University of Groningen Total synthesis of mycolic acids and site-selective functionalization of aminoglycoside antibiotics Tahiri, Nabil

(1)

University of Groningen

Total synthesis of mycolic acids and site-selective functionalization of aminoglycoside

antibiotics

Tahiri, Nabil

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2019

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Tahiri, N. (2019). Total synthesis of mycolic acids and site-selective functionalization of aminoglycoside antibiotics. University of Groningen.

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

(2)

531024-L-sub01-bw-Tahiri 531024-L-sub01-bw-Tahiri 531024-L-sub01-bw-Tahiri 531024-L-sub01-bw-Tahiri Processed on: 3-5-2019 Processed on: 3-5-2019 Processed on: 3-5-2019

Processed on: 3-5-2019 PDF page: 77PDF page: 77PDF page: 77PDF page: 77

Chapter 3:

Completion of the Methoxymycolic Acid

Total Synthesis

Part of this chapter will be submitted for publication:

N. Tahiri, P. Fodran, D. Jayaraman, J. Buter, T. A. Ocampo, I. Van Rhijn, D. B. Moody, M. D. Witte, A. J. Minnaard, manuscript in preparation

(3)

68

3.1 Introduction

Chapter 2 describes the synthesis of all five fragments, that are required for the total synthesis of the four mycolic acid diastereomers. These fragments consist of both enantiomers of fragment C, fragment B and the R,R configured isomer of fragment A (Figure 1). In the current chapter, the combination of the synthesized fragments to complete the synthesis of the methoxymycolic acids is described. The synthetic mycolates were then derivatized to their corresponding R- and S-glycerol mono mycolates (GroMM) and glucose monomycolates (GMM). The four mycolic acid (MA) diastereomers and respective GMMs were tested for their ability to activate T-cells.

Figure 1. Overview of the fragments synthesized in Chapter 2.

3.2 Coupling of fragments B and C

With multiple grams of the required fragments in hand, the endgame of the synthesis was initiated by combining fragments B and C by another Suzuki-Fu cross-coupling. The required alkyl boron species 1 (Scheme 1) could be obtained via straightforward hydroboration of the terminal olefin in fragment C. In order to maintain a high concentration in the hydroboration, we used solid 9-BBN dimer instead of the monomeric form in THF solution. The hydroboration proceeded fast and was generally finished within 3 h (confirmed by 1_{H-NMR), but leaving the reaction to stir overnight}

did not seem to have negative effects on the subsequent coupling step. This alkyl boron intermediate was used without any further manipulation in the cross-coupling reaction. The coupling product was obtained after overnight stirring in the presence of 7.5 mol% of the commercially available Pd(PCy3)2 catalyst and fragment B.[1] Although TLC

indicated good separation between the product and the less polar impurities, the column purification of the crude resulted in coelution of the impurities along with the product. Fortunately, the impurities did not interfere with the subsequent reductive deprotection of the pivaloate group with LiAlH4, and column purification of alcohol 3 resulted in the

isolation of pure material with good isolated yields (73%-78%) over two steps. Alcohol

3 was then transformed to sulfide 4 by means of a Mitsunobu reaction in excellent

(4)

69

Scheme 1. Combination of fragments B and C, and follow-up chemistry towards the sulfone.

The subsequent oxidation towards the sulfone 5, however, proved to be a real bottleneck in this synthesis. Initially, we aimed for an ammonium molybdate catalyzed oxidation employing hydrogen peroxide as the terminal oxidant. Oxidation in alcoholic solvents, as is usually the case in literature,[2,3]_{was not feasible, due to the extremely poor}

solubility of 4 in methanol or ethanol. Cosolvents such as THF[4]_{and DCM had to be}

added to solubilize 4. Of these tested cosolvents, THF gave the best conversions, but the reaction did not go to completion. Although full and clean conversions could eventually be achieved by several successive work-up and resubmission sequences, conversions varied considerably on larger scale, and the reaction often resulted in a mixture of sulfoxide, sulfide and sulfone. Furthermore, monitoring the reaction progress proved to be problematic. The starting sulfide 4 and the desired sulfone 5 coeluted on TLC, excluding TLC as convenient tool for monitoring the reaction progress. Therefore, the conversion had to be determined by 1_{H-NMR analysis of reaction samples. However, in}

some cases the samples taken from the reaction mixture were unreliable, since the reaction mixture was heterogeneous. This led us to believe that full conversion was achieved, while 1_{H-NMR analysis of the crude product, after workup, showed that this}

was not the case.

Therefore, we attempted to improve the oxidation towards a more reliable and practical procedure, by oxidation of recovered sulfoxide (in the synthesis of 5b) towards the sulfone. It seemed to us that this would be a representative model reaction, since in general the oxidation of a sulfoxide to a sulfone is harder than converting a sulfide into

(5)

70

a sulfoxide. Hence, being able to successfully oxidize the sulfoxide would imply that oxidation from sulfide to sulfone would be possible. Moreover, as described before, the sulfoxide and sulfone exhibited a large polarity difference, allowing progress monitoring by TLC.

Table 1. Conditions screening for the oxidation of the sulfoxide to the sulfone.

entry Conditions Yield Remarks

1 MCPBA (1.5 + 2 equiv.), 0 °C, 3 h n.d. no conversion

2 MCPBA (1.5 equiv.), rt, 16 h n.d. complex mixture

3 TAPC (19 mol%), H2O2 (14 equiv.),

THF, rt

n.d. no conversion

4 TAPC (19 mol%), H2O2 (14 equiv.),

THF, rt

n.d. partial conversion into complex mixture 5 RuCl3·H2O (5 mol%) NaIO4 (2 equiv.),

CCl4, MeCN, H2O, rt, 30 min

n.d. Major product was 6 6 RuCl3·H2O (2 mol%) NaIO4 (1 equiv.),

CCl4, MeCN, H2O, rt, 30 min

66% 10% 6 was formed

Surprisingly, oxidation of the sulfoxide in DCM with 1.5 equiv. of MCPBA at 0 °C did not result in any detectable conversion, and addition of 2 equiv. extra oxidant at this temperature did not have any effect (Table 1, entry 1). When the MCPBA oxidation was performed at rt, the sulfoxide decomposed to a complex mixture of unidentifiable products (entry 2). Also oxidation using substoichiometric amounts of TAPC[5]_(Figure

2a and entry 3) or stoichiometric amounts TAPC (entry 4) with H2O2 as terminal

oxidant resulted in no conversion after overnight stirring or gave minimal conversion along with the formation of significant decomposition products after 3 days, respectively. Finally, oxidation using 5 mol% RuCl3·H2O and 2 equiv. NaIO4[6] resulted

in complete oxidation towards the sulfone, but the methoxy functionality in the major product was concomitantly oxidized towards the ketone 6 as evidences by 1_{H- and}13

C-NMR (Figure 2b) (entry 5). This surprising product could probably allow for the late stage transformation of methoxy- towards ketomycolic acids, since it has been shown by Sharpless et al.[7]_{that alpha stereocenters are unaffected under these conditions, and}

therefore the methyl branch should maintain its stereointegrity. Decreasing the RuCl3·H2O loading to 2 mol% and the amount of NaIO4 to 1 equiv. favored formation

of the desired sulfone as the major product, but overoxidation could not be completely suppressed, and led to the formation of 10% keto sulfone 6. Separation of 6 and the desired product was possible by column chromatography and yielded the sulfone in 66% yield. Even though these results are promising, the risk of overoxidation was very likely and hard to suppress. The oxidation towards the sulfone was therefore performed

(6)

71

using the initial ammonium molybdate/H2O2 conditions for the remaining diastereomers

(Scheme 1).

Figure 2. A) molecular structure of TAPC. B) keto sulfone 6 formed by the overoxidation of 5.

3.3 Julia-Kocienski coupling of fragments BC and A

With gram amounts of the sulfones in hand it was time to perform the final key step in this synthesis, which consisted of the modified Julia-Kocienski coupling[8]_{of fragment}

BC and fragment A. We first explored the coupling using LDA as the base, which

resulted in a poor 22% isolated yield. Fortunately, by using LiHMDS, the isolated yield could be improved to 85% (Scheme 2).

Scheme 2. Julia-Kocienski coupling of the final fragments.

Unfortunately, for the synthesis of the three other diastereomers applying LiHMDS stock solutions led to irreproducible yields. In order to eliminate possible errors in the titration of the freshly prepared LiHMDS solutions, we attempted the use of crystallized LiHMDS instead. The LiHMDS crystals could easily be prepared by the addition of 1.6M BuLi to neat HMDS at -20 °C. The formed crystals could be filtered under inert atmosphere and stored in a glovebox for later use. Surprisingly, Julia-Kocienski olefination of 5a and Fragment A at -30 °C resulted in 78% yield on small scale and a gratifying 90% isolated yield of 7a on gram scale (Scheme 3). Julia-Kocienski olefination of 5c resulted in 81% isolated yield of 7d on gram scale. However, the Julia olefination of 7b under identical conditions provided a diminished yield of 54% on gram scale. The recovery of the fragments was difficult as they coeluted on column, but NaBH4 treatment of the aldehyde and sulfone mixture allowed the selective reduction of

the aldehyde to the corresponding alcohol and permitted the isolation of the sulfone and alcohol. The sulfone was resubmitted for an additional Julia-Kocienski olefination and resulted again in a 54% yield, yielding overall a 66% yield for olefin 7b.

