Bachelor Thesis Chemistry
Synthesis, Characterization and Recognition Properties of a New
Water-soluble Zn
II-based Bowl Complex Modeling the Tris-Histidine
Active Site Encountered in Many Mononuclear Metalloenzymes
Marianne Lankelma (Universiteit van Amsterdam 10244107, Vrije Universiteit 2517935) Prof. Dr. Olivia Reinaud
Dr. Olivia Bistri Prof. Dr. Joost Reek
Université Paris Descartes
Laboratoire de Chimie & Biochimie Pharmacologiques et Toxicologiques (LCBPT) Group: Supramolecular Bioinorganic Chemistry (CNRS UMR 8601)
April - June 2014
2
Acknowledgements
I would like to thank Prof. Dr. Olivia Reinaud for giving me the opportunity to do my undergraduate internship in the équipe Reinaud and for wise advice. I am grateful to Dr. Olivia Bistri for her thorough supervision, continuous cheerfulness and numerous wise lessons. Arnaud Parrot MSc is acknowledged for keeping an eye on me and my argon bubbles in the eastwing laboratory and for advice every now and then. To all the others of the équipe Reinaud (Benoît Colasson, Diana Over, Jean-Noël Rebilly, Gaël de Leener, Huy Vo, Marina El Gohary, Doris Eid and Léa van Kerkhoven) I am thankful for answering all my where-can-I-find-questions and for speaking in English for me. Dr. Assia Hessani is acknowledged for recording mass spectra. Prof. Dr. Joost Reek is acknowledged for helping me to arrange this project, for wise advice and for being my supervisor at the University of Amsterdam. To Dr. Sape Kinderman I am grateful for being my independent corrector. The Erasmus Programme is acknowledged for financially supporting Parisian rent.
3
Table of contents
Acknowledgements………....p. 2 List of abbreviations……….. p. 3 Abstract……….p. 4 Nederlandse samenvatting voor scholieren van 6VWO………...……...p. 5 1 Introduction………...p. 6 2 Results & Discussion………..p. 11
2.1 Synthesis of Rim4(OH)4………p. 11
2.2 Complexation to ZnII in polar solvents……….………p. 15
2.3 First trial in catalytic reactivity……… p. 16 3 Conclusions………..………p. 17 References……….………p. 18 Supplementary Information………...p. 19 General experimental methods………p. 19
A) Synthesis of Rim4(OH)4………...p. 19
B) Synthesis of WRim3……….p. 23
List of abbreviations
br = broad d = doublet DCM = dichloromethane, CHCl2 DMA = dimethylacetamide DMF = N,N-dimethylformamide ESI+ = electrospray ionization HR = high resolutionLR = low resolution m = multiplet MeOH = methanol MS = mass spectroscopy
NMR = Nuclear Magnetic Resonance s = singlet
THF = tetrahydrofuran t = triplet
4
Abstract
The synthesis of models for metalloenzyme active sites is the key to the understanding of the fundamental mechanisms of the biocatalytic cycles involved, and to the design of selective and efficient tools for synthetic chemistry and catalysis. Herein is reported the synthesis of a new water-soluble, bowl-shaped model for the tris-histidine active site encountered in many mononuclear metalloenzymes. Both the three histidines as well as the basic amino acid which is often encountered in the proximity of the triad, were mimicked with imidazole. The hydrophobic cavity, which protects the active site and pre-organizes substrates toward the active metallic species, was imitated with a propanol-footed resorcin[4]arene. The resulting
ligand, abbreviated Rim4(OH)4, was proven to coordinate to ZnII and the resulting
organometallic complex was observed to be soluble in polar solvents. The complex of Rim4(OH)4 with ZnII was characterized with NMR spectroscopy (1H, 13C, 1H/1H, 1H/13C
5
Nederlandse samenvatting voor scholieren van 6VWO
De active site van een enzym is de plek waar de katalyse plaatsvindt. Om synthetisch dezelfde reacties te kunnen katalyseren die de natuur zo goed kan katalyseren, proberen chemici wereldwijd om modellen voor active sites te maken.
In veel metalloenzymen bestaat de active site uit drie histidine aminozuren aaneengekoppeld, met daarnaast een basisch aminozuur dat de katalyse helpt door een proton van een zuur substraat op te nemen. In de groep van Prof. Olivia Reinaud (Université Paris Descartes, Frankrijk) waar ik mijn bachelorproject heb gedaan werkt men aan verschillende modellen voor deze active site. En het maken van zo’n model is zo simpel nog niet. Gewoon een potje histidine kopen en er een stel aan elkaar koppelen levert je namelijk geen goed model voor een polyhistidine active site. Er zijn namelijk meerdere kernpunten om bij de synthese van zo’n model belang aan te hechten, want a) de aminozuren moeten nagebootst worden, b) de vouwing van het enzym moet geimiteerd worden, bijvoorbeeld zodat de active site beschermd wordt door een hydrofobe ‘pocket’ en opdat de gewenste substraten richting de active site worden geduwd, en c) je model moet natuurlijk ook nog onder fysiologische condities (dat wil voornamelijk zeggen, bij pH 7.4) werken, net zoals het originele enzym! Wat betreft het imiteren van histidine is puntje a een makkelijke: het molecuul imidazole lijkt namelijk als twee druppels water op histidine. Puntje b is al lastiger, maar niet onhaalbaar: met behulp van macrocyclische, hydrofobe moleculen kun je de hydrofobe pocket van een enzym nabootsen en via die weg met behulp van zwakke interacties controle over je imitatie-aminozuren hebben. Zogeheten calixarenen en resorcinarenen zijn de twee moleculen die in onze groep die hiertoe gebruikt worden. En dan is er als laatste nog puntje c, het laten werken van je model bij pH 7.4: hiervoor is het natuurlijk handig als je molecuul wateroplosbaar is. De enige manier om dit te beinvloeden is eigenlijk door de hydrofobiciteit van de macrocylische kroon te compenseren door er polaire staarten aan vast te zetten.
