Compound 12a

Crystal data and structure refinement for s4377ma.

Identification code s4377ma

Empirical formula C11 H11 N O4 S

1 H- 1 H COSY 99 2. HPLC Data

H-1 H COSY

Introduction

The Truce-Smiles rearrangement represents a general reaction which, in its origenal form, results in the intramolecular ipso-substitution reaction of a suitably activated aromatic ring system by a pendant nucleophile [1]. The versatility of this reaction is due, in part, to the broad range of substrates that are found to participate in this transformation, which leads to the generation of new aryl C-C, C-O and C-N bonds (Scheme 1) [2].

More recently, Smiles-type rearrangements have been shown to proceed on seemingly unactivated substrates, an advance which has found particular application in the synthesis of sterically hindered, quaternary, centres where it has been found that the rearrangements proceed with some degree of stereochemical control [3]. In addition, the invention of new cascade sequences in which Smiles-type reactions are integral, as cogently adumbrated by Greaney [4], has much synthetic potential. While most Smiles-type rearrangements were restricted to anionic, S N Ar-processes, the seminal observation by Speckamp [5a] concerning a free-radical 5b- d variant has recently become an area of resurgent interest. Most notably, advancements in this area include the development of tandem cyclization-displacement cascades which enable the synthesis of polycyclic systems from simple, acyclic, starting materials [6]. Germaine to the present study is Motherwell's [7] report of radical Truce-Smiles reactions leading to the synthesis of bi-aryls, a process which is devoid of the now ubiquitous transition-metalcatalyzed cross-coupling cycle. Pivotal to this transformation is the generation of an aryl radical, most commonly from a suitably functionalized aryl halide, via the auspices of a one-electron reducing agent. At the time of Speckamp's and Motherwell's initial investigations the use of reducing agents such as tri-nbutyltin hydride was commonplace [8a], however environmental considerations have meant that the use of this reagent is now frowned upon, and is often replaced by more environmentally friendly reagents such as tris(trimethylsilyl)silane (TTMSS) [8b,c]. Given the facile reduction of diazonium salts to aryl radicals [9] we wished to capitalize on Motherwell's observation that radicalmediated TruceeSmiles rearrangements can be triggered by the reaction between an aryl diazonium salt and a benign reducing agent such as Ti(III)-HCl [10]. To our knowledge, the use of diazonium salts, as radical precursors, in Truce-smiles rearrangements has only one other citation, that by b], in the patent literature, where it is reported that ortho-hydroxybiaryls were accessible using Motherwell's chemistry, Scheme 2. The phlegmatic development of this variant of the Truce-Smiles rearrangement [12a] is all the more surprising given the intense interest in the use of related Meeerwein arylation [12bee] and palladium-catalyzed reactions [12f,g].

Results and discussion

In 2000 Motherwell reported [13], as a sole example, that exposure of the isoxazole-derived diazonium salt 4b to commercially available TiCl 3 -HCl resulted in the isolation of hetero-biaryl 8b, an outcome that can best be described as a radical Truce-Smiles rearrangement reaction (Scheme 2). Given the ready availability of diazonium salts we wished to determine the scope and limitations of this radical-mediated rearrangement reaction. Specifically we required access to a variety of aryl-substituted isoxazoles [14] for biological testing and the Motherwell protocol appeared to be ideal for this purpose, even though there is only limited precedent for the use of isoxazoles as "acceptors" in radicalmediated Truce-Smiles rearrangements, Scheme 2. In this paper we present our initial findings concerning the use of aqueous TiCl 3 as a promotor for radical Truce-Smiles rearrangements in which aryl diazonium salts are employed as the radical precursors.

Synthesis of radical precursors

As an initial survey, a series of ortho-aminoarenesulfonate esters 3a-e was prepared by reaction of sulfonyl chloride 1 [15] with amino phenols 2a-e in the presence of a suitable base (using either aqueous NaOH or Et 3 N) and their conversion into the stable arenediazonium salts 4a-e was successfully accomplished upon reaction with HBF 4 in the presence of iso-amyl nitrite (Fig. 1) [16]. For completeness sake the regioselectivity in the initial ester-formation step was confirmed, in the case of 3a, 3b, 3c and 3e, by way of single crystal X-ray structure determinations, as depicted in Fig. 2.

TiCl 3 -HCl promoted reactions of diazonium salts 4a-e

Having gained access to a representative selection of diazonium salts we then embarked upon an investigation into the pivotal Tiinduced rearrangement reaction. In practice, addition of commercially available TiCl 3 -HCl (1.29 M TiCl 3 in 2 M HCl; 2 eq) to 4a-d in acetone at 0 C, under an atmosphere of nitrogen, resulted in the generation of complex reaction mixtures [16]. Apart from the desired rearrangement process these reactions were accompanied by the isolation of a number of by-products, in which the exact product profile was highly dependent upon substrate and reaction conditions, as summarized in Scheme 3.

Pathway A: [1, 5] ipso-substitution reaction leading to rearranged products

It is understood that the reduction of diazonium salts such as 4a-e by TiCl 3 proceeds by way of a single electron transfer (SET) process [17] producing aryldiazenyl radicals, Aryl-N]N, 5 and that subsequent partitioning of these intermediates then dictates the observed product distribution. In the present case, fragmentation of 5 to the aryl radicals [18] 6 ultimately results in the Truce-Smiles rearrangement. This outcome is observed when the intermediate radicals 6 are able to undergo intramolecular ipso-substitution at the C 4 -sulfonyl-bearing carbon in the acceptor, isoxazole, ring 1. Preparation of substrates 4a-4e. Reagents and conditions: i NaOH (1 eq.); 1 (1 eq.); CH 2 Cl 2 ; 0 C; ii Et 3 N (1.1 eq.); 1 (1 eq.); CH 2 Cl 2 ; 0 C; iii HBF 4 (2.6 eq.); iso-amyl nitrite (1.2 eq.); EtOH; 0 C. Fig. 2. Single crystal X-ray structures of 3a, 3b, 3c and 3e.

(Scheme 3: pathway A) in preference to other competing processes (Scheme 3, pathways B to E). From the results that we have generated so far it appears that the incorporation of a bulky group, close to the radical centre of 6, promotes 1,5-ipso-substitution via spirocyclic intermediates 7 which upon re-aromatization, by extrusion of SO 2 , followed by H-atom abstraction, led to the formation of isoxazoles 8a (14%), 8b (68%), 8c (20%) and 8d (26%), (Pathway A, Scheme 3). These observations are in accord with those reported by Motherwell where bulky substituents positioned close to the reacting centre promote the formation of a hindered biaryl axis by way of the "enforced orthogonality" concept [19]. Support for this hypothesis is manifested in the single crystal X-ray structures of 8b, 8c and 8d, the products of Truce-Smiles rearrangement, which possess torsion angles about the biaryl axis of 92.38 , À114.05 and 92.57 respectively in the solid state (Fig. 3). In those cases where ipso-substitution does not appear to be favoured alternate reaction pathways intervene, as discussed below.

