Additions to N‐Sulfinylamines as an Approach for the Metal‐free Synthesis of Sulfonimidamides: O‐Benzotriazolyl Sulfonimidates as Activated Intermediates

Abstract Sulfonimidamides are obtained in moderate to very good yields from the key intermediates O‐benzotriazolyl sulfonimidates, which are formed by reacting aryldiazonium tetrafluoroborates, N‐tritylsulfinylamine, and N‐hydroxybenzotriazole hydrate in a process mediated by a tertiary amine. The formation of the sulfonimidate proceeds in inexpensive and environmentally benign dimethyl carbonate as the solvent, it does not require anhydrous conditions, and the product yields generally exceed 70 %. The substrate scope is broad, and a wide range of sensitive organic functionalities is well tolerated. The reactions probably proceed via aryl radicals formed from diazonium cations with assistance from both the tertiary amine and the sulfinylamine.


Introduction
Sulfonamides are among the most abundant and important structural motifs in medicinal and agricultural chemistry. They have arich history in medical circles,with sulfonamides being one of the first commercial antibiotics ever available, for example. [1] Furthermore,s ulfonamide derivatives have exhibited high biological activities,i ncluding antiproliferal, [2] diuretic, [3] antihypertensive, [4] hypoglycemic, [5] antiinflammatory, [6,7] antiviral, [8] and herbicidal properties. [9] Ar ecent report states the prevalence of the sulfamoyl group in sulfur-containing drugs to be as high as 29 %. [10] Among the various sulfonamide derivatives,t heir monoaza analogues, known as sulfonimidamides,h ave recently attracted increasing attention, in particular,b ecause they are regarded as bioisosters of the parent compounds. [11,12] However,t od ate, only relatively few methods for the preparation of sulfonimidamides have been reported, with most of them still having severe synthetic disadvantages that prevent easy access to aw ide variety of such attractive compounds. [13] Most sulfonimidamide syntheses proceed via sulfoximinoyl chlorides as reactive intermediates (Scheme 1), which are commonly not isolated because of their sensitivity to hydrolysis and redox side reactions. [14] In these methods,k ey steps are oxidative chlorinations [14c,15, 16] and/or iminations [14c,17] of lower-valent sulfur compounds,a nd subsequent nucleophilic substitutions of the resulting sulfoximinoyl chlorides.H owever,t he employment of strong oxidation or halogenation agents is often incompatible with sensitive functionalities and poses arisk in itself.Inpart, this is also true for the redox-neutral approach introduced by Chen and Gibson, who used triphenylphosphine dichloride for the deoxychlorination of sulfonamides. [16] Functional-group sensitivity is also an issue when highly basic and nucleophilic organometallic reagents are applied, as demonstrated by Willis and co-workers, [18] because they not only require strictly anhydrous and/or otherwise inert conditions,b ut also do not tolerate important functional groups such as the carbonyl groups of esters,ketones and aldehydes. Grygorenko and co-workers fine-tuned the stability of the reactive intermediates by converting the sulfoximinoyl chlorides to the corresponding N-methylimidazolium triflates. [19] Thes ame can be achieved by using their relatively stable Scheme 1. Overview of approaches towards sulfonimidamides (Trt = trityl).
fluorine-containing counterparts.W hile the preparation of the latter from sulfoximinoyl chlorides by halide exchange has long been known, [14d, 20] they can now directly be accessed by Sharpless SuFEx approach. [21] Unfortunately,t he practicability of this process is limited by the use of thionyl tetrafluoride (SOF 4 ), which is at oxic and corrosive gas. Stockmann, Lücking, and co-workers were able to overcome many of the aforementioned challenges by directly converting sulfinamides into sulfonimidamides through oxidative imination with in situ generated iminoiodinane(III) species. [22] Recently,t his chemistry was extended by Luis,B ull, and coworkers,w ho applied sulfenamides as starting materials. [23] Tw ov ery different approaches circumventing the formation of hydrolytically sensitive intermediates were introduced by us,b oth through the use of copper catalysis.T he first one involved oxidative dealkylations of NH-S-aryl-S-alkyl sulfoximines in the presence of secondary amines,w hich led to tertiary NH-sulfonimidamides,p resumably via S-centered sulfoximinoyl radicals. [24] In the second, fully substituted sulfonimidamides were obtained by reacting S-arylsulfinamides with O-benzoyl hydroxylamine derivatives. [25] Due to the limits of the current methods,w ew ere interested in devising an ew approach towards sulfonimidamides that is flexible and shows high functional-group tolerance.

