You are on page 1of 16

Computational Biology and Chemistry 96 (2022) 107620

Contents lists available at ScienceDirect

Computational Biology and Chemistry


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

In silico analysis and characterization of medicinal mushroom cystathionine


beta-synthase as an angiotensin converting enzyme (ACE)
inhibitory protein
Neng-Yao Goh a, Muhammad Fazril Mohamad Razif a, Yeannie Hui-Yeng Yap b, Chyan
Leong Ng c, *, Shin-Yee Fung a, *
a
Medicinal Mushroom Research Group (MMRG), Department of Molecular Medicine, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia
b
Department of Oral Biology and Biomedical Sciences, MAHSA University, Selangor, Malaysia
c
Institute of Systems Biology (INBIOSIS), Universiti Kebangsaan Malaysia, 43600 UKM Bangi, Selangor, Malaysia

A R T I C L E I N F O A B S T R A C T

Keywords: Angiotensin-converting enzyme (ACE) regulates blood pressure and has been implicated in several conditions
Cystathionine beta-synthase including lung injury, fibrosis and Alzheimer’s disease. Medicinal mushroom Ganordema lucidum (Reishi) cys­
Angiotensin-converting enzyme tathionine beta-synthase (GlCBS) was previously reported to possess ACE inhibitory activities. However, the
Ganordema lucidum
inhibitory mechanism of CBS protein remains unreported. Therefore, this study integrates in silico sequencing,
Tiger milk mushroom
structural and functional based-analysis, protein modelling, molecular docking and binding affinity calculation
Lignosus rhinocerus
Protein-protein docking to elucidate the inhibitory mechanism of GlCBS and Lignosus rhinocerus (Tiger milk mushroom) CBS protein
Chloride blocking (LrCBS) towards ACE. In silico analysis indicates that CBSs from both mushrooms share high similarities in terms
Protein inhibitor of physical properties, structural properties and domain distribution. Protein-protein docking analysis revealed
that both GlCBS and LrCBS potentially modulate the C-terminal domain of ACE (C-ACE) activity via regulation of
chloride activation and/or prevention of substrate entry. GICBS and LrCBS were also shown to interact with ACE
at the same region that presumably inhibits the function of ACE.

1. Introduction presence of chloride ions (Wei et al., 1991). The crystal structure of tACE
shows the existence of two chloride ions embedded in each of the
Angiotensin-converting enzyme (ACE), also known as peptidyl- chloride binding pockets. The chloride ion furthest away from the zinc
dipeptidase A, CD143, kininase II or EC 3.4.15.1, is part of the renin- ion is known to interact with residues Arg186, Trp485, Arg489 and
angiotensin-aldosterone system (RAAS) (Skeggs, 1993; Ruiz-Dueñas water molecules at the CL1 binding pocket (CL1C-domain). The second
et al., 2020). Two distinct isoforms, namely somatic ACE (sACE) and chloride ion, closer to the zinc ion, was shown to interact with Arg522
testicular ACE (tACE) are found in human (Erdös, 1990). sACE is and Tyr224 at a region called the CL2 binding pocket (CL2C-domain) (Guy
composed of 1306 residues; it contains N-terminal (N-ACE) and C-ter­ et al., 2003). Nonetheless, the residues in CL1C-domain was found to not
minal (C-ACE) domains that share similar folding pattern, each con­ be fully conserved in the N-terminal-domain ACE (N-ACE) (Corradi
taining a functional catalytic site that contributes to the ability of ACE to et al., 2006). Hence, only one chloride ion is present in the N-ACE
hydrolyze a variety of peptides such as bradykinin, angiotensin I (AngI), chloride 2 binding pocket (CL2N-domain) coordinated by Arg500 and
substance P, amyloid peptides, tetrapeptide N-acetyl-Ser-Asp-Lys-Pro, Tyr202. C-ACE requires significantly higher chloride concentration for
neurotensin and enkephalins (Bernstein et al., 2011; Cozier et al., optimal catalytic activity compared to N-ACE (Wei et al., 1991),
2020b). tACE found in testis, is a 701-residue protein identical to the attributed to the importance of chloride ion in the stabilization of the
C-terminal domain of sACE. The high-resolution crystal structure of ACE active site (Rushworth et al., 2009; Yates et al., 2014).
tACE is widely used in the drug discovery of C-ACE specific inhibitors Chemically synthesized ACE inhibitors i.e. (captopril and trando­
(Acharya et al., 2003). lapril) have been successfully applied clinically to reduce mortality in
ACE is a substrate- and zinc-dependent peptidase, activated by the patients with high blood pressure and heart failure (De Leo et al., 2009).

* Corresponding authors.
E-mail addresses: clng@ukm.edu.my (C.L. Ng), syfung@um.edu.my (S.-Y. Fung).

https://doi.org/10.1016/j.compbiolchem.2021.107620
Received 16 August 2021; Received in revised form 20 December 2021; Accepted 20 December 2021
Available online 23 December 2021
1476-9271/© 2021 Elsevier Ltd. All rights reserved.
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

Other than blood pressure control, the regulation of ACE and its peptide proteins in the proteome of G. lucidum, one of it being cystathionine
substrates by ACE inhibitors show promising results against conditions beta-synthase (CBS) (Mohamad Ansor et al., 2013). Within the same
such as proteinuria, lung injury, Alzheimer’s disease, oncological dis­ fungi family, tiger milk mushroom or Lignosus rhinocerus is similarly
eases and sepsis (Pacurari et al., 2014; Bernstein et al., 2012; Yn Liu endorsed for its medicinal properties and is commonly consumed raw or
et al., 2019; S. Liu et al., 2019; Regulska et al., 2019; Kutyrina et al., drank as a decoction for the treatment of cough, asthma and fever by
1994). Recently, some ACE inhibitors are predicted with potential to be indigenous communities (Abdullah et al., 2013). L. rhinocerus was pre­
repurpose as an angiotensin-converting enzyme 2 (ACE2) inhibitor viously shown to exhibit immunomodulation, anti-inflammatory,
against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) anti-proliferative, antioxidative and anti-glycation properties (Lee
infection (Choi et al., 2020); attributed to the homology between ACE et al., 2018; Johnathan et al., 2016; Lai et al., 2014; Pushparajah et al.,
and ACE2 with 42% sequence identity and 75% similarity (Lubbe et al., 2016; Yap et al., 2018). Of interest, it was also annotated in both L.
2020). rhinocerus genomic and transcriptomic data the presence of CBS-like
Medicinal mushrooms (MM) are macroscopic fungi that are used to proteins (Yap et al., 2015b; Yap et al., 2014; Yap et al., 2015b).
alleviate and/or prevent diseases (Abdel-Azeem et al., 2019). MM have CBS (EC 4.2.1.22) is a key enzyme in the reverse transsulfuration
been studied for their naturally occurring ACE inhibitors due to the pathway that catalyzes the formation of cystathionine from homocys­
incidence of side effects, such as cough and skin rashes, caused by teine and serine along with the biogenesis of hydrogen sulfide (H2S)
synthetic drugs, (Choi et al., 2001; Hyoung Lee et al., 2004; Gao et al., (Chen et al., 2004). Depending on the organism, CBS adopts different
2012). Reishi mushroom or Ganordema lucidum is a widely recognized magnitudes of domain distribution and oligomeric variation (Fig. 1).
medicinal basidiomycete used in traditional remedy throughout China Based on previous studies, five types of domain distribution (termed
and Southeast Asia (Liu et al., 2015; Wachtel-Galor et al., 2011). Class 1–5) and 3 major oligomers have been identified. Class 1, con­
Mohamad Ansor et al. (2013) reported the presence of ACE inhibitory sisting of one single type II family pyridoxal-5′ -phosphate

Fig. 1. Structural features of various CBS. (A) Diagrammatical representation of catalytic domains predicted by SMART for GlCBS (position 11–329) and LrCBS
(position 11–330). (B) Graphical representation of CBS domain architectures and oligomeric information in different organisms. Class 1, a single catalytic domain
found in Bacillus anthracis (BaCBS), Lactiplantibacillus plantarum (LpCBS), Helicobacter pylori (HpCBS), L. rhinocerus (LrCBs) and G. Lucidum (GlCBS); Class 2, two
tandem catalytic domains reported in C. elegans (CeCBS); Class 3, the combination of a catalytic domain and Bateman module reported in S. cerevisiae (ScCBS) and
T. gondii (TgCBS). Class 4, modulation of 3 domains of heme, catalytic and Bateman Module, reported in Apis mellifera (AmCBS). Class 5, reported in human, contains
the Bateman module with AdoMet binding site. The degree of oligomerization is represented by the number of circles. Single asterisk represents solved crystal
structure. Double asterisk represents predicted structure through modelling protocol.

