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: 137PDF page: 137PDF page: 137PDF page: 137

Chapter 5:

Synthesis of a Library of Neomycin B

Amphiphiles for the Development of Novel

Antimicrobial Agents

Part of this chapter will be submitted for publication:

(3)

128

5.1 Introduction

Since the isolation of streptomycin from Streptomyces griseus, a strain of soil bacteria, by the Waksman laboratory in 1943,[1]_{aminoglycosides have played a vital role in the}

treatment of bacterial infections. Streptomycin was notably the first effective treatment against tuberculosis for which Waksman was awarded the 1952 Nobel Prize in physiology and medicine. Together with penicillin, streptomycin triggered the “antibiotic revolution”, which led to the discovery of many more natural (aminoglycoside) antibiotics by Waksman and others. Two other well-known aminoglycoside antibiotics that have played an important role in combatting pathogenic infections are neomycin[2]_{and kanamycin}[3]_{, which were isolated by Waksman and}

Umezawa, respectively. These natural antibiotics have inspired chemists to develop new semi-synthetic variants with improved activity against resistant pathogens such as amikacin[4]_{and plazomicin}[5]_{, which are derived from kanamycin and sisomicin,}

respectively.

5.1.1 Aminoglycosides: structural composition, mode of action and toxicity

Aminoglycoside antibiotics can contain a wide variety of structural functionalities but share, with a few exceptions, a common 2-deoxystreptamine (2-DOS) motif. The most important exception is streptomycin that, instead of a 2-DOS motif, contains an unusual streptamine motif of which both amines are derivatized as guanidine functionalities (Figure 1).

Figure 1. General overview of aminoglycoside antibiotics classification.

Two molecules that fulfill the minimal structural requirements to be classified as aminoglycoside antibiotics are neamine and paromamine and are members of the 4-mono-substituted class (Figure 1). Neamine[6]_{and paromamine}[7]_{are most probably}

(4)

129

also be obtained as degradation products via acidic methanolysis of neomycin and paromomycin. Although neamine has shown to possess antibacterial activities against both Gram-negative and Gram-positive pathogens, its activity compared to other aminoglycosides is significantly lower.[8]_{It has therefore not fulfilled any significant}

clinical role.

The more common and clinically relevant classes of aminoglycoside antibiotics consist of the 4,5- and 4,6-disubstituted aminoglycosides (Figure 1). Neomycin is a representative example of the 4,5-disubstituted class, and is usually isolated as a mixture of neomycin B and C which are epimeric at the 5”’ carbon. The 4,6-disubstituted class usually consists of at least two amino sugars linked in a 4,6-fashion to the 2-DOS ring. An example of a 4,6-disubstituted aminoglycoside is kanamycin, which is also isolated as a mixture of two compounds. By installing a 4-amino 2-hydroxybutanoyl (AHB) sidechain on N1 of the 2-DOS ring, the semisynthetic antibiotic amikacin was obtained.[4]_{This compound showed to relieve resistance caused}

by aminoglycoside modifying enzymes (AMEs) in several resistant strains, and is still in use for the treatment of multidrug-resistant tuberculosis.

It was seminal work by Erdos and Ullmann[9]_{in 1959 that indicated that}

aminoglycosides interfere with protein synthesis. They showed that streptomycin reduced the incorporation of radioactive amino acids in susceptible strains and that resistant strains remained unaffected. Not much later, Scott and Stanier[10]_{proposed the}

ribosome to be the primary target. After decades of research, the generally accepted mechanism is that aminoglycoside antibiotics bind to the 16S ribosomal subunit, thereby causing a conformational change in the RNA.[11]_{As a result, the number of}

errors in the translation process increases, eventually resulting in cell death. NMR studies[12]_{on paromomycin and high resolution X-ray crystallography on several other}

aminoglycoside RNA co-crystals[13]_{have further elucidated the complex binding}

patterns of these antibiotics with the specific RNA base pairs. From the X-ray studies it is evident that ring I and II, i.e. “the neamine core”, are mainly involved in interactions with the RNA (Figure 2). This binding is largely driving by stabilizing electrostatic interactions between the positively charged amino groups of the aminoglycoside with the negatively charged ribosomal RNA.

The general applicability of aminoglycoside antibiotics has been limited due to their serious side effects, and therefore most of them are nowadays reserved as last resort antibiotics only. They show a dose dependent toxicity against mainly the kidneys (nephrotoxicity) and the ears (ototoxicity), due to non-specific interactions with eukaryotic RNA.

(5)

130

Figure 2. Left: the specific structure adopted by a kanamycin molecule inside the A site of the

RNA-duplex. Right: chemical structure of kanamycin, drawn in the same conformation.

5.1.2 Amphiphilic aminoglycosides: structural features, mode of action and toxicity

The introduction of a single or multiple highly lipophilic residues, usually in the form of a C8 alkyl chain or higher (Figure 3), has been demonstrated to target Gram-negative and Gram-positive bacteria by a completely different mechanism compared to the original aminoglycosides.[14]_{Moreover, some of these amphiphilic aminoglycosides}

(AAGs) have been recently reported to possess antifungal properties. (Figure 3c).[15]_By

varying the functional group used for attaching the lipophilic chain, the parent aminoglycoside and the chain length, AAG could be modulated from targeting bacteria[16]_{and fungi, to targeting fungi only.}[15][17]_{Although the amount of research}

performed is still limited, shorter alkyl chains seem to favor antifungal properties, while longer alkyl residues seem to be in favor of antibacterial properties.[18]_This

demonstrates that minor changes in molecular structure can have a profound change in activity, and highlights the need for thorough structure-activity relationship (SAR) studies that will stimulate the rational design of novel AAGs.

Figure 3. Examples of amphiphilic aminoglycosides.

With their amphiphilic nature, AAGs show great resemblance to many components that constitute the cell membrane in Gram-negative and Gram-positive bacteria, in the sense that they both contain a long linear alkyl chain and an ionized head group. These physiological similarities usually form the basis for antimicrobials designed to target the cell membrane. For example, the cationic lipopeptide colistin targets lipopolysaccharides (LPS) (Figure 4) in Gram negative bacteria by binding to the negatively charged lipid A core of LPS and displaces the calcium and magnesium counterions that keep the LPS together.[19]_{This binding is the result of favorable}

(6)

131

hydrophobic and electrostatic interactions between the LPS and colistin. The displacement of counterions by colistin results in disruption of the cell membrane and an increase in permeability thereof. The latter results in increased passive diffusion rates which eventually leads to cell death. In a similar way to colistin, several AAGs have been shown to have high affinity for LPS.[20,21]_{Since AAGs also show activity against}

Gram-positive bacteria and fungi, the demonstrated antimicrobial activity is most probably a complex interplay of several targets within the cell membrane.

Figure 4. Structure of lipid A, the target of colistin.

This non-specific interaction should limit the emergence of resistance of bacteria against AAGs and gives an advantage for an AAG in the fight against resistant bacteria and fungi.[14]_{Although one would assume this non-specificity of AAGs would lead to}

unwanted interactions with the host cell membrane, the selectivity of AAGs for bacterial and fungal cell membranes over mammalian cells is often quite good. This selectivity is determined by measuring the hemolytic activity of the corresponding AAG,[20–22]_{which is the rupturing (lysis) of red blood cells and the release of their}

content. The concentrations at which significant hemolysis occurs is often one or several orders of magnitude higher than the MIC (minimum inhibitory concentration) value for antimicrobial activity.

For example, the antifungal K20 (Figure 3c) lysed less than 40% of sheep red blood cells at a concentration of 500 μg/ml, which is more than 50-fold higher than the MIC for C. neoformans H99.[15]_{On the other hand, the antibacterial}

4,4’,6-Tri-O-heptylnebramine (figure 3.b) was significantly more hemolytic. This amphiphilic nebramine derivative lysed 50% of rat red blood cells at a concentration of 16 μg/ml and had a MIC range of 1-4 μg/ml for Gram-positive and 4-16 μg/ml for Gram-negative bacteria.[20]_{Fortunately, in a later study, the same group demonstrated the optimization}

of this amphiphilic nebramine derivative, such that antibacterial activity was improved and hemolytic activity was decreased.[21]

(7)

132

The selectivity of AAGs for activity against bacterial over mammalian cells can be attributed to bacteria harboring more negatively charged lipids. Characteristic lipids in bacteria are the negatively charged phospholipids, such as phosphatidylglycerol and cardiolipin (Figure 5a). In contrast, mammalian cell membranes are more abundant in zwitterionic phospholipids, such as phosphatidylcholine (Figure 5b), and this overall neutral charge results in lower interaction with the amphiphiles.

Figure 5. Common phospholipids found in bacterial (a) and mammalian cell membranes (b).

