Bachelor Thesis Chemistry
Biocatalytic Synthesis and Characterization of Fossil Long
Chain Diols
Novel proxies for climate reconstruction
by
Celine Nieuwland
18 July 2017
Student number
10700587
Research Institute
Van 't Hoff Institute for Molecular Sciences
Research group
Synthetic Organic Chemistry and Biocatalysis
Supervising professors
prof. dr. J. H. van Maarseveen and dr. F. Mutti
Daily supervisors
M. J. Wanner and W. Böhmer
Second reviewer
2
Every aspect of the world today – even politics and international relations – is affected by chemistry. – Linus Carl Pauling
3
1.1 Abstract
With climate change being a matter of global concern it is crucial to reconstruct past climate conditions, in order to predict future climate change. Marine sediments turn out to contain chiral long chain diols originating from algae, that offer valuable information about past physiological conditions, as algae adjust their diol distribution to external factors, such as temperature. It is unclear why these organisms do this. In order to shed light on this phenomenon, extensive structural analysis of these long chain diols, and the derived polymeric networks in which these occur, is required. However, isolation and therewith, the analysis of the diol monomers of these robust polymers showed to be challenging. In this research it was proposed to find a suitable method for analysis of the stereochemistry of the long chain diols. HPLC showed to be the most compatible with the nonvolatile character of the diols. The two enantiomers of the synthesized racemic C30 1,13-diol (44% yield, six steps) could be separated by
straight phase chiral HPLC upon chemical derivatization. However, it must be investigated whether this method can be applied on marine sediment samples, because the diols can be isolated in only very small amounts. This research additionally examined enzymatic enantioenrichment of intermediates in the synthesis of the long chain diols. Alcohol dehydrogenases from Sphingobium yanoikuyae and Ralstonia sp. showed to be unsuitable for this purpose. Biotransformations with Amano Lipase (Pseudomonas
cepacia) gave high conversions, but no enantioselectivity due to ambiguous substrate binding.
4
1.2 Populair Wetenschappelijke Samenvatting
Klimaatverandering is een kwestie die ons wereldwijd dagelijks bezighoudt. De toekomst van de natuur en de welvaart van de bevolking hangen sterk af van de maatregelen die op dit moment genomen worden om bijvoorbeeld het broeikaseffect tegen te gaan. Reconstructie van het klimaat door de geschiedenis heen kan gebruikt worden om te voorspellen hoe het klimaat in de toekomst zal veranderen. Dit soort reconstructies gebeurt aan de hand van zogenoemde ‘biomarkers’. Dit zijn moleculen afkomstig uit levende organismen die over tientallen, en soms honderden jaren bewaard zijn gebleven in de aardbodem.
In de jaren 80 ontdekten onderzoekers van het Koninklijk Nederlands Instituut voor Onderzoek der Zee (NIOZ) bestanddelen van celwanden van algen die door de jaren heen geconserveerd zijn gebleven in diepzeesedimenten. Deze vetzuurachtige verbindingen, bestaande uit een lange koolstofketen en twee alcoholgroepen, blijken nuttige informatie op te leveren over het zoutgehalte en de temperatuur van het zeewater waarin de algen leefden (Figuur 1). Algen passen namelijk de ratio van deze verschillende lange diolen in hun celwand aan op deze fysiologische omstandigheden. Om beter te begrijpen waarom algen deze ratio aanpassen, is een uitgebreide analyse nodig van de eigenschappen van deze moleculen. Deze moleculen hebben namelijk een bijzonder kenmerk: het koolstofatoom van de centrale alcoholgroep is omringd door vier verschillende groepen. Het molecuul is hierdoor chiraal, wat betekent dat er twee spiegelbeelden zijn per molecuul. Om de eigenschappen van dit type moleculen te begrijpen is het nodig om deze zowel racemisch (mengsel van de spiegelbeelden), als optisch zuiver (één van de spiegelbeelden) te analyseren. Het is vaak lastig om met de technieken op het laboratorium slechts één van de spiegelbeelden te synthetiseren, maar enzymen kunnen dit maar al te goed.
Figuur 1. Structuur van de lange diolen in celwanden van algen.
In dit onderzoek zijn enzymen gebruikt voor de synthese van één van de spiegelbeelden van de lange diolen. Enzymen afkomstig uit de Gram-negatieve bacteriën Sphingobium yanoikuyae en Ralstonia sp. blijken deze moleculen niet om te zetten en zijn dus niet geschikt voor dit doel. Biotransformaties met een enzym afkomstig uit de bacterie Pseudomonas cepacia, leverde het racemische mengsel in plaats van de optisch zuivere stof, omdat de substraten op verschillende manieren aan de enzymen gebonden kunnen worden. In vervolgonderzoek moet een substraat ontworpen worden dat slechts op één manier kan binden.
Daarnaast werd er in dit onderzoek gezocht naar een geschikte methode om de spiegelbeelden van elkaar te kunnen scheiden. Vloeistofchromatografie onder hoge druk bleek de meest geschikte scheidingsmethode voor de diolen. Bij vloeistofchromatografie kunnen twee spiegelbeelden gescheiden worden met behulp van een speciale kolom. Het principe hiervan staat weergeven in Figuur 2. Hierbij bindt één van de spiegelbeelden aan de korrels in de kolom, terwijl het andere spiegelbeeld niet kan
5
binden. Op deze manier kan het eerste spiegelbeeld geïsoleerd worden door een mengsel van vloeistoffen, het eluens, door de kolom te laten lopen.
Figuur 2. Het principe van een kolom waarmee moleculaire spiegelbeelden gescheiden kunnen worden.
Zodra het ene spiegelbeeld de kolom heeft verlaten, kan de binding van het andere spiegelbeeld met de kolom verbroken worden door de samenstelling van het eluens aan te passen. Zo wordt ook het andere spiegelbeeld zuiver in handen gekregen.
Om de twee spiegelbeelden van de lange diolen te kunnen scheiden, moest eerst het racemische mengsel gemaakt worden door twee kortere koolstofketens aan elkaar te koppelen. Na het bevestigen van een uv-licht absorberend molecuul aan de alcoholgroep op het eerste koolstofatoom van het lange diol, konden de twee spiegelbeelden gescheiden en gedetecteerd worden via vloeistofchromatografie met zo’n speciale kolom die gekoppeld is aan een uv-licht detector. Het moet echter nog uitgezocht worden of deze analysemethode toepasbaar is op monsters uit diepzeesedimenten, omdat de lange diolen lastig en vaak in zeer kleine hoeveelheden geïsoleerd kunnen worden.
