accelerating the discovery of drugs for neglected...

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2020

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1932

Accelerating the discovery of drugsfor Neglected Tropical Diseasesusing biophysical methods

GIULIA OPASSI

ISSN 1651-6214ISBN 978-91-513-0943-9urn:nbn:se:uu:diva-408721

Dissertation presented at Uppsala University to be publicly examined in Room A1:111a,BMC, Husargatan 3, Uppsala, Friday, 12 June 2020 at 13:15 for the degree of Doctorof Philosophy. The examination will be conducted in English. Faculty examiner: JennySandmark (AstraZeneca).

AbstractOpassi, G. 2020. Accelerating the discovery of drugs for Neglected Tropical Diseases usingbiophysical methods. Digital Comprehensive Summaries of Uppsala Dissertations from theFaculty of Science and Technology 1932. 66 pp. Uppsala: Acta Universitatis Upsaliensis.ISBN 978-91-513-0943-9.

Even though they account for 10% of the global disease burden and besides the new recordof funding in 2018, tropical diseases are still neglected by the majority of pharmaceuticalcompanies and public funding. The reason can be mostly found in the fact that for thesediseases a conventional drug discovery process is often too expensive. The development of newapproaches in the early stage of drug discovery has therefore a key role in fighting NeglectedTropical Diseases (NTD). AEGIS is a European network which relates on collaborations fora multidisciplinary approach, in order to “accelerate” drugs development – hence reducing thecosts. As part of the network, this project is focused on a rational exploration of the chemicalspace, together with an in depth-analysis of molecular interactions for a better characterizationof targets involved in NTD.

One cost-efficient way for exploring pharmacologically relevant chemical space is fragmentbased lead discovery (FBLD). This approach requires an extensive understanding of the targetproperties; therefore, a pipeline of orthogonal methods was developed to validate the suitabilityof the target for FBLD.

The choice of a fragment library has a significant impact on the experimental strategy. Notonly the validation of the target, but also practical issues concerning technology, orthogonalvalidation and applicability have to be considered when initiating a fragment screeningcampaign.

Another rational approach for the discovery of new drugs is looking at the target structure,especially when considering protein complexes interaction. Through the structural analysisof the protein-protein interface, several short peptides derived from the binding partner wereanalysed in their interaction with the target both in vitro and in cell. This allowed to identify thekey sequence for the binding to the target and the internalization of the complex, both crucialinformation for a structure-based approach.

The internalization process of the target was characterized by a real-time cell binding assay(RT-CBA), revealing a higher level of complexity than what was previously described.

Implementing RT-CBA in the early stage is a strategy that might improve the successrate of drug development. The possibility to develop an intracellular time resolved molecularinteraction assay between small molecules and a model target fused to a fluorescent protein wastherefore explored.

From a fragment screening campaign to a structure-based approach, culminating with a realtime intracellular validation of hits, all the assays developed address the different possibilitiesfor accelerating the drug discovery process in the early stages.

Keywords: Neglected Tropical Diseases, Drug Discovery, Biochemistry, Biophysics, Kinetic,Fragment Based Lead Discovery, Assay Development

Giulia Opassi, Department of Chemistry - BMC, Box 576, Uppsala University, SE-75123Uppsala, Sweden.

© Giulia Opassi 2020

ISSN 1651-6214ISBN 978-91-513-0943-9urn:nbn:se:uu:diva-408721 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-408721)

“Work it harder, make it betterDo it faster, makes us stronger

More than ever, hour afterHour, work is never over”

Daft Punk, 2001

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Opassi, G., Nordström, H, Lundin, A., Napolitano, V., Magari,

F., Dzus, T., Klebe, G., Danielson, U. H. (2020) Establishing Trypanosoma cruzi farnesyl pyrophosphate synthase as a viable target for biosensor driven fragment based lead discovery Pro-tein Science, 29:991–1003

II FitzGerald, E.A.*, Opassi, G.*, Vagrys, D.*, Klein, H.F.,Ham-ilton, D.J., Talibov, V.O., Davis, B., Hubbard, R.E., Wijtmans, M., O’Brien, P.,Moberg, A., Lindgren, M., Holmgren, C., Dan-ielson, U.H. Experimental strategies for initiating fragment-based drug discovery using SPR biosensors. Manuscript

