Tetracycline as an inhibitor to the coronavirus SARS-CoV-2
TTetracycline as an inhibitor to thecoronavirus SARS-CoV-2
Tom Y. Zhao a1 , Neelesh A. Patankar a1 a Department of Mechanical Engineering, Northwestern University, Evanston, IL 60208
August 14, 2020
Abstract
The coronavirus SARS-CoV-2 remains an extant threat against public health on a global scale. Cell infectionbegins when the spike protein of SARS-CoV-2 binds with the cell receptor, angiotensin-converting enzyme 2(ACE2). Here, we address the role of Tetracycline as an inhibitor for the receptor-binding domain (RBD)of the spike protein. Targeted molecular investigation show that Tetracycline binds more favorably to theRBD (-9.40 kcal/mol) compared to Chloroquine (-6.31 kcal/mol) or Doxycycline (-8.08 kcal/mol) and inhibitsattachment to ACE2 to a greater degree (binding efficiency of 2.98 kcalmol · nm for Tetracycline-RBD, 5.59 kcalmol · nm for Chloroquine-RBD, 5.16 kcalmol · nm for Doxycycline-RBD). Stronger Tetracycline inhibition isverified with nonequilibrium PMF calculations, for which the Tetracycline-RBD complex exhibits the lowestfree energy profile along the dissociation pathway from ACE2. Tetracycline appears to target viral residuesthat are usually involved in significant hydrogen bonding with ACE2; this inhibition of cellular infectioncomplements the anti-inflammatory and cytokine suppressing capability of Tetracycline, and may furtherreduce the duration of ICU stays and mechanical ventilation induced by the coronavirus SARS-CoV-2. I. I ntroduction
The extreme urgency for therapeutics against theacute respiratory syndrome coronavirus 2 (SARS-CoV-2) drives the review of existing drugs for theirability to inhibit the function of this virus .Tetracycline has been proposed as a strong can-didate against SARS-CoV-2 due to its lipophilicnature, anti-inflammatory response, as well as itsability to chelate zinc species on matrix metallopro-teinases (MMPs). Tetracycline class antibiotics havealso been shown to be effective in reducing the dura-tion of ventilatory support and ICU stay from acuterespiratory distress syndrome , and Doxycyclinehas been suggested to be an important componentin combination therapy for its anti-viral properties .Tetracycline as well as a broad band of related an-tibiotics have been approved by the FDA .In this work, we quantify the performance of To whom correspondence may be addressed. Email:[email protected] ; [email protected]
Tetracycline in inhibiting the binding of the SARS-CoV 2 spike protein to ACE2. Tetracycline is foundto bind more favorably to the receptor binding do-main (RBD) of the spike protein compared to Doxy-cycline or Chloroquine, which was included in thisstudy as a baseline. The Tetracycline-RBD complexalso displays lower binding efficiency to the humancell receptor ACE2.