6
1.3 List of Abbreviations
Ac2O Acetic anhydrideADH Alcohol Dehydrogenase
AL-PS Amano Lipase PS, from Burkholderia cepacia Bz(CN) Benzoyl (cyanide)
13C NMR Carbon nuclear magnetic resonance
DCM Dichloromethane
DMAP 4-dimethylaminopyridine
DMF N,N-Dimethylformamide
DMSO Dimethyl sulfoxide
E. coli Escherichia coli
ELSD Evaporative light scattering detector
Et2O Diethyl ether
EtOAc Ethyl acetate
EtOH Ethanol
FID Flame Ionization Detection
GC Gas Chromatography
GDH Glutamate dehydrogenase
1H NMR Proton nuclear magnetic resonance
HPLC High-performance liquid chromatography
IPA Isopropyl alcohol
IPTG Isopropyl β-D-1-thiogalactopyranoside IR Infrared (spectroscopy)
Kan Kanamycin sulphate for molecular biology (antibiotic)
LB Luria Bretani broth Miller (= casein enzymic hydrolysate 10 g/L, yeast extract 5 g/L, NaCl g/L, final pH 7.5 ± 0.2 (25 ˚C)
MeOH Methanol
MM Molecular mechanics
MS Mass Spectroscopy
NAD(P) Nicotinamide adenine dinucleotide (phosphate)
n.d. Not determined
NIOZ Koninklijk Nederlands Instituut voor Onderzoek der Zee (Royal Netherlands Institute for Sea Research)
PDA Photo-Diode Array (detector)
PE Petroleum ether
PG Protecting group
Pi Phosphate
Ras-ADH Ralstonia sp. alcohol dehydrogenase
Sy-ADH Sphingobium yanoikuyae alcohol dehydrogenase
TBAF Tetra-n-butylammonium fluoride TBSCl Tert-butyldimethylchlorosilane
tBu Tert-butyl
TEA Triethylamine
THF Tetrahydrofuran
TLC Thin layer chromatography
7
Table of Contents
1.1 Abstract ... 3
1.2 Populair Wetenschappelijke Samenvatting ... 4
1.3 List of Abbreviations ... 6
2. Introduction ... 8
3. Results and Discussion ... 13
3.1 Racemic synthesis of the C28 and C30 1,13-diols... 13
3.2 Biotransformations ... 18
3.3 Chiral HPLC studies ... 24
6. Acknowledgements ... 28
7. Experimental Section ... 29
7.1 General procedures ... 29
7.2 Racemic synthesis of the C28 1,13-diol ... 29
7.3 Racemic synthesis of the C30 1,13-diol ... 31
7.4 Expression and lyophilization of E. coli cells containing recombinant proteins ... 34
7.5 Enzyme activity test reactions... 34
7.6 Biotransformations with Sy-ADH and Ras-ADH ... 34
7.7 Biotransformations with AL-PS ... 34
7.8 Molecular Mechanics (MM) simulations of 3b ... 34
8. References ... 35
9. Appendix ... 37
9.1 1H NMR Spectra of the Synthesized Compounds ... 37
9.2 13C NMR Spectra of the Synthesized Compounds... 43
9.3 HPLC Chromatograms ... 44
8
2. Introduction
With climate change being a matter of global concern, detailed knowledge on climate processes and feedback mechanisms is crucial. In order to determine which measures are needed to control the effects of climate change, reconstruction of past climate conditions can be used to predict future climate changes. Although palaeoclimate data is particularly limited for continental areas, ancient lipids preserved in marine sediments may provide useful information about past climate conditions, as some organisms adapt their lipid distribution to physiological factors, such as temperature. In this way, these ancient lipids can be used to trace back climate conditions at the time of deposition, a topic of high interest to the researchers associated to the Royal Netherlands Institute for Sea Research (NIOZ). Examples of well-preserved lipids are the long alkyl chain C28 and C30 1,13-, 1,14- and 1,15-diols which
are abundantly present in Quaternary marine sediments (Figure 1).
Figure 1. Structure of the C28 and C30 1,13-, 1,14- and 1,15-diols found in Quaternary marine sediments.
This class of compounds was first identified in 1981 by de Leeuw et al. in the sediments of the Black Sea and are suspected to find their function in cell walls of algae.1 The cell wall of algae consists of an
inner cellulose based layer and an outer layer based on algaenan, a term that consists of long aliphatic hydrocarbons that are cross-linked via ether bonds (Figure 2).2 Algaenans are highly resistant to
hydrolysis and insoluble in aqueous and organic solvents, making them stay preserved after decades in deep sea sediments. The robust character of these polymeric compounds also has a downside, as biochemical characterization is challenging due to the chemical modifications needed for isolation.
9
The diatom genus Proboscia seems to be the main source for C28 and C30 1,14-diols, but these
compounds were also reported in the marine alga Apedinella radians.3–5 On the other hand, the 1,13-
and 1,15-diols originate from the eustigmatophyte eukaryotic algae species.6,7 However, the role of this
alga as major source is questioned, as cultured species do not produce the same distribution of lipids as the wild type and the fact that eustigmatophytes are rarely present in marine environments.6,8
The 1,14-diols showed promising application in reconstruction of, for instance, periods of oceanic upwelling in the past (Figure 3).9,10 Upwelling is
an oceanographic phenomenon that involves wind-driven replacement of warm and usually nutrient-depleted surface water by cooler and nutrient-rich deep seawater.11 Areas with
nutrient-rich water show more extensive Proboscia growth and therefore higher concentrations of the 1,14-diols.
Although there is much less known about the application of the 1,13- and 1,15-diols as climate reconstruction proxies, Versteegh et al. reported in 1997 that the ratio between C30 and C32 1,15-diols
is directly related to the salinity of the seawater.8,12 More C
32 1,15-diols would be present in areas with
fresh and nutrient-rich water influx, due to upwelling. Rampen et al. reported in 2012 that the fractional abundance (F) of C28 and C30 1,13-diols and the C30 1,15-diol is controlled by another physiological
factor, the sea surface temperature (SST).13 The correlation between the SST and the fractional
abundance of these specific diols is depicted in Equation 1.
Equation 1. Correlation between the fractional abundance (F) of 1,13- and 1,15-diols and the sea surface temperature (SST).
𝐿𝑜𝑛𝑔 𝐶ℎ𝑎𝑖𝑛 𝐷𝑖𝑜𝑙 𝐼𝑛𝑑𝑒𝑥 (𝐿𝐷𝐼) = 𝐹𝐶301,15 𝑑𝑖𝑜𝑙
𝐹𝐶301,13 𝑑𝑖𝑜𝑙+ 𝐹𝐶281,13 𝑑𝑖𝑜𝑙+ 𝐹𝐶301,15 𝑑𝑖𝑜𝑙
= 0.033 × 𝑆𝑆𝑇 + 0.095
It is well-known that chain length and degree of saturation of lipids in cell membranes enable organisms to adapt to physiological changes like temperature, in order to maintain constant membrane fluidity. However, this only explains the fact that C30 1,15-diols become more abundant at increased temperature
with respect to the C28,13-diols. Replacement of the C30 1,13-diols by C30 1,15-diols does not affect the
chain length distribution of the membrane. Further research on algae cultures and biochemical characterization of these compounds is needed in order to shed light on the exact relation between the mid-chain hydroxyl group and temperature. Analysis of the melting points of the individual and mixtures of diols might provide useful information as the melting points of the racemic mixture and the enantiomerically pure compounds may differ, explaining their functionality in cell walls and the corresponding temperature relation.
Besides the notable temperature relation, the mid-chain hydroxyl group of the 1,13- and 1,15-diols possesses another special feature: it is a chiral center. Therefore it is of high interest to determine the absolute configuration of these compounds in nature. However, the amounts of long chain diols isolated from marine sediments by NIOZ are very small. De Leeuw and co-workers managed to identify the long chain diols by GC-MS in 1981.1 A great advantage of GC-analysis is that it is a relatively low-cost
method. But there is another problem accompanied with the GC-analysis of long chain diols: the high molecular weight. This feature requires high temperature programs during GC-analysis. Beside the fact that the heavy long chain diols have a long retention time (>65 min. at 275 ˚C) on GC, not all GC-columns are compatible with these higher temperatures. There are non-chiral GC-columns that can stand
10
>400 ˚C, examples are the Agilent J&W Select Mineral Oil (375/400 ˚C) and the CP-Sil PAH CB Ultimetal (400/425 ˚C) columns.14 The problem is that none of these columns contains a chiral phase
which is needed to determine the absolute stereochemistry of the diols. The highest temperature standing chiral column is the Agilent J&W CycloSil-B that works up to 280 °C, but even this column is not compatible with the high boiling points of C28 and C30 compounds, being approximately 430 and
450 ˚C, respectively.14,15
In this research it is proposed to find a suitable analysis method to determine the absolute configuration of the mid-chain hydroxyl group of C28 and C30 1,13- and 1,15-diols. High-performance liquid
chromatography (HPLC) might be more compatible with the non-volatile character and high molecular weight of the diols, because HPLC only requires solubility of the samples.
In order to determine whether the enantiomers of the diols can be separated by chiral HPLC, first the racemic mixtures have to be obtained. However, for analysis of the stereochemistry of the diols, a synthesis route has to be contrived to synthesize the compounds enantioenriched, as biosynthetic pathways of the long chain diols are still unknown.
Attempts to enantioenrichment of aliphatic alcohols have been made by the use of asymmetric organocatalysts in the reduction of prochiral aliphatic ketones.16 The chiral oxazaboralidine catalyst
((R)-Me-CBS) showed to support enantioselective reduction (76-97% ee, absolute configuration (S)) of prochiral long aliphatic ketones due to the stable six-membered transition state that is formed during the reaction (Scheme 1a and b). Besides the fact that the catalyst is quite expensive, another downside of the use of this catalyst is that the reaction must carried out under inert atmosphere, as the reactants are air sensitive.
Scheme 1. a) Asymmetric reduction of prochiral long chain ketones with the (R)-Me-CBS catalyst. b) Coordination of the
prochiral long chain ketone to the (R)-Me-CBS catalyst leading to a six-membered transition state.16
Although, numerous organic chemists tend to perform enantioselective synthesis routes of natural compounds by developing chiral organocatalysts like (R)-Me-CBS, a wide variety of biocatalytic and bio-organic transformations have been developed which make use of enzymes to obtain chiral and enantioenriched compounds.
