Potentiality of Organosulfur Compounds Against SARS-CoV-2-Coinfected Bacteria Streptococcus pyogenes and S pneumoniae : A Cross-Platform Analysis from Computational Chemistry

Abstract

Objective

SARS-CoV-2 coinfection with Streptococcus pyogenes and S pneumoniae increases disease severity, thus deserving more in-depth understanding.

Methods

The potential of garlic extract against the bacteria has been theoretically examined and extended to synergic predictions on SARS-CoV-2 Delta and Omicron variants.

Results

Gas chromatography–mass spectrometry characterized five main components in the extract: benzeneacetaldehyde (A1; 5.65%), diallyl disulfide (A2; 19.27%), 1-allyl-3-methyltrisulfane (A3; 21.30%), methyl salicylate (A4; 7.49%), and diallyl trisulfide (A5; 39.26%). Bioassay-based inhibition zones were 8 ± 2 mm, and MIC values 64 mg·mL⁻¹ (S pneumoniae) and 128 mg·mL⁻¹ (S pyogenes). Density functional theory suggested the organosulfur components as the most promising inhibitors with electrophilic sites at their sulfur atoms. Molecular docking simulation agreed with their potential against S pneumoniae/S pyogenes luxS proteins and Delta/Omicron SARS-CoV-2 spike proteins, given by all docking score values being <−13 kcal·mol⁻¹. QSARIS and ADMET predicted their bio- and pharma-favored properties.

Conclusion

The consistency of a variety of computational outputs justifies the dietary–supplemental potential of garlic essential oil based on its organosulfur ingredients.

Keywords

SARS-CoV-2 experiment computation

Introduction

The respiratory system has been commonly known to be vulnerable to bacterial and viral coinfections, often resulting in significantly higher mortality and morbidity. Because of its severe impacts globally, the current SARS-CoV-2 pandemic and associated coinfections have gained special attention from the scientific community. In particular, for the latter, an estimate indicated that 94.2% of COVID-19-confirmed patients (in Jiangsu Province) were coinfected with one or more (up to 24) respiratory pathogens within 1–4 days of onset of the first infection;¹ Streptococcus pneumoniae was the most common, followed by Klebsiella pneumoniae and Haemophilus influenzae. Pal et al reported in detail three cases with SARS-CoV-2 complicated by S pneumoniae coinfection, revealing that all three patients had respiratory failure and required intensive oxygen supplementation.² Many bacterial species that cause disease in humans and animals belong to the Streptococci group. Various diseases are typically caused by this group in humans, eg arthritis, neonatal sepsis, meningitis, and pneumonia; they are also known mainly to cause mastitis in animals. Therefore, accumulating knowledge about the SARS-CoV-2-related coinfection, in general, and the respiratory potential bacteria, in particular, is of importance.

Streptococcus pyogenes, group A streptococcal (GAS), is a major causal agent of acute bacterial pharyngitis, popularly known as strep throat.³ This illness is known to affect ca 30% of children and over 10% of adults. Nearly 1000 million new cases of pharyngitis are diagnosed globally each year. As a result, S pyogenes infection is one of the leading healthcare costs.⁴ LuxS is essential for various S pyogenes virulence features, especially its capacity to build biofilm. Lyon et al conducted research that involved the mutation of the luxS gene, which resulted in altered virulence activity, increasing expression of the sagA gene and decreasing SpeB proteolytic activity.⁵ In another work on a mutant luxS gene, Kang et al revealed the importance of luxS in S pyogenes pathogenicity.⁶ These findings altogether demonstrated the important role of luxS in S pyogenes activities.⁶ The structure database of luxS protein of S pyogenes can be found and retrieved from UniProtKB with ID P0C0C7 (S-ribosylhomocysteinase from S suis; S-ribosylhomocysteine lyase). The biological assembly of protein P0C0C7 (UniProtKB: P0C0C7 (LUXS_STRPY)) of S pyogenes is shown in Figure 1a.

Figure 1.

Crystal structures of (a) LuxS protein of S pyogenes (UniProtKB: P0C0C7 (LUXS_STRPY)), (b) LuxS protein of S pneumoniae (UniProtKB: Q8DQF8 (Q8DQF8_STRR6)), (d) spike protein of SARS-CoV-2 Delta variant (DOI:10.2210/pdb7V7N/pdb), spike protein of (e) SARS-CoV-2 Omicron variant (DOI:10.2210/pdb7T9J/pdb), and structural formula of (c) benzylpenicillin (penicillin G) (D1) and (f) structural formula of molnupiravir (D2).

Infection with S pneumoniae has caused millions of deaths worldwide over decades. In fact, pneumonia is still the primary cause of juvenile death in poor countries, where infections kill 1.6 million people worldwide each year, ca 0.7–1 million children under the age of five, predominantly recorded in Asia and Africa, according to World Health Organization (WHO).⁷ Like other Gram-positive bacteria, S pneumoniae also releases autoinducers, ie Autoinducer-2 (AI-2) directly produced by the protein luxS. More particularly, AI-2 belongs to a family of signaling molecules activated in quorum sensing (QS) and is essential to build biofilm. The structure database of LuxS protein of S pneumoniae can be found and retrieved from UniProtKB: Q8DQF8 (Q8DQF8_STRR6) of S pneumoniae (strain ATCC BAA-255/R6). The biological assembly of protein Q8DQF8 (UniProtKB: Q8DQF8 (Q8DQF8_STRR6)) of S pneumoniae (strain ATCC BAA-255/R6) is shown in Figure 1b.

Benzylpenicillin (penicillin G), known as the “natural” penicillin, was used as a positive control. Penicillins possess a wide antibacterial spectrum. Billions of doses have been administered since the 1940s in different regions.⁸ S pneumoniae and Lancefield Group A, β-hemolytic streptococcus (S pyogenes) are known to be very sensitive to penicillin G. The structural formula of penicillin G is given in Figure 1c.

The SARS-CoV-2 pandemic has become a worldwide concern to public health because of its high incidence of infection and fatality. Because of unknown effective specific medications or vaccinations, the mortality toll has been escalating at an alarming pace. To control the overrapid spread of life-threatening infectious outbreaks, the idea of drug repositioning and alternative therapy are important strategies. Among the strains encountered, Delta and Omicron variants have gained special attention from the scientific community. Delta variant B.1.167.2, which has 23 mutations (12 are spike-based mutations), is the most well-known and deadliest SARS-CoV-2 variant.⁹ The key mutation of B.1.167.2 is L452R and P681R spike protein mutations. In detail, L452R improves binding affinity to Angiotensin-converting enzyme 2 (ACE2) receptor and evading CD8 T lymphocytes, while P681R cleaves the precursor to form two new spike proteins called S1 and S2, which in turn induce superior fusion on the virus-bound host cells. Omicron has 32 mutations, mainly at receptor binding domain (RBD), N-terminal domain (NTD), and receptor binding motif (RBM). These changes are thought to be able to neutralize antibodies produced by previous infections or vaccinations. In theory, if the inactivation of spike proteins is of efficient success, SARS-CoV-2 itself cannot bind to the host receptors, meaning that the virus is not able to interact with or enter human cells. The crystal structure of spike proteins of SARS-CoV-2 Delta and Omicron variants was recently determined and published in Protein Data Bank under the entries PDB-7V7N (http://doi.org/10.2210/pdb7V7N/pdb) and PDB-7V7N (http://doi.org/10.2210/pdb7T9J/pdb), respectively. Figure 1e–d shows their biological assemblies and that of molnupiravir (Figure 1f), an antiviral medication used to treat COVID-19 in those infected by SARS-CoV-2.

For centuries, garlic (Allium sativum L.), the most abundantly produced bulb after onion, has been serving nutritional and medicinal roles of importance. It is known that garlic essential oil contains a wide variety of special organosulfur compounds, which exhibit a variety of antioxidant, antibacterial, antifungal, antitumor, and antimicrobial properties. Among the main components of garlic essential oil, allicin is a typical reactive sulfur-containing organic compound that has received enormous attention.^10,11 In our preceding work, garlic essential oil was analyzed for chemical characterization, revealing the presence of 17 organosulfur compounds, accounting for 99.4% of its contents. These compounds were predicted to exhibit strong interactions with the amino acids of the protein ACE2 and the main protease PDB6LU7 of SARS-CoV-2 (wild variant).¹¹ This suggests, to a certain degree, that the organosulfur family still retains its potential for the mutated variant of SARS-CoV-2 (eg Delta and Omicron), in particular, and be effective against bacterial microorganisms (eg S pyogenes and S pneumoniae), in general.

In this work, evidence for the antibiotic activity of garlic extract was collected by experimental attempts, while the biological and pharmacological potentiality of individual components was predicted by cross-platform computational iterations. The theoretical consistency obtained can justify the potential of garlic as a supplemental food for the fight against SARS-CoV-2 coinfections.

Results and Discussion

Experimental

Chemical Composition

The chemical composition of the garlic extract, determined by gas chromatography–mass spectrometry (GC-MS) analysis, is summarized in Table 1; five components were characterized, and their structural formulae are shown in Figure 2. In particular, diallyl trisulfide (DTS) (A5; 39.26%) was found as the dominant compound, followed by 1-allyl-3-methyltrisulfane (A3; 21.30%) and diallyl disulfide (DDS) (A2; 19.27%). In addition, methyl salicylate (A4; 7.49%) and benzeneacetaldehyde (A1; 5.65%) comprised the minor components of the garlic extract. The components are consistent with those reported in the literature. For instance, María Molina-Calle et al observed that DDS and DTS are major garlic volatiles with a relative concentration of 51.7% (26.4% DDS and 25.3% DTS) for Chinese garlic, 43.3% (24.8% DDS and 18.5% DTS) for white garlic, and 42.4% (21.0% DDS and 21.4% DTS) for purple garlic.¹² It is noteworthy that this study was carried out through ultrasonic-assisted extraction, instead of conventional percolation, which is thought to induce higher penetration of the solvent into the herbal matrix. This not only encourages the diffusion of the low-polarized solute into the liquid phase without having to use nonpolar solvents but also helps reduce the time of extraction.

Figure 2.

Structural formulae of compounds (A1–A5) from the ethanol 70% extract of A sativum.

Table 1.

Identification of Major Volatile Bioactive Compounds in the Ethanol 70% Extract of Allium sativum.

