Synthesis of Norbornane Bisether Antibiotics via Silver-mediated Alkylation

Synthesis of Norbornane Bisether Antibiotics via Silver-mediated Alkylation Shane M. Hickey, Trent D. Ashton, Jonathan M. White, Jian Li, Roger L. Nation, Heidi Y. Yu Alysha G. Elliott, Mark S. Butler, Johnny X. Huang, Matthew A. Cooper, and Frederick M. Pfeffer* Research Centre for Chemistry and Biotechnology, School of Life and Environmental Sciences, Deakin University, Waurn Ponds, Victoria, 3216, Australia, Bio21 Institute, School of Chemistry, University of Melbourne, Parkville, Victoria, 3010, Australia, Drug Delivery, Disposition and Dynamics, Monash Institute of Pharmaceutical Science, Royal Parade, Parkville, Victoria, 3052, Australia and, # Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland, 4072, Australia.

The direct alkylation of norbornane diol 6 is as yet unreported in the literature, which is presumably because the reaction is troublesome.Indeed, it has been shown that in some cases alkylation of alcohols using typical Williamson protocols gives lower than expected yields, or no desired product. 26,27Indeed, in circumstances where a stabilized leaving group can be formed the basic conditions employed can lead to an elimination/fragmentation event. 28,29An example of this competitive elimination was described by Hergueta et al. (Scheme 2); alkoxide 7, derived from a norbornane fused to a quinoxaline, undergoes elimination to give a carbanion (8) which is stabilized by the aromatic heterocyclic. 30g. 1 Examples of functionalized norbornanes.
The use of silver(I) oxide for the transformation of an alcohol to a methyl ether has been known for well over a century. 31The Irvine-Purdie method has been used to achieve (i) monoalkylations; [32][33][34][35][36] (ii) bisalkylation of unhindered diols such as cycohexanediol; 37 and (iii) per-methylation of carbohydrates such as glucose, galactose and fructose. 31,38Inefficient reaction conditions (e.g.multiple additions of the alkylating agent 34 or carrying the reaction out in neat alkylating agent) means the Irvine-Purdie alkylation is oen overlooked in favour of Williamson methodology.To the best of our knowledge, the use of silver(I) oxide for the bisalkylation of eclipsed syn-1,2-diols has not been previously described.
Herein, we report a protocol for the bisalkylation of the sterically hindered 1,2-diol 6 using silver(I) oxide and a small excess (4.6 equiv.) of suitable alkyl halides.Hydrolysis of the esters gave access to compounds of the general structure 5, which were shown to exist in a sterically congested environment around the two alkyl groups by X-ray crystallography.Subsequent functionalization afforded a series of diguanidines (4, Scheme 1), that were tested against a range of Gram-negative and Gram-positive bacteria (Tables 3 and 4).
Williamson methodology was initially trialled using 6, NaH, MeI, at 55 C (Scheme 2) and aer 16 hours the diol had been completely consumed (as determined by TLC analysis).However, examination of the crude product using 1 H NMR spectroscopy showed no signs of the desired bisether 9a.Instead, a complex mixture (inseparable by column chromatography) was produced.Deprotonation with n-BuLi in THF at À78 to 0 C was also trialled but again the reaction was unsuccessful.Given that similar norbornane diols undergo base-induced ring opening to give stabilized anions (Scheme 2), 30 we propose that following deprotonation, a Grob-type fragmentation occurs to give a stabilized enolate (11, Scheme 3). 42he Sakai group have described the reduction of esters to the corresponding ethers using Et 3 SiH in the presence of catalytic InBr 3 . 43Unfortunately, in the current work, aer acetylation of diol 6 to form diacetate 20 (63%, see Experimental section), 41 reaction with InBr 3 failed to produce the bisether product; instead 1 H NMR analysis indicated that a complex mixture of products had formed.
To further test the scope of these reaction conditions a range of benzyl halides were used.Using benzyl bromide (Table 1, entry 2), bisether diacid 12b was attained over two steps (18% using 300 mg of 6).Pleasingly, when the reaction was performed on a larger scale (900 mg of 6), the yield increased to 37%.A Finkelstein approach 46 using NaI did not increase the yield of bisether diacid 12b.The low yields can be rationalized somewhat by the steric bulk introduced as a result of the rst benzylation-the reaction with the second equivalent of benzyl halide is considerably inhibited.Indeed, an appreciable amount of the monoalkylated regioisomer (37%) was isolated along with the desired bisether product 12b.Similarly, using 2methylbenzyl bromide, 12c was accessed in 19% yield (Table 1, entry 3), and a reasonable quantity of a monoalkylated intermediate was also isolated (39%).Despite the modest yields, this protocol provided access to the desired norbornane bisethersa previously inaccessible family of compounds.
Using 3-bromobenzyl bromide and 4-bromobenzyl bromide the synthesis of bisether diacids 12g and 12h (24 and 28% over two steps respectively, Table 1, entries 7 and 8) was carried out in the same fashion.Crystals suitable for X-ray diffraction were obtained for bisether diacid 12h aer recrystallization from EtOH/pet.spirits.The resulting structure contained a unit cell comprised of two conformational isomers (as shown in Fig. 2).The O/O distance (ca.2.6 Å), clearly illustrates the proximity of the two benzyl groups and their non-symmetric orientation (presumably due to steric constraints).
When allyl bromide was used, stirring for 4 days was required to consume all the starting material; subsequent ester hydrolysis gave diacid 12i in a 25% yield over the two steps (Table 1, entry 9).Unfortunately, reactions with 4-methoxybenzyl chloride and 4-nitrobenzyl bromide (Table 1, entries 10 and 11) failed to give any of the desired product with only minimal consumption of starting material taking place (as evidenced by 1 H NMR analysis of the crude reaction mixture).Also, when 1-iodooctane (Table 1, entry 12) was employed alkylation was unsuccessful.
In six of the alkylation reactions (12b, 12d-h) trans-esterication occurred to a small extent (6-12%, as determined using 1 H NMR spectroscopy).This mixed-ester side-product (example 13 from synthesis of 9b, Fig. 3) was difficult to separate from the desired dimethyl ester product using column chromatography.The presumption that the more sterically accessible exo methyl ester was replaced is based on a previous report by Niwayama and co-workers that illustrated how the exo face is less hindered than the endo face of related norbornane diesters. 40Despite the presence of small amounts of mixed-ester  by-products, in each case hydrolysis proceeded smoothly and pure diacids were isolated in every instance.
A balance of both hydrophobicity and hydrophilicity is essential for the antibacterial activity of structural amphiphiles. 49In light of this, the calculated log P (c log P) values (the log of the octanol/water partition coefficient) were determined for bisether diguanidines (16a-i) using http://www.molinspiration.comsoware (Table 2) and these will be discussed in relation to activity in the following section.

