Citation
Isolation and characterization of taxanes and other compounds from various species of Taxus

Material Information

Title:
Isolation and characterization of taxanes and other compounds from various species of Taxus
Creator:
Davies, Richard Michael, 1959-
Publication Date:
Language:
English
Physical Description:
xi, 115 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Acetates ( jstor )
Acetylation ( jstor )
Carbon ( jstor )
Chromatography ( jstor )
Elution ( jstor )
Esters ( jstor )
Oxidation ( jstor )
Protons ( jstor )
Signals ( jstor )
Taxoids ( jstor )
Antineoplastic Agents, Phytogenic -- chemistry ( mesh )
Antineoplastic Agents, Phytogenic -- isolation & purification ( mesh )
Chromatography, Liquid -- methods ( mesh )
Department of Medicinal Chemistry thesis Ph.D ( mesh )
Dissertations, Academic -- College of Pharmacy -- Department of Medicinal Chemistry -- UF ( mesh )
Paclitaxel -- chemistry ( mesh )
Paclitaxel -- isolation & purification ( mesh )
Plants, Medicinal ( mesh )
Taxus ( mesh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (leaves 110-114).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Richard M. Davies.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
0029897508 ( ALEPH )
51589711 ( OCLC )

Downloads

This item has the following downloads:


Full Text










ISOLATION AND CHARACTERIZATION OF TAXANES AND OTHER COMPOUNDS
FROM VARIOUS SPECIES OF TAXUS














By

RICHARD M. DAVIES













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

UNIVERSITY OF FLORIDA 1998
































This work is dedicated to the memory of Dr. Koppaka V. Rao, an extraordinary scientist, dedicated teacher and very dear friend. I feel blessed to have known and worked with him.














ACKNOWLEDGMENTS


Much of the work in this dissertation was done with the guidance and expertise of the late Dr. K. V. Rao, who passed away on February 20, 1998. Professor Rao and I had known each other since 1981 and I am grateful to him for encouraging me to return for graduate studies and for his true friendship with me. I would also like to thank his wife and children for their much appreciated support, friendship and encouragement.

I wish to thank Dr. John Perrin for assuming the position of chairman of my supervisory committee and for his kind encouragement and guidance. He has helped me in many ways and I am very grateful to him for his persistence in pushing me to complete this work. I would also like to thank Dr. Margaret James, Dr. Jonathan Eric Enholm, Dr. Kenneth Sloan, and Dr. Stephen Schulman for participating on my supervisory committee and for their thoughtful advice and expertise.

I wish to thank my mother and father for their kind encouragement and love, and also my three sisters, brothers-in-law, niece and nephews. I would like to thank the Graduate School and many other University of Florida personnel for all of the kind assistance they have provided, especially Gladys Jan Kalman and Nancy Rosa for all of their helpful assistance.















TABLE OF CONTENTS


page


ACKNOWLEDGMENTS........................................................................ iii

LIST OFTABLES .............................................................................. vii

LIST OF FIGURES ............................................................................ viii

ABSTRACT...................................................................................... x

CHAPTERS

1 HISTORICAL OVERVIEW OF TAXUS .................................................... 1

Background of Research at the University of Florida .................................... 6
Methods ................................................................................... 6

2 A SELECTED REVIEW OF THE LITERATURE ON TAXUS.......................... 11

Earlier Studies ............................................................................. 11
Studies after the Discovery of Taxol ..................................................... 14
Semi-synthesis of Taxol.............................................. ................ 17
Total Synthesis. ..................................................................... 18
Other Synthetic Approaches ......................................................... 20
General Structural Features of Taxanes................................................. 24
Taxa-4(20); 11 -dienes................................................................. 25
4(20)-Epoxides ...................................................................... _25
Oxetanes .............................................................................. 26
Abeotaxanes........................................................................... 28

3 TAXANE CONSTITUENTS OF THE NEEDLES OF TAXUS BREVIFOLIA ......... 29

Fractionation of the Needles of Taxus brevifolia ...................................... ...29
Brevitaxane A (Brevifoliol) [3-1 ] ......................................................... 31
Hydroxyl Functionalities ..................................................... ............. 33
4/20 Unsaturation ..................................*.........* .......... *.... ..***... .... '36
Number and Nature of the Oxygen Substitution ........................................ 37
Experimental .............................................................................. 44
Extraction of the Needles of Taxus brevifolia ....................................... 44
Reverse Phase Column Chromatography:.......................................... 45
B revifoliol [3-1] ... .... ................. ...... ...... .... ..... ... .....46
Brevifoliol-5-Monoacetate [3-2] ...................................................... 48


iv









Brevifoliol-5,13-Diacetate [3-4] ....................................................................... 49
Brevifoliol-13-Ketone [3-6] ............................................................................. 50
Dihydrobrevifoliol [3-7] .................................................................................. 51
Brevifoliol Epoxide [3-8] ................................................................................ 51
Ozonization of Brevifoliol: Brevifoliol-norketone [3-9] .................................... 52
Brevifoliol-4,20-Diol [3-10] .............................................................................. 53
Saponification of Brevifoliol [3-11] ................................................................. 54
Debenzoyl Brevifolio-Pentaacetate [3-12] ..................................................... 54
Periodate Oxidation of [3-11] to [3-13] .......................................................... 55
Form ation of O sazone [3-14] from [3-13] w ith 2,4-DNPH ............................... 55
Debenzoyl Brevifoliol [3-15] ........................................................................... 56

4 SOME UNUSUAL REACTIONS OF BREVIFOLIOL ............................................. 57

1. Acid-Catalyzed Acetylation ....................................................................... 58
2. Oxidation ................................................................................................. 59
3. Action of BF3 on Brevifoliol [4-3] ............................................................... 61
4. Reaction w ith Iodine/Silver Acetate [4-4] ................................................. 65
Experim ental ........................................................................................................ 69
Brevifoliol Triacetate [3-5] .............................................................................. 69
Oxidation w ith Jones Reagent to [4-1] .......................................................... 69
Action of Boron Trifluoride on Brevifoliol [4-3] ............................................... 70
Reaction with Iodine and Silver Acetate [4-4] .............................. 70
Acetylation of [4-4] to [4-5] ........................................................................... 72
Reaction with N-Bromosuccinimide and Silver Acetate [4-6] ......................... 72
Reaction of [4-4] w ith N-Brom osuccinim ide [4-7] .......................................... 73

5 TAXANE CONSTITUENTS OF THE NEEDLES OF TAXUS X MEDIA ................. 74

Brevifoliol [3-1] ............................................................................................... 79
Taxanes 1 [5-1] and II [5-2] ............................................................................ 79
Taxane III [2-1] ............................................................................................. 80
Taxane IV [2-2] ............................................................................................. 80
Taxol [5-3] ..................................................................................................... 80
Ozonolysis of [2-2] ......................................................................................... 80
Experim ental ........................................................................................................ 81
Extraction: ..................................................................................................... 81
Chrom atography: ............................................................................................... 81
Characterization of the Taxane Components of Taxus x Media Hicksii ............... 82
Brevifoliol [2-1] ............................................................................................... 82
Taxanes I and II [5-1] and [5-2] ..................................................................... 83
Taxane III [2-1] ............................................................................................. 84
Taxane IV [2-2] .................................................... 84
Taxol [5-3] ..................................................................................................... 85
Ozonolysis of Com pound [2-2] ........................................................................ 86

6 TAXANE CONSTITUENTS OF TAXUS FLORIDANA .......................................... 87

T a x iflo rin e ................................................................................................................ 8 9
Experim ental ........................................................................................................ 93
Extraction ..................................................................................................... 93


v









Characterization of the Taxane Constituents of Taxus floridana ......................... 95
10-Deacetyl Baccatin III [2-7] ......................................................................... 95
Brevifoliol [3-1] ............................................................................................... 95
Taxiflorine [6-1] ............................................................................................. 96
Baccatin VI [6-2] ........................................................................................... 96
T a x o l [5 -3 ] ................................................................................................... . 9 8
Acetylation of Taxiflorine to [6-3] ................................................................... 98
Benzoylation of Taxiflorine to [6-4] ................................................................. 98
Saponification and Acetylation of [6-1] to [6-5] ............................................... 99

7 NON-TAXANE COMPONENTS FROM THE BARK AND NEEDLE EXTRACTS .... 100

G e n e ra l ................................................................................................................. 1 0 0
E x p e rim e n ta l ......................................................................................................... 1 0 2
F la v o n o id s ............................................................................................................. 1 0 3
Quercetin Rutoside (Rutin) .............................................................................. 103
Q u e rc e tin ......................................................................................................... 1 0 4
S c ia d o p ity s in ................................................................................................... 1 0 4
P-Sitosterol-p3-D-Glucoside ............................................................................. 105
P-Sitosterol-p3-D-Glucoside Tetra-acetate ....................................................... 105
P S ito s te ro l ...................................................................................................... 1 0 6
Phytoecdysteroids ................................................................................................. 107
Ecdysterone & 23, 3P, 22c-Triacetate ............................................................. 107
Ponasterone A and 23, 3P, 22x Triacetate ...................................................... 107
Phenolic Com pounds ............................................................................................ 108
U s n ic A c id ....................................................................................................... 1 0 8
Betuloside (4-(4'-Hydroxyphenyl)-2R-butanol Glucoside)) & Aglycone ............. 109

LIST OF REFERENCES ............................................................................................. 110

BIOGRAPHICAL SKETCH .......................................................................................... 115






















vi














LIST OF TABLES


Table page

3-1 : Proton NMR Spectra of Brevifoliol and Brevifoliol Acetates.....................33

3-2: Carbon NMR Spectra of Brevifoliol and Brevifoliol Acetates ...................34

4-1 : NMR Spectra of Compound [4-3] from BF3 Reaction..............................63

6-1 : Proton NMR Spectra of Compounds [6-3], [6-4] and [6-5]......................91

































vii














LIST OF FIGURES


Figure page

2-1 Early Studies on the Constituents of some Taxus Species ................ 13

2-2: Taxol and some Synthetic Targets .................................................... 16

2-3: Nicolaou's Retrosynthetic Strategy .................................................... 18

2-4: Nicolaou's Taxane Ring Synthesis .................................................... 21

2-5: Nicolaou's Final Synthetic Intermediates ........................................... 22

2-6: Starting Points of Other Synthetic Strategies .................................... 23

3-1 : Proton NMR Spectrum of Brevifoliol ................................................. 32

3-2: Brevifoliol and Reaction Products ................................ 35

3-3: Brevifoliol Hexaol Reaction Products ................................................. 39

4-1 : O xidation P roducts ............................................................................ 6 1

4-2: BF3-etherate Catalyzed Elimination Product .................................... 64

4-3: DEPT Spectra of BF3 Elimination Product [4-3] .................................. 64

4-4: Iodine/Silver Acetate Product [4-4] and Acetate .................................. 66

4-5: H,H-COSY Spectrum of [4-4] ............................................................ 67

4-6: HETCOR Spectrum of [4-4] ............................................................... 68

5-1 : Fractionation of the Extract of Taxus x media Hicksii Needles ........... 77

5-2: HPLC Trace of Taxanes Coeluting with Taxol .................................... 78

5-3: Progress of Elution of Taxanes from Reverse Phase Column ............ 78

6-1 : Taxanes and Analogues from Taxus x media Hicksii ........................ 92


viii








6-2 : Carbon NMR Spectrum of Baccatin VI ............................................... 97



















































ix














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

ISOLATION AND CHARACTERIZATION OF TAXANES AND OTHER COMPOUNDS FROM VARIOUS SPECIES OF TAXUS by

Richard M. Davies

December 1998

Chairman: Koppaka V. Rao
Cochairman: John H. Perrin
Major Department: Medicinal Chemistry

Taxol is a promising antineoplastic agent originally reported in 1971 by Wani and Wall, isolated from the bark of the Pacific yew (Taxus brevifolia). Intensive research in the last decade has demonstrated that this drug possesses exceptional activity in the treatment of many difficult types of cancer.

From the beginning taxol has proven to be a difficult compound to obtain, with very low yields and a highly complex structure with many chiral centers and sensitive moieties. Originally obtained from the bark of a very slow growing tree, the possibility of growing various Taxus (yew) species under hydroponic conditions has been investigated in this project.

One local variety, known as Taxus floridana (Florida yew) was found to grow well and produce taxol and other useful taxanes. During initial investigations a simple and elegant method for the isolation of taxol using reverse phase bonded silica was developed. Generous funding by the University of Florida Division of Sponsored





x









Research made possible the construction of a pilot plant scale facility where these isolation methods were successfully implemented.

Excellent yields and the isolation of many related taxanes have proven that this method is superior to currently approved processes used in the production of taxol. The failure of other researchers to employ bonded silica gel for preparative columns in the past may reflect experiences with analytical columns, but this method has proven to be quite exceptional and should be employed extensively.

This dissertation covers many crystalline and non-crystalline compounds isolated and characterized as a result of this project. Some results from the application of this technique for the isolation of taxanes from the needles of Taxus brevifolia, Taxus x media cultivar Hicksii, and T. floridana are presented. Similar experiments on the bark and wood of T. brevifolia are also described.































xi














CHAPTER 1
HISTORICAL OVERVIEW OF TAXUS


During the late 1950s, the National Cancer Institute initiated a program with the objective of discovering compounds from natural sources, which might prove useful in the treatment of various human cancers. In this program, plant samples from various parts of the world were collected and brought to participating laboratories, where the active principles were isolated, chemically characterized, and subjected to testing in various murine tumors.

It is in such a context that a sample of the bark of the Pacific yew (Taxus brevifolia, Nutt.) was extracted and the active principle called taxol* was reported by Wani and coworkers in 1971 (Wani et al. 1971). Although it exhibited potent cytotoxicity in some tumor assays, many unsuccessful lead compounds also are cytotoxic in these assays. Taxol did not appear exceptional, and the problems of low yield and poor solubility discouraged the pursuit of further research for many years. During the late 1970s, activity against B-16 melanoma in mice, and several human tumors grown in athymic mice was recognized (e.g. MX-1 human mammary xenograft). These activities rekindled interest in taxol as a candidate for cancer treatment, resulting in further studies and human clinical trials. In the ensuing years the pace of research into taxol and taxoids has increased dramatically.



*Taxol is a registered trademark since 1993 by Bristol Myers Squibb Co., with paclitaxel being the generic name. 'Taxol' will be used in this dissertation as the generic name, as this work was started before this change.



1






2


A study of its mode of action revealed that it blocked cell division at the cell cycle through its specific action on the G2/M phase of the tubulin/ microtubule system. Unlike other antitumor drugs such as colchicine, vincristine and vinblastine, which act as tubulin poisons, taxol exhibited a novel mode of action (Schiff et al. 1979). Microtubules are involved in the formation of the mitotic spindle fibers necessary for the replication of DNA and are also integral building blocks within the cell wall. They are generated from a protein known as tubulin, and a dynamic equilibrium exists between tubulin and microtubules in vivo. In the presence of taxol, the polymerization of tubulin produces what are now known as oligo-microtubules. In contrast to the usual microtubules, which can be readily disassembled, these oligo-microtubules resist disassembly to tubulin, thereby preventing cell division (Horwitz, 1992).

Based on potent activity against important experimental tumors and its unique mode of action, interest in taxol became greatly enhanced, and it was approved for human Phase I clinical trials in the early 1980s. Taxol showed significant activity in human tumors in Phase I and Phase II clinical trials, especially in ovarian and breast carcinomas (McGuire et al. 1989; Holmes et al. 1991). The scientific community took a special interest in taxol at that time due to the lack of adequate treatment options available for ovarian cancer.

This knowledge established taxol as an important antitumor drug and stimulated a renewed interest in it. Intensive worldwide studies have reached explosive proportions since 1994 concerning its production, chemistry, biochemistry and many other aspects.

At that point, two problems needed solution before taxol could become a viable alternative as a useful treatment against any type of cancer. First, the lipophilic nature of taxol made it difficult to develop an acceptable dosage form for this drug. Gradually, this was overcome by the introduction of a suitable, "relatively" non-toxic dosage form.






3


Interestingly, the poor solubility characteristics of taxol might prove to be responsible for new discoveries regarding a problem of cross resistance to different classes of chemotherapeutic agents, caused by non-specific drug efflux and referred to as the MDR phenotype. The MDR phenotype is a gene that has been linked to multiple drug resistance (hence MDR); and some studies indicate that the solvent used for the delivery of taxol might have good activity against this common cause for therapeutic failure in the treatment of cancer (Woodcock et al. 1990; Webster et al. 1993; Fjallskog et al. 1993). It is known that this effect can result from the expression of plasmamembrane transport proteins (P-glycoproteins) which can enhance the efflux of structurally unrelated compounds from the cancer cell. At least three reports suggest that the solvent Cremaphore LH might enhance the antitumor actions of taxol when the tumor(s) display the MDR phenotype, and further work with cremaphores alone and in combination with other antitumor agents is needed to clarify this seemingly serendipitous finding.

Cremaphore LH is a form of ethoxylated castor oil and is responsible for many adverse drug reactions during the administration of taxol, and pretreatment with corticosteroids and antihistamines if often required to prevent allergic response up to and including anaphylaxis and death. Perhaps more difficult than this solubility concern was the procurement of an adequate supply of taxol for clinical trials and the anticipated needs for subsequent worldwide clinical use. Reported yields of taxol from the dried bark of T. brevifolia were averaging around 0.01%.

A large-scale process for the isolation of taxol was developed by Polysciences, Inc. (Paul Valley Industrial Park, Warrington, PA 18976); with yields of 0.005-0.01% (Boettner et al. 1979). Under these conditions, one kilogram of the bark could be expected to provide only 50-100 mg of taxol at best (or 30,000 lbs. being required for obtaining one Kg. of taxol). Approximately 2 grams of taxol are needed for one complete






4


course of therapy for a patient and this translates into requiring the bark from five to ten trees based on such reported yields.

Only a few studies on the taxol content of other species of Taxus were published from the time its importance was recognized in 1980 until 1992. From the bark of T. wallichiana were isolated taxol and the closely related cephalomannine, as well as other taxoids (Miller et a!. 1981; Miller, 1980). From the bark of T. baccata L., Senilh et al. isolated nearly 20 different taxoids, including taxol, cephalomannine and a series of xyloside derivatives of these (Senilh et al. 1984). The taxoid content of the needles of Taxus baccata was studied and 10-deacetyl baccatin III isolated in relatively high yields (Chauviere et al. 1981). The fractions from the large scale (Polysciences) process from the bark of T. brevifolia were also investigated to recover any other taxoids with useful activity, or with possible semisynthetic utility such as conversion to taxol. However, only minute yields of 10-deacetylbaccatin Ill, 7-epitaxol and 10-deacetyl-10-oxotaxol were reported in these studies (Huang et al. 1986; Kingston et al. 1982); leaving a strong impression that the bark of T. brevifolia is a poor source for not only taxol, but also for any other useful analogues of taxol.

In spite of these problems, the bark of T. brevifolia has been accepted as the primary source for taxol until recently. However, since at the expected demand for taxol and the yields that can be realized from the bark, the yew tree population would be depleted in a few years, the use of the bark must stop. Among the alternatives that were being, considered to avoid this prospect are the following: 1) the use of needles, which are a renewable source, 2) growing the plant in tissue culture, 3) semi-synthesis from appropriate naturally occurring taxoids and 4) total synthesis. Progress has been made on all of these fronts.

As far as the needles are concerned, the most important candidate selected for direct isolation of taxol is the ornamental yew (Taxus x media cultivar Hicksii). Other






5


than analytical HPLC studies on the taxol content of the needles under various conditions, no practical methodology for the isolation of taxol or other taxoids has been published. Publications from this laboratory which address these issues are essentially the only work available in the literature (Rao et al. 1995; Rao et al. 1996). In addition to direct isolation of taxol, the needles were also examined for the presence of analogues such as 10-deacetyl baccatin Il, since semi-synthesis from such is already an important alternative. The two most important species, T. baccata L. and T. wallichiana Zucc., have become the focus of attention since they were demonstrated to contain the highest concentrations of 10-deacetyl baccatin Ill.

Growing of various tissues of T. brevifolia in plant cell culture has been under development since 1990 and the methods have been standardized in many laboratories. However, the yields, as yet, have not been very attractive. Further research is expected to overcome this problem. Work on this alternative will continue due to the attractiveness of this approach and its potential for large-scale operations.

Starting with 10-deacetyl baccatin Ill, considerable progress was made in the area of semi-synthesis. In the first recorded semi-synthesis of taxol, the 13-cinnamate ester of 7-protected baccatin III was converted to the phenyl isoserine ester through a Sharpless hydroxy-amination (Denis et al. 1988). At this point, as an alternative to benzoylation of the amino group that will yield taxol, a t-BOC group (tertbutoxycarbonyl) was introduced, along with leaving the 10-hydroxyl free, to obtain an analogue known as taxotere. On the basis of its activity, taxotere has also been approved as an antitumor drug. Two important schemes for preparing taxol from 10deacetyl baccatin III have been well developed and used for the large scale semisynthesis of taxol as discussed in Chapter 2 (Denis et al. 1994; Ojima et al. 1991; Holton et al. 1992).






6


Two total syntheses of taxol have been recorded, and several other approaches towards the synthesis have also been reported in the literature as discussed briefly in Chapter 2 (Nicolaou et a/. 1994a, 1994b, 1995a, 1995b, 1995c; Holton et a/. 1994a, 1994b). Although these methods demonstrate remarkable achievements in the field of synthetic organic chemistry, they do not offer a practical method for the large-scale production of taxol or its analogues at this time.

Background of Research at the University of Florida


As part of one of these alternative quests, the National Cancer Institute hoped that instead of using the bark of the Pacific yew, the plant should be grown under hydroponic conditions, and they wanted to know whether plants grown in this manner would produce enough taxol for isolation. This laboratory was approached with this idea in early 1990, and with collaboration from Prof. George Hochmuth, Jr., of WFAS, University of Florida, the project was started. More on this aspect will be discussed in Chapter 6. In order to learn the current knowledge concerning the analysis and isolation of taxol, the pertinent literature was consulted. This yielded only a few papers on the isolation of taxol and taxoids, which were outlined above.

Methods


In general, the methods were found to be too cumbersome for others to repeat. For example, one of these publications in which isolation of taxol and its analogues was described from T. wallichiana, used the following steps, starting with the concentrated ethanolic extract of the plant:

1 Partition between water and hexane
2. Extraction of the aqueous phase with chloroform
3. Silica gel chromatography on the chloroform extract
4. A second silica gel chromatography






7


5. Counter-current distribution
6. HPLC on the appropriate fractions
7. A second HPLC on the appropriate fractions

Similarly, the large-scale process developed for the isolation of taxol by Polysciences Inc. from the bark of T. brevifolia consisted of the following steps, again starting with the alcoholic extract concentrate (Boettner et al. 1979).

1. Solvent partition water and CH2CL2, concentration to a solid

2. Separation of the extract solid into soluble and insoluble fractions
3. Chromatography on the soluble fraction
4. Recovery of taxol and crystallization twice
5. Silica chromatography on the taxol/ cephalomannine mixture
6. Recovery and crystallization of taxol

Thus, it appeared that, although procurement of taxol was of top priority, and many alternative approaches were attempted for solving this problem, one alternative, which was not considered, was to study the existing isolation procedure itself to make it more efficient. Thus the approach pursued at the University of Florida during 1990-91 was to develop a simpler process for the isolation. Over the next few months, a new process was developed based on the use of a single reverse-phase chromatographic column, and consisting of the following steps, starting with the alcoholic extract concentrate (Rao, 1993).

1. Partition between water and chloroform, and concentration
2. Reverse phase column chromatography on the extract directly
3. Harvesting the crystals and recrystallization.

The total chloroform extract of the bark of T. brevifolia, was applied directly to the C18-bonded silica column in 25% acetonitrile in water (i.e., no separation into soluble and insoluble fractions); and the column developed with a step gradient (30-60% acetonitrile). The column fractions were let stand for 3-7 days, whereby taxol and seven






8


of its analogues crystallized out directly from the fractions. These are filtered and purified further by recrystallization, or subjected to a small column.

This process using reverse phase column chromatography, gave not only higher yields of taxol (0.02-0.04% vs. 0.01%) on a pilot plant scale, but also made possible the simultaneous isolation of a number of analogues which have not been obtained from this plant before. These included a 10-deacetyl baccatin 111 (0.02%); and a number of xyloside analogues, chief among which being the 10-deacetyltaxol-7-xyloside, which can now be isolated in yields of 0. 1% or higher.

Based on the successful fractionation of the bark extract of T. brevifolia, this technique was then ready for application to the other extracts such as the needles and wood of T. brevifolia, and to the needles of two other species of Taxus. These applications which gave practical methodology for processing these various extracts, also yielded many interesting taxoid compounds and these experiments are all detailed in this dissertation.

Although this work was started during 1991 and much of the expected work was completed by late 1993, the world-wide interest in taxol research made a "quantum leap" at about this time, with a phenomenal increase in publications dealing with all aspects of taxol chemistry. Some of the compounds which were isolated for the first time in this laboratory, and whose structures were determined, were rediscovered by others and published. In spite of the enormous increase in the number of relevant publications, most of these publications described the isolation of the minimum possible amounts of the compounds, often as amorphous solids. Many determined their structures only through NMVR spectral interpretation, with little or no other physical characterizations, elemental analyses, derivatizations or reactions. In at least a few examples, the assigned structures were found to be wrong and were subsequently corrected once or






9


even twice. In the present work, practical isolation methods were used to obtain gram quantities of many compounds, as crystalline solids, where possible.

The compounds are usually characterized by physical and spectral properties, providing elemental analyses, and carrying out derivatization such as acetylation, oxidation, etc. The structures were elucidated through chemical reactions as well as through spectral data. Thus, even though some of the final structures may have been published, the work described here contains experiments that have not been carried out by these authors.

Brief descriptions of the topics that appear in this dissertation are given below.

Chapter 2 gives a brief and selected summary of the pertinent literature on taxol and taxoids, covering the areas of isolation, elucidation of structures, semi-syntheses and total syntheses. Because the subject matter expanded enormously since 1993, the scope of the review is limited to material that is relevant to the subject matter of the dissertation.

Chapter 3 deals with the taxoid composition of the needles of Taxus brevifolia. It covers the application of the reverse phase column chromatography to the needle extract, isolation of the major taxoid, brevitaxane A (or brevifoliol), along with brevitaxane B, and taxol. It continues with the elucidation of the structure of brevitaxane A by various reactions, as well as by a detailed analysis of the NMR spectral evidence.

Chapter 4 discusses some unusual reactions of brevifoliol. Such reactions have not been reported with this or any other taxoid compounds. In each case, crystalline compounds were obtained and characterized by physical and spectral data.

Chapter 5 deals with fractionation of the extract of the needles of Taxus x media cv. Hicksii by reverse phase column chromatography and isolation of taxol and several other taxoids and their characterization. In spite of the fact that this species (ornamental yew) was declared as the preferred plant for the future isolation of taxol, no publications





10


describing a suitable scheme for isolation of taxol or any other taxoids have appeared so far, other than analytical hplc data on their taxol content.

Chapter 6 is similarly devoted to the fractionation of the extract of the needles of Taxus floridana Nutt. by reverse phase column chromatography. Isolation of taxol, 10deacetyl baccatin Ill, baccatin VI and a new crystalline taxoid compound named taxiflorine, with its structural elucidation are described.

Chapter 7 deals with the isolation of several crystalline non-taxane compounds present in the extracts of the bark and needles of Taxus brevifolia. These were shown to include flavonoids, phenols, and other types of compounds.















CHAPTER 2
A SELECTED REVIEW OF THE LITERATURE ON TAXUS


An overview of the taxol story has been presented in Chapter 1. In this Chapter, a selected review of the literature will be presented on taxol as well as the other taxanes, which are included in this dissertation. Two comprehensive reviews have been published on the subject of taxol, one by Kingston et al. (1993) and one by Miller (1980).

The present review on the genus Taxus may be roughly divided into two parts: studies before and studies after the discovery of taxol.

Earlier Studies


The genus Taxus (N.O. Taxaceae) represents a group of plants (common name, yew) which grow mostly in temperate climates and can be found distributed throughout the world. They are generally slow-growing evergreen trees or shrubs with stiff linear leaves (or needles); and fruits which are small, fleshy and bright red. The common names of the plants are qualified by the place of its origin, as for example, Pacific or western yew (T. brevifolia, Nutt.); European or English yew (T. baccata, Pilg.); Canadian yew (T. canadensis, Willd.); Japanese yew (T. cuspidata, Sieb. et Zucc.); Chinese yew (T. chinensis, Pilg.); Himalayan yew (T. wallichiana, Zucc.); ornamental yew (T. x media "Hicksii", Rehd.) and Florida yew (T. floridana, Nutt.).

The toxic nature of this genus has been recognized for thousands of years, and in modern times, was first investigated chemically using the needles of Taxus baccata (Lucas, 1856). An amorphous mixture of alkaloids was isolated after extraction under acidic conditions and was given the name of "taxine." Further studies on taxine spanned



11






12


many decades and covered many reactions relevant to taxane chemistry. In one such study, Winterstein and Guyer (1923) were the first to show the presence of 3dimethylamino-3-phenylpropanoic acid in the hydrolyzate of taxine and this acid later became known as "Winterstein's acid."

Until the 1960s most of the work on taxanes focused on these acid-extractable alkaloidal substances, which were readily separable from the large quantities of neutral, resinous materials which dominate the extract. Two groups of researchers were able to convert these somewhat unstable alkaloidal mixtures into more stable, non-basic substances in which the 3-dimethylamino-phenylpropanoic ester unit was transformed into a cinnamate ester. This, as well as the development of chromatographic techniques, made it possible to obtain pure compounds rather than mixtures.

Baxter et al. (1958) in England investigated the major cinnamate ester obtained from T. baccata, which they named 5-O-cinnamoyl taxicin-I triacetate [2-1]. Similarly, a Japanese team (Nakanishi & Kurono, 1963; Kurono et al. 1963) studied a cinnamate ester from T. cuspidata, and called it 5-O-cinnamoyl taxicin II triacetate [2-2] and the structures of both these can be seen in Figure 2-1. These two compounds differ only at C-1, where taxicin II lacks the tertiary hydroxyl found in taxicin I. The IUPAC numbering system for taxanes used throughout this dissertation can also be seen [2-3].

A few years earlier, Graf & Betholdt (1957) succeeded in isolating the purified basic alkaloids, taxine A and taxine B from the original taxine mixture. Taxine B was shown to have the structure [2-4] (see Figure 2-1); which corresponded with 5-0cinnamoyl taxicin I triacetate, into which it could be converted via elimination of the dimethylamine moiety.





13




AcO OAc 18 10 919
ii 7
O 12 16 6
0 3' 13 15 8
,A 1 15 ""17 8 4 5
"O 'C65 14 14
R 2 2 20
AcO



[2-1] O-Cinnamoyl Taxicin I Triacetate, R = OH [2-3]-IUPAC [2-2] O-Cinnamoyl Taxicin II Triacetate, R = H Numbering



AcO OH

0 O N

HO
CH CPP


[2-4] Taxine B [2-5] Geranylgeranyl
pyrophosphate



AcO 0 OH R OC


HOOH HO

OAc OCOC6H5 C6H5 0



[2-6a] Proposed Glycol, [2-6b] R = OAc, Baccatin II
later corrected to 6 b [2-7] R = OH 10-DAB-Ill



Figure 2-1 : Early Studies on the Constituents of some Taxus Species






14


Harrison (Harrison & Lythgoe 1966; Harrison et al. 1966) published one of the earliest biogenetic theories for the formation of taxanes starting with geranylgeranyl pyrophosphate and electrophilic cyclization [2-5]. Efforts by many groups to utilize a similar scheme to synthesize the taxane skeleton have been unsuccessful thus far (Kumagai et al. 1981; Hitchcock & Pattenden, 1992). Biogenetic pathways often provide ideas for simplified approaches in the synthesis of natural products.

Many early studies utilized acidic conditions for the extraction which might have hampered the isolation of neutral or acid-labile compounds, Kondo & Takahishi (1925) obtained a non-basic compound from the Japanese yew by using neutral conditions. The cinnamates can also be directly isolated from the plant, indicating that they occur naturally and also as artifacts of processing.

The National Cancer Institute (NCI) and the U.S. Department of Agriculture (USDA) joined forces in 1960 to collect and screen plants for activity in several animal tumor models. Arthur Barclay of the USDA collection team obtained samples from the Pacific yew tree (Taxus brevifolia Nutt., family Taxaceae) from Washington State in 1962. In 1964, the extracts from the bark and stems were found to be active against KB cells in vitro (Wani et al. 1971).

Studies after the Discovery of Taxol


Dr. Monroe Wall had discovered another antitumor agent known as camptothecin using the activity on KB cells for isolation and was interested in any other extracts showing this activity. Thus, work on T. brevifolia by Wani and Wall at the Research Triangle Institute was started and led to the isolation of 500 mg of taxol 2 years later in 1966. Cytotoxic actions (Wani et al. 1971) in KB cells, P388 leukemia, Walker 256 carcinosarcoma and P-1 534 leukemia were present in the extracts from the bark. These





15


assays were all used at various points to monitor the fractionation, resulting in the isolation of taxol as the active principle.

In the 1960s the most straightforward and reliable method for the determination of complex chemical structures was x-ray diffraction analysis of a suitable crystal, also known as crystallography. Taxol crystallizes as thin needles not suitable for x-ray studies, but a tetraol derivative was amenable to x-ray studies. The structure of taxol [28] was determined by methanolysis (Figure 2-2); which yielded two compounds: the methyl ester of N-benzoyl phenylisoserine and an alcohol component shown to be a taxane tetraol [2-10a]. This tetraol skeleton was converted into a 7,10-bis-iodoacetate derivative and, unlike all of the taxanes studied earlier, taxol showed the following unique features:

1. A taxane skeleton with an oxetane ring system involving C-4, C-5 & C-20
2. An ester side chain consisting of N-benzoyl phenylisoserine at C-13
3. A carbonyl function at C-9

Baccatin Ill (Figure 2-1) and its 7c -epimer, baccatin V (Figure 2-2); were shown to be similar to taxol, having the oxetane ring and the C-9 carbonyl function. These epimers yielded better crystals and x-ray crystallography was performed. Baccatin III [26b] lacked the ester side chain present in taxol. Still another analogue, known as 10deacetylbaccatin III (10-DAB, [2-7]) was later found to be much more widely distributed in Taxus spp., especially in the needles of Taxus baccata. This became an important taxane because it could be converted into baccatin III and later to taxol by the reattachment of the N-benzoyl phenylisoserine side chain at C-13 (See Figure 2-2).

Taxol showed significant antitumor activity against a variety of in vivo murine tumors including B-16 melanoma and several human xenografts, which qualified it for





16


0




H 0

OHAc

OBz
[2-8] Taxol, R, = Ac R2 =C61-5

[2-9] Taxotere, Rl= H R2 =OC(CH3)3



RI0 0 R2 R3 RI0 0 OR2






b~z bBz


[2-10a] Tetraol, R, = R3 = H R2 =OH [2-11] 7-0-TES 10-DAB III, R, H [2-l0b] Baccatin V, R, = Ac R2= Triethyl Silyl (TES)
R= HR = OH [2-12] 7-0-TES Baccatin I11, R, Ac
R2= Triethyl Silyl (TES)

0


NH 0 AO 0 0 iE)



CHHO AcO 0

OBz

[2-13] 7-O-TES-13-O-Cinnamoyl Baccatin III

Figure 2-2 :Taxol and some Synthetic Targets





17


clinical trials. A few studies on other species of Taxus have also been published, which are referred to in Chapter 1.

Semi-synthesis of Taxol

The relative ease in ester formation of the three hydroxyl groups in 10-deacetyl baccatin III (10-DAB, [2-7]) are 7>10>>13. Esterification of the C-13 hydroxyl is very challenging due to the "inverted cup"-like folding of the taxane skeleton and strong hydrogen bonding with the carbonyl oxygen on the C-4 acetate. Before the side chain can be attached at C-13, the 7-hydroxyl must first be protected, often accomplished by attachment of a triethylsilyl group to give [2-11]. Next, this compound is acetylated at the 10-position, to form 7-triethylsilyl baccatin III [2-12].

In one method, [2-12] was esterified with cinnamic acid to give [2-13], which was then converted to the phenyl isoserine ester by the Sharpless hydroxyamination procedure (Sharpless et al. 1991) using osmium tetroxide and t-butyl-N-chloro-N-sodiocarbamate (Mangatal et al. 1989). The four isomers were separated and after deprotection of the hydroxycarbamates, N-benzoylation and deprotection of the 7hydroxyl, taxol could be obtained.

During the investigations of Greene and Potier (Denis et al. 1988; Kanazawa et al. 1994) dozens of side chain analogues were synthesized and tested, resulting in the discovery of the taxol analogue known as taxotere [2-9]. Taxotere (docetaxel) was found to be more active than taxol in the tubulin assay and animal tumor systems and has also been approved as an antitumor agent. In an alternative synthesis, the 7protected baccatin III [2-12] was esterified using either the chiral P1-lactam [2-14] or the oxazinone [2-15] derivative to yield taxol. This method or some variation is currently used for the semi-synthesis of taxol and taxotere commercially from 10-deacetyl baccatin III (Ojima et al. 1991, 1992).





18


Total Synthesis

Swindell (1992) published a review on the progress of more than thirty groups and reported "only modest success" in the total synthesis of taxol. Only two years later two separate groups headed by K. C. Nicolaou (Nicolaou et al. 1994b) at the Scripps Research Institute and R. A. Holton (Holton et al. 1994b) at Florida State University would announce almost simultaneously two total syntheses of taxol.

Nicolaou and colleagues designed the strategy for their synthesis based on the one bond disconnection analysis seen in Figure 2-3. After preparation of the fully functionalized A ring [2-16] and C ring [2-17] equivalents, a convergent and flexible Esterification McMurry coupling
0 \
C6HS- AcO O OH
NH 0

OH A B Oxetane
OH formation
H AcO
Oxygenation OBz

Shapiro reaction OBn
OTBSi
TPSiO

A "'
I H "'
NNHSO2Ar O O

[2-16] Aryl sulphonylhydrazone [2-17] Aldehyde


Figure 2-3: Nicolaou's Retrosynthetic Strategy


synthesis of taxol involving 28 more steps allowed the preparation of numerous analogues. While not practical for the commercial production of taxol, synthetic methods





19


provide researchers with a source of analogues for structure-activity relationships and lead to better methods of production in general.

The first carbon-carbon bond between rings A and C was formed using a vinyllithium carbanion generated from the reaction of aryl sulphonylhydrazone [2-16], with n-butyl-lithium in tetrahydrofuran (THF); which was then combined with the aldehyde [2-17] in the Shapiro reaction (Shapiro, 1976) to produce [2-18] (Figure 2-3).

Regioselective epoxidation of the A1'14-double bond was completed in 87% yield with t-butyl peroxide in the presence of VO(acac)2 leading to epoxide [2-19], which was then regioselectively opened with LiAIH4 to give the 1,2-diol [2-20] with a 76% yield. The carbonate introduced between the C-1 and C-2 hydroxyls in the next step served to position the two rings for ring closure and also allowed for the stereo-controlled introduction of the 2ca-benzoate later in the sequence. The dialdehyde [2-21] needed for cyclization of the B ring was obtained after standard deprotection of the two primary hydroxyls and mild oxidation with tetra-n-propylammonium perruthenate (TPAP) and 4-methylmorpholine N-oxide (NMO) in acetonitrile. The previous three steps provided the carbonate dialdehyde in 32% overall yield.

Formation of the B ring was accomplished with the versatile McMurry coupling (McMurry, 1989) under dilute conditions utilizing low valence titanium produced in situ from (TiCI3)2-(DME)3 (10 eq.) and Zn-Cu (20 eq.) in 1,2-dimethoxyethane (DME) at 70 C for 1 hour, giving the tricyclic A/B/C diol [2-22] with a 23% yield.

Selective acetylation of the hydroxyl at C-10 rather than C-9 was expected due to allylic activation and proceeded with 95% yield. Mild oxidation of the C-10 hydroxyl was then carried out with TPAP-NMO in acetonitrile analogous to the oxidation to the dialdehyde with a 93% yield.





20


After removal of the acetonide and protection of the primary hydroxyl at C-20 to make, the benzyl group was removed with catalytic hydrogenation and the 7-O-triethyl silyl protecting group was introduced to give [2-23]. Selective deacetylation of the primary acetate then provided the triol for the formation of the oxetane of ring D, which involves monotosylation at C-20 (primary OH) and triflate formation at C-5 (secondary OH) to produce [2-24]. Oxetane formation with a 60% yield occurs after mild acid treatment with catalytic camphorsulfonic acid (CSA) in methanol, followed by treatment with silica gel in dichloromethane.

Acetylation of the C-4 position (tertiary hydroxyl) was followed by regioselective ring opening of the carbonate to the hydroxybenzoate functionality, both with good yields. The C-13L oxygen is introduced with pyridine chlorochromate in 75% yield followed by stereospecific reduction of the ketone [2-25] using NaBH4 in methanol in excess, for 83% yield. The hydroxyl is esterified using Ojima's j3 lactam synthon [2-14] (Figure 2-2) using the strong base sodium-hexamethyldisilazane for 87% yield based on 90% conversion. Removal of the triethylsilyl groups with hydrogen fluoride in pyridine (HF-Pyr) completes the synthesis of taxol in 80% yield. Other Synthetic Approaches

As previously mentioned Holton's group published a total synthesis of taxol in early 1994 at about the same time as Nicolaou, but their approach was quite different, with only a few reactions in common. Studies involving the fragmentation of bicyclic epoxy alcohols, referred to as "epoxy alcohol fragmentation," were the cornerstone of their syntheses of bicyclo[5.3.1] systems, including the unnatural epimer of (+)-taxusin [2-26], known as (-)-taxusin or ent-taxusin [2-27] (Figure 2-6).





21


TBSiO OTPSi TBSiO OTPSi
OBn n



141 HH
H0 0




[2-18] [2-191









TBSiO OTPSi0 0 n
OBn




HO HO 0

0



[2-20] [2-21]




Figure 2-4: Nicolaou's Taxane Ring Synthesis





22


HO OH OBn AcO 0 OTES



H C
O...H 0" O "OH


0 0
O OO OOc



[2-22] [2-23]






AcO O NOTES AcO O OTES

O
"..OTf A H
H OH HO
O O OAc
/ O 0OTs Bz
0 Bz


[2-24] [2-25]


Figure 2-5 Nicolaou's Final Synthetic Intermediates





23


AcO OAc AcO OAc

AcO ...... AcO
.OAc OAc
H H H
HH

[2-26] natural (+)-Taxusin [2-27] ent-(-)-Taxusin







\/
+ OH o
BF3
AcO OAc


CH30CH20'..... -- AcO
0 OAc
H H

Holton group Patchouline Oxide Fragmentation

0



0&
o \/





[2-28] Wieland-Miescher ketone [2-29] a-Pinene
Danishefsky group Wender group



Figure 2-6 : Starting Points of Other Synthetic Strategies






24


Danishefsky's total synthesis of baccatin III in 1996 (and hence, taxol); borrowed extensively from the experiences of Ojima, Holton, Nicolaou and others. The WeilandMiescher ketone [2-28], available through catalytic asymmetric induction, allowed the installation of all stereochemical requirements to reach baccatin III in a sequential fashion. According to Danishefsky, "Our synthesis, though arduous, involves no relays, no resolutions, and no recourse to awkwardly available antipodes of the chiral pool" (Danishefsky et al. 1996).

Wender's group published a most concise synthesis involving c-pinene [2-29] for construction of the ABC-tricyclic core of the taxanes (Wender & Rawlins 1992). Their approach takes advantage of the tendency for C-7 to undergo facile aldol/reverse aldol epimerization in taxol, allowing for aldol condensation under very mild conditions.

General Structural Features of Taxanes


The taxanes comprise a relatively large group of diterpenoid natural products covering a variety of structural patterns. These are believed to arise from geranyl geraniol [2-16], although the exact biosynthetic route has not been completely elucidated. A brief discussion of the major structural variations of taxanes is relevant to this work because many of these structures have been found in the compounds isolated in this work. A number of different forms that the C-20 diterpene skeleton itself can assume have been isolated. Next, the oxidation states, esterification patterns of the hydroxyls, and presence or absence of basic or neutral side chains allow for the extensive structural variation seen in these compounds.

The taxane skeleton is a specific diterpene structure, consists of 20 carbon atoms arranged in a fused tricyclic system with the 6, 8 and 6 members in rings A, B and C, respectively. The double bonds at 11/12 and 4/20 are part of the basic ring system, although the latter may be modified by oxygenation to an epoxide or more commonly to






25


an oxetane. As in the case of the analogous steroids with the two methyl groups as part of the skeleton, the taxanes have four methyl groups #16, 17, 18 and 19 as part of the taxane ring system. Some examples of the taxane skeleton found in the various species of Taxus are shown in Figures 2-1 and 2-2.

Oxygenation of the taxane ring has been observed to varying extents. The minimum number being 4, distributed at 5, 9, 10 and 13, as seen in taxusin [2-26]. In general, oxygenation may occur at carbons 1, 2, 4, 5, 7, 9, 10 and 13. Instances have been recorded where oxygenation was present at 14 (in place of 13), as well as part of the methyl groups at 19 and 17.

Taxa-4(20):1 1-dienes

This is the most common structural type seen in the taxanes, with a C-4(20) and a A1 double bond. These taxanes are generally referred to taxa-4(20);11-dienes. The alkaloidal Winterstein esters are included in this group as are many of the neutral taxanes. The oxygen at C-9, if present, is usually seen as a secondary alcohol or as an ester. The C-13 position in this group, likewise, exists as an alcohol, ester or oxidized to a carbonyl to form an x, -unsaturated carbonyl. Esterification at C-13 is usually limited to an acetyl or a cinnamoyl, but the side chain (N-acyl phenyl isoserine); as found in taxol, cephalomannine and others has not been reported in this subgroup so far. The 5 position is oxygenated with an a-hydroxyl, which might be free, or esterified by an acetic acid, cinnamic acid or the Winterstein-acid. Some examples of these compounds with a cinnamate ester function are described in Chapter 5. 4(20)-Epoxides

This group is relatively less frequent but examples with different substitution patterns have been isolated. One variation comes from the presence or absence of hydroxyl at C-1. Members of this subgroup also generally contain the 5-ca-hydroxyl,






26


which is esterified in the same fashion as the dienes above to provide further variation. An unusual example is the taxane with the C-9-nicotinoyl ester function, found in Austrotaxus spicata Compton Taxaceae (Ettouati et al. 1988). Oxetanes

This group is characterized by having an oxetane ring system involving the carbons 4, 5 and 20. It may be divided into two subgroups based on whether they contain the phenyl isoserine ester side-chain at C-13 or not. The former contains taxol and all of the other compounds, which are active in the tubulin assay and hence are of much importance. Division into two other subgroups is also possible in those without the C-13 side chain, with one having a carbonyl at C-9 and with a hydroxyl or an esterified hydroxyl at C-9.

The oxetane-containing taxanes are generally highly oxygenated and often have oxygen at C-1, 2, 4, 5, 7, 9, 10, and 13. In some special instances, a hydroxyl has been reported at C-19 (Fuji et al. 1993). The phenylisoserine ester side chain has been seen in the form of at least three different amides that occur in nature. These are taxol, with the N-benzoyl group, cephalomannine, with the N-tiglioyl group and taxol C, with the Nhexanoyl group.

Taxol has a complex structure and knowing what features of this structure are necessary for the activity is of utmost importance and this aspect has been studied using the in vitro tubulin binding, and the cell culture assays and a summary of these data is presented below (Samaranayake etal. 1993).

Acylation of the 2' position of taxol does not destroy cytotoxicity but does stop promotion of microtubule assembly. Bulky acyl groups reduce the activity in the cell culture, thus suggesting that hydrolysis of the 2' position back to a free hydroxyl might be required.






27


Substitution of the 7 position does not appear to significantly decrease the activity. Taxanes with a 7P-O-xyloside moiety are comparably active in both assays when compared to the respective aglycones. Similarly, epimerization at the 7-position does not eliminate activity.

Hydrolysis of the 10-acyl function does not reduce the cytotoxicity significantly in cell culture assays. As with other structural features, this point is being explored in the more recent clinical trials in Europe with taxotere.

The importance of the oxetane ring for activity has been investigated through ring opening via different Lewis acids including Meerwein's reagent (triethyloxonium tetrafluoroborate); acetyl chloride, mesyl chloride and others. The product obtained form the Meerwein's reagent had a primary alcohol at C-20 and secondary C-5-hydroxyl, but no other changes compared to taxol. The activity normally seen with taxol in both assays was lost with the opening of the oxetane ring. This suggests that the oxetane ring is necessary for activity but leaves open questions regarding the effect of ring contractions in ring A.

The properties of the C-13 hydroxyl mentioned above make attachment of a side chain quite difficult. Protection of other free hydroxyls in both the side chain and taxane skeleton are necessary, followed by selective deprotection after the side chain has been attached. Taxotere and taxol have both been synthesized from this taxane and this is currently the starting material for the production of both drugs.

Epimerization of the 7 hydroxyl from 3 to a via a retro-aldol mechanism allows formation of an energetically favorable hydrogen bond with the 4-acetate carbonyl oxygen. This epimerization is a concern in both taxane isolation and synthetic methods, and necessitates the avoidance of acidic or basic conditions. Protection of this C-7 3-






28


hydroxyl with groups such as a chloroacetate avoids both epimerization and unwanted reaction at this position.

Abeotaxanes

A number of taxanes in which the A-ring is isomerized to a 5-membered ring to give a 5/7/6 instead of the 6/8/6 system have been isolated and these are termed abeotaxanes. They are again divided into two groups into a) those with the 4/20 unsaturation and b) those with an oxetane ring at this location. We isolated the first members of each of these groups in our work, e.g. brevifoliol (Chapter 3); and the compounds isolated from the bark of T brevifolia described in Chapter 6. As indicated earlier, treatment of taxol with acidic reagents can isomerize ring A to form such compounds, although these compounds are naturally present in the extract and not artifacts.















CHAPTER 3
TAXANE CONSTITUENTS OF THE NEEDLES OF TAXUS BREVIFOLIA


Taxol was originally isolated from the bark of the Pacific yew (Taxus brevifolia Nutt., N.O. Taxaceae). As indicated in Chapter 1, during 1991-1993 there was a reassessment of the use of the bark as the source. This concern resulted in an intense search for alternative sources for taxol that are renewable, with sources such as the needles of the yew tree instead of the bark. This laboratory was also involved in this search and looked into the needles of three different yew species as a source for taxol: T. brevifolia, T. x media Hicksii and T floridana. The taxane composition of T. brevifolia needles is the subject of this chapter.

Fractionation of the Needles of Taxus brevifolia


A quantity of 100 lbs. of the needles of T. brevifolia was obtained from a supplier in Oregon. They were air-dried and extracted with methanol at room temperature and the extract was concentrated under reduced pressure to a syrup. This was partitioned between water and chloroform, and the organic layer concentrated to give a dark greenish brown semi-solid, called "extract solids", which represented about 5% of the dry weight of the needles.

It was decided to follow the method successfully developed with the bark extract for the fractionation of the extract solids, using preparative scale, reverse phase column chromatography. Direct application of the crude chloroform extract of the needles onto a C-18 bonded reverse phase silica column was accomplished as described in the experimental section. After placing the extract-containing silica onto a 25% acetonitrile



29





30


in water column (1:4 ratio of loaded to clean silica); a step gradient of acetonitrile in water mixtures was performed up to 60% acetonitrile.

Preliminary studies on the extract solids of the needles by TLC and analytical HPLC showed that the sample contained somewhat minor amounts of taxol. A predominant component that was slower moving than taxol in TLC gave a greenish-blue colored spot when sprayed with 1 N sulfuric acid and heated on a hot plate (charring). Likewise, in the analytical HPLC, this component appeared after 10-deacetyl baccatin III as the major constituent judging from the peak heights, but before taxol and at least several times more abundant.

The reverse phase column (C-18 bonded silica gel) on the needle extract concentrate was started with 25% acetonitrile in water. The sample was carefully prepared as a slurry (see experimental) and added to the column. The column was developed using a step gradient of acetonitrile in water 30-60%. Fractions of suitable volume were collected and monitored by absorbance at 275 nm, TLC and analytical HPLC. Four regions were recognized in the elution profile of the column, based on the UV absorbance (275 nm.); which contained the resolved constituents of the extract.

The early fractions contained components, which accounted for the bulk of the UV absorbance of the sample. These appeared to be non-taxane phenolic compounds with or without attached sugars. A description of these will be given in Chapter 7. The first taxane component, which appeared at the 35-40% acetonitrile elution, was also the major component. It was collected from the appropriate fractions, and after concentration, obtained as a crystalline solid. Next, fractions from the 50% acetonitrile elution contained taxol, which was obtained as a crystalline solid directly from the fractions. Following this, the fractions from the 55-60% acetonitrile elution gave another taxane component which gave a greenish blue spot on the TLC (after charring with sulfuric acid) similar to the major constituent referred to above.





31


Brevitaxane A (Brevifoliol) [3-11


The major constituent, which was obtained in a yield of 0.2-0.25%, was named brevitaxane A because the physical and spectral data indicated that it was a new taxane compound (later renamed by others as brevifoliol, which will be used throughout this dissertation). Elemental and FAB-MS analysis (MH+ 557) agreed with the molecular formula of 031 H4009 (Balza et al. 1991).

An examination of the 1H NMR spectrum showed the presence of two acetyl groups (signals at 6 1.76 and 6 2.07); and a benzoate group {6 7.88 (d); 6 7.43 (t) and 8 7.56 (t)}. The spectrum also gave evidence for the presence of a (4/20) exocyclic double bond (two characteristic broad singlets at 8 4.82 (H-20A) and 6 5.20 (H-20B) and signals at 6 112.1 (C-20) and 6 149.0 (C-4) in the 13C NMR spectrum.

Very little information on the various types of taxane structures that are known now was available at that point in time (1991) and even less on their diagnostic spectral characteristics. Based on analogous taxanes and the evidence outlined above it was postulated that this major constituent had the relatively common 4/20,11-taxadiene type skeleton. The presence of an exocyclic 4/20 double bond and absence of an oxetane ring supported our initial assumptions. The next step was to determine the positions of the various substituents in the molecule in order to elucidate the complete structure.

Most of the structural elucidations of taxanes at the time were based on degradative studies. It was decided to follow this lead in establishing the presence of the various functionalities as well as their location in brevifoliol, by actual reactions and/or derivatizations, supplemented by spectral methods.






32










%"A





OD U) (0


T


co
7 co CY) 0
0 E
L)
< 0
0 (D

0 0
< .0 N ne
oil., 7 2
CO z
C:
0
N LO 0
CO
0
;z C\I -Lr)
A
2
2) LL
0
T








E




-00






33


Hydroxyl Functionalities


i) Acetylation: To determine the number and positions of all hydroxyls in the

molecule, the compound was subjected to acetylation. Two products were obtained

under mild conditions (20 0 C, 15 min). These two were separated by chromatography

and both obtained as crystalline solids. One was shown to be a monoacetate and the

other a diacetate.


Table 3-1 : Proton NMR Spectra of Brevifoliol and Brevifoliol Acetates Position Brevifoliol Brevifoliol Brevifoliol Brevifoliol Brevifoliol
(J in Hz) 5-Ac 13-Ac 5,13-Ac 5,13,15-Ac
[3-1] [3-2] [3-3] [3-4] [3-5}
2 1.49 cm 1.46 cm 1.47 cm 1.46 br d (13) 1.53 br d(13)
2.36 dd (9,13) 2.40 dd(9, 13) 2.42 2.41 dd(9, 13) 2.65 dd(9,
3 2.78 d (9) 2.76 br d (9) 2.91 d (9) 2.72 br d (9) 2.71 br d (9)
5 4.45 br s 5.37 4.37 br s 5.39 br s 5.38 dd (4, 2)
6 1.86 cm 1.88 cm 1.85 cm 1.90 cm 1.87 cm
2.02 cm 2.0 cm 1.99 cm 2.00 cm 2.0 cm
7 5.56 dd (5,11) 5.62 dd (5, 11) 5.66 5.61 dd (5, 5.63 dd(5,
9 6.05 br 6.03 br d(10.6) 6.07 d 6.09 br 5.8 d (10.8)
10 6.53 d (10.6) 6.63 d (10.6) 6.66 d 6.65 d (10.6) 6.64 d (10.8)
13 4.38 t (7.5) 4.53 br t (7.2) 5.46 br s 5.54 br t (7.2) 5.61 t (6.9)
14 1.29 1.22 dd* 1.32 cm 1.25 dd* 1.25 dd*
2.46 2.42 dd 2.51 cm 2.51 dd 2.62 dd *
16 1.05s 1.03s 1.09s 1.11 s 1.63s
17 1.35s 1.33 s 1.35s 1.35 s 1.71 s
18 2.01 s 2.06 s 2.02s 2.03s 1.96s
19 0.90s 0.91 s 0.89 br s 0.92s 0.92s
20 A 4.82 br s 4.90 br s 4.80 br s 4.92 br s 4.89 br s 20 B 5.20 br s 5.28 brs 5.15 brs 5.28 brs 5.29 br s
o-Phl 7.88 d (7.5) 7.87 d (7.5) 7.87 d (7.5) 7.87 d (7.5) 7.84 d (7.5)
m-Phl 7.43 t (7.5) 7.43 t (7.5) 7.44 t (7.5) 7.44 t (7.5) 7.42 t (7.5 )
p-Phl 7.56 t (7.5) 7.55 t (7.5) 7.56 t (7.5) 7.56 t (7.5) 7.53 t ( 7.5)
- 1.76s 1.76s 1.76s 1.75s 1.77s
2.07s 206s 2.05s 2.02 s, 2.07 s 2.02s, 2.08s
1 2.13s 2.06s 2.08 s 2.09 s, 2.11 s

NMR were recorded at 600 MHz in CDC3 on a Varian Unity 600 instrument at ambient temperature. Chemical shifts 6 (ppm) are reported with TMS as internal standard.





34




Table 3-2 Carbon NMR Spectra of Brevifoliol and Brevifoliol Acetates Carbon [3-1] [3-2] [3-3] [3-4] [3-5]
Number Brevifoliol 5-Ac 13-Ac 5,13-Di-Ac 1,5,13-Tri-Ac
1 62.4 63.0 63.4 63.0 63.3
2 29.1 29.2 29.4 29.1 28.3
3 37.9 38.8 37.6 38.8 38.9
4 149.0 145.4 147.4 145.2 145.1
5 72.4 74.1 72.7 74.1 73.9.
6 36.0 33.9 36.1 33.9 34.0
7 70.1 69.7 69.8 69.6 69.7
8 45.0 44.8 45.2 44.8 45.0
9 77.1 77.9 79.8 79.3 78.9
10 70.2 70.7 70.3 69.8 68.4
11 133.9 134.0 136.5 136.4 136.5
12 151.5 151.1 150.5 147.3 148.2
13 76.7 76.9 77.8 76.9 78.0
14 47.3 47.1 44.2 44.1 43.3
15 75.9 75.6 75.6 75.6 87.2
16 26.9 27.0 27.0 27.0 23.1
17 24.8 24.8 25.0 24.8 21.8
18 12.0 11.8 12.1 11.9 11.9
19 12.9 12.9 12.9 12.9 13.5
20 112.0 114.1 111.5 114.3 114.3
CO-C6H5 164.3 164.1 164.2 164.1 165.0
Bz-ipso 129.3 129.2 129.4 129.1 129.9
Bz-ortho 129.4 129.4 129.5 129.5 129.3
Bz-meta 128.7 128.7 128.7 128.8 128.4
Bz-para 133.2 133.3 133.2 133.4 133.0
CO-C-3 20.7 20.8 20.7 20.7 20.8
21.4 21.4 21.4 21.4 21.4
21.2 21.1 21.2 21.3
21.0 21.0
21.7
COCH3 169.9 169.9 (X2) 169.9 169.91 169.9
170.5 170.2 170.8 170.5 170.5
169.7 169.6 169.6
169.9 171.0
169.5

13C NMR spectra were recorded at 150 MHz in CDCl3 on a Varian Unity 600
spectrometer at ambient temperature. Chemical shifts 5 (ppm) are reported with TMS as internal standard.





35

BzO OAc R1 R2 R3
Ac [3-1] H H H
[3-2] Ac H H
[3-3]- H Ac H
R20 [3-4] Ac Ac H
R20 ..... OR1 [3-5] Ac Ac Ac

OR3

BzO Ac BzO OAc Ac
SOAc



"OH OH
OH OH H3

[3-6] 13-Ketone [3-7] 4,20-Dhydro

Bz QAc B A
Ac Bz OAc


HO"... HO...
OH 'OH
OH OH 0
[3-8] 4,20-Epoxide [3-9] Norketone


Bz CAc H PR R
Ac H Q2


HO' HO H
H O ....... H O1" .......
HOH OH
OH
OH OH OH
[3-10]- 4,20-Diol [3-11]- Hydrolysate, R1 = R2 = H
[3-15] Debenzoyl, R1 = R2 = Ac


Figure 3-2 : Brevifoliol and Reaction Products





36


Appropriate conditions under which each of these could be obtained as exclusive products were developed. At room temperature in acetic anhydride for 1-2 minutes before quenching the reaction, the monoacetate was the major product (>90%). Likewise, at 80 o C for 30 min. the product was the diacetate.

The 1H NMR spectral data for the monoacetate showed that the signal at 6 4.45 (br s) shifted to 8 5.37 (dd, J=4.2, 2.4 Hz); indicating that acetylation took place at the 5OH, as shown in [3-2]. In the diacetate, besides this shift for the 5-OAc, the signal at 6 4.38 (t, 7.5 Hz) shifted to 8 5.54 (br t, 7.2 Hz); thus showing that the second acetate was located at C-13 [3-4]. A naturally occurring brevifoliol 13-acetate [3-3] was isolated and 1,5,13-brevifoliol triacetate [3-5] produced in this lab will be discussed in Chapter 4.

ii) Oxidation: Brevifoliol was readily oxidized by manganese dioxide (MnO2) in refluxing benzene to yield a ketone product. In the 1H NMR spectrum, a major change was the absence of the triplet at 6 4.38 due to the C-13 proton, thus showing that the oxidation took place at the 13-OH [3-6]. Further evidence was seen by the shift of the signals for the C-14 protons from their normal positions at 6 1.29 (dd, 14.0, 7.6 Hz) and 6 2.46 (dd, 14.0, 7.6 Hz) to 5 2.32 (d, 19 Hz, H-14) and 6 2.48 (d, 19 Hz, H-14P). When brevifoliol was oxidized by Jones reagent, the same 13-keto brevifoliol seen with MnO2 initially formed [3-6]. With time the initial product gradually disappeared, giving rise to a faster moving product. This second oxidation product was shown to be the result of an unusual reaction described in Chapter 4.

4/20 Unsaturation


i) Hydrogenation: When hydrogenated in the presence of 5% Pd/carbon, brevifoliol gave the dihydro derivative [3-7]. In its 1H NMR spectrum, the characteristic signals at 6 4.82 and 5 5.20 due to the C-20 protons were absent and a new methyl





37


doublet and a new methine proton appeared. In the 13C NMR spectrum the characteristic signals from the exocyclic 4/20 double bond were absent, accompanied by the appearance of new methyl and methine signals.

ii) Epoxidation: Brevifoliol was heated in dichloromethane with meta-chloro peroxybenzoic acid (MCPBA); whereby it underwent oxidation to yield the epoxide [3-8], a crystalline compound.

iii) Ozonization: Brevifoliol has two double bonds, one at the 11/12 position and the other at the 4/20 position. Of these, the former is tetra-substituted, while the latter is of an exocyclic methylene type. No information was available in the literature regarding the reactivity of the taxane skeleton to indicate whether one or both double bonds would be cleaved by ozonolysis. In the present work, ozonization was carried out in a mixture of methanol and dichloromethane -70 0 C. After the disappearance of the starting material, the ozonide was decomposed with dimethyl sulfide and the products isolated by chromatography. Two major products were separated. The first was the same as the epoxide [3-8] obtained by reaction with MCPBA. The second was the expected ozonolysis product in which the 4/20 double bond was cleaved to form the ketone [3-9].

iv) Formation of a diol: As one of the characteristic reactions of an ethylenic function, oxidation by osmium tetroxide was attempted with brevifoliol. The reaction proceeded smoothly to give a diol [3-10].

Number and Nature of the Oxygen Substitution


From the preceding discussion it is evident that brevifoliol has two free hydroxyls, two acetoxyls and one benzoyloxy functions. However, in the 13C NMR spectrum of brevifoliol, the number of oxygen substituted carbons was six: 5 70.1, 5 70.2, 6 72.4, 6 75.9, 6 76.7 and 6 77.2. To determine if one of the six is a different type of an ester, or a tertiary hydroxyl, brevifoliol was subjected to saponification in alcoholic KOH to yield the






38


hexaol [3-11], obtained as a crystalline solid. This was then acetylated to a crystalline acetate [3-12]. The 1H- and 13C NMR spectra of [3-12] showed the presence of 5 acetates (1H: 8 21.8, 5 21.7, 8 21.4, 6 21.3, 6 21.0, 6 20,8; and 13C 6 171.0, 8 170.4, 6 169.8, 6 169.6 and 8 169.5); which suggested that brevifoliol contained a tertiary hydroxyl.

In the conventional taxane skeleton, a tertiary hydroxyl is often present at C-1, with the other hydroxyls (or esters) at C-2, C-5, C-7, C-9, C-10 and C-13. Thus, with brevifoliol having five oxygen substituents, one of these positions must be without attached oxygen. Thus, it would be important to know which of these positions does not have an oxygen substituent. For this reason, the hexaol [3-11] was subjected to oxidation by periodate. If there were two pairs of vicinal hydroxyls, e.g. 1,2 and 9,10, the hexaol will be cleaved in such a way as to give smaller molecules which represent the A and C rings. If there is only one such pair, the reaction will produce a product with all of its carbons intact. The hexaol underwent oxidation readily to form a dialdehyde [3-13] without losing any carbon atoms found in original carbon skeleton. Unaware of the unusual A ring structure, it was presumed that the presence of a tertiary hydroxyl at C-1 precluded the presence of oxygen substitution at C-2. Additional evidence for a methylene carbon at C-2 was found in the COSY spectrum from the chemical shifts in the H-3p3-H-2c-H-2p3 isolated spin system.

Thus, brevifoliol has two hydroxyls at 5 and 13. Locating the benzoate group at one of the three choices, 7, 9, or 10 will elucidate the structure. At this point, brevifoliol was required in microbial and fungal biotransformation project in our laboratory. In order to produce an antiseptic sample an aqueous alcoholic solution was sterilized in a steam autoclave at 125 0 C, 20 atm., to see if it is stable. It was found that the compound underwent degradation to give two or three products.





39


AcO OAc OAc OHO OH
H

A cO H O -.
OAc "OH
OH OH

[3-12] lentaacetate from Hexaol [3-13] Periodate Oxidation


02N NO2



02N NH NH NO2
\ /
NH N OH



HO,...
--."OH
OH


[3-14] Periodate Oxidation Osazone Product


Figure 3-3 : Brevifoliol Hexaol Reaction Products


The major component of this mixture was found to be debenzoyl brevifoliol (Figure 3-1 [3-15]). Of the three possible locations, 7, 9, and 10 for the benzoate, only 10 is allylic and hence the ester at this position is more likely to be labile. Taxol with the benzoate at C-2 is completely stable to heat and pressure for hours. This evidence, along with chemical shift arguments concerning the effect of acetylation versus benzoylation, led us to place the benzoate at C-10.

This group presented the isolation and the structural elucidation of brevitaxane A at the International Research Congress on Natural Products held in Chicago, IL in July





40


1991. Balza et al. (1991) published the isolation of a new compound at about that same time from the needles of T. brevifolia, which they named brevifoliol, and an assignment of its structure as shown in [3-16]. The compound appeared to be similar to, if not the same as, brevitaxane A, that was isolated from the needles at the University of Florida. The structure proposed by Balza et al. differed from that of brevitaxane A, with the benzoate group being placed at C-7 instead of at C-10.

That same year (1992); the isolation of taxchinin A was described (Fuji et al. 1992); which was later shown to be 2-acetoxy-brevifoliol. Fuji correctly assigned the structure with a 5-membered ring A, on the basis of NMR spectral data. The authors who isolated brevifoliol and assigned structure with the 7-benzoate (Chu et al. 1993) published a revised structure for brevifoliol, in which the benzoate was moved to C-10, from C-7, but with the skeleton of a conventional taxane.

During 1993, two other publications appeared, one from Georg et al. (1993). and the other from Appendino et al. (1993) reexamining the NMR spectral data of brevifoliol, and arriving at the structure in which the A-ring was 5-membered. Later that year, Chu et al. (1993); on the basis of x-ray crystallographic data, revised the structure of brevifoliol again to the presently accepted structure.

Due to the intense competition in "taxol research", we began a detailed examination and analysis of the NMR spectral data using the 13C NMR, NOESY, HETCOR and other spectral methods to determine if the rearranged (5/7/6) skeleton might be distinguishable from the spectrum of a taxane with a conventional (6/8/6) skeleton. The following is an analysis of the spectral data of brevifoliol.

The carbonyl signal in the 13C NMR spectrum at 6 164.3 indicated the presence of one benzoate, and signals at 5 169.9, 6 170.5, likewise, indicated that 2 acetate ester groups were present. Further support for the benzoate was obtained by the four






41


aromatic signals between 8 128.7 and 8 133.2 (see tables 3-1and 3-2); and for the two acetates, by the methyl signals at 8 20.7 and 8 21.4. Analysis of the 1H NMR and 1H COSY and 1H,13C Heteronuclear Correlation (HETCOR) experiments also gave additional support for the presence of the acetates with signals at 8 1.76 s and 8 2.07 s, as well as benzoate signals at 8 7.88 d (ortho); 8 7.43 t (meta); and 8 7.56 t (para). Next, evidence for the presence of the normally present (11/12) taxane double bond could be seen in the carbon spectrum by the signals at 8 133.9 (C-11) and 8 151.5 (C-12);. Similarly, the existence of a (4/20) exocyclic double bond could be seen by the signal 8 149.0 (C-4) and 8 112.1 for (C-20). In the 1H NMR spectrum the exocyclic 4/20 double bond is also indicated by the two characteristic broad singlets seen at 8 4.82 (H-20A) and 8 5.20 (H-20B).