No.	Retention time (min)	Substance	Formula	Notation	Percentage (%)
1	3.47	Benzeneacetaldehyde (hyacinthin)	C₈H₈O	A1	5.65
2	3.73	Diallyl disulfide	C₆H₁₀S₂	A2	19.27
3	4.24	1-Allyl-3-methyltrisulfane (allyl methyl trisulfide)	C₄H₈S₃	A3	21.30
4	4.63	Methyl salicylate	C₈H₈O₃	A4	7.49
5	5.46	Diallyl trisulfide	C₆H₁₀S₃	A5	39.26

Garlic oil is known to contain ca 20 organosulfur compounds, eg diallyl sulfide (DAS), DDS, DTS, and diallyl tetrasulfide (DATTS), likely to correlate with the biological activities of the plant. Regarding the essential oils (EOs) obtained from fresh bulbs of garlic more particularly, significant antibacterial activity was in vitro proven quantitatively against S aureus (inhibition zone 14.8 mm), P aeruginosa (inhibition zone 21.1 mm), and E coli (inhibition zone 11.0 mm).¹³ In this work, the main constituents determined were diallyl monosulfide, DDS, DTS, and DATTS.

From an experimental standpoint, it seems transparent that there is a solid correlation between garlic extract organosulfur compounds and biological activities of significance; yet, the related explanation for the factual observations has not been looked at in-depth from the standpoint of molecular electronic structures. Therefore, the chemical components determined in this work were also subjected to quantum-based analysis (ie density functional theory [DFT] calculation) in the attempt to provide a better understanding of their biological behavior from the view of the molecular constitution.

Antimicrobial Activity

Garlic extract inhibited the growth of S pneumoniae and S pyogenes, as shown in the agar diffusion method. Inhibition zones of 8 ± 2 mm were recorded, which is regarded as moderate. The positive control, penicillin G, produced an inhibition zone ≥ 30 ± 2 mm. The inhibition zones were observed for 18 h. The final result for S pneumoniae was two inhibitory circles, one showing complete inhibition and the other partial inhibition. This suggests that the bacteria were inhibited for the first few hours in the initial growth phase, but growth renewed after the garlic extract volatile components had vaporized at atmospheric pressure.

Therefore, disc volatilization assays were utilized in an attempt to give a more in-detail view of the antibacterial activity of these volatile components. In brief, garlic extract–impregnated filter paper discs were inoculated with microorganism-cultured agars in Petri dishes and sealed with parafilm to prevent leakage of vaporizable components into the atmosphere. In the absence of garlic extract (negative control), the growth of the two Streptococcus strains could be observed by the naked eye and through the beta-hemolysis phenomena, given by the uniform breakdown of red blood cells leaving clearer regions around the bacterial growth. In contrast, the presence of the garlic extract vapor phase resulted in no colonies of the two test strains on the plates after incubation. This suggests that the bacterial growth was inhibited even at the initial growth phase. In detail, most bacterial cells were killed after treatment with 0.0036 mg garlic extract per cm² in the area unit or 3.6 mg per L in the volume unit. This value is calculated based on the growth area of bacteria in the 90 mm Petri disc, which is 58 cm² in area and 0.058 L in volume. The results observed from the disc volatilization assays also indicate that the nonvolatile compounds did not play a significant role in the observed antibacterial activity.

The efficient susceptibility of the bacteria to the extract is assessed by minimum inhibitory concentration (MIC) values. The results reveal that S pneumoniae was inhibited by the garlic extract at a concentration of 64 mg·mL⁻¹; the corresponding figure for S pyogenes was 128 mg·mL⁻¹. This also indicates the greater susceptibility of the former to the garlic extract, in comparison to that of the latter. However, the MIC value of garlic extract exhibited a large variation compared with that of penicillin G against Streptococcus spp., which was 0.03 μg·mL⁻¹. This value was in line with NCCLS-recommended interpretive criteria for penicillin (ie ≤0.06 μg·mL⁻¹, susceptible; 0.12–1 μg·mL⁻¹, intermediate; and ≥2 μg·mL⁻¹, resistant).

Sulfur-containing compounds found in garlic extract, eg allicin, DDS, and DTS, were reportedly able to interact with enzyme-based thiol groups, eg alcohol dehydrogenase or thioredoxin reductase, and degrade the bacterial membrane stability.¹⁴ In general, garlic extract had already been shown to be active both in vitro and in vivo against S mutans, the principal etiological bacterium in dental caries, with MIC values ranging from 0.5 to 128 mg·mL⁻¹.¹⁵ Herein, the antibacterial activities against two other pathogenic Streptococcus, eg S pyogenes and S pneumoniae, are reported with MIC values ranging from 64 to 128 mg·mL⁻¹.

The effectiveness of the garlic extract against the two Streptococcus spp. is experimentally supported; however, the knowledge of the biological behavior of the candidate–microorganism interactions is still inadequate from the theoretical view. Nevertheless, the evidence from the existing literature seems to direct the responsibility of significance toward sulfur-containing compounds. Therefore, another computational technique (ie molecular docking simulation) was utilized in an attempt to reach an appropriate explanation. Furthermore, due to its infectiousness and transmissibility, in this work, SARS-CoV-2 was not subjected to direct in vitro tests but to in silico predictions. The results can serve as supplemental understanding of the overall potential of garlic extract organosulfur compounds against the dangerous respiratory micropathogens.

Computational

Density Functional Theory–Based Chemical Properties

Results from DFT calculation can be utilized either to check the unusual bonding parameters (if any) in the structural formulae of the bioactive compounds (A1–A5) elucidated from GC-MS or to retrieve their molecular electronic configurations. In short, the computation can provide a certain insight into chemical activities from the standpoint of quantum mechanics.

The optimized structures are shown in Figure 3; in-detail parameters of the structures are presented in Table 2. Overall, the computation can reach a geometry convergence regarding all the formulae, implying the compounds’ in-nature stable existence. No unusual bonding parameters were detected either in lengths or angles. In particular, all the figures meet the characteristic values, eg 1.54 Å for C–C, 1.09 Å for C-H, 1.43 Å for C-O (oxatriquinane), 1.23 Å for C=O (carbonyl compounds), 1.82 Å for C-S, and 1.8–3.0 Å for S-S. The corresponding ground state electronic energies register in the range between −1429.226248 au and −384.931174 au, which are commonly agreed to be highly stable. All the structures are also considered polarized molecules, especially A1 (2.671 Debye) and A4 (2.565 Debye), given by their (nonzero) dipole moment. In essence, high dipole moment values are likely conducive to the formation of polarized forces, such as van de Waals interactions or ionic bonds in ligand–protein inhibition.¹⁶ However, the inhibitory effectiveness might rely heavier on the electronic structure of molecular orbitals because of “covalentlike” bonding formation.

Figure 3.

Optimized structures of A1–A5 calculated by DFT at level of theory M052X/6-311++g (d,p). Bond lengths are given in Å and angles in °.

Table 2.

Ground State Electronic Energy and Dipole Moment Values of A1–A5 Calculated by DFT at Level of Theory M052X/6-311++g(d,p).

No.	Substance	Notation	Ground state electronic energy (au)	Dipole moment (Debye)
1	Benzeneacetaldehyde (hyacinthin)	A1	−384.931174	2.671
2	Allyl disulfide	A2	−1031.024923	2.155
3	1-Allyl-3-methyltrisulfane (allyl methyl trisulfide)	A3	−1351.822529	1.465
4	Methyl salicylate	A4	−535.430094	2.565
5	Allyl trisulfide	A5	−1429.226248	1.499

Figure 4 presents the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of the optimized structures; Table 3 gives the in-detail quantum parameters. Except for A1 and A4, the localization of electron densities of the others distributes rather evenly over their molecular plane regarding both HOMOs and LUMOs. According to the theoretical charge-transferring properties, HOMO represents an intermolecular electron donation tendency, and LUMO represents an electron-accepting ability. In theory, this means that the host molecules are highly flexible for intermolecular interaction. In the case of A1, although there is an aromatic ring in its structure, the low-density LUMO at the region predicts a cation-attraction tendency (by its dominating HOMO). In the case of A4, although the methoxy group is often assigned to an electron affinity site by theoretical elucidation, the computer-calculated output indicates otherwise; in particular, negligible electronic densities were given by either HOMO or LUMO, implicating almost inert chemical activation. Quantitatively, all structures are considered to exhibit electronic stability of very high significance given low E_HOMO, all under −8 eV (while commonly agreed under −5 eV); in addition, their insulation–semiconduction band-gap energy (3.2 eV < ΔE_GAP< 9 eV)¹⁷ might suggest intermolecular binding capability toward polypeptide structures, whose electric conductivity was well-observed and explained based on superexchange theory (or electron tunneling).^18,19

Figure 4.

HOMO and LUMO of A1–A5 calculated by DFT at level of theory M052X/def2-TZVPP.

Table 3.

Quantum Chemical Parameters of Compounds A1–A5 Calculated by NBO at the Level of Theory M052X/def2-TZVPP.

Compound	E_HOMO (eV)	E_LUMO (eV)	ΔE_GAP	I = −E_HOMO	A = −E_LUMO	χ	μ	η	S = 1/η
A1	−8.338	0.335	8.674	8.338	−0.335	4.002	−4.002	4.337	0.231
A2	−8.020	0.073	8.093	8.020	−0.073	3.974	−3.974	4.047	0.247
A3	−8.026	−0.303	7.723	8.026	0.303	4.165	−4.165	3.862	0.259
A4	−8.084	−0.425	7.659	8.084	0.425	4.255	−4.255	3.830	0.261
A5	−7.991	−0.325	7.665	7.991	0.325	4.158	−4.158	3.833	0.261

Energy gap (ΔE_GAP = E_LUMO−E_HOMO); ionization potential I = −E_HOMO; electron affinity A = −E_LUMO; electronegativity χ = (I + A)/2; chemical potential μ = −χ = −(∂E/∂N)_ν₍ _r ₎; hardness η = (I−A)/2; softness S = 1/η.

Molecular electronic potential (MEP) maps of the optimized structures given in Figure 5 can provide another view of their chemical activities based on the distribution of potential energy; by convention, reddish shades represent the negative regions (related to electrophilic reactivity), while bluish shades represent positive regions (related to nucleophilic reactivity). Overall, the structures (A1–A5) can also be categorized into two groups from this standpoint. The organosulfur compounds (A2, A3, A5) possess rather even-distributed MEP maps with low levels of differences between their molecular regions (ca ±2 × 10⁻³ au), implying that the structures are considered flexible to form electron-accepting/donating bonding over their molecular plane. The most pronounced electrophilic sites refer to their organosulfur groups, theoretically predicted as the most active regions when in an intermolecular interaction. In contrast, the quantum-based calculation predicts those of A1 and A4 as highly restricted to their ketone functions, while there is no region showing significant nucleophilicity. These sites in general give information about the regions where the compounds can establish intermolecular interactions, especially via electrophilic links (which are more mobile than those of nucleophilic counterparts).