In addition to hemolytic properties, the non-specific interactions of AAGs can also result in cytotoxic effects.[23,24]_{As with hemolysis, the concentrations at which AAGs}

display significant cytotoxicity are often one or several order of magnitudes higher than its MIC value. On a positive note, the immunomodulatory properties of antibacterial AAGs was recently demonstrated for the first time by Guchheit et al.[25]_{They showed}

that, besides antibacterial properties, AAGs can boost the innate immune response and induce immunomodulatory responses. Unfortunately, research in which ototoxicity and nephrotoxicity for AAGs are studied is still lacking.

5.1.3 Goal of the project

In general, AAGs are obtained via polyalkylation[26]_{of all free alcohols or selective}

alkylation of the only primary alcohol[27,28]_{functionality present in the used}

aminoglycoside. The amines are usually not modified in order to allow for protonation under physiological conditions. Only in very limited cases, AAGs have been obtained by regioselective alkylation of the secondary alcohols, although still providing a mixture of components that had to be separated,[29]_{or by relying on long synthetic}

sequences.[17,30,31]_{Especially for neomycin, regioselective modification on alcohols}

other than the primary one has been very limited.

Recently, our group demonstrated a highly regioselective oxidation of pyranosyl glucosides,[32,33]_{which also proved to be applicable to several N-protected}

aminoglycosides (Scheme 1).[34]

The observed selectivity, under our palladium catalyzed oxidation conditions, can be explained by the following set of rules as proposed by Eisink:[35]

(8)

133

1. Oxidation of secondary vicinal diols > secondary alcohols > primary alcohols 2. Oxidation of cyclic vicinal diols > linear vicinal diols

3. The ring oxygen in glucosides dictates the regioselectivity, the position furthest away is most electron rich and therefore more prone to oxidation.

4. Hyperconjugation with the antibonding orbital of the C-H results in weakening of that H bond and accelerates oxidation. Hyperconjugation of a H > C-OH > C-F.

5. Trans di-axial diols retard the rate of oxidation due to unfavorable interaction with the catalyst

6. Steric hindrance around the alcohols slows down the rate of oxidation

Based on these rules, oxidation of the 3’-OH (see structure 3 in Scheme 1 for ring numbering) would be the most favorable. Since this alcohol is part of a glucose configured ring, the trans di-equatorial conformation allows for optimal coordination with the catalyst, and efficient hyperconjugation of the C2’-H and C4’-H bonds. Furthermore, C3’-OH is furthest located from the ring oxygen and therefore the most electron rich alcohol within this ring.

By exploiting this selective oxidation of neomycin B (Scheme 1), a set of AAGs with varying chain lengths at position 3’ were obtained. Unlike previous reports, we synthesized a set of amphiphilic neomycin derivatives that contained the lipophilic group at a position (3’) not reported before in a concise and selective manner.

(9)

134

5.2 Results and discussion

5.2.1 Synthesis of the amphiphiles

In order to prevent catalyst poisoning by the amine groups in the regioselective oxidation, the free amines first had to be protected. Therefore, the synthesis started with the protection of all six amines of the commercially available neomycin trisulfate (a mixture of neomycin B and C) as benzyl carbamates (Cbz) (Scheme 2). It is known that commercially available neomycin B is contaminated with varying amounts of neomycin C depending on supplier and batch, and UPLC-MS of the crude product showed that the ratio of the protected neomycin B relative to its isomer neomycin C was approximately 3:1. (Cbz)6-neomycin B could be separated from (Cbz)6-neomycin C by column

chromatography. Using ammonia-saturated DCM in the column purification resulted in a slightly better separation of the neomycin B and C, compared to the use of regular DCM or methanol containing ammonia. After two successive column purifications neomycin B could be obtained in 76% yield (based on the neomycin B content in the commercial mixture).

Scheme 2. Protection of neomycin trisulfate.

Subjecting protected neomycin B (2) to the oxidation conditions,[34]_{using 5 mol% as}

catalyst in DMSO at rt, resulted only in partial conversion towards the keto neomycin 3, and a moderate 33% (50% brsm) isolated yield (Scheme 3). Performing the reaction in TFE, instead of DMSO, at 28 °C in the presence of 5 mol% catalyst, did not result in significant improvement. Although the conversion was somewhat higher, part of this increased conversion was due to less selective oxidation of 2, and increased overoxidation of 3, according to UPLC-MS. Although this new procedure led to an improved isolated yield (43%), the side products impeded recovery of the unreacted starting material, making it inferior to the original procedure reported by Jäger.[34]

(10)

135

Scheme 3. Regioselective oxidation of (Cbz)6-neomycin B.

With the keto aminoglycoside in hand, we continued the synthesis by introducing the lipophilic tail. As shown in Scheme 1 we planned to introduce the tails by either a reductive amination or an oxime formation. We started with the reductive amination of

3 with octylamine and NaCNBH3 in methanol using an acetate buffer. In analogy to the

reduction of 3 to the 3’-epi neomycin B[34]_{with NaBH}

4, we expected reduction of the

imine intermediate to result in the axial amine product with great selectivity. However, UPLC-MS analysis clearly showed formation of two main products with identical molecular masses in approximately equal amounts.

Scheme 4. Top: reductive amination of 3. Bottom: UPLC-MS spectrum of the crude.

We speculate that under these conditions, formation of enol 6 followed by reprotonation of the enol could result in 8, i.e. the galactose configuration of 3 (Scheme 5). The epimerized 8 can subsequently undergo a reductive amination. Alternatively, the enol could be protonated at C3, effectively resulting in an isomerization of the ketone from C3 to C4. Subsequent reductive amination of this species also results in an identical molecular mass. This theory is supported by the observation of 8 on UPLC-MS,

8

4? 4?

(11)

136

originating from the enol product 7, that can undergo an elimination of the 2-DOS ring system 10.

Scheme 5. E1cB elimination resulting in cleavage of ring I and the 2-DOS ring.

Fortunately, by omitting the NaOAc from the reaction, we could push the reaction towards one major product. We assumed that this is the desired product (Scheme 5). Although one major product was formed, the crude reaction mixture still contained significant amounts of side-products according to UPLC-MS (Scheme 6).

Scheme 6. Top: reductive amination of 3. Bottom: UPLC-MS spectrum of the crude.

We were curious to see whether the oxime bond formation would give rise to cleaner conversion. Therefore, we prepared the desired alkoxyamines via a straightforward two-step procedure. First, N-hydroxyphthalimide (12) was alkylated with an alkyl bromide (11a-f). The alkoxyamine was subsequently liberated with hydrazine in ethanol (scheme 7). The alkoxyamines decompose over time when stored as the free base. To prevent

8

(12)

137

degradation, the alkoxyamines were converted into the corresponding HCl salts, by precipitation from ether by the addition of 1M HCl in ether.

Scheme 7. Synthesis of the alkoxyamines.

With the set of alkoxyamines 14a-f in hand, we subjected the keto aminoglycoside 3 to two equivalents of alkoxyamine 14a (as the HCl salt) in the presence of excess base. Similar to the reductive amination, the reaction formed six products that had a molar mass which corresponded to the desired product. As can be seen from the UPLC-MS spectrum (Scheme 8), three products clearly dominated and were formed in approximately equal amounts, while the other three were formed in significantly lower quantities. All these peaks possessed a molecular weight that corresponded to the expected product 15a.

Scheme 8. Top: oxime formation under basic conditions. Bottom: UPLC-MS chromatogram of

the crude.

At this point, we realized that the excess base was probably not in favor of the selectivity. Therefore, prior to reaction, the alkoxyamine HCl salts were suspended in DCM and transformed to the free base by washing with aqueous NaHCO3. We were

(13)

138

aniline catalyst,[36]_{UPLC-MS showed the formation of only two main products, namely}

the E and Z isomers of 15a (Scheme 9).

Scheme 9. Top: oxime formation under aniline catalysis. Bottom: UPLC-MS chromatogram of

the crude.

Compared to the reductive amination, the obtained crude reaction mixture was much cleaner. Therefore, we decided to introduce the aliphatic residues by oxime bond formation as depicted in Table 1. All derivatives were obtained as a mixture of E/Z isomers in a good yield.

Table 1. Formation of the oximes.

*_{determined by integrating absorbance chromatograms obtained from LCMS analysis.}

15a 15b 15c 15d 15e 15f

yield (%) 74 73 64 66 71 77

(14)

139

Selective deprotection of the Cbz groups was achieved by catalytic hydrogenolysis using Pd/C,[37]_{while still preserving the oxime functionality. Since the deprotection was}

performed in the presence of TFA, the desired amphiphilic aminoglycosides were obtained as the TFA salts in good yields. 16c and 16f were obtained in slightly lower yields compared to the rest of the set (Table 2). After purification by charcoal column chromatography[38]_{only trace amounts of the product were recovered, probably due to}

the absorption of the free amines on the charcoal. Reversed phase column chromatography using a C18 column resulted in satisfactory separation of the amphiphilic aminoglycosides from traces of over hydrogenated 3’-amino-neomycin B and other impurities. However, the recovery was more than 100%, most likely due to leaching of the C18 column. Fortunately, purification by semi-prep RP-HPLC resulted in satisfying purities. Although analytical amounts of the E/Z isomers could be separated on a semi-preparative RP-HPLC column, only partial separation was achieved when larger amounts were injected. This precluded the separation of the isomers. However, this proved not problematic since the oximes isomerize in solution after separation (vide infra). As a result, two fractions in which either the E (first fraction) or Z (second fraction) isomer was enriched were isolated. Compound 16c was obtained in a 1:1.1 mixture of E/Z isomers after preparative RP-HPLC due to unintentional mixing of the fractions.

