Trifluoromethylation of arylboronic acids by Cu-mediated CF3 radicals

The incorporation of fluorine atoms into pharmaceutical molecules has long been in the medicinal chemist’s book of tricks to improve metabolic stability. The number of recent publications focusing on development of new methodologies for novel fluorination  and trifluoromethyl reactions only reflect the importance of such transformations.

Despite significant progress in the development of trifluoromethylation methodologies, and more recently with the Cu-catalysed trifluoromethylation of boronic acids, many transformations are still limited by either the cost of reagents, high temperature, strong acids / bases or limited scope.

Following on Melanie Sanford’s excellent JACS communication earlier this year, Melanie is back with a further publication addressing the scope and methodology of the trifluoromethylation of arylboronic acids with CF3 radicals. Key advantages to the publication is the generation of a practical source of CF3 radicals under mild conditions.

The CF3 radicals are generated in situ from of readily available NaSO2CF3 (Langlois’ reagent) and tert-butyl hydroperoxide (TBHP) and can undergo Cu-mediated trifluoromethylation with a range of boronic acids.

After an extensive evaluation of copper salts, CuCl proved to work well on reactions with no protodeborylation observed under the mild reaction conditions (Figure 2)

Figure 2

Both electron-rich and electron-neutral boronic acids are tolerated under the conditions described in the publication. However, Melanie reports that electron-deficient boronics acids led to moderate to poor yields. Further Cu screening for electron-deficient boronic acids led to the substitution of CuCl for (MeCN)4CuPF6 and addition of NaHCO3 (acceleration of the transmetalation rate) resulting in an improved yield (Figure 3).

Figure 3

From the scope of the reaction, it is worth noting that sterically hindered boronic acids are tolerated and the reaction compatible with a range of functional groups (esters, ketones, phenols).

Overall, a very mild and general approach for the Cu-catalysed trifluoromethylation of arylboronic acids, demonstrating the feasibility of a radical pathway.

Decarboxylative Esterification of Carboxylic Acids with Dialkyl Dicarbonates

To continue on from the esterification theme of our recent blog (Palladium catalyzed oxidative esterifications of aldehydes and benzylic alcohols) I would like to highlight 2 papers on the subject of decarboxylative esterification of carboxylic acids from professor Gooßen’s group at the Max-Planck-Institut für Kohlenforschung.

Gooßen’s first paper (Adv. Synth. Catal., 2005, 345, 943-947) describes the exploration of conditions for the optimisation of Lewis acid catalysed decarboxylative esterification of carboxylic acids with dialkyl dicarbonates (Scheme 1).

Scheme 1

Gooßen showed that in the presence of Bronsted acids almost no ester formation was observed and that DMAP only yielded a modest catalytic activity using the conditions previously described by Takeda. In contrast to these results it was shown that the all the Lewis acids investigated gave the desired ester with a catalyst loading of only 1 mol % (Table 1).

Table 1

Gooßen’s selected nitromethane as the solvent to further explore the generality of the chemistry, due to the carboxylic acids being readily soluble in it, but the reaction works equally well in methanol, dichloromethane or acetonitrile.

Both magnesium perchlorate and copper triflate are mild enough Lewis acids that transesterification of the sensitive phenol ester did not occur (table 1, entries 25 and 26) but they were both still able to convert the carboxylic acid into a methyl ester in a quantitative yield. Based on these results coupled with both environmental and economical considerations it was decided that magnesium perchlorate was the superior choice of Lewis acids for this reaction.

With these variables fixed the scope of this reaction was further explored with a range of carboxylic acid and dialkyl dicarbonates (table 2). All of the reactions produced the desired esters in high yield except for amino ester 8c. Both the methyl ester 8a and benzyl ester 8b were produced as expected but the t-butyl ester was unstable under the reaction conditions, producing isobutene and the carboxylic acid starting material.

