Oxidative Aliphatic C-H Fluorination with Fluoride Ion...

Oxidative Aliphatic C-H Fluorination with Fluoride Ion Catalyzed by a Manganese Porphryin

Liu, W.; Huang, X.; Cheng, M.; Nielsen, R.; Goddard, W.; Groves, J.Science, 2012, 337, 1322-1325

Carolyn L. LaddNovember 13, 2012

recently reported (24), suggesting that a similarreactive oxo- or dioxo-manganese(V) intermediate(29) is responsible for the H abstraction step inboth reactions.

Likewise, sclareolide, a plant-derived terpenoidwith antifungal and cytotoxic activities, affordedC2 and C3 methylene-fluorinated products in anoverall 58% yield (Fig. 1B). C2-fluorination wasfavored by nearly 3:1, probably because of thesteric hindrance of the gem-dimethyl groups atC4. The products could be separated chromato-graphically. C2 selectivity has been observed forthis substrate by Baran and Eschenmoser for aRh-catalyzed amination (30, 31) and by Whiteand Chen in a Fe(pdp)/H2O2-mediated oxidation(32). In contrast, reaction of this molecule usingSelectfluor (17) afforded an intractable mixture.

F-substituted steroids, such as dexamethasoneand fluasterone, have been found to be beneficialin blocking metabolic pathways (33–35), and18F-fluorodihydrotestosterone has shown prom-ise as a radiotracer for imaging prostate cancer inmen (36). Because a direct, late-stage steroid flu-orination protocol could greatly facilitate suchapplications, we sought to apply this manganese-catalyzed fluorination reaction to simple steroids.We examined the fluorination of 5a-androstan-17-one, which contains 28 unactivated sp3 C-Hbonds (Fig. 1C). Analysis of this molecule sug-gested that the carbonyl group would electron-ically deactivate ring D. Rings B and C aresterically hindered, leaving the methylene groupsof the A ring as the most likely sites for H ab-straction. Consistent with this analysis, and de-

spite the complexity of the molecule, only the C2and C3 positions in the A ring were fluorinatedin an overall yield of 55% (78% of the productdistribution at 70% conversion, withminor amountsof oxygenated products). The products of the re-action could be readily separated by columnchromatography and structurally assigned by thediagnostic 19F–nuclear magnetic resonance (NMR)spectrum and the characteristic proton J-couplings(figs. S19 to S22). A 5:1 a/b diastereoselectivitywas observed for both the C2 and C3 positions,probably reflecting the steric effect of the axialmethyl group at C10.

The reaction of bornyl acetate afforded a 55%yield of a single product, exo-5-fluoro-bornyl ac-etate (Fig. 1D). The characterization of this pro-duct was based on C-H correlation NMR and19F-NMR spectroscopy (figs. S27 to S30) (37).We anticipated that the C5 position of camphorwould also be accessible, in analogy to the selec-tivity of P450cam (CYP101) (38). However,treating camphor under the standard fluorinationconditions resulted in 95% recovered starting ma-terial. We attribute the low reactivity in this caseto the electron-withdrawing carbonyl group,which apparently deactivates the entire moleculetoward fluorination, as with the monofluorideproducts. These results highlight the subtle elec-tronic effects on both the reactivity and selectivityof the fluorination reaction.

We suggest the catalytic cycle shown in Fig.2A for this manganese porphyrin–catalyzedfluorination, although there are numerous aspectsof these transformations that will require furtherelucidation. Oxidation of the resting Mn(TMP)Fcatalyst, formed in situ, would afford a reactiveoxomanganese(V) species (29), O=MnV(TMP)F,which then abstracts a substrate H atom toproduce a C-centered radical and a HO-MnIV-Frebound intermediate. Fluoride binding to sepa-rately prepared MnIV(O)(TMP) was indicated byanultraviolet (UV) spectral shift (423 to427nm) thatwe assign to the formation of [MnIV(O)(F)(TMP)]–,in analogy to the well-characterized coordinationof hydroxide to MnIV(O) (39).

The key step in forming the fluorinated productsis capture of the incipient substrate radicals either byHO-MnIV-F or a trans-difluoro-manganese(IV)species. There is no precedent for such a Fatom transfer. In this important regard, the flu-orination reaction differs from the manganese/hypochlorite chlorinating system we have de-scribed (24). Chloride ion is rapidly and re-versibly oxidized to hypochlorite by oxoMnV

porphyrins (40). Although HOF is known (15),there is no evidence that fluoride is oxidized inthat way under these conditions. The importanceof the hypochlorite in the Mn/–OCl case is illus-trated by the observation of C-H bromination inthe presence of hypobromite, even with a largeexcess of chloride ion present. We attribute theunusual methylene selectivity observed in boththe fluorination and chlorination reactions to stereo-electronically enforced steric clashes between thesubstrate and the approaching oxoMnV catalyst

Table 1. Manganese porphyrin–catalyzed fluorination of simple molecules. Reactions were run for 6 to8 hours at 50°C under N2 in 3:1 CH3CN/CH2Cl2 solvent, 1.5 mmol substrate, 4.5 mmol silver fluoride, 6 to8 mole % catalyst, 0.3 equivalent of tetrabutylammonium fluoride (TBAF) trihydrate, and 6 to 8 equiv-alents of iodosylbenzene. Yields were determined by integration of gas chromatography traces usingnaphthalene as the internal standard. Typical conversions were 70%. Unless otherwise noted, all majorfluorination products were isolated as single compounds.

Entry Substrate Fluorination product Entry Substrate Major fluorination product

Minor sites

1

2, 49%

7

8, 46% dr=6:1

C4 14%

2

3, 51%

8

9, 44% dr=8:1

C4 12%

3

4, 55%

9

10, 42%

C3 11%

4

5, 53% 1:1.4

10

11, 51% dr=1.5:1

C2 <2%

5 6, 49%

exo: endo=5.7

11 12a, 30%

cis/trans=1:1

12b, C3 27%

cis/trans=2:1

6

7, 2:1

12

13, 49% dr=1.6:1

C3 9%

*Rearranged product identified by the characteristic m-(CH2F) peak in the mass spectrum. †Isolated as diastereomers.

www.sciencemag.org SCIENCE VOL 337 14 SEPTEMBER 2012 1323

REPORTS

on

Oct

ober

5, 2

012

www.

scie

ncem

ag.o

rgDo

wnlo

aded

from

1

Fluorine in Nature

-13th most abundant element, yet only 21naturally occurring organofluorine compounds known compared to 3500 for other halogens-Chlorination and bromination reactions well-known to occur via haloperoxidases (electrophilic halogenation).-No fluoroperoxidase known to date (high oxidation potential)-The first natural fluorinating enzyme was 5'-fluoro-5'-deoxyadenosine synthase, found in Streptomyces cattleya, a bacteria.-Mechanism involves nucleophilic attack of F- at the C’5 center.

O'Hagan, D.; Schaffrath, C.; Cobb, S. L.; Hamilton, J. T. G.; Murphy, C. D. Nature 2002, 416, 279–279.

Furuya, T.; Kamlet, A. S.; Ritter, T. Nature 2011, 473, 470–477. 2

N

NN

N

NH2

O

HOH

S+

H3+NO

O-

F-

Fluorinase

N

NN

N

NH2

O

HOH

F O

HF

NAD+O

OHF

S-adensosylmethionine 5'-fluoro-5'-deoxyadenosine Fluoroacetaldehyde Fluoroacetate

Properties of Fluorine

ElementVan der WaalAtomic Radii

(Å)

Bond Length(Å)

(C-X)BDE (kcal/mol)

F 1.43 1.35 105.4

C 1.70 1.54 83.1

H 1.2 1.09 98.8

O 1.52 1.43 84.0

N 1.55 1.47 69.7

Cl 1.74 1.77 78.5

-Forms strongest bonds to carbon.-C-F bonds more electrostatic than covalent with large dipole.-Poor polarizability: forms weak hydrogen bonds, despite 3 lone pairs.-Poor nucleophile: strongly solvated in polar solvents and forms tight ion pairs in aprotic solvents (high solvation energy).

