reaction between grignard reagents and …232682/...title reaction between grignard reagents and...

Reaction Between Grignard Reagents and Heterocyclic N-oxides Synthesis of Substituted Pyridines, Piperidines and Piperazines

Hans Andersson

Department of Chemistry Umeå University Umeå 2009

Title Reaction Between Grignard Reagents and Heterocyclic N-oxides – Synthesis of Substituted Pyridines, Piperidines and Piperazines Author Hans Andersson, Department of Chemistry Umeå University, SE-90187, Umeå, Sweden Abstract This thesis describes the development of new synthetic methodologies for preparation of bioactive interesting compounds, e.g. substituted pyridines, piperidines or piparazines. These compounds are synthesized from commercially available, cheap and easily prepared reagents, videlicet the reaction between Grignard reagents and heterocyclic N-oxides. The first part of this thesis deals with an improvement for synthesis of dienal-oximes and substituted pyridines. This was accomplished by a rapid addition of Grignard reagents to pyridine N-oxides at rt. yielding a diverse set of substituted dienal-oximes. During these studies, it was observed that the obtained dienal-oxmies are prone to ring-close upon heating. By taking advantage of this, a practical synthesis of substituted pyridines was developed. In the second part, an ortho-metalation of pyridine N-oxides using Grignard reagents is dis-cussed. The method can be used for incorporation of a range of different electrophiles, includ-ing aldehydes, ketones and halogens. Furthermore, the importance for incorporation of halo-gens are exemplified through a Suzuki–Miyaura coupling reaction of 2-iodo pyridine N-oxides and different boronic acids. Later it was discovered that if the reaction temperature is kept below -20 °C, the undesired ringopening can be avoided. Thus, the synthesis of 2,3-dihydropyridine N-oxide, by reacting Grignard reagents with pyridine N-oxides at -40 °C followed by sequential addition of aldehyde or ketone, was accomplished. The reaction pro-vides complete regio- and stereoselectivity yielding trans-2,3-dihydropyridine N-oxides in good yields. These intermediate products could then be used for synthesis of either substituted piperidines, by reduction, or reacted in a Diels–Alder cycloaddtion to give the aza-bicyclo compound. In the last part of this thesis, the discovered reactivity for pyridine N-oxides, is applied on pyrazine N-oxides in effort to synthesize substituted piperazines. These substances are ob-tained by the reaction of Grignard reagents and pyrazine N-oxides at -78 °C followed by reduction and protection, using a one-pot procedure. The product, a protected piperazine, that easily can be orthogonally deprotected, allowing synthetic modifications at either nitrogens in a fast and step efficient manner. Finally, an enantioselective procedure using a combination of PhMgCl and (-)-sparteine is discussed, giving opportunity for a stereoselective synthesis of substituted piperazines. Keywords Grignard reagents, pyridine N-oxide, pyrazine N-oxide, dienal-oxime, pyridine, ortho-metalation, piperidine, piperazine, asymmetric

Cathis & Gunnar

Contents List of Papers ................................................................................................. 1 Abbreviations................................................................................................. 3

1. Introduction .................................................................................................... 5

2. A stereodefined synthesis of substituted dienal-oximes .............................. 9 2.1 Introduction.............................................................................................. 9 2.2 Rapid addition of Grignard reagents to pyridine N-oxides................. 10

2.2.1 Synthesis of substituted dienal-oximes ........................................ 10 2.2.2 Addition of alkyl Grignard reagents. ........................................... 11 2.2.3 Further transformations of dienal-oximes.................................... 12

2.3 Conclusion and outlook ........................................................................ 14

3. Complete regioselective synthesis of substituted pyridines ...................... 16 3.1 Introduction............................................................................................ 16

3.1.1 Organometallic addition to activated pyridines........................... 17 3.2 Regiospecific synthesis of substituted pyridines................................. 18

3.2.1 Synthesis of 2-substituted and 2,4-disubstituted pyridines......... 19 3.2.2 Reaction with 2- and 3-substituted N-oxides............................... 20 3.2.3 Synthesis of unsymmetrical 2,6-disubstituted pyridines ............ 21 3.2.4 One-pot synthesis of 4,2-disubstituted pyridines ........................ 22

3.3 Synthesis of 4-pyridones....................................................................... 23 3.3.2. Synthesis of 4 aminopyridines..................................................... 25


4. Synthesis of 2-substituted pyridine N-oxides via directed ortho-metalation (DOM)............................................................................................................... 28

4.1 Introduction............................................................................................ 28 4.1.1 Directed ortho-metalation reaction .............................................. 28

4.2 Metalation of pyridine N-oxides........................................................... 29 4.2.1 Directed ortho-metalation using Grignard reagents.................... 30 4.2.2 DOM comparison between Grignard and lithium reagents ........ 31

4.3 Suzuki–Miyaura coupling of 2-iodo pyridine N-oxides ..................... 34 4.4 The direct coupling of the metalated pyridine N-oxides..................... 35 4.4 Conclusion and outlook ........................................................................ 37

5. Synthesis of trans-2,3-dihydropyridine N-oxides and piperidines ........... 40 5.1 Introduction............................................................................................ 40 5.2 Synthesis of trans-2,3-dihydropyridine N-oxides ............................... 42 5.3 Synthesis of piperidines ........................................................................ 46

5.3.1 Synthesis of 2-substituted piperidines.......................................... 46

5.3.2 Synthesis of trans-2,3 disubstituted piperidines.......................... 47 5.4 Conclusion and outlook ........................................................................ 47

6. Grignard addition to pyrazine N-oxides: towards an efficient enantioselective synthesis of substituted piperazines .................................... 50

6.1 Introduction............................................................................................ 50 6.1.2 Pyrazine N-oxide compared to pyridine N-oxide ........................ 51

6.2 One-pot synthesis of 2-substituted piperazines ................................... 51 6.3 Orthogonal deprotection ....................................................................... 54 6.4 Enantioselective ssynthesis of substituted piperazines ....................... 55

6.4.1 Chiral induction with ligand in combination with Grignard ...... 55 6.4.2 Enantioselective synthesis of substituted piperazines................. 57


7. Concluding Remarks.................................................................................... 60

8. Acknowledgements ...................................................................................... 63

1

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.*

I Reaction of pyridine N-oxides with Grignard reagents: a stereode-

fined synthesis of substituted dienal-oximes Andersson, H.; Wang, X.; Björklund M.; Olsson, R.; Almqvist, F. Tet. Lett, 2007, 48, 6941-6944.

II Synthesis of 2-Substituted Pyridines via a Regiospecific Alkyla-tion, Alkynylation, and Arylation of Pyridine N-Oxides Andersson, H.; Almqvist, F.; Olsson R. Org. Lett, 2007, 9 1335–1337.

III Selective synthesis of 2-substituted pyridine N-oxides via directed ortho-metalation using Grignard reagents Andersson, H.; Gustafsson, M.; Olsson R.; Almqvist, F. Tet. Lett, 2008, 49, 6901-6903. (Highlighted in Synfacts, 2009, 1)

IV The regio- and stereoselective synthesis of trans-2,3-dihydropyridine N-oxides and piperidines Andersson, H.; Gustafsson, M.; Olsson, R.; Almqvist, F. Angew. Chem. Int. Ed. 2009, 48, 3288-3291. (Highlighted in Synfacts, 2009, 7)

V Complete regioselective addition of Grignard reagents to pyrazine N-oxides, towards an efficient enantioselective synthesis of substi-tuted piperazines Andersson H.; Thomas, S.L.; Das, S.; Gustafsson, M.; Olsson R.; Almqvist F. Manuscript, 2009.

Reprints were made with permission from respective publishers.

*The author’s contributions to paper I-V have been: the formation of the projects and research problems, experimental work and contribution in writing the manuscripts.

3

Abbreviations acac acetylacetonate Boc tert-butoxycarbonyl DDQ 2,3-dichloro-5,6-dicyano-1,4-benzoquinone DIPEA diisopropylamine DMD dimethyldioxirane DMF N,N-dimethylformamide DMG direct metalation group DOM directed ortho-metalation DOS diversity-oriented synthesis dr diastereomeric ratio ee enantiomeric excess equiv. equivalent HPLC high performance liquid chromatography hmbc heteronuclear multiple bond coherence i-Pr iso-propyl LC liquid chromatography LiDMAE lithium 2-(dimethylamino)ethoxide mCPBA meta-chloroperoxybenzoic acid Me methyl MeCN acetonitrile MeOH methanol MS mass spectrometry MW microwave n-Bu normal-butyl NMP N-methyl-2-pyrrolidone NMR nuclear magnetic resonance NOESY nuclear overhauser effect spectroscopy OBn benzyloxy Ph phenyl rt. room temperature TCT trichlorotriazine TEA triethylamine THF tetrahydrofuran TLC thin layer chromatography TMEDA N,N,N’,N’-tetramethylethyldiamine UHP urea hydrogen peroxide Å ångström

5

1

Introduction

In recent decades there have been exponential advances in organic chemistry that have resulted in the development of large numbers of new methods and improvement of already known. Nevertheless, the medical and materials sciences continue to require novel drugs and other products, hence there are continuing needs for the development of new methods, and the enhancement of current methods, for synthesizing organic compounds. This thesis is based on studies in which new synthetic methodologies have been developed to synthesize molecules that have interesting bioactivities, such as pyridines, piperidines and piperazines.1 These structures are common fragments in both pharmaceuticals and natural products, and a number of synthetic methods for their preparation are known today. However, the focus in this thesis is on organometallic additions to an already existing, activated pyridine or pyrazine ring. Direct attempts to react organometallic reagents with pyridines are often troublesome and require harsh conditions. To circumvent this problem, acyl-activated pyridines have commonly been used. This strategy has been well studied, but the reported methods often suffer from drawbacks such as the formation of regioisomers or the need for multi-step synthetic protocols. However, as an alternative to these starting materials, we have studied the reaction between pyridine N-oxides 2 and Grignard reagents. Pyridine N-oxide 2 is a cheap, commercially available and bench-stable starting material. Furthermore, differently substituted pyridine N-oxides 2 are easily obtained from the oxidation of pyridines 1, by one of several pub-lished methods (Scheme 1.1).2 Given the number of commercially available pyridines, these methods have the potential to generate a vast diversity of substituted pyridine N-oxides, which potentially could be used as starting materials for multitude of target molecules.

1For review of pyridines (a) Henry, G. D. Tetrahedron 2004, 60, 6043-6061. piperidines (b) Buffat, M. G. P. Tetrahedron 2004, 60, 1701-1729. piperazines (c) Leclerc, J.-P.; Fagnou, K. Angew. Chem. Int. Ed. 2006, 45, 7781-7786. 2For different oxidation procedures see: Youssif, S. Arkivoc 2001, 2, 242-268.

6

N

R

N

O

A; m-CPBAB; UHP

C; MeReO3, H2O2

D; DMD1

2, 80% to quant.

R = OBn, CF3, CN, F, OMe, NO2 etc.

R

Scheme 1.1. Methods for oxidizing pyridines to pyridine N-oxides. Although the pyridine N-oxide 2 has structural resemblance to pyridine 1, the reactivity of the two compounds differs significantly. For example, pyri-dine 1a reacts inefficiently upon nitration, even at high temperatures, and only 6% of the corresponding 3-nitro pyridine is obtained (Scheme 1.2).3 In contrast, the nitration of pyridine N-oxide 2a proceeds smoothly and the corresponding 4-nitro pyridine N-oxide is obtained with a yield of 95% (Scheme 1.2).4 In addition to its lower reactivity, pyridine yields the 3-nitro isomer while the pyridine N-oxide gives the 4-substituted product (Scheme 1.2). Furthermore, the differences in reactivity are reflected in the results of reacting pyridine or pyridine N-oxides with nucleophilic reagents. While pyridine 1a reacts slowly, or requires harsh conditions to react with phenyl lithium (PhLi),5 phenylmagnesium chloride (PhMgCl) readily add to pyri-dine N-oxide 2a, even at -40 °C. Although the main product from the addi-tion occurs at the same position, the 2-substituted pyridine 1c is obtained with PhLi, while the reaction between PhMgCl and pyridine N-oxide gives the ring-opened dienal-oxime 3a (Scheme 1.2).6

N

HNO3, H2SO4

300 oC, 24 h N

NO2

3-nitropyridine, 6%

N

PhLi

Toluene, 110 oC N Ph

1c, 50%

N

O

HNO3, H2SO4

90 oC, 1h N

O

NO2

4-nitropyridine N-oxide, 70%

N

O

PhMgCl, THF

-40 oC to rt N Ph

OH

3a, 45%

prior knowledge before the start of this project

1a

1a

2a

2a

Scheme 1.2. Examples of differences in reactivities of pyridines and pyridine N-oxides.

Furthermore, the distinctive reactivity of pyridine N-oxides can be under-stood by looking at the different possible resonance structures (Figure 1.1). Because of the mesomeric electron release from the oxygen, pyridine N-oxides can easily be reacted with electrophiles, as illustrated in Scheme 1.2

3Joule, J. A.; K., M. Heterocyclic Chemistry 4:th Ed., 75-76. 4Cislak, F. E. J. Ind. Eng. Chem. (Washington, D. C.) 1955, 47, 800-802. 5Evans, J. C. W. A., C.F.H Organic Syntheses, Coll. 1943, 2, 517. 6Kellogg, R. M.; Van Bergen, T. J. J. Org. Chem. 1971, 36, 1705-1708.

7

(structures II and III, Figure 1.1). The situation is some subtle as the same positions are also activated towards nucleophilic additions, however, after coordination to the N-oxygen the electrophilicity of pyridine N-oxide is more pronounced (structures IV and V, Figure 1.1).

N

O

N

O

N

O

N

O

N

O

etc.

I II III IV V Figure 1.1. Different resonances of pyridine N-oxides.

Our interest in the reactivity of pyridine N-oxides and their potential as start-ing materials for the synthesis of a multitude of target molecules, led us to revisit the reaction between Grignard reagents and pyridine N-oxides. As a result from these studies we have been able to apply the same chemistry to include the synthesis of substituted piperazines, from pyrazine N-oxides and Grignard reagents. Figure 1.2 gives a general overview of the chemical transformations discussed in this thesis.

N

X

O

R

N R'

R

N E

R

N R'

E

O

R

N

N

OH

Boc

R'

O

N R'

OH

R

Cahpter II

Chapter III

Chapter IVChapter V

Chapter VI

Figure 1.2. General overview of chemical transformations discussed in this thesis.

9

2

A stereodefined synthesis of substituted dienal-oximes

Paper I

2.1 Introduction The first report on the reaction between Grignard reagents and pyridine N-oxide (2a) was published by Colonna and co-workers in 1936.7 They claimed the reaction to yield 2-phenyl pyridine (1c) when PhMgCl was re-acted with pyridine N-oxide (2a) in diethyl ether (Scheme 2.1). The same reaction was later investigated by Kato et al. in 1965, who instead reported the isolation of 1,2-dihydropyridine 4 in a yield of 60-80% (Scheme 2.1).8 As a result Kellog et al. became interested in the structural aspects of the reaction and therefore reinvestigated the Grignard addition to pyridine N-oxides in 1971. Instead of 2-phenylpyridine (1c) or 1,2-dihydropyridine (4), they reported the isolation of ring-opened dienal-oxime (3a) in 45% yield (Scheme 2.1).6

N

O

PhMgX

X= Cl, Br

N

OH

Ph N Ph

N

OH

Reported by ColonnaReported by Kato

Reported by Kellog

Ph

2a

3a

Et2O

1c

PhMgX

THF

4

PhMgX

THF

Scheme 2.1. Reported reactions between pyridine N-oxides and Grignard reagents.