III Opassi, G., Encarnação, J.C., Napolitano, V., Dubin, G., Popowicz, G.M., Andersson, K., Danielson, U.H. Identification of a Shiga toxin A fragment with an ability to be internalized into Gb3 receptor carrying cells. Manuscript

IV Encarnação, J.C.*, Napolitano V.*, Opassi G., Danielson, U.H., Dubin, G., Popowicz, G, Buijs, J., Andersson, K., Björkelund H. Revealing the dynamic features of STxB-Gb3 co-internalization mechanism of molecular cargo into cancer cells Submitted

V Encarnação, J.C.*, Opassi, G.*, Mourão, A., Napolitano, V., So-rieul, C., Dubin, G., Danielson, U.H., Buijs, J., Andersson, K. Bioengineering living cells for measuring intracellular interac-tions of small-molecules in real-time Manuscript

Reprints were made with permission from the respective publishers. “*” Indicates equal contribution Publication not included in this thesis

VI Opassi, G.*, Gesù, A.*, Massarotti, A. (2018) The hitchhiker’s guide to the chemical-biological galaxy. Drug Discovery Today, 23 (3), 565-574

Contribution report

Paper I Planned the study, produced and purified protein, designed and performed quality evaluation of the protein via nanoDSF and DSL, developed and performed activity luminescence-based assay, developed and performed SPR-biosensor based assay, designed thermal shift screening assay, analyzed data, wrote the manuscript.

Paper II Produced and purified all species of FPPS, designed and per-formed kinetic experiments, analyzed data, co-wrote the man-uscript.

Paper III Planned the study, participated in peptides design, developed and performed SPR-biosensor based assay, developed and performed grating-coupled interferometry assay, analyzed data, wrote the manuscript.

Paper IV Participated in planning the study, produced and purified pro-tein, co-designed the biophysical studies, contributed to the writing of the manuscript.

Paper V Co-planned the study, designed protein production and puri-fication protocol for hCAII-eGFP in E. coli, performed quality evaluation of the protein, designed and performed kinetic ex-periments, analyzed the data, co-wrote the manuscript.

Contents

Introduction ................................................................................................... 11 Neglected Tropical Diseases .................................................................... 12 Rational vs high throughput approaches in ESDD ................................... 14 United we stand ........................................................................................ 16 Approaches and methods to accelerate ESDD ......................................... 17

Aim ............................................................................................................... 26

Present Investigation ..................................................................................... 27 Paper I ...................................................................................................... 27

Establishing Trypanosoma cruzi farnesyl pyrophosphate synthase as a viable target for biosensor driven fragment‐based lead discovery. ... 27

Background ..................................................................................... 27 Results ............................................................................................. 28 Discussion ....................................................................................... 33

Paper II ..................................................................................................... 34 Experimental strategies for initiating fragment-based drug discovery using SPR biosensors ........................................................................... 34


Paper III .................................................................................................... 37 Identification of a Shiga toxin A fragment with an ability to be internalized into Gb3 receptor carrying cells ....................................... 37


Paper IV ................................................................................................... 43 Revealing the dynamic features of STxB-Gb3 co-internalization mechanism of molecular cargo into cancer cells. ................................ 43


Paper V ..................................................................................................... 47 Bioengineering of living cells for measuring intracellular interactions of small molecules in real-time. ....................................... 47


Conclusion .................................................................................................... 51

Future at a glance .......................................................................................... 52

Popular summary/Sammanfattning ............................................................... 53

Acknowledgements ....................................................................................... 57

References ..................................................................................................... 61

Abbreviations

AChBP AEGIS BLI

Acetylcholine Binding Protein Accelerated Early stAge druG dIScovery BioLayer Interferometry

BE Binding Efficiency CA Carbonic Anhydrase CD Circular Dichroism DLS Dynamic Light Scattering DNDi Drugs for Neglected Diseases initiative DSF Differential Scanning Fluorimetry eGFP enhanced Green Fluorescent Protein ESDD Early Stage Drug Discovery EU European Union FBLD FP FPPS GCI

