Uporaba molekulskega modeliranja pri načrtovanju antagonistov DC-SIGN in in vitro vrednotenje njihovega učinka : doktorska disertacija

Jug, Gregor (author), Anderluh, Marko (mentor), Tomašič, Tihomir (co-mentor)

Infekcijske bolezni so na letnem nivoju druge po številu povzročenih smrti, takoj za kardiovaskularnimi boleznimi. Med infekcijskimi boleznimi oz. povzročitelji le-teh največje težave povzročajo virus HIV, Mycobacterium tuberculosis in malarija. Znano je, da različni patogeni v telo vstopajo tudi preko proteina, t.i. za dendritične celice specifičnega, medcelično adhezijsko molekulo-3-vezočega ne-integrina oz. DC-SIGN (Dendritic Cell-Specific Intercellular Adhesion Molecule-Grabbing Non-integrin). Gre za lektin tipa C, ki se nahaja na površini dendritičnih celic (DC). Zanj je značilno, da poleg virusa HIV veže še precej drugih patogenov. Ob vezavi na DC-SIGN pride do internalizacije patogena v fagocitozni mehurček. Nekateri patogeni, vključno z virusom HIV, v tem destruktivnem okolju »preživijo« ter se tako skupaj z zrelimi DC prenesejo po limfi vse do bezgavk in tako naprej po telesu, kar ima za posledico različne vrste infekcij. V sklopu doktorske disertacije smo se osredotočili na načrtovanje novih spojin, kompetitivnih antagonistov DC-SIGN, ki bi zavirale vezavo glikoproteina gp120, ki ga virus HIV izraža na svoji ovojnici. Omenjene spojine bi s svojim delovanjem preprečile vdor virusa HIV preko DC-SIGN v telo ter s tem (verjetno) ubranile organizem pred okužbo ter posledično obolelostjo za AIDS-om. Izvedli smo strukturno podprto in silico načrtovanje antagonistov DC-SIGN. Ker pred začetkom našega raziskovalnega dela ni bilo identificiranega nobenega primernega orodja za molekulsko sidranje v zelo plitko in odprto vezavno mesto DC-SIGN, smo se v prvem koraku odločili za validacijo nam dostopnih orodij za molekulsko sidranje: AutoDock, CDOCKER (vključen v programsko opremo Accelrys Discovery Studio 3.0), FlexX (vključen v programsko opremo LeadIT), FRED (vključen v programsko opremo OpenEye) in GOLD. Validacijo smo izvedli z uporabo treh različnih kristalnih struktur kompleksov DC-SIGN-ligand (PDB oznake 2IT6, 2XR5 in 2XR6), ki smo jih uporabili za ponovno sidranje (t.i. »re-docking«) spojin, ki so vključene v kristalne strukture omenjenih kompleksov. Rezultati validacije kažejo na to, da je najbolj primerno orodje za molekulsko sidranje na izbrani protein FRED, kateremu sledita CDOCKER in FlexX. Pokazali smo tudi, da AutoDock in GOLD nista primerni orodji za uporabo pri molekulskem sidranju v CRD DC-SIGN. Da bi dokazali uporabnost vzpostavljenega protokola za molekulsko sidranje z orodjem FRED, smo izvedli virtualno rešetanje namensko pripravljene knjižnice spojin. Izhodišče je predstavljalo 32 spojin, ki smo jih načrtovali in in vitro testirali v sklopu naše raziskovalne skupine. Za vsako od 32 spojin smo dodatno generirali 50 strukturno podobnih spojin (t.i. »lažnih spojin« oz. »decoys«). Dobili smo knjižnico s 1632 spojinami, s katero smo preverili uporabnost protokola za molekulsko sidranje, kot je navedeno zgoraj. Rezultati so pokazali, da je bilo