This dissertation has been 63-66 microfilmed exactly as received

LEYDA, James Perkins, 1935- THE INTERACTIONS OF SOME BARBITURIC ACID DERIVATIVES WITH METALS AND METAL- CONTAINING SUBSTANCES.

The Ohio State University, Ph.D., 1962 Chemistry, pharmaceutical

University Microfilms, Inc., Ann Arbor, Michigan

— ' ~~ — —' ’ •' THE INTERACTIONS OP SOME BARBITURIC ACID DERIVATIVES

WITH METALS AND METAL-CONTAINING SUBSTANCES

DISSERTATION Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By James Perkins Leyda, B. S., M. Sc.

r * *

The Ohio State University 1962

Approved by

Adviser Department of Pharmacy ACKNOWLEDGMENTS

I wish to acknowledge my appreciation for the advice and assistance rendered by Dr. Loyd E. Harris, Professor of Pharmacy, in the development of this dissertation and for his friendship and guidance during my graduate training. I wish to acknowledge Dr. David E. Guttman, Associate Professor of Pharmacy, for his suggestions and consideration during the course of this research project. I wish to thank the National Institutes of Health (National Institute of Mental Health), division of the Department of Health, Education and Welfare, for financial assistance in the form of a pre-doctoral fellowship. To my fellow graduate students gratitude is expressed for their suggestions and assistance in numerous phases of this research problem. Finally, to my wife, Margaret, I am grateful for her encouragement and help during my graduate training.

June 1962

ii CONTENTS Page

STATEMENT OF PROBLEM...... 1

INTRODUCTION ...... 3 Salts or Complexes of Barbituric Acid Derivatives . 3

Mercury ...... 3 Silver ...... 6 G o l d ...... 8 Transition Metals ...... 9 Copper...... 10 Cobalt...... 11 Alkali Metals ...... 12 Alkaline-Earth Metals ...... 12 Mechanism and Site of Action of the Barbiturates . . 14 Interactions Involving Metalloporphyrlns ...... 16 Interactions Involving Macromolecules ...... 18 Chelates of Structurally Similar Compounds..... 22 EXPERIMENTAL...... 23 I. The Stability of Anionic Complexes of Some Barbituric Acid Derivatives and Silver.... 23 Reagents...... 23

A p p a r a t u s ...... 23 Po tentiome trie Procedure ...... 64 Calculation of Stability Constants ...... 24 ill * Page

Titration Results ...... 27 The Effect of Silver Ion on the Molar Absorptivity of Barbital ...... 28 Hie Effect of Complexation with Silver Ion on the Stability of B a r b i t a l ...... 30

Discussion...... 33 II. The Mercuric Complexes of Some Barbituric Acid Derivatives ...... 36

M a t e r i a l s ...... 37 Colorimetric Assay for Mercury ...... 37 Standard Curve from Amobarbital Complex . . . 41 Molar Absorptivity of Mercuric-Phenobarbital C o m p l e x ...... 42

Preparation of Mercuric Complexes ...... 45 Distribution Coefficients of the Barbiturate- Mercuric Complexes...... 45 The Effect of pH on the Barbital-Mercuric C o m p l e x ...... 47 The Effect of Mercuric Ion on the Distribution of Phenobarbital ...... 48

Discussion...... 52 III. The Stability of Some Cobalt, Amin e and Barbiturate Complexes in a Nonaqueous Solvent. 62 Theory...... 62 Materials ...... 66 Spectrophotometrie Procedure ...... 66 Analysis of the Two Step Sequence ...... 67

The Effect of Barbital Concentration on AQba . 71 The Effect of Amine Concentration on A0t>s . . 71 iv Page

Equilibrium Constant from the One Step Reaction...... 73 The Color Intensities of Various Barbiturate Complexes ...... 73 Discussion ...... 80 IV. The Interaction Between Barbiturates and Naturally Occurring Substances Containing Metal I o n s ...... 87 Materials . . . , ...... 87

The Effect of Barbital on the Spectra of Hemln ...... 88

Analytical Method for Pentobarbital...... 93 Dialysis Studies at pH 3*33 101 Dialysis Studies at pH 7 . 4 2 ...... 104 Discussion and Results ...... 104

SUMMARY AND CONCLUSIONS...... 113 BIBLIOGRAPHY ...... 116 AUTOBIOGRAPHY...... 122

v TABLES Table Page 1. Potentiometric Titration of Some Barbiturates . . 28 2. TOie Effect of Sliver Ion on the Absorbance of B a r b i t a l ...... 29 3. The Effect of Silver Ion on the Degradation of B a r b i t a l ...... • • • • 33 4. Standard Curve for the Mercuric Dlthlzone Chelate at 495 rap...... 41 5. Standard Curve for the Mercuric Dlthozone In Chloroform...... 42 6. The Effect of Mercuric Ion on the Molar Absorp­ tivity of Phenobarbital at 239 mp...... 44 7. Distribution Coefficients of the Barbiturate- Mercuric Complexes ...... 46 8. The Effect of Mercuric Ions on the Distribution Coefficient of Barbital at Two pH Values . . . 49 9. The Effect of Mercuric Ions on the Distribution Coefficient of Phenobarbital ...... 53

10. Comparison of the Distribution Coefficients of Some Barbiturates and ftielr Mercuric Compleses. 56 11. Standard Curve for Cobalt Chloride at Two Wave Lengths ...... 67 12. The Effect of Barbital Concentration on the Observed Absorbance a ...... 73 13. The Effect of Amine Concentration Effect on the Observed Absorbance ...... 74 14. Reagents and Spectral Data for Barbital-Cobalt- Amine Complex ...... 76 1 5. The Effects of 5*5- Disubstitution on the Spectral Characteristics of the Barbiturate Complexes . . . a ...... 80 vi Table Page

16. Standard Curves for Hemln in a Borate Buffer pH 9.4 and In 0.01 Normal Sodium Hydroxide . . 90 17. The Effect of Barbital on the Spectra of Heme . . 94 18. Standard Curve for BSA at 240 mp. and 280 mp. . . . 97 19 • Standard Curve for Cupric Ion at 240 mp. and 280 m p ...... 97 20. The Binding of Pentobarbital to 2% BSA at pH 5.33 and 30°C...... 105 21. The Binding of Pentobarbital to 2# BSA Contain­ ing 3.4 x 10 “5 Molar Cupric Ion at pH 5*33 and 30°C 105 22. The Binding of Pentobarbital to 2% BSA Containing 6 .8 x 10~5 Molar Cupric Ion at pH 5*33 and 30°C...... 106 2 3. Die Binding of Pentobarbital to 2$ BSA Containing 6.75 x 10~4 Molar Cupric Ion at pH 5*33 and 30°C...... 106 24. Die Binding of Pentobarbital to 2$ BSA at pH 7.42 and 309c...... 108 2 5. The Binding of Pentobarbital to 2# BSA Containing 5.74 x 1 0 " 5 Molar Cupric Ion at pH 7*42 and 30OC...... 108 26. The Binding of Pentobarbital to 2% BSA Containing 1.15 x 10“^ Molar Cupric Ion at pH 7*42 and 30°C...... 109 2 7. Die Binding of Pentobarbital to 2% BSA Containing 1.69 x 10“4 Molar Cupric Ion at pH7$.42 and

vii FIGURES Figure Page 1. A Typical Titration Curve for B a r b i t a l ..... 25 2. The Effect of Silver Complexation on the Degradation of Barbital ...... 32 3. Visible Spectrum of Mercuric Dithizone Chelate . 39 4. Standard Curve for Mercuric Dithizone Chelate at 495 mp...... 40 5* Standard Curve for Mercuric Dithizone Chelate at 495 mp.* Extracted as Mercuric-Amobarbital Complex...... 43 6. The Effect of Mercuric Ion on the Distribution Coefficient of Barbital at Two pH Values . . . 50 7. The Increase in Barbital Concentration in Chloroform as a Function of Total Mercuric Ion in the System...... 51 8. The Effect of Mercuric Ion on the Distribution Coefficient of Phenobarbital ...... 54 9. The Increase in Phenobarbital Concentration in Chloroform as a Function of Mercuric Ion in Chloroform...... 55 10. Ultraviolet Spectra of Barbital and the Mercuric- Barbital Complex ...... 61 11. Visible Spectra of Amine-Cobalt-Barbiturate Complex and Cobalt Chloride ...... 68

12. Standard Curve for Cobalt Chloride at Two Wave Lengths ...... 69 13. The Effect of Barbital on Observed Absorbance . . 72 14. The Effect of Amine Concentration on Observed Absorbance...... 7$

viii Figure Page 15. The Effect of Varying Cobalt Concentration on the Observed Absorbance of the Barbital Complex...... 77 16. A Comparison of the Observed Absorbances for Various Barbiturate Complexes ...... 79

17* Visible Spectrum for H e m l n ...... 89 18. Standard Curves for Hemln In Borate Buffer pH 9*4 and 0.01 Normal Sodium Hy d r o x i d e ..... 91 19* Spectra of Hemln in the Presence of Pyridine, Barbital and Various Buffer Solutions ..... 92 20. The Spectra of BSA, Cupric Ion and Pentobarbital in a Phosphate Buffer pH 10.9 ...... 95 21. Standard Curve for BSA at 240 mp. and 280 mix . . . 98 22. Standard Curve for Cupric Ion at 240 mix and 280 mix ...... 99 23* Hie Binding of Pentobarbital to 2$ BSA in the Presence and Absence of Cupric Ion at pH 5.33 and 30°C...... 107 24. The Binding of Pentobarbital to 2# BSA In the Presence and Absence of Cupric Ion at pH 7.42 and 30°C...... 110

ix STATEMENT OP PROBLEM

Barbiturates comprise an Important class of hypnotics and sedatives. The site of action Is believed to be the cortex of the brain and possibly the thalamic portion of the diencephalon (1-2). The mechanism of action of these com­ pounds has never been elucidated; however, the cytochrome system which is Involved in cellular respiration is believed to be effected. The recent awareness of the essential roles which metal ions play in biological reactions has stimulated this investigation. Many enzymes are metal containing substances and numerous biochemical reactions have been shown to require metal ions as catalysts. Interactions between drug molecules and metal ions present in the body, either bound or free, may interfere with the normal physiological responses resulting in a distinct pharmacological action. It is the aim of this investigation to study the interactions between a series of barbituric acid derivatives and metal ions and compounds containing metals. More specifically for this investigation the interactions are divided in four distinct classes. 1. Barbiturate-silver ion interaction in alkaline solution. 1 2. Barbiturate-mercuric ion interaction. 3. Formation of highly colored complexes between barbiturates, amines and transition metals. 4. An investigation of a possible interaction between the barbiturate anion and metal containing compounds present in biological systems, such as metalloporphyrin derivatives or metalloprotelns. INTRODUCTION

The fact that barbituric acid derivatives form salts and/or complexes with many types of metal ions has been recognized for many years. The preparation of these salts, or complexes has Involved different families of metals, as the alkali, alkaline-earth, transition, and heavy metals. Few of the barbiturate-metal compounds are used thera­ peutically because of their insolubility, instability in solution and possible toxicity. The Instability of alkali metal salts in aqueous solution has long been recognized as a problem in the formulation of desirable dosage forms. The major application of barbiturate metal interactions has been in the qualitative and quantitative determination of barbi­ turates in solution, biological systems and pharmaceutical preparations.

Salts or Complexes of Barbituric Acid Derivatives

Mercury

Hie isolation of the mercury "salts" of barbital and phenobarbital was first accomplished, in 1914, by Pio Lami (3)* As a result of this early investigation several analytical procedures have been designed Involving the water insoluble precipitate which is formed by the addition of 4 mercuric salts or organic mercurials to solutions of barbiturates. Bolle and Mirimanoff (4) described a qualitative method for the identification of heterocyclic nitrogen con­ taining compounds. This method Involves the addition of a dilute solution of phenylmercurlborate to a solution contain­ ing the heterocyclic compound, the presence of a yellow color signifies a positive test. This color formation has been shown to appear with barbital and phenobarbital but Is not specific for barbiturates. A quantitative volumetric procedure for barbituric acid derivatives In solution was proposed by Pedley (5)* An excess of mercuric perchlorate was added to the aqueous barbiturate solution. The mercuric complex, which precip­ itates Immediately, was removed by filtration and the excess mercuric Ion present In the filtrate was titrated with a standardized ammonium thiocyanate solution using ferric alum as the Indicator. The 5»5- disubstituted derivatives formed a 1:1 complex while the N-substltuted derivatives gave 2:1 compounds. Kurpiel et al. (6) have recently reported a chelometric titration procedure using the mercuric complex to quantita­ tively remove the barbiturate from solution. The barbi­ turate was precipitated from solution with mercuric acetate, the complex removed by filtration and then dissolved In a standardized sodium versenate (EDTA) solution. Zinc sulfate solution was employed as the tltrant for the excess sodium versenate using erlochrome black as the Indicator. Hie authors gave the following formulas for the mercuric com­ plexes of the 5,5- disubstituted derivates (I) and the N-substituted compounds (II).

so

II Mercuric acetate solution has been successfully used by Cohen and Lordl (?) as the tltrant in the amperometric and potentiometrie titration of phenobarbital elixir. In recent publications by BJorling et al. (8-9) a new colorimetric method has been presented for the determination of trace amounts of barbiturates. The mercury-barbiturate complex was formed and extracted into chloroform or some other suitable organic solvent. The chloroformic solution was then treated with dithizone, a strong chelating agent, which removes the mercury from the barbiturate complex and gives a highly colored solution. The resulting red solution was assayed colorlmetrically. The method has been shown to be sensitive to 30 micrograms of barbiturate. Several other authors have described similar methods for detecting micro quantities of barbituric acid derivatives in biological fluids (10-11). Kalinowski and Baran (12) have described a nephelometric method for the assay of phenobarbital. The dilute barbiturate solution was tasted with mercuric perchlorate and the resulting turbidity was measured and compared to that of a standard curve. The authors reported that this method was accurate within 4 percent. A compound containing the barbital moiety attached to an organic mercurial has been used therapeutically as a mercurial diuretic. The structure for Merbaphen (III) shown below, illustrates an Interaction of an organic mercurial and the oxygen in the 2-position of the barbiturate ring.

o -ch2-c -o “ Na

III

Silver

Budde (13) first described an argentometrlc titration procedure for the assay of barbiturates. Using a solution of silver nitrate as the tltrant and a sodium carbonate solution as the solvent the titration proceeded to the endpoint which was the initial appearance of a permanent 7 white turbidity. Hie turbidity was caused by the formation of the Insoluble dlsllver barbiturate. At the endpoint, one mole of tltrant was equal to one mole of barbiturate. Danlelsson modified this procedure using a potassium metaborate solution as the solvent and potassium chromate as the indicator (14).

Furst (13) has recently proposed a procedure using an excess of silver salt to precipitate the barbiturate, remov­ ing the precipitate by filtration and then titrating the excess silver ions present In the filtrate with an ammonium thlocyanate solution using ferric alum as the indicator. Hie potentiometric titration of barbiturates in alkaline solution, and in pyridine, with silver nitrate solution employing a sllver-calomel electrode system has been described by several Investigators (16-20). The magnitude of the stability constants for the monosilver anionic barbiturate complexes of three barbituric acid derivatives has been reported, as well as the suggested structures for both the monosilver and dlsllver complexes. Hie following structures (IV and V) were proposed by Perelman (1 6 ):

o

o 0

IV V Structures (VI and VII) were proposed by Poethke and Furst (18) as follows:

so

VI A colorimetric assay Involving a reaction between a barbiturate, ammonium thlosulfate and silver nitrate has been proposed by Danlelsson (21). The various barbiturates gave different colors and were not reproducible. This method also was sensitive to only 2 milligrams of material, thus being impractical even for qualitative testing.

