無(wú)底物情況下來(lái)白R(shí)hoclococcus zopfii的腈水解酶中親核進(jìn)攻試劑CYS165的活性狀態(tài)的探究（英文）

關(guān)鍵詞：ONIOM （the Own n?Layered Integrated Molecular Orbital and Molecular Mechanics），腈水解酶，能壘，催化機(jī)制

中圖分類號(hào)：O643 文獻(xiàn)標(biāo)志碼：A

Abstract： Nitrilase，as a class of green biocatalysts of great value in industrial applications，can efficiently catalyze thehydrolysis of nitrile compounds into carboxylic acids. Despite its wide application，the specific catalytic mechanism of nitrilaseremains elusive. Previous studies have revealed that the GLU?LYS?GLU?CYS tetrad in the active center of nitrilase plays apivotal role in the catalysis，where the CYS residue acts as a nucleophile attacking the nitrile，and the ionization of its thiolgroup is a key step in the reaction. However，the process of deprotonation of CYS has not been clearly illustrated. This studyfocuses on the nitrilase from Rhodococcus zopfii （RzNIT） and investigates the protonation state of CYS165 when thesubstrate has not yet entered the enzyme′s active site. Through detailed analysis of possible pathways for CYS165deprotonation，it is confirmed that CYS165 in RzNIT is in its neutral state in the absence of substrate. This finding lays thefoundation for further studies on the catalytic mechanism of RzNIT.

Keywords： ONIOM （the Own n ?Layered Integrated Molecular Orbital and Molecular Mechanics），nitrilase，energy barrier，catalytic mechanism

Nitrilase （EC 3. 5. 5. 1），a key green bioca ?talyst in industry，hydrolyzes nitriles to carboxylicacids and ammonia［1-2］. Nitrile compounds pose aserious threat to human health due to their high toxi?city and potential carcinogenicity［3］. They are wide?ly present in wastewater generated from industrialactivities such as mining，refining，and automobilemanufacturing［4］. Microbial degradation，especiallyhydrolysis by nitrilase，offers an effective methodfor removing highly toxic nitriles from industrialwastewater ［5-8］. Furthermore，the carboxylic acidsproduced by nitrilase hydrolysis，such as （R）?man?delic acid［9-10］，nicotinic acid［11］，and 2?chloronico?tinic acid［12-13］，are key intermediates in the synthe?sis of many compounds and pharmaceuticals［2，12］.Nitrilase，with its mild reaction conditions，environ?ment ? friendly，and sustainable characteristics，hasbecome a biocatalyst of significant commercialvalue［14］. Its potential applications span from envi?ronmental management to industrial manufactur?ing，representing an essential shift towards greenchemistry and sustainable industrial processes［15］.

Nitrilases are widely distributed in nature，in?cluding in plants，animals，and microbes［16］. Basedon the analysis of nitrilase sequence similarity andfunctional domains，Nitrilases have been classifiedinto thirteen different branches［16-17］. The nitrilasesof the first branch mainly hydrolyze nitriles to am ?monia and carboxylic acids，while other branchesexhibit characteristics of hydrolyzing or synthesiz?ing amides［17］. Notably，some nitrilases in the firstbranch not only show hydrolysis activity but alsohydration activity，capable of generating carboxylicacids or amides. For example，the nitrilase fromPseudomonas fluorescens EBC191 can catalyze thetransformation of （S）? mandelonitrile to mande?loamide and （S）? mandelic acid［18-19］. Meanwhile，the nitrilase from Rhodococcus zopfii （RzNIT）demonstrates high catalytic activity towards 2?chlo?ronicotinonitrile，with hydratase activity playing adominant role［12］. The application of these nitrilas?es in actual production is limited by the difficulty inseparating by?products［20］. Additionally，due to thelack of detailed crystal structures and explanationsof catalytic mechanisms，designing enzyme struc?tures to balance catalytic efficiency and reactionspecificity faces challenges［21］. The advent of Al?phaFold in 2021 marked a significant advance?ment，as this artificial intelligence platform can pre?dict protein 3D structures with high precision［22］.

