Structural characterization and substrate specificity of VpAAT1 protein related to ester biosynthesis in mountain papaya fruit

Article abstract of DOI:10.1016/j.jmgm.2010.11.011

The aroma in fruits is an important attribute of quality that influences consumer's acceptance. This attribute is a complex character determined by a set of low molecular weight volatile compounds. In mountain papaya fruit (Vasconcellea pubescens) the aroma is determined mainly by esters, which are produced through an esterification reaction catalyzed by the enzyme alcohol acyltransferase (AAT) that utilizes alcohols and acyl-CoAs as substrates. In order to understand the molecular mechanism involved in the production of esters in this fruit, an AAT gene which has been previously cloned and characterized from mountain papaya (VpAAT1) was expressed in yeasts, and the highest enzyme activity of the recombinant protein was obtained when the enzyme was tested for its ability to produce benzyl acetate. On the other hand, to gain insight the mechanism of action at the molecular level, a structural model for VpAAT1 protein was built by comparative modelling methodology, which was validated and refined by molecular dynamics simulation. The VpAAT1 structure consists of two domains connected by a large crossover loop, with a solvent channel in the center of the structure formed between the two domains. Residues H166 and D170, important for catalytic action, displayed their side chains towards the central cavity of the channel allowing their interaction with the substrates. The conformational interaction between the protein and several ligands was explored by molecular docking simulations, and the predictions obtained were tested through kinetic analysis. Kinetic results showed that the lowest K_M values were obtained for acetyl-CoA and benzyl alcohol. In addition, the most favorable predicted substrate orientation was observed for benzyl alcohol and acetyl CoA, showing a perfect coincidence between kinetic studies and molecular docking analysis.

Full text of DOI:10.1016/j.jmgm.2010.11.011

Journal of Molecular Graphics and Modelling 29 (2011) 635–642
Contents lists available at ScienceDirect
Journal of Molecular Graphics and Modelling
journal homepage: www.elsevier.com/locate/JMGM
Structural characterization and substrate specificity of VpAAT1 protein related to
ester biosynthesis in mountain papaya fruit
Luis Morales-Quintana, Lida Fuentes, Carlos Gaete-Eastman, Raúl Herrera, María Alejandra Moya-León∗
Laboratorio de Fisiología Vegetal y Genética Molecular, Instituto de Biología Vegetal y Biotecnología, Universidad de Talca, Casilla 747, Talca, Chile
a r t i c l e i n f o
a b s t r a c t
Article history:
The aroma in fruits is an important attribute of quality that influences consumer’s acceptance. This
attribute is a complex character determined by a set of low molecular weight volatile compounds. In
mountain papaya fruit (Vasconcellea pubescens) the aroma is determined mainly by esters, which are
produced through an esterification reaction catalyzed by the enzyme alcohol acyltransferase (AAT) that
utilizes alcohols and acyl-CoAs as substrates. In order to understand the molecular mechanism involved
in the production of esters in this fruit, an AAT gene which has been previously cloned and characterized
from mountain papaya (VpAAT1) was expressed in yeasts, and the highest enzyme activity of the recom-
binant protein was obtained when the enzyme was tested for its ability to produce benzyl acetate. On the
other hand, to gain insight the mechanism of action at the molecular level, a structural model for VpAAT1
protein was built by comparative modelling methodology, which was validated and refined by molecu-
lar dynamics simulation. The VpAAT1 structure consists of two domains connected by a large crossover
loop, with a solvent channel in the center of the structure formed between the two domains. Residues
H166 and D170, important for catalytic action, displayed their side chains towards the central cavity of
the channel allowing their interaction with the substrates. The conformational interaction between the
protein and several ligands was explored by molecular docking simulations, and the predictions obtained
were tested through kinetic analysis. Kinetic results showed that the lowest KM values were obtained
for acetyl-CoA and benzyl alcohol. In addition, the most favorable predicted substrate orientation was
observed for benzyl alcohol and acetyl CoA, showing a perfect coincidence between kinetic studies and
molecular docking analysis.
Received 28 September 2010
Received in revised form
10 November 2010
Accepted 15 November 2010
Available online 23 November 2010
Keywords:
Alcohol acyltransferase
Docking simulation
Ester biosynthesis
Molecular modelling
Vasconcellea pubescens
© 2010 Elsevier Inc. All rights reserved.
1. Introduction
The aroma of fruits is determined by a large number of volatile
compounds, and its biosynthesis depends on many factors, such
Mountain or highland papaya fruit is characterized by its intense
aroma [1]. This fruit is native to South America and widely dis-
tributed throughout the Andes [2,3]. The specie was introduced
to Chile more than 50 years ago and is currently cultivated in the
Coquimbo and Valparaiso valleys and along the coast of the Maule
Region. The papaya cultivated in Chile belongs to the Caricaceae
family and corresponds to Vasconcellea pubescens (synonyms of
Vasconcellea cundinamarcensis, Carica pubescens, initially called Car-
ica candamarcensis) [2,3]. The fruit displays a climacteric behavior,
and exhibits the characteristic rise in ethylene production during
ripening, followed by softening, changes in color and rise in aroma
production [1,4].
as cultivar, ripening stage and postharvest conditions [5–7]. V.
pubescens’s aroma is determined mainly by esters and in lower
proportion by alcohols and aldehydes present at the ripe stage [1].