Gold Barbital imldo auric acid was prepared by Kharasch and Isbell (22) by Isolating the white precipitate formed when auric hydroxide, barbital and hydrochloric acid were digested together in an aqueous solution for 48 hours. The authors proposed the following formula for the barbital imldo auric acid. 9

H + • H»0

VIII Transition Metals The coordination of the transition metals with many different types of ligands has been extensively studied within the last twenty years. At one time it was believed that only these metals had the ability to form chelates and complexes with compounds having atoms with an excess of electrons. However, as more accurate methods were developed in the field of coordination chemistry other metals were shown to form complexes. Because of the physical structure of the transition metals, such as the ionic radius and the availability of the electronic orbitals, these metals usually form coordination compounds which are more stable than for other types of metals.

The first appearance of a series of transition metal- barblturate complexes was in 1937 when Cematescu and Vascautanu (23) Isolated the copper, chromium, nickel, manganese and cadmium complexes. These compounds were usually highly colored substances which were insoluble in water and other common solvents. The method of preparation and the empirical formulas of these compounds were listed, but few chemical and physical properties were presented. Copper. A Dutch chemist, Zwikker (24), in 1931* first prepared a crystalline substance from a solution containing barbital, pyridine, and copper sulfate having the formula (barbital) 2CU(;pyr idine)2 • The British Pharmacopeia employs Zwikker*s findings as an official identification test for all 5*5- disubstituted barbituric acid derivatives (25). Recently, Pialkov and Rapaport (26) determined the relative stability and proposed the possible structures of the complex for the disubstituted barbiturates. The struc­ tures (IX and X) are shown below. The stability of the

IX X complex was shown to be greater than that of the

Cu(pyridine)2 which is 1 .2 5 x 10? but ammonia will displace the barbiturate and pyridine, forming the copper ammlne, CutNH^Jip (Ka 5 x 1012). Thus the authors conclude that the barbiturate copper and pyridine complex stability constant is in the range of 10^ to 1012. Levi and Hubley also have applied this complex formation reaction to prepare and isolate the complexes of a series of clinically important barbiturates (27). The infrared spectra of these compounds have been proposed as a tool for the toxicologist and forensic chemist in the micro­ chemical identification and characterization of the parent barbiturate. Also in forensic practice the distinctive crystal formation of these compounds, as well as the ferrous complexes, are valuable for the proof of barbiturate poisoning (28). Numerous other authors have modified the original procedure by substituting various amines for the pyridine, however, the color of the complexes remain the same violet- blue for barbiturates and green for thiobarbiturates (29-3 2) Thus it is believed that the color formation is character­ istic of the malonylurea portion of the molecule and that the substituents in the 5 position have little effect, but they must be present. A unique intermediate was usediln the preparation of N-alkylated barbiturates by Halpern and Jones (33)* To an aqueous solution of the sodium barbiturate a copper salt was added and then alkylation accomplished by the addition of the appropriate alkyl halide. Excellent yields were reported. Cobalt. The highly colored complexes which form by the interaction between cobalt salts, barbiturates and bases 12 usually amine, were initially studied by Parri Zwikker (24), and Bodendorf (35)* The formulas of these complexes are similar to the copper complexes previously described. Koppanyi et al. (36-38) have investigated extensively the use of the cobalt complexes in the analysis of solutions containing barbiturates. This quantitative procedure has been criticized by numerous authors due to the instability of the color formation, and the subsequent non-reproducibil­ ity of the method. Nevertheless, many references are found where it has been continuously used as a qualitative and quantitative method in the analysis of barbiturates in urine, blood, and pharmaceutical dosage forms (39-5 0).

Alkali Metals The alkali salts of barbiturates have received much attention due to their low toxicity and water solubility. The procedures for the preparation of these salts are found mostly in the patent literature.

Alkaline-Earth Metals

Magnesium and calcium have been used in preparing salts of some barbiturates, primarily barbital. In 1921 a British patent was Issued for the calcium and magnesium salts of the barbituric acid derivatives (31). These therapeutically active compounds were obtained by digesting a hot saturated solution of the acid with the theoretical 13 quantity of the appropriate metal hydroxide. More recently, Berggardh (52) prepared the same compounds by the addition of a metal chloride solution to a solution of the sodium barbiturate. The precipitate was allowed to settle and then removed by filtration. Hie stability of the calcium and strontium complexes of some organic acids have been determined by an Ion exchange method utilizing a barbital buffer. It was shown that the buffer component, barbital, did form a weak com­ plex ion with the metals (53)* Chauteau (5*0 has studied the Interactions between copper, calcium, magnesium and barbital In alkaline solution. Hie investigator assumed that the metals were coordinated between the nitrogen atoms, and by infrared analysis showed that the calcium and magnesium compounds were similarly bound, but the copper complex spectra were different between 6 and 8 p. A new analytical procedure for the quantitative determination of tetracycline in biological fluids has recently been developed at the National Cancer Institute. Hie method involves the formation of a fluorescent chelate containing barbital, a calcium salt and tetracycline. Hie calcium salt and barbitalwre added to the fluid contalnirg the tetracycline, the chelate was extracted with an lmmlsslble solvent and the fluorescence was measured in an appropriate manner. The structure of the fluorescent chelate (XI) is shown as reported by Kohn (55)•

Mechanism and Site of Action of the Barbiturates The mechanism of action of the barbituric acid derivatives has been the subject of many investigations, however, the problem is still not answered. 'Rie cytochrome system is believed to be the site of action on the cellular level.

In a series of publications Masserman has attempted to determine the portion of the cat brain which was responsible for the actions caused by sodium amytal. As a result of these works, Masserman has concluded that possibly the thalamic portion of the diencephalon or the brain cortex are sites of action of amytal (1-2). Quastel and Wheatley (56), in 1932, first showed that barbiturates inhibit cellular respiration, thus suggesting 15 that the cytochrome system would be directly Involved. Since this Initial report, numerous Investigators have described the effect of barbiturates on the various cytochrome enzymes. In studies on the site of action of barbiturates In brain metabolism, Orelg (57) determined that Nembutal (pentobarbital) alters the oxidation cf carbohydrates. From the experimental results It was suggested that the compound may act by binding the reduced flavoprotein with cytochrome b (or other intermediates) and that the affinity of the drug for this complex is greater than for the succinic dehydrogenase-cytochrome b complex which Is not effected by low concentrations of narcotics. In a later study using phenobarbital as the narcotic, Greig (5 8) illustrated that the barbiturate does not act directly on cytochrome c but before that step In the biological oxida­ tion, namely by action between the flavoprotein and cytochrome b reaction steps.

The inhibition of the regeneration of high energy phosphate bonds by pentobarbital has been shown by Eiler and McEwen (59) to be only to the extent that it interferes with utilization of oxygen in rat brain homogenates.

Bain (6 0) has found indirect evidence that cytochrome c reductase is the site of action for barbiturates. Barbera (61) has reported that the action of pentothal is 16 modified when it is administered in the presence of cytochrome c and suggests that the narcotic action of pentothal may be to block the nerve cell enzymes, primarily the cytochrome oxidase-cytochrome c system. Hie well known chelating agent, calcium disodium ethylenediaminetetraacetate has been shown to potentiate barbiturate sedation in rabbits (62). Further evidence to support the possibility of metal ion mediation was described by Bonner (6 3) in investigations on the activation of the cytochrome system with EDTA. The EDTA activates the succinic dehydrogenase-cytochrome system while the barbiturates are known to inhibit the same enzyme system.

Interactions Involving Metalloporphyrins The cytochrome system has long been recognized as the pathway of cellular respiration. Metalloporphyrin proteins are the important enzymes responsible for electron transfers through the cytochrome system. Hie prosthetic groups for these enzymes are porphine derivatives coordinated with a transition metal ion, usually iron or copper. Hiese planar chelates possess the unique property which allows the co­ ordinated metal ion to reverslbly undergo oxidation and reduction while remaining in the stable, coordinated form. Electron transfer can be inhibited or prevented when com­ pounds such as amines or other bases Interact with the metal ions. Chromogens are formed when heme or other ferrous complexes of porphyrins react with organic bases. The literature abounds with reports describing the stability of chromggens which contain cyanide anion (64-69) and carbon monoxide (69-70). The lethal effects of carbon monoxide and cyanide are due to their ability to combine with iron containing prosthetic groups and form the appropriate chromogens. Carbon monoxide reacts with hemoglobin to form carboxyhemoglobin which Is much more stable than the cor­ responding oxygen hemoglobin complex. Cyanide does not effect the oxygen-carrying capacity of hemoglobin. The ensyme cytochrome oxidase is attacked by the cyanide anion thus preventing cellular oxidation. Many other chromogens have been investigated which contain nicotine (71-72), pyridine (67-69, 71-74), alpha- picoline (72), caffeine and related xanthines (70), imidazole derivatives (7 3)* histamine and antihistaminic compounds (7 5)* The stability constants characterizing these inter­ actions have been determined qualitatively and more recently, quantitatively by potentiometric titrations using reducing and oxidizing agents as the tltrant (72,7 6), spectra changes under various conditions (64-71* 74), and polarography (77), Barron (76) has also suggested the possibility that buffer components, such as phosphates, borates and barbital, may 18 form complexes with ferriheme hydroxide due to solubility and oxidation-reduction potential changes. Domonkos and Huszak (7 5) have investigated the interaction between hemin and histamine and pyribenzamine as a possible elucidation of the mechanism of action of these therapeutically active amines. The catalase-like activity of hematin upon hydrogen peroxide was shown to increase in the presence of histamine but was reduced when the antihistamine was added to the system. The authors suggest that the inhibition of cytochrome activity in the brain is caused by a complex formation between histamine and the prosthetic group of the cytochrome enzymes.

Interactions Involving Macromolecules The association of small molecules with proteins has been investigated extensively as a key to the understanding of various biochemical reactions. Many dissimilar drug molecules are known to enter combinations with proteins and from these data it has been speculated about the role of these interactions in causing changes in the effective concentration of medicinal agent in vivo, the role of the proteins in transporting of drugs and the influence of protein binding on the excretion of small molecules. It is apparent that these drug-protein Interactions may be of pharmacological, physiological or biochemical significance. 19 Penicillins are known to decrease In activity In the presence of plasma proteins and It has been suggested that this decrease In activity was caused by the formation of a pharmacologically Inert protein-penicillin complex (8 9)- A review of the drug-protein interaction literature previous to 19^9 was presented by Goldstein (8 9) and a more recent list of references has been compiled by Elchman in I960 (90).

Goldbaum and Smith (91) have investigated the binding of clinically important barbiturates with bovine serum albumin employing the ultrafiltration technique and attempted to correlate the fraction bound to other physical properties of these sedatives. Dox (92), previously, illustrated that the pharmacological action increases with an Increase in the length of the alkyl side chains while the duration of action decreases. In a similar manner, the binding was shown to increase with an increase in the length of the alkyl chain. Barbiturates with short duration of action were bound to the greatest extent.

Butler (93) has studied the binding of phenobarbltal to bovine serum albumin in an attempt to correlate the effect of systemic pH changes on drug action, metabolism and excretion. Phenobarbltal binding to bovine serum albumin was shown to increase slightly with an Increase in pH in a maximum at about pH of 8. 20

Serum albumin has been employed in most investiga­ tions due to its ability to associate with many dissimilar small molecules and it is the major fraction of plasma proteins (ca. 60#). Albumin from human and bovine plasma have slight differences, however, their physical properties are similar.

A general review of small raolecule-protein binding has been prepared by Klotz (94) in 1953 and more recently he published a comprehensive article discussing the techniques involved in investigating the metal-protein complexes (95)* Many metal ions are known to associate with albumin in vitro and several functional proteins are present in biological systems as the metal complex. These facts have led to speculation concerning the necessity for these metallo-proteins in biochemical reactions. The role of metals in the various peptidase systems appears to be in forming the complex between the proteins and substrates, thus the metal does not act as an activator but linked in true chemical combination (96). There are other examples of the requirements of metal ions in biochemical reactions. Klotz and Loh Ming (97) first quantitatively investi­ gated the mediating effect of metal ions on the binding of small molecules to protein. The small molecule, an aniline dye, was known not to bind with proteins but was a known chelating agent. In the presence of various transition 21 metals the dye was shown to Interact with the protein. Equilibrium dialysis as well as spectral changes were the techniques used to study the mediating effect of these metals. Hie physiological role of histamine In allergy has been postulated to Involve the binding of histamine to a protein. In vitro experiments have failed to show the Interaction between various purified animal proteins and histamine. Andrews and Lyons succeeded in forming a com­ plex between cupric bovine plasma albumin and histamine (9 8). This metallo-proteln also formed a stable complex with a commercially available antihistamine. The authors postulated that the histamine was bound to the protein through a cupric ion bridge.

Recently, Kohn (99) proposed that the tetracyclines may exert some of their biological effects by complexing with macromolecules through metal Ions. Several of the characteristic biochemical effects of these antibiotics are known to depend on the presence of divalent metal Ions which Albert suggests probably arise by the formation of the tetracycline chelate. Kohn Investigated the interactions of tetracyclines with deoxyribonucleic acid (DNA) and with serum albumin In the presence and absence of divalent cations. Only a slight interaction was noted with the macromolecules In the absence of the metal Ions, however, binding to ONA was apparent In the presence of zinc, calcium, 22

magnesium and manganese but the binding with bovine serum albumin occurred only In the presence of zinc. In all Interactions In which metal Ions have been shown to be a mediating agent, the small molecule has possessed the common property of being able to form coordi­ nation compounds with metal Ions.

Chelates of Structurally Similar Compounds It has been noted that compounds structurally similar to the barbiturates have complexing or chelating tendencies

with metal Ions. Poye et al. (78-7 9) have shown that alloxan (XII) and riboflavin form water Insoluble chelates with some of the transition metals. It has been suggested that alloxan's diabetogenic action Is caused by altering the availability of zinc to the Islet cells. Also some

hypoglycemic agents such as fciguanldes (8 0), sulfonylureas (8 1 ) and related compounds have the ability to bind metal Ions. Ureides, aliphatic substituted, have been protected

from hydrolysis in the presence of hydrogen bonding, termed by the author as chelation (82). Beta-dlketones have also been investigated as chelating agents for the transition elements (83)•

o

XII EXPERIMENTAL

I. The Stability of Anionic Complexes of Some Barbituric Acid derivatives ana Silver The alms of this Investigation were to determine the stabilities of the monosilver barbiturate anionic complexes of a series of nine barbiturates and the effect of the silver complex on the chemical stability of barbital.

Reagents

The barbiturates used In this investigation were recrystallized from ethanol-water mixtures with the exception of butethal which was recrystallized from an ether- petroleum ether (30°-60° B.P.) mixture. Sodium carbonate monohydrate and potassium nitrate were of analytical reagent grade and were not further purified. Hie silver nitrate solution was obtained from the Ohio State University Reagent Laboratory and was standardized against ammonium thlocyanate using acidified ferric alum as the indicator.