Enantioenrichment of alcohols via enzymatic kinetic resolution has been performed by starting with the racemic mixture of (protected) alcohols.17 In this approach, one of the enantiomers reacts, while the
other remains untouched. However, this method has a maximal yield of 50% in the case of perfect enantioselectivity. In the experiments of Yıldız and Yusufoğlu ee-values between 22 and >99% were obtained by using lipases for a wide scope of aryl substituted aliphatic alcohols (Scheme 2). One important advantage of the method is that the reaction can be performed under mild aqueous conditions,
11
making this method more environmental friendly than for instance the enantioselective reaction with the (R)-Me-CBS catalyst.
Scheme 2. Enantioselective synthesis of chiral alcohols via enzymatic hydrolysis.17
In order to perform biotransformations on organic substrates, enzymes need to be obtained in considerable amounts. Several enzymes are commercially available, but it is often much more economic to introduce the gene of the enzyme of interest in a vector, for instance an E. coli bacterium, and clone the gene in the lab. The principles of this approach are shown in Figure 4.18
Figure 4. Principles of bacterial transformation and selection (adapted from ref. 18).
In the first step recombinant circular DNA, called plasmids are added to a culture of E. coli bacteria. These plasmids contain an antibiotic resistance gene, the gene for the desired protein and a promotor sequence to drive transcription of the target gene. After that, a heat shock is given to the bacteria, which causes some of them to take up the plasmid. The plasmid used in cloning, contains an antibiotic resistance gene. So when all bacteria are introduced in an antibiotic inoculated medium, bacteria without the plasmid die, while each bacterium with the plasmid gives rise to a plasmid-containing colony. The colonies are grown in bulk so they can be used for production of the desired enzyme. The bacteria culture can be induced to express the target gene of the plasmid by a adding a gene expression inducing reagent like isopropyl β-D-1-thiogalactopyranoside (IPTG). This compound is a molecular analogue of allolactose, a lactose metabolite that triggers transcription of the lac operon, and can therefore be used to induce protein expression of genes that are controlled by the lac operator.19 If sufficient amounts of
the desired enzyme are produced, the cells can be lyophilized (i.e. freeze dried) to obtain a powder of cells that can be used directly for biotransformations or the enzyme can be isolated upon flash column chromatography.
12
The aim of this bachelor project is to involve enzymes for enantioenrichement in the synthesis of the long chain diols. In this research asymmetric reduction of prochiral ketones is proposed with the use of alcohol dehydrogenases (ADHs). These enzymes belong to the group of oxidoreductases that oxidize substrates by reducing an electron acceptor, usually NA(P)D+ or a flavin coenzyme such as FAD.20
They also catalyze the reverse reaction, for instance the reduction of prochiral ketones to alcohols. The
Sphingobium yanoikuyae DSM 6900 and Ralstonia sp. DSM 6428 ADHs will be used, as these ADHs
showed to catalyze the reduction of prochiral ketones that feature large hydrophobic groups on both sides of the carbonyl group in a high selective fashion (>99% ee). This makes it likely that the ADHs are compatible with the hydrophobic long chain diols.21,22 Additionally, the approach for enzymatic
resolution of Yıldız and Yusufoğlu by using lipases is examined for the long chain diol substrates. An overview of the approach for both the racemic and enantioselective synthesis of the C28 and C30
1,13- and 1,15-diols is depicted in Scheme 3. The total synthesis consists of a parallel six and two-step synthesis. By varying the chain length (n and m) of the reactants, the different diols can be synthesized via the same reaction pathway. The racemic route differs from the enantioselective ones in the way of reduction of the prochiral Friedel-Craft product. The racemic alcohol is obtained by reduction with NaBH4. Asymmetric reduction using ADHs or enzymatic resolution with lipases could yield the
enantioenriched product. The total synthesis is discussed in detail in the Results and Discussion.
13
3. Results and Discussion
3.1 Racemic synthesis of the C
28and C
301,13-diols.
Synthesis of the C28 and C30 1,13 and 1,15-diols was envisioned in an eight-step process starting from
an acyl chloride derivative. In this study only the C28 and C30 1,13 diols were synthesized via the
pathway depicted in Scheme 3. The individual steps of the syntheses are discussed below. Detailed experimental procedures can be found in the Experimental Section and the corresponding 1H and 13C
NMR spectra and HPLC and GC chromatograms can be found in the Appendix.
The synthesis of the long chain diols start with the Friedel-Craft acylation of thiophene. By using palmitoyl chloride (1a) or stearoyl chloride (1b), 2-thienyl ketone precursors can be synthesized for the synthesis of the C28 and C30 1,13-diol, respectively (Table 1).
Table 1. Structures of the acyl chloride starting materials for the Friedel-Craft acylation.
Acyl chloride n Intermediate for
1a 7 C28 1,13-diol
1b 8 C30 1,13-diol
The acetylation on thiophene proceeds in the presence of the Lewis acid AlCl3 at 0 ˚C under inert
atmosphere within 18 hours (Scheme 4).23 Working under inert conditions is crucial as AlCl
3 reacts fast
with water by forming Al(OH)3. Yields of 51 and 96% were obtained for the 2-thienyl ketones 2a and
2b, respectively (1H NMR:Figure 14 and 15 Appendix 9.1).
Scheme 4. Friedel-Craft acylation of thiophene using acyl chloride and AlCl3 yielding the 2-thienyl ketone (2).
The Friedel-Craft acylation of thiophene is an electrophilic aromatic substitution (EAS) that proceeds
via an acylium ion that reacts fast with the double bond of thiophene (Figure 5). Thiophene is selectively substituted on the position as the carbocation is more stabilized by resonance on the 3- than on the 2-position. The aromaticity of the thiophene ring is regenerated after hydrogen abstraction by AlCl4-.
14
In the synthesis of the racemic diols, the next step is the reduction of the thienyl ketone by NaBH4
(Scheme 5). In this reaction, the problematic solubility of the alkyl compounds 2a and b in polar solvents was improved by adding THF. However, the presence of protic solvents, such as EtOH, is crucial for the reaction to obtain the final alcohol.24
After 24 hours at room temperature 3a and 3b were obtained with a yield of 99 and 96%, respectively (1H NMR:Figure 16 and 17 Appendix 9.1, 13C NMR: Figure 25 Appendix 9.2). The NaBH
4
reduction showed to be a clean reaction, because after extraction with 0.1 M NaOH, no further purification steps were required.
Scheme 5. NaBH4 reduction of the Friedel-Craft acylation product (2) yielding the racemic 2-thienyl alcohol (3).
The hydroxyl group of the obtained 2-thienyl alcohol (3) needs to be protected with a protecting group before lithiation with n-BuLi can be performed. Silyl ethers are among the most frequently used protecting groups for alcohols.25 The reason for this is the fact that their reactivity can be tuned sterically
and electronically by changing the substituents on the silicon atom. Another advantage of the silyl ether is that it can be selectively removed by tetra-n-butylammonium fluoride (TBAF), due to strength of the Si-F bond (30 kcal/mol greater than Si-O). It is wise to choose a moderate bulky silyl ether, because bulky substituents make the formed silyl ether more resistant toward hydrolysis, but also make the formation of the silyl ether more difficult. Tert-butyldimethylsilyl (TBS) is a medium sized protecting group and has reasonable stability toward acid and good stability toward base, so tert-butyldimethylchlorosilane (TBSCl) was used for the silylation of 3.
The established procedure for the synthesis of TBS-ethers involves reaction of an alcohol with TBSCl in the presence of imidazole in N,N-dimethylformamide (DMF) at room temperature.26
However, it might be hard to remove traces of DMF afterwards. The reaction was performed in CH2Cl2
with 4-dimethylaminopyridine (DMAP) and triethylamine (TEA) to prevent this problem, following the procedure of Chaudhary and Hernandez (Scheme 6).27 Unfortunately, no conversion of the
secondary alcohol was observed after 48 hours reaction time. The silyl-DMAP intermediate formed in this reaction is likely not converted due to steric hindrance.
Scheme 6. Silylation of the hydroxyl group of 3a by TBSCl, DMAP and TEA in CH2Cl2.
The silylation reaction was then performed via the conventional protocol of Corey and Venkateswarlu by using imidazole in DMF (Scheme 7).26 A yield of 86% was obtained for the silylated alcohol 4b (1H
NMR: Figure 19 Appendix 9.1). The yield of 4a could not be determined, due to spillage of a product containing column fraction (1H NMR: Figure 18 Appendix 9.1).
