Part One, Synthesis of Acyclic-Sugar Nucleosides; Part Two, Cyclization Of

INFORMATION TO USERS

This was produced from a copy of a document sent to us for microfilming. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the material submitted.

The following explanation of techniques is provided to help you understand markings or notations which may appear on this reproduction.

1. The sign or “target” for pages apparently lacking from the document photographed is “Missing Page(s)”. If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure you of complete continuity.

2. When an image on the film is obliterated with a round black mark it is an indication that the film inspector noticed either blurred copy because of movement during exposure, or duplicate copy. Unless we meant to delete copyrighted materials that should not have been filmed, you will find a good image of the page in the adjacent frame.

3. When a map, drawing or chart, etc., is part of the material being photo graphed the photographer has followed a definite method in “sectioning” the material. It is customary to begin filming at the upper left hand comer o f a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again—beginning below the first row and continuing on until complete.

4. For any illustrations that cannot be reproduced satisfactorily by xerography, photographic prints can be purchased at additional cost and tipped into your xerographic copy. Requests can be made to our Dissertations Customer Services Department.

5. Some pages in any document may have indistinct print. In all cases we have Aimed the best available copy.

University Microfilms International 300 N. ZEEB ROAD, ANN ARBOR, Ml 48106 18 BEDFORD ROW, LONDON WC1R 4EJ, ENGLAND 8022307

Liu, Charng-Ming

PART ONE: SYNTHESIS OF ACYCLIC-SUGAR NUCLEOSIDES. PART TWO: CYCLIZATION OF ACYCLIC-SUGAR NUCLEOSIDES

The Ohio State University PH.D. 1980

University Microfilms International300 N. Zeeb Road, Ann Arbor, MI 48106 18 Bedford Row, London WC1R 4EJ, England

In all cases this material has been filmed in the best possible way from the available copy. Problems encountered with this document have been identified here with a check mark .

1. Glossy photographs .

2. Colored illustrations______

3. Photographs with dark background______

4. Illustrations are poor copy______

5. °rint shows through as there is text on both sides of page______

6. Indistinct, broken or small print onseveral pages \/ throughout

7. Tightly bound copy with print lost in spine _

8. Computer printout pages with indistinct print ______

9. Page(s) ______lacking when material received, and not available from school or author______

10. Page(s) ______seem to be missing in numbering only as text follows______

11. Poor carbon copy______

12. Not original copy, several pages with blurred type ___

13. Appendix pages are poor copy______

14. Original copy with light type______

15. Curling and wrinkled pages______

16. Other ______

University M icrdfiim s international

300 ,\ 3=== =0. ANN A .830P Ml A8106 ‘3131 761-4700 PART ONE: SYNTHESIS OF ACYCLIC-SUGAR

NUCLEOSIDES

PART TWO: CYCLIZATION OF ACYCLIC-SUGAR

NUCLEOSIDES

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

Charng-Ming Liu, B.S.

*****

The Ohio State University

1980

Reading Committee: Approved By

. Professor D. Horton

Professor R.M. Mayers.

Professor L.J. Berliner Adviser Department of Chemistry my parents ACKNOWLEDGMENTS

The author thanks professor Derek Horton for his patience and understanding during the course of this work and for his help in the preparation of this thesis.

The author also wishes to thank the past and present members of the sugar alley for their assistance, especially to Mr. Stan Stavinski for assaying of compounds and for useful discussions.

Appreciation is extended to the Department of

Chemistry for a teaching appointment and to the National

Institute of Health for financial support. VITA

August 31, 1949 . . Born, Hsin-Chu, Taiwan

June, 1972 .... B.S., Department of Chemistry Fu Jen Catholic University Taipei, Taiwan

Sept. 1974-June 1976 Teaching Associate, Department of Chemistry, The Ohio State University, Columbus, Ohio

July 1976-June 1980 Research Associate, Department of Chemistry, The Ohio State University, Columbus, Ohio

FIELD OF STUDY

Major Field: Biochemistry TABLE OF CONTENTS

Page

ACKNOWLEDGMENTS ...... iii

VITA ...... iv

LIST OF TABLES ...... x

LIST OF FIGURES ...... •...... xii

LIST OF CHARTS ...... xiv

PART ONE: SYNTHESIS OF ACYCLIC-SUGAR

NUCLEOSIDES

Chapter

I. INTRODUCTION AND HISTORICAL BACKGROUND . 2

A. Historical Perspectives ...... 2

B. Naturally Occurring Nucleosides . . . 6

1. Nucleosides in Nucleic Acids ... 6 2. Nucleosides in Coenzymes ...... 11 3. Nucleosides in Metabolites .... 17 4. Nucleoside Antibiotics ...... 21 5. Other Naturally Occurring Nucleo sides ...... 21

C. Synthesis of Nucleosides ...... 22

1. Fischer— Helferich Method (Silver Salt Coupling) ...... 22 2. The Davoll— Lowy Method (Mercury Salt Coupling) ...... 24 3. The Hilbert— Johnson Method ... 27 4. The Trimethylsilyl Method .... 27 5. Direct Condensation and Miscel laneous Methods ...... 3 0

D. Synthesis, and Biological and Clinical Applications of Acyclic-Sugar Nucleo side Analogs ...... 33 v Page

II. STATEMENT OF THE PROBLEM ...... 48

III. RESULTS AND D I S C U S S I O N ...... 51

A. Synthesis of Acyclic-Sugar Nucleosides 51

1. Study of the Sugar— Base Coupling Reaction ...... 51 (1) Preparation of aldose diethyl dithioacetals ...... 51 (2) Coupling of sugar and base . . 53 2. Separation of the Epimers from the Product of the Coupling Reaction . 65

B. Characterization of the Products . . 70

1. Position of Attachment of the Sugar Chain to the Purine Base . . 70 (1) Ultraviolet spectra ...... 71 (2) Carbon-13 n.m.r. spectra . . . 77 2. Optical Properties and Chirality at C - l 1 ...... 83 3. N.m.r. Studies and Conformational A n a l y s i s ...... 96 i (1) H-n.m.r. spectroscopy .... 96 (2) C-13 n.m.r. spectroscopy . . . 109 4. Mass Spectra ...... 119

IV. EXPERIMENTAL ...... 132

General Methods ...... 132

Preparation of (IS)-2,3,4,5,6-penta-O- acetyl-1-(6-chloropurin-9-yl)-1-S-ethyl- l-thio-g-glucitol (17b) 134

2,3,4,5,6-Penta-O-acetyl-1-bromo-l-S- ethyl-l-thio-g-mannitol (29) ...... 136

(1R)-2,3,4,5,6-Penta-O-acetyl-l-(6- chloropurin-9-yl)-1-S-ethyl-l-thio-g- mannitol (30a) ...... 136

vi Page (lS)-2,3,4,5,6-Penta-O-acetyl-l-(6- chloropurin-9-yl)-1-S-ethyl-l-thio-g- mannitol (39b) ...... 139

2 , 3 , 4 , 5,6-Penta-O-acetyl-l-(6-ethylthiopurin-9-yl) -1-S-ethyl-l-thio-g-mannitol (3J) (C-l1 epimeric mixture) ...... 140

Preparation of methyl 4,6-O-benzylidene- a-g-glucopyranoside (ji) ...... 142

Preparation of methyl 4,6-0-£-tolylsulfonyl-a-g-glucopyranoside ...... 142

Preparation of g-altrose diethyl dithioacetal ...... 142

2.3.4.5.6-Penta-O-acetyl-a-D-altrose diethyl dithioacetal (3J) 142

2.3.4.5.6-Penta-O-acetyl-l-bromo-l-^- ethyl-l-thio-g-altritol (35) ...... 143

(1R )-2,3,4,5,6-Penta-O-acetyl-l-(6- chloropurin-9-yl)-1-S-ethyl-l-thio-g- altritol (36a) ...... 143

Preparation of methyl 4,6-O-benzylidene- 2-0-£-tolylsufonyl-a-g-glucopyranoside . 145

Methyl 4,6-0-benzylidene-3-deoxy-a-D“ arabino-hexopyranoside ...... 145

Preparation of methyl 3-deoxy-a-g- arabino-hexopyranoside (9) (New Method) 146

Preparation of 3-deoxy-D-arabino-hexose diethyl dithioacetal ...... 146

2,4,5,6-Tetra-0-acetyl-3-deoxy-g- arabino-hexose diethyl dithioacetal . . 147

2,4,5,6-Tetra-0-acetyl-l-bromo-3-deoxy- 1-S-ethyl-l-thio-D-arabino-hexitol (37) 147

2,4,5,6-Tetra-O-acetyl-l-(6-chloropurin-9-yl)-3-deoxy-l-S-ethy1-1-thio-g- arabino-hexitol (38) (C-l epimeric mixture) ...... 147 vii Page

(IS)-2,4,5,6-Tetra-O-acetyl-l-(6-chloropurin-9-yl)-3-deoxy-l-S-ethyl-l-thio-D- arabino-hexitol (38fe) 149

(1R)-1-(6-chloropurin-9-yl)-1-S-ethyl- 1-thio-Q-altritol (4j)a) 150

(1R)-1-(6-chloropurin-9-yl)-1-S-ethyl- 1-thio-g-mannitol (52a) 151

(IS)-1- (6-chloropurin-9-yl)-3-deoxy-l- S-ethyl-l-thio-D-arabino-hexitol (54b) . 151

PART TWO: CYCLIZATION OF ACYCLIC-SUGAR

NUCLEOSIDES

Chapter

I. HISTORICAL BACKGROUND AND INTRODUCTION . 154

A. Cyclization of Acyclic-Sugar Nucleosides ...... 154

B. Fermentation by Acetobacter suboxydans 16 0

II. STATEMENT OF THE PROBLEM ...... 173

III. RESULTS AND D I S C U S S I O N ...... 174

A. Study of the Fermentation by Acetobacter suboxydans ...... 174

B. Carbon-13 N.m.r. Spectral Study and the Equilibrium in Solution of (1S)- 6- (6-chloropurin-9-yl)-6-S-ethyl-6- thio-D-ido-hexulose (43) 183

C. Synthesis of 1,2,3,4-tetra-O-acetyl- 6- (6-chloropurin-9-yl)-6-S-ethyl-6- thio-a,g-D-ido-hexulose (44) .... 190

viii Page

1. Mass Spectrum of 44 191 2. 1H-N.m.r. Spectrum ...... 194

IV. EXPERIMENTAL ...... 197

General Methods ...... 197

Oxidation of (IS)-1-(6-chloropurin-9-yl)- 1-S-ethyl-l-thio-g-glucitol (25) to (IS)- 6- (6-chloropurin-9-yl)-6-S-ethyl-6-thio- D-ido-hexulose (43) ...... 199

1,2,3,4-Tetra-0-acetyl-6-(6-chloropurin- 9-yl)-6-S-ethyl-6-thio-a,g-D-ido-hexulofuranose (44) ...... - ...... 201

BIBLIOGRAPHY ...... 203

ix LIST OF TABLES

PART ONE

Table Page

1. Preparation of 17, 30, 36, and 38 ... . 66

2. The uv absorption maxima of some 7- or 9-alkyl-6-substituted purine ...... 72

3. The uv absorption maxima of acyclic- sugar nucleosides ...... 74

4. The uv absorption maxima of 57 and related compounds ...... 76

5. Comparison of C-13 chemical shifts for 25, 49a, 52a, and 5 4 b ...... 82

6. Determination of the chirality at C-l for 6-chloropurine acyclic nucleosides. . . . 85

7. Conformation and coupling constants of 30a and 30b 104

8. 1H-n.m.r. chemical shift data for compounds 17b, 30a, 30b, 36a, and 3 8 £ ...... 106

9. Proton-proton spin-coupling data for compounds 17b, 30a, 30b, 3£a, and 38b . . 107

10. C-13 chemical shifts in compounds 17a, ljb, 30 a , 30b, 31a, 31b, 36a, 38 a , and 38b . . 112

11. C-13 chemical shifts in compounds 25, 49a, 5^a, and 5 4 b , ...... 113

12. C-13 chemical-shift data for nucleosides 25, 49a, and 5 J a ...... 118

13. Mass-spectral data for compounds 17, 30, 31,y-w ' 36,yv/ ' and 3y—' 8 ...... ~ . 129 14. Mass-spectral data for fragments in series C x from 17, 30, 31, 36, and 3 8 ...... 131

x Page

15. Probable structure of prominent smaller ion found in the mass spectra of 17, 30, 36, and 3 8 ...... ~. . . . 131

PART TWO

1. Grouping of Acetobacter by Frateur ... 161

2. Oxidation of D-glucose diethyl dithio acetal and (1J>)-1- (6-chloropurin-9-yl)-1- £y-ethyl-l-thio-g-glucitol ...... 177

3. C-13 n.m.r. chemical-shift data for solutions of D-fructose and nucleoside analogs 2J and 4 3 ...... 186

4. Mass-spectral data for compound 44 . .. 193

5. Toxicity data for 6-chloropurine nucleo side analogs ...... 196

xi LIST OF FIGURES

PART ONE

Figure Page

1. High pressure liquid chromatography of the epimeric mixture of nucleosides 30 . 69

2. C-13 chemical shift diagram for nucleo sides 17a, 17b, 30a, 30b, 31a, 31b, 36a, 38a, and 38b . . . ~.~...... 7 . . 79

3. UV absorption and CD curve of ljb .... 89

4. UV absorption and CD curve of 30a and 30b 9 0

5. UV absorption and CD curve of 36a .... 91

6. UV absorption and CD curve of 38b and CD curve of 54b ...... 9 2

7. Partial n.m.r. spectrum of 3 0 a ...... 98

8. Partial n.m.r. spectrum of 3 0 b , ...... 100

9. C-13 n.m.r. spectrum of 3 0 a ...... 110

10. C-13 n.m.r. spectrum of 49a ...... Ill

11. Probable fragmentation from molecular ion 122

12. Fragmentation of series ...... 123

13. Fragmentation of series ...... 124

14. Fragmentation of C^ s e r i e s ...... 125

15. Fragmentation of D^ s e r i e s ...... 125

PART TWO

1. "Reversed" nucleosides ...... 159

2. Three tautomeric forms of 4J at equili brium in neutral solution ...... 184

xii Page

3. Probable fragmentation of 4 4 ...... 192

4. Partial n.m.r. spectrum of 44 195

xiii LIST OF CHARTS

PART ONE

Chart Page

1. Structure of Q* nucleosides ...... 10

2. Structure of C o A ...... 12

3. Structures of NAD and NADP ...... 13

4. Structures of riboflavin, FMN, and FAD . 14

5. Structure of Vitamin B^2 ...... 15

6. Structure of biopterin ...... 17

7. Structures of ATP, NDP-sugar, and cytidine nucleotides ..... 18

8. Structures of nucleoside antibiotics . . 22

9. X-ray structures of (IS)-1-(6-chloropurin- 9-yl)-1-S-ethyl-l-thio-g-glucitol and 2,3, 4,5-tetra-O-acety1-1-S-ethyl-l-(1,6-dihydro- 6-thioxopurin-9-yl)-1-thio-D-arabinitol . 3 7

10. Structures of some acyclic-sugar nucleo sides ...... 39

11. Synthesis of acyclic-sugar nucleoside . . 55

12. Synthesis of acyclic-sugar nucleosides (Improved Route) ...... 6 0

13. Synthesis of acyclic-sugar nucleosides (New Method) ...... 6 4

14. X-ray structure of (IS)-1-(6-chloropurin- 9-yl)-1-S-ethyl-l-thio-D-glucitol (25) (ORTEP drawing) ....“...... 95

xiv PART ONE:

SYNTHESIS OF ACYCLIC-SUGAR

NUCLEOSIDES

1 I. INTRODUCTION AND HISTORICAL BACKGROUND

A. Historical Perspectives

In 1909, Levene and Jacobs published a paper entitled, 1 "Uber die Hefe-Nucleinsaure", in which they proposed the term "nucleoside" for the purine- and pyrimidine- containing glycosidic compounds from yeast nucleic acid. This work was a direct outgrowth of the pioneering studies of Miescher and 2 coworkers. In 1868, Miescher had isolated a material from

(1) P. A. Levene and W. A. Jacobs, Ber., 4J2, 2474- 2478 (1909).

(2) F. Miescher, Hoppe-Seyler's Med-Chem. Untersuch., 461 (1871). pus-containing surgical bandages and designated it as

"nuclein" on account of its origin. The work of nuclein 3_g isolation was continued by Miescher and many others.

(3) P. Plosz, Hoppe-Seyler1s Med-Chem. Untersuch., 461 (1871).

(4) V. Luvabin, Hoppe-Seyler' s Med-Chem. Untersuch., 465 (1871). 3

(5) F. Hoppe-Seyler, Hoppe-Seyler1s Med-Chem. Untersuch., 486 (1871).

(6) A. Kossel, Z. Physiol. Chem. , 5^, 152 (1811).

7 Altman introduced the term "nucleic acid" in 1889

to describe the protein-free, acidic constituent of nuclein,

Subsequently, nucleic acids were chemically examined by

hydrolysis. The purine bases, being chemically more stable,

were first isolated; later thymine, cytosine, and uracil

were also identified, mainly by Kossel^and A s c o l i . ^

(7) R. Altmann, Arch. Anat. Physiol., Physiol. A b t . 524 (1889).

(8) A. Kossel, Z. Physiol. Chem., 4, 290 (1880).

(9) A. Kossel and H. Steudel, Z. Physiol. Chem., 3J3, 49 (1903).

(10) A. Ascoli, Z. Physiol. Chem., 3^, 161 (1900- 1901).

11 In hxs studies of yeast and thymus nucleic acid, Kossel

observed a specific difference in their hydrolytic products,

This observation leads to the recognition of the fact that

there were two types of nucleic aicd, DNA and RNA. A sugar was also found to be a component of nucleic acid, and 12-14 Levene and Jacobs showed the sugar moiety of yeast

(11) A. Kossel and A. Neumann, Ber. , 26_, 2753- 2756 (1893).

(12) P. A. Levene and W. A. Jacobs, Ber., 41, 2703-2707 (1908). (13) P. A. Levene and W. A. Jacobs, Ber., 42, 1198-1203 (1909).

(14) P. A. Levene and W. A. Jacobs, Ber., 44, 746-753 (1911). nucleic acid to be D-ribose. However, it was not until 1929 that Levene succeeded in establishing the sugar moiety of thymus nucleic acid as 2-deoxy-D-erythro-pentose by enzymic hydrolysis. Thus, the chemical identification of the products of hydrolysis of the two nucleic acids was completed as tabulated here.

Nucleic Acid from Yeast (RNA) Nucleic Acid from Thymus (DNA) Phosphoric acid Phosphoric acid Adenine Adenine Guanine Guanine Cytosine Cytosine Uracil Thymine D-Ribose 2-Deoxy-D-erythro-pentose

Along with the development of methods for selective hydro lysis, the way in which these components are joined together 15-17 was gradually unraveled. Through the efforts of Levene 18 “2 0 and others, the structure of nucleosides in DNA and

RNA were unequivocally established. In the early 1940's

(15) P. A. Levene and L. W. Bass, "Nucleic Acids" New York, 1931.

(16) P. A. Levene and R. S. Tipson, J. Biol. Chem 97, 491-495 (1932).

(17) P. A. Levene and R. S. Tipson, J. Biol. Chem 104, 385-393 (1934). 5

(18) A. R. Todd, J. Chem. Soc., 647-653 (1946).

(19) B. Lythgoe, H. Smith, and A. R. Todd, J. Chem. Soc., 355-357 (1947).

(20) J. Davoll, B. Lythgoe, and A. R. Todd, J. Chem. Soc., 833-839 (1946) and references cited therein. research on the biological roles of nucleic acids and proteins took place and rapid advances were made, high- 21 lighted by the Watson— Crick double helix model of DNA.

(21) J. D. Watson and F. H. C. Crick, Nature, 171, 737-738 (1953).

Concurrently with these advances in the biochemistry of nucleic acids, serious consideration then began to be given to nucleosides as potential antitumor agents. A number of

"rare" nucleosides had, in fact, been isolated during the early part of the twentieth century but were not then examined for medicine applications. Discoveries of such non-nucleic 22~26 27 acid nucleosides as various antibiotics and coenzymes, together with findings that some purine and pyrimidine analogs

(22) K. G. Cunningham, S. A. Hutchinson, W. Manson, and F. S. Spring, J. Chem. Soc., 2299-2330 (1951).

(23) C. W. Waller, P. N. Fryth, B. L. Hutchings, and J. H. Williams, J. Am. Chem. Soc., 7_5* 2025 (1953).

(24) N. Lufgren and B. Luning, Acta Chem. Scand. , 7, 225 (1953).

(25) H. Yiinsten, K. Ohkuma, and Y. Ishii, J. Antibiotics (Tokyo), Ser. A, 9^ 195 (1956). 6

(26) E. H. Flynn, J. W. Hinman, E. L. Caron, and D. 0. Woolf, Jr., J. Am. Chem. Soc., 7_5/ 5867-5871 (1953).

(27) J. Baddiley, "Nucleic Acids", Vol. 1, New York, N. Y. ,1955, pp. 177-187.

28 29 interfere with cell growth ' and 'proved effective against 30 neoplastic disease, stimulated further interest in nucleosides and nucleoside analogs and brought the study of this group of compounds into major prominence. Today, the term nucleoside has been expanded to include all those compounds, either naturally occurring or synthetic, that contain a heterocyclic base linked to a sugar.

(28) G. H. Hitchings, E. A. Falco, and M. B. Sherwood, Science, 102, 251-252 (1945).

(29) R. 0. Robin, Jr., J. O. Lampen, J. P. English, Q. P. Cole, and J. R. Vaughan, Jr., J. Am. Chem. Soc., 67, 290-294 (1945).

(30) G. P. Rhoads, Consulting Ed., Conference on 6-Mercaptopurine, Ann. N. Y. Acad. Sci.,60, 185 (1954).

B. Naturally Occurring Nucleosides

1. Nucleosides in Nucleic Acids— Today, it is known that the four major nucleoside residues found in all types of RNA are adenosine, guanosine, cytidine, and uridine. The second largest class of nucleic acids, DNA, is composed of four major deoxynucleosides, 1- (2-deoxy-g-D-erythro-pent.o-

furanosyl)adenine, guanine, cytosine, and thymine.

The biological functions of nucleic acids are of profound importance in the process of life. These functions

are associated with such important biological phenomena as

cell replication, differentiation, transformation and evolution. DNA was identified as carrier of genetic infor- 31 mation when Avery and coworkers showed that a non-virulent

strain of pneumococcus Type III can be transformed into a virulent strain in a heritable manner by DNA extracted from virulent pneumococci. The conservation of each DNA strain on replication was later established mainly by the ingenious 32 isotope-labelling experiment of Meselson and Stahl.

(31) 0. T. Avery, C. M. McLeod, and M. McCarty, J. Exptl. Med., 79, 137-158 (1944).

(32) M. Meselson and F. W. Stahl, Proc. Natl. Acad. Sci., 44, 671-682 (1958).

The biological roles of RNA are more complicated.

RNA's are further categorized, based on their functions, into three groups: messenger RNA (m-RNA), ribosomal RNA (r-RNA), and transfer RNA (t-RNA). Together, they play an important role in protein synthesis. It is now well understood that

genetic information is transferred from DNA to the

complementary m-RNA during the process of transcription. The 8 genetic information is then translated from m-RNA, with the help of r-RNA and t-RNA, into protein during the process of protein synthesis. The chemical basis of these processes is 33-41 the phenomenon of specific pairing of complementary bases through hydrogen bonding. In DNA replication, the base pairs are adenine— thymine (A-T) and guanine— cytosine (G-C), while in transcription and translation, in which RNAs are involved, the uracil residue is complementary to an adenine residue

(A-U).

(33) R. M. Hamlin, Jr., R. C. Lord, and A. Rich, Science, 148, 1734-1737 (1965).

(34) Y. Kyogoku, R. C. Lord, and A. Rich, Science, 154, 518-520 (1966).

(35) R. R. Shoup, H. T. Miles, and E. D. Becker, Biochem. Biophy. Res. Comm., Y3, 194-201 (1966).

(36) L. Katz and S. Penmann, J. Mol. Biol., 15, 220-231 (1966).

(37) K. H. Scheit, Angew. Chem., 79_, 90 (1967).

(3 8) F. S. Mathews and A. Rich, Nature, 201, 179-180 (1964).

(39) R. F. Steward and L. H. Jensen, J. Chem. Phys., 40, 2071-2075 (1964).

(40) J. Iball and H. R. Wilson, Nature, 198, 1193-1195 (1963).

(41) B. Pullman, P. Claverie, and J. Caillett, Proc. Natl. Acad. Sci., 5j5, 904-912 (1966).

.In addition to the eight major components, certain other nucleosides may also present in nucleic acids. 42 Since Hotchkiss detected the first modified component of

a nucleic acid, 2'-deoxy-5-methylcytosine in a sample of

calf thymus DNA, now over 40 compounds of this type have

been described. Usually, these so-called "minor components"

are formed as derivatives of the major components by such

simple chemical reactions as alkylation, hydrogenation,

and thiation, and are most frequently found in t-RNA.

(42) R. D. Hotchkiss, J. Biol. Chem., 175, 315- 332 (1948).

Examples are pseudouridine, thiouridine, and derivatives of

cytidine methylated at positions 3 or 5. It has been

suggested that these bases may be essential for generating

specific, three-dimensional configurations of t-RNA by preventing base-pairing, and may also render t-RNA less

susceptible to the enzymic hydrolysis by nucleases. Also,

in the DNAs of the T-even phages (T , T , and T ), 2 4 6 2'-deoxycytidine is completely substituted by 2'-deoxy-5- 43 hydroymethylcytidine, and in the DNA of phage SP8 and 44 PBS-1, thymine is replaced by 2 1-deoxyhydroymethyl uridine 45 and 2 1-deoxyuridine, respectively. There are also

(43) G. R. Wyatt and S. S. Cohen, Nature, 170, 1072-1073 (1952).

(44) R. G. Kallen, M. Simon, and J. Marmur, J. Mol. Biol., 5, 248-250 (1962). 10

(45) I. Takahashi and J. Marmur, Nature, 197, 794-795 (1963).

nucleosides structurally more complex than those just men

tioned. A recent example is the isolation of Q* nucleosides 46 from rabbit liver t-RNA. The Q* nucleosides consist of a mixture of a major and a minor component. They are the first

RNA nucleosides found to contain a sugar moiety in the side chain. Their structures are depicted in Chart 1.

HO- R:a-D-mannoovranosvl (v75%) RO CH

Hl4 HO HO

H HN CH20H o

R:S-D-galatopyrancsyl (^25%) HO OH

Chart 1. 11

(46) H. Kasai, K. Nakanishi, R. D. MacFarlane, D. F. Torgerson, Z. Ohashi, J. A. McCloskey, H. J. Gross, and S. Nishim.ura, J. Am. Chem. Soc., 98, 5044-5046 (1976)

The subject of modified nucleosides has been reviewed by 47 Hall and should be consulted for further detail.

(47) R. H. Hall. "The modified Nucleosides in Nucleic Acids", Columbia University Press, New York and London, 1971.

2. Nucleosides in Coenzymes—

(1) Coenzyme A (Co A)— Coenzyme A, coenzyme of

acetylation, is the most important acyl group-transfer

coenzyme in living systems. It functions as an acyl carrier

in metabolism according to the scheme shown here.

Acyl Donor*- HSCoA Acylated Acceptor

Donor Acyl-SCoA — — Acceptor

The structure of CoA (chart 2) was elucidated in 48 the laboratories of Lipmann, Snell, and Baddiley, mainly

by specific enzymic degradation. Total synthesis was

(48) L. Jaenicke and F. Lynen, The Enzymes, Vol. 3, Chapt. 11, P. D. Boyer, H. Lardy, K Myrback, Eds, Academic Press, New York, 1960. 12 accomplished by reaction of the key intermediate, adenosine

2',3'-cyclic phosphate 5 '-phosphomorpholidate (formed by treatment of the nucleoside phosphate with morpholine and

N,N'-dicyclohexylcarbodiimide) with D-pantetheine

4 1-phosDhate.

H O H 0 OH CH-a 0 0 II 1 II I I O Q || VA HS-CHjCHg-N-C-CHgCHgN-C-CH-C-CHg-O-f-O-Ij'-OCHg N|

0 OH _0-P=0 OH

(2) Nicotinamide dinucleotides— These pyridine nucleotides were the first coenzymes to be recognized.

Nicotinamide adenine dinucleotide (NAD) and nicotinamide adenine dinucleotide phosphate (NADP), whose structures are shown in chart3, are involved in a large number of oxidation

— reduction reactions catalyzed by dehydrogenases. The site of reversible oxidation— reduction of the coenzyme was located at the 4 position of the nicotinamide moiety by 49 Colowick et al.

(49) M. E. Pullman, A San Pietro, and S. P. Colowick, J. Biol. Chem., 206, 129-141 (1954). 13

N H Adenine

OH OH This hydroxyl group is HO—P = 0 esterified with phosphate in NADP.

HO—P = 0 H I—CONH,HG .CH Nicotinamide CHj.O.

OH OH

Chart 3

(3) Flavin coenzymes— Among these group of coenzymes are riboflavin (Vitamin B2), flavin adenine dinucleotide (FAD) and riboflavin mononucleotide (FMN).

The structure of riboflavin (chart 4) was established by synthesis5^'^ as 6,7-dimethyl-9-(1-deoxy-D-ribitol-l-yl) isoalloxazine. FMN and FAD (chart 4) are structurally

(50) P. Karrer, K. Schopp, and F. Benz, Helv. Chim. Acta, 18, 426-429 (1935).

(51) R. Kuhn, K. Reinemund, F. Weygand, and R. Strobele, Ber., 68, 1765-1774 (1935). closely related to riboflavin. They are all involved in biological oxidation reactions through the capability of the isoalloxazine ring of undergoing stepwise, reversible oxidation and reduction,

H 0 R=H (Riboflavin) I 1 c h ,— c^°^cr''f\ : ''CS 'ni I II I I R=P03H (FMN) N' I 0 0 f* n n HCOH R=-P-0-P-Adenosine (FAD) 6.7-Dimethylisoalloxazine I «■ ■ D-ribitol i i HCOH I 0. 0 _ HCOH I C H .0 R

Chart 4

(4) Vitamin — This anti-pernicious anemia factor was not isolated pure until 1948. The structure determination was nearly as difficult as its isolation. 52 Chemical studies, together with X-ray diffraction studies, led to the formulation of as shown in chart 5.

(52) D. C. Hodgkin, J. Hamper, J. Lindsey, M. MacKay, J. Pickworth, J. H. Robertson, C. B. Shoemaker, J. G. White, R. J. Prosen, and K. N. Trueblood, Proc. Roy. Soc. (London), Ser. A, 242, 228-263 (1957). 15

OH H

NH,

,N=> ch ; Co The corrin ring system

NHCOCHjCHj CH, CH, I CH,— CH.. .CH; ,OH

CH; OH

5.6-Dimethylbenzimidazole ribonucleoside

Chart 5

The unique feature of the nucleoside involved is that the 5,6-dimethylbenzimidazole ring is attached to ribose through an a-glycosidic linkage. Vitamin catalyzes three different types of rearrangement, namely, dehydration, amino-group migration, and some quite unusual carbon-chain rearrangements.

Vitamin B ^ itself is not active as coenzyme, but there exists a coenzymically active derivative of adenosyl- cobalamin, in which the cyanide ion is replaced as a cobalt

ligand by the 5'-deoxyadenosyl group. In their study with a 16 53 series of analogues, Fukui et al. concluded that both ribose and adenine moiety participate in enzyme— coenzyme interaciton, involving not only the binding of the apoenzyme but the activation of the carbon— cobalt bond.

(53) T. Toraya, K. Ushio, S. Fukui, and H. P. C. Hogenkamp, J. Biol. Chem., 252, 963-970 (1977).

(5) Biopterin— Biopterin is a prosthetic group to the enzyme phenylalanine hydroxylase, found in mammalian 54 liver. Chemical synthesis confirmed that the structure is

2-amino-4-keto-6-(L-erythro-1,2-dihydroxypropyl)pterin

(Chart 6). Biopterin stimulates growth in the flagellate

(54) E. L. Patterson, R. Milstrey, and E. L. R. Stokstad, J. Am. Chem. Soc., 1^, 5868-5871 (1950).

Crithida fasciculata, and probably plays an important role in the process determining the transformation of a bee larva 55 into a queen. It acts as cofactor m hydrogen transfer via 5 6 the system tetrahydrobiopterin £dihydrobiopterin.

(55) H. Rembold, Vitamin Hormone, 23, 359-382 (1965) .

(56) S. Kaufman, J. Biol. Chem., 236, 804-810 (1961) . 17

c— C-CH HN OH OH -

H2N Chart 6

3. Nucleosides in Metabolites—

(1) Adenosine triphosphate (ATP)— ATP is the primary carrier of chemical energy. The free energy derived from hydrolysis of ATP is available, in coupled enzymic reactions, to drive an otherwise unfavorable reaction toward near completion. The known ATP-dependent biochemical processes include, among many others, muscle contraction, photosynthesis, bioluminescence, and nucleic acid and protein synthesis.

The structure has been determined5”7' 5® as depicted in Chart 7.

(57) K. Lohmann, Biochem. Z., 233, 460-469 (1931); 254, 381-397 (1932).

(58) B. Lythgoe and A. R. Todd, Nature, 155, 695- 696 (1945).

(2) Nucleoside glycosyl diphosphates (NDP-sugars)— 59 First isolated by Park from penicillin-treated bacteria, this group of compounds play important role in carbohydrate metabolism, especially in the biosynthesis of polysaccharides and cell walls. 18

r -o— p^o General structure 0 -o— p*o o p a Base 1 o NH, \ OH OH — yH o / — p—o— I p— I O— CHj^O^ - o —p = o N. SUGAR N j L » 11 0 HC. I OH O O 1 \ «*CH hV ^ h CH, N OH OH l / N

OH OH

ATP NDP-Sugars

NH, R I CH, N^C^CH CH, 0 = C . .CH R= N(CH3)3 Choline I OH OH I I I O—P—O—P—O— CH, R= NH* Ethanolamine

" " h \ ^ / h

OH OH

Cytidine Nucleotides

Chart 7 19

(59) J. T. Park, J. Biol. Chem., JL94, 877-904 (1952).

The role of NDP-sugars in enzymic catalysis is

twofold: first, these compounds are involved in various

glycosyl-transfer reactions, in which the sugar residue is

transferred to an acceptor. Secondly, the sugar residue is

chemically transformed into another sugar. Examples include

the interconversion of UDP-D-glucose and UDP-g-galactose as

catalyzed by UDP-D-glucose-4-epimerase, and the formation of

GDP-^-fucose from GDP-D-mannose by sequence: GDP-D-mannose -»■

GDP-"4-keto-6-deoxy-D-mannose" ■* GDP-"4-keto-6-deoxy-L-

glucose" GDP-L-fucose. The general structure for NDP-sugars

is shown in Chart 7.

(3) Cytidine nucleotides— Cytidine-5'-(choline

pyrophosphate) (CDP-choline) and cytidine-5'-(ethanolamine 6 0 pyrophosphate) (CDP-ethanolamine), are essential inter mediates in the synthesis of lecithins. Their structures are

shown in Chart 7.

(60) E. P. Kennedy and S. B. Weiss, J. Am. Chem. Soc., 77, 250-251 (1955); J. Biol. Chem., 222, 193-214 (1956).

(4) S-Adenosy1-methionine (Ado-Met; SAM)—

S-Adenosy1-methionine, classified as a "high-energy" meta bolite, is synthesized biologically (Eq. 1) as an activated

form of methionine. It is capable of 0-, N-, and C-methylation 61 via methy1-cation transfer, and is the primary methyl-

group donor in such biological transmethylation reactions

as the methylation of polynucleotides to form minor nucleosides

in the polymer sequence, and specific methylation of lysine,

arginine, histidine, and the dicarboxylic amino-acid side-

chains of certain proteins.

(61) For a review, see "The Biochemistry of Adenosy1-methionine", F. salvatore, E. Borek, V. Zappiz, H. G. Williams-Ashman, F. Schlenk, Eds., Columbia University Press, New York, 1977. 21

4. Nucleoside Antibiotics— A fourth class of

nucleoside derivatives is the nucleoside antibiotics. These

have proved useful as biological probes in cellular research

and as drugs in clinical application because of their close

structural relationship to nucleoside occurring naturally

in nucleic acids. To illustrate the wide diversity of different structural features of compounds belonging to this group, a few examples are discussed here. There are those nucleosides, such as cordycepin and puromycin, in which modification occur in the ribose residue, or, in which the

ring-oxygen atom in the ribofuranosyl moiety has been replaced by a methylene group, as in aristeromycin. Others are modified

in the aglycon, as in 5-azacytidine, toyocamycin, and showdomycin. The structures of these compounds are shown in

Chart 8.

