 at 600 nm. Expression was induced by adding isopropyl-b-

D-thiogalacto-pyranoside (IPTG) to a final concentration of 1 mM. Cells were then

grown for an additional 2 hours. The bacteria was harvested by centrifugation and stored

at -800C until further use.

Extraction

Recombinant CAT is soluble and was extracted by cell lysis using the French

Press. To purify the protein, the E. coli extract, prepared in 20 mM sodium phosphate

buffer pH 7.6, was added to Ni-nitriloacetate-agarose (Ni-NTA) (Qiagen), which was

equilibrated with the extraction buffer. After a 1 hour incubation the mixture was

transferred to a column and the buffer was allowed to flow through, then collected. The

imidazole ring of the histidine residues of the histidine tag binds to the nickel ions

immobilized by the NTA groups of the matrix (Figure 2-2). Thus the enzyme is retained

on the column. The column was then washed batch wise with buffer containing

increasing concentrations of NaCl from 0 to 2 M. Free imidazole can also bind to the

nickel ions and disrupt the binding of the histidine residues. This mechanism was

employed to elute the enzyme, using an imidazole gradient from 5 to 150 mM, was

employed to elute the enzyme. Next, the enzyme was applied onto a MonoQ column

(Pharmacia) that was equilibrated with 20 mM sodium phosphate buffer pH 7.6. The

enzyme was eluted with a linear salt gradient from 0 to 1 M and then

concentrated/dialyzed using Ultra-free Tangential membranes (Millipore). The purified









enzyme was either used immediately or stored at -800 C in 20 mM sodium phosphate

buffer pH 7.6 containing 40% glycerol.

R

NH

C=0 0
/ PC CH /

/NH -HNi CH2 H 0O /
CH CH -CH **:H CH CH

c,\ "%,/ oC
CN, 0 : 0 || /I

NH
I
R

Figure 2-2. Interaction of histidine residues of 6xHis tag and the Ni-NTA agarose.
Source: Qiagen

Purification

Purity of the purified enzyme was analyzed by SDS-polyacrylamide gel

electrophoresis while the concentration of the enzyme was determined by the BIO-RAD

Bradford Protein Assay with bovine albumin as the standard. Protein expression was

further characterized by Western Blot. In this procedure, the enzyme was

electrophoretically transferred to an Immobilon-P nylon membrane and incubated for

1 hour in blocking buffer consisting of 10 mM Tris-HCl buffer pH 7.4, 150 mM sodium

chloride, 5% nonfat dry milk, and 0.2% Nonidet P-40. After this incubation the

membrane was incubated overnight with an affinity purified anti-CAT antibody diluted in

fresh blocking buffer. The next day the membrane was washed three times with a

solution containing 10 mM Tris-HCl buffer pH 7.4, 150 mM sodium chloride, 0.25%

sodium-7-deoxycholate and 0.1% SDS and then washed three more times with a solution









containing 10 mM Tris-HCl buffer pH 7.4, and 150 mM sodium chloride. The enzyme

was visualized after incubation with a goat anti-rabbit antiserum coupled to alkaline

phosphatase and 5-bromo-4-chloro-3-indolyl phosphatase linked with p-nitroblue

tetrazolium chloride as the color reagent.

Crystallization

Initial crystallization experiments were carried out using the hanging-drop

vapor-diffusion method in a 24-well tissue culture Linbro plate (MP Biomedicals Inc.) at

three different temperatures, 4, 16, and 220C. A sparse-matrix crystallization screen

based on the original report by Jancarik & Kim (1991) [76] was used and promising

conditions were further optimized with respect to pH and precipitants. Each drop of

CAT-L-camitine was formed by mixing 3 uL of 20 mg/mL CAT, 3 uL of

1 mM L-camitine, and 6 uL of reservoir solution that was 50 mM Bis-Tris buffer pH 6.2,

100 mM sodium chloride, and 12-15% PEG 8000.

Mutate Specific Residues of CAT

Specific residues of CAT thought to be involved in substrate binding were

investigated using the QuickChange (Stratagene) site-directed mutagenesis procedure.

This method (Figure 2-3) has proven to be very effective, with an efficiency of

80%. First, oligonucleotide primers containing the desired mutation are incubated with a

double-stranded DNA vector including the gene that encodes CAT and PfuTurbo DNA

polymerase. During temperature cycling, the polymerase extends the primers generating a

mutated plasmid containing staggered nicks. After temperature cycling, the plasmid is

treated with Dpn I endonuclease. This endonuclease is specific for methylated and

hemimethylated DNA and will digest the parental DNA template since DNA isolated










from almost all E. coli strands is dam methylated and thus subject to Dpn I digestion.

Conversely, newly synthesized DNA will not have been methylated and is therefore not



Step 1
"-.'j I.,i 1 Preporation


Step 2
Temperalure ?-ycrin












Step 3





Step 4
"ni. t ir ",'Pl'". .r


P19mers
c











~lMualrd plcirnid
S c'(oa;ns picked

,


Figure 2-3. Overview of the QuickChange site-directed mutagenesis method. Step 1, gene
in plasmid with target site for mutation. Step 2, denature the plasmid and
anneal the oligonucleotide primers containing the desired mutation. Use
PfuTurbo DNA polymerase to extend and incorporate the mutagenic primers
resulting in nicked circular strands. Step 3, Digest the methylated, non-
mutated parental DNA template with Dpn I. Step 4, transform the circular,
nicked dsDNA into XL-1 blue supercompetent cells. After transformation, the
XL-1 blue supercompetent cells repair the nicks in the mutated plasmid.
Source: Stratagene.









susceptible. The nicked vector DNA containing the mutation is then transformed into

XL 1-Blue supercompetent cells, which will repair the nicks in the vector. The colonies

are plated and grown on LB media. Individual colonies are screened to ensure correct

insertion of the mutation by checking for a restriction site that has been introduced by the

primers. The high fidelity of PfuTurbo DNA polymerase and the low number of thermo

cycles contribute to the high mutation efficiency and low potential for random mutations.

However, to ensure that random mutations have not been introduced, the correct

sequence was verified by DNA sequencing (UF DNA Sequencing Core) before further

experiments are completed.

The mutant forms of the enzyme were expressed and purified as previously

described. In addition, expression levels of the mutant forms of the enzyme were

compared to the wild type form by SDS-PAGE and Western Blot, as previously

described.

Evaluate Enzymatic Activity

The kinetic parameters were then determined using a continuous,

spectrophotometric assay. This is an indirect measure of the forward direction of the

reaction. In the forward direction the substrates, acetyl-CoA and L-carnitine, yield the

products acetyl-L-carnitine and CoASH. The release of CoASH is monitored

spectrophotometrically with the reagent 5,5'-dithiobis-2-nitrobenzoic acid (DTNB).

DTNB reacts with the free sulfhydryl group of CoASH to form the yellow colored 5'-

thio-2-nitrobenzoate (NTB), which absorbs strongly at 412 nm (Figure 2-4). Activity was

measured as an increase in absorbance at 412 nm over time.

Kinetic values were used to interpret the data from the mutant and wild type forms

of the enzyme. The initial velocity of the enzymes with different concentrations of









substrates (L-carnitine and acetyl-CoA) was determined. This data was fitted to the

Michaelis-Menton equation using nonlinear regression with the program Prism

(Graphpad). The values of Vmax, Km, kcat and kcat/Km were calculated to determine how the

02N 02N
COOH COOH




S + CoA-SH --
S S HS
S I

CoA


COOH COOH
02N 02N

NTB
DTNB

Figure 2-4. Schematic representation of the CAT assay. In this reaction DTNB reacts
with the free sulfhydryl group of CoA-SH to form the chromogenic NTB.

amino acid mutations effect the enzyme. The Vmax is the theoretical maximum velocity of

the enzyme reaction expressed as the product formed per unit of time. Km is the

concentration of the substrate at half the maximum velocity and is expressed as a

concentration. The turnover number, or the kct value, is the maximum velocity (Vmax)

divided by the concentration of the enzyme and gives information about the efficiency of

the enzyme. The kcat/Km is a measure of the catalytic efficiency. By comparing these

values for the mutant and wild type forms of the enzyme we will be able to understand

how these residues interact with the substrates and determine their role in catalytic

function.









Evaluate Biological Activity

The yeast strain used in this study is derived from the strain FY23. In this strain,

the CIT2 and CAT2 gene were deleted [77]. Yeast were either grown on YPD media,

containing 1% (w/v) yeast extract, 2% (w/v) peptone, and 2% (w/v) glucose, or YND

media, containing 0.67% (w/v) yeast nitrogen base without amino acids and 2% (w/v)

glucose. Media containing ethanol (YNEC) as the carbon source was made with 0.67%

(w/v) yeast nitrogen base without amino acids and 3% (v/v) ethanol. This media was also

supplemented with L-carnitine at a concentration of 10 mg/L (YNEC).

Plasmid containing wild type and mutated CAT was introduced into the mutant

yeast strain by the lithium acetate method. Briefly, yeast cells from an overnight culture

of yeast, grown in YPD media, were harvested and washed in buffer containing 10 mM

Tris, 0.1 mM lithium acetate, and 1 mM EDTA. Next, 50% PEG 4000, 1 M lithium

acetate, water, and plasmid DNA were added to the cells, in the described order, and the

mixture was vortexed. After a 25-minute incubation in a 420C water bath, the cells were

centrifuged and the supernatant was removed. The remaining cells were resuspended in

water, plated onto YND agar plates and grown in a 300C incubator. The yeast cells that

were plated onto YND agar plates were allowed to grow for 2-3 days before being replica

plated onto YNEC and YND agar plates. Growth on YNEC and YND plates was

monitored for several days.

Structural Determination of Choline Acetyltransferase

To determine the structure of ChAT, the enzyme was expressed, purified, and

crystallized following the same general procedure as described previously with CAT and

L-camitine. Briefly, recombinant rat ChAT was expressed in E. coli B834 cells and









extracted by cell lysis with the French Press. The extracted enzyme was applied to a

column containing Ni-NTA agarose (Qiagen) and washed batch wise with buffer

containing increasing concentrations of NaCl from 0 to 2 M and eluted by an imidazole

gradient from 5 to 150 mM. The enzyme was then applied to a MonoQ column

(Pharmacia) and eluted with a NaCl gradient from 0 to 1 M and then

concentrated/dialyzed using Ultra-free Tangential membranes (Millipore). The purified

enzyme was either used immediately or stored at -800C in 20 mM sodium phosphate

buffer pH 7.6 containing 40% glycerol.

