Abstract
Human cytosolic NADP-IDH (IDH1) has recently been found to be involved in tumorigenesis. Notably, the tumor-derived IDH1 mutations identified so far mainly occur at Arg132, and mutation R132H is the most prevalent one. This mutation impairs the oxidative IDH activity of the enzyme, but renders a new reduction function of converting α-ketoglutarate (αKG) to 2-hydroxyglutarate. Here, we report the structures of the R132H mutant IDH1 with and without isocitrate (ICT) bound. The structural data together with mutagenesis and biochemical data reveal a previously undefined initial ICT-binding state and demonstrate that IDH activity requires a conformational change to a closed pre-transition state. Arg132 plays multiple functional roles in the catalytic reaction; in particular, the R132H mutation hinders the conformational changes from the initial ICT-binding state to the pre-transition state, leading to the impairment of the IDH activity. Our results describe for the first time that there is an intermediate conformation that corresponds to an initial ICT-binding state and that the R132H mutation can trap the enzyme in this conformation, therefore shedding light on the molecular mechanism of the “off switch” of the potentially tumor-suppressive IDH activity. Furthermore, we proved the necessity of Tyr139 for the gained αKG reduction activity and propose that Tyr139 may play a vital role by compensating the increased negative charge on the C2 atom of αKG during the transfer of a hydride anion from NADPH to αKG, which provides new insights into the mechanism of the “on switch” of the hypothetically oncogenic reduction activity of IDH1 by this mutation.
Similar content being viewed by others
Introduction
Isocitrate dehydrogenases (IDHs) are a family of enzymes that catalyze the oxidative decarboxylation of isocitrate (ICT) into α-ketoglutarate (αKG) using either NADP or NAD as a cofactor. Human cells express three distinct IDH enzymes: cytosolic NADP-dependent IDH1, mitochondrial NADP-dependent IDH2, and mitochondrial NAD-dependent IDH3. Recently, mutations in IDH1 have been correlated with certain brain tumors, with the prevalence of 54%-94% in grade II-III astrocytomas, oligodendrogliomas, and oligoastrocytomas 1, 2, 3, 4, 5, 6, 50%-88% in secondary glioblastoma multiforme (GBM) 1, 2, 3, 4, 5, 6, 7, and 33%-50% in adult primitive neuroectodermal tumor 1, 7. GBM is the most aggressive brain tumor with almost invariably fatal prognosis, and the association of IDH1 mutation with the proneural subtype of GBM has important implications for future diagnosis and therapies for patients with GBM 8. Besides brain tumors, IDH1 mutations have also been detected in other cancers including acute myeloid leukemia 9, B-acute lymphoblastic leukemia 10, colorectal cancer 11, and prostate cancer 10, with low frequencies. The closely related IDH2 has also been documented to be mutated in several types of tumors albeit with low frequencies (1%-7%) 3, 6.
It is striking to note that all the IDH mutations identified so far occur at Arg132 of IDH1 or the equivalent residue Arg172 of IDH2 heterozygously 1, 2, 3, 4, 5, 6, 7. Among the identified IDH1 mutations, R132H is the most prevalent one in gliomas 1, 2, 3, 4, 5, 6 constituting up to 92.7% of the total mutations, while R132C and R132S were found with a percentage of 3.6% and 1.8%, respectively 1. Recently, two mechanisms by which the Arg132 mutations contribute to tumorigenesis have been proposed 12, 13. One hypothesis is that IDH1 may function as a tumor suppressor, and the impairment of the IDH activity by the Arg132 mutations plays a tumorigenic role in part through activation of the HIF-1 pathway due to decreased αKG production 12. The other is that the R132H mutant gains a new activity of catalyzing NADPH-dependent reduction of αKG to 2-hydroxyglutarate (2HG), which plays a role in the development and progression of malignant brain tumors 13. The two possibilities are not mutually exclusive, and hence the R132 mutation might contribute to the tumorigenesis process by both “off switch” of the ICT dehydrogenase activity and “on switch” of the αKG reduction activity. The molecular mechanisms underlying the “off-on switch” of the activities, however, are still unclear.
Previously, we have determined the crystal structures of the wild-type dimeric IDH1 (IDH1R/R) in complexes with NADP and with NADP, ICT, and Ca2+, and showed that the assembly of the active site is a complicated process involving concerted inter-domain motion and conformational changes of two subunits in a cooperative manner and that the enzyme may employ a self-regulatory mechanism to regulate its activity 14. In the IDH1R/R-NADP-ICT-Ca quaternary complex, Arg132, which is strictly conserved in all NADP-dependent IDHs (NADP-IDHs) 15, interacts with ICT 14, providing an explanation for the substantially decreased Km (ICT) values of the tumor-derived mutants 12. Besides the decreased Km, these mutants also show decreased Vmax 12 or kcat 13 for ICT dehydrogenation and more surprisingly acquire a new function of αKG reduction 13, leading us to speculate that Arg132 may play other functional roles in addition to ICT binding in the catalysis. To further dissect the functional role(s) of Arg132 in the catalytic reaction and to explore the mechanistic role of the R132H mutation in tumorigenesis, we carried out structural and functional analyses of the R132H mutant IDH1. Given that the maternally and paternally derived copies of each gene on autosomes are simultaneously expressed at comparable levels 16, the heterozygous R132H allele of IDH1 may lead to three forms of IDH1 proteins in vivo, namely IDH1R/R, heterodimeric mutant IDH1 (IDH1R/H), and homodimeric mutant IDH1 (IDH1H/H). We report here the crystal structures of IDH1H/H in complexes with NADP and with NADP and ICT, and IDH1R/H in complex with NADP and ICT, respectively. Furthermore, mutagenesis, kinetic, and thermodynamic studies were performed to explore the catalytic mechanisms of the ICT dehydrogenation reaction and the αKG reduction reaction. The structural and biochemical data together not only reveal the molecular basis for the impairment of the IDH activity by the tumor-associated R132H mutation, but also provide insights into the molecular mechanism underlying the gained αKG reduction activity of the mutant IDH1.