O'Hagan, D. Chem. Soc. Rev. 2008, 37, 308. 3

Importance of Fluorine

-Accounts for 30% of agrochemicals and 20% of pharmaceuticals.-Incorporation of fluorine into drug candidates known to increase bioavailability, block phase I metabolism by cytochrome P450 enyzmes, decrease toxicity, and improve interactions with target.-Use of 18F for positron emission tomography (PET) for detection of diseases.

4

years, focusing on the catalytic enantioselective con-struction of carbon-fluorine bonds. Diastereoselectiveprocesses, non-catalytic reactions and functionaliza-

tion of fluorine-containing compounds are beyond thescope of the present review.[3g]

Sylvain Lectard graduatedfrom Ecole Sup!rieure deChimie physique !lectroniquede Lyon in 2002 and obtaineda DEA in organic synthesisunder the guidance of Profes-sor O. Piva the same year. Hethen went to the laboratory ofProfessor H. Lebel at the Uni-versit! de Montr!al, where heobtained his Ph.D. in 2007.Then, he moved to Chiba Uni-versity and worked as a JSPS post-doctoral fellowwith Professor Y. Hamada. In 2009, he joined Pro-fessor M. Sodeoka"s group at RIKEN.

Yoshitaka Hamashima gradu-ated in 1999 from the Univer-sity of Tokyo under the direc-tion of Professor M. Shibasakiand obtained his Ph.D fromthe same laboratory in 2003.In 2001, he joined ProfessorM. Sodeoka"s group as Assis-tant Professor at Tohoku Uni-versity and was promoted toLecturer in 2005. In 2006, hejoined RIKEN with ProfessorM. Sodeoka. In 2010 he joined Professor T. Kan"sgroup at the University of Shizuoka as Associate

Professor. His current research interests include syn-thetic organic chemistry, organometallic chemistry,and asymmetric catalysis.

Mikiko Sodeoka graduatedin 1981 from Chiba Univer-sity, and finished her mas-ter"s degree in 1983. She ob-tained her Ph.D. from thesame university in 1989.After working at SagamiChemical Research Centerfrom 1983 to 1986, shejoined the Faculty of Phar-maceutical Sciences, Hokkai-do University, as a researchassociate. She then worked as a post-doctoral fellowat Harvard University, after which she moved toThe University of Tokyo in 1992. She became agroup leader at Sagami Chemical Research Centerin 1996 and an associate professor at The Universityof Tokyo in 1999. In 2000, she moved to TohokuUniversity as a full professor. Since 2004 she hasbeen chief scientist of RIKEN. In 2008, she was alsoappointed as research director of the ERATO proj-ect (SODEOKA Live Cell Chemistry), JST. Her re-search interests are in the areas of synthetic organicchemistry, medicinal chemistry, and chemical biol-ogy.

Figure 1. Selected fluorinated biologically active compounds.

Adv. Synth. Catal. 2010, 352, 2708 – 2732 # 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim asc.wiley-vch.de 2709

Recent Advances in Catalytic Enantioselective Fluorination Reactions

Biologically active fluorinated molecules Lectard, S.; Hamashima, Y.; Sodeoka, M. Adv. Synth. Catal. 2010, 352, 2708–2732.

Fluorinating Strategies

1. Electrophilic fluorinating reagents (F2, NFSI, Selectfluor)2. Nucleophilic fluorinating reagents (DAST, TBAF, HF)3. Organocatalysis (Barbas, MacMillian, Jorgensen)4. Chemo-enzymatic fluorination via initial P450-mediated hydroxylation5. Decarboxylation routes

5

SN

DAST

N+

F-

TBAF

Sources of F-

Kirk, K. L. Org. Process Res. Dev. 2008, 12, 305–321.

N+N+

F

Cl

2BF4-

Selectfluor

NF

SSO

O

O

O

NFSI

N+

F

OTf-

N-fluoropyridiniumtriflate

Sources of F+

Pd-Catalyzed Electrophilic Fluorination: Hull, K. L.; Anani, W. Q.; Sanford, M. S. J. Am. Chem. Soc. 2006, 128, 7134–7135. Wang, X.; Mei, T.-S.; Yu, J.-Q. J. Am. Chem. Soc. 2009, 131, 7520–7521.Use of organocatalysis: Kwiatkowski, P.; Beeson, T. D.; Conrad, J. C.; MacMillan, D. W. C. J. Am. Chem. Soc. 2011, 133, 1738–1741. Steiner, D. D.; Mase, N.; Barbas, C. F. Angew. Chem. Int. Ed. Engl. 2005, 44, 3706–3710; Brunet, V. A.; O'Hagan, D. Angew. Chem. Int. Ed. 2008, 47, 1179–1182; Lectard, S.; Hamashima, Y.; Sodeoka, M. Adv. Synth. Catal. 2010, 352, 2708–2732


2. Use of nucleophilic fluorinating reagents (DAST, TBAF, AgF)

6

absence of the catalyst, ruling out both classicnucleophilic aromatic substitution (SNAr) andaryne mechanisms (39).

We realized that for some applications, in-cluding PET, it was necessary to have a fasterprocess. We found that the conversion of 9 to 10,using 5 mol% [(COD)Pd(CH2TMS)2] as precat-alyst, 10 mol% of 3 as supporting ligand, and 3equivalents of CsF, was complete in 2.5 hours intoluene at 110°C, yielding 80% of 10. Increasing

the amount of CsF to 6 equivalents and adding30 mol% of the solubilizing agent poly(ethyleneglycol) dimethyl ether (Me2PEG) led to full con-version in less than 30 min, albeit in yield of71%. Similar rates of reaction could be achievedby using [(cinnamyl)PdCl]2; with 5 mol% of thispalladium source and 15% ligand, the reactionproceeds to completion in !2 hours in 79% yield(NMR). We are in the process of identifying con-ditions to achieve faster conversion of the sub-

strate without diminishing the yield, as requiredfor PET applications.

As can be seen in Fig. 7, the fluorinationof aryl triflates has substantial scope. Simplearomatic substrates, such as ortho-biphenyltriflate, react rapidly to provide aryl fluoridesin high yield. Hindered substrates such as4-acetyl-2,6-dimethylphenyl triflate are also ef-ficiently converted to product (13). Electron-deficient arenes can be efficiently transformedby using only 2 mol% of catalyst (14, 18, and19). Important from a practical standpoint, avariety of heterocyclic substrates can also be suc-cessfully fluorinated by using these conditions.Flavones (17), indoles (21), and quinolines (22 to24) were all converted in good yield. More com-plex aryl triflates derived from fluorescein (20)and quinine (25) could also be effectively con-verted to their fluorinated analogs, demonstratingthat this method can be used in the preparation ofpharmaceutically relevant compounds. In somecases, product formed in high yield at 80°C(14, 17, and 24). On a 10 mmol scale, butyl 4-fluorobenzoate 14 was prepared at 80°C withno observable formation of reduced by-product(in general, 2% or less reduction product wasobserved across the range of substrates screened).

Many functional groups are tolerated, an ex-ception being Lewis basic groups such as aminesor carbonyls in the ortho position of the aryltriflate. No reaction takes place in these instances,presumably because the basic group coordinatesthe Pd center, possibly preventing transmetala-tion. As in the Pd-catalyzed formation of Ar-Obonds, the transformation of electron-rich sub-strates was more challenging. We found thatgood yields were obtained at higher temperatures(130°C).

Formation of regioisomers. Unexpectedly,regioisomeric products were formed in a fewcases (Figs. 8 and 9). Because control exper-iments did not yield any product in the absenceof catalyst, we believe that isomer formation isalso a palladium-catalyzed process. Investigat-ing a series of tolyl (26 to 28) and anisole (29to 31) triflates, we found that for substrates 26,27, and 29 the observed selectivities are quitedistinct from those reported for a benzyne process(39) (Fig. 8). Experiments with 2,6-dideuteratedanisole triflate 31 showed a reduced rate of for-mation of the undesired regioisomer, whereasthe rate of formation of the desired productremained largely unchanged, leading to a 2.5-fold increase in selectivity in comparison to thereaction with unlabeled 31. Thus, at least forthis substrate, we conclude that two compet-ing pathways are involved; it is evident thathydrogen abstraction is the rate-limiting or thefirst irreversible step in regioisomer formationand that little or none of the desired isomer isformed from the path that finally leads to theregioisomer.