The purity of the enzyme was evaluated by SDS-polyacrylamide gel

electrophoresis and Western Blot while the concentration of the enzyme was determined

by the BIO-RAD Bradford Protein Assay with bovine albumin as the standard. Purified

ChAT was crystallized using the hanging-drop vapor-diffusion method in a 24-well tissue

culture Linbro plate (MP Biomedicals, Inc.) by mixing 3 uL of 10 mg/mL ChAT with

3 uL of the reservoir solution, containing 50 mM MES buffer pH 6.0, 100 mM sodium

chloride, and 8-10% PEG 8000 at 40C.














CHAPTER 3
STRUCTURE DETERMINATION OF CARNITINE ACETYLTRANSFERASE WITH
L-CARNITINE

The goal of the first specific aim was to identify and characterize essential amino

acid residues of human CAT. These objectives were met by determining the structure of

CAT with its substrate L-camitine, using an enzymatic assay to evaluate the catalytic role

of specific residues, and by developing a genetic selection system to evaluate the

biological role of these residues. This chapter discusses structure determination of the

enzyme complex and residues involved in catalysis.

Introduction

L-camitine is a highly polar zwitterionic quaternary amine carboxylic acid present

in some prokaryotes and all eukaryotes. In prokaryotic cells L-camitine serves either as a

nutrient (such as a carbon and nitrogen source) or as an osmolyte. However, in eukaryotic

cells, L-camitine serves exclusively as a carrier of acyl moieties through various

subcellular compartments [22]. L-carnitine helps to remove toxic accumulations of

poorly metabolized fatty acids in the mitochondria but, more significantly, the camitine

system plays an essential physiological role in fatty acid P-oxidation and in the

maintenance of acyl-CoA pools [78, 79]. L-carnitine and its esters are transported across

the inner mitochondrial membrane to facilitate the entrance of long chain fatty acids into

the mitochondria for P-oxidation [19, 78].

The physiological actions of L-camitine are mediated via a sodium dependent high

affinity carnitine/acylcarnitine plasma membrane transporter, a carnitine/acylcamitine









translocase located on the inner mitochondrial membrane, as well as a group of

L-camitine utilizing enzymes, the camitine acyltransferases [18, 22, 78]. As previously

discussed, carnitine acyltransferases (CAT, COT and CPTs) facilitate the reversible

transfer of acyl groups between CoA and L-carnitine, and fall into three main classes

which differ in their acyl-chain length selectivity, sub-cellular compartmentalization and

physiological function [19, 37, 42]. CAT has a substrate preference for short-chain

acyl-CoAs and is found in the mitochondrial matrix, peroxisome and the endoplasmic

reticulum. COT is predominantly localized in peroxisomes and has a substrate preference

for medium length acyl-CoAs. CPTs are found both in the outer mitochondrial membrane

(CPT I) and the mitochondrial matrix (CPT II) with a substrate preference for long-chain

acyl-CoAs.

Decreases in CAT activity (particularly in the brain) have been described and

present clinical symptoms of progressive ataxic encephalopathy. These symptoms are

most likely due to a decrease in acetyl-CoA available for the citric acid cycle in the

mitochondria [55]. In addition, recent experiments have suggested that CAT also plays a

role in age-associated neurodegeneration. It was found that feeding rats acetyl-L-camitine

and lipoic acid increases CAT activity, improves performance on memory function tasks,

reduces brain mitochondrial structural decay, and reduces oxidative damage in the brain

[58-60]. One explanation is that over time CAT becomes inactivated by aldehydes

formed by lipid peroxidation and that acetyl-L-camitine has a protective effect while

lipoic acid reduces lipid peroxidation and reactivates CAT [60].

In spite of its importance, fundamental questions remain about the structure and

function of the camitine acyltransferases. To further understand the molecular basis of









the benefits of acetyl-L-carnitine and lipoic acid on CAT, camitine acetyltransferase

disorders and amino acid residues involved in catalysis, it is important to know the

structure and mechanism of action of this enzyme. Previous chemical modification,

inactivation, site-directed mutagenesis, and three-dimensional structure studies have

helped to identify residues of CAT that interact with the substrates (L-carnitine and CoA)

and explain the mechanism of this enzyme.

As previously discussed, early studies with CAT isolated from pigeon breast

muscle demonstrated that the enzyme follows a random-order equilibrium reaction

mechanism that includes the formation of an intermediate ternary complex [61]. Further

studies revealed that the reaction product of CAT and radiolabeled

bromoacetyl-L-carnitine is the alkylation of a histidine residue. The alkylation

specifically occurs at the N-3 position and suggests that the active site of CAT contains a

histidine residue that is directly involved in catalysis [62]. The histidine residue is

believed to act as a general base, removing a proton from L-camitine or CoA (depending

on the direction of transfer) to promote the formation of a tetrahedral intermediate [18].

The crystal structure of human peroxisomal and mouse CAT have shown that the

enzyme is composed of 17 helices and 14 strands, that form two equal sized domains [64,

65]. These two domains (Figure 3-1) are tightly associated with each other and form a

central active site tunnel with His322 located in the middle of the tunnel, suggesting that

L-camitine and acetyl-CoA bind to opposite sides of the tunnel [64]. Several amino acid

residues have been implicated in the active site and putative binding residues of

L-camitine and CoA (L-carnitine: His322, Glu326, Tyr431, Arg443, Ser531, Thr532; CoA:









Ser533, Leu142, Lys398, Ser407 and Asp487) [64, 65]. To verify that these residues do indeed

interact with L-camitine and to identify additional L-camitine binding residues, the

structure of CAT with its substrate, L-carnitine, was determined.


S5 S4 S6 S7 SS Sl
514 SI3 01 S 510













N and C termini are indicated.
H917
H12

H 10
HOI



HH14















The experimental procedures involved in this project were described in Chapter 2.
Figure 3-1. Overall structure of human peroxisomal carnitine acetyltransferase [64].
Ribbon diagram representation depicts domains I (blue) and II (orange). The
N and C termini are indicated.

Materials and Methods

The experimental procedures involved in this project were described in Chapter 2.

Briefly, recombinant human peroxisomal CAT was expressed in bacterial cells with a

histidine tag and then purified using a column containing Ni-NTA agarose. The purified

enzyme was then co-crystallized with its substrate (L-carnitine) and the

three-dimensional structure was determined.









Results and Discussion

Structure Determination and Overall Structure of the Enzyme Complex

X-ray data of the CAT-L-carnitine crystal were collected at the Cornell High

Energy Synchrotron Source (CHESS). Data was collected using a crystal-to-detector

distance of 100 mm and a 1 oscillation angle per image. A total of 110 images were

collected from a single crystal at 100 K. Before flash freezing, the crystal was

cryo-protected in the crystallization reservoir solution using 15% PEG 8000 and 40%

glycerol. Data were indexed, integrated and scaled using the HKL2000 suite programs

DENZO and SCALEPACK [80]. Data were then phased using CNS program version 1.1

[81]. The structure was further refined using standard protocols within SHELXL97 [82]

with additional interactive modeling using the program O [83].

The crystal of the CAT-L-carnitine complex was isomorphous with the free CAT

crystals and belonged to the space group P212121 with unit cell dimensions of a = 137.6,

b = 84.7, c = 57.3 k. The structure was refined to a 1.8 k resolution. The final model of

the complex structure consisted of 590 amino acid residues (9-599), 465 water molecules

and the substrate L-carnitine.

Comparison of CAT-L-Carnitine with Free CAT

The crystal structure of the complex is almost identical to that of the free form of

CAT with a backbone root-mean-square deviation of 0.41 A (Figure 3-2) [64]. The

overall structure of the free form of CAT reveals that the enzyme is a monomeric protein

composed of two equally sized domains that are tightly associated with one another to

form a central solvent accessible tunnel, which constitutes the active site of the enzyme.

Examination of the complex structure shows that there are subtle changes in electron









density for the side chains of some active site residues in the L-camitine binding site, but

no major structural changes occur on substrate binding.



Disordered
loop




















Figure 3-2. Structural similarity and superimposition of CAT and CAT-L-carnitine. Coil
diagram shows the superimposition of the overall structure of CAT (gray) and
CAT-L-carnitine (green), the active site residues, His322, Tyr431, Thr444
Arg497, Phe545, Val548, and the water molecules, W621 and W679, with L-
carnitine in the active site. Domain I and II and the N- and C-termini are
indicated.

Comparisons of the structures of the CAT-L-camitine complex and free CAT

crystal structures show that there are slight changes in the conformation of different

surface regions and small displacement of active site residues. However, the only

significant conformational change occurs with Ser433. Upon L-carnitine binding the chi

angles of Ser433 change from 58 to -310, positioning the hydroxyl group 3.03 A from one

of the carboxylic oxygen atoms of L-camitine (Figure 3-3). The orientation of Ser433 and

L-camitine in CAT suggests that this residue may also contribute to transition state

stabilization.






















Figure 3-3. Representation of Ser433 positional change in free and L-camitine bound form
of CAT. The chi angle of Ser433 rotates from 58 to -31 upon substrate binding
to interact with the carboxylic oxygen of L-carnitine and/or contribute to
transition state stabilization.

L-Carnitine Binding Site

The binding site for L-carnitine in CAT is located at one entrance of the active site

tunnel, as was predicted by the previous structure of the free enzyme [64]. The binding

surface of L-carnitine is formed by helix 4, strand 1, strand 7, and strand 8 from domain I

and strand 9 to 10, helix 13 to 14, and strand 11 to 14 from domain II. Six residues,

His322, Tyr431, Thr444, Arg497, Phe545, Val548, and two tightly associated water molecules

(H20621 and H20679), in particular, constitute a distinct pocket to specifically recognize

and interact with L-carnitine (Figure 3-4).

Among these active site residues, His322 is completely conserved in all

L-camitine/choline acyltransferases and is believed to be the catalytic residue that serves

as a general base to catalyze fatty acyl transfer [18, 64, 65]. L-camitine binds to CAT

with its P-hydroxyl moiety aligned directly with His322 to form a hydrogen bond with the

N-3 atom of the imidazole ring (2.76 A) (Figure 3-5). This interaction supports the

catalytic role of His322, which has been proposed to deprotonate the P-hydroxyl group of

L-camitine or the thiol group of CoA to facilitate nucleophilic attack on the ester linkage






47






.. ; -










hr


-am i t ine -',"* 'f i[



Figure 3-4. Electron density map of the carnitine binding site of L-carnitine
acetyltransferase. Residues His322, Tyr431, Thr444, Arg497, Phe545, and Val548
and water molecules W621 and W679 are shown with L-carnitine.

of the acyl group [18, 64, 65]. Additionally, the alignment of L-carnitine and His322 is

stereospecific and explains why D-carnitine is the inactive enantiomer, since the

P-hydroxyl of D-carnitine would point in the opposite direction of His322 and would not

be accessible for acyl group transfer.