(7)

72

Scheme 3. Large scale Julia-Kocienski couplings.

At this point we had successfully upscaled three out of four coupling reactions and only the coupling of 5c, which showed earlier to proceed in good yields on small scale (Scheme 2), had to be performed at larger scale. However, the declining trend in isolated yields hampered us from proceeding with the final large-scale coupling, and we therefore performed a small optimization study. In order to save material of the precious fragment A, butyraldehyde was used as a model substrate.

In order to exclude decomposition of the solid LiHMDS base, in situ prepared LiHMDS was used for the coupling of a small excess of sulfone 5d with butyraldehyde at -30 °C (Table 2, entry 1). 1_{H-NMR analysis of the reaction mixture indicated that 25% of}

product was formed, while still 50% of the original sulfone was intact. When the reaction was performed under identical conditions, but using the crystallized LiHMDS instead, a slightly improved conversion of 33% was obtained with 60% of the sulfone still intact (entry 2).

Table 2. Conditions screening for the Julia-Kocienski coupling.

entry Aldehyde Base Conv. remarks

1 butyraldehyde _{(1.2 equiv.)} _{in solution}LiHMDS 25% 50% sulfone left.

2 butyraldehyde

(1.2 equiv.)

solid

LiHMDS 33% 60% sulfone left.

3 butyraldehyde _{(5.0 equiv.)} _{in solution}LiHMDS 0% NMR of aldehyde showed partial oxidation to carboxylic acid.

4 _{(1.2 equiv.)}fragment A _LiHMDSSolid 50%a _{46% sulfone left}

5 _{(1.2 equiv.)}fragment A MgClTMP 75%

Aldehyde was added until color disappeared, followed by more base. This sequence was repeated

three times total.

(8)

73

In order to eliminate possible self-condensation of the sulfone, which is known to happen at higher temperatures,[9]_{a reversed order of addition of the sulfone to the base}

was performed (entry 3). Unfortunately, after addition of 5 equiv. of aldehyde, 1_H-NMR

showed no conversion towards the desired product, but analysis of the butyraldehyde showed partial oxidation towards the carboxylic acid. This led us to believe that using a large excess of aldehyde must had completely quenched the base. Because of its high tendency towards oxidation, butyraldehyde proved not to be a good model aldehyde and therefore fragment A was used again.

At this point we became aware of several distinct color changes during the reaction. Upon deprotonation of the sulfone, a very bright yellow color characteristic for the anion was formed. However, it was noticed on multiple occasions that the yellow color fully disappeared before complete addition of the aldehyde, indicating that the anion was quenched. Either too little base was added initially or the aldehyde or the sulfone contained a source of protons. With this knowledge, the coupling of sulfone 5d with fragment A was attempted again (entry 4) using freshly crystallized LiHMDS, but 1

H-NMR analysis of the reaction mixture after 30 min indicated only 35% conversion, while 0.75 equiv. of the originally 1.2 equiv. sulfone still being present. Addition of 1 equiv. extra base at this point resulted in an increase in conversion to 50%. Allowing the reaction to warm up overnight did not facilitate further conversion, but also did not interfere with the ester’s chiral alpha position according to 1_{H-NMR. Finally,}

deprotonation using the stronger MgClTMP base (entry 5) was attempted, but upon addition of the aldehyde the bright yellow color disappeared quite rapidly. Therefore, the addition of aldehyde was stopped and an additional 0.6 equiv. base was added in order to accomplish deprotonation of the sulfone. Then, addition of aldehyde was continued until the characteristic color disappeared again, and this sequence was repeated 1 more time. To our surprise an increased conversion of 75% was achieved by this unconventional addition sequence. Although this method worked out on small scale, we were hesitant to apply it on large scale, as it would be difficult to make this procedure reproducible. Fortunately, we had enough sulfone available, and by using a large access (2.8 equiv.) of sulfone compared to the aldehyde coupling partner, a 79% yield on large scale (Scheme 3, compound 7c) could be isolated along with 1.7 equiv. of recovered sulfone.

3.4 Finalization of the synthesis

After having achieved the successful coupling for all four diastereomers (Scheme 4), the synthesis was continued by the hydrogenation of the newly formed double bond. Since hydrogenation of an alkene next to a cyclopropyl group using Pd/C is known to result in cyclopropyl cleavage,[10]_{we used our in house developed diimide reduction protocol.}[11]

Unfortunately, the very hydrophobic nature of 7 induced the formation of micelles in H2NNH2·H2O/EtOH/THF, which reduced the exposure of 7 to the generated diimide

(9)

74

to this solvent mixture prevented micelle formation, and thus more reliable conversion rates were achieved. Separation of 7 and 8 was impossible and the conversion had to be pushed to completion before continuing with the next step. Fortunately, with this improved solvent system, conversion was complete after 5 days. Purification over a plug of silica was sufficient to remove small traces of catalyst that were still left after extractive workup, resulting in near quantitative yields for all diastereomers.

Scheme 4. Completion of the mycolic acid total synthesis.

Unfortunately, TBS deprotection with NBu4F resulted in elimination of the alcohol,

yielding the α,β-unsaturated ester. But the deprotection using HF in pyridine went smoothly and resulted in excellent yields for all isomers. Removal of the chiral auxiliary was unsuccessful by catalytic hydrogenation or hydrolysis using LiOH or LiOOH,[12]

but could be achieved efficiently in the presence of 2 equiv. Bu4NOH at rt and overnight

stirring. An organic MeCN/pentane extraction of the crude resulted in the free carboxylic acids in near quantitative yield and high purities. Although column purification was not necessary, a chloroform/methanol mobile phase proved to yield the best recovery of the mycolic acids. Overall, the total synthesis of mycolic acid was

(10)

75

achieved in 17 steps in the longest linear sequence, with a 12% yield (the average of the four isomers).

3.5 Synthesis of R- and S-GroMMs, and GMMs

The synthesis of glycerol monomycolates (GroMMs) and glucose monomycolates (GMMs) has been described before by nucleophilic substitution of potassium[13]_and

cesium[14,15]_{mycolates on mesyl solketal and tosyl glucose at 70 °C, respectively.}

Nevertheless, the development of a high yielding esterification procedure which proceeds faster and under milder reaction conditions is still highly desired. With this in mind, we attempted a Yamaguchi esterification[16]_{of 10 with racemic solketal. Although}

we realized that the probability for intramolecular esterification was very likely (formation of a β-lactone), we envisioned that by using a large excess of solketal, the reaction towards the intermolecular ester product could be favored. The reaction of 10a with 5 equiv. of racemic solketal seemed to proceed relatively cleanly, but the reaction did not proceed to full conversion after overnight stirring, and resulted in a 36% isolated yield of slightly impure material (Scheme 5).

Scheme 5. Yamaguchi esterification of 10a with solketal.

Our next attempt involved the esterification by means of a Mitsunobu reaction (Scheme 6). Unfortunately, the reaction of 10a with 10 equiv. solketal did not result in the esterified product. On TLC formation of a very apolar spot (Rf = 0.8 in pentane) became

evident and the reaction seemed to be complete within 4 h. Isolation of the compound showed the loss of the carboxylic acid and hydroxy group, and the formation of an olefinic and allylic signal integrating as two and four protons, respectively. Although on

1_{H-NMR the determination of the stereochemistry of the formed alkene was not}

possible due to the overlapping signals, it has been shown previously[17–19] _{that after}

hydroxy group activation decarboxylative decomposition of the anti-periplanar arranged intermediate 12 must result in the cis alkene 13.

(11)

76

Scheme 6. Mitsunobu reaction of 10a leading to the cis alkene side product.

Although the Yamaguchi esterification was promising, we did not pursue any optimization studies for this reaction due to time constrains. Therefore, the required esters were prepared following the existing literature procedures (Scheme 7). Substitution of the mesylate group of both enantiomers of 14 by the cesium mycolates at reflux resulted in the protected glycerol esters in moderate to good yields. The reported[20]_{deprotections using TFA/water/THF mixture at 40 °C resulted in our hands}

in significant amounts of decomposition products. Fortunately, good yields could be obtained in very short reaction times by reducing the amount of TFA.

Scheme 7. The synthesis of R- and S-GroMM diastereomers.

