Role of Conserved Histidine and Serine in the ... - ACS Publications

Subscriber access provided by UNIV OF TEXAS DALLAS

Computational Biochemistry

The role of conserved histidine and serine in the HCXXXXXRS motif of human dual-specificity phosphatase 5 Ankan Gupta, Jaladhi Brahmbhatt, Raulia R. Syrlybaeva, Catherine Bodnar, Natalia Bodnar, Robert D. Bongard, Phani Raj Pokkuluri, Daniel S. Sem, Ramani Ramchandran, Rajendra Rathore, and Marat R. Talipov J. Chem. Inf. Model., Just Accepted Manuscript • DOI: 10.1021/acs.jcim.8b00919 • Publication Date (Web): 05 Mar 2019 Downloaded from http://pubs.acs.org on March 16, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Information and Modeling

The Role of Conserved Histidine and Serine in the HCXXXXXRS Motif of Human Dual-specificity Phosphatase 5 Ankan Gupta1,$, Jaladhi Brahmbhatt2,$, Raulia Syrlybaeva3, Catherine Bodnar4, Natalia Bodnar4, Robert Bongard4, Phani Raj Pokkuluri5, Daniel Sem4, Ramani Ramchandran1,6,$,*, Rajendra Rathore7,$ and Marat R. Talipov3,$,*. Affiliations: 1Department

of Pediatrics, Division of Neonatology, Children’s Research Institute (CRI), Developmental Vascular Biology Program, Translational and Biomedical Research Center, 8701 Watertown Plank Road, P.O. Box 26509, Milwaukee, WI 53226, USA, 2BioTechnology Discovery Research, Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, IN 46221, 3Department of Chemistry, New Mexico State University, 1175 N. Horseshoe Drive, Las Cruces, NM 88003, USA, 4Center for Structure-based Drug Design and Development, Department of Pharmaceutical Sciences, Concordia University Wisconsin, 12800 North Lake Shore Drive, Mequon, WI 53907, 5Biosciences Division, Argonne National Laboratory, Lemont, IL 60439 USA 6Department of Obstetrics and Gynecology, 8701 Watertown Plank Road, P.O. Box 26509, Milwaukee, WI 53226, USA, 7Department of Chemistry, Marquette University, Wehr Chemistry Building, P.O. Box 1881, 535 N. 14th Street, Milwaukee, WI 53201, USA. *Corresponding author: Marat R. Talipov, PhD (for computational-related queries) Ramani Ramchandran, PhD (for DUSP5 biology-related queries) $These authors contributed equally to this work. Address: 3Department of Chemistry, New Mexico State University Email: [email protected] or [email protected] Phone: 575-646-5210 (Marat R. Talipov) or 414-955-2387 (Ramani Ramchandran)

1

ACS Paragon Plus Environment

Journal of Chemical Information and Modeling 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ABSTRACT Background: The mitogen-activated protein kinases (MAPK) pathway is functionally generic and critical in maintaining physiological homeostasis and normal tissue development. This pathway is under tight regulation, which is in part mediated by dual-specific phosphatases (DUSPs), wherein they dephosphorylate serine, threonine and tyrosine residues of ERK family of proteins. DUSP5 is of high clinical interest because of mutations we identified in this protein in patients with vascular anomalies. Unlike other DUSPs, DUSP5 has unique specificity towards substrate pERK1/2. Using molecular docking and simulation strategies, we previously showed that DUSP5 has two pockets, which are utilized in a specific fashion to facilitate specificity towards catalysis of its substrate pERK1/2. Remarkably, most DUSPs share high similarity in their catalytic sites. Studying the catalytic domain of DUSP5 and identifying amino acid residues that are important for dephosphorylating pERK1/2, could be critical in developing small molecules for therapies targeting DUSP5. Results: In this study, we utilized computational modeling to identify and predict the importance of two conserved amino acid residues, H262 and S270 in the DUSP5 catalytic site. Modeling studies predicted that catalytic activity of DUSP5 would be altered if these critical conserved residues were mutated. We next generated independent Glutathione-S-Transferase (GST)-tagged full-length DUSP5 mutant proteins carrying specific mutations H262F and S270A in the phosphatase domain. Biochemical analysis was performed on these purified proteins, and consistent with our computational prediction, we observed altered enzyme activity kinetic profiles for both mutants with a synthetic small molecule substrate (pNPP) and the physiological relevant substrate (pERK) when compared to wild type GST-DUSP5 protein. Conclusion: Our molecular modeling and biochemical studies combined demonstrate that enzymatic activity of phosphatases can be manipulated by mutating specific conserved amino acid residues in the catalytic site (phosphatase domain). This strategy could facilitate generation of small molecules that will serve as agonist of DUSP5 activity.