In the 1H COSY experiment weak but definite interactions between the singlet at 8 4.82 (C-20a) with both the H-3j3 doublet at 8 2.78 (9 Hz.) and the H-2a doublet of doublets at 8 2.36 (9, 13 Hz.) supported the assignments given for the methylene protons. (The designations for the C-20 protons are A and B, since a and 3 do not have the conventional meaning system and could be confusing). Along with the interaction between H-2u. and H-203 the first isolated spin system in the 'H spectrum was established and the relative geometry of the protons.

The region between 8 62.4 and 8 77.1 in the 13C spectrum carbons with hydroxyl or ester oxygen attached to oxygens, and these signals could be further defined in the DEPT experiment (Distortionless Enhancement with Polarization Transfer, NMR) as primary, secondary, tertiary and quaternary carbons. The spectrum showed two quaternary carbon signals and five oxymethine carbon signals. Since the presence of only six signals was expected based on the proposed formula, the quaternary signal at 8 62.4 was intriguing even from the start of the spectral examination. In taxol with its C-9






42


carbonyl, the C-8 signal appears near 6 58, so the signal at 6 62.4 immediately raised questions about the true structure of brevifoliol. This signal did not fit the normal chemical shift pattern of any naturally occurring taxanes known at that time. In the absence of a carbonyl group at C-9, the C-8 carbon usually falls in the region of 6 40-50 ppm.

Unable to satisfactorily explain this unusual peak position, the Chemistry Department was contacted about crystallographic services. X-ray crystallographic analysis was performed by Dr. K. A. Abboud on the 5-monoacetate [3-2]. Surprisingly, the presence of an unusual 5/7/6 ring system was evident, where the normal 6membered A ring of the conventional taxane system was "rearranged" to form a 5membered ring with the carbons 15, 16 and 17 moved out of the ring system to form a hydroxy isopropyl group at C-1. Since the x-ray structure was obtained on brevifoliol-5acetate, it was important to establish whether brevifoliol itself had this rearranged taxane skeleton, or if the rearrangement could have occurred during the acetylation.

This structure represented a departure from the existing naturally occurring taxane structures available at that time, previously seen only as a product of rearrangement under strongly acidic conditions (Samaranayake et al. 1990). Crystallography of the original compound was not done because it failed to yield adequate crystals for analysis without prior acetylation. This made it necessary to determine whether this new ring structure was naturally occurring, or formed during the acetylation.

In one such ring contraction, taxol underwent rearrangement of the A-ring, accompanied by dehydration, to produce an isopropenyl group at C-1, as well as other changes such as the opening of the oxetane ring. Since the 13C NMR spectra of both brevifoliol and its monoacetate showed these signals at 6 62.4 assigned to the C-1






43


carbon and the one at 6 75.9 assigned to the quaternary C-15 containing tertiary hydroxyl, it appeared unlikely that such a rearrangement took place during the acetylation. HETCOR and APT experiments corroborated these conclusions, thereby agreeing with the structure determined by the x-ray crystallographic method.

Further analysis of the 1H COSY spectrum revealed an isolated spin system of two doublets due to H-9p3 at 6 6.05 and H-10occ at 5 6.53, with a pseudo-axial orientation indicated by the degree of splitting (J=10.6 Hz); and significant broadening of the signal at 6 6.05. Some amount of the deshielding of H-10a relative to H-9P3 was expected, due to the adjacent double bond, which makes the C-10 position allylic. The presence of a benzoate at this position would be expected to cause a further downfield shift based on analogous compounds already known (Chu et al. 1992). With a thorough analysis of the 1H NMR and 1H COSY spectra, the signal at 5 4.38 (t, 7.6 Hz) was assigned to the H1313 proton, which coupled strongly with H-14P3 at 6 2.46 (dd, 14.0,7.6 Hz.); as well with H-14t at 6 1.25 (dd, 14.0, 7.6 Hz). Weak long range coupling to the C-18 methyl protons at 8 2.01 was also evident, as the slight broadening of this peak is generally attributed to this long range coupling in other taxanes.

The isolated spin system of H-5p, H-6L, H-6P and H-7ox is easily identified in most taxanes, with a tendency to show a sharp multiplet for H-7L and broader, poorly resolved splitting for H-53, especially if H-5p is not esterified (Della Casa de Marcano & Halsall, 1970; Rao et al. 1995). The H-5p broad singlet at 6 4.45 interacts with the H6c multiplet at 6 1.86, which interacts with the H-613 multiplet at 6 2.02, which in turn interacts with the H-7a signal at 6 5.56 (dd, 5,11 Hz.). In many cases esterification of a hydroxyl causes a deshielding effect on the related proton of about 1 ppm. The chemical shifts and splitting patterns indicated that the acetate groups were at C-7 and C-9, with the benzoate at C-10.






44


The remaining carbons are the 4 methyl groups usually seen in taxanes on C-15 (methyl 16 and methyl 17); at C-12 (methyl 18) and at C-8 (methyl 19). The methyl group located on the 11/12 double bond (methyl 18) is often quite deshielded in the proton spectrum (6 2.01, s) but shielded in the carbon spectrum (6 12.0). This usually aids in its assignment along with further evidence from Heteronuclear NMR experiments. Methyl 19 is usually shielded in both the 1H (5 0.90, s) and 13C (8 12.9, q) spectra, as seen here.

This class of compounds commonly referred to as 11(1->15)-abeo-taxanes or occasionally A-nortaxanes. Many compounds of this type are now known, some containing the 4/20 unsaturation as in brevifoliol and others with a 4/20 oxetane structure as seen in 11 (1->15) abeo baccatin VI.

Experimental


Extraction of the Needles of Taxus brevifolia

The needles obtained from a supplier (Mr. Patrick Connolly, Yew Wood Industries, 6928 North Interstate Avenue, Portland, OR 97217) were air-dried for one week. The dried needles (20 Kg) were extracted by immersing in methanol at room temperature. After two days, the extract was drained, concentrated under reduced pressure at temperatures below 35 o C. The recovered methanol was reused for a second extraction, which was processed the same way. After two more extractions, the combined concentrate was freed from some more of the methanol to obtain a dark green syrup.

The above syrup was partitioned between water (10 gallons) and chloroform (10 gallons). The organic layer was separated and the extraction carried out twice more






45


using 5 and 3 gallons respectively. The combined chloroform extract was concentrated under reduced pressure to reach a dark green semi-solid stage (800-900g). Reverse Phase Column Chromatography:

The column used was a threaded glass column of the Mitchell-Miller type (2.5 x 24") with the appropriate fittings, purchased from Ace Glass Co., Vineland, NJ suitable for low pressure liquid chromatography. A slurry of the C-18 bonded silica (800 g) (Spherisorb, 15-35 micron diameter) purchased from, Phase Separations Inc., Norwalk, CT) in methanol was poured into the column, which was run under a gentle pressure by using a metering pump (Fisher/Eldex) until an adequately packed bed was obtained. The column was then equilibrated with 25% acetonitrile in water, to prepare for the addition of the sample.

The extract solids (200 g) was dissolved in acetonitrile (400 ml) by warming to make sure that no lumps remained. To this was added approximately 200 g equivalent of the equilibrated resin (about 20% of the column packing) with stirring. As the stirring continued, the slurry was diluted with 25% acetonitrile in water (500 ml); followed by water to make up a total volume of approximately 2 L. The stirring was continued with occasional warming to 50-60 0 C for about 15 min. At this point, a sample of the slurry taken into a test tube, showed that the silica settled readily to give a clear supernatant and no green precipitate or oily material was present. The slurry was then filtered using light suction and the solid (silica with the sample) re-slurried using part of the filtrate and the thick slurry added to the column. The rest of the clear supernatant was then pumped on to the top of the column using the metering pump. From time to time, the column feed was checked to see that it remained clear, and if not, to either warm briefly or add minimal amounts of acetonitrile to it until it became clear, so as to prevent precipitate from appearing and blocking the pump.






46


After the sample addition was completed, fresh 25% acetonitrile! water was passed through, followed by the step gradient of acetonitrile! water (30, 35, 40, 45, 50 and 60%) was used. Fractions (200 ml) were collected and monitored by UV absorbance (at 275 nm), TLC and analytical HPLC. The change to the next concentration of solvent was determined by the results of monitoring the fractions. For example, when the absorbance values rose as a result of the previous change, the solvent was continued until a definite trend to lower values was seen. Similarly, when the TLC showed the trend towards decreasing intensity of the major spot, and no new spot had shown a tendency to increase, the solvent was changed to the next level. In general, 2-3 multiples of the hold-up volumes of the column were used.

After the elution with the 60% solvent was completed, the column was washed with 100% methanol, followed by a mixture of methanol! ethyl acetate! ligroin which stripped the column of the chlorophylls, waxes and other lipid soluble components. After this solvent, washing with methanol and equilibration with 25% acetonitrile! water made the column ready for another run.

After the monitoring, fractions with low UV-absorbance values were combined and concentrated into groups, based on the TLC data. Those fractions with relatively stronger UV readings were let stand at room temperature for 3-5 days, whereby crystals appeared in several sections of the fraction sequence. These crude crystals were filtered, dried and purified further either by recrystallization or using a small silica column (normal phase).

Brevifoliol [3-11


The fractions containing this component gave crystals but only a small portion was obtained in this form. Hence, after filtration of the crude crystals, the filtrate was concentrated to dryness and the solid taken up in dichloromethane and passed through






47


a column of normal phase silica, using a ratio of 3-5 g of silica per gram of the solid. The effluent and washes which contained the compound were combined, concentrated to dryness and the solid crystallized from a mixture of acetone and ligroin to obtain brevifoliol as a colorless crystalline solid, yield from 200 g of the chloroform extract solids, 12 g, 0.25% of the dried needles. [a]D23 -27 0 (CHCs; c 1.03); m.p. 220-222 o C (lit. 200-203 OC [Balza et al. 1991]44);

FAB-MS m/z: 557 [MH]*, 539 [MH-H20]*, 479 [MH-AcOH] 435 [MH-PhCO2H]', 417 [MH-PhCO2H-H20]*, 375 [MH-PhCO2H-AcOH]', IR (KBr) vmax cm-': 3370, 1740, 1650, 1600, 1585, 1450, 1370, 1265, 1180. UV X max logs 3.01 (269 nm); logs 4.32 (223 nm).

1H NMR (600 MHz, CHC3, 6) Table 3-1: 0.90, s (H-19); 1.05, s (H-16); 1.30 (dd, J=7.2, 13.8 Hz, H-14c); 1.35, s (H-17); 1.50 (d, J=14.1 Hz, H-2a); 1.76 (s, methyl, 9acetate); 1.80 (m, H-6c ); 2.0 (m, H-613); 2.01 (s, H-18); 2.07 (s, 7-acetate methyl); 2.36 (dd, J=14.1, 9.6 Hz, H-2p); 2.46 (dd, J=7.2, 13.8 Hz, H-14P3); 2.67, br s (C-15 OH, exchangeable with D20); 2.77 (br d, J=9 Hz, H-3a); 4.38 (t, J=7.2 Hz, H-13P3); 4.43 (br s, H-53); 4.82, s (H-20 A); 5.18, s (H-20 B); 5.57 (dd, J=4.8, 11.4 Hz, H-7o); 6.05 (poorly resolved br d, J=10.5 Hz, H-9c); 6.53 (d, J=10.5 Hz, H-103); 7.43 (t, J=7.8 Hz, H-Bzmeta); 7.56 (t, J=7.8 Hz, H-Bz-para); 7.87 (d, J=7.8 Hz, H-Bz-ortho).
13C NMR (CDCl3, 600 MHz, 6) Table 3-2: 12.0 (C-18 methyl, q); 12.9 (C-19 methyl, q); 20.7 (7-0 acetate methyl, q) ; 21.4 (9-0 acetate methyl, q); 24.8 (C-17 methyl, q); 26.9 (C-16 methyl, q); 29.1 (C-2, t); 36.0 (C-6, t); 37.9 (C-3, d); 45.0 (C-8, s); 47.3 (C-14, dd); 62.4 (C-1, s); 70.3 (C-7, d); 70.9 (C-10, d); 72.4 (C-5, d); 75.9 (C-15, s);

76.7 (C-13, d); 77.1 (C-9, d); 112.0 (C-20, t); 128.7 (C-Bz-meta, d); 129.3 (C-Bz-ipso, s);

129.4(C-Bz-ortho, d); 133.3 (C-Bz-para, d); 133.9 (C-12, s); 149.0 (C-4, s); 151.5 (C-11,

s); 164.3 (CO-Ph, s); 169.9 (CO-Acetate, s); 170.5 (CO-Acetate, s).





48


Analysis calculated for C31 H40 09: C, 66.89; H, 7.24. Found: C, 67.12; H, 7.35;



Brevifoliol-5-Monoacetate [3-21

A mixture of brevifoliol (0.2 g); acetic anhydride (2 ml) and pyridine (0.5 ml) was stirred at room temperature for 2-3 min. Water was added to decompose the reagent, and the solid filtered after 15 min. The solid was crystallized from a mixture of acetone and ligroin to obtain the mono acetate as a colorless crystalline solid, yield, 0.18 g; m.p.224-226 C;

1H NMR (CDCI3, 600 MHz, 6) Table 3-1: 0.91, s (H-19); 1.02, s (H-16); 1.22 (dd, J=7.2, 13.8 Hz, H-14c); 1.33, s (H-17); 1.46 (d, J=14.1 Hz, H-2cX); 1.76 (s, 9-0 acetate methyl); 1.88 m, 2.0 m (H-6); 2.06 x 2, s (methyl-18, 5-0 acetate methyl); 2.08 (s, 7-0 acetate methyl); 2.40 (dt, J=14.1, 9.6 Hz, H-213); 2.42 (dd, J=7.2, 13.8 Hz, H-143); 2.75 (d, J=9 Hz, H-3o); 2.83, br s (C-15 OH, exchangeable with D20); 4.53 (t, J=7.2 Hz, H13P3); 4.90, s (H-20 A); 5.28, s (H-20 B); 5.37 (br s, J= H-5p3); 5.65 (dd, J=4.8, 11.4 Hz, H-7c); 6.02 (poorly resolved br d, J=10.5 Hz, H-9a); 6.63 (d, J=10.5 Hz, H-10 3); 7.43 (t, J=7.8 Hz, H-Ph-meta); 7.56 (t, J=7.8 Hz, H-Ph-para); 7.87 (d, J=7.8 Hz, H-Ph-ortho).
13C NMR (CDC13, 600 MHz, 5) Table 3-2: 11.8 (C-18 methyl, q); 12.9 (C-19 methyl, q); 20.8 (7-0 acetate methyl, q); 21.2 (5-0 acetate methyl, q); 21.4 (9-0 acetate methyl, q); 24.8 (C-17 methyl, q); 27.0 (C-16 methyl, q); 29.2 (C-2, t); 33.9 (C-6, t); 38.8 (C-3, d); 44.8 (C-8, s); 47.1 (C-14, dd); 63.0 (C-1, s); 69.7 (C-7, d); 70.7 (C-10, d); 74.1

(C-5, d); 75.6 (C-15, s); 76.9 (C-13, d); 77.9 (C-9, d); 114.0 (C-20, t); 128.7 (C-Ph-meta,

d); 129.2 (C-Ph-ipso, s); 129.4(C-Ph-ortho, d); 133.3 (C-Ph-para, d); 134.0 (C-11, s);

145.2 (C-4, s); 151.1 (C-12, s); 164.1 (CO-Ph, s); 169.9 X 2(CO-Acetate, s); 170.2 (OAcetate, s).

Analysis calculated for C 33H 42010: C, 66.20; H, 7.07. Found: C, 66.38; H, 7.19.






49


Brevifoliol-5,13-Diacetate [3-4]

The above reaction was repeated, except that it was heated at 80-900 C (water bath) for 30 min. After cooling, water was added and the solid filtered after 10 min. The solid was crystallized from acetone/ ligroin to give the diacetate as a colorless crystalline solid, yield, 0.2 g; m.p.241-243oC;

1H NMR (CDCI3, 600 MHz, 8) Table 3-1: 0.92, s (H-19); 1.11, s (H-16); 1.25 (dd, J=7.2, 13.8 Hz, H-14c); 1.35, s (H-17); 1.46 (d, J=14.1 Hz, H-2a); 1.75 (s, 9-0 acetate methyl); 1.90 (m, H-6a) 2.0 (m, H-63); 2.02 (s, 5-0 acetate methyl); 2.03 (s, 13-0 acetate methyl); 2.07 (s, 18 methyl); 2.08 (s, 7-0 acetate methyl); 2.41 (dd, J=14.1, 9.6 Hz, H-23); 2.51 (dd, J=7.2, 13.8 Hz, H-14P); 2.72 (d, J=9 Hz, H-3a); 2.74, brs (C-15 OH, exchangeable with D20); 4.92, s (H-20 A); 5.28, s (H-20 B); 5.39 (br s, J= H-5p3); 5.54 (t, J=7.2 Hz, H-133); 5.61 (dd, J=4.8, 11.4 Hz, H-7a); 6.09 (poorly resolved br d, J=10.5 Hz, H-9c); 6.65 (d, J=10.5 Hz, H-103); 7.43 (t, J=7.8 Hz, H-Ph-meta); 7.56 (t, J=7.8 Hz, HPh-para); 7.87 (d, J=7.8 Hz, H-Ph-ortho).
13C NMR (CDC3, 600 MHz, 5) Table 3-2: 11.9 (C-18 methyl, q); 12.9 (C-19 methyl, q); 20.7 (7-0 acetate methyl, q); 21.0 (13-0 acetate methyl, q); 21.2 (5-0 acetate methyl, q); 21.4 (9-0 acetate methyl, q); 24.8 (C-17 methyl, q); 27.0 (C-16 methyl, q); 29.1 (C-2, t); 33.9 (C-6, t); 38.8 (C-3, d); 44.8 (C-8, s); 44.1 (C-14, dd); 63.0

(C-1, s); 69.6 (C-7, d); 69.8 (C-10, d); 74.1 (C-5, d); 75.6 (C-15, s); 76.9 (C-13, d); 79.3

(C-9, d); 114.3 (C-20, t); 128.8 (C-Ph-meta, d); 129.1 (C-Ph-ipso, s); 129.5(C-Ph-ortho,

d); 133.4 (C-Ph-para, d); 136.4 (C-11, s); 145.2 (C-4, s); 147.3 (C-12, s); 164.1 (CO-Ph,

s); 169.6 (CO-Acetate, s); 169.9 (CO-Acetate, s); 169.91 (CO-Acetate, s); 170.5 (COAcetate, s).

Analysis calculated for C 35H 44011: C, 65.61; H, 6.92. Found: C, 65.68; H, 6.99.





50


Brevifoliol-13-Ketone [3-61

A solution of brevifoliol (0.2 g) in benzene was treated with MnO2 (manganese dioxide, 1 g, Fisher Scientific) and the mixture heated under reflux for 2 hours, at which time, the starting material was consumed and a slightly faster moving product was formed. The mixture was filtered, concentrated and applied to a small silica column (15 g) in dichloromethane. Elution with 1% acetone in dichloromethane gave the major product which was recovered by concentration as a colorless powder, yield, 0.12g. The 1H- and 13C NMR spectra of this faster moving product were quite poorly resolved and only gave usable results at temperatures below -10 o C. Recrystallization and further chromatography failed to improve this situation, and low temperature NMR experiments indicated that a rotameric equilibrium was responsible for the poor resolution seen in these spectra.

1H NMR (CDC13, 600 MHz, -40 'C, 5) major rotamer: 0.92, s (H-19); 0.98, s (H16); 1.35, s (H-17); 1.46 (d, J=14.1 Hz, H-2a); 1.78 (s, 9-0 acetate methyl); 1.90 (m, H6a) 2.0 (m, H-6p); 2.01 (s, 18 methyl); 2.02 (s, 7-0 acetate methyl); 2.32 (d, J=19.0, 14oc); 2.48 (d, J=19.0 Hz, H-14p3); 2.92 (unresolved, H-3x); 2.74, br s (C-15 OH, exchangeable with D20); 4.92, s (H-20 A); 5.28, s (H-20 B); 5.39 (br s, J= H-5p3); 5.54 (t, J=7.2 lz, H-1 3p); 5.61 (dd, J=4.8, 11.4 Hz, H-7x); 6.09 (poorly resolved br d, J=1 0.5 Hz, H-9L); 6.65 (d, J=10.5 Hz, H-10p3); 7.43 (t, J=7.8 Hz, H-Ph-meta); 7.56 (t, J=7.8 Hz, HPh-para); 7.87 (d, J=7.8 Hz, H-Ph-ortho).
13C NMR (CDCI3, 600 MHz, 6): 8.9 (C-18 methyl, q); 12.5 (C-19 methyl, q); 20.7 (7-0 acetate methyl, q); 21.0 (9-0 acetate methyl, q); 26.2 (C-17 methyl, q); 26.6 (C-16 methyl, q); 27.3 (C-2, t); 27.7 (C-6, t); 34.2 (C-3, d); 43.2 (C-8, s); 48.3 (C-14, t); 58.1 (C1, s); 70.8 (C-7, d); 71.0 (C-10, d); 73.2 (C-9, d); 75.6 (C-15, s); 111.4 (C-20, t); 128.8

(Ph-meta, d); 129.1 (Ph-ipso, s); 129.5(Ph-ortho, d); 133.4 (Ph-para, d); 136.4 (C-11, s);






51


145.2 (C-4, s); 144.3 (C-12, s); 163.1 (C-11, s); 165.4 (CO-Ph, s); 169.3 (CO-Acetate, s); 170.3 (CO-Acetate, s); 207.5 (C-13 Ketone, s).

Analysis calculated for C 31H 3809: C, 67.13; H, 6.91. Found: C, 67.48; H, 6.97. Dihydrobrevifoliol [3-71

A solution of brevifoliol (0.2 g) in ethyl acetate (10 ml) was hydrogenated in a Parr apparatus using Platinum oxide (0.05 g) for 16 hours. TLC revealed the formation of a slightly slower moving product. The mixture was filtered and the filtrate concentrated to dryness and purified by chromatography on silica gel in dichloromethane. Elution with 2% acetone in dichloromethane gave a minor product, which was not further investigated. The fractions eluted with 2-5% methanol in dichloromethane gave the major product, which was obtained as a colorless powder, yield, 0.1 g,

Reduction of the 4(20) double bond resulted in significant broadening of most peaks in the 1H- and 13C NMR spectra, but the appearance of additional signals from methyl group at C-20 and methylene at C-4 could be seen, as well as the loss of the two characteristic exocyclic methylene singlets.

Analysis calculated for C31H4209: C, 66.60; H, 7.60. Found: C, 66.89; H, 7.88. Brevifoliol Epoxide [3-81

A mixture of brevifoliol (0.3 g) and meta-chloroperoxybenzoic acid (MCPBA, 0.2 g) in toluene (15 ml) was heated under reflux for 30 min. After cooling, the mixture was diluted with ether, washed successively with aqueous sodium bisulfite, aqueous sodium bicarbonate and saline, and the organic layer concentrated to dryness. The solid was crystallized from acetone/ligroin, to give a colorless crystalline epoxide, yield, 0.15 g; m.p. 227-230 0C.





52


1H NMR (CDCI3, 600 MHz, -40o C, 5) : 1.01 (H-19, s); 1.09 (H-16, s); 1.24 (dd, J=7.2, 13.8 Hz, H-14c); 1.27 (H-17, s); 1.40 (H-2c, br d J=14.1 Hz); 1.76 (s, methyl, 9acetate); 1.80 (m, H-6oc ); 2.0 (m, H-63); 2.01 (s, H-18); 2.07 (s, 7-acetate methyl); 2.36 (dd, J=14.1, 9.6 Hz, H-2); 2.46 (dd, J=7.2, 13.8 Hz, H-14p); 2.67, br s (C-15 OH, exchangeable with D20); 2.64 (H-3c, br d, J=9 Hz); 2.72 (C-20, s); 3.59 (C-20, s); 4.20 (br s, H-5f); 4.46 (H-1 33, br d, J=7.2 Hz); 5.57 (H-7a, br d, J=4.8, 11.4 Hz); 6.05 (H9c, poorly resolved br d, J=10.5 Hz); 6.54 (H-10p, br d, J=10.5 Hz); 7.43 (Ph-meta, t, J=7.8 Hz); 7.56 (Ph-para, t, J=7.8 Hz); 7.87 (Ph-ortho, d, J=7.8 Hz).
13C NMR (CDCI3, 600 MHz, -400 C, 6): 11.9 (C-18 methyl, q); 12.9 (C-19 methyl, q); 20.7 (7-0 acetate methyl, q) ; 21.3 (9-0 acetate methyl, q); 23.9 (C-2, t); 25.0 (C-17 methyl, q); 27.1 (C-16 methyl, q); 34.1 (C-3, d); 34.4 (C-6, t); 45.4 (C-8, s); 46.4 (C-14, t); 50.1 (C-4, s); 60.2 (C-20, t); 62.4 (C-1, s); 69.5 (C-7, d); 70.5 (C-10, d); 71.7 (C-5, d);

75.8 (C-15, s); 76.7 (C-13, d); 77.1 (C-9, d); 128.7 (Ph-meta, d); 129.4 (Ph-ipso, s); 129.5(Ph-ortho, d); 133.2 (Ph-para, d); 134.3 (C-12, s); 151.3 (C-11, s); 164.3 (CO-Ph, s); 169.96 (CO-Acetate, s); 170.0 (CO-Acetate, s).

Analysis calculated for 0311H40010: C, 65.02; H, 7.04. Found: C, 64.72; H, 7.24. Ozonization of Brevifoliol: Brevifoliol-norketone [3-91

A solution of brevifoliol (1 g) in a 9:1 mixture of chloroform and methanol (25 ml) was cooled in a dry ice/ acetone bath and saturated with ozone produced by an ozonizer (Ozone Research and Equipment Co., Phoenix, AZ). After testing for the absence of the starting material by TLC, the mixture was removed from the bath and treated with dimethyl sulfide (1 ml) and let stand at room temperature for 2 h. It was then concentrated to dryness and applied to a silica column prepared in chloroform. Elution with 2% acetone in chloroform gave two bands, which were separated and the fractions concentrated separately.





53


The faster moving fraction [3-9] was obtained as a colorless, amorphous powder, yield, 0.25 g.

1H NMR (CDCl3, 600 MHz, -40 0 C, 6) major rotamer: 1.02, s (H-19); 1.05, s (H16); 1.35, s (H-17); 1.46 (d, J=14.1 Hz, H-2a); 1.78 (s, 9-0 acetate methyl); 1.98 (m, H6c); 2.0 (m, H-6p); 2.01 (s, 18 methyl); 2.02 (s, 7-0 acetate methyl); 2.32 (d, J=19.0, 14c); 2.48 (H-14, dd, J=14, 7.5 Hz); 3.19 (H-3a, d, J=10.2); 4.20 (H-53, br s); 4.48 (H13f3, br t, J=7.2 Hz); 5.78 (H-7ca, br m); 5.92 (H-9a, br); 6.52 (H-10p, br d, J=10.5 Hz);

7.43 (Ph-meta, t, J=7.8 Hz); 7.56 (Ph-para, t, J=7.8 Hz);7.87 (Ph-ortho, d, J=7.8 Hz).
13C NMR (CDCl3, 600 MHz, 8): 12.1 (C-18 methyl); 14.2 (C-19 methyl); 20.6 (7O acetate methyl); 21.2 (5-0 acetate methyl); 21.3 (9-0 acetate methyl); 24.8 (C-17 methyl); 27.0 (C-16 methyl); 29.1 (C-2); 33.9 (C-6); 34.8 (C-3); 46.4 (C-8); 46.4 (C-14);

62.0 (C-1); 69.9 (C-7); 70.7 (C-10); 72.0 (C-5); 75.9 (C-15); 76.9 (C-13); 77.3 (C-9);

128.8 (Ph-meta); 129.1 (Ph-ipso); 129.5(Ph-ortho); 133.4 (Ph-para); 134.4 (C-11); 144.5

(C-4); 147.3 (C-12); 164.3 (CO-Ph, s); 169.8 (CO-Acetate, s); 170.0 (CO-Acetate, s);208.0 (C-13 C=0).

Analysis calculated for C30H38010: C, 64.50; H, 6.86. Found: C, 64.68; H, 6.97.

The slower moving fraction was obtained as a colorless crystalline solid, yield, 0.3 g; m.p.225-232 0 C. It was found to be identical with the epoxide [3-8], described above.

Brevifoliol-4,20-Diol [3-10]

To a solution of brevifoliol (0.4 g) in pyridine (10 ml) was added osmium tetroxide (0.2 g) and the reaction mixture stirred for 1 h, after which time, the starting material was replaced by a much slower moving component. After decomposing the excess reagent with a solution of sodium bisulfite in pyridine, water and dilute sulfuric acid were added and the mixture extracted with dichloromethane. After concentration, the product was






54


placed on a silica column in dichloromethane. Elution with 2% methanol in dichloromethane gave the major band which yielded [3-8] as a white powder, final yield,

0.12 g.

Analysis calculated for C31H42011: C, 63.04; H, 7.17. Found: C, 62.88; H, 7.25. Saponification of Brevifoliol [3-11]

A solution of brevifoliol (1 g) in methanol (20 ml) was stirred with 1N potassium hydroxide (10 ml) for 1 h. TLC showed that the starting material was absent and very slow moving, non UV-absorbing component produced. The reaction mixture was passed through a small column or Amberlite-IR120 ( a sulfonic acid resin) in the H+ form. The column was washed with 1:1 methanol/water. The effluent and washes were concentrated to dryness and the solid crystallized from acetone to give [3-11] as a colorless crystalline solid, yield, 0.45 g; m.p. 290 0C dec.

Analysis calculated for C20H3206+H20: C, 62.15; H, 8.87. Found: C, 62.48; H,

8.99.

Debenzoyl Brevifoliol-Pentaacetate f3-121

Compound [3-11] (0.2 g) was acetylated using acetic anhydride (2 ml) and pyridine (0.5 ml) by heating at 80 o C for 30 min. Water was added to decompose the reagent and the solid filtered. It was crystallized from ether/ ligroin, yield, 0.2 g; m.p. 184-187 C.

1H NMR (CDCl3, 6): 6.36 (H-10, d, J=10.2); 5.88 (H-9, br d, 10.2); 5.54 (H-1313, t, J=7.2); 5.53 (H-7a, q, J=5.4, 10.2); 5.36 (H-5p3, br s); 5.26 (H-20b, s); 4.87 (H-20a, s);

2.65 (OH-15, s); 2.64 (H-3a, d, J=9); 2.48 (H-14p, dd, J=13.8, 7.2); 2.36 (H-2ca, dd, J=14.1, 9.3); 2.07 (AcMe, s); 2.06 (AcMe, s); 2.02 (AcMe, s); 2.00 (AcMe, s); 1.97 (AcMe, s); 1.95 (18-Me, s); 2.0 (H-6a, m); 1.85 (H-6p3, cm); 1.42 (H-2p, d, J=14.4); 1.31 (H-17 Me);1.22 (H-14o, dd, J=13.8, 7.2); 1.13 (H-16 Me); 0.88 (H-19 Me).






55


IR v max (KBr, cm1): 3565 (OH, sharp); 2960, 1740 (C=O); 1370, 1230, 1030.

MS(FAB): 580 [MH+].

Periodate Oxidation of [3-111 to [3-131

A solution of [3-11] (0.2 g) in methanol (5 ml) was treated with sodium periodate (0.3 g) in 1N sulfuric acid (2 ml). After 30 min, TLC showed that the starting material was absent and was replaced by a faster moving product visible under the UV light, unlike the starting material. After dilution with water, the mixture was extracted with ethyl acetate and the extract concentrated to dryness. The crude dialdehyde [5-13] was not further purified before the next reaction, but did exhibit signals for two aldehydes the 1H spectra. The only significant changes from the parent compound showed the loss of the isolated 1H spin system from the protons on C-9 and C-10, and the conversion of two hydroxyl carbons into aldehydes (Guthrie, 1961).

1H NMR (CDC3, 6): 9.96 (CHO, s); 9.42 (CHO, s); 5.36 (H-20b, s); 5.02 (H-20a, s); 4.58 (H-5a, m); 4.40 (H-1 3a, br t); 2.60 (H-313, d); 1.30 (Me, s); 1.25 (Me, s); 1.02 (Me, s); 0.88 (Me, s).

Analysis calculated for C30H42011: C, 62.27; H, 7.32. Found: C, 62.10; H, 7.44. Formation of Osazone [3-141 from [3-131 with 2,4-DNPH

Dialdehyde [3-13] was dissolved in methanol, (2 ml) and heated with a solution of 2,4-dinitrophenyl hydrazine (0.1 g) in 2N hydrochloric acid (2ml) and methanol (2 ml). After 2 h at room temperature, the mixture was extracted with chloroform and the concentrated extract chromatographed on a silica column. The major band [3-14] was obtained as an orange yellow crystalline solid, m.p. 215-218C. Both 'H- and 13C NMR spectra indicated the retention of all twenty of the carbons from the abeo-taxane skeleton.






56


Debenzoyl Brevifoliol [3-151

A solution of brevifoliol (0.3 g) in 30% methanol in water (10 ml) was heated in a sealed tube at 135 0 C for 90 min. The cooled mixture was neutralized with sodium bicarbonate and extracted with chloroform. The extract was purified by chromatography on a C-8 reverse phase column using a step gradient of 25-60% acetonitrile in water in 5% increments of solvent concentration. Elution with 30-35% acetonitrile in water gave the major product, which was obtained as a white powder, yield, 0.1 g. The 1H and 13C NMR spectra revealed the loss of the benzoate from the C-10 position and retention of the acetate substituents at C-9 and C-7

Analysis calculated for C24H3608: C, 63.70; H, 8.02. Found: C, 63.89; H, 7.88.

Benzoic acid was also isolated from the reaction mixture and confirmed using NMR and UV analyses.














CHAPTER 4
SOME UNUSUAL REACTIONS OF BREVIFOLIOL The isolation and structural elucidation of brevifoliol [3-1] was described in detail in Chapter 3. Brevifoliol occurs to the extent of 0.2-0.3 % in the needles of T. brevifolia, and to a lesser extent in the bark of the same species, in the needles of T. x media Hicksii (Rehd.) and in the needles of T. wallichiana. (Georg et al. 1993). Large quantities of the crystalline compound can be readily isolated from the needles of T. brevifolia, which is the best source for the compound.

However, other than acetylation to a diacetate (Balza et al. 1991); and the attachment of the N-benzoyl phenyl-isoserine side chain at the C-13 position (Georg, et al. 1993); no record of its reactions reflecting its various functional groups has been published. This paucity of such information is in keeping with the current trend that, in spite of the virtual explosion of new taxanes that were isolated and characterized structurally by spectral data over the past five years. Very few have been investigated for their chemical reactions, for the relative reactivities of similar functional groups, or for any unusual reactions as a consequence of their stereochemical disposition.

Understanding the products of the various transformations that a compound can be subjected to and the relative rates of reaction can lead to important insight concerning entity, and for the general advancement of chemistry. Although the structure of brevifoliol is now well established as a result of spectral and x-ray crystallographic data, we can gain insight into this molecule through reactions such as those described in Chapter 3. In addition to these, a number of reactions also carried out for the purpose of




57






58


structural elucidation of [3-1] took unexpected courses. The products of these reactions were isolated and characterized structurally, and the details are described here.

1. Acid-Catalyzed Acetylation

In Chapter 3, acetylation of [3-1] by means of acetic anhydride and pyridine to give the crystalline monoacetate [3-2] and the diacetate [3-4] has been described. We isolated from the needles of T. brevifolia another taxane, which resembled brevifoliol, giving a dark greenish blue color when sprayed with sulfuric acid and heated. This same product was also subsequently isolated by others (Barboni et al. 1993). It was similar to but different from the monoacetate [3-2], but when acetylated, gave [3-4], thus showing that it was the 13-acetate [3-3].

Acetylation of brevifoliol using acetic anhydride and BF3 gave a different crystalline product, with almost the same RF as that of the diacetate [3-4]. The BF3 product closely resembled the 5, 13-diacetate in its NMR spectral properties also, but with some differences. In the 1H NMR spectrum, the singlet at 52.74 which is assigned to the 15-hydroxyl in the diacetate [3-4], was no longer present. The two methyl singlets assigned to C-16 and C-17 at 61.11 and 81.35 were deshielded to 61.63 and 61.71. The other two (C-1 8 and C-1 9) showed either no shift or a much smaller one (6 2.03->8 1.96 for C-18). Thus, the significant downfield shift of the signals due to C-16 and C-17 suggested that acetylation of the 15-hydroxyl might have taken place.

Additionally, the signals due to the H-2a and 2P showed slight downfield shifts (H-2cL: 61.46, broad doublet, J=13 Hz to 1.53 ppm; H-2p: 62.41, doublet of doublets, J=9,13 Hz to 2.65 ppm). The spectra of brevifoliol and its acetates generally show the H-1 Oct signal as a sharp and well resolved doublet, but the one due to H-9P as a broad, poorly resolved doublet. In the BF3-reaction product, the signal due to H-9P was not only a well resolved doublet (J=10.8 Hz); but also shielded by 0.29 ppm to 6 5.80. These






59


observations suggest that the C-15 hydroxyl might be responsible for the blurring of the signal of H-913, and acetylation of this hydroxyl has eliminated that interaction.

Further support for the acetylation having taken place at the C-15 hydroxyl is shown by the appearance of another acetyl methyl signal at 82.11 in the 1H NMR spectrum. The appearance of two more signals in the 13C NMR spectrum was also evident, in which 5 acetyl methyl and 5 acetyl-carbonyl signals were present, with the fifth one at 8 21.7 and 6 169.5, respectively. Also in the 13C NMR spectrum, a significant downfield shift from 5 75.6 to 6 87.2 for the C-15, and an upfield shift from 6 27.0 to 6 24.8 and from 5 23.1 to 5 22.0 for the signals due to C-16 and C-17 respectively, completes the evidence to indicate that the 15-hydroxyl was acetylated to give [3-5].

A comparison of the 13C NMR spectral data of brevifoliol, the 5-monoacetate [3-2], the 13-monoacetate [3-3] (naturally occurring and confirmed through semisynthesis); the 5, 13-diacetate [3-4] and of the BF3-catalyzed acetylation product [3-5] were shown in Chapter 3 in Table 3-1.

2. Oxidation

Oxidation of brevifoliol with manganese dioxide gave the monoketone, the NMR spectral data of which showed that the 13-hydroxyl was oxidized, leading to the structure [3-6], as described in Chapter 3. Oxidation of [3-1] with Jone's reagent gave initially, the same 13-monoketone [3-6], but on further reaction, this was replaced by a faster moving compound [4-1], whose spectral data pointed to an unexpected course of reaction.

The molecular formula of the product, C31H3609 (MH+, 553) indicated the loss of 4 protons, as compared to brevifoliol. Although this might indicate that both hydroxyls were oxidized to give the diketone [4-2], certain features suggested otherwise. To begin with, the 1H NMR spectrum showed broad peaks which indicated the existence of a





60


rotameric equilibrium, which was confirmed by a spectrum taken at -40 0 C, in which two sets of peaks with a 5:1 ratio were seen.

In the major rotamer, the coupling between the C-9 and C-10 protons was found to be 4 Hz, in contrast to the value of 10.5 Hz shown by brevifoliol [3-1], (a similar diketone prepared from 2-acetoxy brevifoliol (taxchinin A); described by Fuji et al. (1992) and Appendino et al. (1993) also showed a coupling of 4 Hz. Next, a singlet appeared at 9.4 ppm, which interacted in the HETCOR spectrum with the peak at 194 ppm. The latter showed a negative signal in the Attached Proton Test (APT). These observations indicated the presence of an aldehyde functionality, presumably at C-20. Additionally, in the spectrum of the diketone such as [4-2], the C-5 proton signal was absent, as expected, and the exocyclic methylene protons appeared as two singlets at 8 5.06 and 8

5.94 ppm. The corresponding carbon signals appearing at 5127 and 6143.4 ppm.

However, in the Jones oxidation product, the singlets due to the exocyclic methylene protons were absent, and the characteristic C-20 carbon signal which appears in the 8110-6120 ppm region was also missing. Instead, a signal was found at 6 6.73 ppm, which interacted with the signals at 82.5 and 62.7 (C-6 protons); and in the HETCOR spectrum, with the signal present at 8 147 ppm, and this latter gave a negative signal in the APT spectrum. These data seem to suggest that the product is not the 5,13-diketone [4-2], but an aldehydic product with a double bond present at C-4/C-5, as shown in [4-1].

One possible explanation is that the sulfuric acid in the reagent caused hydration of the 4/20 double bond, and the primary alcohol so generated was oxidized to the aldehyde, followed by dehydration to yield the 4/5-double bond.





61



o 10
oI 0/
o 5


O ...... H 0
H 20
20 U
HO IH 0 HO


[4-11 Jones Oxidation Aldehyde [4-2] MnO2 Diketone Figure 4-1 Oxidation Products


3. Action of BF3 on Brevifoliol [4-31

Before the structure of brevifoliol was fully established, the possibility that the two hydroxyls present in the compound might be vicinal to each other was considered. With the aim of forming an isopropylidene derivative, brevifoliol was reacted with acetone in the presence of Dowex-50 (sulfonic acid resin) as the catalyst. Reaction took place readily with the formation of a faster moving product, with some decomposition also taking place (colors). The reaction proceeded with less decomposition when BF3etherate was used as the catalyst.

Chromatography of the mixture from either reaction yielded the major product as a colorless crystalline solid. Its NMR spectrum quickly ruled out the possibility of its being an isopropylidene derivative. The FAB-mass spectrum gave a value for the MH+ m/z of 481, as compared with 557 for brevifoliol, thus showing a loss of 76 mass units. The elemental analysis, which agreed with C28H3207 showed a loss of C3H802 compared to brevifoliol. This may be interpreted as the loss of the C3H70- side chain attached to C-1, as well as that of H20, possibly through the elimination of the C-1 3-OH (or C-5-OH). The evidence to support this assertion is given below.





62


1. The most striking evidence is that only two of the four methyl signals that are seen in

brevifoliol appear in the BF3-reaction product. This is not only observed in the 1H

spectrum, but also in the 13C NMR thus showing that the oxy-isopropyl chain at C-1

is not present.

2. An examination of the isolated spin system: 5P 6at- 63 7u. in the 'H NMR spectrum

of brevifoliol and that of the product of BF3 reaction indicated that the free hydroxyl at

C-5cx was still present, but that the C-1 3c-OH was absent. Through a study of the interactions in the COSY spectrum, it was possible to assign (and distinguish) the

signals for C-6 and C-14 in the region below 83.

3. A new methine proton signal appeared as a broad singlet at 65.84, assigned to H-13,

which interacted in the HETCOR spectrum with the signal at 6124.1, which

corresponds to C-13.

4. The DEPT spectrum showed the presence of four signals for methyl-type (CH3)

carbons (611.4, 8 13.5 for C-19 and C-18, 5 20.7, 6 21.2 for the two CO-Me); two

signals for methylene-type (CH2) carbons (6 112.4 for C-20, 8 44.6 for C-14, 6 34.1

for C-6 and 6 27.9 for C-2, nine signals for methine-type (R3CH) carbons (6133.0,

6129.7, 6128.4 for the five aromatic carbons, four aliphatic CH-OR type carbons [673.2 (C-9); 672.5 (C-5); 879.6 (C-7); 667.4 (C-10)], one aliphatic methine type

carbon (639.6 (C-3)); one unsaturated methine type carbon (6124.1 (C-13) and nine quaternary carbons (three carbonyls, one aromatic C carrying the carboxyl, C-1 (?);

C-4, C-11, C-12 and C-8); together, account for the 28 carbons present..

5. HETCOR interactions supported the assignments for the o, m, p- positions in the

benzoate, and for the C-13, C-20, C-9, C-5, C-7, C-10, C-3, C-6, C-2, C-18, C-19

and the two acetate methyls.





63


6. In the 1D nOe-difference spectrum, the interaction between the C-14ca,p and C-13 f

protons was the strongest. Crowding of the region around 52.8 ppm made the

spectrum more difficult to interpret, and not very informative.



Table 4-1 : NMR Spectra of Compound [4-3] from BF3 Reaction
Position Proton Carbon APT DEPT
1 ---- 145.9 C
2 cc 1.96 m 28.0 T OH2
2 13 3 .0 0 m ----.. ...
3 c. 2.70 m 39.4 CH
4 ---- 150.4 T C
5 (x 4.42 br s 72.6 CH
6co 1.76 m 34.1 T CH2
6 3 2 .05 m ----.. ...
7 ox 5.48 br d 70.6 CH
8 ---- 46.6 C
913 5.34 d 6.0 73.2 4 CH
10 c 6.30 d 6.0 67.5 CH
11 ---- 146.4 C C
12 ---- 134.2 C
13 5.83 brs 124.0 4 CH
14 c, 3 2.85-3.0 cm 44.6 CH2
18 2.06s 11.4 CH3
19 1.23s 13.4 7, CH3
20A 4.98 br s 112.2 T CH2
20 B 5.20 br s ----. ..
CO-C6H5 ---- 165.2 T C
Bz-ipso ---- 130.2 T C
Bz-ortho 8.01, d 7.5 128.4 CH
Bz-meta 7.44, t 7.5 129.6 T CH
Bz-para 7.56, t 7.5 132.9 CH
COCH3 1.98s 20.6 CH3
2.00s 21.1 OH3
COCH3 ---- 169.9 1 C
170.4 C

1H NMR were recorded at 600 MHz and 13C NMR at 150 MHz in CDC13 on a Varian Unity 600 spectrometer at ambient temperature. Chemical shifts 6 (ppm) are reported relative to TMS as an internal standard.






64


Bz Ac
S O Ac



OH
13 O,




[4-3] BF3 -etherate Product Figure 4-2 : BF3-etherate Catalyzed Elimination Product

Based on this reasoning, the structure of the BF3-reaction product was assigned as shown [4-3] in Figure 4-2 above. The DEPT spectra of [4-3] are given in Figure 4-3.



CH3 Carbons 2xAc] 1819

CH2 Carbons
20 14 6 2




CH Carbons
Acr 1 57
139 10 3 DMSO




AdI Protonated Carbons




I' ' 1 l I I ; I I 1 7 1 F 7 1F t' 1- 1 I i T i i ; I -l 1 1 1 l 1 1 t T F 1 1- 1 i 1 1 1 1 l 1 1-i T i
140 20 100 6 040 20 0


Figure 4-3 : DEPT Spectra of BF3 Elimination Product [4-3]






65


A reaction such as this has not been reported in the taxane series, resulting in the loss of the oxy-isopropyl side chain. In taxol and related compounds containing the conventional taxane skeleton, action of Lewis acids such as BF3 was studied and is shown to produce one or two different changes, depending on the whether protic or aprotic solvent is used. In one case, isomerization of the A-ring from a 6- to a 5membered A ring takes place with the oxy-isopropyl group attached at C-1. In the second instance, the oxetane ring is opened to form a diol, or with the acetate group migrating from C-4 to C-20 to give the 4-hydroxy-20-acetoxy compound. The reaction described here appears to be a continuation of the action of the Lewis acid on the 5membered A ring, with the elimination of the oxy-isopropyl substituent.

4. Reaction with Iodine/Silver Acetate [4-41

Once the structure of brevifoliol had been elucidated, the value and usefulness of this relatively abundant compound in the needles of T. brevifolia was considered. Since the addition of the N-benzoyl isoserine side chain at C-13 did not generate activity in the final product, it was reasoned that the oxetane ring at the 4/20 position might be necessary to produce activity. To this end, one approach was investigated, involving the use of iodine in some form to add across the 4/20 double bond and thereby permitting substitution with other groups.

Brevifoliol was found to react readily with bromine, but the reaction yielded multiple products and considerable decomposition. Reaction with iodine was similarly complex and led to much decomposition and dark colored products. With the idea that addition of a silver salt which can remove the acid that might be produced, but not be too strongly basic (e.g. silver oxide) and hence hydrolyze the ester functions, silver acetate was selected for use with the iodine. The remote possibility that silver acetate might





66


displace the iodine located at C-4 to produce the 4-acetoxy-20-iodo compound was also attractive (Woodard & Brutcher, 1958).

When brevifoliol was stirred with iodine and silver acetate, the course of the reaction was clearly different. No multiple products or decomposition to dark colored products was seen, even when the reaction was continued for 15-30 hours, unlike the reaction with iodine alone that became dark in 1-2 hours. The reaction was continued until the starting material was consumed and a major, faster moving compound was produced. Chromatography on a silica column gave a colorless crystalline solid.

Acetylation with acetic anhydride and pyridine at 700 C for 15 minutes produced the 5-0-monoacetate [4-5], confirming the presence of the 5 hydroxyl with NMR analysis. Treatment of [4-4] with n-Bromosuccinimide produced the 5-0 ketone, further confirming the absence of a free C-13 hydroxyl on the basis of NMR and UV spectral data.






OH- 0 0O
o 0 0 0 -0 0



...... OH ....0 4"

0 0



[4-4] Iodine/Silver Acetate Roduct [4-5] Acetate



Figure 4-4 : Iodine/Silver Acetate Product [4-4] and Acetate





67





S14/14
a



2/2

13/14

20/20

7/6










10 9 20 20 215-OH
[P 7 13 !6)



6 4 2PP



Figure 4-5: H,H-COSY Spectrum of [4-4] The 1H NMR spectrum was similar to brevifoliol in most respects, with a few revealing differences. The AB quartet normally seen from H-2P3 became a clean doublet (J=1 3.8 Hz) deshielded from 6 2.36 to 6 3.31, and coupled to the H-2a. proton, which was deshielded from 6 1.49 to 6 2.34. This deshielding and the coupling patterns indicated the loss of H-3P with the possible formation of a double bond between C-3 and C-4.






68








8 7 6 5 4 3 2 1 0


18192x Ac dp CU
17 -= 16
6

14









7-C
5
-O







13

0
10,o L,






Figure 4-6: HETCOR Spectrum of [4-4]


Recorded on Varian VXR 300 spectrometer at 75.432 MHz in CDC13 with TMS as internal standard. F1: 3000.3 Hz; F2: 15528 Hz; Acquisition time 65.9 sec; Dj: 1 sec; Ambient temperature; Decouple proton; Level 70 high power; PW: 90'; 128 repetitions X 128 increments; Waltz 16 modulation; pseudo echo; FT size: 2K X 512 data points; time:
5 hours.


Elemental analysis and FAB-MS gave the molecular formula C31H3608, which indicated essentially that loss of one molecule of water and dehydrogenation had taken






69


place. The presence of a hydroxyl was indicated by the fact that the compound would still undergo acetylation to form a monoacetate.

The C-15 hydroxyl was still present in both the carbon and proton spectra, and the only other significant change in the spectra occurred with the C-20 signals. The two broad singlets normally seen around 5 ppm were replaced by an isolated spin system of doublets at 4.35 and 3.92 which resemble the H-20 oxetane pattern seen in taxol.

Experimental


Brevifoliol Triacetate [3-51

To a solution of brevifoliol (0.2 g) in acetic anhydride (2 ml) was added 0.2 ml of a 2% solution of boron trifluoride etherate in acetic anhydride to give a final concentration of 0.2% of boron trifluoride etherate. After 20 min at room temperature, the mixture was diluted with water. After another 10 min. the solid was separated, washed, taken up in ether and concentrated to dryness. The solid was crystallized from ether in ligroin to give [3-5] as a colorless crystalline solid, yield, 0.2 g; m.p. 214-216 o C. Oxidation with Jones Reagent to [4-11

A solution of brevifoliol (0.2 g) in acetone (10 ml) was treated with Jones reagent (2 ml) added in small portions and with stirring. Initially TLC analysis of the reaction mixture showed that a yellow color giving spot appeared above that of the starting material. Gradually the first product changed into an even faster moving component. When this latter was the predominant product, the reaction was stopped by the addition of water and extraction with chloroform. After concentration of the solvent, the product was chromatographed on a silica column in 1:1 chloroform/ ligroin. Elution with chloroform gave the major component, which was obtained as a colorless crystalline solid, yield, 0.05 g; m.p. 234-236 0 C.





70


Action of Boron Trifluoride on Brevifoliol [4-3]

Brevifoliol (0.3 g) was dissolved in acetone (10 ml) and to the solution was added 1 ml of 1% boron trifluoride etherate in acetone to make a 0.1% overall concentration of boron trifluoride in the reaction mixture. After 3 h, water was added, the solid filtered and after drying, subjected to chromatography on silica gel in chloroform/ ligroin (1:1). The major band obtained with the same solvent was crystallized from ether/ligroin, yield,

0.1 g, m.p. 162-165 a C.

1H NMR (CDC3, Varian Unity 600 MHz, 5): 1.16, s (H-19); 1.03, s (H-16); 2.08 (cm, H-14a); 1.27, s (H-17); 2.32 (d, J=13.8 Hz, H-2ca); 1.92 (s, methyl, 9-acetate); 1.84 (m, H-6a ); 1.98 (m, H-6p3); 2.26 (s, H-18); 2.05 (s, 7-acetate methyl); 3.30 (d, J=13.8 Hz, H-23); 2.28 (cm, H-14p); 2.95, br s (C-15 OH, exchangeable with D20); 4.38 (t, J=7.2 Hz, H-13P); 4.28 (br s, H-5p3); 4.54 (d, J=13.2 Hz, H-20 B); 4.54 (d, J=13.2 Hz, H-20 B); 5.51 (dd, J=4.8, 11.4 Hz, H-7ca); 6.07 (d, J=10.1 Hz, H-9ca); 6.54 (d, J=10.1 Hz, H-100); 7.45 (t, J=7.5 Hz, H-Bz-meta); 7.57 (t, J=7.5 Hz, H-Bz-para); 7.89 (d, J=7.5 Hz, H-Bzortho).
13C NMR (CDCl3, Varian VXR 300 MHz, 5): 11.4 (C-18 methyl, q); 13.4 (C-19 methyl, q); 20.6 (7-0 acetate methyl, q); 21.1 (9-0 acetate methyl, q); 28.0 (C-2, t); 34.1 (C-6, t); 39.4 (C-3, d); 44.6 (C-14, t); 46.6 (C-8, s); 67.5 (C-10, d); 70.6 (C-7, d); 72.6 (C5, d); 73.2 (C-9, d); 112.2 (C-20, t); 124.0 (C-13, d); 128.4 (C-Bz-ortho, d); 129.6 (C-Bzmeta, d); 130.2 (C-Bz-ipso, s); 132.9 (C-Bz-para, d); 134.2 (C-12, s); 145.9 (C-1, s); 146.4 (C-11, s); 150.4 (C-4, s); 165.2 (CO-Ph, s); 169.9 (CO-Acetate, s); 170.4 (OAcetate, s).

Reaction with Iodine and Silver Acetate [4-41

To a solution of brevifoliol (0.5 g) in benzene (15 ml) were added iodine (0.7 g) and silver acetate (0.75 g) and the mixture stirred at room temperature for 20 h. TLC





71


showed that the starting material was absent and was replaced by two faster moving components. The mixture was filtered and the filtrate washed successively with aqueous sodium bisulfite and water and concentrated to dryness. Chromatography on silica gel in 4:1 chloroform/ligroin gave the major band, which was obtained as a colorless crystalline solid, total yield, 0.12 g; m.p. 250-252 o C.

1H NMR (CDCl3, Varian Unity 600 MHz, 5): 1.16, s (H-19); 1.03, s (H-16); 2.08 (cm, H-14ca); 1.27, s (H-17); 2.32 (d, J=13.8 Hz, H-2a); 1.92 (s, methyl, 9-acetate); 1.84 (m, H-6a ); 1.98 (m, H-6p3); 2.26 (s, H-18); 2.05 (s, 7-acetate methyl); 3.30 (d, J=13.8 Hz, H-213); 2.28 (cm, H-14P3); 2.95, br s (C-15 OH, exchangeable with D20); 4.38 (t, J=7.2 Hz, H-13P3); 4.28 (br s, H-5p); 4.54 (d, J=13.2 Hz, H-20 B); 4.54 (d, J=13.2 Hz, H-20 B); 5.51 (dd, J=4.8, 11.4 Hz, H-7ca); 6.07 (d, J=10.1 Hz, H-9c); 6.54 (d, J=10.1 Hz, H-103); 7.45 (t, J=7.5 Hz, H-Bz-meta); 7.57 (t, J=7.5 Hz, H-Bz-para); 7.89 (d, J=7.5 Hz, H-Bzortho).
13C NMR (CDC13, Varian VXR 300 MHz, 8): 13.1 (C-18 methyl, q); 16.1 (C-19 methyl, q); 20.8 (7-0 acetate methyl, q); 21.5 (9-0 acetate methyl, q); 25.3 (C-17 methyl, q); 26.9 (C-16 methyl, q); 31.7 (C-2, t); 38.2 (C-6, t); 142.7(C-3, d); 45.2 (C-8, s); 38.6

(C-14, t); 65.6 (C-1, s); 67.2 (C-7, d); 76.2 (C-10, d); 69.8 (C-5, d); 74.3 (C-15, s); 84.0

(C-13, d); 72.4 (C-9, d); 64.4 (C-20, t); 128.8 (C-Bz-meta, d); 129.2 (C-Bz-ipso, s);

129.4(C-Bz-ortho, d); 133.4 (C-Bz-para, d); 135.5 (C-12, s); 142.7 (C-4, s); 146.4 (C-11,

s); 164.5 (CO-Ph, s); 169.4 (CO-Acetate, s); 170.3 (CO-Acetate, s).

FAB-MS (dithiothreotol/dithioerythrotol / TFA, m/z): 577 [M+Na]; 537 [M NaOH]; 433 [M+-NaO2C7H5]; 373 [M+-NaO2C7H5 -HOAc]; 313 [M+-NaO2C7H5 2x HOAc]; 253 [M+-NaO2C7H5-3x HOAc].

CI-MS (methane, m/z): 537.9 [MH' -H20]; 373.6 [MH'-H20-HOAc- C6HsCOOH].





72


Acetylation of [4-41 to [4-51

A sample of [4-4] (.05 g) was acetylated in acetic anhydride (2 ml) and pyridine (0.5 ml) at room temperature for 20 h. After addition of water, the solid was filtered and crystallized from ether in ligroin, m.p. 250-254 o C.

1H NMR (CDCl3, Varian Unity 600 MHz, 8): 1.04 (s, H-16); 1.19 (s, H-19); 1.28 (s, H-17); 1.91 (s, methyl, 9-acetate); 1.77 (br dd, H-14o); 2.02 (m, H-6p); 2.05 (s, 7-acetate methyl); 2.27 (s, H-18); 2.19 (s, 5-acetate methyl); 2.22 (cm, H-1413); 2.34 (d, J=13.8 Hz, H-2ca); 3.01, (br s, C-15 OH, exchangeable with D20); 3.31 (d, J=13.8 Hz, H-23); 3.90 (d, J=13.2 Hz, H-20 B); 4.35 (d, J=13.2 Hz, H-20 A); 4.50 (br m, H-13P3); 5.41 (dd, J=3.6, 13.2 Hz, H-7a); 5.46 (d, J=4.2 Hz, H-513); 6.04 (d, J=10.1 Hz, H-9o); 6.57 (d, J=10.1 Hz, H-1013); 7.45 (t, J=7.5 Hz, H-Bz-meta); 7.57 (t, J=7.5 Hz, H-Bz-para); 7.89 (d, J=7.5 Hz,

H-Bz-ortho).

13"C NMR (CDC3, Varian VXR 300 MHz, 5): 12.9 (C-18 methyl, q); 16.1 (C-19 methyl, q); 21.7 (7-0 acetate methyl, q); 21.5 (9-0 acetate methyl, q); 21.7 (5-0 acetate methyl, q); 25.3 (C-16 methyl, q); 27.0 (C-17 methyl, q); 31.5 (C-2, t); 32.9 (C-6, t); 38.3 (C-14, t); 45.2 (C-8, s); 63.5 (C-1, s); 65.4 (C-20, t); 67.6 (C-7, d); 70.3 (C-5, d); 72.3 (C9, d); 74.4 (C-15, s); 76.3 (C-10, d); 83.2 (C-13, d); 127.3 (C-4, s); 128.8 (C-Bz-meta, d);

129.1 (C-Bz-ipso, s); 129.4(C-Bz-ortho, d); 133.4 (C-Bz-para, d); 135.5 (C-12, s);

142.7(C-3, d); 145.6 (C-3, s); 147.6 (C-11, s); 164.6 (CO-Ph, s); 169.6 (CO-Acetate, s); 170.3 (CO-Acetate, s); 170.8 (CO-Acetate, s). Reaction with N-Bromosuccinimide and Silver Acetate [4-6]

A solution of brevifoliol (0.2 g) in benzene (10 ml) was stirred with Nbromosuccinimide {NBS} (0.125 g, recrystallized from water). After 2 h the starting material was absent with two faster moving compounds being present. To the reaction mixture was added silver acetate (0.125 g) and stirred for another 2 h.





73


At this point, the previous major compound moved further to give a new product. The mixture was filtered, the filtrate washed with aqueous sodium bisulfite, followed by water and concentrated to dryness. The product was chromatographed on a silica column in 4:1 chloroform/ ligroin. The major component was obtained as a colorless crystalline solid, yield, 0.1 g. The compound was found to be identical with the product obtained from the reaction of brevifoliol with iodine and silver acetate. Reaction of [4-41 with N-Bromosuccinimide [4-71

A solution of [4-4] (0.04 g) in benzene (5 ml) was stirred with Nbromosuccinimide (25 mg) at room temperature. After 2 h, TLC showed formation of a slightly faster moving compound, which was separable from the starting material only after 2 or 3 developments of the TLC plate. The reaction mixture was washed with aqueous sodium bisulfite, followed by water and concentrated to dryness. The product was crystallized from ether in ligroin, m.p. 185-188 'C.















CHAPTER 5
TAXANE CONSTITUENTS OF THE NEEDLES OF TAXUS X MEDIA


As discussed in Chapter 3, the yield of taxol from the bark of Taxus brevifolia by using the conventional methods of isolation was of the order of 0.01%. It was also shown that through the use of these same methods, no other useful analogues could be isolated in any significant yields. As a consequence of these results and strong ecological considerations, an intense search was started with the aim of finding a source that is renewable, and which can match the bark in the yield of taxol. Many of the available species of Taxus, as well as the various parts of these plants were examined through the use of analytical high performance liquid chromatography (HPLC) and thin layer chromatography (TLC). These searches led to the selection of the needles of the ornamental yew, Taxus x media Hicksii as a possible answer to the problem. The ornamental yew is capable of being grown in a nursery type setting, and on a large scale, so that the needles may be clipped twice a year, and the taxol, which is found to be present to the extent of 0. 0 1% be isolated from them.

At the time of this research (1992-93), almost all of the studies carried out on this species consisted of HPLC analyses. Other than the isolation of taxol by the standard procedure with a total yield of 0.006%, no information had been published either on the taxane constituents, or even a method for the practical isolation of them. In these HPLC analyses, it was recognized that in the extracts of the ornamental yew, taxol was accompanied by other co-eluting taxanes and these could contribute some errors in the total yield calculations. These co-eluting taxanes were isolated in minute yields, in the form of two components (0.8 mg and 1.2 mg); each representing an equilibrium mixture


74






75


of two components. On the basis of NMR spectral evidence, structures were assigned to these two components (Castor & Tyler, 1993).

Due to the presence of pigments, waxes and other impurities, the isolation of taxol and other taxanes from the needles was expected to be more difficult when compared to their isolation from the bark of T brevifolia. A project was started in this laboratory to meet the need for a practical method for the isolation of taxol and other related taxanes from the bark and needles of various Taxus species in spite of these challenges. The application of a preparative scale reverse phase column chromatography technique proved to be surprisingly successful in the processing of the extracts of T. brevifolia.

To begin with, the HPLC analysis of the extract of the needles of the ornamental yew, as shown in Figure 5-2, clearly shows that taxol is accompanied by several major taxane components, which are present in much higher concentrations than taxol itself. In view of such relatively high concentrations of these components, it is surprising that only such minute amounts of two of these mixtures could be isolated earlier, as indicated above. Also, no other characterizing data were provided other than the spectral data. This laboratory's objective was the development of a simpler procedure for the isolation of taxol with potential for large-scale use, in addition to more fully characterizing the major taxanes present in the extract. The needles of the ornamental yew (200 lbs., dried) were received through the courtesy of Hauser Company, Boulder, CO, during May-June 1993.

The extraction was carried out three or four times using methanol and the extract concentrated to a syrup. The resulting concentrate was then partitioned between water and chloroform, and the organic layer containing the taxane fractions was concentrated to a thick semi-solid mass, which was used directly in the next step.





76


The reverse phase column procedure was carried out similar to what was used with the needle extract of T. brevifolia, as described in Chapter 3. Approximately 200 g. of the chloroform extract was dissolved in acetonitrile (see experimental) and stirred with the equilibrated C-18 bonded silica. This slurry was then diluted to the appropriate concentration of the acetonitrile and the added to the column prepared from 800 g of the C-18 silica. Elution was carried out using a step gradient: 30, 35, 40, 45, 50 and 60% acetonitrile in water, and the eluate collected in fractions of 200 ml. As was seen in the case of the columns on the bark extract of T. brevifolia, when the fractions remained at room temperature for about a week, crystals began to separate from the fractions in different regions of the elution. These were filtered and further purified by either recrystallization or a small column of normal phase silica where necessary.

The progress of elution of the column is shown in Figure 5-3. As anticipated, taxol was accompanied by two other taxanes, which were present in higher concentrations than taxol. However, all of these crystallized out of the fractions.

The early fractions contained the bulk of the UV absorbance, and from these could be isolated a crystalline solid, which was a non-taxane compound. The next major component that emerged with the 35-40% acetonitrile in water was shown to be brevifoliol as described in Chapter 3. With the 45-50% acetonitrile and water solvent were eluted taxane I, taxane II, followed by taxol, all of which crystallized from their respective fractions, with some overlap.

The column was finally washed with a mixture of methanol and ethyl acetate/ligroin (2:1:1) which stripped the column of all the waxes, chlorophylls and other pigments. After, washing with methanol, followed by 25% acetonitrile and water the column was made ready for another run. Figure 5-1 shows the steps involved in the fractionation of the extract of Taxus x media Hicksii.





77




Dried Needles of Taxus x media Hicksii





Extract Residue
(Discard)





Chloroform Aqueous




"Extract Solids"


Reverse Phase Column Filter Crystals Recrystallize or chromatograph




Brevifoliol Taxanes I and II Taxol Taxane Ill Taxane IV





Ozonization Chromatography Taxol


Figure 5-1 Fractionation of the Extract of Taxus x media Hicksii Needles






78



























Taxus fioridana Taxus x media Hicksii



Figure 5-2 : HPLC Trace of Taxanes Coeluting with Taxol.




Column Elution, Absorbance vs. Time

6


Es__






0
V o N ( t 0 4 W C e c l Cl 0 C4 t
FRCIO UME


Fiur 5- Prores of___ EltofTxnsfo ees hs oun





79


The reverse phase column run on 600 g of the extract obtained from 12 Kg of the dried needles was applied to a column prepared from 3 Kg of the C-18 bonded silica. The yields of the products obtained were good and important values are listed below. Brevifoliol [3-11

The fractions containing this component were combined, concentrated to dryness and chromatographed on a normal phase silica column. The major component was obtained as a colorless crystalline solid, which was found to be identical with brevifoliol [3-1] on the basis of its spectral data.

Taxanes 1 [5-11 and II [5-21

Additional chromatography of the mixture of the taxanes I and II and taxol on a normal phase silica column gave some separation of the two taxanes. Although they could be further separated and obtained as crystalline solids, they still represented an equilibrium mixture, as was indicated by the 1H- and 13C NMR spectra of the individual crystalline samples. From the spectral data, these two were identified as a mixture of 5-0-cinnamoyl-10-acetyl taxicin I [5-1] and 5-O-cinnamoyl-9-acetyl taxicin I [5-2], which were isolated (Chmurney et al. 1993) from the needles of T. x media Hicksii and from the needles of T. baccata (Appendino et al. 1992). The former authors obtained them in quantities not sufficient for physical properties, and the latter authors obtained them as amorphous powders, by using HPLC and preparative TLC.

The mixture of [5-1] and [5-2] on acetylation with acetic anhydride and pyridine gave the triacetate [2-1], which was obtained as a crystalline solid and was found to be a single entity unlike the starting material. It was also identical with the taxane III (see below).






80


Taxane III [2-11

The crude crystalline solid obtained from the reverse phase column was recrystallized. Its spectral and analytical data agreed with those given for 5-0cinnamoyl-2a,9cx, 10 3-triacetyl taxicin I (Appendino et al. 1993; Baxter et al. 1962).



Taxane IV [2-21

This was also purified by recrystallization of the crude crystals obtained directly from the fractions. It was found to be identical with 5-0-cinnamoyl 2a, 9c, 1013-triacetyl taxicin II, described by Appendino et al. (1992) and Baxter et al (1962). Taxol [5-31

The chromatography using normal phase silica column as described under taxanes I and II yielded taxol, which was purified by crystallization. The sample was still contaminated with some of the taxanes I and II. For complete purification, the mixture was subjected to ozonolysis which converted these two taxanes to more polar compounds from which taxol could be readily separated and obtained pure. Using this method, taxol was obtained in a yield of 0.015% based on the dry needles. This was significantly better than the reported yield of 0.006% (Witherup et al. 1990). Ozonolysis of [2-21

Because of the presence of the cinnamoyl ester function in compounds [5-1], [5-2], [2-1] and [2-21, they all undergo ozonolysis. This method gives a convenient way of separating taxanes [5-1] and [5-2] from taxol, with which they co-elute. In order to determine the nature of the product of ozonolysis, taxane [2-2] was subjected to this reaction and the product recovered and obtained as a crystalline solid. Its NMR spectral characteristics indicated a hydrated aldehyde with the structure shown in [5-6].