Het doel van mijn project was om een resorcin[4]arene (d.w.z. een resorcinarene met vier aromatische ringen) aan zijn bovenkant te functionaliseren met vier imidazole moleculen om vier histidine aminozuren te imiteren en om aan de onderkant vier polaire staarten te hangen, om de wateroplosbaarheid van het totale molecuul te vergroten.
Toen dit molecuul eenmaal gemaakt was kon ik het
complexeren met zink en het verkregen
organometaalcomplex gaan testen op zijn capaciteit om een amide te hydrolyseren.
6
1
Introduction
The synthesis of adequate models for the active sites of metalloenzymes is of fundamental importance for understanding the catalytic cycles facilitated by those active sites and for the
design of new catalysts.1 In the active sites of many metalloenzymes, in particular in many
hydrolytic zinc enzymes, the first coordination sphere surrounding the metal (a tripodal cofacial core) is handled by three histidines featuring two labile
coordination sites in cis-position (Figure 1). 2 The resulting
pentacoordinate organometallic complex is the key to the catalytic activity of the enzyme, e.g. in hydrolytic zinc enzymes, where the metal enhances the hydrolytic activity via electrophilic assistance and
nucleophilic activation through Lewis acid behavior. 1, 2
In the classical approach for the synthesis of biomimetic models for the active sites of metalloenzymes, it is only this first
coordination sphere which is reproduced.2 This strategy, however,
does not allow for a full reproduction of the complexity of metalloenzyme active sites, because the type of reactivity performed by the active site does not depend solely on the organometallic complex, but also on the cavity surrounding the complex, which protects the complex, controls its nuclearity, pre-organizes substrates towards the active metallic species, orientates substrates so as to favor regio-, chemio- and enantioselective processes and favors
the release of products. 2 Therefore, association of the organometallic complex to a cavity is
essential to properly mimic a metalloenzyme environment.
The cavity performs many of its functions through weak interactions (hydrophobic
forces, hydrogen bonding, CH/π interactions) and is hence able to perform his supramolecular influence on the organometallic complex only when the complex is in its proximity. Previously reported approaches to mimic metalloenzyme active sites did not take this important aspect into consideration, yet in the group of Prof. O. Reinaud, various models have been reported in which the complex is genuinely embedded inside a cavity to
enable supramolecular control of the complex. 2
The four main classes of macrocyclic receptors that are widely used to function as this type of cavity are cyclodextrins, cyclotriveratrylenes, calix[n]arenes and resorcin[n]arenes, which differ from each other in terms of shape, size, symmetry patterns, rigidity and hosting
properties, hence each class is valuable for different types of biomimetic models. 3
Over the past decade, the group of Prof. Reinaud has extensively studied the use of calix[n]arenes and resorcin[n]arenes (Figure 2) for the synthesis of supramolecular
biomimetic systems. 4, 5 Receptors based on calix[4]arenes have already been developed and
optimized by other research groups, as they are easily constrained in a certain conformation
through small rim per-O-alkylation. 6 Since their small rim is too narrow for the majority of
organic guest molecules to pass through, they have been applied mainly as platforms for the
Figure 1 - Schematic representation of a polyhistidine active site of a metalloenzyme
7
pre-organization of binding sites outside of the calix[4]arene cavity. 6 Calix[6]arenes,
although highly flexible due to the facile ring inversion of their aromatic units, have been reported by Reinaud et al. to function as hydrophobic cavities for biomimetic models as well.
3 To constrain calix[6]arenes into well-defined cones, their small rim can be functionalized
with three amino arms on alternate phenolic positions, to form a first coordination sphere able to bind a metal ion, resulting in organometallic complexes referred to as ‘funnel complexes’ (Figure 3, left). 3
Figure 2 - Calix[6]arene (left) and resorcin[4]arene (right)
Figure 3 - Schematic representation of a funnel-shaped complex (left) and a bowl-shaped complex (right) These funnel-shaped complexes present a labile coordination site orientated inside the cavity. 3 However, they allow only for trans substrate access, while many enzymes, e.g.
proteases and nucleases, involve in their catalytic cycles the formation of intermediates where two labile metal sites are occupied by reactants, intermediates (often anionic) or
products in cis-arrangement. 7 Therefore, for the modeling of such enzymes, the recently
developed resorcin[4]arene-based complexes provide more outcome than calix[n]arenes, since their geometry facilitates cis substrate access (Figure 3, right).
A bowl-shaped complex recently developed by Reinaud et al. is Rim3, consisting of a
resorcin[4]arene of which the large rim (i.e. the upper rim) is functionalized with three imidazole ligands, as imidazole is quite similar to histidine. First studies on coordination of Rim3 to ZnII, CuI and CuII highlighted remarkable features prone to selective recognition and
reactivity processes. 4, 5, 8 Indeed, the ZnII-complex of this Rim3 ligand presents two labile
coordination sites in cis-position on the metal (Figure 4). 4, 5, 8 First insights into recognition
properties demonstrated that the monocationic ZnII complex was able to bind a large scope
of ligands differing in size, denticity and pKa. 4, 5, 8 This emphasized the outstanding
8 molecules, exo-binding of relatively large substrates, and regioselective coordination of bidentate ligands (Figure 4).
