Supporting online materials For Zhou et al., LeuT-desipramine structure suggests how antidepressants inhibit human neurotransmitter transporters Materials and methods LeuT overexpression and purification. LeuT was overexpressed using the pbad vector in E. coli BL21(pLysS) cells(1). Typically, 1 liter of LB medium with antibiotic was inoculated with 10 ml overnight culture, and grown at 37 C for 2 hours to OD 600 0.5 0.6. Following induction with arabinose, the cells were grown for 2 additional hours at 25 C. Cells were harvested at OD 600 1.0 1.1 by centrifugation at 7,000 g for 12 mins. Fresh cell pellet was resuspended in 5 ml of TBS buffer (400 mm NaCl, 50 mm Tris ph 8.0, 20% glycerol) containing 0.5 mm PMSF and a protease inhibitor cocktail (Sigma) per gram of cells. This was followed by 3 cycles of French Press at 18,000 psi to break the cells. The suspension was centrifuged at 15,000 g for 20 minutes to remove any remaining unbroken cells, and the membrane was harvested by ultracentrifugation at 100,000 g for 2.5 hours. Following solubilization in the detergent β- dodecylmaltoside (DDM), membrane was incubated with Ni 2+ -NTA resin at 0.5 ml per gram of membrane for 2.5 hours. The resin was washed twice with 10 column volumes of buffer (100 mm NaCl, 50 mm Tris ph 8.0, 20% glycerol, 0.1% DDM), containing 0 and 25 mm imidazole, respectively. The protein was eluted in three steps, each with 2 column volumes of the same buffer, containing 250, 500 and 1
1000 mm imidazole, respectively. For crystallization experiments, the protein was further purified in 1.2% octylglucoside (OG) using size-exclusion chromatography. Binding of [ 3 H]leucine to LeuT in the presence of antidepressant compounds. A scintillation proximity assay (SPA) with copper/scintillant containing beads (2) was used for monitoring substrate binding to purified, Histagged LeuT. Briefly, Cu 2+ chelate YSI Scintillation SPA beads (GE Healthcare) were diluted in buffer containing 50 mm Tris/HCl, ph 7.5, 150 mm NaCl, and 0.05% DDM. For screening, test drugs of 90 μm were incubated for 15 minutes at room temperature with 10 nm LeuT purified in DDM, [ 3 H]leucine (60Ci./mmole, Perkin Elmer; final concentration of 100 nm), and beads in 100 μl of total assay volume per well of a 96-well flexible plate. Nonspecific binding was defined by 10 mm leucine, which equaled omission of protein. In the leucine saturation assays, varying concentrations of [ 3 H]leucine were present ranging from 0.3 nm to 1 μm ( varying hot method). For the determination of the mechanism of drug inhibition, desipramine or nortriptyline (final concentration of 100 μm, close to their IC 50 ), purified LeuT, and beads were preincubated for 5 min prior to a 15-min incubation with additional [ 3 H]leucine (varying hot method as above). Experiments with varying drug concentrations for the determination of IC 50 values also included the 5-min preincubation period, with final concentrations of LeuT and [ 3 H]leucine at 1 and 10 nm, respectively. 3 H radioactivity was measured using a 1450 Microbeta Trilux plate counter (Wallac). 2
Transport assays of LeuT in proteoliposomes. Proteoliposomes were prepared using a modification of the method described by Fann et al. (3) by adding 25 µg of purified LeuT to 1 ml of loading buffer (20 mm HEPES-Tris, ph 7.4, 100 mm potassium gluconate) containing 1.5 % (w/v) OG and 10 mg of sonicated lipid (3:1 egg yolk L-α-phosphatidylethanolamine: soybean L-α-phosphatidylcholine) and stirring on ice for 20 min. A 20-fold dilution with loading buffer at room temperature formed the proteoliposomes. Proteoliposomes were isolated by centrifugation at 180,000 g for 1 hour. The pellet surfaces were washed twice in ice-cold transport assay buffer (20 mm HEPES-Tris, ph 7.4, 100 mm NaCl) prior to resuspension in the same buffer to a protein concentration of 250 µg/ml. For transport assays, LeuT activity was measured by diluting