Biological evaluation
The antibacterial activity of these compounds was evaluated against a range of Gram-negative and Gram-positive bacteria, including members of the ESKAPE pathogens; 1 rst using the disk diffusion assay to identify active compounds then microbroth dilution assays to determine minimum inhibitory concentrations (MICs).
Bis-methyl ether 16a and allyl ether 16i, did not show any inhibition in disk diffusion studies at 50 mg per disk (Table 3).For the bis-benzyl ether 16b, a noticeable zone of inhibition (ZOI) was observed (11 mm) against Pseudomonas aeruginosa.The inclusion of small substituents to the phenyl rings, such as 2-methyl (16c), 3-uoro (16e) and 4-uoro (16f), led to improved activity as shown by ZOI's of 14-16 mm against P. aeruginosa and Klebsiella pneumoniae (Table 3).Furthermore, appreciable ZOI's (11-15 mm) were also observed against Gram-positive methicillin-resistant Staphylococcus aureus (MRSA).When larger substituents occupied the 4-position of the phenyl ring, inhibition of vancomycin-resistant Enterococcus faecium (VRE) was also seen in addition to the aforementioned strains.The 4-(triuoromethyl)benzyl (16d) and 4-bromobenzyl (16h) analogs showed ZOI of 18 and 16 mm respectively for VRE (Table 3).The 3-bromobenzyl derivative (16g) was active against all pathogens assessed in this study, and was the sole compound to exhibit activity against Acinetobacter baumannii (ZOI ¼ 15 mm).Comparison of ZOI between 16g and 16h indicates that activity against A. baumannii, K. pneumoniae and E. faecium can be inuenced by subtle changes to substituent locations and may have implications in the design of subsequent compounds.
Given the promising results in the disk diffusion screen the substituted benzyl ethers were subjected to a micro-broth dilution assay to quantify MICs (Table 4).When smaller substituents; 2-methyl (16c), 3-uoro (16e) and 4-uoro (16f) displaying no MIC # 32 mg mL À1 against any bacterial strain tested (Table 4).When larger substituents were included on the benzyl rings such as 4-(triuoromethyl)benzyl (16d) MIC values of 32 mg mL À1 against A. baumannii, and 8 mg mL À1 against each of P. aeruginosa, Escherichia coli and MRSA (Table 4) were  observed.The 3-bromo benzyl-substituted analog (16g) showed a reasonable MIC against both Gram-negative P. aeruginosa (32 mg mL À1 ) and Gram-positive MRSA (16 mg mL À1 ) bacterial strains.Furthermore, an MIC of 32 mg mL À1 was observed for the 4-bromo benzyl-substituted analog (16h) against MRSA (Table 4).A correlation between antibacterial activity and c log P was apparent with compounds with higher c log P values (Table 2) showing stronger antibacterial activity.
Given the encouraging MIC values for compounds 16d and 16g against MRSA (Table 4), both compounds were evaluated  ) against all eight bacterial strains tested.Furthermore, activity was observed by 3-bromobenzyl bisether (16g) against all Gram-positive bacterial strains tested which was highlighted by an MIC of 8 mg mL À1 against S. pneumoniae (Table 4).
The cell viability (% survival) against human embryonic kidney cells (HEK293) and hepatocellular carcinoma (HepG2) was determined aer exposure to compounds 16d or 16g at 100 mM for 24 hours (Table 5).The 4-(triuoromethyl)benzyl bisether (16d) exhibited moderate cytotoxicity with 43 and 56% cell survival observed against HEK293 and HepG2 respectively (Table 5).In the case of bis-3-bromobenzyl ether (16g) the cell viability was determined to be 90% and 96% against HEK293 and HepG2 respectively (IC 50 < 100 mM).The cytotoxicity prole for these compounds is acceptable when compared to previously reported antibacterial agents. 50

