Master Thesis Chemistry
Attempt at a covalent [2]catenane synthesis by a templated
backfolding strategy
The effect of adjacent electron rich functional groups
on the ketal template cleavability of the pre-[2]Catenane
By
Oswald C. Philipsen
09 October 2020
Student number
11165383
Responsible professor
prof. dr. J.H. van Maarseveen
Supervisor
Simone Pilon
Research institute
University of Amsterdam
Research group
Abstract:
In this project an attempt at the synthesis of a pre-[2]catenane by means of a covalent backbonding templated strategy was made. A ketal moiety and a terephthalic acid derivate were used as a template for the pre-organisation of the building blocks. The goal of the project was to see the effect of the addition of electron rich groups adjacent to the ketal on the cleavability. If this modification would increase cleavability it could open a way to catenanes and other mechanically interlocked molecules, which cannot be synthesised by non-covalent means.
This thesis reports on the synthesis of each of the building blocks and their combination towards the desired pre-[2]catenane. For the ketal to be synthesised the ketone ring thread and the diol core had to be obtained. The diol core was based on mannaric acid and was synthesised successfully by two different approaches starting from either mannitol or dimethyl tartrate. Meanwhile the other building blocks were synthesised by the procedure of S. Pilon. Finally, the combination of all the building blocks was reported. Unfortunately, due to time constraints the target goal was not reached. Synthesis of the two macrocycles was achieved, however, no conclusive evidence of their synthesis can be given as unclarity exists within the spectroscopic data.
Table of content
List of abbreviations ... 4
Introduction ... 5
Clipping approach ... 6
Active Metal Centre approach ... 7
Covalent Template Strategy ... 9
Molecularly interlocked Molecule in nature ... 10
Goal of project ... 11
Retrosynthesis ... 12
Chapter 1: Synthesis of the Mannaric acid core ... 14
Oxidation pathways ... 14
Mannose oxidation pathway ... 14
Mannitol oxidation pathway ... 15
Ozonolysis pathways ... 17
1,4-benzoquinone pathway ... 17
Dimethyl tartrate pathway ... 18
Chapter 2: Synthesis of building blocks ... 19
Chapter 3: The road towards the pre-catenane ... 20
Conclusion ... 23
Discussion & Future prospects ... 24
Acknowledgement ... 25
Supporting information ... 26
NMR spectra of synthesised compounds ... 38
List of abbreviations
Abbreviation: Meaning
CSA Camphorsulfonic acid
CuAAC Copper-catalysed Alkyne0Azide Cycloaddition
DCM Methylene Chlorine/Dichloromethane
DCC N,N'-Dicyclohexylmethanediimine
DIBAL-H Diisobutylaluminum hydride
DiPEA N,N-Diisopropylethylamine DMAP N,N-Dimethylpyridin-4-amine DMF N,N-dimethylformaldehyde DMP 2,2-Dimethoxypropane Et2O Diethyl ether EtOH Ethanol
HBTU Hexafluorophosphate Benzotriazole Tetramethyl Uronium
HOBt Hydroxybenzotriazole
LiHMDS Lithium Hexamethyldisiliazide
LDA Lithium diisopropyl Amine
NBS N-bromosuccinimide
NMR Nuclear Magnetic Resonance
PfPO Pentafluorophenol
pTSOH 4-Methylbenzene-1-sulfonic acid
PyBOP (Benzotriazol-1-yloxy)tripyrrolidinophosphonium hexafluorophosphate
TBDMS Tert-Butyldimethyl silyl
TEMPO (2,2,6,6-Tetramethylpiperidin-1-yl)oxyl
THF Tetrahydrofuran
TMS Tetramethyl silyl
Introduction
Mechanically Interlocked Molecules (MIMs) hold a fascinating place in chemistry due to their intriguing shape. MIMs consist of molecular entities, which are held in a certain topology by means of a mechanical bond. This mechanical bond is governed by repulsive forces, preventing covalent bonds from crossing each other. What this means is that the entities cannot be separated without the breaking or distortion of a covalent bond.1 These molecules can be characterised based on the number of entities
and their topological natures. Examples include catenanes2, rotaxanes3, pretzelanes4 and trefoil
molecular knots5 (Fig.1).
Figure 1. Terminology of different Mechanically Interlocked Molecules.
Catenanes are MIMS consisting of two or more interlocked rings. The first synthesis of a catenane was reported as early as 1960 by the group of Wasserman.6 Their method, known as the
statistical approach, relies on the probability of intertwinement occurring between a ring and a linear ring precursor (thread). Formation of the catenane could be achieved by acyloin cyclisation of said precursor, yielding a [2]catenane (Fig. 2). This method gave a yield of only 1% for the desired catenane, exemplifying the inefficiency of a synthesis based on probability.6
Figure 2. Statistical synthesis of a [2]catenane A by Wasserman et al. by an acyloin cyclisation of a threaded precursor yielding the [2]catenane and the separate cyclised thread.6
The work of Wasserman illustrated that without additional assistance of pre-organising both fragments, efficient synthetic approaches could not be obtained. This concept was explored by the group of Schill, who in 1964 reported on the first templated synthesis of a [2]catenane (Fig. 3A).7 A ketal
moiety 1 was used to pre-organise and bring two strands together. An additional amino group directed the macro-cyclisation towards an intertwined structure 2 followed by further modifications giving a [2]catenane 3. This type of synthesis, using functional groups for assembly and pre-organisation was named a directed or templated approach and proved to be more efficient that the statistical one. In 1983 the group of Sauvage reported on a [2]catenane synthesis using a non-covalent template (Fig. 3B).8 This
synthesis employed a Cu(I) ion, which coordinated to two different phenanthroline moieties resulting in a stable, preorganised intermediate 4. A Williamson Ether synthesis with a linear bis-iodine thread yielded a pre-catenane 5 in 42% yield . This was the first example of a template based on non-covalent interactions, in this case metal-ligand interactions. Further research by other groups yielded a variety of templates based on different non-covalent interactions including π-π electrostatic, hydrogen bonding and crownether-ammonium and halogen bonding.3
O O O O 32 + D D D D 1. Na, Xylene 2. HOAc DD DD O OH O OH + A
Figure 3. Template approached by: A. Schill et al. B. Sauvage et al. towards [2]catenanes.7,8
The strategy of templated catenane synthesis can be divided into two approaches; the clipping approach and the active-metal-template approach. The clipping method is older, more thoroughly researched and thus the prevalent applied approach for catenane synthesis. On the other side is the novel active-metal template approach, which takes advantage of the catalytic nature of metal ions templates the ring closing
1.1 The Clipping strategy
The clipping approach involves the formation of an interlocked macrocyclic system by macrocyclization from a linear ring precursor. Firstly, two entities are coordinated perpendicularly towards each other by means of the selected template (in this example a metal ion). Coordination might either involve two free strands which are then cyclised into a [2]catenane, or a free strand that coordinates to another ring. During coordination of the second strand a semi-cyclic form is adopted, achieving pre-organisation of both ends towards a cyclic structure. Once pre-organisation is achieved, macrocyclization can be performed in two ways; either intramolecularly (A) or via an intermolecular ring-closing reaction (B) (Fig.4)
Figure 4. Schematic representation of the clipping strategy by either A) Intramolecular B) intermolecular ring closing
An example of a intramolecular ring closure catenane synthesis was reported by Beer et al. in 2020 (Fig. 5).9 The synthesis of the ring fragments begins with the attachment of a
naphthalene-hydroquinone (7) derivative to a bromoimidazolium (6) derivative in two steps forming the ring fragment (8) as a bromine salt 8-Br. Ion exchange using NH4PF6 gave the PF6 salt analogue. Addition
of 8-Br, 8-PF6 and a Grubbs Catalyst afforced the catenane 9-Br-PF6. The catenane was formed by using a bromine anion as template by means of halogen interactions between the bromoimidazolium and bromide anion. Additional π-π interactions between the electron deficient bromoimidazolium and electron rich hydroquinone stabilised the pre-catenane complex (pre-9-Br-PF6). Ring closure was achieved by a Ring Closing Metathesis (RCM) reaction between the terminal alkene moieties yielding the catenane 9-Br-PF6 in 24% yield.9 Finally, the Br- template could be removed by ion exchange using
NH4PF6. The importance of the Br- template was shown by running the reaction using only 8-PF6, which
OH N N O O O O O O Cu(MeCN)4BF4 N N O O O O O O N N OH Cu+ Cs2CO3/DMF O O O I O I O N N O O O O O O N N O Cu+ O O O O 5 42% HO HO (H2C)25 (CH2)25 NH2 (H2C)25 (CH2)25 NH2 O O (H2C)12 (H2C)12 Cl Cl (H2C)25 (CH2)25 N O O (CH2)8 (H2C)8 (H2C)25 (CH2)25 NAc O HO (CH2)8 (H2C)8 O O A B 1 2 3 4
did not result in the formation of the catenane. This exemplified the importance of the halogen interactions, which held both the ring fragments together.