Alcohol acyltransferase (AAT), that catalyzes the esterification reac-
tion between alcohols and acyl-CoAs, is the key enzyme involved in
ester biosynthesis in this specie [8]. Recently, a gene encoding for
an alcohol acyltransferase (VpAAT1) has been isolated and charac-
terized in ripening mountain papaya fruit [8]. VpAAT1 is member of
the BAHD superfamily and belongs to subfamily III [8], sharing the
group with other AATs that participate in the synthesis of volatile
compounds during ripening of fruits such as melon [9]. The first
member from the BAHD superfamily which has a crystal structure
is vinorine synthase, an acetyl transferase from Rauvolfia serpentine
[10]. Members of the BAHD superfamily share the HXXXD motif,
located in the middle of the protein sequence, which is highly con-
served in higher plants and yeasts. The substitution of the histidine
and aspartic residues from this motif causes the loss of enzyme
function in vinorine synthase [11], which suggests that it could be
∗
Corresponding author at: Instituto de Biologia Vegetal y Biotecnologia,
Universidad de Talca, 2 Norte 685, Talca, Chile. Tel.: +56 71 200 286;
fax: +56 71 200 276.
E-mail addresses: alemoya@utalca.cl, amoyaleon@gmail.com (M.A. Moya-León).
1093-3263/$ – see front matter © 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jmgm.2010.11.011

636
L. Morales-Quintana et al. / Journal of Molecular Graphics and Modelling 29 (2011) 635–642
involved in the transfer of the acyl group from acyl-CoA towards
the alcohol.
2.2. Docking studies
Despite the knowledge generated in the past few years, we still
have a limited understanding of the mechanisms controlling the
synthesis of aroma volatiles. To clarify the molecular mechanism
used by VpAAT1 it is important to know the interaction of the pro-
tein and its ligands, however no structural information has been
reported for any AAT protein until now.
The understanding of the catalytic properties of the enzymes
involved in the production of volatiles in a particular fruit could
allow the design of additional strategies to improve its flavor during
storage, shipping, marketing or processing. In this sense, the aim
of this work was to go further in the structural and biochemical
characterization of VpAAT1 protein, by using both bioinformatics
(molecular modelling and docking simulations) and biochemical
tools (determination of kinetics parameters).
Docking studies were performed to predict the putative bind-
ing of a group of acyl-CoAs and different alcohols. For this purpose,
two different docking investigations using the Internal Coordinate
Mechanics (ICM) program [21] (www.molsoft.com) were carried
out. ICM is based on Monte Carlo simulations in internal coordi-
nates to optimize the position of molecules using a stochastic global
optimization procedure combined with pseudo-Brownian posi-
tional/torsional steps and fast local gradient minimization [21]. The
scoring function includes electrostatic, hydrogen bond, hydropho-
bic, and two van der Waals terms, and the energies are computed
using MMFF partial charges with the ECEPP/3 force field [21]. The
active site of VpAAT1 enzyme was defined as the pocket that con-
tained the HXXXD motif (H166, D170), which is located in the
middle of the protein sequence and is highly conserved in supe-
rior plants and yeasts. Additionally, His166 of the HXXXD motif
has been reported as a general base during catalysis [10,22].
The first docking study was done with acyl-CoA as acyl donor
bound to the enzyme VpAAT1. The second docking run was per-
formed with the alcohol. In order to verify that the order of entrance
of ligands had no significant effect on the stability of the final com-
plex, the docking study was performed in reverse order. The first
one was done with alcohol bound to the enzyme VpAAT1 and the
second docking run was performed with the acyl-CoA as acyl donor.
In each case, 10 different docking arrangements were produced.
The translation/orientation of all-atoms in the ligands was sampled
in a completely unrestricted way. This conformation was obtained
by Monte Carlo movements and local minimization using poten-
tials of interaction pre-calculated on a 3D grid from the receptor
(VpAAT1 protein). The conformations as result of Rigid-body dock-
ing were sorted by total binding energy, which include van der
Waals, electrostatics, and hydrogen bonding potentials. The result
was optimized by flexible docking. The VpAAT1 binding mode with
each substrate pair was employed in docking simulations to assess
the correlation between enzyme activity and the stability of the
substrate–enzyme complex.