81


Thus, in summary, dried needles of Taxus x media Hicksii were extracted and the total chloroform extract applied to a 0-18 reverse phase column. A number of components were separated, such as brevifoliol, taxanes l-V and taxol, which crystallized out directly from the fractions. Separation of taxol from taxanes I and 11 could be carried out directly by ozonolysis of the mixture, followed by chromatography on either a normal phase or reverse phase silica column.

Experimental


Extraction:

Dried needles of Taxus x media Hicksii (50 Ibs) were extracted with methanol as described in Chapter 3. The combined concentrate was partitioned between water and chloroform (10 gallons each). The organic layer was separated and the extraction repeated twice more using 5 and 3 gallons of the solvent, respectively.

The combined chloroform layers were concentrated under reduced pressure to yield a dark green semi-solid, representing approximately 5% of the weight of the dried needles.

Chromatogiraphy:

Approximately 800 g of 0-18 bonded silica gel was poured into a glass MichellMiller type column (2.5 x 24") using methanol (Ace Glass, 1430 North West Blvd., Vineland, NJ 08360). The column was equilibrated with 25% acetonitrile in water. The chloroform extract solids (200 g) was dissolved in acetonitrile (400 ml) in a 4 L stainless steel beaker, by warming in a hot water bath. To this was added approximately 200 g equivalent of the equilibrated 0-1 8 bonded silica (20-25% of the column material). While the mixture is being stirred vigorously, 25% acetonitrile and water 500 ml was added, followed by water (approximately 800 ml). After stirring for 15 min. it was checked for






82


uniformity of the slurry and the absence of oily or waxy material, or lumps. The slurry was filtered under gentle suction and the solid was resuspended in approximately 500 ml of the filtrate to give a thin enough slurry for pouring. It was added to the column, the container rinsed and the rinse transferred to the column.

The remainder of the filtrate was pumped onto the column using a metering pump (Pulsa 680, Pulsafeeder Inc., Rochester, NY). After the sample addition was completed, elution was started using 30% acetonitrile and water. This was followed by 35, 40, 45, 50 and 60% acetonitrile and water. The column was then washed with 100% methanol. Final washing of the column with a mixture of methanol and ethyl acetate and ligroin removed the green pigments and other lipid-soluble components.

Fractions of 200 ml volume were collected and monitored by UV absorbance, TLC and analytical HPLC. After this, those fractions that contained significant UV absorbance and! or components detectable by TLC or HPLC were set aside for 7-10 days, whereby crystals began to appear from a number of fractions. These were filtered in groups, characterized and treated appropriately, as described below.

Characterization of the Taxane Components of Taxus x Media Hicksii Brevifoliol [2-11

Fractions from the 40% acetonitrile and water were concentrated to dryness, the solid taken up in chloroform and applied to a column of normal phase silica (40 g). Elution with 2-5% acetone in chloroform gave the major band. The fractions that contain this component were combined, concentrated to dryness and the solid crystallized from acetone in ligroin. The crystalline product, yield, 0.8 g (0.02%) m.p. 220-222 0 C was found to be identical on the basis of NMVR spectral data with brevifoliol described in Chapter 3.






83


Taxanes I and II [5-11 and [5-21

The crude crystals that separated out from the fractions (8 g) consisting of [5-1], [5-2] and taxol [5-3] were processed by two methods. In one, the mixture (4 g) was taken up in chloroform and ligroin (3:1, 50 ml) and applied to a column of normal phase silica (60 g). The mobile phase was successively changed to chloroform, 2% acetone, 5% acetone, 2% methanol and 5% methanol in chloroform. Compounds [5-1] and [5-2] appeared in the 2-5% acetone and chloroform eluate partially separating from each other. Continuing with 2% methanol in chloroform gave taxol with small amounts of [5-1] and [5-2].

To obtain further purification of [5-1] and [5-2] the mixture was taken up in 40% acetonitrile and water and applied to a column of C-18 bonded silica. The column was eluted with 45 and 50% acetonitrile and water. As the fractions from the 45% acetonitrile and water elution stood for about a week, crystals appeared over a range of tubes and these were filtered in groups. Although [5-1] and [5-2] were separated, such that each contained the other to the extent of 10% or less, recrystallization gave worse mixtures, thus suggesting that isomerization (or equilibration) was taking place during the process. Data obtained on a crystalline (9:1 mixture of [5-1] and [5-2]: m.p. 136-138o C, [Ca]D23 +2140 (c 1.04, CHCI3); (lit. Appendino et al. 1992 on an amorphous sample, m.p. 1631650 C and [XID23 +1850).

Analysis calculated for C31H3808, H20: C, 66.89; H, 7.24. Found: C, 66.51; H,

7.19.

The 1H- and 13C NMR spectra of the crystalline [5-1] and [5-2] gave evidence of mixtures of two compounds. From the spectral data, these two were identified as a mixture of 5-O-cinnamoyl-9-acetyl taxicin I [5-1] and 5-0-cinnamoyl-1O-acetyl taxicin I [52] described by Chmurney et al. (1993) from Taxus x media Hicksii and by Appendino et






84


al. (1992) from Taxus baccata. The former authors isolated insufficient amounts for characterization and the latter authors obtained them as amorphous powders by using HPLC and preparative TLC.

The mixture of [5-1] and [5-2] on acetylation with acetic anhydride and pyridine gave the acetate, readily obtained as a crystalline solid, m.p. 238-2410 C, the NMR spectrum of which showed that it was a single entity, unlike the starting material. It was also identical with taxane Ill (see below).

Taxane III [2-11

The crude crystals of taxane III obtained from the fractions with 50% acetonitrile and water were filtered and recrystallized from acetone in ligroin to obtain colorless needles, yield, 0.8 g (0.02%); m.p. 238-2410 C, [a]D23 +214 (CHCI3, c 1.04); (lit. +218, Baxter, 1962); FAB-MS (m/z): 645 (M +Na); 623 (M' + H); 475 [(MH)* 148 (cinnamoyl)], 415 (475-AcOH); 355 (415-AcOH); 295 (355 AcOH). The spectral data showed that it is the 5-O-cinnamoyl-2L, 9a, 101-triacetyl taxicin I (Appendino et al.1992; Baxter et al. 1962).

Analysis calculated for C35H42010, H20: C, 65.61; H, 6.92. Found: C, 66.00; H,

6.72.

Taxane IV [2-21

This compound also crystallized out directly from the fractions. The crude crystals were purified by recrystallization from acetone and ligroin, yield, 0.8 g (0.02%); m.p. 265-267' C; [a]D23 +1330(C 0.98, CHCI3); (lit. +137, Baxter et al. 1962); FAB-MS: 607 (MH'); 459 (607 148 (cinnamate)); 399 (459 HOAc); 339 (399 HOAc); 279 (339

- HOAc).

Analysis calculated for 035H4209: C, 69.02; H, 7.03. Found: C, 69.29, H, 6.98.





85


The analytical and spectral data of [2-2] indicated that it was identical with 5-0cinnamoyl taxicin I1: 2c,9a,10 -triacetate described by Appendino et al. (1992) and Baxter et al. (1962).

Taxol [5-31

In the silica column described above under the purification of compounds [5-1] and [5-2], taxol (approximately 0.8 g) was eluted by 2-5% methanol in chloroform. A small portion was crystallized from acetone and ligroin to obtain colorless needles of taxol. The 1H NMR spectrum showed that the compound still had appreciable quantities of compounds [5-1] and [5-2]. To remove these compounds completely, ozonization was carried out on the rest of the sample in chloroform and methanol (9:1, 30 ml) at -70o C for 10-15 min. The reaction mixture was treated with dimethyl sulfide (0.5 ml) and let stand at room temperature for 2 h.

After concentration to dryness, the sample was chromatographed on normal phase silica (25 g) in chloroform. Elution with 2% methanol in CHCI3 gave taxol which was crystallized from ligroin to obtain pure taxol, free from compounds [5-1] and [5-2], yield, 0.5 g (0.012%). Its spectral properties agreed with those of an authentic sample.

Alternatively, the crude crystalline solid consisting of compounds [5-1] ,[5-2] and taxol was directly ozonized in chloroform and methanol as before (but without he intermediate silica column purification). After decomposition of the ozonide, and concentration, the sample was subjected to chromatography and taxol isolated from the column. It was crystallized as before to yield 0.75 g (0.015%). The products of ozonization of compounds [5-1] and [5-2] were more polar than the original compounds and separated from taxol in the normal phase silica column.





86


Ozonolysis of Compound [2-21

A solution of compound [2-2] (1 g.) in chloroform and methanol (30 ml, 9:1) was cooled in a dry ice and acetone bath and saturated with ozone for 10-15 min. TLC showed that the starting material was absent and ozonide being formed (detected by spraying with starch and potassium iodide which gave a blue color). After the decomposition of the ozonide by dimethyl sulfide, the reaction mixture was washed with water and concentrated to dryness. The product was crystallized from acetone in ligroin to obtain colorless needles, yield, 0.8 g, m.p. 168-1700 C, [Uc]D23 +130 (c 1.06, pyridine); HRMS: 569.2239, Calc. for C27H36013, 569.2234.














CHAPTER 6
TAXANE CONSTITUENTS OF TAXUS FLORIDANA


Taxus floridana is a species of Taxus, native to Florida. Its distribution is said to be limited to a small area along the Apalachicola River. It is a shrub and used frequently as an ornamental plant. As it is so with the other species of Taxus, the leaves of T. floridana are also reputed to be toxic to livestock and humans.

During the intensive search to find alternative sources for taxol to replace the bark of the Pacific yew, many species of Taxus from the United States, Canada, Europe and Asia were examined. However, there was no study of the taxane constituents of Taxus floridana. Our laboratory undertook this task to evaluate its usefulness as a possible source for taxol.

There was also an impetus for this study from another source. In exploring alternative sources to replace the bark of the Pacific yew, the National Cancer Institute (NCI) was interested in knowing whether the Taxus plants can be grown under hydroponic conditions, as opposed to their growing in their natural state. If these plants can be so grown under hydroponic conditions, which will eliminate the problem of having to harvest the tree bark, the next question was whether they produce taxol in adequate yields. Accordingly, the NCI approached our laboratory, and that of Prof. George Hochmuth Jr. of IFAS, University of Florida, to study this aspect. The hydroponic cultural techniques were studied by the IFAS laboratory and the isolation and characterization of taxanes by our laboratory. It was soon found that the two most wellknown species of Taxus, namely T. brevifolia and T. baccata could not be readily propagated under normal hydroponic conditions, because their growth rate was very


87





88


slow. However, T. floridana responded satisfactorily and could be propagated under available conditions. This species was therefore studied in our laboratory for its taxane constituents.

The needles of T. floridana were collected from the campus and were extracted without drying. After extraction with methanol as before, concentration to remove the solvent, and partition between chloroform and water, the organic layer was concentrated to a dark green semisolid. Fractionation was again carried out using the reverse phase column techniques as was described under the needles of the other Taxus species in Chapters 3 and 5.

The crude chloroform extract was first tested by analytical HPLC to see the elution pattern of the taxane constituents. Taxol was clearly recognizable at its normal location, and in contrast to the observation with the needles of Taxus x media Hicksii, where there were co-eluting taxanes, the taxol from the extract of the needles of T. floridana was relatively free from such interfering taxanes. There were other taxanes situated at different locations.

Elution of the reverse phase column was carried out using a step gradient of 30, 35, 40, 45, 50 and 60% acetonitrile/ water. When the fractions were let stand at room temperature for 3-5 days, taxol and several other taxanes crystallized out as before.

The initial eluates from the column from 25-30% acetonitrile/ water contained highly polar phenolic constituents. The first taxane component to appear from the reverse phase column emerged with the 30-35% acetonitrile/water solvent, and crystallized almost immediately. This was found to be 10-deacetyl baccatin III [2-7]. The next taxane was eluted with the 40% acetonitrile/ water and it was found to be identical with brevifoliol [3-1]. With the 45% acetonirile/ water, was eluted another crystalline compound which was found to be a new compound, and was named taxiflorine [6-1]. Continued elution with 50% acetonitrile/ water gave two crystalline compounds in






89


succession. One of these was identified as baccatin VI [6-2], and the second one was taxol [5-3].

Taxiflorine


Taxiflorine [6-1] was readily obtained as a colorless crystalline solid. Its elemental analysis agreed with the molecular formula C35H44013. Its 1H NMR spectrum in CDC13 showed broad and rounded peaks with poor resolution. In DMSO-d6, the spectrum gave sharper signals but showed double the number of peaks in certain positions. The 13C spectrum also exhibited extra peaks, which suggested that the compound was a mixture of rotamers in equilibrium. One could infer the presence of ester functions from the spectra, with four acetates and one benzoate, and an oxetane ring.

Acetylation of taxiflorine gave a monoacetate [6-3], which gave sharp signals in its 1H NMR spectrum, with the expected number of peaks, thus showing that it is a single compound, unlike the starting material. Although the acetate was isomeric with baccatin VI, it was different. The most striking difference between the two spectra was seen with the signal for the H-13. In the acetate of taxiflorine, this signal was at 8 5.60, while the same was found at 6 6.3 in baccatin VI. A comparison with other related taxanes showed that in those with the 6-membered A-ring, the H-13 signal appears at 6 6.2-6.5, whereas in taxanes with a 5-membered A-ring, as in the 11(15-->1)-abeotaxanes, it appears at 8 5.4-5.7 (Appendino et aL 1993B).

Positions 9 and 10 in taxiflorine carry a free hydroxyl and a benzoate function. To locate the benzoate, a comparison of the signals due to H-9 and H-10 in taxiflorine were compared with the corresponding signals in the monoacetate. With the two signals at 5 6.30 and 6 5.90 in taxiflorine, the latter undergoes a down-field shift from 6 5.90 to




Full Text
79
The reverse phase column run on 600 g of the extract obtained from 12 Kg of the
dried needles was applied to a column prepared from 3 Kg of the C-18 bonded silica.
The yields of the products obtained were good and important values are listed below.
Brevifoliol 3-11
The fractions containing this component were combined, concentrated to dryness
and chromatographed on a normal phase silica column. The major component was
obtained as a colorless crystalline solid, which was found to be identical with brevifoliol
[3-1] on the basis of its spectral data.
Taxanes I [5-11 and II [5-21
Additional chromatography of the mixture of the taxanes I and II and taxol on a
normal phase silica column gave some separation of the two taxanes. Although they
could be further separated and obtained as crystalline solids, they still represented an
equilibrium mixture, as was indicated by the 1H- and 13C NMR spectra of the individual
crystalline samples. From the spectral data, these two were identified as a mixture of
5-0-cinnamoyl-10-acetyl taxicin I [5-1] and 5-0-cinnamoyl-9-acetyl taxlcin I [5-2], which
were isolated (Chmurney et at. 1993) from the needles of T. x media Hicksii and from the
needles of T. baccata (Appendino et at. 1992). The former authors obtained them in
quantities not sufficient for physical properties, and the latter authors obtained them as
amorphous powders, by using HPLC and preparative TLC.
The mixture of [5-1] and [5-2] on acetylation with acetic anhydride and pyridine
gave the triacetate [2-1], which was obtained as a crystalline solid and was found to be a
single entity unlike the starting material. It was also identical with the taxane III (see
below).


58
structural elucidation of [3-1] took unexpected courses. The products of these reactions
were isolated and characterized structurally, and the details are described here.
1. Acid-Catalyzed Acetylation
In Chapter 3, acetylation of [3-1] by means of acetic anhydride and pyridine to
give the crystalline monoacetate [3-2] and the diacetate [3-4] has been described. We
isolated from the needles of T. brevifolia another taxane, which resembled brevifoliol,
giving a dark greenish blue color when sprayed with sulfuric acid and heated. This same
product was also subsequently isolated by others (Barboni et al. 1993). It was similar to
but different from the monoacetate [3-2], but when acetylated, gave [3-4], thus showing
that it was the 13-acetate [3-3],
Acetylation of brevifoliol using acetic anhydride and BF3 gave a different
crystalline product, with almost the same RF as that of the diacetate [3-4], The BF3
product closely resembled the 5, 13-diacetate in its NMR spectral properties also, but
with some differences. In the 1FI NMR spectrum, the singlet at 52.74 which is assigned
to the 15-hydroxyl in the diacetate [3-4], was no longer present. The two methyl singlets
assigned to C-16 and C-17 at 51.11 and 51.35 were deshielded to 51.63 and 51.71. The
other two (C-18 and C-19) showed either no shift or a much smaller one (5 2.03->5 1.96
for C-18). Thus, the significant downfield shift of the signals due to C-16 and C-17
suggested that acetylation of the 15-hydroxyl might have taken place.
Additionally, the signals due to the H-2a and 2(3 showed slight downfield shifts
(H-2a: 51.46, broad doublet, J=13 Hz to 1.53 ppm; H-2p: 52.41, doublet of doublets,
J=9,13 Hz to 2.65 ppm). The spectra of brevifoliol and its acetates generally show the
H-10a signal as a sharp and well resolved doublet, but the one due to H-9p as a broad,
poorly resolved doublet. In the BF3-reaction product, the signal due to H-9p was not only
a well resolved doublet (J=10.8 Hz); but also shielded by 0.29 ppm to 5 5.80. These


5.
6.
7.
7
Counter-current distribution
HPLC on the appropriate fractions
A second HPLC on the appropriate fractions
Similarly, the large-scale process developed for the isolation of taxol by
Polysciences Inc. from the bark of T. brevifolia consisted of the following steps, again
starting with the alcoholic extract concentrate (Boettner et al. 1979).
1. Solvent partition water and CH2CL2, concentration to a solid
2. Separation of the extract solid into soluble and insoluble fractions
3. Chromatography on the soluble fraction
4. Recovery of taxol and crystallization twice
5. Silica chromatography on the taxol/ cephalomannine mixture
6. Recovery and crystallization of taxol
Thus, it appeared that, although procurement of taxol was of top priority, and
many alternative approaches were attempted for solving this problem, one alternative,
which was not considered, was to study the existing isolation procedure itself to make it
more efficient. Thus the approach pursued at the University of Florida during 1990-91
was to develop a simpler process for the isolation. Over the next few months, a new
process was developed based on the use of a single reverse-phase chromatographic
column, and consisting of the following steps, starting with the alcoholic extract
concentrate (Rao, 1993).
1. Partition between water and chloroform, and concentration
2. Reverse phase column chromatography on the extract directly
3. Harvesting the crystals and recrystallization.
The total chloroform extract of the bark of T. brevifolia, was applied directly to the
C18-bonded silica column in 25% acetonitrile in water (i.e., no separation into soluble
and insoluble fractions); and the column developed with a step gradient (30-60%
acetonitrile). The column fractions were let stand for 3-7 days, whereby taxol and seven


I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Koppaka V. Rao, Chair (deceased)
Professor of Medicinal Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
V-/ZL
Perrin, Cochair
ifssor of Medicinal Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Margaretp. James
Professor of Medicinal Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Kenneth B. Sloan
Professor of Medicinal Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Jonathan Eric Enholm
Associate Professor of Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate^wyscop
quality, as a dissertation for the degree of DoctortoF^ilosojahy.
Stephen G. Scbiilman
Professor of Medicinal Chemistry


6-2 : Carbon NMR Spectrum of Baccatin VI
97
ix


48
Analysis calculated for C31 H40O9: C, 66.89; H, 7.24. Found: C, 67.12; H, 7.35;
Brevifoliol-5-Monoacetate 3-21
A mixture of brevlfoliol (0.2 g); acetic anhydride (2 ml) and pyridine (0.5 ml) was
stirred at room temperature for 2-3 min. Water was added to decompose the reagent,
and the solid filtered after 15 min. The solid was crystallized from a mixture of acetone
and ligroin to obtain the mono acetate as a colorless crystalline solid, yield, 0.18 g;
m.p.224-226 C;
1H NMR (CDCIa, 600 MHz, 6) Table 3-1: 0.91, s (H-19); 1.02, s (H-16); 1.22 (dd,
J=7.2, 13.8 Hz, H-14a); 1.33, s (H-17); 1.46 (d, J=14.1 Hz, H-2a); 1.76 (s, 9-0 acetate
methyl); 1.88 m, 2.0 m (H-6); 2.06 x 2, s (methyl-18, 5-0 acetate methyl); 2.08 (s, 7-0
acetate methyl); 2.40 (dt, J=14.1, 9.6 Hz, H-2P); 2.42 (dd, J=7.2, 13.8 Hz, H-14p); 2.75
(d, J=9 Hz, H-3a); 2.83, br s (C-15 OH, exchangeable with D20); 4.53 (t, J=7.2 Hz, H-
13p); 4.90, s (H-20 A); 5.28, s (H-20 B); 5.37 (br s, J= H-5p); 5.65 (dd, J=4.8, 11.4 Hz,
H-7a); 6.02 (poorly resolved br d, J=10.5 Hz, H-9a); 6.63 (d, J=10.5 Hz, H-10p); 7.43 (t,
J=7.8 Hz, H-Ph-mefa); 7.56 (t, J=7.8 Hz, H-Ph-para); 7.87 (d, J=7.8 Hz, H-Ph-ortho).
13C NMR (CDCI3, 600 MHz, 5) Table 3-2: 11.8 (C-18 methyl, q); 12.9 (C-19
methyl, q); 20.8 (7-0 acetate methyl, q); 21.2 (5-0 acetate methyl, q); 21.4 (9-0 acetate
methyl, q); 24.8 (C-17 methyl, q); 27.0 (C-16 methyl, q); 29.2 (C-2, t); 33.9 (C-6, t); 38.8
(C-3, d); 44.8 (C-8, s); 47.1 (C-14, dd); 63.0 (C-1, s); 69.7 (C-7, d); 70.7 (C-10, d); 74.1
(C-5, d); 75.6 (C-15, s); 76.9 (C-13, d); 77.9 (C-9, d); 114.0 (C-20, t); 128.7 (C-Ph-mefa,
d); 129.2 (C-Ph-ipso, s); 129.4(C-Ph-orfbo, d); 133.3 (C-Ph-para, d); 134.0 (C-11, s);
145.2 (C-4, s); 151.1 (C-12, s); 164.1 (CO-Ph, s); 169.9 X 2(CO-Acetate, s); 170.2 (CO-
Acetate, s).
Analysis calculated for C 33H 42O10: C, 66.20; H, 7.07. Found: C, 66.38; H, 7.19.


CHAPTER 2
A SELECTED REVIEW OF THE LITERATURE ON TAXUS
An overview of the taxol story has been presented in Chapter 1. In this Chapter,
a selected review of the literature will be presented on taxol as well as the other taxanes,
which are included in this dissertation. Two comprehensive reviews have been
published on the subject of taxol, one by Kingston et al. (1993) and one by Miller (1980).
The present review on the genus Taxus may be roughly divided into two parts:
studies before and studies after the discovery of taxol.
Earlier Studies
The genus Taxus (N.O. Taxaceae) represents a group of plants (common name,
yew) which grow mostly in temperate climates and can be found distributed throughout
the world. They are generally slow-growing evergreen trees or shrubs with stiff linear
leaves (or needles); and fruits which are small, fleshy and bright red. The common
names of the plants are qualified by the place of its origin, as for example, Pacific or
western yew (T. brevifolia, Nutt.); European or English yew (7. baccata, PHg.)\ Canadian
yew (7. canadensis, Willd.); Japanese yew (7. cuspidata, Sieb. et Zucc.); Chinese yew
(7. chinensis, Pilg ); Himalayan yew (7 wallichiana, Zucc.); ornamental yew (7. x media
Hicksii, Rehd.) and Florida yew (7 floridana, Nutt.).
The toxic nature of this genus has been recognized for thousands of years, and
in modern times, was first investigated chemically using the needles of Taxus baccata
(Lucas, 1856). An amorphous mixture of alkaloids was isolated after extraction under
acidic conditions and was given the name of "taxine. Further studies on taxine spanned
11


24
Danishefskys total synthesis of baccatin III in 1996 (and hence, taxol); borrowed
extensively from the experiences of Ojima, Holton, Nicolaou and others. The Weiland-
Miescher ketone [2-28], available through catalytic asymmetric induction, allowed the
installation of all stereochemical requirements to reach baccatin III in a sequential
fashion. According to Danishefsky, Our synthesis, though arduous, involves no relays,
no resolutions, and no recourse to awkwardly available antipodes of the chiral pool
(Danishefsky etal. 1996).
Wenders group published a most concise synthesis involving a-pinene [2-29] for
construction of the ABC-tricyclic core of the taxanes (Wender & Rawlins 1992). Their
approach takes advantage of the tendency for C-7 to undergo facile aldol/reverse aldol
epimerization in taxol, allowing for aldol condensation under very mild conditions.
General Structural Features of Taxanes
The taxanes comprise a relatively large group of diterpenoid natural products
covering a variety of structural patterns. These are believed to arise from geranyl
geraniol [2-16], although the exact biosynthetic route has not been completely
elucidated. A brief discussion of the major structural variations of taxanes is relevant to
this work because many of these structures have been found in the compounds isolated
in this work. A number of different forms that the C-20 diterpene skeleton itself can
assume have been isolated. Next, the oxidation states, esterification patterns of the
hydroxyls, and presence or absence of basic or neutral side chains allow for the
extensive structural variation seen in these compounds.
The taxane skeleton is a specific diterpene structure, consists of 20 carbon
atoms arranged in a fused tricyclic system with the 6, 8 and 6 members in rings A, B and
C, respectively. The double bonds at 11/12 and 4/20 are part of the basic ring system,
although the latter may be modified by oxygenation to an epoxide or more commonly to


98
13C NMR (CDCI3i 300.075 MHz, 8): 75.0 (C-1); 78.8 (C-2); 47.5 (C-3); 81.6 (C-4);
83.8 (C-5); 34.3 (C-6); 73.3 (C-7); 45.8 (C-8); 71.9 (C-9); 69.7 (C-10); 133.5 (C-11);
142.0 (C-12); 70.4 (C-13); 35.3 (C-14); 42.8 (C-15); 22.4 (C-16); 28.3 (C-17); 12.9 (C-
18); 15.0 (C-19); 76.5 (C-20); 20.7; 20.9, 21.2, 21.4, 22.7 (5x OAc Me); 168.6, 168.9,
169.6, 169.9, 170.2 (5X OAc CO); 166.8 (Bz CO); 129.2 (Bz ipso); 128.5 (Bz ortho);
129.9 (Bz meta); 133.5 (Bz para).
Analysis calculated for CsrH^O^: C, 62.18; H, 6.49. Found: C, 61.83; H, 6.45.
Taxol [5-31:
Crystals (4.5 g) from the second part of the peak which contained mostly taxol.
were combined dissolved in chloroform (60 ml) and chromatographed on Florisil(40 g).
Elution with chloroform gave more of [6-3] and subsequent elution with 5% acetone in
chloroform gave taxol, which was recovered by concentration of the appropriate fractions
and crystallized from acetone and ligroin to obtain taxol, yield, 1.98 g (0.01%); m.p. 220-
222 0 C. The spectral and chromatographic properties of the sample agreed with those
of taxol.
Acetylation of Taxiflorine to [6-31:
An aliquot of [6-1] (0.05 g) was acetylated by acetic anhydride (2 ml) and pyridine
(0.5 ml) at room temperature for 16 h. Water was added, the solid filtered and purified
by chromatography on a silica column using chloroform / ligroin (2:1) to obtain [6-3] as a
white powder.
Benzoylation of Taxiflorine to [6-4]:
To a solution of [6-1] (0.05 g) in pyridine (2 ml) was added dropwise with stirring
at 0-5 C, benzoyl chloride (0.1 ml). After 20 h water was added followed by 2N sulfuric
acid and the solid filtered. The product was purified by chromatography as given under
[6-3], to obtain [6-4] as a white powder.


72
Acetylation of [4-41 to [4-51
A sample of [4-4] (.05 g) was acetylated in acetic anhydride (2 ml) and pyridine
(0.5 ml) at room temperature for 20 h. After addition of water, the solid was filtered and
crystallized from ether in ligroin, m.p. 250-254 C.
1H NMR (CDCIa, Varian Unity 600 MHz, 5): 1.04 (s, H-16); 1.19 (s, H-19); 1.28 (s,
H-17); 1.91 (s, methyl, 9-acetate); 1.77 (br dd, H-14oc); 2.02 (m, H-6p); 2.05 (s, 7-acetate
methyl); 2.27 (s, H-18); 2.19 (s, 5-acetate methyl); 2.22 (cm, H-14p); 2.34 (d, J=13.8 Hz,
H-2a); 3.01, (brs, C-15 OH, exchangeable with D20); 3.31 (d, J=13.8 Hz, H-2(3); 3.90 (d,
J=13.2 Hz, H-20 B); 4.35 (d, J=13.2 Hz, H-20 A); 4.50 (br m, H-13|3); 5.41 (dd, J=3.6,
13.2 Hz, H-7a); 5.46 (d, J=4.2 Hz, H-5P); 6.04 (d, J=10.1 Hz, H-9a); 6.57 (d, J=10.1 Hz,
H-10P); 7.45 (t, J=7.5 Hz, H-Bz-mefa); 7.57 (t, J=7.5 Hz, H-Bz-para); 7.89 (d, J=7.5 Hz,
H-Bz-ortho).
13C NMR (CDCIa, Varian VXR 300 MHz, 5): 12.9 (C-18 methyl, q); 16.1 (C-19
methyl, q); 21.7 (7-0 acetate methyl, q); 21.5 (9-0 acetate methyl, q); 21.7 (5-0 acetate
methyl, q); 25.3 (C-16 methyl, q); 27.0 (C-17 methyl, q); 31.5 (C-2, t); 32.9 (C-6, t); 38.3
(C-14, t); 45.2 (C-8, s); 63.5 (C-1, s); 65.4 (C-20, t); 67.6 (C-7, d); 70.3 (C-5, d); 72.3 (C-
9, d); 74.4 (C-15, s); 76.3 (C-10, d); 83.2 (C-13, d); 127.3 (C-4, s); 128.8 (C-Bz-meta, d);
129.1 (C-Bz-/pso, s); 129.4(C-Bz-ortho, d); 133.4 (C-Bz-para, d); 135.5 (C-12, s);
142.7(C-3, d); 145.6 (C-3, s); 147.6 (C-11, s); 164.6 (CO-Ph, s); 169.6 (CO-Acetate, s);
170.3 (CO-Acetate, s); 170.8 (CO-Acetate, s).
Reaction with N-Bromosuccinimide and Silver Acetate [4-61
A solution of brevifoliol (0.2 g) in benzene (10 ml) was stirred with N-
bromosuccinimide {NBS} (0.125 g, recrystallized from water). After 2 h the starting
material was absent with two faster moving compounds being present. To the reaction
mixture was added silver acetate (0.125 g) and stirred for another 2 h.


62
1. The most striking evidence is that only two of the four methyl signals that are seen in
brevifoliol appear in the BF3-reaction product. This is not only observed in the 1H
spectrum, but also in the 13C NMR thus showing that the oxy-isopropyl chain at C-1
is not present.
2. An examination of the isolated spin system: 5p 6a- 6p 7a in the 1H NMR spectrum
of brevifoliol and that of the product of BF3 reaction indicated that the free hydroxyl at
C-5a was still present, but that the C-13a-OFI was absent. Through a study of the
interactions in the COSY spectrum, it was possible to assign (and distinguish) the
signals for C-6 and C-14 in the region below S3.
3. A new methine proton signal appeared as a broad singlet at S5.84, assigned to H-13,
which interacted in the FIETCOR spectrum with the signal at 5124.1, which
corresponds to C-13.
4. The DEPT spectrum showed the presence of four signals for methyl-type (CPI3)
carbons (811.4, 5 13.5 for C-19 and C-18, 5 20.7, 8 21.2 for the two CO-Me); two
signals for methylene-type (CFb) carbons (8 112.4 for C-20, 8 44.6 for C-14, 8 34.1
for C-6 and 8 27.9 for C-2, nine signals for methine-type (R3CPI) carbons (5133.0,
8129.7, 5128.4 for the five aromatic carbons, four aliphatic CH-OR type carbons
[873.2 (C-9); 572.5 (C-5); 579.6 (C-7); 867.4 (C-10)], one aliphatic methine type
carbon (539.6 (C-3)); one unsaturated methine type carbon (5124.1 (C-13) and nine
quaternary carbons (three carbonyls, one aromatic C carrying the carboxyl, C-1(?);
C-4, C-11, C-12 and C-8); together, account for the 28 carbons present..
5. HETCOR interactions supported the assignments for the o, m, p- positions in the
benzoate, and for the C-13, C-20, C-9, C-5, C-7, C-10, C-3, C-6, C-2, C-18, C-19
and the two acetate methyls.


39
[3-14] Fteriodate Oxidation Osazone Product
Figure 3-3 : Brevifoliol Hexaol Reaction Products
The major component of this mixture was found to be debenzoyl brevifoliol
(Figure 3-1 [3-15]). Of the three possible locations, 7, 9, and 10 for the benzoate, only
10 Is allylic and hence the ester at this position is more likely to be labile. Taxol with the
benzoate at C-2 is completely stable to heat and pressure for hours. This evidence,
along with chemical shift arguments concerning the effect of acetylation versus
benzoylatlon, led us to place the benzoate at C-10.
This group presented the isolation and the structural elucidation of brevitaxane A
at the International Research Congress on Natural Products held in Chicago, IL in July


CHAPTER 6
TAXANE CONSTITUENTS OF TAXUS FLORIDANA
Taxus floridana is a species of Taxus, native to Florida. Its distribution is said to
be limited to a small area along the Apalachicola River. It is a shrub and used frequently
as an ornamental plant. As it is so with the other species of Taxus, the leaves of T.
floridana are also reputed to be toxic to livestock and humans.
During the intensive search to find alternative sources for taxol to replace the
bark of the Pacific yew, many species of Taxus from the United States, Canada, Europe
and Asia were examined. However, there was no study of the taxane constituents of
Taxus floridana. Our laboratory undertook this task to evaluate its usefulness as a
possible source for taxol.
There was also an impetus for this study from another source. In exploring
alternative sources to replace the bark of the Pacific yew, the National Cancer Institute
(NCI) was interested in knowing whether the Taxus plants can be grown under
hydroponic conditions, as opposed to their growing in their natural state. If these plants
can be so grown under hydroponic conditions, which will eliminate the problem of having
to harvest the tree bark, the next question was whether they produce taxol in adequate
yields. Accordingly, the NCI approached our laboratory, and that of Prof. George
Hochmuth Jr. of IFAS, University of Florida, to study this aspect. The hydroponic
cultural techniques were studied by the IFAS laboratory and the isolation and
characterization of taxanes by our laboratory. It was soon found that the two most well-
known species of Taxus, namely T. brevifolia and T. baccata could not be readily
propagated under normal hydroponic conditions, because their growth rate was very
87


33
Hydroxyl Functionalities
i) Acetylation: To determine the number and positions of all hydroxyls in the
molecule, the compound was subjected to acetylation. Two products were obtained
under mild conditions (20 C, 15 min). These two were separated by chromatography
and both obtained as crystalline solids. One was shown to be a monoacetate and the
other a diacetate.
Table 3-1 : Proton NMR Spectra of Brevifoliol and Brevifoliol Acetates
Position
Brevifoliol
Brevifoliol
Brevifoliol
Brevifoliol
Brevifoliol
(J in Hz)
5-Ac
13 Ac
5,13 Ac
5,13,15 Ac
[3-1]
[3-2]
[3-3]
[3-4]
[3-5}
2
1.49 cm
1.46 cm
1.47 cm
1.46 brd (13)
1.53 br d(13)
2.36 dd (9,13)
2.40 dd(9, 13)
2.42
2.41 dd(9, 13)
2.65 dd(9,
3
2.78 d (9)
2.76 brd (9)
2.91 d (9)
2.72 brd (9)
2.71 brd (9)
5
4.45 brs
5.37
4.37 brs
5.39 br s
5.38 dd (4, 2)
6
1.86 cm
1.88 cm
1.85 cm
1.90 cm
1.87 cm
2.02 cm
2.0 cm
1.99 cm
2.00 cm
2.0 cm
7
5.56 dd (5,11)
5.62 dd (5, 11)
5.66
5.61 dd (5,
5.63 dd(5,
9
6.05 br
6.03 br d(10.6)
6.07 d
6.09 br
5.8 d (10.8)
10
6.53 d (10.6)
6.63 d (10.6)
6.66 d
6.65 d (10.6)
6.64 d (10.8)
13
4.38 t (7.5)
4.53 brt (7.2)
5.46 br s
5.54 brt (7.2)
5.61 t (6.9)
14
1.29
1.22 dd *
1.32 cm
1.25 dd *
1.25 dd *
2.46
2.42 dd *
2.51 cm
2.51 dd *
2.62 dd *
16
1.05 s
1.03 s
1.09 s
1.11 s
1.63 s
17
1.35 s
1.33 s
1.35 s
1.35 s
1.71 s
18
2.01 s
2.06 s
2.02 s
2.03 s
1.96 s
19
0.90 s
0.91 s
0.89 br s
0.92 s
0.92 s
20 A
4.82 brs
4.90 br s
4.80 br s
4.92 brs
4.89 br s
20 B
5.20 brs
5.28 br s
5.15 br s
5.28 br s
5.29 brs
o-Ph1
7.88 d( 7.5)
7.87 d (7.5)
7.87 d (7.5)
7.87 d (7.5)
7.84 d (7.5 )
m-Ph1
7.43 t( 7.5)
7.43 t (7.5)
7.44 t (7.5)
7.44 t (7.5)
7.42 t ( 7.5 )
p-Ph1
7.56 t ( 7.5)
7.55 t (7.5)
7.56 t (7.5)
7.56 t (7.5)
7.53 t( 7.5)
-
1.76 s
1.76 s
1.76 s
1.75 s
1.77 s
2.07 s
2 06 s
2.05 s
2.02 s, 2.07 s
2.02 s, 2.08 s
2.13 s
2.06 s
2.08 s
2.09 s, 2.11 s
NMR were recorded at 600 MHz in CDCI3on a Varian Unity 600 instrument at
ambient temperature. Chemical shifts 8 (ppm) are reported with TMS as internal
standard.


61
[4-1 ] Jones Oxidation Aldehyde [4-2] Mn02 Diketone
Figure 4-1 : Oxidation Products
3. Action of BF3 on Brevifoliol [4-31
Before the structure of brevifoliol was fully established, the possibility that the two
hydroxyls present in the compound might be vicinal to each other was considered. With
the aim of forming an isopropylidene derivative, brevifoliol was reacted with acetone in
the presence of Dowex-50 (sulfonic acid resin) as the catalyst. Reaction took place
readily with the formation of a faster moving product, with some decomposition also
taking place (colors). The reaction proceeded with less decomposition when BF3-
etherate was used as the catalyst.
Chromatography of the mixture from either reaction yielded the major product as
a colorless crystalline solid. Its NMR spectrum quickly ruled out the possibility of its
being an isopropylidene derivative. The FAB-mass spectrum gave a value for the MFI+
m/z of 481, as compared with 557 for brevifoliol, thus showing a loss of 76 mass units.
The elemental analysis, which agreed with C28FI32O7 showed a loss of C3FI802 compared
to brevifoliol. This may be interpreted as the loss of the C3H7O- side chain attached to
C-1, as well as that of FI2O, possibly through the elimination of the C-13-OH (or C-5-OFI).
The evidence to support this assertion is given below.


91
Table 6-1 : Proton NMR Spectra of Compounds [6-3], [6-4] and [6-5],
H#
Compound [6-3]
Compound [6-4]
Compound [6-5]
2
6.19, d, J=7.8 Hz
6.26, d, J=7.8 Hz
6.07, d, J=7.8 Hz
3
2.99, d, J=7.8 Hz
3.06, d, J=7.8 Hz
2.92, d, J=7.8 Hz
5
4.98, d, J=7.5 Hz
5.01, d, J=7.5 Hz
4.98, d, J=7.5 Hz
6a
2.68, m
2.70, m
2.52, m
6p
1.84, m
1.84, m
1.84, m
7
5.52, m
5.64, m
5.49, t, J=7.8 Hz
9
6.32, d, J=10.8HZ
6.48, d, J=10.8Hz
6.04, d, J=10.8Hz
10
6.37, d, J=10.8 Hz
6.72, d, J=10.8 Hz
6.27, d, J=10.8, Hz
13
5.62, t, J=7.8Hz
5.64, m
5.61, t, J=7.8Hz
14a
2.30, dd,J=7.4,14.2Hz
2.34, dd,
2.30, m
14(3
1.72, dd,J=7.4,14.2Hz
1.78, m
1.72, m
16
1.16, s
1.24, s
1.15, s
17
1.19, s
1.21, s
1.13, s
18
1.72, s
1.72, s
1.83, s
19
1.64, s
1.95, s
1.66, s
20
4.50, 4.42, d, J=7.9 Hz
4.52, 4.44, d, J=7.2
4.47,4.38, d,J=7.5
Ph(2,6)
7.93, d
Unresolved

Ph(3,5)
7.45, t
7.24, mm

Ph(4)
7.62, t
7.37, m

Ph(2,6)

7.63, d, J=7.2Hz
Ph(3,5)

7.24, m
Ph(4)

7.37, m

OAc Me
2.02,s
2.14, s
2.11,s
OAc Me
2.14, s, (2x)
2.05, s
2.10, s
OAc Me
1.86, s

2.08, s
OAc Me
1.80, s

2.03, s
OAc Me


2.01, s
OAc Me


1.95, s
1H NMR were recorded at 600 MHz in CDCI3 on a Varian Unity 600 spectrometer
at ambient temperature. Chemical shifts 5 (ppm) are reported relative to TMS as an
internal standard.


Figure 3-1 : Proton NMR Spectrum of Brevifoliol


16
O
[2-9] Taxotere, RfHR2= OC(CH3)3
[2-10a] Tetraol, R1 = R3 = H R2 = OH
[2-10b] BaccatinV, R-, = Ac
R2 = H, R3 = OH
[2-11] 7-O-TES 10-DAB III, R, = H
R2 = Triethyl Silyl (TES)
[2-12] 7-O-TES Baccatin III, = Ac
R2 = Tri ethyl Silyl (TES)
[2-13] 7-0-TES-13-0-Cinnamoyl Baccatin III
Figure 2-2 : Taxol and some Synthetic Targets


84
al. (1992) from Taxus baccata. The former authors isolated insufficient amounts for
characterization and the latter authors obtained them as amorphous powders by using
HPLC and preparative TLC.
The mixture of [5-1] and [5-2] on acetylation with acetic anhydride and pyridine
gave the acetate, readily obtained as a crystalline solid, m.p. 238-241 C, the NMR
spectrum of which showed that it was a single entity, unlike the starting material. It was
also identical with taxane III (see below).
Taxane III f2-11
The crude crystals of taxane III obtained from the fractions with 50% acetonitrile
and water were filtered and recrystallized from acetone in ligroin to obtain colorless
needles, yield, 0.8 g (0.02%); m.p. 238-241 C, [a]D23 +214 (CHCI3, c 1.04); (lit. +218,
Baxter, 1962); FAB-MS (m/z): 645 (M+ +Na); 623 (M+ + H); 475 [(MH)+ 148
(cinnamoyl)], 415 (475-AcOH); 355 (415-AcOH); 295 (355 AcOH). The spectral data
showed that it is the 5-0-cinnamoyl-2a, 9a, 10p-triacetyl taxicin I (Appendino et at. 1992;
Baxter et at. 1962).
Analysis calculated for C35H42O10, H20: C, 65.61; H, 6.92. Found: C, 66.00; H,
6.72.
Taxane IV [2-2]
This compound also crystallized out directly from the fractions. The crude
crystals were purified by recrystallization from acetone and ligroin, yield, 0.8 g (0.02%);
m.p. 265-267 C; [a]D23 +133(C 0.98, CHCI3); (lit. +137, Baxter et al. 1962); FAB-MS:
607 (MH+); 459 (607 148 (cinnamate)); 399 (459 HOAc); 339 (399 HOAc); 279 (339
- HOAc).
Analysis calculated for C35H4209: C, 69.02; H, 7.03. Found: C, 69.29, H, 6.98.


28
hydroxyl with groups such as a chloroacetate avoids both epimerization and unwanted
reaction at this position.
Abeotaxanes
A number of taxanes in which the A-ring is isomerized to a 5-membered ring to
give a 5/7/6 instead of the 6/8/6 system have been isolated and these are termed
abeotaxanes. They are again divided into two groups into a) those with the 4/20
unsaturation and b) those with an oxetane ring at this location. We isolated the first
members of each of these groups in our work, e.g. brevifoliol (Chapter 3); and the
compounds isolated from the bark of T brevifolia described in Chapter 6. As indicated
earlier, treatment of taxol with acidic reagents can isomerize ring A to form such
compounds, although these compounds are naturally present in the extract and not
artifacts.


45
using 5 and 3 gallons respectively. The combined chloroform extract was concentrated
under reduced pressure to reach a dark green semi-solid stage (800-900g).
Reverse Phase Column Chromatography:
The column used was a threaded glass column of the Mitchell-Miller type (2.5 x
24) with the appropriate fittings, purchased from Ace Glass Co., Vineland, NJ suitable
for low pressure liquid chromatography. A slurry of the C-18 bonded silica (800 g)
(Spherisorb, 15-35 micron diameter) purchased from, Phase Separations Inc., Norwalk,
CT) in methanol was poured into the column, which was run under a gentle pressure by
using a metering pump (Fisher/Eldex) until an adequately packed bed was obtained.
The column was then equilibrated with 25% acetonitrile in water, to prepare for the
addition of the sample.
The extract solids (200 g) was dissolved in acetonitrile (400 ml) by warming to
make sure that no lumps remained. To this was added approximately 200 g equivalent
of the equilibrated resin (about 20% of the column packing) with stirring. As the stirring
continued, the slurry was diluted with 25% acetonitrile in water (500 ml); followed by
water to make up a total volume of approximately 2 L. The stirring was continued with
occasional warming to 50-60 0 C for about 15 min. At this point, a sample of the slurry
taken into a test tube, showed that the silica settled readily to give a clear supernatant
and no green precipitate or oily material was present. The slurry was then filtered using
light suction and the solid (silica with the sample) re-slurried using part of the filtrate and
the thick slurry added to the column. The rest of the clear supernatant was then pumped
on to the top of the column using the metering pump. From time to time, the column
feed was checked to see that it remained clear, and if not, to either warm briefly or add
minimal amounts of acetonitrile to it until it became clear, so as to prevent precipitate
from appearing and blocking the pump.


109
IR, v max (KBr, cm"1): 1690 (dienone carbonyl at C-2); 1630 (aromatic C-acetyl);
1610 (enol ether and aromatic double bonds); 1540 (conjugate carbonyl at C-3).
Analysis calculated for Ci8H1607: C, 62.77; H, 4.69. Found: C, 62.47; H, 4.75.
Betuloside (4-(4-IHvdroxvphenvl)-2R-butanol Glucoside)) & Aqlycone
M.p. : 191-193 0 (Lit. 187-190 Khan et al. 1976).
1H NMR (CDCI3, 5): 1.17 (3H, d, Me, J=5.7 Hz); 1.8 (2H, cm, H-3); 2.58 (2H, t, H-
4, J=7.5 Hz); 3.6-4.5 (glucosyl); 6.8 & 7.0 (4H, d, A2B2 Aromatic, J=8.2 Hz); 8.84 (1H, s,
4-phenol).
13C NMR (CDCI3, aglycone, 8): 153.9 (C-4); 133.7 (C-1 ); 129.4 (C-2, 6); 115.3
(C-3, 5); 67.8 (C-2); 40.7 (C-3); 31.2 (C-4); 23.4 (C-1 Me).
IR, v max (KBr, cm"1): 3370, 2930, 2860, 1610, 1590, 1510, 1435, 1445, 1370,
1230.
Analysis calculated for C10H14O2: C, 72.26; H, 8.49. Found: C, 72.12; H, 8.56.


37
doublet and a new methine proton appeared. In the 13C NMR spectrum the
characteristic signals from the exocyclic 4/20 double bond were absent, accompanied by
the appearance of new methyl and methine signals.
ii) Epoxidation: Brevifoliol was heated in dichloromethane with meta-chloro
peroxybenzoic acid (MCPBA); whereby it underwent oxidation to yield the epoxide [3-8],
a crystalline compound.
¡ii) Ozonization: Brevifoliol has two double bonds, one at the 11/12 position and
the other at the 4/20 position. Of these, the former is tetra-substituted, while the latter is
of an exocyclic methylene type. No information was available in the literature regarding
the reactivity of the taxane skeleton to indicate whether one or both double bonds would
be cleaved by ozonolysis. In the present work, ozonization was carried out in a mixture
of methanol and dichloromethane -70 0 C. After the disappearance of the starting
material, the ozonide was decomposed with dimethyl sulfide and the products isolated
by chromatography. Two major products were separated. The first was the same as the
epoxide [3-8] obtained by reaction with MCPBA. The second was the expected
ozonolysis product in which the 4/20 double bond was cleaved to form the ketone [3-9],
iv) Formation of a diol: As one of the characteristic reactions of an ethylenic
function, oxidation by osmium tetroxide was attempted with brevifoliol. The reaction
proceeded smoothly to give a diol [3-10],
Number and Nature of the Oxygen Substitution
From the preceding discussion it is evident that brevifoliol has two free hydroxyls,
two acetoxyls and one benzoyloxy functions. Plowever, in the 13C NMR spectrum of
brevifoliol, the number of oxygen substituted carbons was six: 5 70.1, 5 70.2, 8 72.4, 5
75.9, 5 76.7 and 8 77.2. To determine if one of the six is a different type of an ester, or a
tertiary hydroxyl, brevifoliol was subjected to saponification in alcoholic KOFI to yield the


10
describing a suitable scheme for isolation of taxol or any other taxoids have appeared so
far, other than analytical hplc data on their taxol content.
Chapter 6 is similarly devoted to the fractionation of the extract of the needles of
Taxus floridana Nutt, by reverse phase column chromatography. Isolation of taxol, 10-
deacetyl baccatin III, baccatin VI and a new crystalline taxoid compound named
taxiflorine, with its structural elucidation are described.
Chapter 7 deals with the isolation of several crystalline non-taxane compounds
present in the extracts of the bark and needles of Taxus brevifolia. These were shown
to include flavonoids, phenols, and other types of compounds.


30
in water column (1:4 ratio of loaded to clean silica); a step gradient of acetonitrile in
water mixtures was performed up to 60% acetonitrile.
Preliminary studies on the extract solids of the needles by TLC and analytical
HPLC showed that the sample contained somewhat minor amounts of taxol. A
predominant component that was slower moving than taxol in TLC gave a greenish-blue
colored spot when sprayed with 1 N sulfuric acid and heated on a hot plate (charring).
Likewise, in the analytical HPLC, this component appeared after 10-deacetyl baccatin III
as the major constituent judging from the peak heights, but before taxol and at least
several times more abundant.
The reverse phase column (C-18 bonded silica gel) on the needle extract
concentrate was started with 25% acetonitrile in water. The sample was carefully
prepared as a slurry (see experimental) and added to the column. The column was
developed using a step gradient of acetonitrile in water 30-60%. Fractions of suitable
volume were collected and monitored by absorbance at 275 nm, TLC and analytical
HPLC. Four regions were recognized in the elution profile of the column, based on the
UV absorbance (275 nm.); which contained the resolved constituents of the extract.
The early fractions contained components, which accounted for the bulk of the
UV absorbance of the sample. These appeared to be non-taxane phenolic compounds
with or without attached sugars. A description of these will be given in Chapter 7. The
first taxane component, which appeared at the 35-40% acetonitrile elution, was also the
major component. It was collected from the appropriate fractions, and after
concentration, obtained as a crystalline solid. Next, fractions from the 50% acetonitrile
elution contained taxol, which was obtained as a crystalline solid directly from the
fractions. Following this, the fractions from the 55-60% acetonitrile elution gave another
taxane component which gave a greenish blue spot on the TLC (after charring with
sulfuric acid) similar to the major constituent referred to above.


19
provide researchers with a source of analogues for structure-activity relationships and
lead to better methods of production in general.
The first carbon-carbon bond between rings A and C was formed using a
vinyllithium carbanion generated from the reaction of aryl sulphonylhydrazone [2-16],
with /7-butyl-lithium in tetrahydrofuran (THF); which was then combined with the
aldehyde [2-17] in the Shapiro reaction (Shapiro, 1976) to produce [2-18] (Figure 2-3).
Regioselective epoxidation of the A1,14-double bond was completed in 87% yield
with f-butyl peroxide in the presence of VO(acac)2 leading to epoxide [2-19], which was
then regioselectively opened with LiAIH4 to give the 1,2-diol [2-20] with a 76% yield. The
carbonate introduced between the C-1 and C-2 hydroxyls in the next step served to
position the two rings for ring closure and also allowed for the stereo-controlled
introduction of the 2a-benzoate later in the sequence. The dialdehyde [2-21] needed for
cyclization of the B ring was obtained after standard deprotection of the two primary
hydroxyls and mild oxidation with tetra-n-propylammonium perruthenate (TPAP) and
4-methylmorpholine N-oxide (NMO) in acetonitrile. The previous three steps provided
the carbonate dialdehyde in 32% overall yield.
Formation of the B ring was accomplished with the versatile McMurry coupling
(McMurry, 1989) under dilute conditions utilizing low valence titanium produced in situ
from (TiCI3)2-(DME)3 (10 eq.) and Zn-Cu (20 eq.) in 1,2-dimethoxyethane (DME) at 70 0
C for 1 hour, giving the tricyclic A/B/C diol [2-22] with a 23% yield.
Selective acetylation of the hydroxyl at C-10 rather than C-9 was expected due to
allylic activation and proceeded with 95% yield. Mild oxidation of the C-10 hydroxyl was
then carried out with TPAP-NMO in acetonitrile analogous to the oxidation to the
dialdehyde with a 93% yield.


51
145.2 (C-4, s); 144.3 (C-12, s); 163.1 (C-11, s); 165.4 (CO-Ph, s); 169.3 (CO-Acetate, s);
170.3 (CO-Acetate, s); 207.5 (C-13 Ketone, s).
Analysis calculated for C 31H 3809: C, 67.13; H, 6.91. Found: C, 67.48; H, 6.97.
Dihydrobrevifoliol [3-71
A solution of brevifoliol (0.2 g) in ethyl acetate (10 ml) was hydrogenated in a
Parr apparatus using Platinum oxide (0.05 g) for 16 hours. TLC revealed the formation
of a slightly slower moving product. The mixture was filtered and the filtrate
concentrated to dryness and purified by chromatography on silica gel in
dichloromethane. Elution with 2% acetone in dichloromethane gave a minor product,
which was not further investigated. The fractions eluted with 2-5% methanol in
dichloromethane gave the major product, which was obtained as a colorless powder,
yield, 0.1 g,
Reduction of the 4(20) double bond resulted in significant broadening of most
peaks in the 1H- and 13C NMR spectra, but the appearance of additional signals from
methyl group at C-20 and methylene at C-4 could be seen, as well as the loss of the two
characteristic exocyclic methylene singlets.
Analysis calculated for C3iH420g: C, 66.60; H, 7.60. Found: C, 66.89; H, 7.88.
Brevifoliol Epoxide [3-81
A mixture of brevifoliol (0.3 g) and meta-chloroperoxybenzoic acid (MCPBA, 0.2
g) in toluene (15 ml) was heated under reflux for 30 min. After cooling, the mixture was
diluted with ether, washed successively with aqueous sodium bisulfite, aqueous sodium
bicarbonate and saline, and the organic layer concentrated to dryness. The solid was
crystallized from acetone/ligroin, to give a colorless crystalline epoxide, yield, 0.15 g;
m.p. 227-230 C.


107
(C-2, t); 29.2 (C-25, q); 28.2 (C-16, t); 26.1 (C-23, t); 24.3 (C-15, q); 23.0 (C-28, t); 21.0
(C-11, t); 19.8 (C-27, q); 19.3 (C-19, q); 19.0 (C-26, q); 18.7 (C-21, q); 11.8 (C-18, q);
11.9 (C-29, q).
Analysis calculated for C29H50O: C, 83.99; H, 12.15. Found: C, 83.16; H, 12.42.
Phytoecdysteroids
Ecdysterone & 2p, 3B. 22a-Triacetate
(22R)-2p,3p,14a,20p,22a,25-hexahydroxycholest-7-en-6-one, m.p.: 237-240 0 C
(Lit. 240 0 C, Takemoto etal. 1967).
1H NMR (CDCI3, Triacetate, 8): 5.85 (1H, s, H-7); 5.31 (1H, br s, H-3a); 5.04 (1H,
dt=9, 3 Hz, H-2a); 4.79 (1H, d=9 Hz, H-22p); 3.10 (1H, br t=7.8 Hz, H-9a); 2.10 (6H, s,
2X OAc); 1.99 (3H, s, OAc); 1.26 (3H, s, Me-21); 1.23 (3H, s, Me-26*); 1.21 (3H, s, C-
27); 1.04 (3H, s, Me-19); 0.85 (3H, s, Me-18).
13C NMR (CDCI3, Triacetate, 5): 202.1 (C-6); 172.2 (OAc); 170.3 (OAc); 170.0
(OAc); 164.8 (C-8); 121.6 (C-7); 84.4 (C-14); 79.8 (C-22); 77.0 (C-20); 70.4 (C-25); 68.7
(C-3); 67.2 (C-2); 51.0 (C-5); 49.6 (C-17); 47.6 (C-13); 40.4 (C-24); 38.4 (C-1*); 38.3 (C-
10*); 34.1 (C-4); 33.7 (C-9); 31.5 (C-15); 31.2 (C-12); 30.2 (C-27); 29.2 (C-16); 28.5 (C-
26); 24.7 (C-23); 23.8 (C-19); 20.6 (C-211); 20.5 (C-111); 17.5 (C-18).
IR, v max (KBr, cm'1): : 3400, 2960, 2870, 1643, 1450, 1370, 1050, 880.
UV max (ethanol): 243 nm (e 10,400)
Analysis calculated for C27H44O7: C, 67.47; H, 9.23. Found: C, 67.08; H, 9.30.
Ponasterone A and 2B. 38, 22a Triacetate
(22R)-2p, 3p, 14a, 20p, 22a-pentahydroxycholest-7-en-6-one. M.p. : 256-262 C
(Lit. 259-260 0 C, Nakanishi et a/. 1966)


53
The faster moving fraction [3-9] was obtained as a colorless, amorphous powder,
yield, 0.25 g.
1H NMR (CDCI3i 600 MHz, -40 0 C, 8) major rotamer: 1.02, s (H-19); 1.05, s OH-
16); 1.35, s OH-17); 1.46 (d, J=14.1 Hz, H-2cc); 1.78 (s, 9-0 acetate methyl); 1.98 (m, H-
6a); 2.0 (m, H-6p); 2.01 (s, 18 methyl); 2.02 (s, 7-0 acetate methyl); 2.32 (d, J=19.0,
14a); 2.48 (H-14p, dd, J=14, 7.5 Hz); 3.19 (H-3a, d, J=10.2); 4.20 (H-5p, br s); 4.48 (H-
13p, br t, J=7.2 Hz); 5.78 (H-7a, br m); 5.92 (H-9a, br); 6.52 (H-10p, br d, J=10.5 Hz);
7.43 (Ph-meta, t, J=7.8 Hz); 7.56 (Ph-para, t, J=7.8 Hz);7.87 (Ph-ortho, d, J=7.8 Hz).
13C NMR (CDCIa, 600 MHz, 5): 12.1 (C-18 methyl); 14.2 (C-19 methyl); 20.6 (7-
O acetate methyl); 21.2 (5-0 acetate methyl); 21.3 (9-0 acetate methyl); 24.8 (C-17
methyl); 27.0 (C-16 methyl); 29.1 (C-2); 33.9 (C-6); 34.8 (C-3); 46.4 (C-8); 46.4 (C-14);
62.0 (C-1); 69.9 (C-7); 70.7 (C-10); 72.0 (C-5); 75.9 (C-15); 76.9 (C-13); 77.3 (C-9);
128.8 (Ph-meta); 129.1 (Ph-ipso); 129.5(Ph-ortho); 133.4 (Ph-para); 134.4 (C-11); 144.5
(C-4); 147.3 (C-12); 164.3 (CO-Ph, s); 169.8 (CO-Acetate, s); 170.0 (CO-Acetate,
s);208.0 (C-13 C=0).
Analysis calculated for C3oH38010: C, 64.50; H, 6.86. Found: C, 64.68; H, 6.97.
The slower moving fraction was obtained as a colorless crystalline solid, yield,
0.3 g; m.p.225-232 0 C. It was found to be identical with the epoxide [3-8], described
above.
Brevifoliol-4,20-Diol [3-10]
To a solution of brevifoliol (0.4 g) in pyridine (10 ml) was added osmium tetroxide
(0.2 g) and the reaction mixture stirred for 1 h, after which time, the starting material was
replaced by a much slower moving component. After decomposing the excess reagent
with a solution of sodium bisulfite in pyridine, water and dilute sulfuric acid were added
and the mixture extracted with dichloromethane. After concentration, the product was


8
of its analogues crystallized out directly from the fractions. These are filtered and
purified further by recrystallization, or subjected to a small column.
This process using reverse phase column chromatography, gave not only higher
yields of taxol (0.02-0.04% vs. 0.01%) on a pilot plant scale, but also made possible the
simultaneous isolation of a number of analogues which have not been obtained from this
plant before. These included a 10-deacetyl baccatin III (0.02%); and a number of
xyloside analogues, chief among which being the 10-deacetyltaxol-7-xyloside, which can
now be isolated in yields of 0.1% or higher.
Based on the successful fractionation of the bark extract of T. brevifolia, this
technique was then ready for application to the other extracts such as the needles and
wood of T. brevifolia, and to the needles of two other species of Taxus. These
applications which gave practical methodology for processing these various extracts,
also yielded many interesting taxoid compounds and these experiments are all detailed
in this dissertation.
Although this work was started during 1991 and much of the expected work was
completed by late 1993, the world-wide interest in taxol research made a quantum leap
at about this time, with a phenomenal increase in publications dealing with all aspects of
taxol chemistry. Some of the compounds which were isolated for the first time in this
laboratory, and whose structures were determined, were rediscovered by others and
published. In spite of the enormous increase in the number of relevant publications,
most of these publications described the isolation of the minimum possible amounts of
the compounds, often as amorphous solids. Many determined their structures only
through NMR spectral interpretation, with little or no other physical characterizations,
elemental analyses, derivatizations or reactions. In at least a few examples, the
assigned structures were found to be wrong and were subsequently corrected once or


65
A reaction such as this has not been reported in the taxane series, resulting in
the loss of the oxy-isopropyl side chain. In taxol and related compounds containing the
conventional taxane skeleton, action of Lewis acids such as BF3 was studied and is
shown to produce one or two different changes, depending on the whether protic or
aprotic solvent is used. In one case, isomerization of the A-ring from a 6- to a 5-
membered A ring takes place with the oxy-isopropyl group attached at C-1. In the
second instance, the oxetane ring is opened to form a diol, or with the acetate group
migrating from C-4 to C-20 to give the 4-hydroxy-20-acetoxy compound. The reaction
described here appears to be a continuation of the action of the Lewis acid on the 5-
membered A ring, with the elimination of the oxy-isopropyl substituent.
4. Reaction with lodine/Silver Acetate [4-41
Once the structure of brevifoliol had been elucidated, the value and usefulness of
this relatively abundant compound in the needles of T. brevifolia was considered. Since
the addition of the N-benzoyl isoserine side chain at C-13 did not generate activity in the
final product, it was reasoned that the oxetane ring at the 4/20 position might be
necessary to produce activity. To this end, one approach was investigated, involving the
use of iodine in some form to add across the 4/20 double bond and thereby permitting
substitution with other groups.
Brevifoliol was found to react readily with bromine, but the reaction yielded
multiple products and considerable decomposition. Reaction with iodine was similarly
complex and led to much decomposition and dark colored products. With the idea that
addition of a silver salt which can remove the acid that might be produced, but not be too
strongly basic (e.g. silver oxide) and hence hydrolyze the ester functions, silver acetate
was selected for use with the iodine. The remote possibility that silver acetate might


CHAPTER 5
TAXANE CONSTITUENTS OF THE NEEDLES OF TAXUS X MEDIA
As discussed in Chapter 3, the yield of taxol from the bark of Taxus brevifolla by
using the conventional methods of isolation was of the order of 0.01%. It was also
shown that through the use of these same methods, no other useful analogues could be
isolated in any significant yields. As a consequence of these results and strong
ecological considerations, an intense search was started with the aim of finding a source
that is renewable, and which can match the bark in the yield of taxol. Many of the
available species of Taxus, as well as the various parts of these plants were examined
through the use of analytical high performance liquid chromatography (HPLC) and thin
layer chromatography (TLC). These searches led to the selection of the needles of the
ornamental yew, Taxus x media Hicksii as a possible answer to the problem. The
ornamental yew is capable of being grown in a nursery type setting, and on a large
scale, so that the needles may be clipped twice a year, and the taxol, which is found to
be present to the extent of 0.01% be isolated from them.
At the time of this research (1992-93), almost all of the studies carried out on this
species consisted of HPLC analyses. Other than the isolation of taxol by the standard
procedure with a total yield of 0.006%, no information had been published either on the
taxane constituents, or even a method for the practical isolation of them. In these HPLC
analyses, it was recognized that in the extracts of the ornamental yew, taxol was
accompanied by other co-eluting taxanes and these could contribute some errors in the
total yield calculations. These co-eluting taxanes were isolated in minute yields, in the
form of two components (0.8 mg and 1.2 mg); each representing an equilibrium mixture
74


21
[2-18] [2-19]
[2-20] [2-21]
Figure 2-4 : Nicolaous Taxane Ring Synthesis


43
carbon and the one at 5 75.9 assigned to the quaternary C-15 containing tertiary
hydroxyl, it appeared unlikely that such a rearrangement took place during the
acetylation. HETCOR and APT experiments corroborated these conclusions, thereby
agreeing with the structure determined by the x-ray crystallographic method.
Further analysis of the 1H COSY spectrum revealed an isolated spin system of
two doublets due to H-9p at 5 6.05 and H-10a at 8 6.53, with a pseudo-axial orientation
indicated by the degree of splitting (J=10.6 Hz); and significant broadening of the signal
at 5 6.05. Some amount of the deshielding of H-10a relative to H-9p was expected, due
to the adjacent double bond, which makes the C-10 position allylic. The presence of a
benzoate at this position would be expected to cause a further downfield shift based on
analogous compounds already known (Chu et al. 1992). With a thorough analysis of the
1H NMR and 1H COSY spectra, the signal at 5 4.38 (t, 7.6 Hz) was assigned to the H-
13p proton, which coupled strongly with H-14p at 8 2.46 (dd, 14.0,7.6 Hz.); as well with
H-14a at 5 1.25 (dd, 14.0, 7.6 Hz). Weak long range coupling to the C-18 methyl
protons at 8 2.01 was also evident, as the slight broadening of this peak is generally
attributed to this long range coupling in other taxanes.
The isolated spin system of H-5p, H-6a, H-6p and H-7a is easily identified in
most taxanes, with a tendency to show a sharp multiplet for H-7a and broader, poorly
resolved splitting for H-5p, especially if H-5p is not esterified (Della Casa de Marcano &
Halsall, 1970; Rao et al. 1995). The H-5p broad singlet at 8 4.45 interacts with the H-
6a multiplet at 8 1.86, which interacts with the H-6p multiplet at 5 2.02, which in turn
interacts with the H-7a signal at 8 5.56 (dd, 5,11 Hz.). In many cases esterification of a
hydroxyl causes a deshielding effect on the related proton of about 1 ppm. The
chemical shifts and splitting patterns indicated that the acetate groups were at C-7 and
C-9, with the benzoate at C-10.


113
Nicolaou, K. C.; Yang, Z.; Liu, J.-J.; Nantermet, P. G.; Claiborne, C. F.; Renaud, J.; Guy,
R. K.; & Shibayama, K. J. Am. Chem. Soc. 117:645-652 (1995d).
Nicolaou, K. C.; Yang, Z.; Liu, J. J.; Lleno, H.; Nantermet, P. G.; Guy, R. K.; Claiborne,
C. F.; Renaud, J.; Couladouros, E. A.; Paulvannan, K.; & Sorenson, E. J.
Nature, 367:630-634 (1994b).
Ojima, I.; Habus, I.; Zhao, M.; Georg, G. I.; & Jayasinghe, L. R. J. Org. Chem. 56:1681-
1683 (1991).
Ojima, I.; Habus, L; Zhao, M.; Zueco, M.; Park, Y. H.; Sun, C. M.; & Brigaud, T.
Tetrahedron, 48:6895-7012 (1992).
Rao, K. V. J. Heterocyclic Chem. 34:675-680 (1997).
Rao, K. V. Pharm. Res. 10:521-524 (1993).
Rao, K. V.; Bhakuni, R. S.; Juchum, J.; & Davies, R. M. J. Liq. Chrom. & Rel. Technol.
19:427-447 (1996).
Rao, K. V.; Hanuman, J. B.; Alvarez, C.; Stoy, M.; Juchum, J.; Davies, R. M.; & Baxley,
R. Pharm. Res. 12:1003-1010 (1995).
Samaranayake, G.; Magri, N. F.; Jitrangsri, C.; & Kingston, D. G. I. J. Org. Chem.
56:5114-5119 (1990).
Samaranayake, G.; Neidigh, K.; & Kingston; D. G. I. J. Nat. Prod. 56:884-898 (1993).
Schiff, P. B.; Fant, J.; & Horwitz, S. B. Nature, 277:665-667 (1979).
Schopf, C.; & Ross, F. Naturwissenschaften, 26:772-776 (1938).
Senilh, V.; Blechert, S.; Colin, M.; Gunard, D.; Picot, F.; Potier, P.; & Varenne, P. J.
Nat. Prod. 47:131-137 (1984).
Shapiro, R. H. Org. React. 23:405-507 (1976).
Sharpless, K. B.; Amberg, W.; Beller, M.; Chen, H.; Hartung, J.; Kawanami, Y.; Lbben,
D.; Manoury, E.; Ogino, Y.; Shibata, T.; & Ukita, T. J. Org. Chem. 56:4585-4588
(1991).
Sucrow, W. Chem. Ber. 99:2765-2777 (1966).
Swift, L. J. J. Am. Chem. Soc. 74:1099-1100 (1952).
Swindell, C. S. Org. Prep. Proced. Inti. 23:467-543 (1992).
Wani, M. C.; Taylor, H. L.; Wall, M. E.; Coggon, P.; & McPhail, A. I. J. Am. Chem. Soc.
93:2325-2327 (1971).
Webster, L.; Linsenmeyer, M.; Millward, M.; Morton, C.; Bishop, J.; & Woodcock, D. J.
Natl. Cancer Inst. 85:1685-1690 (1993).