7Colonna, M. Chem Abstr. 1936, 30, 3420-3421. 8Kato, T.; Yamanaka, H. J. Org. Chem. 1965, 30, 910-913. 6Kellogg, R. M.; Van Bergen, T. J. J. Org. Chem. 1971, 36, 1705-1708.

10

2.2 Rapid addition of Grignard reagents to pyridine N-oxides In 2003, Almqvist and co-workers published a method for the synthesis of 4-substituted piperidines, based on copper-catalyzed organozinc addition to N-acyl pyridinium salts.9 During these studies, the potential utility of pyridine N-oxides for the synthesis of 2-substituted piperidines was investigated. The use of organozinc reagents or organolithium reagent in combination with pyridine N-oxides resulted in no reaction and in a complex mixture, respec-tively. However, when Grignard reagents were used, a ring-opened product was observed, and later the dienal-oxime 3a was confirmed by structural elucidation using NMR (Figure 2.1). Furthermore, it could be conducted that only one isomer was formed suggesting the reaction to be a conserted elec-trocyclic ring-opening reaction. (Figure 2.1).

N

O

PhMgCl

THF, rtN

OClMg

possible intermediate

N

OH3a

2a

5-phenyl-(1E,2Z,4E)-penta-2,4-dienal-oxime Figure 2.1. Electrocyclic ring-opening to dienal-oxime 3a. 2.2.1 Synthesis of substituted dienal-oximes The formation of dienal-oximes, with a defined diene-system and oxime functionality present in the structure, appeared as an attractive intermediate for further transformations. Therefore, the reaction between Grignard rea-gents and pyridine N-oxides was studied further. When PhMgCl was added slowly to pyridine N-oxide (2a) at -40 °C, only a moderate 38% yield of dienal-oxime 3a was isolated. However, if the addition rate was increased and the reaction performed at room temperature (rt.), the yield of dienal-oxime 3a increased to 85% (entry 1, Table 2.1). This protocol was therefore used to prepare a small set of dienal-oximes 3a-3l starting from differently substituted pyridine N-oxides 2a-2g (Table 2.1). 2-substituted and 4-substituted pyridine N-oxides 2b, 2d-2g reacted smoothly with aryl and alkynyl Grignard reagents to form dienal-oximes 3b, 3d-3l in yields between 66-95% (entries 2, 4-12, Table 2.1). However, when the reaction between PhMgCl and 3-picoline N-oxide (2c) was performed, dienal-oxime was not observed by crude-NMR or isolated (entry 2, Table 2.1). Instead, the 2,3-disubstituted pyridine was obtained in a 43% yield (see Chapter 3 for a more detailed discussion).

9Wang, X.; Kauppi, A. M.; Olsson, R.; Almqvist, F. Eur. J. Org. Chem. 2003, 4586-4592.

11

Table 2.1. Synthesis of 3,5-substituted dienal-oximes.

N

R3

R4MgX

THF, rt

R4

N

OH

R3

O

2a-g3a-l

X= Br or Cl

R1

R2 R2

R1

entry N-oxide R1 R2 R3 R4 oxime yield (%)a

1 2a H H H Ph 3a 85 2 2b Me H H Ph 3b 79 3 2c H Me H Ph 3c 0b

4 2d H H Me Ph 3d 81 5 2e H H Ph Ph 3e 83 6 2e H H Ph PhCC 3f 77c

7 2e H H Ph 2-Thiophene 3g 66 8 2f H H OBn Ph 3h 95 9 2f H H OBn Naphthyl 3i 83

10 2f H H OBn PhCC 3j 78c

11 2f H H OBn 2-Thiophene 3k 76 12 2g H H Cl Ph 3l 86

Reactions conditions: pyridine N-oxide (1 equiv.) in THF, Grignard reagent 1.2 (equiv.) at rt. aIso-lated yields. bNo dienal-oxime observed instead the corresponding 2,3-substituted pyridine was iso-lated. cTo consume starting material the reaction mixture was heated gently. PhCC = phenylethynyl.

2.2.2 Addition of alkyl Grignard reagents. To broaden the scope for the synthesis of dienal-oximes, we studied the pos-sibility of synthesizing 5-alkyl substituted dienal-oximes by using alkyl-Grignard reagents. Iso-propylmagnesium chloride (i-PrMgCl) and methyl-magnesium chloride (MeMgCl) were reacted with pyridine N-oxides 2a and 2e as previously described. Unfortunately only low yields (< 10%) were isolated after aqueous work-up and purification by column chromatography. This was surprising since LC-MS analysis indicated that a large amount of the product was present in the crude reaction mixture, together with only minor amounts of by-products. Although this could have been due to differ-ences in the MS response factors of the products, it indicated that the alky-lated dienal-oxime formed in the reaction is less stable than the previously isolated 5-aryl dienal-oximes (Table 2.1). Therefore, we set out to perform a transformation of the ring-opened dienal-oxime into a potentially more sta-ble product before purification. A well-known transformation of oximes is the Beckmann rearrangement.10 Whereas ketoximes result in the correspond-ing amides, aldoxime normally gives the corresponding nitrile by the elimi-

10Smith, M. B. M., J. Advanced Organic Chemistry, 5th ed.; John Wiley &Sons: New York, 2001; pp 1415.

12

nation of water. Since this is quite a straightforward transformation, and the corresponding nitrile is, potentially, a more stable product, we investigated the possibility of transforming the aldoximes into the corresponding nitriles before purification. Initially, the reaction was performed using trichlorotriaz-ine (TCT) in dimethylformamide (DMF), followed by addition of dienal-oxime 3e.11 The reaction was complete after 30 minutes at rt. and the nitrile 5a was isolated in a 59% yield. However, using the commercially available Vilsmeier salt 6 gave a more straightforward protocol and a cleaner reac-tion.12 Salt 6 was dissolved in dichloromethane (DCM) and added to the crude mixture obtained after Grignard addition, which increased the yield of isolated nitrile 5a to 74% (entry 1,Table 2.2). The reaction was also per-formed on alkylated dienal-oximes, and the nitriles 5b and 5c from the addi-tion of i-PrMgCl and MeMgCl to pyridine N-oxides 2a and 2e, were isolated with yields of 49% and 64%, respectively (entries 2 and 3, Table 2.2). Table 2.2. Conversion to nitriles.

N

R

O

R'MgCl

THF, rt

6, DCM, rt R'

CN

R

2a, e

5a-c

N

Cl

H

Cl6 =

entry N-oxide R R’ nitrile yield(%)a

1 2a H Ph 5a 74b

2 2a H CH(Me)2 5b 49 3 2e Ph Me 5c 64

Reaction conditions: pyridine N-oxide (1 equiv.) in THF, Grignard rea-gent (1.2 equiv.) at rt. Vilsmeier salt (2 equiv.) in DCM at rt. aIsolated yields. bIsolated as a mixture of cis-trans and trans-trans-diene in 20:3 ratio, respectively.

2.2.3 Further transformations of dienal-oximes. Having developed this efficient method for the preparation of dienal-oximes, we addressed the possibility that we might be able to transform the interme-diate products further into other compounds. If possible, this could constitute a platform for diversity-oriented synthesis (DOS)13 – a strategy based on small molecules with potential for further transformations giving a range of diverse structures. Basically, there are two ways for planning DOS path-ways, the reagent-based and the substrate-based approach. The substrate-based approach take advantage of different starting materials, whereas in the

11De Luca, L.; Giacomelli, G.; Porcheddu, A. J. Org. Chem. 2002, 67, 6272-6274. 12Vilsmeier, A.; Haack, A. Ber. Dtsch. Chem. Ges. B 1927, 60B, 119-122. 13Burke, M. D.; Schreiber, S. L. Angew. Chem., Int. Ed. 2004, 43, 46-58.

13

reagent-based approach the same starting material is treated with different reagents to generate a set of diverse compounds. In our case the reagent-based strategy would be the most suitable for planning DOS via dienal-oximes, starting from pyridine N-oxides (Figure 2.2).

Y

X

Z

X,Y and Z being different reagents

Figure 2.2. Reagent-based approach towards DOS.

Perhaps the most straightforward transformation would be the synthesis of the corresponding saturated primary amine 9 (Scheme 2.2). Reduction of 3e, using the practical hydrogen transfer method, palladium on charcoal, Pd/C, with ammonium formate in methanol (MeOH), gave the secondary amine 7 in a 48% yield and not the expected primary amine 9 (Scheme 2.2).14 A higher yield (62%) of the secondary amine 7 was obtained when using t-BuNH2-BH3 and Pd/C; but again, the primary amine 9 was not observed (Scheme 2.2).15 To synthesize the primary amine, a two-step protocol was tested: reduction of the oxime 3e followed by hydrogenation to give the saturated primary amine 9 (Scheme 2.2). Nucleophilic hydride reagents such as NaCNBH3

16 and LiAlH4

17 only returned the starting material; and the electrophilic hy-dride reagent BH3-SMe2

18 gave complex reaction mixtures. Our attention was therefore turned instead to zinc dust in acetic acid, a method often used to reduce nitroso functionalities to amines.19 Zn dust in acetic acid gave a rapid and clean conversion to amine 8 without reduction of the double bonds (Scheme 2.2). Although the E/Z isomers 8a and 8b were obtained in a 2:1 mixture, the subsequent reduction gave the primary saturated amine 9 in a 86% yield calculated over two steps (Scheme 2.2). Reacting 4-benzyloxypyridine N-oxide (2f) with PhMgCl gave 3-benzyloxy substituted dienal-oxime 3h in a 95% yield (entry 8, Table 2.1). Compound 3h can be considered as a masked enaminone, hence debenzylation followed by reduction of the oxime functionality would render the enaminone. Enaminones are versatile intermediates that combine the nucleophilic prop-

14Searcey, M.; Grewal, S. S.; Madeo, F.; Tsoungas, P. G. Tetrahedron Lett. 2003, 44, 6745-6747. 15Couturier, M.; Tucker, J. L.; Andresen, B. M.; Dube, P.; Brenek, S. J.; Negri, J. T. Tetrahedron Lett. 2001, 42, 2285-2288. 16Gribble, G. W. Chem. Soc. Rev. 1998, 27, 395-404. 17Wang, S. S.; Sukenik, C. N. J. Org. Chem. 1985, 50, 5448-5450. 18Feuer, H.; Braunstein, D. M. J. Org. Chem. 1969, 34, 1817-1821. 19Hatanaka, M.; Ishimaru, T. J. Med. Chem. 1973, 16, 978-984.

14

erties of enamines and the eletrophilic properties of enones, which makes them very important intermediates for the synthesis of various heterocyclic compounds.20 To our delight no dimerization – which had been seen in pre-vious reductions, i.g. 3e (Scheme 2.2) – was observed when 3h was sub-jected to Pd/C and ammonium formate reduction, and the enaminone 10 was isolated in a 78% yield (Scheme 2.2). To demonstrate further transforma-tions using this intermediate, enaminone 10 was reacted with hydrazine hy-drate under microwave irradiation to yield pyrazole 11 in a yield of 71%.

A or BPd/C

MeOH, rtPh N

HPh

PhPhA = NH4HCO2, 48%B = t-Bu-NH4-BH3, 62%

7

Zn, dust

AcOH MeOH, rtPh NH2

Ph

9, 86%

PhMgCl,

N

O

R

Ph

N

R

OH

THF, rt.

Ph

NH2

Ph

Ph

Ph

NH2

8a

8b

Ph

NH2O

10, 78%

Ph

NHN

11, 71%

Pd/C, H2

NH4HCO2

Pd/C

MeOH, rt

N2H2*H2OMeOH

MW, 120 oC

2 min

R = Ph, 3e

R = Ph, 3e

R = OBn, 3h

2e, 2f

3e, 3h

3e R = Ph3h R = OBn

Scheme 2.2. Further transformation of dienal oxime 3e and 3h.

2.3 Conclusion and outlook

In this chapter, the synthesis of dienal-oximes by the addition of Grignard reagents to pyridine N-oxides at rt. has been discussed. The addition of aryl and alkynyl Grignard reagents gave the dienal-oximes in good to excellent yields, whereas an in situ transformation of the resulting oxime to a more stable intermediate, its corresponding nitrile, was necessary after the reaction with alkyl Grignard reagents. In addition, these intermediates can potentially be converted into a diverse set of compounds e.g. nitriles, amines, enami-nones and pyrazoles. Thus, the generation of dienal-oximes from the reac-tion between Grignard reagents and pyridine N-oxides, affords an excellent platform for the design and synthesis of diversity-oriented synthesis (DOS).

20(a) Katritzky, A. R.; Hayden, A. E.; Kirichenko, K.; Pelphrey, P.; Ji, Y. J. Org. Chem. 2004, 69, 5108-5111. (b) Singh, V.; Saxena, R.; Batra, S. J. Org. Chem. 2005, 70, 353-356 and references therein.

16

3

Complete regioselective synthesis of substituted

pyridines

Paper II

3.1 Introduction

Due to the importance of pyridines in bioactive compounds and materials (Figure 3.1), a considerable number of methods have been developed for the synthesis of substituted pyridines.21 Typically, these methods are based on cyclization reactions, i.e. from an aldehyde, 1,5- or 1,3-dicarbonyls, and ammonia: also known as Hantzsch pyridine synthesis.22 However, the addi-tional functionalization of already existing pyridines still poses a significant synthetic challenge. This chapter discusses the synthesis of substituted pyri-dines derived from pyridine N-oxides, a transformation that proceeds via the previously described dienal-oximes (Chapter 2).

N

N

NH2

O

O

O

F N

N

N

OO

N

141312 Figure 3.1. Examples of important pyridines. 12 is a promising new sodium channel inhibitor,23 13 acts as an allosteric antagonist at the metabotropic glutamate receptor mGlu524 and 14 a chiral catalyst used in asymmetric synthesis.25 21Joule, J. A.; K., M. Heterocyclic Chemistry 4:th Ed., 103-110. 22(a) Hantzsch, A. Ann. Chem. 1882, 1, 215. for review see (b) Stout, D. M.; Meyers, A. I. Chem. Rev. 1982, 82, 223-243. 23Shao, B.; Victory, S.; Ilyin, V. I.; Goehring, R. R.; Sun, Q.; Hogenkamp, D.; Hodges, D. D.; Islam, K.; Sha, D.; Zhang, C.; Nguyen, P.; Robledo, S.; Sakellaropoulos, G.; Carter, R. B. J. Med. Chem. 2004, 47, 4277-4285. 24Kulkarni, S. S.; Zou, M.-F.; Cao, J.; Deschamps, J. R.; Rodriguez, A. L.; Conn, P. J.; Hauck Newman, A. J. Med. Chem. 2009, 52, 3563-3575. 25Usuda, H.; Kuramochi, A.; Kanai, M.; Shibasaki, M. Org. Lett. 2004, 6, 4387-4390.

