new strategy for the synthesis of functionalized phosphonic acids

Heteroatom Chemistryq 1997 John Wiley & Sons, Inc. 1042-7163/97/020103-20 103

Heteroatom ChemistryVolume 8, Number 2, 1997

New Strategy for the Synthesis ofFunctionalized Phosphonic Acids*Chengye Yuan,† Shusen Li, Chaozhong Li, Shoujun Chen,Weisheng Huang, Guoquan Wang, Chun Pan, and Yixin ZhangShanghai Institute of Organic Chemistry, Chinese Academy of Sciences, 354 Feng Ling Lu,Shanghai 200032, China

Received 8 July 1996; revised 6 September 1996

ABSTRACT

Various approaches leading to mono-, di-, and poly-functionalized phosphonates as well as phosphorylheterocycles are reported for the systematic study ofrelationships between chemical structure and biolog-ical activity of this most important class of organo-phosphorus compounds. q 1997 John Wiley & Sons,Inc.

INTRODUCTION

Functionalization of phosphonic acid moleculesopens enormous possibilities for structural variationof this important class of organophosphorus com-pounds with potential biological activity. This is par-ticularly true for 1-aminoalkylphosphonic acids andrelated peptides for which have been found wide ap-plications as both agrochemicals and medicinalproducts. Numerous synthetic approaches havebeen described for the formation of monofunction-alized phosphonic acids, which usually require mul-tistage manipulations. Introduction of additionalfunctionalized groups to these molecules by tradi-

*This article was presented as an invited lecture at 4th Inter-national Conference on Heteroatom Chemistry, Seoul, Korea, Au-gust 1995. The abstract was published in Pure and Applied Chem-istry, 68, 1996, 907.

†To whom correspondence should be addressed.

tional methods is complicated due to inter- or intra-molecular interactions. Consequently, the design ofsyntheses of mono-, di-, and polyfunctionalizedphosphonic acids and study of the structural effectsof these compounds on their biological activity areof great interest both in synthetic organophosphoruschemistry and for the development of leads for thestructures for biologically active molecules. In thisarticle, our emphasis is placed on the description ofnew and convenient methods of synthesis.

RESULTS AND DISCUSSION

Monofunctionalized Phosphonic Acids

Alkylidenebisphosphonates. Condensation ofaldehydes with diethyl phosphite, followed by sul-fonylation of the 1-hydroxy group thereby formedwith methanesulfonyl chloride, and subsequent sub-stitution by another molecule of diethyl phosphiteprovides the corresponding alkylidenebisphospho-nates in one pot procedures with good yields [1].Upon acid hydrolysis, the resultant alkylidenebis-phosphonates are converted smoothly to 1-alkyl-methylenebisphosphonic acids [2] containing a ba-sic skeleton for compounds with potentialtherapeutic effects in the treatment of osteoporosis[2,3].

1-Heteroatom-Substituted Phosphonic Acids.

1-Heteroatom-substituted phosphonic acids cover awide spectrum of important organophosphorus

104 Yuan et al.

SCHEME 1

SCHEME 2

compounds. Unfortunately, the conventional syn-thetic method based on nucleophilic substitution ofa 1-haloalkylphosphonate is under great challengeassociated with low reactivity of the 1-haloalkylgroup, the narrow scope of applications, and therelative inaccessibility of the intermediates. As foundby us, substitution of a 1 (methanesulfonyl)al-kylphosphonate by a phosphite anion provides an al-kylidenebisphosphonate. Since the -OMs group be-haves as a good leaving group, this nucleophilic sub-stitution process was extended as a general methodof preparation of 1-heteroatom-substituted alkyl-phosphonates (Scheme 1).

The yields in the reactions were increased withthe enhancement of the nucleophilicity of the rea-gents as shown by the order:

n-BuSH . PhSH . NH NH . PhNHNH2 2 2

; PhCH NH . (EtO) POH . PhCONHNH2 2 2 2

k KCN k KF

Diethyl 1-hydroxypropylphosphonate providedthe products in much higher yields than the corre-sponding 1-phenyl-1-hydroxymethylphosphonatesdue to the greater steric hindrance of phenyl overethyl in these nucleophilic substitution reactions.

1-Aminophosphonic Acids.

As phosphorus analogues of a-aminocarboxylic ac-ids, 1-aminoalkylphosphonic acids are the most im-portant members of the 1-heteroatom-substitutedphosphonic acids because of their potential biolog-ical activity. As the mimetic tetrahedral intermedi-ates of hydrolyzed esters, amides, and peptides, 1-aminophosphonic acid derivatives have been used asantibiotics, medicaments, or enzyme inhibitors [4].

Among numerous synthetic methods for thepreparation of 1-aminoalkylphosphonic acids, athree-component condensation involving an alde-

New Strategy for the Synthesis of Functionalized Phosphonic Acids 105

hyde, an alkyl carbamate, and a trivalent phospho-rus reagent is the most attractive. A systematic studyis being carried on in our laboratory with the aim ofdeveloping new and convenient methods for the syn-thesis of 1-aminoalkylphosphonic acids based on theinvestigation of the influence of substrate structure,nature of the catalyst and the type of solvent on theyield of product [5–7].

By using acetyl chloride as the solvent, the scopeof applications of such three-component condensa-tions can be expanded to include the convenient syn-thesis of derivatives of 1-aminoalkylphosphonates,key intermediates for phosphonopeptide synthesis[8–10] (Scheme 2).

The application of N-protected alkylphosphonicmonoesters and diesters in phosphonopeptide syn-thesis has also been demonstrated by us [11]. Thus,dialkyl 1-(N-benzyloxycarbonylamino)alkylphos-phonates were converted to 1-aminoalkylphosphon-ates in almost quantitative yields upon catalytic hy-drogenolysis using palladium on carbon as the cat-alyst. The latter was directly coupled, withoutisolation, with an N-protected amino acid using di-cyclohexylcarbodiimide (DCCI) plus N-hydroxyben-zotriazole (HOBt) as condensation agents, produc-ing phosphonopeptides containing a C-terminalaminophosphonic moiety. On the other hand, alkylhydrogen 1-(N-benzyloxycarbonylamino)alkyl-phosphonates, upon treatment with thionyl chloridein the presence of triethylamine, followed by con-densation with an aminoalkylcarboxylate or ami-noalkylphosphonate, afforded phosphonopeptidesbearing an N-terminal aminophosphonic moiety(Scheme 3).

