I

.

.4 ‘ \
\

‘l l

I
1

I
L

llxll

I
w

W

l
l

I
l *.

 

HTHS

REACTIONS 0F MHHYL 2‘3uANHYDRO-RIBOSlDES
WR'H HYD-RQGEN FLUQRIQE

3.52538 for Hm Degree of M. S
MICHIGAN STATE UNIVERSETY

33213 C. Justice

1959

'I Nil-.515

C,L

   

  
 

LIBRAR Y
Michigan Stats
University

 

 

 

 

a, , 1’11"-“N'WOW

 

 

 

nun—n _.... . .

_‘

 

 

1|!ch gum MM"
R E C E I V E D

om. op mums!"

 

 

REACTIONS or METHYL 2.3-ANHYDRO-RIBOSIDES
wma HIDROGEN FLUORIDE

3!
John 0; Justice

A THESIS
Submitted to the School for Advanced Graduate Studies of
Michigan State university of Agriculture and Applied Science
in partial fulfillment of the requirements

for the degree of

MASTER OF SCIENCE
Department of Chemistry
1959

To

PM

'IIACKNOWLEDGMENT
The author expresaée gratitude

to Dr. John c._ Speck. Jr... for. _

"hie many helpful and directing

discussions, and for the undfirr..

standing and interest in.his
students which led to the
author's present feelings about
an academic life.

nnnnnnnnnnn

Gratitude is felt toward the__
Lansing Chapter of the
American.Cano¢r Society

whose grant supported this
investigation.

\\\\\

ooooo

. ~~~~~~~

3333

........

TABLE OF CONTENTS

PAGE
HISTORICAL INTRODUCTXONooooooQquoecoooooootooeooooooooo 1
CHEMICAL»BACKGROUNDeooooeoooooooooenovice-ooooooototoooo 9
KATERIALS AND METHODsooeooOQtooOOOOotovooooooebteoehcoio 15
EXPERIMEHTALooOIoiooooOcolcode...ooooooooo0.060.000.0000 16
Methal fi-L-erabinopyranoeide..nu.nun... coco-coo. .n 16
Methyl 3.4-0-leepropylidenoqflvLeerebinopyranoside.o..... 16
Methyl 2aoo'roeylc-3,A-O-ioopropylidone—fi-lpurabino-

_pyrnnoeide............u......u................... 1?
Methyl 2~0~TooquߴLnorEbinooyranoeide....s............. '7
Methyl 3,4-Di-o-acotyl-e-o-toequfLLnerEbinopyrenceido.. 1?
Methyl 2,S-AnhydroqflLLpercbinopyranoaide.g.............. 18

It: utilization.................................. 19
1.233.hnDi-O-ieopropylidene~D-xylofuranoee...o.......... 20
I.2-0-Inapropylidene-D-xylofuranoee..................... 21
i,2-0-leaprOpylidene-Sac-methoxycarbonyl-D—xyiofuranoee. 22

1.2~O¢Inopropylidens-B-O-neeyl~5-O~methoxycarbon31~
D91310fur8n0IO¢oooooon...oooooooooooeooooooooocco 23

1.2~Di~0~ccetyl-3~O~neeyi-S-O-methoxyeerbonylc
D'XyloruanOBOeonoooooooootoQOOIooooeooooouse...o 2}

Methyl Sva-Mceyl-S-Oomcthoxycerbonyl-a(and[9)u
D’xylorurnnaaldofloooooooooaoeooooooooo«0.09000... 24

Methyl 2.3-»Anhydro-adcnd fl )‘D’fibfiflflhOBldOo o o o‘ e o o o i o o o 25
Its ”$11138tiflnoeoielooooooooooeleoooooeooooooooo 26

DISOUSSIONoooeooOIoQno«coco-bebo-ooooolooooooonecoco-too 3'

LITERATURE CITEnneoooooooeoinnocuo-boonuooodoboooccooooe 32

Moissan. who discovered elemental fluorine in 1886,
also investigated its reactions with many organic compounds
and was the first to test the pharmacological activity of
a fluoro compound. He found that fluoroethans, in
contrast to ethyl chloride, ethyl bromide, and ethane,
is very toxic. (I) In low concentrations in air it pro-
duces acceleration of breathing rate in guinea pigs. At
increasingly higher concentrations the animal becomes ex-
cited: the breath comes spasmodically; there are convulo
sive Jerks; the rear extremities become lame; and, at a
concentration of six to seven percent. death results.

The value of fluoride ion in the construction of
tooth enamel has elicited much controversy in the past
decade and is part of the increasing interest and awareness
of the anomalous behavior of fluorine, fluorides. and
fluoro-compounds. The chemistry of this element has
enjoyed a logarithmic exploitation as notable as any other.

In the field of chemotherapy. interest in pseudo.
metabolites has brought attention to fluoro-compounds,
among others. Closely related to the object of this thesis
is the pseudcmetabolite eefluoroadenosine. prepared by the
Schiemann reaction (2) by Montgomery and Henson. (3)

2-Fluoroadsnosine in a concentration of 10-8

g. per ml.
inhibited human epidermoid carcinoma; azaserins and 6-diazo-

Scone-Lenorleucine inhibit this tissue at 10-7 g. per ml.

2
6-Methyl purine has shown marked effects on neoplastic
growth. (&.5) In prevention of mitosis in.both sarcoma
$80 and normal skin cells in tissue culture, it is effective
at a concentration of 0.05 umole per milliliter, but
6~trif1ucromethyl purine shows no toxicity at five hundred
times this concentration. (6)

The antitumor activity of purine itself and of 6-chloro-
purine ($1.6) aroused interest in 6-fluoropurine, but at—
tempts to prepare this compound by several routes were
without success. This was interpreted as due to the in-
stability of the compound because similar’difficulty was
encountered by Roe and Rankine in the unsuccessful Schiemann
synthesis of AefluorOPyridine, although 2~ and 3-f1uoro-
pyridine were preparable. (12)

many other purine and pyrimidine derivatives have been
‘prepared and tested for antitumor activity. Among those
that have had some success are 6-mercapt0purine, 5-fluoro-
uracil. 5-bis(2-chlorcethyl)-aminouracil, and 8-azaguanine.

