Synthesis of 4-O-(4-Amino-4-deoxy-β-D-xylopyranosyl)paromomycin and 4-S-(β-D-Xylopyranosyl)-4-deoxy-4’-thio-paromomycin and Evaluation of their Antiribosomal and Antibacterial Activity

Abstract

The design, synthesis and antiribosomal and antibacterial activity of two novel glycosides of the aminoglycoside antibiotic paromomycin are described. The first carries of 4-amino-4-deoxy-β-D-xylopyranosyl moiety at the paromomycin 4’-position and is approximately two-fold more active than the corresponding β-D-xylopyranosyl derivative. The second is a 4’-(β-D-xylopyranosylthio) derivative of 4’-deoxyparomomycin that is unexpectedly less active than the simple β-D-xylopyranosyl derivative, perhaps because of the insertion of the conformationally more mobile thioglycosidic linkage.

Graphical Abstract

1. Introduction

The ever-expanding spread of multidrug resistant infectious diseases necessitates a constant search for new anti-infective agents. With this in mind and focusing on Gram-negative infections, in our laboratories we have focused on the development of next generation aminoglycoside antibiotics (AGAs) because of their many favorable attributes which include i) a well-established mechanism of action, ii) broad spectrum activity, iii) minimal impact on the intestinal microbiota, and iv) absence of allergic reactions [1–4]. We took our lead from the atypical 2-deoxystreptamine (DOS) type aminoglycoside apramycin 1 (Figure 1) that features an unusual 4-aminoglucosyl substituted octadiosyl moiety at the DOS 4-position, but lacks additional substitution on the DOS ring unlike the more common 4,5- and 4,6-disubstituted DOS AGAs. The structural features of apramycin 1 are such that it is not inactivated by the enormous majority of aminoglycoside modifying enzymes (AMEs), the most common mechanism of AGA resistance [5–8], nor by the increasingly prevalent ribosomal methyltransferases that block action of all AGAs in clinical use by methylation of G1405 in the drug binding pocket of the target bacterial ribosome [9, 10], and consequently retains activity against many ESKAPE pathogens [11–25]. Furthermore, unlike most AGAs in clinical use apramycin 1 displays minimal drug-induced hearing loss or ototoxicity [3, 11, 26], making it an ideal starting point for optimization [27]. One of the avenues that we have pursued is to use the widely available and somewhat more active 4,5-AGA paromomycin 3 as starting material and modify it to more closely resemble apramycin 1 with the goal of incorporating the more favorable attributes of 1 and minimizing the less favorable ones of 2, which include greater toxicity and more widespread resistance. To this end we have extensively modified the 4’-position of paromomycin preparing an extensive series of paromomycin 4’-O-glycosides 4 resulting in considerable SAR and the model depicted in Figure 2, with the optimal glycoside being the β-D-xylopyranoside 5 [28, 29].

Figure 1.

Apramycin, Saccharosin, Paromomycin 4’-Derivatives and the Target Molecules 6 and 10.

Figure 2.

SAR Model for the Paromomycin 4’-Glycosides

Building on this model and noting that the presence of the amino group in the 4-aminoglucosyl residue of apramycin 1 confers greater activity with respect to its biosynthetic precursor saccharosin 2 [30–33], which differs only by the presence of a glucosyl ring in place of the aminoglucosyl one, we designed the 4’-O-(4-amino-4-deoxy-β-D-xylopyranosyl) paromomycin derivative 6 on which we now report. In parallel with the paromomycin glycosides we have also investigated an extensive series of 4’-O-alkyl ethers and hydroxyalkyl ethers of paromomycin and especially their 4’-deoxy-4’-alkylthio- and 4’-deoxy-4’-alkyl derivatives analogs. The optimal ether was the 4’-O-ethyl derivative 7, which displayed excellent selectivity for the target prokaryotic ribosome over its eukaryotic counterparts despite a modest loss in activity. The selectivity and activity of 7 were surpassed however by 4’-deoxy-4’-ethylthioparomomycin 8 and eventually 4’-deoxy-4’-propylparomomycin, or propylamycin, 9 leading to the conclusion that selectivity and activity is greater with a short substituent linked to the 4’-position by a less electronegative atom (S or C) than the native oxygen [34–36]. Combining the beneficial influences of i) the 4’-O-β-D-xylopyranosyl modification, and ii) the beneficial influence of 4’-deoxygenation, we designed the 4’-deoxy-4’-β-D-xylopyranosylthio paromomycin derivative 10 on which we also report here.

2. Results

2.1. Synthesis.

To prepare the 4-amino-4-deoxy-β-D-xylopyranoside 6 with the help of stereodirecting neighboring group participation we required donor 14. To this end, in a rare application of the Shoda method [37, 38] to a pentose sugar, L-arabinose was treated with 2-chloro-N,N′-dimethylimidzolinium chloride, thiophenol, and triethylamine in aqueous acetonitrile at room temperature. After complete conversion of the substrate the reaction mixture was concentrated to dryness, filtered through a plug of silica gel, and heated with 2,3-butanediol, trimethyl orthoformate, and camphor-10-sulfonic acid [39, 40] to give, after work-up and chromatographic purification the desired bisacetal 12 in 23% yield and a number of minor byproducts of which only the most abundant isomer 13 was isolated in 4% yield (Scheme 1). The structure of 13 was inferred from nuclear Overhauser difference spectroscopy measurements and ultimately confirmed by X-ray crystallography (Supplementary Data; CCDC 2238003). Subsequent triflation followed by displacement with lithium azide gave the D-xylo derivative 14, that was converted to the desired donor 15 by cleavage of the bisacetal and subsequent acetylation under standard conditions (Scheme 1).