102
culture (Stierle et al. 1993). It is a mystery how these two organisms that are so different
are able to make a complex diterpene like taxol. In spite of this taxol producing fungus,
it is most likely that the usnic acid was actually produced by lichens growing on the bark,
and is not produced by the tree. Usnic acid is mentioned here because it was difficult to
characterize and not uncommon in samples processed in this lab. Usnic acid is also
known to be quite toxic and should be handled carefully.
Betuloside is a simple glycoside first isolated from the plant Betula pendata
(Khan 1966). Animal studies using hepatotoxic agents indicate that betuloside has
significant hepatoprotective activity. The mechanism by which this compound protects
the liver is not known, but teas made from plants containing betuloside have been used
in India for centuries for various problems. Betuloside is just one more example of the
usefulness of preparative scale reverse phase chromatography.
Experimental
Analytical HPLC was performed using two different systems. For determinations
of purity and quantitative information on composition, a setup with a Waters 600 E pump
with gradient control, a Waters 996 photodiode array detector, and a Waters 717
autosampler, coupled with an NEC-386 computer and printer was used. Waters
Millennium 1.1 program was used with the photodiode array system. For routine use, a
combination of a Waters 501 pump with a U6K injector, a 486 tunable absorbance
detector and a Goerz Servogor 120 recorder was used.
For analytical purposes, standard columns packed with C-8 bonded silica
(Whatman Partisil, 4.6 mm x 25 cm, 5pi) were used with either of the solvents: 50%
acetonitrile in water, or a 5:4:1 mixture of acetonitrile, water and methanol.
For preparative scale purposes, stainless steel columns of two sizes were used:
4 x 4 and 6 x 6, fabricated by Fluitron Inc. (Ivyland, PA) and rated to 200 psi. The


56
Debenzoyl Brevifoliol [3-151
A solution of brevifoliol (0.3 g) in 30% methanol in water (10 ml) was heated in a
sealed tube at 135 0 C for 90 min. The cooled mixture was neutralized with sodium
bicarbonate and extracted with chloroform. The extract was purified by chromatography
on a C-8 reverse phase column using a step gradient of 25-60% acetonitrile in water in
5% increments of solvent concentration. Elution with 30-35% acetonitrile in water gave
the major product, which was obtained as a white powder, yield, 0.1 g. The 1H and 13C
NMR spectra revealed the loss of the benzoate from the C-10 position and retention of
the acetate substituents at C-9 and C-7
Analysis calculated for C24H3608: C, 63.70; H, 8.02. Found: C, 63.89; H, 7.88.
Benzoic acid was also isolated from the reaction mixture and confirmed using
NMR and UV analyses.


BIOGRAPHICAL SKETCH
Richard Michael Davies was born in Cocoa, Florida on March 21, 1959 to Dan
and Ruth Lewis Davies. He attended Rockledge High School where he participated in
the cross country and track teams, the schools concert and marching bands, and other
extracurricular activities and societies. He attended Brevard Community College for one
year before enrolling at the University of Florida to complete bachelors degrees in
chemistry and then pharmacy. While studying pharmacy he also worked on projects
with Professor Rao in the laboratory. Interest in the chemistry and research of natural
products and the development of new anticancer therapies brought him back for
graduate studies.
115


76
The reverse phase column procedure was carried out similar to what was used
with the needle extract of T. brevifolia, as described in Chapter 3. Approximately 200 g.
of the chloroform extract was dissolved in acetonitrile (see experimental) and stirred with
the equilibrated C-18 bonded silica. This slurry was then diluted to the appropriate
concentration of the acetonitrile and the added to the column prepared from 800 g of the
C-18 silica. Elution was carried out using a step gradient: 30, 35, 40, 45, 50 and 60%
acetonitrile in water, and the eluate collected in fractions of 200 ml. As was seen in the
case of the columns on the bark extract of T. brevifolia, when the fractions remained at
room temperature for about a week, crystals began to separate from the fractions in
different regions of the elution. These were filtered and further purified by either
recrystallization or a small column of normal phase silica where necessary.
The progress of elution of the column is shown in Figure 5-3. As anticipated,
taxol was accompanied by two other taxanes, which were present in higher
concentrations than taxol. However, all of these crystallized out of the fractions.
The early fractions contained the bulk of the UV absorbance, and from these
could be isolated a crystalline solid, which was a non-taxane compound. The next major
component that emerged with the 35-40% acetonitrile in water was shown to be
brevifoliol as described in Chapter 3. With the 45-50% acetonitrile and water solvent
were eluted taxane I, taxane II, followed by taxol, all of which crystallized from their
respective fractions, with some overlap.
The column was finally washed with a mixture of methanol and ethyl
acetate/ligroin (2:1:1) which stripped the column of all the waxes, chlorophylls and other
pigments. After, washing with methanol, followed by 25% acetonitrile and water the
column was made ready for another run. Figure 5-1 shows the steps involved in the
fractionation of the extract of Taxus x media Hicksii.


86
Ozonolvsis of Compound [2-21
A solution of compound [2-2] (1 g.) in chloroform and methanol (30 ml, 9:1) was
cooled in a dry ice and acetone bath and saturated with ozone for 10-15 min. TLC
showed that the starting material was absent and ozonide being formed (detected by
spraying with starch and potassium iodide which gave a blue color). After the
decomposition of the ozonide by dimethyl sulfide, the reaction mixture was washed with
water and concentrated to dryness. The product was crystallized from acetone in ligroin
to obtain colorless needles, yield, 0.8 g, m.p. 168-170 C, [a]D23 +130 (c 1.06, pyridine);
HRMS: 569.2239, Calc, for C27H36013, 569.2234.


20
After removal of the acetonide and protection of the primary hydroxyl at C-20 to
make, the benzyl group was removed with catalytic hydrogenation and the 7-O-triethyl
silyl protecting group was introduced to give [2-23], Selective deacetylation of the
primary acetate then provided the triol for the formation of the oxetane of ring D, which
involves monotosylation at C-20 (primary OH) and triflate formation at C-5 (secondary
OH) to produce [2-24], Oxetane formation with a 60% yield occurs after mild acid
treatment with catalytic camphorsulfonic acid (CSA) in methanol, followed by treatment
with silica gel in dichloromethane.
Acetylation of the C-4 position (tertiary hydroxyl) was followed by regioselective
ring opening of the carbonate to the hydroxybenzoate functionality, both with good
yields. The C-13a oxygen is introduced with pyridine chlorochromate in 75% yield
followed by stereospecific reduction of the ketone [2-25] using NaBH4 in methanol in
excess, for 83% yield. The hydroxyl is esterified using Ojimas (3 lactam synthon [2-14]
(Figure 2-2) using the strong base sodium-hexamethyldisilazane for 87% yield based on
90% conversion. Removal of the triethylsilyl groups with hydrogen fluoride in pyridine
(HF-Pyr) completes the synthesis of taxol in 80% yield.
Other Synthetic Approaches
As previously mentioned Holtons group published a total synthesis of taxol in
early 1994 at about the same time as Nicolaou, but their approach was quite different,
with only a few reactions in common. Studies involving the fragmentation of bicyclic
epoxy alcohols, referred to as epoxy alcohol fragmentation, were the cornerstone of
their syntheses of bicyclo[5.3.1] systems, including the unnatural epimer of (+)-taxusin
[2-26], known as (-)-taxusin or enf-taxusin [2-27] (Figure 2-6).


83
Taxanes I and II [5-11 and [5-21
The crude crystals that separated out from the fractions (8 g) consisting of [5-1],
[5-2] and taxol [5-3] were processed by two methods. In one, the mixture (4 g) was
taken up in chloroform and ligroin (3:1, 50 ml) and applied to a column of normal phase
silica (60 g). The mobile phase was successively changed to chloroform, 2% acetone,
5% acetone, 2% methanol and 5% methanol in chloroform. Compounds [5-1] and [5-2]
appeared in the 2-5% acetone and chloroform eluate partially separating from each
other. Continuing with 2% methanol in chloroform gave taxol with small amounts of [5-1]
and [5-2],
To obtain further purification of [5-1] and [5-2] the mixture was taken up in 40%
acetonitrile and water and applied to a column of C-18 bonded silica. The column was
eluted with 45 and 50% acetonitrile and water. As the fractions from the 45% acetonitrile
and water elution stood for about a week, crystals appeared over a range of tubes and
these were filtered in groups. Although [5-1] and [5-2] were separated, such that each
contained the other to the extent of 10% or less, recrystallization gave worse mixtures,
thus suggesting that isomerization (or equilibration) was taking place during the process.
Data obtained on a crystalline (9:1 mixture of [5-1] and [5-2]: m.p. 136-138 C, [a]D23
+214 (c 1.04, CHCI3); (lit. Appendino et at. 1992 on an amorphous sample, m.p. 163-
165 C and [a]D23 +185 ).
Analysis calculated for C31H3808, H20: C, 66.89; H, 7.24. Found: C, 66.51; H,
7.19.
The 1H- and 13C NMR spectra of the crystalline [5-1] and [5-2] gave evidence of
mixtures of two compounds. From the spectral data, these two were Identified as a
mixture of 5-0-cinnamoyl-9-acetyl taxicin l [5-1] and 5-0-cinnamoyl-10-acetyl taxicin I [5-
2] described by Chmurney et al. (1993) from Taxus x media Hicksii and by Appendino et


88
slow. However, T. floridana responded satisfactorily and could be propagated under
available conditions. This species was therefore studied in our laboratory for its taxane
constituents.
The needles of T. floridana were collected from the campus and were extracted
without drying. After extraction with methanol as before, concentration to remove the
solvent, and partition between chloroform and water, the organic layer was concentrated
to a dark green semisolid. Fractionation was again carried out using the reverse phase
column techniques as was described under the needles of the other Taxus species in
Chapters 3 and 5.
The crude chloroform extract was first tested by analytical HPLC to see the
elution pattern of the taxane constituents. Taxol was clearly recognizable at its normal
location, and in contrast to the observation with the needles of Taxus x media Hicksii,
where there were co-eluting taxanes, the taxol from the extract of the needles of T.
floridana was relatively free from such interfering taxanes. There were other taxanes
situated at different locations.
Elution of the reverse phase column was carried out using a step gradient of 30,
35, 40, 45, 50 and 60% acetonitrile/ water. When the fractions were let stand at room
temperature for 3-5 days, taxol and several other taxanes crystallized out as before.
The initial eluates from the column from 25-30% acetonitrile/ water contained
highly polar phenolic constituents. The first taxane component to appear from the
reverse phase column emerged with the 30-35% acetonitrile/water solvent, and
crystallized almost immediately. This was found to be 10-deacetyl baccatin III [2-7], The
next taxane was eluted with the 40% acetonitrile/ water and it was found to be identical
with brevifoliol [3-1], With the 45% acetonirile/ water, was eluted another crystalline
compound which was found to be a new compound, and was named taxiflorine [6-1].
Continued elution with 50% acetonitrile/ water gave two crystalline compounds in


94
conditions. For convenience, they were extracted in fresh state. Several batches were
collected from the plants on the campus ranging from 1-20 Kg. With the hydroponically
grown plants, the smallest amount was 54 g of the fresh needles, and the largest, 2.5
Kg. The method of extraction, concentration, partition between water and chloroform
were the same as was described in Chapter 3. The yield of the chloroform extract varied
from 20-25 g per Kg of the fresh leaves and twigs.
Before the conditions for the reverse phase column chromatography were fully
developed, an alternative procedure was tested for the purpose of making the sample
preparation easier. The chloroform extract, especially of the needles, usually contains a
higher amount of waxes, chlorophylls and other lipid-soluble components and can
potentially pose problems in the preparation of the sample for applying to the column.
For this reason, a study was made to see at what concentration of methanol or
acetonitrile would be needed to obtain an essentially clear solution that can be applied to
the column.
It was found that at least an 80% methanol in water would be necessary to
prepare a 10% solution of the chloroform extract solids. However, at this concentration
of the solvent, taxol and most of the other taxanes do not remain on the column. It was
also found that if such a solution is passed through a column of C-18 bonded silica,
almost all of the chlorophylls, waxes etc., remain on the column, while taxol and other
taxanes appear in the effluent and washes. There was also a reduction of the solids
content by 50-60%, which meant that the waxes and other lipid-soluble components
account for this much of the chloroform extract and can be readily removed from the
sample.
The material obtained by concentrating the effluent and washes was much less
difficult to prepare as the slurried sample for applying to the column. In order to carry
out this wax-removal operation, it was only necessary to use a ratio of 3 g of the C-18


TABLE OF CONTENTS
page
ACKNOWLEDGMENTS iii
LIST OF TABLES vii
LIST OF FIGURES vii:
ABSTRACT x
CHAPTERS
1 HISTORICAL OVERVIEW OF TAXUS 1
Background of Research at the University of Florida 6
Methods 6
2 A SELECTED REVIEW OF THE LITERATURE ON TAXUS 11
Earlier Studies 11
Studies after the Discovery of Taxol 14
Semi-synthesis of Taxol 17
Total Synthesis 18
Other Synthetic Approaches 20
General Structural Features of Taxanes 24
Taxa-4(20);11-dienes 25
4(20)-Epoxides 25
Oxetanes 26
Abeotaxanes 28
3 TAXANE CONSTITUENTS OF THE NEEDLES OF TAXUS BREVIFOLIA 29
Fractionation of the Needles ofTaxus brevifolia 29
Brevitaxane A (Brevifoliol) [3-1] 31
Hydroxyl Functionalities 33
4/20 Unsaturation 36
Number and Nature of the Oxygen Substitution 37
Experimental 44
Extraction of the Needles of Taxus brevifolia 44
Reverse Phase Column Chromatography: 45
Brevifoliol [3-1] 46
Brevifoliol-5-Monoacetate [3-2] 48
iv


59
observations suggest that the C-15 hydroxyl might be responsible for the blurring of the
signal of H-9(3, and acetylation of this hydroxyl has eliminated that interaction.
Further support for the acetylation having taken place at the C-15 hydroxyl is
shown by the appearance of another acetyl methyl signal at 5 2.11 in the 1FI NMR
spectrum. The appearance of two more signals in the 13C NMR spectrum was also
evident, in which 5 acetyl methyl and 5 acetyl-carbonyl signals were present, with the
fifth one at 5 21.7 and 5 169.5, respectively. Also in the 13C NMR spectrum, a significant
downfield shift from 5 75.6 to 5 87.2 for the C-15, and an upfield shift from 5 27.0 to 5
24.8 and from 5 23.1 to 5 22.0 for the signals due to C-16 and C-17 respectively,
completes the evidence to indicate that the 15-hydroxyl was acetylated to give [3-5],
A comparison of the 13C NMR spectral data of brevifoliol, the 5-monoacetate
[3-2], the 13-monoacetate [3-3] (naturally occurring and confirmed through semi
synthesis); the 5, 13-diacetate [3-4] and of the BF3-catalyzed acetylation product [3-5]
were shown in Chapter 3 in Table 3-1.
2. Oxidation
Oxidation of brevifoliol with manganese dioxide gave the monoketone, the NMR
spectral data of which showed that the 13-hydroxyl was oxidized, leading to the structure
[3-6], as described in Chapter 3. Oxidation of [3-1] with Jones reagent gave initially, the
same 13-monoketone [3-6], but on further reaction, this was replaced by a faster moving
compound [4-1], whose spectral data pointed to an unexpected course of reaction.
The molecular formula of the product, C31FI3609 (MH+, 553) indicated the loss of
4 protons, as compared to brevifoliol. Although this might indicate that both hydroxyls
were oxidized to give the diketone [4-2], certain features suggested otherwise. To begin
with, the 1FI NMR spectrum showed broad peaks which indicated the existence of a


71
showed that the starting material was absent and was replaced by two faster moving
components. The mixture was filtered and the filtrate washed successively with
aqueous sodium bisulfite and water and concentrated to dryness. Chromatography on
silica gel in 4:1 chloroform/ligroin gave the major band, which was obtained as a
colorless crystalline solid, total yield, 0.12 g; m.p. 250-252 0 C.
1H NMR (CDCI3, Varan Unity 600 MHz, 5): 1.16, s (H-19); 1.03, s (H-16); 2.08
(cm, H-14a); 1.27, s (H-17); 2.32 (d, J=13.8 Hz, H-2a); 1.92 (s, methyl, 9-acetate); 1.84
(m, H-6a); 1.98 (m, H-6(3); 2.26 (s, H-18); 2.05 (s, 7-acetate methyl); 3.30 (d, J=13.8 Hz,
H-2p); 2.28 (cm, H-14P); 2.95, br s (C-15 OH, exchangeable with D20); 4.38 (t, J=7.2
Hz, H-13P); 4.28 (br s, H-5p); 4.54 (d, J=13.2 Hz, H-20 B); 4.54 (d, J=13.2 Hz, H-20 B);
5.51 (dd, J=4.8, 11.4 Hz, H-7a); 6.07 (d, J=10.1 Hz, H-9a); 6.54 (d, J=10.1 Hz, H-10p);
7.45 (t, J=7.5 Hz, H-Bz-mefa); 7.57 (t, J=7.5 Hz, H-Bz-para); 7.89 (d, J=7.5 Hz, H-Bz-
ortho).
13C NMR (CDCI3, Vahan VXR 300 MHz, 5): 13.1 (C-18 methyl, q); 16.1 (C-19
methyl, q); 20.8 (7-0 acetate methyl, q); 21.5 (9-0 acetate methyl, q); 25.3 (C-17 methyl,
q); 26.9 (C-16 methyl, q); 31.7 (C-2, t); 38.2 (C-6, t); 142.7(C-3, d); 45.2 (C-8, s); 38.6
(C-14, t); 65.6 (C-1, s); 67.2 (C-7, d); 76.2 (C-10, d); 69.8 (C-5, d); 74.3 (C-15, s); 84.0
(C-13, d); 72.4 (C-9, d); 64.4 (C-20, t); 128.8 {C-Bz-meta, d); 129.2 (C-Bz-/pso, s);
129.4(C-Bz-ortho, d); 133.4 (C-Bz-para, d); 135.5 (C-12, s); 142.7 (C-4, s); 146.4 (C-11,
s); 164.5 (CO-Ph, s); 169.4 (CO-Acetate, s); 170.3 (CO-Acetate, s).
FAB-MS (dithiothreotol/dithioerythrotol / TFA, m/z): 577 [M+Na]; 537 [M+ NaOH];
433 [M+-Na02C7H5]; 373 [M+-Na02C7H5 -HOAc]; 313 [M+-Na02C7H5 2x HOAc]; 253
[M+-Na02C7H5-3x HOAc],
CI-MS (methane, m/z): 537.9 [MH+ -H2Oj; 373.6 [MH+-H20-HOAc- C6H5COOH],


CHAPTER 4
SOME UNUSUAL REACTIONS OF BREVIFOLIOL
The isolation and structural elucidation of brevifoliol [3-1] was described in detail
in Chapter 3. Brevifoliol occurs to the extent of 0.2-0.3 % in the needles of T. brevifolia,
and to a lesser extent in the bark of the same species, in the needles of T. x media
Hicksii (Rehd.) and in the needles of T. wallichiana. (Georg et al. 1993). Large
quantities of the crystalline compound can be readily isolated from the needles of T.
brevifolia, which is the best source for the compound.
However, other than acetylation to a diacetate (Balza et al. 1991); and the
attachment of the N-benzoyl phenyl-isoserine side chain at the C-13 position (Georg, et
al. 1993); no record of its reactions reflecting its various functional groups has been
published. This paucity of such information is in keeping with the current trend that, in
spite of the virtual explosion of new taxanes that were isolated and characterized
structurally by spectral data over the past five years. Very few have been investigated
for their chemical reactions, for the relative reactivities of similar functional groups, or for
any unusual reactions as a consequence of their stereochemical disposition.
Understanding the products of the various transformations that a compound can
be subjected to and the relative rates of reaction can lead to important insight
concerning entity, and for the general advancement of chemistry. Although the structure
of brevifoliol is now well established as a result of spectral and x-ray crystallographic
data, we can gain insight into this molecule through reactions such as those described in
Chapter 3. In addition to these, a number of reactions also carried out for the purpose of
57


ACKNOWLEDGMENTS
Much of the work in this dissertation was done with the guidance and expertise of
the late Dr. K. V. Rao, who passed away on February 20, 1998. Professor Rao and I
had known each other since 1981 and I am grateful to him for encouraging me to return
for graduate studies and for his true friendship with me. I would also like to thank his
wife and children for their much appreciated support, friendship and encouragement.
I wish to thank Dr. John Perrin for assuming the position of chairman of my
supervisory committee and for his kind encouragement and guidance. He has helped
me in many ways and I am very grateful to him for his persistence in pushing me to
complete this work. I would also like to thank Dr. Margaret James, Dr. Jonathan Eric
Enholm, Dr. Kenneth Sloan, and Dr. Stephen Schulman for participating on my
supervisory committee and for their thoughtful advice and expertise.
I wish to thank my mother and father for their kind encouragement and love, and
also my three sisters, brothers-in-law, niece and nephews. I would like to thank the
Graduate School and many other University of Florida personnel for all of the kind
assistance they have provided, especially Gladys Jan Kalman and Nancy Rosa for all of
their helpful assistance.


35
[3-6] 13-Ketone
[3-8] 4,20-Epoxide
R-i R? R3
[3-1] H H H
[3-2] Ac H H
[3-3] H Ac H
[3-4] Ac Ac H
[3-5] Ac Ac Ac
[3-11] Hydrolysate, R1 = R2 = H
[3-15] Debenzoyl, R1 = R2 = Ac
Figure 3-2 : Brevifoliol and Reaction Products


95
silica for 1 g of the extract. The chlorophylls and waxes that were held up on the column
could be readily removed by washing with a mixture or methanol/ ethyl acetate/ ligroin
(2:1:1). Examination by TLC showed that this wash did not contain any taxane
constituents. This procedure was not used in later trials, as methods were found for a
successful and convenient preparation of the sample slurry made it unnecessary, as
described in Chapter 3.
Characterization of the Taxane Constituents of Taxus floridana
The results given here represent the work carried out on a 20-Kg batch of the
fresh needles.
10-Deacetvl Baccatin III [2-71:
Elution with 30% acetonitrile in water gave this component which crystallized
almost immediately. After a week, the crystals were filtered off, dried and recrystallized
from methanol/ chloroform, yield, 12 g (0.06%); m.p. 232-234 C. The spectral
properties were identical with those described in the literature (Chauviere et at. 1981,
Appendino et at. 1993b).
Brevifoliol [3-11:
Fractions from the 35-40% acetonitrile eluate, which contained this component
but did not give a crystalline solid directly, were combined, concentrated to dryness and
the solid (3 g) was applied to a normal phase silica column (120 g) in chloroform.
Elution with 2% methanol in chloroform gave the major band, the fractions from which
were combined, concentrated and the solid crystallized from acetone / ligroin to give 1 g
of [3-1], m.p. 220-222 C. Its spectral data proved to be identical with those described in
Chapter 3.


170.4-AcCO 83.7
170.1-AcCO 81.4
169.8-AcCO 78.7
169.1-AcCO 76.3
168.8-AcCO 74.9
166.8-BzCO 73.1
141.1-12 71.7
133.6-Bz(p) 70.3
133.5- 11 69.6
129.9 -Bz(m) 47.2
129.2-Bz(q) 45.7
128.5-Bz(o) 42.7
5 35.0- 14*
4 34.4 6*
2 28.2- 17 Me
20 22.6-Ac Me
1 22.2-16 Me
7 21.3-Ac Me
9 21.1-AcMe
13 20.8-Ac Me
13 20.7-Ac Me
3 14.9 18 Me
8 12.7 19 Me
15 Interchangeable
Figure 6-2 : Carbon NMR Spectrum of Baccatin VI


85
The analytical and spectral data of [2-2] indicated that it was identical with 5-0-
cinnamoyl taxicin II: 2a,9a,1 Op-triacetate described by Appendino et al. (1992) and
Baxter et al. (1962).
Taxol 5-31
In the silica column described above under the purification of compounds [5-1]
and [5-2], taxol (approximately 0.8 g) was eluted by 2-5% methanol in chloroform. A
small portion was crystallized from acetone and ligroin to obtain colorless needles of
taxol. The 1H NMR spectrum showed that the compound still had appreciable quantities
of compounds [5-1] and [5-2]. To remove these compounds completely, ozonization
was carried out on the rest of the sample in chloroform and methanol (9:1, 30 ml) at -70
C for 10-15 min. The reaction mixture was treated with dimethyl sulfide (0.5 ml) and let
stand at room temperature for 2 h.
After concentration to dryness, the sample was chromatographed on normal
phase silica (25 g) in chloroform. Elution with 2% methanol in CHCI3 gave taxol which
was crystallized from ligroin to obtain pure taxol, free from compounds [5-1] and [5-2],
yield, 0.5 g (0.012%). Its spectral properties agreed with those of an authentic sample.
Alternatively, the crude crystalline solid consisting of compounds [5-1] ,[5-2] and
taxol was directly ozonized in chloroform and methanol as before (but without he
intermediate silica column purification). After decomposition of the ozonide, and
concentration, the sample was subjected to chromatography and taxol isolated from the
column. It was crystallized as before to yield 0.75 g (0.015%). The products of
ozonization of compounds [5-1] and [5-2] were more polar than the original compounds
and separated from taxol in the normal phase silica column.


60
rotameric equilibrium, which was confirmed by a spectrum taken at -40 0 C, in which two
sets of peaks with a 5:1 ratio were seen.
In the major rotamer, the coupling between the C-9 and C-10 protons was found
to be 4 Hz, in contrast to the value of 10.5 Hz shown by brevlfoliol [3-1], (a similar
diketone prepared from 2-acetoxy brevifoliol (taxchinin A); described by Fuji et al. (1992)
and Appendino et al. (1993) also showed a coupling of 4 Hz. Next, a singlet appeared at
9.4 ppm, which interacted in the HETCOR spectrum with the peak at 194 ppm. The
latter showed a negative signal in the Attached Proton Test (APT). These observations
indicated the presence of an aldehyde functionality, presumably at C-20. Additionally, in
the spectrum of the diketone such as [4-2], the C-5 proton signal was absent, as
expected, and the exocyclic methylene protons appeared as two singlets at 5 5.06 and 8
5.94 ppm. The corresponding carbon signals appearing at 5127 and 8143.4 ppm.
However, in the Jones oxidation product, the singlets due to the exocyclic
methylene protons were absent, and the characteristic C-20 carbon signal which
appears in the 8110-8120 ppm region was also missing. Instead, a signal was found at
5 6.73 ppm, which interacted with the signals at 82.5 and 52.7 (C-6 protons); and in the
HETCOR spectrum, with the signal present at 8 147 ppm, and this latter gave a negative
signal in the APT spectrum. These data seem to suggest that the product is not the
5,13-diketone [4-2], but an aldehydic product with a double bond present at C-4/C-5, as
shown in [4-1],
One possible explanation is that the sulfuric acid in the reagent caused hydration
of the 4/20 double bond, and the primary alcohol so generated was oxidized to the
aldehyde, followed by dehydration to yield the 4/5-double bond.


6
Two total syntheses of taxol have been recorded, and several other approaches
towards the synthesis have also been reported in the literature as discussed briefly in
Chapter 2 (Nicolaou et at. 1994a, 1994b, 1995a, 1995b, 1995c; Holton et at. 1994a,
1994b). Although these methods demonstrate remarkable achievements in the field of
synthetic organic chemistry, they do not offer a practical method for the large-scale
production of taxol or its analogues at this time.
Background of Research at the University of Florida
As part of one of these alternative quests, the National Cancer Institute hoped
that instead of using the bark of the Pacific yew, the plant should be grown under
hydroponic conditions, and they wanted to know whether plants grown in this manner
would produce enough taxol for isolation. This laboratory was approached with this idea
in early 1990, and with collaboration from Prof. George Hochmuth, Jr., of IFAS,
University of Florida, the project was started. More on this aspect will be discussed in
Chapter 6. In order to learn the current knowledge concerning the analysis and isolation
of taxol, the pertinent literature was consulted. This yielded only a few papers on the
isolation of taxol and taxoids, which were outlined above.
Methods
In general, the methods were found to be too cumbersome for others to repeat.
For example, one of these publications in which isolation of taxol and its analogues was
described from T. wallichiana, used the following steps, starting with the concentrated
ethanolic extract of the plant:
1. Partition between water and hexane
2. Extraction of the aqueous phase with chloroform
3. Silica gel chromatography on the chloroform extract
4. A second silica gel chromatography


42
carbonyl, the C-8 signal appears near 5 58, so the signal at 8 62.4 immediately raised
questions about the true structure of brevifoliol. This signal did not fit the normal
chemical shift pattern of any naturally occurring taxanes known at that time. In the
absence of a carbonyl group at C-9, the C-8 carbon usually falls in the region of 8 40-50
ppm.
Unable to satisfactorily explain this unusual peak position, the Chemistry
Department was contacted about crystallographic services. X-ray crystallographic
analysis was performed by Dr. K. A. Abboud on the 5-monoacetate [3-2], Surprisingly,
the presence of an unusual 5/7/6 ring system was evident, where the normal 6-
membered A ring of the conventional taxane system was rearranged to form a 5-
membered ring with the carbons 15, 16 and 17 moved out of the ring system to form a
hydroxy isopropyl group at C-1. Since the x-ray structure was obtained on brevifoliol-5-
acetate, it was important to establish whether brevifoliol itself had this rearranged taxane
skeleton, or if the rearrangement could have occurred during the acetylation.
This structure represented a departure from the existing naturally occurring
taxane structures available at that time, previously seen only as a product of
rearrangement under strongly acidic conditions (Samaranayake et al. 1990).
Crystallography of the original compound was not done because it failed to yield
adequate crystals for analysis without prior acetylation. This made it necessary to
determine whether this new ring structure was naturally occurring, or formed during the
acetylation.
In one such ring contraction, taxol underwent rearrangement of the A-ring,
accompanied by dehydration, to produce an isopropenyl group at C-1, as well as other
changes such as the opening of the oxetane ring. Since the 13C NMR spectra of both
brevifoliol and its monoacetate showed these signals at 8 62.4 assigned to the C-1


This dissertation was submitted to the Graduate Faculty of the College of
Pharmacy and to the Graduate School and was aco
requirements for the degree of Doctor of Philosop^
December 1998
Dean, Graduate School


15
assays were all used at various points to monitor the fractionation, resulting in the
isolation of taxol as the active principle.
In the 1960s the most straightforward and reliable method for the determination
of complex chemical structures was x-ray diffraction analysis of a suitable crystal, also
known as crystallography. Taxol crystallizes as thin needles not suitable for x-ray
studies, but a tetraol derivative was amenable to x-ray studies. The structure of taxol [2-
8] was determined by methanolysis (Figure 2-2); which yielded two compounds: the
methyl ester of N-benzoyl phenylisoserine and an alcohol component shown to be a
taxane tetraol [2-10a], This tetraol skeleton was converted into a 7,10-bis-iodoacetate
derivative and, unlike all of the taxanes studied earlier, taxol showed the following
unique features:
1. A taxane skeleton with an oxetane ring system involving C-4, C-5 & C-20
2. An ester side chain consisting of N-benzoyl phenylisoserine at C-13
3. A carbonyl function at C-9
Baccatin III (Figure 2-1) and its 7a -epimer, baccatin V (Figure 2-2); were shown
to be similar to taxol, having the oxetane ring and the C-9 carbonyl function. These
epimers yielded better crystals and x-ray crystallography was performed. Baccatin III [2-
6b] lacked the ester side chain present in taxol. Still another analogue, known as 10-
deacetylbaccatin III (10-DAB, [2-7]) was later found to be much more widely distributed
in Taxus spp., especially in the needles of Taxus baccata. This became an important
taxane because it could be converted into baccatin III and later to taxol by the
reattachment of the N-benzoyl phenylisoserine side chain at C-13 (See Figure 2-2).
Taxol showed significant antitumor activity against a variety of in vivo murine
tumors including B-16 melanoma and several human xenografts, which qualified it for


55
IR i/ max (KBr, cm'1): 3565 (OH, sharp); 2960, 1740 (C=0); 1370, 1230, 1030.
MS(FAB): 580 [MH+].
Periodate Oxidation of [3-111 to [3-131
A solution of [3-11] (0.2 g) in methanol (5 ml) was treated with sodium periodate
(0.3 g) in 1N sulfuric acid (2 ml). After 30 min, TLC showed that the starting material
was absent and was replaced by a faster moving product visible under the UV light,
unlike the starting material. After dilution with water, the mixture was extracted with ethyl
acetate and the extract concentrated to dryness. The crude dialdehyde [5-13] was not
further purified before the next reaction, but did exhibit signals for two aldehydes the 1H
spectra. The only significant changes from the parent compound showed the loss of the
isolated 1H spin system from the protons on C-9 and C-10, and the conversion of two
hydroxyl carbons into aldehydes (Guthrie, 1961).
1H NMR (CDCI3,5) : 9.96 (CHO, s); 9.42 (CHO, s); 5.36 (H-20b, s); 5.02 (H-20a,
s); 4.58 (H-5a, m); 4.40 (H-13a, br t); 2.60 (H-3p, d); 1.30 (Me, s); 1.25 (Me, s); 1.02
(Me, s); 0.88 (Me, s).
Analysis calculated for C30H42OH: C, 62.27; H, 7.32. Found: C, 62.10; H, 7.44.
Formation of Osazone [3-141 from [3-131 with 2,4-DNPH
Dialdehyde [3-13] was dissolved in methanol, (2 ml) and heated with a solution of
2,4-dinitrophenyl hydrazine (0.1 g) in 2N hydrochloric acid (2ml) and methanol (2 ml).
After 2 h at room temperature, the mixture was extracted with chloroform and the
concentrated extract chromatographed on a silica column. The major band [3-14] was
obtained as an orange yellow crystalline solid, m.p. 215-218 C. Both 1H- and 13C NMR
spectra indicated the retention of all twenty of the carbons from the abeo-taxane
skeleton.


LIST OF REFERENCES
Appendino, G.; Barboni, L.; Gariboldi, P.; Bombardelli, E.; Gabetta, B.; Viterbo, D.
J. Med. Soc., Chem. Commun. 20:1587-1592 (1993).
Appendino, G.; Garibaldi, P.; Pisetta, V.; Bombardelli, E.; & Gabetta, B. Phytochem.
31:4253-4257 (1992).
Appendino, G.; Tagliapietra, S.; Ozen, H. C.; Gariboldi, P.; Gabetta, B.; Bombardelli, E.
J. Nat Prod. 56:514-520 (1993).
Baker, W.; Finch, A. C. M.; Oilis, W. D.; & Robinson, K. W. J. Chem. Soc. 94:1477-
1490 (1963).
Balza, F.; Tachibana, S.; Barrios, H.; & Towers, G. FI. N. Phytochem. 30:1613-1614
(1991).
Barboni, L; Gariboldi, P.; Torregiani, E.; Appendino, G.; Gabetta, B.; Zini, G.;
Bombardelli, E. Phytochem. 33:145-149 (1993).
Baxter, J. N.; Lythgoe, B.; Scales, B.; Scrowston, R. M.; & Trippett, S. J. Chem. Soc.
93:2964-2971 (1962).
Baxter, J. N.; Lythgoe, B.; Scales, B.; Trippett, S.; & Blount, B. K. Proc. Chem. Soc.
90:9-10 (1958).
Boettner, F. E.; Williams, T. M.; Boyd, R.; & Halpern, B. D. Preparation Report 9 (1979).
Polysciences, Inc., Paul Valley Industrial Park, Warrington, PA 18976.
Castor, T. P.; & Tyler, T. A. J. Liq. Chrom. 16:723-731 (1993).
Chauviere, G.; Gunard, D.; Picot, F.; Senilh, V.; & Potier, P. C. R. Acad. Sci. Paris,
Serie II, 293:501-503 (1981).
Chmurney, G. N.; Paukstelis, J. V.; Alvarado, B.; McGuire, M. T.; Snader, V.; Muschik,
G. M.; & Hilton, B. D. Phytochem. 34:477-483 (1993).
Chu, A.; Zajicek, J.; Davin, L. B.; Lewis, N. G.; & Croteau, R. B. Phytochem. 31:4249-
4252 (1992).
Danishefsky, S. J.; Masters, J. J.; Young, W. B.; Link, J. T.; Snyder, L. B.; Magee, T. V.;
Jung, D. K.; Isaacs, R. C. A.; Bornmann, W. G.; Alaimo, C. A.; Coburn, C. A.; &
Di Grandi, M. J. J. Am. Chem. Soc. 118:2843-2859 (1996). For the quote see
the first paragraph of Conclusions, p.2853.
110


17
clinical trials. A few studies on other species of Taxus have also been published, which
are referred to in Chapter 1.
Semi-synthesis of Taxol
The relative ease in ester formation of the three hydroxyl groups in 10-deacetyl
baccatin III (10-DAB, [2-7]) are 7>1013. Esterification of the C-13 hydroxyl is very
challenging due to the inverted cup-like folding of the taxane skeleton and strong
hydrogen bonding with the carbonyl oxygen on the C-4 acetate. Before the side chain
can be attached at C-13, the 7-hydroxyl must first be protected, often accomplished by
attachment of a triethylsilyl group to give [2-11]. Next, this compound is acetylated at the
10-position, to form 7-triethylsilyl baccatin III [2-12],
In one method, [2-12] was esterified with cinnamic acid to give [2-13], which was
then converted to the phenyl isoserine ester by the Sharpless hydroxyamination
procedure (Sharpless et al. 1991) using osmium tetroxide and t-butyl-N-chloro-N-sodio-
carbamate (Mangatal et al. 1989). The four isomers were separated and after
deprotection of the hydroxycarbamates, N-benzoylation and deprotection of the 7-
hydroxyl, taxol could be obtained.
During the investigations of Greene and Potier (Denis et al. 1988; Kanazawa et
al. 1994) dozens of side chain analogues were synthesized and tested, resulting in the
discovery of the taxol analogue known as taxotere [2-9]. Taxotere (docetaxel) was
found to be more active than taxol in the tubulin assay and animal tumor systems and
has also been approved as an antitumor agent. In an alternative synthesis, the 7-
protected baccatin III [2-12] was esterified using either the chiral p-lactam [2-14] or the
oxazinone [2-15] derivative to yield taxol. This method or some variation is currently
used for the semi-synthesis of taxol and taxotere commercially from 10-deacetyl
baccatin III (Ojima et al. 1991, 1992).


54
placed on a silica column in dichloromethane. Elution with 2% methanol in
dichloromethane gave the major band which yielded [3-8] as a white powder, final yield,
0.12 g.
Analysis calculated for C31H42011: C, 63.04; H, 7.17. Found: C, 62.88; H, 7.25.
Saponification of Brevifoliol [3-111
A solution of brevifoliol (1 g) in methanol (20 ml) was stirred with 1N potassium
hydroxide (10 ml) for 1 h. TLC showed that the starting material was absent and very
slow moving, non UV-absorbing component produced. The reaction mixture was
passed through a small column or Amberlite-IR120 ( a sulfonic acid resin) in the H+
form. The column was washed with 1:1 methanol/water. The effluent and washes were
concentrated to dryness and the solid crystallized from acetone to give [3-11] as a
colorless crystalline solid, yield, 0.45 g; m.p. 290C dec.
Analysis calculated for C20H32O6+H2O: C, 62.15; H, 8.87. Found: C, 62.48; H,
8.99.
Debenzoyl Brevifoliol-Pentaacetate [3-12]
Compound [3-11] (0.2 g) was acetylated using acetic anhydride (2 ml) and
pyridine (0.5 ml) by heating at 80 0 C for 30 min. Water was added to decompose the
reagent and the solid filtered. It was crystallized from ether/ ligroin, yield, 0.2 g; m.p.
184-187 C.
1H NMR (CDCI3, §): 6.36 (H-10, d, J=10.2); 5.88 (H-9, br d, 10.2); 5.54 (H-13(3, t,
J=7.2); 5.53 (H-7cx, q, J=5.4, 10.2); 5.36 (H-5p, br s); 5.26 (H-20b, s); 4.87 (H-20a, s);
2.65 (OH-15, s); 2.64 (H-3a, d, J=9); 2.48 (H-14p, dd, J=13.8, 7.2); 2.36 (H-2a, dd,
J=14.1, 9.3); 2.07 (AcMe, s); 2.06 (AcMe, s); 2.02 (AcMe, s); 2.00 (AcMe, s); 1.97
(AcMe, s); 1.95 (18-Me, s); 2.0 (H-6a, m); 1.85 (H-6p, cm); 1.42 (H-2p, d, J=14.4); 1.31
(H-17 Me);1.22 (H-14a, dd, J=13.8, 7.2); 1.13 (H-16 Me); 0.88 (H-19 Me).


31
Brevitaxane A (Brevifoliol) [3-11
The major constituent, which was obtained in a yield of 0.2-0.25%, was named
brevitaxane A because the physical and spectral data indicated that it was a new taxane
compound (later renamed by others as brevifoliol, which will be used throughout this
dissertation). Elemental and FAB-MS analysis (MH+ 557) agreed with the molecular
formula of C31H40O9 (Balza et al. 1991).
An examination of the 1H NMR spectrum showed the presence of two acetyl
groups (signals at 5 1.76 and 5 2.07); and a benzoate group {5 7.88 (d); 5 7.43 (t) and 8
7.56 (t)}. The spectrum also gave evidence for the presence of a (4/20) exocyclic double
bond (two characteristic broad singlets at 8 4.82 (H-20A) and 8 5.20 (H-20B) and signals
at 8 112.1 (C-20) and 8 149.0 (C-4) in the 13C NMR spectrum.
Very little information on the various types of taxane structures that are known
now was available at that point in time (1991) and even less on their diagnostic spectral
characteristics. Based on analogous taxanes and the evidence outlined above it was
postulated that this major constituent had the relatively common 4/20,11-taxadiene type
skeleton. The presence of an exocyclic 4/20 double bond and absence of an oxetane
ring supported our initial assumptions. The next step was to determine the positions of
the various substituents in the molecule in order to elucidate the complete structure.
Most of the structural elucidations of taxanes at the time were based on
degradative studies. It was decided to follow this lead in establishing the presence of
the various functionalities as well as their location in brevifoliol, by actual reactions
and/or derivatizations, supplemented by spectral methods.


This work is dedicated to the memory of Dr. Koppaka V. Rao, an extraordinary
scientist, dedicated teacher and very dear friend. I feel blessed to have known and
worked with him.


63
6. In the 1D nOe-difference spectrum, the interaction between the C-14a,p and C-13 p
protons was the strongest. Crowding of the region around 52.8 ppm made the
spectrum more difficult to interpret, and not very informative.
Table 4-1 : NMR Spectra of Compound [4-3] from BF3 Reaction
Position
Proton
Carbon
APT
DEPT
1

145.9
"T"
C
2 a
1.96 m
28.0
~T~
ch2
2 P
3.00 m



3 a
2.70 m
39.4
~T
CH
4
150.4
~T~
C
5 a
4.42 brs
72.6
nr
CH
6 a
1.76 m
34.1
ch2
6 P
2.05 m


7a
5.48 brd
70.6
nr
CH
8

46.6
C
9 P
5.34 d 6.0
73.2
n-
CH
10a
6.30 d 6.0
67.5
CH
11

146.4
C
12

134.2
n-
C
13 p
5.83 br s
124.0
CH
14 a, p
2.85-3.0 cm
44.6
ch2
18
2.06 s
11.4
CO
X
0
19
1.23 s
13.4
nr
ch3
20 A
4.98 br s
112.2
ch2
20 B
5.20 br s


CO-C6H5

165.2
nr
C
Bz-ipso

130.2
T
C
Bz-ortho
8.01, d 7.5
128.4
CH
Bz-meta
7.44, t 7.5
129.6
nr
CH
Bz-para
7.56, t 7.5
132.9
nr
CH
COCH3
1.98 s
20.6
ch3
2.00 s
21.1
i
ch3
COCH3

169.9
T
C
170.4
T
C
1H NMR were recorded at 600 MHz and 13C NMR at 150 MHz in CDCl3on
a Varian Unity 600 spectrometer at ambient temperature. Chemical shifts 5 (ppm)
are reported relative to TMS as an internal standard.


LIST OF FIGURES
Figure page
2-1 : Early Studies on the Constituents of some Taxus Species 13
2-2 : Taxol and some Synthetic Targets 16
2-3 : Nicolaous Retrosynthetic Strategy 18
2-4 : Nicolaous Taxane Ring Synthesis 21
2-5 : Nicolaous Final Synthetic Intermediates 22
2-6 : Starting Points of Other Synthetic Strategies 23
3-1 : Proton NMR Spectrum of Brevifoliol 32
3-2 : Brevifoliol and Reaction Products 35
3-3 : Brevifoliol Hexaol Reaction Products 39
4-1 : Oxidation Products 61
4-2 : BF3-etherate Catalyzed Elimination Product 64
4-3 : DEPT Spectra of BF3 Elimination Product [4-3] 64
4-4 : lodine/Silver Acetate Product [4-4] and Acetate 66
4-5 : H,FI-COSY Spectrum of [4-4] 67
4-6 : FIETCOR Spectrum of [4-4] 68
5-1 : Fractionation of the Extract of Taxus x media Hicksii Needles 77
5-2 : FIPLC Trace of Taxanes Coeluting with Taxol 78
5-3 : Progress of Elution of Taxanes from Reverse Phase Column 78
6-1 : Taxanes and Analogues from Taxus x media Hicksii 92
viii


18
Total Synthesis
Swindell (1992) published a review on the progress of more than thirty groups
and reported only modest success in the total synthesis of taxol. Only two years later
two separate groups headed by K. C. Nicolaou (Nicolaou et at. 1994b) at the Scripps
Research Institute and R. A. Holton (Holton et at. 1994b) at Florida State University
would announce almost simultaneously two total syntheses of taxol.
Nicolaou and colleagues designed the strategy for their synthesis based on the
one bond disconnection analysis seen in Figure 2-3. After preparation of the fully
functionalized A ring [2-16] and C ring [2-17] equivalents, a convergent and flexible
Esterification
McMurry coupling
_ Oxetane
q formation
Shapiro reaction
OBn
A
[2-16] Aryl sulphonylhydrazone [2-17] Aldehyde
Figure 2-3 : Nicolaous Retrosynthetic Strategy
synthesis of taxol involving 28 more steps allowed the preparation of numerous
analogues. While not practical for the commercial production of taxol, synthetic methods


5
than analytical HPLC studies on the taxol content of the needles under various
conditions, no practical methodology for the isolation of taxol or other taxolds has been
published. Publications from this laboratory which address these issues are essentially
the only work available in the literature (Rao et al. 1995; Rao et al. 1996). In addition to
direct isolation of taxol, the needles were also examined for the presence of analogues
such as 10-deacetyl baccatin III, since semi-synthesis from such is already an important
alternative. The two most important species, T. baccata L. and T. wallichiana Zucc.,
have become the focus of attention since they were demonstrated to contain the highest
concentrations of 10-deacetyl baccatin III.
Growing of various tissues of T. brevifolia in plant cell culture has been under
development since 1990 and the methods have been standardized in many laboratories.
However, the yields, as yet, have not been very attractive. Further research is expected
to overcome this problem. Work on this alternative will continue due to the
attractiveness of this approach and its potential for large-scale operations.
Starting with 10-deacetyl baccatin III, considerable progress was made in the
area of semi-synthesis. In the first recorded semi-synthesis of taxol, the 13-cinnamate
ester of 7-protected baccatin III was converted to the phenyl isoserine ester through a
Sharpless hydroxy-amination (Denis et al. 1988). At this point, as an alternative to
benzoylation of the amino group that will yield taxol, a t-BOC group (tert-
butoxycarbonyl) was introduced, along with leaving the 10-hydroxyl free, to obtain an
analogue known as taxotere. On the basis of its activity, taxotere has also been
approved as an antitumor drug. Two important schemes for preparing taxol from 10-
deacetyl baccatin III have been well developed and used for the large scale semi
synthesis of taxol as discussed in Chapter 2 (Denis et al. 1994; Ojima et al. 1991; Holton
et al. 1992).


41
aromatic signals between 8 128.7 and 5 133.2 (see tables 3-1 and 3-2); and for the two
acetates, by the methyl signals at 8 20.7 and 8 21.4. Analysis of the 1H NMR and 1H
COSY and 1H,13C Heteronuclear Correlation (HETCOR) experiments also gave
additional support for the presence of the acetates with signals at 5 1.76 s and 8 2.07 s,
as well as benzoate signals at 8 7.88 d (ortho); 8 7.43 t (meta); and 8 7.56 t (para). Next,
evidence for the presence of the normally present (11/12) taxane double bond could be
seen in the carbon spectrum by the signals at 5 133.9 (C-11) and 8 151.5 (C-12);.
Similarly, the existence of a (4/20) exocyclic double bond could be seen by the signal 5
149.0 (C-4) and 8 112.1 for (C-20). In the 1H NMR spectrum the exocyclic 4/20 double
bond is also indicated by the two characteristic broad singlets seen at 8 4.82 (H-20A)
and 8 5.20 (H-20B).
In the 1H COSY experiment weak but definite interactions between the singlet at
8 4.82 (C-2Qa) with both the H-3(3 doublet at 8 2.78 (9 Hz.) and the H-2a doublet of
doublets at 8 2.36 (9, 13 Hz.) supported the assignments given for the methylene
protons. (The designations for the C-20 protons are A and B, since a and p do not have
the conventional meaning system and could be confusing). Along with the interaction
between H-2a and H-2p the first isolated spin system in the 1H spectrum was
established and the relative geometry of the protons.
The region between 8 62.4 and 8 77.1 in the 13C spectrum carbons with hydroxyl
or ester oxygen attached to oxygens, and these signals could be further defined in the
DEPT experiment (Distortionless Enhancement with Polarization Transfer, NMR) as
primary, secondary, tertiary and quaternary carbons. The spectrum showed two
quaternary carbon signals and five oxymethine carbon signals. Since the presence of
only six signals was expected based on the proposed formula, the quaternary signal at
8 62.4 was Intriguing even from the start of the spectral examination. In taxol with its C-9


9
even twice. In the present work, practical isolation methods were used to obtain gram
quantities of many compounds, as crystalline solids, where possible.
The compounds are usually characterized by physical and spectral properties,
providing elemental analyses, and carrying out derivatization such as acetylation,
oxidation, etc. The structures were elucidated through chemical reactions as well as
through spectral data. Thus, even though some of the final structures may have been
published, the work described here contains experiments that have not been carried out
by these authors.
Brief descriptions of the topics that appear in this dissertation are given below.
Chapter 2 gives a brief and selected summary of the pertinent literature on taxol
and taxoids, covering the areas of isolation, elucidation of structures, semi-syntheses
and total syntheses. Because the subject matter expanded enormously since 1993, the
scope of the review is limited to material that is relevant to the subject matter of the
dissertation.
Chapter 3 deals with the taxoid composition of the needles of Taxus brevifolia. It
covers the application of the reverse phase column chromatography to the needle
extract, isolation of the major taxoid, brevitaxane A (or brevifoliol), along with brevitaxane
B, and taxol. It continues with the elucidation of the structure of brevitaxane A by
various reactions, as well as by a detailed analysis of the NMR spectral evidence.
Chapter 4 discusses some unusual reactions of brevifoliol. Such reactions have
not been reported with this or any other taxoid compounds. In each case, crystalline
compounds were obtained and characterized by physical and spectral data.
Chapter 5 deals with fractionation of the extract of the needles of Taxus x media
cv. Hicksii by reverse phase column chromatography and isolation of taxol and several
other taxoids and their characterization. In spite of the fact that this species (ornamental
yew) was declared as the preferred plant for the future isolation of taxol, no publications


73
At this point, the previous major compound moved further to give a new product.
The mixture was filtered, the filtrate washed with aqueous sodium bisulfite, followed by
water and concentrated to dryness. The product was chromatographed on a silica
column in 4:1 chloroform/ ligroin. The major component was obtained as a colorless
crystalline solid, yield, 0.1 g. The compound was found to be identical with the product
obtained from the reaction of brevifoliol with iodine and silver acetate.
Reaction of [4-41 with N-Bromosuccinimide [4-71
A solution of [4-4] (0.04 g) in benzene (5 ml) was stirred with N-
bromosuccinimide (25 mg) at room temperature. After 2 h, TLC showed formation of a
slightly faster moving compound, which was separable from the starting material only
after 2 or 3 developments of the TLC plate. The reaction mixture was washed with
aqueous sodium bisulfite, followed by water and concentrated to dryness. The product
was crystallized from ether in ligroin, m.p. 185-188 0 C.


14
Harrison (Harrison & Lythgoe 1966; Harrison et al. 1966) published one of the
earliest biogenetic theories for the formation of taxanes starting with geranylgeranyl
pyrophosphate and electrophilic cyclization [2-5], Efforts by many groups to utilize a
similar scheme to synthesize the taxane skeleton have been unsuccessful thus far
(Kumagai et at. 1981; Hitchcock & Pattenden, 1992). Biogenetic pathways often provide
ideas for simplified approaches in the synthesis of natural products.
Many early studies utilized acidic conditions for the extraction which might have
hampered the isolation of neutral or acid-labile compounds, Kondo & Takahishi (1925)
obtained a non-basic compound from the Japanese yew by using neutral conditions.
The cinnamates can also be directly isolated from the plant, indicating that they occur
naturally and also as artifacts of processing.
The National Cancer Institute (NCI) and the U.S. Department of Agriculture
(USDA) joined forces in 1960 to collect and screen plants for activity in several animal
tumor models. Arthur Barclay of the USDA collection team obtained samples from the
Pacific yew tree (Taxus brevifolia Nutt., family Taxaceae) from Washington State in
1962. In 1964, the extracts from the bark and stems were found to be active against KB
cells in vitro (Wani et al. 1971).
Studies after the Discovery of Taxol
Dr. Monroe Wall had discovered another antitumor agent known as camptothecin
using the activity on KB cells for isolation and was interested in any other extracts
showing this activity. Thus, work on T. brevifolia by Wani and Wall at the Research
Triangle Institute was started and led to the isolation of 500 mg of taxol 2 years later in
1966. Cytotoxic actions (Wani et al. 1971) in KB cells, P388 leukemia, Walker 256
carcinosarcoma and P-1534 leukemia were present in the extracts from the bark. These


89
succession. One of these was identified as baccatin VI [6-2], and the second one was
taxol [5-3],
Taxiflorine
Taxiflorine [6-1] was readily obtained as a colorless crystalline solid. Its
elemental analysis agreed with the molecular formula C35H44013. Its 1H NMR spectrum
in CDCI3 showed broad and rounded peaks with poor resolution. In DMSO-d6, the
spectrum gave sharper signals but showed double the number of peaks in certain
positions. The 13C spectrum also exhibited extra peaks, which suggested that the
compound was a mixture of rotamers in equilibrium. One could infer the presence of
ester functions from the spectra, with four acetates and one benzoate, and an oxetane
ring.
Acetylation of taxiflorine gave a monoacetate [6-3], which gave sharp signals in
its 1H NMR spectrum, with the expected number of peaks, thus showing that it is a single
compound, unlike the starting material. Although the acetate was isomeric with baccatin
VI, it was different. The most striking difference between the two spectra was seen with
the signal for the H-13. In the acetate of taxiflorine, this signal was at 6 5.60, while the
same was found at 8 6.3 in baccatin VI. A comparison with other related taxanes
showed that in those with the 6-membered A-ring, the H-13 signal appears at 8 6.2-6.5,
whereas in taxanes with a 5-membered A-ring, as in the 11(15->1)-abeotaxanes, it
appears at 8 5.4-5.7 (Appendino et al 1993B).
Positions 9 and 10 in taxiflorine carry a free hydroxyl and a benzoate function.
To locate the benzoate, a comparison of the signals due to H-9 and H-10 in taxiflorine
were compared with the corresponding signals in the monoacetate. With the two signals
at 8 6.30 and 8 5.90 in taxiflorine, the latter undergoes a down-field shift from 8 5.90 to


34
Table 3-2 : Carbon NMR Spectra of Brevifoliol and Brevifoliol Acetates
Carbon
Number
[3-1]
Brevifoliol
[3-2]
5-Ac
[3-3]
13-Ac
[3-4]
5,13-Di-Ac
[3-5]
1,5,13-Tri-Ac
1
62.4
63.0
63.4
63.0
63.3
2
29.1
29.2
29.4
29.1
28.3
3
37.9
38.8
37.6
38.8
38.9
4
149.0
145.4
147.4
145.2
145.1
5
72.4
74.1
72.7
74.1
73.9.
6
36.0
33.9
36.1
33.9
34.0
7
70.1
69.7
69.8
69.6
69.7
8
45.0
44.8
45.2
44.8
45.0
9
77.1
77.9
79.8
79.3
78.9
10
70.2
70.7
70.3
69.8
68.4
11
133.9
134.0
136.5
136.4
136.5
12
151.5
151.1
150.5
147.3
148.2
13
76.7
76.9
77.8
76.9
78.0
14
47.3
47.1
44.2
44.1
43.3
15
75.9
75.6
75.6
75.6
87.2
16
26.9
27.0
27.0
27.0
23.1
17
24.8
24.8
25.0
24.8
21.8
18
12.0
11.8
12.1
11.9
11.9
19
12.9
12.9
12.9
12.9
13.5
20
112.0
114.1
111.5
114.3
114.3
CO-C6H5
164.3
164.1
164.2
164.1
165.0
Bz-ipso
129.3
129.2
129.4
129.1
129.9
Bz-ortho
129.4
129.4
129.5
129.5
129.3
Bz-meta
128.7
128.7
128.7
128.8
128.4
Bz-para
133.2
133.3
133.2
133.4
133.0
CO-CH3
20.7
20.8
20.7
20.7
20.8
21.4
21.4
21.4
21.4
21.4
21.2
21.1
21.2
21.3
21.0
21.0
21.7
COCH3
169.9
169.9 (X2)
169.9
169.91
169.9
170.5
170.2
170.8
170.5
170.5
169.7
169.6
169.6
169.9
171.0
169.5
13C NMR spectra were recorded at 150 MHz in CDCI3on a Varian Unity 600
spectrometer at ambient temperature. Chemical shifts 5 (ppm) are reported with TMS as
internal standard.


27
Substitution of the 7 position does not appear to significantly decrease the
activity. Taxanes with a 7p-0-xyloside moiety are comparably active in both assays
when compared to the respective aglycones. Similarly, epimerization at the 7-position
does not eliminate activity.
Hydrolysis of the 10-acyl function does not reduce the cytotoxicity significantly in
cell culture assays. As with other structural features, this point is being explored in the
more recent clinical trials in Europe with taxotere.
The importance of the oxetane ring for activity has been investigated through ring
opening via different Lewis acids including Meerwein's reagent (triethyloxonium
tetrafluoroborate); acetyl chloride, mesyl chloride and others. The product obtained form
the Meerwein's reagent had a primary alcohol at C-20 and secondary C-5-hydroxyl, but
no other changes compared to taxol. The activity normally seen with taxol in both
assays was lost with the opening of the oxetane ring. This suggests that the oxetane
ring is necessary for activity but leaves open questions regarding the effect of ring
contractions in ring A.
The properties of the C-13 hydroxyl mentioned above make attachment of a side
chain quite difficult. Protection of other free hydroxyls in both the side chain and taxane
skeleton are necessary, followed by selective deprotection after the side chain has been
attached. Taxotere and taxol have both been synthesized from this taxane and this is
currently the starting material for the production of both drugs.
Epimerization of the 7 hydroxyl from p to a via a retro-aldol mechanism allows
formation of an energetically favorable hydrogen bond with the 4-acetate carbonyl
oxygen. This epimerization is a concern in both taxane isolation and synthetic methods,
and necessitates the avoidance of acidic or basic conditions. Protection of this C-7 p-


50
Brevifoliol-13-Ketone (3-61
A solution of brevifoliol (0.2 g) in benzene was treated with Mn02 (manganese
dioxide, 1 g, Fisher Scientific) and the mixture heated under reflux for 2 hours, at which
time, the starting material was consumed and a slightly faster moving product was
formed. The mixture was filtered, concentrated and applied to a small silica column (15
g) in dichloromethane. Elution with 1% acetone in dichloromethane gave the major
product which was recovered by concentration as a colorless powder, yield, 0.12g. The
1H- and 13C MMR spectra of this faster moving product were quite poorly resolved and
only gave usable results at temperatures below -10 0 C. Recrystallization and further
chromatography failed to improve this situation, and low temperature NMR experiments
indicated that a rotameric equilibrium was responsible for the poor resolution seen in
these spectra.
1H NMR (CDCI3i 600 MHz, -40 C, 5) major rotamer: 0.92, s (H-19); 0.98, s (H-
16); 1.35, s (H-17); 1.46 (d, J=14.1 Hz, H-2a); 1.78 (s, 9-0 acetate methyl); 1.90 (m, H-
6a) 2.0 (m, H-6(3); 2.01 (s, 18 methyl); 2.02 (s, 7-0 acetate methyl); 2.32 (d, J=19.0,
14a); 2.48 (d, J=19.0 Hz, H-14P); 2.92 (unresolved, H-3a); 2.74, br s (C-15 OH,
exchangeable with D20); 4.92, s (H-20 A); 5.28, s (H-20 B); 5.39 (br s, J= H-5P); 5.54 (t,
J=7.2 Hz, H-13P); 5.61 (dd, J=4.8, 11.4 Hz, H-7a); 6.09 (poorly resolved br d, J=10.5 Hz,
H-9a); 6.65 (d, J=10.5 Hz, H-10p); 7.43 (t, J=7.8 Hz, H-Ph-meta); 7.56 (t, J=7.8 Hz, H-
Ph-para); 7.87 (d, J=7.8 Hz, H-Ph-ortho).
13C NMR (CDCIg, 600 MHz, 5): 8.9 (C-18 methyl, q); 12.5 (C-19 methyl, q); 20.7
(7-0 acetate methyl, q); 21.0 (9-0 acetate methyl, q); 26.2 (C-17 methyl, q); 26.6 (C-16
methyl, q); 27.3 (C-2, t); 27.7 (C-6, t); 34.2 (C-3, d); 43.2 (C-8, s); 48.3 (C-14, t); 58.1 (C-
1, s); 70.8 (C-7, d); 71.0 (C-10, d); 73.2 (C-9, d); 75.6 (C-15, s); 111.4 (C-20, t); 128.8
(Ph-meta, d); 129.1 (Ph-ipso, s); 129.5(Ph-ortho, d); 133.4 (Ph-para, d); 136.4 (C-11, s);


70
Action of Boron Trifluoride on Brevifoliol 4-31
Brevifoliol (0.3 g) was dissolved in acetone (10 ml) and to the solution was added
1 ml of 1% boron trifluoride etherate in acetone to make a 0.1% overall concentration of
boron trifluoride in the reaction mixture. After 3 h, water was added, the solid filtered
and after drying, subjected to chromatography on silica gel in chloroform/ ligroin (1:1).
The major band obtained with the same solvent was crystallized from ether/ligroin, yield,
0.1 g, m.p. 162-165 0 C.
1H NMR (CDCI3, Varian Unity 600 MHz, 8): 1.16, s (H-19); 1.03, s (H-16); 2.08
(cm, H-14a); 1.27, s (H-17); 2.32 (d, J=13.8 Hz, H-2a); 1.92 (s, methyl, 9-acetate); 1.84
(m, H-6a ); 1.98 (m, H-6p); 2.26 (s, H-18); 2.05 (s, 7-acetate methyl); 3.30 (d, J=13.8 Hz,
H-2P); 2.28 (cm, H-14P); 2.95, br s (C-15 OH, exchangeable with D20); 4.38 (t, J=7.2
Hz, H-13P); 4.28 (br s, H-5P); 4.54 (d, J=13.2 Hz, H-20 B); 4.54 (d, J=13.2 Hz, H-20 B);
5.51 (dd, J=4.8, 11.4 Hz, H-7a); 6.07 (d, J=10.1 Hz, H-9a); 6.54 (d, J=10.1 Hz, H-10p);
7.45 (t, J=7.5 Hz, H-Bz-mefa); 7.57 (t, J=7.5 Hz, H-Bz-para); 7.89 (d, J=7.5 Hz, H-Bz-
ortho).
13C NMR (CDCI3, Varian VXR 300 MHz, 8): 11.4 (C-18 methyl, q); 13.4 (C-19
methyl, q); 20.6 (7-0 acetate methyl, q); 21.1 (9-0 acetate methyl, q); 28.0 (C-2, t); 34.1
(C-6, t); 39.4 (C-3, d); 44.6 (C-14, t); 46.6 (C-8, s); 67.5 (C-10, d); 70.6 (C-7, d); 72.6 (C-
5, d); 73.2 (C-9, d); 112.2 (C-20, t); 124.0 (C-13, d); 128.4 (C-Bz-ortho, d); 129.6 (C-Bz-
meta, d); 130.2 (C-Bz-ipso, s); 132.9 (C-Bz-para, d); 134.2 (C-12, s); 145.9 (C-1, s);
146.4 (C-11, s); 150.4 (C-4, s); 165.2 (CO-Ph, s); 169.9 (CO-Acetate, s); 170.4 (CO-
Acetate, s).
Reaction with Iodine and Silver Acetate [4-41
To a solution of brevifoliol (0.5 g) in benzene (15 ml) were added iodine (0.7 g)
and silver acetate (0.75 g) and the mixture stirred at room temperature for 20 h. TLC


75
of two components. On the basis of NMR spectral evidence, structures were assigned
to these two components (Castor & Tyler, 1993).
Due to the presence of pigments, waxes and other impurities, the isolation of
taxol and other taxanes from the needles was expected to be more difficult when
compared to their isolation from the bark of T. brevifolia. A project was started in this
laboratory to meet the need for a practical method for the isolation of taxol and other
related taxanes from the bark and needles of various Taxus species in spite of these
challenges. The application of a preparative scale reverse phase column
chromatography technique proved to be surprisingly successful in the processing of the
extracts of T. brevifolia.
To begin with, the HPLC analysis of the extract of the needles of the ornamental
yew, as shown in Figure 5-2, clearly shows that taxol is accompanied by several major
taxane components, which are present in much higher concentrations than taxol itself.
In view of such relatively high concentrations of these components, it is surprising that
only such minute amounts of two of these mixtures could be isolated earlier, as indicated
above. Also, no other characterizing data were provided other than the spectral data.
This laboratorys objective was the development of a simpler procedure for the isolation
of taxol with potential for large-scale use, in addition to more fully characterizing the
major taxanes present in the extract. The needles of the ornamental yew (200 lbs.,
dried) were received through the courtesy of Hauser Company, Boulder, CO, during
May-June 1993.
The extraction was carried out three or four times using methanol and the extract
concentrated to a syrup. The resulting concentrate was then partitioned between water
and chloroform, and the organic layer containing the taxane fractions was concentrated
to a thick semi-solid mass, which was used directly in the next step.