Results and Discussion
Overall structures of the R132H mutant complexes
The structures of IDH1H/H in complexes with NADP and with NADP and ICT were determined at 3.4 Å and 2.8 Å resolutions, respectively. The structure of IDH1R/H in complex with NADP and ICT was determined at 3.2 Å resolution. All the three structures belong to space group P43212 with similar cell parameters and contain two IDH1 subunits forming a dimer in the asymmetric unit (Table 1). Each IDH1 subunit consists of three domains: a large domain (residues 1-103 and 286-414), a small domain (residues 104-136 and 186-285), and a clasp domain (residues 137-185; Figure 1A). In the structure of the IDH1H/H-NADP complex, there is an NADP bound at the NADP-binding site of each subunit even though no NADP was added in the crystallization solution, indicating that similar to the wild-type IDH1 14, the mutant IDH1 was co-purified with a tightly bound NADP from the expression system. In the structures of both IDH1H/H-NADP-ICT and IDH1R/H-NADP-ICT complexes, in each subunit, there is an NADP bound at the NADP-binding site and an ICT bound at a site composed of NADP and residues Thr77, Ser94, and Arg100 of the large domain (Figure 1A and 1B) which is distinct from the ICT-binding site in the IDH1R/R-NADP-ICT-Ca complex 14 (see discussion later). In all the mutant IDH1 complexes, two key structural elements encompassing residues 132-141 and residues 271-286 (designated as seg1 and seg2, respectively) exhibit great flexibility with the majority of them being disordered (Supplementary information, Table S1; see discussion later). In the IDH1R/H-NADP-ICT complex, residue 132 cannot be distinguished as Arg or His, which may be due to the medium resolution of the structure or reflect a random distribution of the wild-type and mutant IDH1 in both subunits.
The R132H mutant complexes assume asymmetric open conformations
Previously, we have shown that in the IDH1R/R-NADP complex, the two subunits of the dimeric IDH1 adopt different conformations: subunit A assumes an open conformation, while subunit B adopts a quasi-open conformation 14. Further binding of ICT induces substantial conformational changes of the large domain towards the small and clasp domains, leading to the formation of a closed conformation of both subunits in the IDH1R/R-NADP-ICT-Ca complex 14. Structural analysis of the three mutant IDH1 complexes shows that their overall structures are similar (with an RMSD of 0.5–1.1 Å for 776 Cα atoms), and intriguingly, the two subunits in each complex are asymmetric (with an RMSD of 1.2–1.7 Å for 388 Cα atoms). Superposition of the corresponding subunits among the IDH1 structures shows that in all the mutant IDH1 complexes, subunit A takes an open conformation and subunit B a quasi-open conformation as in the IDH1R/R-NADP complex (Figure 1C) rather than a closed conformation as in the IDH1R/R-NADP-ICT-Ca complex (Figure 1D). Correspondingly, the width of the active site entrance (19.6–20.7 Å in subunit A and 17.5–18.8 Å in subunit B, defined as the distance between residue 76 of one subunit and residue 250 of the adjacent subunit) in the mutant IDH1 complexes is also comparable to that in the IDH1R/R-NADP complex (21.2 Å and 18.7 Å, respectively). For simplicity, in the following discussion, we will use the structure of the IDH1H/H-NADP-ICT complex as a representative of the mutant IDH1 complexes owing to its relatively higher resolution.
The R132H mutation renders an increased flexibility of two key structural segments
It is intriguing to observe that two regions, seg1 and seg2, are largely disordered in all three mutant IDH1 complexes. In the wild-type IDH1 structures, seg1 forms a hinge region that dictates the conformational changes of the large domain induced by ICT binding, and seg2 forms a “regulatory segment” that undergoes substantial conformational change during the catalytic reaction and was proposed to regulate the enzymatic activity 14. Specifically, seg1 adopts a coil structure in both subunits of the IDH1R/R-NADP complex and a one-turn α-helix in the IDH1R/R-NADP-ICT-Ca complex similar to that observed in the other eukaryotic NADP-IDH structures 17, 18. Seg2 takes a completely unwound loop structure in subunit A and a partially unwound one-turn α-helix structure in subunit B in the IDH1R/R-NADP complex, while it adopts a long α-helix structure in the IDH1R/R-NADP-ICT-Ca complex similar to that observed in all other NADP-IDH structures 17, 18, 19, 20, 21.