Besides His322 a water molecule (H20679) is within hydrogen bonding distance with

the P-hydroxyl moiety of L-carnitine. This water also interacts with the side chain

hydroxyl group of Ser533, which is the last residue in a Ser-Thr-Ser motif that is

conserved among all carnitine/choline acyltransferases. Ser533 is believed to play an

essential role in transition state stabilization during catalysis and may also be involved in

L-camitine binding through its interaction with H20679 [64, 84, 85].


















&

v-I


Figure 3-5. Close up figure of the active site of L-carnitine acetyltransferase-L-Carnitine.
This three-dimensional representation depicts the interaction between His322
and L-carnitine.
Hydrogen bond interactions seem to be the primary force for the recognition of
L-camitine since a hydrogen bond network is formed between the carboxylate oxygens
(01 and 02) of L-carnitine and a number of active site residues (Figure 3-6) [65, 85].
Specifically, Thr444 and an ordered active site water molecule, H20621, interact with one
oxygen (02) of the carboxylate group. The water molecule is further coordinated by three
residues in the active site, Trp81, Tyr86, and Glu326. The other carboxylic oxygen appears
to primarily interact with Tyr431 through a strong hydrogen bond (2.53 A). In addition to
the numerous hydrogen bond interactions with L-carnitine there is also an electrostatic
interaction between the negatively charged carboxylate group and the positively charged
guanidinium group of Arg497. Furthermore, the side chain NH of Arg497 forms hydrogen


F, I~








bonds with the hydroxyl groups of Tyr431 and Thr444, which may further influence L-

carnitine binding.


he 545


Figure 3-6. Two-dimensional representation of the interactions observed between L-
carnitine and L-carnitine acetyltransferase. Dashed lines denote hydrogen
bonds and distances in k. Hydrophobic interactions are shown as arcs and
radial spokes.
In contrast to the negatively charged carboxylate group of L-camitine, the

positively charged quaternary nitrogen does not appear to have electrostatic interactions

with negatively charged residues in the active site. However, the aromatic side chain of

Phe545 is close to the methyl carbon of the quaternary amine head group of L-camitine









(3.6 A). Thus, Phe545 appears to be the primary residue responsible for interacting with

the quaternary nitrogen of L-carnitine through pi-cation interactions (Figure 3-4 and 3-6)

[86, 87]. The de-localized electron clouds on the side chain benzene ring can interact with

and stabilize the positively charged quaternary amine head group of L-carnitine. Val548

may also contribute to the binding of the trimethyl amino head group of L-carnitine

through van der Waals interactions.

Comparison of CAT-L-Carnitine with Mouse CAT

Recently the crystal structure of mouse CAT in complex with L-carnitine was

determined to 1.9 A [65]. Based on sequence homology, human CAT and mouse CAT

share 90% identity at the amino acid level. Interestingly, the two proteins crystallize in

different space groups, human CAT-L-camitine crystallized in the orthorhombic space

group P212121 with one monomer in the asymmetric unit whereas mouse

CAT-L-carnitine crystallized in the monoclinic space group C2 which contains two

molecules in the crystallographic asymmetric unit [65, 85]. This may reflect a difference

in the unit cell packing arrangement of the proteins themselves, the different

crystallization conditions [65, 88] or both. Despite this difference, the structures are very

similar with a root-mean-square deviation of 0.6 A over 581 Coc atoms. Furthermore,

there is a slight variation in atomic distances between L-carnitine and the active site

residues in human and mouse CAT. This difference is within experimental error but may

be attributable to subtle structural differences between the two proteins. Overall, the

conformation of L-carnitine and the amino acid residues that form its binding site are

almost identical between these two enzymes.









Conclusions

The structure of the complex of CAT and L-carnitine has been determined and

reveals several key residues that recognize and interact with the substrate L-camitine.

This structure also supports the proposed mechanism that His322 plays an essential role in

catalysis by abstracting the proton from the P-hydroxyl of L-carnitine or the thiol of CoA

depending on the direction of the reaction [64, 65]. Additionally, active site residues

involved in binding L-camitine and catalysis were identified. Specifically, Tyr431, Thr444,

and Arg497 form a binding pocket for the carboxylate group while Phe545 and Val548

interact with the quaternary amine head group of L-carnitine. This study also provides

important information about L-carnitine binding and the catalytic mechanism of other

carnitine acyltransferases since His322, Tyr431, Thr444, and Phe545 are highly conserved

among the carnitine acyltransferase family.














CHAPTER 4
FUNCTIONAL AND BIOLOGICAL CHARACTERIZATION OF ESSENTIAL
RESIDUES OF CARNITINE ACETYLTRANSFERASE

Introduction

The structure of CAT in complex with L-carnitine indicated several residues

involved in catalysis, His322, Tyr431, Thr444, Arg497, Phe545, and Val548. To confirm these

residues are essential enzymatic residues and determine their role in catalysis, they were

functionally and biologically characterized. To functionally characterize these residues,

they were replaced with amino acid residues of varying charge, side-chain length,

aromaticity and polarity. The kinetic properties of the mutant forms of the enzymes were

determined by an enzymatic assay, as described in Chapter 2. To biologically

characterize active site residues a novel genetic system was developed in the yeast,

Saccharomyces cerevisiae. In this system an endogenous CAT is disrupted rendering the

cells unable to grow on media with short-chain carbon groups, such as ethanol, as the

sole energy source. However, human CAT complements these mutant cells and enables

the cells to grow on ethanol media by providing a pathway for short-chain carbon groups

to enter the mitochondria for oxidation.

Degradation of fatty acids in yeast differs from mammals. In yeast, long-chain fatty

acid oxidation occurs exclusively in the peroxisomes, where they are oxidized to

acetyl-CoA [89, 90]. For further metabolism the acetyl-CoA must be transported from the

peroxisome to the mitochondria either by the peroxisomal glyoxylate cycle or the

carnitine shuttle (Figure 4-1) [91]. In the five steps of the glyoxylate cycle, acetyl-CoA









forms succinate which is then transferred to the mitochondria where it enters the citric

acid cycle [91]. In the first step of the glyoxylate cycle, acetyl-CoA is combined with

oxaloacetate to form citrate by the action of the enzyme peroxisomal citrate synthase. The

enzyme aconitrase catalyzes the conversion of citrate to isocitrate, which is then

hydrolyzed by isocitrate lysase 1 to glyoxylate, which in turn forms malate by the

enzyme malate synthase, and succinate, which is transported to the mitochondria. Once in

the mitochondria the succinate will enter the citric acid cycle and release malate, which

will start the process of gluconeogenesis.

Mitochondria
Peroxisome -C02 -
TCA
Fatty acyl-CoA A\Cycle
P-oxidation Camitine
Shuttle t
Acetyl-CoA + L-Carnitine L-Carnitine + Acetyl-CoA
Glyoxylate
Cycle CAT; CATt
Cycle

Oxaloacetate Citrate Acetyl-L Acetyl-L-Carnitine
CIT2 AC arnitine
Malate Isocitrate
TCA
1 MLS ICL2le
Glyoxylate CSuccinate-- Succinate- CyC



Malate

Figure 4-1. Schematic representation of the two pathways for the metabolism of fatty
acids. Long-chain fatty acids are oxidized in the peroxisome by P-oxidation to
acetyl-CoA. The acetyl-CoA can use either the carnitine shuttle or glyoxylate
cycle for further metabolism, as described in the text. CAT, carnitine
acetyltransferase; CIT2, peroxisomal citrate synthetase, ACO, aconitase, ICL,
isocitrate lysase; MLS, malate synthase.









With the carnitine shuttle, the acetyl group of acetyl-CoA is transferred to

L-camitine and the resulting acetyl-L-carnitine is transferred across intracellular

membranes to the mitochondria. Once in the mitochondria the acetyl group is transferred

from L-carnitine to a free CoA where it can then enter the citric acid cycle [91]. CAT

catalyzes the reversible transfer of L-camitine to acetyl-CoA since yeast does not contain

endogenous COT or CPT activity.

Three genes encoding CAT in S. cerevisiae have been identified, CAT2, YAT1 and

YAT2. The first identified CAT gene, CAT2, is translated to CAT, which is localized to

the peroxisomes and inner mitochondrial membrane and is responsible for more than

99% of total CAT activity in galactose grown cells [92]. The second gene, YAT1, codes

for a CAT located in the outer surface of the mitochondria that contributes to 5% of the

total CAT activity in acetate and ethanol grown cells [93]. The most recently identified

gene, YAT2, codes for a cytosolic CAT that is responsible for 50% of total CAT activity

in ethanol grown cells [77].

Disruption of any one of the three genes that code for CAT (CA T2, YA T1 and

YAT2) allows for growth similar to wild type cells on all carbon sources [77]. However,

in yeast cells devoid of the glyoxylate cycle (by disruption of the CIT2 gene and

disruption of any one of the genes encoding CAT) the cells cannot grow on media with

fatty acids, such as oleic acid, or non-fermentable carbon sources, such as ethanol or

acetate, as the sole carbon source. However, these cells grow as normal with glucose as

the carbon source [77]. This indicates that in the absence of the glyoxylate cycle, all three

CATs are essential for growth. We hypothesized that human CAT would complement the









mutant yeast strain and allow growth on media containing ethanol as the sole carbon

source.

To develop the genetic system, we obtained a strain of S. cerevisiae in which

CIT2, the gene that encodes the first enzyme of the glyoxylate cycle, and one of the CAT

genes, CA T2, was disrupted, thus blocking both the glyoxylate cycle and the carnitine

shuttle. These cells could not grow on media with ethanol as the sole carbon source since

they were not capable of utilizing ethanol as an energy source. We hypothesized that

human CAT would complement the mutant yeast strain and that we could use this strain

to evaluate the biological function of CAT and confirm the role of amino acid residues

thought to be involved in substrate recognition/binding and catalysis.

Materials and Methods

The general procedures used in these experiments were described in Chapter 2.

Briefly, recombinant wild type and mutant human peroxisomal CAT was expressed in E.

coli B834 cells with a histidine tag and then purified using a column containing Ni-NTA-

agarose. The kinetic parameters of the purified enzyme were determined using a

spectrophotometric assay that measures the forward direction of the reaction.

The biological activity of the wild type and mutant forms of the enzyme were

determined using an assay in the yeast. Plasmids containing wild type and mutant forms

of CAT were introduced to a mutant strain of yeast. These yeast cells were able to grow

with media containing glucose as the carbon source, but were unable to grow with non-

fermentable carbon sources, such as ethanol. Yeast cells expressing wild type and mutant

forms of CAT were screened for growth on media containing ethanol or glucose as the

sole carbon source.








