Copyright WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001. Supporting Information for Angew. Chem. Int. Ed. Z18552 Click Chemistry in situ: Acetylcholinesterase as a Reaction Vessel for the Selective Assembly of a Femtomolar Inhibitor from an Array of Building Blocks

Warren G. Lewis, Luke G. Green, Flavio Grynszpan, Zoran Radic, Paul R. Carlier, Palmer Taylor, M.G. Finn*, and K. Barry Sharpless*

Supporting Information

CAUTION! All of the compounds described here (and especially the most potent bivalent inhibitors) are potentially neurotoxic. They must be handled with extreme care by trained personnel.

Introductory Material

Known inhibitors of acetylcholinesterase. Active-center ligands tacrine (1), huperzine A (2) and huprine X (3) (Figure S1) bind to the choline binding site.1,2 Inhibitors such as propidium (4)1,3 generally have lower affinity for AChE and bind selectively to the peripheral site, interacting with aromatic residues at the mouth of the gorge. It has also been shown that inhibitors which span the two sites exhibit tighter binding than the individual components; examples include bis-quaternary species like decamethonium (5),1 decidium (6),4 ambenonium (7),5 as well as dimeric tacrine derivatives such as 8,6 and others7 (Figure S1).

1 Z. Radic, P. Taylor, J. Biol. Chem. 2001, 7, 4622-4633. 2 P. Camps, B. Cusack, W. D. Mallender, R. El Achab, J. Morral, D. Muñoz-Torrero, T. L. Rosenberry, Mol. Pharmacol. 2000, 57, 409-417 3 S. Lappi, P. Taylor, Biochemistry 1975, 14, 1989-1997. 4 H. A. Berman, M. Baker, M. McCauley, K. J. Leonard, M. W. Nowak, M. M. Decker, P. Taylor, Mol. Pharmacol. 1987, 31, 610-616. 5 A. S. Hodge, D. R. Humphrey, T. L. Rosenberry, Mol. Pharmacol. 1992, 41, 937-942. 6 Y.-P. Pang, P. Quiram, T. Jelacic, F. Hong, S. Brimijoin, J. Biol. Chem. 1996, 271, 23646- 23649. 7 (a) P. R. Carlier, Y. F. Han, E. S.-H. Chow, C. P.-L. Li, H.-S. Wang, T. X. Lieu, H. S. Wong, Y.-P. Pang, Bioorg. Med. Chem. 1999, 7, 351-357. (b) P. R. Carlier, E. S.-H. Chow, Y. Han, J. Liu, J. El Yazal, Y.-P. Pang, J. Med. Chem. 1999, 42, 4225-4231. (c) P. R. Carlier, D.-M. Du, Y.-F. Han, J. Liu, E. Perola, I. D. Williams, Y.-P. Pang, Angew. Chem. Int. Ed. 2000, 39, 1775-1777.

– S1 – Figure S1. Selected inhibitors of AChE. (a) mouse, (b) human, (c) Torpedo californica, (d) rat brain.

Me Et N H2N NH 2 H N N O Cl N Me NH 2 H2N NH2 N 1 2 3 4 a a b a Kd = 18 nM Kd = 5.2 nM Ki = 26 pM Kd = 1.1 mM

Cl N NMe 3

Me3N N NH

NH O O HN

NH N N NMe3 H 2N NH2 Cl N

5 6 7 8 a c b b Kd = 460 nM Kd = 21 nM Ki = 120 pM Ki = 1.4 nM b d Kd = 250 pM IC50 = 400 pM

Additional Experimental Results Control experiments relative to the AChE-directed cycloaddition of TZ2 and PA6. – Under otherwise identical conditions, mixtures of TZ2 and PA6 in the absence of AChE (buffer alone) or in the presence of 10 mM bovine serum albumin failed to give adduct detectable by either DIOS or LC-MS. – Pre-treatment of AChE with 100 mM tacrine, an active center ligand, followed by incubation with TZ2 and PA6 also failed to give the triazole. – Enzyme exposed to diisopropyl fluorophosphate (DIFP), which is known to inactivate AChE by covalent phosphorylation of the active center serine,8 afforded no detectable cycloadduct after incubation with TZ2 and PA6. When treated with pralidoxime chloride,9 the reactivated enzyme again induced formation of the triazole.