Figure 5. Clipping approach towards a [2]catenane with an intramolecular ring closing reaction.9
Intermolecular ring closing reactions involve the use of a third fragment which is covalently attached to both sides of a ring strand in order to form a ring. Examples include the synthesis by Megiatto et al. who reported in 2008 a synthesis of a [2]catenane using a Cu(I) template (Fig. 6).10 The
strategy begins with the synthesis of a phenanthroline ring fragment 10, previously developed by Sauvage et al.. Alkylation with 3-bromoprop-1-yne gave the desired ring precursor 11. The ring precursors could be preorganised by the addition of a copper(I) complex, resulting in interactions between the phenanthroline and copper. Finally, the rings were closed by a “click” reaction using diazide (13) to yield the [2]catenane 14 after the removal of the copper ion. The [2]catenane was obtained in yields of 80-92% depending on the diazide used.10
Figure 6. Clipping approach towards a [2]catenane with an “Click” ring closing reaction.10
1.2 Active Metal Centre approach
In the clipping approach the metal ion template, plays a passive role during the ring closing reaction. This is in contrast to the Active Metal Centre approach, in which the template plays an active role in the ring closing reaction. A short overview of this strategy is given in Figure 7. The synthesis generally begins with the separate synthesis of one of the two rings. The ring incorporates a specific motif required for the coordination of the metal ion, for example pyridine derivatives. This is followed by the introduction and coordination of the metal ion. Next the second ring thread coordinates to the metal ion from thread end, followed by the coordination of the other thread end. Since the first coordination occupies one face of the metal ion the latter one has to occur from the other side of the
N NH Br Br O O O + 6 7 NaOH r.t MeCN N N Br O O O Br O O O + 60°C, MeCN N N Br O O O O O O Br 8-Br N N Br O O O O O O PF6 8-PF6 NH4PF6 8-Br +8-PF6 N N H Br O O O O O O O O O N Br H O H O O Br pre-9-Br-PF6 PF6 -N N H Br O O O O O O O O O N Br H O H O O Br 9-Br-PF6 PF6 -Grubbs II DCM O N N O O O O OH O O OH Br NaH, THF N N O O O O O O O O Cu(I) N N O O O O O O O O O O O O O O N N O Cu 10 11 12 R N3 N3 O N N O O O O O O O O O O O O O O N N O 14 N N N N N N R N N N N N N R 1. CuI R = CH=CH2 R = O O 13 2. NH4OH/H2O
ring. This ensures that an intertwined system is created. Then, the ring closing reaction occurs, which is being catalysed by the metal ion itself. Lastly, the metal ion dissociates, which can either be spontaneous or chemically facilitated.
Figure 7. Schematic representation of the Active metal centre approach.
An example of a catenane synthesis by the Active metal centre approach was reported by Goldup at al. In their synthesis the ring closing reaction was performed by means of a Cadiot-Chodkiewicz reaction with copper(I) as the catalyst (Fig. 8). To the pre-synthesised ring 15 the ring precursor 16 was added, followed by the addition of Cu(I) and a base. The copper coordinated to the bipyridine motif of 15 while the base resulted in the deprotonation and coordination of the terminal alkyne to the copper. The alkyne-bromide then undergoes oxidative addition to the copper(I) species forming the pre-catenane 17. Since one face of the copper ion was occupied by the alkyne, oxidative addition had to occur on the other face, resulting in a threaded system. Finally, a reductive elimination resulted in the formation of the C-C bond of the between the alkynes yielding the catenane 18 in 21% over the whole catalytic cycle.
Figure 8. Active metal centre approach towards a [2]catenane by Goldup et al.
N N O O O O N N O O O O O O Br O O Cu L 2. CuI, LiHMDS THF -78°C N N O O O O O O Cu Br N N O O O O O O Br Oxidative additon Reductive Elimination 15 16 18 17
1.3 Covalent Template strategy
In the previous examples, pre-organisation was achieved by means of a variety of non-covalent interactions, including halogen interactions, coordination to a copper(I) ion, π-π interactions. While reliable, these approaches pose a specific challenge. For non-covalent interactions to occur specific motifs must be introduced into the ring fragments. For example, coordination to a copper(I) ion requires the presence of a crown ether or bi-pyridine moiety; π-π interactions require the presence of large aromatic groups. These moieties cannot be removed post ring closing, limiting the scope of variety for the ring strands within the synthesis.
As an alternative for non-covalent templates covalently bound templates have been developed. In these cases, covalent bonds are used to conjunct and pre-organise the separate ring strands. While these approaches have been less studied than the templated ones examples include, as earlier mentioned the famous synthesis by Schill et al. (Fig. 9).7 An important aspect of this synthesis was the ketal group
19, which was responsible for the conjunction of the ring and thread-precursor. The tetrahedral
geometry of the ketal played a vital role as it ensured a perpendicular arrangement of the fragments, which is essential for efficient ring closure. The second vital feature was the amine group. It was placed in such a fashion that the halogenated ring strands had to fall back onto themselves 20 for the alkylation to occur. This resulted in the characteristic intertwined structure of the [2]catenane 21
Figure 9: Short schematic of the covalent template synthesis by Schill et al., employing a ketal moiety.
A different covalent template was reported by Schweez et al. in 2018.11 In this synthesis a
para-terphenyl template featuring two aryl iodides was attached to a macrocycle by means of phenolic ester bonds 22. The ring threads were consequently attached by means of a palladium coupling yielding ring-precursor 23. The iodides and phenolic ester groups were positioned in such a way that the resulting macrocycle and ring strand were arranged in a perpendicular fashion. Compared to the synthesis of Schill et al., during this synthesis intertwinement was achieved by means of steric bulk on the ring threads (blue). The steric demand of the attached acetylenes of 23 resulted that only the intertwined precursor was formed. The synthesis was followed by a ring-closing reaction 24 and finally the templated was detached by a transamination yielding catenane 25.11
Figure 10. Short Schematic overview off the covalent templated synthesis by Schweez et al.
.
11While the covalent template approach has the disadvantage of requiring additional steps for the implementation and detachment of the template, it show great promise as result of the ability to modify the template post-catenane formation, which possess a challenge for non-covalent template approaches. Furthermore, templates that mimic motives found in natural compounds may be developed, allowing access to natural mechanically interlocked molecules.
1.4 Molecularly interlocked Molecules in Nature
Although all compounds mentioned in the previous sections are of synthetic origin, examples of natural occurring catenane-like structures have been reported. Examples of such natural cases include DNA catenanes, in which two rings of DNA are intertwined.12 Certain proteins have also been found to
possess catenane structures. For example the capsid of the HK97 bacteriophage named the “protein chainmail”, consisting out a network of intertwined protein rings.13
A particular interesting family of MIMs are the “lasso peptides”, which possess two distinct features; a cyclic peptide ring and peptide thread, which is threaded and attached to the ring via a side chain residue.14,15 These lasso-peptides are an interesting case as they possess superior biological
stability compared other types of peptides.14 The biological stability is attributed to two factors; firstly,
the cyclic peptide structure possess a higher stability to peptidases and secondly, the thread fragment is shielded by the surrounded ring fragment. Due to the conformational strain, degrading enzymes cannot reach the inner thread amino acid, resulting in higher stability. Furthermore, the conformational stain inhibits movement of the peptide thread leading to an increased thermal stability.14 Lasso peptides have
been found to exhibit biological antimicrobial activity. Combined with their high biological stability and bacterial origin, they have been target of interest as potential drug scaffolds.16
One such lasso-peptide is Microcin J25 (MccJ25), which is considered as a kind of archetype, since it was one of the earliest discovered lasso-peptides and the first to have its gene cluster identified. The structure of MccJ25 is shown in figure 11. The macrocycle (red) consists out of eight amino acids and is biosynthesised by the cyclisation between Gly1 with Glu8. The thread (blue) is made out of
thirteen amino acids and is attached to the residue of Glu8. This thread is kept in place by two bulky
residues Phe19 and Tyr20 on both sides of the ring, which prevent dethreading.14
Figure 11. Structure of Microcin J25.
