Docking studies on MexB efflux pump inhibitors of Pseudomonas aeruginosa.pdf

The virtual screening results of the first binding site the distal pocket...43 3.5.2.. LIST OF ACRONYMSABC The ATP binding cassette ABI-pp Inhibitors Pryridopyrimidine derivative EPIs Ef

MINISTRY OF EDUCATION ANDTRAINING

NGUYEN TAT THANH UNIVERSITY

SCIENTIFIC RESEARCH PROJECT OF STUDENT IN 2020

NAME OF DISSERTATION: DOCKING STUDIES ON MEXB EFFLUX

PUMP INHIBITORS OF PSEUDOMONAS AERUGINOSA

Code of dissertation:

Author of dissertation: LE THI KHANH VY

Scientific instructor: M.S Pharm PHAN THIEN VY

Faculty: Faculty of Pharmacy

Student ’s name: LE THI KHANH VY

Ho Chi Minh City, October 2020

TABLE OF CONTENTS

LIST OF ACRONYMS iii

LIST OF FIGURE iv

LIST OF TABLE vii

CHAPTER 1 LITERATURE REVIEW 1

1.1 Antibiotic resistance and multidrugefflux systems 1

1.1.1 Introduction 1

1.1.2 Antibiotic resistance 2

1.1.3 Multidrug efflux systems 3

1.2 Pseudomonas aeruginosa 4

1.2.1 Structure of MexAB- OprM 5

1.2.2 Functional rotationmechanismMexB 8

1.2.3 Drug efflux mechanism 10

1.3 Inhibitionof MexBefflux pump 12

1.3.1 Synthetic compounds 12

1.3.2 Natural compounds 14

1.3.3 The methods determined efflux pumps inhibitors activity 15

1.4 Virtual screening 16

1.4.1 Structural-based virtual screening 16

1.4.2 FlexX software 17

CHAPTER 2 SUBJECTS - RESEARCH METHOD 19

2.1 Subjects 19

2.1.1 Data sets 19

2.1.2 MexB pump 23

2.2 Docking studyprocess 23

2.3 Docking molecular 24

2.3.1 Preparation oftheprotein 24

2.3.2 Preparation ofthe ligands 24

2.3.3 Binding site determination 25

2.3.4 Docking 26

2.3.5 Analyzing the dockingresults 26

2.4 Virtual screening 27

CHAPTER 3 RESULT AND DISCUSSION 28

3.1 The result of preparing protein 28

3.2 The binding site result 28

3.3 Redock 32

3.4 The dockingresults 33

3.4.1 The first binding site (the distal pocket) 33

3.4.2 The second binding site 38

3.5 Thevirtual screening results 42

3.5.1 The virtual screening results of the first binding site (the distal pocket) 43

3.5.2 The virtual screening results of the second binding site 45

CHAPTER 4 CONCLUSION AND SUGGESTION 47

4.1 Conclusion 47

4.2 Suggestion 48

REFERENCES

APPENDIX Appendix-1

11

LIST OF ACRONYMS

ABC The ATP binding cassette

ABI-pp Inhibitors Pryridopyrimidine derivative

EPIs Effluxpump inhibitors

LMNG Lauryl Maltose Neopentyl Glycol

MATE The multidrug and toxic compound extrusion family

MESs Multi-drug efflux systems

MFS The majorfacilitator superfamily

MIC The minimuminhibitory concentration

MPC8 Concentration thatdecreases the MIC by 8-fold.PACE The proteobacterial antimicrobial compound efflux

familyRND Resistance -Nodulation - Division

SMR The small multidrug resistance family

LIST OF FIGURE

Figure 1.1 Sixmajorfamilies of efflux pump transporters (Chapman et al., 2019)4

Figure 1.2 The overall structure of MexAB - OprM (Ding et al.,2014) 5

Figure 1.3 MexB efflux pump structure (Sennhauser et al., 2009) 6

Figure 1.4 Proximal pocket (red) and distal pocket (blue) in the porter domain (Ramaswamyet al., 2018) 7

Figure 1.5 (A) Functional rotation mechanism and (B) Substrate transportpathway fromthe AP totheExit Gate going through theDPin RND transporters(Ramaswamy et al., 2018) 9 Figure 1.6 Drug access and pathway in each step of the functional rotation mechanism 10 Figure 1.7 Complex formation and drug efflux by MexAB-OprM (Tsutsumi et al., 2019) 11 Figure 1.8 2D structure ofsome synthetics EPIs 14

Figure 1.9 FlexX software 17

Figure 1.10 Incremental Construction Algorithm (Moitessieret al., 2008) 18

Figure 2.1 3D structure of 6IIA in MOE 2008.10 23

Figure 2.2 Docking studyprocess 23

Figure 2.3 The virtual screeningprocess 27

Figure 3.1 Chain B ofthe 6IIA proteinin MOE 2008.10 28

Figure 3.2 Amino acids are determined at binding cavity by docking function in MOE 2008.10 28

Figure 3.3 Two binding sitesdetermined by docking resultsin MOE 2008.10 29

Figure 3.4 Amino acids aredetermined at the first binding site 30

Figure 3.5 Three importantamino acids atthe first binding site 30

Figure 3.6 Amino acids aredetermined at the second binding site 31

Figure 3.7 Three importantamino acids atthe second binding site 32

Figure 3.8 Lauryl Maltose Neopentyl Glycol Molecularweight: 1,005.19 33

Figure 3.9 Amino acids aredetermined at the distal pocket in the docking result 34 Figure 3.10 3D & 2D poseview images of BMRJ 201616 PA4MBN (-42.49 KJ/mol) atthe distal pocket 34