vseh 32 spojin, katerim smo že uspešno dokazali vezavo v vezavno mesto DC-SIGN tekom in vitro testiranja, uvrščenih med zgornjih 5,6% rangiranih spojin. Poleg tega sta bili spojini z največjo afiniteto vezave (IC50 = 40 μM oz. 50 μM) rangirani kot četrta oz. peta rangirana spojina. Omenjeno potrjuje, da je vzpostavljen protokol primeren za izvedbo začetne faze, t.j. virtualnega rešetanja in strukturno podprtega načrtovanja ligandov – antagonistov DC-SIGN. Dodatno smo pokazali, da je vzpostavljen protokol za molekulsko sidranje z orodjem FRED, primeren tudi za uporabo pri molekulskem sidranju na druge, DC-SIGN podobne lektine. Za potrditev smo izvedli sidranje liganda v vezavno mesto proteina Langerin (PDB oznaka 3P5F), ki spada v isto skupino lektinov tipa C, ter ima podobno strukturo domene za prepoznavanje ogljikovih hidratov (CRD) kot DC-SIGN. Vezavno mesto za ogljikove hidrate se nahaja na ekstracelularni domeni DC-SIGN, točneje v sklopu domene za prepoznavanje ogljikovih hidratov (CRD). Najbolj pomemben za vezavo ogljikovih hidratov na DC-SIGN je Ca2+ ion, ki se nahaja v vezavnem mestu in je ključen za tvorbo koordinacijskih vezi med ligandom in CRD. Poleg Ca2+ iona imajo pomembno vlogo za vzpostavitev interakcij ligand-protein tudi aminokislinski ostanki DC-SIGN (npr. vzpostavitev H-vezi). Spojine smo načrtovali na podlagi podatkov iz temeljite študije vezavnega mesta, ki je bila v preteklosti narejena v sklopu naše raziskovalne skupine. Vezavno mesto, ki je sicer zelo plitko in iz tega stališča zelo težavno za strukturno podprto načrtovanje potencialnih ligandov z molekulskim sidranjem, ponuja nekatere neizkoriščene strukturne atribute, ki smo jih tekom načrtovanja želeli izkoristiti za namen vzpostavitve dodatnih interakcij med proteinom in ligandom. Prvo možnost predstavlja hidrofobna votlinica, locirana za Phe313. Druga možnost, ki se je tekom našega dela izkazala kot precej zanimiva, je sam Phe313, ki zaradi lege aromatskega obroča omogoča tvorbo π-π interakcij z ligandi, ki imajo v svoji strukturi aromatske obroče. Tretjo možnost pa predstavlja raven hidrofoben del vezavnega mesta, ki ga velja izkoristiti za izboljšanje jakosti vezave liganda v vezavno mesto. Tekom načrtovanja novih potenicalnih antagonistov DC-SIGN smo izhajali iz osnovne manoze, ki je ključna za prepoznavo in vzpostavitev interakcij med spojino in Ca2+ ionom ter nekaterimi aminokislinskimi ostanki v vezavnem mestu proteina. Manozi smo na podlagi izsledkov iz literature in lastnega znanja ter poznavanja strukture proteina dodajali primerne distančnike in funkcionalne skupine. Usmerili smo se v strukture, ki nudijo primerne geometrijske in elektrokemijske lastnosti za uspešno vezavo osnovne manoze, kakor tudi možnost interakcij z dodatnimi področji proteina, predvsem s Phe313. Pri raziskovalnem delu smo potrebovali protein DC-SIGN, ki ni komercialno dostopen. S tem namenom smo del raziskovalnega dela opravlili na inštitutu IBS (»Institut de Biologie Structurale«) v Grenoblu v Franciji. Pripravljen rekombinantni protein DC-SIGN ECD smo uporabili pri in vitro testiranju antagonistov DC-SIGN