Apparatus

A 5 ml. buret graduated in 0 .1 ml.; magnetic stirrer; Rubicon Model WBW and a Leeds-Northrup Potentiometer with a General Electric galvanometer as thfe null Instrument;

23 24 silver electrode, prepared from 24 gauge silver wire: calomel electrode, containing a saturated potassium chloride solu­ tion; Beckman DU Spectrophotometer; and Beckman pH meter, Model OS.

Potentlometrlc Procedure An accurately weighed sample (between 15 and 80 mg.) was dissolved in a sufficient amount of an alkaline carbonate solution (pH 11.0, 0.16 Molar sodium carbonate and 0.10 Molar potassium nitrate) to make 100 ml. A 25 ml. aliquot

of the sample solution was placed in a 180 ml. titration vessel and titrated with the standardized 0.1 Normal silver nitrate solution. After each addition of tltrant the E.M.F. was measured, using a silver electrode as Indicator electrode and a calomel reference electrode. The titration was con­ tinued until the permanent turbidity appeared, which tes the visual endpoint described by Budde. A plot of Amv./Aml. versus ml. of titrant determined the endpoint used in the calculation of titration efficiency and stability constants. At the endpoint the average Amv./Aml. was about 300. A typical titration curve for barbital is shown in Figure 1.

Calculation of Stability Constants Die reactions which lead to the formation of the monosilver barbiturate anionic complex in alkaline solution is shown by the following sequence:

BarbH + Na2 0 0 3 5 ^ = 2 Barb” Na+ + NaHCO^ Barb" Na+ + AgN0^+ Na2C0^^: £ [BarbAg]“ Na+ + NaNO^ + NaHCO^ 250

200

150 A MVS A MLS 100

50

1.5 2*0 3*0 3.5 4.0

Milliliters of Silver Nitrate

Fig* 1* A typical titration curve for barbital. Hie equilibrium constant (Instability constant) which defines the reaction is

^ [Barb~][Ag+] inB [BarbAg"] The stability constant of this anionic complex is the reciprocal of the Klns»

[BarbAg"] Ks " [Barb"][Ag+ ]

Kolthoff (84) has given an equation for the calculation of free silver ion concentration by measuring the potential using a silver electrode and calomel electrode system.

0.7991 “Eg - Ec * 0.0591 Where Ec is the E.M.F. of the calomel electrode which is 0.2444 volt when the electrode contains a saturated solution of potassium chloride at 25°C. (8 5). Es is the measured E.M.F. of the system. The concentration of anionic complex and free barbiturate was determined by the procedure described by Perelman (16). Molar concentration of the free barbiturate was calculated by the equation „ (c - a)*N molar a + b . The molar concentration of the anionic complex was calculated in a similar manner, by the equation

r. _ a x N ''molar a + b . 27 Where a la the number of milliliters of titrant added: N la the normality of titrant (silver nitrate): b designates the number of milliliters in the original sample and c is the number of milliliters of titrant at the endpoint, determined by the maximum Amv./Aml. Due to the low concentration of free silver ion in the solution the activity coefficient was assumed to be unity and molar concentrations were used in all calculations.

Titration Results ftie data presented in Table 1 were obtained from the potentiometric titration procedure. The per cent recovery (titration efficiency) and stability data are the average values from a series of determinations plus the standard deviation from the mean. Perelman (16-17) calculated the instability constants for diallylbarbituric acid, K - 1.1 x 10~7(KS * 9 .1 x 106 at 20° C.) and barbital, Klns - 4.7 x 10“8(K3 - 2.1 x 107). Poethke and Furst (18) have reported the instability con­ stants for barbital and phenobarbltal respectively, K^ns « 4.0 x 10"7(Ks - 2.6 x 106 ) and 2.7 x 10“7(KS - 3.7 x 106 ). The data presented in Table 1 agrees with the limited number of values given by Poethke and Furst. A general trend is noted between the stability con­ stants of these anionic complexes and the pKa values of the parent acids. Due to the relatively narrow range which the pKa values span for the barbituric acids, it would be 28 expected that the stability constants would be of similar magnitude. The stability of the complexes are the same magnitude and slightly Increase with the decrease in the pKa of the parent acid.

TABLE 1 POTENTIOMETRIC TITRATION OP SOME BARBITURATES

Per Cent Compound Samples Recovery 4 mv.Ave. Ks x 10”6 pKa Barbital 18 1 0 1 .1 + 1 .7 170.7 3.46 + 0.48 8 .0 6 Phenobarbltal 18 98.5 + 1.5 163.4 5.03 + 1.03 7.54 Diallyl- barbituric acid 18 9 8 .9 + 1.3 166.6 4.60 + 0.55 7.88 Aprobarbltal 12 1 0 0 .8 + 1 .9 I6 9.O 3 .8 1 + 0.48 $.09

Vinbarbital 18 100.7 + 1-7 I6 5.O 3 .8 5 + 0.74 7.72

Probarbital 12 100.3 + 2.4 1 7 0 .7 3.39 ± 0.75 8.17 Butethal 12 98.2 + 3.4 164.8 4.09 + 1.04 8.10

Butabarbital 12 100.5 + 2.3 171.5 3.46 + 0.59 8.16 Amobarbltal 6 9 8 .6 + 3*1 1 6 5.6 3 .9 6 + 0.20 8.02

The Effect of Silver Ion on the Molar Absorptivity of karbltal Danlelsson (14) reported that the molar absorptivity of the silver-barbital anion at the pH of 10.9 appears to be 18,820 when determined at 2381mx. Previously, the molar absorptivity of barbital, at 239d*ji and the pH 10.9, has been shown to be 10.2 x 10^ (86). In order to assay barbital in 29 the presence of sliver Ion and silver-barbltal complex the absorbances of these various species were determined. TSie absorbances of 3*89 x 10“^ Molar solutions of barbital (total) were measured by a Beckman DU Spectro­ photometer at 239n|x. The solutions contained varying amounts of silver Ions and were buffered at pH 10.9* The data are presented In Table 2.

TABLE 2 THE EFFECT OF SILVER ION ON THE ABSORBANCE OF BARBITAL

Barbital <- Silver Complex Free Barbital . Sample Molarity x 105 MolaritykL05 Molarity x 10^ Aobs

1 3-89 0 3.89 0.398

2 3.89 0 3.89 0.397 3 3-89 0.841 3.05 0.400 4 3.89 0.841 3.05 0.395 5 3-89 1.680 2.21 0.394 6 3.89 1.680 2.21 0.395 7 3-89 2.520 1.37 0.395 8 3.89 2 .5 2 0 1.37 0.394

9 3-89 3.360 0.53 0.398 10 3-98 3.360 0.53 0.388

The following equation was used in the calculation of the molar absorptivity ( A ^ of each species present in the solution. It was assumed that all silver ions added to 30 the system Interact completely with the enollzed barbital.

Aobs " ^(barb)* (Concentration of free barbital) + ^(barbAg)•(Concentration of complex) The calculated absorbance (O.3 9 7) for this concentration was obtained from a previously derived equation.

°n®lar * °'021 * 10-5 +.(9-7511 * 10-5 • Aobs) The average molar absorptivity of the complex is 10.1 x 1 CW, which correlated within 99«0# of the literature value (86) for the molar absorptivity of the enollzed barbital. Thus it appears that the spectrophotometrlc assay is efficient for determining the total amount of barbital in the system which contains silver ions.

The Effect of Complexation with Silver Ion on the Stability ot Barbital The decomposition of aliphatic substituted ureides in alkali has been shown to be retarded when intermolecular hydrogen bonding is present. Werner (82) described this as chelate formation. Since barbiturates are cyclic ureides which hydrolize in alkaline solution, silver ions are believed to form a silver-barblturate anionic complex which resists hydrolysis. A 0.2 Molar solution of sodium carbonate (pH 11.0) was employed as the alkaline media. Solutions were prepared to contain approximately 4.0 x 10”3moles/liter of barbital and varying amounts of silver nitrate. Samples were Sealed in 2 ml. "Neutraglas" ampuls and placed in an oil bath at 31 42.5°C. At various time intervals samples were removed from the constant temperature bath and immediately frozen in a chloroform-dry ice mixture. The samples were then assayed spectrophotometrlcally as previously described. The degradation rate (apparent) is defined by the following equation.

kapp “ Nbarbkbarb + NbarbAgkbarbAg

Where Nbarb is the mole fraction of free barbital, NbarbAg the mole fraction of silver-barbital complex (silver mde fraction), kbarb the pseudo-first order rate constant of free barbital and kbarbAg the psudo-first order rate constant of the complex. The results are presented in Table 3, and Figure 2 illustrates the relationship between the pseudo-first order reaction rate constant and the concentration of anionic sllver-barbital complex. The reaction rate constants for the control samples and the samples containing known amounts of silver ion agree within 97*7# of the calculated rate of hydrolysis (assuming that only the free barbital degrades). Calculated data (solid line) had a slope -5*43 x 10~5 and the experimental data (broken line) had a slope -5*56 x 10"5, as determined by the method of least squares. 32

5*0

4.0

9*0

2*0

1*0

0 0 0 . 2 0.4 0.6 0*8 Mole Fraotion of Complex

Fig. 2# The effeot of silver oomplexatlon on the degradation of barbital. 33 TABLE 3 THE EFFECT OF SILVER ION ON THE DEGRADATION OF BARBITAL

Sample Mole Fraction kapp x 3-°^ kbarb x Silver Complex hour"-1 (calc.) 1 0 5.42 5.42

2 0 5.49 5.49

3 0.268 3.78 3.97 4 0.576 1.93 2.31 5 0 .8 3 6 0.82 0.89

Discussion Perelman (1 6) has described the reaction sequence for the formation of the sliver complexes of barbiturates. In an alkaline solution, barbiturates react with silver Ions In a step-wise fashion which Is shown by the following formulas.

IV V Gautier et al. (19) have followed these reactions In pyridine, ftiey have shown that titration of the barbituric acid liberated two protons and that only one proton Is liberated when the sodium salts are employed. These results support the fact that the reaction occurs step-wise.

Poethke and Furst (Id) use the following chelate structures formulas to represent the anionic complex VI and the water Insoluble dlsllver barbiturate (VII) structures.

VIVII Goyan et al. (8 7) have written the resonance forms of the anionic species In the following manner.

XIV XIII

Spectral data have supported the view that species XIII Is not the resonance form which Is attacked by the silver Ion.

The enolate forms which are present In alkaline solution are the species responsible for the Increase In absorbance and are represented by structures XIV and XV. The writer has shown that the sllver-barbltal complex does not appear to 35 alter the absorptivity of the barbiturate. Thus the silver ion does not change the electron distribution in the enollzed form. Qoyah et al. (8 7) have suggested that barbital degra­ dation in ammonia buffer is similar to ammonolysls of an ester and that hydrolysis of the ionized species would prob­ ably be at the 4 (or 6) position. By virtue of infrared spectra of the p-nitrobenzyl derivatives of a series of barbiturates, Chatten and Levy (88) have proposed that the mechanism of the reaction Involves enollzatlon of the carbonyl at position 4 (or 6). The silver-barbiturate anionic complex has been shown to be stable in alkaline solution. The silver ion, when attacking the barbiturate anion, liberates a proton, thus the nitrogen atom which is not involved in the enolization appears to be attacked by the silver ion, as shown by structures XVI and IV.

XVI IV These formulas do hot appear to explain the stability of the complex to alkaline hydrolysis, especially if the 36

4 (or 6) carbonyl is the position vulnerable to hydroxyl ion attack. Barbiturates which have substituents on the nitrogen atom, i.e., a methyl group, are known to hydrolize in the same manner as the unsubstituted compounds. Hie monosilver barbiturate complexes are believed to be stabllzed by the formation of a chelate structure previously described by Poethke and Furst. Hiis chelate is described by structure

"O

XVII

II. The Mercuric Coarolexes of Some Barbituric ~ Acid Derivatives" One of the well-known incompatabillties of the barbituric acid derivatives in alkaline solution is precipi­ tation caused by the presence of heavy metal ions, especially mercuric ions. This incompatablllty has been the basis of several analytical procedures involving gravimetric, tltrimetrlc and spectrophotmetric methods. The mercuric complex is slightly soluble in water and also has a limited solubility in common organic solvents. 37 It was the aim of this Investigation to study the effect of mercuric Ions on the distribution coefficients of a series of barbiturates.

Materials The barbiturates were purified by recrystallization from alcohol-water mixtures as previously described. Diphenylthlocarbazone (dithizone), potassium chloride and mercuric acetate were of reagent grade and were not further purified. ©le chloroform was of reagent grade or redistilled after being neutralized. Trls (hydroxymethyl) amino methane (trls) was recrystallized by the method described by

Fo b sum (100).

Colorimetric Assay for Mercury Several Investigators have analyzed for the mercuric ion In solution both qualitatively and quantitatively with dithizone in an organic solvent. This strong chelating agent has the ability to form a highly colored chelate with the mercuric ion and the color intensity of this complex Is proportional to the Ion concentration when the dithizone Is maintained In excess. Chloroform was chosen as the ozganlc solvent to be used In preparing the standard curve. Dithizone was dis­ solved in chloroform (13 mg./lOO ml.). This deep blue reagent solution was prepared, stored in a refrigerator and protected from light. A solution of mercuric acetate was prepared by dissolving an appropriate amount of mercuric acetate in distilled water (ca. 300 mg./lOO ml.) containing a small amount of acetic acid. The acetic acid was added to prevent hydrolysis of the mercuric salt to the hydroxide.

Various aliquots of the mercuric acetate solution were extracted twice by the dithizone solution. The extracts were combined from each sample and diluted with chloroform to the desired volume. A Perkin-Elmer Spectracord was used to determine the wave length of maximum absorption for the mercury-dithizone chelate when.using a blank of the dithizone solution. Hie wave length of maximum absorbance was found to be 495

(Figure 3)- Hie absorbance values used to construct the standard curve were determined at 495 mu with a Beckman DU Spectro­ photometer using a tungsten lamp as the light source and with the dithizone solution as solvent blank. Table 4 and Figure 4 illustrate that the color intensity of the mercury-dithizone chelate does follow Beer's Law. Hie data reported in Table 4 was obtained from duplicate runs. An equation for the straight line was calculated by the method of least squares. This equation was used to calculate the molar concentration of mercuric ion from the spectral data.

Molar Mercuric Ion - A°bB— T., Q,*?Q5 4.564 x 104 0.5

0.4 8. IS x 1 0 “ ° Molar

obs

O.S

S.20 x 10"® Molar

450 475 500 525 550 Wavelength, Mlllimiorons

Tig* 3. Visible Speotrum of Merourio Dithizone Chelate 0.8

0.7

0.6

Obs 0 # 5

0.5

0.2

0 0.4 0 . 8 1.2 Merourio Concentration, Molar x 10®

Fig. 4. Standard curve for merourio dithizone ohelate at 495 mu. 41

TABLE 4 STANDARD CURVE FOR THE MERCURIC-DITOIZONE CHELATE AT 495 W

Mercuric Ion Sample Molar x 10© Aobs

1 1.36 0.050 2 2.70 0.144

3 4.05 0.177 4 5.40 0.250

5 6.25 0.311 6 8.10 0.378

7 9.45 0.431 8 10.80 0.503

9 12.15 0.539 10 13.50 0.633 11 16.20 0.740

Standard Curve from Amobarbltal Complex A mercuric acetate solution of known concentration (ca. 300 mg/100 ml.) was prepared by dissolving mercuric acetate In distilled water. Various aliquots of this solu­ tion were added to amobarbltal In "trls" buffer solutions and repeatedly extracted with 10 mis. of chloroform. The extracts were combined with each sample, dithizone reagent solution added and diluted to volume with chloroform. As previously described, the absorbances were determined at 42 495 np. Table 5 and Figure 5 show the standard curve obtained for the mercuric ion when extracted as the amobarbital complex.