2
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

(PLP)-dependent enzyme, has a catalytic domain that functions as a PLP Expresso multiple sequences (Armougom et al., 2006) on the T-COFFEE
co-factor binding site via the presence of a Schiff base aided by a Server (http://tcoffee.crg.cat/) (Notredame et al., 2000) and were dis­
conserved lysine (Evande et al., 2004). Class 1 CBS exhibits played via ESPript 3.0 (https://espript.ibcp.fr/ESPript/ESPript/) (Rob­
self-assembly properties into an active dimer and are usually found in ert and Gouet, 2014). The domain architecture was investigated by
bacteria (Matoba et al., 2017). Class 2 CBS are monomer in nature, SMART genomic mode (http://smart.embl-heidelberg.de/) (Letunic
found in the nematode Caenorhabditis elegans (CeCBS) and consists of et al., 2020).
two tandem catalytic domains with only one PLP binding site (Vozdek
et al., 2012). 2.4. Secondary structure analysis
Class 3 CBS are characterized by the presence of a catalytic domain
followed by a C-terminal Bateman module consisting of two consecutive PSIPRED 4.0 (Jones, 1999) and DISOPRED3 (Jones and Cozzetto,
cystathionine β-synthase motifs (CBS1, CBS2) that are associated with its 2015) in PSIPRED server (http://bioinf.cs.ucl.ac.uk/psipred/) (Buchan
regulatory function and tetramer formation (Tu et al., 2018; and Jones, 2019) were used for secondary structure and disorder region
Oliveriusová et al., 2002). Class 4, typically found in insects such as prediction (Suppl. Fig. S1).
honeybee and fruit fly (Giménez et al., 2017; Su et al., 2013), is a CBS
dimer comprised of three separate domains with a middle conserved 2.5. Protein modelling of CBS proteins
catalytic domain flanked by an N-terminal heme-binding domain and a
C-terminal Bateman module. Finally, Class 5, the most complex of them The protein structure of both CBS proteins remains unresolved.
all, are found in mammals (Su et al., 2013). It is a tetrameric CBS with Therefore, composite protein modelling approaches using (1) contact
one functioning heme-binding domain, a catalytic domain and a Bate­ assisted protein structure prediction (C-I-TASSER) (https://zhanglab.
man module. ccmb.med.umich.edu/C-I-TASSER/) (Zhang et al., 2018); (2)
Current studies are greatly inclined towards protein hydrolysate or transform-restrained Rosetta (TrRosetta) and (3) Comparative model­
short inhibitory peptides (Boschin et al., 2014). Therefore, general un­ ling based Rosetta (RosettaCM) (https://robetta.bakerlab.org/) (Kim
derstanding on the inhibitory mechanism of large bulky proteins, such et al., 2004) were employed to obtain the structural model of GlCBS and
as CBS, towards ACE is lacking. This is because large bulky proteins were LrCBS proteins. Default modelling protocol was employed in
previously speculated to be weak ACE inhibitors due to the presence of C-I-TASSER and Robetta pipeline and the model with the highest C-score
ACE α1 (position 40–71), α2 (position 74–107), and α3 (position and confident score were selected, respectively (Suppl. Fig. S2). Roset­
109–120) helices which hinder the entry of bulky polypeptides to the taCM supports oligomeric modelling where GlCBS and LrCBS structural
substrate-binding site (Fang et al., 2019). Nonetheless, identification of models were obtained as a dimer while C-I-TASSER and TrRobetta
significant ACE inhibition by several mushroom ACE inhibitory proteins outputs the model as a single subunit. The selected models are provided
with high molecular weight (>30 kDa) prompts the need to further in (Suppl. Fig. S3) and were further evaluated in Sections 2.6 and 2.7 to
understand the inhibitory mechanism of those proteins (Mohamad obtain GlCBS and LrCBS models with the highest reliability. Upon
Ansor et al., 2013; Ibadallah et al., 2015; Geng et al., 2015). release of AlphaFold2 algorithm (Thornton et al., 2021), ab initio
Our study demonstrates the in silico identification, characterization, structural prediction of both GlCBS and LrCBS were carried out using
structural modelling of GlCBS and LrCBS. The molecular docking anal­ AlphaFold2 at ColabFold server (https://colab.research.google.com/
ysis of GlCBS, LrCBS and ACE are employed to elucidate the potential github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb).
interaction and inhibitory mechanism of CBSs towards ACE. The iden­
tification of CBS-interacting residues can lead to the discovery of a new 2.6. Model structural analysis and superimposition
druggable site on ACE for the design of peptide binding epitopes derived
from CBS protein for ACE inhibition. The generated models underwent DALI PDB search (http://ekhidna
2.biocenter.helsinki.fi/dali/), to detect structurally similar models.
2. Material and methods Structural comparisons were performed via superimposition using
PyMol (Schrödinger, www.pymol.org) to compare structural conserva­
2.1. Genomic data mining of potential ACE inhibitory CBS-like protein in tion. LrCBS and GlCBS models generated as a single subunit, i.e., C-I-
Lignosus rhinocerus TASSER and TrRosetta, were further evaluated to detect potential
clashing of dimeric interface prior to further dimerization of the models
The protein sequences of GlCBS (NCBI ID: AUN37950.1) was (Suppl. Fig. S4). Structural comparison of RosettaCM and Colabfold
retrieved from the NCBI database (https://www.ncbi.nlm.nih.gov/ models were also conducted (Suppl. Fig. S5).
protein/). The GlCBS sequence was used as the query for sequence
similarity (“homology”) search with L. rhinocerus genomic- 2.7. Model stereochemical evaluation
transcriptomic in-house database (Yap et al., 2015a) using locally
installed blast-2.4.0 + software (BLASTP cut-off e-value ≤1e-4) to The structure models of LrCBSRosettaCM, GlCBSRosettaCM and GlCBSC-I-
TASSER
identify LrCBS (Suppl. Table S1). were subjected for validation using three independent servers:
SAVES, SWISS-Model and ProSa to aid final model selections. Swiss-
2.2. Determination of physical parameters Model structural assessment tool was employed (https://swissmodel.
expasy.org/assess) to obtain QMEAN Z-score and Ramachandran plots
The physiochemical characteristics of GlCBS and LrCBS enzymes using MolProbity version 4.4 (Suppl. Fig. S6 & S7), SAVES 6.0 server
were analyzed via the ExPASy-ProtParam tool (http://web.expasy.org/ (https://saves.mbi.ucla.edu/) for ERRAT (Suppl. Fig. S8) (Colovos and
protparam) to explore its characteristics such as molecular weight, Yeates, 1993) and Verify3D (Suppl. Fig. S9) (Lüthy et al., 1992) analysis
theoretical pI value, instability index, aliphatic index, and grand average and ProSa (Suppl. Fig. S10 & S11) (Wiederstein and Sippl, 2007)
of hydrophobicity (GRAVY). (https://prosa.services.came.sbg.ac.at/prosa.php).

2.3. Sequence alignment and domain 2.8. Model Refinement and dimerization model construction of GlCBS
and LrCBS
GlCBS and LrCBS sequences were subjected to BLASTP search under
non-redundant database. Five sequences with the highest sequence Dimerized model of GlCBSC-I-TASSER was generated using Galax­
identity to the CBS protein from NCBI protein blast were aligned with yHomomer (Baek et al., 2017) (Suppl. Fig. S12). LrCBSRosettaCM and

3
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

GlCBSRosettaCM models were refined with GalaxyRefineComplex (Heo Table 1


et al., 2016) with the option of symmetric refinement for homo-oligomer Protparam physicochemical characterization of LrCBS and GlCBS.
(http://galaxy.seoklab.org/refinecomplex). The results were provided Parameter LrCBS GlCBS
as 10 refined models with five lowest-energy models from two different
Sequence length (A.A) 399 396
refinement protocols: (1) distance restraints; (2) both distance and po­ MW (Da) 42049.94 42141.79
sition restraints (Heo et al., 2016). GlCBSRosettaCM and GlCBSC-I-TASSER Theoretical pI: 5.81 5.81
model quality were further evaluated to choose the best model for Total number of negatively charged residues (Asp + 46 49
GlCBS. LrCBS and GlCBSRosettaCM models with the most improved ste­ Glu)
Total number of positively charged residues (Arg + Lys) 41 44
reochemical quality were selected as shown in Table 3. The dimer The instability index (II) 36.87 37.82
interface of the final selected structural model of GlCBS and LrCBSRo­ Aliphatic index 97.54 92.63
settaCM
were further investigated and validated by PISA (Krissinel and Grand average of hydropathicity (GRAVY) -0.084 -0.230
Henrick, 2007) to examine the buried interface area between CBS pro­ Note: Protein with a stability index less than 40 is stable; Positive aliphatic index
tomers, the solvation free-energy gain (ΔiG), P-value of ΔiG (statistical indicates the increase of thermostability of globular proteins and is defined as
probability to validate the protein interfaces as biological importance), the relative volume occupied by aliphatic side chains (Ala, Val, Ile, and Leu).
amount of hydrogen bonds and salt bridges. Negative GRAVY indicates hydrophilic proteins.

2.9. Protein-protein docking and interaction evaluation 3.3. Sequence alignment, domain, disordered region and secondary
structure analysis
The full crystal structure of human somatic angiotensin-converting
enzyme (sACE) was not available. Therefore, human testicular ACE, LrCBS shares high sequence identity (%) with mushroom CBS i.e.,
representing the C-domain of sACE (C-ACE, PDB Id: 1O8A and N domain G. lucidum (82.20), Dichomitus squalens (86.75), Polyporus brumalis
of sACE (N-ACE, PDB Id: 2C6F) crystal structure of protein in native state (85.96), Polyporus arcularius HHB13444 (85.96), Lentinus tigrinus
were obtained separately from the Research Collaboratory for Structural (85.21), Trametes coccinea (85.60) (Fig. 3). Both LrCBS and GlCBS se­
Bioinformatics (RCSB) Protein Data Bank (http://www.rcsb.org) for quences were similarly predicted to contain 12 extended strands and 19
subsequent docking study. Before docking analysis, polar hydrogen was α helices (Suppl. Fig. S1) with the presence of intrinsically disordered
added along with the removal of irrelevant water and ligand molecules regions (IDRs) at position (1− 3) and (366− 399) for LrCBS as well as
using Chimera 1.15 (Pettersen et al., 2004) software (University of position (1–4, 389 and 396) for GlCBS (Suppl. Fig. S1).
California, San Francisco, CA, USA), whereas the zinc and chloride ions
were retained. The GlCBSRosettaCM and LrCBSRosettaCM protein models 3.3.1. LrCBS and GlCBS are class I CBS
were docked individually onto the C- and N-ACE receptors using (1) SMART domain search revealed that both LrCBS and GlCBS were
Cluspro 2.0 (Suppl. Fig. S13) (Kozakov et al., 2017) (https://cluspro.org similarly annotated with the presence of a single type II family pyri­
/home.php), (2) HADDOCK 2.4 (Suppl. Table S2) (van Zundert et al., doxal-5′ -phosphate (PLP)-dependent catalytic domain, which contains
2016) (https://wenmr.science.uu.nl/HADDOCK2.4/) and (3) Hdock the PLP-cofactor binding sites (Fig. 1). The C-terminal Bateman module
(Suppl. Fig. S14) (Yan et al., 2020b) under default setting. The required present in the C-terminal of Toxoplasma gondii (TgCBS) (position
active and passive residues by HADDOCK pipeline are predicted by 366–419) was non-existent in both mushroom CBSs (Fig. 3). Previous
CPORT (Suppl. Table S3), which incorporate six different interface tools studies speculated that the ACE inhibitory activity of GlCBS was
into a consensus predictor (de Vries and Bonvin, 2011). The binding potentially derived from the zinc-binding capability demonstrated by
poses of the top docking complexes were selected and compared. The the bateman module on CBS C-terminal domain by interfering with the
interacting residues of CBS-ACE were evaluated by Ligprot+ Dimplot zinc ion active centre of ACE (Mohamad Ansor et al., 2013). Unlike other
function (Laskowski and Swindells, 2011) to generate schematic dia­ CBS proteins in higher-order organisms with greater complexity i.e.,
grams of protein-protein interactions (Suppl. Fig. S16). Potential containing N-terminal heme binding and/or C-terminal Bateman mod­
C-domain specific inhibition of CBS was studied using molecular blind ule, simplicity of the domains can be observed in both LrCBS and GlCBS
docking of substrate Hip-His-Leu (HHL) (.sdf file obtained from Pub­ along with other closely similar mushrooms CBSs (Fig. 1). Hence, ACE
Chem) into the selected CBS-C-ACE complexes with CB-Dock server, a inhibitory mechanism of GlCBS is not Bateman module dependent.
cavity guided protein-ligand docking server coupled with Autodock Based on sequence and domain analysis, the native structure of both
Vina docking algorithm(Yn Liu et al., 2019; S. Liu et al., 2019). The mushroom CBS showed greater similarity with that of Class 1 CBS and
binding affinity and dissociation constant of the plausible solution were may exist predominantly in the dimeric form. The potential ACE
predicted with contact-based method, PRODIGY (PROtein binDIng en­ inhibitory mechanisms of GlCBS and LrCBS proteins are investigated in
erGY prediction) using non-interface surface intermolecular contacts of Section 3.6.
the CBS-ACE and C-ACE-HHL interactions (Xue et al., 2016). Also, the lysine residue (position 50) that makes a Schiff base with
PLP (pyridoxal 5’-phosphate) (indicated by an asterisk in Fig. 3), crucial
3. Results and discussion for the catalytic activity of CBS are conserved in both L. rhinocerus and
G. lucidum CBS. The important conserved catalytic residue Lys50 sug­
3.1. Genomic data mining of potential ACE inhibitory CBS-likened protein gests that both LrCBS and GlCBS proteins are likely functionally active as
in Lignosus rhinocerus a key regulator in the reverse transsulfuration pathway. This is in
agreement with a study reported by Tian et al. (2019) whereby inhibi­
Blastp homology search of in-house L. rhinocerus genomic- tion of GlCBS resulted in the loss of ~60% of endogenous H2S biosyn­
transcriptomic database revealed a protein sequence (Seq. ID thesis. GlCBS transcription was reported to influence H2S biogenesis
GME6470_g) that shares 82% identity and 89% similarity with GlCBS which regulates the synthesis of bioactive secondary metabolites such as
protein, term LrCBS (Suppl. Table S1). ganoderic acid. The conserved nature of LrCBS and GlCBS domain ar­
chitecture and catalytic residue infers the possibility of similar H2S
regulation in secondary metabolite biosynthesis present in L. rhinocerus.
3.2. Determination of physical parameters

Protparam analysis (Table 1) showed that LrCBS and GlCBS possess


similar physical parameters and amino acid constituents (Fig. 2).

4
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

Fig. 2. Amino acid composition profile of LrCBS and GlCBS.