Figure 5.

Molecular electrostatic potential (MEP) formed by mapping of total density over the electrostatic potential of A1–A5 calculated by DFT at level of theory M052X/def2-TZVPP.

The Fukui function can predict the reaction sites favored by nucleophilic (f⁺), electrophilic (f⁻), and radical (f⁰) attacks. In principle, reactivity of atoms is measured by the corresponding Fukui indices, which are summarized in Table 4. Similar to those retrieved from MEP analysis, organosulfur compounds (A2, A3, A5) center their reactivity at their sulfur atoms, either for nucleophilic (f⁺ > +0.2), electrophilic (f⁻ > +0.1), or radical attacks (f⁰ > +0.1). Regarding others, the aromatic ring and ketone groups seem to be the primary reactive sites of A4 given by the f⁺, f⁻, and f⁰ indices registering moderate nonnegative values; meanwhile, significantly high figures for A1 might encourage further theoretical argument and experimental validation from extended work. However, it is noteworthy that unlike the population analyses carried out solely on neutral species, the Fukui indices represent the “regional disturbance” in the electronic distribution if a molecule is added/removed an infinitesimal amount of charge units (by original definition) or a whole electron (by technical algorithm via the differences in Mulliken atomic charges in this work).

Table 4.

Mulliken Atomic Charges and Fukui Indices of A1–A5.

	Atom	q(N)	q(N + 1)	q(N−1)	f⁺	f⁻	f⁰
A1	1 O	-0.231194	−0.187082	−0.308096	0.044112	0.076902	0.060507
	2 C	1.022901	0.921867	−2.244509	−0.101034	3.267410	1.583188
	3 C	−0.264169	−0.239174	2.961293	0.024995	−3.225462	−1.600234
	4 C	0.032887	0.148756	−2.006459	0.115869	2.039346	1.077608
	5 C	−0.838169	−0.748629	1.288316	0.089540	−2.126485	−1.018473
	6 C	−0.305252	−0.257819	−0.002595	0.047433	−0.302657	−0.127612
	7 C	−0.363245	−0.323050	0.438741	0.040195	−0.801986	−0.380896
	8 C	−0.152273	−0.065155	0.890974	0.087118	−1.043247	−0.478065
	9 C	−0.169195	−0.158945	1.740916	0.010250	−1.910111	−0.949931
A2	1 S	0.012224	0.396122	−0.266872	0.383898	0.279096	0.331497
	2 S	0.012221	0.396080	−0.266860	0.383859	0.279081	0.331470
	3 C	−0.472560	−0.632160	−0.148721	−0.159600	−0.323839	−0.241720
	4 C	−0.472503	−0.632093	−0.148649	−0.159590	−0.323854	−0.241722
	5 C	0.261107	0.251097	0.175092	−0.010010	0.086015	0.038003
	6 C	0.261044	0.251020	0.174979	−0.010024	0.086065	0.038021
	7 C	−0.548797	−0.515803	−0.700996	0.032994	0.152199	0.092597
	8 C	−0.548784	−0.515774	−0.700940	0.033010	0.152156	0.092583
A3	1 S	−0.024702	0.225600	−0.384557	0.250302	0.359855	0.305079
	2 S	0.065258	0.279636	−0.069818	0.214378	0.135076	0.174727
	3 S	0.138913	0.413849	−0.183110	0.274936	0.322023	0.298480
	4 C	−0.522006	−0.656988	−0.366158	−0.134982	−0.155848	−0.145415
	5 C	0.277705	0.306236	0.338457	0.028531	−0.060752	−0.016111
	6 C	−0.543719	−0.505143	−0.638074	0.038576	0.094355	0.066466
	7 C	−0.603290	−0.657236	−0.508045	−0.053946	−0.095245	−0.074596
A4	1 O	−0.071862	−0.074122	−0.074837	−0.002260	0.002975	0.000358
	2 O	−0.155218	−0.009028	−0.182616	0.146190	0.027398	0.086794
	3 O	−0.226780	−0.152675	−0.367973	0.074105	0.141193	0.107649
	4 C	0.496871	0.611257	0.476804	0.114386	0.020067	0.067227
	5 C	−0.157806	−0.113288	−0.232662	0.044518	0.074856	0.059687
	6 C	−0.503699	−0.511671	−0.559133	−0.007972	0.055434	0.023731
	7 C	−0.279488	−0.192265	−0.330633	0.087223	0.051145	0.069184
	8 C	−0.400133	−0.256695	−0.407335	0.143438	0.007202	0.075320
	9 C	−0.082015	−0.085881	−0.244791	−0.003866	0.162776	0.079455
	10 C	0.184974	0.126713	0.147170	−0.058261	0.037804	−0.010229
	11 C	−0.205410	−0.202669	−0.215285	0.002741	0.009875	0.006308
A5	1 S	0.006169	0.288375	−0.344174	0.282206	0.350343	0.316275
	2 S	0.006178	0.288439	−0.344175	0.282261	0.350353	0.316307
	3 S	0.065914	0.277142	−0.014015	0.211228	0.079929	0.145579
	4 C	−0.464926	−0.629786	−0.399705	−0.164860	−0.065221	−0.115041
	5 C	−0.464922	−0.629798	−0.399672	−0.164876	−0.065250	−0.115063
	6 C	0.266006	0.301172	0.445377	0.035166	−0.179371	−0.072103
	7 C	0.266005	0.301175	0.445371	0.035170	−0.179366	−0.072098
	8 C	−0.555511	−0.503880	−0.708410	0.051631	0.152899	0.102265
	9 C	−0.555520	−0.503895	−0.708435	0.051625	0.152915	0.102270

This means that the reference is more useful for chemical reactions (based on electronic reconfiguration) than for intermolecular interaction (based on dipole–dipole forces); meanwhile, the latter is commonly accepted as the principal driving force for ligand–protein inhibition.

Docking-Based Inhibitory Properties

Retrieval from molecular docking simulation can be utilized either to verify the inhibitory effectiveness observed by experimental bioassays or to study the ligand–protein inhibitory behavior. In principle, the total docking score (DS) values and the number of hydrogen bonds are the most important parameters as the former represents Gibbs free energy of the inhibitory complexes, while the latter gives information on strong intermolecular bonds; besides, root-mean-square deviation (RMSD) (ie the average distance between backbone atoms; bioconformational rigidity or ligand–protein fitting) and van de Waals interactions (ie hydrophobic forces) may be considered to a certain degree. In each ligand–protein duo, only the most stable intermolecular structure is selected for more in-depth discussion since it most likely corresponds to the dominant product in reality. In brief, the computation can provide a certain insight into ligand–protein interaction potential from the standpoint of classical mechanics.

LuxS Proteins (Q8DQF8 and P0C0C7)

Figure 6 presents the most susceptible sites (sites 1–4) of the targeted proteins (Q8DQF8 and P0C0C7) to the ligands (A1–A5); the main screening data are summarized in Table 5; the control drug (D1) is penicillin G. Overall, although the proteins show their susceptibility differently toward particular ligands, the screening still predicts the organosulfur compounds (A2, A3, A5) as promising inhibitors against the protein structures in general given lower chemical potential (DS < −14 kcal·mol⁻¹) and higher number of strong interactions (hydrogen bonds > 4). In-bold inhibitory systems are selected for more in-depth analysis.

Figure 6.

Quaternary structures of protein Q8DQF8 and protein P0C0C7 with approachable sites by A1–A5 and D1: site 1 (gray), site 2 (yellow), site 3 (blue), and site 4 (orange).

Table 5.

Prescreening Results on Inhibitability of A1–A5 and Controlled Drug (Penicillin G) (D1) toward the sites of proteins Q8DQF8 and P0C0C7.

P	L	Site 1		Site 2		Site 3		Site 4
P	L	E	N	E	N	E	N	E	N
Q8DQF8	A1	−12.1	4	−10.9	2	−10.4	2	−10.8	2
	A2	−11.7	3	−13.7	5	−10.3	3	−9.6	2
	A3	−9.0	2	−13.9	5	−8.9	2	−10.4	3
	A4	−8.4	2	−7.0	1	−8.3	2	−11.8	3
	A5	−11.9	3	−14.2	7	−11.0	3	−10.8	3
	D1	−11.4	3	−10.6	2	−13.0	4	−9.3	2
P0C0C7	A1	−11.7	3	−10.0	2	−9.7	2	−8.9	2
	A2	−14.4	7	−10.5	3	−11.6	4	−10.8	3
	A3	−11.9	4	−14.1	7	−12.0	4	−10.4	3
	A4	−9.5	2	−9.7	2	−12.0	4	−10.3	3
	A5	−11.0	4	−11.9	5	−15.0	10	−12.0	4
	D1	−8.7	2	−11.0	3	−11.8	4	−14.0	6

P: protein; L: ligand; E: DS value (kcal·mol−¹); N: number of hydrophilic interactions.

The selected data are summarized in Table 6 (ligand–Q8DQF8) and Table 7 (ligand–P0C0C7). According to the theoretical interpretation as given, the most effective ligand–protein inhibitory structures regarding S pneumoniae are in the order: A5–Q8DQF8 (DS −14.2 kcal·mol⁻¹; RMSD 1.10 Å) > A3–Q8DQF8 (DS −13.9 kcal·mol⁻¹; RMSD 1.13 Å)≈A2–Q8DQF8 (DS −13.7 kcal·mol⁻¹; RMSD 1.42 Å) > A1–Q8DQF8 (DS −12.1 kcal·mol⁻¹; RMSD 1.22 Å)≈A4–Q8DQF8 (DS −11.8 kcal·mol⁻¹; RMSD 1.41 Å). In terms of S pyogenes, the corresponding order is A5–P0C0C7 (DS −15.1 kcal·mol⁻¹; RMSD 1.04 Å) > A2–P0C0C7 (DS −14.4 kcal·mol⁻¹; RMSD 1.06 Å)≈A3–P0C0C7 (DS −14.1 kcal·mol⁻¹; RMSD 0.87 Å) > A4–P0C0C7 (DS −12.0 kcal·mol⁻¹; RMSD 1.74 Å) > A1–P0C0C7 (DS −11.7 kcal·mol⁻¹; RMSD 1.20 Å). In summary, the first three candidates (A5, A2, and A3) are thought able to induce effective inhibitory effects with DS values < −13 kcal·mol⁻¹; in fact, works on single-protein structures (ie diabetes-related proteins) of ours correlated with these ranges with assay-based IC₅₀ values < 10 μM (control drug IC₅₀ ca 300 μM).^20,21 However, the comparison can be justified loosely since there are many proteins serving different living roles in a bacterium; in other words, the correlation between the inhibition of the proteins selected in in silico simulations and the inhibition of the bacterial growth in in vitro bioassays is still unclear. Therefore, these compounds, ie DDS (A2), 1-allyl-3-methyltrisulfane (A3), and DTS (A5), deserve further experimental attempts for analytical graded isolation, structural redetermination, and in vitro bioassays. They also account for the major components of the garlic extract, quantified by GC-MS.