TABLE 5 STANDARD CURVE FOR MERCURIC-DITHIZONE CHELATE IN CHLOROFORM

Sample Mercuric Ion Aobs Molar x 10? 495 mix 1 0 0 2 0.26 0.101

3 0.52 0.210 4 0 .7 8 0.318

5 1.04 0.427

6 1.30 0.535 7 1.56 0.644

The equation for the line in Figure 5 which can be used in the calculation of molar concentration of mercuric ion is shown below.

...... _ Aobs + 0.007 Molat* Mercuric Ion ■ --- ;-- ;----- 4174.7

Molar Absorptivity of Mercurlc- Phenobarbltal Complex

Hie molar absorptivity of phenobarbltal at pH 10.9 •a and 239 mi* has been shown to be 11.0 x 10 (86). In order to assay for phenobarbltal in the presence of mercuric ion and mercuric-phenobarbital complex, the molar absorptlvltles 0.8

0.6

0*2

0

0 0.4 0*8 1*2 1*6 Molar Merourio Ion z 10s Fig. 5. Standard ourve for merourio dithizone ehelate at 495 mu, extraoted as amobarbital complex. 44 of these species were determined In a manner similar to the

sliver anionic complex of barbital. The absorbances of a series of phenobarbltal solu­ tions (4.0 x 10"5 Molar) containing varying concentrations of mercuric acetate were measured with a Beckman DU Spectrophotometvr at 239 mix and a pH = 10.9. The data are presented In Table 6. Tfte absorbance calculated from molar

TABLE 6

THE EFFECT OF MERCURIC ION ON THE MOLAR ABSORPTIVITY OF PHENOBARBITAL AT 239 m|X

Mercuric Ion Sample Molar x 105 Aobs 1 0 0.441

2 0.54 0.437

3 1.08 0.443 4 1.52 0.447 3 2.16 0.446 6 2.43 0.436

7 2.70 0.445 8 2.97 0.454

9 3-24 0.445 10 3-78 0.446

absorptivity was 0.443* The observed absorbance varied slightly from this value but the average A0bs was 0.443 which agreed within ’99#- Thus, it appears that mercuric ion does not Interfere with the assay of total phenobarbltal in a system with the pH of 10.9.

Preparation of the Mercuric Complexes Pedley (5) has shown that the mercuric complexes of the barbiturates do form stolchlometrlcally. The addition of a solution of a mercuric salt to an alkaline sodium barbiturate solution produces a white, amorphous precipitate which has a 1:1 ratio of mercury to barbiturate. The complexes of nine barbiturates were prepared as described above, Isolated by filtration and washed with water until the filtrate was neutral, then air-dried and finally dried in a vacuum oven at 60° C. over phosphorus pentoxlde.

Distribution Coefficients of the Barbiturate-Mercuric Complexes

The distribution coefficients of the complexes were determined between chloroform and a trls (hydroxymethyl) amino methane (tris) buffer with a pH 8.90 value. A stock solution was prepared by dissolving the mercuric complex in chloroform. Ten milliliter samples of each of the chloro­ form stock solution and 10 ml. of the tris buffer were placed into ten dram Optlclear vials. The vials were sealed with polyethlene caps and placed on paddles in a constant temper­ ature bath at 30° C. The samples were rotated for 12 to 18 hours, then removed and shaken for 1 minute. When the phases 46 were completely separated an aliquot of the chloroform phase was removed and assayed for mercury content as previously described. Distribution coefficient, kc, is defined as the con­ centration of the complex in chloroform/concentration of the complex in the aqueous phase. The kc was calculated by the following equation.

^ Asample c Astock “ Asample Where Asample was the absorbance of the sample at 495 mp. and Astock was the absorbance of the stock solution. The data listed in Table 7 are the averages of ten determinations for each complex and the standard deviation from the mean. TABLE 7 DISTRIBUTION COEFFICIENTS OF THE BARBITURATE- MERCURIC C0MPU3XES

Parent Barbiturate Kc Molecular Weight of Mercuric Complex Phenobarbltal O.O87 + 0.019 431.84 Dlallylbarblturic Acid 0.049 +0.010 407.82

Pentobarbital 2.78 + 0 .16 425.88 Amobarbltal 2.80 + 0.16 425.88

Butethal 1.11 + 0 .05 411.85

Butabarbltal 1.21 + 0.05 411.85 Vinbarbital 0.326+ 0.027 423.86

Probarbital 0 .016+ 0.004 392.83 The Effect of pH on the Barbital- Mercuric Complem

The distribution coefficient of barbital was deter­ mined between chloroform and acidic buffer (pH 1) and a tris buffer (pH 8.0) at 25° C. Varying amounts of mercuric Ions were added to the aqueous phase to Investigate what effect the presence of an extractable barbital complex would have. The data in Table 8 and Figure 6 illustrate the effect of mercuric ions on the distribution coefficient of barbital at two pH values. Previously, the complex was shown to form only at alkaline pH values. From the results reported in Table 8 this pH dependence was confirmed. At pH 1 the distribution coefficient of barbital was not altered In the presence of mercuric ions while at the pH 8 the mercuric-barbital com­ plex did form and a portion of it was extracted into the chloroform phase, thus causing the change in the distribution coefficient. Die distribution coefficients were determined by assaying the aqueous phase spectrophotometrlcally for barbital. Hie following equation was used in the calculation of the k value. « Astock - Asample * ^sample Where Astoclc was the absorbance of the aqueous barbital stock solution and Asample the absorbance of the sample following distribution. 48

Hie increase of barbital concentration In the organic phase due to extraction of the mercuric complex was determined from the following equation and the results are Illustrated in Table 8 and Figure 7*

Cl - Cw(k* - k) Where C* represents the Increase in barbital concentration o in the chloroform phase due to formation and extraction of the mercuric complex, Cw represents the concentration of barbital in the aqueous phase, k* the apparent distribution coefficient and k the intrinsic distribution coefficient in the absence of mercuric ion. The Increased concentration of barbital in the chloroform was proportional to the total amount of mercuric ions added to the system, as shown in Figure 7 the slope of the line was equal to 1.05* These results suggest that the barbital complex was extracted quantitatively, and a 1:1 complex was confirmed.

The Effect of Mercuric Ion on the distribution of Phenobarbltal Hie effect of mercuric ions on the distribution coefficient of phenobarbltal between a tris buffer (pH 8.90) and chloroform was determined at 30° C. as previously des­ cribed.

t 49 TABLE 8

THE EPPECT OP MERCURIC IONS ON THE DISTRIBUTION COEFFICIENT OP BARBITAL AT TVO pH VALUES

Barbital . Mercuric Ion Sample Molar x 104 Mofta* x 1©4 k* Molar x 104 pH 1.0

1 4.385 0 O .69 0 2 4.385 0.29 0.72 0 3 4.385 0.58 0.65 0 4 4.385 0.83 O .65 0

5 4.385 1.43 O .69 0 6 4.385 1.75 O.69 0 7 4.385 2.04 0.66 0 8 4.385 2.33 0.69 0

9 4.385 2.91 0.68 0 pH 8 .0

1 9.950 00 0.28 0

2 9.950 0.70 0.43 1.05

3 9.950 1.39 0.52 1.52 4 9.950 2.09 0.72 2.49

5 9.950 2.78 0.91 3.21 6 9.950 3.48 3.96 1 , 1 5 7 9.950 4.17 1.47 4.71 8 9.950 4.87 1.86 5.39

9 9.950 5.56 2.36 6.04 10 9-950 6.95 4.13 7.32 50

3

2

1

0 0 1 2 3 4 5 6 7 8 Molar Merourio Ion x 104 Fig. 6. The effect of merourio Ion on the distribution ooeffioient of barbital at two jE values* 51

8

6

4

2

0 0 2 4 6 8 Molar Merourio Ion x 104 Fig* 7* The increase in barbital oonoentration in ohloroform as a funotion of total merourio ion in the system* 52 The aqueous phase was assayed spectrophotometrlcally at 2k0 mil for phenobarbltal and the chloroform phase assayed at 495 mil for the mercuric complex by the dithlzone method. Table 9 and Figures 8 and 9 present the data from this experiment. As greater concentrations of mercuric Ions were added to the system more phenobarbltal was extracted Into the organic phase. Because of the similarity In the slopes of the lines describing the Increase of phenobarbltal concen­ tration In the ozganlc phase as a function of the total mercuric Ions added to the system as well as the mercuric complex present In the chloroform these data suggest that the complex was almost quanltltatively extracted, and, since the slopes approach unity a 1:1 complex appears to have been formed. Deviations of the slope from unity would be a direct Indication of the extraction of a complex which does not have a 1 to 1 ratio of metal to ligand. Due to the large concentration of barbital as compared to mercuric ion in the system the formation of a higher complex as well as the expected complex may cause the slope to deviate from unity.

Discussion

Barbituric acid derivatives are precipitated by mercuric ions, which are soluble In many organic solvents. The distribution coefficients of a series of barbiturate- mercuric complexes were determined between an alkaline 53 TABLE 9 THE EFFECT OF MERCURIC IONS ON THE DISTRIBUTION COEFFICIENT OF PHENOBARBITAL (Phenobarbltal Concentration 1.133 x 10 “3 Molar)

Mercuric Total Mercuric Organic co •. Sample Molar x 10**_____ Molar x 102*_____ K* Molar x 10 1 0 0 0.15 0 2 0.52 0.46 0.17 0.28

3 0.78 - 0.21 - 4 1.04 0.96 0.21 0.73 5 1.30 - 0.26 - 6 1.56 1.32 0.31 1.69

7 1.82 - 0.32 - 8 2.08 1.58 0.36 2.04

9 2.34 - 0.44 - 10 2.59 2.19 0.45 2.71 11 2.84 - 0.46 -

12 3.11 3.28 0.61 3.71 13 3.37 wm 0.69 - 14 3.63 3.20 0.72 4.33 15 3.89 - O.76 - 16 4.15 4.18 0.87 4.95 17 4.41 - 0.91 - 18 4.70 4.46 0.95 5.29 19 4.93 1.06 - 20 5.19 4.77 1.12 5.88 54

1.2

1*0

0*8

0.2

0 1 5 4 5 Molar Merourio x 10* Fig* 8. The effeot of merourio Ion on the distribution ooeffiolent of phenobarbltal* 66

8

6

4

2

0

0 1 2 6 4 5 Molar Merourio in Organic Phase z 10* Fig* 9. The Increase In phenobarbltal oonoentration In ohloroform as a function of merourio ion in chloroform. buffer and chloroform. It was of Interest to note that the distribution coefficients of these complexes followed the same general trends In variation as did the distribution coefficients of the corresponding undissociated parent acids (Table 10). The more polar the parent acid the lower the distribution coefficient, while as the polarity increased the distribution coefficient Increased In a analogous manner. Polarity changes in this series are caused by the substitu­ ents in the five position of the pyrimidine ring.

TABUS 10 COMPARISON OF DISTRIBUTION COEFFICIENTS OF SOME BARBITURATES AND THEIR MERCURIC COMPIEXES

Parent Barbiturate Kc k* k** Phenobarbltal 0.087 1.54 0.63

Barbital m m 0.50 0.11

Dlallylbarblturlc acid 0.049 1.77 0.43 Pentobarbital 2.78 6.37 5.88 Amobarbltal 2.80 15.38 6.25 Buthethal 1.11 - 2.13 Butabarbltal 1.21 - 2.17 Vinbarbital 0.326 - 1.56 Probarbital 0.016 - 0.40 k* calculated from the data of Dybllng (101) with chloroform as the organic phase, k** reciprocal of the distribution 57 coefficients given for the free barbiturates between an acidic buffer and a 60# chloroform-40# isooctane organic phase (36). The barbital complex had an extremely low solubility in chloroform thus the absorbancy readings of the organic phase following distribution were insignificant, making the determination of the distribution coefficient virtually impossible. These results possibly would be expected by analysis of the data present in Table 10. BJorling et al. (9) were also confronted by this same problem when extract­ ing the barbital complex from a buffer solution with chloro­ form. Less than 20# of the complex was removed after four extractions. The organic phase was made more polar by the addition of benzyl alcohol and then complete extraction was possible.

The distribution coefficients of barbital and phenobarbltal between an alkaline buffer and chloroform were shown to vary with the addition of mercuric ion. Both barbiturate complexes were found to be extracted into the organic phase in a quantitative manner as shown by the amount of barbiturate increase in the chloroform phase. This increase was equivalent to mercuric ion added to the system and to the divalent mercury found by the dlthizone method.

The drastic effect which the mercuric ion had on the distribution coefficients of these barbiturates would not be expected from the distribution data for the complexes alone. Other components of the system are known to react with mercuric ion, such as tris (hydroxymethyl) amino methane. Uils basic buffer constituent has been shown to form a complex ion with the formula (tris^Hg and having a stability constant of 101^*1(102). This compound is soluble only in the aqueous phase and would not assist in distribu­ tion of free barbiturate into the chloroform. Pedley (5) has stated that the barbiturate complex does not form in the presence of a mineral acid, and hydrochloric acid was a constituent of the buffer solution. These interactions Involving the buffer substances would tend to give low results due to the type of compounds that would be formed. Since the barbiturate Increase in the chloroform was directly proportional to the mercuric ion concentration, the complex was formed and must have a stability constant larger than 1016*1. In the system, the mercuric ion concentration was small in comparison to the barbiturate and buffer concentra­ tion, thus forming a barbiturate complex which was in low concentration and was the least soluble substance present. This complex could have been "salted out" by the many more soluble substances present in the buffer and in turn causing a decrease in barbiturate in the aqueous phase. 59 Die reactions Involved In the formation of the mercuric complex of the barbiturates can be described by the following reaction sequence. BarbH + OH" - Barb” + 1^0

Barb" + H g ^ — - ^ (BarbHg) complex Infrared spectra of some barbituric acids and their com­ pounds containing sodium, barium and mercuric ions Indicated the mercuric compound was a complex rather than a salt as with the sodium and barium (8). Evidence of various types has Indicated that the nitrogen atoms are Involved in the complex formation. Barbiturates having no alkyl substitu­ ents on the nitrogen atoms form 1:1 complexes, two moles of the monosubstituted compounds react with one mole of mercuric ion, and the dlsubstltuted barbiturates do not react. Bjorling et al. (9) noted that nitrogen containing compounds having a structure similar to barbituric acid, such as cer­ tain hydantolns and cyclic imides also form these complexes which are soluble in organic solvents. A chelate structure for this complex has recently been suggested by Kurplel et al. (6) and is given below (I).

o

I 60

Figure 10 illustrates the ultraviolet absorption spectra for the mercuric complex of barbital in chloroform and dioxane as compared to the spectrum for the enollzed barbital in alkaline solution. It appears that complex formation with mercuric ion causes the compound to remain in the enollzed form as shown by the characteristic absorp­ tion peak around 240 mix. nils was further verified by the fact that the molar absorptivity of the complex in aqueous solution was nearly the same as that of the enollzed barbital. ftie reaction yielding the complex has also been shown to be dependent on the pH of the solution, as illustrated by the distribution coefficient changes in Figure 6. The com­ plex was not formed at the pH of 1 while at the pH of 8.0 the complex was formed and extracted into the organic phase. It appears that enollzaticn was required for the formation of the mercuric complex. Mercuric ions appear to attack the nitrogen atom not Involved in enolization and causing the release of an equlmolar amount of protons which can be detected by the initial pH drop in an unbuffered solution. The metal ion also coordinates with the second nitrogen giving a neutral complex. In this chelate a coorindation number of two would be expected from other data involving complexes containing mercuric ion and nitrogen containing substances. Complex In chloroform

0*6 Complex in dloxane

0.4

Barbital(enollzed)

S00 220 £60 £80 500 320 340 Wave length, Millimicrons

Fig* 10* Ultraviolet speotra of barbital and merourio-barbital complex* 62 The structure proposed by Kurplel et al. (6) appears to satisfy all the data which has been presented except for the spectral evidence for the enollzed carbonyl group.