Results and Discussion
In 2017, Willis and co-workers reported groundbreaking findings on aone-pot-four-step reaction sequence (Scheme 2, top) [18,26] that enables the preparation of sulfonimidamides through the addition of (hetero)aryl-or alkylmagnesium halides to the hydrolytically stable and easily accessible reagent N-tritylsulfinylamine (1), followed by oxidation of the resulting sulfinimidate salts with tert-butylhypochlorite to give the corresponding sulfoximinoyl chlorides.S ubsequent treatment of the latter products with primary and secondary amines yielded N-tritylsulfonimidamides,w hich were deprotected with strong acids to give S-aryl and S-alkyl-NHsulfonimidamides in high yields.T he redox cascade,inwhich the oxidation number of sulfur changed from + IV to + II back to + IV,i sp articularly conspicuous here.W ew ere also interested by ar eport by Wu and co-workers (Scheme 2, middle), [27] who described the formation of sulfonamides via O-aminosulfonates,w hich were obtained from diazonium tetrafluoroborates,the solid sulfur dioxide surrogate DABSO (DABCO·2SO 2 ), and N,N-disubstituted hydroxylamines,i ncluding 1-hydroxybenzotriazole (HOBt). Theproposed reaction mechanism included DABCO-assisted extrusion of molecular nitrogen from the diazonium salt to yield an aryl radical and the DABCO radical cation. Thef ormer attacks the SO 2 to form as ulfur-centered sulfonyl radical, while the latter abstracts ah ydrogen atom from the hydroxylamine to form the DABCO cation and an oxygen-centered hydroxylamine radical. Finally,c ombination of the two radicals gives the O-aminosulfonates in ar emarkably exergonic process. When HOBt was employed as the hydroxylamine,the method enabled the one-pot formation of sulfonamides by substitution with primary or secondary amines present in the reaction mixture.Asimilar mechanism was reported for the formation of sulfonamides from aryl diazonium salts,D ABSO,s odium azide,and triphenylphosphine. [28] Considering 1 as aprotected monoaza analogue of SO 2 ,wewondered about its use for the synthesis of O-aminosulfonimidates and sulfonimidamides (Scheme 2, bottom). Ther ealization of this strategy is reported here.
With the intention to use ar epresentative starting material that could be easily detected and followed by NMR spectroscopy,t he study was initiated by applying 4fluorophenyldiazonium tetrafluoroborate (2a)a st he aryl source.T oour delight, 2a reacted with equimolar amounts of N-tritylsulfinylamine (1)and HOBt hydrate in acetonitrile as hypothesized, affording the targeted O-Bt-N-Trt-sulfonimidate 3a in 42 %y ield. Ther eaction occurred within seconds, as revealed by the rapid termination of nitrogen evolution. Va rying the amount of DABCO (Table 1; with a5:1 mixture of MeCN and THF as solvent for 20 min at ambient temperature;analysis by 19 FqNMR spectroscopy) showed that it acts as mediator and that at least one equivalent of this base had to be added to achieve high conversion of 2a (> 95 %), providing 3a in 60 %y ield (Table 1, entry 5). With less DABCO (Table 1, entries 1a nd 3-4) or with one equiv of CaCO 3 and acatalytic amount of DABCO (Table 1, entry 6), the product formation was not more efficient. [29] Performing the reaction at 60 8 8Cyielded N-tritylsulfonamide 3aa as the main product (Table 1, entry 6). Other detectable side products were detritylated sulfonamide 3ab, N-trityl-sulfinamide 3ac, and O-Bt-sulfonate 3ad (for details,s ee the Supporting Information).