Nitrilases are commonly found as homopoly?mers in nature［23］. Studies on the crystal structuresof microbial nitrilases，such as the nitrilase fromPyrococcus abyssi （PDB：3KLC）［24］ and Synecho?cystis sp （PDB： 3WUY）［25］，have revealed that ni?trilases possess an α?β?β?α sandwich structure. Ad?ditionally，homodimers exhibit a superstructure ofα?β?β?α?α?β?β?α［23，26-28］. Research has shown thatthe GLU ?LYS?GLU ?CYS catalytic tetrad plays adecisive role in the enzyme′s catalytic activity［25，29］.Mutation experiments on various nitrilases haveshown that altering any residue in the catalytic tet?rad leads to a loss or significant reduction in en?zyme activity［30-32］. While these experiments un?derscore the importance of the catalytic tetrad，theprecise mechanism of catalysis remains unclear.The current understanding of the nitrilase catalyticmechanism is primarily based on experimental re?sults and analogy studies with other enzyme mecha?nisms. In the mechanism study of N ?carbamyl?D ?amino acid amidohydrolase，it was proposed thatthe nucleophilic attack and deprotonation of CYSoccur simultaneously［33］. Meanwhile，the study ofthe hydrolysis mechanism of mouse Nit2 （PDB：2W1V） in converting α ? ketosuccinamate （KSM）to oxaloacetate suggested that under physiologicalconditions with substrate，a steady?state process ex?ists，where the catalytic tetrad′s CYS can be in twostates：protonated and ionized. Being in the ionizedstate is a prerequisite for the nucleophilic attack［34］.The deprotonation of the CYS residue is a key stepin the nitrilase catalytic reaction，affecting the en?zyme′s activity and the substrate selectivity. Sincethe deprotonation of CYS is directly related to theenzyme catalytic mechanism，determination of thetiming of deprotonation is crucial for understandingthe enzyme efficiently.

The study aims to investigate the protonationstate of the CYS residue of the catalytic tetrad inRzNIT in the absence of a substrate. The dimericstructure and catalytic tetrad are shown in Fig. 1.

1 Methods

1. 1 AlphaFold2 prediction of RzNIT 3D struc?ture The sequence for the RzNIT （WP_138999863） was obtained by searching the NCBIdatabase （https：//www. ncbi. nlm. nih. gov/）.The dimeric structure of the RzNIT was predictedusing AlphaFold2［22］. Certain protein regions ofthe predicted structure of the full length have in?complete folds，as evidenced by Integrated Dis?tance Difference Test （IDDT） scores lower than40. Consequently，these regions were excised fromthe sequence，leaving only residues 1 to 310 for asecond round of prediction. The structure with thehighest IDDT score from the second predictionwas selected as the structure for subsequent molec?ular dynamics（ MD） simulations.

1. 2Molecular dynamics simulations The pro?tein was assigned the Amber ff19SB force field［35］The protein was placed in a cubic TIP3P waterbox［36］ with the distance between the protein andthe boundary of the box no less than 15 ?. Na+ions were added for charge neutralization. Thesystem was optimized using 500 steps of steepest"descent followed by 500 steps of conjugate gradi?ent optimization with the protein structure re?strained with a force constant of 10 kcal·mol-1·?-2on the backbone atoms. In the second stage of en?ergy minimization，the entire system was optimizedusing 500 steps of steepest descent followed by500 steps of conjugate gradient optimization with?out any restraint applied. Then，the temperaturewas gradually raised to 300 K in 100 ps，with aforce constant of 5 kcal·mol-1· ?-2 to the proteinbackbone atoms. A 100 ? ps simulations was per?formed under the NPT ensemble to further relaxthe system with the restraint removed. After that，a 200 ? ns production run was conducted under theNPT ensemble. The SHAKE algorithm［37］ wasused to constrain all bonds involving hydrogen at?oms. The integration time step was 2 fs. Temper?ature regulation was achieved using a Langevinthermostat［38］ with a collision frequency of 1. 0ps-1. Pressure regulation was managed by theBerendsen barostat，with long ? range electrostaticinteractions treated using particle mesh Ewald witha real ? space cutoff of 12 ?. All the simulationswere performed using Amber20［ 39］.