Scheme 1.

Preparation of the 4-azido-4-deoxy-D-xylopyranosyl donor 15.

The glycosyl acceptor 16, obtained from paromomycin in 6 steps as described previously [36], was coupled to donor 15 with activation by N-iodosuccinimide and trifluoroacetic acid at −78 °C in dichloromethane giving, after quenching and chromatographic purification, giving the desired equatorial glycoside 17 in 35% yield (Scheme 2). Finally, deprotection of 17 was achieved by controlled ester hydrolysis with magnesium methoxide in methanol, followed by cleavage of the trifluoroacetamide groups with hot aqueous barium hydroxide, and finally hydrogenolysis of the azido groups giving the desired glycoside 6 as its hexaacetate salt in 50% overall yield after filtration over Sephadex C-25 resin and lyophilization from acetic acid.

Scheme 2.

Synthesis of glycoside 6.

To obtain the thioglycoside 10 we prepared per-O-acetyl β-D-thioxylopyranose 18 according to the literature procedure [41], and briefly exposed it to sodium methoxide in methanolic dichloromethane at −25 °C to selectively remove the acetyl group from the anomeric position. After concentration the so-obtained sodium salt was stirred with the previously described triflate 19 [42] in DMF at 0 °C. The crude reaction mixture was then treated with sodium methoxide in methanol leading, after chromatographic work-up, to the thioxyloside 20 in 89% yield. Staudinger reaction of 20 with trimethylphosphine [42] was followed by hydrogenolysis over palladium hydroxide, filtration through Sephadex C-25 and lyophilization from acetic acid to give the target thioglycoside 10 in the form of its pentaacetate salt in 42% yield (Scheme 3).

Scheme 3.

Synthesis of thioglycoside 10.

2.2. Cell-Free Translation Assays:

Glycosides 6 and 10 were subject to cell-free translation assays to assess the disruption of protein synthesis by bacterial ribosomes (Table 1) along with comparators apramycin 1, saccharosin 2, paromomycin 3, 4’-O-β-D-xylopyranosyl paromomycin 5, 4’-O-ethylparomomycin 7, and 4’-deoxy-4’-ethylthioparomomycin 8 (Table 1). The ability of glycosides 6 and 10 and all comparators to inhibit protein synthesis by hybrid bacterial ribosomes carrying the complete decoding A site of the human mitochondrial ribosome (Mit13), the A1555G mutant of the human mitochondrial ribosome (A1555G), and the human cytoplasmic ribosome (Cyt14) [43] were also investigated, leading to the activities and eukaryotic/prokaryotic selectivities displayed in Table 1.

Table 1.

Antiribosomal activities and selectivities of studied derivatives^[a]

Compound	IC50 (μM)				Selectivity
	wt	Mit13	A1555G	Cyt14	Mit13	A1555G	Cyt14
Apramycin, 1	0.144	114	105	158	791	729	1097
Saccharosin, 2	0.28	580	567	612	2071	2025	2186
Paromomycin, 3	0.035	139	13	34	3971	371	971
Xyloside, 5	0.045	192	203	127	4267	4511	2822
4’-O-Ethyl parom, 7	0.12	154	177	160	1283	1475	1333
4’-Ethylthioparom, 8	0.053	166	76	104	3132	1434	1962
4-Aminoxyloside, 6	0.040	16	20	25	400	500	625
Thioxyloside, 10	0.105	149	141	116	1419	1343	1105

^[a]Ran in duplicate

The aminoxyloside 6 had activity similar to xyloside 5 and comparable to the parent paromomycin 3 for the inhibition of the wild-type bacterial ribosome. This result reflects the minor two-fold difference in activity observed on replacement of the pendant aminoglucosyl ring of apramycin by a simple glucosyl ring as in saccharosin 2. As seen in the apramycin 1/saccharosin 2 pair, the aminoxyloside 6 displayed greater inhibitory activity for the eukaryotic ribosomes than the simple xyloside 5, leading to an across-the-board reduction in selectivity. The replacement of the xylopyranosyloxy appendage of xyloside 5 by the corresponding thioglycoside in 10 leads to an approximately two-three-fold loss of activity in inhibition of the wild-type bacterial ribosome and stands in contrast to the increased activity observed on going from ether 7 to thioether 8. The thioxyloside modification 10 is also accompanied by a reduction in selectivity compared to the simple xyloside 5.

2.3. Antibacterial Activity:

Glycosides 6 and 10 were also investigated for their ability to inhibit growth of various Gram-negative pathogens and methicillin-resistant Staphylococcus aureus (MRSA), again in comparison to 1, 2, 3, 5, 7, and 8 (Table 2)

Table 2.

Antibacterial Activity MIC (μg/mL) ^[a]

	E. coli	P. aerug	A. baum	MRSA	E. coli
Strain	AG001	AG220b	AG309	AG038	AG173c
Apramycin, 1	4	4	4	4	256
Saccharosin, 2	8	8–16	8	16–32	>256
Paromomycin, 3	2–4	>256	2–4	2	4–8
Xyloside, 5	16	16–32	16	16–32	64
4’-O-Ethyl parom, 7	16	64–128	16–32	16	128–256
4’-Ethylthioparom, 8	4–8	32	8–16	4	16–32
4-Aminoxyloside, 6	8–16	8–16	16	8	16–32
Thioxyloside, 10	32–64	128	32–64	32	128–256

^a)Ran in duplicate

^b)Carries APH(3’)-II

^c)Carries AAC(3)-IV

Consistent with the trends seen in the inhibition of the wild-type bacterial ribosomes the aminoglycoside 6 is a marginally better inhibitor of bacterial growth than the xyloside 5, just as apramycin 1 is a marginally better antibacterial agent than its desamino-hydroxy analog saccharosin 2. Likewise consistent with the trend observed for inhibition of the wild-type bacterial ribosomes, the thioxyloside 10 is a somewhat less effective antibacterial agent than the simple xyloside 5.

