Graduate Category: Physical & Life Science Degree Level: Ph.D ...
Graduate Category: Physical & Life Science Degree Level: Ph.D. Abstract ID# 708
Ornithine Transcarbamylase Has a Spatially Extended Active Site
Third shell Second shell First shell 5Å 5Å
Lisa Ngu, Kevin E. Ramos, Nicholas A. DeLateur, Penny J. Beuning, and Mary Jo Ondrechen Department of Chemistry & Chemical Biology ABSTRACT
RESULTS
Partial Order Optimum Likelihood (POOL), developed at Northeastern University, is a machine learning technique to predict catalytically important residues based on the tertiary structure of a protein and on computed electrostatic properties. POOL has been shown to predict accurately the catalytic residues and discern between a compact and extended active site in enzymes. Classic studies of enzymes typically identify catalytic residues that are in direct contact with the substrate. POOL predicts a spatially extended active site for Escherichia coli ornithine transcarbamylase (OTC), for which dynamic processes are believed to play a role in its catalytic mechanism. OTC is reported to undergo induced-fit conformational changes upon binding carbamoyl phosphate, followed by binding of ornithine. POOL predicts OTC to have an extended triple-layer active site. Kinetics assay of Asp140, His272, Glu299 and Arg57 variants show these POOL-predicted remote residues, located in the second and third layers, are important for catalysis. Specifically, variants D140N, H272L, E299Q, Y160S and R57A (positive control) have reduced catalytic efficiency relative to wild type for ornithine, and all but Y160S have reduced activity relative to wild type for carbamoyl phosphate. Negative controls, which are mutations of conserved residues not predicted by POOL to be catalytically important, were S61A and Q104L and had insignificant differences in catalytic efficiencies for ornithine and carbamoyl phosphate relative to wild type. Our observations indicate the importance of remote residues for OTC activity and the power of POOL to predict accurately the catalytically important residues.
OTC Mechanism and Kinetics Assay11:
Carbamoyl phosphate
Wild-type OTC Kinetics: Carbamoyl Phosphate
Structure
Charge
THEMATICS Electrostatic Properties
INTREPID
Geometric Features
Phylogenic Score
POOL
Machine Learning
pH
Functional Residue Prediction
Normalized POOL Score
POOL: OTC HAS TRIPLE LAYER ACTIVE SITE 1.0 0.8 0.6 0.4 0.2 0.0
0.2 0.1
OTC Normalized POOL Scores
0.2 0.1 0.0
0.5
1.0
0
1.5
70 65
1
2
Catalytic Efficiency (x105 M-1min-1)
700 600 500 400 300 200 100 0
OTC Thermofluor Apo Enzyme + Carbamoyl Phosphate
60 55 50 45
- OTC variants had decreased Tm with differences up to 18 ᵒC, but kinetics assays were carried out at a temperature that was well below these melting temperatures - Carbamoyl Phosphate stabilizes OTC (Tm increased up to 5 ᵒC)
• • • •
20
30
POOL Rank
Residues with normalized POOL scores > 0.01 are predicted to be catalytically important
Ornithine Transcarbamoylase from E. coli (PDB entry 1DUV)9 Negative Controls: Low POOL score & High ConSurf Score10 PSQ: Nδ-(N'-Sulphodiaminophosphinyl)-L-ornithine, transition state analog
2.5
Carbamoyl Phosphate Ornithine
