that that starting

H . BREDERECK, R . GOMPPER,* H . G. V. SCHTJH,* AND G. THEILIG*

Institut filr Organische Chemie und Organisch-chemische Technologie der Technischen Hochschule Stuttgart

Introduction

In 1938 we succeeded in cleaving yeast nucleic acid to the nucleoside stage by the action of enzyme preparations of vegetable origin. This enabled us to develop a simple and productive method of preparing ribonucleosides, especially guanosine and adenosine (1). Nucleosides, in particular adenosine, were for many years manufactured on a commer­

cial scale by this method (2). A short time later, we were able to effect a chemical cleavage of yeast nucleic acid by means of aqueous pyridine (3). The readily accessible nucleosides then served as starting materials for further syntheses.

The well-known pharmacological properties of the methylxanthines (caffeine, theophylline, theobromine) on the one hand, and of the nucleo­

sides or nucleotides (adenosine, muscle adenylic acid, adenosine triphos­

phate) on the other, induced us to attempt the synthesis of methylated nucleosides and nucleotides in view of their possible pharmacological action. Apart from diazomethane, we used mainly dimethyl sulfate as the methylating agent. We then established that the methylation of nucleo­

sides with dimethyl sulfate yields different methyl derivatives depending on the pH of the solution, and that complete methylation is preferen­

tially effected in a weakly alkaline medium (4). Methylation of xanthine according to this method yielded over 90% of caffeine (5), while a more strongly alkaline medium caused the cleavage of the caffeine formed.

This facile synthesis of caffeine caused us to seek a simple method of preparing xanthine. In 1942/43 we succeeded in obtaining very good yields of xanthine from uric acid and formamide (5). This, our first contact with formamide, already gave us an inkling of the exceptional capabilities of this small organic molecule. The difficult working condi­

tions during the last years of the war, the loss of all the equipment essen­

tial to our nucleic acid work and then the inadequate working conditions under which our research was resumed after an interruption of several

* M y collaborators, Doz. Dr. R. Gompper, Dr. H. G. v. Schuh, and Dr. G. Theilig have played a decisive part in the development of this field. I am also indebted to many others of my co-workers, whose names appear in the references. Bredereck

241

years, all these caused the work on nucleic acid to recede in favor of the formamide reactions—restricted to the purine and pyrimidine field at first—which could be carried out with simple materials.

Syntheses in the Purine a n d Pyrimidine Series

X a n t h i n e from Uric A c i d — S y n t h e s i s o f C a f f e i n e

If uric acid (I) is heated in formamide, xanthine (II) is formed in very good yield {5). Snake excrement may be utilized instead of pure uric acid. Since xanthine can readily be converted into caffeine (III) (5) (see above), this method represents the simplest caffeine synthesis available.

CO CO HN'

VNH

HCONH,

,C-NH OC C-N NH

o cx /

NH

CO CH, (CH,),SO

".C-N'

|-^

pH 8 to 9 '

U

S /

N

CH,

III

If uric acid is replaced by 1-methyl- (6), 3-methyl-, or 1,3-dimethyl- uric acid {5,6), 1-methyl-, 3-methyl-, or 1,3-dimethylxanthine ( = theo­

phylline) are obtained. 8-Thiouric acid can similarly be converted into xanthine with formamide, though in lower yield (6). Uric acids contain­

ing an N-methyl group at position 7 or 9 (7-methyl-, 9-methyl-, 1,3,7-, and 1,3,9-trimethyluric acid) yield no xanthines (6). The reaction is by no means restricted to formamide; on the contrary, formanilide can also be used, though the yields obtained are then lower. The reaction of uric acid with acetamide gives 8-methylxanthine and that with propionamide 8-ethylxanthine (5).

REACTION MECHANISM

The first obvious assumption, namely that the CO or COH group in the 8-position is reduced to CH by the formyl group of formamide, is in­

correct. As mentioned earlier, acetamide and propionamide give 8-alkyl- substituted xanthines. Since rupture of the alkyl-CONH2 bond in these acid amides is out of the question, the C8 atom of the xanthine formed must originate from the CONH2 group of the acid amide. In identical

manner, the C8 of the xanthine formed from uric acid is the C of the formamide GONH2 group. The acid amides consequently cause the fis­

sion of the —NH—CO—NH— grouping in the imidazolone ring of the uric acid.

An indication of the nature of the cleavage is given by the following experiments :

1. The action of formamide on mono- and disubstituted ureas and thioureas results in a fission to give N-substituted formamides (7).

H C O N H .

R — N H — C O — N Ht R - N H - C H O H C O N H ,

R - N H - C O - N H - R * 2 R - N H - C H O

Biuret and cyanuric acid can also be detected in the reaction mixture.

In our own experiments, we used mono- and diphenylurea, mono- and di-p-ethoxyphenylthiourea and 5-ureidouracil. This last compound yields 5-formylaminouracil.

The action of acetamide on substituted ureas correspondingly results in the formation of N-substituted acetamides (acetylamines). The ability of acyl amides to cleave a

—NH—CO—NH bond is demonstrated by these experiments.

2. 4-Amino-5-ureidouracil (IV) reacts with formamide to give xan­

thine and 8-formaminoxanthine (V) (6).

CO

HN^/ V N H - C O - N H , O Cx / C - N H j

N H I V

NH,

\

CO

H N^{7 X}C - N H - C H O

OC

/C-NH,

N H CO

HN

O Cx / C - N N H

HN

C-NH

1 II >

N H „ CO HN^{7 N}C - N H

I II >

OC C - N

NH

C - N H - C H O

Since 8-aminoxanthine does not give xanthine with formamide—the imidazole ring is already present—4-amino-5-ureidouracil cannot be an intermediate in the practically quantitative formation of xanthine from uric acid (6).