2


Page 2 of 22

Page 3 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


INTRODUCTION The mitogen-activated protein kinases (MAPKs) is the output for various extracellular stimuli in the cell, which determines physiological responses such as gene expression, cell proliferation and other cellular functions1, 2. MAPKs are activated by dual phosphorylation of tyrosine and threonine residues in their activation loops3, 4. Proteins that dephosphorylate both at serine/threonine and tyrosine residues are referred to as dual-specificity phosphatases (DUSPs), which have evolved to regulate MAPK pathways3, 5. DUSPs fall under the sub-family of type I-cysteine based phosphatases, which consists of 61 members6. Our group have focused on the fifth member in this group, DUSP5 (UniProtKB Q16690) that was reported to have ‘serine to proline’ mutation (S147P) as identified in patients diagnosed with vascular anomalies7. Unlike other members of this family, DUSP5 is uniquely conserved for its physiological substrate pERK 1 or 2 (pERK1/2) and does not act on any other MAPKs8, 9. However, DUSP5 has been recently reported to inhibit NF-κβ activity in vitro10. In contrast to dephosphorylation activity in the nucleus, DUSP5 relieves the inhibitors of MAPK and thereby promotes phosphorylation of ERK in the cytoplasm11. Depending on its cellular localization and activation state, DUSP5 has been implicated in various pathological states. DUSP5 acts as a tumor suppressor in some cases such as skin cancer where it negatively regulates the oncogenic potential of activated Ras12, while in other cancers (prostate and breast), it acts a tumor promoter13. Recently we showed that DUSP5 is implicated in the expansion of virus-specific T cells14. DUSP5 has also been implicated in cardiac hypertrophy15. In mouse model, transgenic expression of DUSP5 has been shown to prevent thymocyte development and hastens autoimmune symptoms16. Together, these studies suggest that DUSP5 is important target for tumor development and also for diseases, influencing cardiovascular as well as immune system. Our laboratory has made a systematic attempt to decipher the underlying mechanism through which DUSP5 dephosphorylates pERK1/217. The active site of DUSP5 includes the phosphatase signature motif HCXXXXXRS that is essential for the catalytic activity of DUSP5 and known as the catalytic P-loop18-20. Compared to other DUSPs, DUSP5 has a uniquely positioned glutamic acid (E264), adjacent to the enzymatically nucleophilic cysteine (or catalytic cysteine) at 263rd amino acid position. It also possesses other unique features such as an additional substrate-binding site or secondary site, which we have hypothesized to influence the accessibility to the active site9, 17. In our proposed “lock-and-key” mechanism9, E264 forms a salt bridge (ionic interaction) with R269 residue thus preventing access to catalytic C263. On occupation of secondary site with an anion residue (phosphate group), the ionic interaction is broken, thus facilitating entry of second anion (phosphate group) into the catalytic pocket thus initiating catalysis by C263 nucleophilic attack. This mechanism is hypothesized to explain the unique substrate specificity of DUSP5 to pERK. Dephosphorylation of a substrate by DUSPs is an energetically challenging process that requires co-localization of a double negative charge of the incoming phosphate group and negative charge of the catalytic cysteine thiolate21, 22. Analysis of the experimental structures of various DUSPs (Table 3) shows that the catalytic P-loop plays an important role in the stabilization of this overall triple negative charge. In addition, the dianion substrate in DUSPs is stabilized by the guanidine group of a well-conserved Arg R269 (+6) amino acid residue18, while the catalytic cysteine (C263) thiolate is stabilized by the hydrogen bonds with the hydroxyl group of the Ser (S270) (+7)22,23. Also, the presence of a hydrogen bond between His (H262) (-1) side chain and the backbone of the catalytic cysteine in several phosphatase structures further suggested23, 25 that His (-1) also serves to stabilize the negative charge of the catalytic cysteine. In this light, it was surprising to find in the X-ray structure of DUSP5 that His (-1) did not form a hydrogen bond with the catalytic cysteine24 and therefore did not participate in electrostatic stabilization of the negative charge in DUSP5. Absence of the proposed interaction between His (-1) and catalytic cysteine in DUSP5 was indicative of the existence of an alternative mechanism of action of His (-1) in DUSP5.