ter pri izvedbi vrednotenja vezave z uporabo metode STD NMR. Kot dodaten cilj smo si zadali pripravo rekombinantnega fluorescenčno označenega proteina DC-SIGN ECD, ki bi ga uporabili pri razvoju novega in vitro testa za preverjanje afinitete vezave načrtovanih antagonistov DC-SIGN. Načrtovane spojine, za katere smo predvideli potencialno visoko afiniteto vezave, smo v sklopu raziskovalnega dela Katedre za farmacevtsko kemijo uporabili pri izvedbi in vitro testa. Uporabili smo robusten in vitro test, ki temelji na kompetitivni vezavi spojine in gp120 na DC-SIGN ECD, ki smo ga pred tem vezali na primerno podlago. Samo uspešnost vezave naših načrtovanih spojin smo preverili z reakcijo z biotinom ter z merjenjem kemiluminiscence. V sklopu raziskovalne skupine smo izvedli in vitro testiranje ter tako preverili afiniteto vezave za 32 novih spojin. Pri dveh spojinah smo dobili IC50 vrednosti v mikromolarnem območju (40 μM oz. 50 μM). V zadnjem sklopu raziskovalnega dela smo podrobneje proučili interakcije štirih izbranih monovalentnih ligandov na osnovi manoze z vezavnim mestom DC-SIGN. Najprej smo z molekulskim sidranjem predvideli vezavno pozo vseh diastereoizomerov spojin, pri čemer smo uporabili validirana protokola za molekulsko sidranje v programih FlexX in FRED. Rezultati sidranja kažejo, da je program FRED osnovno manozo pravilno umestil v vezavno mesto, ter da je dodatno pravilno predvidel tudi hidrofobne interakcije in/ali π-π interakcije med ligandom in stransko verigo Phe313. Da bi odpravili pomanjkljivosti, ki so del molekulskega sidranja, smo v nadaljevanju dva kompleksa ligand-protein uporabili pri Gregor Jug Uporaba molekulskega modeliranja pri načrtovanju antagonistov simulaciji molekulske dinamike (MD). Z metodama STD NMR in MD smo ponovno potrdili pomen osrednje manoze pri interakciji s Ca2+ ionom ter tako za vezavo v vezavno mesto proteina DC-SIGN. Prav tako smo določili pomembnost strukturnih elementov v stranskih verigah, ki dodatno prispevajo k vezavi na protein. Podatki kažejo, da so sicer te interakcije precej šibke, kar gre pripisati zelo plitvemu in slabo definiranemu vezavnemu mestu, kar predstavlja precejšnjo oviro pri snovanju novih močnejših antagonistov DC-SIGN. Zanimivost, ki smo jo razbrali iz rezultatov, so interakcije ligandov s hidrofobno ravnino v vezavnem mestu. Le-ta predstavlja dodatno možnost za vzpostavitev interakcij ter tako predstavlja smernico za nadaljnji razvoj antagonistov DC-SIGN. Predvidevamo, da bo na osnovi rezultatov doktorskega dela možno v prihodnje načrtovati antagoniste DC-SIGN z večjo afiniteto, pri čemer se bo kot uporabna metoda lahko uporabljal validiran protokol za molekulsko sidranje z orodjem FRED. Kot dobro izhodišče za nadaljnji razvoj bodo vsekakor služili tudi podatki in vitro testiranj načrtovanih spojin, kakor tudi študija interakcij antagonistov DC-SIGN z vezavnim mestom, ki smo jo izvedli s kombinacijo metod molekulskega modeliranja in STD NMR. Pri slednjem smo določili tudi novo področje vezavnega mesta, ki ga velja pri načrtovanju novih ligandov izkoristiti in tako izboljšati afiniteto vezave.