proteoliposomes with a 200-fold excess of assay buffer and initiating transport by addition of [ 3 H]leucine (60 Ci/mmol) to 1.0 nm. Aliquots of 100 µl were taken at various time points covering the range from 0 to 10 mins and the transport reaction was terminated by harvesting the proteoliposomes on 0.22 µm nitrocellulose filters under vacuum, followed by washing with two 5 ml aliquots of ice-cold assay buffer. The filters were incubated overnight in scintillation fluid prior to measuring radioactivity with a liquid scintillation counter. To measure LeuT transport activity in the presence of desipramine, proteoliposomes were incubated with 200 µm of the drug for 1 hour on ice prior to assay. All experiments were performed in triplicate at 23 C. 3
Crystallization and structure determination. Fractions of LeuT purified in OG from the size-exclusion column were concentrated to 6 mg/ml. After incubation in 10 mm desipramine for 30 mins, crystallization drops were set up by hanging-drop vapor diffusion (4) at 20 C in reservoir solution that contained 0.1 M HEPES-NaOH (ph 6.8), 18% PEG 550 MME and 0.2 M NaCl. Crystals were harvested in 25% PEG 550 MME and frozen in liquid nitrogen. X-ray diffraction data were collected from frozen crystals at NSLS beamline X29 at Brookhaven National Laboratory. The diffraction data was initially refined against the LeuT model with the ligands omitted (PDB ID 2A65). The LeuT-desipramine complex model was manually adjusted using O (5) and refined using CNS (6) with R free sets containing 5% of the reflections. Ribbon diagrams and molecular surfaces were generated with Pymol (7), and surface electrostatic calculations were performed with GRASP (8) assuming an ionic strength of 100mM. Homology modeling of hsert, hnet and hdat. Amino acid sequence alignment was carried out following the scheme as described in (9). Homology models were generated from a manually adjusted sequence alignment containing the LeuT, hsert, hnet and hdat sequences using SWISS-MODEL (10) with our LeuT-desipramine structure as the starting model. Despipramine was kept in the same position as in the template. Molecular contacts were analyzed with CONTACT in the CCP4 suite (11). 4
Uptake assays of hsert and hdat in human HEK293 cells in the presence of desipramine. HEK-293 cells were cultured and transiently transfected with the hdat or hsert gene using LipofectAMINE 2000 (Invitrogen, Carlsbad, CA) (12). Cells were grown in DMEM/F12 medium containing 10% bovine calf serum (hdat), or fetal bovine serum (hsert), and 2 mm L-glutamine at 37 o C and 5% CO 2. To measure uptake of tritiated substrate, suspended, intact HEK-293 cells and inhibitor were preincubated at room temperature. Following the addition of tritiated substrate and further incubation, the assay was terminated by rapid filtration through glass fiber filtermats (Wallac) with a Brandel 96-pin harvester (Brandel, Gaithersburg, MD) (13). Radioactivity on filters was estimated by liquid scintillation counting in a 1450 Microbeta Trilux liquid scintillation counter. The conditions for assays with HEK-293 hdat and hsert, respectively, were as follows: Radioactive substrate: [ 3 H]dopamine (55.1 Ci/mmol, Perkin Elmer), and [ 3 H]serotonin (30.0 Ci/mmol, Perkin Elmer); preincubation time with desipramine: 5 and 5 min; and incubation time with desipramine plus radioactive substrate: 4 and 10 min. The buffer was cell culture balanced salts buffer DPBS (Sigma No. D5652) containing 4.15 mm K +, 153 mm Na +, 140 mm Cl -, and 0.1 mm tropolone for inhibition of COMT (14), at ph 7.4. Nonspecific uptake was defined by 100 µm cocaine (hdat) or citalopram (hsert). The significance of observed differences for desipramine inhibition between mutants and the control was analyzed using the Dunnett multiple comparisons test and Student t-test. 5