Although we do not have a complete mech-anistic understanding of the pathway leading to theregioisomers, we have found that the product ratio

NBoc

F

N

F

N

F

75% (1%)110 °C

78% (2%)2 mol% "Pd"

120 °C

N

Me

CF3

83% (<1%)80 °C

F

O

O

PhF

6 mol% tBuBrettPhos (3)2 mol% [(cinnamyl)PdCl]2

(=4 mol% "Pd"), Pd:L = 1:1.5

toluene, temp, 12 h

73% (4%)110 °C

63% (2%)80 °C

OTfR

FR

(Me)2N F

84% (1%)%*C6H12, 130 °C

83% (2%)2 mol% "Pd"

110 °C

N

F

AcON

H

70% (n/o)10 mol% "Pd"

110 °C

O

O

O

F F

73% (n/o)8 mol% "Pd"

110 °C

nBuO2C

F

77% (1%)2 mol% "Pd"

80 °C

10 mmol scale85% (n/o)

2 mol% "Pd"80 °C, 24 h

F

57% (2%)*10 mol% "Pd"C6H12, 130 °C

BnO

19

1514 16

232221 24

17

20

25

F

Me

O2N

F

Ph

1282% (2%) 110 °C

O2N

F

Me

80% (<1%)2 mol% "Pd"

110 °C

18

F

Me

83% (2%)110 °C

Me

13

MeOC

* Cyclohexane (C6H12) was used as a solvent to decrease the amount of reduced product.

CsF+

Fig. 7. Fluorination of aryl triflates. Isolated yields are an average of at least two independent reactions.Values in parentheses denote the amount of reduced starting material based on the isolated product yield (n/oindicates not observed). In cases with different palladium loading, the ligand amount was adjusted accordingly(“Pd”:L = 1:1.5). Quotation marks around Pd symbols indicate the amount of Pd, not of the Pd dimer.

OTf

Me

OTfOTf

MeMe

o/m=>99:1*(57:43)

o/m/p=0:95:5†

(26:53:21)

m/p=36:64†

(63:37)

o/m=98:2*

(0:100)

m/p=98:2*

(100:0)

m/p=70:30*(62:38)

ArOTf

5 mol% [(COD)Pd(CH2TMS)2]6 mol% 3

130 °C, toluene, 12 h

26, 42% 27, 63% 28, 48%

OTf

OMe

OTfOTf

OMe MeO

29, 49% 30, 65% 31, 25%

*Product ratios and yield determined by GC. † Product ratios and yield determined by 19F NMR Spectroscopy.

F

R

F

R

F

R

o m p

+ +

Fig. 8. Regioisomers in tolyl and anisole fluorination. Chemical yields are given as the sum of ArFproducts. Traces of reduction products were also observed. Ratios are compared with those of a reportedfluorination of bromoaryls that proceeds via a benzyne intermediate (in parentheses) (39).


RESEARCH ARTICLE

on

Oct

ober

31,

201

2ww

w.sc

ienc

emag

.org

Down

load

ed fr

om

-CsF used as fluoride source.

-Steric demand of tBuBrettPhoskey for challenging reductive elimination of Ar-F

in dichloromethane (Fig. 3). The isolated [2·PdAr(F)] complexes exhibit a characteristic doublet in31P and 19F NMR spectra with a couplingconstant 2JPF of ~175 Hz, depending on the arylgroup. The x-ray structure of 4 (Ar is 2-methyl-4-trifluoromethylphenyl) confirmed the mono-meric nature of the complex (Fig. 4).

We next examined the thermolysis of 4 and5 at 100°C in toluene and found that reductiveelimination to form 6 and 7 occurs in yields of15% and 25%, respectively (Fig. 3), which dem-onstrates that Ar-F bond formation is possiblewith use of these complexes. The yields in thesereactions could be increased to 45% of 6 and55% of 7 if the reductive eliminations wereconducted in the presence of an excess of thecorresponding aryl bromide (33, 34). In thesecases, the 31P NMR spectra of the reaction mix-tures showed that the oxidative-addition com-plexes [2·PdAr(Br)] were formed as the onlyphosphorous-containing products, suggestingLPd(0) is formed along with the ArF product.

Having demonstrated that oxidative addi-tion, halide exchange (transmetalation), and re-ductive elimination were all possible, we nextexamined the catalytic conversion of ArBr 8 toArF 7. Upon treatment of 8 with AgF (1.5equivalents) and 10 mole percent (mol%) eachof 2 and [(COD)Pd(CH2TMS)2] (COD is 1,5-cyclooctadiene) (35) at 110°C for 18 hours, a52% yield of 7 was observed, which was in-creased to 74% after optimization of the re-action conditions (Fig. 5). No product was detectedin control experiments that omitted 2 and/or thePd precatalyst. This result demonstrates that Pdcomplexes supported by BrettPhos can catalyzethe conversion of an aryl bromide to an aryl fluo-ride. However, the scope of aryl bromides thatcould be effectively transformed is to date limitedto electron-poor substrates bearing an ortho sub-stituent, in line with the observation that no re-ductive elimination took place from [2·PdAr(F)]complexes with Ar being 3,5-dimethylphenyl or4-n-butylphenyl.

Efficient catalysis with aryl triflates. Con-currently, we also examined the use of aryltriflates as substrates. Although initial reactionsof the triflate of 1-naphthol (9) provided only atrace of product 10, the use of CsF in place ofAgF as the fluoride source gave 10 in 30% yield(Fig. 6). In addition, 5% of naphthalene (11)was also observed. We also found that the read-

ily prepared and more stable [(cinnamyl)PdCl]2could be used as the Pd precatalyst with a sim-ilar outcome. Overall, this result is importantbecause it demonstrates that the fluorination canbe carried out without needing a stoichiometicquantity of a noble metal component while stillusing a nucleophilic fluoride source.

In order to optimize the Ag-free reaction, abroad range of ligands were examined; underthese conditions (Fig. 6), only ligands closelyrelated to 2 provided more than a trace amount

of 10 (36, 37). Best results were obtained withuse of tBuBrettPhos (3) (Fig. 2) as ligand; a71% yield of 10 was realized, with only 1% ofreduction product 11 observed. Further optimi-zation of the reaction conditions increased theyield of 10 to 79%. Because the reaction provedto be sensitive to water (38), commercially ob-tained CsF was dried at 200°C under vacuumovernight and handled in a nitrogen-filled glove-box. Replacing CsF with spray-dried KF afforded10 in 52% yield. No reaction was observed in the

Fig. 3. Preparation of and reduc-tive elimination from [2·PdAr(F)]complexes.

L PdBr

MeR

L PdF

MeR

F

Me R

4 R = CF35 R = CN

74%70%

6 7

15%25%

L = 2. a 5 equiv. AgF, CH2Cl2, 25 °C, exclusion of light, 12 to 24 h.b toluene, 100 °C, 2 h, yields determined by 19F NMR spectroscopy.

a b

Fig. 4. X-ray structure of complex4 [ORTEP (www.ornl.gov/sci/ortep/ortep.html) drawing at 50% prob-ability, hydrogen atoms omitted forclarity].

OMe

i-Pr i-Pr

i-Pr

MeO PR2

i-Pr i-Pr

i-Pr

P(t-Bu)2

tBuXPhos (1) BrettPhos (2, R = Cyclohexyl)tBuBrettPhos (3, R = t-Bu)

Fig. 2. Ligands for the successful reductive elim-ination of Ar-F.

Fig. 5. Catalytic conversionof aryl bromide 8 to aryl flu-oride 7. Br

Me

F

MeCN CN

Yield was determined by NMR spectroscopy due to volatility of product. (1% of reduced substrate, m-tolunitrile, was also observed.)

