Tetrahedron Letters,Vol.23,No.49,pp 5107-5110,1982 oo40-4039/82/495107-04$03.00/O Printed in Creat Brltaln 01982 Pergamon Press Ltd.

PERACID-MEDIATED N-OXIDATION AND REARRANGEMENT OF DIMETHYLPHOSPHORAMIDES -

PLAUSIBLE MODEL FOR OXIDATIVE BIOACTIVATION OF THE CARCINOGEN HEXAMETHYLPHOSPHORAMIDE (HMPA)

Ian Holden*, YOffl Segalla, Ella C Klmmel and John E Caslda

Pestlclde Chemistry and Toxicology Laboratory, Department of Entomological Sciences

Unlverslty of Callfornla, Berkeley, Callfornla 94720

Summary Dlmethylphosphoramldes react with m-chloroperoxybenzolc acid

(MCPBA) In anhydrous acetone to yield the previously unknown P-dlmethylamlno- -

oxyphosphonous derlvatlves via N-oxldatlon and rearrangement Further MCPBA --

oxldatlon yields formaldehyde and nltrosomethane, Isolated as its trans-dlmer

These reactlons provide a possible blomlmetlc model for the metabolic actlva-

tlon of hexamethylphosphoramlde as a mutagen and carcinogen

HexamethylphosphoramIde (HMPA) (la), an extensively-used solvent In organic synthesis, 1s

a powerful rodent carcinogen 1

It 1s detected as a mutagen In a cell transformation assay but

only when activated by a liver enzyme preparation 2 Metabolism of HMPA In rats and liver

preparations Involves E-demethylatlon with llberatlon of formaldehyde, possibly via the F-

hydroxymethyl, y-methyl derlvatlve from rearrangement of an x-oxide 394 Analogous reactlons

occur on metabolic conversion of the lnsectlclde octamethyldlphosphoramlde (schradan) [(Me2N)2-

P(0)OP(O)(NMe2)2] to a potent cholinesterase InhibItor 334 These metabolic reactlons and

actlvatlon of the lnsectlclde are reproduced In part on oxldatlon with permanganate or a per-

acid 3'4 The existence of dlmethylphosphoramlde N-oxldes 1s not established despite their -

potential Importance as bloactlvated metabolltes or Intermediates In the blologlcal activity

of dlmethylphosphoramldes We therefore examined the reactlons of HMPA and related compounds

with m-chloroperoxybenzolc acid (MCPBA) as a possible blomlmetlc model (Scheme)

The reactlon of HMPA with MCPBA (> 99%) In anhydrous d6-acetone at 25'C was monltored

directly by 'H and 13

C NMR (Internal tetramethylsilane) and 31

P NMR [referenced from external

(Me0)3P(O) In CDC13] with parallel studies on _G5 (which has the advantage of a single dl-

methylamino group) and on their possible oxldatlon products. Treatment of lb for 24 hr with

up to two equivalents of MCPBA gave strong 31P NMR (6 +3.48) and 'H NMR signals (6 2 81,

slnglet) These chemical shifts and particularly the lack of 31 1

P- H coupling In the 'H NMR

suggested the formatlon of dlmethylamlnooxyphosphonous derivative 2, Instead of dlmethyl-

phosphoramlde E-oxide 2b, a speculation conflrmed by synthesis and spectral comparison. 6

The analogous product 3a from HMPA was slmllarly ldentlfled 6

- Thus, the dlmethylphosphoramlde

z-oxide (2) 1s strongly lmpllcated as an lntermedlate undergolng a rearrangement reactlon to the

dlmethylamlnooxyphosphonous derlvatlve (3) 7

aCurrent address Israel Institute for Blologlcal Research, Ness-Zlona, P 0 B 19, 70450, Israel

5107

5108

lo,lb 2 3o,3b

0

cHj>N-BR, HOCH2

13b 12b

kH20 /

[ l-

CHS\ f -CH20

HyN-o-PR2

‘M LOI

CH3\ 8

CH30 /N-O-PRO

5 1 ’ 0

CH3\N /!R 8 R [ 1

w H/- 2 CH3’N-O-PR2

H’ Ar PRO

14b 7 6b

‘1 @I ‘\ P

+

CH3~N-4p Ar =

- [CH,NO] + HO- P

:I

0’ ‘CH3 CH3NHOH

I

0 0

[ 1 CH3\t ,N-0-!R2 CH3

4

II IO

t id 9

J

o R = N(CH3)2

b R = OC,H,

Scheme Oxldatlon of hexamethylphosphoramIde (la) and g,g-dlethyl &&dlmethylphosphoramlde

(Ib) to the corresponding E-dlmethylamlnooxyphosphonous derlvatlves (&, 32 and

nltrosomethane dlmer (11)

One dlmethylamlno group of HMPA and this moiety of lb form two terminal products in near -~

quantitative yield (lH NMR) wlthln 24 hr on treatment of &and &wl'ch a five-fold excess of

