Rapid Iododeboronation with and without Gold Catalysis: Application to Radiolabelling of Arenes

Abstract Radiopharmaceuticals that incorporate radioactive iodine in combination with single‐photon emission computed tomography imaging play a key role in nuclear medicine, with applications in drug development and disease diagnosis. Despite this importance, there are relatively few general methods for the incorporation of radioiodine into small molecules. This work reports a rapid air‐ and moisture‐stable ipso‐iododeboronation procedure that uses NIS in the non‐toxic, green solvent dimethyl carbonate. The fast reaction and mild conditions of the gold‐catalysed method led to the development of a highly efficient process for the radiolabelling of arenes, which constitutes the first example of an application of homogenous gold catalysis to selective radiosynthesis. This was exemplified by the efficient synthesis of radiolabelled meta‐[125I]iodobenzylguanidine, a radiopharmaceutical that is used for the imaging and therapy of human norepinephrine transporter‐expressing tumours.


Introduction
Aryl iodides are versatile key building blocks in organic synthesis and have found widespread applicationsi na reas such as cross-coupling reactions and the generation of free-radical intermediates. [1] In additiont ot heir applicationsi nsynthesis, aryl iodides are also found in naturalp roducts and pharmaceutically important compounds. [2,3] More recently,r adiolabelled aryl and heteroaryl iodides have been increasingly reported in medicala pplications; in particular,i ns ingle-photon emission computed tomography (SPECT) imaging for drugd evelopment and the clinical diagnosis of disease, and in targeted radionuclide therapy. [4] As ar esult of these applications,m uch effort has gone into developing methods for the efficient synthesis of aryl iodides. [5] One such method is the ipso-substitutiono fa rylboronic acids by using N-iodosuccinimide (NIS), which wasd eveloped by Olah. [6] However, one limitation of this uncatalysed method is the substrate scope:d eactivated arylboronic acids with electron-withdrawing substituents performed poorly,e ven after extended reactiont imes. For this reason, severalb ase-or phasetransfer-mediated [7] and Cu-catalysed [7c, 8, 9] ipso-iodination reactions of boronic acids have emergedi nr ecent years. Despite the much-improved substrates copes afforded by these recent developments, there are still limitations,s uch as the need for base/additives, environmentally damaging solvents, or long reactiont imes. [10] Therefore, improvement is still required, especially if this transformation is to find widespread utility in fields such as total synthesis, material science, and medical imaging.
Our interest in this area arose from our previousw ork on gold catalysis; [11] in particular, the mild gold-catalysed [12] protoand deuterodeboronations [13] and cross-couplings (Scheme 1A(a) and (b), respectively). [14] We hypothesisedt hat the organogold intermediate I [15] should reactw ith NIS to rapidly form iodoarenes 3.T herefore,o ur aims were to 1) compare the gold-catalysed andu ncatalysed reactions to ascertain whether gold catalysis can be used to overcome some of the substrate scopel imitations of the uncatalysed reaction, 2) develop a much faster iodination protocol, both catalysed and uncatalysed,3 )utilise environmentally friendly solvents to significantly improvet he practicality of the reaction( the original reaction times were 1.5-25 hi na cetonitrile), [6] and 4) to apply the methodology to radioiododeboronations.
Although there are early isolatedr eports of unselective gold-mediated radioiodinations [16] and am ore recent elegant use of stoichiometricg old substrates for radiofluorinations, [17] as far as we are aware,t here are no examples of selective radioiodinations [5] or radiofluorinations [18] that use homogenous gold catalysis. Therefore, one of the main aims of this work was to demonstrate the first such application of gold catalysis to radiosynthesis. To the besto fo ur knowledge,g old catalysis has never been used for iododeboronations, even in cold (nonradiolabelled) procedures. [19] However,t he groups of Wang and Frontier have reported mechanistically distinct iodinations of arenest hrough the gold-catalysed activation of NIS for electrophilic aromatic substitutions. [20,21] Therefore, another challenge was to developt he intended ipso-iododeboronation in as elective manner and avoid any over-iodination, which would be caused by electrophilic aromatic substitutions.
Herein,w er eport ar apid ipso-substitution of arylboronic acids by using NIS in the environmentally friendly solventd imethyl carbonate (DMC) [22] under microwaveh eating [23] or thermalh eating (Scheme1B). Both the gold-catalysed andu ncatalysed reactions were investigated concurrently,a nd we demonstrated that Au I catalysis can be used to greatlyimprove the yields in cases where the uncatalysed reaction is poor.Conversely, the uncatalysed reaction is often successful in cases where gold catalysis fails, so both protocols complemente ach other nicely.W ea lso describe the use of this mild and general transformation for the radioiodination of arenes( Scheme 1C). Incorporation of the 125 Ir adioisotope was fast and highly efficient, which allowed the preparation of the radiopharmaceutical meta-[ 125 I]iodobenzylguanidine in high molar activity and excellent radiochemical purity.