5 mol% [(COD)Pd(CH2TMS)2]10 mol% 2, 1.5 eq AgF

toluene, 130 °C, 18 hexclusion of light

74%

8 7

Fig. 6. Optimization of aryl triflatefluorination. Conversion and yield weredetermined by GC.

OTf F

9 10

+

11

10 11

[cinnamylPdCl]2ligand 2 or 3AgF or CsF

tol, 110 °C, 18 h†

1010102

2 (10)2 (10)3 (10)3 (3)

Conversion

AgF (1.5)CsF (1.5)CsF (1.5)CsF (2.0)

trace30%71%79%

‡5%1%1%

‡90%100%100%

* mol% of palladium equivalents ("Pd"), † time not optimized, ‡ not determined

Pd(mol%)*

Ligand(mol%)

F- source(eq)

25 SEPTEMBER 2009 VOL 325 SCIENCE www.sciencemag.org1662

RESEARCH ARTICLE

on

Oct

ober

31,

201

2ww

w.sc

ienc

emag

.org

Down

load

ed fr

om

Watson, D. A.; Su, M.; Teverovskiy, G.; Zhang, Y.; Garcia-Fortanet, J.; Kinzel, T.; Buchwald, S. L. Science 2009, 325, 1661–1664 See also for nucleophilic fluorination of benzylic substrates: McMurtrey, K. B.; Racowski, J. M.; Sanford, M. S. Org. Lett. 2012, 14, 4094–4097.


4. Chemo-enzymatic fluorination via initial P450-mediated hydroxylation

7

Chemo-enzymatic fluorination ofunactivated organic compoundsAndrea Rentmeister1, Frances H Arnold1 & Rudi Fasan1,2

Fluorination has gained an increasingly important role in drugdiscovery and development. Here we describe a versatilestrategy that combines cytochrome P450–catalyzedoxygenation with deoxofluorination to achieve mono- andpolyfluorination of nonreactive sites in a variety of organicscaffolds. This procedure was applied for the rapididentification of fluorinated drug derivatives with enhancedmembrane permeability.

Fluorination has become an increasingly important tool for fine-tuning the pharmacokinetic and pharmacological properties of drugsand lead compounds, thus leading to a growing number of fluorine-containing pharmaceuticals on the market1. Benefits of hydrogen-to-fluorine substitutions arise principally from their effects onmembrane permeability, metabolic stability and/or receptor-bindingproperties of bioactive molecules1–3. In many cases, fluorination ofmuch less active precursors has yielded potent drugs with enhancedbioavailability, reduced toxicity or improved affinity for the targetreceptor3. A number of methods have been developed for synthesis offluorinated compounds4,5, including asymmetric fluorination strate-gies6,7 and chemo-enzymatic approaches8,9. Despite this progress,selective incorporation of fluorine at non-activated or weakly reactivesites of a target scaffold remains difficult and may require severalsynthetic steps.

Here we describe a facile two-step procedure for the selectivefluorination of one or more non-activated sites in an organic molecule.This approach couples the exceptional ability of cytochrome P450monooxygenases to selectively insert oxygen into nonreactive C-Hbonds with a deoxofluorination reaction in which the newly generatedhydroxyl group is substituted by fluorine by means of a nucleophilicfluorinating reagent (Fig. 1). To test the validity of this approach, wetargeted various classes of small molecules, including marketed phar-maceuticals (Supplementary Fig. 1 online). For the enzymatic step, weused variants of the bacterial long-chain fatty acid hydroxylaseP450BM3 from Bacillus megaterium. Catalytic self-sufficiency, highmonooxygenase activity and high expression level in Escherichia colirender P450BM3 an attractive catalyst for in vitro and in vivo applica-tions10. For this work, we assembled a panel of 96 P450s derived from acatalytically promiscuous P450BM3 variant identified in the early stagesof the directed evolution of a proficient alkane monoxygenase11. Thesevariants were found to exhibit good activity and various degrees ofselectivity on alkanes and non-alkane substrates11.

The first group of test molecules (1, 2 and 3; Fig. 2a,b) contain acyclopentenone moiety found in several natural products (for exam-ple, jasmonoids, cyclopentanoid antibiotics and prostaglandins). Thesynthesis and functionalization of these scaffolds is not trivial12. Theactivities of the enzymes toward these substrates were probed inmultiwell format using gas chromatography and GC-MS (Fig. 2a).Depending on the substrate, approximately 30 to 50% of the 96enzyme variants displayed useful activity (4800 turnovers), while 30to 50% of this active subset showed moderate to excellent regioselec-tivity (50–100%). The most active and selective variants were appliedin preparative scale reactions (100–300 mg) using B0.05 mol%catalyst. Compared with 96-well plate reactions, three to four timeshigher turnover numbers could be obtained using purified enzymeand longer reaction times (24–48 h). After flash chromatographypurification, the hydroxylated products were subjected to deoxo-fluorination using the common nucleophilic fluorinating agentdiethylaminosulfur trifluoride (DAST, 4). The identities and puritiesof the fluorinated products were established by GC-MS, HRMS and1H, 13C and 19F NMR (Supplementary Methods online). Using thisstrategy, we were able to target two to three different sites on eachsubstrate with good to excellent regioselectivity (55–100%), therebyaffording the fluorinated derivatives 5, 6, 7, 8, 9, 10 and 11 with yieldsof up to 80% over the two steps.

Next we tested this fluorination strategy on a methylester pro-drugof the anti-inflammatory drug ibuprofen (12; Fig. 2a,c). Althoughpreparation of a-fluoro derivatives of this compound is straightfor-ward13, incorporating fluorine atoms in the poorly reactive isobutylgroup is not. The general procedure described above enabled us toidentify two chemo-enzymatic routes to achieve this goal in a selective(position 1, 75%; position 2, 100%) and efficient manner (yields overtwo steps for 15 and 16 were 62% and 84%, respectively) and atpreparative scales (150–200 mg).

R–H

R–OH R–FDF

R–OCH3P450

monooxygenase

DF F

F

F

DF

DF

Variant 1

Variant 3

Variant 2Drug or

chemicalbuilding block

a b

Figure 1 Cytochrome P450-based approach for selective fluorination oforganic molecules. (a) Hydroxyl groups are introduced (via direct oxygeninsertion) or exposed (via hydroxylation-demethylation) in a targetscaffold using a P450 monooxygenase and substituted for fluorine usinga nucleophilic fluorinating agent. DF, deoxofluorination. (b) Differentfluorinated derivatives of a molecule of interest are obtained using P450variants with different regioselectivities.

Received 11 August; accepted 27 October; published online 16 November 2008; doi:10.1038/nchembio.128

1Department of Chemistry and Chemical Engineering, California Institute of Technology 210-41, 1200 East California Boulevard, Pasadena, California 91125, USA.2Present address: Department of Chemistry, University of Rochester, Hutchison Hall, RC Box 270216, Rochester, New York 14627, USA. Correspondence should beaddressed to R.F. ([email protected]).

26 VOLUME 5 NUMBER 1 JANUARY 2009 NATURE CHEMICAL BIOLOGY

BR I E F COMMUNICAT IONS

© 2

009

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

Step 1: Insertion of oxygen using P450 monooxygenaseStep 2: Deoxofluorination using stoichiometric DAST as a nucleophilic fluorine source.

-Different variants of P450 produce different regioselectivities.

-Useful for molecules within the range110-450 g/mol.

-Issues with highly polar substrates, solubility in aqueous media, competing elimination reactions

-Yield ranged from 45%-98%

Rentmeister, A.; Arnold, F. H.; Fasan, R. Nat Chem Biol 2008, 5, 26–28.

We then investigated whether two sequential chemo-enzymatictransformations could be used to fluorinate multiple sites of thesame molecule. P450 variant B4 (var-B4)—which was used to convert12 to 13—was found to retain comparable activity on 16, providing apossible route to the desired 17 intermediate. Re-screening of 12-active variants on 16, however, led to the identification of a moresuitable candidate, var-B2, with higher activity than var-B4 toward 16and excellent (100%) 2-regioselectivity. Using var-B2, the synthesis offluorohydroxy derivative 17 was afforded in higher yields (93% versus72% for conversion of 12 to 13) and required less catalyst (0.06 mol%versus 0.1 mol% for conversion of 12 to 13). 17 was then convertedquantitatively to the desired difluoroderivative 18.