2. Materials and methods
2.1. The 3D structure of VpAAT1 using homology modelling
The cDNA sequence of VpAAT1 was isolated and characterized
recently (GenBank accession number FJ548611) [8]. The 3D model
of VpAAT1 protein was built by homology modelling based on
high-resolution crystal structures of homologous proteins. BLAST
(Basic local alignment search tool) search and PSIPRED (Protein
Structure Prediction Server) [12] were used for selecting the 3D
models with the closest homology available in the Brookhaven
Protein Data Bank (PDB). It resulted in the selection of the crys-
tal structure of vinorine synthase protein (PDB: 2BGH) as template,
that shares a moderate sequence identity with VpAAT1. The first
requirement in the construction of VpAAT1 model is the sequence
alignment with the template, focussed on the identification of
structurally conserved regions (SCR) common to the template
and the target. The pair-wise sequence alignment was obtained
by using ClustalW software to build the initial model. The com-
parative model of VpAAT1 structure was generated by using the
software MODELLER 9v5 (http://salilab.org/modeller/) [13]. Fifty
models were generated and the model showing the lowest mod-
eller objective function and rms deviation with respect to trace (C␣
atoms) of the crystal structure of the template, was saved for further
refinement and validation. Further, refinement was performed in
order to obtain the best conformation of VpAAT1 structure result-
ing from MODELER. Structural refinements were accomplished by
molecular dynamics simulations (MDS) and equilibration meth-
ods using Nano Molecular Dynamics (NAMD 2.6) software [14]
and Chemistry of Harvard Molecular Modelling (CHARMM27) force
field for lipids and proteins [15,16] along with the TIP3P model
for water [17]. The tautomeric states of histidine residues in the
model were assigned according to the local environment. After an
initial minimization, the system was subjected to a short molec-
ular dynamics run in order to remove wrong contacts and to fill
empty pockets. All backbone atoms were restrained using a har-
2.3. Expression of recombinant VpAAT1
The full-length sequence of VpAAT1 was initially cloned into
pYES2.1 TOPO-TA cloning vector containing a C-terminal polyhis-
tidine tag following the instructions provided by the manufacturer
(Invitrogen). The construct was cloned in Escherichia coli TOP 10
One Shot (Invitrogen) and sequenced by Macrogen Inc. (Seoul,
Korea). Once the right orientation of insert was confirmed, the con-
struction was used to transform the Saccharomyces cerevisiae cell
line INVSc1. Transformed yeasts were grown at 30 ◦C in selective
medium (SC-U) with 2% galactose as inducer of the expression of
recombinant protein, with constant stirring and bubbling of sterile
air until OD600 of the culture reached <1 U.
2.4. Purification of recombinant VpAAT1
−2
˚
A
monic force constant of 5 kcal mol−1
and the loops were left
The purification of recombinant protein was carried out accord-
ing to the method of El-Sharkawy et al. [9] with modifications [8].
Yeast cells collected were suspended in 2 mL of phosphate buffer
(50 mM sodium phosphate (pH 7.5), 10% (v/v) glycerol, 0.3 M NaCl)
containing 2 mM ␤-mercaptoethanol, and mechanically ground in
liquid nitrogen. The crude extract obtained was purified through
a Talon Metal Affinity column (BD Biosciences), an affinity col-
umn designed to purify polyhistidine-tagged proteins, according to
the manufacturer’s protocol. The recombinant protein was eluted
with phosphate buffer containing 150 mM imidazole. Proteins were
quantified according to Bradford [23] and visualized through 10%
SDS-PAGE gels.
free during relaxations. All MDS were done using a time step of
˚
1 fs, with a 12 A spherical cut off for non-bonding interactions
˚
and a switching function of 10 A for the van der Waals term. The
structure obtained was embedded into a periodic bordering con-
˚
˚
˚
dition (PBC) box (92 A × 86 A × 94 A) contained in a 150 mM NaCl
solution at 300 K. Under these conditions, a 2 ns MDS was per-
formed using NAMD v2.6 program. Particle Mesh Ewald (PME) was
employed to calculate long-range electrostatic interactions. Finally,
to evaluate the accuracy of the model ProSA [18], PROCHECK
(http://nihserver.mbi.ucla.edu/SAVS/) [19] and Verify3D [20] were
employed.

L. Morales-Quintana et al. / Journal of Molecular Graphics and Modelling 29 (2011) 635–642
637
Table 1
2.5. Assay of AAT activity and kinetic parameters
Validation of VpAAT1 protein structure using PROCHECK program (Ramachandran
Plot).
AAT activity of recombinant VpAAT1 protein was quantified by
its ability to convert alcohols and acyl-CoAs into the corresponding
ester [24,25]. AAT activity was assayed in 500 ␮L of total volume
in the presence of 2 mM benzyl alcohol and 250 ␮M acetyl-CoA
in 50 mM Tris–HCl (pH 7.5) buffer containing 10% (v/v) glycerol
and 1 mM dithiothreitol (DTT) [8,9]. The reaction was initiated
by addition of 35 ␮g purified protein, and the mixture was incu-
bated at 30 ◦C for 2.5 h in sealed Eppendorf tubes. The reaction
was stopped by the addition of 50 mg of citric acid and 185 mg
of KCl, and after mixing during several minutes, the supernatant
was transferred to a glass vial, which was sealed after the addition
of 0.5 ␮L of 1,2-dichlorobenzene as internal control. The solution
was stirred during 15 min at room temperature; meanwhile the
volatiles produced during the enzymatic reaction were released
into the headspace and adsorbed onto an SPME fiber (PDMS/DVB).
The separation and quantification of each ester was done by a gas
chromatograph fitted with a flame ionization detector (GC-FID)
(Perkin-Elmer, Clarus 500) as described previously [1]. A calibration
curve was prepared for benzyl acetate. AAT enzyme activity was
expressed as pmol h−1 per milligram of protein. Determinations
were performed in triplicate and expressed as mean SE.
Core (%)a
90.1
Allow (%)b
7.8
Gener (%)c
2.1
Disall (%)d
0
VpAAT1
a
Most favorable region.
Additional allowed regions.
Generously allowed regions.
