Expression and purification of Dat
A truncated Dat, Dat21 − 230, was used to remove disorder region in N-terminus and C-terminus (Cheng et al., 2012). This truncated sequence is consistent with the published crystal structures that exhibited a similar function and secondary structure to that of full-length Dat (Cheng et al., 2012). Dat21 − 230 was cloned into a pET28a vector with an N-terminal His-tag, a TEV cleavage site, and kanamycin resistance and was expressed in Escherichia coli BL21(DE3) cells. Cells were grown in M9 medium containing 15N-NH4Cl and 13C-glucose at 37°C until the OD600 reached 0.7, and then protein expression was induced by the addition of 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) at 25°C for 16 hours. Cells were harvested by centrifugation at 6,000×g at 4°C for 30 min, resuspended in buffer A (50 mM Tris − HCl, 100 mM NaCl, 14 mM β-mercaptoethanol, pH 8.0) and lysed by sonication. Cell debris were removed by centrifugation at 16,000×g and 4°C for 25 min. The 0.22 µm-filtered lysate was loaded on a Ni2+ column (GE Healthcare), washed, and eluted with buffer A containing 20 mM and 500 mM imidazole, respectively. The eluent containing Dat21 − 230 was dialyzed against buffer A and then was treated with TEV protease to remove the N-terminal His-tag. After cleavage, Dat was further purified with a 1 ml HiTrap Q XL column (GE Healthcare) with a gradient concentration of NaCl from 0.1 to 1 M in buffer A followed purification on a Superdex 75 column (GE Healthcare) with buffer B (50 mM Tris − HCl, 100 mM NaCl, 1 mM DTT, and 1 mM EDTA, pH 7.0). Purified Dat was concentrated to 24 mg/ml, flash-frozen in liquid nitrogen, lyophilized, and stored at -80°C.
NMR spectroscopy
The lyophilized 15N, 13C-labeled protein was dissolved in H2O/D2O, and the final buffer conditions were 50 mM Tris-HCl, 100 mM NaCl, 1 mM DTT, and 1 mM EDTA in 90% H2O/10% D2O containing 0.2 mM DSS (2,2-dimethyl-2-silapentane-5-sulfonate). All NMR experiments were performed with 1 mM Dat at 20°C for apo-Dat and 25°C for the Dat/Ac-CoA binary complex (in a ratio of 1:4) and data were collected on an Avance III HD 850-MHz NMR spectrometer (Bruker, Germany) equipped with a 1H/13C/15N cryoprobe. Non-uniform sampled (NUS) 3D-heteronuclear NMR experiments, HNCO, HNCACO, HNCA, HNCOCA, HNCB, and HNCACB were acquired for the sequential assignments of 1H, 15N, and 13C backbone chemical shifts. Assignments were confirmed and improved by selective unlabeling of different types of amino acids. Amino acid selective unlabeling can provide the exact amino acid types; hence, it can accelerate to deal with the connections of each residue and rule out the alternative possibility of a connection. Spectra were processed by NMRpipe (Delaglio et al., 1995) and analyzed with NMRFAM-SPARKY (Lee et al., 2015). Chemical shift differences, Δδ, were calculated according to the equation below, where α was 0.14 for most residues but 0.2 for glycine (Williamson, 2013):
$$\:{\Delta\:}{\delta\:}=\sqrt{\frac{1}{2}\:[{{\delta\:}}_{H}^{2}+\left(\alpha\:\bullet\:{{\delta\:}}_{N}^{2}\right)]}\:$$
Assignments and data deposition
The 15N-1H HSQC spectra of apo-Dat and Dat/Ac-CoA are shown in Fig. 1. Out of the 197 non-proline residues, the amide signals of 184 residues (93.4%) and 192 residues (97.5%) were assigned in apo-Dat and in Dat/Ac-CoA, respectively, with complete or partial C’, Cα, and Cβ resonances. We also confirmed the assignments by amino acid selective unlabeling. The missing assignments in apo-Dat were N50, K59, Y75, I85, M93, G154, G156, I157, H184, F195, Q209, G210, and E211, in which G154, G156, and I157 are in the P-loop sequence (R153-A158). The P-loop, which is the sequence Gln/Arg-x-x-Gly-x-Gly/Ala, is a conserved signature in GNAT proteins that is responsible for Ac-CoA binding (Ud-Din et al., 2016). While, unassigned I85 is in the loop before β3, which has two backbone hydrogen bonds with A77 (on β2) and D149 (on the loop before the P-loop sequence). Another set of missing assignments of the consecutive residues Q209, G210, and E211 was located at the loop connecting α8 and β7. Unassigned N50 is at the α1 terminus, which provides two hydrogen bonds to the hydroxyl groups of dopamine and was suggested to be responsible for the binding of dopamine. In contrast, most of these unassigned residues could be assigned in the Dat/Ac-CoA binary complex, and the missing resonances in Dat/Ac-CoA were S182, H184, E191, K192, and K215. Most of these residues are located in α7 (184–192), except K215, which is located at the loop between α8 and α9, and S182, which is located at the loop adjacent α7 and is one of the catalytic residues (Cheng et al., 2012). K192 can form two salt bridges to the 3’-adenosine phosphate and pyrophosphate groups of Ac-CoA. The missing resonances of the P-loop in the free state that are shown in the Ac-CoA-bound state revealed that the binding of Ac-CoA reduced the fluctuation of the P-loop region. The loss of the resonance of S182 in the Ac-CoA-bound state implied a conformational exchange on the millisecond timescale, which might be related to catalysis, as the turnover rate of Dat is 17.6 s− 1 for dopamine (Dempsey, Jeffries, Bond, et al., 2014).