The GMMs were obtained, in an analogous fashion to the GroM synthesis, by nucleophilic substitution of the primary tosylate 16 with the cesium mycolates at reflux,[21]_{resulting in the benzyl protected GMM in moderate yields (Scheme 8).}

Catalytic hydrogenolysis of the benzyl groups using a Pd/C and Pd(OH)2/C

combination[22]_{resulted in the final GMMs as a mixture of alpha and beta isomers after}

(12)

77

Scheme 8. The synthesis of GMMs diastereomers.

3.6 Comparison of [α]

D

and

[Φ]

D

values

With the mycolic acids in hand we were able to compare our specific rotations ([α]D)

with the ones obtained by Baird and coworkers for their synthetic material,[23]_{and our}

specific molar rotations ([Φ]D = specific rotation corrected for the concentration in

mol/100 ml) with those determined by Asselineau et al.[24]_{in their comparison study. In}

order to do so, we converted a small quantity of the free carboxylic acids to their methyl esters by esterification with TMS diazomethane in toluene/methanol at room temperature (Scheme 9).

Scheme 9. Transformation of the free MAs to their corresponding methyl esters.

The specific rotations for both MAs with S,S stereochemistry in the α-methyl methoxy segment (18a and 18d) were [α]D = 0°, which corresponds with the specific rotation

that has been reported before for the natural sample.[24]_{This result also confirms that}

indeed the stereochemistry at the cyclopropyl ring does not have an influence on the specific rotation. Consequently, also the [Φ]Dvalue was therefore 0°. Based on the work

of Asselineau et al.[24]_{this is the correct stereochemistry at this position.}

The specific rotations ([α]D) for the diastereomers with the R,R stereochemistry in the

α-methyl methoxy segment were +7.8° for 18b and +7.9° for 18c. These values are in approximation similar to the value found by Baird and coworkers for diastereomer 18c.[23]_{Having synthetic material available in which the the α-methyl methoxy segment}

contains the R,R stereochemistry allowed us to verify whether the principle of optical superposition is allowed to be used for mycolic acids, and thereby verify the work of Asselineau et al. This principle dates back to a hypothesis of Van’t Hoff and states that stereocenters that are remotely located do not influence each other’s optical activity.[25]

(13)

78

contribution of each moiety to the overall optical rotation. However, mycolic acids are known to be able to fold in the cell wall in such a way that although these functional groups would be far apart through bonds, the relative distance in space would be very proximate.[26]_{If MAs would be able to adopt a similar fold in chloroform, this principle,}

and therefore the relative stereochemistry in the α-methyl methoxy segment would become questionable. The specific molar rotations ([Φ]D)of 18b and 18c were + 99 and

+ 100, respectively, which is comparable with the sum of the values of the specific molar rotations of the α-methyl methoxy and the β-hydroxy acid segments (approximately + 85 to 90). The fact that the latter is slightly lower may be explained by the low concentrations used to determine the specific rotation of 18b and 18c, which could result in larger error margins. Furthermore, the specific molar rotation of R,R fragment C was determined at a concentration comparable to the one used by Asselineau et al., which resulted in a value of + 47. For a very comparable structure Asselineau et al. obtained a specific molar rotation of + 48 (Chapter 1, Scheme 4). With the specific molar rotations finally being determined for the diastereomers with R,R stereochemistry in the α-methyl methoxy moiety, it is safe to state that the principle of optical superposition is allowed to be used for mycolic acids.

3.7 Biological assays

GMM and MA have shown in the past to stimulate T cells when bound to the antigen presenting molecule CD1b.[27,28]_{For GMM, the acyl chain length and type of mycolate}

does not influence its T cell stimulatory capacity, but the identity and linkage of the sugar does.[29]_{For MA, acyl chain length and type of functional groups determine the}

level of T cell activation.[30,31]_{Short chain MAs have never been demonstrated to}

stimulate T cells. Among long chain MA-specific T cells, some clones are more readily activated by a certain type of MA (D, keto, or methoxy), and other clones by others, but only natural isolates were tested. The exact stereochemistry of these natural isolates is unknown. In addition to T cell activation assays, GMM and MA can also be loaded into CD1b in vitro. Tetramerized, fluorescently labeled CD1b-lipid complexes (“tetramers”) can be used to study GMM[32]_{and MA}[31]_{specific T cells at single cell level, in}

combination with fluorescently labeled antibodies, in flow cytometry. The tetramer fluorescence level is an indication of the strength of the interaction between T cell receptor and antigenic target. With the synthesis of four MA diastereomers that differ only in the configuration of their cyclopropyl and methoxy groups, but not in their overall length, we can now test the influence of these structures. The four synthetic diastereomers of MA and their respective GMMs were tested for biological activation of T cells using previously validated T cell lines[28,31]_{and in flow cytometric validation}

experiments using lipid loaded CD1b-tetramers.

All four synthetic GMMs activated the LDN5 T cell line in a dose-dependent manner (Figure 3a). The level of activation by the four different synthetic diastereomers or by natural long GMM (M. phlei) or short chain GMM (R. equi) was comparable.

(14)

79

All four synthetic MAs stimulated the T cell clone 11 (Figure 1b). The two synthetic MAs with S,S configuration at the α-methyl methoxy segment were better agonists than the those with a R,R configuration, and better than natural M. tuberculosis-derived MA. No significant difference was found for the stereochemistry at the cyclopropyl group. All synthetic GMMs, when loaded into CD1b tetramers, stained LDN5 well (Figure 3c). A minor but noticeable difference in the Mean Fluorescence Intensity (MFI) is seen depending on the stereochemistry of the α-methyl methoxy unit. The MFI for the GMMs with S,S configuration is twice as high as in those with a R,R configuration. No noticeable difference is seen between the tetramers loaded with the GMMs that differ in the stereochemistry of their cyclopropyl group.

For the MAs, several differences can be noted between the diastereomers (Figure 3d). First, a major difference among the compounds is that tetramers loaded with MAs with the α-methyl methoxy unit with S,S configuration give much higher MFIs than their R,R counterparts. Secondly, MAs with R,S configuration at the cyclopropyl (10c and 10d) cause a moderately increased MFI compared to their S,R counterparts (10a and 10b).

(15)

80

3.8 Conclusion

In conclusion, this chapter describes the completion of the mycolic acid total synthesis. From the individual fragments, a total of eight additional steps were necessary to finish the synthesis in 50% yield. Overall, the total synthesis of these methoxymycolic acids was achieved in 17 steps in the longest linear sequence, with a 12% yield (the average of the 4 isomers). Because of this high yielding synthesis, we were able to obtain gram quantities for all diastereomers, except for 10c of which slightly less than 1 gram could be prepared. Derivatization of the free mycolic acids towards GroMMs and GMMs was achieved by using existing methodology and resulted in a set of 12 additional mycolates, of which a minimal amount of 30 mg for each compound was obtained. All synthetic MAs and GMMs were bioactive. The comparable T cell activation by all tested GMMs suggests that the stereochemistry at the chiral centers in the mycolate tail has a minimal effect on the ability of GMM to activate LDN5 T cells. These results are consistent with previous findings that the head group of GMM is the predominant factor leading to T cell activation and not the mycolate chain.[29,33]

Differences in staining using CD1b tetramers loaded with GMMs that differ in the stereochemistry of their methoxy group were detected, and most likely reflect small differences in in vitro loading efficiencies, which are not translated into differences in outcome of T cell stimulation assays.

Unlike the GMMs, the MA diastereomers do show considerable differences in T cell activation. For MAs, the S,S configuration of the α-methyl methoxy group gives rise to a more potent antigen for clone 11, both in tetramer staining and biological assays. A smaller effect but in the same direction is seen for the R,S stereochemistry (as in 10c and 10d) of the cyclopropyl ring. The importance of the stereochemistry of the mycolate tail is a not unexpected, considering that short chain variants of MA cannot stimulate T cells, and different T cell clones exhibit preference for certain functional groups.[30][31]

Availability of these synthetic mycolates is an invaluable resource for in vitro, ex vivo, and in vivo studies of the role of GMM and MA-specific T cells in tuberculosis. Actual contamination or the possibility of undetectable contamination present in purified natural material could greatly hamper biological experimentations and interpretation of results. In addition, knowing the exact chemical nature of the synthetic compounds, as opposed to existing uncertainties in the structures of natural mycolates, is a big advantage in biological experiments. Also, the synthetic variants could be used as standards in the elucidation of the structure of natural mycolates and help the interpretation of crystallographic data of CD1b-mycolates. Last, synthetic molecules provide a first step in translating these results into clinically useful diagnostic tests or using the immunogens as vaccines, where use of molecules of defined structures are needed to pass certain regulatory hurdles.