When an a-halocarboxamide was used in placeof a benzyl carbamate in the above tricomponentcondensation, a backbone for the formation of oli-gophosphonopeptides was established [12]. Thus,condensation of 2-halogenated amides, substitutedbenzaldehydes, and a dialkyl phosphite in the pres-ence of acetic anhydride containing ethanolic hydro-gen chloride provided a-(2-haloacylamino)-substi-tuted benzylphosphonates, which were reacted withphthalimide in the presence of potassium carbonate,using tetraethylammonium bromide as a phasetransfer catalyst, followed by hydrazinolysis, to givedipeptides with a free amino group. This type of di-peptide was coupled with another N-protectedamino acid to form a tripeptide. It is obvious thatanalogous couplings on both N- and P-terminal moi-eties should furnish oligophosphonopeptides(Scheme 4).

Since the biological activities of 1-aminophos-phonic acids are largely dependent on their absolute

configurations, the asymmetric synthesis of thesecompounds has aroused the interest of organicchemists [13]. The nucleophilic addition of a dialkylphosphite to chiral imines or oxoiminium deriva-tives have constituted the majority of asymmetricsyntheses of 1-aminophosphonic acids to date, al-though several additional interesting syntheticmethods have been developed [14–16].

Gilmore and McBride [17] have briefly investi-gated the stereochemical behavior of the addition ofdiakyl phosphites to aldimines resulting from thecondensation of substituted benzaldehydes with (R)or (S)-1-phenylethylamine at 1408C. Since a hightemperature is not favorable for asymmetric induc-tion, we studied this reaction at ambient tempera-ture in the presence of various catalysts in order toimprove stereoselectivity. We found that the natureof catalysts and solvents influenced remarkably thediastereomeric excess (de) values attained and theinduced direction of the asymmetric addition [18].The addition reaction of diethyl phosphite to N-ben-zylidene-(S)-1-phenylethylamine was examined as atypical example. With AlCl3 or BF3 •Et2O as catalystand CH2Cl2 as solvent, the de value was found to be70% or 61%, respectively, based on 31P NMR spec-troscopic measurements. Upon hydrolysis to cleavethe ester and hydrogenolysis to remove the a-phen-ylethyl group, the adduct formed by the use ofBF3Et2O as the catalyst was converted into the a-aminobenzylphosphonic acid with [a]D 1 18.3,which was determined to have the S-configurationby comparison with the literature [19] reportedvalue of [a]D 1 17.6. This indicates that the (1S, 18S)diastereoisomer is the major product of the additionreaction. However, with ZnCl2 or TsOH as catalystand CH2Cl2 as solvent, or BF3 •Et2O as catalyst andtoluene as solvent, or without use of a catalyst, thisreaction gave the (1R, 18S) diastereoisomers pre-dominantly. These results were reasonably explainedby our molecular mechanics studies. Calculation in-dicates that three stable conformations of the imineare involved in this reaction and the relative energiesare shown to be in the order C , A , B. In the ab-sence of catalyst or in the presence of ZnCl2 or TsOH,attack of diethyl phosphite on the favored conformerC takes place from the re face of the imine due to asteric interaction, leading to mainly the (1R, 18S)diastereoisomer. When BF3Et2O or ZnCl2 is used asthe catalyst, the empty d-orbital of the central atomis used to form an intramolecular chelate complexwith the lone pair electrons of nitrogen and thephenyl at C-18, which renders conformer A predom-inant. Attack of diethyl phosphite at conformer Afrom the less hindered direction, namely, the si face

106 Yuan et al.

SCHEME 3

SCHEME 4


SCHEME 5

SCHEME 6

of the imine, provides the (1S, 18S) diastereoisomers.The opposite stereofacial selectivity, with toluene assolvent, can be rationalized by the strong coordina-tion of BF3 with the solvent that prevents the for-mation of the intramolecular coordination complex(Scheme 5).

Based on the above results, we reasoned that if

the methyl group were replaced by the methoxy-methylene or methoxycarbonyl group, introductionof catalysts that coordinated strongly with an oxygenatom would bring about a greater contribution ofconformer C, and the asymmetric induction effectwould be improved significantly. Consequently, weinvestigated the reaction using (R)-2-methoxy-1-

108 Yuan et al.

phenylethylamine or (R)-phenylglycine methyl esteras chiral auxiliaries instead of (S)-1-phenylethyl-amine. Here, the relative spatial arrangements of thegroups linked to the chiral carbon remain un-changed, and the conformational analysis men-tioned previously should be applicable.

With (R)-phenylglycine methyl ester as the chiralauxiliary, this asymmetric addition with BF3Et2O ascatalyst and CH2Cl2 as solvent gave a de value of 89%with (1R, 18R) as the major configuration of theproduct, in addition to an 84% chemical yield. Othercatalysts, such as AlCl3, TiCl4, and ZnCl2, furnished(1R, 18R)-diastereoisomers predominantly withgood to excellent de values regardless of the exis-tence or nonexistence of an empty d-orbital of thecentral metal atom. These results can be explainedby predominant contribution of conformer C as aresult of the coordination of the Lewis acid with thecarbonyl group. The de values are dependent on thecoordination ability of the catalysts, and therefore,BF3 •Et2O gives the best de values.

Analogous results were obtained with (R)-2-methoxy-1-phenylethylamine as chiral auxiliary. Ourinvestigations represent the first successful trial ofmolecular mechanics calculations as a guide inasymmetric syntheses of 1-aminophosphonic acids[19,20]. These experimental data were strongly sup-ported by Smith’s report [14].

It is interesting to note that, as one of the het-erobimetallic asymmetric catalysts, the Lanthaniod-Potassium-BINOL complex was used in an effectiveasymmetric hydrophosphonylation of an imine withexcellent, up to 96%, de values [21].

1- and/or 2-Hydrazinoalkylphosphonic Acids.

As derivatives of aminoalkylphosphonic acids, hy-drazinoalkylphosphonic acids are of considerableinterest as compounds with potential biological ac-tivity. These compounds have been prepared by theaddition of dialkyl phosphites to aldazines, followedby acid hydrolysis [22]. However, this method is notapplicable to the synthesis of the esters of 1-hydra-zinoalkylphosphonic acids since acid hydrolysisleads to the cleavage of the ester linkage, while cat-alytic hydrogenolysis to remove the N-protectivegroup usually results in N–N bond splitting to formthe 1-aminoalkylphosphonic ester [23]. Another syn-thetic route to these compounds was based on theaddition of a dialkyl phosphite to the hydrazone re-sulting from condensation of formaldehyde with N-(benzyloxycarbonyl)hydrazine. The adducts thusformed, upon removal of the benzyloxycarbonylgroup by catalytic hydrogenation, gave the 1-hydra-

zinomethylphosphonates [24]. Unfortunately, othersubstituted hydrazones are inert to nucleophilic ad-dition of a dialkyl phosphite [25].