3. Discussion.

The cell-free translation assays of ribosomal protein synthesis and the antibacterial assays reveal that the incorporation of an equatorial amino group into the 4-position of a pyranosyl residue affixed to either the 4’-position of paromomycin or the analogous 8′-position in apramycin results in a minor increase in activity, but one that is unfortunately offset by a reduction in selectivity for prokaryotic over eukaryotic ribosomes. As inhibition of the eukaryotic mitochondrial ribosomes Mit13 and especially A1555G is associated with ototoxicity and that of the cytoplasmic ribosome Cyt13 with systemic toxicity [1, 26, 44–49], there is little benefit to be gained from installation of the additional amino group. The changes in activity and selectivity between 5 and 6 are consistent with those between 1 and 2, and are therefore supportive of the overall model. The reduction in selectivity on going from 5 to 6 arises because of the moderately increased inhibition of the eukaryotic ribosomes: this is presumably because the increased electrostatic interaction with the ribosomes on addition of an extra amino group (protonated at physiological pH) has greater impact on the relatively weak complexes with the eukaryotic ribosomes than on the already highly optimized complex with the bacterial ribosome. The reduction in activity and selectivity on going from the xyloside 5 to the thioxyloside 10 stand in contrast to those seen between ether 7 and thioether 8. This difference in effect of replacement of the 4’- oxygen atom by a sulfur atom between the two pairs of compounds presumably reflects the fact that the change between the xyloside 5 and the thioxyloside 10 is a more complex one than that between ether 7 and thioether 8. In the latter case an ether oxygen is replaced by a thioether sulfur, which has the effect of lengthening two bonds and, as previously discussed [36] of increasing the electron density and hydrogen bond accepting ability of the ring oxygen (O-5′) in ring I. In going from 5 to 10 however the change involves replacing a glycosidic oxygen by a thioglycosidic sulfur, which not only results in the lengthening of two bonds but also relaxes the conformational constraints of the glycosidic bond [50, 51] suggesting that thioglycoside 10 is a less than ideal fit for the model of Figure 2. Presumably, the looser complexes with the eukaryotic ribosomes are better able to accommodate the changes in structure due to the insertion of the thioglycoside in 10 than the tighter complex with the bacterial ribosome, resulting in the observed reduction in selectivity.

4. Conclusion.

Two novel paromomycin glycosides have been designed and prepared by combination of aspects of previous SAR models. The effects observed on incorporation of an additional amino group into the xylose ring of 4’-O-xyloside 5 were consistent with expectations based on the differences in structure and activity of apramycin 1 and saccharosin 2 and generally validate the design hypothesis. In contrast, the changes observed on replacement of the glycosidic oxygen in 5 by sulfur in 10 do not reflect those seen by the analogous replacement of an ether oxygen in 7 by a sulfur atom in thioether 8 most likely reflecting the more complex nature of glycosidic, and thioglycosidic bonds compared to ethers and thioethers.

5. Experimental information

5.1. General information.

All reagents and solvents were purchased from commercial suppliers and were used without further purification unless otherwise specified. All experiments were carried out under a dry argon atmosphere unless otherwise specified. Progress of reactions was monitored by silica gel thin-layer chromatography with plates visualized by UV irradiation (254 nm) and by charring with sulfuric acid in ethanol, or potassium permanganate solution. Flash column chromatography was carried out over silica gel. Optical rotations were measured at 589 nm and 20 °C on a digital polarimeter with a path length of 10 cm. ¹H and ¹³C NMR spectra of all compounds were recorded using at 500 MHz or 600 MHz instruments unless otherwise stated and assignments made with the help of COSY, HMBC, and HSQC spectra. ESI-HRMS were recorded using a time-of-flight mass spectrometer or an Orbitrap mass spectrometer, both fitted with an electrospray source.

5.2. Phenyl 2,3-O-(2’,3’-dimethoxybutane-2’,3’-diyl)-1-thio-β-L-arabinopyranoside (12) and Phenyl 3,4-O-(3’,4’-dimethoxybutane-3’,4’-diyl)-1-thio-α-L-arabinopyranoside (13)

A stirred solution of L-arabinose (2.0 g, 13 mmol) in 1:1 water: acetonitrile (5 mL), triethylamine (18 mL, 129 mmol) and thiophenol (6.4 mL, 53.8 mmol) was cooled to 0 °C and treated portion-wise with 2-chloro-1,3-dimethylimidazolinium chloride (6.8 g, 40 mmol). Stirring was continued for 2 h before the solvent was evaporated, and the crude mixture was filtered through silica gel (eluent: methanol: ethyl acetate 1:99) and concentrated to give a residue (900 mg) that was taken up in methanol (90 mL) under Ar atmosphere, treated with trimethyl orthoformate (3.0 mL, 29.7 mmol), butane-2,3-dione (0.8 mL, 9.3 mmol) and camphorsulfonic acid (105 mg, 0.34 mmol) and heated to reflux with stirring for 18 h. The reaction mixture was brought to room temperature and quenched with triethylamine (0.1 mL .07 mmol) and concentrated under vacuum. The crude mixture was purified by flash column chromatography (ethyl acetate: hexane 1:4) to give 12 as a yellow oil (1.1 g, 3.09 mmol, 23% two steps) and isomer 13 as a white solid (180 mg, 0.51 mmol, 4% two steps).