G-Mercaptopurins is thought to be a competitive in.
hibitor‘between the conversion of inosinic acid to adenine
or guanine polynucleotides. (7) It has given temporary
relief from the symptoms of leukemia. (8)

S-Fluorouracil is a potent uracil competitor which is
not greatly differentially toxic to tumor and normal tissue.

8~Azaguanine has been shown to enter the ribonucleic
acid of cells. Among the azapurine analogues. it has been

more interesting than the adenine, hypoxanthine. and zan-
thine analogues because of its inhibition 01’ a variety of
living systems. R. E. F. Matthews gives a broad review of
the literature. (9) Incorporation seems to be into shorter
polynucleotides according to Matthews' work and that of
others in the discussion following the paper. An Interest-
ing aspect is the work of Crosser (13) which showed that
the enzyme adaptation of WW gums, toward
lactose, which necessitates the production of fi-galactcsid-
ass, was prevented in whole growing cells at relatively
low concentrations of Boazaguanine.

5-131“2~chloroethyl)-aminouracil is a thgen mustard
derivative prepared bylyttle and Petering. (M) It is not
a uracil competitor; the uracil moiety serves as a carrier
for the nitrogen mustard which acts as an alkylating agent.
The nitrogen mustards have been used for some time in leu-
kemia therapy with temporary remission of symptoms. The
new compound has produced complete repression of experie
mental growing tumors of several types in rats and is being
tested in human patients at present. The ratio of its
effective dose to toxicity is favorable.

The compomxd 9(fi-D-xylofuranosyl)adenine gives further
hope for the rationale behind this thesis. This compound
has been shown to be highly toxic by Biesele at Sloan-
Kettering Institute. In as much as it has been strongly
indicated that the nucleic acids are polymers of ribose or

4
of 2-deoxyribose linked by phosphate ester'bridges between
the 3- and Suhydroxyls, (16,i7) the l-ccnfiguration on the
S-carbcn of xylcse in xylcfuranosyladenine may mean that a
stable termination of a nucleic acid is produced which,
because of the orientation of the 3-hydroxyl. is incapable
of further extension and completion of the normal nucleic
acid. Arabinofuranosyladenine is not toxic. In this case
the 3-hydroxyl has the same configuration as in the natural
ribosyl- or deoxyribosyladenine.

Work with.purcmycin. 6~dimethylaminc~9-(3'«p-methoxy-
Lpphenylanylamino-B'«decxy-D-ribcfurancsyl)pumine. a compound
produced.by‘fitzgntomycgn.glhgniggz, has shown it to have a
broad-spectrum activity against bacteria inhxitzg (18) and
against a number of protozoa in m. (to) Puromycin also
displays activity against higher animals and against
mammary adenccarcinoma in.the 63H strain of mice. Although
it has undesirable side effects. these are remitted on dis-

continuance of administration and are not sufficiently severe

to prevent its use in combating W whims (19)

and MW (20) in man.
A.number of structural modifications of puromycin have

been made. Hutchings (to) states that the structural
requirements for activity vary with the organisms studied.
"The intact molecule is required for the antibacterial
activity in m. antismoebic activity in 1119, and anthol-

mintic activity. The aminonucleoside is more effective on

5

a molar basis than the parent compound when tested against
trypanosomes in,1119 and the mammary adenocarcinoma of the
C38 mouse.” Because in several studies (21,22) purines
have reversed the action of puromycin, it is suspected that
it may interfere uith.purine metabolism. With respect to
substituents on the purine ring, a 6-dipropylamincnuclsoside
was eight times more effective than the nucleoside and
thirty-two times more effective than.puromycin against
(Izznangsgma.egninezdnm in the mouse. (10a)

Again the 3’ position of the nucleoside is indicated as
a significant point in the molecule. B. R. Baker and his
group at Stanford Research Institute are concentrating on
syntheses of pentofurancses of potential interest deriving
from his work with puromycin. The methyl 3-daoxy-3-amino-
xxlnrfuranoside has been synthesized (23) by ammonolysis of
methyl 2.B-anhydro~D~ribcfuranosids. The 3'-deoxy~3‘e
mercaptcribcnuclecsides are now holding their attention. (2h)

One might interdect comments on the surprising activity
of various flucrosteroids at this point. Adrenal cortical
hormones, e.g. cortisone. have found use in the treatment
of allergies. the prevention of scarring. treatment of
rheumatoid arthritis, and the treatment cf shock. Synthetic
methods for the preparation of cortisone provided routes and
material for various derivatives of the parent compound.
9~abFluor0hydroccrtisone has elicited considerable interest
since its gluccrticoid activity. i.e., the extent of deposition

of liver glycogen from metabolized protein, is about ten
times that of hydrccortisone, the'lithium aluminum hydride
reduction product of cortisone which.has the other two -
ketc groups protected. 94x-Fluorohydrccortisone'was pre-
pared (25) by Opening_a.9fi; tificepoxide with hydrogen.
fluoride. It was found that the activity of the compound
'wss inversely proportional to the size of the substituted
halogen. 9-drFluorohydrocortisone had undesirable side
effects which included the production of edema. Intro—
duction of c.16de or*~@~methyl has overcome some of the
side effects of the parent fluorosteroid. 16deMethy1«
9d§fluoroprednisolone 21-acétate. prepared by the Shering
group, has enhanced glucocorticoid activity and antiein-
flammatory activity, with complete avoidance of salt and
water‘retaining characteristics found in the parent
9arflucropredniselone. (26)

Ringold and Bowers (27) have used Henbest and Wrigley's
synthesis (28) to open a 5,6—epoxide‘by reaction with
boron.trifluoride-ether complex in a‘benzene medium to
produce 6-flucrosteroids. They report that 6drfluorocortis-
one acetate, 6d¢fluorohydrocortisone acetate. Gdefluoro-
prednisolcne acetate, and odefluorOPrednisone acetate
had respective antieinflammatcry and thymolytio activities
of (10, 6). (to. 8). (20. 62). and (20. 23) times that of
hydrccortisone acetate. In addition the latter three promoted
sodium excretion in non-salt-lodded adrenalectomized rats.