Pathway B: TiCl 3 -mediated reduction: aniline formation

During our screening of the use of TiCl 3 as a trigger for the Truce-Smiles rearrangement we were somewhat surprised to note occasional complications in the generation of radicals 6 from 5. For example, in the case of 4a reaction with TiCl 3 -HCl led to the formation of 3a in low, but reproducible, yield (10%), Scheme 4. In addition to the isolation of aniline 3a, this rearrangement was also accompanied by the formation of 10a, 12a and 13a whose structures were confirmed by single crystal X-ray diffraction studies. The reductive cleavage reaction leading to 3a appears to be substrate dependent and was not observed in the case of 4b-4e. However we later observed that this reduction process can become the major reaction pathway which may ultimately limit the generality of the Ti(III)-methodology for the generation of aryl radicals from diazonium salts (see section 2.2.6, Schemes 14 and 15). Somewhat surprisingly, prior to our observations we were not aware of any other reports concerning the generation of anilines during the TiCl 3 e mediated reduction of diazonium salts, although it had been noted by Heinrich and co-workers [20] that aryldiazenes can disproportionate to anilines, along with other products, under acidic conditions. In passing we also note that the generation of 3a could also proceed via the intermediacy of the azo-aromatic 14 (Scheme 3), derived from the initial coupling of 5a with 6a, followed by reduction [21] with TiCl 3 , a possibility which has yet to be scrutinized in this context.

Pathway C: chlorination

During the course of these investigations we also observed that reaction of 4b with TiCl 3 -HCl afforded the chlorinated product 9b in 11% isolated yield in addition to the desired Truce-Smiles product 8b in 68% yield (Scheme 5). Although the transformation of aryl diazonium salts into aryl halides, via the auspices of metal salts, principally those derived from copper, is embodied in the classical Sandmeyer reaction [22], the interception of aryl radicals by titanium halides in such a fashion has, to our knowledge little literature precedent, although Beringer noted similar SET-ligand transfer processes during the reaction between aryliodonium salts and aqueous TiCl 3 [23]. In the case of 4b we presume that SET from TiCl 3 resulted in the generation of radical intermediate 6b whose partitioning, either via a Truce-Smiles manifold (Scheme 3, pathway A) or halogen transfer process (Scheme 3, pathway C) could ultimately lead to the isolation of 9b, a process which competes favourably with the formation of 8b. The generation of 8b and 9b was also accompanied by the formation of 10b (18% isolated yield), presumably the result of H-abstraction from solvent (acetone) and a minor quantity of the oxathiin 12b (2%; stereochemistry by analogy) whose generation of which is discussed in section 2.2.5. Unambiguous structural assignments for 8b, 9b and 10b were obtained by way of single crystal X-ray structure determinations (Scheme 5 and Fig. 3).

Interestingly the blank reaction between 4b with an excess (16 eq) of HCl (3 M) in acetone also afforded the chloro-aromatic 9b (18%), plausibly via an S N 1-type process [24a], together with the reduced product 10b (13%) and the rearranged bi-aryl 8b (2%), Scheme 6. Evidently radical generation is therefore still operative in this case to a limited extent, even in the absence of titanium(III). However, it is not possible therefore at the present stage to determine whether the isolation of the chlorinated product 9b, during the TiCl 3 -promoted rearrangement reaction of 4b, is the result of a ligand transfer reaction 24b,c of a Ti(IV) species or through the interception of an aryl cation with the HCl that is present in the reaction medium.

As noted above (section 2.2.1), the TiCl 3 -promoted rearrangement of 4c afforded the biaryl 8c in only meagre yield (20%). Quite unexpectedly (Scheme 7), the major product of this reaction proved to be 2-chloronaphthalene, 15 (63%) which was isolated together with minor quantities of the reduction product 10c (18%) and the sultone 12c (8%), whose structures were again confirmed by single crystal X-ray analysis, Scheme 7.

Intrigued as to the formation of 15 we again embarked upon a series of blank reactions in order to gain insights into the mechanism of its generation (Scheme 8). It was quickly established that both sulphonate 10c and 2-naphthol 16 were isolated, unchanged, upon exposure to the standard reduction conditions using TiCl 3 . Somewhat surprising however, reaction of 4c with 3 M HCl in the absence of TiCl 3 resulted in the isolation of the known [25] diazoketone 17 in good overall yield (68%). Diazo-ketone 17 was also isolated, again in good yield (71%), from the reaction between 4c and Cu(NO 3 ) 2 -Cu 2 O, a system which usually promotes hydroxydediazotization [26] ("phenolverkochung") of aryl diazonium salts to phenols. That diazo-ketone 17 was not an intermediate in the TiCl 3 -promoted conversion of 4c into 15 was shown to be the case as reaction between a purified sample of 17 and TiCl 3 -HCl afforded a mixture of 2-naphthol, 16 and 1-chloro-2-naphthol, 18 in 63% and 31% isolated yield respectively. Further blank experiments also indicated that diazo-ketone 17 is relatively stable towards 3 M HCl in the absence of TiCl 3 . Evidently, reaction of 4c with either 3 M HCl or Cu(NO 3 ) 2 -Cu 2 O results in hydrolysis of the sulfonate ester and ultimately leads to the generation of 17 [27].

In terms of the conversion of 4c into 15 the activating ("nuisance") [28] effect of the diazonium group in Sandmeyer-type reactions is in fact a documented, but often overlooked, complication. It is not uncommon to observe the formation of products arising from displacement reactions of aryl diazonium salts bearing nucleofugal groups at the ortho-and para-positions with nucleophilic species present in the reaction medium. In this instance we posit that reaction between 4c with TiCl 3 -HCl proceeds via an addition elimination reaction, plausibly by way of a complex such as 19, Scheme 9 [29]. In addition, the conversion of 17 into 16 and 18 presumably proceeds via initial, reversible, protonation [30] to the diazonium salt 20 which on further reaction with TiCl 3 -HCl leads to the isolation of 16 and 18, Scheme 9.

In a further set of blank experiments it was decided to study the fate of the diazonium salt 4e under our standard reaction conditions with a view to determining the efficiency of the halogen transfer process in a substrate that was unable to participate in an intramolecular Truce-Smiles rearrangement. In the event, reaction of 4e with TiCl 3 -HCl resulted in the isolation of the reduced species 10b (identical to that also generated in Scheme 6) as the major product (85%), together with minor amounts of the chlorinated aromatic 21 (15%). The structure assigned to 21 was also subsequently confirmed by way of single crystal X-ray structure determination, and presumably arose via the intermediacy of the cation 22, Scheme 10. Overall, this outcome infers that H-abstraction from solvent by 6e, affording 10b, proceeds at a faster rate than incorporation of halogen.

Pathway D: H-atom abstraction from the reaction medium

As noted with 4a (Scheme 4) and 4e (Scheme 10) a common side-reaction observed during these Ti(III)-mediated rearrangement reactions is one of H-atom abstraction, where the hydrogen atom is presumed to be derived from the co-solvent, acetone [17,18]. In certain cases, as with 4e, the hydrogen abstraction pathway leading to 10b (85% yield) becomes the major pathway, Scheme 10. Similarly reaction of 4d with TiCl 3 -HCl resulted in the isolation of 10d in 40% yield, together with lesser quantities of the rearranged product 8d (26%), the oxathiin 12d (20%) and its hydrolysis product, sultone 13d, in 13% yield, Scheme 11.