Conclusions
The previously inaccessible norbornane bisether diacids 12a-i, were successfully prepared using Ag 2 O and a suitable alkyl or benzyl halide as the key step.An X-ray crystal structure of bisether diacid 12h highlighted the sterically crowded environment of the ethers; which presumably hindered the second etherication step and resulted in lower yields.Nevertheless, the protocol presented here provides a viable alternative for the alkylation of congested syn-diols or base-sensitive alcohols where typical Williamson ether synthesis conditions fail.Further functionalization of bisether diacids (12a-i) gave a series of bisguanidines as hydrochloride salts (16a-i).
Several of the compounds (16d, 16g and 16h) displayed antibacterial activity, with MIC values as low as 8 mg mL À1 , against a range of problematic bacterial species including P. aeruginosa, E. coli, S. pneumonia, E. faecalis and several strains of S. aureus.The results presented here reinforce the notion that the activity of cationic antimicrobial peptides can be mimicked by relatively small, structurally rigid amphiphiles.Indeed, when compared to other synthetic scaffolds (such as calixarenes) which are used to generate antibacterial amphiphiles, 9 the low molecular weight of these compounds and their reasonable antibacterial activity make them an attractive class of compounds worthy of further investigation.

General information
All microwave reactions were conducted using a CEM Discover S-Class Explorer 48 Microwave Reactor, operating on a frequency of 50/60 Hz and continuous irradiation power from 0-300 W. All reactions were performed in sealed reaction vessels.All melting points are uncorrected.All 1 H, 13  High resolution mass spectral data was collected on using a QTOF mass spectrometer (LC-1200 series) under the following conditions: gas temperature (300 C), nitrogen drying gas (10.0 L min À1 ), capillary voltage (3500 V), fragmentor (140 V), and nebuliser (45 psi) in a 80% MeCN in H 2 O solvent system.Analyte solutions were prepared in HPLC grade methanol (conc.$1 mg mL À1 ).
All chemicals and solvents were used as received without further purication unless otherwise stated.Column chromatography was performed on silica gel (230-400 mesh).

