ANALOGS OF NICOTINAMIDE ADENINE DINUCLEOTIDE AS PROBES OF THE STEREOCHEMICAL ‘ RELATIONSHIPS IN THE TERNARY COMPLEX OF ALCOHOL DEHYDROGENASES Dissertation for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY DANIEL JOHN KROON '_ - 1976 I I ‘34:: I ff I :33 g; 51-: fi'rg.‘g\ kio‘ . . IKE 5"; do 9 CLJ§;Y f“ This is to certify that the thesis entitled Analogs of Nicotinamide Adenine Dinucleotide.As Probes of the Stereochemical Relationships in the Ternary Complex of Alcohol Dehydrogenases Date 0-7 639 presented by Daniel John Kroon has been accepted towards fulfillment of the requirements for Chemistry jI/‘kvx J. Karabatsos Major professor Ph.D. degree in June 11, 1976 1’ Jun Jun; "0 \u J‘Y‘a 8 Am ABSTRACT ANALOGS OF NICOTINAMIDE ADENINE DINUCLEOTIDE AS PROBES OF THE STEREOCHEMICAL RELATIONSHIPS IN THE TERNARY COMPLEX OF ALCOHOL DEHYDROGENASES BY Daniel John Kroon Nicotinamide adenine dinucleotide is the coenzyme for a large number of enzyme catalyzed oxidation-reduction reactions. Many of these reactions are stereospecific and much work has gone into the elucidation of the factors controlling this stereospecificity. One such reaction is the interconversion of alcohols and carbonyl compounds catalyzed by the alcohol dehydrogenases (E.C.l.1.1.1.). Prelog, Karabatsos, and others have proposed various models for the spatial arrangements of substrate and coenzyme in the enzyme active site to account for the observed stereochemistry of the reaction products. In recent studies using purified enzymes and coenzymes, Karabatsos, Nunez, and Stamoudis established that hydrophobic-hydrophilic interactions are important in determining the stereochemical outcome of the reactions catalyzed by alcohol dehydrogenases. To design a general predictive model, however, more information is required as Daniel John Kroon to the location of these points of interaction in the enzyme active site and their stereochemical relationships to the coenzyme and substrate. The available evidence for a particular orientation of the substrate with respect to the coenzyme in the enzyme ternary complex, e.g. the spectral studies of Kosower, is rather meager. With the aim of furnishing more data relevant to this issue, the analogs of nicotinamide adenine dinucleo- tide (I, IS and IR) were synthesized and their activity with yeast and horse liver alcohol dehydrogenase was investigated. This analog study is a step toward elimi- nating the problem from which all of the previous product studies have suffered, namely, the inability to describe accurately the overall stereochemistry because of the presence of too many variables in the arrangement of enzyme, coenzyme, and substrate. The analogs of this study have the substrate covalently attached to the coenzyme, thus restricting the number of ways that the substrate can align itself with the coenzyme. C) (DH \ .. 3'- NH (CH2)5(': CH3 H N/ I I‘aCCmiC é‘I‘ IS S-configuration IR R-configuration Z :. ribosc-diphosphote.ribosc-odcninc Daniel John Kroon The investigation of the stereochemistry of the reaction of the analog with the enzyme required the syn- thesis and resolution of 7-amino-2-heptanol. The amino- alcohol was resolved by crystallization of its tartrate salt. Deamination to 2-heptanol and straight-forward cor- relations allowed the conclusion that (-)-7-amino-2- heptanol has the R configuration. The nicotinamide adenine dinucleotide analogs func- tioned as coenzymes in the oxidation of ethanol with liver alcohol dehydrogenase, but the oxidation proceeded at a much slower rate than with the natural coenzyme. The analogs showed no activity with ethanol and yeast alcohol dehydrogenase. The coenzyme-substrate analogs underwent oxidation-reduction in the presence of liver alcohol dehydrogenase as determined by UV spectroscopy. The analog prepared from the 7-amino-2-heptanol with the S configuration reacted more than the one prepared from (R)- 7-amino-2-heptanol. It was determined that this reaction could be intermolecular, leaving the question of the spatial orientation of the substrate and coenzyme in the alcohol dehydrogenase ternary complex unanswered. ANALOGS OF NICOTINAMIDE ADENINE DINUCLEOTIDE AS PROBES OF THE STEREOCHEMICAL RELATIONSHIPS IN THE TERNARY COMPLEX OF ALCOHOL DEHYDROGENASES BY Daniel John Kroon A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1976 ACKNOWLEDGMENTS The author wishes to express his appreciation to Professor Gerasimos J. Karabatsos for his guidance in this investigation. Appreciation is also extended to Dr. Mark M. Green and Michael K. May for their helpful discussions and encouragement which aided in the completion of this research. The financial support provided by the Department of Chemistry, Michigan State University, is gratefully acknowledged. ii LIST OF LIST OF TABLE OF CONTENTS TABLES O O O O O O O O I O O O FIGURES O O O O O O O O O O O 0 INTRODUCTION 0 O O O O O O O O O O O 0 EXPERIMENTAL O O O O O O O O O O O O 0 RESULTS Instrumentation . . ’ . . . . . Synthesis of 7-amino-2- -heptanol . . . Synthesis of ethyl (3-cyanopropyl) acetoacetate . . . . . . Synthesis of 6-cyano-2-hexanone . Synthesis of 6-cyano-2-hexanol . . Synthesis of 7-amino-2-heptanol . Synthesis of 6-amino-2-hexanol . . . Synthesis of S-cyano-Z-pentanone . Synthesis of 5-cyano-2-pentanol Synthesis of 6-amino-2-hexanol . Resolution of 7-amino-2-heptanol . Resolution with d-tartaric acid Resolution with l-tartaric acid . . Determination of’the Absolute Configur- ation of (-)-7-amino-2-heptanol . NMR Method for the Determination of Enantio- meric Purities . . . . . . . .. Resolution of 6-amino-2- hexanol . Synthesis of the N-(hydroxyalkyl)-nicotin- amides . . . . . . . . . Preparation of NAD+ Analogs . . . . . Activity Determinations in Enzyme Systems AND DISCUSSION . . . . . . . . . Resolution of 7-amino-2-heptanol . . . Enantiomeric Purity of the Resolved 7-amino-2-heptanols . . . . . Optical Rotations of the Resolved 7-amino-2-heptanols . . . . . iii Page vi Page Absolute Configurations of (—)- and (+)- -7- amino-2- -heptanol . . . . . 47 Resolution of 6- amino-Z- hexanol . . . . 48 Synthesis of Optically Active N-(6- -hydroxy- l- -heptyl)- -nicotinamides . . . . . . 49 Preparation of the NAD+ Analogs . . . . . 50 Activity of the NAD+ Analogs in Enzyme Systems . . . . . . . . . . . 53 Substrate Specificities and Stereospecifici- ties of Liver Alcohol Dehydrogenase . . 63 Conclusion . . . . . . . . . . . . 70 LIST OF REFERENCES . . . . . . . . . . . . 73 iv LI ST OF TABLES Table Page 1. Crystallizations of the d-tartrate of 7-amino- 2-heptan01 o o o o o o o o o o o 24 2. Optical rotations of (-)-7-amino-2-heptanol . 25 3. Crystallizations of the letartrate of 7-amino- 2-heptan01 o o o o o o o o o o o 27 4. Optical rotations of (+)-7-amino-2-heptanol . 28 5. Crystallizations of the dibenzoyl-d-tartrate Of 6-aminO—2-hexan01 o o o o o o o o 32 6. Optical rotations of (-)-N-(6-hydroxy-l- heptyl)-nicotinamide . . . . . . . . 37 7. Equilibrium constants for the oxidation- reduction of N-(hydroxyalkyl)-nicotin- amides . . . . . . . . . . . . 66 8. Equilibrium constants for the oxidation- reduction of aminoalcohols . . . . . . 68 9. Oxidation—reduction equilibrium in optically active systems . . . . . . . . . . 70 LIST OF FIGURES Figure 1. Structure of oxidized nicotinamide adenine dinucleotide . . . . . . . . . 2. Stereochemistry of hydrogen transfer from NADH. 3. Karabatsos model . . . . . . . . . 4. Prelog model . . . . . . . . . . . 5. Reduction of hydroxyacetone by NADH with glycerol dehydrogenase . . . . . . 6. Reduction of methyl ethyl ketone by NADH With LADH I O O O O O O O O O 7. Reduction of l-chloro-Z-propanone by NADH With LADH I O O I O O O O O O 8. Arrangement suggested by Graves . . . . 9. Model of Cervinka and Hub . . . . . . 10. Model for the mechanism of lactate dehydrog- enase O O O O O O O I O I O O 11. Stereochemistry of the reaction of NAD+ analogs . . . . . . . . . . . 12. NAD+ analogs (I) . . . . . . . . . 13. NMR spectrum of (S)—N-(6-hydroxy-1-hepty1)- nicotinamide . . . . . . . . . l4. ORD curve of (-)-7-amino-2-heptanol . . . 15. ORD curve of (-)-N-(6-hydroxy-l-hepty1)- nicotinamide . . . . . . . . . 16. Preparation of NAD+ analogs . . . . . . l7. Column separation of NAD+ analog . . . . vi Page 03:5 10 12 12 12 14 15 36 46 46 51 52 Figure Page 18. NAD+ analogs (II) . . . . . . . . . . 61 19. N-(hydroxyalkyl)-nicotinamides . . . . . . 65 20. Possible intramolecular hydrogen bonding . . 