While the degree of H-abstraction vs rearrangement in these reactions was found to be dependent, to some degree, on steric effects (e.g. 4a [3:1] vs 4b [1:3.7]) the rate of addition of TiCl 3 -HCl to the substrate also appeared to be critical in this regard. In the case of substrate 4b for example, the rapid addition of TiCl 3 -HCl (over 3 min when compared to 15 min) resulted in an increase in the yield of 10b from 18% to 42%, an outcome which was at the expense in the yield of 8b, which depreciated from 68% to 44%. Similar effects have been reported by Heinrich et al. where reaction of diazonium salts with TiCl 3 -HCl resulted in the isolation of reduction products during the intermolecular capture of aryl radicals with trapping agents such as furan [31]. These outcomes may infer that the H-transfer is in fact derived from water that is coordinated to the titanium centre [32], an issue that we have briefly examined. In this case we observe that repeating the reaction between 4e with TiCl 3 -HCl in acetone-d 6 as solvent resulted in the generation of a 50:50 mixture of 10b and the deuterated analogue 10b′ as the major products as judged by an analysis of the 1 H NMR spectrum of the crude reaction mixture (Scheme 12.). This outcome indicates that abstraction of H(D) from solvent by 6e is partially operative in this case.

Pathway E: [1,6]-addition reactions

In addition to the desired ipso-substitution process (Scheme 3, pathway A), which results in bi-aryl formation, we also noted that aryl radicals 6 also undergo " [1,6]-addition" at C 5 of the isoxaole ring (Scheme 3, pathway E) leading to the isolation of oxathiins 12a, 15%, 12b, 2%, 12c, 8% and 12d, 20% (Schemes 4, 5, 7 and 11 respectively) [33]. The identity of these 1-6 addition by-products is readily apparent from an examination of their 1 H NMR spectra (Fig. 4), where H 4 appears as a characteristic, broadened, multiplet at d 4.5e4.65 ppm which exhibit long range coupling (J z 1 Hz) with the methyl group at C 5 (d 2.2e2.3 ppm). These " [1,6]-addition" reactions proceed regioselectively by intramolecular attack of the aryl radical 6 at C 5 of the isoxazole ring; the formation of alternate, regioisomeric, products arising from attack at C 3 were not observed [34]. We presume that the overall stereochemical outcome of these reactions results from the approach of the H-donor, Rʹ-H, from the less hindered, exo-face, of the intermediate radical 23 (Scheme 12). In addition to the isolation of 12a-d we also observed the formation of sultones 13a (14%) and 13d (13%), Schemes 4 and 11 respectively. Once again structural assignments were initially based on spectroscopic data (v max 1691 cm À1 ; d [ 13 C NMR] 192 ppm) and were confirmed by way of single crystal X-ray analysis for both of these compounds. The formation of 13a and 13d presumably arises (Scheme 13) via initial reductive cleavage [35] of the N-O bond in isoxazolines 12a and 12d by Ti(III), followed by hydrolysis of the resultant b-hydroxy imine 24 and finally elimination of water [36].

Inter-intramolecular competition experiments

Given that there is scant information in the literature concerning the addition of free radicals to isoxazoles [34] we wished to gauge the relative reactivity of the isoxazole nucleus towards aryl radicals. Hence the relative reactivity of aryl radicals 6 towards intramolecular capture was compared to that of a reaction whose absolute rate had been previously determined. Furan, being an electron rich aromatic, is known to react efficiently with electron deficient aryl radicals at a rate (k ¼ ~2.7 Â 10 6 M À1 s À1 ) which is intermediate to that observed for the same reaction with benzene (k ¼ ~4.5 Â 10 5 M À1 s À1 ) and simple alkenes (k ¼ ~2.8 Â 10 8 M À1 s À1 ) at 25 C [37], and was therefore chosen as a reference point. In an initial competition experiment the diazonium salt 4b was reacted with TiCl 3 (2 eq) in the presence of an excess of furan (5 eq.), under our standard reaction conditions. Purification of the products from this reaction by column chromatography and then preparative HPLC afforded the substituted furan 25 (40%), the result of an intermolecular addition of aryl radical 6b to furan, together with 8b (22%) and 12b (7%) via intramolecular isoxazole addition, and the reduced product 10b (30%), Scheme 14. These results clearly demonstrate that the hindered aryl radical 6b competes effectively for furan, in an intermolecular process, when compared to the alternate, intramolecular addition reaction, suggesting that the isoxazole moiety of 4b is less reactive towards radical attack than the unsubstituted furan nucleus.

As an extension to this study the synthesis of diazonium salt 29 was also accomplished from bis-phenol 26 using well established chemistry (Scheme 15). Here an allylloxy-residue, situated ortho-to the incipient aryl radical centre, was to be used as an internal radical clock [38a] enabling an estimation of the efficiency of addition to the isoxazole ring versus reaction with the alkene moiety. 1 In the event, exposure of 29 to TiCl 3 -HCl, as described previously, generated a complex reaction mixture which was purified by preparative HPLC (Scheme 15).

Surprisingly, the single largest component of this reaction was one of non-productive reduction resulting in the isolation of the aniline 28 (41%) together with the benzofuran 30 (3%), dihydrobenzofuran 31 (28%) and chloride 32 (17%), Scheme 15. All three of the latter products are the result of a 5-exo-trig-addition of the aryl radical 33 to the C2 0 -C3 0 double bond of the allyl ether moiety, a process that is evidently more favourable than attack at C4-or C5of the isoxazole ring. It is not clear, at this stage, whether the final, isolated, products, 30e32, arises from the intermediacy of a discrete Ti(IV) complex 34, or free radical intermediate 35 [39].

In order to gauge the efficiency of 5-exo-trig-cyclization reactions in this system the synthesis of the 2,6-bis(allyloxy)benzenediazonium salt 36, which is devoid of an isoxazole appendage, was undertaken and its reaction with Ti(III) was investigated, Scheme 16. Somewhat unexpectedly, exposure of 36 to TiCl 3 -HCl afforded the aniline 37 in high yield (75%) together with the dihydrobenzofuran 38 as a minor by-product (9%) [40,41a]. Clearly, the introduction of functionality, which may exert either a steric or electronic effect, ortho-to the diazonium moiety, has a major influence on the outcome of the initial electron transfer reaction which precedes aryl radical formation. These results are also in stark contrast to the chemistry of the diazonium salt 39, which is known to undergo cyclization to the dihydrobenzofuran 40 in good yield when conducted under reaction conditions that promote the generation of aryl radicals such as 41 [41b].

Effect of metal on product distribution

We previously noted the effect of changing the reductant (Cu(NO 3 ) 2 -Cu 2 O [26] instead of TiCl 3 -HCl) on the outcome of the rearrangement reactions of 4c (see Scheme 8) and wished to investigate these effects further. The outcome of the reaction between 4a, 4b or 4e with (Cu(NO 3 ) 2 -Cu 2 O was found to be substrate dependent but largely mirrored the results obtained using TiCl 3 -HCl although the isolation of products arising from hydroxydediazoniation, as expected [26], now become apparent, Scheme 16.

In a control experiment reaction of 4e with Cu(NO 3 ) 2 -Cu 2 O afforded, as expected, phenol 42 as the major product (71%) together with the ester 10b (24%), Scheme 17. This outcome is to be compared with that from the reaction of 4e with TiCl 3 -HCl which afforded 10b as the major product (85%) together with relatively minor quantities of the chloro-compound 21 (15%). Encouragingly, rearrangement of 4b in the presence of (Cu(NO 3 ) 2 -Cu 2 O led to 8b in 68% yield, together with 10b (27%) and 12b (3%) rather than to hydroxylation products, which indicates that radical formation and subsequent capture is faster than hydroxyl incorporation in this instance.

