Phd Thesis J.C. Van Der Toorn

The use of aromatic selenides as oxidation catalysts

John C. van der Toorn “You can see forever Look inside of your mind Find a sense, another wonder Just release the fears you left behind”

Cover design: Brigitte van Loon The use of aromatic selenides as oxidation catalysts

Proefschrift

ter verkrijging van de graad van doctor aan de Technische Universiteit Delft, op gezag van de Rector Magnificus Prof. ir. K.C.A.M. Luyben voorzitter van het College voor Promoties, in het openbaar te verdedigen op 28 oktober 2011 om 15.00 uur

door

Johannes Cornelis VAN DER TOORN

Master of Science in de Chemie Geboren te Ridderkerk Dit proefschrift is goedgekeurd door de promotoren:

Prof. dr. R.A. Sheldon Prof. dr. I.W.C.E. Arends

Samenstelling promotiecommissie:

Prof. ir. K.C.A.M. Luyben Rector Magnificus, voorzitter Prof. dr. R.A. Sheldon Technische Universiteit Delft, promotor Prof. dr. I.W.C.E. Arends Technische Universiteit Delft, promotor Prof. dr. J.I. Garcia Universidad de Zaragoza, Prof. dr. F. Kapteijn Technische Universiteit Delft Prof. dr. R.J.M. Klein Gebbink Universiteit Utrecht Prof. dr. K. Lammertsma Vrije Universiteit van Amsterdam Dr. G. Kemperman (voorheen) MSD, Oss, adviseur (thans) Axxence Aroma Chemicals, Emmerich Prof. dr. ir. H. van Bekkum Technische Universiteit Delft, reservelid

ISBN: 978-90-5335-461-2

The research described in this thesis was financially supported by the Netherlands Ministry of Economic affairs and the B-Basic partner organizations (www.b- basic.nl) through B-Basic, a public private NWO-ACTS program (ACTS = Advanced Chemical Technologies for Sustainability).

All rights reserved. No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system, without written permission from the author. Contents

Chapter 1: Introduction to oxidations with selenium compounds ...... 1 The use of selenium in chemistry, from old to new ...... 2 Epoxidations with selenium compounds ...... 3 Baeyer-Villiger reactions catalyzed by selenium compounds ...... 7 The oxidation of alcohols with organoselenium compounds ...... 8 Additional selenium catalysed oxidations ...... 11 Selenium and hydroperoxides in Nature ...... 13 Selenosubtilisin: a chemically modified protease as GPx mimic ...... 13 Hybrid enzymes ...... 16 Scope of the thesis ...... 20 References ...... 21

Chapter 2: Diphenyl diselenide-catalyzed selective oxidation of alcohols with tert-butyl hydroperoxide ...... 25 Abstract ...... 25 Introduction ...... 26 Results and discussion ...... 28 Diphenyl diselenide mediated oxidation of benzyl alcohol; catalytic and stoichiometric reactions ...... 28 Reaction calorimetry ...... 31 Conclusions ...... 35 Experimental section ...... 37 References ...... 38

Chapter 3: Spectroscopic investigations on the activation of diphenyl diselenide by tert-butyl hydroperoxide39 Abstract ...... 39 Introduction ...... 40 Results and discussion ...... 40 UV-Vis spectroscopy ...... 40 In-situ infrared and Raman ...... 41 Solvent influence on the oxidation of diphenyl diselenide by TBHP ...... 45 Conclusions ...... 48 Experimental section ...... 49 References ...... 50 Chapter 4: Substituted aromatic diselenides in the catalytic oxidation of alcohols ...... 51 Abstract ...... 51 Introduction ...... 52 Results and discussion ...... 53 Substituted diselenides as precatalysts in oxidations with TBHP ...... 53 Substituted diselenides in the catalytic oxidation of benzyl alcohol ...... 55 Oxidation of an aliphatic alcohol: 1-decanol ...... 56 Conclusions ...... 59 Experimental section ...... 61 References ...... 63

Chapter 5: Glycerol-based solvents as green reaction media in selenium catalyzed epoxidations with hydrogen peroxide ...... 65 Abstract ...... 65 Introduction ...... 66 Results and discussion ...... 68 Epoxidation of cyclohexene ...... 71 Conclusions ...... 74 Experimental section ...... 75 References ...... 75

Chapter 6: The design and synthesis of selenium catalysts capable of selective and irreversible attachment to enzymes ...... 77 Abstract ...... 77 Covalent versus non-covalent approach in hybrid catalysts ...... 79 Choice of enzyme class: hydrolases ...... 81 Probing the active site ...... 82 Serine hydrolases ...... 82 Design of functionalized aromatic selenium compounds for enzyme inhibition ...... 84 Results and discussion ...... 87 Compatibility of the selenoether with oxidation catalysis ...... 87 Towards synthesis of phosphonate esters ...... 89 Synthesis of Se-compounds for selective functionalization of cysteine hydrolases ...... 92 Conclusions ...... 93 Experimental section ...... 95 References ...... 101 Chapter 7: Enzymes as scaffolds for aromatic selenium catalysts ...... 105 Abstract ...... 105 Introduction ...... 106 Selenosubtilisin ...... 106 Results and discussion ...... 108 Testing the hybrid in oxidation catalysis ...... 111 The use of selenium hybrids in oxidative bromination reactions ...... 113 Conclusions ...... 114 Experimental section ...... 116 References ...... 117

Summary of the thesis ...... 119

Samenvatting ...... 123

Dankwoord ...... 127

List of Publications ...... 131

Curriculum Vitae ...... 132

Chapter 1: Introduction to oxidations with selenium compounds

John C. van der Toorn, Roger A. Sheldon and Isabel W. C. E. Arends

Selenium is an abundant main group element which can be found almost everywhere in the world in different inorganic forms, but the most important form 2- is the selenate anion SeO4 . The element was discovered by Berzelius and Gahn in 1817 while they were studying the chemicals formed during the production of sulfuric acid. First they thought that they had found tellurium (earth), but closer inspection by Berzelius himself revealed that the nature of the found element was slightly different and he renamed the element to the Greek word for moon: σεληνη.1 The first major application for selenium was in selenium rectifiers, which were able to convert alternating electricity current into direct current. Selenium is used currently in glass manufacturing, either to discolor the glass by neutralizing the effect of iron (which gives glass a green color) or simply to color the glass ruby red. For animals and humans, regular intake of selenium is essential even though it can be toxic. Chronic toxicity of selenium is better known as selenosis and the symptoms are garlic breath, hair and nail loss, gastrointestinal problems, skin rash and nervous system abnormalities.2 Even though the numbers are somewhat controversial, the recommended dietary allowance for selenium is 0.055 mg per day. An upper limit is defined with an intake of 0.4 mg per day, which shows that the window of daily intake is rather narrow.3 Low levels of intake of selenium may Chapter 1 contribute to a type of juvenile cardiomyopathy, known as Keshan disease. There is definite proof that selenium plays a major role in cell cycle and apoptosis. The form and doses are a key element in the targets and effectiveness of selenium compounds.4,5 In the human body there are three main groups of selenium containing enzymes: specific selenocysteine-containing selenoproteins, non-specifically selenium incorporated enzymes and specific selenium-binding proteins. Selenium is in tissues mainly present as selenocysteine, selenomethionine and in lower quantities as the metabolic precursors of these compounds. The incorporation of selenomethionine as a substitute for methionine is random in enzymes. For the incorporation of selenocysteine a sophisticated trick in the translation machinery is needed, as this amino acid is coded as UGA which is normally a stop codon.6 A serine-bearing tRNA is transformed in situ to a selenocysteine-bearing tRNA via a reactive selenium donor compound, known as monoselenophosphate.7 In this introduction, we will only focus on the aforementioned first class of selenoproteins. The proteins reported which have a specified function are several glutathione peroxidases, iodinases, thioredoxin reductase and selenophosphate synthetase-P. The majority of these proteins are involved in redox reactions. Due to the pKa value of selenocysteine, which is 5.2 at physiological pH, the selenols in selenocysteine-containing proteins are present as selenolates. The oxidation states are mainly in the selenol and selenenic acid form and reactive diselenides have not yet been found in proteins.8 The best known selenoprotein is glutathione peroxidase, which protects cells from oxidative damage by catalyzing the reduction of peroxides using glutathione as the reductant. Later in this chapter this subject will be discussed further.

The use of selenium in chemistry, from old to new As described above, selenium compounds had already been discovered in the nineteenth century, but it took quite a while before selenium was used on a regular basis by organic chemists. The only applications were in using selenium dioxide as an oxidant and elemental selenium as a dehydrogenating reagent. In the 1970’s the interest grew in using organoselenium compounds in synthetic chemistry when the syn-elimination reaction was discovered. Selenium has the ability to easily oxidize from the Se(II) to the Se(IV) species, without overoxidizing 2 Introduction to Se(VI) when the circumstances are controlled.9 Because selenenyl halides and pseudohalides (RSeX) are soft, powerful electrophiles, these compounds are highly suitable for introducing the selenium moiety in unsaturated substrates by electrophilic addition.10 Selenols and their conjugate bases on the other hand, are soft nucleophiles and can thus be added to electrophilic substrates.11 This shows the versatility of selenium compounds in synthetic organic chemistry (figure 1).

Addition of electrophilic organoselenium reagents R RSeX Se X- or Nu- SeR SeR or X Nu

Addition of nucleophilic organoselenium reagents LG - SeR' R'Se - EWG RSe RSe EWG R - LG R

Figure 1 Application of organoselenium reagents in organic synthesis. LG = leaving group, EWG = electron withdrawing group.

For more recent advances in organoselenium chemistry, we would like to refer to the general literature.9,12,13 The focus in this thesis, and thus in this introduction, will be on oxidations with selenium compounds.14

Epoxidations with selenium compounds One of the most important reactions in organoselenium chemistry is the selenoxide elimination reaction.15,16 When the β-carbon of one substituent of a selenoxide bears a hydrogen atom, the structure becomes unstable and an olefin is formed. Because of the mild conditions this reaction has led to many synthetic applications. One of the most important findings is that the elimination runs in a syn-fashion, which was most apparent in the selenoxide elimination of 6-β- phenylseleninylcholestane as the reaction yielded only cholest-6-ene (figure 2).17

3 Chapter 1

C H C H 8 17 8 17 Ph Se Me H H O H H Se(O)Ph

Figure 2 The selenoxide syn elimination yielding cholest-6-ene

When Grieco and co-workers used this reaction in one of their synthetic schemes, they found that with eight equivalents of H2O2, not only their desired elimination product, but also epoxide was formed in a rather large yield. They attributed this to the formation of phenylperseleninic acid, which is formed under these conditions (figure 3).18

SePh H2O2 H H O H O O O O O O PhSeOH O O H2O2 H2O2 PhSeOH PhSeOOH Ph2Se2

Figure 3 Epoxide formation during selenoxide elimination

This turned out to be a very attractive reaction for further exploration. Classically, perbenzoic acids such as meta-chloroperbenzoic acid (mCPBA) are used for these reactions and the subject of catalytic epoxidations with organoselenium compounds was further explored in the late 1970’s by the groups of Reich and Sharpless.19-21 Both groups found that o-nitrophenylseleninic acid and 2,4-dinitrophenylseleninic acid were the best catalysts. Hydrogen peroxide was used successfully and the best results were obtained when the water was scavenged by means of MgSO4 (figure 4).

4 Introduction

10% cat, 30 % H2O2 o (MgSO4), CHCl2, 0 C to RT O

NO2

SeO H SeO H SeO2H 2 2

91% 79% 0% CO2H 99%a NO2

SeO H O2N SeO2H O2N SeO2H 2 10 0% 96% 50% 99%a

SeO2H Cl SeO2H SeO2

65% 40% 2%

a Figure 4 Epoxidation yield of cyclooctene with selenium catalysts. MgSO4 was added. The yield of the reaction is shown in bold below the catalysts.

Two groups have tried to apply the recovery of the selenium catalyst. One way was to immobilize the selenium catalyst on a polymer.22 Unfortunately, the epoxidation reaction only yielded vicinal diols, but the catalyst could be used several times without any loss in activity. Another way of recycling selenium catalysts is via fluorous biphasic catalysis, which has been performed by Knochel and co-workers. By adding fluorous ponytails to the phenylseleninic acid, they were able to perform the reaction in a bromoperfluorooctane / benzene mixture with 23 60% H2O2 as the oxidant (figure 5).

O SeOH

F C C F 17 8 5 mol% 8 17 O 60% H2O2, C8F17Br / benzene 63-93% 70 oC, 1-12 h

Figure 5 Selenium catalyzed epoxidations in perfluorinated solvents 5 Chapter 1

More recently, this catalytic epoxidation has been revisited. A systematic study for the best solvent and the best arylseleninic catalyst learned that 2,2,2- trifluoroethanol in conjunction with bis[3,5-bis(trifluoromethyl)phenyl diselenide was the best catalytic system for the epoxidation of a variety of substrates.24 The initial rates of these reactions had TOF (Turn Over Frequency) values up to 420 per hour. Solvents have a large impact on the yield and the selectivity of the reaction. 2,2,2-Trifluoroethanol was by far the best solvent for this reaction and it was found later that this solvent activates H2O2 through a hydrogen bond network with several solvent molecules.25 The best epoxidation catalysts in terms of substitution pattern and groups on the aromatic ring turned out to be the trifluoromethyl- and nitrosubstituted perseleninic acids (figure 6).

1% mol cat, H O O R' 2 2 R' R R CF3CH2OH, RT, 1h

F3C F3C O2N

SeO2H SeO2H SeO2H 98% 81% 80% F3C CF3 CF3

F3C SeO2H F3C SeO2H O2N SeO2H 85% 82% 75% CF3 NO2

SeO2H SeO2H F SeO2H

73% 83% 64% F F

Cl SeO H 2 F SeO2H SeO2 70% 2% F 2% F

Figure 6 Selenium catalyzed epoxidations with various aromatic catalysts. Yield of 1,2- epoxycyclohexane (formed from cyclohexene) is shown below the catalyst in bold.

6 Introduction

Recently, diphenyl diselenide was reinvestigated for the olefin dihydroxylation with hydrogen peroxide.26 The formation of the diol diastereomers was one of the main interests and depending on whether the substrate could stabilize the tertiary carbocation, there was a preference for one of the products. The use of a chiral diselenide yielded, at the appropriate conditions, an excellent enantiomeric excess of one of the products (figure 7). The authors do state that under these circumstances, a large amount of the catalyst is necessary and one can question the catalytic scope of this reaction.

S Ph Se Ph OH Ph OH 2 OH OH 0.5 eq

H2O2, 1 H2O / MeCN 3/1 2 3 T = 23 oC Yield 78% dr 80/20 ee (2) 12% T = -10 oC Yield 56% dr 68/32 ee (2) 92%

Figure 7 Selenium catalyzed dihydroxylation with a chiral diselenide

Baeyer-Villiger reactions catalyzed by selenium compounds Following epoxidations with perseleninic acids, it is not surprising that Baeyer- Villiger reactions, that require similar activation of hydrogen peroxide, have been studied as well. It was found that during the syn-elimination reaction of α- phenylselenocycloalkanones a Baeyer-Villiger reaction could take place under certain circumstances.27 The oxygen is introduced on the side of the more substituted carbon atom. The only exception is the oxidation of aromatic aldehydes as the electronic substituents (and the solvent) dominate the outcome of the reaction towards the phenol or the acid.28,29 The Baeyer-Villiger reaction only works under the influence of hydrogen peroxide. Many selenium catalysts have been tested in this reaction and the best catalysts have electron withdrawing groups on the aromatic ring.29-31 This reaction has also been performed with the polystyrene immobilized catalyst (vide supra) in order to make recyclability of the catalyst more straightforward.22 More recently, the oxidation was tested under fluorous biphasic and triphasic conditions which also showed that the recyclability 7 Chapter 1 of the especially designed selenium catalyst was excellent.32 In synthetic organic chemistry and in biocatalysis, the enantioselective Baeyer-Villiger reaction has always received attention.33-37 There is one example in the scientific literature of an optically active selenium compound which can mediate a very modest enantioselectivity in the oxidation of 3-phenylcyclobutanone (figure 8).38 Also, the reaction required a Lewis acid and only the presence of ytterbium triflate provided the reaction outcome with enantioselectivity, which should lead to the conclusion that this metal is most probably part of the catalytic reaction. Even though many chiral diselenides have been developed, examples of catalytic enantioselective oxidation reactions with these selenium compounds have been scarce in literature.

O O 1 mol% cat, O 2 mol % Yb(OTf) 3 O Se N 30% H2O2, THF, rt Ph 2 Ph Yield 54% ee 19% OTBDMS

Figure 8 Selenium catalyzed Baeyer-Villiger oxidation with a chiral diselenide

The oxidation of alcohols with organoselenium compounds Benzeneseleninic anhydride (BSA) is an affective oxidizing agent for the selective oxidation of alcohols to the aldehyde or the ketone.39 The mechanism of this reaction is comparable to the selenoxide syn-elimination. In the oxidation of cyclohexanols, the oxidation can yield the cyclohexenone. In the presence of a catalytic amount of BSA and stoichiometric quantities of iodosobenzene, the oxidation of cyclohexanones to a conjugated double bond can be performed twice.40 The catalytic oxidation of alcohols with selenium compounds was studied by Kuwajima and co-workers. They found that simple diphenyl diselenide in combination with stoichiometric quantities of tert-butyl hydroperoxide (TBHP) was capable of the oxidation of benzylic and allylic alcohols with high yields.41 In a follow-up article, the same authors also studied effects of substituents on the aromatic ring of the selenium catalysts.42 In the oxidation of trans-2-hexen-1-ol, bis (p-chlorophenyl)diselenide gave the best results. Neither electron donating substituents (o,p-dimethoxy or o- or p-methoxy) nor electron withdrawing substituents (o-nitro or m-trifluoromethyl) influenced the reaction positively. The 8 Introduction authors proposed two possible mechanisms for the catalytic oxidation via the seleninate or the selenenate ester (figure 9). In another article, the same authors suggest that mixing an arylselenide with a stoichiometric amount of TBHP leads to a high concentration of the seleninic anhydride.43 However, this suggestion was opposed when this oxidation was followed by 19F NMR spectroscopy in the oxidation of p-fluorophenyl diselenide. It was found that this oxidation only yielded a mixture of the parent diselenide and the seleninic anhydride.44

O R R R1 R R 1 R 1 1 1 ArSeO 1 OH ArSeO O OH O R R2 R2 2 R2 R2 R2

O 1 ArSeOH ArSeOH /2 ArSeOSeAr ArSeH

tBuOH tBuOOH tBuOH tBuOOH

Figure 9 Proposed mechanisms for the catalytic oxidation of alcohols

In the oxidation of saturated aliphatic alcohols, the reaction could not be performed catalytically, but 0.5 equivalents of dimesitylene diselenide gave satisfying results in these oxidations using 1.5 equivalents of TBHP. The previously described polystyrene-immobilized selenium catalyst was also tested in the oxidation of alcohols and indeed the selective oxidation of activated alcohols was viable using TBHP as the terminal oxidant in carbon tetrachloride.22 It was not until 1996 that the catalytic oxidation of alcohols with selenium compounds was revisited. When diphenyl diselenide is mixed with chloramine-T, an analog of benzeneseleninic anhydride is supposedly formed (figure 10).45 O O Tos S Cl N NTos N NTos Se Se Se 2 Na Ph Ph

Figure 10 Oxidation of Ph2Se2 with Chloramine-T

9 Chapter 1

In preliminary experiments, it was found that the oxidation of 2-octen-4-ol gave

48% yield of the corresponding ketone in the presence of 10 mol% Ph2Se2. By changing the oxidant to N-chloro-4-chlorobenzenesulfonamide sodium salt and varying the diselenide, the catalytic oxidation of secondary alcohols could be carried out with only 1-3 mol% of the catalyst. The best catalysts were diselenides bearing an ester group on the ortho-position, and the position of these ester groups was of vital importance. If the ester group was located on the para- position the catalytic activity declined drastically. This catalytic oxidation was further optimized by changing the ester group to a pyridyl group, of which the nitrogen could coordinate to the selenium.46 This catalyst could be used in even lower quantities (down to 0.1 mol%) in the oxidation of octan-2-ol, 1- phenylethanol and geraniol. The oxidation of 1-decanol remained problematic, even though conversions were better with this system. The proposed mechanism is depicted in figure 11.

ArSO2NH2 R1 Se O R1 N SO2Ar OH O H R2 R2 R1 R2

N N

Se Se N SO Ar NH 2 NH

SO2Ar SO2Ar NaCl ArSO2NClNa

Figure 11 Catalytic cycle of alcohol oxidations with bis[2-(2-pyridyl)phenyl] diselenide and N- chloro-4-chlorobenzenesulfonamide sodium salt.

Even though selenium compounds have been used as ligands in copper catalyzed oxidations,47 there have been no other publications using selenium compounds as the sole catalysts for the oxidation of alcohols.

10 Introduction

Additional selenium catalysed oxidations Allylic oxidations with selenium compounds The oxidation of activated methyl groups and of allylic carbons has traditionally been carried out with selenium dioxide. However, the selectivities and yields of these reactions are generally quite low.48 One of the most problematic issues is that there are a lot of selenium-containing by-products formed and the use of 49 catalytic amounts of SeO2 only partially solves this problem. In later research it was found that the use of organoselenium compounds as catalysts gave much less selenium-containing by-products. With the use of 2-(N-oxido)pyridineseleninic anhydride as a catalyst and TBHP as the oxidant, β-(-)-pinene was oxidized to either the alcohol or the ketone (or a mixture), depending on the reaction conditions.50 Unfortunately the anhydride was highly explosive. A safe alternative proved to be the use of bis(2-pyridyl) diselenide as catalyst in combination with the terminal oxidant iodoxybenzene. Also pentafluorobenzeneseleninic acid was capable of the same reaction and other compounds could be oxidized catalytically when TBHP was used as the terminal oxidant (figure 12).51

OH O (cat) Se-compound and / or R R' R R' R R' F F O O SeO2 or F Se or Se O or Se OH N N F F O 2 2 PhIO2

Figure 12 (Organo)selenium compounds capable of catalytic allylic oxidations

Benzeneseleninic anhydride is capable of oxidizing activated C-H bonds such as benzylic hydrocarbons at elevated temperatures.52 In some cases, the aromatic group becomes selenylated which is presumably caused by the PhSeO+ cation. In more recent years, Crich and co-workers have been working on a recyclable fluorous seleninic acid for allylic oxidations and the oxidations of aryl alkyl ketones to the corresponding ketoacids (figure 13).53-55

11 Chapter 1

R R1 R R1 1. C8F17SeO2H (10 mol%) PhIO , PhCF , !T O 2 3 O 2. Na2S2O5 R 3. Fluorous extraction OH O O

Figure 13 Catalytic oxidations with fluorous recyclable selenium catalysts

Another catalytic option is the oxyselenenylation of unsaturated compounds in combination with oxidative deselenylation. In this way it is possible to make double-bond transpositioned allylic alcohols and ethers from alkenes. This strategy has been employed by several groups and in all cases 10 mol% of a diselenide was used in conjunction with peroxidisulfates for these reactions.56

Oxidative halogenation reactions With arylseleninic acids and -selenoxides it is also possible to oxidize halide salts such as bromide and chloride,57-59 but the use of N-bromo- or N-chloro- succinimide as the halogen source has also been reported.60

- O Br HO- O O ArSeOBr Br2 CO2H Br - O O Br O H2O2 Ph Ph O ArSeOH ArSeOOH Br CO H 2 Ph HOBr Br- OMe Ph OMe Br HO OBr HO- Br- HO- ArSeR' Br MeO OMe MeO OMe 2 Br- O H2O2 HO OOH ArSeR' ArSeR' N O Br N O HO- HOBr Br-

Figure 14 Selenium catalyzed bromination, proposed mechanisms and reactions

When arylseleninic acids are used as the catalysts, it was found that for bromo- lactonization and electrophilic bromination the catalysts needed to have a slightly different substitution pattern than for the Baeyer-Villiger and epoxidation

12 Introduction reactions. This is most probably due to the fact that the peroxyseleninic acid is not the active species in this reaction, but this option can not be ruled out (figure 14).

Selenium and hydroperoxides in Nature As described earlier in this chapter, selenium in the human body is incorporated as selenomethionine and selenocysteine. The latter is often incorporated in enzymes to make the protein backbone redox active. The most well studied selenium- containing enzyme is glutathione peroxidase. The catalytic cycle of glutathione peroxidase (GPX) is considered to involve the oxidation of the selenocysteine residue by hydroperoxide, forming selenenic acid, which is then further converted to a selenenylsulfide by reacting with glutathione (figure 15). This intermediate reacts with another glutathione, which delivers the disulfide and the reduced selenol in the enzyme.61 Together with catalases, glutathione peroxidases protect cells from oxidative damage as many types of peroxides can be present in the cell, formed via oxidative stress. Many synthetic selenium compounds show GPX- activity and in recent years, much research has been devoted to the optimization of the synthetic structure for GPX mimics.62

H O GSH 2 Enz SeSG

GSH GSSG

Enz SeOH Enz SeH

ROH ROOH

Figure 15 Crystal structure of glutathione peroxidase, with selenocysteine in space-filling mode, in oxidized form (pdb: 1gp1). The catalytic cycle is shown (GSH = glutathione).

Selenosubtilisin: a chemically modified protease as GPx mimic Before serious effort was put into the synthesis of small selenium containing compounds as GPX-mimics, a different approach was tested for bringing the redox activity of the selenium moiety into biomolecules. Subtilisin is a serine protease which can be selectively inhibited by phenyl methyl sulfonyl fluoride (PMSF). The 13 Chapter 1 intermediate sulfonic ester can undergo an SN2 displacement by small nucleophiles. After the inhibition of subtilisin with PMSF, Hilvert and co-workers found that selenide (NaSeH) was capable of displacing the sulfon ester, thus creating selenosubtilisin (SeSub, figure 16). 63 This semi-synthetic enzyme turned out to be a poor catalyst for amide hydrolysis, but it is still capable of cleaving activated acyl derivatives. The intermediate selenol ester undergoes aminolysis faster than normal esters and thiol esters and it is thus possible to use SeSub for the synthesis of amide bonds.

PMSF O NaSeH Sub OH Sub O S Sub SeH O

Figure 16 Synthesis of selenosubtilisin

Besides the synthesis of amide bonds, SeSub was capable of the reduction of hydroperoxides, thus acting as a glutathione peroxidase mimic.64 The reduction of

TBHP with SeSub was almost 70,000 times more effective than with Ph2Se2. There have been numerous studies on the activity of SeSub.65,66 The crystal structure of SeSub was resolved at a resolution of 2 Å and this showed that the chemical modification had only altered the enzyme active site and the structure of the enzyme was completely intact.67 By means of molecular modeling it was shown that once the Se-thiol intermediate is formed, the active site is very crowded, making the insertion of another molecule from solution sterically demanding. The use of other types of subtilisin showed that there is indeed a great influence of the choice of enzyme source on the kinetics of the reduction.68 Even though this work was impressive, the authors neglected to test the potential of SeSub for enantioselective reactions. It was found later that SeSub was capable of the kinetic resolution of a range of racemic hydroperoxides and kinetic data was derived (figure 17 and table 1).69 OOH OOH OH SeSub 2 ArSH ArSSAr H2O R1 R2 R1 R2 R1 R2

Figure 17 Kinetic resolution of racemic hydroperoxides with SeSub 14 Introduction

Table 1 Kinetic resolution of racemic hydroperoxides with SeSub

-1 Entry Substrate t (min) R:S (peroxide) R:S (alcohol) Km (mM) kcat (min ) OOH A 12 76:24 20:80 15,7 2125

OOH

B 18 74:26 22:78 6,0 1723

Cl OOH C OH 3 1:99 99:1 2,1 2443

OOH D 7 30:70 71:29 18,0 1745

In later work, SeSub was synthesized from different sources such as detergent additives,70 and differences in enantioselectivity between enzyme sources were also tested.71 The immobilization of the enzyme in CLECs (cross-linked enzyme crystals) showed that the enzyme could be reused several times without a (substantial) loss in activity and selectivity.72 Finally, the SeSub was tested with even more substrates and showed great enantioselectivity in certain cases.73 This semi- synthetic enzyme is a great example of how a chemical modification of an enzyme converts the natural activity into another useful function. In this example, the enantioselectivity of the enzyme is complementary to that of enzymes found in Nature. The chemical modification of proteins on the outer shell of the enzyme is a standard laboratory technique nowadays, and selectively targeting enzyme active sites is also reasonably straightforward. However, modifying the active site in such a way that it will perform other types of reactions compared to the native activity of the enzyme is something different and SeSub is one of the most underestimated examples in this respect.

15 Chapter 1

Hybrid enzymes The concept of combining chemical and enzyme catalysts has been demonstrated as early as 1978 when George Whitesides and co-workers already established the proof of principle for this approach. They inserted a non-chiral Rh-hydrogenation catalyst into (strept)avidin by means of a biotin handle.74 This new hybrid enzyme was capable of the enantioselective reduction of α-acetamidoacrylic acid with an enantiomeric excess of 44%. More recently, this field has received renewed attention by other groups and has been denoted as hybrid enzymes or artificial (metallo)enzymes. This field of research combines redox-active compounds (metals and / or organocatalysts) with enzymes as the chiral scaffold.75,76 The principle idea is that one could use the chiral cavity of an enzyme in combination with a non- chiral chemocatalyst in order to create a new chiral catalyst. We will focus here on artificial metal cofactors only and within this approach different strategies can be followed. It is possible to introduce artificial metal cofactors (A) by the incorporation of non-natural amino acids, (B) by supramolecular anchoring, (C) by dative anchoring and (D) by covalent anchoring. Each of these four approaches will be concisely explained and illustrated by a literature example. In this way a brief overview is given of the possibilities and limitations in the use of hybrid enzymes.

(A) The introduction of non-natural amino acids This technique can be performed in several ways such as solid phase peptide synthesis or by a variation in genetic encoding. One recent example from literature shows the possibility of encoding (2,2’-bipyridin-5-yl)alanine into a catabolite activator protein.77 When this moiety is combined with transition metal ions which are capable of oxidations, this non-natural amino acid enriched protein is capable of site-specific cleavage of DNA (Ala-bpy, figure 18). It must be noted that this approach of hybrid enzymes is still in its infancy.

16 Introduction

CAP-Lys26Ala-bpy, Cu2+ or Fe2+, 3-MPA, cAMP, MOPS buffer, pH 7.3

N N Ala-bpy

H2N CO2H

Figure 18 The use of Ala-bpy encoded CAP enzyme in oxidative cleavage of DNA

(B) Supramolecular anchoring Many enzymes can have high affinities for compounds which interact with the protein backbone. One could exploit this by attaching a metal complex to the corresponding compound which, upon binding, would introduce the redox active compound in the chiral cavity of the host protein. The classic example of this approach is the Biotin-(strept)avidin combination which was pioneered by Whitesides (figure 19).74 Due to the formidable association constant of biotin to the protein host (Kd ≈ 10-15 M) one can assume that there will be no free metal complex in solution. In recent years, the group of Ward and co-workers (and some other groups) have been exploring the use of this strategy and in principle three enantioselective reactions have been optimized: the hydrogenation of N-protected dehydroamino acids,78 the transfer hydrogenation of ketones79 and the allylic alkylation of 1,3-diphenyl-allylacetate.80 Even though much research has been performed towards this specific technique, this seems to be the only successful example for the supramolecular anchoring approach in artificial metalloenzymes.