5. Other Naturally Occurring Nucleosides— In addition to the nucleosides just discussed, a number of nucleosides, whose biological function does not fit into the previous classification, has been isolated. Spongosine, shown to be

6-amino-2-methoxy-9-6-D-ribofuranosyl-purine, was first isolated from sponges of the species Cryptotethia crypta by 6 2 Bergmann and Feeney. Crotonside (isoguanosine, 6-amino-2- hydroxy-9-B-D-ribofuranosylpurine) was isolated from seeds

(62) W. Bergmann and R. J. Feeney, J. Am. Chem. Soc., 72, 2809-2810 (1950). 22

H3Cv CH3 NH2

HOHoC o HOHoC

o=c-ch-ch2-c 6h4

|vIH2 0CH3 P'

PUROMYCIN CORDYCEPIN

NH 0 II /-NlNH

0 HOHoC

HO OH HO OH ARISTEROMYCIN SHOWDOMYCIN

nh2 cn n h 2

N ^ N

0 LJ h o h 2 c 0 H °H 2 C 0

Q HO OH HOH OH TOYOCAMYCIN 5-AZACYTIDINE

Chart 8 23 of Croton tiglium and is a naturally occurring structural analog of guanosine. Oxoformycin B (3-8“D-ribofuranosyl- pyrazolo[4,3d]-4-(H),6(H)-5,7-pyrimidone) has been isolated from two strains of Norcadia interforma and found in the 63 64 urine of mice and rabbits after injection of the anti biotics formycin or formycin B. It is assumed to be an end-product of detoxification.

(63) M. R. Sheen, B. K. Kim, and R. E. Parks, Jr., Proc. Am. Assoc. Cancer Res., j), 63 (1968).

(64) M. Ishizuka, T. Sawa, G. Koyama, T. Takeuchi, and H. Umezawa, J. Antibiotics, 21A, 1-5 (1968).

C. Synthesis of Nucleosides

1. The Fischer— Helferich Method (Silver Salt

Coupling)— The first report of a laboratory synthesis of a nucleoside was made by Fischer and Helferich in 1914 when they successfully coupled the silver salt of 2,8-dichloro- adenine with tetra-O-acetyl-D-glucopyranosyl bromide and deacetylated the product to give 7-B-D-glucopyranosyltheo- 6 5 phylline. The use of other purine derivatives led to 66-68 9-glycosylpurines. This same method, however, is not applicable to the coupling of pyrimidines because of lactam— 2 H -*■ 9H lactim tautomerism (— c— A — C=N-), and the procedure has progressively been superseded by the Davoll— Lowy method. 24

(65) E. Fischer and B. Helferich, Ber. , 4_7, 210- 233 (1914).

(66) G. M. Hitchings and L. Goodman, U. S. Pat. 3074929 (1963).

(67) J. M. Gulland, E. R. Holiday, and T. F. Macrae, J. Chem. Soc., 1639-1644 (1934).

(68) J. M. Gulland and L. F. Story, J. Chem. Soc., 259-261 (1938).

2. The Davoll— Lowy Method (Mercury Salt Coupling)—

Essentially a modification of the Fischer— Helferich method, the Davoll— Lowy method was first developed by employing monochloromercuri derivatives of purines instead of the corresponding silver salts. A wide range of purine nucleosides and nucleoside analogs has been synthesized by this 69-74 procedure. In several cases, higher yields have been

(69) M. L. Wolfrom and P. McWain, J. Org. Chem., 30, 1099-1101 (1965).

(70) J. Farkas and F. £3orm, Collect. Czech. Chem. Comm., 32, 2663-2667 (1967).

(71) E. Walton, S. R. Jenkins, R. F. Nutt, M. Zimmermann, and F. W. Holly, J. Am. Chem. Soc., 88, 4524-4525 (1966).

(72) D. Horton and C. G. Tindall, Jr., Carbohydr. Res., 17, 240-244 (1971).

(73) D. Horton and S. S. Kokrady, Carbohydr. Res., 24, 333-342 (1972).

(74) J. Defaye and Z. Machon, Carbohydr. Res., 24, 235-245 (1972). 25 obtained by using O-benzoyl blocking groups instead of 7 c 76 O-acetyl groups. ' Fox and coworkers extended the method to the synthejsis of pyrimidine nucleosides and improved it by devising a procedure for the preparation of chloromercuri 77 . . . derivatives. In most of the syntheses of pyrimidine nucleosides by the mercury salt method, N-glycosylpyrimidines are formed, although there are a few examples where O-glycosyl 78 79 derivatives ' have been produced; however, these could be rearranged to N-glycosylpyrimidines by treatment with mercury salts.

(75) B. R. Baker, J. P. Joseph, and R. E. Schaub, J. Am. Chem. Soc., 11_, 5905-5910 (1955).

(76) H. M. Kissman, C. Pidacks, and B. R. Baker, J. Am. Chem. Soc., 11_, 18-24 (1955).

(77) J. J. Fox, N. Yung, J. Davoll, and G. B. Brown, J. Am. Chem. Soc., 78, 2117-2122 (1956).

(78) C. Ukita, H. Hayatsu, and Y. Tomita, Chem. Pharm. Bull. (Tokyo), 11, 1068-1073 (1963).

(79) M. L. Wolfrom and H. B. Bhat, J. Org. Chem., 2757-2759 (1967).

The anomeric configuration of the product of coupling of a per-O-acetylated glycosyl halide is observed to be B in 80 81 most cases. The "trans-rule", ' which states that

"condensation of a heavy metal salt of a purine (or pyrimidine) with an acylated glycosyl halide will form a nucleoside with a C-l— C-2 trans configuration in the sugar moiety; regardless of the original configuration at C-l— C-2", was formulated to 26 sum up the observation. Anchimeric assistance of the 81 82 2-acyloxy group ' is the common explanation for the observed stereochemistry.

(80) R. S. Tipson, J. Biol. Chem., 130, 55-59 (1939) .

(81) B. R. Baker, J. P. Joseph, R. E. Schaub, and J. H. Williams, J. Org. Chem., _19, 1780-1785 (1954).

(82) B. R. Baker, in "Ciba Foundation Symposium on the Chem. and Biol, of Purines", Ed. by G. E. W. Wolstenholme and C. M. O'Connor, Little, Brown and Co., Boston, 1957, p. 120.

A modification of Davoll— Lowy method was developed 8 3 by Baker and coworkers during their synthetic studies on puromycin. The modification, known as the titanium tetra chloride procedure, in which acylated sugar is converted into its chloride by TiCl^ and then coupled to the chloromercuri salt of a heterocyclic base, makes possible the coupling of sugars that are unstable under the usual conditions of glycosyl halide synthesis. This procedure was extensively used in the synthesis of various 3'-amino- 83 8 4 3'-deoxy sugar nucleosides ' and was later applied to couple O-isopropylidene sugar derivatives with 6-benzamido- chloromercuripurine.

(83) B. R. Baker, R. E. Schaub, J. P. Joseph, and J. H. Williams, J. Am. Chem. Soc., 77, 12-15 (1955). 27

(84) B. R. Baker, R. E. Schaub, and H. M. Kissman, J. Am. Chem. Soc., 11_, 5911-5915 (1955).

(85) L. M. Lerner and Y. Y. Cheng, Carbohydr. Res., 14, 297-303 (1970).

3. The Hilbert--Johnson Method— Hilbert and Johnson,

recognizing the lactam— lactim tautomerism as being res

ponsible for the failure of the Fischer— Helferich attempt

to extend their silver— salt method to pyrimidines, converted

pyrimidines into 2,4-dialkylpyrimidines and successfully

coupled them with per-O-acylated glycosyl halides to yield 8 6 the desired N-l coupling-product. Removal of substituents

from the resulting blocked derivative afforded the nucleoside.

A number of pyrimidine nucleosides have been synthesized by 87 88 this procedure, ' and the preponderant anomer is generally Q88 ,89 ii m -ni ii 80,81 8 as predicted by "Trans Rule". '

(86) G. E. Hilbert and T. B. Johnson, J. Am. Chem. Soc., 52, 2001-2007; 4489-4494 (1930).

(87) J. J. Fox and I. Wempen, Adv. Carbohydr. Chem., 14, 283-380 (1959) and references therein.

(88) J. Pliml and M. PrystaS, Adv. Heterocycl. Chem., 8, 115-142 (1967).

(89) M. Prystas and F. Sorm, Collect. Czech. Chem. Comm., 29, 2965-2970 (1964).

4. Trimethylsilyl Method— The Hilbert— Johnson method was modified in an improved procedure by using trimethylsily-

lated derivatives of a purine or pyrimidine. It was first used 28

90 by Birkofer for the synthesis of 3-a-g-ribofuranosylunc acid. Nishimura and coworkers prepared various trimethylsilyl derivatives of purine and pyrimidine, which were subsequently 91-94 used for the synthesis of a number of nucleosides.

(90) L. Birkofer, A. Ritter, and H. P. Kiihlthau, Angew. Chem., 75, 209-210 (1963).

(91) T. Nishimura, B. Shimizu, and I. Iwai, Chem.- Pharm. Bull. (Tokyo), 3JL, 1470-1477 (1963).

(92) T. Nishimura, and I. Iwai, Chem. Pharm. Bull. (Tokyo), 12, 352-356; 357-361 (1964).

(93) T. Nishimura, B. Shimizu, and I. Iwai, Chem. Pharm. Bull. (Tokyo), 12, 1471-1478 (1964).

(94) T. Nishimura and B. Shimizu, Agr. Biol. Chem., 28, 224-229 (1964).

In case of pyrimidines, the trimethylsilyl groups are attached to 0-2 and 0-4, or 0-2 and N-4, whereas in case of purines, bis and tris(trimethylsilyl) derivatives are used having trimethylsilyl groups attached to N-9 and to the oxo or amino groups in the ring. One apparent advantage is the direct applicablity of this procedure for synthesis of cytosine nucleosides, which had previously been prepared by aminolysis of the corresponding preformed, uracil nucleo-

(95) E. Wittenburg, Z. Chem., 4_, 303-304 (1964).

(96) M. W. Winkley and R. K. Robins, J. Org. Chem., 33, 2822-2827 (1968). 29

In contrast to the Hilbert— Johnson method, the products obtained by this method are generally obtained as mixture of a and 6 anomers, although condensations that gave exclusively a or g nucleosides has been 9 7 reported. For purines, a significant proportion of

7-substituted nucleoside has been observed under certain conditions, although this product could be isomerized by 9 8 heating to give the 9-substituted nucleoside.

(97) T. J. Bardos, M. D. Kotick, and C. Szantay, Tetrahedron Lett., 1759-1764 (1966).

(98) B. Shimizu and A. Saito, Agr. Biol. Chem., 33, 119-121 (1969).

Various reaction conditions have been examined; Birkofer used silver perchlorate as catalyst whereas Nishimura and coworkers conducted the fusion in an inert atmosphere. 99 Wittenburg found that the use of HgO/HgBr2 as a Lewis acid catalyst and benzene as inert solvent gave a better result. These "Wittenburg conditions", were later applied to the synthesis of 6,7-diphenyl-l-g-D-ribofuranosyl- , . 100 l u m a z m e .

(99) E. Wittenburg, Chem. Ber., 101, 1095-1114 (1968).

(100) G. Ritzmann and W. Pfleiderer, Chem. Ber., 106, 1401-1417 (1973). 30

101 More recently, Vorbriiggen modified the method so that 1-O-acyl-glycoses rather than the (less stable) glycosyl halides could be used. The method, named the "SnCl^— catalyzed silyl Hilbert— Johnson reaction", is similar to the titanium tetrachloride modification of the Davoll—

Lowy method. Although other Friedel— Crafts catalysts and solvents are also effective, the combination of SnCl^ and

1,2-dichloroethane has been shown to be the most simple and practical. This procedure was originally developed for the 102 synthesis of such pyrimidine nucleosides as 6-methyluridme 103 and 5-azacytidine. , and was later extended to the synthesis of4T purine nucleosides.1 -A 104

(101) U. Niedballa and H. Vorbriiggen, J. Org. Chem., 39, 3654-3660 (1974).

(102) U. Niedballa and H. Vorbriiggen, J. Org. Chem., 39, 3660-3663 (1974).

(103) U. Niedballa and H. Vorbriiggen, J. Org. Chem., 39, 3672-3674 (1974).

(104) F. W. Lichtenthaler, P. Voss, and A. Heerd, Tetrahedron Lett., 24, 2141-2144 (1974).

5. Direct Condensation and Miscellaneous Methods—

The direct condensation of base and sugar, if effective, is the simplest and shortest of all condensation because there is no activation involved. One such attempt was reported by 105 Schramm in a synthesis of 2 '-deoxyadenine, with a claimed

3 0% yield, by treating 2-deoxy-D-erythro-pentose with 31 adenosine in the presence of "polyphosphoric ester". However, other workers^were unable to confirm the result. A pro cedure, often referred to as the "-fusion method", was 107 introduced by Helferich and developed by Sato and co- 108 workers. The latter investigators fused, in the presence

(105) G. Schramm, H. Grotsch, and W. Pollmann, Angew. Chem., _73, 619 (1961).

(106) J. A. Carbon, Chem. Ind. (London), 529 (1963).

(107) B. Helferich and E. Schmitz-Hillebrecht, Ber., 66, 378-383 (1933).

(108) T. Shimadate, Y. Ishido, and T. Sato, Nippon Kagaku Zasshi, 82, 938-940 (1961); T. Sato, T. Shimadate, and Y. Ishido, ibid. ,8_1, 1442 (1960). of an acid catalyst, a number of peracylated sugars with 2,6- dichloropurine and other chloropurines to give the protected nucleosides. A variety of acid catalysts have been used, including aluminum chloride, sulfur trioxide, magnesium sulfate, p-toluenesulfonic acid,and chloroacetic 111 acid. in general, 9-substituted isomers are obtained, with the exception that 7-isomer is the exclusive product in the 112 case of 2,8-dichloropurine.

(109) Y. Ishido, A. Hosono, K. Fujii, Y. Kikuchi, and T. Sato, Nippon Kagaku Zasshi, _87, 752 (1966); Chem. Abstr. 65, 17034 (1966).

(110) K. Onodera and H. Fukumi, Agr. Biol. Chem., 27, 526-529; 864-869 (1963). 32

(111) M. J. Robins and R. K. Robins, J. Am. Chem. Soc., 87, 4934-4940 (1965).

(112) K. Antonakis and M. J. Arror, C. R. Acad. Sci. Ser. C, 272, 1982-1984 (1971).

There are other methods that involve direct conden sation of a glycosyl halide with a purine or pyrimidine base.

Yamaoka and coworkers carried out the reaction in hot nitromethane with mercuric cyanide as acid scavenger, and prepared g-glucopyranosyl and D-ribofuranosyl derivatives of adenine, theophylline, 2,6-dichloropurine, 6,8-dichloro- 113 purine, and benzimidazole with high yield. Fletcher and 114 coworkers reported a simple way to synthesize "cis- nucleosides", in which they coupled per-O-benzylated g-arabinofuranosyl chloride with 6-benzamidopurine in dichloromethane without any catalyst except for a molecular sieve added as acid scavenger. The reaction was carried out for 1 week at room temperature. This method was also applied in an improved synthesis of 9-(B-D-arabinofuranosyl)-2- 115 chloroadenine. Another method involved use of the sodium salt of a pyrimidine or purine base, obtained by the action of sodium methoxide, hydride, or hydroxide, to couple with 116 117 a sugar epoxide or iodoalkylether.

(113) N. Yamaoka, K. Aso, and K. Matsuda, J. Org. Chem., 30, 149-152 (1965).

(114) C. P. J. Glaudemans and H. G. Fletcher, Jr., J. Org. Chem., 28, 3004-3006 (1963). 33 (115) F. Keller, I. J. Botvinick, and J. E. Bunker, J. Org. Chem., 32, 1644-1646 (1967).

(116) A. Holy and F. Sorm, Collect. Czech. Chem. Comm., 34, 3383-3401 (1969).

(117) G. E. Keyser, J. D. Bryant, and J. R. Barrio, J. Org. Chem., 44_, 3733-3734 (1979).

87 118-121 Numerous reviews ' cover the synthesis of nucleosides, and to those the reader is directed for a

comprehensive account.

(118) J. A. Montgomery and H. J. Thomas, Adv. Carbohydr. Chem., 17_, 311-369 (1962).

(119) A. M. Michelson, "The Chemistry of Nucleosides and Nucleotides", Academic Press, New York, 1968, Chapter 2.

(120) L. Goodman and C. A. Dekker in "The Carbohy drate" Vol. IIA, Ed. by W. Pigman and D. Horton, Academic Press, New York, 1970, Chapter 29.

(121) "Nucleic Acid Chemistry", Part I and II, Ed. by L. B. Townsend and R. S. Tipson, Wiley— Interscience, New York, 1978, Topic III.

D. Synthesis, and Biological and Clinical Applications

of Acyclic-Sugar Nucleoside Analogs

The first nucleoside derivative in which the sugar 122 portion is acyclic was prepared by Wolfrom and coworkers.

In this first synthesis, D-galactose diethyl dithioacetal pentaacetate was converted into the reactive halide 2,3,4,5,6- penta-O-acetyl-l-bromo-l-O-methyl-D-galactitol, which was 34 then condensed with 6-acetamido-9-chloromercuripurine. The

product was O-deacetylated to yield 1-(adenin-9-yl)-1-S-ethyl-

1-thio-D-galactitol as a pair of C-l' epimers. Both products

(122) M. L. Wolfrom, A. B. Foster, P. McWain, W. von Bebenburg, and A. Thompson, J. Org. Chem. , _26, 3095- 3097 (1961). were obtained crystalline and designated first form and

second form. Since then, many other acyclic-sugar nucleosides have been prepared. The heterocyclic bases used have included *7*3 1 90-1 Ofi such purine derivatives as adenine, ' 6-chloro- . ^ 127-130,134 _ . . 127,129,134 , purine, ' 6-mercaptopunne, ’ ' and

6-(mathylthio)purine as well as such pyrimidine derivatives

as thymine,126,131,132 uracil,128,137 5-fluorouracil,131•138

and cytosine.133,135" 137

(123) M. L. Wolfrom, P. McWain, and A. Thompson, J. Org. Chem., 27, 3549-3551 (1962).

(124) M. L. Wolfrom, H. G. Garg, and D. Horton, J. Org. Chem., 29, 3280-3283 (1964).

(125) M. L. Wolfrom, H. G. Garg, and D. Horton, J. Org. Chem., 30, 1096-1098 (1965).

(126) M. L. Wolfrom, W. von Bebenburg, R. Pagnucco, and P. McWain, J. Org. Chem., 3_0, 2732-2735 (1965).

(127) M. L. Wolfrom, P. McWain, H. B. Bhat, and D. Horton, Carbohydr. Res., 2_3, 296-300 (1972); D. Horton, Pure Appl. Chem., £2, 301-325 (1975).

(128) D. C. Baker and D. Horton, Carbohydr. Res., 69, 117-134 (1979). 35

(129) K. C. Blieszner, D. Horton, and R. A. Markovs, Carbohydr. Res., 8_0, 241-262 (1980).

(130) D. Horton and R. A. Markovs, Carbohydr. Res., 80, 356-363 (1980).

(131) M. L. Wolfrom, H. B. Bhat, P. McWain, and D. Horton, Carbohydr. Res., 2_3, 289-295 (1972).

(132) M. L. Wolfrom and P. J. Conigliaro, Carbohydr. Res., 20, 369-374 (1971).

(133) D. Horton and S. S. Kokrady, Carbohydr. Res., 80, 364-374 (1980).

(134) D. C. Baker, K. Blieszner, and D. Horton, "Nucleic Acid Chemistry", Vol. 2, L. B. Townsend, R. S. Tipson, Eds., Wiley, New York, 1978, pp. 627-637.

(135) D. Horton, S. S. Kokrady, "Nucleic Acid Chemistry", Vol. 1, L. B. Townsend and R. S. Tipson, Eds., Wiley, New York, 1978, pp. 267-272.

(136) D. C. Baker, S. S. Kokrady, and D. Horton, Ann. N. Y. Acad. Sci., 255, 131-150 (1975).

(137) S. S. Kokrady, Ph.D. Dissertation, The Ohio State University (1972) ; Diss. Abstr. Int. B, 3_4, 597-B (1973).

(138) R. A. Markovs, Ph.D. Dissertation, The Ohio State University (1975); Diss. Abstr. Int. B, 3j6, 3946-B (1976).

Sugars that have been coupled to these bases

including g-galactose,122'123'126'131 g-glucose,73'122-125'127'131'132 2-amino-2-deoxy-g-glucose,132

2-deoxy-2- (2 , 4-dinitroanilino) -D-glucose,134 ' 13~*

D-arabinose,125'128'133'134'137 g-xylose,73'128'133-137 and D-ribose,128'133'134'137 as well as 2-deoxy-D-erythro- 1 -2 1 pentitol and 2-deoxy-D-arabino-hexitol. The C-l’ alkyl groups have included O-methyl,73'122'123'125'127'131 36 , 126,131 „ J.u , 130 _ .. , 73,122-124,127-138 O-benzyl, ' S-methyl, S-ethyl, ' 13 7 and S-isobutyl. In most of these syntheses, only one C-l’ epimer has been isolated. However, there are instances, including the first acyclic-sugar nucleosides synthesized, where epimeric pairs have been separated from each other, 131 133 137 either by fractional crystallization ' ’ or by 132 133 137 thin-layer chromatography. ' ' In each of these cases, one epimer was found to be strongly levorotatory and the other strongly dextrorotatory. The absolute configuration of these epimers was mostly not established, and epimers were assigned

(+) or (-) in accordance with their sign of optical rotation.

In the earlier work, the site of glycosylation of the product was not established or was assigned solely by ultraviolet- spectroscopic data. It was not until X-ray crystallographic structure analysis was performed on compounds 2,3,4,5-tetra-

0-acetyl-l-S-ethyl-1-(1,6-dihydro-6-thioxopurin-9-yl)-1-thio-

D-arabinitol139 and (IS)-1-(6-chloropurin-9-yl)-1-S-ethyl-

1-thio-D-glucitol,which were prepared by Baker et al. 128,139 an^ gt al. 129,138 reSpectively, that both the site of glycosylation and configuration at C-l' were firmly established for certain key examples (Chart 9).

(139) D. C. Baker, A. Ducruix, D. Horton, and C. Pascard-Billy, Chem. Commun., 729-732 (1974); A. Ducruix and C. Pascard-Billy, Acta Crystallogr. Ser. B, 3JL, 2250- 2256 (1975).

(140) A. Ducruix and C. Pascard-Billy, Acta Crystallogr. Ser. B, 3J3, 2501-2503 (1977) . 37

Cli cr

02 i

C21 >C3 104 03 C4

C5, OS

(IS)-1- (6-chloropurin-9-yl)-1-S-ethyl-l-thio-g-glucitol

IS (6)

13B

1-3 0 ’

146 1-BSj ,C(3)

iC(5) sm C(2) O s ® 0 ® N O c

2,3,4,5-tetra-O-acetyl-l-S-ethyl-l-(1,6 dihydro-6-thioxo- purine-9-yl)-1-thio-D-arabinitol

Chart 9 38 Significant in vivo biological activity has been demonstrated by 1-(lH-6-thioxopurin-9-yl)-1-S-ethyl-l-thio- 127 D-glucitol (NSC 115963) against L-1210 mouse lymphoid

leukemia, -whereas the g-galacto analog has no activity. The g-gluco compound has also been shown to be a more potent

inhibitor of Escherichia coli K-12 (an in vitro model for 136 antitumor activity) than the D-galacto compound. This dependence of activity on the configuration has also been 127 observed in other acyclic-pentose nucleosides, and appears to be a general phenomenon for acyclic-sugar nucleosides as

indicated from more examples to be discussed later.

In addition to those compounds just discussed, there are other types of compound that may be classified, in the broadest sense of the term nucleoside, as acyclic-sugar nucleosides. Willardiin and eritadenine (Chart 10) are naturally occurring ones. 4-(6-Amino-9H-purin-9-yl)-2(R),3(R)- dihydroxybutanoic acid (eritadenine) was first crystallized 141 by Chibata et ad. from caps of the dried mushroom,

Lentinus edodes; their study was stimulated by earlier reports that plasma cholesterol levels are markedly decreased by a substance present in the edible mushroom "Shiitake" (Lentinus edodes), and which is a traditional Japanese delicacy.

(141) I. Chibata, K. Okumura, S. Taneyama, and K. Kotera, Experientia, ^25, 1237-1238 (1969) . 39

HN'

CHs-CH-COOH 2 I HO— i NHo Willardiin Acycloguanosine NH2 \ NH ch2

NH H OH H. OH

H,_C„ CH HOHoC HOHpC

H-C-OH I CH- i HO OH EHNA Coformycin 2'-Deoxycoformycin

Chart 10 40

The structure of eritadenine was established by a 141 total synthesis in which 4-amino-6-chloro-5-nitropyrimidine was condensed with erythro-4-amino-2,3-dihydroxybutanoic acid, with subsequent reduction of the 5-nitro group and formylation

to generate the purine ring. The synthesis was later improved by direct condensation of sodium salt of adenine with either 142 (2R),(3R)-O-protected dihydroxybutyrolactone or 2,3-0- 143 isopropylidene-D-erythronolactone. Two other substances,

(142) K. Okumura, T. Oine, Y. Yamada, M. Tonie, T. Adachi, T. Hagura, M. Kawaza, T. Mizoguchi, and I. Inoue, J. Org. Chem., 3_6, 1573-1579 (1971).

(143) T. Kamiya, Y. Saito, M. Hashimoto, and H. Seki, J. Heterocycl. Chem., 9_, 359-362 (1972). deoxyeritadenine and 9-(carboxypropyl) adenine (Chart 10), 144 . . were also isolated from same source but showed no activity.

Eritadenine has been shown to decrease the levels of all lipid components of plasma lipoprotein in animals and man, and its toxicity in rats is very low. Its biological effects have 145 been studied; the data indicate that the hypocholesterolemia is induced by increasing the excretion of cholesterol without affecting the de novo synthesis of cholesterol from acetate.

(144) Y. Saito, M. Hashimoto, H. Seki, and T. Kamiya, Tetrahedron Lett., _56, 4863-4866 (1970); F. Tokita, N. Shibukawa, T. Yasumoto, and T. Kaneda, J. Jpn. Soc. Food Nutr., 24, 92-95 (1971). 41

(145) S. Tokuda, A. Tagiri, E. Kano, and T. Kaneda, J. Jpn. Soc. Food Nutr., 2_4, 477-480 (1971); S. Tokuda, E. Kane, and T. Kaneda, ibid., 2_5, 608-613 (1972).

146 Okumura et ad. synthesized 124 derivatives of eritadenine and evaluated their hypocholesterolemic activities. The most active derivatives, which were as much as 50 times more effective than eritadenine at a dose of

0.0001% in the diet of rats, were found to be the carboxylic acid esters formed with short-chain monohydroxy alcohols.

(14 6) L. Okumura, K. Matsumoto, M. Fukamizu, H. Yasao, Y. Taguchi, Y. Sugihara, I. Inoue, M. Seto, Y. Sato, N. Takamura, T. Kanno, M. Kawazu, T. Mizoguchi, S. Saito, K. Takashima, and S. Takeyama, J. Med. Chem., 17, 846-855 (1974).

147 A series of aliphatic analog of nucleosides having a 2,3-dihydroxypropyl chain were also synthesized.

The compounds having the (S) configuration were synthesized by condensation of 2,3-O-isopropylidene-l-O-p-tolylsulfonyl-

D-glycerol with the sodium salt of the base, followed by acid hydrolysis. The (R) enantiomers were prepared from a preformed homonucleoside by cleavage of the ribose ring with sodium periodate, followed by reduction with sodium borohydride. Some ribonucleases are able to cleave 2',3'-cyclic phosphate of the (£3) enantiomer because only the (S) configuration can imitate the conformation of the naturally 147 occurring 6-D-ribofuranosyl derivatives. These compounds were also found to be active as inhibitors of enzymes of nucleic acid metabolism and effective as antimicrobial

(147) A. Holy, Collect. Czech. Chem. Comm., 40, 187-214 (1975); A. Holy and G. S. Iranova, Nucleic Acid Res., 1, 19-34 (1974).

(148) A. Holy, Proc. Int. Conf. Ribonucleic Acids and Their Components, Poznan, Poland, 1976, p. 134; A. Holy, Proc. Int. Conf. Recent Developments in Oligo nucleotide Syn. and Chem. of Minor Bases of tRNA, Poznan, Poland, 1974, p. 223.

Similar compounds were prepared earlier by different 149 methods. Ueda et ad. synthesized 3- and 9-(2,3-dihydroxy-

propyl)adenines by condensation of glycerol a-chlorohydrin with the sodium salt of adenine, which gave a 1:4 mixture of

9- and 3-isomers, or by treating adenine with 2,3-epoxy-l-

propanol in N,N-dimethylformamide with potassium carbonate

as catalyst, which gave a 3:1 mixture in favor of the

9-isomer. The (S) enantiomer has been shown to display 150 broad-spectrum antiviral activity and low acute toxicity.

It suppressed the cytopathic effect of vaccinia, herpes

simplex virus type I (HSV-1), measles, and vesicular stomatis virus in cell culture. The (R,S) form is also active, whereas

(R) enantiomer is inactive. The (S) enantiomer is also an

inhibitor of adenosine deaminase from calf thymus mucosa and acts synergistically with 9-(B-D-arabinofuranosyl)-adenine

(Ara-A) against vaccinia virus. 43

(149) N. Ueda, T. Kawabata, and K. Takemoto, J. Heterocycl. Chem., J3, 827-829 (1971).

(150) Stichting REGA, Belg. Pat. 871, 366; Chem. Abstr., 9JL, 117516t (1979).

Adenosine deaminase (adenosine aminohydrolase

EC 3.5.4.4) is an enzyme that catalyzes the deamination of adenosine to inosine. This enzyme is of chemotherapeutic

importance because it is responsible for inactivation of

some antineoplastic agents that are analogs of adenine.

A noted example is ara-A. Ara-A has marked in vitro activity against a broad spectrum of DNA viruses implicated in human and veterinary diseases. Becuase ara-A is rapidly deaminated to 9-(g-D-arabinofuranosyl)hypoxanthine (ara-Hx) in the 151 mouse, the effectiveness of ara-A is limited because ara-Hx is inactive.

(151) J. J. Brink and G. A. LePage, Cancer Res., 24, 1042 (1964).

One approach to minimize the deamination process is to use inhibitors of adenosine deaminase. The discovery of inhibitors is indeed one of the most exciting developments on the therapeutic efficacy of the adenosine nucleoside analogs. Besides the two most potent, naturally occurring nucleoside analogs, coformycin (CF) and 2'-deoxycoformycin

(dCF, Pentostatin) (Chart 10) the total synthesis of the 152 latter having been described by Baker and Putt, there are 44

several artificially produced inhibitors, erythro-9- (2- hydroxy-3-nonyl)adenosine (EHNA, Chart 10) is one of them.

As part of their studies of a"series of l-hydroxy-2-alkyl and 2-hydroxy-3-alkyl derivatives of adenine as adenosine 153 deaminase inhibitors, Schaffer and Schwender prepared this compound by a scheme analogous to that used for the 154 synthesis of eritadenine. They observed the (R)-9-(l- hydroxy-2-alkyl)adenine and (S)-9-(2-hydroxypropyl)adenine were better inhibitors of the enzyme than were their 155 . . corresponding enantiomers. This stereoselectivity was explained on the basis of formation of an enzyme— inhibitor

(El)complex. With 9-(l-hydroxy-2-alkyl)adenine, the chirality favored for the El complex is R, whereas with 9-(2-hydroxy propyl ) adenine , the S enantiomer is bound more tightly to the enzyme than the compound with R configuration. Similarly, erythro-9-(2-hydroxy-3-alkyl)adenine is a better inhibitor than its threo isomer.

(152) D. C. Baker and S. R. Putt, J. Am. Chem. Soc., 101, 6127-6128 (1979).

(153) H. J. Schaffer and C. F. Schwender, J. Med. Chem. , 11_, 6-8 (1974) .

(154) T. Kamiya, Y. Saito, M. Hashimoto, and H. Seki, Tetrahedron, ^28, 899-906 (1972).

(155) H. J. Schaffer and R. Vince, J. Med. Chem., 10, 689-691 (1967).; H. J. Schaffer, R. N. Johnson, M. A. Schwarz, and C. F. Schwender, ibid., 15, 456-458 (1972). 45

156 EHNA is a potent semi-tight binding inhibitor. -9 -11 Its Ki is 1.6 x 10 M, as compared with 1 x 10 M for 157 coformycin. The therapeutic potential of EHNA was 158 demonstrated by Cohen and coworkers. They have shown that

the therapeutic efficacy of ara-A is increased by the

simultaneous administration of EHNA. For example, EHNA

dramatically increases the anti-herpes simplex virus (anti- 158 HSV) effect of ara-A and increases the lethality of ara-A 159 — 6 in mice bearing the Ehrlich ascites carcinoma. EHNA (10 M)

alone inhibits HSV production by 30%.

(156) R. P. Agarwal, S. Cha, G. W. Crabtree, and R. E. Parks, Jr., in "Symp. Chem. Biol, of Nucleosides and Nucleotides", R. K. Robins and R. E. Harmon, Eds., Academic Press, New York, 1978, pp. 159-197.

(157) R. P. Agarwal and R. E. Parks, Jr., Biochem. Pharmacol., 2j5, 663-666 (1977).

(158) T. W. North and S. S. Cohen, Proc. Natl. Acad. Sci., 75, 4684-4688 (1978).

(159) W. Plunkett and S. S. Cohen, Cancer Res., 35, 1547-1554 (1975).

Recently, a new antiviral agent has been reported and has attracted substantial attention. The compound,

9 - (2-hydroxyethoxymethyl)guanine (acycloguanosine, abbr. 16 fl Acyclo-Guo) (Chart 10) was synthesized by condensation of 2 ,6-dichloropurine with 2-benzoyloxyethoxymethyl chloride, followed by stepwise displacement of chloride by the amino group. Acyclo-Guo has potent antiviral activity toward HSV-1 46

161 and HSV-2 viruses. This compound is essentially,nontoxic

to host cells. The LD,.- index in mice given the drug orally 5U 160 was greater than 10,000 mg per Kg. The most prominent

feature of the compound is its selectivity. Acyclo-Guo was

found to be phosphorylated at a 30- to 120-fold higher rate with extracts of Vero cells (host cells) infected with HSV

than extracts of uninfected cells. The enzyme responsible for

this reaction was identified as virus-induced thymidine

kinase (TK). The host-cell TK does not take acyclo-Guo as

substrate. This difference in specificity of virus-induced and

host-cell enzymes is the basis of selectivity. The induction

of acyclo-Guo phosphorylating activity appeared to be

essential for the compound to be effective as an inhibitor of

(160) H. J. Schaffer, L. Beauchamp, P. deMiranda, G. B. Elion, D. J. Bauer, and P. Collins, Nature, 272, 583-583 (1978).

(161) G. B. Elion, P. A. Furman, J. A. Fyfe, P. deMiranda, L. Beauchamp, and H. J. Schaffer, Proc. Natl. Acad. Sci., 74, 5716-5720 (1977).

(162) J. A. Fyfe, P. M. Keller, P. A. Furman, R. L. Miller, and G. B. Elion, J. Biol. Chem., 253, 8721- 8727 (1978).

161 virus replication. Kinetic studies showed that acyclo-GTP

is a competent inhibitor of HSV-1 polymerase, having Ki = 0.08

± 0.03 yM, which is about 1/30 that for the a-polymerase of the host cells. This implies that there is about 30-fold greater sensitivity of the viral DNA polymerase to acyclo-GTP than that of the host cellular polymerase. There is also indication that acyclo-GTP is a substrate for HSV-1 spe cified DNA polymerase. Incorporation of acyclo-GTP into viral DNA could terminate the chain, because no hydroxyl group corresponding to the 3 '-hydroxyl group of deoxyribose is available. II. STATEMENT OF THE PROBLEM

The synthesis, characterization, and biological properties of acyclic-sugar nucleosides has been pursued

in this laboratory as part of a long-standing general

interest in acyclic carbohydrate derivatives. The general procedure developed has involved the conversion of

acetylated acyclic sugars into 1-bromo derivatives and

subsequent condensation of these reactive a-halothioethers with suitably activated purines or pyrimidines to afford

l'-epimeric mixtures of acyclic-sugar nucleoside derivatives.

Although technical aspects of this procedure have been

improved greatly, difficulties are sometimes still

encountered in repeating the procedure. Furthermore, the

bromination reaction does not work for 2-deoxy sugar deri vatives, presumably because of changes in the electronic

environment at C-l caused by the absence of an electron- withdrawing acyloxy group at C-2. Therefore, it was judged

desirable to develop a new procedure that would give more-

consistent results and be of broader applicability.

The products of these coupling reactions of acyclic derivatives are commonly encountered as mixtures of two diastereomers differing in configuration at C-l. Outside

48 of crystallization, which generally resolves only one epimer in pure form, there is no uniformly effective method for separation of both epimers in pure form. Since the dependence of biological activity on the configuration has been observed and appeared to be a general phenomenon for acyclic-sugar nucleosides, separation of these epimers is essential for understanding the structure-activity relationship of these compounds with respect to the stereochemistry at C-l. Therefore, it was proposed to explore the feasibility of high pressure liquid chromato graphy as a technique for effecting such separation.

Complete structural identification of the products requires proof of the position of attachment of the sugar chain to the heterocycle. Definitive proof involves such non-routine methods as X-ray analysis or unambiguous chemical 13 synthesis. Recently, C-NMR spectroscopy has been found a very useful tool for characterization of nucleosides. Its application with acyclic-sugar nucleosides was to be studied.