3



Page 4 of 22

Because of the indication that the catalytic cysteine anion of DUSP5 might be stabilized by a nonconventional mechanism, we hypothesized that H262 and S270 amino acid residues might play a specific role for the catalytic site of DUSP5, which might differ from the traditional catalytic mechanism shown in Scheme 1. Accordingly, in this study we have used combined computational and experimental approaches to decipher the role of His (-1) and Ser (+7) in the catalytic activity of DUSP5 towards the model (pNPP) and native (pERK) substrates. The computational studies included the long time-scale non-reactive molecular dynamics simulations, which were used to identify the prevalent interactions in the catalytic site of wild-type and mutant DUSP5. The experimental studies included expression and purification of the WT and mutant DUSP5, combined with the dynamic light scattering, circular dichroism, pNPP and pERK activity assays. Based on the obtained results, we proposed a kinetic model of the catalytic activity of DUSP5. The details of this study are discussed herein. Scheme 1. Schematics showing typical mechanism of catalytic activity of phosphatases as exemplified in DUSP5.

RESULTS AND DISCUSSION Analysis of the S270A mutation on the catalytic activity of DUSP5 To investigate the role of the Ser (+7) and His (-1) amino acids for the catalytic efficiency of DUSP5, we performed a series of MD simulations on the phosphatase domain (PD) of the wild-type (WT), S270A, and H262F mutant DUSP5, with total simulation time being 15 microseconds (Table 1). A 1.8 microsecond-long MD simulation of WT PD, with catalytic cysteine being in its thiolate form (Cyscat-S–), showed a single conformation of the catalytic P-loop consistent with that in the X-ray structure of DUSP5 PD (Fig 1A). Interestingly, during 95% of the MD simulation time, the negative charge of the Cyscat-S– was stabilized by the hydrogen bonding with the hydroxyl group of Ser270 and the N-H bonds of the P-loop backbone (Fig 1B). The rest 5% of the MD simulation corresponded to a different Cyscat-S– stabilization mechanism, in which Cyscat-S– was interacting with the positively charged guanidine group of Arg269 (+6) (Fig 1C). In agreement with the possible favorable thiolate stabilization mechanism (Fig 1B vs 1C), the analysis of the MD trajectories using the continuous solvation model showed that the 95:5 ratio between the different stabilization mechanisms (Fig 1 B vs 1C) arises from the free energy difference of 1.7 kcal/mol between the corresponding configurations of the active site.

A

CO

HN

NH2

A

NH

S

SO

HN NH R269

I 267

H G H 266 O N

S270

NH2 NH

HN

H262

NH N

I 267

E 264

CO 97 no c % on of ta A C263 sim ct i . ti n – HN me S

HN

HN NH R269

S268NH

C

E 264

H2N

C263 –

24

S268NH

B

E 264

H2N

H G H 266 O N

H262

NH N

S270

HN

A NH

NH2 NH2

NH R269

S268NH

I 267

–

S

HN

CO 1 n 00 o c o C263 % of ntac sim t in HN . ti me

H262

H G 266 N

NH N

S270 OH

Accounts for 95% of MD simulation time

Accounts for 5% of MD simulation time

Fig 1. Schematics showing molecular dynamics (MD) simulation analysis of DUSP5 phosphatase domain (PD). Structure of the catalytic pocket of DUSP5 PD as derived from crystallographic analysis

4


Page 5 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(A). Two distinct mechanisms (B & C) of the catalytic thiolate stabilization, as observed during MD simulations of DUSP5 PD (Cys263-S–). The MD simulations of the S270A DUSP5-PD showed that absence of hydroxyl group in the S270A mutant resulted in a complete switch to a mechanism of the Cyscat-S– stabilization, which previously played only a minor role in WT DUSP5-PD. Thus, in S270A PD, the catalytic thiolate was stabilized exclusively by the side chain of Arg269 and the NH bonds of the P-loop backbone (Fig 1C).

Table 1. Summary of the MD simulations of DUSP5 PD.

a b

Ro w

Mutation s

1

Protonation state 2-)

N(SO4

Ions

b

N(H2O )

Total simulation time, ns c

RMSDmax, Åd

Cys263

His262 a

-

-SH

HID

-

-

6826

1819

2.3

2

-

-SH

HIE

-

-

6825

906

2.0

3

-

-SH

HIP

-

1 Cl–

6826

879

1.8

4

H262F

-SH

-

-

-

6824

1730

2.1

5

S270A

-SH

HID

-

-

6825

900

1.3

6

-

-S–

HID

-

1 Na+

6824

1827

3.3

7

-

-S–

HIE

-

1 Na+

6827

1688

1.8

8

-

-S–

HIP

-

-

6824

905

1.5

9

H262F

-S–

-

-

1 Na+

6825

1818

2.3

10

S270A

-S–

HID

-

1 Na+

6828

904

2.0

11

-

-S–

HID

1e

3 Na+

6813

15 f

-

12

H262F

-S–

-

1e

3 Na+

6817

3f

-

13

-

-S–

HID

2g

5 Na+

6820

600 f

2.0

14

H262F

-S–

-

2g

5 Na+

6821

80 f

1.8

15

-

-SH

HID

2g

4 Na+

6821

789 f

1.5

HID, HIE, and HIP denote histidine protonated at the N(), N(), or both at N()+N(), respectively; Used to neutralize the charge of the periodic cell;