Table 2. Yields and E/Z ratios for the purified 3’-alkoxyimino-neomycin B derivatives.

Product 16a 16b 16c 16d 16e 16f

Yield

(%) 74 73 57 69 88 45

Fraction 1 2 1 2 mix 1 2 1 2 1 2

E/Z

ratio* 3.5:1 1:3 2.6:1 1:4 1:1.6 1:1 1:2.4 2.4:1 1:2.1 2:1 1:3.2

*_{determined by integrating absorbance chromatograms obtained from LCMS analysis}

The stability of the synthesized amphiphiles is very high. Storage in water at -20 °C over six months did not result in any noticeable hydrolysis of the amphiphiles. However, under these conditions the oxime did isomerize to an E/Z ratio of

(15)

140

approximately 1:1.7. Full equilibration was observed after storing the oxime in water for one day at room temperature. Furthermore, storage of the solid amphiphiles at -20 °C over six months resulted in noticeable isomerization, but to a much lesser extent than when in solution. The tendency for isomerization towards an equilibrium obviates the need to separate the E and Z isomers, since in solution (during the biological assays), the isomerization cannot be suppressed.

With the help of 1_H,13_{C, APT, gCOSY, TOCSY and gHSQC NMR, we were able to}

assign the most relevant peaks and differentiate between the E and Z isomers (Figure 6 for 16c). The peaks assigned with a blue label correspond to the E isomer, red to the Z isomer and a black label to both isomers. Deviation of the chemical shifts between the isomers is small, except for the protons at the 2’ and 4’ positions (circled). A downfield displacement of the chemical shift is observed when the 2’ or 4’ proton is syn to the oxygen of the oxime ether and is consistent with previously reported data[39,40]

Figure 6. HSQC-NMR of 11c with peak assigment.

5.2.1 Viability assay

The antibiotic activity of the synthesized 3’-alkoxyimino-neomycin B derivatives was evaluated on the Gram-positive B. subtilis using a viability assay. To B. subtilis in growth medium were added different concentrations of each fraction of the purified amphiphiles, and the bacteria were cultured for 4 h at 37 °C. Neomycin trisulfate and water were used as controls. Thereafter, resazurin was added to assess the viability of the cells. Resazurin is reduced by viable cells into resorufin, which fluoresces at 590

(16)

141

nm. By measuring the fluorescence intensity, the amount of living cells was determined (Figure 7).

Figure 7. Fluorescence emission at different concentrations of the amphiphilic neomycins.

Figure 7 shows that all amphiphiles display antibacterial activity at sufficiently high concentrations. The derivatives 16a, 16b and 16f have a MIC value of approximately 8 μg/ml and are less active against B. subtilis than amphiphiles 16c, 16d and 16e. These more active derivatives have a MIC value in the 2-4 μg/ml range. Although neomycin showed the best activity (MIC value of 0.5 μg/ml), its molecular weight is much lower than that of the synthetic amphiphiles. The comparison of the MIC values of 16e and neomycin in μM reveals that neomycin sulfate (MIC = 0.6 μM) only has a twofold better antibiotic activity compared to 16e (MIC = 1.3 μM).

The observation that a lipophilic chain of 12-16 carbons (16c-e) has better antibacterial activity than shorter and longer aliphatic chains is in line with the general trend in literature. No significant difference in E and Z isomers was observed in the assay, most probably because isomerization cannot be suppressed under the applied conditions. To the best of our knowledge, only one study has previously tested AAGs on B. subtilis and none of the tested amphiphiles had a MIC value of ≤64 μg/ml.[16]_{In that regard, our}

AAGs showed better results.

5.3 Conclusion

In conclusion, we have synthesized a set of amphiphilic neomycin derivatives with an alkyl residue ranging from C8 up to C18. Starting from the commercially available neomycin trisulfate, the amphiphilic aminoglycosides were obtained in only four synthetic steps and good yields. Because introduction of the lipophilic tail proved to proceed better via oxime bond formation compared to reductive amination, the former was used to prepare the set of amphiphiles. The consequence of this method was that a

0,000 0,200 0,400 0,600 0,800 1,000 1,200 1,400 0,1 1 10 Relativ e v iab ility Concentration (μg/ml) 16aF1 16aF2 16bF1 16bF2 16c 16dF1 16dF2 16eF1 16eF2 16fF1 16fF2 neomycin

(17)

142

mixture of E/Z isomers was obtained, but a preliminary viability assay demonstrated that these isomers did not display significant differences in antimicrobial activity, most probably due to isomerization under the applied conditions. We demonstrated that C12, C14 and C16 residues (16c-e) display better activity against B. subtilis compared to C8 (16a), C10 (16b) and C18 (16f). In contrast to other AAGs,[16]_{our compounds showed}

activity against B. subtilis. Further testing might be useful to determine the activity of our novel AAGs on neomycin-resistant strains.

5.4 Experimental section

General remarks

All solvents used for extraction, filtration and chromatography were of commercial grade, and used without further purification. Neomycin trisulfate was purchased from Sigma-Aldrich. The Pd/C was purchased from Alfa Aesar (Palladium, 10% on carbon, Type 487, dry). [(neocuproine)PdOAc]2OTf2 was prepared according to literature

procedure.[41]_{Benzoquinone was purified prior to use, by sublimation according to a}

literature procedure.[42]_{Other reagents were purchased from Sigma-Aldrich, Acros and}

TCI and were used without further purification.

Flash chromatography was performed manually with silica gel from Silicycle (Sila Flash 40-63 μm, 230-400 mesh) or with automated using a Reveleris flash purification system purchased from Grace Davison Discovery Sciences. Residual palladium was removed employing a MP-TMT (macroporous polystyrene-bound trimercaptotriazine) resin supplied by Biotage.[43]_{Preparative HPLC was performed on a Shimadzu}

LC-20AD HPLC instrument equipped with a FRC-10A fraction collector, using an xTerra® Prep MS C18 10 μm column, H2O + 0.1% TFA/acetonitrile + 0.1% TFA as the eluent

and a flow of 1 ml/min.

TLC was performed with Merck silica gel 60, 0.25 mm plates and visualization was done by staining with potassium permanganate stain (a mixture of KMnO4 (3 g), K2CO3

(10 g), water (300 ml)), with anisaldehyde stain (AcOH : H2SO4 : anisaldehyde 300:6:3)

or with ninhydrin stain (a mixture of ninhydrin (1.5 g), n-butanol (100 ml), and acetic acid (3.0 ml)). LCMS analysis was performed using either a Thermo Scientific Vanquish UHPLC equipped with a LCQ-Fleet orbitrap or an Waters Acquity UPLC TOF, and employing for both machines an Acquity® UPLC HSS T3 1.8μm column with H2O + 0.1% FA/Acetonitrile + 0.1% FA or H2O + 0.1% TFA/Acetonitrile + 0.1%

TFA as the eluent. High-resolution mass spectra (HRMS) were recorded on a Thermo Scientific LTQ Orbitrap XL. Optical rotations were measured on a Schmidt+Haensch polarimeter (Polartronic MH8) with a 10 cm cell (c given in g/100mL) at ambient temperature (±20 °C).

1_H-,13_{C-, APT-, HMQC-, HSQC, gCOSY, TOCSY, were recorded on a Varian}

(18)

143

DMSO-d6, MeOD-d4, CDCl3 or D2O as solvent. Chemical shift values are reported in

ppm with the solvent resonance as the internal standard (DMSO-d6: δ 2.50 for 1_{H, δ}

39.52 for 13_{C; MeOD-d4: δ 3.31 for}1_{H, δ 49.00 for}13_{C; CDCl}

3: δ 7.26 for 1H, δ 77,16

for 13_{C; D}

2O: δ 4.79 for 1H) or with TFA as the internal standard for 13C with D2O (13C

NMR (101 MHz, D2O): δ 162.54 (q, JCF = 37.5 Hz), 116.42 (q, JCF = 289.3 Hz)). The

latter was determined according to a literature method.[44]_{Data are reported as follows:}

chemical shifts (δ), multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, p = pentet, m = multiplet, br = broad), coupling constants J (Hz), and integration.