Fragment-Based Lead Discovery Fluorescence Polarization Farnesyl Pyrophosphate Synthase Grating-Coupled Interferometry

HTS High Throughput Screening ITC Isothermal Titration Calorimetry LE MD MS

Ligand Efficiency Molecular Dynamic Mass Spectrometry

MSCA Marie Skłodowska-Curie Actions MST MicroScale Thermophoresis MW Molecular Weight NMR NTDs

Nuclear Magnetic Resonance Neglected Tropical Diseases

PPDK Pyruvate Phosphate Dikinase PPi Pyrophosphate PPI Protein-Protein Interaction R&D Research & Development SAR Structure-Activity Relationship SPR Surface Plasmon Resonance STx Shiga Toxin STxA Shiga Toxin subunit A STxB Shiga Toxin subunit B tb Trypanosoma Brucei

tc TR-FRET

Trypanosoma Cruzi Time-Resolved Fluorescence Energy Transfer

WHO World Health Organization

11

Introduction

Through millennia the search for the Elixir of Life has been the supreme quest for many. This elixir was described as a potion that supposedly cures all dis-eases, granting eternal life. Alchemists in various ages and cultures sought the resources for formulating the elixir, and the French scribe Nicolas Flamel was one of the few believed to have achieved this important task during the 15th century. 1 From that point in history, mankind’s quest for longevity has been maybe less legendary but still quite successful.

Human lifespan has been calculated to be maximally 122 years, however only two hundred years ago our life expectancy was around 40. 2 But because of important discoveries and development, this number has increased dramat-ically since then.

New engineered ways to get clean water, knowledge about germs causing diseases, soap for washing hands and clothes, and new ways to preserved food improved both life quality and nutrition. 3–5 Pharmaceutical Sciences raised up, and the Elixir of Life started to look more like a combination of well dosed poisons – a φάρμακον indeed – than a mystical potion. Moreover, the discov-ery of antibiotics and vaccines development knocked out some of the worst killers that mankind has ever seen. 6,7

Penicillin discovery caused actually a shift towards patenting drugs and thereafter, pharmaceuticals began to invest heavily in the discovery of new products. Public funds also provided a boost to Research and Development (R&D), starting the golden age of pharmaceutical industry with its assets of drugs for each disease that afflicted humans for centuries. 8

From the 1800 life expectancy has therefore more than doubled, up to an average of 72 years in the current age. Yet the main reason for this huge change is the lower probability to die of an infection disease before reaching old age. Or at least, this is what happens in First World countries. 9

In the Tropics there are still a number of endemic infectious diseases that cannot almost be found anywhere else, and that are often quite horrific. Ken-neth S Warren, an American medical researcher in 1977 described them with terms as tropical and neglected, leading to the current definition of Neglected Tropical Diseases (NTDs). 10,11

12

Neglected Tropical Diseases There is some debate among the different agencies over which diseases are classified as neglected tropical diseases. The World Health Organization (WHO) defines NTDs as a diverse group of communicable diseases that pre-vail in tropical and subtropical conditions, affecting more than one billion peo-ple and that cost developing economies billions of dollars every year. 12 This definition includes infections from a wide range of different pathogens, such as bacteria, protozoa, viruses, amoeba and worms. NTDs are strongly corre-lated with poor hygienic conditions, lack of infrastructures and limited access to healthcare. The term neglected is used to indicate that there is no existing product to treat/prevent the disease, or that those products are poorly suited for use in low- and middle-income countries. 13 Most importantly there is no commercial market to stimulate R&D since NTDs are hidden among the ne-glected population of the world, and they have therefore been ignored for dec-ades by the pharmaceutical industry and public health coverage in general. 14

Moreover, there is an uneven funding distribution making some NTDs more neglected than others. HIV/AIDS, malaria and tuberculosis collectively received around two third of the total R&D funding (69%) for neglected dis-eases, while strictly endemic diseases, such as kinetoplastids and diarrhoeal diseases accounted for only 8%. 13 These “big three” infections are character-ized by a higher mortality and public awareness rates, which actually put a spotlight on them as global health priorities while other NTDs remain un-known to the masses. 15