Detailed analyses of the wild-type IDH1 structures indicate that these two segments have extensive interactions with each other and Arg132 plays an important role in the interaction network (Figure 2A and 2B, and Supplementary information, Table S2). In subunit A of the IDH1R/R-NADP complex, the Nη1 atom of Arg132A (residues and structural elements of subunits A and B are denoted with superscripted letters A and B, respectively) imposes a salt bridge with the Oδ2 atom of Asp275A (3.0 Å), while the O atom of Asp275A forms a hydrogen bond with the N atom of Tyr135A (2.9 Å) and the O atom of Tyr135A interacts with the N atom of Gln277A (2.7 Å; Figure 2A left panel and Supplementary information, Table S2). In subunit B, the Nη1 atom of Arg132B forms a hydrogen bond with the Oδ2 atom of Asn271B (3.3 Å) and its Nη2 atom forms a salt bridge with the Oδ2 atom of Asp279B (3.1 Å; Figure 2A right panel and Supplementary information, Table S2). At the dimer interface, hydrogen bonds form between the Oη atom of Tyr139B and the Oη atom of Tyr272A (3.6 Å), between the Nδ2 atom of Asn271A and the Oδ2 atom of Asp275B (2.8 Å), and between the N atom of Tyr272A and the Oδ2 atom of Asp273B (3.4 Å; Figure 2B and Supplementary information, Table S2). Moreover, the aromatic rings of Tyr272A, Tyr135B, Tyr139B, and Tyr272B form a hydrophobic cluster, further strengthening the interactions of the two regions (Figure 2B). In the IDH1R/R-NADP-ICT-Ca complex, the interactions between the two segments are even more intense (Supplementary information, Table S2). The Nε atom of Arg132A forms two hydrogen bonds with the O and Oδ1 atoms of Asn271A (3.1 Å and 2.9 Å, respectively) and its Nη2 atom forms a hydrogen bond with the Oδ1 atom of Asn271A (2.9 Å). In addition, the O atom of Gly136A forms a hydrogen bond with the Nδ2 atom of Asn271A (3.3 Å), and the Oη atom of Tyr139A forms a hydrogen bond with the Oδ1 atom of Asp275A (2.7 Å). Corresponding interactions also exist among the residues of subunit B. At the dimer interface, the Oη atom of Tyr272A forms two hydrogen bonds with the Oδ1 and Oδ2 atoms of Asn273B (2.7 Å and 3.4 Å, respectively), and the Oδ1 atom of Asp275A forms a salt bridge with the Nζ atom of Lys212B (2.8 Å), and vice versa. These results underscore the importance of Arg132 in the establishment of the interaction network and the stabilization of the two structural segments and thus, mutation of Arg132 to His would disrupt its interactions with the other residues and subsequently interactions between the two segments, leading to an increased flexibility of the two segments in the mutant IDH1 complexes.
It has been shown that the homodimeric IDH1 mutant has a significantly decreased IDH activity and intriguingly, the heterodimeric IDH1 mutant displays a comparable low IDH activity, suggesting that the two subunits of the dimeric IDH1 work in a cooperative manner 12. Our structures strengthen this notion by showing that although ICT is bound with both subunits, the wild-type subunit of the heterodimer could not accomplish the conformational changes and concurrently, the two key structural segments exhibit remarkably increased flexibility. As in the wild-type enzyme, the conformational changes of the two segments and the inter-domain motion are induced by ICT binding, and a number of residues in these two segments are involved in ICT binding, metal ion coordination, and catalysis 14, 22; we speculate that the increased flexibility of these two regions rendered by the R132H mutation might interfere with the ICT binding and/or the overall conformational changes of the enzyme, leading to the impairment of the IDH activity.
The R132H mutant complexes reveal a novel ICT-binding site
Comparison of the IDH1H/H complexes shows that in contrast to the wild-type IDH1, the ICT binding does not cause any significant domain movement in the mutant. In the IDH1H/H-NADP-ICT complex, although the two subunits of the homodimer display certain asymmetry, ICT binds at the same site which is far from the small domain and is distinct from the ICT-binding site in the IDH1R/R-NADP-ICT-Ca complex (Figure 1D). Correspondingly, the interactions between ICT and the surrounding residues are different (Figure 2C and 2D). Specifically, the O2 atom of α-carboxyl and O5 atom of β-carboxyl of ICT form salt bridges with the Nη2 and Nη1 atoms of Arg100 (3.0 Å and 3.2 Å), respectively, and the O6 atom of β-carboxyl forms a hydrogen bond with the Oγ atom of Ser94 (3.3 Å). In addition, the γ-carboxyl forms two hydrogen bonds with the 2′-OH group of NADP (3.0 Å) and the Oγ1 atom of Thr77 (3.6 Å; Figure 2C). Notably, in the mutant IDH1 complexes, the residues responsible for ICT binding (Thr77, Ser94, and Arg100) all belong to the large domain, and the respective interactions between ICT and the residues of the small domain in the wild-type IDH1 complexes are absent (Figure 2C and 2D). For convenience, we designate the ICT-binding site in the IDH1H/H-NADP-ICT complex as “site I” and that in the IDH1R/R-NADP-ICT-Ca complex as “site II”.
So far, there is no structure of IDH1 in complex with ICT alone available. For porcine and yeast (Saccharomyces cerevesiae) mitochondrial NADP-IDHs (PmIDH and Idp1), the structures of the enzymes in complexes with ICT show that the binding of ICT, but not NADP, induces transformation of the enzyme from an open conformation to a quasi-closed conformation 17, 18. In these structures, ICT binds at a site close to site II and interacts with residues of both large and small domains in a similar way. Besides, seg1 assumes an ordered structure containing a one-turn α-helix and seg2 adopts a long α-helix structure, similar to those observed in the IDH1R/R-NADP-ICT-Ca complex 14. These results suggest that for eukaryotic IDHs, the binding of ICT at site I appears to be associated with an open conformation of the enzyme with disordered structures of seg1 and seg2, while the binding of ICT at site II seems to be correlated to a closed conformation with ordered structures of the two segments.