As found by us, starting from diethyl 1-keto-phosphonates, 1-hydrazonophosphonates can beprepared smoothly. In this reaction, the presence ofglacial acetic acid is critical since direct condensa-tion of hydrazine or hydrazine hydrate with 1-keto-phosphonates usually results in cleavage of the P–Cbond to form acyl hydrazines and the dialkyl phos-phite. 1-Hydrazonophosphonates, upon treatmentwith NaBH3CN, were conveniently reduced to thecorresponding hydrazino derivatives that were iso-lated as their oxalates due to their air sensitivity. Thefree 1-hydrazinoalkylphosphonic acids were ob-tained from the hydrazinophosphonates upon acidhydrolysis [26] (Scheme 6).

The 2-hydrazinoethylphosphonates were pre-pared analogously from 2-ketoethylphosphonatesusing BH3THF as the reducing reagent, which is ableto convert 2-hydrazono derivatives to 2-hydrazino-ethylphosphonates in high yield.

As shown by us, a direct synthetic route to 2-hydrazinoethylphosphonic acids was achieved basedon the addition of the carbanion derived from di-ethyl methylphosphonate to the aldimine preparedconveniently by condensation of an aromatic alde-hyde with hydrazine. Difficulty was encounteredduring the hydrogenolysis of the N4C bond since itis usually accompanied by N–N bond cleavage. How-ever, removal of the 4CHR group can be achievedby acid treatment with 1N HCl, although the yield ispoor (around 30%). The corresponding 2-hydrazino-2-alkyl(aryl)ethylphosphonic acids are easily ob-tained from the diethyl esters by reaction withMe3SiBr and subsequent treatment with methanol inthe usual manner (Scheme 7).

1-Hydroxyamino-alkylphosphonic Acids.

1-(Hydroxyamino)alkylphosphonic acids are a classof important compounds possessing strong antibac-terial activity [27]. A number of syntheses have beendeveloped to provide a convenient route to this classof compounds, including the controlled reduction of1-nitroalkylphosphonates with zinc and ammoniumchloride [28], the addition of a dialkyl phosphite toa 1-oxoaldoxime at an elevated temperature [29], thenucleophilic addition of a lithium or potassium di-alkyl phosphite to N-glycosylnitrone followed by gly-coside cleavage and hydrolysis [30], and the conden-sation of 1-(benzyloxyamino)alkylphosphonic acidwith an O-alkylisourea, followed by treatment withboron tris(trifluoroacetate) [31]. These methods are,


SCHEME 7

SCHEME 8

SCHEME 9

SCHEME 10

however, limited by their narrow scope, other com-petitive reactions, and relatively inaccessible start-ing materials. We established a new and facile ap-proach to 1-(hydroxyamino)alkylphosphonic acidsby controlled reduction of 1-(hydroxyimino)alkylphosphonates with the borane-pyridine com-plex [32].

Standard Arbuzov reactions of acyl chlorideswith trialkyl phosphites gave 1-oxophosphonates,

which were converted into 1-(hydroxyimino)-alkylphosphonates in almost quantitative yield byoximation with hydroxylamine hydrochloride in thepresence of pyridine. Reduction of the oximes inalcoholic hydrogen chloride with an excess of theborane-pyridine complex provided 1-(hydroxy-amino)alkylphosphonates as crystalline hydrochlo-ride salts that were converted into the correspondingfree acids either by hydrolysis with 6N HCl at 908C

110 Yuan et al.

for 4 hours in 45–86% yield or by dealkylation withMe3SiBr in acetonitrile, followed by treatment withaqueous methanol (Scheme 8).

DIFUNCTIONALIZED PHOSPHONIC ACIDS

1-Hydroxy-2-aminoalkylphosphonic Acid andDerivatives Thereof

As a phosphorus analogue of an isomer of serine, 1-hydroxy-2-aminoethylphosphonic acid has been iso-lated from a living organism [33]. This awakened anincreased interest in the synthesis of this unusualaminophosphonic acid containing a hydroxyl group.1-Alkyl derivatives of 1-hydroxy-2-aminoethylphos-phonic acid were prepared either by addition of di-methyl phosphite to N-(2-oxoethyl)phthalimide, fol-lowed by hydrazinolysis [34], or by aminativecleavage of the ethylene oxide ring of phosphomycinderivatives, followed by subsequent hydrolysis [35].A serious drawback of these methods is associatedwith starting materials attainable only with diffi-culty, and therefore, a new approach is necessary. Wefound that careful reduction of 1-hydroxy-2-nitroal-kylphosphonates could achieve this goal.

Preparation of 1-hydroxy-2-nitroethylphosphon-ates by nucleophilic addition of nitromethane toacetylphosphonates in the presence of di- or trieth-ylamine has been reported [36]. We found that thisprocedure is not applicable to the synthesis of di-ethyl 1-hydroxy-2-nitro-2-phenylethylphosphonatessince the electron-withdrawing ability of the ben-zene ring destabilizes the molecule toward alkali andresults in C–P or C–C bond cleavage. We achieved aneffective and general procedure for the preparationof 1-hydroxy-2-nitroalkylphosphonates based on theaddition of nitromethane to acyl- or aroylphospho-nates under phase transfer catalysis conditions, us-ing potassium carbonate/tetrabutylammonium bro-mide as the catalyst [37].

The conversion of the nitro group into an aminofunctionality by catalytic hydrogenation over Raney-Nickel in acetic acid provided 1-hydroxy-2-aminoal-kylphosphonates. Here, acetic acid plays an impor-tant and interesting role. It serves not only as solventbut also as an accelerator of the hydrogenation pro-cess. It also prevents the decomposition of the re-action products since the basicity of the formedamino group is strong enough to initiate the C–P andC–C cleavage of the resultant molecule. Without iso-lation, each generated 1-hydroxy-2-aminoalkylphos-phonate was directly refluxed in 8N HCl at 1008C togive the corresponding 1-hydroxy-2-aminoalkyl-phosphonic acid [38].