IV

Figure 3.11 3D & 2D poseview images of BMCL_2006_14_8506_4C (-42.04KJ/mol) atthe distal pocket 35

Figure 3.12 The BMCL 2006 14 8506 derivated compounds:BMCL_2006_14_8506_4B (pink); BMCL_2006_14_8506_4C (blue);BMCL_2006_14_8506_4P (red); BMCL_2006_14_8506_4W (yellow) 35

Figure 3.13 2D poseview images of BMCL_2006_14_l 993-2004-37 andBMCL_2006_14_1993_2004_20at the distal pocket 36

Figure 3.14 2D poseview of MCR_2017_26_414_430_PRENAG20 at the distal

Figure 3.15 2D poseview images of BMCL-2004-14-2493-2497-32 (tetrazolesubstituents) and BMCL_2004_14_2493_2497_26 (carboxylic acid substituents) at the distal pocket 37

Figure 3.16 3D & 2D poseview images of BMCL_2007_15_7087_D139001 at the

Figure 3.22 Thevirtual screening process 42

Figure 3.23 3D & 2D viewpose images of DB01145 (-34.32 KJ/mol) at the distal

Figure 3.26 3D & 2D viewpose images of DB07325 (-21.75 KJ/mol) at the second

LIST OF TABLE Table 2.1 Potential compounds, which weretestedbyin vitrobiological assay (MIC,

MPC8) 19

Table 2.2 Set A, created from the agent have to inhibit the transport of an efflux pump substrate 20

Table 3.1 Interaction ofSet A atthe first pocket 30

Table 3.2 Interaction ofSet A atthe second pocket 32

Table 3.3. 10 compounds have good docking score at the distal pocket 33

Table 3.4. 10 compounds have good docking scoreatthe second binding site 38

Table 3.5 Comparing docking score at two binding of BMCL 2006 14 8506 derivatives 41

CHAPTER 1 LITERATURE REVIEW

1.1 Antibiotic resistance and multidrug efflux systems

1.1.1 In traduction

Antibiotics have been used extensively during several decadesand we are now facing the emergence of multidrug resistant strains Furthermore, multidrug - resistant (MDR) in Gram - negative pathogens, poses significant threat to our ability to effectively treatinfections caused by organisms The inabilityofexisting therapies to treat multidrugresistantpathogens has been recognized as an important challenge ofthe 21st century A major component in the development of the MDR phenotype in Gram-negative bacteria is overexpression of Resistance - Nodulation - Division (RND) - type efflux pumps, which actively pump antibacterial agents and biocidesfrom the periplasm to the outside of the cell Consequently, bacterial efflux pumps are an important target for developing novel antibacterial treatments Potent effluxpump inhibitors (EPIs)could be used as adjunctive therapies that would increase thepotency ofexisting antibiotics and decrease the emergence of multidrug resistance bacteria (Reza et al., 2019)

A particularly problematic pathogen in the clinical setting is Pseudomonas aeruginosa (P aeruginosa), an opportunistic Gram - negative pathogen characterized by intrinsic resistance to a wide variety of antimicrobial agents At present, MexAB- OprM playavery important role in extruding antibioticsfromthecell in p aeruginosa. It is widely overexpressed in clinical isolates and contributes

to the resistance ofwild - type strains to antibiotic drugs MexAB - OprM exhibits the broadest substrate specificity of all known multidrug efflux system of

p aeruginosa.

The list of substrates includes various antimicrobial agents (Chloramphenicol, Macrolides, B-lactams, Trimethoprim, older and newer Fluoroquinolones), dyes (Acriflavine, Acridine orange,Crystal violet, and Ethidium bromide), detergents, teatree oil components(a-terpineol), organic solvents, andtriclosan(Lamutetal.,2019)

In addition, elimination of RND pumps in p aeruginosa by inhibition with a potent efflux pump inhibitor decreases the frequency of resistance to Levofloxacin (Griffith

et al., 2001)

1

With the development of science and technology, computational therapeutics is a dynamic and rapid growing in the global drug research and development because these method can save alotof time and money So that, the direction of research is

- Building database of MexB efflux pump inhibitors (EPIs) inp aeruginosa

from scientific research articles;

- Buildinga moleculardocking basedonthe X - ray crystal structure of MexBfromp aeruginosa;

- Analyzing and giving the connection between structure and bindingpotentiation;

- Screening and finding out potential EPIs from Drugbank, ZINC, Traditional Chinese Medicine, Natural plant

1.1.2 Antibiotic resistance

Antibiotic resistance has become a serious threat to human health (Organization,2017) According to an exhaustive review commissioned by the UK government,globally,approximately 700,000 deaths can be attributed to antibiotic resistanceeach year Furthermore, based onincreasing incidence ofdrugresistanceamong bacterial infections, this toll is expected to exceed 10 million by 2050, at a cumulative global costof 100 trillion US dollars Additionally, global surveillance datafrom the WHO reported awidespreadoccurrence of antibioticresistanceamong 500,000people with suspected bacterial infectionsacross 22 countries Moreover, resistancetopenicillin,P-lactam antibiotic widely used for decades to treat a range of different bacterial infections, was reported amongst 51% of the surveyed countries (Reza et al., 2019)

p aeruginosa is an opportunistic pathogen that is a leading cause of morbidity andmortality in cystic fibrosis patients and immunocompromised individuals Eradication of p. aeruginosa has become increasingly difficult due to itsremarkable capacity to resist antibiotics Strains ofp aeruginosa are knownto utilize their highlevels of intrinsic and acquired resistance mechanisms to counter most antibiotics (Pang et al., 2019)