Kinetic studies of the wild-type and mutant IDH1
To discern how the changes rendered by the R132H mutation, including the increased flexibility of seg1 and seg2 and the distinct binding mode of ICT, lead to impaired IDH activity, we carried out mutagenesis and biochemical studies towards Arg132 and other residues at the active site that have interactions with ICT and the metal ion. As shown in Table 2, the Vmax and Km (ICT) values of the wild-type IDH1 are comparable to that of PmIDH and E. coli NADP-IDH (EcIDH) 22, 23, 24, 25, 26. Compared with the wild-type IDH1, the homodimeric R132H mutant manifests a 200-fold increase in the Km whereas its specific activity is decreased to about 5%; similarly, the Km of the R132A mutant is also substantially increased by 106-fold while its specific activity is decreased to about 18% (Table 2). There is discrepancy in the IDH activity of the R132H mutant as Zhao et al. 12 showed a 2-fold decrease in the catalytic rate while Dang et al. 13 reported a 1000-fold reduction, and our data are in better agreement with that reported by Zhao et al. In addition, the kinetic parameters of the R132C, R132S, and R132L mutants are similar to that of the R132H mutant (Table 2). The Y139A and K212A mutants display < 1% activity (Table 2), indicating that Tyr139 and Lys212, like their counterparts in EcIDH and PmIDH 22, 25, 27, play critical roles in the catalytic reaction. Our data also show that mutations D252A and D275A have severe effects on the Km and/or kcat due to their involvements in both ICT binding and catalysis (Table 2).
Site I is an initial ICT-binding site
As discussed above, in the IDH1H/H-NADP-ICT structure, ICT binds at site I and has interactions with residues Thr77, Ser94, and Arg100 of the large domain. Lys212 retains a position similar to that in the IDH1R/R-NADP-ICT-Ca structure and thus is unable to interact with ICT (Figure 2C and 2D). In addition, Tyr139 of seg1 and Asp275 of seg2 appear to lose their interactions with ICT due to the increased flexibility of the two segments. As proper positioning of these key residues at the active site is critical for the enzymatic activity, the IDH1H/H-NADP-ICT structure is unlikely to represent a catalytically competent pre-transition state for the R132H mutant that retains very low IDH activity (our data and 12). Thus, we propose that formation of an active pre-transition state similar to the IDH1R/R-NADP-ICT-Ca structure is a pre-requisite for the catalysis to take place. This leads us to the following questions: is site I an initial ICT-binding site and if yes, how does IDH1 transform from the initial ICT-binding state to the pre-transition state?
In the IDH1H/H-NADP complex, seg2 is disordered in both subunits and thus, site I is easily accessible and residues Thr77, Ser94, and Arg100 are available for ICT binding (Figure 2E). However, in the IDH1R/R-NADP complex, seg2 is ordered but assumes different conformations in the two subunits, and site I is available only in subunit B, as in subunit A it is occupied by Asp279A of seg2 which makes hydrogen bonds with both Thr77A and Ser94A (Figure 2E). Among the three residues, Arg100 is of particular interest as it is available for ICT binding in both subunits of the wild-type or mutant IDH1 binary complex and appears to play an important role in the interactions of the large domain with ICT due to its contribution of two salt bridges with ICT (Figure 2C and 2D). Our biochemical data show that the R100A mutation has drastic effects on both Km (120-fold) and kcat (1%; Table 2), underscoring its pivotal roles in both ICT binding and catalysis. The T77A and S94A mutants possess about 26% and 22% activity, whereas their Km values are increased by about 15- and 10-fold, respectively; and the double mutation shows a synergetic effect (Table 2), indicating that these two residues play more important roles in ICT binding than in catalysis. These results are consistent with the biochemical studies of the equivalent residues (Thr78, Ser95, and Arg101) of PmIDH 24, 26.
Considering that the Km value can be affected by other factors such as the rate of conversion of the bound substrate to the product and hence may not accurately reflect the binding affinity between the substrate and the enzyme, we further analyzed the thermodynamics of ICT binding by the wild-type and mutant IDH1 using isothermal titration calorimetry (ITC; Figure 3). The ITC data show that the wild-type enzyme binds ICT with a KD value (1.13 μM) (Figure 3A and Table 2) comparable to the Km value (6.39 μM) as determined by the kinetics method (Table 2). In contrast, while the kinetic data show that the Km values of the T77A and S94A mutants are only increased by 15- and 10-fold, respectively (Table 2), the ITC data show that the KD value of the T77A mutant is increased by 153-fold and the S94A mutant shows a distinct profile with no measurable KD value (Figure 3B and 3C, and Table 2), suggesting that Thr77 and Ser94 might not be very critical in the ICT binding in the pre-transition state but play important roles in the initial binding of ICT. Although the Km values of the R100A and R132H mutants are increased to the same order of magnitude (120- and 200-fold, respectively; Table 2), the two mutants have distinct thermodynamic profiles: the R132H mutant shows some binding ability with ICT, whereas the R100A mutant has no detectable binding (Figure 3D and 3E), indicating that mutation of Arg100 has much greater impact on the KD compared with mutation of Arg132. These results together indicate that site I is an initial ICT-binding site and residues Thr77, Ser94, and Arg100 of the large domain are responsible for the binding of ICT at site I.
After the initial binding of ICT, the enzyme needs to convert from the open conformation to the closed conformation with the corresponding movement of ICT from site I to site II to reach the pre-transition state. A detailed analysis of the IDH1H/H-NADP-ICT structure shows that the Nη1 atom of Arg109 may interact with the α-carboxyl of ICT (4.6 Å and 4.9 Å from the α-carboxyl oxygen atoms, respectively), which may trigger such conformational changes. As the IDH1H/H-NADP-ICT complex assumes a conformation similar to that of the IDH1R/R-NADP complex, Arg109 may exert such a function in the wild-type enzyme. It is also possible that Arg132 may play a similar role in the wild-type enzyme. Analysis of the IDH1R/R-NADP complex shows that Arg132 forms salt bridges with Asp275 in subunit A and Asp279 in subunit B. At these positions, Arg132 would be too far to interact with ICT at site I. However, the ICT binding to site I might lead to disruption of the interactions between Arg132 and seg2 and thus, the “freed” Arg132 might be able to extend its side chain to interact with ICT and therefore initiate and/or participate in the conformational changes. The notion that Arg109 and Arg132 function in triggering the conformational changes is supported by the biochemical data that mutations of both Arg109 and Arg132 cause drastic increase in the Km (Table 2).