7

Pattison and his students at -the University of Western
Ontario have written many papers in their series on toxic
fluoro-compounds. The papers mentioned here give reference
to their work.

a-Fluorcaklyl nitrogen compounds have been prepared to
lead to new barbiturates. in this case, S-ethyl-S-(h'ofluoro-
butyl)barbiturste and S-ethy1e5(5'rfluoroamyl)berbiturate. (29)
Their activity as central nervous system depressants was
lower than that of neonsl and smytal. and their toxicity was
low as. according to similar data from Bruce and Huber. (30)
could be expected.

can-Fluoro fatty acids shd ketones show higher toxicity
for members having an even masher of carbons such as
ethyl 8-flucro-3-oxoecctancate and ethyl 12-f1uoro-3-oxo-
dodecanoste than for members having an odd number of carbons
such as ethyl 9-fluoro-3voxc-nonoate. (3H- This is in
accord with the betaeoxidation theory of fatty acid
metabolism, which would yield fluoroacetic acid as the end
product of the acids with an even number of carbons.

Even chain iefluoroalkanes. ealkenes. and ealkynes
are very toxic to mice. Ethyl 3-fluoroa-lactate. discussed
in the same paper. was non toxic. (32)

1 B-Fluoro-‘l 0-methyloctad ecanoio so id . IBoflucrootubeh
culostearic acid, was prepared to test it as an antietubero
oular compound. it showed complete inhibition of W...

isms tuberculosis var: honinis at a concentration of

8

1.25 gamma per ml. in m in the absence of bovine serwn,
but was inactive in the presence of ten percent bovine
serum at a concentration or 20 gamma per ml. In nice it
had no effect on tuberculosis at a diet concentration or
25 mg. per kg. per day. Intmperitoneally, the lethal dose
for riftypercent of mice was 2.7 mg. per kg. (33)

The activity or fluoro-compmmds in biological systems
and the possible strong effect or a 2- or 3-substituted
pentose on nucleic acid metabolism prompted this thesis.

CHEEICAL BACKGROUHD I ,

2-Deoxyv2-f1uoro~D,L-glyceraldehyde has been prepared
by Taylor and Kent. (3h), Diethyl fluoro»oxalacetate was
reduced to 2-decxy-2-fluorotetritols with potassium boro-
hydride and lithium aluminum hydride. then oxidized with
sodium periodate to the fluoroglyceraldehyde, The fluoro
atom is removed by reaction with dinitrophenylhydrazine in
2.3 hydrochloric acid.

Taylor and Kent have also prepared 6-daoxy—6-fluoro-
D-galactose and 5—deoxy~5~f1uoro-D~ribose by reaction of
protected 6a or 5-mesy1 derivatives with anhydrous
potassium fluoride in ethylene glycol at reflux. (35)

6-Fluoro-6~deoxy~D~glucose has been prepared by
Helferich and Gnfichtel (36) and the compound has been used
by Blakeley and Boyer (37) in a study of its effect on
yeast fermentation and on hexokinase. Its competitive
inhibition of glucose and fructose was greater in intact
cells than in cell free extracts and in its competitive
effect on hexokinase. Its toxicity toward rats was moderate
and it did not affect virus synthesis or behavior in tissue
culture. Its effects on glycolysis were not as marked as
found by mcDonald's co~workers, Cramer'and Woodward. (38)
in their studies of the effects of 2-deoxy-D-glucose on the
rate of yeast glycolysis.

Micheal and his students at the University of Munster

have synthesized a number of fluorosugars with the fluoro

10
substituent on the glycosidic carbon or on the primary
carbon. Michael states that‘"of the halosugars with free
hydroxyl groups which carry the halogen on the glycosidic
carbon atom, only those of fluorine are stable.” (39) ;

Other than the 2-fluoro-D,L~glycerose of Taylor and
Kent (34) no other sugars are recorded in the literature
which have fluorine on a secondary carbon.

The route to the synthesis of s 2. or 3-substituted
fluorosugar was selected to use methyl 2,3-anhydro-D-
riboruranoside as the terminal intermediate. This inter-
mediate should sllow the preparation of methyl 3~deoxyo
3-f1uoro-D—zylg-furanoside or methyl 2-deoxy-2~f1uoro-
Deepeningrfuranoside. The production of these as the
result of scission of the 2.3Oepoxide ring is illustrated
by the acid fission of methyl 3,4-anhydro- -D-galactoside (I)
(so) and the alkaline fission of 3.4-anhydro-i,2~O-i30pro-

      

pylidene-gtfuagatose (11).: w (M ) a...» “‘3.
.~°" u on
I '%fifiHP '* o
a u N
I I 0 ——;.(CH,),_ O "" (-‘(C "A.
' Md1> ‘
H S" 6‘. CH‘O _H$ H v 06,0

0":
Interestingly the compound II reacts with base to give

principally the sorbose derivative, but with acetic acid,
acetic anhydride, and a little pyridine the product is
3.4.5-triacetyl~1.2-0-1sopropylidene-D-fructose. This might

11
result from an orthoester type of interaction between an

acetylated 0-5 and C-h. (42,45)
‘ ""C (CH5):

  

cu,

Post (as) has given a thorough review of the chemistry
of the anhydro sugars through 1946. Goodman, Benitez, and
Baker have studied epoxypentene as a model compound for
syntheses involving reactions compatible with 2',3'-anhydro-
xihgrnucleosides. (45)

The formation of anhydro sugars results from the attack
of a base on the proton of a hydroxyl situated Lynne to a
cationogen such as a sulfonic ester. This is displaced in
a cission between the carbon of the sugar and the oxygen
of the ester by the attack of the deprotonated hydroxyl. (d6)

“‘0“ Cute” drum!
0 cm 0 0M.

0 -
4- M50-

 

0H

 

No Acres" 0” COMP°“”J (’47)

Anhydro sugars of several types can be formed: the
ethylene oxide type illustrated above. the propylene oxide,
hydrofuranol, and five member rings containing two oxygens.
(M)

Anhydro rings can be cleaved by either acids or bases,

12
including those of the Lewis definition. Four cases will
be illustrated. : '

With sodium methoxide, i ,2-0oisopropylidene-3, 4-
anhydroaD-psicose is converted principally to 1.2-0-
isopropylidene~4~0emethyl~D-sorbose. whereas with sodium
hydroxide 1.2-0-isopr0pylidene-D-fructose was detected in
the mixture of products. (48) Stereochemical consideration
of the pyranose ring would predict the reaction with a
large base to result in inversion on 0-4. since an enter~
ing group of any size could be alternately equatorial rith
the other substituents in a sorbose configuration.