Antibiotic resistance p aeruginosa accounts for 13-19% of healthcare-associated infections each year in the US The increasing level of resistance in antibiotic resistance p aeruginosa is often attributed to patient-to-patient transmission ofresistant strains as well as newly acquired resistance owing to previous antibiotic

exposure (Raman et al., 2018) The World HealthOrganization places resistant p aeruginosa as a critical priority pathogen that desperately requires newtreatment options (Tacconelli et al., 2017).

carbapenem-Increasing rates of multidrug - resistant (MDR) p aeruginosa in healthcare - associated infections and among hospitalized patients is a major public health problem (Obritsch et al., 2004)

1.1.3 Multidrug efflux systems

Bacterial efflux pumps play an important role in expelling toxic compounds out ofthe cell They consist of cytoplasmic membrane transporters, periplasmic linkerproteins and outer membrane porin channel proteins (Pang et al., 2019)

There are various ways in which bacteriumcan developits resistance to an antibiotic, including target alteration, enzymatic inactivation, permeable change and effluxeffect Over - expression of efflux pump has been progressively discovered to be related to multi-drug resistance in clinical isolates These proteins are found in bothGram - positive and negative bacteria as well as in eukaryotic organisms such as

Candida albicans, Plasmodium falciparum Efflux pumps are transport proteins involved inthe extrusion of toxic substrates, such as antibiotic, biocides and dye from within cells into external environment Some bacterial efflux pumps extrude selectively specific antibiotics, others may transport arange of structurally differentcompounds, creating aMDRphenotype (Blanco et al., 2016)

Multidrug efflux systems (MESs) are classified in six families, according to theirstructure(amino acidsequence homology) and energy requirements Theseare (i) themajor facilitator superfamily (MFS), (ii) the small multidrug resistance family (SMR), (iii) the multidrug and toxic compound extrusion family (MATE), (iv) theresistancenodulation - cell division superfamily (RND), (v) the adenosine - triphosphate (ATP)-binding cassette superfamily (ABC), and (vi) theproteobacterialantimicrobial compound efflux family (PACE) (Lamut et al., 2019)

PACE family is just determined nearly in some of negative bacteria Therefore, its molecular mechanism in transportation, proton energy hasn’t determined yet.However, it could be involve to electrochemical gradient (Hassan et al., 2015).Except forthe RND superfamily, which is only existed in Gram - negative bacteria,other efflux systems are widely distributed in both Gram - negative and positive

3

bacteria Moreover, all of these families utilize theproton motive force as anenergy source, exceptfortheABC superfamily, whichusesATP forthe subtracted extrusion (Blanco etal.,2016) TheRND transporters are found in theinner membrane of Gram

- negative bacteria which possess membrane - spanning domains RND transporters are homo or hetero - trimers consisting of promoters of about 1100 - amino acid residues in size Each promoter consist of the 12 transmembrane helices (TMHs) and

a largeperiplasmic domain The most common studied members of this family areE coll AcrAB, p aeruginosa MexB,Neisseria gonorrhoeae MtrD andCampylobacter jejuni CmeB forwhich high -resolution structures are available (Du et al., 2018).Bacteria can express its MES from morethan one family and/ or more than one type

of efflux pump belongingtothe same family Furthermore, efflux pumps canconsist

of either single components,namely NorA ofs aureus, or multiple components Forexample, efflux pumps of the RND family, which are associated with clinicallysignificant MDR, are organized as tripartite system such as AcrAB - TolC of E coli,

MexAB - OprM of p aeruginosa and BpeAB- OprB ofBurkholderiapseudomallei.

Figure 1.1 Six major families of efflux pump transporters

(Chapman et al., 2019)

In 1872, p aeruginosa was isolated for the first time by Schroeter, and in 1882Gessard made a complete descriptionof thepathogen

p aeruginosa is a ubiquitous Gram - negative bacterium, can exist in awide range

ofdifferent environmental niches including soil and large bodies of water, due to itsadaptability Clinically, it is a leading cause of incurable nosocomial infections among immunocompromised patients and chronic infections in cystic fibrosis

patients, p aeruginosa has naturally resistant to antibiotics due to a highly impermeable outer membrane On top of that, a synergy between a low-levelexpression ofthe porins and anincrease of the efflux pump’s activity leads to a very efficient expulsion of the antibiotics outside of the cell, evenbefore thedrags couldreach the target And so, it has ability to exhibit resistance to different classes ofsubstrates, which know as multidrug efflux system ofp aeruginosa (Reza et al.,2019).

Several substrates were known as dyes for example acriflavine, acridine orange,crystal violet and Ethidiumbromide; antimicrobial agents such as Fluoroqunolones,P-lactam, Macrolides, Chloramphenicol and Novobiocin; detergents andtea tree oil components (Lamut et al., 2019) Notably, carbapenem -resistant p aeruginosa hasbeendesignated as the second most threateningMDR pathogen to the human health,

by the WHO (Reza et al., 2019)

1.2.1 Structure ofMexAB - OprM

RND type efflux pumps play an important role in extruding antibiotics from the cell

inp aeruginosa. MexAB - OprM is themain efflux systems responsible for bacterialsurvival during infection and intrinsic antibiotic resistance It is composed of threeparts an inner membrane protein MexB,anouter membrane proteinOprM,andMexA(Lamut et al., 2019)

The overall structure of MexAB - OprM is a vertically elongated rod shape, ~320 Ảalong the long axis MexB does not directly contact OprM; instead, MexA joinsMexB and OprM byforming a funnel-like hexamer(Tsutsumi et al., 2019)

Figure 1.2 The overall structure of MexAB - OprM (Ding et al., 2014)

5

MexB contains 1046 amino acids MexB are symmetrical homotrimers and each monomer is composed of 12 transmembrane a-helices and two large hydrophilicloops The protein has a three - fold symmetryaxis perpendicular to the membraneplane In their trimeric organization, three domains can be established: atransmembranedomain, embedded inthe inner membrane; apore domain, located in theperiplasm; a docking domain (Ramaswamyet al., 2018).