The controlled reduction of 1-hydroxy-2-nitroal-

kylphosphonates to 1-hydroxy-2-(hydroxylamino)-alkylphosphonates was also realized by treatment ei-ther with aluminum amalgam in ethyl acetate orwith stannous chloride in hydrochloric acid solution[39]. The yield is very much dependent on the struc-ture of the substrate and the reactivity of the Al–Hg.However, reduction with SnCl2 usually provided bet-ter yields at low temperature (,108C), regardless ofthe structure of the substrates (Scheme 9).

Aminophosphonic Acids Bearing aTrifluoromethyl Moiety

Introduction of a fluorine atom, a difluoromethylenegroup, or a trifluoromethyl group into organic mol-ecules is the subject of active investigation due tosome special properties of the fluorine atom com-pared with the hydrogen atom. For the purpose ofchanging the biological activity, a series of trifluo-romethyl-amino-alkylphosphonic acids was synthe-sized by us. Our synthetic approach is based on thechemical reactivity of trifluoroacetimidoyl chloride.Thus, Arbuzov-type reactions of trialkyl phosphiteswith this reagent, followed by subsequent reductionto the imine intermediates, provided 1-(N-aryl/alky-lamino)-2,2,2-trifluoroethylphosphonates, the tri-fluoromethylated phosphorus analogues of N-pro-tected alanine [40] (Scheme 10).

The reaction of triethyl phosphite with N-phen-yltrifluoroacetimidoyl chloride was carried out at808C without solvent. The 19F NMR chemical shift forthe starting imidoyl chloride is 4.8 while that for thecompound that was formed is 9.7. After 6 hours, theimidoyl chloride disappeared completely, and themixture was worked up to give the iminophospho-nate in 95% yield. Signals in the 31P NMR spectra atd 4 7.37 (q, J 4 9.2 Hz) definitely revealed the for-mation of the Arbuzov reaction product. The sub-sequent reduction was achieved by treatment withNaBH3CN as indicated by the 31P NMR resonance atd 4 15.2 (q, J 4 8.7 Hz), much downfield from thatof the iminophosphonate, because the phosphonatemoiety in the product is linked to an sp3-hybridizedcarbon, while in the iminophosphonate, it is linkedto an sp2 carbon.

We also extended the above reaction to variousN-substituted trifluoroacetimidoyl chlorides andother trivalent phosphorus reagents. N-substituentsin the imidoyl chlorides caused a remarkable effecton the reaction. When R1 was alkyl, prolongation ofreaction time was found to be necessary. When R1

was aryl, the more powerful the electron-withdraw-ing group on the benzene ring, the more rapidlythe reaction occurred. When N-(p-nitrophen-


yl)trifluoroacetimidoyl chloride was used, the reac-tion was performed under 08C because of its highreactivity. Diethyl phenylphosphonite reacted morerapidly than a typical trialkyl phosphite. These re-sults indicate that the Arbuzov-type reaction is ini-tiated by the nucleophilic attack of the trivalentphosphorus species on the imidoyl chloride. Highlyelectrophilic imidoyl chlorides and more nucleo-philic phosphorus reagents promote the reactionmarkedly. Dialkyl phosphites also reacted with theimidoyl chlorides. In such cases, addition of Et3Nwas necessary in order to remove HCl generated.Monoethyl phenylphosphonite reacted similarly.

On the other hand, as a reactive intermediate,trifluoroacetimidoyl chloride underwent a substitu-tion reaction with the carbanion obtained by depro-tonation of an appropriate alkyl phosphonate. Sub-sequent reduction gave the 2-trifluoromethyl-2-(substituted amino)ethylphosphonate, a tri-fluoromethyl-2-AEP [41] (Scheme 11).

Depending on the nature of R2, an appropriatebase should be used. Replacement of the chlorine ofthe imidoyl chloride by a carbanion provides, as arule, a mixture of the iminophosphonate and its iso-meric enaminophosphonate, which are very difficultto separate by column chromatography but are eas-ily identified by 1H NMR spectroscopy. The ratio ofthe imine to the enamine is dependent on the struc-ture of the substituent. The mixture of intermediateswas subjected to reduction without isolation. The re-duction proceeded smoothly with NaBH3CN in ace-tic acid. Besides its role as a solvent, HOAc enhancedthe electrophilicity of the substrate by the protona-tion of the imino and amino nitrogens.

POLYFUNCTIONALIZED PHOSPHONICACIDS

Phosphonostatine-Containing DifluoromethylMoiety

As an unusual amino acid 4(S)-amino-3(S)-hydroxy-5-methylheptanoic acid (statine) is an essential com-ponent of pepstatin, so it is a naturally occurringpeptide possessing an inhibitory effect on proteolyticenzymes such as renin [42]. Some synthetic peptidesderived from difluorostatin are potent renin inhibi-tors and show promising new therapeutic possibili-ties for the treatment of high blood pressure [43].Replacement of a carboxyl group in a biologicallyactive molecule by a phosphonic moiety often resultsin an interesting change in the biological activity. Re-cently, the difluoromethylphosphonate moiety hasattracted much attention because it offers significantadvantages over its nonfluorinated counterpart as an

enzyme inhibitor due to a small steric disparity buta great polarity difference between the difluorometh-ylene and methylene group [44]. This situationaroused our interest in the synthesis of a phosphorusanalogue of statine bearing a difluoromethylenemoiety [45].

Thus, an N-phthalimido-protected L-amino acidwas converted into its corresponding acyl chloridethat was reacted with BrZnCF2P(O)(OEt)2, derivedby the action of zinc powder on diethyl 1-bromo-1,1-difluoromethylphosphonate, providing the 3-(N-phthalimido)-2-oxo-1,1-difluoroalkylphosphonate40. Owing to the chiral induction of C-3, diastereo-selective reduction of compound 40 with NaBH3CNin acidic medium afforded a mixture of 2R,3S- (41)and 2S,3S-(42) diastereoisomers. The diastereo-meric excess value was determined by HPLC. Depro-tection at both the N and P terminals of these dias-tereoisomers in the usual manner furnished thea,a-difluorophosphonic analogues of statine(Scheme 12).