Mechanism of the impaired IDH activity of the R132H mutant
As discussed above, the IDH1H/H-NADP-ICT structure unexpectedly captures a snapshot representing a transient state of the ICT dehydrogenation reaction in which ICT is initially bound and the active site is partially assembled (Figure 4A). For the complete assembly of the active site and for the catalysis to take place, the enzyme has to convert from the open conformation of the initial ICT-binding state to the closed conformation of the pre-transition state represented by the IDH1R/R-NADP-ICT-Ca structure, accompanied by the movement of ICT from site I to site II (Figure 4A). During the catalytic reaction, Arg132 exerts multiple functional roles via direct interactions with ICT, facilitation of the conformational changes, and formation of the active site (this study and the study by Xu et al. 14). Mutation of Arg132 to His does not prohibit the binding of ICT, instead it constrains the enzyme in the inactive, initial ICT-binding state and prevents it from undergoing the conformational changes required to form the active, pre-transition state (Figure 4A).
Intriguingly, our results also show that the structure of the IDH1R/H-NADP-ICT complex resembles the corresponding structure of the homodimeric IDH1H/H rather than the wild-type enzyme, supporting the notion that trap of the mutant subunit in the initial ICT-binding state has a negative impact on the wild-type subunit 12 due to the cooperation of the two subunits in ICT binding and conformational changes via the two key structural segments (seg1 and seg2) 14. Our findings indicate that the heterodimeric IDH1R/H would function in a way similar to the homodimeric IDH1H/H, which accounts for the substantially reduced IDH activity of IDH1R/H at the physiological concentration of ICT and hence provides the mechanism underlying the observed dominant inhibitory effect of the R132H mutation 12. Recently, Dang et al. 13 showed that the NADPH production is decreased when the wild-type and R132H mutant IDH1 are co-expressed and thus proposed that the mutant subunit might consume αKG and NADPH produced locally by the wild-type subunit. Direct transfer of the produced αKG and NADPH from the wild-type subunit to the mutant subunit seems unlikely because the two active sites are located on the opposite sides of the dimer and the dimeric interface is very tight and compact; instead, the reduction in NADPH production most likely results from the inhibited activity of the heterodimeric mutant.
Mechanism of the gained reduction activity of the R132H mutant
Recently, Dang et al. 13 reported that the R132H mutation confers IDH1 a new function of catalyzing the conversion of αKG to 2HG, which might be important in the formation and progression of gliomas. Here, we show that the homodimeric R132H mutant harbours a weak reduction activity (Figure 4B) over a wide range of αKG concentration (10 μM-1 mM; Supplementary information, Figure S1A), and that Mn2+ but not Ca2+ facilitates the reaction (Figure 4C). Spontaneous turnover of αKG to 2HG is insignificant unless the αKG concentration is very high (e.g., 5 mM; Figure 4B). The structure of the homodimeric R132H mutant in complex with NADP, αKG, and Ca2+ (IDH1H/H-NADP-αKG-Ca) reported by Dang et al. 13 represents a post-transition state of the ICT dehydrogenation reaction and the pre-transition state of the αKG reduction reaction. It has been noted that the IDH1H/H-NADP-αKG-Ca structure 13 is very similar to the IDH1R/R-NADP-ICT-Ca structure 14 except that Tyr139 displays a marked positional displacement 13. This conformational difference should be rendered by the R132H mutation only because the active sites of the quaternary complexes of ICT-bound IDH1 14 and αKG-bound Idp1 18 are very similar, while αKG-bound IDH1H/H 13 and Idp1 18 differ mainly at this position (Supplementary information, Figure S1B). Our data prove the necessity of Tyr139 for the gained function by showing that the reduction activity of the R132H mutant is almost abolished when Tyr139 is substituted with either Phe or Ala (Figure 4B).
We further hypothesize that the reduction of αKG is initiated by the transfer of a hydride anion from NADPH to the C2 atom of αKG, and Tyr139 may play a vital role by compensating the increased negative charge on the C2 atom at the intermediate state (Figure 4D). Detailed structural analyses show that the hydroxyl group of the displaced Tyr139 is at a distance of 3.7 Å from the C2 atom of αKG in the IDH1H/H-NADP-αKG-Ca complex 13, while the corresponding distance is 5.8 Å in the Idp1-NADP-αKG-Ca complex 18 which would be too far for a similar interaction. Therefore, appropriate positioning of the phenolic hydroxyl of Tyr139 would be required for its interaction with the C2 atom of αKG and the reduction of the keto group of αKG, which is achieved in the R132H mutant. The metal ion may also be involved in the stabilization of the intermediate state by neutralizing the negative charge on the oxygen atom of the keto group (Figure 4D).
Taken together, our structural and biochemical data reveal a transient, initial ICT-binding state undiscovered before, implicate multiple functional roles of Arg132 in the catalytic reaction, and provide new insights into the catalytic mechanism of the ICT dehydrogenation reaction by IDH1. In particular, our results demonstrate that mutation of Arg132 to His does not prohibit the ICT binding, instead, it causes the enzyme to be constrained in the inactive, initial ICT-binding state and prevents the assembly of the active, pre-transition state, providing in part the molecular basis for the impaired IDH activity of the R132H mutant. As the wild-type IDH1 may act as a tumor suppressor in part through suppression of the HIF1α pathway by αKG 12, our results provide insights into the mechanism of the “off switch” of the potential tumor suppressive function of IDH1 by the R132H mutation. Furthermore, our structural and biochemical data also show that mutation of Arg132 to His causes a positional displacement of Tyr139, which is optimal to be involved in the αKG reduction reaction. Since the reduction product 2HG is implicated to play a role in the pathogenesis of brain tumors 13, our results also provide insights into the mechanism of the “on switch” of the hypothetically oncogenic function of IDH1 by the R132H mutation. These findings may have important implications in diagnosis of cancers associated with the IDH1 mutations and in the design of therapeutic strategies for these cancers.