Table 2

Gooßen believes that the reaction proceeds via an intramolecular decarboxylation due to methyl ester 18a (table 2) being formed in a high yield (92%). If only an intermolecular reaction of the mixed anhydride occurred then a statistical mixture of the methyl ester, lactone and oligomeric esters should be formed.  In an attempt to confirm this hypothesis Gooßen reacted phenylpropionic acid with dimethyl dicarbonate and Mg(ClO4)2 in the presence of 1 equivalent of n-butanol. The desired methyl ester formed (81%) but also a quantity of the n-butyl ester (19%).  Gooßen is investigating the mechanism of this reaction and will report his finding in due course.

In Gooßen’s second paper (Synlett 2004, 263–266) he describes a convenient protocol for the esterification of carboxylic acids with alcohols in the presence of di-t-butyl dicarbonate (Boc2O). By using a stoichiometric amount of a carboxylic acid, a primary or secondary alkyl alcohol and Boc2O, in the presence of a catalytic amount of N,N-dimethylaminopyridine (DMAP), gave high yields of the corresponding esters (Scheme 2).

Scheme 2

Gooßen investigated a range of catalysts (Hf(OTf)4, Mg(ClO4)2, Sc(OTf)3, Yb(OTf)3 and DMAP) as well as a selection of solvents (nitromethane, THF, toluene, DMF and MeCN) before arriving at his standard conditions (1.0 mmol 3-phenylpropionic acid, 1.3 mmol Boc2O, 1.3 mmol benzyl alcohol, 5 % DMAP, 4 mL nitromethane, 50°C, 16 h). It was observed that the reaction gave the same yield regardless of the solvent used to run the reaction in.

Scheme 3

Gooßen exemplified this procedure by using a selection of alcohols that he coupled with 3-phenylpropionates (Scheme3, Table 3). The reaction produced the desired products in high yields with both primary and secondary alcohols. Groups including cyano, keto, ester and less reactive alcohols were tolerated.

Table 3

Gooßen then chose a selection of carboxylic acids that he then coupled to benzyl alcohol (Table 4). The reaction was able to be performed on aryl, alkyl, vinyl and heterocyclic carboxylic acids which contain ester, ketone, amide, ether and nitrile functional groups. Unfortunately the reaction failed to yield the desired product and resulted in decomposition of the starting material for o-fluorobenzoic acid and m-NO2-benzoic acid.

This procedure presents an advantage over the use of alkylating reagents such as alkyl halides for esterifications due to the relative costs of reagents and the larger range of alcohols that are available.

Table 4

In summary both papers offer alternative approaches for synthesising esters from carboxylic acids. The reactions are mild enough to tolerate a large range of functional groups and use cheap, readily available starting materials. The work up of these reactions is also simplified by the fact that most of the by-products are either volatile or water-soluble and this make these methods very practical for a synthetic chemist to use.

Modelling CNS with inducible pluripotent stem cells

The challenges of accurately modelling diseases in vitro are great. A prime example is found in the central nervous system (CNS), where the complexities of the multicellular systems increase the difficulty in obtaining primary cells, especially those from patients. The use of stem-cell derived cultures is changing this, enabling neurons to be cultured from progenitor cells. More recently it has been possible to obtain human induced pluripotent stem cells (iPSC) derived from patients, effectively opening up a new source of patient cells. A number of publications have come out using these cells, and a recent double-publication from a group in Cambridge looking at Alzheimer’s disease illustrate not only the technical process, but how well the disease state can be modelled using this technology.

The first, published in Nature Neuroscience (2), describes the multi-step differentiation of human cerebral cortical neurons from pluripotent stem cells. The authors describe how they developed a multistep process for the differentiation of human cortical cells from embryonic stem (ES) and induced pluripotent stem cells (iPSC). The publication goes into some depth regarding the different phases these cells undergo during this differentiation, but critically after two months, these cells form multicellular cultures including both cortical neurons as well as astrocytes. These neurons form glutamatergic synapses that contain NMDA and AMPA receptors and whole cells patch clamp demonstrating? they have mature electrical properties and form functional excitatory synapses.