67 vii INTRODUCTION One truly amazing property of enzymes as chemical reactants is their extreme selectivity in reaction partners. Living matter is very dependent on this specificity, there being many enzymes that have been developed to sustain par- ticular reaction sequences that are vital to the life pro- cess. Some of these enzymes operate on a single substrate and many are known that react with only one stereoisomer of a substrate. For example, in the biosynthesis of squalene from six molecules of mevalonate, Popjak and Cornforth (1) determined that there were 16,384 (214) pos— sible stereochemical paths that the sequence of reactions could follow, yet the enzyme-catalyzed process follows one of these paths exclusively! Indeed, in living systems stereospecificity appears to be the rule rather than the exception. During his years of work on sugar stereochemistry and fermentation reactions, Emil Fischer recognized the importance of enzyme-substrate interactions in the regula- tion of the specificity of enzymes. Ogston (2) extended this concept to the ability of enzymes to distinguish between two chemically identical groups bonded to the same carbon atom, at centers now termed "prochiral." The prev— alence of stereospecificity in enzyme-catalyzed reactions suggests that precise spatial orientations of substrates to the enzyme and coenzyme are an integral part of the catalysis. It is with the hope of obtaining information about these stereochemical relationships, and therefore about the catalytic mechanism, that many studies of enzyme stereospecificity are conducted. Stereospecificity is commonly observed in the bio- logical interconversion of alcohols and carbonyl compounds catalyzed by alcohol dehydrogenases (E.C.l.1.1.1.). The alcohol dehydrogenases isolated from yeast and horse liver (YADH and LADH) have been fully characterized and their reactions with a large number of compounds have been studied (3). This class of enzymes exhibits stereospeci- ficity towards both substrate and the coenzyme necessary for the reaction, nicotinamide adenine dinucleotide (NAD). The structure of the oxidized form of nicotinamide adenine dinucleotide (NAD+), confirmed by synthesis (4), is shown in Figure 1. In the oxidation of alcohols catalyzed by YADH and LADH, a hydrogen from the a carbon of the alcohol is trans- ferred to NAD+ to produce the reduced form of the coenzyme (NADH). This was elegantly demonstrated by the use of deuterium labeled substrates and coenzymes (5,6). In the same studies it was shown that the transfer of the hydrogen \I I??? <’ 'N’J +NH HCHgO—P-O—P—OCHQ N I I O WON” OH on W H H HOHHOH NAD+ Figure 1.--Structure of oxidized nicotinamide adenine dinucleotide. is stereospecific with respect to the coenzyme, as chemi- cally labeled NADH lost only 50% of the label to the sub— strate, whereas enzymatically prepared NADH-d lost 100% of the label. The site of reduction of NAD+ in the enzymatic reaction was shown by Colowick and coworkers (7) to be at C-4 of the nicotinamide ring. The reduction produces a methylene carbon that is a prochiral center and it was evident that the yeast and liver alcohol dehydrogenases could distinguish between the two enantiotopic hydrogens. By using tritium and deuterium to determine which enantio- topic hydrogen was transferred, various investigators classified alcohol dehydrogenases as either type A or type B. YADH and LADH are both type A enzymes. By using a degradation method on enzymatically labeled NADH-d, Cornforth and coworkers (8) determined the absolute config- uration at C-4 of the coenzyme. They concluded that type A dehydrogenases utilize the pro-R hydrogen of NADH and type B enzymes utilize the pro-S hydrogen, as pictured in Figure 2. These hydrogens are alternatively labeled as HA(HR) and HB(HS). Hs HR Hs ’ coNH2 / CONH2 N' + RCR’+H* ; L‘ M + R “a“ l 'Ype A I z z z : ribose . diphosphote -ribose-adenine Figure 2.--Stereochemistry of hydrogen transfer from NADH. Westheimer, Vennesland, and Loewas (9) demonstrated that YADH reduces acetaldehyde and oxidizes ethanol stereo- selectively. This stereoselectivity was later shown to be absolute (10). To completely describe the stereochemical outcome of the enzymatic reduction of acetaldehyde-l-d it was necessary to know the absolute configuration of the optically active ethanol-l-d product. This was accomplished by Lemieux and Howard (11) who determined that (+)-1- deuterioethanol has the R configuration by using a corre- lation scheme involving the degradation of S-deuterio-D- xylose. Since the deuterated ethanol produced by the action of YADH and LADH on acetaldehyde-l-d was levorota- tory, they concluded that the product from the enzyme reaction has the S configuration. In the Hansen nomen- clature, the hydrogen is transferred to the re face of the carbonyl. Thus, it had been verified that the two alcohol dehydrogenases were stereospecific towards both their coenzyme and at least one of their natural substrates. Studies (12,13) on the product stereospecificities of a large number of ketone reductions by fermenting yeast preparations revealed a pronounced predilection for the generation of the alcohol with the S configuration. Using purified YADH, Van Eys and Kaplan (14) examined the kine- tics of oxidation of a variety of primary and secondary alcohols. From the results with (d,l)-2-butanol and (+)-2-octanol, they concluded that the enzyme utilizes only the S enantiomer of the alcohols as a substrate. From all of these studies, it is apparent that some mechanism exists whereby the enzyme orients the substrate molecule with respect to the coenzyme so that a stereochemically defined transfer of hydrogen occurs. Steric effects were proposed as the controlling factor in references 12 and 14. Hydro- phobic and hydrophilic interactions between enzyme and substrate were considered to be important by other workers. As evidence of a hydrophobic binding region at the active site of alcohol dehydrogenase, the binding of long chain fatty amides was cited by Winer and Theorell (15). ' ° enzyme Figure 3.--Karabatsos model. Based on the stereochemical results of the alcohol dehydrogenase reactions, Karabatsos and coworkers (16) proposed the model shown in Figure 3 for the spatial rela- tionship of the substrate to the coenzyme in the enzyme complex. In this model the carbonyl of the substrate is lying in a plane above the plane of the NADH nicotinamide ring. To account for the substrate stereoselectivity it was suggested that a hydrophobic region of the enzyme preferentially binds the larger alkyl group of the sub- strate. This region is pictured as being on the same side of the nicotinamide ring'as the carbamido group which would produce the alcohol with the S configuration after transfer of the pro-R hydrogen of the coenzyme. This model is simi- lar to the one proposed by Prelog (17) as a result of his group's experiments with cyclic ketone reductions by Curvularia falcata, an organism that contains a type B dehydrogenase. Prelog's model is shown in Figure 4. B-type enzyme 5 L 7 \ \\ \ \\ H‘ (3H S-configuration Figure 4.--Prelog model. These two models differ in that the carbamido group of the coenzyme is placed beneath the larger carbonyl substituent of the substrate in the Karabatsos Model and below the smaller alkyl group in the Prelog Model. In devising his model, Prelog considered non-bonded interactions between the carbamido group of the coenzyme and the alkyl groups of the substrate to be important in aligning the molecules. This model, however, predicted that type A dehydrogenases would reduce ketones to alcohols with the R configuration, just the opposite of what is observed. Karabatsos felt that the binding of the large group of the substrate to a hydrophobic pocket of the enzyme is a more important factor in determining the stereochemical outcome of the reaction. That hydrophobic and hydrophilic interactions between substrate and enzyme are important in the reactions of dehydrogenases was convincingly demonstrated by the work of Nunez (18) and Stamoudis (19) in this laboratory. Karabatsos and Nunez determined that the product from the reduction of hydroxyacetone with purified glycerol dehydro- genase, a type A enzyme, was 100% optically pure (R)-1,2- propanediol. The stereochemistry of the product suggests a transition state for the reduction that resembles the substrate-coenzyme configurational relationship shown in Figure 5 in which the hydrophilicity of the CHZOH group has reversed the positions of the large and small group of the Karabatsos Model. glycerol-Dig 2 H (DH 100% R—configuratinn Figure 5.--Reduction of hydroxyacetone by NADH with glycerol dehydrogenase. Using purified LADH and NAD+, Karabatsos and Stamoudis studied the reduction of methyl ethyl ketone and chloro- acetone. Methyl ethyl ketone yielded (+)—2—butanol that was 44% optically pure (72% S enantiomer) and chloroacetone gave (+)-l-chloro-2-propanol that was only 6.8% optically pure (S enantiomer slightly in excess). The larger size and greater hydrophobicity of the ethyl group over that of the methyl group lead to a prediction based on the Karabatsos Model that is in accord with the experimental results for methyl ethyl ketone (see Figure 6). A similar product stereoselectivity of 64-67% had been observed (12) in the reduction of methyl ethyl ketone by fermenting yeast. 727$S-configuration Figure 6.--Reduction of methyl ethyl ketone by NADH with LADH. The carbonyl substituents of chloroacetone are approxi- mately the same in size as those of methyl ethyl ketone but the hydrophilicity of the chloromethyl group is much greater than that of the ethyl group. The nearly racemic product mixture of the reduction of chloroacetone implies the configurations shown in Figure 7. Apparently the effects of steric and hydrophobic-hydrophilic interactions are nearly balanced in this system. 10 c Hg’ H“ ONH’ CONH, CHIC' —-—-* s R d-——— CICH | CH l 2 J Kim l-chloro-Z-propanol Ir (3 z 53% 47% Figure 7.