Experimental procedures

Viability assay:

Discrimination between active and non-active concentrations of compounds (μg/ml) was carried out in 96-microtiter plates in triplicate using a viability assay. The final volume of each experiment was 0.2 mL consisting of 0.15 mL of Mueller-Hinton broth, 0.04 mL of antimicrobial compound in water and 0.01 mL inoculum of B. subtilis (from an exponentially growing culture on Mueller-Hinton medium at OD = 0.4, final OD = 0.02). Neomycin and water were used as controls. The B. subtilis strain 168 was grown in Mueller-Hinton broth, with the antimicrobial compound, at 37 °C for 4 h. Resazurin (0.01 mL of a 0.01% (wt/vol) solution water) was added and the plate was incubated at room temperature for 20 min. The fluorescence emission at 590 nm (excitation 560 nm) was measured to determine the antimicrobial activity.

N-(Cbz)6-neomycin B (2):

To a stirring solution of neomycin trisulfate (12.15 g, 13.36 mmol, 1.0 equiv.) in H2O (43 ml) K2CO3 (6.37 g, 46.9 mmol,

3.5 equiv.) in water (70 ml) was added. Next, THF (146 ml) was added to the solution followed by the dropwise addition of N-(benzyloxycarbonyloxy)-succinimide (31.3 g, 120 mmol, 9.0 equiv.) in THF (82 ml) over 30 min. The suspension was stirred for 18 h at room temperature. The reaction was quenched after completion with glycine (120 mmol, 9.1 g, 9.0 equiv.) in saturated aqueous NaHCO3

(100 ml). After the resulting mixture was stirred for one hour, THF was removed in vacuo. Then, the mixture was basified to pH ±9 with 1 M NaOH. This mixture was extracted with EtOAc (3 × 100 ml). The combined organic layers were dried with MgSO4 and concentrated in vacuo. The obtained white foam was purified by column

chromatography employing 6% MeOH/DCM as the eluent. The DCM was first saturated with ammonia beforehand, by washing the DCM with 25% aqueous ammonia. This resulted in an improved separation of neomycin B & C. The product (10.79 g, 75%) was obtained as a white foam.

(19)

144

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

3OD): δ 7.45 – 7.19 (m, 30H), 5.22 – 4.97 (m, 14H), 4.81 (s,

1H), 4.08 – 3.85 (m, 5H), 3.82 (s, 1H), 3.80 – 3.67 (m, 3H), 3.66 – 3.51 (m, 4H), 3.51 – 3.39 (m, 4H), 3.38 – 3.27 (m, 3H, overlaps with MeOH peak), 3.26 – 3.12 (m, 2H), 1.96 (br d, J = 12.7 Hz, 1H), 1.38 (br dd, J = 23.2, 11.7 Hz, 1H). 13_{C-NMR (101 MHz,} CD3OD, 40°C): δ 159.1, 158.9, 158.7, 158.5, 158.2, 138.2, 138.1, 138.0, 129.6, 129.5, 129.5, 129.4, 129.4, 129.4, 129.3, 129.2, 129.1, 129.0, 128.9, 128.9, 128.9, 128.8, 128.8, 128.8, 128.7, 128.5, 110.7, 100.5, 100.4, 87.1, 83.8, 81.8, 78.2, 75.8, 75.5, 74.5, 73.4, 72.9, 72.3, 71.5, 69.2, 67.9, 67.8, 67.7, 67.6, 67.5, 63.2, 57.8, 54.1, 52.6, 52.5, 43.2, 42.6, 35.4. 3’-keto-N-(Cbz)6-neomycin B (3): General method 1:[34]_N-(Cbz) 6-neomycin B (2) (1.5 g, 1.06

mmol, 1.0 equiv.) and benzoquinone (340 mg, 3.17 mmol, 3.0 equiv.) were dissolved in DMSO (9 ml) and a drop of water was added. [(neocuproine)PdOAc]2OTf2 (26 mg, 26 μmol, 2.5

mol%) was added and the red solution was stirred for 1 hour. A second batch of [(neocuproine)PdOAc]2OTf2 (26 mg, 26

μmol, 2.5 mol%) was added and the mixture was stirred overnight. The reaction was quenched by addition of water (75 ml) and the resulting suspension was filtered and washed with water. The residue was dissolved in DCM and dried with MgSO4. Concentration in vacuo resulted in a crude product that was purified

using automated column chromatography (80 g silica column, DCM/MeOH gradient: 0% for 3 CV, 0-2% in 3 CV, 2-4% in 4 CV, 4% for 3 CV 4-5% in 3 CV, 5% for 5 CV). The product (498 mg, purity 94%, 33% yield, 50% yield based on recovered starting material) was obtained as a yellow foam.

General method 2: N-(Cbz)6-neomycin B (18) (2.67 g, 1.88 mmol, 1.0 equiv.) and

benzoquinone (620 mg, 5.74 mmol, 3 equiv.) were dissolved in 2,2,2-trifluoroethanol (7.5 ml) and a drop of water was added. [(Neocuproine)PdOAc]2OTf2 (94 mg, 94 μmol,

5 mol%) was added and the red solution was stirred for 70 h at 28°C. The mixture was then concentrated in vacuo and co-evaporated twice with DCM to remove all trifluoroethanol. The resulting brown foam was purified using automated column chromatography (80 g silica column, DCM/MeOH gradient: 0% for 3 CV, 0-5% in 8 CV, 5% 10 CV). Mixed fractions with product were recovered and purified again with the same method (40 g silica column). Combined product fractions gave the 3’-keto-(Cbz)6-Neomycin B (1152 mg, purity 93%, yield 43%) as a yellow foam.

1_{H-NMR (400 MHz, CD} 3OD): δ 7.44 – 7.20 (m, 30H), 5.69 (br s), 1H), 5.24 – 4.97 (m, 14H), 4.75 – 4.58 (m, 1H), 4.10 (br d), J = 10.1 Hz, 1H), 4.03 (br s), 2H), 3.98 – 3.79 (m, 5H), 3.79 – 3.33 (m, 12H), 1.95 (d, J = 12.9 Hz, 1H), 1.50 – 1.31 (m, 1H). 13 C-NMR (101 MHz, CD3OD): δ 204.3 (C=O), 159.3, 159.0, 158.5, 158.3, 158.1, 138.3, 138.2, 138.0, 137.7, 129.6, 129.5, 129.4, 129.3, 129.2, 129.0, 129.0, 129.0, 128.9,

(20)

145

128.9, 128.8, 128.7, 109.3, 102.2, 100.3, 85.7, 83.5, 80.7, 78.1, 75.8, 75.7, 75.4, 74.7, 74.6, 71.6, 69.2, 68.3, 67.8, 67.7, 67.6, 67.5, 62.9, 61.8, 54.1, 52.6, 51.9, 43.1, 42.6, 35.2. HRMS (ESI) calculated for C71H81N6O25 ([M+H]+): 1417.52, found: 1417.51.

Preparation of O-alkylhydroxylamine hydrochloride (14a-f):

General method 1: To a solution of N-hydoxyphthalimide (6.78 mmol, 1.5 equiv.) and

NaHCO3 (6.78 mmol, 1.5 equiv.) in DMF (9 ml) at 60 °C the alkyl bromide (4.52

mmol, 1 equiv.) was added. After stirring the red mixture for 16 h, most of the DMF was removed in vacuo. Next, the residue was diluted with DCM (50 ml), washed with water (1× 50 ml), 1 M aqueous NaHCO3 (3× 50 ml) and again with water (2× 50 ml).

The organic layer was dried over MgSO4 and concentrated in vacuo to give the

N-alkoxyphthalimide as a crystalline product, which was used in the next step without further purification. N-alkoxyphthalimide (3.43 mmol, 1.0 equiv.) was dissolved in EtOH (3.5 ml) and hydrazine hydrate (3.77 mmol, 1.1 equiv.) was added to the mixture. The mixture was refluxed for one hour and allowed to cool to room temperature. The obtained suspension was filtered and the residue was washed with Et2O. The filtrate was

concentrated in vacuo and dissolved in Et2O (15 ml). The organic layer was filtered

again if a suspension was obtained. The organic layer was acidified by HCl (g), which was produced by adding concentrated sulfuric acid dropwise to NaCl. This resulted in a suspension, which was filtered to obtain the O-alkylhydroxylamine HCl salt as a white powder.

General method 2: To a solution of N-hydroxyphthalimide (5.0 mmol, 1.5 equiv.) and

NaHCO3 (5.0 mmol, 1.5 equiv.) in DMF (6.6 ml) at 60 °C the alkyl bromide (3.33

mmol, 1 equiv.) was added. After stirring the red mixture for 16 h the mixture was allowed to cool to room temperature and diluted with Et2O (75 ml) and water (50 ml).