The value of the present approach as synthetic tool for asymmetricfluorination was also examined. In the absence of anchimeric groupparticipation, the deoxofluorination reaction generally preserves theenantiopurity of the enzymatic products through inversion of config-uration14. Chiral gas chromatography analysis showed appreciablestereoselectivity during preparation of 7 (78% enantiomeric excess(ee), 9 (diasteromeric ratio (dr) 1:8.5), 10 (dr 4:96), 15 (dr 1:3.2) and18 (dr 1:3.7) (see Supplementary Figs. 2–6 online for gas chromato-graphy traces). We then extended our previous investigations on2-phenylacetic acid esters15, carrying out the asymmetric synthesisof the corresponding 2-fluoro-2-arylacetic acid derivatives at100-mg scale (19a, 20a, 21a, 22a and 23a; Supplementary Table 1online). In this case, up to 89% ee in up to 60% yield (two steps)was achieved.

P450-catalyzed hydroxylation of methoxy groups leads to exposureof a free hydroxyl group through decomposition of the hemi-acetalproduced and release of formaldehyde. We reasoned that our chemo-enzymatic strategy could be extended to substitute a methoxy group forfluorine, a challenging transformation for traditional chemical meth-ods. This approach was first tested on the 5-phenyloxazoline derivative24 (Fig. 3a,b). The demethylation activities of the P450 variants couldbe easily assessed using a colorimetric screen (Fig. 3a). The most activevariants from the screen were further analyzed with respect to theregioselectivity of oxidation using GC-MS. The highly selective P450variant var-H1 (95%) was thus applied in combination with deoxo-fluorination to afford the desired fluorine-containing compound (26).

The same approach was tested on a set of derivatives of thesynthetically important building block Corey lactone (Fig. 3c). Theuse of various Corey lactones (27a, 28a and 29a) enabled us toinvestigate the tolerance of the enzymatic transformation to structuralvariations within the target scaffold. Based on the colorimetric screen,30 variants displayed activity on at least one Corey lactone (12 on 27a,17 on 28a, 5 on 29a). Twelve variants were found to accept both 27aand 28a, five were found to accept 28a and 29a and five were found toaccept 27a and 29a. Notably, four variants (B10% of the Coreylactone–active variants) could be used to activate the substrate forsubsequent fluorination, regardless of the size of the variable sub-stituent. Using the most active and selective enzymes toward each ofthe Corey lactones, the desired fluorine-containing compounds 27c,28c and 29c were synthesized and isolated.

a

cb

R2

O

R1

R3

O

O

O

O

O

O

i, ii

iii, iv

v, vi

vii, viiiix, x

xi, xii

xiii, xiv

5

6

7

8

10

11

O

1

O

2

O

3

F

F

F

F

F

F

OO

F

OO

i

OO

HO

OO

OH

13

14

15

16

18

12

12

iii

v

vi

ii

iv

OO

F

OO

F

HO 17

OO

F

F

O

2

1,500

0

TON

1P450 variant library

96

WT

OO

12

1,500

01

P450 variant library96

TON

WT

O

9

F

Figure 2 Chemo-enzymatic fluorination of organic molecules. (a) Screening of P450 library in 96-well format. Reactions were carried out in the presence ofthe substrate, P450 enzyme from cell lysate and a glucose-6-phosphate dehydrogenase–based NADPH regeneration system. TON, turnover number. Standarderror is within 15%. WT, wild-type P450BM3. (b) Selective fluorination of cyclopentenone derivatives. Reagents and conditions: (i) 1, 0.04 mol% var-H3,88%; (ii) DAST (1.2 equiv), CH2Cl2, –78 1C, 12 h, 90% (20% ee); (iii) 1, 0.04 mol% var-G6, 45%; (iv) DAST (1.2 equiv), CH2Cl2, –78 1C, 12 h, 85%;(v) 2, 0.05 mol% var-H3, 85%; (vi) DAST (1.3 equiv), CH2Cl2, –78 1C, 12 h, 92% (78% ee); (vii) 2, 0.05 mol% var-G4, 42%; (viii) DAST (1.5 equiv),CH2Cl2, –78 1C, 12 h, 89%; (ix) 3, 0.05 mol% var-D10, 69%; (x) DAST (1.2 equiv), CH2Cl2, –78 1C, 3 h, 88% (dr 1:8.5; major, 5% ee; minor, 71% ee);(xi) 3, 0.05 mol% var-G4, 62%; (xii) DAST (1.2 equiv), CH2Cl2, –78 1C, 5 h, 92% (dr 4:96; major, 0% ee; minor, 57% ee); (xiii) 3, 0.07 mol% var-G5,32%; (xiv) DAST (1.2 equiv), CH2Cl2, –78 1C, 5 h, 90% (dr not measurable). (c) Selective mono- and difluorination of pro-drug ibuprofen methylester.Reagents and conditions: (i) 0.1 mol% var-B4, 72%; (ii) DAST (1.4 equiv), CH2Cl2, –78 1C, 12 h, 86% (dr 1:3.2; major, 19% ee; minor, 44% ee);(iii) 0.05 mol% var-G4, 88%; (iv) DAST (1.4 equiv), CH2Cl2, –78 1C, 12 h, 95%; (v) 0.06 mol% var-B2, 93%; (vi) DAST (1.2 equiv), CH2Cl2, –78 1C,12 h, 98% (dr 1:3.7; major, 9% ee; minor, 9% ee). All experimental procedures are described in detail in Supplementary Methods. The sequences of theP450 variants are described in Supplementary Table 2 online. Yields refer to the isolated products. Enantiomeric excess values were determined by chiralgas chromatography analysis.

BR I E F COMMUNICAT IONS

NATURE CHEMICAL BIOLOGY VOLUME 5 NUMBER 1 JANUARY 2009 27

© 2

009

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.


5. Decarboxylation routes

8removed while the benzoic carboxyl groups remained intact,providing alkyl fluorides 28 and 29, respectively (eqs 4 and 5).The above relative reactivities of carboxylic acids strongly

suggest that the reaction proceeds by an oxidative radicaldecarboxylation mechanism. To provide further evidence of theradical mechanism, cyclopropylacetic acidA-30was designed as aradical probe.17 The silver-catalyzed reaction of A-30 withSELECTFLUOR® reagent in water afforded the ring-openingproduct 30 in 40% yield as a 4:1 mixture of two stereoisomersalong with the recovery of A-30 in 21% yield (eq 6). Nocorresponding fluoromethylcyclopropane derivative could bedetected by 1H NMR analysis. This result strongly supports theinvolvement of a free radical mechanism in the silver-catalyzedfluorodecarboxylation.Although the detailed mechanism is still not clear, a tentative

one can be proposed (Figure 1). The oxidation of Ag(I) bySELECTFLUOR® reagent generates an Ag(III)!F intermedi-ate, presumably via oxidative insertion. The trivalent silverspecies then undergoes single electron transfer (SET) with acarboxylate anion to give the divalent silver intermediate Ag(II)!F and a carboxyl radical. The fast decarboxylation of the carboxylradical provides the corresponding alkyl radical, which thenabstracts the fluorine atom of the adjacent Ag(II)!F to afford thealkyl fluoride product and regenerate the Ag(I) catalyst. Thus,the silver-catalyzed decarboxylative fluorination likely involvesSET followed by fluorine atom transfer. The inactiveness ofNFSI in the fluorodecarboxylation might be ascribed to its failureto generate the high-valent silver species, as NFSI is a muchweaker oxidant than SELECTFLUOR® reagent.

