The reductive decyanation reaction: an overview and recent developments

This review presents an overview of the reductive decyanation reaction with a special interest for recent developments. This transformation allows synthetic chemists to take advantages of the nitrile functional group before its removal. Mechanistic details and applications to organic synthesis are provided.


Introduction
Many strategies in organic synthesis involve the removal of a beneficial functional group. The electron-withdrawing properties of the nitrile functional group appear beneficial in a variety of reactions [1,2]. This group coordinates metal complexes and can be used as a directing group for C-H bond activation reactions catalyzed by transition metals [3][4][5][6]. The α-deprotonation of alkylnitriles generates active α-cyano carbanion nucleophiles. Recent investigations have resulted in different modes of alkylnitrile activations and in the development of new catalytic cyanoalkylation methodologies [7]. While Fleming and Zhang first focused on the removal of the cyano group from cyclic substrates [8], in 2006 we published a review reporting various methods allowing the reductive decyanation reaction that transforms organic nitriles into the parent alkane [9]. Even if chemical procedures previously described are still of relevance in organic synthesis, it is noteworthy that new methods have now emerged with the aim to develop mild reaction conditions that allow reduction of a wide scope of substrates and tolerate a variety of functional groups. This review attempts to be complementary to the paper published in 2006 and proposes an overview of the reductive decyanation reaction that focuses on more modern methods.

Review Alkali-metal-promoted decyanation
Since the article by Arapakos in 1967 [10], decyanations using alkali metal dissolving conditions, typically Li or Na/NH 3 , are widely used in organic synthesis [11][12][13][14][15]. The mechanism proposed involves an electron transfer with formation of a radical anion intermediate, and then a cyanide anion is eliminated with Scheme 1: Mechanism for the reduction under metal dissolving conditions. Scheme 2: Example of decyanation in metal dissolving conditions coupled with deprotection [30]. TBDMS = tert-butyldimethylsilyl.
Scheme 3: Preparation of α,ω-dienes [18,33]. concomitant formation of a radical. The latter is then reduced to a carbanion which can be in situ protonated by ammonia or, depending on the conditions used, another proton source (Scheme 1).
Using Na/NH 3 or Li/EtNH 2 solutions, Arapakos obtained the best yields for the decyanation of phenyl-substituted acetonitriles, tertiary alkyl, and aromatic nitriles. However, primary and secondary nitriles also led to the reduction to the amine [10,16]. This drawback can be overcome using K/HMPA/ t-BuOH [17,18] or K/dicyclohexano-18-crown-6/toluene [19,20]. In the latter case, the toluene radical anion is believed to be the reactive species. LiDBB (lithium di-tert-butylbiphenylide) and Li naphthalenide are also common electron donors [15,21,22]. Because of the mechanism described in Scheme 1, the nature of the medium and the substrate strongly influence the course of the reaction. Then, in the absence of a proton source, the organolithium intermediate can cyclize or react with an electrophile giving the expected coupling products [23][24][25][26][27]. Metal dissolving conditions allow the reduction of various other functional groups [28]. Rychnovsky took advantage of this reactivity and achieved reductive decyanations with concomitant Birch reduction or benzyl ether cleavage [29][30][31]. An example related to the synthesis of polyene macrolides is described in Scheme 2.
Rojas et al. proposed a convenient two-step pathway for the preparation of alkyl α,ω-dienes 3. These dienes are well-known precursors in ring-closing metathesis (RCM) and acyclic diene metathesis (ADMET) chemistry [32]. They first reported the quantitative α-alkylation of primary nitriles 1 [33]. In a second part of their work, they developed adequate conditions to carry out the decyanation reaction without olefin isomerization [18]. They explored several methods for the preparation of 12-butyltricosa-1,22-diene 3 (R = n-C 4 H 9 ). The reaction was carried out in a slurry of K/Al 2 O 3 in hexane, hexane/toluene (1:1) and toluene giving 20%, 63% and 75% of olefin isomerization (from NMR and GC), respectively, for each solvent [34]. This isomerization was attributed to the translocation of the tertiary radical intermediate to a more stable allyl radical leading to the double bond migration. This rearrangement was avoided using K/ Ph 3 CH in hexane/ether (3 R = n-C 4 H 9 , 41% yield) or K/HMPA/ t-BuOH in ether (3 R = n-C 4 H 9 , 99% yield). The latter optimized conditions allow the decyanation of alkylcyano α,ωdienes 2 in quantitative yields with no detection of olefin isomerization (Scheme 3).
Radical intermediates were trapped using the radical probe 4 (Scheme 4). Replacing t-BuOH with t-BuOD in Scheme 3 (R = CH 3 case), yielded the decyanation product 3 with 92% deuterium incorporation. With respect to the mechanism de-Scheme 4: Cyclization reaction using a radical probe [18].