The course of the formation of xanthine from uric acid consequently is as follows: The attack by formamide occurs at the CO group ( C8) , in a corresponding manner to the initial step in the action of formamide on ketones (Leuckart reaction). This is followed by the cleavage of the carbinol la (7a) between N7 and C8 and formylation of the amino com­

pound l b thus formed. The ureide structure at C4 (Ic) immediately breaks down to give an amino group (Id) which then reacts with the

formylamino group accompanied by the elimination of water to form xanthine.

In conformity with this reaction scheme, interruption of the reaction between uric acid and acetamide at the proper time allows the isolation of 4-amino-5-acetylaminouracil. Furthermore, l-methyl-4-amino-5- formylaminouracil (corresponding to Id) can be detected chromatog- raphically in the reaction between 1-methyluric acid and formamide (6).

The question arises, whether the conversion of a —NH—CO—NH—

grouping forming part of an imidazolone ring into a —NH—CH=N—

grouping by means of formamide is restricted to uric acid, or whether other ring systems will also undergo a reaction of this type.

C O

H N⁷ V- N R

VNH.

N H

l a

VNH,

C-NH-CHO

c! - N H - C O - - N H - C H O J I- N H - C O - N H - C H O

l b I c C O

XC - N H - C H O H N^/ V N H

- I I - I II >

C - N H , O C / C - N N H

Id II

The incomplete experiments to date have shown the following:

Benzimidazolone cannot be converted into benzimidazole with form­

amide. Xanthine, which also contains a —NH—CO—NH— grouping in its pyrimidine ring, yields hypoxanthine (VI) when the reaction is carried out in an autoclave (200°) (8).

C O C O H N⁷ V N H H N⁷

VNH

I ii ^ i ii \ _

l 11 > — i | ;c

> C , C - N H CV Jt-N

O C , N H

II V I

Uric acid itself can be converted into hypoxanthine directly under similar conditions (9). The formation of decomposition and well-defined reaction products from formamide itself (see below) however, renders the separation of the hypoxanthine both laborious and wasteful.

2,4-Dioxoquinazoline (VII), which results in a good yield of 4-quin- azolone (VIII) by reaction with formamide in an autoclave, possesses

the same arrangement in the pyrimidine ring as do xanthine and uric acid.

VII VIII

The experimental data to date are inadequate for the determination of which structural prerequisites are essential for ensuring the success of the formamide reaction. It may be of decisive importance that the readi­

ness of the cyclic —NH—CO—NH— grouping to add via linkage with electrophilic groups (e.g., CO) be sufficiently great. It may furthermore be essential that the ring systems possess quasi-aromatic character both before and after the "reduction." We are at present determining experi­

mentally the validity of these considerations.

4 , 5- D i a m i n o p y r i m i d i n e s a n d A c i d A m i d e s

The isolation or detection of 4-amino-5-acylaminouracils as inter­

mediates during the conversion of uric acids into xanthines leads one to expect that the reaction of o-diamines with formamide and acid amides generally results in an imidazole ring closure.

Y"

^{RCONH, NpNH -COR ^ Y ™ .}

JLNH, -

» ' /C-NH, -

« ° /LN'

Whereas the reaction between o-phenylenediamines and acid amides to give benzimidazoles has long been known (10), we were the first

(1942/43) to apply this reaction in the pyrimidine series (5). A brief period of boiling with formamide converts 4,5-diaminouracil ( I X ) into xanthine (5). Since that time, numerous diaminopyrimidines have been made to react with formamide, both by us and by other workers.

c o

HN

C-NH.

I

il OC C-NH, NH

IX

After a longer reaction time, 4,5-diaminouracil and acetamide or propionamide give 8-methyl- or 8-ethylxanthine, respectively (5). If the reaction is interrupted, 4-amino-5-acetylamino- or 4-amino-5-pro- pionylaminouracil (X, R = C H3 or C2H5) , respectively, can be isolated as intermediates.

If a 4-amino-5-acylaminouracil is allowed to react with formamide, xanthine is obtained and the acyl group is eliminated as acyl amide (11).

CO CO CO HN

C-NH-COR

HCONH

'

C-NH-CHO HN^

C-NH

' 1-NH, -RCONH,'' 0I | _N H s I 1 ^ ^{C H} oc,

NH

X a

N H " \'H ' ^

If a 4-amino-5-acylaminouracil is allowed to react with a higher amide instead of formamide, the 8-alkylxanthine corresponding to the amide is formed. In general, the amide added in excess will determine the nature of the 8-substituted xanthine formed during the initial acyla- tion, irrespective of the nature of the acyl group on the 5-amino group

(11). As a result of this reaction of acylaminouracils we have devised new syntheses of the purine alkaloids, especially theobromine and theo­

phylline, based on uric acid (see below).

Thiouramil a n d F o r m a m i d e

Thiouramil ( X I ) reacts with formamide to give a good yield of 4,6- dioxo-4,5,6,7-tetrahydrothiazolo-(5,4-d)-pyrimidine (XII) (12).

CO CO HN'

C-NH, HN

C-N

' >H

0

D

J: £-SH OC. I -SH OC

l - s NH NH

X I X I I

The 1,3-dimethyl compound obtained by methylation (dimethyl sulfate, pH 8) is an S analog of theophylline (12).

S y n t h e s e s o f C a f f e i n e , T h e o p h y l l i n e , a n d T h e o b r o m i n e

If uric acid is heated with acetic anhydride in the presence of pyri­

dine, a good yield of the compound designated below as "triacetate" is obtained (13). Mild alkaline saponification or boiling with water causes the loss of one acetyl group to give a "diacetate"; warm alkali causes the loss of another acetyl group from the latter to form a "monoacetate."

Finally, saponification with alkali will also effect the removal of the last acetyl group, resulting in the formation of 4,5-diaminouracil.

