Significantly improved catalytic efficiency of caffeic acid O-methyltransferase towards N-acetylserotonin by strengthening its interactions with the unnatural substrate's terminal structure

Article abstract of DOI:10.1016/j.enzmictec.2019.02.005

O-Methylation of N-acetylserotonin (NAS) has been identified as the bottleneck in melatonin biosynthesis pathway. In the present paper, caffeic acid O-methyltransferase from Arabidopsis thaliana (AtCOMT) was engineered by rational design to improve its catalytic efficiency in conversion of NAS to melatonin. Based on the notable difference in the terminal structure of caffeic acid and NAS, mutants were designed to strengthen the interactions between the substrate binding pocket of the enzyme and the terminal structure of the unnatural substrate NAS. The final triple mutant (C296F-Q310L-V314T) showed 9.5-fold activity improvement in O-methylation of NAS. Molecular dynamics simulations and binding free energy analysis attributed the increased activity to the higher affinity between the substrate terminal structure and AtCOMT, resulting from the introduction of N–H?π interaction by Phe296 substitution, hydrophobic interaction by Thr314 substitution and elimination of electrostatic repulsion by substitution of Gln310 with Leu310. This work provides hints for O-methyltransferase engineering and meanwhile lays foundation for biotechnological production of melatonin.

Full text of DOI:10.1016/j.enzmictec.2019.02.005

Accepted Manuscript
Title: Significantly improved catalytic efficiency of caffeic
acid O-methyltransferase towards N-acetylserotonin by
strengthening its interactions with the unnatural substrate’s
terminal structure
Authors: Wenya Wang, Sisi Su, Shizhuo Wang, Lidan Ye,
Hongwei Yu
PII:
S0141-0229(19)30021-3
DOI:
Reference:
https://doi.org/10.1016/j.enzmictec.2019.02.005
EMT 9312
To appear in:
Enzyme and Microbial Technology
Received date:
Revised date:
Accepted date:
10 December 2018
12 February 2019
13 February 2019
Please cite this article as: Wang W, Su S, Wang S, Ye L, Yu H, Significantly improved
catalytic efficiency of caffeic acid O-methyltransferase towards N-acetylserotonin by
strengthening its interactions with the unnatural substrate’s terminal structure, Enzyme
and Microbial Technology (2019), https://doi.org/10.1016/j.enzmictec.2019.02.005
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Significantly improved catalytic efficiency of caffeic acid O-
methyltransferase towards N-acetylserotonin by strengthening its
interactions with the unnatural substrate’s terminal structure
a#*
a#
a
b*
b*
Wenya Wang , Sisi Su , Shizhuo Wang , Lidan Ye , Hongwei Yu
aCollege of Life Science and Technology, Beijing University of Chemical Technology, Beijing
00029, China
1
bInstitute of Bioengineering, College of Chemical and Biological Engineering, Zhejiang
University, Hangzhou 310027, China
#These authors contribute to the work equally.
*
Corresponding author:
wangwy@mail.buct.edu.cn, tel：+86-10-64421335;
yelidan@zju.edu.cn, tel: +86-571-88273997;
yuhongwei@zju.edu.cn, tel: +86-571-88273997.
Highlights


A caffeic acid O-methyltransferase (AtCOMT) was engineered by rational design to
improve its catalytic efficiency to produce melatonin with N-acetylserotonin (NAS).
The AtCOMT activity was improved by 8.5-fold by applying the strategy of
strengthening the interactions between enzyme binding-pocket and NAS terminal
structure.


The strengthening strategies include the introduction of N—H···π interaction,
hydrophobic interaction and elimination of electrostatic repulsion.
This work provides hints for O-methyltransferase engineering and lays foundation for
biotechnological production of melatonin.

