Computational Biology and Chemistry 76 (2018) 79–86

Contents lists available at ScienceDirect

Computational Biology and Chemistry

journal homepage: www.elsevier.com/locate/cbac

Research Article

Computational study of FaEXPA1, a strawberry alpha expansin protein, T

through molecular modeling and molecular dynamics simulation studies
⁎⁎ ⁎
Felipe Valenzuela-Riffoa, Patricio Ramosb,e, , Luis Morales-Quintanac,d,
a
Phytohormone Research Laboratory, Instituto de Ciencias Biológicas, Universidad de Talca, Chile
b
Instituto de Ciencias Biológicas, Universidad de Talca, Chile
c
Multidisciplinary Agroindustry Research Laboratory, Universidad Autónoma de Chile, Chile
d
Instituto de Ciencias Biomédicas, Universidad Autónoma de Chile, Talca, Chile
e
Núcleo Científico Multidiciplinario-DI, Universidad de Talca, Chile

A R T I C LE I N FO A B S T R A C T

Keywords: Changes in the cellulose-hemicellulose fraction take place during ripening of strawberry fruit and are associated
Expansin protein with the activity of a set of proteins and hydrolytic enzymes. Expansins are proteins located in the cell wall with
Molecular dynamic simulation no catalytic activity. In this context, FaEXPA1 was previously reported to have a high accumulation rate during
Strawberry fruit ripening in three different strawberry cultivars. In order to understand at the molecular level the expansin
Plant cell wall
mechanism mode, a 3D model of FaEXPA1 protein was built by comparative modeling. FaEXPA1 protein model
Cellulose
displayed two domains, a cellulose-binding domain with a β-sandwich structure, and a second domain that
included a HFD motif with a similar structure to the catalytic core of endoglucanase V from Humicola insolens.
Additionally, in the center of the structure, an open groove was formed. Finally, using a cellulose polymer as a
ligand, the protein-ligand interaction was evaluated by molecular dynamic (MD) simulation. Two MD simula-
tions showed that FaEXPA1 can interact with cellulose via the flat aromatic surface of its binding domain D2,
composed mainly of residues Trp99 and Trp225. In addition, FaEXPA1 formed a high number of hydrogen bonds
with the glycan chain and the Asn81, Phe114 and Asn211 residues.

1. Introduction
Plant cell walls are a complex and dynamic supra-molecular as-
sembly composed of crystalline cellulose microfibrils surrounded by an
amorphous matrix of polysaccharides such as hemicellulose and pectins
as well as inorganic molecules and proteins (Chundawat et al., 2011;
Johansson et al., 2004; Vaaje-Kolstad et al., 2010). The plant cell wall
regulates growth and development, mechanical support, and cell shape
and acts as a barrier against biotic and abiotic stresses (Pilling and
Höfte, 2003; Nardi et al., 2015), and changes in cell wall architecture
have been well described (Marga et al., 2005).
Regarding plant cell wall metabolism, disassembly is the main
process that leads to fruit softening during ripening and postharvest
(Vicente et al., 2007). For this reason, several enzymes and proteins
have been widely studied (Goulao and Oliveira, 2008; Quesada et al.,
2009; Mercado et al., 2011). One class of these proteins, which are
actively implicated in cell wall disassembly, is the expansins (Bennett,
2002).
The plant-specific superfamily of expansin proteins constitutes an
ancient and major gene family known to have roles in regulating di-
verse biological processes in plants (Li et al., 2016). Expansins char-
acteristically loosen the plant cell wall by weakening the non-covalent
bonding of polysaccharides to one another (McQueen-Mason et al.,
1992; McQueen-Mason and Cosgrove, 1994; Cosgrove, 2000, 2005; Li
et al., 2003; Sampedro and Cosgrove, 2005). In this way, the expansin
proteins exert their molecular function by disrupting hydrogen bonds
between cellulose microfibrils and xyloglucans that hold them jointly
within plant cell walls (McQueen-Mason and Cosgrove, 1994; Mcqueen-
Mason and Cosgrove, 1995; Whitney et al., 2000) to promote the re-
laxation of the cell wall structure in a pH-dependent manner, allowing
access of different hydrolases to their substrates (McQueen-Mason
et al., 1992; McQueen-Mason and Cosgrove, 1994; Cosgrove, 2000,
2005 Cho and Cosgrove, 2000; Choi et al., 2003; Gray-Mitsumune et al.,
2008).
The expansin proteins contain 250–275 amino acids and are com-
posed of two domains. Domain 1 (D1) is in the N-terminal region and
includes a series of conserved cysteines and three residues (His, Phe,
and Asp) that are important for its protein activity (the HFD motif)

⁎
Corresponding author at: Instituto de Ciencias Biomédicas, Universidad Autónoma de Chile, Chile.
⁎⁎
Corresponding author.
E-mail addresses: pramos@utalca.cl (P. Ramos), luis.morales@uautonoma.cl (L. Morales-Quintana).