Investigation of solvent systems revealed that both acetone and dimethyl carbonate (DMC) are significantly better than MeCN or the MeCN/THF mixture used before and that both are equally suitable if applied as singlecomponent solvent ( Table 2, entries 7-10). A1 :1 mixture of the two was not more effective ( Table 2, entry 11). On the basis of these results and in light of its "green" solvent properties (non-toxic,water-immiscible,and environmentally benign), [30] DMC was selected as medium for the subsequent studies.
As expected for ar adical reaction, the presence of dioxygen diminished the yield, whereas al arge excess of water did not. Furthermore,t he reaction proved sensitive to scale,c oncentration of the reactants,a nd addition sequence. Thus,a10-fold increase in scale (from the initially used 0.05 mmol to 0.5 mmol) reduced the yield of 3a from more than 95 %t o6 3% (as determined by 19 Fq NMR spectroscopy). Ac loser inspection revealed that the order in which the reactants are added is important. Hence,i fasolution of HOBt hydrate and N-methylpiperidine in DMC was added dropwise to as uspension of 1 and 2a (0.1 mmol L À1 )i nt he same solvent, as pectroscopic yield of 3a of 77 %w as observed. Inverting the addition mode gave 3a in only 56 % yield. Thei mportance of ac arefully adjusted reactant concentration became even more apparent when the diazonium salt concentration was further changed from the commonly used 0.1 mol L À1 to either 0.5 mol L À1 or 0.02 mol L À1 ,i nw hich case 3a was isolated in 46 %a nd 69 %y ield, respectively.D egassing the solvent led to 3a in 78 %y ield (at ac oncentration of 0.02 mol L À1 ). Drying the solvent prior to use had no effect. Those final conditions proved robust for scale-up,leading to 73 %yield of 3a (after column chromatography) on a1 0mmol scale.
With the optimized conditions in hand, the substrate scope was investigated (Scheme 3). Pleasingly,t he reaction conditions were applicable to awide range of diazonium salts. Phenylsulfonimidate 3b was isolated in 89 %y ield. Methyl substituents at the para or meta position of the aryl group reduced the yield of the corresponding sulfonimidate (3c and 3d,respectively) slightly.More significant was the effect of an ortho-methyl group,w hich led to product 3e in only 61 % yield. In the series of compounds with para-halo substituents, the yields of 3a, 3f-h dropped with increasing atomic number of the halogen atom. This may be due to the fact that the  halogen-bonding affinity of aryl halides increases with increasing polarizability from fluorine to iodine, [31] and therefore halogen-bonding effects between the nitrogen of one diazonium cation and the halogen substituent of another may lead to undesired coordination and thus to side reactions.This already suggests that electrostatic interactions play ar ole in the product formation, ah ypothesis that was later substantiated in control experiments. As generally expected in radical reactions,e lectronic effects had only am inor impact. Thus,d iazonium salts with both electron-donating and electron-withdrawing substituents reacted well, providing the corresponding sulfonimidates 3i-o in yields ranging from 65 %( for 3o with a para-SF 5 group) to 79 %( for 3j with a para-CN substituent). Surprisingly,acetyl-containing product 3p was only obtained in 29 % yield. Due to the lack of reactivity of diazonium salts 2q (bearing af ree phenolic OH-group) and 2x (containing ap entafluorophenyl moiety), sulfonimidates 3q and 3x remained inaccessible.I nc ontrast, diazonium salts 2r and 2s,w hich bear fused arenes,r eacted well, providing 1naphthyl or benzo[d] [1,3]dioxol-5-yl derivatives 3r and 3s in yields of 72 %a nd 62 %, respectively.F or heteroaromatic sulfonimidates,t he yields strongly diverged, and the individual nature of the heterocycle appeared to play ar ole.T hus, while 3-bromopyridin-5-ylsulfonimidate 3t was obtained in 69 %yield, thiazol-2-yl, 3-phenylpyrazol-1-yl, and quinolin-6yl derivatives 3u-w were only formed in the 30 %yield range. At least in part, the latter results might be due to chemical instabilities of the diazonium salts,a sp articularly observed for 2u,w hich decomposed at temperatures above À10 8 8C during its synthesis.