Streptavidin O

HN NH Biotin Y H H M Spacer Y S Biotin O

Figure 19 Schematic principle of the Biotin-(stept)avidin technique in hybrid enzymes

17 Chapter 1

(C) Dative anchoring Enzymes which bear a transition metal (compound) as the cofactor for their catalytic activity use coordinating ligands for binding the cofactor. Under the right circumstances, this cofactor can be extracted from the protein backbone and replaced by a different redox-active compound. One example of this approach is replacing the zinc from carbonic anhydrase with manganese or rhodium, yielding either a new enzyme capable of epoxidation or hydrogenation/isomerization respectively (figure 20).81,82 This method has also been used in other systems.83,84

COO-

His96 N His96 2+ His94 Zn 1. His94 M COO- His119 His119 2. M Native carbonic anhydrase O 0.6 mol% Mn-CA ee 67% HCO -, H O Cl 3 2 2 Cl TON 7

1 mol% Rh-CA9xHis->Ala

H2, 0.1 M MES, rt

Figure 20 Metal exchange in carbonic anhydrase and subsequent activity of these enzymes

(D) Covalent anchoring The final way of adding organometallic compounds to proteins is by the method of covalent anchoring. This uses the nucleophilicity of certain amino acid side chains for attachment. One of the best examples is the work of Van Koten and co- workers who directed their effort to attaching organometallic compounds to the active site of an enzyme, in this case a cutinase. They performed covalent anchoring of metallopincer ligands by means of p-nitrophenyl phosphonate esters which were tethered covalently to the pincers (figure 21). 85,86 18 Introduction

Figure 21 Phosphonate ECE-metallopincer complexes

Interestingly, even though such hybrid systems should be promising catalysts for a variety of chemical transformations, reports on their catalytic use are missing so far. The crystal structures for two metallopincer-modified cutinases, one bearing compound A and one cutinase bearing compound D, were resolved and showed that the metallopincer sticks out of the cutinase molecule and is thereby mostly solvent-exposed.87 Under certain conditions, even Pt-Cl-Pt mediated dimerization of the enzyme in the crystal structure was observed (figure 22) suggesting very little steric constraints around the Pt-central atom. As a consequence, any chiral induction by the cutinase on the Pt-reaction centre is rather unlikely, thereby severely impairing the attractiveness of this concept for enantioselective hybrid catalysts.

It is evident from the above examples that hybrid catalysts indeed offer new possibilities for enantioselective reactions for chemocatalysts, but that it is by no means an easy route for success.

19 Chapter 1

Figure 22 Complex dimerization of cutinase metallopincer hybrids

Scope of the thesis Oxidations are still of vital importance in chemistry, both on a small and on a large (process) scale. As has been introduced in this chapter, (organo)selenium compounds can be important in this oxidative toolbox. Even though a lot of research has already been conducted, there are a number of questions which will be addressed in this thesis. For the catalytic oxidation of alcohols with diselenides as pre-catalysts, one of the most interesting questions is the actual catalytic cycle of this reaction, which is still uncertain (see figure 9). This question is addressed in chapter 2 and the outcome has been the basis for the development of a truly catalytic process for the selenium mediated oxidation of alcohols. In chapter 3, the mechanism is further substantiated by spectroscopic and calorimetric studies. Furthermore, in chapter 4 the investigation of the influence of substituents at the aromatic ring on the activity of the catalyst is reported. Even though many terminal oxidants have been used, our focus will lie on the use of tert-butyl hydroperoxide as the terminal oxidant, which is cheap and commercially available. The second subject in the thesis will be the use of organoselenium catalysts for epoxidations in new types of solvents which are based on glycerol as the main building block, which is discussed in chapter 5. Chapters 6 and 7 cover the final subject, which is the attachment of organoselenium catalysts to enzymes, in order to create new hybrid enzymes, 20 Introduction which in principle should be able to catalyze oxidations with hydrogen peroxide as the terminal oxidant.

References (1) Jorpes, J. E. Jac. Berzelius, His life and work; Almqvist & Wiksell: Stockholm, 1966. (2) Yang, G. Q.; Wang, S. Z.; Zhou, R. H.; Sun, S. Z. Am. J. Clin. Nutr. 1983, 37, 872-881. (3) Goldhaber, S. B. Regul. Toxicol. Pharm. 2003, 38, 232-242. (4) Whanger, P. D. Br. J. Nutr. 2004, 91, 11-28. (5) Zeng, H. W. Molecules 2009, 14, 1263-1278. (6) Stadtman, T. C. Annu. Rev. Biochem. 1996, 65, 83-100. (7) Glass, R. S.; Singh, W. P.; Jung, W.; Veres, Z.; Scholz, T. D.; Stadtman, T. C. Biochemistry 1993, 32, 12555-12559. (8) Jacob, C.; Giles, G. L.; Giles, N. M.; Sies, H. Angew. Chem.-Int. Edit. 2003, 42, 4742-4758. (9) Paulmier, C. Selenium reagents and intermediates in organic synthesis; Pergamon Press: Oxford, 1986. (10) Iwaoka, M.; Tomoda, S. In Organoselenium Chemistry 2000; Vol. 208, p 55-80. (11) Tiecco, M. In Organoselenium Chemistry 2000; Vol. 208, p 7-54. (12) Freudendahl, D. M.; Santoro, S.; Shahzad, S. A.; Santi, C.; Wirth, T. Angew. Chem.- Int. Edit. 2009, 48, 8409-8411. (13) Back, T. G. Organoselenium Chemistry; A Practical Approach; Oxford University Press: Oxford, 1999. (14) Mlochowski, J.; Brzaszcz, M.; Giurg, M.; Palus, J.; Wojtowicz, H. Eur. J. Org. Chem. 2003, 4329-4339. (15) Back, T. G. In Organoselenium Chemistry: A Practical Approach; Harwood, L. M., Moody, C. J., Eds.; Oxford University Press: New York, 1999, p 7-34. (16) Paulmier, C. In Selenium Reagents and Intermediates in Organic Synthesis; Baldwin, J. E., Ed.; Pergamon Press: Oxford, 1986; Vol. 4, p 125-161. (17) Jones, D. N.; Mundy, D.; Whitehouse, R.D. J. Chem. Soc. D 1970, 86-87 (18) Grieco, P. A.; Yokoyama, Y.; Gilman, S.; Nishizawa, M. J. Org. Chem. 1977, 42, 2034-2036. (19) Reich, H. J.; Chow, F.; Peake, S. L. Synthesis 1978, 299-301. (20) Reich, H. J.; Wollowitz, S.; Trend, J. E.; Chow, F.; Wendelborn, D. F. J. Org. Chem. 1978, 43, 1697-1705. (21) Hori, T.; Sharpless, K. B. J. Org. Chem. 1978, 43, 1689-1697. (22) Taylor, R. T.; Flood, L. A. J. Org. Chem. 1983, 48, 5160-5164. (23) Betzemeier, B.; Lhermitte, F.; Knochel, P. Synlett 1999, 489-491. (24) ten Brink, G. J.; Fernandes, B. C. M.; van Vliet, M. C. A.; Arends, I. W. C. E.; Sheldon, R. A. J. Chem. Soc.-Perkin Trans. 1 2001, 224-228. (25) de Visser, S. P.; Kaneti, J.; Neumann, R.; Shaik, S. J. Org. Chem. 2003, 68, 2903-2912. (26) Santoro, S.; Santi, C.; Sabatini, M.; Testaferri, L.; Tiecco, M. Adv. Synth. Catal. 2008, 350, 2881-2884. (27) Grieco, P. A.; Yokoyama, Y.; Gilman, S.; Ohfune, Y. J. Chem. Soc.-Chem. Commun. 1977, 870-871. (28) Choi, J. K.; Chang, Y. K.; Hong, S. Y. Tetrahedron Lett. 1988, 29, 1967-1970. (29) Syper, L. Synthesis 1989, 167-172.

21 Chapter 1

(30) ten Brink, G. J.; Vis, J. M.; Arends, I. W. C. E.; Sheldon, R. A. J. Org. Chem. 2001, 66, 2429-2433. (31) Ichikawa, H.; Usami, Y.; Arimoto, M. Tetrahedron Lett. 2005, 46, 8665-8668. (32) ten Brink, G. J.; Vis, J. M.; Arends, I. W. C. E.; Sheldon, R. A. Tetrahedron 2002, 58, 3977-3983. (33) Kamerbeek, N. M.; Janssen, D. B.; van Berkel, W. J. H.; Fraaije, M. W. Adv. Synth. Catal. 2003, 345, 667-678. (34) Clouthier, C. M.; Kayser, M. M.; Reetz, M. T. J. Org. Chem. 2006, 71, 8431-8437. (35) Reetz, M. T.; Brunner, B.; Schneider, T.; Schulz, F.; Clouthier, C. M.; Kayser, M. M. Angew. Chem.-Int. Edit. 2004, 43, 4075-4078. (36) ten Brink, G. J.; Arends, I. W. C. E.; Sheldon, R. A. Chem. Rev. 2004, 104, 4105-4123. (37) Colladon, M.; Scarso, A.; Sgarbossa, P.; Michelin, R. A.; Strukul, G. J. Am. Chem. Soc. 2006, 128, 14006-14007. (38) Miyake, Y.; Nishibayashi, Y.; Uemura, S. Bull. Chem. Soc. Jpn. 2002, 75, 2233-2237. (39) Barton, D. H. R.; Brewster, A. G.; Hui, A. H. F.; Lester, D. J.; Ley, S. V.; Back, T. G. J. Chem. Soc.-Chem. Commun. 1978, 952-954. (40) Barton, D. H. R.; Morzycki, J. W.; Motherwell, W. B.; Ley, S. V. J. Chem. Soc.-Chem. Commun. 1981, 1044-1045. (41) Shimizu, M.; Kuwajima, I. Tetrahedron Lett. 1979, 2801-2804. (42) Kuwajima, I.; Shimizu, M.; Urabe, H. J. Org. Chem. 1982, 47, 837-842. (43) Shimizu, M.; Takeda, R.; Kuwajima, I. Tetrahedron Lett. 1979, 419-422. (44) Gancarz, R. A.; Kice, J. L. Tetrahedron Lett. 1981, 22, 1661-1662. (45) Onami, T.; Ikeda, M.; Woodard, S. S. Bull. Chem. Soc. Jpn. 1996, 69, 3601-3605. (46) Ehara, H.; Noguchi, M.; Sayama, S.; Onami, T. J. Chem. Soc.-Perkin Trans. 1 2000, 9, 1429-1431. (47) Paine, T. K.; Weyhermuller, T.; Wieghardt, K.; Chaudhuri, P. Dalton Trans. 2004, 2092-2101. (48) Paulmier, C. In Organic Chemistry Series; Baldwin, J. E., Ed.; Pergamon Press: Oxford, 1986, p 353-386. (49) Umbreit, M. A.; Sharpless, K. B. J. Am. Chem. Soc. 1977, 99, 5526-5528. (50) Barton, D. H. R.; Crich, D. Tetrahedron 1985, 41, 4359-4364. (51) Barton, D. H. R.; Wang, T. L. Tetrahedron Lett. 1994, 35, 5149-5152. (52) Barton, D. H. R.; Hui, R.; Lester, D. J.; Ley, S. V. Tetrahedron Lett. 1979, 3331-3334. (53) Crich, D.; Barba, G. R. Org. Lett. 2000, 2, 989-991. (54) Crich, D.; Zou, Y. K. Org. Lett. 2004, 6, 775-777. (55) Crich, D.; Zou, Y. K. J. Org. Chem. 2005, 70, 3309-3311. (56) Nishibayashi, Y.; Uemura, S. Organoselenium Chemistry 2000, 208, 235-255. (57) Drake, M. D.; Bateman, M. A.; Detty, M. R. Organometallics 2003, 22, 4158-4162. (58) Goodman, M. A.; Detty, M. R. Organometallics 2004, 23, 3016-3020. (59) Francavilla, C.; Drake, M. D.; Bright, F. V.; Detty, M. R. J. Am. Chem. Soc. 2001, 123, 57-67. (60) Mellegaard-Waetzig, S. R.; Wang, C.; Tunge, J. A. Tetrahedron 2006, 62, 7191-7198. (61) Johansson, L.; Gafvelin, G.; Arner, E. S. J. Biochim. Biophys. Acta-Gen. Subj. 2005, 1726, 1-13. (62) Mugesh, G.; Singh, H. B. Chem. Soc. Rev. 2000, 29, 347-357. (63) Wu, Z. P.; Hilvert, D. J. Am. Chem. Soc. 1989, 111, 4513-4514. (64) Wu, Z. P.; Hilvert, D. J. Am. Chem. Soc. 1990, 112, 5647-5648. (65) Bell, I. M.; Hilvert, D. Biochemistry 1993, 32, 13969-13973. (66) Bell, I. M.; Fisher, M. L.; Wu, Z. P.; Hilvert, D. Biochemistry 1993, 32, 3754-3762. (67) Syed, R.; Wu, Z. P.; Hogle, J. M.; Hilvert, D. Biochemistry 1993, 32, 6157-6164.

22 Introduction

(68) Peterson, E. B.; Hilvert, D. Biochemistry 1995, 34, 6616-6620. (69) Haring, D.; Herderich, M.; Schuler, E.; Withopf, B.; Schreier, P. Tetrahedron: Asymmetry 1997, 8, 853-856. (70) Haring, D.; Schreier, P. Biotechnol. Bioeng. 1998, 59, 786-791. (71) Haring, D.; Hubert, B.; Schuler, E.; Schreier, P. Arch. Biochem. Biophys. 1998, 354, 263-269. (72) Haring, D.; Schreier, P. Angew. Chem.-Int. Edit. 1998, 37, 2471-2473. (73) Haring, D.; Schuler, E.; Adam, W.; Saha-Moller, C. R.; Schreier, P. J. Org. Chem. 1999, 64, 832-835. (74) Wilson, M. E.; Whitesides, G. M. J. Am. Chem. Soc. 1978, 100, 306-307. (75) Letondor, C.; Ward, T. R. ChemBioChem 2006, 7, 1845-1852. (76) Heinisch, T.; Ward, T. R. Curr. Opin. Chem. Biol. 2010, 14, 184-199. (77) Lee, H. S.; Schultz, P. G. J. Am. Chem. Soc. 2008, 130, 13194-13195. (78) Klein, G.; Humbert, N.; Gradinaru, J.; Ivanova, A.; Gilardoni, F.; Rusbandi, U. E.; Ward, T. R. Angew. Chem.-Int. Edit. 2005, 44, 7764-7767. (79) Creus, M.; Pordea, A.; Rossel, T.; Sardo, A.; Letondor, C.; Ivanova, A.; LeTrong, I.; Stenkamp, R. E.; Ward, T. R. Angew. Chem.-Int. Edit. 2008, 47, 1400-1404. (80) Pierron, J.; Malan, C.; Creus, M.; Gradinaru, J.; Hafner, I.; Ivanova, A.; Sardo, A.; Ward, T. R. Angew. Chem.-Int. Edit. 2008, 47, 701-705. (81) Okrasa, K.; Kazlauskas, R. J. Chem.-Eur. J. 2006, 12, 1587-1596. (82) Jing, Q.; Okrasa, K.; Kazlauskas, R. J. Chem.-Eur. J. 2009, 15, 1370-1376. (83) Ohashi, M.; Koshiyama, T.; Ueno, T.; Yanase, M.; Fujii, H.; Watanabe, Y. Angew. Chem.-Int. Edit. 2003, 42, 1005-1008. (84) Correia, I.; Aksu, S.; Adao, P.; Pessoa, J. C.; Sheldon, R. A.; Arends, I. J. Inorg. Biochem. 2008, 102, 318-329. (85) Kruithof, C. A.; Casado, M. A.; Guillena, G.; Egmond, M. R.; van der Kerk-van Hoof, A.; Heck, A. J. R.; Klein Gebbink, R. J. M.; van Koten, G. Chem.-Eur. J. 2005, 11, 6869-6877. (86) Kruithof, C. A.; Dijkstra, H. P.; Lutz, M.; Spek, A. L.; Egmond, M. R.; Gebbink, R.; van Koten, G. Eur. J. Inorg. Chem. 2008, 4425-4432. (87) Rutten, L.; Wieczorek, B.; Mannie, J.; Kruithof, C. A.; Dijkstra, H. P.; Egmond, M. R.; Lutz, M.; Klein Gebbink, R. J. M.; Gros, P.; van Koten, G. Chem.-Eur. J. 2009, 15, 4270-4280.

Chapter 2: Diphenyl diselenide-catalyzed selective oxidation of alcohols with tert-butyl hydroperoxide

John C. van der Toorn, Gerjan Kemperman, Roger A. Sheldon and Isabel W.C.E. Arends

This chapter has been published in: Journal of Organic Chemistry 2009, Vol. 74, No. 8, 3085-3089

Abstract Diselenides are known precatalysts for catalytic oxidations. Even though the use of these widely available compounds as catalysts for the oxidation of alcohols has previously been investigated, many mechanistic questions are still to be answered. By using a range of analytical techniques, we found evidence for the involvement of benzeneseleninic anhydride in the catalytic mechanism. The activation of the diselenide was carefully studied, which had implications on the oxidation reaction in terms of selectivity, conversion and reaction kinetics. Based on these findings, an improved protocol for the selective oxidation of activated alcohols was devised resulting in significantly decreased catalyst loadings (<1% Ph2Se2) in the oxidation of activated alcohols. Chapter 2

Introduction Alcohol oxidations are highly relevant reactions in synthetic organic chemistry.1 The development of catalytic systems for these types of transformations, affording sustainable and atom efficient methodologies, is still of great interest.2 Excellent transition metal catalysts, based on e.g. Cu, Pd, Ru and Pt have been developed which allow the use of molecular oxygen as the terminal oxidant.3 However these metal catalysts are often unsuitable for highly functionalized molecules, which are used as intermediates in the pharmaceutical industry. In these cases, organocatalytic methodologies can offer a distinct advantage, due to their high selectivities. A good example of an industrially applied organocatalyst capable of alcohol oxidations is the 2,2,6,6-tetramethylpiperidinyloxy radical (TEMPO), using bleach (sodium hypochlorite) or a hypervalent iodine (III) species as the most popular terminal oxidants.4,5 Major drawbacks of these systems are the formation of halogenated side-products and the use of halogen-rich solvents. Even though iodine (III) reagents can be used chloride- and bromide-free, their use is accompanied by the formation of two equivalents of acid. In search for other main group elements capable of performing these types of redox reactions with benign oxidants, our interest focused on the use of organoselenium compounds as catalysts.6-8 The classical organoselenium reagent for alcohol oxidations is benzeneseleninic anhydride (BSA) which was introduced by Barton and co-workers in the late 70’s.9 They proposed a mechanism involving the selenenate ester (compound 2) as the intermediate species (figure 1).

O O O H O H Se Se Se O O R R 1 OH 2 SeOH O R Se OH 3

Figure 1 The mechanism of alcohol oxidation by BSA

26 Ph2Se2 as a catalyst in the selective oxidation of alcohols

Interestingly, if one performs the oxidation of benzyl alcohol with BSA, a yellow color immediately appears upon addition of the alcohol which is attributed to the formation of Ph2Se2. The formation of Ph2Se2 can be explained by an equilibrium between three equivalents of benzeneselenate (4) which gives H2O, Ph2Se2 and one equivalent of benzeneseleninic acid (figure 2).10

O SeOH Se 3 OH Ph2Se2 H2O 4 3 5

Figure 2 The rearrangement of benzeneselenate to benzeneseleninic acid and Ph2Se2

We reasoned that diselenides could be introduced as precursors for the active oxidant as several oxidants are known to oxidize aromatic diselenides to the corresponding anhydrides. If this could be done in situ, this would yield a new catalytic process for the oxidation of alcohols. So far, the most optimized system involved the use of dimesitylene diselenide (0.5 equiv) and aqueous tert-butyl hydroperoxide (TBHP) (1.5 equiv). This combination was applied to oxidize a variety of primary (benzylic and aliphatic) alcohols to the corresponding aldehydes in high selectivity.11,12 Two possible mechanisms were proposed but neither was substantiated with experimental evidence. The first mechanism involved a selenenate ester, the second mechanism was based upon a seleninate ester being the active intermediate responsible for the dehydrogenation. The catalytic use of simple diphenyl diselenide was also mentioned, but the authors stated that in this case only benzylic (also secondary benzylic) and primary allylic alcohols could be fully oxidized and that using less than 10% mol of this (standard) diselenide did not result in complete conversion of the alcohol.

Herein we report our investigation on the mechanism of Ph2Se2 catalyzed alcohol oxidations which led to the development of a new procedure for alcohol oxidations employing as little as 1 mol% of diphenyl diselenide as the catalyst and TBHP as the terminal oxidant.

27 Chapter 2

Results and discussion Diphenyl diselenide mediated oxidation of benzyl alcohol; catalytic and stoichiometric reactions

Initially we repeated the Ph2Se2 catalyzed oxidation of benzyl alcohol as described by Kuwajima et al.12 This involved reactions in chlorobenzene at 80 °C, employing

10% mol Ph2Se2 and 1.5 equivalents of aqueous TBHP. In agreement with their results we found that when less than 10% mol Ph2Se2 was used, the result was an incomplete conversion of benzyl alcohol and the formation of a significant amount of benzoic acid. We also found that the use of TBHP in decane (rather than aqueous TBHP) led to an increased yield and selectivity of the reaction, depending on the solvent used. In the older reports, chlorobenzene or benzene were the preferred solvents. As we were not interested in these solvents we tested various alternatives (figure 3).

*))"#

&)"# @ABC#:3#D-8=3-# =E#@ABC# '!"# (&"# ')"# ()"# ()"# !!"#!("# !$"# !$"# !*"# !)"# !*"#!*"# !)"# +'"#

!"#$%&'#()*$%#+,%#&& %&"# +)"# %'"#

%)"#

,-./# 012-3-# 45678# ..2%# ;->5=3-# 5<2?-3-# 59:8;2<9<-5;=3-#

Figure 3 Yields of benzaldehyde (aldehyde) and benzoic acid (acid) in Ph2Se2 (5%) catalyzed oxidation of benzyl alcohol with 1.1 eq of TBHP at 80 °C, 8 h.

The best results were obtained by using an apolar solvent such as heptane or toluene and with the temperature of the reaction around 80 °C. At higher temperatures, the reactions became sluggish as a lot of by-products were formed and the mass balance of the reaction seriously deteriorated. This could be attributed to the fact that TBHP is unstable at higher temperatures. DSC measurements indeed showed the instability of TBHP at higher temperatures

28 Ph2Se2 as a catalyst in the selective oxidation of alcohols

(figure 4). From approximately 70 °C there is a slight incline in temperature (heat evolved from the crucible) which becomes even more extreme from 90 °C onwards.13 It is worth noting that the gold-plated crucibles have a different behavior in activation of the hydroperoxide. In order to avoid decomposition of TBHP in situ, the temperature of the reaction should thus not exceed 80 °C.

Figure 4 DSC measurement of THBP in decane. Gold plated or stainless steel crucibles were filled with TBHP in decane (5.5 M) and heated. The heat of the reaction was determined by subtracting the temperature of heating from the temperature of the crucible (y-axis, ΔK)

When using 5% mol of catalyst and aqueous TBHP, an induction period of ca. one hour was observed suggesting that a pre-activation of the Ph2Se2 catalyst was required. Therefore we performed a series of experiments under stoichiometric conditions to investigate the origin of this induction period. We found that premixing Ph2Se2 with one equivalent of TBHP in toluene for two hours followed by the addition of one equivalent of benzyl alcohol resulted in complete conversion to benzaldehyde within 40 minutes (figure 5). When two or three equivalents of TBHP were added, the rate of benzyl alcohol oxidation increased only slightly. In theory, the use of three equivalents of TBHP should result in the formation of one equivalent of BSA relative to benzyl alcohol. A further increase of

29 Chapter 2 the amount of TBHP to four or five equivalents resulted in oxidation of toluene to a mixture of benzyl alcohol and benzaldehyde which was confirmed in blank experiments in the absence of alcohol substrate.

100%

90%

80% 1 eq TBHP 70% 2 eq TBHP 3 eq TBHP 60% 4 eq TBHP Yield benzaldehyde benzaldehyde Yield 5 eq TBHP 50% 0 10 20 30 40 50 60 t (min)

Figure 5 Oxidation of benzyl alcohol with Ph2Se2 pre-activated with different amounts of TBHP in toluene

The oxidation of toluene has been reported before using BSA as the oxidant, but these reactions are thought to occur at higher temperatures and with longer reaction times.14 We assume that the excess of TBHP could undergo (possibly selenium catalyzed) homolytic decomposition to afford tert-butoxy and tert- butylperoxy radicals which can initiate the autoxidation of toluene. It is important to keep the TBHP concentration low to avoid the competing homolytic decomposition of TBHP which would initiate this autoxidation.15

To further understand the pre-activation of Ph2Se2, it was activated with two equivalents of TBHP at different time-lengths, followed by the addition of substrate (figure 6). From this figure it is easy to deduce that there is a very distinct amount of time required for the activation of the diselenide. Obviously, a pre-activation of less than 20 minutes is deleterious for the oxidation reaction.

30 Ph2Se2 as a catalyst in the selective oxidation of alcohols

100%

80%

30 min 60% 25 min 20 min 40% 15 min 10 min 20% Yield benzaldehyde benzaldehyde Yield 5 min

0% 0 10 20 30 40 50 60 t (min)

Figure 6 Influence of activation time-length on the oxidation of benzyl alcohol with Ph2Se2 and two equivalents of TBHP in toluene

Reaction calorimetry The activation period was further investigated by reaction calorimetry. First the activation of Ph2Se2 (0.1 M) in toluene with one equivalent of TBHP was studied (figure 7).

&# addition 1 equivalent %# TBHP

"# !"#$#!%#&'(#

$# time from addition to activation maximum: 26 min

!"# '%$$# '($$# %%$$# %($$# )%$$# )*+#&,(#

Figure 7 Reaction calorimetry profile of the activation of 0.1 M Ph2Se2 by one equivalent TBHP in toluene

Upon the addition of the cold TBHP there was an expected drop in temperature but after 20-25 minutes a sudden increase in reaction temperature was observed. This showed an exothermic reaction with an enthalpy of approximately 150 kJ/mol 31 Chapter 2 and careful analysis of the mixture showed that it consisted for one-third of BSA. For the exotherm of this reaction a rough estimation of 196 kJ/mol could be made, based on a calculated heat of formation of BSA and known gas phase values for the other compounds (see the experimental section). As there was a huge acceleration in the last part of the reaction, the question arose whether or not this could be attributed to the formed BSA. Therefore, a mixture of Ph2Se2 / BSA (95/5) was activated with TBHP. This revealed that BSA is indeed influencing the oxidation of Ph2Se2 to BSA as in this case the activation period was virtually absent and the enthalpy had the same value (150 kJ/mol). Presumably, mixing BSA with TBHP gives the intermediate tert-butyl perseleninic ester which is possibly a stronger oxidant than either TBHP or BSA.16 However, there was a significant difference in selectivity. BSA alone gave no overoxidation of benzaldehyde to the benzoic acid. In contrast, the BSA-TBHP mixture showed substantial conversion to the acid (5% in 20 minutes). Yet, in the presence of Ph2Se2, the aldehyde was not overoxidized, but the diselenide was converted to the anhydride, thereby showing the selectivity of the oxygen transfer. In order to investigate whether the in situ formed BSA could be used for a shorter activation period in a second round of diselenide activation, an experiment was performed in which Ph2Se2 was activated with one equivalent of TBHP and was subsequently reacted with two-third equivalent of benzyl alcohol. In this case some of the BSA or other oxidized selenium intermediates should remain in the reaction mixture. Two-thirds of an equivalent of TBHP was subsequently added and indeed the pre-activation of Ph2Se2 was instantaneous as indicated by the exotherm (figure 8). In a following experiment the activation of Ph2Se2 by TBHP was investigated in the presence of one equivalent of substrate. As can be deduced from figure 9, in the presence of substrate the activation of Ph2Se2 is significantly slower. This can be explained by taking into account that in the presence of benzyl alcohol BSA is immediately consumed by oxidation of benzyl alcohol to benzaldehyde. Therefore, the perester formation is inhibited, thus reducing or precluding its involvement in the oxidation of Ph2Se2 by TBHP. The latter result also supports the hypothesis that the perester is involved in the formation of BSA from Ph2Se2 and TBHP.

32 Ph2Se2 as a catalyst in the selective oxidation of alcohols

'+"

'*"

')"

''" addition TBHP '("

&" !"#!$%&'(% %"

$" time from addition to activation #" maximum: 5 min

!" +%((" !*((" !%((" )*+%&,(%

Figure 8 Reaction calorimetry of the activation of Ph2Se2 in the presence of BSA

*# addition )# 2/3 eq addition addition (# benzyl one eq 2/3 eq alcohol '# TBHP TBHP

!"#$#!%#&'(# calibration %#

$# activation time: activation time: 26 min 2,5 min !"# &$$$# '($$# )$$$# *($$# +$$$# "$($$# )*+#&,(#

Figure 9 Reaction calorimetry of activation of 0.1 M Ph2Se2 in the presence of one equivalent of benzyl alcohol

In summary, a number of the above mentioned findings are of importance for the development of an effective Ph2Se2 mediated oxidation of benzyl alcohol with TBHP as the terminal oxidant. First, exposure of more than three equivalents of

TBHP relative to the amount of Ph2Se2 causes overoxidation of benzaldehyde to benzoic acid and the oxidation of toluene to benzyl alcohol. We attribute this to the formation of the perseleninic ester which can oxidize benzaldehyde easily to benzoic acid. In addition, a relative high concentration of TBHP could undergo homolytic cleavage, thereby promoting the latter two reactions via a radical- 33 Chapter 2 autoxidation process. Therefore the concentration of TBHP should remain low throughout the process. Second, the pre-activation of Ph2Se2 by reaction with TBHP to afford BSA is autocatalytic, that is the BSA catalyzes its own formation which we attribute to the formation of the perester and thus regeneration of the active species is most effective when a small amount of BSA remains present in the reaction mixture. These findings have led us to the development of a protocol (figure 10) in which first Ph2Se2 is pre-treated with TBHP in order to generate BSA. Next, an amount of an activated alcohol (such as benzyl alcohol or cinnamyl alcohol) is added that is insufficient to consume all of the BSA. Subsequently, after the oxidation of the substrate, another portion of TBHP is added in order to regenerate all of the BSA.