15
Scheme 7. Silylation of the hydroxyl group of 3a and 3b by TBSCl and imidazole in DMF.
The mechanism of the silylation reaction with imidazole is presented in Figure 6. Imidazole attacks the silicon atom of TBSCl via a SN2 mechanism. After substitution of imidazole by the alcohol and
hydrogen abstraction by imidazole, the secondary silyl ether is obtained. Notable is that the imidazole ring never loses its aromaticity during the mechanism, due to the interplay of the nitrogen lone pairs.
Figure 6. Mechanism of the silylation of alcohols by TBSCl in the presence of imidazole.
The synthesis of the second fragment for the lithiation reaction starts with the monobromination of the commercially available 1,8-octanediol (Scheme 8). After refluxing overnight in toluene and HBr (48% (v/v) in H2O), the monobrominated product 5 was obtained with a yield of 93%.28 In this reaction one
of the hydroxyl groups is protonated by HBr, making the resulting oxonium ion a good leaving group for SN2 substitution by Br-. The di-bromo product is barely formed under these reaction conditions.
There is evidence that long chain bromo alcohols behave like surfactants and are less reactive due to the formation of aggregates, shielding the hydroxyl groups from the reagent.29
Scheme 8. Monobromination of 1,8-octanediol.
The hydroxyl group of the monobrominated product (5) was also silylated with TBSCl under the same conditions as for 4, yielding 6 with 96% yield (Scheme 9). The same protecting group was used, because this way simultaneous deprotection of both hydroxyl groups can be performed later on in the synthesis.
Scheme 9. Silylation of the hydroxyl group of 5 by TBSCl and imidazole in DMF.
The next step of the synthesis involves the alkylation of the thienyl group by lithiation with n-BuLi in THF (Scheme 10).30
16
Scheme 10. Regioselective thienyl alkylation of 4a via lithiation with n-BuLi in THF.
This reaction consists of two parts. First, the 2-thienyl silyl ether (4a) has to be deprotonated by n-BuLi at -78 ˚C (Figure 7). This yields a carbanion that is highly nucleophilic toward the electron poor carbon atom next to the bromide of the monobrominated compound 6, that is added at room temperature in the second step of the reaction. The lithiation of the thienyl compound occurs only on the 5-position, because this proton is most acidic by inductive effect of the sulfur atom.
Figure 7. Mechanism of the regioselective thienyl alkylation of 4a via lithiation with n-BuLi in THF.
Compound 7a was formed with 50% NMR yield after 48 hours. However, isolation of the compound was problematic, due to the extreme apolar character of compound 7a. Flash column chromatography had to be performed in pure PE, but even this did not yield sufficient retention on the column. Additionally to the polarity problems, partial cleavage of the primary silyl ether of the product was observed. This could have been caused by addition of an acidic chloroform-d NMR sample to the product before purification. Eventually, 7a was isolated with a yield of 11% (1H NMR:Figure 20
Appendix 9.1). From here on there was not enough material of 7a and previous intermediates to finish the synthesis of the C28 1,13-diol during this research. For further research it might be a good alternative
to deprotect compound 7, before the purification.
With this information, the lithiation reaction was performed also with the intermediate 4b (Scheme 11 a). NMR showed a conversion ratio of 1.80:1 (7b:4b) after 20 hours. It is hard to see whether enough
n-BuLi is added to the reaction, as the lithiated product is not visible on TLC. Addition of
tetramethylethylenediamine (TMEDA) to the reaction can be tried in follow-up experiments. TMEDA can coordinate with its nitrogen atom to the lithium atom, making n-BuLi more reactive than the tetra- or hexamer that n-BuLi normally adapts.31
The crude product 7b was deprotected before purification by TBAF in THF at room temperature (Scheme 11b).32 The driving force of this reaction is the formation of the strong Si-F bond.
8b was isolated by flash column chromatography in a 56% yield over two steps (1H NMR:Figure 21
17
Scheme 11. a) Regioselective thienyl alkylation of 4b via lithiation. b) TBS removal of 7b by TBAF in THF.
The final step of the synthesis of the C30 1,13-diol (9b) involves the reduction of the thiophene ring in
compound 8b. This reaction was performed with Raney-Nickel under H2 pressure (Scheme 12).33 For
this reaction addition of THF was required to dissolve 8b at elevated temperature.
Scheme 12. Reduction of the thiophene ring of 8b by Raney Ni and H2.
After 24 hours, TLC showed a wide scope of half-products (i.e. partially reduced), so an extra batch of Raney Ni was added whereupon a white precipitation appeared. This white precipitation turned out to be the desired product. 9b was isolated by filtration with warm THF in a 99% yield, making the overall yield of the synthesis of the diol 44% over six steps (Scheme 13; 1H NMR: Figure 22 Appendix 9.1; 13C NMR: Figure 26 Appendix 9.2).
18
9b has the feature to be highly insoluble in any solvent at room temperature. The product dissolves inwarm nonpolar solvents, but immediately recrystallizes after cooling down. This problematic solubility makes sense as the long chain diols are part of the cell wall layer of algae, that shields algae from toxic compounds from the environment. Solid phase 9b molecules showed to rather form aggregate kind of networks, than crystalline structures, as can be seen from the microscopic image of the 9b in Figure 8. During the synthesis of 9b it was notable that the oily products became
crystalline after cooling down from for instance rotary evaporation. Therefore it might be useful for further research to analyze the melting points of the individual and mixtures of diols, as the melting points of the racemic mixture and the enantiomerically pure compounds may differ. This analysis might provide useful information regarding the temperature relation of the long chain diol distribution within algae. A melting trajectory for the racemic C30 1,13-diol was measured
between 91.2–95.3 ˚C, indicating that the compound is indeed not crystalline, but amorphous.
3.2 Biotransformations
In this research the genes of the Sphingobium yanoikuyae DSM 6900 (Sy) and Ralstonia sp. DSM 6428 (Ras) alcohol dehydrogenases (ADHs) were expressed via bacterial transformation in E. coli. These ADHs showed to catalyze the reduction prochiral ketones that feature large hydrophobic groups on both sides of the carbonyl group in a high enantioselective fashion and are likely to be compatible with the hydrophobic long chain prochiral Friedel-Craft products (2a and b).21,22 Man et al. revealed the
structural basis for the recognition of bulky ketones by the Ras- and Sy-ADH. Both enzymes possess a hydrophobic tunnel near the surface of the enzymes that leads the bulky hydrophobic substrates to the active site (Figure 9).
Figure 9. The active sites of a) Ras-ADH and b) Sy-ADH illustrating amino acid residues of the hydrophobic tunnel (copied
from ref. 21).
After sufficient expression of the desired ADHs by bacterial cloning, the E. coli cells were lyophilized to obtain a powder of cells that could be used directly for biotransformations. Before the enzymes were exposed to the C16 and C18 substrates (2a and b), the activity of both ADHs was examined on test
substrates. The activity of Ras-ADH and Sy-ADH was tested in the reduction of the test substrates acetophenone and 2-hexanone, respectively (Scheme 14a and b). Both ADHs require a cofactor Figure 8. Microscopic image of solid 9b
illustrating the aggregating behavior of the long chain diols.
19
(NAD(P)H) to operate. Because NADH and NADPH are quite expensive, the cofactors were regenerated via the coupled recycling system comprising glucose and glutamate dehydrogenase (GDH).
Ras-ADH did convert acetophenone to phenyl ethanol with 95% in 24 hours. Sy-ADH converted 2-hexanone to 2-hexanol with 88% conversion within this time (GC Chromatograms: Figure 30 and 31, Appendix 9.4). From the results of these test reactions it can be concluded that the ADHs are active.
Scheme 14. Activity test reactions for the expressed a) Ras-ADH and b) Sy-ADH in a 50 mM tris buffer (pH 7) at 30 ˚C.
Next, the ADHs were used for the reduction of the substrates of interest, 2a and b (Scheme 15). The reaction was performed under the same conditions as the test reactions with the same coupled glucose-GDH recycling system. Because of solubility problems of the substrates, it was necessary to add a co-
Scheme 15. The reduction of 2a (n=7) and b (n=8) by Ras- and Sy-ADH in a tris buffer (pH 7) and 3% (v/v) DMF or DMSO
at 30 ˚C.
solvent. Both DMF and DMSO were examined as co-solvent. The substrates were thereupon dissolved in DMSO or DMF and were added to the reaction mixture, yielding a final concentration of 3% (v/v) co-solvent. Solubility problems were still faced for dissolving the substrates in this small amounts of organic solvent. The substrate-DMSO mixture even had to be heated to avoid precipitation of substrate before adding it to the reaction mixture.