The organic layer was washed with water (3× 50 ml), 1 M aqueous NaHCO3 (2× 50

ml), water (1× 50 ml) and brine (1× 50 ml). The organic layer was dried over MgSO4

and concentrated in vacuo to give the N-alkoxyphthalimide as a crystalline product, which was used in the next step without further purification. N-alkoxyphthalimide (3.0 mmol, 1 equiv.) was dissolved in Et2O (18 ml, 0.16 M) and hydrazine hydrate was

added (4.5 mmol, 1.5 equiv.) to the mixture. The suspension was refluxed for 30 min followed by filtration of the suspension. The filtrate was concentrated in vacuo and redissolved in Et2O (12 ml) followed by addition of HCl (1 M in Et2O, 6 ml, 2 equiv.).

The immediately formed suspension was stirred for 30 min and filtered to obtain the O-alkylhydroxylamine HCl salt as a white powder.

O-octylhydroxylamine hydrochloride (14a):

O-octylhydroxylamine was synthesized according to the first general method starting from octyl bromide (1.05 g,

(21)

146

5.47 mmol, 1.0 equiv.), N-hydroxyphthalimide (1.27 g, 7.81 mmol, 1.4 equiv.) and NaHCO3 (0.66 g, 7.81 mmol, 1.4 equiv.). The product (427 mg, 2.36 mmol, 43%) was

obtained as a white powder over two steps.

1_{H-NMR (400 MHz, DMSO-d6): δ 10.95 (s, 3H), 3.98 (t, J = 6.5 Hz, 2H), 1.56 (p, J =}

6.8 Hz, 2H), 1.35 – 1.20 (m, 10H), 0.86 (t, J = 6.7 Hz, 3H). 13_{C-NMR (101 MHz,}

DMSO-d6): δ 73.98, 31.20, 28.55, 28.53, 27.10, 25.11, 22.06, 13.95. HRMS (ESI) calculated for C8H20NO ([M+H]+): 146.154, found: 146.154.

O-decylhydroxylamine hydrochloride (14b):

O-decylhydroxylamine was synthesized according to the first general method starting from decyl bromide (1.00 g, 4.52 mmol, 1.0 equiv.), N-hydroxyphthalimide (1.11 g, 6.78 mmol, 1.5 equiv.) and NaHCO3 (0.57 g, 6.78 mmol, 1.5 equiv.). The product (466 mg, 2.13 mmol, 47%)

was obtained as a white powder over two steps.

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

3OD): δ 4.02 (t, J = 6.5 Hz, 2H), 1.73 – 1.64 (m, 2H), 1.48 –

1.24 (m, 14H), 0.90 (t, J = 6.7 Hz, 3H). 13_{C-NMR (101 MHz, CD}

3OD): δ 76.4, 33.1,

30.6, 30.6, 30.4, 30.3, 28.8, 26.7, 23.7, 14.4. HRMS (ESI) calculated for C10H24NO

([M+H]+_{): 174.185, found: 174.185.}

O-dodecylhydroxylamine hydrochloride (14c):

O-dodecylhydroxylamine was synthesized according to the second general method starting from dodecyl bromide (1.00 g, 4.01 mmol, 1.0 equiv.), N-hydroxyphthalimide (980 mg, 6.02 mmol, 1.5 equiv.) and NaHCO3 (505 mg, 6.02

mmol, 1.5 equiv.). The product (468 mg, 1.97 mmol, 49%) was obtained as a white powder over two steps.

1_{H-NMR (400 MHz, DMSO-d6): δ 10.96 (s, 3H), 3.99 (t, J = 6.5 Hz, 2H), 1.60 – 1.52}

(m, 2H), 1.34 – 1.21 (m, 18H), 0.85 (t, J = 6.8 Hz, 3H). 13_{C-NMR (101 MHz,}

DMSO-d6): δ 74.0, 31.3, 29.0, 29.0, 29.0, 28.9, 28.7, 28.6, 27.1, 25.1, 22.1, 14.0. HRMS (ESI) calculated for C12H28NO ([M+H]+): 202.217, found: 202.216.

O-tetradecylhydroxylamine hydrochloride (14d):

O-tetradecylhydroxylamine was synthesized according to the second general method starting from tetradecyl bromide (1.00 g, 3.61 mmol, 1.0 equiv.), N-hydroxyphthalimide (882 mg, 5.41 mmol, 1.5 equiv.) and NaHCO3 (454 mg, 5.41

(22)

147

1_{H-NMR (400 MHz, DMSO-d6): δ 10.93 (s, 3H), 3.98 (t, J = 6.5 Hz, 2H), 1.56 (p, J =}

6.6 Hz, 2H), 1.35 – 1.18 (m, 22H), 0.85 (t, J = 6.8 Hz, 3H). 13_{C-NMR (101 MHz,}

DMSO-d6): δ 74.0, 31.3, 29.1, 29.05, 29.0, 29.0, 29.0, 29.0, 28.7, 28.6, 27.1, 25.1, 22.1, 13.9. HRMS (ESI) calculated for C14H32NO ([M+H]+): 230.248, found: 230.248.

O-hexadecylhydroxylamine hydrochloride (14e):

O-hexadecylhydroxylamine was synthesized according to the second general method starting from hexadecyl bromide (1.02 g, 3.33 mmol, 1.0 equiv.), N-hydroxyphthalimide (814 mg, 4.99 mmol, 1.5 equiv.) and NaHCO3 (419 mg, 4.99

(m, 2H), 1.35 – 1.15 (m, 26H), 0.89 – 0.81 (m, 3H). 13_{C-NMR (101 MHz, DMSO-d6):}

δ 74.1, 31.3, 29.0, 29.0, 29.0, 28.9, 28.7, 28.6, 27.1, 25.1, 22.1, 14.0. HRMS (ESI) calculated for C16H36NO ([M+H]+): 258.279, found: 258.279.

O-octadecylhydroxylamine hydrochloride (14f):

O-octadecylhydroxylamine was synthesized according to the first general method starting from octadecyl bromide (1.05 g, 3.13 mmol, 1.0 equiv.), N-hydroxyphthalimide (733 mg, 4.50 mmol, 1.44 equiv.) and NaHCO3 (380 mg, 4.50 mmol, 1.44 equiv.). The

product (504 mg, 1.57 mmol, 50%) was obtained as a white powder over two steps.

(m, 2H), 1.33 – 1.13 (m, 30H), 0.83 (t, J = 6.4 Hz, 3H). 13_{C-NMR (101 MHz,}

DMSO-d6): δ 74.1, 31.3, 29.0, 29.0, 29.0, 28.9, 28.7, 28.6, 27.1, 25.1, 22.1, 14.0. HRMS (ESI) calculated for C18H40NO ([M+H]+): 286.310, found: 286.310.

General method for the oxime formation:

Saturated aqueous NaHCO3 (10 ml) was added to the O-alkylhydroxylamine HCl

(14a-f, 288 μmol, 2 equiv.), and the mixture was extracted with DCM (10 ml). The organic

layer was concentrated in vacuo. To the resulting yellow oil were added MeOH (0.6 ml, 0.25 M), acetic acid (432 μmol, 3 equiv.), aniline (14.4 μmol, 10 mol%) and the 3’-keto-N-(Cbz)6-Neomycin B (3) (144 μmol, 1.0 equiv.). The yellow mixture was stirred

for 48 h at room temperature. After completion, the mixture was concentrated in vacuo and purified by column chromatography (manual column, 15 cm height, 11 g silica, starting at 2% MeOH/DCM and the polarity was increased by 0.5% every 6 CV). The products were obtained as white foams and as a mixture of E/Z stereoisomers.

(23)

148

3’-octoxyimino-N-(Cbz)6-neomycin B (15a):

3’-Octoxyimino-N-(Cbz)6-neomycin B (15a) was

synthesized according to the general method starting from 3’-keto-N-(Cbz)6-neomycin B (3) (199 mg, 140

μmol, 1.0 equiv.), O-octylhydroxylamine hydrochloride (14a) (51 mg, 282 μmol, 2.0 equiv.), acetic acid (25 mg, 423 μmol, 3.0 equiv.) and aniline (1.3 mg, 14 μmol, 10 mol%). The product (161 mg, 104 μmol, 74%) was obtained as a mixture of E/Z stereoisomers (1/1.48, respectively).