While the structure, biosynthesis and gene cluster of MccJ25 have been elucidated, no successful synthesis has been reported yet. While the dethreaded isomer and the fully cyclic head to tail MccJ25 analogues have been synthesised they show no bioactivity compared to natural MccJ25 exemplifies unique nature of the intertwined strand.15,17 The synthetic challenge lies in the correct preorganisation
of the ring fragment precursors, firstly due to the inherit chirality of the molecule and secondly, the absence of moieties for conventional templates, making the use of non-covalent interaction bases templates impossible. Therefore, the covalent templated approach becomes the more feasible strategy towards the synthesis of such unique compounds.
1.5 Goal of the project
In 2017 a paper was published by L. Steemers et al. reporting on a synthesis of a pre-[2]catenane 26 by means of a covalent template.18 The goal of that project was to establish a strategy of
catenane synthesis, excluding the use of supramolecular motifs, which are generally applied in modern synthetic approaches. Ultimately, such covalent strategy was explored for the potential to aid in the synthesis of Lasso-peptides, which were seen as the far future end goal.
The paper described the use of a ketal functional group as the covalent template. This group was chosen as it is a well know functional group, facile to install, rigid and can easily be removed and was therefore a prominent candidate for a template. More importantly, the central ketal carbon atom possesses a tetrahedral geometry, which pre-organises the ring strands in a perpendicular fashion (90°), assuring optimal configuration for the formation of interlocked rings. The pre[2]catenane 26 in question is shown in figure 12. For simplicity of characterisation, the two rings were symmetrically designed and ring closure was based on robust chemistry, such as “click” reactions and ring closing metathesis. Amide groups were included as to mimic peptide chemistry as close as possible. Tartaric acid was taken as building block for the ketal. The stereochemistry of the hydroxyl groups allowed for the formation of a symmetrical ketal with the two ring strands perpendicular to each other.
Figure 12. The pre-[2]catenane designed by Steemers et al.
Besides the use of a ketal covalent template a second template was included, which directed the ring formation by means of a backfolding effect. This effect is represented in figure 11. Formation of the of the ketal yields the tetrahedral precursor 27A. However, at this stage ring closing reactions of
27A would not yield the desired interlocked. Instead, a spiro-bicycle 28 would be formed (Fig. 13).
Cleavage of the ketal moiety would then result in two separate macrocycles. In order to force the formation of interlocked rings the backfolding template has to be present, which is attached to the ring strands by means of temporary linkages 27B.
Figure 13. Schematic representation of the Backfolding template strategy.
During the first cyclisation 27C the template forces ring strands to fold back onto themselves, with the ends reacting with the template itself forming the first macro cycle. The second macro cyclisation is also forced into a backfolding cyclisation resulting in an inverted-spiro conformation 27D. Removal of the temporary linkages affords the desired pre-[2]catenane 27E. Cleavage of the ketal bonds would result in the [2]catenane 29.
Unfortunately, attempts at cleavage of the ketal 26 using acidic conditions were unsuccessful. It was reasoned that the electron withdrawing amides hindered the protonation of the dioxolane oxygen. Additionally, the steric hinderance resulted in the shielding of the ketal from protonation. In order to improve cleavability of the ketal two modification to the design of 26 were made. Since it was reasoned that the amides inhibited protonation, one of the rings containing amides (red) was transformed into an
O O O N H O HN O NH N N N CO2Me MeO2C NN N HN O 26
Cyclisation Backfolding Template First BackfoldingCyclisation
Second Backfolding
Cyclisation Removal of temporaryLinkages Ketal cleavage
27D 27A
aliphatic strand. Secondly, an additional carbon atom was placed in between the amides and the hydroxyl functionalities on the tartaric acid core (Fig. 14). This yielded pre-[2]catenane 30 (Fig. 12) as the target molecule of this project. While methoxide groups are inductively electron withdrawing, they were included for synthetic feasibility.
Figure 14:.Design of target pre-[2]catenane.
Retrosynthesis:
The retrosynthesis analysis for the desired pre-[2]catenane starts by removal of temporary likages and hydrogenation of the macro-cycle compound 31 (Fig. 15). Formation of 31 can be envisioned by a ring closing metathesis forming the second macrocycle starting from 32, which is the product of a Copper catalysed Alkyne-Azide cyclodaddition (CuAAC) macrocyclisation of 33. Compound 33 is made by the combination of ring-precursor 34 and the terephtalic backbonding template 35.
Figure 15. Retrosynthetic analysis of the pre-[2]catenane 30.
The ring precursor 34 can be synthesised by attaching the temporary linkages to carboxylic acid ketal core as amides of the amine 37. In turn the carboxylic acid groups are obtained by saponification of 36. The ketal is made by attaching the ketone ring thread 39 to the mannaric acid building block 38. Amine
37 is known from the procedure of L. Steemers and is made by the reductive amination of commercially
available aldehyde 40 and the primary amine 41. This amine can be synthesis from the alcohol derivative 42, which is commercially available.
O O N N N CO2Me MeO2C NN N OMe OMeNH O O N H 30 O O O O O N H NH O O O N N N N N N OMe OMe O O O O O N N O O O O O O O N N N N N N O O O O O N N O O O O O O O N N N N N N O O O O 9 9 O N N O O O O O 9 9 O O N3 N3 9 9 O O O O 9 9 O N N O OH HO O O 9 9 O O N3 N3 OPfp OPfp + 30 31 32 33 34 35
Figure 16: Retrosynthetic analysis of the ring precursor and temporary linkage
The ketone ring thread 39 can be made by oxidation of the corresponding secondary alcohol 43, which is envisioned to be the product of the reaction between ethyl formate and the corresponding Grignard reagent. However, since the bromo-alkyne is not commercial available the chosen starting material is the bromo-alkene 45. Transformation of the alkene 44 to alkyne 43 can be performed after the reaction with ethyl formate.
Figure 17: Retrosynthetic analysis of the ketone ring thread
In order to synthesise the mannaric acid core 38 the 3,4-hydroxyl moieties (red) had to be protected before alkylation on the 2,5 position (blue) could be performed. A strategy was devised that by making the di-lactone of mannaric acid 46 the desired protection could be achieved, which could be trans esterified to obtain 38. The dilactone would be synthesised by oxidation and condensation of mannose, allowing for protection and to the di-acid in one step. Mannose and mannaric acid were choses due to their inherit stereochemistry as ketal possess a trans-configuration for hydroxyl moiety relative to each other. Most importantly, the configuration of the 3,4 hydroxyl moieties mimics the configuration of tartaric acid used in the procedure of L. Steemers.
Figure 18: Retrosynthetic analysis of mannaric acid core.
O O O O 9 9 O N N O OH HO O O 9 9 O O O O 9 9 O NH O O OH O 9 O O O OH HO O O O O O 9 9 34 37 36 39 38 NH OH O 9 37 O OH O 40 NH2 41 OH 42 O 39 OH 43 OH 44 Br 45 OH HO O O O O O O O O O O O O O O OH HO O O O OH OH OH OH HO d-mannose 38 46
Chapter 1: Synthesis of the Mannaric acid core
In this chapter the synthesis approaches towards the synthesis of the 2,5-di-O-methyl Mannaric acid core 38 will be discussed. While Mannaric acid is a known compound, the 2,5-di-O-methyl derivative has not been synthesised before and is rarely mentioned in literature. At the end of this chapter the steps towards the transformation towards the 2,5-di-O-methyl-dimethyl-L-mannarate will be discussed.
1.1 Oxidation Pathways
1.1.1: Mannose oxidation pathway
The first attempted strategy involved the oxidation of mannose towards the L-Mannaro-1,4:3:6-di-𝛾-lactone. The formation of this lactone has the advantage of instantly protecting two hydroxyl moieties as lactones while leaving the desired 2,5-hydroxyl groups free. This would allow for a regioselective methylation in only two steps. Trans-esterification of both lactones would the result in the desired core in only 3 steps.