Scheme 5:
Synthesis of (±)-xanthorrhizol (8) [39]. scribed in Scheme 1, the authors suggest that t-BuOH (or Ph 3 CH) could act as H-atom donor that quenches the radical intermediate. However, this interpretation is opened to discussion because olefin functionalities also can undergo isomerization via anionic intermediates [35] and radicals usually abstract hydrogen atoms preferentially from the alkyl groups of t-BuOH [36][37][38]. Therefore t-BuOH could act as a proton donor and so prevent the olefin isomerization.
Alkali metals can also be used in suspension. As mentioned above, highly dispersed potassium over neutral alumina (K/Al 2 O 3 ) in hexane is able to effect the reductive cleavage of alkylnitriles [18,34]. Zárraga et al. described an efficient synthesis of (±)-xanthorrhizol (8) [39]. The authors prepared the intermediate 7 by dialkylation of 6 and attempted to carry out a one-pot decyanation and demethylation [40] with a suspension of lithium in THF. The target compound 8 was obtained in 74% yield together with 24% of the byproduct 9 (Scheme 5). This compound seems to be formed by a cross-linked ether cleavage of the methoxy group induced by the anion intermediate resulting from the decyanation pathway (Scheme 1). The authors increased the yield to 99% by adding NH 4 Cl as proton source that immediately reacts with the anion before the ether cleavage.
Aluminium-and borohydrides and the use of sodium hydride
Chuang et al. prepared a set of α-aminoacrylonitriles 11 by a cyano-promoted aza-Diels-Alder cycloaddition [71]. The cyano groups were then removed in high yields by treatment with NaBH 4 in 2-propanol by using both basic and nucleophilic properties of the hydride ion. The proposed mechanism involves a double-bond isomerization to the α-aminonitrile intermediate which is then reduced by the hydride ion in a classical way (Scheme 7). Interestingly, deuterium-labelling experiments indicate that one of the methylene hydrogens of the formed allylamine 12 is derived from the protic solvent and the other comes from the reducing agent. Finally, an oxidative aryl-aryl coupling promoted by vanadium oxytrifluoride (VOF 3 ) afforded phenanthroindolizidines (13, n = 1) and phenanthroquinolizidines (13, n = 2). Anticancer activities of these 18 compounds were evaluated against three human cancer cell lines. BH 3 ·THF containing NaBH 4 has been used for the reduction of diimines [72,73] and was studied in-depth by Zhang and co-workers in the reductive decyanation reaction. In their work, the cyano group activates the [3 + 2] cycloaddition of azomethine ylides and is then removed to yield 5-unsubstituted pyrrolidines [74]. These substructures appear in several biologically active natural products and drugs [75]. A range of decyanation conditions were screened such as NaBH 4 in THF or MeOH, NaBH 4 /AgBF 4 in THF and NaBH 3 CN in MeOH/ AcOH. They also explored BH 3 alone in THF and with varying amounts of NaBH 4 . They found that the addition of a catalytic amount of NaBH 4 was very efficient for the reductive decyanation reaction. Scheme 8 shows the scope of this two-step transformation. This protocol is successfully used with various electron-deficient substituted phenyl groups (15b-e) and with heterocycles (15f-h). The double-bond of cyanopyrrolidine 14l is preserved from the hydroboration reaction (15l). Olefins bearing bulky ester, sulfone or amide groups afforded good to excellent yields (15i-k). A series of aliphatic α-iminonitriles also gave good results (15n,o). Notably the labile dimethylacetal group tolerates this two-step transformation (15m).