17

3.1.1 Organometallic addition to activated pyridines

Even though pyridine is much more electron-poor than benzene, nucleo-philic additions to pyridines are slow and requires harsh reaction conditions in order to react with organometallic reagents with reasonable efficiency (Chapter 1, Scheme 1.2). One way to address this problem is to activate the pyridine prior to the nucleophilic attack. Consequently, Fraenkel and co-workers presented a method based on ready addition of Grignard reagents to pyridines in the presence of ethylchloroformate to yield the corresponding substituted dihydropyridine intermediates.26 These intermediates were easily oxidized, and the carbamate was hydrolyzed to yield the substituted pyri-dines. Comins and co-workers further developed this strategy into an effi-cient method for the synthesis of substituted pyridines, including activated acyl- and alkyl-pyridines and the use of other organometallic reagents (i.e. Li and Zn) (eq. 1, Scheme 3.1).27 However, the formation of a mixture of re-gioisomers, which results from addition at the 2- or 4-positions, has limited the applicability of this methodology (Scheme 3.1). To achieve selective addition at either position, protection of the unwanted addition site has been necessary, which limits the scope of the method.

N

R Cl

O

N

RO

N

RO

N R'

NMePh

N

R'

NMePh

R'

R'

(eq 1)

(eq 2)DDQ

N R'Ph N

H

O1) Tf2O, pyridine

2) R'MgX, -78 oC

R'MgX

95:5

84% after seperation

Scheme 3.1. Examples of methods for synthesis of substituted pyridines.

Nevertheless, this strategy is attractive, and additional methods using acti-vated pyridine derivatives with directing groups have been reported. For example, Charette and co-workers presented a method in 2001 for the syn-thesis of substituted pyridines and piperidines.28 Their method relies on the stereoselective formation of N-pyridinium imidate from the reaction between amide and pyridine. In this case, the nitrogen imidate lone pair is oriented so

26Fraenkel, G.; Cooper, J. W.; Fink, C. M. Angew. Chem., Int. Ed. 1970, 9, 523. 27(a) Lyle, R. E.; Marshall, J. L.; Comins, D. L. Tetrahedron Lett. 1977, 1015-1018. (b) Comins, D. L.; Abdullah, A. H. J. Org. Chem. 1982, 47, 4315-4319. (c) Comins, D. L.; Stolze, D. A.; Thakker, P.; McArdle, C. L. Tetrahedron Lett. 1998, 39, 5693-5696. (d) Comins, D. L.; Kuethe, J. T.; Hong, H.; Lakner, F. J.; Concolino, T. E.; Rheingold, A. L. J. Am. Chem. Soc. 1999, 121, 2651-2652. 28Charette, A. B.; Grenon, M.; Lemire, A.; Pourashraf, M.; Martel, J. J. Am. Chem. Soc. 2001, 123, 11829-11830.

18

as to direct the addition of an organometallic reagent at the 2-position (eq. 2, Scheme 3.1). In general, the regioselectivity is good, favoring the formation of 1,2-dihydropyridine, which can then be oxidized with 2,3-dichloro-5,6-dicyanobezoquinone (DDQ) to the corresponding 2-substituted pyridine. Despite the large number of reports on the preparation of substituted pyri-dines from organometallic reagents and N-activated pyridines, previously described methods often result in the formation of isomeric mixtures of 2- and 4-substitued products. With this in mind, and the knowledge of the ex-cellent regiocontrolled synthesis of dienal-oximes described in Chapter 2, we focused on the use of these intermediates in the synthesis of substituted pyri-dines.

3.2 Regiospecific synthesis of substituted pyridines The formation of dienal-oximes from the reactions between Grignard rea-gents and pyridine N-oxides, and their subsequent transformations into a range of different compounds was discussed in Chapter 2. Now, we wanted to explore these compounds further, to also include the synthesis of substi-tuted pyridines. Kellogg et al. have earlier reported a ring closure of dienal-oximes in presence of acetic anhydride to yield the substituted pyridine, (ex-emplified with two reactions).6 However, as a result from the low isolated yields of dienal-oximes, the pyridines were only obtained in 17 and 24% yields calculated from the pyridine N-oxides (Scheme 3.2).

N

O

N

OH

R

RMgX Ac2O

N R

N

O

R

O

N

OO

H

R

!

-AcOH

I II

1

Ac2O

Scheme 3.2. Synthesis of pyridines from dienal-oximes. The transformation is probable to proceed via the formation of intermediate I, upon reaction with actetic anhydride. The reaction is suggested to be an electrocyclic ring-closure reaction forming one new σ-bond (intermediate II, Scheme 3.2). Subsequent elimination of acetic acid provides the substituted pyridine 1 (Scheme 3.2).

6Kellogg, R. M.; Van Bergen, T. J. J. Org. Chem. 1971, 36, 1705-1708.

19

3.2.1 Synthesis of 2-substituted and 2,4-disubstituted pyridines In initial studies, the isolated dienal-oxime 3e (Chapter 2, entry 5, Table 2.1) was dissolved in acetic anhydride, and then subjected to microwave irradia-tion at 100 °C for 2 minutes. According to LC-MS analysis of the crude re-action mixture, the corresponding pyridine 1g was the major product with considerable amount of starting material still present in the reaction mixture. To consume the starting material, both the reaction temperature and reaction time were increased to 120 °C and 4 minutes. These adjustments of the reac-tion conditions gave the corresponding pyridine 1g in an isolated yield of 92% calculated from the dienal-oxime. However, to increase the practicality of the method, without isolating the dienal-oxime, the unsubstituted pyridine N-oxide 2a was reacted with PhMgCl, quenched, worked-up using extrac-tion, and concentrated. The crude residue was then dissolved in acetic anhy-dride and heated under microwave irradiation at 120 °C for 4 minutes. After work-up and purification, 2-phenyl pyridine (1c) was isolated in a 63% yield (entry 1, Table 3.1). As expected, when studying the crude mixture by NMR, the 4-substituted regioisomer was not observed. The scope of the reaction was further studied by reacting p-Me- and p-OMe phenylmagnesium chlo-ride to yield corresponding 2-substituted pyridines 1d and 1e, both in 83% yield (entries 2 and 3, Table 3.1). In addition, 2,4-disubstituted pyridines were prepared by starting from 4-substituted pyridine N-oxides 2e-2g (Table 3.1). Aryl, alkynyl and heteroaryl Grignard reagents reacted well with 4-phenyl and 4-benzyloxy substituted pyridine N-oxides, 2e and 2f to form the corresponding 2,4-disubstituted pyridines 1g-1m in yield of 73-86% (entries 5-11, Table 3.1). To challenge the regioselectivity in the reaction further, and enable use of starting materials that are predisposed for further transforma-tions, the 4-chloropyridine N-oxide (2g) was reacted with PhMgCl (entry 14, Table 3.1). In this case no 4-substitution was observed and pyridine 1p was isolated in a yield of 74%. This is a higher yield than reported in previous studies in which phenoxy-carbonyl activated 4-chloropyridines have been reacted with Grignard reagents to yield 55% of the corresponding 2-substituted pyridine.29 As can be seen in Table 3.1, the low yields obtained when alkyl Grignard reagents were used to synthesize dienal-oximes are also reflected in attempts to synthesize alkyl substituted pyridines (entries 4, 12 and 13, Table 3.1). Thus, alkyl Grignard reagents such as benzyl magnesium chlodirde, i-PrMgCl and MeMgCl chloride resulted in low yields (37-45%), when re-acted with pyridine N-oxides 2a and 2f (For a more detailed discussion of alkyl Grignard reagents and N-oxides see Chapter 4).

29Bonnet, V.; Mongin, F.; Trecourt, F.; Queguiner, G.; Knochel, P. Tetrahedron 2002, 58, 4429-4438.

20

Table 3.1. Synthesis of 2-substituted and 2,4-disubstituted pyridines.

N

O

R

N R'

R

2a, e-g1a-n

1) R'MgCl

THF, rt

2) Ac2O

MW, 120 oC,

4 min

entry N-oxide R R’ product yield (%)a

1 2a H Ph 1c 63 2 2a H p-MePh 1d 83 3 2a H p-OMePh 1e 83 4 2a H Bn 1f 38 5 2e Ph Ph 1g 86 6 2e Ph 2-naphthyl 1h 75 7 2e Ph PhCC 1i 78 8 2e Ph Cy-propylCC 1j 86 9 2e Ph 2-thienyl 1k 73

10 2f OBn Ph 1l 82 11 2f OBn 2-naphthyl 1m 79 12 2f OBn Iso-propyl 1n 45 13 2f OBn Me 1o 37 14 2g Cl Ph 1p 74

Reaction conditions: pyridine N-oxide (1 equiv.) in THF, Grignard reagent 1.2 (equiv.) at rt. Crude residue was dissolved in acetic anhydride and heated in microwave for 4 minutes at 120 °C. aIsolated yields.

3.2.2 Reaction with 2- and 3-substituted N-oxides The pyridine N-oxides used in the reaction described as yet have either been unsubstituted, or substituted at the 4-position. To extend the usefulness of the reaction further, the scope for using 2- and 3-substituted pyridine N-oxides was explored. In the case of 2-picoline N-oxide (2b), one α-position is blocked which could promote the formation of 4-addition products. Fur-thermore, pyridine N-oxide 2b has the potential of deprotonation of the methyl in the 2-position upon addition of the Grignard reagent. Despite this, pyridine 1q was isolated in an excellent yield of 87% and the 4-substituted regioisomer was not observed (Scheme 3.3). In the first addition of PhMgCl to 3-picoline N-oxide (2c), attempts were made to isolate the corresponding dienal-oxime (Chapter 2, entry 2, Table 2.1). Surprisingly, no dienal-oxime was observed; instead the 2,3-disubstituted pyridine 1r was formed directly, and was isolated in a moder-ate yield of 43% (Scheme 3.3.). However, the regioselectivity was excellent and only trace amounts of the 2,5-addition product were observed. This re-gioselectivity was somewhat unexpected, since the steric hindrance caused by the methyl group at the ortho-position would be expected to give a domi-

21

nance of a 2,5-disubstituted product if any regioselectivity at all. However, similar results have been previously reported, in studies that have shown the ortho-positions to be more susceptible towards nucleophilic addition than the para-position (with respect to the methyl group).30 Furthermore, the addition of acetic anhydride to the reaction mixture was not necessary. Elimination of magnesium chloride occurred instantaneously, which allowed the corre-sponding pyridine 1r to form directly (Scheme 3.3). Corresponding results were also observed when 3,5-picoline N-oxide (2h) was reacted, and the trisubstituted pyridine 1s was isolated in an excellent yield of 91% (Scheme 3.3).

N

O

N

N

O

N N

N

O

1) PhMgCl THF, rt

2) Ac2O, MW

120 oC, 4 min

2b

1q, 87%

2c

1r, 43%only trace amounts observed

PhMgCl

THF, rt

PhMgCl

THF, rtN

2h1s, 91%

Scheme 3.3. Synthesis of substituted pyridines. 3.2.3 Synthesis of unsymmetrical 2,6-disubstituted pyridines

Oxidation of the ring-cycled intermediate 15, instead of reduction, should open up for a second addition of Grignard reagents (Scheme 3.4). If possible this would serve as an attractive method for the synthesis of 2,6-disubstituted pyridines (Scheme 3.5).

R'MgCl,

N

O

THF, rtN

OH

R'

!

N

OH

H

R'

RR

oxidation

R

2 3 15 Scheme 3.4. Oxidation of dienal oximes To oxidize dihydropyridines, reagents such as DDQ, potassium permanga-nate (KMnO4) or Pd/C are typically used. However, both DDQ and KMnO4,

30Yamada, S.; Toshimitsu, A.; Takahashi, Y. Tetrahedron 2009, 65, 2329-2333.

22

in combination with heating, gave sluggish reaction mixtures that produced only traces of the oxidized product. The use of Pd/C and microwave irradia-tion gave slightly better, but still unsatisfactory results. Our attention was therefore turned to heating the reaction mixture in DMF in the presence of air. By refluxing the crude intermediate dienal-oxime, from reacting pyridine N-oxide 2f with PhMgCl, dissolved in DMF in the presence of oxygen, the corresponding 2-phenylpyridine N-oxide (2i) was obtained in 86% yield. The addition of PhMgCl or p-tolyl magnesium chloride, followed by treat-ment with acetic anhydride and microwave irradiation, produced the corre-sponding pyridines 1t and 1u with yields of 73% and 63%, respectively (Scheme 3.5).

N

O

OBn

1) PhMgCl, THF, rt

2) DMF, refluxing air, 8 hrs

N

O

OBn

Ph

1) PhMgCl THF, rt

2) Ac2O, Mw

120 oC, 4 min

1) p-tolylMgCl THF, rt

2) Ac2O, Mw

120 oC, 4 min

N

OBn

Ph

N

OBn

Ph2f 2i, 86%

1t, 73%

1u, 63% Scheme 3.5. Synthesis of 2,6-disubstituted pyridines.

3.2.4 One-pot synthesis of 4,2-disubstituted pyridines In the method described above, the Grignard reagent was added to the pyri-dine N-oxide at rt., but to obtain good yields of the substituted pyridines, liquid-liquid extractive work-up, followed by the addition of acetic anhy-dride and microwave irradiation, was essential. However, to get a more ro-bust and straightforward synthesis of substituted pyridines we set-out to improve this method. To avoid the work-up and the need to transfer products to another reaction flask, the Grignard addition was performed in the micro-wave vial, and the pH was adjusted to 6-8 using aqueous NaHCO3 after con-sumption of the pyridine N-oxide. After addition of 10 equivalents of acetic anhydride, the resulting slurry was irradiated at 120 °C for 4 minutes. Upon completion of the reaction, the product was purified using solid phase ex-traction, which makes the method amenable for parallel synthesis. Using this protocol, pyridine 1l was isolated in a yield of 69% compared to an 82% yield (entry 10, Table 3.1) when using the liquid-liquid extraction between dienal-oxime and pyridine formation (entry 1, Table 3.2). The lower yield can be explained by the use of solid-phase extraction, in which the focus is on the purity of the products rather than their yield. This procedure gave similarly good yields when used to synthesize additional 4 examples of disubstituted pyridines, 1v – 1y (entries 2-5, Table 3.2).

23

Table 3.2. One-pot procedure for synthesis of substituted pyridines.

N

O

OBnR'MgCl, THF, rt

then NaHCO3

Ac2O, MW

120 oC, 4 minR N R'

OBn

R

2f, 2i

1l, 1v-1y

entry R R’ pyridine

(yield %)a

1 H Ph 1l (69) 2 H 4-Cl-Ph 1v (73) 3 H N-Me-indole 1w (72) 4 H 2-thienyl 1x (69) 5 Ph 4-OMe-Ph 1y (76)

Reaction conditions: pyridine N-oxide (1 equiv.) in THF, Grignard reagent 1.2 (equiv.) at rt, pH 6-8, Ac2O (10 equiv.) 4 minutes at 120 °C. aIsolated yields.

3.3 Synthesis of 4-pyridones (Appendix I) In investigations outlined here we extended the scope of 4-benzyloxy pyri-dine N-oxide to the synthesis of substituted 4-pyridones and 4-aminopyridinium salts. Both of these compound classes are frequently used in pharmaceuticals and are therefore of considerable interest.31 The previous chapter described a one-pot synthesis of substituted pyridines from 4-benzyloxypyridine N-oxides. We postulated that this method, with removal of the benzyl group (e.g. via application of Pd/C in combination with hydrogen gas), could provide easy access to the corresponding substi-tuted 4-pyridione. Indeed, thin layer chromatography (TLC) and LC-MS analysis indicated good results when pyridine 1l was hydrogenated under atmospheric pressure (eq 1, Scheme 3.6). However, during attempts to iden-tify product 16 a brown insoluble solid was formed. The identity of this compound is still unclear, but one possibility is that the 4-pyridone is prone to aggregate, radically changing both its chemical reactivity and solvation (eq 1, Scheme 3.6).32 Due to the difficulties in identifying product 16, we searched for other meth-ods and found one recently reported by Dudley and co-workers. They re-

31For 4-pyridones see (a) Kitagawa, H.; Ozawa, T.; Takahata, S.; Iida, M.; Saito, J.; Yamada, M. J. Med. Chem. 2007, 50, 4710-4720. (b) Clatworthy, A. E.; Pierson, E.; Hung, D. T. Nat. Chem. Biol. 2007, 3, 541-548. For 4-aminpyridinium salts see (c) Qin, D.; Sullivan, R.; Berkowitz, W. F.; Bittman, R.; Roten-berg, S. A. J. Med. Chem. 2000, 43, 1413-1417. 32Reimers, J.R.; Lachlan, E. H.; Hush, N. S. J. Phys. Chem. A 2000, 104, 5087-5092.