III. The Stability of Some Cobalt, Amine and Barbiturate Complexes i n a tttonaqueous Solvent- During the last three decades many investigators have employed Koppanyl1 s method for the qualitative and quantita­ tive determination of barbiturates. However, several Investigators have criticized this procedure in quantitative determinations due to the nonreproduclbllity and the Instability of the highly colored complex formed between the cobaltous salt, amine and barbiturate.

The alms of this investigation were to determine the relative stability of the complex compounds formed during the Koppanyl procedure and to more extensively elucidate the causes for the nonreproduclbllity of the color formation.

Theory

If the reaction between the barbiturate, amine and cobaltous salt which causes the color formation Is assumed to be an equilibrium, then the reaction or reaction sequence can be defined by a series of equations. From these mass action equations the system can be described and the appropriate equilibrium constants defined. Since cobaltous salts In organic solvents have a characteristic visible spectra and the amine and barbiturate are colorless the systems are defined In terms of total cobaltous salt. The reactions Involved In the formation of the highly colored complex may proceed either In a stepwise manner (two step sequence) or by a one step reaction. A two step sequence could arise by two alternate routes as described In equations 1, 2, 3 and 4. Let M - Cobaltous A * Amine B » Barbiturate M + 2A ~ MAg (1)

MA2 + 2B a-... JrAgMBg (2) or

M + 2B — .... MBg (3) MBg + 2A s ■ — AgMBg (4) Using the reaction sequence illustrated by equations 1 and 2 the following derivation was obtained. Step 1

■ h - T ^ (5) Cm ][a ] Step 2 t v j H ] (6) [MA2 ][b] Where total cobaltous concentration is equal to [M]b.

[Mlt - CM] + [MAg] + [AaMBg] (7) The observed absorbance, Aob8.

Aobs - E[M] + E1[MA21 + E2[A2MB2] (8) 64

The molar absorptlvitles E, E^ and Eq are measurable, as well as, the AQbs when the concentrations of amine and barbiturate are known quantities.

ftie [MA2] and [A2MB2 ] are determined In terms of total cobalt, [M]t and then by combining equations 5> 6, 7 and 8 the observed absorbance Is desdrlbed by equation 9*

rif1 k + E ^ A ] 2 + E2kik2 [A]2[B]2 Aobs " tM]t ------(9) \ 1 + ki[A]2 + k1k2[A]2[B]2

From equation 9 a plot of A0t>8 versus [M]t gives a straight line with the Intercept equal to zero and the slope

E + EjkilA]2 + E2k1k2[A]2[B]2

1 + kitA]2 + kikaU^tB]2 If the alternate reaction sequence was Involved an analogous derivation using reactions 3 and 4 gives equation 10.

. - r«l /E + E3k3tB]2 + E4k3k2|[B]2[A]2 \ obs tMJt ------—Z------2 J (10) \ 1 + k3[B]2 + k3k4 [B]2[A]2 / Where k3 and k4 are the respective equilibrium constants for reactions 3 and 4, and E and E4, the molar absorptivi- ties for the products of these reactions. From equation 10, a plot of Aobfl versus [M]^ gives a straight line with the Intercept of zero and the slope

E + E3k3rB]2 + E4k3k4 [B]2[A]2

1 + k3[B]2 - k3k4[B]2[A]2 In a similar manner a one step reaction can be described. 2A + M + 2B A2MB2

k _ i—[A2MB2] _— L- (li), . U 1 2[b ]2[m ] Total cobalt Is equal to [M)b.

[M]t - [M] + [AgMBg] (12)

Observed absorbance Is equal to Aob8.

Aobs - E[M] + E1[A2MB2] (13)

The concentration of free metal Ion and of the complex were determined In terms of total metal. Combining equations 11, 12, and 13 gives equation 14 which describes the absorbance of the colored complex.

/is + ®1 ^ (W) \ 1 + 1k[A]2[B]2/

From equation 14, a plot of Aobs versus [M] t gives a straight line with an Intercept of zero and the slope

. B + Ejk[A]2[B]2

1 + k[Als[B]2 . 66

Materials Isopropylamlne (Eastman Chemical Company), cobalt chloride (Baker Chemical Company), anhydrous chloroform and anhydrous methanol were of reagent grade and were not further purified. Barbiturates were purified by recrystallization as previously described.

Spectrophotometric Procedure The reagent solutions were prepared fresh for each series of determinations. Isopropylamine was rapidly weighed into an appropriate volumetric flask and sufficient chloroform was added to volume. Cobalt chloride solution was prepared employing methanol as the solvent and the appropriate barbiturate was dissolved in chloroform. Barbital was the compound used in all preliminary investiga­ tions- as the standard barbiturate. The desired quantities of the various reagent solu­ tions were added to a volumetric flask with the aid of a buret or pipet and sufficient chloroform was added to volume. A Perkln-Elmer Spectracord, Model 4000, was the instrument used to determine the visible spectra of the colored complex

in the various solutions. The maximum absorbance of the cobalt chloride was shown to appear at 525 rap. and at that

wave length follows Beer's Law (Table 11 and Figures 11 and 12). The maximum absorbance of the amine-cobalt-barbiturate

complex appears near 563 mu and follows Beer's Law at this wave length, as well as at 323 mp.. In all determinations the absorbances were measured with the Spectracord or Beckman DU Spectrophotometer at

525 mix or 563 mix and In some cases at both wave lengths using chloroform as the blank.

TABIE 11 BEER'S LAW DATA FOR COBALT CHLORIDE AT TWO WAVE LENGTHS

Cobalt Chloride Aobs E Aobs E Molar x 103 525 mix 525 mix 563 mix 563 mix 0 0 0 0 0 0.47 0.064 13.6 0.086 18.3 0.94 0.122 12.9 0.126 13.4 1.41 0.180 12.8 0.160 11.3 1.88 0.239 12.7 0.190 10.1 2.34 0.288 12.3 0.210 9.0

Cobalt chloride follows Beer's Law only at 523 mix and hot at the maximum absorbance of the complex.

Analysis of the Two Step Sequence The validity of the previously derived equations 9 and 10, describing the two possible alternate reaction sequences leading to the formation of the highly colored complex was based on the assumption that the starting material (cobalt chloride), the intermediate complex (amine-cobalt or barbiturate-cobalt) and the final highly colored complex have characteristic visible spectra and that their molar ^.bsorptlvltles are different and measurable. 1.0

Complex

0.6 obs

Cobalt salt

0.8

400 450 500 550 600 650 700 Wave length, Mllllmiorons Fig* 11* Visible speotra for amine-cobalt-barbiturate oomplex and oobalt ohloride. 69

0*30

0.25 525mn

0.20 56Smn

0.10

0.05

0 0.5 1.0 1.5 Cobaltous Concentration, Molar x 10s

Fig. 12. Standard ourva for oobaltoua ohloride at two wave lengths. 70 Hie measurement of E and B2 (or E4) was shown to be acceptable but upon the addition of the amine to the cobalt chloride solution* In the determination of E2* a brilliant blue color formed Immediately Which rapidly faded to a colorless* slightly colloidal solution. In the analogous reaction sequence the determination of tie molar absorptiv­ ity of the cobalt-barblturate Intermediate (E3) was not possible due to Identical spectra for cobalt chloride In the presence and the absence of barbiturate. Thus, It became Impossible to determine because of the formation of an Insoluble material and E^ was shown to be zero. Since the basic assumptions were shown not to be true both equation 9 and 10 are Invalidated In the calculation of the equilibrium constants for the reactions. It was of Interest to note that the order of mixing the reagents was critical. When the cobalt and amine solu­ tions were mixed the Insoluble product which formed remained unchanged in the presence of barbiturate and no color was produced. However* when the cobalt and barbiturate solutions were Initially mixed and then the amine added* the highly colored complex did form In the usual manner. Also the complex formed when the cobalt solution was added to an amlne-barblturate mixture • 7 1 The Effect of Barbital Concentration on A ^ ,

In the preceding section it was noted that the molar absorptivity may vary slightly as the barbiturate concentra­ tion varies, it was necessary to determine how the concen­ tration of barbital alters the observed absorbance of the highly colored complex. The cobalt chloride (1.95 x 10 "3 Molar) and isopropylamine (7*53 x 10 Molar) concentrations were maintained constant while the barbital concentration was varied up to 1.51 x 10"2 Molar. The absorbances of the samples were determined at 525 mp. and 563 mix and the results obtained are reported in Table 12 and Illustrated in

Figure 13.

The Effect of Amine Concentration on Aobs

In all determinations concerned with the quantitative aspects of Koppanyl's test the amine concentration was maintained in excess. Since the amine concentration may vary between various sample series it was necessary to determine the effect of changing amine concentration on the observed absorbance of the colored complex. The cobalt chloride (2.06 x 10-3 Molar) and barbital (9 .5 1 x 10*3 Molar) concentrations were maintained constant while varying the amine concentration. The observed 78

1.0

0*8

563 xmx

obs 525 mu

0*4

0 0.4 1.6 2.0 Uolar Barbital x 10s

Fig* IS* The effeot of barbital on observed absorbanee* TABLE 12 i VARIATION OF BARBITAL CONCENTRATION ON OBSERVED ABSORBANCE

Barbital Sample Molar x 102 A525 A563

1 0 0 0

2 0.168 0.132 0.177 3 0.336 0.322 0.536 4 0.502 0.433 0.681 5 0.672 0.443 0.683 6 0.840 0.461 O .703

7 1.008 0.461 0 .700 8 1.176 0.460 0.701

9 1.344 0.451 0.688 10 1.512 0.453 0.685 absorbance were measured as previously described and the results are listed In Table 13 and illustrated In Figure 14. It is apparent from the results reported In the table that the observed absorbance Increased with an Increase In amine concentration. In additional runs similar type curves resulted.

Equilibrium Constant from the One Step Reaction Previously, It has been shown that the molar absorp- tlvltles for the cobalt chloride and the barbiturate-cobalt- amlne complex could be determined In a convenient manner. 74 TABLE 13 VARIATION OF ISOPROPYLAMINE CONCENTRATION EFFECT ON Aobs

Sample Isoprppylamine A525 m

1 0.33 0 .0 3 6 0.055 2 0.67 0 .0 5 8 0.089 3 1.00 0.132 0.202

4 1.33 0.193 0.295 5 1.67 0.264 0.404 6 2 .0 0 O .327 0 .5 0 1

7 2.33 O .369 0.565 8 2.66 0.418 0.640

9 3.00 0.418 0.640 10 3.33 0.431 0.660

11 3.66 0.439 0.672 12 4.00 0.446 0.683

13 4.33 0.453 0.694 14 4.66 0.451 0.690

15 5.00 0.449 0.687

16 6.66 0.468 0.717 17 8.33 0.465 0.712

18 9-99 0.479 0.733 11.66 1 9 0.468 0.717 20 13-32 0.465 0.712 21 14.99 0.476 0.729 22 16.65 0.476 0.729 23 1 8 .3 2 0.481 0.736 0 . 8

563 mu

0.6 lobs 5E5 mu

0 4 8 12 16 20 Molar Isopropylamine x 10% Pig* 14* The effect of amine concentration on observed absorbanoe* 7 6 ftiese values £ and E ^ as well as the concentration of amine and barbiturate are essential In the calculations Involving equation 14. Table 14 and Figure 15 list the quantities of react­ ants employed and illustrate the results obtained from the visible spectral measurements.

TABLE 14

REAGENTS AND SPECTRAL DATA FOR BARBITAL-COBALT-AMINE COMPIEX

Cobalt . Barbital ' Isopropylamine Absorbance (AobB) Molar x 103 Molar x 102 Molar x 102 525 mp. 563 mp. 0 .6 1 7.25 8 .1 0 0.125 0.229 1 .2 2 7.25 8 .1 0 0.251 0.432 i m 7.25 8.10 0.422 0.700 2 .4 4 7.25 8.10 0.530 0.878 3 .0 6 7.25 8.10 0.665 1.110

3 .6 7 7.25 8.10 0.797 1.315 4.28 7.25 8.10 0.945 1.535 4.89 7.25 8.10 1.063 1.722

A plot of Aoba versus [M]t gave a straight line with the Intercept of approximately zero. Equation 14 was manipulated algebraically to give the following equation which was In terms of the slope. slope - B [A]2 [B]2 (£ 2 - slope) 77

1.6 563 mu

1*2 obs 525 mu 0*8

0

0 1 2 3 4 5 Molar Cob<ous z 103

Fig* 15. The effeot of varying oobalt oonoentratlon on the observed absorbanoa of the barbital complex* 78 The E was obtained from the spectral data for cobalt chloride In Table 11 and Figure 12. Ej was obtained from the obserited absorbance at a solution containing an excess of amine and barbital. At both wave lengths the k vdues were similar, about 6.8 x 10^.

The Color Intensities of Various barbiturate complexes A series of six barbiturates were used to determine the effect of the dlsubstltutlon In the 5 position on the intensity of the colored barbiturate-cobalt-amine complexes. Hie barbiturate and isopropylamine (0.065 Molar) concen­ tration was varied. A Beckman DU Spectrophotometer and a Perkln-Elmer Spectracord, Model 4000, equipped with tungsten lamps were used to measure the absorbances of the various a simples at two wave lengths. Representative of the results obtained are listed In Table 15 and Illustrated In Figure 16. Different cobalt chloride concentrations were used with barbital than for the other barbiturates to give a more comprehensive picture of their color intensity as compared to that of barbital. All barbiturate concentrations were of similar magnitude and the El value for each compound was determined in the presence of vast excesses of both Isopropylamine and the barbiturate. Figure 16 shows that the color Intensities of the com­ plexes are very similar to that of barbital (line). This would be expected due to the structure similarities between 79

• Vinbarbital o Phenobarbital a Amobarbital a Pentobarbital 1 . 6 x Butabarbital

obs

0 . 8

0 1 2 3 4 5 Molar Cobalt z 1 0 * : Pig. 16. A oomparison of the observed absorbances for various barbiturate complexes. 80

TABLE 15 THE EFFECT OF 5,5- DISUBSTITUTION ON THE SPECTRAL CHARACTERISTICS OF THE BARBITURATE COMPLEXES

Barbiturate Ei Slope k x 10-6 Barbiturate Molar x 102 525 m 525 mil 525 mp.