Using this protocol, the authors described a total synthesis of (±)-isoretronecanol (19), a pyrrolizidine alkaloid (Scheme 9). In this case the two-step protocol is followed by a lactamization and the one-pot reduction of ester and lactam groups of 18.
Scheme 8: Two-step synthesis of 5-unsubstituted pyrrolidines (25 examples and 1 synthetic application, see below). The two yields refer successively to the cycloaddition (14) and decyanation (15) steps [74]. Boc = tert-butoxycarbonyl. EWG = electron-withdrawing group. The authors proposed a reductive anionic chain mechanism described in Scheme 10. The exposure of 14a to BH 3 generates a borane-amine complex 20 whose fragmentation could be promoted by NaBH 4 . The resulting imine 21 is reduced by BH 3 with the help of the cyanoborohydride anion. The formed anion 22 abstracts a proton from complex 20 to produce 23 or 24 and regenerate 21 and BH 3 CN − . A set of experiments supports this proposal. Notably, borane is the major hydride source for the reduction and 22 (derived from the reaction of 23 and NaH) is efficient for this chain reaction. Stable intermediates 23 and 24 were fully characterized while the unstable complex 20 was only characterized with 1 H NMR.
Chiba et al. accidently discovered the reductive decyanation of aryl substituted tertiary nitriles (Scheme 11) [81]. The protocol involves NaH in THF in the presence of LiI at 85 °C and appears suitable for the construction of the important 1,1-diarylalkane (26k-m) and triarylmethane (26n) derivatives. Strained cyclobutylarenes (26g,h) and heterocycles (26i,j) are prepared in this way and the electron-rich 4-methoxyphenyl group decreases the reaction rate (compare 26e,g with 26f,h). The decyanation of nitriles with the NaH-NaI system gives comparable yields but much longer reaction times are generally required.
The reduction of radical probes (no rearranged products formed from 25d,l,m) and deuterium-labelling experiments (no deuterium incorporation using THF-d 8 and quenching with D 2 O) discard the possibility of a single-electron transfer pathway. Other reductions suggest a hydride addition with formation of an iminyl anion intermediate. Particularly, when the reaction of 25f was quenched after 2.5 h, the corresponding Scheme 12: Proposed mechanism for the reduction by NaH [81]. aldehyde was formed in 42% yield together with 37% of the decyanated product 26f.
DFT calculations conducted on nitrile 25o support the hydride addition to the CN triple bond with formation of an iminyl anion intermediate 27. The latter easily isomerizes to its isomer 28 where a sodium cation-π-interaction occurs. The last step involves a C-C bond cleavage and proton transfer with elimination of NaCN (Scheme 12). This proton transfer occurs with retention of configuration as experimentally observed. Indeed, the kinetic profiles show that the decyanation reactions include an induction period (0.5 h and 2 h, respectively, for NaH-LiI and NaH-NaI systems) suggesting the formation of a new inorganic composite. These materials consist of metal iodide interspersed with activated NaH resulting in an unique hydridedonor reactivity [82]. An addition-elimination mechanism has been previously proposed for the LiAlH 4 promoted decyanation of 2,2-diphenylpropionitrile and related nitriles. In such pathways, the phenyl groups probably favor the C-C bond cleavage by stabilizing the incipient negative charge on the carbon adjacent to the cyano group [76,77].