5



Page 6 of 22

Aggregated from three independent simulations. The aggregated simulation time was chosen to be at least 1600 ns for the simulations of primary interest (i.e., rows 1, 4, 6, 7, and 9) and at least 850 ns for other simulations; d Calculated from the positions of the DUSP5 PD backbone atoms; e Placed in the catalytic pocket; f Simulation was manually interrupted because of the dissociation of the complex between the dianion substrate and DUSP5 PD; g Second sulfate group was placed in the binding site in proximity to the catalytic pocket, as observed in the X-ray structure of DUSP5 PD (PDB code: 2G6Z). c

Analysis of the role of H262 in DUSP5 PD The purported role of His in stabilization of the negative charge of the catalytic cysteine is based on the observation that its imidazole ring forms a hydrogen bond with the backbone of the catalytic cysteine in various tyrosine phosphatases 23, 25. However, no such bonding was observed in the X-ray structure of DUSP524. Our prolonged MD simulations of DUSP5-PD also showed the absence of such His(-1)– N()H...O=C–Cyscat interaction. Also, the MD simulations of H262F DUSP5-PD showed that the dynamics of the catalytic site was similar in H262F and WT PD (Table 2). Table 2. Time fraction (in %) of the presence of various short-range (within 2.5 Å) contacts that stabilize the thiolate form of the catalytic cysteine 263 in DUSP5

WTWTH262F HID262 HIE262

Short-range contact a,b

a b

His(-1)-NH...O=Cbb-Cys-S-

3

-

-

Cys(0)-NHbb...S-

1

4

0

Glu(+1)-NHbb...S-

1

1

1

Ala(+2)-NHbb...S-

10

4

13

Gly(+3)-NHbb...S-

64

45

62

Ile(+4)-NHbb...S-

3

5

8

Ser(+5)-NHbb...S-

0

5

0

Arg(+6)-NHbb...S-

0

5

0

Ser(+7)-NHbb...S-

62

63

50

Ser(+7)-OH...S-

95

96

89

Arg(+6)-NH2...S-

4

2

11

The interacting atoms are marked by the underscoring line; 'bb' denotes backbone atoms forming a peptide bond.

6


Page 7 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


These observations indicate that H262 does not directly stabilize the catalytic cysteine via the electrostatic interactions. Indeed, the space occupied by H262 might influence the structural arrangement of the catalytic site and as a result, could change the catalytic activity of DUSP5. It is likely that such role of H262 should be best evident from the dynamics of the most flexible part of the active site of DUSP5, the P-loop. In this light, it is interesting to note that our MD simulations of WT, S270A, and H262F PD with Cyscat-S– (Fig 2) showed an identical conformation of the P-loop, which was rigid during the MD course and consistent with that in the X-ray structure of DUSP5. Intriguingly, similar MD simulations with neutral Cyscat-SH showed that the P-loop becomes flexible, with the ratio of conformation distribution varying in WT vs H262F PD (Fig 2). The presence of such differences indicates that H262F might affect the P-loop in a manner that influences the conformational mobility of the active site.