24

ported that the 2-benzyloxy pyridinium salt 17 could be used as a benzyla-tion reagent for alcohols providing the benzylated alcohol together with the 2-hydroxyl pyridinium salts 18 (eq. 2, Scheme 3.6).33

N Ph

OBn

Pd/C, H2

MeOHNH

O

Ph

N OBn

R-OH R-OBn

N OH

(eq 1.)

(eq 2.)

TfO TfO

1l16

problem encoutered during identification

17 18 Scheme 3.6. (eq 1.) Attempt to synthesize 4-pyridone from Pd/C and H2 gas. (eq 2.) Dudley’s benzylation of alcohols from 2-benzyloxy pyridinium salts. Hence, this approach could potentially serve as a method for the synthesis of 2-pyridones (eq 2, Scheme 3.6). Inspired by this method for debenzylation, we started to investigate possible quartenization methods for pyridines. The addition of methyl iodine to pyridine 1l dissolved in acetone yielded the corresponding iodopyridinium salt. However this reaction was slow (requir-ing stirring overnight); but higher yields were obtained, more quickly, using methyl triflate as the alkylating agent in the reaction with pyridine 1l. In the next step, a 2M aqueous sodium hydroxide solution was added to the gener-ated pyridinium salt 19a, and the corresponding 4-pyridone 20a was isolated in a 81% yield (entry 1, Table 3.3). Table 3.3. Synthesis of 4-pyridones.

N R'

OBn

R N

O

R'

MeOTf, tol

0 oC - rt 2M NaOH (aq)

R

1l, 1v-1y20a-e

N R'

OBn

R

19a-e

TfO

entry R R’ pyridone

(yield %)a

1 H Ph 20a (81) 2 H 1-napthyl 20b (74) 3 H 4-OMe-Ph 20c (79) 4 H 2-thienyl 20d (82) 5 Ph 4-OMe-Ph 20e (70)

Reaction conditions: MeOTf (1.05 equiv.) in toluene at 0 °C to rt. 30 min, then crude residue treated with 2M NaOH at rt. aIsolated yields.

33Poon, K. W. C.; Dudley, G. B. J. Org. Chem. 2006, 71, 3923-3927.

25

Using the method described above, a small set of substituted 4-pyridones 20a - 20e was synthesized (Table 3.3). As can be seen, the corresponding 4-pyridones was all obtained in isolated yields between 70% and 82% (entries 1 to 5, Table 3.3). 3.3.2. Synthesis of 4 aminopyridines During the debenzylation investigation, we also studied the use of ammonia-saturated THF solution. The aim was to generate the 4-pyridone by adding a pre-saturated ammonia-THF solution to the crude pyridinium salt, followed by removal of reagents and solvent by concentration under reduced pressure. However, when ammonia was allowed to react with the 2-phenylpyridinium salt 19a, no 4-pyridone was observed. Instead the 4-aminopyridnium salt 21a was obtained. Therefore, in parallel with the synthesis of the 4-pyridone, the same set of pyridinium salts were reacted with ammonia to generate the 4-amino-pyridnium salts 21a-21e (Scheme 3.7).

N R'

OBn

R NR

NH2

R'

19a-e

R = H, R' = Ph (83%) 21a

R = H, R' = 1-naphthyl (78%) 21b

R = H, R' = 4-OMe-Ph(80%) 21c

R = H, R' = 2-thienyl (81%) 21d

R = Ph, R', = 4-OMe-Ph (75%) 21e

NH3 saturated

21a-e

N

OBn

Ph

NH

O

NH

MW, 100 oC,

4minN

N

Ph

O

N

N

PhMW, 100 oC,

4min

19a22a, 50% 22b, 60%

TfOTfO

TfO TfOTfO

MeOH

Scheme 3.7. Synthesis of 4-aminopyridinium salts. The pattern seen in Table 3.3 can also be observed in scheme 3.7 and corre-sponding pyridinium salts were isolated in similar yields as the pyridones (Scheme 3.7). In addition to ammonia, the potential use of other amines were also investigated, and the simple addition of either morpholine or piperidine to the pyridinium salt 19a gave the corresponding 4-amino substituted pyridinium salts 22a and 22b with reasonable yields of 50 and 60%, respec-tively (Scheme 3.7).

3.4 Conclusion and outlook In this chapter, the synthesis of substituted pyridines via an electrocyclic ring-closure reaction of dienal-oximes has been presented. The method pro-vides a complete regioselective synthesis of substituted pyridines starting

26

from the reaction between Grignard reagents and pyridine N-oxides. The method is robust, allowing use of alkyl, alkynyl, aryl and heteroaryl Grig-nard reagents to synthesize diverse pyridines. In addition, we have demon-strated ring-closure followed by oxidation to the corresponding 2-substituted pyridine N-oxide, which allows the synthesis of unsymmetrical 2,6-disubstituted pyridines. Furthermore, a one-pot procedure of substituted pyridines has been demonstrated, which is more suitable for library synthe-sis. Finally, if the 4-benzyloxy substituted N-oxide is used as the starting material, both 4-pyridones and 4-amino pyridinium salts can be obtained in a fast and simple manner.

28

4

Synthesis of 2-substituted pyridine N-oxides via

directed ortho-metalation (DOM)

Paper III

4.1 Introduction

In the reactions between Grignard reagents and pyridine N-oxides, discussed in Chapter 2 and 3, alkyl Grignard reagents mainly resulted in low to moder-ate yields. This chapter presents the optimization of a side-reaction – a meta-lation reaction that competes with the nucleophilic addition – as one reason for the low yields observed earlier for the alkylations.

4.1.1 Directed ortho-metalation reaction

In about 1940, Henry Gilman and Georg Wittig reported the first directed ortho-metalation (DOM) reaction between anisole and n-BuLi. This discov-ery has since been developed into what is now a fundamental method for the construction of substituted aromatic and heteroaromatic compounds.34 The general principle for DOM is the chelation of an organometallic reagent (II, Scheme 4.1) to a direct metalation group (DMG), followed by a proton ab-straction ortho to the DMG. The metalated species so created (III, Scheme 4.1) is then reacted with an electrophile to form a new carbon-carbon- or carbon-heteroatom bond (IV, Scheme 4.1).

DMG DMG

R-MH

M-RDMG

ME+

DMG

E

I II III IV

R-M = Organometallic, R-Li, R-MgX etc. Scheme 4.1. Principle of directed ortho-metalation (DOM).

34For reviews see: (a) Snieckus, V. Chem. Rev. 1990, 90, 879-933. (b) Schlosser, M.; Mongin, F. Chem. Soc. Rev. 2007, 36, 1161-1172.

29

Despite intensive studies of ortho-metalation reactions, problems frequently arise when the method is applied to heterocyclic compounds, e.g. pyridines. In these cases one of the greatest challenges is to circumvent the formation of regioisomers. To this end a number of papers have reported successful outcomes when symmetrical pyridines have been used in the reaction (eq. 1, Scheme 4.2).35 However, similar successes have not been achieved when unsymmetrical pyridines have been used (eq. 1, Scheme 4.2).36 To obtain regioselectivity with these compounds, the starting material often needs to contain a directing substitutent, or to have a halogen substituent that under-goes metal–halogen exchange when reacted with organometallic reagents (eqs. 2 and 3, Scheme 4.2). However, the limited availability of such suitable starting materials makes these procedures impractical for the synthesis of substituted pyridines.

N

n-BuLi,LIDMAE

E+N E

(eq 1)

(eq. 2)

N

BrR-M

E+N

E(eq. 3)

N

NHCO2t -Bu

n-BuLi

then MeI N

NHCO2t -Bu

N N E

n-BuLi,LIDMAE

E+NE N

E

ratio: 78:5:17

Scheme 4.2. Different DOM reactions of pyridines.

4.2 Metalation of pyridine N-oxides Pyridine N-oxides are more prone to undergo C-2 metalation than pyridines. This can be explained by the presence of the electron withdrawing N-O functionality, which not only activates the ring towards deprotonation but also allows chelation of the organometallic reagent. The deprotonation of pyridine N-oxides using n-BuLi as the reagent (eq. 1, Scheme 4.3) has been thoroughly studied.37 However, only low to moderate yields of between 14% and 44% have been reported, with the disubstituted products being among the most abundant by-products observed. An alternative method to deprotonation of pyridine N-oxide was developed by Fagnou and co-workers.38 They took advantage of the reactivity of pyridine N-oxides for the 35Caubere, P. Chem. Rev. 1993, 93, 2317-2334. 36Schlosser, M. Angew. Chem., Int. Ed. 2005, 44, 376-393. 37(a) Abramovitch, R. A.; Saha, M.; Smith, E. M.; Coutts, R. T. J. Am. Chem. Soc. 1967, 89, 1537-1538. (b) Abramovitch, R. A.; Coutts, R. T.; Smith, E. M. J. Org. Chem. 1972, 37, 3584-3587. (c) Taylor, S. L.; Lee, D. Y.; Martin, J. C. J. Org. Chem. 1983, 48, 4156-4158. (d) Mongin, O.; Rocca, P.; Thomas-dit-Dumont, L.; Trecourt, F.; Marsais, F.; Godard, A.; Queguiner, G. J. Chem. Soc., Perkin Trans. 1 1995, 2503-2508. 38Campeau, L.-C.; Rousseaux, S.; Fagnou, K. J. Am. Chem. Soc. 2005, 127, 18020-18021.

30

took advantage of the reactivity of pyridine N-oxides for the synthesis of substituted pyridines via C-H activation using Pd as the metal. Although their synthetic procedure is elegant, it is restricted to arylation, and requires an excess of up to 4 equivalents of the pyridine N-oxide (eq. 2, Scheme 4.3).

N

O

n-BuLi

E+N

O

E N

O

EE

(eq. 1)

R R R

N

O

R1) Pd(OAc)2, P

tBu3-HBF4

K2CO3, Tol. 110 oC

2) Pd/C, HCOONH4

MeOH, rt.N

R

Br(eq. 2)

Scheme 4.3. DOM and C-H activation of pyridine N-oxides. Our studies of the synthesis of alkyl substituted pyridines, suggested that the low yields of dienal-oxime 3 (Chapter 2) resulted from a competing deproto-nation of the pyridine N-oxide. Given the previous problems encountered with n-BuLi for deprotonation, we became interested in investigating the possibility of using Grignard reagents to synthesize substituted pyridine N-oxide via a deprotonation reaction. 4.2.1 Directed ortho-metalation using Grignard reagents In our preliminary investigations n-BuMgCl was added drop-wise to pyri-dine N-oxide 2f dissolved in THF at -78 °C. After 60 minutes the reaction was quenched by the addition of MeOD. NMR studies of the crude reaction mixture indicated that more than 90% had achieved a regioselective incorporation of deuterium at the 2-position (Scheme 4.4).

N

O

N

O

MgCl N

O

D

2f 23a 24a

OBn

n-BuMgCl

OBn

MeOD

OBn

Scheme 4.4. DOM using n-BuMgCl followed by addition of MeOD.

Inspired by this result, we changed the electrophile to benzaldehyde but, disappointingly, this resulted in a significant decrease of the isolated yield to 45% of pyridine N-oxide 24d (entry 3, Table 4.1). However, if n-BuMgCl was exchanged to i-PrMgCl an improvement of the yield to 65% of 24d was accomplished (entry 3, Table 4.1). A set of five different pyridine N-oxides were therefore reacted with i-PrMgCl followed by quenching with benzalde-hyde or iodine. These initial results are summarized in Table 4.1.

31

Table 4.1. Deprotonation of pyridine N-oxides followed by addition of benzaldehyde or iodine.

N

O

R3

R2

R1 i-PrMgCl,

THF, -78 oC

then E+

-78 oC to rt

N

O

R3

R2

R1

E

2 24E+ = PhCHO, I2

entry N-oxide R1 R2 R3 E product yield (%)a

1 2c H H Me PhCHOH 24b 61b

2 2e H Ph H PhCHOH 24c 38 3 2f H OBn H PhCHOH 24d 65c

4 2f H OBn H I 24e 67 5 2h Me H Me I 24f 92 6 2j H H OMe PhCHOH 24g 86

Reaction conditions: pyridine N-oxide (1 equiv.) in THF, Grignard reagent (1.7 equiv.) at -78 °C stirred for 60 minutes. Benzaldehyde (2.0 equiv.) added at -78 °C and allowed to attain rt and stirred for additional 30 minutes. aIsolated yields. bIsomeric mixture of 2,3 and 2,5 disubstituted N-oxide obtained in 1:1.5 ratio respectively. cA lower yield 45% was obtained when n-BuMgCl was used.

The isolated yields varied between 38% and 92% depending on the pyridine N-oxide used in the reaction (Table 4.1). Generally, high yields were ob-tained when the pyridine N-oxide was substituted with electron donating groups Me, OBn and OMe (Table 4.1) When deprotonated and reacted with benzaldehyde, both 3-picoline N-oxide (2c) and 3-metoxy pyridine N-oxide (2j) resulted in good yields of 61% and 86% of the disubstituted pyridine N-oxides 24b and 24g, respectively (entries 1 and 6, Table 4.1). As expected, the reaction with 3-picoline N-oxide (2c) gave a mixture of isomers (24b), comprising 2,3- and 2,5-disubstituted pyridine N-oxides in a 1:1.5 ratio, whereas the 3-metoxy N-oxide 2j resulted in 2,3-disubstituted pyridine N-oxide 24g as the sole product (entries 1 and 6, Table 4.1). The 3,5-picoline N-oxide (2h) was also reacted to give the corresponding trisubstituted pyri-dine N-oxide 24f (entry 5, Table 4.1) in an excellent yield of 92%. Further-more, pyridine N-oxide 2f gave approximately the same results whether io-dine or benzaldehyde were used as the electrophile (entries 3 and 4, Table 4.1). The 4-phenylpyridine N-oxide (2e) formed the expected 2,4-disubstituted pyridine N-oxide 24c (entry 2, Table 4.1), but the product was isolated in only a low yield of 38%. The major product isolated in a 56% yield was 2-iso-propyl substituted N-oxide, as a result from the addition of i-PrMgCl followed by oxidation (entry 2, Table 4.1). 4.2.2 DOM comparison between Grignard and lithium reagents In the protocol described above, 1.7 equivalents of Grignard reagents were used with 2.0 equivalents of the electrophile. To further improve and explore

32

the directed ortho-metalation of pyridine N-oxides, we aimed to reduce the number of equivalents and use more demanding electrophiles. We also wanted to be able to compare our results with previously reported using n-BuLi.39 Cyclohexanone was therefore chosen as the electrophile, which po-tentially impact the reaction in two ways, increased bulkiness and the pres-ence of acidic α-protons. The results are summarized in table 4.2. Table 4.2. DOM followed by trapping with cyclohexanone.