0 O‘C‘Cfls‘s

O “C\(CH3L a a ‘
' IVQOH
Or
a O CH‘ \

o o-c\(CH.h

H c mo

H
In aqueous solution methyl 3,h-anhydro-p-D-galactoside

gave on hydrolysis with sulfuric acid a mixture of sugars in
which Duglucose and D-gulose were identified. No galactose

is formed. (40)

"‘0“ H.0u H.0u . "‘0"
(3 (”we (3 O
H OH
C) 8 ANJJVO
H
u a

In an earlier publication on the-reactions of this
same methyl 3,hnanhydrofifi-D-galactoside, Muller reported the

introduction of chlorine into the sugar on reaction with

13
normal aqueous hydrogen chloride at steam bath temperature.
(49) This would indicate that the hydronium ion and chloride

ion were responsible for the epoxide opening in this case.

"‘8?! H "bog-l HgO H H H;OH
‘ .,
O O ”.0; ‘1@“cl-a g 0 0‘:- O OH
3 C‘
OH OH H cl H

In as much as Muller did not prove the configuration
of the compounds he obtained, the work of Robertson and
DunlOp (50) and of Pest and Wiggins (46) should be mentioned.
In the former paper it was stated that two methyl chloro-
hexosides were produced when methyl benzylidine-2,3-anhydro-
D-alloside was treated with a very dilute solution of
hydrogen chloride in acetone. These produced the same
methyl 2,3-anhydro-Dnalloside when treated with silver
oxide. The latter paper concerned work done on methyl
2,B-anhydro~dimethyl~D~alloside and methyl 3,4-anhydro-
dimethyl-D-alloside which each cpened to give two chloro-
dimethylhexoses when treated with hydrochloric acid, all
of which were unique.

Another mechanism for Opening snhydrorings is evidenced
in the amination with ammonia in dry alcohol or ether, or,
as in a recent paper (23). with concentrated ammonium
hydroxide. In none of these three cases can the attack of
a base other than ammonia result in the observed product,
and an acid catalysed opening is improbable in view of the
magnitude of the equilibrium EH; m H+.FS’NH3.

14

A Lewis acid Opening of epoxide rings has been described
by Henbest and Wrigley (28) in the preparation of 6,5—f1uoro—
hydrins of sterols. This is a very rapid reaction with
boron trifluoride-ether complex in benzene which has been
followed polarimetrically and terminated by shaking with
aqueous sodium bicarbonate. No mechanism for the donation
of fluorine from the very stable boron trifluoride is given
although discussions of the effect or configurational and]
electrical factors in relation to the fig, §6~epoxids is

treated.

15
‘NATERIALS AND METHODS
All reagents used in this research were of reagent

grade except those used in extractions, which were of .

USP grade. _
Pyridine was dried over calcium hydride and distilled.
p-Toluenesulfonyl chloride was purified by the method

of Pelletier. (57) I
Derivatives which were purified by distillation were

distilled using the distilling head designed by Dr. John

C. Speck, Jr. A
Solvent removal inflxaggg‘was accomplished in most

cases with the flash evaporator Obtainable from Arthur S.

Lapins Company.

Melting points were determined in capillary tubes in

a stirred oil bath.

Intermediates were identified and experimental results
were interpreted with a Perkin-Elmer Model 21 Infrared

Spectrophotometer.

16
EXPmIE'r’IENTAL

The first attempts at a preparation of a 2-(or 3-)
fluoro-e-(or 3-)deoxy-pentose were made using the 2.3-an-
hydrojfl-methyl-ribOpyranoside of Boneyman. (5!)

1.231111 fl-L-WMw-Beginning with 54 g. of
Loarabinose in one liter of 1% methanolic hydrogen chloride
the sugar was converted to methyl fi-L—arabinopyranoside by
refluxing for seven hours. The product (32 g.) crystallized
from the solution on cooling.

mm i.i-Q~immxlidens-p-L-srahinmmmaida
was made by Levenc and Raymond's method (52) for'preparing
diacetone xylose because Honeyman's preparation using
phosphorus pentoxide as the catalyst failed to work.
Twenty-five g. of methylufiLLparabinapyranosids was stirred
with 1 ml. of sulfuric acid (sp. 5. 1.84) and 40 g. of
anhydrous cupric sulfate in 1/2 1. of dry acetone for
65 hours. At this time the mixture began to darken consid-
erably. The acid was neutralized by the addition of 25 g.
of sodium carbonate with continued stirring for 24 hours.
The solution was then separated from the inorganic residue
by drawing it through a sintered glass filterstick, of
medium porosity, into an evaporating flask where it was
concentrated to a syrup. The inorganic residue was thorough-
ly rinsed with acetone a number of times. Considerable
P-methyl arabinosids crystallized out in the evaporator.
The syrup was separated from these crystals and transferred

17
to another flask for distillation. The product distilled
close to the value recorded by Boneyman (82°/b.1 mm.) at
749/0.03 mm.. 809/0.05 mm., and BBQ/0.05 mm. to give a
yield of 16 g.. or 51%.

mm remade-sausammmmeA—L-mn—
p12gngs1de.--The whole amount of the distillate was treated
with 33 g. of p~toluenesulfonyl chloride in 100 m1. of dry
pyridine. After standing at room temperature for five days,
the mixture was poured onto 400 g. or ice (cdbes or about
1 g. each) and allowed to stand for two hours until the
ice melted. The crude 2-O-tosyl-3,Aoo—isOpropylidene«
Fbmethyl-Lnarabinepyrenoside weighed 21.7 g. after drying
over phosphorus pentcxide in Iggng and represented a 76%
yield.

W a-Q-Tnsxl-p-Lramhinmmmide.~-A Sires. .
sample or the compound prepared in the previous step was
dissolved in 275 ml. of methanol to which 5.7 ml. or
0.5 n hydrochloric acid had been added and the solution was
refluxed for 9 hours according to the directions of Honey-
man. Following treatment with 7 g. of silver carbonate. filo
tration, and evaporation, the product was Obtained as a
glass which was exhaustively pumped with the water aspir-
ator and dried forth hours over phosphorus pentoxide at
15 mm . '

hem}. l.&-21~Q~sseixl~2-Q-icsyl-fi-L-W. ~-
The compound obtained above was treated in‘sitn'with 6 ml.