https://doi.org/10.1016/j.compbiolchem.2018.05.018
Received 2 January 2018; Received in revised form 26 April 2018; Accepted 15 May 2018
Available online 26 May 2018
1476-9271/ © 2018 Elsevier Ltd. All rights reserved.
F. Valenzuela-Riffo et al.
(Cosgrove, 2000); the second domain (D2) in the C-terminal region is
responsible for carbohydrate binding, which is distantly related to
group-2 grass pollen allergens (Cosgrove, 2000). Based on phylogenetic
and conserved sequence analyses, the expansin superfamily has been
organized into four expansin sub-families: α-expansin and β-expansin
are found in all groups of land plants and have diverse functions, while
Like α-expansin and Like β-expansin proteins are found in both plants
and bacteria (Sampedro and Cosgrove, 2005).
Strawberry (Fragaria × ananassa) is one of the most economically
and nutritionally important fruit crops in the world, and as a member of
the rosaceae family, it is used as a model species for molecular biology
and genomic studies. Strawberry fruit is characterized by a high soft-
ening rate, short postharvest life and fast decay (Bustamante et al.,
2009). Its fast softening rate during fruit ripening is coincident with a
decreased content of cellulose and hemicellulose in the cell wall (Rosli
et al., 2004). The rapid decrease in fruit firmness observed in straw-
berry correlates with a large increase in the accumulation of transcripts
from different genes including FaEXPA1, which codes for an alpha ex-
pansin protein (Dotto et al., 2006).
Although well studied, there is still a lack of understanding of the
molecular mechanisms involved in cell wall remodeling in strawberry.
Although the effect of expansins on cell walls is known, the molecular
mechanism underlying their biophysical function is poorly understood.
In the present study, we focused on molecular aspects related to
strawberry fruit softening. A 3-D model of the FaEXPA1 protein was
built and used to evaluate its interaction with cellulose polymers as a
ligand. Finally, the mechanism of expansin action at the molecular level
was analyzed, suggesting a putative interaction between FaEXPA1 and
cellulose as the substrate.
2. Materials and methods
2.1. Obtaining the structure of FaEXPA1 using molecular modeling
FaEXPA1 structural model was obtained by comparative modeling
methodology using MODELLER 9v12 software (http://salilab.org/
modeller/) (Šali and Blundell, 1993). The comparative model was
generated according to the method employed by Morales-Quintana
et al. (2011). The crystal structure of a β-expansin from Zea mays (PDB
code: 2HCZ) was selected as the template based on the sequence
identity between both sequences. Five models were generated for the
FaEXPA1 structure, and the model with the lowest RMSD value with
respect to trace (Cα atoms) of the crystal structure of the template was
saved for further refinement and validation. An SPC water model was
used in the molecular dynamics (MD) simulation. Particle Mesh Ewald
was used to calculate long-range electrostatic interactions. The system
was neutralized by adding ions (Na+ or Cl−) until reaching 150 mM
NaCl (specifically 105 Na+ and 106 Cl− atoms were added). After
that, SCHRÖDINGER suite with the OPLS v2005 force field (Jorgensen
et al., 1996) was used to equilibrate the system in a MD simulation of
10 ns. The protein protonation state was set to pH 4.5 using the
PROPKA program and SCHRÖDINGER suite, and as according to Wang
et al. (2008), who described it as the pH at which the expansins have
the highest activity. To evaluate the quality of the model, both PROC-
HECK (Laskowski et al., 1993) and ProSA-Web (Sippl, 1993;
Wiederstein and Sippl, 2007) programs were employed.
2.2. Protein-ligand interaction
A molecular dynamics (MD) simulation of a cellulose chain com-
posed of 30 glucose monomers linked by β1-4 bonds was used as the
ligand. The cellulose structure was built using the GLYCAM web server
(http://glycam.org). At the beginning of the MD simulation, the ligand
was located at 15 Å of the protein and the system was constructed so
that each atom movement was completely free. The complex was em-
bedded into a pre-equilibrated SPC water model and neutralized by
80
Computational Biology and Chemistry 76 (2018) 79–86
adding ions (Na+ or Cl−) until reaching 150 mM NaCl (specifically
105 Na+ and 106 Cl− atoms were added). Each MD simulation was
performed at a constant temperature (300 K) and pressure
(1.01325 bar), with an NPT ensemble. The MD simulation was run over
100 ns, and the motion equations were integrated with 2 fs, while the
data were collected for every 50 ps of trajectory. SCHRÖDINGER suite
with the OPLS v2005 force field (Jorgensen et al., 1996) was used in the
MDS. Finally, the MDS was analyzed using VMD software (Humphrey
et al., 1996).
2.3. In silico site-directed mutagenesis of FaEXPA1
To gain insight into the action mechanism of FaEXPA1 and to clarify
the role of the residues Asn81, Phe114, and Asn211 in their interaction
with cellulose, the mutants Asn81Ala, Phe114Ala and Asn211Ala were
generated from the FaEXPA1 wild-type structural model using alanine
screening methodology (Morrison and Weiss, 2001). A short molecular
dynamics was run in order to remove wrong contacts and to fill empty
pockets using SCHRÖDINGER suite with OPLS v2005 force field
(Jorgensen et al., 1996) according to Morales-Quintana et al. (2013).
All MD simulations were done using a time step of 1 fs during 1 ns of
MD simulation.
Additionally, PROCHECK (Laskowski et al., 1993) and ProSA-Web
(Sippl, 1993; Wiederstein and Sippl, 2007) programs were employed to
evaluate the mutant models generated. Finally, MD simulations were
performed to predict the effect of the mutant on the protein-ligand
interaction using a similar methodology to that previously described
above for the native model.
3. Results and discussion
3.1. 3-D structure of FaEXPA1
Plant and microbial expansins share a higher degree of structural
similarity than their low sequence identity (around 20%) might suggest
(Georgelis et al., 2012). To determine the structural properties of the
FaEXPA1 sequence, a FaEXPA1 protein model was built using the in-
formation provided by the sequence alignment between the template
2HCZ and the FaEXPA1 sequence (Fig. S1). The sequence identity value
was 27.8%, and similarity was 47.8% (Fig. S1). A complete evaluation
(geometric and energetic) of the model was performed. The RMSD
value of the backbone calculated between the strawberry expansin and
the template was 3.09 Å, and the model showed similar secondary
structure; however, small differences were found in the orientation of
the loops present in FaEXPA1 compared to the template (Fig. 1a). The
stereochemical quality of the model was evaluated using the PROCH-
ECK program indicating that 78.6% of the residues were classified in
the most favored regions, and only one residue was considered as badly
modeled (the Gln225 residue, located far away from the catalytic motif
and open groove) (Table 1). ProSA analysis of individual residues of the
FaEXPA1 showed a low z-score in most of the structurally conserved
regions. However, a small region with an unfavorable z-score in the N-
terminal region (the first eight residues) was identified, which is distant
from the active site and out of the protein open groove (data not
shown). Additionally, the global z-score obtained from the ProSA ana-
lysis over all structures of FaEXPA1 was −5.46, while the global z-score
of the template was −6.68. Both values were within the accepted range
for proteins of this size according to the program parameters
(Wiederstein and Sippl, 2007). Previously, a partial structural similarity
has been described between D1 of expansins and the catalytic site of
endoglucanases that belong to family 45 (Sampedro and Cosgrove,
2005). For this, a structural alignment between D1 of FaEXPA1 and
endoglucanase V from Humicola insolens (EGV) (PDB code of 2ENG) was
evaluated (Fig. 1b). A good overlap between the two β-barrels was
observed, but not in the rest of the structure, which mainly consisted of
loops and short α-helices according to a previous report by Gaete-
F. Valenzuela-Riffo et al. Computational Biology and Chemistry 76 (2018) 79–86