46
After the sample addition was completed, fresh 25% acetonitrile/ water was
passed through, followed by the step gradient of acetonitrile/ water (30, 35, 40, 45, 50
and 60%) was used. Fractions (200 ml) were collected and monitored by UV
absorbance (at 275 nm), TLC and analytical HPLC. The change to the next
concentration of solvent was determined by the results of monitoring the fractions. For
example, when the absorbance values rose as a result of the previous change, the
solvent was continued until a definite trend to lower values was seen. Similarly, when
the TLC showed the trend towards decreasing intensity of the major spot, and no new
spot had shown a tendency to increase, the solvent was changed to the next level. In
general, 2-3 multiples of the hold-up volumes of the column were used.
After the elution with the 60% solvent was completed, the column was washed
with 100% methanol, followed by a mixture of methanol/ ethyl acetate/ ligroin which
stripped the column of the chlorophylls, waxes and other lipid soluble components. After
this solvent, washing with methanol and equilibration with 25% acetonitrile/ water made
the column ready for another run.
After the monitoring, fractions with low UV-absorbance values were combined
and concentrated into groups, based on the TLC data. Those fractions with relatively
stronger UV readings were let stand at room temperature for 3-5 days, whereby crystals
appeared in several sections of the fraction sequence. These crude crystals were
filtered, dried and purified further either by recrystallization or using a small silica column
(normal phase).
Brevifoliol 3-11
The fractions containing this component gave crystals but only a small portion
was obtained in this form. Hence, after filtration of the crude crystals, the filtrate was
concentrated to dryness and the solid taken up in dichloromethane and passed through


Brevifoliol-5,13-Diacetate [3-4] 49
Brevifoliol-13-Ketone [3-6] 50
Dihydrobrevifoliol [3-7] 51
Brevifoliol Epoxide [3-8] 51
Ozonization of Brevifoliol: Brevifoliol-norketone [3-9] 52
Brevifoliol-4,20-Diol [3-10] 53
Saponification of Brevifoliol [3-11] 54
Debenzoyl Brevifoiiol-Pentaacetate [3-12] 54
Periodate Oxidation of [3-11] to [3-13] 55
Formation of Osazone [3-14] from [3-13] with 2,4-DNPH 55
Debenzoyl Brevifoliol [3-15] 56
4 SOME UNUSUAL REACTIONS OF BREVIFOLIOL 57
1. Acid-Catalyzed Acetylation 58
2. Oxidation 59
3. Action of BF3 on Brevifoliol [4-3] 61
4. Reaction with lodine/Silver Acetate [4-4] 65
Experimental 69
Brevifoliol Triacetate [3-5] 69
Oxidation with Jones Reagent to [4-1] 69
Action of Boron Trifluoride on Brevifoliol [4-3] 70
Reaction with Iodine and Silver Acetate [4-4] 70
Acetylation of [4-4] to [4-5] 72
Reaction with N-Bromosuccinimide and Silver Acetate [4-6] 72
Reaction of [4-4] with N-Bromosuccinimide [4-7] 73
5 TAXANE CONSTITUENTS OF THE NEEDLES OF TAXUS X MEDIA 74
Brevifoliol [3-1] 79
Taxanes I [5-1] and II [5-2] 79
Taxane III [2-1] 80
Taxane IV [2-2] 80
Taxol [5-3] 80
Ozonolysis of [2-2] 80
Experimental 81
Extraction: 81
Chromatography: 81
Characterization of the Taxane Components of Taxus x Media Hicksii 82
Brevifoliol [2-1] 82
Taxanes I and II [5-1] and [5-2] 83
Taxane III [2-1] 84
Taxane IV [2-2] 84
Taxol [5-3] 85
Ozonolysis of Compound [2-2] 86
6 TAXANE CONSTITUENTS OF TAXUS FLORIDANA 87
Taxiflorine 89
Experimental 93
Extraction 93
v


66
displace the iodine located at C-4 to produce the 4-acetoxy-20-iodo compound was also
attractive (Woodard & Brutcher, 1958).
When brevifoliol was stirred with iodine and silver acetate, the course of the
reaction was clearly different. No multiple products or decomposition to dark colored
products was seen, even when the reaction was continued for 15-30 hours, unlike the
reaction with iodine alone that became dark in 1-2 hours. The reaction was continued
until the starting material was consumed and a major, faster moving compound was
produced. Chromatography on a silica column gave a colorless crystalline solid.
Acetylation with acetic anhydride and pyridine at 70 C for 15 minutes produced
the 5-O-monoacetate [4-5], confirming the presence of the 5 hydroxyl with NMR
analysis. Treatment of [4-4] with n-Bromosuccinimide produced the 5-0 ketone, further
confirming the absence of a free C-13 hydroxyl on the basis of NMR and UV spectral
data.
[4-4] lodine/Silver Acetate FToduct
[4-5] Acetate
Figure 4-4 : lodine/Silver Acetate Product [4-4] and Acetate


103
columns were packed with C-18 or C-8 bonded silica gel (Spherisorb, 15-35 p diameter,
Phase Separations Inc., Norwalk, CT) as a slurry in methanol After thorough washing
with methanol, the columns were equilibrated with 25% acetonitrile in water.
Thin-layer chromatography was carried out using silica gel HF-60, 254+366 (EM
Science/Fisher). Visualization was by a UV-lamp and by charring with 1 N H2S04.
Column chromatography was performed using silica gel (Fisher, 100-200 and 235-425
mesh) or Florisil (Fisher F-101, 100 mesh) with a solvent sequence consisting of
ligroin/CHCI3i CHCI3, 2-5% acetone and finally, 2-10% MeOPI in CHCI3.
Melting points were determined on a Fisher-Johns hot stage apparatus and are
uncorrected. The following instrumentation was used for the spectra recorded here:
UV, Perkin Elmer A.3B; IR, Perkin Elmer PE-1420; and NMR, General Electric QE-300,
Nicolet NY-300, Varan VXR-300, Varian Gemini-300 and Vahan Unity-600
spectrometers. Mass spectra (FAB) were obtained on a Finnegan Mat 95Q
spectrometer using a cesium gun operated at 15 KeV of energy.
Flavonoids
Quercetin Rutoside (Rutin)
3-[[6-0-(6-deoxy-a-L-mannopyranosyl)-p-D-glucopyranosyl]oxy]-2-(3,4-
dihydroxyphenyl)-5,7-dihydroxy-4H-1-benzopyran-4-one. m.p. : 186-189 C (dec., turns
brown at~127C); [a]D23 +14.06 (ethanol, c 1.02); [a]D23 -39.76 (pyridine, c 1.06).
1H NMR (DMSO-d0, 300 MFIz, 8): 12.6 (C-5 hydroxyl, br s, D20 exchangeable);
7.57 (H-21, d, 2.2); 7.53 (H-6, dd, 8.6, 2.2); 6.86 (H-5, d, 8.3); 6.40 (H-8, br s); 6.21 (H-6,
br s); 5.34 (H-1 glucosyl, br s); 4.42 (H-7 rhamnosyl, br s); 3.1-3.75 (CH-OPI glycosyl,
6H, mult.); 1.03 (H-12 methyl, d, 6.0).


78
Figure 5-2 : HPLC Trace of Taxanes Coeluting with Taxol.
Column Elution, Absorbance vs. Time
E
c
10
h-
CM
o
o
c
ro
-Q
s
G
if)
SI
<
Figure 5-3 : Progress of Elution of Taxanes from Reverse Phase Column.


23
Holton group Patchouline Oxide Fragmentation
[2-28] Wieland-Miescher ketone
Danishefsky group
[2-29] a-Pinene
Wender group
Figure 2-6 : Starting Points of Other Synthetic Strategies


93
From the present studies, which compared the taxane composition of both
species, it became clear that T. floridana would have been a much better choice, for the
reasons given below:
1. Taxol is isolated more easily from the Florida yew than from the ornamental yew,
because, in the former there are no co-eluting taxanes to interfere and which have to
be removed as an additional step. The yields of taxol (0.01% from fresh leaves) are
accordingly better with the Florida yew, than from that of the ornamental yew
(0.015% from dried leaves).
2. Besides taxol, the Florida yew gives relatively high yields (0.05-0.06% from fresh
leaves) of 10-deacetyl baccatin III, which is the most commonly used intermediate for
the semi-synthesis of taxol. This compound is present in the ornamental yew in
exceedingly low concentrations.
3. The other components that contain the oxetane ring, such as baccatin VI and
taxiflorine, are also useful compounds for the synthesis of taxol analogues. In
contrast, the taxanes from the ornamental yew that are 11,4/20 diene taxanes have
no currently documented use.
The work described here is the first definitive investigation of the taxane
constituents of Taxus floridana on a preparative scale giving the actual recovered yields
of the crystalline compounds.
Experimental
Extraction
The needles and small twigs of Taxus floridana were collected from several
bushes growing at different locations of the campus of the University of Florida.
Likewise, they were made available from the plants growing under hydroponic


40
1991. Balza et al. (1991) published the isolation of a new compound at about that same
time from the needles of T. brevifolia, which they named brevifoliol, and an assignment
of its structure as shown in [3-16], The compound appeared to be similar to, if not the
same as, brevitaxane A, that was isolated from the needles at the University of Florida.
The structure proposed by Balza et at differed from that of brevitaxane A, with the
benzoate group being placed at C-7 instead of at C-10.
That same year (1992); the isolation of taxchinin A was described (Fuji et al.
1992); which was later shown to be 2-acetoxy-brevifoliol. Fuji correctly assigned the
structure with a 5-membered ring A, on the basis of NMR spectral data. The authors who
isolated brevifoliol and assigned structure with the 7-benzoate (Chu et al. 1993)
published a revised structure for brevifoliol, in which the benzoate was moved to C-10,
from C-7, but with the skeleton of a conventional taxane.
During 1993, two other publications appeared, one from Georg et al. (1993).
and the other from Appendino et al. (1993) reexamining the NMR spectral data of
brevifoliol, and arriving at the structure in which the A-ring was 5-membered. Later that
year, Chu et al. (1993); on the basis of x-ray crystallographic data, revised the structure
of brevifoliol again to the presently accepted structure.
Due to the intense competition in taxol research, we began a detailed
examination and analysis of the NMR spectral data using the 13C NMR, NOESY,
HETCOR and other spectral methods to determine if the rearranged (5/7/6) skeleton
might be distinguishable from the spectrum of a taxane with a conventional (6/8/6)
skeleton. The following is an analysis of the spectral data of brevifoliol.
The carbonyl signal in the 13C NMR spectrum at 6 164.3 indicated the presence
of one benzoate, and signals at 8 169.9, 5 170.5, likewise, indicated that 2 acetate ester
groups were present. Further support for the benzoate was obtained by the four


106
over Na2S04, concentrated, then crystallized from ethyl acetate in ligroln, yield 440 mg
(68%); first crop.
M.p.: 165-167 0 C (lit. 171 0 C, Sucrow 1966). [a]D23 -35.0 (pyridine, c 1.3); (lit.
pyridine, c 1.33,-33.7, Swift 1952).
1H NMR (CDCI3, 300 MHz, 5): 5.37 (H-6, br d, J= 4.8); 5.21 (H-3, t, J=9.3); 5.08
(H-4, t, J=9.2); 4.95 (H-2, dd, J=9.3 and 8.1); 4.60 (H-1'f d, J=8.1); 4.26 (H-6b, dd,
J=4.8 and 12); 4.11 (H-6a, dd, J=2.4, 12); 3.70 (H-5, cm); 3.5 (H-3, cm); 2.07, 2.05,
2.02, 2.00 (4 X Ac CH3, s); 1.0 (H3C-19, s); 0.84 s; 0.82 s; 0.68 (H3C-18, s).
13C NMR (CDCI3, 75 MHz, 5): 170.6, 170.3, 169.3, 169.2 (4 X AcC=0, s); 140.3
(C-5, s); 122.1 (C-6, d); 99.6 (C-1\ d); 80.0 (C-3, s); 72.9 (C-3, d); 71.7 (C-5, d); 71.5
(C-2', d); 68.5 (C-4, d); 62.1 (C-6, t); 56.7 (C-14, d); 56.0 (C-17, d); 50.1 (C-9, s); 45.8
(C-24, s); 42.3 (C-13, s); 39.7 (C-12, t); 38.9 (C-4, d); 37.2 (C-1, d); 36.7 (C-10, s); 36.1
(C-20, d); 33.9 (C-22, t); 31.9 (C-8, d); 31.8 (C-7, t); 29.4 (C-2, t); 29.1 (C-25, d); 28.2 (C-
16, t); 26.1 (C-23, t); 24.3 (C-15, d); 23.0 (C-28, t); 21 0(C-11, d); 20.6, 20.5, 20.44,
20.42 (4 X AcMe); 19.7 (C-27, q); 19.3 (C-19, q); 19.0 (C-26, q); 18.7 (C-21, q); 11.8 (C-
18, q); 11.8 (C-29, q).
IR, v max (KBr, cm'1): 1755, 1220, 1045, 910.
B-Sitosterol
M.p.: 138-140 0 C (lit. 137-138 0 C, Swift 1952); [a]D23 -37.2 (chloroform, c 1.1);
(lit. -38.2, chloroform, c 5.1, Swift 1952).
1H NMR (CDCI3, 300 MHz, 5): 5.36 (H-6, d); 3.50 (H-3, m); 2.30 m, 2.0 m, 1.0 s,
0.82 d, 0.7 s.
13C NMR (CDCI3, 75 MHz, 5): 140.7 (C-5, s); 121.6 (C-6, d); 71.8 (C-3, s); 56.7
(C-14, d); 56.0 (C-17, d); 50.1 (C-9, d); 45.8 (C-24, d); 42.3* (C-13, s); 43.2 (C-4, t); 39.7
(C-12, t); 37.2 (C-1, t); 36.4 (C-10, s); 36.1 (C-20, d); 33.9 (C-22, t); 31.9 (C-8, d); 31.6


92
[6-1] Taxiflorine
R-i = H R2 = C6H5CO
[6-3] Taxiflorine Acetate
R-i =CH3CO R2 = C6H5CO
[6-4] Taxiflorine Benzoate
Ri = R2 = C0H5CO
[6-5] Hexa-Acetate
R-i = R2 = CH3CO
Figure 6-1 : Taxanes and Analogues from Taxus x media Hicksii


25
an oxetane. As in the case of the analogous steroids with the two methyl groups as part
of the skeleton, the taxanes have four methyl groups #16, 17, 18 and 19 as part of the
taxane ring system. Some examples of the taxane skeleton found in the various species
of Taxus are shown in Figures 2-1 and 2-2.
Oxygenation of the taxane ring has been observed to varying extents. The
minimum number being 4, distributed at 5, 9, 10 and 13, as seen in taxusin [2-26], In
general, oxygenation may occur at carbons 1, 2, 4, 5, 7, 9, 10 and 13. Instances have
been recorded where oxygenation was present at 14 (in place of 13), as well as part of
the methyl groups at 19 and 17.
Taxa-4(20):11-dienes
This is the most common structural type seen in the taxanes, with a C-4(20) and
a A11 double bond. These taxanes are generally referred to taxa-4(20);11 -dienes. The
alkaloidal Winterstein esters are included in this group as are many of the neutral
taxanes. The oxygen at C-9, if present, is usually seen as a secondary alcohol or as an
ester. The C-13 position in this group, likewise, exists as an alcohol, ester or oxidized to
a carbonyl to form an a,p-unsaturated carbonyl. Esterification at C-13 is usually limited
to an acetyl or a cinnamoyl, but the side chain (N-acyl phenyl isoserine); as found in
taxol, cephalomannine and others has not been reported in this subgroup so far. The 5
position is oxygenated with an a-hydroxyl, which might be free, or esterified by an acetic
acid, cinnamic acid or the Winterstein-acid. Some examples of these compounds with a
cinnamate ester function are described in Chapter 5.
4(20)-Epoxides
This group is relatively less frequent but examples with different substitution
patterns have been isolated. One variation comes from the presence or absence of
hydroxyl at C-1. Members of this subgroup also generally contain the 5-a-hydroxyl,


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
ISOLATION AND CHARACTERIZATION OF TAXANES AND OTHER COMPOUNDS
FROM VARIOUS SPECIES OF TAXUS
by
Richard M. Davies
December 1998
Chairman: Koppaka V. Rao
Cochairman: John H. Perrin
Major Department: Medicinal Chemistry
Taxol is a promising antineoplastic agent originally reported in 1971 by Wani and
Wall, isolated from the bark of the Pacific yew (Taxus brevifolia). Intensive research in
the last decade has demonstrated that this drug possesses exceptional activity in the
treatment of many difficult types of cancer.
From the beginning taxol has proven to be a difficult compound to obtain, with
very low yields and a highly complex structure with many chiral centers and sensitive
moieties. Originally obtained from the bark of a very slow growing tree, the possibility of
growing various Taxus (yew) species under hydroponic conditions has been investigated
in this project.
One local variety, known as Taxus floridana (Florida yew) was found to grow well
and produce taxol and other useful taxanes. During initial investigations a simple and
elegant method for the isolation of taxol using reverse phase bonded silica was
developed. Generous funding by the University of Florida Division of Sponsored
x


38
hexaol [3-11], obtained as a crystalline solid. This was then acetylated to a crystalline
acetate [3-12], The 1H- and 13C NMR spectra of [3-12] showed the presence of 5
acetates (1H: 5 21.8, 8 21.7, 5 21.4, 8 21.3, 8 21.0, 8 20,8; and 13C 8 171.0, 8 170.4, 8
169.8, 8 169.6 and 8 169.5); which suggested that brevifoliol contained a tertiary
hydroxyl.
In the conventional taxane skeleton, a tertiary hydroxyl is often present at C-1,
with the other hydroxyls (or esters) at C-2, C-5, C-7, C-9, C-10 and C-13. Thus, with
brevifoliol having five oxygen substituents, one of these positions must be without
attached oxygen. Thus, it would be important to know which of these positions does not
have an oxygen substituent. For this reason, the hexaol [3-11] was subjected to
oxidation by periodate. If there were two pairs of vicinal hydroxyls, e.g. 1,2 and 9,10, the
hexaol will be cleaved in such a way as to give smaller molecules which represent the A
and C rings. If there is only one such pair, the reaction will produce a product with all of
its carbons intact. The hexaol underwent oxidation readily to form a dialdehyde [3-13]
without losing any carbon atoms found in original carbon skeleton. Unaware of the
unusual A ring structure, it was presumed that the presence of a tertiary hydroxyl at C-1
precluded the presence of oxygen substitution at C-2. Additional evidence for a
methylene carbon at C-2 was found in the COSY spectrum from the chemical shifts in
the H-3p-H-2a-H-2p isolated spin system.
Thus, brevifoliol has two hydroxyls at 5 and 13. Locating the benzoate group at
one of the three choices, 7, 9, or 10 will elucidate the structure. At this point, brevifoliol
was required in microbial and fungal biotransformation project in our laboratory. In order
to produce an antiseptic sample an aqueous alcoholic solution was sterilized in a steam
autoclave at 125 0 C, 20 atm., to see if it is stable. It was found that the compound
underwent degradation to give two or three products.


CHAPTER 7
NON-TAXANE COMPONENTS FROM THE BARK AND NEEDLE EXTRACTS
General
Some of the benefits of using reverse phase rather than normal phase
chromatography have been described in previous chapters. Two important
disadvantages of normal phase silica gel chromatography are the acidic nature of silica
and the tendency for irreversible adsorption to occur. Both of these problems can lead
to significant loss of the compound(s) of interest. Fortunately, almost all free acidic
groups are capped during the bonding process used to make reverse phase silica,
followed by a final capping with trimethylsilyl groups. This process effectively eliminates
these problems of acidity and irreversible adsorption.
These properties allow recovery of many compounds that would normally not be
amenable to silica gel chromatography. Glycosides, phenols, steroids and hydrophilic
compounds are often difficult to chromatograph using normal phase silica gel. During
the processing thousands of pounds of bark and lesser amounts of needles many
interesting compounds were isolated. The taxane glycosides are most notable among
these, especially now that the efficient removal of the glycosyl moiety has become
possible (Rao, 1997). The relative abundance of taxane glycosides amenable to
conversion into taxol gives further support to the use of reverse phase columns. Large
amounts of valuable precursors are lost with the normal Polysciences isolation process
(Boettner et al. 1979).
100


67
Figure 4-5 : H,H-COSY Spectrum of [4-4]
The 1H NMR spectrum was similar to brevifoliol in most respects, with a few
revealing differences. The AB quartet normally seen from H-2p became a clean doublet
(J=13.8 Hz) deshielded from S 2.36 to 5 3.31, and coupled to the H-2a proton, which was
deshielded from 5 1.49 to 5 2.34. This deshielding and the coupling patterns indicated
the loss of H-3p with the possible formation of a double bond between C-3 and C-4.


CHAPTER 1
HISTORICAL OVERVIEW OF TAXUS
During the late 1950s, the National Cancer Institute initiated a program with the
objective of discovering compounds from natural sources, which might prove useful in
the treatment of various human cancers. In this program, plant samples from various
parts of the world were collected and brought to participating laboratories, where the
active principles were isolated, chemically characterized, and subjected to testing in
various murine tumors.
It is in such a context that a sample of the bark of the Pacific yew (Taxus
brevifolia, Nutt.) was extracted and the active principle called taxol* was reported by
Wani and coworkers in 1971 (Wani et al. 1971). Although it exhibited potent cytotoxicity
in some tumor assays, many unsuccessful lead compounds also are cytotoxic in these
assays. Taxol did not appear exceptional, and the problems of low yield and poor
solubility discouraged the pursuit of further research for many years. During the late
1970s, activity against B-16 melanoma in mice, and several human tumors grown in
athymic mice was recognized (e.g. MX-1 human mammary xenograft). These activities
rekindled interest in taxol as a candidate for cancer treatment, resulting in further studies
and human clinical trials. In the ensuing years the pace of research into taxol and
taxoids has increased dramatically.
*Taxol is a registered trademark since 1993 by Bristol Myers Squibb Co., with
paclitaxel being the generic name. Taxol will be used in this dissertation as the generic
name, as this work was started before this change.
1


4
course of therapy for a patient and this translates into requiring the bark from five to ten
trees based on such reported yields.
Only a few studies on the taxol content of other species of Taxus were published
from the time its importance was recognized in 1980 until 1992. From the bark of T.
wallichiana were isolated taxol and the closely related cephalomannine, as well as other
taxoids (Miller et at. 1981; Miller, 1980). From the bark of T. baccata L, Senilh et al.
isolated nearly 20 different taxoids, including taxol, cephalomannine and a series of
xyloside derivatives of these (Senilh et al. 1984). The taxoid content of the needles of
Taxus baccata was studied and 10-deacetyl baccatin III isolated in relatively high yields
(Chauviere et al. 1981). The fractions from the large scale (Polysciences) process from
the bark of T. brevifolia were also investigated to recover any other taxoids with useful
activity, or with possible semisynthetic utility such as conversion to taxol. Flowever, only
minute yields of 10-deacetylbaccatin III, 7-epitaxol and 10-deacetyl-10-oxotaxol were
reported in these studies (Huang et al. 1986; Kingston et al. 1982); leaving a strong
impression that the bark of T. brevifolia is a poor source for not only taxol, but also for
any other useful analogues of taxol.
In spite of these problems, the bark of T. brevifolia has been accepted as the
primary source for taxol until recently. However, since at the expected demand for taxol
and the yields that can be realized from the bark, the yew tree population would be
depleted in a few years, the use of the bark must stop. Among the alternatives that were
being, considered to avoid this prospect are the following: 1) the use of needles, which
are a renewable source, 2) growing the plant In tissue culture, 3) semi-synthesis from
appropriate naturally occurring taxoids and 4) total synthesis. Progress has been made
on all of these fronts.
As far as the needles are concerned, the most important candidate selected for
direct isolation of taxol is the ornamental yew (Taxus x media cultivar Hicksii). Other


Characterization of the Taxane Constituents of Taxus floridana 95
10-Deacetyl Baccatin III [2-7] 95
Brevifoliol [3-1] 95
Taxiflorine [6-1] 96
Baccatin VI [6-2] 96
Taxol [5-3] 98
Acetylation of Taxiflorine to [6-3] 98
Benzoylation of Taxiflorine to [6-4] 98
Saponification and Acetylation of [6-1] to [6-5] 99
7 NON-TAXANE COMPONENTS FROM THE BARK AND NEEDLE EXTRACTS .... 100
General 100
Experimental 102
Flavonoids 103
Quercetin Rutoside (Rutin) 103
Quercetin 104
Sciadopitysin 104
p-Sitosterol-p-D-Glucoside 105
p-Sitosterol-p-D-Glucoside Tetra-acetate 105
(I Sitosterol 106
Phytoecdysteroids 107
Ecdysterone & 2(3, 3(3, 22a-Triacetate 107
Ponasterone A and 2(3, 3p, 22a Triacetate 107
Phenolic Compounds 108
Usnic Acid 108
Betuloside (4-(4-Hydroxyphenyl)-2R-butanol Glucoside)) & Aglycone 109
LIST OF REFERENCES 110
BIOGRAPHICAL SKETCH 115
vi


ISOLATION AND CHARACTERIZATION OF TAXANES AND OTHER COMPOUNDS
FROM VARIOUS SPECIES OF TAXUS
By
RICHARD M. DAVIES
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1998

This work is dedicated to the memory of Dr. Koppaka V. Rao, an extraordinary
scientist, dedicated teacher and very dear friend. I feel blessed to have known and
worked with him.

ACKNOWLEDGMENTS
Much of the work in this dissertation was done with the guidance and expertise of
the late Dr. K. V. Rao, who passed away on February 20, 1998. Professor Rao and I
had known each other since 1981 and I am grateful to him for encouraging me to return
for graduate studies and for his true friendship with me. I would also like to thank his
wife and children for their much appreciated support, friendship and encouragement.
I wish to thank Dr. John Perrin for assuming the position of chairman of my
supervisory committee and for his kind encouragement and guidance. He has helped
me in many ways and I am very grateful to him for his persistence in pushing me to
complete this work. I would also like to thank Dr. Margaret James, Dr. Jonathan Eric
Enholm, Dr. Kenneth Sloan, and Dr. Stephen Schulman for participating on my
supervisory committee and for their thoughtful advice and expertise.
I wish to thank my mother and father for their kind encouragement and love, and
also my three sisters, brothers-in-law, niece and nephews. I would like to thank the
Graduate School and many other University of Florida personnel for all of the kind
assistance they have provided, especially Gladys Jan Kalman and Nancy Rosa for all of
their helpful assistance.

TABLE OF CONTENTS
page
ACKNOWLEDGMENTS iii
LIST OF TABLES vii
LIST OF FIGURES vii:
ABSTRACT x
CHAPTERS
1 HISTORICAL OVERVIEW OF TAXUS 1
Background of Research at the University of Florida 6
Methods 6
2 A SELECTED REVIEW OF THE LITERATURE ON TAXUS 11
Earlier Studies 11
Studies after the Discovery of Taxol 14
Semi-synthesis of Taxol 17
Total Synthesis 18
Other Synthetic Approaches 20
General Structural Features of Taxanes 24
Taxa-4(20);11-dienes 25
4(20)-Epoxides 25
Oxetanes 26
Abeotaxanes 28
3 TAXANE CONSTITUENTS OF THE NEEDLES OF TAXUS BREVIFOLIA 29
Fractionation of the Needles ofTaxus brevifolia 29
Brevitaxane A (Brevifoliol) [3-1] 31
Hydroxyl Functionalities 33
4/20 Unsaturation 36
Number and Nature of the Oxygen Substitution 37
Experimental 44
Extraction of the Needles of Taxus brevifolia 44
Reverse Phase Column Chromatography: 45
Brevifoliol [3-1] 46
Brevifoliol-5-Monoacetate [3-2] 48
iv

Brevifoliol-5,13-Diacetate [3-4] 49
Brevifoliol-13-Ketone [3-6] 50
Dihydrobrevifoliol [3-7] 51
Brevifoliol Epoxide [3-8] 51
Ozonization of Brevifoliol: Brevifoliol-norketone [3-9] 52
Brevifoliol-4,20-Diol [3-10] 53
Saponification of Brevifoliol [3-11] 54
Debenzoyl Brevifoiiol-Pentaacetate [3-12] 54
Periodate Oxidation of [3-11] to [3-13] 55
Formation of Osazone [3-14] from [3-13] with 2,4-DNPH 55
Debenzoyl Brevifoliol [3-15] 56
4 SOME UNUSUAL REACTIONS OF BREVIFOLIOL 57
1. Acid-Catalyzed Acetylation 58
2. Oxidation 59
3. Action of BF3 on Brevifoliol [4-3] 61
4. Reaction with lodine/Silver Acetate [4-4] 65
Experimental 69
Brevifoliol Triacetate [3-5] 69
Oxidation with Jones Reagent to [4-1] 69
Action of Boron Trifluoride on Brevifoliol [4-3] 70
Reaction with Iodine and Silver Acetate [4-4] 70
Acetylation of [4-4] to [4-5] 72
Reaction with N-Bromosuccinimide and Silver Acetate [4-6] 72
Reaction of [4-4] with N-Bromosuccinimide [4-7] 73
5 TAXANE CONSTITUENTS OF THE NEEDLES OF TAXUS X MEDIA 74
Brevifoliol [3-1] 79
Taxanes I [5-1] and II [5-2] 79
Taxane III [2-1] 80
Taxane IV [2-2] 80
Taxol [5-3] 80
Ozonolysis of [2-2] 80
Experimental 81
Extraction: 81
Chromatography: 81
Characterization of the Taxane Components of Taxus x Media Hicksii 82
Brevifoliol [2-1] 82
Taxanes I and II [5-1] and [5-2] 83
Taxane III [2-1] 84
Taxane IV [2-2] 84
Taxol [5-3] 85
Ozonolysis of Compound [2-2] 86
6 TAXANE CONSTITUENTS OF TAXUS FLORIDANA 87
Taxiflorine 89
Experimental 93
Extraction 93
v

Characterization of the Taxane Constituents of Taxus floridana 95
10-Deacetyl Baccatin III [2-7] 95
Brevifoliol [3-1] 95
Taxiflorine [6-1] 96
Baccatin VI [6-2] 96
Taxol [5-3] 98
Acetylation of Taxiflorine to [6-3] 98
Benzoylation of Taxiflorine to [6-4] 98
Saponification and Acetylation of [6-1] to [6-5] 99
7 NON-TAXANE COMPONENTS FROM THE BARK AND NEEDLE EXTRACTS .... 100
General 100
Experimental 102
Flavonoids 103
Quercetin Rutoside (Rutin) 103
Quercetin 104
Sciadopitysin 104
p-Sitosterol-p-D-Glucoside 105
p-Sitosterol-p-D-Glucoside Tetra-acetate 105
(I Sitosterol 106
Phytoecdysteroids 107
Ecdysterone & 2(3, 3(3, 22a-Triacetate 107
Ponasterone A and 2(3, 3p, 22a Triacetate 107
Phenolic Compounds 108
Usnic Acid 108
Betuloside (4-(4-Hydroxyphenyl)-2R-butanol Glucoside)) & Aglycone 109
LIST OF REFERENCES 110
BIOGRAPHICAL SKETCH 115
vi

LIST OF TABLES
Table page
3-1 : Proton NMR Spectra of Brevifoliol and Brevifoliol Acetates 33
3-2 : Carbon NMR Spectra of Brevifoliol and Brevifoliol Acetates 34
4-1 : NMR Spectra of Compound [4-3] from BF3 Reaction 63
6-1: Proton NMR Spectra of Compounds [6-3], [6-4] and [6-5] 91
vii

LIST OF FIGURES
Figure page
2-1 : Early Studies on the Constituents of some Taxus Species 13
2-2 : Taxol and some Synthetic Targets 16
2-3 : Nicolaous Retrosynthetic Strategy 18
2-4 : Nicolaous Taxane Ring Synthesis 21
2-5 : Nicolaous Final Synthetic Intermediates 22
2-6 : Starting Points of Other Synthetic Strategies 23
3-1 : Proton NMR Spectrum of Brevifoliol 32
3-2 : Brevifoliol and Reaction Products 35
3-3 : Brevifoliol Hexaol Reaction Products 39
4-1 : Oxidation Products 61
4-2 : BF3-etherate Catalyzed Elimination Product 64
4-3 : DEPT Spectra of BF3 Elimination Product [4-3] 64
4-4 : lodine/Silver Acetate Product [4-4] and Acetate 66
4-5 : H,FI-COSY Spectrum of [4-4] 67
4-6 : FIETCOR Spectrum of [4-4] 68
5-1 : Fractionation of the Extract of Taxus x media Hicksii Needles 77
5-2 : FIPLC Trace of Taxanes Coeluting with Taxol 78
5-3 : Progress of Elution of Taxanes from Reverse Phase Column 78
6-1 : Taxanes and Analogues from Taxus x media Hicksii 92
viii

6-2 : Carbon NMR Spectrum of Baccatin VI
97
ix

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
ISOLATION AND CHARACTERIZATION OF TAXANES AND OTHER COMPOUNDS
FROM VARIOUS SPECIES OF TAXUS
by
Richard M. Davies
December 1998
Chairman: Koppaka V. Rao
Cochairman: John H. Perrin
Major Department: Medicinal Chemistry
Taxol is a promising antineoplastic agent originally reported in 1971 by Wani and
Wall, isolated from the bark of the Pacific yew (Taxus brevifolia). Intensive research in
the last decade has demonstrated that this drug possesses exceptional activity in the
treatment of many difficult types of cancer.
From the beginning taxol has proven to be a difficult compound to obtain, with
very low yields and a highly complex structure with many chiral centers and sensitive
moieties. Originally obtained from the bark of a very slow growing tree, the possibility of
growing various Taxus (yew) species under hydroponic conditions has been investigated
in this project.
One local variety, known as Taxus floridana (Florida yew) was found to grow well
and produce taxol and other useful taxanes. During initial investigations a simple and
elegant method for the isolation of taxol using reverse phase bonded silica was
developed. Generous funding by the University of Florida Division of Sponsored
x

Research made possible the construction of a pilot plant scale facility where these
isolation methods were successfully implemented.
Excellent yields and the isolation of many related taxanes have proven that this
method is superior to currently approved processes used in the production of taxol. The
failure of other researchers to employ bonded silica gel for preparative columns in the
past may reflect experiences with analytical columns, but this method has proven to be
quite exceptional and should be employed extensively.
This dissertation covers many crystalline and non-crystalline compounds isolated
and characterized as a result of this project. Some results from the application of this
technique for the isolation of taxanes from the needles of Taxus brevifolia, Taxus x
media cultivar Hicksii, and T. floridana are presented. Similar experiments on the bark
and wood of T. brevifolia are also described.
xi

CHAPTER 1
HISTORICAL OVERVIEW OF TAXUS
During the late 1950s, the National Cancer Institute initiated a program with the
objective of discovering compounds from natural sources, which might prove useful in
the treatment of various human cancers. In this program, plant samples from various
parts of the world were collected and brought to participating laboratories, where the
active principles were isolated, chemically characterized, and subjected to testing in
various murine tumors.
It is in such a context that a sample of the bark of the Pacific yew (Taxus
brevifolia, Nutt.) was extracted and the active principle called taxol* was reported by
Wani and coworkers in 1971 (Wani et al. 1971). Although it exhibited potent cytotoxicity
in some tumor assays, many unsuccessful lead compounds also are cytotoxic in these
assays. Taxol did not appear exceptional, and the problems of low yield and poor
solubility discouraged the pursuit of further research for many years. During the late
1970s, activity against B-16 melanoma in mice, and several human tumors grown in
athymic mice was recognized (e.g. MX-1 human mammary xenograft). These activities
rekindled interest in taxol as a candidate for cancer treatment, resulting in further studies
and human clinical trials. In the ensuing years the pace of research into taxol and
taxoids has increased dramatically.
*Taxol is a registered trademark since 1993 by Bristol Myers Squibb Co., with
paclitaxel being the generic name. Taxol will be used in this dissertation as the generic
name, as this work was started before this change.
1

2
A study of its mode of action revealed that it blocked cell division at the cell cycle
through its specific action on the G2/M phase of the tubulin/ microtubule system. Unlike
other antitumor drugs such as colchicine, vincristine and vinblastine, which act as tubulin
poisons, taxol exhibited a novel mode of action (Schiff et al. 1979). Microtubules are
involved in the formation of the mitotic spindle fibers necessary for the replication of DNA
and are also integral building blocks within the cell wall. They are generated from a
protein known as tubulin, and a dynamic equilibrium exists between tubulin and
microtubules in vivo. In the presence of taxol, the polymerization of tubulin produces
what are now known as oligo-microtubules. In contrast to the usual microtubules, which
can be readily disassembled, these oligo-microtubules resist disassembly to tubulin,
thereby preventing cell division (Horwitz, 1992).
Based on potent activity against important experimental tumors and its unique
mode of action, interest in taxol became greatly enhanced, and it was approved for
human Phase I clinical trials in the early 1980s. Taxol showed significant activity in
human tumors in Phase I and Phase II clinical trials, especially in ovarian and breast
carcinomas (McGuire et al. 1989; Holmes et al. 1991). The scientific community took a
special interest in taxol at that time due to the lack of adequate treatment options
available for ovarian cancer.
This knowledge established taxol as an important antitumor drug and stimulated
a renewed interest in it. Intensive worldwide studies have reached explosive proportions
since 1994 concerning its production, chemistry, biochemistry and many other aspects.
At that point, two problems needed solution before taxol could become a viable
alternative as a useful treatment against any type of cancer. First, the lipophilic nature of
taxol made it difficult to develop an acceptable dosage form for this drug. Gradually, this
was overcome by the introduction of a suitable, relatively non-toxic dosage form.

3
Interestingly, the poor solubility characteristics of taxol might prove to be
responsible for new discoveries regarding a problem of cross resistance to different
classes of chemotherapeutic agents, caused by non-specific drug efflux and referred to
as the MDR phenotype. The MDR phenotype is a gene that has been linked to multiple
drug resistance (hence MDR); and some studies indicate that the solvent used for the
delivery of taxol might have good activity against this common cause for therapeutic
failure in the treatment of cancer (Woodcock et al. 1990; Webster et al. 1993; Fjllskog
et al. 1993). It is known that this effect can result from the expression of plasma-
membrane transport proteins (P-glycoproteins) which can enhance the efflux of
structurally unrelated compounds from the cancer cell. At least three reports suggest
that the solvent Cremaphore LH might enhance the antitumor actions of taxol when the
tumor(s) display the MDR phenotype, and further work with cremaphores alone and in
combination with other antitumor agents is needed to clarify this seemingly serendipitous
finding.
Cremaphore LH is a form of ethoxylated castor oil and is responsible for many
adverse drug reactions during the administration of taxol, and pretreatment with
corticosteroids and antihistamines if often required to prevent allergic response up to
and including anaphylaxis and death. Perhaps more difficult than this solubility concern
was the procurement of an adequate supply of taxol for clinical trials and the anticipated
needs for subsequent worldwide clinical use. Reported yields of taxol from the dried
bark of T. brevifolia were averaging around 0.01%.
A large-scale process for the isolation of taxol was developed by Polysciences,
Inc. (Paul Valley Industrial Park, Warrington, PA 18976); with yields of 0.005-0.01%
(Boettner et al. 1979). Under these conditions, one kilogram of the bark could be
expected to provide only 50-100 mg of taxol at best (or 30,000 lbs. being required for
obtaining one Kg. of taxol). Approximately 2 grams of taxol are needed for one complete

4
course of therapy for a patient and this translates into requiring the bark from five to ten
trees based on such reported yields.
Only a few studies on the taxol content of other species of Taxus were published
from the time its importance was recognized in 1980 until 1992. From the bark of T.
wallichiana were isolated taxol and the closely related cephalomannine, as well as other
taxoids (Miller et at. 1981; Miller, 1980). From the bark of T. baccata L, Senilh et al.
isolated nearly 20 different taxoids, including taxol, cephalomannine and a series of
xyloside derivatives of these (Senilh et al. 1984). The taxoid content of the needles of
Taxus baccata was studied and 10-deacetyl baccatin III isolated in relatively high yields
(Chauviere et al. 1981). The fractions from the large scale (Polysciences) process from
the bark of T. brevifolia were also investigated to recover any other taxoids with useful
activity, or with possible semisynthetic utility such as conversion to taxol. Flowever, only
minute yields of 10-deacetylbaccatin III, 7-epitaxol and 10-deacetyl-10-oxotaxol were
reported in these studies (Huang et al. 1986; Kingston et al. 1982); leaving a strong
impression that the bark of T. brevifolia is a poor source for not only taxol, but also for
any other useful analogues of taxol.
In spite of these problems, the bark of T. brevifolia has been accepted as the
primary source for taxol until recently. However, since at the expected demand for taxol
and the yields that can be realized from the bark, the yew tree population would be
depleted in a few years, the use of the bark must stop. Among the alternatives that were
being, considered to avoid this prospect are the following: 1) the use of needles, which
are a renewable source, 2) growing the plant In tissue culture, 3) semi-synthesis from
appropriate naturally occurring taxoids and 4) total synthesis. Progress has been made
on all of these fronts.
As far as the needles are concerned, the most important candidate selected for
direct isolation of taxol is the ornamental yew (Taxus x media cultivar Hicksii). Other

5
than analytical HPLC studies on the taxol content of the needles under various
conditions, no practical methodology for the isolation of taxol or other taxolds has been
published. Publications from this laboratory which address these issues are essentially
the only work available in the literature (Rao et al. 1995; Rao et al. 1996). In addition to
direct isolation of taxol, the needles were also examined for the presence of analogues
such as 10-deacetyl baccatin III, since semi-synthesis from such is already an important
alternative. The two most important species, T. baccata L. and T. wallichiana Zucc.,
have become the focus of attention since they were demonstrated to contain the highest
concentrations of 10-deacetyl baccatin III.
Growing of various tissues of T. brevifolia in plant cell culture has been under
development since 1990 and the methods have been standardized in many laboratories.
However, the yields, as yet, have not been very attractive. Further research is expected
to overcome this problem. Work on this alternative will continue due to the
attractiveness of this approach and its potential for large-scale operations.
Starting with 10-deacetyl baccatin III, considerable progress was made in the
area of semi-synthesis. In the first recorded semi-synthesis of taxol, the 13-cinnamate
ester of 7-protected baccatin III was converted to the phenyl isoserine ester through a
Sharpless hydroxy-amination (Denis et al. 1988). At this point, as an alternative to
benzoylation of the amino group that will yield taxol, a t-BOC group (tert-
butoxycarbonyl) was introduced, along with leaving the 10-hydroxyl free, to obtain an
analogue known as taxotere. On the basis of its activity, taxotere has also been
approved as an antitumor drug. Two important schemes for preparing taxol from 10-
deacetyl baccatin III have been well developed and used for the large scale semi
synthesis of taxol as discussed in Chapter 2 (Denis et al. 1994; Ojima et al. 1991; Holton
et al. 1992).

6
Two total syntheses of taxol have been recorded, and several other approaches
towards the synthesis have also been reported in the literature as discussed briefly in
Chapter 2 (Nicolaou et at. 1994a, 1994b, 1995a, 1995b, 1995c; Holton et at. 1994a,
1994b). Although these methods demonstrate remarkable achievements in the field of
synthetic organic chemistry, they do not offer a practical method for the large-scale
production of taxol or its analogues at this time.
Background of Research at the University of Florida
As part of one of these alternative quests, the National Cancer Institute hoped
that instead of using the bark of the Pacific yew, the plant should be grown under
hydroponic conditions, and they wanted to know whether plants grown in this manner
would produce enough taxol for isolation. This laboratory was approached with this idea
in early 1990, and with collaboration from Prof. George Hochmuth, Jr., of IFAS,
University of Florida, the project was started. More on this aspect will be discussed in
Chapter 6. In order to learn the current knowledge concerning the analysis and isolation
of taxol, the pertinent literature was consulted. This yielded only a few papers on the
isolation of taxol and taxoids, which were outlined above.
Methods
In general, the methods were found to be too cumbersome for others to repeat.
For example, one of these publications in which isolation of taxol and its analogues was
described from T. wallichiana, used the following steps, starting with the concentrated
ethanolic extract of the plant:
1. Partition between water and hexane
2. Extraction of the aqueous phase with chloroform
3. Silica gel chromatography on the chloroform extract
4. A second silica gel chromatography

5.
6.
7.
7
Counter-current distribution
HPLC on the appropriate fractions
A second HPLC on the appropriate fractions
Similarly, the large-scale process developed for the isolation of taxol by
Polysciences Inc. from the bark of T. brevifolia consisted of the following steps, again
starting with the alcoholic extract concentrate (Boettner et al. 1979).
1. Solvent partition water and CH2CL2, concentration to a solid
2. Separation of the extract solid into soluble and insoluble fractions
3. Chromatography on the soluble fraction
4. Recovery of taxol and crystallization twice
5. Silica chromatography on the taxol/ cephalomannine mixture
6. Recovery and crystallization of taxol
Thus, it appeared that, although procurement of taxol was of top priority, and
many alternative approaches were attempted for solving this problem, one alternative,
which was not considered, was to study the existing isolation procedure itself to make it
more efficient. Thus the approach pursued at the University of Florida during 1990-91
was to develop a simpler process for the isolation. Over the next few months, a new
process was developed based on the use of a single reverse-phase chromatographic
column, and consisting of the following steps, starting with the alcoholic extract
concentrate (Rao, 1993).
1. Partition between water and chloroform, and concentration
2. Reverse phase column chromatography on the extract directly
3. Harvesting the crystals and recrystallization.
The total chloroform extract of the bark of T. brevifolia, was applied directly to the
C18-bonded silica column in 25% acetonitrile in water (i.e., no separation into soluble
and insoluble fractions); and the column developed with a step gradient (30-60%
acetonitrile). The column fractions were let stand for 3-7 days, whereby taxol and seven

8
of its analogues crystallized out directly from the fractions. These are filtered and
purified further by recrystallization, or subjected to a small column.
This process using reverse phase column chromatography, gave not only higher
yields of taxol (0.02-0.04% vs. 0.01%) on a pilot plant scale, but also made possible the
simultaneous isolation of a number of analogues which have not been obtained from this
plant before. These included a 10-deacetyl baccatin III (0.02%); and a number of
xyloside analogues, chief among which being the 10-deacetyltaxol-7-xyloside, which can
now be isolated in yields of 0.1% or higher.
Based on the successful fractionation of the bark extract of T. brevifolia, this
technique was then ready for application to the other extracts such as the needles and
wood of T. brevifolia, and to the needles of two other species of Taxus. These
applications which gave practical methodology for processing these various extracts,
also yielded many interesting taxoid compounds and these experiments are all detailed
in this dissertation.
Although this work was started during 1991 and much of the expected work was
completed by late 1993, the world-wide interest in taxol research made a quantum leap
at about this time, with a phenomenal increase in publications dealing with all aspects of
taxol chemistry. Some of the compounds which were isolated for the first time in this
laboratory, and whose structures were determined, were rediscovered by others and
published. In spite of the enormous increase in the number of relevant publications,
most of these publications described the isolation of the minimum possible amounts of
the compounds, often as amorphous solids. Many determined their structures only
through NMR spectral interpretation, with little or no other physical characterizations,
elemental analyses, derivatizations or reactions. In at least a few examples, the
assigned structures were found to be wrong and were subsequently corrected once or

9
even twice. In the present work, practical isolation methods were used to obtain gram
quantities of many compounds, as crystalline solids, where possible.
The compounds are usually characterized by physical and spectral properties,
providing elemental analyses, and carrying out derivatization such as acetylation,
oxidation, etc. The structures were elucidated through chemical reactions as well as
through spectral data. Thus, even though some of the final structures may have been
published, the work described here contains experiments that have not been carried out
by these authors.
Brief descriptions of the topics that appear in this dissertation are given below.
Chapter 2 gives a brief and selected summary of the pertinent literature on taxol
and taxoids, covering the areas of isolation, elucidation of structures, semi-syntheses
and total syntheses. Because the subject matter expanded enormously since 1993, the
scope of the review is limited to material that is relevant to the subject matter of the
dissertation.
Chapter 3 deals with the taxoid composition of the needles of Taxus brevifolia. It
covers the application of the reverse phase column chromatography to the needle
extract, isolation of the major taxoid, brevitaxane A (or brevifoliol), along with brevitaxane
B, and taxol. It continues with the elucidation of the structure of brevitaxane A by
various reactions, as well as by a detailed analysis of the NMR spectral evidence.
Chapter 4 discusses some unusual reactions of brevifoliol. Such reactions have
not been reported with this or any other taxoid compounds. In each case, crystalline
compounds were obtained and characterized by physical and spectral data.
Chapter 5 deals with fractionation of the extract of the needles of Taxus x media
cv. Hicksii by reverse phase column chromatography and isolation of taxol and several
other taxoids and their characterization. In spite of the fact that this species (ornamental
yew) was declared as the preferred plant for the future isolation of taxol, no publications

10
describing a suitable scheme for isolation of taxol or any other taxoids have appeared so
far, other than analytical hplc data on their taxol content.
Chapter 6 is similarly devoted to the fractionation of the extract of the needles of
Taxus floridana Nutt, by reverse phase column chromatography. Isolation of taxol, 10-
deacetyl baccatin III, baccatin VI and a new crystalline taxoid compound named
taxiflorine, with its structural elucidation are described.
Chapter 7 deals with the isolation of several crystalline non-taxane compounds
present in the extracts of the bark and needles of Taxus brevifolia. These were shown
to include flavonoids, phenols, and other types of compounds.

CHAPTER 2
A SELECTED REVIEW OF THE LITERATURE ON TAXUS
An overview of the taxol story has been presented in Chapter 1. In this Chapter,
a selected review of the literature will be presented on taxol as well as the other taxanes,
which are included in this dissertation. Two comprehensive reviews have been
published on the subject of taxol, one by Kingston et al. (1993) and one by Miller (1980).
The present review on the genus Taxus may be roughly divided into two parts:
studies before and studies after the discovery of taxol.
Earlier Studies
The genus Taxus (N.O. Taxaceae) represents a group of plants (common name,
yew) which grow mostly in temperate climates and can be found distributed throughout
the world. They are generally slow-growing evergreen trees or shrubs with stiff linear
leaves (or needles); and fruits which are small, fleshy and bright red. The common
names of the plants are qualified by the place of its origin, as for example, Pacific or
western yew (T. brevifolia, Nutt.); European or English yew (7. baccata, PHg.)\ Canadian
yew (7. canadensis, Willd.); Japanese yew (7. cuspidata, Sieb. et Zucc.); Chinese yew
(7. chinensis, Pilg ); Himalayan yew (7 wallichiana, Zucc.); ornamental yew (7. x media
Hicksii, Rehd.) and Florida yew (7 floridana, Nutt.).
The toxic nature of this genus has been recognized for thousands of years, and
in modern times, was first investigated chemically using the needles of Taxus baccata
(Lucas, 1856). An amorphous mixture of alkaloids was isolated after extraction under
acidic conditions and was given the name of "taxine. Further studies on taxine spanned
11

12
many decades and covered many reactions relevant to taxane chemistry. In one such
study, Winterstein and Guyer (1923) were the first to show the presence of 3-
dimethylamino-3-phenylpropanoic acid in the hydrolyzate of taxine and this acid later
became known as "Winterstein's acid.
Until the 1960s most of the work on taxanes focused on these acid-extractable
alkaloidal substances, which were readily separable from the large quantities of neutral,
resinous materials which dominate the extract. Two groups of researchers were able to
convert these somewhat unstable alkaloidal mixtures into more stable, non-basic
substances in which the 3-dimethylamino-phenylpropanoic ester unit was transformed
into a cinnamate ester. This, as well as the development of chromatographic
techniques, made it possible to obtain pure compounds rather than mixtures.
Baxter et al. (1958) in England investigated the major cinnamate ester obtained
from T. baccata, which they named 5-O-cinnamoyl taxicin-l triacetate [2-1], Similarly, a
Japanese team (Nakanishi & Kurono, 1963; Kurono et al. 1963) studied a cinnamate
ester from T. cuspidata, and called it 5-O-cinnamoyl taxicin II triacetate [2-2] and the
structures of both these can be seen in Figure 2-1. These two compounds differ only at
C-1, where taxicin II lacks the tertiary hydroxyl found in taxicin I. The IUPAC numbering
system for taxanes used throughout this dissertation can also be seen [2-3],
A few years earlier, Graf & Betholdt (1957) succeeded in isolating the purified
basic alkaloids, taxine A and taxine B from the original taxine mixture. Taxine B was
shown to have the structure [2-4] (see Figure 2-1); which corresponded with 5-O-
cinnamoyl taxicin I triacetate, into which it could be converted via elimination of the
dimethylamine moiety.

13
[2-1] 0-C¡nnamoyl Taxicin I Triacetate, R = OH [2-3]-IUPAC
[2-2] O-Cinnamoyl Taxicin II Triacetate, R = H Numbering
[2-4] Taxine B
[2-5] Geranylgeranyl
pyrophosphate
[2-6a] Proposed Glycol,
later corrected to 6 b
[2-6b] R = OAc, Baccatin II
[2-7] R = OH 10-DAB-lll
Figure 2-1 : Early Studies on the Constituents of some Taxus Species

14
Harrison (Harrison & Lythgoe 1966; Harrison et al. 1966) published one of the
earliest biogenetic theories for the formation of taxanes starting with geranylgeranyl
pyrophosphate and electrophilic cyclization [2-5], Efforts by many groups to utilize a
similar scheme to synthesize the taxane skeleton have been unsuccessful thus far
(Kumagai et at. 1981; Hitchcock & Pattenden, 1992). Biogenetic pathways often provide
ideas for simplified approaches in the synthesis of natural products.
Many early studies utilized acidic conditions for the extraction which might have
hampered the isolation of neutral or acid-labile compounds, Kondo & Takahishi (1925)
obtained a non-basic compound from the Japanese yew by using neutral conditions.
The cinnamates can also be directly isolated from the plant, indicating that they occur
naturally and also as artifacts of processing.
The National Cancer Institute (NCI) and the U.S. Department of Agriculture
(USDA) joined forces in 1960 to collect and screen plants for activity in several animal
tumor models. Arthur Barclay of the USDA collection team obtained samples from the
Pacific yew tree (Taxus brevifolia Nutt., family Taxaceae) from Washington State in
1962. In 1964, the extracts from the bark and stems were found to be active against KB
cells in vitro (Wani et al. 1971).
Studies after the Discovery of Taxol
Dr. Monroe Wall had discovered another antitumor agent known as camptothecin
using the activity on KB cells for isolation and was interested in any other extracts
showing this activity. Thus, work on T. brevifolia by Wani and Wall at the Research
Triangle Institute was started and led to the isolation of 500 mg of taxol 2 years later in
1966. Cytotoxic actions (Wani et al. 1971) in KB cells, P388 leukemia, Walker 256
carcinosarcoma and P-1534 leukemia were present in the extracts from the bark. These

15
assays were all used at various points to monitor the fractionation, resulting in the
isolation of taxol as the active principle.
In the 1960s the most straightforward and reliable method for the determination
of complex chemical structures was x-ray diffraction analysis of a suitable crystal, also
known as crystallography. Taxol crystallizes as thin needles not suitable for x-ray
studies, but a tetraol derivative was amenable to x-ray studies. The structure of taxol [2-
8] was determined by methanolysis (Figure 2-2); which yielded two compounds: the
methyl ester of N-benzoyl phenylisoserine and an alcohol component shown to be a
taxane tetraol [2-10a], This tetraol skeleton was converted into a 7,10-bis-iodoacetate
derivative and, unlike all of the taxanes studied earlier, taxol showed the following
unique features:
1. A taxane skeleton with an oxetane ring system involving C-4, C-5 & C-20
2. An ester side chain consisting of N-benzoyl phenylisoserine at C-13
3. A carbonyl function at C-9
Baccatin III (Figure 2-1) and its 7a -epimer, baccatin V (Figure 2-2); were shown
to be similar to taxol, having the oxetane ring and the C-9 carbonyl function. These
epimers yielded better crystals and x-ray crystallography was performed. Baccatin III [2-
6b] lacked the ester side chain present in taxol. Still another analogue, known as 10-
deacetylbaccatin III (10-DAB, [2-7]) was later found to be much more widely distributed
in Taxus spp., especially in the needles of Taxus baccata. This became an important
taxane because it could be converted into baccatin III and later to taxol by the
reattachment of the N-benzoyl phenylisoserine side chain at C-13 (See Figure 2-2).
Taxol showed significant antitumor activity against a variety of in vivo murine
tumors including B-16 melanoma and several human xenografts, which qualified it for

16
O
[2-9] Taxotere, RfHR2= OC(CH3)3
[2-10a] Tetraol, R1 = R3 = H R2 = OH
[2-10b] BaccatinV, R-, = Ac
R2 = H, R3 = OH
[2-11] 7-O-TES 10-DAB III, R, = H
R2 = Triethyl Silyl (TES)
[2-12] 7-O-TES Baccatin III, = Ac
R2 = Tri ethyl Silyl (TES)
[2-13] 7-0-TES-13-0-Cinnamoyl Baccatin III
Figure 2-2 : Taxol and some Synthetic Targets

17
clinical trials. A few studies on other species of Taxus have also been published, which
are referred to in Chapter 1.
Semi-synthesis of Taxol
The relative ease in ester formation of the three hydroxyl groups in 10-deacetyl
baccatin III (10-DAB, [2-7]) are 7>1013. Esterification of the C-13 hydroxyl is very
challenging due to the inverted cup-like folding of the taxane skeleton and strong
hydrogen bonding with the carbonyl oxygen on the C-4 acetate. Before the side chain
can be attached at C-13, the 7-hydroxyl must first be protected, often accomplished by
attachment of a triethylsilyl group to give [2-11]. Next, this compound is acetylated at the
10-position, to form 7-triethylsilyl baccatin III [2-12],
In one method, [2-12] was esterified with cinnamic acid to give [2-13], which was
then converted to the phenyl isoserine ester by the Sharpless hydroxyamination
procedure (Sharpless et al. 1991) using osmium tetroxide and t-butyl-N-chloro-N-sodio-
carbamate (Mangatal et al. 1989). The four isomers were separated and after
deprotection of the hydroxycarbamates, N-benzoylation and deprotection of the 7-
hydroxyl, taxol could be obtained.
During the investigations of Greene and Potier (Denis et al. 1988; Kanazawa et
al. 1994) dozens of side chain analogues were synthesized and tested, resulting in the
discovery of the taxol analogue known as taxotere [2-9]. Taxotere (docetaxel) was
found to be more active than taxol in the tubulin assay and animal tumor systems and
has also been approved as an antitumor agent. In an alternative synthesis, the 7-
protected baccatin III [2-12] was esterified using either the chiral p-lactam [2-14] or the
oxazinone [2-15] derivative to yield taxol. This method or some variation is currently
used for the semi-synthesis of taxol and taxotere commercially from 10-deacetyl
baccatin III (Ojima et al. 1991, 1992).

18
Total Synthesis
Swindell (1992) published a review on the progress of more than thirty groups
and reported only modest success in the total synthesis of taxol. Only two years later
two separate groups headed by K. C. Nicolaou (Nicolaou et at. 1994b) at the Scripps
Research Institute and R. A. Holton (Holton et at. 1994b) at Florida State University
would announce almost simultaneously two total syntheses of taxol.
Nicolaou and colleagues designed the strategy for their synthesis based on the
one bond disconnection analysis seen in Figure 2-3. After preparation of the fully
functionalized A ring [2-16] and C ring [2-17] equivalents, a convergent and flexible
Esterification
McMurry coupling
_ Oxetane
q formation
Shapiro reaction
OBn
A
[2-16] Aryl sulphonylhydrazone [2-17] Aldehyde
Figure 2-3 : Nicolaous Retrosynthetic Strategy
synthesis of taxol involving 28 more steps allowed the preparation of numerous
analogues. While not practical for the commercial production of taxol, synthetic methods

19
provide researchers with a source of analogues for structure-activity relationships and
lead to better methods of production in general.
The first carbon-carbon bond between rings A and C was formed using a
vinyllithium carbanion generated from the reaction of aryl sulphonylhydrazone [2-16],
with /7-butyl-lithium in tetrahydrofuran (THF); which was then combined with the
aldehyde [2-17] in the Shapiro reaction (Shapiro, 1976) to produce [2-18] (Figure 2-3).
Regioselective epoxidation of the A1,14-double bond was completed in 87% yield
with f-butyl peroxide in the presence of VO(acac)2 leading to epoxide [2-19], which was
then regioselectively opened with LiAIH4 to give the 1,2-diol [2-20] with a 76% yield. The
carbonate introduced between the C-1 and C-2 hydroxyls in the next step served to
position the two rings for ring closure and also allowed for the stereo-controlled
introduction of the 2a-benzoate later in the sequence. The dialdehyde [2-21] needed for
cyclization of the B ring was obtained after standard deprotection of the two primary
hydroxyls and mild oxidation with tetra-n-propylammonium perruthenate (TPAP) and
4-methylmorpholine N-oxide (NMO) in acetonitrile. The previous three steps provided
the carbonate dialdehyde in 32% overall yield.
Formation of the B ring was accomplished with the versatile McMurry coupling
(McMurry, 1989) under dilute conditions utilizing low valence titanium produced in situ
from (TiCI3)2-(DME)3 (10 eq.) and Zn-Cu (20 eq.) in 1,2-dimethoxyethane (DME) at 70 0
C for 1 hour, giving the tricyclic A/B/C diol [2-22] with a 23% yield.
Selective acetylation of the hydroxyl at C-10 rather than C-9 was expected due to
allylic activation and proceeded with 95% yield. Mild oxidation of the C-10 hydroxyl was
then carried out with TPAP-NMO in acetonitrile analogous to the oxidation to the
dialdehyde with a 93% yield.

20
After removal of the acetonide and protection of the primary hydroxyl at C-20 to
make, the benzyl group was removed with catalytic hydrogenation and the 7-O-triethyl
silyl protecting group was introduced to give [2-23], Selective deacetylation of the
primary acetate then provided the triol for the formation of the oxetane of ring D, which
involves monotosylation at C-20 (primary OH) and triflate formation at C-5 (secondary
OH) to produce [2-24], Oxetane formation with a 60% yield occurs after mild acid
treatment with catalytic camphorsulfonic acid (CSA) in methanol, followed by treatment
with silica gel in dichloromethane.
Acetylation of the C-4 position (tertiary hydroxyl) was followed by regioselective
ring opening of the carbonate to the hydroxybenzoate functionality, both with good
yields. The C-13a oxygen is introduced with pyridine chlorochromate in 75% yield
followed by stereospecific reduction of the ketone [2-25] using NaBH4 in methanol in
excess, for 83% yield. The hydroxyl is esterified using Ojimas (3 lactam synthon [2-14]
(Figure 2-2) using the strong base sodium-hexamethyldisilazane for 87% yield based on
90% conversion. Removal of the triethylsilyl groups with hydrogen fluoride in pyridine
(HF-Pyr) completes the synthesis of taxol in 80% yield.
Other Synthetic Approaches
As previously mentioned Holtons group published a total synthesis of taxol in
early 1994 at about the same time as Nicolaou, but their approach was quite different,
with only a few reactions in common. Studies involving the fragmentation of bicyclic
epoxy alcohols, referred to as epoxy alcohol fragmentation, were the cornerstone of
their syntheses of bicyclo[5.3.1] systems, including the unnatural epimer of (+)-taxusin
[2-26], known as (-)-taxusin or enf-taxusin [2-27] (Figure 2-6).

21
[2-18] [2-19]
[2-20] [2-21]
Figure 2-4 : Nicolaous Taxane Ring Synthesis

22
[2-24]
[2-25]
Figure 2-5 : Nicolaous Final Synthetic Intermediates

23
Holton group Patchouline Oxide Fragmentation
[2-28] Wieland-Miescher ketone
Danishefsky group
[2-29] a-Pinene
Wender group
Figure 2-6 : Starting Points of Other Synthetic Strategies

24
Danishefskys total synthesis of baccatin III in 1996 (and hence, taxol); borrowed
extensively from the experiences of Ojima, Holton, Nicolaou and others. The Weiland-
Miescher ketone [2-28], available through catalytic asymmetric induction, allowed the
installation of all stereochemical requirements to reach baccatin III in a sequential
fashion. According to Danishefsky, Our synthesis, though arduous, involves no relays,
no resolutions, and no recourse to awkwardly available antipodes of the chiral pool
(Danishefsky etal. 1996).
Wenders group published a most concise synthesis involving a-pinene [2-29] for
construction of the ABC-tricyclic core of the taxanes (Wender & Rawlins 1992). Their
approach takes advantage of the tendency for C-7 to undergo facile aldol/reverse aldol
epimerization in taxol, allowing for aldol condensation under very mild conditions.
General Structural Features of Taxanes
The taxanes comprise a relatively large group of diterpenoid natural products
covering a variety of structural patterns. These are believed to arise from geranyl
geraniol [2-16], although the exact biosynthetic route has not been completely
elucidated. A brief discussion of the major structural variations of taxanes is relevant to
this work because many of these structures have been found in the compounds isolated
in this work. A number of different forms that the C-20 diterpene skeleton itself can
assume have been isolated. Next, the oxidation states, esterification patterns of the
hydroxyls, and presence or absence of basic or neutral side chains allow for the
extensive structural variation seen in these compounds.
The taxane skeleton is a specific diterpene structure, consists of 20 carbon
atoms arranged in a fused tricyclic system with the 6, 8 and 6 members in rings A, B and
C, respectively. The double bonds at 11/12 and 4/20 are part of the basic ring system,
although the latter may be modified by oxygenation to an epoxide or more commonly to

25
an oxetane. As in the case of the analogous steroids with the two methyl groups as part
of the skeleton, the taxanes have four methyl groups #16, 17, 18 and 19 as part of the
taxane ring system. Some examples of the taxane skeleton found in the various species
of Taxus are shown in Figures 2-1 and 2-2.
Oxygenation of the taxane ring has been observed to varying extents. The
minimum number being 4, distributed at 5, 9, 10 and 13, as seen in taxusin [2-26], In
general, oxygenation may occur at carbons 1, 2, 4, 5, 7, 9, 10 and 13. Instances have
been recorded where oxygenation was present at 14 (in place of 13), as well as part of
the methyl groups at 19 and 17.
Taxa-4(20):11-dienes
This is the most common structural type seen in the taxanes, with a C-4(20) and
a A11 double bond. These taxanes are generally referred to taxa-4(20);11 -dienes. The
alkaloidal Winterstein esters are included in this group as are many of the neutral
taxanes. The oxygen at C-9, if present, is usually seen as a secondary alcohol or as an
ester. The C-13 position in this group, likewise, exists as an alcohol, ester or oxidized to
a carbonyl to form an a,p-unsaturated carbonyl. Esterification at C-13 is usually limited
to an acetyl or a cinnamoyl, but the side chain (N-acyl phenyl isoserine); as found in
taxol, cephalomannine and others has not been reported in this subgroup so far. The 5
position is oxygenated with an a-hydroxyl, which might be free, or esterified by an acetic
acid, cinnamic acid or the Winterstein-acid. Some examples of these compounds with a
cinnamate ester function are described in Chapter 5.
4(20)-Epoxides
This group is relatively less frequent but examples with different substitution
patterns have been isolated. One variation comes from the presence or absence of
hydroxyl at C-1. Members of this subgroup also generally contain the 5-a-hydroxyl,

26
which is esterified in the same fashion as the dienes above to provide further variation.
An unusual example is the taxane with the C-9-nicotinoyl ester function, found in
Austrotaxus spicata Compton Taxaceae (Ettouati et at. 1988).
Oxetanes
This group is characterized by having an oxetane ring system involving the
carbons 4, 5 and 20. It may be divided into two subgroups based on whether they
contain the phenyl isoserine ester side-chain at C-13 or not. The former contains taxol
and all of the other compounds, which are active in the tubulin assay and hence are of
much importance. Division into two other subgroups is also possible in those without the
C-13 side chain, with one having a carbonyl at C-9 and with a hydroxyl or an esterified
hydroxyl at C-9.
The oxetane-containing taxanes are generally highly oxygenated and often have
oxygen at C-1, 2, 4, 5, 7, 9, 10, and 13. In some special instances, a hydroxyl has been
reported at C-19 (Fuji et at. 1993). The phenylisoserine ester side chain has been seen
in the form of at least three different amides that occur in nature. These are taxol, with
the N-benzoyl group, cephalomannine, with the N-tiglioyl group and taxol C, with the N-
hexanoyl group.
Taxol has a complex structure and knowing what features of this structure are
necessary for the activity is of utmost importance and this aspect has been studied using
the in vitro tubulin binding, and the cell culture assays and a summary of these data is
presented below (Samaranayake et ai. 1993).
Acylation of the 2' position of taxol does not destroy cytotoxicity but does stop
promotion of microtubule assembly. Bulky acyl groups reduce the activity in the cell
culture, thus suggesting that hydrolysis of the 2' position back to a free hydroxyl might be
required.

27
Substitution of the 7 position does not appear to significantly decrease the
activity. Taxanes with a 7p-0-xyloside moiety are comparably active in both assays
when compared to the respective aglycones. Similarly, epimerization at the 7-position
does not eliminate activity.
Hydrolysis of the 10-acyl function does not reduce the cytotoxicity significantly in
cell culture assays. As with other structural features, this point is being explored in the
more recent clinical trials in Europe with taxotere.
The importance of the oxetane ring for activity has been investigated through ring
opening via different Lewis acids including Meerwein's reagent (triethyloxonium
tetrafluoroborate); acetyl chloride, mesyl chloride and others. The product obtained form
the Meerwein's reagent had a primary alcohol at C-20 and secondary C-5-hydroxyl, but
no other changes compared to taxol. The activity normally seen with taxol in both
assays was lost with the opening of the oxetane ring. This suggests that the oxetane
ring is necessary for activity but leaves open questions regarding the effect of ring
contractions in ring A.
The properties of the C-13 hydroxyl mentioned above make attachment of a side
chain quite difficult. Protection of other free hydroxyls in both the side chain and taxane
skeleton are necessary, followed by selective deprotection after the side chain has been
attached. Taxotere and taxol have both been synthesized from this taxane and this is
currently the starting material for the production of both drugs.
Epimerization of the 7 hydroxyl from p to a via a retro-aldol mechanism allows
formation of an energetically favorable hydrogen bond with the 4-acetate carbonyl
oxygen. This epimerization is a concern in both taxane isolation and synthetic methods,
and necessitates the avoidance of acidic or basic conditions. Protection of this C-7 p-

28
hydroxyl with groups such as a chloroacetate avoids both epimerization and unwanted
reaction at this position.
Abeotaxanes
A number of taxanes in which the A-ring is isomerized to a 5-membered ring to
give a 5/7/6 instead of the 6/8/6 system have been isolated and these are termed
abeotaxanes. They are again divided into two groups into a) those with the 4/20
unsaturation and b) those with an oxetane ring at this location. We isolated the first
members of each of these groups in our work, e.g. brevifoliol (Chapter 3); and the
compounds isolated from the bark of T brevifolia described in Chapter 6. As indicated
earlier, treatment of taxol with acidic reagents can isomerize ring A to form such
compounds, although these compounds are naturally present in the extract and not
artifacts.