Materials and Methods
Cloning, expression, and purification
The pRSF-Duet-IDH1R132H plasmid was kindly provided by Dr Yue Xiong at Fudan University and used for expression of the homodimeric IDH1H/H. Protein purification was carried out with Ni-NTA affinity chromatography and the target protein was eluted with an elution buffer (20 mM Tris-HCl (pH 7.4), 200 mM imidazole, and 500 mM NaCl). The protein was subsequently purified by gel filtration in a storage buffer (20 mM Tris-HCl (pH 7.4) and 100 mM NaCl) using a Hiload 16/60 Superdex G200 column (GE Healthcare).
The wild-type IDH1 was expressed with the pET-22b(+)-IDH1 plasmid as described previously 14. For expression of the heterodimeric IDH1R/H, the gene fragment of the wild-type IDH1 was cloned to pGEX-4T1, and E. coli BL21(DE3) cells were co-transformed with the pRSF-Duet-IDH1R132H and pGEX-4T1-IDH1 plasmids.
Purification of the heterodimer was performed as described above except that the elution fractions from the Ni-NTA affinity chromatography were subjected to further purification using glutathione affinity column (GE Healthcare) and subsequent enzymatic digestion using thrombin (GE Healthcare), followed by final purification using gel filtration in a storage buffer (20 mM Tris-HCl (pH 7.4) and 100 mM NaCl) on a Hiload 16/60 Superdex G200 column (GE Healthcare). Further analysis of the purified protein with native-PAGE showed that it has a migration rate different from that of the homozygous proteins, confirming that it is the heterozygous enzyme. The constructs of the mutant IDH1 containing point mutations except for R132H were generated using the pET-22b-IDH1 plasmid as the template with QuikChange Site-directed Mutagenesis kit (Stratagene). Expression and purification of these IDH1 mutants were carried out as described previously 14.
Crystallization and diffraction data collection
Crystallization conditions were screened using the Hampton Research Screen kits using the sitting-drop vapor diffusion method by Mosquito (TTP Labtech) at both 4 °C and 20 °C. The crystals of the IDH1H/H-NADP complex were grown in drops consisting of equal volumes (0.5 μl) of the protein solution (10 mg ml−1) and the reservoir solution (1.0 M Na2HPO4/KH2PO4 (pH 8.2)) without the addition of NADP at 4 °C. For crystallization screening of the IDH1H/H and IDH1R/H mutants in complexes with NADP and ICT, the protein solution was supplemented with 10 mM NADP, 100 mM ICT, and 100 mM CaCl2. The crystals of the IDH1H/H-NADP-ICT complex were grown in drops containing the reservoir solution of 0.1 M Tris-HCl (pH 8.5) and 2.0 M (NH4)2SO4, and that of the IDH1R/H-NADP-ICT complex were obtained from drops containing the reservoir solution of 0.1 M HEPES-Na (pH 7.5), 2% PEG400, and 2.0 M (NH4)2SO4. Diffraction data of the IDH1H/H binary complex were collected at beamline NW12 of Photon Factory, Japan, and that of the IDH1H/H and IDH1R/H ternary complexes at beamline BL17U1 of SSRF, China. All the diffraction datasets were processed using the HKL2000 program 28. A summary of the diffraction data statistics is given in Table 1.
Structure determination and refinement
The crystal structures of the R132H mutants were solved with the molecular replacement method implemented in the program PHASER of the CCP4 suite 29 using the structure of the IDH1R/R-NADP complex (PDB code 1T09) 14 as the search model. The initial structure refinement was carried out with the program CNS using the standard protocols 30 and the later structure refinement was performed with the programs REFMAC5 31 and Phenix 32. In addition, the IDH1H/H-NADP complex was further refined with the newly released CNS v1.3. A free R factor monitor calculated with 5% randomly chosen reflections was applied throughout the refinement. In all three structures, there are two monomers in the asymmetric unit, which were refined independently. Model building was performed with the program COOT 33 and guided by SIGMAA-weighted 2Fo-Fc and Fo-Fc maps. The stereochemistry qualities of the structure models were analyzed using the program PROCHECK 34 and the structural analyses were carried out using programs of the CCP4 suite 29. Figures were prepared using the program Pymol (http://www.pymol.org). A summary of the structure refinement statistics is given in Table 1.
IDH activity assay
The IDH activities of both wild-type and mutant IDH1 were assayed spectrophotometrically by monitoring the reduction of NADP to NADPH, which has an extinction coefficient of 6 220 M−1 cm−1 at 340 nm. The standard reaction buffer (1 ml) consisted of 30 mM Tris-HCl (pH 7.4), 0.1 mM NADP, 4 mM DL-ICT, and 2 mM MnSO4 25. The reaction was initiated by adding 5 μl of the wild-type or mutant IDH1 at appropriate concentrations. For determining the Km of ICT, the concentration of ICT was varied, whereas NADP and Mn2+ were maintained at the standard concentrations. All kinetic data were measured using a Beckman DU800 spectrophotometer (Beckman Coulter) and analyzed using the program Prism 4.0 (Graphpad Software). The kinetic data were fitted to the Scatchard plot for apparent Km and approximate Vmax which was used to calculate the kcat. All the experiments were performed in triplicates.
