Conformational changes in the catalytic region are responsible for heat-induced activation of hyperthermophilic homoserine dehydrogenase

When overexpressed as an immature enzyme in the mesophilic bacterium Escherichia coli, recombinant homoserine dehydrogenase from the hyperthermophilic archaeon Sulfurisphaera tokodaii (StHSD) was markedly activated by heat treatment. Both the apo- and holo-forms of the immature enzyme were successively crystallized, and the two structures were determined. Comparison among the structures of the immature enzyme and previously reported structures of mature enzymes revealed that a conformational change in a flexible part (residues 160–190) of the enzyme, which encloses substrates within the substrate-binding pocket, is smaller in the immature enzyme. The immature enzyme, but not the mature enzyme, formed a complex that included NADP+, despite its absence during crystallization. This indicates that the opening to the substrate-binding pocket in the immature enzyme is not sufficient for substrate-binding, efficient catalytic turnover or release of NADP+. Thus, specific conformational changes within the catalytic region appear to be responsible for heat-induced activation.


Results and discussion
Activation by heat treatment. Recombinant StHSD was overproduced in E. coli grown at 310 K. The activity of the crude extract (2.0 U/mL, 0.18 U/mg) was increased approximately 2.1 times (4.2 U/mL, 4.7 U/mg) by heat-treatment (343 K for 2 h), which precipitated the major proteins from E. coli. The immature StHSD was purified using TOYOPEARL DEAE-650M and Butyl-650M (Tosoh Corp.) columns without heat treatment. From a structural analysis of the purified enzyme, binding of NADP + and 1,4-butanediol (BU) to the enzyme was detected, even though NADP + was not added during the crystallization (Fig. 1). BU was found at the binding site for homoserine (Fig. 1b). This structure was designated the IM/NADP/BU structure. An earlier report showed that NAD(P) + is absent from the M structure formed by heat-treatment at 343 K for 3 h before purification 8 .
The immature enzyme was further purified using a Blue Sepharose column which is an affinity column for purification of enzymes requiring adenyl group-containing molecules, such as NAD(P) + and ATP. Since StHSD binds to a Blue Sepharose column instead of NAD(P) + , the enzyme in the eluant after binding to the column should be the apo-form. The absence of NAD(P) + from the purified enzyme was also confirmed by structural analysis of the immature enzyme. The structure of the immature apoenzyme was designated the IM structure. The mature enzyme showed significantly higher specific activity at alkaline pHs (Fig. 2) at 303 K where both enzymes are stable during the assay (5 min) and showed higher specific activity at all of the tested temperatures (Fig. 3) as compared to those of immature enzyme. V max of the mature enzyme was approximately two times higher than that of the immature enzyme at pH 8.0 and 303 K (Table 1).
Conformational changes. The mature (M and M/NAD/Cys) and immature (IM and IM/NADP/BU) structures were compared to determine the effect of heat-induced maturation on the structures (Fig. 4). The M/NAD/Cys structure has structurally unique residues 22-26, mainly due to their contact with other molecules within the crystal which belongs to space group I4 1 , and peak signals are seen by the comparison (Fig. 4b, 5b) with the other three structures solved in space group P2 1 .
Heat treatment-induced conformational changes were more obvious in a region of residues 181-190 in the apo-form (Fig. 4a) than those observed in the substrate analog-bound forms (Fig. 4b). Although the substrate analogs were distinct in the M/NAD/Cys and IM/NADP/BU structures, they share similar structures (Fig. 4b), except for the residues having peaks at 23 due to crystal packing. This suggests that these substrate analogs induce similar conformational change of StHSD by their binding.
The catalytic region (residues 146-255) changed its conformation upon NAD(P) + binding (Fig. 5). This conformational change induced by ligand binding was smaller in the immature enzyme ( Fig. 5a) than that in the mature enzyme (Fig. 5b). A part (residues 160-190) of the catalytic region showed greater flexibility in its responses to heat treatment and NAD(P) + binding, and was therefore designated the flexible part. Because conformational changes induced by NAD(P) + binding to the flexible part were smaller in the immature enzyme than in the mature enzymes (Fig. 5), heat treatment was able to induce a larger conformational change in that region. The flexible part (residues 160-190) is located at the entrance of the substratebinding pocket (Fig. 6a). In the immature apoenzyme, the substrate-binding pocket assumes a more closed form than in the mature apoenzyme. The more opened form of the substrate-binding pocket in the mature enzyme can accelerate the catalytic turnover that results from its activation. In addition, this impediment for the catalytic turnover of the immature enzyme could stabilize ligand-binding form to show the significantly lower K m values as compared to those of the mature enzyme ( Table 1).