Peptides Containing the 1-Hydroxy-2-aminophosphonic Acid Residue

1-Hydroxy-2-aminoethylphosphonic acid (HAEP) isanother naturally found C–P compound; the 2-alkyl-substituted HAEP and its peptide derivatives werereported to be an inhibitor of renin [46]. Therefore,we devoted our efforts to the synthesis of a new typeof phosphonopeptide containing a 1-alkyl-substi-tuted HAEP unit [23]. Since the mixed carboxylic-carbonic acid anhydride method (MCCA), the mostpopular procedure for peptide bond formation, isunsuitable for the synthesis of phosphonopeptidesbearing a free hydroxyl function, we tried to use anactive ester in situ, one formed by the reaction of N-hydroxyphthalimide (HONPht) and the N-protectedamino acid, in the preparation of phosphonopep-tides. In this reaction, apart from its role in accel-erating the coupling process via an active ester,HONPht also acted as an acid to prevent the alkali-labile 44 from decomposing. The free peptides 46were obtained after deprotection of the phosphono-peptides 45 in the usual manner [47] (Scheme 13).

PHOSPHORYLATED HETEROCYCLES

Phosphorylated heterocycles can be regarded as aspecial class of polyfunctionalized phosphonic acids.The successful clinical application of Fosfomycineand Cyclophosphonamide as wide spectrum antibi-otics and anti-cancer drugs, respectively, arousedour interest in structure-activity studies of hetero-

112 Yuan et al.

SCHEME 11

cyclic compounds bearing the phosphonate moietyfor evaluation of their biological activities. At thesame time, a synthetic study of phosphorylated het-erocyclic compounds deserves considerable atten-tion because of their extensive applications in or-ganic synthesis.

Phosphorylated Heterocycles fromIsocyanomethylphosphonates

Although alkyl or aryl phosphonates can be preparedfrom alkyl or aryl halides and trivalent phosphorusreagents, unfortunately, these methods cannot beused for the synthesis of heterocyclic compounds inwhich the phosphoryl group is directly linked to thering. However, the building-block strategy can be ex-ploited to solve this problem. Isocyanomethylphos-phonates serve as a promising building block owingto the versatile reactivity of the isocyano group.Schollkopf and co-workers [48,49] synthesized phos-phorylated oxazoline, oxazole, and thiazole by cy-cloaddition of an isocyanomethylphosphonate to analdehyde, acyl chloride or carbon disulfide, namely,the C4O or C4S bond, respectively. We first re-ported cycloaddition of an isocyanomethylphos-phonate to electron-deficient alkenes or imidoylchlorides, namely, to the C4C or C4N bond, to pre-pare phosphorylated pyrroles and imidazoles.

Phosphorylated Pyrrole and Dihydropyrrole.Reactions of nitroalkenes with diethyl isocyanome-thylphosphonate using t-BuOK, n-BuLi, or LDA asbase provided 3,4-disubstituted pyrrole-2-phosphon-ates in moderate yields [50]. Reaction of b-nitrostyr-ene with diethyl isocyanomethylphosphonate gave acomplicated mixture that demonstrated that the ex-istence of a substituent geminal to the nitro groupwas essential in this reaction. Aliphatic nitroalkenesfurnished the products in lower yields than aromatic

nitroalkenes owing to the base-catalyzed polymeri-zation of the alkenes (Scheme 14). A tentative reac-tion mechanism was postulated by us. The additionof a carbanion derived from diethyl isocyanome-thylphosphonate to a nitroalkene generated a nitron-ate anion that cyclized by intramolecular addition tothe isocyano group. Elimination of the nitro groupon the tertiary carbon, followed by subsequent allylicproton transfer, afforded the 3,4-disubstituted pyr-role-2-phosphonates. This is an interesting reactionin which the nitro function acts as both an activatingand a leaving group (Scheme 15).

Cuprous oxide catalyzed reactions of diethyl is-ocyanomethylphosphonate with methyl methacry-late or methacrylonitrile produced 3,4-dihydro-2H-pyrrole-2-phosphonates. In this case, Cu2O activatesthe methylene group or the isocyanomethylphos-phonate to form a species with a C–Cu bond, which,being similar to a carbanion, attacks the Michael ac-ceptor and then cyclizes to give the product [51]. Inspite of its low stereoselectivity, this reaction is agood example of atom economy (Scheme 16).

Phosphorylated Imidazoles. We also investi-gated the reaction of diethyl isocyanomethylphos-phonate with the C4N bond. Unsuccessful trials onreaction with imines or O-trimethylsilyl oxime im-pelled us to study the reaction with imidoyl chlo-rides. Keeping the role of the nitro group in theaforementioned formation of the pyrrole-2-phos-phonate in mind, we supposed that the presence ofa leaving group at the imino carbon should facilitatethe reaction with the C4N bond. However, imidoylchlorides are not easy to handle due to their highsensitivity to moisture. We chose trifluoroacetimi-doyl chlorides as starting materials because of theirstability resulting from the presence of the strongelectron-withdrawing trifluoromethyl group.

As shown in Scheme 17, the carbanion derived


SCHEME 12

SCHEME 13

from diethyl isocyanomethylphosphonate at 1708Cdisplaced smoothly the chlorine of N-substituted tri-fluoroacetimidoyl chlorides to form an imine inter-mediate, which, upon rearrangement, followed bycyclization, gave 1-substituted 5-trifluoromethylim-idazole-4-phosphonates in moderate to good yields.Signals at d 4 7.60–7.79 (N4CHN) in 1H NMR spec-tra definitely revealed the formation of the imidaz-ole. When R was an alkyl group, the yields of imid-

azoles were somewhat lower than when R was anaryl group, possibly owing to the base-induced isom-erization of imidoyl chlorides in the former case[52]. Usually, the addition of amines to an isocyanocarbon requires the participation of a catalyst. Thedriving force of such cyclization seems to be the ten-dency toward aromatization. Attempts to isolate theimine or the enamine intermediate were notsuccessful.

114 Yuan et al.

SCHEME 14

SCHEME 15

Replacement of BuLi by NaH in this reactionfailed to give imidazole compounds. Quenching ofthe reaction produced a complex that was difficultto purify by column chromatography.

The regiochemistry of this reaction deservesmore attention. The single chemical shift in the 31PNMR and 19F NMR spectra demonstrated the solestructure of the products. The regioisomers, 1-sub-stituted 5-trifluoromethylimidazole-2-phospho-nates, which conceivably could have resulted from a1,3-dipolar cycloaddition, were not detected. The 13CNMR spectra confirmed the regiochemistry. Thedoublets assigned to C-2 and C-4 and the dq signalsof C-5 indicated that the phosphoryl group waslinked to C-4.