--Reduction of l-chloro-Z-propanone by NADH with LADH . The product-controlling ternary complex of the alcohol dehydrogenases has been discussed up to this point in terms of representations in which the oxygen of the sub- strate is located over the nitrogen of the coenzyme pyridine ring. Prelog (17) had favored this orientation because he believed that it involved minimal non-bonded interactions between substrate and coenzyme. However, as the stereo- chemical results of the reactions with type A dehydrogenases have shown, these non-bonded interactions are probably not the dominant factor in orientating the substrate. The model proposed by Karabatsos also envisions the carbon-oxygen bond of the substrate as pointing towards the pyridine ring nitrogen. This assumption was adopted from a similar model advanced by Kosower (20) based on the observation of a shift in the absorption spectrum of NADH from 340nm to 325nm upon binding to LADH. Studies on the spectra of 11 model systems supported by calculations on their spectral transition energies led to his suggestion that this shift was caused by the presence of a positively charged nitrogen atom, such as a lysine residue of the enzyme, in the vicinity of the nitrogen of the dihydrOpyridine ring. Kosower further postulated that this quaternary nitrogen was hydrogen bonded to the ribose of the coenzyme and to the oxygen of the substrate, thus positioning this oxygen over the pyridine ring nitrogen. No direct evidence sup- porting this arrangement has been reported and without such confirmation, any proposal of a particular structure of the enzyme-coenzyme-substrate complex based on substrate stereospecificity must be considered as only speculation. Examination of the literature dealing with the mechanism of alcohol dehydrogenases reveals that there is no general agreement on the structure of the ternary com- plex. Graves, Clark, and Ringold (21) preferred an arrange- ment of the coenzyme and cyclic ketone substrates such as illustrated in Figure 8. Cervinka and Hub (13), from their studies on fermenting yeast reductions of ketones, drew the pictorial representation of the substrate-coenzyme orientation shown in Figure 9. More recent studies (22) done with addition products of NAD+ also support a model for the ternary complex in which the carbonyl of the 12 C) Ha HA CONH2 CONH2 'C I I z/T L N \N s I I Z Z Figure 9.--Mode1 of Figure 8.--Arrangement Cervinka and Hub. suggested by Graves. substrate is pointing away from the pyridine ring nitrogen of the coenzyme. The author's representation for the mechanism of lactate dehydrogenase is pictured in Figure 10. N.— Figure 10.--Model for the mechanism of lactate dehydrogenase. 13 The stumbling block that this confusion has created is attested to by the following quote from this paper: To the best of our knowledge, there is no informa- tion available at the present time concerning the position of the substrate molecule on the enzyme, except that the substrate is located somewhere near the nicotinamide ring. However, a more detailed knowledge of the spatial orientation of the coenzyme and the substrate is important to the understanding of the reaction mechanism of the dehydrogenases. Our approach to the solution of this difficulty was to covalently attach one end of a substrate molecule to + . . . . NAD , thereby reduc1ng the number of p0331ble or1entat10ns that the substrate could assume with respect to the coenzyme. If the substrate-coenzyme analog exhibits reactivity with alcohol dehydrogenase, information concerning the spatial relationships of the species participating in the reaction could be inferred. Furthermore, studies on such analogs that are enantiomerically purified at the substrate alcohol center might yield data on the location of the hydrophobic and other binding regions of the enzyme. As illustrated in Figure 11, if only one of the stereoiosmers of the NAD+ analog exhibited reactivity with alcohol dehydrogenase, this would be direct evidence for a particular placement of the carbon-oxygen bond of the substrate in the enzyme reaction. Initial studies of several NAD+ analogs by J. Miedema (23) demonstrated the feasibility of using these covalently joined substrate-coenzyme molecules as probes of the alcohol dehydrogenase active sites. l4 IADH R-configuration Figure ll.--Stereochemistry of the reaction of NAD+ analogs. This thesis describes the synthesis and reactivity studies of the NAD+ analogs (I), (IR), and (IS) shown in Figure 12. These studies required the synthesis and resolution of 7-amino—2-heptanol and the determination of the resolved aminoalcohols' enantiomeric purity and absolute configurations. In the course of this investigation, 6-amino-2-hexanol and its corresponding nicotinamide were 15 C) (3H \ .. 1' .. / H ’ I4 ;+ I racemic IS S-Configuration IR R-configuration Z : ribose- diphosphote -ribosc-odenine Figure 12.--NAD+ analogs (I). also synthesized and these two compounds along with several other aminoalcohols and nicotinamides were tested as sub- strates for liver alcohol dehydrogenase. EXPERIMENTAL Instrumentation Proton NMR spectra were recorded on a Varian T-60 NMR spectrometer. Fluorine NMR spectra were obtained with a Varian A-56/60D spectrometer. All infrared spectra were recorded on a Perkin Elmer 237B IR grating spectrophoto- meter. Melting points were determined with a Hoover capil- lary melting point apparatus. A Bodenseewerk model 141 polarimeter with auto- matic readout was used to measure optical rotations. The instrument was equipped with sodium and mercury lamps to allow readings at wavelengths of 589nm, 578nm, 546nm, 436nm, and 365nm. The pH measurements of buffer solutions and enzyme reaction solutions were done with an Instrumentation Lab- oratory model 245 pH meter. A Sorvall RCZ-B refrigerated centrifuge was used to isolate the products of the NAD analog synthesis. Column separations were monitored by passing the eluate through a 0.1 cm flow cell in a Beckman DB-G grat- ing spectrophotometer connected to a Sargent SR recorder. 16 l7 Fractions were collected with a Buchler Fractomette 200 collector that had an output to the recorder to record tube changes. Ultraviolet spectra for the reactivity studies were recorded on a Unicam SP800 spectrophotometer. Synthesis of 7-amino-2-heptanol Synthesis of ethyl (Begyanopropyl)- acetoacetate Ethanol 800ml was distilled from magnesium meth- oxide directly into a 2 liter three-necked round? bottomed flask that had been oven-dried. To this flask was added 239 of sodium metal under\a flow of dry nitro- gen. When all of the sodium had reacted, 509 of powdered potassium iodide and 220g of freshly distilled ethyl acetoacetate was added to the flask. The flask was fitted with a reflux condenser and mechanical stirrer and the reaction mixture heated to reflux for ten minutes. The contents of the flask were allowed to cool somewhat, then 1009 of 4-chlorobutyronitrile (Aldrich) was added slowly with stirring. The nitrogen flow was discontinued at this point and the reaction mixture heated at reflux for seven hours. During this time a white precipitate of sodium chloride formed and the pH of the solution decreased to about 9. The condenser was then replaced with a distil- lation head and the ethanol was distilled. After addition 18 of 600ml of a 50/50 mixture of benzene and water to the cooled residue, the organic layer was separated, and the aqueous layer was extracted three times with 75ml of ethyl ether. The combined organic layers were washed with saturated sodium chloride solution and dried over anhydrous magnesium sulfate. Following filtration the solvent was removed from the reaction products by rotary evaporation. Fractional distillation of the residue under vacuum yielded several fractions. The first fraction consisted mainly of ethyl acetoacetate; the second fraction, b.p. 72-75° at 0.16mm, was a pale yellow liquid; the third fraction, b.p. 96-122° at 0.14mm, was a mixture; and the fourth fraction, b.p. 122-125° at 0.14mm, was the desired product. The NMR spectrum of this fraction was consistent with the struc- ture of ethyl (3-cyanopropyl)-acetoacetate. The yield was 103.99, 55% of theoretical. A second reaction using the same procedure gave another 1109 of product. Synthesis of 6-cyano- 2-hexanone To a solution of 2009 of sodium carbonate in 1600m1 of water in a 2000m1 round—bottomed flask equipped with a reflux condenser was added 213.79 (1.08 mol) of ethyl- (3-cyanopropy1)-acetoacetate. The reaction flask was con- nected to a gas bubbler filled with barium hydroxide 19 solution to detect the evolution of carbon dioxide. The reaction mixture was refluxed for six hours, by which time carbon dioxide formation had ceased. After cool- ing, the reaction mixture was saturated with potassium carbonate. The two layers that formed were separated. The aqueous layer was extracted four times with ether and the combined organic layers were dried over anhydrous magnesium sulfate. Solvent removal and distillation yielded only one fraction, b.p. 78-79° at 0.5mm. The NMR spectrum of this material had only a methyl singlet (62.05ppm) and peaks for the