CHAPTER 3
TAXANE CONSTITUENTS OF THE NEEDLES OF TAXUS BREVIFOLIA
Taxol was originally isolated from the bark of the Pacific yew (Taxus brevifolia
Nutt., N.O. Taxaceae). As indicated in Chapter 1, during 1991-1993 there was a
reassessment of the use of the bark as the source. This concern resulted in an intense
search for alternative sources for taxol that are renewable, with sources such as the
needles of the yew tree instead of the bark. This laboratory was also involved in this
search and looked into the needles of three different yew species as a source for taxol;
T. brevifolia, T. x media Hicksii and T. floridana. The taxane composition of T. brevifolia
needles is the subject of this chapter.
Fractionation of the Needles of Taxus brevifolia
A quantity of 100 lbs. of the needles of T. brevifolia was obtained from a supplier
in Oregon. They were air-dried and extracted with methanol at room temperature and
the extract was concentrated under reduced pressure to a syrup. This was partitioned
between water and chloroform, and the organic layer concentrated to give a dark
greenish brown semi-solid, called extract solids, which represented about 5% of the dry
weight of the needles.
It was decided to follow the method successfully developed with the bark extract
for the fractionation of the extract solids, using preparative scale, reverse phase column
chromatography. Direct application of the crude chloroform extract of the needles onto a
C-18 bonded reverse phase silica column was accomplished as described in the
experimental section. After placing the extract-containing silica onto a 25% acetonitrile
29

30
in water column (1:4 ratio of loaded to clean silica); a step gradient of acetonitrile in
water mixtures was performed up to 60% acetonitrile.
Preliminary studies on the extract solids of the needles by TLC and analytical
HPLC showed that the sample contained somewhat minor amounts of taxol. A
predominant component that was slower moving than taxol in TLC gave a greenish-blue
colored spot when sprayed with 1 N sulfuric acid and heated on a hot plate (charring).
Likewise, in the analytical HPLC, this component appeared after 10-deacetyl baccatin III
as the major constituent judging from the peak heights, but before taxol and at least
several times more abundant.
The reverse phase column (C-18 bonded silica gel) on the needle extract
concentrate was started with 25% acetonitrile in water. The sample was carefully
prepared as a slurry (see experimental) and added to the column. The column was
developed using a step gradient of acetonitrile in water 30-60%. Fractions of suitable
volume were collected and monitored by absorbance at 275 nm, TLC and analytical
HPLC. Four regions were recognized in the elution profile of the column, based on the
UV absorbance (275 nm.); which contained the resolved constituents of the extract.
The early fractions contained components, which accounted for the bulk of the
UV absorbance of the sample. These appeared to be non-taxane phenolic compounds
with or without attached sugars. A description of these will be given in Chapter 7. The
first taxane component, which appeared at the 35-40% acetonitrile elution, was also the
major component. It was collected from the appropriate fractions, and after
concentration, obtained as a crystalline solid. Next, fractions from the 50% acetonitrile
elution contained taxol, which was obtained as a crystalline solid directly from the
fractions. Following this, the fractions from the 55-60% acetonitrile elution gave another
taxane component which gave a greenish blue spot on the TLC (after charring with
sulfuric acid) similar to the major constituent referred to above.

31
Brevitaxane A (Brevifoliol) [3-11
The major constituent, which was obtained in a yield of 0.2-0.25%, was named
brevitaxane A because the physical and spectral data indicated that it was a new taxane
compound (later renamed by others as brevifoliol, which will be used throughout this
dissertation). Elemental and FAB-MS analysis (MH+ 557) agreed with the molecular
formula of C31H40O9 (Balza et al. 1991).
An examination of the 1H NMR spectrum showed the presence of two acetyl
groups (signals at 5 1.76 and 5 2.07); and a benzoate group {5 7.88 (d); 5 7.43 (t) and 8
7.56 (t)}. The spectrum also gave evidence for the presence of a (4/20) exocyclic double
bond (two characteristic broad singlets at 8 4.82 (H-20A) and 8 5.20 (H-20B) and signals
at 8 112.1 (C-20) and 8 149.0 (C-4) in the 13C NMR spectrum.
Very little information on the various types of taxane structures that are known
now was available at that point in time (1991) and even less on their diagnostic spectral
characteristics. Based on analogous taxanes and the evidence outlined above it was
postulated that this major constituent had the relatively common 4/20,11-taxadiene type
skeleton. The presence of an exocyclic 4/20 double bond and absence of an oxetane
ring supported our initial assumptions. The next step was to determine the positions of
the various substituents in the molecule in order to elucidate the complete structure.
Most of the structural elucidations of taxanes at the time were based on
degradative studies. It was decided to follow this lead in establishing the presence of
the various functionalities as well as their location in brevifoliol, by actual reactions
and/or derivatizations, supplemented by spectral methods.

Figure 3-1 : Proton NMR Spectrum of Brevifoliol

33
Hydroxyl Functionalities
i) Acetylation: To determine the number and positions of all hydroxyls in the
molecule, the compound was subjected to acetylation. Two products were obtained
under mild conditions (20 C, 15 min). These two were separated by chromatography
and both obtained as crystalline solids. One was shown to be a monoacetate and the
other a diacetate.
Table 3-1 : Proton NMR Spectra of Brevifoliol and Brevifoliol Acetates
Position
Brevifoliol
Brevifoliol
Brevifoliol
Brevifoliol
Brevifoliol
(J in Hz)
5-Ac
13 Ac
5,13 Ac
5,13,15 Ac
[3-1]
[3-2]
[3-3]
[3-4]
[3-5}
2
1.49 cm
1.46 cm
1.47 cm
1.46 brd (13)
1.53 br d(13)
2.36 dd (9,13)
2.40 dd(9, 13)
2.42
2.41 dd(9, 13)
2.65 dd(9,
3
2.78 d (9)
2.76 brd (9)
2.91 d (9)
2.72 brd (9)
2.71 brd (9)
5
4.45 brs
5.37
4.37 brs
5.39 br s
5.38 dd (4, 2)
6
1.86 cm
1.88 cm
1.85 cm
1.90 cm
1.87 cm
2.02 cm
2.0 cm
1.99 cm
2.00 cm
2.0 cm
7
5.56 dd (5,11)
5.62 dd (5, 11)
5.66
5.61 dd (5,
5.63 dd(5,
9
6.05 br
6.03 br d(10.6)
6.07 d
6.09 br
5.8 d (10.8)
10
6.53 d (10.6)
6.63 d (10.6)
6.66 d
6.65 d (10.6)
6.64 d (10.8)
13
4.38 t (7.5)
4.53 brt (7.2)
5.46 br s
5.54 brt (7.2)
5.61 t (6.9)
14
1.29
1.22 dd *
1.32 cm
1.25 dd *
1.25 dd *
2.46
2.42 dd *
2.51 cm
2.51 dd *
2.62 dd *
16
1.05 s
1.03 s
1.09 s
1.11 s
1.63 s
17
1.35 s
1.33 s
1.35 s
1.35 s
1.71 s
18
2.01 s
2.06 s
2.02 s
2.03 s
1.96 s
19
0.90 s
0.91 s
0.89 br s
0.92 s
0.92 s
20 A
4.82 brs
4.90 br s
4.80 br s
4.92 brs
4.89 br s
20 B
5.20 brs
5.28 br s
5.15 br s
5.28 br s
5.29 brs
o-Ph1
7.88 d( 7.5)
7.87 d (7.5)
7.87 d (7.5)
7.87 d (7.5)
7.84 d (7.5 )
m-Ph1
7.43 t( 7.5)
7.43 t (7.5)
7.44 t (7.5)
7.44 t (7.5)
7.42 t ( 7.5 )
p-Ph1
7.56 t ( 7.5)
7.55 t (7.5)
7.56 t (7.5)
7.56 t (7.5)
7.53 t( 7.5)
-
1.76 s
1.76 s
1.76 s
1.75 s
1.77 s
2.07 s
2 06 s
2.05 s
2.02 s, 2.07 s
2.02 s, 2.08 s
2.13 s
2.06 s
2.08 s
2.09 s, 2.11 s
NMR were recorded at 600 MHz in CDCI3on a Varian Unity 600 instrument at
ambient temperature. Chemical shifts 8 (ppm) are reported with TMS as internal
standard.

34
Table 3-2 : Carbon NMR Spectra of Brevifoliol and Brevifoliol Acetates
Carbon
Number
[3-1]
Brevifoliol
[3-2]
5-Ac
[3-3]
13-Ac
[3-4]
5,13-Di-Ac
[3-5]
1,5,13-Tri-Ac
1
62.4
63.0
63.4
63.0
63.3
2
29.1
29.2
29.4
29.1
28.3
3
37.9
38.8
37.6
38.8
38.9
4
149.0
145.4
147.4
145.2
145.1
5
72.4
74.1
72.7
74.1
73.9.
6
36.0
33.9
36.1
33.9
34.0
7
70.1
69.7
69.8
69.6
69.7
8
45.0
44.8
45.2
44.8
45.0
9
77.1
77.9
79.8
79.3
78.9
10
70.2
70.7
70.3
69.8
68.4
11
133.9
134.0
136.5
136.4
136.5
12
151.5
151.1
150.5
147.3
148.2
13
76.7
76.9
77.8
76.9
78.0
14
47.3
47.1
44.2
44.1
43.3
15
75.9
75.6
75.6
75.6
87.2
16
26.9
27.0
27.0
27.0
23.1
17
24.8
24.8
25.0
24.8
21.8
18
12.0
11.8
12.1
11.9
11.9
19
12.9
12.9
12.9
12.9
13.5
20
112.0
114.1
111.5
114.3
114.3
CO-C6H5
164.3
164.1
164.2
164.1
165.0
Bz-ipso
129.3
129.2
129.4
129.1
129.9
Bz-ortho
129.4
129.4
129.5
129.5
129.3
Bz-meta
128.7
128.7
128.7
128.8
128.4
Bz-para
133.2
133.3
133.2
133.4
133.0
CO-CH3
20.7
20.8
20.7
20.7
20.8
21.4
21.4
21.4
21.4
21.4
21.2
21.1
21.2
21.3
21.0
21.0
21.7
COCH3
169.9
169.9 (X2)
169.9
169.91
169.9
170.5
170.2
170.8
170.5
170.5
169.7
169.6
169.6
169.9
171.0
169.5
13C NMR spectra were recorded at 150 MHz in CDCI3on a Varian Unity 600
spectrometer at ambient temperature. Chemical shifts 5 (ppm) are reported with TMS as
internal standard.

35
[3-6] 13-Ketone
[3-8] 4,20-Epoxide
R-i R? R3
[3-1] H H H
[3-2] Ac H H
[3-3] H Ac H
[3-4] Ac Ac H
[3-5] Ac Ac Ac
[3-11] Hydrolysate, R1 = R2 = H
[3-15] Debenzoyl, R1 = R2 = Ac
Figure 3-2 : Brevifoliol and Reaction Products

36
Appropriate conditions under which each of these could be obtained as exclusive
products were developed. At room temperature in acetic anhydride for 1-2 minutes
before quenching the reaction, the monoacetate was the major product (>90%).
Likewise, at 80 0 C for 30 min. the product was the diacetate.
The 1H NMR spectral data for the monoacetate showed that the signal at 5 4.45
(br s) shifted to 8 5.37 (dd, J=4.2, 2.4 Hz); indicating that acetylation took place at the 5-
OH, as shown in [3-2], In the diacetate, besides this shift for the 5-OAc, the signal at 8
4.38 (t, 7.5 Hz) shifted to 8 5.54 (br t, 7.2 Hz); thus showing that the second acetate was
located at C-13 [3-4], A naturally occurring brevifoliol 13-acetate [3-3] was isolated and
1,5,13-brevifoliol triacetate [3-5] produced in this lab will be discussed in Chapter 4.
ii) Oxidation: Brevifoliol was readily oxidized by manganese dioxide (Mn02) in
refluxing benzene to yield a ketone product. In the 1H NMR spectrum, a major change
was the absence of the triplet at 8 4.38 due to the C-13 proton, thus showing that the
oxidation took place at the 13-OH [3-6], Further evidence was seen by the shift of the
signals for the C-14 protons from their normal positions at 8 1.29 (dd, 14.0, 7.6 Hz) and 8
2.46 (dd, 14.0, 7.6 Hz) to 5 2.32 (d, 19 Hz, H-14cc) and 8 2.48 (d, 19 Hz, H-14p). When
brevifoliol was oxidized by Jones reagent, the same 13-keto brevifoliol seen with Mn02
initially formed [3-6], With time the initial product gradually disappeared, giving rise to a
faster moving product. This second oxidation product was shown to be the result of an
unusual reaction described in Chapter 4.
4/20 Unsaturation
i) Hydrogenation: When hydrogenated in the presence of 5% Pd/carbon,
brevifoliol gave the dihydro derivative [3-7], In its 1H NMR spectrum, the characteristic
signals at 8 4.82 and 8 5.20 due to the C-20 protons were absent and a new methyl

37
doublet and a new methine proton appeared. In the 13C NMR spectrum the
characteristic signals from the exocyclic 4/20 double bond were absent, accompanied by
the appearance of new methyl and methine signals.
ii) Epoxidation: Brevifoliol was heated in dichloromethane with meta-chloro
peroxybenzoic acid (MCPBA); whereby it underwent oxidation to yield the epoxide [3-8],
a crystalline compound.
¡ii) Ozonization: Brevifoliol has two double bonds, one at the 11/12 position and
the other at the 4/20 position. Of these, the former is tetra-substituted, while the latter is
of an exocyclic methylene type. No information was available in the literature regarding
the reactivity of the taxane skeleton to indicate whether one or both double bonds would
be cleaved by ozonolysis. In the present work, ozonization was carried out in a mixture
of methanol and dichloromethane -70 0 C. After the disappearance of the starting
material, the ozonide was decomposed with dimethyl sulfide and the products isolated
by chromatography. Two major products were separated. The first was the same as the
epoxide [3-8] obtained by reaction with MCPBA. The second was the expected
ozonolysis product in which the 4/20 double bond was cleaved to form the ketone [3-9],
iv) Formation of a diol: As one of the characteristic reactions of an ethylenic
function, oxidation by osmium tetroxide was attempted with brevifoliol. The reaction
proceeded smoothly to give a diol [3-10],
Number and Nature of the Oxygen Substitution
From the preceding discussion it is evident that brevifoliol has two free hydroxyls,
two acetoxyls and one benzoyloxy functions. Plowever, in the 13C NMR spectrum of
brevifoliol, the number of oxygen substituted carbons was six: 5 70.1, 5 70.2, 8 72.4, 5
75.9, 5 76.7 and 8 77.2. To determine if one of the six is a different type of an ester, or a
tertiary hydroxyl, brevifoliol was subjected to saponification in alcoholic KOFI to yield the

38
hexaol [3-11], obtained as a crystalline solid. This was then acetylated to a crystalline
acetate [3-12], The 1H- and 13C NMR spectra of [3-12] showed the presence of 5
acetates (1H: 5 21.8, 8 21.7, 5 21.4, 8 21.3, 8 21.0, 8 20,8; and 13C 8 171.0, 8 170.4, 8
169.8, 8 169.6 and 8 169.5); which suggested that brevifoliol contained a tertiary
hydroxyl.
In the conventional taxane skeleton, a tertiary hydroxyl is often present at C-1,
with the other hydroxyls (or esters) at C-2, C-5, C-7, C-9, C-10 and C-13. Thus, with
brevifoliol having five oxygen substituents, one of these positions must be without
attached oxygen. Thus, it would be important to know which of these positions does not
have an oxygen substituent. For this reason, the hexaol [3-11] was subjected to
oxidation by periodate. If there were two pairs of vicinal hydroxyls, e.g. 1,2 and 9,10, the
hexaol will be cleaved in such a way as to give smaller molecules which represent the A
and C rings. If there is only one such pair, the reaction will produce a product with all of
its carbons intact. The hexaol underwent oxidation readily to form a dialdehyde [3-13]
without losing any carbon atoms found in original carbon skeleton. Unaware of the
unusual A ring structure, it was presumed that the presence of a tertiary hydroxyl at C-1
precluded the presence of oxygen substitution at C-2. Additional evidence for a
methylene carbon at C-2 was found in the COSY spectrum from the chemical shifts in
the H-3p-H-2a-H-2p isolated spin system.
Thus, brevifoliol has two hydroxyls at 5 and 13. Locating the benzoate group at
one of the three choices, 7, 9, or 10 will elucidate the structure. At this point, brevifoliol
was required in microbial and fungal biotransformation project in our laboratory. In order
to produce an antiseptic sample an aqueous alcoholic solution was sterilized in a steam
autoclave at 125 0 C, 20 atm., to see if it is stable. It was found that the compound
underwent degradation to give two or three products.

39
[3-14] Fteriodate Oxidation Osazone Product
Figure 3-3 : Brevifoliol Hexaol Reaction Products
The major component of this mixture was found to be debenzoyl brevifoliol
(Figure 3-1 [3-15]). Of the three possible locations, 7, 9, and 10 for the benzoate, only
10 Is allylic and hence the ester at this position is more likely to be labile. Taxol with the
benzoate at C-2 is completely stable to heat and pressure for hours. This evidence,
along with chemical shift arguments concerning the effect of acetylation versus
benzoylatlon, led us to place the benzoate at C-10.
This group presented the isolation and the structural elucidation of brevitaxane A
at the International Research Congress on Natural Products held in Chicago, IL in July

40
1991. Balza et al. (1991) published the isolation of a new compound at about that same
time from the needles of T. brevifolia, which they named brevifoliol, and an assignment
of its structure as shown in [3-16], The compound appeared to be similar to, if not the
same as, brevitaxane A, that was isolated from the needles at the University of Florida.
The structure proposed by Balza et at differed from that of brevitaxane A, with the
benzoate group being placed at C-7 instead of at C-10.
That same year (1992); the isolation of taxchinin A was described (Fuji et al.
1992); which was later shown to be 2-acetoxy-brevifoliol. Fuji correctly assigned the
structure with a 5-membered ring A, on the basis of NMR spectral data. The authors who
isolated brevifoliol and assigned structure with the 7-benzoate (Chu et al. 1993)
published a revised structure for brevifoliol, in which the benzoate was moved to C-10,
from C-7, but with the skeleton of a conventional taxane.
During 1993, two other publications appeared, one from Georg et al. (1993).
and the other from Appendino et al. (1993) reexamining the NMR spectral data of
brevifoliol, and arriving at the structure in which the A-ring was 5-membered. Later that
year, Chu et al. (1993); on the basis of x-ray crystallographic data, revised the structure
of brevifoliol again to the presently accepted structure.
Due to the intense competition in taxol research, we began a detailed
examination and analysis of the NMR spectral data using the 13C NMR, NOESY,
HETCOR and other spectral methods to determine if the rearranged (5/7/6) skeleton
might be distinguishable from the spectrum of a taxane with a conventional (6/8/6)
skeleton. The following is an analysis of the spectral data of brevifoliol.
The carbonyl signal in the 13C NMR spectrum at 6 164.3 indicated the presence
of one benzoate, and signals at 8 169.9, 5 170.5, likewise, indicated that 2 acetate ester
groups were present. Further support for the benzoate was obtained by the four

41
aromatic signals between 8 128.7 and 5 133.2 (see tables 3-1 and 3-2); and for the two
acetates, by the methyl signals at 8 20.7 and 8 21.4. Analysis of the 1H NMR and 1H
COSY and 1H,13C Heteronuclear Correlation (HETCOR) experiments also gave
additional support for the presence of the acetates with signals at 5 1.76 s and 8 2.07 s,
as well as benzoate signals at 8 7.88 d (ortho); 8 7.43 t (meta); and 8 7.56 t (para). Next,
evidence for the presence of the normally present (11/12) taxane double bond could be
seen in the carbon spectrum by the signals at 5 133.9 (C-11) and 8 151.5 (C-12);.
Similarly, the existence of a (4/20) exocyclic double bond could be seen by the signal 5
149.0 (C-4) and 8 112.1 for (C-20). In the 1H NMR spectrum the exocyclic 4/20 double
bond is also indicated by the two characteristic broad singlets seen at 8 4.82 (H-20A)
and 8 5.20 (H-20B).
In the 1H COSY experiment weak but definite interactions between the singlet at
8 4.82 (C-2Qa) with both the H-3(3 doublet at 8 2.78 (9 Hz.) and the H-2a doublet of
doublets at 8 2.36 (9, 13 Hz.) supported the assignments given for the methylene
protons. (The designations for the C-20 protons are A and B, since a and p do not have
the conventional meaning system and could be confusing). Along with the interaction
between H-2a and H-2p the first isolated spin system in the 1H spectrum was
established and the relative geometry of the protons.
The region between 8 62.4 and 8 77.1 in the 13C spectrum carbons with hydroxyl
or ester oxygen attached to oxygens, and these signals could be further defined in the
DEPT experiment (Distortionless Enhancement with Polarization Transfer, NMR) as
primary, secondary, tertiary and quaternary carbons. The spectrum showed two
quaternary carbon signals and five oxymethine carbon signals. Since the presence of
only six signals was expected based on the proposed formula, the quaternary signal at
8 62.4 was Intriguing even from the start of the spectral examination. In taxol with its C-9

42
carbonyl, the C-8 signal appears near 5 58, so the signal at 8 62.4 immediately raised
questions about the true structure of brevifoliol. This signal did not fit the normal
chemical shift pattern of any naturally occurring taxanes known at that time. In the
absence of a carbonyl group at C-9, the C-8 carbon usually falls in the region of 8 40-50
ppm.
Unable to satisfactorily explain this unusual peak position, the Chemistry
Department was contacted about crystallographic services. X-ray crystallographic
analysis was performed by Dr. K. A. Abboud on the 5-monoacetate [3-2], Surprisingly,
the presence of an unusual 5/7/6 ring system was evident, where the normal 6-
membered A ring of the conventional taxane system was rearranged to form a 5-
membered ring with the carbons 15, 16 and 17 moved out of the ring system to form a
hydroxy isopropyl group at C-1. Since the x-ray structure was obtained on brevifoliol-5-
acetate, it was important to establish whether brevifoliol itself had this rearranged taxane
skeleton, or if the rearrangement could have occurred during the acetylation.
This structure represented a departure from the existing naturally occurring
taxane structures available at that time, previously seen only as a product of
rearrangement under strongly acidic conditions (Samaranayake et al. 1990).
Crystallography of the original compound was not done because it failed to yield
adequate crystals for analysis without prior acetylation. This made it necessary to
determine whether this new ring structure was naturally occurring, or formed during the
acetylation.
In one such ring contraction, taxol underwent rearrangement of the A-ring,
accompanied by dehydration, to produce an isopropenyl group at C-1, as well as other
changes such as the opening of the oxetane ring. Since the 13C NMR spectra of both
brevifoliol and its monoacetate showed these signals at 8 62.4 assigned to the C-1

43
carbon and the one at 5 75.9 assigned to the quaternary C-15 containing tertiary
hydroxyl, it appeared unlikely that such a rearrangement took place during the
acetylation. HETCOR and APT experiments corroborated these conclusions, thereby
agreeing with the structure determined by the x-ray crystallographic method.
Further analysis of the 1H COSY spectrum revealed an isolated spin system of
two doublets due to H-9p at 5 6.05 and H-10a at 8 6.53, with a pseudo-axial orientation
indicated by the degree of splitting (J=10.6 Hz); and significant broadening of the signal
at 5 6.05. Some amount of the deshielding of H-10a relative to H-9p was expected, due
to the adjacent double bond, which makes the C-10 position allylic. The presence of a
benzoate at this position would be expected to cause a further downfield shift based on
analogous compounds already known (Chu et al. 1992). With a thorough analysis of the
1H NMR and 1H COSY spectra, the signal at 5 4.38 (t, 7.6 Hz) was assigned to the H-
13p proton, which coupled strongly with H-14p at 8 2.46 (dd, 14.0,7.6 Hz.); as well with
H-14a at 5 1.25 (dd, 14.0, 7.6 Hz). Weak long range coupling to the C-18 methyl
protons at 8 2.01 was also evident, as the slight broadening of this peak is generally
attributed to this long range coupling in other taxanes.
The isolated spin system of H-5p, H-6a, H-6p and H-7a is easily identified in
most taxanes, with a tendency to show a sharp multiplet for H-7a and broader, poorly
resolved splitting for H-5p, especially if H-5p is not esterified (Della Casa de Marcano &
Halsall, 1970; Rao et al. 1995). The H-5p broad singlet at 8 4.45 interacts with the H-
6a multiplet at 8 1.86, which interacts with the H-6p multiplet at 5 2.02, which in turn
interacts with the H-7a signal at 8 5.56 (dd, 5,11 Hz.). In many cases esterification of a
hydroxyl causes a deshielding effect on the related proton of about 1 ppm. The
chemical shifts and splitting patterns indicated that the acetate groups were at C-7 and
C-9, with the benzoate at C-10.

44
The remaining carbons are the 4 methyl groups usually seen in taxanes on C-15
(methyl 16 and methyl 17); at C-12 (methyl 18) and at C-8 (methyl 19). The methyl
group located on the 11/12 double bond (methyl 18) is often quite deshielded in the
proton spectrum (§2.01, s) but shielded in the carbon spectrum (5 12.0). This usually
aids in its assignment along with further evidence from Heteronuclear NMR experiments.
Methyl 19 is usually shielded in both the 1H (5 0.90, s) and 13C (5 12.9, q) spectra, as
seen here.
This class of compounds commonly referred to as 11(1-15)-a6eo-taxanes or
occasionally A-nortaxanes. Many compounds of this type are now known, some
containing the 4/20 unsaturation as in brevifoliol and others with a 4/20 oxetane structure
as seen in 11(1 >15) abeo baccatin VI.
Experimental
Extraction of the Needles of Taxus brevifolia
The needles obtained from a supplier (Mr. Patrick Connolly, Yew Wood
Industries, 6928 North Interstate Avenue, Portland, OR 97217) were air-dried for one
week. The dried needles (20 Kg) were extracted by immersing in methanol at room
temperature. After two days, the extract was drained, concentrated under reduced
pressure at temperatures below 35 0 C. The recovered methanol was reused for a
second extraction, which was processed the same way. After two more extractions, the
combined concentrate was freed from some more of the methanol to obtain a dark green
syrup.
The above syrup was partitioned between water (10 gallons) and chloroform (10
gallons). The organic layer was separated and the extraction carried out twice more

45
using 5 and 3 gallons respectively. The combined chloroform extract was concentrated
under reduced pressure to reach a dark green semi-solid stage (800-900g).
Reverse Phase Column Chromatography:
The column used was a threaded glass column of the Mitchell-Miller type (2.5 x
24) with the appropriate fittings, purchased from Ace Glass Co., Vineland, NJ suitable
for low pressure liquid chromatography. A slurry of the C-18 bonded silica (800 g)
(Spherisorb, 15-35 micron diameter) purchased from, Phase Separations Inc., Norwalk,
CT) in methanol was poured into the column, which was run under a gentle pressure by
using a metering pump (Fisher/Eldex) until an adequately packed bed was obtained.
The column was then equilibrated with 25% acetonitrile in water, to prepare for the
addition of the sample.
The extract solids (200 g) was dissolved in acetonitrile (400 ml) by warming to
make sure that no lumps remained. To this was added approximately 200 g equivalent
of the equilibrated resin (about 20% of the column packing) with stirring. As the stirring
continued, the slurry was diluted with 25% acetonitrile in water (500 ml); followed by
water to make up a total volume of approximately 2 L. The stirring was continued with
occasional warming to 50-60 0 C for about 15 min. At this point, a sample of the slurry
taken into a test tube, showed that the silica settled readily to give a clear supernatant
and no green precipitate or oily material was present. The slurry was then filtered using
light suction and the solid (silica with the sample) re-slurried using part of the filtrate and
the thick slurry added to the column. The rest of the clear supernatant was then pumped
on to the top of the column using the metering pump. From time to time, the column
feed was checked to see that it remained clear, and if not, to either warm briefly or add
minimal amounts of acetonitrile to it until it became clear, so as to prevent precipitate
from appearing and blocking the pump.

46
After the sample addition was completed, fresh 25% acetonitrile/ water was
passed through, followed by the step gradient of acetonitrile/ water (30, 35, 40, 45, 50
and 60%) was used. Fractions (200 ml) were collected and monitored by UV
absorbance (at 275 nm), TLC and analytical HPLC. The change to the next
concentration of solvent was determined by the results of monitoring the fractions. For
example, when the absorbance values rose as a result of the previous change, the
solvent was continued until a definite trend to lower values was seen. Similarly, when
the TLC showed the trend towards decreasing intensity of the major spot, and no new
spot had shown a tendency to increase, the solvent was changed to the next level. In
general, 2-3 multiples of the hold-up volumes of the column were used.
After the elution with the 60% solvent was completed, the column was washed
with 100% methanol, followed by a mixture of methanol/ ethyl acetate/ ligroin which
stripped the column of the chlorophylls, waxes and other lipid soluble components. After
this solvent, washing with methanol and equilibration with 25% acetonitrile/ water made
the column ready for another run.
After the monitoring, fractions with low UV-absorbance values were combined
and concentrated into groups, based on the TLC data. Those fractions with relatively
stronger UV readings were let stand at room temperature for 3-5 days, whereby crystals
appeared in several sections of the fraction sequence. These crude crystals were
filtered, dried and purified further either by recrystallization or using a small silica column
(normal phase).
Brevifoliol 3-11
The fractions containing this component gave crystals but only a small portion
was obtained in this form. Hence, after filtration of the crude crystals, the filtrate was
concentrated to dryness and the solid taken up in dichloromethane and passed through

47
a column of normal phase silica, using a ratio of 3-5 g of silica per gram of the solid. The
effluent and washes which contained the compound were combined, concentrated to
dryness and the solid crystallized from a mixture of acetone and ligroin to obtain
brevifoliol as a colorless crystalline solid, yield from 200 g of the chloroform extract
solids, 12 g, 0.25% of the dried needles. [a]D23 -2 7 0 (CHCI3; c 1.03); m.p. 220-222 0 C
(lit. 200-203 C [Balza etaL 1991]44);
FAB-MS m/z: 557 [MH]+, 539 [MH-H20]+, 479 [MH-AcOF)]+, 435 [MH-PhCOzH]+,
417 [MH-PhC02H-H20]+, 375 [MH-PhC02H-AcOH]+, IR (KBr) vmax cm'1: 3370, 1740,
1650, 1600, 1585, 1450, 1370, 1265, 1180. UV 7 max log s 3.01 (269 nm); log s 4.32 (223
nm).
1H NMR (600 MHz, CHCI3, 5) Table 3-1: 0.90, s (H-19); 1.05, s (H-16); 1.30 (dd,
J=7.2, 13.8 Hz, H-14a); 1.35, s (H-17); 1.50 (d, J=14.1 Hz, H-2a); 1.76 (s, methyl, 9-
acetate); 1.80 (m, H-6a ); 2.0 (m, H-6p); 2.01 (s, H-18); 2.07 (s, 7-acetate methyl); 2.36
(dd, J=14.1, 9.6 Hz, H-2p); 2.46 (dd, J=7.2, 13.8 Hz, H-14p); 2.67, br s (C-15 OH,
exchangeable with D20); 2.77 (br d, J=9 Hz, H-3a); 4.38 (t, J=7.2 Hz, H-13p); 4.43 (br s,
H-5p); 4.82, s (H-20 A); 5.18, s (H-20 B); 5.57 (dd, J=4.8, 11.4 Hz, H-7a); 6.05 (poorly
resolved br d, J=10.5 Hz, H-9a); 6.53 (d, J=10.5 Hz, H-10p); 7.43 (t, J=7.8 Hz, H-Bz-
meta); 7.56 (t, J=7.8 Hz, H-Bz-para); 7.87 (d, J=7.8 Hz, H-Bz-ortfro).
13C NMR (CDCI3i 600 MHz, 5) Table 3-2: 12.0 (C-18 methyl, q); 12.9 (C-19
methyl, q); 20.7 (7-0 acetate methyl, q) ; 21.4 (9-0 acetate methyl, q); 24.8 (C-17
methyl, q); 26.9 (C-16 methyl, q); 29.1 (C-2, t); 36.0 (C-6, t); 37.9 (C-3, d); 45.0 (C-8, s);
47.3 (C-14, dd); 62.4 (C-1, s); 70.3 (C-7, d); 70.9 (C-10, d); 72.4 (C-5, d); 75.9 (C-15, s);
76.7 (C-13, d); 77.1 (C-9, d); 112.0 (C-20, t); 128.7 (C-Bz-mefa, d); 129.3 (C-Bz-/pso, s);
129.4(C-Bz-ort/?o, d); 133.3 (C-Bz-para, d); 133.9 (C-12, s); 149.0 (C-4, s); 151.5 (C-11,
s); 164.3 (CO-Ph, s); 169.9 (CO-Acetate, s); 170.5 (CO-Acetate, s).

48
Analysis calculated for C31 H40O9: C, 66.89; H, 7.24. Found: C, 67.12; H, 7.35;
Brevifoliol-5-Monoacetate 3-21
A mixture of brevlfoliol (0.2 g); acetic anhydride (2 ml) and pyridine (0.5 ml) was
stirred at room temperature for 2-3 min. Water was added to decompose the reagent,
and the solid filtered after 15 min. The solid was crystallized from a mixture of acetone
and ligroin to obtain the mono acetate as a colorless crystalline solid, yield, 0.18 g;
m.p.224-226 C;
1H NMR (CDCIa, 600 MHz, 6) Table 3-1: 0.91, s (H-19); 1.02, s (H-16); 1.22 (dd,
J=7.2, 13.8 Hz, H-14a); 1.33, s (H-17); 1.46 (d, J=14.1 Hz, H-2a); 1.76 (s, 9-0 acetate
methyl); 1.88 m, 2.0 m (H-6); 2.06 x 2, s (methyl-18, 5-0 acetate methyl); 2.08 (s, 7-0
acetate methyl); 2.40 (dt, J=14.1, 9.6 Hz, H-2P); 2.42 (dd, J=7.2, 13.8 Hz, H-14p); 2.75
(d, J=9 Hz, H-3a); 2.83, br s (C-15 OH, exchangeable with D20); 4.53 (t, J=7.2 Hz, H-
13p); 4.90, s (H-20 A); 5.28, s (H-20 B); 5.37 (br s, J= H-5p); 5.65 (dd, J=4.8, 11.4 Hz,
H-7a); 6.02 (poorly resolved br d, J=10.5 Hz, H-9a); 6.63 (d, J=10.5 Hz, H-10p); 7.43 (t,
J=7.8 Hz, H-Ph-mefa); 7.56 (t, J=7.8 Hz, H-Ph-para); 7.87 (d, J=7.8 Hz, H-Ph-ortho).
13C NMR (CDCI3, 600 MHz, 5) Table 3-2: 11.8 (C-18 methyl, q); 12.9 (C-19
methyl, q); 20.8 (7-0 acetate methyl, q); 21.2 (5-0 acetate methyl, q); 21.4 (9-0 acetate
methyl, q); 24.8 (C-17 methyl, q); 27.0 (C-16 methyl, q); 29.2 (C-2, t); 33.9 (C-6, t); 38.8
(C-3, d); 44.8 (C-8, s); 47.1 (C-14, dd); 63.0 (C-1, s); 69.7 (C-7, d); 70.7 (C-10, d); 74.1
(C-5, d); 75.6 (C-15, s); 76.9 (C-13, d); 77.9 (C-9, d); 114.0 (C-20, t); 128.7 (C-Ph-mefa,
d); 129.2 (C-Ph-ipso, s); 129.4(C-Ph-orfbo, d); 133.3 (C-Ph-para, d); 134.0 (C-11, s);
145.2 (C-4, s); 151.1 (C-12, s); 164.1 (CO-Ph, s); 169.9 X 2(CO-Acetate, s); 170.2 (CO-
Acetate, s).
Analysis calculated for C 33H 42O10: C, 66.20; H, 7.07. Found: C, 66.38; H, 7.19.

49
Brevifoliol-5,13-Diacetate 3-41
The above reaction was repeated, except that it was heated at 80-90 C (water
bath) for 30 min. After cooling, water was added and the solid filtered after 10 min. The
solid was crystallized from acetone/ ligroin to give the diacetate as a colorless crystalline
solid, yield, 0.2 g; m.p.241-243C;
1H NMR (CDCI3, 600 MHz, 5) Table 3-1: 0.92, s (H-19); 1.11, s (H-16); 1.25 (dd,
J=7.2, 13.8 Hz, H-14a); 1.35, s (H-17); 1.46 (d, J=14.1 Hz, H-2a); 1.75 (s, 9-0 acetate
methyl); 1.90 (m, H-6a) 2.0 (m, H-6p); 2.02 (s, 5-0 acetate methyl); 2.03 (s, 13-0
acetate methyl); 2.07 (s, 18 methyl); 2.08 (s, 7-0 acetate methyl); 2.41 (dd, J=14.1, 9.6
Hz, H-2p); 2.51 (dd, J=7.2, 13.8 Hz, H-14P); 2.72 (d, J=9 Hz, H-3a); 2.74, br s (C-15 OH,
exchangeable with D20); 4.92, s (H-20 A); 5.28, s (H-20 B); 5.39 (br s, J= H-5p); 5.54 (t,
J=7.2 Hz, H-13P); 5.61 (dd, J=4.8, 11.4 Hz, H-7a); 6.09 (poorly resolved br d, J=10.5 Hz,
H-9a); 6.65 (d, J=10.5 Hz, H-1 Op); 7.43 (t, J=7.8 Hz, H-Ph-mefe); 7.56 (t, J=7.8 Hz, H-
Ph-para); 7.87 (d, J=7.8 Hz, H-Ph-ortho).
13C NMR (CDCI3, 600 MHz, 5) Table 3-2: 11.9 (C-18 methyl, q); 12.9 (C-19
methyl, q); 20.7 (7-0 acetate methyl, q); 21.0 (13-0 acetate methyl, q); 21.2 (5-0
acetate methyl, q); 21.4 (9-0 acetate methyl, q); 24.8 (C-17 methyl, q); 27.0 (C-16
methyl, q); 29.1 (C-2, t); 33.9 (C-6, t); 38.8 (C-3, d); 44.8 (C-8, s); 44.1 (C-14, dd); 63.0
(C-1, s); 69.6 (C-7, d); 69.8 (C-10, d); 74.1 (C-5, d); 75.6 (C-15, s); 76.9 (C-13, d); 79.3
(C-9, d); 114.3 (C-20, t); 128.8 (C-Ph-mefa, d); 129.1 (C-Ph-ipso, s); 129.5(C-Ph-or#)0,
d); 133.4 (C-Ph-para, d); 136.4 (C-11, s); 145.2 (C-4, s); 147.3 (C-12, s); 164.1 (CO-Ph,
s); 169.6 (CO-Acetate, s); 169.9 (CO-Acetate, s); 169.91 (CO-Acetate, s); 170.5 (CO-
Acetate, s).
Analysis calculated for C 35H 44On: C, 65.61; H, 6.92. Found: C, 65.68; H, 6.99.

50
Brevifoliol-13-Ketone (3-61
A solution of brevifoliol (0.2 g) in benzene was treated with Mn02 (manganese
dioxide, 1 g, Fisher Scientific) and the mixture heated under reflux for 2 hours, at which
time, the starting material was consumed and a slightly faster moving product was
formed. The mixture was filtered, concentrated and applied to a small silica column (15
g) in dichloromethane. Elution with 1% acetone in dichloromethane gave the major
product which was recovered by concentration as a colorless powder, yield, 0.12g. The
1H- and 13C MMR spectra of this faster moving product were quite poorly resolved and
only gave usable results at temperatures below -10 0 C. Recrystallization and further
chromatography failed to improve this situation, and low temperature NMR experiments
indicated that a rotameric equilibrium was responsible for the poor resolution seen in
these spectra.
1H NMR (CDCI3i 600 MHz, -40 C, 5) major rotamer: 0.92, s (H-19); 0.98, s (H-
16); 1.35, s (H-17); 1.46 (d, J=14.1 Hz, H-2a); 1.78 (s, 9-0 acetate methyl); 1.90 (m, H-
6a) 2.0 (m, H-6(3); 2.01 (s, 18 methyl); 2.02 (s, 7-0 acetate methyl); 2.32 (d, J=19.0,
14a); 2.48 (d, J=19.0 Hz, H-14P); 2.92 (unresolved, H-3a); 2.74, br s (C-15 OH,
exchangeable with D20); 4.92, s (H-20 A); 5.28, s (H-20 B); 5.39 (br s, J= H-5P); 5.54 (t,
J=7.2 Hz, H-13P); 5.61 (dd, J=4.8, 11.4 Hz, H-7a); 6.09 (poorly resolved br d, J=10.5 Hz,
H-9a); 6.65 (d, J=10.5 Hz, H-10p); 7.43 (t, J=7.8 Hz, H-Ph-meta); 7.56 (t, J=7.8 Hz, H-
Ph-para); 7.87 (d, J=7.8 Hz, H-Ph-ortho).
13C NMR (CDCIg, 600 MHz, 5): 8.9 (C-18 methyl, q); 12.5 (C-19 methyl, q); 20.7
(7-0 acetate methyl, q); 21.0 (9-0 acetate methyl, q); 26.2 (C-17 methyl, q); 26.6 (C-16
methyl, q); 27.3 (C-2, t); 27.7 (C-6, t); 34.2 (C-3, d); 43.2 (C-8, s); 48.3 (C-14, t); 58.1 (C-
1, s); 70.8 (C-7, d); 71.0 (C-10, d); 73.2 (C-9, d); 75.6 (C-15, s); 111.4 (C-20, t); 128.8
(Ph-meta, d); 129.1 (Ph-ipso, s); 129.5(Ph-ortho, d); 133.4 (Ph-para, d); 136.4 (C-11, s);

51
145.2 (C-4, s); 144.3 (C-12, s); 163.1 (C-11, s); 165.4 (CO-Ph, s); 169.3 (CO-Acetate, s);
170.3 (CO-Acetate, s); 207.5 (C-13 Ketone, s).
Analysis calculated for C 31H 3809: C, 67.13; H, 6.91. Found: C, 67.48; H, 6.97.
Dihydrobrevifoliol [3-71
A solution of brevifoliol (0.2 g) in ethyl acetate (10 ml) was hydrogenated in a
Parr apparatus using Platinum oxide (0.05 g) for 16 hours. TLC revealed the formation
of a slightly slower moving product. The mixture was filtered and the filtrate
concentrated to dryness and purified by chromatography on silica gel in
dichloromethane. Elution with 2% acetone in dichloromethane gave a minor product,
which was not further investigated. The fractions eluted with 2-5% methanol in
dichloromethane gave the major product, which was obtained as a colorless powder,
yield, 0.1 g,
Reduction of the 4(20) double bond resulted in significant broadening of most
peaks in the 1H- and 13C NMR spectra, but the appearance of additional signals from
methyl group at C-20 and methylene at C-4 could be seen, as well as the loss of the two
characteristic exocyclic methylene singlets.
Analysis calculated for C3iH420g: C, 66.60; H, 7.60. Found: C, 66.89; H, 7.88.
Brevifoliol Epoxide [3-81
A mixture of brevifoliol (0.3 g) and meta-chloroperoxybenzoic acid (MCPBA, 0.2
g) in toluene (15 ml) was heated under reflux for 30 min. After cooling, the mixture was
diluted with ether, washed successively with aqueous sodium bisulfite, aqueous sodium
bicarbonate and saline, and the organic layer concentrated to dryness. The solid was
crystallized from acetone/ligroin, to give a colorless crystalline epoxide, yield, 0.15 g;
m.p. 227-230 C.

52
1H NMR (CDCI3, 600 MHz, -40 C, 8) : 1.01 (H-19, s); 1.09 (H-16, s); 1.24 (dd,
J=7.2, 13.8 Hz, H-14a); 1.27 (H-17, s); 1.40 (H-2a, br d J=14.1 Hz); 1.76 (s, methyl, 9-
acetate); 1.80 (m, H-6a ); 2.0 (m, H-6p); 2.01 (s, H-18); 2.07 (s, 7-acetate methyl); 2.36
(dd, J=14.1, 9.6 Hz, H-2p); 2.46 (dd, J=7.2, 13.8 Hz, H-14p); 2.67, br s (C-15 OH,
exchangeable with D20); 2.64 (H-3a, br d, J=9 Hz); 2.72 (C-20, s); 3.59 (C-20, s); 4.20
(br s, H-5p); 4.46 (H-13p, br d, J=7.2 Hz); 5.57 (H-7a, br d, J=4.8, 11.4 Hz); 6.05 (H-
9a, poorly resolved br d, J=10.5 Hz); 6.54 (H-10p, br d, J=10.5 Hz); 7.43 (Ph-meta, t,
J=7.8 Hz); 7.56 (Ph-para, t, J=7.8 Hz); 7.87 (Ph-ortho, d, J=7.8 Hz).
13C NMR (CDCI3, 600 MHz, -40 C, 5): 11.9 (C-18 methyl, q); 12.9 (C-19 methyl,
q); 20.7 (7-0 acetate methyl, q) ; 21.3 (9-0 acetate methyl, q); 23.9 (C-2, t); 25.0 (C-17
methyl, q); 27.1 (C-16 methyl, q); 34.1 (C-3, d); 34.4 (C-6, t); 45.4 (C-8, s); 46.4 (C-14, t);
50.1 (C-4, s); 60.2 (C-20, t); 62.4 (C-1, s); 69.5 (C-7, d); 70.5 (C-10, d); 71.7 (C-5, d);
75.8 (C-15, s); 76.7 (C-13, d); 77.1 (C-9, d); 128.7 (Ph-meta, d); 129.4 (Ph-ipso, s);
129.5(Ph-ortho, d); 133.2 (Ph-para, d); 134.3 (C-12, s); 151.3 (C-11, s); 164.3 (CO-Ph,
s); 169.96 (CO-Acetate, s); 170.0 (CO-Acetate, s).
Analysis calculated for C31H4o010: C, 65.02; H, 7.04. Found: C, 64.72; H, 7.24.
Ozonization of Brevifoliol: Brevifoliol-norketone 3-91
A solution of brevifoliol (1 g) in a 9:1 mixture of chloroform and methanol (25 ml)
was cooled in a dry ice/ acetone bath and saturated with ozone produced by an ozonizer
(Ozone Research and Equipment Co., Phoenix, AZ). After testing for the absence of the
starting material by TLC, the mixture was removed from the bath and treated with
dimethyl sulfide (1 ml) and let stand at room temperature for 2 h. It was then
concentrated to dryness and applied to a silica column prepared in chloroform. Elution
with 2% acetone in chloroform gave two bands, which were separated and the fractions
concentrated separately.

53
The faster moving fraction [3-9] was obtained as a colorless, amorphous powder,
yield, 0.25 g.
1H NMR (CDCI3i 600 MHz, -40 0 C, 8) major rotamer: 1.02, s (H-19); 1.05, s OH-
16); 1.35, s OH-17); 1.46 (d, J=14.1 Hz, H-2cc); 1.78 (s, 9-0 acetate methyl); 1.98 (m, H-
6a); 2.0 (m, H-6p); 2.01 (s, 18 methyl); 2.02 (s, 7-0 acetate methyl); 2.32 (d, J=19.0,
14a); 2.48 (H-14p, dd, J=14, 7.5 Hz); 3.19 (H-3a, d, J=10.2); 4.20 (H-5p, br s); 4.48 (H-
13p, br t, J=7.2 Hz); 5.78 (H-7a, br m); 5.92 (H-9a, br); 6.52 (H-10p, br d, J=10.5 Hz);
7.43 (Ph-meta, t, J=7.8 Hz); 7.56 (Ph-para, t, J=7.8 Hz);7.87 (Ph-ortho, d, J=7.8 Hz).
13C NMR (CDCIa, 600 MHz, 5): 12.1 (C-18 methyl); 14.2 (C-19 methyl); 20.6 (7-
O acetate methyl); 21.2 (5-0 acetate methyl); 21.3 (9-0 acetate methyl); 24.8 (C-17
methyl); 27.0 (C-16 methyl); 29.1 (C-2); 33.9 (C-6); 34.8 (C-3); 46.4 (C-8); 46.4 (C-14);
62.0 (C-1); 69.9 (C-7); 70.7 (C-10); 72.0 (C-5); 75.9 (C-15); 76.9 (C-13); 77.3 (C-9);
128.8 (Ph-meta); 129.1 (Ph-ipso); 129.5(Ph-ortho); 133.4 (Ph-para); 134.4 (C-11); 144.5
(C-4); 147.3 (C-12); 164.3 (CO-Ph, s); 169.8 (CO-Acetate, s); 170.0 (CO-Acetate,
s);208.0 (C-13 C=0).
Analysis calculated for C3oH38010: C, 64.50; H, 6.86. Found: C, 64.68; H, 6.97.
The slower moving fraction was obtained as a colorless crystalline solid, yield,
0.3 g; m.p.225-232 0 C. It was found to be identical with the epoxide [3-8], described
above.
Brevifoliol-4,20-Diol [3-10]
To a solution of brevifoliol (0.4 g) in pyridine (10 ml) was added osmium tetroxide
(0.2 g) and the reaction mixture stirred for 1 h, after which time, the starting material was
replaced by a much slower moving component. After decomposing the excess reagent
with a solution of sodium bisulfite in pyridine, water and dilute sulfuric acid were added
and the mixture extracted with dichloromethane. After concentration, the product was

54
placed on a silica column in dichloromethane. Elution with 2% methanol in
dichloromethane gave the major band which yielded [3-8] as a white powder, final yield,
0.12 g.
Analysis calculated for C31H42011: C, 63.04; H, 7.17. Found: C, 62.88; H, 7.25.
Saponification of Brevifoliol [3-111
A solution of brevifoliol (1 g) in methanol (20 ml) was stirred with 1N potassium
hydroxide (10 ml) for 1 h. TLC showed that the starting material was absent and very
slow moving, non UV-absorbing component produced. The reaction mixture was
passed through a small column or Amberlite-IR120 ( a sulfonic acid resin) in the H+
form. The column was washed with 1:1 methanol/water. The effluent and washes were
concentrated to dryness and the solid crystallized from acetone to give [3-11] as a
colorless crystalline solid, yield, 0.45 g; m.p. 290C dec.
Analysis calculated for C20H32O6+H2O: C, 62.15; H, 8.87. Found: C, 62.48; H,
8.99.
Debenzoyl Brevifoliol-Pentaacetate [3-12]
Compound [3-11] (0.2 g) was acetylated using acetic anhydride (2 ml) and
pyridine (0.5 ml) by heating at 80 0 C for 30 min. Water was added to decompose the
reagent and the solid filtered. It was crystallized from ether/ ligroin, yield, 0.2 g; m.p.
184-187 C.
1H NMR (CDCI3, §): 6.36 (H-10, d, J=10.2); 5.88 (H-9, br d, 10.2); 5.54 (H-13(3, t,
J=7.2); 5.53 (H-7cx, q, J=5.4, 10.2); 5.36 (H-5p, br s); 5.26 (H-20b, s); 4.87 (H-20a, s);
2.65 (OH-15, s); 2.64 (H-3a, d, J=9); 2.48 (H-14p, dd, J=13.8, 7.2); 2.36 (H-2a, dd,
J=14.1, 9.3); 2.07 (AcMe, s); 2.06 (AcMe, s); 2.02 (AcMe, s); 2.00 (AcMe, s); 1.97
(AcMe, s); 1.95 (18-Me, s); 2.0 (H-6a, m); 1.85 (H-6p, cm); 1.42 (H-2p, d, J=14.4); 1.31
(H-17 Me);1.22 (H-14a, dd, J=13.8, 7.2); 1.13 (H-16 Me); 0.88 (H-19 Me).

55
IR i/ max (KBr, cm'1): 3565 (OH, sharp); 2960, 1740 (C=0); 1370, 1230, 1030.
MS(FAB): 580 [MH+].
Periodate Oxidation of [3-111 to [3-131
A solution of [3-11] (0.2 g) in methanol (5 ml) was treated with sodium periodate
(0.3 g) in 1N sulfuric acid (2 ml). After 30 min, TLC showed that the starting material
was absent and was replaced by a faster moving product visible under the UV light,
unlike the starting material. After dilution with water, the mixture was extracted with ethyl
acetate and the extract concentrated to dryness. The crude dialdehyde [5-13] was not
further purified before the next reaction, but did exhibit signals for two aldehydes the 1H
spectra. The only significant changes from the parent compound showed the loss of the
isolated 1H spin system from the protons on C-9 and C-10, and the conversion of two
hydroxyl carbons into aldehydes (Guthrie, 1961).
1H NMR (CDCI3,5) : 9.96 (CHO, s); 9.42 (CHO, s); 5.36 (H-20b, s); 5.02 (H-20a,
s); 4.58 (H-5a, m); 4.40 (H-13a, br t); 2.60 (H-3p, d); 1.30 (Me, s); 1.25 (Me, s); 1.02
(Me, s); 0.88 (Me, s).
Analysis calculated for C30H42OH: C, 62.27; H, 7.32. Found: C, 62.10; H, 7.44.
Formation of Osazone [3-141 from [3-131 with 2,4-DNPH
Dialdehyde [3-13] was dissolved in methanol, (2 ml) and heated with a solution of
2,4-dinitrophenyl hydrazine (0.1 g) in 2N hydrochloric acid (2ml) and methanol (2 ml).
After 2 h at room temperature, the mixture was extracted with chloroform and the
concentrated extract chromatographed on a silica column. The major band [3-14] was
obtained as an orange yellow crystalline solid, m.p. 215-218 C. Both 1H- and 13C NMR
spectra indicated the retention of all twenty of the carbons from the abeo-taxane
skeleton.

56
Debenzoyl Brevifoliol [3-151
A solution of brevifoliol (0.3 g) in 30% methanol in water (10 ml) was heated in a
sealed tube at 135 0 C for 90 min. The cooled mixture was neutralized with sodium
bicarbonate and extracted with chloroform. The extract was purified by chromatography
on a C-8 reverse phase column using a step gradient of 25-60% acetonitrile in water in
5% increments of solvent concentration. Elution with 30-35% acetonitrile in water gave
the major product, which was obtained as a white powder, yield, 0.1 g. The 1H and 13C
NMR spectra revealed the loss of the benzoate from the C-10 position and retention of
the acetate substituents at C-9 and C-7
Analysis calculated for C24H3608: C, 63.70; H, 8.02. Found: C, 63.89; H, 7.88.
Benzoic acid was also isolated from the reaction mixture and confirmed using
NMR and UV analyses.

CHAPTER 4
SOME UNUSUAL REACTIONS OF BREVIFOLIOL
The isolation and structural elucidation of brevifoliol [3-1] was described in detail
in Chapter 3. Brevifoliol occurs to the extent of 0.2-0.3 % in the needles of T. brevifolia,
and to a lesser extent in the bark of the same species, in the needles of T. x media
Hicksii (Rehd.) and in the needles of T. wallichiana. (Georg et al. 1993). Large
quantities of the crystalline compound can be readily isolated from the needles of T.
brevifolia, which is the best source for the compound.
However, other than acetylation to a diacetate (Balza et al. 1991); and the
attachment of the N-benzoyl phenyl-isoserine side chain at the C-13 position (Georg, et
al. 1993); no record of its reactions reflecting its various functional groups has been
published. This paucity of such information is in keeping with the current trend that, in
spite of the virtual explosion of new taxanes that were isolated and characterized
structurally by spectral data over the past five years. Very few have been investigated
for their chemical reactions, for the relative reactivities of similar functional groups, or for
any unusual reactions as a consequence of their stereochemical disposition.
Understanding the products of the various transformations that a compound can
be subjected to and the relative rates of reaction can lead to important insight
concerning entity, and for the general advancement of chemistry. Although the structure
of brevifoliol is now well established as a result of spectral and x-ray crystallographic
data, we can gain insight into this molecule through reactions such as those described in
Chapter 3. In addition to these, a number of reactions also carried out for the purpose of
57

58
structural elucidation of [3-1] took unexpected courses. The products of these reactions
were isolated and characterized structurally, and the details are described here.
1. Acid-Catalyzed Acetylation
In Chapter 3, acetylation of [3-1] by means of acetic anhydride and pyridine to
give the crystalline monoacetate [3-2] and the diacetate [3-4] has been described. We
isolated from the needles of T. brevifolia another taxane, which resembled brevifoliol,
giving a dark greenish blue color when sprayed with sulfuric acid and heated. This same
product was also subsequently isolated by others (Barboni et al. 1993). It was similar to
but different from the monoacetate [3-2], but when acetylated, gave [3-4], thus showing
that it was the 13-acetate [3-3],
Acetylation of brevifoliol using acetic anhydride and BF3 gave a different
crystalline product, with almost the same RF as that of the diacetate [3-4], The BF3
product closely resembled the 5, 13-diacetate in its NMR spectral properties also, but
with some differences. In the 1FI NMR spectrum, the singlet at 52.74 which is assigned
to the 15-hydroxyl in the diacetate [3-4], was no longer present. The two methyl singlets
assigned to C-16 and C-17 at 51.11 and 51.35 were deshielded to 51.63 and 51.71. The
other two (C-18 and C-19) showed either no shift or a much smaller one (5 2.03->5 1.96
for C-18). Thus, the significant downfield shift of the signals due to C-16 and C-17
suggested that acetylation of the 15-hydroxyl might have taken place.
Additionally, the signals due to the H-2a and 2(3 showed slight downfield shifts
(H-2a: 51.46, broad doublet, J=13 Hz to 1.53 ppm; H-2p: 52.41, doublet of doublets,
J=9,13 Hz to 2.65 ppm). The spectra of brevifoliol and its acetates generally show the
H-10a signal as a sharp and well resolved doublet, but the one due to H-9p as a broad,
poorly resolved doublet. In the BF3-reaction product, the signal due to H-9p was not only
a well resolved doublet (J=10.8 Hz); but also shielded by 0.29 ppm to 5 5.80. These

59
observations suggest that the C-15 hydroxyl might be responsible for the blurring of the
signal of H-9(3, and acetylation of this hydroxyl has eliminated that interaction.
Further support for the acetylation having taken place at the C-15 hydroxyl is
shown by the appearance of another acetyl methyl signal at 5 2.11 in the 1FI NMR
spectrum. The appearance of two more signals in the 13C NMR spectrum was also
evident, in which 5 acetyl methyl and 5 acetyl-carbonyl signals were present, with the
fifth one at 5 21.7 and 5 169.5, respectively. Also in the 13C NMR spectrum, a significant
downfield shift from 5 75.6 to 5 87.2 for the C-15, and an upfield shift from 5 27.0 to 5
24.8 and from 5 23.1 to 5 22.0 for the signals due to C-16 and C-17 respectively,
completes the evidence to indicate that the 15-hydroxyl was acetylated to give [3-5],
A comparison of the 13C NMR spectral data of brevifoliol, the 5-monoacetate
[3-2], the 13-monoacetate [3-3] (naturally occurring and confirmed through semi
synthesis); the 5, 13-diacetate [3-4] and of the BF3-catalyzed acetylation product [3-5]
were shown in Chapter 3 in Table 3-1.
2. Oxidation
Oxidation of brevifoliol with manganese dioxide gave the monoketone, the NMR
spectral data of which showed that the 13-hydroxyl was oxidized, leading to the structure
[3-6], as described in Chapter 3. Oxidation of [3-1] with Jones reagent gave initially, the
same 13-monoketone [3-6], but on further reaction, this was replaced by a faster moving
compound [4-1], whose spectral data pointed to an unexpected course of reaction.
The molecular formula of the product, C31FI3609 (MH+, 553) indicated the loss of
4 protons, as compared to brevifoliol. Although this might indicate that both hydroxyls
were oxidized to give the diketone [4-2], certain features suggested otherwise. To begin
with, the 1FI NMR spectrum showed broad peaks which indicated the existence of a

60
rotameric equilibrium, which was confirmed by a spectrum taken at -40 0 C, in which two
sets of peaks with a 5:1 ratio were seen.
In the major rotamer, the coupling between the C-9 and C-10 protons was found
to be 4 Hz, in contrast to the value of 10.5 Hz shown by brevlfoliol [3-1], (a similar
diketone prepared from 2-acetoxy brevifoliol (taxchinin A); described by Fuji et al. (1992)
and Appendino et al. (1993) also showed a coupling of 4 Hz. Next, a singlet appeared at
9.4 ppm, which interacted in the HETCOR spectrum with the peak at 194 ppm. The
latter showed a negative signal in the Attached Proton Test (APT). These observations
indicated the presence of an aldehyde functionality, presumably at C-20. Additionally, in
the spectrum of the diketone such as [4-2], the C-5 proton signal was absent, as
expected, and the exocyclic methylene protons appeared as two singlets at 5 5.06 and 8
5.94 ppm. The corresponding carbon signals appearing at 5127 and 8143.4 ppm.
However, in the Jones oxidation product, the singlets due to the exocyclic
methylene protons were absent, and the characteristic C-20 carbon signal which
appears in the 8110-8120 ppm region was also missing. Instead, a signal was found at
5 6.73 ppm, which interacted with the signals at 82.5 and 52.7 (C-6 protons); and in the
HETCOR spectrum, with the signal present at 8 147 ppm, and this latter gave a negative
signal in the APT spectrum. These data seem to suggest that the product is not the
5,13-diketone [4-2], but an aldehydic product with a double bond present at C-4/C-5, as
shown in [4-1],
One possible explanation is that the sulfuric acid in the reagent caused hydration
of the 4/20 double bond, and the primary alcohol so generated was oxidized to the
aldehyde, followed by dehydration to yield the 4/5-double bond.

61
[4-1 ] Jones Oxidation Aldehyde [4-2] Mn02 Diketone
Figure 4-1 : Oxidation Products
3. Action of BF3 on Brevifoliol [4-31
Before the structure of brevifoliol was fully established, the possibility that the two
hydroxyls present in the compound might be vicinal to each other was considered. With
the aim of forming an isopropylidene derivative, brevifoliol was reacted with acetone in
the presence of Dowex-50 (sulfonic acid resin) as the catalyst. Reaction took place
readily with the formation of a faster moving product, with some decomposition also
taking place (colors). The reaction proceeded with less decomposition when BF3-
etherate was used as the catalyst.
Chromatography of the mixture from either reaction yielded the major product as
a colorless crystalline solid. Its NMR spectrum quickly ruled out the possibility of its
being an isopropylidene derivative. The FAB-mass spectrum gave a value for the MFI+
m/z of 481, as compared with 557 for brevifoliol, thus showing a loss of 76 mass units.
The elemental analysis, which agreed with C28FI32O7 showed a loss of C3FI802 compared
to brevifoliol. This may be interpreted as the loss of the C3H7O- side chain attached to
C-1, as well as that of FI2O, possibly through the elimination of the C-13-OH (or C-5-OFI).
The evidence to support this assertion is given below.

62
1. The most striking evidence is that only two of the four methyl signals that are seen in
brevifoliol appear in the BF3-reaction product. This is not only observed in the 1H
spectrum, but also in the 13C NMR thus showing that the oxy-isopropyl chain at C-1
is not present.
2. An examination of the isolated spin system: 5p 6a- 6p 7a in the 1H NMR spectrum
of brevifoliol and that of the product of BF3 reaction indicated that the free hydroxyl at
C-5a was still present, but that the C-13a-OFI was absent. Through a study of the
interactions in the COSY spectrum, it was possible to assign (and distinguish) the
signals for C-6 and C-14 in the region below S3.
3. A new methine proton signal appeared as a broad singlet at S5.84, assigned to H-13,
which interacted in the FIETCOR spectrum with the signal at 5124.1, which
corresponds to C-13.
4. The DEPT spectrum showed the presence of four signals for methyl-type (CPI3)
carbons (811.4, 5 13.5 for C-19 and C-18, 5 20.7, 8 21.2 for the two CO-Me); two
signals for methylene-type (CFb) carbons (8 112.4 for C-20, 8 44.6 for C-14, 8 34.1
for C-6 and 8 27.9 for C-2, nine signals for methine-type (R3CPI) carbons (5133.0,
8129.7, 5128.4 for the five aromatic carbons, four aliphatic CH-OR type carbons
[873.2 (C-9); 572.5 (C-5); 579.6 (C-7); 867.4 (C-10)], one aliphatic methine type
carbon (539.6 (C-3)); one unsaturated methine type carbon (5124.1 (C-13) and nine
quaternary carbons (three carbonyls, one aromatic C carrying the carboxyl, C-1(?);
C-4, C-11, C-12 and C-8); together, account for the 28 carbons present..
5. HETCOR interactions supported the assignments for the o, m, p- positions in the
benzoate, and for the C-13, C-20, C-9, C-5, C-7, C-10, C-3, C-6, C-2, C-18, C-19
and the two acetate methyls.

63
6. In the 1D nOe-difference spectrum, the interaction between the C-14a,p and C-13 p
protons was the strongest. Crowding of the region around 52.8 ppm made the
spectrum more difficult to interpret, and not very informative.
Table 4-1 : NMR Spectra of Compound [4-3] from BF3 Reaction
Position
Proton
Carbon
APT
DEPT
1

145.9
"T"
C
2 a
1.96 m
28.0
~T~
ch2
2 P
3.00 m



3 a
2.70 m
39.4
~T
CH
4
150.4
~T~
C
5 a
4.42 brs
72.6
nr
CH
6 a
1.76 m
34.1
ch2
6 P
2.05 m


7a
5.48 brd
70.6
nr
CH
8

46.6
C
9 P
5.34 d 6.0
73.2
n-
CH
10a
6.30 d 6.0
67.5
CH
11

146.4
C
12

134.2
n-
C
13 p
5.83 br s
124.0
CH
14 a, p
2.85-3.0 cm
44.6
ch2
18
2.06 s
11.4
CO
X
0
19
1.23 s
13.4
nr
ch3
20 A
4.98 br s
112.2
ch2
20 B
5.20 br s


CO-C6H5

165.2
nr
C
Bz-ipso

130.2
T
C
Bz-ortho
8.01, d 7.5
128.4
CH
Bz-meta
7.44, t 7.5
129.6
nr
CH
Bz-para
7.56, t 7.5
132.9
nr
CH
COCH3
1.98 s
20.6
ch3
2.00 s
21.1
i
ch3
COCH3

169.9
T
C
170.4
T
C
1H NMR were recorded at 600 MHz and 13C NMR at 150 MHz in CDCl3on
a Varian Unity 600 spectrometer at ambient temperature. Chemical shifts 5 (ppm)
are reported relative to TMS as an internal standard.

64
Bz
Ac
[4-3] BF3-etherate Product
Figure 4-2 : BF3-etherate Catalyzed Elimination Product
Based on this reasoning, the structure of the BF3-reaction product was assigned
as shown [4-3] in Figure 4-2 above. The DEPT spectra of [4-3] are given in Figure 4-3.
CH3 Carbons
CH2 Carbons
20
2x Ac 18 19
14 6 2
CH Carbons
13
5
9
7
10
3 DMSO
i
.... .
-- -

Adi Protonated Carbons
i t i ] r 1 i r -j i i T'i ]' i n i j i m 7 rrrrriTT rrpr'n T]-| i t ;-) n r r j n m m r r r yr ¡-mi tt trt -j- t ti i ?
140 120 100 00 60 40 20 6
Figure 4-3 : DEPT Spectra of BF3 Elimination Product [4-3]

65
A reaction such as this has not been reported in the taxane series, resulting in
the loss of the oxy-isopropyl side chain. In taxol and related compounds containing the
conventional taxane skeleton, action of Lewis acids such as BF3 was studied and is
shown to produce one or two different changes, depending on the whether protic or
aprotic solvent is used. In one case, isomerization of the A-ring from a 6- to a 5-
membered A ring takes place with the oxy-isopropyl group attached at C-1. In the
second instance, the oxetane ring is opened to form a diol, or with the acetate group
migrating from C-4 to C-20 to give the 4-hydroxy-20-acetoxy compound. The reaction
described here appears to be a continuation of the action of the Lewis acid on the 5-
membered A ring, with the elimination of the oxy-isopropyl substituent.
4. Reaction with lodine/Silver Acetate [4-41
Once the structure of brevifoliol had been elucidated, the value and usefulness of
this relatively abundant compound in the needles of T. brevifolia was considered. Since
the addition of the N-benzoyl isoserine side chain at C-13 did not generate activity in the
final product, it was reasoned that the oxetane ring at the 4/20 position might be
necessary to produce activity. To this end, one approach was investigated, involving the
use of iodine in some form to add across the 4/20 double bond and thereby permitting
substitution with other groups.
Brevifoliol was found to react readily with bromine, but the reaction yielded
multiple products and considerable decomposition. Reaction with iodine was similarly
complex and led to much decomposition and dark colored products. With the idea that
addition of a silver salt which can remove the acid that might be produced, but not be too
strongly basic (e.g. silver oxide) and hence hydrolyze the ester functions, silver acetate
was selected for use with the iodine. The remote possibility that silver acetate might

66
displace the iodine located at C-4 to produce the 4-acetoxy-20-iodo compound was also
attractive (Woodard & Brutcher, 1958).
When brevifoliol was stirred with iodine and silver acetate, the course of the
reaction was clearly different. No multiple products or decomposition to dark colored
products was seen, even when the reaction was continued for 15-30 hours, unlike the
reaction with iodine alone that became dark in 1-2 hours. The reaction was continued
until the starting material was consumed and a major, faster moving compound was
produced. Chromatography on a silica column gave a colorless crystalline solid.
Acetylation with acetic anhydride and pyridine at 70 C for 15 minutes produced
the 5-O-monoacetate [4-5], confirming the presence of the 5 hydroxyl with NMR
analysis. Treatment of [4-4] with n-Bromosuccinimide produced the 5-0 ketone, further
confirming the absence of a free C-13 hydroxyl on the basis of NMR and UV spectral
data.
[4-4] lodine/Silver Acetate FToduct
[4-5] Acetate
Figure 4-4 : lodine/Silver Acetate Product [4-4] and Acetate

67
Figure 4-5 : H,H-COSY Spectrum of [4-4]
The 1H NMR spectrum was similar to brevifoliol in most respects, with a few
revealing differences. The AB quartet normally seen from H-2p became a clean doublet
(J=13.8 Hz) deshielded from S 2.36 to 5 3.31, and coupled to the H-2a proton, which was
deshielded from 5 1.49 to 5 2.34. This deshielding and the coupling patterns indicated
the loss of H-3p with the possible formation of a double bond between C-3 and C-4.