Figure 1.3 MexB efflux pump structure (Sennhauser et al., 2009)

The transmembrane (TM) domain part is 50 Ảlongand isformed of 36a - helices, where the chemical to mechanical energy conversion takes place Each monomer consists of 12 transmembrane helices (TMH1 to TMH12), defining a central cavity accessible from the cytoplasm Each monomer contains an extra - membrance a - helix located between TMH6 and TMH7 attached to the cytoplasmic membranesurface TMH4 and TMH10, at the core of the domain consist of charged residuesAsp407, Asp408 (TMH4) and Lys939 in MexB, which mediate the proton translocation Indeed, Lys939 (or 940) is engaged in a salt-bridge interaction with Asp407 and Asp408 This domain is loosely packed and therefore probably filledwith phospholipids (Sennhauser et al., 2009)

The porter domain located in the periplasm, where substrate recruitment andtransport occur This domain has pivotal role because this is where the substrate binding takes place The porter domain, is composed of four subdomains ofPN1 and PN2, located in the N - terminal, half between first and second TMH (TMH1 and

TMH2) and the repeat correspondent subdomains of PC 1 and PC2, situated in the c

- terminal half between TM7 and TM8 These subdomains have a characteristicstructural motif (a p - strand - a - helix - p -strandrepeat) that is sandwiched with one another PN1 and PCI are composed offive p - strand In PN2 and PC2, therepetition of the p - strand - a - helix - p - strand is interrupted by the dockingdomain, DN and DC These four subdomains are packed with their p sheets back to back in the Center and their a-helices onthe outside (Sennhauser et al., 2009)

The foursubdomains (PN1, PN2, PCI,PC2) packto formtwoproximal (access) anddistal (deeper) substrate binding pockets, which are separated by a switch glycine- rich loop, namely G-loop The pockets are enriched in aromatic, polar and chargedamino acid residues that form favorable interactions with thetransportsubstrates The proximal and distal pockets have substratepreferences Thedistal pocket is separatedinto two parts, ahydrophobic trap and an almost hydrophilic substrate translocation channel The distal pocket is situated in the periplasmic domain of the bindingconformer The channel is spacious so that it makes a convenient space for multisitedrug binding Also a narrow and deep cleavage has been formed by the hydrophilictrap G - loop controls the access of substrates to the distal pocket by forming aboundarybetweenthe proximal and distal-binding pockets (Jamshidi et al., 2016)

Figure 1.4 Proximal pocket (red) and distal pocket (blue) in the porter domain

(Ramaswamy et al., 2018)

7

The docking domain which interacts with the outer membrane channel It is theperiplasmic component that is responsible for the specificity of the componentinteractions A protrusion from the docking region into the adjacent monomer holdsthe three monomers together and possibly stabilisesthe complex The subdomains of

DN and DC belong to the docking domain.These subdomains are each comprised of

a four - stranded mixed p - sheet, whereby two antiparallel p - strands are located parallel to an additional p - sheet hairpin structure incorporated from the other half

ofthe protomer or one of the other monomers Additionally, a vertical hairpin motif

is locatedatthe apex ofeach of the subdomains As the three OprM docking domains,one from each monomer, come together, a funnel-like structure is formed withlarger opening at the top Intriguingly, this opening is similar in diameter to thebottom ofthe outer membrane channel OprM.The funnel narrowsas it extends down the OprM docking domain and leads to the central pore of the porter domain(Sennhauser etal., 2009)

1.2.2 Functional rotation mechanism MexB

MexB is situated in the bacterium cytoplasmic inner membrane and possess membrane - spanning domains and a large periplasmic domain They mediatetransport ofmany diverse substrates, whichcan be anionic, cationic, zwitterionic orneutral, and include antibiotics, chemotherapeutics, othertoxins, detergents and bile salts Whether they are energized by electrochemical gradients or the binding andhydrolysis of nucleoside triphosphates, the inner membrane proteins bear TMdomains that can adoptdistinct conformations,which are necessaryto cycle through substrate transport states (Jamshidi et al., 2016) Substrate transport is characterized

by the typical “functional rotation mechanism” in which concerted cycling of theprotomer occurs through all ofthe so far identified asymmetric states: Loose (L) in which a substrate binds to a peripheral site termed access pocket (proximal pocket);Tight (T) in which the substrate binds to a deeperpocket (distal pocket); and Open (O) in which the substrate is releasedinto the central funnel leading toward the OprM

(Figure 1.5) (Ramaswamy et al., 2018)

❖ The loose state (L): substrate access

When a monomer adopts this conformation, there are three possible entry points forthe substrate: (Figure 1.6) i) A hydrophobic cavity, highly accessible, about 15 Â

above the membrane plan This cavity, calledthe vestibule, which corresponds to agroove between TMH8 and TMH9, is presumably used by highly hydrophobicsubstratescoming from theouter leaflet of the inner membrane This groove leads to the proximal pocket but, inthe loose state, a switchloop prevents the substrate frombinding to the deepbinding pocket, ii) A periplasmic cleft between PCI and PC2 is probably used by less hydrophobic substrates which are soluble intheperiplasm.Thiscleft leads as well to the proximal pocket The switch loop also precludes thebinding

of substrate inthe deep binding pocket, iii) A tunnel thatwould link the large centralcavity of MexB to the access pocket In this conformation, the side of the bindingpocketthat isconnected totheOprM docking domain is close Inthe loosestate, PN2and PCI strongly interact, which creates a hydrophobic pocket at the PN2/PC1 interface