N

O

R1

R2

R3

R4

i-PrMgCl, THF, -78 oC

then cyclohexanone,

THF, -78 oC to rtN

O

R1

R2

R3

R4

OH

2 24 entry N-oxide R1 R2 R3 R4 product yield (%)a

1 2a H H H H 24h 38 2 2b Me H H H 24i 32 3 2d H H Me H 24j 36 4 2f H H OBn H 24k 62 5 2g H H Cl H 24l 20 6 2h H Me H Me 24m 81 7 2k OMe H H H 24n 0 8 2j H H H OMe 24o 90 9 2l H H OMe H 24p 22 10 2m Cl H H H 24q 0

Reaction conditions: pyridine N-oxide (1 equiv.) in THF, Grignard reagent (1.2 equiv.) at -78 °C stirred for 60 minutes. Cyclohexanone (1.5 equiv.) was added at -78 °C and al-lowed to attain rt and stirred for additional 30 minutes. aIsolated yields.

A regioselective incorporation of cyclohexanone was achieved with pyridine N-oxide (2a) to give the substituted pyridine N-oxide 24h in a 38% yield (entry 1, Table 4.2). Although the isolated yield is rather low, compared with the previously reported 7% yield using n-BuLi, this is still a marked im-provement.39 As a by-product, the corresponding 2-substituted iso-propyl pyridine N-oxide was isolated in a 15% yield. Similar results were obtained with 2- and 4-picoline N-oxide (2b and 2d), which gave yields of 32% and 36% of 29i and 29j, respectively (entries 2 and 3, Table 4.2), compared with yields of 0% and 21% when n-BuLi was used.39 Furthermore, with 2-chloro and 2-metoxy pyridine N-oxide 2k and 2m, no products were observed after LC-MS or crude-NMR analysis (entries 7 and 10, Table 4.2). However, by switching to 4-chloro- and 4-metoxypyridine N-oxides (2g and 2l), the yields were improved to 20% and 22% respectively (entries 5 and 9, Table 4.2). Finally, pyridine N-oxides 2f, 2h, and 2j were deprotonated, followed by the 39Abramovitch, R. A.; Smith, E. M.; Knaus, E. E.; Saha, M. J. Org. Chem. 1972, 37, 1690-1696.

33

addition of cyclohexanone, to yield the corresponding tertiary alcohols 24k (62%), 24m (81%) and 24o (90%), respectively, (entries 4, 6 and 8, Table 4.2). In conclusion, Table 4.2 indicates that the best results are achieved when the pyridine N-oxides are substituted with an electron-donating group in the 3- or the 3- and 5-positions. We therefore decided to explore further the incor-poration of different electrophiles by reacting 3-metoxy-pyridine N-oxide (2j) with the three different electrophiles, piperidinone 25, iodine, and phen-ylisocyanate, each of which has different useful properties. The reactions were carried out in the same manner as previously described (Scheme 4.5).

N

O

i-PrMgCl

THF, -78 oC

OMe

N

O

OMe

MgCl

PhNCO

I2

N

O

Boc

N

O

OMe

NBoc

OH

N

O

OMe

I

N

O

OMe

O

HN

2j

24r, 93%

24s, 96%

24t, 81%

23b

25 =

25

Scheme 4.5. Direct trapping of intermediate 23b with piperidinone, iodine or phenylisocyanate. The direct trapping of the intermediate 23b using commercially available N-methyl piperidinone, was performed first. After the addition of N-methyl piperidinone to the solution of 23b, a white precipitation was formed and the reaction congested resulting in only a 36% yield of isolated product. How-ever, this yield was improved to an excellent 93% when N-Boc-protected piperidinone 25 was used instead (Scheme 4.3). Furthermore, the addition of iodine to the intermediate 23b gave the corresponding 2-iodo pyridine N-oxide 24s in a 96% yield. This compound, being predisposed to be used in further transformations such as Suzuki-Miyaura (see Chapter 4.3), Heck40 and Stille41 couplings. Finally, reacting phenylisocyanate with intermediate 23b gave 2-phenylcarbamoyl pyridine N-oxide 24t in 81% yield (Scheme 4.5).

40(a) Heck, R. F.; Nolley, J. P., Jr. J. Org. Chem. 1972, 37, 2320-2322. (b) Beletskaya, I. P.; Cheprakov, A. V. Chem. Rev. 2000, 100, 3009-3066. 41(a) Milstein, D.; Stille, J. K. J. Am. Chem. Soc. 1978, 100, 3636-3638. (b) Espinet, P.; Echavarren, A. M. Angew. Chem. Int. Ed. 2004, 43, 4704-4734.

34

4.3 Suzuki–Miyaura coupling of 2-iodo pyridine N-oxides (Appendix II) The Suzuki-Miyaura cross-coupling reaction was first reported in 1979 and has since developed into one of the most popular methods for carbon-carbon bond formation.42 Although tremendous progress has been made such that couplings can now be performed with more challenging substrates,43 only a few reports can be found in the literature concerning Suzuki–Miyaura cou-pling of halo-pyridine N-oxides.44 We therefore decided to investigate the possibility of coupling N-oxide 24s (Scheme 4.5) in a Suzuki–Miyaura cou-pling reaction. We started to investigate the reaction using conditions, previously reported successful on similar substrates. A mixture of 24s, phenyl boronic acid, tetrakis (triphenylphosphine) palladium (Pd(PPh3)4) and potassium carbonate (K2CO3) in DMF/water was irradiated with microwaves for 10 minutes at 100 °C to give the corresponding 2,3-substituted N-oxide 26a in an excellent 99% yield (entry 1, Table 4.3). When phenyl boronic acid was changed for 3-methyl- or 3-nitrophenyl boronic acids, the corresponding pyridine N-oxides 26b and 26c were isolated in yields of 97% and 90%, respectively (entries 2, 3, Table 4.3).

Table 4.3. Suzuki–Miyaura coupling of 2-iodo pyridine N-oxide 24s.

N

O

OMe

I

R-B(OH)2, Pd(PPh3)4 (10 mol%)

K2CO3, DMF/H2O

MW 100 oC, 10 min

N

O

OMe

R

24s 26 entry R product yield (%)a

1 Ph 26a 99 2 3-Me-Ph 26b 97 3 3-NO2-Ph 26c 90

Reaction conditions: pyridine N-oxide (1 equiv.) in DMF:water 4:1, boronic acid (3 equiv.), Pd(PPh3)4 (10 mol%), K2CO3 (3 equiv). Irradiated 10 minutes at 100 °C. aIsolated yields.

During attempts to improve the reaction, it was found that when MeOH was used as the solvent, the reaction time could be reduced from 10 to 2 minutes, and the temperature could be decreased slightly from 100 °C to 80 °C. In

42(a) Miyaura, N.; Suzuki, A. Chem. Commun. 1979, 866-867. (b) Miyaura, N.; Suzuki, A. Chem. Rev. 1995, 95, 2457-2483. 43Littke, A. F.; Fu, G. C. Angew. Chem. Int. Ed. 2002, 41, 4176-4211. 44(a) Lohse, O.; Thevenin, P.; Waldvogel, E. Synlett 1999, 45-48. (b) Gong, Y.; Pauls, H. W. Synlett 2000, 829-831.

35

addition, to further ease the work-up, the amount of catalyst was reduced from 10 mol% to 1 mol% by using a solid supported palladium catalyst: VersaCatTM Pd.45 Eight different substituted pyridine N-oxides were synthe-sized in parallel using these changes to the method. Thus, the two pyridine N-oxides 24e and 24s, were reacted with four boronic acids i.e. phenyl, 4-methoxy, 3-tolyl, and 3-nitro phenyl boronic acids. After irradiation the cata-lyst was filtered off and the solvent removed under reduced pressure. The eight crude mixtures were then purified by parallel column chromatography to give the corresponding pyridine N-oxides 26a-26h in 80% to 92% isolated yields (entries 1-8, Table 4.4).

Table 4.4. Parallel synthesis of pyridine N-oxides using Suzuki–Miyaura coupling.

N

O

I

R1-B(OH)2

Ps-Pd (1 mol%)

K2CO3, MeOH

MW 80 oC, 2 min

N

O

R1

24 26

R R

entry N-oxide R R1 product yield(%)a

1 24s 3-OMe Ph 26a 84 2 24s 3-OMe 3-Me-Ph 26b 86 3 24s 3-OMe 3-NO2-Ph 26c 80 4 24s 3-OMe 4-OMe-Ph 26d 88 5 24e 4-OBn Ph 26e 90 6 24e 4-OBn 3-Me-Ph 26f 93 7 24e 4-OBn 3-NO2-Ph 26g 80 8 24e 4-OBn 4-OMe-Ph 26h 92

Reaction conditions: pyridine N-oxide (1 equiv.) in MeOH, boronic acid (3 equiv.), VersaCat (1 mol%), K2CO3 (3 equiv). Irradiated 2 min at 80 °C. a Isolated yields.

In summary we were able to isolate eight new pyridine N-oxides in high yields in only 7 hours, including reactions and the final purification.

4.4 The direct coupling of the metalated pyridine N-oxides The direct cross-coupling of Grignard reagents and halides has recently been the focus of much attention. Several different methods have been reported, including the use of an iron catalyst as an economical and green alternative to palladium and nickel in cross-coupling reactions.46 Aryl, vinyl and alkyl halides have been efficiently coupled with different aryl and alkyl magne-sium halides using iron as catalyst. In 2004, Fürstner and co-workers re-ported a cross-coupling of alkyl halides with aryl Grignard reagents, that was catalyzed by low-valency iron complexes such as FeCl3, Fe(acac)2 or

45Basu, B.; Das, S.; Das, P.; Mandal, B.; Banerjee, D.; Almqvist, F. Synthesis 2009, 1137-1146. 46Frisch, A. C.; Beller, M. Angew. Chem. Int. Ed. 2005, 44, 674-688.

36

Fe(acac)3 (where acac = acetylacetonate).47 Interested in this method, we decided to investigate the use of metalated pyridine N-oxides, with an iron catalyst, as a method for the direct coupling of Grignard reagents with vari-ous halides (Scheme 4.6). Unfortunately neither 1-bromopentane nor bromo-cyclohexane, in combination with FeCl3 or Fe(acac)2 in THF or a THF:NMP (9:1) mixture, reacted successfully: only the starting material was recovered (Scheme 4.6). Other reports have shown that the addition of N,N,N’,N’-tetramethylethyldiamine (TMEDA) facilitates the coupling.48 However, equimolar amounts of Grignard reagent and TMEDA, in the presence of 5 mol% FeCl3 again returned only the starting material. Cahiez et al. reported that the addition of 1.5 equivalents of TMEDA to a solution of FeCl3 gener-ates a [(FeCl3)2(tmeda)3] complex that can be efficiently used for cross-coupling reactions.49 Unfortunately, also with this additive only the starting material 2j was isolated (Scheme 4.6).

N

O

O i-PrMgCl,

THF/NMP, -78 oC

N

O

O

AlkylFe-catalyst

Alkyl-X

Fe-catalyst: FeCl3 Fe(acac)2

[(FeCl3)2(tmeda)3]

Alkyl-X: 1-bromopentane bromocyclohexane

N

O

O i-PrMgCl,

THF, -78 oC

N

O

O

CuI or CuCN

1-bromopentane

2j 2j

Scheme 4.6. Attempts for direct cross coupling of pyridine N-oxide 2j.

Because of the unsuccessful results using iron as catalyst, our attention was turned to copper instead.50 Pyridine N-oxide 2j was therefore first deproto-nated, and then transferred to a solution of 1-bromopentane with catalytic amounts of either CuI or CuCN (5% in both cases) (Scheme 4.6). Again, these attempts failed to cross-couple and only the starting material 2j was recovered. The failure of both iron and copper in the cross-coupling reactions, led us to study palladium as the catalyst. Therefore the palladium-catalyzed direct-coupling of Grignard, known as the Kumada coupling was investigated.51 Since coupling with alkyl, vinyl or aryl has been reported, the metalated pyridine N-oxide 23b was reacted with either iodobenzene in the presence of Pd(PPh3), or with benzyl chloride in the presence of palladium acetate

47Martin, R.; Fuerstner, A. Angew. Chem. Int. Ed. 2004, 43, 3955-3957. 48Nakamura, M.; Matsuo, K.; Ito, S.; Nakamura, E. J. Am. Chem. Soc. 2004, 126, 3686-3687. 49Cahiez, G.; Habiak, V.; Duplais, C.; Moyeux, A. Angew. Chem. Int. Ed. 2007, 46, 4364-4366. 50Terao, J.; Todo, H.; Begum, S. A.; Kuniyasu, H.; Kambe, N. Angew. Chem. Int. Ed. 2007, 46, 2086-2089. 51(a) Tamao, K.; Sumitani, K.; Kumada, M. J. Amer. Chem. Soc. 1972, 94, 4374-4376. (b) Frisch, A. C.; Shaikh, N.; Zapf, A.; Beller, M. Angew. Chem. Int. Ed. 2002, 41, 4056-4059. for review see (c) Corbet, J.-P.; Mignani, G. Chem. Rev. 2006, 106, 2651-2710.

37

(Pd(OAc2)). Unfortunately, in both cases no product was observed and only the starting material 2j was recovered (Scheme 4.7).

N

O

i-PrMgCl

THF, -78 oC

OMe

N

O

OMe

MgCl

2j 23b

Catalyst

HalideN

O

OMe

R

Catalyst: Pd(PPh3)4 Pd(OAc)2 with PCy3

Halide: iodobenzene benzylchloride

R: phenyl benzyl

Scheme 4.7. Kumada coupling of pyridine N-oxide 2j. Next our attention was turned to diaryliodonium salts 28. These salts have proven to be excellent alternatives to aryl halides for the introduction of aryl substitutents.52 The diarylidonium salt 28 was prepared in a 79% yield ac-cording to a published method (Scheme 4.8).53 Howeve, the uncatalyzed reaction between Grignard reagents and diaryliodnium salts has been re-ported to be troublesome.54 Also in our case the arylated pyridine N-oxide was not observed when iodnium salt 28 was added directly to intermediate 23c. However, repeating the reaction in presence of catalytic amount of zinc chloride (ZnCl2) and Pd(PPh3)4 gave trisubstituted pyridine N-oxide 26i in 10% yield. This yield was increased to 51% when the reaction was irradiated for 10 minutes at 70 °C in a microwave reactor (Scheme 4.8).55

N

O

i-PrMgCl

THF, -78 oC N

O

MgCl

2h 23c

Pd(PPh3)4

ZnCl2, 28 THF

MW 70 oC

10 min

N

O

26i, 51%

Im-MCPBA

TfOH, DCM, rt.

I

28, 79%

TfO

Scheme 4.8. Double metal catalyzed arylation of metalated N-oxide 6h.

4.4 Conclusion and outlook In this chapter, a complete regioselective deprotonation of pyridine N-oxides using i-PrMgCl has been described. Previously reported deprotonations, using lithium reagents, have suffered from drawbacks such as low yields and the formation of the disubstituted product as the major product. This prob- 52(a) Kalyani, D.; Deprez, N. R.; Desai, L. V.; Sanford, M. S. J. Am. Chem. Soc. 2005, 127, 7330-7331. (b) Aggarwal, V. K.; Olofsson, B. Angew. Chem. Int. Ed. 2005, 44, 5516-5519. 53Bielawski, M.; Olofsson, B. Chem. Commun. 2007, 2521-2523. 54Wang, L.; Chen, Z.-C. Synth. Commun. 2000, 30, 3607–3612. 55Further optimization of the reaction was not performed.