Chemical shift predicted secondary structure using TALOS-N (Shen & Bax, 2013) showed that apo-Dat was similar to Dat/Ac-CoA, and both were consistent with the secondary structures found in the crystal structures (PDB code: 3V8I and 3TE4 for apo-Dat and Dat/Ac-CoA, respectively), except for a new β-strand at the C-terminus of Dat/Ac-CoA (Fig. 2). The short α5 (near the P-loop) and α9 defined in the crystal structures were predicted as loops in solution states. The predicted order parameters (S2 values) calculated from the RCI values showed that all the flexible regions were consistent with the loop regions of the crystal structures (Fig. 2). A significant flexible loop (N79-I85) connecting β2 and β3 was observed and presented conformational heterogeneity in both the apo-Dat and the Dat/Ac-CoA complex. In the Dat/Ac-CoA complex, E83 and I84 showed two conformers with a relative ratio of intensity of ~ 80:20. Another resonance near the P-loop sequence, N151, presented the same proportion of intensities of two signals. Among these three resonances, E83 showed much differences in the two conformations because relatively large differences in 1H-15N chemical shift correlations were found (~ 0.2 ppm of 1H resonance) (Fig. 1). Similar conformational heterogeneity was also found in apo-Dat, in which a minor conformer of I84 was found (relative ratio of ~ 85:15). Additionally, two hydrogen bonds formed within the side chains of the residues E83-N79-Y152 in the crystal structures, which connect the flexible loop (N79-I85) and the loop adjacent to the P-loop sequence. It seems that conformational heterogeneity simultaneously exists in the two loops but is independent of the binding of Ac-CoA because they exist in both the free and bound states of Dat.
The similarity in the predicted secondary structures for the apo-Dat and Dat/Ac-CoA binary complex revealed no major structural changes upon binding. However, the chemical shift differences indicated significant changes in α1, β4-P-loop-α6 and α7 (Fig. 3), suggesting conformational changes between these two forms. According to the crystal structure, the binding site of Ac-CoA was major at β4, P-loop, β5, and α7, in which several hydrogen bonds were formed by L146-A158 (on the β4-P-loop) and two salt bridges were formed by K192 (on α7). An unexpected and significant change was found in the α1 region, which was greater than 2 standard deviations of the mean of the chemical shift differences, indicating the important conformational regulation of Dat. Consistent with our results, F42-T51 on α1 showed backbone hydrogen bonds changes between the apo-Dat and the Dat/Ac-CoA complex in crystal structures (Supplementary Table 1). The changes in α1 may be induced by Ac-CoA, which switches the structure from the open to the closed form in the β4-α6 region and creates an interhelix interaction between D46 (on α1) and R153 (on the P-loop) (Wu et al., 2020). This result was also similar to that found for AANAT3 from Bombyx mori, in which the change in α1 was important for further formation of the substrate binding site (Aboalroub, Bachman, et al., 2017). Because F42-D46 is conserved among insect AANATs, changes in the conformation of the α1 region in the Ac-CoA-bound state could be an important signature for substrate binding and selection.
The 1HN, 15N, 13C', 13Cα, and 13Cβ chemical shifts for the free and Ac-CoA-bound forms of Dat have been deposited in the BioMagResBank (http://www.bmrb.wisc.edu/) under accession numbers 50445 and 28139, respectively.