The Ag(II)- or Ag(III)-mediated decarboxylation of carboxylicacids is well-documented.18,19 It is reasonable to assume thatAg(III)!F has a reactivity similar to that of Ag(II) indecarboxylation since the Ag!F bonding is substantially covalentin Ag(III) and Ag(II) fluorides.20 While the fluorine atomtransfer from Ag(II)!F to an alkyl radical is unprecedented,transition-metal-assisted halogen (Cl or Br) atom transfer, that is,trapping of an alkyl radical by a complexed metal halide in ahigher oxidation state (Mn+1LmX) to give the alkyl halide and themetal ion in a lower oxidation state (MnLm), is well-known.

21 Analternative explanation for the fluorine transfer would be an SETmechanism involving oxidation of the alkyl radical to acarbocation by Ag(II) with subsequent capture of thecarbocation by a fluoride anion. However, this is unlikelybecause solvated F! is much less nucleophilic than H2O.Another possibility for the fluorination would be fluorine

transfer from SELECTFLUOR® reagent to an alkyl radical. Totest this hypothesis, we designed the following experiments.When 1-adamantanecarboxylic acid was treated with Ag-(BPy)2S2O8 (200 mol %) and SELECTFLUOR® reagent (200mol %) in acetone/H2O at room temperature, decarboxylationoccurred. However, only a trace amount of 1-fluoroadamantanecould be detected by GC!MS, as the major product wasadamantan-1-ol in 43% yield. This showed that the oxidation ofadamantyl radical by the Ag(II) complex is much faster than thefluorine transfer to the adamantyl radical from SELECT-FLUOR® reagent. In another set of experiments, tert-butyl 2-ethyltetradecaneperoxoate (PE-1), the perester of acid A-1, wasused as the radical precursor (eq 7). The reaction of PE-1 with

SELECTFLUOR® reagent (200 mol %) in dry acetone wascarried out in a sealed tube at 120 °C. All of the PE-1 wasconsumed within 2 h. The 1H NMR and GC-MS analyses of theresulting mixture showed the formation of pentadecane (30%)and fluoride 1 (22%) along with a number of unidentifiedbyproducts, implying that the rate of F abstraction fromSELECTFLUOR® reagent is comparable to that of H abstractionfrom the solvent acetone for secondary alkyl radicals. When thesame reaction was carried out in the 1:1 (v:v) acetone/watersolution, pentadecane was obtained in 26% yield, but only a 4%yield of fluoride 1 was observed. Instead, pentadecan-3-ol wasformed in 7% yield as determined by GC!MS (see the

Figure 1. Proposed Mechanism of Silver-Catalyzed DecarboxylativeFluorination.

Journal of the American Chemical Society Communication

dx.doi.org/10.1021/ja3048255 | J. Am. Chem. Soc. 2012, 134, 10401!1040410403

Yin, F.; Wang, Z.; Li, Z.; Li, C. J. Am. Chem. Soc. 2012, 134, 10401–10404.

removed while the benzoic carboxyl groups remained intact,providing alkyl fluorides 28 and 29, respectively (eqs 4 and 5).The above relative reactivities of carboxylic acids strongly

suggest that the reaction proceeds by an oxidative radicaldecarboxylation mechanism. To provide further evidence of theradical mechanism, cyclopropylacetic acidA-30was designed as aradical probe.17 The silver-catalyzed reaction of A-30 withSELECTFLUOR® reagent in water afforded the ring-openingproduct 30 in 40% yield as a 4:1 mixture of two stereoisomersalong with the recovery of A-30 in 21% yield (eq 6). Nocorresponding fluoromethylcyclopropane derivative could bedetected by 1H NMR analysis. This result strongly supports theinvolvement of a free radical mechanism in the silver-catalyzedfluorodecarboxylation.Although the detailed mechanism is still not clear, a tentative

one can be proposed (Figure 1). The oxidation of Ag(I) bySELECTFLUOR® reagent generates an Ag(III)!F intermedi-ate, presumably via oxidative insertion. The trivalent silverspecies then undergoes single electron transfer (SET) with acarboxylate anion to give the divalent silver intermediate Ag(II)!F and a carboxyl radical. The fast decarboxylation of the carboxylradical provides the corresponding alkyl radical, which thenabstracts the fluorine atom of the adjacent Ag(II)!F to afford thealkyl fluoride product and regenerate the Ag(I) catalyst. Thus,the silver-catalyzed decarboxylative fluorination likely involvesSET followed by fluorine atom transfer. The inactiveness ofNFSI in the fluorodecarboxylation might be ascribed to its failureto generate the high-valent silver species, as NFSI is a muchweaker oxidant than SELECTFLUOR® reagent.

The Ag(II)- or Ag(III)-mediated decarboxylation of carboxylicacids is well-documented.18,19 It is reasonable to assume thatAg(III)!F has a reactivity similar to that of Ag(II) indecarboxylation since the Ag!F bonding is substantially covalentin Ag(III) and Ag(II) fluorides.20 While the fluorine atomtransfer from Ag(II)!F to an alkyl radical is unprecedented,transition-metal-assisted halogen (Cl or Br) atom transfer, that is,trapping of an alkyl radical by a complexed metal halide in ahigher oxidation state (Mn+1LmX) to give the alkyl halide and themetal ion in a lower oxidation state (MnLm), is well-known.

21 Analternative explanation for the fluorine transfer would be an SETmechanism involving oxidation of the alkyl radical to acarbocation by Ag(II) with subsequent capture of thecarbocation by a fluoride anion. However, this is unlikelybecause solvated F! is much less nucleophilic than H2O.Another possibility for the fluorination would be fluorine

transfer from SELECTFLUOR® reagent to an alkyl radical. Totest this hypothesis, we designed the following experiments.When 1-adamantanecarboxylic acid was treated with Ag-(BPy)2S2O8 (200 mol %) and SELECTFLUOR® reagent (200mol %) in acetone/H2O at room temperature, decarboxylationoccurred. However, only a trace amount of 1-fluoroadamantanecould be detected by GC!MS, as the major product wasadamantan-1-ol in 43% yield. This showed that the oxidation ofadamantyl radical by the Ag(II) complex is much faster than thefluorine transfer to the adamantyl radical from SELECT-FLUOR® reagent. In another set of experiments, tert-butyl 2-ethyltetradecaneperoxoate (PE-1), the perester of acid A-1, wasused as the radical precursor (eq 7). The reaction of PE-1 with

SELECTFLUOR® reagent (200 mol %) in dry acetone wascarried out in a sealed tube at 120 °C. All of the PE-1 wasconsumed within 2 h. The 1H NMR and GC-MS analyses of theresulting mixture showed the formation of pentadecane (30%)and fluoride 1 (22%) along with a number of unidentifiedbyproducts, implying that the rate of F abstraction fromSELECTFLUOR® reagent is comparable to that of H abstractionfrom the solvent acetone for secondary alkyl radicals. When thesame reaction was carried out in the 1:1 (v:v) acetone/watersolution, pentadecane was obtained in 26% yield, but only a 4%yield of fluoride 1 was observed. Instead, pentadecan-3-ol wasformed in 7% yield as determined by GC!MS (see the

Figure 1. Proposed Mechanism of Silver-Catalyzed DecarboxylativeFluorination.

Journal of the American Chemical Society Communication

dx.doi.org/10.1021/ja3048255 | J. Am. Chem. Soc. 2012, 134, 10401!1040410403

-Involves oxidation of Ag(I) to Ag(III) by Selecfluor.-SET to generate a carboxyl radical, followed bydecarboxylation generates an alkyl radical.-The alkyl radical then abstracts the fluorine atom,producing the fluorinated product.

A Brief Intro to Porphyrins

9

-Name derived from Greek word for “purple.”-Highly conjugated aromatic 18 𝛑 macrocycles.-Comprised of 4 pyrrole subunits.-Well-known example is heme (pigment in red blood cells).-Can tune properties by installing different substituents.-Useful for applications such as artificial photosynthesis, photodynamic therapy, dye-sensitized photovoltaic cells, and catalysis.-Versatile ligands: complexes of almost every metal and metalloid have been isolated.