38

lem was not encountered when Grignard reagents were used. Although elec-tron-donating groups in the third position of the ring were crucial in order to obtain good yields, the reaction allowed incorporation of a range of different electrophiles. We also demonstrated the efficacy of 2-iodo-pyridine N-oxides in a Suzuki–Miyaura coupling. This method was used to synthesize eight different aryl substituted pyridine N-oxides in a parallel manner. Finally, we demonstrated a direct cross-coupling reaction between metalated pyridine N-oxide and diaryliodonium salt under zinc and palladium catalysis.

40

5

Synthesis of trans-2,3-dihydropyridine N-oxides

and piperidines

Paper IV

5.1 Introduction The main focus in Chapter 5, is the synthesis of substituted piperidines, a frequently occurring motif in pharmaceuticals and natural products (Figure 5.1).56 For example, Paroxetine 29, which is used in the treatment of depres-sion, contains a 3,4-disubstituted piperidine (Figure 5.1). Another interesting compound is the naturally occurring akaloid coniine 30, which, despite its simple structure, is very toxic, and due to its stereochemistry, is a great chal-lenge to synthesize. As a final example, compound 31 is the analgesic mor-phine, which is a multi-ring fused compound with piperidine embedded in its structure.

NH

O

O

OF

NH

O

HO

N

H

HO

Paroxetine, 29 Coniine, 30 Morphine, 31 Figure 5.1. Structures of pharmaceuticals and natural products containing piperidine.

The importance of piperidines has stimulated intense research into the de-velopment of new methods for their preparation.57 One recently reported method included a reductive cross-coupling between homoallylic alcohols and imines followed by a cyclisation to the substituted piperidines. (eq. 1, Scheme 5.1).58 This method resulted in good regio- and stereocontrol of the reaction (dr 95:5). An alternative method, to cyclisation, that has proven

56(a) Watson, P. S.; Jiang, B.; Scott, B. Org. Lett. 2000, 2, 3679-3681. (b) Fore review see Kaellstroem, S.; Leino, R. Bioorg. Med. Chem. 2008, 16, 601-635. 57For reviews see (a) Laschat, S.; Dickner, T. Synthesis 2000, 1781-1813. (b) Buffat, M. G. P. Tetrahe-dron 2004, 60, 1701-1729. 58Takahashi, M.; Micalizio, G. C. J. Am. Chem. Soc. 2007, 129, 7514-7516.

41

useful is the asymmetric hydrogenation of activated pyridines (eq. 2, Scheme 5.1). For example, Charette and co-workers developed an iridium-catalyzed asymmetric hydrogenation of N-iminopyridinium ylides for the synthesis of enantiomerically enriched piperidines (Scheme 5.1).59

H Ph

NBn

-70-40 oC

OH

Ph

HN

Et

Bn

PPh3, imidazole

N

R3

Ph

Et

(eq 1)

N

NBz

Me N Me

NHBz

cc (2mol%)I2 (2mol%)

H2 (400 psi)16-20 hourstoluene, rt

cc=

N

O

t-Bu

Ar2P

IrCOD

(eq 2)

Et

OTi(Oi-Pr)4Li

c-C5H9MgCl

CCl4

dr: 95/5

ee: 90 Scheme 5.1. Examples of stereoselective synthesis of piperidines. Chiral N-acyl activated pyridinium salts 32 have also been used for the asymmetric synthesis of substituted piperidines (Scheme 5.2). In the synthe-sis of a range of different piperidine containing alkaloids, Comins and cowerkers used 4-dihydropyridones 33 as common intermediates formed from the addition of Grignard reagents to N-acyl-4-methoxypyridinium salts.60

N

OMe

TIPS

CO2R*

H3O+

N

O

TIPS

R

CO2R*X

R*= chiral auxiliary

RMgXPiperidine containing alkaloids

32 33

N

O

RMgX

rt

RMgX

-40 oC

then E+

or NaBH4

MeOH

N

OH

R N

OH

R

OH

EH

or

N R

OH

2

prior knowledge

discussed in this chapter

Scheme 5.2. Synthesis of substituted piperidines from activated pyridines.

Pyridine N-oxides 2, on the other hand, have earlier been shown to undergo ring-opening when reacted with Grignard reagents (Chapter 2 and 3) and have not been used as precursors to piperidines. However, since our previous results indicated that ring-opening can be avoided by modifying the reaction conditions, we were interested to see whether this approach could be used for the synthesis of substituted piperidines (Scheme 5.2).

59Legault, C. Y.; Charette, A. B. J. Am. Chem. Soc. 2005, 127, 8966-8967. 60Comins, D. L.; Joseph, S. P.; Goehring, R. R. J. Am. Chem. Soc. 1994, 116, 4719-4728.

42

5.2 Synthesis of trans-2,3-dihydropyridine N-oxide During our studies of the previously discussed metalation reaction (Chapter 4), PhMgCl was reacted with pyridine N-oxide (2a) at -78 °C, followed by the addition of MeOD. The expected incorporation of deuterium at the 2-position was, not observed. Instead, the 2-phenyldihydropyridine N-oxide 34 was observed as the major product according to LC-MS analysis (Scheme 5.3). However, attempts to isolate this compound by silica column chroma-tography only resulted in its decomposition. To circumvent this, the reaction was repeated at -40 °C; but immediately after the addition of MeOD, a slurry of NaBH4 in MeOH was added (Scheme 5.3). This time, after work-up and purification, the reaction resulted in N-hydroxyl 2-phenyl tetrahydropyridine 35, which was isolated in a quantitative yield. NMR studies revealed that the deuterium was assigned to the 5-position; not the 2-position as had been expected. The experiment was then repeated using MeOH instead of MeOD and the corresponding tetrahydropyridine 36 was isolated in a 94% yield (Scheme 5.3).

N

O

PhMgCl

THF, -40 oC

then MeOD

then MeOH,

-78 oCN

O

Ph

34

N

O

PhMgCl

THF, -40 oC

then NaBH4

in MeOH

-40 oC

N

OH

Ph

2a

2a 36, 94%

Not stable

N

O

PhMgCl

THF, -40 oC

then MeOD

then NaBH4

in MeOH

-40 oC

N

OH

Ph

2a 35, quant.

D

Scheme 5.3. Synthesis of 2-substituted tetrahydropyridine N-oxide from pyridine N-oxide.

To study the reaction further, benzaldehyde was used instead of MeOH. Thus, after the addition of PhMgCl to pyridine N-oxide (2a), benzaldehyde was added at -78 °C. The reaction mixture was stirred at -78 °C for 30 min, and, according to LC-MS, a disubstituted product was formed. Repeating the experiment at -40 °C gave a total conversion of the starting material and the disubstituted dihydropyridine N-oxide 37a was isolated in a 79% yield (Scheme 5.4). Data from 1H-NMR and 13C-NMR analysis confirmed the 2,3-disubstitution pattern, and not the expected 2,5-disubstituted product that was obtained when MeOD or MeOH was used as the electrophiles (Scheme 5.3). This finding was further confirmed by 2D-NMR, noesy and hmbc ex-periments, and the elucidation of a crystal structure. Furthermore, H-NMR analysis of the crude reaction mixture confirmed that the 2,3-trans isomer

43

was formed exclusively, with the cis isomer not being detected at all. The diastereomeric mixture formed in the reaction results from the benzylic al-cohol substituted in the third position. Although an isomeric mixture is ob-tained, the diastereomeric ratio (dr) of 82/18, is rather impressive, especially considering that three stereocentres are formed in one reaction step.

N

O

PhMgCl

THF, -40 oC

then benzaldehyde,

-40 oCN

O

Ph

OH

PhH

2a 37a, 79%dr= 82/12

Scheme 5.4 Syntheisis of trans 2,3-dihydropyridine N-oxide.

To further support the diene-structure, dimethyl acetylene-dicarboxylate was reacted with the trans-2,3-dihydropyridine N-oxide in a Diels–Alder reac-tion. Thus, heating 37a and dimethyl acetylenedicarboxylate in toluene for 30 min gave the aza-bicyclo compound 38 in an 86% yield (Scheme 5.5).

dimethyl acetylene-dicarboxylate

Tol. 60 oC, 30 min

37a

38, 86%

NPh

O

HO

Ph

CO2Me

CO2MeNPh

H

HO

Ph

O

Scheme 5.5 Diels–Alder reaction and structures of aza-bicyclo compound 38.

The structure of the Diels–Alder product 38 was determined by a noesy ex-periment, which showed a strong cross-coupling-peak between the proton in the 3-position and the bridged proton. Two energy minimized structures of the product, using molecular mechanics (MM2) are shown in Scheme 5.5. Based on these calculations, the distance between the two protons in each structure should be 3.2 Å and 4.3 Å, respectively. However, because 4.3 Å is not likely to give a strong coupling in the noesy spectra (Scheme 5.5), the diasteromer 38 shown in Scheme 5.5 is supported by the results of the analy-sis.

Further investigation of the scope for the reaction was performed with dif-ferent pyridine N-oxides and PhMgCl followed by sequential addition of

44

benzaldehyde, butyraldehyde or cyclohexanone. These results are summa-rized in Table 5.1. Table 5.1. Synthesis of 2,3-dihydropyridine N-oxides.

N

O

PhMgCl

THF, -40 oC

then E+= 39,

40 or 41N Ph

E

OH

O

39= butyraldehyde40 = benzaldehyde41 = cyclohexanone

237

R R

entry N-oxide E+ product dra yield (%)b

entry N-oxide E+ product dra yield (%)b

1 2a

N

O 39

N Ph

O

OHH

37b 82/18 59 6 2n

N

O

Ph

40

N Ph

O

OH

Ph

Ph

H

37g 78/22 98

2

2e

N

O

Ph

40

N Ph

O

OH

PhPhH

37c 89/11 81 7 2n

N

O

Ph

39

N Ph

O

OH

Ph

H

37h 90/10 96

3

2e

N

O

Ph

39

N Ph

O

OHPhH

37d 8911 80 8

2oN

O

Ph

40

N Ph

O

OH

Ph

Ph

H

37i 50/50 56c

4

2f

N

O

OBn

40

N Ph

O

OH

PhBnOH

37e 88/12 86 9

2oN

O

Ph

39

N Ph

O

OH

Ph

H

37j 60/40 71c

5 2f

N

O

OBn

39

N

OH

Ph

O

BnOH

37f 72/28 71 10

2e

N

O

Ph

41

N Ph

O

Ph

OH

H

37k

- 50d

Reaction conditions: pyridine N-oxide (1 equiv.) in THF, PhMgCl (1.2 equiv.), electrophile (1.5 equiv.). aDiastereomeric mixture (dr) reported as major/minor with respect to the alcohol. Estimated from crude NMR analysis. bIsolated yields. cReaction performed at -20 °C. dReaction performed in toluene.

Pyridine N-oxide (2a) and the substituted 4-phenyl and 4-benzyloxy pyridine N-oxides (2e and 2f) gave the corresponding trans-2,3-dihydropyridine N-oxides 37b-37f using either benzaldehyde or butyraldehyde as electrophiles (entries 2-5, Table 5.1). 2-Phenylpyridine N-oxide (2n) gave excellent yields of the 2,3-substituted products 37g and 37h, of 98% and 96%, respectively, (entries 6 and 7, Table 5.1). Pyridine N-oxide 2o, which is already substi-tuted in both the 2- and 3-positions, was also examined. In this case the nu-cleophilic addition, and thus the consumption of the pyridine N-oxide, was slower. The reaction was therefore conducted at -20 °C61 instead of -40 °C. Using this procedure, benzaldehyde gave product 37i, which was isolated in a 56% yield (entry 8, Table 5.1), whereas the corresponding reaction with butyraldehyde gave the product 37j in a 71% yield (entry 9, Table 5.1). It

61A series of temperature experiments at -78, -40, -20, 0 °C proved that -20 °C was the upper limit. At temperatures above -20 °C a substantial amount of dienal-oxime was observed.

45

was also observed that the diastereomeric mixture ratios differed, depending on the temperature. The low drs. of 50/50 for 37i and 60/40 for 37j could be due to the fact that they were reacted at -20 °C (entries 8 and 9, Table 5.1). This tendency was also observed in the preparation of compounds 37c and 37h. Changing temperatures from -40 °C to -78 °C improved the dr. from 89/11 to 98/2 and from 78/22 to 92/8, for 37c and 37h respectively. The incorporation of other functionalities was also investigated. Unfortu-nately, using cyclohexanone, iodine, or phenylisocyanate as electrophiles only resulted in complex product mixtures. However, these electrophiles are less reactive than aldehydes, and to address this, the temperature was in-creased, but unfortunately without success. However, changing THF to tolu-ene gave a better result when cyclohexanone was used, and the correspond-ing tertiary alcohol 37k was obtained in a 50% yield (entry 10, Table 5.1). The rationale for this could be a tighter coordination of the reactive species to the carbonyl group, thereby increasing the reactivity of the carbonyl, i.e. ketone. However, these changes in the method was not successful when ei-ther iodine or phenylisocyanate were used as electrophiles, which might indicate that the strong coordination to the oxyphilic magnesium is neces-sary. Table 5.2. Synthesis of 2,3-dihydropyridine N-oxides.

N

O

R-MgX

THF,-40 oC

then 40

-40 oC N

O

R

OH

Ph

2a 37X = Cl or Br

H

entry RMgX product dra yield (%)b

1 O

MgCl

N

O

OH

Ph

O

H

37l 81/19 72

2 F

MgCl

N

H

O

Ph

OH

F37m

79/21 76

3 MgBr

N

H

O

Ph

OH

37n 81/19 28

4 MgBr

N

H

O

Ph

OH

37o 81/19 60

Reaction conditions: pyridine N-oxide (1 equiv.) in THF, PhMgCl (1.2 equiv.), electrophile (1.5 equiv.). aDiastereomeric mixture (dr) reported as major/minor with respect to the alcohol. Estimated from crude NMR analysis. bIsolated yields.

46

The reactivity of different Grignard reagents in combination with benzalde-hyde was studied next. Both electron rich and electron poor arylmagnesium chloride reacted in a similar manner to PhMgCl to give the corresponding products 37l and 37m in yields of 72% and 76% respectively (entries 1 and 2, Table 5.2). As might be expected, reacting the alkyl Grignard reagent n-BuMgCl with pyridine N-oxide (2a) gave 37n in only a 28% yield. This further confirms the previous results obtained when pyridine N-oxides were reacted with al-kyl Grignard reagents at low temperature (see section 4.2.1). The addition of vinyl magnesium chloride to 2a followed by benzaldehyde, gave product 37o in 60% yield (entry 4, Table 5.2).

5.3 Synthesis of piperidines

From the synthesis of substituted tetrahydropyridine N-hydroxyls 36 and 2,3-dihydropyridine N-oxides 37 described above, we now wanted to use these intermediates in the preparation of substituted piperidines. 5.3.1 Synthesis of 2-substituted piperidines Performing atmospheric hydrogenation using catalytic amounts of Pd/C (10 mol%) in MeOH gave the desired piperidine, but a long reaction time (up to three days) was necessary unless the amount of Pd/C was greatly increased. The reaction time could be reduced by the addition of catalytic amounts of acetic acid, but 48 hours or more were still necessary for all the starting ma-terial to be consumed. Increasing the pressure of hydrogen gas to 35-50 psi reduced the reaction time further, but still, more than 24 hours were required and from a practical point of view this is not optimal. From these experi-ments, we observed that the double bond was easily reduced, but that the removal of the N-hydroxyl was more troublesome. In order to address this, we applied a double reduction strategy. First the tetrahydropyridine N-hydroxyl 36 was reacted with Zn dust in acetic acid and MeOH (1:1). Only after complete removal of the hydroxyl group, which took 4 hours at rt., was Pd/C (10 mol%) added and the mixture subjected to hydrogen gas. To facili-tate the purification of the amine, a N-Boc protection was performed and the 2-phenyl piperidine 42 was isolated in a 79% yield (Scheme 5.6).