The reactions were performed at 30 ˚C in a shaking incubator and the conversion was followed by TLC. The entries for and the results of the (control) reactions are depicted in Table 2.
a.
20
Table 2. Conditions and results of the (control) reactions of the biocatalytic reduction of 2a and b by lyophilized E. coli cells
containing Ras- and Sy-ADH. The reactions were performed at 30 ˚C for 4 days in a shaking incubator. Substrate (20 mM) ADH (30 mg) GDH (10 mg) Tris buffer (50 mM/pH 7) Glucose (60 mM) Cofactor (1 mM) Co-solvent 3% (v/v) Conversion (%)
1 2a Ras ✓ ✓ ✓ NADPH DMF No conversion
2 2a Ras ✓ ✓ ✓ NADPH DMSO No conversion
3 2a – ✓ ✓ ✓ NADPH DMF No conversion
4 2a – ✓ ✓ ✓ NADH DMSO No conversion
5 2a – – ✓ ✓ NADH DMF No conversion
6 2a – – ✓ ✓ NADPH DMSO No conversion
7 2a Sy ✓ ✓ ✓ NADH DMF No conversion
8 2b Sy ✓ ✓ ✓ NADH DMSO No conversion
9 2b Ras ✓ ✓ ✓ NADPH DMF No conversion
10 2b Ras ✓ ✓ ✓ NADPH DMSO No conversion
11 2b – ✓ ✓ ✓ NADPH DMF No conversion
12 2b – ✓ ✓ ✓ NADH DMSO No conversion
13 2b – – ✓ ✓ NADH DMF No conversion
14 2b – – ✓ ✓ NADH DMSO No conversion
15 2b Sy ✓ ✓ ✓ NADH DMF No conversion
16 2b Sy ✓ ✓ ✓ NADH DMSO No conversion
After four days the reactions were quenched because TLC did not show conversion for any of the experiments. After work-up, NMR did not show a trace of product formed. This bad conversion can be the result of the poor solubility of the substrates, 2a and b. In follow-up experiments, higher volumes of co-solvent must be examined. However it is know that organic solvent can obstruct the enzyme activity.34
Another reason for the lack of conversion can be the limited substrate acceptance of the enzymes. It is known that the Ras- and Sy-ADH accept ‘bulky-bulky’ ketones, but a n-pentyl phenyl ketone was the most bulky alkyl-aryl substrate that was converted with these enzymes in the research of Lavandera and co-workers.22 It is possible that there is a limit to the tolerance of the hydrophobic
tunnel of the Ras- and Sy-ADH for the length of the alkyl chain. This way the induced fit of the enzyme upon substrate binding may be obstructed, leading to no conversion.
The substrates might also stick in the cell membrane by passing through it. Because the substrates feature lipid kind of properties, it is possible that the substrates may stick in the lipid bilayer by nonpolar interactions and therefore cannot reach the enzymes within the E. coli cells.
A fourth possible explanation for no conversion can be the condition of the cell walls of the lyophilized cells. It is possible that cells face extreme cell wall damage after freeze-drying.35,36 There
are even examples that indicate that cell walls had collapsed and become impermeable after lyophilization. It is possible that lyophilization led to limited membrane permeability for the bulky substrates, as the test reactions with the smaller substrates did work (Scheme 14a and b).
21
In order to exclude obstruction of the biotransformations by cell wall or membrane conditions, the E.
coli cells need to be disrupted instead of lyophilized after gene expression. During disruption, the
molecules inside the cells are released by bursting of the cell wall and membrane. However, this technique requires again the process of bacterial cloning and gene expression, additional to the disruption procedure. Due to the limited time scope of the project and the uncertainty about the role of the membranes in the problematic conversion, it was decided to leave the approach of asymmetric reduction by ADHs for now.
Because there is a possibility that the ADHs are not compatible with the alkyl bulk of the substrates 2a and b, it was decided to get back on with the enzymatic resolution via lipases, as was performed by Yıldız and Yusufoğlu (vide supra Scheme 2).17 Lipases are enzymes that catalyze the hydrolysis of ester
bonds in lipids and are more likely to be compatible with the bulky hydrophobic substrates, then ADHs that usually comprise smaller substrates. In order to use lipases for the enantioselective ester hydrolysis in the synthesis of the long chain diols, the biotransformations must be performed on the acetylated form of the racemic alcohol 3b instead of the prochiral ketone 2b. 3b was acetylated with 88% yield using acetic anhydride (Ac2O), DMAP and TEA in CH2Cl2 at room temperature (Scheme 16; 1H NMR:
Figure 23 Appendix 9.1).
Scheme 16. Acetylation of the racemic alcohol 3b.
DMAP catalyzes the acetylation of alcohols by reacting with Ac2O with formation of the
acylpyridinium cationic intermediate (Scheme 17).37 The alcohol then reacts with the acylated DMAP
by forming the ester product. Without an auxiliary base, such as TEA, DMAP would be deactivated upon deprotonation of the ester product.
Scheme 17. Mechanism for the DMAP catalyzed acetylation of secondary alcohols in the presence of Ac2O and TEA.
After acetylation of 3b, yielding 10b, the acetylated substrate was examined in the biocatalytic hydrolysis with the Amano lipase (AL-PS) from the Gram-negative bacterium Burkholderica cepacia (Pseudomonas cepacia). This reaction was performed in a pH 7, 50 mM phosphate (Pi) buffer at 30 ˚C in a shaking incubator by using acetone as co-solvent for the solubility of the substrate (Scheme 18).17
22
Scheme 18. Enzymatic hydrolysis of the racemic acetylated alcohol 10b by AL-PS.
The reaction was first performed overnight. TLC indicated conversion of 10b to 3b. After work-up, the conversion could not be determined by NMR, but chiral HPLC showed 11% ee. Because it was uncertain whether this ee-value was real or just an error due to the detection limits of the HPLC apparatus, the reaction was repeated. This time with a duration of four days with four times more lipase. After four days NMR showed 45% conversion of 10b to 3b, but chiral HPLC showed that the product was racemic (HPLC Chromatograms: Figure 27 to 29 Appendix 9.3). A control experiment without lipase showed that the conversion of 10b was indeed caused by the enzyme. Thus, AL-PS is able to recognize the substrate and convert it to the chiral alcohol, but cannot do this in an enantioselective way. To explain this, it is needed to take a closer look into the active site of the lipase and the conformational movement of the alkyl-thienylsubstrate.
Molecular dynamics studies from 1997 show that the active site of the lipase consists of three well-defined pockets: a large hydrophobic (HL) pocket of circa 7 x 6.6 x 4.4 Å3, to accommodate large
hydrophobic substituents like aromatic rings, a more hydrophilic pocket (HH) of approximately 1.8 x
1.8 x 1.5 Å3 to bind the hydroxyl group and a smaller tunnel-shaped hydrophobic pocket (H
s) of 2 Å
wide and 1.9 Å high, able to interact with long side chains (Figure 10).38
Figure 10. Dimensions and shape of the active site of AL-PS with three important binding pockets for substrate specificity:
the large hydrophobic pocket (HL), a smaller tunnel-shaped hydrophobic pocket (Hs), and a more hydrophilic cavity for the
hydroxyl group (HH) a) front perspective; b) side perspective; c) three dimensional representation (copied from ref. 38).
From this active site model it looks trivial that 3b fits perfectly with the thiophene ring in the HL-pocket,
the hydroxyl group in the HH-pocket and the long alkyl chain in the Hs-region, leading to unambiguous
substrate binding and therefore to perfect enantioselectivity of the enzyme. However, the enzyme is apparently not capable of enantiodifferentiation of substrate 10b, leading to the racemic mixture.
a. b.