• Language: Slovenian
• Keywords: DC-SIGN, lektini tipa C, antagonisti, računalniško podprto načrtovanje, molekulsko modeliranje, sidranje, validacija, patogeni, virus HIV, infekcije, disertacije
• Work type: Doctoral dissertation
• Typology: 2.08 - Doctoral Dissertation
• Organization: FFA - Faculty of Pharmacy
• Place of publishing: Ljubljana
• Publisher: [G. Jug]
• Year: 2016
• Number of pages: VI, 221 str.
• PID: 20.500.12556/RUL-143763
• UDC: 004.9:615.2(043.3)
• COBISS.SI-ID: 285154048
• Publication date in RUL: 11.01.2023

Secondary language

• Language: English
• Title: Use of molecular modelling for design of DC-SIGN antagonists and their in vitro evaluation

Abstract

Infectious diseases are ranked as the second most common cause of deaths on the yearly basis, right behind the cardiovascular diseases. The most frequent among infections or causes for their development are HIV virus, Mycobacterium tuberculosis and malaria. It is known that different pathogens can enter the body via protein, named Dendritic Cell-Specific Intercellular Adhesion Molecule-Grabbing Non-integrin (DC-SIGN). The former is a C-type lectin located on the surface of dendritic cells (DCs), which besides HIV also binds various other pathogens. Upon binding to DC-SIGN pathogen is internalised in the phagocytotic bubble. Some pathogens, including HIV can »survive« in this destructive environment and consecutivelly travel with mature DCs through the lymph all the way to the lymphatic glands and further on through the body, which results in varios kinds of infections. In the scope of our PhD thesis we focused on the design of novel antagonists of DC-SIGN that would obstruct binding of glycoprotein gp120, which is expressed on the HIV surface, to DC-SIGN. Upon their binding these compounds would prevent HIV to enter the body via DC-SIGN and as such (probably) defend the organism from infection and consecutivelly from AIDS. We performed a structure-based in silico design of DC-SIGN antagonists. At the start of our research work there were no suitable molecular docking tools identified that could be used for molecular docking to the very shallow and featureless DC-SIGN binding site. Thus, as a first step of our experiment we decided to validate several molecular docking tools for docking to DC-SIGN: AutoDock, CDOCKER (included in the software Accelrys Discovery Studio 3.0), FlexX (included in the software LeadIT), FRED (included in the software package OpenEye) and GOLD. Validation was performed using three different crystal structures of DC-SIGN-ligand complexes (PDB codes 2IT6, 2XR5 and 2XR6), which were used for re-docking of ligands included in the crystal structures of above stated complexes. Results have shown that the most suitable molecular docking tool to be used for docking of ligands to DC-SIGN is FRED, followed by CDOCKER and FlexX. We showed that AutoDock and GOLD are not appropriate molecular docking tools to be used for docking to DC-SIGN CRD. To proof the usefulness of the established molecular docking protocol in FRED we performed a virtual screening of rationally designed library of potential DC-SIGN antagonists. As a starting point we selected 32 compounds, which were designed and tested using in vitro assay within our research group. For each of 32 compounds we additionaly generated 50 structurally similar compounds (»decoys«). As a result we got a library of 1632 compounds, which was used to evaluate the usefulness of molecular docking protocol, as stated above. Results have shown that all 32 compounds for which we have also successfully proven their binding to DC-SIGN ECD in the in vitro assay, were placed amongst the top 5.6% of all ranked library compounds. In addition, two compounds with the highest binding affinity (IC50 = 40 μM and 50 μM) were ranked fourth and fifth. This confirms that the established docking protocol is appropriate for use in the initial phase, i.e. virtual screening and structure-based design of ligands – DC-SIGN antagonists. We have demonstrated that the established molecular docking protocol in FRED is also convenient for docking to other lectins similar to DC-SIGN. To confirm this we performed docking of the ligand to the binding site of the protein Langerin (PDB code 3P5F), which is a member of the same group of C-type lectins and poses similar structure of carbohydrate recognition domain (CRD) as DC-SIGN. Binding site for carbohydrates is located on the extracellular domain of DC-SIGN and represents a part of carbohydrate recognition domain (CRD). Ca2+ ion that is present at the binding site is the most important element for carbohydrate bidning to DC-SIGN as it is crucial for establishment of coordination bonds between the ligand and CRD. In addition to Ca2+ ion, amino acid residues play an important role when it comes to establishment of ligand-protein interactions (e.g. establishment of H-bonds). Compounds were designed on the basis of data gained during a thorough study of binding site, which was done within our research group in the past. Binding site, which is very shallow and thus very problematic