Kinetoplastid infections for instance include two vector-borne diseases en-demic in Latin America and sub-Saharan Africa, which for this reason are known as American (or Chagas disease) and African Trypanosomiasis (or Sleeping sickness) respectively. 16,17 The causative agent of both Chagas dis-ease and Sleeping sickness is a parasitic trypanosome, which starts its life cy-cle in an insect host – the triatomine beetle and tsetse fly, respectively – and is then transmitted to humans. Both infections are characterized by local mul-tiplication in the human host, followed by dissemination and localization in target organs, in which they cause potentially lethal damage. 18

WHO placed both trypanosomiases in the list of candidates for control and elimination by 2020, 19,20 and, even if the first oral treatment for Sleeping sickness (fexinidazole) was released at the end of 2018 21, the situation is far from being resolved. 22–24 More 16 diseases, besides American and African Trypanosomiases, are included in the WHO list. 25 However, some diseases are not mentioned as they represent “minor threats” in terms of morbidity and mortality, although they still have a huge impact on the world population (Fig 1). 26 An example is given by diarrhoeal diseases, a group of illnesses that spread through contaminated food or water and affect mostly children under the age of five and immunocompromised individuals in developing world. 27,28

13

Shigellosis is one of the most contagious bacteria mediated diarrhoeal dis-eases, causing about 700,000 deaths per year worldwide. 29 The etiological agent of this NTD is Shigella dysenteriae, an enterobacteria that through fecal-oral route invades epithelial cells in the large intestine and produces entero-toxins, leading to cramps and severe diarrhea. 30,31 The first drugs used to treat severe shigellosis were antibiotic and antimicrobial agents (sulfonamides) for which Shigella developed resistance, and the treatment is therefore now shifted to antibiotic cocktails. 32 Multidrug-resistant strains have been reported as responsible for Shigella infection outbreaks, whereby new treatments for shigellosis have become a therapeutic challenge. 33

Understanding the biochemistry of causative pathogenic agents have al-lowed the identification of several potential biological targets for NTDs, set-ting the basis for developing safe and effective new drugs. A big problem in this scenario is the little incentive that pharmaceutical companies have to in-vest in NTD research, as the new drugs cannot be sold at a price that would cover the high R&D costs. One reason for high costs is almost certainly the high failure rate of selected molecules in pre-clinical studies. 14 Hence, accel-erating the early stage of the drug discovery (ESDD) process while imple-menting a rational optimization of multidisciplinary assays development plays

Fig 1 Estimates percentages and numbers for deaths globally in 2016 sorted by neglected dis-ease. The preparation of these statistics was undertaken by the WHO Department of Infor-mation Evidence and Research. 26 *Respiratory infections cases are reported here since Bac-terial pneumonia is part of the neglected diseases list from G-FINDER report, although the number includes a wider panel of respiratory infectious diseases.

14

a key role in reducing the number of false positive hits, and the consequent costs.

Rational vs high throughput approaches in ESDD The search for a new drug begins by finding molecules that “hit” the target of choice, or in other words that are able to interact with the target. The hit mol-ecule is then progressed towards a lead candidate through a series of structural optimizations, with the aim of optimizing affinity, selectivity and pharmaco-kinetic properties. This entire process – from hit discovery to lead optimiza-tion – defines the early stage drug discovery. When no hit or lead molecules are available for a determined target, or there is an interest in looking for new chemical scaffolds and/or allosteric binders, the first step usually consists of developing a screening assay for testing libraries of different compounds. 34

High throughput screening (HTS) is a well-established approach for an ef-ficient exploration of drug-like chemical space, an awkwardly vast universe of around 1060 unique structures. 35 The first papers on HTS appeared in Pub-Med in 1991, 36 and the concept quickly became dominant for lead-discovery in pharmaceutical industry. They started investing millions in building huge libraries (≥ 1.5 millions) of diverse drug-size compounds, coming from in house medicinal chemistry resources, combinatorial chemistry approaches and/or purchased from commercially available sources. 37 However, inde-pendent on the amount of money and time invested, the success rate of an HTS campaign is generally not much more than 50%, especially for challenging targets. 38 Moreover, even if initial expectations were that it would be possible to find a drug directly from a screening campaign, it was soon discovered that HTS typically identifies chemical starting points. Finally, the subsequent hit to lead medicinal chemistry process is often characterized by optimization dif-ficulties due to the often large and relatively lipophilic nature of screening hits. 39