Transition-metal-catalyzed reductive decyanation Hydrogenation of α-aminonitriles
The decyanation of α-aminonitriles with hydrogen present in an excess of Raney nickel was described by Husson and co-workers on oxazolidine derivatives [83,84]. The authors logically proposed that the decyanation occurred via the reduction of an iminium ion intermediate. The hydrogenation of α-aminonitriles catalyzed by nickel nanoparticles results in reductive decyanation and yields 29-31. The colloid solution of nickel is prepared in situ via reduction of anhydrous NiCl 2 with NaBH 4 in t-BuOH, iPrOH or n-butanol, the reactions are then performed upon bubbling of hydrogen at atmospheric pressure through the reaction mixture (Scheme 13). The reductive decyanation of 2-hydroxyadamantane-2-carbonitrile is successful by this method and yields alcohol 32 [85].
For less reactive nitriles, the C-CN bond cleavage usually requires hydrogen under pressure and high temperatures [86]. Opatz et al. developed the enantioselective syntheses of various alkaloids using the rhodium catalyst developed by Noyori [88] for the asymmetric transfer hydrogenation of imines. Interestingly, imines are formed from unstable α-aminonitrile intermediates which spontaneously eliminate HCN [89][90][91].

Iron-catalyzed reductive decyanation
In 1982, Yamamoto et al. disclosed the C-CN bond cleavage promoted by an electron-rich cobalt complex [92]. Since that time, the activation of inert C-CN bonds by transition metals has been widely investigated. Improvements towards mild and Scheme 15: Simplified pathways involved in transition-metal-promoted reductive decyanations [93,95]. green conditions with a large substrate scope have been achieved in many reactions including decyanation. Several reviews account for the richness of this chemistry [93][94][95][96][97][98][99]. Two major pathways for the C-CN bond activation have emerged (Scheme 15). One is the oxidative addition of the C-CN bond to a low-valent metal center (Ni case). The other pathway involves a silylmetal-assisted carbon cyano bond cleavage through an iminoacyl intermediate (Rh, Fe cases). In both pathways, the combination with a reducing agent gives a catalytic reductive system for the removal of the cyano group.
The discussion below on transition-metal-catalyzed reactions focuses on selected examples that show the wide scope of reduced substrates including alkyl cyanides, challenging substrates due to their propensity to undergo β-hydride elimination from alkylmetal intermediates. Nakazawa et al. reported the photoinduced C-CN bond cleavage catalyzed by iron complexes of a few primary alkyl cyanides and aryl cyanides in the presence of Et 3 SiH [100,101]. Decyanated species were formed together with silyl cyanide. Both aliphatic and aromatic nitriles were successfully reduced (Scheme 16) but an electron-withdrawing, bulky or coordinating substituent on the C atom linked to the cyano group disfavors the reductive decyanation.

Rhodium-catalyzed reductive decyanation
Chatani and co-workers have investigated the rhodium-catalyzed carbon-cyano bond cleavage reactions using organosilicon reagents [94]. They reported a rhodium-catalyzed reductive decyanation with a hydrosilane as reducing agent. They selected [RhCl(cod)] 2 (cod = 1,5-cyclooctadiene) as catalyst and found that the use of triisopropylsilane and P(OBu) 3 or P(OiPr) 3 as ligands led to the best results [102,103]. Aryl, heteroaryl (Scheme 17), benzyl and alkyl cyanides (Scheme 18) are applicable to this decyanation reaction. The use of hydrosilane as a mild reducing agent allows a high functional group tolerance (34b-e, 35, 39b, 43) and sterically hindered cyano groups are successfully removed (36, 41, 43). Reactions of benzyl cyanides proceed smoothly even in the absence of the phosphite ligands (39b, 43). Interestingly, an ester group in the α-position increases the reactivity (43). The reaction works well with simple primary and secondary alkyl cyanides but requires a higher reaction temperature. Aliphatic nitriles containing β-hydrogen atoms are successfully reduced (40, 42a-c, 43, 44a). Even if alkenes are not formed, the authors show the occurrence of a β-hydride elimination/ hydrometalation sequence. However, the reduction of the more hindered substrate 44b requires a larger excess of hydrosilane to prevent the formation of alkene. Finally, unactivated tertiary alkyl cyanides lead to a complex mixture.