Results and Discussion
We initiated our studies by carrying out as olvent screen on the gold-catalysedi ododeboronation of arylboronic acid 1a (Table 1). Pleasingly,t he expected iododeboronationp roduct 3a was observeda st he major product in various solvents, althought he over-iodination product 4a,w hich resulted from electrophilic aromatic substitution of gold-activated NIS, [20a] was also observed as am inor product. In addition, the proto-deboronation product 2a was also observed when the reaction was carriedo ut in toluene (entry 4), and it is unsurprisingly the main product of the reaction performed in water (entry 8). The two best results were obtained in chloroform (entry 2) and dimethyl carbonate (entry 3);t herefore, dimethyl carbonate was taken forward for optimisation owing to its "green" credentials. [22] With an eco-friendly solvent in hand, our next aim was to significantly reduce the reaction times. To this end, the goldcatalysed iododeboronationo fa rylboronic acid 1b was investigated by using various times (2-5 min), temperatures (70-100 8C), and equivalents of NIS (Table 2). Higher temperatures (entries 1v s. 2) and al ower equivalent of NIS (entries3 vs. 5) reduced the amount of unwanted over-iodination product 4b. The conditions shown in entry 6( 90 8C, 5min, 1.0equiv of NIS) produced the best compromise between yield and 3b/4b selectivity,a nd therefore, theyw ere taken forward as the optimal conditions.   It is worth noting that the reaction is not sensitive to air or moisture, and thus, it is av ery practical, as well as fast, procedure. In fact, wet solvente nhanced the yields for the gold-catalysed reaction (see the Supporting Information). [13] Next, the substrate scope of the gold-catalysed and uncatalysed reactions were investigated under these optimised conditions (Table 3). Both the catalysed and uncatalysed reactions were run concurrently for all substrates to ascertain whether gold catalysis offered anya dvantage over the uncatalysed reaction or vice versa. For strongly electron-richa rylboronic acids 1a-b,t he catalysed reactiong ave ab etter combined yield but often poorer selectivity for the desired ipso-iodination (3) versuso ver-iodination (4)p roducts (18:1 and 12:1 forc atalysed reactiono f1a and b,r espectively,v s. > 20:1 for uncatalysed). Nevertheless, we observed that thermal heatingc an be used to improve the selectivity in the catalysed reaction (3b/ 4b, > 20:1); furthermore, over-iodination was not ap roblem with less electron-rich substrates (1c-l). This trend was expected, as very electron-rich aryl substrates are more likely to undergo competitive electrophilic aromatic substitution than the less electron-rich counterparts. With mildly electron-rich arylboronic acid 1d,t he gold-catalysedr eaction was significantly more efficient (100 vs. 62 %) andc lean (inseparable side-product observed in uncatalysed reaction). However, the presence of aphenolic proton in the substrate 1e caused both catalysed and uncatalysed reactions to produce as ignificant amounto f ap rotodeboronateds ide product 2e. [24] However,T HP-protected phenol substrate 1f was tolerated in the uncatalysed reaction, and the acid-sensitive THP group remained intact during the iododeboronation reactiont oy ield product 3f in 60 % yield. Therefore, for electron-rich boronic acids, the gold-catalysed reaction generally provided higher yields, but the lower yieldingu ncatalysed reaction may arguably still be preferred owing to itscheaper cost.