As shown in Scheme 18, another pathway inwhich cyclization precedes the elimination of chlo-rine may be postulated [53]. This addition-cycliza-tion-prototropic-elimination process, however, wasprecluded since the reaction of diethyl 1-isocyano-ethylphosphonate with N-phenyltrifluoroacetimi-doyl chloride gave a 1-isocyano-2-iminophosphon-ate rather than the cyclized imidazoline derivative.

This result means that the elimination of chlorideion from the N-anion is more rapid than the cycli-zation (Scheme 19).

It is noted that this reaction provides 1-arylimi-dazoles [54] that are otherwise not easily attainabledue to the fact that nucleophilic alkylation at N-1 byan aryl halide is not possible. Moreover, fluorinatedheterocycles are the subject of active investigation,and some fluorinated imidazoles have been found tofunction as xanthine oxidase inhibitors or have beenused as drugs.

Nucleophilicity of Isocyanomethylphosphonate.As demonstrated by Huckel, MNDO, and ab initioMO calculations, the predominant contribution tothe valence-bond description of alkyl isocyanides ismade by the polar structure RN`4C1, an iminocar-bene resonance hybrid [55]. The isocyano carbonatom thus carries a nonbonding (n) electron pairwith a formal negative charge. A nucleophilic char-acter is therefore not unexpected. An evidence forsuch a property is the [1 ` 4] cycloaddition reactionof alkyl isocyanides with conjugated nitroalkenes


SCHEME 16

SCHEME 17

[56]. However, we found that diethyl isocyanome-thylphosphonate does not react similarly with ni-troalkenes, which is attributed to the reduced nu-cleophilicity of its isocyano carbon atom caused bythe electron-withdrawing phosphoryl group. It canbe envisaged that the weak nucleophilicity of diethylisocyanomethylphosphonate can be demonstratedby its reaction with a strong electrophile. Thus, we

investigated the a-addition reaction of trifluoroac-etic anhydride and acyl chlorides to diethyl isocy-anomethylphosphonate [57].

Diethyl isocyanomethylphosphonate was first al-lowed to react with the highly electrophilic trifluo-roacetic anhydride. As shown in Scheme 20, this ex-othermic reaction occurred at room temperature indichloromethane, giving (N-diethoxyphosphoryl-

116 Yuan et al.

SCHEME 18

SCHEME 19

SCHEME 20

methyl)trifluoropyruvamide hydrate after hydrolysisof the a-adduct. The resonances at dH 3.24 (br, 3H,NH, 2OH), dF 16.3 (s), and combustion microanal-ysis confirmed the structure.

We next examined the reaction of acyl chlorideswith an isocyanomethylphosphonate ester. As shownin Scheme 21, the acyl chlorides, derived from the

corresponding primary, secondary, and tertiary al-kanoic acids, reacted with the isocyanide functionalgroup in refluxing dichloromethane over 20 hours(via TLC monitoring). The a-adduct with the struc-ture of an a-ketoimidoyl chloride can be easily dis-tinguished from the starting materials on a TLCplate by its ultraviolet absorption due to the conju-


SCHEME 21

SCHEME 22

gation between the carbonyl and the imine group.The observed difference between trifluoroacetic an-hydride and acyl chlorides when they reacted withdiethyl isocyanomethylphosphonate, was attributedto the weaker electrophilicity of acyl chlorides asagainst that of trifluoroacetic anhydride.

1,3-Dipolar cycloaddition of an a-phosphoryl ni-trile oxide to olefins has been reported [58]; however,reactions of phosphorylated nitrile ylides have, tothe best of our knowledge, not yet been studied. Wedemonstrated that the a-ketoimidoyl chlorides, gen-

erated by the a-addition reaction of acyl chlorides tothe isocyanomethylphosphonate ester, may serve asprecursors to phosphorylated nitrile ylides.

Thus, the a-ketoimidoyl chlorides formed in theabove manner were treated with triethylamine with-out isolation, providing a 1,3-dipolar species upon1,3-dehydrochlorination. This 1,3-dipolar interme-diate was then trapped in situ with methyl acrylateto give the cycloadducts as a mixture of 3- and 4-methoxycarbonylpyrrolines, showing a lack of regio-selectivity (Scheme 21).

118 Yuan et al.

SCHEME 23

SCHEME 24

SCHEME 25

SCHEME 26


SCHEME 27

SCHEME 28

When nitroalkenes were used as dipolarophilesto capture the above 1,3-dipolar intermediates, 5-acylpyrrole-2-phosphonates were obtained in mod-erate yields, together with some unidentified com-ponents. As illustrated in Scheme 22, the pyrroles 59are produced via a regioselective 1,3-dipolar cy-cloaddition followed by elimination of nitrous acidand subsequent aromatization (Scheme 22).

Phosphorylated 4,5-Dihydroisoxazoles

In our continuing study of the chemistry of nitroal-kenes, we observed that reaction of tetraethyl meth-ylenebisphosphonate with b-methyl-b-nitrostyrenein the presence of n-BuLi in refluxing THF afforded

a 7% yield of an ethenylidenebisphosphonate in ad-dition to the normal Michael addition product. Wereasoned that the equilibrium shown in Scheme 23may be involved in this Michael addition. This sug-gestion was verified by the following: (1) Addition ofp-nitrobenzaldehyde to the reaction mixture pro-vided the Horner-Emmons reaction product solely.(2) Addition of trimethylchlorosilane (TMSiCl), fol-lowed by immediate quenching with acid, furnishedthe ethenylidenebisphosphonate quantitatively, be-cause TMSiCl reacted with lithium nitromethane toform the silyl nitronate, shifting the equilibrium tothe right side completely. (3) Addition of TMSiCl fol-lowed by allowing the mixture to stand for two daysand subsequent quenching with acid produced the

120 Yuan et al.

SCHEME 29

SCHEME 30

2-isoxazoline-5,5-dilylbisphosphonate, resultingfrom 1,3-dipolar addition of the silyl nitronate to theethenylidenebisphosphonate. Extension of this re-action to other nitroalkenes led to the developmentof a useful route to 3,4-disubstituted 4,5-dihydro-isoxazole-5,5-diylbisphosphonates [59] (Scheme 23).

Other compounds with an active methylenegroup reacted analogously with nitroalkenes andthen TMSiCl, giving 4,5-dihydroisoxazoles with twoelectron-withdrawing substituents at C-5 of the ring[60]. Generation of alkenes and nitronates in thesereactions was thought to result from a 1,7-silicon mi-gration, which is, however, quite different behaviorfrom that of bisphosphonates (Scheme 24).