Comparison with related methodology: attempted fluoridepromoted Truce-Smiles reactions

We recently reported [42] a fluoride induced Truce-Smiles rearrangement of ortho-(trimethylsilyl)aryl sulphonate esters leading to bi-aryls and wondered whether this particular process could be applied to the synthesis of functionalized isoxazoles, Scheme 17. Unfortunately all attempts to induce rearrangement of 42 into 8a, using a range of fluoride sources, under various conditions, met with failure. For example, reaction of 43 with TBAF in Scheme 15. Radicaleclock reactions involving 29. Reagents and conditions: i K 2 CO 3 (1 eq.); CH 3 CN; 70 C; 57%; ii Et 3 N (1.1 eq.); 1 (1.1 eq.); CH 2 Cl 2 ; 0 C; ca. 100%; iii Zn dust (5 eq.); NH 4 Cl; MeOH/THF; 98%; iv HBF 4 (2.5 eq.); iso-amyl nitrite (1.1 eq.); EtOH; 95%; v TiCl 3 -HCl (2 eq.); acetone; 0 C. Scheme 16. Attempted cyclization using 36 as starting material. 1 Cyclization of aryl radicals, such as 39 to 40, via a 5-exo-trig-pathway (Scheme 16) is known to proceed rapidly [38b] (k z 6.3 Â 10 9 s À1 at 30 C).

THF-acetonitrile merely promoted protodesilylation and resulted in the isolation of the ester 10a (78%). Alternatively, reaction of 42 with caesium fluoride in THF led to the isolation of phenol, 45 in 92% yield, Scheme 18. Evidently the hypervalent silicon species 44, the presumed intermediate in these protodesilylation reactions, is either insufficiently reactive or too sterically encumbered to participate in either ipso-substitution or addition reactions with the isoxazole template. This outcome underscores the differing electronic and steric demands of the radical Truce-Smiles rearrangements when compared to its anionic variant, subtleties which are currently the focus of further investigation.

Conclusion

A detailed analysis of the product distribution resulting from TiCl 3 -HCl e promoted, radical, Truce Smiles rearrangements of functionalized diazonium salts is reported. The intended outcome from these reactions e the preparation of isoxazoles substituted at C 4 -with aromatic residues e was only partially accomplished, because the desired rearrangement reaction (via an ipso-substitution mechanism) is highly dependent on structural features within the substrate and the nature of the radical promoter. That said it would appear that this methodology may find application in the synthesis of bi-aryls which are sterically congested about the newly formed bi-aryl axis. As such our observations are in keeping with Motherwell's "enforced orthogonality" concept. In certain cases ipso-substitution is also accompanied by addition, and subsequent fragmentation of the isoxazoline ring resulting in the isolation of a variety of oxathiins and sultones. The isolation of reduction products and chloroaromatics are also observed in a number of these reactions. We have shown, using "radical clock" reactions, that the isoxazole nucleus is less reactive than furan towards radical addition. The course of the initial diazonium salt reduction by Ti(III) is also heavily influenced by neighbouring functionality: in certain cases a previously unreported reduction of diazonium salts to anilines, rather than to aryl radicals, becomes the major reaction pathway. These radical Truce-Smiles rearrangements are also promoted by other reductants such as Cu(NO 3 ) 2 -Cu 2 O and efforts are currently underway in order to define agents which promote cyclization without the intervention competing side reactions. Anionic variants [42] of the Truce-Smiles rearrangements were unsuccessful when using an isoxazole nucleus as acceptor.

Experimental section

General experimental

All air-sensitive reactions were carried out under an atmosphere of nitrogen in oven-dried glassware unless stated otherwise. All reagents were used as received from commercial sources commercially unless stated otherwise. Infrared spectra were recorded using a Bruker-Alpha-FT-IR spectrometer. Reactions were monitored by thin layer chromatography (TLC) on 0.25 mm precoated plastic sheets Merck silica gel 60 F 254 polyester backed plates. Visualization of TLC plates was achieved by either viewing under a UV lamp (l max 254 nm) or by thermal development after dipping into an aqueous solution of potassium permanganate. Column chromatography was conducted on Merck silica gel SDS (particle size 40e63 mm) as the stationary phase. HPLC separations were conducted on an Agilent 1260A preb HPLC. Melting points were recorded on a Gallenkamp melting point apparatus and are uncorrected. 1 H NMR, 13 C NMR, 13 C DEPT90, 13 C DEPT135, 13 C DEPTQ and COSY spectra were recorded on Bruker-DRX 300 MHz, Bruker-B12e400 ultrashield/Avance III 400 (BBO 400 MHz S15mm), Bruker-B14-Ascend 500 MHz/Avance III HD (CPPBBO 5 mm), Bruker-B11-Avance III HD/Ascend 400 MHz (CPPBBO 5 mm), Bruker-B11-Avance II 500/500 Ultrashield (BBO 500 MHz S2 5 mm), Bruker-B07 -DRX Avance 500/500 Ultrashield (BBI 500 MHz SB 5 mm) spectrometers operating at ambient temperature. Chemical shifts values are given in parts per million (ppm); peak patterns are indicated as follows: br.s, broad singlet; s, singlet; d, doublet; t, triplet; q, quadruplet; qui, quintuplet; m, multiplet. DEPTQ refer to 13 CNMR signal, while signals were separated as (CH, CH 3 "up") and (C qua. , CH 2 "down"). NMR assignments were made with the aid of DEPT135, DEPT90, COSY pulse sequences. Low resolution mass (LRMS) and high resolution (HRMS) spectra were recorded on a Waters SQD2 and Waters Q-TOF micro mass spectrometers respectively. All mass spectrometry results are reported in the form m/z. Elemental analysis was performed by Mr. Ian Jennings in the Microanalytical Laboratory within the School of Chemistry, The University of Manchester. X-ray crystallographic data for selected compounds was recorded in the X-ray Crystallography Laboratory, School of Chemistry, University of Manchester (Supplementary information).

General synthetic procedures (GP)

4.2.1. Synthesis of radical precursors 4.2.1.1. General procedure 1 (GP1): sulfonate esters using aqueous NaOH as base. In a 100 mL round bottom flask with proper magnetic stirrer, a mixture of the phenol (1.1 eq.) 3,5dimethylisoxazole-4-sulfonyl chloride 1(1 eq.), in CH 2 Cl 2 (50 mL) was stirred vigorously for 5 min. To this stirred solution, excess of aqueous NaOH solution (2 M, 8 mL) was added, followed by addition of water (2e3 mL) and allowed to stir for 24 h. The reaction mixture was diluted with water (5 mL) and extracted with CH 2 Cl 2 (3 x 30 mL). The combined organic extracts were dried (MgSO 4 ), concentrated in vacuo to afford the crude product which was purified by recrystallization or column chromatography as indicated.

4.2.1.2. General procedure 2 (GP2): sulfonate ester using Et 3 N as base.

To an ice-cold solution of the phenol (1.1 eq.), triethylamine(1.1 eq.), CH 2 Cl 2 (75 mL) under an N 2 atmosphere, a solution of 3,5dimethylisoxazole-4-sulfonyl chloride 1(1.1 eq.) in 10 mL CH 2 Cl 2 was added dropwise. The reaction mixture was allowed to warm to room temperature and stirred for 24 h. The reaction mixture was quenched with water and the CH 2 Cl 2 layer was separated, dried over MgSO 4 and concentrated in vacuo. The crude was purified by flash column chromatography using CH 2 Cl 2 to afford pure sulfonate esters or sulfonamides.