II. M ethods
The SARS-CoV 2 RBD, ACE2, Tetracycline,and Chloroquine molecular structures were ob-tained from RCSB PDB (6M0J, 2UXO, 4V2O,2XRL) . Missing hydrogen atoms were ap-pended, after which structural preparation andmolecular docking with full ligand and proteinbackbone flexibility were carried out using theRosetta suite . The resulting complexes wereinspected manually, after which the binding affini-ties of the best-scoring complexes were gauged us-1 a r X i v : . [ q - b i o . B M ] A ug ng MM/PBSA calculations after 100 ns equilibriummolecular dynamics simulations . The poten-tials of mean force (PMF) along the dissociationpathway of these RBD complexes from ACE2 werefound in LAMMPS using steered molecular dy-namics after parameterization with CHARMM .Jarzynski’s equality was employed to calculate thefree energy profile for each RBD complex from 10statistically independent trajectories . III. R esults and D iscussion Tetracycline exhibits higher binding affinity to theRBD in both blind and site-specific docking (-9.40 kcal/mol) compared to Doxycycline (-8.08kcal/mol) or Chloroquine (-6.63 vs 6.31 kcal/mol)as delineated in Table 1. The amino acid residuesof the RBD involved in hydrogen bonding withthe Tetracycline molecule are Tyr 449, Asn 501, Gly496, and Tyr 505 (Fig. 1), which have been shownto be crucial for the SARS-CoV 2 RBD in bindingto ACE2 for cellular access . These four residuescomprise major hot spots that form persistent hy-drogen bonds with ACE2. Meanwhile, the aminoacids of RBD that interact with Chloroquine in thesite-specific configuration are Lys 356, Arg 454, Arg466 and Arg 355, of which none are involved inextended hydrogen bonding with ACE2.Tetracyline appears to bind preferably to polaror slightly lipophilic RBD residues, which com-prise the majority of amino acids that form persis-tent hydrogen bonds with ACE2 . Other tetracy-cline derivatives as Doxycycline or Minocycline areknown to be more lipophilic and may thereforeprefer nonpolar residues that are often buried be-neath the solvent accessible surface area of the spikeprotein. Indeed, the RBD residues that have highestbinding affinity to Doxycycline are Tyr 449, Gly 447,Val 445, Gly 496, of which only two overlap withRBD amino acids that engage in extended hydro-gen bonding with ACE2. On the other end of thespectrum, Chloroquine targets clusters of chargedresidues on the RBD that do not actively participatein hydrogen bonding with the cell receptor ACE2.The binding efficiency (magnitude of bindingenergy normalized by contact interface area) ofthe SARS-CoV2 RBD-ACE2 complex was foundto be 7.58 kcal/(mol · nm ). In the presence of the Figure 1:
Interaction map of amino acid residues of the SARS-CoV 2 receptor binding domain (RBD) that have thehighest binding affinity with Tetracycline. The Tyr449, Asn 501, Gly 496 and Tyr 505 residues havealso been shown to form persistent hydrogen bondsin maintaining the RBD-ACE2 complex . Inhibitor ∆ G bind (kcal/mol) ∆ G bind (kcal/mol)to RBD This work Prior literature
Tetracycline − − − Table 1:
The binding free energy of small-molecule inhibitorsto the SARS-CoV2 receptor binding domain (RBD).Tetracycline binds preferably to the RBD. yr 505 Tyr 489 Gln 498 Thr 500 Gln 493 Asn 48700.10.20.30.40.50.60.7 N o r m a li z ed H - bond li f e t i m e Chloroquine-RBDDoxycycline-RBDTetracycline-RBD
Figure 2:
The hydrogen bonding lifetimes of binding siteresidues of the inhibited RBD with ACE2 sustainedin ns of simulation time, normalized by hydro-gen bonding lifetimes in the uninhibited RBD-ACE2complex. protein-ligand complex Tetracycline-RBD, the bind-ing efficiency with ACE2 (2.98 kcal/(mol · nm )) issignificantly lower than that for Chloroquine-RBD(5.59 kcal/(mol · nm )) and Doxycycline-RBD (5.16kcal/(mol · nm )) as displayed in Table 2. A surveyof hydrogen bonding lifetimes between the impor-tant binding site residues in the RBD with ACE2 shows that the Tetracycline inhibited RBD exhibitsthe least hydrogen bonding activity with ACE2 (Fig.2). This suggests that not only does Tetracyclinebind more favorably to the receptor binding domainof the spike protein, it also inhibits the binding ofthe RBD to ACE2 to a greater degree.To verify this