Disallowed regions.
b
c
d
three residues, located between the two catalytic ones, have their
side chains oriented inward of the enzyme core and not exposed
to the solvent channel.Another highly conserved region within the
BAHD acyltransferases is the DFGWG motif near the C terminus.
The analysis of VpAAT1 structure reveals that the DFGWG motif is
remote from the active site, and therefore it is unlikely to have a
direct role in either substrate binding or catalysis.
3.2. Model evaluation
To evaluate the geometric and energetic stability of the VpAAT1
model, 10 conformers with the lowest total energy were selected
after energy minimization. The best conformer was determined
by using different evaluation methods. The stereochemical qual-
ity of the VpAAT1 model was analyzed with PROCHECK. The most
favored regions correspond to 90.1% of the structure and no badly
modelled residues were found (Table 1) that validates the stereo-
chemistry of the model. On the other hand, ProSA analysis displayed
a local model quality by plotting energies as a function of amino
acid sequence position I. This analysis showed a low ProSA energy
in most structurally conserved regions, however there is a small
region with unfavorable ProSA energy: this region corresponds to a
loop distant from the active site and out of the protein solvent chan-
nel (Fig. 4a). Finally, an analysis using Verify3D program showed
favorable scores for the entire structure of the three-dimensional
structure of VpAAT1 protein and no residues showing values under
0 (Fig. 4b). The final structure of VpAAT1 protein was accepted for
subsequent analysis.
For determination of KM and Vmax values standard assay
mixtures were prepared using 35 ␮g of purified protein. The
Michaelis–Menten saturation curve was built using substrate con-
centrations between 0 and 8 mM for benzyl alcohol and between 0
and 4 mM for acetyl-CoA. The KM and Vmax values were calculated
from Lineweaver–Burk plots.
3. Results
3.1. The homology modelling of VpAAT1
The initial pair-wise alignment of the protein template (2BGH)
against the target sequence (VpAAT1) with 21.5% of sequence iden-
tity was optimized manually, incorporating information on the
secondary structure of the BAHD superfamily, especially in rela-
tion to conserved motifs (e.g., the catalytic HXXXDG and DFGWG
motifs) which were spatially restricted during the modelling to
avoid any distortion of them (Fig. 1). The final unperturbed con-
formations of the model were acquired by progressively relaxing
parts of the initial model (Fig. 2). The model obtained showed that
the protein structure is composed of 15 ␤-sheets, 14 ␣-helices
(11 ␣-helices and three 310-helices) and 23 loops. Furthermore,
the VpAAT1 structure consists of two approximately equal-sized
domains (Fig. 3). The domains are connected with a large crossover
3.3. Ligand binding analysis
The refined VpAAT1 model was used to study protein–ligand
conformations by automatic docking for seven substrates (three
acyl-CoAs and four alcohols). Substrates were chosen based on
the results of activity previously described [8]. As shown in
Table 2 the lowest ligand binding energies were found for acyl-
CoAs (values between −203.5 and −212.0 kcal mol−1), and binding
energy for alcohols were less favorable (ranging from −41.4 to
−22.8 kcal mol−1). The strongest binding interaction was found
between VpAAT1 protein and hexanoyl-CoA.
˚
loop (residues 210–232) that spans nearly 37 A. Domain 1 con-
tains a mixed 7-stranded ␤-sheets (␤1–␤7), which is covered on
both sides by 7 helices (five ␣-helices, H␣1–H␣5; two 310-helices,
To obtain the ternary complex, the acyl-CoAs substrates were
used to form a complex with VpAAT1 first, and then, the alco-
310H1–310H2), and domain 2 contains 7 helices (six ␣-helices,
H␣6–H␣11; one 310-helice, 310H3) and a mixed 8-stranded ␤-
sheets (␤8–␤15) (Fig. 3). Additionally, the model showed a solvent
channel in the center of the 3D structure of VpAAT1 that is formed
between the two domains by two loops which protrudes from
domain 2 to contact domain 1 (Fig. 3). The first loop is located
between the ␤-sheets 9 and 10 of domain 2. The second loop is
situated between the ␤-sheets 12 and 13 also in domain 2. In this
channel the important residues for catalytic action are located;
these residues are in the HTMSD segment and located in the loop
13 between the ␤-sheet 7 and the ␣-helice 4 in the center of this
solvent channel. The catalytic site is formed by the residues H166
and D170 with their side chains oriented to the central cavity of
the channel in the protein, and they most probably interact with
the substrates alcohols and acyl-CoAs (Fig. 2). Whereas the other
Table 2
Affinity energy of VpAAT1 protein model towards different acyl-CoAs and alcohols
as substrates.
Substrate
Affinity energy (kcal mol−1
)
rmsd (Å)a
Ethanol
Butanol
Benzyl alcohol
Hexanol
−22.8
−33.8
−35.7
−41.4
0.1
1.6
2.7
2.2
Acetyl-CoA
Butanoyl-CoA
Hexanoyl-CoA
−203.5
−206.9
−212.0
1.8
0.2
2.1
a
rms deviation was calculated for the backbone (C␣, C, N) atoms.