The Michael addition of carbanions derivedfrom methylphosphonates to nitroethylene, followedby reaction with TMSiCl, gave the corresponding si-lyl nitronates that reacted with electron-deficient al-

kenes to give 4,5-dihydroisoxazoles in a one-pot pro-cedure [61]. Electron-rich alkenes such as styrene or1-hexene did not undergo this reaction due to thelow reactivity of the silyl nitronate. Other carbanionsstabilized by an electron-withdrawing group be-haved analogously (Scheme 25). Similar reactionsusing lithium diethyl phosphites as the nucleophileoffered 3-diethoxyphoshorylmethyl-4,5-dihydroisox-azole derivatives [62], which have been reported byothers to be useful synthetic intermediates (Scheme26).

Intramolecular silyl nitronate-olefin cycloaddi-tions (ISOC) and intramolecular nitrile oxide-olefincycloadditions (INOC) were also investigated by us.The Michael addition of carbanions derived fromalk-3-en-1-ylphosphonates to 2-aryl-1-nitroethyle-nes, followed by reactions with TMSiCl, gave the cor-responding intermediates of silyl nitronates, which


underwent the ISOC reaction. This led to the stereo-selective synthesis of 6-aryl-3,3a,4,5,6-pentahydro-cyclopent[C]isoxazole-5-ylphosphonates [63]. Thespatial structures of the products were unambigu-ously characterized by 1H, 13C, and 31P NMR,NOESY, and 1H–13C COSY spectra. The hydrogen atC-3a is trans to both the phosphonate group at C-5and the aryl group at C-6. The Michael additionproducts were isolated in the absence of TMSiCl andwere then allowed to undergo the INOC reaction bytreatment with phenyl isocyanate and triethylamineto provide the products where the hydrogen at C-3ais cis to both the phosphonate group at C-5 and thearyl group at C-6 [64]. The Michael addition pro-ceeded via the transition state 66 of least steric hin-drance in both reactions. The opposite stereoselec-tivity of ISOC and INOC reactions may berationalized by the difference of their transitionstates. In the transition state 67 of an ISOC reaction,the aryl and trimethylsilyl groups have to be transdue to the steric interaction, which renders the hy-drogen at C-3a trans to the aryl group at C-6; on theother hand, the transition state 68 of INOC leads tothe cis relationship between the hydrogen at C-3aand the aryl group at C-6 (Scheme 27).

Phosphorylisoxazoline Derivatives

1,3-Dipolar addition of nitrones was established byHuisgen [65] as one of the important routes to het-erocycles. By the reaction of 1-(hydroxylam-ino)alkylphosphonates with aldehydes we obtainedphosphoryl nitrones with exclusively the Z-configu-ration that gave (Z)N-phosphorylisoxazolines via 4` 2 dipolar cycloadditions with maleic anhydride(Scheme 28).

The configuration of the product was demon-strated by 2D NOESY NMR spectroscopy and ra-tionalized by FMO theory [66].

Phosphorylated Indoles

Reactions of dialkyl phosphites with 2-nitrostyrenesin the presence of triethylamine, trimethylsilyl chlo-ride and hexamethyldisilazane provides a one-potprocedure for the synthesis of N-hydroxy-indole-3-phosphonates (70) and indole-2-one-3-phosphona-tes (71) as well as the normal addition products, 1-aryl-2-nitroalkylphosphonates [67] (Scheme 29). Atentative reaction mechanism was postulated by us(Scheme 30).

An active nucleophile, a tri-coordinated silylatedphosphite, is formed by reaction of a dialkyl phos-phite with TMSCl/Et3N. This species attacks the b-carbon of a 2-nitrostyrene to give a dipolar inter-

mediate (72), which undergoes fast 1,6-silylmigration to form the intermediate 73. a-Proton ab-straction by Et3N from the latter gives 74. Nucleo-philic attack of an ortho-carbanionic center of 74 onthe nitrogen atom with loss of TMSO, followed by a1,3-shift of hydrogen from carbon to oxygen, yieldsan N-hydroxy-indole-3-phosphonate.

ACKNOWLEDGMENT

This project was supported by the National NaturalScience Foundation of China.

REFERENCES

[1] C. Z. Li, C. Y. Yuan, Tetrahedron Lett., 34, 1993, 1515.[2] F. H. Ebetino, S. M. Kaas, R. J. Crawford, Phospho-

rus, Sulfur and Silicon, 76, 1993, 151.[3] F. H. Ebetino, A. V. Bayless, J. Amburgey, K. J. Ibbot-

son, S. Danserean, A. Ebrahimpour, Phosphorus, Sul-fur and Silicon, 109–110, 1996, 24.

[4] P. Kafarski, B. Lejczak, Phosphorus, Sulfur and Sili-con, 63, 1991, 193.

[5] C. Y. Yuan, Y. M. Qi, Acta Chim. Sinica, 1986, 93.[6] C. Y. Yuan, Y. M. Qi, Synthesis, 1986, 821.[7] C. Y. Yuan, Y. M. Qi, Synthesis, 1988, 472.[8] C. Y. Yuan, G. H. Wang, Synthesis, 1990, 256.[9] C. Y. Yuan, G. H. Wang, S. J. Chen, Synthesis, 1990,

522.[10] C. Y. Yuan, S. J. Chen, G. H. Wang, Synthesis, 1991,

490.[11] C. Y. Yuan, G. H. Wang, Phosphorus, Sulfur and Sili-

con, 65, 1992, 207.[12] C. Y. Yuan, S. J. Chen, Synthesis, 1992, 1124.[13] V. P. Kukhar, V. A. Soloshonok, V. A. Solodenko,

Phosphorus, Sulfur and Silicon, 92, 1994, 239.[14] A. B. Smith III, K. M. Yager, C. M. Taylor, J. Am.

Chem. Soc., 117, 1995, 10879.[15] R. Hamilton, B. Walker, B. J. Walker, Tetrahedron

Lett., 36, 1995, 4451.[16] A. Bongini, R. Camerini, S. Hofman, M. Panunzio,

Tetrahedron Lett., 35, 1994, 8045.[17] W. F. Gilmore, A. McBride Hilmer, J. Am. Chem. Soc.,

94, 1972, 4361.[18] C. Y. Yuan, S. H. Cui, Phosphorus, Sulfur and Silicon,

55, 1991, 159.[19] S. S. Li, G. Q. Wang, Y. H. Li, C. Y. Yuan, Acta Chim.

