Crystal structure and catalytic activity of the PPM1K N94K mutant

Protein Phosphatase Mg²⁺/Mn²⁺-Dependent 1K (PPM1K), also known as branched-chain α-ketoacid dehydrogenase complex phosphatase, is a member of the metal-dependent protein phosphatase (PPM) family, and an important metabolic regulator located within the mitochondria that modulate a plethora of fundamental biological activities. PPM1K deficiency has been reported to closely correlate with type 2 diabetes, obesity, and cardiovascular and neurological disease. For instance, PPM1K gene polymorphisms that reduce PPM1K level in pancreatic β cells significantly reduced insulin secretion, and are associated with type 2 diabetes and obesity (Kettunen et al. 2012; Taneera et al. 2012; Xu et al. 2013). PPM1K expression was significantly decreased in the myocardium of the heart failure patients and inactivated PPM1K in zebrafish embryos led to aberrant neural and heart development (Lu et al. 2007).

Moreover, mounting evidence has shown that branched-chain amino acid tightly regulated by the branched-chain α-ketodehydrogenase (BCKD) has a critical function in the maintenance of correct brain function. Disorders in branched-chain amino acids homeostasis cause multiple neurological abnormalities such as toxicity and deterioration, encephalopathy, cerebral edema, inflammation, autism, epilepsy, intellectual disability, Alzheimer's disease, and seizures (Joshi et al. 2006; Novarino et al. 2012; Garcia-Cazorla et al. 2014; Conway and Hutson 2016; Scaini et al. 2017, 2018; Gruenbaum et al. 2018; Hull et al. 2018; Li et al. 2018). As PPM1K is a critical regulator of branched-chain α-ketoacid, through direct dephosphorylation of the pS²⁹³ site of the E1α subunit of the BCKD complex (Lu et al. 2009), several studies have linked PPM1K defects with neurological diseases in both rodents and humans (Cronin et al. 2009; Oyarzabal et al. 2013). Recently, a number of human PPM1K single nucleotide polymorphisms (SNPs) have been identified in the public database which have been reported to produce PPM1K defects, inducing inactive or constitutively active mutants. A frameshift mutation c.417_418TA (p.His139 fs) of PPM1K is highly associated with maple syrup urine disease that leads to severe neurological defects (Oyarzabal et al. 2013). Another SNP (rs10024717) of PPM1K is associated with sporadic amyotrophic lateral sclerosis in a genome-wide association study (Cronin et al. 2009). These studies suggest that the genetic polymorphism and activity of PPM1K play a pivotal role in the central nervous system.

PPM1K N94K (rs17853762) is an identified missense mutant produced by one of the SNPs in the human PPM1K coding sequence (Zhou et al. 2012). Recently, in a broad survey of PPM1K homologues among different species, we found that the residue Asn94 (in H. sapiens) is prevalent among mammals, birds, and reptiles, but the Lys94 variant is less prevalent and only detected in some species, such as C. canadensis, M. lucifugus, L. africana, G. gallus, C. livia, and O. hannah. However, the Lys94 variant is prevalent among amphibians and fishes (Fig. 1), suggesting that the Asn94 may possess evolutionarily conserved but diverse functions across different species. However, the effects of the N94K mutant on its activity and structural property have not been defined.

In this study, we investigated the catalytic and structural property of the N94K mutant of PPM1K. Whereas the PPM1K-N94K mutant displays normal activity toward pNPP and a phospho-segment derived from the pS²⁹³ site of the branched-chain α-ketoacid decarboxylase (E1) alpha subunit (BCKDE1α) using Mn²⁺ as the catalytic ion, it significantly impaired its catalytic activity using Mg²⁺ as the catalytic ion. The crystallographic study demonstrated that the N94K mutation caused the conformational change in the key residue in coordinating the Mg²⁺ in the active site. Specifically, three Mg²⁺ were located in the active site instead of two Mg²⁺ in the wild type. Our results provide a structure basis for the metal ion-dependent PPM1K-N94K phosphatase activity.

Materials and methods

Results

Crystal structure of the PPM1K-N94K mutant indicated the presence of the third Mg²⁺ in the active site

To understand the molecular mechanism underlying how N94K affect the catalytic activity of PPM1K, we expressed the truncated version of PPM1K-N94K (84-360) (Fig. S1). The successful substitution of PPM1K Asn94 with Lysine was confirmed by tandem mass spectrometry analysis (Fig. S2). The PPM1K-N94K (84-360) was crystallized in a condition similar to the wild type, diffraction quality crystals were obtained only in the presence of Mg²⁺ but not Mn²⁺. Similar to wild-type PPM1K, the space group of PPM1K-N94K is P3₁21 and the structure was solved at 2.6 Å resolution with a single molecule in one asymmetric unit (Table 5). Although the replacement of Asn94 with Lys did not disturb the local residue coordination in the PPM1K-N94K crystal (Fig. S3), two important metal ion–coordinated catalytic residues, including the D298 and E108, shift toward the active site by 0.7 Å and 0.6 Å, respectively (Fig. 2A). Whereas residues direct coordinate to the two conserved metal ion 1 and metal ion 2, including D127, G128, D298, and D337, are maintained, a water bridging the metal ion 1 and D298 is not visible. Moreover, the crystal structure unambiguously identified the third metal ion (M3) in the active site, which was coordinated by D208, D298, and several water molecules (Figs. 2B and S4). These observed Mg²⁺ coordination changes in the active site could provide a structural basis for the decreased Mg²⁺-dependent phosphatase activity of PPM1K-N94K compared to the wild-type PPM1K.