(16)

81

3.9 Future prospects

With these molecules in hand, we hope that biology will enable us to better understand the role of mycolic acids in virulence and immunology. The synthesis of the various diastereomers will also allow us to embark on a formidable challenge that still remains; the determination of the absolute stereochemistry of the cyclopropyl unit in mycolic acids (see Chapter 6). With all four diastereomers in hand (the stereochemistry of the hydroxy acid unit is known), comparison to the natural compound is now a possibility. Because of the long distance between the functional groups, NMR is not of great help as the spectra of the synthetic compounds turned out to be superimposable. Therefore, other tools such as HPLC will be necessary to distinguish between the diastereomers. This is complicated by the high lipophilicity of the compounds that complicates finding conditions in which the diastereomers dissolve and can be separated. Furthermore, the natural sample is isolated as a mixture of homologues. These natural methoxy MAs possess different chain lengths, and separation of the compounds by prep-HPLC will be necessary, in order to allow for direct comparison with the synthetic material. Although by no means a proof, the mycolic acid (MA) and glucose monomycolate (GMM) tetramer binding studies allow us to anticipate on the absolute stereochemistry of the cyclopropyl unit. For both the MA and GMM holds that the S,S-stereochemistry of the α-methyl methoxy unit leads to considerably stronger binding. This is the natural stereochemistry as determined by Asselineau et al. and us. In the same vein, one can interpret these results as an indication that the cis cyclopropyl unit in the natural material should therefore be present in R,S stereochemistry (as in 10c and 10d).

The endgame synthesis of mycolic acid proceeded generally well, but some steps in this synthesis will need improvement in any future synthesis. In particular, the Julia-Kocienski coupling resulted in irreproducible yields, and up to this point it remains still unclear what actually caused this problem. As a result of the extensive screening, we are tempted to conclude that the base is quenched by an acidic NMR silent impurity. Repeating the Julia-Kocienski coupling with sulfone that resulted earlier in good yields resulted eventually in low conversions as well, and based on this observation it is reasonable to exclude the sulfone as the problem. We showed that by using more base, better conversion could be achieved. However, the safest option is always to have at least an equivalent amount of sulfone present so that deprotonations elsewhere in the molecule, such as alpha to the ester, are prevented. In addition, the coupling reactions were performed at -30 °C. Because of solubility issues a lower temperature was not feasible, and it has been shown that the phenyltetrazole-derived sulfone anion has a limited stability at higher temperatures and is prone to self-condensation.[9]

(17)

82

3. 10 Experimental section

General remarks

All moisture sensitive reactions were performed using flame-dried glassware under nitrogen atmosphere using standard Schlenk techniques and dry solvents. Reaction temperatures below 0 °C refer to internal temperatures, while reaction temperatures higher than rt refer to heating bath temperatures. Dry solvents were taken from a MBraun solvent purification system (SPS-800). All other reagents were purchased from Sigma-Aldrich, Acros, TCI Europe, Strem chemicals or Fluorochem and used without further purification unless noted otherwise.

TLC analysis was performed in Merck silica gel 60/Kieselguhr F245, 0.25 mm. Compounds were visualized using either a KMnO4 stain (K2CO3 (40g) KMnO4 (6g),

H2O (600 ml) and 10% NaOH (5ml)), anisaldehyde stain (EtOH (135 ml) H2SO4 (5 ml)

AcOH (1.5 ml) p-anisaldehyde (3.7 ml)), PMA stain (phosphomolybdic acid (10 g) in ethanol (100 ml)) or elemental iodine.

Flash chromatography was performed using SiliCycle silica gel type SiliaFlash P60 (230-400 mesh) as obtained from Screening Devices.

1_{H- and}13_{C-NMR spectra were recorded on a Agilent MR400 (400 and 100 MHz,}

respectively) or a Bruker Avance NEO 600 (600 and 150 MHz, respectively). CDCl3

was used as solvent unless stated otherwise. Chemical shift values are reported in ppm with the solvent resonance as the internal standard (CDCl3: δ7.26 for 1H, δ77.16 for 13_{C). Data are reported as follows: chemical shifts, multiplicity (s = singlet, d = doublet,}

dd = double doublet, ddd = double double doublet, dt = double triplte, td = triple doublet, t = triplet, q = quartet, p = pentet, b = broad, m = multiplet), coupling constants J (Hz), and integration.

High resolution mass spectra (HRMS) were recorded on a Thermo Scientific LTQ Orbitrap XL.

CsHCO3 used for the formation of GMMs and GrMMs was dried overnight under high

vacuum (0.01 mbar) at 120 °C and stored in the glovebox for later use.

Solid LiHMDS was obtained by addition of HMDS (1.1 equiv.) to BuLi (1.6 M in hexanes, 1.0 equiv.) at -20 °C. The formed crystals were filtered under inert atmosphere and stored in the glovebox for later use.

(18)

83

Experimental procedures

Biological assays:

GMMs were tested using a T cell line named LDN5 and a CD1b transfected C1R cell line (C1R.CD1b) as an antigen presenting cell. 100,000 LDN5 T cells were co-incubated with 50,000 C1R.CD1b cells in T cell medium[34]_{containing GMM in}

concentrations between 2.0 - 0.016 μg/mL or medium without GMM. The following day the supernatant was examined for the presence of the cytokine IL-2. MAs were tested in an IFNɣ ELISPOT assay using a monoclonal cell line named clone 11, derived from human blood.[31]_{800 Clone 11 T cells were co-incubated with 50,000 C1R.CD1b}

cells in T cell medium containing MA in concentrations of 5.0, 0.08, and 0.0013 μg/mL. All conditions were tested in biological duplicates. Organic solvent was evaporated in a stream of N2 gas and lipids were sonicated into T cell medium at room temperature.

Along with the synthetic mycolates, natural long and short chain GMM isolated from Mycobacterium phlei and Rhodococcus equi respectively (isolated in house as described[29]_{), purchased methoxy MA from Avanti Lipids (#79128), and natural long}

chain MA isolated from Mycobacterium tuberculosis (Sigma) were included as positive controls in these functional assays.

For tetramer studies, GMMs and MAs were loaded into monomeric, biotinylated CD1b protein (NIH tetramer facility). Organic solvent was evaporated in a stream of N2 gas

and lipids were sonicated in tetramer loading buffer at 37 qC for 2 hours. GMMs were loaded in a citrate buffer of pH 7.4 with 0.5 % 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) detergent (Sigma). MAs were loaded in a citrate buffer of pH 4.5 with 0.6% CHAPS detergent. CD1b protein was added and incubated overnight at 37 qC. After loading MAs, the pH was adjusted to 7.4 using 1 M tris(hydroxymethyl)aminomethane pH 8.5. Lipid-loaded CD1b monomers were then tetramerized with streptavidin conjugated to a phycoerythrin (PE) (Molecular probes). GMM or MA tetramers were used to stain LDN5 or Clone 11 cells, respectively, and were analyzed in a 5-laser BD LSRFortessa flow cytometer.

Compound 3a:

Hydroboration:

Fragment C (2.82 g, 5.90 mmol, 1.2 equiv.) was dissolved in anhydrous THF (9 ml) and solid 9-BBN dimer (780 mg, 3.19 mmol, 0.65 equiv.) was added in one portion. The

(19)

84

resulting solution was stirred at rt for 3.5 h, after which 1_{H-NMR showed complete}

disappearance of the olefinic signals. Cross-coupling:

To the hydroboration solution was added Pd(PCy3)2 (225 mg, 0.34 mmol 6.9 mol%),

and K3PO4·H2O (1.69g, 7.37 mmol, 1.5 equiv.) and stirring was continued for 15 min.

Then, R,S-fragment B (1.50 g, 4.91 mmol, 1.0 equiv.) was added, and THF (3 ml) was used to rinse the syringe. The reaction mixture was vigorously stirred at rt overnight.

1_{H-NMR indicated full consumption of the bromide and the reaction mixture was coated}

on celite and purified by flash column chromatography using a 1-4% ether in pentane gradient, yielding a dark black oil. The obtained (impure) product was used as such in the next step.

Note: The bromide and the product coelute on TLC, and therefore 1_{H-NMR is essential}

to determine the conversion. Although TLC showed easy separation of the product, upon purification a less polar side-product coeluted. All fractions containing the product were combined and used in the next step.