For the last case, 400 mL AChE solution was treated with 2 mL DIFP (Sigma) in a well- ventilated hood. The resulting 27 mM solution of DIFP was left open in the hood. After three hours at room temperature, unreacted and hydrolyzed DIFP was removed by size-exclusion chromatography (filtration through washed BioRad P6 resin), after which only traces of remaining activity were observed by addition of indoxyl acetate. A sample of AChE treated identically except for the omission of DIFP retained its activity.

8 T.L. Rosenberry, Enzymol. Relat. Areas Mol. Biol. 1975, 43, 103-218. 9 D. M. Quinn, Chem. Rev. 1987, 87, 955-979

– S2 – The DIFP-inactivated enzyme regained much of its activity upon treatment with 72 mM pralidoxime chloride (Sigma).1

Preliminary survey of AChE inhibition using other triazoles. Preliminary measurements were made of the relative inhibitory potency of 11 regioisomeric pairs of triazoles derived from simple thermally-promoted cycloadditions of the azides and alkynes used in the screen for in situ assembly. These studies, using an indoxyl acetate assay described below, allow us to group the triazoles into two categories. A set of three compounds (TZ2/PA6, TA2/PZ6, and TA1/PZ7), all having a total of eight methylene units in the chain linking tacrine and phenanthridinium groups and tested as a mixture of regioisomers, was found to be at least 500 times higher in affinity for Electrophorus AChE than tacrine alone. A second set (TZ6/PA2, TZ5/PA3, TZ4/PA4, TZ6/PA3, TZ5/PA2, TZ4/PA2, TZ3/PA2, TZ2/PA2), containing compounds with a total of four to nine methylene spacers, was found to also exceed the inhibitory power of tacrine, but less dramatically than the first set. A more comprehensive analysis is required to discriminate between the members of each class.

Experimental Details Caution! Acetylcholinesterase inhibitors are potentially highly toxic. Acetylcholinesterase Concentrations. Commercially-available Electrophorus AChE was found to contain approximately 10% active enzyme. Its concentration was determined from a quantitative measurement of activity by Ellman assay, using the kcat value of 5 –1 10 4.8x10 min . We also indirectly confirmed this kcat by titrating Electrophorus and mouse enzymes with comp syn-1 and then by titrating mouse enzyme with known concentrations of an organophosphate (echothiophate). All calculations used the titrated value of active center concentration, which is equivalent to the subunit concentration.

Inhibitor asssembly in situ. To ~1mM solutions of Electrophorus AChE (Type V-S, Sigma) in 2 mM ammonium citrate buffer (pH 7.3-7.5), tacrine components (TZ2-TZ6 and TA1- TA3) in water were added and allowed to stand for 90 minutes at room temperature. The corresponding coupling partner was then added and the solutions were thoroughly mixed. The final concentrations were: solid containing AChE, 1.0 mg/mL; tacrine component, 30 mM, phenanthridinium component, 66 mM). Each reaction mixture was allowed to stand at room temperature for a week. At daily intervals, 0.25-0.5 mL drops were deposited on the photopatterned wells of a DIOS chip, allowed to dry, and analyzed by DIOS-MS.

Indoxyl acetate assay. Appropriate concentrations of the inhibitor of interest (ranging from 1-100 nM) were incubated with 6.7 mg/mL acetylcholinesterase in 5 mM HEPES buffer pH 7.0 with 0.01% BSA for 45 minutes at room temperature in a 96-well plate. A solution of indoxyl acetate (Acros) in DMSO (100 mM, 1/100 vol) was then added to each well and thoroughly mixed by pipetting. After 30 minutes, the absorbance at 560

10 (a) D. M. Quinn, Chem. Rev. 1987, 87, 955-979. (b) I.B. Wilson, S. Ginsburg, Biochem et Biophys. Acta. 1955, 18, 168-170.