Table 5. Crystallographic data and refinement statistics

	PPM1K-N94K
Data collection
Space group	P3121
Cell dimensions
a (Å)	122.784
b (Å)	122.784
c (Å)	61.58
α (deg)	90
β (deg)	90
γ (deg)	120
Resolution	50–2.6 (2.60 –2.69)a
Unique observations	16 613 (1624)
Completeness (%)	99.60 (98.72)a
Redundancy (σ)	9.4 (9.8)
R merge	0.3 (1.69)a
Structure refinement
Resolution (Å)	35.45–2.61
Reflections used in refinement	16 568 (1615)
Rwork/Rfreeb (%)	0.19/22
RMSD ideal bonds (Å)	0.008
RMSD ideal angles (deg)	1.21
Ramchandran plot (%)
Most favored	98.04
Allowed
Generously allowed	1.96
Disallowed	3.3

• ^a Each dataset was obtained from a single crystal.
• The values in parentheses correspond to the highest resolution shell. RMSD, root-mean-square deviation.

A potential structural basis for substrate specificity of PPM1K-N94K

Similar to all other PPM phosphatase member, the PPM1K wild-type and PPM1K-N94K harbor a catalytic cleft, which contains residues that may contribute to their substrate specificity. In addition to D298, we identified that N247 in the catalytic cleft of N94K undergoes a larger-scale conformational change with root-mean-square deviation value of 1.0 in PPM1K-N94K mutant relative to other residues, compared to the PPM1K wild type (Fig. 3A and B). We, therefore, measured the phosphatase activities of wild-type PPM1K and N94K mutant toward a phospho-segment derived from the pS²⁹³ site of the BCKDE1α. Whereas the PPM1K-N94K only displays a slightly higher k_cat/K_m toward phospho-pS²⁹³- BCKDE1α segment compared with wild-type PPM1K (Fig. 4A and B) that is similar to previously published results by cellular studies (Zhou et al. 2012), it does not exclude the observed conformational change as a result of N247 that contribute to the substrate specificity of PPM1K-N94K toward another protein substrate of PPM1K.

Discussion

In this study, we measured the phosphatase activity of PPM1K-N94K mutant toward small artificial substrate pNPP and a phospho-segment derived from the pS²⁹³ site of the BCKDE1α. Whereas the Mn²⁺-dependent phosphatase activity of PPM1K-N94K using pNPP as the substrate was slightly decreased when compared to wild type, the catalytic activity toward phospho-segment is slightly elevated (Table 3 and Fig. 4b). These observations are similar to previously published results by cellular studies (Zhou et al. 2012). Most strikingly, we found that PPM1K exhibited moderate catalytic activity toward pNPP when using Mg²⁺ as the catalytic ion (Table 4), whereas all previous reports of its activity had been in the presence of Mn²⁺ (Wynn et al. 2012).

Mg²⁺ participates in many biochemical processes that are crucial for neuronal properties and synaptic plasticity (Xu et al. 2014; Nayak et al. 2018). Recent studies have demonstrated that the serum and brain magnesium levels were decreased in Alzheimer's disease (AD), cerebral injury, and other neurological abnormalities (Daher et al. 2018; Meloni et al. 2011). Therefore, it is important to understand how Mg²⁺ regulated activities of key enzymes with relevant functions in neurological activities. It was reported that the phosphatase activity of PPM1K is intimately related to the normal function of nerve cells (Garcia-Cazorla et al. 2014; Gruenbaum et al. 2018). Here, our studies revealed that PPM1K N94K mutant decreased Mg²⁺-dependent phosphatase activity, suggesting that PPM1K N94K mutant may be involved in the pathological processes of aforementioned neurological abnormalities. Future in vivo studies, such as creating a PPM1K-N94K knock in mice model, will be beneficial to test such hypothesis.