18
of acetic anhydride in 50 m1. of pyridine and allowed to
stand at room temperature for three days. The acetylated'
compound was then obtained by pouring the reaction mixture
onto 1/2 1. of ice and allowing the ice to melt. The
methyl 3,4-di-O-acetyl~2~0~tosyl~arabinoside was oolledted
on a sintered glass filter in 92% yield from the acetone
derivative. Its melting point of 114° compared with
Honeyman's 116°. ,

Hamel 2.1-Anhxm-fl-L-ambimmemidsw-The above
compound (5.9 g.) was dissolved in 75 ml. of methanol and
6 g. of sodium methoxide was added with cooling. .The mix-
ture was allowed to stand at room temperature for two days.
It was then carefully brought to neutrality with acetic
acid using an external indicator and diluted with water
to 500 ml. Extraction with three SO-ml. portions of
chloroform failed to give an evaporation more than a few
drops of 2.B-anhydrojflLmethyl-riboside. A fact which was
discovered too late to be of any use and a mistake which
was repeated in following Baker's preparation of methyl
2.3-anhydroribofuranoside is the inordinately unfavorable
distribution coefficient of the methyl anhydroglycosides
between water and organic solvents. For example, in a
paper by Mothers and Rdbertson (53) washing of a 0.15 3.
sample of methyl 4,6-di-0—methylwe.3-anhydro~dphexoside
with water almost quantitatively eluted the compound into
the water phase, which had Optical activity corresponding to

19
0.13 g. of starting material. The benzene solution had
very little optical activity.

The water phase of the extraction above was evaporated
to dryness and extracted with chloroform to giVe about two
grams of an oil which had the same infrared spectrum as
that obtained from the chloroform extract of the a...
solution and presented an epoxide ~CH2- peak at 3.92 p
and epoxide «0- peaks at 11.6 and 12.1 p.in films between
salt plates.

This material was used in several experiments in
treatment with hydrogen fluoride in reagent grade chloro-
form which had been allowed to stand in contact with an-
hydrous calcium chloride for a week to remove the
preservative amounts of ethanol and had then been purified
by filtration and distillation.

Oneahalf g. was dissolved in 89 g. of chloroform in
a polyethylene bottle and weighed. HydrOgen fluoride gas
was added from a cylinder with protection from moisture
in the air and frequent checks of weight until the equivalent
weight (70 mg.) of hydrogen fluoride had been added. The
mixture was allowed to stand for four hours at room tem-
perature and evaporated to a syrup in a glass beaker within
a filtering bell. Because silicon tetrafluoride would not
interfere with infrared examination of the product. con-

tamination from this source was not of any consequence.

The spectrum was identical with that of the reactant.

More rigorous treatment was given 500 mg. ofthe
epoxide in chloroform with hydrogen fluoride in equimolar‘
concentration at the temperature of boiling acetone over~
night plus two days at room temperature. A syrup which
still retained the epoxide peaks at 11.7 and 12.15‘p
was obtained on evaporation.

A like quantity of the anhydro compound was treated
with an equimolar amount of potassium acid fluoride (REF?)
in 10 ml. of glacial acetic acid at room temperature for
12 hours. Removal of the acetic acid in yagng left a paste
which retained the epoxide peaks.

The chloroform-hydrogen fluoride treatment was repeated

once more withthe same result.

Anderson. Goodman, and Baker‘s preparation (23) of a
methyl anhydrc~riboside was followed for subsequent exper~
iments because it is prepared unequivocally in higher yield
than is Honeyman's and produces methyl 2,3-anhydro~d.and p-
D-ribofuranosides which have widely separated boiling points.
They may be resolved by vacuum distillation through a short
Vigreaux column to give the pure anomers, which are in the
correct (furan) ring for subsequent reactions terminating
in nucleosides.

1.2:}.5-m-Q-ismmndsns-D-mmrsmsswno 1 1. or
acetone, which had been dried with magnesium sulfate and

21
distilled, were added 150 g. of D-xylose and 300 5. of
anhydrous cupric sulfate. Ten ml. of sulfuric acid (sp.
5r. 1.84) were added and the mixture was stirred for 25
hours. Twenty—five g. of anhydrous sodium carbonate was
added at that time with continued stirring overnight. The.
solution was then filtered off and the residue washed with
acetone. The solution was evaporated to a syrup which
distilled at 135-138° at 12 mm. to yield 173 g. of light-
brown diacetone xylose.

1s2rQ?IEQDIQEIIIQEBSPDhlllfliflkfinflfifln*‘Th8 contents of
the flask were washed into 1100 ml. of 0.2% hydrochloric
acid (6 m1. of 37.6% H01 diluted to 1:00 ml.) and shaken
vigorously for 25 minutes. The hydrOgen ion concentration
was then reduced to that corresponding to a pH of 7-8 with
6.5 s. of sodium.bicarhonate and the water*uas removed
overnight at less than 40° C. in,1agug. The residual syrup
was then taken up in 500 ml. of chloroform, treated with a
little “Dares" and filtered on to 20 g. of sodium sulfate
to dry.

The chloroform solution was filtered into the distilling
apparatus and pumped to a syrup with heating to 80° at the
end of the concentration process. Following pumping for
4 hours at 80° and 12 mm.. the pressure was lowered to 0.2
to 0,3 mm. The pot was heated at about too-110° until the
thermometer in the distilling head hogan to rise. Slow
distillation at 82° proceeded until liquid stepped coming

22
over. About 10 5. distilled under these conditions, which
are characteristic of diacetone xylose. The head temper-
ature then rose to 132° and a few drOps were distilled to
flush the side arm of the distilling head. The receiver was
changed and distillation was resumed. At a pot temperature
of 170° and a pressure of 0.3 mm., the product distilled-
at 332-1340. host of the material was distilled and the
receiver again was changed. Approximately 10 a. was dis-
tilled into the third receiver. The product in both the
second and third flasks crystallized. The net weight of
the second flask was 97.5 g. This weight plus about 10 g.
in the third receiver and 10 g. of recovered reactant allows
calculation of a yield of 80% for the conversion and
purification of monoacetone xylose from diacetone xylose.