eight methylene protons. IR (cc14): 1720cm'1, 2245cm‘1. Yield: 114.79 (0.92mol), 85% theoretical. Synthesis of 6-cyano- 2—hexanol The 114.79 of 6-cyano-2-hexanone was dissolved in 400ml of methanol in a 1000ml three-necked round- bottomed flask equipped with a thermometer, mechanical stirrer, and addition funnel. The contents of the flask were chilled to 0° with a salt-ice bath. To it was added a solution of 11.59 (0.30mol) of sodium borohydride in cold water, slowly with stirring so as to maintain the temperature of the reaction mixture below 10°. After com- pletion of the addition, stirring was continued for one hour. This was followed by the addition of 25% sulfuric acid until the reaction mixture was slightly acidic. The 20 reaction mixture was allowed to stand overnight, then the precipitate was filtered. Most of the methanol was removed from the filtrate by rotary evaporation. Addition to the residue of 300ml of ether resulted in the formation of two layers. The organic layer was washed with saturated sodium chloride solution and dried over anhydrous potassium carbo- nate. Distillation of the solvent, followed by vacuum distillation of the residue, produced 101.69 (0.80mol) of the alcohol, b.p. 92-94° at 0.25mm. Yield: 87% of theo- retical. Synthesis of 7-amino- 2-hgptanol Raney nickel W-2 was prepared by the standard method (24). Approximately 99 of the catalyst was added to a 500ml Parr bomb containing a solution of 349 of 6-cyano-2-hexanol in 300ml of lON ammonia in methanol. The bomb was fitted to a Parr hydrogenation apparatus and hydrogen introduced to a pressure of 50psi. After twelve hours of shaking, the pressure in the bomb had fallen by the theoretical amount (43psi). The nickel was removed by filtration through celite. The filtrate was distilled yielding a single product, b.p. 96-97° at 0.9mm, which solidified in the receiver. Two more portions of the cyanoalcohol were hydrog- enated to give a total yield of 95.99 (0.732mol) 7-amino- 2-heptanol, 92% of theoretical. NMR (CDC13): 21 61.1 doublet (3), 61.4 broad multiple (8), 61.95 singlet (3-OH and NHZ)’ 62.6 triplet (2), and 63.65 multiplet (1). Synthesis of 6-amino-2-hexanol Synthesis of 5-cyano- 2:pentanone Ethyl (2-cyanoethy1)-acetoacetate was prepared by a Michael addition of ethyl acetoacetate and acrylonitrile using the method of Albertson (25). Of this ester, 154g (0.842mol) was hydrolyzed and decarboxylated by refluxing with a solution of 1609 of sodium carbonate in 1400m1 of water. After four hours of refluxing the solution had become homogeneous. Saturating with potassium carbonate caused an organic layer to separate. The aqueous layer was extracted five times with ether and the organic layers were combined. After drying over anhydrous magnesium sulfate, the ether was removed from the product by rotary evaporation. The residue was vacuum distilled through a short Vigreaux column to give 68.99 of product. The NMR and IR spectra of the product were consistent with the structure of 5-cyano-2-pentanone. The yield was 74% of theoretical. Synthesis of Segyano- 2-pentanol Into a 200ml three-necked round-bottomed flask equipped with a thermometer and stir bar was added a 22 solution of 63.59 (0.572mol) of S-cyano-Z-pentanone in 200 m1 of methanol. After chilling the contents of the flask to 0° with an ice bath, a solution of 7.29 (0.19mol) of sodium borohydride in a small amount of cold water was added at a rate such that the temperature of the reaction mixture did not rise above 10°. When all of the boro- hydride had been added, the reaction mixture was allowed to warm solwly to room temperature. The reaction was com- pleted by acidifying with 25% sulfuric acid. The precipi- tated salts were removed by filtration and the bulk of the methanol was removed from the filtrate by rotary evapor— ation. The residue was dissolved in ether; this solution was washed with water and saturated sodium chloride solu- tion, then dried over anhydrous magnesium sulfate. Solvent removal and vacuum distillation yielded 60.89 of product, 94% of theoretical. IR (CC14): 2240cm-1, 3480cm-l, 3630cm-1, no carbonyl absorption. Synthesis of 6-amino- 2-hexanol Into a 500ml glass Parr bomb were added 19.36 (0.170mol) of 5-cyano-2-pentanol, 4g of Raney nickel W-2, and 300ml of approximately 10N ammonia in methanol. Hydrogen was introduced into the bomb to a pressure of 45psi. After shaking for twenty hours the pressure had decreased the theoretical amount (27psi). The beautiful blue-green reaction mixture was filtered through celite. 23 Distillation of the filtrate gave a single product, b.p. 63-65° at 0.08mm. The NMR and IR spectra confirmed that this was 6-amino-2-hexanol. The yield was 18.39 (0.156mol), 92% of theoretical. Resolution of 7-amino-2-heptanol Resolution with d-tartaric acid Resolution of 7-amino-2-heptanol was accomplished by the method of repeated recrystallizations of its tar- trate salt. A mixture of methanol and ethyl acetate was found to be a suitable solvent for crystallization. 102.99 of 7-amino-2-heptanol was dissolved in 100ml of methanol and this solution added to a solution of 118.19 of d-tartaric acid (Eastman Organic) in 200ml warm methanol in an Erlenmeyer flask. The solution was stirred for one half hour, then 150 m1 of hot ethyl acetate was added and the flask set aside for crystallization. After standing in the refrigerator overnight, the crystals that had formed were collected by suction filtration and washed with a little cold methanol. This tartaric acid salt was repeatedly recrystallized from 80/20 (v/v) methanol/ethyl acetate. Melting points and optical rotations of the salt were used to follow the progress of the resolution. “Table 1 summarizes the data from the successive recrystal- lizations. 24 TABLE 1.--Crystallizations of the d—tartrate of 7-amino- 2-heptanol. Volume of Product Melting [d]D(°) Solvent Weight Po1nt (°) I 450ml 2009 122-125 II 400ml 1679 123-125 III 310ml 134g 124.5-126 +16.7 IV 380ml 120.59 125-126.5 V 200ml 112.59 126-127 ,+15.2 VI 200ml 104g 127-128.5 VII 250ml 87.59 127.5-128.5 +14.3 VIII 175ml 74g 128-129.5 IX 150ml 609 129-130 +13.2 X* 150ml 68.79 128.5-130 XI 150ml 59.89 129.5—130.5 +12.9 XII 150ml 51.69 130-13l.5 +12.4 XIII 125ml 49.19 130.5-131.5 +12.2 XIV 125ml 42.39 131.5-132 XV 100ml 38.39 l32-132.5 +12.l it For this recrystallization the product from the ninth crystallization was combined with another 14.79 of tartrate obtained by taking a second crop of crystals from the mother liquors of the fourth cyrstallization. It was then carried through successive crystallizations using as a solvent the mother liquors of the original crystalliza- tions, lizations. then from fresh solvent for the last two crystal- 25 The final product from the resolution was dissolved in 100ml of warm methanol. To this solution was added a solution of 15.39 of potassium hydroxide in 75ml of methanol. After stirring for twenty minutes, the precipi- tate was filtered and washed with 50ml of methanol. The filtrate was distilled yielding 13.59 of 7-amino-2-heptanol, b.p. 86—87° at 0.6mm. The optical rotation of a 10.0m1 solution of 0.34629 of this 7—amino-2-heptanol in purified ethanol was measured at five wavelengths using ethanol as a blank. The measured rotations at 20° and the specific rotations are reported in Table 2. TABLE 2.--Optical rotations of (-)-7-amino-2-heptanol. Wavelength a(°) [a1200(°) 589nm -0.271 - 7.83 578nm -0.282 - 8.16 546nm -O.321 - 9.28 436nm -0.528 -15.25 365nm -0.797 -23.0 Another crop of the d-tartrate crystals was obtained by evaporating the mother liquors of the last three recrystallizations to a smaller volume. These ‘1! W-P'cjffll . 0 ‘1’“. n' ‘QTE f a «Wt-'9. 26 crystals were recrystallized twice from fresh solvent to give a product, m.p. 129-130°, which was saved for work on the determination of the absolute configuration of the resolved aminoalcohol. Resolution with V l-tartaric acid The mother liquors of the recrystallizations of the d-tartrate were evaporated to obtain the salt. An aqueous solution of the salt was treated with sodium hydroxide, then saturated with sodium chloride and extracted six times with methylene chloride. After drying over anhydrous potassium carbonate, the organic layers were distilled to yield 63.49 of 7-amino-2-heptanol. The 7-amino-2-heptanol was dissolved in 100ml of methanol and to this solution was added a solution of 72.69 of lftartaric acid (Aldrich) in 150ml of warm methanol. After stirring for one half hour, 100ml of warm ethyl acetate was added and the solution was allowed to cool slowly, then refrigerated overnight. The crystal— lized product was collected by suction filtration. The tartrate was dissolved in fresh methanol/ethyl acetate, the solution was filtered while hot, and then cooled for crystallization. Table 3 reports the weights and melting points of the products of the successive crystallizations. TABLE 3.