MCPBA. Disappearance of the N,E-dlmethyl doublet (lH 6 -2 7) IS accompanied by appearance of

singlets corresponding to formaldehyde (IH 6 8.15) and trans-nltrosomethane dlmer (u (lH

6 3.85). Formaldehyde (steam dlstlllatlon) was further characterized as Its methylene bls- __

dlmedone derlvatlve (m p 189-190°C) The IdentIty of dlmer z was conflrmed by 13C NMR'( 6

47.2) and chemical lonlzatlon (methane)-mass spectrometry (CI-MS) (M+l+ 91) on the crude reac-

tlon mixture and by preparative TLC lsolatlon and comparison with synthetic CIS- and trans-

dlmers 8

Nltromethane (lH 6 4 35) 1s not present In the reactlon mixture and would be detected

If formed since It 1s reslstant to MCPBA oxldatlon Dlmer 1 and formaldehyde were the prlncl-

pal terminal products onMCPBA oxldatlon not only of &and l&but also of schradan and tetra-

5109

methylphosphorodlamldlc chloride [(Me2N)2P(O)Cl] This dlmer was not formed from O,O-

dlphenyl N,E-dlmethylphosphoramlde Phosphorus-contalnlng products from @ evident by CI-MS

were &b and (J,g,O,O-tetraethyl pyrophosphate

Potential IntermedIates In the conversIon of g and j& to 11 and formaldehyde were examined -

by subJectlng them to MCPBA oxldatlon under the same condltlons DImethylamlnooxyphosphonous

derlvatlves &and 3b quantltatlvely ylelded &II wlthln 2 hr - Other potential organophosphorus

Intermediate? (12b, 13b and 14b) were treated with excess MCPBA but gave no dlmer (II) E-

Methylhydroxylamlne @) (free base in acetone) reacted Instantaneously with MCPBA, glvlng an

Intense, transient blue flash due to nltrosomethane (10) and quantltatlve formatlon of dlmer 11

These flndlngs establish two portlons of the pathway In the Scheme, 1 e 2 + 2 + 2 and 4 -f J& +

11 Formaldehyde llberatlon might Involve formatlon and degradation of the E-hydroxymethyl, N- -

methylphosphoramlde (2) but this would generate the corresponding monomethylphosphoramlde (2) -~

which was not detected , & reacts slowly with MCPBA formlng small amounts of 11 but only after

several days lndlcatlng It IS not an IntermedIate In this oxldatlon The proposed rearrangement

reactlon for conversIon of 5 to 2 accorunodates the subsequent formaldehyde llberatlon on forming

2 and provides for an oxldatlve pathway In addltlon to that from lnltlal methyl hydroxylatlon

The reaction sequence for conversion of 3 to 11 IS uncertain but clearly requires two further _ --

oxldatlon steps Alternatlve pathways are shown In the Scheme, one :ia N-methylhydroxylamine _-

(10) and the other via N-oxide 7, in each case with concommltant formation of the appropriate

phosphoric acid or anhydrlde (8b or tetraethyl pyrophosphate)

HMPA readily undergoes bIologIca y-demethylatlon to J& and formaldehyde 4

Unidentified

metabolltes may Include some of the compounds shown In the Scheme The carclnogenlclty of

formaldehyde" may contrlbute to that of HMPA 11

but seems InsuffIcIent to account for the high

potency of this phosphoramlde !J-Methylhydroxylamlne 1s a mutagen 12

and might react directly 13

or require further oxldatlon, e g to nltrosomethane (10) Thus, 9 gives >I with MCPBA (this _

study) or cls-11 with periodate 14

via 10 -= Although not an alkylating agent itself, nitroso-

methane might be expected to react mth nucleophlles at nitrogen Nltrosomethane tautomerlzes

to formaldehyde oxlme, 15

which might also contribute to the overall bIologIca actlvlty

We have used the Ames mutagenesis assay 16

with Salmonella fyphlmurlum strain TA-100 to

evaluate the mutagenic actlvlty of all compounds (except 8) indicated without brackets In the =

Scheme and also formaldehyde oxlme, formaldoxlme HCl and nltromethane None of these compounds

was detected as a mutagen (< 0 05 revertants/ug) even on addltlon of the microsomal actlvatlon

system The Ames assay may not be an appropriate lndlcator of carclnogenlclty In the present 2

series or the ultimate mutagen and carcinogen may be a compound other than those tested, eg.

proposed Intermediate 5 or 10 = _ In addltlon, the reactlon mixture of HMPA and MCPBA IS highly

mutagenic due to a single trace component (- 7000 revertants/pg) as we ~111 detail elsewhere

Care should therefore be taken In using HMPA with peraclds and possibly other oxidants

Acknowledgment Supported in part by Natlonal Institutes of Health Grant PO1 ES00049. Helpful

comments were provided by L 0 Ruzo and W M Draper of this laboratory

5110

6

7

a

9

10

11.