Table 6.

Molecular Docking Simulation Results for Ligand–Q8DQF8 Inhibitory Complexes.

Ligand–protein			Hydrogen bond						van der Waals interaction
Name	DS	RMSD	L	P		T	D	E	van der Waals interaction
A1–Q8DQF8	−12.1	1.22	O	N	Arg B318	H-acceptor	2.85	−1.5	Leu A109. Phe B6, Glu B321, Ser B10, Gln B338, Ser B9, Leu A98, Lys B5
			O	N	Arg B325	H-acceptor	3.14	−1.5
			6-ring	C	Leu A101	π-H	3.96	−0.7
			6-ring	N	Arg B325	π-cation	3.43	−0.6
A2–Q8DQF8	−13.7	1.42	S	O	Glu B317	H-donor	3.57	−0.5	Phe A96, Arg B320, Leu A97
			S	O	Glu B321	H-donor	3.70	−0.8
			S	C	Arg B318	H-acceptor	3.42	−0.6
			S	C	Leu A98	H-acceptor	3.76	−0.8
			S	N	Ser A99	H-acceptor	3.79	−2.7
A3–Q8DQF8	−13.9	1.13	S	O	Gly B173	H-donor	3.59	−0.3	Gly A173, Gln A93, Gln B93
			S	O	Gln A94	H-donor	3.79	−0.7
			S	C	Ala A174	H-acceptor	3.84	−0.9
			S	C	Ala B174	H-acceptor	3.55	−0.9
			S	N	Gln B94	H-acceptor	3.78	−1.0
A4–Q8DQF8	−11.8	1.41	O	N	Glu B321	H-donor	3.01	−0.8	Leu A101, Ala B13, Leu A98, Lys A100, Phe B6, Ser B9, Ser B10
			O	N	Arg B325	H-acceptor	2.77	−2.5
			O	N	Arg B325	H-acceptor	2.97	−1.2
A5–Q8DQF8	−14.2	1.10	S	O	Ser A99	H-donor	3.26	−1.1	Arg B320, Lys A100, His B14, Leu A97
			S	O	Gly B317	H-donor	3.45	−1.4
			S	O	Glu B321	H-donor	3.66	−2.0
			S	N	Ser A99	H-acceptor	3.49	−2.5
			S	C	Arg B318	H-acceptor	3.15	−0.8
			S	C	Leu A98	H-acceptor	3.81	−0.8
			S	N	Ser A99	H-acceptor	3.71	−1.2
D1–Q8DQF8	−13.0	0.92	S	C	Ala A119	H-acceptor	2.27	−0.8	Asn B118, Asn B130, Ala B119, Phe B229, Phe A193, Asn A191, Thr A117, Asn A118, Ile A120
			O	N	Ile B120	H-acceptor	4.44	−1.5
			O	N	Asn B191	H-acceptor	3.76	−1.3
			N	6-ring	Phe A229	H-π	4.34	−0.9

DS: Docking score energy (kcal·mol⁻¹); RMSD: root-mean-square deviation (Å); L: ligand; P: protein; T: type; D: distance (Å); E: energy (kcal·mol⁻¹).

Table 7.

Molecular Docking Simulation Results for Ligand–P0C0C7 Inhibitory Complexes.

Ligand–protein			Hydrogen bond						van der Waals interaction
Name	DS	RMSD	L	P		T	D	E	van der Waals interaction
A1–P0C0C7	−11.7	1.20	O	N	Arg A42	H-acceptor	3.39	−1.4	His A14, Cys B127, Ser B126, Cys A82, His B57, Glu B60
			O	N	Gly B128	H-acceptor	3.40	−0.7
			6-ring	5-ring	His B61	π-π	3.63	−0.0
A2–P0C0C7	−14.4	1.06	S	O	Ser B126	H-donor	3.11	−0.7	His A14, Cys A82, His B57
			S	O	Glu B61	H-donor	3.48	−0.3
			S	N	His B61	H-acceptor	3.34	−1.5
			S	C	Cys B127	H-acceptor	3.80	−0.8
			S	S	Cys B127	H-acceptor	3.70	−1.0
			S	N	Gly B128	H-acceptor	3.08	−2.2
			S	N	His B61	H-acceptor	3.47	−3.0
A3–P0C0C7	−14.1	0.87	S	O	Asn B129	H-donor	2.98	1.9	Asn A49, Leu C144, Glu A48, Ala A50, Trp C141, Ala C50
			S	O	Asp D52	H-donor	3.20	−0.4
			S	N	Asn B49	H-acceptor	2.78	−0.8
			S	C	Asn A47	H-acceptor	3.63	−1.0
			S	C	Asn B129	H-acceptor	3.47	−0.6
			S	N	Lys B131	H-acceptor	2.96	−2.2
			S	C	Lys B131	H-acceptor	3.46	−1.1
A4–P0C0C7	−12.0	1.74	O	N	His A61	H-donor	3.05	−3.6	Gly A128, His A57, Glu A60, Asp A76, Phe B80, Ser A78, Arg B42, His B14, Ser A126
			C	S	Cys A127	H-donor	3.22	−1.2
			O	N	Gly B81	H-acceptor	3.14	−1.3
			6-ring	S	Cys B82	π -H	3.48	−0.6
A5–P0C0C7	−15.1	1.04	S	O	Glu A60	H-donor	3.02	−0.6	Ser A126, His B14, Arg B23, Leu B12, Phe B80. Gly B81
			S	N	His A61	H-donor	3.56	−1.1
			S	S	Cys A127	H-donor	3.71	−0.1
			S	N	Gly A128	H-acceptor	3.81	−1.9
			S	N	Arg B42	H-acceptor	3.80	−3.8
			S	N	Arg B42	H-acceptor	3.93	−2.4
			S	C	His A57	H-acceptor	3.97	−0.8
			S	N	Cys B82	H-acceptor	3.81	−2.0
			S	C	His A57	H-acceptor	3.04	−1.7
			S	N	Gly A128	H-acceptor	3.96	−1.5
D1–P0C0C7	−14.3	1.20	S	O	Glu C48	H-donor	3.49	−0.1	Ile B124, Asn C47, Pro C46, Asn A49, Glu A48, Leu A43, Ala A19, Ile A44, Leu A144
			O	N	Glu A148	H-acceptor	2.90	−2.0
			O	N	Lys B131	Ionic	3.97	−0.6
			O	N	Lys B131	Ionic	3.30	−2.8
			6-ring	N	Gln A45	π -H	3.65	−0.8
			6-ring	N	Gln A148	π -H	3.58	−1.0

DS: docking score energy (kcal·mol⁻¹); RMSD: root-mean-square deviation (Å); L: ligand; P: protein; T: type; D: distance (Å); E: energy (kcal·mol⁻¹).

The corresponding visual projections (3D in-pose morphology and 2D interaction map) are presented in Figure 7 (ligand-Q8DQF8) and Figure 8 (ligand-P0C0C7). Regarding A1, its aromatic ring only plays a role as a nucleophilic attractor in hydrogen-like bonding (ie π-H and π-cation); meanwhile, that of A4 almost only contributes to the ligand–protein construction by van de Waals interactions via its compositional atoms. Regarding A4, its methoxy group exhibits no intermolecular activity despite the high reactivity of an individual oxygen atom. Furthermore, A2, A3, and A5 can form up to 10 hydrophilic bonds with their targeted protein structures via only 2–3 sulfur atoms in their molecules; on the other side, the ketone groups serve as the primary sites of reactivity of A1 and A4, being involved in all the related complexes. Finally, the sites are rather large and open, in comparison to the size of the inhibitors, meaning there is room for further functionalization/modification on these ligands regarding the same host proteins. For example, one approach could be to improve the electrophilic reactivity of the sulfur-based regions even more.

Figure 7.

Visual presentation and in-pose interaction maps with ligand-Q8DQF8 (A1–A5 and D1).

Figure 8.

Visual presentation and in-pose interaction maps with ligand-P0C0C7 (A1–A5 and D1).

Spike Proteins (7V7N and 7T9J)

Figure 9 presents the most susceptible sites (sites 1–4) of the targeted proteins (7V7N and 7T9J) to the ligands (A1–A5); the main screening data are summarized in Table 8; the control drug (D2) was molnupiravir. It seems that the protein structures are more vulnerable to the inhibitors at site 3, which is specified in blue, although the tendency is not explicitly consistent. The organosulfur compounds (A2, A3, A5) still outweigh their counterparts in terms of inhibitability. In-bold inhibitory systems are selected for more in-depth analysis.

Figure 9.

Quaternary structures of protein 7V7N and protein 7T9J with approachable sites by A1–A5 and D2: site 1 (gray), site 2 (yellow), site 3 (blue), and site 4 (orange).

Table 8.

Prescreening Results on Inhibitability of A1–A5 and Controlled Drug Molnupiravir (D2) Toward the Sites of Proteins 7V7N and 7T9J.

P	L	Site 1		Site 2		Site 3		Site 4
P	L	E	N	E	N	E	N	E	N
7V7N	A1	−10.9	3	−9.8	2	−13.1	5	−10.1	2
	A2	−11.2	4	−10.7	3	−14.9	8	−10.9	4
	A3	−14.0	6	−11.2	3	−11.0	3	−11.8	4
	A4	−8.6	2	−9.4	3	−13.6	5	−9.0	3
	A5	−9.8	3	−9.3	3	−13.9	6	−10.7	4
	D2	−9.7	3	−8.1	2	−9.0	3	−13.4	5
7T9J	A1	−7.6	1	−8.3	2	−7.4	1	−11.0	3
	A2	−9.0	3	−8.9	3	−14.0	7	−11.4	4
	A3	−11.8	4	−9.9	3	−14.2	7	−10.0	3
	A4	−7.9	1	−8.7	2	−11.7	3	−7.0	1
	A5	−11.3	4	−14.8	8	−10.2	3	−11.5	4
	D2	−13.8	5	−9.7	3	−10.0	3	−8.6	2

P: protein; L: ligand; E: DS value (kcal·mol⁻¹); N: number of hydrophilic interactions.