Fig 1. Structural alignment of FaEXPA1 with the template. A) Structural superposition of the FaEXPA1 model and the 2HCZ structure. The cartoon model of the
template is colored in red and that of FaEXPA1 in blue. B) Structural superposition of the FaEXPA1 protein model and the EGV structure. The cartoon model of
endoglucanase V from Humicola insolens (EGV) (PDB: 2ENG) is colored in blue and that of FaEXPA1 domain 1 in red. (For interpretation of the references to colour in
this figure legend, the reader is referred to the web version of this article.)

Table 1 Eastman et al. (2015), indicating a good structural quality of D1.

Validation of FaEXPA1 protein structure using PROCHECK program Consequently, the final structure of the FaEXPA1 expansin protein was
(Ramachandran plot). accepted for subsequent analysis.
Core (%) a
Allow (%) b
Gener (%) c
Disall (%) d The FaEXPA1 comparative model was included in the Protein Model
DataBase (PMDB; https://bioinformatics.cineca.it/PMDB/) under the
FcEXPA1 78.6 17.9 2.6 0.9 PMDB code PM0081200.
a The model obtained showed the structural characteristics proposed
Most favourable region.
b for the expansin protein family (Sampedro and Cosgrove, 2005). Re-
Additional allowed regions.
c garding this structure, the ‘catalytic’ domain (D1) showed a highly
Generously allowed regions.
d conserved β-barrel fold formed by residues between the Arg1 and
Disallowed regions.
Arg159 residues (the protein was numbered according to the mature
protein) (Fig. 2a, green structure). The second domain is the