Conversely, the lower acetyl compounds can be reacetylated to the higher stages. The following simplified scheme summarizes the results of these experiments (13):

U r i c a c i d

T r i a c e t a t e

M o n o a c e t a t e D i a c e t a t e

4 , 5 - D i a m i n o u r a c i l

The monoacetate has the constitution of 4-amino-5-acetylaminouracil ( X I I I ) , the diacetate that of 8-hydroxy-2,6-dioxo-8-methyl-7-acetyl- 1,2,3,6,8,9-hexahydropurine (XIV) and finally the triacetate that of 8- acetoxy-2,6-dioxo-8-methyl-7-acetyl-l,2,3,6,8,9-hexahydropurine ( X V ) .

The monoacetate and, under the right conditions of temperature (see below), also the di- and triacetate yield xanthine after a short period of warming with formamide (11). All the acetyl groups are eliminated as acetamide during this process. If formamide is replaced by acetamide, the three acetates yield 8-methylxanthine.

The monoacetate reacts with formamide to give largely pure xan­

thine, suitable for further reactions. The di- and triacetate, on the other hand, yield both xanthine as the major product and, resulting from the elimination of acetic acid, 8-methylxanthine (ca. 10%), which renders the further processing of the xanthine more difficult. Under suitable tem­

perature conditions (1 hr at a bath temperature of 100°, followed by boiling for 30 min under reflux) and the use of formamide containing NH3, the formation of 8-methylxanthine can, however, be obviated.

The reaction with formamide discussed above can thus be effected with acetates methylated in positions 1 and 3. Suitable methylation will convert the monoacetate into both the 3-methyl- and 1,3-dimethylmono- acetate (11). It is worth noting that these compounds can also be pre­

pared in very good yield (80-90%) directly from the triacetate in a single operation.

The 1,3-dimethylmonoacetate (XHIb) and the 3-methylmonoacetate ( X l l l a ) can be converted directly into theophylline ( X V I ) and 3-methylxanthine ( X V I I ) , respectively, by boiling with formamide (11).

The latter yields theobromine (XVIII) with dimethyl sulfate (pH 7 to

CO CO C H/ C 3O / O / CO C H , H N C - N H - C O C H , H N C - N C H , H N C - N ^ ^ H , OC C - N H j OC C - N H O H OC C - N H O C O C H ,

N H N H N H

X I I I X I V X V

7.5) (11). The preparation of theophylline can further be simplified by the fact that when the triacetate is used as the starting material, neither the monoacetate nor the dimethylmonoacetate is isolated, but the theo­

phylline prepared in one step (11).

/ O C O C H , H N C - Nx / C H , U r i c a c i d (I) > | || /^Cx

OC C - N H O C O C H , , N H

i

CO CO CO H N^X V N H H N^/ V - N H - C O C H . H . C - N ^ ^ C - N H - C O C H ,

I II > I II I II

OC C - N O CN / C - N H , OC C-NH,

N H N N

I I

I II C H , X H I a C H , X H I b

1 i 1

CO C H CO CO H . C - N C - N H N C - N H H , C - N C - N H

I II > | || > | |i >

O CN / C - N O CN / C - N O Cx / C - N

N N N

I I I

C H , I I I C H , X V I I C H , X V I

^ CO C H ,

I II > ^H

O Cx / C - N N

I

C H , X V I 1 1

The above scheme represents the syntheses which we have developed, based on uric acid, of caffeine, theobromine, and theophylline.

The commercial execution of these syntheses is hindered by the fact that it is at present not possible to prepare uric acid economically from guano. We have therefore attempted to apply our experience in the

formamide field to the total synthesis of the purine alkaloids.

S i m p l i f i c a t i o n o f the T r a u b e X a n t h i n e S y n t h e s i s

The Traube synthesis essentially constitutes the present-day prepara­

tion of xanthines, especially caffeine, theobromine, and theophylline.

This method consists in the nitrosation of the 4-aminopyrimidine deriva­

tive obtained by condensation, reduction of the 4-amino-5-nitroso com­

pound with, e.g. reducing sulfur compounds, to give the 4,5-diamino product, and formylation of the latter followed by ring closure. We were able to simplify this method considerably as follows: The 4-amino com­

pound (e.g., X I X , X X I I I ) in formamide was nitrosated with nitrite-

formic acid and the intermediate products not isolated, the 5-nitroso com­

pound (e.g. X X ) reduced by means of a small quantity (approximately y5 of the theoretical amount) of dithionite, the diamine (e.g., X X I ) formylated (e.g., X X I I ) and cyclized to the xanthine (e.g., X V I , X X I V ) (14)- The yield of xanthine (calculated with respect to 4-aminouracil) amounts to 70-75%, that of theophylline (calculated with respect to l,3-dimethyl-4-aminouracil) 80-85%. The synthetic routes followed in the preparation of theophylline (14) and adenine (15) are reproduced schematically.

S y n t h e s i s of theophylline (XVI):

H.C-N' CO

CH

1 II

OC

CH, N

X I X

CO

N a N O , H C O O H

H , C - N O Cs

C - N H , II C - N H , C H , X X I

H C O N H ,

C O

H,C-N' C - N O^N I Na , S , 04 O CN / C - N H ,

N I C H , X X

C O

H , C - N^{X N}C - N H - C H O oi / C - N H ,

N C H , X X I I C O

H.C-N' \ - N H

I I I >

O C

C-N

N

Besides xanthine, theophylline, and adenine, guanine and 2-thioxan- thine have also been prepared by this route in our Institute.

S y n t h e s i s of a d e n i n e (XXV):

N C H , N ^NC H ,

I + I S = CX C N

N H ,

N H , N CH

A

I II H S - C v yC - N H ,

N X X I I I

N H ,

I

N ^AC H I II C H , - S - CX / C - N H ,

N X X I I I a

N H , N H , 1. H N O , C C

\ H r & S l N C - N H Rv aNi N ' C - N H n e

3. H C O N r if> 1 || xC H R a n e y ^{N i}> J || - C H C H . - S - C . / C - N HC^ C - N

N N X X I V X X V

The synthesis described above can be interrupted at various inter­

mediate stages between the 4-amino compound and the final product.