12: ${[\alpha ]}_D^{23}=-\mathit{69.5}$ (c = 0.49, chloroform). ¹H NMR (600 MHz, CDCl₃) δ 7.67−7.42 (m, 2H, SPh), 7.42 − 7.14 (m, 3H, SPh), 4.69 (dd, J = 9.8, 1.3 Hz, 1H, H1), 4.09 (dd, J = 12.9, 1.9 Hz, 1H, H5-eq), 4.04 (td, J = 9.9, 1.3 Hz, 1H, H2), 3.94 (dt, J = 3.3, 1.6 Hz, 1H, H4), 3.78 (dd, J = 9.8, 3.0 Hz, 1H, H3), 3.60 (dd, J = 12.8, 1.6 Hz, 1H, H5-ax), 3.25 (d, J = 1.2 Hz, 3H, OCH₃), 3.17 (d, J = 1.1 Hz, 3H, OCH₃), 1.31 (s, 3H, CH₃), 1.31 (s, 3H, CH₃). ¹³C NMR (150 MHz, CDCl₃) δ 133.7 (SPh), 132.0 (SPh), 128.9 (SPh), 127.5 (SPh), 100.58 (C- quaternary), 100.55 (C-quaternary), 86.3 (C1), 71.5 (C3), 70.6 (C5), 67.9 (C4), 65.5 (C2), 48.2 (OCH₃), 48.2 (OCH₃), 17.9 (CH₃), 17.7 (CH₃). ESI-HRMS: m/z calcd for C₁₇H₂₄O₆SNa [M+Na]⁺ 379.1191, found 379.1192.

13: m.p = 133-134 °C. ${[\alpha ]}_D^{23}=+\mathit{121.3}$ (c 0.21, chloroform). ¹H NMR (500 MHz, CDCl₃) δ 7.43 − 7.39 (m, 2H, SPh), 7.25 − 7.15 (m, 3H, SPh), 5.56 (d, J = 5.4 Hz, 1H, H1), 4.86 (ddd, J = 10.3, 9.1, 5.4 Hz, 1H, H2), 4.33 (dd, J = 13.4, 2.5 Hz, 1H, H5-eq), 4.05 (ddd, J = 3.9, 2.4, 1.3 Hz, 1H, H4), 3.74 (dt, J = 13.3, 1.0 Hz, 1H, H5-ax), 3.54 (dd, J = 10.4, 4.0 Hz, 1H, H3), 3.34 (s, 3H, OCH₃), 3.18 (s, 3H, OCH₃), 2.07 (d, J = 9.1 Hz, 1H, OH), 1.28 (s, 3H, CH₃), 1.26 (s, 3H, CH₃). ¹³C NMR (126 MHz, CDCl₃) δ 134.4 (SPh), 131.5 (SPh), 129.2 (SPh), 127.5 (SPh), 100.7 (C-quaternary), 99.0 (C-quaternary), 92.7 (C1), 72.2 (C3), 68.4 (C2), 65.2 (C4), 63.3 (C5), 49.2 (OCH₃), 48.1 (OCH₃), 18.9 (CH₃), 18.0 (CH₃). ESI-HRMS: m/z calcd for C₁₇H₂₄O₆SNa [M+Na]⁺ 379.1191, found 379.1172.

5.3. Phenyl 2,3-O-(2’,3’-dimethoxybutane-2’,3’-diyl)-4-azido-4-deoxy-1-thio-β-D-xylopyranoside (14).

A stirred solution of 12 (200 mg, 0.56 mmol) in dichloromethane (1.5 mL) and pyridine (0.4 mL, 5.6 mmol) under Ar atmosphere was cooled to −20 °C, treated with trifluoromethanesulfonic anhydride (0.4 mL, 0.8 mmol) and stirred for 1 h. The reaction mixture was brought to −5 °C and diluted with dichloromethane and quenched with saturated aqueous NaHCO₃. The organic layer was washed with saturated NaHCO₃ (30 mL × 3), 1 N HCl (30 mL × 3) and saturated NaCl (20 mL), then concentrated and dried under high vacuum for 0.5 h. The residue was dissolved in N,N-dimethylformamide (2 mL) and was treated with LiN₃ (64 mg, 1.40 mmol). The reaction mixture was stirred for 24 h at room temperature before the solvent was evaporated, and the residue was dissolved in ethyl acetate and washed with water (20 mL × 3) and brine (20 mL). The organic layer was concentrated, and purified by column chromatography over silica gel (ethyl acetate: hexane 1: 39) to afford 14 (180 mg, 0.47 mmol, 84%) as a white solid. ${[\alpha ]}_D^{23}=-\mathit{78.9}$ (c = 0.78, chloroform). ¹H NMR (600 MHz, CDCl₃) δ 7.64 − 7.40 (m, 2H, SPh), 7.38 − 7.18 (m, 3H, SPh), 4.68 (d, J = 9.8 Hz, 1H, H1), 4.03 (dd, J = 11.7, 5.3 Hz, 1H, H5-eq), 3.80 (t, J = 9.7 Hz, 1H, H3), 3.72 (td, J = 10.2, 5.3 Hz, 1H, H4), 3.65 (t, J = 9.6 Hz, 1H, H2), 3.33 (s, 3H, OCH₃), 3.21 (s, 3H, OCH₃), 3.17 (dd, J = 11.8, 10.4 Hz, 1H, H5-ax), 1.34 (s, 3H, CH₃), 1.33 (s, 3H, CH₃). ¹³C NMR (150 MHz, CDCl₃) δ 132.9 (SPh), 131.9 (SPh), 128.9 (SPh), 127.6 (SPh), 100.4 (C- quaternary), 100.0 (C- quaternary), 85.9 (C1), 73.9 (C3), 68.3 (C2), 68.1 (C5), 57.7 (C4), 48.2 (OCH₃), 48.1 (OCH₃), 17.57 (CH₃), 17.55 (CH₃). ESI-HRMS: m/z calcd for C₁₇H₂₃O₅N₃SNa [M+Na]⁺ 404.1256, found 404.1261.