Large substrate molecules tightly bind to the proximal pocket in the loose monomer.Then, they are forced into the deepbinding pocket, uponthe loose to tight transition and thanks to the motion ofthe switch loop By contrast, small molecules cross theproximal pocket without specific interactions and directly bind to the deep bindingpocket (Murakamiet al., 2006)

Figure 1.5 (A) Functional rotation mechanism and (B) Substrate transport pathway from the AP to the Exit Gate going through the DP in RND

transporters (Ramaswamy et al., 2018)

❖ The tightstate (T): substratebinding

In the tight state, the geometry of the hydrophobic pocket is changed, which nowadopts a conformation that does not exist inthe open and loose states Indeed, inthe

9

tight state, the binding pocket is expanded As a result, the aromatic rings of thephenylalanine residues favour hydrophobicinteractions with the substrate In the tight state the cavity is still partially open to the periplasm but the switch loop is in adifferent position compared to that in the loose state and thereby prevents thesubstrate to go backto the accesspockets (Murakami et al., 2006).

❖ The open state (O): substrate extrusion

In this state, the lateral entrance of the hydrophobic cavity is closed (which is wideopen in the loose stateand accessible inthe tightstate) During thetransconfonnationfromtight to open, the substrate is squeezed out of its hydrophobic cavity because thespace in the cavity is greatly decreased The tunnel pathway is opened to the funnel­like partof MexB, allowing the substrate to access the OprM docking domain Intheopen state, substrate binding from the periplasm is impossible This mechanism suggests an interdependence ofthe monomers (Murakamiet al., 2006)

Figure 1.6 Drug access and pathway in each step of the functional rotation

mechanism The purple color is the substrate Numbers to 3 localize the

several drug access pockets: 1 is the groove; 2 is the cleft and 3 is the large

central cavity (Nakashima et al., 2011).

1.2.3.Drug efflux mechanism

p aeruginosahasnaturallyresistantto antibiotics due to ahighly impermeable outermembrane On top ofthat, synergy between a low-level expression of the porinsand anincrease of the efflux pump’s activity leads toa very efficient expulsion of the

antibioticsoutside of the cell, evenbeforethe drugs could reach the target And so, it has ability to exhibit resistance to different classes of substrates, which know as Multi-drug efflux system of p aeruginosa.

Inthe absence of drugs, theMexB protomers are closed towards theoutsideand these are therefore unable to bind drugs Hence, this state probably resembles the resting state of thecomplex In the presence ofdrugs,however, some of the MexB protomersare opened towards the outside, thereby enabling the binding ofdrugs in the distal binding pocket When the surrounding drug concentration becomes high, the gate loop of MexB shifts downward, and the binding pocket is opened to the molecularsurface After that, MexB ejects drugs into the tunnel of MexA - OprM via afunctional rotation mechanism When the concentration of drugs in this tunnel becomes higherthantheir concentration outside the cell, they diffuse out ofthe cell viathe concentration gradient As the concentration of drug inthe cell decreases, thedrug entrance ofthe binding protomer of MexB is closed by the structural change ofthe gate loop, and the complex shifts from the Binding state (Tight state) to theResting state Under these conditions, because the pump is completely closed to theperiplasm, backflow of drugs is prevented When the concentration of drug in theenvironment rises, the resting protomer again undergoes a structural change to theBinding state, and drugs are taken in and released (Tsutsumi et al., 2019)

Figure 1.7 Complex formation and drug efflux by MexAB-OprM

(Tsutsumi et al., 2019)

11

1.3 Inhibition of MexB efflux pump

During this decade, various compounds have been designed and synthesized to inhibitthe activity ofp aeruginosa drug efflux pumps These compounds, called effluxpump inhibitors (EPIs), exhibitnointrinsic antibacterialeffect at a concentration thatinduces a restoration of antibiotic activity in an efflux pump mediated MDR bacterium Effluxpump activities can be blocked by EPIs This can be achieved by interferingwith the functional assemblies of efflux pumps’ components, and effluxinhibition can play an important role in preventingthe development ofresistance inbacterial pathogens (Pages etal., 2009)

EPIs act as therapeutic agents since they have the potential to restore the activity ofconventional antibiotics The combinationof EPIs alongwithantibioticsis expected

to reduce the intrinsic resistance ofbacteria towards a varying range of antibiotics, decrease the frequency ofemergingresistant mutant strains and facilitate the reversal

ofacquired resistance in strains thathave multiple targetedmutations Consequently, there has been considerable interest in developing EPIs of the RND family pumps(Lamut et al., 2019)

1.3.1 Synthetic compounds

❖ Peptidomimetics

In 1999, Renau et al characterized a group of peptidomimetic molecules which areactive against p aeruginosa strains that overexpress the MexAB - OprM efflux pump The leadcompound was extensively studied and several derivativemoleculeshave been made allowing a structure - activity relationship Among this group ofpeptidomimetics is thepioneerlead compound,MC-207,110 (PA0N) that effectivelyinhibits the efflux mechanism that is responsible for pumping out quinolones, especiallylevofloxacin, the fluoroqinolone used for the screening The MexB effluxpumpof p aeruginosa recognizes PApNas apump substrate The competitivenature

of PA0N with antibiotic substrates of an efflux pump system means that while thepump preferentially pumps out PApN, the antibiotic remains in the cell, and itsamount continuestobuild, ultimately tothepoint where it canreachthe concentration required forits activity on the target The final effect of PApNis to reduce resistance

or completely reverse resistance to a given antibiotic to which the bacterium was initially resistant However, PApN can compete with some antibiotics and notothers

for access to a given over-expressed efflux pump This suggests a differential relationship between PApN and antibiotic substrates which is based uponthe nature

ofthepump and the large substrate binding pocket (Pages & Amaral, 2009)