56



Results and Discussion


Functional Characterization of L-Carnitine Binding Residues


The structure of the binary complex of CAT and L-camitine reveals several


important residues responsible for the recognition and/or binding of L-carnitine. These


residues are conserved among all carnitine acyltransferases with the exception of Phe545


and Val548 that become valine and methionine respectively in carnitine


octanoyltransferases (Figure 4-2). In order to determine their contribution to L-carnitine


binding, four residues, Tyr431, Thr444, Arg497, and Phe545, have been investigated using


site-directed mutagenesis.


humanCAT
mouseCAT
humanCOT
ratCOT
humanCPT1A
ratCPT1A
humanCPT1B
ratCPT1B
humanChAT
ratChAT


TYE SASLRMFHLGRTD
TYE SASLRMFHLGRTD
CYETAMTRHFYHGRTE
CYETAMTRYFYHGRTE
TYEASMTRLFREGRTE
TYEASMTRLFREGRTE
TYEASMTRMFREGRTE
TYEASMTRMFREGRTE
TYE SASIRRFQEGRVD
TYE SASIRRFQEGRVD


GEAFDRHLLGL
GEAFDRHLLGL
GKGFDRHLLGL
GKGF DRHLLGL
GS GIDRHLFCL
GAGI DRHLFCL
GAGIDRHLFCL
GAGIDRHLFCL
GMAIDNHLLAL
GMAIDNHLLAL


DCVMFFGPWP
DCVMFFGPWP
RVQGVVVPMVH
RIQGVVVPMVH
SSGGGFGPVAD
SCGGGFGPVAD
GAGGGFGPVAD
GAGGGFGPVAD
EMFCCYGPWP
EMFCCYGPWP


Figure 4-2. Sequence alignment of carnitine and choline acyltransferases. The amino acid
sequences of human and mouse hpCAT, human and rat carnitine
octanoyltransferase, human and rat camitine palmitoyltransferase 1 liver
isoform, human and rat carnitine palmitoyltransferase 1 muscle isoform and
human and rat choline acetyltransferase were aligned using ClustalW [94].
The residue numbers bordering each segment are indicated. The residues in
bold typeface, Tyr431, Thr444, Arg497 and Phe545, were mutated in this study
using site directed mutagenesis.


Substitution at these positions creates mutant forms of CAT with altered steady-


state kinetic constants. Generally, there is a large increase in Km values for L-carnitine,


with minor changes in Km values for acetyl-CoA compared to L-camitine (Table 4-1). In


almost all cases the catalytic efficiency of L-carnitine with the mutant forms compared to









the wild type enzyme was decreased. A decrease in turnover rates was also observed,

although in two cases, Tyr431Ala and Arg497Gln, the ka, value increased (Table 4-1).

Replacement of Tyr431 with phenylalanine produced the most defective mutant

enzyme, which increased the L-carnitine Km value by 326 fold and decreased the kcat

value by 49 fold compared to wild type CAT (Table 4-1). This mutation also results in

the most dramatic decrease in catalytic efficiency, 2.2 x 101 compared to 3.5 x 105 for

wild type (Table 4-1). The substitution of tyrosine with phenylalanine removes a phenolic

hydroxyl group, which is involved in a strong hydrogen bond with one of the oxygen

atoms of the carboxylic group of L-carnitine (2.53 A). Replacing Tyr431 with alanine also

increased the L-carnitine Km constant by 100 fold, but this increase was three fold lower

than Tyr431Phe (Table 4-1). The decreased enzymatic activity and increased Km and

lowered catalytic efficiency towards L-carnitine for both of the Tyr431 mutants

demonstrates the importance of this residue and the requirement of a hydroxyl group for

high affinity binding.

Table 4-1. Kinetic parameters of recombinant wild type and mutant forms of camitine
acetyltransferase.
Enzyme Vmax L-Carnitine Km L-Carnitine Km ACoA kcat kca,/Km
(tmol/min*mg) (mM) ([LM) (sec-) (M-^sec^)
Wild-Type 33.1 + 1.4 0.11 + 0.02 42.3 2.7 38.5 3.5 x 105
Tyr431Ala 38.0 3.2 11.1 1.4 87.1 + 1.6 44.2 4.0 x 103
Tyr431Phe 0.67 + 0.07 35.9 + 6.4 497.9 + 83.5 0.78 2.2 x 101
Thr444Ala 5.0 + 0.3 7.5 + 0.9 188.3 18.7 5.76 7.7 x 102
Arg497Gln 81.5 + 4.9 25.6 + 3.0 120.2 + 15.7 94.8 3.7 x 103

Phe545Ala 2.9 + 0.1 2.0 + 0.2 177.9 + 15.2 8.57 4.3 x 103
Phe545Tyr 26.2 + 0.8 0.15 + 0.01 63.7 + 6.1 30.5 2.0 x 105









It is interesting that substitution of Phe, rather than Ala, at this position produced a

more defective mutant, since Phe is a more conservative mutation. This may be explained

by a couple of factors. First, charge repulsion most likely occurs between the aromatic

side chain and the negatively charged carboxylate group of L-carnitine, reducing

substrate affinity. Furthermore, this substitution may produce unfavorable interactions

between the Phe residue and adjacent amino acids, which could distort the geometry of

the L-carnitine binding pocket. This may explain the high Km as well as extremely low

specific activity of this mutant. Alternatively, the finding that insertion of Ala at this

position yields a less drastic change in kinetic parameters can perhaps be explained by the

fact that substitution of Tyr by a much smaller Ala residue may allow a water molecule to

fit into the vacant space to form hydrogen bond interactions with L-camitine (Figure 4-3).


Figure 4-3. Three-dimensional representation of the active site with Tyr431. A, Wild type
with His322, L-camitine and Tyr431; B, Mutant form of the enzyme with His322,
L-carnitine and Ala modeled into position 431; C, Mutant form of the enzyme
with His322, L-camitine and Phe modeled into position 431.

Another residue that forms putative interactions with L-carnitine was also

investigated. Thr444 is 2.82 A from the other carboxylic oxygen of L-camitine and can

form a hydrogen bond with the side chain hydroxyl group. Substitution of Thr444 with an


L-Carnitine
B 431
Ala
Hi s 322









alanine increases the L-carnitine Km value by 68 fold and decreases the kcat value by

6 fold (Table 4-1). In addition, Ala444 CAT is not as effective as a catalyst compared to

the wild type enzyme since the catalyst efficiency of Ala444 CAT is more than 450 fold

less than wild type (Table 4-1). The substitution of threonine by alanine removes the side

chain hydroxyl group and presumably the hydrogen bond between Thr444 and L-carnitine,

resulting in a large increase in the L-camitine Km value and a corresponding decrease in

the catalytic efficiency utilizing L-carnitine (Table 4-1). The above mutations and kinetic

results clearly establish the importance of hydrogen bonds among Tyr431, Thr444, and the

carboxylic group of L-camitine.

Next the role of Arg497 was examined. This residue was proposed to influence

L-camitine binding through both electrostatic interactions and its participation in a

hydrogen-bonding network with Tyr431 and Thr444. In addition, Arg497 is conserved

among all carnitine acyltransferases [64], but the analogous position in all known choline

acetyltransferases is an asparagine (Figure 4-2) [95]. Since choline does not have a

carboxylate group as L-carnitine does, it was hypothesized that this arginine residue is

partially responsible for the ability of camitine acyltransferases to distinguish L-camitine

from choline [95]. In fact, substitution of this residue with an asparagine residue in the

bovine form of COT increased the L-carnitine Km value -1650 fold and enabled the

mutant form of the enzyme to use choline as a substrate [95].

Substitution of Arg497 with glutamine in CAT increases the L-carnitine Km value by

232 fold and the kcat value by 2.5 fold (Table 4-1). By replacing the positively charged

Arg497 with an uncharged residue the binding pocket for the carboxylate group of

L-camitine is most likely disrupted, which may account for a decreased apparent affinity









towards L-camitine and a decreased catalytic efficiency of 95 fold compared with the

wild type CAT (Table 4-1). These results are consistent with the previous study

conducted with bovine COT [95] and the structural analysis reported in this work

suggesting a critical role for the electrostatic interactions in binding and recognition of

L-carnitine.

Finally, the interaction of the positively charged quaternary amine head group of

L-camitine in the active site of CAT was examined. Phe545 is hypothesized to be the

residue primarily responsible for the binding of the quaternary amine of L-carnitine. The

structure of the binary complex shows that the aromatic side chain of Phe545 is 3.6 A from

the methyl carbon of the quaternary nitrogen of L-camitine. Substitution of Phe545 with

alanine increased the L-camitine Km value by -18 fold and decreased the turnover rate by

4.5 (Table 4-1). On the other hand, substitution with tyrosine caused only minor changes

in the kinetic parameters compared to wild type. These results demonstrate the

importance of an aromatic ring for L-camitine interactions, since removal of it led to a

more drastic change than replacement with a phenolic side chain. It is interesting to note

that the phenylalanine at this position is conserved in all camitine acyltransferases except

for COT. Mutations may have occurred in COT to compensate for the loss of an aromatic

group at this position. Furthermore, these results are consistent with the hypothesis that

the pi-electrons of the aromatic group of Phe545 interact with the positively charged

nitrogen of L-carnitine through pi-cation interactions. This type of interaction is also

observed with acetylcholinesterase, an enzyme associated with the cholinergic nerve

terminals that hydrolyzes acetylcholine to terminate neurotransmission. The crystal

structure of acetylcholinesterase and photoaffinity labeling revealed that the aromatic ring









of Phe330 as well as Trp84 interacts with the quaternary amine group of acetylcholine in

the active site [96].

Based on the crystal structure of CAT, the residues that were investigated using

site-directed mutagenesis appear to be primarily involved in binding L-camitine. The

kinetic results from this study presented here further support this conclusion. For the most

part, the substitutions at these positions produce only minor effects on the Km values for

acetyl-CoA (1.5-11 fold) compared to the dramatic affects on L-carnitine (1.5-330 fold)

(Table 4-1). The largest change in Km value for acetyl-CoA occurred with the Tyr431 Phe

substitution, which also had the largest Km value for L-carnitine. This mutant also showed

the lowest specific activity and turnover number compared to any other substitution. Thus

it seems likely that this mutation causes a significant perturbation in the active site

leading to decreased activity and ligand affinity. Overall, the substitutions at these

positions in the active site of CAT reveal the key role these residues play in high affinity

binding of L-carnitine and catalysis.