Abstract
O-methylation of N-acetylserotonin (NAS) has been identified as the bottleneck in melatonin
biosynthesis pathway. In the present paper, caffeic acid O-methyltransferase from Arabidopsis
thaliana (AtCOMT) was engineered by rational design to improve its catalytic efficiency in
conversion of NAS to melatonin. Based on the notable difference in the terminal structure of
caffeic acid and NAS, mutants were designed to strengthen the interactions between the
substrate binding pocket of the enzyme and the terminal structure of the unnatural substrate
NAS. The final triple mutant (C296F-Q310L-V314T) showed 9.5-fold activity improvement
in O-methylation of NAS. Molecular dynamics simulations and binding free energy analysis
attributed the increased activity to the higher affinity between the substrate terminal structure
and AtCOMT, resulting from the introduction of N—H···π interaction by Phe296 substitution,
hydrophobic interaction by Thr314 substitution and elimination of electrostatic repulsion by
substitution of Gln310 with Leu310. This work provides hints for O-methyltransferase
engineering and meanwhile lays foundation for biotechnological production of melatonin.
Key words: N-acetylserotonin, Melatonin, O-methyltransferase, Rational Design, O-
methylation

1
. Introduction
Melatonin is a hormone mainly synthesized in mammals’pineal gland and is used as a medicine
for sleep disorders, with a huge market demand. Currently, it is produced mainly by chemical
synthesis, which requires costly equipment and toxic or inflammable substrates and solvents
[
1]. Therefore, it is essential to find safer and more cost-effective routes for melatonin
production. In 2016, Byeon et al. and Germann et al. respectively constructed de novo
melatonin biosynthetic pathways in Escherichia coli and Saccharomyces cerevisiae, however,
both with low yields. The O-methyltransferase-catalyzed conversion of N-acetylserotonin
(NAS) to melatonin was found to be a major bottleneck in melatonin biosynthesis [2]. Thus,
improving the activity of O-methyltransferase towards NAS would be of great significance for
biotechnological production of melatonin.
Taken into account the difficulty in functional expression of animal-sourced
acetylserotonin O-methyltransferases (ASMTs) in E. coli, plant enzymes harboring ASMT
activity may be good candidates for protein engineering. Among the plant O-
methyltransferases that are able to catalyze the conversion of NAS to melatonin, the caffeic
acid O-methyltransferase from Arabidopsis thaliana (AtCOMT) showed the highest catalytic
efficiency under physiological conditions [3] and was selected as the engineering target.
AtCOMT is a multifunctional enzyme, responsible for lignin and flavonoid biosynthesis
in plants with caffeic acid as the natural substrate [4-5]. The hundred-fold difference of Vmax
-1
-1
value between NAS and caffeic acid for AtCOMT (1.8 vs. 564 nmol·min ·mg protein) [6]

inspired us to explore the molecular basis for this huge difference and thereby develop a
rational design strategy to enhance its O-methylation activity towards NAS. In this study,
molecular docking was adopted to comparatively analyze the enzyme-substrate complexes, so
as to guide the design of mutants for better catalysis of NAS O-methylation. After kinetic
analysis, molecular dynamics simulations were conducted for the positive mutants and the wild
type to reveal the molecular mechanism of the activity improvement.

2
. Materials and methods
2
.1 Materials
N-acetylserotonin was purchased from TCI (Tokyo, Japan). S-adenosyl-L-methionine (SAM)
and melatonin were purchased from Sangon Biotech (Shanghai, China). The enzymes for
molecular manipulations such as PrimeSTARTM HS DNA polymerase, Taq DNA polymerase,
restriction endonucleases (BamHI and HindIII) and T4 DNA ligase were purchased from
TaKaRa (Dalian, China). Genomic DNA purification kit was obtained from Dongsheng
Biotech (Guangzhou, China). Gel extraction kit and plasmid miniprep kit were obtained from
Axygen (Wujiang, China). Primers were synthesized by Invitrogen (Suzhou, China). Ni-NTA
column was purchased from GE Healthcare Bio-Science (Uppsala, Sweden). Other chemical
reagents were all of analytical grade and obtained from local companies.
2
.2 Cloning of caffeic acid O-methyltransferase (AtCOMT) from Arabidopsis thaliana
The full-length sequence of AtCOMT (At5g54160, GenBank ID: 835504) was amplified by
PCR with 1 unit of Prime STAR polymerase in 25 μL PCR mixtures containing 0.2 mM of
each dNTP, and 5 pmol each of the forward (5’-ATG GGT TCAACG GCA GAG ACA CAA-
3
’) and reverse (5’-TTA GAG CTT CTT GAG TAA CTC AAT AAG G-3’) primers, using
cDNA of A. thaliana as the template. The pET-30a plasmid harboring the recombinant COMT
gene and N-terminal hexahistidine-tag was transformed into competent cells of E. coli strain
BL21 (DE3).
2
.3 Mutants construction