--Crysta11izations of the i—tartrate of 7-amino- 2-heptanol. Volume of Solvent Product Weight Melting Point 1 350ml 1159 126.5-128 II 300ml 899 127-128 111 210ml 629 128-128.5 1v 175ml 53.59 128.5-129.5 v 150ml 47.59 129-130 v1 100ml 43.49 129.8-130.5 VII 90ml 39.29 130-131 v111* 100ml 40.19 131—131.8 1x 90ml 34.99 131-132 x 80ml 32.29 132-132.5 *A second crop of crystals was obtained from the mother liquors of the fourth crystallization and recrystal- lized from the mother liquors of the succeeding crystal— lizations. Finally, it was recrystallized three times from fresh solvent. This afforded another 5.69 of tar- trate which was combined with the product of the original seventh cyrstallization. The product from the tenth crystallization was dis- solved in 50ml of water. A solution of 9.209 of sodium hydroxide in 25ml of water was added with stirring to the tartrate solution. The solution was then saturated with sodium chloride and extracted with chloroform until the extract gave only a faint color with ninhydrin spray. The 28 chloroform extracts were dried over anhydrous sodium sul- fate, then distilled to give 13.2859 of 7-amino-2-heptanol. The Optical rotation of this product was determined using a 10.0ml solution of 0.18269 of the aminoalcohol in pure ethanol. The observed and specific rotations are listed in Table 4. TABLE 4.--Optical rotations of (+)-7-amino-2-heptanol. Wavelength a(°) [alzo°(°) 589nm +0.142 +7.76 578nm +0.148 +8.12 546nm +0.168 +9.22 436nm +0.281 +15.4 365nm +0.428 +23.4 Determination of the Absolute Configuration of (-)—7—amino- S:h§ptanol The (-)-7-amino-2-heptanol was reductively deaminated by a procedure modified from Nickon and Hill's method (26). The aminoalcohol for this reaction was obtained from the second crop of salt from the resolution with d—tartaric acid. The (-)-7-amino-2-heptanol was first converted to its sulfonamide by dissolving 1.2579 (9.61mmol) of it in 29 10ml of 10% sodium hydroxide solution and adding 1.24m1 (9.63mmol) of benzenesulfonyl chloride. The mixture was stirred vigorously for one hour, then acidified and extracted three times with chloroform. The extracts were dried over anhydrous magnesium sulfate and the solvent removed to yield 2.259 (87% of theoretical yield) of an almost colorless oil. The sulfonamide was dissolved in 100ml of 10% sodium hydroxide solution and this solution added to a 200ml round—bottomed flask equipped with a reflux con- denser and stir bar. The contents of the flask were heated to about 60°. Over the period of one hour, 129 of hydroxylamine-O-sulfonic acid (Alfa) was added in portions to the flask with vigorous stirring. The reflux condenser was then replaced with a distillation head and distilla- tion was begun. When 50ml of steam distillate had been collected, 25ml more water was added to the reaction flask and the distillation was resumed until another 50ml of distillate had been collected. The distillates were neutralized with 10% hydrochloric acid, saturated with sodium chloride, and extracted four times with ether. The combined extracts were dried over anhydrous sodium sulfate. The solvent was removed by distillation through a Vigreaux column leaving a residue with the characteristic odor of 2-heptanol. GLC analysis of the residue showed that it 30 contained ether, 2-heptanol (identified by mixed injection with a known sample), and very little else. The product was short-path vacuum distilled to give 0.10189 of 2-heptanol, GLC pure. The 2-heptanol was dissolved in 6.0m1 of pure ethanol for the determination of its optical activity. The observed optical rotation at 589nm was -0.109° and at 365nm was -0.289°. The calculated specific rotations at ambient temperature are -6.43° and -l7.0° at 589nm and 365nm, respectively. NMR Method for the Determination of Enantiomeric Purities The aminoalcohols were converted to their sulfon- amides by reaction with an equivalent amount of benzene- sulfonyl chloride in 10% sodium hydroxide solution. The sulfonamides were obtained pure by acidification of the reaction mixture and extraction with chloroform. Yields were generally 75-80% of theoretical. (+)- and (-)-d-methoxy-a-trifluoromethylphenyl— acetic acid were obtained from Aldrich Chemical Co. The acid chlorides were prepared by refluxing the acid with purified thionyl chloride and a small amount of sodium chloride for fifty hours followed by vacuum distillation. About 0.35mmol of the sulfonamide was dissolved in 10 drops of dry chloroform. To this solution was added 31 a slight molar excess of the acid chloride, ten drops of dry carbon tetrachloride, and ten drops of pyridine. The solution was allowed to stand in the stoppered flask for twelve hours. Two ml of water was added and the reaction mixture transferred to a separatory funnel with 25ml of ether. The ether solution was washed twice with 10% hydrochloric acid solution, then once with sodium carbonate solution, water, and saturated sodium chloride solution. After drying over anhydrous magnesium sulfate, the ether was removed from the product. The residue was taken up in deuteriochloroform for NMR analysis, using approximately 15% trifluoroacetic acid as an internal standard for the fluorine NMR spectra. Resolution of 6-amino-2-hexanol The 6-amino-2-hexanol was partially resolved by using dibenzoyl-d-tartaric acid. A quantity of 23.959 (0.205mol) of the aminoalcohol was added to a warm solu- tion of 79.19 (0.210mol) of dibenzoyl-d-tartaric acid monohydrate (Aldrich) in 150ml of methanol. After stir- ring for one half hour the solution was filtered hot, 400ml of ether was added to the filtrate, and the flask was set in a refrigerator overnight. The crystals that had formed were collected and recrystallized from 75/25 ethyl acetate/methanol. A summary of the crystalliza- tions is given in Table 5. 32 TABLE 5.--Crystallizations of the dibenzoyl-d-tartrate of 6-amino-2-hexanol. Volume of Solvent Product Weight Melting Point (°) I 1129 147-150 dec. II 300ml 989 151-153.5 III 200ml .669 155-156.5 IV 150ml 389 156-157 v 100ml 25.99 157—158 vr 75ml 11.59 159-160.5 VII 55ml 7.69 160.5-161.6 The product from the seventh cyrstallization was added to 2.49 of sodium hydroxide in water. The solu- tion was saturated with potassium fluoride and extracted ten times with chloroform. The chloroform extracts were dried over anhydrous sodium sulfate. Solvent removal and vacuum distillation yielded 1.139 of 6-amino-2-hexanol. The specific rotation of the product in ethanol was [a]§°-—2.6°. Synthesis of the N-(hydroxyalkyl)- nicotinamides Nicotinyl chloride was prepared by the slow addi- tion of 36ml of purified thionyl chloride to a suspension of 64.59 of potassium nicotinate in 400ml of dry carbon tetrachloride. The reaction mixture was refluxed for 33 twelve hours, then the solids were filtered and the sol- vent removed from the filtrate under vacuum. Vacuum dis- tillation gave one fraction, b.p. 38-39° at 0.3mm. The synthesis of (-)-N-(6-hydroxy-l-hepty1)- nicotinamide is described herein as representative of the procedure used to prepare all of the N-(hydroxyalkyl)- nicotinamides. To an oven-dried 500ml four-necked round—bottomed flask was added 12.629 of the (-)-7-amino-2-heptanol obtained from the resolution with d-tartaric acid. The flask was fitted with a mechanical stirrer, reflux con- denser, rubber septum, and nitrogen inlet. All glass connections were sealed with Teflod:>sleeves. After flushing the flask with dry nitrogen, the aminoalcohol was heated until it melted. A catalytic amount of ammonium chloride was added, followed by 10.1ml of hexamethyldisilazane (Aldrich). Stirring was begun, the flask was heated, and ammonia began to evolve. The contents of the flask were heated at 100° for one hour, by which time the evolution of ammonia had just about ceased. To the reaction flask was added 70ml of pyridine and 20 ml of triethylamine, both having been dried and distilled under nitrogen. An amount of 10.6m1 of nicotinyl chloride was then added slowly from a dis- posable syringe. The reaction being exothermic, the 34 heat source was removed from the flask. The colorless solution became yellow, then orange. When the nicotinyl chloride had been completely added the septum was replaced with a glass stopper and the reaction mixture was refluxed for one hour. The reaction mixture was cooled and the precipi- tated hydrochloride salt was filtered and washed with a little pyridine. The filtrate was transferred to a 250ml round-bottomed flask and the remaining triethylamine was distilled off. A reflux condenser was attached to the flask, 15ml of water was added, and the solution was refluxed for two hours. All of the solvent was then removed under vacuum, leaving a viscous brown oil. The NMR of the product indicated it to be mainly the desired N-(6-hydroxy-1-heptyl)-nicotinamide with a few minor impurities. TLC of the product on silica gel with 10/1 CHCl3/MeOH produced five spots: the major component with an Rf of 0.11, the others with Rf's of 0.92, 0.56, 0.36, and 0.0. The N-(hydroxyalkyl)-nicotinamides were purified by chromatography on a 2.5 X 80cm column of alumina (type F-l, 80-100 mesh) with 100/2 chloroform/methanol as the eluant. A 10ml aliquot of a solution of the nicotinamide product in 30ml of chloroform was loaded onto the column and eluted at a flow rate of about two ml per minute, while the column eluate was continuously 35 monitored at 254nm with a Beckman DB-G UV spectrophotom- eter connected to a recorder. Several fractions were observed before the main component began coming off the column. This fraction was collected and the solvent distilled off, leaving a slightly yellow oil. The N-(hydroxyalkyl)-nicotinamides were further purified by dissolving the oil in 40ml of acetone, boil- ing with a little Norit-A, then gravity filtering. The filtrate was evaporated to a volume of about 15ml and set in a freezer at -20°. Scratching led to the slow formation of crystals. The products were recrystallized one or two more times until a constant melting point was obtained. Crystallization was greatly facilitated by seeding with a crystal from the previous recrystalliza- tion. The N-(6-hydroxy-l-heptyl)-nicotinamides syn- thesized from the racemic and optically active 7-amino- 2-heptanols were characterized as follows: The melting points were (R,S) 39-41°, (R) 48-49.5°, and (S) 48.5- 50°. The NMR spectrum of (8)-N-(6-hydroxy-1-heptyl)- nicotinamide is shown in Figure 13. By deuterium exchange, the singlet at 63.1ppm was identified as the OH peak and the triplet at 67.4ppm as the NH peak. Irradiation of the peak at 67.4ppm caused decoupling of the peak at 63.3ppm. IR spectrum (CHZClZ): 3610, 3450, 3325, and 1660cm’1. The optical rotations of 36 .ooAEoeHoooacuLampoonuausxooomnuovuzuxmv “0 ssuuoomm mzzn-.MH ouomfim o 9. on o. 9. 