A

B

C

WT PD: 66% H262F: 34%

WT PD: 4% H262F: 36%

D

WT PD: 31% H262F: 22%

Fig 2. Geometry of the DUSP5 catalytic P-loop. Distinct conformations of the catalytic P-loop observed during MD simulations of DUSP5 PD (Cys263-SH). Numbers represent fraction of the MD simulation time specific for each conformation for the cases of PD from WT DUSP5 and H262F mutant. Based on the observed lack of electrostatic contribution of His (-1) into stabilization of Cyscat-S– and the role of the P-loop-mediated conformational mobility of active site in Cyscat-SH form of H262F vs WT, we speculated that the role of His (-1) might be to stabilize a dianion ligand in the catalytic pocket by organizing the proper structural arrangement of the catalytic site. The MD simulations of WT DUSP5-PD and H262F PD, where the catalytic site had Cyscat-S– and was occupied by a dianion ligand, were undertaken using a sulfate dianion as a surrogate for the native substrate of DUSP5. This choice of dianion was motivated by the comparable size/shape of the phosphate and sulfate groups and the fact that the position in the catalytic pocket has been reliably established by the X-ray crystallography (PDB entry 2G6Z)24. Interestingly, the MD simulations showed that the WT DUSP5-PD complex with sulfate was kinetically unstable and dissociated on a time scale of 3-6 ns (Fig S1 in the Supporting Information). Similar MD simulations for the H262F mutant showed even lower dissociation time of 0.1-1.5 ns. The reduced dissociation time for the dianion/PD complex suggest that H262F DUSP5-PD has lower binding affinity toward dianions than WT PD. This observation was substantiated by the calculations of ligand binding energy using the MM/PBSA and MM/GBSA approaches. These calculations showed that the free energy of the sulfate group binding was reduced by ∆G = 1-2 kcal/mol due to the H262F mutation. Such a difference in the binding energy corresponds to the 4- to 10-fold reduction in the binding constant in the H262F mutant at room temperature according to the equation for the equilibrium constant Keq = exp(∆G/RT). A clustering analysis of the pre-dissociation dynamics of the complex of catalytic pocket of DUSP5-PD with dianion ligand showed two conformations of the catalytic site (Fig 3). In the first 7



conformation (Fig 3A), the N-H bond of Glu264 backbone forms a short-range (2.8 Å) contact with the sulfate group and a 2.3 Å contact with the negatively charged sulfur atom of Cyscat-S–. These short-range contacts were absent in the second conformation and compensated by a hydrogen bond between the backbone C=O group of Cys263 and the backbone N-H group of Ala191 (Fig 3B). The second conformation is expected to be energetically unfavorable as it reorients the Cys263 sidechain and pushes the negatively charged sulfur atom toward the dianion substrate from 4.5 Å to 4.0 Å. Interestingly, the first conformation (Fig 3A) was found to be more prevalent in WT DUSP5-PD (observed during 61% of the MD simulation), while the second conformation (Fig 3B) was more prevalent in the H262F DUSP5-PD mutant (55%). Since the side imidazole ring-chain of His262 was turned away from Cys263 in WT DUSP5PD, the switch in the conformation prevalence in H262F mutant could not be attributed to the loss of electrostatic stabilization. Instead, it is likely that the switch in the conformational prevalence switch in the H262F mutant DUSP5-PD likely arises from the larger volume occupied by phenylalanine as compared with histidine.

Fig 3. Geometry of catalytic pocket of DUSP5 PD. Distinct conformations (A, B) of the catalytic pocket as observed during short (6 ns) MD simulations of DUSP5 PD (Cys263-S–) in presence of a di-anionic substrate. Numbers represent fraction of the MD simulation time specific for each conformation for the cases of PD from WT DUSP5 and H262F mutant. Generation, purification and characterization of GST-tagged WT and DUSP5 mutants. To confirm the MD simulation predictions, we investigated the role of His262 and Ser270 residues in DUSP5 by generating point mutations in DUSP5 and investigating its biochemical activity towards the substrates pNPP and pERK. First, we generated GST-tagged H262F and S270A mutant constructs in the backbone of the WT DUSP5 construct published previously26 (Fig 4A and 4B) in E. coli bacteria. Point mutations (nucleotide changes to reflect S270A and H262F) were introduced in the GST-DUSP5 construct using a quick-change mutation kit. The sequence-confirmed plasmids were transfected into competent cells for protein generation. The IPTG method of protein induction was utilized as described in our previous studies26. Clones showing robust protein expression were subsequently utilized in this study. We then purified the GST-tagged proteins on GST column, and subjected the purified proteins to Mass Spectrometry analysis to confirm the mutation (Fig S2). Because previous studies showed minimal difference between GST-DUSP5 and DUSP5 protein in terms of activity26, we did not cleave the GST tag in these studies and therefore, used the fusion proteins in subsequent characterization and activity assays.

8


Page 8 of 22

Page 9 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


B

75

A

50

25

WT

H262F S270A

69 kDa

*

**

Conceptualized from: Talipov et al; Biochemistry 2016, 55, 6187−6195

Fig 4. Generating point mutants of DUSP5. Schematics of the GST-DUSP5 fusion protein with various domains (A). Length of each domain is indicated by the number of amino acid residues (not in scale). The mutations are indicated at the respective locations in phosphatase domain. SDS−PAGE gel showing the band of the two GST tagged point mutants along with WT-DUSP5 (B). The DUSP5 monomer (~42kDa), along with GST tag (~27 kDa) was expected to yield a product of 69 kDa (black arrow). Trace amount of untagged enzyme (42 kDa) or GST (27 kDa) alone were also detected in the gel (B). We performed Dynamic Light Scattering (DLS) and Circular Dichroism (CD) spectroscopy analysis on the fusion proteins (Fig 5). The DLS data (Fig 5A-C) suggested the samples are heterogeneous (several peaks and high polydispersity). The hydrodynamic radius of the main peak is larger for the WT (Rh = 22.1 nm) and smaller for the mutants (H262F Rh = 15.8 nm; S270A Rh = 16.5 nm), which suggests that the WT forms larger aggregates than the mutants (Fig 5D). The CD spectra collected on the GST-fused WT and H262F and S270A mutant samples revealed no significant changes in the secondary structure of the proteins (Fig 5E).