The mutants were site-directed mutated by the QuickChangeTM method (Stratagene, La Jolla,
CA). The primers are listed in the Supplementary material (Table S1). The PCR reaction
mixture contained the following: 10 μL of 5× PrimerSTARTM buffer, 4 μL of dNTPs (2.5 mM
each), 15 ng of template (pET-30a-COMT), 1.0 μL of forward primer (10 μM), 1.0 μL of
reverse primer (10 μM), 0.5 μL ofPrimerSTARTM HS DNA polymerase (2.5 U/μL) in a total
volume of 50 μL. The PCR program consisted of a denaturation step at 95 ℃ for 1 min,
followed by 20 cycles of denaturation at 98℃for 10 s, annealing at 55℃for 15 s, and elongation
at 72℃ for 7 min. The PCR product was purified and then digested with Dpn I. The digested
mixture was transformed into competent cells of E. coli BL21 (DE3). Finally, the cells were
incubated overnight at 37°C on Luria-Bertani agar plates supplemented with 50 μg/mL of
kanamycin. The resulting mutants were sent for sequencing (Sangon Biotech, Shanghai, China)
to verify the target mutations.
2
.4 Overexpression and purification of recombinant enzymes
Expression was induced using 0.1 mM isopropyl β-D-1-Thiogalactopyranoside (IPTG) for 16
h at 16℃. Cells were harvested by centrifugation (4000×g, 4°C, 10 min) and washed with
normal saline. Cells were suspended in 10 mL phosphate buffered saline (100 mM, pH 8.0)
and disrupted by sonication. The supernatant containing crude enzyme was obtained by
centrifugation (12,000×g, 4°C, 15 min). After going through the filter (0.45 µm), the enzyme
was purified by HisTrap affinity chromatography (GE Healthcare) and desalted by HiTrap gel
chromatography (GE Healthcare).

2
.5 Enzyme assay and kinetic analysis
Enzymatic methylation was carried out at 37°C in 100 µL potassium phosphate buffer (100
mM, pH 7.8, containing 1 mM substrate and 2 mM S-adenosyl-L-methionine) for 30 min. The
reaction was stopped by adding the same volume of methanol. Kinetic analysis was conducted
at 37°C in 100 µLpotassium phosphate buffer (100 mM, pH 7.8, contained 0.1-9 mM substrate,
2
mM S-adenosyl-L-methionine and 0.05 mg/mL purified enzyme). The enzyme activity for
N-acetylserotonin methylation was calculated on the basis of HPLC results.
.6 Analytical methods
2
The reaction mixture was analyzed by HPLC (SHIMADZU LC-20 AT) equipped with an
Amethyst C18 column (4.6×150 mm, 5 μm, Sepax Technologies, Inc.), and the signals were
detected using a UV detector at 280 nm. The mobile phase was composed of methanol and
water (45/55, v/v) at a flow rate of 0.2 ml/min.
2
.7 Molecular homology modeling and molecular dynamics simulations
The homology model of AtCOMT was built by SWISS-MODE [7-8], using COMT structure
Protein Data Bank code 1KZY) [9] as the template. The three-dimensional homology models
(
of mutants were constructed by Swiss-Pdb Viewer 4.03. The docking models of substrate and
protein were set up using Autodock 4.0. Molecular dynamics simulation was performed with
AMBER 11. The substrate and His267 were constrained within the distance of proton transfer,
and the distance between the substrate and the co-enzyme SAM was constrained to 2.5 Å.
During simulation, 2500 cycles of steepest descent followed by 2500 cycles of conjugate

gradient minimization were first applied in a three-step procedure to realize energy
minimization. After gradually heating to 310 K over the course of 50 ps, the system was
equilibrated at a constant temperature of 303 K and a constant pressure of 1×105 Pa. Finally, a
5
ns production run in the constant-pressure and constant-temperature ensemble was performed.
The nonbonding cutoff distance was set to 8.0 Å and simulation snapshots were saved every 2
ps for analysis. SHAKE algorism was used to constrain H-containing aromatic bonds, whereas
PME (particle mesh Ewald) method was used to handle remote electrostatic force. The
conformation of each simulation was analyzed using the ptraj program in the AMBER package.
The mmpbsa program in AMBER was used to calculate the binding free energy of each
complex and to decomposite the binding free energy into individual residues. The hydrophobic
effect of AtCOMT and NAS was analyzed by LigPlot+.