1. C wt 0 . .4 o a 9. 1I_II I— I‘M 1.1 j—Ij.d ‘Id IuidefiIlaldI.I—d1 Q d8}.\l’{" J O 7". . f . . 1 sWII-I----..I IIII; 1:. 2.113.791 1 ,1, ~ \ a, .35 .63. 1. 1 I ,4 L. L 5. .2 __ .1, .1 i... ,2; n1. 1 : 37 (R)-N-(6-hydroxy-l-heptyl)-nicotinamide, as measured on a 10ml solution of 0.27859 of it in ethanol, are given in Table 6. TABLE 6.-—Optical rotations of (-)-N-(6-hydroxy-l-heptyl)- nicotinamide. o 23o Wavelength dI ) [a] I ) 589nm -0.189 - 6.78 578nm -0.196 - 7.04 546nm -0.223 - 8.01 436nm -O.368 -13.2 365nm -0.563 -20.2 Prepgration of NAD+ Analogs To a reaction flask with Teflon© seal cap was added 0.1809 (0.25mmol) of e-NAD+ (Sigma, grade 111) and 0.2359 (1.00mmol) of N-hydroxyalkyl-nicotinamide. Twelve milliliters of 0.1M phosphate buffer, pH 7.5, was added and the solids were dissolved with gentle stirring. The pH of the solution was adjusted to 7.5 with 0.2M sodium hydroxide solution. The reaction flask was equilibrated in a constant temperature bath at 37.0°. Then, 0.309 (4.5 units) of pig brain NADase (Sigma) was added and the reaction vial was put back in the bath. The reaction 38 mixture was stirred vigorously for three hours at 37°. The reaction was stopped by chilling in an ice bath and adding 0.59 of trichloroacetic acid. The contents of the flask were transferred to a centrifuge tube and centri- fuged at 10,000rpm for ten minutes in a refrigerated centrifuge. The supernatant liquid was decanted and diluted with 60ml of cold acetone. After standing a few minutes, the mixture was centrifuged at 5000rpm for fif- teen minutes. The supernatant liquid was discarded. The small amount of solid lining the bottom of the centrifuge bottle was dried under a flow of nitrogen, then dissolved in 5ml of deionized water and the solution transferred to a sealed vial which was stored at 3°. A 0.9x 5-cm column was constructed of PEI cellu- lose anion exchange resin (Sigma, 1.17meq/9) and 0.003OM ammonium bicarbonate in deionized water. The column, which was kept in a refrigerator, was fully equilibrated before use. A lml aliquot of the analog product solution was loaded onto the column and eluted with 0.003M ammonium bicarbonate solution using a polystatic pump to obtain a flow rate of 1.4ml per minute. The eluate was monitored by passing it through a 0.1cm flow cell in a UV spec- trophotometer. Four fractions were separated. 39 Activity Determinations in Enzyme Systems Horse liver alcohol dehydrogenase (LADH) and yeast alcohol dehydrogenase (YADH) were obtained in vials of lyophilized protein from Sigma Chemical Co. The enzyme was reconstituted prior to use by adding 1.0m1 of phos- phate buffer, pH 7.5. These solutions were stored at -15°. B—NAD (Grade III) was also obtained from Sigma. The buffer stock solutions used in the enzyme studies were prepared as described in Methods in Enzy- mology (27). The following buffers were used: pH 7.5 phosphate, pH 10 glycine/NaOH, and pH 8.8 glycine/NaOH. All water used in these studies was house- supplied distilled water that was passed through an ion exchange column and boiled. The extent of the enzyme reactions were measured as the increase in the UV spectrum at 340nm due to the reduction of the pyridine ring of the nicotinamide adenine dinucleotide. Blank solutions were used as references and care was taken to establish the same base- line for spectra taken over long time intervals. The activity of the NAD+ analogs with the alcohol dehydrogenases was determined in solutions made with 1.0ml of the column eluate containing the analog and 2.0ml of the buffer stock solution. One unit of LADH or ’ wh‘d‘jln'r Lug-4 L 40 25 units of YADH was added and the UV spectrum taken immediately and again at various time intervals. The solutions were incubated in the UV cuvettes either at ambient temperature or at 30° in a constant temperature 2M solution of bath. Measured amounts of a 3.00x10- ethanol in water were added to the enzyme reaction solu- tion for testing the coenzyme function of the analogs. More concentrated solutions of the NAD+ analogs were obtained by two processes: lyophilization and molecular filtration. The molecular filtration method © involved the use of a Millipore molecular filtration apparatus with a Pellicon© PSAC membrane to concentrate the analog fractions from several column runs. The remain- ing ammonium bicarbonate was removed by twice diluting with a buffer solution and reconcentrating. These concentrated solutions of analog and buffered solutions of the lyophil- ized analogs were tested for reactivity with LADH. Column fractions containing NAD+ were also concentrated by these methods as a check of their capability to produce intact coenzymes. The general procedure for determining a substance's activity as a substrate for LADH was to add 1.00m1 of an aqueous solution of the compound to 2.00ml of a stock solution of NAD+, about 5.2x10-4M, in the buffer of interest. One or one-half unit of LADH was added and the UV spectrum determined at various time intervals. Relative 41 initial rates of NADH production were measured as the increase in absorbance at 340nm. Equilibrium concentra- tions of NADH were determined after incubating for 24 hours or more. When no further increases in the peak at 340nm were noted, fresh enzyme was added to insure that a true equilibrium had been reached before enzyme denatur- ation. Alternatively, ethanol was added to check the activity of the enzyme system. Equilibrium constants for the general reaction: NAD+ + RCHOHR' e=ée NADH + RCOR' + H+ were calculated by using the expression X2 . [H+] Keq = ([NAD]0 - x)([RCH0HR']O - x) where x = [NADH] = [RCOR'] at equilibrium and [NAD+]O = initial concentration of NAD+ and [RCHOHR']0'= the initial concentration of the alcohol. The concentration of NADH was determined from the absorbance at 340nm by using the reported (28) extinction coefficient of 6.22x 106 cmZ/mol. The [H+] was fixed by the pH of the buffer solution. For equilibrium constants determined on the analog systems, the extinction coefficients at 260nm and 340nm were assumed to be equivalent to those of NAD+ for calculating the necessary concentrations. RESULTS AND DISCUSSION Resolution of 7-amino-2-heptanol Enantiomeric Purity of the Resolved 7-amino- 2-heptanols Since interpretation of the type of results which were the goal of this research project depended on having NAD+ analogs of known stereoisomeric composition, it was necessary to have some means of determining the extent of resolution of the aminoalcohols from which the analogs were synthesized. Two methods were investigated. The first attempt at solving this problem was to study the effect of an optically active shift reagent on the NMR spectrum of the aminoalcohol. Europium (III)- tris-(3-heptafluoropropyl-hydroxymethylene)-d-camphor (Optishift II, Willowbrook Laboratories) was added in portions to a chloroform-d solution of 7-amino-2-heptanol and the NMR spectrum was recorded after each addition. Since the europium reagent complexes with both alcohols and amines, it was necessary to add large amounts of this shift reagent to the solution (l/l molar ratios of shift reagent to alcohol were found to be needed to produce significant chemical shift differences for enantiomers of simple a1cohols)(29). The peak broadening that resulted 42 43 from the use of two equivalents of shift reagent made the integration of any slightly separated peaks impractical. Because of this and the expense of the optically active shift reagents, this approach was abandoned. The next approach proved more successful. Use was made of fluorine NMR to measure the relative amounts of the two diastereomeric esters formed from the reaction of optically active d-methoxy-a-trifluoromethylphenylacetyl chloride with the sulfonamide of the resolved amino- alcohols. In 1969 Mosher, Dale, and Dull (30) reported the successful use of this reagent in the determination of the