Several years later, derivatives molecular of PaPN have been created and tested to increasethe stability of theEPI in biologicalfluids.MC-04,124, which presents goodsynergistic effect on Levofloxacine against p aeruginosa strains overproducingefflux pump in an animal model ofexperimental infection, has both greater stabilityand less toxicity than original molecule (Askoura et al., 2011)

❖ Pyridopyrimidinones

Among pyridopyrimidine derivatives, the relatively water - soluble inhibitor ofpryridopyrimidine derivative (ABI - PP) is a promising drug candidate ABI - pp binds to the distal pocket, therefore, only the distal-binding drugs are affected by binding ofABI - pp tothe pump and ABI - pp has no effect on proximal -bindingdrugs due to its lack of interaction with the proximal region Molecular dynamics simulations also show that ABI - pp binds tightly to the phenylalanine - rich trap, and inhibits the functional rotation mechanism of inner membrane protein (IMP)monomers Therefore, it increases the activity of antibiotics, exported by IMP, whether they interact with the distal - binding or proximal - binding sites The inhibitorspecificity is determined bythe smalldifferenceofhydrophobictrap space Yoshida et al generated and characterised a series of pyridopyrimidine analogues, by incorporatinghydrophilic substituents onto the aryl core Among them, D13-9001, amorpholine analogue, not only retained good in vitro activity but also potentiated Levofloxacin and Aztreonam activity in vivo, in animal models infected with p.aeruginosa overexpressing MexAB - OprM mutants The compound and itsderivatives exhibited an excellent solubility and a favourable safety profile.Futhermore, in comparison with PaPN, DI3-9001 has been shown to significantly reduce the invasiveness of p aeruginosa mutants that highlighting the potential ofthe pyridopyrimidine to manage p aeruginosa virulence. However, unlike PApN,

DI3-9001 had no influence on p aeruginosa growth, possibly due to its reduced cytotoxicity (Reza et al., 2019)

13

inhibitory capability on MexAB - OprM efflux pump ofp aeruginosa (Pesingi etal., 2019).

Futhermore, Conessine - a streroidal alkaloid compound can be found inHolarrhena antidysenterica, which is usually used for the treatment of gastrointestinalinfection such as dysentery and diarrhoea, is discovered as a EPI by inhibiting efflux pumpsystem in p aeruginosa and restoring antibiotic activity In the research, Conessinecan decreased MICs ofLevofloxacin, Novobiocin and Rifampicin in all strain of p aeruginosa similar to PaPN With Cefotaxime, Erythromycin and Tetracycline,Conessine could Lover the MICs in all trains better than PaPN (Siriyong et al., 2017) Epigallocatechin - - gallate (EGCG), a polyphenolic compound extracted from green tea, exhibits antibacterial activity Our results also suggestthat EGCG inhibits the efflux pump (MexAB - OprM)as shownby the results thatEGCG synergistically interacted with PApN and enhanced susceptibility toantibioticsatmuch higher levelsthan that of either EGCG or PApN The efflux pump inhibitor(PApN) is specific to

the effluxpump (MexAB - OprM) inp aeruginosa. Therefore, EGCG, as well as PAỊ3N, inhibits the same efflux pumpwith an accumulationofmore antibiotics thanthat of either one However, further calcification for EGCG as an efflux pumpinhibitor requires(Kanagaratnamet al., 2017).

Microorganisms have evolved to produce some interesting compounds that targetMDR efflux pumpsas apart oftheir chemical arsenalfor combating othercompetingspecies Two ofthese are EA - 371« and EA - 371Ô, which were identified throughscreening of a libraryof 78,000 microbial fermentation extracts They are produced

by a strain ofthe Streptomyces family closely related to Streptomyces velosus , andwere demonstrated to be potent and specific p aeruginosa MexAB - OprM effluxsystem inhibitors, with MPC8 values of 4.29 pM (EA - 371«) and 2.15 pM (EA- 371Ỗ) for levofloxacin against PAM1032 They did not show intrinsic antibacterialactivity in any ofthe strainstested In an efflux pump substrate accumulation assay with the substrate that is enzymatically hydrolyzed inside the cells to fluorescent product, EA - 371awas capable ofincreasing the accumulation ofthe substrate in adose- dependent mannerin PAM1032 (Lamut et al., 2019)

1.3.3 The methods determined efflux pumps inhibitors activity

In the studies, there are many methods determined EPIs activity known as theminimum inhibitory concentration (MIC) value, the value ofminimal concentration

of an EPIs required to decreasethe MIC of an antibiotic by 8-fold (MPC8), andthelastone usually is basedon the fluorometric efflux assay suchasEthidium bromide

assay,Nile red efflux assay andH33342 accumulation assay, which has drawn many attention fromresearchers

In drug accumulationassays the difference inrateof accumulationof the fluorescentcompound between cells with and without an active efflux pump are used as anindication of efflux, since efflux will result in lower accumulation of compound Indrug efflux assays, the de-energized cells are preloaded with the fluorescentcompound and then energized by the addition of glucose to catalyze drug efflux Inaddition, all the samples must be preloaded to the same level of fluorescence to avoiddifferencesin efflux rate as aresult of differences in the concentrationofdrug insidethe cell (Venter et al., 2015)