81
F. Valenzuela-Riffo et al.
carbohydrate-binding module (CBM) (D2), which showed an anti-
parallel ß-sandwich fold (Phe171–Val262 residues) (Fig. 2a, orange
structure). The domains are connected through a short loop (Va-
l160–Arg170 residues) that spans 15.4 Å (Fig. 2a, blue loop). Ad-
ditionally, in the center of the protein, an open groove was formed
(Fig. 3b, cyan surface), which is a highly conserved region and dis-
tinctive of the EXPA protein family (Gaete-Eastman et al., 2015). Six
conserved cysteine residues in positions 33, 61, 64, 69, 76 and 139 (in
the mature protein) allowed the formation of three disulfide cross-lin-
kages (C33–C61, C64–C139, and C69–C76) (Fig. 1, red asterisks)
(Fig. 2a, yellow residues) in agreement with other expansins previously
reported (Gaete-Eastman et al., 2015; Cosgrove et al., 1997; Mateluna
et al., 2017). Among those, three-disulfide bonds are strictly conserved
in both the alpha and beta expansin families, suggesting that they are
critical for protein folding maintenance of expansins (Gaete-Eastman
et al., 2015; Mateluna et al., 2017). Interestingly, even though there is a
fourth cysteine pair (highly conserved in α-expansins) (Fig. 3), this
bond could not be modeled due to the long distance between these
residues (data not shown), similar to the VpEXPA2, PrEXPA1, PrEXPA2,
PrEXPA3 and PrEXPA4 expansin models previously described (Gaete-
Eastman et al., 2015; Mateluna et al., 2017) (Fig. 3, red asterisks).
Additionally, the HFD motif (named as the ‘catalytic’ motif) (Fig. 1a,
red residues) is located in the center of the structure of D1 and oriented
to the groove between the two loops, which corresponds to residues
21–45 and 70–82, respectively. This orientation was previously
82
Computational Biology and Chemistry 76 (2018) 79–86
Fig. 2. FaEXPA1 protein model. The 3D-structure was built
using the 2HCZ protein as a template (27.8% identity). A) The
FaEXPA1 structure is composed of two domains: Domain 1 (D1)
shown in green that has 6 β-strands and 1 α-helix forming a β-
barrel with the helices to the outside, while Domain 2 (D2) shown
in orange has 8 β-strands, assembled in two anti-parallel β-sheets.
The red residues correspond to the HFD motif; the cysteine is
shown in yellow, and aromatic residues of D2 are shown in purple.
B) Surface of the FaEXPA1 protein model shown in dark gray,
while the open groove is shown in cyan. (For interpretation of the
references to colour in this figure legend, the reader is referred to
the web version of this article.)
Fig. 3. Alignment of the deduced full-length amino acid se-
quence of FaEXPA1 with various expansin sequences. Protein
sequence alignment of FaEXPA1 and different expansin proteins.
Identical residues are indicated by black boxes and similar re-
sidues by gray boxes. Red asterisks (*) show the three pairs of Cys
residues that form the three disulfide bonds, which are preserved
in both alpha and beta expansins, while the two green asterisks
show the pair of Cys residues present only in alpha expansins that
cannot form disulfide bonds. Additionally, the symbols (+) show
the HFD motif concerned in alpha and beta expansins. (For in-
terpretation of the references to colour in this figure legend, the
reader is referred to the web version of this article.)
described in the template (pdb code: 2HCZ) structure (Gaete-Eastman
et al., 2015) and in other alpha and beta expansins obtained by mole-
cular modeling methodology (Morrison and Weiss, 2001; Georgelis
et al., 2012). The HFD motif has been described as important for the
protein-ligand interaction by in silico approaches that showed the im-
portance of the aspartate residue in the interaction with a cellulose
octamer (Morrison and Weiss, 2001). Additionally, the HFD motif was
described as important for the activity of the EGV protein, and EGVs are
evolutionarily related to expansins (Cosgrove et al., 1997). The muta-
tion of the aspartate residue of the HFD motif into asparagine (As-
p114Asn) in EGV showed that the EGV-D114N mutant protein reduced
the activity and the Kcat/Km ratio by 20-fold and 160-fold less, re-
spectively (Cosgrove et al., 1997).
The ‘glycine-rich loop’ located between the Phe73 and Pro91 re-
sidues, which was previously described as important in the protein-li-
gand interaction and participates in the conformational change during
the binding of the ligand (Morrison and Weiss, 2001), displayed a low
similarity compared to the template (Fig. 3a, black arrow).
The D2 domain is a CBM type A according to Kerff et al. (2008)
(Morales-Quintana et al., 2013), and D2 has been recognized as an
independent family of CBMs assigned to the CAZy database family 63,
mainly because D2 showed a low sequence similarity to other CBMs
from hydrolytic enzymes, and additionally it has a different biological
role and does not need metal ions as co-factors (Gilbert et al., 2013;