68
JjLjULjL-JjJLjikj
w
Q.
Q.
OJ
U-
Figure 4-6 : HETCOR Spectrum of [4-4]
Recorded on Varian VXR 300 spectrometer at 75.432 MHz in CDCI3 with TMS as
internal standard. F-,: 3000.3 Hz; F2: 15528 Hz; Acquisition time 65.9 sec; 1 sec;
Ambient temperature; Decouple proton; Level 70 high power; PW: 90; 128 repetitions X
128 increments; Waltz 16 modulation; pseudo echo; FT size: 2K X 512 data points; time:
5 hours.
Elemental analysis and FAB-MS gave the molecular formula C31H3608, which
indicated essentially that loss of one molecule of water and dehydrogenation had taken

69
place. The presence of a hydroxyl was indicated by the fact that the compound would
still undergo acetylation to form a monoacetate.
The C-15 hydroxyl was still present in both the carbon and proton spectra, and
the only other significant change in the spectra occurred with the C-20 signals. The two
broad singlets normally seen around 5 ppm were replaced by an isolated spin system of
doublets at 4.35 and 3.92 which resemble the H-20 oxetane pattern seen in taxol.
Experimental
Brevifoliol Triacetate [3-51
To a solution of brevifoliol (0.2 g) in acetic anhydride (2 ml) was added 0.2 ml of
a 2% solution of boron trifluoride etherate in acetic anhydride to give a final
concentration of 0.2% of boron trifluoride etherate. After 20 min at room temperature,
the mixture was diluted with water. After another 10 min. the solid was separated,
washed, taken up in ether and concentrated to dryness. The solid was crystallized from
ether in ligroin to give [3-5] as a colorless crystalline solid, yield, 0.2 g; m.p. 214-216 0 C.
Oxidation with Jones Reagent to T4-11
A solution of brevifoliol (0.2 g) in acetone (10 ml) was treated with Jones reagent
(2 ml) added in small portions and with stirring. Initially TLC analysis of the reaction
mixture showed that a yellow color giving spot appeared above that of the starting
material. Gradually the first product changed into an even faster moving component.
When this latter was the predominant product, the reaction was stopped by the addition
of water and extraction with chloroform. After concentration of the solvent, the product
was chromatographed on a silica column in 1:1 chloroform/ ligroin. Elution with
chloroform gave the major component, which was obtained as a colorless crystalline
solid, yield, 0.05 g; m.p. 234-236 0 C.

70
Action of Boron Trifluoride on Brevifoliol 4-31
Brevifoliol (0.3 g) was dissolved in acetone (10 ml) and to the solution was added
1 ml of 1% boron trifluoride etherate in acetone to make a 0.1% overall concentration of
boron trifluoride in the reaction mixture. After 3 h, water was added, the solid filtered
and after drying, subjected to chromatography on silica gel in chloroform/ ligroin (1:1).
The major band obtained with the same solvent was crystallized from ether/ligroin, yield,
0.1 g, m.p. 162-165 0 C.
1H NMR (CDCI3, Varian Unity 600 MHz, 8): 1.16, s (H-19); 1.03, s (H-16); 2.08
(cm, H-14a); 1.27, s (H-17); 2.32 (d, J=13.8 Hz, H-2a); 1.92 (s, methyl, 9-acetate); 1.84
(m, H-6a ); 1.98 (m, H-6p); 2.26 (s, H-18); 2.05 (s, 7-acetate methyl); 3.30 (d, J=13.8 Hz,
H-2P); 2.28 (cm, H-14P); 2.95, br s (C-15 OH, exchangeable with D20); 4.38 (t, J=7.2
Hz, H-13P); 4.28 (br s, H-5P); 4.54 (d, J=13.2 Hz, H-20 B); 4.54 (d, J=13.2 Hz, H-20 B);
5.51 (dd, J=4.8, 11.4 Hz, H-7a); 6.07 (d, J=10.1 Hz, H-9a); 6.54 (d, J=10.1 Hz, H-10p);
7.45 (t, J=7.5 Hz, H-Bz-mefa); 7.57 (t, J=7.5 Hz, H-Bz-para); 7.89 (d, J=7.5 Hz, H-Bz-
ortho).
13C NMR (CDCI3, Varian VXR 300 MHz, 8): 11.4 (C-18 methyl, q); 13.4 (C-19
methyl, q); 20.6 (7-0 acetate methyl, q); 21.1 (9-0 acetate methyl, q); 28.0 (C-2, t); 34.1
(C-6, t); 39.4 (C-3, d); 44.6 (C-14, t); 46.6 (C-8, s); 67.5 (C-10, d); 70.6 (C-7, d); 72.6 (C-
5, d); 73.2 (C-9, d); 112.2 (C-20, t); 124.0 (C-13, d); 128.4 (C-Bz-ortho, d); 129.6 (C-Bz-
meta, d); 130.2 (C-Bz-ipso, s); 132.9 (C-Bz-para, d); 134.2 (C-12, s); 145.9 (C-1, s);
146.4 (C-11, s); 150.4 (C-4, s); 165.2 (CO-Ph, s); 169.9 (CO-Acetate, s); 170.4 (CO-
Acetate, s).
Reaction with Iodine and Silver Acetate [4-41
To a solution of brevifoliol (0.5 g) in benzene (15 ml) were added iodine (0.7 g)
and silver acetate (0.75 g) and the mixture stirred at room temperature for 20 h. TLC

71
showed that the starting material was absent and was replaced by two faster moving
components. The mixture was filtered and the filtrate washed successively with
aqueous sodium bisulfite and water and concentrated to dryness. Chromatography on
silica gel in 4:1 chloroform/ligroin gave the major band, which was obtained as a
colorless crystalline solid, total yield, 0.12 g; m.p. 250-252 0 C.
1H NMR (CDCI3, Varan Unity 600 MHz, 5): 1.16, s (H-19); 1.03, s (H-16); 2.08
(cm, H-14a); 1.27, s (H-17); 2.32 (d, J=13.8 Hz, H-2a); 1.92 (s, methyl, 9-acetate); 1.84
(m, H-6a); 1.98 (m, H-6(3); 2.26 (s, H-18); 2.05 (s, 7-acetate methyl); 3.30 (d, J=13.8 Hz,
H-2p); 2.28 (cm, H-14P); 2.95, br s (C-15 OH, exchangeable with D20); 4.38 (t, J=7.2
Hz, H-13P); 4.28 (br s, H-5p); 4.54 (d, J=13.2 Hz, H-20 B); 4.54 (d, J=13.2 Hz, H-20 B);
5.51 (dd, J=4.8, 11.4 Hz, H-7a); 6.07 (d, J=10.1 Hz, H-9a); 6.54 (d, J=10.1 Hz, H-10p);
7.45 (t, J=7.5 Hz, H-Bz-mefa); 7.57 (t, J=7.5 Hz, H-Bz-para); 7.89 (d, J=7.5 Hz, H-Bz-
ortho).
13C NMR (CDCI3, Vahan VXR 300 MHz, 5): 13.1 (C-18 methyl, q); 16.1 (C-19
methyl, q); 20.8 (7-0 acetate methyl, q); 21.5 (9-0 acetate methyl, q); 25.3 (C-17 methyl,
q); 26.9 (C-16 methyl, q); 31.7 (C-2, t); 38.2 (C-6, t); 142.7(C-3, d); 45.2 (C-8, s); 38.6
(C-14, t); 65.6 (C-1, s); 67.2 (C-7, d); 76.2 (C-10, d); 69.8 (C-5, d); 74.3 (C-15, s); 84.0
(C-13, d); 72.4 (C-9, d); 64.4 (C-20, t); 128.8 {C-Bz-meta, d); 129.2 (C-Bz-/pso, s);
129.4(C-Bz-ortho, d); 133.4 (C-Bz-para, d); 135.5 (C-12, s); 142.7 (C-4, s); 146.4 (C-11,
s); 164.5 (CO-Ph, s); 169.4 (CO-Acetate, s); 170.3 (CO-Acetate, s).
FAB-MS (dithiothreotol/dithioerythrotol / TFA, m/z): 577 [M+Na]; 537 [M+ NaOH];
433 [M+-Na02C7H5]; 373 [M+-Na02C7H5 -HOAc]; 313 [M+-Na02C7H5 2x HOAc]; 253
[M+-Na02C7H5-3x HOAc],
CI-MS (methane, m/z): 537.9 [MH+ -H2Oj; 373.6 [MH+-H20-HOAc- C6H5COOH],

72
Acetylation of [4-41 to [4-51
A sample of [4-4] (.05 g) was acetylated in acetic anhydride (2 ml) and pyridine
(0.5 ml) at room temperature for 20 h. After addition of water, the solid was filtered and
crystallized from ether in ligroin, m.p. 250-254 C.
1H NMR (CDCIa, Varian Unity 600 MHz, 5): 1.04 (s, H-16); 1.19 (s, H-19); 1.28 (s,
H-17); 1.91 (s, methyl, 9-acetate); 1.77 (br dd, H-14oc); 2.02 (m, H-6p); 2.05 (s, 7-acetate
methyl); 2.27 (s, H-18); 2.19 (s, 5-acetate methyl); 2.22 (cm, H-14p); 2.34 (d, J=13.8 Hz,
H-2a); 3.01, (brs, C-15 OH, exchangeable with D20); 3.31 (d, J=13.8 Hz, H-2(3); 3.90 (d,
J=13.2 Hz, H-20 B); 4.35 (d, J=13.2 Hz, H-20 A); 4.50 (br m, H-13|3); 5.41 (dd, J=3.6,
13.2 Hz, H-7a); 5.46 (d, J=4.2 Hz, H-5P); 6.04 (d, J=10.1 Hz, H-9a); 6.57 (d, J=10.1 Hz,
H-10P); 7.45 (t, J=7.5 Hz, H-Bz-mefa); 7.57 (t, J=7.5 Hz, H-Bz-para); 7.89 (d, J=7.5 Hz,
H-Bz-ortho).
13C NMR (CDCIa, Varian VXR 300 MHz, 5): 12.9 (C-18 methyl, q); 16.1 (C-19
methyl, q); 21.7 (7-0 acetate methyl, q); 21.5 (9-0 acetate methyl, q); 21.7 (5-0 acetate
methyl, q); 25.3 (C-16 methyl, q); 27.0 (C-17 methyl, q); 31.5 (C-2, t); 32.9 (C-6, t); 38.3
(C-14, t); 45.2 (C-8, s); 63.5 (C-1, s); 65.4 (C-20, t); 67.6 (C-7, d); 70.3 (C-5, d); 72.3 (C-
9, d); 74.4 (C-15, s); 76.3 (C-10, d); 83.2 (C-13, d); 127.3 (C-4, s); 128.8 (C-Bz-meta, d);
129.1 (C-Bz-/pso, s); 129.4(C-Bz-ortho, d); 133.4 (C-Bz-para, d); 135.5 (C-12, s);
142.7(C-3, d); 145.6 (C-3, s); 147.6 (C-11, s); 164.6 (CO-Ph, s); 169.6 (CO-Acetate, s);
170.3 (CO-Acetate, s); 170.8 (CO-Acetate, s).
Reaction with N-Bromosuccinimide and Silver Acetate [4-61
A solution of brevifoliol (0.2 g) in benzene (10 ml) was stirred with N-
bromosuccinimide {NBS} (0.125 g, recrystallized from water). After 2 h the starting
material was absent with two faster moving compounds being present. To the reaction
mixture was added silver acetate (0.125 g) and stirred for another 2 h.

73
At this point, the previous major compound moved further to give a new product.
The mixture was filtered, the filtrate washed with aqueous sodium bisulfite, followed by
water and concentrated to dryness. The product was chromatographed on a silica
column in 4:1 chloroform/ ligroin. The major component was obtained as a colorless
crystalline solid, yield, 0.1 g. The compound was found to be identical with the product
obtained from the reaction of brevifoliol with iodine and silver acetate.
Reaction of [4-41 with N-Bromosuccinimide [4-71
A solution of [4-4] (0.04 g) in benzene (5 ml) was stirred with N-
bromosuccinimide (25 mg) at room temperature. After 2 h, TLC showed formation of a
slightly faster moving compound, which was separable from the starting material only
after 2 or 3 developments of the TLC plate. The reaction mixture was washed with
aqueous sodium bisulfite, followed by water and concentrated to dryness. The product
was crystallized from ether in ligroin, m.p. 185-188 0 C.

CHAPTER 5
TAXANE CONSTITUENTS OF THE NEEDLES OF TAXUS X MEDIA
As discussed in Chapter 3, the yield of taxol from the bark of Taxus brevifolla by
using the conventional methods of isolation was of the order of 0.01%. It was also
shown that through the use of these same methods, no other useful analogues could be
isolated in any significant yields. As a consequence of these results and strong
ecological considerations, an intense search was started with the aim of finding a source
that is renewable, and which can match the bark in the yield of taxol. Many of the
available species of Taxus, as well as the various parts of these plants were examined
through the use of analytical high performance liquid chromatography (HPLC) and thin
layer chromatography (TLC). These searches led to the selection of the needles of the
ornamental yew, Taxus x media Hicksii as a possible answer to the problem. The
ornamental yew is capable of being grown in a nursery type setting, and on a large
scale, so that the needles may be clipped twice a year, and the taxol, which is found to
be present to the extent of 0.01% be isolated from them.
At the time of this research (1992-93), almost all of the studies carried out on this
species consisted of HPLC analyses. Other than the isolation of taxol by the standard
procedure with a total yield of 0.006%, no information had been published either on the
taxane constituents, or even a method for the practical isolation of them. In these HPLC
analyses, it was recognized that in the extracts of the ornamental yew, taxol was
accompanied by other co-eluting taxanes and these could contribute some errors in the
total yield calculations. These co-eluting taxanes were isolated in minute yields, in the
form of two components (0.8 mg and 1.2 mg); each representing an equilibrium mixture
74

75
of two components. On the basis of NMR spectral evidence, structures were assigned
to these two components (Castor & Tyler, 1993).
Due to the presence of pigments, waxes and other impurities, the isolation of
taxol and other taxanes from the needles was expected to be more difficult when
compared to their isolation from the bark of T. brevifolia. A project was started in this
laboratory to meet the need for a practical method for the isolation of taxol and other
related taxanes from the bark and needles of various Taxus species in spite of these
challenges. The application of a preparative scale reverse phase column
chromatography technique proved to be surprisingly successful in the processing of the
extracts of T. brevifolia.
To begin with, the HPLC analysis of the extract of the needles of the ornamental
yew, as shown in Figure 5-2, clearly shows that taxol is accompanied by several major
taxane components, which are present in much higher concentrations than taxol itself.
In view of such relatively high concentrations of these components, it is surprising that
only such minute amounts of two of these mixtures could be isolated earlier, as indicated
above. Also, no other characterizing data were provided other than the spectral data.
This laboratorys objective was the development of a simpler procedure for the isolation
of taxol with potential for large-scale use, in addition to more fully characterizing the
major taxanes present in the extract. The needles of the ornamental yew (200 lbs.,
dried) were received through the courtesy of Hauser Company, Boulder, CO, during
May-June 1993.
The extraction was carried out three or four times using methanol and the extract
concentrated to a syrup. The resulting concentrate was then partitioned between water
and chloroform, and the organic layer containing the taxane fractions was concentrated
to a thick semi-solid mass, which was used directly in the next step.

76
The reverse phase column procedure was carried out similar to what was used
with the needle extract of T. brevifolia, as described in Chapter 3. Approximately 200 g.
of the chloroform extract was dissolved in acetonitrile (see experimental) and stirred with
the equilibrated C-18 bonded silica. This slurry was then diluted to the appropriate
concentration of the acetonitrile and the added to the column prepared from 800 g of the
C-18 silica. Elution was carried out using a step gradient: 30, 35, 40, 45, 50 and 60%
acetonitrile in water, and the eluate collected in fractions of 200 ml. As was seen in the
case of the columns on the bark extract of T. brevifolia, when the fractions remained at
room temperature for about a week, crystals began to separate from the fractions in
different regions of the elution. These were filtered and further purified by either
recrystallization or a small column of normal phase silica where necessary.
The progress of elution of the column is shown in Figure 5-3. As anticipated,
taxol was accompanied by two other taxanes, which were present in higher
concentrations than taxol. However, all of these crystallized out of the fractions.
The early fractions contained the bulk of the UV absorbance, and from these
could be isolated a crystalline solid, which was a non-taxane compound. The next major
component that emerged with the 35-40% acetonitrile in water was shown to be
brevifoliol as described in Chapter 3. With the 45-50% acetonitrile and water solvent
were eluted taxane I, taxane II, followed by taxol, all of which crystallized from their
respective fractions, with some overlap.
The column was finally washed with a mixture of methanol and ethyl
acetate/ligroin (2:1:1) which stripped the column of all the waxes, chlorophylls and other
pigments. After, washing with methanol, followed by 25% acetonitrile and water the
column was made ready for another run. Figure 5-1 shows the steps involved in the
fractionation of the extract of Taxus x media Hicksii.

77
Dried Needles of
Taxus x media Hicksii
Extract
Residue
(Discard)
Chloroform
Aqueous
Extract Solids
Reverse Phase Column
Filter Crystals
Recrystallize or chromatograph
Brevifoliol Taxanes I and II Taxol Taxane III Taxane IV
Ozonization
Chromatography
Taxol
Figure 5-1 : Fractionation of the Extract of Taxus x media Hicksii Needles

78
Figure 5-2 : HPLC Trace of Taxanes Coeluting with Taxol.
Column Elution, Absorbance vs. Time
E
c
10
h-
CM
o
o
c
ro
-Q
s
G
if)
SI
<
Figure 5-3 : Progress of Elution of Taxanes from Reverse Phase Column.

79
The reverse phase column run on 600 g of the extract obtained from 12 Kg of the
dried needles was applied to a column prepared from 3 Kg of the C-18 bonded silica.
The yields of the products obtained were good and important values are listed below.
Brevifoliol 3-11
The fractions containing this component were combined, concentrated to dryness
and chromatographed on a normal phase silica column. The major component was
obtained as a colorless crystalline solid, which was found to be identical with brevifoliol
[3-1] on the basis of its spectral data.
Taxanes I [5-11 and II [5-21
Additional chromatography of the mixture of the taxanes I and II and taxol on a
normal phase silica column gave some separation of the two taxanes. Although they
could be further separated and obtained as crystalline solids, they still represented an
equilibrium mixture, as was indicated by the 1H- and 13C NMR spectra of the individual
crystalline samples. From the spectral data, these two were identified as a mixture of
5-0-cinnamoyl-10-acetyl taxicin I [5-1] and 5-0-cinnamoyl-9-acetyl taxlcin I [5-2], which
were isolated (Chmurney et at. 1993) from the needles of T. x media Hicksii and from the
needles of T. baccata (Appendino et at. 1992). The former authors obtained them in
quantities not sufficient for physical properties, and the latter authors obtained them as
amorphous powders, by using HPLC and preparative TLC.
The mixture of [5-1] and [5-2] on acetylation with acetic anhydride and pyridine
gave the triacetate [2-1], which was obtained as a crystalline solid and was found to be a
single entity unlike the starting material. It was also identical with the taxane III (see
below).

80
Taxane III [2-11
The crude crystalline solid obtained from the reverse phase column was
recrystallized. Its spectral and analytical data agreed with those given for 5-0-
cinnamoyl-2a,9a, 10p-triacetyl taxicin I (Appendino et al. 1993; Baxter etal. 1962).
Taxane IV 2-21
This was also purified by recrystallization of the crude crystals obtained directly
from the fractions. It was found to be identical with 5-0-cinnamoyl 2a, 9a, 10p-triacetyl
taxicin II, described by Appendino et al. (1992) and Baxter et al (1962).
Taxol r5-31
The chromatography using normal phase silica column as described under
taxanes I and II yielded taxol, which was purified by crystallization. The sample was still
contaminated with some of the taxanes I and II. For complete purification, the mixture
was subjected to ozonolysis which converted these two taxanes to more polar
compounds from which taxol could be readily separated and obtained pure. Using this
method, taxol was obtained in a yield of 0.015% based on the dry needles. This was
significantly better than the reported yield of 0.006% (Witherup et al. 1990).
Ozonolysis of [2-21
Because of the presence of the cinnamoyl ester function in compounds [5-1],
[5-2], [2-1] and [2-2], they all undergo ozonolysis. This method gives a convenient way
of separating taxanes [5-1] and [5-2] from taxol, with which they co-elute. In order to
determine the nature of the product of ozonolysis, taxane [2-2] was subjected to this
reaction and the product recovered and obtained as a crystalline solid. Its NMR spectral
characteristics indicated a hydrated aldehyde with the structure shown in [5-6],

81
Thus, in summary, dried needles of Taxus x media Hicksii were extracted and
the total chloroform extract applied to a C-18 reverse phase column. A number of
components were separated, such as brevifoliol, taxanes l-IV and taxol, which
crystallized out directly from the fractions. Separation of taxol from taxanes I and II
could be carried out directly by ozonolysis of the mixture, followed by chromatography
on either a normal phase or reverse phase silica column.
Experimental
Extraction:
Dried needles of Taxus x media Hicksii (50 lbs) were extracted with methanol as
described in Chapter 3. The combined concentrate was partitioned between water and
chloroform (10 gallons each). The organic layer was separated and the extraction
repeated twice more using 5 and 3 gallons of the solvent, respectively.
The combined chloroform layers were concentrated under reduced pressure to
yield a dark green semi-solid, representing approximately 5% of the weight of the dried
needles.
Chromatography:
Approximately 800 g of C-18 bonded silica gel was poured into a glass Michell-
Miller type column (2.5 x 24) using methanol (Ace Glass, 1430 North West Blvd.,
Vineland, NJ 08360). The column was equilibrated with 25% acetonitrile in water. The
chloroform extract solids (200 g) was dissolved in acetonitrile (400 ml) in a 4 L stainless
steel beaker, by warming in a hot water bath. To this was added approximately 200 g
equivalent of the equilibrated C-18 bonded silica (20-25% of the column material). While
the mixture is being stirred vigorously, 25% acetonitrile and water 500 ml was added,
followed by water (approximately 800 ml). After stirring for 15 min. it was checked for

82
uniformity of the slurry and the absence of oily or waxy material, or lumps. The slurry
was filtered under gentle suction and the solid was resuspended in approximately 500 ml
of the filtrate to give a thin enough slurry for pouring. It was added to the column, the
container rinsed and the rinse transferred to the column.
The remainder of the filtrate was pumped onto the column using a metering
pump (Pulsa 680, Pulsafeeder Inc., Rochester, NY). After the sample addition was
completed, elution was started using 30% acetonitrile and water. This was followed by
35, 40, 45, 50 and 60% acetonitrile and water. The column was then washed with 100%
methanol. Final washing of the column with a mixture of methanol and ethyl acetate and
ligroin removed the green pigments and other lipid-soluble components.
Fractions of 200 ml volume were collected and monitored by UV absorbance,
TLC and analytical HPLC. After this, those fractions that contained significant UV
absorbance and/ or components detectable by TLC or HPLC were set aside for 7-10
days, whereby crystals began to appear from a number of fractions. These were filtered
in groups, characterized and treated appropriately, as described below.
Characterization of the Taxane Components of Taxus x Media Hicksii
Brevifoliol [2-11
Fractions from the 40% acetonitrile and water were concentrated to dryness, the
solid taken up in chloroform and applied to a column of normal phase silica (40 g).
Elution with 2-5% acetone in chloroform gave the major band. The fractions that contain
this component were combined, concentrated to dryness and the solid crystallized from
acetone in ligroin. The crystalline product, yield, 0.8 g (0.02%) m.p. 220-222 0 C was
found to be identical on the basis of NMR spectral data with brevifoliol described in
Chapter 3.

83
Taxanes I and II [5-11 and [5-21
The crude crystals that separated out from the fractions (8 g) consisting of [5-1],
[5-2] and taxol [5-3] were processed by two methods. In one, the mixture (4 g) was
taken up in chloroform and ligroin (3:1, 50 ml) and applied to a column of normal phase
silica (60 g). The mobile phase was successively changed to chloroform, 2% acetone,
5% acetone, 2% methanol and 5% methanol in chloroform. Compounds [5-1] and [5-2]
appeared in the 2-5% acetone and chloroform eluate partially separating from each
other. Continuing with 2% methanol in chloroform gave taxol with small amounts of [5-1]
and [5-2],
To obtain further purification of [5-1] and [5-2] the mixture was taken up in 40%
acetonitrile and water and applied to a column of C-18 bonded silica. The column was
eluted with 45 and 50% acetonitrile and water. As the fractions from the 45% acetonitrile
and water elution stood for about a week, crystals appeared over a range of tubes and
these were filtered in groups. Although [5-1] and [5-2] were separated, such that each
contained the other to the extent of 10% or less, recrystallization gave worse mixtures,
thus suggesting that isomerization (or equilibration) was taking place during the process.
Data obtained on a crystalline (9:1 mixture of [5-1] and [5-2]: m.p. 136-138 C, [a]D23
+214 (c 1.04, CHCI3); (lit. Appendino et at. 1992 on an amorphous sample, m.p. 163-
165 C and [a]D23 +185 ).
Analysis calculated for C31H3808, H20: C, 66.89; H, 7.24. Found: C, 66.51; H,
7.19.
The 1H- and 13C NMR spectra of the crystalline [5-1] and [5-2] gave evidence of
mixtures of two compounds. From the spectral data, these two were Identified as a
mixture of 5-0-cinnamoyl-9-acetyl taxicin l [5-1] and 5-0-cinnamoyl-10-acetyl taxicin I [5-
2] described by Chmurney et al. (1993) from Taxus x media Hicksii and by Appendino et

84
al. (1992) from Taxus baccata. The former authors isolated insufficient amounts for
characterization and the latter authors obtained them as amorphous powders by using
HPLC and preparative TLC.
The mixture of [5-1] and [5-2] on acetylation with acetic anhydride and pyridine
gave the acetate, readily obtained as a crystalline solid, m.p. 238-241 C, the NMR
spectrum of which showed that it was a single entity, unlike the starting material. It was
also identical with taxane III (see below).
Taxane III f2-11
The crude crystals of taxane III obtained from the fractions with 50% acetonitrile
and water were filtered and recrystallized from acetone in ligroin to obtain colorless
needles, yield, 0.8 g (0.02%); m.p. 238-241 C, [a]D23 +214 (CHCI3, c 1.04); (lit. +218,
Baxter, 1962); FAB-MS (m/z): 645 (M+ +Na); 623 (M+ + H); 475 [(MH)+ 148
(cinnamoyl)], 415 (475-AcOH); 355 (415-AcOH); 295 (355 AcOH). The spectral data
showed that it is the 5-0-cinnamoyl-2a, 9a, 10p-triacetyl taxicin I (Appendino et at. 1992;
Baxter et at. 1962).
Analysis calculated for C35H42O10, H20: C, 65.61; H, 6.92. Found: C, 66.00; H,
6.72.
Taxane IV [2-2]
This compound also crystallized out directly from the fractions. The crude
crystals were purified by recrystallization from acetone and ligroin, yield, 0.8 g (0.02%);
m.p. 265-267 C; [a]D23 +133(C 0.98, CHCI3); (lit. +137, Baxter et al. 1962); FAB-MS:
607 (MH+); 459 (607 148 (cinnamate)); 399 (459 HOAc); 339 (399 HOAc); 279 (339
- HOAc).
Analysis calculated for C35H4209: C, 69.02; H, 7.03. Found: C, 69.29, H, 6.98.

85
The analytical and spectral data of [2-2] indicated that it was identical with 5-0-
cinnamoyl taxicin II: 2a,9a,1 Op-triacetate described by Appendino et al. (1992) and
Baxter et al. (1962).
Taxol 5-31
In the silica column described above under the purification of compounds [5-1]
and [5-2], taxol (approximately 0.8 g) was eluted by 2-5% methanol in chloroform. A
small portion was crystallized from acetone and ligroin to obtain colorless needles of
taxol. The 1H NMR spectrum showed that the compound still had appreciable quantities
of compounds [5-1] and [5-2]. To remove these compounds completely, ozonization
was carried out on the rest of the sample in chloroform and methanol (9:1, 30 ml) at -70
C for 10-15 min. The reaction mixture was treated with dimethyl sulfide (0.5 ml) and let
stand at room temperature for 2 h.
After concentration to dryness, the sample was chromatographed on normal
phase silica (25 g) in chloroform. Elution with 2% methanol in CHCI3 gave taxol which
was crystallized from ligroin to obtain pure taxol, free from compounds [5-1] and [5-2],
yield, 0.5 g (0.012%). Its spectral properties agreed with those of an authentic sample.
Alternatively, the crude crystalline solid consisting of compounds [5-1] ,[5-2] and
taxol was directly ozonized in chloroform and methanol as before (but without he
intermediate silica column purification). After decomposition of the ozonide, and
concentration, the sample was subjected to chromatography and taxol isolated from the
column. It was crystallized as before to yield 0.75 g (0.015%). The products of
ozonization of compounds [5-1] and [5-2] were more polar than the original compounds
and separated from taxol in the normal phase silica column.

86
Ozonolvsis of Compound [2-21
A solution of compound [2-2] (1 g.) in chloroform and methanol (30 ml, 9:1) was
cooled in a dry ice and acetone bath and saturated with ozone for 10-15 min. TLC
showed that the starting material was absent and ozonide being formed (detected by
spraying with starch and potassium iodide which gave a blue color). After the
decomposition of the ozonide by dimethyl sulfide, the reaction mixture was washed with
water and concentrated to dryness. The product was crystallized from acetone in ligroin
to obtain colorless needles, yield, 0.8 g, m.p. 168-170 C, [a]D23 +130 (c 1.06, pyridine);
HRMS: 569.2239, Calc, for C27H36013, 569.2234.

CHAPTER 6
TAXANE CONSTITUENTS OF TAXUS FLORIDANA
Taxus floridana is a species of Taxus, native to Florida. Its distribution is said to
be limited to a small area along the Apalachicola River. It is a shrub and used frequently
as an ornamental plant. As it is so with the other species of Taxus, the leaves of T.
floridana are also reputed to be toxic to livestock and humans.
During the intensive search to find alternative sources for taxol to replace the
bark of the Pacific yew, many species of Taxus from the United States, Canada, Europe
and Asia were examined. However, there was no study of the taxane constituents of
Taxus floridana. Our laboratory undertook this task to evaluate its usefulness as a
possible source for taxol.
There was also an impetus for this study from another source. In exploring
alternative sources to replace the bark of the Pacific yew, the National Cancer Institute
(NCI) was interested in knowing whether the Taxus plants can be grown under
hydroponic conditions, as opposed to their growing in their natural state. If these plants
can be so grown under hydroponic conditions, which will eliminate the problem of having
to harvest the tree bark, the next question was whether they produce taxol in adequate
yields. Accordingly, the NCI approached our laboratory, and that of Prof. George
Hochmuth Jr. of IFAS, University of Florida, to study this aspect. The hydroponic
cultural techniques were studied by the IFAS laboratory and the isolation and
characterization of taxanes by our laboratory. It was soon found that the two most well-
known species of Taxus, namely T. brevifolia and T. baccata could not be readily
propagated under normal hydroponic conditions, because their growth rate was very
87

88
slow. However, T. floridana responded satisfactorily and could be propagated under
available conditions. This species was therefore studied in our laboratory for its taxane
constituents.
The needles of T. floridana were collected from the campus and were extracted
without drying. After extraction with methanol as before, concentration to remove the
solvent, and partition between chloroform and water, the organic layer was concentrated
to a dark green semisolid. Fractionation was again carried out using the reverse phase
column techniques as was described under the needles of the other Taxus species in
Chapters 3 and 5.
The crude chloroform extract was first tested by analytical HPLC to see the
elution pattern of the taxane constituents. Taxol was clearly recognizable at its normal
location, and in contrast to the observation with the needles of Taxus x media Hicksii,
where there were co-eluting taxanes, the taxol from the extract of the needles of T.
floridana was relatively free from such interfering taxanes. There were other taxanes
situated at different locations.
Elution of the reverse phase column was carried out using a step gradient of 30,
35, 40, 45, 50 and 60% acetonitrile/ water. When the fractions were let stand at room
temperature for 3-5 days, taxol and several other taxanes crystallized out as before.
The initial eluates from the column from 25-30% acetonitrile/ water contained
highly polar phenolic constituents. The first taxane component to appear from the
reverse phase column emerged with the 30-35% acetonitrile/water solvent, and
crystallized almost immediately. This was found to be 10-deacetyl baccatin III [2-7], The
next taxane was eluted with the 40% acetonitrile/ water and it was found to be identical
with brevifoliol [3-1], With the 45% acetonirile/ water, was eluted another crystalline
compound which was found to be a new compound, and was named taxiflorine [6-1].
Continued elution with 50% acetonitrile/ water gave two crystalline compounds in

89
succession. One of these was identified as baccatin VI [6-2], and the second one was
taxol [5-3],
Taxiflorine
Taxiflorine [6-1] was readily obtained as a colorless crystalline solid. Its
elemental analysis agreed with the molecular formula C35H44013. Its 1H NMR spectrum
in CDCI3 showed broad and rounded peaks with poor resolution. In DMSO-d6, the
spectrum gave sharper signals but showed double the number of peaks in certain
positions. The 13C spectrum also exhibited extra peaks, which suggested that the
compound was a mixture of rotamers in equilibrium. One could infer the presence of
ester functions from the spectra, with four acetates and one benzoate, and an oxetane
ring.
Acetylation of taxiflorine gave a monoacetate [6-3], which gave sharp signals in
its 1H NMR spectrum, with the expected number of peaks, thus showing that it is a single
compound, unlike the starting material. Although the acetate was isomeric with baccatin
VI, it was different. The most striking difference between the two spectra was seen with
the signal for the H-13. In the acetate of taxiflorine, this signal was at 6 5.60, while the
same was found at 8 6.3 in baccatin VI. A comparison with other related taxanes
showed that in those with the 6-membered A-ring, the H-13 signal appears at 8 6.2-6.5,
whereas in taxanes with a 5-membered A-ring, as in the 11(15->1)-abeotaxanes, it
appears at 8 5.4-5.7 (Appendino et al 1993B).
Positions 9 and 10 in taxiflorine carry a free hydroxyl and a benzoate function.
To locate the benzoate, a comparison of the signals due to H-9 and H-10 in taxiflorine
were compared with the corresponding signals in the monoacetate. With the two signals
at 8 6.30 and 8 5.90 in taxiflorine, the latter undergoes a down-field shift from 8 5.90 to

90
8 6-20, whereas the peak at 8 6.30 remains essentially unchanged (8 6.37). With the
reasoning that the allylic H-10 must be more down-field than H-9, the signal at 8 6.30
may be assigned to H-9 and the one at 8 5.90 for the H-10. This leads to the
assignment of the structure of taxiflorine as [6-1], with the hydroxyl at 9 and the
benzoate at 10. Based on the 1H,1H-COSY and 1H,13C-HETCOR spectral data, the four
acetate functions were assigned as 2a, 4a, 7p and 13a.
Benzoylation of [6-1] was carried out to yield the monobenzoate [6-4], which also
gave a 1H NMR spectrum that indicated that it was a single entity. Taxiflorine was also
saponified to the heptaol and re-acetylated to the hexa-acetate [6-5], The 1H NMR
spectra of [6-3], [6-4] and [6-5] are shown in Table 6-1.
The structure of [6-1] with the hydroxyl at C-10 and benzoate at C-9 has the
potential for intramolecular transesterification occurring between the 9-benzoate, as well
as the 7-acetate (Lewis et al. 1993). The fact that the monoacetate and the
monobenzoate gave single products discounted the possibility that transesterification
was responsible for the anomalous NMR spectra of [6-1], To verify if the appearance of
the spectrum is due to an rotameric equilibrium, the spectrum was taken in DMSO-d6 at
temperatures between -20 and 100. At lower temperatures, the spectra were sharper
and showed two sets of peaks for some protons. At higher temperatures, the peaks
coalesced into a single set, as well as became broad to the extent that they were barely
seen. This behavior suggested that the conformational equilibrium between the
rotamers is responsible and that the presence of the 10-OH facilitates this process. As
indicated earlier, during the exploration of possible alternative sources for taxol to
replace the bark of the Pacific yew, scant, if any attention was paid to Taxus floridana.
The species T. x media Hicksii (the ornamental yew) was selected as a possible source
for taxanes in the future.

91
Table 6-1 : Proton NMR Spectra of Compounds [6-3], [6-4] and [6-5],
H#
Compound [6-3]
Compound [6-4]
Compound [6-5]
2
6.19, d, J=7.8 Hz
6.26, d, J=7.8 Hz
6.07, d, J=7.8 Hz
3
2.99, d, J=7.8 Hz
3.06, d, J=7.8 Hz
2.92, d, J=7.8 Hz
5
4.98, d, J=7.5 Hz
5.01, d, J=7.5 Hz
4.98, d, J=7.5 Hz
6a
2.68, m
2.70, m
2.52, m
6p
1.84, m
1.84, m
1.84, m
7
5.52, m
5.64, m
5.49, t, J=7.8 Hz
9
6.32, d, J=10.8HZ
6.48, d, J=10.8Hz
6.04, d, J=10.8Hz
10
6.37, d, J=10.8 Hz
6.72, d, J=10.8 Hz
6.27, d, J=10.8, Hz
13
5.62, t, J=7.8Hz
5.64, m
5.61, t, J=7.8Hz
14a
2.30, dd,J=7.4,14.2Hz
2.34, dd,
2.30, m
14(3
1.72, dd,J=7.4,14.2Hz
1.78, m
1.72, m
16
1.16, s
1.24, s
1.15, s
17
1.19, s
1.21, s
1.13, s
18
1.72, s
1.72, s
1.83, s
19
1.64, s
1.95, s
1.66, s
20
4.50, 4.42, d, J=7.9 Hz
4.52, 4.44, d, J=7.2
4.47,4.38, d,J=7.5
Ph(2,6)
7.93, d
Unresolved

Ph(3,5)
7.45, t
7.24, mm

Ph(4)
7.62, t
7.37, m

Ph(2,6)

7.63, d, J=7.2Hz
Ph(3,5)

7.24, m
Ph(4)

7.37, m

OAc Me
2.02,s
2.14, s
2.11,s
OAc Me
2.14, s, (2x)
2.05, s
2.10, s
OAc Me
1.86, s

2.08, s
OAc Me
1.80, s

2.03, s
OAc Me


2.01, s
OAc Me


1.95, s
1H NMR were recorded at 600 MHz in CDCI3 on a Varian Unity 600 spectrometer
at ambient temperature. Chemical shifts 5 (ppm) are reported relative to TMS as an
internal standard.

92
[6-1] Taxiflorine
R-i = H R2 = C6H5CO
[6-3] Taxiflorine Acetate
R-i =CH3CO R2 = C6H5CO
[6-4] Taxiflorine Benzoate
Ri = R2 = C0H5CO
[6-5] Hexa-Acetate
R-i = R2 = CH3CO
Figure 6-1 : Taxanes and Analogues from Taxus x media Hicksii

93
From the present studies, which compared the taxane composition of both
species, it became clear that T. floridana would have been a much better choice, for the
reasons given below:
1. Taxol is isolated more easily from the Florida yew than from the ornamental yew,
because, in the former there are no co-eluting taxanes to interfere and which have to
be removed as an additional step. The yields of taxol (0.01% from fresh leaves) are
accordingly better with the Florida yew, than from that of the ornamental yew
(0.015% from dried leaves).
2. Besides taxol, the Florida yew gives relatively high yields (0.05-0.06% from fresh
leaves) of 10-deacetyl baccatin III, which is the most commonly used intermediate for
the semi-synthesis of taxol. This compound is present in the ornamental yew in
exceedingly low concentrations.
3. The other components that contain the oxetane ring, such as baccatin VI and
taxiflorine, are also useful compounds for the synthesis of taxol analogues. In
contrast, the taxanes from the ornamental yew that are 11,4/20 diene taxanes have
no currently documented use.
The work described here is the first definitive investigation of the taxane
constituents of Taxus floridana on a preparative scale giving the actual recovered yields
of the crystalline compounds.
Experimental
Extraction
The needles and small twigs of Taxus floridana were collected from several
bushes growing at different locations of the campus of the University of Florida.
Likewise, they were made available from the plants growing under hydroponic

94
conditions. For convenience, they were extracted in fresh state. Several batches were
collected from the plants on the campus ranging from 1-20 Kg. With the hydroponically
grown plants, the smallest amount was 54 g of the fresh needles, and the largest, 2.5
Kg. The method of extraction, concentration, partition between water and chloroform
were the same as was described in Chapter 3. The yield of the chloroform extract varied
from 20-25 g per Kg of the fresh leaves and twigs.
Before the conditions for the reverse phase column chromatography were fully
developed, an alternative procedure was tested for the purpose of making the sample
preparation easier. The chloroform extract, especially of the needles, usually contains a
higher amount of waxes, chlorophylls and other lipid-soluble components and can
potentially pose problems in the preparation of the sample for applying to the column.
For this reason, a study was made to see at what concentration of methanol or
acetonitrile would be needed to obtain an essentially clear solution that can be applied to
the column.
It was found that at least an 80% methanol in water would be necessary to
prepare a 10% solution of the chloroform extract solids. However, at this concentration
of the solvent, taxol and most of the other taxanes do not remain on the column. It was
also found that if such a solution is passed through a column of C-18 bonded silica,
almost all of the chlorophylls, waxes etc., remain on the column, while taxol and other
taxanes appear in the effluent and washes. There was also a reduction of the solids
content by 50-60%, which meant that the waxes and other lipid-soluble components
account for this much of the chloroform extract and can be readily removed from the
sample.
The material obtained by concentrating the effluent and washes was much less
difficult to prepare as the slurried sample for applying to the column. In order to carry
out this wax-removal operation, it was only necessary to use a ratio of 3 g of the C-18

95
silica for 1 g of the extract. The chlorophylls and waxes that were held up on the column
could be readily removed by washing with a mixture or methanol/ ethyl acetate/ ligroin
(2:1:1). Examination by TLC showed that this wash did not contain any taxane
constituents. This procedure was not used in later trials, as methods were found for a
successful and convenient preparation of the sample slurry made it unnecessary, as
described in Chapter 3.
Characterization of the Taxane Constituents of Taxus floridana
The results given here represent the work carried out on a 20-Kg batch of the
fresh needles.
10-Deacetvl Baccatin III [2-71:
Elution with 30% acetonitrile in water gave this component which crystallized
almost immediately. After a week, the crystals were filtered off, dried and recrystallized
from methanol/ chloroform, yield, 12 g (0.06%); m.p. 232-234 C. The spectral
properties were identical with those described in the literature (Chauviere et at. 1981,
Appendino et at. 1993b).
Brevifoliol [3-11:
Fractions from the 35-40% acetonitrile eluate, which contained this component
but did not give a crystalline solid directly, were combined, concentrated to dryness and
the solid (3 g) was applied to a normal phase silica column (120 g) in chloroform.
Elution with 2% methanol in chloroform gave the major band, the fractions from which
were combined, concentrated and the solid crystallized from acetone / ligroin to give 1 g
of [3-1], m.p. 220-222 C. Its spectral data proved to be identical with those described in
Chapter 3.

96
Taxiflorine f6-11:
The crude crystalline solid (2.5 g) that separated from the fractions from 45%
acetonitrile/ water was filtered and purified by recrystallization from acetone/ ligroin to
give [6-1] as a colorless crystalline solid, yield, 1.2 g (0.006%); m.p. 254-255 C, [a]D23
-26.1.
Analysis calculated for C35H44013: C, 62.48; H, 6.59. Found: C, 62.12; H, 6.63.
Baccatin VI [6-2]:
Eluates from the 50% acetonitrile/ water gave crystals in a number of fractions.
These were filtered into groups and tested by TLC and analytical HPLC. The earlier
fractions contained mostly baccatin VI, with gradually increasing amounts of a slower
compound, shown to be taxol. The crystals from the first group containing mostly [6-3]
(3.5 g) were dissolved in chloroform (50 ml) and passed through a column of Florisil (20
g) for the purpose of decolorization. The effluents and washes were combined,
concentrated to dryness and the solid crystallized from acetone/ ligroin to give pure [6-3]
yield, 1.6 g. Together with the amount obtained from the next fraction, the total yield
was 1.95 g (0.01%); m.p. 250-252 0 C (lit. 248-250 0 C decomp., Senilh et al. 1984);
[a]D23 -11 (chloroform, c 0.98) (lit. -5, chloroform, c 1.3, Senilh et al. 1984); MS(FAB);
737 [M+Na]\ 697 [M-HzO]+.
1H NMR (CDCI3i 300 MHz, 8): 1.22 (17-Me, s); 1.60 (19-Me, s); 1.78 (16-Me, s);
1.87 (6a, cm); 1.99 (OAc-Me, s); 2.02 (OAc-Me, s); 2.04 (C-14P, unresolved mult.); 2.10
(18-Me, s); 2.10 (OAc-Me, s); 2.19 (OAc-Me, s); 2.20 (C-14a, unresolved mult.); 2.28
(OAc-Me, s); 2.50 (C-6(3, cm); 3.18 (C-3, d, 6 Hz); 4.13 (C-20, d, 8.4 Hz); 4.34 (C-20, d,
8.4 Hz); 4.97 (C-5, d, 8.4 Hz); 5.55 (C-7, dd, 7.5, 9.3 Hz); 5.87 (C-2, d, 6.0 Hz); 5.99 (C-
9, d, 11.1 Hz); 6.17 (C-13, dd, 7.6, 9.3 Hz); 6.22 (C-10, d, 11.1 Hz): 7.48 (Ar-meta, t, 7.8
Hz); 7.61 (Ar-para, t, 7.5 Hz); 8.09 (Ar-ortho, dd, 7.2, 1.3 Hz).

170.4-AcCO 83.7
170.1-AcCO 81.4
169.8-AcCO 78.7
169.1-AcCO 76.3
168.8-AcCO 74.9
166.8-BzCO 73.1
141.1-12 71.7
133.6-Bz(p) 70.3
133.5- 11 69.6
129.9 -Bz(m) 47.2
129.2-Bz(q) 45.7
128.5-Bz(o) 42.7
5 35.0- 14*
4 34.4 6*
2 28.2- 17 Me
20 22.6-Ac Me
1 22.2-16 Me
7 21.3-Ac Me
9 21.1-AcMe
13 20.8-Ac Me
13 20.7-Ac Me
3 14.9 18 Me
8 12.7 19 Me
15 Interchangeable
Figure 6-2 : Carbon NMR Spectrum of Baccatin VI

98
13C NMR (CDCI3i 300.075 MHz, 8): 75.0 (C-1); 78.8 (C-2); 47.5 (C-3); 81.6 (C-4);
83.8 (C-5); 34.3 (C-6); 73.3 (C-7); 45.8 (C-8); 71.9 (C-9); 69.7 (C-10); 133.5 (C-11);
142.0 (C-12); 70.4 (C-13); 35.3 (C-14); 42.8 (C-15); 22.4 (C-16); 28.3 (C-17); 12.9 (C-
18); 15.0 (C-19); 76.5 (C-20); 20.7; 20.9, 21.2, 21.4, 22.7 (5x OAc Me); 168.6, 168.9,
169.6, 169.9, 170.2 (5X OAc CO); 166.8 (Bz CO); 129.2 (Bz ipso); 128.5 (Bz ortho);
129.9 (Bz meta); 133.5 (Bz para).
Analysis calculated for CsrH^O^: C, 62.18; H, 6.49. Found: C, 61.83; H, 6.45.
Taxol [5-31:
Crystals (4.5 g) from the second part of the peak which contained mostly taxol.
were combined dissolved in chloroform (60 ml) and chromatographed on Florisil(40 g).
Elution with chloroform gave more of [6-3] and subsequent elution with 5% acetone in
chloroform gave taxol, which was recovered by concentration of the appropriate fractions
and crystallized from acetone and ligroin to obtain taxol, yield, 1.98 g (0.01%); m.p. 220-
222 0 C. The spectral and chromatographic properties of the sample agreed with those
of taxol.
Acetylation of Taxiflorine to [6-31:
An aliquot of [6-1] (0.05 g) was acetylated by acetic anhydride (2 ml) and pyridine
(0.5 ml) at room temperature for 16 h. Water was added, the solid filtered and purified
by chromatography on a silica column using chloroform / ligroin (2:1) to obtain [6-3] as a
white powder.
Benzoylation of Taxiflorine to [6-4]:
To a solution of [6-1] (0.05 g) in pyridine (2 ml) was added dropwise with stirring
at 0-5 C, benzoyl chloride (0.1 ml). After 20 h water was added followed by 2N sulfuric
acid and the solid filtered. The product was purified by chromatography as given under
[6-3], to obtain [6-4] as a white powder.

99
Saponification and Acetylation of [6-11 to [6-51:
A solution of taxiflorine (0.1 g) in methanol (5 ml) was treated with 2N potassium
hydroxide (1 ml) and the mixture let stand at room temperature for 2 h. TLC showed that
the starting material was no longer present, along with the appearance of a very slow
moving component. The solution was acidified and extracted with chloroform (3x) and
the combined extracts concentrated to dryness. The residue was dissolved in acetic
anhydride (2 ml) and pyridine (0.5 ml) and let stand for 16 h. Addition of water, filtration
of the solid and chromatography on a normal phase silica gave [6-5] as a white powder.

CHAPTER 7
NON-TAXANE COMPONENTS FROM THE BARK AND NEEDLE EXTRACTS
General
Some of the benefits of using reverse phase rather than normal phase
chromatography have been described in previous chapters. Two important
disadvantages of normal phase silica gel chromatography are the acidic nature of silica
and the tendency for irreversible adsorption to occur. Both of these problems can lead
to significant loss of the compound(s) of interest. Fortunately, almost all free acidic
groups are capped during the bonding process used to make reverse phase silica,
followed by a final capping with trimethylsilyl groups. This process effectively eliminates
these problems of acidity and irreversible adsorption.
These properties allow recovery of many compounds that would normally not be
amenable to silica gel chromatography. Glycosides, phenols, steroids and hydrophilic
compounds are often difficult to chromatograph using normal phase silica gel. During
the processing thousands of pounds of bark and lesser amounts of needles many
interesting compounds were isolated. The taxane glycosides are most notable among
these, especially now that the efficient removal of the glycosyl moiety has become
possible (Rao, 1997). The relative abundance of taxane glycosides amenable to
conversion into taxol gives further support to the use of reverse phase columns. Large
amounts of valuable precursors are lost with the normal Polysciences isolation process
(Boettner et al. 1979).
100

101
Flavonoids are chemicals generally found in plants that are ubiquitous and have
been studied for hundreds of years. These compounds are generally yellow to red in
color and are usually responsible for the colors seen in flowers and plants. Research
Into the possibility that flavonoids might possess useful biological activities has
undergone a renaissance in recent years, after decades of neglect. Quercetin and its
most common glycoside (rutin) are probably the most ubiquitous of the flavonoids, and
have been used for many years to enhance immune response to pathogens.
Bioflavonoid complex is mostly rutin and another flavonoid known as hesperidin, and
can be purchased in most drug stores and herbal shops.
These flavonoid compounds were easily purified on the reverse phase columns.
The future usefulness of flavonoids as bioactive compounds remains to be seen, but the
evidence is growing. Some of these compounds have been reported to inhibit the
reverse transcriptase enzyme of HIV virus in vitro, but further work is needed concerning
the mechanism.
Insect molting hormones, commonly classed as edcysones, are responsible for
the maturation of larvae into adult form and have interested scientists for many decades.
Originally, silk worms were extracted to obtain these compounds for research and in
very low yields indeed. Efforts to find better sources for these compounds led to their
discovery in some plants, including Taxus baccata (Hoffmeister 1966). Substantial
amounts of p-ecdysterone, ponasterone A and other ecdysones were easily isolated
during the workup of fractions from the reverse phase columns. All samples tested in
this project had substantial amounts of these hormones, and were easy to isolate.
Usnic acid is a bright yellow compound which is known to grow in lichens. Due
to its structure all signals in the proton NMR are singlets. It is commonly classed as a
tricarbonyl and has peaks as high as 18.8, 13.3 and 11 ppm. An endophytic fungus of
the pacific yew known as Taxomyces andreanae is capable of producing taxol in cell

102
culture (Stierle et al. 1993). It is a mystery how these two organisms that are so different
are able to make a complex diterpene like taxol. In spite of this taxol producing fungus,
it is most likely that the usnic acid was actually produced by lichens growing on the bark,
and is not produced by the tree. Usnic acid is mentioned here because it was difficult to
characterize and not uncommon in samples processed in this lab. Usnic acid is also
known to be quite toxic and should be handled carefully.
Betuloside is a simple glycoside first isolated from the plant Betula pendata
(Khan 1966). Animal studies using hepatotoxic agents indicate that betuloside has
significant hepatoprotective activity. The mechanism by which this compound protects
the liver is not known, but teas made from plants containing betuloside have been used
in India for centuries for various problems. Betuloside is just one more example of the
usefulness of preparative scale reverse phase chromatography.
Experimental
Analytical HPLC was performed using two different systems. For determinations
of purity and quantitative information on composition, a setup with a Waters 600 E pump
with gradient control, a Waters 996 photodiode array detector, and a Waters 717
autosampler, coupled with an NEC-386 computer and printer was used. Waters
Millennium 1.1 program was used with the photodiode array system. For routine use, a
combination of a Waters 501 pump with a U6K injector, a 486 tunable absorbance
detector and a Goerz Servogor 120 recorder was used.
For analytical purposes, standard columns packed with C-8 bonded silica
(Whatman Partisil, 4.6 mm x 25 cm, 5pi) were used with either of the solvents: 50%
acetonitrile in water, or a 5:4:1 mixture of acetonitrile, water and methanol.
For preparative scale purposes, stainless steel columns of two sizes were used:
4 x 4 and 6 x 6, fabricated by Fluitron Inc. (Ivyland, PA) and rated to 200 psi. The

103
columns were packed with C-18 or C-8 bonded silica gel (Spherisorb, 15-35 p diameter,
Phase Separations Inc., Norwalk, CT) as a slurry in methanol After thorough washing
with methanol, the columns were equilibrated with 25% acetonitrile in water.
Thin-layer chromatography was carried out using silica gel HF-60, 254+366 (EM
Science/Fisher). Visualization was by a UV-lamp and by charring with 1 N H2S04.
Column chromatography was performed using silica gel (Fisher, 100-200 and 235-425
mesh) or Florisil (Fisher F-101, 100 mesh) with a solvent sequence consisting of
ligroin/CHCI3i CHCI3, 2-5% acetone and finally, 2-10% MeOPI in CHCI3.
Melting points were determined on a Fisher-Johns hot stage apparatus and are
uncorrected. The following instrumentation was used for the spectra recorded here:
UV, Perkin Elmer A.3B; IR, Perkin Elmer PE-1420; and NMR, General Electric QE-300,
Nicolet NY-300, Varan VXR-300, Varian Gemini-300 and Vahan Unity-600
spectrometers. Mass spectra (FAB) were obtained on a Finnegan Mat 95Q
spectrometer using a cesium gun operated at 15 KeV of energy.
Flavonoids
Quercetin Rutoside (Rutin)
3-[[6-0-(6-deoxy-a-L-mannopyranosyl)-p-D-glucopyranosyl]oxy]-2-(3,4-
dihydroxyphenyl)-5,7-dihydroxy-4H-1-benzopyran-4-one. m.p. : 186-189 C (dec., turns
brown at~127C); [a]D23 +14.06 (ethanol, c 1.02); [a]D23 -39.76 (pyridine, c 1.06).
1H NMR (DMSO-d0, 300 MFIz, 8): 12.6 (C-5 hydroxyl, br s, D20 exchangeable);
7.57 (H-21, d, 2.2); 7.53 (H-6, dd, 8.6, 2.2); 6.86 (H-5, d, 8.3); 6.40 (H-8, br s); 6.21 (H-6,
br s); 5.34 (H-1 glucosyl, br s); 4.42 (H-7 rhamnosyl, br s); 3.1-3.75 (CH-OPI glycosyl,
6H, mult.); 1.03 (H-12 methyl, d, 6.0).

104
13C NMR (DMSO-d6, 300 MHz, 5): 177.5 (C-4, carbonyl); 164.1 (C-7, s); 161.4
(C-5, s); 156.9 (C-2, s); 156.7 (C-9, s); 148.5 (C-4, s); 144.8 (C-3,s); 133.5 (03, s);
121.9 (C-6\ d); 121.4 (C-1,s); 116.4 (C-2,d); 115.4 (C-5,d); 104.2 (C-10, s) ;100.6 (C-
1 ,d); 98.6 (C-6,d); 93.5 (C-8, d); 76.5; 76.0; 74.2; 71.9; 70.6; 70.55; 70.49; 70.1; 68.4;
67.2 (C-6, t); 18.0 (C-12 methyl, q).
IR, \/ max (KBr, cm'1): 3340 (OH, bonded); 2920 (CH stretch); 1655 (C=0); 1620
(C=C); 1510 (aromatic); 1355 (C-O-C); 1290 (C-O-C); 1200 (C-O-C); 1055 (C-O-C); 970,
880, 810 (subst. aromatic); 730, 695.
Analysis calculated for C27H3o016: C, 53.12; H, 4.95. Found: C, 52.88; H, 5.06.
Quercetin
2-(3,4-dihydroxyphenyl)-3,5,7-trihydroxy-4H-1-benzopyran-4-one or 3,3,4, 5,7-
pentahydroxyflavone. Refluxed 300 mg of quercetin glycoside in 2 N H2S04 for 3 hours,
filtered, washed with water, yellow needles crystallized from aqueous ethanol, 150 mg
(77%).
1H NMR (DMSO-de, 300 MHz, 6): 12.5 (C-5 hydroxyl, br s, exchangeable with
D20); 7.68 (H-2\ d, J=2.2); 7.54 (H-6\ dd, J= 2.0 and 8.4); 6.90 (H-5\ d, J=8.4); 6.41
(H-8, d, J=2.0); 6.19 (H-6, d, J=2.4).
13C NMR (DMSO-d6, 300 MHz, 5): 175.7 (C-4, s); 163.8 (C-7, s); 160.7 (C-5, s);
156.9 (C-9, s); 147.6 (C-4, s); 146.9 (C-2, s); 145.0 (C-3, s); 122.0 (C-1, s); 120.0 (C-
6, d); 115.6 (C-5\ d); 115.3 (C-2, d); 103.0 (C-10, s); 98.2 (C-6, d); 93.4 (C-8, d).
Analysis calculated for C15H10O7+ 2 H20: C, 53.26; H, 4.17. Found: C, 53.64; H,
4.04.
Sciadopitysin
This biflavonoid is a member of the amentoflavone group, m.p. : 302-304 C.

105
1H NMR (DMSO-d6, 300 MHz, 5): 13.05 (1H, s, 5-OH); 12.90 (1H, s, 5-OH); 8.31
(1H, s, 7-OH); 8.17 (1H, dd, J=9.0, 2.4, 6-H); 8.08 (1H, d, J=2.4, 2-H); 7.60 (2H, d,
J=9.0, 2 H, 6 H); 7.37 (1H, d, J=8.7, H-5); 6.98 (1H, s. H-3); 6.93 (2H, d, J=9, H-3,
H-5); 6.89 (1H, s); 6.78 (1H, d, J=2.4, H-8); 6.42 (1H, s, H-6); 6.36 (1H, s, 6~H); 3.83
(3H, s); 3.80 (3H, s); 3.76 (3H, s).
13C NMR (DMSO-d6, 300 MHz, 5): 181.9; 181.8; 165.1; 163.5; 162.9; 162.1;
161.0; 160.5; 160.4; 157.2; 154.2; 130.8; 128.2; 127.7; 122.7; 122.3; 121.5; 114.4;
111.6; 104.1; 103.7; 103.55; 103.48; 103.1; 98.6; 98.0; 92.6; 55.9; 55.8; 55.4.
IR, v max (KBr, cm"1) : 1660, 1650, 1620, 1610, 1570, 1510, 1430, 1370, 1240,
1180, 1160, 1050, 1030, 960, 910, 880, 830, 760.
Analysis calculated for C33H24O10: C, 68.27; H, 4.17. Found: C, 67.94; H, 4.26.
B-Sitosterol-B-D-Glucoside
Identity of glycoside confirmed using authentic sample with mixed melting point,
TLC, and IR spectrum, as well as 1H- and 13C NMR of the tetra-acetate, the aglycone,
and the aglycone acetate. M.p. : 288-290 0 C (lit. varies from 280-300 0 C, 298 0 C
Sucrow 1966). EI-MS: 414(2%) [M+ glucosyl], 396(34%) [MH+ glucosyl H20]. Cl-
MS (methane): 413(9%) [MH+ glucosyl H2], 397(100%) [MH+ glucosyl H20],
[a]D23 -40.1 (pyridine, c 1.1); (lit. -41.0, c 1.33, pyridine, Swift 1952).
IR v max (KBr, cm'1) : 3400 (-OH); 3090 (C=CH2); 1650, 890.
Analysis calculated for C35H6o06: C, 72.87; H, 10.48. Found: C, 72.48; H, 10.61.
B-Sitosterol-B-D-Glucoside Tetra-acetate
Acetylation of p-sitosterol-p-D-glucoside: Dissolved 500 mg in 5 ml acetic
anhydride with 0.1 ml pyridine, then placed in hot water bath for 1 hour with stirring.
After TLC indicated the reaction was complete the mixture was stirred with water for 15
minutes to decompose the anhydride, then extracted with chloroform 3 x at pH 4. Dried

106
over Na2S04, concentrated, then crystallized from ethyl acetate in ligroln, yield 440 mg
(68%); first crop.
M.p.: 165-167 0 C (lit. 171 0 C, Sucrow 1966). [a]D23 -35.0 (pyridine, c 1.3); (lit.
pyridine, c 1.33,-33.7, Swift 1952).
1H NMR (CDCI3, 300 MHz, 5): 5.37 (H-6, br d, J= 4.8); 5.21 (H-3, t, J=9.3); 5.08
(H-4, t, J=9.2); 4.95 (H-2, dd, J=9.3 and 8.1); 4.60 (H-1'f d, J=8.1); 4.26 (H-6b, dd,
J=4.8 and 12); 4.11 (H-6a, dd, J=2.4, 12); 3.70 (H-5, cm); 3.5 (H-3, cm); 2.07, 2.05,
2.02, 2.00 (4 X Ac CH3, s); 1.0 (H3C-19, s); 0.84 s; 0.82 s; 0.68 (H3C-18, s).
13C NMR (CDCI3, 75 MHz, 5): 170.6, 170.3, 169.3, 169.2 (4 X AcC=0, s); 140.3
(C-5, s); 122.1 (C-6, d); 99.6 (C-1\ d); 80.0 (C-3, s); 72.9 (C-3, d); 71.7 (C-5, d); 71.5
(C-2', d); 68.5 (C-4, d); 62.1 (C-6, t); 56.7 (C-14, d); 56.0 (C-17, d); 50.1 (C-9, s); 45.8
(C-24, s); 42.3 (C-13, s); 39.7 (C-12, t); 38.9 (C-4, d); 37.2 (C-1, d); 36.7 (C-10, s); 36.1
(C-20, d); 33.9 (C-22, t); 31.9 (C-8, d); 31.8 (C-7, t); 29.4 (C-2, t); 29.1 (C-25, d); 28.2 (C-
16, t); 26.1 (C-23, t); 24.3 (C-15, d); 23.0 (C-28, t); 21 0(C-11, d); 20.6, 20.5, 20.44,
20.42 (4 X AcMe); 19.7 (C-27, q); 19.3 (C-19, q); 19.0 (C-26, q); 18.7 (C-21, q); 11.8 (C-
18, q); 11.8 (C-29, q).
IR, v max (KBr, cm'1): 1755, 1220, 1045, 910.
B-Sitosterol
M.p.: 138-140 0 C (lit. 137-138 0 C, Swift 1952); [a]D23 -37.2 (chloroform, c 1.1);
(lit. -38.2, chloroform, c 5.1, Swift 1952).
1H NMR (CDCI3, 300 MHz, 5): 5.36 (H-6, d); 3.50 (H-3, m); 2.30 m, 2.0 m, 1.0 s,
0.82 d, 0.7 s.
13C NMR (CDCI3, 75 MHz, 5): 140.7 (C-5, s); 121.6 (C-6, d); 71.8 (C-3, s); 56.7
(C-14, d); 56.0 (C-17, d); 50.1 (C-9, d); 45.8 (C-24, d); 42.3* (C-13, s); 43.2 (C-4, t); 39.7
(C-12, t); 37.2 (C-1, t); 36.4 (C-10, s); 36.1 (C-20, d); 33.9 (C-22, t); 31.9 (C-8, d); 31.6

107
(C-2, t); 29.2 (C-25, q); 28.2 (C-16, t); 26.1 (C-23, t); 24.3 (C-15, q); 23.0 (C-28, t); 21.0
(C-11, t); 19.8 (C-27, q); 19.3 (C-19, q); 19.0 (C-26, q); 18.7 (C-21, q); 11.8 (C-18, q);
11.9 (C-29, q).
Analysis calculated for C29H50O: C, 83.99; H, 12.15. Found: C, 83.16; H, 12.42.
Phytoecdysteroids
Ecdysterone & 2p, 3B. 22a-Triacetate
(22R)-2p,3p,14a,20p,22a,25-hexahydroxycholest-7-en-6-one, m.p.: 237-240 0 C
(Lit. 240 0 C, Takemoto etal. 1967).
1H NMR (CDCI3, Triacetate, 8): 5.85 (1H, s, H-7); 5.31 (1H, br s, H-3a); 5.04 (1H,
dt=9, 3 Hz, H-2a); 4.79 (1H, d=9 Hz, H-22p); 3.10 (1H, br t=7.8 Hz, H-9a); 2.10 (6H, s,
2X OAc); 1.99 (3H, s, OAc); 1.26 (3H, s, Me-21); 1.23 (3H, s, Me-26*); 1.21 (3H, s, C-
27); 1.04 (3H, s, Me-19); 0.85 (3H, s, Me-18).
13C NMR (CDCI3, Triacetate, 5): 202.1 (C-6); 172.2 (OAc); 170.3 (OAc); 170.0
(OAc); 164.8 (C-8); 121.6 (C-7); 84.4 (C-14); 79.8 (C-22); 77.0 (C-20); 70.4 (C-25); 68.7
(C-3); 67.2 (C-2); 51.0 (C-5); 49.6 (C-17); 47.6 (C-13); 40.4 (C-24); 38.4 (C-1*); 38.3 (C-
10*); 34.1 (C-4); 33.7 (C-9); 31.5 (C-15); 31.2 (C-12); 30.2 (C-27); 29.2 (C-16); 28.5 (C-
26); 24.7 (C-23); 23.8 (C-19); 20.6 (C-211); 20.5 (C-111); 17.5 (C-18).
IR, v max (KBr, cm'1): : 3400, 2960, 2870, 1643, 1450, 1370, 1050, 880.
UV max (ethanol): 243 nm (e 10,400)
Analysis calculated for C27H44O7: C, 67.47; H, 9.23. Found: C, 67.08; H, 9.30.
Ponasterone A and 2B. 38, 22a Triacetate
(22R)-2p, 3p, 14a, 20p, 22a-pentahydroxycholest-7-en-6-one. M.p. : 256-262 C
(Lit. 259-260 0 C, Nakanishi et a/. 1966)

108
1H NMR (CDCI3, Triacetate, 5): 5.87 (1H, s, H-7); 5.31 (1H, br s, H-3a); 5.04 (1H,
dt=9, 3 Hz, H-2a); 4.84 (1H, d=9 Hz, H-22P); 3.12 (1H, br t=7.8 Hz, H-9a); 2.11 (6H, s,
2X OAc); 2.01 (3H, s, OAc); 1.25 (3H, s, Me-21); 1.03 (3H, s, Me-19); 0.89 (3H, d=2.7
Hz, Me-26*); 0.87 (3H, d=2.7 Hz, Me-27*); 0.85 (3H, s, Me-18).
13C NMR (CDCI3, Triacetate, 8): 201.9 (C-6); 172.3 (OAc); 170.4 (OAc); 170.1
(OAc); 164.7 (C-8); 121.5 (C-7); 84.4 (C-14); 79.5 (C-22); 76.9 (C-20); 68.7 (C-3); 67.1
(C-2); 50.9 (C-5); 49.6 (C-17); 47.5 (C-13); 38.3 (C-1); 35.7 (C-10); 34.1 (C-4); 33.6 (C-
9); 31.5 (C-15); 31.2 (C-12); 29.2 (C-16); 27.9 (C-23); 27.7 (C-25); 23.8 (C-21); 22.9 (C-
19); 22.1, 21.1 (C-27*); 21.0 (C-26*); 20.6 (C-24f); 20.5 (C-111); 17.4 (C-18).
IR, v max (KBr, cm'1) : 3400, 2960, 2870, 1643, 1450, 1380, 1050, 870.
Analysis calculated for C27H4406: C, 69.79; H, 9.54. Found: C, 69.41; H, 9.72.
Phenolic Compounds
Usnic Acid
m.p. : 208-213 C, yellow orthorhombic prisms from ligroin/ethyl acetate (lit. 204,
acetone, Schopf & Ross, 1938). [a]D25 -510, CHCI3, c 0.62 (lit. -509, CHCI3, c 0.679,
Schopf & Ross, 1938).
1H NMR (CDCI3, 300 MHz, 5): 1.76 (3H, s, C-4 angular methyl); 2.10 (3H, s, C-6
aromatic methyl); 2.67 (6H, s, C-3 & C-8 acetyl methyls); 5.97 (1H, s, C-1); 11.02 (1H, s,
C-5 phenol); 13.30 (1H, s, C-7 phenol); 18.83 (1H, s, C-4 enol).
13C NMR (CDCI3, 75 MHz, 5): 7.5 (q, C-6 methyl); 27.9 (q, C-4 angular methyl);
31.3 (q, C-8 acetyl methyl); 32.1 (q, C-3 acetyl methyl); 59.0 (s, C-4); 98.3 (d, C-1);
101.5 (s, C-3); 103.9 (s, C-8); 105.2 (s, C-6); 109.2 (s, C-5); 155.1 (s, C-8); 157.4 (s, C-
7); 163.8 (s, C-5); 179.0 (s, C-1); 191.6 (s, C-8 acetyl CO); 198.0 (s, C-2 CO); 200.3 (s,
C-3 acetyl CO); 201.7 (s, C-4 CO).