3
. Results and discussion
To provide hints for the design of superior mutants with improved activity towards the
unnatural substrate NAS, the structural models of caffeic acid-AtCOMT complex and NAS-
AtCOMT complex were comparatively analyzed. Molecular docking indicated significant
difference in the terminal structure of the substrates (Fig. 1). The obviously longer "tail" of
NAS as compared to caffeic acid may constrain its binding with AtCOMT. Therefore, the
residues Cys296, Gln310 and Val314 of AtCOMT that locate near the NAS "tail" were selected
as the hot spots, and mutagenesis was conducted by introducing residues with relatively long
side chain and the ability to form hydrogen bond (C296N, C296S, Q310K), or neutral residues
with hydroxyl groups capable of stabilizing the substrate conformation by formation of various
interactions with the solvent or other residues (C296F, C296W, C296Y, C296H, Q310L, Q310V,
Q310T, Q310F, V314T, V314F, V314H, V314N), so as to enhance the interactions between the
substrate-binding pocket and NAS. In addition, residues locating in the proximity of the
substrate "head" such as Asn129, His321 and Ile317, as well as Asp268 near the phenolic
hydroxyl group were substituted with long-chain residues to enhance the forming probability
of hydrogen bonds between the enzyme and the substrate. Trp269, Phe174 and Phe161 located
at the entry of substrate binding pocket were substituted with short-chain residues to reduce
steric hinderance.
Among all mutants designed, C296F, C296H, C296Y, Q310L, V314N and V314T
showed activity improvement (Table 1), implying that the sites near the substrate "tail" indeed

play important roles in the activity of AtCOMT towards NAS. The activity improvement was
presumably due to the formation of cation-π interaction and hydrogen bonds with the substrate
terminal structure by introduction of residues with aromatic rings or hydroxyl groups, as well
as the stabilization of the conformation of the enzyme-substrate complex by substitution of
hydrophilic residues with hydrophobic ones or vice versa. Based on the positive single mutants,
combinatorial mutagenesis was conducted. Among the double mutants, C296F-Q310L,
C296H-Q310L, C296Y-V314T and Q310l-V314T showed obviously enhanced activity over
the parent mutants. Further integration of these positive mutations generated the best triple
mutant C296F-Q310L-V314T with 9.5-fold activity improvement as compared to the wild type.
Interestingly, the single mutant C296H showed the highest activity among all mutants at site
2
96, whereas the best double and triple mutants were respectively C296Y-V314T and C296F-
Q310L-V314T, indicating the synergy of the single mutations.
In order to illustrate the mechanism behind activity improvement of mutant C296F-
Q310L-V314T, enzymatic kinetics was analyzed (Table 2). The results showed both improved
affinity (shown as the decrease in K
mutants, leading to obvious increased kcat/K
V314T was improved by over 40-fold. In consistence, the binding free energy analysis (Table
) found the highest binding energy for C296F-Q310L-V314T, explaining its high binding
m
) and conversion rate towards the substrate (NAS) for all
m
. Particularly, the kcat/K for C296F-Q310L-
m
3
capacity and high catalytic efficiency.
The catalytic reaction of AtCOMT has been postulated to proceed in two steps: first, the