As though it wasn’t enough to publish the technical achievement differentiating? a cortical neuronal culture from iPSC and ES, the group went on to publish a follow-up in Science Translational Medicine (1), immediately demonstrating the application of this technology. They used the iPSC derived from Down syndrome patients from early Alzheimer’s pathology to model the neuronal cortex of a patient with Alzheimer’s disease.

The classic pathological hallmarks of Alzheimer’s disease are amyloid plaques composed of the amyloid Aβ (Aβ) peptide, (formed from the amyloid precursor protein) and neurofibrillary tangles comprising hyperphosphorylated forms of Tau. The amyloid precursor protein (APP) gene is encoded on chromosome 21 and therefore patients with Down syndrome (caused by trisomy of chromosome 21) develop early onset Alzheimer’s, commonly by 35 years of age.

The group took iPSC from a patient with Down syndrome (DS) and using the method set out in the neuroscience paper differentiated them into cortical neurons in the same way as they did in the first paper with control iPSCs and ES.  The group proceeded to measure whether there was Alzheimer’s pathology in these cultures.

They measured levels of Aβ40 and Aβ42 peptide production from 2-4 weeks after the onset of neuronal differentiation. In DS cultures, the levels of both Aβ40 and Aβ42 were significantly higher than in control iPSC.  In the DS cultures, these Aβ peptides also form amyloid aggregates that are not visible in the control cultures. They also measured the abundance of Tau phosphorylated at Ser202 and Thr205. This demonstrated that although phosphorylated Tau was expressed in both DS and control neurons, in control cultures phospho Tau was diffusely localised in the primary axons whereas in DS neurons it was aberrantly localised into linear foci in the cell bodies and dendrites in the DS cells. This matched the distribution found within the patient CNS.

The recent failure of so many late-stage Alzheimer’s drugs highlights the need to bring new compounds through the pipeline. Being able to elegantly model such a complex disease is perfect for a drug screening program, or phenotypic screen, to either hit Aβ or phosphorylated Tau. However, the screen can also be used for basic science to monitor disease progression potentially unravelling new targets, e.g. to elucidate the link between Aβ aggregation and hyperphosphorylated Tau. The development of such an elegant in vitro culture system such as this makes all of these more possible.


1.           Shi Y, Kirwan P, Smith J, MacLean G, Orkin SH, Livesey FJ. A human stem cell model of early Alzheimer’s disease pathology in Down syndrome. Science translational medicine 4: 124ra29, 2012.

2.           Shi Y, Kirwan P, Smith J, Robinson HPC, Livesey FJ. Human cerebral cortex development from pluripotent stem cells to functional excitatory synapses. Nature neuroscience 15: 477–86, S1, 2012.

Eph in ALS

A recent publication from Van Hoecke and colleagues (Van Hoecke et al., 2012) suggests a novel therapeutic approach to the treatment of amyotrophic lateral sclerosis (ALS; also called Lou Gehrig’s disease). ALS is a progressive degenerative disorder that affects 1-2/100,000 people per year and results in death, normally by respiratory failure, 3-5 years after onset. It is caused by a loss of the motor neurons that control muscle movement. There is a hereditary component in about 10% of all ALS cases and in these familial ALS subjects, a variety of genes have been implicated (Al Chalabi et al., 2012), including SOD1, TARDBP and FUS and which encode superoxide dismutase 1 (SOD-1; which was the initial gene reported to be associated with familial ALS in 1993) TAR DNA binding protein (TDP-43) and the Fused in Sarcoma protein, respectively. Recently, a hexanucleotide repeat expansions that occurs in the chromosome 9 open-reading frame 72 gene (C9ORF72) has been described to be associated with ALS with frontotemporal dementia (DeJesus-Hernandez et al, 2011; Renton et al, 2011).

The huge advances in our understanding of the genetics underlying the familial form of ALS have yet to result in breakthrough therapies for this disorder and Riluzole remains the only FDA-approved treatment for ALS. It was approved in 1995 on the basis of clinical studies that demonstrated that it increased survival times in patients, yet the effects are relatively modest and there is a clear need for new and improved treatments for ALS. Since Riluzole was approved, there have been over 30 clinical trials of new treatments but for a variety of reasons (including poor clinical trial design and drug delivery or dose selection issues) none have reached the market, although dexpramipexole, which enhances mitochondrial function, is currently undergoing Phase III trials sponsored by Knapp (Cudcowicz et al, 2011).