1_{H-NMR (400 MHz, CD} 3OD): δ 7.37 – 7.20 (m, 30H), 5.47 (appears as br s, 1H), 5.27 – 4.91 (m, 14H), 4.56 (appears as s, 1H), 4.41 (br d), J = 6.3 Hz, 1H), 4.34 – 3.86 (m, 9H), 3.81 – 3.26 (m, 12H), 2.05 (appears as br s, 1H), 1.59 (appears as d, J = 6.5 Hz, 2H), 1.35 (appears as br s, 1H), 1.35 – 1.16 (m, J = 21.8 Hz, 10H), 0.86 (t, J = 6.8 Hz, 3H). 13_{C-NMR (101 MHz, CD} 3OD): δ 159.1, 159.1, 158.8, 158.3, 158.1, 158.0, 158.0, 157.8, 153.1 (C=N-O), 151.0 (C=N-O), 138.1, 138.0, 137.8, 129.5, 129.4, 129.4, 129.0, 128.9, 128.9, 128.8, 128.7, 128.5, 128.5, 109.7, 108.9, 100.3, 98.3, 98.0, 86.2, 83.5, 78.5, 76.1, 76.0, 75.4, 74.6, 71.5, 69.8, 69.2, 68.8, 67.9, 67.7, 67.5, 67.4, 62.6, 54.7, 54.0, 54.0, 53.2, 52.5, 51.4, 51.4, 42.6, 42.1, 35.0, 32.9, 32.9, 30.5, 30.4, 30.3, 30.2, 29.9, 29.9, 26.9, 26.9, 23.6, 23.6, 14.5, 14.5. HRMS (ESI) calculated for C79H97N7NaO25 ([M+Na]+): 1566.64, found: 1566.65. [α]D20 +20 (c 0.55, MeOH). 3’-decoxyimino-N-(Cbz)6-neomycin B (15b):

3’-Decoxyimino-N-(Cbz)6-neomycin B (15b) was

μmol, 1.0 equiv.), O-decylhydoxylamine hydrochloride (14b) (90 mg, 428 μmol, 2.0 equiv.), acetic acid (39 mg, 641 μmol, 3.0 equiv.) and aniline (2 μl, 21.4 μmol, 10 mol%). The product (245 mg, 156 μmol, 73%) was obtained as a mixture of E/Z stereoisomers (1/1.51, respectively).

1_{H-NMR (400 MHz, CD} 3OD): δ 7.41 – 7.00 (m, 30H), 5.58 – 5.39 (m, 1H), 5.28 – 4.91 (m, 14H), 4.57 (br s, 1H), 4.41 (br d, J = 6.0 Hz, 1H), 4.35 – 3.87 (m, 9H), 3.83 – 3.33 (m, 12H), 2.14 – 1.95 (m, 1H), 1.59 (appears as br d, J = 6.3 Hz, 2H), 1.48 – 1.14 (m, 15H), 0.87 (t, J = 6.8 Hz, 3H). 13_{C-NMR (101 MHz, CD} 3OD): δ 159.1, 159.0, 158.9, 158.8, 158.3, 158.3, 158.2, 158.0, 153.1 (C=N-O), 151.0 (C=N-O), 138.1, 137.8, 129.4, 129.4, 129.0, 128.9, 128.8, 128.7, 128.5, 109.6, 108.9, 100.3, 98.4, 98.0, 86.2, 83.5, 78.5, 76.1, 76.0, 75.4, 74.6, 71.4, 69.1, 67.9, 67.8, 67.7, 67.7, 67.6, 67.5, 67.4, 67.4, 62.7, 54.8, 54.1, 54.0, 53.4, 52.6, 51.5, 51.4, 42.7, 42.6, 35.0, 33.0, 32.9, 30.6, 30.6, 30.5, 30.5, 30.4, 30.3, 29.9, 29.8, 26.9, 23.6, 14.5. HRMS (ESI) calculated for C81H101N7NaO25 ([M+Na]+): 1594.67, found: 1594.68. [α]D20 +20 (c 0.81, MeOH).

(24)

149

3’-dodecoxyimino-N-(Cbz)6-neomycin B (15c):

3’-Dodecoxyimino-N-(Cbz)6-neomycin B (15c) was

μmol, 1.0 equiv.), O-dodecylhydroxylamine hydrochloride (14c) (62 mg, 260 μmol, 2.0 equiv.), acetic acid (23 mg, 389 μmol, 3.0 equiv.) and aniline (1.2 mg, 13 μmol, 10 mol%). The product (134 mg, 82 μmol, 63%) was obtained as a mixture of E/Z stereoisomers (1/1.40, respectively).

1_{H-NMR (400 MHz, CD} 3OD, 40oC): δ 7.41 – 7.19 (m, 30H), 5.40 (appears as br s, J = 10.6 Hz, 1H), 5.21 – 4.91 (m, 14H), 4.55 (br s, 1H), 4.42 (br d, J = 7.1 Hz, 1H), 4.34 – 3.83 (m, 9H), 3.81 – 3.32 (m, 12H), 2.07 (br t, J = 13.6 Hz, 1H), 1.60 (br dd, J = 13.1, 6.3 Hz, 2H), 1.46 – 1.18 (m, 19H), 0.88 (t, J = 6.6 Hz, 3H). 13_{C-NMR (101 MHz,} CD3OD, 40oC): δ 159.1, 158.9, 158.8, 158.4, 158.1, 158.0, 157.8, 153.1 (C=N-O), 151.2 (C=N-O), 138.1, 138.0, 137.9, 129.5, 129.4, 129.4, 129.2, 129.0, 129.0, 128.9, 128.9, 128.9, 128.8, 128.8, 128.7, 128.5, 128.5, 109.8, 109.2, 100.4, 98.6, 98.0, 86.2, 86.1, 83.6, 78.7, 76.5, 76.2, 76.0, 75.5, 75.3, 74.6, 71.6, 69.9, 69.3, 68.7, 67.9, 67.6, 67.5, 62.7, 54.8, 54.1, 52.6, 51.6, 51.5, 42.6, 42.2, 35.1, 32.9, 30.7, 30.7, 30.6, 30.6, 30.6, 30.6, 30.5, 30.5, 30.3, 30.3, 29.9, 29.8, 26.9, 26.9, 23.6, 14.4. HRMS (ESI) calculated for C83H105N7NaO25 ([M+Na]+): 1622.71, found: 1622.71. [α]D20 +18 (c 0.77, MeOH). 3’-tetradecoxyimino-N-(Cbz)6-neomycin B (15d):

3’-Tetradecoxyimino-N-(Cbz)6-neomycin B (15d)

was synthesized according to the general method starting from 3’-keto-N-(Cbz)6-neomycin B (3) (200

mg, 141 μmol, 1.0 equiv.), O-tetradecylhydroxylamine hydrochloride (14d) (75 mg, 282 μmol, 2.0 equiv.), acetic acid (25 mg, 432 μmol, 3.0 equiv.) and aniline (1.3 mg, 14 μmol, 10 mol%). The product (149 mg, 86 μmol, 61%) was obtained as a mixture of E/Z stereoisomers (1/1.36, respectively). 1_{H-NMR (400 MHz, CD} 3OD, 40oC): δ 7.43 – 7.16 (m, 30H), 5.42 (appears as br s, 1H), 5.25 – 4.84 (m, 14H), 4.56 (br d, J = 2.6 Hz, 1H), 4.42 (br d, J = 7.3 Hz, 1H), 4.34 – 3.85 (m, 9H), 3.82 – 3.33 (m, 12H), 2.07 (t, J = 14.3 Hz, 1H), 1.60 (br dd, J = 13.3, 6.6 Hz, 2H), 1.47 – 1.17 (m, 23H), 0.88 (t, J = 6.7 Hz, 3H). 13_{C-NMR (101 MHz, CD} 3OD, 40o_{C): δ 159.1, 158.8, 158.8, 158.3, 158.1, 158.0, 157.7, 153.1 O), 151.1} (C=N-O), 138.1, 138.01, 137.9, 129.4, 129.4, 129.4, 129.0, 129.0, 128.9, 128.9, 128.9, 128.8, 128.8, 128.7, 128.5, 128.5, 109.8, 109.1, 100.3, 98.5, 98.0, 86.2, 86.1, 83.6, 78.6, 76.2, 76.0, 75.5, 75.3, 74.6, 71.5, 69.8, 69.3, 68.8, 67.9, 67.7, 67.6, 67.5, 62.7, 54.8, 54.1, 52.5, 51.6, 51.5, 42.6, 42.2, 35.1, 32.9, 30.7, 30.6, 30.6, 30.6, 30.6, 30.5, 30.4, 30.3,

(25)

150

29.9, 29.8, 26.9, 23.6, 14.4. HRMS (ESI) calculated for C85H109N7NaO25 ([M+Na]+):

1650.74, found: 1650.74. [α]D20 +18 (c 0.93, MeOH). 3’-hexadecoxyimino-N-(Cbz)6-neomycin B (15e):

3’-Hexadecoxyimino-N-(Cbz)6-neomycin B (15e) was

μmol, 1.0 equiv.), O-hexadecylhydroxylamine hydrochloride (14e) (76 mg, 260 μmol, 2.0 equiv.), acetic acid (23 mg, 389 μmol, 3.0 equiv.) and aniline (1.2 mg, 13 μmol, 10 mol%). The product (154 mg, 93 μmol, 71%) was obtained as a mixture of E/Z stereoisomers (1/1.59, respectively).