With this new ligand, the group of Prof. O. Reinaud now disposes of an adequate active site model, consisting of a metal ion coordinated to a facial triad that presents two coordination sites in cis-position, under the control of a hydrophobic cavity, as it is also the case for the above cited metalloenzymes (Figure 1).
Figure 4 - Coordination and substrate recognition properties of the Rim3 cavitand
Research currently carried out by A. Parrot (2nd year Ph.D. candidate) concerns the
investigation of an organosoluble tetra-imidazole substituted resorcin[4]arene (abbreviated
Rim4)(Figure 5), in order to mimic the basic amino acid often encountered in the proximity of
the classical triad coordination pattern.
Cavity with an intrinsic base
O O OO O OO C5H11 C5H11 O N N N N O O N N C5H11 C5H11 O O N N N N N ZnII Lin N Zn(II) + HX N N N ZnII X NH Rim4
Figure 5 - Rim4, a new bowl-shaped ligand presenting an intrinsic base
As was previously explained, coordination of a metal ion both to a first coordination sphere mimicking the amino acids and to a hydrophobic cavity provides many advantages for the modeling of active sites. It allows for control of the nuclearity of the metal ion, constrains the cavity into a well-defined shape, favors coordination of exogenous ligands and results in functional selectivity for interactions between the metal ion and exogenous
ligands. 2 However, one aspect is still difficult to imitate. Enzymes function under
physiological conditions, thus in an aqueous environment. Therefore, solubility of an active site model in water is highly desirable, yet most of the biomimetic complexes that have been reported to reproduce a specific active site are not soluble in water. The main cause for this is the hydrophobic cavity, which strongly reduces the water-solubility of the complex. Besides
9 that, it is difficult to control the reactivity of a metal ion in water due to its propensity to form hydro/oxo species that readily undergo dimerization, polymerization and precipitation. Furthermore, water prevents recognition of guest molecules and is highly competitive with all non-covalent interactions, including electrostatic charges, dipole-dipole bonding and hydrogen bonding. And self-evidently, the synthesis of a receptor simultaneously allowing for metal coordination, protection against water and water-solubility is a difficult task both from a design and synthetic point of view. It is because of all of these reasons that only few
water-soluble active site models have been reported so far by the group of Prof. Reinaud. 9, 10
Recently, the group of Prof. Reinaud has obtained a water-soluble analogue of the
previously discussed Rim3 ligand (abbreviated WRim3)(Figure 6), which demonstrated
coordination to zinc in water. The recognition properties of the resulting organometallic complex are currently under study.
Figure 6 – WRim3, the water-soluble analogue of the Rim3 ligand
The primary goal of this project was to synthesize also a water-soluble analogue of the
previously discussed Rim4 ligand (abbreviated Rim4(OH)4). To accomplish this
water-solubility, the eight-step synthesis (Figure 7) starts with condensation of resorcinol with 2,3-hydrofuran, in order to a start from a propanol-footed resorcinarene instead of one which has alkyl pendant groups. Subsequently, the protons positioned at the large rim are
substituted with bromides, the large rim is rigidified and the pendant groups are protected.11
In the three steps that follow, the large rim is functionalized with four imidazoles, which are
connected to the resorcinarene via four –CH2-O-CH2- linkers, because this type of linker has
been reported to be sufficiently long to allow all imidazoles to bind the same metal ion, but
short enough to favor intracavity coordination. 4 Finally, the propanol pendant groups can be
deprotected.
This thesis discusses the synthesis of Rim4(OH)4), as well as studies on its solubility in
polar solvents, its ability to coordinate to zinc in polar solvents and a first trial in reactivity in polar solvents.
10 Figure 7 - Overview of the synthesis of Rim4(OH)4
11
2
Results & Discussion
2.1 Synthesis of Rim4(OH)4
2.1.1 Synthesis of the dodecol
The 8-step synthesis of Rim4(OH)4 starts with condensation of resorcinol with
2,3-dihydrofuran as described by Sherman and co-workers. 11 In this reaction, hydrochloric acid
functions as a catalyst. The reason why not three, nor five, nor six molecules of resorcinol are coupled is the fact that the formation of the resorcin[4]arene is thermodynamically
favourable. Due to the meta-positioning of the hydroxyl groups of resorcinol, a C2v dodecol
also forms, but fortunately, this side product has different solubility properties than the desired dodecol and is therefore facile to discard in the work-up. After the work-up, it is useful to dry the dodecol thoroughly in a lyophilizator, since the ligand is very polar at this stage and therefore likely to encapsulate water in its cavity, yet it was proven not to be essential in order to brominate the dodecol successfully in the next step.
2.1.2 Benzylic bromination of the dodecol
In the second step, also described by Sherman et al., the four aromatic protons at the large
rim (Ar-H up) are replaced by four bromides via electrophilic aromatic substitution. 11
Bromides are supplied by N-bromosuccinide (NBS), in which the bromide atom possesses over a partial positive charge, making it a good electrophile. NBS is sensitive to light, thus the reaction must be carried out in the dark.