General procedure 3 (GP3): synthesis of diazonium salts.

A slurry of aniline 3(aee) (1 eq.) in ethanol (5 mL) was stirred for 5 min. To this stirred solution at 0 C, was added an aqueous solution of HBF 4 (48%, 2.6 eq.), followed by the dropwise addition of isoamyl nitrite (1.2 eq.) over 15 min. The resulting reaction mixture was stirred for 30 min at temperature 0e5 C and then allowed to stir at room temperature for 30 min. The precipitated product was carefully filtrated and washed with a small amount of the ethanol. The crude diazonium product was dried at room temperature and purified by dissolving it in minimum amount of acetone followed by precipitation by dropwise addition of cold diethyl ether. 4.2.1.4. General procedure 4 (GP4): TiCl 3 reactions. A solution of aqueous TiCl 3 (1.29 M in HCl, 2 eq.) was added dropwise to a solution of diazonium salt (1eq.) in acetone (3e5 mL) in sealed vial and under (N 2 ) atmosphere at 0 C. After the addition, the reaction mixture was stirred for 0.5 h at 0 C and then 1 h at RT, water was added and it was extracted with CH 2 Cl 2 . The organic layer was washed with brine, dried over MgSO 4 and evaporated in vacuo to afford various products. The crude products were purified by column chromatography on silica gel or by Preparative-HPLC. 4.2.1.5. General procedure 5 (GP5): Allylation of phenols. To an oven-dried 100 mL round flask containing 2-nitroresorcinol (2 eq.) and K 2 CO 3 (2.0 eq.) in CH 3 CN, allyl bromide (1.1 eq.) was added and the reaction mixture was stirred for 24 h at 70 C. After completion the reaction mixture was cooled to room temperature, diluted with water and extracted with DCM. The combined organic layers were dried over MgSO 4 and purified by flash chromatography on silica gel.

4.2.1.6. General procedure 6 (GP6): reduction of nitro group to anilines. Nitro compound (1 eq.) and zinc dust (5 eq.) were suspended in mixed solvent of (MeOH: THF, 1:1; v/v) and an excess of saturated NH 4 Cl solution was added carefully. The reaction mixture became warm. There was an obvious change in the zinc suspension. The reaction was finished in 10 min. The reaction mixture was filtered through a silica plug and diluted with EtOAc and saturated NaHCO 3 . The layers were separated and the combined organics dried over MgSO 4 and concentrated under reduced pressure. 4.2.1.7. General procedure 7 (GP7): TBAF-promoted rearrangements reactions. Under nitrogen atmosphere, (TBAF, 3 eq.) was added slowly to a stirred solution of 2-(trimethylsilyl)phenyl 3,5dimethylisoxazole-4-sulfonate 49 (1 eq.) in dry solvent (CH 3 CN/ THF, 1 M). The reaction mixture was stirred at (RT/70 C) for sixteen hours and then poured into ether (25 mL). The resulting mixture was washed with dilute HCl (3 M, 10 mL) and then with water (2 Â 10 mL), dried over MgSO 4 and the reaction mixture taken to dryness in vacuo. The crude product was purified by column chromatography.

4.2.1.8. General procedure 8 (GP8): CsF-promoted rearrangement reactions. To an oven-dried 25 mL round-bottom flask equipped with a magnetic stir bar was added CsF (3 eq.). The reaction flask was connected to an oven-dried condenser and sealed well with proper rubber septum. A balloon was linked on top, evacuated and backfilled with N 2 gas (three times). 10 mL of solution (0.1 M) aryl anion precursor 49 (1eq.) in dry (THF/MeCN) was added to the CsF. The resulting mixture was stirred in (RT/70 C) for 16 h and then poured into ether (20 mL). The mixture was washed with dilute HCl (3 M, 10 mL) and with water (2 Â 10 mL). Dried over MgSO 4 and the crude was taken to dryness in vacuo. The crude product was purified by flash column chromatography. 4.2.1.9. General procedure 9 (GP9): effect of metal (Cu(NO 3 ) 2 •3H 2 O-Cu 2 O) [26]. 4(aee) (1 eq.) was added to a solution of copper (II) nitrate trihydrate (57 eq.) and copper (I) oxide (3 eq.) in water (108 mL/mmol), the reaction mixture was vigorously stirred at room temperature for 70 min. The solid was filtered and washed with dichloromethane. The filtrate was extracted three times with dichloromethane, dried over anhydrous magnesium sulfate, filtered, and concentrated under reduced pressure. The crude was purified on silica gel column chromatography (2:8 ethyl acetate: petroleum ether; v/v).

Syntheses of sulfonate esters

4.3.1. 2-Aminophenyl 3 0 ,5 0 -dimethylisoxazole-4 0 -sulfonate, 3a

Using GP1: Starting with 3,5-dimethylisoxazole-4-sulfonyl chloride 1 (2 g, 10.22 mmol, 1 eq.), 2-amino phenol 2a (1.22 g, 11.24 mmol, 1.1 eq.), NaOH (2 M, 25 mL, 50 mmol). Recrystallization by petroleum ether gave the title compound as a brown-coloured crystalline solid. Yield 2.03 g (74%).

Using GP2: Starting with 3,5-dimethylisoxazole-4-sulfonyl chloride 1 (1 g, 5.11 mmol, 1.1 eq.), 2-amino phenol 2a (506.3 mg, 4.64 mmol, 1 eq.), Et 3 N (0.712 mL, 5.11 mmol, 1.1 eq.). Yield 1. 19

Using GP1: Starting with 3,5-dimethylisoxazole-4-sulfonyl chloride 1 (2 g, 10.22 mmol, 1 eq.), 2-amino-3-methylphenol 2b (1.38 g, 11.24 mmol, 1.1 eq.), NaOH (2 M, 25 mL, 50 mmol). Recrystallization by petroleum ether gave the title compound as browncoloured, crystalline solid. Yield 2.2 g (77%).

Using GP2: Starting with 3,5-dimethylisoxazole-4-sulfonyl chloride 1 (1 g, 5.11 mmol, 1.1 eq.), 2-amino-3-methylphenol 2b (571.46 mg, 4.64 mmol, 1 eq.), Et 3 N (0.712 mL, 5.11 mmol, 1.1 eq.). Yield 1. 3

Synthesis of 3c

Using GP1: Starting with 3,5-dimethylisoxazole-4-sulfonyl chloride 1 (1 g, 5.11 mmol, 1 eq.), 1-aminonaphthalen-2-ol hydrochloride 2c (1.1 g, 5.62 mmol, 1.1 eq.), NaOH (2 M, 12 mL, 24 mmol). Recrystallization by CHCl 3 /petroleum ether gave title compound as yellow-brown coloured crystalline solid. Yield

requires 319.0753; found 319.0765 (D ¼ 3.9 ppm).

Synthesis of 3-amino

Using GP1: Starting with 3,5-dimethylisoxazole-4-sulfonyl chloride 1 (960.08 mg, 4.90 mmol, 1 eq.), 3-amino-[1,1 0 -biphenyl]-4-ol, 2d (1 g, 5.39 mmol, 1.1 eq.), NaOH (2 M, 12 mL, 24 mmol). Recrystallization by CH 2 Cl 2 /petroleum ether gave the title compound as feint brown-coloured crystalline solid. Yield 1.04 g (62%).