2.0 1.5 1.0 0.5 0.0
Carbamoyl Steady-State Kinetic ParametersPhosphate of OTC and Its Variants with RespectOrnithine to Carbamoyl Phosphate Shell kcat (x102 s-1) KM (μM) kcat/KM (x105 M-1s-1) Relative kcat Relative kcat/KM Wild---57 ± 23 0.10 ± 0.042 590 ± 86 1 1 type C273A first 140 ± 20 0.52 ± 0.051 260 ± 14 2.4 0.44 D140N second 33 ± 4 0.17 ± 0.018 200 ± 13 0.56 0.33 Y229F second 140 ± 15 0.58 ± 0.073 240 ± 62 2.4 0.41 H272L second 2.1 ± 0.32 0.059 ± 0.015 36 ± 3.9 0.037 0.06 H272N second 54 ± 6.9 0.092 ± 0.015 600 ± 140 0.95 1 E299D second 190 ± 73 0.37 ± 0.067 530 ± 53 3.3 0.89 E299Q second 2.9 ± 0.27 0.078 ± 0.031 40 ± 10 0.051 0.067 Y160S third 82 ± 9.5 0.15 ± 0.0001 550 ± 64 1.4 0.93 Y160F third 100 ± 5.7 0.48 ± 0.098 210 ± 36 1.8 0.36 R57A (+) third 1.2 ± 0.076 0.064 ± 0.026 20 ± 5.5 0.02 0.033 S61A (-) second 230 ± 17 0.17 ± 0.025 1400 ± 140 4.1 2.3 Q104L (-) third 72 ± 8.6 0.092 ± 0.012 790 ± 170 1.3 1.3 Steady-State Kinetic Parameters of OTC and Its Variants with Respect to Ornithine Shell kcat (x102 s-1) KM (μM) kcat/KM (x105 M-1s-1) Relative kcat Relative kcat/KM Wild---58 ± 23 0.35 ± 0.11 170 ± 52 0 1 type C273A first 130 ± 18 0.53 ± 0.11 260 ± 27 2.3 1.5 D140N* second 130 ± 18 9.2 ± 1.6 14 ± 2.5 2.2 0.083 Y229F second 130 ± 28 0.42 ± 0.027 320 ± 45 2.3 1.9 H272L second 2.6 ± 0.40 0.59 ± 0.055 4.5 ± 0.5 0.045 0.027 H272N second 58 ± 10 0.46 ± 0.082 130 ± 7.9 1 0.75 E299D second 180 ± 64 0.65 ± 0.066 280 ± 96 3.2 1.7 E299Q second 2.8 ± 0.33 0.35 ± 0.029 8.2 ± 1.6 0.048 0.049 Y160S* third 160 ± 32 4.1 ± 0.26 40 ± 5.3 2.8 0.23 Y160F third 96 ± 11 0.46 ± 0.12 220 ± 32 1.6 1.3 R57A (+) third 1.5 ± 0.10 0.45 ± 0.14 3.6 ± 1.3 0.026 0.021 S61A second(-) negative 230 ± 10controls, 0.40 ±(+) 0.0032 570 ± 29 3.9 3.4 * KM (-) apparent, positive control Q104L (-) third 82 ± 17 0.28 ± 0.010 290 ± 48 1.4 1.7 Reduced catalytic efficiency relative to wild-type
CONCLUSION Distal residues Arg 57, Asp 140, Tyr 160, His 272 & Glu 299 play a significant role in catalysis, as predicted Charge on Glu 299 and polarity of His 272 are important Understanding how nature builds enzyme active sites has important implications for protein engineering Future work includes SAXS and MD simulations of variants to probe conformational changes
• NSF: MCB-0843603 & MCB-1158176 (MJO, PJB); NSF-REU & LSAMP (KR) 10
4
OTC (Ornithine) Catalytic Efficiency
Funding and Acknowledgments 0
3
Ornithine (mM)
OTC (Carbamoyl Phosphate) Catalytic Efficiency
75
Sequence
Concavity
Velocity (mM/min)
Velocity (mM/min)
3D Structure
3.0
0.3
Carbamoyl Phosphate (mM)
1800 1600 1400 1200 1000 800 600 400 200 0
3.5
OTC Relative Catalytic Efficiency
0.4
0.3
0.0 0.0
Melting Temperature (◦C)
THEMATICS: Theoretical Titration Curves Not an Active-Site residue Active-Site residues
inorganic phosphate
Wild-type OTC Kinetics: Ornithine
0.4
Catalytic Efficiency (x105 M-1min-1)
POOL and THEMATICS (1) accurately predict catalytic residues and (2) discern between compact and extended active sites. Predictions biochemically confirmed include: • Compact active sites: alkaline phosphatase5 and ketosteroid isomerase.6 • Extended active sites: phosphoglucose isomerase6, nitrile hydratase7, and DinB8.