Biological Characterization of L-Carnitine Binding Residues

Functional characterization studies have confirmed four residues to be intimately

involved in L-camitine recognition and binding. Next, the biological importance of these

residues was evaluated with a novel genetic selection system.

The genetic system is based on the observation that there are only two pathways for

peroxisomally produced acetyl-CoA to be further metabolized: the glyoxylate cycle and

the carnitine shuttle. The first step of the glyoxylate cycle is the formation of citrate from

acetyl-CoA and oxaloacetate, catalyzed by the enzyme CIT2 [91]. Disruption of the CIT2

gene blocks the glyoxylate cycle and forces the cell to employ the carnitine shuttle for

further metabolism of acetyl-CoA. The camitine shuttle utilizes carnitine









acetyltransferase activity to transport acetyl-CoA formed in the peroxisome to the

mitochondria where it can enter the citric acid cycle for energy production. CAT also

facilitates in the transfer of cytosolic acetyl-CoA formed by the catablolism of ethanol

and/or acetate to the mitochondria. As previously discussed, disruption of any one of the

three genes that code for CAT blocks the carnitine shuttle [77]. Yeast cells devoid of the

glyoxylate cycle (by disruption of the CIT2 gene) and the carnitine shuttle (by disruption

of the CAT2 gene) cannot grow on media with fatty acids, such as oleic acid, or

non-fermentable carbon sources, such as ethanol or acetate, as the sole carbon source.

However, they grow as normal with glucose as the carbon source [77]. This indicates that

in the absence of the glyoxylate cycle, all three CATs are essential for growth. Therefore,

it was hypothesized that human CAT would complement the mutant yeast strain and

allow growth on media containing ethanol as the sole carbon source.

To test this hypothesis mutant yeast cells with the CIT2 and CAT2 genes disrupted

were screened on ethanol plates to determine their viability. The mutant yeast cells grew

on YND media, but when replica plated onto media with ethanol as the sole carbon

source there were no viable cells, as expected. However, when human CAT was

introduced into the cells the mutant yeast cells were able to grow on both YND and

YNEC media (Figure 4-4). This indicates that human CAT was able to provide a

mechanism for the transport of cytosolic acetyl-CoA to the mictochondria for energy

production and therefore allow growth on non-fermentable carbon sources such as

ethanol. This system suggests that the mutant yeast phenotype can function as a genetic

screen for the verification of residues of CAT important for the biological function of the

enzyme.









To test the feasibility of the genetic system and verify that human CAT does

complement the mutant yeast strain, a mutant form of CAT was introduced to the cells. In

this form of CAT the active site histidine, His322, was substituted with alanine, serine and

glutamine, rendering the enzyme inactive. The resulting yeast cells were able to grow on

YND media; however, no viable cells were observed on the ethanol media, as shown in

Figure 4-5. Since the mutant forms of CAT were inactive, cytosolic acetyl-CoA was

unable to cross the mitochondrial membrane and could not be used as an energy source.

Since ethanol was the only carbon source available the mutant yeast cells could not

survive without metabolizing it for energy. These results show that the mutant form of

CAT could not complement the yeast strain. This provides further evidence that human

CAT complements the mutant yeast strain and verifies that this genetic system can be

used to evaluate the biological activity of essential enzymatic residues of human CAT.

Next, the biological activity of the L-carnitine binding residues was evaluated.

Changing these residues were all shown to decrease the enzyme's effectiveness by

displaying altered catalytic efficiencies. However the distorted kinetic values for each

mutation were different. To further understand how these residues affect L-carnitine

binding and to what degree, they were characterized biologically using the genetic system

in the yeast.

The mutant forms of the enzyme were sub-cloned into the mutant yeast strain and

screened for growth on ethanol and glucose plates. All yeast cells grew on glucose plates

indicating the cells were viable (Table 4-2). However, yeast cells with Thr444Ala,

Phe545Ala or Phe545Tyr CAT grew on ethanol media while cells with the substitution of









Tyr431 by either alanine or phenylalanine did not grow with ethanol. Kinetic analysis of

Tyr431 illustrates how essential hydrogen bonding is to L-camitine binding by producing


N EC' r YND .





lhCA T









Figure 4-5. Human carnitine acetyltransferase complements mutant S. cerevisiae. Shows
a replica plate of yeast media with glucose as carbon source (YND) and yeast
media with ethanol as carbon source (YNEC). Growth of the mutant yeast
strain with endogenous CAT that has been disrupted is shown as the Mutant
section of the YND and YNEC plates. The hCAT section of the YND and
YNEC plates shows growth of the mutant yeast strain with human CAT
introduced. The His32Ala section of the YND and YNEC plates shows
growth of the mutant yeast strain with mutant human CAT, containing the
substitution of the active site histidine with alanine, introduced.

the most defective mutants (Table 4-1). The genetic screen furthers this point by showing

that the enzyme is not active enough to provide a mechanism for the transport of

acetyl-CoA to the mitochondria for energy production, thus inhibiting growth of the cells.

Although Thr444 and Phe545 are also essential to L-camitine binding, mutation at

these positions resulted in less defective mutants (Table 4-1). The Km value towards

L-carnitine for the Thr444Ala, Phe545Ala, and Phe545Tyr forms of CAT were increased 68,

18, and 1.3-fold, respectively while the acetyl-CoA Km values for Thr444Ala, Phe545Ala,

and Phe545Tyr CAT were closer to the wild type form, only increasing 4.4, 4.2, and









1.5-fold, respectively. However, despite these altered kinetic parameters the mutant forms

of CAT were able to complement the mutant yeast and allow growth on ethanol media

(Table 4-2).

Table 4-2. Growth of mutant yeast cells with glucose and ethanol media.

Glucose Ethanol
Tyr431Ala Yes No
Tyr431Phe Yes No
Thr444Ala Yes Yes
Phe545Ala Yes Yes
Phe545Tyr Yes Yes

Characterization of Glu326
Glu326 is located in close proximity to the active site histidine and coordinates with

a water molecule that is involved with binding the carboxylate group of L-camitine

(Figure 4-5). Due to its close proximity to His322, it is thought that this negatively charged

residue helps to stabilize the protonated histidine during catalysis (Figure 4-6).





















Figure 4-5. Three-dimensional representation of the active site of CAT. Shows L-
carnitine, His322 and Glu326. Green and yellow dashes represent hydrogen
bonding potential.






















Figure 4-6. Schematic of the reaction mechanism of CAT. The nitrogen of His322
deprotonates L-carnitine, which facilitates the nucleophilic attack on acyl-
CoA, forming a tetrahedral intermediate. The intermediate falls apart releasing
coenzyme A and acyl-L-camitine. It appears as though the carboxylate group
of Glu326 stabilizes the quaternary nitrogen of histidine during the transition
state.

To evaluate this residue, Glu326 was replaced by amino acids that would alter the

physical properties of this residue: specifically, serine, asparagine, histidine, lysine,

glutamine, alanine, and aspartate. The mutant forms of the enzyme were functionally and

biologically characterized using the enzymatic assay and genetic selection system. As

shown in Table 4-3, the only mutant form of the enzyme enzymatically active was Asp326

CAT. Replacement of glutamate by an aspartate retains the negative charge, but

decreases the side chain length by one carbon atom. The other amino acid substitutions

eliminated the negative charge. As theorized by the three-dimensional structure and

reaction mechanism, these results provide additional support that the negative charge is

necessary during the transition state for stabilizing the protonated histidine.

To determine the biological role of this residue the mutant forms of the enzyme

were introduced to the mutant yeast strain and screened for growth on ethanol and

glucose media. Interestingly, the mutant yeast strain was able to grow with the Asp326

form of CAT, but not with any of the other substitutions (Table 4-3). Apparently this









mutant form of CAT was the only form able to complement the mutant yeast strain. This

means that the Glu326Asp CAT was active enough to provide energy to sustain growth

while the other mutant forms of CAT were not. These results agree with the functional

characterization studies and further support the role of this residue.

Table 4-3. Enzymatic activity and growth with glucose and ethanol of Glu326 mutants.
Growth Activity
Glucose Ethanol
Glu326Gln Yes No None
Glu326Asp Yes Yes Yes
Glu326Asn Yes No None
Glu326His Yes No None
Glu326Lys Yes No None


Conclusions

The structure of the complex of CAT and L-carnitine reveals several key residues

of CAT that recognize and interact with the substrate L-carnitine. To further evaluate the

role of these residues, they were functionally and biologically characterized. Substitution

of four active site residues produced CAT mutants with altered kinetic properties that

support their functional role in binding L-carnitine and catalysis. Specifically, Tyr431

Thr444, and Arg497 form a binding pocket for the carboxylate group while Phe545 interacts

with the quaternary amine head group of L-carnitine. This study also provides important

information about L-carnitine binding and the catalytic mechanism of other camitine

acyltransferases since His322, Tyr431, Thr444, Arg497, and Phe545 are highly conserved

among the carnitine acyltransferase family.














CHAPTER 5
STRUCTURE OF CHOLINE ACETYLTRANSFERASE

The goal of the second specific aim is to further understand the L-carnitine/choline

acyltransferase family by solving the structure of ChAT. Given that ChAT and CAT have

a high sequence homology (- 40%) and the active site residues of CAT are shared with

ChAT, it can be presumed that these two enzymes will share several structural features.

The x-ray crystal structure of ChAT can confirm these estimations and can also be used

to understand several disease causing mutations.

Introduction

As previously discussed, ChAT is the fourth member of the L-carnitine/choline

acyltransferase family and catalyzes the synthesis of the neurotransmitter acetylcholine

(ACh). ACh was the first neurotransmitter to be reported [97] and has since been shown

to play a pivotal role in fundamental brain processes such as learning, memory, and sleep

[98-100]. ACh functions in the cholinergic neurons of the peripheral and central nervous

systems. In the peripheral nervous system ACh stimulates muscle contraction and in the

central nervous system ACh facilitates learning and short-term memory formation.

In 1943 ChAT was identified as the enzyme responsible for the synthesis of ACh

[101]. Similar to other members of the L-carnitine/choline acyltransferase family, ChAT

catalyzes the reversible transfer of an acetyl group from CoA to choline (Figure 5-1).

Although ChAT has been known as the enzyme responsible for the synthesis of

ACh for more than 60 years, its mechanism of action is still unclear [47]. Comparably to

studies with CAT, several inactivation and modification studies of ChAT have suggested









several histidine and arginine residues that are critical for the enzyme to function (Table

5-1) [102-108].

0
HC NOH 0
H3--"N+ H3CN+O^ CH3 + CoASH
H3C H3 CoAS CH3 H3



Figure 5-1. Schematic of ChAT reaction mechanism.