The immature structures tend to have more disordered residues than their mature counter parts (Supplementary Table 1). Increase of folding rate of protein L at higher temperature 12 supports a hypothesis that heat-treatment accelerates proper folding of StHSD. Disordered parts in the IM/NADP/BU structure are found in the flexible part or a part adjacent to the flexible part (Fig. 1A). It is possible that their stabilization leads to the heat-induced activation by facilitating conformational change of the flexible part. On the other hands, the immature enzyme which has more closed form than in the mature apoenzyme may have higher affinity to NAD(P) + (Table 2 and Supplementary  Table 2).
The flexible part. As mentioned above, the flexible part is situated within the catalytic region. Alignment of the stable α-helix structure (residues 194-220) adjacent to the flexible part revealed the displacement of the flexible part (Fig. 6b). The flexible part includes two α-helices and a β-turn harboring Glu186, which encloses the bound NAD(P) + . Alignment of the area that includes the two α-helices and β-turn (residues 160-180) showed that this area is rigid and was therefore designated the rigid part (Fig. 6c). The area adjacent to the region that includes the β-turn structure in the rigid part was designated the lid part (residues 181-190). Upon NAD(P) + binding, the lid part interacts with the rigid part and changes its location substantially to form a hydrogen bond between E186 and R38. The distance between the rigid and lid parts within the IM structure (cyan in Fig. 6c) is larger than within the M structure (represented in pink in Fig. 6c). Both hydrophobic interactions and hydrogen bonds are thought to strengthen the interaction between the lid and rigid parts. The hydrophobic effect increases with increasing temperature, and hydrophobic interactions are known to play major roles in strengthening of thermostability of thermophilic enzymes [13][14][15] . This strengthened interaction may explain why the substrate-binding pocket is more open in the M structure than     the IM structure (Fig. 6a). In addition, the side chain of F195, which is adjacent to the lid part, may interfere with effective opening of the substrate-binding pocket in the immature enzyme. The location of F195 within the immature enzyme was slightly altered by substrate binding, whereas it was markedly altered in the mature enzyme (Fig. 6c).
NADP + binding. Arg38 forms two hydrogen bonds with Glu186 to enclose the bound NAD(P) + within the ligand-bound structures (the IM/NADP/BU and M/NAD/Cys structures) (Fig. 7). These hydrogen bonds appear to stabilize the closed state of StHSD (the IM/NADP/BU structure) and support tight binding of NAD(P) + within the nucleotide-binding pocket. The side chains of Arg38 and Arg39 form hydrogen bonds with the phosphate group at the 2′-position of the ribose moiety of NADP + (Fig. 7), but not NAD + , within the ligand-bound structures. This tight binding of NADP + can restrain turnover, resulting in the inactivation of the NADP + -bound immature enzyme. Both the immature and mature enzymes were competitively inhibited by NADP + with apparent K i of 26 ± 1.7 and 48 ± 1.2 μM, respectively at pH 7.0 at 303 K (Supplementary Table 2). This low apparent K i can explain that the immature enzyme forming a complex with NADP + probably derived from the host E. coli cells. HSD from P. horikoshii is also strongly inhibited by NADP + , and the crystal structure of the enzyme shows the presence of NADP + despite its absence during crystallization as does StHSD 16 . In HSD from Thermus thermophilus, Arg44 and Ala181 respectively correspond to Arg38 and Glu186 in StHSD, but no corresponding hydrogen bond is found 16 . Nor was it found in HSDs from Thermoplasma volcanium (PDB 3JSA: NAD + complex) or Thermoplasma acidophilum (PDB 3ING: NAD + complex), judging from the ligand-bound structures.
Disulfide bond of the structures. The mature enzyme is activated via reductive cleavage of the disulfide bond between the respective cysteine (C303) residues in the C-terminal nucleotide-binding region of each subunit of the homodimeric enzyme 8 . The immature enzyme was less stable than the mature enzyme in the presence of 10 mM DTT or 2-mercaptoethanol.   The immature enzyme decreased its activity approximately half of the initial activity after 12 h incubation at 277 K, though the mature enzyme increased the activity approximately 1.2 times at the same conditions. The immature enzyme probably requires the disulfide bond for the stability due to the inappropriate folding especially in the flexible part. All the structures obtained here had a disulfide bond between the C303 residues in the C-terminal regions of the two subunits. The distance between the two sulfur atoms in both the IM and IM/NADP/BU structures was 2.1 Å, which is similar to that in the oxidized enzyme (the M structure; 2.2 Å) but less than that in the reduced enzyme (2.9 Å) 8 .