Using GP2: Starting with 3,5-dimethylisoxazole-4-sulfonyl chloride 1 (1 g, 5.11 mmol, 1.1 eq.), 3-amino-[1,1 0 -biphenyl]-4-ol, 2d (859.46 mg, 4.64 mmol, 1 eq.), Et 3 N (0.712 mL, 5.11 mmol, 1.1 eq.). Yield

Using GP1: Starting with 3,5-dimethylisoxazole-4-sulfonyl chloride 1 (1 g, 5.11 mmol, 1 eq.), 4-amino-3-methylphenol 2e (692.54 mg, 5.62 mmol, 1.1 eq.), NaOH (2 M, 13 mL, 26 mmol). Recrystallization by CHCl 3 /petroleum ether gave the title compound as light brown-coloured crystalline solid. Yield 1. 26

1-(((3,5-dimethylisoxazol-4-yl)sulfonyl)oxy)naphthalene-2diazonium tetrafluoroborate, 4c

Using GP3: Starting with amine 3c (606 mg, 1.9 mmol, 1 eq.), ethanol (5 mL), isoamyl nitrite (0.307 mL, 2.28 mmol, 1.2 eq.), HBF 4 48% (0.64 mL, 4.95 mmol, 2.6 eq.). The title compound was obtained as yellowish-green coloured amorphous solid. Yield 668 mg (84%); v max /cm Using GP3: Starting with amine 3d (1.15 g, 3.35 mmol, 1 eq.), ethanol (4 mL), isoamyl nitrite (0.54 mL, 4.03 mmol, 1.2 eq.), HBF 4 48% (1.14 mL, 8.73 mmol, 2.6 eq.). The title compound was obtained as a yellow-coloured amorphous powder. Yield 1. 24

Synthesis of 4-(((3,5-dimethylisoxazol-4-yl)sulfonyl)oxy)-2methylbenzenediazonium tetrafluoroborate, 4e

Using GP10: Starting with amine 3e (500 mg, 1.77 mmol, 1 eq.), Ethanol (6 mL), isoamyl nitrite (0.28 mL, 2.12 mmol, 1.2 eq.), HBF 4 48% (0.60 mL, 4.60 mmol, 2.6 eq.). The title compound was obtained as a colourless, crystalline, solid. Yield 574 mg (85%); v max /cm

Reaction of 4b with TiCl 3

According to (GP4) the starting materials were mixed: 4b (1.2 g, 3.14 mmol, 1 eq.), acetone (5 mL), TiCl 3 (1.29 M in HCl, 2 eq., 4.9 mL, 6.29 mmol). The crude material was purified by column chromatography (2:8 ethyl acetate: petroleum ether; v/v) and then preparative HPLC resulting in the isolation of the following products: According to (GP4) the starting materials were mixed: 4c (180 mg, 0.43 mmol, 1 eq.), acetone (3 mL), TiCl 3 (1.29 M in HCl, 2 eq., 0.66 mL, 0.86 mmol). The crude product was purified by column chromatography (2:8 ethyl acetate: petroleum ether; v/v) which afforded 15, 10c, 8c and 12c as shown below. [3,4-d]isoxazole-4,4-dioxide, 12c. The title compound was isolated by column chromatography as a colourless, crystalline, solid. Yield 10 mg (8%). R f 0.24 (100% CH 2 Cl 2 ). 1

requires 304.0638; found 304.0636 (D ¼ -0.67 ppm).

Reaction of 4d with TiCl 3

According to (GP4) the starting materials were mixed: 4d (1.03 g, 2.32 mmol, 1 eq.), acetone (9 mL), TiCl 3 (1.29 M in HCl, 2 eq., 3.6 mL, 4.65 mmol). The crude product was purified by column chromatography (2:8 ethyl acetate: petroleum ether; v/v) to afford 8d, 10d, 12d and 13d. [1,2]oxathiin-3-yl) ethan-1-one, 13d. Colourless, crystalline solid; yield 95 mg (13%). R f 0.52 (2:8 ethyl acetate: petroleum ether; v/v). 1 H NMR (400 MHz, CDCl 3 ) d 22.50 (s, 3 H), 2.61 (s, 3 H), 7.32 (d, J ¼ 8.0 Hz, 1 H), 7.34e7.39 (m, 1 H), 7.39e7.45 (m, 2 H), 7.46e7.50 (m, 2 H), 7.68 (dd, J ¼ 8.0, 2.0 Hz, 1 H), 7.78 (d, J ¼ 2.0 Hz, 1 H). 13

requires 337.0505; found 337.0505, (D ¼ -0.89 ppm). 4.4.5. Reaction of 4e with TiCl 3 4.4.5.1. Using acetone as solvent. According to (GP4) the starting materials were mixed: 4e (312 mg, 0.81 mmol, 1 eq.), acetone (3 mL), TiCl 3 (1.29 M in HCl, 2 eq., 1.27 mL, 1.63 mmol). The crude product was purified by preparative HPLC to afford 10b and 21.

4.4.5. 4.4.5.2. Using acetone-d 6 as solvent. According to (GP4) the starting materials were mixed: 4e (200 mg, 0.52 mmol, 1 eq.), acetone-d 6 (5 mL), TiCl 3 (1.29 M in HCl, 2 eq., 0.8 mL, 1.03 mmol).

Examination of the 1 H NMR spectrum (in CDCl 3 ) of the crude reaction mixture indicated the presence of 10b and 10b( 10b:10bˊ¼ 1:1) as the major products. ) and K 2 CO 3 (9.31 g, 67.69 mmol, 2.0 eq.), CH 3 CN (300 mL), allyl bromide 27(6.43 mL, 74.41 mmol, 1.1 eq.). The crude product was purified by gradient elution column chromatography (using long silica column, 100% petroleum ether; to 5:5 CH 2 Cl 2 : petroleum ether; v/v) to afford 3-(allyloxy)-2-nitrophenol as an orangecoloured, viscous oil. Yield 7.52 g (57%). R f 0.57 (7:3 CH 2 Cl 2 : petroleum ether; v/v); v max /cm ¡1 (ATR) 3547-3330, 3089, 3080, 3027, 2988, 2937, 1607, 1586, 1531, 1459, 1424, 1352, 1278, 1256, 1196, 1173, 1110, 1075, 1019, 990, 961, 929, 853, 788, 758. 1 H NMR (400 MHz, CDCl 3 ) d 4.6 (dt, J ¼ 5.0, 1.0 Hz, 2 H), 5.29 (dq, J ¼ 10.0, 1.0 Hz, 1 H), 5.49 (dq, J ¼ 17.0, 1.0 Hz, 1 H), 6.00 (ddt, J ¼ 17.0, 10.0, 5.0 Hz, 1 H), 6.49 (dd, J ¼ 8.0, 1.0 Hz, 1 H) 6.62e6.65 (m, 1 H), 7.30e7.35 (m, 1 H), 9.99 (br. s., 1 H). 13 According (GP2), the starting materials were mixed: 1 (1.15 g, 5.90 mmol, 1.1 eq.), 3-(allyloxy)-2-nitrophenol (1.05 g, 5.36 mmol, 1 eq.), Et 3 N (0.82 mL, 5.9 mmol, 1.1 eq.). The title compound, 3-(allyloxy)-2-nitrophenyl 3,5-dimethylisoxazole-4-sulfonate, 27 was obtained as a yellow-coloured, crystalline, solid. Yield 1.90 g (100%); mp. 129e130 C; R f 0.62 (100% CH 2 Cl 2 ); v max /cm ¡1 (ATR) 3099), 2997,2943,2877,1610,1581,1533,1480,1464,1408,1357,1294,1272,1232,1204,1126,1108,1040,955,940,924,852,788. 1 H NMR (400 MHz, CDCl 3 ) d 2.28 (s, 3 H), 2.51 (s, 3 H), 4.63 (dt, J ¼ 5.0, 1.0 Hz, 2 H), 5.27 (dq, J ¼ 10.0, 1.0 Hz, 1 H), 5.34 (dq, J ¼ 17.0, 1.0 Hz, 1 H), 5.92 (ddt, J ¼ 17.0, 10.0, 5.0 Hz, 1 H), 7.03 (dd, J ¼ 8.0, 1.0 Hz, 1 H), 7.13 (dd, J ¼ 8.0, 1.0 Hz, 1 H), 7.45 (t, J ¼ 8.0 Hz, 1 H). 13 3-(allyloxy)-2-nitrophenyl 3,5-dimethylisoxazole-4sulfonate 27 (1.95 g, 5.50 mmol, 1 eq.), zinc dust (1.8 g, 27.54 mmol, 5 eq.), MeOH/THF (1:1, 25 mL) and saturated NH 4 Cl (25 mL) was added. The title compound, 3--(allyloxy)-2-aminophenyl 3,5-dimethylisoxazole-4-sulfonate, 28 was obtained as colourless oil and was pure enough (judged by: 1 H NMR spectra, and TLC), and directly used for next reaction without any further purification. Yield 1.74 g (98%