Besides the desired tetrabromoresorcinarene 2, a small amount of tribromo-substituted derivative is also likely to form, yet it is impossible to separate these two
12 products using silica gel flash column chromatography, because product 2 is so polar that it would probably not elute with classical eluents, but instead stay behind on the silica gel used as the stationary phase of the column. Therefore, the side product is not removed until the next step, after which the polarity of both products is reduced.
2.1.3 Rigidification by formation of methylene bridges at the large rim
Subsequently, methylene bridges are formed between the eight hydroxyl groups positioned at the large rim, in order to rigidify the resorcinarene bowl. 11 The final ring closure (i.e. the
formation of the fourth methylene bridge) requires a relatively high energy input, therefore temperature is gradually increased from 25°C to 45°C and finally to 65°C.
2.1.4 Protection of the propanol pendant groups
In order for the propanol pendant groups not to react with any of the reagents used in the following steps, it is essential to protect them. The protecting group should be resistant to both n-butyllithium and lithium aluminum hydride. Additionally, deprotection should be possible without affecting the first coordination sphere. Silyl ethers fit both of these criteria and were therefore used as temporary substituents of the primary alcohols of the pendant groups.
13
2.1.5 Substitution of the four bromides by four methyl esters
Four bromide-lithium exchanges occur in this reaction. The resulting carbanion attacks the carbonyl of methyl chloroformate, to form the tetraester depicted above.
In a first trial, 5 equivalents of n-butyllithium and 15 equivalents of methyl chloroformate were used, successfully yielding the tetraester derivative but also triester as a side product. In a second trial, higher equivalents (10.5 and 11.1 eq. respectively) were applied based on work by Bryant et al., upon which pure tetraester was obtained in a yield of
99%.12
The resulting tetraester-functionalized resorcinarene can be used without further purification in the next step.
2.1.6 Reduction of the methyl esters
Using lithium aluminum hydride, the methyl esters are reduced to primary alcohols. After the reduction, the excess of lithium aluminum hydride is allowed to react with ethyl acetate.
A remarkably low yield was obtained, as has always been the case for this step in previous trials. This could be due to difficulties to extract the product from lithium and/or aluminum salts.
Therefore, it would be desirable to omit this reaction step and its associated purification column. This could be achieved by exchange of the bromides of the tetrabromo derivative with lithium ions (using n-BuLi as usual), directly followed by primary alcohol
functionalization using paraformaldehyde, the polyoxymethylene polymer of
14
2.1.7 Imidazole functionalization of the tetraol resorcinarene
Using sodium hydride, the four primary alcohols are deprotonated in DMF, upon which each oxygen atom is supposed to attack 2-chloromethyl-1-methyl-1H-imidazole at its relatively electropositive CH2. In practice, even when a large excess of
2-chloromethyl-1-methyl-1H-imidazole is added, not all four positions are easily functionalized with imidazole at once. The procedure of deprotonation followed by imidazole functionalization had to be repeated in order to obtain solely Rim4, without Rim2 and/or Rim3 as side products.
2.1.8 Deprotection of the propanol pendant groups
Tetrabutylammoniumfluoride (TBAF) is a common reagent to remove silyl ethers, yet several difficulties have been encountered when it was previously used as a deprotection
agent for Rim4. The problems mainly concerned the difficult removal of TBAF from the
product and the partial water-solubility of Rim4(OH)4, which impeded full recovery of the
product after the aqueous extraction required for the TBAF deprotection method.
Trifluoroacetic acid (TFA) was tested as a new deprotection agent, since it produces less side products and does not require an aqueous extraction.
15 After deprotection of the propanol pendant groups with TFA, it is important to treat
the ligand in methanol with a basic resin such as DOWEX OH-, in order to deprotonate the
imidazoles, which will have been most likely protonated by TFA.
2.2 Complexation to ZnII in polar solvents
Upon treatment of a solution of Rim4(OH)4 in deuterated methanol ([Rim4(OH)4]
= 3 mM) with one equivalent of zinc nitrate, ZnII was observed to coordinate to
Rim4(OH)4. This could be derived using 1H-NMR (Figure 9) from an upfield shift
of the singlet of Hα of the imidazoles (Figure 8), a downfield shift of the singlet of
Hβ of the imidazoles and a downfield shift of the singlet of the methyl group of
the imidazoles, indicating that zinc coordinated to Nα, consequently pulling Hα
into the cavity (hence making it more shielded by electrons) and directing Hβ and the
methyl group more outwards (hence making them more deshielded).
Figure 9 - Lower spectrum: 1D 1H-NMR of Rim4(OH)4 (500 MHz, CD3OD, 300 K, [Rim4(OH)4] = 3 mM)
Upper spectrum: 1D 1H -NMR after addition of 1 eq. Zn(NO3)2.6H2O(500 MHz, CD3OD, 300 K, [Rim4(OH)4] = 3 mM)
After evaporation of deuterated methanol, deuterated water was added and ZnII was
observed to still coordinate to Rim4(OH)4, or at least, at the natural pD of this solution, which
was measured to be 6.68. In a more acidic environment, however, no more zinc coordination was observed, which is most likely because the imidazoles become protonated at their Nα.
This was derived from the fact that the singlets of Hα and Hβ return to similar chemical shifts
upon lowering of the pD. Zinc coordination was neither observed to occur at a pD of 7.54, at
which a white precipitate forms (most likely Zn(OH)x).