αKG reduction activity assay
The αKG reduction activities of the R132H, R132H/Y139F, and R132H/Y139A IDH1 mutants were assayed by measuring the conversion of NADPH to NADP. Briefly, 500 nM of the IDH1 mutant was added to 1 ml of an assay solution consisting of 30 mM Tris-Cl (pH 7.4), 5 mM MnSO4, 250 μM β-NADPH, and 5 mM αKG, and the decrease of absorbance at 340 nm over time was measured. To assay the reduction activity under the physiological condition, 100 μM β-NADPH and 25 μM αKG were used whereas Mn2+ was maintained at the standard concentration. All the experiments were performed in triplicates.
ITC analysis
Thermodynamic studies of both wild-type and mutant IDH1 were carried out by ITC experiments at 25 °C using an ITC200 calorimeter (GE healthcare). The protein samples were dialyzed against an assay buffer (30 mM Tris-HCl (pH 7.4), 100 mM NaCl, and 5% glycerol), and DL-ICT was dissolved in the same buffer. For the wild-type IDH1, titrations were conducted with 20 injections of 1 mM ICT, 1 μl each, into 0.2 ml of 0.044 mM IDH1 at an interval of 100 s. For the T77A, S94A, R100A, and R132H mutants, titrations were performed in a similar way except that 2 μl ICT was injected each time with an interval of 120 s, and for the T77A, S94A, and R100A mutants, the concentrations of the protein (0.075 mM) and ICT (2 mM) were increased for higher heat release. The thermodynamic parameters were obtained by nonlinear least-squares fitting of the data using the Microcal ORIGIN software. Heats of the reactions were corrected for the effect of dilution.
Protein data bank accession codes
The structures of the IDH1H/H-NADP, IDH1H/H-NADP-ICT, and IDH1R/H-NADP-ICT complexes have been deposited with the RCSB Protein Data Bank under accession codes 3MAR, 3MAP, and 3MAS, respectively.
Abbreviations
- ICT:
-
(isocitrate)
- αKG:
-
(α-ketoglutarate)
- 2HG:
-
(2-hydroxyglutarate)
- IDH:
-
(isocitrate dehydrogenase)
- NADP-IDH:
-
(NADP-dependent IDH)
- EcIDH:
-
(E. coli NADP-IDH)
- IDH1:
-
(human cytosolic NADP-IDH)
- PmIDH:
-
(porcine mitochondrial NADP-IDH)
- Idp1:
-
(S. cerevesiae mitochondrial NADP-IDH)
- GBM:
-
(glioblastoma multiforme)
References
Balss J, Meyer J, Mueller W, Korshunov A, Hartmann C, von Deimling A . Analysis of the IDH1 codon 132 mutation in brain tumors. Acta Neuropathol 2008; 116:597–602.
Bleeker FE, Lamba S, Leenstra S, et al. IDH1 mutations at residue p.R132 (IDH1R132) occur frequently in high-grade gliomas but not in other solid tumors. Hum Mutat 2009; 30:7–11.
Yan H, Parsons DW, Jin G, et al. IDH1 and IDH2 mutations in gliomas. N Engl J Med 2009; 360:765–773.
Watanabe T, Nobusawa S, Kleihues P, Ohgaki H . IDH1 mutations are early events in the development of astrocytomas and oligodendrogliomas. Am J Pathol 2009; 174:1149–1153.
Ichimura K, Pearson DM, Kocialkowski S, et al. IDH1 mutations are present in the majority of common adult gliomas but rare in primary glioblastomas. Neuro Oncol 2009; 11:341–347.
Hartmann C, Meyer J, Balss J, et al. Type and frequency of IDH1 and IDH2 mutations are related to astrocytic and oligodendroglial differentiation and age: a study of 1,010 diffuse gliomas. Acta Neuropathol 2009; 118:469–474.
Hayden JT, Fruhwald MC, Hasselblatt M, Ellison DW, Bailey S, Clifford SC . Frequent IDH1 mutations in supratentorial primitive neuroectodermal tumors (sPNET) of adults but not children. Cell Cycle 2009; 8:1806–1807.
Verhaak RGW, Hoadley KA, Purdom E, et al. Intergrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell 2010; 17:98–110.
Mardis ER, Ding L, Dooling DJ, et al. Recurring mutations found by sequencing an acute myeloid leukemia genome. N Engl J Med 2009; 361:1058–1066.
Kang MR, Kim MS, Oh JE, et al. Mutational analysis of IDH1 codon 132 in glioblastomas and other common cancers. Int J Cancer 2009; 125:353–355.
Sjoblom T, Jones S, Wood LD, et al. The consensus coding sequences of human breast and colorectal cancers. Science 2006; 314:268–274.
Zhao S, Lin Y, Xu W, et al. Glioma-derived mutations in IDH1 dominantly inhibit IDH1 catalytic activity and induce HIF-1a. Science 2009; 324:261–265.
Dang L, White DW, Gross S, et al. Cancer-associated IDH1 mutations produce 2-hydroxyglutarate. Nature 2009; 462:739–744.
Xu X, Zhao J, Xu Z, et al. Structures of human cytosolic NADP-dependent isocitrate dehydrogenase reveal a novel self-regulatory mechanism of activity. J Biol Chem 2004; 279:33946–33957.
Nekrutenko A, Hillis DM, Patton JC, Bradley RD, Baker RJ . Cytosolic isocitrate dehydrogenase in humans, mice, and voles and phylogenetic analysis of the enzyme family. Mol Biol Evol 1998; 15:1674–1684.
Gimelbrant A, Hutchinson JN, Thompson BR, Chess A . Widespread monoallelic expression on human autosomes. Science 2007; 318:1136–1140.
Ceccarelli C, Grodsky NB, Ariyaratne N, Colman RF, Bahnson BJ . Crystal structure of porcine mitochondrial NADP+-dependent isocitrate dehydrogenase complexed with Mn2+ and isocitrate. Insights into the enzyme mechanism. J Biol Chem 2002; 277:43454–43462.