1 eq 1 eq alcohol TBHP

Ph2Se2 Activate 2 eq TBHP 30x stir 30 min

remove H2O, redissolve in toluene

Figure 10 Experimental protocol for the oxidation of alcohols with Ph2Se2

Initial experiments according to this protocol showed that substrate/catalyst ratios above 10 could be achieved. However, the water formed in the oxidation during these cycles had a deleterious effect on the reaction, which can be attributed to the hydrolysis of BSA, thus competing with the alcohol substrate. This was overcome by adding a drying agent to the reaction mixture (anhydrous MgSO4 or MS 4Å) enabling substrate to catalyst ratios of more than 100 (table 1).

34 Ph2Se2 as a catalyst in the selective oxidation of alcohols

Table 1 Results of optimized reaction conditions; a TTN = mmol product * mmol catalyst-1 Substrate Product TTN a Selectivity Benzyl alcohol Benzaldehyde 115 99,1% Cinnamyl alcohol Cinnamaldehyde 180 > 99,9 %

Based on the above observations, we propose the following tentative mechanism for the Ph2Se2 catalyzed oxidation of alcohols with TBHP (figure 11).

R OH O O Se Se O tBuOH + H2O O Se OH

O H TBHP H Se O R SeOH

Figure 11 Catalytic cycle in the oxidation of alcohols with TBHP as the terminal oxidant and

Ph2Se2 as catalyst

Conclusions

The Ph2Se2 catalyzed oxidation of activated alcohols with TBHP as the terminal oxidant is a reaction that has been known for a long time. The main problem of the methodology was the high loading of catalyst required. We have shown that benzeneseleninic anhydride (BSA) is the active oxidant and that it catalyzes its own formation most likely via the in situ formation of a perester. If the concentration of TBHP is too high, homolytic decomposition of this perester results in the autoxidation of solvents containing reactive C-H bonds, such as toluene, and the 35 Chapter 2 autoxidation of benzaldehyde to benzoic acid. Based upon these findings we developed a new catalytic protocol, which included the in situ removal of water. A drying agent could be added but alternatively, the reaction can be performed under Dean-Stark conditions, which is more convenient on a larger scale. Two activated alcohols, benzyl alcohol and cinnamyl alcohol, were oxidized by TBHP to the corresponding aldehydes employing <1% of Ph2Se2 as the catalyst. Compared to other methods with main group elements, our method is advantageous in that all compounds are cheap, readily available and the method is completely halogen- free. However, the use of Ph2Se2 as a catalyst precursor has the disadvantage that only activated alcohols are oxidized smoothly, which was already found earlier.12 Improvements to this system can then only be expected if the diselenide catalyst is modified, in order to improve the substrate scope and possibly the rate of oxidation.

36 Ph2Se2 as a catalyst in the selective oxidation of alcohols

Experimental section Materials

Ph2Se2 (99%), TBHP in decane (5.5M), benzyl alcohol (99+%), 1,2-dimethoxybenzene (99+%), molecular sieves (4 Å), MgSO4 (anhydrous 97%), benzeneseleninic anhydride (98%), toluene (reagent grade, stored over molecular sieves 4 Å), ethyl acetate (p.a.) were all used as received.

Catalytic reactions To a stirred solution of benzyl alcohol (2 mmol, 216 mg), 1,2-dimethoxybenzene (0,5 mmol, 69 mg, internal standard) and Ph2Se2 (0.1 mmol, 31 mg) in 10 mL toluene was added at 80 °C TBHP (2.2 mmol, 400 µL). At several intervals, 50 µL aliquots were withdrawn, quenched with Na2SO3 (100 mg in 1.5 mL EtOAc) and the solids were filtered off. The mixture was then analyzed by GC.

Stoichiometric reactions

To a stirred solution of Ph2Se2 (1 mmol, 312 mg), 1,2-dimethoxybenzene (0.5 mmol, 69 mg, internal standard) in 10 ml toluene at 80 °C was added TBHP (5.5 M solution in decane) and after the appropriate amount of time, benzyl alcohol was added. At several intervals, 50 µL aliquots were withdrawn, quenched with Na2SO3 (100 mg in 1.5 mL EtOAc) and the solids were filtered off. The mixture was then analyzed by GC.

Catalytic protocol for the use of low concentrations of catalyst

To a solution of Ph2Se2 (1 mmol, 312 mg) and 1,2-dimethoxybenzene (5 mmol, 690 mg) in 15 mL toluene at 80 °C was added first 2 mmol TBHP (364 µL) with a pump system (syringe pump equipped with a glass syringe and PFTE 1/16” tubing into the reaction) and the mixture was allowed to stand for 1 hour. After this, 1 mmol of benzyl alcohol (103 µL) was added with a pump system (programmable syringe pump equipped with a glass syringe and PFTE 1/16” tubing into the reaction) which was allowed to react for 30 minutes. After this, 1 mmol TBHP (182 µL) was added with the pump system and the mixture was allowed to stand for 30 minutes after which benzyl alcohol was added and this sequence was allowed to loop 30 times. Then either molecular sieves or anhydrous

Na2SO4 were added to dry the reaction. The solids were then filtered off, the solvent evaporated and the residue redissolved in 15 mL toluene after which the sequence was restarted. This allowed a total of 115 mmol benzyl alcohol to be converted (0,86% Ph2Se2) into benzaldehyde after which benzoic acid formation could no longer be suppressed. Another way of water elimination was performing a Dean-Stark reflux after 30 cycles of the reaction. By using this sequence also cinnamyl alcohol could be oxidized to cinnamaldehyde (180 mmol was converted; 0.55% Ph2Se2) without any by-product formation. In this case, cinnamyl alcohol was dissolved in toluene as a standard solution (2.0 M).

To a 100 mM solution of Ph2Se2 at 80 °C was added the appropriate amount of TBHP in decane after which the temperature difference between the internal sensor (Tr; temperature of the reaction) and the external sensor (Tj; temperature of the jacket) was monitored.

37 Chapter 2

Calculation of heat of reaction of oxidation of diphenyl diselenide with tert-butyl hydroperoxide The heat of reaction for the oxidation of diphenyldiselenide into benzeneseleninic acid anhydride (BSA) with tert-butyl hydroperoxide can be roughly estimated from the respective heats of formation of the different compounds involved, assuming that solvation enthalpies for substrates and products are equal. For this calculation the ΔHf (g) for BSA was calculated with MOPAC© using the semi- empirical AM1 theory as -117 kJ/mol (in a comparison of known and calculated values for diphenyl diselenide, AM1 gave far better results than PM3; the difference between known and calculation amounted to an overestimation of only 42 kJ/mol in this case). Other values were taken from the NIST Chemistry Webbook.17

Thus the ΔHr for the reaction: 1/3 Ph2Se2 (ΔHf = 237 kJ/mol) + 1 C4H9OOH (ΔHf = -235 kJ/mol)

→ 1/3 PhSe(O)OSe(O)Ph (ΔHf,calc = -117 kJ/mol) + C4H9OH (ΔHf = -313 kJ/mol) can be estimated as -196 kJ/mol. Taking the calculated value of Ph2Se2 instead of the true value leads to -211 kJ/mol. Given the uncertainty in calculated and liquid phase values, this is in good agreement with the observed heat of reaction of 150 kJ/mol TBHP.

References

(1) Bäckvall, J.-E. Modern Oxidation methods; Wiley, 2004. (2) Anastas, P. T.; Williamson, T. C. Green Chemistry; Oxford University Press: Oxford, 1998. (3) Sheldon, R. A.; Arends, I.; Dijksman, A. Catal. Today 2000, 57, 157-166. (4) Anelli, P. L.; Biffi, C.; Montanari, F.; Quici, S. J. Org. Chem. 1987, 52, 2559-2562. (5) DeMico, A.; Margarita, R.; Parlanti, L.; Vescovi, A.; Piancatelli, G. J. Org. Chem. 1997, 62, 6974-6977. (6) ten Brink, G. J.; Vis, J. M.; Arends, I. W. C. E.; Sheldon, R. A. Tetrahedron 2002, 58, 3977-3983. (7) ten Brink, G. J.; Vis, J. M.; Arends, I. W. C. E.; Sheldon, R. A. J. Org. Chem. 2001, 66, 2429-2433. (8) ten Brink, G. J.; Fernandes, B. C. M.; van Vliet, M. C. A.; Arends, I. W. C. E.; Sheldon, R. A. J. Chem. Soc.-Perkin Trans. 1 2001, 224-228. (9) Barton, D. H. R.; Brewster, A. G.; Hui, A. H. F.; Lester, D. J.; Ley, S. V.; Back, T. G. J. Chem. Soc.-Chem. Commun. 1978, 952-954. (10) Hori, T.; Sharpless, K. B. J. Org. Chem. 1978, 43, 1689-1697. (11) Shimizu, M.; Kuwajima, I. Tetrahedron Lett. 1979, 2801-2804. (12) Kuwajima, I.; Shimizu, M.; Urabe, H. J. Org. Chem. 1982, 47, 837-842. (13) Hiatt, R. In Organic Peroxides II; 9th ed.; Swern, D., Ed.; Wiley: New York, 2007 (14) Barton, D. H. R.; Hui, R.; Lester, D. J.; Ley, S. V. Tetrahedron Lett. 1979, 3331-3334. (15) Sheldon, R. A.; Kochi, J. K. Metal-Catalyzed Oxidations of Organic Compounds; Academic Press, New York: New York, 1981. (16) Bloodworth, A. J.; Lapham, D. J. J. Chem. Soc.-Perkin Trans. 1 1983, 471-473. (17) Afeefy, H. V.; Liebman, J. F.; Stein, S. E. In NIST Chemistry Webbook.

38 3

Chapter 3: Spectroscopic investigations on the activation of diphenyl diselenide by tert-butyl hydroperoxide

John C. van der Toorn, Gerjan Kemperman, Roger A. Sheldon and Isabel W.C.E. Arends

This chapter has been published in: European Journal of Organic Chemistry 2011, Issue 23, 4345-4352

Abstract We have investigated the diphenyl diselenide catalyzed oxidation of alcohols with tert-butyl hydroperoxide (TBHP) using a variety of spectroscopic techniques. The pre-activation of Ph2Se2 to benzeneseleninic anhydride (BSA), which is the actual dehydrogenation reagent, was the subject of study. A catalytic cycle describing the oxidation of alcohols with TBHP as oxidant and diselenides as pre-catalysts could thus be constructed. In addition, reaction calorimetry was used to study in depth the formation of BSA as a function of solvent. A correlation between the catalytic activity of BSA in specific solvents and the formation of BSA could be demonstrated. Chapter 3

Introduction

Recently we revisited the oxidation of alcohols using Ph2Se2 as the catalyst and TBHP as the terminal oxidant. Results of a mechanistic study indicated the involvement of an autocatalytic step with benzeneseleninic anhydride as the key intermediate. Based on these results we were able to develop a procedure for the catalytic oxidation of activated alcohols, with TBHP and <1 mol% of Ph2Se2 as the catalyst.1 The formation of benzeneseleninic anhydride (BSA) was detected using reaction calorimetry in the absence of alcohol substrate. Its formation exhibited a distinct exothermic profile, which we attributed to some sort of autocatalysis. Driven by the potential of aromatic diselenides as catalysts, there is a need to gain a deeper insight in the mechanism of the reaction.

Our first goal was to further extend our knowledge on how Ph2Se2 is activated by TBHP at elevated temperatures by using a range of spectroscopic techniques. We found reaction calorimetry to be a useful technique to monitor the course of the reaction. It revealed very clearly that there was an activation time needed during the reaction to form the actual dehydrogenation species, now identified as BSA. It also revealed the autocatalytic character of this step, as instantaneous exothermicity occurs. Previously it was observed that there is a large influence of the solvent in the catalytic oxidation. In this study, we will also use reaction calorimetry to study the solvent dependency of this important first step in the catalytic reaction, namely the oxidation of Ph2Se2 to BSA (figure 1).

O O Se Se Ph2Se2 + TBHP O

Figure 1 Oxidation of Ph2Se2 to BSA

Results and discussion UV-Vis spectroscopy We started our tests with UV-Vis spectroscopy which is highly suitable for determining the initial oxidation as oxidized selenium species (such as diphenylselenoxide) do not absorb light above 300 nm.2 Samples were withdrawn from the reaction mixture and were consecutively diluted, thus allowing for

40 Spectroscopic investigations in the activation of Ph2Se2 by TBHP appropriate absorption measurements (figure 2). After mixing the warm sample with cold toluene, the signal was constant for several hours. The absorption maximum around 330 nm almost immediately decreased to one third of its original intensity upon exposure of one equivalent of Ph2Se2 to one equivalent of TBHP. After 10 minutes no change in intensity occurred anymore. We assign the change in the spectrum to the first oxidation of the diselenide. We know that at this stage the BSA is not formed yet as we have observed that the oxidation of benzyl alcohol did not start when it was added after 10 minutes.2 In contrast, if we added the benzyl alcohol after 40 minutes, an equimolar amount of the alcohol was selectively oxidized to benzaldehyde (95% yield).

(#'"

+,,-./0" (#&" 560789" +9:/;/9" (#%"

(#$"

!#'"

!"#$%"&'()*+!,* !#&"

!#%" +,,-./0" +,,-./0" !#$" 1234" 560789" +9:/;/9" !" )!!" )$!" )%!" )&!" )'!" %!!" %$!" %%!" %&!" %'!" *!!" -*+'.,*

Figure 2 UV-Vis spectral changes of Ph2Se2 upon oxidation with TBHP and subsequent addition of benzyl alcohol (after 40 min)

In-situ infrared and Raman To further search for the active species, we turned to in situ Raman and IR spectroscopy. The IR data are relevant in view of the Se=O stretch vibration, which has a complicated and intensive absorption area around 850 cm-1.3 The IR data correlated with the UV data, as there was again a change in the spectrum after 5 minutes. The time plots of two distinct selenium-oxygen vibration frequencies (860 and 840 cm-1) are shown in figures 3 and 4.

41 Chapter 3

-1 Figure 3 Time profile of IR at 860 cm during the activation of Ph2Se2 by TBHP. At t = 39 min, one equivalent of benzyl alcohol was added.

-1 Figure 4 Time profile of IR at 840 cm during the activation of Ph2Se2 by TBHP. At t = 39 min, one equivalent of benzyl alcohol was added.

From the combination of the UV and IR data we hypothesize that after 5 minutes the first Se=O bond is formed (figure 3). When this is almost complete, the second Se=O stretch appears in the spectrum (figure 4). In Raman spectroscopy, the Se-Se stretch vibration has a distinct frequency, possibly providing us with information on the time of breakage of the diselenide bond. Unfortunately, the spectra could only be recorded in heptane, which had a main drawback; after 23 minutes the formed BSA started to precipitate from the solution. However, from our previous experiments we expected a change of composition between 10 to 15 minutes after addition of the TBHP and we had

42 Spectroscopic investigations in the activation of Ph2Se2 by TBHP found that the activation in heptane and toluene gave almost identical data in reaction calorimetry (vide infra), so we felt that the change of solvent would not greatly influence our spectroscopic studies. As expected, after 15 minutes a large drop in the Se-Se stretch vibration was observed (figure 5). We presume that this event marks the formation of BSA, which correlates with the calorimetric profile that shows strong exothermicity after 15 minutes of the reaction.

Figure 5 Normalized data Raman spectroscopy; upper line Se-Se stretch vibration (310 cm-1), lower line C=O stretch vibration (1710 cm-1). TBHP was added at t = 0 min and benzyl alcohol (0.9 equivalents) at t = 39 min

By using NMR, we anticipated to find other oxidized selenium compounds besides

BSA at the end of the reaction. We studied the activation of Ph2Se2 by one equivalent of TBHP using NMR. The time course of the aromatic region of the NMR signals is shown in figure 6. The spectra revealed two new sets of peaks only after 15 minutes, which increased in intensity up to 25 minutes. The newly found 1 2 signals corresponded to BSA, in a quantity of /3 of the mixture, while the other /3 of the mixture was the remaining Ph2Se2. We were not able to find any distinct signals that we could attribute to other intermediates on the path from Ph2Se2 to BSA. This is consistent with a literature report on the oxidation of bis(p- fluorophenyl) diselenide followed by 19F-NMR.4 This is a remarkable feature of the oxidation of these (two) aromatic diselenides. Apparently, the second and third oxidation of Ph2Se2 proceeds much more easily than the first as suggested by the formation of a 2:1 mixture of Ph2Se2 and BSA, rather than a more complex mixture of selenides in different oxidation states.

43 Chapter 3

Figure 6 NMR spectra of the activation of Ph2Se2 by TBHP, data shown in the region 8.5-7.5 ppm. After t = 30 min, one equivalent of benzyl alcohol was added.

From all of the above results, we can deduct a scheme for the activation of Ph2Se2 (figure 7). The first oxidation (Ox I) is underlined by UV and IR spectroscopy and occurs within 5-10 minutes. After this, the second oxidation takes place (Ox II) as evidenced by IR. The final oxidation of compound 3 to compound 4 would then occur between 15-20 minutes, as shown by Raman spectroscopy and reaction calorimetry (see chapter 2). In the scheme we have also included the formation of compound 5: in our previous study we postulated that this intermediate is most likely responsible for the autocatalysis that occurs when BSA (compound 4) is formed. Tert-butyl benzeneperoxyseleninate (5) is most likely a more effective oxygen transfer compound than TBHP itself.5 The synthesis of this compound has so far not been successful in our hands, but we were able to show that the oxidation of benzyl alcohol with a mixture of BSA and TBHP proceeded at the same rate as with BSA alone (results not shown).

44 Spectroscopic investigations in the activation of Ph2Se2 by TBHP

Ox ! O TBHP, 5-10 min Se Se Se Se 2 1

Ox !! O O Se O TBHP, TBHP, 5 10-12 min 15-25 min

TBHP, O O O O 15-20 min Se O Se Se Se Ox !!! 4 3

Figure 7 Path of (autocatalytic) activation of Ph2Se2 by TBHP

Solvent influence on the oxidation of diphenyl diselenide by TBHP During our initial studies, we tested a variety of solvents in the catalytic oxidation of benzyl alcohol. Interestingly, we observed that some solvents gave faster initial turnovers in this reaction, often leading to higher yields. It is very likely that the solvent greatly influences the activation of the diselenide to the formation of the anhydride, the actual dehydrogenation species in the reaction.

Table 1 Reaction calorimetry data of the activation of Ph2Se2 by TBHP in different solvents a Defined as time from addition of TBHP to the time of temperature maximum (Tr-Tj) b Defined as time between temperature increase and restoration of the equilibrium Solvent Activation (min)a Temp (°C) Peak width (min)b ΔTr-Tj / Δt (K/min) max Toluene 26 80 13 2.406 Xylene 28 80 13 1.110 Heptane 30 80 14 1.956

PhCF3 35 80 18 1.212 MeCN 20 75 16 0.780

CCl4 80 70 23 0.030 EtOAc 34 70 18 0.108 TCA 142 65 53 0.020

45 Chapter 3

To better understand this behavior, we used reaction calorimetry to observe the differences of activation of Ph2Se2 in different solvents (table 1). The solvents that gave the most exothermic profile in the reaction calorimetry, which is represented by ΔTr-Tj/Δt, were heptane and toluene. Solvents which are generally regarded as good solvents for oxidation, like acetonitrile and trifluoromethylbenzene, also give reasonable steep exothermic profiles. It should be noted that there was no direct correlation in the activation time and the peak width of the exothermic profile. In a next step, we compared these activation profiles with the catalytic activity of

BSA in the same range of solvents. The catalytic action of Ph2Se2 in the diverse solvents in the oxidation of benzyl alcohol is shown in figure 8. The solvents that initially (during the first two hours) show the slowest turnover, namely heptane, toluene and xylene, are the three solvents that show the highest exothermic profile in the activation. The exception here is PhCF3, which actually shows the fastest initial turnover.

,!"# ./01232#

+!"# 450232# 627.832# *!"# 990'# :29;# )!"# <.=>?#

(!"# .@A?60/@/2.6832# B69C&# '!"#

&!"# !"#$%&'#()*$%#+,%#&

%!"#

$!"#

!"# !-!# $-!# %-!# &-!# '-!# (-!# )-!# *-!# -".#&/+0&

Figure 8 Catalytic oxidation of benzyl alcohol (2 mmol) by Ph2Se2 (5 mol%) in the selected solvent at 80 °C with 1,1 eq of THBP (in decane).

If we look at the selectivity of the reaction after six hours, the picture changes drastically (table 2). The mass balance stays good in only acetonitrile and BTF which seem to be inert under the reaction conditions. 46 Spectroscopic investigations in the activation of Ph2Se2 by TBHP

Table 2 Reaction composition after 6 hours in the Ph2Se2 (5 mol%) catalyzed oxidation of benzyl alcohol at 80 °C with 1,1 eq of THBP (in decane). Solvent Alc (t=0) (mM) Aldehyde (mM) Acid (mM) Mass balance (%) Toluene 18.3 (196.8) 174.8 7.2 102 Xylene 20.6 (197.8) 126.7 9.2 79 Heptane 16.4 (196.9) 118.0 5.7 71

CCl4 50.3 (218.6) 105.6 21.0 81 MeCN 64.2 (195.2) 98.1 28.7 98 EtOAc 43.0 (214.7) 106.0 20.4 77 TCA 80.0 (217.5) 86.6 25.0 87

PhCF3 62.5 (209.5) 68.8 73.2 94

In all other solvents there are side-products formed. In toluene, the side reaction is the oxidation of the methyl group and thus some extra benzyl alcohol is formed (2%). This is also the case with xylene, but as there are other products formed than benzyl alcohol, the selectivity drops. The main side product in heptane is the formation of the benzyl ester of benzyl alcohol. In CCl4 and TCA, the main side reaction is chlorination of starting materials and products leading to a complex reaction mixture. In EtOAc there is a high percentage of transesterification as acetylation of the starting material is observed. It is also striking that the selectivity of the oxidation in toluene is so high. In PhCF3, which has been successfully used in selenoxide elimination reactions, the selectivity drops significantly as the aldehyde is further oxidized to the acid.6 The reaction merely stops because the reaction eventually runs out of oxidant. If at this stage more oxidant is added, the oxidation proceeds only to the acid and no more alcohol is oxidized (results not shown). This is probably due to the fact that there is too much water present and the BSA is not formed anymore due to hydrolysis. It could be that in solvents which are prone to oxidation, the formation of the perester 5 leads to a loss in selectivity and the occurrence of side reactions. This seems to correlate with the reaction calorimetry data of the activation of Ph2Se2. The solvents which show a low selectivity have a longer activation time and a less steep exothermic profile. We think that the formation of the perester in these solvents accounts for the side reactions which occur. Thus, the autocatalytic activation of the diselenide is less prominent and the activation will take more time or will be less exothermic. The choice of solvent is thus of paramount

47 Chapter 3 importance. Overall we conclude that trifluoromethyl benzene is the solvent of choice, in view of its eminent properties in terms of stability and high activity. As shown previously,1 the selectivity towards aldehyde can be attained by using a substrate and oxidant feed protocol.

Conclusions Using a variety of spectroscopic techniques and reaction calorimetry we have shown that the oxidation of Ph2Se2 to BSA follows a number of distinct steps that have been described in figure 7. This is consistent with the anhydride being the active oxidant in the (catalytic) oxidation of alcohols. Reaction calorimetry has been used to verify the behavior of activation of the diselenide in different solvents. Little or no side-reactions occur in solvents that are characterized by a large exothermic profile for activation of diphenyl diselenide with TBHP. In conclusion, we found reaction calorimetry to be a useful technique for rationalizing the course of these catalytic oxidation reactions.

48 Spectroscopic investigations in the activation of Ph2Se2 by TBHP

Experimental section

Ph2Se2 (99%), TBHP in decane (5.5 M), benzyl alcohol (99+%), 1,2-dimethoxybenzene (99+%), toluene (reagent grade, stored over molecular sieves 4 Å), ethyl acetate (p.a.) acetonitrile (HPLC grade), H2O (doubly distilled) were all used as received.

UV-Vis

To a heated mixture of 0.1 M Ph2Se2 in toluene (10 mL) was added at 80 °C an equimolar amount of TBHP in decane. After 40 minutes, an equimolar amount of benzyl alcohol was added to the mixture. At selected time intervals, 50 μL aliquots were withdrawn from the reaction and diluted with 2.95 mL toluene (room temperature). The UV spectrum was recorded instantly on a UV-Vis Hewlett Packard 8452A Diode Array Spectrofotometer.

In-situ IR The IR data were recorded by means of a Bruker Matrix-MF FT-IR in combination with a fiber optic diamond ATR probe. A 0.1 M solution of Ph2Se2 in toluene was heated to 80 °C and two equivalents of TBHP in decane was added. After 40 minutes, two equivalents of benzyl alcohol was added. Besides the 2D-plots shown in the text, a 3D-plot of the reaction has also been created (see figure 10).

-1 -1 Figure 10 3D plot of the IR region from 800 cm to 1000 cm in the activation of Ph2Se2 by TBHP in toluene

In-situ Raman The Raman data were recorded by means of RamanRXN1 analyser from Kaiser Optical systems with a standard immersion probe (MR probe filtered Probe Head). A 0.1 M solution of Ph2Se2 in heptane was heated to 80 °C and two equivalents of TBHP in decane was added. After 40 minutes, two equivalents of benzyl alcohol was added.

49 Chapter 3

NMR 1H NMR 400 Hz and 13C 100 Hz spectra were recorded on a Bruker AC 400 spectrometer using tetramethylsilane as an external standard. The kinetic reactions were performed the same way as the UV-Vis analysis (vide supra), but the solvent for the reactions and for the dilutions was deuterated toluene.

Reaction calorimetry This was performed in a Multimax apparatus: a programmable 4 parallel reactor box, reaction volume from 25 to 70 ml with overhead stirring, temperature range from -25 to 150 °C, reflux cooler and inter gas purging. Each reactor can be set individually for temperature and stirring. Temperature control modes: Jacket and reactor contents. Reaction calorimetry is done by adding a known amount of heat to the reaction mixture using a calibration probe, followed by integration of the signals obtained (150Ω, 24 volts). To a solution of a specific concentration of the diselenide at the selected temperature was added the appropriate amount of TBHP in decane after which the temperature difference between the internal sensor (Tr; temperature of the reaction) and the external sensor (Tj; temperature of the jacket) was monitored.

Mass spectrometry Data were collected on a Shimadzu GCMS-QP2010 Ultra, equipped with a non-polar CP Sil-5 CB 50 m * 0.53 mm colomn.

Gas chromatography Analysis was carried out on a Shimadzu GC2010 instrument equipped with a non-polar CP Sil-5 CB 50 m * 0.53 mm colomn, using a FID detector.

Catalytic reactions To a stirred solution of substrate (2 mmol), 1,2-dimethoxybenzene (0,5 mmol, 69 mg, internal standard) and the selected diselenide (0.1 mmol) in 10 mL solvent was added at 80 °C TBHP (2.2 mmol). At several intervals, 50 µL aliquots were withdrawn, quenched with Na2SO3 (100 mg in 1.5 mL EtOAc) and the solids were filtered off. The mixture was subsequently analyzed by GC.

References

(1) van der Toorn, J. C.; Kemperman, G.; Sheldon, R. A.; Arends, I. J. Org. Chem. 2009, 74, 3085-3089. (2) McCulla, R. D.; Jenks, W. S. J. Am. Chem. Soc. 2004, 126, 16058-16065 (3) Socrates, G. Infrared and Raman Characteristic Group Frequencies; third ed.; Wiley, 2004. (4) Gancarz, R. A.; Kice, J. L. Tetrahedron Lett. 1981, 22, 1661-1662. (5) Bloodworth, A. J.; Lapham, D. J. J. Chem. Soc.-Perkin Trans. 1 1983, 471-473. (6) Ogawa, A.; Curran, D. P. J. Org. Chem. 1997, 62, 450-451.

50 4

Chapter 4: Substituted aromatic diselenides in the catalytic oxidation of alcohols

John C. van der Toorn, Gerjan Kemperman, Roger A. Sheldon and Isabel W.C.E. Arends

This chapter has been published in: European Journal of Organic Chemistry 2011, Issue 23, 4345-4352

Abstract In this chapter the influence of aromatic substituents on the activity of aromatic diselenides as catalysts in alcohol oxidations was systematically studied. For this purpose a small library of aromatic diselenides was synthesized. Using reaction calorimetry studies, a correlation was found between the time of activation of substituted diselenides with TBHP and their catalytic activity in the final oxidation of alcohols: more easily activated diselenides are more active catalysts in the catalytic oxidation of benzyl alcohol. Experiments were performed to expand the scope of the diselenide-catalyzed oxidation of alcohols towards an aliphatic alcohol, notably 1-decanol. This aliphatic alcohol was oxidized fully by three of the diselenides in our library, namely dimesitylene diselenide, bis(pentafluorophenyl) diselenide and bis(2-methoxyphenyl) diselenide. However no direct correlation between activation and the actual dehydrogenating activity of our various catalysts could be delineated. Full selectivity towards the aldehyde could be achieved by using an oxidant and substrate feed protocol. Chapter 4

Introduction Diphenyl diselenide has been reported by us and others as an efficient and selective catalyst for the oxidation of activated alcohols. Its aromatic ring gives the possibility to further explore the frontiers and limitations of these compounds. Many groups have reported the use of different aromatic diselenides in alcohol oxidations.1-4 However, a systematic investigation of the use of different diselenides in combination with an easily accessible oxidant has not been published. Improvements are needed, in order to enhance the selectivity of these catalysts when hydroperoxides are used as terminal oxidants. In addition, higher activities are needed for non-activated alcohols. One of the best selenium catalysts in terms of substrate scope is dimesitylene diselenide, but the catalyst loading is usually 50 mol%.1 In the oxidation of allylic alcohols Kuwajima et al. reported the effective use of bis(p-chlorophenyl) diselenide with trans-2-hexen-1-ol as the substrate and TBHP as the oxidant, however no catalyst loading was reported.2 In certain cases highly specific and costly oxidants are required. By using sulfonamides such as Chloramine-T sodium salt or N-chloro-4-chlorobenzenesulfonamide sodium salt, Kuwajima and coworkers found a large improvement for ortho-substituted diphenyl diselenides with ketone or ester groups as the catalysts. Again, primary aliphatic alcohols were too demanding substrates for this approach even though activated alcohols were oxidized fully. In a more recent article, the use of bis[2-(2-pyridyl)phenyl] diselenide in combination with a sulfonamide was described as a more efficient catalyst, and catalyst loadings could be reduced to 0.2 mol%.4 These findings demonstrate that aromatic diselenides still bear great promise for catalytic oxidations. Our goal here was to make a small library of substituted diselenides and to test these compounds as oxidation catalysts. For this we investigated the influence of the substituents on the first step of the reaction, the oxidation of diaryl diselenide to the anhydride with TBHP, as well as the overall activity in the oxidation of alcohols. Both benzyl alcohol as well as 1-decanol, which is an example of a non- activated alcohol, were used as substrates.

52 Substituted aromatic diselenides in the catalytic oxidation of alcohols

Results and discussion Substituted diselenides as precatalysts in oxidations with TBHP The influence of aromatic substituents on the potential of diselenides as precatalysts in oxidations with TBHP was investigated. The goal was to extend the substrate scope of the current method towards non-activated alcohols. For this purpose a small library of different diselenides was selected (figure 1).