3.4. Structural and stereochemical quality-based protein model selection et al., 1992). QMEAN Z-score or QMEAN4 is a linear combination of
statistical potential terms, and a value less than four indicates a reliable
Homology modelling servers HHpred and Swiss-modelling generated model. ProSa evaluates experimental structures by comparing its
incomplete protein models due to the lack of template similarity at the knowledge-based score to random structures with the same sequence
IDR, which is known to be highly variable (Not shown). Therefore, (Wiederstein and Sippl, 2007).
protein modelling servers C-I-TASSER, TrRosetta and RosettaCM with Both LrCBSRosettaCM and GlCBSRosettaCM model were unable to satisfy
inclusion of ab initio protocol were employed to obtain the full protein the Verify-3D (%) cutoff point of 80%, with only 79.95% and 79.29% of
structure for GlCBS and LrCBS. DaliPDB search showed the obtained the amino acids scoring > 0.2 in the 3D/1D profile, respectively
LrCBS models bear significant structural resemblance with TgCBS while (Table 3). This indicates poor compatibility of the amino acid sequences
GlCBS models share significant similarity with either ScCBS or TgCBS, with its 3D model. Hence, further refinement was performed with Gal­
all with significant Z-scores > 50 (Z-score >2 represents significance) axyRefineComplex. Refinement of LrCBSRosettaCM model improved the
(Table 2). Molprobity score of 2.98–1.58, RC favoured (%) of 96.47–98.24%
Unlike RosettaCM, both C-I-TASSER and TrRosetta pipeline do not (>98%) (Elsliger and Wilson, 2012), clashscore of 195.74–11.76 with a
support oligomeric modelling. Further oligomerization modelling of the tolerable RC outlier (%) of 0.63%. QMEAN value of 0.19 (close to zero)
protein models were required to be performed via GalaxyHomomer. indicates a model of high quality. Additionally, Verify-3D analysis with
Nonetheless, additional structural superimposition of dimeric models an improvement from 79.95% to 81.70% of the amino acids scoring
for GlCBSTrRosetta, LrCBSTrRosetta and LrCBSC-I-TASSER with TgCBS (PDB > 0.2 in the 3D/1D profile and an ERRAT score of 95.24 (greater than 5)
ID: 6xyl) revealed clashing dimerization interface at the C-terminal re­ indicates a good quality model. Besides, ProSA Z-score of − 10.01 shows
gion with IDR characteristics (Suppl. Fig. S4). Hence, only the GlCBSC-I- the Z-value of the protein was similar to native protein of equivalent
TASSER
model without any notable structural clashing in the dimerization sizes (Suppl. Fig. S10). The reliability of the model was also shown by
interface during structural comparison was selected for further dimer­ the ProSA energy plot with no obvious problematic regions with a
ization and refinement. Structural comparison revealed two protein positive value in the ProSA energy plot (Suppl. Fig. S11).
model candidates for GlCBS i.e. GlCBSC-I-TASSER and GlCBSRosettaCM; The two GlCBS models, GlCBSC-I-TASSER and GlCBSRobettaCM, were
Conversely, only one for LrCBS that is LrCBSRosettaCM. further evaluated. GlCBSRobettaCM refined model revealed significant
The inaccuracy of a protein 3D model can arise from incorrect improvement in Molprobity score of 3.01–1.58, RC Favored (%)
template alignment or prediction mistakes. Those inaccuracies can be 96.07–98.24%, clash score from 189.32 to 11.76 with acceptable RC
detected based on high conformational energies or quality assessment outliers of 0.63(%). QMEAN Z-score of 0.70 to − 0.40 also indicate the
scores (Bhattacharya et al., 2007). GlCBSC-I-TASSER, GlCBSRosettaCM and closeness of the model with experimental structures (Feig, 2016).
LrCBSRosettaCM models were evaluated with five structure validation Moreover, model quality improvement was also observed in the
tools including Molprobity, QMEAN, ProSa, Verify-3D and ERRAT using Verify-3D analysis, from 79.29% to 81.57% of the amino acids scoring
3 independent servers (Table 3). Further comparisons of model qualities > 0.2 in the 3D/1D profile. ProSA Z-score improvements from − 9.01 to
were carried out between the dimerized GlCBSC-I-TASSER model and − 9.06 indicated that its energy value was similar to those found for
GlCBSRosettaCM for the final GlCBS model selection. SWISS-Model struc­ native protein of equivalent sizes (Suppl. Fig. S10) with no obvious
ture assessment server employs MolProbity version 4.4 to generate a problematic regions with positive value in the ProSA energy plot (Suppl.
Ramachandran (RC) plot and clashscore analysis. MolProbity is an Fig. S11). Although there is a decrease in the ERRAT value, from 94.57
all-atom measurement of structural correctness based on steric clashes, to 92.65, the value is still higher than the generally accepted range of
rotamer outliers, and Ramachandran-plot while Molprobity score is a > 50 for a high-quality model (Laskowski et al., 1993).
log-weighted combination of those functions that reflect crystallography GlCBSC-I-TASSER dimerized model was evaluated with Molprobity
resolution (Davis et al., 2004). ERRAT is mainly used to verify protein score (2.05), RC favoured (94.16%), RC outlier (1.27%), clash score
via the study of non-bonded atom-atom interactions (Colovos and (14.13), QMEAN Z-score (− 1.98), ERRAT score (87.61), Verify-3D score
Yeates, 1993). On the other hand, Verify-3D functions to determine the (82.20%) and ProSA Z-score (− 8.62), all of which signifies an acceptable
compatibility of a 3D atomic model with its amino acid sequence (Lüthy model with good quality. Although GlCBSC-I-TASSER and GlCBSRobettaCM

5
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

Fig. 3. Comparison of amino acid sequences of LrCBS and GlCBS with other CBS proteins. Multiple sequence alignments were visualized by ESPript 3.0 (Robert and
Gouet, 2014) subsequent to alignment of protein sequences using Expresso T-Coffee (Notredame et al., 2000). Secondary structure features of RosettaCM generated
LrCBS model and DALI homolog, Toxoplasma gondii CBS, (TgCBS, PDB ID: 6XYL) are shown above and below the alignments, respectively. α-helices and β-strands are
represented as helices and arrows, respectively. β-turns are marked with TT while η symbol refers to a 310-helix. Fully conserved areas are shaded in red. Conserved
sequences of greater than 70% similarity are indicated by a box. Dichomitus squalens (TBU30616.1), Polyporus brumalis (RDX54216.1), Polyporus arcularius HHB13444
(TFK86983.1), Lentinus tigrinus ALCF2SS1–6 (RPD64447.1), Trametes coccinea (OSD05344.1), Ganoderma lucidum (AUN37950.1) and Toxoplasma gondii (6xyl).
Asterisk depicts the conserved Lysine residue responsible for PLP cofactor binding. Single red dot represents conserved amino acid residues involved in hydrogen
bonding or salt bridging by TgCBS dimer interface, respectively. Black triangle represents recurring interacting residues in both C-ACE-GlCBS and C-ACE-LrCBS
interactions predicted by Cluspro and HADDOCK. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of
this article.)

Table 2
DALI structural homolog with highest Z-score with the predicted CBS models.
Protein type Modelling pipelines Oligomeric state Z-score PDB ID & chain Protein name %ID Species

LrCBS C-I-TASSER – 52.8 6xyl-A Cystathionine beta-synthase 53 Toxoplasma gondii


TRrosetta – 51.6 6xyl-A Cystathionine beta-synthase 53 Toxoplasma gondii
Rosetta CM Dimer 50.2 6xyl-A Cystathionine beta-synthase 53 Toxoplasma gondii
GlCBS C-I-TASSER – 54.0 6c2h-A Cystathionine beta-synthase 49 Saccharomyces cerevisiae
TrRosetta – 51.8 6xyl-A Cystathionine beta-synthase 54 Toxoplasma gondii
Rosetta CM Dimer 50.2 6xyl-A Cystathionine beta-synthase 54 Toxoplasma gondii

Note: %ID = Percentage identity.

models were observed to possess similar structural resemblance, ho­ (Pereira et al., 2021)., ab initio model of GlCBS and LrCBS were
mologous to the TgCBS structure, nonetheless, the GlCBSRobettaCM model generated using Colabfold server. Structural comparison between the
analyzed had better stereochemical qualities compared to GlCBSC-I- RosettaCM and AlphaFold2 models showed high similarity at the core
TASSER
across Molprobity, QMEAN, ERRAT and ProSa scores. Quality structure of CBS with low RMSD value (<1.34 Å) (Suppl. Fig S5). The
evaluation of the GlCBS models revealed that the GlCBSRosettaCM model AlphaFold2 predicted models showed low structural confidence at the
was superior overall, with better stereochemical, and geometrical CBS C-terminal (IDR), shown in the predicted aligned error (PAE) plot
qualities compared to the GlCBSC-I-TASSER model. Hence, the GlCBSRo­ and is possibly the reason for the observed conformation differences
settaCM
and LrCBSRosettaCM were selected for the following docking study when compared with RosettaCM models.
(Fig. 4A). DALI server and all of the quality assessment analysis from SAVES,
Upon the release of the recent Alphafold2 modelling algorithm Swiss-Model structural assessment and ProSa servers revealed that both

6
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

Table 3
Stereochemical validation of pre/post refinement CBS proteins models.
Protein Modelling server SWISS-Model structure assessment server SAVES ProSA Z-
score
MolProbity RC Favored (%) RC Outliers (%) Clash QMEAN ERRAT Verify- 3D (%)
Score score

LrCBS RosettaCM 2.98 96.47 0.25 195.74 0.77 97.03 79.95 -9.86
RosettaCM (refined) 1.58 98.24 0.63 11.76 0.19 95.24 81.70 -10.01
GlCBS C-I-TASSER (dimerized and 2.05 94.16 1.27 14.13 -1.98 87.61 82.20 -8.62
refined)
RosettaCM 3.01 96.07 0.38 189.32 0.70 94.57 79.29 -9.01
RosettaCM (refined) 1.58 98.24 0.63 11.76 -0.40 92.65 81.57 -9.06

Note = RC: Ramachandran.

Fig. 4. 3D structural model. (A) Two protomers


of LrCBS (left) and GlCBS (right) obtained from
RosettaCM. Chain A and B are coloured in blue
and cyan, respectively. N- and C-terminals are
represented with symbol N and C, respectively.
Structural superimposition of (B) LrCBSRo­
settaCM
-TgCBS; (C) GlCBSRosettaCM-TgCBS; (D)
LrCBSRosettaCM-GlCBSRosettaCM Chain. Cyan
arrow indicates the longer helices (position
174–184) observed in LrCBSRosettaCM model.
Black circle represents the region with notable
structural divergence between the LrCBS and
GlCBSRosettaCM model. Chain A and B of TgCBS
are coloured in pink and red, respectively. (E)
Interdimeric residues in TgCBS (Red & pink,
chain A and B) showing the formation of non-
covalent bonding between Asp15-Arg28 and
Asp15-Arg31, represented with stick models.
Equivalent residues in LrCBSRosettaCM and
GlCBSRosettaCM are shown and coloured as blue
and orange respectively. (For interpretation of
the references to colour in this figure legend,
the reader is referred to the web version of this
article.)

LrCBSRosettaCM and GlCBSRosettaCM model refined by GalaxyR­


efineComplex protocol were more appropriate for protein modelling of
dimeric CBS structures.