The selected data are summarized in Table 9 (ligand–7V7N) and Table 10 (ligand–7T9J). Regarding the former, the difference in inhibitory effectiveness is rather marginal, varying between −13 and −15 kcal·mol⁻¹. In detail, the order is A2–7V7N (DS −14.9 kcal·mol⁻¹; RMSD 1.18 Å) > A3–7V7N (DS −14.0 kcal·mol⁻¹; RMSD 1.23 Å)≈A5–7V7N (DS −13.9 kcal·mol⁻¹; RMSD 1.61 Å) > A4–7V7N (DS −13.6 kcal·mol⁻¹; RMSD 1.18 Å)≈A1–7V7N (DS −13.1 kcal·mol⁻¹; RMSD 1.18 Å). Although similar, there is a slight change in order in terms of the latter, ie A5–7T9J (DS −14.8 kcal·mol⁻¹; RMSD 1.58 Å) > A3–7T9J (DS −14.2 kcal·mol⁻¹; RMSD 1.82 Å)≈A2–7T9J (DS −14.0 kcal·mol⁻¹; RMSD 1.61 Å) > A4–7T9J (DS −11.7 kcal·mol⁻¹; RMSD 1.39 Å)≈A1–7T9J (DS −11.0 kcal·mol⁻¹; RMSD 0.96 Å). The bacterium-effective inhibitors, ie A5, A2, and A3, are also predicted as most promising against the SARS-CoV-2 proteins, ie S-Delta (PDB-7V7N) and Omicron (PDB-7T9J) spike glycoproteins. The findings are comparable to our work on the wild type¹¹; unfortunately, our means still do not allow an experimental verification for either. Nevertheless, this computer-based prediction would justify further experimental attempts using the organosulfur compounds, DDS (A2), 1-allyl-3-methyltrisulfane (A3), and DTS (A5).

Table 9.

Molecular Docking Simulation Results for Ligand–7V7N Inhibitory Complexes.

Ligand–protein			Hydrogen bond						van der Waals interaction
Name	DS	RMSD	L	P		T	D	E	van der Waals interaction
A1–7V7N	−13.1	1.54	O	N	Arg A1037	H-acceptor	3.06	−3.6	Phe C1040, Thr C1025, Leu C1022, Thr B1025
			O	N	Arg A1037	H-acceptor	3.03	−1.6
			O	N	Arg B1037	H-acceptor	2.97	−1.6
			O	N	Arg C1037	H-acceptor	3.48	−1.0
			6-ring	N	Arg C1037	π-cation	3.99	−0.6
A2–7V7N	−14.9	1.18	S	N	Arg A1037	H-acceptor	3.83	−3.4	Leu B1022, Thr B1025, Phe B1040, Leu A1022, Phe A1040, Thr A1025
			S	N	Arg A1037	H-acceptor	3.54	−4.2
			S	N	Arg B1037	H-acceptor	3.90	−2.0
			S	N	Arg B1037	H-acceptor	3.51	−4.5
			S	N	Arg C1037	H-acceptor	3.77	−3.9
			S	N	Arg C1037	H-acceptor	3.67	−1.1
			S	N	Arg B1037	H-acceptor	3.42	−0.8
			S	N	Arg C1037	H-acceptor	3.31	−3.7
A3–7V7N	−14.0	1.23	S	O	Gln B952	H-donor	3.05	−0.5	Ala B956, Thr B959, Tyr B1005, Gln B1008, Thr B1004
			S	N	Arg B1012	H-acceptor	3.75	−3.6
			S	N	Arg B1012	H-acceptor	3.58	−4.3
			S	N	Gln A760	H-acceptor	3.12	−1.6
			S	N	Arg A763	H-acceptor	3.28	−1.8
			S	C	Gln B955	H-acceptor	3.77	−0.8
A4–7V7N	−13.6	1.57	O	N	Arg A1037	H-acceptor	3.30	−1.7	Leu B1022, Thr B1025, Phe B1040, Phe A1040, Thr A1025
			O	N	Arg A1037	H-acceptor	3.22	−2.7
			O	N	Arg B1037	H-acceptor	3.09	−1.1
			O	N	Arg C1037	H-acceptor	3.47	−1.3
			O	N	Arg C1037	H-acceptor	3.57	−0.6
A5–7V7N	−13.9	1.61	S	O	Asp C1039	H-donor	3.68	−0.5	Asn A1021, Ala A1024, Lys B1026, Val B1038, Leu B1022, Glu B723, Gln A782
			S	O	Thr A1025	H-donor	3.74	−0.9
			S	O	Ser A1028	H-acceptor	3.52	−1.0
			S	C	Thr A1025	H-acceptor	3.47	−0.6
			S	O	Ser A1028	H-acceptor	3.51	−0.8
			S	C	Phe B1040	H-acceptor	3.20	−0.7
D2–7V7N	−13.4	1.10	O	O	Ile B971	H-donor	3.01	−1.4	Ser B972, His A517, Leu A516, Asp B977
			O	O	Ile B971	H-donor	3.14	−0.9
			O	N	Arg B981	H-acceptor	3.38	−0.9
			N	N	Arg B981	H-acceptor	3.00	−1.4
			6-ring	C	Leu A51	π-H	3.20	−0.6

DS: docking score energy (kcal·mol⁻¹); RMSD: root-mean-square deviation (Å); L: ligand; P: protein; T: type; D: distance (Å); E: energy (kcal·mol⁻¹).

Table 10.

Molecular Docking Simulation Results for Ligand–7T9J Inhibitory Complexes.

Ligand–protein			Hydrogen bond						van der Waals interaction
Name	DS	RMSD	L	P		T	D	E	van der Waals interaction
A1–7T9J	−11.0	0.96	O	N	Arg B457	H-acceptor	2.96	−2.7	Asn A234, Glu B465
			O	N	Lys B462	H-acceptor	2.04	−1.9
			6-ring	N	Lys B462	π-cation	3.81	−1.0
A2–7T9J	−14.0	1.61	S	O	Ile B569	H-donor	3.78	−0.4	Val A47, Leu A48, Asp B568, Arg A44
			S	O	Asp B571	H-donor	3.01	−0.4
			S	C	His A49	H-acceptor	3.15	−0.9
			S	N	Ser A50	H-acceptor	3.17	−1.0
			S	N	Lys A304	H-acceptor	3.46	−7.2
			S	N	Lys A964	H-acceptor	3.69	−5.4
			S	C	Lys A964	H-acceptor	3.58	−1.4
A3–7T91	−14.2	1.82	S	O	Leu A48	H-donor	3.59	−0.5	Val A47, Arg A44, Ile B569, Asp B571, Asp B568
			S	O	Ser A50	H-donor	3.36	−1.7
			S	C	His A49	H-acceptor	3.20	−0.9
			S	C	Lys A964	H-acceptor	3.74	−0.9
			S	N	Lys A304	H-acceptor	3.35	−6.1
			S	N	Lys A964	H-acceptor	3.40	−6.6
			S	N	Ser A50	H-acceptor	3.30	−3.6
A4–7T9J	−11.7	1.39	O	O	Ser A50	H-donor	3.06	−1.9	Asp B571, Ile B569, Val A47, Ser A697, Leu A48, His A49
			O	N	Lys A304	H-acceptor	3.16	−5.8
			O	N	Lys A964	H-acceptor	3.11	−1.8
A5–7T9J	−14.8	1.58	S	N	Arg B1091	H-acceptor	3.82	−3.9	Arg A1091, Asp A1118, Asp B1118
			S	N	Arg B1091	H-acceptor	3.37	−1.2
			S	N	Arg C1091	H-acceptor	3.62	−1.2
			S	N	Arg C1091	H-acceptor	3.42	−1.9
			S	O	Thr C1117	H-acceptor	3.35	−0.7
			S	C	Asp C1118	H-acceptor	3.21	−0.8
			S	N	Arg C1091	H-acceptor	3.48	−3.4
			S	N	Arg C1091	H-acceptor	3.46	−2.5
D2–7T9J	−13.9	1.45	N	O	Asp B745	H-donor	2.93	−5.7	Thr A573, Val B976, Asn B976, Ser B975, Ser B746, Ace B855, Gly B744, Cys B743, Met B740, Lys B856, Thr A549, Pro A589, Thr A572, Leu B966, Thr A573
			C	O	Asp B745	H-donor	3.26	−0.7
			O	N	Leu B977	H-acceptor	3.32	−2.1
			O	N	Arg B1000	H-acceptor	3.26	−1.0
			O	N	Arg B1000	H-acceptor	3.05	−5.6

DS: docking score energy (kcal·mol⁻¹); RMSD: root-mean-square deviation (Å); L: ligand; P: protein; T: type; D: distance (Å); E: energy (kcal·mol⁻¹).

The corresponding visual projections (3D in-pose morphology and 2D interaction map) are presented in Figure 10 (ligand-7V7N) and Figure 11 (ligand-7T9J). Similarly, the consistency between the results retrieved from the quantum-based calculation (DFT) and that of Newton's mechanics (docking) is explicitly noticeable. The aromatic ring of A1 serves as a nucleophilic attractor while that of A4 is not; their ketone groups are still the primary sites of reactivity; the A4 methoxy group is still inactive. Regarding sulfur–protein interactions, A2, A3, and A5 can form up to eight hydrophilic bonds via these atoms. The sites are also rather large and open to their inhibiting compounds.

Figure 10.

Visual presentation and in-pose interaction maps with ligand-7V7N (A1–A5 and D2).

Figure 11.

Visual presentation and in-pose interaction maps with ligand-7T9J (A1–A5 and D2).

It is also noteworthy that quantum computations are purely based on the wavelike behavior of molecular electrons, while those of classical mechanics are established on the assumption that each molecular atom is a rigid particle and assigned with arbitrary coefficients representing its individual physical properties. Therefore, the consistency seen should be of great interest; yet, more observations collected from this approach are still needed in order to reach a solid conclusion.