NE2 in His267 (catalytic amino acid as a general base) takes a proton (H1) away from the
phenolic hydroxyl of substrate, resulting in formation of the phenol anion in substrate (NAS);
second, the phenol anion attacks the C14 of methyl in SAM to methylate NAS at the phenolic
hydroxyl position by SN reaction (Fig. 2) [10]. In order to get further hints for the molecular
2
mechanism behind the activity improvement of mutant C296F-Q310L-V314T, the interactions
between NAS and His267/SAM were analyzed by comparing the molecular models of WT and
C296F-Q310L-V314T using molecular dynamics simulation. The results indicated no
significant difference in the H-bond of H1···NE2 between NAS and His267, as well as the
distance and angle between NAS and C14 position of active methyl in SAM (Table 4). All
these results demonstrated that the mutations in C296F-Q310L-V314T did not alter the
conformation of the substrate, catalytic amino acids and cofactor in the enzyme-substrate
complex, only causing the improvement of binding capacity between enzyme and substrate.
Molecular dynamics simulations provided further insight into the mechanism of higher
catalytic efficiency for the mutant. C296F mutation substituting the original Cys with Phe
introduced an aromatic group to position 296. As a result, an N—H···π interaction was formed
between the aromatic ring at position 296 and the H atom of NH group at NAS "tail", with a
distance of 2.21 Å (Fig. 3) [10]. N—H···π interaction could stabilize the amido bond in the
terminal chain of NAS, and thus increase the affinity of AtCOMT towards NAS, in consistence
with the decrease of K
m
value in C296F mutant (Table 2), resulting in the higher catalytic
efficiency (Table 1). The same phenomenon could also be observed in the C296Y mutant,

which also introduced an aromatic group at position 296 by substituting Cys with Tyr. In the
case of position 310, the original amino acid is Gln, containing an amido group with negative
charge. Taken together, with the existence of an amido group in the terminal chain of NAS,
strong electrostatic repulsion would occur when NAS tail enters the substrate binding pocket
of AtCOMT, preventing the binding of NAS with the catalytic pocket. After substitution with
a neutral amino acid (Leu), the binding free energy decreased from -0.05 kcal/mol to -0.50
kcal/mol, indicating the elimination of electrostatic repulsion [11], which is in favor of the
binding between AtCOMT and NAS. This result was also confirmed by the lower K and
m
higher Kcat of mutant Q310L as compared to the wild type. Substitution of V314T is expected
to introduce strong hydrophobic effect via the interaction between Thr and the methyl group in
NAS "tail" [12], resulting in higher catalytic activity. Before the introduction of V314T
mutation, the H-bond length of N—H···π in mutant C296F-Q310Lwas 3.36 Å, and the binding
free energy of Phe296 was 0.007 kcal/mol in mutant C296F-Q310L. In contrast, the H-bond
length of N—H···π and the binding free energy of Phe296 were reduced respectively to 2.21
Å and -0.59 kcal/mol in C296F-Q310L-V314T (Table 5). All these data demonstrated the
introduction and stabilization of the N—H···π interaction as a result of synergistic effect of the
three mutations, which improved enzyme-substrate binding and enhanced the catalytic
efficiency [13].

4
. Conclusion
The catalytic center of enzyme has varied binding affinity towards compounds with different
structures, resulting in variation of catalytic efficiency towards different substrates [14].
Consequently, by comparing the active center conformation of enzyme-substrate complexes
for the natural and unnatural substrates, key clues may be acquired on the limited catalytic
efficiency towards the target reaction. This assumption inspired us to develop a rational design
strategy for AtCOMT towards improved NAS O-methylation efficiency by strengthing the
binding between the enzyme and the substrate terminal structure which is obviously different
from the natural substrate caffeic acid. The final triple mutant showed 9.5-fold activity
improvement as compared to the wild type, demonstrating the efficiency of this design method.
Molecular dynamics simulations suggested the introduction of N—H···π interaction and
hydrophobic interaction in addition to elimination of electrostatic repulsion as the mechanism
for activity improvement. Considering the rate-limiting role of NAS O-methylation in
melatonin biosynthetic pathway, this work may provide a basis for biotechnological production
of melatonin.

Acknowledgements
This work was financially supported by the National Key Research and Development Program
of China (No. 2016YFD0400604), Natural Science Foundation of China (Grant Nos. 21576234
and 21776244) and Zhejiang Provincial Natural Science Foundation of China (Grant No.
LY18B060001).
Conflicts of interest
There are no conflicts of interest to be declared.