1. 3 ONIOM calculations Two representativestructures were extracted from the MD trajectory.The protein and the water molecules within 3 ?from the protein were preserved. Na+ cations wereadded using the leap module in AMBERTools toneutralize the charge. Following this treatment，the two structures contained 14038 and 14122atoms，respectively. These structures served asthe initial configurations for the ONIOM［40-42］ cal?culations using Gaussian16［43］. The quantum me?chanics（QM） region contains 45 atoms in total，en?compassing the side chains of GLU138，LYS131，GLU48，and CYS165，as well as two water mole?cules. One water molecule acts as a medium forproton transfer，while the other imitates the aque?ous environment surrounding the active site. Struc?ture optimization for the QM region was performedat the B3LYP/6 ? 31G（d） level of theory［44］. Themolecular mechanical （MM） region was modeledusing the Amber ff99SB force field parameters.An electrostatic embedding scheme was employedto implement the electrostatic polarization effect onthe QM region from the MM region.

Due to the complexity of the system，an inde?pendent optimization of the atoms within the QMregion was initially performed ，using a clustermodel to optimize the transition states. During thisprocess，main chain atoms were removed，the CBatoms of each residue were fixed ，and hydrogenatoms were added to saturate the CB atoms. Oncethe transition state structure was confirmed，it wasreintegrated into the initial structure for ONIOMcalculations. The transition state structure was fur?ther optimized under the condition that atoms be?yond 6 ? from the catalytic tetrad were frozen.Based on the optimized transition state structure，all atoms in the MM region were frozen，and thestructures of the transition state，reactant，and pro?duct were optimized again to minimize energy fluc?tuations caused by movements of atoms in the MMregion，ensuring the accuracy of the energy differ?ences between the QM region，reactant，product，and transition states. Finally，the correctness ofthe transition state structure and the feasibility ofthe reaction path were determined by calculatingthe intrinsic reaction coordinate（ IRC）.

2 Results and discussion

The AlphaFold 2 IDDT scores of the predict?ed structure is shown in Fig. 2. The predictedstructure of the complete sequence shows that resi?dues 311 to 366 are structureless，which are，there?fore，truncated in the second?round of structure pre?diction to avoid unrealisitc interaction between thisshort squence and the ordered region. Among thefive structures obtained in the second?round of pre?diction，the one with the highest confidence，therank_1 structure，was selected for the subsequentstudy.

The root mean square deviation （RMSD） of the protein backbone atoms along the 200 ns simu?lations is shown in Fig. 3. It was observed that dur?ing the first 60 ns the RMSD plateaued at 2. 0 ?，but it increased to 2. 5 ? at around 80 ns. At around 80 ns，an influx of water into the active sitewas observed， which caused the increase inRMSD. Despite this，the protein maintained rela?tively stable from 80 ns to 200 ns，corroboratingthe stability of RzNIT structure as predicted by Al?phaFold2.

Further analysis of the trajectory showed thatthe active site of the protein was enriched with wa?ter molecules. Especially a stable water molecule（WAT3030） was found within the catalytic tetrad，of which the Root Mean Square Fluctuation（RMSF） is 0. 50 ? within 120～200 ns. Thestructural characteristics of WAT3030 are show inFig. 4. The distance between the carboxyl oxygenatom of GLU48 and the oxygen atom ofWAT3030 centered at 2. 6 ? and the distance be?tween the carboxyl oxygen atom of GLU138 andthe oxygen atom of WAT3030 centered at 2. 7 ?.These distances indicate hydrogen bonding interac?tions between WAT3030 and the carboxyl groupsof GLU48 and GLU138，anchoring WAT3030centrally within the catalytic tetrad. The oxygenatom of WAT3030 is 3. 1 ? away from the sulfuratom of CYS165，indicating a relatively weak inter?action between them.

Additionally，GLU48 and GLU138 were iden?tified as the only amino acid residues potentially ca?pable of abstracting a proton from CYS165. Thedistance distribution between the sulfur atom of"CYS165 and the carboxyl oxygen atom ofGLU138 and the distance distribution between thesulfur atom of CYS165 and the carboxyl oxygen at?om of GLU48 are shown in Fig. 5. The distancebetween the sulfur atom of CYS165 and the car?boxyl oxygen atom of GLU138 centered at 5. 6 ?，and that between the sulfur atom of CYS165 andthe carboxyl oxygen atom of GLU48 centered at4. 6 ?，respectively. Both of them are larger than 3?，suggesting that proton transfer from CYS165 toGLU48 and GLU138 cannot occur directly. Giventhe discovery of a stable WAT3030 within the cat?alytic tetrad，it was hypothesized that the protontransfer is mediated through WAT3030. Thus，WAT3030，involved in the reaction，and a watermolecule representing the aqueous environmentwere included in the QM region. Two representa?tive structures were selected from the MD trajecto?ry. The QM regions of the two structures areshown in Fig. 6.