Silveira and Skaf, 2016). The structure shows a β-sandwich fold (which
F. Valenzuela-Riffo et al.
is a common fold among CBMs) with polar (Arg, Glu, Lys, Thr, and Asp)
and aromatic (Trp, Phe and Tyr) residues oriented to the surface
(Fig. 2a, purple residues). We also corroborated that the aromatic re-
sidues are organized in an extended strip at the edges of the open
groove forming a planar surface, similar to other alpha expansins
(Morrison and Weiss, 2001; Mateluna et al., 2017). These residues are
located along the open groove, which could interact with poly-
saccharides that cover both protein domains, as in ZmEXPB1 and car-
bohydrate-binding module (CBM) proteins (Yennawar et al., 2006;
Gilbert et al., 2013). Also, these aromatic residues could facilitate the
movement of the expansin protein on the cellulose surface, similar to
the mechanism proposed for cellulases (Jervis et al., 1997) and the
EXLX1 protein, an expansin-like protein from Bacillus subtilis (Silveira
and Skaf, 2016). In this line, the MD simulation results showed that
FaEXPA1 interacts with cellulose polymers via the two Trp residues in
CBM. The importance of this aromatic residue is well described
(Georgelis et al., 2012; Gaete-Eastman et al., 2015; Mateluna et al.,
2017; Jervis et al., 1997; Silveira and Skaf, 2016; Georgelis et al.,
2011), while in D1, the ‘glycine-rich-like’ loop (in D1) helps to orientate
the ligand along the open groove.
3.2. Protein-ligand interaction analysis
Two independent MD simulations were performed for a system with
one cellulose fibril monomer and FaEXPA1 in a water box. FaEXPA1
was placed close (around 15 Å) to the glycan chain (Fig. 4b). In both
MD simulations, the FaEXPA1 showed the shortest distances between
the glycan chain and D1 and D2 domains. However, in both MD si-
mulations, D2 was the first to approach the glycan chain (Fig. 4a, green
line). Consistently in both cases, the interaction between FaEXPA1 and
the glycan chain was led by the Trp99 residue (5 ns). After this, the
aromatic residues of D2 (Trp225 and Tyr114) approach the glycan
chain. Finally, after the interaction of D2, D1 also began to interact with
Fig. 4. Interaction between FaEXPA1 protein and cellulose polymer as ligands. A) Minimum distance obtained from two independent molecular dynamic (MD)
simulations (MD simulations 1 and 2) between FaEXPA1 D1 and D2 domains and the cellulose polymer. B) Protein-ligand complexes at different times during the MD
simulation.
83
Computational Biology and Chemistry 76 (2018) 79–86
the cellulose ligand (Fig. 4). Interestingly, during this process, all water
molecules were driven out of the interfacial region between the cellu-
lose and planar aromatic residues of D2. These results are in agreement
with previous reports showing that the binding affinities of D2 from the
B. subtilis (BsEXLX1) expansin and other type-A CBMs are 1000-fold
lower compared to soluble cellooligosaccharides than cellulose, mainly
as a consequence of high affinity for more ordered cellulose regions (as
crystalline microfibrils) (Georgelis et al., 2011). For BsEXLX1, crystal-
lographic data of the bond to cellohexose show hydrophobic bonding of
the aromatic residues Trp101 and Trp126 aligned to the plane of the
pyranose rings of the carbohydrate, while Tyr149 interacts partially
with the glucan chain (Georgelis et al., 2011) in the FaEXPA1, and a
similar orientation was found between Trp99; however, Trp225 and
Tyr149 interact inexactly with the glucan chain and do not form a plane
with the pyranose rings of the carbohydrate in the cellulose chain.
Georgelis et al. (2011) replaced the three aromatic residues present on
the surface of D2 (W125, W126 and Y157) with alanine in the BsEXLX1
protein to generate a triple mutant with no expansin activity. The au-
thors suggested that these residues on the conserved D2 surface are
important for BsEXLX1 activity. Other important D1 residues for the
BsEXLAX1 activity described were Thr14, Asp71 and Tyr73, and the
crystallographic structures of BsEXLAX1 in complex with cellohexose
showed that D1 cannot interact with this molecule. These observations
of the crystallographic data are in line with experiments where they
showed that D1 was unable to bind the substrate when it is expressed
separately from D2 (Georgelis et al., 2011; Martinez-Anaya, 2016). In
our case, we observed a small distance (less than 3 Å) in the FaEXPA1-
cellulose complex during more than 40% of the time-course simulations
in both of the MD simulations analyzed, displaying 19 residues in-
cluding the three equivalent aromatic residues (F114, F178 and W212)
identified in the D2 domain of FaEXPA1 and the HFD motif (Fig. 5a, red
residues). With respect to the ligand, the cellulose is positioned along
the open groove (Fig. 5b).
F. Valenzuela-Riffo et al. Computational Biology and Chemistry 76 (2018) 79–86