638
L. Morales-Quintana et al. / Journal of Molecular Graphics and Modelling 29 (2011) 635–642
Fig. 1. Sequence alignment and secondary structure predictions of VpAAT1 protein. VpAAT1 amino acid sequence was aligned with that of vinorine synthase (PDB: 2BGH)
from Rauvolfia serpentina. Gaps are indicated by dashes, letters in green color are identical residues, and letters in other colors are similar residues. Red sticks and blue arrows
represent ␣-helix and ␤-strand structures, respectively.
hol was incorporated into the acyl-CoA VpAAT1 complex. Fig. 5a
shows a general view of the position of both ligands into the sol-
vent channel and the interaction with the two catalytic residues
(His 166 and Asp 170). In the acetyl-CoA benzyl alcohol VpAAT1
complex His166 is protonated at the nitrogen ␧, which forms a
hydrogen bond with the hydroxyl group of acetyl-CoA (Fig. 5b).
Additionally, the distance between the carboxylic group of Asp170
localization into the solvent channel and the distance of the
ligands with the residues His166 and Asp170. In the first com-
plex, benzyl alcohol was found at 2.73 A from residue D170
˚
into the solvent channel. In the second complex, ethanol was
˚
located between the residues D170 and H166 at 1.6 A of the car-
˚
boxylic group of D170 and 1.8 A of the nitrogen ␧ of H166. The
acetyl-CoA and hexanoyl-CoA are oriented to similar distance
˚
˚
˚
and the hydroxyl group of benzyl alcohol was 2.73 A, in an
from nitrogen ␧ of H166 (3.51 A and 2.13 A, respectively). Inter-
estingly, another complex acetyl-CoA ethanol VpAAT1 showed a
similar binding to the complex formed by acetyl-CoA benzyl-
alcohol VpAAT1, with a distance between the carboxylic group
adequate position to stabilize this substrate through a hydro-
gen bond (Fig. 5b). In the case of hexanoyl-CoA and ethanol,
a shorter distance could be observed for both substrates, sug-
gesting a stronger interaction and high affinity (Fig. 5c). The
main difference between acetyl-CoA benzyl-alcohol VpAAT1 com-
plex and hexanoyl-CoA ethanol VpAAT1 complex is the alcohol
˚
of Asp170 and the hydroxyl group of ethanol of 3.02 A, mean-
while the distance between the carbonyl group of acetyl-CoA
˚
and nitrogen ␧ of His166 was 2.6 A (Fig. 5d).Interestingly, the

L. Morales-Quintana et al. / Journal of Molecular Graphics and Modelling 29 (2011) 635–642
639
Fig. 2. (a) Structural model of VpAAT1. The ␣-helices are shown in purple, the 310 helices are shown in blue, and the ␤-strands are shown in yellow. Image obtained with VMD
software. NewCartoon representation. (b) The catalytic residues of VpAAT1 protein model: the H166 and D170 residues are exposed to the solvent channel of the enzyme,
allowing their interaction with substrates.
two complexes acetyl-CoA ethanol VpAAT1 and acetyl-CoA benzyl
alcohol VpAAT1, showed a similar binding position for the alco-
hol (which interacts with residue D170) and the acetyl-CoA
(which interacts with residue His166) (Figs. 5b and d). Albeit,
in the hexanoyl-CoA ethanol VpAAT1 complex, the ethanol is
located between residues His166 and Asp170, showing a shorter
distance to both residues than the other complexes. However,
when the interaction energies were calculated for the complexes,
there is a great difference between the complexes (Table 3). The
acetyl-CoA benzyl alcohol VpAAT1 complex, that produces benzyl
acetate, showed the lowest interaction energy (−266 kcal mol−1),
and the complex acetyl-CoA ethanol VpAAT1 (that produces ethyl
acetate) has the highest energy value (−73.3 kcal mol−1). When the
acetyl-CoA hexanol VpAAT1 and butanoyl-CoA ethanol VpAAT1
complexes were modelled the alcohol took a position apart from
His166 and close to Asp170, similar to what was observed for ben-
zyl alcohol and ethanol when acetyl-CoA was the other substrate
(data not shown).
3.4. Kinetic studies
Kinetic properties of the recombinant VpAAT1 enzyme were
determined to validate ‘in silico’ measurements. The KM and Vmax
values were determined for benzyl alcohol using acetyl-CoA as co-
substrate at saturating concentration (1.5 mM acetyl-CoA), and for
acetyl-CoA using benzyl alcohol as co-substrate at saturating con-
centration (2 mM benzyl alcohol) (Table 4). A big difference in KM
values was observed, with lower KM value for acetyl-CoA (80 ␮M)
Fig. 3. The two domains of the VpAAT1 structure. (a) and (b) represent orthogonal views of the VpAAT1 structure as depicted in NewCartoon representation. N and C denote
the ends of VpAAT1. The secondary structure elements are labeled (H␣1–H␣11, 310H1–310H3 and ␤1–␤15), and domains 1 and 2 are indicated. The ␣-helices are shown in
purple, the 310 helices are shown in blue, and ␤-strands are shown in yellow. The large crossover loop (amino acids 210–232) that connects both domains is marked in red.