NO2

Se Se Cl Se 2 2 2 1 2 3

Se Se Se 2 2 5 6 2 4

OMe NMe2 F F

Se Se F Se 2 2 7 2 8 F F 9

Figure 1 Selected substituted diselenides

Compound 1 is by definition our reference catalyst. Compound 2 bis(2-nitrophenyl) diselenide is a well-known and often applied epoxidation and Baeyer- Villiger catalyst.5,6 Compound 3 has been referred to as an effective catalyst in the oxidation of allylic alcohols.2 Compound 4 is known to be a very effective (pre) catalyst for alcohol oxidations, although the catalyst loading was often very high.2 Compounds 5 and 6 were selected for comparison to compound 4. Compounds 7 and 8, in contrast to compound 2, posses electron donating groups on the ortho position of the aromatic ring. The bis(pentafluorophenyl) diselenide 9 was selected because the corresponding seleninic acid was previously reported as an active reagent in dehydrogenations of non-activated alcohols.7 The non-commercially available compounds were made via the corresponding Grignard reagents, except 9, which was made via direct lithiation of pentafluorobenzene, followed by reaction 53 Chapter 4 with selenium.8 During the synthesis, we noticed that the Grignard reactions often led to a large amount of the monoselenide besides the diselenide.9 The only way these could be separated was via reversed-phase silica chromatography. The bis(2- dimethylaminophenyl) diselenide 8 was successfully prepared via the Knochel Mg / Br exchange, after which the higher ate complex was created by adding one equivalent of dioxane.10 This complex could be reacted with elemental selenium, yielding the target compound. The reaction of compounds 1-9 with TBHP was studied using calorimetry. In this way the so-called activation time, which is the time needed to form the anhydride as catalytic species, and the activation energy could be measured. The influence of the substituents on these parameters is listed in table 1. Note that the diselenide concentration is lower (6.3 mM) than in previous experiments, which leads to longer activation times (i.e. 50 minutes for Ph2Se2 instead of the aforementioned 25 minutes).

Table 1 Time of activation of (substituted) diselenides. 0.25 mmol of diselenide was dissolved in 40 mL toluene and 0.5 mmol of TBHP was added at 80 ºC. The activation time is defined as the time between the addition of the TBHP and the highest peak in the exothermic profile. The activation energy is calculated by integrating the peak surface. Diselenide Substitution pattern Activation (min) Activation energy (kJ/mol) 1 - 50 150 2 2-nitro not observed not observed 3 4-chloro 52 230 4 2,4,6-trimethyl 22 120 5 4-methyl 38 150 6 3-methyl 42 145 7 2-methoxy not observed not observed 8 2-dimethylamino 49 345 9 pentafluoro 52 190

It turns out that Ph2Se2 is one of the slowest compounds in activation. The fastest of the library is compound 4, but it delivers the lowest heat in activation of all compounds. Compound 8 shows the same profile as 1 but it has the highest activation energy, which is more than double compared to most of the other diselenides. Very remarkable is the series of compounds 4-6: the mesitylene-

54 Substituted aromatic diselenides in the catalytic oxidation of alcohols diselenide is faster in activation than 4-methyl and the 3-methyl group slows the oxidation even further.

Substituted diselenides in the catalytic oxidation of benzyl alcohol After having established the activation behavior of the compounds in the absence of substrate, the overall reaction efficiency for the catalytic oxidation of benzyl alcohol using 5 mol% of compounds 1-8 was studied. The results are presented in figure 2 for the compounds with the lowest rates, and in figure 3 for the compounds with the highest rates.

Figure 2 Oxidation of benzyl alcohol with diselenides (5 mol%) at 80 °C and TBHP (1.1 eq)

We observed that the slightly electron donating methyl groups seem to promote a fast initial turnover of the substrate. All methyl substituted diselenides are among the fastest in conversion of the substrate. Second, the ortho position seems to be very important as contradictory results are obtained with the diselenides substituted at this position. Compound 2 (the original epoxidation catalyst) is the one with the lowest activity. When comparing calorimetry data during pre- activation with activity data, a clear correlation is observed between the time of activation and the initial catalytic activity of the catalyst. The two compounds that did not show activation in the calorimetric experiments (compounds 2 and 7), 55 Chapter 4 show the lowest activity in the oxidations. The main exception is compound 8 but the 2-dimethylamino group could possibly act as a base in this reaction. The calorimetry suggests it does not have an effect in activation, but in the catalytic activity it could play an important role.

Figure 3 Oxidation of benzyl alcohol with diselenides (5 mol%) at 80 °C and TBHP (1.1 eq)

The above results demonstrate that formation of the active intermediate is highly influenced by the substituents on the aromatic ring. A strong electron-withdrawing or electron-donating substituent on the ortho position diminishes the activity of the diselenide in catalysis. A nitrogen coordinating substituent on the other hand, can possibly enhance the catalytic mode of action, a phenomenon that has been found in selenium based glutathione peroxidase mimics.11

Oxidation of an aliphatic alcohol: 1-decanol Previously we have shown that using an oxidant and substrate feed in a loop cycle, high selectivity and efficiency can be reached with diphenyl diselenide as the oxidation catalyst.12 To further explore the scope of our substituted diselenides, a catalytic test was performed with 5 mol% diselenide and 1-decanol as the substrate. Unfortunately, the yield of decanal was maximum 8% which declined over time again as it was further oxidized to capric acid. One of the other by-

56 Substituted aromatic diselenides in the catalytic oxidation of alcohols products was the decyl ester of capric acid. Therefore we decided to switch to the oxidant and substrate feed technique. First the behavior of different diselenides in the dehydrogenation of 1-decanol was studied in a stoichiometric fashion: pre- activation with one equivalent of TBHP was followed by the addition of a stoichiometric quantity of 1-decanol. The conversion of the substrate was followed in time (figure 4). In these experiments toluene was replaced with trifluoromethyl benzene to prevent oxidation of the solvent. Compounds 5 and 6 were not tested in these experiments because of their lack of activity in the catalytic oxidation of 1-decanol.

Figure 4 Yield of decanal at different time points with pre-activated diselenides (0.5 mmol diselenide and 1.1 mmol TBHP premixed for two hours at 80 °C in PhCF3, followed by adding 1 mmol 1-decanol)

Compounds 4, 7 and 9 all completely converted the alcohol to the aldehyde. The speed of oxidation is considerably slower compared to the same experiment with

Ph2Se2 and benzyl alcohol, which gave full conversion to benzaldehyde within 10 minutes. In the case of the oxidation of 1-decanol this takes at least 120 minutes. In these experiments the effect of pre-activation is excluded, and what we observe is the primary dehydrogenating activity of the corresponding seleninic anhydride with 1-decanol. It can be envisaged that the dehydrogenation step is favored by strongly inductive and electron donating substituents, such as o-methoxy 7 and the three methyl substituents in 4. Surprisingly also bis(pentafluorophenyl) diselenide 9

57 Chapter 4

- once pre-activated - shows good activity. These results show that the actual rate determining step during oxidation is not well understood, and we presume that a variety of factors is important in the dehydrogenation of alcohols by BSA and its derivatives. We decided to test the most promising catalysts in the batch wise oxidation of 1- decanol, using our previously described protocol (see chapter 2). The protocol was changed in order to prolong the period between the addition of 1-decanol and a new amount of TBHP for regeneration of the catalyst. The stepwise addition of substrate and oxidant is visualized in figures 5 and 6.

Figure 5 Oxidation of 1-decanol with different diselenides (0.25 mmol) following the oxidant and substrate feed protocol

All of the selected diselenides turned out to be good oxidation pre-catalysts using the oxidant and substrate feed protocol. One of the main problems is the overoxidation of the product leading to capric acid, which is shown in figure 6. This reaction is also selenium catalyzed as a blank run with TBHP and 1-decanal did not show capric acid formation. The formation of the intermediate seleninic perester is most likely one of the main problems as this can oxidize aldehydes to the corresponding acids. In the presence of the seleninic acid, the perester formation is also possible and thus the formation of by-product cannot be suppressed. One of the main problems in this protocol is most probably the reformation of the

58 Substituted aromatic diselenides in the catalytic oxidation of alcohols intermediate anhydride. From the above results it seems most likely that the anhydride is the active dehydrogenating species, but it could be that this reactivation is more difficult for substituted diselenides.

Figure 6 Capric acid formation during the oxidant and substrate feed protocol with different diselenides in the oxidation of 1-decanol

It is known that benzeneseleninic acid dehydrates to the anhydride at relatively low temperatures (around 70 °C), but for the other compounds we have used here, the dehydration temperature is probably higher. Even though it is possible to heat up the mixture to higher temperatures, TBHP is unstable at temperatures above 90 °C so this is not a viable option. However, compared to previous results in literature, our system has the lowest catalyst loading (25%) and highest yield (98%) in the oxidation of 1-decanol.

Conclusions We must conclude that for the mechanism of oxidation of alcohols by substituted diselenides there are two important parameters. First, there is the mode of activation of the diselenide by the oxidant (see chapter 3). Certain diselenides such as dimesitylene diselenide 4 are easily activated by TBHP whereas others are not activated at all. Secondly, the oxidation potential of the activated substituted diselenides differs greatly. Not all activated diselenides are capable of

59 Chapter 4 dehydrogenating non-activated alcohols. Finally, the regeneration of the anhydride of the active species is also expected to change when the substituents are varied. The traditional catalytic oxidation of 1-decanol with aromatic diselenides was hampered by formation of side-products due to overoxidation and transesterification. However, when a pre-activation of the diselenides was performed, followed by addition of 1-decanol, compounds 4, 7 and 9 were capable of converting this substrate selectively to the aldehyde. Obviously, the steric and electronic demands for the hydrogen abstraction are optimal for these catalysts where other diselenides are not capable of quantitative conversion of the substrate. The aforementioned diselenides were tested in the catalytic oxidation of 1-decanol using our previously optimized oxidant and substrate feed protocol. It was shown that this was indeed a good way of improving the catalyst performance. In conclusion, we have shown that by introducing substituents on aromatic diselenides, catalytic properties could be improved. However, more studies are needed in order to predict these properties and to develop better dehydrogenating selenium catalysts.

60 Substituted aromatic diselenides in the catalytic oxidation of alcohols

Experimental section

Ph2Se2 (99%), Bis(2-nitrophenyl)diselenide (97%), Bis(4-chlorophenyl)diselenide (98%) TBHP in decane (5.5 M), benzyl alcohol (99+%), 1,2-dimethoxybenzene (99+%), selenium (powder 99%), i- propylmagnesiumchloride lithium chloride complex (1.3 M in THF), dodecane (99%), dioxane (reagent grade), p-tolylmagnesium bromide (1.0 M in THF), m-tolylmagnesium bromide (1.0 M in THF), pentafluorobenzene (98%), 2-bromo-N,N-dimethylaniline (97%), 2-methoxyphenylmagnesium bromide (1.0 M in THF), trifluoromethylbenzene (reagent grade), 1-decanol (99%), toluene (reagent grade, stored over molecular sieves 4 Å), ethyl acetate (p.a.) acetonitrile (HPLC grade), H2O (doubly distilled) were all used as received.

NMR 1H NMR 400 Hz and 13C 100 Hz spectra were recorded on a Bruker AC 400 spectrometer using tetramethylsilane as an external standard.

Mass spectrometry Data was collected on a Shimadzu GCMS-QP2010 Ultra, equipped with a non-polar CP Sil-5 CB 50 m * 0.53 mm colomn.

Gas chromatography Analysis was carried out on a Shimadzu GC2010 instrument equipped with a non-polar CP Sil-5 CB 50 m * 0.53 mm colomn, using a FID detector.

Reaction calorimetry This was performed in a Multimax apparatus: a programmable 4 parallel reactor box, reaction volume from 25 to 70 ml with overhead stirring, temperature range from -25 to 150 °C, reflux cooler and inter gas purging. Each reactor can be set individually for temperature and stirring. Temperature control modes: jacket and reactor contents. Reaction calorimetry is done by adding a known amount of heat to the reaction mixture using a calibration probe, followed by integration of the signals obtained (150Ω, 24 volts). To a solution of a specific concentration of the diselenide at the designated temperature was added the appropriate amount of TBHP in decane after which the temperature difference between the internal sensor (Tr; temperature of the reaction) and the external sensor (Tj; temperature of the jacket) was monitored.

Stoichiometric reactions To a stirred solution of the diselenide (0.5 mmol), dodecane (0.25 mmol, internal standard) in 10 mL trifuoromethylbenzene at 80 °C was added TBHP (1 mmol from 5.5 M solution in decane) and after the appropriate amount of time, 1-decanol (0,5 mmol) was added. At several intervals, 50 µL aliquots were withdrawn, quenched with Na2SO3 (100 mg in 1.5 mL EtOAc) and the solids were filtered off. The mixture was subsequently analyzed by GC.

61 Chapter 4

Oxidant and substrate feed reactions To a stirred solution of the diselenide (0.25 mmol), dodecane (0.25 mmol, internal standard) in 5 mL trifuoromethylbenzene at 80 °C was added TBHP (0.5 mmol from 5.5 M solution in decane) and after the appropriate amount of time, 1-decanol (0.25 mmol) was added. After 2 hours 1-decanol was added (0.25 mmol 1-decanol). After another hour, TBHP (0.25 mmol) was added. These steps were repeated four times. At several intervals, 50 µL aliquots were withdrawn, quenched with Na2SO3 (100 mg in 1.5 mL EtOAc) and the solids were filtered off. The mixture was subsequently analyzed by GC.

Synthesis of diselenides

Bis(2,4,6-trimethylphenyl) diselenide (4) This compound was prepared following the literature procedure. Starting from 10 mL of 1M of 2,4,6- trimethylphenylmagnesium bromide, the yield was 77%. m/z (EI) 398 (with a typical diselenide pattern). δH (400 MHz; CDCl3) 6.48 (s, 2H), 2.27 (s, 3H), 2.24 (s, 6H); δC (100 MHz; CDCl3) 144.1 (Cq), 139.5 (Cq), 129.5 (Cq), 128.7, 24.6, 21.4;

Bis(4-methylphenyl) diselenide (5) To 30 mL of a 1 M solution of p-tolylmagnesium bromide, Se powder (30 mmol) was added portion wise and after complete addition, the mixture was heated for 30 min to 40 °C. The color of the solution changed in this time from light-brown to bright green. After cooling the mixture to room temperature it was poured onto a mixture of ice and saturated NH4Cl. This was stirred for 10 minutes and the water layer was extracted 5 times with Et2O. The combined organic layers were dried over MgSO4 and the solvent was evaporated in vacuo. The residue was taken up in 75 mL of EtOH and 100 mg of KOH was added. A stream of air was passed over the solution for 45 minutes after which the mixture was stirred in an open flask overnight. A white compound had crystallized out which was filtered off (dimer) and the mother liquor was subjected to RP18-silica gel chromatography which could separate the diselenide from the monoselenide (eluent: MeCN:H2O 8:2, Rf 0.32). Yield: 2.57 mmol (17%) m/z (EI) 342 (with a typical diselenide pattern). δH (400 MHz;

CDCl3) 7.48 (d, 2H, J=6.4 Hz), 7.07 (d, 2H, J=7.6 Hz), 2.34 (s, 3H); δC (100 MHz; CDCl3) 138.4 (Cq), 132.7, 130.3, 128.1 (Cq), 21.5

Bis(3-methylphenyl) diselenide (6) To 10 mL of a 1 M solution of m-tolylmagnesium bromide, Se powder (10 mmol) was added portion wise and after complete addition, the mixture was stirred for another three hours during which a thick slurry formed. The reaction mixture was poured onto a mixture of ice and saturated NH4Cl. During this work-up, large amounts of selenium red formed. The mixture was stirred for 10 minutes and the water layer was extracted 5 times with Et2O. The combined organic layers were dried over

MgSO4 and the solvent was evaporated in vacuum. The residue was subjected to RP18-silica gel chromatography which could separate the diselenide from the monoselenide (eluent: MeCN:H2O 8:2, Rf 0.35). Yield: 1.07 mmol (21%); m/z (EI) 342 (with a typical diselenide pattern). δH (400 MHz;

CDCl3) 7.34 (m, 1H), 7.25 (m, 3H), 2.31 (s, 3H); δC (100 MHz; CDCl3) 138.9 (Cq), 133.0, 129.3, 129.0, 128.7, 128.1, 21.3

Bis(2-methoxyphenyl) diselenide (7) To 40 mL of a 1M solution in THF of the Grignard, Se powder (40 mmol) was added portion wise in such a rate that the temperature did not exceed 35 °C in the reaction. Ten minutes after the 62 Substituted aromatic diselenides in the catalytic oxidation of alcohols complete addition of the Se, the mixture started to polymerize and thus the mixture was hydrolyzed with saturated NH4Cl. The mixture was extracted three times with EtOAc, the organic layers were combined, dried over MgSO4 and subsequently the solids were filtered off. All volatiles were removed in vacuo and the compound was purified over a silica column (PE:EtOac = 95:5). The orange liquid was subjected to kugelrohr distillation, and two fractions were obtained, one fraction at 170 °C (0.1 mbar) and one fraction at 210 °C (0.1 mbar). The diselenide was recrystallized from EtOH. Yield: 9.59 mmol (48%); m/z (EI) 374 (with a typical diselenide pattern). δH (400 MHz; CDCl3) 7.54 (d, 1H, J = 7.6 Hz), 7.21 (t, 1H, J = 8 Hz), 6.87 (t, 1H, J = 7.2 Hz), 6.82 (d, 1H, J = 8 Hz), 3.91 (s, 3H); δC (100

MHz; CDCl3) 157.3 (Cq), 131.0, 128.6, 122.3, 119.0 (Cq), 110.6, 56.4

Bis(2-dimethylaminophenyl) diselenide (8) To 5 mmol of the starting compound in 10 mL THF was added 1,1 eq iPrMgCl • LiCl complex and after full addition, one equivalent of dioxane was added. A precipitate was observed and conversion of the starting material was followed by taking samples. After overnight stirring, conversion was complete and Se-powder was added gradually. The mixture was stirred for another two hours and was then hydrolyzed with NH4Cl. The mixture was purified over a silica column (PE only). Yield: 1.10 2+ mmol (44%); m/z (EI) 199 (M ). δH (400 MHz; CDCl3) 7.52 (d, 1H, J = 8.1 Hz), 7.15 (2 H, m), 7.00 (1

H, t, J = 6.4 Hz), 2.80 (s, 6H); δC (100 MHz; CDCl3) 152.8 (Cq), 129.3, 127.6, 126.3, 120.8, 45.8

References (1) Shimizu, M.; Kuwajima, I. Tetrahedron Lett. 1979, 2801-2804. (2) Kuwajima, I.; Shimizu, M.; Urabe, H. J. Org. Chem. 1982, 47, 837-842. (3) Onami, T.; Ikeda, M.; Woodard, S. S. Bull. Chem. Soc. Jpn. 1996, 69, 3601-3605. (4) Ehara, H.; Noguchi, M.; Sayama, S.; Onami, T. J. Chem. Soc.-Perkin Trans. 1 2000, 9, 1429-1431. (5) Mlochowski, J.; Brzaszcz, M.; Giurg, M.; Palus, J.; Wojtowicz, H. Eur. J. Org. Chem. 2003, 4329-4339. (6) Giurg, M.; Syper, L. Phosphorus Sulfur Silicon Relat. Elem. 2008, 183, 970-985. (7) Barton, D. H. R.; Wang, T. L. Tetrahedron Lett. 1994, 35, 5149-5152. (8) Klapotke, T. M.; Krumm, B.; Polborn, K. Eur. J. Inorg. Chem. 1999, 1359-1366. (9) Muller, J.; Terfort, A. Inorg. Chim. Acta 2006, 359, 4821-4827. (10) Krasovskiy, A.; Straub, B. F.; Knochel, P. Angew. Chem.-Int. Edit. 2006, 45, 159-162. (11) Mugesh, G.; du Mont, W. W.; Sies, H. Chem. Rev. 2001, 101, 2125-2179. (12) van der Toorn, J. C.; Kemperman, G.; Sheldon, R. A.; Arends, I. J. Org. Chem. 2009, 74, 3085-3089.

Chapter 5: Glycerol-based solvents as green reaction media in selenium catalyzed epoxidations with hydrogen peroxide

John C. van der Toorn, Héctor García-Marín, José A. Mayoral, Isabel W. C. E. Arends and José I. García

Parts of this chapter have been published in: Green Chemistry 2009, Vol. 11, 1605-1609 Journal of Molecular Catalysis A: Chemical 2011, Vol. 334, 83-88

Abstract We found that fluorinated members of a new class of glycerol based solvents, such as 1,3-bis(2,2,2-trifluoroethoxy)-2-propanol and 1,3-bis(2,2,3,3,3-pentafluoropropoxy)-2-propanol, are highly suitable for epoxidation reactions catalyzed by aromatic selenium compounds with hydrogen peroxide as the oxidant. Similar to trifluoroethanol, these new solvents have sufficient hydrogen bond donating capacity to activate hydrogen peroxide as the oxidant. In the epoxidation of cyclooctene and cyclohexene, high yields of the corresponding epoxides were observed and, in addition, the hydrolysis of the epoxides was largely prevented in the case of cyclohexene. Due to the higher boiling point of these solvents, the design of a process was achieved in which simple distillation was sufficient to separate the product from the catalytic phase. This opens the way to efficient and straightforward downstream processing, which is usually cumbersome in fine chemical synthetic sequences. Chapter 5

Introduction One of the twelve principles of Green Chemistry is to circumvent the use of toxic solvents. Solvents are responsible for a large part of the waste generated by chemical processes. In many cases, organic solvents cannot simply be left out or replaced by water. There is therefore a definite need to develop sustainable and non-toxic organic solvents,1 which have solvating properties comparable to i.e. dichloromethane, namely polar and non-coordinating. In general, organic solvents are chemical substances derived from petroleum. Five of the ten contaminants most abundant in the atmosphere are organic solvents and most solvents have been labelled as toxic or hazardous substances by the European program REACH (Registration, Evaluation and Authorization, and Restrictions of Chemicals). Both environmental restrictions in chemical compounds and a steady decline in petrol sources have propelled chemistry to develop new and renewable sources of materials. Current research is aimed at trying to find new renewable solvents. A new solvent will have to meet the demands for sustainability in terms of renewability, (non)toxicity, (non)volatility and should exhibit a small environmental impact.2,3 Glycerol-based solvents are highly promising in this respect. Glycerol is currently produced as a co-product in biodiesel preparation. This increase in availability of glycerol has initiated new research and industrial processes for making chemicals from glycerol.4 A wide family of glycerol derivatives, consisting of 1,3-dialkoxy-2-propanols and 1,2,3- trialkoxypropanes, both symmetrically and unsymmetrically substituted at the terminal positions, has recently been synthesized, and the possible role of these glycerol based solvents as solvents has been evaluated through physico-chemical measurements.5 In this work a selected group of this family of glycerol derivatives, consisting of eighteen 1,3-dialkoxy-2-propanols and 1,2,3-trialkoxypropanes (figure 1) was tested for their potential to act as solvent in catalyzed epoxidation reactions, using aqueous hydrogen peroxide as oxidant and bis[3,5-bis(trifluoromethyl)phenyl] diselenide as the catalyst.6 This reaction is usually performed in dichloromethane or 2,2,2-trifluoroethanol (TFE) as solvents.

66 Glycerol based solvents for selenium catalyzed epoxidations

OR2 R1O OR2 R groups Example: 3i, 0, 1 0 = H 4t = tBu 1 = Me 4i = iBu OH 2 = Et 3F = CF3CH2 iPr O OMe 3i = iPr 5F = CF3CF2CH2 4 = Bu 7F = CF3CF2CF2CH2

Figure 1 Structures and simplified nomenclature for glycerol based solvents

The aryl seleninic acid derived from compound 1 is one of the most active catalysts reported for epoxidations with HOOH.6 In this catalytic reaction, the diselenide catalyst is oxidized by hydrogen peroxide to give the corresponding aryl seleninic acid which is the actual catalytic species for epoxidation (figure 2).7

H2O2 H2O F C F3C F C 3 3 O H2O2 O Se Se Se OH OOH F C F3C 3 1 2 F3C

O R R

Figure 2 Catalytic epoxidation with arylseleninic acids and hydrogen peroxide

The success of this catalyst is largely due to the beneficial influence of trifluoroethanol on hydrogen peroxide. Physicochemical and theoretical studies have indicated that fluorinated alcohols act as a template, activating oxygen transfer through multiple hydrogen bonding.8,9 It thus makes hydrogen peroxide a more active oxidant which can even oxidize olefins in the absence of catalysts. However, trifluoroethanol is highly volatile, toxic and expensive, and a sustainable alternative for this solvent would therefore be highly desirable. A novel solvent would have to exclude these negative properties, while activating hydrogen peroxide at the same time. In addition, we were interested in high-boiling solvents,

67 Chapter 5 that can facilitate down-stream processing steps such as distillation. In this way easy separation of reagents and products from the catalytic phase should become more straightforward. The catalytically active phase can thus be reused for more reaction cycles. For instance, in the case of ionic liquids, it has been shown that product and reagents can be distilled out of the catalytic active phase.10 In this chapter we explore the possibility of combining the best of all of the above described: a green oxidant, a homogeneous catalyst and the use of solvents from renewable resources. In addition, a proof-of-principle for recyclable catalytic phases will be given.

Results and discussion We started by studying the epoxidation of cyclooctene using diselenide 1 as the precatalyst and H2O2 as the oxidant in different solvents. Cyclooctene was chosen as the model substrate as this is a relatively reactive alkene and the epoxide is rather stable. A variety of solvents was selected to get a heterogeneous and representative group. Therefore, 10 common solvents and 18 glycerol derivatives were tested (table 1; for full names of the glycerol solvents, see the experimental section).

Table 1 Solvents tested in the epoxidation of cyclooctene

Traditional solvents (abbreviation) Glycerol based solvents 2,2,2-trifluoroethanol (TFE) 101 3i14 dichloromethane (DCM) 104t 114t n-butanol (n-BuOH) 404t 414 n-hexanol (n-HxOH) 404 444 ethanol (EtOH) 3F03F 3i13F 2-propanol (2-PrOH) 5F05F 4t13F

di-i-propyl ether (i-Pr2O) 7F07F 413F diethylene glycol dibutyl ether (DEGDBE) 111 3F13F ethylene glycol dimethyl ether (EGDME) 113I 114I 1,4-dioxane (Diox)

The highest initial turnover frequency (TOF0) and the fastest total conversion in the reaction was achieved using trifluoroethanol as solvent, but some glycerol- based solvents, mostly bearing fluorinated alkyl chains (3F03F, 5F05F, 7F07F and 3F13F) also showed fast conversions and high turnover frequencies (table 2). Some

68 Glycerol based solvents for selenium catalyzed epoxidations of these solvents performed better than dichloromethane, which was considered the second best solvent for this system. No other by-products than cyclooctan-1,2-diol were observed in the reaction. With these results of the epoxidation of cyclooctene using diselenide 1 as precatalyst and H2O2 as oxidant in different reaction media, solvent properties effects were investigated through QSPR equations relating solvent polarity and hydrophobicity parameters with catalytic activity.11 From all tests, three common organic solvents and four glycerol derivatives were chosen for further optimization. The set of solvents used is the following: TFE, DCM, n-BuOH, 404, 3F03F, 3F13F, 5F05F. The results of these tests are shown in table 2.

Table 2 Rate of cyclooctene epoxidation in different solvents (see experimental section); a -1 -1 -1 b TOF0 (h ), initial rate in mmol product * mmol catalyst h ; GC-conversion, mmol converted olefin * mmol olefin start-1; c Selectivity, mmol epoxide * mmol converted olefin-1; d Blank test without catalyst, reaction conducted at 57 °C

N -1 a b b c Solvent E T β TOF0 (h ) Conv (%) (20 min) Conv (%) (120 min) Select. (%) TFE 0.898 0.000 491 99 100 97 TFEd - 1 26 91 DCM 0.309 0.000 153 54 93 93 n-BuOH 0.586 0.880 100 18 59 93 3F03F 0.701 0.700 155 61 >99 99 3F03Fd - 1 4 - 5F05F 0.699 0.700 214 65 >99 93 5F05Fd - 1 1 - 3F13F 0.553 0.500 130 48 88 88 3F13Fd - 0 3 - 404 0.450 0.600 46 16 52 90 404d - 0 2 -

It can be seen that there is a direct relation between polarity of the solvent and the reaction kinetics of the epoxidation. Using the data in table 2, the following N regression equation can be established (with E T denoting the hydrogen bond donor capacity and β denoting the hydrogen bond acceptor capacity):

N 2 TOF0 = 591.6 ⋅ E T - 263.3 β - 43.2 (N = 7; R = 0.970; s = 30.3)

69 Chapter 5

The main feature of the solvents should be that they have a good hydrogen bond N donor capacity (E T) while the hydrogen bond acceptor capacity (β) should be low even though the influence of this feature is smaller. These results agree with the activation of HOOH by multiple hydrogen bonds as proposed by Berkessel et al.8 The two best-performing catalytic phases, namely those with TFE and 3F03F as solvents, were recharged with fresh hydrogen peroxide and cyclooctene, in order to continue the reaction. The catalyst remained fully active and the reaction could take place for up to eleven cycles without loss of activity (figure 3).

Epoxidation of cyclooctene in 3F03F is somewhat slower than in TFE (see TOF0 in Table 2), so whereas almost complete cyclooctene conversion is achieved after 2 h of reaction time in TFE, some unreacted cyclooctene remains in the reaction medium after each cycle (Figure 3b). In all cases there was full selectivity towards cyclooctene oxide as the sole product. These results show the stability of the catalytic system, which could allow the possibility of recovery and recycling from a catalytically active phase.

%#" %#" &&'("'()*+" (+*$ &&'("'()*+" (+*$ ,',-("&&'("'()*+" ,',-("&&'("'()*+" &-.."/-(-+0)" &-.."/-(-+0)" %!" %!" &&'(")1'234)" &&'(")1'234)"

$#" $#" !!"#$ !!"#$ $!" $!"

#" #"

!" !" !" #" $!" $#" %!" !" #" $!" $#" %!" %&!'$()*$ %&!'$()*$ Figure 3 Recharging cycles in the epoxidation of cyclooctene with hydrogen peroxide using

TFE (a) and 3F03F (b) as solvents (1.5 eq H2O2, 0.9 mol% catalyst).

As the boiling point of TFE (79 ºC) is lower than that of cyclooctene oxide (189 ºC), the direct distillation of the product from the reaction medium is impossible, which prevents the option of recovery of the phase containing the catalyst. In contrast, when 3F03F (197 ºC) is used as the solvent in this reaction, the distillation of the cyclooctene oxide from the reaction should in principle be possible.