– S3 – nm was read on a plate reader; results are summarized in Figure S2.

Figure S2. Results of indoxyl acetate assay of inhibitory power of candidate triazoles for Electrophorus AChE, showing those compounds that were found to be more potent than tacrine.

Comparison of heterodimeric inhibitors. Selected tacrine-phenanthridinium adducts were compared for their ability to inhibit the production of indigo in the presence of AChE. A Dixon-type plot of the reciprocal absorbance versus inhibitor concentration shows that TA1/PZ7, TA2/PZ6, and 1 are the most potent inhibitors (Figure S2).

Measurement of cycloaddition background reaction rate. Kinetic measurements were performed for two reactions between TZ2 and PA6 , at 10 mM concentration in each reagent and at 5 mM PA6 and 10 mM TZ2. Both reactions were performed in n-butanol at 18 ˚C. Product formation was quantified by single ion monitoring ([MH]2+ = 331) of 5 mL reaction aliquots injected directly into the LC-MS instrument, using a 30 x 2.1 mm Zorbax SB-C8 column and Zorbax C8 guard column employing a solvent gradient of 30%–40% MeCN in water with 0.05% trifluoroacetic acid (over 10 minutes). A calibration curve of 5 mL injections was performed using a solutions containing 10 mM TZ2 and varying amounts of 1 from 500 mM to 8 mM to convert the observed peak areas to absolute concentrations. The reactions were monitored over two weeks and proceeded to less than 0.5% over that time. Plots of the integrated second-order kinetics equations appropriate to each case (Figure S3) gave the observed rate constants of 1.9 ± 0.4 x 10–5 M–1min–1 and 1.8 ± 1.0 x 10–5 M–1min–1.

– S4 – Figure S3. Determination of background rate of cycloaddition reaction between TZ2 and PA6 in the absence of AChE. (A) 10 mM PA6 + 10 mM TZ2 yielded reaction rate constant k = 1.9 x 10–5 M–1min–1 (r = 0.951). (B) 5 mM PA6 + 10 mM TZ2 yielded reaction rate constant k = 1.8 x 10–5 M–1min–1 (r = 0.887).

A 100.3 100.25

100.2 1 [PY6] 100.15 100.1

100.05

100 0 2000 4000 6000 8000 10000 12000 14000 time (min)

B 0.6949 0.6947

0.6945

[TZ2]0 – [PA6]0 0.6943 ln + 1 [PA6]t 0.6941

0.6939

0.6937

0.6935 0 2000 4000 6000 8000 10000 12000 14000 time (min)

Quantitation of inhibitor produced. Using the calibration curve described above, the peak intensity of 1 from four independent reactions containing ~1 mM AChE revealed the formation of 2.0±1.0 mM of 1.

Determination of AChE - inhibitor association and dissociation rate constants.

The association rate constants (kon) were determined by both, direct measurements of inhibitor binding to AChE and by measurements of time-dependent loss of AChE activity in reaction with inhibitor. The stopped-flow technique was used to measure rates of quenching of intrinsic AChE tryptophan fluorescence upon binding of inhibitor at micromolar concentrations as previously described.11 The time-dependent loss of AChE activity was measured upon mixing AChE with picomolar concentrations of inhibitor in ten-fold excess. The AChE activity in aliquots of the reaction mixture was

11 Z. Radic, P. Taylor, J. Biol. Chem. 2001, 7, 4622-4633.

– S5 – determined by Ellman assay12 at intervals of several minutes to an hour. The second order rate constants of inhibitor association were obtained by linear fit of first order decay rates of either AChE fluorescence or its activity, at several inhibitor concentrations.