15

1.4 Virtual screening

Virtual screening, sometimes is also known as in silico screening, is to computationallyscreen large libraries of chemicals for compounds that complement targets of known structure, and experimentally test those that are predicted to bind well With the widespread development of science and technology, computational therapeutics is a dynamic and rapid growing in the global drug research and development(Ekins et al., 2007)

Several approaches can be used for virtual screening, for instance those based onstructural homologies with respect to aligandorto a substrate (ligand- based virtual screening), but also based on the structure of a protein by performing docking experiments (structure-based virtualscreenings) (Shoichet, 2004)

1.4.1 Structural - based virtual screening

Structural - based virtual screeningis applicable when the 3D structure of the protein

is available and then virtual screening is done on the basis of the existing structures;duetothis, anovel compound can be identified mimicking biological targetstructure

It eases the process of identification of lead compound by quick, fast and effectiveproperty (Leach et al., 2007)

cost-Moleculardocking is a major class of Structural-based virtual screening, which isbased onpredicting a binding energy between the ligand and the target proteinwiththe help of apredefinedbindingpocket site inthe targetprotein Docking also helps

in revealing different pocket sites of the ligand due to conformational changes in thereceptor(Shin et al., 2015)

Thanks to the incessant development of information technology in biology, either many databases of small molecules and objective structures as protein data banks; docking method is widely used in virtual screening Docking methods has a lot ofapplications as rational drug design, studying aboutthestructureactivity relationships (SAR), optimizing the lead compound, searching for potential compounds throughvirtual screening, giving some theory about interacting to make a guess formutagenesis studies, supporyingfor making substratesandinhibitors become suitable into electrondensity map in X-ray diffractioncrystals structures, investigatingabout chemical mechanism Mostfrequentlyused docking software are FlexX, AutoDock,GOLD, Glide, Dock (Guvench et al., 2008)

1.4.2 FlexX software

Themolecular docking process was done by usingfragmentation algorithm of FlexXsoftwareintergratedin LeadIT 2.1.8 The fragment ofligands wereseparately bindingwith protein, and then they were collected and calculated their docking scores The dockingscore was indicatedbinding affinitieswith the unit is KJ/mol

Flex X is program which predicts the interaction between protein-ligand FlexXpredicts geometrically structures of protein-ligand complex as well as the bindingenergy Docking method in FlexX works without any manual intervention FlexXworks based on protein-ligand and screening a big collection of compounds to find out the direction for rational drug design In short, FlexX is very useful in case ofhaving 3D structures of proteins and exact location ofavailable working areas; either having an association of compounds molecule FlexX would analyze the interactiveability (could or couldn’t, andhow to do that) ofevery single compoundwithproteins

(https://www.biosolveit.de/)

FlexX supports effectivelyforthe rational drug design programebased on structures and giving reliable prediction, building up an interactive way for compounds in protein’s interactive point exactly, the ability of docking a compound molecule from

an available biological structure with the deviation no more than 2 Ả Moreover,FlexX is thought tohave a high speed, suitable for highentered virtual sreening

The docking principles in FlexX: protein is rigid, ligand is flexible Flex X usesIncremental Construction Algorithm as: dividing ligand into hard small pieces, docking them into interactive pocket of protein, finally, assembling ligand from all pieces in low- energy types (Moitessier et al., 2008).

According to the interation of protein - molecular, the scoring function of FlexX is developed by Bohm The lower of docking scores are, the stronger affinity bondachive The binding energy caculation is based on the energy created hydrogenbondings, ion, aromatic and Vai der Waals interactionsbetween target structure andmoleculars

Bind fragment Select n best poses

Bind fragment Select n best poses

Bind fragment Select n best poses

Figure 1.10 Incremental Construction Algorithm (Moitessier et al., 2008)

CHAPTER 2 SUBJECTS - RESEARCH METHOD 2.1 Subjects

2.1.1 Data sets

According to 23 scientific articles, were selected the source of databases Total 66potential compounds, which were tested by in vitro biological assay (MIC, MPC8), have good inhibitory activity on p aeruginosa MexAB - OprM efflux pump

(Table 2.1; 2D - structure in Appendix 1) In 66 potential compounds have 15compounds (Set A), created from the agent have to inhibit the transport ofan efflux pump substrate Etbr (Table 2.2)

Table 2.1 Potential compounds, which were tested by in vitro biological assay

mpc8

(Renau et al., 2001)

Accumulation (Renauet al., 2002)

(Nakayama, Ishida, Ohtsuka,

Hosono, et al., 2003)

(Nakayama, Ishida, Ohtsuka,

(Yoshida, Nakayama, Kuru, et al.,

(Yoshida et al., 2007)

Accumulation (Renauet al., 1999)

19

No Scientific articles Number of

Accumulation (Lamutet al., 2019)

* Research article annotation: the acronym of research journal _year publication _ volume _ issue _ the page begins; MIC: The minimum inhibitory concentration; MPCg: Concentration that decreases the MIC by

8-fold.