7
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

3.5. Structural comparison and dimer interface analysis low-energy docked structures. Model cluster size was emphasized
instead of docking energy for the selection of probable cluster solution
3.5.1. LrCBS and GlCBS dimer interface analysis and structural (Kozakov et al., 2017). Similarly, Hdock will output 10 docking solu­
comparison with TgCBS tions as the most plausible binding positions based on its docking scoring
Structural comparisons between LrCBSRosettaCM and GlCBSRosettaCM and algorithm. These are applicable for docking studies where the
model revealed that the catalytic domain structure was highly conserved reference template possesses a sequence identity of greater than 30%
in comparison to TgCBS (Fig. 4B & 4C). One major insertion at position (Yan et al., 2020b). Besides, HADDOCK docking complexes are indicated
174–184 of LrCBSRosettaCM was observed, contributing to the formation by its Z-score which provides an indication of the amount of standard
of longer helices (Fig. 2). The structural variants were observed to not deviation from the average cluster. Cluster with a positive Z-score are
impose any structural obstructions that might affect the dimeric inter­ usually unfavourable and not selected for further investigation (van
face and oligomeric state of the protein (Fig. 4B). In addition, there are Zundert et al., 2016).
some minor differences due to small insertions or deletions in the
mushroom CBS relative to TgCBS. Nonetheless, secondary elements 3.6.1. Docking studies uncover two potential binding site on C-ACE
were predominantly unaltered concerning the equivalent region in the Our preliminary docking study revealed a possible interaction be­
TgCBS (Fig. 4B & 4C). tween GlCBS with entrance of CL1/CL2 binding pocket (CL1/2 C-domain)
The inter-dimeric formation of TgCBS at the catalytic core was of C-ACE (not shown). Therefore, a subsequent docking study between
supported by the formation of a hydrogen bond and salt bridge between C-ACE and CBSs was conducted with removal of the chloride ion in both
Asp15-Arg28, Asp15-Arg31 (Fig. 4E). Similar conservation of those CL1/2 C-domain to affirm the ability of the CBS protein to block and
residues in LrCBS and GlCBS suggests that the interdimeric interactions prevent entrance of Cl- ion into the C-ACE. All of the selected binding
in the dimer interface of LrCBS and GlCBS were retained (indicated by positions (Fig. 5), binding energies and affinities (Table 5) and inter-
red dot in Fig. 3). However, further structural comparisons revealed the molecular interacting residues (Suppl. Table S4) are provided. Cluspro
presence of Asp9-Arg22 interaction in the equivalent position of LrCBS server unveiled that GlCBS interacts with C-ACE at CL1C-domain with
and GlCBS, but Asp9-Arg25 interaction was not observed (Fig. 4E). binding affinity (BA) and dissociation constant (Kd) of − 14.2 kcal/ mol
Nonetheless, the oligomeric state of the predicted LrCBSRosettaCM and and 3.7E-11 mol/L, respectively. Comparable binding positions for
GlCBSRosettaCM models was further validated via PISA dimer interface LrCBS was also observed at cluster 6 with BA and Kd of − 13.7 kcal/mol
analysis (Table 4) where dimer formation for LrCBS and GlCBS models and 8.9E-11 mol/L, respectively.
are favoured, gaining solvation energy during interface formation with Intriguingly, similar binding interaction with the CL1C-domain (Fig. 5A
Δi G values of − 39.4 and − 41.7 kcal/mol, along with interface area of & 5B) can be observed in the HADDOCK top clusters for both GlCBS (BA:
3324.4 and 3811.7 Å2, respectively. A total of 40 and 44 hydrogen − 12.2 kcal/mol; Kd: 1.1E-0.9 mol/L) and LrCBS (BA: 13.1 kcal/mol; Kd:
bonds as well as 4 and 12 salt bridges, formed during the dimeric 2.6E-10 kcal/mol), where stronger binding affinity was observed for
interface formation of LrCBS and GlCBS, respectively. Moreover, the HADDOCK C-ACE-LrCBS complexes compared to GlCBS. Although the
obtained P-value > 0.5 also indicates that the interface was interaction- detailed mechanism is still unknown, it was postulated that conforma­
specific and biologically important. PISA analysis supports the notion of tion changes near the surface of CL1 pocket assist the acceptance of
homodimer formation between CBS protomers in LrCBS and GlCBS, chloride ion into the pocket (Fig. 6) and both CBS proteins are found to
respectively. similarly interact and conceal the supposed entrance of CL1C-domain (Guy
et al., 2003; Rushworth et al., 2009).
3.5.2. Structural comparison of LrCBS with GlCBS HADDOCK complexes of LrCBS and GlCBS with CL1C-domain pre­
Further structural comparisons between the LrCBSRosettaCM and dicted several conserved inter-molecular interactions with the formation
GlCBSRosettaCM models indicated some variations can be observed in the of hydrogen bonding between C-ACE Lys174 and Asp235 in both CBSs.
coiled regions as well as the IDR in the C-terminal region, but the pre­ There were also similar electrostatic interactions between Glu176 and
dicted secondary elements in the main catalytic domain were mostly the Ile230 (GlCBS) and Ile229 (LrCBS), Gln262 with Gly243, Pro245
same (Fig. 4D). The high structural resemblance of the main catalytic (GlCBS) and Gly242, Pro244 (LrCBS) (Fig. 7A) found in both CBSs. The
domain implies great possible conservation of protein functionalities partake of recurrent residues Lys174, Glu176, Gln262, His263 in C-ACE
and interactions. with the CBS proteins via Cluspro and HADDOCK signifies the impor­
tance of those residues in the inhibitory mechanism of CBS proteins.
Top Hdock docking solution for C-ACE-GlCBS showed a different
3.6. Protein-protein docking and interactions between CBS proteins and
binding location compared to both Cluspro and HADDOCK (Fig. 5A),
ACE
where binding occurs at the supposed entrance of the CL2C-domain with
BA − 17.3 kcal/mol and Kd 2.0E-13 mol/L. Nonetheless, similar LrCBS-
Protein-protein docking algorithms can achieve successful predictive
CACE docking interaction (Fig. 5B) was observed at solution 3 with BA
rates in many complexes, as shown in the Critical Assessment of Pre­
− 12.9 kcal/mol and Kd 3.5E-10 mol/L. Both CBS proteins were
dicted Interactions (CAPRI). Protein docking servers that possess great
revealed to similarly interact with CL2C-domain surrounding residues of
reliability, based on recent CAPRI community assessment experiments,
Gln231, Asp232, Glu234, Arg235, Gln238, Gln239, Pro242, Gln579,
are Cluspro 2.0, HADDOCK 2.4 and Hdock (Lensink et al., 2020). All
Pro585, Asn586, Ser590, Leu593, Arg604 (Fig. 7B). Similar hydrogen
three were used to elucidate the ACE inhibitory mechanism of GlCBS
bonding formation between C-ACE and both CBS protein were observed
and LrCBS with the C- and N-domain of ACE. The outputs of Cluspro
in Asp 232-Tyr166. The conserved negatively charged Glu174 in CBS
involved ten models defined by centres of highly populated clusters of
protein could also be an important residue in the PPI between C-ACE
CL2 binding site and CBS, as it actively engages in salt bridge formation
Table 4 with a positively charged arginine on C-ACE of Arg604 and Arg235 in
PISA analysis of CBS protein dimeric interface.
LrCBS and GlCBS docking complexes, respectively. Two conserved res­
Dimer Interface area ΔiG (kcal/ ΔiG P- No. of H- No. of Salt idues, Tyr166 and Glu174, in CBS proteins could be the crucial residues
(Å2) mol) value bond bridge assisting in blocking the entrance of the CL2C-domain.
LrCBS 3324.4 -39.4 0.018 40 4 CL1C-domain showed greater possibility as the most plausible inter­
GlCBS 3811.7 -41.7 0.021 44 12 acting site compared to CL2C-domain with mutual agreement of binding
Note: Å2 = interface area; ΔiG = solvation free energy gains upon formation of location predicted by Cluspro and HADDOCK top clusters. However,
the interface. based on the binding affinities between the three docking algorithms,

8
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

Fig. 5. Protein-protein docking solutions predicted by Cluspro, HADDOCK and Hdock. (A) Cluspro and HADDOCK top clusters predicted GlCBS interaction with the
C-ACE CL1 chloride binding site (CL1C-domain); Hdock predicted GlCBS binding at the entrance of C-ACE CL2 chloride binding site (CL2C-domain). (B) Cluspro and
HADDOCK predicted similar mechanism of LrCBS interaction with the CL1C-domain; while Hdock predicted similar GlCBS binding at the entrance of CL2C-domain
binding site. (C) Cluspro, Haddock and Hdock top docking complexes all showed GlCBS binding to the entrance of N-ACE CL1 binding pocket (CL1N-domain). (D) Top
docking complexes of Cluspro, Haddock and Hdock revealed LrCBS binds similarly to CL1N-domain. Magenta-coloured atoms represent the supposed removed chloride
ions in C-ACE and are only shown to show the location of the binding site, the chloride atom in N-ACE is coloured in green. ACE are coloured in orange. Chain A and
chain B of CBS proteins are coloured in blue and cyan, respectively. The possible entry way of chloride 1 and 2 into its respective binding pocket in C-ACE (circled in
black and orange, respectively) could be blocked upon CBS binding. (For interpretation of the references to colour in this figure legend, the reader is referred to the
web version of this article.)

Hdock predicted C-ACE-GlCBS interacting interface has the highest Despite both Cluspro and Hdock top docking solution showing different
binding affinity at the CL2C-domain. The observed discrepancy possibly interacting sites, Arg144 on GlCBS was observed to actively partake in
arises due to the template-dependent nature of Hdock that sets itself hydrogen bond formation in both cases with Gln183 and Gln483 of C-
apart from the template-free docking algorithm of HADDOCK and Clu­ ACE (Fig. 7C), respectively. This indicates the potential importance of
spro. Hdock algorithm with the inclusion of biological information may Arg144 in C-ACE-GlCBS interactions.
be the source of these differently predicted docking outputs (Yan et al., The two predicted potential binding sites on C-ACE are potentially
2020b). linked with the attenuation of chloride activation activity in C-ACE. ACE
Nevertheless, two potential binding sites on C-ACE were uncovered was determined to be chloride-dependent and substrate-specific, espe­
by the C-ACE-GlCBS protein-protein interaction studies. (1) Cluspro and cially in the catalysis of Angiotensin I (Ehlers and Kirsch, 1988). Certain
HADDOCK predicted binding onto the surrounding residues of CL1C- substrates (eg. AngI) require the initial binding of a chloride ion to ACE
domain
; (2) Hdock proposed interaction with the entrance of CL2C-domain. prior to Angiotensin II (AngII) formation (Shapiro et al., 1983). Chloride

9
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

Table 5
Comparison of top docking cluster or solutions for GlCBS against LrCBS with similar binding conformation in C-ACE and N-ACE interaction.
Receptors Protein ligand Cluspro HADDOCK Hdock

C. no. LES BA Kd C. no. H. S. Z-score BA Kd S. no. D. S. BA Kd

C-ACE GlCBS 1 -895.4 -14.2 3.7E-11 11 25.3 + /- 6.0 -2.5 -12.2 1.1E-09 1 -286.69 -17.3 2.0E-13
LrCBS 6 -797.3 -13.7 8.9E-11 1 -21.7 + /- 5.9 -2.6 -13.1 2.6E-10 3 -221.69 -12.9 3.5E-10
N-ACE GlCBS 1 -856.9 -14.0 5.2E-11 1 45.2 + /- 9.9 -2.2 -15.7 3.0E-12 1 -291.56 -14.4 2.7E-11
LrCBS 1 -762.6 -12.6 5.9E-10 3 -4.3 + /- 12.6 -1.3 -13.4 1.5E-10 1 -259.15 -14.7 1.6E-11

Note: BA = Binding affinity/ΔG (kcal/mol), Kd = dissociation constant (mol/L) C. no. = Cluster number, D. S. = Docking score, H. S. = HADDOCK score LES= Lowest
energy score, S. no. = Solution number. The docking solution with the highest reliability in Cluspro, HADDOCK and Hdock are decided by cluster number/size
(Kozakov et al., 2017), Z-score (van Zundert et al., 2016) and docking score (Yan et al., 2020b). BA and Kd were calculated by PRODIGY.