The QM region for the study of proton trans?fer from CYS165 to GLU138 is shown in Fig. 6a.In this structure，the distance between the oxygenatoms of WAT2900 and WAT3030 is 3. 1 ?.The distance between the sulfur atom of CYS165and the oxygen atom of WAT3030 is 3. 0 ?，andthe distance between carboxyl oxygen atom ofGLU138 and the oxygen atom of WAT3030 is2. 6 ?. These distances are consistent with theaforementioned distributions，indicating that thisstructure represents a characteristic structure with?in the MD trajectory. Additionally，the angle madeby the thiol hydrogen atom of CYS165 and the ox?ygen atom and one of the hydrogen atoms inWAT3030 is 109 degrees，which closely approxi?mates the bond angle of water molecules. This isconducive for the water ? bridged proton transferfrom CYS165 to GLU138. Therefore，this struc?ture has been selected for studying the proton trans?fer pathway from CYS165 to GLU138. The QMregion for the study of the proton transfer fromCYS165 to GLU48 is shown in Fig. 6b. The hy?drogen bonding network in this structure can ac?commodate the proton transfer from CYS165 toGLU48.

In the proton transfer from CYS165 toGLU138 via WAT3030，a concerted process wasconsidered. While the thiol hydrogen atom ofCYS165 is being transferred to WAT3030 ，a"hydrogen atom in WAT3030 is abstracted bycarboxyl group of GLU138 simultaneously. Thequasi ? stable structures for this reaction are shownin Fig. 7. In the reactant，the distance between thethiol hydrogen atom of CYS165 and the oxygenatom in WAT3030 is 1. 9 ?. The hydrogen bonddistance between WAT3030 and one of the oxy?gen atoms in the carboxyl group of GLU138 is1. 7 ?. The distance between the positivelycharged α?amino group in LYS131 and the carbox?yl group in GLU48 is 1. 5 ?（ hydrogen?oxygen dis?tance），forming a strong salt bridge that stabilizesthe entire structure. The hydrogen bond anglebetween CYS165 and WAT3030 and thatbetween WAT3030 and GLU138 are close to 180degrees，indicating a favorable configuration forproton transfer.

The transition state（ TS） has an imaginary vi?bration model of 611. 55 cm-1 linking the reactantand the product. In the transition state，the hydro?gen bond distance between CYS165 andWAT3030 decreases to 1. 1 ?，the bond length be?tween WAT3030 and GLU138 （hydrogen?oxygendistance） shrinks to 1. 2 ?. This structuralchange results in a tetrahedral configuration involv?ing the thiol hydrogen atom of CYS165 and all thethree atoms in WAT3030.

In the product structure，the hydrogen bondlength between WAT3030 and GLU48 is 1. 7 ?.The hydrogen bond length between the oxygen ofWAT3030 and the hydrogen transferred fromWAT3030 to GLU138 is 1. 4 ?. The stronghydrogen bond interaction between the side chainsof LYS131 and GLU48 still exists，stabilizing theentire product structure.

Energy calculations for the CYS165 protontransfer reaction yielded an energy barrier of 16. 29kcal·mol-1 between the transition state and reactant（shown in Fig. 8），and the product is 15. 88 kcal ·mol-1 higher in energy than that of the reactant.These calculations suggest that although the reac?tion is theoretically feasible from an energy per?spective，the significant energy difference betweenthe product and reactant renders the probability ofthe product of this reaction pathway extremely lowunder physiological conditions. This outcome indi?cates that GLU138 does not effectively abstract aproton from CYS165 in the absence of the sub?strate binding.

The water ? bridged proton transfer fromCYS165 to GLU48 also follows a concerted path?way. As the thiol hydrogen atom of CYS165 is"relocated to WAT3030，a hydrogen atom inWAT3030 is abstracted concurrently by the car?boxyl group in GLU48. The reactant，transitionstate and product structures of this reaction areshown in Fig. 9. In the reactant ，the distancebetween the thiol hydrogen atom of CYS165 andthe oxygen atom in WAT3030 is 1. 8 ?. Thehydrogen bond distance between WAT3030 andone of the oxygen atoms in the carboxyl group ofGLU48 is 1. 8 ?，and the distance between thepositively charged α ? amino group in LYS131 andthe carboxyl group in GLU48 is 1. 5 ? （hydrogen?oxygen distance），constructing a strong salt bridgethat stabilizes the entire structure. The hydrogenbond angle between CYS165 and WAT3030 andthat between WAT3030 and GLU48 are near 180degrees，suggesting a favorable configuration forproton transfer.