The synthesis of the di-𝛾-lactone 46 from mannose had been reported in literature a number of times by using nitric acid as an oxidant. Unfortunately, most of these procedures used some form of reactor or alternatively, a very expensive catalyst, neither of which were available. Nonetheless, an attempt was made using the conditions described in the paper of Carpenter et al. (Fig. 19).19 Heating
the mixture up at 30°C overnight followed by condensation of the intermediate mannaric acid gave only a yield of 5%. Increasing the temperature to 40°C did not result in any significant change.
Figure 19: Synthesis of the L-Mannaro-1,4:3:6-di-𝛾-lactone 1 and further alkylation attempt
Fortunately, a paper by Alterman et al. reporting on the synthesis of the di-𝛾-lactone starting from L-mannonic 𝛾-lactone, using nitric acid was found.20 While this specific starting material was not
available, by using the condition described in this paper a yield of 33% was achieved. The crude product could be obtained in acceptable purity by condensation of the reaction mixture and washing the impurities off using Et2O and EtOH. The next step involved methylation of the free hydroxyl groups
using sodium hydride and methyl iodide in DMF. However, after purification no desired product was obtained, instead, a product resulting from the elimination of one of the lactone rings 47a was isolated (Fig. 19). This base promoted elimination was indeed reported by the group of Alterman, in their case by using sodium hydroxide on the di-lactone 46. Using silver(II)oxide and methyl iodide did not yield the desired results.21
In the paper of Alterman the use of an acid promoted alkylation was recommended, in their case benzyl-2,2,2-trichloroacetimidate.20 As the methyl-2,2,2-trichloroacetimidate was not available,
the benzyl derivative was used for a test run. Unfortunately, the desired product could not be obtained. It was noticed that after a couple of days a colour change in the starting material from white to pink was observed, however this discolouration did not affect the 1H-NMR spectrum. It was though that potential
traces of nitric acid/water might still be present, therefore, the di-lactone was filtered through dry K2CO3
and dried with MgSO4 in order to remove potential traces of nitric acid and water. However, this
additional workup did not result in the formation of the desired benzylated product.
An alternative procedure using TMS-diazomethane and aqueous tetrafluoroboronic acid (48%) as catalyst did neither yield the product.22 The failure of this reaction could be attributed to solubility
issues. In literature alkylation with diazomethane is primarily performed in DCM since other solvents generally give lower yields.22 However, due to the high polarity of 46, the di-lactone would not dissolve
O O O OH OH OH OH HO D-mannose NaNO2 cat. O O HO OH T = 30°C 5% 40°C 5% 85°C 33% 46 T, 65% HNO3, NaH (2.5eq.), MeI (3.0eq.) DMF, r.t., o.n. O O O O O O 47a 34% + O O O O O O 47b 0%
in DCM and accumulate in the tiny water droplets from the tetrafluoroboronic acid hindering the reaction severely. Exchanging the aqueous acid for the ether complex or changing the solvent did not change the outcome of the reaction
1.1.2: Mannitol selective protection pathway
In the paper of Alterman a second pathway was reported using mannitol in the second step. This pathway involved numerous steps, orthogonally and selectively protecting each pair of alcohol groups, ending with an oxidation to form the di-acid (Fig. 20)
Figure 20: Orthogonal protecting strategy for the synthesis of the mannaric acid core.
The paper described L-mannonic 𝛾-lactone as starting material, however, in order to save time and due to the readily availability mannitol was chosen in instead. Nevertheless, mannitol could be synthesised by reducing mannose using sodium borohydride in quantitative yield.20 The first part of the
pathway involved selectively protecting the 3,4-hydroxyl groups (red) of mannitol (Fig. 20-21), which was done in 2 steps. An acetonide was chosen as the desired protecting group. For its installation, firstly a tri-acetonide 48 was prepared using DMP as solvent and reagent, and camphorsulfonic acid (CSA) as acid catalyst, giving a yield of 86%.23 The same compound could also be obtained using acetone, DMP
and H2SO4 as acid catalyst but due to lower yields of 75% the former route was favoured.24 The next
step was the selective removal of the outer acetonide groups while leaving the internal one intact. Literature research showed that this could be performed using acetic acid, however, the reaction time and acid concentration varied in each procedure. It was found that by varying the scale the reaction time had to be altered. This is likely due to the purification involving evaporation of the water and acetic acid in a rotary evaporator with the deprotection continuing during the process. Eventually a yield of 84% of the 3,4-mono-acetonide 49 for a batch of 10g starting material was obtained (Fig. 21). For this the tri-acetonide was suspended in 70% acetic acid and heated at 40°C for 1:30 hours, following by rapid evaporation (2mbar) of both the water and acetic acid at 40°C on the rotary evaporator.25
Following, d-mannitol and the di-acetonide by-products, which were formed by over and under-deprotection were removed by extraction and recrystallisation. It must be noted that the reaction time had to be altered to ~105min for batches of 3g, reflecting the time needed to evaporate all the solvent.
Figure 21: Selective protection of the 3,4-hydroxy groups of D-mannitol
For the methylation of the 2,5-hydroxyl groups (black), first the 1,6-hydroxyls (blue) (Fig. 16) had to be protected. It was decided to use large groups as steric hinderance would lead towards selectivity of the primary alcohols. In literature, tert-butyl-di-methyl-silyl groups (TBDMS) were used, however, the protocols differed with each article. The chosen protocol was based on the highest yield and consisted of slow addition of TBDMSCl to 49 at 0°C in the presence of imidazole (Fig. 18).26 This
gave a yield of 88% of a mixture of 1,6-silyl-protected 50 and a second compound of identical mass. It’s suspected that this compound is the 1,2-silyl-protected compound based on the presence of a second pair of peaks in the 1H-NMR spectrum in the TBDMS region. This by-product constituted around 10%
of the total yield and could sadly not be removed by column purification. The triple protected compound was also detected but it could be isolated. Usage of an even larger silyl protecting groups such as tert-butyl-di-phenyl silyl resulted in a lower yield while not improving the selectivity for the primary alcohols. OH O OH OH O HO O O Alkyl Alkyl OH O OR1 OR1 O HO Alkyl Alkyl OR2 OH OR1 OR1 OH R2O OH OH OH OH OH HO Mannose [O] OH OH OH OH OH HO D-mannitol (quant.) O O O O O O CSA DMP, r.t. 48 86% AcOH 70% 40°C O O OH HO HO OH 49 84% O OH OH OH OH HO D-mannose NaBH4 H2O, r.t.
With only the desired 2,5-hydroxyl groups unprotected, methylation could be performed using methyl iodide and sodium hydride giving a yield 78% for the di-methylated product 51 (Fig. 22).20 The
1,2-silyl-by-product could neither be removed in this step. Following the methylation, the silyl groups had to be removed for oxidation to proceed. This was done using tetra-butylammonium fluoride (TBAF) in THF, luckily by tedious column purification the previously named by-product could finally be separated. This gave a yield 64% for the pure 2,5-Di-O-methyl-3,4-O-isopropylidene-L-mannitol 52 as an oil that solidified upon standing.20
Figure 22: Selective protection of primary alcohols and synthesis of diol 52.
The free primary alcohols could now be oxidised to an acid. For this the procedure of Alterman was followed. It described a bi-phasic multi-component oxidation using TEMPO and a sodium hypochlorite/bromide system. The oxidation mechanism is shown in figure 23.27,28
An excess of hypochlorite (7.8 equiv.) is slowly added in order to oxidise the catalytic amount of bromide (0.186equiv.) into hypobromite, which in turn, oxidises the TEMPO (0.05 equiv.) into the oxoammonium salt. This oxoammonium salt then oxidises the alcohol towards the aldehyde. Comproportionation of the formed hydroxylamine with another equivalent of oxoammonium results in the recovery of TEMPO finishing the catalytic cycle. The aldehyde then oxidised by a second molecule of hypobromite into the carboxylate. Addition of base and phase transfer catalyst ensures proper distribution of hypohalites in the organic phase while the temperature is kept at 0°C to ensure stability of the reactive oxidative species. The diacid 53 was obtained in a yield a 43% yield with acceptable purity (Fig. 24). It is unknown what the impurities could be.
Figure 23. Mechanism of TEMPO mediated alcohol oxidation by a hypochlorite-bromide system.