for computer-aided design using molecular docking, offers several unexploited structural attributes that we were planning to target to establish additional interactions between the protein and the ligand. The first option is a hydrophobic cave, which is located behind the Phe313. The second option, which proved as quite interesting during our research, is Phe313 itself because of the position of its aromatic ring which allows formation of π-π interactions with ligands possessing the aromatic rings. The third option represents a flat hydrophobic area of the binding site, which could be exploited to enhance the binding of the ligand to the binding site. The basis for the design of novel ligands was a core mannose, which is crucial for recognition and establishing interactions between the compound and Ca2+ ion and also interactions with some of the amino acid residues in the protein binding site. Based on the literature data and our own expertise and knowledge about the binding site topology, we added adequate linkers and functional groups to the mannose. We have designed ligands that have proper geometrical and electrochemical properties for a successful binding of the core mannose as well as good chance for interactions of substituents with the other relevant areas of the protein, especially Phe313. To perform the research work we needed a DC-SIGN protein, which is not commercially available. Thus, one part of the research work was carried out at the IBS institute (»Institut de Biologie Structurale«) in Grenoble, France. Prepared recombinant protein DC-SIGN ECD was used for execution of in vitro assay for evaluation of potential DC-SIGN antagonists and also during binding evaluation using STD NMR method. Our second goal was preparation of fluorescently labeled protein DC-SIGN ECD, which would be utilized during the development of a novel in vitro assay for evaluation of binding affinity of designed DC-SIGN antagonists. Designed compounds for which we had foreseen potentially strong binding affinity were synthesized and used for in vitro testing at the Chair for Pharmaceutical chemistry. A robust in vitro test was used which is based on the competitive binding of the compound and gp120 to the DC-SIGN ECD that was prior bound to a proper surface. Success of the designed compounds binding was checked with a biotin reaction and chemiluminiscence measurement. We calculated IC50 values for each individual ligand based on the strenght of its response. Within our research group we performed in vitro testing to determine the binding affinity of 32 novel compounds. Results for two compounds gave the IC50 values in the micromolar range (40 μM and 50 μM). At the last step of research work we performed detailed evaluation of interactions between four selected mannose based monovalent ligands and DC-SIGN CRD binding site. In the first step we used validated molecular docking protocols for FlexX and FRED to predict binding pose of all ligands' diastereisomers. Docking results have showed that FRED managed to correctly place the core mannose to the binding site and that it successfully predicted also hydrophobic and/or π-π interaction between the ligand and the side chain of Phe313. To suppres the deficiencies of molecular docking, we further performed a molecular dynamics (MD) simulations for the two selected ligand-protein complexes. Using STD NMR and MD we have once again confirmed the importance of core mannose for interaction with Ca2+ ion and consequently for the binding to the DC-SIGN binding site. Furthermore we have defined the importance of side chain structural elements which additionally contribute to the binding of the ligand to the protein. Data have shown that those interactions are rather weak, which could be related to the shallow and poorly defined binding site, which represents quite an obstacle for the design of novel more potent DC-SIGN antagonists. An interesting observation was made when reviewing the results, namely that ligands establish interactions with the hydrophobic flat part of the binding site. This represents possibility for additional interactions and could thus be used as a guide for future development of DC-SIGN antagonists. We believe that the results of this PhD thesis could be the basis for the future design of even more potent DC-SIGN antagonists during which the FRED validated docking protocol could be utilized as a useful structure-based design method. In addition, the results obtained from in vitro testing of designed ligands as well as study of interactions between DC-SIGN antagonists and binding site, performed using a combination of molecular modelling and STD NMR are a good starting point for future development. Using the latter we managed to define an additional part of the DC-SIGN CRD binding site, which should be taken into account during design of novel ligands and in such way possibly improve the binding affinity of novel ligands.

Projects

• Funder: ARRS - Slovenian Research Agency
• Project number: P1-0208
• Name: Farmacevtska kemija: načrtovanje, sinteza in vrednotenje učinkovin