9



B

GST-H262F DUSP5

R (nm) D

***

R (nm)

***

C

GST-S270A DUSP5

% Mass

% Mass

GST-WT DUSP5

% Mass

A

R (nm)

E

R (nm)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 22

WT

H262F

S270A

Fig 5. Characterizing proteins by dynamic Light Scattering (DLS) and circular dichroism spectra (CD). Representative DLS plot showing WT (A) or mutant (B-C) DUSP5 scattergram and aggregation. Hydrodynamic radius (R) was measured in triplicates, for all respective WT or mutant enzymes (D). The circular dichroism spectra of GST-WT, GSTH262F and GST-S270A samples (E). ***p50% drop in the velocity of the enzymatic reaction (Fig 6B). To determine which steps in the catalytic mechanism of DUSP5 which substrate the mutations might influence, we developed a kinetic model of pNPP hydrolysis by DUSP5 (Scheme 2). This model was largely based on the well-studied catalysis reaction of another closely related phosphatase, VH1-related kinase (VHR)29. DUSP5, like other DUSPs and protein tyrosine phosphatases including VHR, shares the active-site HCXXXXXRS motif (Table 3)18, 29. The catalytic mechanism of DUSPs was shown to go through a formation of a thiol-phosphate intermediate, which is subsequently hydrolyzed to release orthophosphate. Thus, there are two distinct covalent catalytic events; phosphorylation of enzyme (k3 or cysteine nucleophile) and dephosphorylation of enzyme (k5 or water nucleophile). Scheme 2. Kinetic model of pNPP hydrolysis by DUSP5. pNPP= para-nitrophenol phosphate, pNP= para-nitrophenol, Pi= orthophosphate

10


Page 11 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 3. Summary of the experimental DUSP structures Name