Many members of the PPM superfamily have been structurally characterized and all contain the αββα sandwich arrangement of the catalytic domain (Egloff et al. 1995; Goldberg et al. 1995; Das et al. 1996). In the active site of PPM1K, a binuclear metal ions (Mn²⁺ or Mg²⁺) bridged by the conserved D127 (corresponding to D60 in PPM1A) and a water molecular played critical roles in catalysis (Wynn et al. 2012; Pan et al. 2015). Recently, our studies have shown whereas replacing the catalytic metal ion in M1 site of PPM1A or PPM1G by cadmium abolished the phosphatase activity (Pan et al. 2013), deleting the catalytic metal in M2 site through a D38 mutation in PPM1A impaired the phosphatase activity through losing the ability in neutralization of the negative charge developed by the leaving group of the substrates (Pan et al. 2013). Moreover, some crystal structures of bacterial, plant, and human PPM phosphatases have revealed an additional, third metal ion (M3) along with two tightly bound metal ions in the active site (Pullen et al. 2004; Rantanen et al. 2007; Wehenkel et al. 2007; Schlicker et al. 2008; Su et al. 2011; Tanoue et al. 2013; Debnath et al. 2018). Studies have shown loss of M3 binding has no effect on the binding of the first metal (M1) and the second metal (M2) in the catalytic center but notably impaired the catalytic activity in Thermosynechococcus elongatus and necessity of M3 in PPM-catalyzed reactions (Su et al. 2011). Higher enzyme activity with random order, bi-substrate mechanism was observed toward phospho-peptide substrate when M3 binds a low-affinity site in presence of M1 and M2 which allows solvent access to the third metal in the catalytic domain of active site in the human phosphatases PP2Cα (PPM1A) and Wip1 (PPM1D) (Tanoue et al. 2013; Debnath et al. 2018). By determination of the PPM1K-N94K crystal structure, we identified that the PPM1K-N94K mutant accommodates a third Mg²⁺ in the active site, accompanied with the slight change in the coordination state of two known conserved metal ion (Figs. 2A and S4). The metal ion coordination changes in the active site could provide the possible structural basis for the decreased phosphatase activity of PPM1K-N94K toward pNPP compared to the wild-type PPM1K (Tables 3 and 4). In this study, we mainly focused on the correlation between the Mg²⁺ coordination changes and the decreased Mg²⁺-dependent activity of PPM1K-N94K. The decreased phosphatase activity of PPM1K-N94K in the presence of Mg²⁺ but not Mn²⁺ is possibly attributed to the differences in the distance between Mg²⁺-water and Mn²⁺-water dative covalent bonds because the atomic radius of manganese is lower than that of magnesium. It would be particularly important to analyze a Mn-bound structure in order to evaluate whether the same active site coordination changes are not observed in the presence of manganese. Moreover, the moderately tighter binding of Mg²⁺ to PPM1K-N94K may have contributed to the presence of the third metal ion (M3) in the active site in the crystal of the mutant rather than the crystal of the wild-type enzyme, which displays only two divalent ions in the active site. We also observed the conformational changes in the N247 and D298 in the catalytic cleft of the PPM1K-N94K compared with wild-type PPM1K (Fig. 3). As the N94K mutation has a minor effect on the overall structure, another explanation is that the active site conformation may be dependent on the presence/absence of Mg²⁺/Mn²⁺ ions, with different ‘states’ occurring in the PPM1K wild type or N94K mutant. Currently, all crystals of PPM1K-N94K/Mn²⁺ complex diffracted poorly, which impeded us to give a full interpretation of the above questions. Further studies using different constructs of PPM1K that produce better quality crystals may help in solving these problems.

Conclusion

In brief, our present study unraveled that PPM1K-N94K mutant significantly impaired its Mg²⁺-dependent catalytic activity while inducing a conformational change in the key residue in coordinating the Mg²⁺ in the active site. These observations might provide a structure basis for the metal ion-dependent PPM1K-N94K phosphatase activity and its substrate specificity.

Acknowledgments and conflict of interest disclosure

We would like to thank Xiang Ding and Mengmeng Zhang from Laboratory of Proteomics, Core Facilities for Protein Science, at the Institute of Biophysics (IBP), Chinese Academy of Sciences, for their help with the mass spectrometry analysis. This work was supported by the National Natural Science Foundation of China (31470789, 31611540337, and 81773704 to J.-P.S.; 31471102 and 31671197 to X.Y.; 31870781 to P.-J.Z.; and 31700692 to P.X.), the National Science Fund for Distinguished Young Scholars (81825022 to J.-P.S.), the Shandong Natural Science Fund for Distinguished Young Scholars (JQ201517 to J.-P.S.), the Shandong Natural Science Fund (ZR2016CQ07 to P.X.), the Key Research and Development Program of Shandong Province (2018GSF118073 to P.-J.Z.; 2018GSF118147 to P.X.), the Fundamental Research Fund of Shandong University (2016JC017 to J.-P.S.), and the China Postdoctoral Science Foundation (2015M582082 to P.X.) and the Rolling program of ‘ChangJiang Scholars and Innovative Research Team in University’ (IRT_17R68 to Y.-M.S.). The authors have no conflict of interest to declare.

Institutional approval

No institutional approval was required for this study.

Data availability

The authors declare that data supporting the finding of this study are available within the article and its supplementary information and additional data are available from the corresponding author upon reasonable request.