23
Problems with enantioselectivity of AL-PS were encountered by Grabuleda et al. for small chiral alcohols, such as 2-pentanol. The excess of room left in the active site by binding of both enantiomers, implies no clear preference in the binding mode of each, leading to a racemic mixture of the (S)- and (R)-product. 3b is not a such as small alcohol as 2-pentanol, so there might be some other factors involved that lead to the lack of enantioselectivity. As mentioned before in the case of the ADHs, the bulk of the alkyl chain might affect the binding of the substrate in the active site of the enzyme. In order to analyze conformational effects of the alkyl chain, Molecular Mechanics simulations were performed by using a MM2 force field. This simple force field is mainly used for conformational analysis of hydrocarbons and other small organic molecules, so good for this purpose.39
The MM simulation showed that at 30 ˚C there is a lot of flexibility within the alkyl chain (Figure 11). This enables the molecule to adopt a wide scope of conformations. The alkyl chain only fits within the HS-pocket of the active site if it is completely linear. The thiophene ring points then inside the larger
hydrophobic region. This leads after the reaction to one of the enantiomers of 3b. If the alkyl chain is bended it probably fits better in the HL binding pocket. This means that the thiophene ring has to point
inside the HS-pocket. To motivate this theory the results of Yıldız and Yusufoğlu are reviewed.17 Table
3 shows some of the results of the hydrolysis by AL-PS of different substrates in that research.
24
Table 3. Results of Yıldız and Yusufoğlu of the enzymatic resolution of acetylated secondary alcohols by AL-PS.17
Product X n Conversion (%) ee¤ (%)
p-tolyl 10 20 >99
phenyl 17 20 92
2-thienyl 13 19 70
n-propyl 11 45 57
¤ The ee was determined by chiral HPLC.
It can be seen from the results that the steric bulk of the X-substituent influences the enantioselectivity of the biocatalytic reaction. The bulky p-tolyl and phenyl substituents led to almost perfect enantioselectivity, whereas the significant smaller n-propyl, and even thienyl substituents experience a notable decrease in ee-values. Figure 12 shows the dimensions of the 2-thienyl and phenyl substituents. With a diameter of 2.78 Å it is unlikely for a phenyl group to fit in the small hydrophobic pocket of AL-PS that has a diameter of approximately 2.76 Å. The thienyl group, on the other hand, with a diameter of 2.31 Å fits more readily in this binding pocket. Ambiguous binding of the thienyl substituent in both hydrophobic regions of the active site, may give rise to diminished enantioselectivity of the enzyme.
Figure 12. Dimensions of the 2-thienyl and phenyl substituent.
However, this does not explain the fact that the C15-2-thienyl alcohol was converted with 70% ee by
Yıldız and Yusufoğlu, but the C18-2-thienyl alcohol (3b) was converted with no enantiomeric excess
(racemic) in this experiment. This large difference is probably not be due to the small difference in reaction conditions at which the experiments were performed (25 vs 30 ˚C in this experiment) or the by two carbon atoms extended alkyl chain of the substrate used in this experiments. However, Yıldız and Yusufoğlu only optimized the enzymatic reaction on pH, type of lipase and reaction time. It might be useful to evaluate the effect of temperature on the enantioselectivity of AL-PS in sequel research.
3.3 Chiral HPLC studies
Chiral HPLC was used for the analysis of the enantioselectivity of the AL-PS biotransformations, as well as for the separation of the enantiomers of the final C30 1,13-diol. The HPLC samples of the
AL-PS reactions were obtained by preparative TLC and showed, as mentioned before, that the enzymatic reaction yields the racemic product.
In this research it is proposed to find a suitable analysis method to determine the absolute configuration of the mid-chain hydroxyl group of 1,13- and 1,15-diols. Chiral HPLC might be more compatible with the non-volatile character of the diols than chiral GC, because HPLC only requires solubility of the samples. However, the photo-diode array (PDA) detector coupled to the chiral HPLC system present at this research institute, requires derivatization of the diols with an UV-active group.
For this purpose the primary hydroxyl group was selectively benzoylated with benzoyl cyanide (BzCN) and TEA with a yield of 47% (Scheme 19; 1H NMR:Figure 24 Appendix 9.1).40 The reaction was
25
performed on milligram scale in order to relate this protecting method to the applicability for the small amounts of diol in marine samples. The relatively low yield was mainly caused by di-benzoylation of the diol. This reaction can be optimized by choosing milder conditions, working more diluted or by shortening of the reaction time. Performing the reaction at lower temperature is not possible with respect to the solubility of 9b. It also possible to double benzoylate the diol intentionally, but this might be not advantageous for the retention time of the compound on the straight phase chiral HPLC column.
Scheme 19. Benzoylation of 9b by BzCN and TEA.
11b could be baseline separated on chiral (AD) HPLC (n-heptane:IPA, 99.5:0.5) (Figure 13). Up to
now, chiral HPLC looks like the only method to analyze the absolute stereochemistry of the non-volatile long chain diols, because chiral GC columns are not compatible with the high boiling points of these compounds.
Figure 13. HPLC-chromatogram of 11b (Chiralpak® AD (0.46 cm ⌀ x 25 cm) straight phase chiral column; 25 ° C; mobile phase n-heptane:IPA, 99.5:0.5; flow rate 1.0 mL/min; detection, 254 nm.)
Because it showed to be very difficult to isolate the individual diols from marine sediment samples, most analysis of the diols is only performed with GC-MS on the small amounts (micrograms) of diols isolated from nature.1,3,4,7 Benzoylation of small amounts of marine diols might be challenging. If the
derivatization of the diols is unsuccessful, chiral HPLC coupled to an evaporative light scattering detector (ELSD) might offer a solution. ELSD enables detection of compounds that do not absorb UV-light, such as the long chain diols.41 However, this detection method is less sensitive than a
26
4. Conclusion
The racemic C30 1,13-diol 9b could be synthesized with an overall yield of 44%. The two enantiomers
were separable by straight phase chiral HPLC upon benzoyl derivatization of the primary hydroxyl group. Hitherto does chiral HPLC look like the only method to analyze the absolute stereochemistry of the nonvolatile long chain diols, but in order to use this method for marine sediment samples of NIOZ, it is crucial that the chemical derivatization is applicable on microgram scale.
Enzymatic enantioenrichment by of the intermediates in the synthesis of the long chain diols showed to be unsuccessful. The reduction of the prochiral ketones 2a and 2b by the Ras- and Sy-ADHs showed no conversion after four days. Possible explanations for this are 1) solubility problems of the substrate; 2) bulkiness of the substrates; 3) hydrophobic interactions of the substrate with the cell membrane; 4) diminished cell wall permeability upon lyophilization.
The enzymatic hydrolysis of the acetylated chiral alcohol 10b by AL-PS yielded 45% conversion after four days, but with no enantioselectivity (racemic). Apparently there is no clear preferred binding mode of the substrate under these conditions. This ambiguous substrate binding might be explained by the dimensions of the active site of the lipase and the by MM-simulations supported flexibility of the bulky alkyl side chain of the substrate.
27
5. Outlook
This research has led to the insight that there is much more to investigate with respect to the synthesis and characterization of the long chain diols. First of all, analysis of the absolute stereochemistry of the marine samples by chiral HPLC upon benzoylation must be performed. The fact that the marine diols can only be isolated in quite small amounts (micrograms), can make chemical derivatization upon benzoylation challenging. The benzoylation of marine samples needs to be examined in sequel research. If the derivatization does not succeed, chiral HPLC coupled to an ELSD might offer a solution. The second aspect of improvement is the use of enzymes for enantioenrichment of the diols. In further research it might be useful to test whether the removal of the cell walls of the ADH containing E. coli cells, leads to successful reduction of the prochiral ketones. For the reaction with the Amano Lipase to succeed it seems to be required to improve the substrate design. Biocatalysis on 2-thienyl acetylated alcohols with less bulky alkyl chains, might result in improved enantiodiscrimination by the enzyme. A new approach for enantioenrichment of the long chain diol synthesis via enzymatic resolution is proposed in Scheme 20. This synthesis starts with the formation of the homoallylic thienyl alcohol from the commercially available allylmagnesium chloride and thiophene carboxaldehyde.42 The required
reagents are cheap and the reaction showed to give high yields. The next step is the lipase catalysed resolution of the racemic alcohol. Amano Lipase PS-C II showed to give 47% conversion of this exact substrate in a high enantioselective fashion (84% ee (R)-OH; 96% ee (S)-OAc) within 4 hours.43 The
alkyl chain can be extended in the next step via an olefin cross-metathesis reaction with the second generation Grubbs catalyst.44 Cross-metathesis in the presence of a thienyl group showed to give no
problems with the Grubbs catalyst.45 However, the selectivity of the Grubbs catalyst in the presence of
two homoallylic species must be examined. From the metathesis product, the synthesis proceeds via the same procedures as was performed in this research, starting with the lithiation reaction. The unsaturation resulting from the cross-metathesis reaction is unlikely to influence the next steps of the synthesis and is removed upon Raney Nickel reduction. The different diols can be easily synthesized upon varying of the length (n) of the alkyl chain.