5.4. Phenyl 2,3-di-O-acetyl-4-azido-4-deoxy-1-thio-β-D-xylopyranoside (15).

A solution of 14 (180 mg, 0.47 mmol) in trifluoroacetic acid: water, 9:1 (2 mL) was stirred for 5 min at room temperature. The solvent was evaporated and the residue was dried under high vacuum for 2 h then was dissolved in pyridine (1 mL) and acetic anhydride (1 mL) and stirred for 3 h at room temperature. The solvent was evaporated and the residue purified by column chromatography over silica gel (ethyl acetate: hexane 1:4) to give 15 (157 mg, 0.45 mmol, 95%) as a yellow oil. ${[\alpha ]}_D^{23}=-\mathit{23.5}$ (c = 0.60, chloroform). ¹H NMR (600 MHz, CDCl₃) δ 7.49 − 7.42 (m, 2H, SPh), 7.34 − 7.28 (m, 3H, SPh), 5.09 (t, J = 8.9 Hz, 1H, H3), 4.88 (t, J = 8.9 Hz, 1H, H2), 4.70 (d, J = 9.2 Hz, 1H, H1), 4.16 (dd, J = 11.9, 5.2 Hz, 1H, H5-eq), 3.67 − 3.62 (m, 1H, H4), 3.31 (dd, J = 11.9, 10.1 Hz, 1H, H5-ax), 2.08 (s, 6H, COCH₃, COCH₃). ¹³C NMR (150 MHz, CDCl₃) δ 169.8 (CO), 169.5 (CO), 132.9 (SPh), 131.8 (SPh), 129.0 (SPh), 128.4 (SPh), 86.6 (C1), 74.0 (C3), 69.9 (C2), 66.8 (C5), 58.6 (C4), 20.7 (CH₃), 20.6 (CH₃). ESI-HRMS: m/z calcd for C₁₅H₁₇O₅ N₃SNa [M+Na]⁺ 374.0787, found 374.0775.

5.5. 6,3’,6′,2″,5″,3″′,4″′-Hepta-O-benzoyl-1,3,2′,2″′,6″′-penta-N-trifluoroacetyl-4’-O-(2,3-di-O-acetyl-4-azido-4-deoxy-β-D-xylopyranosyl)-paromomycin (17).

A solution of glycosyl acceptor 16 (380 mg, 0.21 mmol), glycosyl donor 15 (88 mg, 0.25 mmol, 1.2-equiv), N-iodosuccinimide (140 mg, 0.43 mmol, 3-equiv) and activated 4 Å molecular sieves in dichloromethane was stirred at room temperature under an argon atmosphere for 3 h, then cooled down to −72 °C. Trifluoromethanesulfonic acid (TfOH, 24 μL, 0.27 mmol, 1.3 eq) was added slowly and the resulting solution was stirred at −72 °C for 5 h. The reaction was quenched by addition of saturated aqueous NaHCO₃, and then diluted with ethyl acetate. The organic phase was washed with saturated aqueous NaHCO₃ (10 mL × 3), aqueous Na₂S₂O₃ (10 mL × 3), brine (10 mL) and dried over Na₂SO₄, concentrated and purified by column chromatography (toluene: acetone 9:1) over silica gel to give product 17 as a white solid (170 mg 0.08 mmol, 35%). ${[\alpha ]}_D^{23}=+\mathit{64.2}$ (c = 1.90, chloroform). ¹H NMR (600 MHz, CD₃OD) δ 8.18 − 8.11 (m, 2H), 8.08 − 7.97 (m, 8H), 7.75 (m, 3H), 7.65 − 7.57 (m, 5H), 7.54 − 7.44 (m, 9H), 7.39 − 7.31 (m, 2H), 7.23 (m 4H), 7.10 − 7.03 (m, 1H), 6.95 (m, 2H), 6.21 (d, J = 4.0 Hz, 1H, H1′), 5.49 (dd, J = 10.9, 8.7 Hz, 1H, H3′), 5.41 (s, 1H, H1″), 5.34 (td, J = 9.9, 4.5 Hz, 1H, H6), 5.24 (t, J = 3.1 Hz, 1H, H3″′), 5.15 (m, 2H, H2″, H1″′), 5.06 (t, J = 2.3 Hz, 1H, H4″′), 4.75 (t, J = 9.2 Hz, 1H, H3^iv), 4.70 − 4.61 (m, 3H, H3″, H5″, H2^iv), 4.54 (dd, J = 11.0, 4.0 Hz, 1H, H2′), 4.48 (d, J = 7.4 Hz, 1H, H1^iv), 4.43 − 4.18 (m, 6H, H1, H3, H5, H6′-eq, H6′-ax, H5″), 4.15 − 3.97 (m, 4H, H4, H4’, H5′, H4″), 3.92 (d, J = 2.8 Hz, 1H, H2″′), 3.48 − 3.32 (m, 2H, H6″′, H4^iv), 3.14 − 3.04 (m, 2H, H6″′, H5^iv-eq), 2.76 (dd, J = 11.9, 10.4 Hz, 1H, H5^iv-ax), 2.23 (q, J = 12.8 Hz, 1H, H2-ax), 2.01 (dt, J = 12.8, 4.5 Hz, 1H, H2-eq), 1.95 (s, 3H, COCH₃), 1.94 (s, 3H, COCH₃). ¹³C NMR (150 MHz, CD₃OD) δ 171.4, 170.9, 167.9, 167.6, 167.3, 167.3, 166.3, 166.0, 165.4, 159.6, 159.4, 159.2, 158.6, 158.4, 134.9, 134.8, 134.6, 134.5, 134.3, 134.2, 133.9, 131.4, 131.1, 130.96, 130.94, 130.9, 130.8, 130.7, 130.6, 130.4, 129.9, 129.8, 129.7, 129.6, 129.5, 129.4, 129.37, 129.3, 129.1, 118.9, 118.7, 118.6, 118.5, 116.0, 110.8 (C1″), 102.8 (C1^iv), 98.5 (C1″′), 96.6 (C1′), 87.2 (C5), 80.6 (C4″), 78.6 (C4′), 77.4 (C2″), 76.8 (C6), 76.7 (C3″), 75.5 (C4), 75.0 (C3^iv), 73.2 (C5″′), 73.1 (C2^iv), 72.7 (C3′), 69.7 (C5′), 69.5 (C3″′), 67.3 (C4″′), 63.8 (C1), (C5^iv), 63.7 (C6′), 63.0 (C5″), 59.6 (C4^iv), 53.0 (C2′), 50.7 (C2″′), 50.0 (C3), 40.0 (C6″′), 31.8 (C2), 20.6 (CH₃), 20.5 (CH₃). ESI-HRMS: m/z calcd for C₉₁H₇₉O₃₁N₈F₁₅Na [M+Na]⁺ 2087.4509, found 2087.4543.