Fig. 6. Important regions and residues in C terminal domain of ACE (C-ACE). (A) (tACE crystal structure, 1O8A, coloured in orange). Two chloride ions can be found
in the C-ACE (coloured in green). The binding site of GlCBS on entrance of CL1C-domain. The pocket with chloride ion bound which furthest away from the zinc ion
(coloured in grey) is termed CL1 binding pocket binding pocket. The chloride ion closest to zinc ion is termed CL2 binding pocket. Six flexible hinge regions involved
in hinge bending activity of C-ACE were coloured in red (Watermeyer et al., 2006), Glu403-Lys118 salt bridging that controls the movement of chloride ion into the
CL2 pocket (Masuyer et al., 2014) is represented as blue sticks. (B) Frontal view of the proposed entrance of CL1 binding site, showing the residues affecting chloride
sensitivity of C-ACE. (C) Overview of GlCBS-C-ACE interacting surface that may be involve in blocking the entrance of CL2 into C-ACE. (D) Entrance of C-ACE CL2
binding pocket is circled. Critical salt bridging residues that control the entry of chloride into CL2 binding pocket colored in blue. (For interpretation of the references
to colour in this figure legend, the reader is referred to the web version of this article.)

was studied to modulate the affinity of ACE inhibitors i.e. (enalaprilat, sites on C-ACE, thus, inhibiting C-ACE catalytic activity. Therefore,
lisinopril and captopril) towards ACE domains (Wei et al., 1992). A based on our docking analysis, it is likely that the inhibitory mechanism
mutagenesis study by Rushworth et al. (2009) showed that both of mushroom CBS proteins is dependent on the modulation of chloride
CL1C-domain and CL2C-domain modulate the chloride dependence of activation in C-ACE via prevention of chloride ion entry into either one
C-ACE, hence it is relevant to catalysis. It was also proposed by Tzakos of the CL1/CL2C-domain.
et al. (2003) that CL1 is a synergistic anion for high-affinity substrate The acceptance of substrate entry into ACE is characterized by mouth
binding. The absence of chloride ion in the CL1C-domain would result in opening and closing aided by a hinge-bending motion (Cozier et al.,
an unfavorable shift of residue Lys511, required for the stabilization of 2020b). ACE complex is known to undergo various open, semi-open and
the substrate C-terminal carboxylate in C-ACE. Likewise, the absence of closed conformations in the presence of AngII (Vy et al., 2020). An
chloride ion in the CL2C-domain was found to greatly affect catalytic site inhibitory mechanism was proposed for the mixed-non-competitive in­
stability via chloride coordination of Arg522 (Yates et al., 2014). The hibitor hexapeptide (Thr-Met-Glu-Pro-Gly-Lys-Pro), in which the pep­
interacting interface between C-ACE-CBS interface potentially results in tide is able to stabilize the closed and semi-open states of ACE while
the blockage of chloride ion entry into either one of the chloride binding preventing the open state, resulting in a dead-end complex that holds

10
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

Fig. 7. Consolidated protein-protein interaction studies of C-ACE residues engaging in non-covalent interactions with GlCBS/LrCBS. (A) Four surrounding residues
on C-ACE CL1 binding pocket (CL1C-domain) partake in interactions with several residues on GlCBS/LrCBS, predicted by Cluspro and Haddock, in the selected protein-
protein complexes. (B) Surrounding residues on C-ACE CL2 binding pocket (CL2C-domain) interacts with both GlCBS and LrCBS predicted by Hdock. Chain A and B on
CBS proteins are represented by (1) and (2), respectively. (C) Arg144 on GlCBS was found to be actively engaging in C-ACE-GlCBS intermolecular interactions among
the top complexes predicted by Cluspro and Hdock.

AngII in the active site. Watermeyer et al. (2006) reported that hinging interactions with C-ACE hinge bending residues could potentially
of C-ACE was closely related to the flexibility of residues at positions sequester the hinge bending motion and help to stabilize the
98–125, 296–297, 400–409, 434–439, 535–537 and 569–578 (Fig. 6A). Glu403-Lys118 interaction that blocks chloride ion entry into the
In line with this, the Hdock PPIs analysis of GlCBS against CL2C-domain CL2C-domain. This eventually reduces the chloride affinity of CL2C-domain
identified several key hinging residues (i.e., Arg100, Asn105, Gln106, (Yates et al., 2014).
Gln108, Pro575, Glu576 and Gln579) that potentially could restrain the Our docking analysis revealed that the inhibitory action of GlCBS
open-close transition of ACE and prevent substrate entry and indirectly and LrCBS most likely occur via (1) the regulation of chloride activation
(Suppl. Table S4). Disruption of the Glu403-Lys118 salt bridge was re­ by blocking either CL1C-domain or CL2C-domain, or (2) formation of CBS-
ported to promote entry of chloride ion into the CL2 pocket. Hence, ACE complex that prevents the entry of substrate into the active site

11
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

via an allosteric interaction. likely to result in an inhibitory effect on N-ACE catalytic activity.

3.6.2. Inhibition activity of LrCBS and GlCBS are likely C-domain specific 3.6.3. C-ACE inhibition of GlCBS are likely conserved in LrCBS and
Astoundingly, Cluspro, HADDOCK and Hdock top docking solutions potentially in a non-competitive manner
(Fig. 5C) unequivocally predicted GlCBS to interact with the surround­ LrCBS demonstrated binding at the entrance of CL1 and CL2 binding
ing residues of N-ACE CL1 binding pocket (CL1N-domain) with binding pocket of C-ACE with good resemblance against GlCBS top docking so­
energies − 14.0 kcal/mol, − 15.7 kcal/mol and − 14.4 kcal/mol and lution, in Cluspro and HADDOCK, and Hdock, respectively (Section
dissociation constant 5.2E-11 mol/L, 3.0E-12 mol/L and 2.7E-11 mol/L, 3.6.1). Several equivalently conserved residues of GlCBS engage in inter-
respectively (Table 5). Similarly, Cluspro, HADDOCK, and Hdock top molecular interactions with CL1C-domain such as Asp235-Lys174, Ile230-
docking solutions (Fig. 5D) showed LrCBS able to interact with Glu176 and Gly243 & Pro245-Gln262, respectively, shown in
convergent binding site of CL1N-domain with binding energy of HADDOCK and Cluspro top docking complexes, are found in LrCBS with
− 12.6 kcal/mol, − 13.4 kcal/mol and − 14.7 kcal/mol and dissocia­ similar interplay against CL1C-domain (i.e. Asp235-Lys174, Ile229-
tion constant of 5.9E-10 mol/L, 1.5E-10 mol/L and 1.6E-11 mol/L, Glu176, Gly242 & Pro244-Gln262, respectively). Interestingly, residues
respectively. Gly243 and Pro245 in GlCBS are not present in TgCBS and are only
Cluspro and HADDOCK N-ACE-CBS top docking solutions also present in some mushroom CBS proteins including LrCBS and Lentinus
revealed several conserved residues on the PPi of both LrCBS and GlCBS, tigrinus (LtCBS) (Fig. 3). Cluspro and HADDOCK revealed some over­
such as Thr4, Tyr166 and Glu174, that contribute to the stabilization of lapping residues present in both GlCBS and LrCBS that interact with
protein complexes, mostly via electrostatic interactions (Fig. 8A). The CL1C-domain; these are also only conserved in some mushroom CBSs.
overlap in convergent binding site across all top docking complexes The presence of overlapping convergent binding sites on PPI can
revealed that Leu603 and Asn606 on N-ACE appear to be critical resi­ potentially signify the intrinsic characteristics of protein-protein in­
dues that contribute to the interaction of LrCBS and GlCBS with N-ACE, teractions (DeLano, 2002; Cukuroglu et al., 2014). Hence, Lys174,
engaging in various non-covalent bonding interactions (Fig. 8B). Glu176 and Gln262 located at the surrounding entrance of the
All docking algorithms for GlCBS and LrCBS reported a similar CL1C-domain could potentially serve as a druggable site for selective in­
binding mechanism to the surrounding residues of CL1N-domain. None­ hibition of C-ACE due to its repetitive formation and intermolecular
theless, these interactions may not result in significant inhibition of N- interactions with both CBS proteins as predicted by Haddock. Stronger
ACE catalytic activity. Unlike C-ACE, only one chloride ion can be found binding affinity was observed for C-ACE-LrCBS compared to
in the crystal structure of N-ACE at the CL2N-domain, due to the residue C-ACE-GlCBS in HADDOCK top docking solutions. Hence, this suggests
difference between His164 of N-ACE and the equivalent Arg186 residues that LrCBS may exert similar or potentially stronger ACE inhibitory
of C-ACE, that restrict the entry of chlorine into the CL1N-domain (Natesh properties than reported for GlCBS.
et al., 2003). Therefore, the binding and blocking of CL1N-domain will There are three main active site pockets in an ACE molecule,
likely not affect chloride activation mechanism on N-ACE. Moreover, including the S1 pocket (i.e., Ala 354, Glu 384, and Tyr 523), S2′ pocket
N-ACE was reported to exhibit lower chloride requirement and can (i.e., Gln 281, His 353, Lys 511, His 513, and Tyr 520), and S1′ pocket
remain catalytically active even without the presence of chloride ions (Glu 162) (Suppl. Fig. S15A) (Ma et al., 2019). Further molecular
(Masuyer et al., 2014; Jaspard et al., 1993), possibly attributed to the docking simulation was employed to study the potential effect of
presence of only one chloride ion at the CL2N-domain (Fang et al., 2019). absence of chloride ion in the CL1/CL2 binding pocket of C-ACE. Mo­
Hence, any interactions towards the CL1N-domain of N-ACE-CBS are less lecular docking simulation revealed C-ACE substrate, Hip-His-Leu

Fig. 8. Graphical representation of important residues on ACE-CBs interactions. (A) Conserved residues on LrCBS and GlCBS are denoted in blue boxes with recurrent
interactions with surrounding residues of N-ACE CL1 binding site predicted by Cluspro and HADDOCK. (B) Conserved residues on N-ACE denoted in blue boxes
engaging in non-covalent interactions with GlCBS and LrCBS top docking complexes predicted by Cluspro, HADDOCK and Hdock. Chain A and B on CBS proteins are
represented by (1) and (2), respectively.