Methods
Purification. E. coli BL21 strain was transformed with pSTHSD and recombinant StHSD was overexpressed at 310 K after 3 h induction by 0.1 mM isopropyl β-D-1thiogalactopyranoside before harvesting culture 8 . The crude extract prepared from 12 g (wet weight) of E. coli cells was centrifuged for 20 min at 10,000 g and 277 K. The supernatant was dialyzed against 10 mM Tris-HCl (pH 8.5) and was then applied to a DEAE-TOYOPEARL column. The column was washed with 10 mM Tris-HCl (pH 8.5) containing 50 mM NaCl, and the fractions were eluted with 10 mM Tris-HCl (pH 8.5) containing 75 mM NaCl. The active fractions were collected and applied to a Butyl-TOYOPEARL column in the presence of 40% saturated ammonium sulfate in 10 mM Tris-HCl (pH 8.0). The enzyme was eluted with a gradient of 40 to 0% saturated ammonium sulfate. For purification of the apo-form without NADP + , the purified enzyme was applied to a Blue Sepharose column, and the apoenzyme was eluted with a gradient of 50 mM to 1 M NaCl in 10 mM Tris-HCl (pH 8.5). The purified enzyme (about 14 mg) was dialyzed against 10 mM phosphate buffer (pH 7.0) and concentrated using an Amicon Ultra 10 K filter unit (Millipore). The homogeneity of the final preparation was confirmed with SDS-PAGE.
Enzyme and protein assay. HSD activity was determined in a reaction mixture containing 0.1 M Tris-HCl buffer (pH 8.0), 10 mM L-homoserine, and 10 mM NAD + and the enzyme. Activity was assessed based on the initial velocity of NADH production measured at 340 nm (molar extinction coefficient of 6220 cm −1 M −1 ) during the first 120 s of incubation. The standard assay conditions were carried out at low temperature (303 K) at which the immature and mature enzymes are stable at pHs from 4.5 to 11.0 for 5 min. The immature and mature enzymes were partially inactivated after heat treatment at 353 K for 10 min at some pHs ( Supplementary Fig. 1). Protein concentrations were measured using a Pierce ® BCA protein assay kit (Thermo Scientific Pierce). One unit of enzyme activity was defined as the amount of enzyme that catalyzed the formation of 1 μmol NADH per minute. The types of inhibition were determined using Lineweaver-Burk plots. The V max and K m , values and the apparent K i values were calculated by fitting Michaelis-Menten ( Supplementary Fig. 2) and Morrison ( Supplementary Fig. 3) equations, respectively, to the assay data using Solver in Excel. DMSO. Crystals were grown using the hanging-drop vapor diffusion method with 100 μL of reservoir solution. Diffraction data were collected on beamlines at the Photon Factory, Tsukuba, Japan. All data sets were collected at a wavelength of 1.0000 Å at 95 K. All images were indexed and integrated using the program HKL2000 17 , and the data sets were phased with molecular replacement using the program Phaser 18 in the CCP4 program package. The M structure was used as an initial phasing model for the IM structure, whereas the M/NAD/Cys structure was used for the IM/NADP/BU structure. The models were built using the program COOT 19 and refined using Refmac5 20 . The two subunits in the asymmetric unit were refined without crystallographic symmetry restraints. Both the main chain and side chains were clearly identified in the 2Fo-Fc electron density map, and the final difference Fourier maps contained no significant peaks. The programs RAMPAGE 21 and SFCHECK 22 in the CCP4 package were used for stereochemical analyses of all models, for calculation of the RMSD, and for calculation of the average error using the Luzzati plot. The Ramachandran plot analysis of the IM/ NADP/BU and IM structures show that 96.5 and 95.6%, respectively, of the residues lie within the most favored region, 3.3 and 4.3%, respectively, of the residues within allowed region and no residues with in disallowed region. Data collection and refinement statistics are presented in Table 2. All figures with illustrated structures were prepared using the program CCP4mg 23 and PDBsum 24 . The coordinates of the IM and IM/NADP/BU structures have been deposited in the PDB under entry numbers 7F4B and 7F4C, respectively.
Statistics and reproducibility. The same enzyme was assayed repeatedly more than three times.