Site-directed mutagenesis studies in Drosophila ChAT have lead to the proposal

that a highly conserved histidine residue, histidine 426, (histidine 334 in the rat form of

ChAT) is essential for activity and may serve as an acid/base catalyst [109]. Additional

mutagenesis studies in rat ChAT have also suggested that another highly conserved

residue, arginine 452, interacts with acetyl-CoA [63]. Interestingly, an equivalent

mutation in human ChAT, arginine 560 to histidine, has also been shown to significantly

reduce the affinity for both substrates, choline and acetyl-CoA [56].

The ChAT gene appears to have multiple splice variants that result in different

mRNA species. Three non-coding exons, R, N and M, have been identified upstream of

the coding exon 4 in rat, mouse and human. However, the difference between these

species is that the human form of exon M contains a sequence that can be translated to

amino acids [47]. At least seven, five and four different mRNA species have been

isolated from the mouse, rat and human, respectively [47]. These mRNA species are

formed from different combinations of the non-coding exons. Since the exons are non-

coding in the mouse and rat, a single form of ChAT protein is generated in these species.

In contrast, the human M exon can be translated to amino acids and thus generates two

forms of ChAT in humans, small and large.









Table 5-1. Reported mutations of choline acetytlransferase.*
Animal species Mutation Corresponding
residue in rChAT


Site-Directed
Mutations
Drosophila


Rat


Disease Cauing
Mutations
Human


His393Leu
His436Leu
His426Gln
Arg452Ala


Arg452Gln


Arg452Glu


Arg453Glu


Arg452Gln/
Arg453Glu


Glu441Lys

Arg560His


His301
His334
His334


Glu331

Arg452
Arg


Temperature
Sensitive
Mutations
Drosophila Met403Lys Leu318
*Adapted from information reviewed in [56, 63, 109, 11


Inactive
Inactive
Inactive
Increase Km for CoA (12-fold)
and ACh (8-fold) and increase
kc, (7-fold)
Increase Km for CoA (12-fold)
and ACh (8-fold) and increase
kc, (7-fold)
Increase Km for CoA (54-fold)
and ACh (60-fold) and
increase kcat (15-fold)
Increase Km for CoA (7-fold)
and ACh (12-fold) and
increase kcat (5-fold)
Increase Km for CoA (170-
fold) and ACh (70-fold) and
increase kat (17-fold)


Low expression, no catalytic
activity
Increase in Km toward
acetyl-CoA and choline



Decrease activity


0].


The small form of ChAT in humans is the primary form observed and is composed

of 630 amino acids with a molecular weight of 67 68 kDa. The observed molecular

weight is less than the calculated molecular weight, most likely due to post-translational

modification or proteolysis during enzyme purification [47]. The large form of ChAT is


Effect


r'









composed of 748 amino acids and contains a 118 amino acid N-terminal extension. It is

not clear whether this form is actually synthesized in vivo [47].

ChAT is localized in the cholinergic nerve terminals (Figure 1-6) in two forms,

soluble (80-90% of total enzyme activity) and membrane bound forms (10-20%) [47,

111]. The significance of the membrane association is unknown; however,

phosphorylation appears to be required for membrane association [112]. ChAT is thought

to exist as an extremely low abundance and monomeric globular protein in nature, with

its mechanism of action still unknown.

Decreases in ChAT activity are an indicator of cholinergic neuron damage and have

been observed with patients who have neurodegenerative disorders, such as Alzheimer's

disease, Huntington's disease and amyotrophic lateral sclerosis. Decreases in ChAT

activity have also been correlated with schizophrenia, Rett syndrome and sudden infant

death syndrome [47]. Recently, direct sequencing of the ChAT gene from five patients,

exhibiting congenital myasthenic syndrome with episodic apnea (CMS-EA), revealed

10 recessive mutations, which result in severe muscle weakness and respiratory

insufficiency. Kinetic analysis of nine of these mutants (the tenth was a frameshift null

mutation), demonstrated that one mutant (Glu441Lys; residue 332 in rChAT) lacked

catalytic activity while the others (Leu21Pro, Pro211Ala, Ile305Thr, Arg420Cys, Arg482Gly,

Ser498Leu, Val506Leu, and Arg560His; residues 101, 102, 196, 311, 373, 388, 397, and 452

in rChAT, respectively) had significantly impaired catalytic efficiencies [56].

Understanding the basis of these mutations, however, has been hindered since no three

dimensional structure was available.









Recently, the crystal structures of human and mouse carnitine acetyltransferase

have been reported [64, 65]. These enzymes show significant amino acid sequence

homology (- 40% identity) with ChAT, and therefore can be used for the molecular

replacement solution of the structure of the rat form of ChAT (rChAT). The structure,

together with the modeled substrate (choline and acetyl-CoA) complexes has given

rational to understanding the mechanisms of catalysis of the enzyme and understanding to

disease causing mutations of ChAT.

Materials and Methods

The experimental procedures involved in this project were described in Chapter 2.

Briefly, recombinant rat ChAT was expressed in bacterial cells with a histidine tag and

then purified using a column with Ni-NTA-agarose. The purified enzyme was then

crystallized and the three-dimensional structure was determined.

Results and Discussion

Structure Determination and Overall Structure of the Enzyme

RChAT was over-expressed in Escherichia coli, purified by affinity

chromatography and crystallized (Figure 5-2). A single crystal was cryoprotected by a

60 second soak in a reservoir solution containing 15% PEG 8000 and 40% glycerol. The

crystal was then mounted in a thin fiber nylon loop (Hampton Research) and flash frozen

prior to data collection at 100 K. The diffraction data were collected at the Cornell High

Energy Synchrotron Source (CHESS) using the ADSC Quantum 4 CCD detector system.

Data were collected using a 0.2 mm collimator, detector-to-crystal distance of 150 mm,

1 oscillation angle with an exposure time of 120 seconds for each image. Reflection

intensities were indexed and evaluated with DENZO and scaled using SCALEPACK

[80].




















Figure 5-2. Photograph of recombinant rat choline acetyltransferase crystals.
Approximate dimensions are 0.4 x 0.2 x 0.2 mm.

A total of 70 of data were collected (70 images) from a single crystal. Figure 5-3

shows a typical 1.00 oscillation diffraction image. The rChAT crystal was shown to be

orthorhombic P212121 with unit cell parameters a = 139.0, b = 77.7, c = 59.7 A. The

structure was refined to 1.55 A resolution [113].

The structure of ChAT was solved by molecular replacement procedures [114],

with the coordinates of human peroxisomal CAT (with the side-chains of all non-glycine

residues truncated to alanines) as a search model [64] using the CNS program package

[81]. A sequence alignment between rChAT and CAT showed there was ~ 40% amino

acid identity indicating the overall structural fold of the two enzymes would be similar.

The structure consists of -40 % helix (20 helices: H1-H20) and 15% strand

(16 strands: S1-S16) arranged into two domains, I (residues 89-400) and II (residues 20-

88 and 401-616; Figure 4.4). These domains are interconnected and create a tunnel,

approximately 16 A in diameter that passes through the center of the enzyme, which

forms the active site.

Domain I is composed of a six-stranded mixed P-sheet (S1, S8, S7, S6, S4, and

S5) that is interconnected on both flanks by eight c-helices (Figure 5-4). Domain II also































Figure 5-3. X-ray diffraction pattern of rat choline acetyltransferase crystal. Data were
collected on the Fl beam line (k = 0.9504 A) with ADSC Quantum 4 CCD
detector system at the Cornell High Energy Synchrotron Source. This data
was collected with a crystal-to-detector distance of 150 mm with an exposure
time of 120 seconds/image and an oscillation range of 1.00 at 100 K. The
black concentric rings show the demarcation of the 5.0, 3.5, 2.0 A resolution
diffraction shells. The highest observed diffraction extended beyond 1.55 A
resolution.

consists of six-mixed P-sheets (S10, S16, S15, S13, Sll, and S12) that are connected

with 11 a-helices (seven from the C-terminal and four from the N-terminal). Helices H1

and H2 generate an anti-parallel bundle that elongates to an extended structure composed

of helices H3 and H4, which crosses over to domain I to form the front entrance (choline)

of the tunnel. In domain I residues 322-335 generate anti-parallel P-sheets (S7 and S8)

with the catalytic His334 located at the end of strand S8 positioned at the center of the

active site tunnel (Figure 5-4). This turn (residues 333-358) extends to form helix H12







































Figure 5-4. Overall Structure of Rat Choline Acetyltransferase. Ribbon diagram
representation depicting ca-helices H1-H20 (red) and P-strands S1-S16 (blue).
Also shown in ball and stick representation the active site His334 and docked
substrates choline and coenzyme A.

that lines one side of the active site tunnel. The long helix H13 is arranged in front of the

P-sheet of domain I and connects to P-strand S10 that links domain I to domain II. Strand

S10 leads into helix H14 that is parallel with the P-sheet of domain II. The P-sheet of

domain II facilitates the flipping out of strand S14 (between S13 and S15), which forms

an anti-parallel interaction with the strand S1 of domain I. Five helices, H15-H20, are

between strands S11 and S16, generate a supporting scaffold that holds domain II P-sheet

open face towards the active site of the enzyme. The P-strand S16 connects with the









C-terminal helix H20 via a loop that terminates domain II. The two domains of rChAT

adopt a similar topology with each other in the absence of any recognizable amino acid

sequence homology (- 11% identity) between them.

The overall surface area of the molecule is 25,000 A2. Out of the 644 amino acids,

89 (mostly hydrophobic residues) are buried at the interface between the two domains

(-4000 A2). In domain I residues Ala109, Ile111, and Glu347 have hydrogen bond

interactions with domain II amino acids Thr555, Met556, Phe559, and Cys561. The carbonyl

oxygens of Glu347 hydrogen bond with the side chain of Thr555 (3.1, 2.8 A) and the

main-chain nitrogen of Met556 (2.8 A). Both residues, Ala109 and Ile11, have hydrogen

bond interactions with Cys561 (3.0 A) and Phe559 (2.8 A) main chain atoms. These

interactions are predominantly main-chain and therefore there is no evolutionary pressure

to conserve them between species (Figure 5-5).

Comparison of CAT-L-Carnitine with Free CAT

The amino acid sequence of rChAT is highly homologous with ChATs from other

species and has 95, 86, and 85% identity with mouse, human, and pig respectively. In

addition, the amino acid sequence of rChAT has 40% identity compared to human and

mouse CAT [64, 65] and shares a common structure with a RMS deviation of 1.30 and

1.25 A, respectively, for approximately 570 Ca atoms. The rChAT structure has 2

additional P-strands (S9 and S10) and an additional helix (H14) compared to the

structures of human and mouse forms of CAT. Most of the main-chain structural

variation (3.0 to 6.0 A) occurs between residues 209-216 and 233-240 of rChAT

compared to human and mouse CAT.


