109
IR, v max (KBr, cm"1): 1690 (dienone carbonyl at C-2); 1630 (aromatic C-acetyl);
1610 (enol ether and aromatic double bonds); 1540 (conjugate carbonyl at C-3).
Analysis calculated for Ci8H1607: C, 62.77; H, 4.69. Found: C, 62.47; H, 4.75.
Betuloside (4-(4-IHvdroxvphenvl)-2R-butanol Glucoside)) & Aqlycone
M.p. : 191-193 0 (Lit. 187-190 Khan et al. 1976).
1H NMR (CDCI3, 5): 1.17 (3H, d, Me, J=5.7 Hz); 1.8 (2H, cm, H-3); 2.58 (2H, t, H-
4, J=7.5 Hz); 3.6-4.5 (glucosyl); 6.8 & 7.0 (4H, d, A2B2 Aromatic, J=8.2 Hz); 8.84 (1H, s,
4-phenol).
13C NMR (CDCI3, aglycone, 8): 153.9 (C-4); 133.7 (C-1 ); 129.4 (C-2, 6); 115.3
(C-3, 5); 67.8 (C-2); 40.7 (C-3); 31.2 (C-4); 23.4 (C-1 Me).
IR, v max (KBr, cm"1): 3370, 2930, 2860, 1610, 1590, 1510, 1435, 1445, 1370,
1230.
Analysis calculated for C10H14O2: C, 72.26; H, 8.49. Found: C, 72.12; H, 8.56.

LIST OF REFERENCES
Appendino, G.; Barboni, L.; Gariboldi, P.; Bombardelli, E.; Gabetta, B.; Viterbo, D.
J. Med. Soc., Chem. Commun. 20:1587-1592 (1993).
Appendino, G.; Garibaldi, P.; Pisetta, V.; Bombardelli, E.; & Gabetta, B. Phytochem.
31:4253-4257 (1992).
Appendino, G.; Tagliapietra, S.; Ozen, H. C.; Gariboldi, P.; Gabetta, B.; Bombardelli, E.
J. Nat Prod. 56:514-520 (1993).
Baker, W.; Finch, A. C. M.; Oilis, W. D.; & Robinson, K. W. J. Chem. Soc. 94:1477-
1490 (1963).
Balza, F.; Tachibana, S.; Barrios, H.; & Towers, G. FI. N. Phytochem. 30:1613-1614
(1991).
Barboni, L; Gariboldi, P.; Torregiani, E.; Appendino, G.; Gabetta, B.; Zini, G.;
Bombardelli, E. Phytochem. 33:145-149 (1993).
Baxter, J. N.; Lythgoe, B.; Scales, B.; Scrowston, R. M.; & Trippett, S. J. Chem. Soc.
93:2964-2971 (1962).
Baxter, J. N.; Lythgoe, B.; Scales, B.; Trippett, S.; & Blount, B. K. Proc. Chem. Soc.
90:9-10 (1958).
Boettner, F. E.; Williams, T. M.; Boyd, R.; & Halpern, B. D. Preparation Report 9 (1979).
Polysciences, Inc., Paul Valley Industrial Park, Warrington, PA 18976.
Castor, T. P.; & Tyler, T. A. J. Liq. Chrom. 16:723-731 (1993).
Chauviere, G.; Gunard, D.; Picot, F.; Senilh, V.; & Potier, P. C. R. Acad. Sci. Paris,
Serie II, 293:501-503 (1981).
Chmurney, G. N.; Paukstelis, J. V.; Alvarado, B.; McGuire, M. T.; Snader, V.; Muschik,
G. M.; & Hilton, B. D. Phytochem. 34:477-483 (1993).
Chu, A.; Zajicek, J.; Davin, L. B.; Lewis, N. G.; & Croteau, R. B. Phytochem. 31:4249-
4252 (1992).
Danishefsky, S. J.; Masters, J. J.; Young, W. B.; Link, J. T.; Snyder, L. B.; Magee, T. V.;
Jung, D. K.; Isaacs, R. C. A.; Bornmann, W. G.; Alaimo, C. A.; Coburn, C. A.; &
Di Grandi, M. J. J. Am. Chem. Soc. 118:2843-2859 (1996). For the quote see
the first paragraph of Conclusions, p.2853.
110

111
Della Casa de Marcano, D. P.; & Halsall, T. G. J. Chem. Soc., Chem. Commun.
20:1282-1283 (1970).
Denis, J. N.; Greene, A. E.; Gunard, D.; Gueritte-Vogelein, F.; Mangatal, L; & Potier, P.
J. Am. Chem. Soc. 110:5917-5919 (1988).
Denis, J. N.; Kanazawa, A. M.; & Greene, A. E. Tetrahedron, 35:105-109 (1994).
Ettouati, L. A.; Ahond, A.; Convert, O.; Laurent, D.; Poupat, C.; & Potier, P. Bull. Soc.
Chim. France, 125:749-754 (1988).
Fjllskog, M.-L; Frii, L; & Bergh, J. Lancet, 342:873-874 (1993).
Fuji, K.; Tanaka, K.; Li, B.; Shingu, T.; Sun, H.; & Taga, T. J. Nat. Prod. 56:1520-1531
(1993).
Fuji, K.; Tanaka, K.; Li, B.; Shingu, T.; Sun, H.; & Taga, T. Tetrahedron Lett. 33:7915-
7916 (1992).
Georg, G. I.; Cheruvallath, Z. S.; Vander Velde, D.; Ye, Q.-M.; Mitscher, L. A. Biorg.
Med. Chem. Lett. 3:1349-1350 (1993).
Georg, G. I.; Gollapudi, S. R.; Grnewald, G. L.; Gunn, C. W.; Himes, R. H.; Rao, B. K.;
Liang, X.-Z.; Mirhom, Y. W.; Mitscher, L. A.; Vander Velde, D. G.; & Ye, Q.-M.
Biorg. Med. Chem. Lett. 3:1345-1348 (1993).
Graf, V. E.; & Bertholdt, H. Pharm. Zentralhalle, 96:385-395 (1957).
Guthrie, R. D. Adv. Carbohydr. Chem. 16:105-116 (1961).
Harrison, J. W.; & Lythgoe, B. J. Chem. Soc. C, 1932-1945 (1966).
Hitchcock, S. A.; & Pattenden, G. Tetrahedron Lett. 33:4843-4852 (1992).
Holmes, F. A.; Walters, R. S.; Theriault, R. L.; Forman, A. D.; Newton, L. K.; Raber, M.
N.; Buzdar, A. U.; Frye, D. K.; & Hortobagyi, G. N. J. Natl. Cancer Inst. 83:1797-
1805 (1991).
Holton, R. A.; Kim, H.-B.; Somoza, C.; Liang, F.; Beidiger, R. J.; Boatman, P. D.; Shindo,
M.; Smith, C. C.; Kim, S.; Nadizadeh, H.; Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.;
Zhang, P.; Murthi, K. M.; Gentile, L. N.; & Liu, J. H. J. Am. Chem. Soc.
116:1599-1600 (1994a).
Holton, R. A.; Liu, J. H.; Gentile, L. N.; & Beidiger, R. J. Semi-synthesis of Taxol.
Second National Cancer Institute Workshop on Taxol and Taxus (1992).
Holton, R. A.; Somoza, C.; Kim, H.-B.; Liang, F.; Beidiger, R. J.; Boatman, P. D.; Shindo,
M.; Smith, C. C.; Kim, S.; Nadizadeh, H.; Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.;
Zhang, P.; Murthi, K. M.; Gentile, L. N.; & Liu, J. H. J. Am. Chem. Soc.
116:1597-1598 (1994b).
Horwitz, S. B. Trends Pharmacol. Sci. 52:134-136 (1992).

112
Huang, C. H. O.; Kingston, D. G. I.; Magri, N. F.; Samaranayake, G.; & Boettner, F. E.
J. Nat. Prod. 49:665-669 (1986).
Kanazawa, A. M.; Denis, J.-N.; & Greene, A. E. J. Org. Chem. 59:1238-1245 (1994).
Kingston, D. G. I.; Hawkins, D. R.; & Ovington, L. J. Nat. Prod. 46:466-470 (1982).
Kingston, D. G. I.; Molinero, A. A.; & Rimoldi, J. M. The Taxane Diterpenoids, in
Progress in the Chemistry of Natural Products. 1-206 (1993), W. Herz, G. W.
Kirby, R. E. Moore, W. Steglich, and Ch. Tamm, Eds. Springer-Verlag, New York.
Kondo, H.; &Takahishi, H. J. Pharm. Soc. Japan, 524:821-824 (1925).
Kumagai, T.; Ise, F.; Uyehara, T.; & Kato, T. Chem. Lett. 25-32 (1981).
Kurono, M.; Nakadaira, Y.; Onuma, S.; Sasaki, K.; & Nakanishi, K. Tetrahedron Lett.
30:2153-2160 (1963).
Lewis, N. G.; Chu, A.; Davin, L. B.; & Croteau, R. Phytochem. 34:473-478 (1993).
Lucas, H. Arch. Pharm. 95:145-149 (1856).
Mangatal, L.; Adeline, M.-T.; Gunard, F.; Guritte-Vogelein; & Potier, P. Tetrahedron,
45:4177-4190 (1989).
McGuire, W. P.; Rowinsky, E. K.; Rosenshein, N. B.; Grumbine, F. C.; Ettinger, D. S.;
Armstong, D. K.; & Donehower, R. C. Ann. Int. Med. 111:273-279 (1989).
McMurry, J. E. Chem. Rev. 89:1513-1524 (1989).
Miller, R. W. J. Nat. Prod. 43:425-434 (1980).
Miller, R. W.; Powell, R. G.; Smith, C. R., Jr.; & Clardy, J. J. Org. Chem. 46:1469-1474
(1981).
Nakanishi, K.; Koreeda, M.; Sasaki, S.; Chang, M. L.; & Hsu, H. Y. Chem. Commun.
86:915-919 (1966).
Nakanishi, K.; & Kurono, M. Tetrahedron Lett. 30:2161-2163 (1963).
Nicolaou, K. C.; Dai, W.-M.; & Guy, R. Angew. Chem. Int. Ed. Engl. 33:15-44 (1994a).
Nicolaou, K. C.; Liu, J.-J.; Yang, Z.; Ueno, H.; Sorenson, E. J.; Claiborne, C. F.; Guy, R.
K.; Hwang, C.-K.; Nakada, M.; & Nantermet, P. G. J. Am. Chem. Soc. 117:634-
644 (1995a).
Nicolaou, K. C.; Nantermet, P. G.; Lleno, H.; Guy, R. K.; Couladouros, E. A.; &
Sorenson, E. J. J. Am. Chem. Soc. 117:624-633 (1995b).
Nicolaou, K. C.; Ueno, H.; Liu, J.-J.; Nantermet, P. G.; Yang, Z.; Renaud, J.;
Paulvannan, K.; & Chadha, R. J. Am. Chem. Soc. 117:653-659 (1995c).

113
Nicolaou, K. C.; Yang, Z.; Liu, J.-J.; Nantermet, P. G.; Claiborne, C. F.; Renaud, J.; Guy,
R. K.; & Shibayama, K. J. Am. Chem. Soc. 117:645-652 (1995d).
Nicolaou, K. C.; Yang, Z.; Liu, J. J.; Lleno, H.; Nantermet, P. G.; Guy, R. K.; Claiborne,
C. F.; Renaud, J.; Couladouros, E. A.; Paulvannan, K.; & Sorenson, E. J.
Nature, 367:630-634 (1994b).
Ojima, I.; Habus, I.; Zhao, M.; Georg, G. I.; & Jayasinghe, L. R. J. Org. Chem. 56:1681-
1683 (1991).
Ojima, I.; Habus, L; Zhao, M.; Zueco, M.; Park, Y. H.; Sun, C. M.; & Brigaud, T.
Tetrahedron, 48:6895-7012 (1992).
Rao, K. V. J. Heterocyclic Chem. 34:675-680 (1997).
Rao, K. V. Pharm. Res. 10:521-524 (1993).
Rao, K. V.; Bhakuni, R. S.; Juchum, J.; & Davies, R. M. J. Liq. Chrom. & Rel. Technol.
19:427-447 (1996).
Rao, K. V.; Hanuman, J. B.; Alvarez, C.; Stoy, M.; Juchum, J.; Davies, R. M.; & Baxley,
R. Pharm. Res. 12:1003-1010 (1995).
Samaranayake, G.; Magri, N. F.; Jitrangsri, C.; & Kingston, D. G. I. J. Org. Chem.
56:5114-5119 (1990).
Samaranayake, G.; Neidigh, K.; & Kingston; D. G. I. J. Nat. Prod. 56:884-898 (1993).
Schiff, P. B.; Fant, J.; & Horwitz, S. B. Nature, 277:665-667 (1979).
Schopf, C.; & Ross, F. Naturwissenschaften, 26:772-776 (1938).
Senilh, V.; Blechert, S.; Colin, M.; Gunard, D.; Picot, F.; Potier, P.; & Varenne, P. J.
Nat. Prod. 47:131-137 (1984).
Shapiro, R. H. Org. React. 23:405-507 (1976).
Sharpless, K. B.; Amberg, W.; Beller, M.; Chen, H.; Hartung, J.; Kawanami, Y.; Lbben,
D.; Manoury, E.; Ogino, Y.; Shibata, T.; & Ukita, T. J. Org. Chem. 56:4585-4588
(1991).
Sucrow, W. Chem. Ber. 99:2765-2777 (1966).
Swift, L. J. J. Am. Chem. Soc. 74:1099-1100 (1952).
Swindell, C. S. Org. Prep. Proced. Inti. 23:467-543 (1992).
Wani, M. C.; Taylor, H. L.; Wall, M. E.; Coggon, P.; & McPhail, A. I. J. Am. Chem. Soc.
93:2325-2327 (1971).
Webster, L.; Linsenmeyer, M.; Millward, M.; Morton, C.; Bishop, J.; & Woodcock, D. J.
Natl. Cancer Inst. 85:1685-1690 (1993).

114
Wender, P. A.; & Rawlins, D. B. Tetrahedron, 48:7033-7048 (1992).
Winterstein, E.; & Guyer, A. Hoppe-Seylers Z. Physiol. Chem. 128:175-179 (1923).
Witherup, K. M.; Look, S. A.: Stasko, M. W.; Muschil, G. M.; & Cragg, G. M. J. Nat.
Prod. 53:1249-1255 (1990).
Woodcock, D. M.; Jefferson, S.; Linsenmeyer, M. E.; Crowther, P. J.; Chojnowski, G. M.;
Williams, B.; & Bertoncello, I. Cancer Res. 50:4199-4157 (1990).
Woodward, R. B.; & Brutcher, Jr., F. V. J. Am. Chem. Soc. 80:209-211 (1958).

BIOGRAPHICAL SKETCH
Richard Michael Davies was born in Cocoa, Florida on March 21, 1959 to Dan
and Ruth Lewis Davies. He attended Rockledge High School where he participated in
the cross country and track teams, the schools concert and marching bands, and other
extracurricular activities and societies. He attended Brevard Community College for one
year before enrolling at the University of Florida to complete bachelors degrees in
chemistry and then pharmacy. While studying pharmacy he also worked on projects
with Professor Rao in the laboratory. Interest in the chemistry and research of natural
products and the development of new anticancer therapies brought him back for
graduate studies.
115

I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Koppaka V. Rao, Chair (deceased)
Professor of Medicinal Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
V-/ZL
Perrin, Cochair
ifssor of Medicinal Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Margaretp. James
Professor of Medicinal Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Kenneth B. Sloan
Professor of Medicinal Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Jonathan Eric Enholm
Associate Professor of Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate^wyscop
quality, as a dissertation for the degree of DoctortoF^ilosojahy.
Stephen G. Scbiilman
Professor of Medicinal Chemistry

This dissertation was submitted to the Graduate Faculty of the College of
Pharmacy and to the Graduate School and was aco
requirements for the degree of Doctor of Philosop^
December 1998
Dean, Graduate School



99
Saponification and Acetylation of [6-11 to [6-51:
A solution of taxiflorine (0.1 g) in methanol (5 ml) was treated with 2N potassium
hydroxide (1 ml) and the mixture let stand at room temperature for 2 h. TLC showed that
the starting material was no longer present, along with the appearance of a very slow
moving component. The solution was acidified and extracted with chloroform (3x) and
the combined extracts concentrated to dryness. The residue was dissolved in acetic
anhydride (2 ml) and pyridine (0.5 ml) and let stand for 16 h. Addition of water, filtration
of the solid and chromatography on a normal phase silica gave [6-5] as a white powder.


81
Thus, in summary, dried needles of Taxus x media Hicksii were extracted and
the total chloroform extract applied to a C-18 reverse phase column. A number of
components were separated, such as brevifoliol, taxanes l-IV and taxol, which
crystallized out directly from the fractions. Separation of taxol from taxanes I and II
could be carried out directly by ozonolysis of the mixture, followed by chromatography
on either a normal phase or reverse phase silica column.
Experimental
Extraction:
Dried needles of Taxus x media Hicksii (50 lbs) were extracted with methanol as
described in Chapter 3. The combined concentrate was partitioned between water and
chloroform (10 gallons each). The organic layer was separated and the extraction
repeated twice more using 5 and 3 gallons of the solvent, respectively.
The combined chloroform layers were concentrated under reduced pressure to
yield a dark green semi-solid, representing approximately 5% of the weight of the dried
needles.
Chromatography:
Approximately 800 g of C-18 bonded silica gel was poured into a glass Michell-
Miller type column (2.5 x 24) using methanol (Ace Glass, 1430 North West Blvd.,
Vineland, NJ 08360). The column was equilibrated with 25% acetonitrile in water. The
chloroform extract solids (200 g) was dissolved in acetonitrile (400 ml) in a 4 L stainless
steel beaker, by warming in a hot water bath. To this was added approximately 200 g
equivalent of the equilibrated C-18 bonded silica (20-25% of the column material). While
the mixture is being stirred vigorously, 25% acetonitrile and water 500 ml was added,
followed by water (approximately 800 ml). After stirring for 15 min. it was checked for


101
Flavonoids are chemicals generally found in plants that are ubiquitous and have
been studied for hundreds of years. These compounds are generally yellow to red in
color and are usually responsible for the colors seen in flowers and plants. Research
Into the possibility that flavonoids might possess useful biological activities has
undergone a renaissance in recent years, after decades of neglect. Quercetin and its
most common glycoside (rutin) are probably the most ubiquitous of the flavonoids, and
have been used for many years to enhance immune response to pathogens.
Bioflavonoid complex is mostly rutin and another flavonoid known as hesperidin, and
can be purchased in most drug stores and herbal shops.
These flavonoid compounds were easily purified on the reverse phase columns.
The future usefulness of flavonoids as bioactive compounds remains to be seen, but the
evidence is growing. Some of these compounds have been reported to inhibit the
reverse transcriptase enzyme of HIV virus in vitro, but further work is needed concerning
the mechanism.
Insect molting hormones, commonly classed as edcysones, are responsible for
the maturation of larvae into adult form and have interested scientists for many decades.
Originally, silk worms were extracted to obtain these compounds for research and in
very low yields indeed. Efforts to find better sources for these compounds led to their
discovery in some plants, including Taxus baccata (Hoffmeister 1966). Substantial
amounts of p-ecdysterone, ponasterone A and other ecdysones were easily isolated
during the workup of fractions from the reverse phase columns. All samples tested in
this project had substantial amounts of these hormones, and were easy to isolate.
Usnic acid is a bright yellow compound which is known to grow in lichens. Due
to its structure all signals in the proton NMR are singlets. It is commonly classed as a
tricarbonyl and has peaks as high as 18.8, 13.3 and 11 ppm. An endophytic fungus of
the pacific yew known as Taxomyces andreanae is capable of producing taxol in cell


13
[2-1] 0-C¡nnamoyl Taxicin I Triacetate, R = OH [2-3]-IUPAC
[2-2] O-Cinnamoyl Taxicin II Triacetate, R = H Numbering
[2-4] Taxine B
[2-5] Geranylgeranyl
pyrophosphate
[2-6a] Proposed Glycol,
later corrected to 6 b
[2-6b] R = OAc, Baccatin II
[2-7] R = OH 10-DAB-lll
Figure 2-1 : Early Studies on the Constituents of some Taxus Species


90
8 6-20, whereas the peak at 8 6.30 remains essentially unchanged (8 6.37). With the
reasoning that the allylic H-10 must be more down-field than H-9, the signal at 8 6.30
may be assigned to H-9 and the one at 8 5.90 for the H-10. This leads to the
assignment of the structure of taxiflorine as [6-1], with the hydroxyl at 9 and the
benzoate at 10. Based on the 1H,1H-COSY and 1H,13C-HETCOR spectral data, the four
acetate functions were assigned as 2a, 4a, 7p and 13a.
Benzoylation of [6-1] was carried out to yield the monobenzoate [6-4], which also
gave a 1H NMR spectrum that indicated that it was a single entity. Taxiflorine was also
saponified to the heptaol and re-acetylated to the hexa-acetate [6-5], The 1H NMR
spectra of [6-3], [6-4] and [6-5] are shown in Table 6-1.
The structure of [6-1] with the hydroxyl at C-10 and benzoate at C-9 has the
potential for intramolecular transesterification occurring between the 9-benzoate, as well
as the 7-acetate (Lewis et al. 1993). The fact that the monoacetate and the
monobenzoate gave single products discounted the possibility that transesterification
was responsible for the anomalous NMR spectra of [6-1], To verify if the appearance of
the spectrum is due to an rotameric equilibrium, the spectrum was taken in DMSO-d6 at
temperatures between -20 and 100. At lower temperatures, the spectra were sharper
and showed two sets of peaks for some protons. At higher temperatures, the peaks
coalesced into a single set, as well as became broad to the extent that they were barely
seen. This behavior suggested that the conformational equilibrium between the
rotamers is responsible and that the presence of the 10-OH facilitates this process. As
indicated earlier, during the exploration of possible alternative sources for taxol to
replace the bark of the Pacific yew, scant, if any attention was paid to Taxus floridana.
The species T. x media Hicksii (the ornamental yew) was selected as a possible source
for taxanes in the future.


22
[2-24]
[2-25]
Figure 2-5 : Nicolaous Final Synthetic Intermediates


26
which is esterified in the same fashion as the dienes above to provide further variation.
An unusual example is the taxane with the C-9-nicotinoyl ester function, found in
Austrotaxus spicata Compton Taxaceae (Ettouati et at. 1988).
Oxetanes
This group is characterized by having an oxetane ring system involving the
carbons 4, 5 and 20. It may be divided into two subgroups based on whether they
contain the phenyl isoserine ester side-chain at C-13 or not. The former contains taxol
and all of the other compounds, which are active in the tubulin assay and hence are of
much importance. Division into two other subgroups is also possible in those without the
C-13 side chain, with one having a carbonyl at C-9 and with a hydroxyl or an esterified
hydroxyl at C-9.
The oxetane-containing taxanes are generally highly oxygenated and often have
oxygen at C-1, 2, 4, 5, 7, 9, 10, and 13. In some special instances, a hydroxyl has been
reported at C-19 (Fuji et at. 1993). The phenylisoserine ester side chain has been seen
in the form of at least three different amides that occur in nature. These are taxol, with
the N-benzoyl group, cephalomannine, with the N-tiglioyl group and taxol C, with the N-
hexanoyl group.
Taxol has a complex structure and knowing what features of this structure are
necessary for the activity is of utmost importance and this aspect has been studied using
the in vitro tubulin binding, and the cell culture assays and a summary of these data is
presented below (Samaranayake et ai. 1993).
Acylation of the 2' position of taxol does not destroy cytotoxicity but does stop
promotion of microtubule assembly. Bulky acyl groups reduce the activity in the cell
culture, thus suggesting that hydrolysis of the 2' position back to a free hydroxyl might be
required.


36
Appropriate conditions under which each of these could be obtained as exclusive
products were developed. At room temperature in acetic anhydride for 1-2 minutes
before quenching the reaction, the monoacetate was the major product (>90%).
Likewise, at 80 0 C for 30 min. the product was the diacetate.
The 1H NMR spectral data for the monoacetate showed that the signal at 5 4.45
(br s) shifted to 8 5.37 (dd, J=4.2, 2.4 Hz); indicating that acetylation took place at the 5-
OH, as shown in [3-2], In the diacetate, besides this shift for the 5-OAc, the signal at 8
4.38 (t, 7.5 Hz) shifted to 8 5.54 (br t, 7.2 Hz); thus showing that the second acetate was
located at C-13 [3-4], A naturally occurring brevifoliol 13-acetate [3-3] was isolated and
1,5,13-brevifoliol triacetate [3-5] produced in this lab will be discussed in Chapter 4.
ii) Oxidation: Brevifoliol was readily oxidized by manganese dioxide (Mn02) in
refluxing benzene to yield a ketone product. In the 1H NMR spectrum, a major change
was the absence of the triplet at 8 4.38 due to the C-13 proton, thus showing that the
oxidation took place at the 13-OH [3-6], Further evidence was seen by the shift of the
signals for the C-14 protons from their normal positions at 8 1.29 (dd, 14.0, 7.6 Hz) and 8
2.46 (dd, 14.0, 7.6 Hz) to 5 2.32 (d, 19 Hz, H-14cc) and 8 2.48 (d, 19 Hz, H-14p). When
brevifoliol was oxidized by Jones reagent, the same 13-keto brevifoliol seen with Mn02
initially formed [3-6], With time the initial product gradually disappeared, giving rise to a
faster moving product. This second oxidation product was shown to be the result of an
unusual reaction described in Chapter 4.
4/20 Unsaturation
i) Hydrogenation: When hydrogenated in the presence of 5% Pd/carbon,
brevifoliol gave the dihydro derivative [3-7], In its 1H NMR spectrum, the characteristic
signals at 8 4.82 and 8 5.20 due to the C-20 protons were absent and a new methyl


108
1H NMR (CDCI3, Triacetate, 5): 5.87 (1H, s, H-7); 5.31 (1H, br s, H-3a); 5.04 (1H,
dt=9, 3 Hz, H-2a); 4.84 (1H, d=9 Hz, H-22P); 3.12 (1H, br t=7.8 Hz, H-9a); 2.11 (6H, s,
2X OAc); 2.01 (3H, s, OAc); 1.25 (3H, s, Me-21); 1.03 (3H, s, Me-19); 0.89 (3H, d=2.7
Hz, Me-26*); 0.87 (3H, d=2.7 Hz, Me-27*); 0.85 (3H, s, Me-18).
13C NMR (CDCI3, Triacetate, 8): 201.9 (C-6); 172.3 (OAc); 170.4 (OAc); 170.1
(OAc); 164.7 (C-8); 121.5 (C-7); 84.4 (C-14); 79.5 (C-22); 76.9 (C-20); 68.7 (C-3); 67.1
(C-2); 50.9 (C-5); 49.6 (C-17); 47.5 (C-13); 38.3 (C-1); 35.7 (C-10); 34.1 (C-4); 33.6 (C-
9); 31.5 (C-15); 31.2 (C-12); 29.2 (C-16); 27.9 (C-23); 27.7 (C-25); 23.8 (C-21); 22.9 (C-
19); 22.1, 21.1 (C-27*); 21.0 (C-26*); 20.6 (C-24f); 20.5 (C-111); 17.4 (C-18).
IR, v max (KBr, cm'1) : 3400, 2960, 2870, 1643, 1450, 1380, 1050, 870.
Analysis calculated for C27H4406: C, 69.79; H, 9.54. Found: C, 69.41; H, 9.72.
Phenolic Compounds
Usnic Acid
m.p. : 208-213 C, yellow orthorhombic prisms from ligroin/ethyl acetate (lit. 204,
acetone, Schopf & Ross, 1938). [a]D25 -510, CHCI3, c 0.62 (lit. -509, CHCI3, c 0.679,
Schopf & Ross, 1938).
1H NMR (CDCI3, 300 MHz, 5): 1.76 (3H, s, C-4 angular methyl); 2.10 (3H, s, C-6
aromatic methyl); 2.67 (6H, s, C-3 & C-8 acetyl methyls); 5.97 (1H, s, C-1); 11.02 (1H, s,
C-5 phenol); 13.30 (1H, s, C-7 phenol); 18.83 (1H, s, C-4 enol).
13C NMR (CDCI3, 75 MHz, 5): 7.5 (q, C-6 methyl); 27.9 (q, C-4 angular methyl);
31.3 (q, C-8 acetyl methyl); 32.1 (q, C-3 acetyl methyl); 59.0 (s, C-4); 98.3 (d, C-1);
101.5 (s, C-3); 103.9 (s, C-8); 105.2 (s, C-6); 109.2 (s, C-5); 155.1 (s, C-8); 157.4 (s, C-
7); 163.8 (s, C-5); 179.0 (s, C-1); 191.6 (s, C-8 acetyl CO); 198.0 (s, C-2 CO); 200.3 (s,
C-3 acetyl CO); 201.7 (s, C-4 CO).


104
13C NMR (DMSO-d6, 300 MHz, 5): 177.5 (C-4, carbonyl); 164.1 (C-7, s); 161.4
(C-5, s); 156.9 (C-2, s); 156.7 (C-9, s); 148.5 (C-4, s); 144.8 (C-3,s); 133.5 (03, s);
121.9 (C-6\ d); 121.4 (C-1,s); 116.4 (C-2,d); 115.4 (C-5,d); 104.2 (C-10, s) ;100.6 (C-
1 ,d); 98.6 (C-6,d); 93.5 (C-8, d); 76.5; 76.0; 74.2; 71.9; 70.6; 70.55; 70.49; 70.1; 68.4;
67.2 (C-6, t); 18.0 (C-12 methyl, q).
IR, \/ max (KBr, cm'1): 3340 (OH, bonded); 2920 (CH stretch); 1655 (C=0); 1620
(C=C); 1510 (aromatic); 1355 (C-O-C); 1290 (C-O-C); 1200 (C-O-C); 1055 (C-O-C); 970,
880, 810 (subst. aromatic); 730, 695.
Analysis calculated for C27H3o016: C, 53.12; H, 4.95. Found: C, 52.88; H, 5.06.
Quercetin
2-(3,4-dihydroxyphenyl)-3,5,7-trihydroxy-4H-1-benzopyran-4-one or 3,3,4, 5,7-
pentahydroxyflavone. Refluxed 300 mg of quercetin glycoside in 2 N H2S04 for 3 hours,
filtered, washed with water, yellow needles crystallized from aqueous ethanol, 150 mg
(77%).
1H NMR (DMSO-de, 300 MHz, 6): 12.5 (C-5 hydroxyl, br s, exchangeable with
D20); 7.68 (H-2\ d, J=2.2); 7.54 (H-6\ dd, J= 2.0 and 8.4); 6.90 (H-5\ d, J=8.4); 6.41
(H-8, d, J=2.0); 6.19 (H-6, d, J=2.4).
13C NMR (DMSO-d6, 300 MHz, 5): 175.7 (C-4, s); 163.8 (C-7, s); 160.7 (C-5, s);
156.9 (C-9, s); 147.6 (C-4, s); 146.9 (C-2, s); 145.0 (C-3, s); 122.0 (C-1, s); 120.0 (C-
6, d); 115.6 (C-5\ d); 115.3 (C-2, d); 103.0 (C-10, s); 98.2 (C-6, d); 93.4 (C-8, d).
Analysis calculated for C15H10O7+ 2 H20: C, 53.26; H, 4.17. Found: C, 53.64; H,
4.04.
Sciadopitysin
This biflavonoid is a member of the amentoflavone group, m.p. : 302-304 C.


82
uniformity of the slurry and the absence of oily or waxy material, or lumps. The slurry
was filtered under gentle suction and the solid was resuspended in approximately 500 ml
of the filtrate to give a thin enough slurry for pouring. It was added to the column, the
container rinsed and the rinse transferred to the column.
The remainder of the filtrate was pumped onto the column using a metering
pump (Pulsa 680, Pulsafeeder Inc., Rochester, NY). After the sample addition was
completed, elution was started using 30% acetonitrile and water. This was followed by
35, 40, 45, 50 and 60% acetonitrile and water. The column was then washed with 100%
methanol. Final washing of the column with a mixture of methanol and ethyl acetate and
ligroin removed the green pigments and other lipid-soluble components.
Fractions of 200 ml volume were collected and monitored by UV absorbance,
TLC and analytical HPLC. After this, those fractions that contained significant UV
absorbance and/ or components detectable by TLC or HPLC were set aside for 7-10
days, whereby crystals began to appear from a number of fractions. These were filtered
in groups, characterized and treated appropriately, as described below.
Characterization of the Taxane Components of Taxus x Media Hicksii
Brevifoliol [2-11
Fractions from the 40% acetonitrile and water were concentrated to dryness, the
solid taken up in chloroform and applied to a column of normal phase silica (40 g).
Elution with 2-5% acetone in chloroform gave the major band. The fractions that contain
this component were combined, concentrated to dryness and the solid crystallized from
acetone in ligroin. The crystalline product, yield, 0.8 g (0.02%) m.p. 220-222 0 C was
found to be identical on the basis of NMR spectral data with brevifoliol described in
Chapter 3.


114
Wender, P. A.; & Rawlins, D. B. Tetrahedron, 48:7033-7048 (1992).
Winterstein, E.; & Guyer, A. Hoppe-Seylers Z. Physiol. Chem. 128:175-179 (1923).
Witherup, K. M.; Look, S. A.: Stasko, M. W.; Muschil, G. M.; & Cragg, G. M. J. Nat.
Prod. 53:1249-1255 (1990).
Woodcock, D. M.; Jefferson, S.; Linsenmeyer, M. E.; Crowther, P. J.; Chojnowski, G. M.;
Williams, B.; & Bertoncello, I. Cancer Res. 50:4199-4157 (1990).
Woodward, R. B.; & Brutcher, Jr., F. V. J. Am. Chem. Soc. 80:209-211 (1958).


3
Interestingly, the poor solubility characteristics of taxol might prove to be
responsible for new discoveries regarding a problem of cross resistance to different
classes of chemotherapeutic agents, caused by non-specific drug efflux and referred to
as the MDR phenotype. The MDR phenotype is a gene that has been linked to multiple
drug resistance (hence MDR); and some studies indicate that the solvent used for the
delivery of taxol might have good activity against this common cause for therapeutic
failure in the treatment of cancer (Woodcock et al. 1990; Webster et al. 1993; Fjllskog
et al. 1993). It is known that this effect can result from the expression of plasma-
membrane transport proteins (P-glycoproteins) which can enhance the efflux of
structurally unrelated compounds from the cancer cell. At least three reports suggest
that the solvent Cremaphore LH might enhance the antitumor actions of taxol when the
tumor(s) display the MDR phenotype, and further work with cremaphores alone and in
combination with other antitumor agents is needed to clarify this seemingly serendipitous
finding.
Cremaphore LH is a form of ethoxylated castor oil and is responsible for many
adverse drug reactions during the administration of taxol, and pretreatment with
corticosteroids and antihistamines if often required to prevent allergic response up to
and including anaphylaxis and death. Perhaps more difficult than this solubility concern
was the procurement of an adequate supply of taxol for clinical trials and the anticipated
needs for subsequent worldwide clinical use. Reported yields of taxol from the dried
bark of T. brevifolia were averaging around 0.01%.
A large-scale process for the isolation of taxol was developed by Polysciences,
Inc. (Paul Valley Industrial Park, Warrington, PA 18976); with yields of 0.005-0.01%
(Boettner et al. 1979). Under these conditions, one kilogram of the bark could be
expected to provide only 50-100 mg of taxol at best (or 30,000 lbs. being required for
obtaining one Kg. of taxol). Approximately 2 grams of taxol are needed for one complete


52
1H NMR (CDCI3, 600 MHz, -40 C, 8) : 1.01 (H-19, s); 1.09 (H-16, s); 1.24 (dd,
J=7.2, 13.8 Hz, H-14a); 1.27 (H-17, s); 1.40 (H-2a, br d J=14.1 Hz); 1.76 (s, methyl, 9-
acetate); 1.80 (m, H-6a ); 2.0 (m, H-6p); 2.01 (s, H-18); 2.07 (s, 7-acetate methyl); 2.36
(dd, J=14.1, 9.6 Hz, H-2p); 2.46 (dd, J=7.2, 13.8 Hz, H-14p); 2.67, br s (C-15 OH,
exchangeable with D20); 2.64 (H-3a, br d, J=9 Hz); 2.72 (C-20, s); 3.59 (C-20, s); 4.20
(br s, H-5p); 4.46 (H-13p, br d, J=7.2 Hz); 5.57 (H-7a, br d, J=4.8, 11.4 Hz); 6.05 (H-
9a, poorly resolved br d, J=10.5 Hz); 6.54 (H-10p, br d, J=10.5 Hz); 7.43 (Ph-meta, t,
J=7.8 Hz); 7.56 (Ph-para, t, J=7.8 Hz); 7.87 (Ph-ortho, d, J=7.8 Hz).
13C NMR (CDCI3, 600 MHz, -40 C, 5): 11.9 (C-18 methyl, q); 12.9 (C-19 methyl,
q); 20.7 (7-0 acetate methyl, q) ; 21.3 (9-0 acetate methyl, q); 23.9 (C-2, t); 25.0 (C-17
methyl, q); 27.1 (C-16 methyl, q); 34.1 (C-3, d); 34.4 (C-6, t); 45.4 (C-8, s); 46.4 (C-14, t);
50.1 (C-4, s); 60.2 (C-20, t); 62.4 (C-1, s); 69.5 (C-7, d); 70.5 (C-10, d); 71.7 (C-5, d);
75.8 (C-15, s); 76.7 (C-13, d); 77.1 (C-9, d); 128.7 (Ph-meta, d); 129.4 (Ph-ipso, s);
129.5(Ph-ortho, d); 133.2 (Ph-para, d); 134.3 (C-12, s); 151.3 (C-11, s); 164.3 (CO-Ph,
s); 169.96 (CO-Acetate, s); 170.0 (CO-Acetate, s).
Analysis calculated for C31H4o010: C, 65.02; H, 7.04. Found: C, 64.72; H, 7.24.
Ozonization of Brevifoliol: Brevifoliol-norketone 3-91
A solution of brevifoliol (1 g) in a 9:1 mixture of chloroform and methanol (25 ml)
was cooled in a dry ice/ acetone bath and saturated with ozone produced by an ozonizer
(Ozone Research and Equipment Co., Phoenix, AZ). After testing for the absence of the
starting material by TLC, the mixture was removed from the bath and treated with
dimethyl sulfide (1 ml) and let stand at room temperature for 2 h. It was then
concentrated to dryness and applied to a silica column prepared in chloroform. Elution
with 2% acetone in chloroform gave two bands, which were separated and the fractions
concentrated separately.


47
a column of normal phase silica, using a ratio of 3-5 g of silica per gram of the solid. The
effluent and washes which contained the compound were combined, concentrated to
dryness and the solid crystallized from a mixture of acetone and ligroin to obtain
brevifoliol as a colorless crystalline solid, yield from 200 g of the chloroform extract
solids, 12 g, 0.25% of the dried needles. [a]D23 -2 7 0 (CHCI3; c 1.03); m.p. 220-222 0 C
(lit. 200-203 C [Balza etaL 1991]44);
FAB-MS m/z: 557 [MH]+, 539 [MH-H20]+, 479 [MH-AcOF)]+, 435 [MH-PhCOzH]+,
417 [MH-PhC02H-H20]+, 375 [MH-PhC02H-AcOH]+, IR (KBr) vmax cm'1: 3370, 1740,
1650, 1600, 1585, 1450, 1370, 1265, 1180. UV 7 max log s 3.01 (269 nm); log s 4.32 (223
nm).
1H NMR (600 MHz, CHCI3, 5) Table 3-1: 0.90, s (H-19); 1.05, s (H-16); 1.30 (dd,
J=7.2, 13.8 Hz, H-14a); 1.35, s (H-17); 1.50 (d, J=14.1 Hz, H-2a); 1.76 (s, methyl, 9-
acetate); 1.80 (m, H-6a ); 2.0 (m, H-6p); 2.01 (s, H-18); 2.07 (s, 7-acetate methyl); 2.36
(dd, J=14.1, 9.6 Hz, H-2p); 2.46 (dd, J=7.2, 13.8 Hz, H-14p); 2.67, br s (C-15 OH,
exchangeable with D20); 2.77 (br d, J=9 Hz, H-3a); 4.38 (t, J=7.2 Hz, H-13p); 4.43 (br s,
H-5p); 4.82, s (H-20 A); 5.18, s (H-20 B); 5.57 (dd, J=4.8, 11.4 Hz, H-7a); 6.05 (poorly
resolved br d, J=10.5 Hz, H-9a); 6.53 (d, J=10.5 Hz, H-10p); 7.43 (t, J=7.8 Hz, H-Bz-
meta); 7.56 (t, J=7.8 Hz, H-Bz-para); 7.87 (d, J=7.8 Hz, H-Bz-ortfro).
13C NMR (CDCI3i 600 MHz, 5) Table 3-2: 12.0 (C-18 methyl, q); 12.9 (C-19
methyl, q); 20.7 (7-0 acetate methyl, q) ; 21.4 (9-0 acetate methyl, q); 24.8 (C-17
methyl, q); 26.9 (C-16 methyl, q); 29.1 (C-2, t); 36.0 (C-6, t); 37.9 (C-3, d); 45.0 (C-8, s);
47.3 (C-14, dd); 62.4 (C-1, s); 70.3 (C-7, d); 70.9 (C-10, d); 72.4 (C-5, d); 75.9 (C-15, s);
76.7 (C-13, d); 77.1 (C-9, d); 112.0 (C-20, t); 128.7 (C-Bz-mefa, d); 129.3 (C-Bz-/pso, s);
129.4(C-Bz-ort/?o, d); 133.3 (C-Bz-para, d); 133.9 (C-12, s); 149.0 (C-4, s); 151.5 (C-11,
s); 164.3 (CO-Ph, s); 169.9 (CO-Acetate, s); 170.5 (CO-Acetate, s).


49
Brevifoliol-5,13-Diacetate 3-41
The above reaction was repeated, except that it was heated at 80-90 C (water
bath) for 30 min. After cooling, water was added and the solid filtered after 10 min. The
solid was crystallized from acetone/ ligroin to give the diacetate as a colorless crystalline
solid, yield, 0.2 g; m.p.241-243C;
1H NMR (CDCI3, 600 MHz, 5) Table 3-1: 0.92, s (H-19); 1.11, s (H-16); 1.25 (dd,
J=7.2, 13.8 Hz, H-14a); 1.35, s (H-17); 1.46 (d, J=14.1 Hz, H-2a); 1.75 (s, 9-0 acetate
methyl); 1.90 (m, H-6a) 2.0 (m, H-6p); 2.02 (s, 5-0 acetate methyl); 2.03 (s, 13-0
acetate methyl); 2.07 (s, 18 methyl); 2.08 (s, 7-0 acetate methyl); 2.41 (dd, J=14.1, 9.6
Hz, H-2p); 2.51 (dd, J=7.2, 13.8 Hz, H-14P); 2.72 (d, J=9 Hz, H-3a); 2.74, br s (C-15 OH,
exchangeable with D20); 4.92, s (H-20 A); 5.28, s (H-20 B); 5.39 (br s, J= H-5p); 5.54 (t,
J=7.2 Hz, H-13P); 5.61 (dd, J=4.8, 11.4 Hz, H-7a); 6.09 (poorly resolved br d, J=10.5 Hz,
H-9a); 6.65 (d, J=10.5 Hz, H-1 Op); 7.43 (t, J=7.8 Hz, H-Ph-mefe); 7.56 (t, J=7.8 Hz, H-
Ph-para); 7.87 (d, J=7.8 Hz, H-Ph-ortho).
13C NMR (CDCI3, 600 MHz, 5) Table 3-2: 11.9 (C-18 methyl, q); 12.9 (C-19
methyl, q); 20.7 (7-0 acetate methyl, q); 21.0 (13-0 acetate methyl, q); 21.2 (5-0
acetate methyl, q); 21.4 (9-0 acetate methyl, q); 24.8 (C-17 methyl, q); 27.0 (C-16
methyl, q); 29.1 (C-2, t); 33.9 (C-6, t); 38.8 (C-3, d); 44.8 (C-8, s); 44.1 (C-14, dd); 63.0
(C-1, s); 69.6 (C-7, d); 69.8 (C-10, d); 74.1 (C-5, d); 75.6 (C-15, s); 76.9 (C-13, d); 79.3
(C-9, d); 114.3 (C-20, t); 128.8 (C-Ph-mefa, d); 129.1 (C-Ph-ipso, s); 129.5(C-Ph-or#)0,
d); 133.4 (C-Ph-para, d); 136.4 (C-11, s); 145.2 (C-4, s); 147.3 (C-12, s); 164.1 (CO-Ph,
s); 169.6 (CO-Acetate, s); 169.9 (CO-Acetate, s); 169.91 (CO-Acetate, s); 170.5 (CO-
Acetate, s).
Analysis calculated for C 35H 44On: C, 65.61; H, 6.92. Found: C, 65.68; H, 6.99.


112
Huang, C. H. O.; Kingston, D. G. I.; Magri, N. F.; Samaranayake, G.; & Boettner, F. E.
J. Nat. Prod. 49:665-669 (1986).
Kanazawa, A. M.; Denis, J.-N.; & Greene, A. E. J. Org. Chem. 59:1238-1245 (1994).
Kingston, D. G. I.; Hawkins, D. R.; & Ovington, L. J. Nat. Prod. 46:466-470 (1982).
Kingston, D. G. I.; Molinero, A. A.; & Rimoldi, J. M. The Taxane Diterpenoids, in
Progress in the Chemistry of Natural Products. 1-206 (1993), W. Herz, G. W.
Kirby, R. E. Moore, W. Steglich, and Ch. Tamm, Eds. Springer-Verlag, New York.
Kondo, H.; &Takahishi, H. J. Pharm. Soc. Japan, 524:821-824 (1925).
Kumagai, T.; Ise, F.; Uyehara, T.; & Kato, T. Chem. Lett. 25-32 (1981).
Kurono, M.; Nakadaira, Y.; Onuma, S.; Sasaki, K.; & Nakanishi, K. Tetrahedron Lett.
30:2153-2160 (1963).
Lewis, N. G.; Chu, A.; Davin, L. B.; & Croteau, R. Phytochem. 34:473-478 (1993).
Lucas, H. Arch. Pharm. 95:145-149 (1856).
Mangatal, L.; Adeline, M.-T.; Gunard, F.; Guritte-Vogelein; & Potier, P. Tetrahedron,
45:4177-4190 (1989).
McGuire, W. P.; Rowinsky, E. K.; Rosenshein, N. B.; Grumbine, F. C.; Ettinger, D. S.;
Armstong, D. K.; & Donehower, R. C. Ann. Int. Med. 111:273-279 (1989).
McMurry, J. E. Chem. Rev. 89:1513-1524 (1989).
Miller, R. W. J. Nat. Prod. 43:425-434 (1980).
Miller, R. W.; Powell, R. G.; Smith, C. R., Jr.; & Clardy, J. J. Org. Chem. 46:1469-1474
(1981).
Nakanishi, K.; Koreeda, M.; Sasaki, S.; Chang, M. L.; & Hsu, H. Y. Chem. Commun.
86:915-919 (1966).
Nakanishi, K.; & Kurono, M. Tetrahedron Lett. 30:2161-2163 (1963).
Nicolaou, K. C.; Dai, W.-M.; & Guy, R. Angew. Chem. Int. Ed. Engl. 33:15-44 (1994a).
Nicolaou, K. C.; Liu, J.-J.; Yang, Z.; Ueno, H.; Sorenson, E. J.; Claiborne, C. F.; Guy, R.
K.; Hwang, C.-K.; Nakada, M.; & Nantermet, P. G. J. Am. Chem. Soc. 117:634-
644 (1995a).
Nicolaou, K. C.; Nantermet, P. G.; Lleno, H.; Guy, R. K.; Couladouros, E. A.; &
Sorenson, E. J. J. Am. Chem. Soc. 117:624-633 (1995b).
Nicolaou, K. C.; Ueno, H.; Liu, J.-J.; Nantermet, P. G.; Yang, Z.; Renaud, J.;
Paulvannan, K.; & Chadha, R. J. Am. Chem. Soc. 117:653-659 (1995c).


80
Taxane III [2-11
The crude crystalline solid obtained from the reverse phase column was
recrystallized. Its spectral and analytical data agreed with those given for 5-0-
cinnamoyl-2a,9a, 10p-triacetyl taxicin I (Appendino et al. 1993; Baxter etal. 1962).
Taxane IV 2-21
This was also purified by recrystallization of the crude crystals obtained directly
from the fractions. It was found to be identical with 5-0-cinnamoyl 2a, 9a, 10p-triacetyl
taxicin II, described by Appendino et al. (1992) and Baxter et al (1962).
Taxol r5-31
The chromatography using normal phase silica column as described under
taxanes I and II yielded taxol, which was purified by crystallization. The sample was still
contaminated with some of the taxanes I and II. For complete purification, the mixture
was subjected to ozonolysis which converted these two taxanes to more polar
compounds from which taxol could be readily separated and obtained pure. Using this
method, taxol was obtained in a yield of 0.015% based on the dry needles. This was
significantly better than the reported yield of 0.006% (Witherup et al. 1990).
Ozonolysis of [2-21
Because of the presence of the cinnamoyl ester function in compounds [5-1],
[5-2], [2-1] and [2-2], they all undergo ozonolysis. This method gives a convenient way
of separating taxanes [5-1] and [5-2] from taxol, with which they co-elute. In order to
determine the nature of the product of ozonolysis, taxane [2-2] was subjected to this
reaction and the product recovered and obtained as a crystalline solid. Its NMR spectral
characteristics indicated a hydrated aldehyde with the structure shown in [5-6],


ISOLATION AND CHARACTERIZATION OF TAXANES AND OTHER COMPOUNDS
FROM VARIOUS SPECIES OF TAXUS
By
RICHARD M. DAVIES
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1998


Research made possible the construction of a pilot plant scale facility where these
isolation methods were successfully implemented.
Excellent yields and the isolation of many related taxanes have proven that this
method is superior to currently approved processes used in the production of taxol. The
failure of other researchers to employ bonded silica gel for preparative columns in the
past may reflect experiences with analytical columns, but this method has proven to be
quite exceptional and should be employed extensively.
This dissertation covers many crystalline and non-crystalline compounds isolated
and characterized as a result of this project. Some results from the application of this
technique for the isolation of taxanes from the needles of Taxus brevifolia, Taxus x
media cultivar Hicksii, and T. floridana are presented. Similar experiments on the bark
and wood of T. brevifolia are also described.
xi


2
A study of its mode of action revealed that it blocked cell division at the cell cycle
through its specific action on the G2/M phase of the tubulin/ microtubule system. Unlike
other antitumor drugs such as colchicine, vincristine and vinblastine, which act as tubulin
poisons, taxol exhibited a novel mode of action (Schiff et al. 1979). Microtubules are
involved in the formation of the mitotic spindle fibers necessary for the replication of DNA
and are also integral building blocks within the cell wall. They are generated from a
protein known as tubulin, and a dynamic equilibrium exists between tubulin and
microtubules in vivo. In the presence of taxol, the polymerization of tubulin produces
what are now known as oligo-microtubules. In contrast to the usual microtubules, which
can be readily disassembled, these oligo-microtubules resist disassembly to tubulin,
thereby preventing cell division (Horwitz, 1992).
Based on potent activity against important experimental tumors and its unique
mode of action, interest in taxol became greatly enhanced, and it was approved for
human Phase I clinical trials in the early 1980s. Taxol showed significant activity in
human tumors in Phase I and Phase II clinical trials, especially in ovarian and breast
carcinomas (McGuire et al. 1989; Holmes et al. 1991). The scientific community took a
special interest in taxol at that time due to the lack of adequate treatment options
available for ovarian cancer.
This knowledge established taxol as an important antitumor drug and stimulated
a renewed interest in it. Intensive worldwide studies have reached explosive proportions
since 1994 concerning its production, chemistry, biochemistry and many other aspects.
At that point, two problems needed solution before taxol could become a viable
alternative as a useful treatment against any type of cancer. First, the lipophilic nature of
taxol made it difficult to develop an acceptable dosage form for this drug. Gradually, this
was overcome by the introduction of a suitable, relatively non-toxic dosage form.


69
place. The presence of a hydroxyl was indicated by the fact that the compound would
still undergo acetylation to form a monoacetate.
The C-15 hydroxyl was still present in both the carbon and proton spectra, and
the only other significant change in the spectra occurred with the C-20 signals. The two
broad singlets normally seen around 5 ppm were replaced by an isolated spin system of
doublets at 4.35 and 3.92 which resemble the H-20 oxetane pattern seen in taxol.
Experimental
Brevifoliol Triacetate [3-51
To a solution of brevifoliol (0.2 g) in acetic anhydride (2 ml) was added 0.2 ml of
a 2% solution of boron trifluoride etherate in acetic anhydride to give a final
concentration of 0.2% of boron trifluoride etherate. After 20 min at room temperature,
the mixture was diluted with water. After another 10 min. the solid was separated,
washed, taken up in ether and concentrated to dryness. The solid was crystallized from
ether in ligroin to give [3-5] as a colorless crystalline solid, yield, 0.2 g; m.p. 214-216 0 C.
Oxidation with Jones Reagent to T4-11
A solution of brevifoliol (0.2 g) in acetone (10 ml) was treated with Jones reagent
(2 ml) added in small portions and with stirring. Initially TLC analysis of the reaction
mixture showed that a yellow color giving spot appeared above that of the starting
material. Gradually the first product changed into an even faster moving component.
When this latter was the predominant product, the reaction was stopped by the addition
of water and extraction with chloroform. After concentration of the solvent, the product
was chromatographed on a silica column in 1:1 chloroform/ ligroin. Elution with
chloroform gave the major component, which was obtained as a colorless crystalline
solid, yield, 0.05 g; m.p. 234-236 0 C.


44
The remaining carbons are the 4 methyl groups usually seen in taxanes on C-15
(methyl 16 and methyl 17); at C-12 (methyl 18) and at C-8 (methyl 19). The methyl
group located on the 11/12 double bond (methyl 18) is often quite deshielded in the
proton spectrum (§2.01, s) but shielded in the carbon spectrum (5 12.0). This usually
aids in its assignment along with further evidence from Heteronuclear NMR experiments.
Methyl 19 is usually shielded in both the 1H (5 0.90, s) and 13C (5 12.9, q) spectra, as
seen here.
This class of compounds commonly referred to as 11(1-15)-a6eo-taxanes or
occasionally A-nortaxanes. Many compounds of this type are now known, some
containing the 4/20 unsaturation as in brevifoliol and others with a 4/20 oxetane structure
as seen in 11(1 >15) abeo baccatin VI.
Experimental
Extraction of the Needles of Taxus brevifolia
The needles obtained from a supplier (Mr. Patrick Connolly, Yew Wood
Industries, 6928 North Interstate Avenue, Portland, OR 97217) were air-dried for one
week. The dried needles (20 Kg) were extracted by immersing in methanol at room
temperature. After two days, the extract was drained, concentrated under reduced
pressure at temperatures below 35 0 C. The recovered methanol was reused for a
second extraction, which was processed the same way. After two more extractions, the
combined concentrate was freed from some more of the methanol to obtain a dark green
syrup.
The above syrup was partitioned between water (10 gallons) and chloroform (10
gallons). The organic layer was separated and the extraction carried out twice more


CHAPTER 3
TAXANE CONSTITUENTS OF THE NEEDLES OF TAXUS BREVIFOLIA
Taxol was originally isolated from the bark of the Pacific yew (Taxus brevifolia
Nutt., N.O. Taxaceae). As indicated in Chapter 1, during 1991-1993 there was a
reassessment of the use of the bark as the source. This concern resulted in an intense
search for alternative sources for taxol that are renewable, with sources such as the
needles of the yew tree instead of the bark. This laboratory was also involved in this
search and looked into the needles of three different yew species as a source for taxol;
T. brevifolia, T. x media Hicksii and T. floridana. The taxane composition of T. brevifolia
needles is the subject of this chapter.
Fractionation of the Needles of Taxus brevifolia
A quantity of 100 lbs. of the needles of T. brevifolia was obtained from a supplier
in Oregon. They were air-dried and extracted with methanol at room temperature and
the extract was concentrated under reduced pressure to a syrup. This was partitioned
between water and chloroform, and the organic layer concentrated to give a dark
greenish brown semi-solid, called extract solids, which represented about 5% of the dry
weight of the needles.
It was decided to follow the method successfully developed with the bark extract
for the fractionation of the extract solids, using preparative scale, reverse phase column
chromatography. Direct application of the crude chloroform extract of the needles onto a
C-18 bonded reverse phase silica column was accomplished as described in the
experimental section. After placing the extract-containing silica onto a 25% acetonitrile
29


LIST OF TABLES
Table page
3-1 : Proton NMR Spectra of Brevifoliol and Brevifoliol Acetates 33
3-2 : Carbon NMR Spectra of Brevifoliol and Brevifoliol Acetates 34
4-1 : NMR Spectra of Compound [4-3] from BF3 Reaction 63
6-1: Proton NMR Spectra of Compounds [6-3], [6-4] and [6-5] 91
vii


105
1H NMR (DMSO-d6, 300 MHz, 5): 13.05 (1H, s, 5-OH); 12.90 (1H, s, 5-OH); 8.31
(1H, s, 7-OH); 8.17 (1H, dd, J=9.0, 2.4, 6-H); 8.08 (1H, d, J=2.4, 2-H); 7.60 (2H, d,
J=9.0, 2 H, 6 H); 7.37 (1H, d, J=8.7, H-5); 6.98 (1H, s. H-3); 6.93 (2H, d, J=9, H-3,
H-5); 6.89 (1H, s); 6.78 (1H, d, J=2.4, H-8); 6.42 (1H, s, H-6); 6.36 (1H, s, 6~H); 3.83
(3H, s); 3.80 (3H, s); 3.76 (3H, s).
13C NMR (DMSO-d6, 300 MHz, 5): 181.9; 181.8; 165.1; 163.5; 162.9; 162.1;
161.0; 160.5; 160.4; 157.2; 154.2; 130.8; 128.2; 127.7; 122.7; 122.3; 121.5; 114.4;
111.6; 104.1; 103.7; 103.55; 103.48; 103.1; 98.6; 98.0; 92.6; 55.9; 55.8; 55.4.
IR, v max (KBr, cm"1) : 1660, 1650, 1620, 1610, 1570, 1510, 1430, 1370, 1240,
1180, 1160, 1050, 1030, 960, 910, 880, 830, 760.
Analysis calculated for C33H24O10: C, 68.27; H, 4.17. Found: C, 67.94; H, 4.26.
B-Sitosterol-B-D-Glucoside
Identity of glycoside confirmed using authentic sample with mixed melting point,
TLC, and IR spectrum, as well as 1H- and 13C NMR of the tetra-acetate, the aglycone,
and the aglycone acetate. M.p. : 288-290 0 C (lit. varies from 280-300 0 C, 298 0 C
Sucrow 1966). EI-MS: 414(2%) [M+ glucosyl], 396(34%) [MH+ glucosyl H20]. Cl-
MS (methane): 413(9%) [MH+ glucosyl H2], 397(100%) [MH+ glucosyl H20],
[a]D23 -40.1 (pyridine, c 1.1); (lit. -41.0, c 1.33, pyridine, Swift 1952).
IR v max (KBr, cm'1) : 3400 (-OH); 3090 (C=CH2); 1650, 890.
Analysis calculated for C35H6o06: C, 72.87; H, 10.48. Found: C, 72.48; H, 10.61.
B-Sitosterol-B-D-Glucoside Tetra-acetate
Acetylation of p-sitosterol-p-D-glucoside: Dissolved 500 mg in 5 ml acetic
anhydride with 0.1 ml pyridine, then placed in hot water bath for 1 hour with stirring.
After TLC indicated the reaction was complete the mixture was stirred with water for 15
minutes to decompose the anhydride, then extracted with chloroform 3 x at pH 4. Dried


96
Taxiflorine f6-11:
The crude crystalline solid (2.5 g) that separated from the fractions from 45%
acetonitrile/ water was filtered and purified by recrystallization from acetone/ ligroin to
give [6-1] as a colorless crystalline solid, yield, 1.2 g (0.006%); m.p. 254-255 C, [a]D23
-26.1.
Analysis calculated for C35H44013: C, 62.48; H, 6.59. Found: C, 62.12; H, 6.63.
Baccatin VI [6-2]:
Eluates from the 50% acetonitrile/ water gave crystals in a number of fractions.
These were filtered into groups and tested by TLC and analytical HPLC. The earlier
fractions contained mostly baccatin VI, with gradually increasing amounts of a slower
compound, shown to be taxol. The crystals from the first group containing mostly [6-3]
(3.5 g) were dissolved in chloroform (50 ml) and passed through a column of Florisil (20
g) for the purpose of decolorization. The effluents and washes were combined,
concentrated to dryness and the solid crystallized from acetone/ ligroin to give pure [6-3]
yield, 1.6 g. Together with the amount obtained from the next fraction, the total yield
was 1.95 g (0.01%); m.p. 250-252 0 C (lit. 248-250 0 C decomp., Senilh et al. 1984);
[a]D23 -11 (chloroform, c 0.98) (lit. -5, chloroform, c 1.3, Senilh et al. 1984); MS(FAB);
737 [M+Na]\ 697 [M-HzO]+.
1H NMR (CDCI3i 300 MHz, 8): 1.22 (17-Me, s); 1.60 (19-Me, s); 1.78 (16-Me, s);
1.87 (6a, cm); 1.99 (OAc-Me, s); 2.02 (OAc-Me, s); 2.04 (C-14P, unresolved mult.); 2.10
(18-Me, s); 2.10 (OAc-Me, s); 2.19 (OAc-Me, s); 2.20 (C-14a, unresolved mult.); 2.28
(OAc-Me, s); 2.50 (C-6(3, cm); 3.18 (C-3, d, 6 Hz); 4.13 (C-20, d, 8.4 Hz); 4.34 (C-20, d,
8.4 Hz); 4.97 (C-5, d, 8.4 Hz); 5.55 (C-7, dd, 7.5, 9.3 Hz); 5.87 (C-2, d, 6.0 Hz); 5.99 (C-
9, d, 11.1 Hz); 6.17 (C-13, dd, 7.6, 9.3 Hz); 6.22 (C-10, d, 11.1 Hz): 7.48 (Ar-meta, t, 7.8
Hz); 7.61 (Ar-para, t, 7.5 Hz); 8.09 (Ar-ortho, dd, 7.2, 1.3 Hz).


77
Dried Needles of
Taxus x media Hicksii
Extract
Residue
(Discard)
Chloroform
Aqueous
Extract Solids
Reverse Phase Column
Filter Crystals
Recrystallize or chromatograph
Brevifoliol Taxanes I and II Taxol Taxane III Taxane IV
Ozonization
Chromatography
Taxol
Figure 5-1 : Fractionation of the Extract of Taxus x media Hicksii Needles


12
many decades and covered many reactions relevant to taxane chemistry. In one such
study, Winterstein and Guyer (1923) were the first to show the presence of 3-
dimethylamino-3-phenylpropanoic acid in the hydrolyzate of taxine and this acid later
became known as "Winterstein's acid.
Until the 1960s most of the work on taxanes focused on these acid-extractable
alkaloidal substances, which were readily separable from the large quantities of neutral,
resinous materials which dominate the extract. Two groups of researchers were able to
convert these somewhat unstable alkaloidal mixtures into more stable, non-basic
substances in which the 3-dimethylamino-phenylpropanoic ester unit was transformed
into a cinnamate ester. This, as well as the development of chromatographic
techniques, made it possible to obtain pure compounds rather than mixtures.
Baxter et al. (1958) in England investigated the major cinnamate ester obtained
from T. baccata, which they named 5-O-cinnamoyl taxicin-l triacetate [2-1], Similarly, a
Japanese team (Nakanishi & Kurono, 1963; Kurono et al. 1963) studied a cinnamate
ester from T. cuspidata, and called it 5-O-cinnamoyl taxicin II triacetate [2-2] and the
structures of both these can be seen in Figure 2-1. These two compounds differ only at
C-1, where taxicin II lacks the tertiary hydroxyl found in taxicin I. The IUPAC numbering
system for taxanes used throughout this dissertation can also be seen [2-3],
A few years earlier, Graf & Betholdt (1957) succeeded in isolating the purified
basic alkaloids, taxine A and taxine B from the original taxine mixture. Taxine B was
shown to have the structure [2-4] (see Figure 2-1); which corresponded with 5-O-
cinnamoyl taxicin I triacetate, into which it could be converted via elimination of the
dimethylamine moiety.


111
Della Casa de Marcano, D. P.; & Halsall, T. G. J. Chem. Soc., Chem. Commun.
20:1282-1283 (1970).
Denis, J. N.; Greene, A. E.; Gunard, D.; Gueritte-Vogelein, F.; Mangatal, L; & Potier, P.
J. Am. Chem. Soc. 110:5917-5919 (1988).
Denis, J. N.; Kanazawa, A. M.; & Greene, A. E. Tetrahedron, 35:105-109 (1994).
Ettouati, L. A.; Ahond, A.; Convert, O.; Laurent, D.; Poupat, C.; & Potier, P. Bull. Soc.
Chim. France, 125:749-754 (1988).
Fjllskog, M.-L; Frii, L; & Bergh, J. Lancet, 342:873-874 (1993).
Fuji, K.; Tanaka, K.; Li, B.; Shingu, T.; Sun, H.; & Taga, T. J. Nat. Prod. 56:1520-1531
(1993).
Fuji, K.; Tanaka, K.; Li, B.; Shingu, T.; Sun, H.; & Taga, T. Tetrahedron Lett. 33:7915-
7916 (1992).
Georg, G. I.; Cheruvallath, Z. S.; Vander Velde, D.; Ye, Q.-M.; Mitscher, L. A. Biorg.
Med. Chem. Lett. 3:1349-1350 (1993).
Georg, G. I.; Gollapudi, S. R.; Grnewald, G. L.; Gunn, C. W.; Himes, R. H.; Rao, B. K.;
Liang, X.-Z.; Mirhom, Y. W.; Mitscher, L. A.; Vander Velde, D. G.; & Ye, Q.-M.
Biorg. Med. Chem. Lett. 3:1345-1348 (1993).
Graf, V. E.; & Bertholdt, H. Pharm. Zentralhalle, 96:385-395 (1957).
Guthrie, R. D. Adv. Carbohydr. Chem. 16:105-116 (1961).
Harrison, J. W.; & Lythgoe, B. J. Chem. Soc. C, 1932-1945 (1966).
Hitchcock, S. A.; & Pattenden, G. Tetrahedron Lett. 33:4843-4852 (1992).
Holmes, F. A.; Walters, R. S.; Theriault, R. L.; Forman, A. D.; Newton, L. K.; Raber, M.
N.; Buzdar, A. U.; Frye, D. K.; & Hortobagyi, G. N. J. Natl. Cancer Inst. 83:1797-
1805 (1991).
Holton, R. A.; Kim, H.-B.; Somoza, C.; Liang, F.; Beidiger, R. J.; Boatman, P. D.; Shindo,
M.; Smith, C. C.; Kim, S.; Nadizadeh, H.; Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.;
Zhang, P.; Murthi, K. M.; Gentile, L. N.; & Liu, J. H. J. Am. Chem. Soc.
116:1599-1600 (1994a).
Holton, R. A.; Liu, J. H.; Gentile, L. N.; & Beidiger, R. J. Semi-synthesis of Taxol.
Second National Cancer Institute Workshop on Taxol and Taxus (1992).
Holton, R. A.; Somoza, C.; Kim, H.-B.; Liang, F.; Beidiger, R. J.; Boatman, P. D.; Shindo,
M.; Smith, C. C.; Kim, S.; Nadizadeh, H.; Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.;
Zhang, P.; Murthi, K. M.; Gentile, L. N.; & Liu, J. H. J. Am. Chem. Soc.
116:1597-1598 (1994b).
Horwitz, S. B. Trends Pharmacol. Sci. 52:134-136 (1992).


68
JjLjULjL-JjJLjikj
w
Q.
Q.
OJ
U-
Figure 4-6 : HETCOR Spectrum of [4-4]
Recorded on Varian VXR 300 spectrometer at 75.432 MHz in CDCI3 with TMS as
internal standard. F-,: 3000.3 Hz; F2: 15528 Hz; Acquisition time 65.9 sec; 1 sec;
Ambient temperature; Decouple proton; Level 70 high power; PW: 90; 128 repetitions X
128 increments; Waltz 16 modulation; pseudo echo; FT size: 2K X 512 data points; time:
5 hours.
Elemental analysis and FAB-MS gave the molecular formula C31H3608, which
indicated essentially that loss of one molecule of water and dehydrogenation had taken


64
Bz
Ac
[4-3] BF3-etherate Product
Figure 4-2 : BF3-etherate Catalyzed Elimination Product
Based on this reasoning, the structure of the BF3-reaction product was assigned
as shown [4-3] in Figure 4-2 above. The DEPT spectra of [4-3] are given in Figure 4-3.
CH3 Carbons
CH2 Carbons
20
2x Ac 18 19
14 6 2
CH Carbons
13
5
9
7
10
3 DMSO
i
.... .
-- -

Adi Protonated Carbons
i t i ] r 1 i r -j i i T'i ]' i n i j i m 7 rrrrriTT rrpr'n T]-| i t ;-) n r r j n m m r r r yr ¡-mi tt trt -j- t ti i ?
140 120 100 00 60 40 20 6
Figure 4-3 : DEPT Spectra of BF3 Elimination Product [4-3]