Furthermore, the rationale for synthesizing the acyclic- sugar nucleoside is that, according to the particular configuration of the sugar moiety, the portion of the acyclic chain may fold in such a way as to mimic the conventional furanosyl nucleoside. This hypothesis seemed valid by support from many examples, in which the observed activity could be explained by the ability (or inability) 50 of the acyclic portion to assume a structure that is isosteric with natural furanosyl ring. The importance of the conformation of acyclic-sugar nucleoside is obvious. If the conformation of the sugar chain in these compounds can be predicted with reasonable certainty, it will be very helpful in the future design to improve biological activity.

Therefore, the conformations of the acyclic sugar chains in these products was to be examined as a continuation of efforts to correlate biological activity and chemical reactivity to structural features of the nucleosides. III. RESULTS AND DISCUSSION

A. Synthesis of Acyclic-Sugar Nucleosides

The purpose for the preparation of acyclic-sugar nucleosides in this study was to provide compounds for the study of their cyclization by microbiological oxidation

(See Part Two). The sturctural requirements for substrate to be oxidized by Acetobacter suboxydans may be written as OH OH -C— C— CH„OH. D-Glucose, D-mannose, D-altrose, and 3-deoxy- H H D-arabinohexose were chosen for this study because they all possess the configuration that meet the requirements as substrate for oxidation by A. suboxydans. The reason for including 3-deoxy sugar is because, after cyclization, it will afford a nucleoside that is an analog of a 2-deoxy nucleoside.

1. Study of the Sugar— Base Coupling Reaction—

(1) Preparation of aldose diethyl dithioacetals—

D-Glucose and g-mannose were converted into their diethyl dithioacetals 1 and 2 , respectively, by treatment with ethanethiol in concentrated hydrochloric acid. Each compound had physical properties in good agreement with those reported

51 52

in the literature.

(16 3) M. L. Wolfrom and A. Thompson, Methods Carbohydr. Chem., 2, 427-430 (1963).

(164) E. Fischer, Ber., 27, 673-679 (1894).

D-Altrose diethyl dithioacetal (3) was prepared

from methyl a-D-glucopyranoside; the latter was converted

stepwise^®^ into methyl 4,6-0 -benzylidene-2 ,3-di-0-£-

tolylsulfonyl-a-D-glucopyranoside and thence into methyl

2,3-anhydro-4,6-O-benzylidene-a-D-altropyranoside, which 166 was then converted in one step into 3 in 70% yield.

(165) N. K. Richtmyer, Methods Carbohydr. Chem., 1, 107-113 (1962).

(166) B. Coxon and L. Hough, Carbohydr. Res., 8, 379-397 (1968).

3-Deoxy-D-arabino-hexose diethyl dithioacetal (4J was also prepared from methyl a-D-glucopyranoside. Following the 167 16 8 scheme of Wiggins and of Rembarz, the starting compound was converted into methyl 4,6-0-benzylidene-3-deoxy-a-D- arabino-hexopyranoside in four steps. The final product was O deprotected by boiling for 3 h at 100 in 30% acetic acid under reflux to afford the free sugar, 3-deoxy-D-arabino- hexose, which was converted into its diethyl dithioacetal by the method mentioned earlier. 3-Deoxy-D-arabino-hexose diethyl dithioacetal, which has been reported as colorless syrup, was isolated as needle-like crystals having 53 O m.p. 69-71 . All other physical properties were in good 168 agreement with literature values.

(167) L. F. Wiggins, Methods Carbohydr. Chem., 2, 188-189 (1963).

(168) G. Rembarz, Ber., 9J3, 622-625 (1960).

Each of these compounds was acetylated conventionally by

use of pyridine— acetic anhydride. The physical properties 163 of penta-O-acetyl-D-glucose diethyl dithioacetal (16), 169 penta-O-acetyl-D-mannose diethyl dithioacetal (28), and

tetra-0-acetyl-3-deoxy-D-arabino-hexose diethyl dithioacetal 168 (32) agree with those reported in the literature.

(169) N. W. Pirie, Biochem. J. , 3_0, 374-376 (1936).

Penta-O-acetyl-D-altrose diethyl dithioacetal (34), which 170 was synthesized for gas-chromatographic purposes but not

properly characterized, was prepared in 74% yield by

conventional acetylation and subsequent crystallization from

95% ethanol.

(170) D. J. Williams and J. K. N. Jones, Can. J. Chem., 44, 412-415 (1966).

(2) Coupling of sugar and base— The coupling of

acetylated sugar dithioacetals with purine bases to afford

acyclic-sugar nucleosides was first reported by Wolfrom 122 et al. in 1961. The synthesis utilized a reactive, 54 acyclic 1-bromo derivative prepared from a peracetylated 171 sugar dithioacetal by Weygand's adaptation of the 172 Gauthier procedure of direct bromination. The unstable bromo derivative was then condensed with a nucleophilically activated base to form the desired nucleoside. The base of

(171) F. Weygand, H. Ziemann, and H. J. Bestmann, Chem. Ber., 91, 2534-2537 (1958).

(172) B. Gauthier, Ann. Pharm. Fr., 12, 281-285 (1954) . particular interest in this work is 6-chloropurine, which was activated by converting into its chloromercuri salt, namely, 6-chloro-9-(chloromercuri)purine. This procedure has been applied for preparation of many acyclic-sugar nucleosides, and several modifications of the procedure have been made to improve the ease of isolation and/or the yield in our laboratory. However, difficulties were encountered in repeating the preparation of (IS)-2,3,4,5,6- penta-O-acetyl-1-(6-chloropurine-9-yl)-1-S-ethyl-l-thio-

D-glucitol (ljfcj) by this procedure (Chart 11) as described 129 138 in Markovs' thesis, ' wherein the 1-bromo derivative

16 was boiled in toluene for 4 h under reflux with 6-chloro-

9-chloromercuripurine in the presence of cadmium carbonate and Celite. The procedure was repeated many times, and either the syrup obtained failed to crystallize even with seed crystals added, or when it did crystallize, the yields fluctuated and never approached the 58% claimed. The original Cl

Br H I EtSCH EtSCSE-t HCSEt I I HCOR HCOAc HCOAc I I I ROCH AcOCH 0r2 AcOCH I I HCOR HCOAc EtgO HCOAc I I I HCOR HCOAc HCOAc I I c h 2o r CH2OAc CH2 OAc

rjb r = Ac (4966) 2$ (515.4) (589.0) R=H (378.8)

Chart 11 56 plan of this work requires the use of 17b as a starting compound. In order to obtain this compound in a more consistent and predictable manner, the reaction scheme was re-examined. Factors that may affect the outcome of this reaction include, among many others, the quality of the metal salt, the extent of completion of the bromination step, and the quality of the cadmium carbonate used. Evidence that the need for prior formation of a metal salt of the purine can be dispensed withare abundant in the literature.

One important example is the work of Yamaoka and asso- 113 . ciates, m which they successfully coupled purine bases with acylglycosyl halides, in the presence of mercuric cyanide as acid scavenger, in inert solvents. The bases employed were 6-chloro-, 2 ,6-dichloro-, and 2 ,6 ,8-trichloro- purine and the reaction was claimed to give improved yields, some 20-40% higher than the conventional processes. Studies of the condensation with 6-chloropurine of the 1-bromo derivatives obtained from sugar dithioacetals were made.

Bromination of the peracetylated diethyl dithioacetals was conducted by dissolving them in anhydrous ethyl ether; bromine, which was dried with concentrated sulfuric acid and dissolved in anhydrous ethyl ether, was then added dropwise O during a period of 20 min at 0 . The solution was then warmed to room temperature. The bromination reaction was monitored by t.l.c. with 1:1 benzene— ethyl acetate. The unstable bromo compounds were used immediately in the coupling reaction 57 without being characterized. Thus, the freshly prepared

2,3,4,5,6-penta-O-acetyl-l-bromo-l-S-ethyl-l-thio-D-glucitol was condensed with 6-chloropurine in nitromethane and in the presence of mercuric cyanide, cadmium carbonate, Celite, and calcium sulfate. In attempting to find optimal conditions for the coupling, the amounts and combination of the reagents were varied, whereas the bromo sugar and base were kept at approximately equimolar ratio. For no known reason, the amount of mercuric cyanide used in the syntheses by Yamaoka et al. of many nucleosides appeared to vary randomly, and was generally not equimolar to the amount of the reactant. In the synthesis of 6-chloropurine derivatives, the amount of mercuric cyanide .used was about 2/3 equiv. with respect to the base. This same ratio was adapted in most of the trials, and no dramatic improvement of the yield was observed when the amount of mercuric cyanide was increased to 1 equiv. 173 Cadmium carbonate is a reagent that was used originally for demercaptalation of the peracetate of D-glucose diethyl dithioacetal. This reagent was found to be a very effective acid acceptor in this type of reaction, and it is noteworthy that attempts to remove the thio group failed when cadmium 174 carbonate was not used at all or was not present m excess.

Subsequently, this reagent was also used successfully for the preparation of acyclic-sugar nucleosides by the chloromercuri salt method. Because of these considerations, cadmium carbonate was included as one of the reagents. In most of the 58 trials, the amount of cadmium carbonate was kept in equimolar portion to that of mercuric cyanide. When it was used in excess, the reaction was retarded somewhat as shown by t.l.c., and was evidenced by the increased yield of a major side-product. When cadmium carbonate was completely removed, the reaction still proceeded without difficulties. It was also found that Celite could be omitted without affecting the outcome of the reaction.

(173) M. L. Wolfrom, J. Am. Chem. Soc., 51, 2188-2193 (1929).

(174) P. Brigl and H. Muehlschlegel, Ber., 63, 1551-1557 (1930); W. Schneider, J. Sepp, and 0. Stiehler, 53., 220-234 (1918).

Comparative studies were also made on the conditions for this coupling reaction of bromo compounds. Temperature O is an important factor. When the mixture was stirred at 0 , virtually no coupling occurred, whereas at room temperature, the reaction did proceed, but extremely slowly. This result 175 contrasts with the observation that 1.5 h at room temper- O ature or at 0 was sufficient to afford an optimal yield of a peracetylated 1-S-methyl-l-thio-D-glucitol nucleoside derivative by the chloromercuri salt method. On the other

(175) K. Blieszner, Ph.D. Dissertation, The Ohio State University, (1978), p. 127; Diss. Abstr. Int. B. 39, 4888-B (1979). 59 hand, the reaction was complete within 4 h in boiling O nitromethane (110 ). However, when the reaction was performed O at an intermediate temperature (60-70 ), the reaction product was relatively less complex, as judged by t.l.c. examination.

The reaction yield generally was lower than that obtained by 129 the former method. One unexpected advantage of the present procedure is the relative ease of crystallization directly

from the reaction mixture. In many instances, crystals formed

immediately, without seeding from the syrup obtained after processing. The crystals obtained from these preparations 129 had the same physical properties as reported. This modified reaction scheme (Chart 12) was also applied for preparation of penta-O-acetyl-1-(6-chloropurin-9-yl)-S-ethyl-l-thio-D- mannitol (30), -g-altritol (36), and tetra-O-acetyl-1-

(6-chloropurin-9-yl)-3-deoxy-l-S-ethyl-l-thio-D-arabino- hexitol (38). None of these syntheses allowed direct crystal lization of the product. Nevertheless, all of these reactions proceeded satisfactorily except one. In a scaled-up preparation

(5-g scale) of 30, in the presence of mercuric cyanide, cadmium carbonate, Celite, and calcium sulfate, two coupling products were obtained in low yield. The desired nucleoside

30 was obtained in 15% and another product, later identified as the 9-substituted derivative of 6- (ethylthio)purine, was obtained in 15% yield.

Careful review of the reaction conditions offered no apparent explanation for this unusual behavior, and the Br H Cl EtSCSEt HCSEt CHSEt I I I Ri CRo r|C r 2 r ,Cr 2 I R 3 ^ R4 Br2 r,cr4 r 3 Cr 4 HCOAc Et^O hioAc HCOAc I Hg(CN)2 H < W HCOAc HCOAc I CH2 0Ac I CH2OAc CH2OAc

R 1 R2 R3 R4

17: D-gluco H OAc OAc H

30: D-manno OAc H OAc H

36 : D-altro OAc H H OAc

38: 3-deoxy- OAc H H H D-arabino

Chart 12 61 only noted difference of this reaction from the others is that the syrup obtained after processing was dark-colored as contrasted with the normal light-yellow or orange product mixture. This result was not reproducible. The same reaction was later repeated several times only to give the desired product 30 in good yield. The 6-ethylthio nucleoside presumably was formed by attack of the ethylthio mercury complex on the

6-position of the purine.

The isolation procedure used for the reaction mixture 129 was the same as that described, except for the minor modification that the suspension in nitromethane was filtered, the filtrate evaporated to dryness, the residue extracted with warm chloroform, and the insoluble solid filtered off. The advantage of this modification is that the substantial amount of EtSHgCN formed, and trace of unreacted 6-chloropurine, were removed; this made the subsequent washings with potassium iodide less troublesome.

To further simplify the procedure, the next logical target of choice was the bromination step, which does not always go to completion and hence lowers the reaction yield.

Elimination of the bromination reaction would necessitate conditions that would remove one ethylthio group (demercaptalation) other than by use of bromine. The fact that the dithioacetal groups can be desulfurized by heavy metals is 176 well known. In 1916, Schneider and Sepp prepared thio- glucosides by partial hydrolysis of glucose dithioacetals 62 173 with one mole of mercuric chloride. Later, Wolfrom successfully demercaptalized the peracetate of glucose diethyl dithioacetal by mercuric chloride with cadmium carbonate as acid acceptor. Mercuric chloride and yellow 177 mercuric oxide have been used for the same purpose. 178 Furthermore, Pedersen and Fletcher reported the conden sation of 5-O-benzoyl-2-deoxy-D-erythro-pentose diisopropyl dithioacetal with 6-benzamido-9-(chloromercuri)purine, followed by removal of the acyl groups, to give 2 '-deoxy- adenosine and its anomer, although the yield was very low.

(176) W. Schneider and J. Sepp, Ber. , 4_9, 2054- 2057 (1916).

(177) J. W. Green and E. Pacsu, J. Am. Chem. Soc., 59, 1205-1210 (1937).

(178) C. Pedersen and H. G. Fletcher, Jr., J. Am. Chem. Soc., 82, 5210-5211 (1960).

All of these results suggested that prior activation of the sugar dithioacetal by bromination may not be necessary.

This assumption was tested by the following experiments. 173 At first, procedures similar to Wolfrom's, namely, mercuric cyanide and cadmium carbonate were used, but no coupling occurred. Next, conditions similar to those used 177 by Pacsu and Green were tried. 6-Chloropurine and 16 were dissolved in nitromethane and boiled under reflux in the presence of mercuric cyanide and yellow mercuric oxide plus drying agent (calcium sulfate). The reaction proceeded 63 well and gave a reasonably good yield of the desired nucleo side. Obviously, yellow mercuric oxide functioned in this reaction more than as a simple acid acceptor, and caused the reaction to proceed. This reaction was studied as before, under a range of conditions. The amount of mercuric cyanide and mercuric oxide used was equal and was, as in the previous procedure, 2/3 mole equivalent of the reactant in all trials.

The temperature was again an important factor. As expected, O the reaction did not proceed at all at 0 and room temperature, O and occurred very slowly at 70 . The best result was obtained O under reflux in boiling nitromethane (110 ). The reaction was complete within 4 h. Prolonged heating did not improve the yield. Acetonitrile could also be used as solvent in this coupling reaction. The reaction proceeded much slower than with nitromethane and the yield was lower. This result is 113 quite in accordance with Yamaoka's observations. 179 As Ishido observed a very unusual kinetic formation of a 7-substituted isomer, which was rapidly rearranged to the 9-isomer in the coupling reaction of theophylline, the acyclic-nucleoside coupling reaction was monitored very closely, especially at early stages, by t.l.c., and no similar rearrangement was detected. This method (Chart 13) was also

(179) Y. Ishido, personal communication. applied for synthesis of compounds 30, 36, and 38. The reaction was conducted on both large (5-g) and small H I EtSCSEt CHSEt I R, C r 2 Hg(CN)2 R l ^ R 2 R,a C R ^ HgO r 3 C r 4 Hi0Ac CHgNOg HCOAc I H io A c HCOAc I I CH2OA c CH20A c

R1 R2 R3 R4

17: D-qluco H OAc OAc H

30: D-raanno OAc H OAc H

36: D-altro OAc H H OAc

38: 3-deoxy- OAc H H H D-arabino

Chart 13 65

(0.5- or 0.25-g) scale. The yields from small-scale synthesis were about 20% higher than those on the larger scale. Iso

lation of crystalline nucleoside derivatives from ethano,lic

solutions of the purified, syrupy 11, 30, and 36, was achieved.

The syrup of 38 failed to crystallize from various solvent systems tried.

As shown in Table 1, use of either one of the reactants in large or slight excess did not give better yields than when equimolar amount reactants were used.

The advantages of this procedure, besides the short ened reaction scheme, is the consistency in yield (Table 1) probably because of the fact that the procedure is simple and straightforward, and the conditions are less sensitive to moisture. The only factor that may seriously lower the yield is inadequate stirring, because charring may occur if suffi cient stirring is not maintained.

2. Separation of the Epimers from the Product of the

Coupling Reaction— The products of the foregoing coupling reactions were obtained as mixtures of two epimers differing only in configuration at C-l. The ^H-n.m.r. spectra of these chromatographically pure, syrupy mixtures in chloroform-d were not fully interpretable, but did reveal the presence of two products, as shown by the doubling of H-2,

H-8 proton signals, ethylthio proton signals, and especially of the doublet for anomeric protons (H-l'). For the pairs of

1-epimers 17, the H-l1 signals are observed at 6 6.20 (minor) Table 1. Preparation of 17, 30, 36, and 38.

Acetylated _ . b Compound sugar dithioacetal 6-C1--Purine Reaction Yield[ Epimer— Q. g mmol g mmol conditions— g *5 ratio

17 5 10.08 1.38 8.92 70° 5 h & 1 0 0 ° 24 h 2.58 49.5 5:1 5 10.08 1.56 10.08 1 20° 6 h 3.51 59 0.5 1.01 0.08 0.51 1 10° 4 h 0.42 70 0.25 0.504 0.156 1.01 1 2 0 ° 6 h 0.22 75 5.2:1 0.25 0.504 0.156 1.01 100° 72 h (in ch3cn) 0.18 61

30 5 10.08 1.56 10.08 1 2 0° 4 d 2.9 50 5 10.08 1.56 10.08 80° 3 h & 1 2 0° 3 d 3.33 56 2.7:1 5 1 0 . 08 2 13.07 120° 42 h 3.76 63 2.7:1 O A p C 0.25 0.504 0.156 1.01 110 5 h 0.13 45- 2 .8:1 o 0.25 0.504 0.156 1.01 110 5 h 0.20 68

36 5 10.08 1.56 10.08 120° 40 h 3.43 58 2.5 5.04 0.78 5. 04 12 0° 6 h 1.8 60 0.5 1.01 0.16 1.01 1 1 0 ° 4 h 0.47 79

38 5 11.42 1.56 10.08 1 2 0 ° 6 h 2.9 54 4:1 2.5 5.71 0.78 5. 04 120° 72 h 1.46 55 4.6:1

— In nitromethane, unless specified otherwise, b 1 — Determined by H-NMR spectroscopy, c — Charring of the reactant is responsible for low yield. 67

and 6 5.93 (major), with , 2 , 3.0 Hz, and , 2 , 3.5 Hz,

respectively. The corresponding H-l' signals of 30 resonate

at 6 5.92 ( , 2 , 2.6 Hz, major) and 6 6.13 ( , 2 , 4.5 Hz, minor), whereas those of 38 fell at 6 5.98 ( , 2 , 3.8 Hz, major), and 6 5. 89 ( , 2 , 4.3 Hz, minor). The ratios of

epimers were determined from integration of these two signals

(Table 1). The two H-l1 signals of 36 fell at 6 5.96 ( , 2 ,

3.0 Hz, minor) and 6 6.05 ( , 2 , 5.7 Hz, major). These two

sets of doublets are partially overlapped and make accurate measurements of the ratio impossible.

In order to characterize the physical properties of the products fully, isolation of pure epimers from the mixtures is essential. Unfortunately, conventional t.l.c. and column chromatography are generally not suitable to achieve that goal, although it has been used to separate a pair of 132 amino sugar nucleosides. The most direct method is the

isolation of fully acetylated isomers by crystallization.

In all cases of acyclic-sugar nucleosides synthesized, this

is indeed the method most frequently used for isolation of components from the mixture. The major C-l1 epimer of 17 has been crystalline from the syrupy mixture in benzene or abs. ethanol. This crystalline compound was unequivocally esta blished as having the IS configuration by X-ray crystallographic analysis^®' of its deprotected analog. Similarly, one epimer was successfully isolated crystalline from the

syrupy, epimeric mixtures of 30 and 36 by dissolution of the 68

syrup in abs. ethanol and keeping the solution at room tem perature. These crystalline products were the major epimers 1 formed, as shown by their H-n.m.r. spectra, which revealed the signal of the major component as the only H-l1 signal.

In contrast, all attempts to crystallize the syrupy nucleoside analog, 38, having the 3-deoxy-D-arabino-hexose sugar chain failed. Interestingly all coupling products of 2-deoxy-D- erythro-pentose and 2-deoxy-D-arabino-hexose with 6-chloro- 130 purine and 5-fluorouracil were reported as syrupy mixtures of IS and 1R epimers. Nevertheless, when the syrupy mixtures were deprotected by methanolic ammonia to afford the free- hydroxyl sugar chain derivative, one of the epimers crystal lized out. This epimerically pure compound was then re- acetylated for purposes of physical characterization.

Isolation by this method generally afforded one epimer pure, whereas the other was left enriched, but inseparable from the remaining syrup. Only in a rare few 131 137 instances ’ have both epimers been obtained m pure form directly from the reaction mixture by fractional crystal lization or chromatography. As enzymes are known to be able to recognize stereochemical features in a substrate molecule, useful information might be obtained regarding the structure— activity relationship and determination of anomeric configu ration if all epimeric pairs thus far prepared were isolated.

There are many reports of separation of anomers, epimers, and isomers of nucleosides by high pressure liquid chromatography Column: yBondapak C10 (7.8mm ID x 30cm) 1 O Solvent: Acetonitrile/water (40/60) Flow rate: 1 mL/min Detector: RI 32x (-), uv 254 nm

Figure 1: High pressure liquid chromatography of the epimeric mixture of nucleosides 30. 70

(HPLC). This technique was applied for the first time for

the separation of the epimers of an acyclic-sugar nucleoside.

The epimeric pair 30 was successfully separated after extensive

studies. Separation was carried out on yBondapak C^g, an

exceedingly nonpolar, microparticulate, and chemically-

bonded reverse-phase packing material. A semi-preparative

column of 7.8 mm inside diameter and 30 mm long was employed

and about 20 mg of sample, which is lower than the maximum

load that such a column can normally handle, was injected

each time. Separation was achieved by recycling and "shaving", with the solvent system of 3:2 water— acetonitrile. Repro

ductions of some actual chromatograms are shown (Fig. 1).

The two components were collected separately. The slower- 1 eluting component was crystallized and shown by H-n.m.r.

spectroscopy to be epimerically pure and identical with the crystalline compound obtained directly from the mixture.

The faster-eluting component, obtained with recycling, was

shown (‘''H-n.m.r. ) to be epimerically pure also.

B. Characterization of the Products

1. Position of Attachment of the Sugar Chain to the

Purine Base— Determination of the position of attachment of the sugar moiety to the heterocyclic base in

synthetic nucleosides is sometimes a most difficult problem as there are several possible sites for alkylation available. 71

Definitive proof of the structure involves such procedures as X-ray analysis of a single crystal or unambiguous chemical transformations. As mentioned earlier, an X-ray crystallographic analysis has been made on (IS)-1-(6-chloropurin-9-yl)-

1-S-ethyl-l-thio-D-glucitol (17b) and the position of glycosylation established to be at N-9 of the 6-chloropurine ring.

This compound thus serves as a very useful point of reference for the study of compounds prepared in this work. To ascertain with reasonable confidence the position of glycosylation without resorting to the less accessible methods just mentioned, several methods have been developed and widely used to solve the problems.

(1) Ultraviolet spectra— The ultraviolet spectrum of a nucleoside product and that of the corresponding N-alkyl derivative of the aglycon is very useful for structural correlation. Comparison of the intensity and wavelength of maxima and minima under different pH conditions is usually 180 sufficient to establish the site of substitution.

(180) L. B. Townsend, R. K. Robins, R. N. Leoppky, and N. J. Leonard, J. Am. Chem. Soc., 86_, 5320-5325 (1964).

There are several reports of ultraviolet spectral 181-183 data for alkylated 6-chloropurines. For the purpose of comparison, they are listed in Table 2. The ultraviolet spectra of all acetylated nucleosides prepared in this work were measured in methanol. The absorption maximum of 30, 36, Table 2. The ultraviolet absorption maxima of some 7 or 9-alkyl-6-substituted purine.

acidic neutral basic Compound Ref. X . . eXlO-3 X , . eXlO '3 X , , eXlO-3 max (nm) max (nm) max (nm)

6-Chloro-9-ethylpurine 265 9.46 266 9.40 266 9.40 181a

6-Chloro-9-methylpurine 265 8.1 ------268 21.1 183

6-Chloro-7-ethylpurine 267 8.0 270 8.1 ------181b

6-Chloro-7-methylpurine 268 8.4 271 8.5 ------182 73

and 38 (264.5 nm, e 5,000; 265 nm, £ 7,500, and 266 nm,

£ 1,380, respectively), together with that of VJ (264 nm,

e 13,000), correspond closely to the literature values for

9-alkyl-6-chloropurines. Compound 31, whose mass-spectral 1 13 and H- and C-n.m.r. spectroscopic data indicated that the

base in this compound is 6- (ethylthio)purine, displayed in

methanol an absorption maximum at 284 nm (e 10,750) with a

shoulder at 29 0 nm.

(181) (a) J. A. Montgomery and C. Temple, Jr., J. Am. Chem. Soc., 79_, 5238-5242 (1957); (b) ibid., 83, 630-635 (1961).

(182) R. N. Prasad and R. K. Robins, J. Am. Chem. Soc., 79, 6401-6407 (1957).

(183) R. K. Robins and H. H. Lin, J. Am. Chem. Soc., 79, 490-494 (1957).

Compounds 30a, 31, 36a, and 38 were readily

deacetylated by methanolic ammonia. The uv spectra of the corresponding deprotected nucleosides 5Ja, 57, 49a, and 5Jb, were measured at three different pH values, i.e. pH 1, 7,

and 12. Their absorption maxima are listed in Table 3.

The relative insensitivity to pH and the wavelength of the maxima shown in the spectra of 49a, 52a, and 54b are quite consistent with those observed for (1£>)-6- (chloropurin-9-yl) -

1-S-ethyl-l-thio-D-glucitol and 6-chloro-9-alkylpurines and are at variance with those of 6-chloro-7-alkylpurines (See

Tables 2 and 3). These uv data thus support the assignment that the sugar chain is attached to the purine at N-9. Table 3. The ultraviolet absorption maxima of acyclic-sugar nucleosides.

pH 1 pH 7 pH 12 Compound eXlO 3 eXlO 3 eXlO 3 ^max (nm) * max (nm)/ * \nax (nm)

(IS)-1- (6-Chloropurin-9-yl)- 1-S-ethy1-1-thio-D- 261 6.9 264 7.9 264 6.9 glucitol (17b)

49a 262.5 5.9 263 5.9 263 5.9

52a 263. 5 5.2 263 5.2 263 5.2

54b 262.5 18.1 265 18.1 262 18.5 75

The spectrum of 57 is totally different from that of others, because of the different substituent on purine ring.

There are several reports of uv data for 9-alkyl-6-(methylthio)purines but none could be found for their 7-substituted isomers. However, uv spectral data have been reported for the 18 5 nucleosides 6- (methylthio)-7-£-D-ribofuranosylpurine, 186 6- (methylthio)-9-g-D-arabinofuranosyl- and 6- (ethylthio)- 181 9-B-D-ribofuranosylpurine (See Table 4). Also a series of

6- (alkylthio)-9-(hydroxymethyl)purines, in which the alkyl group ranged from methyl to decyl, has been prepared and 187 their uv spectra studied. Although detailed uv data for individual purines were not given, it was concluded that all of these compounds had similar spectra, with within the range of 288-292 nm (e max 17,900-12,000). When these data are compared with those of 57 and a 4-nm wavelength variation is allowed, it is clear that the uv spectrum of

57 corresponds more closely to the 9- rather than the 7- substituted isomer, and is almost identical with that of

9-8-D-arabinofuranosyl-6-(methylthio)purine. Therefore, the site of glycosylation for compounds 31 and 57 is assigned as N-9.

(184) J. A. Montgomery, T. P. Johnson, A. Gallahger, C. R. Stringfellow and F. M. Schabel, Jr., J. Med. Pharm. Chem., 3, 265-288 (1961).

(185) R. J. Rousseau, R. P. Panzica, S. M. Reddick, R..K. Robins and L. B. Townsend, J. Org. Chem., 35, 631-635 (1970). Table 4. The ultraviolet absorption maxima of 57 and related compounds.

acidic neutral basic Compound -3 Ref, X eXlO~3 X , . eXlO-3 X , . eXlO max (nm) max (nm) max (nm)

9-Methyl-6-(methylthio)purine 295 19.5 ------287 20.8 183

9-Ethyl-6-(methylthio)purine 296 16. 5 286 17.6 286 17.7 181a 290 (sh) 2 9 0 (sh) 6- (Methylthio)-9-g-D- 293 17.4 289 18.9 289 18.9 184 ribofuranosylpurine MeOH 6- (Methylthio)-7-g-D- 300 12.5 291 13.7 293 13.7 185 ribofuranosylpurine 6- (Ethylthio)-9-3~g- 294 16.8 292 18.2 292 18.4 183 ribofuranosylpurine

9-S-D-Arabinofuranosyl-6- 293.5 15.4 287.5 16.3 287 17.6 186 Tmethylthio)purine 2 8 8 (sh) 292.5 (sh) 292.5(sh)

57 293 28.5 288 29.9 288 29.9 2 8 8 (sh) 292 (sh) 292 (sh)

(187) C. P. Bryant and R. E. Harmon, J. Med. Chem., 10, 104-106 (1967).

(2) Carbon-13 n.m.r. spectra— In a recent report 188 by Fischer and coworkers, carbon-13 n.m.r. spectroscopy was shown to be useful for differentiating between purines substituted at the 7- and 9-position. It was observed that 13 the C-n.m.r. spectra of 2,6-dichloro-9-methylpurine is completely different from that of 2,6-dichloro-7-methylpurine.

This same distinctively different pattern is also found displayed in several 7- and 9-substituted purine nucleosides. 13 The following natural-abundance, C-n.m.r. spectroscopic data thus furnish a strong, additional line of evidence concerning the site of glycosyl attachment assigned on the basis of uv spectroscopy.

(188) P. Fischer, G. Losch, and R. R. Schmidt, Tetrahedron Lett., Yl_, 1505-1508 (1978).

13 The C-n.m.r. spectra of the l'-epimeric mixtures

17, 30, 31, and 38 as well as those of the pure epimers

17b, 30a, and 36a, were recorded in chloroform-d at ambient temperature with Me^Si as internal standard. All mixtures, with the exception of 31, were prepared in such a manner that one of the epimers preponderate by more than twofold over the other, so that the carbon signals for each isomer could be separately identified simply on the basis of peak 78 intensity. The spectra and carbon assignments will be discussed in detail later. Here, attention will be focused on the chemical shifts of carbon atoms in the 6-chloropurine component only. There are five purine carbon atoms altogether.

Their range of chemical shifts is from about 130 to 152 ppm downfield from Me^Si. The C-2 and C-8 carbon atoms are tertiary and can be readily differentiated by off-resonance decoupling from the remaining three carbon atoms, which have no hydrogen attached. A chemical-shift diagram for the purine- base region for 17a, 17b, 30a, 30b, 3JLa, 31b, 36a, 38a, and

38b is presented in Figure 2. The two distinctly different patterns of 7- and 9-substituted derivatives, as represented by 2,6-dich.loro-7-methyl- (B) and 2 , 6-dichloro-9-methyl purine (A), are also shown for comparison. The diagram clearly reveals the similarity of shift pattern among the acyclic- sugar nucleosides prepared in this study. It also illustrates that this pattern, with but small differences of individual shift, is in accord with the assignment of substitution at the 9-position.

In the case of 31a and 3Jb, the shift pattern at first seems to be similar to that of 7-substituted derivatives.

However, upon close scrutiny, several discrepancies are found. First of all, the C-5 carbon atoms are shifted upfield by less than 1.5 ppm, as compared with the large (^10 ppm) shift observed for 7-substituted purines. Secondly, a large downfield shift of C-4 (>10 ppm) is not observed. Instead, 79

C-2 C-4 c-6 c- g c-5

| i7b MO ISO MO IM IM

17a , 11 , 1

30a/%/ /■*-■ i * i1 1

30b II 1

| | 36a

| 38a ,

| 38b , • • A , 1

... 1 . . r : ... i

C-6| c-Zl C-8 c-5j 31a ( I

| 31b ... . | . |

■■■» » 1 » , .. | i 1*0 ISO MO IM 110

Figure 2: C-13 chemical shift diagram (CDC1.,) for

nucleosides 17a,17b,30a.3Ob,36a,38a,38b*** * /"V /V /V * /V /V ’ ^ /w* 2,6-dichloro-9-methylpurine (A) and 2,6- dichloro-7-methylpurine (B). 80

the C-4 carbon atoms are slightly shifted upfield (v3 ppm).

Finally, the large downfield shift of C-6 , which is the cause

of the deceptive superficial resemblance to the 7-substituted

pattern, actually arises because of replacement of chlorine

by an ethylthio group at the 6-position of the purine ring.

Therefore, if the C -6 shift is ignored, the pattern of shifts

of the remaining carbon atoms is more closely accord with the 13 9-substitution pattern. Another C-n.m.r. spectral method

for the assignment of glycosylation site in nitrogen hetero

cycles has also been developed. The methodology is based on 189 190 reports by Pugmire and Grant ' that, in a nitrogen

heterocycle system, when the free pair of electrons on nitrogen

(189) R. J. Pugmire and D. M. Grant, J. Am. Chem. Soc., 91), 697-706, 4232-4238 (1968).

(190) R. J. Pugmire, D. M. Grant, L. B. Townsend, and R. K. Robins, J. Am. Chem. Soc., 9^5, 2791-2796 (1973).

in the anion is protonated, an upfield shift for the carbon

a to the protonated nitrogen and downfield shift for the 6

and y carbons is observed. The a-substitution shifts have been explained on the basis of a decrease in bonding between

N“C , whereas the 6 and y shifts are the result of charge- 18 9 polarization effects. The large upfield a shifts and small downfield 6 shifts, termed the "protonation parameters",

are also observed in various heterocyclic base systems when compared with the base anion. 81

To determine whether the effect of N-glycosylation in these systems is similar to the previously reported five-, six-membered, and other fused-ring systems, the carbon-13 chemical shifts of 25 were examined and compared with those of the anion of 6-chloropurine, formed by neutra lization with lithium hydroxide in Me_SO-dr. The site of Z — D glycosylation in 25 has been firmly established by X-ray analysis. The acetylated nucleosides 49a, 52a, and 54b were also studied. All chemical shifts were recorded in Me_SO-dr 2 — 6 with reference to Me^Si, and are summarized in Table 5.

The assignment of chemical shifts for the nucleosides will be explained later. The assignments for the 6-chloropurine 190 anion are based on the model compound, purine anion and on peak intensities. The C-2 and C -8 resonances were readily identified by the presence of C-H splitting in the off- resonance decoupled spectrum. The resonance positions of C-2 and C-8 were assigned the same relative order (C-8 > C-2) as noted for the purine anion. This assignment was confirmed from the peak intensities, the C-2 signal being more intense than C -8 peak. The assignments for C-4 and C -6 were not definitive and again were based on the relative order reported for the purine anion. However, the following results support the correctness of these assignments.

In compound 25, the C-4 and C-8 carbon atoms are a to the glycosylated nitrogen (N-9) and C-5 is 8 and C-6 is y. A corresponding large, upfield shift of 8.44 ppm for C-4 Table 5. Comparison of carbon-13 chemical shifts for the nucleosides 25, 49a, 52a, and 54b.

Compound Chemical shifts3 C-2 C-4 C-5 C-6 C-8

6-Chloropurine anion (I) 149.03 159.80 131.35 145.87 152.67 (IS)-1-(6-Chloropurin-9-yl)-1- 151.36 151.36 130.33 148.94 146.52 S-ethyl-l-thio-D-glucitol (25) (1R)-1-(6-Chloropurin-9-yl)-1- 151.94 151.94 130.92 149.49 147.04 S-ethyl-l-thio-D-altritol (49a) (1R)-1-(6-Chloropurin-9-yl)-1- 151.84 151.60 130.73 149.33 146.89 S-ethyl-l-thio-D-mannitol (52a) (IS)-1-(6-Chloropurin-9-yl)-3- deoxy-l-S-ethyl-l-thio-D- 151.26 151.86 130.20 148.76 146.39 arabino-hexitol (54b)

A6I - 25 -2.33 + 8.44 + 1.02 -3.07 + 6.15

A(SI - 49a -2.91 + 7.86 + 0.43 -3.62 +5.63

A6I - 52a -2.81 + 8 . 20 + 0.62 -3.46 + 5.78

A6I - 54b -2.23 +7.94 + 1.15 -2.89 + 6.28

£ Shifts given in part per million downfield from Me^Si in Me 2S0 -dg. and 6.15 ppm for C-8 is observed (See Table 5). An upfield shift of 1.02 ppm, which contradicts the predicted downfield

6 shift, was observed for C-5. This reversal in trend was 189 also observed by Pugmire for bridgehead carbon atoms.