At first esterification of the carboxylic acids and removal of the acetonide in one step was attempted using methanolic HCl at 60°C. However, this led to a mixture of unknown compounds and no product could be isolated. Therefore, a twostep approach was taken (Fig. 24). First the carboxylic acid groups were methylated using iodomethane and potassium carbonate 54, which gave the purified desired ester in a 40% yield. The acetonide could then be deprotected using methanolic HCl with a yield of 50% for
38
Figure 24. Oxidation of diol 33 and final steps towards the core 38
O O OH HO HO OH 49 TBDMSCl, Imidazole DCM, 0° —> r.t. O O OH TBDMSO HO OTBDMS 50 88% NaH, MeI THF, r.t. O O O TBDMSO O OTBDMS TBAF 51 78% THF, r.t. O O O HO O OH 52 64% Br -OBr -OCl -Cl -Br -OBr -N+ O N O N OH OH R R O R O O
-Alcohol Aldyhyde Carboxylate
O O O HO O OH 52 O O O HO O OH 53 43% O O TEMPO, KBr TBAB, NaHCO3 NaOCl, NaCl DCM, 0°, 2h. MeI, K2CO3 DMF O O O O O O O O 54 40% MeOH HO OH O O O O O O 38 50% 7% HCl
1.2. Ozonolysis pathways:
Despite successful synthesis of the core 38 using the mannitol route an alternative pathway was explored due to poor yields/selectivity during the silylation, oxidation and following steps. Ozonolysis is an alternative frequently employed method for the formation of carboxylic acids. Despite requiring an oxidative workup in order to obtain the carboxylic acid the procedures found in literature were easier to perform compared to the TEMPO-oxidation, requiring only one oxidant instead of a complicated biphasic system. In this section, two pathways employing ozonolysis for the carboxylic acid formation are described. The first pathway involves ozonolysis of a cyclic alkene while the second one that of a terminal alkene.
1.2.1: 1,4-benzoquinone pathway
In literature a stereoselective 2 step procedure (Fig. 26) for the formation of compound 55 (Fig. 25) was described starting from 1,4-benzoquinone.29 The first step involved
brominating of one of the double bonds. This was performed at 0°C using only 1equivalent of bromine to ensure bromination of only one of the double bonds
56. This reaction was followed by a stereoselective reduction of the alcohols
using sodium borohydride at -15°C, which gave a yield of 94% (Fig. 26). Slow addition of sodium borohydride and the low temperature ensure that only the
all-trans compound 55 was obtained.
Figure 26. Synthesis of compound 57 starting from 1,4-benzoquinone.
The goal of the following steps was to selectively alkylate the alcohol groups and replace the bromides with orthogonal protected alcohol groups. After protecting of the newly formed hydroxyl groups ozonolysis could be performed resulting in a ring opening to yield the mannaric acid core 53. In the paper of L. Kelebekli such a procedure was described by transformation of the alcohols into methoxide groups and the bromides into hydroxyl groups 57 in one step using freshly prepared anhydrous sodium methoxide.29 The mechanism is shown in figure 27. Firstly, the alcohol groups are
deprotonated and an intramolecular nucleophilic substitution occurs forming an epoxide. This epoxide is then opened by a methoxy anion on the more electrophilic, allylic side resulting in the desired product
57 and the inversion of all stereocenters.
Figure 27: Mechanism for the formation of compound 57.
Unfortunately, the compound 57 was not obtained. It is unknown what exactly went wrong however, no pure product could be isolated and the 1H-NMR showed no characteristic peaks. It was therefore
decided to abandon this pathway, since a more feasible alternative had been found.
O O Br2 0°C O O Br Br NaBH4 Et2O, -15°C —> r.t. OH OH Br Br 56 (quant.) 55 (94%) O O OH OH 57 (0%) Na MeOH 0°C —> r.t. OH Br Br HO NaOMe O -Br Br -O O O MeO -OH HO MeO OMe 57 Figure 25: Compound 55 OH Br Br HO
1.2.2: Dimethyl-tartrate pathway
In literature a paper reported on the synthesis of an all trans-tetra-hydroxy-terminal-octadiene 58 starting with the acetonide of L-dimethyl-tartrate (DMT). Synthesis of this compound followed by ozonolysis could lead to the formation of di-acid 53 (Fig. 28)
The first step in this pathway was the formation of the acetonide. This was performed using by
L-dimethyl-tartrate (DMT) and pTsOH and yielded compound 59 in a 85% yield.30 L-Diethyl tartrate (DET) could
also be used as a starting material but DMT was preferred due to the formation of mixed ester acetonides when using DET. In the next step the ester groups were transformed into vinylic alcohols in a One-Pot reaction (Fig. 29). Firstly the esters were carefully reduced into aldehydes using dropwise addition of DIBAL-H at -78°C.31,32 By keeping the temperature low and limiting the reaction time to 2-2.5h.
overreduction could be avoided. This was followed by dropwise addition of divinyl zinc, which was freshly prepared by addition of vinyl magnesium bromide to anhydrous zinc(II) chloride. By slow addition of divinyl zinc at -78°C, it was ensured that high diastereoselectivtiy was achieved (>94% according to NMR) for the nucleophilic addition of the vinyl group.32 This gave the divinyl carbinol
compound 58 in 56% yield. It is hypothesised that the high diastereoselectivity is attributed to three factors. Firstly, and secondly, the lower reactivity of the organozinc reagents compared to the Grignard reagent combined with the slow addition and low temperature of the reaction ensure kinetic control. Finally, the presence of the acetonide is thought to play a role. It inhibits twisting of the backbone resulting in the presence of only one conformer of the aldehyde intermediate thus improving the selectivity.
Figure 29: Synthesis of the divinyl carbinol compound 60
The next step involved alkylation of the vinyl alcohols using sodium hydride and methyl iodide. A yield of 99% for compound 60 could be obtained by double addition of the alkylating reagents, otherwise the yield dropped to 75-80%. The alkylation was the followed by ozonolysis. Two different approaches were attempted only varying by the workup. Firstly, an oxidative workup procedure using sodium chlorite, which gave a yield of 55%. The second one was a two-step procedure with a reductive workup followed by an oxidation, giving a yield of 93% for diacid 53 (Fig. 30).18 In this procedure
ozone was bubbled through until saturation ensuring complete conversion of the alkene to the trioxolane. This was followed by the addition of triphenyl phosphine, reducing the trioxolane to the aldehyde and CO2. The aldehyde could then be oxidised towards the carboxylate using sodium chlorite.
2-Methyl-2-butene was added as hypochlorite scavenger. The high yield (93%) and high purity demonstrated the superiority of this procedure over the TEMPO oxidative one. The free acid groups could then me converted towards methyl esters. Instead of using K2CO3 and methyl iodide as before
TMS-diazomethane in DCM and MeOH was used (Fig. 30).33 There was no real reason for this change
except the simplicity of procedure. This gave a yield of 62% for the diester 54 compared to the previous 43%. The acetonide could then be removed using diluted 37% HCl with MeOH (7% HCl) giving a yield of 70% of the 38.20
Figure 30: Synthesis of diol 38 by ozonolysis. O O OH HO O O O O O O 1. DIBALH, PhMe, -78°C 2. Zn(II)Cl2, C2H3MgBr, THF, -78°C —> 0°C DMP, CSA DCM, r.t. HO OH O O O O L-dimethyl-tartrate 59 58 NaH, MeI THF, r.t. O O O O 60 1. O3, PPh3, DCM 2. 2-methyl-2-butene, NaH2PO4, NaOCl2, tBuOH, H2O O O O O O O OH HO O O O O O O O O 53 (73%) 54 (62%) TMSCHN2 DCM:MeOH, r.t. O O O O 60 HO OH O O O O O O 38 (70%) 7% HCl MeOH, r.t.
Figure 28: Proposed synthesis of the diacid 53 from an di-hydroxy-allyl acetonide 58.
O O OH HO O O O HO O OH O O 58 53
Chapter 2: Synthesis of the ring building blocks
18In this chapter the synthesises of the building blocks for the formation of the pre-catenane are described. These include the terephthalic acid template, the temporary linkages and the ketone ring strand. All of these compounds were synthesised following the procedures of L. Steemers and S. Pilon
et al..