Nickel-catalyzed reductive decyanation
Maiti et al. established that the Ni(acac) 2 complex (acac = acetylacetonate) of PCy 3 (Cy, cyclohexyl) in combination with TMDS (tetramethyldisiloxane) as hydride source can catalyze the reductive decyanation efficiently (method A, Scheme 19) [104]. The use of AlMe 3 appeared beneficial by facilitating the oxidative addition of the C-CN bond to Ni. Soon after, the same team developed the use of cheaper, greener and milder hydrogen gas as the hydride source. They found that catalytic Ni(cod) 2 in combination with PCy 3 and AlMe 3 under 1 bar pressure of H 2 gas was the protocol of choice (method B) [105]. As well as the activation of the C-CN bond, AlMe 3 helps to remove and consume the HCN gas produced under the reaction conditions. These protocols were successfully applied to a wide range of nitriles containing various other functional groups (Scheme 19). The ether bonds (33b, 49), ester (45), hydroxy (39c) and keto groups (46) are tolerated. With method B, aryl cyanides with amide (54), trifluoromethyl (55) and fluoro groups (47c) are reduced in attractive yields. For substrates bearing dicyano groups, monodecyanation is the major pathway (47b) and decyanations are efficient with sterically demanding ortho-substituents (54,57) or with ortho-directing groups (45,48,50). Aliphatic nitriles underwent decyanation without β-hydride elimination (42a, 52b, 53).
Enthaler et al. reported the application of nickel complex 58 as precatalyst and tert-butylmagnesium chloride as reducing agent (method A) in the decyanation of alkyl and aryl cyanides [106]. In the same year, to easily generate nickel hydride complexes, the authors explored the possibility to apply more reactive hydride donors such as LiBH 4 (method B) [107,108]. Finally, after investigating the reaction conditions and a precatalyst screening, the conditions described in Scheme 20 were selected. With method B, lower catalyst loading and shorter reaction times are necessary. With method A, benzonitrile yields 51% of benzene (34a) and an increase in yield is obtained with the methoxy and methyl groups (34c, 59, 34j para). For some substrates, the replacement of t-BuMgCl by n-Bu 2 Mg improves the yield (34b,c, 59). With method B, excellent yields are obtained for methoxy (34c, 59) and methyl substituted aryl cyanides 34j while with method A, changing the position of the methyl substituent from the para to the ortho or meta positions decreases the yield in toluene. Dimethylamino and thioether groups display a lower yield with method B (34b,g) contrary to the tri-

Scheme 21:
Proposed catalytic cycle for the nickel complex 58 catalyzed decyanation (method A). Only the cycle for 63 is shown [106]. fluoromethyl group (34i). With p-fluorobenzonitrile the decyanation is observed with both methods (34h) but, with the bromo counterpart, dehalogenation also occurs. 2-Cyanothiophene results in quantitative yield of thiophene 60 with both methods but cyanopyridines are not converted, probably due to the coordination abilities of the pyridine. Finally, decanenitrile is decyanated with method A (61) but aliphatic nitriles are not converted to the desired product with method B.
Trapping of radicals with method A led the authors to propose the electron transfer mechanism described in Scheme 21. The complex 62 reacts with the Grignard reagent to form the nickelmagnesium hydride intermediate 64 via a β-hydride elimination from 63. A single-electron transfer (SET) to the nitrile oxidizes the complex at the metal center into 65 and generates an aryl radical. The electron can also be located in the ligand (noninnocent ligand, not represented in Scheme 21). Then, elimination of MgXCN and radical recombination with the nickel species produce 66. Finally, an elimination of Ar-H closes the catalytic cycle. A similar mechanism was proposed with method B.