70 Glycerol based solvents for selenium catalyzed epoxidations

After the distillation of the epoxide in high vacuum at low temperature, it turned out that the catalyst was slightly deactivated (Table 3); moreover the complete separation of cyclooctene oxide from 3F03F was impossible by fractional distillation as some solvent (30–40%) was lost after each recovery cycle. Therefore it deemed logical to explore this strategy using a more volatile epoxide which should solve the fractional distillation issue. To test this hypothesis we chose cyclohexene for the recycling experiments. Furthermore the epoxide of cyclohexene is more sensitive to hydrolysis than that of cyclooctene oxide, thus this substrate constitutes a more valuable probe regarding catalyst application.

Table 3 Recycling of the catalytic phase by direct distillation of the cyclooctene oxide from the reaction medium, in the case of 3F03F. Conditions: 0.5 mol% diselenide 1, 2 mmol of cyclooctene, 4 mmol H2O2 (50%), 2 mL solvent, 25 °C, 2h. After the first cycle, cyclooctene oxide was removed by high vacuum distillation and fresh oxidant and substrate was added. a -1 -1 -1 b c TOF0 (h ): initial rate in mmol product * mmol catalyst h ; GC-yield of epoxide; pure cyclooctene oxide obtained, but it was contaminated with traces of solvent.

-1 a b,c Solvent Cycle TOF0 (h ) Yield (%) 3F03F 1 160 82 2 90 85

Epoxidation of cyclohexene The epoxidation of cyclohexene was carried out using six of the seven solvents employed in the cyclooctene epoxidation experiments. Although good conversions and selectivities were obtained after relatively short reaction times, after some time the epoxide decomposed slowly and the selectivity of the reaction decreased (Table 4). Figure 4 shows the evolution of epoxide yield with time in each solvent tested (without a base such as NaOAc added). Water miscibility seemed to be the key point controlling the selectivity. A low miscibility of water with the solvent slows the hydrolysis of the epoxide by reducing the concentration of water in the catalytically active phase. Water miscibility depends on both hydrogen bond donor (HBD) ability and lipophilicity of the solvent. In this regard, it is worth noting that 3F03F gathers both a high HBD ability and immiscibility with water which results in high conversions from the alkene to the product, but also a low hydrolysis rate of the epoxide. This unusual combination of physicochemical properties has also recently been reported for other fluorinated derivatives of glycerol,5 and this 71 Chapter 5 allows the development of special applications in catalytic systems, unattainable for common organic solvents. One of these applications has been recently reported.12

Table 4 Rate of cyclohexene epoxidation in different solvents (see experimental section); a -1 -1 -1 b TOF0 (h ), initial rate in mmol product * mmol catalyst h ; GC-conversion, mmol converted olefin * mmol olefin start-1; c Selectivity, mmol epoxide * mmol converted olefin-1; d cyclohexan-1,2-diol was the only by-product; e 0.5 mol% NaOAc was added

-1 a b c b c Solvent TOF0 (h ) Conv. (%) (20 min) Select. (%) Conv. (%) (120 min) Select. (%) TFE 398 100 89 100 16 DCM 133 39 88 80 75 n-BuOH 44 11 95 59 35 3F03F 183 46 86 96 73 3F13F 98 22 70 66 8 404 10 11 69 38 70 TFEe 240 - - 100 98 3F03Fe 84 - - 95 97

#!!" +!" ,-." /01" 234567" %-!%-" %-#%-" &!&" *!" )!" (!" '!"

!"#$%&'()& &!" %!" $!" #!" !" !" '!" #!!" #'!" $!!" *"+#&'+",)&

Figure 4 Decay of epoxide yield in time in different solvents

The reaction was also carried out with the two best solvents (TFE and 3F03F) in the presence of traces of sodium acetate to increase the pH of the reaction medium and to minimize the hydrolysis suffered by the epoxide (final two entries table 4). Although the reaction was initially slower under these conditions, the hydrolysis was totally avoided. Similar to the case of cyclooctene, it was not

72 Glycerol based solvents for selenium catalyzed epoxidations possible to directly distill the product from the reaction media when TFE was used, because its boiling point (79 ºC) is lower than that of cyclohexene oxide (129 ºC). However, when 3F03F (197 ºC) was used as the solvent, the fractional distillation of the cyclohexene oxide from the reaction was possible (recovering ca. 80% of the produced epoxide) with no loss of solvent after each distillation. Unfortunately, we found that after every distillation which was performed in high vacuum at low temperature, the catalyst was progressively deactivated (table 5).

Table 5 Recycling of the catalytic phase by direct distillation of the cyclohexene oxide from the reaction medium, in the case of 3F03F. Conditions: 0.5 mol% diselenide 1, 2 mmol of cyclooctene, 4 mmol H2O2 (50%), 2 mL solvent, 25 °C, 2h. After the first cycle, cyclooctene oxide was removed by high vacuum distillation and fresh oxidant and substrate was added. a -1 -1 -1 b c TOF0 (h ): initial rate in mmol product * mmol catalyst h ; GC-yield of epoxide; pure cyclooctane oxide obtained

-1 a b c Solvent Cycle TOF0 (h ) Yield (%) Isolated (%) 3F03F 1 95 73 57 2 20 20 16 3 14 8 7

The remaining hydrogen peroxide was responsible for catalyst deactivation during the vacuum distillation. This deactivation is not merely due to the presence of excess hydrogen peroxide as the catalyst is stable under these circumstances. To avoid the deactivation, the reaction was carried out with a sub-stoichiometric amount of oxidant. Although this obviously results in lower and slower olefin conversion, the stability of catalyst and product is higher under distillation conditions, which allows catalyst and solvent recovery by direct distillation of the cyclohexene oxide and the unreacted cyclohexene. By working at low cyclohexene conversions, three catalytic cycles were carried out (Table 6). Complete recovery of the catalytic medium (catalyst and the glycerol-derived solvent, 3F03F) was possible. Only in the third catalytic cycle a small amount of diol was observed. Cyclohexene oxide could be easily isolated in pure form by distillation.

73 Chapter 5

Table 6 Catalyst reuse in epoxidation of cyclohexene with 3F03F as solvent; Conditions: 0.5 a mol% diselenide 1, 4 mmol of cyclooctene, 2 mmol H2O2 (50%), 2 mL solvent, 25 °C, 2h. -1 -1 -1 b TOF0 (h ): initial rate in mmol product * mmol catalyst h ; GC-yield of epoxide, based on the amount of oxidant; c Total turnover number, mmol epoxide * mmol catalyst-1; d cyclohexan-1,2-diol was the only by-product observed.

-1 a b c d Cycle Time (min) TOF0 (h ) Yield (%) TON Diol (%) 1 5 28 14 0 10 129 48 25 0 120 >99 54 0 Distillation of the product and additions of fresh reagents for the second cycle 2 5 31 84 0 10 88 52 91 0 120 >99 110 0 Distillation of the product and addition of fresh reagents for the third cycle 3 5 26 135 4 10 121 44 145 4 120 76 155 15

Conclusions A new group of renewable solvents has been successfully tested in the selenium- catalyzed epoxidation of cyclooctene and cyclohexene with hydrogen peroxide as oxidant. The results were in some cases comparable to standard organic solvents but sometimes even better results were achieved. Especially the prevention of epoxide hydrolysis was remarkable in certain members of these designer solvents (see figure 4). Moreover, a quantitative relationship between solvent polarity properties and rate of epoxidation has been established, leading to the conclusion that the best solvents for this transformation should have a high hydrogen bond donor ability, but a low hydrogen bond acceptor ability. Catalyst stability tests show that the catalytic medium (solvent with the seleninic acid) can be repeatedly recharged with reactants without loss of catalytic activity and selectivity. Optimization of the reaction conditions results in the preparation of a recoverable catalytic phase allowing direct distillation of the epoxide product, with further recycling and reuse of the solvent and the catalyst. This recovery strategy could, in principle, be extrapolated to other catalytic transformations carried out in this kind of green solvents.

74 Glycerol based solvents for selenium catalyzed epoxidations

Experimental section Catalytic experiments The epoxidation of alkenes was carried out at 25 ºC, using the following conditions: 0.5 mol % of selenide compound 1, bis[3,5-bis(trifluoro-methyl)phenyl] diselenide, was dissolved in 2 mL of solvent. Then hydrogen peroxide (50%) was added. After the solution became colorless, 0.4 mmol of veratrole or 2 mmol of ethylene glycol dimethyl ether was added (internal standard) followed by 4 mmol of olefin. Reactions were followed by taking samples at regular time intervals. Those samples were dissolved in ethyl acetate, and some manganese dioxide was added to quench the excess of hydrogen peroxide in the sample. Samples were dried over sodium sulphate and analyzed by GC.

Synthesis of glycerol based solvents Glycerol-based solvents were obtained by ring opening of either the appropriate glycidol ether (non- symmetric glycerol based solvents) or epichlorohydrin (symmetric glycerol-based solvents) with the corresponding alkoxide in alcoholic media.6 After this, the opening glycerol-based solvents were purified by vacuum distillation.

Systematic names of glycerol based solvents (table 1) 1,3-dimethoxy-2-propanol (101), 1-methoxy-3-tert-butoxy-2-propanol (104t), 1-n-butoxy-3-tert- butoxy-2-propanol (404t), 1,3-di-n-butoxy-2-propanol (404), 1,3-bis(2,2,2-trifluoroethoxy)-2- propanol (3F03F), 1,3-bis(2,2,3,3,3-pentafluoropropoxy)-2-propanol (5F05F), 1,3-bis(2,2,3,3,4,4,4- heptafluorobutoxy)-2-propanol (7F07F), 1,2,3-trimethoxypropane (111), 1,2-dimethoxy-3- isopropoxypropane (113i), 1,2-dimethoxy-3-isobutoxypropane (114i), 1-butoxy-2-methoxy-3- isopropoxypropane (3i14), 1,2-dimethoxy-3-tert-butoxypropane (114t), 1,3-di-n-butoxy-2- methoxypropane (414), 1,2,3-tri-n-butoxypropane (444), 1-isopropoxy-2-methoxy-3-(2,2,2- trifluoroethoxy)propane (3i13F), 1-tert-butoxy-2-methoxy-3-(2,2,2-trifluoroethoxy)-propane (4t13F), 1-n-butoxy-2-methoxy-3-(2,2,2-trifluoro-ethoxy)propane (413F), and 2-methoxy-1,3-bis (2,2,2-trifluoroethoxy)propane (3F13F).

Recycling experiments After certain reaction times the epoxidation product was distilled off during 1–2 h at 50 ºC and 10 mbar. Then more hydrogen peroxide was added to the remaining distillation mixture (solvent and catalyst). After a few minutes more olefin was added. Reactions were followed by GC as was previously described.

Stability of solvents In no case by-products coming from solvent oxidation in the reaction conditions were found. Furthermore, neither peroxide compounds nor by-products from decomposition were observed when solvents were put under 50 bar O2 (8%) at 100 ºC for 4 hours. These results indicate the remarkable stability of the solvents employed under oxidation conditions, even in the case of secondary alcohols.

References (1) Anastas, P. T.; Williamson, T. C. Green Chemistry; Oxford University Press: Oxford, 1998. (2) Capello, C.; Fischer, U.; Hungerbuhler, K. Green Chem. 2007, 9, 927-934. (3) Hellweg, S.; Fischer, U.; Scheringer, M.; Hungerbuhler, K. Green Chem. 2004, 6, 418-427. 75 Chapter 5

(4) Pagliaro, M.; Ciriminna, R.; Kimura, H.; Rossi, M.; Della Pina, C. Angew. Chem.-Int. Edit. 2007, 46, 4434-4440. (5) Garcia, J. I.; Garcia-Marin, H.; Mayoral, J. A.; Perez, P. Green Chem. 2010, 12, 426-434. (6) ten Brink, G. J.; Fernandes, B. C. M.; van Vliet, M. C. A.; Arends, I. W. C. E.; Sheldon, R. A. J. Chem. Soc.-Perkin Trans. 1 2001, 224-228. (7) Syper, L.; Mlochowski, J. Tetrahedron 1987, 43, 207-213. (8) Berkessel, A.; Adrio, J. A. J. Am. Chem. Soc. 2006, 128, 13412-13420. (9) de Visser, S. P.; Kaneti, J.; Neumann, R.; Shaik, S. J. Org. Chem. 2003, 68, 2903-2912. (10) Dupont, J.; de Souza, R. F.; Suarez, P. A. Z. Chem. Rev. 2002, 102, 3667-3691. (11) Garcia-Marin, H.; van der Toorn, J. C.; Mayoral, J. A.; Garcia, J. I.; Arends, I. Green Chem. 2009, 11, 1605-1609. (12) Aldea, L.; Fraile, J. M.; Garcia-Marin, H.; Garcia, J. I.; Herrerias, C. I.; Mayoral, J. A.; Perez, I. Green Chem. 2010, 12, 435-440.

76 6

Chapter 6: The design and synthesis of selenium catalysts capable of selective and irreversible attachment to enzymes

John C. van der Toorn, Ron van Klaveren, Frank Hollmann, Roger A. Sheldon and Isabel W.C.E. Arends

Abstract Aryl seleninic compounds are highly active epoxidation catalysts with HOOH. The general aim is therefore to anchor these compounds in a well-defined chiral environment, in order to design a chiral oxidation catalyst. In this chapter, the synthesis of aromatic selenium compounds for selective and irreversible binding to enzymes is presented. First, the attempted synthesis of a phosphonate based Se- catalyst for selective binding to serine hydrolases is presented. Unfortunately the challenging synthesis of this complex precursor could not be accomplished. In the second part of this chapter, a successful strategy for maleimide based Se-catalysts is presented. The latter Se-based catalyst precursors should be capable of selective and irreversible binding to enzymes containing a cysteine group in the active site, via a Michael-type addition of the active-site thiol to the maleimide. Chapter 6

Introduction The search for selective oxidative transformations is ongoing in organic chemistry, especially in view of the twelve commandments of Green Chemistry.1,2 One of the obvious choices for sustainable and selective oxidation methodologies is the use of enzymes. Enzyme catalyzed reactions often operate under ambient reaction conditions, in a relatively harmless solvent (water). The change of enzyme properties such as enantiospecificity, thermal stability or activity and stability issues under non-natural conditions can nowadays be addressed through enzyme engineering. Both directed evolution as well as semi-rational approaches have already proven their capabilities.3-5 Hence, integration of enzyme catalysis into organic synthesis is steadily gaining importance.6,7 Still there is a need for new approaches in the application of biocatalysts in oxidative transformations.8 Oxidoreductases which perform redox-reactions are often co-factor dependent.9 Co-factors are complex, expensive molecules, which can only be applied in catalytic quantities in (larger) processes via an in situ regeneration system.10,11 Many systems have been developed to circumvent the problems related to cofactor regeneration. Examples of successful regeneration systems include i.e. enzyme coupled systems, photochemical approaches and electrochemical regeneration.12 On industrial scale however, the use of whole-cell systems for redox biotransformations is still preferred. This ‘cofactor-challenge’ may be circumvented by the use of heme peroxidases which can use simple and cheap oxidants such as hydrogen peroxide directly via the peroxide shunt pathway.13 Unfortunately, the stability of these enzymes is generally very limited. There are chemical catalysts which can use hydrogen peroxide directly for oxidation reactions,14 but only a limited number of these catalysts are capable of enantioselective reactions even though this is a very active research field.15-20 Selenium compounds, both organic and inorganic, are well-established compounds for catalyzing oxidation reactions with various oxidants.21-23 They are still among the fastest catalysts known to date for selective epoxidations with HOOH. Ongoing research in our group has established aromatic selenium compounds as (pre)catalysts for epoxidations and Baeyer-Villiger oxidations with hydrogen peroxide as the terminal oxidant.24-26 Besides trifluoroethanol, also the use of various glycerol-based solvents have favorable properties as solvents for these

78 Design and synthesis of selenium catalysts for binding to enzymes catalyst systems.27 However, the synthesis and application of chiral arylseleninic catalysts for enantiospecific oxidation reactions is not straightforward, mostly due to the difficult synthesis procedures involved. Only one example in literature of enantioselective oxygen transfer with aromatic selenium catalysts has been reported.28 In recent years, several groups have been focusing on the use of enzymes as chiral scaffolds for enantioselective transformations, the so-called hybrid enzyme approach.29 Inspired by this technology, we wondered if we could utilize the well- defined active site residues of enzymes for creating chiral selenium catalysts. Here we will outline in more detail the approach followed in our case (figure 1).

Enantioselective enzyme Enantioselective (not redox-active) redox-enzyme

+ Redox-active catalyst (non-enantioselective) R O O 2 R2 R1 R Se 1 OH H2O2 H2O

Figure 1 General strategy of combining enzymes and chemical redox catalysts

Covalent versus non-covalent approach in hybrid catalysts Roughly there are two different methods for the design of hybrid enzymes with the attachment approach: covalent versus non-covalent attachment (figure 2). The non-covalent approach uses non-covalent (ionic) interactions of e.g. vanadate anions or manganese cations. In this way redox activity could be introduced into non-redox active enzymes. For examples of these we refer to various excellent reviews that have been published.29-31

79 Chapter 6

cat cat

Figure 2 Covalent (left) versus non-covalent (right) binding of catalysts in hybrid enzymes

Especially worth mentioning is the so-called Biotin-Avidin approach. This technology has been reasonably successful in hybrid catalysis. In this case the transition metal based catalyst is covalently linked to the biotin scaffold, with subsequent (non-covalent) complexation of biotin to avadin. This strategy has been successfully adopted in recent years by the group of Ward.32 The (strept)avidin acts as a chiral pocket in the reactions they perform. Due to the high binding constant of biotin with the enzyme, there is no non-bonded catalyst in solution and all of the redox-metal is embedded in the (strept)avidin polypeptide (figure 3). In a more direct approach it can be foreseen to directly and covalently link the redox active chemocatalyst to a specific position in or on the enzyme. In principle numerous ways of attaching compounds to the outer shell of a protein can be envisaged.33 Such an approach however is not likely to yield enantiospecific hybrid catalysts for two reasons: (1) the attached chemical catalysts would be significantly solvent-exposed with presumably little influence of the chiral protein; (2) due to the multiplicity of potential anchoring sites, the individual chemocatalysts would be exposed to different chiral environments, possibly neutralizing each other with respect to chiral induction. Hence, selective attachment of a chemocatalyst to one well-defined site within the protein shell is highly desirable. One obvious choice for this is to target the enzyme active site.34 Usually, the active-site architecture of a given enzyme (super)family is highly conserved thus enabling a wide range of proteins to be targeted with one type of suicide substrate. In this chapter our recent progress in developing a redox-active selenium catalyst fulfilling the requirements outlined before is described.

80 Design and synthesis of selenium catalysts for binding to enzymes

Figure 3 The inclusion of a Ru-transfer hydrogenation catalyst following the biotin-avidin approach. This image was reproduced from the crystallographic data (2QCB)

Choice of enzyme class: hydrolases Hydrolases are available in large quantities at reasonable prices, which is not surprising. Proteases and lipases are commonly used in the food industry for several purposes. Furthermore, the use of these types of enzymes in detergents is popular and widespread. Many of these enzymes have been optimized by biotech companies for their activity, selectivity and stability and they can sustain large pH ranges, endure high temperatures and work under oxidative conditions.35 Production of these biocatalysts is up to kiloton scales. Therefore, one can expect that the expression systems for these enzymes have already been optimized. Thus, one common pitfall en route to designing an enzyme engineering strategy for enantioselective hybrid catalysts is circumvented. Often, promising protein scaffolds turn out to be difficult to produce and purify to near-homogeneity thus complicating the sequence of diversity generation, selective covalent modification and screening.

81 Chapter 6

Probing the active site As explained before, we believe that the most promising approach for hybrid enzymes is to address the enzyme active site. Therefore one would need to exploit a class of compounds capable of irreversible binding to the active site. These compounds are generally better known as irreversible inhibitors, suicide substrates, or warheads. Depending on the active site amino acid which needs to be addressed by the warhead, the suicide substrate has a different design.

Serine hydrolases Several types of irreversible inhibitors for serine hydrolases have been reported. Coumarins are one example of these types of compounds. Even though the initial attack of the enzyme is believed to start from the active site serine36 the eventual inhibition of the enzyme is established through an electrophilic quinone methide. Another nucleophilic residue in or near the active site could attack this, whereby the enzyme active would be permanently blocked (figure 4).

O O R R X Y Y O O Nu OH Nu O O O O R Enzyme X Y Enzyme

Nu O O O OH O R R Nu Y Enzyme Nu Y O OH O O O

Enzyme OH Enzyme

Figure 4 Mechanistic pathways for serine hydrolase inhibition by coumarin compounds

Another class of serine hydrolase inhibitors are phosphonate esters bearing a good leaving group such as fluorine or p-nitrophenol. A schematic representation of the covalent attachment mechanism is depicted in figure 5. It is important to notice that this synthetic strategy is limited to phosphonate esters containing one good leaving group. Otherwise the enzyme active site can expel the phosphonate 82 Design and synthesis of selenium catalysts for binding to enzymes and restore the original activity by releasing the chemocatalyst into the reaction medium. This problem has already been encountered in the field of hybrid enzymes.37

O Enzyme OH R P L

R1 L = leaving group R = alkyl, O-alkyl or amid R1 = O-alkyl or amid

Inhibition O reactivation Enzyme OH R P OH O R1 P R Enzyme O 1 O R P OH aging Enzyme O R

Figure 5 Inhibition of serine hydrolases by organophosphorus compounds

The phosphonate esters have already been applied in synthesizing hybrid enzymes. Van Koten and co-workers established the synthesis of various phosphonate metallopincer complexes with different transition metals (see chapter 1). Their approach utilizes p-nitrophenyl phosphonate esters. Here, the p-nitrophenol group serves as an excellent leaving group (L, figure 5) and furthermore, the p- nitrophenolate released colors the reaction yellow and allows quantification of the reaction progress. Particularly, the authors described the covalent modification of cutinase with a range of pincer-complexes.38,39

Cysteine hydrolases The class of cysteine hydrolases contains (as already suggested by the name) a nucleophilic cysteine within the active site. There are several practical approaches to irreversibly inhibit cysteine hydrolases. As the majority of these strategies is based on covalent modification of the catalytically active cysteine, they also fulfill the requirements outlined previously for successful hybrid catalysts. Generally α-

83 Chapter 6 functionalized ketones comprising a good leaving group or a terminal epoxide are good irreversible inhibitors (see figure 6, top row and lower left).

O O O O O O R R 1 F S R N R R R N O acyloxymethylketone fluoromethylketone vinylsulfone imidazolide O O O R N R R S S O S O O N S R

epoxyketone thiosulfonate ester maleimide 2-pyridyldisulfide

Figure 6 Types of inhibitors for cysteine hydrolases

After attack by the cysteine on the ketone group, the enzyme-inhibitor complex collapses, expelling the activated leaving group or opening the epoxide ring. The inhibitor is then irreversibly bound to the enzyme. Michael acceptors, such as the maleimide group of compounds, are also good inhibitors in this respect. Upon attack by the enzyme, the complete compound stays bound. The final class of inhibitors for cysteine hydrolases are the activated (di)sulfide compounds, such as 2-pyridyl disulfide or thiosulfonate derivatives. In this case, the bond created is a disulfide (figure 6, lower row). These compounds are not always stable under strongly reducing or oxidizing conditions and are therefore not the most logical choice, even though they have been applied successfully for hybrid enzymes.40

Design of functionalized aromatic selenium compounds for enzyme inhibition

We envisioned selenium-protein hybrid catalysts for enantiospecific H2O2-driven oxidation reactions by combining the intrinsic chirality of a protein combined with the rich oxidation chemistry of aromatic selenium catalysts. In order to obtain stable and efficient catalysts, the strategy of design is divided in two parts: firstly

84 Design and synthesis of selenium catalysts for binding to enzymes the design of the selenium catalyst and secondly, the design and synthesis of aromatic selenium functionalized inhibitors (figure 7).

electron- withdrawing groups

EWG

Y SeR funtionalized Inhibitor selenium

Figure 7 General structure of functionalized aromatic selenium compounds

The introduction of selenium on the aromatic ring poses restrictions on the synthetic scheme. The most common procedures comprise of (1) the reaction of elemental selenium with Grignard reagents followed by aqueous work-up and aerobic oxidation of the selenol yielding the aromatic diselenide, (2) nucleophilic 41,42 substitution of aryl halides with Li2Se2 or Na2Se2, or (3) reacting an aryldiazonium compound with KSeCN.43 Using a functionalized diselenide results, upon oxidative activation of the pre-catalyst, in the homolytic cleavage of the former Se-Se bond. Thus, one equivalent of catalytically active selenide would be released in solution (figure 8), necessitating further purification.

R R

Inhibitor Se Se Inhibitor

oxidation

R R O O Inhibitor SeOH HOSe Inhibitor

Figure 8 Schematic drawing of the diselenide-inhibitor problem

85 Chapter 6

Therefore it would be a better option to temporarily protect the selenium. Using a selenocyanide could be an option, but deprotection would require harsh basic conditions (pH > 11) which might severely impair the structural integrity of the protein. The best option is to protect the selenide as a selenoether, which will eliminate upon introduction of the oxidant (figure 9).

R R R H O SeOH oxidant Se Se

Figure 9 Selenoxide elimination reaction for an aromatic butyl selenoether

The second part of the synthetic strategy concerns the anchoring part of the selenium compound (Y in figure 7). For serine hydrolases, the coumarin-route was excluded mainly due to two reasons: firstly, coumarin synthesis is not straightforward as many functionalities need to be introduced at a later stage of the synthetic sequence. Secondly, due to the size of the coumarin, the selenium catalyst would most likely be positioned outside the enzyme. This problem would be avoided when phosphonates would be applied as a short linker group between the protein scaffold and the catalyst. For cysteine hydrolases, a wider range of potential anchoring groups is accessible. It should be kept in mind that, for the reasons discussed above, larger linker groups do not seem advisable. From a synthetic point of view, activated α-halogen methylketones seem like a viable strategy as described previously by the group of Fraaije.44 Alternatively, maleimide- bearing selenium catalysts appear attractive, especially as various transition metal hybrid catalysts based on the maleimide strategy have been reported with papain. 45 In figure 10, the general structures of our target compounds are presented.

EWG EWG O EWG O O EtO P SeBu SeBu N SeBu L X O

Figure 10 General structures for the inhibitor arylseleninic pre-catalysts

86 Design and synthesis of selenium catalysts for binding to enzymes

Results and discussion Compatibility of the selenoether with oxidation catalysis To confirm at an early stage of synthesis that selenoethers indeed are suitable precatalysts we evaluated a range of bromo-functionalized aromatic selenoethers

(figure 11) towards their reactivity in the H2O2-driven epoxidation of alkenes. We reasoned that the bromide could provide us with a suitable point of attachment for functionalization in the later stage of synthesis. This would thus give us the option of attaching different types of inhibitors to the aromatic ring. Also, we were interested to see whether the bromide would have a large influence on the oxidation activity of the catalysts. This has been addressed before, and a substituent effect is to be expected.

NO2 O2N

Br SeBu Br SeBu Br SeBu

1 2 3 Br CF3 F3C

O2N SeBu Br SeBu Br SeBu

4 5 6

Figure 11 Bromo-functionalized aromatic butyl selenoethers

All compounds were premixed with hydrogen peroxide for one hour, in which time the mixture decolorized. Subsequently cyclohexene was added and conversion was followed in time (figure 12).26 Especially compounds 1, 2 and 5 showed significant activity and selectivity in the epoxidation of cyclohexene (table 1). It is interesting to note that compound 1 is the fastest catalyst, even though this was not expected. Most importantly, the selenoether does not seem to impair the catalytic oxidation, as long as the compounds are pre-mixed with oxidant before addition of the substrate. Interestingly, in some reactions (especially those with catalyst 4), a significant pressure build-up was observed during the course of the reaction and activation. This may be indicative of H2O2-disproportionating activity of the Se-catalysts, forming H2O and O2. In literature, the controversial “catalase- like” activity of aromatic selenium compounds is extensively discussed.46,47 87 Chapter 6

Figure 12 Oxidation of cyclohexene (2 mmol) with various butyl-protected aromatic selenides (1 mol%) with H2O2 (4 mmol 50%) in 2 mL trifluoroethanol at 20° C. Mass balance was in all reactions 100%. The dark graphs refer to conversion, the lighter graphs to yield of cyclohexene oxide.

Therefore, we performed control experiments by incubating only the catalyst and

H2O2 under otherwise identical conditions. Also the effect of added base was 26 studied. At intervals, samples were withdrawn and analyzed for the residual H2O2 concentration (figure 13). The catalysts show a minor catalase activity. After five hours, a 10% loss of H2O2 was observed which gradually increased overnight. This seems to be in line with the experiments conducted by Reich et al., who speculated that the intermediate selenoxide is responsible for this behavior, where certain types of aromatic selenium compounds do not eliminate the ether group readily.48 Overall it can be concluded that the selenoether can be a good protection group for the selenium, but it must be kept in mind that when the intermediate selenoxide does not eliminate fast enough, there could be problems with hydrogen peroxide decomposition.

88 Design and synthesis of selenium catalysts for binding to enzymes

)""#$

(!#$

('#$

(&#$

(%#$

("#$ -./01$ !!#$ 2/3$&$ 9;<=$%&$6>?=@A$ !'#$ 2/3$&$4536$7/89:$ -./01$(!#$ 2/3$&$B%#$ !&#$ 2/3$&$4536$7/89:$''#$ !"#$%&'()*'$%+,#')-%(-'(.$/0%() !%#$

!"#$ "$ "*+$ )$ )*+$ %$ %*+$ ,$ ,*+$ &$ &*+$ +$ 1,2')345)

Figure 13 Concentration of hydrogen peroxide upon exposure to selenium catalyst 4.

Conditions: 0.01 mmol 4, 4 µmol NaOAc, 3.6 mmol H2O2, 2 mmol trifluoroethanol, 20 °C

Towards synthesis of phosphonate esters There are several strategies for the synthesis of aromatic phosphonate esters as depicted in figure 14. Transition-metal catalyzed phosphorylations of aromatic compounds are widespread and established. Popular methods are NiCl2 or Pd catalyzed reactions with aryl iodides or triflates. In addition a microwave assisted coupling to aryl bromides has been published.49-52 More recently, Goossen et al. disclosed that the Pd(II) catalyzed reaction is capable of coupling aryl bromides and diethyl phosphite in high yield in alcoholic solvents at rather mild conditions.53 The second general strategy is to use an electrophilic handle on the aromatic compound which, after selective lithiation or magnesation, could be quenched with an electrophilic phosphonate. We anticipated that the harsh conditions in the

NiCl2 catalyzed reaction (route III) could be problematic for our selenium catalysts. The main problem in route V is the putative incompatibility of the acidic conditions necessary to form the diazonium salt with the sensitivity of the selenium moiety. Also the amount of steps is too large for an efficient synthetic route. Therefore we decided to follow synthetic route I first.