QSARIS-Based Physicochemical Properties

Table 11 summarizes the QSARIS-based physicochemical properties of the investigated compounds (A1–A5) and the number of hydrogen bonds referenced from the molecular docking simulation. The parameters are subjected to druglike evaluation based on Lipinski's rule of five. All the candidates possess molecular mass significantly lower (from 121.0 to 180.8 amu) than that required by the first criterion (500 amu). No compound registers a partition coefficient higher than the threshold (logP < +5); in particular, all values are under +3. Given the docking-based output, the maximum number of hydrogen-bond acceptors is 3; that of the hydrogen-bond donors is 8. This means their hydrophilic behavior appears to satisfy the last two criteria viewed from the scope of ligand–protein interaction. In summary, from Lipinski's perspective, the physicochemical properties of A1–A5 signify their biocompatibility for drug-developing applications.

Table 11.

Physicochemical Properties of Studied Compounds A1–A5 and D1, D2.

Compound	Mass (amu)	Polarizability (Å³)	Size (Å)	Dispersion coefficients		Hydrogen bond (Q8DQF8/ P0C0C7/7V7N/7T9J)
Compound	Mass (amu)	Polarizability (Å³)	Size (Å)	logP	logS	H-donor	H-acceptor	H-π
A1	121.0	15.9	201.3	1.97	−1.93	2/2/4/2	0/0/0/0	1/0/0/0
A2	147.2	19.4	248.3	2.41	−2.17	3/5/8/5	2/2/0/2	0/0/0/0
A3	153.1	18.6	225.7	2.45	−2.67	3/5/5/5	2/2/1/2	0/0/0/0
A4	152.9	16.8	210.3	2.64	−2.31	2/1/5/2	1/2/0/1	0/1/0/0
A5	180.8	22.7	278.5	2.76	−2.91	4/7/4/8	3/3/2/0	0/0/0/0
D1	335.1	35.2	396.7	1.43	−2.83	3/1	0/1	1/2
D2	330.5	28.7	338.9	−1.59	−1.21	2/3	2/2	1/0

ADMET-Based Pharmacokinetics and Pharmacology

Table 12 summarizes the absorption, distribution, metabolism, excretion, and toxicity (ADMET)-based output for the candidates (A1–A5), including chemical absorption, distribution, metabolism, excretion, and toxicity. This model is built using statistical regression based on experimental measurements; hence, the results should be considered a preliminary probability-based assessment.

Table 12.

ADMET-Based Pharmacokinetics and Pharmacology of the Studied Compounds A1–A5 and D1, D2.

Property	A1	A2	A3	A4	A5	D1	D2	Unit
Absorption
Water solubility	−1.783	−3.181	−2.885	−0.851	−3.791	−2.841	−2.403	Numeric (log mol·L⁻¹)
Caco-2 permeability	1.364	1.255	1.248	1.231	1.258	0.42	−0.413	Numeric (log Papp 10⁻⁶ cm·s⁻¹)
Intestinal absorption (human)	100	96.881	95.524	86.91	95.162	68.928	56.571	Numeric (% absorbed)
Skin permeability	−1.724	−1.323	−1.432	−2.773	−1.351	−3.109	−3.193	Numeric (log Kp)
P-glycoprotein substrate	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Categorical (yes/no)
P-glycoprotein I inhibitor	No	No	No	No	No	No	No	Categorical (yes/no)
P-glycoprotein II inhibitor	No	No	No	No	No	No	No	Categorical (yes/no)
Distribution
VDss (human)	−0.017	0.298	0.197	−0.394	0.314	−0.582	−1.032	Numeric (log L·kg⁻¹)
Fraction unbound (human)	0.411	0.458	0.495	0.46	0.429	0.417	0.629	Numeric (log L·kg⁻¹)
BBB permeability	0.01	0.704	0.478	−0.194	0.7	−0.74	−1.45	Numeric (log BB)
CNS permeability	−1.304	−1.836	−2.084	−1.775	−1.935	−2.944	−4.061	Numeric (log PS)
Metabolism
CYP2D6 substrate	No	No	No	No	No	No	No	Categorical (yes/no)
CYP3A4 substrate	No	No	No	No	No	Yes	No	Categorical (yes/no)
CYP1A2 inhibitor	Yes	Yes	Yes	Yes	Yes	No	No	Categorical (yes/no)
CYP2C19 inhibitor	No	No	No	No	No	No	No	Categorical (yes/no)
CYP2C9 inhibitor	No	No	No	No	No	No	No	Categorical (yes/no)
CYP2D6 inhibitor	No	No	No	No	No	No	No	Categorical (yes/no)
CYP3A4 inhibitor	No	No	No	No	No	No	No	Categorical (yes/no)
Excretion
Total clearance	0.342	0.554	0.354	0.636	0.453	0.198	0.175	Numeric (log mL·min⁻¹·kg⁻¹)
Renal OCT2 substrate	No	No	No	No	No	No	No	Categorical (yes/no)
Toxicity
AMES toxicity	No	No	No	No	No	Yes	Yes	Categorical (yes/no)
Max. tolerated dose (human)	1.754	1.502	1.421	1.625	1.437	0.373	0.642	Numeric (log mg·kg⁻¹·day⁻¹)
hERG I inhibitor	No	No	No	No	No	No	No	Categorical (yes/no)
hERG II inhibitor	No	No	No	No	No	No	No	Categorical (yes/no)
Oral rat acute toxicity (LD50)	1.898	2.597	2.975	1.957	2.917	2.354	2.133	Numeric (mol/kg)
Oral rat chronic toxicity (LOAEL)	1.788	1.691	1.58	2.665	1.694	1.674	1.265	Numeric (log mg·kg⁻¹_bw·day⁻¹)
Hepatotoxicity	No	No	No	No	No	Yes	Yes	Categorical (yes/no)
Skin sensitization	Yes	Yes	Yes	No	Yes	No	No	Categorical (yes/no)
T pyriformis toxicity	0.001	0.711	0.658	−0.066	1.378	0.683	0.181	Numeric (log µg·L⁻¹)
Minnow toxicity	1.373	0.693	0.938	1.94	0.358	1.883	3.627	Numeric (log mM)

First, a compound is considered to have poor intestinal absorption if less than 30%. This means that all the candidates possess significant effectiveness of intestine absorbance regarding any compound by over 85%; especially, the model predicts complete absorption (ie 100%) regarding A1. Nevertheless, the relatively low Caco-2 permeability values, under 1.5 × 10⁻⁶ cm·s⁻¹, suggest that the overall intestinal processes might meet certain resistance, eg through transcellular transport, paracellular transport, and efflux-active transport. The model also predicts that all the compounds are likely to be a substrate of P-glycoprotein, thus reducing the extrusion of the toxins and xenobiotics out of cells. Secondly, A4 is more likely to accumulate in tissue (log VDss < −0.15); otherwise, A1–A3 and A5 would see a balance in plasma–tissue distribution. The organosulfur compounds (A2, A3, A5) are considered to cross the blood–brain barrier readily (logBB > 0.3) and are also able to penetrate the central nervous system (log PS < −3). Thirdly, the potential effects on the inhibitors and substrates relating to the cytochrome P450 family should be insignificant for all the candidates; this means they can be effectively oxidized in the liver. Predicted with a total clearance of ca 0.1–0.5 log mL·min⁻¹·kg⁻¹, all the compounds can be excreted by Organic Cation Transporter 2. Given toxicity, all the compounds are expected to be safe: no mutagenic potentials (AMES toxicity), no potential for fatal ventricular arrhythmia, and no potential for hepatotoxicity. They are considered toxic to T pyriformis (pIGC50 > −0.5 log μg·L⁻¹), a protozoa bacteria, yet safe to fathead minnows (LC₅₀> −0.3), a species of temperate freshwater fish. To our knowledge, there is no explicit report on the adverse effects regarding all these natural compounds, which is also consistent with this in silico retrieval; meanwhile, benzylpenicillin (D1) was associated with anaphylaxis in horses,²² and molnupiravir (D2) was found to have mutagenic potential from in vitro and in vivo evidence.²³ Overall, this preliminary assessment of the garlic extract composition (A1–A5) shows the need to perform appropriate pharmacokinetics and pharmacology for drug-developing efforts.

Conclusions

This study, for the first time, includes both experimental and computational approaches to examine the potentiality of garlic extract against S pyogenes and S pneumoniae, extended to prediction against SARS-CoV-2 Delta and Omicron Spike proteins. GC-MS characterized five bioactive compounds: benzeneacetaldehyde (A1; 5.65%), DDS (A2; 19.27%), 1-allyl-3-methyltrisulfane (A3; 21.30%), methyl salicylate (A4; 7.49%), and DTS (A5; 39.26%). Bioassays produced inhibition zones of 8 ± 2 mm, corresponding to MIC values of 64 mg·mL⁻¹ (S pneumoniae) and 128 mg·mL⁻¹ (S pyogenes). Regarding DFT-based calculations, geometry optimization found no abnormal structural constraints; molecular electronic structures suggest a cation-attraction tendency of the A1 aromatic ring and an inert chemical activation of the A4 methoxy group; electronic potential maps predict ketone groups and organosulfur groups as the center of chemical reactivity. Docking-based simulations specify the organosulfur components (A2, A3, A5) as the promising bioactive inhibitors against protein structures (DS < −13 kcal·mol⁻¹), including those of S pneumoniae/S pyogenes luxS proteins and SARS-CoV-2 Delta/Omicron spike proteins; meanwhile, in-depth analysis of the ligand–protein intermolecular interactions confirms the quantum-based unusual retrieval from DFT. QSARIS and ADMET preliminarily studies related the compounds with bio- and pharma-favored suitability. From a theoretical standpoint, the cross-platform analysis highly encourages the use of garlic as a synergic therapeutic food for SARS-CoV-2-related coinfections, in general, and S pneumoniae/S pyogenes, in particular.

Material and Methods

Materials

Plant Materials

Fresh peeled garlic was milled, then incubated at 4 °C for 10 min; afterward, ethanol (70%) was used for a further 15-min soaking, with a ratio of material: solvent of 1:10 g/mL. The mixture was then extracted with the assistance of ultrasonic radiation (frequency 37 kHz; temperature < 30 °C; duration 30 min). The obtained extract was homogenized in a flask (ca 1–2 min), then diluted with the original solvent to appropriate material/solvent ratios. The supernatant was collected after centrifugation (speed 3500 rpm; duration 10 min; temperature 4 °C) and subjected to a secondary extraction with similar experimental settings. These alcoholic extracts were collected and low-temperature evaporated (below 30 °C; atmospheric pressure) to obtain the final extract. The volatile part isolated from the total extract was further subjected to compositional analysis (GC-MS) and antibacterial tests.