With the goal of shedding light on the reaction mechanism, various control experiments were performed. First, the reaction between 1 and 2a was performed in the absence of additional base.InMeCN (for the result with a5:1 mixture of MeCN and THF as olvent, see Table 1, entry 1), product 3a was obtained in low yield (15 %) after 1h at ambient temperature,t hus indicating the critical role of the tertiary base.Whenthe reaction was performed in the presence of two equiv of TEMPO as radical scavengers,t he formation of 3a was completely suppressed, thus suggesting the intermediacy of radicals as key components.T rapping of such radical by 4phenylstyrene (instead of HOBt) proved impossible.A ttempts to substitute N-tritylsulfinylamine (1)w ith bis(trimethylsilyl)sulfur diimide (4)i nt he coupling with 2a to target ar epresentative of the virtually unknown arylsulfondiimidates 5 or arylsulfondiimidamides 6 [14d, 32] remained unsuccessful ( Figure 1). No reaction occurred, thus indicating the importance of the oxygen in reagent 1.Neither the reaction of 2a with 1 and HOBt nor the analogous one with 4 instead of 1 were catalyzed by the addition of copper(I) chloride (10 mol %) with the intention of promoting aS andmeyertype coupling reaction via radicals.
Thea forementioned observations led us to propose the mechanistic scenario depicted in Scheme 4. In ah ighly organized (transition) state,b oth the tertiary amine and sulfinylamine 1 coordinate to the nitrogen of the diazonium salt. Upon electron transfer from the tertiary amine to give radical cation B,d initrogen is expelled and aryl radical A is formed. Being close to 1,a ryl radical A adds to the sulfur reagent to give sulfoximiminoyl radical C.Hydrogen abstraction from HOBt by radical cation B generates BtO radical D and the HBF 4 salt of the tertiary amine.I na ne xergonic process,w hich provides the driving force of the process, combination of radicals C and D leads to product 3a.
Theisolated N-trityl-O-Bt sulfonimidates 3 were white to yellowish solids that could be purified by conventional flash column chromatography in air at room temperature.T he decomposition rate on silica is low,and they can be stored at À18 8 8Co ver months without significant signs of degradation, thus rending them highly attractive as intermediate for subsequent synthetic applications.H ere,w edeveloped their use in the preparation of sulfonimidamides 8 (Scheme 5).
Fort he initial optimization, morpholine (7a)a nd 1butylamine (7b)were selected as representative nucleophiles. To our delight, both reacted well with sulfonimidate 3a, provided that an additional base was added and that acetonitrile was applied as the solvent. Form orpholine, sulfonimidamide 8a was obtained in up to 77 %y ield (as determined by 19 Fq NMR spectroscopy), and 8b,s temming from 1-butylamine,w as formed in 86 %y ield. As additional bases, N-methylpiperidine and triethylamine were applied, respectively.Attempts to use other solvents than acetonitrile (DMC,a cetone,D MF,o rp yridine) led to lower conversions and yields.I nt he reaction of 3a with 1-butylamine (7b), triethylamine could be substituted with potassium carbonate  without affecting the yield of 8b.A dding catalytic amounts (10 mol %) of potassium iodide or DMAP,w hich are known catalysts for the aminolysis of methyl esters, [33] to amixture 3a and 7b did not improve the yield 8b.Notably,the method did not require dry conditions,a nd hydrolysis products were not detected.