A key challenge to the development of new drugs based upon the genetic information derived from familial ALS, as well as genes associated with sporadic ALS, is to understand how mutations in the various genes produce a similar clinical and pathological phenotype. In other words, what is the final common pathway by which these genetic mutations produce ALS? Generic explanations such as mitochondrial dysfunction or alterations in protein degradation pathways have been suggested but how these processes are affected by genetic influences remain vague. However, it is not necessary to understand the mechanism if one can develop a screen that rescues the phenotype produced by different mutations, and this is what Van Hoecke and colleagues did. Hence, they screened for different morpholinos (antisense oligos in which ribose or deoxyribose is replaced by a morpholine ring) that rescued a SOD-1 induced axonopathy in zebra fish. The most protective morpolino targeted the zebra fish Rtk2 gene, which has 67% identity to the human EPHA4 gene that encodes for the Epha4 receptor tyrosine kinase that can bind both type A and type B ephrins. Knock down of the Rtk2 gene rescued the phenotype in zebra fish with various SOD1 mutants (A4V, G37R and G93A) and SOD-1-induced axonopathy could also be rescued pharmacologically by inhibition of Epha4 using 2,5-dimethylpyrrolyl benzoic acid.  Importantly, knockdown of Rtk1, which is a paralog with 83% identity to human Epha4, was able to rescue the axonopathy induced by either mutant SOD-1, TDP-43 or knockdown of Smn1 in zebra fish, indicating that inhibition of EphA4 is protective against motor neuron degeneration irrespective of the genetic determinant of vulnerability. Having identified Epha4 as a potential modifier of SOD1-mediated pathology, the authors also studied the effects of a deletion of the Epha4 gene in mice overexpressing the G93A mutant SOD1 and were able to show that in heterozygotes, a 50% reduction in Epha4 was able to prolong survival.

As regards ALS itself, EphA4 mRNA expression in total blood was inversely collected to the age of onset such that patients with lower levels of EphA4 expression had an age of onset older than those with higher levels of expression. Suggesting that reduced EphA4 expression is associated with a reduced disease severity. Collectively, these data shed light onto an intriguing pathway in which rescue of the axonopathy is achieved irrespective of the genetic cause. A further understanding of the mechanism by which Epha4 exerts these effects could provide the basis for novel therapeutic approaches to treating ALS.


Al-Chalabi, A., Jones, A., Troakes, C., King, A., Al-Sarraj, S. and van den Berg, L.H. (2012) The genetics and neuropathology of amyotrophic lateral sclerosis. Acta Neuropathol., 124:339-352.

Cudkowicz, M., Bozik, M.E., Ingersoll, E.W., Miller, R., Mitsumoto, H., Shefner, J., Moore, D.H., Schoenfeld, D., Mather, J.L., Archibald, D., Sullivan, M., Amburgey, C., Moritz, J. and Gribkoff, V.K. (2011) The effects of dexpramipexole (KNS-760704) in individuals with amyotrophic lateral sclerosis. Nat. Med., 17:1652-1656.

DeJesus-Hernandez, M., Mackenzie, I.R., Boeve, B.F., et al. (2011) Expanded GGGGCC hexanucleotide repeat in noncoding region of C9ORF72 causes chromosome 9p-linked FTD and ALS. Neuron, 72:245-256.

Renton, A.E., Majounie, E., Waite, A., et al., A hexanucleotide repeat expansion in C9ORF72 is the cause of chromosome 9p21-linked ALS-FTD. Neuron, 72:257-268.