Bioassay Materials

Microorganisms S pyogenes (ATCC 19615) and S pneumoniae (ATCC 49619) were supplied by the Microbiology and Parasitology Department, Faculty of Pharmacy, Nguyen Tat Thanh University.

Chemicals and reagents include ethanol (OPC Pharmaceutical Company); n-hexane (VN-Chemsol, Co., Ltd); dimethyl sulfoxide (DMSO), Mueller–Hinton broth (MHB), and Mueller–Hinton agar (MHA) (Merck Co., Ltd, Germany); and penicillin G (Nam Khoa Biotek Co., Ltd). Equipment includes a test tube, micropipette, Eppendorf, Petri dish, 96-well microplate, water bath, and biosafety cabinet class II. All other chemical reagents and solvents, purchased from Sigma–Aldrich (USA), were of analytical grade and used without further purification.

Experimental Methods

Chemical Characterization

The garlic extract was analyzed by GC-MS, with an Agilent DB-5MS column (30 m × 0.25 mm × 0.25 µm). Helium was used as the carrier gas. In a typical procedure, the sample (1.0 µL) was injected with a split ratio of 20:1 and at a temperature of 250 °C. The GC temperature configuration was (i) initiation at 80 °C and (ii) linear increase to 300 °C (rate 20 °C min⁻¹). The MS scanning was set from 29 to 650 amu for mass analysis. The compounds in the garlic extract were identified by reference to the NIST-17 database.

Antimicrobial assay

Microorganisms

The antibacterial activity of garlic extract on S pyogenes (ATCC 19615) and S pneumoniae (ATCC 49619) was evaluated by well diffusion assay and broth dilution methods. The procedures followed the standards issued by the Clinical and Laboratory Standards Institute, with modifications (CLSI supplement M100, 2019).²⁴ The bacteria were cultured in MHB (Merck, Germany) medium (temperature 37 °C; duration 5–8 h) until the 0.5 McFarland turbidity (approximately 5.10⁸ CFU·mL⁻¹). This provided the standardized bacterial suspension containing a concentration of 5.10⁸ CFU·mL⁻¹.²⁴

Well Diffusion Test

The standardized bacterial suspension was diluted to 1:10 in PBS before (100 µL used) being swabbed evenly over the surface of blood agar; the medium was allowed to dry (ca 5 min). Circular holes were punched in the agar for the diffusion of the antibacterial samples. Garlic extract (volume 70 µL; diluted 1:5 w/v in DMSO 10%) was injected into a hole for each experiment. Penicillin G (20 U) and DMSO (10% w/v) were used as the positive and negative controls, respectively. Afterward, the dishes were incubated (duration 24 h; temperature 37 °C) before the inhibition zone diameter (IZD) was measured.²⁵ The procedure was in triplicate in order for the mean IZD to be determined^26,27 and interpreted as 0 mm for no sensitivity or resistance, 1–6 mm for low sensitivity, 7–14 mm for moderate sensitivity, and >14 for high sensitivity.

Vapor Phase Test

The in vitro vapor phase test was based on the method described by Kloucek et al with some adjustments.²⁸ The test system was carried out in three-section Petri dishes (90 mm diameter), consisting of a blood agar medium. The standardized bacterial suspension (100 µL) was swabbed evenly over the surface of the blood agar with two sections of the plate and allowed to dry (duration ca 5 min); the two different strains were spread into two different sections; the third section was left uninoculated as the negative control. Garlic extract (210 μL) was distributed on the surface of a sterile filter paper disc (thickness 0.18 mm; diameter 84 mm); the paper disc was placed on the separating wall of the plate. This means the distance between the disc and the inoculated agar surface was ca 2 mm. The plates were hermetically closed with parafilm adhesive tape to avoid evaporation and incubated (temperature 37 °C; duration 16–18 h). The visible growth of the test bacteria (if observed) was captured. All tests were carried out in a class II microbiological safety cabinet and in triplicate.

Agar Dilution Assay

Agar dilution involves the incorporation of different concentrations of the extract into an agar medium followed by a standardized microbial inoculum onto the surface of the agar plate.²⁹ In this study, the method was utilized for the evaluation of garlic extract MIC, ie the lowest concentration of the drug that inhibited the organism's growth. The assay followed a twofold dilution of the garlic extracts, ie 1, 2, 4, 8, 16, 32, 64, 128, and 256 mg·mL⁻¹, in molten blood agar media; the standardized bacterial suspension was under 50-fold dilution (concentration ca 10⁷ CFU·mL⁻¹) before 2 mL was applied onto the agar surface (approximately 10⁴ CFU per spot). One agar plate was used as a control seeded without garlic extract. After incubation (temperature 37 °C; duration 16–18 h), the plates were observed for visual comparison of microbial growth. Regarding this experimental scheme, MICs were determined as the lowest concentration of garlic extract that stopped the growth of bacteria.

Computational Methods

Quantum Chemical Calculation

Molecular quantum properties of the studied compounds (A1–A5) and their geometry were optimized based on DFT using Gaussian 09 without symmetry constraints³⁰ with functional M052X and 6–311++G(d,p) basis set.³¹ Their potential energy surface (PES) was performed at the same level of theory. The stationary points on the PES were confirmed as energy minima by frequency calculations which were obtained at the same level for each species to achieve the zero-point vibrational energy (ZPE) corrections. A larger basis set def2-TZVPP³² was used for frozen-core approximation of non-valence-shell electrons, resulting in single-point energies. Resolution-of-identity (RI) approximation was applied for each optimization run. Frontier orbital analysis at the level of theory M052X/def2-TZVPP²⁹ was implemented by NBO 5.1. At the same level of theory as optimization, molecular electrostatic potential (MEP) in the molecule was calculated and visualized by GaussView 05 software. The HOMO energy, E_HOMO, can be assigned to the tendency of intermolecular electron donation; in contrast, the value E_LUMO (for LUMO) might represent the electron-accepting capability; their difference, ie energy gap ΔE = E_LUMO − E_HOMO, can suggest molecular electronic excitability of the host molecule. Koopman's theorem³³ was used for the calculation of ionization potential (I = −E_HOMO) and electron affinity (A = −E_LUMO), which in turn are useful in evaluating the reactivity of a molecule via electronegativity (χ = (I + A)/2), hardness (η = (I − A)/2), softness (S = 1/η), and other molecular orbital parameters. Condensed Fukui functions were computed using the Mulliken atomic charge given by the population analysis method.³⁴

Molecular Docking Simulation

A typical procedure for molecular docking simulation (by MOE 2015.10³⁵) follows four steps,^36–38 providing predictions on ligand–protein complex structures.

Predocking preparation: Sources for simulation input: crystal structures of LuxS proteins of S pneumoniae (UniProtKB: Q8DQF8 (Q8DQF8_STRR6)) and of S pyogenes (UniProtKB: P0C0C7 (LUXS_STRPY)) were referenced from UniProt under the respective entry ID; those of spike proteins of SARS-CoV-2 Delta (PDB-7V7N) and Omicron (PDB-7T9J) variants were downloaded from Protein Data Bank; chemical formulae of ligands were from experimental findings by GC-MS in this work; structural formulas of benzylpenicillin (penicillin G) (D1) and molnupiravir (D2) were used as the docking references. Configuration for determination of protein active areas: amino acids within 4.5 Å to ligands were defined potentially active; water or small molecules (detected by the experimental work) were removed; tether–receptor with the strength of 5000; refine of 0.0001 kcal·mol⁻¹·Å⁻¹; output format in *.pdb. Configuration for optimization of ligand structures: Conj Grad for minimal energy, termination for energy change = 0.0001 kcal·mol⁻¹; max interactions = 1000; Gasteiger–Huckel method for empirical charge-assigning.

Docking investigation: Configuration for docking simulation: poses retaining for intermolecular interaction probing = 10; maximum solutions per iteration = 1000; maximum solutions per fragmentation = 200; output format for ligand–protein inhibitory structures in *.sdf.

Redocking iteration: The reliability of the docking protocol was validated by redocking iterations and recalculating RMSD. To implement, the ligand was removed from its inhibitory complexes and redocked to the protein structures. To validate, the accuracy of the docking protocol is justified if RMSD values (of docked and redocked conformations) are all under 2 Å.

Postdocking analysis: In principle, the inhibitory effectiveness of a ligand–protein inhibitory system was primarily indicated by DS energy, the pseudo-Gibbs free energy. The value is contributed by attractive forces (hydrophilic binding and hydrophobic interaction) and geometrical complementarity (RMSD, ie the average between backbone atoms).

QSARIS-Based Analysis

Druglikeness properties of phytochemicals were under in silico analysis by the QSARIS system using the Gasteiger–Marsili method;³⁹ as output, the main parameters were molecular mass (Da), polarizability (Å³) and volume or size (Å), and dispersion coefficients (logP: partition coefficient, and logS: aqueous solubility). Afterwards, they were subjected for assessment of orally pharmacological compatibility based on Lipinski's rule of five.⁴⁰ Accordingly, a good membrane-permeable candidate should satisfy the following criteria: (1) Molecular mass < 500 Da, (2) no more than five groups for hydrogen bonds, (3) no more than 10 groups receiving hydrogen bonds, and (4) the value of logP is less than +5 (logP < 5).^41,42

ADMET-Based Analysis

ADMET properties were obtained from a web-based regressive model developed and maintained by the Molecular Modeling Group, Swiss Institute of Bioinformatics, ie SwissADME (http://www.swissadme.ch/; May 17, 2023). The theoretical interpretations of output pharmacokinetic parameters were described by Pires et al.⁴³

Supplemental Material

sj-docx-1-npx-10.1177_1934578X231194440 - Supplemental material for Potentiality of Organosulfur Compounds Against SARS-CoV-2-Coinfected Bacteria Streptococcus pyogenes and S pneumoniae: A Cross-Platform Analysis from Computational Chemistry

Supplemental material, sj-docx-1-npx-10.1177_1934578X231194440 for Potentiality of Organosulfur Compounds Against SARS-CoV-2-Coinfected Bacteria Streptococcus pyogenes and S pneumoniae: A Cross-Platform Analysis from Computational Chemistry by Nguyen Minh Thai, Thanh Q. Bui, Phan Tu Quy, Nguyen Thi Thanh Hai, Dao-Cuong To, Duong Tuan Quang, Nguyen Quang Co, Nguyen Thanh Triet, Nguyen Thi Ai Thuan, Nguyen Thanh To Nhi, Vo Mong Tham and Nguyen Thi Ai Nhung in Natural Product Communications

Footnotes

Authors’ Contribution

All authors contributed to the study conception and design. All authors read and approved the final manuscript. Nguyen Minh Thai worked on methodology (experimental), formal analysis (experimental), and writing—original draft (experimental). Thanh Q. Bui worked on conceptualization (computational), methodology (computational), and writing—review and editing. Phan Tu Quy worked on investigation (computational) and software (computational). Nguyen Thi Thanh Hai worked on investigation (extraction) and visualization. Dao-Cuong To worked on conceptualization (experimental), formal analysis (experimental), and writing—original draft (experimental). Duong Tuan Quang worked on conceptualization (experimental), formal analysis (experimental), writing—review and editing (experimental), and supervision (experimental). Nguyen Quang Co worked on data curation (computational) and writing—original draft (computational). Nguyen Thanh Triet worked on resources (protein structures) and writing—original draft (computational). Nguyen Thi Ai Thuan worked on investigation (factionalization) and visualization. Nguyen Thanh To Nhi worked on investigation (bioassay) and visualization. Vo Mong Tham worked on investigation (bioassay) and visualization. Nguyen Thi Ai Nhung worked on conceptualization (computational), investigation (computational), data curation (computational), writing—review and editing (experimental), and supervision (computational).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethical Approval

Ethical approval is not applicable to this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Vietnam National Foundation for Science and Technology Development (NAFOSTED) (Grant number 02/2020/ĐX) and Hue University (Grant number DHH2022-01-198; Grant number NCM.DHH.2020.04).