The reduction effected with dithionite (at 80°) is immediately followed by formylation with the formamide. If the reaction mixture is cooled after the reduction, a good yield of the 4-amino-5-formylamino com­

pound (e.g., X X I I ) is often obtained. The formyl group is readily elimi­

nated with alcoholic hydrochloric acid, yielding the 4,5-diaminopyrimi- dine. Both l,3-dimethyl-4,5-diaminouracil ( X X I ) (14) and 2,4,5-tri- amino-6-hydroxypyrimidine (16), used in syntheses in the pteridine

series, have been obtained by this route.

X a n t h i n e S y n t h e s e s from 5- S u l f a m i n o u r a c i l s

Even though the total syntheses mentioned above and their wide ap­

plicability may constitute the simplest and most economical method, an­

other synthesis of xanthine by means of acid amides should be men­

tioned, particularly because of its especially peculiar nature. Fischer and his co-workers (17) obtained xanthine (80%) or 8-methylxanthine

(80%) by allowing formamide or acetamide, respectively, to react with 5-sulfaminouracil ( X X V I ) (17a), prepared by reducing 5-nitrouracil with dithionite in alkaline solution [for theory cf. ref. (17)].

CO

H f / V- N H S O 3 H H C O N H2

X X V I

The reaction between 3-methyl-5-sulfaminouracil and formamide or acetamide correspondingly yields 3-methylxanthine (85%) or 3,8-di- methylxanthine (70%) (14)- l,3-Dimethyl-5-sulfaminouracil, on the other hand, only gives l,3-dimethyl-5-formylaminouracil with form­

amide, and l,3-dimethyl-5-acetaminouracil with acetamide; i.e., the sul­

fonic acid is merely replaced by an acyl group (14)- The study of this reaction led us to another specific reaction undergone by formamide.

A C - f o r m y l a t i o n : 4- A m i n o u r a c i l —> 4- A m i n o- 5- f o r m y l u r a c i l

In addition to the reaction of the 5-sulfamino compound, we at­

tempted to prepare a 4-sulfaminouracil and to allow the latter to react with formamide. By treating 4-aminouracil with chlorosulfonic acid, we obtained a compound which we initially regarded as being uracil-4-sulf- aminouracil. We believed we had converted this compound with form­

amide into 4-formylaminouracil (14) and that methylation of the latter had given us l,3-dimethyl-4-formylaminouracil, which can also be pre-

co

H N^{7 X}C - N H CH O Cx / C - N

N H II

pared from l,3-dimethyl-4-aminouracil and formamide or formic acid.

More recent investigations have, however, shown the alleged 1,3-di- methyl-4-formylaminouracil to be 2,4,6,8-tetramethyl-l,3,7,9-tetraoxo- 6ctahydrobipyrimido-[4,5-b,4',5'-e]-pyridine ( X X V I I I ) (18). Its forma­

tion from l,3-dimethyl-4-aminouracil ( X I X ) can be explained as follows:

c o c o c o

H.C-N' ^NC H H C O N H . H,C-N' ^XC - C H O He' ^XN - C H , I II -> I II || I 0 CN / C - N H , - N H , O CN / C - N H , H2N - C CO

N N N I

I I

C H , C H3 C H ,

X I X X X V I I

J-NH,

CO C H CO CO C CO

H j C - N⁷ \ ' ^ C^{7 X}N - C H , H n H a C - N⁷ V \ ' ^XN - C H ,

i ii i i r

° i ii ii i

0 \ /C ^C\ / ^{C O} x /^Cx /^cx /^{C O 0 C} N N N N N H N

I I I I

C H , C H , C H , C H , X X V I I I

The first step therefore consists of a C-formylation at position 5. 1,3- Dimethyl-4-amino-5-formyluracil (XXVII) can in fact be isolated from the filtrate of X X V I I I . This 5-formyluracil reacts further with unchanged l,3-dimethyl-4-aminouracil to give X X V I I I .

If the reaction between l,3-dimethyl-4-aminouracil and formamide is carried out around 140°, 2,4-dimethyl-l,3-dioxotetrahydropyrimido-[4, 5-d]-pyrimidine is obtained.

According to this, the alleged 4-formylaminouracil and 4-sulfamino- uracil could in reality be the unmethylated bipyrimidopyridine deriva­

tive (tetraoxooctahydrobipyrimidopyridine) and 4-aminouracil-5-sulfonic acid, respectively. It is worth noting that the formylation occurs not at the amino, but at the CH group. This reaction nevertheless becomes comprehensible when it is realized that the amino group is linked to a vinylene-homologous acid amide grouping. It is nonetheless surprising that the C5 atom undergoes formylation rather than condensation to an aminomethylene compound (see below). On the basis of the reactions of dimethylbarbituric acid, to be discussed next, such a reaction lies well within the realm of possibility.

D i m e t h y l b a r b i t u r i c A c i d a n d A c i d A m i d e s

Even though reactions analogous to the experiments described in this section were already known, our results are nevertheless briefly stated again, firstly, since our investigations were originally undertaken from

a different viewpoint and secondly, in order to point out the fresh reac­

tion possibilities displayed by the acid amides.

As reported below in greater detail, ketones can be brominated in a single-stage reaction in formamide; the a-bromoketone formed can be converted by means of an excess of formamide into an imidazole by raising the temperature. We extended this reaction to dimethylbarbituric acid in the expectation that here too an imidazole ring would be formed

{19). The reaction (170°) however, resulted in the formation of 5-amino- methylene-l,3-dimethylbarbituric acid and a compound which for the time being we consider to be N,C-bis[l,3-dimethylbarbituryl-5]-diamino- methane ( X X X ) (19).