Figure 8 – Imidazole and its Hα, Hβ and Nα
16 It would be interesting to determine what the maximum pD value is at which zinc still coordinates to Rim4(OH)4. In the ideal case, this pD value would be as close as possible to
physiological pH (=7.4).
2.3 First trial in reactivity
Now that zinc was proven to coordinate to Rim4(OH)4, the next step was to investigate
whether any hydrolytic reactivity was possible with the resulting complex. Since the hydrophobic cavity had been previously demonstrated to have a high affinity for the acetate anion in organic solvents, the hydrolysis of 4-nitroacetanilide was studied, because acetate is one of the two products which is supposed to form in this reaction. The hydrolysis test was performed in a 1:1 mixture of deuterated water and deuterated methanol ([Rim4(OH)4] = 3.2 mM), as was the blanco test in which methyl imidazole
replaced the Rim4(OH)4 ligand. This study was started, but will be pursued
by A. Parrot (2nd year Ph.D. candidate).
Figure 9 – 4-nitroacetanilide
17
3
Conclusions
Product 8, the intended tetraimidazole-functionalized resorcin[4]arene Rim4(OH)4, was
successfully obtained in a final quantity of 138 mg. Many of the steps of the synthesis have been carried out multiple times, hence several batches of intermediates have been left, too. The yields obtained in the first three steps were comparable to those obtained in the literature 11and the same accounts for the yields obtained in step 4 – 7 relative to those
obtained before in the group of Prof. Reinaud.
Several aspects of the synthesis still call for improvement (in particular an increase of yield in the reduction of the methyl esters would be highly desirable), but the following small improvements have been achieved:
Separation of product 1 from the undesired C2v dodecol which forms as a side
product, by taking the crude product into THF and filtering off the insoluble material (i.e. the desired C4v dodecol).
The use of the equivalents of n-butyllithium and methyl chloroformate as reported by
Bryant et al. in the step of methyl ester introduction, which resulted in a significant improvement of the yield.
The use of TFA instead of TBAF for the step of deprotection of the propanol pendant
groups, which resulted in less loss of product and less side products.
The Rim4(OH)4 ligand itself was observed not to be soluble in deuterated water, but
solubility both in deuterated water as well as in deuterated methanol was achieved upon complexation with ZnII. A decrease in pD resulted in de-coordination of ZnII from Rim4(OH)4,
and increasing the pD to basic values resulted in the formation of a precipitate which is most likely Zn(OH)x.
Further studies are required to fully explore the properties of the ZnII- complex of
Rim4(OH)4. Most importantly, the reactivity trials will be continued in order to determine the
scope as well as the performance of the complex, as well as to ascertain the optimal conditions for reactivity. Besides that, it would be beneficial to experimentally confirm the hypothesis that ZnII indeed coordinates to all four imidazoles just like it coordinated to all
four imidazoles of the organosoluble Rim4 ligand. It would also be interesting to investigate
complexation of Rim4(OH)4 to copper. Thus, there remain many possibilities open for future
studies, which could possibly result in valuable applications for synthetic chemistry and catalysis, as well as for biological and pharmaceutical sciences.
18
References
1. Reinaud, O.; Le Mest, Y.; Jabin, I. Calixarenes in the Nanoworld, Chapter 13 – Models of
Metallo-Enzyme Active Sites, Springer Science & Business Media, 2006.
2. Rebilly, J.-N.; Reinaud, O. Supramolecular Chemistry : From Molecules to Nanomaterials -
Supramolecular Bioinorganic Chemistry, Wiley, 2012.
3. Coquière, D.; Le Gac, S.; Darbost, U.; Sénèque, O.; Jabin, I.; Reinaud, O. Org. Biomol.
Chem. 2009, 7, 2485 – 2500.
4. Višnjevac, A.; Gout, J.; Ingert, N.; Bistri, O.; Reinaud, O. Org. Lett. 2010, 12, 2044 – 2047.
5. Gout, J.; Višnjevac, A.; Rat, S.; Bistri, O.; Le Poul, N.; Le Mest, Y.; Reinaud, O. Eur. J.
Inorg. Chem. 2013, 29, 5171 – 5180.
6. A) Ikeda, A.; Shinkai, S. Chem. Rev. 1997, 97, 1713 – 1734.
B) Homden, D.M.; Redshaw, C. Chem. Rev. 2008, 108, 5086 – 5130. 7. Lipscomb, W. N.; Sträter, S. Chem. Rev. 1996, 96, 2375 – 2434.
8. Gout, J. ; Rat, S. ; Bistri, O. ; Reinaud, O. Eur. J. Inorg. Chem. 2014, 17, 2819 – 2828. 9. Bistri, O.; Colasson, B.; Reinaud, O. Chem. Sci. 2012, 3, 811 – 818.
10. Rondelez, Y.; Bertho, G.; Reinaud, O. Angew. Chem. Int Edit. 2002, 41, 1044 – 1046. 11. Gibb, B.C.; Chapman, R.G.; Sherman, J.C. J. Org. Chem. 1996, 61, 1505 – 1509. 12. Bryant, J.A.; Blanda, M.T.; Vincenti, M.; Cram, D.J. J.Am.Chem.Soc. 1991, 113, 2167 –
2172.
13. A) Marshall, J.A.; Yanik, M.M; Adams, N.D.; Ellis, K.C.; Chobanian, H.R. Org. Synth.,
2005, 81, 157 – 170.