The long-range cross-ring effects noted at position C-2 in purine^^^ are also preserved in this system (-2.33 ppm for

25 as compared with -2.59 ppm and -2.37 ppm for 7-methylpurine and 9-methylpurine, respectively). This long-range electronic effect was also observed in the chemical shift for C-6 , which is 8 to the glycosylated nitrogen N-9. This downfield shift for 25 (-3.07 ppm) is comparable with that in 9-methylpurine (-3.75 ppm). Values for compound 25 confirmed that the protonation parameters in this system are no exception to those already observed. The chemical shifts for 49

(See Table 5), indicating that the position of glycosylation of 49a, 52a, and 54b are at N-9 of 6-chloropurine, the same as that in 25.

All of these three lines of evidence unanimously lead to the same conclusion and establish beyond doubt that N-9 is the site of glycosylation for 30a, 30b, 31a, 31b, 36a,

38a, 38b, 49a, 52a, and 54b.

2. Optical Properties and Chirality at C-l'— It has been proposed the sign of the optical rotation of a sugar- chain substituted, heterocyclic compound is determined by the 84 chiral center to which the heterocyclic base group is attached and is irrespective of the configurations at the other asymmetric carbon atoms. It is on this basis that the 191 Generalized Heterocycle Rule was formulated. The rule states

(191) H. El Khadem and Z. M. El-Shafei, Tetrahedron Lett., 27, 1887-1889 (1963). that, if the asymmetric carbon atom attached to the hetero cycle ring is represented in the Fischer projection with the ring on top, then the rotation is positive when the oxygenated group is to the right and negative when it is to the left.

The rule further states that methylation and acetylation of the chain hydroxyl groups have no influence on the sign of rotation. This Generalized Heterocycle Rule presumably can be extended to include acyclic-sugar nucleosides, because it may be argued that the replacement of oxygen by sulfur would represent a relatively minor change, as the two elements belong to the VI-A family of elements and sulfur is directly below oxygen in the periodic table. This argument is valid for several acyclic-sugar purine nucleosides whose absolute configurations have been established directly or indirectly by X-ray crystallographic analysis. Among them is the depro tected analog of 17b, referred to as compound 25, which is predicted from the rule and confirmed by X-ray analysis to possess the S configuration at C-l'. Table 6. Determination of the chirality at C-l for 6-chloropurine acyclic nucleosides

Compound Chirality Sign of C.D. curve [a] Generalized C.D. Generalized Heterocycle Measure- C.D. Observed Rule ment Rule

1-(6-Chloropurin-9-yl)- 0 1-S-ethyl-l-thio-D-glucitol -112 (25) 1-S 1-S— (25)

1-(6-Chloropurin-9-yl)-1-S- 0 ethyl-1-thio-D-mannitol (52a) + 87 (23) 1-R 1-R

1-(6-Chloropurin-9-yl)-1-S- 0 ethyl-l-thio-D-altritol (49a) + 89 (23) 1-R 1-R . +

1-(6-Chloropurin-9-yl)-1-S- 0 ethyl-1-thio-D-arabino-hexitol - 95 (29) 1-S 1-S (54b)

2.3.4.5.6-Penta-O-acetyl-l-S- ethyl-l-thio-1-(6-chloropurin- 0 9-yl)-D-glucitol (17b) -105 (25) 1-S 1-S—

2.3.4.5.6-Penta-O-acetyl-l-S- ethyl-l-thio-1-(6-chloropurin- 0 9-yl)-Q-mannitol (30a) + 62 (29) 1-R 1-R

2.3.4.5.6-Penta-O-acetyl-l-S- ethyl-l-thio-1-(6-chloropurin- 0 9-yl)-D-mannitol (30b) - 94 (20) 1-S 1-S Table 6. (Cont'd)

2.3.4.5.6-Penta-O-acetyl-l-S- ethyl-l-thio-1-(6-chloropurin- 0 9-yl)-D-altritol (36a) +104 (23) 1-R 1-R +

2.3.4.5.6-Tetra-O-acety1-1-S- ethyl-l-thio-1-(6-chloropurin- 0 9-yl)-D-arabino-hexitol (38b) - 74 (20) 1-S 1-S

— Numbers in parentheses indicate temperature; Solvent: 25, 52a, 49a Water; 54b MeOH; 17b, 30a,30b, 36a,38b CHC1_. b 3 . — Established by X-ray crystallographic analysis.

00 <31 87

The optical rotations of 30a, 36a, and 38b as well as their deprotected analogs 49a,, 52a, and 54b, were determined at sodium g line (589 nm) and are recorded in

Table 6 . The manno derivatives 30ja and 52a both show large O O positive rotations (+62 and +87 respectively). So do the altro derivatives 36ja and 49a,. in contrast, the pair of 3- deoxy-arabino derivatives display large negative rotations.

Evidently, acetylation has no influence on the sign of the rotation as would be predicted by the rule. According to this

Generalized Heterocycle Rule, the pairs of 30a, 52a. and 36a,

49a are assigned the 1'(R) configuration whereas the pairs

38b, 54b as well as 30b are 1'(S ).

Unlike the problem of site of chain attachment dis cussed earlier, the problem of routinely determining the absolute configuration at C-l' is still not completely solved.

Optical rotarory power is the only generally accessible physical property by which the pure enantiomers may be distinguished. If the relative configuration can be determined, the absolute configuration follows by simple chemical corre lation with the original sugar used in the synthesis. As the absolute stereochemistry at the chiral center C-l' of 25 and its acetylated analog 3/7b has been established as (S) , examination of the circular dichroism (c.d.) spectra should yield the relative and absolute configurations of the nucleo sides prepared in this study, because of the close structural similarity of these series of compounds. The c.d. spectra were.measured in the neighborhood of the uv absorption maxima

(^265 nm) from 350— 220 nm, in methanol at ambient temperature.

The c.d. spectrum, together with uv absorption data, of 17b is shown in Fig. 3, which serves as an important basis of reference. The c.d. spectra of 30a, 30b, 36a and 38b were L ./•>*— /V A- A/ A A-'A- measured and are shown in Fig. 4-6. Simply by comparison of these spectra, it is clear that the overall shapes of the spectra of 30

that of 17b.A- /V It may be concluded that 30b and 38b —' A- have the same relative configuration as that of 3^7b and are then assigned the absolute configuration (S) at C-l1, whereas

30a and 36a have the opposite relative configuration and are assigned (l'R). As the free-hydroxyl acyclic-sugar nucleosides give c.d. spectra similar to those of their acetylated counterparts, the same assignments are also made for compounds

49a, 52a and 54b as for their acetylated precursors. These assignments agree very well with those predicted by Generalized

Heterocycle Rule (See Table 6 ). Disregarding the rest of sugar chain, which contributes negligibly to the c.d. curve, the R and S conformers may then be viewed as mirror images to each other, as depicted below: 89

UV — u—u—&— CD —o-o—o—

EtSCH HCOAc I AcOCH I HCOAc

HCOAc e x lO I CH2OAc

~m>----- Z5T5----- Z8T5----- UoTT 1 7 0 (nrnT

Figure 3: UV absorption and CD curve of 17b uv CD 30 b 30o

o 2

-2

2 6 0 " 280 300 i_ W absorption and CD curve of 30a and 30b -10 5 - 5 1 te]xio 20 - U absorption and C curve of 36a 6 3 f o e v r u c CD d n a n o i t p r o s b a UV : 5 e r u g i F 0 4 2 0 8 2 0 6 2 uv CD 0 0 3 AcOCH HCOAc HCOAc HCSEt HCOAc CHoOAc I I I (nm) 0 2 3 m 91 EX 10 te]xio V D D f o e v r u c CD d n a b 8 3 f o e v r u c CD d n a n o i t p r o s b a UV : 6 e r u g i F 240 ~— b 4 5 0 6 2 0 8 2 — — 0 A-&-A - 0-0 0 0 3 - 38b ] — AcOCH ] EtSCH EtSCH HOCH HCOAC HCOAc HCOH HCOH HCH 54b CHgOAc ch I I I

(nm) 0 2 3 h o

Cl Cl SEt ; Ets

c C s R

This picture may be used to explain the near mirror image relationship of the c.d. spectra of the R and S 1-epimers.

Recently, another rule has been proposed by El Khadem.

This rule, termed the Generalized Circular Dichroism

is applicable for nitrogen heterocycles attached

(192) H. S. El Khadem, Carbohydr. Res., 59_, 11-18 (1977).

(193) H. S. El Khadem, in "Synthetic Methods for Carbohydrates", H. S. El Khadem Ed., ACS Symposium Series, 39, 1977, pp. 77-89. to the glycosyl chains, and correlates the sign of the Cotton effect induced by the heterocycle with the following vari ables: (a) configuration of the chiral center attached to the heterocycle, (b) direction of the C-l'-K) (e ) , and (c) posi- D — tion of the glycosyl chain relative to the dipole-moment vector (eB ) of the heterocycle (Rgg)• This complicated rule predicts that, for a ompound having (R) configuration at C-l 1 94

and having the glycosyl chain lying "below" the base, the sign

of the Cotton effect will be negative if, in. the most-stable

conformers, the C-l' ->-0 vector (eg) and the dipole-moment vector (e ) of the base both point in the same direction; the B sign will be positive if these vectors are pointed in opposite

directions. For the S configuration, the sign of the Cotton

effect is the reverse of that for the R configuration. One

approach to establish the dipole moment of the purine ring

system is to treat the fused ring as one entity. Such well-

studied compounds as adenine and guanine nucleosides were used to align the fused ring along the X-axis so that they gave the expected sign of the Cotton effect. An example that is closely related to a 6-chloropurine derivative is depicted below.

This above chart with the predicted sign of the Cotton effect is for the R configuration. Before the rule can be applied to acyclic-sugar nucleoside, two assumptions have to be made: (JL) replacement of oxygen by sulfur causes little or no change to the system and (2 ) the dipole moment vector of 6- chloropurine is about the same as that of adenine. To investi gate the applicability of this c.d. rule, compound 25 was chosen because of the stereochemical information available. N oH

Chart 14. X-ray structure of (IS)-1-(6-chloropurin-9-y1)-1-S-ethyl-l-thio D-glucitol (25)

i_n 96

According to the most stable conformer of 25, as determined

by X-ray analysis and ‘''H-n.m.r. spectroscopy (Chart 14), the

Generalized Circular Dichroism Rule predicts that the sign of

the Cotton effect will be positive for the S configuration,

and this is contrary to what is observed (See Fig. 3). El

Khadem noted that "this rotation rule is not applicable to 192 all acyclic compounds, because of hydrogen bonding". The

acyclic-sugar nucleosides prepared in this study, thus, fall

into non-applicable category, although not necessarily because

of hydrogen bonding.

3. N.m.r. Studies and Conformational Analysis— The

conformations of the acyclic-sugar chain is of great potential

importance in dictating the biological activity of these 127 acyclic-sugar nucleosides. It has been proposed that

biological activity depends on the ability of the sugar chain

to achieve a conformation that is isosteric with that of

naturally occurring nucleosides. A better understanding of the

acyclic-sugar chain conformations would be very helpful in the

design of this class of compound.

(1) ^H-n.m.r. spectroscopy— The proton-n.m.r.

spectra of acyclic-sugar nucleosides generally exhibit exten

sively overlapping multiplets that produce no useful informa

tion. Peracetylation of these nucleosides, however, may permit well separated signals if the proper solvent is chosen. The

compound (IS)-2,3,4,5,6-penta-0-acetyl-l-(6-chloropurin-9-yl) 97

-1-S-ethyl-l-thio-D-glucitol {17b) gave first order spectra in dimethyl sulfoxide-d^ containing 5% of chloroform-d. The 129 spin-coupling data obtained from this spectrum indicates that this compound assumes a conformation almost identical with that of its deprotected analog (25), whose solid-state conformation is established by X-ray analysis (Chart 14).

(1R)-2,3,4,5,6-penta-0-acetyl-l-(6-chloropurin-9-yl)-

1-S-ethyl-l-thio-D-mannitol (30a.) .

The 90 MHz n.m.r. spectrum of 30a in chloroform-d showed two singlets at 6 8.74 and 8.57 for H-2 and H-8 respectively. A quartet at 6 2.43 and a triplet at 1.14 corresponds to the methylene and methyl protons of ethylthio group. Among the sugar-chain protons (See Fig. 7), H-l' resonates at 6 5.92 as a doublet whose spacing gives (2.64 Hz). The signal for H-3' is a doublet of doublets at 6 5.63 from which the

J , value (1.72 Hz) was measured. The signals of H-2' and J r 4 H - 4 1 are two doublets of doublets at 6 5.31 and 5.42, which are partially overlapped. However, the coupling constant can be extracted as 9.48 Hz. The H-5' proton signal appears at 6 4.98, whereas H- 6 1 and H-6 " resonate at 6 4.20 and 3.99 respectively. The H-6 1 and H-6 " signals, which comprise the

AB portion of ABX system, are partially overlapping, four- line patterns. The coupling constants J,_, g, = 2.93 Hz,

— 5',6 " = 4.40 Hz and Jg , g„ = 12.33 Hz were measured. The coupling constant J4 , was calculated as 9.83 Hz based on HCSEt AcOCH

AcOCH

HCOAc HCOAc CHpOAc

LvM 1_/A axAiLv ± 6.0 5.0 4.0 Figure 7. Partial n.m.r. spectrum (chloroform-d) of 30a. 99

the fact that the total width of H-5' multiplet, which is the

X portion of ABX system, is equal to , g, + , g„

(IS)-2,3,4,5,6-penta-0-acetyl-l-(6-chloropurin-9-yl)-

1-S-ethyl-l-thio-D-mannitol (30b).

The 90 MHz n.m.r. spectrum of 30b, in chloroform-d, also gave

a nearly first-order resolution. The two purine protons H-2

and H-8 are observed as singlets at 6 8.72 and 8.49, respec

tively. Resonances of the ethylthio protons fall at <5 2.3 8 (q)

and 1.26 (t), whereas those of acetyl-methyl protons appeared at 6 2.12, 2.07, 2.03, 1.99, and 1.97. The H-l' signal (See

Fig. 8 ) is a doublet at 6 6.13 and H-2' gives a doublet of doublets at 6 5.74. The H-3 1 signal at 6 5.12 and H-4 1 signal at 5.42 are two partially overlapping doublets of doublets.

The ABX system of H-51, H-6 ', and H-6 " is very similar to that displayed in the spectrum of 30a. Couplings of J., = J- t ^ 4.45 Hz, J 2 , 3 , = 8.27 Hz, J 3 , 4 , = 2.54 Hz, , 5 , = 8.58 Hz,

Jc, = 3.50 Hz, J_. = 4.87 Hz, and 3C , = 12.55 Hz were b / O j ^ u D / u measured from analyses analogous to that described for 30a.

(1R)-2,3,4,5,6-penta-O-acetyl-l-(6-chloropurin-9-yl)-

1-S-ethyl-l-thio-D-altritol (36a).

The 90 MHz spectrum of 36a was recorded in chloroform-d and dimethyl sulfoxide-dr. The singlets for H-2 and H-8 are — b observed at 6 8.91 and 8.82 in dimethyl sulfoxide. The acetyl- methyl protons are observed at 6 2.07, 2.02, 1.99, 1.85, and

1.83. The ethylthio protons resonates at 6 2.44 (q) and 1.00 k EtSCH H-3 AcOCH

AcOCH H-6 ' H-5' HCOAc HCOAc

CH2OAc

6.0 5.0 4.0 3TCT

Figure 8 . Partial n.m.r. spectrum (chloroform-d) of 30b, 101

(t). The H-l' proton resonates at 6 5.85 as doublet, giving

J , , = 8.51 Hz. The H-2' proton resonance, which is a 1 f 2 doublet of doublets, fall at $ 6.03. The H-4 1 proton signal

is also a doublet of doublets. The spacings of these signals

give J , , = 2.05 Hz, J_, = 7.34 Hz, and J . , = 4.40 Hz. £ f -3 f rk A broad multiplet from 6 4.94 to 4.78 covers the signals for

H - 3 1 and H-5'. The signals of H-6 ' and H-6 " are merged into

each other to the extent that the two four-line pattern are

no longer recognizable. The spectrum in chloroform displays

extensively overlapping pattern.

(IS)-2,4,5,6-tetra-O-acetyl-l-(6-chloropurin-9-yl)-

3-deoxy-l-S-ethyl-l-thio-D-arabino-hexitol (38b).

The spectrum of 38b was recorded in benzene-d and chloroform ed. Because of the presence of methylene protons at C-3', the

spectra were more complicated and not completely resolved.

The spectrum in chloroform-d showed the H-2 and H -8 singlets

at 6 8.76 and 8.50. The coupling constant (4.28 Hz) can be measured from the H-l' doublet at 6 5.89. A clearly defined,

eight-line pattern at 6 5.57 is assigned to H-4' by elimination.

The spacings of the octet give three coupling values: 3.23,

5.25, and 10.56 Hz. Among three protons that would give an octet pattern, H-2' is excluded because the J., coupling 1 , z of 4.28 Hz is not among the measured couplings. H-5' is also eliminated for the same reason. A multiplet ranging from 6

5.17 to 4.87 contains the signals for H-2' and H-5'. The H-6 ' 102 and H-6 " signals are two four-line patterns that are slightly overlapped. The coupling constants J,., gl (3.81 Hz), J,., 6„

(7.20 Hz), and . r „ (11.64 Hz) are measured. The H-3' — D / D methylene proton signals appear upfield, from 6 1.74 to 1.43, as a complex multiplet, and no coupling data can be extracted.

Among the coupling constants measured, the J^, 3 , of

30a and J-, of 36a may deviate slightly from the absolute first-order values because of the second-order effect caused by the proximity of signals of coupled protons (6/J -v3) . The first-order spin-coupling data so obtained may be used for determining the most favorable conformations of the acyclic- sugar chain of these nucleosides. The following well founded 139 194 195 assumptions ' ' were made:

(19 4) P. L. Durette and D. Horton, Adv. Carbohydr. Chem. Biochem. , 2_6, 49-125 (1971).

(195) P. L. Durette, D. Horton, and J. D. Wander, Adv. Chem. Ser., 117_, 147-176 (1973).

(a) the sugar chain tends to stretch away from the

substituent at C-l1;

(b) vicinal pairs of protons in an antiparellel arrangement

have spin-coupling interactions of ^9 Hz;

^3 Hz.

For compound 30a, the observed J_ , and J , Z j 5 4 ,5 values (9.48 and 9.83 Hz, respectively), demonstrate 103 the predominance of the conformation having H-2 1 and H - 3 ', and H - 4 1 and H-5', antiparallel. On the other hand, the small values of , 2'' — 3 ' 4' (2.64 and 1.72 Hz) indicate a gauche relationship between H-l' and H-2', and H-3' and

H-4'. The conformation that is consistent with these vicinal coupling data is a planar, zig-zag arrangement of carbon 196 atoms 1-6 (Table 7). The symbolism P is assigned to indicate that this conformer is virtually the exclusive form. 197 There is no unfavorable parallel 1,3-interaction of acetoxyl group present in this arrangement.

(196) D. Horton and J. D. Wander, J. Org. Chem., 39, 1859-1863 (1974).

(197) H. S. El Khadem, D. Horton, and T. F. Page, Jr., J. Org. Chem., 33, 734-740 (1968).

In 30b, the coupling data are also entirely con sistent with the P backbone of extended planar, zig-zag arrangement advanced for £0a. The large couplings of ^ 1 2 ' and , (8.27 and 8.42 Hz, respectively) and small coupling of J^, 4 , (2.45 Hz) are similar to that of 30a. The value of

, 21 (4.45 Hz), however, is intermediate between that expected for the antiparallel and the gauche arrangement, suggesting that the dihedral angle of the H-l' and H-2' is O somewhat smaller than 60 . This is probably due to the parallel 1,3-interaction between the acetoxyl group at C-3' and the ethylthio group at C-l'. Alternatively, there may be 104

Table 7. Conformation and coupling constants for 1-R

and 1-S 2,3,4,5,6-penta-O-acetyl-l- (6-chloro-

purin-9-yl)-1-S-ethyl-l-thio-D-mannitols. (30a

and ^Ob,)

1-(R) 1-(S)

^ 1 -,2 ' 2.64 4.45

9.48 8.27 — 2 ',3' 1.71 2. 54 — 3' t 4 1 9.83 8.58 — 4 1,5' 2.93 3.50 — 5 ',6 ' 4.40 4.87 — 5 1,6 " 12.33 12.55 — 6 'f 6 ” OAc H OAC H SEt

D-manno-l-(R)

OAc OAc

AcO SEt OAc OAc

D-manno-l-(S) 105 substantial simultaneous population of both the H-l— H-2

antiparallel and one gauche conformer. By comparison with

30a, whose , 2 , coupling suggesting exclusive gauche

relationship, the latter seems less likely the case.'

For compound 36a, the coupling data measured are

, 2 i (8.51 Hz), J2 , (2.05 Hz), J3 , 4 , (7.34 Hz), and

— 4' 5' (4.40 Hz). Although no information was secured on the

couplings between H-5 1 and H-6 ', 6 ", the data are sufficient

to project the conformation between C-l' and C-51. The two

large couplings ( , 2, and , ^,) establish the antiparallel relationship between H-l' and H-2' and H-3 1 and H-41, whereas the small coupling (2.05 Hz) indicates the gauche relationship between H-2' and H-3'. The intermediate value of , 5 , (4.40.Hz) indicates that there is rotation about C-4— C-5 bond, which is expected because of the presence of unfavorable 1,3-inter action between 0-3 and 0-5 in the planar, zig-zag conformation.

This ^G+ conformation may be regarded as virtually the exclu sive form, which is free from parallel 1,3-interaction. The approximate conformation of 36a, is pictured below: Table 8. 90-MHz ^H-NMR Chemical-shift Data for 17b, 30a, 30b, 36a, and 38b.

C h e m ica l shifts (6) Acetate Ethyl Compd. Sol H-2 v e n t t h i o H - l ' H-2* H -3 ' H-4' H-5' H-6’ H-6" H-8

1 7 b a (CD3 ) 2 SO 5 .9 3 d 5.75t 5.3 8dd 5.56dd 5.05m 4 . 26dd 4 . 08dd 2.24s,2.08s 2 . 63q 8 . 8 8 s, 2.01s,1.98s 1 . 2 1 t 8 ■ 82 s -5%CDC13 1 .9 6 s

3 0 a CDC13 5.9 2d 5.3ldd 5.6 3dd 5.42dd 4.98m 4.20dd 3.99dd 2.26s,2.07s 2 . 43q 8 . 7 4 s , 2.03s,2.01s 1 . 1 4 t 8 .5 7 s

30b CDC13 6 .1 3 d 5 . 74dd 5.12dd 5.42dd 4.93m 4,20dd 3.98dd 2.12s,2.09s 2. 38q 8 . 7 2 s , 1.97s,1.87s 1 . 2 6 t 8 .4 9 s 1 .8 4 s

36a CDC13 6.05d 5.48—5.04m 5.72dd 5.48—5.04m 4.38-— 4 . 05m 2.12s,2.09s 2. 4 7q 8 . 6 9 s, 1.87s,1.84s 1 . 1 6 t 8 .4 5 s

(c d 3 ) 2 so 5.85d 6.03dd 4.94—4.87m 5.lOdd 4.94— 4. 78m 4.18-—3.85m 2.07s,2.02s 2. 44q 8 . 9 1 s , 1.99s,1.85s l.O O t 8 .8 2 s 1 .8 3 s

38b CDC13 5. 89d 5.17— 4.87m 1.74— 1.43m 5.57o 5.17— 4.87m 4.19dd 4.Oldd 2.04s,2.02s 2 .3 9q 8 . 7 6 s, 1.99s,1.93s 1 . 1 7 t 8 .5 0 s

3 From reference 129. Table 9. Proton-proton Spin-coupling Data for 17b, 30a, 30b, 36^, and 3j3b.

First--order couplings (Hz) Compd. Solvent “ 1 1 2 ~ 2 , 3 *3,4 -3' ,4 -4,5 -5,6 ~ 5 ,6 ' - 6 ,6 '

17b^ (cd3 )2so- 4.75 6.1 3.5 7.0 3.0 5.3 12.5 5% CDC13

30a cdci3 2.64 9.48 1.72 9.83 2.93 4.40 12.33

30b cdci3 4.45 8.27 2.54 8.58 3.50 4.87 12.55

36a cdci3 5.72 3.50 6 . 04

(c d 3 )2s o 8.51 2.05 7.34 4.40

38b CDC13 4.28 3.23 5.25 10.56 3.81 7.20 11.64

— From reference 129. 108

1 When the conformations, determined by H-n.m.r., of ll]Oj ^0a,

30b, and 36a are compared with those of their corresponding 19 8 peracetylated diethyl dithioacetals, striking similarities between them are observed. Both 17b, and D-glucose diethyl dithioacetal pentaacetate adopts the "sickle" conformation

(2G ) with C-l exo-planar, whereas 30a, 30b, and g-mannose diethyl dithioacetal pentaacetate assume the same extended, zig-zag conformation (P) . The same can be said of and

D-altrose diethyl dithioacetal pentaacetate as evidenced by the conformational identity of their C-l— C-5 fragment (^g "*”) .

(198) M. Blanc-Muesser, J. Defaye, and D. Horton, in press.

Furthermore, the same conformations also are adopted by their 199 corresponding alditois, both m solid-state and in solution.i 4.- 2 0 0

(199) G. A. Jeffrey and H. S. Kim, Carbohydr. Res., 14, 207-216 (1970).

(200) G. W. Schnarr, D. M. Vyas, and W. A. Szarek, J. Chem. Soc., Perkin I, 496-503 (1978).

128 129 The conformation of a number of other purine ' 137 201 and pyrimidine ' acyclic-sugar nucleosides synthesized in this laboratory have also been studied. The same kind of relationships also are found in these examples. Based on these studies, 137,194 202 ^ can conciU(je(j that 109 substitution at C-l by heterocyclic base and ethylthio group has little influence on the acyclic-sugar conformation, neither does O-acetylation of the free hydroxyl groups on sugar chain.

(201) D. Horton and R. A. Markovs, Carbohydr. Res., 80, 263-275 (1980).

(202) D. Horton and J. D. Wander, Carbohydr. Res., 10, 279-288 (1969); ibid., 13, 33-47 (1970); J. Defaye, D. Gagnaire, D. Horton, and M. Muesser, Carbohydr. Res., 21, 207-416 (1972).

(2) C-13 n.m.r. spectroscopy— The natural- abundance carbon-13 n.m.r. spectra of the epimeric mixtures

17, 30, 31, and 38 as well as those of the pure epimers

17b, 30a, and 36

(Figures 9, 10) and the chemical-shift data tabulated in

Tables 10 and 11. The spectra of pure epimers 17,b, 30a, 36a, and mixtures 17, 30, and 38 display very similar patterns.

The observed resonances for all of the compounds clearly separate into four groups: (a) those arising from the I I 67.45 67.79 68.18 I ! 71.04 152.01 k 61.68 26.23 144.35 HCSEt 60.08

AcOCH 43.97 AcOCH HCOAc HCOAC CHoOAc

131.70 151.39 W'Y’MvrtV

_l i I —J---1---1---1___I___i___I___i___lil. j I i I i

Figure 9. Carbon-13 n.m.r. spectrum (chloroform-d) of 30a. /X -A ' HCSEt I HOCH 71.11 I 71.54 HCOH 71.98 I 73.34 63.00 151.94 HCOH 24.94 I HCOH 14.55 I 62.37 c h 2o h 147.04

149.99 130.92

i____ I____ i____ I____ L-!___ i ' Figure 10. Carbon-13 n.m.r. spectrum (dimethyl sulfoxide-dg) of 49

Compd Carbon positions -sch2ch3 C-l' C-2 ’ C-3 • C-4 ' C - 5 ' C-6 ' C-2 C-4 C-5 C-6 C-8 ch2 ch3

17b 59.10 (71.76 67.61 69.43)- 68.17 61. 26 152.67 151.18 131.78 150.99 143.62 25.54 14.14 17a- 57.73 (71.79 67.85 69.64)- 68.07 61.29 151.97 151.97 131.75 151.12 144.85 25.86 14.17 30a 60.08 71.04 68.18 67.45 67.79 61. 68 152.01 152.01 131.70 151.39 144.35 26.23 13.97 3 0b^ 57.42 70.40 67.98 67.42 67.73 61.67 152.08 151.89 131.47 151.33 143.42 25.63 14.17 3 la- 56.74 70.43 68.05 67.42 67.95 61.50 151.99 148.93 129.82 161.76 141.75 25.42-14.65- 23.27-13.97- 31 b— 58.95 70.96 68.22 67.47 67.71 61.59 151.99 148.64 130.72 161.76 141.68 25.96-14.70- 23.27-13.90- 36a 59.97 (69.01 69.18 69.69 70.64)- 61.40 151.94 151.53 131.59 151.36 143.75 26.05 14.00 3 8a- 61.42 70.15--30.62 71.67- 67.66 61.78 151.12 152.12 132.05 151.79 144.00 26.31 14.19 3 8b- 61.12 69.66-32.50 71.67- 67.66 61.78 151.12 152.12 131.72 151.79 144.72 25.65 14.19

a — Shifts given in ppm downfield from Me.Si in chloroform-d. b — From spectra of 17, 30, 31, and 38 mixtures. Q — Not assigned. — Assignments may be reversed, s — Ethylthio group at C-l1 position. — Ethylthio group at C-6 position. Table 11. Carbon-13 Chemical-shif ts— for 25, 49a, 52a, and 54b.

CompdI. Carbon positions -sch2 CH3 C-l' C-2 ' C-3 1 C - 4 1 C - 5 1 C-6 1 C-2 C-4 C-5 C-6 C-8 ch2 ch3

25 61.46 73.48 69.99 71.44-71.25-63.20 151.36 151.36 130.33 148.94 146.52 24.48 14.30

49a 62.37 71.11 71.54 71.98 73.34 63.00 151.94 151.94 130.92 149.99 147.04 24.94 14.55

52a 63.13 72. 95 70.18 69.28 71.32 63.91 151.84 151.60 130.73 149.33 146.89 24.83 14.55

54b 64.40 67.67-^ d 68.49-74.81 63.05 151.26 151.86 130.20 148.76 146.39 24.69 14.38

—Shifts given in ppm downfield from Me.Si in dimethyl sulfoxide-d_. 4 —b —Assignments may be reversed. Q —Not assigned.

—The resonance overlapped with those of the solvent. alkyl carbon atoms (14-27 ppm), (b) those of the sugar-chain carbon atoms (57-73 ppm), (c) those of carbon atoms of the purine base (130-152 ppm, with the exception of 31), and

(d) those from the carbonyl carbon atoms (169-171 ppm).

Assignments were made by various considerations, which are to be discussed below:

(a) Alkyl region: The resonances that fall within this region include the C-l' ethylthio group and methyl carbons of the acetate groups. The assignments are rather straightforward.

The peak at highest field is CH^ of the SEt group because of its quartet splitting in the off-resonance decoupled spectrum, as well as its characteristic chemical shift, whereas the other peak is assigned to CE^ of SEt group because of its triplet splitting and chemical shift. The cluster of peaks in between arises, of course, from the CH^ carbons of the acetate groups. In the spectrum of 31, there are two additional peaks in this region, which correspond to the ethylthio group at the 6-position of purine. They may be assigned by their splittings and chemical shifts. In case of 3§, the methylene carbon at C-3' of the sugar chain also appears in this region, about 5 ppm downfield from the CI^ carbon signal of the SEt group.

(b) Carbohydrate region: The C-l' carbon atom, on account of the shielding effect of the ethylthio group,is shifted upfield away from the rest of the sugar carbons to the neigh borhood where the terminal C-6 ' carbon atom resonates. These 115

two carbons are assigned based on facts that C-l 1 is a

tertiary carbon (doublet splitting) and C-6 ' is secondary

(triplet splitting). The remaining carbon atoms C-2 1,3',4 1 ,5 1

(C-2',4',5' in 3J) are all tertiary, and are not specifically

assignable by off-resonance decoupling. Nevertheless, the

assignments in the spectra of 25a,, 49a,, and 52a, were readily made by comparing the spectra with those of the diethyl dithioacetal of D-glucose, p-mannose,and D-altrose, whose carbon signals are assigned in turn by comparing them with those of D-altritol^*^ as shown in Table 12.

The substitution effect of the 6-chloropurine group

is manifested at the substituted a-carbon (C-l1) and the

adjacent 6-carbon (C-21), whereas the more remote carbon atoms

experience negligible shielding changes.

The assignments for the acetylated nucleosides were not as straightforward because of the lack of suitable model

compounds. Definitive assignments can, however, be made by

single-frequency proton decoupling, if the chemical shift of

the proton attached to the specific carbon to be assigned can i be measured from the H-n.m.r. spectrum. In case of 30a, C-2",

C-31, and C-5' were so assigned by decoupling in turn the H-21,

H-31, and H-5' protons. The remaining resonance was then

assigned to C-4'. Using 30a as a model compound, 30b, 31a,

and 31b were assigned accordingly (See Table 10). The C - 5 1

signals of 38ja and 38b were assigned based on the observation^*^

that substitution of a hydroxyl group by a proton causes 116 negligible shift for the carbon atom y to the substituted carbon. The resonance at 67.7 ppm is the only one that closely matches the C-5' shift of 30a/V A/ (67.8 *■ ppm) and therefore is assigned as C-5'. All other carbon atoms were left unassigned,

(c) Purine-base region: There are five purine carbon atoms altogether. Two of them (C-2 and C-8 ) can be readily differ entiated from the remaining three by their peak intensities, because they are attached to hydrogen. The assignment is confirmed by their doublet splittings. Atoms C-4, C-5, and

C-6 are all carbon atoms having no attached hydrogen, and remained singlets under off-resonance decoupling. Model compounds were used to provide further assignments. The lone resonance, which appears especially well shielded (6 < 135 ppm) is assigned as C-5. This is a well noted fact for purines and 203 204 is validated by theoretical evaluations. ' The resonances for C-8 and C-2 were assigned by analogy with the assignment in the model compound 6-chloro-21,3'-0-isopropylidene-9-8-D- ribofuranosylpurine, i.e. C-2 downfield from C-8 . This same order has also been observed for various other purines,20^ '2C^5

(203) R. J. Pugmire, D. M. Grant, R. K. Robins, and G. W. Rhodes, J. Am. Chem. Soc., 8_7, 2225-2228 (1965).

(204) R. J. Pugmire and D. M. Grant, J. Am. Chem. Soc., 1880-1887 (1971).

(205) A. J. Jones, D. M. Grant, M. W. Winkley, and R. K. Robins, J. Am. Chem. Soc., 92, 4079-4087 (1970). 117

and it is quite clear that substitution at 6-position by

such functional groups as amino, chloro, oxo, and thio does not perturb this relative order. The remaining two resonances,

namely C-4 and C-6 were assigned from totally different .

consideration. In the model compounds, the resonance between

C-2 and C-8 was assigned as C-4, whereas the resonance that merged into that of C-2 was assigned as C-6 . This convergence

of two resonances were also observed in most of the spectra

studied. The peak hidden under the much more intense C-2 peak

can be checked of its presence, in off-resonance decoupled-

spectra, as a singlet between the doublet splitting of C-2.

When the spectrum of 25 is compared with that of 6-chloropurine

anion, the a-shift at C-4, as discussed in section 1 (p. 80 ), would be negative, as opposed to the expected large positive value, if the assignments of model compound are followed.

However, if the C-4 and C-6 assignments are reversed, all

shifts fit very well with those predicted. Therefore, the resonance at higher field was assigned as C-6 and the merged resonance C-4.

The spectrum of 31 is somewhat different because of the replacement by ethylthio of the 6-chloro group. A large 205 downfield shift is generated by this sulfur substituent, while the rest of the purine carbons are relatively unper turbed. The result of this shift is the crossing over of the shifts of C-6 and C-4 so that the relative order is now

C-6 > C-2 > C-4 > C -8 > C-5. 13 Table 12. C chemical -shift data— for nucleosides 25., 49a, and 52a.

Compound C-l' C-2 ' C-3 ' C-4 ' C-5' C-6 1 o| r- i-H

25 61.5 73.5 70.0 • 71.3- 63.2 (+7.4)- (-1 .8 ) (+0 .2 ) (-0.7) (-0 .1 ) (-0 .2 )

49a 62.4 71.1 71.5 72.0 73.3 63.0 (+7.9) (-1 .2 ) (+0 .1 ) (+0 .1 ) ( 0 .0 ) (+0 .1 )

52a 63.1 73.0 70.2 69.3 71.3 63.9 (+8 .1 ) (-0 .8 ) (+0.4) (-0 .1 ) (-0 .1 ) (+0 .1 )

£ — In ppm downfield from Me^Si in (CD^^SO. — Values in parentheses denote shielding with respect to the corresponding carbon of the parent aldose diethyl dithioacetal; (+) denotes shielding and (-) denotes deshielding. — Assignments for these peak positions may be reversed. 119

(d) Carbonyl region: All resonances in this region are attributable to the carbonyl carbons of acetate. The number of acetate groups present may be counted if all the acetate signals are separated from each other. This is the case for the spectra of 17b and 36a, in which all five acetate peaks are not overlapped.