LiAlH4 reduction:

To a chilled solution at 0 °C of the pivaloyl ester in THF (25 ml) was added dropwise a solution of LiAlH4 (1M in THF, 4.4 ml, 4.4 mmol, 1.1 equiv.). After 15 min TLC (5%

ether in pentane) showed complete consumption of the ester. The reaction mixture was diluted with ether (30 ml), and water (0.18 ml), NaOHaq (15%, 0.18 ml) and water (0.50

ml) were carefully added. The reaction mixture was then stirred for 15 min at rt, followed by the addition of enough MgSO4 in order to remove the excess water. After

stirring for 15 min, the suspension was filtered and the filtrate was concentrated in vacuo, yielding the crude as a brown oil. Flash column chromatography using a 10-15% EtOAc in pentane gradient yielded the product (2.383 g, 3.84 mmol, 78% over 2 steps) as a yellow oil. 1_{H-NMR (400 MHz, CDCl} 3) δ 3.65 (dd, J = 11.3, 7.1 Hz, 1H), 3.58 (dd, J = 11.3, 8.1 Hz, 1H), 3.34 (s, 3H), 3.00 – 2.91 (m, 1H), 1.67 – 1.58 (m, 1H), 1.57 – 1.03 (m, 67H), 0.94 – 0.79 (m, 8H), 0.70 (td, J = 8.3, 4.4 Hz, 1H), -0.04 (q, J = 5.3 Hz, 1H). 13_C-NMR (101 MHz, CDCl3) δ 85.56, 63.28, 57.76, 35.46, 32.49, 32.06, 30.60, 30.30, 30.11, 30.07, 29.84, 29.82, 29.80, 29.71, 29.50, 28.70, 27.70, 26.27, 22.81, 18.24, 16.26, 14.98, 14.22, 9.59. HRMS (ESI) Calcd. for C42H84NaO2 ([M + Na]+): 643.636, found:

(20)

85

Compound 3b:

Compound 3b was obtained in a 73% isolated yield by subjecting R,R-Fragment C (2.82 g, 5.90 mmol, 1.2 equiv.) and R,S-Fragment B (1.50 g, 4.91 mmol, 1.0 equiv.) to the aforementioned procedures. 1_{H-NMR (400 MHz, CDCl} 3) δ 3.65 (dd, J = 11.3, 7.1 Hz, 1H), 3.58 (dd, J = 11.3, 8.0 Hz, 1H), 3.34 (s, 3H), 3.01 – 2.90 (m, 1H), 1.68 – 1.57 (m, 1H), 1.57 – 1.01 (m, 67H), 0.95 – 0.78 (m, 8H), 0.70 (td, J = 8.3, 4.5 Hz, 1H), -0.04 (q, J = 5.2 Hz, 1H). 13_C-NMR (101 MHz, CDCl3) δ 85.59, 63.43, 57.82, 35.48, 32.51, 32.08, 30.63, 30.33, 30.13, 30.09, 29.85, 29.84, 29.83, 29.81, 29.73, 29.52, 28.72, 27.72, 26.31, 22.84, 18.30, 16.30, 15.02, 14.25, 9.62. HRMS (ESI) Calcd. for C42H84NaO2 ([M + Na]+): 643.636,

found: 643.634.

Compound 3c:

Compound 3c was obtained in a 77% isolated yield by subjecting R,R-Fragment C (2.59 g, 5.41 mmol, 1.2 equiv.) and S,R-Fragment B (1.42 g, 4.65 mmol, 1.0 equiv.) to the aforementioned procedures. 1_{H-NMR (400 MHz, CDCl} 3) δ 3.69 – 3.61 (m, 1H), 3.61 – 3.54 (m, 1H), 3.34 (s, 3H), 3.00 – 2.91 (m, 1H), 1.68 – 1.56 (m, 1H), 1.50 – 1.01 (m, 67H), 0.94 – 0.79 (m, 8H), 0.70 (td, J = 8.3, 4.5 Hz, 1H), -0.04 (q, J = 5.3 Hz, 1H). 13_{C-NMR (101 MHz, CDCl} 3) δ 85.58, 63.36, 57.79, 35.48, 32.51, 32.07, 30.63, 30.32, 30.12, 30.08, 29.85, 29.80, 29.72, 29.51, 28.71, 27.71, 26.29, 22.82, 18.28, 16.29, 15.01, 14.23, 9.61. HRMS (ESI) Calcd. for C42H84NaO2 ([M + Na]+): 643.636, found: 643.634.

(21)

86

Compound 3d:

Compound 3d was obtained in a 75% isolated yield by subjecting S,S-Fragment C (2.70 g, 5.65 mmol, 1.2 equiv.) and S,R-Fragment B (1.44 g, 4.71 mmol, 1.0 equiv.) to the aforementioned procedures. 1_{H-NMR (400 MHz, CDCl} 3) δ 3.70 – 3.62 (m, 1H), 3.61 – 3.54 (m, 1H), 3.34 (s, 3H), 3.02 – 2.89 (m, 1H), 1.69 – 1.57 (m, 1H), 1.51 – 1.00 (m, 67H), 0.94 – 0.78 (m, 8H), 0.70 (td, J = 8.3, 4.5 Hz, 1H), -0.04 (q, J = 5.2 Hz, 1H). 13_{C-NMR (101 MHz, CDCl} 3) δ 85.58, 63.40, 57.81, 35.47, 32.51, 32.08, 30.62, 30.32, 30.13, 30.09, 29.85, 29.83, 29.81, 29.73, 29.52, 28.71, 27.72, 26.30, 22.83, 18.29, 16.30, 15.02, 14.25, 9.61.

HRMS (ESI) Calcd. for C42H84NaO2 ([M + Na]+): 643.636, found: 643.635.

Compound 4a:

A solution of alcohol 3a (2.35 g, 3.79 mmol, 1.0 equiv.) in THF (10 ml) was cooled to 0 °C, followed by the addition of PPh3 (1.49 g, 5.69 mmol, 1.5 equiv.) and

1-phenyl-1H-tetrazole-5-thiol (878 mg, 4.93 mmol, 1.3 equiv.). Then, DIAD (970 μl, 4.93 mmol, 1.3 equiv.) was added dropwise and the resulting orange mixture was stirred for 45 min at rt. After TLC (10% EtOAc in pentane) indicated full conversion, the reaction mixture was poured in water (50 ml) and extracted with ether (3× 50 ml). The combined organic layer was washed with brine (1× 50 ml), dried over MgSO4 and evaporated in vacuo.

The crude residue was purified by flash column chromatography using a 5-10% ether in pentane gradient, yielding the product (2.86 g, 3.66 mmol, 96%) as a colorless oil that solidified upon standing.

1_{H-NMR (400 MHz, CDCl} 3) δ 7.67 – 7.48 (m, 5H), 3.48 (d, J = 7.9 Hz, 2H), 3.34 (s, 3H), 2.99 – 2.91 (m, 1H), 1.68 – 1.58 (m, 1H), 1.53 – 1.01 (m, 66H), 1.00 – 0.78 (m, 9H), 0.07 (q, J = 5.3 Hz, 1H). 13_{C-NMR (101 MHz, CDCl} 3) δ 154.56, 133.91, 129.97, 129.76, 123.76, 85.43, 57.71, 35.39, 35.06, 32.43, 32.00, 30.54, 30.09, 30.05, 30.02, 29.78, 29.76, 29.74, 29.61, 29.44, 28.53, 27.65, 26.22, 22.76, 18.02, 14.95, 14.71, 14.18, 12.51. HRMS (ESI) Calcd. for C49H89N4OS ([M + H]+): 781.675, found:

(22)

87

Compound 4b:

Compound 4b was obtained in a 96% isolated yield by subjecting alcohol 3b (2.21 g, 3.56 mmol) to the aforementioned procedure.

1_{H-NMR (400 MHz, CDCl} 3) δ 7.65 – 7.49 (m, 5H), 3.48 (d, J = 7.9 Hz, 2H), 3.34 (s, 3H), 2.99 – 2.91 (m, 1H), 1.67 – 1.58 (m, 1H), 1.52 – 1.03 (m, 66H), 1.00 – 0.78 (m, 9H), 0.07 (q, J = 5.3 Hz, 1H). 13_{C-NMR (101 MHz, CDCl} 3) δ 154.65, 133.94, 130.05, 129.82, 123.85, 85.50, 57.78, 35.43, 35.12, 32.47, 32.03, 30.58, 30.13, 30.09, 30.05, 29.81, 29.77, 29.65, 29.47, 28.57, 27.68, 26.26, 22.80, 18.07, 14.99, 14.74, 14.22, 12.56. HRMS (ESI) Calcd. for C49H89N4OS ([M + H]+): 781.675, found: 781.674.

Compound 4c:

Compound 4c was obtained in a 93% isolated yield by subjecting alcohol 3c (2.20 g, 3.54 mmol) to the aforementioned procedure.

1_{H-NMR (400 MHz, CDCl} 3) δ 7.64 – 7.50 (m, 5H), 3.48 (d, J = 7.9 Hz, 2H), 3.33 (s, 3H), 3.00 – 2.91 (m, 1H), 1.67 – 1.58 (m, 1H), 1.54 – 1.01 (m, 66H), 0.98 – 0.78 (m, 9H), 0.07 (q, J = 5.3 Hz, 1H). 13_{C-NMR (101 MHz, CDCl} 3) δ 154.68, 133.96, 130.07, 129.84, 123.88, 85.53, 57.80, 35.46, 35.14, 32.49, 32.04, 30.61, 30.14, 30.10, 30.07, 29.82, 29.78, 29.66, 29.48, 28.58, 27.69, 26.28, 22.81, 18.08, 15.01, 14.77, 14.23, 12.58. HRMS (ESI) Calcd. for C49H89N4OS ([M + H]+): 781.675, found: 781.674.

Compound 4d:

Compound 4d was obtained in a 96% isolated yield by subjecting alcohol 3d (2.18 g, 3.51 mmol) to the aforementioned procedure.