Rat
Bat

M~busa
Pig
Hu maplpil.
C-Men rbditi.










Rat

buet


Hr ,aphilm.
Cenarhbdit i



iat

Bat








Pig
Hro maphila
Caenarmbditit



Rat

Bat
Eman











~bume
Pig








fIl mphila



Caencrhabditia
Rat

Bat
bf--
Pig






E mma phila
Cen aItbitia.
Rat

Pat












hman
Pig






flr naphila






Cenc r bditi.
Rat

Pat


Pig
Dr- aphila
Caenrxbd3i t



Rat

Pat


Pig
Dr a.phila




Rat




Pig
Era cphila
Ge^ncrTTT-ia4


Figure 5-5. Sequence alignment of choline acetyltransferases. Shows the amino acid

sequence alignment of rat, human, mouse, pig, D. melanogaster, and C.

elegans ChAT. The amino acid numbering and secondary structural elements


4l E T J
RD lOD 3lD 140 50 40 D7
BP R PVK ls ..LT V D S r ;s DT1. 0 BA r S P G
TF T B TPNl L ...I. ES I F QgT 0 C EfRSrG P













T"F T B "Nfl T DTL [ o A r B5E IE.SG S I 3C GQ
nP ; PVQAS C E r..L. rM rPE LT V A I iESPE BE
MPI L E AAKPSE s .-F.PGITgB. E ER
DE3 z N E TAEBV s P F DT Tl



H4 51 to






o0 90 ll 11 1f0 D130 10 i0



Il IG: S1 T I Il






I I TI I











M6 __ E 5 T Sl_ T TT
... .....
IiaT sTD .n ,. 11 *. ...


-S GTD VL" T.a r T V VV
C VR
.2B T







IT33 D 33 IEP v.........3 K... .. ........ ..... C





Q 31 0 320 33 0 40 40 430 f20
1ER D D Y.. l ... .... .. .......


nit SS HS lrTP









AA AAAAAAAA -----r C VA PP
S GDEf Ir IfDl PPD~ c7LQ QP3 VAZR RnQ 5 3C
A HMGMCB 49"VM. GR V F? VIIVQr jS IKH IMT iITKKLVRV DSV!3

... BlE TcGr SPEGI Jl^ SCk c E lVV4ILr~4N rI rfl TP P D1TilL P B



SY 14 kH 1117





3 T [1 BD 3SQ PL D 4
T C T i lBE E PRI [ v0 iBe ilMrw A eB KT H IafB Sl



-_A ,,,,,,,,.nAflA SSS2ZQUA A ------ ---- --AAAAASU9 mAAAA9agaA
4ED 460 4D50 4BD 40








13 HN EH D INLA E I A [ FA P. . P.. ..-M
HAI






HIS t. 1 514 S13
AAKK AAASA AAif ------ --- "OL
E0 B0 E3D0 E40 E 3E 0 5EDB I7
';SB ..n .x [i S Bj~ LK rTIE B H .BCCrIi



laI.E fE yELTI-:L C5T Q .E BG IE|TT C



His 514M230




ST PAKAV S SQ GM& SEP?;TS lTHPS3VSP .

s~ r ~a iE V. S .. .









are based on the structure of rat ChAT. Strands S1-S16 are represented by
blue arrows while helices H1-H20 are depicted by red spirals. Identical and
conserved residues are indicated by red and yellow blocks, respectively.
Residues mutated using site-directed mutagenesis are indicated by purple
triangles [63, 95, 109]. Residues mutated in patients with congenital
myasthenic syndrome with episodic apnea (CMS) are shown by green
triangles [56]. A blue star indicates the residue, Arg312, that was mutated by
site-directed mutagenesis and shown to be mutated in patients with CMS.

Interestingly, E2p, an enzyme that catalyzes a similar reaction to that of rChAT,

in that it transfers an acetyl group from acetyl-CoA to an organic substrate, shares an

8% sequence identity with rChAT. When E2p is structurally aligned with domain I of

rChAT it has an RMS deviation of 1.87 A for 117 Ca atoms [115].

Active Site

The catalytic center of rChAT is a 16 A long tunnel, formed between the interface

of domains I and II. Chemical modification, inactivation, and mutagenesis studies have

identified the catalytic histidine of rChAT as His334, which is located at the center of this

tunnel between 3-strand S8 and helix H8 in domain I (Figure 5-4) [18, 116]. This

observation is consistent with sequence homology comparison with the known catalytic

histidine residues of other acyl group transfer enzymes including His322 in human CAT

[64], His343 in mouse CAT [65], His610 in E2p [115], and His195 in chloramphenicol

transferase [117].

In rChAT, His334 assumes an unusual conformational arrangement (X1 = -146;

X2 = -49) to form a hydrogen bond (3.0 A) interaction between the imidazole nitrogen

N61 and its main chain carbonyl oxygen. In the crystal structure of human and mouse

CAT, E2p, and chloramphenicol transferase, the catalytic histidine also shows a similar

conformation [64, 65, 115, 117]. Interestingly, this conformation of His334 allows the N3









nitrogen to point towards the hydroxyl group of choline to extract a proton, as shown in

Figure 5-6 with His322 of human CAT and L-carnitine.

















Figure 5-6. Three-dimensional representation of the conformation arrangement of the
active site histidine, His322 in human CAT. Dashed line (yellow) represents
hydrogen bond interaction between imidazole nitrogen and backbone carbonyl
group (green) represents hydrogen bond interaction between N3 of His322 and
hydroxyl group of L-carnitine. These interactions are common with mouse
CAT, E2p, chloramphenicol transferase and ChAT.

A highly conserved glutamate, Glu338, interacts with His334 and may be important

for catalytic activity. The carbonyl oxygen 01 of Glu338 is hydrogen bonded (3.7 A) with

the N3 of the active site His334 and the 02 of Glu338 hydrogen bonds (2.8 A) with Arg458

Most likely these interactions contribute to potentiate the catalytic activity of His334 and

also stabilize the binding of choline.

Choline Binding Site

The docking of choline into the active site of rChAT indicates five residues (Tyr95,

His334, Ser548, Tyr562, and Val565) that may be involved in its binding (Figure 5-7, Table

5-2). The hydroxyl group of choline interacts with His334 and forms a hydrogen bond (3.0

A). This hydrogen bond formation supports the catalytic role of His334. Tyr95 is also









within hydrogen bonding distance to the hydroxyl group of choline (2.0 A). The

quaternary amine head group of choline is positioned close to the side chain of Tyr562,

separated by 4.6 A, and has hydrophobic interactions with the side chain residues Ser548

and Val565 (3.0 A and 3.2 A, respectively) (Figure 5-7, Table 5-2). Interestingly, in other

highly conserved members of the L-camitine/choline acyltransferase family, CAT, the

quaternary amine of L-camitine appears to be positioned close to a phenylalanine residue,

Phe545 and Phe566 respectively. However, in the case of rChAT the quaternary amine of

choline interacts with a tyrosine residue (Tyr562) [64, 65]. In CAT, the highly conserved

residues within carnitine acyltransferases, Thr444 and Arg497 (Val459 and Asn514 in rChAT)

play critical roles by interacting with the carboxylate group of L-carnitine [65, 85]. Since

choline does not have a carboxylate group at this position, these residues are not

conserved in rChAT.

All members of the L-camitine/choline acyltransferase family contain a conserved

Ser-Thr-Ser (STS) motif near the carboxy terminal: Ser548, Thr549, and Ser550 in rChAT

[84]. The first serine of this motif, Ser548, is positioned close to the quaternary amine

group of choline and may play a critical role in transition state stabilization during

catalysis.

Coenzyme A Binding Site

The CoA binding site was modeled on the opposite side of His334 from the choline

binding site. The strands S11 and S12 of domain II are separated from each other and

create a space that can accommodate the pantothenic arm of CoA, which extends towards

the active site. Highly conserved residues in L-carnitine/choline acyltransferases surround

the CoA binding site at the C-terminal end of the enzyme (Table 5-3).
































CoA


Figure 5-7. Close up view of the modeled binding site of choline and CoA. The Ca trace
of domain I (blue) and domain II (red). RChAT residues putatively involved
in binding shown in ball and stick. Red dash lines indicate possible hydrogen
bond interactions

Residues Lys413 and Lys417 may interact with the 3'-phosphate group of CoA and a

clustered group of residues (Asp424, Glu447, and Gln551) possibly interact with the

pantothenic arm of CoA (Figure 5-7, Table 5-3). The carboxyl oxygens of Asp424 are

hydrogen bonded with Gln551 (3.0 A) and Glu447 (2.6 A). This would imply at least some

effect on the protonation state of at least one of the acidic residues in the active site.









Table 5-2. Choline substrate binding sites for rat choline acetyltransferase (rChAT),
human peroxisomal L-camitine acetyltransferase (hCAT), and mouse L-
carnitine acetyltransferase (mCAT).
Atom rChAT hCAT** mCAT***
03 His334 (3.02) His322 (2.76) His434 (2.68)
N(CH3)3 Ser548 (3.04) Ser533 (3.61) Ser552 (3.62)
Tyr562 (4.6) Phe545 (3.6) Phe545 (3.96)
Val565 (3.16) Val548 (3.83) Val569 (3.86)
01 Tyr431 (2.55) Tyr452 (2.59)
02 Thr444 (2.82) Thr465 (2.63)
*Residue (Distance (A))
** Interactions taken from the crystal structure of hCAT,
PDB number, 1S50 [85].
***Interactions taken from the crystal structure of mCAT,
PDB number, 1NDF [65].

Table 5-3. Coenzyme A substrate binding sites for rat choline acetyltransferase (rChAT),
human peroxisomal L-camitine acetyltransferase (hCAT), and mouse L-
carnitine acetyltransferase (mCAT).
Atom rChAT hCAT** mCAT***
P3(09) Lys413 (3.23) Lys398 (3.27) Lys419 (3.64)
Lys417 (2.72) Lys402 (2.67) Lys423 (3.58)
*Residue (Distance (A))
** Interactions taken from the crystal structure of hCAT,
PDB number, 1S50 [85].
***Interactions taken from the crystal structure of mCAT,
PDB number, 1NDF [65].