Considering the premises, pharmaceuticals started to go smaller, shifting to what became an important alternative to HTS: Fragment Based Lead Discov-ery (FBLD). 40 Since fragments have a lower molecular weight, a fragment library can sample a greater portion of chemical diversity even when using small libraries (typically <2,000 compounds) 41,42, hence reducing expenses and screening time. The fundamental principles behind the design of drugs using fragments representing small building block can be traced back when William P. Jencks in 1981 described the non-additivity of the binding energies for different component parts of a molecule binding to a protein, and pointing out that “It is not unusual to find that the binding of individual molecules A and B is weak or negligible but A-B binds well, so that the whole appears to be greater than the sum of its parts”. 43 However, it was not until the technol-

15

ogies developed in the mid-1990s – able to detect low affinity/activity inter-actions – that it was finally possible to evolve drugs from a low-molecular-weight entities, via a step-wise addition of groups. 44 A rational process, which often involved structure-based design.

Although very influential and effective, HTS has often been considered as an approach that has undermined the creativity in drug design, the opposite to a “rational” approach. Elaborating a fragment into a drug is a very different process from making analogues from HTS hits, and requires a different level of creativity within medicinal chemistry. However, truth to be told, HTS is an intellectually neutral approach. Technologies originally developed for this massive screenings approaches have now been integrated also downstream with a lower throughput and suitable for lead optimization. 36 FBLD has today also become mainstream within pharmaceutical companies and even aca-demic groups, especially for its possibility to build in desirable properties. Still, the tendency – where possible – is to use a combination of approaches and integrated knowledge from all methods. 45 (Fig 2)

Fig 2 (a) Contributions of different techniques in Drug Discovery. Numbers obtained by a PubMed search of keywords containing terms referring to high-throughput screening, frag-ment-screening, and virtual screening focusing only in research article (b) Approved FDA drugs that were originated from FBLD from 2017 and development methods, collected from KEGG-DRUG website (http://www.genome.jp/kegg/drug/br08319.html?id=D01441); (c) Pubmed entries and published for 2017. NMR, Nuclear magnetic resonance; MD, Molecular dynamics; MS, Mass Spectrometry; SPR, surface plasmon resonance; DSF, differential scan-ning fluorimetry; BLI, biolayer interferometry; TR-FRET, Time-Resolved Fluorescence En-ergy Transfer; X-ray, Crystallography. Adapted from 46

16

United we stand There is a constant need for more rapid and cost-efficient methods for lead discovery and characterization. Multidisciplinary approaches have been rec-ognized as the best when addressing challenging targets, since they overcome the limitation of individual methods and reduce the number of false positive hits progressed to pre-clinical trials. Often this strategy has been carried through collaborations and partnerships, especially for unprofitable diseases such as NTDs. 47,48

A prominent example is the Drugs for Neglected Diseases initiative (DNDi), a non-profit R&D organization for driving medical innovation. It was launched in 1999 by the Nobel Prize winner “Doctors Without Borders”, in collaboration with WHO and public health research institutions from a num-ber of countries. 49 Their partnership with public, private, academic, non-profit and philanthropic sectors demonstrated the power of collaborative efforts by delivering eight new treatments for NTDs, including the first oral drug ever developed for Sleeping Sickness. 50

Public institutions, such as the European Union (EU) have also supported research into NTDs, bringing together large multinational research networks. 51 International networks are widely encouraged, and a clear example is given by Marie Skłodowska-Curie Actions (MSCA), a well-known EU research as-sociation that provides grants for all stage of researcher’s career. Accelerating Early staGe drug dIScovery (AEGIS) is a MSCA innovative training network to implement drug discovery in neglected and infectious diseases that are reemerging in Europe and are prevalent globally. The main resource of the AEGIS consortium is to combine a panel of different methods, developing new assays and techniques against challenging but promising targets. Moreo-ver, all the partners involved – academic and industrial – adhere to the open science movement, making research outcomes completely accessible (open access), and spreading awareness through popular science initiatives. 52,53 Both these efforts combined represent the means of accelerating early stage drug discovery.