640
L. Morales-Quintana et al. / Journal of Molecular Graphics and Modelling 29 (2011) 635–642
Fig. 4. Evaluation of structural model of VpAAT1 enzyme. ProSA energy profile (a) and Verify3D graph (b) drawn for the VpAAT1 model.
than for benzyl alcohol (130 ␮M), indicating a higher affinity of the
enzyme for acetyl-CoA. This result is coincident with the informa-
tion provided by docking simulations (Table 2).
its protein sequence shares all the conserved regions described for
other acyltransferases belonging to the BAHD superfamily [8]. A
large number of acyltransferases have been described in plants,
nevertheless in fruit species such as melon and apple a small gene
family has been identified [9,26,27].The BAHD superfamily mem-
bers have been divided into three subfamilies (I, II and III) [27]
and VpAAT1 has been clustered into subgroup III [8]. Members of
subfamily III have been described for their great capacity to pro-
duce benzyl acetate [28], and interestingly, benzyl acetate has not
been described as an aroma volatile in mountain papaya fruit [1],
4. Discussion
Mountain papaya is a climacteric fruit that develops an interest-
ing and characteristic aroma during ripening, which is mainly due
to esters [1], which are synthesized by alcohol acyltransferases. A
gene (VpAAT1) has been isolated from mountain papaya fruit and
Fig. 5. Ligand binding analysis based on the refined VpAAT1 model. (a) The 3D structure of acetyl-CoA benzyl alcohol VpAAT1 complex is shown. (b) Docking of benzyl
alcohol and acetyl-CoA to VpAAT1 model. (c) Docking of ethanol and hexanoyl-CoA to VpAAT1 model. (d) Docking of ethanol and acetyl-CoA to VpAAT1 model. Dash lines
indicate distance between atoms.

L. Morales-Quintana et al. / Journal of Molecular Graphics and Modelling 29 (2011) 635–642
641
Table 3
Affinity energies for the different acyl-CoA alcohol VpAAT1 modelled complexes.
a
Acyl-CoA
Alcohol
Ester produced
AAT activity (pkat mg−1
)
Affinity energy (kcal mol−1
)
rmsd (Å)b
Benzyl alcohol
Hexanol
Ethanol
Ethanol
Ethanol
Benzyl acetate
Hexyl acetate
Ethyl acetate
Ethyl butanoate
Ethyl hexanoate
29,275
20
tr
83
790
−266.3
−222.9
−73.3
2.4
4.1
4.4
5.7
1.9
Acetyl-CoA
Butanoyl-CoA
Hexanoyl-CoA
−230.9
−250.2
a
AAT activity values of VpAAT1 recombinant protein towards different acyl-CoAs and alcohols as substrates were obtained from [8].
rms deviation was calculated for the backbone (C␣,C,N) atoms. tr, traces.
b
Table 4
the recombinant protein and the affinity energy determined for
each protein–ligand complex: a higher VpAAT1 activity was mea-
sured when the complex has higher stability (lower affinity energy)
(Table 3).
Kinetic parameters obtained for purified recombinant VpAAT1 protein using cosub-
strate at saturating concentration.
Substrate
KM (␮M)
130
Vmax (pmol h−1 mg−1
)
Hexyl acetate, ethyl butanoate and ethyl hexanoate are volatiles
produced by mountain papaya fruit, and their production increased
during ripening progress. These esters are characterized for its
fruity notes and low odour threshold values in water (1–2 ppb), and
in mountain papaya fruit ethyl butanoate and ethyl hexanoate are
the most potent odourants [1]. Their increment during ripening is
coincident with the increase in the expression level of VpAAT1 tran-
scripts and total AAT activity in fruit extracts [8]. The low affinity
energies determined for the complex between the VpAAT1 protein
and each corresponding pair of ligands indicates that the ester-
ification reaction could be performed by the VpAAT1 protein. In
fact, the VpAAT1 recombinant protein is able to synthesize them,
suggesting that VpAAT1 protein is directly involved in the biosyn-
thesis of these important esters. On the other hand, ethyl acetate
is also profusely produced by mountain papaya fruit, however its
production starts late during ripening of the fruit and its increment
continues even after the climacteric phase. The VpAAT1 recom-
binant protein is unable to produce ethyl acetate and also the
complex VpAAT1 acetyl-CoA ethanol has an unfavorable affinity
binding energy. Therefore, it can be proposed that ethyl acetate is
most probably produced in vivo by another alcohol acyltransferase
and not by the VpAAT1 protein in mountain papaya fruit.
In summary, our study strongly suggests that the struc-
tural model of VpAAT1 enzyme is a useful tool to study the
protein–substrate interaction, and reveals that the catalytic site is
located into the solvent channel at the center of the protein struc-
ture. Moreover, docking simulations determine that the enzyme
has a lower binding energy for benzyl alcohol and acetyl-CoA than
for other pair of substrates, which correlates well with in vitro
enzymatic activity data. As it has been shown, the combination of
bioinformatics and kinetics assays are necessary to validate ‘in sil-
ico’ predictions and to determine the more favorable interaction
between an enzyme and its substrates.
Benzyl alcohol
Acetyl-CoA
27,950
30,578
80
Saturation curves were prepared for each substrate, while the other was remained
constant at a saturating concentration (1.5 mM acetyl-CoA, 2 mM benzyl alcohol).
35 ␮g of purified VpAAT1 protein were employed in each enzymatic assay.
although the recombinant protein is able to synthesize this ester
(Table 3).