However,adifferent pattern emerged for electron-poor and sterically demandinga rylboronic acid substrates (Table 3, 1gl). These deactivated speciesr eactede xtremelys luggishlyi n the uncatalysed reaction and provided poor yields, even after extended reactiont imes. To our delight, gold catalysis provided as ignificant improvement in the yields as well as reaction times. For example, aryl iodides 3g, h,a nd i were formed in 71, 78, and 62 %y ields, respectively,a fter five minutes in the presence of PPh 3 AuNTf 2 ,w hereas only 33, 25, and1 3% yields were obtained without the catalyst. The comparison was even starker with the nitro-substituted substrate 1j:p roduct 3j was furnishedi n7 8% yield after ten minutes under gold catalysis, whereas the uncatalysed reaction provided av ery poor 10 % yield. Extending the reaction time to one hour did not significantly improve the yield of product 3j in the absence of the catalyst( 16 %). The free carboxylic acid 1k wasn ot tolerated under the reaction conditions;i nf act, substrates that bear acidic protons (e.g., 1e, k)were generally observed to be detrimentalt ot he reactions uccess. The acetylf unctionality in substrate 1l was tolerated in the uncatalysed reaction, although the desired product 3l was yieldedi namodest4 2% after 30 min. In contrast, the faster gold-catalysed reaction yielded a 1:1r atio of the desired product 3l and the undesired a-iodination product. The fluoro-substituted arylboronic acid 1m reacted sluggishly and produced as ignificant amount of the homocouplingp roduct (4,4'-difluoro-1,1'-biphenyl) in both the goldcatalysed and uncatalysed reactions.P leasingly,t he extremely sterically demanding substrate 1n underwent iododeboronation smoothly in the presence of the gold catalyst to give the desired product 3n in ag ood yield of 78 %( vs. 14 %u ncatalysed yield), albeit with al ongerr eaction time of three hours to account for the steric hindrance. For electron-poorand sterically hindereda rylboronic acids, the uncatalysed reactionw as generally low yielding and extremelys luggish, and gold catalysis could be used to significantly improvet he reactiont imes and yields.
The opposite pattern emerged for heterocyclic and N-containing arylboronic acid substrates (Table 3, 1o-r). Both heterocyclic boronic acids 1o and p produced ac omplex mixture of products under gold catalysis, but they iododeboronated cleanly under uncatalysed conditions to give 51 and 78 %o f the desired aryl iodides, respectively.S imilarly,a mine-substituted arylboronic acids 1q and r provided ac omplex mixture under gold catalysis, but they were successfully iododeboronated under uncatalysed conditions to give 3q and r in 100 and 32 %y ield, respectively.T he complex mixture that resulted from gold catalysis is most likely due to over-iodination [20a] of these highly electron-rich aryl substrates. Unsurprisingly,t he more electron-rich amine in 1t also resulted in ac omplexm ixture, but this time, under both the catalysed and uncatalysed procedures. The amide functionality in substrate 1s seemed to shut the reaction down under both conditions. Although there are somel imitations to the type of N-substituents that are tolerated, the uncatalysed procedure wasg enerally preferred for heterocyclic and N-substituted arylboronica cids.
As our studies demonstrated that gold-catalysed ipso-substitution of arylboronic acids by using NIS could be performed rapidly and efficientlyu nder mild conditions, we werek een to investigate the applicationo ft his methodi nto the radioiodination of arenes. Radioactive iodine is normally supplied in the form of NaI;t herefore, initial studies involved the iodination of 4-methoxybenzeneboronic acid (1b)w ith NIS, which was prepared in situ by pre-stirring NaI and N-chlorosuccinimide (NCS). [25] Under our standard gold-catalysed reaction conditions, the desired product 3b was formed in 80 %y ield. These general conditions were investigated for the radioiodination of boronic acid 1b,i nw hich radiochemical yields (RCY) were determined by radio-HPLCa nalysis of the crude product. [26] At a micromolar scale, the gold-catalysed ipso-substitution reaction was easily modified for radioiodination.[ 125 I]NaI (4-6 MBq solution in water) was used as the limitingr eagent, and the reaction gave the desired radiolabelledp roduct 125 I-3b in 63 %R CY (Table 4, entry 1). Extendingt he reaction time to 20 min was found to be optimal and gave aq uantitative RCY of product 125 I-3b (entry 3). The corresponding radio-HPLCf or this transformation showedaparticularly clean reactionw ith no other radiolabelled by-products ( Figure 1). It should be noted that radioiodinationc an be done without the use of PPh 3 AuNTf 2 .N otably,i np arallelw ith the cold studies, the thermally-mediated reactionw as less efficient and gave only 47 %R CY after 20 min (entry 4).
In radioiodination reactions, ar apid transformation that gives the product cleanly is crucial for generating the target radiolabelled compound in high radiochemicalp urity,r adioactive yield, and molar activity.T herefore, the use of gold catalysis for these radioiodinations rather than the thermally mediated reaction is particularly advantageous because the reactions are faster,m ore efficient, and easier to purify.