PDB entry

catalytic motif (from pdb) a

Anion Type b

H-C bonding c

P-loop distorted? d

R(S/O...A–), Å e

DUSP2

1m3g f

HSQAGISRS

-

-

Yes

-

DUSP3

1vhr

HCREGYSRS

S

Yes

-

3.6

DUSP4

3ezz

HSQAGISRS

S

-

-

3.5

DUSP5

2g6z

HSEAGISRS

S

-

-

3.7

DUSP6

1mkp

HSLAGISRS

Cl

-

Yes

3

DUSP7

4y2e

HSLAGISRS

P

Yes

-

3.4

DUSP8

4jmk

HSLAGISRS

S

Yes

-

3.2

DUSP9

2hxp

HSLAGVSRS

P

Yes

-

3.4

DUSP9

3lj8

HCLAGVSRS

-

-

Yes

-

DUSP10

2oud

HCQAGVSRS

Cl

Yes

-

3.3

DUSP10

1zzw

HCQAGVSRS

S

Yes

-

3.6

DUSP11

4jmj

HSTHGLNRT

P

Yes

-

3.3

DUSP12

4ki9

HSHAGVSRS

P

Yes

-

3.6

DUSP13b

2gwo

HCAMGVSRS

-

Yes

-

-

DUSP14

2wgp

HCAAGVSRS

P

Yes

-

3.7

DUSP15

1yz4

HSFAGISRS

S

-

-

3.5

DUSP18

2esb

HCAAGVSRS

A

Yes

-

3.6

DUSP19

3s4e

HSNAGVSRA

S

Yes

-

3.3

DUSP19

4d3p

HANAGVSRA

S

Yes

-

-

DUSP22

1wrm

HCLAGVSRS

MES

Yes

-

3.6

DUSP22 E24A/K28A/ K30A

4woh

HSLAGVSRS

P

Yes

-

3.4

DUSP23a

4erc

HCALGFGRT

V

Yes

-

2.4

DUSP23a

2img

HCALGFGRT

MLT

Yes

-

3.2

DUSP26

2e0t

HCAVGVSRS

-

-

Yes

-

DUSP26

4b04

HSAVGVSRS

-

-

Yes

-

DUSP26

4hrf

HSAVGVSRS

-

-

Yes

-

DUSP26

5gtj

HSAVGVSRS

P

Yes

-

3.3

DUSP27

2y96

HCVMGRSRS

S

Yes

-

-

a

Underscored is the position of the catalytic cysteine; S denotes sulfate, Cl–chloride, P–phosphate, A–acetate, MES–2-(N-morpholino)-ethanesulfonate, V–vanadate, and MLT–D-malate anion; c Presence of the short-range His(-1)-N...O=C-Cys cat contact, based on the criterion R(N...O) ≤ 3.2 Å; d 'Yes' if the conformation of the P-loop was dissimilar from that in majority of DUSPs; e Distance between the sulfur (or oxygen) atom of the catalytic cysteine (or serine) and the central atom of the anion (e.g. S in case of sulfate); f Obtained from NMR spectroscopy, the rest of the structures were obtained by X-ray crystallography. b

11



A Full length DUSP5

DUSP5 PD(WT)

GST-H262F

3.5e-8

2.5e-8

GST-S270A

2.0e-8

1.5e-9

3.0e-8

1.0e-8 Trial 1 Trial 2 Trial 3 Trial 4 Michaelis-Menten fit

5.0e-9

10

20

30

40

1.5e-8 1.0e-8

1.0e-8

5.0e-9

0.0

50

0.0 0

10

pNPP (mM)

20

30

40

50

9.0e-10 6.0e-10 3.0e-10

5.0e-9

0.0 0

2.0e-8

1.2e-9

1.5e-8

Velocity (M/s)

1.5e-8

2.5e-8

Velocity (M/s)

Velocity (M/s)

2.0e-8

Velocity (M/s)

0.0 0

10

pNPP (mM)

20

30

40

50

pNPP (mM)

0

10

20

50

pNPP (mM)

B

Velocity (M/s)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 22

Fig 6. Reaction velocities in pNPP assay. Reaction velocities were monitored as a function of pNPP concentration for WT versus mutant DUSP5s. Four trials were performed for each protein (A). Lines represent non-linear regression least square fits to the Michaelis-Menten equation. Mean velocities for each enzymatic reaction were also compared (B). pNPP (mM)

The kinetic studies (Fig 7A) showed that the S270A mutation reduced dramatically (~16-fold) the kcat constant of DUSP5, but had little affect on kcat/Km. Km, which is defined as (kr + kcat)/kf, also measured kcat/Km, the second order rate constant indicating the catalytic efficiency of the enzyme. kcat/Km reflects only steps of DUSP5 enzyme (E) to substrate (S) pNPP-binding and the first chemical reaction, which is phosphorylation of the enzyme to form an enzyme-S-phosphate intermediate (k3 in Scheme 1). While the S270A mutation decreased kcat, it had no significant effect on kcat/Km (Fig 7B) suggesting that S270 residue catalyzes a kinetic step that occurs after substrate binding (k1) and the first irreversible step (k3), and yet is important for catalysis (k5 within kcat). By analogy to the previously published mechanism29, S270 is likely acting as a general acid to catalyze cleavage of the phospho-enzyme intermediate (k5, Scheme 2). Further, general acids and general bases are important for enzyme kinetics catalysis in phosphatases29, because they are influenced by ionization states of active site amino acid residues, which is a pH-dependent phenomenon. When we measured the initial velocities (kcat and kcat/Km values) of WT DUSP5 PD over a range of pNPP concentrations at various pH values (pH 5-9), we observed that kcat (Fig S3A) and kcat/Km (Fig S3B) generated bell-shaped curves over a wide range of pH values. The pH-dependency of kcat and kcat/Km values of WT DUSP5 PD are consistent with ionizable groups that must be deprotonated for activity, and others that must be protonated. The kcat/Km pH profile is bell-shaped and consistent with a “reverse protonation” ionization state preferred for substrate binding and catalysis, with Asp232 protonated and Cys263 deprotonated (perturbed pKa values of 5.5 and 5.7), in WT DUSP5 PD. This is consistent with the mechanism in Scheme 1, where the Cys263 thiolate anion attacks substrate (k3).