(23)

88 1_{H-NMR (400 MHz, CDCl} 3) δ 7.70 – 7.45 (m, 5H), 3.48 (d, J = 7.9 Hz, 2H), 3.34 (s, 3H), 3.00 – 2.88 (m, 1H), 1.68 – 1.56 (m, 1H), 1.52 – 1.02 (m, 66H), 0.99 – 0.78 (m, 9H), 0.07 (q, J = 5.3 Hz, 1H). 13_{C-NMR (101 MHz, CDCl} 3) δ 154.49, 133.88, 129.92, 129.72, 123.71, 85.38, 57.67, 35.36, 35.02, 32.41, 31.98, 30.51, 30.07, 30.03, 29.99, 29.76, 29.72, 29.59, 29.42, 28.50, 27.62, 26.19, 22.73, 17.99, 14.92, 14.69, 14.16, 12.48. HRMS (ESI) Calcd. for C49H89N4OS ([M + H]+): 781.675, found: 781.674.

Compound 5a:

To a solution of 4a (2.70 g, 3.57 mmol, 1.0 equiv.) in EtOH (40 ml) and THF (30 ml) was added ammonium molybdate (1.28 g, 1.04 mmol, 30 mol%) followed by H2O2

(30% aqueous solution, 8.2 ml, 86.39 mmol, 25 equiv). The resulting mixture was stirred overnight at rt. 1_{H-NMR showed complete consumption of the sulfide, resulting}

in a 5:1 sulfone/sulfoxide mixture. More catalyst (650 mg, 0.53 mmol, 15 mol%) and peroxide (4.5 ml, 47.51 mmol, 14 equiv.) were added and the reaction mixture was allowed to stir for another day. Then, the volatiles were evaporated and the residue was diluted with water (100 ml) and extracted with ether (3× 75 ml). The combined organic layer was washed with a saturated aqueous solution of Na2SO4 (1× 100 ml), water (1×

100 ml) brine (1× 100 ml), dried over MgSO4 and evaporated in vacuo. The crude

residue was purified by flash column chromatograph using 10% ether in pentane, yielding the product (2.77 g, 3.41 mmol, 99%) as a colorless oil that solidified upon standing. 1_{H-NMR (400 MHz, CDCl} 3) δ 7.72 – 7.67 (m, 2H), 7.64 – 7.57 (m, 3H), 3.96 (dd, J = 14.7, 5.6 Hz, 1H), 3.56 (dd, J = 14.7, 9.3 Hz, 1H), 3.34 (s, 3H), 3.01 – 2.90 (m, 1H), 1.67 – 1.57 (m, 1H), 1.50 – 0.93 (m, 68H), 0.92 – 0.81 (m, 7H), 0.24 (q, J = 5.5 Hz, 1H). 13_{C-NMR (101 MHz, CDCl} 3) δ 153.84, 133.24, 131.51, 129.75, 125.29, 85.54, 57.82, 57.21, 35.45, 32.49, 32.06, 30.61, 30.11, 30.08, 29.83, 29.80, 29.77, 29.74, 29.57, 29.50, 29.20, 27.70, 26.29, 22.82, 16.00, 15.02, 14.25, 11.48, 8.12. HRMS (ESI) Calcd. for C49H92N5O3S ([M + NH4]+): 830.692, found: 830.690.

(24)

89

Compound 5b was obtained in a 78% isolated yield by subjecting sulfide 4b (2.66 g, 3.41 mmol) to the aforementioned procedure. After complete elution of 5b from the column, eluting with 25% ether in pentane resulted in the recovery of sulfoxide (575 mg, 0.72 mmol, 21%). The sulfoxide was used for conditions screening, but by resubmission according to the aforementioned procedure extra sulfone could have been acquired. 1_{H-NMR (400 MHz, CDCl} 3) δ 7.73 – 7.66 (m, 2H), 7.67 – 7.56 (m, 3H), 3.96 (dd, J = 14.7, 5.6 Hz, 1H), 3.56 (dd, J = 14.7, 9.3 Hz, 1H), 3.34 (s, 3H), 3.00 – 2.91 (m, 1H), 1.68 – 1.59 (m, 1H), 1.50 – 0.93 (m, 68H), 0.93 – 0.79 (m, 7H), 0.24 (q, J = 5.6 Hz, 1H). 13_{C-NMR (101 MHz, CDCl} 3) δ 153.69, 133.17, 131.21, 129.51, 125.16, 85.31, 57.59, 56.97, 35.33, 32.39, 31.95, 30.48, 30.01, 29.97, 29.75, 29.71, 29.68, 29.66, 29.64, 29.46, 29.42, 29.40, 29.03, 27.59, 26.15, 22.70, 15.85, 14.88, 14.12, 11.28, 7.99.

HRMS (ESI) Calcd. for C49H92N5O3S ([M + NH4]+): 830.692, found: 830.690

Compound 5c:

To a solution of 4c (2.56 g, 3.28 mmol, 1.0 equiv.) in n-BuOH (35 ml) and THF (30 ml) was added ammonium molybdate (1.24 g, 1.00 mmol, 31 mol%) followed by H2O2

(30% aqueous solution, 5.0 ml, 52.8 mmol, 16 equiv.). The resulting mixture was stirred at rt overnight. 1_{H-NMR showed complete consumption of the sulfide, resulting in a 5:1}

sulfone/sulfoxide mixture. The volatiles were evaporated and the residue was diluted with water (100 ml) and extracted with ether (3× 75 ml). The combined organic layer was washed with a saturated aqueous solution of Na2SO4 (1× 100 ml), water (1× 100

ml), brine (1× 100 ml), dried over MgSO4 and evaporated in vacuo. The crude residue

was purified by flash column chromatograph using 10% ether in pentane to elute the sulfone (2.16 g) and 30% ether in pentane to elute the sulfoxide (0.47 g). The sulfoxide was resubmitted under the same conditions, but the reaction was stirred for 3 d. The obtained sulfone was combined with the previously obtained batch, resulting in the product (2.60 g, 3.20 mmol, 98% yield) as a colorless oil that solidified upon standing). Note: It is not essential for the reaction to be homogeneous, but no visible micelles should be present in order for the reaction to proceed to completion. A small amount of the formed precipitate was isolated, but this material did not dissolve in ether. Therefore, it was concluded that the formed precipitate is not organic material and must therefore originate from the molybdate catalyst. This procedure (in which n-BuOH is used as solvent) should be used in future synthesis.

(25)

90 1_{H-NMR (400 MHz, CDCl} 3) δ 7.72 – 7.66 (m, 2H), 7.65 – 7.56 (m, 3H), 3.96 (dd, J = 14.7, 5.6 Hz, 1H), 3.56 (dd, J = 14.7, 9.3 Hz, 1H), 3.34 (s, 3H), 3.00 – 2.90 (m, 1H), 1.67 – 1.57 (m, 1H), 1.51 – 0.94 (m, 68H), 0.93 – 0.80 (m, 7H), 0.24 (q, J = 5.5 Hz, 1H). 13_{C-NMR (101 MHz, CDCl} 3) δ 153.86, 133.24, 131.51, 129.75, 125.29, 85.55, 57.82, 57.22, 35.47, 32.50, 32.06, 30.62, 30.11, 30.08, 29.83, 29.80, 29.77, 29.74, 29.57, 29.49, 29.21, 27.70, 26.29, 22.82, 16.01, 15.02, 14.24, 11.48, 8.13. HRMS (ESI) Calcd. for C49H92N5O3S ([M + NH4]+): 830.692, found: 830.690.

Compound 5d:

Compound 5d was obtained in 97% yield, by using the same procedure as for 5a, but with two workup/resubmission sequences.

1_{H-NMR (400 MHz, CDCl} 3) δ 7.73 – 7.65 (m, 2H), 7.65 – 7.55 (m, 3H), 3.96 (dd, J = 14.7, 5.6 Hz, 1H), 3.56 (dd, J = 14.6, 9.3 Hz, 1H), 3.34 (s, 3H), 3.00 – 2.90 (m, 1H), 1.68 – 1.57 (m, 1H), 1.51 – 0.93 (m, 68H), 0.96 – 0.79 (m, 7H), 0.24 (q, J = 5.6 Hz, 1H). 13_{C-NMR (101 MHz, CDCl} 3) δ 153.84, 133.23, 131.50, 129.75, 125.29, 85.54, 77.48, 77.16, 76.84, 57.81, 57.20, 35.45, 32.49, 32.05, 30.61, 30.11, 30.07, 29.83, 29.79, 29.76, 29.74, 29.57, 29.49, 29.20, 27.70, 26.29, 22.82, 16.00, 15.02, 14.24, 11.47, 8.12. HRMS (ESI) Calcd. for C49H92N5O3S ([M + NH4]+): 830.692, found:

830.690.