Barbital 3-37 2 5 4 .6 246.3 3.91 Vinbarbital 3.66 2 5 0 .7 241.0 5.47 Phenobarbltal 3*07 248.2 237.1 5 .0 8

Amobarbltal 3.25 249.9 237.9 4.21 Pentobarbital 2.80 236.7 221.9 4.21 Butabarbltal 2.96 232.8 219.1 4.10 the compounds. The calculated k values also appear to be similar as shown In Table 15.

Discussion The highly colored complex which forms between a barbiturate, a cobaltous salt and an amine has been defined by the following formula, Barblturate2-Cobalt-Amlne2 * This complex forms only In anhydrous organic solvents. Many other nitrogen containing materials give the same color reaction, for example, phthallmlde, theobromine and theophylline. Thus, the reaction Is not specific for barbituric acid derivatives. The amine Is used as an alkalizing agent and can be replaced by barium oxide, potassium hydroxide or sodium ethoxlde, however, the amine (usually Isopropylamlne) has been shown to give the most stable color. 81

In all previous Investigations the barbiturates have been the compounds upon which the studies were based. This was the logical approach when attempting to establish a method for the quantitative and qualitative analysis for a specific barbiturate or for, the series of barbiturates. In this study a new approach has been undertaken by studying the effect of cobaltous salt concentration on the color Intensity. The equations which were previously derived are based upon the spectral properties of the cobaltous salts In the presence and absence of the barbiturates and amine. From the spectral data It had been expected that the reaction sequence for the formation of the highly colored complex would be verified. However, due to the several alternate reactions which are possible only one sequence has been shown not to be Involved. There remains three possible routes, each Involving the Initial presence of the barbiturate.

1) M + 2B 3 = MBg MB2 + 2A — * A2MB2

2) 2A 2B s = " *2^2 V®2 M 3rr-— T A2MB2 3) 2A + 2B + M -— — a 2m b 9

Reaction sequence 1 cannot be studied by application of equation 10 because there appears to be no detectable Intermediate complex formed, which was a requirement for 82 this method. Since all equations were based upon the intermediate complex formation with cobalt, reaction sequence 2 was not applicable to this procedure.

Ohe presence of an alkalizing agent has been shown to be essential for the formation of the highly colored complex and this base must be maintained in excess. In Figure 14, it is noted that when the concentrations of cobalt and barbital were maintained constant and the Isopropylamine concentration varied, the break in the curve illustrated the main reactiai was completed when the ratio of amine to cobalt was not 2:1 but greater than 10:1. This type of curve reappeared during each replication of the run. Thus, it appears that the amine concentration must remain in vast excess, but again in Figure 14, it is noted that the absorptivity of the complex slowly Increased with an increase in amine concentration. When the cobalt and Isopropylamine concentrations were maintained constant and the barbital varied (Figure 13), it was noted that at the completion of the reaction the ratio of barbital to cobalt was 2.1:1 which was in good agreement with the expected ratio 2:1. Figure 13 also shows that there was no Increase in the absorptivity of the cepplex with an excess of barbital present.

From the above discussion it appears that the barbital did react in a quantitative manner with the cobalt when in the presence of the lsopropylamlne. The amine appears to be a requirement In this system for two reasons. First, as a required ligand with an electron rich nitrogen atom to stabilize the system to some extent and secondly, as a base which removes an acidic hydrogen from pyrimidine nitrogen of the barbiturate thus making possible the use of the barbiturate as a ligand. From this evidence it appears that the barbiturate coordinates with the cobalt through a nitrogen. This Is further verified by the fact that thio­ pental produces the same colored complex as do the barbi­ turates with an oxygen in the two position. It would be difficult to suggest a more specific explanation for this color reaction due to the use of the nonaqueous solvents. With this information It would be expected that the complex would arise either through reaction sequences 2 or 3 bicause we have shown that no reaction occurs In the absence or presence of cobalt until the Isopropylamine converts the barbiturate into the coordinating species. It is a known fact that in aqueous solutions barbiturates only interact with silver ions and mercuric ions when the ligand Is enolized.

Reaction sequence 3 has been mathematically defined in terms of the cobaltous salt and observed absorbance. This relationship is described by equation 14. This equation was based on the assumption that an equilibrium was estab­ lished and that Ei was an accurate and obtainable value. was determined experimentally by measuring the absorbance 84 of a known concentration of cobalt In the presence of an excess of amine and barbiturate. However, the molar absorptivity of the complex varies with the amine concen­ tration thus, this was not a true constant. When the flat portion of Figure 14 was extrapolated to aero amine concentration an abnormally low value for was obtained. From equation 14, E^ must be greater than the slope of the line In order to calculate the equilibrium constant (k), when Ei varies k also varies. To compare the relative Intensities or stabilities for the complexes of a series of barbiturates, the same molar concentrations of barbiturates and amine must be used as described In Table 15• It appears that the six barbi­ turates studied form similar complexes of the same magnitude and that the color formed was of the same Intensity (Figure 16). The barbiturates differ by the substltutents In the 5 position but these groups seem to have little effect on the formation of the complexes. Some authors have attempted to differentiate between the barbiturates by the varying color Intensities on the milligram basis. This type of an approach would have merit If the molecular weights of the compounds within this series varied to a greater extent. Many Investigators have criticized the use of Kop- panyi's method for the quantitative analysis of the barbi­ turates due to the nonreproduclblllty. In systems

containing reagent grade materials and by maintaining 85 precise reagent quantities this color reaction could become a useful analytical tool. However, there are many practical drawbacks when the system cannot be predesigned. The main application of this procedure has Involved the assay of barbiturates in dosage forms, biological fluids and tissues. Each of these systems require numerous purification pro­ cedures that may cause further deviation from an ideal system. Previously, it has been shown that the amine must be present in excess amounts and that variation in this con­ centration does cause alterations in the observed absorp­ tivity of the complex. In the assay for the barbiturates the amine concentration effects greatly reduce the repro­ ducibility and accuracy of the method. Hie cobalt salts do absorb at the wave length where the complex has its maximum absorbance and cobalt does not obey Beer's Law at this wave length. Too great an excess of cobalt in the system would have two distinct disadvantages, first, causing an alteration in the spectra of the complex and second, the possibility of reacting with the excess amine to cause a precipitate, another color and reducing the amount of alkallnlzing agent present. Two main disadvantages for the solvent system which was employed are, first its volatility causing erroneous spectral data due to cacentration changes arising from evaporation and second, the requirement 'for it to be anhydrous. The moral absorptivity of the highly colored complex

(A2MB2 ) has been shown to be about 350 at 363 mix and about 230 at 523 mix* which are extremely low as compared to the molar absorptivity of the enollzed barbiturates In the ultraviolet range which was about 10 x 103 around 239 mp. - 240 rap.. Assuming that the procedure of Kbppanyi was Improved to overcome Its many disadvantages the aBsay would still not have the accuracy or the range Which are available with the ultraviolet assay. In this Investigation the solutions were analyzed Immediately following mixing to prevent the color fading from adding any further complications to the results. Preliminary phases of a study to determine the rates at which the various colors faded were Instigated and were later discarded due to the Inconsistent results obtained. 87 IV. The Interaction between Barbiturates and Naturally Occurring substances Containing Metal Ions Barbituric acid derivatives have been used thera­ peutically as sedatives and hypnotics since the turn of the century. The mechanism of action and the site of action has never been elucidated, however, some Investigations have led to the belief that the physiological action of these substances Is Intimately related to alterations In the oxidation processes within the cytochrome system. Many cytochrome enzymes contain a porphyrin-Iron chelate prosthetic group attached to two protein moieties. The In vitro Interactions between a barbiturate and similar enzyme materials may give an Insight Into the In vivo mechanism and possible site of action for these Important drug substances. It was the alms of this Investigation to determine whether an Interaction between hemln and a barbiturate was detectable by standard spectrophotometrie procedures and also to determine the effect of cuprlc Ions on the binding of pentobarbital to a protein.

Materials Hemln purchased from K and K Laboratories (Lot #20241) was purified by recrystalllzatlon from pyridine solutions as described by Fischer (103). Protein: Bovine serum albumin (Fraction V) (BSA) was purchased from Armour Laboratories 88

(Lot W19109) and used throughout this investigation. Since the BSA contains several per cent of impurities and some moisture, it was assumed to be 95% pure (90). The copper metal and all substances used in preparing the appropriate buffer solutions were of reagent grade and were not further purified. Hie barbiturates were recrystal- lized as previously described.

Hie Effect of Barbital on the Spectra of Hemln Hie visible spectra of hemin in alkaline buffer solu­ tions are characterizued by the appearance of a peak between 580 mix and 610 mix, then at a lower wave length the solutions become opaque (Figure 17). A standard visible curve at 580 mu was obtained for hemin in a borate buffer (pH 9»*0 and at 605 mix in 0.01 Normal sodium hydroxide using a Beckman DU Spectrophotometer equipped with a tungsten lamp. Hie results are listed in Table 16 and illustrated in Figure 18. In the borate buffer at 580 mix. and in 0.01 Normal sodium hydroxide at 605 mix the hemin spectra appear to obey Beer's Law and have molar absorptlvlties of 4.49 x 10^ and

4.63 x 10^, respectively. The presence of coordinating substances such as pyridine and caffeine have been shown to alter the spectral curves of hemin. Figure 19 illustrates the spectra for hemin in the presence of pyridine, borate buffer (pH 9*4), 1.0

0.0 obs

0.2

400 450 500 550 600 650 700 Wave length, Mllllmiorona Pig* 17. Visible speotrum of hemin. 90 TABIE 16

STANDARD CURVE FOR HEMIN IN A BORATE BUFFER pH 9*4 AND IN 0.01 NORMAL SODIUM HYDROXIDE

Borate Buffer pH 9*4 Sodium Hydroxide 0.01 Normal

Hemln Aobs Hemln Aobs Molar x 10^ 580 mil Molar x 105 605 mix 2.21 0.099 2.33 0.110

4.43 0.197 4.66 0.217 6.64 0.296 6.99 0.316 8.87 0.396 9.32 O .430

11.07 0.502 11.65 0.535 13.30 0.600 13.98 0.650

15.57 0.699 16.31 0.752 17.73 0.802 18.64 0.864 19.94 O .905 20.97 0.964 22.16 0.998 23.30 1.060 sodium hydroxide solution (0.01 Normal), phosphate buffer

(pH 9) and barbital In the borate buffer. All curves are similar In shape except In the presence of pyridine and

sodium hydroxide solutions, fryridine coordination causes a hypsochromlc shift In the peak of maximum absorbance to 522 mp. and the sodium hydroxide solution causes a less drastic bathochromic shift in the peak to 610 mix. The appearance of the peak at 610 mp. In the sodium hydroxide solutions was caused by the formation of heme (hydroxyhemin). Slight deviations In the spectra In the presence of the 1*1 1 . 0

0*9

0*8

0.7

0.6

0.4

* at 605 mu In 0.01 NaOH • at 580 mu in borate buffer

0.1 ;

0 2 6 10 14 18 22 Molar Hemln x 105 Pig. 18. Standard ourve for hemin in borate buffer pH 9.4 and 0.01 Normal sodium hydroxide. 1*0

0*8

0,6

obs 0,01 NaOH Barbital and- Borate Buffer0.4

Phosphate Buffer

400 450 500 550 600 650 Wave length, Millimicrons Pig, 19, Spectra of hemin in the presenoe of pyridine, barbital and various buffer solutions. 93 other buffer components were caused by the varying amounts of sodium hydroxide in the buffer systems* The phosphate buffer contained no added sodium hydroxide, thus the lower spectral curve. Tbe curve for the spectra of borate buffer- hemln system and borate buffer-barbital-hemin system are superimposable, thus it appears there are no alterations of the hemin spectra caused by an interaction with barbital* The effect of sodium barbital on the spectra of heme was determined in 0.001 N sodium hydroxide, The spectral curves of the stock solution containing no barbital and the solutions containing varying amounts of barbital were 3uperlmposable between 700 mix - 480 mix, below 480 mix slight deviations were noted. Obese deviations increased with an increase in barbital, however, the changes were not propor­ tional to the barbital concentration. Table 17 lists the results obtained at two wave lengths, 610 mix and 450 mix. From the data presented in Table 17 it is believed that these spectral changes for heme are not significant to indicate an interaction between the barbital and heme 4s was noted for pyridine and other complexing agents. If barbital does interact with hemin or heme this interaction has not been evident from this procedure. Analytical Method for Pentobarbital The concentration of pentobarbital in the dialysis solutions were determined spectrophotometrlcally at 240 mix with a Beckman DU Spectophotometer. Previously, it has been 94

TABUS 17

THE EFFECT OF BARBITAL ON THE SPECTRA OF HEME

Heme Barbital Aobs A0bs Molar x 104 Molar x 102 610 mu 450 mu

0.93 0 0.425 0.715 0.93 0.45 0.425 0.724 0.93 0.91 0.425 0.729 0.93 2.27 0.425 0.735 0.93 4.54 0.425 0.755 1.86 0 0 .8 5 6 1.432

1.86 0.45 0.856 1.440

1.86 0.91 0.856 1*457 1.86 2.27 0.856 1.471 1.86 4.54 O .856 1*510

shown that the molar absorptivity of pentobarbital In an alkaline buffer (pH 10.9) 240 mu was 10.1 x 10^ (86). Experimentally It was shown that the BSA and cuprlo salt solutions Interfere with the spectrophotometrlc assay

of pentobarbital. Figure 20 Illustrates the spectra of the Individual Interfering substances and pentobarbital under the Identical conditions In the phosphate buffer (pH 10.9). The molar absorptlvltles for these substances were determined at two wave lengths, 240 mu and 280 mix* The maximum absorbance for pentobarbital appears at 240 mu> BSA maximum absorbance was at 280 m|x> while the cuprlc Ion 1.0

0.8

0.6

BSA Obs Pentobarbital

Cuprio salt 0.8

800 850 500 Wave length, Millimlorons

Pig. 80. The speotra of BSA, ouprlo salt and pentobarbital In a phosphate buffer |fl 10.9. 96 had a nonspecific ultraviolet spectra. Tables 18 and 19 list the results for the preparation of a standard curve for the Individual Interfering substances and Figures 21 and 22 Illustrate the results. The BSA (assumed 95# pure) followed Beer's Law at both wavelengths and the molar absorptivitles at 280 mix and 240 mix are 4.51 x 10 ^ and 16.51 x 10 ^, respectively. Eichman (90) has reported the molar absorptivity for BSA at 278 mix In water to be 4.60 x 10 ^. nils value was similar to the value found at 280 mix. In both cases the molecular weight of BSA was assumed to be 69*000. The cuprlc salt obeys Beer's Law at 240 mix and up to about 8 x 10 “5 Molar at 280 mix. The concentration of cuprlc salts used in the dialysis studies fell within the accept­ able portion of the curve for the metal salt. At 240 mix the molar absorptivity was found to be 2319 and at 280 mix it was

found to be 3194 at the lower portion of the curve. Pentobarbital has been foundd to have a very low nonspecific absorbance at 280 mix, but did obey Beer's Law at 240 mix with themolar absorptivity of 1 0 .1 x 10 ^ (8 6). At each wave length It was found that the observed absorbance was additive for the components present. At 280 mix only the copper and BSA were measured while at 240 mix all three components absorbed. From these facts and the standard curves the following formula was used to calculate 97 TABLE 18 STANDARD CURVE FOR BSA AT 240 rap. AND 280 mp.