1.2-Q-W5-Q-W-D-W - --

The 97.5 g. of monoacetone xylose was washed out of the
flash with 200 g.fof chloroform into a three-necked flask
fitted with a drOpping funnel. Two hundred g. of pyridine
was added and the whole mixture was cooled to ~5° C. in an
iceosalt bath. Fifty-five g. of methyl chloroformate were
added dropwise over a period of 1.5 hours with stirring and
maintenance of temperature within the range -5° to 0°. The
mixture was then stored in the refrigerator for three days.
The contents of the flask were then rinsed into 1300 m1. of
water in a separatory funnel and extracted with 150 ml. of

chloroform in four extractions until 600 ml. was used. This

23
extract was washed twice with #00 ml. of water, dried over
sodium sulfate, and evaporated to dryness in,xacng. It
was then taken up in.300 ml. of benzene warmed to boiling
and 300 ml. cf petroleum ether was added. The collected
crystals weighed ‘01 g. and represented a yieldcf 80%.

i .a-Q-Wl~fl~mn-5-Q-mmmml~
Denalgznzanoss.~~The material Obtained above was dissolved
in 300 ml. of dry pyridine and cooled to 0° in an ice bath.
Kinety~five g. of methanesulfcnyl chloride were added. After
allowing it to stand for 30 hours in a closed flask, the
mixture was poured into 1 1. of ice water and extracted
with 1 1. of chloroform in 250~ml. portions. The combined
chloroform extracts were washed with four 250wml. portions
of 1% sodium'bicarhonate solution and twice with 250-ml.
portions of water. The extract use then dried with mag~
nesium sulfate and evaporated to a residue (46.5 g.) which

was used in the next step.

It had been found that solvolysis of the isOpropylidene
group in.methanolic hydrogen chloride resulted in formation
of the xylcpyranosids derivative as well as the desired
furanoside. Intermediate solvolysis to a $.2-di-o-scetyl
derivative followed by methanolic solvolysis circumvented

this undesirable consequence.

1.2-Ei-Q-sseixl-i-Q'mlfifl-W-D- _
zzlgzuzangne.~-The mesylated compound Obtained above (46.5 g.)

was dissolved in 640 ml. of glacial acetic acid which
contained 71 ml. of acetic anhydride. A 39-mli volume

or concentrated culturic acid was then added in small
portions, with mechanical stirring andice cooling. to
maintain the temperature‘below 20°. The flask was covered
and allowed to stand at room temperature for 20 hours.

The contents were then poured into 3 l. of ice-water
mixture and extracted with two 500~ml. portions of ohloro- ‘
form. The combined extracts were washed quickly with a liter
of 1% sodium‘bicarbonate solution in 250-ml. portions
followed by one wash with 250 ml. or water. The ohloro~
term solution was then dried with magnesium sulfate and
evaporated to an amber gum (53.5'3.) which was used in the
next step.

mm i-Q-heafl-fi-Q—W-at any) ~n-mn-
Immangside.-Tho gum was solvolysed to methyl 3-0~mesyl-
5-meethoxycarbonyl-D-xylofuranoside by dissolving it in
1 l. of i% methanolio hydrogen chloride and allowing the
mixture to stand at room temperature for 20 hours. This
mixture was then neutralized by the addition of 32 g. of
sodium bicarbonate. After filtering, the methanol was
removed from this mixture inbzggug. The residue was
extracted with two soc-m1. and one too-ml. portions of
boiling methylene chloridot It was then filtered with the
aid of Celite and evaporated to dryness inbxsgng to give
a residue weighing 35.5 g.

25

=4 e fi,1-fipb‘{d:§-§(§m fi)~2—:12Qf;u:3angsigl,§.--The '
35.5 g. of the methyl glycoside was converted to the
anhydro derivative according to the literature directions
by dissolution in 75 ml. of dry methanol and treatment at
ice temperature with a solution of seven grams of sodium
methoxide in 65 ml. of methanol.“ The flask was stored i
the refrigerator for four days at which time four grams
of Colite was added and the mixture was filtered, neutral-
ized with 3 m1. of acetic acid, and evaporated to dryness
in,yngug. The residue was dissolved in 500 ml. of water
(the literature did not specify the amount) and extracted
with seven 80-ml. portions of chloroform. The chloroform
extract was dried over magnesium sulfate and evaporated to‘
a residue in yogug.

Theta» andlfi-methyl 2,3-anhydro-D~rihofuranoside mixture
was distilled through a short Vigreaux column to obtain a
small yield of methyl 2.3-anhydronflkD-rihofuranoside at
5 p and 50° and of methyl 2,3-anhydro-qu-rihofuranoside
at 3 p and 90°. The original authors obtained clear white
syrups, the more volatile of which was shown by its Optical
rotation and previous deductions concerning its physical
characteristics to he the named compound. Thecxvmethyl
glycoside would be expected to he less fluid because of
hydrogen bonding between molecules and the F-anomer more
fluid because of intramolecular hydrogen bonding between the
5-hydroxyl and the glycosidic oxygen.

26

Numerous experiments were conducted with the methyl
anhydroribofuranosides as follows: V

A 3.35-3. sample was dissolved in 100 ml. of a stand~
ardized solution of hydrogen fluoride in chloroform which
contained 0.2524 millimole of hydrogen fluoride per g.
Samples of the solution were titrated from time to time
to determine whether the epoxide reacted with the hydrogen
fluoride. The solution was contained in a polyethylene
wash bottle. The amount of hydrogen fluoride was deters
mined by weighing a sample into an excess of 0.2099 2
aqueous sodium hydroxide. The flask was shaken vigorously
and the excess of base was titrated with 0.1198 E hydro-
chloric acid to the faintest discernahle pink of the
indicator phenolphthalein. It was thought that overrunning
this end point with acid required a larger than equivalent
amount of base to return to the pink basic form of phenol-
phthalein. This effect was,however, too close to the limit
of experimental error. The discrepancy was thought to be
due to an acid catalysed hydrolysis of a glycosidic fluoro
substituent. (54)

One hundred mg. of the anhydro sugar was treated with
the same weight of potassium acid fluoride in 25 ml. of dry
methanol, in which the salt is slightly soluble. Evaporation,
after 2 days at room temperature, gave a syrup which indicat~
ed no diminution of epoxide absorption at 11.6 p.

The experiment was repeated at the boiling point of

27
methanol at 60° 3 5° for 36 hours with the same result.