Sinica, 51, 1993, 713.[20] C. Y. Yuan, S. S. Li, G. Q. Wang, Y. L. Ma, Phospho-

rus, Sulfur and Silicon, 81, 1993, 27.[21] S. H. Sasai, S. Arai, Y. Tahara, M. Shibasaki, J. Org.

Chem., 66, 1995, 6656.[22] J. Rachon, C. Wasielewski, Rocz. Chem., 50, 1976,

477.[23] J. Rachon, C. Wasielewski, Rocz. Chem., 49, 1975,

397.[24] P. J. Diel, L. Maier, Phosphorus, Sulfur and Silicon,

36, 1988, 85.[25] K. Afasinkkia, C. W. Rees, J. I. G. Cadogen, Tetrahe-

dron, 46, 1990, 7175.[26] C. Y. Yuan, S. J. Chen, R. Y. Xie, H. Z. Feng, L. Maier,

Phosphorus, Sulfur and Silicon, 106, 1995, 115.[27] K. Hemmi, H. Takeno, M. Hashimoto, T. Kamiya,

Chem. Pharm. Bull., 30, 1982, 111.

122 Yuan et al.

[28] K. A. Petrov, L. V. Treshchalina, V. M. Chizov, Zh.Obshch. Khim., 49, 1979, 590.

[29] M. P. Osipova, P. M. Lukin, V. A. Kukhtin, Zh.Obshch. Khim., 52, 1982, 449.

[30] R. Huber, A. Knierzinger, J. P. Obrecht, A. Vasella,Helv. Chim. Acta, 68, 1985, 1730.

[31] M. Elhaddadi, R. Jacquier, F. Petrus, C. Petrus, Phos-phorus, Sulfur and Silicon, 45, 1989, 161.

[32] C. Y. Yuan, S. J. Chen, H. Zhou, L. Maier, Synthesis,1993, 955.

[33] E. D. Korn, D. G. Dearbon, E. A. Sokdoski, J. Biol.Chem., 248, 1973, 2257.

[34] T. Tone, Y. Okamoto, H. Sakuvai, Chem. Lett., 1978,1349.

[35] J. Zyymunt, Pol. J. Chem., 52, 1978, 227; 54, 1980,223.

[36] T. A. Mastryukova, G. M. Baranov, V. V. Perekalin,M. I. Kabachinick, Dokl. Akad. Nauk SSSR, 171,1966, 1341.

[37] C. Y. Yuan, S. H. Cui, G. H. Wang, H. Z. Feng, Syn-thesis, 1992, 258.

[38] C. Y. Yuan, G. H. Wang, S. J. Chen, Phosphorus, Sul-fur and Silicon, 63, 1991, 111.

[39] C. Y. Yuan, G. H. Wang, H. Z. Feng, J. P. Chen, L.Maier, Phosphorus, Sulfur and Silicon, 81, 1993, 149.

[40] W. S. Huang, Y. X. Zhang, C. Y. Yuan, Phosphorus,Sulfur and Silicon, 107, 1995, 21.

[41] W. S. Huang, C. Y. Yuan, Phosphorus, Sulfur and Sil-icon, 106, 1995, 163.

[42] R. J. Workman, D. S. Burkitt, Arch. Biochem. Bio-phys., 194, 1977, 157.

[43] S. Thaisrivongs, D. T. Pals, W. M. Kati, S. R. Turner,L. M. Thomasco, J. Med. Chem., 28, 1985, 1153.

[44] S. Halazy, A. Ehrhard, C. Danzin, J. Am. Chem. Soc.,113, 1991, 315.

[45] S. J. Chen, C. Y. Yuan, Phosphorus, Sulfur and Sili-con, 82, 1993, 73.

[46] D. V. Patel, K. Rielly-Gauvin, D. E. Ryono, Tetrahe-dron Lett., 31, 1990, 5587.

[47] S. J. Chen, C. Y. Yuan, Synthesis, 1993, 1074.[48] U. Schollkopf, T. Wintel, Synthesis, 1984, 1033.[49] U. Schollkopf, J. Rachon, Liebigs Ann. Chem., 1981,

1186; 1981, 1693.[50] C. Y. Yuan, W. S. Huang, Synthesis, 1993, 473.[51] C. Y. Yuan, W. S. Huang, Chinese Chem. Lett., 5, 1994,

565.[52] K. Tanaka, H. Daikaku, K. Mitsuhashi, Chem. Lett.,

1983, 1463.[53] R. Meyer, U. Schollkopf, P. Bohme, Liebigs Ann.

Chem., 1977, 1183.[54] W. S. Huang, C. Y. Yuan, Synthesis, 1996, 511.[55] D. J. Carini, A. T. Chiu, P. C. Wong, A. L. Johnson,

R. R. Timmermans, P. M. B. W. Timmermans, Bioorg.Med. Chem. Lett., 3, 1993, 895.

[56] J. B. Moffat, Chem. Phys. Lett., 55, 1978, 125.[57] W. S. Huang, Y. X. Zhang, C. Y. Yuan, J. Chem. Soc.,

Perkin Trans. I, in press.[58] R. Y. Zhang, J. Chen, Synthesis, 1990, 817.[59] C. Y. Yuan, C. Z. Li, Phosphorus, Sulfur and Silicon,

69, 1992, 75.[60] C. Y. Yuan, C. Z. Li, Phosphorus, Sulfur and Silicon,

78, 1993, 47.[61] C. Z. Li, C. Y. Yuan, Synthesis, 1993, 471.[62] C. Z. Li, D. B. Zhou, C. Y. Yuan, Chinese Chem. Lett.,

4, 1993, 391.[63] C. Y. Yuan, C. Z. Li, Phosphorus, Sulfur and Silicon,

103, 1995, 133.[64] C. Y. Yuan, C. Z. Li, Tetrahedron Lett., 34, 1993, 5959.[65] R. Huisgen, Angew. Chem. Int. Eng. Ed., 7, 1968, 321;

J. Org. Chem., 33, 1968, 2291.[66] S. J. Chen, Ph.D. Dissertation, SIOC 1993.[67] G. H. Wang, C. Y. Yuan, Heteroatom Chem., 3, 1992,

521.

new strategy for the synthesis of functionalized phosphonic acids

Documents