5.6. 4’-O-(4-Amino-4-deoxy-β-D-xylopyranosyl)-paromomycin (6).

To a stirred solution of 17 (55 mg 26.6 μmol) in methanol (1 mL) under an Ar atmosphere at room temperature was added 6-10% magnesium methoxide in methanol (1 mL, 0.56-0.93 mmol). After stirring for 48 h at ambient temperature with monitoring by ESIMS, the reaction mixture was concentrated under reduced pressure and taken up in water (1 mL), stirred and treated with barium hydroxide (92 mg 0.54 mmol) at room temperature. The solution was heated to 80 °C for 24 h with monitoring by ESIMS then was cooled to 0 °C with an ice bath, and treated with dry ice was added portion-wise until pH = 8-9. The compound was taken up in water (2 mL) and treated with Pd(OH)₂ (45 mg, 300%) and stirred under H₂ at 50 psi for 2 h. The reaction mixture was filtered, and the filtrate was treated with acetic acid (10 μL), then was loaded onto a Sephadex C-25 column (100 g, packed in water) and eluted with gradually increasing concentrations of ammonium hydroxide up to 0.12%. The fractions containing the title compound, detected with 0.8% ethanolic ninhydrin, were combined and concentrated. The residue was dissolved in water (1.5 mL) and treated with 2 drops of glacial acetic acid and then lyophilized to give the hexaacetate salt 6 as a white amorphous solid (10 mg, 13.4 μmol, 50 %). ${[\alpha ]}_D^{23}=+\mathit{32.0}$ (c = 0.45, H₂O). ¹H NMR (600 MHz, D₂O) δ 5.62 (d, J = 3.9 Hz, 1H, H1′), 5.24 (d, J = 2.6 Hz, 1H, H1″), 5.16 (d, J = 1.9 Hz, 1H, H1″′), 4.39 (dd, J = 6.7, 4.9 Hz, 1H, H3″), 4.34 (d, J = 7.5 Hz, 1H, H^iv), 4.25 (dd, J = 5.0, 2.5 Hz, 1H, H2″), 4.18 (ddd, J = 6.4, 3.9, 1.5 Hz, 1H, H5″′), 4.10 (t, J = 3.1 Hz, 1H, H3″′), 4.09 − 4.04 (m, 2H, H4″, H5^iv), 3.89 (dd, J = 10.5, 8.8 Hz, 1H, H3′), 3.83 − 3.77 (m, 3H, H5′, H6′, H5″), 3.76 − 3.71 (m, 3H, H4, H5, H6′), 3.69 (dt, J = 3.0, 1.3 Hz, 1H, H4″′), 3.64 (ddd, J = 12.4, 4.7, 1.9 Hz, 1H, H5″), 3.58 (t, J = 9.2 Hz, 1H, H4′), 3.53–3.51 (m, 2H, H6, H3^iv), 3.48 − 3.39 (m, 3H, H5^iv, H1, H2″′), 3.32 − 3.24 (m, 4H, H2′, H6″′-eq, H6″′-ax, H2^iv), 3.23–3.16 (m, 2H, H3, H^iv), 2.24 (dq, J = 12.6, 3.9 Hz, 1H, H2-eq), 1.58 (p, J = 12.4 Hz, 1H, H2-ax). ¹³C NMR (150 MHz, D₂O) δ 181.0, 109.9 (C1″), 103.0 (C1^iv), 95.6 (C1′), 95.4 (C1″′), 84.6 (C5), 81.2 (C4″), 78.8 (C4), 77.4 (C4′), 75.1 (C3″), 73.4 (C2″), 73.1 (C^iv), 72.7 (C2″), 72.5 (C3^iv), 71.7 (C6), 70.2 (C5″′), 67.7 (C3′), 67.6 (C3″′), 67.2 (C4″′), 61.6 (C5^iv), 60.0 (C5″), 59.3 (C6′), 53.8 (C2′), 50.9 (C2″′), 50.8 (C1), 50.0 (C^iv), 48.8 (C3), 40.4 (C6″′), 29.6 (C2). ESI-HRMS: m/z calcd for C₂₈H₅₅O₁₇N₆ [M+Na]⁺ 747.3624, found 747.3630.