Hydrogen fluoride in methanol at a concentraticn'of
655 mg. in 50 m1. of methanol which contained 250 mg. of
the anhydro sugar failed to show altered absorption in the
infrared after 2 days at room temperature.

Since Eflller (49) Opened a sugar epcxide with hydrogen
chloride in aqueous solution under conditions in which it
would have high concentration and ionic character, an
analogous experiment with hydrogen fluoride was trded.

One and one-tenth g. of the anhydrc sugar was dissolved in

5 g. of 9.1% hydrofluoric acid and heated in the steam

bath for 3 hours. It was then neutralized with calcium
carbonate, filtered with the aid of.a little Celite,
evaporated and dried over phosphorus pentoxide at less than
15 mm. overnight. This was also unsuccessful by the infra-
red criterion. The syrup was strongly reducing to Benedict's
reagent, however, and hence had lost its glycosidic methoxyl.

Success appeared imminent when the anhydro sugar was
treated with a slight excess of hydroaen fluoride and a
catalytic trace of boron trirluoride.

A $.55~3. sample of the anhydro sugar was dissolved in
100 ml. of anhydrous methanol. The polyethylene tubing which
led from the valve to the reaction vessel was inserted through
a cork which allowed for some volume change. Hydrogen
fluoride was added from the cylinder. the cylinder valve was

turned off, the needle valve was turned off, then removed

from the cylinder without disconnecting the receiving
bottle. The exposed end of the needle valve was then c1ean~
ed and sucked repeatedly with the aspirator to remove con-
densed hydrOgen fluoride and water. It was then screwed
into the boron triflucridc cylinder and enough boron trin
fluoride was added to make a fog about one-quarter of an
inch deep above the surface or the liquid. The cylinder
valve and needle valve were closed and the bottle was seal-
ed with a polythene cap and left at room temperature for
two days. For lack of any other means of removing solvent
and acid, phosphorus pentoxide was used in a vacuum des-
iccator. The solution was reduced to a viscous syrup much
different from the mobile fluid of the reactant. A small
sample had been removed earlier and dried to a hard glass
which was free of epoxide peaks in the infrared. On fur»
ther drying in the desiccator. the main portion or the
product was converted to a glass which charred badly. This
material was taken up in about 25 ml. of methanol. treated
with a little ”Darco”, and filtered. It was insoluble in
chloroform and hydrocarbons and crystallized to a very
slight extent in the cold from methanol, in which it was
very soluble. Dissolution in a minimal amount of n-pro-
panel produced white crystals on cooling which grew on
standing. These were filtered out with the aid or a small
Hirsch funnel and washed with a small amount of methylene
chloride, whereupon they became a sticky gum. They were

taken up again in prepanol, but were not recovered.

29

A 1.95-3. sample of anhydro sugar was dissolved in
80 g. of anhydrous methanol and 1.73 g. of hydrogen fluoride
including a trace of boron trifluoride was added as de-
scribed above. Subsequent removal of solvent over phos-
phorus pentcxide after two days again resulted in charring.
The charcoal treatment was repeated and on nearing fruition
in the crystallization step the compound was lost through
accident.

A 3.5-5. sample or the snhydro sugar was dissolved in
100 g. of methanol and treated with 1.000 5. or hydrogen
fluoride including a trace of boron trifluoride. After
2 days, 5 g. of anhydrous sodium carbonate was added to
this mixture. carbon dioxide was removed under reduced
pressure. and the solution was filtered and evaporated in
glass on the flash evaporator to Obtain a thin syrup which
resembled the reactant and was shown to have its composition-
by infrared examination.

A 1.75-5. sample of the oL-anomer of the anhydro sugar
was dissolved in 40 g. or methanol and treated with 0.85 g.
of hydrogen fluoride, including a trace or‘boron trirlucride.
This was left for two days at room temperature, split into
two portions, the first or which was treated with 1/2
equivalent (! 5.) of calcium carbonate, evacuated, filtered
with Celite, and evaporated to a syrup which began turning.
dark immediately. The other portion was treated with shout
100 mg. of sodium fluoride and evaporated to dryness over

phosphorus pentoxide at first and finally over crushed

30
magnesium turnings which appeared to remove both hydrogen
fluoride and methanols The residue was a white crystalline
mass. This was extracted with 25 m1. of hot prepanol.
Filtration through a small Hirsch funnel was attempted but
evaporaticn or the solvent precluded this._ By adding a
little methanol, the hot solution was made filterable;
but on partial evaporation the carbohydrate did not crys-
‘ttllize and on evaporation to dryness over phosphorus
pentoxide, the material again charred very badly.

Eleven 5. ot’partially’purified.methyl 2,3-anhydr0c
cL~D~rihofuranoside was dissolved in about 100 ml or an-
hydrous methanol snd treated with 6 g. of hydrogen fluoride
and a trace of boron trifluoride; After tno days at room
temperature the solution was treated with 30 g. or calcium
carbonate and evacuated. The effect of the calcium car—
bonste was more noticable on this scale and it was found
that the solution was still acid after overnight contact.
Twenty-five g. or sodium‘hicarbonate was then added. The
bottle was placed in the vacuum desiccator and evacuated,
whereupon carbon dioxide‘bubbled off vigorously. The solu»
tion was then filtered through Celite and evaporated to a
thin syrup which was not the compound sought. (Infrared

examination showed prominent epoxide peaks.)

31
DISCUSSIos '

It seems feasible to prepare a fluorosugar using an
anhydro sugar intermediate which is subjected to ionic
ring-Opening conditions in methanol. Boron trirluoride
in trace amounts was used to enhance the ionic character
of hydrogen fluoride. Presumably the ringéOpening agent
was fluOboric acid, which is much more ionic than hydrogen
fluoride. I

The compound which was obtained was readily decomposed
by residual boron trifluoride and easily converted to the
anhydro precursor‘by basic materials in methanol, even
such mild bases as sodium‘bicarbonatc. Its value as an
antimetabolite thus is open to question since its existence
in solutions buffered near a‘pH of‘7 might be ephemeral.

If an ionic mechanism of epoxide Opening is necessary,
the failure of short term exposures in aqueous and methan~
olic solutions to hydrOgen fluoride alone is readily underu
standable. since Deussen's data (55) quoted in holler (56)
gives the molor electrical canductivity of aqueous hydrogen
fluoride at a concentration of 0.25‘E,at 25° as 29.6 com-
pared to 366, 377. and 376 for H01, HBr, and HI respectively.