To understand the structural and functional characteristics, the
3D structure of VpAAT1 protein was built by homology modelling
and molecular dynamic simulations. The model obtained fulfils
the structural characteristics proposed for the BAHD superfamily,
including the solvent channel in where the active site is located,
and the globular shape composed by two equally sized domains
connected by an extensive loop (residues 210–232). The VpAAT1
structure presented in this work explains the functional importance
of His166, a residue located in loop 13 between helix 5 and ␤-strand
7, and directly situated in the center of the solvent channel. This
structural arrangement allows the incorporation of different acyl-
CoAs and alcohols as substrates to approach the active site, at the
front face (acyl-CoA binding) and the back face (alcohols binding)
of the enzyme. In the HTMSD motif, the Asp170 is located close
to His166 and it is part of the active site. The importance of the
residues His and Asp in the HXXXD motif has also been demon-
strated by chemical modification and site-directed mutagenesis
experiments for other members of the BAHD superfamily [10,29].
The BAHD superfamily enzymes might therefore have a similar
conformation of the catalytic His and use a similar reaction mech-
anism to that proposed for VpAAT1.Molecular docking analysis
indicates that interaction between VpAAT1 enzyme with acetyl-
CoA and benzyl alcohol showed the most favorable binding energy,
and the recombinant protein displayed a high activity towards the
synthesis of benzyl acetate, so the 3D models show specific inter-
actions between the substrates benzyl alcohol and acetyl-CoA and
key amino acid residues in the active site, and thus these inter-
actions are consistent with the previously reported experimental
data concerning the catalytic activity of VpAAT1 [8]. Moreover, the
analysis performed supports the hypothesis that VpAAT1 enzyme
shares similar structural and functional characteristics to members
of the acyltransferase subfamily III with high activity for the pro-
duction of benzyl acetate. Therefore, the fact that mountain papaya
fruit does not produce benzyl acetate may be due to the absence of
substrates during fruit ripening and not to the inability of VpAAT1
protein to interact with the substrates.
Acknowledgements
L. Morales-Quintana acknowledges CONICYT for a doctoral
fellowship. L. Fuentes and C. Gaete-Eastman acknowledge the
PBCT Postdoctoral programme for financial support. This project
was supported by a PBCT Anillo ACT-41 programme. The group
acknowledges Mauricio Arenas and the Center of Bioinformatics
and Molecular Simulations (CBSM) of the University of Talca for
providing ICM programs.
Recently, we described that VpAAT1 recombinant protein was
able to use different alcohols and acyl-CoAs as substrates for the
synthesis of esters [8]. We now show that the binding complex
energies between the VpAAT1 protein and its different substrates
(acyl-CoAs and alcohols) were significantly different. A good cor-
relation was observed between the VpAAT1 activity recorded for
References
[1] C. Balbontín, C. Gaete-Eastman, M. Vergara, R. Herrera, M.A. Moya-León, Treat-
ment with 1-MCP and the role of ethylene in aroma development of mountain
papaya fruit, Postharvest Biol. Technol. 43 (2007) 67–77.
[2] V.M. Badillo, Carica L. vs. Vasconcella St. Hil. (Caricaceae): con la rehabilitación
de este último, Ernstia 10 (2000) 74–79.

642
L. Morales-Quintana et al. / Journal of Molecular Graphics and Modelling 29 (2011) 635–642
[3] B. Van Droogenbroeck, P. Breyne, P. Goetghebeur, E. Romeijn-Peeters, T. Kyndt,
[16] M. Schlenkrich, J. Brickmann, A.D. MacKerell Jr., M. Karplus, in: K.M. Merz,
B. Roux (Eds.), A Molecular Perspective from Computation and Experiment,
Birkhauser, Boston, MA, 1996, pp. 31–81.
[17] W.L. Jorgensen, J. Chandresekhar, J.D. Madura, R.W. Impey, M.L. Klein, Compar-
ison of simple potential functions for simulating liquid water, J. Chem. Phys. 79
(1983) 926–935.
G. Gheysen, AFLP analysis of genetic relationships among papaya and its wild
relatives (Caricaceae) from Ecuador, Theor. Appl. Genet. 105 (2002) 289–297.
[4] M.A. Moya-León, M. Moya, R. Herrera, Ripening of the Chilean papaya fruit
(Vasconcellea pubescens) and ethylene dependence of some ripening events,
Postharvest Biol. Technol. 34 (2004) 211–218.
[5] C.F. Forney, W. Kalt, M.A. Jordan, The composition of strawberry aroma is influ-
enced by cultivar, maturity and storage, HortScience 35 (2000) 1022–1026.
[6] M.A. Moya-León, M. Vergara, C. Bravo, M.E. Montes, C. Moggia, 1-MCP treat-
ment preserves aroma quality of ‘Packham’s Triumph’ pears during long-term
storage, Postharvest Biol. Technol. 42 (2006) 185–197.
[18] M.J. Sippl, Recognition of errors in three-dimensional structures of proteins,
Proteins 17 (1993) 355–362.
[19] R.A. Laskowski, M.W. MacArthur, D.S. Moss, J.M. Thornton, PROCHECK: a
program to check the stereochemical quality of protein structures, J. Appl.
Crystallogr. 26 (1993) 283–291.
[7] M.A. Moya-León, M. Vergara, C. Bravo, M. Pereira, C. Moggia, Development of
aroma compounds and sensory quality of ‘Royal Gala’ apples during storage, J.