Proceeding from the fact that dibromodimethylbarbituric acid ( X X I X ) is partially formed first, the formation of X X X can be ex­

plained by the following reaction scheme:

CO CO CO CO

H.C-N' ^XC B r , H C O N H , 3^HC~^{N / X}C ( O H )2 H,C— ^XC O H C O NH2 HJC-N' ^XC H - N H - C H O O CN 7C 0 ^> O CN / C O ~ * O Cx / C O ^> O Cx / C O

N N

N

I I I .

C H , C H , C H , \ C H , X X I X

CO CO CO CO \ CO H , C - N^{X X}C H - N H - C H - C H ^XN - C H , M„ H , C - N^{/ X}C H - N H - C H - C H ^XN - C H , \ H-C^ ^XN — C H ,

I I I I I £T± | i I I I I I O Cx / C O

H

N

N N N N

N

I I I ! I

C H , X X X C H , C H , C H , C H ,

The aminomethylene compound X X X I I a simultaneously produced can also be obtained directly from dimethylbarbituric acid ( X X X I ) and formamide (19).

CO CO H ^ - N_{OC / C O}^{7 X}C H , H C O N H , ^{> H} OC / C O

9

-^ V

~

*

N N I I C H , C H , X X X I X X X I I a

Barbituric acid had already been converted in this manner by Papini and Cimmarusti (20) and S. Hunig (21).

™ ?

H a C - f / ^XC = C - N H , b : R = C H , oi CO ^: R = ^Cc*^H«

V d : R » n - C , H7

CIH,

X X X I I b - d

The reaction between dimethylbarbituric acid and higher acid amides yields the corresponding aminomethylene compounds (19) X X X I I b - d .

l,3-Dimethyl-4-chlorouracil ( X X X I I I ) reacts with acid amides to give the same aminomethylene compounds (19):

co

H.C-N'

CH

ocx C - C l

N

I C H , X X X I I I

C H , C O N H ,

CO

H.C-N' CH ^N

0<t / C - O H

N

CH.

C O

C O C H ,

H.C-N' C = C - N H , ^X I I O C ^ / C O

I X X X I I b C H ,

C O

H . C - N C H , = C H - N H - C , H0 § «H^C~ N O - C H - N H - C . H , I I

O CX / C O N C H , X X X I

r I I

ocx C O N

I C H , X X X I V C O

H . C - N⁷ C H - C H = N - C « H^X 5

I I

O CX / C O

I X X X I V a C H ,

This conversion readily fits into the scheme of the reactions between halogen compounds and acid amides (see below). If formanilide, being an N-substituted amide, is allowed to react with 1,3-dimethylbarbituric acid, 5-anilinomethylene-1,3-dimethylbarbituric acid ( X X X I V ) is ob­

tained, though the examination of its IR spectrum shows this compound to exist in the form of a Sehiff base ( X X X I V a ) (19). Barbituric acid reacts in an identical manner (20).

Reaction with dimethylformamide on the other hand results in the formation of the dimethylammonium salt of l,3-dimethyl-5-(r,3'-di- methylbarbituryl-5')-rnethylenebarbituric acid ( X X X V ) (19). In this instance too, Hiinig (21) had previously obtained the corresponding compound in the reaction with barbituric acid.

If the initial step in the course of the reaction between acid amides and compounds containing an acidic CH group is assumed to be the formation of the adduct X X X V I , it is noteworthy that route A (elimina­

tion of water) is followed in the reaction with formamide, formanilide, and higher acid amides, whereas route B or C is followed in the reaction with dimethylformamide. The formation of X X X V can be explained

CO CO H.C-N

C—CH=C

N-CH,

I I I I I

OC C-OH OC /CO N N (!H, CH,

X X X V

either by the direct route ( C ) , or via the hydroxymethylene compound X X X V I I (see below).

X X X I H C O N R ,

CO O H H a C - ^ ^NC H - C H - N R ,

o cx CO

| X X X V I C H ,

- R , N H

CO

H , C - N C = C H - N RM 0_ H , C - N C = C H O HH XI X X

oc / i o

N I C H

2. HCI T y p e XXXIIa

OC / C O N

- > X X X V

C H . ^{X X}" ^{X V}

The 5-aminomethylene-l,3-dimethylbarbituric acid (XXXIIa) is in its turn capable of undergoing numerous reactions. Sodium hydroxide, for example, causes the loss of N H3 to give the sodium salt of 5-hydroxy- methylene-l,3-dimethylbarbituric acid ( X X X V I I ) ; acidification of the latter yields compound X X X V via the tautomeric 5-formyl compound (19). X X X V had previously been obtained by the condensation of di­

methylbarbituric acid and formic acid (22).

The amino group in the aminomethylene compound can also be re­

placed by the hydrazine group (19).

co

H.C-N' C = C H - N H , ^X

I I

O CX / C O N

N H . N H ,

X X X I I a

CH.-N' CO

C=CH-NHNH,

o ix co

N CH,

The hydrazine compound X X X V I I I is itself capable of undergoing numerous reactions (19).

Acid Amides a n d o-Phenylenediamine

In addition to the reactions between 4,5-diaminouracil and acid amides, we have also undertaken corresponding reactions with other 1,2- diamines. Reactions between aromatic o-diamines or their hydrochlorides

and acid amides were already known (10). We have supplemented these experiments by reactions with dicarboxylic acid amides (23,24).

If o-phenylenediamine is heated with oxamide, malonamide, or suc- cinamide, 2,2'-dibenzimidazolyl ( X X X I X a ) , 2-dibenzimidazolylmethane

( X X X I X b ) , or 2- (1,2-dibenzimidazolyl)ethane ( X X X I X c ) is obtained (H).

O-NHI ^{+ H}TN - C O - ( C H2 )- C O - N HN 2 +

"'^Q

^ ^ - N H H N , / : - ( C H2)N- < ;

- N N

+ 2 N H , + 2 H , 0

X X X I X

: n = 0 ; b : n = l ; c : n = 2

Formamide a n d Compounds Containing a

— N H — C O — G r o u p i n g

The reaction of mono- and disubstituted ureas with formamide, yield­

ing formylamines via the fission of the CO—NH— bond, has been men­

tioned earlier in connection with the discussion on the mechanism of the urea -»xanthine reaction. This consists in effect of a transacylation,

during the course of which the carbaminyl moiety is replaced by a formyl group.