12
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

(HHL), established various interactions with the S1 (Ala 354, Glu 384, expression of LrCBS protein in Lignosus rhinocerus is required.
Tyr 523) and S2 pocket (Gln 281, His 353, Lys 511, His 513, Tyr 520) of Protein-protein interaction studies on mushroom CBS with two ACE
ACE when both chloride ions are present (Suppl. Fig. S17A). Loss of HHL domains have revealed preliminary clues on its probable ACE inhibitory
interaction with the S1 pocket residues were observed in Cluspro, mechanism, as demonstrated by GlCBS in vitro [Graphical summary of
Haddock and Hdock predicted GlCBS-C-ACE complexes (Suppl. all predicted mushroom CBS binding sites on ACE are provided in
Fig. S17B), with decreased binding affinity in Haddock and Hdock Fig. 9]. Several conserved residues in LrCBS and GlCBS were found to
(Suppl. Table S5). On the other hand, most interactions with the S1 and actively interact with residues in ACE, that could set up the foundation
S2 pockets are retained in HHL-LrCBS-C-ACE complexes predicted by for the future discovery of novel peptide binding epitope based-drug
Cluspro, Haddock and Hdock (Suppl. Fig. S17C), but with decreasing discovery. Nevertheless, the potential inhibitory activity of LrCBS re­
binding affinity compared with uncomplexed C-ACE (Suppl. Table S5). quires further validation in vitro, along with confirmation of the po­
This further demonstrates that CBS binding to C-ACE may inhibit C-ACE tential chlorine-activation and domain-specific inhibitory mechanisms
activity. However, this hypothesis requires further in vitro in­ in both CBS proteins.
vestigations to examine the hydrolysis ratio of Z-Phe-His-Leu (ZPHL)
and HHL, in order to determine its domain selective activity (Danilov 4. Conclusion
et al., 2008).
At present, C-ACE specific drugs such as phosphinic peptide Genomic data mining of L. rhinocerus revealed a CBS protein (LrCBS)
RXPA380 inhibits C-ACE competitively by occupying the S2’, S2, S1’, with prominent structural similarity (>80%) with ACE inhibitory pro­
and S1 subsites of the C-ACE substrate binding site (Corradi et al., 2007). tein GlCBS. Domain analysis revealed the conservation of the catalytic
Past studies comparing RXPA380 and N-domain specific RXP407 pep­ domain in both LrCBS and GlCBS with high structural resemblance.
tides also revealed that the domain selectivity of those peptides are Protein-protein docking analysis revealed an ACE inhibitory mecha­
mainly governed by interactions to the S2 subsite (Phe 391, Glu 403) nism, whereby both GlCBS and LrCBS potentially bind to the entrance of
(Kröger et al., 2009). Discovery of drugs with C-domain inhibitory the CL1/CL2 binding pocket, thereby moderating C-ACE activity. Resi­
specificity are crucial for blood pressure regulation due to its ability to dues Lys174, Glu176 and Gln262, and Asp232 on tACE (C-ACE) are
bypass unfavourable side effects such as cough and angioedema, which potentially important residues that stabilize the C-ACE-CBS interface on
arise from N-ACE inhibition (Burger et al., 2014). Based on CL1C-domain and CL2C-domain, respectively. All top binding solutions for
protein-protein docking prediction, C-domain selective inhibitory both N-ACE-CBS studies predicted by Cluspro, HADDOCK and Hdock
mechanism of GlCBS and LrCBS are different compared to RXPA380. revealed convergent binding at the CL1 binding pocket of N-ACE with
This is likely because (i) CBS was predicted to interact with potential recurring residues Leu603 and Asn606 on N-ACE. However, the
allosteric sites of C-ACE i.e., CL1/2 C-domain, but not substrate binding blockage of CL1N-domain is proposed to have minimum effect on the
site, and (ii) the bulky structure of CBS proteins may pose a challenge for catalytic activity of N-ACE as the binding site serves no purpose in
CBS to enter ACE substrate binding site, unlike the small peptide of chloride activation activity. Our results highlight the potential of G.
RXPA380. lucidum and L. rhinocerus as a natural source of ACE inhibitory proteins,
The high molecular weight (>30Kda) fraction of L. rhinocerus scle­ potentially in a C-domain specific manner. However, further in vitro
rotium cold water extract was previously shown to possess anti- investigations are required to affirm the presence of such ACE inhibitory
inflammatory and anticancer properties (H.-Y.Y. Yap et al., 2018; H.Y. activities in LrCBS and GlCBS. The identification of recurring interacting
Y. Yap et al., 2018; Lee et al., 2014). It was also reported to mediate the residues between ACE and CBS proteins can potentially be exploited for
activity of cell surface Gαq-coupled protein receptor (GPCR), contrib­ further studies on C domain-selective ACE inhibitors in order to develop
uting to airway relaxation. High molecular weight proteins, such as prospective therapeutics.
LrCBS, with excellent predicted solubility and stability (Table 1) are an
excellent candidate for future drug discovery due to its potential func­ CRediT authorship contribution statement
tion as an ACE inhibitor, alleviating conditions related to the over­
expression of the ACE/AngII axis. However, further studies on the actual Neng-Yao Goh: Conceptualization, Data curation, Formal analysis,

Fig. 9. Graphical representation of sACE as an


integral-membrane protein with two active sites
of N- and C- domain. C-domain contains two
chloride binding pockets of chloride 1 (CL1C-
domain
) and 2 (CL2C-domain), while N-domain
contains only 1 chloride ion at the CL2N-domain
binding pocket. Docking analysis revealed three
possible binding modes. I: Binding of CBS at
CL1N-domain binding. II: Binding of CagBS at
CL1C-domain binding pocket, leading to blockage
of CL1 entrance. III: Binding of CBS at CL2C-
domain
binding pocket, leading to blockage of
CL2 entrance. Prevention of either CL1/CL2 C-
domain
binding leads to destabilization of the
active site and prevents high-affinity substrate
binding to the S1, S2 and S1’ pocket. Chloride
ion is coloured in green. Zinc ion is represented
as a grey circle. (For interpretation of the ref­
erences to colour in this figure legend, the
reader is referred to the web version of this
article.)

13
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

Investigation, Methodology, Resources, Visualization, Validation, Colovos, C., Yeates, T.O., 1993. Verification of protein structures: patterns of nonbonded
atomic interactions. Protein Sci. 2 (9), 1511–1519.
Writing – original draft, Writing – review & editing. Shin-Yee Fung:
Corradi, H.R., Schwager, S.L.U., Nchinda, A.T., Sturrock, E.D., Acharya, K.R., 2006.
Conceptualization, Supervision, Resources, Project administration, Crystal structure of the N domain of human somatic angiotensin I-converting enzyme
Writing – review & editing. Muhammad F M Razif: Conceptualization, provides a structural basis for domain-specific inhibitor design. J. Mol. Biol. 357 (3),
Funding acquisition, Project administration, Resources, Supervision, 964–974. https://doi.org/10.1016/j.jmb.2006.01.048.
Corradi, H.R., Chitapi, I., Sewell, B.T., Georgiadis, D., Dive, V., Sturrock, E.D.,
Writing – review & editing. Chyan Leong Ng: Supervision, Resources, Acharya, K.R., 2007. The structure of testis angiotensin-converting enzyme in
Project administration, Writing – review & editing. Hui-Yeng Y Yap: complex with the C domain-specific inhibitor RXPA380. Biochemistry 46 (18),
Supervision, Resources, Project administration, Writing – review & 5473–5478.
Cozier, G.E., Lubbe, L., Sturrock, E.D., Acharya, K.R., 2020b. Angiotensin-converting
editing. enzyme open for business: structural insights into the subdomain dynamics. FEBS J.
n/a (n/a) https://doi.org/10.1111/febs.15601.
Cukuroglu, E., Engin, H.B., Gursoy, A., Keskin, O., 2014. Hot spots in protein–protein
interfaces: towards drug discovery. Prog. Biophys. Mol. Biol. 116 (2–3), 165–173.
Declaration of Competing Interest
Danilov, S.M., Balyasnikova, I.V., Albrecht 2nd, R.F., Kost, O.A., 2008. Simultaneous
determination of ACE activity with 2 substrates provides information on the status of
The authors declare that they have no known competing financial somatic ACE and allows detection of inhibitors in human blood. J. Cardiovasc
interests or personal relationships that could have appeared to influence Pharm. 52 (1), 90–103. https://doi.org/10.1097/FJC.0b013e31817fd3bc.
Davis, I.W., Murray, L.W., Richardson, J.S., Richardson, D.C., 2004. MOLPROBITY:
the work reported in this paper. structure validation and all-atom contact analysis for nucleic acids and their
complexes. Nucleic Acids Res 32 (suppl_2), W615–W619.
De Leo, F., Panarese, S., Gallerani, R., Ceci, L.R., 2009. Angiotensin converting enzyme
Acknowledgements (ACE) inhibitory peptides: production and implementation of functional food. Curr.
Pharm. Des. 15 (31), 3622–3643. https://doi.org/10.2174/138161209789271834.
This work was supported by Faculty Research Grant, University of DeLano, W.L., 2002. Unraveling hot spots in binding interfaces: progress and challenges.
Curr. Opin. Struct. Biol. 12 (1), 14–20. https://doi.org/10.1016/s0959-440x(02)
Malaya, Malaysia [Grant numbers: GPF003A-2020 and GPF003B-2020].
00283-x.
Ehlers, M.R., Kirsch, R.E., 1988. Catalysis of angiotensin I hydrolysis by human
Appendix A. Supporting information angiotensin-converting enzyme: effect of chloride and pH. Biochemistry 27 (15),
5538–5544.
Elsliger, M.A., Wilson, I.A., 2012. 1.8 Structure Validation and Analysis. In: Egelman, E.
Supplementary data associated with this article can be found in the H. (Ed.), Comprehensive Biophysics. Elsevier, Amsterdam, pp. 116–135.
online version at doi:10.1016/j.compbiolchem.2021.107620. Erdös, E.G., 1990. Angiotensin I converting enzyme and the changes in our concepts
through the years. Lewis K. Dahl memorial lecture. Hypertension 16 (4), 363–370.
https://doi.org/10.1161/01.HYP.16.4.363.
References Evande, R., Ojha, S., Banerjee, R., 2004. Visualization of PLP-bound intermediates in
hemeless variants of human cystathionine beta-synthase: evidence that lysine 119 is
a general base. Arch. Biochem. Biophys. 427 (2), 188–196. https://doi.org/10.1016/
Abdel-Azeem, A.M., Abdel-Azeem, M.A., Khalil, W.F., 2019. Chapter 21 - Endophytic
j.abb.2004.04.027.
Fungi as a New Source of Antirheumatoid Metabolites. In: Watson, R.R., Preedy, V.R.
Fang, L., Geng, M., Liu, C., Wang, J., Min, W., Liu, J., 2019. Structural and molecular
(Eds.), Bioactive Food as Dietary Interventions for Arthritis and Related
basis of angiotensin-converting enzyme by computational modeling: insights into the
Inflammatory Diseases (Second Edition). Academic Press, pp. 355–384.
mechanisms of different inhibitors. e0215609-e0215609 PloS One 14 (4). https://
Abdullah, N., Haimi, M.Z.D., Lau, B.F., Annuar, M.S.M., 2013. Domestication of a wild
doi.org/10.1371/journal.pone.0215609.
medicinal sclerotial mushroom, Lignosus rhinocerotis (Cooke) Ryvarden. Ind. Crops
Feig, M., 2016. Local protein structure refinement via molecular dynamics simulations
Prod. 47, 256–261. https://doi.org/10.1016/j.indcrop.2013.03.012.
with locPREFMD. J. Chem. Inf. Model. 56 (7), 1304–1312.
Acharya, K.R., Sturrock, E.D., Riordan, J.F., Ehlers, M.R.W., 2003. Ace revisited: a new
Gao, X.J., Yan, P.S., Wang, J.B., Yu, J.J., 2012. ACE inhibitory, antitumor and
target for structure-based drug design. Nat. Rev. Drug Disco 2 (11), 891–902.
antioxidant activities of submerged culture materials of three medicinal mushrooms.
https://doi.org/10.1038/nrd1227.
Appl. Mech. Mater. 145, 179–183. https://doi.org/10.4028/www.scientific.net/
Armougom, F., Moretti, S., Poirot, O., Audic, S., Dumas, P., Schaeli, B., Keduas, V.,
AMM.145.179.
Notredame, C., 2006. Expresso: automatic incorporation of structural information in
Geng, X., Tian, G., Zhang, W., Zhao, Y., Zhao, L., Ryu, M., Wang, H., Ng, T.B., 2015.
multiple sequence alignments using 3D-Coffee. Nucleic Acids Res. 34 (Web Server
Isolation of an angiotensin I-converting enzyme inhibitory protein with
issue), W604–W608. https://doi.org/10.1093/nar/gkl092.
antihypertensive effect in spontaneously hypertensive rats from the edible wild
Baek, M., Park, T., Heo, L., Park, C., Seok, C., 2017. GalaxyHomomer: a web server for
mushroom leucopaxillus tricolor. Molecules 20 (6), 10141–10153.
protein homo-oligomer structure prediction from a monomer sequence or structure.
Giménez, P., Majtan, T., Oyenarte, I., Ereño-Orbea, J., Majtan, J., Klaudiny, J., Kraus, J.,
Nucleic Acids Res 45 (W1), W320–W324. https://doi.org/10.1093/nar/gkx246.
Martínez-Cruz, L.A., 2017. Crystal structure of cystathionine β-synthase from
Bernstein, K.E., Shen, X.Z., Gonzalez-Villalobos, R.A., Billet, S., Okwan-Duodu, D.,
honeybee Apis mellifera. J. Struct. Biol. 202 https://doi.org/10.1016/j.
Ong, F.S., Fuchs, S., 2011. Different in vivo functions of the two catalytic domains of
jsb.2017.12.008.
angiotensin-converting enzyme (ACE). Curr. Opin. Pharm. 11 (2), 105–111. https://
Guy, J.L., Jackson, R.M., Acharya, K.R., Sturrock, E.D., Hooper, N.M., Turner, A.J., 2003.
doi.org/10.1016/j.coph.2010.11.001.
Angiotensin-converting enzyme-2 (ACE2): comparative modeling of the active site,
Bernstein, K.E., Ong, F.S., Blackwell, W.-L.B., Shah, K.H., Giani, J.F., Gonzalez-
specificity requirements, and chloride dependence. Biochemistry 42 (45),
Villalobos, R.A., Shen, X.Z., Fuchs, S., Touyz, R.M., 2012. A modern understanding
13185–13192. https://doi.org/10.1021/bi035268s.
of the traditional and nontraditional biological functions of angiotensin-converting
Heo, L., Lee, H., Seok, C., 2016. GalaxyRefineComplex: refinement of protein-protein
enzyme. Pharmacol. Rev. 65 (1), 1–46. https://doi.org/10.1124/pr.112.006809.
complex model structures driven by interface repacking. Sci. Rep. 6 (1), 32153.
Bhattacharya, A., Tejero, R., Montelione, G.T., 2007. Evaluating protein structures
https://doi.org/10.1038/srep32153.
determined by structural genomics consortia. Protein.: Struct., Funct., Bioinforma.
Hyoung Lee, D., Ho Kim, J., Sik Park, J., Jun Choi, Y., Soo Lee, J., 2004. Isolation and
66 (4), 778–795.
characterization of a novel angiotensin I-converting enzyme inhibitory peptide
Boschin, G., Scigliuolo, G.M., Resta, D., Arnoldi, A., 2014. ACE-inhibitory activity of
derived from the edible mushroom Tricholoma giganteum. Peptides 25 (4),
enzymatic protein hydrolysates from lupin and other legumes. Food Chem. 145,
621–627. https://doi.org/10.1016/j.peptides.2004.01.015.
34–40. https://doi.org/10.1016/j.foodchem.2013.07.076.
Ibadallah, B., Abdullah, N., Shuib, A., 2015. Identification of angiotensin-converting
Buchan, D.W.A., Jones, D.T., 2019. The PSIPRED protein analysis workbench: 20 years
enzyme inhibitory proteins from mycelium of pleurotus pulmonarius (Oyster
on. Nucleic Acids Res. 47 (W1), W402–W407. https://doi.org/10.1093/nar/gkz297.
Mushroom). Planta Med. 81 https://doi.org/10.1055/s-0034-1383409.
Burger, D., Reudelhuber, T.L., Mahajan, A., Chibale, K., Sturrock, E.D., Touyz, R.M.,
Jaspard, E., Wei, L., Alhenc-Gelas, F., 1993. Differences in the properties and enzymatic
2014. Effects of a domain-selective ACE inhibitor in a mouse model of chronic
specificities of the two active sites of angiotensin I-converting enzyme (kininase II).
angiotensin II-dependent hypertension. Clin. Sci. 127 (1), 57–63. https://doi.org/
Studies with bradykinin and other natural peptides. J. Biol. Chem. 268 (13),
10.1042/cs20130808.
9496–9503.
Chen, X., Jhee, K.H., Kruger, W.D., 2004. Production of the neuromodulator H2S by
Johnathan, M., Gan, S.H., Ezumi, M.F.W., Faezahtul, A.H., Nurul, A.A., 2016.
cystathionine beta-synthase via the condensation of cysteine and homocysteine.
Phytochemical profiles and inhibitory effects of Tiger Milk mushroom (Lignosus
J. Biol. Chem. 279 (50), 52082–52086. https://doi.org/10.1074/jbc.C400481200.
rhinocerus) extract on ovalbumin-induced airway inflammation in a rodent model of
Choi, H.S., Cho, H.Y., Yang, H.C., Ra, K.S., Suh, H.J., 2001. Angiotensin I-converting
asthma, 167-167 BMC Complement Alter. Med. 16. https://doi.org/10.1186/
enzyme inhibitor from Grifola frondosa. Food Res. Int. 34 (2), 177–182. https://doi.
s12906-016-1141-x.
org/10.1016/S0963-9969(00)00149-6.
Jones, D.T., 1999. Protein secondary structure prediction based on position-specific
Choi, Y., Shin, B., Kang, K., Park, S., Beck, B.R., 2020. Target-centered drug repurposing
scoring matrices. J. Mol. Biol. 292 (2), 195–202. https://doi.org/10.1006/
predictions of human angiotensin-converting enzyme 2 (ACE2) and transmembrane
jmbi.1999.3091.
protease serine subtype 2 (TMPRSS2) interacting approved drugs for coronavirus
disease 2019 (COVID-19) treatment through a drug-target interaction deep learning
model. Viruses 12 (11). https://doi.org/10.3390/v12111325.