1_{H-NMR (400 MHz, CD} 3OD, 40oC): δ 7.41 – 7.16 (m, 30H), 5.49 – 5.34 (m, 1H), 5.24 – 4.92 (m, 14H), 4.56 (appears as br s, 1H), 4.46 – 4.35 (m, 1H), 4.32 – 3.88 (m, 9H), 3.81 – 3.32 (m, 12H), 2.15 – 1.99 (m, 1H), 1.60 (br dd, J = 12.3, 5.8 Hz, 2H), 1.47 – 1.17 (m, 27H), 0.89 (t, J = 6.6 Hz, 3H). 13_{C-NMR (101 MHz, CD} 3OD, 40oC): δ 159.0, 158.8, 158.7, 158.2, 158.0, 157.9, 157.9, 157.7, 153.0 (C=N-O), 151.0 (C=N-O), 138.0, 137.9, 137.8, 129.4, 129.4, 129.3, 128.9, 128.9, 128.9, 128.8, 128.7, 128.7, 128.5, 128.4, 109.7, 109.0, 100.3, 98.5, 98.0, 86.1, 83.5, 78.6, 76.2, 76.0, 75.4, 74.5, 71.5, 69.2, 67.8, 67.66, 67.6, 67.4, 62.7, 54.8, 54.1, 53.2, 52.5, 51.6, 51.4, 42.5, 42.3, 35.1, 32.9, 30.6, 30.6, 30.6, 30.5, 30.5, 30.4, 30.3, 29.8, 29.8, 26.9, 23.6, 14.5. HRMS (ESI) calculated for C87H113N7NaO25 ([M+Na]+): 1678.77, found: 1678.77. [α]D20 +17 (c 0.66,

MeOH).

3’-octadecoxyimino-N-(Cbz)6-neomycin B (15f):

3’-Octadecoxyimino-N-(Cbz)6-neomycin B (15f) was

μmol, 1.0 equiv.), O-octadecylhydroxylamine hydrochloride (14f) (91 mg, 282 μmol, 2.0 equiv.), acetic acid (25 mg, 423 μmol, 3.0 equiv.) and aniline (1.3 mg, 14 μmol, 10 mol%). The product (178 mg, 105 μmol, 75%) was obtained as a mixture of E/Z stereoisomers (1/1.48, respectively).

1_{H-NMR (400 MHz, CD} 3OD, 40°C): δ 7.38 – 7.18 (m, 30H), 5.39 (br s, 1H), 5.22 – 4.91 (m, 14H), 4.55 (appears as br s, 1H), 4.41 (br d, J = 7.2 Hz, 1H), 4.34 – 3.85 (m, 9H), 3.79 – 3.32 (m, 12H), 2.07 (br t, J = 13.5 Hz, 1H), 1.60 (dd, J = 13.3, 6.5 Hz, 2H), 1.49 – 1.19 (m, 31H), 0.89 (t, J = 6.8 Hz, 1H). 13_{C-NMR (101 MHz, CD} 3OD, 40 °C): δ 159.1, 158.8, 158.8, 158.4, 158.3, 158.1, 158.0, 158.0, 157.7, 153.1 (C=N-O), 151.1 (C=N-O), 138.1, 138.0, 137.9, 129.4, 129.4, 129.4, 129.0, 128.9, 128.9, 128.9, 128.8, 128.7, 128.7, 128.5, 128.5, 109.7, 109.1, 100.4, 98.5, 98.0, 86.2, 86.1, 83.6, 78.6, 76.2,

(26)

151

76.0, 75.5, 74.6, 71.5, 69.9, 69.3, 67.9, 67.7, 67.6, 67.5, 62.7, 54.8, 54.1, 53.2, 52.6, 51.6, 51.5, 42.6, 42.3, 35.1, 32.9, 30.7, 30.6, 30.6, 30.5, 30.5, 30.3, 29.9, 29.8, 26.9, 26.9, 23.6, 14.5. HRMS (ESI) calculated for C89H117N7NaO25 ([M+Na]+): 1706.80,

found: 1706.80. [α]D20 +17 (c 0.53, MeOH). General procedure for the Cbz deprotection:

The 3’-alkoxyimino-N-(Cbz)6-neomycin B derivative (15a-f, 64 μmol, 1.0 equiv.) was

dissolved in a mixture of MeOH : water (12 : 1, 6.4 ml, 0.01 M), TFA (514 μmol, 8.0 equiv.) was added, and the mixture was degassed by three freeze-pump-thaw cycles. After addition of palladium on carbon (10% w/w, 10 mg), the mixture was allowed to stir for 3-4 h under a 1 atm. pressure of H2. After completion, the mixture was filtered

over a plug of celite and washed with MeOH and water. The MeOH was removed in vacuo and MeOH (1 ml) was added. The residual palladium was removed by overnight stirring in the presence of MP-TMT resin (40 mg). Next day, the resin was removed by filtration and the filtrate was concentrated in vacuo, redissolved in water and freeze dried to obtain the crude 3’-alkoxyimino-neomycin B TFA salt and as a white fluffy powder. The crude was further purified by preparative RP-HPLC, which resulted in the partial separation of E/Z stereoisomers.[45]

3’-octoxyimino-neomycin B hexa-trifluoroacetic acid (16a):

3’-Octoxyimino-neomycin B (16a) was prepared according to the general method starting from 3’-octoxyimino-N-(Cbz)6-neomycin B (15a) (101 mg,

71μmol, 1.0 equiv.), TFA (57 mg, 498 μmol, 8.0 equiv.) and Pd/C (10% w/w, 11 mg). The mixture was hydrogenated for 4 h and worked up to obtain the crude 3’-decoxyimino-neomycin B (98 mg, E/Z = 1/1.77, crude yield: 97%). The crude (91 mg, 64 μmol) was purified by preparative RP-HPLC (from 18% to 20% acetonitrile in 3.5 CV) which resulted in three fractions: 16aF1 (22 mg, 15 μmol, E/Z: 3.5:1), 16aF2 (42 mg, 29 μmol, E/Z: 1:3.0) and a mixed fraction (5 mg, 3.5 μmol, E/Z: 1:1.3). Overall yield: 74% HRMS (ESI) calculated for C31H62N7O13

([M+H]+_{): 740.440, found: 740.439.} 16aF1: 1_{H-NMR (400 MHz, D} 2O): δ 5.75 (d, J = 2.6 Hz, 0.33H), 5.66 (d, J = 2.9 Hz, 0.67H), 5.32 (br s, 1H), 5.29 (d, J = 2.2 Hz, 0.33H), 5.25 (d, J = 2.6 Hz, 0.67H), 5.04 (br s, 0.33H), 5.00 (d, J = 2.8 Hz, 0.67H), 4.57 – 4.44 (m, 3H), 4.44 – 4.31 (m, 2H), 4.31 – 4.12 (m, 5H), 3.96 – 3.65 (m, 5H), 3.62 (br s, 1H), 3.60 – 3.13 (m, 6H), 2.57 – 2.48 (m, 1H), 1.97 – 1.81 (m, 1H), 1.78 – 1.65 (m, 2H), 1.43 – 1.23 (m, 10H), 0.87 (t, J = 6.5 Hz, 3H). 3’-E-octoxyimino-neomycin B: 13_{C-NMR (101 MHz, D} 2O): δ 163.0 (q, JCF =

35.5 Hz, TFA), 146.0 (C=N-O), 116.4 (q, JCF = 292.1 Hz, TFA), 110.7, 95.6, 93.5, 84.3,

81.7, 77.4, 76.1, 76.1, 75.6, 73.4, 71.7, 70.2, 67.7, 67.4, 66.1, 60.8, 50.9, 49.8, 48.5, 45.3, 40.5, 39.0, 31.1, 28.6, 28.4, 28.2, 27.8, 25.0, 22.1, 13.4.

(27)

152 16aF2: 1_{H-NMR (400 MHz, D} 2O): δ 5.75 (d, J = 2.6 Hz, 0.78H), 5.66 (d, J = 2.7 Hz, 0.22H), 5.31 (br s, 1H), 5.29 (d, J = 1.1 Hz, 0.78H), 5.24 (d, J = 2.1 Hz, 0.22H), 5.03 (br s, 0.78H), 4.99 (d, J = 2.7 Hz, 0.22H), 4.57 – 4.44 (m, 3H), 4.43 – 4.38 (m, 1H), 4.34 (br s, 1H), 4.30 – 4.11 (m, 5H), 3.93 (dd, J = 12.2, 2.7 Hz, 1H), 3.86 (br s, 1H), 3.84 – 3.65 (m, 3H), 3.62 (br s, 1H), 3.59 – 3.22 (m, 5H), 3.22 – 3.12 (m, 1H), 2.57 – 2.46 (m, 1H), 1.97 – 1.81 (m, 1H), 1.78 – 1.64 (m, 2H), 1.43 – 1.22 (m, 10H), 0.87 (t, J = 6.3 Hz, 3H). 3’-Z-octoxyimino-neomycin B: 13_{C-NMR (101 MHz, D} 2O): δ 162.9 (q, JCF =

35.4 Hz, TFA), 145.7 (C=N-O), 116.4 (q, JCF = 292.1 Hz, TFA), 111.1, 95.6, 95.0, 84.9,

81.7, 76.5, 76.2, 76.1, 75.3, 73.6, 72.0, 70.1, 67.7, 67.4, 61.2, 61.0, 50.9, 49.9, 49.8, 48.6, 40.5, 39.4, 31.1, 28.5, 28.4, 28.2, 27.8, 25.0, 22.0, 13.4.