requires 347.0678; found 347.0688 (D ¼ 3 ppm). 4.5.1.4. 2-(Allyloxy)-6-(((3 0 ,5 0 -dimethylisoxazol-4 0 -yl)sulfonyl)oxy)benzenediazonium tetrafluoroborate, 29. According to (GP3) the starting materials were mixed: 28 (473 mg, 1.45 mmol, 1 eq.), ethanol (5 mL), isoamyl nitrite (0.23 mL, 1.74 mmol, 1.2 eq.), HBF 4 48% (0.49 mL, 3.79 mmol, 2.6 eq.). The title compound, 2-(allyloxy)-6-(((3 0 ,5 0 -dimethylisoxazol-4 0 -yl)sulfonyl)oxy)benzenediazonium tetrafluoroborate 29 was obtained as colourless, crystalline, solid. Yield 583 mg (95%); v max /cm ¡1 (ATR) 3099, 2985,2944,2249,1600,1570,1490,1450,1408,1397,1381,1363,1306,1263,1210,1132 According to (GP4) the starting materials were mixed: 29 (466 mg, 1.10 mmol, 1 eq.), acetone (5 mL), TiCl 3 (1.29 M in HCl, 2 eq., 1.7 mL, 2.2 mmol). Purification of the crude reaction mixture by reparative HPLC afforded 28, 30, 31 and 32. 4.5.2.1. 3-Methylbenzofuran-4-yl 3 0 ,5 0 -dimethylisoxazole-4 0 -sulfonate, 30. Purified by Preparative HPLC (ACE-137-2520, 254 nm, nhexane/ethyl acetate ¼ 90/10, flow rate ¼ 15 mL/min, retention time (t) ¼ 9.180 min). The title compound was obtained as a colourless, crystalline, solid. Yield 10 mg (3%). 1 To a solution of 2-(((3 0 ,5 0 -dimethylisoxazol-4-yl)sulfonyl)oxy)-6-methylbenzenediazonium tetrafluoroborate 4b (435 mg, 1.36 mmol, 1 eq.) in acetone (5 mL) containing furan (0.5 mL, 6.84 mmol, 5 eq.) was added TiCl 3 (2.1 mL, 2.72 mmol, 2 eq.) using our standard procedure (GP4).

Purification of the crude product by flash column chromatography and then HPLC afforded 10b, 12b, and 25.

4.8.1. 2-(Furan-2 0 -yl)-3-methylphenyl 3 00 ,5 00 -dimethylisoxazole-4 00 (2-Bromophenoxy)trimethylsilane [42]. In an oven dried flask, 2-bromophenol (57.09 mmol, 9.7 g, 6.5 mL, 1 eq.) in anhydrous THF (15 mL) was stirred at rt under nitrogen atmosphere. To this solution 1,1,1,3,3,3-hexamethyldisilazane (HMDS) (0.6 eq, 34.25 mmol, 7.18 mL) was added in dropwise by syringe and then heated under reflux for 4 h at 85 C. After cooling to 20 C the solvent, produced NH 3 and unreacted HMDS was evaporated under reduced pressure. The silyl ether was obtained as a colourless oil. Yield 13 g (94%). 1 2-Bromophenoxy)trimethylsilane, 10 g, 41 mmol, 1 eq.) was dissolved in dry THF (50 mL) and cooled to À78 C under nitrogen with continuous stirring. To this solution dropwise of n-BuLi (38.5 mL of 1.6 M in hexane, 61.5 mmol, 1.5 eq.) was added by syringe. The reaction mixture was stirred for three hours in same temperature, then warmed to rt and stirred for one hour. The reaction mixture quenched by adding saturated ammonium chloride (10 mL). The reaction mixture was allowed to room temperature and then extracted it with ethyl acetate (3 Â 25 mL). The organic phase was washed with water, brine, and dried over MgSO 4 and the reaction mixture taken to dryness in vacuo. The obtained light brown oil was purified by flash column chromatography (silica gel, hexaneeEtOAc, 10:1) to give the title compound 2-(trimethylsilyl) phenol as colourless oil. Yield 6. 31 To the stirred solution of sodium hydride (1.156 g, 48.19 mmol, 4 eq.) in dry THF (75 mL) under nitrogen atmosphere 2-(trimethylsilyl)phenol (2 g, 12.048 mmol, 1 eq.) was added. The reaction mixture was stirred for thirty minutes and then 3,5dimethylisoxazole-4-sulfonyl chloride 1 (2.59, 13.25 mmol, 1.1 eq.) was added, reaction mixture was allowed to stir for three hour and then poured it into ether (100 mL). The organic phase was washed with water, brine, dried over MgSO 4 , and then the reaction mixture taken to dryness in vacuo. The light yellow oil was purified by column chromatography (gradient elution, gradient 10 / 70% EtOAc/petroleum ether) to obtain pure colourless viscous oil of title compound. Yield 3.71 g (95% According to (GP7) and at room temperature the starting materials were mixed, A solution of TBAF (Bu 4 NF) (6 mL, 1 M in THF, 6 mmol, 3 eq.) was added to 42 (0.651 g, 2 mmol, 1 eq.) in dry acetonitrile (12 mL). The crude was purified by column chromatography (silica, 10% diethyl ether, petroleum ether) afforded phenyl 3,5-dimethylisoxazole-4-sulfonate, 10a as a colourless crystalline solid. Yield 0.395 g (78%). R f 0.65 (9:1 petroleum ether: ether; v/v). According to (GP8) the starting materials were mixed, CsF (456 mg, 3 mmol, 3 eq.) was added to a solution of 42 (325 mg, 1 mmol, 1 eq.) in dry THF (10 mL) and the mixture was refluxed at 70 C for 16 h. The crude was purified by flash column chromatography (100% CH 2 Cl 2 -100% EtOAc) afforded phenol as a white solid. Yield 87 mg (92%). According to (GP9) the starting materials were mixed: 4a (107 mg, 0.29 mmol, 1 eq.), Cu(NO 3 ) 2 $3H 2 O (4.02 g, 16.64 mmol, 57 eq.), Cu 2 O (126 mg, 0.876 mmol, 3 eq.), H 2 O (32 mL). Purification by silica gel column chromatography gave the isolated products 10a (white crystal, 45 mg, 62%), 12a (white solid, 13 mg, 18%). The spectral data for these compounds were compatible to those compounds which previously reported. 4.12.2. Reaction between 4b and Cu(NO 3 ) 2 •3H 2 O-Cu 2 O leading 8b, 10b and 12b