12

13

14

15

16.

References and Notes

Nasal turblnate tumors are Induced in rats exposed to a vapor level of 50 ppb HMPA for one

year J A Zapp,Jr,& Ind a - Ass J 36, 916 (1975), K P Lee, H J --

Trochlmowlcz, and C F Relnhardt, The Toxlcologlst 1, 128 (1981)

J Ashby, J A. Styles, and D Anderson, Br. 3 Cancer 36, 564 (1977), J Ashby, J. A --

Styles, and D Paton, Br J Cancer 2, 418 (1978) --

B W Arthur and J E Caslda, J Econ Entomol 2, 49 (1958) -___

A R Jones and H Jackson, Blochem Pharmacol 17, 2247 (1968)

Compound lb [(Et0)2P(0)Cl/excess gaseous dlmethylamlne/Et20 at O'C], b.p 38"C/O 2 mm Hg, lH

6 4 04 (4H, dq, J 7 Hz each), 2 70 (6~, d, J 10 HZ), 1 34 (6~, t, J 7 HZ), 31~ 6 +a 22

Compounds & and & were synthesized by reacting the relevant chlorldate In tetrahydrofuran -

for 5 hr at 25°C with N,E-dlmethylhydroxylamlne (free base from refluxlng HCl salt with

equlv NaH In tetrahydrofuran) In the presence of Et3N 3&, b p 8%87'C/O 3 mm Hg, 'H 6

2 75 (6~, s), 2 68 (12H, d, J 10 HZ) 3h b p 68"C/O 05 mm Hg, 'H S 4 19 (4H, dq, J 7

Hz each), 2 al (6~, s), i 36 (6~, dt, J 1 HZ, 7 HZ), 31~ 6 +3 48.

Analogous rearrangements are proposed for phosphorothlolate S-oxldes to phosphlnyloxysulfen- _

ates, 1 e =P(O)-S(O)-alkyl + =P(O)-0-S-alkyl [Y Segall and 3 E Caslda, Tetrahedron

Lett 3, 139 (1982)i On a slmllar basis, oxldatlon of Me2N-SO*-NMe2, m p 69-70°C (lH

6 2 83), with 3 equlv MCPBA gives a single product (6 2 83, 3 10, 3H each) tentatively

suggested to be Me2N-0-S02-NMe2

cls-Nltrosomethane dlmer ('H d CDC13 4 2) was obtalned by vapor phase photolysls of t-butyl _

nltrlte (254 nm/quartz) and trans-dlmer (lH 6 CDC13 3 9) by thermolysls of cls-dlmer

[(C S Coe and T F Doumanl, J Am Chem Sot _--_ 70, 1516 (1948)] or peracld oxldatlon of

E-(benzylldene)methylamlne [K G Taylor, M -S Chl, and M S Clark, Jr , J a m

41, 1131 (1976)]

Compound &z& (made as for 2 and & with N,g-dlmethylhydroxylamlne), b p 63-64'C/O 3 mm Hg,

1 H 6 4 20 (4H, dq, J 7 Hz each), 3 50 (3H, s), 2 92 (3H, d, J 12 Hz),1 28 (6H,t,J7Hz)

mm (from Ub_/l equlv 40% aqueous formaldehyde/trace Na2C03 for 18 hr at 25OC),

decomposltlon on attempted dlstlllatlon, 'H 6 4 62 (2H, d, J 15 Hz), 4 06 (4H, dq, J 7 Hz

each), 2 75 (3~, d, J 9 HZ), 1 30 (6~, t, J 7 HZ) 14b (made as for $ll: with excess

gaseous methylamlne), b p 84"C/O 2 mm Hg, 1 H d 4 10 (4H, dq, J 7 Hz each), 2 70 (6H, d,

J 10 Hz), 1 30 (6~, t, J 7 Hz)

F Perera and C Petlto, Science 216, 1285 (1982) ___

A R Dahl, W M Hadley, F F Hahn, .I M Benson, and R 0 McClellan, Science 16, 57 ~ ~___

(1982)

P Marfey and E RobInson, Mutation Res 86, 155 (1981) _ -=

C Janlon and D Shugar, In Molecular Mechanisms of Genetlc Processes, N B Dublnln and

D M Gol'dfarb, Ed , pp 106, Halsted Press, Wiley, New York, 1975

T Emery and J B Nellands, J Am Chem Sot 82, 4903 (1960) - __-

K A Jensen and A Helm, Mat -Fys Medd -K Dan Vldensk Selsk 40, 1 (1978), Chem __~_~ __-

Abstr 89, ii04aob 0978)

B N Ames, 3 McCann, and E Yamasakl, Mutation Res 3l, 347 (1975)

(Received In USA 30 August 1982)