As clearly shown in Table 12, the chemical shifts of

C-3 to C-6 in the spectra of nucleosides 25, 49a, and 52a closely match those of their corresponding aldose diethyl dithioacetal, indicating close conformational similarities of the sugar chain. The large positive a-shift of C-l 1 and small negative 6-shift at C-2' arises from substitution at

C-l" by 6-chloropurine. These results of C-13 n.m.r. studies are in good agreement with the preceding ^H-n.m.r. observations, and strongly support the generalization that the most favored conformation of the sugar chain of acyclic-sugar nucleosides will be the one similar to that of the corresponding alditol.

This generalization should prove very useful in the design of this class of compound, because it can be used to determine in advance the capability of the sugar chain to mimic the sugar moiety of the naturally occurring nucleosides and possibly to predict its properties at a particular binding site in a biological process.

4. Mass Spectra— The electron-impact mass spectra of the fully acetylated, acyclic-sugar nucleosides 17, 30, 31,

36, and 38 were recorded and analyzed. Published works on 120 sugar diethyl dithioacetals^*^ and sugar acetate"2^ ' were useful guidelines in interpretation of these spectra.

Comparison of these spectra was made, and it was found that all of these compounds displayed very similar fragmentation patterns, which were distinctively different from those of 209 acyclxc-sugar nucleosides. Mass spectra of 1-epimeric

(206) D. C. DeJongh, J. Am. Chem. Soc., 8j>, 3.149- 3151 (1964).

(207) K. Biemann, D. C. DeJongh, and H. K. Schnoes, J. Am. Chem. Soc., 85, 1763-1771 (1963).

(208) D. C. DeJongh and K. Biemann, J. Am. Chem. Soc., 85, 2289-2295 (1963).

(209) S. J. Shaw, D. M. Desiderio, Y. Tsuboyama, and J. A. McCloskey, J. Am. Chem. Soc., 92, 2510-2522 (1970). mixtures and the pure epimers were recorded and examined.

Virtually no difference was found between them, which leads to the conclusion that the stereochemistry at C-l 1 has no effect on the fragmentation process. It is evident, when the spectra of 17, 30, and 36 are compared, that the spatial arrangement of acetate group also has no effect. Furthermore, the fact that 31 displayed a fragmentation pattern similar to those of 17, 30, 36, and 38 seems to suggest that substi tution on the heterocyclic ring also has no influence on fragmentation process. Despite these facts, mass spectro scopy proved to be a very useful tool for characterizing gross structural features of these acetates. Each example 121

displayed the molecular ion, together with M+l and M+2 ions

caused by isotopes. The most intense peak in the spectra is

at m/e 43, corresponding to the acetyl ion CH 3C0+ . Loss of

acetic acid (60 m.u.), ketene (42 m.u.), and acetic anhydride

(10-2 m.u.) are important modes. Other important ions are

those representing the base: B, B+H, and B+2H. These are the

most generally characteristic ions in nucleoside and are

very informative about the structural feature of base. For

example, in the case of 6-ethylthiopurine derivatives (31),

the base ions have m/e 179, 180, and 181, corresponding to

6-ethylthiopurine. The following scheme (See Figs. 11-15)

is postulated for the probable fragmentation mode for 1/7, 30,

and ^6 . A detailed assignment of all fragments is not intended

here. Instead, with careful studies and comparisons of these

spectra, fragmentation processes characteristic of this class

of compound were elucidated and all peaks of higher m/e value

were accounted for. Many of the fragmentation proposed were

corraborated by metastable peaks, as marked with asterisks.

Each nucleoside displays a sizable molecular-ion peak, which allows direct determination of the molecular weight.

Several fragmentations derived directly from the molecular

ion are shown in figure 11. Loss of base by carbon— nitrogen

bond-cleavage and subsequent elimination of thioester leads

to fragment A^. The fragmentation series of A^ (See Fig. 12) 207 has been well studied by D. C. DeJongh and coworkers with

the aid of deuterated compounds. It is worth mentioning that 122

S E t A cO I^ C I HO CHjCOSEt (C H O A cL AcO— H I * CHgOAc AcO OAc

m/e 435 m/e 331 (A^)

- B-

AcOHgC

C H SE t SEt H O “ CHoCO I (C H 0A c)4 *(472.3) (CHOAcK*446-447 I H (446.3) CH^OAc CH20A c M + m /e 588 ‘ m/e 527 HOAc -E t *474 -HOAc ( 4 7 4 . 1 )

B I C SE t II AcO OAc

AcO ?H m/e 485 (B^) COAc

iAc (CHOAc)5 I * CHgOAc m/e 439 (D

m/e 528 (C^)

* Corresponding metastable peaks (calculated value in parentheses)

Figure 11. Probable fragmentation from molecular ion. 123

R^HgC

-2H 0A C A, C H f ° . / T \ V h

OAc OR, R,+R2 m/e 211 m/e 169 = H ♦ CH -jC O or AcOHoC AcOHgC

Figure 12. Fragmentation of series 124

A cO H^C -H O A c -H O A c B. AcO B

AcO H m/e 425 m/e 365

Figure 13. Fragmentation of series *327-329 (327.2) *327-329 -329 -329 (328.3) *415 *415 (414.8) HOAc A O -H - HOAc h c iue 5 Famnain f series of Fragmentation 15: Figure iue 4Famnain f series of ]4:Fragmentation Figure 2 o c OH^C H cO A % " — b H / 379 m/e ^ C N SEt / 366 m/e / 468 m/e OAc CH? COAc c A O C h c II o = c c COAc CH CHgOAc II 1 1 1 I 1 1 / ® 1

! - 0 = 0 * 2 1 m B

J j ♦ + *356-358 (355.7) *356-358 *256(255.8) HOAc - HOAc A O -H - C0 ^C H C 1

2C H O 2 R - m/e 337 337 m/e S t SE B " k '

b * V / c X Et SE OR! / 408 m/e / 306 m/e h c CH h c c COAc o = c II II c II CH CH II COAc 1 11 1 1 II 1 «h*CH 2 2 j+ g Rj R + J 3 C0 125 4

126 this series is also observed in the spectra of methyl and phenyl glucoside tetraacetate as well as of the peracetylated hexoses.

Loss of the alkylthio radical by fission of the carbon— sulfur bond, followed by the elimination of ketene, forms fragment B^. This is a very unusual process, because loss of ketene is generally preceded by loss of acetic acid, i.e. formation of vinyl acetate. Nevertheless, this sequence is substantiated by the detection of a metastable peak at m/e 446, and no other simple, obvious mechanism can be formulated to account for the peak at m/e 485. The fragmen tation series starting with (See Fig. 13) is just a logical and apparent sequence of eliminations based on the generali zations obtained from numerous studies of peracetates of pentoses and hexoses. 2°6-208 > 23-0

(210) N. K. Kochetkov and 0. S. Chizhov, Adv. Carbohydr. Chem. Biochem., 21, 39-93 (1966).

The fragment C^ results from loss of a molecule of neutral acetic acid. This is also substantiated by an observed metastable peak at m/e 474 (theoretical value: 474.1).

The proposed fragmentation series C^ is shown in figure 14.

The uniqueness of this series is that all of the fragments preserve the acyclic nature of their carbohydrate moiety.

This series proceeds with consecutive loss of three molecules of acetic acid, followed with a loss of ketene, and ends with 127 the loss of the remaining molecule of acetic acid. The supporting evidence for this proposed series is rather strong.

First of all, every single metastable peak expected from proposed steps of elimination is detected. Two of the metastable peaks coincide with the two characteristic, intense and broad metastable peaks at m/e 356-358 and 327-329, which are observed in all of the nucleosides studied here (m/e 381-

383 and 352-354 for 3J.) . Secondly, the intensities of fragments in this series are distinctly stronger than those in other series (See Table 14). This is in agreement with the observed intense metastable peaks described earlier and is a strong indication that the series is the major mode of fragmen tation for this type of molecule.

Finally, overall loss of two molecules of acetic acid and an ethyl radical leads to fragment D^. This series is similar to a series in the mass spectra of peracetylated 211 dialkyl dithioacetals. This fragment can be drawn as a

(211) S. Hanessian, Methods Biochem. Anal., 19^, 105 (1971). stabilized six-membered ring. This series (See Fig. 15) consists of sequential loss of acetic acid and ketene, which is typical of these type of molecules.

This proposed fragmentation sequence is also applicable for interpretation of the spectra of 31. The spectra of 38 differs from the others, as would be expected, but the 128 difference is surprisingly small. The most noticeable differ ence is the position of the characteristic metastable peaks discussed earlier. In 38, one of the metastable transitions fell at m/e 299-301, as expected from the proposed scheme, while the other was found at m/e 239-241, which is far lower than the predicted value of 269-271. This shift is best explained by a slight modification of the series such that three consecutive losses of acetic acid takes place, rather than loss of two followed by elimination of a ketene molecule hypothesis. This is supported by the intensity of the fragment at m/e 290 as opposed to the one at m/e 308

(See Table 14). Therefore, in both the original and modified

series, sequential loss of three molecules of acetic acid is favored.

Several prominent fragments of lower m/e value are listed in Table 15 together with their probable structure.

No useful information could be obtained from the electron- impact mass spectra of the deprotected nucleoside analogs because they decompose prior to ionization. However, moclecular ion and simpler fragmentations were observed in some of these 13 8 deprotected nucleosides by the technique of chemical- ionization mass spectrometry (c.i.m.s.). 129

Table 13. Relative intensities of fragmentations in the mass-spectral decomposition of acyclic-sugar nucleosides 17, 30, 31, 36, and 38.

m/e 17 30 36 m/e 31 m/e 38

590 0.7 0.4 0.6 616 0.1 532 0.1 589 0.2 0.3 0.5 615 0.2 531 0.2 588 1.5 1.0 1.0 614 0.6 530 0.5 528 2.4 1.4 1.0 554 0.5 470 0.4 527 2.2 3.0 1.2 ' 553 0.4 469 0.5 485 1.6 0.9 0.8 511 0.2 427 0.2 468 10.0 8.0 7.0 494 3.0 410 2.0 439 1.0 0.2 0.4 465 0.1 381 0.3 435 0.5 0.3 0.3 435 0.1 376 0.1 425 1.8 --- 0.9 451 0.5 367 0.3 408 24. 0 16.8 9.6 434 0.3 350 6.0 379 1.9 1.0 1.4 419 0.3 343 0.4 367 3.8 2.5 2.0 405 0.1 321 1.5 366 2.7 2.0 1.4 393 0.5 308 0.3 365 4.0 3.0 2.0 392 1.8 307 0.2 348 0.8 0.3 0.5 375 0.4 290 5.4 337 3.3 0.1 2.0 363 0.1 279 P. 6 331 2.9 1.0 1.8 333 0.6 273 2.3 323 1.2 0.4 1.0 332 2.0 249 0.4 306 4.7 3.2 2.8 331 1.5 248 0.6 293 1.3 1.0 0.8 290 1.1 228 3.3 269 2.0 1.2 1.2 269 0.5 . 227 2.9 263 1.5 1.0 1.2 263 0.1 205 3.4 259 1.7 1.3 1.0 259 0.9 201 1.4 251 5.1 4.0 3.4 251 1.4 194 1.2 239 3.8 2.4 2.2 239 2.6 181 0.8 229 5.2 3.6 4.0 229 2.6 171 1.5 130 Table 13. (Cont'd)

m/e 17 30 36 m/e 31 m/e 38

227 12.0 8.4 9.4 227 0.1 168 3.6 217 6.5 6.0 5.2 217 1.5 157 4.5 198 0.8 0.8 0.8 198 0.8 155 11.3 197 5.0 4.0 3.8 197 2.0 149 2.6 189 0.7 0.8 0.8 189 0.4 142 1.5 188 0.4 1.0 0.4 188 4.0 130 1.0 187 1.6 2.8 1.4 187 5.0 129 1.3 169 7.4 4.6 8.8 180 1.9 111 1.5 158 2.7 3.1 2.6 179 7.0 103 10.0 157 8.3 8.0 7.4 169 15.0 100 1.1 156 2.3 2.0 2.6 157 15.0 94 2.1 155 19.0 16.0 16.0 127 6.2 43 100.0

167 3.3 3.4 4.2 115 14.9 — ------

115 6.8 10.0 8.0 103 25.0 — ------

103 3.2 5.0 5.0 43 100.0 — ------

43 100.0 100.0 100.0 ------131

Table 14. Relative intensities of fragments in series C-, from 17, 30, 31, 36, and 38.

m/e 17 30 36 m/e 31 m/e 38

528 2.4 1.4 1.0 554 0.5 470 0.4 468 10.0 8.0 7.0 494 3.0 410 2.0 408 24.0 16.8 9.6 434 0.3 350 6.0 366 2.7 2.0 1-4 392 1.8 (308 0.3) 306 4.7 3.2 2.8 332 2.0 290 5.4

Table 15. Probable structure of prominent smaller ions formed in the mass spectral fragmentation of 17, 30, 36, and 38.

m/e Probable structure of ion

227 B-CH=SC9H,- s 5 209 B-C=C+ or

43 ch3co+ IV. EXPERIMENTAL

General Methods Infrared spectra were recorded with

a Perkin-Elmer Infracord spectrophotometer and ultraviolet

spectra with a Cary 15 u.v. recording spectrophotometer.

Melting points were determined by a Thomas-Hoover "Unimelt"

apparatus and uncorrected. X-Ray powder diffraction data give O interplanar spacings in A for CuKa radiation (camera diameter

= 114.59 cm). Relative intensities were estimated visually;

m, moderate; s, strong; vw, very weak; w, weak. Optical

rotation, was measured with a Perkin-Elmer Model 141 recording

polarimeter. Electron-impact mass spectra were recorded by

C. R. Weisenberger with an AEI MS-9 double-focusing mass

spectrometer at an ionization potential of 70 eV and an 1 accelerating potential of 8 kV. H-N.m.r. spectra were

recorded at 100 or 90 MHz with Varian HA-100 or Bruker HX-90 O spectrometers at ^25 . All samples were dissolved in chloro-

form-d with tetramethylsilane (Me^Si) as the lock signal

unless otherwise noted. The following conventions are used

to refer to n.m.r. spectra: d, doublet; dd, doublet of doublet; m, multiplet; o, octet; q, quartet; s, singlet; t, triplet. 13 C-N.m.r. spectra were recorded by C. Cottrell with a

132 133

Bruker WP-80 spectrometer operating at 20 MHz in the Fourier- O transform mode at ^25 . Chemical shifts were reported in ppm relative to Me^Si. O.r.d. and c.d. spectra were recorded with a Durrum-Jasco o.r.d.-c.d. spectrometer at ambient temperature with a 1-cm optical cell. T.l.c. was performed on silica gel 60 F-254 (#5765, E. Merck). Preparative t.l.c. was performed on chromatoplates (200 X 200 X 2.5 mm) of silica gel 60PF-254 (#7747, E. Merck) or microcrystalline cellulose

(Avicel-TG-101, FMC Corporation), containing 1% LUmilux Green

25. Silica gel 60 (#7734, E. Merck) was used for all liquid- column chromatography. High pressure liquid chromatography was performed on a Waters Associates ALC-244 system equipped with U 6K injector, M-6000 pump, 440 u.v. detector, and

R 401 R.I. detector. Microanalyses were mainly performed by

W. N. Rond, with some analyses by Galbraith Analytical

Laboratories, Knoxville, Tenn.' 134

Preparation of (IS)-2,3,4,5,6-penta-0-acetyl-l-

(6-chloropurin-9-yl)-1-S-ethyl-l-thio-g-glucitol(17b)---

Method A ; improved route. 2,3,4,5,6-penta-0-acetyl-l- 129 bromo-l-£>-ethyl-l-thio-g-glucitol (5.5 g, 11.09 mmol) O was added to a mixture, preheated to 60 , of 6-chloropurine

(1.60 g, 10.35 mmol), mercuric cyanide (1.65 g, 6.55 mmol), cadmium carbonate (1.16 g, 6.74 mmol), calcium sulfate (1 g), and nitromethane (120 mL). The solution was stirred vigorously O for 4 h at 60 . The mixture was filtered hot and the filter cake washed with several portions of nitromethane. The filtrate O was evaporated under diminished pressure at 50 , the resulting syrup was extracted with hot chloroform and the insoluble, yellow precipitate filtered off. The filtrate was washed successively with 30% aqueous potassium iodide (three times) and water (three times), dried over sodium sulfate, and evaporated to a syrup. Elution with 1:1 benzene— ethyl acetate from a column (2 X 205 cm) of silica gel yielded the syrup 17

(3.6 g, 59%) as a 5:1 mixture of the IS and 1R anomers.

The (IS) epimer (17b) was isolated crystalline upon refrigeration of the foregoing syrup in abs. ethanol; yield

1.73 g (29%). m.p. 138 , [a]^5 -105 (c 1.0, chloroform);

1H-n.m.r. (100 MHz): 6 8.62 (H-2, purine H), 8.47 (H-8, purine H), 5.91 d ( , 2, 3.5 Hz, H-l1), 5.74— 5.48 m

(H-2',3*,4'), 5.05 o (H-5'), 4.22 dd, 4.01 dd (Jg, gI 3.0 Hz,

Jg, g„ 6.0 Hz, J g , g„ 13 Hz, H-6 ', H-6"); 13C-n.m.r. (chloro form): 14.14 (SCH2CH3), 25.54 (SCH2CH3), 59.10 (C-l') 135

68.17 (C-5'), 61.26 (C-6 '), 67.61, 69.43, 71.76 (C-2',3'4'),

131.78 (C-5), 14.3.62 (C-8), 150.99 (C-6), 151.18 (C-4), and

151.67 (C-2). 129 ° 2S ° Lit. data: m.p. 137-138 , ta]£ -105 (c 1.0, chloroform); ^H-n.m.r. (chloroform-d, 100 MHz), 8.70 s, 8.54 s

(H-2,8), 5.98 d ( , 2, 3.5 Hz, H-l1), 5.76 — 5.58 m

(H-2',3',41), 5.12 o (H-51), 4.30 dd, 4.10 dd (Jc, CI 3.0 Hz, — 5 , b J 5 ,^6„ 6.0 Hz, J6,f6„ 12.5 Hz).

The foregoing physical characterizations established beyond doubt that this crystalline compound is (lS)-2,3,4,5,6- penta-O-acetyl-1-(6-chloropurin-9-yl)-1-S-ethyl-l-thio-g- glucitol.

Method B ; new route. A mixture of 2,3,4,5,6-penta- 163 O-acetyl-g-glucose diethyl dithioacetal (5.0 g, 10.08 mmol),

6-chloropurine (1.56 g, 10.09 mmol), mercuric cyanide (2.55g,

10.12 mmol), mercuric oxide (1.37 g, 6.34 mmol), calcium sulfate (1 g), and nitromethane (120 mL) was boiled under O reflux for 4 h at 110 . The mixture was then treated as described in Method A . Elution with 1:1 benzene— ethyl acetate from a column (3.5 X 60 cm) of silica gel yielded the syrup

17 (3.5 g, 59%), whose epimeric ratio was 5:1 in favor of the

(IS) epimer. The (IS) epimer was isolated from the syrup by the same method as described in Method A ; yield 1.6 g (27%).

This compound was found to be identical with the compound prepared in Method A .

Small-scale syntheses (0.5 g and 0.25 g) by this method (Method B) gave consistently better yields (v70%). 136

The epimeric ratios of the products obtained from different trials are all close to 5:1 in favor of the (IS) epimer.

When acetonitrile was used as solvent in place of nitromethane, the yield was slightly lower but the reaction progressed at a much lower rate.

2 ,3 ,4 ,5 ,6-Penta-O-acetyl-l-bromo-l-S-ethyl-l-thio-

D-mannitol(29)---2,3,4,5,6-Penta-O-acetyl-g-mannose diethyl

16 4 dithioacetal (1.92 g, 3.87 mmol), was dissolved in anhydrous ethyl ether (30 mL in 15 mL of ether). The solution O was cooled to 0 and bromine (0.2 mL in 15 mL of ether) was added dropwise during 20 min with stirring. The solution was warmed to room temperature and stirred for another 3 0 min.

Ether was then removed at ^25 torr at a bath temperature of O 35 . The dark syrup was washed with fresh ether and evaporated again. The process was repeated two more times to give 2j) as a light-yellow syrup; yield 1.91 g (95%).

This reaction was monitored on t.l.c. with the solvent system of 1:1 benzene— ethyl acetate. The completeness of the reaction, judged by the appearance of 29, which is faster migrating than the starting compound, and the disappearance of the peracetate, was found to vary with each trial.

(JLR) -2 , 3 , 4 , 5, 6-Penta-O-acetyl-l- (6-chloropurin-9-yl) -

1-S-ethyl-l-thio-D-mannitol(30a) Method A . Compound 29

(1.91 g, 3.67 mmol), freshly prepared by the foregoing procedure, was added to a mixture of 6-chloropurine (0.58 g,

3.77 mmol), mercuric cyanide (0.6 g, 2.38 mmol), cadmium 137 carbonate (0.42 g, 2.4 mmol), calcium carbonate (1 g) , and nitromethane (60 mL). The solution was boiled for 12 h at O 110 under reflux with vigorous stirring. The mixture was then treated by the general isolation procedure described for preparation of the glucose derivative 17. The resulting syrup was eluted with 1:1 benzene— ethyl acetate through a column

(2 X 205 cm) of silica gel, and the desired product 30 was obtained as a chromatographically homogeneous, pure syrup

(1.28 g, 58%). The light-yellow syrup was taken up in the minimum amount of abs. ethanol and kept at room temperature.

Granular crystals were formed after one day; yield 0.44 g

(20%), m.p. 107-108 ; ta]^3 +61.2 (c 1.04, chloroform);

V Dv- R„ 0.38 (1:1 benzene— ethyl acetate); v 2980 (C-H), — I? msx 1750 (C=0 of acetate), 1590, 1570 (purine ring), 1375, 1215,

1050 cm-1 (C-O-C); c.d.: (0]33O 0°, [0]282 -5,294°, [0]274 °°/

*-0-*265 +12'941 <■ 16] 255 0 ’ ^ 250 -6'470 ' ^ 243 0 '

E 0] 237 +6'176 ' f0^231 0 ' and E 221 ~16/471 > 1H-n.m.r. (90 MHz): 8.74 s (H-2), 8.57 s (H-8), 5.92 d (J , 2.64 Hz, 1 , H-l1), 5.31 dd (J2 , 3 , 9.48 Hz, H-21), 5.63 dd (J3 , 4 , 1.72 Hz,

H-3 1) , 5.42 dd (J4 , 5 , 9.83 Hz, H-4 ’) , 5.10 — 4.90 m (H-5 1 ) ,

4.20 dd, 3.99 dd (Jc , CI 2.93 Hz, J c , cu 4.40 Hz, ^ „ D ^ O J ^ D D f O 12.33 Hz, H-6 ', H-6"), 2.43 q (SCH2CH3), 2.26 s, 2.07 s,

2.03 s, 2.01 s (acetates), and 1.14 t (SCH^CH^) ; ^C-n.m.r.

(chloroform-d): 13.97 (SCH2CH3), 26.23 (SCH2CH3), 60.08 (C-l'),

71.04 (C-2'), 68.18 (C-3'), 67.45 (C-4'), 67.79 (C-5'), 61.68

(C-6'), 131.70 (c-5), 144.35 (C-8), 151.39 (C-6), and 152.01 (C—4 and C-2); X-ray powder diffraction data: 3.82 w,

3,89 w, 4.11 m, 4.26 m, 4.43 s (1), 6.08 vw, 6.40 w, 6.69 w,

7.76 m (3), 8.88 s (2), and 11.21 w.

Anal. Calc, for C 23H 2gClN4O 10S : C, 46.93; H, 4.97;

Cl, 5.59; N, 9.52; S, 5.44. Found: C, 46.96; H, 5.11; Cl, 6.24

N, 9.49; S, 5.62.

Method B . A mixture of 2,3,4,5,6-penta-O-acetyl-D-

mannose diethyl dithioacetal (5.0 g, 10.08 mmol),

6-chloropurine (1.56 g, 10.09 mmol), mercuric cyanide (1.58 g,

6.27 mmol), mercuric oxide (1.33 g, 6.16 mmol), and nitro- O methane (100 mL) was boiled for 4 h at 110 under reflux.

The mixture was processed as in Method A . The crude syrup was

purified on a column (2.5 X 200 cm) of silica gel. Elution with 1:1 benzene— ethyl acetate gave a syrup; yield 3.76 g 1 (63%), whose H-n.m.r. spectrum showed it to be a 2.7:1

epimeric mixture.

Likewise, the (1R) epimer was isolated from the syrup

by crystallization from abs. ethanol, yield 1.31 g (23%).

This compound was identical with the material prepared by

Method A . With seeding, the (1R) epimer could also be isolated

from the crude syrup obtained by either method prior to

chromatographic purification. After the crystallization, the epimeric ratio of the remaining mother liquor was about 2:1

in favor of the (1R) epimer as shown by 1H-n.m.r. A negli gible amount of crystals was obtained by crystallization of the concentrated mother liquor from such less polar solvent 139 systems as hexane— chloroform, ether— benzene.

(IS)-2,3,4,5,6-Penta-O-acetyl-l-(6-chloropurin-9-yl)-

1-S-ethyl-l-thio-p-mannitol(30b) The mother liquor from the preceeding crystallization was further purified by high- pressure liquid chromatography to separate the (IS) from the

(1R) epimer. The syrup (20 mg portions) was injected into a y-Bondapak Cno column (30 cm X 7.8 mm I.D., Waters Associates) ± O and eluted with 2:3 acetonitrile— water at a flow rate of -2 1 mL/min and at a pressure of 1500— 2000 lb. in . The eluate was monitored by both uv absorbance (254 nm) and refractive index. Separation was achieved by recycling and "shaving".

After six or seven recycles, the faster-eluting peak was collected. Removal of solvent, first by evaporation in vacuo, followed by freeze-drying, gave a syrup that was further purified by applying it to a preparative t.l.c. plate

(200 X 200 X 2.5 mm) of silica gel, which was developed twice with 2:1 hexane— ethyl acetate. The major zone was excised and extracted with chloroform to give, upon evaporation of solvent, pure 30b as a homogeneous syrup; 8 mg. The yield could not be determined because the epimers were not completely resolved 20 0 and pure 30b was obtained by "shaving". [ ex] D -93.9

(c 0.3, chloroform); R„ 0.38 (1:1 benzene— ethyl acetate); — — F 264.5 nm (e 5,000); vKBr 2980 (C-H) , 1750 (C=0 of iucix nicix acetate), 1590, 1570 (purine ring), 1375, 1215, 1050 cm-^

(C-O-C); c .d.: [6] 305 0°, [0] 284 +5,637, [ 0]2?7 0°,

[e] 265 -12'745' [9] 260 °°' [61 248 +19'608' [6] 240 + 1 3 '4 8 0 ' 140

[0] 230 + 2 2 '794' 1H-n.m.r. (90 MHz): 8.72 s (H-2), 8.49 s

(H-8), 6.13 d (Jx , 2 , 4.45 Hz, H - l ’), 5.74 dd (£2 . f3 . 8-27 H z '

H-2’), 5.42 dd (J4 , 5 , 8.58 Hz, H-4'), 5.12 dd (J 3 ,^4 .

2.54 Hz, H - 3 1) , 5.05 — 4.89 m (H-5') , 4.20 dd, 3.98 dd

(J5 , 6 , 3.50 Hz, J 5 , 6 „ 4.87 Hz, Jg , glI 12.55 Hz, H-6 ', H-6"),

2.38 q (SCH2CH3), 2.12 s, 2.09 s, 1.97 s, 1.87 s, 1.84 s 13 (acetate), 1.26 t (SCH^CHt ); C-n.m.r. (chloroform-d):

14.17 (SCH2CH3), 25.63 (SCH2CH3), 57.42 (C-l'), 70.40 (C-2'),

67.98 (C-3'), 67.42 (C-4'), 67.73 (C-5'), 61.67 (C-6 '),

131.47 (C-5), 143.42 (C-8), 151.33 (C-6), 151.89 (C-4),

152.08 (C-2).

Anal. Calc, for C ^ H ^ C I N ^ q S (588.1293). Found:

588.1309 (exact mass). The mass spectrum of 30b displayed the fragmentation pattern identical with that of 210a.

2,3,4,5,6-Penta-O-acetyl-1-(6-ethylthiopurin-9-yl)-1-

S-ethyl-l-thio-D-mannitol(31) (C-l' epimeric mixture)---

In a scale-up synthesis of above compound by Method A , the following unexpected result was observed. Compound 29

(5.48 g, 10.54 mmol) was added to a mixture of 6-chloropurine

(1.65 g, 10.68 mmol), mercuric cyanide (1.71 g, 6.79 mmol), cadmium carbonate (1.19 g, 6.78 mmol), calcium sulfate (2 g),

Celite (1.42 g ) , and nitromethane (120 m L ) . The mixture was O boiled for 22 h at 120 under reflux. Following the procedure described in Method A for the preparation of 30, a dark syrup was obtained. T.l.c. showed the presence of two major products.

Elution from a column (4.5 X 80 cm) of silica gel with 1:1 benzene— ethyl acetate partially separated the two components

Fractions containing these two products were pooled and

evaporated. The resulting, orange syrup was applied again to

a column (3.5 X 50 cm) of silica gel and eluted with same

solvent system. The faster-eluting portion was collected and

evaporated to afford a dark-orange syrup (1.04 g, 15%); 9 Q o R 0.45 (1:1 benzene— ethyl acetate); [aln -1.1 (o 0.7, — r U chloroform); x^?e0H 284 nm (e 10,750), 290 nm (sh) ; max yfilm 2943 (C-H), 1740 (C=0 of acetate), 1575, 1555 (purine), max 1360, 1205, 1040 cm ^ (C-O-C); ^H-n.m.r. (90 MHz): 8.56 s,

(H-2), 8.31 s, 8.26 s (H-8), 6.03 d, 5.86 d, ( , 2, 4.0 Hz,

, 2 i 2.5 Hz, H-l1, 1:1 epimeric mixture), 4.33— 3.77 m

(H-6 ', H-6"), 3.36 q (6-SCH2C H 3), 2.38 q (11-SCH2CH3), 2.11 s

2.09 s, 2.08 s, 2.05 s, 1.88 s (acetates), 1.45 t (6-SCH2CH3 )

1.27— 1.20 m (l'-SCH2CH ), 14.65, 14.70 (11-SCH2CH3 ) ,

23.27 (6-SCH2CH3), 25.42, 25.96 (1-SCH2CH3), 56.74, 58.95

(C-l1), 70.43, 70.96 (C-21), 68.05, 68.22 (C-3'), 67.42,

67.47 (C-4'), 67.95, 67.71 (C-5'), 61.50, 61.59 (C-6 '),

129.82, 130.72 (C-5), 141.75, 141.68 (C-8), 148.64, 148.93

(C-4), 151.99 (C-2), 161.76 (C-4); m/e 614 (M+ ), 179 (Base+ ).

Anal. Calc. C_cH_.N.CL + 1/7 C^Hc : C, 49.64; 25 24 4 10 2 6 b H, 5.58; N, 8.96; S, 10.24. Found: C, 49.60; H, 5.65; N, 8.61

S, 10.16.

The slower-eluting component was obtained (0.99 g,

15%) as a light-yellow syrup, which was found to be identical with 30 (3:1 epimeric mixture). The (1R) epimer of 30 142 crystallized readily from the syrup after seeding.

This reaction result was not reproducible. The same reaction condition was later used to give the desired product

30 in good yield.

Preparation of methyl 4, 6-0-benzylidene-a~D- glucopyranoside (5,) This compound was prepared by the 165 procedure of Richtmyer from 120 g of methyl a-D-glucopyranoside. The crude product was used without further purification for next step.

Preparation of methyl 4,6-O-p-tolylsulfonyl-a-D- glucopyranoside The compound was prepared from 40 g of 165 5 by the modified procedure of Danilov and Lishansky ; ° 165 ° yield 81 g (98%), m.p. 152-154 (lit. m.p. 154-155 ).

Preparation of D-altrose diethyl dithioacetal---

The foregoing compound was treated with sodium methoxide to yield methyl 2,3-anhydro-4,6-O-benzylidene-a-D-allopyranoside

(33 g, 92%), which was then converted into methyl 4,6-0- 16 5 benzylidene-a-D-altropyranoside as described by Richtmyer ; yield, 17.4 g (66%), m.p. 167-169 (lit. m.p. 169-170 ).

Shaken with ethanethiol and conc. hydrochloric acid by the 166 procedure of Coxon and Hough, the blocked altropyranoside

(13.2 g) was converted in one step into D-altrose diethyl O dithioacetal; yield 9.35 g (70%), m.p. 96-99 (lit. m.p.

98-100°).

2,3,4,5,6-Penta-O-acetyl-a-D-altrose diethy1- dithioacetal(34) Acetylation of D-altrose diethyl dithioacetal (9 g, 31.5 mmol), by the conventional pyridine—

acetic anhydride procedure gave the crystalline product; 170 yield 11.6 g (74%). This compound has been prepared before O for gas chromatography but not characterized; m.p. 9 0-9 2 ,

[a]J4 +29.2° (c 3.1, chloroform), 3000 (C-H), 1750 jj — HlaX (C=0 of acetate); 1450, 1430 (C-S), 1370, 1230, 1055, 1040 cm'

(C-O-C); X-ray powder diffraction data: 4.81 s (2), 4.40 w,

4.54 m, 5.12 w, 5.57 m, 5.79 w, 6.91 vw, 7.54 s' (3,3),

7.94 s (3,3), 8.60 m, 9.16 s (1).

Anal. Calc, for C20H32°10S2: C ' ^8.38; H' 6.50;

S, 12.90. Pound: C, 48.49; H, 6.67; S, 12.78.

2,3,4,5,6-Penta-O-acetyl-l-bromo-l-S-ethyl-l-thio-D-

altritol (35) This compound was obtained from 34 (2.0 g,

4.03 mmol) by the procedure used for preparing the mannose derivative 29. This compound was very unstable and was used

immediately in the subsequent reaction. The reaction was monitored by t.l.c. with the solvent system of 1:1 benzene—

ethyl acetate. The completeness of the reaction also found to vary with each trial.

(3JR) -2 , 3 , 4 , 5 , 6-Penta-O-acetyl-l- (6-chloropurin-9-yl) -

1-S-ethyl-l-thio-D-altritol(36a) Method A . The bromo syrup

35 was added to a mixture containing 6-chloropurine (0.62 g,

4.01 mmol), mercuric cyanide (0.64 g, 2.54 mmol), cadmium carbonate (0.45 g, 2.57 mmol), calcium sulfate (0.6 g), and nitromethane (120 mL). The solution was stirred for 4 h at O 65 . Again, the mixture was treated in the same manner as 144 described in Method A for the preparation of 17. The crude syrup was applied to a column (2.5 X 100 cm) of silica gel that was eluted with 1:1 benzene— ethyl acetate. The fractions containing the product were combined and evaporated to give 36 as a light-yellow syrup (1.0 g, 46%). The (1R) epimer (36a) was later crystallized from the syrup with seeding, as described in detail in Method B .

Method B . A mixture of the dithioacetal of 34 (5 g,

10.08 mmol), 6-chloropurine (1.56 g, 10.09 mmol), mercuric cyanide (1.58 g, 6.27 mmol), mercuric oxide (1.34 g, 6.21 mmol), calcium sulfate (1 g), and nitromethane (100 mL) was O boiled under reflux for 12 h at 120 . Following the procedure used in Method A , a crude syrup was obtained that was applied to a column (4.5 X 80 cm) of silica gel, which was eluted with

1:1 benzene— ethyl acetate. Fractions containing the desired product were pooled and evaporated to give 36 as a pale-yellow syrup (3.43 g, 58%). The syrup was taken up the minimal amount of abs. ethanol and the solution kept overnight at room tem perature to induce crystallization. The solution was then refrigerated. The resulting crystals were filtered off and dried to give 36a as a white solid; yield, 1.01 g (17%); m.p. 111.5-112.5 , [ct]^ +104.2 (c 1.0, chloroform); 0.36

(1:1 benzene— ethyl acetate); \ 265 nm (e 7,500); ITlclX vtjr* 2980 (C-H), 1750 (C=0 of acetate), 1580, 1540 (purine), UlcLX 1375, 1220, 1050 (C-O-C); c.d.: [9]310 0°, t ©3 2g4 “5 4 r800°,

[9] 276 °°' [0] 265 +20'940°/ I®3 257 °°' t0]25O "88'210°/ 145

[ 0] 238 °°' [0] 235 + 8 '020°' [e] 234 °°' [e]221 " 132'3 20'' 1H-n.m.r. (90 MHz, Me^O-dg): 8.91 s (H-2), 8.82 (H-8),

5.85 d (Jlt 2 , 8.51 Hz, H-l1), 6.03 dd (J2 , 3 , 2.05 Hz, H-2’),

5.10 dd (J3 , 4 , 7.34 Hz), 4.94 — 4.87 m (H-3',5'), 4.18 — 3.85 m

(H-6 ',6"), 2.44 q (SCH2CH3), 2.07 s, 2.02 s, 1.99 s, 1.85 s,

1.83 s (acetates), 1.00 t (SCH2CH3) ; ^C-n.m.r. (chloroform-d):

14.00 (SCH0CH_), 26.05 (SCH„CH_), 59.97 (C-l’), 61.40 (C-6'), /— j — z * J 68.88, 69.05, 69.56, 70.51 (C-2•,3•,4',51), 131.59 (C-5),

143.75 (C-8), 151.36 (C-6), 151.53 (C-4), 151.94 (C-2); X-ray

powder diffraction data: 3.69 w, 3.90 s (3), 4.10 vw, 4.33 m,

4.65 s (2), 5.02 w, 5.51 vw, 5.79 w, 6.22 m, 7.03 w, 7.98 m,

9.22 s (1) .