89 Chapter 6

O X P(OEt)3 MgX P Cl EtO NiCl , OEt X = Br or I 2 150-160 oC III IV 1. PCl OTf O 3 5 mol% Pd(PPh3)4 2. Mg P 1.2 eq Cs2CO3, THF EtO 3. EtOH Y o 4. m-CPBA O µW, 120 C, 10 min + Y = OEt or N2 P H II V EtO Y = N(R2)2 (route VI) OEt Pd(OAc)2 I VI O O Br P H P N(R2)2 EtO EtO OEt Cl Li / MgX

Figure 14 General routes for phosphorylation of aromatic compounds

We first tested the coupling of an aromatic selenocyanide to see if protection of the selenium as an ether should be performed beforehand. Thus, we attempted the Pd(II) catalyzed coupling of 4-bromophenyl selenocyanate 1a with diethylphosphite. However, instead of the desired coupling via the aromatic C-Br bond, we observed substitution of the cyanide functionality (figure 15), suggesting pseudo-halogenide reactivity of the cyanide group.

O Br SeCN P H EtO OEt 1a

Pd(OAc)2, PPh3, X TEA, EtOH O O P SeCN P Se Br EtO EtO OEt OEt 7 8

Figure 15 Attempted synthesis of a phosphonate modified aryl selenocyanate

90 Design and synthesis of selenium catalysts for binding to enzymes

Even though these types of compounds are highly interesting from a synthetic point of view,54 it is clear that the cyanide is not well-suited as a protecting group, which led us to protect the selenium with a n-butyl group. The introduction of the phosphonate ester mediated by Pd(II) following synthetic procedure I was attempted next. The reaction was successful and the product was isolated in reasonable yield (figure 16). In the next step, one of the two ethoxy groups should be replaced by a good leaving group. The synthetic strategy would be to react the compound with

SOCl2, which would yield the dichlorophosphonate. This could then be reacted with for instance an alcohol, followed by reacting it with a good leaving group. Upon attempting this synthetic option, we found that the reaction of compound 9 with SOCl2 did not yield the desired product.

O O Pd(OAc)2, PPh3, Br SeBu P H TEA, EtOH P SeBu EtO EtO OEt OEt 1 9 60%

Figure 16 Phosphorylation of compound 9 using Pd-catalyzed reaction

In fact, no product could be isolated and mass spectral analysis of the reaction mixture did not show any identifiable compounds. A literature search revealed that most probably the chlorination of the selenide occurred, but we think that in our case the product was not stable.55 Faced with these difficulties in our initially preferred synthetic scheme, we had to reconsider our strategy. First of all, the phosphonate should probably be partially protected, in order to introduce a good leaving group in a later stage of the synthesis. One option comprised a phosphonamidic chloride (route VI, figure 14), which could be introduced by reacting it with a Grignard reagent or a lithiate of the aromatic selenide 1, a well established strategy for the synthesis of enzyme inhibitors.56-59 To our demise, the magnesation of compound 1 was unsuccessful, as the bromoaryl was activated neither by metallic Mg, nor by other Grignard reagents such as iPrMgCl.60 The lithiation of the compound by n-BuLi, followed by addition of the electrophile was also unsuccessful. With one equivalent of n-BuLi added, 91 Chapter 6 there was no conversion of the substrate after addition of the electrophile. With two equivalents of n-BuLi added, there was conversion of the substrate, albeit forming a complex product mixture. Two of the identified compounds (by mass) are depicted in figure 17. The α-selenol alkyllithium is the most stable nucleophile in this case and this can react with any BuBr which has formed in the reaction.61,62 To our surprise, reacting the compound with tBuLi also failed to selectively lithiate the aromatic compound. At this stage, the synthetic procedure for acquiring Se-catalyst derivatized phosphonate compounds for inhibiting serine hydrolases became too laborious. Therefore we decided to attempt the synthesis of selective target reagents for cysteine hydrolases.

SeBu BuLi Br SeBu

1 Br Se

Figure 17 Outcome of lithiation of compound 1

Synthesis of Se-compounds for selective functionalization of cysteine hydrolases As discussed in the introduction, there are two reasonable possibilities for compounds capable of selective functionalization for cysteine hydrolases, α- functionalized ketones and Michael-acceptors. As we expected that the synthesis of, for instance, an α-bromoketone would give synthetic problems during the introduction of the bromide using elemental bromine,55 we decided not to evaluate this strategy but to immediately aim at maleimide-based anchoring groups at the aromatic ring of the Se-catalyst. Direct reduction of nitro groups on selenides is known.63 To our delight, the reduction of butyl-(4-nitrophenyl)selenide (10) with SnCl2 in EtOH yielded the target aniline in good yield. Next, introduction of the maleimide was successful by reacting the free amine with maleic anhydride, followed by ring closure with sodium acetate in acetic anhydride, providing the target compound in high yield

92 Design and synthesis of selenium catalysts for binding to enzymes

(figure 18). As has been established before (see chapter 1), the catalytic activity of the Se-catalysts is highest in the presence of electron-withdrawing substituents.

Therefore, we also synthesized an α–CF3-substituted, maleimide containing Se- catalyst. The commercially available starting compound 4-nitro-2-trifluoro- methylaniline was diazotized, followed by the introduction of the SeCN. Next, the butyl group was introduced using the Grieco method,64 followed by reduction of the nitro group. The free amine was subsequently reacted with maleic anhydride and after ring closure the target compound was obtained (figure 16).

1. HCl, NaNO2 R 2. KSeCN R 3. Bu3P, nBuOH O2N NH2 O2N SeBu

R = H 10 48% R = CF3 13 36%

SnCl2, EtOH HCl O R 1. maleic anhydride R 2. NaOAc, Ac2O N SeBu H2N SeBu

O R = H 12 93% R = H 11 77% R = CF 15 96% 3 R = CF3 14 92%

Figure 18 Synthesis of maleimide enhanced compounds 12 and 15

Conclusions The synthesis of functionalized aromatic selenium compounds which could be capable of binding selectively to enzymes is not an easy task. First, the selection of the type of enzyme is of vital importance, after which the right inhibitor should be selected. Phosphonate esters are the most promising compounds for the inhibition of serine hydrolases, as this would allow us to tap into the reservoir of available esterases, lipases and peptidases. These have been developed by many companies as food additives or simply as washing additives. Unfortunately, we were unable to successfully synthesize the phosphonate target compounds and thus we had to change our strategy. The synthesis of the maleimide enhanced aromatic selenides was successful, as all of the synthetic steps were in this case compatible with the 93 Chapter 6 selenide. The synthesis of these compounds is simple and straightforward, thus opening the way to the synthesis of Se-based hybrid catalysts as described in Chapter 7.

94 Design and synthesis of selenium catalysts for binding to enzymes

Experimental section 1H and 13C NMR spectra were recorded with a Bruker avance 400 (400 or 100 Mhz). Chemical shifts are denoted in ppm (δ) in deuterated solvents with TMS as an internal standard. GCMS spectra were recorded with a Shimadzu GCMS-QP2010S, equipped with an auto injector AOC-20I and a Four VF5ms column (0.25 µm, length 30 m and diameter 0,25 mm). GC analysis was performed on a Shimadzu GC-2014, equipped with an autoinjector AOC-20I, an autosampler AOC-20S with a CP Sil 5a-CB column (0.40 µm, length 35 m and diameter 0,25 mm). For catalytic experiments, a thermo twister comfort (Quantifoil Instruments) was used with custom made alumina sample holders. Melting points were determined with a Büchi Melting Point B-540. Preparative column chromatography was performed with Fluka silica gel 60 (0,2 – 0,5 mm). TLC analyses were performed with Uniplate silica gel G. Solvents used in column chromatography were of technical quality.

Bu3P and diethyl phosphite were freshly distilled under argon and stored under an argon atmosphere. 1-Butanol was also freshly distilled and stored under a nitrogen atmosphere over pre- activated molecular sieves (4 Å). All other reagents were purchased with the highest possible purity from chemical suppliers and used as received.

Catalytic experiments To a shaking solution (thermo shaker position 9) of the required amount of catalyst, NaOAc (if necessary) and 2 mL of solvent was added 4 mmol of H2O2 (50%, 0.272 g, 245 μL). After complete de-coloration of the reaction mixture was added 0.8 mmol of dibutyl ether (0.1042 g, 137 μL) as an internal standard and subsequently 2 mmol of cyclohexene (0.1643 g, 203 μL) was added. At several time intervals, 50 μL samples were withdrawn from the mixture and added to a suspension of ± 20 mg of Na2SO3 in 1 mL of EtOAc. The mixture was filtrated and analyzed by GC.

Iodometric titration for determination of H2O2 concentration To a shaking solution (thermo shaker position 9) of a mixture with or without catalyst 4 (0.01 mmol, 3.4 mg) in presence or absence of NaOAc in 2mL of trifluoroethanol was added 4 mmol of

H2O2 (50%, 0.272 g, 245 μL). At several time intervals, 50 μL samples were withdrawn from the mixture and added to an erlenmeyer flask which was wrapped with aluminum foil in which there was a mixture of 15 mL AcOH:CHCl3 (2:1) and 1.5 mL of a 2.4 M KI solution. This mixture was stirred for 10 minutes in the dark and titrated with a 0.02 M solution of Na2S2O3.

General procedure of diazotation

In a round bottom flask, 50 mmol of the aniline was added to 30 g of H2SO4 en 50 mL of H2O. Subsequently, 60 mmol of NaNO2 (30 mL of a 2M solution) was added drop-wise. After the complete addition, the temperature was lowered to 0 °C and the pH was slowly raised to 5-7 with a saturated solution of NaOAc. After this, 50 mmol KSeCN (50 mL of a 1M solution) was slowly added whilst the reaction was vigorously stirred with a mechanical stirrer. The product was extracted with

EtOAc. The organic layer was washed with brine and dried over anhydrous MgSO4.

General procedure for butylation of selenocyanides Under Schlenk conditions, 15 mmol of starting material and 19.5 mmol n-BuOH (1.8 mL) was dissolved in 75 mL THF and stirred under an argon atmosphere. To this mixture was drop-wise added 16,5 mmol Bu3P (4.1 mL). The reaction mixture was stirred until TLC analysis showed complete consumption of the starting material.

95 Chapter 6

1) H2SO4, NaNO2, R.T. NaOH, THF 2) KSeCN, 0 oC BuBr Br NH2 Br SeCN Br SeBu 1a Butyl (4-bromophenyl) selenide (1) The starting 4-bromoaniline 100 mmol (17.097 g) was dissolved in 60 mL concentrated HCl (instead of H2SO4) and the general procedure of diazotation was followed.The solvent was subsequently evaporated and the product was purified over a silica column (PE:EtOAc 99:1). The brownish oil was further purified via sublimation under low pressure yielding long white needles (18.88 g, 72.4 mmol,

72%) with a purity of 99% (GC). m/z (EI) 261 (with a typical Se-pattern). δH (400 MHz; CDCl3) 7.54

(d, 2H, J = 8.4 Hz), 7.50 (d, 2, J = 8.4 Hz); δC (100 MHz; CDCl3) 134.2, 133.5, 124.6, 119.8.

4-bromophenyl selenocyanate (1a) To 1 mmol of compound 1 which was dissolved in 6 mL of absolute EtOH was added 1.1 mmol NaOH (44 mg) and the mixture was heated to reflux until the starting material had disappeared on TLC. Then 1.1 mmol butyl bromide (1.1 mmol, 128 μL) was added and the mixture was refluxed overnight. The solvent was evaporated and the remaining slurry was taken up in EtOAc and washed twice with brine. The organic layer was dried over MgSO4 and the solvent was evaporated. The compound was purified over a silica column (PE only) yielding a yellow colored oil (266 mg, 0.91 mmol, 91%) with a purity of 99% (GC). m/z (EI) 292 (with a typical Se pattern). δH (400 MHz;

CDCl3) 7.31 (m, 4H), 2.87 (2H, t, J = 7.2 Hz), 1.63 (2H, p), 1.39 (2H, m), 0.89 (t, J = 7.6 Hz); δC (100 MHz; CDCl3) 133.2, 132.1, 131.9, 129.5, 120.7, 32.0, 27.7, 21.4, 13.5.

NO NO NO 2 1) H2SO4, NaNO2, R.T. 2 2 o 2) KSeCN, 0 C Bu3P, BuOH Br NH Br SeCN Br SeBu 2 (38%) THF, Argon 2a 2 4-bromo-2-nitrophenyl selenocyanate (2a) The general procedure for diazotation was followed with 50 mmol of the starting material. After work-up, the solvent was evaporated and the product was purified over a silica column (PE:EtOac = 100:0 to 96:4) which gave a orange-yellow compound (5.91 g, 19.3 mmol, 38%) with a purity of 99,9%

(GC). Melting point is 140-142 °C; m/z (EI) 306 (with a typical Se-pattern); δH (400 MHz; CDCl3)

8.58 (1H), 8.07 (1H, d, J = 8,8 Hz), 7.90 (1H, d, J = 8,8 Hz); δC (100 MHz; CDCl3) 139.0, 132.2, 129.3, 122.6, 103.8.

Butyl (4-bromo-2-nitrophenyl) selenide (2) The general procedure for butylation of selenocyanides was followed. The solvent was subsequently evaporated and the compound was immediately purified over a silica column (PE:EtOac = 100:0 to 98:2) which gave a yellow solid(2.98 g, 8.9 mmol, 59%) with a purity of 99.9% (GC). Melting point is

79-81 °C; m/z (EI) 337 (with a typical Se-pattern); δH (400 MHz; CDCl3 and MeOD) 8.12 (1H,s), 7.38 (1H,m), 7.18 (1H, m), 2.68 (2H, m), 1.48 (2H, m), 1.23 (2H, m), 0.68 (3H, m); δC (100 MHz;

CDCl3, MeOD) 136.0, 130.2, 128.4, 29.8, 25.5, 22.6, 12.7.

96 Design and synthesis of selenium catalysts for binding to enzymes

O2N O N 18-crown-6, Bu Se O2N MeOH, HBF4, 2 2 2 NaNO KOAc, CHCl3 2 Br SeBu Br NH2 Br N2BF4 (44%) (10%) 3a 3b 3 4-bromo-3-nitroaniline (3a)

In an erlenmeyer flask, 225 mmol 4-bromoaniline (38.71 g) was added to 340 mL H2SO4 and to this mixture, 225 mmol guanidinium nitrate (CH5N3•HNO3) (27.50 g) was added and the mixture was stirred until all solids were dissolved. The mixture was then poured in 2,2 L of 20% NaOH solution which was cooled to 0 °C. Compound 3a was filtered from the mixture, washed with water and dried in vacuo. This gave a dark brown solid (38.0 g, 175 mmol, 78%) with a purity of 99.9% (GC). m/z (EI) 216; δH (400 MHz; CDCl3) 7.31 (1H, d, J = 9.2 Hz), 7.19 (1H, s) 6.78 (1H, d, J = 9.2 Hz), 4.93

(2H, bs); δC (100 MHz; CDCl3) 144.0, 129.3, 120.7 (C3).

4-bromo-3-nitrophenyldiazonium boron tetrafluoride (3b) In a round bottom flask 10 mmol of 3a (2.15 g) was dissolved in 30 mL methanol. To this solution was added 20 mmol HBF4 (2,5 mL of a 8 M solution) drop-wise. The mixture was cooled to 0 °C and 12 mmol NaNO2 (6 mL of a 2M solution) was slowly added. Methanol was removed by rotary evaporation and the remaining solid was filtered and washed with cold methanol. The crystalline compound 3b was dried in vacuo which yielded a dark brown powder (1.39 g, 4.39 mmol, 44%). Melting point is 150-152 °C. δH (400 MHz; CDCl3) 7,83 (1H, s), 7.68 (1H, d, J = 8.4 Hz), 7.44 (1H, d, J = 9.2 Hz).

Butyl (4-bromo-3-nitrophenyl) selenide (3) In a round bottom flask 2 mmol of 3b (0.627 g) was added to a solution of 0.2 mmol 18-crown-6 (10 mol%, 0.056 g) and 2.25 mmol Bu2Se2 (0.616 g) in 6.5 mL CHCl3. The temperature was lowered to 0 °C and in a time period of 10 minutes 4 mmol of KOAc (0.4 g) was added to the mixture which was then allowed to stir for 4 hours. The product was purified over a silica column (PE:EtOac = 100:0 to 96:4) gave an orange-red liquid (0.07 g, 0.2 mmol, 10%). m/z (EI) 337 (with a typical Se pattern); δH

(400 MHz; CDCl3) 7.79 (1H, s), 7.49 (1H,d, J = 8.0 Hz), 7.41 (1H, d, J = 8.4 Hz), 2.91 (2H, t, J = 7.6

Hz), 1.64 (2H, m), 1.38 (2H, m), 0,86 (3H, t, J = 7.2 Hz); δC (100 MHz; CDCl3) 136.0, 135.1, 127.8, 31.8, 20.0, 22.9, 13.5.

Br Br 1) H2SO4, NaNO2, R.T. Br o 2) KSeCN, 0 C Bu3P, BuOH O2N NH2 O2N SeCN THF, Argon O2N SeBu 4a 4 2-bromo-4-nitrophenyl selenocyanate (4a) The general procedure for diazotation was followed with 50 mmol of the starting material. After work-up, the solvent was evaporated and the product was purified over a silica column (PE:EtOac = 100:0 to 96:4) which gave a light yellow compound (8.38 g, 27.4 mmol, 54%). Melting point is 127-129

°C. m/z (EI) 306 (with a typical Se pattern); δH (400 MHz; CDCl3) 8.35 (1H, s), 8.18 (1H, d, J = 8.8 Hz), 7.94 (1H, d, J = 8.8 Hz); δC (100 MHz; CDCl3) 130.4, 127.8, 124.0.

Butyl (2-bromo-4-nitrophenyl) selenide (4) The general procedure for butylation of selenocyanides was followed with 20 mmol of the starting 97 Chapter 6 compound. The solvent was subsequently evaporated and the compound was immediately purified over a silica column (PE:EtOac = 100:0 to 98:1) which gave a yellow solid(3.32 g, 9.9 mmol, 49%) with a purity of 99.9% (GC). Melting point is 44-46 °C. m/z (EI) 337 (with a typical Se pattern); δH

(400 MHz; CDCl3) 8.27 (1H, s), 8.00 (1H, d, J = 8.8 Hz), 7.27 (1H, d, J = 8.8 Hz), 2.96 (2H, t, J = 7.2

Hz), 1.74 (2H, m), 1.46 (2H, m), 0.91 (3H, t, J = 7.2 Hz); δC (100 MHz; CDCl3) 146.5, 145.6, 128.0, 127.4, 123.4, 122.2, 30.7, 27.3, 23.2, 13.6.

CF3 CF3 CF 1) H2SO4, NaNO2, R.T. 3 2) KSeCN, 0 oC Bu3P, BuOH Br NH Br SeCN Br SeBu 2 THF, Argon 5a 5 4-bromo-2-trifluoromethylphenyl selenocyanate (5a) The general procedure for diazotation was followed with 50 mmol of the starting material. After work-up, the solvent was evaporated and the product was purified over a silica column (PE only) which gave a orange-brown liquid (5.91 g, 19.3 mmol, 38%) with a purity of 94% (GC). m/z (EI) 329

(with a typical Se Pattern); δH (400 MHz; CDCl3) 7.86 (1H, s), 7.81 (1H, d, J = 8.0 Hz), 7.73 (1H, d, J =

8.0 Hz); δC (100 MHz; CDCl3) 136.7, 135.4, 131.0, 123.7.

Butyl (4-bromo-2-trifluoromethylphenyl) selenide (5) The general procedure for butylation of selenocyanides was followed with 2 mmol of the starting compound. The solvent was subsequently evaporated and the compound was immediately purified over a silica column (PE only) which gave a yellow liquid (3.32 g, 9.9 mmol, 49%) with a purity of

99.9% (GC). m/z (EI) 360 (with a typical Se pattern); δH (400 MHz; CDCl3) 7.71 (1H, s), 7.44 (1H, d, J = 8.4 Hz), 7.39 (1H, d, J = 8.4 Hz), 2.88 (2H, t, J = 7.2 Hz), 1.62 (2H, p), 1.36 (2H, h), 0.87 (3H, t); δC (100 MHz; CDCl3) 134.8 (C1), 129.9, 129.8, 31.5, 28.1, 22.9, 13.5.

F3C 1) H2SO4, NaNO2, R.T. F3C F3C o 2) KSeCN, 0 C Bu3P, BuOH Br NH2 Br SeCN THF, Argon Br SeBu 6a 6 4-bromo-3-trifluoromethylphenyl selenocyanate (6a) The general procedure for diazotation was followed with 50 mmol of the starting material. After work-up, the solvent was evaporated and the product was purified over a silica column (PE: EtOAc = 100:0 to 98:2) which gave a light yellow compound (7.59 g, 23.1 mmol, 46%) with a purity of 99.9%

(GC). Melting point is 36-38 °C. m/z (EI) 329 (with a typical Se pattern); δH (400 MHz; CDCl3) 7.94

(1H, s), 7.78 (1H, d, J = 8.4 Hz), 7.71 (1H, d, J = 8.4 Hz); δC (100 MHz; CDCl3) 136.9, 136.8, 123.3, 122.4, 121.4, 120.6.

Butyl (4-bromo-3-trifluoromethylphenyl) selenide (6) Under Schlenk conditions, 20 mmol of compound 6a (6.57 g) and 26 mmol n-BuOH (0.24 mL) was dissolved in 100 mL THF and stirred under an argon atmosphere. To this mixture was drop-wise added 22 mmol Bu3P (5.5 mL). The reaction mixture was stirred until TLC analysis showed complete consumption of the starting material. The solvent was subsequently evaporated and the compound was immediately purified over a silica column (PE only) which gave a yellow-orange solid liquid (0.72 g, 1.99 mmol, 10%) with a purity of 99.9% (GC). m/z (EI) 360 (with a typical Se pattern); Melting point

98 Design and synthesis of selenium catalysts for binding to enzymes is 57-59 °C. m/z (EI) 360. δH (400 MHz; CDCl3) 7.90 (1H, s), 7.60 (2H, m), 2.89 (m, 2H), 1.60 (2H, m), 1.35 (2H, m), 0.81 (3H, m)

4-bromophenyl diethyl selenophosphate (8)

A 10 mL round bottom flask was charged (under argon) with 0.05 mmol of Pd(OAc)2 (11 mg) and 0.15 mmol PPh3 (40 mg). Next, 5 mL deoxygenated EtOH was added, followed by 0.5 mmol of compound 1a (130 mg), 0.75 mmol triethylamine (54 μL) and 0.6 mmol of diethyl phosphite (77 μL) and the mixture was heated to reflux. When TLC analysis showed complete conversion of the starting material, the solution was cooled to room temperature, diluted with EtOAc and subsequently washed with 1 M HCl, saturated NaHCO3 and brine. The organic layer was dried over MgSO4 and the residue was purified by column chromatography yielding a yellow liquid (131 mg, 0.35 mmol, 70%). m/z (EI) 372 (with a typical Se-pattern). δH (400 MHz; CDCl3) 7.50 (d, 2H, J = 8.4 Hz), 7.43 (2H, d, J = 8.4 Hz), 4.18 (4H, dq), 1.31 (6H, t, J = 7.2 Hz).

4-butylphenylselenide diethylphosphate (9)

A 10 mL round bottom flask was charged (under argon) with 0.10 mmol of Pd(OAc)2 (22 mg) and 0.30 mmol PPh3 (80 mg). Next, 5 mL deoxygenated EtOH was added, followed by 0.91 mmol of compound 1a (292 mg), 1.35 mmol triethylamine (98 μL) and 1.1 mmol of diethylphosphite (141 μL) and the mixture was heated to reflux. After the addition of compound 1a, the mixture became very dark, but this color disappeared after 30 minutes. After overnight reflux, the mixture was dark again and TLC analysis showed complete conversion of the starting material. The solution was cooled to room temperature, diluted with EtOAc and subsequently washed with 1 M HCl, saturated NaHCO3 and brine. The organic layer was dried over MgSO4 and the residue was purified by column chromatography yielding a yellow liquid (98 mg, 0.26 mmol, 28 %). m/z (EI) 350 (with a typical Se pattern). δH (400 MHz; CDCl3) 7.66 (2H, m), 7.50 (2H, m), 4.11 (2H, dq), 2.98 (2H, t, J = 7.6 Hz), 1.72 (2H, p), 1.44 (2H, m), 1.32 (6H, t, J = 6.8 Hz), 0.92 (3H, t, J = 7.6 Hz).

HCl, NaNO2 Bu3P, BuOH O N NH KSeCN O N SeCN O N SeBu 2 2 2 THF, pyridine 2 10a 10

4-nitrophenyl selenocyanate (10a)

The starting aniline 40 mmol (5.552 g) was dissolved in 30 mL concentrated HCl (instead of H2SO4) and the general procedure of diazotation was followed.The solvent was subsequently evaporated and the product was purified via bulb-to-bulb destillation (200 °C, 0.2 mbar) which yielded (after cooling down) a brown solid (5.603 g, 24.7 mmol, 62%) with a purity of 99% (GC). m/z (EI) 228 (with a typical Se-pattern). δH (400 MHz; CDCl3) 8.26 (2H, d, J = 8.8 Hz), 7.79 (2H, d, J = 8.8 Hz).

Butyl (4-nitrophenyl) selenide (10) A solution of 5 mmol of compound 10a (1.135 g) in 10 mL THF, 10 mL pyridine and 6 mmol nBuOH

(552 μL) was treated drop-wise with 6 mmol Bu3P (1.50 mL) and this was stirred at room temperature for three days. The reaction mixture was diluted with EtOAc and washed two times with 1M of HCl and water. The organic layers was dried over Na2SO4, the solids were filtered off and the solvent was evaporated. The compound was purified over a silica column (PE:EtOAc = 99:1) which gave a yellow liquid (1.005 g, 3.91 mmol, 78%) with a purity of 99.9% (GC). m/z (EI) 258 (with

99 Chapter 6 a typical Se pattern). δH (400 MHz; CDCl3) 8.07 (2H, d, J = 8.8 Hz), 7.50 (2H, d, J = 8.8 Hz), 3.04

(2H, t, J = 7.2 Hz), 1.75 (2H, p), 1.47 (2H, m), 0.95 (3H, t, J = 7.6 Hz); δC (100 MHz; CDCl3) 142.6, 130.5, 129.3, 124.1, 123.4, 31.7, 26.8, 23.0, 13.1.

O SnCl2, EtOH HCl Bu3P, BuOH O2N SeBu H N SeBu N SeBu 2 THF, pyridine 10 11 12 O

Butyl (4-aminophenyl) selenide (11) Compound 10 (5.88 mmol, 1.511g) was dissolved in a mixture of 30 mL EtOH and 4 mL of concentrated HCl. To this mixture was added 18 mmol of SnCl2 and the mixture was heated to 50 °C. After completion of the reaction (4 hours), the mixture was cooled to room temperature and brought to neutral pH with 5 M NaOH. The solids were filtered off and the liquids were extracted with EtOAc. The organic layer was washed with brine, dried with Na2SO4 and the solvent was evaporated. The product was purified via column chromatography (PE:EtOAc = 95:5) and this gave the product as a a dark red liquid (640 mg, 2.8 mmol, 48%) which was used directly in the next reaction. m/z (EI) 229 (with a typical Se pattern).

1-(4-butylphenylselenide)-1H-pyrrole-2,5-dione (12)

Compound 11 (600 mg, 2.62 mmol) was dissolved in 20 mL Et2O and maleic anhydride (2.65 mmol, 260 mg) was added. The mixture was stirred for 1 hour and the solvent was evaporated. Acetic anhydride 10 mL and 100 mg NaOAc were added to the solids and the mixture was heated to reflux for 1 hour. The mixture was cooled to room temperature and poured on crushed ice. This was extracted three times with Et2O. The combined organic layers were washed with saturated NaHCO3 and water, dried with MgSO4 and the solvent was evaporated. The product was purified via column chromatography (PE:EtOAc = 95:5) and this gave the product as a a light yellow solid (733 mg, 2.37 mmol, 90%). m/z (EI) 308 (with a typical Se pattern). δH (400 MHz; CDCl3) 7.55 (2H, d, J = 8.0 Hz), 7.23 (2H, d, J = 8.0 Hz) 6.83 (2H, s), 2.93 (2H, t, J = 7.2 Hz), 1.70 (2H, m), 1.44 (2H, m), 0.89 (3H, t);

δC (100 MHz; CDCl3) 169.3, 135.2, 135.0, 134.7, 134.0, 133.8, 133.5, 132.3, 126.7, 125.9, 32.1, 29.7, 27.7, 22.9.

CF3 CF3 CF3 HCl, NaNO2 Bu3P, BuOH O N NH KSeCN O N SeCN O N SeBu 2 2 2 THF, pyridine 2 13a 13

2-trifluoromethyl-4-nitrophenyl selenocyanate (13a) The starting aniline 100 mmol (20.612 g) was dissolved in 75 mL concentrated HCl (instead of

H2SO4) and the general procedure of diazotation was followed.The solvent was subsequently evaporated and the product was purified via bulb-to-bulb distillation (230 °C, 0.2 mbar) which yielded (after cooling down) a red crystalline solid (20.863 g, 70.7 mmol, 71%) with a purity of 99%

(GC). m/z (EI) 295 (with a typical Se-pattern). δH (400 MHz; CDCl3) 8.59 (1H, s), 8.44 (1H, d, J = 8.4 Hz), 8.22 (1H, d, J = 8.8 Hz).

100 Design and synthesis of selenium catalysts for binding to enzymes

Butyl (2-trifluoromethyl-4-nitrophenyl) selenide (13) The general procedure for butylation was followed with 10 mmol of compound 13a (2.95 g) and 10 mL pyridine was added extra to the mixture. After work-up, the mixture was filtered through a silica plug (flushed with PE:EtOAc = 90:10) and the fraction was subjected to bulb-to-bulb distillation (180 °C, 0.1 mbar) which gave a light yellow oil. The compound was finally purified over a silica column (PE:EtOAc = 100:0 to 98:2) which gave a yellow liquid (1.652g, 5.05 mmol, 51%) with a purity of

99.9% (GC). m/z (EI) 326 (with a typical Se pattern). δH (400 MHz; CDCl3) 8.46 (1H, s), 8.23 (1H, d, J = 8.8 Hz), 7.64 (1H, d, J = 8.4 Hz), 3.10 (2H, t, J = 7.2 Hz), 1.78 (2H, m), 1.51 (2H, m), 0.96 (3H, t); δC

(100 MHz; CDCl3)131.5, 130.2, 126.2, 125.0, 122.2, 118.7, 30.9, 27.5, 23.0, 13.1.