The ketone ring strand could be synthesised in three steps (Fig. 31). Starting from ethyl formate and two equivalents Grignard reagents of 9-bromo-1-nonene a symmetrical alkene alcohol 44 was obtained in 81% yield. This was followed by the transformation of the alkenes into alkynes in two steps. This was done by bromination of the alkenes towards the dibromides followed by elimination with LDA to yield the di-alkyne alcohol 43 in 69% yield. Despite full conversion for the bromination (confirmed by 1H-NMR) the dialkene and the unsymmetrical alkene-alkyne alcohol could be detected during
purification. It is unknown how dibromide is converted back to the alkene, though it is suspected this might go via lithiation of the intermediate vinyl bromide followed by formation of the alkene during workup. Nevertheless, 43 could obtained and was oxidised to the pure ketone 39 using PCC in a 73% yield.
Figure 31: Synthesis of the ketone ring thread 39
Likewise, the temporary linkages were synthesised over three steps (Fig. 32). Firstly, 10-Undecen-1-ol was converted into 10-undecen-1-amine 41 in two steps by means of a Gabriel amine synthesis with an overall yield of 40%. In this way a primary amine could be made without the formation of an ammonium salt. This was followed by a reductive amination of the corresponding commercially available aldehyde 40 yielding the temporary linkage 37 in 98%.
Figure 32: Synthesis of the temporary linkages 37.
For the synthesis of the terephthalic acid template lab synthesised 1,4-bis(bromomethyl)-2,5-dimethylbenzene 61 was oxidised towards 2,5-dimethylterephthalaldehyde 62 by performing a Hass-Bender oxidation (Fig. 33). This oxidation uses the sodium salt of 2-nitropropane to transform the bromides into aldehyde groups in a 46% yield. This reaction was followed by a second oxidation using vanadium oxide and hydrogen peroxide, in order to oxidise the aldehydes into ethyl esters 63 in a 45% yield. The next step involved the bromination of the benzylic methyl groups using NBS, which gave a yield of 59% for the diethyl 2,5-bis(bromomethyl)terephthalate 64. Next, the bromines were substituted by azides 65 overnight using sodium azide in a yield of 82%. In order to make an activated ester 65 was hydrolysed using Lithium hydroxide in a quantitative yield and then transformed into pentafluorophenol (Pfp) esters 35. The low yield is likely due to the pentafluorophenol ester degrading on the column and rotary evaporator during workup.
Br THF, 70° 1. Mg, I2 2. Ethyl Formate OH 1. Br2, DCM 2. LDA, THF OH PCC, SiO2 O DCM, r.t. 44 (81%) 43 (69%) 39 (73%) OH N O O NH2 H N H2NNHMe HO OMe 1. 46 2. NaBH4 MeO OH O 40 41 (46%) 37 (98%) undec-10-en-1-ol Phtalimide, PPh3, DIAD THF, r.t. (85%) THF:EtOH MeOH
Figure 33: Synthesis of the terephtalic acid template 35.
Chapter 3: The road towards the pre-catenane.
With all the building blocks at hand it was possible to combine them step by step in order to synthesise the desired pre-catenane. Firstly, the ketal of 39 and diol 38 was made (Fig. 26), for this, activation of the ketone was required by formation of a dimethoxy acetal 67. The formation of this active species had to be done separately as attempts at in-situ formation of acetal did not yield the desired ketal product. The activated acetal was quantitatively made using trimethyl orthoformate under acidic conditions (Fig. 34). Addition of the diol-core 38 and 67 with CSA as acid catalyst at 50°C in DCM resulted in the formation of 36 in a 70% yield. Control of temperature and reaction time was important, as was observed that letting the reaction run in THF at 70°C over-weekend resulted in; a. racemisation of the methyl ether stereocenters, b. elimination of one of the methyl ethers (Fig. 35A) of the product.
Figure 34: Two step synthesis of the ketal 36 via a dimethoxy acetal.
Next, was the installation of the temporary linkages. Firstly, the methyl esters were hydrolysed towards the carboxylic acids. Under the conditions described in the paper of L. Steemers an undesirable secondary reaction was occurring. Instead of saponification, the entire ketal moiety would undergo elimination, yielding a hydroxy-a,b-unsaturated ester and 39 (Fig. 35B). However, by using more dilute conditions and carefully controlling the pH during workup, elimination could significantly be avoided and yielding the di-acid (68) in 95% yield.
Figure 35: Acid (A) and Base (B) promoted degradation pathways of 36
The carboxylic acids could now be transformed into amides with amine 37. The first attempt was performed using DiPEA and PyBOP as the coupling agent and letting the reaction stir at 30°C, unfortunately, only the amine 37 was isolated. It was decided to use the procedure used by L. Steemers
Br Br O O O O OEt EtO O O OEt EtO Br Br O O OEt EtO N3 N3 O O OH HO N3 N3 O O OPfp PfpO N3 N3 Pfp = F F F F F 61 62 (46%) 63 (45%) 64 (59%) 65 (82%) 35 (10%) DMF, 0°C 2-nitropropane NaH V2O5, H2O2 EtOH, HCl 0°C —> r.t NBS, 500W DCM, 50°C NaN3 DMF, r.t. LiOH THF:MeOH :H2O 66(quant.) DCM, r.t. HBTU. DiPEA PfpOH O O O O O O O O 7 7 CSA THF, 70°C, over weekend O O O O O O O O 7 7 O O O O O O O 7 7 + 1:1 O O O O O O O O 7 7 NaOH THF:MeOH:H2O OH O O O O O O H HO A B 7 7
et al. using HOBt, DCC and DMAP for the coupling. This reaction creates in situ an activated ester of
HOBt and 68, which in turn reacts with the amine forming 34 in 60% yield (Fig. 36).
Figure 36. Installation of the temporary linkages 37 onto ring precursor 68
The amide formation was followed by the installation of the terephthalic acid template (Fig. 37). Addition of the PfP-ester template 35 to 34 in the presence of Cs2CO3 as the base and heating the
suspension at 40°C overnight gave compound 34 in a 57% yield. Molecular sieves were added to ensured water-free conditions. Additionally, the reaction was run at high dilution (2.3mM), which was done in order to avoid oligomerization between the starting materials thus ensuring that one equivalent molecule 35 would react with only one equivalent of 34.
Figure 37: Attachment of the terephtalic acid template 35 onto 34
However, when the 1H-NMR spectrum was recorded a surprising observation was made (Fig. 38).
When observing the proton spectrum, a number of signals were missing, foremostly the three signals originating from the mannaric acid core (methyl ethers and the backbone), and the benzylic CH2 signals
of the backfolding template and temporary linkage. Meanwhile the signals assigned to the alkene group (7.4-5.0ppm), aromatic methyl ether (3.9ppm), terminal alkyne proton (1.9ppm) and the CH2 protons
neighbouring the alkyne (2.04ppm) were present, showing that the ketal group was still present in some form. Furthermore, conformation from Infra-Red spectroscopy (IR), showed a strong absorption at 2100cm-1 characteristic for azide groups, from which their presence could be determined. Mass
spectroscopy showed the presence of a compound with isotope peaks at m/z of 1309.7331 and 1310.7331,corresponding to the desired product, from which could be concluded that the template was successfully installed. Currently, the only explanation for the disappearance of the 1H-NMR signals that
can be given is that due to the formation of a cyclic system, the signals have been broadened out, to the point that they are hard to recognise, however, further analysis must be performed to achieve a conclusive answer. O OPfp O PfpO N3 N3 35 OH O N O O O O HO O N O O 7 7 9 9 34 35, CsCO3. O O N O O O O O O N O O O O N3 N3 33 (57%) 7 7 9 9 ACN, 40°C
Figure 38. Attachment of the terephtalic acid template comparison of 1H-NMR spectra of 34 and 33
The next steps involved the formation of two additional ring systems (Fig. 39) Firstly, by means of the templated backfolding Copper(I)-catalysed Azide-Alkyne Cycloaddition (CuAAC). This was once again performed using highly diluted conditions (2.06mM) as to avoid poly/oligomerization using Cu(I)(MeCN)4BF4 as the catalyst and TBTA as a ligand for enhanced catalytic activity. The reaction
was performed under an oxygen free environment as to avoid oxidising the Cu(I) into Cu(II) and followed by means of IR-spectroscopy as the absence of the 2100cm-1 azide peak indicated full
conversion of the starting material. This gave a yield of 13% of compound 32. Since 1H-NMR was
inconclusive, a mass analysis was performed, which showed the presence of a compound with m/z 1309.7876, confirming the formation of the desired product. The CuAAC was followed by a ring closing metathesis (RCM) for the formation of the second ring using a Grubbs second generation catalyst. This gave a brown oil in a 33% yield 31, which was analysed using mass spectrometry giving a mass of 1281.7184 (calculated mass 1281.7539). While 1H-NMR was still inconclusive, from the
mass it is hypothesised that the desired ring was indeed formed, however, further analysis is required for a final conclusion. Unfortunately, as time was running out and the small amount of remaining product, it was decided to end the project here only 3 steps away from the final goal.