The first order dissociation constants (koff) were determined by measurements of the return of AChE activity by Ellman assay upon 5000-fold dilution of 50 – 100 nM concentrations of AChE•1 complex into 250 mg/ml solution of herring sperm DNA (Boehringer) in buffer. The dissociation constant was determined by nonlinear fit of first order increase in enzyme activity up to 70 – 80 % of the AChE control activity in mixture containing no inhibitor. All experiments were performed in at least triplicate with the standard error of determination smaller than 20% of the mean value. The measurements were performed in 0.1 M phosphate buffer pH 7.0 at 22 °C on a SX.18 MV stopped-flow instrument (Applied Photophysics) or Cary 1E UV/VIS spectrophotometer (Varian). Examples of determinations of koff (left) and kon (right) are shown in Figure S4.

Figure S4. Determination of kinetic parameters for interaction of 1 with Electrophorus AChE.

enzyme fluorescence activity (V) (fraction)

time (days) time (ms)

DIOS analysis. Silicon wafers (0.005-0.02 W-cm resistivity, Silicon Quest) were etched for 1 minute at a current density of 5 mA/cm2 under white light illumination (35 mW/cm2) through a 100-spot mask. The surface was then oxidized with ozone for 40 seconds, then immersed in 5% aqueous HF for one minute. An analogous procedure has been previously described.13 Similar chips are now available commercially (see http://www.masscons.com). DIOS-MS was performed on a Perceptive Biosystems MALDI-TOF instrument, with 337-nm laser pulses. After a 25 ns delay, ions are accelerated with a 20 kV potential pulse. Laser energy was routinely increased from sub-threshold levels to above the threshold intensity for appearance of starting material

12 L. G. Ellman, K. D. Courtney, V. J. Andres, R. M. Featherstone, Biochem. Pharmacol. 1961, 7, 88-95. 13 Z. Shen, J. J. Thomas, C. Averbuj, K. M. Broo, M. Engelhard, J. E. Crowell, M. G. Finn, G. Siuzdak, Anal. Chem. 2001, 73, 612-619.

– S6 – ions. The laser intensity employed was somewhat higher than used in conventional MALDI or DIOS analyses due to the presence of large amounts of protein. The laser spot was continuously swept across the sample deposition area to avoid ablation of the underlying material, which contributes to background ions. At the highest laser intensities, triazoles, and to a greater extent, azides, were observed to lose molecular nitrogen. The DIOS-MS spectrum showing the formation of 1 is shown in Figure S5.

Figure S5. DIOS-MS spectrum of the reaction mixture including PA6, TZ2, and AChE. Unreacted TZ2 is detected with low sensitivity under these conditions, as determined by independent measurements.

N

N HN

H N NH N N 2 2 N PA6 counts

N

H2N NH2

1 m/z = 661

m/z

LC-MS analysis. DIOS results were confirmed by the more laborious LC-MS technique (Figure S6), performed on a Hewlett-Packard MSD-1100 instrument; reverse-phase HPLC was accomplished with a 30 x 2.1 mm Zorbax SB-C8 column, preceded by a C18 filter (Phenomenex) and Zorbax SB-C8 guard column. The solvent system consisted of 35% MeCN in water with 0.05% trifluoroacetic acid. For quantitative analysis of 1 (Figure S7), the electrospray ionization conditions were optimized for its doubly charged ion (m/z = 331).

Figure S6. Ion-extraction LC-MS of the reaction mixture including PA6, TZ2, and AChE.

– S7 – Figure S7. Optimized electrospray ionization mass spectrum of a larger-scale reaction mixture containing TZ2, PA6, and AChE.