Peng Y, Zhong C, Huang W, Ding J . Structural studies of Saccharomyces cerevesiae mitochondrial NADP-dependent isocitrate dehydrogenase in different enzymatic states reveal substantial conformational changes during the catalytic reaction. Protein Sci 2008; 17:1542–1554.
Hurley JH, Thorsness PE, Ramalingam V, Helmers NH, Koshland DE, Jr ., Stroud RM . Structure of a bacterial enzyme regulated by phosphorylation, isocitrate dehydrogenase. Proc Natl Acad Sci USA 1989; 86:8635–8639.
Yasutake Y, Watanabe S, Yao M, Takada Y, Fukunaga N, Tanaka I . Structure of the monomeric isocitrate dehydrogenase: evidence of a protein monomerization by a domain duplication. Structure 2002; 10:1637–1648.
Karlstrom M, Stokke R, Steen IH, Birkeland NK, Ladenstein R . Isocitrate dehydrogenase from the hyperthermophile Aeropyrum pernix: X-ray structure analysis of a ternary enzyme-substrate complex and thermal stability. J Mol Biol 2005; 345:559–577.
Kim TK, Lee P, Colman RF . Critical role of Lys212 and Tyr140 in porcine NADP-dependent isocitrate dehydrogenase. J Biol Chem 2003; 278:49323–49331.
Dean AM, Koshland DE, Jr . Kinetic mechanism of Escherichia coli isocitrate dehydrogenase. Biochemistry 1993; 32:9302–9309.
Soundar S, Danek BL, Colman RF . Identification by mutagenesis of arginines in the substrate binding site of the porcine NADP-dependent isocitrate dehydrogenase. J Biol Chem 2000; 275:5606–5612.
Grodsky NB, Soundar S, Colman RF . Evaluation by site-directed mutagenesis of aspartic acid residues in the metal site of pig heart NADP-dependent isocitrate dehydrogenase. Biochemistry 2000; 39:2193–2200.
Kim TK, Colman RF . Ser95, Asn97, and Thr78 are important for the catalytic function of porcine NADP-dependent isocitrate dehydrogenase. Protein Sci 2005; 14:140–147.
Lee ME, Dyer DH, Klein OD, Bolduc JM, Stoddard BL, Koshland DE Jr . Mutational analysis of the catalytic residues lysine 230 and tyrosine 160 in the NADP+-dependent isocitrate dehydrogenase from Escherichia coli. Biochemistry 1995; 34:378–384.
Otwinowski Z, Minor W . Processing of X-ray diffraction data collected in oscillation mode. Methods in Enzymology 1997; 276:307–326.
Collaborative Computational Project, Number 4. The CCP4 suite: programs for protein crystallography. Acta Crystallogr D Biol Crystallogr 1994; 50(Pt 5):760–763.
Brunger AT, Adams PD, Clore GM, et al. Crystallography & NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr D Biol Crystallogr 1998; 54(Pt 5):905–921.
Murshudov GN, Vagin AA, Dodson EJ . Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr D Biol Crystallogr 1997; 53(Pt 3):240–255.
Adams PD, Afonine PV, Bunkoczi G, et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr 2010; 66(Pt 2):213–221.
Emsley P, Cowtan K . Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr 2004; 60(Pt 12 Pt 1):2126–2132.
Laskowski RA, Moss DS, Thornton JM . Main-chain bond lengths and bond angles in protein structures. J Mol Biol 1993; 231:1049–1067.
Acknowledgements
We thank the staff members at Photon Factory of Japan and Shanghai Synchrotron Radiation Facility (SSRF) of China for technical supports in diffraction data collection, and Sheng Li, Jian Wu, and Xianchi Dong of our group for technical assistance and helpful discussion. This work was supported by Grants from the Ministry of Science and Technology of China (2006AA02A313 and 2007CB914302), the National Natural Science Foundation of China (30730028), and the Chinese Academy of Sciences (KSCX2-YW-R-107 and SIBS2008002).
Author information
Authors and Affiliations
Corresponding author
Additional information
( Supplementary information is linked to the online version of the paper on the Cell Research website.)
Supplementary information
Supplementary information, Table S1
The undetected regions of seg1 and seg2 in the R132H mutant complexes (PDF 7 kb)
Supplementary information, Table S2
Hydrogen-bonding interactions between the two flexible regions in the wild-type IDH1 complexes (PDF 17 kb)
Supplementary information, Figure S1
Mechanisms of “on switch” of the αKG reductase activity of IDH1 by the tumor-associated mutation R132H. (PDF 93 kb)
Rights and permissions
About this article
Cite this article
Yang, B., Zhong, C., Peng, Y. et al. Molecular mechanisms of “off-on switch” of activities of human IDH1 by tumor-associated mutation R132H. Cell Res 20, 1188–1200 (2010). https://doi.org/10.1038/cr.2010.145
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/cr.2010.145
Keywords
This article is cited by
-
Roles of metal ions in the selective inhibition of oncogenic variants of isocitrate dehydrogenase 1
Communications Biology (2021)
-
Identification of a selective inhibitor of IDH2/R140Q enzyme that induces cellular differentiation in leukemia cells
Cell Communication and Signaling (2020)
-
Molecular mechanism of the dual regulatory roles of ATP on the αγ heterodimer of human NAD-dependent isocitrate dehydrogenase
Scientific Reports (2020)
-
IDH mutation in glioma: molecular mechanisms and potential therapeutic targets
British Journal of Cancer (2020)
-
In vivo efficacy of mutant IDH1 inhibitor HMS-101 and structural resolution of distinct binding site
Leukemia (2020)