5.7. 1,2′,2″′,3,6″′-Pentaazido-2″,3′,3″′,4″′,5″,6,6′-hepta-O-benzyl-1,2′,2″′,3,6″′-pentadeamino-4’-thio-4’-S-(β-D-xylopyranosyl)-paromomycin (20).

To a solution of 1-S-acetyl-2,3,4-tri-O-acetyl-1-thio-β-D-xylopyranose 18 (35.0 mg, 105 μmol) in dichloromethane/methanol (1:4, 500 μL) was added sodium methoxide (5.7 mg, 105 μmol) at −25°C under an argon atmosphere. The mixture was stirred for 20 min and warmed to room temperature. After concentration under reduced pressure, 19 (79.0 mg, 52.0 μmol) and N,N-dimethylformamide (300 μL) were added to the residue in an ice bath. The mixture was stirred for 1 h in an ice bath. Methanol (2 mL) and sodium methoxide (35.7 mg, 659 μmol) were added and the mixture was further stirred for 1 h in an ice bath. The mixture was neutralized with Amberlite IR-120 (H⁺ form) and the resin was filtered off. After concentration of the filtrate under reduced pressure, the residue was purified by flash column chromatography on silica 23 ¹H NMR (600 gel (chloroform:methanol = 30:1) to afford 20 (71.1 mg, 89%). ${[\alpha ]}_D^{\mathit{23}}=+\mathit{63.1}$ (c = 1.12, methanol). MHz, CD₃OD) δ 7.46 − 7.09 (m, 35H: aromatic), 6.25 (d, J = 3.7 Hz, 1H: H1′), 5.59 (d, J = 4.7 Hz, 1H: H1″), 4.96 (d, J = 10.4 Hz, 1H: PhCH₂O-C3′), 4.88 (d, J = 10.9 Hz, 1H: PhCH₂O-C6), 4.84 (d, J = 1.8 Hz, 1H: H1″′), 4.72 − 4.65 (m, 2H: PhCH₂O-C6, PhCH₂O-C3′), 4.60 (s, 2H: PhCH₂O-C6′), 4.58 (d, J = 11.6 Hz, 1H: PhCH₂O-C4″′), 4.53 (d, J = 9.2 Hz, 1H: H1″″), 4.49 − 4.40 (m, 4H: PhCH₂O-C5″, PhCH₂O-C2″, PhCH₂O-C3″′, PhCH₂O-C5″), 4.38 − 4.32 (m, 2H: PhCH₂O-C2″, PhCH₂O-C4″′), 4.27 − 4.20 (m, 2H: H5′, H3″), 4.19 (q, J = 3.4 Hz, 1H: H4″), 4.04 (dd, J = 11.1, 4.5 Hz, 1H: H6′a), 4.01 − 3.96 (m, 2H: H6′b, H2″), 3.87 (t, J = 10.2 Hz, 1H: H3′), 3.85 − 3.74 (m, 5H: H3″′, H5″″eq, H5″′, H5, H5″a), 3.67 (t, J = 9.4 Hz, 1H: H4), 3.58 − 3.50 (m, 2H: H5″b, H6″′a), 3.49 − 3.42 (m, 4H: H3, H1, H4″″), 3.35 − 3.26 (m, 3H: H2″′, H4″′, H3″″), 3.24 − 3.18 (m, 2H: H2″″, H6), 3.09 − 3.06 (m, 1H: H6″′b), 3.06 − 3.02 (m, 2H: H5″″ax, H2′), 2.99 (t, J = 10.9 Hz, 1H: H4′), 2.13 (dt, J = 12.7, 4.6 Hz, 1H: H2_eq), 1.25 (q, J = 12.6 Hz, 1H: H2_ax). ¹³C NMR (150 MHz, CD₃OD) δ 138.5, 138.2, 138.1, 138.1, 137.7, 137.5, 137.4, 128.2, 128.1, 128.1, 128.0, 128.0, 128.0, 127.9, 127.8, 127.8, 127.8, 127.7, 127.7, 127.6, 127.5, 127.4, 127.3, 127.2, 127.2, 127.2, 106.8 (C1″), 98.6 (C1″′), 95.6 (C1′), 86.2 (C1″″), 83.9 (C6), 82.1 (C4″), 82.1 (C5), 81.5 (C2″), 77.6 (C3″″), 76.7 (C3′), 75.9 (C3″), 75.0 (C4), 74.7 (PhCH₂O-C6), 74.2 (C5″′, PhCH₂O-C3′), 73.4 (C2″″), 73.3 (C3″′), 73.0 (PhCH₂O-C5″), 72.8 (PhCH₂O-C2″, PhCH₂O-C6′), 72.5 (C5′), 72.2 (PhCH₂O-C3″′), 72.1 (C4″′), 71.7 (PhCH₂O-C4″′), 70.2 (C5″), 69.7 (C6′), 69.5 (C4″″), 68.8 (C5″″), 63.9 (C2′), 60.4 (C1), 60.2 (C3), 57.2 (C2″′), 50.9 (C6″′), 46.7 (C4′), 31.9 (C2). ESI-HRMS: m/z calcd for C₇₇H₈₅N₁₅O₁₇SNa [M+Na]⁺ 1546.5866, found 1546.5844.

5.8. 4’-Thio-4’-S-(β-D-xylopyranosyl)-paromomycin (10).