The transition state is characterized by animaginary vibration model of 895. 65 cm-1. In thetransition state，the hydrogen bond distance be?tween CYS165 and WAT3030 decreases to 1. 2 ?and the hydrogen bond length between WAT3030and GLU48 decreases to 1. 3 ?. The hydrogenbond distance between the side chains of LYS131and GLU48 increases to 1. 7 ?，denoting a weak?ening of the hydrogen bond interaction.

In the product，the hydrogen bond length be?tween WAT3030 and GLU138 is 1. 7 ?. The hy?drogen bond length between the oxygen ofWAT3030 and the hydrogen transferred fromWAT3030 to GLU48 is 1. 5 ?. The salt bridgedistance between LYS131 and GLU48 increasesto 1. 8 ?，denoting a further weakening of the Cou?lomb interaction.

The energy involved in the CYS165 protontransfer reaction were calculated （show inFig. 10）. The results show an energy barrier of10. 86 kcal·mol-1 between the transition state andreactant，and the product is 8. 04 kcal·mol-1 higher in energy than that of the reactant. These calcula?tions indicate that the probability of product forma?tion along this pathway is extremely low underphysiological conditions.

3 Summary

Upon analysis of the active site of the RzNITprotein，two pathways of water? mediated protontransfer from CYS165 to GLU residues nearbywere investigated. The computational results dem ?onstrated that in the absence of its substrate，the"nucleophilic agent （CYS165） of the RzNIT pro?tein remains in a neutral state under physiologicalconditions. These insights are important for under?standing the characteristics of the RzNIT protein′sactive site in a substrate?free state and its behaviorunder physiological conditions，laying the ground?work for further studies on the protein′s reactionmechanism and functional activity.

Acknowledgement CPU time was provided bythe Supercomputer Center of East China NormalUniversity （ECNU Public Platform for InnovationNo.001）.

References

［1］ Yamamoto K，Ueno Y，Otsubo K，et al. Productionof S ? （+ ） ? ibuprofen from a nitrile compound byAcinetobacter sp. Strain AK226. Applied andEnvironmental Microbiology，1990，56（10）：3125-3129.

［2］ Shen J D，Cai X，Liu Z Q，et al. Nitrilase：Apromising biocatalyst in industrial applications forgreen chemistry. Critical Reviews in Biotechnology，2021，41（1）：72-93.

［3］ Alvillo?Rivera A，Garrido?Hoyos S，Buitrón G，et al.Biological treatment for the degradation of cyanide：Areview. Journal of Materials Research andTechnology，2021，12：1418-1433.

［4］ Kuyucak N，Akcil A. Cyanide and removal optionsfrom effluents in gold mining and metallurgicalprocesses. Minerals Engineering， 2013， 50-51：13-29.

［5］ Gong J S，Lu Z M，Li H，et al. Nitrilases in nitrilebiocatalysis： Recent progress and forthcomingresearch. Microbial Cell Factories，2012，11（1）：142.

［6］ Dash R R，Gaur A，Balomajumder C. Cyanide inindustrial wastewaters and its removal：A review onbiotreatment. Journal of Hazardous Materials，2009，163（1）：1-11.

［7］ Menashe O， Raizner Y， Kuc M E， et al.Biodegradation of the endocrine ? disrupting chemical17α ? ethynylestradiol （EE2） by Rhodococcus zopfiiand Pseudomonas putida encapsulated in small"bioreactor platform （SBP） capsules. AppliedSciences，2020，10（1）：336.

［8］ Holtze M S，S?rensen S R，S?rensen J，et al.Microbial degradation of the benzonitrile herbicidesdichlobenil， bromoxynil and ioxynil in soil andsubsurface environments – insights into degradationpathways， persistent metabolites and involveddegrader organisms. Environmental Pollution，2008，154（2）：155-168.

［9］ Singh R V，Sambyal K. Green synthesis aspects of（R） ? （- ） ? mandelic acid; a potent pharmaceuticallyactive agent and its future prospects. Critical Reviewsin Biotechnology，2023，43（8）：1226-1235.