N

1) Zn dust, AcOHPd/C, H2

Ph

42, 79%

2) Boc2O, TEAdioxane/water 4:1

Boc

N

O

PhMgCl

THF, -78 oC

then NaBH4

in MeOH

-78 oC to rt

N

OH

Ph

2a 36, 94% Scheme 5.6. Synthesis of 2-substituted piperidine.

47

This is the first reported synthesis of piperidines derived from pyridine N-oxide. The method is complete regioselective and provides an easy access to 2-substituted piperidines. 5.3.2 Synthesis of trans-2,3 disubstituted piperidines In addition to 2-substituted piperidines, we also wanted to synthesize 2,3- disubstituted piperidines from the previously described 2,3-dihydropyridine N-oxides. With the stereochemistry of the 2- and 3-position already defined, a reduction of the diene system, followed by removal of the hydroxyl group, would render a diastereoselective synthesis of 2,3-disubstituted piperidines. Being interested in a recently reported hydrogenation, using Pd/C and 1,4-cyclohexadiene as the hydrogen source in combination with microwave irra-diation, we decided to try the same procedure.62 After optimization of the reaction time and temperature, dihydropyridine N-oxide 37a was irradiated for 40 min at 145 °C in the presence of 10 equivalents of 1,4-cyclohexadiene, which resulted in a complete conversion to the secondary amine. After Boc protection, the corresponding piperidine 43 was isolated in an excellent 90% yield (Scheme 5.7). Surprisingly, this method only gave trace amounts of piperidine 44 when dihydropyridine N-oxide 37o was re-acted in the same manner. However, this transformation was performed in-stead, by using Raney-Nickel as the catalyst. Thus, Raney-Nickel was added to 37o dissolved in MeOH and the resulting slurry was subjected to atmos-pheric pressure of hydrogen gas for 45 minutes at rt. to give piperidine 44 in a 71% yield (Scheme 5.7).

N Ph

O

OH

Ph 1) Pd/C, MeOH

1,4-cyclohexadiene

MW, 145 oC, 40 min

2) Boc2O, TEA dioxane:water 4:1

N Ph

OH

Ph

Boc

43, 90%

H H

N

O

OH

Ph

Raney-Nickel

H2, 1 atmMeOH

NH HCl

OH

Ph

44, 71%

H H

37a, 79%

37o, 60%

N

O

PhMgCl

THF, -40 oC

then benzaldehyde

2a

N

O

vinylMgCl

THF, -40 oC

then benzaldehyde

2a Scheme 5.7. Synthesis of trans-2,3-disubstituted piperidines.

5.4 Conclusion and outlook In this chapter we have shown that readily available pyridine N-oxides can be used with advantage for the complete regio- and stereoselective synthesis of substituted piperidines. We have proven that, if the reaction is kept cold during the addition of Grignard reagents to pyridine N-oxides, the ring re- 62Quinn, J. F.; Razzano, D. A.; Golden, K. C.; Gregg, B. T. Tetrahedron Lett. 2008, 49, 6137-6140.

48

mains intact. The method can be used for the synthesis of both 2-substituted piperidines and 2,3-subsituted piperidines. The latter compounds are formed via a sequential addition of Grignard reagents and aldehydes or ketones, to pyridine N-oxides at -40 °C, yielding the trans-2,3-dihydropyridine N-oxide addition. Furthermore, we have shown that these novel intermediates formed in the reaction, can be used for further transformations, such as a Diels–Alder reaction, to give a substituted aza-bicyclo scaffold.

50

6

Grignard addition to pyrazine N-oxides: towards an efficient

enantioselective synthesis of substituted piperazines

Paper V

6.1 Introduction Substituted piperazines are motifs often found in bioactive compounds. The scaffold is considered to be a privileged structure,63 and a recent survey showed that 60 out of 1000 orally administrated drugs contained a piperazine fragment.64 Synthetic strategies for substituted piperazines often rely on cyclisation pro-cedures. Madsen and co-workers recently reported an iridium catalyzed syn-thesis of piperazines from diamines and diols (eq. 1, Scheme 6.1).65 Another efficient strategy is the hydrogenation of pyrazines. This procedure can al-low asymmetric control of the reaction, although few examples have been reported to date, and most of these report only modest enantiomeric excess e.g. 78% ee (eq. 2, Scheme 6.1).66

H2N NH2 HO

Ph

OH

0.5 mol%[Cp'IrCl2]2

5 mol% NaHCO3

Water

NH

HN

Ph

(eq 1)

N

N Rh/Josiphos (4%)

H2 (50 atm), 70 oCCONH-t-Bu N

H

HN

CON-t-Bu

(eq 2)

78% ee

quant.

Scheme 6.1. Different methods to prepare substituted piperazines. 63(a) Evans, B. E.; Rittle, K. E.; Bock, M. G.; DiPardo, R. M.; Freidinger, R. M.; Whitter, W. L.; Lundell, G. F.; Veber, D. F.; Anderson, P. S.; et al. J. Med. Chem. 1988, 31, 2235-2246.(b) Horton, D. A.; Bourne, G. T.; Smythe, M. L. Chem. Rev. 2003, 103, 893-930. 64Vieth, M.; Siegel, M. G.; Higgs, R. E.; Watson, I. A.; Robertson, D. H.; Savin, K. A.; Durst, G. L.; Hipskind, P. A. J. Med. Chem. 2004, 47, 224-232. 65Nordstrom, L. U.; Madsen, R. Chem. Commun. 2007, 5034-5036. 66Zhou, Y.-G. Acc. Chem. Res. 2007, 40, 1357-1366.

51

The organometallic additions to activated pyrazines for the synthesis of piperazines, in analogues to activated pyridines, have not been reported. This is surprising since activated pyrazines could potentially constitute an excel-lent precursor for the preparation of substituted piperazines. In this chapter we discuss discoveries regarding Grignard reagent addition to pyrazine N-oxides for the synthesis of substituted piperazines. 6.1.2 Pyrazine N-oxide compared to pyridine N-oxide In contrast to pyridine N-oxide 2a the pyrazine N-oxide 45 contains two nitrogens (Figure 6.1). As a result, the ring is more π-electron deficient and is therefore more susceptible to nucleophiles. When pyridine N-oxide is re-acted with nucleophiles, we have shown excellent regioselectivity towards the 2- and 6-positions. However, the nucleophilic reaction with pyrazine N-oxide could potentially result in addition at either the 2- and 6- positions, or at the 3- and 5-postions.

N

O

N

N

O

more electron deficient potentiallymore demanding regioselectivity

vs

2a 45 Figure 6.1. Comparison between pyridine and pyrazine N-oxide.

As with pyridine N-oxides, the pyrazine N-oxides are commercially avail-able, or can be prepared by the oxidation of the corresponding pyrazine.

6.2 One-pot synthesis of 2-substituted piperazines Assuming that the pyrazine N-oxide reacts in a similar manner to the pyri-dine N-oxide, the rationale was that after the addition of the Grignard rea-gent, a sequential addition of a hydride source would generate the saturated hydroxyl piperazine 46 (Scheme 6.2). Protection of this intermediate would then result in a short and efficient route to substituted piperazines 47, which are protected so that the nitrogens can be orthogonally deprotected and thus selectively functionalized (Scheme 6.2).

N

N

O

RMgX

N

N

R

OMgX

possible intermediate

red.

N

HN

R

OH

protection

N

N

OH

R

4546 47

Pg

Scheme 6.2. Overview of the synthetic route to orthogonally protected piperazine.

52

Initially, different reaction temperatures were investigated. Bearing in mind the fact that the reaction between Grignard reagents and pyridine N-oxides is fast, even at -78 °C (30-60 minutes, see chapter 5), the reaction with pyrazine N-oxide might be expected to be even faster. We therefore reacted PhMgCl with pyrazine N-oxide at -78 °C, -40 °C, -17 °C and 0 °C. Immedi-ately (1 minute) after the addition of the Grignard reagent, the reaction mix-tures were analyzed by TLC. In all cases total consumption of the starting material was observed. However, when the reaction was performed at tem-peratures above -40 °C, a black insoluble solid was seen to form almost di-rectly after the addition of PhMgCl. It was also noticed that DCM was a more suitable solvent due to its solvation properties being better than THF. In a second reaction, PhMgCl was added to pyrazine N-oxide 45 in DCM at -78 °C followed by a reduction using NaBH4 in MeOH and subsequent pro-tection of the secondary amine using di-tert-butyl dicarbonate (Boc2O) in a one-pot sequence (Scheme 6.3). This gave the corresponding piperazine 47a in an overall yield of 42%. LC-MS analysis indicated that the low yield was mainly due to the formation of the possible dimer 48 as a by-product (Scheme 6.3).

N

N

O

1) PhMgCl,

DCM, -78 oC

2) NaBH4, MeOH

3) Boc2O, -78 oC N

N

OH

Boc

45 47a, 42%

N

N

O

1) PhMgCl,

DCM, -78 oC

2) NaBH4, MeOH

3) Boc2O, -78 oC

45

possible dimer

48

PhMgCl,

DCM, -78 oC

N

N

OH

PhN

N

O

N

N

Ph

N

N

HO

Boc

Boc

OH

red. and protection

Scheme 6.3. Synthesis of N,N-diprotected substituted piperazines. We hypothesized that the dimer 48 (Scheme 6.3) had been formed by the reaction between the intermediate and the unconsumed pyrazine N-oxide, in analogy with the results discussed in Chapter 5. To overcome this problem the addition was reversed. Instead of adding the Grignard reagent to pyrazine N-oxide, pyrazine N-oxide was added drop-wise, to 2.5 equivalents of PhMgCl at -78 °C. Furthermore, the reaction of the reduction and protection time was prolonged to 25 and 60 minutes, respectively. By using this proto-col a better result was obtained and to our delight piperazine 47a was iso-lated in 91% overall yield (entry 1, Table 6.1).

53

We then followed this protocol in further investigations of reactions between different Grignard reagents and pyrazine N-oxide 45. These results are sum-marized in Table 6.1. Table 6.1. Synthesis of N-Boc, N-hydroxyl 3-substituted piperazines.

N

N

O

1) RMgX,

DCM, -78 oC

2) NaBH4, MeOH

3) Boc2O, -78 oCN

N

R

Boc

OH

45 47 entry RMgX product yield

(%)a entry RMgX product yield

(%)a

1 MgCl

N

N

Boc

OH47a

91b 8 MgCl

N

N

Boc

OH47h

67

2 MgCl

F N

N

Boc

OHF

47b 83 9

MgCl

N

N

Boc

OH47i

33

3 MgCl

MeO N

N

Boc

OHO

47c 81 10 MgBr

N

N

Boc

OH47j

76

4 MgBr

N

N

Boc

OH47d

74 11 SMgBr N

N

Boc

OH

S

47k 55

5 MgBr

F N

N

Boc

OHF

47e 72 12 MgBr

N N

N

Boc

OHN

47l

52

6 MgClO

O N

N

Boc

OHO

O

47f 87 13

MgCl N

N

Boc

OH47m

40

7 MgBr

TMS N

N

Boc

OH

TMS

47g

53 14 MgCl N

N

Boc

OH47n

54

Reaction conditions: pyridine N-oxide (1 equiv.) in DCM, Grignard reagent (2.5 equiv.), NaBH4 (1.1 equiv.) in 2 mL MeOH, Boc2O (3.0 equiv.). aYields of isolated products. bAlso performed in gram-scale resulting in piperazine 47a in a isolated yield of 76%.

Electron rich and electron poor aryls (entries 2-6, Table 6.1) gave similar good isolated yields of the the protected piperazines 47b-47f. Also a TMS protected alkyne was succefully incorporated, and yielded corresponding piperazine 47g in a 53% yield. A sterical influence could be observed when 4-biphenyl, 2-biphenyl and naphthyl magnesium chloride was reacted result-ing in 67%, 33% and 76% yields, respectively (entries 8, 9 and 10, Table 6.1). The heteroaromatic, thienyl and indole Grignard reagents were also

54

successfully reacted resulting in 55% and 52% isolated yields (entries 11 and 12, Table 6.1). Furthermore, the formation of an ortho-metalated derivative was not observed after pyrazine N-oxide was reacted with n-BuMgCl fol-lowed by trapping experiments with beenzaldehyde. Instead, the correspond-ing n-butyl substituted piperazine 47m was isolated in a 40% yield. We also reacted propenyl magnesium chloride that gave the corresponding piperazine 47n in 54% yield. Finally, we performed the reaction between pyrazine N-oxide 45 and PhMgCl in gram-scale (10 mmol). This time the correspond-ing piperazine was isolated in a 76% yield.

6.3 Orthogonal deprotection Having developed a good method for the synthesis of N-Boc protected 3-substituted hydroxyl piperazines, we went on to investigate the possibility of selectively removing the N-hydroxyl or the N-Boc-group, thus making the corresponding amine accessible for further transformation. This can be im-portant since the incorporation of piperazine fragments in pharmaceuticals often requires selective transformations on either one of the two nitrogens (e.g. Scheme 6.4).

NH

HN

HO

O

1) BocON, H2O pH= 11.0

2) CbzCl pH= 9.5

N

N

HO

O

Boc

Cbz

1) EDC,HOBt, t-Butylamine, Et3N, DMF

2) H2 Pd/C 10% MeOH

NH

N

N

O

Boc

N

N

N

HN O

tBu

tBu

HN

OHPh

OH

N

N

HN O

tBu

HN

OHPh

OHBoc1) 8N HCl

i-PrOH

2) TEA, DMF, 49

N

Cl

several steps

49=Indinavir

Scheme 6.4. Orthogonally protection and deprotection included in the synthesis of indinavir.67 As discussed previously, the N-hydroxyl can be reduced to the correspond-ing amine when treated with zinc dust in acetic acid (Chapter 5, Scheme 5.4). However, to circumvent the possible removal of the acid-labile N-Boc-group at the same time as the hydroxyl, we used MeOH as the solvent, to-gether with a few drops of acetic acid and zinc dust. In this way, the loss of the Boc group was avoided and the deprotection was completed after 4 hours at rt. Performing this reaction in the microwave at 80 °C reduced the reac-

67Dorsey, B. D.; Levin, R. B.; McDaniel, S. L.; Vacca, J. P.; Guare, J. P.; Darke, P. L.; Zugay, J. A.; Emini, E. A.; Schleif, W. A.; et al. J. Med. Chem. 1994, 37, 3443-3451.

55

tion time to only 6 minutes. The dehydroxylated piperazine was reacted, without further purification, with benzylbromide to give piperazine 50 in a 92% yield over two steps (Scheme 6.5). The selective N-Boc deprotection was accomplished by dissolving piperazine 47a in a 4M HCl saturated di-oxane solution, and stirring for 30 minutes at rt. After evaporation of the solvent, the crude residue was reacted with benzylbromide resulting in piperazine 51 in an 89% yield (Scheme 6.5). The deprotection of both the N-hydroxyl and the N-Boc protecting groups was then investigated further. The most straightforward reaction of zinc dust in 4M HCl saturated dioxane, did not yield the diamine. Instead, diamine 52 was obtained by first reacting piperazine 47a with zinc dust at pH 4, followed by adjusting the pH to 1 with a 4M HCl saturated solution of dioxane, which resulted in the isolation of piperazine 52 in a 90% yield (Scheme 6.5).