Regioisomer determination. AChE at 20-fold higher concentration than normal (20 mg solid/mL, 70 mL) was incubated with 30 mM TZ2 for 30 minutes, followed by 66 mM PA6, and allowed to stand for five days. An equal volume of acetonitrile was added, and the resulting suspension was filtered by centrifugation at 6000 g for 30 min through

– S8 – a 10,000 Da molecular weight cutoff filter (Amicon), after washing twice with buffer containing 50% acetonitrile. Aliquots of 30 mL were then analyzed by HPLC (55%

MeOH in water with 1% Me3N/formic acid buffer, 0.7 ml/min) using a 100 x 4.6 mm Cyclobond I 2000 DMP column, preceded by a C18 filter (Phenomenex) and Cyclobond I 2000 DMP guard column, monitoring absorbance at 290 nm. Fractions containing the cycloadduct were collected, desalted using a Walters C18 cartridge eluting with H2O followed by a 9:1 mixture of MeOH and 1M Et3N/H2CO3 buffer (pH 7.5) and re- analyzed by DIOS-MS to confirm the identity of the isolated compound. The dominant peak observed displayed m/z = 661.05 (expected 661.38).

Representative Synthetic Procedures

Caution: Azide-containing compounds, particularly those lower in saturated carbon and oxygen content, are potentially explosive and should be handled accordingly. 1-Azido-ethyl-2-amino-N'-9'-(1',2',3'4,'-tetrahydroacridinyl) hydrochloride (TZ2; Fig. 1). A mixture of 9-chloro-1,2,3,4-tetrahydroacridine14 1.00 g, 4.61 mmol) and ethanolamine (0.56 mL, 9.22 mmol) in pentanol was heated at reflux for 12 h. The product, 1-hydroxyethyl-2-amino-N'-9'-(1',2',3'4,'-tetrahydroacridinyl) hydrochloride, precipitated upon cooling and was isolated as a grey amorphous powder (1.28 g, quant.) by filtration, washing with Et2O. This material was used without further purification: 0.85 g (3.33 mmol) was dissolved in thionyl chloride (5 mL) and heated at 50°C for 30 minutes. Evaporation of the reaction mixture in vacuo was followed by dissolution in a mixture of 5 mL each of MeOH and water. The pH was made neutral with 10% NaOH, and NaN3 (0.43 g, 6.67 mmol) added. The reaction mixture was heated at 90°C for 12 h after which time the cooled reaction was extracted with CH2Cl2 (3 x 20 mL). The combined organic extracts were dried (MgSO4) and concentrated. The residue was purified by flash column chromatography (20% MeOH in CH2Cl2) and then converted to its hydrochloride salt by treatment with methanolic HCl to give TZ2 as a 1 yellow amorphous solid (0.77 g, 73%). H NMR (500 MHz, CDCl3) d=8.18 (2 H, dd, 8.4, 3.7 Hz), 7.58 (1H, t, 8.4 Hz), 7.39 (1H, t, 7.0 Hz), 6.23 (1H, br s), 3.90 (2H, dd, 11.4, 5.9 Hz), 3.72 (2H, t, 5.9 Hz), 3.15 (1H, br s), 2.78 (1H, br s) and 1.86 (4H, br s); 13C NMR (125

MHz, CDCl3) d=154.6, 153.8, 142.2, 130.8, 125.1, 123.9, 123.8, 118.3, 114.8, 51.7, 47.5, 30.9, + 24.9, 22.5 and 21.6; m/z (MALDI-FTMS) 268.1560 (C15H18N5= [MH] ).

3,6-Chloroacetamido-6-phenylphenanthridine. To a mixture of 3,6-diamino-6- phenylphenanthridine15 (3.0 g, 10.5 mmol) and chloroacetic anhydride (3.6 g, 21.1 mmol) in THF (50 mL) was added Et3N (3.6 mL, 26.3 mmol) dropwise with stirring, and the resulting mixture was stirred for 30 min. The reaction was then diluted with CH2Cl2 (200 mL), washed with water and saturated NaHCO3 (50 mL each), dried (MgSO4), and concentrated to afford 3,6-chloroacetamido-6-phenylphenanthridine as a yellow amorphous solid (4.31 g, 94%); 1H NMR (500 MHz, MeOD) d=10.62 (2H, s), 8.78 (1H, d, 8.8 Hz), 8.68 (1H, d, 8.8 Hz), 8.39 (1H, d, 2.2 Hz), 8.30 (1H, d, 1.8 Hz), 8.16 (1H, m dd, 8.8, 1.8 Hz), 7.84 (1H, dd, 8.8, 2.2 Hz), 7.69-7.67 (2h, m), 7.59-7.55 (3H, m), 4.32 (2H, s)