［10］ Zhang X H，Liu Z Q，Xue Y P，et al. Production of R?mandelic acid using nitrilase from recombinant E. colicells immobilized with tris （hydroxymethyl）phosphine. Applied Biochemistry and Biotechnology，2018，184（3）：1024-1035.

［11］ Mathew C D，Nagasawa T，Kobayashi M，et al.Nitrilase?catalyzed production of nicotinic acid from3 ? cyanopyridine in Rhodococcus rhodochrous J1.Applied and Environmental Microbiology，1988，54（4）：1030-1032.

［12］ Dai A D，Tang X L，Wu Z M，et al. Rationalregulation of reaction specificity of nitrilase forefficient biosynthesis of 2 ? chloronicotinic acidthrough a single site mutation. Applied andEnvironmental Microbiology，2022，88（5）：e0239721.

［13］ Jin L Q，Li Y F，Liu Z Q，et al. Characterization of anewly isolated strain Rhodococcus erythropolis ZJB ?09149 transforming 2 ? chloro ? 3 ? cyanopyridine to2 ? chloronicotinic acid. New Biotechnology，2011，28（6）：610-615.

［14］ Wohlgemuth R. Biocatalysis ? key to sustainableindustrial chemistry. Current Opinion in Biotech ?nology，2010，21（6）：713-724.

［15］ Alcántara A R，Domínguez De María P，Littlechild JA，et al. Biocatalysis as key to sustainable industrialchemistry. ChemSusChem，2022，15（9）：e202102709.

［16］ Pace H C，Brenner C. The nitrilase superfamily：Classification， structure and function. GenomeBiology，2001，2（1）：REVIEWS0001.

［17］ Brenner C. Catalysis in the nitrilase superfamily.Current Opinion in Structural Biology，2002，12（6）：775-782.

［18］ Sosedov O，Matzer K，Bürger S，et al. Constructionof recombinant Escherichia coli catalysts whichsimultaneously express an （S） ? oxynitrilase anddifferent nitrilase variants for the synthesis of （S） ?mandelic acid and （S） ? mandelic amide frombenzaldehyde and cyanide. Advanced Synthesis amp;Catalysis，2009，351（10）：1531-1538.

［19］ Sosedov O，Stolz A. Improvement of the amidesforming capacity of the arylacetonitrilase fromPseudomonas fluorescens EBC191 by site ? directedmutagenesis. Applied Microbiology andBiotechnology，2015，99（6）：2623-2635.

［20］ Han C，Yao P F，Yuan J，et al. Nitrilase ? catalyzedhydrolysis of 3 ? aminopropionitrile at high concen ?tration with a tandem reaction strategy for shifting thereaction to β ?alanine formation. Journal of MolecularCatalysis B：Enzymatic，2015，115：113-118.

［21］ Gong J S，Li H，Lu Z M，et al. Engineering of afungal nitrilase for improving catalytic activity andreducing by ? product formation in the absence ofstructural information. Catalysis Science amp;Technology，2016，6（12）：4134-4141.

［22］ Jumper J，Evans R，Pritzel A，et al. Highly accurateprotein structure prediction with AlphaFold. Nature，2021，596（7873）：583-589.

［23］ Thuku R N，Brady D，Benedik M J，et al. Microbialnitrilases： Versatile， spiral forming， industrialenzymes. Journal of Applied Microbiology，2009，106（3）：703-727.

［24］ Raczynska J E，Vorgias C E，Antranikian G，et al.Crystallographic analysis of a thermoactive nitrilase.Journal of Structural Biology， 2011， 173（2）：294-302.

［25］ Zhang L J，Yin B，Wang C，et al. Structural insightsinto enzymatic activity and substrate specificitydetermination by a single amino acid in nitrilase fromSyechocystis sp. PCC6803. Journal of StructuralBiology，2014，188（2）：93-101.

［26］ Barglow K T，Saikatendu K S，Bracey M H，et al.Functional proteomic and structural insights intomolecular recognition in the nitrilase family enzymes.Biochemistry，2008，47（51）：13514-13523.

［27］ Liu H，Gao Y，Zhang M，et al. Structures of enzyme–intermediate complexes of yeast Nit2：Insights intoits catalytic mechanism and different substratespecificity compared with mammalian Nit2. ActaCrystallographica Section D：Biological Crystallo ?graphy，2013，69（8）：1470-1481.