Scheme 6s ummarizes the results of the substrate-scope evaluation in reactions with 3a.Asanadditional base,1equiv of triethylamine was used in acetonitrile for 24 ha ta mbient temperature.T he nucleophile quantities were 2equiv for primary amines and 1.2 equiv for secondary amines.M ost reactions proceeded well, affording the targeted sulfonimidamide 8 in good yields.Inthe series of primary amines,ethyl-, iso-propyl-, and benzylamines led to the corresponding products 8d-f in yields of 79 %, 75 %, and 71 %, respectively. Applying methylamine (7c)p roved less effective under the standard reaction conditions,g iving 8c in only 42 %y ield. This result could be improved by carrying out the substitution in an 8 m ethanolic solution of methylamine (corresponding to an excess of about 28 equiv), which led to 8c in 60 %y ield. Sulfonimidamides 8g and 8h remained inaccessible,presumably for steric reasons in the case of tert-butyl amine (7g)and due to low nucleophilicity for aniline (7h). Attempts to apply 7h in DMF instead of MeCN,athigher temperatures or after prior deprotonation with sodium hydride,r emained unsuccessful. Considering that benzyl amine (7f)r eacted well to give 8fwhile aniline (7h)w as inactive,2 -aminobenzylamine (7i), which contains both ab enzylic and an anilinic nitrogen, was applied. As hypothesized, the reaction with 3a was chemoselective,providing sulfonimidamides 8i in 37 %yield. Alicyclic amines reacted almost uniformly,p roviding products in yields of about 80 %. As revealed by the results for sulfonimidamides 8j-l stemming from reactions with pyrrolidine (7j), piperidine (7k)and azepane (7l), the ring size had no significant influence.I na ddition to morpholido sulfonimidamide 8a,t hiomorpholido and N-methylpiperazido sulfonimidamides 8m and 8n could be prepared, and again the yields of all products in this series were in the 80 %r ange. Particularly noteworthy is the 80 %yield for the formation of sulfonimidamide 8o,w hich is derived from 3a and 4hydroxypiperidine (7o), since it revealed as trong reactivity preference for the N-o ver the O-nucleophilic site of the bifunctional amine.T his pronounced chemoselectivity is not only of synthetic interest but also explains the observed high resistance of the sufonimidates towards hydrolysis.T he employment of other nucleophiles (thiolate,f luoride,a nd azide) led to product mixtures in which only the corresponding sulfinamide (4-fluoro-N-tritylbenzenesulfinamide) could be identified.
In order to demonstrate the practicability of the new sulfonimidamide synthesis,t wo analogues of pharmacologically relevant sulfonamides were prepared (Scheme 7). First, endo-N-n-butyl-substituted azasaccharine derivative 10 was obtained from 2-carboxymethylphenyldiazonium tetrafluoroborate 9 in at otal yield of 48 %o ver three steps.T his product is of interest because Chen and co-workers have shown that saccharine aza bioisosters such as compounds of type 10 exhibit promising preclinical properties. [34] Thesecond molecule in this series was an analogue of sildenafil, which is ac ommercial PDE5 inhibitor with as ulfonamide core. Diazonium salt 12 was prepared in an overall yield of 72 % over several steps from the commercially available 11.Using the aforementioned optimized conditions for the sulfonimidate formation, followed by treatment with N-methylpiperazine and acidic deprotection of the trityl group, 11 was converted into sulfonimidate 13 in 53 %yield over three steps.

Conclusion
We have introduced new methods for the mild preparation of activated sulfonimidates and sulfonimidamides.F irst, aromatic or heteroaromatic diazonium salts are reacted with the bench-stable reagent N-tritylsulfinylamine (1)a nd 1-hydroxybenzotriazole hydrate in the presence of N-methylpiperidine to give ab road range of N-trityl-O-benzotriazolylsulfonimidates 3.N on-toxic, environmentally benign dimethyl carbonate is the solvent. Ther eaction takes place at room temperature without the requirement for dry conditions.D egassing of the solvent and al ow reactant concentration have ab eneficial effect on the yields,b ut neither is critical for the success of the transformations.M any functional groups are tolerated, and the electronic properties of the substituents have only am inor effect on the product yields.I nasubsequent step,t he sulfonimidates 3 can be converted into the corresponding sulfonimidamides 8 through simple reactions with primary or secondary amines,w hereby only aliphatic amines react. This chemoselectivity is without precedent. Thef ormation of the sulfonimidates probably takes place through ar adical mechanism involving preorganized aggregates of the reactants.Anovel azasaccharine derivate and an unprecedented sulfonimidamide analogue of sildenafil were prepared by using these methods.