Fig. 5. Protein-ligand interaction mode of

FaEXPA1 protein models with cellulose as
the substrate. A) A scheme indicating the re-
sidues that interact with cellulose at a 3 Å or
shorter distance. B) Surface representation of
the residues involved in the protein-ligand in-
teraction.

Fig. 6. Hydrogen bonding between FaEXPA1 and cellulose

polymer. A) On the left, a graph indicating the percentage of
time that a hydrogen bond (H-bond) between particular amino
acid residues and the ligand is established in each MD simu-
lation. On the right, a general view of the interaction of the
protein and cellulose as the ligand. B) The number of H-bonds
established during each MD simulation between the ligand
and different amino acid residues located in the open groove
of the protein model.

Fig. 7. MDS analyses of the FaEXPA1-mu-

tant interaction with cellulose polymer. A)
A general view of the interaction of the corre-
sponding FaEXPA1-mutant protein and cellu-
lose as the ligand. B) The number of hydrogen
bonds established during MDS between the li-
gand and different amino acid residues located
in the open groove of each FaEXPA1-mutant
model.

84
F. Valenzuela-Riffo et al.
The disruption of hydrogen bonds within the cell wall structure is
the most probable mechanism of action of expansin proteins (Cosgrove,
2000). A large number of polar (charged or uncharged) residues were
found in D1, which indicates that they could interact with the glycan
chain, providing evidence to support that D1 has a similar structure as
the catalytic core of endoglucanase V (EGV) from Humicola insolens, and
in our case, this domain has a disrupted functionality. In this sense,
Mateluna et al. (2017) showed that differences in the binding energy
interaction obtained among six different pine expansin proteins can be
explained by changes in some residues that generate differences in the
electrostatic surface at the open groove region that contains both do-
mains, suggesting that these differences may help with the substrate
specificity of pine expansins.
Interestingly, we did not observe that the Asp100 residue (part of
the HFD motif) was an important residue for the establishment of H-
bonds with the cellulose chain, forming a H-bond during 34.5% and
38% of the trajectories of both MD simulations. In contrast, the
VpEXPA2 protein showed that Asp104 forms a H-bond with cellulose as
a ligand, suggesting that it is crucial in protein-ligand interactions
(Yennawar et al., 2006). Similar results were found in BsEXLX1 in
which the replacement of Asp82 by alanine induces complete in-
activation of the expansin-like activity (Silveira and Skaf, 2016).
However, other residues were found to be important for protein-ligand
interactions, and three residues were shown to form H-bonds for more
than 50% of the time-course of the MD simulation (Fig. 6a, left panel).
A closer view of the interactions between the cellulose polymer and
FaEXPA1 reveals that Asn81, Phe114 and Asn211 residues establish H-
bonds with the twisted glycan chain (Fig. 6a, right panel). The H-
bonding frequency between FaEXPA1 and the cellulose chain indicates
that the Asn81, Phe114 and Asn211 residues play the most important
roles in the FaEXPA1-cellulose complex and are more persistent in H-
bond formation compared to the other residues shown in Fig. 5a.
Recently, a short cellulose ligand (8 glycan monomers) and two
different short XGs (XXXGXXXG and XXFGXXFG structures) were used
to describe the interaction mode between this ligand and VpEXPA2 (the
first α-expansin protein model described) (Gaete-Eastman et al., 2015).
The authors showed that VpEXPA2 displays better binding energy for
cellulosic substrates than for xyloglucans, and the principal interaction
occurs between D1 and the ligands (Gaete-Eastman et al., 2015). The
small size of the ligands could explain the absence of an interaction
with D2. However, a recent study showed that a functional domain (D2)
from FaEXPA2 (another expansin isoform from strawberry previously
described) not only has affinity for cellulose but also for pectin and
xylan, although with significantly less affinity for the last two compared
to cellulose (Nardi et al., 2013). The authors suggested that expansins
could bind not only cellulose but also a wide range of cell wall polymers
(Nardi et al., 2013).
3.3. Analysis of different FaEXPA1-mutants
Using the FaEXPA1 structural model, in silico site-directed muta-
genesis experiments were carried out to understand the role of the re-
sidues present in the open groove that could interact with the cellulose
as a ligand (Fig. 7). Three different mutants were generated and eval-
uated in silico (Fig. 7).
Firstly, the 3D protein models of FaEXPA1 and FaEXPA1-mutants
were superimposed, and the proteins displayed similar 3D structures
with D2 and D1, with the same arrangement and disposition. Their root
mean square deviation (RMSD) values, obtained from the superposition
of the proteins, showed a value of 1.18 Å between the native and the
three mutant structures.
The protein-ligand conformations were studied by MD simulation
with cellulose as the ligand in each mutant protein (Fig. 7). All
FaEXPA1-mutant protein structures were able to interact with the li-
gands; however, the MD simulation showed differences for FaEXPA1-
Asn81Ala compared to the orientation of the cellulose with respect to
85
Computational Biology and Chemistry 76 (2018) 79–86
the wild-type protein and the other two mutant models because the
orientation of the ligand was out of the open groove and far from the
HFD motif, probably in an unspecific pocket (Fig. 7A). For FaEXPA1-
Phe114Ala, the ligand was oriented near the open groove but distant
from the HFD motif (around 7 Å) (Fig. 7A) and approached only for
short periods during the simulation, and thus, this interaction cannot be