14 P. R. Carlier, Y. F. Han, E. S.-H. Chow, C. P.-L. Li, H.-S. Wang, T. X. Lieu, H. S. Wong, Y.-P. Pang, Bioorg. Med. Chem. 1999, 7, 351-357. 15 T. I. Watkins, J. Chem. Soc. 1952, 3059-3061.

– S9 – and 4.24 (2H, s): 13C NMR (125 MHz, MeOD) d=174.6, 170.3, 152.9, 148.9, 148.3, 147.0, 139.2. 138.6. 138.4, 137.9, 134.1, 133.1, 132.7, 129.4, 129.0, 127.8, 126.4, 124.4, 53.3 and 53.2; + m/z (MALDI-FTMS) 438.0772 (C23H18Cl2N3O2 = [MH] ).

3,6-Chloroacetamido-5-oct-7'-ynyl-6-phenylphenanthridinium triflate. To a partial suspension of 3,6-chloroacetamido-6-phenylphenanthridine (0.4 g, 0.92 mmol) in nitrobenzene (15 mL) and saturated NaHCO3 (0.1 mL) was added oct-7-ynyl-1-triflate (0.7 g, 2.76 mmol) prepared according to the published method.16 The reaction was stirred for 16 h, followed by flash chromatography (gradient from pure CH2Cl2 to 10% MeOH in CH2Cl2) to afford the desired triflate salt as an orange amorphous solid (33 mg, 5%). 1H NMR (500 MHz, MeOD) d=9.21 (1H, d, 1.8 Hz), 9.04 (1H, d, 9.0 Hz), 8.98 (1H, d, 9.2 Hz), 8.43 (1H, dd, 9.2, 2.2 Hz), 8.07-8.04 (2H, m), 7.86-7.80 (3H, m), 7.71-7.70 (2H, m), 4.71 (2H, t, 8.1 Hz), 4.31 (2H, s), 4.12 (2H, s), 2.17 (1H, t, 2.6 Hz), 2.11 (2H, td, 7.0, 2.6 Hz), 2.07 (2H, br s) and 1.46-1.27 (6H, m); 13C NMR (125 MHz, MeOD) d=167.4, 166.9, 164.1, 141.9, 139.6, 134.8, 132.0, 131.9, 130.3, 129.9, 128.5, 126.2, 125.6, 123.9, 123.0, 122.9, 120.8, 108.9, 83.8, 68.7, 55.3, 43.1, 42.9, 28.9, 28.1, 27.6, 26.0 and 17.8; m/z (MALDI- + FTMS) 546.1729 (C31H30Cl2N3O2 = [M] ).

3,6-Diamino-5-oct-7'-ynyl-6-phenylphenanthridinium triflate (PA6; Fig. 1). A mixture of thiourea (50 mg, 658 mmol) and 3,6-chloroacetamido-5-oct-7'-ynyl-6- phenylphenanthridinium triflate (33 mg, 47 mmol) in ethanol (2 mL) was heated at reflux for 9 h. Upon cooling, the reaction was concentrated, redissolved in CH2Cl2 (25 ml) and washed with water (5 ml). The aqueous phase was repeatedly extracted with CH2Cl2 (3 x 10 ml) and the combined organic phase was dried (MgSO4) and concentrated. This afforded the title compound (25 mg, 98%) as a deep red amorphous solid; 1H NMR(500 MHz, MeOD) d=8.56 (1H, d 9.2 Hz), 8.50 (1H, d, 9.2 Hz), 7.81-7.70 (3H, m), 7.61-7.52 (3H, m), 7.36-7.32 (2H, m), 6.45 (1H, d, 2.2 Hz), 4.51 (2H, t, 8.5 Hz), 2.19 (1H, t, 2.6 Hz), 2.10 (2H, td, 7.0, 2.6 Hz), 1.94 (2H, br s), 1.40 (2H, qn, 7.4 Hz) and 1.33-1.25 (4H, m); 13C (125 MHz, MeOD) d=159.1, 151.8, 148.5, 135.3, 132.6, 131.2, 129.7, 128.9, 128.7, 128.4, 125.6, 124.9, 122.7, 120.4, 118.6, 109.3, 98.7, 83.8, 68.7, 53.8, 28.6, 28.2, + 27.7, 25.9 and 17.8; m/z (MALDI-FTMS) 394.2261 (C27H28N3 = [M] )