enantiomeric composition of a number of secondary alcohols. Since the acid chloride also reacts with amines, this complication in the present study was avoided by pro- tecting the amino group by conversion to its sulfonamide before reaction with the acid chloride. Using the pro- cedure described in the experimental section, the Sulfon- amides were obtained cleanly in good yields. The NMR spectrum (triplet at 5.1ppm, doublet of triplets centered at 2.9ppm, singlet at 1.8ppm in CDC13) and IR spectrum 1 1 (sulfonamide absorptions at 1325cm- and 1160cm- and 1 hydroxy absorptions at 3530cm- and 3620cm-1) confirmed the identity of the products. The reaction of the MTPA chloride with the sulfon- amides was clean and yielded nearly theoretical yields of 44 the esters. The product of the reaction of (R)~d-methoxy- a—trifluoromethylphenylacetyl chloride with the sulfon- amide of unresolved 7—amino-2-heptanol exhibited two single peaks of equal intensity, separation lOHz, in its fluorine NMR spectrum. The chemical shifts and peak separation varied somewhat with the concentration of trifluoroacetic acid used as an internal standard. The relative amount of the two diastereomers resulting from the reaction of an optically active acid chloride with an alcohol depends on the enantiomeric purity of the acid chloride as well as the alcohol. Since the MTPA purchased from Aldrich is presumably produced by a resolution, it is not safe to assume that it is optically pure. The optical purity of the (+)-MTPA was checked by fluorine NMR analysis of the esters of lementhol (EKC Chemicals) and 2—octanol obtained from the reduction of 2-octanone with fermenting yeast. The l—menthol is extracted from a natural source and thus is likely to be 100% optically pure; and the reduction of 2-octanone by yeast has been shown (14) to be at least 98% stereo- specific. Both of these esters had two peaks in the fluorine NMR spectrum with an area ratio of 94/6, indicat- ing that the (R)-MTPA chloride was only 88% optically pure. This was taken into account in determining the enantiomeric purity of the resolved aminoalcohols. 45 The fluorine NMR spectrum of the (R)-MTPA esters of (-)-7-amino-2-heptanol and (+)-7-amino-2—heptanol showed two peaks with the area ratios of 80/20 and 81/19 respectively. The downfield peak was the major one in the ester of (-)—7-amino-2—heptanol; the upfield one was the major peak in the spectrum of the (+)-7-amino-2-heptanol ester. Correcting for the enantiomeric purity Of the acid chloride, we calculated that the (-)-7—amino-2-heptanol was 68% optically pure (84/16 ratio of enantiomers) and the (+)-7-amino-2—heptanol was 70% optically pure (85/15 ratio of enantiomers). In the following discussions (-) and (+)-7-amino-2-heptanol refer to the compounds Of these enantiomeric purities. thical Rotations of the Resolved 7-amino- 2-hgptanols The optical rotatory dispersion curve of the (-)-7-amino—2-heptanol obtained from the resolution with d—tartaric acid is shown in Figure 14. The (+)-7-amino- 2-heptanol gave a similar curve. By using the Optical purities of these aminoalcohols, the specific rotation for 100% Optically pure 7-amino-2-heptanol was calculated to be 11.3° and the molecular rotation to be 14.8° at 589nm and 20°. This can be compared with the molecular rotation of (+)-2-heptanol which has been reported (31) to be +12.0° at 20°. no 0 ’5 0 (Ow [a] (o) \\ 5 r : e e <9— 350 goo 450 500 $50 600 A (nm) Figure l4.--ORD curve of (-)-7-amino-2-heptanol. 20 4' [Sq ’0: I31] (°) 5 4 3 A L L 350 900 450 {on 350 600 A.(hml Figure 15.--ORD curve of (-)-N-(6-hydroxy- l-heptyl)-nicotinamide. 47 Absolute Configurations of (-)- and (+)-7-amino- 2-hgptanol The absolute configurations Of the methyl n-alkyl carbinols have been established as being (S):(+) and (R).(-) (32,33). The absolute configuration of the resolved 7-amino-2-heptanols could be assigned by reduc- tively deaminating them to 2-heptanol and correlating the Optical rotations. Very few methods have been reported for carrying out this type Of reaction, but use was made of the procedure described by Nickon and Hill (26) wherein the amine was first converted to its sulfonamide and then reductively deaminated with hydroxylamine-O-sulfonic acid in basic solution. For this reaction the (-)-7-amino-2- heptanol obtained from a second crop of the d—tartrate was used since it was only necessary to have an excess of one enantiomer for the determination of the sign of rotation Of the 2-heptanol product. Since the 2-heptanol isolated from the reaction mixture had a specific rotation at 589nm Of -6.43°, the (-)-7—amino-2-heptanol has the R configu- ration and the (+)-7-amino-2-heptanol has the S configu- ration. As a further indication that these configurational assignments are correct, it was noted that with all of the methyl alkyl carbinols studied in the experiments of Mosher (30) and the present work, the ester diastereomer 48 made with the (S)-alcohol and (R)-MTPA chloride produced the trifluoromethyl peak with a lesser downfield shift than the one made with the (R)-alcohol. This NMR method might prove useful in assigning absolute configurations to Optically active alcohols and amines. Resolution of 6-amino-2-hexanol The reaction of 6-amino-2-hexanol with d-tartaric acid and (+)-mandelic acid produced oils that could not be crystallized. The 6-amino-2-hexanol did form a crystalline salt with d-lO-camphorsulfonic acid, but after six recrystallizations the recovered aminoalcohol was still nearly racemic. Considerable difficulty was encountered in recovering the aminoalcohol from the salt in good yield. Apparently 6-amino-2—hexanol is more soluable in water than in most organic solvents; salting out and numerous extrac- tions were necessary to recover a reasonable amount of the aminoalcohol. Better success was Obtained by using dibenzoyl-d- tartaric acid as the resolving agent. Several grams of partially resolved 6-amino-2—hexanol were Obtained which had a specific rotation of -2.6° at 598nm. Analysis of the enantiomeric purity of this product by the fluorine NMR method indicated it to be 34% Optically pure (67/33 ratio of enantiomers). 49 Synthesis of Optically Active N-(6- gydroxy-l-heptyl)-nicotinamides (+)-, (-), and (d,l)-N-(6-hydroxy-l-hepty1)- nicotinamides were prepared from the resolved aminoalcohols with the same sign of optical rotation. The products were identified by their NMR and IR spectra. Synthesis of these compounds involved protection of the alcohol group of the aminoalcohol by formation of its trimethylsilyl ether and subsequent hydrolysis of the ether. The possibility that racemization occurs in this reaction can be eliminated on the basis of experiments done with 018 exchange (34) and optically active dimethyldi-Q—2-butoxysilane (35) which demonstrated that hydrolysis of silyl ethers proceed with breakage of the Si-O bond giving alcohols with retention of configuration. It can be assumed then that the enantio- meric purities and absolute configurations of the N-(6- hydroxy-l-heptyl)-nicotinamides are the same as the 7- amino-Z-heptanols from which they were made. An optical rotatory dispersion curve of the (-)-N-(6-hydroxy-1-heptyl)-nicotinamide (predominately R configuration) is presented in Figure 15. Assuming an optical purity of 68%, the [MSGo for 100% (R)-N-(6- hydroxy-l-heptyl)-nicotinamide is -9.98° and the molec- ular rotation is -23.5°. 50 Preparation of the NAD+ Analogs The NAD+ analogs (I), (IR), and (IS) were prepared by using the procedures developed by Kaplan and Ciotti (36), wherein pig brain NADase catalyzes the exchange of the nicotinamide portion of the coenzyme as shown in Figure 16. Excess amounts of NAD+ were used to drive the equilibrium to greater production of the analog. After separating the protein from the products, analysis Of the product mixture by TLC on polyethyleneimine cellulose plates with fluorescence indicator (Brinkman) using 0.15M aqueous ammonium chloride solution as the developing solvent showed two spots of about equal size, Rf 0.50 and 0.37, and two smaller spots with Rf 0.045 and Rf 0.023. The NAD analogs were isolated by anion exchange chromatography on polyethyleneimine cellulose as described in the Experimental section. A UV absorption-time trace of one column separation is shown in Figure 17. The first two fractions off the column were nicotinamide and N- alkylnicotinamide. The fourth fraction was identified as NAD+ by thin layer chromatography against a known sample of NAD+ and by its rate of reduction by ethanol with LADH. Fraction number three was identified as the NAD+ analog. This fraction had a peak in the UV spectrum at 260nm, as does NAD+, and formed an addition compound with cyanide ion as evidenced by a decrease in the absorbance at 260nm 51 mcmcmcm .mmOHmcm +Q OOH \ O. (ii-(CH3 \ raw-(“2% // bl Figure 20.