Figure 39. Two backfolding cyclisation in order to form an intertwined ring system.
O O N O O O O O O N O O O O N3 N3 33 (57%) 7 7 9 9 TBTA, Cu(CN)4BF4 DCM, reflux, on O O O O O N O N O O O O O O O N N N N N N 32 (15%) Grubbs Gen II O O O O O N O N O O O O O O O N N N N N N 31 (33%) DCM
Conclusion
During this project, an attempt at the synthesis of pre-[2]catenane was made, with as goal to see the effect of making the central ketal template electron richer on the cleavability.
Originally, the synthesis envisioned the oxidation and condensation of mannose towards the di-lactone as the strategy towards the mannaric acid core building block. However, due to the inability to alkylate the di-lactone alternative pathways had to be dived. Successful synthesis of the mannaric acid core was achieved in two ways starting either from mannitol and a performing multiple regioselective protection and deprotection steps ending with a TEMPO-NaOCl oxidation. Alternatively, L-dimethyl tartrate could be used as starting material. This pathway involved a selective reduction of the ester towards aldehydes following by a stereoselective reaction with an organozinc reagent. In this approach the carboxylic acid groups were formed by ozonolysis.
The other building blocks were synthesised following the procedures of L. Steemers and S. Pilon. The ketone ring thread was formed in three steps starting from a reaction between ethyl formate with a Grignard reagent, followed by the transformation of the alkynes into alkenes and ending with an oxidation. Similarly, the temporary linkages were synthesised by means of a Gabriel amine synthesis and a reductive amination. Finally, the backfolding template was obtained in six steps starting from 1,4-bis(bromomethyl)-2,5-dimethylbenzene.
Attachment of the ketone thread to the mannaric acid core, forming the ketal moiety was successfully performed, followed by the installation of the temporary linkages in two steps. Caution had to be taken as the ketal was subject to base and acid catalysed elimination. The attachment of the terephthalic backbonding template however, proceeded with less certainty. Based on certain signals in the 1H-NMR spectrum, the IR azide peak and Mass spectrometry data it can be assumed that the
template was successfully attached. However, the lack of specific characteristic peaks in the 1H-NMR
spectrum shines doubt upon these results. Therefore, re-synthesis and further analysis of this compound must be performed. The same can be concluded for the CuAAC macrocycle and the RCM macrocycle. While the masses coincide with the desired product the 1H-NMR spectra are unclear and therefore no
Discussion & Future plans
In order to understand the uncertainty regarding the installation of the terephthalic template it would be worthwhile to resynthesise compound 33. Confirmation of successful synthesis would simplify the data analysis of the CuAAC and Grubbs macrocyclization steps. These steps would be followed by the removal of the temporary linkages and hydrogenation of the macrocycle, which would result in the desired target compound. Finally, the cleavability of the ketal may be tested using different acidic conditions.
Due to the inherit chirality of lasso peptides asymmetry into the template has to be introduced. In a yet unpublished paper, it was shown that the tartaric acid core could be cleaved, in the absence of amide on the second ring. To mimic the directional chirality of the peptides the synthesis of an asymmetric ketal may be investigated. As alternative to tartaric acid threonic acid may be employed. This sugar acid possesses similar hydroxyl conformation as that of tartaric acid in combination with the asymmetric properties that of an amino acid. Transformation of the terminal hydroxyl group into a tertiary amine may be employed in order to attach the temporary linkages.
Other templates that resemble amino acids may also be of interests. For example, a 4-imidazolidinones moiety could be used as the tetrahedral template, instead of the ketal moiety. It possesses a similar geometry ensuring proper organisation. Replacing the terephthalic template with something that resembles more natural occurring motifs might also be worth investigating. A hydantoin
69 may be a viable alternative, which possess a flat geometry similarly to an aromatic ring. Hydantions
may be synthesised from basic amino acids and (iso)cyanates, allowing for flexible variation in the side chain residue of the amino acid. Furthermore, acid hydrolysis results in the liberation of the amino acid and an amine, which can alternatively be performed enzymatically. This would allow for simple transformation of the template into an amino acid chain, mimicking the structure of lasso peptides. Additionally, it would result in cleavage of the ring, resulting in a lasso peptide structure.
Figure 40. Hydantion based backbonding template strategy. With R1 and R2 being chains containing
groups, which partake in the ring closing. N N O O R 2 R1 R3 HO O NHR1 R3 hydantion template 69
Amino acid (R3 = residue)
Hydrolysis HO O NHR1 R3 + HNR2 R2 N C O +
Acknowledgement
I would like to thank Jan van Maarseveen for giving me the opportunity to do my Master Research in the Synthetic Organic Chemistry group. Thank you Simone for being my daily supervisor, giving me advice, answering my question and helping me set up various reaction apparatus. Additionally, I want to thank Joost Reeks for being my second reviewer of this thesis.
A big thank you for Nick and Bas for the support on the lab and assisting me when required. Thank you Bart for advice on reactions and for the automatic column machine. Lastly I wish to thank the entire SOC group for the good and fun time during this project.
Supporting information
1.1 Experimental
1.1 General procedures:
Oxygen and/or water sensitive reactions were performed in oven 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-NMR spectra were recorded on a Bruker DRX 300, 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, dd = double doublet.
1.2 Synthesis of the Diol core:
L-Mannaro-1,4:3:6-di-𝛾-lactone (46)20:
Mannose (2g, 11,1mmol, 1equiv.) and NaNO2 (55mg, 0.79mmol, 0.07equiv.) in H2O (2.5mL) were
added to 65% HNO3 (9mL) and heated at 85°C with evolution of nitrous gasses over 16h. The solution
was concentrated at 50°C to a thick syrup then dissolved in 20mL H2O and concentrated again this was
repeated a second time. The concentrated was suspended in Et2O (30mL) and EtOH (5mL) overnight.
The precipitate was filtered, washed with Et2O and dried, yielding a white solid (0.524g, 3.70mmol,
33%). 1H-NMR (DMSO-d6, 400MHz) ẟ 6.42 (d, OH), 5.03 (d, 2H), 4.79(d, 2H).
2,5-Di-O-methyl-L-Mannaro-1,4:3:6-di-𝛾-lactone (47a-b):
To NaH (60%, 50mg, 1.2mmol, 2.5equiv.) in DMF (4mL) under N2 at 0°C. 46 (69mg, 0.48mmol,
1equiv.) in DMF (1mL) added dropwise and allowed to warm up to r.t. Added MeI (0.1mL, 1.44mmol, 3equiv.) dropwise and stirred at r.t. overnight. Quenched with H2O & extracted with Et2O (3×15mL),
washed with brine, dried and concentrated. Column 1:3 PE:EA. (43.7mg, 0.54mmol, 45%) O O O O OH HO H H O O O O OH HO H H NaH, MeI DMF, r.t. O O O O O O H H O OH O O O O H NaH, MeI DMF, r.t.
1,2:3,4:5,6-Tri-O-isopropylidene-L-mannitol (48)34
Mannose (2.0g, 11.2mmol, 1equiv), in H2O (30mL). Added NaBH4 (0.53g, 14mmol, 1.2equiv) and
stirred at r.t. overnight. Acidified with AcOH and concentrated, redissolved in MeOH (30mL), acidified and concentrated, yielding white crystals in quantitative yield. D-Mannitola (10g) was suspended in
DMP (110mL) and p-TSOH was added and stirred at r.t. overnight. Neutralised with Et3N and
concentrated. H2O (15mL) was added to residue and extracted with CHCl3 (3x 40mL), dried and
concentrated. Recrystalised in PE to yield white crystals (14.3g, 47.4 mmol, 86%).b1H-NMR (CDCl 3,
400MHz) ẟ 4.20 (m, 2H), 4.09 (m, 2H), 3.98 (m, 4H), 1.44 (s, 6H), 1.41 (s, 6H), 1.37 (s, 6H).
a) Commercially purchased D-mannitol was preferably used. b) Procedure using H2SO4 in Acetone and DMP gave
yields of 75%.35
3,4-O-isopropylidene-L-mannitol (49)36
48 (10ga, 33mmol, 1equiv) was suspended in 200mL 70% HOAc and stirred for 1.5h at 40°C. The
solved was evaporated as quickly as possible (~35mbar, 40°C) and remaining HOAc was coevaporated with toluene. The residue was dissolved in hot acetone and filtered. Filtrate was concentrated and the syrup was recrystalised in PhMe to yield white/clear needle like crystals. (6.1g, 27.4mmol, 84%). 1
H-NMR (MeOD-d3, 400MHz) ẟ 3.96 (m, 2H), 3.79 (m, 2H), 3.63 (m, 4H), 1.39 (s, 6H). a) Performing this
reaction on a different scale required different reaction time in order to achieve a similar yield.