ORCID iDs

Dao-Cuong To

Nguyen Thanh Triet

Nguyen Thi Ai Nhung

Statement of Human and Animal Rights

This article does not contain any studies with human or animal subjects.

Statement of Informed Consent

There are no human subjects in this article, and informed consent is not applicable.

Supplemental Material

Supplemental material for this article is available online.

References

Zhu

, et al. Co-infection with respiratory pathogens among COVID-2019 cases. Virus Res. 2020;285:198005.

Pal

Przydzial

Chika-Nwosuh

Shah

Patel

Madan

. Streptococcus pneumoniae coinfection in COVID-19: a series of three cases. Case Rep Pulmonol. 2020;2020:8849068.

Shulman

Bisno

Clegg

, et al. Clinical practice guideline for the diagnosis and management of group A streptococcal pharyngitis: 2012 update by the Infectious Diseases Society of America. Clin Infect Dis. 2012;55(10):e86-e102.

Avire

Whiley

Ross

. A review of Streptococcus pyogenes: public health risk factors, prevention and control. Pathogens. 2021;10(2):248.

Lyon

Madden

Levin

Stein

Caparon

. Mutation of luxS affects growth and virulence factor expression in Streptococcus pyogenes. Mol Microbiol. 2001;42(1):145-157.

Kang

Caparon

Cho

. Virulence gene regulation by CvfA, a putative RNase: the CvfA-enolase complex in Streptococcus pyogenes links nutritional stress, growth-phase control, and virulence gene expression. Infect Immun. 2010;78(6):2754-2767.

Tuomanen

Austrian

Masure

. Pathogenesis of pneumococcal infection. N Engl J Med. 1995;332(19):1280-1284.

Deshpande

Srivastava

Bendet

, et al. Antibacterial and sterilizing effect of benzylpenicillin in tuberculosis. Antimicrob Agents Chemother. 2018;62(2):e02232-17.

Kumar

Malviya

Sharma

. Corona virus: a review of COVID-19. EJMO. 2020;4(1):8-25.

10.

Yingngam

Navabhatra

Brantner

, et al. One-pot extraction and enrichment of diallyl trisulfide in garlic oil using an eco-friendly solvent-free microwave extraction method. Sustain Chem Pharm. 2022;27:100655.

11.

Thuy

BTP

TTA

Hai

NTT

, et al. Investigation into SARS-CoV-2 resistance of compounds in garlic essential oil. ACS Omega. 2020;5(14):8312-8320.

12.

Molina-Calle

Priego-Capote

de Castro

MDL

. HS–GC/MS volatile profile of different varieties of garlic and their behavior under heating. Anal Bioanal Chem. 2016;408(14):3843-3852.

13.

Casella

Leonardi

Melai

Fratini

Pistelli

. The role of diallyl sulfides and dipropyl sulfides in the in vitro antimicrobial activity of the essential oil of garlic, Allium sativum L., and leek, Allium porrum L. Phyther Res. 2013;27(3):380-383.

14.

Rouf

Uddin

Sarker

, et al. Antiviral potential of garlic (Allium sativum) and its organosulfur compounds: a systematic update of pre-clinical and clinical data. Trends Food Sci Technol. 2020;104:219-234.

15.

Golestannejad

Mohammadi

Motamedi

, et al. Chemical composition and antibacterial activity of some herbal essential oils against Streptococcus mutans. Futur Nat Prod. 2016;2(1):1-8.

16.

Rad

Ardjmand

Esfahani

Khodashenas

. DFT calculations towards the geometry optimization, electronic structure, infrared spectroscopy and UV–vis analyses of Favipiravir adsorption on the first-row transition metals doped fullerenes; a new strategy for COVID-19 therapy. Spectrochim Acta Part A Mol Biomol Spectrosc. 2021;247:119082.

17.

Rosenberg

. Electrical conductivity of proteins. Nature. 1962;193(4813):364-365.

18.

Kharkyanen

Petrov

Ukrainskii

. Donor-acceptor model of electron transfer through proteins. J Theor Biol. 1978;73(1):29-50.

19.

Cordes

Giese

. Electron transfer in peptides and proteins. Chem Soc Rev. 2009;38(4):892-901.

20.

Thao

TTP

Bui

Hai

NTT

, et al. Newly synthesised oxime and lactone derivatives from Dipterocarpus alatus dipterocarpol as anti-diabetic inhibitors: experimental bioassay-based evidence and theoretical computation-based prediction. RSC Adv. 2021;11(57):35765-35782.

21.

Thao

TTP

Bui

Quy

, et al. Isolation, semi-synthesis, docking-based prediction, and bioassay-based activity of Dolichandrone spathacea iridoids: new catalpol derivatives as glucosidase inhibitors. RSC Adv. 2021;11(20):11959-11975.

22.

English

Roberts

. Adverse reactions to antimicrobial agents in the horse. Vet Res Commun. 1983;7(1):207-210.

23.

Painter

Natchus

Cohen

Holman

Painter

. Developing a direct acting, orally available antiviral agent in a pandemic: the evolution of molnupiravir as a potential treatment for COVID-19. Curr Opin Virol. 2021;50:17-22.

24.

Clinical and Laboratory Standards Institute. In: Weinstein

Patel

, eds. Performance standards for antimicrobial susceptibility testing. 29th ed. Vol 8. Clinical and Laboratory Standards Institute; 2019. http://www.emeraldinsight.com/doi/10.1108/08876049410065598

25.

Castilho

Caleffi-Ferracioli

Canezin

Siqueira

VLD

de Lima Scodro

Cardoso

. Detection of drug susceptibility in rapidly growing mycobacteria by resazurin broth microdilution assay. J Microbiol Methods. 2015;111:119-121.

26.

Mehmood

Dar

Ali

, et al. In vitro assessment of antioxidant, antibacterial and phytochemical analysis of peel of Citrus sinensis. Pak J Pharm Sci. 2015;28(1):231-239.

27.

Ghous

Akhtar

Andleeb

, et al. Evaluation of α-glucosidase inhibition, antioxidant and antibacterial effects of Gymnema sylvestre R. Br. Bangladesh J Bot. 2021;50(1):61-68.

28.

Kloucek

Smid

Frankova

Kokoska

Valterova

Pavela

. Fast screening method for assessment of antimicrobial activity of essential oils in vapor phase. Food Res Int. 2012;47(2):161-165.

29.

Reed

Weinstock

Weinhold

. Natural population analysis. J Chem Phys. 1985;83(2):735-746.

30.

Frisch

Schlegel

Scuseria

, et al. Gaussian 09. Gaussian Inc. Published online 2009.

31.

Markovi

Dimitri

. Mechanistic study of the structure – activity relationship for the free radical scavenging activity of baicalein. J Mol Model. 2011;17:2575-2584.

32.

Weigend

Ahlrichs

. Balanced basis sets of split valence, triple zeta valence and quadruple zeta valence quality for H to Rn: design and assessment of accuracy. Phys Chem Chem Phys. 2005;7(18):3297-3305.

33.

Koopmans

. Über die Zuordnung von Wellenfunktionen und Eigenwerten zu den Einzelnen Elektronen Eines Atoms. Physica. 1934;1(1-6):104-113.

34.

Ayers

Yang

Bartolotti

. Fukui function. In: Chattaraj

, ed. Chemical reactivity theory: a density functional view. CRC Press; 2004:255-267.

35.

Molecular Operating Environment (MOE), 2015.02 Chemical Computing Group ULC, 1010 Sherbooke St. West, Suite #910, Montreal, QC, Canada, H3A 2R7, 2015.

36.

Tarasova

Poroikov

Veselovsky

. Molecular docking studies of HIV-1 resistance to reverse transcriptase inhibitors: mini-review. Molecules. 2018;23(5):11-13.

37.

Thai

Tran

NVK

Nguyen

TTH

Tran

. Computational assay of Zanamivir binding affinity with original and mutant influenza neuraminidase 9 using molecular docking. J Theor Biol. 2015;385:31-39.

38.

Ngo

Tran

Thai

. Computational predictive models for P-glycoprotein inhibition of in-house chalcone derivatives and drug-bank compounds. Mol Divers. 2016;20(4):945-961.

39.

Gasteiger

Marsili

. Iterative partial equalization of orbital electronegativity-a rapid access to atomic charges. Tetrahedron. 1980;36(22):3219-3228.

40.

Lipinski

Lombardo

Dominy

Feeney

. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv Drug Deliv Rev. 1997;23:3-25.

41.

Ahsan

Samy

Khalilullah

, et al. Molecular properties prediction and synthesis of novel 1,3,4-oxadiazole analogues as potent antimicrobial and antitubercular agents. Bioorganic Med Chem Lett. 2011;21(24):7246-7250.

42.

Mazumdera

Chakraborty

Sena

Vadrab

Dec

Ravi

. Synthesis and biological evaluation of some novel quinoxalinyl triazole derivatives. Der Pharma Chem. 2009;1(2):188-198.

43.

Pires

DEV

Blundell

Ascher

. pkCSM: predicting small-molecule pharmacokinetic and toxicity properties using graph-based signatures. J Med Chem. 2015;58(9):4066-4072.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

2.63 MB