citrulline
2,3-butanedione monoxime (in the presence of acidic condition & iron (III) chloride) absorbance read at 530 nm (pink color formation)
BACKGROUND POOL (Partial Order Optimum Likelihood) is a machine learning computational technique that predicts important catalytic residues based on the 3D structure of a protein.1 Input features: (1) THEMATICS, predicts catalytic residues by calculating theoretical microscopic titration curves of ionizable residues.1, 4 (2) INTREPID, identifies functionally important residues based on phylogenic tree information2 (3) ConCavity, predicts catalytic sites based on surface geometric properties of a protein structure3
ornithine
Relative Catalytic Efficiency
Substrate
References 1. Tong, W., et al. (2009) PLoS Comput Biol, 5(1), e1000266. 2. Sankararaman, S., et al. (2009) Nucleic Acids Res, 37(suppl 2), W390. 3. Capra, J.A., et al. (2009) PLoS Comput Biol, 5(12), e1000585. 4. Wei, Y., Ko, et al. (2007) BMC Bioinformatics, 8,119. 5. Brodkin, H. et al. (2015) Protein Science, in press, doi: 10.1002/pro.2648. 6. Somarowthu, S., et al. (2011) Biochemistry,50(43), 9283. 7. Brodkin, H.R et al. (2011) Biochemistry, 50(22), 4923. 8. Walsh, J., et al. (2012) Env Mol Mut, 53(9), 766. 9. Langley, D. B. , et al. (2000) J. Biol. Chem. 275(26) 20012. 10.Landau, M. et al. (2005) Nucl. Acids Res. 33: W299-W302. 11.Boyde, T. R. C. and Rahmatullah, M. (1980) Anal Biochem, 107(2), 424.
Ornithine Transcarbamylase Has a Spatially Extended Active Site
Third shell Second shell First shell 5Å 5Å
Lisa Ngu, Kevin E. Ramos, Nicholas A. DeLateur, Penny J. Beuning, and Mary Jo Ondrechen Department of Chemistry & Chemical Biology ABSTRACT
RESULTS
Partial Order Optimum Likelihood (POOL), developed at Northeastern University, is a machine learning technique to predict catalytically important residues based on the tertiary structure of a protein and on computed electrostatic properties. POOL has been shown to predict accurately the catalytic residues and discern between a compact and extended active site in enzymes. Classic studies of enzymes typically identify catalytic residues that are in direct contact with the substrate. POOL predicts a spatially extended active site for Escherichia coli ornithine transcarbamylase (OTC), for which dynamic processes are believed to play a role in its catalytic mechanism. OTC is reported to undergo induced-fit conformational changes upon binding carbamoyl phosphate, followed by binding of ornithine. POOL predicts OTC to have an extended triple-layer active site. Kinetics assay of Asp140, His272, Glu299 and Arg57 variants show these POOL-predicted remote residues, located in the second and third layers, are important for catalysis. Specifically, variants D140N, H272L, E299Q, Y160S and R57A (positive control) have reduced catalytic efficiency relative to wild type for ornithine, and all but Y160S have reduced activity relative to wild type for carbamoyl phosphate. Negative controls, which are mutations of conserved residues not predicted by POOL to be catalytically important, were S61A and Q104L and had insignificant differences in catalytic efficiencies for ornithine and carbamoyl phosphate relative to wild type. Our observations indicate the importance of remote residues for OTC activity and the power of POOL to predict accurately the catalytically important residues.