Hort. Sci. Biotechnol. 82 (2007) 403–413.
[8] C. Balbontin, C. Gaete-Eastman, L. Fuentes, C.R. Figueroa, R. Herrera, D. Man-
riquez, A. Latche, J.C. Pech, M.A. Moya-León, VpAAT1, a gene encoding an alcohol
acyltransferase, is involved in ester biosynthesis during ripening of mountain
papaya fruit, J. Agric. Food Chem. 58 (2010) 5114–5121.
[20] R. Lüthy, J.U. Bowie, D. Eiserberg, Assessment of protein models with three-
dimensional profiles, Nature 356 (1992) 83–85.
[21] R. Abagyan, M. Totrov, D. Kuznetsov, ICM—a new method for protein modeling
and design: applications to docking and structure prediction from the distorted
native conformation, J. Comput. Chem. 15 (1994) 488–506.
[22] J.C. D’Auria, Acyltransferases in plants: a good time to be BAHD, Curr. Opin.
Plant Biol. 9 (2006) 331–340.
[9] I. El-Sharkawy, D. Manriquez, F.B. Flores, F. Regad, M. Bouzayen, A. Latche,
J.C. Pech, Functional characterization of a melon alcohol acyl-transferase gene
family involved in the biosynthesis of ester volatiles. Identification of the cru-
cial role of a threonine residue for enzyme activity, Plant Mol. Biol. 59 (2005)
345–362.
[10] X. Ma, J. Koep, S. Panjikar, G. Fritzch, J. Stockigt, Crystal structure of vinorine
synthase, the first representative of the BAHD superfamily, J. Biol. Chem. 280
(2005) 13576–13583.
[23] M.M. Bradford, A rapid and sensitive method for the quantification of micro-
gram quantities of protein utilizing the principle of protein dye binding, Anal.
Biochem. 72 (1976) 248–254.
[24] J.K. Fellman, D.S. Mattinson, B.C. Bostick, J.P. Mattheis, M.E. Patterson, Ester
biosynthesis in ‘Rome’ apples subjected to low oxygen atmospheres, Posthar-
vest Biol. Technol. 3 (1993) 201–214.
[25] M. González, C. Gaete-Eastman, M. Valdenegro, C.R. Figueroa, L. Fuentes, R.
Herrera, M.A. Moya-León, Aroma development during ripening of F. chiloensis
fruit and participation of an alcohol acyltransferase (FcAAT1) gene, J. Agric.
Food Chem. 57 (2009) 9123–9132.
[26] E.J. Souleyre, D.R. Greenwood, E.N. Friel, S. Karunairetnam, R.D. Newcomb, An
alcohol acyltransferase from apple (cv. Royal Gala), MpAAT1, produces esters
involved in apple fruit flavor, FEBS J. 272 (2005) 3132–3144.
[27] D. Li, Y. Xu, G. Xu, L. Gu, D. Li, H. Shu, Molecular cloning and expression of a gene
encoding alcohol acyltransferase (MdAAT2) from apple (cv. Golden Delicious),
Phytochemistry 67 (2006) 658–667.
[28] J. D’Auria, F. Chen, E. Pichersky, Characterization of an acyltransferase capable of
synthesizing benzylbenzoate and other volatile esters in flowers and damaged
leaves of Clarkia breweri, Plant Physiol. 130 (2002) 466–476.
[29] H. Suzuki, T. Nakayama, T. Nishino, Proposed mechanism and functional
amino acid residues of malonyl-CoA:anthocyanin 5-O-glucoside-6ꢀꢀꢀ-O-
malonyltransferase from flowers of Salvia splendens, a member of the versatile
plant acyltransferase family, Biochemistry 42 (2003) 1764–1771.
[11] A. Bayer, X.Y. Ma, J. Stockigt, Acetyltransfer in natural product
biosynthesis—functional cloning and molecular analysis of vinorine synthase,
Bioorg. Med. Chem. 12 (2004) 2787–2795.
[12] D.T. Jonnes, Protein secondary structure prediction based on position specific
scoring matrices, J. Mol. Biol. 292 (1999) 195–202.
ˇ
[13] A. Sali, T.L. Blundell, Comparative protein modelling by satisfaction of spatial
restraints, J. Mol. Biol. 234 (1993) 779–815.
[14] J.C. Phillips, R. Braun, W. Wang, J. Gumbart, E. Tajkhorshid, E. Villa, C. Chipot, R.D.
Skeel, L. Kale, K. Schulten, Scalable molecular dynamics with NAMD, J. Comput.
Chem. 26 (2005) 1781–1802.
[15] A.D. MacKerell Jr., D. Bashford, M. Bellott, R.L. Dunbrack Jr., J. Evanseck, M. Field,
J.S. Fischer, J. Gao, H. Guo, S. Ha, D. Joseph, L. Kuchnir, K. Kuczera, F.T.K. Lau, C.
Mattos, S. Michnick, T. Ngo, D.T. Nguyen, B. Prodhom, B. Roux, M. Schlenkrich,
J. Smith, R. Stote, J. Straub, M. Watanabe, J. Wiorkiewicz-Kuczera, D. Yin, M.
Karplus, Self-consistent parameterization of biomolecules for molecular mod-
eling and condensed phase simulations, FASEB J. 6 (1992) A143.