12


Page 13 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Fig 7. Kinetics for enzymatic activities of DUSP5. Michaelis-Menten kinetics were monitored for WT DUSP5 versus various mutants (H262F and S270A), using para-nitrophenylphosphate (pNPP) as a model substrate. Table showing various parameters as monitored during pNPP assay (A). Data represents mean ± SEM; n=4. kcat / Km was monitored for WT versus mutant DUSP5, as a function of catalytic activity of various enzymes (B). Final DUSP5 PD(WT), Full length DUSP5, GST-H262F, and GST-S270A assay concentrations were 0.97, 0.92, 0.84, and 0.71 µM, respectively. The effect of our S270A mutation, which is pronounced on kcat but not on kcat/Km, is consistent with the mutation affecting a step that occurs after the first irreversible step, indicating that the effect is likely on k5 (Scheme 2). A similar mutation (S131A) in the VHR enzyme has been characterized before 29 and it was concluded that the mutant greatly impacted the ability of VHR to hydrolyze the thiol-phosphorylated enzyme intermediate, shifting the rate-limiting step from phospho-intermediate formation to phosphointermediate hydrolysis. These findings, combined with the proximity and favorable mutual orientation of Cys263 and Ser270 for the formation of hydrogen bonding between them, endorse the current understanding that Ser270 serves to stabilize the thiolate anion of the catalytic cysteine as needed to enable efficient hydrolysis of the phosphorylated cysteine intermediate. We also measured the catalytic activity of the purified GST-tagged H262F point mutant DUSP5 protein towards pNPP substrate (Fig 7A). We observed only a moderate ~50% drop in mean kcat as compared with wild-type DUSP5, combined with ~ 3-fold increase in Km, which is dissimilar to the effect observed for the S270A mutant. In contrast to S270A, the H262F mutation has a profound effect on the second order rate constant kcat/Km (Fig 7B). This dramatic effect on kcat/Km but not kcat suggests that this mutation impairs the ability of the PD to bind with and dephosphorylate its substrate, affecting steps k1 and k3. In summary, H262F lowers mostly catalytic efficiency (kcat/Km) because it affects either substrate binding or dephosphorylation (k3), but it has little effect on the rate of dephosphorylation of the enzyme-Sphosphate intermediate. In contrast, the S270A mutation does not affect these early steps, but rather affects the rate of enzyme-S-phosphate dephosphorylation (k5); and, this mutation now makes k5 the rate limiting step. A major shift in kcat/Km, along with minimal impact on kcat alone, which was observed in H262 compared to S270 suggests that H262F mutation influences steps up to the first irreversible step in the kinetics (Scheme 2). The k constants in the equation for kcat/Km dictate this interpretation. This step could be the binding step and/or formation of the phospho-enzyme intermediate due to the attack of the phosphorylated pNPP substrate by catalytic cysteine (k1 or k3 in Scheme 2). Our proposed kinetic model (Scheme 2) suggests that H262 might affect earlier steps in the mechanism, whereas S270 might affect later steps in the catalytic reaction. Such a disparity in the effects of the H262F vs S270A mutations show that His262 and Ser270 utilize distinct mechanisms to control the catalytic activity of DUSP5. Comparing catalytic activity of of S270A, H262F to WT DUSP5 using physiological substrate ERK We next investigated S270A and H262F mutant proteins effect on the dephosphorylation of the physiological substrate of DUSP5, pERK. We measured the enzymatic activity of WT versus mutant DUSP5s by monitoring the dephosphorylation of ERK in western blots (Fig 8). Total ERK to pERK ratio was considered as function of enzymatic activity. Higher the catalytic activity of enzyme; higher will be the conversion of pERK to ERK. So increased ratio of ERK:pERK would indicate increased catalytic activity of the enzyme. Although we did not observe any difference of ERK:pERK within the groups at higher dilutions (10-6 μM); the ratio trended higher for H262F compared to WT or S270A mutant at 10-5 μM and significantly higher at 10-4 μM (Fig 8A & C). Enzymatic activity at 0 min was considered as background, which was consistent within the groups (Fig 8B). When normalized against background and ‘no enzyme’ internal control; H262F showed ~50% increased enzymatic activity, compared to WT and S270A (Fig 8D). To understand the implications of this finding, we constructed a step-by-step 13



dephosphorylation sequence reaction of DUSP5, starting with its binding to pERK in the nucleus, which is largely based on information from the literature, our MD simulation studies, and previous data from our lab.

A

B

D

C

Fig 8. Representative blot showing dephosphorylation of recombinant ERK protein by various WT or mutant DUSP5 enzymes at graded concentrations as indicated (A). ERK to pERK ratio was considered as the read out of enzymatic activity and was shown as relative to ‘no enzyme’ control. Kinetics of relative enzymatic activities were monitored at 0 (B) or 15 min (C). For each enzyme, the assay was replicated at least 3 times. Activity of various enzymes at 15 min was also normalized to 0 min (D). *p

Role of Conserved Histidine and Serine in the ... - ACS Publications

Recommend Documents