BSA Aob« Aobs Molar x 106 280 rap. 240 mp. 0.44 0.029 0.073 0.89 0.049 0.147 1.33 0.071 0.214 1.77 0 .0 8 6 0 .2 9 2 2 .2 2 0 .1 0 1 0.353 2 .6 6 0.119 0.437 3.11 0.141 0.504 3.55 0 .1 6 0 0-582 3.99 0.179 0 .6 6 8 4.44 0.199 0.746

TABLE 19 STANDARD CURVE FOR CUPRIC ION AT 240 rap. AND 280 mp.

Cuprlc Ion Aobs Aobs Molar x lO2* 280 mp. 240 rap. O .2 7 0 .0 9 0 0.069 0.54 0.175 0 .1 2 8 0 .8 1 0 .2 5 2 0 .1 8 1 1 .0 8 0 .3 0 1 0.244 1.35 0.335 0.307 1 .6 2 0.365 0.370 1.89 0.408 0.445 2 .1 6 0.445 0 .5 0 1 2.43 0.473 0.571 2.70 0 .5 0 8 0 .6 3 2 0.8

240 mu 0.6

0.2 280 mu

\ 0 1 2 3 4 Molar BSA X 10® Fig. 21. Standard ourve for BSA at 240 mu and 280 mu* 0*5

0*3 obs 280 mu

0.1

0 4 8 12 16 20 24 28 Molar Cuprlo Ion x 105 Fig. 22• Standard curve for Cuprlo Ion at 240 mu and 280 mu. 100 the amount of pentobarbital present In the system by knowing the absorbances of the sample solutions and blanks at the two wave lengths. In solutions containing only pentobarbital and slight amounts of BSA following dialysis the following equation was used In the calculations.

(A28O ~ Ab28o) x 3.64 » BSA correction at 240 mp..

Agijo - (BSA correction + Ag24o) ■ Acorr

ACorr —s « Concentration of pentobarbital Am

Where ^ 8 0 “ Observed absorbance at 280 mp..

aB280 “ Observed absorbance of blank at 280 mp.. 3.64 * Ratio of molar absorptivities for BSA at

240 mp. to 280 mp..

a240 “ Observed absorbance at 240 mp.. AB240 18 Observed absorbance of blank at 240 mp.. Throughout each series of dialysis experiments the cuprlc ion concentration (total) was maintained constant within a series but varied between series. The following equations describe the correction applied to the readings obtained for systems which contain pentobarbital, BSA and cuprlc Ions after dialysis.

(a Cu 280 " AB28o) x 0.726 . ACu24o 101

The ACu24q was shown to remain constant within a series.

(A280 ” Ar»«Toftr>) x 3*64 - BSA correction at 240 mp..

A240 ~ correction + ACu240 + A^q) - Acorr

Where ACu 2q o * Observed absorbance of cuprlc ion solution at 280 mil free of protein.

a c u B280 * Observed absorbance of blank at 280 mp. plus

ACu280* 0.726 « Ratio of the molar absorptivities of the cuprlc Ion at 240 mp. to 280 mp.. In cases where these corrections were applied, the cuprlc ion effects were small but significant, ftie protein material which was removed during the dialysis procedure was corrected for, In the manner described above, and it was assumed that this material had the same molar absorptivity as the BSA. When large corrections were necessary the samples were dis­ carded in as much as it was possible to have an inferior dialysis bag which may have allowed more than just dialyzable material to pass through.

Dialysis Studies at pH 5.33 An equilibrium dialysis procedure described by Hughes and Klotz (95) was used to study the effect of cuprlc ion on the association of pentobarbital with bovine serum albumin in an acetate buffer at pH 5*33 and 1/2 » 0.20. Ionic strength following equilibrium was 0.10. To 20 Pyrex test tubes (25 mm. x 20 mm) was added ten ml. of the acetate buffer containing the pentobarbital. Throughout each series the pentobarbital concentration was varied. Within each tube a dialysis bag (Viscose casing,

2 1 /3 2 Inch diameter) was placed which contained 10 ml. of the 2$ BSA solution In water for the control series and In the sample series the dialysis bag contained the 2 % BSA solution with known amounts of cuprlc Ion. The tubes were sealed with polyethylene caps and placed In the shaking apparatus. Equilibrium was assumed to be established by gentle shaking for twenty-four hours at 30°C. At equilibrium, the external phase was assayed spectrophotometrically for pentobarbital as previously described. Appropriate blank solutions were run to determine the external concentration of pentobarbital In the labsence of BSA, as well as the spectral contributions from the dialysis membrane, buffer components, cuprlc ion and BSA components. The standard cuprlc ion solution were prepared by weighing out a known amount of copper metal and dissolving it In a minimum amount of a mixture of nitric acid and sulfuric acid. Die acid solution was heated to dryness on a hot plate after the reaction had ceased. The residue was > dissolved In water, transferred to a volumetric flask and sufficient distilled water was added to give the desired final volume. 103 Die number of moles of pentobarbital bound In the albumin tube was determined by the difference In molar con­ centration of pentobarbital In the external phase In the presence and absence of the BSA solution within the dialysis bag. Since the BSA solution was prepared by dissolving a known weight of BSA In sufficient water to make the desired volume, the number of moles of BSA within the dialysis bag was determined when the molecular weight was assumed to be 69,000. The association constants and the number of binding sites (maximum) on the BSA were evaluated by the following equation developed by Klotz (104). n (A) K + (A) The derivation of this equation was based on the law of mass action and on the assumption that association occurred In a stepwise fashion. The above equation relates the number of moles of pentobarbital bound to either the moles of BSA or BSA-Cuprlc Ion complex, r, with the concentration of the unbound pentobarbital, (A). K Is the dissociation constant and n Is the maximum number of moles of pentobarbital that could be bound per mole of protein. A rearranged form of the equation was used In determining .the numerical values for

K and n. 1 . K 1 + 1 r n (A) n 104 l _L_ A plot of r versus (5), gave a straight line with the slope e^uax to f and the mtercept eauaX to ±.

Tables 20, 21, 22 and 23 list the results for the effect of cuprlc Ion on the association of pentobarbital to bovine serum albumin at pH 5 »33• Figure 23 illustrates the results graphically.

Dialysis Studies at pH 7>42 Equilibrium dialysis was used to study the effect of cuprlc Ion on the association of pentobarbital with bovine serum albumin in a phosphate buffer pH 7*42 and f/2. « 0.20. Following dialysis the ionic strength of the system was 0.10. The procedure and method of evaluation for the results have been described in the previous section. The results are listed In Tables 24, 25* 26 and 27 and illustrated in Figure 24.

Discussion and Results

Barbital appeared to have little effect on the spectra of hemin in various systems. Some interactions have no effect on the spectrA of the components, thus It is only possible to state that there was no apparent interaction between barbital and hemin detectable from the spectro- photometrlc procedure.

Pentobarbital was chosen as the barbituric acid derivative to be used in determining the effect of cuprlc ion on the binding of a barbiturate to bovine serum albumin. 105 TABLE 20

THE BINDING OP PENTOBARBITAL TO 2 % BSA AT pH 5-33 AND 30°C.

(A) x 106 Moles bound x 10^ r 1 1 x 10-5 I W 0 .2 8 0.39 0.14 7.14 35.7 1 .0 8 0.39 0.14 7.14 9.26 1.97 0.41 0.15 6.67 5.08 2.60 0.48 0.17 5.88 3.85 3.05 0.82 0.29 3.45 3.28 4.27 0 .6 1 0.22 4.55 2.34 4.89 0.68 0.24 4.17 2.04 5.35 0.31 0.88 3.22 1.87 5.35 1.11 0.40 2.50 1.87 6.03 1.17 0.42 2.38 1.67 8.21 0.54 1.51 1.87 1.22 17.89 0.83 2.35 1.20 0.56 20.18 1.17 1 .6 0 1.75 0.50 23.46 1.18 3.33 0.85 0.43 25.60 1.99 5.62 0.50 0.39

TABLE 21 THE BINDING OP PENTOBARBITAL TO 2 $ BSA CONTAINING 3.4 x 10”5 MOLAR CUPRIC ION AT pH 5.33 AND 30°C.

(A) x 106 Moles bound x 10^ r 1 * x 10”5 r Ta T 1.27 0.20 0.07 14.3 7.87 1.96 0.42 0.15 6.67 5.10 2.86 0.22 0 .0 8 12.5 3.50 3.20 0.67 0.24 4.17 3.13 3.98 0.90 0.33 3.03 2.51 4.68 0.89 0.33 3.03 2.14 5.39 1.07 O .38 2.63 1.86 6.14 1.06 O .3 8 2.63 1.63 6.87 1.15 0.41 2.44 1.46 106

TABLE 22 THE BINDING OF PENTOBARBITAL TO 2# BSA CONTAINING 6.8 x 10*3 MOLAR CUFRIC ION AT pH 5-33 AND 30°C.

1 (A) x 106 Moles bound x !06 r r 1 S T x 19'5 0.26 0.41 0.15 6.67 38.5 0.75 0.72 0.26 3.85 13.3 1.24 1.14 0.41 2.44 8.06 2.64 0.44 0 .1 6 6.25 3.79 3.46 0.41 0.15 6.67 2 .8 9 4.11 0.77 0.28 3.57 2.43 4.94 O .63 0.23 4.35 2.02 5.54 0.92 0.33 3.03 1 .8 1 6.13 1.07 0.38 2.63 1.63 6.88 1.14 0.4l 2.44 1.45

TABLE 23 THE BINDING OF PENTOBARBITAL TO 2 % BSA CONTAINING 6.75 x 10“4 MOLAR CUPRIC ION AT pH 5*33 AND 30°C.

1 (A) x 106 Moles bound x 106 r 1 x 10-5 r TST 1.86 0.94 0.33 3.03 5.43 7.07 2.65 0.94 1.06 1.41 12.84 0.27 0.10 10.00 0.78 16.75 3.49 1.24 0 .8 1 0 .6 0 20.08 3.54 1.26 0.79 0.50 24.76 2.03 0.72 1.39 0.40 25.49 5.73 2.03 0.49 0.39 2 4 I- 8 HlH 6 Pig. 23. The binding of pentobarbital to to pentobarbital of binding The 23. Pig. absence of ouprio ion at pH 5.33 and 30°0. and 5.33 pH at ion ouprio of absence • no Cuprlo Ion Cuprlo no • ■ 6.8 6.8 ■ a o 6.8 x 10”4' x Cuprlo 6.8 Ion o Molar 3*4 x 10-5 Molar Cuprlo Cuprlo Ion 10-5 Molar x 3*4 0.5 x 10“5 Molar Cuprlo Cuprlo Ion 10“5 Molar 1.0

1.5 (A) _ 1 5 — Q T j

8 1 2.0 i

BSA in the presence and presence the in BSA . 30 3.5 3.0 2.5 H © i 108

TABLE 24 THE BINDING OP PENTOBARBITAL TO 2% BSA AT pH 7*42 and 30°C.

1 (A) x 106 Moles bound x 106 r r j j j x i * T 5 1.15 0.50 0 .18 5.56 8.70 2.01 0.42 0.15 6.67 4 .9 8 2.33 0.59 0.21 4.76 4.29 2.85 O .7 8 0.28 3.57 3.51 4.20 1.57 0.57 1.76 2 .3 8 4.73 1.08 0.38 2.63 2.11 6.70 4.00 1.41 0.71 1.49 7-32 1.73 0.62 1.61 1.37 9.77 2.76 1.00 1.00 1.02 11.40 7.30 2 .58 0.39 0.88 11.49 3.90 1.41 0.71 0.87 15.21 4.6O 1.63 0 .61 0.66 18.40 3.80 1.37 0.73 0.54 18.92 7.00 2.47 0.40 0.53 23.40 5.10 1.84 0.54 0.43 26.31 8.61 3.04 0.33 0 .3 8 26.90 7.20 2.60 0.38 0.49 29.40 3.10 1.10 0.91 0.34

TABLE 25 THE BINDING OP PENTOBARBITAL TO 2 % BSA CONTAINING 5*74 x 10“5 MOLAR CUPRIC ION AT pH 7-42 AND 30°C.

1 (A) x 106 Moles bound x 106 r r U T x10”5 0 .8 7 0 .9 4 0 .38 2 .6 3 11.5 1 .7 6 1 .1 6 0 .47 1.91 5.70 2 .8 0 1 .6 1 0.67 1.49 3.60 4 .7 8 1 .5 1 0.61 1.64 2.09 4 .6 7 2 .3 3 0.94 1.06 2.11 20.4 4 5.76 2 .32 0.43 0.50 109 TABLE 26

THE BINDING OP PENTOBARBITAL TO 2# BSA CONTAINING 1.15 x 10"24 MOLAR CUPRIC ION AT pH 7-42 AND 30°C.

1 (A) x 106 Moles bound x 10^ r r TIT x 10-5

0.14 O .3 6 0.15 6 .6 7 71.4 0 .6 1 1.20 0.48 2 .0 8 16.4 2.02 0.90 O .36 2.78 4.95 2.96 1.50 0.60 1.67 3.38 3.84 2.45 0.99 1.01 2.60 5.56 1.44 O .58 1.72 1.80 11.91 2.69 1.08 0.93 0.84 14.25 6.35 2.56 0.39 0.70 23.05 3.15 1.27 0.79 0.43

TABLE 27 THE BINDING OP PENTOBARBITAL TO 2% BSA CONTAINING 1.69 x 10"2* MOLAR CUPRIC ION AT pH 7*42 and 30°C.

1 (A) x 106 Moles bound x 10^ r V TIT * 10"5

1.39 1.06 0.38 2 .6 3 7 .1 9 4.52 1.25 0.45 2.22 2.21 7 .8 0 1.25 •0.45 2.22 1.28 8 .7 8 3.75 1.35 0.74 1.14 13.20 2.20 0.79 1.27 O .76 16.91 2.00 0.72 1.39 0.59 18.18 4.00 1.44 0.69 0.55 24.50 4.00 1.44 0.69 0.41 21.80 6 .3 0 2.?7 0.44 0.46 0 1 H IH 2 3 Pig. 24. The binding of pentobarbital to to pentobarbital of binding The 24. Pig. absence of cuprlo ion at pH 7.42 and 30°C. and 7.42 pH at ion cuprlo of absence 0 a 0 5.74 x X0~® Molar Cuprlo Cuprlo Ion 5.74 X0~® Molar x 0 ■ 1.15 x 10"4 Molar Cuprlo Cuprlo Ion 1.15 Molar 10"4 x ■ no • 1.69 x 10"4 Molar Cuprlo Ion Cuprlo 1.69 Molar 10"4 x 0.5 1.0 on 1.5 L x J 10-5 (A) 2.0 2fo BSA in the presence and presence the in BSA 3.0

3.5 Ill

Pentobarbital was bound to BSA to a greater extent than most barbiturates (9 1)* For simplicity, the lines In Figures 23 and 24 repre­ sent only the binding of pentobarbital to BSA In the absence of cuprlc Ion. The other data which was determined In the presence of the metal Ion gave results which were similar, thus suggesting that cuprlc Ion has little or no effect on the binding of pentobarbital. The deviations which are apparent In Figure 24 for the data corresponding to 5.74 x 10“5 an

Hie data obtained at pH 5*33 gave a straight line with the slope of 1 .5 9 x 10“5 and the intercept of 0 .0 7* From these data, n - 14.29 and K * 2.27 x 10"4. In the similar study at pH 7*42 the slope of the line was

9 .4 9x 10“^ and the Intercept of 0.07. From these data the following values were calculated, n - 14.29 and K »

1.37 x 10-4* Probably only by coincidence the n values were the same at each pH. Karush (105) has pointed out that the values for n should not be given too much significance. This Is particularly true In these studies where the data

I 4 112 varied to such a large extent that statistical methods could not be applied In determining the slope or Intercept of the line.