H C O N H2

R - N H - C O - N H , - + R - N H - C H O

Thus phenyl- and diphenylurea, p-tolylurea, and a-naphthylurea yield formanilide, p-formtoluidide, and a-formnaphthalide, respectively

(25). The urea simultaneously produced during the course of the reac­

tion is partially converted into biuret and cyanuric acid.

Piperazinedione reacts with formamide at 130-145° (the presence of a small quantity of water is advantageous) to give a good yield of formylglycylglycinamide ( X L ) . The piperazinedione ring is consequently unilaterally opened (26, 28).

/ C O N H , H , C N H H C O N H , H2C N H - C H O

I I • I I

H N C H , H N C H2

C O C O X L

Micheel (27) observed an N-formylation only, during the reaction of amino acids and peptides with formamide or formamide-formic acid at 60°.

e-Caprolactam undergoes ring opening to give equally good yields of c-formylaminocaproamide ( X L I ) (28).

( C Ht) » H C O N H , O C H - N H ( C Ha) , - C O N H , H N CO * X L I

Stable lactams, e.g. dihydrocarbostyril, and butyrolactone, do not undergo fission under these conditions (28). Reaction of the polyamide from adipic acid and hexamethylenediamine (Ultramide A) gives a mix­

ture of adipic acid diamide and N,N'-diformylhexamethylenediamine, from which both of these compounds can be isolated (28).

Syntheses of Imidazoles

From A c y l o i n s

Novelli (29) had allowed benzoin to react with NH3-containing formamide (prepared by the distillation of ammonium formate) and obtained diphenylimidazole with pyrazine as by-product; the formation of the latter is accounted for by the ammonia contained in the form­

amide. We allowed a large number of aliphatic, aromatic, and hetero­

cyclic acyloins to react with formamide at 150-180° and invariably ob­

tained good yields of 4,5-disubstituted imidazoles (30) (Table 1).

T A B L E 1

imidazoles from Acyloins and Formamide

Acyloln (XLII) Imidazole (XLIV) Yield (%)

Acetoin 4,5-Dimethy 1- 55

Propionoin 4,5-Diethyl- 67

n -Butyroin 4,5-Di-n -propyl- 81

Iso^yityroin 4,5-Diiaopropyl- 85

n -Valeroin 4,5-Di-w-butyl- 72

Isovaleroin 4,5-DiUobutyl- 61

n -Caproin 4,5-Di-«-arayl- 51

n -Enanthoin 4,5-Di-n-hexyl- 45

1,6-Diphenylpropionoin 4,5-BisO-phenylethyl)- 66

Cyclodecan- l-ol-2-one

(Sebacoin) 4,5-Octamethylene- 87

Cyclohexadecan-1 -ol-2-one

(Thapsoin) 4,5-Tetradecamethylene- 91

Benzoin 4,5-Diphenyl- 91

Furoin 4,5-Di-a-f ury 1- 89

/>-Dimethylaminobenzoin 4 (5) - (p -Dimethy laminopheny 1) -

5(4)-phenyl- 62

a - Hydroxybuty rophenone 4(5)-Phenyl-5(4)-ethyi- 70

As illustrated by sebacoin and thapsoin, the —CH(OH)—CO—

grouping can also form part of a large ring, to which the imidazole ring then becomes joined.

In accordance with Novelli (29), we assume the following reaction path for the formation of the imidazoles (30):

R - C O H C O N H2 R - C

| ^> | N H C H O _H jO⁾ R ' - C H O H R ' - C H O H

X L I I

R - C H - N H C H O R - C H - N H C H O

| H C O N Ht | / OH >

R ' - C O R ' - C - HtO

X L I I I N H C H O R - C - N H C H O R - C - N H

^ H R ' - C - N H C H O - H C O O H ' _CR_ N

X L I V

The correctness of this scheme is supported by two empirical observa­

tions: Firstly, it is possible to isolate the intermediate X L I I I (a-form- aminoketone) and to convert it into imidazole with formamide (31)

(see below). Secondly, by the fact that the theoretically possible route from the formaminoketone via an oxazole XLIIIa can be excluded since a-formaminoketones give oxazoles to but a minor extent if heated in the absence of a condensing agent. The most convincing argument, however, lies in the fact that oxazoles are not converted into imidazoles at 150°

(32) (see below).

R - C - N H C H O R - C - N R-C-1ST

II - II > " > || > «

R - C - O H R - C - O R - C - N H X L I I I X L I I I a X L I V

From a-Diketones

Since substituted benzoins are frequently prepared via the a-dike- tones (i.e. benzils), we checked whether «-diketones could not be uti­

lized directly in the synthesis of imidazole. The importance of the direct conversion of a-diketones into imidazoles follows from the fact that aromatic a-hydroxyketones are often more difficult to prepare directly than are the a-diketones obtained with oxalyl chloride via a Friedel- Crafts reaction.

The action of formamide and formaldehyde on a benzil at 180-200°

results in a good yield of the corresponding 4,5-disubstituted imidazole (33) (Table 2).

Formaldehyde does not, as originally thought, act as a reducing agent in this case; no a-hydroxyketone can be detected during the reaction.

The formation of the imidazole must rather be explained by the genera­

tion of ammonia from formamide at 180-200°, and consequent reaction between the diketone, ammonia, and formaldehyde.

O C H , C H

If formaldehyde is replaced by other aldehydes, imidazoles substi­

tuted in the 2-position are formed (Table 2).

T A B L E 2

Imidazoles from a-Diketonesf Aldehydes, and Formamide

Benzil Aldehyde Imidazole Yield (%) M.p.