Anal. Calc, for C ^ H ^ C I N ^ O ^ S : C, 46.93; H, 4.97;

Cl, 5.92; N, 9.52; S, 5.44. Found: c, 47.03; H, 5.06;

Cl, 7.00; N, 9.31; S, 5.34.

With seeding, the (1R) epimer could also be isolated,

prior to chromatographic purification, from the crude syrup

obtained by either method. The major side product of this

reaction (< 10%) was aldehydo-D-altrose pentaacetate, formed

by the demercaptalation of the starting compound 3,4.

Preparation of methyl 4,6-0-benzylidene-2-0-p-

tolylsulfonyl-g-D-glucopyranoside Prepared from 5. in 4 7% 167 yield, by the procedure of Wiggins, this compound had ° 167 0 m.p. 153-154 (lit. m.p. 153-154 ).

Methyl 4,6-0-benzylidene-3-deoxy-a-D-arabino-

hexopyranoside was treated with sodium methoxide to yield 146 methyl 2,3-anhydro-4,6-0-benzylidene-a-D-mannopyranoside (79%); ° 167 ° m.p. 145-146 (lit. 145-147 ). This compound was then converted into methyl 4,6-0-benzylidene-3-deoxy-a~D- ara'bino-hexopyranoside by lithium aluminum hydride reduction, 168 following the literature procedure of Rembarz; yield,

° 1 f\ R ° 15.5 g (78%), m.p. 109-111 (lit. m.p. 111-112 ).

Preparation of methyl 3-deoxy-g-D-arabino- hexopyranoside (9) New Method The following procedure gave superior yields and eased the problem of crystallization as 168 compared with the literature procedure. Methyl 4,6-0- benzylidene-3-deoxy-a-D-arabino-hexopyranoside (15.5 g,

58.27 mmol) was dissolved in 30% aqueous acetic acid and O boiled for 3 h at 100 under reflux. The solution was cooled to room temperature and washed with, chloroform (75 mL, 2x) .

The aqueous portion was freeze-dried and dissolved in the minimal amount of methanol to induce crystallization.

Refrigeration overnight afforded a mass of white crystals; yield, 9.0 g (84%), m.p. 140-141° (lit.168 m.p. 141-142°).

Preparation of 3-deoxy-D-arabino-hexose diethyl dithioacetal This compound was prepared from the foregoing compound (9.04 g, 50.8 mmol) by shaking with ethanethiol

(10 mL) and conc. hydrochloric acid (10 mL). The solution was diluted with MeOH and made neutral with lead carbonate, the mixture filtered, and the filtrate washed with methanol.

The filtrate was evaporated to a syrup. 3.5 g of the syrup was applied to a column (2.5 X 100 cm) of silica gel. 147

Elution with 4:1 chloroform— methanol gave the desired product, which immediately followed a yellow band. Both the pure syrup and the remaining crude syrup solidified into O fine needles upon standing; yield, 11.7 g (85%), m.p. 70-71 1 c O O A ° (lit. reported as syrup); [a]p -35.0 (c 1.1, methanol)

1 ft R 1 o o [lit. fot]D -38.0 (c 2.1, methanol)].

2.4.5.6-Tetra-0-acetyl-3-deoxy-D-arabino-hexose diethyl dithioacetal(32) This syrupy compound was prepared by the conventional acetic anhydride— pyridine procedure in

80% yield; [a]p4 -2.75 (c 1.8, chloroform), [lit.168

[a]^1 +0 (c 1.7, chloroform)].

2.4.5.6-Tetra-0-acetyl-l-bromo-3-deoxy-l-S-ethyl-

1-thio-g-arabino-hexitol(37) This compound was obtained as an unstable syrup essentially by the procedure used for preparing the mannose derivative 2J. This compound was used immediately in the subsequent reaction.

2.4.5.6-Tetra-O-acetyl-l-(6-chloropurin-9-yl)-3- deoxy-l-S-ethyl-l-thio-D-arabino-hexitol(38) (C-l epimeric mixture) Method A . The bromo compound 37 (0.62 g, 0.57 mmol), was added to a mixture of 6-chloropurine (0.062 g, 0.40 mmol), mercuric cyanide (0.064 g, 0.25 mmol), cadmium carbonate

(0.051 g, 0.29 mmol), calcium sulfate (0.1 g), and nitro- O methane (15 mL). The mixture was heated for 4 h at 70 with vigorous stirring. The mixture was filtered and processed in a manner analogous to the procedure used for 17. The resulting, crude syrup was applied to a preparative t.l.c. plate 148

(200 X 200 X 2.5 nun) of silica gel, which was developed twice with 1:1 benzene— ethyl acetate. The major zone was excised and extracted with chloroform to give, upon evaporation of solvent, compound 38 as a homogeneous, syrup 4:1 mixture of

1-epimers; yield, 0.103 g (49%); +26.1 (c 0.64, chloro- MeOH form); 0.32 (1:1 benzene— ethyl acetate); A. 266 nm — F J max (e 1,380); vfllm 2980, 2918 (C-H), 1750 (C=0 of acetate), ITlclX -1 1 1580, 1540 (purine), 1374, 1220, 1050 cm (C-O-C); H-n.m.r.

8.69 s (H-2), 8.49 s (H-8), 5.96 d, 5.85 d ( , 2, 5.0 Hz,

H-l', 4:1 epimeric mixture), 5.76— 4.93 m (H-21,4',51),

4.46 — 3.86 m (H-6 ',6"), 2.51 q (SCH2CH3), 2.05 — 1.97 (C-3 ' ) ,

1.17 t (SCH2CH3); ^C-n.m.r. (chlorof orm-d) ; 14.19 (SC H ^ H ^ ) ,

25.26, 26.31 (SCH2CH3 ), 30.69, 32.50 (C-3'), 61.12, 61.42

(C-l'), 61.78 (C-6 '), 67.66 (C-5'), 69.66, 70.15 (C-2' or 4'),

71.67 (C-4' or 2 ’), 132.05 (C-5), 144.00, 144.72 (C-8),

151.79 (C-6), 152.11 (C-2,4).

Anal. Calc, for C^H^ClN^OgS: C, 47.34; H, 5.13;

Cl, 6.60; N, 10.56; S, 6.04. Found: C, 47.34; H, 5.45;

Cl, 6.81; N, 10.29; S, 6.08.

Method B . A mixture of 32 (5.0 g, 11.42 mmol),

6-chloropurine (1.56 g, 10.09 mmol), mercuric cyanide (1.58 g,

6.27 mmol), mercuric oxide (1.33 g, 6.16 mmol), and nitro methane (120 mL) was boiled for 12 h under reflux with O vigorous stirring at 110 . The hot mixture was filtered and treated as in Method A . The crude syrup was applied to a column (3.5 X 80 cm) of silica gel and eluted with 1:1 149 benzene— ethyl acetate. The fractions containing the product was pooled and evaporated to afford an orange syrup; yield,

2.90 g (54%). ^H-N.m.r. spectroscopy showed that the syrup was an epimeric mixture in 4.6:1 ratio and was identical with that prepared by Method A .

Attempts to separate the epimers by using the same

HPLC condition for the separation of epimeric pair 30 were unsuccessful. The epimers were only slightly resolved after several recycles.

(JLS) -2,4,5, 6-Tetra-O-acetyl-l- (6-chloropurin-9-yl) -3- deoxy-l-S-ethyl-l-thio-D-arabino-hexitol (38b)--- 54b, (15 mg,

0.041 mmol) was dissolved in 0.2 mL dry pyridine and cooled in acetone-dry ice bath. Acetic anhydride (0.2 mL) was then O added. The solution was kept at 4 overnight. Then water was added dropwise while the solution was kept in acetone— dry ice bath. The solvents were removed by high vaccum pump.

The resulting syrup was chromatographed on silica gel plate and developed with 1:1 benzene— ethyl acetate. The major band was extracted with chloroform. Evaporation of the solvent gave light yellow syrup, yield, 19 mg (87%).

The product possesses the same physical properties

20 ° as mixture except the following: [a ]D -74.3 (c 0.9, chloroform); c.d.: [ 0 3 3^0 ® ' ^ 285 +3,356 > ^ 278 0 ’

[6] 264 "15,160°, [0] 254 0°, [0] 246 +10,372°, [0] 235 +7,447°,

[0]225 +26,596°; 1H-n.m.r. (90 MHz): 8.76 s, 8.50 s (H-2,8),

5.89 d (Jlt 2 , 4.28 Hz, H - l ’), 5.17--4.87 m (H-2',5'), 150

5.57 o (H-4 1) , 1.74 — 1.43 m (H-3' ) , 4.19 dd, 4.01 dd (J5 , gI

3.81 Hz, J 5 , gI 7.20 Hz, J g , 6 „ 11.64 Hz, H-6,6"), 2.39 q

(SCH CH ), 2.04 s, 2.02 s, 1.99 s, s.93 s (acetates), 1.17 t

(SCH2OS3 ); l3C-n.m.r.: 14.19 (SCH2CH3), 25.65 (SCH2CH3 ),

61.12 (C-l'), 69.66, 71.67 (C-21,4'), 32.50 (C-3’), 67.66

(C-5'), 61.78 (C-6 '), 131.72 (C-5), 144.72 (C-8), 151.79

(C-6), 152.12 (C-2 and C-4).

U R ) -1-(6-Chloropurin-9-yl)-1-S-ethyl-l-thio-D- altritol(49a) Compound 36a (400 mg, 0.68 mmol) was dissolved O in 20 mL of methanol and the solution cooled to 0 in an ice- bath. Gaseous ammonia was bubbled through the solution for

3 0 min, and then the flask stoppered and refrigerated over night. The solvent was then evaporated off and the residual syrup chromatographed on two silica gel plates by using 4:1 chloroform— methanol as the eluant. The major band was isolated 23 0 as colorless syrup; yield, 0.25 g (97%); [a]D +88.9 (c 1.07, water); R_ 0.47 (4:1 chloroform— methanol); A^2^ pH 7, Jb Iucix 263 nm (e 5,900); pH 1, 262.5 nm (e 5,900); pH 12, 263 nm

(e 5,900); vfllm 3300 (OH), 1590, 1560 (purine ring), 1400, ITlclX 1340, 1180 cm-1 (C-O-H); 13C-n.m.r. (Me2SO-dg): 14,55

(SCH2CH3), 24.94 (SCH2CH3 ), 62.37 (C-l'), 63.00 (C-6 '),

71.11 (C-2'), 71.54 (C-3'), 71.98 (C-4*), 73.34 (C-5'),

130.92 (C-5), 147.04 (C-8), 149.99 (C-6), and 151.94 (C-2 and

C-4) .

Anal. Calc. C^H^gClN^Oj-S: c, 41.26; H, 5.07; Cl, 9.25;

N, 14.82; S, 8.46. Found: C, 41.41; H, 5.35; Cl, 9.13; N,

15.16; S, 8.53. 151

(_1R) -1- (6-Chloropurin-9-yl) -1-S-ethyl-l-thio-D- mannitol(52a) Gaseous ammonia was bubbled for 30 min through solution of compound 30a (325 mg, 0.55 mmol) in methanol o o (20 mL) at 0 . The solution was kept overnight at 4 and then evaporated to dryness. The residual syrup was chromatographed on silica gel plate by using 4:1 chloroform— methanol as the developer. The major zone was extracted with hot ethanol and evaporated to afford an amorphous glass (180 mg, 86%); 23 ° [al +87 (c 0.4, water); R_ 0.54 (4:1 chloroform— methanol); D — XH2° pH 7, 263 nm (e 5,200); pH 1, 263.5 nm (e 5,200); max ^ r KRy vpH 12, / 263 nm (e 5,200);t v max 3340 (OH), 2900 (CH), 1595,

1560 (purine ring), 1390, 1340, 1190, 1075 cm-1 (C-O-H);

13C-n.m.r. (Me„SO-dr): 14.55 (SCH„CH_), 24.83 (SCH0CH_), 2 — 6 2-- 3 -- ^ 3 63.13 (C-l1), 72.95 (C-2’), 70.18 (C-3’), 69.28 (C-4'),

71.32 (C-5'), 63.91 (C-6 '), 130.73 (C-5), 146.89 (C-8),

149.33 (C-6), 151.60 (C-4), and 151.84 (C-2).

Anal. Calc, for C 13H igClN40 5S: C, 41.26; H, 5.07;

Cl, 9.25; N, 14.82; S, 8.46. Found: C, 41.37; H, 5.05;

Cl, 9.28; N, 14.58; S, 8.38.

(IS)- 1-(6-Chloropurin-9-y1)-3-deoxy-l-S-ethy1-1-thio-

D-arabino-hexitol(54b) A sample of 3J (epimeric mixture,

400 mg) was saponified by the same procedure described for

49a and the resultant solution was evaporated to give a thin syrup. The syrup was chromatographed on preparative t.l.c. plate of silica gel by using 4:1 chloroform— methanol as the developer. The major zone was extracted with ethanol. 152

Evaporation of the solvent afforded 200 mg of a light-yellow syrup (74%). The syrup was crystallized from methanol to yield

40 mg (15%) of 54b,. After several recrystallization from O methanol, a white solid was obtained; m.p. 138-139 ; 29 ° [a] -95 (c 0.3, methanol); 0.61 (4:1 chloroform— u — — r methanol); \Me0H 265 nm (e 18,100); vKBr 3310 (OH), 2910 (CH), max ' max 1580, 1500 (purine ring), 1390, 1340, 1190, 1075 cm ^

(C-O-H); 13C-n.m.r. (Me„SO-dc): 14.38 (SCH0CH_), 24.69 2 — 6 2--3 (SCH2CH3 ), 63.05 (C-6 ’), 64.40 (C-l'), 67.67, 68.49 (C-2',4');

C-3' signal overlapped with that of the solvent, 74.81 (C-51),

130.20 (C-5), 146.39 (C-8), 148.76 (C-6), 151.26 (C-2), and

151.86 (C-4).

Half of the already small quantity of 5^b was re- acetylated to afford the (IS) epimer of the starting compound

(38b). Not enough sample was left for microanalysis. PART TWO:

CYCLIZATION OF ACYCLIC-SUGAR

NUCLEOSIDES

153 I. HISTORICAL BACKGROUND AND INTRODUCTION

A. Cyclization of Acyclic-Sugar Nucleosides

The quest for a method to convert the carbohydrate moiety of acyclic-sugar nucleosides into cyclized (furanosyl or pyranosyl) forms without loss of the nucleosides base has long been part of the acyclic-sugar nucleoside study in our laboratory. Attempts to cyclize by such chemical methods as 1 2 2 catalytic hydrogenation, demercaptalation and S-methylation were unsuccessful.

(1) M. L. Wolfrom, W. von Bebenburg, R. Pagnucco, and P. McWain, J. Org. Chem., 30, 2732-2735 (1965).

(2) Unpublished result.

It is suspected that many synthetic analogs of nucleosides are inactivated in vivo by the action of glycosidases. Insertion of a carbon atom between the sugar 3 4 and the base has been proposed ' as one solution to this problem of enzymic inactivation. This type of compound, having a carbon atom inserted into the base-sugar linkage, is termed a "homonucleoside". Synthesis of homonucleosides

154 155 may be approached by initial cyclization of a suitable

sugar derivative, followed by coupling to the base. Defaye 3-7 and coworkers have reported the synthesis of a series of

such compounds by this approach. Another approach is by

(3) J. Defaye, M. Naumberg, and T. Reyners, J. Heterocycl. Chem. , 229-234 (1969).

(4) J. Defaye and T. Reyners, Bull. Soc. Chem. Biol., 50, 1625-1635 (1968).

(5) G. Giovannietti, L. Nobile, M. Amorosa, and J. Defaye, Carbohydr. Res., 2_1, 320-325 (1972).

(6) J. Defaye and Z. Machon, Carbohydr. Res., 24, 235-245 (1972).

(7) V. Zecchi, I. Garuti, G. Giovannietti, L. Rodriguez, M. Amorosa, and J. Defaye, Bull. Soc. Chim. Fr., 1389-1394 (1974). cyclization of a pre-formed acyclic-sugar nucleoside. The existence of marked differences in behavior between dialkyl dithioacetals of pentoses on being treated with one molar equivalent of p-toluenesulfonyl chloride in pyridine is well recognized. Dialkyl dithioacetals of D-arabinose give g 5-p-toluenesulfonates under these conditions, whereas the 9 10 dixsobutyl dithioacetals of g-lyxose, g-ribose, and g-xylose"*"^ give 2,5-anhydro-pentose diisobutyl dithioacetals in good yield under the same conditions. A rationale was developed"^ to explain this difference by assuming that the rate of conversion into 2,5-anhydride is controlled .by the difference in energy between the ground state and the tran sition state for forming the anhydro ring. In the transition 156 state, the acyclic-sugar chain adopts a conformation in

(8) H. Zinner, H. Brandhoff, and H. Kristen, Chem. Ber., .92, 1618-1623 (1959).

(9) J. Defaye, Bull. Soc. Chim. Fr. , 2686-2689 (1964) .

(10) H. Zinner, H. Brandhoff, H. Schmandke, H. Kristen, and R. Haun, Chem. Ber., 9_2, 3151-3155 (1959).

(11) J. Defaye and D. Horton, Carbohydr. Res., 14, 128-132 (1970). which 0-2 can effect an internal SN2 type attack on C-5 to form the oxolane ring. In case of arabino derivatives, it was noticed that extremely unfavorable steric interactions would arise among the substituents on C-2, C-3, and C-4, because all of them would be on the same side of the ring.

Based on this stereochemical rationale for ring formation, the acyclic-sugar nucleoside 1-S-ethyl-l-thio-(uracil-l-yl)- 12 D-xylitol was converted into its 5-0-£-toluenesulfonyl derivative, which was rapidly cyclized, as expected, to

(12) J. Defaye, D. Horton, S. S. Kokrady, and Z. Machon, Carbohydr. Res., 4_3, 265-280 (1975). give 3,4-di-0-acetyl-2,5-anhydro-l-S-ethyl-l-thio-l-

(uracil-l-yl)-D-xylitol upon acetylation. This is the first example of successful cyclization of an acyclic-sugar nucleoside. However, the usefulness of this method is severely limited by the fact that only sugar chains that are stereo- chemically able to assume a transition state with no severe 157 steric destabilization by cis substituents will form their

2,5-anhydrides upon treatment with £-toluenesulfonyl chloride.

Another type of nucleoside that is structurally close related to this study is the "reversed" nucleoside. 13 This is a term originally proposed for the ribose derivative

(13) N. J. Leonard and K. L. Carraway, J. Hetero- cycl. Chem., 3, 485-489 (1966). having a heterocyclic base bonded at C-5 of the sugar moiety instead of at the anomeric C-l. Later, more "reversed" 14 15 nucleosides were synthesized ' as part of a program to

(14) J. Hildesheim, J. Cleophax, S. D. Gero, and R. D. Guthrie, Tetrahedron Lett., A9_, 5013-5016 (1967).

(15) N. J. Leonard, F. C. Sciavolino, and V. Nair, J. Org. Chem., _33, 3169-3174 (1968). determine the cytokinin activity and chemical properties of compounds closely related to kinetin, a cell-division factor. 16 17 Other "reversed" nucleosides were synthesized ' as useful

(16) M. Kawazu, T. Kanno, S. Yamamura, T. Mizoguchi, and S. Saito, J. Org. Chem., 2887-2890 (1973).

(17) N. Takamura, N. Tag, T. Kanno, and M. Kawazu, J. Org. Chem., 3JB, 2891-2895 (1973). synthetic intermediates in an improved, scaled-up synthesis of the hypercholesteremic agent, eritadenine. Still others 158 18 were synthesized in search of suitable functional monomers.

Nucleoside-containing polymers have been synthesized by

copolymerization of acrylamide with monomers that are acryloyl 19 derivative of nucleosides, and studied for their possible

applications in the separation of mixtures of nucleic acid bases and as antiviral, antineoplastic agents, as well as 21 interferon inducers.

(18) N. Ueda, Y. Nakatani, S. Terada, K. Konda. and K. Takemoto, Techno. Report, Osaka Univ., ^3, 713-714 (1973).

(19) M. Imoto and K. Takemoto, Synthesis, 1, 173- 179 (1970).

(20) Von G. Greber and H. Schott, Angew. Chem., 82, 82 (1970).

(21) K. Takemoto, J. Polym. Sci., Polym. Symp., 25, 105-125 (1976).

There are also "reversed" nucleosides obtained either 22 for the purpose of preparing new nucleoside and analogs or as a result of intramolecular alkylation reaction from 23 normal nucleosides. Some "reversed" nucleosides are depicted in Fig. 1.

(22) S. Fakatsu, Y. Takeda, and S. Umezawa, Bull. Chem. Soc. Jap., 46, 3165-3168 (1973).

(23) K. Kobayashi and W. Pfleiderer, Chem. Ber., 109, 3175-3183 (1976). 159

OH NH

CH CH

H,OH

OH OH OH 5 (6-aminopurin-9-yl) 5'-deoxy-5-(uracil-l-yl) -5-deoxy-D-arabinofuranoside -D-ribofuranoside

o c h 2c h 3

c h 2

OCH OCH- OH Methyl 5 (6-aminopurin-9 Methyl 6 '-deoxy-6 (1,2- y l )-21,51-dideoxy-a-D- dihydro-2-keto-4-ethoxy- ribofuranoside “ pyrimidin-l-yl)-a-D- glucopyranoside -

Figure 1: "Reversed" nucleosides 160 B. Fermentation by Acetobacter suboxydans

Man had made vinegar from ethanol-containing

solutions by natural fermentation for many many centuries.

It was gradually realized that vinegar is formed by living 24 organisms— "Mother of vinegar". Persoon was the first to

study this living organism and named it Mycoderma, but it 25 was Pasteur who carried out systematic studies on acetic

acid fermentation by mother of vinegar. Later, Knieriem and 2 6 Mayer recognized Mycoderma as a bacterium. The now univer

sally accepted generic name Acetobacter was coined in 1900 27 by Beijennck. Since that time, many attempts to classify

(24) C. H. Persoon, Mycologia europaea, 1, 960 (1822).

(25) L. Pasteur, "Etudes sur la vinaergre", Paris 1868.

(2 6) W. von Knieriem and A. Mayer, Landw. Versuchsstation, _16, 305 (1973).

(27) M. W. Beijerinck, Proc. Acad. v. Wetenshapp., Amsterdam. 2, 495 (1900).

acetic acid bacteria have been made and more than 50 different 2 8 "species" names were give. In 1950, Frateur greatly reduced

the number by proposing a new system of classification based

solely on biochemical properties of species. In this system,

Acetobacter was divided into four groups: peroxydans, oxydans, mesoxydans, and suboxydans, as shown in Table 1.

A strain of Acetobacter first isolated by Kluyver 29 and de Leeuw from beer falls into the suboxydans category Table 1. Grouping of Acetobacter by Frateur.

2nd 3rd criterion: 1st 4th criterion: 5th criterion: criterion: oxidation criterion: ketogenic gluconic acid Group oxidation of lactic acid to catalase activities production acetic acid C 0 2+ H 20

peroxydans - + + -- oxydans + + + - or + mesoxydans + + or ± + + -> ++ + -*-++• suboxydans + - - ++ ++ 161 162 and is accordingly named Acetobacter suboxydans Kluyver and de Leeuw. A less commonly used name, according to DeLey's 30 classification, is Gluconobacter oxydans.

(28) J. Frateur, Cellule Res. Cytol. Histol., 53, 285-396 (1950).

(29) A. J. Kluyver and F. J. G. Leeuw, Tijdschr. Vergelijk. Geneesk., 1JD, 170-182 (1924).

(30) J. DeLey, J. Gen. Microbiol., Zl, 352-365 (1959).

Many acetic acid bacteria are noted for their unsurpassed ability to oxidize a great variety of carbohydrate and derivatives. Most of these oxidations occur in one or two discrete steps, resulting in accumulation of end products, often with a near-quantitative yield. Therefore, these bacteria are of considerable academic interest and of practical importance in both preparative organic chemistry and industry.

Acetobacter suboxydans is of particular interest because its limited oxidizing power (See Table 1) and its stereospecific oxidation of polyols. The oxidization of polyols having particular steric configurations was first observed in growing 31-33 cultures of Acetobacter xylinum.

(31) G. Bertrand, C. R. Acad. Sci. Paris, 126, 846, 984 (1898).

(32) G. Bertrand, Ann. Chim. Phys., [8] 3_, 181, (1904). (33) G. Bertrand, Compt. Rend., 149, 225-227 (1909). 163 The Bertrand rule was enunciated from these studies; it states that the favorable configuration for oxidation by

A. xylinum is the cis arrangement of two secondary hydroxyl groups adjacent to a primary alcohol grouping. Extending this 3 4 . rule to A. suboxydans, Hann, Tilden, and Hudson found this organism more selective and indicated that the configuration

(34) R. M. Hann, R. B. Tilden, and C. S. Hudson, J. Am. Chem. Soc., 60, 1201-1203 (1938). that is most readily oxidized by this organism might be OH OH written -C— C— CH„OH. H H The oxidation of ribitol best exemplifies the selec tivity of this organism. Ribitol is oxidized by A. suboxydans . . 35 specifically to L-erythro-pentulose, as would be expected from the well-known Bertrand— Hudson rule. The nature of the oxidation product was established by the following chemical conversions:

1. When the product is refluxed with pyridine for 2 h,

it is isomerized into ^-arabinose.

2. When treated with phenylhydrazine, the product gives an

osazone that is identical with that derived from L-

arabinose.

3. The product reacts with acetone to form 1,2:3,4-di-O-

isopropylidene-L-erythro-pentofuranoside. This asymmetric

transformation turns an optically-inactive compound into

an optically-active one by introducing chirality at C-3. 164

Similarly, allitol is transformed into L-psicose

(L-ribo-hexulose) by the action of this organism.

(35) T. Reichstein, Helv. Chim. Acta, 17_, 996-1002 (1934).

(36) M. Steiger and T. Reichstein, Helv. Chim. Acta, 18, 790-799 (1935).

The oxidation of other carbohydrates by suspension of A. suboxydans has been the subject of a number of inves- 37 51 51 tigations. The polyols studied included octitols, such heptitols as perseitol"^' (D-glycero-D-galacto- 39 40 41 heptitol), Q-erythro-g-gulo-heptitol, volemitol ' 47 (p-glycero-D-manno-heptitol), g-sedoheptitol (D-glycero- 48 g-gluco-heptitol), and D-glycero-D-ido-heptitol, hexitols^^'^ and such w-deoxy sugar alcohols as 34 45 46 L-fucitol ' (1-deoxy-g-galactitol), 5-deoxy-L-lyxitol, 44 46 1-deoxy-D-mannitol, 1-deoxy-D-glucitol and 1,6-dideoxy- 45 galactitol.

(37) K. R. Butlin, Biochem. J., 30, 1870-1877 (1936).

(38) E. B. Tilden, J. Bacteriol., 37, 629-637 (1939).

(39) W. D. Maclay, R. M. Hann, and C. S. Hudson, J. Am. Chem. Soc., 64, 1606-1609 (1942).

(40) L. C. Steward, N. K. Richtmyer, and C. S. Hudson, J. Am. Chem. Soc., 7JL, 3532-3534 (1949).

(41) V. Ettel and J. Liebster, Collect. Czech. Chem. Comm., 14, 80-90 (1949).

(42) A. J. Kluyver and A. G. J. Boezaardt, Rec. Trav. Chim., 57, 609-615 (1938). (43) E. L. Totton and H. A. Lardy, J. Am. Chem. Soc., 71, 3076-3078 (1949).

(44) L. Anderson and H. A. Lardy, J. Am. Chem. Soc., 70, 594-597 (1948).

(45) N. K. Richtmyer, L. C. Steward, and C. S. Hudson, J. Am. Chem. Soc., 7_2' 4934-4937 (1950).

(46) G. N. Bollenback and L. A. Underkofler, J. Am. Chem. Soc., 12_, 741-745 (1950).

(4 7) L. C. Steward, N. K. Richtmyer, and C. S. Hudson, J. Am. Chem. Soc., 7_4, 2206-2210 (1952).

(48) J. W. Pratt, N. K. Richtmyer, and C. S. Hudson, J. Am. Chem. Soc., 7_4' 2210-2214 (1952).

(49) T. E. King and V. H. Cheldelin, J. Biol. Chem., 198, 127-133, 135-141 (1952).

(50) T. E. King and V. H. Cheldelin, J. Bacteriol., 6 6 , 581-584 (1953).

(51) N. K. Richtmyer, Carbohydr. Res., 2_3, 319-322 (1972) .

From the results of those studies on w-deoxy sugar alcohols, the Bertrand— Hudson rule was further extended to predict the action of this organism on this type of compound by considering the CH^CHOH group as simply an elongated C^OH group. Therefore, only the w-deoxy sugar alcohols having OH OH H OH OH H the configuration -C— C— C— CH_ or -C— C— C— CH_ are * OH * H oxidizable by A. suboxydans and at the.position indicated by an asterisk.

The study of this organism has been further extended to functionally more complicated carbohydrate derivatives. 52-54 These results show that the presence of acetamido,

S-ethyl,^5 dimethyl acetal, diethyl dithioacetal,and 57 branching hydroxymethyl substituents does not affect the 166

specificity of the organism, provided that the substituents

are located at some distance from the site of oxidation.

(52) J. K. N. Jones, M. B. Perry, and J. C. Turner, Can. J. Chem., 3_9, 965-971 (1961).

(53) J. K. N. Jones, M. B. Perry, and J. C. Turner, Can. J. Chem., 3_9, 2400-2410 (1961).

(54) J. K. N. Jones, M. B. Perry, and J. C. Turner, . Can. J. Chem., 40_, 503-510 (1962).

(55) L. Hough, J. K. N. Jones, and D. L. Mitchell, Can. J. Chem. ,' 3_7, 725-735 (1959).

(56) R. T. Williams, J. K. N. Jones, N. J. Dennis, R. J. Ferrier, and W. G. Overend, Can. J. Chem., 4_3, 955- 959 (1965).

(57) W. A. Szarek, G. W. Schnarr, H. C. Jarrell, and J. K. N. Jones, Carbohydr. Res., 5_3, 101-108 (1977).

A series of aliphatic glycols ranging from ethylene glycol to 1 ,7-heptane-diol has also been found to be oxidized by this microorganism to the corresponding acids, which is the 5 8 subject of a review. It is noteworthy that some of these oxidation (e.g. that of ethylene glycol and 2 ,3-butanediol) follows the Bertrand— Hudson rule and has been described as 59 demonstrations of the rule; however, enzymic studies showed

(58) J. DeLey and K. Kersters, Bacteriol. Rev., 28, 164-180 (1964).

(59) K. Kersters and J. DeLey, Biochem. Biophys. Acta, 71, 311-331 (1963). that different enzymes are involved in these oxidation and have no connection with the rule. 167 Regardless of the enzymes involved, Fulmer and

Underkofler^ classified polyols into the following four groups according to the way in which they are oxidized by

A. suboxydans:

1. oxidized at high concentrations (25% or above)— e.g.

D-glucitol, D-mannitol.

1 . oxidized at relatively low concentrations— e.g. glycerol

and erythritol.

3. require additional assimilable substrate for subculture

— e.g. myo-inositol, meso-2 ,3-butanediol.

4. not oxidized at all— e.g. galactitol, L-rhammitol.

(60) E. I. Fulmer and L. A. Underkofler, Iowa State Coll. J. .Sci., 21, 251-270 (1947).

The Bertrand— Hudson rule works well in most cases thus-far studied, but, like many other rules, there are a few exceptions. D-Galactose diethyl dithioacetal and 61 D-galactose dimethyl acetal both gave ketose products when oxidized by A. suboxydans. According to the rule, the galacto configuration does not satisfy the requirements for oxidation and therefore can not be oxidized or, if -CH(SR)2 and

-CH(OR)2 group is viewed as an elongated CH^ group, the oxidation is anticipated to occur at C-3 to give hex-3-ulose derivatives. However, the products were identified as

L-arabino-hex-5-ulose 1,1-diethyl dithioacetal and L-arabino- hex-5-ulose 1,1-dimethyl acetal, respectively. It was 168 suggested that a new enzyme system was present in

A. suboxydans that was not reported earlier. The proposed dehydrogenases appear to be highly specific for the D-qalacto configuration.

(61) D. T. Williams and J. K. N. Jones, Can. J. Chem., 45, 741-744 (1967).

As there are many enzymes involved in the oxidation process, it is not surprising that in many cases, more than 61 one oxidation product are obtained. In case of L-fucitol, two ketose products were obtained. One was the expected l-deoxy-D-threo-D-qlycero-3-hexulose and the second minor product was 6-deoxy-L-lyxo-hexulose, the product expected from the proposed "D-galacto" dehydrogenase. Another inter esting example is the oxidation of (±)-threitol. Whistler and 6 2 Underkofler reported that the yield of oxidation product

L-glycero-tetrulose (L-erythrulose) was practically quanti tative in seven days when the concentration of substrate did not exceed 4.5%. When a 5% solution was oxidized by 6 3 A. suboxydans, three products were obtained. The main product was the expected L-erythro-tetrulose and the two minor products were D-altro-heptulose and L-threitol. It was surmised that heptulose arose from the action of transaldolase, preceded by a sequence of isomerization and phosphorylation, whereas the L-threitol was derived from the stereospecific reduction of L-erythro-tetrulose. 169

(62) R. L. Whistler and L. A. Underkofler, J. Am. Chem. Soc., 6j), 2507-2508 (1938).

(63) C. L. Hu, E. A. McComb, and V. V. Rendig, Arch. Biochem. Biophys., 110, 350-353 (1965).

Other substrates that are also subjected to the oxidation by A. suboxydans include cyclic polyols, aldoses, ketoses, and aldonic acid. The detail of these oxidations 64 has been reviewed.

(64) T. Asai, "Acetic Acid Bacteria", University of Tokyo Press, 1968, Part II.

The enzyme systems involved in the oxidation of polyols have been studied rather extensively. Edson first 6 5 suggested that highly selective oxidation of polyols by

A. suboxydans could be explained by the presence of an enzyme having substrate specificity defined by Bertrand—

Hudson rule, and he later examined^ cell-free extracts of

A. suboxydans. Two polyol dehydrogenase activities having different substrate specificities and pH optima were found.

One of them has the pH optimum of about 5.5 and is most active within pH range of 5— 6 , which is also optimal for 6 7 the growth of the organism. This constitutive enzyme is associated with cytochrome-containing particles, which were 6 7 probably derived from the cytoplasmic membrane. The enzyme catalyzes the oxidation of polyols (with the exception of 170

Q L-threitol ) possessing the required Bertrand— Hudson configuration. The methods used in these studies failed to release the enzyme from the particles and therefore, it is not yet known whether oxidation of the polyol by the particle is due to one or to several enzymes. Also not known is the 6 6 cofactor requirement of the enzyme. It was proposed to name this enzyme "Bertrand— Hudson enzyme" or "cytochrome- linked g-mannitol dehydrogenase" and was called "acid enzyme" 6 8 or "particulate dehydrogenases". This is the enzyme that accounts for the unique oxidative behavior of whole cells of acetic acid bacteria. The presence and specificity of this particulate enzyme system has been confirmed by many other , 67-70 workers.

(65) N. L. Edson, Rept. Australian New Zealand Assoc. Advance Sci., 29[, 281-298 (1953).

(6 6 ) A. C. Arcus and N. L. Edson, Biochem. J., 64, 385-394 (1956).

(67) J. DeLey and R. Dochy, Biochem. Biophys. Acta, 40, 277-289 (1960).

(6 8 ) K. Kersters, W. A. Wood, and J. DeLey, J. Biol. Chem., 240, 965-974 (1964).

(69) J. A. Fewster, Biochem. J. , 6_5, 14p (1957).

(70) C. Widmer, T. E. King, and V. H. Cheldelin, J. Bacteriol., Tl_, 737 (1956).

6 6 The other enzyme systems found by Arcus and Edson in the cell-free extracts of the organism are soluble, NAD- linked dehydrogenases. As there is no configuration common 171

to all the polyols oxidized by this enzyme system, it was 6 6 suspected that more than one NAD-linked polyol dehydro genase was present in extracts of A. suboxydans. It was

6 8 71 — 73 found ' later that at least five different polyol dehydrogenases were present that were collectively named 6 8 soluble dehydrogenase. These dehydrogenases are (1) NADP- 6 8 linked xylitol dehydrogenase which oxidizes xylitol to

L-threo-pentulose, (2) NAD-linked D-erythro dehydrogenase^ which oxidizes pentitols having the g-erythro configuration,

(3) NAD-linked D-glucitol dehydrogenase ' ' (NAD-linked 6 8 g-xylo dehydrogenase ) which oxidizes pentitols, hexitols and heptitols having the g-xylo configuration, (4) NADP- 71 72 linked D-mannitol dehydrogenase ' (NADP-linked D-xylo 6 8 dehydrogenase ) which oxidizes pentitols and hexitols having the g-xylo configuration, and (5) NADP+-linked D-mannitol 6 6 73 dehydrogenase, ' which oxidizes g-glucitol and D-arabinitol as well as D-mannitol. The functions of these enzymes in the living cell are yet to be determined. Their activities are not displayed in growing cells. This follows from the fact that the cells contain at least three different xylitol dehydrogenases in the cytoplasm, but xylitol is not oxidized by growing cultures. On the other hand, when resting cell suspensions were incubated at pH 7.5, D-glucitol was oxidized 6 6 to D-fructose, indicating that NAD-linked D-glucitol dehydrogenase was active under this condition. More work has 172 to be done before any enzymic explanation of this range of behavior can be given.