B) Mori, M.; Tonogaki, K.; Kinoshita, A. Org. Synth., 2005, 81, 1 – 13.
Figures
Fig. 1 – 3 Višnjevac, A.; Gout, J.; Ingert, N.; Bistri, O.; Reinaud, O. Org. Lett. 2010,
12, 2044 – 2047.
Fig, 4 – 6 Dr. O. Bistri
19
Supplementary Information
General experimental methods
All reagents and solvents were commercially obtained and used without further purification, except for N-bromosuccinimide, which was recrystallized from water, and THF, which was freshly distilled over sodium in the presence of the colour indicator benzophenone under an inert atmosphere, and butanone, which was freshly distilled over sodium sulfate under an
inert atmosphere. Anhydrous ‘extra-dry’ DMF and DMA (H2O < 30 ppm, Acros Organics)
were used as received and kept over molecular sieves under argon. At all times, ´water´ or
´H2O´ signifies distilled water. n-Butyllithium was titrated regularly with diphenylacetic acid
to measure its exact molarity. Diphenylacetic acid was recrystallized from toluene.
One- and two-dimensional 1H and 13C-NMR spectra were recorded with a Bruker ARX250
spectrometer (250 MHz) and a Bruker AvanceII 500 spectrometer (500 MHz). 1H- and 13C
chemical shifts (δ) were referred to SiMe4. Standard COSY, HSQC and HMBC experiments
were used for peak assignments.
MS (ionization via ESI+) analyses were carried out with a ThermoFinnigen LCQ Advantage spectrometer using methanol or dichloromethane as a solvent.
A) Synthesis of Rim4(OH)4
Note
The major part of the steps has been carried out multiple times. The experimental method given corresponds to the trial which resulted in the highest percent yield.
Dodecol (R=(CH2)3OH) (1) 11
Under an inert atmosphere, resorcinol (20.02 g, 181.82 mmol) was dissolved in 4:1 MeOH/HCl (150 mL). Over a period of 24h, 2,3-dihydrofuran (13.8 mL) was added to the solution whilst stirring. Stirring was proceeded for another 4h at RT, then for 7 ½ days at 50°C, after which the mixture was allowed to cool to RT. The precipitate that had formed was filtered off over a frit filter (pore size nº 3) and washed with ±55°C water (~600 mL). The residue was taken into cold water (1 L) and sonicated (~10 min). The solid was again filtered off over a frit filter (pore size nº 3) and washed with ~55°C water (~250 mL), then dried for nights at the filter, then for 1 night in a desiccator filled with P2O5 and then for 24 h in a
lyophilizator, yielding product 1 as a crème-coloured powder (18.8 g, 57%).
1H-NMR (250 MHz, (CD3)2SO, 300K) δ (ppm): 8.90 (OH, s, 8H), 7.21 H, s, 4H), 6.13
(Ar-H, s, 4H), 4.32 (-CH-, t, J = 5.2Hz , 4H), 4.18 (-CH2-OH, t, J = 7.8 Hz, 4H), 3.37 (-CH2-OH, m,
20
Tetrabrominated dodecol (R=(CH2)3OH) (2) 11
Dodecol 1 (1.91 g, 2.65 mmol) was taken into 7:3 butanone/MeOH (38 mL) under an inert atmosphere. To this suspension, freshy recrystallized NBS (2.11 g, 11.94 mmol) was added and the resulting solution was stirred for 6h in the dark at RT. Additional NBS (0.94 g, 5.31) was added and the mixture was stirred overnight (still at RT, under inert atmosphere, in the dark). The precipitate that had formed was filtered off over a frit filter (pore size nº 4), washed with cold butanone (100 mL) and dried, yielding product 2 as a crème-coloured powder (1.96 g, 71%).
1H-NMR (250 MHz, (CD3)2SO, 300K) δ (ppm) : 9.10 (Ar-OH, s, 8H), 7.37 (Ar-H, s, 4H), 4.30
(-CH-, t, J= 5.4 Hz, 4H), 3.13 (CH2-OH, m, 8H), 2.41 (CH2-CH2-CH, m, 8H), 1.31 (CH2-CH2-OH,
m, 8H).
Bridged tetrabromoresorcinarene (R=(CH2)3OH) (3) 11
Tetrabromododecol 2 (5.7 g) was dissolved in DMA (26 mL) under an inert atmosphere.
K2CO3 (9.88 g) and CH2BrCl (1.6 mL, 24.76 mmol) were suspended in DMA (~185 mL) under
an inert atmosphere. Under an inert atmosphere at RT and whilst stirring, the solution of
tetrabromododecol was added to the solution of K2CO3/CH2BrCl during five days. One more
day of stirring at RT under an inert atmosphere followed. Additional CH2BrCl (1.6 mL, 24.76
mmol) was added and stirring was proceeded at 45°C for 3 days. The final portion of CH2BrCl (1.6 mL, 24.76 mmol) was added and the mixture was stirred at 65°C for seven
days, after which the mixture was allowed to cool to RT and all DMA was evaporated under vacuum. The resulting solid was dissolved in water 150 mL and at 0°C, HCl (2M, 60 mL) was added slowly. After 15 min of stirring, the mixture was filtered over a frit filter (pore size nº 4) and the residue was washed with water (4 x 50 mL). After drying under air, the residue
was taken into 150 mL THF (non-distilled) and dried over MgSO4. MgSO4 was filtered off and
THF was evaporated. The crude product was purified by flash column chromatography
(SiO2, CHCl3/MeOH 9:1), yielding product 3 as a white solid (2.44 g, 41%).