stable. The mutant protein forms a small number of H-bonds with cel-
lulose as the ligand with respect to the native protein, and therefore,
any interaction between the protein and the cellulose was less possible
(Fig. 7B). Finally, FaEXPA1-Asn211Ala showed a similar protein-ligand
interaction for FaEXPA1-Phe114Ala, with the principal changes being
that the native protein was the distance of the ligand from the open
groove and the HFD motif (Fig. 7A).
From this study, it can be concluded that residues Asn81, Phe114
and Asn211 are important for maintaining the protein-ligand interac-
tion. The three mutations could avoid the protein-ligand interaction of
FaEXPA1 with cellulose, an issue that needs to be experimentally
proven in future work.
4. Conclusion
In summary, the FaEXPA1 protein model contains two domains
consistent with the structure of other expansins. The protein structure
has an open groove covering both domains that allows the interaction
of the ligand with the protein. MD simulations are consistent in pre-
dicting that the cellulose chain can interact first and quickly with D2
and after that with the D1 domain of the FaEXP1 protein. This evidence
strongly suggests that the specificity and the ligand recognition of ex-
pansin proteins with their ligands could reside in the open groove
structure, which mainly consists of Asn81 and Phe114 (in the D1 do-
main) and the Asn214 residue (in the D2 domain) involved in protein-
ligand stability. In silico site-directed mutagenesis experiments high-
lighted that these key residues (Asn81, Phe114 and Asn211) were in-
volved in maintenance of the protein-ligand interaction. Nevertheless,
this issue should be addressed in future works.
Funding
This work was supported by FONDECYT [grant number 11150543].
The funders had no role in the study design, data collection and ana-
lysis, decision to publish, or preparation of the manuscript.
Competing financial interests
The authors declare no competing financial interests.
Acknowledgements
F.V-R acknowledges Universidad de Talca for a doctoral scholar-
ship. P.R. acknowledges ‘Núcleo Científico Multidisciplinario’ from
Universidad de Talca. We acknowledge the helpful comments and
suggestions made by the three anonymous reviewers of this manuscript.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at https://doi.org/10.1016/j.compbiolchem.2018.05.
018.
References
Šali, A., Blundell, T.L., 1993. Comparative protein modeling by satisfaction of spatial
restraints. J. Mol. Biol. 234, 779–815.
Bennett, A., 2002. Biochemical and genetic of cell wall disassembly in ripening fruit: a
general model. HortScience 37, 447–450.
Bustamante, C., Civello, P.M., Martìnez, G.A., 2009. Cloning of the promoter region of b-
xylosidase (FaXyl1) gene and effect of plant growth regulators on the expression of
F. Valenzuela-Riffo et al.
FaXyl1 in strawberry fruit. Plant Sci. 177, 49–56.
Cho, H.T., Cosgrove, D.J., 2000. Altered expression of expansin modulates leaf growth
and pedicel abscission in Arabidopsis thaliana. Proc. Natl. Acad. Sci. U. S. A. 97,
9783–9788.
Choi, D., Lee, Y., Cho, H.T., Kende, H., 2003. Regulation of expansin gene expression
affects growth and development in transgenic rice plants. Plant Cell 15, 1386–1398.
Chundawat, S.P.S., Beckman, G.T., Himmel, M.E., Dale, B.E., 2011. Deconstruction of
lignocellulosic biomass to fuels and chemicals. Annu. Rev. Chem. Biomol. Eng. 2,
121–145.
Cosgrove, D.J., Bendinger, P., Durachko, D., 1997. Group I allergen of grass pollen as cell
wall-loosening agents. Proc. Natl. Acad. Sci. U. S. A. 94, 6559–6564.
Cosgrove, D.J., 2000. Expansive growth of plant cell walls. Plant Physiol. Biochem. 38,
109–124.
Cosgrove, D.J., 2005. Growth of the plant cell wall. Nat. Rev. Mol. Cell Biol. 6, 850–861.
Dotto, M.C., Martínez, G.A., Civello, P.M., 2006. Expression of expansin genes in straw-
berry varieties with contrasting fruit firmness. Plant Physiol. Biochem. 44, 301–307.
Gaete-Eastman, C., Morales-Quintana, L., Herrera, R., Moya-León, M.A., 2015. In-silico
analysis of structure and binding site feature of expansin protein from mountain
papaya fruit (VpEXPA2), through molecular modelling, docking and dynamics si-
mulation studies. J. Mol. Model. 21. http://dx.doi.org/10.1007/s00894-015-2656-7.
Georgelis, N., Tabuchi, A., Nikolaidis, N., Cosgrove, D.J., 2011. Structure-function ana-
lysis of the bacterial expansin EXLX1. J. Biol. Chem. 286, 16814–16823.
Georgelis, N., Yennawar, N.H., Cosgrove, D.J., 2012. Structural basis for entropy-driven
cellulose binding by a type–a cellulose-binding module (CBM) and bacterial expansin.
Proc. Natl. Acad. Sci. U. S. A. 109, 14830–14835.
Gilbert, H.J., Knox, J.P., Boraston, A.B., 2013. Advances in understanding the molecular
basis of plant cell wall polysaccharide recognition by carbohydrate-binding modules.
Curr. Opin. Struct. Biol. 23, 669–677.
Goulao, L.F., Oliveira, C.M., 2008. Cell wall modification during fruit ripening: when a
fruit is not the fruit. Trends Food Sci. Technol. 19, 4–25.
Gray-Mitsumune, M., Blomqvist, K., McQueen-Mason, S., Teeri, T.T., Sundberg, B.,
Mellerowicz, E.J., 2008. Ectopic expression of a wood-abundant expansin PttEXPA1
promotes cell expansion in primary and secondary tissues in aspen. Plant Biotechnol.
J. 6, 62–72.
Humphrey, W., Dalke, A., Schulten, K., 1996. VMD: visual molecular dynamics. J. Mol.
Graph. 14, 33–38.
Jervis, E.J., Haynes, C.A., Kilburn, D.G., 1997. Surface diffusion of cellulases and their
isolated binding domains on cellulose. J. Biol. Chem. 272, 24016–24023.
Johansson, P., Brumer, H., Baumann, M.J., Kallas, A.M., Henriksson, H., Denman, S.E.,
Teeri, T.T., Jones, T.A., 2004. Crystal structures of a poplar xyloglucan en-
dotransglycosylase reveal details of transglycosylation acceptor binding. Plant Cell
16, 874–886.
Jorgensen, W.L., Maxwell, D.S., Tirado-Rives, J., 1996. Development and testing of the