OTC Mechanism and Kinetics Assay11:
Carbamoyl phosphate
Wild-type OTC Kinetics: Carbamoyl Phosphate
Structure
Charge
THEMATICS Electrostatic Properties
INTREPID
Geometric Features
Phylogenic Score
POOL
Machine Learning
pH
Functional Residue Prediction
Normalized POOL Score
POOL: OTC HAS TRIPLE LAYER ACTIVE SITE 1.0 0.8 0.6 0.4 0.2 0.0
0.2 0.1
OTC Normalized POOL Scores
0.2 0.1 0.0
0.5
1.0
0
1.5
70 65
1
2
Catalytic Efficiency (x105 M-1min-1)
700 600 500 400 300 200 100 0
OTC Thermofluor Apo Enzyme + Carbamoyl Phosphate
60 55 50 45
- OTC variants had decreased Tm with differences up to 18 ᵒC, but kinetics assays were carried out at a temperature that was well below these melting temperatures - Carbamoyl Phosphate stabilizes OTC (Tm increased up to 5 ᵒC)
• • • •
20
30
POOL Rank
Residues with normalized POOL scores > 0.01 are predicted to be catalytically important
Ornithine Transcarbamoylase from E. coli (PDB entry 1DUV)9 Negative Controls: Low POOL score & High ConSurf Score10 PSQ: Nδ-(N'-Sulphodiaminophosphinyl)-L-ornithine, transition state analog
2.5
Carbamoyl Phosphate Ornithine
2.0 1.5 1.0 0.5 0.0
Carbamoyl Steady-State Kinetic ParametersPhosphate of OTC and Its Variants with RespectOrnithine to Carbamoyl Phosphate Shell kcat (x102 s-1) KM (μM) kcat/KM (x105 M-1s-1) Relative kcat Relative kcat/KM Wild---57 ± 23 0.10 ± 0.042 590 ± 86 1 1 type C273A first 140 ± 20 0.52 ± 0.051 260 ± 14 2.4 0.44 D140N second 33 ± 4 0.17 ± 0.018 200 ± 13 0.56 0.33 Y229F second 140 ± 15 0.58 ± 0.073 240 ± 62 2.4 0.41 H272L second 2.1 ± 0.32 0.059 ± 0.015 36 ± 3.9 0.037 0.06 H272N second 54 ± 6.9 0.092 ± 0.015 600 ± 140 0.95 1 E299D second 190 ± 73 0.37 ± 0.067 530 ± 53 3.3 0.89 E299Q second 2.9 ± 0.27 0.078 ± 0.031 40 ± 10 0.051 0.067 Y160S third 82 ± 9.5 0.15 ± 0.0001 550 ± 64 1.4 0.93 Y160F third 100 ± 5.7 0.48 ± 0.098 210 ± 36 1.8 0.36 R57A (+) third 1.2 ± 0.076 0.064 ± 0.026 20 ± 5.5 0.02 0.033 S61A (-) second 230 ± 17 0.17 ± 0.025 1400 ± 140 4.1 2.3 Q104L (-) third 72 ± 8.6 0.092 ± 0.012 790 ± 170 1.3 1.3 Steady-State Kinetic Parameters of OTC and Its Variants with Respect to Ornithine Shell kcat (x102 s-1) KM (μM) kcat/KM (x105 M-1s-1) Relative kcat Relative kcat/KM Wild---58 ± 23 0.35 ± 0.11 170 ± 52 0 1 type C273A first 130 ± 18 0.53 ± 0.11 260 ± 27 2.3 1.5 D140N* second 130 ± 18 9.2 ± 1.6 14 ± 2.5 2.2 0.083 Y229F second 130 ± 28 0.42 ± 0.027 320 ± 45 2.3 1.9 H272L second 2.6 ± 0.40 0.59 ± 0.055 4.5 ± 0.5 0.045 0.027 H272N second 58 ± 10 0.46 ± 0.082 130 ± 7.9 1 0.75 E299D second 180 ± 64 0.65 ± 0.066 280 ± 96 3.2 1.7 E299Q second 2.8 ± 0.33 0.35 ± 0.029 8.2 ± 1.6 0.048 0.049 Y160S* third 160 ± 32 4.1 ± 0.26 40 ± 5.3 2.8 0.23 Y160F third 96 ± 11 0.46 ± 0.12 220 ± 32 1.6 1.3 R57A (+) third 1.5 ± 0.10 0.45 ± 0.14 3.6 ± 1.3 0.026 0.021 S61A second(-) negative 230 ± 10controls, 0.40 ±(+) 0.0032 570 ± 29 3.9 3.4 * KM (-) apparent, positive control Q104L (-) third 82 ± 17 0.28 ± 0.010 290 ± 48 1.4 1.7 Reduced catalytic efficiency relative to wild-type