N

NH N

HN

porphyrine

meso

beta

N

NH N

HN

TMP, Tetramesitylporphyrin

N

NH N

HN

TPP, Tetraphenylporphyrin

Shinokubo, H.; Osuka, A. Chem. Commun. 2009, 1011.

Scope











Minor sites

1

2, 49%

7

8, 46% dr=6:1

C4 14%

2

3, 51%

8

9, 44% dr=8:1

C4 12%

3

4, 55%

9

10, 42%

C3 11%

4

5, 53% 1:1.4

10

11, 51% dr=1.5:1

C2 <2%

5 6, 49%

exo: endo=5.7

11 12a, 30%

cis/trans=1:1

12b, C3 27%

cis/trans=2:1

6

7, 2:1

12

13, 49% dr=1.6:1

C3 9%



REPORTS

on

Oct

ober

5, 2

012

www.

scie

ncem

ag.o

rgDo

wnlo

aded

from

Substrate Product Yield (%)

49

51

53

49

F

F

FF

Fexo:endo= 5.7

10

1: 1.4

Compare to:

R-HMn(TPP)Cl (2.0 mol%), NaOCl (2.0mL, 0.33M)

TBACl (4.0 mol%) DCM, 12h, rt

(2.0 mmol)R-Cl

Cl Br

49 %57%

Cl

74 %

R-HPhIO (6.0-8.0 equiv), AgF (4.5 mmol)

Mn(TMP)Cl (6.0-8.0 mol%), TBAF.3H2O (0.3 equiv)3:1 MeCN: DCM, 6-8 h, 50oC

(1.5 mmol)R-F

Cl

31 %

with Mn(TMP)Cl

with NaOBr

Liu, W.; Groves, J. T. J. Am. Chem. Soc. 2010, 132, 12847–12849, (Liu, W.; Huang, X.; Cheng, M. J.; Nielsen, R. J.; Goddard, W. A.; Groves, J. T. Science 2012, 337, 1322–1325.

Scope


Mn(TMP)Cl (6.0-8.0 mol%, TBAF.3H2O (0.3 equiv)(1.5 mmol)R-F











Minor sites

1

2, 49%

7

8, 46% dr=6:1

C4 14%

2

3, 51%

8

9, 44% dr=8:1

C4 12%

3

4, 55%

9

10, 42%

C3 11%

4

5, 53% 1:1.4

10

11, 51% dr=1.5:1

C2 <2%

5 6, 49%

exo: endo=5.7

11 12a, 30%

cis/trans=1:1

12b, C3 27%

cis/trans=2:1

6

7, 2:1

12

13, 49% dr=1.6:1

C3 9%



REPORTS

on

Oct

ober

5, 2

012

www.

scie

ncem

ag.o

rgDo

wnlo

aded

from


46 + 14 (60)

44 + 12 (56)

42 + 11 (53)

O OMe O OMe

F

OH OH

F

O O

F

dr=6:1

dr=8:1

O OMe

FOH

F

O

F

11


30 + 27 (57)

49 + 9 (58)

51 + trace

OAc OAc

F

OBz OBz

F

cis:trans=1:1

dr=1.6:1

N

O

F3C N

O

F3C

Fdr=1.5:1

N

O

F3C

F

OAc

F

cis:trans=2:1

OBz

F

More Complex Substrates...(Fig. 2B). The lowest unoccupied molecular or-bitals in a low-spin, d2 oxoMnV complex are ex-pected to be the two orthogonal Mn-O p* orbitals,whichwould direct the approach of the scissile C-Hbond into a bent p*-approach trajectory (29, 41).

We conducted a number of experiments toexamine this mechanistic hypothesis. Initial C-Hhydroxylation was ruled out by controls showingthat cyclohexanol was oxidized to cyclohexanoneunder these conditions. No cyclohexylfluoridewas detected. Also, the hydroxyl group of 1-methylcyclohexanol is stable to the reaction condi-tions (Table 1, entry 8). Deuterium kinetic isotopeeffects (KIEs) were evaluated by the reaction of a1:1 mixture of cyclohexane and cyclohexane-d12,producing an intermolecular competitive KIEof 6.1. A similar value (5.7) was observed with amixture of ethylbenzene and ethylbenzene-d10.The large KIE indicates that C-H bond cleavageis the rate-limiting step in the reaction, consistentwith typical manganese porphyrin–catalyzed hy-droxylation reactions. Furthermore, reaction ofnorcarane, a diagnostic radical clock substrate (2),afforded 2-fluoronorcaranes and a significantamount of the rearranged fluorinated product, 3-fluoromethylcyclohexene (7), which is indicativeof a C radical ring-opening process (Table 1, entry6). The 2:1 ratio of these cyclopropylcarbinyl andhomoallyl fluorides indicates a short radical life-time of 2.5 ns and rapid trapping of the substrateradicals, given the ring-opening rate constant forthe 2-norcaranyl radical of 2 ! 108 M!1 s!1 (42).We have shown that the chlorination of norcaranewith t-butyl hypochlorite, also involving diffus-ing C radical intermediates, gave a similar ratio ofrearranged and unrearranged products. Further,the yields of alkyl fluorides were reduced whenthe reactions were run in air, indicating substrateradical trapping by O2.

The identification of trans-difluoroMnIV(TMP)as the likely fluorinating agent was made possi-ble by its isolation and structural characteriza-tion. We were able to obtain pure crystals of theMnIV(TMP)F2 by treating MnIV(TMP)Cl2 (43)with excess AgF. The molecular structure of thiscompound showed two axially bound fluorideions with F-MnIV-F bond lengths of 1.7931(17)and 1.7968(16) Å (Fig. 2C and tables S1 to S5).These distances are very close to those of di-ammonium hexafluoromanganate(IV), the onlyother fluoromanganese(IV) species to be struc-turally characterized to date (44).

We found that stoichiometric amounts ofMnIV(TMP)F2 could replace silver fluoride in asingle-turnover C-H fluorination of cyclooctaneusing Mn(TMP)Cl and iodosylbenzene. A 43%yield of cyclooctyl fluoride was obtained basedon added MnIV(TMP)F2. Thermal decomposi-tion of azo-bis-a-phenylethane to generate thephenethyl radical in the presence of MnIV(TMP)F2 led to a 41% yield of 1-fluoroethylbenzene.These observations indicate that after initial Habstraction, MnIV(TMP)F2 can trap the substrateradicals in the fluorine delivery step (Fig. 2A).The moderate fluorination yields from these rad-

ical trapping experiments are probably due to thefalling concentration of the manganese(IV) di-fluoride under these conditions. Crucial rolesfor silver fluoride in this scenario under cata-lytic conditions are first to convert the addedMn(TMP)Cl to the manganese(III) fluoride formof the catalyst and then to replenish the inventoryofmanganese(IV) fluoride during turnover (26,27).Although a direct reaction between the substrate

radicals and AgF might also be considered, thereaction between AgF and phenethyl radicals gen-erated in situ from azo-bis-a-phenylethane af-forded only trace amounts of fluorinated products.

We have explored the potential energy land-scape and electronic structures of the intermedi-ates and transition states proposed in Fig. 2Ausing density functional theory (DFT) computa-tions and a polarizable continuum solvationmodel.

MnV

L

H

approachA B

C

MnIV(TMP)F2 MnIV(TMP)F2

Fig. 2. (A) Posited catalytic cycle for manganese porphyrin–catalyzed C-H fluorination reactions. (B)Inferred stereoelectronics for H abstraction. (C) X-ray crystal structure of trans-MnIV(TMP)F2 drawn at 50%probability of the electron density. Highlighted atoms are F (yellow), Mn (magenta), and N (blue) (H atomsare omitted for clarity).

OH

Me

Me

Me

O

OMe

42% / =3.1/ =7.8

16%C3-fluoride

F

O

H

H3.5

FF

: 1

A

B

51%H

H

Mn(TMP)Cl (6 mol%)AgF (3 equiv.) TBAF (0.3 equiv.)


PhIO (6 equiv.)

PhIO (10 equiv.)