［28］ Pace H C，Hodawadekar S C，Draganescu A，et al.Crystal structure of the worm NitFhit Rosetta Stoneprotein reveals a Nit tetramer binding two Fhitdimers. Current Biology，2000，10（15）：907-917.

［29］ Soriano?Maldonado P，Martínez?Gómez A I，Andújar?Sánchez M，et al. Biochemical and mutational studiesof the Bacillus cereus CECT 5050T formamidasesupport the existence of a CEEK tetrad in severalmembers of the nitrilase superfamily. Applied andEnvironmental Microbiology，2011，77（16）：5761-5769.

［30］ Chien C H，Gao Q Z，Cooper A J L，et al. Structuralinsights into the catalytic active site and activity ofhuman Nit2/ω ?amidase：Kinetic assay and moleculardynamics simulation. Journal of BiologicalChemistry，2012，287（31）：25715-25726.

［31］ Novo C ，F(xiàn)arnaud S ，Tata R ，et al. Support for athree?dimensional structure predicting a Cys?Glu?Lyscatalytic triad for Pseudomonas aeruginosa amidasecomes from site ? directed mutagenesis and mutationsaltering substrate specificity. Biochemical Journal，2002，365（3）：731-738.

［32］ Liu D J，Xi L J，Han D Y，et al. Cloning，expression，and characterization of a novel nitrilase，PaCNit，fromPannonibacter carbonis Q4.6. BiotechnologyLetters，2019，41（4）：583-589.

［33］ Nakai T，Hasegawa T，Yamashita E，et al. Crystalstructure of N ? carbamyl ? D ? amino acid amido ?hydrolase with a novel catalytic framework commonto amidohydrolases. Structure，2000，8（7）：729-738.

［34］ Silva Teixeira C S，Sousa S F，Cerqueira N M F SA. An unsual Cys ? Glu ? Lys catalytic triad isresponsible for the catalytic mechanism of thenitrilase superfamily：A QM/MM study on Nit2.ChemPhysChem，2021，22（8）：796-804.

［35］ Tian C，Kasavajhala K，Belfon K A A，et al. ff19SB：Amino ? acid ? specific protein backbone parameterstrained against quantum mechanics energy surfaces insolution. Journal of Chemical Theory andComputation，2020，16（1）：528-552.

［36］ Jorgensen W L，Chandrasekhar J，Madura J D，et al.Comparison of simple potential functions for"simulating liquid water. The Journal of ChemicalPhysics，1983，79（2）：926-935.

［37］ Miyamoto S，Kollman P A. Settle：An analyticalversion of the SHAKE and RATTLE algorithm forrigid water models. Journal of ComputationalChemistry，1992，13（8）：952-962.

［38］ Davidchack R L，Handel R，Tretyakov M V.Langevin thermostat for rigid body dynamics. TheJournal of Chemical Physics，2009，130（23）：234101.

［39］ Case D A，Aktulga H M，Belfon K，et al. Amber2021. San Francisco：University of California，2021.

［40］ Svensson M，Humbel S，F(xiàn)roese R D J，et al.ONIOM： A multilayered integrated MO+ MMmethod for geometry optimizations and single pointenergy predictions. A test for Diels ? Alder reactionsand Pt（P（t?Bu）3）2+ H2 oxidative addition. The Journalof Physical Chemistry，1996，100（50）：19357-19363.

［41］ Dapprich S，Komáromi I，Byun K S，et al. A newONIOM implementation in Gaussian98. Part I. Thecalculation of energies， gradients， vibrationalfrequencies and electric field derivatives. Journal ofMolecular Structure：THEOCHEM，1999，461-462：1-21.

［42］ Bakowies D，Thiel W. Hybrid models for combinedquantum mechanical and" molecular mechanicalapproaches. The Journal of Physical Chemistry，1996，100（25）：10580-10594.

［43］ Frisch M E，Trucks G，Schlegel H B，et al. Gaussian16. Wallingford，CT：Gaussian，Inc.，2016.

［44］ Tirado ? Rives J，Jorgensen W L. Performance ofB3LYP density functional" methods for a large set oforganic molecules. Journal of Chemical Theory andComputation，2008，4（2）：297-306.

（責(zé)任編輯 楊可盛）