--Possible intramolecular hydrogen bonding. The aminoalcohols: CH3CHOH(CH2)nNH2 where n= 3,4,5,6 were also tested as substrates for LADH, and a similar correlation of reactivity to alkyl chain length was found. Over the first twenty minutes of reaction, the relative reaction rates were about l.0/l.5/2.0/34 for 68 n=3,4,5, and 6 respectively. The equilibrium constants at 30° are reported in Table 8. Except for the eight carbon aminoalcohol, the equilibrium constants were smaller for the aminoalcohols than the corresponding nicotinamides. Perhaps this is a reflection of greater hydrogen bonding stabilization in the aminoalcohols, but certainly there are other factors operating in the equilibrium. TABLE 8.--Equilibrium constants for the oxidation- reduction of aminoalcohols. aminoalcohol [RCHOHR'] [NADH] K S-amino-Z-pentanol 1.00x10'3M 1.37x10‘5M 5.6x10' 6-amino-2-hexanol 1.05x10'3m 2.09x10‘5M 1.3x10‘ 7-amino-2-heptanol 1.02x10'3M 5.22x10‘5M A 9.0x10' 8-amino-2-octanol 1.02x10'3M 3.06x10'4M 3.0 4 3 3 In the reactions of liver alcohol dehydrogenase with analog (IS) and analog (IR), a marked stereoselectiv— ity was observed. It was relevant to the interpretation of these results to determine if this stereoselectivity also existed in the reaction of LADH and NAD+ with N-(6-hydroxy-l-heptyl)-nicotinamide (IIIc) and with 7-amino-2-heptanol. Solutions of known concentration of 69 (+), (-), and (d,l) 7-amino-2—heptanol and N-(6—hydroxy- l-heptyl)-nicotinamide were incubated with 0.5 units of 4M NAD+ in pH 10 buffer at 23°. The LADH and 3.47x10' reaction with the aminoalcohols was also run at 30°. In all three experiments the (+)—substrate produced NADH at an initial rate about five times greater than the (-)- substrate. The equilibrium amounts of NADH in the (+)- substrate solutions were 2-2.5 times as much as in the (-)-substrate solutions. The rate and equilibrium mea- surements suggest that the enzyme reacts only with the S enantiomer of the substrate. The slower reactions with the (-)-substrates were probably due to the minor amount of the S enantiomer in the mixture. As discussed pre- viously, the actual concentrations of the S enantiomer of the resolved aminoalcohols and nicotinamides was 85% of the total concentration of the (+) compoundsand 16% of the (-) compounds. Assuming that the R enantiomer of the compounds is unreactive with the enzyme, the equi- librium constants for the reactions were calculated using the concentrations of the 8 components as the substrate concentrations. Tabulation of these results appears in Table 9. The good agreement of the values from the (+) and (-) systems suggests that the assumption that the enzyme is stereospecific for the S enantiomer of both N-(6-hydroxy-l-heptyl)-nicotinamide and 7-amino-2-heptanol is valid. '- l! l‘ 1| 'Illll Illlll‘. III. III *I I .II‘ I ll Ill ‘1' ll" 70 TABLE 9.--Oxidation-reduction equilibrium in optically active systems. System [(S)-RCHOHR'] K (+)-IIIc 8.95x10'4M 1.73 (—)-IIIc 1.76x10'4M 1.77 (d,l)-IIIc 5.15x10'4M 1.89 (+)-7-amino-2-heptanol 1.25x10-2M 1.44x10-2(23°) (-)-7-amino-2-heptanol 2.39x10‘3M 1.30x10'2(23°) (+)-7-amino-2-heptanol 1.25x1o'2M 1.88x10'2(30°) (-)-7-amino-2-heptanol 2.39x10'3M 1.79x10'2(30°) Conclusion This thesis has described the synthesis of two aminoalcohols and their corresponding nicotinamides. Resolution of these aminoalcohols gave (+)- and (-)-7- amino-Z-heptanol and (-)-6-amino-2-hexanol. In the course of this work a general method for the determination of the extent of resolution and the absolute configuration of optically active aminoalcohols was developed. Analogs of nicotinamide adenine dinucleotide were prepared from the optically active 7-amino-2-heptanols and these were used to explore the stereochemical relationships in the active site of alcohol dehydrogenases. It was found that the analog (I) does function as a coenzyme for liver alcohol 71 dehydrogenase but not for yeast alcohol dehydrogenase. This was interpreted as evidence of a more sterically hindered coenzyme binding site in the yeast enzyme. The ability of the NAD+ analog to oxidize ethanol when in the presence of LADH indicates that the amino group of the amide portion of the coenzyme is not essential for the coenzyme function. The NAD+ analog (I), which consists of both a coenzyme and a substrate portion, displayed a small amount of activity with LADH. The analog (IS) exhibited a much greater amount of reaction with LADH than did the analog (IR). If this stereoselectivity represents an intramolecular oxidation-reduction reaction of the sub- strate and coenzyme portions of the analog, it would pro- vide support for the Karabatsos model for the ternary complex of the enzyme. However, intramolecularity could not be established. In fact, evidence was obtained that an intermolecular mode of reaction is available to the system. It is possible that the observed reaction of the analogs is the oxidation of the N-(hydroxyalkyl)- nicotinamides produced under the reaction conditions by cleavage of the nicotinamide-ribose bond of the analog, since these substituted nicotinamides were also found to serve as substrates for LADH. Stereospecificity for the enantiomers with the S configuration was observed in the reaction of N-(6-hydroxy-l-heptyl)-nicotinamide and :4‘...‘.; ‘9‘" 72 7-amino-2—heptanol with NAD+ and LADH, confirming the hypothesis that, in substrates with alkyl groups of greatly differing size, the enzyme has some mechanism of discriminating between the two stereoisomers, probably by binding of the larger group to a hydrophobic region. The question of the actual spatial orientation of the sub- strate and coenzyme in the ternary complex remains unanswered. Determination of the location of the hydro- phobic region and other binding sites of the enzyme, perhaps with the aid of other analog systems, would help to clarify this matter. In conclusion, it is hOped that the results of this study will stimulate further research into the elucidation of the structure of the active site of alcohol dehydrogenases and the mechanism by which these structural features control the stereospecificities of enzyme reactions. LIST OF REFERENCES 73 l‘l":.‘lly ll!.l‘l‘nl ’3‘. 10. 11. 12. 13. G. LIST OF REFERENCES POpjak and J. W. Cornforth, Biochem. J., 101, 553 (1966). — G. Ogston, Nature London, 162, 963 (1948). For a review see H. Sund and H. Theorell in "The Enzymes," 2nd ed., Vol. 7, pg. 25, ed. by P. D. Boyer, H. Lardy, and K. Myrback, Academic Press, New York (1963). . A. Hughes, G. W. Kenner, and Sir Alexander Todd, J. Chem. Soc., 3733 (1957). F. Fisher, E. E. Conn, B. Vennesland, and F. H. Westheimer, J. Biol. Chem., 206, 129 (1954). R. Levy and B. Vennesland, J. Biol. Chem., 228, 85 (1957). E. Pullman, A. San Pietro, and S. Colowick, J. Biol. Chem., 206, 129 (1954). W. Cornforth, G. Ryback, G. Popjak, C. Donninger, and G. Schroepfer Jr., Biochem. Biophys. Res. Comm., 2, 371 (1962). A. Loewus, F. H. Westheimer, and B. Vennesland, JACS, 15, 5018 (1953). R. Levy, F. A. Loewus, and B. Vennesland, JACS, 12' 2949 (1957). U. Lemieux and J. Howard, Can. J. Chem., 41, 308 (1963). MacLeod, H. Prosser, L. Fikentscher, J. Lanyi, and H. S. Mosher, Biochem., 3, 838 (1964). Cervinka and L. Hub, Coll. Czech. Comm., 31, 2613 (1966). 74 1'1 Illlllllll Ill .lll‘llj‘lillll' ‘Ix ltll IIII. Ila )I) II. I ll ill I I! 14. 15. 16. 17. l8. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. J. 75 van Eys and N- 0. Kaplan, JACS, 79, 2783 (1957). D. Winer and H. Theorell, Acta Chem. Scand., 13, 1729 (1960). J. Karabatsos, J. S. Fleming, N. Hsi, and R. H. Abeles, JACS, 88, 849 (1966). Prelog, Pure Appl. Chem., 8 (1), 119 (1969). Nunez, unpublished results (1971). C. Stamoudis, Ph.D. dissertation, Michigan State University (1973). . M. Kosower, Biochim. Biophys. Acta, 88, 474 (1962). Graves, A Clark, H. J. Ringold, Biochem., 8, 2655 (1965). Everse, E. C. 2011, L. Kahan, and N. 0. Kaplan, Biorg. Chem., 8, 207 (1971). Miedema, Ph.D. dissertation, Michigan State Uni- versity (1974). Mozingo, Organic Synthesis, Coll. Vol. III, 181 (1955). F. Albertson, JACS, 18, 2594 (1950). Nickon and A. S. Hill, JACS, 88, 1152 (1964. Gomori in "Methods in Enzymology," Vol. I, 138, ed. by S. Colowick and N. Kaplan, Academic Press, New York (1955). L. Horecker and A. Kornberg, J. Biol. Chem., 175, 385 (1948). L. Goering, J. Eikenberry, G. Koerwer, and C. Lattimer, JACS, 88, 1493 (1974). Dale, D. Dull, and H. S. Mosher, J. Org. Chem., .88, 2543 (1969). Pickard and Kenyon, J. Chem. Soc., 88, 45 (1911). P. A. Levene and H. L. Haller, J. Biol. Chem., 28, 475 (1928). 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 76 von E. Doering and R. Young, JACS, 18, 2997 (1952). G. Khaskin, Chem. Abs., 49, 4505 (1955). H. Krielde and C. A. Burkhard, JACS, 88, 2689 (1947). 0. Kaplan and M. Ciotti, J. Biol. Chem., 221, 823 (1956). Colowick, N. Kaplan, and M. Ciotti, J. Biol. Chem., 197, 447 (1951). Theorell and B. Chance, Acta Chem. Scand., 8, 1127 (1951). Dalziel and F. M. Dickinson, Biochem J., l 0, 34 (1966). Dalziel and F. M. Dickinson, Biochem J., 04, 165 (1967). Theorell in "Advances in Enzymology," Vol. 20, pg. 31, ed. by F. Nord, Interscience, New York (1958). I. Backlin, Acta Chem. Scand., 88, 1279 (1958). Waller, H. Theorell, and J. Sjovall, Arch. Biochem. Biophys., 111, 671 (1965). Hansch, J. Schaeffer, and R. Kerley, J. Biol. Chem., 247, 4703 (1972). I. Branden, H. Eklund, B. Nordstrom, T. Boiwe, G. Soderlund, E. Zeppezauer, I. Ohlsson, and A. Akeson, Proc. Nat. Acad. Sci. USA, 18, 2493 (1973). TTTTTT M((1111(111111)