Fig. S1. Screening for binding of various antidepressants and other compounds to LeuT using the SPA assay. Bars show [ 3 H]leucine binding in the presence of amino acids (0.8 mm) and other compounds (90 μm) as percent of binding with vehicle. Data are shown as mean ± SEM (horizontal bars, N = 3). 6
Fig. S2. LeuT transport activity measured in the absence and presence of desipramine. Leucine transport by LeuT in reconstituted proteoliposomes was completely inhibited by desipramine at a concentration of 200 µm. Data are shown as mean ± SEM (vertical bars, N = 3). 7
Fig. S3. Structure viewed of the LeuT-desipramine complex from the extracellular space. An F o -F c map contoured at 3 σ (blue, desipramine) is superimposed on the structural model. The EL4 hairpin is colored green, and the rest of the protein pink. 8
Fig. S4. Schematic drawing showing the proposed inhibition mechanism of LeuT transport activity by desipramine. 9
Fig. S5. Amino acid sequence alignment of LeuT, hsert, hnet, hdat and C. elegans and Drosophila DAT proteins. Sequences for two regions that are responsible for differences in desipramine-binding affinity are shown, the EL4 loop and transmembrane TM10. The dots and square indicate the positions where gain-of-function with respect to desipramine binding was observed for hdat and hsert, respectively, and the circle indicates the position of the control experiment. 10
Table S1. LeuT-desipramine complex structure determination and refinement Crystal Parameters: Space group Cell constants at 100 K C2 88.41Å, 86.76Å, 81.19Å (90, 95.72, 90 ) Data quality: Resolution 40.0 2.9 Å R sym 5.6% (22.0)* Mean redundancy 6.1 Completeness 94.4% (74.7) Mean I/σ(I) 19.1 (3.1) Residuals (F 2σ f ): R free 22.1% R work 19.9% Model contents: Protein residues 509 Ligands 1 Leu, 2Na +, 1Cl -, 1 desipramine, 4 β-og Water molecules 15 * Note: the number for the highest resolution shell is given in parenthesis. 11
Table S2. LeuT-desipramine intermolecular contacts LeuT residue Atom Distance (Å) Desipramine atom* R30 CG 3.56 C8 CG 3.63 C7 NE 3.44 C9 NE 3.62 C10 CZ 3.41 C9 CZ 3.23 C10 NH1 3.76 C14 NH1 3.48 C1 NH1 3.33 C10 NH2 3.42 C9 NH2 3.42 C10 Q34 NE2 2.92 C8 3.49 C9 F253 CZ 3.74 C2 A319 CB 3.78 C17 3.50 C6 3.49 C12 3.51 C7 F320 CE 3.71 C6 L400 CD2 3.63 C15 D401 OD1 2.76 N2 OD2 3.68 N2 * Note: Desipramine atoms are named according to (15). 12
Table S3. Potency of desipramine in inhibiting [ 3 H]dopamine and [ 3 H]serotonin uptake by human DAT and SERT constructs (N = 3-5) Construct IC 50 (μm) hdat WT 47.9 ± 7.8 P387A 50.5 ± 9.3 I390V 17.7 ± 3.8 * F472L 9.9 ± 2.3 * hsert WT 0.108 ± 0.015 K490T 0.0524 ± 0.0034 * * Note: P < 0.05 (compared with corresponding WT, one-way analysis of variance followed by Dunnett multiple comparisons test for hdat, and Student t-test for hsert) 13
Supporting references 1. D. N. Wang et al., Biochim. Biophys. Acta 1610, 23 (2003). 2. M. Quick, J. Javitch, Proc. Natl. Acad. Sci. U. S. A. 104, 3603 (2007). 3. M. C. Fann, A. Busch, P. C. Maloney, J. Bacteriol. 185, 3863 (2003). 4. A. Yamashita, S. K. Singh, T. Kawate, Y. Jin, E. Gouaux, Nature 437, 215 (2005). 5. T. A. Jones, J. Y. Zou, S. W. Cowan, Kjeldgaard, Acta Crystallogr. A 47, 110 (1991). 6. A. T. Brünger et al., Acta Crystallogr. D 54, 905 (1998). 7. W. L. DeLano, The PyMOL User's Manual (DeLano Scientific, San Carlos, CA, 2002), pp. 1. 8. A. Nicholls, GRASP: Graphic Representation Analysis Surface Properties (Columbia University, New York, 1992), pp. 1. 9. T. Beuming, L. Shi, J. A. Javitch, H. Weinstein, Mol. Pharmacol. 70, 1630 (2006). 10. T. Schwede, J. Kopp, N. Guex, M. C. Peitsch, Nucleic Acids. Res. 31, 3381 (2003). 11. Collaborative-Computational-Project-Number-4, Acta Crystallogr. D 50, 760 (1994). 12. N. Chen, M. E. Reith, J. Neurochem. 101, 377 (2007). 13. N. Chen, J. Zhen, M. E. Reith, J. Neurochem. 89, 853 (2004). 14
14. N. Chen, J. Rickey, J. L. Berfield, M. E. Reith, J. Biol. Chem. 279, 5508 (2004). 15. M. L. Post, O. Kennard, A. S. Horn, Acta Crystallogr. B 31, 1008 (1975). 15