To a stirred solution of 20 (68.3 mg, 44.8 μmol) in tetrahydrofuran (68 μL) were added a solution of 1 M trimethylphosphine in tetrahydrofuran (358 μL, 358 μmol) and 0.1 M NaOH (452 μL) at room temperature. After 11 h of stirring at 50 °C, the reaction mixture was concentrated under reduced pressure and the residue was purified by flash column chromatography on silica gel (chloroform:2-propanol:25% aqueous ammonia solution = 5:3:0.2). The resulting material was dissolved in water/methanol/acetic acid (2:1:0.3, v/v/v, 3.3 mL) and Pd(OH)₂/C (20% loading, 108 mg) was added at room temperature. The mixture was stirred for 33 h under H₂ atmosphere (48 psi) at room temperature. The reaction mixture was filtered through a Celite pad and the cake was washed with water. The combined filtrate was concentrated under a current of air. The crude product was purified by CM Sephadex C-25 ion exchange column (stepwise elution with water, 0.1% aqueous ammonia, and 0.5% aqueous ammonia) to give 10 (12.5 mg, 42%). This compound was dissolved in water (1 mL) and acetic acid (30 μL) and lyophilized to afford the corresponding pentaacetate salt 23 ¹H NMR (600 MHz, D₂O) δ 5.66 (d, J = 3.8 Hz, 1H: H1′), 5.22 (17.1 mg, 42%). ${[\alpha ]}_D^{\mathit{23}}=+\mathit{23.7}$ (c = 0.6, water). (d, J = 2.6 Hz, 1H: H1″), 5.14 (d, J = 1.8 Hz, 1H: H1″′), 4.41 (d, J = 9.7 Hz, 1H: H1″″), 4.39 − 4.34 (m, 1H: H3″), 4.22 (dd, J = 5.0, 2.6 Hz, 1H: H2″), 4.16 (t, J = 5.5 Hz, 1H: H5″′), 4.10 − 4.07 (m, 1H: H3″′), 4.06 (d, J = 7.7 Hz, 1H: H4″), 3.94 − 3.74 (m, 6H: H6′a, H5′, H5″″_eq, H6′b, H3′, H5″a), 3.71 − 3.65 (m, 3H: H5, H4, H4″′), 3.63 (dd, J = 12.4, 4.8 Hz, 1H: H5″b), 3.45 − 3.41 (m, 2H: H6, H4″″), 3.32 − 3.10 (m, 8H: H3″″, H6″′a, H2′, H6″′b, H2″″, H5″″_ax), 2.82 (t, J = 10.8 Hz, 1H: H4′), 2.21 (dt, J = 12.9, 3.8 Hz, 1H: H2_eq), 1.76 (s, 15H: CH₃CO₂H), 1.55 (q, J = 12.6 Hz, 1H: H2_ax). ¹³C NMR (150 MHz, D₂O) δ 181.1 (CH₃CO₂H), 110.0 (C1″), 96.1 (C1′), 95.4 (C1″′), 85.5 (C1″″), 84.7 (C5), 81.2 (C4″), 78.9 (C4), 76.9 (C3″″), 75.1 (C3″), 73.9 (C5′), 73.3 (C2″), 72.7 (C6), 72.4 (C2″″), 70.2 (C5″′), 68.9 (C4″″), 68.6 (C5″″), 67.6 (C3″′), 67.2 (C4″′), 66.2 (C3′), 60.9 (C6′), 60.0 (C5″), 55.2 (C2′), 50.8 (C2″′), 50.0 (C3), 49.0 (C1), 47.5 (C4′), 40.4 (C6″′), 29.7 (C2), 23.1 (CH₃CO₂H). ESI-HRMS: m/z calcd for C₂₈H₅₄N₅O₁₇S [M+H]⁺ 755.3412, found 755.3406.

5.9. Cell-free luciferase translation assays.

Cell-free in-vitro translation inhibition assays were performed using luciferase mRNA and bacterial S30 extracts containing either wild-type bacterial or human hybrid ribosomes. In brief, firefly luciferase mRNA was transcribed in vitro using T7 RNA polymerase (Thermo) using a plasmid as template in which the mammalian promoter in pGL4.14 (Promega) has been replaced by theT7 bacteriophage promoter. Test articles in aqueous solution containing 0.3% Tween20 were dispensed into white 96-well plates (Eppendorf) using the TECAN D300e digital dispenser. The test article dispension volume was balanced to a total of 1.5 μL by 0.3% Tween20 in water. The reaction volume was brought to 15 μL by addition of 13.5 μL Translation Master Mix comprised of bacterial S30 extract, 0.2 mM amino acid mix, 6 μg tRNA (Sigma), 0.4 μg hFluc mRNA, 0.3 μL protease inhibitor (cOmplete, EDTA-free, Roche), 12 U RNAse inhibitor (Ribolock, Thermo Scientific), and 6 μL S30 premix without amino acids (Promega). Dispension and mixing of reagents was performed on ice prior to incubating the sealed plates at 37 °C. After 1 h of incubation, the reaction was stopped on ice and 75 μL of luciferase assay reagent (Promega) was added to each well. Luminescence was recorded with a plate reader (BIOTEK FLx800, Witec AG, Littau, Switzerland).

5.10. Antibacterial inhibition assays.

The minimal inhibitory concentrations (MIC) of synthesized compounds were determined by broth microdilution assays according to CLSI reference methodology M07 [52] as described previously [12] and using strains described previously [53]. Clinical bacterial isolates were obtained from the diagnostic laboratories of the Institute of Medical Microbiology, University of Zurich.

5.11. X-ray crystallography.

Crystallographic data (excluding structure factors) for the structures in this paper have been deposited with the Cambridge Crystallographic Data Centre as supplementary publication no. CCDC 2238003. Copies of the data can be obtained, free of charge, on application to CCDC, 12 Union Road, Cambridge CB2 1EZ, UK, (fax: +44-(0)1223–336033 or e-mail: deposit@ccdc.cam.ac.Uk).