1,6-Di-O-tert-butyldimethylsilyl-3,4-O-isopropylidene-L-mannitol (50)
49 (6.1g, 27.4mmol, 1equiv.), imidazole (4.29g, 63.02mmol, 2.3equiv.) were dissolved in dry DCM
(33mL) and cooled to 0°C. A solution of TBDMSCl in dry DCM (16.8mL) was added dropwise and the solution was stirred for 2h at 0°C. Poured into 30mL H2O and extracted with DCM (3x 40mL),
washed with H2O (2x15mL), brine (10mL), dried and concentrated. Column purified by hexane:EA
9.8:0.2. yielded a clear oil. (11.4g. 25.0mmol, 88%)a. 1H-NMR (CDCl
3, 400MHz) ẟ 3.90 (m, 4H), 3.72
(m, 2H), 3.66 (m, 2H), 1.38 (s, 6H), 0.91 (s, 18H), 0.11 (s, 12H).
a) Minor impurity of 1,2-silyl Regi-isomers
O O O O O O HO OH O O OH OH TBDMSO OH O O OH OTBDMS
2,5-Di-O-methyl-1,6-di-O-tert-butyldimethylsilyl-3,4-O-isopropylidene-L-mannitol (51)5
NaH (60%, 2.3g, 54mmol, 2.5equiv.) in dry THF (130mL) at 0°C under N2. 50 (9.82g, 21.6mmol, 1
equiv.) in 8mL dry THF added dropwise and stirred for 15min at r.t.. MeI (4.1mL, 64.8mmol, 3equiv.) added ad stirred overnight at r.t.. H2O (20mL) added and THF evaporated, extracted with Et2O
(3x45mL), washed with H2O (15mL), brine, dried and concentrated. Column purified PE:EA 17:1
yielded 51 as a clear oil (8.18g, 16.9mmol, 78%). 1H-NMR (CDCl
3, 400MHz) ẟ 4.08 (m, 2H), 3.86 (m,
2H), 3.68 (m, 2H), 3.50 (s, 6H), 3.36 (m, 2H), 1.38 (s, 6H), 0.94 (s, 18H), 0.07 (s, 12H).
2,5-Di-O-methyl-3,4-O-isopropylidene-L-mannitol (52)5
51 (1.90g, 3.95mmol, 1equiv.) in dry THF (20mL). TBAF (1M THF) added and stirred at r.t. overnight.
Concentrated and columned with MeCN. The crude product was purified by column chromatography PE:EA 1:1 à 1:5 à 0:1 to yield a clear oil that crystalised overnight to form wax like white-clear solid (0.631g, 2.52mmol, 64%). 1H-NMR (CDCl 3, 400MHz) ẟ 4.02 (m, 2H), 3.72-3.55 (m, 4H), 3.25 (s, 6H), 3.25 (s, 2H), 3.14 (m, 2OH), 1.25 (s, 6H). 13C-NMR (CDCl 3, 75MHz) ẟ 109.8, 82.2, 78.3, 60.6, 58.0, 27.1; IR (cm-1): 3405, 2935, 1371, 1213, 1054, 874, 844, 509. HR-MS Calcd for C 11H22O6 (M+H) 251.1495 found: 251.1489.
2,5-Di-O-methyl-3,4-O-isopropylidene-L-mannaric acid (53) by alcohol oxidation
To 52 (0.631g, 2.52mmol, 1equiv) in DCM (34mL) TEMPO (20mg, 0.126mmol, 5 mol%), KBr (54.7mg, 0.47mmol, 0.186 equiv.), TBAB (85mg, 0.263mmol, 10 mol%) and 10.3mL sat NaHCO3
were added and cooled to 0°C. A solution of NaHCO3 (5.55mL), 1.2M NaOCl (19.63, 7.8 equiv.) and
brine (11.1mL) was added dropwise over 45min. After addition stirred for another 45min at 0°C. Layers separated and DCM extracted with H2O (3x20mL) then aqueous layer acidified with 2M HCl and
extracted with EA (3x30mL), dried and concentrated. A clear oil was obtained (0,312g, 1.11mmol, 44%). 1H-NMR (DMSO-d6, 400MHz) ẟ 4.26 (d, 2H), 3.80 (d, 2H), 3.29 (s, 6H), 1.29(s, 6H); 13C-NMR
(DMSO-d6, 75MHz) ẟ 170.1, 110.5, 80.9, 78.1, 58.2, 26.7; IR (cm-1): 2990, 2938, 1733, 1457, 1375,
1216, 1102, 851, 697; HR-MS Calcd for C11H18O8 (M-H) 277.0923 found: 277.0941.
TBDMSO O O O O OTBDMS HO O O O O OH HO O O O O OH O O
Compound 55
1,4-Benzoquinone (3g, 27.57mmol, 1equiv.) wad dissolved in CHCl3 (25mL) and cooled down to 0°C.
Br2 (1.43mL, 27.57mmol, 1equiv.)in CHCl3 (11mL) was added dropwise at 0°C over 1h then stirred
for 2h at r.t.. Concentrated gave a brown solid (56) (7.37g, 27.50mmol, 98%). %). 1H-NMR (CDCl 3,
400MHz) ẟ 6.74(2H, s), 4.82(2H, s).
56 (7.37g, 27.50mmol, 1equiv.) was dissolved Et2O (100mL) and cooled down to -15°C (salt
ice bath) NaBH4 (2.46g, 65.06mmol, 2.2equiv.) in 41mL H2O was added dropwise do the vigerous
stirred solution then stirred for another 2h at r.t.. The phases were separated and the aqueous phase was extracted with EA (4x40mL), dried and concentrated, gave a pink solid 55 (7.55g, 26.09mmol, 94%).
1H-NMR (CDCl
3, 400MHz) ẟ 5.74(2H, s), 4.44(2H, m), 4.12 (2H, m)
Compound 57
Sodium (2g, 87.31, 3.15equiv) was cooled down under N2 in a flask to 0°C and dry MeOH (100mL)
was added and stirred under all sodium was dissolved and bubble formation seized. 55 was added in small portion then stirred at 0°C for 1h and 2h at r.t.. NH4Cl (25mL) was added and the aquesous layer
was extracted with EA (3x300mL), dried and concentrated. Recrystalisation from DCM:Et2O yielded
a brown oil. No product found.
Dimethyl-3,2-O-isopropylidene-tartrate (59)
To Dimethyltartrate (3.06g, 17.2mmol, 1equiv.) in dry DCM (40mL) was DMP (17mL, 137.6mmol, 8 equiv.) and pTSOH (0.65g, 3.4mmol, 0.2equiv.) added and refluxed overnight under N2. The dark red
solution was neutralised with Et3N and concentrated. Added sat. NaHCO3 (15mL), and extracted with
brine (3x40mL), washed with sat. NaHCO3 (15mL), brine, dried and concentrated. Column purified
PE:EA 10:1 to give a yellow oil (3.19g, 14.62mmol, 85%). 1H-NMR (CDCl
3, 400MHz) ẟ 4.74 (s, 2H),
3.75 (s, 6H), 1.41 (s, 6H).
4,5-O-Isopropylidene-1,7-octadiene-3,6-diol (58).
To a solution of 59 (2.5g, 11.45mmol, 1 equiv.) in 30mL dry PhMe at -78°C was DIBAL (1M in toluene) (26.4mmol, 2.3 equiv.) added dropwise and stirred at -78°C for 2.25h. The mixture developed
OH OH Br Br O O OH OH O O O O O O O O OH OH