CF 3 CF3 O CF3 SnCl2, EtOH HCl Bu3P, BuOH O2N SeBu H N SeBu N SeBu 2 THF, pyridine 13 14 15 O

Butyl (4-amino-2-trifuoromethylphenyl) selenide (14) Compound 13 (2 mmol, 654 mg) was dissolved in a mixture of 10 mL EtOH and 1.3 mL of concentrated HCl. To this mixture was added 6 mmol of SnCl2 and the mixture was heated to 50 °C. After completion of the reaction (3 hours), the mixture was cooled to room temperature and brought to neutral pH with 5 M NaOH. The solids were filtered off and the liquids were extracted with EtOAc. The organic layer was washed with brine, dried with Na2SO4 and the solvent was evaporated. The product was purified via column chromatography (PE:EtOAc = 99:1) and this gave the product as a a dark red oil (497 mg, 1.85 mmol, 92%). m/z (EI) 296 (with a typical Se pattern). δH

(400 MHz; CDCl3) 7.47 (1H, d, J = 8.4 Hz), 6.97 (1H, s), 6.70 (1H, d, J = 6.4 Hz), 3.76 (2H, bs), 2.82 (2H, t, J = 7.2 Hz), 1.61 (2H, m), 1.40 (2H, m), 0.88 (3H, t, J = 7.2 Hz)

1-(4-butyl-2-trifluoromethylphenylselenide)-1H-pyrrole-2,5-dione (15)

Compound 14 (498 mg, 1.85 mmol) was dissolved in 10 mL Et2O and maleic anhydride (1.85 mmol, 181 mg) was added. The mixture was stirred for 1 hour and the solvent was evaporated. Acetic anhydride 5 mL and 100 mg NaOAc were added to the solids and the mixture was heated to reflux for 1 hour. The mixture was cooled to room temperature and poured on crushed ice. This was extracted three times with EtOAc. The combined organic layers were washed with saturated

NaHCO3 and water, dried with MgSO4 and the solvent was evaporated. The product was purified via column chromatography (PE:EtOAc = 90:10 to 70:30) and this gave the product as a a light yellow solid (667 mg, 1.78 mmol, 96%). m/z (EI) 376 (with a typical Se pattern). δH (400 MHz; CDCl3) 7.66 (2H, m), 7.43, (1H, d, J = 8.4 Hz), 2.99 (2H, t, J = 7.6 Hz), 1.71 (2H, m), 1.46 (2H, m), 0.92 (3H, t, J =7.6

Hz); δC (100 MHz; CDCl3) 168.9, 135.3, 135.1, 134.8, 134.0, 31.5, 27.9, 22.9, 13.8.

References

(1) Poliakoff, M.; Fitzpatrick, J. M.; Farren, T. R.; Anastas, P. T. Science 2002, 297, 807-810. (2) Sheldon, R. A. Pure Appl. Chem. 2000, 72, 1233-1246. (3) Powell, K. A.; Ramer, S. W.; del Cardayre, S. B.; Stemmer, W. P. C.; Tobin, M. B.; Longchamp, P. F.; Huisman, G. W. Angew. Chem.-Int. Edit. 2001, 40, 3948-3959. (4) Adamczak, M.; Krishna, S. H. Food Technol. Biotechnol. 2004, 42, 251-264. (5) Arnold, F. H. Nature 2001, 409, 253-257. 101 Chapter 6

(6) Davis, B. G.; Borer, V. Nat. Prod. Rep. 2001, 18, 618-640. (7) Tao, J. H.; Zhao, L. S.; Ran, N. Q. Org. Process Res. Dev. 2007, 11, 259-267. (8) Leresche, J. E.; Meyer, H. P. Org. Process Res. Dev. 2006, 10, 572-580. (9) Hollmann, F.; Arends, I. W. C. E.; Buehler, K.; Schallmey, A.; Buhler, B. Green Chem. 2010, 13, 226-265. (10) Hollmann, F.; Hofstetter, K.; Schmid, A. Trends Biotechnol. 2006, 24, 163-171. (11) Zhao, H.; van der Donk, W. A. Curr. Opin. Biotechnol. 2003, 14, 583-589. (12) Hollmann, F.; Arends, I.; Buehler, K. Chemcatchem 2010, 2, 762-782. (13) van Rantwijk, F.; Sheldon, R. A. Curr. Opin. Biotechnol. 2000, 11, 554-564. (14) Lane, B. S.; Burgess, K. Chem. Rev. 2003, 103, 2457-2473. (15) Colladon, M.; Scarso, A.; Sgarbossa, P.; Michelin, R. A.; Strukul, G. J. Am. Chem. Soc. 2006, 128, 14006-14007. (16) Sawada, Y.; Matsumoto, K.; Kondo, S.; Watanabe, H.; Ozawa, T.; Suzuki, K.; Saito, B.; Katsuki, T. Angew. Chem.-Int. Edit. 2006, 45, 3478-3480. (17) Sawada, Y.; Matsumoto, K.; Katsuki, T. Angew. Chem.-Int. Edit. 2007, 46, 4559-4561. (18) Gelalcha, F. G.; Bitterlich, B.; Anilkumar, G.; Tse, M. K.; Beller, M. Angew. Chem.-Int. Edit. 2007, 46, 7293-7296. (19) Shitama, H.; Katsuki, T. Chem.-Eur. J. 2007, 13, 4849-4858. (20) Yamaguchi, T.; Matsumoto, K.; Saito, B.; Katsuki, T. Angew. Chem.-Int. Edit. 2007, 46, 4729-4731. (21) Paulmier, C. Selenium reagents and intermediates in organic synthesis; Pergamon Press: Oxford, 1986. (22) Nishibayashi, Y.; Uemura, S. In Organoselenium Chemistry 2000; Vol. 208, p 201-233. (23) Giurg, M.; Syper, L. Phosphorus Sulfur Silicon Relat. Elem. 2008, 183, 970-985. (24) ten Brink, G. J.; Vis, J. M.; Arends, I. W. C. E.; Sheldon, R. A. Tetrahedron 2002, 58, 3977-3983. (25) ten Brink, G. J.; Vis, J. M.; Arends, I. W. C. E.; Sheldon, R. A. J. Org. Chem. 2001, 66, 2429-2433. (26) ten Brink, G. J.; Fernandes, B. C. M.; van Vliet, M. C. A.; Arends, I. W. C. E.; Sheldon, R. A. J. Chem. Soc.-Perkin Trans. 1 2001, 224-228. (27) Garcia-Marin, H.; van der Toorn, J. C.; Mayoral, J. A.; Garcia, J. I.; Arends, I. Green Chem. 2009, 11, 1605-1609. (28) Miyake, Y.; Nishibayashi, Y.; Uemura, S. Bull. Chem. Soc. Jpn. 2002, 75, 2233-2237. (29) Thomas, C. M.; Ward, T. R. Chem. Soc. Rev. 2005, 34, 337-346. (30) Kohler, V.; Wilson, Y. M.; Lo, C.; Sardo, A.; Ward, T. R. Curr. Opin. Biotechnol. 2010, 21, 744-752. (31) Reetz, M. T. Tetrahedron 2002, 58, 6595-6602. (32) Letondor, C.; Ward, T. R. ChemBioChem 2006, 7, 1845-1852. (33) Davis, B. G. Curr. Opin. Biotechnol. 2003, 14, 379-386. (34) Qi, D. F.; Tann, C. M.; Haring, D.; Distefano, M. D. Chem. Rev. 2001, 101, 3081-3111. (35) Saeki, K.; Ozaki, K.; Kobayashi, T.; Ito, S. J. Biosci. Bioeng. 2007, 103, 501-508. (36) Pochet, L.; Dieu, M.; Frédérick, R.; Murray, A.-M.; Kempen, I.; Pirotte, B.; Masereel, B. Tetrahedron 2003, 59, 4557-4561. (37) Reetz, M. T.; Rentzsch, M.; Pletsch, A.; Maywald, M. Chimia 2002, 56, 721-723. (38) Kruithof, C. A.; Casado, M. A.; Guillena, G.; Egmond, M. R.; van der Kerk-van Hoof, A.; Heck, A. J. R.; Klein Gebbink, R. J. M.; van Koten, G. Chem.-Eur. J. 2005, 11, 6869-6877.

102 Design and synthesis of selenium catalysts for binding to enzymes

(39) Kruithof, C. A.; Dijkstra, H. P.; Lutz, M.; Spek, A. L.; Egmond, M. R.; Klein Gebbink, R. J. M.; van Koten, G. Eur. J. Inorg. Chem. 2008, 4425-4432. (40) Kuang, H.; Brown, M. L.; Davies, R. R.; Young, E. C.; Distefano, M. D. J. Am. Chem. Soc. 1996, 118, 10702-10706. (41) Syper, L.; Mlochowski, J. Tetrahedron 1988, 44, 6119-6130. (42) Deryagina, E. N.; Russavskaya, N. V.; Papernaya, L. K.; Levanova, E. A.; Sukhomazova, E. N.; Korchevin, N. A. Russ. Chem. Bull. 2005, 54, 2473-2483. (43) Sharpless, K. B.; Young, M. W. J. Org. Chem. 1975, 40, 947-949. (44) Panella, L.; Broos, J.; Jin, J. F.; Fraaije, M. W.; Janssen, D. B.; Jeronimus-Stratingh, M.; Feringa, B. L.; Minnaard, A. J.; de Vries, J. G. Chem. Commun. 2005, 5656-5658. (45) Haquette, P.; Salmain, M.; Svedlung, K.; Martel, A.; Rudolf, B.; Zakrzewski, J.; Cordier, S.; Roisnel, T.; Fosse, C.; Jaouen, G. ChemBioChem 2007, 8, 224-231. (46) Reich, H. J.; Renga, J. M.; Reich, I. L. J. Am. Chem. Soc. 1975, 97, 5434-5447. (47) Umbreit, M. A.; Sharpless, K. B. J. Am. Chem. Soc. 1977, 99, 5526-5528. (48) Reich, H. J.; Wollowitz, S.; Trend, J. E.; Chow, F.; Wendelborn, D. F. J. Org. Chem. 1978, 43, 1697-1705. (49) Tavs, P. Chem. Ber.-Recl. 1970, 103, 2428-&. (50) Kohler, M. C.; Stockland, R. A.; Rath, N. P. Organometallics 2006, 25, 5746-5756. (51) Gasparini, F.; Inderbitzin, W.; Francotte, E.; Lecis, G.; Richert, P.; Dragic, Z.; Kuhn, R.; Flor, P. J. Bioorg. Med. Chem. Lett. 2000, 10, 1241-1244. (52) Kalek, M.; Ziadi, A.; Stawnski, J. Org. Lett. 2008, 10, 4637-4640. (53) Goossen, L. J.; Dezfuli, M. K. Synlett 2005, 445-448. (54) Han, L. B.; Choi, N.; Tanaka, M. J. Am. Chem. Soc. 1996, 118, 7000-7001. (55) Ternon, M.; Outurquin, F.; Paulmier, C. Tetrahedron 2001, 57, 10259-10270. (56) Green, M.; Hudson, R. F. J. Chem. Soc. 1958, 3129-3133. (57) Burger, A.; Dawson, N. D. J. Org. Chem. 1951, 16, 1250-1254. (58) Harger, M. J. P.; Hurman, B. T. J. Chem. Soc.-Perkin Trans. 1 1998, 1383-1388. (59) Rumthao, S.; Lee, O.; Sheng, Q.; Fu, W.; Mulhearn, D. C.; Crich, D.; Mesecar, A. D.; Johnson, M. E. Bioorg. Med. Chem. Lett. 2004, 14, 5165-5170. (60) Krasovskiy, A.; Straub, B. F.; Knochel, P. Angew. Chem.-Int. Edit. 2006, 45, 159-162. (61) Ponthieux, S.; Paulmier, C. In Organoselenium Chemistry 2000; Vol. 208, p 113-142. (62) Huang, X. A.; Xu, W. M. Tetrahedron Lett. 2002, 43, 5495-5497. (63) Carland, M. W.; Martin, R. L.; Schiesser, C. H. Org. Biomol. Chem. 2004, 2, 2612-2618. (64) Grieco, P. A.; Gilman, S.; Nishizawa, M. J. Org. Chem. 1976, 41, 1485-1486.

103

Chapter 7: Enzymes as scaffolds for aromatic selenium catalysts

John C. van der Toorn, Frank Hollmann, Roger A. Sheldon and Isabel W.C.E Arends

Abstract Hybrids of maleimide functionalized phenyl butyl selenides were synthesized using papain as the protein scaffold. Inhibition of the protolytic activity of the enzyme could be observed upon incubation with the selenide precursor. MALDI-TOF was used to analyze the selenium-papain hybrids. These hybrids were envisaged to act as precursors for seleninic acid catalyzed oxidations with hydrogen peroxide. A variety of oxidation reactions were conducted, however no catalytic activity of the selenium hybrid could be observed in aqueous solution. Chapter 7

Introduction In the field of oxidation catalysis using hydrogen peroxide as the oxidant there is a drive to develop enzyme mimics that are more stable and active than the peroxidase enzymes themselves. Seleninic acid derived catalysts are among the fastest catalysts to date for epoxidations with hydrogen peroxide in specific organic solvents (see chapter 1 and chapter 5). Therefore the use of selenium protein hybrids as precatalysts for oxidation reactions is expected to have a considerable potential. The protein cavity can induce chirality and can possibly stabilize the selenium moiety. Even though many redox-active groups have been attached in and to enzymes, there are very few systems which really are capable of joining “the best of both worlds”.1 Most of the hybrid systems do not show any activity, or at best a significantly reduced activity compared to the original system. In some cases the substrate preference is highly influenced, like in the Rh-catalyzed hydroformylation reaction with human serum albumin.2,3 In other cases, the enzyme provides selectivity in the hybrid system, but the reaction is not catalytic.4 There are some successful examples of hybrid enzymes, and one of those contain a selenium as prosthetic group.

Selenosubtilisin Selenosubtilisin (SeSub) was originally reported by Hilvert and co-workers.5 The synthesis of SeSub is rather straightforward. Subtilisin is inhibited quite rapidly by phenylmethanesulfonylfluoride, giving the intermediate sulfonate ester. A rather simple substitution reaction with NaHSe yields the SeSub (figure 1).

O PMSF PMSF Subtilisin OH Subtilisin O S Subtilisin SeH O

Figure 1 Synthesis of selenosubtilisin

Initially SeSub was designed to transform a protease into a more efficient acyl transferase as it was a poor catalyst for amide hydrolysis. The selenol ester, which is formed after attack of SeSub on activated esters, undergoes aminolysis faster

106 Enzymes as scaffolds for aromatic selenium catalysts than hydrolysis explaining this acyl transferase activity. Enzymes which bear a selenocysteine residue are very often redox-active enzymes and the classical example of a selenocysteine-bearing enzyme is Glutathione Peroxidase.6 SeSub was indeed capable of reducing hydroperoxides in the presence of sulfides (figure

2). Compared to Ph2Se2, the efficiency is notably accelerated in the reduction of TBHP with 3-carboxy-nitrobenzenethiol as the terminal reductant: 1 µM SeSub had a turn-over frequency of 1,5 min-1, a 70,000 fold increase compared to 7 Ph2Se2.

H O ArSH 2 Enz SeSAr ArSH ArSSAr

Enz SeOH Enz SeH

ROH ROOH

Figure 2 Cycle for GPx and selenosubtilisin

Hilvert and co-workers neglected to investigated if the enzyme would provide a chiral environment for the reduction of hydroperoxides. It was not until the end of the 90’s that Schreier and co-workers explored this possibility (figure 3).8 It turned out that a kinetic resolution of several aromatic hydroperoxides was indeed possible (see chapter 1).

OOH OOH OH SeSub ArSSAr H2O

2 ArSH Figure 3 Kinetic resolution of racemic hydroperoxides by Se-subtilisin

Other groups have also reported the incorporation of selenium as its selenocysteine derivative in glyceraldehyde 3-phosphate dehydrogenase or in catalytic antibodies.9,10 Indeed, these enzymes are active as GPx mimics, but no enantioselective reductions have been reported.

107 Chapter 7

In the previous chapter we have discussed the synthetic route of aromatic selenium compounds capable of selective binding to enzyme active sites. In this chapter the synthesis of the selenium-hybrid using papain as protein scaffold is reported. Secondly, these hybrids were probed as precatalysts in oxidations with hydrogen peroxide.

Results and discussion Papain is a commercially available cysteine hydrolase with a molecular weight of approximately 23 kDa. The inhibitors 1, 2, and negative control 3 were reacted with papain (figure 4). The inhibition of the enzyme with all compounds was rather slow ( >16 hours), after which no hydrolytic activity was observed any more (hydrolysis of p-nitrophenyl acetate).

O O

Cys25 S- N R Cys25 S N R H+ O O

O O CF3 O

N SeBu N SeBu N

O O O 1 2 3

Figure 4 Inhibition mechanism and structure of the selenium functionalized cysteine hydrolase inhibitors

All of the hybrids and non-modified papain were analyzed using Maldi-TOF analysis to check whether attachment of these inhibitors to the enzyme occurred. In figure 5, the mass analysis of native papain is shown. Despite the broadness of the peak the parent mass of 23 kDa could be distinguished. The main peak is at 23500 Da, but other maxima of 23600, 23800 and 24000 Da were observed as well. Interestingly, SDS-page did indeed show a broad band for papain (results not shown). 108 Enzymes as scaffolds for aromatic selenium catalysts

Figure 5 Maldi-TOF analysis of commercially available papain

This shows that the enzyme preparation is not very pure. Most probably the enzyme is polluted with low mass peptides which possibly bind to the enzyme, thus explaining the tailing. Next, the papain which was inhibited with compound 1 was measured and the mass distribution is shown in figure 6.

Figure 6 Maldi-TOF analysis of papain inhibited with compound 1

Even though there are peaks which still seem to coincide with the original enzyme preparation, a new peak with a mass of almost 23800 can be seen. This gives the indication that there is indeed a new compound attached to the enzyme with a mass of approximately 300 Da. Compound 1 has a molecular mass of 310 Da so the assumption could be made that the compound is indeed bound to the enzyme. Unfortunately, the resolution of the spectrum is not sufficiently resolved to be completely sure. 109 Chapter 7

The Maldi-TOF spectrum of papain-adduct 2 is somewhat more pronounced. As can be seen in figure 7, there is a distinct mass peak at 23850 Da, which is approximately 380 Da higher in mass than the native papain. This mass spectrum most convincingly shows that the inhibition of papain with the maleimide linkers indeed is a viable strategy.

Figure 7 Maldi-TOF analysis of papain with compound 2

Unfortunately, the spectrum of the enzyme with the negative control compound 3 (figure 8) very much resembles the original spectrum of papain (figure 5), so any conclusion regarding the attachment of this compound to the enzyme is not viable.

Figure 8 Maldi-TOF analysis of papain with compound 3

The conclusion from the above results is that the protein used should in principle be very pure to be able to be analyzed properly. The purification of papain can be performed via various methods. There is the method of covalent binding of the 110 Enzymes as scaffolds for aromatic selenium catalysts papain to thiopropyl sepharose 6B which seems a suitable method.11 The necessity to reactivate the chromatography material after every use makes this strategy tedious. Another promising method is the use of an affinity chromatography column (Gly-Gly-Tyr-Arg immobilized on cyanogen-activated sepharose).12,13 This was performed and indeed, the enzyme seemed to have a large fraction of protein products that were not active in the hydrolysis of p-nitrophenyl acetate (figure 9). There is some hydrolytic activity in the first few fractions of the preparation, which is most likely explained by a slight overloading of the column. Papain is eluted from the column upon reducing the ionic strength of the eluent.Therefore, one is certain that the enzyme solution is in this case pure and will yield very active enzyme without contaminants. The flow-through 34-41 mL was combined and the active thiol content was determined which showed a thiol to protein ratio of 0.95, indicating the purity of the enzyme.

1.4 0.5

1.2 A280 0.4 1

0.8 Activity 0.3

A280 0.6 0.2 Activity 0.4 0.1 0.2

0 0 0 5 10 15 20 25 30 35 40 45 50 flow through (mL)

Figure 9 Affinity chromatography of papain with functionalized agarose gel

Even though the purity of the enzyme is important to consider in later stages of the application, at this point the use of the non-purified papain should be sufficient to obtain a proof-of-principle for the selenium modified papain.

Testing hybrids in oxidation catalysis Butyl phenyl selenides are potential precursors of seleninic acids. As shown in previous studies,14 the reaction of PhSeBu with HOOH led to the formation of 111 Chapter 7 the seleninic acid, which can then perform an electrophilic oxygen transfer, i.e. an epoxidation or a Baeyer-Villiger oxidation. Several types of oxidations were tested with the papain adduct of compound 2 (figure 10 and table 1). Methanol and acetonitrile were added in order to solubilize the substrates. In all cases no conversion of the substrate was observed compared to the background reaction.

O O R2

N R1 O O

O O

R1 = H, R2 = SeBu (1) O S R1 = CF3, R2 = SeBu (2) S R1 = R2 = H (3)

Figure 10 Oxidation tests with hybrid enzyme papain adducts

Table 1 Oxidation tests with hybrid enzyme papain adducts 1 and 2 Reaction Enzyme (mg) Solvent (2 mL total volume) Temperature (°C) Time Epoxidation 10 / 20 / 30 MeOH : KPi 50 mM pH 7 1:1 20 24 h Epoxidation 10 / 20 / 30 MeOH : KPi 50 mM pH 7 1:1 40 24 h Epoxidation 30 MeOH : KPi 50 mM pH 7 1:1 60 8 h Epoxidation 30 MeCN : KPi 50 mM pH 7 1:1 40 24 h Epoxidation 30 MeCN : KPi 50 mM pH 7 3:1 40 24 h Epoxidation 10 mg SeSub MeCN : KPi 50 mM pH 7 3:1 40 24 h Epoxidation 10 mg SeSub MeCN : KPi 50 mM pH 7 1:1 40 24 h BV 30 MeOH : KPi 50 mM pH 7 1:1 4 72 h BV 30 MeOH : KPi 50 mM pH 7 1:1 20 24 h BV 30 MeOH : KPi 50 mM pH 7 1:1 40 24 h BV 10 mg SeSub MeOH : KPi 50 mM pH 7 1:1 40 24 h Sulfoxidation 30 MeOH : KPi 50 mM pH 7 1:1 20 24 h Sulfoxidation 30 MeCN : KPi 50 mM pH 7 3:1 20 24 h Sulfoxidation 30 MeCN : KPi 50 mM pH 7 5:1 30 24 h Sulfoxidation 10 mg SeSub MeCN : KPi 50 mM pH 7 3:1 20 24 h

As we were at this point uncertain whether the solvents were influencing the reactions we tested the sulfoxidation of thioanisole with catalyst 4 in a selection

112 Enzymes as scaffolds for aromatic selenium catalysts of organic solvents (figure 11). In dioxane, a water miscible non-protic solvent, the sulfoxidation reaction still works fine, but already in isopropanol, the reaction stalls after ± 25% conversion. It seems that the activation of the selenium catalyst in protic solvents is problematic, which is then reflected in the oxygen transfer reaction. For the reactions we have tested so far with the hybrid catalyst, it seems that these reactions are not compatible with the conditions suitable for the hybrid.

600

500

400 dioxane 10 mM cat 300 nBuOH 10 mM cat EtOH 10 mM cat 200 MeOH 10 mM cat iPrOH 10 mM cat 100 Concentrationthioanisole (mM) 00 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Time (h)

Figure 11 Oxidation of thioanisole with catalyst 4 and HOOH in several solvents

The use of selenium hybrids in oxidative bromination reactions The number of examples of selenium catalyzed reactions which are performed in water besides the glutathione peroxidase reactions are limited. The group of Tiecco published the dihydroxylation of several prochiral alkenes with diphenyl diselenide and hydrogen peroxide in a mixture of acetonitrile and water.15 However, the catalyst loadings were rather high (10% mol) which would not be a viable option with our hybrids. Another reaction which has received some attention in the recent years is the selenium catalyzed oxidative halogenation. The group of Detty first started with arylseleninic acids as catalysts for the bromination of 4-pentenoic acid with hydrogen peroxide as the oxidant (figure 12).16

113 Chapter 7

Figure 12 Oxidative halogenation using arylseleninic catalysts

In a following article, they reported that selenoxides were more effective catalysts in oxidative bromination than seleninic (per)acids.17 More recently, the selenium- catalyzed oxidative halogenation with PhSeCl was investigated.18 The oxidative halogenation can be easily monitored by the formation of bromophenol blue from phenol red (figure 13). We have also conducted this latter reaction with our hybrid catalysts. Unfortunately we did not see any conversion compared to the background.

O O O O S S PhSeCl, H O , Br O 2 2 O NaBr, buffer pH 7 OH OH

Br Br Br HO HO

Figure 13 Selenium catalyzed oxidative bromination of phenol red

Conclusions The attachment of the maleimide modified aromatic selenium compounds to papain, a well-known cysteine hydrolase was successful. Maldi-TOF mass spectrometry showed an increase of mass of respectively 300 or 380 Da, corresponding to the mass of the inhibitors. Full analysis was hampered by the impurity of the enzyme and the resulting hybrid enzyme preparations. Affinity chromatography could be used for purifying the crude papain. The crude hybrid enzymes were screened as catalysts for selenium catalyzed oxidation reactions, but in no case was any catalytic activity observed. Tests with isolated selenium catalysts showed that the activation of the catalyst and the activity were rather poor in protic solvents. This could be due to the the fact that the intermediate selenoxide 114 Enzymes as scaffolds for aromatic selenium catalysts can be responsible for hydrogen peroxide breakdown. These results thus show that selective attachment of arylselenium compounds to the active site of an enzyme was indeed possible, but that its activity as pre-catalyst is still to be confirmed. Further studies with pure protein and possibly other oxidants will need to be carried out to see whether future applications of these Se-hybrids are possible.

115 Chapter 7

Experimental section 1H and 13C NMR spectra were recorded with a Bruker avance 400 (400 or 100 Mhz). Chemical shifts are denoted in ppm (δ) in deuterated solvents and TMS as an internal standard. GCMS spectra were recorded with a Shimadzu GCMS-QP2010S, equipped with an auto injector AOC-20I and a Four VF5ms column (0.25 µm, length 30 m and diameter 0,25 mm). GC analysis was performed on a Shimadzu GC-2014, equipped with an autoinjector AOC-20I, an autosampler AOC-20S with a CP Sil 5a-CB column (0.40 µm, length 35 m and diameter 0,25 mm). MALDI-ToF mass spectra were recorded with an Applied Biosystems Voyager System 6020. UV-Vis spectroscopy was recorded on a UV-Vis Hewlett Packard 8452A Diode Array Spectrofotometer. Melting points were determined with a Büchi Melting Point B-540. Preparative column chromatography was performed with Fluka silica gel 60 (0,2 – 0,5 mm). TLC analyses were performed with Uniplate silica gel G. Solvents used in column chromatography were of technical quality. All reagents were purchased with the highest possible purity from chemical suppliers and used as received.

Inhibition of papain with compounds 1-3 All reactions with enzyme were performed under a nitrogen atmosphere. A solution of papain (3 mg / ml) was activated in 0.1 M phosphate buffer pH 7.0 with 1 mM of DTT. This was allowed to stir for 30 minutes and DTT was subsequently removed from the solution by gel filtration. To the remaining solution was added a stock solution (10 μL/mL of the activated protein) of 5.75 μM in dioxane of compounds 1, 2 or 3. To test for (residual) hydrolytic activity, all reactions were tested for the hydrolysis of Z-Gly-p-nitrophenyl (see below). After 23 hours, full inhibition of papain by all compounds was complete, where non-inhibited papain still showed 88% residual activity. The enzymes were all purified by ultracentrifugation or extensive dialysis and were subsequently lyophilized.

Affinity chromatography purification of papain13 Cyanogen bromide activated sepharose (1 g) was swollen in cold 1mM HCl for 30 minutes while the supernatant was actively refreshed (total 200 mL). After this, the resin was washed ten times with 10 mL of H2O and then quickly added to 42 μmol of peptide (Gly-Gly-Tyr-Arg) which was dissolved in 1.5 mL 0.5M NaHCO3. This slurry was allowed to react for 16 hours at 4 ºC during shaking. The gel was applied in a 5 mL syringe and washed with 0.1 M NaHCO3 and H2O. The washings were checked by UV for unbound peptide and it was found that 6 mg of peptide was bound to the gel. Purification of papain was performed by preactivating the papain (10 mg of papain in 2 mL) in 25 mM EDTA and 10 mM DTT at pH 4.3. The column was equilibrated with 20 mM EDTA and 10 mM DTT at pH 4.3. The enzyme was then applied to the column and the column was flushed with 20 mL of the equilibration buffer. Then the column was flushed with 10 mL of 5mM EDTA at pH 6.0. After this, the protein was eluted with milliQ water. The activity of all 1 mL fractions was checked by the hydrolysis of Z-Gly-p-nitrophenyl at 410 nm. To 2.9 mL of phosphate buffer 50 mM at pH 7.0 to which was added 50 μL of the activated ester (6.3 mM), followed by 50 μL of enzyme solution. The activity is defined as ΔA/min (410 nm).

Papain-hybrid oxidation tests (sulfoxidation, Baeyer-Villiger and epoxidation) For the exact reaction conditions, see table 1. The reactions were performed in 4 mL glass vials and shaken on a Thermotwister Comfort (Quantifoil Instruments) at 50 rpm. In the sulfoxidation and epoxidation experiments, 50 μL samples were withdrawn and analyzed directly on HPLC using a

Speedrod RP-18 column. The eluent was 60:40 H2O:MeOH with a flow rate of 1 mL/min. Detection 116 Enzymes as scaffolds for aromatic selenium catalysts was performed with a UV-Vis detector at 210 nm. In the Baeyer-Villiger reactions, 100 μL samples were withdrawn, diluted with 1 mL ethyl acetate and quenched with Na2SO3. The solids were filtered off and analyzed by GC.

Synthesis of butyl (4-bromophenyl) selenide (4) 4-Bromoaniline (20 mmol, 3.460 g) was dissolved in 10 mL water and 5 mL concentrated HCl was added. This mixture was cooled to 0 °C and NaNO2 (1.48 g, 21.5) dissolved in 8 mL H2O was added drop-wise. After the complete addition, the temperature was lowered to 0 °C and the pH was slowly raised to 5-7 with a saturated solution of NaOAc. After this, 20 mmol KSeCN (20 mL of a 1M solution) was slowly added whilst the reaction was vigorously stirred with a mechanical stirrer. The product was extracted with EtOAc. The organic layer was washed with brine and dried over anhydrous MgSO4. The raw product was filtered through a silica plug (hexanes) and was then purified via bulb-to-bulb distillation (200 °C, 0.2 mbar) which yielded (after cooling down) 4-bromophenyl selenocyanate as a light brown solid (3.747 g, 14.36 mmol, 72%) with a purity of 97% (GC). m/z (EI)

261 (with a typical Se-pattern). δH (400 MHz; CDCl3) 7.89 (2H, d, J = 8.8 Hz), 7.47 (2H, d, J = 8.8 Hz). To a solution of 4-bromophenyl selenocyanate (1 mmol, 261 mg) in 6 mL EtOH absolute was added NaOH (1.1 mmol, 44 mg) and the mixture was heated to reflux until all starting material has disappeared as judged by TLC analysis. Then 1-bromobutane (1.1 mmol, 128 μL) was added and the mixture was heated to reflux overnight. All volatiles were evaporated in vacuo and the remaining slurry was dissolved in EtOAc and washed with 1 M HCl and water. The organic layers were dried over Na2SO4 and the solvent was subsequently evaporated. The compound was purified over a silica column (PE only) which gave a yellow solid (266 mg, 0.91 mmol, 91%) with a purity of 99.9% (GC). m/z (EI) 292 (with a typical Se pattern).