Table 2.2 Set A, created from the agent have to inhibit the transport of an

efflux pump substrate

No Compounds References

(Aparna et al., 2014)

21

No Compounds References

9

(Lee et al., 2001)

(Yoshida et al., 2007)

BMCL 2007 15_7087_D 139001

(Aparna et al., 2014)

12

(Renau et al., 2002)

No Compounds References

B, c, D, E, F) with resolution 2.91 Ả Chain B is chosen for docking because co - crystal in chain B

Chain B [

Figure 2.1 3D structure of 6IIA in MOE 2008.10

2.2 Docking study process

In this study, the moleculardockingwasproceeded by FlexX software intergrated inLeadIT 2.1.8 The docking process were described inFigure 2.2

Preparing protein

MOE 2008.10

SYBYL-X 2.0

Determining binding site

LeadIT 2.1.8

Preparing ligands

Docking

Analyzing the docking results

Figure 2.2 Docking study process

23

2.3 Docking molecular

2.3.1 Preparation of the protein

At first, the MexB protein 6IIA was dowloaded from ProteinDataBank and saved asformat *.pdb After that, used the Sequence Editor function in MOE 2008.10 to eliminate redundant chains and retain chain B Finally, the interaction between protein and ligands were discovered by using LigX function in MOE 2008.10, thisprocess was described as follow:

- Protonate: Protonated andcharged the amino acids ofprotein

- Tether and Minimize-. Specified the movingregion of protein atoms and conducted energy minimization

- Deleted co - crystallization ligands

- Saved protein as format pdb.*

Figure 2.3 MOE 2008.10 software interface of the protein preparing process

2.3.2 Preparation of the ligands

Having drawn the 3D - structure ofligands in MOE 2008.10 and saved as format

*.mo!2 and transferred to Sybyl - X 2.0 The ligands were minimized their energy by Sybyl - X 2.0 to make stabilized structures The ligand preparation process were describe as followed:

- Conducted 1st energy minimized with selected parameters such as Method: Conj Grad; Termination: Energy Change 0.001 kcal/(mol A); Max Iterations: 10000; Charges: Gasteriger — Huckel. The others parameters were set up by defaultvalues

*

- Run molecular dynamics by using Compute - Dynamics — Setup Simulated Annealing - Run: 5.

- Similarly, conducted 2ndenergy minimized and saved as format mol*

- Collected ligands, saved as format sdf* andtransferred to LeadIT

Figure 2.4 Draw the 3D - structure of ligands in MOE 2008.10 software

interface

Figure 2.5 Sybyl-X 2.0 software interface in the energy minimize of ligands

2.3.3 Binding site determination

Determinatedthe bindingsiteby using docking function inMOE.2008.10

In LeadIT software, chose Molecules/Prepare Receptor and uploaded prepared protein After that, chose the main amino acids in biding site in Project Tree andtransformed them into spheres Then, specified binding cavity and extended theradius ofbiding cavity to 12.5 Ả Finally, saved thebinding cavity as format *.fxx

25

fcepRIT Baecues gocttfiR * (ore ịconng Oistfay Jpndow 1)4®

• □ LEU-30

s □ MO-31 s- □ MM.-32 s- □ ASH-33

t □ ALA-54

»0>»ct toe H k «>(0« (t_m * x6_4ock)

Figure 2.6 FlexX/LeadIT software in the binding site determination

2.3.4 Docking

At first, downloaded necessary ligand by Docking - Docking Library - Load File.

Next, conducted docking process byDocking - Define FlexXDocking. The selected parameters were described as followed:

- Number of Poses to Keep- 10 (with Training Set) and 1 (with Test Set)

- Maximum Number of Solutions per Iteration-. 1000;

- Maximum Number of Solutions per Fragmentation- 200;

- The others parameters were setup by default values

After finished docking process, saved the docking scores as format *.sdfcorresponding with binding structure ofligands

Q Ẽ3 a a ® #

fl dock_G17_15 lead IT 2.1.8 (17.09.14)

Figure 2.7 FlexX/LeadIT 2.1.8 software in the docking process

2.3.5 Analyzing the docking results

The docking score was validatedbased on the interation between ligand and protein suchas ion bonding, hydrogen bonding, Van der Waals, 71-71bonding The docking

result indicated not only the affinity bond between protein and ligand but also theinteraction between ligand and amino acid of biding site Moreover, the dockingresultwasused to researchthe accordant binding cavity.

Q polar ‘ 'Sidechain acceptor

o acidic *■' ■ sidechain donor

o basic ‘ * backbone acceptor

o greasy • * backbone donor

proximity ligand

Q solvent residue ©íẫtarene-arene

o rnetal corn pl ex ©+arene-caticiri solvent contact

(http://zinc 15.docking.org/) The virtual screening process were describe in

Figure 2.9.

plants TCM

Drug like Pharmacophore models (Dinh Nguyen Thuy Duyen, 2020) 2D - QSAR models (Nguyen Truong Khanh Vy, 2020)

Potential compounds

Figure 2.9 The virtual screening process

27

CHAPTER 3 RESULT AND DISCUSSION 3.1 The result of preparing protein

The crystal structure is downloaded from Protein Data Bank https://www.rcsb.org/

(PDB Code: 6IIA) Chain B included 1030 amino acids, is chosen for docking because its LMNG co -crystal in chain B (Figure 3.1).

Figure 3.1 Chain B of the 6IIA protein in MOE 2008.10

3.2 The binding site result

Using the Docking functionin MOE 2008.10, it isreportedthat we have 30pose foreach compound in a set A Two amino acids are determined to have the closet interaction is Phe617 (253/450 ligand) and Lysl34 (212/450 ligand)

253/450(562%)

J Ji -■ I [.IJ, Ilk 1.1 iJlLdni I I l I.I l hlL ll

33 46 77 79 81 87 69 91 93 128 132134135175176 296328564566575616 62W45664667 670672 675676677 : '681 714/1/ 718 816 829 861

Figu re 3.2 Amino acids are determined at binding cavity by docking function

in MOE 2008.10