Benzil Par af or m aldehyde 4, 5-Diphenyl- 67 231°

2, 2' -Dichloro- Paraformaldehyde 4. 5-Bis-(2'-chlorophenyl)- 45 237°

4, 4 ' - D i b r o m o - Paraformaldehyde 4, 5 - B i s - ( 4 ' -bromophenyl)- 42 261°

4, 4'-Dimethoxy- Paraformaldehyde 4. 5-Bis-(4'-methoxyphenyl)- 52 184°

Benzil Benzaldehyde 2, 4, 5-Triphenyl- 61 274°

4, 4' -Dibromo- Benzaldehyde 2-Phenyl-4, 5 - b i s - ( 4 ' -bromophenyl)- 36 297°

Benzil 3, 4-Dichloro- benzaldehyde

4, 5-Diphenyl-2-(3\ 4'-dichloro-

phenyl)- 52 243°

Benzil 2-Chlorobenzaldehyde 4, 5-Diphenyl-2-(2'-chlorophenyl)- 20 192°

Benzil 4-Diethylamino- 4, 5-Diphenyl-2-(4'-diethylamino- 27 250°

benzaldehyde phenyl) -

This synthesis therefore resembles the well-known synthesis from a-diketones, ammonia, and aldehydes, devised by Radziszewski (34).

The latter, however, does not proceed uniformly; the action of the am­

monia causes the partial decomposition of the diketone, with the result that difficult-to-separate mixtures of di- and trisubstituted imidazoles are obtained. The yields are correspondingly unsatisfactory. The advan­

tage of our method resides in the fact that the ammonia becomes avail­

able only gradually by the decomposition of the formamide. Further­

more, our reaction is complete within 3 hr, whereas the method of Radziszewski requires several days.

Benzil and substituted benzils could be converted into imidazoles with formamide in the presence of formic acid/sodium bisulfite (34a) (Table 3), in a manner analogous to the preparation of imidazoles from

T A B L E 3

Imidazoles from a-Diketones and Formamide in the Presence of NaHSO,/HCOOH

Benzil Imidazole Yield (%) M.p.

Benzil 4. 5-Diphenyl- 78 231°

2. 2'-Dichloro- 4. 5-Bis(2-chlorophenyl)- 40 237°

4. 4'-Dibromo- 4, 5-Bis(4^r-bromophenyl)- 36 261°

4. 4 '-Diphenoxy- 4. 5-Bis(4'-phenoxyphenyl)- 43 214°

4. 4'-Bis-phenylmercapto- 4. 5-Bis(4 '-phenylmercaptophenyl)- 33 174°

2, 2'-Dimethoxy- 4, 5 - B i s ( 2 m e t h o x y phenyl)- 57 193°

isonitroso compounds and formamide in the presence of reducing agents (see below).

The reaction proceeds via the benzoin stage; if it is interrupted soon after the addition of the reducing agent, a 95% yield of benzoin is iso­

lated when the starting compound is benzil: The imidazole is then pro­

duced via the mechanism followed by the reaction between an a-hy- droxyketone and formamide (see above). Sodium bisulfite and sodium dithionite proved to be the most suitable reducing agents. In conjunction with formamide-formic acid, a 5-20% equimolar quantity of the sulfur compound is adequate. This "catalytic" action of the bisulfite is due to the reduction by formamide-formic acid to the sulfide; this then effects the reduction of the benzil to benzoin, being itself oxidized to elementary sulfur. The latter is then once more reduced by formamide-formic acid to sulfide.

H S O J + H C O O H / H C O N H2 — >

Benzil - > B e n z o i n

s 2_

H C O N H2/ H C O O H

Good yields of imidazole are accordingly also obtained with ele­

mentary sulfur or the sulfides of sodium or potassium instead of bisul­

fite. The formamide-formic acid mixture is also capable of reducing the diketone alone, but the yields do not then exceed 20-30%.

From tt-Aminoketones

Following the detection of the a-formaminoketone as intermediate in the synthesis of imidazole from a-hydroxyketone and formamide, we were able to convert several a-hydroxyketones (priopionoin, butyroin, benzoin) into the corresponding a-formaminoketones in yields of ap­

proximately 50% with formamide under certain conditions (31).

The conversion of a-aminoketones into imidazoles by means of form­

amide, briefly sketched earlier, proceeds extremely readily. Desylamine, for example, gives 4,5-diphenylimidazole in 90% yield (30). It is ad­

vantageous that the a-aminoketones (XLVI), accessible only with dif­

ficulty, can be replaced by their precursors, namely the a-isonitroso- ketones ( X L V ) . The isonitrosoketones are reduced in formamide at 70- 100°, and the imidazole ring then closed at a more elevated temperature.

Hydrogen and platinum oxide or sodium bisulfite (in catalytic amounts) can be used as reducing agent.

H2 H C O N H .

R - C O - C - R ' —4 R - C O - C H - R ' - > R - C C - R'

II I I I

N O H N Ha H NX / N

X L V X L V I C H X L I V

From a- H a l o k e t o n e s

The use of a-haloketones offers a further possibility in the synthesis of imidazoles. Unlike the reaction involving a-hydroxyketones or a-di- ketones, this method also permits the preparation of 4(5)-monosubsti- tuted imidazoles (30). It furthermore enables the imidazole ring to be­

come linked to cycloaliphatic a-haloketones [35). Table 4 shows the imidazoles which we have prepared by this method.