Activation test of butyl (4-bromophenyl) selenide (4)

To a solution of compound 4 in 3 mL solvent (see figure 11) was added three equivalents of H2O2. Reactions were followed by taking 50 μL samples at regular times. Those samples were dissolved in ethyl acetate (1.5 mL) and MnO2 (10 mg) was added to quench the excess of hydrogen peroxide in the sample. Samples were dried over sodium sulphate and analyzed by GC.

Oxidation of thioanisole with compound 4

To a solution of 10 mM of compound 4 in 10 mL of solvent was added two mmol of H2O2 and the reaction was allowed to stand for one hour at room temperature. After this, 1 mmol of thioanisole and 0.5 mmol veratrole (internal standard) were added. Reactions were followed by taking 50 μL samples at regular times. Those samples were dissolved in ethyl acetate (1.5 mL) and MnO2 (10 mg) was added to quench the excess of hydrogen peroxide in the sample. Samples were dried over sodium sulfate and analyzed by GC.

References (1) Thomas, C. M.; Ward, T. R. Chem. Soc. Rev. 2005, 34, 337-346. (2) Marchetti, M.; Mangano, G.; Paganelli, S.; Botteghi, C. Tetrahedron Lett. 2000, 41, 3717-3720. (3) Bertucci, C.; Botteghi, C.; Giunta, D.; Marchetti, M.; Paganelli, S. Adv. Synth. Catal. 2002, 344, 556-562. (4) Kuang, H.; Brown, M. L.; Davies, R. R.; Young, E. C.; Distefano, M. D. J. Am. Chem. Soc. 1996, 118, 10702-10706. 117 Chapter 7

(5) Wu, Z. P.; Hilvert, D. J. Am. Chem. Soc. 1989, 111, 4513-4514. (6) Stadtman, T. C. Annu. Rev. Biochem. 1996, 65, 83-100. (7) Bell, I. M.; Hilvert, D. Biochemistry 1993, 32, 13969-13973. (8) Haring, D.; Herderich, M.; Schuler, E.; Withopf, B.; Schreier, P. Tetrahedron: Asymmetry 1997, 8, 853-856. (9) Boschi-Muller, S.; Muller, S.; Van Dorsselaer, A.; Böck, A.; Branlant, G. FEBS Lett. 1998, 439, 241-245. (10) Lian, G.; Ding, L.; Chen, M.; Liu, L.; Zhao, D.; Ni, J. Biochem. Biophys. Res. Commun. 2001, 283, 1007-1012. (11) Brömme, D.; Nallaseth, F. S.; Turk, B. Methods 2004, 32, 199-206. (12) Haquette, P.; Salmain, M.; Svedlung, K.; Martel, A.; Rudolf, B.; Zakrzewski, J.; Cordier, S.; Roisnel, T.; Fosse, C.; Jaouen, G. ChemBioChem 2007, 8, 224-231. (13) Blumberg, S.; Schechte.I; Berger, A. Eur. J. Biochem. 1970, 15, 97 (14) Betzemeier, B.; Lhermitte, F.; Knochel, P. Synlett 1999, 489-491. (15) Santoro, S.; Santi, C.; Sabatini, M.; Testaferri, L.; Tiecco, M. Adv. Synth. Catal. 2008, 350, 2881-2884. (16) Drake, M. D.; Bateman, M. A.; Detty, M. R. Organometallics 2003, 22, 4158-4162. (17) Goodman, M. A.; Detty, M. R. Organometallics 2004, 23, 3016-3020. (18) Mellegaard-Waetzig, S. R.; Wang, C.; Tunge, J. A. Tetrahedron 2006, 62, 7191-7198.

118 Summary

Summary of the thesis The use of selenium compounds in organic chemistry is widespread and well documented. Even though selenium compounds are frowned upon by many chemists due to the fact that they can be quite malodorous, their efficiency and activity in oxidation reactions make them quite unique. Especially in the activation of hydrogen peroxide, organoseleninic acids are almost unsurpassed in terms of selectivity and activity.

Chapter 1 starts with a general introduction on the use of selenium compounds in chemistry. An overview is given of the large array of oxidative transformations that can be performed with organoselenides. Special emphasis is directed to organoselenium compounds in combination with hydrogen peroxide as a sustainable oxidation technology. In order to expand this technology, the design of hybrid enzymes containing selenium moieties can be considered. An entry into this research field is added in this chapter.

Chapter 2 is dedicated to the optimization of the Ph2Se2 catalyzed oxidation of activated alcohols with TBHP. In literature two competing mechanisms were presented for this reaction. We have shown that benzeneseleninic anhydride (BSA) is the active oxidant and that it catalyzes its own formation most likely via the in situ formation of a perester. With high concentrations of TBHP, homolytic decomposition of this perester can result in the autoxidation of solvents containing reactive C-H bonds, such as toluene. In addition the autoxidation of benzaldehyde to benzoic acid could be observed. Based upon these findings we developed a new catalytic protocol, which included the in situ removal of water. In particular for large scale reactions, a Dean-Stark set-up turned out to be an elegant approach. Two activated alcohols, benzyl alcohol and cinnamyl alcohol, were oxidized by TBHP to the corresponding aldehydes employing <1% of Ph2Se2 as the catalyst. Compared to other methods with main group elements, our method is advantageous in that all compounds are cheap, readily available and the method is completely halogen-free.

In chapter 3, a variety of spectroscopic techniques plus reaction calorimetry were used to monitor the reaction. Thus, we have shown that the oxidation of

119 Summary

Ph2Se2 to BSA follows a number of distinct steps. This is consistent with the anhydride being the active oxidant in the (catalytic) oxidation of alcohols. Reaction calorimetry has been used to verify the behavior of activation of Ph2Se2 in different solvents. Little or no side-reactions occur in solvents that are characterized by a large exothermic profile for activation of Ph2Se2 with TBHP. In conclusion, we found reaction calorimetry to be a useful technique for rationalizing the course of catalytic oxidation reactions.

Chapter 4 is dedicated to the use of subsituted aromatic diselenides as precursors for the oxidation of alcohols. For the mechanism of oxidation of alcohols by these compounds there are two important parameters. First, there is the mode of activation of the diselenide by the oxidant. Certain diselenides such as dimesitylene diselenide are easily activated by TBHP whereas others are not activated at all. Secondly, the oxidation potential of the activated substituted diselenides differs greatly. Not all activated diselenides are capable of dehydrogenating non-activated alcohols. Finally, the regeneration of the anhydride of the active species is also expected to change when the substituents are varied. Some of the diselenides were tested in the semi-stoichiometric oxidation of 1- decanol and the most promising diselenides were tested in the catalytic oxidation of 1-decanol using our previously optimized oxidant and substrate feed protocol. It was shown that this was indeed a good way of improving the catalyst performance. However, more studies are needed in order to predict these properties and to develop better dehydrogenating selenium catalysts.

In chapter 5, a new group of renewable solvents has been successfully tested in the selenium-catalyzed epoxidation of cyclooctene and cyclohexene with hydrogen peroxide as oxidant. The results were comparable or better relative to standard organic solvents. Especially the prevention of epoxide hydrolysis was remarkable in certain members of these designer solvents. Moreover, a quantitative relationship between solvent polarity properties and rate of epoxidation has been established. It was concluded that the best solvents for this transformation should be strong hydrogen bond donors, but weak hydrogen bond acceptor ability. Catalyst stability tests show that the catalytic medium (solvent with the seleninic acid) can be repeatedly recharged with reactants without loss of catalytic activity and

120 Summary selectivity. Optimization of the reaction conditions results in the preparation of a recoverable catalytic phase allowing direct distillation of the epoxide product, with further recycling and reuse of the solvent and the catalyst. This recovery strategy could, in principle, be extrapolated to other catalytic transformations carried out in this class of green solvents.

Chapter 6 describes the synthesis of functionalized aromatic selenium compounds which could be capable of selective binding to enzymes. Specific enzyme-inhibitor combinations are considered. Both cysteine as well as serine based hydrolases were chosen as promising enzyme scaffolds. Synthesis of maleimide enhanced aromatic selenides was successful, as all of the synthetic steps were compatible with the selenide moiety. The synthesis of these compounds is simple and straightforward, thus opening the way to the synthesis of Se-based hybrid catalysts.

Finally in chapter 7 it was shown that attachment of maleimide modified aromatic selenium compounds to papain was partly successful. Mass analysis showed the impurity of the enzyme preparation, a problem which can be solved by using affinity chromatography. The crude hybrid enzymes were screened as catalysts for selenium catalyzed oxidation reactions, but in no case was any catalytic activity observed. Tests with isolated selenium catalysts showed that the activation of the catalyst and the activity were rather poor in protic solvents. It could be that the intermediate selenoxide induces hydrogen peroxide breakdown. These results thus show that selective attachment of arylselenium compounds to the active site of an enzyme is indeed possible, but that its activity as pre-catalyst is limited. Further studies with pure protein and possibly others oxidants will need to be carried out to see whether future applications of these Se-hybrids are possible.

121

Samenvatting

Samenvatting Verbindingen die het element seleen bevatten zijn tegenwoordig in de organische chemie in algemeen gebruik en er is dan ook veel wetenschappelijk onderzoek naar gedaan. Hoewel seleenverbindingen geen grote populariteit genieten bij chemici doordat ze vaak onwelriekend zijn, zijn ze uniek in oxidatiereacties. Met name in de activatie van waterstofperoxide zijn organoseleenzuren ongeëvenaard qua effectiviteit en selectiviteit.

In hoofdstuk 1 wordt er een algemene inleiding gegeven op het gebruik van seleenverbindingen in de chemie. Organoseleenverbindingen kunnen een grote verscheidenheid van oxidatieve reacties katalyseren en deze worden hier op een rij gezet. Er is extra aandacht voor de combinatie van deze verbindingen met waterstofperoxide als een duurzame manier om oxidaties uit te voeren. Om deze technische toepassing nog verder uit te breiden, kunnen hybride enzymen, waarin kunstmatig seleen wordt gezet, overwogen worden als nieuwe strategie. In dit hoofdstuk wordt een kort overzicht geschetst van dit concept en het onderzoeksveld van hybride enzymen.

Hoofdstuk 2 staat in het teken van de optimalisatie van de door Ph2Se2 gekatalyseerde oxidatie van zogenaamde geactiveerde alcoholen met tert-butyl hydroperoxide (TBHP) als oxidant. In de literatuur zijn er twee mogelijke mechanismen geponeerd. Wij hebben aangetoond dat benzeenseleenanhydride (BSA) het deeltje is dat de oxidatie uitvoert en dat het zijn eigen formatie uit

Ph2Se2 katalyseert middels een perester die gedurende de reactie wordt gevormd. Bij hoge concentraties van TBHP kan de homolytische decompositie van deze perester zorgen voor een autoxidatie van oplosmiddelen die reactieve C-H bindingen bevatten. Een voorbeeld van zo een oplosmiddel is tolueen. Bovendien zagen we ook nog de oxidatie van benzaldehyde naar benzoezuur. Op basis van al deze experimenten hebben we een nieuw katalytisch protocol ontworpen, waarbij het water werd onttrokken aan de reactie tijdens het uitvoeren van de reactie. Met name bij reacties op grote schaal was het gebruik van Dean-Stark condities een handige keuze. Twee geactiveerde alcoholen, namelijk benzylalcohol en kaneelalcohol, konden selectief geoxideerd worden tot hun overeenkomstige aldehydes, waarbij minder dan 1% Ph2Se2 als katalysator kon worden gebruikt. In

123 Samenvatting vergelijking met andere katalysatoren die ook gebaseerd zijn op verbindingen bestaande uit alleen hoofdgroepelementen, is onze methode erg aantrekkelijk omdat alle verbindingen goedkoop en makkelijk verkrijgbaar zijn. De methode bevat tevens in het geheel geen halogenen.

Omdat de vorming van het katalytisch actieve deeltje nog niet geheel duidelijk was wordt daar in hoofdstuk 3 dieper op ingegaan. Met behulp van een aantal spectroscopische technieken laten we zien dat de vorming van BSA uit Ph2Se2 een aantal specifieke stappen volgt. Dit komt ook volledig overeen met de eerdere conclusie dat BSA het katalytisch actieve deeltje is in de oxidatie van alcoholen.

Het gedrag van Ph2Se2 in verschillende oplosmiddelen is bestudeerd met behulp van reactiecalorimetrie. De oplosmiddelen die een sterk exothermisch profiel kennen tijdens de activatie van Ph2Se2 met TBHP zijn doorgaans inert. Al met al blijkt calorimetrie een zeer krachtige techniek om het verloop van dit soort katalytische oxidatiereacties beter te begrijpen.

In hoofdstuk 4 worden er diverse gesubstitueerde diseleniden gebruikt voor de katalytische oxidatie van alcoholen. Er zijn drie belangrijke parameters voor het mechanisme van oxidaties met behulp van deze verbindingen. Ten eerste wordt ieder diselenide anders geactiveerd door de oxidant. Een goed voorbeeld is dimesityleen diselenide dat zeer makkelijk wordt geactiveerd terwijl andere diseleniden in het geheel niet geactiveerd worden. Ten tweede zijn de geactiveerde seleenverbindingen niet allemaal even actief in de omzetting van verschillende alcoholen. Dat wordt met name duidelijk bij de oxidatie van niet geactiveerde alcoholen. Ten derde blijkt de regeneratie van iedere seleenverbinding anders. Enkele verbindingen zijn getest in de semi-kwantitatieve oxidatie van 1-decanol en degenen die het meeste potentiaal lieten zien zijn getest met het protocol dat in hoofdstuk 2 beschreven is, maar nu voor de oxidatie van 1-decanol. Het bleek dat dit een prima manier was om de hoeveelheid katalysator te verminderen. Desalniettemin is het noodzakelijk om meer onderzoek te doen naar het gedrag van deze verbindingen om zodoende betere katalysatoren te ontwikkelen voor dehydrogeneringsreacties.

124 Samenvatting

In hoofdstuk 5 is er een nieuwe groep van oplosmiddelen getest in de selenium gekatalyseerde oxidatie van cycloocteen en cyclohexeen. Deze oplosmiddelen zijn gebaseerd op de hernieuwbare grondstof glycerol. Bij deze reacties is waterstofperoxide gebruikt als oxidant en de resultaten waren vergelijkbaar of beter met dezelfde oxidaties in normale organische oplosmiddelen. Wat opviel was dat in sommige van deze speciaal gemaakte oplosmiddelen de hydrolyse van het epoxide onderdrukt werd. Er werd een correlatie gevonden tussen specifieke kenmerken van deze oplosmiddelen en de snelheid waarmee de reactie plaatsvond. De beste oplosmiddelen moeten sterke waterstofbrugdonors zijn, maar zwakke waterstofbrugacceptors. De stabiliteit van de katalysator was opvallend goed en het is aangetoond dat het oplosmiddel met daarin de katalysator (de zogenaamde katalytisch actieve fase) iedere keer weer nieuwe reagentia kan accepteren zonder dat de activiteit of selectiviteit van de reactie verloren gaat. Door optimalisatie van de reactiecondities kon het proces zodanig ontworpen worden dat het product (het epoxide) direct uit het mengsel gedestilleerd kon worden waarna het mengsel weer opnieuw een reactie kon katalyseren. In principe zou deze strategie ook toegepast kunnen worden in andere katalytische reacties met dit soort groene oplosmiddelen.

Hoofdstuk 6 staat in het teken van de synthese van gefunctionaliseerde aromatische seleenverbindingen die selectief aan enzymen zouden kunnen binden. In het hoofdstuk worden diverse ontwerpmogelijkheden voor dit soort verbindingen geschetst. De enzymen die wij op het oog hebben als doel om onze verbindingen aan vast te maken zijn zogenaamde cysteine en serine hydrolases. De synthese van aromatische verbindingen waaraan een maleimide groep vastzit was succesvol omdat alle synthetische stappen niet interfereerden met de seleen- groep. De uiteindelijke synthese van deze verbindingen is relatief eenvoudig waardoor het mogelijk moet zijn om seleen-gefunctionaliseerde enzymen op eenvoudige wijze te maken.

In hoofdstuk 7 wordt aangetoond dat de eerder genoemde maleimide gemodificeerde aromatische seleenverbindingen inderdaad aan specifieke eiwitten kunnen binden. Met behulp van massa spectroscopie was te zien dat het gebruikte enzym nogal onzuiver was. Dit probleem kon opgelost worden door middel van

125 Samenvatting affiniteitschromatografie. Desondanks werden de nieuwe hybride enzymen in ruwe vorm getest in seleengekatalyseerde reacties, maar helaas was er geen enkele reactie waarbij katalytische activiteit aantoonbaar was. Bij testen met de katalysator waaraan geen eiwit vastzat bleek dat deze verbindingen weinig activiteit vertoonden in protische oplosmiddelen. Een van de mogelijke oorzaken is dat een van de intermediairen in de reactie, het seleenoxide, voor de afbraak van waterstofperoxide zorgt. Het is dus mogelijk om aromatische seleenverbindingen vast te maken aan de actieve plek in het enzym, maar deze hybride enzymen vertonen (nog) geen katalytische activiteit. Het zou zeer interessant zijn om zuiver enzym en eventuele andere oxidanten te gebruiken om verder te onderzoeken of deze selenium hybride enzymen interessante toepassingen op kunnen leveren.

126 Dankwoord

Dankwoord Een proefschrift is niet compleet zonder een dankwoord aan allen die op enigerlei wijze betrokken zijn geweest bij het maken van dit boekje. Dat zijn nog aardig wat mensen geweest, maar dat kan ook niet anders als je er iets langer over doet dan eigenlijk gepland staat.

Roger, ik kan me de dag nog goed herinneren dat ik uitgenodigd was op jouw kantoor in Delft. Ik was ontzettend zenuwachtig en ik weet nog dat ik meerdere formules van allerlei suikers waarmee ik gewerkt had in Leiden verkeerd op het bord schreef. Ik wil je bedanken voor de kans die je me toen gegeven hebt om in de vakgroep te komen werken. Jouw enthousiasme en frisse blik op onderzoek doen zal ik niet vergeten. Isabel, met jou heb ik het meeste contact gehad van alle begeleiders. Je moest me vaak afstoppen in mijn enthousiasme die ik onbewust besteedde in allerlei zijprojecten en ik heb heel veel van je geleerd. Ik heb een zeer groot respect voor jouw doorzettingsvermogen, ook op de momenten als het helemaal niet gaat zoals het moet en dan heb ik het natuurlijk niet alleen over onderzoek doen. Ik hoop dat je in de komende jaren de vakgroep op kunt bouwen zoals jij voor ogen ziet. Bedankt dat ik jouw aio mocht zijn. Fred, ik heb altijd genoten van de discussies die wij voerden aan de allang opgeheven koffietafel in de Boesekenzaal. Jouw onmetelijke kennis is van grote waarde geweest voor velen die in de donkere labzaal door mochten brengen. Mieke, je hebt het al vaker gehoord, maar het kan niet vaak genoeg gezegd worden: BOC zou niet kunnen bestaan zonder al jouw inspanningen. Mijn dank is groot voor alle tijd en moeite die jij in alle administratieve processen stopt binnen en buiten de groep. Frank, ik maakte deel uit van de selectiecommissie toen jij solliciteerde voor de tenure track. Jij hebt een geheel eigen wijze van het managen van je projecten en ik verwacht dat dat jou nog veel succes zal brengen in de toekomst. Kristina, bedankt voor het vertrouwen dat je in mij had om zelf met de NMR aan de slag te gaan, ik geloof dat ik destijds een van de eersten was buiten de mensen van de NMR club om die aan de apparatuur mocht zitten. Ik vind het meer dan terecht dat je een vaste aanstelling hebt gehad en ik wens je alle goeds in de toekomst. Ulf, bij de algemene werkbesprekingen, maar vooral toen eerst Isabel en later Frank bij jou op het kantoor zaten hebben wij regelmatig interessante wetenschappelijke discussies

127 Dankwoord gevoerd en die heb ik altijd zeer gewaardeerd. Ik kon ook altijd bij je binnenlopen voor een handtekening als ik weer eens chemicaliën wilde bestellen, ik heb dat altijd gewaardeerd. Remco, wat had ik toch moeten doen tijdens mijn promotie als je mij niet had geholpen met alle problemen van kapotte pompen, extra organisch glaswerk en andere materialen waarmee je me altijd uit de brand hielp. Ik heb nog altijd goede herinneringen aan de practica die we samen draaiden, ook al hadden we wel eens woorden. Ik hoop dat alle veranderingen in jouw leven je ten goede zullen komen. Lars, dank je wel voor alle hulp bij de metingen of als een GC(MS) weer eens niet deed wat ik wilde.

Een speciaal stuk in dit dankwoord is natuurlijk gericht aan Marcel Schreuder- Goedheijt en Gerjan Kemperman die mij de mogelijkheid gegeven hebben om een deel van mijn promotieonderzoek uit te voeren in de labs in Oss. Ik heb intens genoten van die speciale tijd die ik als buitenstaander bij DPC mocht ervaren en voor het feit dat ik zo ontzettend veel vrijheid heb genoten op de afdeling. Ik denk dat uit dit boekje wel blijkt hoe effectief die tijd voor mij geweest is. Ook het feit dat jullie altijd tijd voor mij hadden en de (niet) wetenschappelijke discussies die wij over het onderzoek gevoerd hebben zijn van cruciaal belang geweest voor mijn ontwikkeling. Ook al heb ik er zo kort gezeten, mijn hart doet zeer bij de gedachte dat het niet meer is wat het geweest is. Ik wil ook de mannen bedanken die mij op weg geholpen hebben op het lab. Marco, Gijs en Roy, bedankt voor jullie hulp en de leuke discussies. Hans en Emiel, bedankt voor het hulp met de analytische (GC) vraagstukken. Peter, dank je wel voor de hulp met de robotica en de reactie calorimetrie. Zonder dat apparaat hadden een paar hoofdstukken er anders uitgezien. Daniel, thank you for the nice discussions and fun we had and helping me getting acquainted in the lab. Tenslotte wil ik Martin Ostendorf bedanken voor de discussies in de B-Basic vergaderingen.

Een speciale sectie moet natuurlijk ingericht worden voor de studenten die stage bij mij hebben gelopen. Prieyé, het is jammer dat jouw enthousiasme er niet altijd uitkwam in de resultaten en helaas heb ik geen van jouw resultaten kunnen insluiten in dit boekje. De andere student die nog een tijd bij mij heeft rondgelopen is Ron. Dank je wel voor je inspanningen, het was voor mij een

128 Dankwoord prettige tijd om met je samen te kunnen werken, ook al heb je besloten om een ander carrièrepad in te slaan.

Hector, I would like to thank you for the collaboration we had on the glycerol based solvents project. You have shown to be a very hard worker and a good scientist. I wish you all the best for the future.

Er zijn nog een aantal mensen uit Delft die hier een speciale plek verdienen. Sander kerel, ik weet niet hoe ik die vier jaar in de bieb had moeten overleven zonder jouw aanwezigheid. Je bent geen prater met betrekking tot je emoties, maar ik heb groot respect voor de manier waarop jij je door de moeilijke tijd heen hebt geslagen. Verder waren de uitjes naar het noorden, de congressen en als klap op de vuurpijl de reis naar New York en Boston echte hoogtepunten. De diverse etentjes en borrels mogen natuurlijk ook niet vergeten worden en ook daar had ik af en toe je hulp nodig. Master Chmura, (I should say doctor Chmura, but that does not sound the same), thank you for the wonderful time we had in the library. You made life in the lab a lot of fun, especially with the sandwiches on Fridays. HP, ouwe klaagbaas van me, dank je wel voor het lekker ongegeneerd klagen en onze verbazing delen gedurende onze projecten. Seda, speciaal voor jou in het Nederlands, ik had gehoopt dat we meer zouden samenwerken, maar ik wil je toch bedanken voor de tijd die we samen hebben doorgebracht in de vakgroep. Furthermore I would like to thank the people from the oxidation subgroup in the lab. Special memories I still have to our work meetings. Aleksandra, your maternal guidance was of essence in our group. Inga, thank you for your humor and the good discussions we always had. Jeroen, je zat niet beneden, maar toch was het altijd erg gezellig. Jij bent de volgende en ik wens je alle succes hiermee. Daniela, please keep an eye out for Sander. May your hopes and dreams come true. De mannen van CLEA verdienen ook een vermelding hier. Mike en Menno, dank je wel voor de mooie tijd. Chrétien, ik hoop dat we de game avondjes in ere kunnen houden. I would like to thank everyone else involved in or outside of the group: Marco N, Hilda, Pedro, Monica, Maria, Luigi, Christophe, Luuk, Tobias, Selvedin, Jin, Ksenia, Marina, Aida, Frederico, Dani, Ton, Maarten, Daniel and everyone else I am forgetting here.

129 Dankwoord

De mensen buiten Delft verdienen zeker een plaats in dit dankwoord. Annemiek, dank je wel dat je mijn stapmaatje bent geweest in al die tijd. Ik weet zeker dat we nog een hoop mooie momenten mee gaan maken. Remon, dank je wel dat ik haar altijd mee mag nemen!

De mannen van de tafeltennis zijn van onbeschrijflijke waarde in mijn leven. Iedere week mag ik met jullie mooie en minder mooie momenten beleven. Sam en Bart, hoewel jullie geen paranimf zijn wisten jullie op eigen wijze altijd een mening te vormen over het boekje. Dank jullie wel voor de steun. Dat geldt natuurlijk voor iedereen bij Taveri.

Kristel, jij bent al 12 jaar mijn scheikundemaatje en we hebben al aardig wat meegemaakt in die tijd. Jij hebt besloten een andere richting op te gaan en daarom is het ook zo mooi dat we elkaar nog steeds steunen. Ik ben vereerd dat je mijn paranimf bent.

De familie moet natuurlijk niet vergeten worden. Pa en ma, dank jullie wel voor alles en voor het feit dat ik altijd mijn eigen keuzes heb mogen maken, ook al ging ik nogal eens rommelig om met mijn vrijheid. Zonder jullie was ik hier niet gekomen. Andy, ik ben enorm trots op je, ook al laat ik dat niet altijd blijken. Jij en Daisy gaan een mooie toekomst tegemoet, dat weet ik zeker. Leo, Margreet en Wes, dank jullie wel voor mij te accepteren wie ik ben en jullie interesse in het hele traject. Binnenkort horen we ook officieel bij elkaar, nu alvast op papier in deze alinea.

Als laatste wil ik mijn betere helft bedanken voor alle geduld en liefde die ze met mij heeft. Zij vult mij aan in alles wat ik niet heb of ben. Sabine, als er iemand is zonder wie dit proefschrift er nooit was gekomen, dan ben jij het wel. Het is tijd voor een nieuw hoofdstuk, maar dat ligt buiten dit boekje.

John

130 List of Publications

“Aromatic selenides for selective attachment to enzymes: towards selenium enhanced hybrid enzyme catalysts” John C. van der Toorn, Ron van Klaveren, Frank Hollmann, Roger A. Sheldon and Isabel W.C.E. Arends; manuscript in preparation

“Studies on (substituted) aromatic diselenides as catalysts for selective alcohol oxidation using tert-butyl hydroperoxide” John C. van der Toorn, Gerjan Kemperman, Roger A. Sheldon and Isabel W.C.E. Arends; European Journal of Organic Chemistry 2011, 4345-4352

“Epoxidation of cyclooctene and cyclohexene with hydrogen peroxide catalyzed by bis[3,5- bis(trifluoromethyl)-diphenyl] diselenide: Recyclable catalyst-containing phases through the use of glycerol-derived solvents” Héctor García-Marín, John C. van der Toorn, José A. Mayoral, José I. Garcia and Isabel W.C.E. Arends; Journal of Molecular Catalysis A: Chemical 2011, Vol. 334, 83-88 . “Photoenzymatic reduction of C=C double bonds” Maria Mifsud Grau, John C. van der Toorn, Linda G. Otten, Peter Macheroux, Andreas Taglieber, Felipe E. Zilly, Isabel W. C. E. Arends and Frank Hollmann; Advanced Synthesis and Catalysis 2009, Vol. 351, 3279-3286.

“Glycerol-based solvents as green reaction media in epoxidations with hydrogen peroxide catalyzed by bis[3,5-bis(trifluoromethyl)phenyl] diselenide” Héctor García-Marín, John C. van der Toorn, José I. Garcia, José A. Mayoral and Isabel W.C.E. Arends; Green Chemistry 2009, Vol. 11, No. 10, 1605-1609.

“Diphenyldiselenide-catalyzed selective oxidation of activated alcohols with tert- butylhydroperoxide; new mechanistic insights” J.C. van der Toorn, G. Kemperman, R.A. Sheldon, I.W.C.E. Arends; Journal of Organic Chemistry 2009, Vol. 74 No. 8, 3085-3089.

“Selective, catalytic aerobic oxidation of alcohols using CuBr2 and bifunctional triazine- based ligands combining both a bipyridine and a TEMPO group” Zhengliang Lu, Tim Ladrak, Olivier Roubeau, John van der Toorn, Simon J. Teat, Chiara Massera, Patrick Gamez and Jan Reedijk; Dalton Transactions 2009, 3359-3570.

“Thioglycuronides: synthesis and applications in the assembly of acidic oligosaccharides” L.J. van den Bos, J.D.C. Codée, J.C. van der Toorn, T.J. Boltje, J.H. van Boom, H.S. Overkleeft and G.A. van der Marel; Organic Letters 2004, Vol. 6 No. 13, 2165-2168. Curriculum Vitae John C. van der Toorn werd geboren op 20 september 1979 te Ridderkerk. In 1999 behaalde hij het atheneumdiploma aan het Farelcollege te Ridderkerk. In datzelfde jaar werd er begonnen aan de opleiding Scheikunde aan de Universiteit Leiden. Er werd een minor gevolgd in homogene katalyse en een major in synthetische organische chemie, waarbij het synthetiseren van een biologisch relevant oligosaccharide het doel van afstuderen was. In 2005 koos hij ervoor om een promotieonderzoek te starten aan de Technische Universiteit Delft in de vakgroep Biokatalyse en Organische Chemie onder toezicht van Prof. dr. Roger A. Sheldon en Prof. dr. Isabel W. C. E. Arends naar het gebruik van seleen-gemodificeerde enzymen als mogelijke oxidatieve hybride katalysatoren. Gedurende het onderzoek heeft hij ook gewerkt bij Schering-Plough (tegenwoordig Merck, Sharp and Dohme) in Oss in de groep van Dr. Marcel Schreuder-Goedheijt en Dr. Gerjan Kemperman. Hier heeft hij onderzoek gedaan naar aromatische seleenverbindingen als katalysatoren voor de oxidatie van alcoholen. De resultaten van het onderzoek zijn in dit proefschrift beschreven. Sinds 1 september 2009 is hij werkzaam als docent analytische chemie bij de opleiding Chemie en Life Sciences aan de Hogeschool Inholland te Amsterdam.