The scope of the optimisedg old-mediated radioiodination reactionw as next examined forar ange of arylboronic acids (Table 5). Undert hese conditions, electron-rich and electrondeficient arenes with various substitution patterns were found to be suitable substrates for the reaction, and they reacted to give the 125 I-labelled products in excellent RCY (92-100 %). Only two arylboronic acids required furthero ptimisation: The initial reactions of ethyl ester and trifluoromethyla nalogues 1h and i under the optimised conditions (90 8C, 20 min) produced radio-HPLC traces with severalr adiolabelled by-products. Repeating the reactions at al ower temperature of 80 8C allowed ac leanert ransformation, and despite the need for a slightly longer reactiont ime (30 min), this gave the products 125 I-3h and i in excellent RCY.O nly sterically hinderedb oronic acid 1n showed no reactionu nder these conditions, even after 30 min. As observed for the cold iododeborination of this compound, as ignificantly longer reaction time was likely required for this substrate.  Having shown that this transformation could be used for the radioiododeboronation of simple arylboronic acids, we next investigated the application of this method for radiolabelling biologically active compounds and imaging agents ( Table 5). The first target was meta-[ 125 I]iodobenzylguanidine (MIBG, 5). In the 123 I-form, MIBG is ac ommercially available radiopharmaceuticalt hat is used for the SPECT imaging of humann orepinephrinet ransporter-expressing cancers. [27] In the 131 I-form, MIBG is used for targeted radionuclide therapy. [28] Adi-Boc-protected boronica cid analogue of MIBG was efficiently prepared in one step by the coupling of 3-(aminomethyl)benzeneboronic acid with di-Boc-protected 1H-pyrazole-1-carboxamidine. [29] After someo ptimisation, gold-mediated ipso-substitution of this arylboronic acid was found to be most effective at 80 8C with ar eaction time of 30 min. This produced the corresponding 125 I-labelled compound in quantitative RCY.T he reaction mixture was then treated with hydrochlorica cid to removet he Boc protecting groups,a nd this gave [ 125 I]MIBG (5)i n9 1% RCY over the two steps. Gold-mediated radioiododeboronation was also effective for the preparation of phthalazinone 6,ananomolar inhibitora nd SPECTi maging agent of poly(ADP-ribose) polymerase-1 (PARP-1), which is ad iagnostic and therapeutic target for cancer. [30] Despite the complexs tructure of this sub-strate,w hich contained the amide and N-heterocycle moieties, gold-mediated radioiododeborination was achieved by using a reactiont emperature of 100 8Ca nd gave 125 I-6 in 41 %R CY. [31] Following these results, we decided to validate the radioiododeboronation methodw ith the synthesis and purification of [ 125 I]MIBG (Scheme 2). Boronic acid 7 was treated with [ 125 I]NaI (10.16 MBq) by using our previously optimised goldmediated ipso-radioiodination reaction. Following the removal of the Boc protecting groupsu nder acidic conditions and HPLC purification,[ 125 I]MIBG (5)w as isolated in 28 %r adioactivity yield. Compound 5 had ar adiochemicalp urity of > 98 % and am olar activity of > 2.73 GBq mmol À1 .I dentificationo ft he product was confirmed by HPLC,w ith co-elution of as ample of unlabelled MIBG. These results compared favourably with other approaches for the preparation of radiolabelled MIBG, which demonstrates the potentialo ft his methodology for widespreadu se in generating radioiodine-labelledt racers. [9a,b, 27a, 32] Conclusions Fast, air and water stable ipso-iododeboronation reactions in the presence of NIS and the green solvent dimethylcarbonate have been developed. The gold-catalysed reaction was significantly preferred for electron-deficient and sterically hindered arylboronic acid substrates, whereas the uncatalysed reaction provided very poor yields. For electron-richb oronic acids,t he gold-catalysed reactiona lso provided much higher yields than the uncatalysed procedure. However,h eterocyclic and N-containinga rylboronic acids reacted more favourably under uncatalysed conditions. Moreover,t he reaction wast olerant of halogen moieties (Br and Cl), which is complementary to the commonly used halogen-exchange method to form aryl iodides. [33] The gold-catalysed iododeboronation reaction was shown to be ah ighly amenable procedure for the 125 I-labelling of arenes, which constitutes the first example of an application of homogenous gold catalysis to selectiver adiosynthesis. Under the optimised radiochemistry conditions, both electron-rich and electron-poor aryl boronic acids were rapidly convertedt ot he radioiodinated products in excellent RCY.T his general method was validated with the efficient synthesis and isolation of radioiodinated MIBG, at racer that is used for the imaging of cancer.C urrents tudies are ongoing to investigate the extension of this methodology for the preparation of existing and novel SPECT imaging agents.
General procedure B: uncatalysed reaction Boronic acid 1 (0.10 mmol, 1.0 equiv), NIS (0.10 mmol, 1.0 equiv), and DMC (0.4 mL) were added to am icrowave tube and heated under microwave irradiation at 90 8Cf or 5min. The resulting solution was passed through asilica plug and washed with hexanes/diethyl ether (20:1) to yield product 3.T he crude product was purified by column chromatography as needed.