T A B L E 4

Imidazoles from a-Haloketones and Formamide

a-Haloketone Imidazole M.p. Yield (%)

u>-Bromoacetophenone 4(5)-Phenyl- 128° 90

a -Bromo-isovalerophenone 4(5) - Phenyl- 5(4) - isopropyl- 200° 69

a -Bromo-n - propyl ketone 4(5)-Ethyl-5(4)-» -propyl- 162°

(Hydrochloride) 49 3 -Bromobutanone 4,5-Dimethylimidazole

+ 4,5-Dimethyloxazole

265°

(Hydrochloride) 47 22 m -Nitro-tu -bromoacetophenone 4(5)-(w -Nitrophenyl)- 224° 72 P - Met hoxy - -br o moac et ophe none 4(5) - (/>-Methoxyphenyl)- 137° 62 a -Chlorocyclohexanone 4,5,6,7-Tetrahydrobenz-

(= 4,5-Cyclotetramethylene-)

149° 63

a -Chlorocycloheptanone 4,5-Cyclopentamethylene- 202° 85

a -Chlorocyclooctanone 4,5-Cyclohexamethylene- 166° 70

A substantial simplification of this process consists in brominating the ketones in formamide, and directly converting the a-bromoketones so formed into the imidazoles by heating with an excess of formamide.

T A B L E 5

Imidazoles from Ketones, Bromine, and Formamide

Ketone Imidazole Yield (%)

Acetophenone 4(5)-Phenyl- 61

Propiophenone 4(5)-Phenyl-5(4)-methyl- 80

Butyrophenone 4(5)-Phenyl-5(4)-ethyl- 8 1 . 5

Methyl benzyl ketone 4(5)-Phenyl-5(4)-methyl- 43

Deoxybenzoin 4,5-Diphenyl- 67

The initial step in the formation of imidazole from a-haloketones is the replacement of the halogen by an OH group (see below); the a-hydroxyketone is then converted into the imidazole according to the mechanism depicted earlier.

R - C H - C O - R ' — > R - C H - C O - R ' — > R - C = C - R '

I I I I

Hal O H H N N C H

The imidazole syntheses described succeed only with formamide, and not with higher acid amides.

From O x a z o l e s

Blumlein (36) and Lewy (37) obtained a small yield of 4-phenyl- oxazole from w-bromoacetophenone and a two molar quantity of formamide at 130°. After we had obtained a 90% yield of 4 (5)-phenyl- imidazole from w-bromoacetophenone and formamide at 150-180°, the assumption naturally suggested itself that oxazoles constitute an inter­

mediate in the imidazole synthesis. This assumption has, however, proved to be incorrect.

We discovered that oxazoles can be converted into imidazoles by heating at 180-190° in formamide or by reaction with formamide/am- monia in an autoclave (32,38). Since oxazoles can be prepared by a variety of processes (see next section), this method constitutes yet a further possible synthesis of imidazoles. The yields lie between 70 and 90%, and the method has thus far only failed in the case of benzoxazole and 2,4,5-triethyloxazole. The latter decomposes as a result of the vigor­

ous reaction conditions. Inasmuch as the oxazoles themselves are stable therefore, the conversion of oxazoles into imidazoles is theoretically pos­

sible in all cases.

The replacement of the oxazole oxygen by NH presupposes a ring opening as the first step. This can be regarded as either an ammonolytic

(by the action of the ammonia liberated by the formamide on boiling or the ammonia added as reactant) or a hydrolytic reaction (formamide invariably contains small quantities of water; water is furthermore lib­

erated during the course of the reaction). A number of oxazoles had been converted into imidazoles earlier, by means of alcoholic ammonia or alcoholic aniline in a sealed tube at 300° (39). As far as comparable data regarding yields are available, those obtained by the formamide method are found to be substantially higher. The ammonolytic fission is certainly not the only effect influencing this reaction. This is indicated by the fact that while 2,4,5-triphenyloxazole remains unchanged in con­

tact with ammonia in an autoclave at 210°, it affords an 85% yield of triphenylimidazole in the presence of formamide. Formamide appears on the one hand to allow salt formation with the oxazole [quaternary oxazolium salts are converted into imidazoles far more readily than are the free oxazoles (39a)] by the formation of formic acid, and on the other, to supply the nucleophilic agent simultaneously by decomposition to ammonia. To this must be added the solvent and condensing proper­

ties of formamide. We postulate the following reaction course on the basis of an ammonolytic fission (without protonation):

R - C

N H ,

N H .

R - C H — C I Ns

R - C = C - R ' N ^ O H I I

C - I R "

- N H ,

N H , C - N H , R "

R - C =

V I R

R - C H - C - R ' I II N ^ O

C - N H , R "

C - R ' I N H

Synthesis o f I m i d a z o l e

The imidazole synthesis, described above, from a-haloketones and formamide, allows the preparation of the parent substance of the series.

In order to protect the aldehyde group in the bromoacetaldehyde we converted the latter into the acetal with ethylene glycol, and allowed the acetal (XLVII) to react with formamide at 180°, while ammonia was passed through the mixture. Imidazole (XLVIII) is formed in around 60% yield (40).

C H , - B r

| O - C H , H C O N H , C H - N H H CX |

• i| >

O - C H , C H - N X L V I I X L V I I I

This method constitutes the simplest and most productive imidazole synthesis. The preparation from tartaric acid, via dinitrotartaric acid and imidazole-4,5-dicarboxylic acid, for example, gives a yield of 29-30%

(calculated with respect to tartaric acid) (41)-

Further possibilities of synthesizing imidazoles are discussed later.

Syntheses of Oxazoles

From A c y l o i n Esters a n d the A m m o n i u m S a l t s o f C a r b o x y l i c A c i d s

The conversion, discussed earlier, of oxazoles into imidazoles, for which we required a large number of oxazoles, induced us to develop the existing oxazole syntheses and to supplement them with new ones. Some of these syntheses also utilize formamide, and others are indirectly con­

nected with it. These methods are therefore briefly discussed.

Davidson and co-workers (42) had obtained 2-methyl-4,5-diphenyl- oxazole from benzoin acetate and ammonium acetate in glacial acetic acid; the corresponding reaction with benzoin benzoate gave them 2,4,5- triphenyloxazole. This method enabled us to prepare a series of hitherto largely unknown oxazoles (L) (32) using the esters of aliphatic or mixed