Previous chemical modification, inactivation, and site-directed mutagenesis

studies have suggested that Arg452 of rChAT interacts with the 3'-phosphate group of

CoA to stabilize CoA binding [63, 105]. Although substitution of this residue by an

alanine resulted in a 7-12 fold increase in Km for both CoA and acetyl-CoA as well as an

increase in the kc, value [63], based on observations from the structure of rChAT it

appears that this residue is too far from the 3'-phosphate group of CoA (- 11 A) to form

an interaction. Perhaps the mutation of Arg452 with alanine disrupts the overall structure

of the active site or the stability of the enzyme to cause such a decreased affinity towards

the substrates.









Disease Associated Mutations

Congenital myasthenic syndrome (CMS) is an inherited disease that affects the

neuromuscular junction and causes muscle weakness from birth or early childhood. The

neuromuscular junction is the junction between a terminal motor neuron and a muscle

fiber. Nerve impulses travel down the motor neuron to cause the muscle fiber to contract.

Once the action potential reaches the neuron terminal ACh, synthesized by ChAT and

stored in synaptic vesicles, is released into the synaptic cleft where it binds to its

receptors to elicit neurotransmission. The action of ACh is terminated by

acetylcholinesterase, which is the target of several Alzheimer's disease therapeutics and

chemical warfare.

CMS is classified into three categories based on the site of the mutation. Pre-

synaptic defects, which account for 8% of cases, are related to decreased production or

release of ACh. Synaptic basal lamina associated defects are responsible for 16% of cases

and are due to endplate acetylcholineterase deficiency that leads to excessive muscle

stimulation and ultimately damages the muscle. The third category accounts for the

largest proportion of cases and is a result of deficiency or mechanical changes of the ACh

receptor.

Interestingly, the cause of the defect determines the symptoms. With the pre-

synaptic form of the disorder patients experience dyspnea (shortness of breath) and

bulbar (muscles of the face, jaw, pharynx, larynx and tongue) weakness. The

distinguishing feature of this form of CMS is that the dyspnea and bulbar weakness is

sudden and can lead to apnea or temporary cessation of breathing. This symptom is

precipitated by infection, fever or excitement. The only treatment of this form of CMS is









acetylcholinesterase inhibitors. However, recent studies on the molecular basis of the

disease have revealed mutations of the ChAT encoding gene.

An analysis of the ChAT encoding gene from five patients with CMS-EA revealed

recessive mutations, one (523insCC) null mutation, three mutations (Ile305Thr, Arg420Cys,

and Glu441Lys) that reduce ChAT expression, one mutant (Glu441Lys) devoid of catalytic

activity and eight mutants (Leu210Pro, Pro211Ala, Ile305Thr, Arg420Cys, Ser49Leu,

Val506Leu, and Arg560His) that have decreased catalytic efficiencies towards CoA

(Table 5-1, Figures 5-5 & 5-7) [56]. Further mutational analysis of the ChAT encoding

gene from three patients with CMS-EA from two Turkish families shows an additional

mutation (Ile336Thr) and studies from five patients from three independent families shows

three further mutations (Vall94Leu, Arg548Stop, and Ser694Cys) [118, 119]. Figure 5-8

shows that most of these residues are located in the periphery of the structure, away from

the active site tunnel and perhaps are important for stability or maintaining the

conformation of the active site.

The most severe mutation reported, Glu441Lys, drastically reduces ChAT

expression in COS cells (29% compared to wild type) and the mutant form of the enzyme

is devoid of catalytic activity [56]. This conserved residue corresponds to Glu333 in

rChAT and is located in strand S8, adjacent to the catalytic histidine (Figure 5-5).

Although this residue points in the opposite direction of the active site and appears to not

be involved directly in the catalytic mechanism of the enzyme, mutation to a lysine could

introduce interactions with a nearby arginine, Arg250, (4.0 A). These interactions could

disrupt the active site or the orientation of the active site histidine and thus eliminate

catalytic activity.



















CoA


Figure 5-8. Structural comparison of rat choline acetyltransferase (rChAT) and human
peroxisomal carnitine acetyltransferase (hCAT) and mutations in human
ChAT that cause myasthenic syndrome associated with episodic apnea (CMS-
EA). Coil diagram shows the superimposition of rChAT (red) onto hCAT
(blue). Arrows show structural variation (Ca RMS deviation > 2.0 A), residues
209-216 and 233-241, between rChAT and hCAT. Shown in ball and stick
His334, choline/L-carnitine, and coenzyme A. Residues mutated in patients
with CMS-EA are shown by green spheres [56]. Residues mutated using site
directed mutagenesis are indicated by purple spheres [63, 95, 109]. A blue
sphere indicates the residue, Arg312, that was mutated by site directed
mutagenesis and shown to be mutated in patients with CMS-EA. (Refer to
text, Table 5-1, and Figure 5-5).

Another interesting mutation is Arg560His. This mutation causes a decreased

affinity towards both acetyl-CoA and choline [56]. Arg560 is absolutely conserved and

corresponds to Arg452 in rChAT (Figure 5-5). As previously discussed, chemical









modification, inactivation and site-directed mutagenesis studies have shown that this

residue interacts with the 3'-phosphate group of CoA [63, 105]. However, based on the

structure, Arg452 does not interact with either substrate. Therefore, it is quite perplexing

that a mutation of this residue can impair catalytic activity to such a degree that it causes

a disease state. Based on the structure, it appears that a histidine located at position 452,

could interact with Gln164 (-2.5 A), which could disrupt the conformation of the active

site histidine. A histidine residue at position 452 could also interact with Glu338 (-4.0 A),

which, as previously discussed, forms a hydrogen bond interaction with the active site

histidine.

The other mutations, Leul02Pro, Prol03Ala, Ile197Thr, Arg312Cys, Arg374Gly,

Ser390Leu, and Val398Leu, have been shown to cause a decrease in the affinity for acetyl-

CoA and therefore decrease catalytic efficiency when compared to wild type [56].

Temperature Sensitive Mutations

The structure of rChAT can also be used to understand the temperature sensitive

mutations in Drosophila that were generated to determine the role of the neurotransmitter

acetylcholine in development and mediating behavioral function in the brain [120]. These

mutations (Chats1 and Chats2) show a conditional temperature sensitive phenotype in

which at increased temperature, 300C, ChAT activity decreases. Drosophila expressing

either mutation become paralyzed and subsequently die when incubated at 300C [120]. At

the higher temperature it was shown that the mutant phenotype appears more rapidly with

the Chats2 mutant, indicating that the Chats2 mutation is more severe than the Chats' [120].

Recently these mutations have been identified as Met403Lys for Cha"s, which

corresponds to Leu318 in rChAT, and Arg397His for Chas2, which corresponds to Arg312 in

rChAT (Figure 5-5) [121]. Arg312 is absolutely conserved within all known ChATs while









Leu318 is not conserved. In fact, this residue is a leucine residue in rat, human, mouse and

pig, an isoleucine in C. elegans, and a methionine in the Drosophila ChAT.

Both of these residues are close to the active site histidine. The replacement of

Leu318 by lysine appears to affect the overall stability of the enzyme since it appears that

this mutation does not introduce deleterious interactions with surrounding residues. In

contrast, substitution of Arg312 by histidine may change the choline binding site. It

appears that a histidine residue located at position 312 may interact with Glu333, which is

located adjacent to the active site histidine. As previously discussed, this residue faces the

opposite direction of the active site. However, it may be critical in formation of the

choline binding pocket, such that an unfavorable interaction will disrupt choline binding.

Conclusion

The structure of rChAT has been determined to a 1.55 A resolution and exhibits

conserved structural features found in human and mouse CAT [64, 65]. RChAT is a

monomeric protein containing two domains that interconnect to form an active site

tunnel. Within the active site the catalytic residue His334 is located at the center of the

tunnel, suggesting a common catalytic mechanism for the entire L-carnitine/choline

acyltransferase family. The substrates, choline and acetyl-CoA, have been modeled in the

active site based on the structure of mouse CAT, and reveal several residues that may be

involved in substrate recognition and ligand interactions and provide insight into the

molecular basis of disease causing mutations.














CHAPTER 6
CONCLUSIONS

The camitine/choline acyltransferase family plays an important role in fat

metabolism by facilitating the transfer of acyl groups across intracellular membranes.

These enzymes catalyze the reversible transfer of acyl groups between CoA and

L-camitine. There are three known classes of camitine acyltransferases that differ in their

acyl group specificity, subcellular location, tissue distribution and physiological function.

Of these carnitine acetyltransferase (CAT) is of particular interest. CAT has a substrate

preference for short chain (2-6 carbon atoms) acyl groups and is found in the

mitochondrial matrix, the endoplasmic reticulum and peroxisomes. Chemical

modification and site-directed mutagenesis experiments have identified an active site

histidine that acts as a general base to catalyze the reaction. However, in spite of its

importance, fundamental questions remain about the structure of CAT.

The structure of CAT with its substrate L-carnitine was determined and reveals

several residues involved in catalysis. The structure supports the putative mechanism of

the enzyme and identified the active site histidine as His322. His322 plays an essential role

in catalysis by abstracting a proton from the P-hydroxyl group of L-camitine or the thiol

group of CoA depending on the direction of the reaction.

In addition, specific L-camitine binding residues were identified. It was shown that

three residues, Tyr431, Thr444 and Arg497, interact with the carboxylate group of

L-camitine while two residues, Phe545 and Val548, interact with the quaternary amine









group. These residues are highly conserved within the carnitine/choline acyltransferase

family.

To confirm their role in catalysis these residues were functionally and biologically

characterized. Functional characterization was carried out by a spectrophotometric

enzymatic assay. A novel genetic assay with Saccharomyces cerevisiae as the model

organism was developed to biologically characterize these residues. In this system an

endogenous yeast CAT gene is disrupted rendering the cells incapable of using short

chain carbon groups as an energy source and thus unable to grow on ethanol media.

Human CAT complements the mutant yeast strain and allows growth on short chain

carbon sources while an inactive form of CAT is unable to complement the yeast strain.

Substitution of the four residues (Tyr431, Thr444, Arg497 and Phe545) produced mutant

forms of the enzyme with altered kinetic parameters. The mutant forms of the enzyme

were also unable to complement the mutant yeast strain. Taken together these studies

support the role of these residues in binding L-camitine and catalysis.

In addition to these studies, the role of Glu326 was evaluated using the enzymatic

and biological assays. This residue is located in close proximity to the active site histidine

and is thought to stabilize the protonated histidine during catalysis. Replacement of

Glu326 by other amino acids, except by aspartate, produced mutant forms of the enzyme

that were catalytically and biologically inactive. This provides additional support that the

negative charge is necessary during the transition state for stabilization.

The structure of the fourth member of the camitine/choline acyltransferase family,

choline acetyltransferase (ChAT), was also determined. It was found that rat ChAT and

human CAT share several structural features; both enzymes are composed of two