1H-NMR (250 MHz, (CD3)2SO, 300K) δ (ppm): 7.64 (Ar-H, s, 4H), 5.99 (O-CH2-O out, d, J = 7.5Hz, 4H), 4.68 (-CH-, t, J = 8.0Hz, 4H), 4.50 (CH2-OH, m, 4H), 4.29 (O-CH2-O in, d, J = 7.5Hz,
4H), 3.50 (CH2-OH, m, 8H), 2.44 (CH2-CH2-CH, m, 8H), 1.43 (CH2-CH2-OH, m, 8H).
Tetrabromoresorcinarene (R=(CH2)3OTIPS) (4)
Under an inert atmosphere, product 3 (3.22 g, 2.97 mmol) was dissolved in anhydrous DMF (31 mL). Imidazole (3.03 g, 44.55 mmol) was added, followed by TIPS-Cl (8.7 mL, 40.7 mmol). The mixture was allowed to stir at RT under an inert atmosphere for 28h. DMF was evaporated and the crude product was dissolved in DCM (60 mL) and washed with water.
21
evaporated. The crude was purified by flash column chromatography (SiO2,
cyclohexane/DCM 3:1 then 3:2), yielding product 4 as a white foam (3.28 g, 65%).
1H-NMR (250 MHz, CDCl3, 300K) δ (ppm): 7.04 (Ar-H, s, 4H), 5.93 (O-CH2-O out, d, J = 7.5 Hz, 4H), 4.87 (-CH-, t, J = 8.2 Hz, 4H), 4.37 (O-CH2-O in, d, J=7.2 Hz, 4H), 3.75 (CH2-OTIPS, t,
J=6.2 Hz, 8H), 2.26 (CH2-CH2-CH, m, 8H), 1.61 (CH2-CH2-OTIPS, m, 8H), 1.15-1.08
(O-Si-{CH-(CH3)}3, m, 84H).
Tetraesterresorcinarene (R=(CH2)3OTIPS) (5)
Tetrabromoresorcinarene 4 (1.0 g, 0.59 mmol) was dried by dissolving it in freshly distilled THF and allowing the THF to evaporate at 80°C under vacuum. This procedure was repeated twice. Under an inert atmosphere at RT, the dried tetrabromoresorcinarene was dissolved in THF (35 mL). At -78°C, n-BuLi (3.6 mL, 5.47 mmol) was added, 2 hours later followed by ClCOOCH3 (0.45 mL, 5.78 mmol) and after 15 min of stirring at -78°C, the
mixture was allowed to stir under an inert atmosphere at RT for 19 ½ h. At 0°C, water (5 mL) was added, after which all solvent was evaporated under vacuum. The solid was dissolved in DCM (60 mL) and the organic layer was washed twice with water (40 then 30 mL). The
organic layer was dried over MgSO4, MgSO4 was filtered off, all DCM was evaporated and
the product was dried at the vacuum ramp, yielding product 5 as a white foam (0.94 g, 99%). 1H-NMR marl 0519 1401 (250 MHz, CDCl3, 300 K) δ (ppm): 7.18 (s, 4H, Ar-H down), 5.65 (d, J= 7.4 Hz , 4H, -O-CH2out-O-), 4.78 (t, J= 7.7 Hz, 4H, CH2-CH), 4.57 (d, J= 7.2 Hz, 4H, -O-CH2in
-O-), 3.84 (s, 12H, COOCH3) 3.77 (t, J= 7.8 Hz, 8H, CH2-OTIPS), 2.29 (m, 8H, CH2), 1.58 (m, 8H,
CH2), 1.09 (s, 72H, Si(CH3)2), 1.06 (s, 12H, SiCH)
Tetraolresorcinarene (R=(CH2)3OTIPS) (6)
Tetraesterresorcinarene 5 (0.94 g, 0.58 mmol) was dried by dissolving it in freshly distilled THF and allowing the THF to evaporate at 80°C under vacuum. This procedure was repeated twice. Under an inert atmosphere at 0°C, a solution of the dried tetraester in THF
(28 mL) was added to a solution of LiAlH4 (0.28 g, 7.35 mmol) in THF (34 mL). The mixture
was stirred at 0°C for 15 min, then at RT for 48h. Over a period of 30 min, ethyl acetate (30
mL) was added to react with unreacted LiAlH4 and this was stirred for 5 ½ h at 0°C. Then,
still at 0°C, MeOH (2 mL) was added, followed by water (2.5 mL). The mixture was stirred for 1h (half at 0°C half at RT), then dried over Na2SO4. Na2SO4 and residual salts were filtered
off over Büchner and the filtrate was concentrated. The crude was purified by flash column
chromatography (SiO2, DCM:MeOH 94:6 then 90:10), yielding product 6 as a white foam
(0.31 g, 35%).
1H-NMR 16062014-chot-20 (500 MHz, CDCl3, 300 K) δ (ppm): 7.14 (s, 4H, Ar-H down), 5.9 (d, J= 7.6 Hz , 4H, -O-CH2out-O-), 4.82 (t, J= 8.3 Hz, 4H, CH2-CH), 4.55 (s, 8H, CH2OH), 4.42 (d,
22
J= 7.6 Hz, 4H, -O-CH2in-O-), 3.78 (t, J= 6.4 Hz, CH2-OTIPS), 2.30 (m, 8H, CH2-CH), 1.60 (m, 8H,
CH2-CH2), 1.09 (s, 72H, Si(CH3)2), 1.07 (s, 12H, SiCH).