H

26 unactivated sp3 C-H bonds

Methylene selectivity

O

H

A B

C D

H

Me OMe

Me OMe

F

32% 23%/ =4.5 / =6.2F

C

H


PhIO (8 equiv.)

Selective A ring fluorination

OAc

D

H


PhIO (8 equiv.)

P450-like selectivity

OAcF

57%

23

Fig. 1. Manganese porphyrin–catalyzed selective C-H fluorina-tions. Yields were determined bythe integration of gas chroma-tography traces using naphtha-lene as the internal standard. (A)

Methylene-selective fluorination of trans-decalin. (B) Selective fluorination of sclareolide. (C) SelectiveA-ring fluorination of 5a-androstan-17-one. (D) Selective 5-exo-fluorination of bornyl acetate.


REPORTS

on

Oct

ober

5, 2

012

www.

scie

ncem

ag.o

rgDo

wnlo

aded

from

(Fig. 2B). The lowest unoccupied molecular or-bitals in a low-spin, d2 oxoMnV complex are ex-pected to be the two orthogonal Mn-O p* orbitals,whichwould direct the approach of the scissile C-Hbond into a bent p*-approach trajectory (29, 41).







MnV

L

H

approachA B

C



OH

Me

Me

Me

O

OMe

42% / =3.1/ =7.8

16%C3-fluoride

F

O

H

H3.5

FF

: 1

A

B

51%H

H



PhIO (6 equiv.)

PhIO (10 equiv.)

H



O

H

A B

C D

H

Me OMe

Me OMe

F

32% 23%/ =4.5 / =6.2F

C

H


PhIO (8 equiv.)


OAc

D

H


PhIO (8 equiv.)


OAcF

57%

23




REPORTS

on

Oct

ober

5, 2

012

www.

scie

ncem

ag.o

rgDo

wnlo

aded

from








MnV

L

H

approachA B

C



OH

Me

Me

Me

O

OMe

42% / =3.1/ =7.8

16%C3-fluoride

F

O

H

H3.5

FF

: 1

A

B

51%H

H



PhIO (6 equiv.)

PhIO (10 equiv.)

H



O

H

A B

C D

H

Me OMe

Me OMe

F

32% 23%/ =4.5 / =6.2F

C

H


PhIO (8 equiv.)


OAc

D

H


PhIO (8 equiv.)


OAcF

57%

23




REPORTS

on

Oct

ober

5, 2

012

www.

scie

ncem

ag.o

rgDo

wnlo

aded

from








MnV

L

H

approachA B

C



OH

Me

Me

Me

O

OMe

42% / =3.1/ =7.8

16%C3-fluoride

F

O

H

H3.5

FF

: 1

A

B

51%H

H



PhIO (6 equiv.)

PhIO (10 equiv.)

H



O

H

A B

C D

H

Me OMe

Me OMe

F

32% 23%/ =4.5 / =6.2F

C

H


PhIO (8 equiv.)


OAc

D

H


PhIO (8 equiv.)


OAcF

57%

23




REPORTS

on

Oct

ober

5, 2

012

www.

scie

ncem

ag.o

rgDo

wnlo

aded

from

12

-Similar reactivity and selectivities were observed for aliphatic chlorinationfor reactions A and B

Newhouse, T.; Baran, P. S. Angew. Chem. Int. Ed. 2011, 50, 3362–3374.

Mechanistic Studies

13

Control Result

without catalyst and iodosylbenzene no products

without TBAF2:1

oxygenated:fluorinated products

without AgFonly oxygenated

products

FF

Substrate Product

2:1


Mn(TMP)Cl (6.0-8.0 mol%, TBAF.3H2O (0.3 equiv)(1.5 mmol)R-F

Mechanistic Clues Result

Reaction of cyclohexanol gives cyclohexanone (no fluorinated substrate detected).

Initial hydroxylation does not occur

Intermolecular competitive KIE= 6.1 for cyclohexane and deuterated cyclohexane

C-H cleavage is the rdsParallel to other Mn porphyin catalyzed hydroxylation reactions

Rearranged product observed in norcarane radical clock experiment

Proof of C radical ring opening

Proposed Catalytic Cycle








MnV

L

H

approachA B

C



OH

Me

Me

Me

O

OMe

42% / =3.1/ =7.8

16%C3-fluoride

F

O

H

H3.5

FF

: 1

A

B

51%H

H



PhIO (6 equiv.)

PhIO (10 equiv.)

H



O

H

A B

C D

H

Me OMe

Me OMe

F

32% 23%/ =4.5 / =6.2F

C

H


PhIO (8 equiv.)


OAc

D

H


PhIO (8 equiv.)


OAcF

57%

23




REPORTS

on

Oct

ober

5, 2

012

www.

scie

ncem

ag.o

rgDo

wnlo

aded

from

14

1. MnIII (TMP)(F) is formed in situ.2. This is oxidized to O=MnIV(TMP)(F) by iodosylbenzene.3. H-atom abstraction occurs, generating the alkyl radical and MnIV(OH)(F) rebound intermediate.4. AgF interacts with the catalyst to form MnIV(TMP)(F2), the fluorinating reagent.5. F-atom transfer to the alkyl radical produces the fluorinated product, regenerating the catalyst.

-Similar to proposed Oxygen Rebound Mechanism observed in metalloporphyrin catalyzed C-H oxidations.








MnV

L

H

approachA B

C



OH

Me

Me

Me

O

OMe

42% / =3.1/ =7.8

16%C3-fluoride

F

O

H

H3.5

FF

: 1

A

B

51%H

H



PhIO (6 equiv.)

PhIO (10 equiv.)

H



O

H

A B

C D

H

Me OMe

Me OMe

F

32% 23%/ =4.5 / =6.2F

C

H


PhIO (8 equiv.)


OAc

D

H


PhIO (8 equiv.)


OAcF

57%

23




REPORTS

on

Oct

ober

5, 2

012

www.

scie

ncem

ag.o

rgDo

wnlo

aded

from

-Crystal structure indicated F-MnIV-F bond lengths to be close to diammonium hexafluoromanganate (IV), only fluoromanganese species to be characterized to date.

MnIV(TMP)F2MnIV(TMP)Cl2 benzene, rt, 2hr84% yield

AgF (1.6 mmol)

0.033mmol

Proof of MnIV (TMP) F2

15

F

43%

PhIO (0.05 mmol), MnIV (TMP)F2 (0.034 mmol)

Mn(TMP)Cl (0.034mmol), TBAF.3H2O (0.3mmol)3:1 MeCN: DCM, 30 min, 50oC1.5 mmol

Single-turnover experiment

MnIV(TMP)F2MnIV(TMP)Cl2benzene, rt, 2hr

AgF (1.6 mmol)

0.033mmol

NN

105oC, 4 min

F41%

0.4 equiv

Thermal Decomposition Experiment

-Use of AgF under the same conditions gave trace amounts of product.

Concluding Remarks

-Liu et al. developed a direct aliphatic C-H fluorination process using a Mn porphyrin system.-This process showed that the traditional Oxygen Rebound Mechanism could be redirected to deliver F-atom.-Process gave moderate to good yields on a wide range of substrates including a complex terpenoid and a steroid (late-stage fluorination and 18F incorporation applications)-Characterized a Mn IV (TMP)(F2) species capable of capturing substrate radicals. -This type of F-atom transfer is unprecedented! -Mechanistic studies supported by DFT calculations identified this Mn IV (TMP)(F2) complex as the fluorinating agent.-Potential for other transition metal fluorides?

-Overall: 1. Fluorination= hot topic! 2. Porphyrins= fun and super cool ligands! 3. Manganese= a metal that does not get enough attention! 4. There is both a lack and need for more catalytic fluorination methods. Exciting field to jump into!

16

“The best scientist is open to experience and begins with romance - the idea that anything is possible.” ― Ray Bradbury

http://www.goodreads.com/author/show/1630.Ray_Bradbury

http://www.goodreads.com/author/show/1630.Ray_Bradbury

Oxidative Aliphatic C-H Fluorination with Fluoride Ion...

Documents

Transcript of Oxidative Aliphatic C-H Fluorination with Fluoride Ion...