N

N

Boc

OH

Ph

1) Zn dust, MeOH, AcOH, pH = 4

2) Benzylbromid DiPEA, MeCNN

N

Ph

Boc

Ph

1) 4M HCl sat dioxan sol.

2) Benzylbromid DiPEA, MeCN N

N

OH

Ph

Ph

50, 92%51, 89%

2) 4M HCl sat dioxan sol.

1) Zn dust, MeOH, AcOH, pH = 4

47a

NH

HN

Ph

52, 90% Scheme 6.5. Orthogonal deprotection of piperazine 47a.

6.4 Enantioselective ssynthesis of substituted piperazines In an attempt to synthesize enantioenriched monosubstituted piperazines, our attention turned next to the asymmetric reaction between Grignard reagents and pyrazine N-oxides. Principally, there are two concrete alternatives to achieve this: first, by using a chiral auxillary attached to the starting mate-rial, which then directs the Grignard reagent; or second, by using a combina-tion of a chiral ligand together with the Grignard reagent. We decided to investigate the latter alternative.

6.4.1 Chiral induction with ligand in combination with Grignard

While enantioselective addition using chiral ligands in combination with zinc and/or lithium reagents, has been reported, far fewer publications ad-

56

dress the reaction with organomagnesium reagents.68 This limited success is mainly due to the high reactivity of Grignard reagents, combined with de-creased reactivity upon forming the complex with chiral ligands. Neverthe-less, a few successes have been reported concerning this topic. For example, Harada and co-workers presented a method for the asymmetric alkylation and arylation of aldehydes using a combination of Grignard, titanium tetrai-sopropoxide and the chiral ligand 3-(3,5-diphenylphenyl)-2,2’-dihydroxy-1,1’-binaphthyl (DPP-binol) (eq. 1, Scheme 6.5).69 In 2002, Chong and co-workers reported an enantioslective alkylation of aldehydes using chiral organomagnesium amides (eq. 2 Scheme 6.5).70 They took advantage of the Schlenk equilibrium to form a dialkyl magnesium species that was used in combination with chiral secondary amines. As a final example, Fu and co-workers reported a de-symmetrization of anhydrides by using Grignard rea-gents and (-)-sparteine 53 (eq. 3, Scheme 6.5).71

R H

O R'MgX, Ti(OiPr)4

DPP-binol, DCM,

-78 - 0 oCR R'

OH

OH

OH

Ph

Ph

DPP-binol =

O

Ph

O O

PhMgCl,

(-)-sparteine

tol. -78 oC Ph OH

OPhO

*

ee 88%

(eq 1)

(eq 2)

NH

N

Ph R2Mg N

N

Mg

Ph

R

R'CHO

R' R

OH

N

NH

H

(-)-sparteine 53

(eq 3)

Scheme 6.5. Different successfull enantioselective reactions using Grignard reagents and a chiral ligand.

(-)-Sparteine 53 appeared as suitable ligand, commercially available and cheap ligand (1g cost approximately 7 EUR) to start the investigation for synthesis of enantioenriched phenyl substituted piperazines. 68(a) Hoppe, D.; Hense, T. Angew. Chem. Int. Ed. 1997, 36, 2282-2316. (b) O'Brien, P. Chem. Commun. (Cambridge, U. K.) 2008, 655-667. (c) Schütz, T. Synlett 2003, 901-902. 69For alkylation see (a) Muramatsu, Y.; Harada, T. Angew. Chem. Int. Ed. Engl. 2008, 47, 1088-1090. For arylation see (b) Muramatsu, Y.; Harada, T. Chem. Eur. J. 2008, 14, 10560-10563. 70Yong, K. H.; Taylor, N. J.; Chong, J. M. Org. Lett. 2002, 4, 3553-3556. 71Shintani, R.; Fu, G. C. Angew. Chem. Int. Ed. 2002, 41, 1057-1059.

57

6.4.2 Enantioselective synthesis of substituted piperazines

PhMgCl and (-)-sparteine 53 were stirred together in DCM at rt. for 30 min-utes before cooling to -78 °C. To this mixture, pyrazine N-oxide 45 was added drop-wise. After further stirring for 10 minutes, NaBH4 in MeOH was added, followed by protection with Boc2O. This protocol gave a very prom-ising result as chiral HPLC analysis showed an 82% ee. However, due to unconsumed pyrazine N-oxide 45, piperazine 54 was isolated in a low yield of only 21% (entry 1, Table 6.2). Although low yields, the ee for the reaction prompted us to investigate further different reaction conditions. The results are summarized in Table 6.2. Table 6.2 Enantioselective synthesis of orthogonally protected 3-substituted piperazine

N

N

O

1) PhMgCl, (-)-sparteine

2) NaBH4, MeOH

3) Boc2O,

Solvent, Temp oCN

N

Boc

OH*

entry Equiv.

Grinard/sparteine solvent temperature °C time hr ee % yield (%)a

1 1.2/1.2 DCM -78 2.5 82 21 2 1.2/1.2 DCM -78 22 83 18 3 1.5/1.5 DCM -78 16 78 28 4 1.5/1.5 DCM -78 2.5 rac 37

5 1.5/1.5 DCM -78 to rt 2.5 nd - 6 1.5/1.5 Toluene -78 2.5 75 24 7 1.5/1.5 THF -78 2.5 rac 24 8 3/3 DCM -78 2.5 62 26 9 4/4 DCM -78 2.5 41 18

10 3/1.5 DCM -78 2.5 rac 36 11 3/3 DCM -78 2.5 rac 17 12 1.5/1.5 DCM/THF -78 2.5 rac 54

General procedure exemplified with entry 1: PhMgCl (1.2 equiv.) and (-)-sparteine (1.2 equiv.) was diluted with 4 mL DCM and stirred at rt. After 30 minutes the mixture was cooled to -78 °C and stirred additional 15 minutes at -78 °C, whereupon pyrazine N-oxide (1 equiv.) was added dropwise. NaBH4 (1.1 equiv.) in 2 mL MeOH was added and stirred 25 min before Boc2O (3 equiv) was added and stirred for 60 minutes. aIsolated yields.

Increasing the reaction time did not improve the isolated yield, and the per-centage ee remained unchanged (entry 2, Table 6.2). A slightly better yield was obtained when the Grignard/sparteine complex equivalent was increased to 1.5, but still not satisfactory (entry 3, Table 6.2). However, when pyrazine N-oxide was pre-treated with (-)-sparteine, followed by addition to PhMgCl at -78 °C, a 37% yield was obtained, but unfortunately only as a racemic mixture (entry 4, Table 6.2). When the temperature was allowed to reach rt., a complex reaction mixture was obtained (entry 5, Table 6.2), and by chang-ing the solvent to toluene, similar results were obtained to those when DCM was used as the solvent (entry 6, Table 6.2). However, this was not the case when using THF, which resulted in the isolation of the racemate in a 24%

58

yield (entry 7, Table 6.2). We also studied the effect of increasing the Grig-nard/sparteine equivalents by reacting 3 and 4 equivalents, but this decreased enantioselectivity and did not change the yields (entries 8 and 9, Table 6.2). We then tried to force the consumption of the pyrazine N-oxide by a sequen-tial addition of another 1.5 equivalents of Grignard and/or the Grig-nard/sparteine complex. By doing so, we managed to increase the yields slightly but, to our surprise, we obtained the racemate (entries 10 and 11, Table 6.2). This finding suggests that the intermediate so formed, epimerizes upon the sequential addition of more Grignard or Grignard/sparteine com-plexes. Finally, we decided to use a combination of sparteine and PhMgCl*LiCl, the latter generated from the addition of LiCl to PhMgCl. In this instance, the total consumption of the starting material was achieved, but again, only the racemate was obtained (entry 12, Table 6.2). The results reported here are only representative of some initial studies from an on-going project in our laboratory. There are still many things that can be tested in an effort to improve the ee and the isolated yield of the reaction.

6.5 Conclusion and outlook In this chapter, we have described the first synthesis of 2-substituted piperazines by reacting Grignard reagents and pyrazine N-oxides. In fact, this is, to our knowledge, the first reported reaction between Grignard rea-gents and pyrazine N-oxides giving piperazines. The reaction is a scalable one-pot strategy that includes the addition of pyrazine N-oxide to a Grignard reagent at -78 °C followed by reduction with NaBH4 and then protection of the secondary amine with Boc-anhydride. The use of aryl, vinyl and alkyl Grignard reagents in the reaction, resulted in differentially protected, 3-substituted piperazines. Furthermore, this compound could be orthogonally de-protected, which allowed further derivatisation on either of its two nitro-gens. Finally, we presented some initial studies of the enantioselective synthesis of these orthogonally protected substituted piperazines. While the initial results regarding the enantiomeric excess are promising, the isolated yields are unfortunately still to low.

60

7

Concluding Remarks

This study of compounds derived from reactions between Grignard reagents and pyridine and pyrazine N-oxides, that has formed the basis of this thesis, has clearly demonstrated these reactions to have been underestimated since reports in the 1970’s established them to be simply low yielders of dienal-oximes. Initial confirmation of these previously reported results was fol-lowed by our improvements of the method that enabled us to establish high yielding synthesis of a diverse range of substituted dienal-oximes. Further studies of these compounds confirmed them to be excellent intermediates in the preparation of other interesting compounds such as saturated amines, nitriles or enaminones. During the development of our method of synthesiz-ing dienal-oximes, we observed them to be prone to ring-closure upon heat-ing. This led us to explore further the usefulness of dienal-oximes in the synthesis of substituted pyridines. We found that if dienal-oximes were treated with acetic anhydride and irradiated with microwaves, the corre-sponding substituted pyridines could be obtained. Furthermore, the abun-dance of commercially available pyridine N-oxides allowed us to use 4-benzyloxy pyridine N-oxide, which meant that the scope for the reaction could be expanded even further to include the synthesis of substituted 4-pyridones and 4-aminopyridinium salts. Aryl Grignard reagents consistently performed better than alkylmagnesium halides in the reactions. The reason for this proved to be a competitive meta-lation reaction. Earlier problems encountered in the deprotonation of pyri-dine N-oxides using n-BuLi led us to investigate the possibility of executing this proton abstraction with alkylmagnesium chlorides. However, it appeared that good yields were only obtainable when the pyridine N-oxide was substi-tuted with an electron-donating group in the third position, although, the reaction tolerated several different electrophiles e.g. aldehydes, ketones, isocyanates or halogens. Furthermore, the usefulness of incorporating iodine was realized in a Suzuki-Miayura coupling reaction when eight different arylsubstituted N-oxides were synthesized. We made our most noteworthy discovery during the investigation of metalation. We found that attempts to metalate pyridine N-oxide with PhMgCl, followed by the addition of deuter-ated methanol, did not result in the expected 2-deuterated pyridine N-oxide. Instead, the reaction yielded an incorporation of deuterium in the 5-position. We discovered that this reaction could be used to synthesize 2-substituted piperidines after reduction of the obtained tetrahydropyridine. Further stud-

61

ies of this reaction showed that, if the electrophile was changed to an alde-hyde or ketone, a regio- and stereoselective synthesis of trans-2,3-dihydropyridine N-oxides could be obtained. This novel heterocycle that, with an additional reduction gave the corresponding piperidine, could also be reacted in a Diels-Alder reaction to yield the aza-bicyclo system. In chapter 6, we showed that the method developed for the synthesis of sub-stituted pyridines and piperidines, could also be used for the synthesis of substituted piperazines. In that chapter we reported how pyrazine N-oxides were reacted with different Grignard reagents at -78 °C, followed by reduc-tion and N-Boc protection, to yield the N,N-diprotected piperazines. This differentially protected piperazine could be orthogonally de-protected, giv-ing opportunities for synthetic modifications at either nitrogen. Finally, we have shown promising results for the enantioselective synthesis of substi-tuted piperazines using a combination of PhMgCl and (-)-sparteine.

63

8 Acknowledgements

Det är många som ska tackas, här är några:

Fredrik Almqvist, för att du trodde på mig och vågade låta mig driva detta projekt. Din idéri-kedom och inspirerande teorier har ständigt drivit mig framåt. Du har till och med lyckats inspirera mig till att heja på Skellefteå AIK.

Roger Olsson, som ständigt tagit sig tid att hjälpa mig med forskningen. Vet inte hur många veckor och helger som gått bara på att svara på mina majl och konstiga frågor. Ditt enorma kunnande och intresse för kemi har verkligen boostat tillvaron. Ett extra stort tack för att ni på Acadia-Pharmaceuticals har sponsrat min forskning och att ni lät mig få möjligheten att labba där ett halvår.

Även ett stort tack till sidekickern från Acadia, Magnus Gustafsson, du kan allt om google.

Nils Pemberton som korrekturläst avhandlingen. Jag är så sjukt glad att jag började i Umeå och därför fick möjligheten att lära känna dig. Ditt enorma pepp har betytt massor under åren. Glädjer mig dessutom att jazzen kryper sig närmare varje dag.

Erik Chorell som hjälpt till med avhandlingen, fast mer viktigt influerat med sitt lugn och därför fått en skåning att förstå att även vintern i norrland tar slut. Tack också till Elin Chorell för otroligt trevliga middagar, fester och dylikt (du e grym på Guitar hero).

Alla labb-kompisar genom åren TACK: Thomas Gustafsson för fiske och kemikunskaper (jag brukar få mer fisk än dig), David Blomberg för den bästa separertratten som finns, Veronica Åberg för härliga surströmmings sessioner, Hans Emtenäs för att jag fick göra exjobb, Andre-as Larsson som kanske inte riktigt vet hur man släcker en eld, Matthias Hedenström all hjälp med NMR, Magnus Sellstedt för att du har flest konstiga idéer, Ida Andersson för att ditt knä mår bra, Christoffer Bengtsson för att du alltid gör störst skala, Sofie Edvinsson för ny syn-tespepp.

Thanks Sajal Das and Thomas Banchelin (Cheese).

Dan Johnels för mycket kemi snack men också massa fina foton. Mikael Elofsson som man alltid kan snacka jazz med. Anton Lindström som verkligen har förmågan att göra Umeå så mycket roligare. David An-dersson för beräknings hjälp.

De gamla vännerna från söder. Utan er hade jag totalt tappat bort mig själv i kemin. Tack Paddy & Madde och Thomas och Hedvig för att ni påminner mig om vad enkelt livet är.

Min bror Johan som är en stor anledning till att jag valde Umeå framför Lund? Hursom, tack för stöd tankar och vettiga resonemang. Nu gäller det bara att ni (Sara, Axel, Hugo) hittar till rätt kust.

Mina föräldrar, tack pappa för att du är kemiingenjör men ändå alltid får mig att känna sig som professor i kemi. Mamma för enorma tryggheten som du ger mig. Ni båda har förmå-gan att när ni såväl besöker som välkomnar en, får en att glömma allt som har med kemi att göra. Ni är fantastiska föräldrar.

Min underbara familj Cathis och Gunnar. Ni har tagit hand om mig under skrivandet på ett sätt som gjort detta möjligt! Ni är det viktigaste för mig och jag älskar er över allt annat.

reaction between grignard reagents and …232682/...title reaction between grignard reagents and...

Documents