13C-NMR 16062014-chot-21 (500 MHz, CDCl3, 300 K) δ (ppm): 153.8, 138.1, 126.8, 120.2, 100.1, 77.2, 63.3, 55.2, 36.6, 31.2, 26.4, 18.2, 12.2.
Tetraimidazoleresorcinarene (R=(CH2)3OTIPS) (7)
Tetraolresorcinarene 6 (0.21 g, 0.14 mmol) was dried by dissolving it in freshly distilled THF (5 mL) and allowing the THF to evaporate at 80°C under vacuum. This procedure was repeated twice. Under an inert atmosphere at 0°C, a solution of the dried tetraol in anhydrous DMF (4 mL) was added to a solution of NaH (60 mass% in oil, washed with pentane, 0.09 g, 4.16 mmol) and the resulting mixture was stirred for 15 min at 0°C, then for 1h at RT. At 0°C, 2-chloromethyl-1-methyl-1H-imidazole hydrochloride (0.23 g, 1.39 mmol) was added in three portions, with 30 min of stirring between each portion. The mixture was allowed to stir at RT for 12h and then poured in portions into water (150 mL) at 0°C to precipitate the product. The solid was filtered off over Büchner, washed with water (3 x 15 mL), dissolved in DCM (70 mL), washed with water twice (40 mL then 30 mL). The organic
layer was dried over Na2SO4, after which Na2SO4 was filtered off and DCM was evaporated.
The procedure as described until this point was repeated once again. The crude product was triturated with pentane, yielding product 7 as a light yellow powder (0.19 g, 72%).
1H-NMR marl 0430 1402 (250 MHz, CDCl3, 300 K) δ (ppm): 7.07 (s, 4H, Ar-H down), 6.99 (br
s, Im-H), 6.95 (br s, Im-H), 5.62 (d, J=7.3 Hz , 4H, -O-CH2out-O-), 4.75 (t, J=8.2 Hz, 4H, CH2
-CH), 4.63 (s, CH2O, 8H), 4.25 (s, 8H, OCH2Im), 4.20 (d, J=7.1 Hz, 4H, -O-CH2in-O-), 3.74 (t, J=6.1 Hz, 8H, CH2-OTIPS), 3.66 (s, 12H, ImCH3), 2.20 (m, 8H, CH2-CH), 1.55 (m, 8H, CH2
-CH2), 1.09 (s, 72H, Si(CH3)2), 1.06 (s, 12H, SiCH)
Tetraimidazoleresorcinarene (R=(CH2)3OH) i.e. Rim4(OH)4 (8)
At RT, not under an inert atmosphere, product 7 (0.052 g, 0.028 mmol) was dissolved in 1:1
THF/H2O (3 mL) and TFA (0.3 mL, 3.89 mmol) was added and the mixture was stirred
overnight. TFA was evaporated, after which the resulting solid was suspended in toluene. Solvents were evaporated and the resulting solid was dried at the vacuum ramp. It was then
dissolved in MeOH (5 mL) and stirred with DOWEX OH- resin at the rotary evaporator
(since magnetic stirring beans are known to destroy the resin). The resin was filtered off over a frit filter (pore size nº 3) and washed with MeOH. The filtrate was concentrated, suspended in a (1:1) mixture of toluene/MeOH, solvents were evaporated and the resulting solid was dried under vacuum, yielding product 8 as a light yellow powder (0.041 g, 79%).
1H-NMR 05062014-chot-10 (500 MHz, CD3OD, 300 K) δ (ppm): 7.37 (s, 4H, Ar-H down), 7.06 (br s, Im-H), 6.90 (br s, Im-H), 5.52 (d, J=7.2 Hz , 4H, -O-CH2out-O-), 4.76 (t, J=8.2 Hz, 4H, CH2
-23 CH), 4.48 (s, CH2O, 8H), 4.32 (s, 8H, OCH2Im), 4.19 (d, J=7.1 Hz, 4H, -O-CH2in-O-), 3.67 (t, J=6.1 Hz, 8H, CH2-OH), 3.55 (s, 12H, ImCH3), 2.38 (m, 8H, CH2-CH), 1.55 (m, 8H, CH2-CH2).
13C-NMR 16052014-chot-11 (500 MHz, CD3OD, 300 K) δ (ppm): 155.7, 139.3, 127.5, 125.3, 124.0, 122.8, 101.4, 64.8, 63.0, 49.2, 38.2, 33.4, 32.2, 27.5
HR-MS (ESI+, MeOH) m/z expected for [Rim4(OH)4H]+: 1265.578, m/z obtained : 1265.5765
(Δ = -2.10 ppm)
COSY (1H/1H) 05062014-chot-13 (500 MHz, CD3OD, 300 K)
HSQC (1H/13C) 05062014-chot-12 (500 MHz, CD3OD, 300 K)
HMBC (1H/13C) 05062014-chot-11 (500 MHz, CD3OD, 300 K)
B) Synthesis of WRim3
Alongside the synthesis of Rim4(OH)4, an attempt was made to reproduce WRim3, the
water-soluble tris-imidazole-functionalized resorcin[4]arene earlier obtained by Reinaud et al.The synthesis of WRim3 was brought to a halt at the stage of triol (i.e. before imidazole