Crystallography
Intensity data were collected with an CCD diffractometer using Cu-Ka radiation, the temperature during data collection was maintained at 130.0(1) using an Oxford Cryosystems cooling device.The structure was solved by direct methods and difference Fourier synthesis. 53Thermal ellipsoid plots were generated using the program ORTEP-3 (ref.54) integrated within the WINGX 55 suite of programs.Disordered solvent, assumed to be ethanol was removed using the Squeeze procedure. 56sk diffusionzone of inhibition assay A stock solution of 10 mg mL À1 was made for each compound under observation using DMSO as a solvent.Each of these stock solutions was then diluted by a factor of 1 : 2 to bring the concentration to 5 mg mL À1 .The diluted solutions were then lter-sterilized using a 0.2 mm Nylon lter, and 10 mL of the 5 mg mL À1 stock was pipetted onto a blank disk (i.e.50 mg per disk; Oxoid Limited, Hampshire, UK).All bacterial isolates were matched to a 0.5 McFarland standard (in 0.9% NaCl) before they were swabbed onto nutrient agar.The controls used were a 10 mg colistin disk (Oxoid), 10 mL of DMSO and a plate swabbed with saline from the dispenser used.

Minimum inhibitory concentration (MIC) determination
Bacteria were obtained from American Type Culture Collection (ATCC; Manassas, VA, USA) as listed in Table S2 (see ESI †).Bacteria were cultured in Nutrient broth (NB; Bacto Laboratories, catalog no.234000) or Muller-Hinton broth (MHB; Bacto Laboratories, catalog no.211443) at 37 C overnight with shaking ($180 RPM).A sample of each culture was diluted 50fold in fresh MHB and incubated at 37 C for 1.5-3 h with shaking ($180 RPM).Compound stock solutions were prepared as 10 mg mL À1 in DMSO and colistin was dissolved in Milli-Q water at 5.12 mg mL À1 .The compounds, at twice the nal desired concentration, were serially diluted 2-fold across the wells of 96-well plates (Non-Binding Surface, Corning, catalog no.3641).Mid-log phase bacterial cultures (aer 1.5-3 h incubation) were diluted to a nal concentration of 5 Â 10 5 colony forming units (CFU) per mL, and 50 mL was added to each well giving a nal compound concentration range of 32 mg mL À1 to 0.015 mg mL À1 (DMSO # 1%).MICs were determined visually aer 20 h of incubation at 37 C, with the MIC dened as the lowest compound concentration at which no bacterial growth was visible.

Cytotoxicity evaluation
HEK293 (ATCC CRL-1573) and HepG2 (ATCC HB-8065) cells were seeded as 3000 cells per well in a 384-well plate in DMEM medium (GIBCO-Invitrogen #11995-073), in which 10% of FBS was added.Cells were incubated for 24 h at 37 C, 5% CO 2 to allow cells to attach to the plates.A concentration series of compounds was then added into each well.The cells were incubated with the compounds for 24 h at 37 C, 5% CO 2 .Aer the incubation, 10 mM resazurin (dissolved in PBS) was added to each well.The plates were then incubated for 2 h at 37 C, 5% CO 2 .The uorescence intensity was read using Polarstar Omega with 560/590.The data was analysed by Prism soware.Results are presented as the average percentage of control AE SD for each set of duplicate wells using the following equation: Percentage viability ¼ (FITEST À FI Negative /FI UNTREATED À FI Negative ) Â 100.

Fig. 2
Fig. 2 Thermal ellipsoid plot of the two independent molecular confirmations (ellipsoids at 20% probability level) of compound 12h.

Fig. 3
Fig. 3 Representative structure of proposed mixed-ester 13 formed during reaction with BnBr.

Table 1
Synthesis of bisether diacids 12a-i a Reaction was stirred for 4 days.b N.R ¼ no reaction.c Yield calculated over two steps.

Table 2
Formation of HCl salts 16a-i and associated c log P values a Reaction was performed in CHCl 3 .b Product was isolated as an HCl salt.c Calculated using http://www.molinspiration.comsoware.

Table 4
MIC values (mg mL À1 ) a NT ¼ not tested due to lack of activity against the primary Gram-positive strain (MRSA ATCC 43300).bCOL ¼ colistin sulphate.cVAN¼  vancomycin.