(71) J. T. Cummins, T. E. King, and V. H. Cheldelin, J. Biol. Chem., 224, 323-329 (1957).

(72) J. T. Cummins, V. H. Cheldelin, and T. E. King, J. Biol. Chem., 226, 301-306 (1957).

(73) D. R. D. Shaw and F. L. Bygrave, Biochim. Biophys. Acta, 113, 608-610 (1966). II. STATEMENT OF THE PROBLEM

The purpose of this undertaking was to cyclize the carbohydrate chain of an acyclic-sugar nucleoside. It was proposed that, if the microorganism Acetobacter suboxydans could be adapted to oxidize an acyclic-sugar nucleoside according to the Bertrand— Hudson rule, then the sugar chain would cyclize to a furanose ring. It was further proposed that, when using (IS)-1-(6-chloropurin-9-yl)-1-S- ethyl-l-thio-D-glucitol as the substrate, the fermentation product would be a novel "reversed" nucleoside sterically resembling the clinically important 9-(g-D-arabinofuranosyl) adenine (Ara-A) structure, with the reducing center at the opposite end of the sugar ring. The biological activity of this product was to be evaluated.

173 III. RESULTS AND DISCUSSION

A. Study of the Fermentation by Acetobacter suboxydans

The microbiological oxidation of the sugars and their derivatives by the acetic acid bacterium, Acetobacter suboxydans, has been studied extensively. In this study, adaptation of this microorganism to oxidize such unfavorable substrate as a nucleoside was attempted for the first time.

At the outset, the equipment used in this study is explained in detail as it is critical for success of the reaction. The containers used for fermentation are 250-mL

Erlenmeyer flasks. Each flask contains 50 mL of liquid

(1/5 of the volume of the flask) because this is the optimal 5 8 ratio for best results. The flask is capped with a lint 5 8 square instead of the conventional cotton plug, as it allows better gas and heat transfer. It is less likely to be wetted by the culture medium, and it reduces the chance of contamination during transfer. These lint pads consist of layers of cotton sandwiched between gauze and are clamped around necks of the flasks by means of spring wire clamps.

Baffles made from an aluminum baking pin were sometimes used,

174 175 because the oxygen transfer rate (OTR) was found increased 74 75 substantially by this device. ' Antifoam spray was applied

(74) E. L. Gaden, Jr., Biotech. Bioeng., 99-103 (1962) .

(75) G. L. Solomons, "Methods and Materials in Fermentation", Academic Press, 19 69, Chapter 1. whenever the baffle was used to avoid the severe foaming that the baffle caused. The shaker used is the rotary type, with speeds adjustable between 0-500 rpm.

The culture technique most frequently used is the proliferating-cell method, in which the substrate broth is inoculated with a culture of A. suboxydans grown on

D-glucitol. This technique has been used for the oxidation of simple alditols and sugars. These single-step oxidations by

A. suboxydans are well documented and the optimal conditions 5 8 are established. Therefore, this method was chosen for preliminary trials.

It was decided to check the oxidation of D-glucose diethyl dithioacetal first for two reasons. First of all, two different groups have reported successful attempts to 34 76 77 oxidize the compounds after others had failed. ' '

(76) B. Iselin, J. Biol. Chem., 175, 997-998 (1948).

(77) L. Hough, J. K. N. Jones, and D. L. Mitchell, Can. J. Chem., 37, 725-730 (1959). 176

Jones et al.^ oxidized the compound by the proliferating culture method without agitation, and obtained the product in 60% yield as syrup, whereas Overend et al.^ used isolated cells method under oxygen and obtained the crystal- 7 8 line procudt in 62% yield. Recently, Szarek et a_l. reported, without detail, that the yield was improved to 1 0 0 % by the isolated-cell method. Secondly, the oxidation may be used as reference with which the result of oxidation of acyclic- sugar nucleosides may be compared.

Several trials were carried out by the proliferating culture method (See Table 2). In each instance, 2— 3 day old cells of A. suboxydans (ATCC 621 H) were used, because the 79 cell growth reaches the exponential phase at this stage 80-82 and it is found that exponentially growing cells have

(78) G. W. Schnarr, W. A. Szarek, and J. K. N. Jones, Appl. Environ. Microbiol., _33, 732-734 (1977).

(79) B. L. Batzing and G. W. Claus, J. Bacteriol., 113, 1455-1461 (1973).

(80) K. R. Butlin, Biochem. J., _32, 508-512 (1938).

(81) Z. Fencl, J. Ricica, and J. Simmer, Symp. Int. Congress Microbiol., IX, Moscow, 1966, pp. 159-167.

(82) T. Kono and T. Asai, Biotech. Bioeng., 11, 293-32,1 (1969). the highest activity. The inoculated media, contained in either baffled or plain Erlenmeyer flasks, were incubated O with continuous rotary shaking (^350 rpm) at 34 . The oxidation was carefully monitored and adjusted to keep the pH range Table 2. Oxidation of g-glucose diethyl dithioacetal and (IS)-1-(6-chloropurin- 9-yl) -1-S-ethyl-l-thio-D-glucitol.

Rp- Substrate Yield Time Method— Substrate C one. Substrate Product (%) (days) (% in 50 mL)

D-Glucose diethyl 0.52 0.81 56 6 A (baffled) 4 dithioacetal (1 ) 50~ - 10 A (baffled) 2 89 10 B ( 3 mL) 2

(IS)- 1 - (6-Chloro- purin-9-yl)-1-S- 0.21 0.47 0_ 10 A (baffled) 1 ethyl-l-thio-D glucitol (25) 40 6 B ( 2 mL) 1 31 7 B ( 2 mL) 0.4 33 10 B ( 2 mL ) 0.6 70 8 B ( 6 mL, 0.6 under oxygen)

— Solvent system: butanone saturated with water for 1, 3:2 acetone— benzene for 25. — A: proliferating culture method, B: isolated-cells method, the wet-packed volume of cells suspended is indicated in parenthesis. — The bracket indicated that the experiments were performed simultaneously. 178 within 5-6. The best yield obtained was 56%, which is comparable to the 60% reported by Jones56 but the time of oxidation is decreased by nearly half. However, under the same conditions and using the same cells, no oxidation of the potential substrate (l£>)-1-(6-chloropurin-9-yl)-1-S- ethyl-l-thio-g-glucitol (25) occurred.

Apparently, more-vigorous conditions had to be used to adapt the organism to oxidize an unfavorable 7 8 substrate. The recent work of Szarek et al. clearly suggested the superiority of the isolated-cell method over the proliferating-culture method for this purpose. This proposal is confirmed by the results in this study. As shown in Table 2, D-glucose diethyl dithioacetal was oxidized in

89% yield by this method, in which cells isolated from 300 mL of standard D-glucitol broth were used. The oxidation product was obtained as syrup, in accord with results reported by

Williams and Jones.56

This method was also applied to study the oxidation of 25. The reaction was monitored by t.l.c. with the solvent system of 3:2 acetone— benzene. After 2— 3 days, a new faster-migrating spot appeared, indicating that the substrate was being oxidized to a new compound. The reaction was complete in 6 days and the yield was 40%. Attempts to raise the yield by increasing the number of added cells failed.

However, when the suspension with about 6 mL of wet-packed cells was shaken under an atmosphere of pure oxygen, 179 the yield was significantly raised, to 70%. The simplest explanation for the effect of pure oxygen is that, under the pure oxygen, the oxygen-transfer rate is raised about five-fold and thus ensure a sufficient supply of oxygen to keep all or most of the cells alive and active. It is known that vigorous aeration is required for successful oxidation, and it is especially true at this unusually high concentration 10 9 of cells (^10 cells/mL) as compared with ^10 cells/mL in 6 3 proliferating culture. This account is consistent with the observation that a suspension of ^6 mL of wet-packed cells gave almost identical yield when vl/3 the amount of cells were used under a normal atmosphere.

During the course of this study with A. suboxydans, it was noticed that the oxidizing power of the bacteria deteriorated with each subculture, and after about one year, the oxidative power was lost completely in some subcultures and obvious changes in colonies were observed. The mutant frequency of acetic acid bacteria has been the subject of 83 84 a great controversy. Schimwell, ' on the one hand, reported the extraordinary facility with which the bacteria were found to give rise to spontaneous mutants which had lost or 8 5 gained some major biochemical property. DeLey, on the other hand, found the frequency comparable with that in other 8 6 genera, and Leisinger et al. determined that the mutant “ 6 frequency was 10 . Whichever report is correct, it is important that the microorganism be checked before oxidation 180 study is carried out. It is found that ribitol serves this

(83) J. L. Schimwell, Antonie van Leeuw, 26_, 169- 181 (1961).

(84) J. L. Schimwell and J. G. Carr, Antonie van Leeuw, 26, 383-396 (1961).

(85) J. Schell and J. DeLey, Antonie van Leeuw, 28, 445-465 (1962).

(8 6 ) T. Leisinger, W. Jagger, P. Weber, and L. Ettlinger, Arch. Mikro., 5_7, 76-92 (1967). purpose satisfactorily for several reasons: (1) High sensiti vity. Ribitol is oxidized near quantitatively to L-erythro- pentulose. Any decrease in oxidative power can be detected by the presence of the starting material after oxidation.

(2) Short incubation time. The oxidation of ribitol (2%) is complete within 30 h in proliferating culture. (3) Avai- 78 87 lability of a detection system. ' The oxidation may be monitored by paper chromatography developed with the solvent

(87) T. Yamada, M. Hisamatsu, and M. Taki, J. Chromatogr., 103, 390-391 (1975). system of 7:1:2 (v/v/v) n-propanol— ethyl acetate— water.

The compounds are detected by spraying the chromatogram successively with NalO^, AgNO^, and NaOH solution.

The best result was obtained by the following procedure: The most active cells, checked by ribitol oxidation, were used to inoculate 15 Erlenmeyer flasks containing 50 mL of standard D-glucitol broth each. 181 O The broths were shaken for 2 days at 34 . The actively O growing cells were harvested by centrifuge at 25 . The

collected cells were washed with phosphate buffer (pH v6 )

and centrifuged. This was repeated several times to remove the nutrients and metabolites. These cells were suspended in the substrate solution immediately. The substrate solution was prepared by dissolving the substrate in demineralized water and stored in a ground glass Erlenmeyer flask with

24/40 joint. The oxygen reservoir, which was made of a balloon and a two-way stopcock, was attached to the flask through a 24/40 adaptor. This arrangement minimizes the evaporation of water molecules. Since the whole system is not completely airtight, oxygen has to be replenished everyday to make up for the loss through diffusion. The whole apparatus was shaken on a rotatory shaker for 8 days.

After the reaction is complete, the cells can be removed by centrifugation, but better results were obtained by filtration with activated charcoal. The solvent was removed by freeze-drying. The resulting syrup was extracted with hot methanol and the mixture filtered to remove an insoluble precipitate. Because the hydrophilic nature of the compound, a substantial amount of the product was lost when purified by column chromatography with 4:1 chloroform— methanol. However, preparative thin-layer chromatography worked well with small-scale (100 mg) purification to give a 70% yield, although it is suspected that loss of product 182 also occurred in the process, because the starting compound was present only in small amount as'judged by t.l.c. examination, and less than 10% of the starting compound was recovered. Attempts to purify the mixture by paper and thin- layer chromatography on micro-crystalline cellulose were unfruitful for no suitable solvent system was found. The product was isolated as a chromatographically pure syrup.

The 90-MHz ^H-n.m.r. of the product in deuterium oxide showed three resonances at 6 8.82, 8.77, and 8.71 for H-2, and one resonance at 8.59 for H- 8 , giving evidence that three tautomeric forms were present. The H- 6 ' signal of the preponderant tautomer at 5.87 give Jr , 5.28 Hz. D f u Another H-6 ' doublet (minor) overlapped with the major one and made it impossible to measure the coupling. Quartets at 2.36 and triplets at 0.94 arose from the ethylthio group.

All other proton signals are uninterpretable. From the integrals of the three H-2 signals, the ratio of three tautomers is determined as 12:12:76.

The c.d. spectrum and optical rotation of compound

43 indicated that the configuration at C-6 1 remained the same as that at C-l' of the starting compound 25. The electron- impact mass spectrum provided no useful information because the compound decomposed prior to ionization. 183

B. Carbon-13 N.m.r. Spectral Study and the Equilibrium

in Solution of (IS)-6- (6-chloropurin-9-yl)-6-S-

ethyl-6-thio-D-ido-hexulose (43)

The carbon-13 spectrum in dimethyl sulfoxide-dg of the fermentation product 43 was studied for two purposes; first of all, to determine the site of oxidation in the fermentation product, and secondly, to determine the tautomeric composition of this product in solution.

In earlier studies, characterizations of the ketose products were often made by irreversible and sometimes lengthy chemical transformations such as hydrazone or osazone formation, and sodium borohydride reduction, which converted products back into starting material, together with their 13 epimers. With the advancement of C-n.m.r. spectroscopy, most of these transformations become ununcessary and the site of oxidation can be unequivocally established, sometimes with the help of related compounds. Because of the great sensitivity, the method allows detection of minor components, and the large range of chemical shifts observed for the anomeric carbon atom of various tautomers makes assignments and quantitative estimation possible.

In this study, the site of oxidation was expected to be C-51, the carbon atom next to the terminal hydroxymethyl group. If this is the case, the product should exist, in neutral solution, in three tautomeric forms (See Fig. 2), Cl 1

OH EtSCH 6 ) 2 I 5H C0H c h 2o h 1 H0CH 4 HCSEt I 3 HCOH HCSEt 2 C=0 I lCHgOH

Cl 6-form acyclic-form a-form

Figure 2: Three tautomeric forms of 43 at equilibrium in neutral solution 184 185

with the cyclic form preponderant at equilibrium, according

to the general observation that the acyclic forms of reducing

sugars are present only in minute proportions at equili

brium.8 8 '89

(8 8 ) W. Pigman and H. S. Isbell, Advan. Carbohydr. Chem., 23, 11-57 (1968).

(89) E. L. Eliel, N. L. Allinger, S. J. Angyal, and G. A. Morris, "Conformational Analysis", Interscience, N. Y., 1965, Chapter 6 .

13 The C-n.m.r. spectrum showed a complicated pattern

of signals, as would be expected, were there three tautomers

present. To simplify the picture, the spectrum was divided

into four regions; (1 ) alkyl region, (2 ) carbohydrate region,

(3) purine region, and (4) carbonyl region. No attempt was made to assign all the signals.

(1) Alkyl region: There are only three signals. The

resonance at 6 14.3 was assigned to the methyl carbon and

the resonances at 24.9 and 24.5 were assigned to the methylene

carbon of the ethylthio group for the same reasons discussed

in part one. The signal at 24.5 was assigned to the acyclic

form by comparison with the starting compound 25 (See Table 3).

The fact that the methyl carbons of the ethylthio group in all three isomers have identical chemical shifts suggests that the S-ethyl group extends away from the rest of the sugar moiety, as in the acyclic nucleosides. 186

13 3 Table 3. C-N.m.r. chemical-shift data— for solutions of D-fructose and nucleoside analogues and 43.“

------J_------Isomer atom— D-Fructose- 43 "25

C-l ' 61.4 62.6 C-2 ' 104.4 106.7 C-3 ' 81.2] 79.5 a-f C-4 C - 4 ' ' 83.11-83.1] ■ 80.1 C-5 • 76.0 n . s, C -6 ' 61.4 60.4 sch2ch3 24.9

SCH2— 3 1 4 .3 C-l' 63.3- 63.9 C-2 1 102.2 103.6 C-3 ' 76.0 78 g-f C-4 C-4' ' 82.182.1 76 C-5' 75.6, 75.5 C-6 ' 6 3.4— 59.7 sch2ch3 24.9 sch2ch3 14.3

C-l ' 66.2 63.22 C-2 1 211.6 71.3 ~ ^ C-3 ’ „ n.s. 71.4 e C-4C-4' ’ n.a.— n.s. 70.0 chain C-5 ’ 72.5 73.5 C -6 • 61.2 61.5 soh2ch3 24.5 24.5 sch2ch3 14.3 14.3

— ppm downfield from Me^Si in dimethyl sulfoxide-dg. — From Ref. 90; the numbering for 43 is shown in Fig. 2. ~ — Not assigned, cl — Assignments could be reversed. — Not available. — Not seen (not separated from major signals). 2 The numbering for 25 is reversed to correspond to that for 4J). ~ 187

(2) Carbohydrate region: The signals appeared in 90 this region were assigned based on the data available

(90) W. Funcke and A. Kleiner, Justus Liebigs Ann. Chem., 1232-1236 (1975). for the a- and g-furanose forms of D-fructose (See Table 3) and for acyclic-sugar nucleoside 25.

The two resonances at 6 106.7 and 103.6 were assigned to C-2' (anomeric carbon) of the ct- and g-forms, respectively, following the same order observed for a-D-fructofuranoside

(104.4) and S-D-fructofuranose (102.2) and assuming that this relative order was not altered by the configuration of the group at the other end of the molecule because in D-fructose this group (CH^OH) is on top of the furanose ring, whereas in 43, which is a L-xylo derivative, the group (SEtCHB) is below the ring. Although the chemical shift data in dimethyl sulfoxide are scarce, there are numerous studies done in deuterium oxide. The chemical shift of anomeric carbon of g-D-fructofuranose (a, 101.7) and that of a-L-sorbofuranose 91 (b, 101.9) supported the validity of the assumption.

OH OH a b 188 92 There are also report ' that the chemical shift of anomeric

carbon depends on the configuration of the neighboring

hydroxyl group (cis or trans), but is insensitive to the

configuration elsewhere on the ring.

(91) L. Que, Jr. and G. R. Gray, Biochemistry, 13, 146-153 (1974).

(92) A. J. Angyal and G. S. Bethell, Aust. J. Chem., 2j), 1249-1265 (1975).

(93) G. R. Gray, Accounts Chem. Res., 9^, 418-424 (1976) .

The integrals of these two anomeric peaks clearly

indicate that 6 form is the preponderant component. Therefore,

the set of resonances at 59.7, 63.9, 75.5, 76.2, and 78.1 were assigned to the 8 form from their intensities. Among

them, the signal at 59.7 was readily assigned to C-6 1 and that

at 63.9 to C-l’ by their off-resonance splitting patterns.

The assignments of signals for the acyclic form were made by

comparison with those of 25, assuming that the oxidaiton of hydroxyl group at C-5 1 has a negligible effect on remote

carbon atoms (6 or e). Thus, the signal at 61.2 was assigned

to C-6 1. The remaining C-6 ' signal at 61.2 was assigned to the a form. There were also three signals arising from the

C - l 1 carbon atoms. The one at 63.9 was already assigned to the 6 form from its peak intensity. Another one at 62.6 was assigned to the a form in accordance with the assignments made for the a- and B-furanose forms of D-fructose 189

(See Table 3). The remaining one at 66.2 was then assigned to the acyclic form. These assignments are in agreement with their relative integrals. To quantitate the ratios of the three forms, the integrals of the signals for C-l' and C-6 1 were averaged. The result showed that the equilibrium compo sition is 7% acyclic form, 11% a form and 82% 6 form. It is somewhat unexpected to find the 8 form preponderant, because of the hydroxymethyl group at the one end and the bulky purine group at the other end are on the same side of the furanose ring. This observation provides another line of evidence that the S-ethyl group extends away from the rest of the sugar moiety. Also, it is surprising to observe such a high proportion of the open-chain form at equilibrium at 0 25 .

(3) Purine region: There are altogether five peaks in this area. They were assigned readily by following the same arguments discussed earlier (p. 116). in principle, more peaks were to be expected. There are two possibilities why this is not so. Either the shifts of the purine carbon atoms in all three tautomeric forms are identical and appear as one, or that the a form and open-chain form are present in such low concentration that their signals are buried under the more-intense 8 peaks for C -8 and C-2, or not detected for the tertiary carbons C-4, C-5. The latter possibility was ruled out because the carbonyl carbon peak of the acyclic form was detected with moderate intensity. 190 It was concluded that the three tautomeric forms had identical or very close purine carbon shifts.

(4) Carbonyl region: The lone resonance here was assigned to C-5 of the acyclic form.

From the spectrum, the site of oxidation is esta blished beyond doubt as C-5' of compound 25. The C-2 1 site is ruled out as the chemical shifts of C-l 1 are only changed slightly. The C-3' site is ruled out for the same reason.

The C-4' site is ruled out for its inability to assume a cyclic form if oxidized. The possibility at C- 6 ' is also ruled out because if it were oxidized, there should be no carbon that gives triplet splitting in the proton-decoupled spectrum.

In summary, the site of oxidation is established as C-5' of the sugar chain, which is capable of cyclizing to give only the furanose ring. The equilibirum composition of (l£>)-6-

(6-chloropurin-9-yl)-6-S-ethyl-6-thio-D-ido-hexulose in 0 13 dimethyl sulfoxide-d^ at 25 , as determined from C-n.m.r. — 6 spectrum, is 7:11:82 acyclic:a-furanose:6-furanose.

C. Synthesis of 1,2,3,4-tetra-0-acetyl-6-(6- chloropurin-9-yl)-6-S-ethyl-6-thio-a,8-D-ido-hexulose (44)

The fermentation product 43 was acetylated by the conventional pyridine— acetic anhydride method. The reaction mixture turned dark being kept at room temperature for a few hours. The reason for the discoloration is unknown, but the 191 possibility of decomposition of purine base was excluded.

When water was added dropwise to the solution, orange precipitate formed. The precipitate was filtered and the filtrate was extracted with several portions of chloroform to recover the acetylated product remained in the filtrate.

The product was purified by preparative plates of silica gel and appeared as a faint orange, but strongly uv-active band immediately above a yellow band, which is not uv active.

The band was excised and extracted with acetone, in which the compound is most soluble. The reaction gave 28% yield of the acetylated compound as light-yellow glass. The discoloration is responsible for the low yield of the O reaction. The compound is strongly levorotatory (-6 9 ), indicating the stereochemistry at C-6 1 was not altered.

1. Mass Spectrum of 44— The electron-impact mass spectrum of this peracetylated "reversed" nucleoside was recorded. The general appearance of the spectrum is distinctly different from that of its acyclic counterpart (p. 119).

Loss of acetate (m/e 43) generated the most-intense peak, as expected. The molecular ion (m/e 54 4) and the subsequent losses of acetic acid giving m/e 484 and 424 were the only detectable peaks beyone m/e 400. Other work 94 95 on nucleoside fragmentation and peracetylated furanose were consulted in interpreting the spectrum (See Fig. 3). ♦0S-4-H

BH AcOH^O HOAc AcOH^i AcO, m/e 154 I----- OAc OAc HOAc. m/e 4 84 m/e 424

\ B B AcO o E1S— H AcO / 0 EtS+ H E tS -j-H < * c0r c AcO, A cO H £ AcOHjjC . AdV OAc / OAc OAc m/e 544 \ m/.e 384 -HOAc ' SEt C -l/C -2 B AcO , 0 .6 Ac

AcOH2C

kOAc m/e 317 m/e 324 m/e 323 192

Figure 3: Probable fragmentation of 44. 193

Table 4. Relative intensity in the fragment of mass spectrum of 44.

Relative , Relative intensity — — intensity

545 0.1 231 1.1 544 0.2 230 0.6 485 0.6 229 2.0 484 1.0 228 2.1 425 0.4 227 2.4 424 1.5 224 0.7 384 0.2 223 1.8 378 0.2 212 0.5 377 0.4 211 2.4 360 0.2 210 0.4 359 0.4 209 2.4 349 0.2 199 1.2 348 0.5 198 0.5 332 0.5 197 1.0 331 0.7 196 0.8 326 0.3 195 0.8 324 1.0 193 0.9 323 0.8 192 0.8 310 0.6 168 3.0 309 0.8 167 10.0 308 0.5 156 4.5 307 0.8 155 6.5 306 0.6 154 10.0 305 1.0 149 26.0 271 1.5 119 9.0 261 1.3 111 15.0 252 . 0.4 60 26.0 251 1.4 43 100.0 194

(94) S. J. Shaw, D. M. Desiderio, Y. Tsuboyama, and J. A. McCloskey, J. Am. Chem. Soc., 9_2, 2510-2522 (1970).

(95) N. K. Kochetkov and 0. S. Chizhov, Adv. Carbohydr. Chem. Biochem., 2^, 39-93 (1966).

1 1 2. H-N.m.r. Spectrum— The 90-MHz H-n.m.r. spectrum

of 44 was recorded in acetone-dg (See Fig. 4) and chloroform-d.

In acetone, the H-6 ' signal resonates at 6 6.05,

thus resembling the behavior in acyclic-sugar nucleosides.

The H-5' signal is a doublet of doublets, showing ,

9.5 Hz, which indicates the eclipsed relationship between 9 6 H - 4 1 and H-5' and is in agreement with the value calculated

(96) F. E. Hruska, A. A. Grey, and I. C. P. Smith, J. Am. Chem. Soc., 92, 214-215 (1970).

from the Karplus equation. The H-3 1 and H - 4 1 signals over

lapped. The H-l' and H-l" protons interacted only with each

other because of the lack of a neighboring proton at C-2 1

(anomeric carbon). This simplified the signals to two doublets having spacings of 11.55 Hz. The acetate peaks were covered under the more-intense solvent peaks, and no information

regarding the number of acetates could be extracted. However,

the spectrum in chloroform-d showed four acetate peaks at

6 2.21, 2.12, 2.10, and 2.07, respectively. Because of the

lack of anomeric proton, the anomeric configuration at C-2 1 was not determined. H-3 ' ,4

H-l' H-l H-6 1 II-5

_L j . i k__k. JL 6.0 5.0 4.0 <5

Figure 4: Partial n.m.r. spectrum (acetone-dg) of 1,2,3,4-tetra-0-aectyl- 195 b- (6-chloropurin-9-yl)-6-S-ethyl-6-thio-D-ido-hexulofuranose (44) Table 5. Toxicity data for 6-chloropurine nucleoside analogs.

Molarity for Compound Cells 50% Inhibition

(1R,IS)-2,3,4,5,6-Penta-O-acetyl-l- (6-chloropurin-9-yl)-1-S-ethy1-1-thio- L-1210 > io“ 4 D-glucitol (17)

(1R,IS)-2,3,4,5,6-Penta-O-acetyl-l- i i— o A

(6-chloropurin-9-yl)-1-S-ethyl-1-thio- L-1210 i

D-mannitolZ (30) Ay

(1R)-1-(6-Chloropurin-9-yl)-1-S-ethyl- 1-thio-D-altritol (49a) L-1210 > 10 4

(1R)-1-(6-Chloropurin-9-yl)-1-S-ethy1- ~ 5 1-thio-D-mannitol= (52a)/~\s L-1210 5xl0

(1R,IS)-1-(6-Chloropurin-9-yl)-3-deoxy- 1 rH o 1-S-ethyl-l-thio-D-arabino-hexitol (5.4) L-1210 A

(IS)-6- (6-Chloropurin-9-yl)-6-S-ethyl-

6-thio-D-ido-hexulose— ----- (43) A/ L-1210 > 10~4 196 IV. EXPERIMENTAL

General Methods Infrared spectra were recorded with a Perkin— Elmer Infracord spectrophotometer and ultra violet spectra with a Cary 15 u.v. recording spectrophoto meter. Optical rotation was measured with a Perkin--Elmer

Model 141 recording polarimeter. Electron-impact mass spectra were recorded by C. R. Weisenberger with an AEI MS-9 double-focusing mass spectrometer at an ionization potential of 70 eV and an accelerating potential of 8 kV. ‘*'H-N.m.r. O were recorded with a Bruker HX-90 spectrometer at v25 .

The following conventions are used to refer to n.m.r. spectra: d, doublet; dd, doublet of doublet; m, multiplet; q, quartet; 13 t, trxplet. C-N.m.r. spectra were recorded by C. Cottrell with Bruker WP-80 spectrometer operating at 20 MHz in the O Fourier-transform mode at ^25 . Chemical shifts were reported in ppm relative to Me^Si. C.d. spectra were recorded with a

Durrum-Jasco o.r.d.— c.d. spectrometer at ambient temperature with a 1-cm optical cell. T.l.c. was performed on silica gel 60 F-254 (#5765, E. Merck). Preparative t.l.c. was performed on chromatoplates (200 X 200 X 2.5 mm) of silica gel 60 PF-254 (#7747, E. Merck), containing 1% of Lumilux

197 198

Green 25.

Organism— Acetobacter suboxydans (ATCC No. 621H) was purchased from the American Type Culture Collection, O Rockville, Md., maintained at 4 on an agar slant containing

D-glucitol (5%, w/v), yeast extract powder (0.5%, w/v), potassium dihydrogenphosphate (0.05%, w/v), D-glucose

(0.05%, w/v), and agar (1.5%, w/v). The cells were subcultured at about 3-month intervals by inoculating a broth containing the same ingredients as described in agar slant. The broth

O was shaken for 2— 3 days at 34 . Fresh slants were inoculated O from the broth and kept for 2--3 days at 34 until cream- colored colonies became visible. The slants were then stored O at 4 C for future use.

All cells used were grown by heavily inoculating standard D-glucitol broth (50 mL of broth in 250 mL Erlen- meyer flask) containing D-glucitol (7.2%, v/v, 70% g-glucitol solution, ICI America, Inc.), yeast extract powder (0.5%, w/v), potassium dihydrogenphosphate (0.05%, w/v), and D-glucose

(0.05%, w/v). The flasks were covered with lint square and sterilized. Inoculations were carried out in sterilized air hood to avoid contamination. The shaker used was the portable gyratory shaker (Model G-2, New Brunswick Scientific Co., O Inc.). The broths were shaken (300 rpm) at 34 for 2— 3 days and then harvested and used immediately. 199

Oxidation of (IS)-1-(6-chloropurin-9-yl)-1-S- ethyl-l-thio-p-glucitol (25.) to (IS) -6- (6-chloropurin-9-yl) -

6-S-ethyl-6-thio-D-ido-hexulose(43) Method A . Standard

D-glucitol broths (600 mL with 50 mL in each of 250-mL

Erlenmeyer flask) were inoculated and grown for 3 days, and the cells were collected by centrifugation for 30 min O at 2,000 g at 25 . The cells were then washed with 0.01 M phosphate buffer and recollected by centrifugation. The procedure was repeated two to three times to remove L-sorbose and nutrients. The collected bacterium (^6 mL of wet-packed cells) was suspended in substrate-containing solution, which was prepared by dissolving the substrate (25, 3 00 mg) in distilled water (50 mL) in a 250-mL Erlenmeyer flask without O sterilization. The solution was then incubated at 34 with continuous rotatory shaking (300 rpm) for a period of 10 days.

The oxidation was terminated by adding 150 mL of ethanol and the mixture was left for 1 h. A small amount of activated charcoal was added and the solution boiled for a few minutes and filtered. The filtrate was evaporated under diminished pressure to remove ethanol and the remaining water was removed by freeze-drying. The resulting crude syrup was applied to a preparative t.l.c. plate (200 X 200 X 2.5 mm), which was developed twice with 3:2 acetone— benzene. The faster- migrating band was excised and extracted first with warm ethanol then with cold methanol. The other band was the starting compound 25. The extract was evaporated to give 200

43 as a homogeneous syrup, yield, 0.10 g (33%); Rp 0.47 Me OH — (3:2 acetone— benzene), A 267 nm (e 3,500); in 3.x vfllm 3200-3500 (OH), 2980, 2940 (C-H), 1730 (C=0, weak), max 1595, 1570 (purine ring), 1490, 1400, 1345, 1150— 1050 cm ^

(C-O-H); [a]p7 -53 (c 0.58, methanol); c.d.: [9]300 0 '

[0]285 + 1 '406 ' [0]28O ° ' [0]269 ” 6 '563 ' [0]257 0 '

[0 ^ 248 + 4 '375°' [0] 232 + 6' 8 7 5 °'' 1H-n.m.r. (D20 ): 8.81, 8.77,

8.72 (H-2), 8.59 (H-8 ), 5.87 d (J5 , gI 5.28 Hz, H-6 1),

4.13 — 4.00 m (H-3 1 , 4 1 ) , 3.51 m (H-l’,1"), 2.36 q (SOH2CH 3) ,

0.94 t (SCH2CH3 ); l3C-n.m.r. (in Me^O-dg); 14.27 (SCH2CH3 ),

24.46, 24.88 (SCH2CH3 ), 59.70, 60.37, 61.16 (C-6 ’), 62.55,

63.89, 66.19 (C-l1), 72.50, 75.47, 76.20, 78.08 (C-3'-5'),

75.47, 74.48, 80.08 (C-3'-5•, 3 form), 103.56 (C-2', a form),

106.71 (C-2', 8 form), 211.59 (C-2', chain form), 130.73 (C-5),

146.81 (C-8 ) , 149.17 (C-4), and 151.54 (C-2,6 ).

Anal. Calc, for C^H^ClN^Oj-S: C, 41.48; H, 4.56;

Cl, 9.30; N, 14.89; S, 8.50. Found: C, 41.56; H, 4.72;

Cl, 9.47; N, 15.03; S, 7.66.

Method B The substrate (300 mg) was dissolved in

50 mL of distilled water in a 250 mL Erlenmeyer flask that was tightly stoppered and connected to an oxygen reservoir.

The same amount of cells as described in Method A was suspended and the incubation was performed under an atmo sphere of oxygen instead of air. The oxygen was replenished at 1— 2 day intervals. The oxidation was terminated after

8 days. Following the same procedure of purification, 201

a homogeneous syrup was obtained; yield, 0.21 g (70%),

identical with that produced by Method A .

1,2,3,4-Tetra-0-acetyl-6-(6-chloropurin-9-yl)-6-S-

ethyl-6-thio-a,g-D-ido-hexulofuranose(44) A solution of

compound 43 (50 mg, 0.13 mmol) in pyridine (2 mL) was cooled O to 0 , and then acetic anhydride (0.25 mL) was added. The

solution was kept overnight at room temperature and turned

dark. Cold water was then added dropwise to the solution, which was cooled in ice bath, until heavy precipitation

occurred. The orange-colored precipitate was collected by

filtration and the filtrate extracted with portions of

chloroform. The precipitate was dissolved in the chloroform

extract. The chloroform solution was concentrated and applied

to a preparative chromatoplate of silica gel, which was

developed with 85:15 chloroform— acetone. The uv active,

orange band, which was immediately above a uv inactive yellow band, was excised and extracted with several portions of acetone. Evaporation of the solvent gave 44 as a light yellow glass; yield, 20 mg (28%); R 0.48 (6:1 chloroform— —F 0 7 o T C R t acetone); [ot] - 68.6 (c 0.05, acetone); v 2880, 2800 i-j max (C-H), 1740 (C=0 of acetate), 1575, 1540 (purine ring), 1360,

1200, 1090— 1000 cm ^ (C-O-C); ^H-n.m.r. (90 MHz, acetone-dr): — 0 8.74 (H-2,8), 6.05 d (H-6 ', Jg , gI 5.40 Hz), 5.27 d (H-5',

£ 4 . 5 . 9.5 Hz), 5.66 — 5.47 m (H-3',4'), 4.74 d (H-l'),

4.45 d (H-l", J,, 11.55 Hz), 2.55 q (SCH0CH0), 1.18 t 0^0 £ (SCH2CH3 ). 202 Anal Calc, for C ^ H ^ C I N ^ S (544.1031). Found:

544.1041 (exact mass). BIBLIOGRAPHY PART ONE:

1. P. A. Levene and W. A. Jacobs, Ber., 42!, 2474-2478 (1909) .

2. F. Miescher, Hoppe-Seyler1s Med-Chem. Untersuch., 461 (1871).

3. P. Plosz, Hoppe-Seyler's Med-Chem. Untersuch., 461 (1871).

4. V. Luvabin, Hoppe-Seyler's Med-Chem. Untersuch., 465 (1871).

5. F. Hoppe-Seyler, Hoppe-Seyler1s Med-Chem. Untersuch., 486 (1871).

6 . A. Kossel, Z. Physiol. Chem., 5, 152 (1811).

7. R. Altmann, Arch. Anat. Physiol., Physiol. Abt. 524 (1889) .

8 . A. Kossel, Z. Physiol. Chem., 4_, 290 (1880).