Compound 7a:

A solution of sulfone (1.57 g, 1.93 mmol, 1.20 equiv.) in anhydrous THF (40 ml) was cooled to -40 °C. LiHMDS (crystallized, 0.31 g, 1.85 mmol, 1.15 equiv.) in THF (7 ml) was added while keeping the temperature between -40 °C and -30 °C. The reaction mixture turned bright yellow and after stirring for 15 min, fragment A (1.91 g, 1.61 mmol, 1.0 equiv.) in THF (8 ml) was added over 5 min and the reaction mixture was stirred for an additional 45 min. Then, the reaction was quenched by the addition of HClaq (1M, 50 ml) and water (100 ml) and the mixture was allowed to reach rt. The

mixture was extracted with ether (3× 50 ml), and the combined organic layer was washed with brine (1× 50 ml), dried over MgSO4 and evaporated in vacuo. The crude

(26)

91

residue was purified by flash column chromatograph using 5% ether in pentane, yielding the product (2.57 g, 1.45 mmol, 90%, E/Z = 7:3) as a colorless oil.

1_{H-NMR (400 MHz, CDCl} 3) δ 7.33 (d, J = 6.6 Hz, 2H), 7.28 – 7.20 (m, 3H), 7.17 (t, J = 7.4 Hz, 1H), 7.08 (t, J = 7.5 Hz, 2H), 6.87 – 6.76 (m, 4H), 5.70 (d, J = 6.1 Hz, 1H), 5.51 (dt, J = 14.3, 6.8 Hz, 0.7H), 5.40 (dd, J = 11.0, 7.6 Hz, 0.3H), 5.17 (dd, J = 15.2, 8.5 Hz, 0.7H), 5.03 (t, J = 10.1 Hz, 0.3H), 4.81 (d, J = 16.2 Hz, 1H), 4.42 (d, J = 16.2 Hz, 1H), 4.14 (p, J = 6.6 Hz, 1H), 3.86 (br s, 1H), 3.34 (s, 3H), 2.99 – 2.92 (m, 1H), 2.48 (dt, J = 10.4, 5.5 Hz, 1H), 2.39 (s, 6H), 2.28 (s, 3H), 2.19 – 2.10 (q, 0.6H), 1.99 (q, J = 6.9 Hz, 1.4H), 1.68 – 0.74 (m, 165H), 0.11 (q, J = 5.3 Hz, 1H), 0.04 (s, 3H), 0.03 (s, 3H). 13_{C-NMR (101 MHz, CDCl} 3) δ 172.76, 142.40, 140.50, 138.39, 138.34, 133.26, 132.17, 130.53, 130.29, 129.59, 129.54, 128.43, 128.34, 128.19, 127.92, 127.40, 126.93, 85.50, 77.87, 77.36, 72.83, 57.77, 56.53, 51.68, 48.28, 35.46, 33.64, 32.89, 32.51, 32.08, 30.61, 30.13, 30.09, 29.94, 29.87, 29.82, 29.77, 29.75, 29.73, 29.71, 29.66, 29.59, 29.53, 29.48, 29.33, 29.29, 27.75, 27.72, 27.69, 27.10, 26.29, 26.00, 25.02, 22.97, 22.84, 20.98, 18.57, 18.38, 18.18, 15.07, 15.02, 14.31, 14.26, 14.03, 12.44, -4.27, -4.51. HRMS (ESI) Calcd. for C116H211N2O6SSi ([M + NH4]+): 1788.58,

found: 1788.57.

Compound 7b:

Compound 7b was obtained in a 54% yield by submitting sulfone 5b (2.16, 2.66 mmol, 1.26 equiv.), fragment A (2.5 g, 2.11 mmol, 1.0 equiv.) and LiHMDS (crystallized, 0.42g, 2.49 mmol, 1.2 equiv.) to the aforementioned procedure.

A sulfone and aldehyde mixture could be recovered after the product eluted from the column by eluting with 30% ether in pentane. The crude mixture was dissolved in EtOH/THF (8:1, 9 ml) and treated with NaBH4 (100 mg) at 0 °C. After 30 min the

reaction mixture was quenched by the addition of HClaq (1M, 50 ml) and the mixture

was extracted with ether (3× 25 ml). The combined organic layer was concentrated and the sulfone (450 mg, 0.55 mmol) could be recovered after flash column chromatography using 20% ether in pentane.

1_{H-NMR (400 MHz, CDCl} 3) δ 7.33 (d, J = 6.5 Hz, 2H), 7.28 – 7.20 (m, 3H), 7.17 (t, J = 7.4 Hz, 1H), 7.08 (t, J = 7.5 Hz, 2H), 6.87 – 6.76 (m, 4H), 5.70 (d, J = 6.1 Hz, 1H), 5.51 (dt, J = 14.3, 6.8 Hz, 0.7H), 5.40 (dt, J = 10.7, 7.3 Hz, 0.3H), 5.17 (dd, J = 15.2, 8.5 Hz, 0.7H), 5.03 (t, J = 10.1 Hz, 0.3H), 4.81 (d, J = 16.2 Hz, 1H), 4.42 (d, J = 16.2 Hz, 1H), 4.14 (p, J = 6.7 Hz, 1H), 3.86 (br s, J = 6.1, 4.9 Hz, 1H), 3.34 (s, 3H), 3.02 – 2.90 (m, 1H), 2.48 (dt, J = 10.6, 5.6 Hz, 1H), 2.39 (s, 6H), 2.28 (s, 3H), 2.14 (q, J = 7.6,

(27)

92 7.1 Hz, 0.6H), 1.99 (q, J = 6.9 Hz, 1.4H), 1.70 – 0.73 (m, 165H), 0.11 (q, J = 5.3 Hz, 1H), 0.04 (s, 3H), 0.04 (s, 3H). 13_{C-NMR (101 MHz, CDCl} 3) δ 172.79, 142.44, 140.51, 138.39, 138.34, 133.26, 132.19, 130.55, 130.31, 129.60, 129.56, 128.44, 128.34, 128.21, 127.94, 127.41, 126.94, 85.53, 77.90, 77.36, 72.85, 57.80, 56.54, 51.71, 48.29, 35.47, 33.65, 32.90, 32.51, 32.08, 30.62, 30.45, 30.13, 30.10, 29.94, 29.87, 29.82, 29.78, 29.76, 29.73, 29.72, 29.67, 29.59, 29.53, 29.50, 29.49, 29.33, 29.30, 27.77, 27.73, 27.70, 27.09, 26.30, 26.01, 25.05, 22.99, 22.84, 21.00, 18.59, 18.40, 18.19, 15.07, 15.03, 14.32, 14.26, 14.04, 12.45, -4.26, -4.49. HRMS (ESI) Calcd. for C116H211N2O6SSi ([M + NH4]+): 1788.58, found: 1788.57.

Compound 7c:

Compound 7c was obtained in a 79% yield by the reaction of sulfone 5c (1.90 g, 2.33 mmol, 2.9 equiv.), fragment A (0.95 g, 0.80 mmol, 1.0 equiv.) and LiHMDS (crystallized, 0.377 g, 0.63 mmol, 2.8 equiv.) at -45 °C in an otherwise similar way to

7a. Eluting with 20% ether in pentane after the product completely eluted from the

column resulted in the recovery of sulfone (1.10 g, 1.68 mmol).

1_{H-NMR (400 MHz, CDCl} 3) δ 7.35 – 7.31 (m, 2H), 7.26 – 7.14 (m, 4H), 7.08 (t, 2H), 6.86 – 6.77 (m, 4H), 5.70 (d, J = 6.1 Hz, 1H), 5.51 (dt, J = 15.3, 6.8 Hz, 0.7H), 5.40 (dd, J = 15.2, 8.5 Hz, 0.3H), 5.17 (dd, J = 15.2, 8.5 Hz, 0.7H), 5.03 (t, J = 9.7 Hz, 0.3H), 4.80 (d, J = 16.2 Hz, 1H), 4.42 (d, J = 16.2 Hz, 1H), 4.15 (p, J = 6.8 Hz, 1H), 3.86 (br s, 1H), 3.34 (s, 3H), 2.99 – 2.92 (m, 1H), 2.51 – 2.44 (m, 1H), 2.39 (s, 6H), 2.28 (s, 3H), 2.14 (q, J = 6.8 Hz, 0.6H), 1.99 (q, J = 6.8 Hz, 1.4H), 1.66 – 0.75 (m, 165H), 0.10 (q, J = 5.4 Hz, 1H), 0.04 (s, 3H), 0.03 (s, 3H). 13_{C-NMR (101 MHz, CDCl} 3) δ 172.79, 142.45, 140.52, 138.40, 138.36, 133.29, 132.20, 130.56, 130.32, 129.61, 129.56, 128.45, 128.35, 128.22, 127.95, 127.42, 126.95, 85.55, 77.91, 72.86, 57.80, 56.56, 51.73, 48.30, 35.49, 33.66, 32.90, 32.53, 32.11, 32.09, 30.64, 30.14, 30.10, 29.95, 29.91, 29.87, 29.82, 29.78, 29.76, 29.74, 29.72, 29.67, 29.59, 29.53, 29.50, 29.34, 29.30, 27.79, 27.73, 27.71, 27.08, 26.31, 26.02, 25.08, 22.99, 22.84, 21.00, 18.59, 18.41, 18.20, 15.07, 15.03, 14.33, 14.26, 14.05, 12.46, -4.26, -4.48. HRMS (ESI) Calcd. for C116H211N2O6SSi ([M + NH4]+): 1788.58, found: 1788.57.