3’-decoxyimino-neomycin B hexa-trifluoroacetic acid (16b):

3’-Decoxyimino-neomycin B (16b) was prepared according to the general method starting from 3’-decoxyimino-N-(Cbz)6-neomycin B (15b) (101 mg, 64

μmol, 1.0 equiv.), TFA (59 mg, 514 μmol, 8.0 equiv.) and Pd/C (10% w/w, 10 mg). The mixture was hydrogenated for 4 h and worked up to obtain the crude 3’-decoxyimino-neomycin B (91 mg, E/Z = 1/2.04, crude yield: 98%). The crude (86 mg, 59 μmol) was purified by preparative RP-HPLC (from 24.5% to 26.5% acetonitrile in 3.5 CV), which resulted in three fractions: 16bF1 (18 mg, 12 μmol, E/Z: 2.6:1), 16bF2 (44 mg, 30 μmol, E/Z: 1:3.85) and a mixed fraction (2 mg, 1.3 μmol, E/Z: 1:1.15). Overall yield: 73% HRMS (ESI) calculated for C33H66N7O13 ([M+H]+): 768.471, found: 768.469 16bF1: 1_{H-NMR (400 MHz, D} 2O): δ 5.75 (d, J = 2.9 Hz, 0.21H), 5.66 (d, J = 3.0 Hz, 0.79H), 5.32 (br s, 1H), 5.30 (br s, 0.21H), 5.25 (d, J = 2.7 Hz, 0.79H), 5.04 (d, J = 1.6 Hz, 0.21H), 4.99 (d, J = 3.0 Hz, 0.79H), 4.56 – 4.43 (m, 3H), 4.43 – 4.30 (m, 2H), 4.30 – 4.12 (m, 5H), 3.97 – 3.65 (m, 5H), 3.62 (br s, 1H), 3.60 – 3.50 (m, 1H), 3.50 – 3.13 (m, 5H), 2.53 (dt, J = 12.5, 4.2 Hz, 1H), 1.97 – 1.82 (m, 1H), 1.78 – 1.66 (m, 2H), 1.43 – 1.23 (m, 14H), 0.87 (t, J = 6.7 Hz, 3H). 3’-E-decoxyimino-neomycin B: 13_C-NMR

(101 MHz, D2O): δ 163.0 (q, JCF = 35.4 Hz, TFA), 146.0 (C=N-O), 116.4 (q, JCF =

292.1 Hz, TFA), 110.7, 95.7, 93.5, 84.3, 81.7, 77.5, 76.1, 76.1, 75.6, 73.4, 71.8, 70.2, 67.7, 67.4, 66.1, 60.8, 51.0, 49.8, 48.6, 45.4, 40.6, 39.1, 31.3, 28.8, 28.6, 28.6, 28.2, 27.8, 25.0, 22.1, 13.5. 16bF2: 1_{H-NMR (400 MHz, D} 2O): δ 5.74 (d, J = 2.5 Hz, 0.84H), 5.66 (d, J = 2.8 Hz, 0.16H), 5.30 (appears as d, J = 7.8 Hz, 1.84H), 5.25 (d, J = 2.2 Hz, 0.16H), 5.01 (br s, 0.84H), 4.96 (d, J = 2.6 Hz, 0.16H), 4.55 – 4.43 (m, 3H), 4.43 – 4.37 (m, 1H), 4.34 (br s), 1H), 4.29 – 4.18 (m, 4H), 4.13 (t, J = 9.7 Hz, 1H), 3.92 (dd, J = 11.9, 2.5 Hz, 1H), 3.88 – 3.64 (m, 4H), 3.61 (br s, 1H), 3.55 – 3.30 (m, 4H), 3.26 (dd, J = 13.6, 3.0 Hz, 1H), 3.21 – 3.11 (m, 1H), 2.56 – 2.44 (m, 1H), 1.94 – 1.78 (m, 1H), 1.76 – 1.64 (m,

(28)

153

2H), 1.44 – 1.20 (m, 14H), 0.87 (t, J = 6.3 Hz, 3H). 3’-Z-decoxyimino-neomycin B:

13_{C-NMR (101 MHz, D}

2O): 163.0 (q, JCF = 35.5 Hz, TFA), 145.9 (C=N-O), 116.4 (q,

JCF = 291.8 Hz, TFA), 111.1, 95.6, 95.2, 84.9, 81.7, 76.7, 76.2, 76.1, 75.2, 73.7, 72.1,

70.1, 67.7, 67.5, 61.3, 61.2, 51.0, 50.0, 49.8, 48.6, 40.5, 39.5, 31.3, 28.8, 28.8, 28.6, 28.6, 28.2, 27.9, 25.0, 22.1, 13.5.

3’-dodecoxyimino-neomycin B hexa-trifluoroacetic acid (16c):

3’-Dodecoxyimino-neomycin B (16c) was prepared according to the general method starting from 3’-dodecoxyimino-N-(Cbz)6-neomycin B (15c) (88 mg,

55μmol, 1.0 equiv.), TFA (50 mg, 440 μmol, 8.0 equiv.) and Pd/C (10% w/w, 8 mg). The TFA was added 4 h after starting the reaction. The mixture was hydrogenated for an extra 2.5 h, and worked up to obtain the crude 3’-decoxyimino-neomycin B (66 mg, E/Z = 1/1.77, crude yield: 81%). The crude (60 mg, 41 μmol) was purified by preparative RP-HPLC (from 29.5% to 32.5% acetonitrile in 3.5 CV) which resulted in three fractions: 16cF1 (4 mg, 2.7 μmol, E/Z: 2.5:1), 16cF2 (28 mg, 19 μmol, E/Z: 1:2.7) and a mixed fraction (10 mg, 6.8 μmol, E/Z: 1.2:1). Later during analysis fractions F1 and F2 were accidently mixed (E/Z: 1:1.57). Overall yield: 57% HRMS (ESI) calculated for C35H70N7O13 ([M+H]+):

796.503, found: 796.502. 1_{H-NMR (400 MHz, D} 2O): δ 5.74 (d, J = 2.6 Hz, 0.64H, Z), 5.66 (d, J = 2.9 Hz, 0.36H, E), 5.31 (br s, 1H), 5.29 (br s, 0.64H, Z), 5.25 (d, J = 2.5 Hz, 0.36H, E), 5.01 (d, J = 1.4 Hz, 0.64H, Z), 4.94 (d, J = 2.8 Hz, 0.36H, E), 4.55 – 4.42 (m, 3H), 4.42 – 4.30 (m, 2H), 4.29 – 4.17 (m, 4H), 4.12 (t, J = 9.6 Hz, 1H), 3.96 – 3.72 (m, 4H), 3.72 – 3.64 (m, 1H), 3.61 (br s, 1H), 3.58 – 3.12 (m, 6H), 2.50 (d, J = 12.5 Hz, 1H), 1.94 – 1.78 (m, 1H), 1.78 – 1.64 (m, 2H), 1.41 – 1.20 (m, 18H), 0.87 (t, J = 6.7 Hz, 3H). 13_{C-NMR (101}

MHz, D2O): δ 163.0 (q, JCF = 35.3 Hz, TFA), 146.5 (3’, Z), 146.1 (3’, E), 116.4 (q, JCF =

291.8 Hz, TFA), 111.0 (1’’, Z), 110.6 (1’’, E), 95.7 (1’’’, E), 95.6 (1’’’, Z), 95.3 (1’, Z), 93.7 (1’, E), 84.8 (5, Z), 84.3 (5, E), 81.8 (4’’, Z), 81.7 (4’’, E), 77.6 (4, E), 76.7 (4, Z), 76.2 (3’’), 76.1 (1alk_{), 76.0 (1}alk_{), 75.4 (5’, E), 75.1 (5’, Z), 73.7 (2’’, Z), 73.4 (2’’, E),}

72.1 (6, Z), 71.9 (6, E), 70.2 (5’’’, E), 70.2 (5’’’, Z), 67.7 (3’’’), 67.5 (4’’’, Z), 67.4 (4’’’, E), 66.2 (4’, E), 61.3 (4’, Z), 61.2 (5’’, Z), 60.8 (5’’, E), 51.0 (2’’’), 50.0 (2’, Z), 49.8 (1), 48.7 (3, Z), 48.6 (3, E), 45.8 (2’, E), 40.6 (6’’’), 39.6 (6’, Z), 39.2 (6’, E), 31.3 (alk_{), 28.9 (}alk_{), 28.9 (}alk_{), 28.8 (}alk_{), 28.8 (}alk_{), 28.6 (}alk_{), 28.2 (}alk_{), 28.0 (2), 25.1 (}alk_{), 22.1}