Radical reactions
This part is related to radical reactions not involving alkali metals and focuses on the reduction of malononitriles and cyanoacetates (α-cyanoesters). These compounds are particularly versatile reagents in organic synthesis including multicompo-nent reactions [109][110][111][112][113]. However, like decarboxylation for related malonic ester or acetoacetates, an easy access to the removal of the cyano group should encourage future developments using these compounds.

Bu 3 SnH and N-heterocyclic carbene boranes
The reductive decyanation of malononitriles to mononitriles using tributyltin hydride/AIBN in benzene was unexpectedly discovered by Curran and Seong [114]. Later, they made a full study and successfully reduced to mononitriles a variety of mono-and dialkylated malononitriles but under these conditions, the reduction of cyanoacetates failed [115]. Synthetic applications of this methodology were later described [116][117][118]. Chiba et al. have developed a concise and stereoselective methodology for the preparation of highly substituted carbocycles [119,120]. An example in described in Scheme 22. The 5-membered ring is formed via a K 2 CO 3 mediated S N 2 -conjugate addition sequence between malononitrile and 6-bromo-2hexenoate derivatives 67a,b. Dicyanocyclopentanes 68a,b are treated with tributyltin hydride/AIBN giving the monodecyanated products 69a,b. Bicyclic lactones 70a,b are then obtained in 3 steps in 41% and 51% yields, respectively, from 69a,b.

Scheme 23:
Reductive decyanation of malononitriles and cyanoacetates using NHC-boryl radicals (9 examples). For 74 and 75, isolated or NMR yields are given from 71 [121]. dinitriles and applied this reaction for the synthesis of new NHC-boryl nitrile and dinitrile compounds [121]. They observed that malononitrile was the most efficient donor to produce boryl nitriles and concluded that substituted malononitriles would be decyanated by NHC-boranes. For this transformation, malononitriles are reacting with a slight excess of NHC-borane 71, in refluxing t-BuOH with DTBP (di-tertbutylperoxide) as radical initiator. The yields are attractive while roughly comparable amounts of boryl nitriles 74 and 75 are formed (Scheme 23). Malononitriles 72d,e are successfully reduced to 73d,e while with photoactivated 78 such substrates afforded complex mixtures (see below). Contrary to the reductive decyanation with Bu 3 SnH [115], the NHC borane reduces the α-cyanoester 72g to 73g while the aryl chloride and bromide moieties are preserved (73h,i).
The use of a 5-hexenyl radical probe led to the formation of 99% of cyclized products (relative yield). The authors proposed a radical chain mechanism similar to that proposed in the reaction with Bu 3 SnH. While the tin radical was proposed to add to nitrogen [115], here the evidence points more to the addition of the boryl radical 76 on the nitrile carbon to form the nitrogen centered radical 77. β-Fragmentation leads to NHC-boryl nitrile 74 and a carbon centered radical. A hydrogen atom transfer reaction between the electrophilic α-cyano radical and the nucleophile NHC-borane achieves the chain propagation (Scheme 24). The isolation of the NHC-boryl nitrile 74, EPR spectroscopy observations [122], and polar effects fit with this proposition.