Ooldbaum and Smith (91) have reported an n value of 22, and that there are two types of binding sltes> one small numerically with a large association constant and the other large numerically but with a less affinity for the pentobarbital. When taking the data for these two sites and their appropriate association constants, the calculation for an over-all dissociation constant at pH %.4 gives K - 2.5 x 10~^ which Is of similar magnitude to the values found above, when assuming one type of binding site. At both pH values the per cent of pentobarbital bound to the bovine serum albumin was of similar magnitude. SUMMARY AND CONCLUSIONS

1. A series of nine barbiturates were potentio- metrically titrated In a sodium carbonate solution with a sliver nitrate solution as tltrant. From these data, the titration efficiency was determined to be quantitative and the stability constants for nine anionic silver-barbiturate complexes were calculated. 2. The molar absorptivity of the sliver anionic complex for barbital was found to be the same as for the free barbital In alkaline solution. 3. Hie rate of alkaline hydrolysis of barbital at 42.5°C. was shown to decrease In the presence of silver ion; this decrease in reaction rate was proportional to the mole fraction of silver Ion. These results indicate that the complex resisted alkaline hydrolysis. 4. A general structure for these anionic complexes has been proposed. 5. Mercuric complexes were prepared for a series of nine barbiturates. 6. Distribution coefficients for the barbiturate- mercuric complexes-were determined between chloroform and a tris (hydroxymethyl) amlnomethane buffer. The mercuric Ion

113 114 was assayed as the dithizone chelate at 495 mix In a chloroform solution. 7. The presence of mercuric Ion was shown not to Interfere with the ultraviolet spectra of barbital.

8 . The effect of mercuric Ion on the distribution coefficient of barbital was determined at two hydrogen Ion concentrations. In acid solution the distribution coeffi­ cient was not altered but In a slightly alkaline solution a complex formed which was soluble In the organic phase. 9. The effect of mercuric Ion on the distribution coefficient of phenobarbltal was determined In a slightly alkaline buffer solution. The amount of barbiturate extracted Into the organic phase was shown to be a function of mercuric ion concentration In the aqueous phase before distribution. 10. Equations were derived in an attempt to describe the formation of the highly colored complex between a barbiturate, a cobaltous salt and isopropylamine in a nonaqueous solvent. In general, the spectrophotometrie data agreed with the expected trend in results, however, no quantitative Interpretations were possible. 11. The molar absorptivity of the complexes for six barbiturates were determined at 563 mu and were found to be similar. 115 12. Numerous reasons are discussed Which deal with the nonreproduclbility and Inaccuracy of the analytical method based on the visible spectra of the barbiturate- cobalt- isopropylamine complexes.

13* The visible spectra for hemin were found not to be altered in the presence of barbital in alkaline solution. An interaction between barbital and hemin was not apparent from the results of these spectral studies. 14. Cuprlc Ion was shown to have little or no effect on the binding of pentobarbital to bovine serum albumin at two hydrogen Ion concentrations and at 30°C. The dissoci­ ation constants for the pentobarbital-albumin complexes were calculated as 1.37 x 10“^ at pH 7.42 and 2.27 x 10”^ at pH 5.33. BIBLIOGRAPHY 1. Masserraan, J. H., Arch. Neurol, and Psychiat., 37* 617 (1937). “ “ 2. Masserman, J. H., ibid., 1250 (1938). 3. Lami, P., Boll. Chim. Farm., 53, 193 (1914). Through Chemical Abstracts, 10, 2613 (1916). 4. Bolle, A. and Mlrimanoff, A., Festschrift Paul Casparls, 154 (1949). Through Chemical Abstracts, 44, 9630 (1950). ~ 5. Pedley, E., J. Pharm. and Pharmacol., 2, 39 (1950). 6. Kurplel, I., et al., Acta Polon. Pharm., Id, 221 (1961). 7* Cohen, E. M. and Lordl, N. G., J. Pharm. Scl., 50, 661 (1961). 8. BJorling, C. 0., et al., J. Pharm. and Pharmacol.. 11. 297 (1959). 9* Bjorllng, C. 0., et al., Acta Chem. Scand., 12. 297 (1959). ' “ 10. Lubran, M., Clin. Chlm. Acta, 6, 594 (1961). 11. Zaar, B. and Gronwall^ A., Scand. J. Clin, and Lab. Invest., 13, 225 (1961).

12. Kallnoskl, K. and Baran, H., Acta Polon. Pharm., 15, 327 (1958).

13. Budde, H., Apt. Z., 4£, 695 (1934). 14. Danlelsson, B., Svenk. Farm. Tldskr., ^5, 125 (1901).

15. Furst, W., Pharmazie, 16, 24 (1961).

16. Perelman, Ya. M., J. Anal. Chem. U.S.S.R., 11, 24l (1956). ~ 17. Perelman, Ya. M., Ibid., 11, 466 (1956).

116 117 18. Poethke, W. and Furst, W., Pharmazie, 15, 673 (I960).

19* Gautier, J. A., et al., Ann. pharm. franc., 16, 625 (1958). ““ 20. Papaport, L. I. and Vaiaman, G. A., Ukrain, Khlm. Zhur., 20, 429 (1954). 21. Danlelsonn, B., Svenk. Farm. Tidskr., 58, $11 (1954). 22. Kharasch, M. S. and Isbell, H. S., J. Amer. Chem. Soc., 52* 3059 (1931). 23. Cernatescu, R. and Vascautanu, Mne. S., Ann. Scl. Univ. Jassy, 2£, 356 (1937). 24. Zwikker, J. J. L., Pharm. Weekblad, 975 (1931). 2 5. The British Pharmacopia, The Pharmaceutical Press, London (1958). 26. Fialkov, Ya. A. and Rapaport, L. I., Zhur. Obshchel Khim, 25, 1914 (1955). Through Chemical Abstracts, 20, 3 9 ^ (1956).

27. Levi, L. and Hubley, C. E., Anal. Chem., 28, 1591 (1956). 28. Paulus, W., Mikrcehemie ver. Mikrochim. Acta, 38, 566 (1951). 29* Gomahr, H. and Kresbach, H., Sclentia Pharm., 19, 154 (1951). Through Ehemical Abstracts, 46, 2237(1952). 30. Cooper, P., Pharm. J., 177, 53 (1956). 31* Valov. A., Industr. Engng. Chem., Anal. Ed., 18, 456 (1946). “

32. Cooper, P., Pharm. J., 173, 481 (1954). 33> Halpem, A. and Jones, J. W., J. Amer. Pharm. Assoc., Sci. Ed., 3 8, 352 (1949). 34. Parri, Boll. Chim. Farm., 6£, 4011 (1924). 35. Bodendorf, Arch d. Pharm., 270, 290 (1932). 36. Koppanyl, T., et al., Arch, intern. Pharm. Therap.. 46 76 (1933). “ 118

37* Koppanylj T., et al., Proc. Expt. Biol, and Med.* 31. 373 U933). 38. Koppanyl. T., et al.. J. Amer. Pharm. Assoc.. Scl. Ed., 22, iorJTTI534).

39* Kozelka, F. L. and Tatum. H. J., J. Pham, and Expt. Therap., 52, 54 (1937).

40. L e w y , 0. A., Biochem. J., ^4, 73 (1940). 41. Denis. X. and Lambrechts. A.. Rev. Belg. Scl. Med.. 8. 247 (1940). 42. Iversen. M., Acta Pharmacol., 1, 255 (1945). 43* Turfitt, Q. E., Quart. J. Pham, and Pharmacol., 21, 1 (1948). “ 44. Mattson, L. N. and Holt, W. L., J. Amer. Pham. Assoc., Scl. Ed., £8, 55 (1949).

4 5. Ljungberg, S., Fam. Revy., jjl, 569 (1952).

46. Raventos, J., Brit. J. Pharmacol., 1, 210 (1946). 47. Paulus, W. and Prlbllla, 0., Arzneimlttel-Forschung, 3, 478 (1953). 48. Cohen, E. L., Amer. J. of Pham., 118, 40 (1946).

4 9. Younger, M. J., Ann. Pham. Franc., 1£, 553 (1957). 50. Qomahr, H. and Kresbach, H., Sc lent la Pham., 19, 148 (1951). “ 51. Sefton-Jones, H., British Patent 190,286, October 8, 1921. 52. Berggardh, C., Flnska Kemlstoamfundets Medd., 58, 88 (1949). Through Chemical Abstracts, 45, 102T5I (1951). 53* Schubert, J. and Llndenbaum, A., J. Amer. Chem. Soc., 14,3529(1952). 54. Chauteau, J., Ann.fac. scl. Marseille, 23, 11 (1954). Through Chemical Abstracts, 4g, 14838 TI955). 55. Kohn, K. W., Anal. Chem., 862 (1962). 119 5 6. Quastel, J. H. and Wheatley, A. H. M., Proc. Roy. Soc. Lond., B112, 60 (1932). 57* Greig, M. S., J. Pharmacol, and Expt. Therap., 87, 185 (1946). —

58. Greig, M. E., Ibid., £1, 317 (1947). 59* Eller, J. J. and McEwen, W. K., Arch. Biochem., 20, 163 (1949). ~ 60. Bain, J. A., Federation Proc., 11, 653 (1952).

61. Barbera, G., Olorn. ital. chir., 11, 1500 (1955). 62. Bersin, T., et al., Arzneimlttel-Porsch., _4, 199 (1954).

6 3. Bonner, L., Biochem. J., 5 6, 274 (1954). 64. Anson, M. L. and Mirsky, A. E., J. Gen. Physiol., it, 43 (1930). 6 5. Hogness, T. R., et a l ., J. Bio. Chem., 118, 1 (1937). 66. Prabkin, D. L., ibid., 142, 855 (1942).

6 7. Mori, T., Science (Japan), 1£, 334 (1947). 68. Mori, T., ibid., 18, 55 (1948). 6 9. Drabkin, D. L., J. Bio. Chem., 146, 605 (1942). 70. Keilin, J., Biochem. J., 3X, 281 (1943). 71. Taylor, J. P., J. Bio. Chem., 135, 569 (1940). 72. Davies, T. H., ibid., 13£, 597 (1940). 73. Beetling, C. S., ibid., 135, 623 (1940). 74. Clark, W. M. and Perkins, M. E., ibid., 135, 643 (1940). 75. Domonkos, J. and Huszak, I., Acta Physiol. Acad. Scl. Hung., 4, 225 (1 9 5 3). Through Chemical Abstracts, 48, 1449 (1954). 7 6. Barron, E. S. G., J. Bio. Chem., 121, 285 (1937). 77. Brdlcka, R. and Wlesner, K., Chem. Listy, 40, 66 (1946). 7 8. Lange, W. E. and Foye, W. 0., J. Amer. Pharm. Assoc., Sci. Ed., 4£, 699 (1956). 120 79* Foye, W. 0. and Lange, W. E., J. Amer. Chem. Soc., 76, 2199 (1956). 80. Weinberg, E. D., et al.* J. Amer. Fharm. Assoc., Scl. Ed., Ss, m ( 1 W 81. Morozowlch, W., Ph.D. Dissertation, Ohio State Univer­ sity (1959)* 82. Werner, A. E. A., Soc. Proc. Roy. Dublin Soe., 33, 131 (1943). 83* Van Ultert, W., et al., J. Amer. Chem. Soc., 75, 451 (1953). 84. Kolthoff, I. M. and Furman, N. H., Potentlometrie Titrations, John Wiley and Sons, Inc., New York, 1931.

8 5. Kolthoff, I. M. and Elving, P. J., Treatise on Analytical Chemistry, Vol. 1, Part 1, The Inter science Encyclo­ pedia, Inc., New Yor#, N. Y., 1959, P* 338. 86. Leyda, J. P., Lamb, D. J. and Harris, L. E., J. Amer. Pharm. Assoc., Scl. Ed., 49, 581 (i9 6 0). 8 7. Ooyan, J. E., et al ., ibid., 4£, 627 (i9 6 0). 88. Chatten, L. Q. and Levi, L., Applied Spectroscopy, 11, 177 (19571• ” 8 9. Goldstein, A., Pharmacol. Rev., 1, 102 (1949). 90. Elchman, M. L., Ph.D. Dissertation, Ohio State Univer­ sity (1961). 91. Goldbaum, L. R. and Smith, P. K., J. Pharmacol, and Exper. Therap., Ill, 197 (1954). 9 2. Dox, A. W. and Yoder, L., J. Am. Chem. Soc., 44, 1141 (1922). “ 93. Waddell, W. J. and Butler, T. C., J. Clin. Invest., 3 6, 1 2 1 7 (1957). “ 9 4. Klotz, I. M., in Tfte Proteins, edited by Neurath, H. and Bailey, K., Vol. l, Part B, Academic Press, Inc., New York, N. Y . , 1953, pp. 727-806. 95. Hughes, T. R. and Klotz, I. M., in Methods of Biockemi- cal Analysis, edited by Glick, D.7 VoTT III, tnter- sclence Publishers, Inc., New York, N.Y., pp. 265*300. 121

96. Smith, E. L., Fed. Proc., 8, 581 (1949). 97* Klotz, I. M. and Loh Ming, W. C., J. Am. Chem. Soc., 16, 805 (1954). 98. Andrews, A. C. and Lyons, T. D., Science, 126, 561 (1957). — 99. Kohn, K. W., Nature, 191, 1156 (1961). 100. Possum, J. H., Markunas, C. C. and Riddick, J. A., Anal. Chem., 2£, 491 (1951). 101. Dybling, P., Scand. J. Clin, and Lab. Invest., T§> Suppl. 20, 127 (1955). 102. Rlelly. C. N. and Schmid, R. W., Anal. Chem., 30, 947 (1958). ““ 103. Fischer, H., In Organic Synthesis, edited by N. L. Drake, Vol. 21, John Wiley and Sons, Inc., New York, N. Y., 1941, p. 53. 104. Klotz, I. M., Arch. Biochem., % 109 (1946). 105* Karush, P., J. Amer. Chem. Soc., 72, 2705 (1950). AUTOBIOGRAPHY

I, James Perkins Leyda, was born In Youngstown, Ohio, October 2, 1935* I received my secondary education In Youngstown, graduating from South High School In 1933* I obtained a Bachelor of Science degree in Pharmacy from Ohio Northern University, in 1957, and a Master of Science degree in Pharmacy here at Ohio State University, in 1959* In 1957, I was granted an asslstantship at Hie Ohio State University which I held until i9 6 0. At that time I was granted a pre-doctoral fellowship by the National Institutes of Health (National Institute of Mental Health), division of Health, Education and Welfare. I held this fellowship until I received the degree Doctor of Philosophy.

122