1-(3',6'-Diamino-5'-hexyl-6'-phenylphenanthridinium triflate)-5-[ethyl-2-amino-N''- 9''-(1'',2'',3''4,''-tetrahydroacridinyl)]-triazole and 1-(3',6'-diamino-5'-hexyl-6'- phenylphenanthridinium triflate)-4-[ethyl-2-amino-N''-9''(-1'',2'',3''4,''- tetrahydroacridinyl)]-triazole (regioisomeric mixture of 1). TZ2 (6.8 mg, 22 mmol) and PA6 (12.0 mg, 22 mmol) were dissolved in MeOH (5 mL). The solvent was removed by evaporation and the residue heated at 80 ˚C in an oven for 7 days. This furnished the adduct (18.8 mg) as a 1:1 mixture, as measured by 1H NMR. The regioisomers could be partially separated by reverse-phase HPLC using a Zorbax Sil column (9.6 x 250 mm, flow rate 7 mL/min, 30% MeCN in H2O with 1% Me3N/formic acid buffer, pH 7.5). The collected fractions were concentrated by rotary evaporation at less than 40°C, and the resulting concentrates desalted using a Walters C18 cartridge eluting with H2O

16 S. A. Ross, M. Pitié, B. Meunier, J. Chem. Soc., Perkin Trans. 1 2000, 571-574.

– S10 – followed by a 9:1 mixture of MeOH and 1M Et3N/H2CO3 buffer (pH 7.5), to afford the products as deep red powders. The syn-isomer (1,5-triazole) elutes first and was isolated cleanly; the anti-isomer was isolated as a 6:1 anti:syn mixture. The structures were assigned by the nOe results shown below. Syn-1: 1H NMR (500 MHz, MeOD) d 8.63 (1H, d, 9.2 Hz), 8.57 (1H, d, 9.2 Hz), 7.89-7.16 (13H, m), 6.42 (1H, d, 2.2 Hz), 4.52 (2H, t, 5.1 Hz), 4.45-4.41 (2H, m), 4.15 (2H, t, 5.5 Hz), 2.91 (2H, t, 5.5 Hz), 2.61 (2H, t, 6.6Hz), 2.34 (2H, t, 7.7 Hz), 1.94-1.74 (6H, m) and 1.40-1.15 (6H, m); m/z (MALDI-FTMS) – 1 661.3780 (C42H45N8 = [M] ). Anti-1: H NMR (500 MHz, MeOD) d 8.62 (1H, d, 9.2 Hz), 8.56 (1H, d, 9.2 Hz), 7.81-7.25 (13 H, m), 6.43 (1H, d, 2.2 Hz), 4.60 (2H, t, 5.5 Hz), 4.38- 4.35 (2H, m), 4.03 (2H, t, 5.9), 2.86 (2H, br s), 2.63 (2H, t, 5.9 Hz), 2.47 (2H, t, 7.3 Hz), 1.91- + 1.70 (6H, m) and 1.40-1.20 (6H, m); m/z (MALDI-FTMS) 661.3780 (C42H45N8 = [M] ).

N

N

HN

HN N H N N N H N 1.7% H 2.5% N H

NH2 N N

H2N NH2

H2N

– S11 –