OPLS all-atom force field on conformational energetics and properties of organic li-
quids. J. Am. Chem. Soc. 118, 11225–11236.
Kerff, F., Amoroso, A., Herman, R., Sauvage, E., Petrella, S., Filée, P., Charlier, P., Joris,
B., Tabuchi, A., Nikolaidis, N., Cosgrove, D.J., 2008. Crystal structure and activity of
Bacillus subtilis YoaJ (EXLX1), a bacterial expansin that promotes root colonization.
Proc. Natl. Acad. Sci. U. S. A. 105, 16876–16881.
Laskowski, R.A., MacArthur, M.W., Moss, D.S., Thornton, J.M., 1993. PROCHECK: a
program to check the stereochemical quality of protein structures. J. Appl. Cryst. 26,
283–291.
Li, Y., Jones, L., McQueen-Mason, S., 2003. Expansins and cell growth. Curr. Opin. Plant
Biol. 6, 603–610.
Li, N., Pu, Y., Gong, Y., Yu, Y., Ding, H., 2016. Genomic location and expression analysis
of expansin gene family reveals the evolutionary and functional significance in
Triticum aestivum. Genes Genomics 38, 1021–1030.
Marga, F., Grandbois, M., Cosgrove, D.J., Baskin, T.I., 2005. Cell wall extension results in
the coordinate separation of parallel microfibrils: evidence from scanning electron
microscopy and atomic force microscopy. Plant J. 43, 181–190.
Martinez-Anaya, C., 2016. Understanding the structure and function of bacterial ex-
pansins: a prerequisite towards practical applications for the bioenergy and agri-
cultural industries. Microb. Biotechnol. 727–736.
86
Computational Biology and Chemistry 76 (2018) 79–86
Mateluna, P., Valenzuela-Riffo, F., Morales-Quintana, L., Herrera, R., Ramos, P., 2017.
Transcriptional and computational study of expansins differentially expressed in the
response to inclination in radiata pine. Plant Physiol. Biochem. 115, 12–24.
McQueen-Mason, S., Cosgrove, D.J., 1994. Disruption of hydrogen bonding between plant
cell wall polymers by proteins that induce wall extension. Proc. Natl. Acad. Sci. U. S.
A. 91, 6574–6578.
McQueen-Mason, S., Durachko, D.M., Cosgrove, D.J., 1992. Two endogenous proteins
that induce cell wall expansion in plants. Plant Cell 4, 1425–1433.
Mcqueen-Mason, S., Cosgrove, D.J., 1995. Expansin mode of action on cell walls. Analysis
of wall hydrolysis, stress relaxation, and binding. Plant Physiol. 107, 87–100.
Mercado, J.A., Pliego-Alfaro, F., Quesada, M.A., 2011. Fruit shelf life and potential for its
genetic improvement. In: Jenks, M.A., Bebeli, P.J. (Eds.), Breeding for Fruit Quality.
John Wiley & Sons, Oxford, pp. 81–104.
Morales-Quintana, L., Fuentes, L., Gaete-Eastman, C., Herrera, R., Moya-León, M.A.,
2011. Structural characterization and substrate specificity of VpAAT1 protein related
to ester biosynthesis in mountain papaya fruit. J. Mol. Graph. Model. 29, 635–642.
Morales-Quintana, L., Nuñez-Tobar, M.X., Moya-León, M.A., Herrera, R., 2013. Molecular
dynamics simulation and site-directed mutagenesis of alcohol acyltransferase: a
proposed mechanism of catalysis. J. Chem. Inf. Model. 53, 2689–2700.
Morrison, K.L., Weiss, G.A., 2001. Combinational alanine-scanning. Curr. Opin. Chem.
Biol. 5, 4321–4326.
Nardi, C., Escudero, C., Villarreal, N., Martínez, G., Civello, P.M., 2013. The carbohy-
drate-binding module of Fragaria × ananassa expansin 2 (CBM-FaExp2) binds to cell
wall polysaccharides and decreases cell wall enzyme activities “in vitro”. J. Plant Res.
126, 151–159.
Nardi, C., Villarreal, N., Rossi, F., Martínez, S., Martínez, G., Civello, P.M., 2015.
Overexpression of the carbohydrate-binding module of strawberry expansin2 in
Arabidopsis thaliana modifies plant growth and cell wall metabolism. Plant Mol. Biol.
88, 101–117.
Pilling, E., Höfte, H., 2003. Feedback from the wall. Curr. Opin. Plant Biol. 6, 611–616.
Quesada, M.A., Blanco-Portales, R., Pose, R.S., Gracía-Gago, J.A., Jiménez-Bermúdez, S.,
Muñóz-Serrano, A., Caballero, J.L., Pilego-Alfaro, F., Mercado, J.A., Muñóz-Blanco,
J., 2009. Antisense down-regulation of the FaPG1 gene reveals an unexpected central
role for polygalacturonase in strawberry fruit softening. Plant Physiol. 150,
1022–1032.
Rosli, H.G., Civello, P.M., Martínez, G.A., 2004. Changes in cell wall composition of three
Fragaria × ananassa cultivars with different softening rate during ripening. Plant
Physiol. Biochem. 42, 823–831.
Sampedro, J., Cosgrove, D.J., 2005. The expansin superfamily. Genome Biol. 6, 242.
Silveira, R.L., Skaf, M.S., 2016. Molecular dynamics of the Bacillus subtilis expansin
EXLX1: interaction with substrate and structural basis of the lack of activity of mu-
tants. Phys. Chem. Chem. Phys. 18, 3510–3521.
Sippl, M.J., 1993. Recognition of errors in three-dimensional structures of proteins.
Proteins 17, 355–362.
Vaaje-Kolstad, G., Farkaš, V., Hrmova, M., Fincher, G.B., 2010. Xyloglucan xyloglucosyl
transferases from barley (Hordeum vulgare L.) bind oligomeric and polymeric xy-
loglucan molecules in their acceptor binding sites. Biochim. Biophys. Acta 1800,
674–684.
Vicente, A.R., Saladié, M., Rose, J.K.C., Labavitch, J.M., 2007. The linkage between cell
wall metabolism and fruit softening: looking to the future. J. Sci. Food Agric. 87,
1435–1448.
Wang, C.X., Wang, L., McQueen-Mason, S.J., Pritchard, J., Thomas, C.R., 2008. pH and
expansin action on single suspension-cultured tomato (Lycopersicon esculentum)
cells. J. Plant Res. 121, 527–534.
Whitney, S.E.C., Gidley, M.J., McQueen-Mason, S.J., 2000. Probing expansin action using
cellulose/hemicellulose composites. Plant J. 22, 327–334.
Wiederstein, M., Sippl, M.J., 2007. ProSA-web: interactive web service for the recognition
of errors in three-dimensional structures of proteins. Nucleic Acids Res. 35,
W407–W410.
Yennawar, N.H., Li, L.C., Dudzinski, D.M., Tabuchi, A., Cosgrove, D.J., 2006. Crystal
structure and activities of EXPB1 (Zea m1), a β-expansin and group-1 pollen allergen
from maize. Proc. Natl. Acad. Sci. U. S. A. 103, 14664–14671.