14
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

Jones, D.T., Cozzetto, D., 2015. DISOPRED3: precise disordered region predictions with Pacurari, M., Kafoury, R., Tchounwou, P.B., Ndebele, K., 2014. The renin-angiotensin-
annotated protein-binding activity. Bioinformatics 31 (6), 857–863. https://doi.org/ aldosterone system in vascular inflammation and remodeling. Int. J. Inflamm. 2014,
10.1093/bioinformatics/btu744. 689360 https://doi.org/10.1155/2014/689360.
Kim, D.E., Chivian, D., Baker, D., 2004. Protein structure prediction and analysis using Pereira, J., Simpkin, A.J., Hartmann, M.D., Rigden, D.J., Keegan, R.M., Lupas, A.N.,
the Robetta server. Nucleic Acids Res 32 (Web Server issue), W526–W531. https:// 2021. High-accuracy protein structure prediction in CASP14. Protein.: Struct.,
doi.org/10.1093/nar/gkh468. Funct., Bioinforma. 89 (12), 1687–1699.
Kozakov, D., Hall, D.R., Xia, B., Porter, K.A., Padhorny, D., Yueh, C., Beglov, D., Pettersen, E.F., Goddard, T.D., Huang, C.C., Couch, G.S., Greenblatt, D.M., Meng, E.C.,
Vajda, S., 2017. The ClusPro web server for protein–protein docking. Nat. Protoc. 12 Ferrin, T.E., 2004. UCSF Chimera—A visualization system for exploratory research
(2), 255–278. https://doi.org/10.1038/nprot.2016.169. and analysis. J. Comput. Chem. 25 (13), 1605–1612. https://doi.org/10.1002/
Krissinel, E., Henrick, K., 2007. Inference of macromolecular assemblies from crystalline jcc.20084.
state. J. Mol. Biol. 372 (3), 774–797. Pushparajah, V., Fatima, A., Chong, C.H., Gambule, T.Z., Chan, C.J., Ng, S.T., Tan, C.S.,
Kröger, W.L., Douglas, R.G., O’Neill, H.G., Dive, V., Sturrock, E.D., 2009. Investigating Fung, S.Y., Lee, S.S., Tan, N.H., Lim, R.L.H., 2016. Characterisation of a new fungal
the domain specificity of phosphinic inhibitors RXPA380 and RXP407 in immunomodulatory protein from tiger milk mushroom, lignosus rhinocerotis,
angiotensin-converting enzyme. Biochemistry 48 (35), 8405–8412. https://doi.org/ 30010-30010 Sci. Rep. 6. https://doi.org/10.1038/srep30010.
10.1021/bi9011226. Regulska, K., Regulski, M., Karolak, B., Murias, M., Stanisz, B., 2019. Can cardiovascular
Kutyrina, I.M., Tareeva, I.E., Shestakova, M.V., Gerasimenko, O.I., Zverev, K.V., drugs support cancer treatment? The rationale for drug repurposing. Drug Discov.
Rogov, V.A., Sheremet’eva, O.V., 1994. [The antiproteinuric action of angiotensin- Today 24 (4), 1059–1065. https://doi.org/10.1016/j.drudis.2019.03.010.
converting enzyme inhibitors in chronic glomerulonephritis and diabetic Robert, X., Gouet, P., 2014. Deciphering key features in protein structures with the new
nephropathy]. Ter. Arkh 66 (6), 19–22. ENDscript server. Nucleic Acids Res. 42 (W1), W320–W324. https://doi.org/
Lai, W.H., Zainal, Z., Daud, F., 2014. Preliminary study on the potential of 10.1093/nar/gku316.
polysaccharide from indigenous Tiger’s Milk mushroom (Lignosus rhinocerus) as Ruiz-Dueñas, F.J., Barrasa, J.M., Sánchez-García, M., Camarero, S., Miyauchi, S.,
anti-lung cancer agent. AIP Conf. Proc. 1614 (1), 517–519. https://doi.org/10.1063/ Serrano, A., Linde, D., Babiker, R., Drula, E., Ayuso-Fernández, I., Pacheco, R.,
1.4895252. Padilla, G., Ferreira, P., Barriuso, J., Kellner, H., Castanera, R., Alfaro, M.,
Laskowski, R.A., Swindells, M.B., 2011. LigPlot+: multiple ligand-protein interaction Ramirez, L., Pisabarro, A.G., Riley, R., Kuo, A., Andreopoulos, W., LaButti, K.,
diagrams for drug discovery. J. Chem. Inf. Model 51 (10), 2778–2786. https://doi. Pangilinan, J., Tritt, A., Lipzen, A., He, G., Yan, M., Ng, V., Grigoriev, I.V., Cullen, D.,
org/10.1021/ci200227u. Martin, F., Rosso, M.N., Henrissat, B., Hibbett, D., Martínez, A.T., 2020. Genomic
Laskowski, R.A., MacArthur, M.W., Moss, D.S., Thornton, J.M., 1993. PROCHECK: a analysis enlightens agaricales lifestyle evolution and increasing peroxidase diversity.
program to check the stereochemical quality of protein structures. J. Appl. Mol. Biol. Evol. https://doi.org/10.1093/molbev/msaa301.
Crystallogr. 26 (2), 283–291. Rushworth, C., Guy, J., Turner, A., 2009. Residues affecting the chloride regulation and
Lee, M.K., Lim, K.H., Millns, P., Mohankumar, S.K., Ng, S.T., Tan, C.S., Then, S.M., substrate selectivity of the angiotensin-converting enzymes (ACE and ACE2)
Mbaki, Y., Ting, K.N., 2018. Bronchodilator effects of Lignosus rhinocerotis extract identified by site-directed mutagenesis. FEBS J. 275, 6033–6042. https://doi.org/
on rat isolated airways is linked to the blockage of calcium entry. Phytomedicine 42, 10.1111/j.1742-4658.2008.06733.x.
172–179. https://doi.org/10.1016/j.phymed.2018.03.025. Shapiro, R., Holmquist, B., Riordan, J.F., 1983. Anion activation of angiotensin
Lee, S.S., Tan, N.H., Fung, S.Y., Sim, S.M., Tan, C.S., Ng, S.T., 2014. Anti-inflammatory converting enzyme: dependence on nature of substrate. Biochemistry 22 (16),
effect of the sclerotium of Lignosus rhinocerotis (Cooke) Ryvarden, the Tiger Milk 3850–3857. https://doi.org/10.1021/bi00285a021.
mushroom, 359-359 BMC Complement Alter. Med. 14. https://doi.org/10.1186/ Skeggs Jr., L.T., 1993. Discovery of the two angiotensin peptides and the angiotensin
1472-6882-14-359. converting enzyme. Hypertension 21 (2), 259–260. https://doi.org/10.1161/01.
Lensink, M.F., Nadzirin, N., Velankar, S., Wodak, S.J., 2020. Modeling protein-protein, hyp.21.2.259.
protein-peptide, and protein-oligosaccharide complexes: CAPRI 7th edition. Protein.: Su, Y., Majtan, T., Freeman, K.M., Linck, R., Ponter, S., Kraus, J.P., Burstyn, J.N., 2013.
Struct., Funct., Bioinforma. 88 (8), 916–938. Comparative study of enzyme activity and heme reactivity in drosophila
Letunic, I., Khedkar, S., Bork, P., 2020. SMART: recent updates, new developments and melanogaster and homo sapiens cystathionine β-Synthases. Biochemistry 52 (4),
status in 2020. Nucleic Acids Res. 49 (D1), D458–D460. https://doi.org/10.1093/ 741–751. https://doi.org/10.1021/bi300615c.
nar/gkaa937. Thornton, J.M., Laskowski, R.A., Borkakoti, N., 2021. AlphaFold heralds a data-driven
Liu, S., Ando, F., Fujita, Y., Liu, J., Maeda, T., Shen, X., Kikuchi, K., Matsumoto, A., revolution in biology and medicine. Nat. Med. 27 (10), 1666–1669.
Yokomori, M., Tanabe-Fujimura, C., Shimokata, H., Michikawa, M., Komano, H., Tian, J.-L., Ren, A., Wang, T., Zhu, J., Hu, Y.-R., Shi, L., Yu, H.-S., Zhao, M.-W., 2019.
Zou, K., 2019. A clinical dose of angiotensin-converting enzyme (ACE) inhibitor and Hydrogen sulfide, a novel small molecule signalling agent, participates in the
heterozygous ACE deletion exacerbate Alzheimer’s disease pathology in mice. regulation of ganoderic acids biosynthesis induced by heat stress in Ganoderma
J. Biol. Chem. 294 (25), 9760–9770. https://doi.org/10.1074/jbc.RA118.006420. lucidum. Fungal Genet. Biol. 130, 19–30. https://doi.org/10.1016/j.
Liu, Y.J., Du, J.L., Cao, L.P., Jia, R., Shen, Y.J., Zhao, C.Y., Xu, P., Yin, G.J., 2015. Anti- fgb.2019.04.014.
inflammatory and hepatoprotective effects of Ganoderma lucidum polysaccharides Tu, Y., Kreinbring, C.A., Hill, M., Liu, C., Petsko, G.A., McCune, C.D., Berkowitz, D.B.,
on carbon tetrachloride-induced hepatocyte damage in common carp (Cyprinus Liu, D., Ringe, D., 2018. Crystal structures of cystathionine β-synthase from
carpio L.). Int Immunopharmacol. 25 (1), 112–120. https://doi.org/10.1016/j. saccharomyces cerevisiae: one enzymatic step at a time. Biochemistry 57 (22),