statement, steered molecular dy-namics simulations were carried out to find thepotential of mean force (PMF) along a singular dis-sociation pathway for the inhibited and uninhibitedRBD-ACE2 complexes. Figure 3 shows that thePMF for unbinding of the Tetracycline-RBD com-plex from ACE2 was lowest of the three structurestested, which is in agreement with the binding effi-ciencies found from equilibrium simulations. Thisdisruption of the RBD-ACE2 interface may there-fore inhibit the signaling cascade initiated duringbinding of the viral spike protein. Complex Binding efficiencyto ACE2 (kcal/(mol · nm ))RBD 7.58Chloroquine-RBD 5.59Doxycycline-RBD 5.16Tetracycline-RBD 2.98 Table 2:
The binding efficiency (magnitude of binding en-ergy normalized by contact interface area) of thespike protein RBD as well as the Tetracycline-RBD,Doxycycline-RBD and Chloroquine-RBD complexesto the human cell receptor ACE2. Binding efficiencyis lowest for the Tetracycline-RBD complex, indicat-ing that Tetracycline is a more effective inhibitor. -2 0 2 4 6 8 10 12 Distance (Å) -1001020304050607080 P M F ( kc a l / m o l ) RBDChloroquine-RBDDoxycycline-RBDTetracycline-RBD
Figure 3:
The potential of mean force (PMF) as a functionof the distance between the centers of masses of thespike protein RBD complexes and the cell receptorACE2. The Tetracycline-RBD complex exhibits thelowest free energy profile along the dissociation path-way. V. C onclusion
The tetracycline class of antibiotics, including Tetra-cycline, Oxytetracycline, and Doxycycline may behelpful in the fight against the coronavirus SARS-CoV-2, due to its preferential association with theimportant residues in the viral receptor bindingdomain and the resulting strong inhibition of theRBD-ACE2 complex. Further experimental studiesare recommended to validate how this reductionof cellular infection complements or enhances theanti-inflammatory and anti-viral properties of tetra-cyclines in their role as treatment for SARS-CoV-2. A uthor contributions T.Y.Z conceived and planned the research, as wellas performed calculations. N.A.P. and T.Y.Z. per-formed analysis and wrote the manuscript. C ompeting interests The authors have no competing financial interests orother interests that might be perceived to influencethe results and/or discussion reported in this paper. R eferences [1] Omar, S.; Bouziane, I.; Bouslama, Z.; Djemel, A.In-Silico Identification of Potent Inhibitors ofCOVID-19 Main Protease (Mpro) and An-giotensin Converting Enzyme 2 (ACE2) fromNatural Products: Quercetin, Hispidulin, andCirsimaritin Exhibited Better Potential In-hibition than Hydroxy-Chloroquine AgainstCOVID-19 Main Protease Active Site andACE2. ,[2] Amin, M.; Abbas, G. Docking study of chloro-quine and hydroxychloroquine interactionwith RNA binding domain of nucleocapsidphospho-protein - an in silico insight into thecomparative efficacy of repurposing antiviraldrugs. J Biomol Struct Dyn , 1–13.[3] Sodhi, M.; Etminan, M. Therapeutic Potentialfor Tetracyclines in the Treatment of COVID-19.
Pharmacotherapy: The Journal of Human Pharma-cology and Drug Therapy , , 487–488.[4] Yates, P. A.; Leone, A. M.; Reichel, E. A Pro-posed Randomized, Double Blind, PlaceboControlled Study Evaluating Doxycycline forthe Prevention of COVID-19 Infection andDisease In Healthcare Workers with OngoingHigh Risk Exposure to COVID-19. medRxiv ,[5] J. D. Byrne, R. Shakur, J. Collins, S. L. Becker,C. C. Young, H. Boyce, and C. Traverso. Pro-phylaxis with tetracyclines in ards: Potentialtherapy for covid-19-induced ards? medRxiv ,2020.[6] A. E. Malek, B. P. Granwehr, and D. P. Kon-toyiannis. Doxycycline as a potential partnerof covid-19 therapies. IDCases , 21:e00864, 2020.[7] Andrei, S.; Droc, G.; Stefan, G. FDA ap-proved antibacterial drugs: 2018-2019.
Discov-eries (Craiova, Romania) , , e102–e102.[8] Veeramachaneni, G. K.; Thunuguntla, V. B.S. C.; Bobbillapati, J.; Bondili, J. S. Structuraland simulation analysis of hotspot residuesinteractions of SARS-CoV 2 with human ACE2receptor. Journal of Biomolecular Structure andDynamics , 1–11.49] Du, Q.-S.; Li, D.-P.; He, W.-Z.; Chou, K.-C. Heuristic molecular lipophilicity potential(HMLP): Lipophilicity and hydrophilicity ofamino acid side chains.