References
1
. T. Sun, L. Chen, W. Zhang, Microbial production of mammalian melatonin – a promising
solution to melatonin industry, Biotechnol. J. 11 (2016) 601-602.
2
. S.M. Germann, S.A. Baallal Jacobsen, K. Schneider, S.J. Harrison, N.B. Jensen, X. Chen,
S.G. Stahlhut, I. Borodina, H. Luo, J. Zhu, J. Maury, J. Forster, Glucose‐based microbial
production of the hormone melatonin in yeast Saccharomyces cerevisiae, Biotechnol. J., 11
(2016)717-724.
3
4
. K. Back, D. X. Tan, R. J. Reiter, Melatonin biosynthesis in plants: multiple pathways
catalyze tryptophan to melatonin in the cytoplasm or chloroplasts, J. Pineal Res., 61 (2016)
4
26-37.
. H.Y. Lee, Y. Byeon, K. Lee, H-J Lee, K. Back. Cloning of Arabidopsis serotonin N‐
acetyltransferase and its role with caffeic acid O‐methyltransferase in the biosynthesis of
melatonin in vitro despite their different subcellular localizations, J. Pineal Res., 57 (2014)
4
18-426.
5
6
. K.C. Lam, R.K. Ibrahim, B. Behdad, S. Dayanandan, Structure, function, and evolution of
plant O-methyltransferases, Genome, 50 (2007) 1001-1013.
. Y. Byeon, H.Y. Lee, K. Lee, K. Back, Caffeic acid O‐methyltransferase is involved in the
synthesis of melatonin by methylating N‐acetylserotonin in Arabidopsis, J. Pineal Res.,
5
7 (2014) 219-227.
7
. L.S. Arantes, L.G.V. Nova, B.C. Resende, M. Bitar, I.E.V. Coelho, A. Miyoshi, V.A.
Azevedoc, L.L. Santos, Carlos Renato Machado, D.O. Lopes, The corynebacterium
pseudotuberculosis genome contains two formamidopyrimidine-DNA glycosylase
enzymes, only one of which recognizes and excises 8-oxoguanine lesion, Gene 575 (2016)
2
33-243.
8
9
. K. Arnold, L. Bordoli, T. Schwede, The SWISS-MODEL workspace: a web-based
environment for protein structure homology modelling, Bioinformatics, 22 (2006)195-201.
. C. Zubieta, P. Kota, J.L. Ferrer, R. A. Dixon, J. P. Noel, Structural basis for the modulation
of lignin monomer methylation by caffeic acid/5-hydroxyferulic acid 3/5-o-
methyltransferase, Plant Cell, 14 (2002)1265-1277.
1
1
0. E.B. Starikov, T. Steiner, Structural and computational analysis of published neutron
diffraction data show that crystalline vitamin B12 coenzyme contains a strong
intramolecular N-H...Phe hydrogen bond, Acta Crystallogr. B 54 (1998) 94–96.
1. S. Miranda-Rojas, I. Fernández, J. Kästner, A. Toro‐Labbé, F. Mendizábal, Unraveling
the nature of the catalytic power of fluoroacetate dehalogenase, Chemcatchem 10
(2018)1052-1063

1
2. T. Hakoshima, M. Tanaka, T. Itoh, K-k. Tomita, T. Amisaki, S. Nishikawa, H. Morioka, S-
i Uesugi, E. Ohtsuka, M. Ikehara, Hydrophobic effects on protein/nucleic acid interaction:
enhancement of substrate binding by mutating tyrosine 45 to tryptophan in ribonuclease Tl,
Protein Eng. Des. Sel., 4 (1991) 793–799.
1
1
3. M.A. Tararina, S. Xue, L.C. Smith, S.N. Muellers, P.O. Miranda, K.D. Janda, K.N. Allen,
Crystallography coupled with kinetic analysis provide mechanistic underpinnings of a
nicotine-degrading enzyme, Biochemistry 57 (2018) 3741-3751.
4. S.G. Moinuddin, M. Jourdes, D.D. Laskar, C. Ki, C.L. Cardenas, K-W. Kim, D. Zhang,
L.B. Davin, N.G. Lewis. Insights into lignin primary structure and deconstruction from
Arabidopsis thaliana COMT (caffeic acid O-methyl transferase) mutant Atomt1, Org.
Biomol. Chem. 8 (2010) 3928-3946.