intimp.2015.01.023. 3134–3145. https://doi.org/10.1021/acs.biochem.8b00092.
Liu, Yn, Grimm, M., Dai, W.-t, Hou, M.J., Xiao, Z.-X., Cao, Y., 2019. CB-Dock: a web Tzakos, A.G., Galanis, A.S., Spyroulias, G.A., Cordopatis, P., Manessi-Zoupa, E.,
server for cavity detection-guided proteiŽ ligand blind docking. Acta Pharmacol. Sin. Gerothanassis, I.P., 2003. Structure–function discrimination of the N- and C-
41, 138–144. catalytic domains of human angiotensin-converting enzyme: implications for Cl–
Lubbe, L., Cozier, G.E., Oosthuizen, D., Acharya, K.R., Sturrock, E.D., 2020. ACE2 and activation and peptide hydrolysis mechanisms. Protein Eng., Des. Sel. 16 (12),
ACE: structure-based insights into mechanism, regulation and receptor recognition 993–1003. https://doi.org/10.1093/protein/gzg122.
by SARS-CoV. Clin. Sci. 134 (21), 2851–2871. https://doi.org/10.1042/ van Zundert, G.C.P., Rodrigues, J.P.G.L.M., Trellet, M., Schmitz, C., Kastritis, P.L.,
CS20200899. Karaca, E., Melquiond, A.S.J., van Dijk, M., de Vries, S.J., Bonvin, A.M.J.J., 2016.
Lüthy, R., Bowie, J.U., Eisenberg, D., 1992. Assessment of protein models with three- The HADDOCK2.2 web server: user-friendly integrative modeling of biomolecular
dimensional profiles. Nature 356 (6364), 83–85. complexes. J. Mol. Biol. 428 (4), 720–725. https://doi.org/10.1016/j.
Ma, F.-F., Wang, H., Wei, C.-K., Thakur, K., Wei, Z.-J., Jiang, L., 2019. Three novel ACE jmb.2015.09.014.
inhibitory peptides isolated from Ginkgo biloba seeds: purification, inhibitory Vozdek, R., Hnízda, A., Krijt, J., Kostrouchová, M., Kožich, V., 2012. Novel structural
kinetic and mechanism. Front. Pharmacol. 9, 1579. arrangement of nematode cystathionine β-synthases: characterization of
Masuyer, G., Yates, C.J., Sturrock, E.D., Acharya, K.R., 2014. Angiotensin-I converting Caenorhabditis elegans CBS-1. Biochem. J. 443 (2), 535–547. https://doi.org/
enzyme (ACE): structure, biological roles, and molecular basis for chloride ion 10.1042/bj20111478.
dependence. Biol. Chem. 395 (10), 1135–1149. https://doi.org/10.1515/hsz-2014- de Vries, S.J., Bonvin, A.M., 2011. CPORT: a consensus interface predictor and its
0157. performance in prediction-driven docking with HADDOCK. PloS One 6 (3), e17695.
Matoba, Y., Yoshida, T., Izuhara-Kihara, H., Noda, M., Sugiyama, M., 2017. Vy, T.T., Heo, S.-Y., Jung, W.-K., Yi, M., 2020. Spontaneous hinge-bending motions of
Crystallographic and mutational analyses of cystathionine β-synthase in the H(2) S- angiotensin i converting enzyme: role in activation and inhibition. Molecules 25 (6),
synthetic gene cluster in Lactobacillus plantarum. Protein Sci. 26 (4), 763–783. 1288.
https://doi.org/10.1002/pro.3123. Wachtel-Galor, S., Yuen, J., Buswell, J.A., Benzie, I.F.F., 2011. Ganoderma lucidum
Mohamad Ansor, N., Abdullah, N., Aminudin, N., 2013. Anti-angiotensin converting (Lingzhi or Reishi): A Medicinal Mushroom. In: Benzie, I.F.F., Wachtel-Galor, S.
enzyme (ACE) proteins from mycelia of Ganoderma lucidum (Curtis) P. Karst. In: (Eds.), Herbal Medicine: Biomolecular and Clinical Aspects. CRC Press/Taylor &
BMC Complement Altern Med, 13. https://doi.org/10.1186/1472-6882-13-256. Francis, Boca Raton (FL).
Natesh, R., Schwager, S.L., Sturrock, E.D., Acharya, K.R., 2003. Crystal structure of the Watermeyer, J.M., Sewell, B.T., Schwager, S.L., Natesh, R., Corradi, H.R., Acharya, K.R.,
human angiotensin-converting enzyme-lisinopril complex. Nature 421 (6922), Sturrock, E.D., 2006. Structure of testis ACE glycosylation mutants and evidence for
551–554. https://doi.org/10.1038/nature01370. conserved domain movement. Biochemistry 45 (42), 12654–12663. https://doi.org/
Notredame, C., Higgins, D.G., Heringa, J., 2000. T-Coffee: a novel method for fast and 10.1021/bi061146z.
accurate multiple sequence alignment. J. Mol. Biol. 302 (1), 205–217. https://doi. Wei, L., Alhenc-Gelas, F., Corvol, P., Clauser, E., 1991. The two homologous domains of
org/10.1006/jmbi.2000.4042. human angiotensin I-converting enzyme are both catalytically active. J. Biol. Chem.
Oliveriusová, J., Kery, V., Maclean, K.N., Kraus, J.P., 2002. Deletion mutagenesis of 266 (14), 9002–9008.
human cystathionine beta-synthase. Impact on activity, oligomeric status, and S- Wei, L., Clauser, E., Alhenc-Gelas, F., Corvol, P., 1992. The two homologous domains of
adenosylmethionine regulation. J. Biol. Chem. 277 (50), 48386–48394. https://doi. human angiotensin I-converting enzyme interact differently with competitive
org/10.1074/jbc.M207087200. inhibitors. J. Biol. Chem. 267 (19), 13398–13405.

15
N.-Y. Goh et al. Computational Biology and Chemistry 96 (2022) 107620

Wiederstein, M., Sippl, M.J., 2007. ProSA-web: interactive web service for the metabolite pathways and small cysteine-rich proteins in the sclerotium of lignosus
recognition of errors in three-dimensional structures of proteins. Nucleic Acids Res rhinocerotis. PLOS ONE 10 (11), e0143549. https://doi.org/10.1371/journal.
35 (suppl_2), W407–W410. https://doi.org/10.1093/nar/gkm290. pone.0143549.
Xue, L., Rodrigues, J., Kastritis, P., Bonvin, A., Vangone, A., 2016. PRODIGY: a web Yap, H.-Y.Y., Fung, S.-Y., Ng, S.-T., Tan, C.-S., Tan, N.-H., 2015b. Genome-based
server for predicting the binding affinity of protein-protein complexes. proteomic analysis of Lignosus rhinocerotis (Cooke) Ryvarden sclerotium. Int. J.
Bioinformatics 32, btw514. https://doi.org/10.1093/bioinformatics/btw514. Med. Sci. 12 (1), 23–31. https://doi.org/10.7150/ijms.10019.
Yan, Y., He, J., Feng, Y., Lin, P., Tao, H., Huang, S.Y., 2020b. Challenges and Yap, H.-Y.Y., Tan, N.-H., Ng, S.-T., Tan, C.-S., Fung, S.-Y., 2018. Inhibition of protein
opportunities of automated protein-protein docking: HDOCK server vs human glycation by tiger milk mushroom [lignosus rhinocerus (cooke) ryvarden] and search
predictions in CAPRI Rounds 38-46. Proteins 88 (8), 1055–1069. https://doi.org/ for potential anti-diabetic activity-related metabolic pathways by genomic and
10.1002/prot.25874. transcriptomic data mining. Front. Pharmacol. 9 (103) https://doi.org/10.3389/
Yap, H.Y.Y., Tan, N.H., Ng, S.T., Tan, C.S., Fung, S.Y., 2018. Molecular attributes and fphar.2018.00103.
apoptosis-inducing activities of a putative serine protease isolated from Tiger Milk Yates, C.J., Masuyer, G., Schwager, S.L.U., Akif, M., Sturrock, E.D., Acharya, K.R., 2014.
mushroom (Lignosus rhinocerus) sclerotium against breast cancer cells in vitro. Molecular and thermodynamic mechanisms of the chloride-dependent human
PeerJ 6, e4940. https://doi.org/10.7717/peerj.4940. angiotensin-I-converting enzyme (ACE). J. Biol. Chem. 289 (3), 1798–1814. https://
Yap, H.-Y.Y., Chooi, Y.-H., Firdaus-Raih, M., Fung, S.-Y., Ng, S.-T., Tan, C.-S., Tan, N.-H., doi.org/10.1074/jbc.M113.512335.
2014. The genome of the Tiger Milk mushroom, Lignosus rhinocerotis, provides Zhang, C., Mortuza, S.M., He, B., Wang, Y., Zhang, Y., 2018. Template-based and free
insights into the genetic basis of its medicinal properties. BMC Genom. 15 (1), 635. modeling of I-TASSER and QUARK pipelines using predicted contact maps in
https://doi.org/10.1186/1471-2164-15-635. CASP12. Protein.: Struct., Funct., Bioinforma. 86 (S1), 136–151. https://doi.org/
Yap, H.-Y.Y., Chooi, Y.-H., Fung, S.-Y., Ng, S.-T., Tan, C.-S., Tan, N.-H., 2015a. 10.1002/prot.25414.
Transcriptome analysis revealed highly expressed genes encoding secondary

16

You might also like