That the initial step in opening the epoxide ring may
be the attack of a proton on the epcxido oxygen would explain
the failure of potassium acid fluoride to‘bring about the
desired reaction. This reagent was used in the hepe that an
electron-sharing mechanism such as is found with Opening by

ammonia would be operative.

32
LILudA“bKM CIT ED

1. Ho 1301881311, £1111. d6 1.1’5icado dB 113913901178, (3). 21'
296 (1890)

2. G. Balz and G. Schiemann, Ber.. 59, 1186 (1929)

3. J. A. Kontgomery and K. hewson, J. Am. Chem. 800.,
12.21559 (1957)

4. G. H! HitChlnga and Cd PD FuhoadB’ Ann. N- Y. ACE-do 501..
Ego art. 2 (1954)

5. A.(Ge1horn and E. Hirschberg. Cancer Res., Suppl. 3,
1 1955

6. A. Bendich. A. Giner~Sorolla, and J. J. Fox, Chemistry
and Biology of the Purines. Ciba Foundation Symposium,
p. 3, Little, Brown, and Company, Boston, 1957

7. G. B. Elion and G. H. Hitchings, inig.. p. 293

8. D; A. Go Galton’ MO, p. 299

9. R. E. F. Matthews, 111g., pp. 270—285

10. B. L. fiutchings. 1h;fi.. p. 177

10a. B. L. Hutchings, 1:1Q.. p. 182; data are those of
Goldman, Karsico. and Angler. unpublished.

1“ A. Bendich 23 file, J. M- CheIBo 5°C.. 15’ 6073 (‘QSAJ

12. %. i0? and G. F. Hawkins, J. Am. Chem. 500., 52, 2443
19 7

13. E. H. Crosser. Kature, 115, 899 (1955)

14. D. A. Lyttle and H. G. Petering, J. Am. Chem. Soo.,
so, 6459 (1958)

15. G. H. HitChinr’ 5 and G. B, 1311011, Ann. L. Y. Acad. 501..
59. 195 (19543

16. R- Karkham and J. D. Smith, 810011613. Jo, 5.2! 558 (1952)
17. E. Volkin and W. E. Cohn. J. Biol. Chem., 295. 767 (1953)

18. J. K. Porter,2&‘g1.. Antibiotics and Chemotherapy,
2, 409 (1952)

19. C. Trincao §L_g1,. Antibiotics Annual, Kedical Encyclo-
pedia, Inc.. flew York, 1956

30.

)8.

39.
no.
41.

33

J. Faiiuenboum and K. Alba, B01. Parisit. Ch11.,
2. 94 31954)

L. Bortle and J. J. Cleson, Antibiotics Annual,
Kedical EnconOpedia, Inc.. Kew York, 1955

R. I. Hewitt .. Antibiotics and Chemotherapy.
g, 1222 (1954

Go Do Anderson, Le Goodman and Bo R; Baker, J. Am.
Chem. 500., go, 6578 (1958)

B. R. Baker, K. Henson, L. Goodman, and A. Benitez,
J. Am. Chem. 8°C., m, 6578 (1958)

J. Fried and E. Saba, J. Am. Chem. 800.. 12. 1130 (1957)
P. L. Perlmsn e; 31., J. Am. Chem. 800., fig, 6687 (1958)
A. Bowers and H. J. Ringold. 151g., p. hh23

H. B. Her best and T. I. Erigley, J. Chem. Soc.. 4765(1957)

G. J. O'Iieill and F. L. K. Pattison, J. Am. Chem. Soc.,
12. 19 56 (1957)

W. F. Pruce and R. deV. IIUber, J. Am. Chem. 800..
‘15, 4668 (1953)

R. R. Fraser, J. E. Killington, and F. L. K. Pattison,
J. Am. Chem. 300.. 19; 1959 (‘95?)

F. L. K. Pottison and J. J. Norman. 1h1fi.. p.2311

F. L. K. Pettison and R. G. Woolford. ihifi.. p. 2306

H. F. Taylor and P. w. Kent, J. Chem. Soc.. 2150 (1956)
I. F. Taylor-end P. W. Kent. J. Chem. Soc., 872 (1958)
B. Helferioh and A. Gnfichtel. Ber., 15. 1035 (1941)

E. R. Blakeley end P. D. Boyer, Biochem. et BiOphys.
Acta. 1&9576 (1955)

F. B. Crsmer and G. E. Woodward, J. Franklin Inst..
r”. 35h (1952)

F. Kicheel £3 31., Ber., 29, 1612 (195?
a hfiller’ Bern. mg 1094 (1935)
H. Chis and C. A. Schultz, Ber.. 112, 2302 (1938)

42.

94.

45.

34

H. L. Frush and H. 3. Iebell, J. Res. Kat. Bur. Stand.,
214 413 (1941)

R. K. EGGS, H. G. FletCher, and C. S. fiudson, Jo Afl.
Chem. 8000, n: 959 (1951)

8. Peat, Advances in Carbohydrate Chemistry, 2,
38-77 (1946)

L. Goodman, A. Benitez, and B. R. Baker, J. Am. Chem.
8°C., £94 1680 (1958)

8. Peat and L. F. Wiggins, J. Chem. 800., 1088 (1938)

A. I-Zfiller, K. Koricz, and GVemer, Ber”, 12;}, 7215 ((1939)
H. ohle and F. Just, Ber..1§§2, 601 (1935)

A. Kflller, Ber.,.612, 421 (1934)

G. J. Rdberteon and H. G. Dunlap, J. Chem. 8°C., 972 (1938)
J. Honeyman, J. Chem. 500.. 987 (1946)

P. A. Levene and A. L. Raymond, J. 8101. Chem..,1Q£,
317 (1933)

D. S. Kathera and G. J. Robertson, J. Chem. Soc..
1076 (1933) ,

B. Ielferich and s. Bettger. Ann.. 316, 150 (1929)

E. Deueeen, Z. anorg. Chem., 35, 300, 408 (1905)

J. 9. Keller, Treatise on Inorganic and Theoretical
Chemistry, Vol. II, p. 131. Loncmane, Green and Company,
London, 1922

S. W. Pelletier, Chem. and Ind.,‘1043 (1953)

-m». 1"" F7“
-._..‘..€'.Y Li’iukuy