CONCLUSION Distal residues Arg 57, Asp 140, Tyr 160, His 272 & Glu 299 play a significant role in catalysis, as predicted Charge on Glu 299 and polarity of His 272 are important Understanding how nature builds enzyme active sites has important implications for protein engineering Future work includes SAXS and MD simulations of variants to probe conformational changes
• NSF: MCB-0843603 & MCB-1158176 (MJO, PJB); NSF-REU & LSAMP (KR) 10
4
OTC (Ornithine) Catalytic Efficiency
Funding and Acknowledgments 0
3
Ornithine (mM)
OTC (Carbamoyl Phosphate) Catalytic Efficiency
75
Sequence
Concavity
Velocity (mM/min)
Velocity (mM/min)
3D Structure
3.0
0.3
Carbamoyl Phosphate (mM)
1800 1600 1400 1200 1000 800 600 400 200 0
3.5
OTC Relative Catalytic Efficiency
0.4
0.3
0.0 0.0
Melting Temperature (◦C)
THEMATICS: Theoretical Titration Curves Not an Active-Site residue Active-Site residues
inorganic phosphate
Wild-type OTC Kinetics: Ornithine
0.4
Catalytic Efficiency (x105 M-1min-1)
POOL and THEMATICS (1) accurately predict catalytic residues and (2) discern between compact and extended active sites. Predictions biochemically confirmed include: • Compact active sites: alkaline phosphatase5 and ketosteroid isomerase.6 • Extended active sites: phosphoglucose isomerase6, nitrile hydratase7, and DinB8.
citrulline
2,3-butanedione monoxime (in the presence of acidic condition & iron (III) chloride) absorbance read at 530 nm (pink color formation)
BACKGROUND POOL (Partial Order Optimum Likelihood) is a machine learning computational technique that predicts important catalytic residues based on the 3D structure of a protein.1 Input features: (1) THEMATICS, predicts catalytic residues by calculating theoretical microscopic titration curves of ionizable residues.1, 4 (2) INTREPID, identifies functionally important residues based on phylogenic tree information2 (3) ConCavity, predicts catalytic sites based on surface geometric properties of a protein structure3
ornithine
Relative Catalytic Efficiency
Substrate
References 1. Tong, W., et al. (2009) PLoS Comput Biol, 5(1), e1000266. 2. Sankararaman, S., et al. (2009) Nucleic Acids Res, 37(suppl 2), W390. 3. Capra, J.A., et al. (2009) PLoS Comput Biol, 5(12), e1000585. 4. Wei, Y., Ko, et al. (2007) BMC Bioinformatics, 8,119. 5. Brodkin, H. et al. (2015) Protein Science, in press, doi: 10.1002/pro.2648. 6. Somarowthu, S., et al. (2011) Biochemistry,50(43), 9283. 7. Brodkin, H.R et al. (2011) Biochemistry, 50(22), 4923. 8. Walsh, J., et al. (2012) Env Mol Mut, 53(9), 766. 9. Langley, D. B. , et al. (2000) J. Biol. Chem. 275(26) 20012. 10.Landau, M. et al. (2005) Nucl. Acids Res. 33: W299-W302. 11.Boyde, T. R. C. and Rahmatullah, M. (1980) Anal Biochem, 107(2), 424.