Table 1. Activity of positive AtCOMT mutants towards NAS
-
1
-1
Mutant
WT
Activity (nmol·min ·mg )
6.09 ± 0.08
C296F
9.82 ±0.63
C296H
19.85 ±0.68
11.09 ±0.95
11.31 ±0.80
9.45 ±0.63
C296Y
Q310L
V314N
V314T
16.97 ±0.37
27.01 ±1.29
21.86 ±2.03
34.62 ±1.90
33.42 ±0.59
57.87 ±3.49
39.13 ±1.27
46.94 ±0.56
C296F-Q310L
C296H-Q310L
C296Y-V314T
Q310L-V314T
C296F-Q310L-V314T
C296H-Q310L-V314T
C296Y-Q310L-V314T

Table 2. Kinetic analysis of AtCOMT WT and mutants towards NAS
-
1
-3 -1 -1
Mutant
WT
K
m
(mM)
k
cat (s )
kcat/K
m
(10 s M )
0.74
0.58
0.36
0.58
0.26
2.94
3.99
V314T
11.32
4.76
19.53
13.11
9.37
Q310L
C296F
5.41
C296F-Q310L-V314T
43.79
171.06

Table 3. Individual energy term of the binding free energies of AtCOMT WT or C296F-
Q310L-V314T binding with NAS (kcal/mol)
Mutant
WT
VDWAALS*
-16.51
EEL*
-33.84
-28.21
EPB*
52.39
40.22
ECAVITY*
-2.41
ΔGbinding*
-0.36 ± 2.94
-2.24 ± 2.51
C296F-Q310L-V314T
-11.86
-2.38
*
VDWAALS: van der Waals energy as calculated by the MM force field; EEL: Electrostatic energy as
calculated by the MM force field; EPB: the electrostatic contribution to the salvation free energy calculated
by PB; ECAVITY: nonpolar contribution to the salvation free energy calculated by an empirical model;
ΔGbinding: final estimated binding free energy.

Table 4. Structural analysis of AtCOMT WT and C296F-Q310L-V314T towards His267 and
SAM*
Distance(Å)
His267@NE2-
NAS@H1
Angle（°）
His267@NE2-
NAS@H1-
Distance(Å)
SAM@C14-
NAS@O
Angle（°）
SAM@S-
SAM@C14-
NAS@O
NAS@O
WT
1.78 ± 0.03
1.78 ± 0.03
166.47 ± 6.93
166.09 ± 6.79
2.56 ± 0.02
2.55 ± 0.02
171.87 ± 4.16
171.46 ± 4.92
C296F-Q310L-V314T
*The bond distances and angles were calculated using the AMBER ptraj program.

Table 5 Decomposition of binding free energy for mutated sites and NAS (kcal/mol)
Site
WT
C296F-Q310L-V314T
C296F-Q310L
van der
Waals
-0.32
Electros
tatic
Polar
Solvation
-0.15
TOTAL
van der Electros
Polar
Solvation
0.75
TOTAL van der Electrost
Polar
Solvation
0.083
TOTAL
Waals
-1.05
-0.53
-0.51
tatic
-0.30
-0.27
0.08
Waals
-0.015
-0.015
-0.130
atic
2
3
3
96
10
14
0.14
-0.33
-0.05
-0.93
-0.59
-0.50
0.60
-0.061
-0.056
-0.284
0.007
-0.022
0.064
-0.06
-0.04
-0.06
0.05
0.29
0.049
-1.02
0.16
1.03
0.477

Figure legends:
Fig. 1 Structural model comparison of NAS-AtCOMT complex (golden) and caffeic acid-
AtCOMT complex (blue). Cys296, Gln310 and Val314 are marked in orange.
Fig. 2 Postulated catalysis of NAS O-methylation by AtCOMT. Yellow arrows indicate the
direction of atoms attack.
Fig. 3 Enzyme-substrate complexes of the wild-type and mutant AtCOMT. Purple arc
indicates hydrophobic effect, and dotted red line indicates N—H···π interaction.
Fig.1. Structural model comparison of NAS-AtCOMT complex (golden) and caffeic acid- AtCOMT
complex (blue). Cys296, Gln310 and Val314 are marked in orange.

Fig.2. Mechanism of AtCOMT O-methylating NAS
(Yellow arrows show as the direction of atoms attack)

Fig.3.Mechanism for activity improvement in the mutants
Purple arc indicates hydrophobic effect, dotted red line indicates N—H···π interaction.