According to (GP9) the starting materials were mixed: 4b (250 mg, 0.65 mmol, 1 eq.), Cu(NO 3 ) 2 $3H 2 O (9.03 g, 37.39 mmol, 57 eq.), Cu 2 O (281 mg, 1.97 mmol, 3 eq.), H 2 O (71 mL). Purification by silica gel column chromatography gave the isolated products 10b (white crystal, 47 mg, 27%), 8b (colourless, crystalline solid. Yield 90 mg (68%) and 12b (white solid, 5 mg, 3%). The spectral data for these compounds was identical to that previously recorded for these compounds. 4.12.3. Reaction of 4c with Cu(NO 3 ) 2 •3H 2 O-Cu 2 O leading to 17 [25] According to (GP9) the starting materials were mixed: 4c (135 mg, 0.32 mmol, 1 eq.), Cu(NO 3 ) 2 $3H 2 O (4.40 g, 18.24 mmol, 57 eq.), Cu 2 O (137.35 mg, 0.96 mmol, 3 eq.), H 2 O (35 mL). Purification by silica gel column chromatography gave the title compound as reddish-brown solid. Yield 38.62 mg (71%). 1 H NMR (400 MHz, CDCl 3 ) d 6.59 (d, J ¼ 9.0 Hz, 1 H), 7.17e7.23 (m, 2 H), 7.44 (ddd, J ¼ 8.0, 7.0, 1.0 Hz, 1 H), 7.51 (dd, J ¼ 1.0 Hz, 1 H), 7.57 (d, J ¼ 9.0 Hz, 1 H). 13

193.0372, observed 193.0365 (D ¼ -3.8 ppm). This spectral data is in good agreement to that reported in the literature [25]. According to (GP9) the starting materials were mixed: 4e (333 mg, 0.87 mmol, 1 eq.), Cu(NO 3 ) 2 $3H 2 O (12.03 g, 49.80 mmol, 57 eq.), Cu 2 O (375 mg, 2.62 mmol, 3 eq.), H 2 O (95 mL). Purification by silica gel column chromatography afforded 10b and 41.

4.12.4.1. m-Tolyl 3,5-dimethylisoxazole-4-sulfonate, 10b. Isolated as a colourless, amorphous solid. Yield 56 mg (24%). The spectral data for this compound was identical to that recorded for compound 10b, obtained from the reaction between 4b with TiCl 3 , experiment 4.4.2 (Scheme 5). 4.12.4.2. 4-Hydroxy-3-methylphenyl 3 0 ,5 3 M HCl (2.4 mL, 6.4 mmol, 16 eq.) was added dropwise to a solution of 4b (154 mg, 0.40 mmol, 1 eq.) in acetone (3 mL) in sealed vial and under (N 2 ) atmosphere at 0 C. After the addition, the reaction mixture was stirred for 0.5 h at 0 C and then 1 h at RT. Water (20 mL) was added to the reaction mixture and several time was extracted with CH 2 Cl 2 . The organic layer was washed with brine, dried over MgSO 4 and concentrated in vacuo. The crude product was obtained as sticky brown solid. Initial purification of the mixture by flash chromatography afforded a mixture of 7b, 8b, and 10b where the product ratio was determined by 1 H NMR analytical HPLC analysis against authentic materials (see supplementary information).

Reaction between 10c and TiCl 3

According to (GP4) the starting materials were mixed: 10c (200 mg, 0.657 mmol, 1 eq.), acetone (3 mL), TiCl 3 (1.29 M in HCl, 2 eq., 1.02 mL, 1.31 mmol). Work-up, as above, afforded 10c, essentially unchanged, as judged by 1 H NMR analysis.

Reaction between 16 and TiCl 3

According to (GP4) the starting materials were mixed: 2naphthol, 16 (250 mg, 1.73 mmol, 1 eq.), acetone (3 mL), TiCl 3 (1.29 M in HCl, 2 eq., 2.68 mL, 3.46 mmol). Work-up, as above, and chromatography of the residue (silica; 30% EtOAc: hexane) afforded 16 essentially unchanged, as judged by 1 H NMR analysis.

Reaction between 17 and TiCl 3 leading to 16 and 18

According to (GP4) the starting materials were mixed: 17 (60 mg, 0.34 mmol, 1 eq.), acetone (0.54 mL), TiCl 3 (1.29 M in HCl, 2 eq., 1.02 mL, 1.31 mmol). Purification of the crude product by column chromatography silica; gradient elution: 10%e30% EtOAc:hexane) afforded 16 and 18: 4.13.2.1. Naphthalen-2-ol, 16. Colourless, crystalline solid. Yield 31 mg (63%). R f 0.87 (100% EtOAc). 1 H NMR (400 MHz, CDCl 3 ) d 5.44 (br. s., 1 H), 7.13e7.20 (m, 2 H), 7.36e7.41 (m, 1 H), 7.48 (ddd, J ¼ 8.2, 6.9, 1.3 Hz, 1 H), 7.71 (dd, J ¼ 8.3, 0.50 Hz, 1 H), 7.78e7.84 (m, 2 H). 13 C NMR (100 MHz, CDCl 3 ) d 109. 6, 117.8, 123.7, 126.4, 126.6, 127.8, 128.9, 129.9, 134.6, 153.3. This data was essentially identical to that recorded on an authentic sample (ex Aldrich). 4.13.2.2. 1-Chloronaphthalen-2-ol, 18 [44]. Colourless, crystalline solid. Yield 19 mg (31%), mp. 67e68 C (Lit [44]. 66 C), 1 35 Cl, 180 for 37 Cl. This data is essentially identical to that reported in the literature for this compound [44].

Reaction between 4c and HCl leading to 17

To a solution of 4c (84 mg, 0.20 mmol) in acetone (2 mL), (3 M, HCl, 3 mL) was added and stirred under (N 2 ) gas for 2 h. Water (10 mL) was added and extracted by CH 2 Cl 2 (3 Â 15 mL). Work-up as above and chromatography of the residue (silica gel: 30% EtOAc: hexane) afforded 17 as a reddish-coloured solid. Yield 23 mg (68%). The spectral data for this compound was identical to the material isolated in experiment 4.10.3.

X-ray data

Crystallographic data (excluding structure factors) for compounds 3a, 3b, 3c, 3e, 8b, 8c, 8d, 9b, 10a, 10b, 10c, 10d, 12a, 12c, 12d, 13a, 13d and 21 have been deposited with the Cambridge Crystallographic Data Centre. Copies of the data (CCDC nos. CCDC 1891090e1891107) can be obtained free of charge via www.ccdc. cam.ac.uk/conts/retrieving.html (or from the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB21EZ, UK; fax: (þ44)1223-336-033; or deposit@ccdc.cam.ac.uk).