(28)

93

Compound 7d (81% yield) was obtained in a similar way to 7a by reacting sulfone 5d (2.09, 2.57 mmol, 1.25 equiv.), fragment A (2.5 g, 2.1 mmol, 1.0 equiv.) and LiHMDS (crystallized, 0.41 g, 2.4 mmol, 1.2 equiv.).

1_{H-NMR (400 MHz, CDCl} 3) δ 7.33 (d, J = 7.3 Hz, 2H), 7.24 (d, J = 13.5 Hz, 3H), 7.17 (t, J = 7.5 Hz, 1H), 7.08 (t, J = 7.5 Hz, 2H), 6.87 – 6.76 (m, 4H), 5.70 (d, J = 6.1 Hz, 1H), 5.51 (dt, J = 14.2, 6.8 Hz, 0.7H), 5.39 (q, J = 8.8 Hz, 0.3H), 5.17 (dd, J = 15.3, 8.5 Hz, 0.7H), 5.03 (t, J = 10.1 Hz, 0.3H), 4.81 (d, J = 16.2 Hz, 1H), 4.42 (d, J = 16.2 Hz, 1H), 4.14 (p, J = 6.7 Hz, 1H), 3.86 (br s, 1H), 3.34 (s, 3H), 2.96 (br s, 1H), 2.47 (dd, J = 9.9, 5.3 Hz, 1H), 2.39 (s, 6H), 2.28 (s, 3H), 2.14 (q, J = 7.8, 7.3 Hz, 0.6H), 1.99 (q, J = 7.0 Hz, 1.4H), 1.81 – 0.74 (m, 165H), 0.14 – 0.06 (m, 1H), 0.04 (s, 6H). 13_{C-NMR (101} MHz, CDCl3) δ 172.67, 142.28, 140.46, 138.37, 138.32, 133.28, 132.12, 130.47, 130.23, 129.53, 129.48, 128.37, 128.34, 128.14, 127.86, 127.35, 126.92, 85.42, 77.80, 77.36, 72.80, 57.69, 56.47, 51.61, 48.23, 35.44, 33.61, 32.87, 32.50, 32.07, 30.58, 30.12, 30.08, 29.86, 29.76, 29.72, 29.63, 29.57, 29.52, 29.48, 29.45, 29.31, 29.28, 27.71, 27.67, 27.13, 26.26, 25.98, 24.93, 22.93, 22.82, 20.93, 18.54, 18.34, 18.13, 15.07, 14.98, 14.28, 14.23, 12.38, -4.31, -4.55. HRMS (ESI) Calcd. for C116H211N2O6SSi ([M + NH4]+): 1788.58, found: 1788.57.

Compound 8a:

Compound 7a (3.20 g, 1.81 mmol, 1.0 equiv.) was dissolved in n-BuOH/THF (5:1 v/v, 24 ml) and hydrazine monohydrate (2.2 ml, 25 equiv.) followed by the flavine catalyst (0.34 g, 0.84 mmol, 0.46 equiv.) were added. EtOH (5 ml) was added followed by enough THF so that the liquids formed a single phase (parts of the catalyst did not dissolve). The reaction mixture was stirred at rt, and the volatiles were removed in vacuo after full conversion according to 1_{H-NMR (around 5 d). To the crude residue}

was added water (75 ml), and the aqueous phase was extracted with pentane (3× 75 ml). The combined organic layer was washed with brine (1× 50 ml), dried over MgSO4 and

concentrated in vacuo. The crude product was filtered over a plug of silica using 10% ether in pentane, yielding the product (3.14 g, 1.77 mmol, 98%) as a colorless oil.

1_{H-NMR (400 MHz, CDCl} 3) δ 7.33 (d, J = 7.2 Hz, 2H), 7.28 – 7.20 (m, 3H), 7.17 (t, J = 7.4 Hz, 1H), 7.09 (t, J = 7.5 Hz, 2H), 6.87 – 6.76 (m, 4H), 5.71 (d, J = 6.1 Hz, 1H), 4.81 (d, J = 16.2 Hz, 1H), 4.43 (d, J = 16.2 Hz, 1H), 4.15 (p, J = 6.7 Hz, 1H), 3.91 – 3.83 (m, 1H), 3.34 (s, 3H), 3.00 – 2.92 (m, 1H), 2.52 – 2.44 (m, 1H), 2.40 (s, 6H), 2.29 (s, 3H), 1.69 – 1.57 (m, 1H), 1.57 – 1.02 (m, 147H), 1.02 – 0.92 (m, 2H), 0.92 – 0.80 (m, 18H), 0.70 – 0.61 (m, 2H), 0.60 – 0.52 (m, 1H), 0.04 (s, 6H), -0.33 (q, J = 4.9 Hz,

(29)

94 1H). 13_{C-NMR (101 MHz, CDCl} 3) δ 172.66, 142.27, 140.46, 138.38, 138.32, 133.28, 132.12, 128.37, 128.35, 128.13, 127.86, 127.35, 126.93, 85.42, 77.79, 77.36, 72.80, 57.69, 56.47, 51.60, 48.23, 35.45, 33.62, 32.51, 32.08, 30.59, 30.40, 30.37, 30.13, 30.09, 29.88, 29.84, 29.81, 29.78, 29.76, 29.72, 29.68, 29.64, 29.53, 29.46, 28.86, 27.71, 27.67, 27.15, 26.26, 25.98, 24.91, 22.93, 22.83, 20.92, 18.13, 15.89, 15.08, 14.98, 14.23, 11.07, -4.31, -4.55. HRMS (ESI) Calcd. for C116H213N2O6SSi ([M +

NH4]+): 1790.59, found: 1790.59.

Compound 8b:

Compound 8b was obtained in 95% isolated yield by subjecting 7b to the aforementioned procedure for 5 days.

1_{H-NMR (400 MHz, CDCl} 3) δ 7.33 (d, J = 7.2 Hz, 2H), 7.29 – 7.13 (m, 4H), 7.08 (t, J = 7.5 Hz, 2H), 6.88 – 6.76 (m, 4H), 5.70 (d, J = 6.1 Hz, 1H), 4.81 (d, J = 16.2 Hz, 1H), 4.42 (d, J = 16.2 Hz, 1H), 4.14 (p, J = 6.7 Hz, 1H), 3.86 (br s, 1H), 3.34 (s, 3H), 3.01 – 2.88 (m, 1H), 2.52 – 2.44 (m, 1H), 2.39 (s, 6H), 2.28 (s, 3H), 1.69 – 1.57 (m, 1H), 1.57 – 1.03 (m, 147H), 1.03 – 0.91 (m, 2H), 0.91 – 0.81 (m, 18H), 0.69 – 0.60 (m, 2H), 0.60 – 0.52 (m, 1H), 0.04 (s, 3H), 0.03 (s, 3H), -0.33 (q, J = 5.0 Hz, 1H). 13_{C-NMR (101} MHz, CDCl3) δ 172.69, 142.30, 140.47, 138.38, 138.33, 133.28, 132.14, 128.39, 128.35, 128.15, 127.88, 127.36, 126.93, 85.44, 77.81, 77.36, 72.81, 57.71, 56.49, 51.63, 48.24, 35.46, 33.62, 32.51, 32.08, 30.59, 30.37, 30.13, 30.09, 29.88, 29.83, 29.79, 29.77, 29.72, 29.69, 29.64, 29.53, 29.46, 28.87, 27.72, 27.69, 27.14, 26.27, 25.98, 24.94, 22.94, 22.83, 20.94, 18.14, 15.89, 15.08, 14.99, 14.24, 11.07, -4.30, -4.54.

HRMS (ESI) Calcd. for C116H213N2O6SSi ([M + NH4]+): 1790.59, found: 1790.59.

Compound 8c:

Compound 8c was obtained in 96% isolated yield by subjecting 7c to the aforementioned procedure for 7 days.

1_{H-NMR (400 MHz, CDCl}

3) δ 7.35 – 7.30 (m, 2H), 7.25 – 7.14 (m, 4H), 7.08 (t, J = 7.8

Hz, 2H), 6.85 – 6.77 (m, 4H), 5.70 (d, J = 6.2 Hz, 1H), 4.81 (d, J = 16.2 Hz, 1H), 4.42 (d, J = 16.2 Hz, 1H), 4.14 (p, J = 6.7 Hz, 1H), 3.90 – 3.80 (m, 1H), 3.34 (s, 3H), 3.00 –