Scheme 20. New approach for the enantioselective synthesis of the C28 and C30 1,13- and 1,15-diols. The different long chain
diols can be easily obtained by varying the length (n) of the alkyl chain. After the cross-metathesis reaction, the synthesis proceeds via the same procedures as was performed in this research, starting with the lithiation reaction.
28
6. Acknowledgements
This work was supported by the Synthetic Organic Chemistry (SOC) and Biocatalysis (Biocat) research group of the Van ‘t Hoff Institute for Molecular Sciences (HIMS) and the University of Amsterdam Faculty of Science. I thank Martin Wanner and Wesley Böhmer for their helpful guidance, discussion and technical support, dr. Tiddo Mooibroek for being the second reviewer of this project and prof. dr. Jan van Maarseveen and dr. Francesco Mutti for giving me the opportunity to work at their inspiring departments of the SOC and Biocat research groups. At last, I want to thank all the members of the SOC and Biocat research groups for the great working environment and making it for me a pleasure to go to the lab every day. This project gave me the opportunity to explore the interplay of different fields of chemistry and made me become a more independent and competent chemist.
29
7. Experimental Section
7.1 General procedures
Oxygen and/or water sensitive reactions were performed in flame dried glassware under inert atmosphere with dry solvents. All reagents were purchased from commercial suppliers and were used without further purification. Corresponding spectra and chromatograms can be found in the Appendix. Thin layer chromatography was performed on Merck TLC silica gel 60 F245 aluminium plates and flash
column chromatography was carried out using SiliCycle SiliaFlash® P60 silica gel (40-63 μm; 230-400
mesh). 1H- and 13C-NMR spectra were recorded on a Bruker DRX 400 MHz instrument and calibrated
on residual undeuterated solvent signals as internal standard. 1H-NMR signal multiplicities were
abbreviated as s = singlet, d = doublet, t = triplet, q = quartet, p = quintet and m = multiplet. IR spectra were recorded on a Bruker Alpha-P FTIR instrument. Melting points were determined with a Wagner & Munz Polytherm A apparatus. Enantiomeric excesses (ee) of the chiral alcohols were determined with a Shimadzu/LC20AD (Photo-diode Array detector SPD-M20A) HPLC apparatus fitted with a Chiralpak® AD (0.46 cm ⌀ x 25 cm) straight phase chiral column. GC-FID measurements were
recorded on a Agilent Technologies 7890 B GC system fitted with an Agilent J&W DB-1701 GC column (25 m x 0.25 mm x 0.25 µm in a 7 inch cage). Enzymatic reactions were performed in a Stuart®
Orbital shaking incubator SI500.
7.2 Racemic synthesis of the C28 1,13-diol
1-(thiophen-2-yl)hexadecan-1-one (2a)
Thiophene (0.50 mL; 6.3 mmol; 1.0 equiv.) and palmitoyl chloride (1.9 mL; 6.3 mmol; 1.0 equiv.) were dissolved in dry DCM (20 mL) at 0 ˚C under inert atmosphere.AlCl3 (anhydrous; 1.3 g; 9.5 mmol; 1.5
equiv.) was added in four batches. The resulting solution was stirred overnight under inert atmosphere at 0 ˚C. The reaction was quenched with saturated NH4Cl (20 mL) and HCl (2M; 25 mL) and the mixture
was extracted three times with DCM. The combined organic layers were dried over MgSO4, filtered
and concentrated in vacuo. The white solid product (2a) was isolated upon purification by flash column chromatography (PE:EtOAc, 8:1) with a yield of 51% (1.04 g; 3.22 mmol). 1H NMR (400 MHz,
Chloroform-d) δ 7.73 (dd, J = 3.8, 1.1 Hz, 1H), 7.64 (dd, J = 4.9, 1.1 Hz, 1H), 7.15 (dd, J = 4.9, 3.8 Hz, 1H), 2.91 (t, J = 7.5 Hz, 2H), 1.76 (p, J = 7.4 Hz, 2H), 1.53 – 1.15 (m, 24H), 0.90 (t, J = 6.7 Hz, 3H). 1-(thiophen-2-yl)hexadecan-1-ol (3a)
2a (0.73 g; 2.3 mmol; 1.0 equiv.) was dissolved in EtOH (15 mL) and THF (10 mL) at room temperature
under inert atmosphere. NaBH4 (50 mg; 1.9 mmol; 0.80 equiv.) was added in two equal batches at 0 ˚C.
The resulting solution was stirred overnight, allowing to warm up to room temperature. TLC (PE:EtOAc, 5:1, KMnO4 staining) showed that the reaction was not completed so another batch of
NaBH4 (10 mg; 0.38 mmol; 0.20 equiv.) was added and the mixture was stirred overnight again at room
temperature. The reaction was quenched with NaOH (0.1 M) and the mixture was extracted with diethyl ether. The combined organic layers were dried on MgSO4, filtered and concentrated in vacuo. The white
solid product (3a) was isolated with a yield of 99% (0.733 g; 2.26 mmol). 1H NMR (400 MHz,
Chloroform-d) δ 7.27 (dd, 1H), 7.07 – 6.89 (m, 2H), 4.94 (t, J = 6.7 Hz, 1H), 1.98 – 1.74 (m, 3H), 1.40 – 1.18 (m, 36H), 0.90 (t, J = 6.7 Hz, 5H).13C NMR (101 MHz, Chloroform-d) δ 148.95, 126.31, 124.10,
30
123.45, 77.32, 77.00, 76.69, 70.05, 39.18, 31.85, 30.19, 29.64, 29.60, 29.54, 29.48, 29.39, 29.34, 29.30, 29.18, 25.70, 22.61, 14.02.
tert-butyldimethyl((1-(thiophen-2-yl)hexadecyl)oxy)silane (4a)
Small scale 3a (0.16 g; 0.50 mmol; 1.0 equiv.) was dissolved in dry DMF (2 mL). Imidazole (0.17 g; 2.5 mmol; 5.0 equiv.) and TBSCl (0.19 g; 1.3 mmol; 2.5 equiv). were added in that order. The resulting solution was stirred over the weekend at room temperature under inert atmosphere. The reaction was quenched with sat. NH4Cl (10 mL) and water (10 mL). The mixture was extracted three times with
diethyl ether. The combined organic layers were washed with brine, dried on MgSO4, filtered and
concentrated in vacuo. Purification was performed upon flash column chromatography (PE:EtOAc, 30:1) (see big scale). Big scale 3a (0.32 g; 0.97 mmol; 1.0 equiv.) was dissolved in dry DMF (10 mL). Imidazole (0.33 g; 4.9 mmol; 5.0 equiv.) and TBSCl (0.37 g; 2.4 mmol; 2.5 equiv.) were added in that order. The resulting solution was stirred overnight at room temperature under inert atmosphere. TLC (PE:EtOAc, 9:1) showed that the reaction was not finished, so more imidazole (0.03 g; 0.5 mmol; 0.5 equiv.) and TBSCl (0.07 g; 0.5 mmol; 0.5 equiv.) was added. The mixture was stirred again overnight and was then quenched with sat. NH4Cl (10 mL) and water (10 mL). The mixture was extracted three
times with diethyl ether. The combined organic layers were washed with brine, dried on MgSO4, filtered
and concentrated in vacuo. The brown oily product (4a) was isolated upon flash column chromatography (PE:EtOAc, 30:1). The yield could not be determined, due to spillage of a product containing fraction. 1H NMR (400 MHz, Chloroform-d) δ 7.19 (dd, J = 5.2 Hz, 1H), 6.93 (dd, J = 5.0,
3.5 Hz, 1H), 6.86 (d, J = 3.4 Hz, 1H), 4.92 (t, J = 6.2 Hz, 1H), 1.88 – 1.66 (m, 2H), 1.45 – 1.19 (m, 34H), 0.91 (s, 14H), 0.16 – 0.02 (m, 4H), -0.05 (s, 3H).