Journal of ComputationalChemistry , , 685–692.[10] Gautieri, A.; Beeg, M.; Gobbi, M.; Rigoldi, F.;Colombo, L.; Salmona, M. The Anti-Amyloidogenic Action of Doxycycline: AMolecular Dynamics Study on the Interactionwith AB42. Int J Mol Sci , .[11] Nimgampalle, M.; Devanathan, V.; Saxena, A.Screening of Chloroquine, Hydroxychloro-quine and its derivatives for their binding affin-ity to multiple SARS-CoV-2 protein drug tar-gets. Journal of biomolecular structure & dynamics , 1–13.[12] Rohan Narkhede R., J. A. P., RameshwarCheke S.; D., S. S. The Molecular DockingStudy of Potential Drug Candidates ShowingAnti-COVID-19 Activity by Exploring of Ther-apeutic Targets of SARS-CoV-2.
EJMO , ,185–195, doi: 10.14744/ejmo.2020.31503.[13] Day, E. S.; Cote, S. M.; Whitty, A. Binding effi-ciency of protein-protein complexes. Biochem-istry , , 9124–9136.[14] Lan, J.; Ge, J.; Yu, J.; Shan, S.; Zhou, H.;Fan, S.; Zhang, Q.; Shi, X.; Wang, Q.; Zhang, L.;Wang, X. Structure of the SARS-CoV-2 spikereceptor-binding domain bound to the ACE2receptor. Nature , , 215–220.[15] Alguel, Y.; Meng, C.; Terán, W.; Krell, T.;Ramos, J. L.; Gallegos, M.-T.; Zhang, X. Crystalstructures of multidrug binding protein TtgRin complex with antibiotics and plant antimi-crobials. Journal of molecular biology , ,829–840.[16] Huta, B. P.; Mehlenbacher, M. R.; Nie, Y.;Lai, X.; Zubieta, C.; Bou-Abdallah, F.;Doyle, R. P. The Lysosomal Protein SaposinB Binds Chloroquine. ChemMedChem , ,277–282.[17] Palm, G. J.; Buchholz, I.; Werten, S.; Gir-bardt, B.; Berndt, L.; Delcea, M.; Hinrichs, W. Thermodynamics, cooperativity and stabil-ity of the tetracycline repressor (TetR) upontetracycline binding. Biochimica et BiophysicaActa (BBA) - Proteins and Proteomics , ,140404.[18] Davis, I. W.; Baker, D. RosettaLigand dockingwith full ligand and receptor flexibility. J MolBiol , , 381–392.[19] Lemmon, G.; Meiler, J. Rosetta Ligand dockingwith flexible XML protocols. Methods Mol Biol , , 143–155.[20] Meiler, J.; Baker, D. ROSETTALIGAND:protein-small molecule docking with full side-chain flexibility. Proteins , , 538–548.[21] Berendsen, H. J. C.; van der Spoel, D.; vanDrunen, R. GROMACS: A message-passingparallel molecular dynamics implementation. Computer Physics Communications , , 43–56.[22] Lindahl, E.; Hess, B.; van der Spoel, D. GRO-MACS 3.0: a package for molecular simulationand trajectory analysis. Molecular modeling an-nual , , 306–317.[23] Kumari, R.; Kumar, R.; Lynn, A. g_mmpbsa—A GROMACS Tool for High-Throughput MM-PBSA Calculations. Journal of Chemical Informa-tion and Modeling , , 1951–1962.[24] Chen, P.-C.; Kuyucak, S. Accurate determi-nation of the binding free energy for KcsA-charybdotoxin complex from the potential ofmean force calculations with restraints. Bio-physical journal , , 2466–2474.[25] Plimpton, S. Fast Parallel Algorithms for Short-Range Molecular Dynamics. Journal of Compu-tational Physics , , 1 – 19.[26] Brooks, B. R. et al. CHARMM: the biomolecu-lar simulation program. J Comput Chem , , 1545–1614.[27] Dellago, C.; Hummer, G. Computing Equilib-rium Free Energies Using Non-EquilibriumMolecular Dynamics. Entropy ,16