Van Hoecke, A., Schoonaert, L., Lemmens, R., Timmers, M., Staats, K.A., Laird, A.S., Peeters, E., Philips, T., Goris, A., Dubois, B., Andersen, P.M., Al-Chalabi, A., Thijs, V., Turnley, A.M., van Vught, P.W., Veldink, J.H., Hardiman, O., Van Den Bosch, L., Gonzalez-Perez, P., Van Damme, P., Brown, R.H. Jr., van den Berg, L.H. and Robberecht, W. (2012) EPHA4 is a disease modifier of amyotrophic lateral sclerosis in animal models and in humans. Nat. Med., Aug 26. doi: 10.1038/nm.2901. [Epub ahead of print]

Palladium catalyzed oxidative esterifications of aldehydes and benzylic alcohols

Over the past 10 years there has been a successful effort on the direct esterification of aldehydes with alcohols  using palladium catalysts mainly using molecular oxygen as the oxidant as the most environmentally friendly variation.

The common proposed mechanism for the transformation is shown below

The first step is the reduction of the Pd(II) species to Pd(0) and recently Xu and co-workers (Chem. Com. 2012, 48, 8592-8594) have taken this in consideration and added hydrosilanes to the reaction mixture in order to accelerate the first catalytic step in the aerobic oxidation of benzylic acohols.


Polymethylhydrosiloxane (PMHS) was found to be the best performing hydrosilane and gave high chemoselective oxidation for benzyl alcohol.  However, the reaction was found to be substrate dependent and for other substituted benzylic alochols, the aldehyde was the major product (Table 1)

Further attempts to improve the conversion and selectivity on o-tolylmethanol lead to the discovery that addition of BiCl3 as co-catalyst resulted in complete conversion and excellent ester selectivity and reduction of the reaction time. The optimized bimetallic Pd/Bi catalyst in the presence of PMHS proceeded in high yield and excellent selectivities also on different benzylic alcohols (Scheme 1).

Scheme 1.

 In the former oxidative esterification there was no reported issue of dimeric esterification presumably because methanol is used as solvent and therefore in large excess.  A highly selective cross-oxidative esterification was reported earlier this year by Lei (Angew. Chem. Int. Ed. 2012, 51, 5662) where the esterification of an aldehyde with a benzylic alcohol occurred before any alcohol oxidation. This selective transformation was achieved by using benzyl chloride as the oxidant. Other organohalides selectively converted the benzylic alcohol to the corresponding aldehyde without any ester formation. (Table 2) The use of aryl halides for the oxidation of alcohols is well documented and Lei’s paper cites a review by Muzart from 2003 (Tetrahedron 2003, 59, 5789), although there have been more recent publications improving this reaction  (J. Org. Chem. 2011, 76, 1390; Org. Lett. 2003, 5, 2485)…but not cited.

The use of 2-chloroactophenone as oxidant of alcohols has not been reported before but a similar reactivity with a-bromo sulfoxide has been previuosly described by Asensio (Adv. Synth. Catal. 2007, 349, 987-991). Lei, however, describes the alcoholisis step to generate I2 in Scheme 1 to occcur at the Pd-enolate bond of I1 (A-B = Cl-enolate) whereas Asensio proposes the alcoholisis to occur at the Pd-Cl bond. In the case of benzyl chloride, the alcoholysis is proposed to occur at the Pd-Cl bond by Lei, and he confirms it by isotopic labeling experiments.

The benzyl group seems therefore to be the key for the selective esterification of aldehydes with alcohols and Lei proposes that it forms a n3-coordination to palladium and facilitates dissociation of one of the triphenyphosphine ligands favouring the coordination of the aldehyde to the palladium.

The high selectivity of this reaction allows the authors to use the reagents (aldehyde and alcohol) in a 1:1 ratio, rather than using the alcohol in excess as originally described by the authors who first described the oxidative esterification using benzyl chloride (Tetrahedron Lett. 2011, 52, 5319). By the way…no proper citation (again) to this reference in Lei’s paper!!.

As the original publication (Table 3), Lei’s procedure is suitable for electron-rich and electron-deficient aromatic aldehydes and also aliphatic aldehydes and a,b-unsaturated aldehyes, their novelty, however, is the use of various aliphatic, benzylic and allylic alcohols  without alcohol oxidation observed (Table 4), making this piece of work a general and selective oxidative esterification of aldehyde with alcohol in a 1:1 ration.