1-bromo-8-octanol (5)
To 1,8-octanediol (5.2 g; 36 mmol; 1.0 equiv.) in toluene (80 mL), concentrated HBr (48% v/v in water, 5.0 mL; 45 mmol; 1.3 equiv.) was added dropwise. The resulting mixture was put under reflux for 20 hours. Next, water (80 mL) was added and the mixture was extracted with diethyl ether. The combined organic layers were washed with NaOH (0.1 M in H2O; 100 mL), dried on MgSO4, filtered and
concentrated in vacuo. The product (5) was isolated upon purification by flash column chromatography (PE:EtOAc, 1:1, I2 staining) with a yield of 93% (6.94 g; 33.2 mmol). The analytical data was consistent
with literature.28
((7-bromoheptyl)oxy)(tert-butyl)dimethylsilane (6)
5 (6.9 g; 33 mmol; 1.0 equiv.), imidazole (5.7 g; 83 mmol; 2.5 equiv.) and TBSCl (6.0 g; 40 mmol; 1.2
equiv.) were dissolved in DMF (10 mL). The resulting mixture was stirred overnight at room temperature under inert atmosphere. TLC showed that the reaction was not finished, so more TBSCl (2.5 g; 17 mmol; 0.50 equiv.) was added and the mixture was stirred for another three hours. The reaction was quenched with sat. NH4Cl and the mixture was extracted with diethyl ether. The combined
organic layers were washed with brine, dried on MgSO4, filtered and concentrated in vacuo. The crude
product was purified upon flash column chromatography (PE:EtOAc, 95:5, I2 staining), yielding a
viscous colourless liquid (6) with a yield of 96% (10.2 g; 31.5 mmol). The analytical data was consistent with literature.46
31
t-butyl((8-(5-(1-((tert-butyldimethylsilyl)oxy)hexadecyl)thiophen-2-yl)octyl)oxy)dimethylsilane (7a)
n-BuLi (in hexanes; 0.32 mL; 0.81 mmol; 1.1 equiv.) was added dropwise to a mixture of 4a (0.32 g;
0.74 mmol; 1.0 equiv.) in THF (2 mL) under inert atmosphere at -78 ˚C. The mixture was stirred for an hour at -78 ˚C. Next, 6 (0.24 mL; 0.74 mmol; 1.0 equiv.) was added, after which the mixture was allowed to warm up to room temperature and the mixture was stirred for 48 hours. The reaction was quenched with water and extracted three times with diethyl ether. The combined organic layers were dried on MgSO4, filtered and concentrated in vacuo. . The crude product was purified upon flash column
chromatography (PE:EtOAc, 100:0 to 50:1), yielding a yellow oily product (7a) with a yield of 11% (60.6 mg; 31.5 mmol). 1H NMR (400 MHz, Chloroform-d) δ 6.76 (d, J = 3.5 Hz, 1H), 6.61 (d, J = 3.3
Hz, 1H), 4.83 (t, J = 8.9, 6.9 Hz, 1H), 3.60 (t, J = 6.6 Hz, 2H), 2.77 (t, J = 7.5, 6.6 Hz, 2H), 1.93 – 1.73 (m, 3H), 1.73 – 1.58 (m, 2H), 1.54 – 1.46 (m, 1H), 1.43 – 1.16 (m, 44H), 0.89 (s, 15H), 0.05 (s, 5H).
7.3 Racemic synthesis of the C30 1,13-diol
1-(thiophen-2-yl)octadecan-1-one (2b)
Thiophene (0.50 mL; 6.3 mmol; 1.0 equiv.) and stearoyl chloride (2.1 mL; 6.3 mmol; 1.0 equiv.) were dissolved in dry DCM (20 mL) at 0 ˚C under inert atmosphere. AlCl3 (anhydrous; 1.3 g; 9.5 mmol; 1.5
equiv.) was added in four batches. The resulting solution was stirred overnight under inert atmosphere at 0 ˚C. The reaction was quenched with HCl (2M; 40 mL) and the mixture was extracted three times with DCM. The combined organic layers were dried over MgSO4, filtered and concentrated in vacuo.
and analyzed by 1H-NMR. The white solid product (2b) was isolated upon purification by flash column
(PE:EtOAc, 10:1) with a yield of 96% (2.13 g; 6.08 mmol). Rf 0.74 (PE:EtOAc, 10:1). 1H NMR (400 MHz, Chloroform-d) δ 7.73 (d, J = 3.8 Hz, 1H), 7.64 (d, J = 4.8 Hz, 1H), 7.15 (t, J = 4.4 Hz, 1H), 2.91 (t, J = 7.5 Hz, 2H), 1.76 (p, J = 7.4 Hz, 2H), 1.46 – 1.23 (m, 38H), 0.90 (t, J = 6.6 Hz, 3H).
1-(thiophen-2-yl)octadecan-1-ol (3b)
2b (0.50 g; 1.4 mmol; 1.0 equiv.) was dissolved in EtOH (15 mL) and THF (10 mL) at room temperature
under inert atmosphere. NaBH4 (43 mg; 1.1 mmol; 0.80 equiv.) was added in two equal batches at 0 ˚C.
The mixture was stirred overnight, allowing to warm up to room temperature. The reaction was quenched with NaOH (0.1 M) and the mixture was extracted with diethyl ether. The combined organic layers were dried on MgSO4, filtered and concentrated in vacuo. The white solid product (3b) was
isolated with a yield of 96% (0.846 g; 1.37 mmol). Rf 0.51 (PE:EtOAc, 10:1). 1H NMR (400 MHz, Chloroform-d) δ 7.26 (dd, J = 4.7, 1.8 Hz, 1H), 7.04 – 6.94 (m, 2H), 4.94 (q, J = 6.2, 5.7 Hz, 1H), 1.94 – 1.77 (m, 2H), 1.42 – 1.20 (m, 26H), 0.90 (t, J = 6.6 Hz, 3H). HPLC (Chiralpak® AD (0.46 cm ⌀ x 25
cm) straight phase chiral column; 25 ° C; mobile phase n-heptane:IPA, 99.0 : 1.0; flow rate 1.0 mL/min; detection, 254 nm.): 11.7 and 12.9 min.
32
tert-butyldimethyl((1-(thiophen-2-yl)octadecyl)oxy)silane (4b)
3b (1.5 g; 4.2 mmol; 1.0 equiv.) was dissolved in dry DMF (20 mL), whereupon imidazole (1.4 g; 21
mmol; 5.0 equiv.) and TBSCl (1.6 g; 11 mmol; 2.5 equiv.). The resulting mixture was stirred overnight at room temperature under inert atmosphere. TLC (PE:EtOAc, 9:1) showed no full conversion, so more imidazole (0.14 g; 2.1 mmol; 0.50 equiv.) and TBSCl (0.32 g; 2.1 mmol; 0.50 equiv.) was added. The mixture was stirred for another two hours and after that, the reaction was quenched with sat. NH4Cl (10
mL) and water (20 mL). The mixture was extracted three times with diethyl ether. The combined organic layers were washed with brine, dried on MgSO4, filtered and concentrated in vacuo. The yellow
oily product (4b) was isolated upon purification by flash column chromatography (PE:EtOAc, 100:0 to 100:1) with a yield of 86% (1.68 g; 3.61 mmol). 1H NMR (400 MHz, Chloroform-d) δ 7.19 (d, J = 4.9
Hz, 1H), 6.93 (dd, J = 5.1, 3.4 Hz, 1H), 6.86 (d, J = 3.4 Hz, 1H), 4.93 (t, J = 6.3 Hz, 1H), 1.93 – 1.65 (m, 2H), 1.46 – 1.19 (m, 31H), 1.03 – 0.82 (m, 13H), 0.07 (s, 3H), -0.04 (s, 3H).
t-butyl((8-(5-(1-(tert-butyldimethylsilyl)oxy)octadecyl)thiophen-2-yl)octyl)oxy)dimethylsilane (7b)
n-BuLi (in hexanes; 0.82 mL; 2.0 mmol; 1.2 equiv.) was added dropwise to a mixture of 4b (0.80 g;
1.7 mmol; 1.0 equiv.) in THF (5 mL) under inert atmosphere at -78 ˚C. The mixture was stirred for an hour at -78 ˚C. Next, 6 (0.55 mL; 1.7 mmol; 1.0 equiv.) was added. The resulting mixture was allowed to warm up to room temperature and was stirred overnight. The reaction was quenched with water and extracted with three times with diethyl ether. The combined organic layers were dried on MgSO4,
filtered and concentrated in vacuo, yielding a yellow oily crude product. No further purification steps were performed.
1-(5-(8-hydroxyoctyl)thiophen-2-yl)octadecan-1-ol (8b)