9. A. Kossel and H. Steudel, Z. Physiol. Chem., 3_8, 49 (1903) .

10. A. Ascoli, Z. Physiol. Chem., 3_1, 161 (1900-1901).

11. A. Kossel and A. Neumann, Ber., 26_, 2753-2756 (1893).

12. P. A. Levene and W. A. Jacobs, Ber., 41, 2703-2707 (1908) .

13. P. A. Levene and W. A. Jacobs, Ber., 42, 1198-1203 (1909) .

14. P. A. Levene and W. A. Jacobs, Ber., 4_4, 746-753 (1911).

15. P. A. Levene and L. W. Bass, "Nucleic acids", New York, 1931. 203 204

16. P. A. Levene and R. S. Tipson, J. Biol. Chem., 97, 491-495 (1932).

17. P. A. Levene and R. S. Tipson, J. Biol. Chem., 104, 385-393 (1934).

18. A. R. Todd, J. Chem. Soc., 647-653 (1946).

19. B. Lythgoe, H. Smith, and A. R. Todd, J. Chem. Soc., 355-357 (1947).

20. J. Davoll, B. Lythgoe, and A. R. Todd, J. Chem. Soc., 833-839 (1946) and references cited therein.

21. J. D. Watson and F. H. C. Crick, Nature, 171, 737-738 (1953).

22. K. G. Cunningham, S. A. Hutchinson, W. Manson, and F. S. Spring, J. Chem. Soc., 2299-2330 (1951).

23. C. W. Waller, P. N. Fryth, B. L. Hutchings, and J. H. Williams, J. Am. Chem. Soc., 7_5, 2025 (1953).

24. N. Lufgren and B. Luning, Acta Chem. Scand. , 1_, 225 (1953) .

25. H. Yunsten, K. Ohkuma, and Y. Ishii, J. Antibiotics (Tokyo), Ser. A, 9_, 195 (1956).

26. E. H. Flynn, J. W. Hinman, E. L. Caron, and D. 0. Woolf, Jr., J. Am. Chem. Soc., T5_, 5867-5871 (1953).

27. J. Baddiley in "Nucleic Acids", Vol. 1., E. Chargaff and J. N. Davidson, Ed. Academic Press, New York, N. Y. 1955, pp. 177-187.

28. G. H. Hitchings, E. A. Falco, and M. B. Sherwood, Science, 102, 251-252 (1945).

29. R. 0. Robin, Jr., J. O. Lampen, J. P. English, Q. P. Cole, and J. R. Vaughan, Jr., J. Am. Chem. Soc., 67, 290-294 (1945).

30. G. P. Rhoads, Consulting Ed., Conference on 6-Mercapto- purine, Ann. N. Y. Acad. Sci., 6_0, Art. 2, 185 (1954).

31. 0. T. Avery, C. M. McLeod, and M. McCarty, J. Exptl. Med., 79, 137-158 (1944).

32. M. Meselson and F. W. Stahl, Proc. Natl. Acad. Sci., 44, 671-682 (1958). 205

33. R. M. Hamlin, Jr., R. C. Lord, and A. Rich, Science, 148, 1734-1737 (1965).

34. Y. Kyogoku, R. C. Lord, and A. Rich, Science, 154, 518-520 (1966).

35. R. R. Shoup, H. T. Miles, and E. D. Becker, Biochem. Biophy. Res. Comm., 23_, 194-201 (1966).

36. L. Katz and S. Penmann, J. Mol. Biol., L5, 220-231 (1966).

37. K. H. Scheit, Angew. Chem., 79_, 90 (1967).

38. F. S. Mathews and A. Rich, Nature 201, 179-180 (1964).

39. R. F. Stewart and L. H. Jensen, J. Chem. Phys., 40, 2071-2075 (1964).

40. J. Iball and H. R. Wilson, Nature, 198, 1193-1195 (1963).

41. B. Pullman, P. Claverie, and J. Caillett, Proc. Natl. Acad. Sci., _55, 904-912 (1966).

42. R. D. Hotchkiss, J. Biol. Chem., 175, 315-332 (1948).

43. G. R. Wyatt and S. S. Cohen, Nature, 170, 1072-1073 (1952).

44. R. G. Kallen, M. Simon, and J. Marmur, J. Mol. Biol., 5, 248-250 (1962).

45. I. Takahashi and J. Marmur, Nature, 197, 794-795 (1963).

46. H. Kasai, K. Nakanishi, R. D. MacFarlane, D. F. Torgerson, Z. Ohashi, J. A. McCloskey, H. J. Gross, and S. Nishimura, J. Am. Chem. Soc., 9J3, 5044-5046 (1976).

47. R. H. Hall, "The Modified Nucleosides in Nucleic Acids", Columbia University Press, New York and London, 19 71.

48. L. Jaenicke and F. Lynen, The Enzymes, Vol. 3, Chapt. 11, P. D. Boyer, H. Lardy, K. Mybrack, Eds., Academic Press, New York, 19 60.

49. M. E. Pullman, A. San Pietro, and S. P. Colowick, J. Biol. Chem., 206, 129-141 (1954). 206

50. P. Karrer, K. Schopp, and F. Benz, Helv. Chim. Acta, 18, 426-429 (1935).

51. R. Kuhn, K. Reinemund, F. Weygand, and R. Strobele, Ber., 6_8 , 1765-1774 (1935).

52. D. C. Hodgkin, J. Kamper, J. Lindsey, M. MacKay, J. Pickworth, J. H. Robertson, C. B. Shoemaker, J. G. White, R. J. Prosen, and K. N. Trueblood, Proc. Roy. Soc. (London), Ser. A, 242, 228-263 (1957).

53. T. Toraya, K. Ushio, SI Fukui, and H. P. C . Hogenkamp J. Biol. Chem., 252, 963-970 (1977).

54. E. L. Patterson, R. Milstrey, and E. L. R. Stokstad, J. Am. Chem. Soc., 78, 5868-5871 (1950).

55. H. Rembold, Vitamin Hormone, 23, 3 59-38 2 (1965).

56. S. Kaufman, J. Biol. Chem., 236, 804-810 (1961).

57. K. Lohmann, Biochem. Z., 233, 460-469 (1930); 254, 381-397 (1932).

58. B. Lythgoe and A. R. Todd, Nature, 155, 695-696 (1945).

59. J. T. Park, J. Biol. Chem., 194, 877-904 (1952).

60. E. P. Kennedy and S. B. Weiss, J. Am. Chem. Soc., 77, 250-251 (1955); J. Biol. Chem., 222, 193-214 (1956).

61. For a review, see "The Biochemistry of Adenosyl- methionine", F. Salvatore, E. Borek, V. Zappiz, H. G. Williams-Ashman, F. Schlenk, Eds., Columbia University Press, New York, 1977.

62. W. Bergmann and R. J. Feeney, J. Am. Chem. Soc., 72, 2809-2810 (1950).

63. M. R. Sheen, B. K. Kim, and R. E. Parks, Jr., Proc. Am. Assoc. Cancer Res., £, 63 (1968).

64. M. Ishizuka, T. Sawa, G. Koyama, T. Takeuchi, and H. Umezawa, J. Antibiotics, 21A, 1-5 (1968).

65. E. Fischer and B. Helferich, Ber., 4_7, 210-233 (1914).

6 6 . G. M. Hitchings and L. Goodman, U.S. Pat. 3074929 (1963).

67. J. M. Gulland, E. R. Holiday, and T. F. Macrae, J. Chem. Soc., 1639-1644 (1934). 207

68. J. M. Gulland and L. F. Story, J. Chem. Soc., 259- 261 (1938).

69. M. L. Wolfrom and P. McWain, J. Org. Chem., _30, 1099- 1101 (1965).

70. J. Farkas and F. Sorm, Collect. Czech. Chem. Comm., 32, 2663-2667 (1967).

71. E. Walton, S. R. Jenkins, R. F. Nutt, M. Zimmermann, and F. W. Holly, J. Am. Chem. Soc., £ 8 , 4524-4525 (1966).

72. D. Horton and C. G. Tindall, Jr., Carbohydr. Res., 17, 240-244 (1971).

73. D. Horton and S. S. Kokrady, Carbohydr. Res., 24, 333-342 (1972).

74. J. Defaye and Z. Machon, Carbohydr. Res., 2A_, 235-245 (1972).

75. B. R. Baker, J. P. Joseph, and R. E. Schaub, J. Am. Chem. Soc., 77, 5905-5910 (1955).

76. H. M. Kissman, C. Pidacks, and B. R. Baker, J. Am. Chem. Soc., 77, 18-24 (1955).

77. J. J. Fox, N. Yung, J. Davoll, and G. B. Brown, J. Am. Chem. Soc., 78, 2117-2122 (1956).

78. C. Ukita, H. Hayatsu, and Y. Tomita, Chem. Pharm. Bull. (Tokyo), 11, 1068-1073 (1963).

79. M. L. Wolfrom and H. B. Bhat, J. Org. Chem., 2757-2759 (1967).

80. R. S. Tipson, J. Biol. Chem., 130, 55-59 (1939).

81. B. R. Baker, J. P. Joseph, R. E. Schaub, and J. H. Williams, J. Org. Chem., 19_, 1780-1785 (1954).

82. B. R. Baker, in "Ciba Foundation Symposium on the Chem. and Biol, of Purines", Ed. by G. E. W. Wolstenholme and C. M. O'Connor, Little, Brown and Co. Boston, 1957, p. 120.

83. B. R. Baker, R. E. Schaub, J. P. Joseph, and J. H. Williams, J. Am. Chem. Soc., 77, 12-15 (1955). 208

84. B. R. Baker, R. E. Schaub, and H. M. Kissman, J. Am. Chem. Soc., 11_, 5911-5915 (1955).

85. L. M. Lerner and Y. Y. Cheng, Carbohydr. Res., 14, 297-303 (1970).

8 6 . G. E. Hilbert and T. B. Johnson, J. Am. Chem. Soc., 52, 2001-2007; 4489-4494 (1930).

87. J. J. Fox and I. Wempen, Adv. Carbohydr. Chem., 14, 283-380 (1959) and references therein.

8 8 . J. Pliml and M. Prystas, Adv. Heterocycl. Chem., 8^, 115-142 (1967).

89. M. Prystas and F. Sorm, Collect. Czech. Chem. Comm., 29_, 2965-2970 (1964).

90. L. Birkofer, A. Ritter, and H. P. Kuhlthau, Angew. Chem., 75, 209-210 (1963).

91. T. Nishimura, B. Shimizu, and I. Iwai, Chem. Pharm. Bull. (Tokyo), 11, 1470-1477 (1963).

92. T. Nishimura and I. Iwai, Chem. Pharm. Bull. (Tokyo), 12, 352-356; 357-361 (1964).

93. T. Nishimura, B. Shimizu, and I. Iwai, Chem. Pharm. Bull. (Tokyo), 12, 1471-1478 (1964).

94. T. Nishimura and B. Shimizu, Agr. Biol. Chem., 28, 224-229 (1964).

95. E. Wittenburg, Z. Chem., A_, 303-304 (1964).

96. M. W. Winkley and R. K. Robins, J. Org. Chem., 33, 2822-2827 (1968).

97. T. J. Bardos, M. D. Kotick, and C. Szantay, Tetrahedron Lett., 1759-1764 (1966).

98. B. Shimizu and A. Saito, Agr. Biol. Chem., 33_, 119- 121 (1969).

99. E. Wittenburg, Chem. Ber., 101, 1095-1114 (1968).

100. G. Ritzmann and W. Pfleiderer, Chem. Ber., 106, 1401- 1417 (1973).

101. U. Niedballa and H. Vorbruqgen, J. Org. Chem., 39, 3654-3660 (1974). 209

102. U. Niedballa and H. Vorbruggen, J. Org. Chem., 39, 3660-3663 (1974).

103. U. Niedballa and H. Vorbruggen, J. Org. Chem., 39, 3672-3674 (1974).

104. F. W. Lichtenthaler, P. Voss, and A. Heerd, Tetrahedron Lett., 24, 2141-2144 (1974).

105. G. Schramm, H. Grotsch, and W. Pollmann, Angew, Chem., 73, 619 (1961).

106. J. A. Carbon, Chem. Ind. (London), 529 (1963).

107. B. Helferich and E. Schmitz-Hillebrecht, Ber., 6 6 , 378-383 (1933).

108. T. Shimadate, Y. Ishido, and T. Sato, Nippon Kagaku Zasshi, E[2, 938-940 (1961) ; T. Sato, T. Shimadate, and Y. Ishido, ibid., 8_1, 1442 (1960); Chem. Abstr. , 56, 11692 (1961).

109. Y. Ishido, A. Hosono, K. Fujii, Y. Kikuchi, and T. Sato, Nippon Kagaku Zasshi, 8_7, 752 (1966); Chem. Abstr., 65, 17034 (1966).

110. K. Onodera and H. Fukumi, Agr. Biol. Chem., 21_, 526- 529, 864-869 (1963).

111. M. J. Robins and R. K. Robins, J. Am. Chem. Soc., 87, 4934-4940 (1965).

112. K. Antonakis and M. J. Arror, C. R. Acad. Sci. Ser. C, 27_2, 1982-1984 (1971).

113. N. Yamaoka, K. Aso, and K. Matsuda, J. Org. Chem., 30, 149-152 (1965).

114. C. P. J. Glaudemans and H. G. Fletcher, Jr., J. Org. Chem., 28, 3004-3006 (1963).

115. F. Keller, I. J. Botvinick, and J. E. Bunker, J. Org. Chem., 32., 1644-1646 (1967).

116. A. Holy and F. Sorm, Collect. Czech. Chem. Comm., 34, 3383-3401 (1969).

117. G. E. Keyser, J. D. Bryant, and J. R. Barrio, J. Org. Chem., 44, 3733-3734 (1979). 210

118 - J. A. Montgomery and H. J. Thomas, Adv. Carbohydr. Chem., ]/7, 311-369 (1962).

119. A. M. Michelson, "The Chemistry of Nucleosides and Nucleotides", Academic Press, New York, 1968, Chapt. 2.

120. L. Goodman and C. A. Dekker in "The Carbohydrate" Vol. IIA, Ed. by W. Pigman and D. Horton, Academic Press, New York, 1970, Chapt. 29.

121. "Nucleic Acid Chemistry", Part I and II, Ed. by L. B. Townsend and R. S. Tipson, Wiley— Interscience, New York, 19 78, Topic III.

122. M. L. Wolfrom, A. B. Foster, P. McWain, W. von Bebenburg, and A. Thompson, J. Org. Chem., , 3095-3097 (1961).

123. M. L. Wolfrom, P. McWain, and A. Thompson, J. Org. Chem., 27, 3549-3551 (1962).

124. M. L. Wolfrom, H. G. Garg, and D. Horton, J. Org. Chem., 29, 3280-3283 (1964).

125. M. L. Wolfrom, H. G. Garg, and D. Horton, J. Org. Chem., 20, 1096-1098 (1965).

126. M. L. Wolfrom, W. von Bebenburg, R. Pagnucco, and P. McWain, J. Org. Chem., 22.' 2732-2735 (1965).

127. M. L. Wolfrom, P. McWain, H. B. Bhat, and D. Horton, Carbohydr. Res., 23_, 296-300 (1972); D. Horton, Pure Appl. Chem., £2, 301-325 (1975).

128. D. C. Baker and D. Horton, Carbohydr. Res., 22., 117- 134 (1979).

129. K. C. Blieszner, D. Horton, and R. A. Markovs, Carbohydr. Res., 22' 241-262 (1980).

130. D. Horton and R. A. Markovs, Carbohydr. Res., 80, 356-363 (1980).

131. M. L. Wolfrom, H. B. Bhat, P. McWain, and D. Horton, Carbohydr. Res., 23, 289-295 (1972).

132. M. L. Wolfrom and P. J. Conigliaro, Carbohydr. Res., 20, 369-374 (1971).

133. D. Horton and S. S. Kokrady, Carbohydr. Res., 80, 364-374 (1980). 211

134. D. C. Baker, K. Blieszner, and D. Horton, "Nucleic Acid Chemistry", Vol. 2, L. B. Townsend, R. S. Tipson, Eds., Wiley, New York, 1978, pp. 627-637.

135. D. Horton, S. S. Kokrady, "Nucleic Acid Chemistry", Vol. 1, L. B. Townsend and R. S. Tipson, Eds., Wiley, New York, 1978, pp. 267-272.

136. D. C. Baker, S. S. Kokrady, and D. Horton, Ann. N. Y. Acad. Sci., 255, 131-150 (1975).

137. S. S. Kokrady, Ph.D. Dissertation, The Ohio State University (1972); Diss. Abstr. Int. B, 3_4' 597-B (1973).

138. R. A. Markovs, Ph.D. Dissertation, The Ohio State University (1975); Diss. Abstr. Int. B, 3_6' 3946-B (1976).

139. D. C. Baker, A. Ducruix, D. Horton, and C. Pascard- Billy, Chem. Commun., 729-732 (1974); A. Ducruix and C. Pascard-Billy, Acta Crystallogr., Ser. B, 31, 2250-2256 (1975).

140. A. Ducruix and C. Pascard-Billy, Acta Crystallogr., Ser. B, 33, 2501-2503 (1977).

141. I. Chibata, K. Okumura, S. Taneyama, and K. Kotera, Experientia, 25, 1237-1238 (1969)."

142. K. Okumura, T. Oine, Y. Yamada, M. Tonie, T. Adachi, T. Hagura, M. Kawaza, T. Mizoguchi, and I. Inoue, J. Org. Chem., 36_, 1573-1579 (1971).

143. T. Kamiya, Y. Saito, M. Hashimoto, and H. Seki, J. Heterocycl. Chem., 9_, 359-362 (1972).

144. Y. Saito, M. Hashimoto, H. Seki, and T. Kamiya, Tetrahedron Lett., .56, 4863-4866 (1970); F. Tokita, N. Shibukawa, T. Yasumoto, and T. Kaneda, J. Jpn. Soc. Food Nutr., 2A_, 92-95 (1971).

145. S. Tokuda, A. Tagiri, E. Kano, and T. Kaneda, J. Jpn. Soc. Food Nutr., 2_4, 477-480 (1971); S. Tokuda, E. Kane, and T. Kaneda, ibid., 25, 608-613 (1972).

146. L. Okumura, K. Matsumoto, M. Fukamizu, H. Yasao, Y. Taguchi, Y. Sugihara, I. Inoue, M. Seto, S. Saito, K. Takashima, and S. Takeyama, J. Med. Chem., 17, 846-855 (1974). 212 147. A. Holy, Collect. Czech. Chem. Comm., 4_0, 187-214 (1975); A. Holy and G. S. Iranova, Nucleic Acid Res., 1, 19-34 (1974).

148. A. Holy, Proc. Int. Conf. Ribonucleic Acids and Their Components, Poznan, Poland, 1976, p. 134; Proc. Int. Conf. Recent Developments in Oligonucleotide Syn. and Chem. of Minor Bases of tRNA, Poznan, Poland, 1974, p. 223.

149. N. Ueda, T. Kawabata, and K. Takemoto, J. Heterocycl. Chem., 8 , 827-829 (1971) .

150. Stichting REGA, Belg. Pat. 871,366; Chem. Abstr., 91, 117516t (1979).

151. J. J. Brink and G. A. LePage, Cancer Res., 2^4, 1042 (1964) .

152. D. C. Baker and S. R. Putt, J. Am Chem. Soc., 101, 6127-6128 (1979).

153. H. J. Schaffer and C. F. Schwender, J. Med. Chem., 17, 6-8 (1974).

154. T. Kamiya, Y. Saito, M. Hashimoto, and H. Seki, Tetrahedron, 2J3, 899-906 (1972).

155. H. J. Schaffer and R. Vince, J. Med. Chem., 10, 689-691 (1967); H. J. Schaffer, R. N. Johnson, M. A. Schwarz, and C. F. Schwender, J. Med. Chem., 15, 456-458 (1972).

156. R. P. Agarwal, S. Cha, G. W. Crabtree, and R. E. Parks, Jr., in "Symp. Chem. Biol, of Nucleosides and Nucleotides", R. K. Robins and R. E. Harmon, Eds., Academic Press, New York, 1978, pp. 159-197.

157. R. P. Agarwal and R. E. Parks, Jr., Biochem. Pharmacol., 26, 663-666 (1977).

158. T. W. North and S. S. Cohen, Proc. Natl. Acad. Sci., 75, 4684-4688 (1978).

159. W. Plunkett and S. S. Cohen, Cancer Res., 3_5, 1547- 1554 (1975).

160. H. J. Schaffer, L. Beauchamp, P. deMiranda, G. B. Elion, D. J. Bauer, and P. Collins, Nature, 272, 583-585 (1978). 213

161. G. B. Elion, P. A. Furman, J. A. Fyfe, P. deMiranda, L. Beauchamp, and H. J. Schaffer, Proc. Natl. Acad. Sci., 74, 5716-5720 (1977).

162. J. A. Fyfe, P. M. Keller, P. A. Furman, R. L. Miller, and G. B. Elion, J. Biol. Chem., 253, 8.721-8727 (1978).

163. M. L. Wolfrom and A. Thompson, Methods Carbohydr. Chem., 2, 427-430 (1963).

164. E. Fischer, Ber., 2_7, 673-679 (1894).

165. N. K. Richtmyer, Methods Carbohydr. Chem., 1_, 107-113 (1962).

166. B. Coxon and L. Hough, Carbohydr. Res., 8 , 379-397 (1968).

167. L. F. Wiggins, Methods Carbohydr. Chem., 2, 188-189 (1963).

168. G. Rembarz, Ber., 9_3, 622-625 (1960).

169. N. W. Pirie, Biochem. J. , 3_0, 374-376 (1936).

170. D. J. Williams and J. K. N. Jones, Can. J. Chem., 44, 412-415 (1966).

171. F. Weygand, H. Ziemann, and H. J. Bestmann, Chem. Ber., 91, 2534-2537 (1958).

172. B. Gauthier, Ann. Pharm. Fr., 1_2, 281-285 (1954).

173. M. L. Wolfrom, J. Am. Chem. Soc., 5_1, 2188-2193 (1929).

174. P. Brigl and H. Muehlschlegel, Ber., 6_3, 1551-1557 (1930); W. Schneider, J. Sepp, and O. Stiehler, ibid., 51, 220-234 (1918).

175. K. Blieszner, Ph.D. Dissertation, The Ohio State University, (1978). p. 127; Diss. Abstr. Int. B. 3 9 , 4888-B (1979).

176. W. Schneider and J. Sepp, Ber., 4_9, 2054-2057 (1916).

177. J. W. Green and E. Pacsu, J. Am. Chem. Soc., 59, 1205-1210 (1937).

178. C. Pedersen and H. G. Fletcher, Jr., J. Am. Chem. Soc., 82, 5210-5211 (1960). 214

179. Y. Ishido, personal communication.

180. L. B. Townsend, R. K. Robins, R. N. Loeppky, and N. J. Leonard, J. Am. Chem. Soc., £H5, 5320-5325 (1964).

181. (a) J. A. Montgomery and C. Temple, Jr., J. Am. Chem. Soc., 79, 5238-5242 (1957); (b) ibid., 83, 630-635 (1961).

182. R. N. Prasad and R. K. Robins, J. Am. Chem. Soc. , 79, 6401-6407 (1957).

183. R. K. Borins and H. H. Lin, J. Am. Chem. Soc., 79, 490-494 (1957).

184. J. A. Montgomery, T. P. Johnson, A. Gallagher, C. R. Stringfellow and F. M. Schabel, Jr., J. Med. Pharm. Chem., 3, 265-288 (1961).

185. R. J. Rousseau, R. P. Panzica, S. M. Reddick, R. K. Robins, and L. B. Townsned, J. Org. Chem., 3_5, 631- 635 (1970).

186. E. J. Reist, A. Benitez, L. Goodman, B. R. Baker, and W. W. Lee, J. Org. Chem., _27, 3274-3279 (1962).

187. C. P. Bryant and R. E. Harmon, J. Med. Chem., 10, 104-106 (1967).

188. P. Fischer, G. Losch, and R. R. Schmidt, Tetrahedron Lett., r7, 1505-1508 (1978).

189. R. J. Pugmire and D. M. Grant, J. Am. Chem. Soc. , 90, 697-706, 4232-4238 (1968).

190. R. J. Pugmire, D. M. Grant, L. B. Townsned, and R. K. Robins, J. Am. Chem. Soc., 95^, 2791-2796 (1973).

191. H. El Khadem and Z. M. El-Shafei, Tetrahedron Lett., 27, 1887-1889 (1963).

192. H. S. El Khadem, Carbohydr. Res., _59, 11-18 (1977).

193. H. S. El Khadem, in "Synthetic Methods for Carbo hydrates", H. S. El Khadem ed., ACS Symposium Series, 39, 1977, pp. 77-89.

194. P. L. Durette and D. Horton, Adv. Carbohydr. Chem. Biochem., 26, 49-125 (1971). 215

195. P. L. Durette, D. Horton, and J. D. Wander, Adv. Chem. Ser., 117, 147-176 (1973).

196. D. Horton and J. D. Wander, J. Org. Chem., 39^, 1859- 1863 (1974).

197. H. S. El Khadem, D. Horton, and T. F. Page, Jr., J. Org. Chem., 33, 734-740 (1968).

198. M. Blanc-Muesser, J. Defaye, and D. Horton, in press.

199. G. A. Jeffrey and H. S. Kim, Carbohydr. Res., 14, 207-216 (1970).

200. G. W. Schnarr, D. M. Vyas, and W. A. Szarek, J. Chem. Soc., Perkin I, 496-503 (1978).

201. D. Horton and R. A. Markovs, Carbohydr. Res., 80, 263-275 (1980).

202. D. Horton and J. D. Wander, Carbohydr. Res., 10, 279-288 (1969); ibid., 13, 33-47 (1970); J. Defaye, D. Gagnaire, D. Horton, and M. Muesser, Carbohydr. Res., 2_1, 207-416 (1972).

203. R. J. Pugmire, D. M. Grant, R. K. Robins, and G. W. Rhodes, J. Am. Chem. Soc., 8J7, 2225-2228 (1965).

204. R. J. Pugmire and D. M. Grant, J. Am. Chem. Soc., 9J3, 1880-1887 (1971).

205. A. J. Jones, D. M. Grant, M. W. Winkley, and Robins, J. Am. Chem. Soc., 92, 4079-4087 (1970).

206. D. C. DeJongh, J. Am. Chem. Soc., 8j6 , 3149-3151 (1964).

207. K. Biemann, D. C. DeJongh, and H. K. Schnoes, J. Am. Chem. Soc., 85, 1763-1771 (1963).

208. D. C. DeJongh and K. Biemann, J. Am. Chem. Soc., 85, 2289-2295 (1963).

209. S. J. Shaw, D. M. Desiderio, Y. Tsuboyama, and J. A. McCloskey, J. Am. Chem. Soc., 92, 2510-2522 (1970).

210. N. K. Kochetkov and 0. S. Chizhov, Adv. Carbohydr. Chem. Biochem. , 23^, 39-93 (1966).

211. S. Hanessian, Methods Biochem. Anal., 19, 105 (1971). 216

PART TWO:

1. M. L. Wolfrom, W. von Bebenburg, R. Pagnucco, and P. McWain, J. Org. Chem., 30, 2732-2735 (1965).

2. Unpublished result.

3. J. Defaye, M. Naumberg, and T. Reyners, J. Heterocycl. Chem., 6 , 229-234 (1969).

4. J. Defaye and T. Reyners, Bull. Soc. Chem. Biol., 50, 1625-1635 (1968).

5. G. Giovannietti, L. Nobile, M. Amorosa, and J. Defaye, Carbohydr. Res., Zl, 320-325 (1972).

6 . J. Defaye and Z. Machon, Carbohydr. Res., 24^ 235-245 (1972).

7. V. Zecchi, I. Garuti, G. Giovannietti, L. Rodriguez, M. Amorosa, and J. Defaye, Bull. Soc. Chim. Fr., 1389-1394 (1974).

8 . H. Zinner, H. Brandhoff, and H. Kristen, Chem. Ber., 92, 1618-1623 (1959).

9. J. Defaye, Bull. Soc. Chim. Fr., 268602689 (1964).

10. H. Zinner, H. Brandhoff, H. Schmandke, H. Kristen, and R. Haun, Chem. Ber., 9_2, 3151-3155 (1959).

11. J. Defaye and D. Horton, Carbohydr. Res., 14, 128-132 (1970).

12. J. Defaye, D. Horton, S. S. Kokrady, and Z. Machon, Carbohydr. Res., 43, 265-280 (1975).

13. N. J. Leonard and K. L. Carraway, J. Heterocycl. Chem., 3, 485-489 (1966).

14. J. Hildesheim, J. Cleophax, S. D. Gero, and R. -D. Guthrie, Tetrahedron Lett., 4_9, 5013-5016 (1967).

15. N. J. Leonard, F. C. Sciavolino, and V. Nair, J. Org. Chem., 33, 3169-3174 (1968).

16. M. Kawazu, T. Kanno, S. Yamamura, T. Mizoguchi, and S. Saito, J. Org. Chem., 38, 2887-2890 (1973). 217 17. N. Takamura, N. Tag, T. Kanno, and M. Kawazu, J. Org. Chem., 38, 2891-2895 (1973).

18. N. Ueda, Y. Nakatani, S. Terada, K. Kondo, and K. Takemoto, Technol. Report, Osaka Univ., 23_, 713-714 (1973).

19. M. Imoto and K. Takemoto, Synthesis, 1, 173-179 (1970).

20. Von G. Greber and H. Schott, Angew. Chem., 82, 82 (1970).

21. K. Takemoto, J. Polym. Sci., Polym. Symp., 2_5, 105-165 (1976).

22. S. Fakatsu, Y. Takeda, and S. Umezawa, Bull. Chem. Soc. Jap., _46, 3165-3168 (1973).

23. K. Kobayashi and W. Pfleiderer, Chem. Ber., 109, 3175- 3183 (1976).

24. C. H. Persoon, Mycologia europaea, 1, 960 (1822).

25. L. Pasteur, "Etudes sur la vinaergre”, Paris, 1868.

26. W. von Knieriem and A. Mayer, Landw. Versuchsstation, 16, 305 (1973).

27. M. W. Beijerinck, Proc. Acad. v. Wetenshapp., Amsterdam. 2, 495 (1900).

28. J. Frateur, Cellule Res. Cytol. Histol. , 5j3, 285-396 (1950).

29. A. J. Kluyver and F. J. G. Leeuw, Tijdschr. Vergelijk. Geneesk., 10, 170-182 (1924).

30. J. DeLey, J. Gen. Microbiol., 21, 352-365 (1959).

31. G. Bertrand, C. R. Acad. Sci. Paris, 126, 846, 984 (1898).

32. G. Bertrand, Ann. Chim. Phys., [8] 3, 181, (1904).

33. G. Bertrand, Compt. Rend., 149, 225-227 (1909).

34. R. M. Hann, R. B. Tilden, and C. S. Hudson, J. Am. Chem. Soc., 6J), 1201-1203 (1938).

35. T. Reichstein, Helv. Chim. Acta, _17, 996-1002 (1934).

36. M. Steiger and T. Reichstein, Helv. Chim. Acta, 18, 790799 (1935). 218 37. K. R. Butlin, Biochem. J. , !30, 1870-1877 (1936).

38. R. B. Tilden, J. Bacteriol., 37, 629-637 (1939).

39. W. D. Maclay, R. M. Hann, and C. S. Hudson, J. Am. Chem. Soc., 6A_, 1606-1609 (1942).

40. L, C. Steward, N. K. Richtmyer, and C. S. Hudson, J. Am. Chem. Soc., 7JL, 3532-3534 (1949).

41. V. Ettel and J. Liebster, Collect. Czech. Chem. Comm., 14, 80-90 (1949).

42. A. J. Kluyver and A. G. J. Boezaardt, Rec. Trav. Chim., 57, 609-615 (1938).

43. E. L. Totton and H. A. Lardy, J. Am. Chem. Soc., 71, 3076-3078 (1949).

44. L. Anderson and H. A. Lardy, J. Am .Chem. Soc., 70, 594-597 (1948).

45. N. K. Richtmyer, L. C. Steward, and C. S. Hudson, J. Am. Chem. Soc., 72^, 4934-4937 (1950).

46. G. N. Bollenback and L. A. Underkofler, J. Am. Chem. Soc., 72, 741-745 (1950).

47. L. C. Steward, N. K. Richtmyer, and C. S. Hudson, J. Am. Chem. Soc., 2206-2210 (1952).

48. J. W. Pratt, N. K. Richtmyer, and C. S. Hudson, J. Am. Chem. Soc., 74, 2210-2214 (1952).

49. T. E. King and V. H. Cheldelin, J. Biol. Chem., 198, 127-133, 135-141 (1952).

50. T. E. King and V. H. Cheldelin, J. Bacteriol., 6 6 , 581-584 (1953).

51. N. K. Richtmyer, Carbohydr. Res., 23, 319-322 (1972).

52. J. K. N. Jones, M. B. Perry, and J. C. Turner, Can. J. Chem., 39, 965-971 (1961).

53. J. K. N. Jones, M. BY-Perry, and J. C. Turner, Can. J. Chem., 39, 2400-2410 (1961).

54. J. K. N. Jones, M. B. Perry, and J. C. Turner, Can. J. Chem., 40, 503-510 (1962). 219 55. L. Hough, J. K. N. Jones, and D. L. Mitchell, Can. J. Chem., 31_, 725-730 (1959).

56. R. T. Williams, J. K. N. Jones, N. J. Dennis, R. J. Ferrier, and W. G. Overend, Can. J. Chem., 43, 955- 959 (1965).

57. W. A. Szarek, G. W. Schnarr, H. C. Jarrell, and J. K. N. Jones, Carbohydr. Res., 53_, 101-108 (1977).

58. J. DeLey and K. Kersters, Bacteriol. Rev., 23^, 164- 180 (1964).

59. K. Kersters and J. DeLey, Biochim. Biophys. Acta, 71, 311-331 (1963).

60. E. I. Fulmer and L. A. Underkofler, Iowa State Coll. J. Sci., 21, 251-270 (1947).

61. D. T. Williams and J. K. N. Jones, Can. J. Chem., 45, 741-744 (1967).

62. R. L. Whistler and L. A. Underkofler, J. Am. Chem. Soc., 60, 2507-2508 (1938).

63. C. L. Hu, E. A. McComb, and V. V. Rendig, Arch. Biochem. Biophys., 110, 350-353 (1965).

64. T. Asai, "Acetic Acid Bacteria", University of Tokyo Press, 1968, Part II.

65. N. L. Edson, Rept. Aust. New Zealand Assoc. Adv. Sci., 29, 281-298 (1953).

6 6 . A. C. Arcus and N. L. Edson, Biochem. J. , 6_4, 385-394 (1956).

67. J. DeLey and R. Dochy, Biochim. Biophys. Acta, 40, 277-289 (1960).

6 8 . K. Kersters, W. A. Wood, and J. DeLey, J. Biol. Chem., 240, 965-974 (1964).

69. J. A. Fewster, Biochem. J., 6j>, 14p (1957).

70. C. Widmer, T. E. King, and V. H. Cheldelin, J. Bacteriol., 71_, 737 (1956).

71. J. T. Cummins, T. E. King, and V. H. Cheldelin, J. Biol. Chem., 224, 323-329 (1957). 220

72. J. T. Cummins, V. H. Cheldelin, and T. E. King, J. Biol. Chem., 226, 301-306 (1957).

73. D. R. D. Shaw and F. L. Bygrave, Biochim. Biophys. Acta, U 3 , 608-610 (1966).

74. E. L. Gaden, Jr., Biotech. Bioeng., £, 99-103 (1962).

75. G. L. Solomons, "Methods and Materials in Fermentation", Academic Press, 1969, Chapt. 1.

76. B. Iselin, J. Biol. Chem., 175, 997-998 (1948).

77. L. Hough, J. K. N. Jones, and D. L. Mitchell, Can. J. Chem., 37, 725-730 (1959).

78. G. W. Schnarr, W. A. Szarek, and J. K. N. Jones, App l . Environ. Microbiol., 3^, 732-734 (1977).

79. B. L. Batzing and G. W. Claus, J. Bacteriol., 113, 1455-1461 (1973).

80. K. R. Butlin, Biochem. J. , 3_2, 508-512 (1938).

81. Z. Fencl, J. Ricica, and J. Simmer, Symp. Int. Congress Microbiol., IX, Moscow, 1966, pp. 159-167.

82. T. Kono and T. Asai, Biotech. Bioeng., JL1, 293-321 (1969).

83. J. L. Schimwell, Antonie van Leeuw. , _26, 169-181 (1961).

84. J. L. Schimwell and J. G. Carr, Antonie van Leeuw., 26, 383-396 (1961).

85. J. Schell and J. DeLey, Antonie van Leeuw., _2£, 445- 465 (1962).

8 6 . T. Leisinger, W. Jagger, P. Weber, and L. Ettlinger, Arch. Mikro., 57, 76-92 (1967).

87. T. Yamada, M. Hisamatsu, and M. Taki, J. Chromatogr., 103, 390-391 (1975).

88. W. Pigman and H. S. Isbell, Adv. Carbohydr. Chem., 23, 11-57 (1968).

89. E. L. Eliel, N. L. Allinger, S. J. Angyal, and G. A. Morris, "Conformational Analysis", Interscience, N. Y., 1965, Chapt. 6. 221

90. W. Funcke and A. Klemer, Justus Liebigs Ann. Chem., 1232-1236 (1975).

91. L. Que, Jr. and G. R. Gray, Biochemistry, _13, 146-153 (1974) .

92. S. J. Angyal and G. S. Bethell, Aust. J. Chem., 29, 1249-1265 (1975).

93. G. R. Gray, Accounts Chem. Res., 9^, 418-424 (1976).

94. S. J. Shaw, D. M. Desiderio, Y. Tsuboyama, and J. A. McCloskey, J. Am. Chem. Soc., 9_2, 2510-2522 (1970).

95. N. K. Kochetkov and 0. S. Chizhov, Adv. Carbohydr. Chem. Biochem., 2JL, 39-93 (1966).

96. F. E. Hruska, A. A. Grey, and I. C. P. Smith, J. Am. Chem. Soc., 92, 214-215 (1970).