Neutral organic electron donors
Powerful single-electron transfer reagents have been described. Kang et al. reported the decyanation of both malononitriles and α-cyanoesters using samarium(II) iodide/THF/HMPA at respectively 0 °C and room temperature [123]. Metallic samarium can also promote the decyanation [124].
Doni and Murphy have reported the reductive decyanation of malononitriles and α-cyanoesters by using the neutral organic electron donor 78 (Scheme 25) under photoactivation (method A, Scheme 26) [125]. The first observation was that surprisingly, the reaction of 2,2-dibenzylmalononitrile (72b) provided Scheme 24: Proposed mechanism for the reduction by NHC-boryl radicals. The other possible pathway (addition of 76 to the nitrogen) is not represented here [121].
both debenzylated (2-benzylmalononitrile, 19%) and decyanation (73b, 75%) products. In contrast, the corresponding dibenzylcyanoacetate led exclusively to the debenzylation product [126]. Selected examples are presented in Scheme 26. Excellent yields are obtained even if decyanation of cyanoacetates requires higher amounts of 78 (6 equiv) with extended reaction times (72 h). In line with this work, the authors prepared the tetra(iminophosphorano)-substituted bispyridinylidene 80, a new highly efficient neutral organic electron donor (Scheme 25) [127]. This compound is able to reduce aryl halides and appears as the only reductant able to reduce dialkylarenesulfonamides as well as malononitriles without photoexcitation. Electron donor 80 can be isolated but is more conveniently generated in situ by deprotonation with KHMDS of its pyridinium ion precursor 79 in refluxing toluene (method B).

Scheme 25:
Structures of organic electron-donors. Only the major Z isomer of 80 is shown [125,127].

Scheme 28:
Examples of decyanation promoted in acid or basic media [129,131,134,135].
With method B, malononitriles are reduced to mononitriles with comparable yields to that achieved with photoactivated 78 (73k,m,n). No allylic bond cleavage is observed and alkene moieties are preserved in the decyanated product (73m,n,q). No radicals are trapped from the reduction of the 5-hexenyl radical probe 72m with both methods. Mononitrile 2,2-dimethylhexadecanenitrile appears inert with both methods, and organic bromides are not tolerated. The proposed mechanism is similar to Scheme 1. First, a SET from the electron donor to the nitrile forms a radical anion. This radical anion can fragment into a cyanide ion and a radical which is rapidly reduced into a stabilized carbanion before protonation. This carbanion appears as a key intermediate leading to complex mixtures when method A is applied to 72d or a substrate similar to 72e.
More recently, the reaction of dibenzylmalononitrile 72b with the hydroxyaryl-substituted benzimidazoline derivative 81 as photo-reductant in a basic medium has been described and led to 73b in fair yield (Scheme 27) [128]. Acid, base or organometallic-induced reductive decyanation Among the other procedures mentioned in our previous review [9], the base [129,130] or acid-induced [131][132][133][134] hydro-lysis-decarboxylation sequence appears as a classical pathway (Scheme 28).
In the particular case of diphenylacetonitriles, an addition-elimination mechanism is proposed (Scheme 29) [135,136]. Such a pathway could also apply for the reductive decyanation of diphenylacetonitriles induced by organolithiums or Grignard reagents [137,138]. This reaction, applied to nitriles substituted with suitable leaving groups appears as a cyanation method of organometallic reagents or other nucleophiles [139]. Nambo et al described the preparation of triarylacetonitriles using sequential Pd-catalyzed arylations. Triarylacetonitriles obtained can be transformed into various species including triarylmethanes by treatment with MeMgCl (Scheme 30) [140]. In this case a SET mechanism could operate [141].

Conclusion
The reductive decyanation reaction appears useful in organic synthesis because it offers the possibility to temporarily use the advantages from the nitrile functional group. While our previous review showed that several methods had a narrow substrate scope [9], new and convenient synthetic methods have now emerged. Classical metal dissolving conditions are still used but the method, while it works, shows that the course of the reaction may strongly depend on reaction conditions. The transition-metal-catalyzed defunctionalization reactions cover a wide range of nitriles including the most challenging alkanenitriles and allow a large functional group tolerance. The decyanation of α-aminonitriles by aluminium-and borohydrides has been widely investigated and offers synthetic applications in heterocyclic chemistry. Recent developments using N-heterocyclic carbene boranes and super electron donors provide new procedures for the reduction of malononitriles or α-cyanoesters. This opens the possibility to synthetic applications by using these intermediates in the future.