17

Approaches and methods to accelerate ESDD Critical for a successful early stage drug discovery project is the ability to detect and validate hit molecules through the application of multiple orthogo-nal methods. The AEGIS platform combines recombinant proteins and in cell assays, in silico screening, biophysical and biochemical methods into rational approaches to accelerate the hit discovery-optimization on intended targets. Two different strategies are applied based on the availability of the 3D struc-ture of the targeted binding site (Fig 3).

Fig 3 Flowchart illustrating the AEGIS approach to the early stage drug discovery pipeline.

When the 3D structure of the potential drug binding site is available, it is pos-sible to specifically address the design of the new molecule with a structure-based approach. Virtual screening on large collections of chemical compounds enable the identification of hits that will be then verified experimentally. 54 Literature presents many successful examples of this approach, even for NTDs and challenging targets, such as Protein-Protein Interactions (PPI). 55

18

Alternatively, if the structure is unknown, or the aim is to identify ligands binding to previously undefined allosteric sites, the initial approach for hit identification is generally mediated by biophysical-based fragment library screening. 56 In this case, a library of compounds is tested against the target of interest and identified according to specific criteria defined specifically for the chosen technology. The molecular weight, availability and characteristics of the target, fragments library size and features, are all parameters which influ-ence the potential success of a screening campaign, and consequently the method of choice (Table I). 57 Since each technology relies on different spe-cific change that is observed in the read out – angle of resonance, magnetic resonance, thermal stability, depolarization of light, X-ray diffraction – it is quite common that there is little overlap among hits identified by different screening methods. For this reason, it is important to understand which kind of information can be given by the technology used, and what its limitations are. 58 In general, it is possible to identify five different types of information at this stage, which can be addressed with biophysical methods: 1. Does the ligand bind (Yes/No)? 2. How is the ligand binding? 3. Where is the ligand binding? All the biophysical methods described can detect binding, within certain dy-namic range. Among them, X-ray crystallography is the most elegant and in-formative way since it provides direct structural information, but it is also lim-ited to targets for which a robust crystallographic system is available. 59

For all the other methods, a general parameter for hit ranking and selection in the initial screening phase is the affinity of the ligand (L) for the target (T), expressed as dissociation constant (KD) at the equilibrium.

1

Where and are the free concentration of target and ligand, respec-tively, and is the concentration of the binary target-ligand complex. This interesting way of biochemists to express the affinity relationships in terms of dissociation – rather than a more intuitive association – found its rea-sons in the constant units. KD has actually units of molarity, and can therefore be equated with a specific ligand concentration that leads to half-maximal sat-uration of the available target binding sites. 60

The rationale behind ranking compounds based on the equilibrium dissoci-ation constant, calculated in vitro, is that KD can be sometimes directly related to the in vivo efficacy of a ligand. However, it has also been observed that a

19

crucial factor for sustained drug efficacy in vivo is not the apparent affinity of the drug for its target per se, but rather the duration of this binding, identified as instead the dissociation rate constant, koff. 61 The introduction of this “resi-dence time” concept, with all its limitations, still underlines the importance of understanding the kinetic of the mechanism. 62

Binding kinetics it is undeniably an important element in drug discovery, and knowing the individual on- and off-rates at a screening stage, may help understand modes of action, providing a qualitative analysis that gives useful information for hit prioritization, and further optimization.

In depth-characterization of the binding of hits to the target is the next step towards lead optimization. Considering the lower throughput needed, this pro-cess can be performed using a wider variety of technologies, which range from in vitro analysis to in cell studies (Table IIa-IIb). Applying an orthogonal validation strategy to get a better overview at this point is crucial, since the strengths of one biophysical method can compensate the weaknesses of an-other. Structural studies are often applied at this stage, in order to locate the binding site for the hit in the protein. Moreover, when the target of interest is an enzyme, activity assays can be applied to confirm the functional effect (typ-ically inhibition) of a compound already at the early stage of the discovery process. 63 The implementation of enzymatic assays already in screening cam-paigns seems actually to be the most straightforward approach in the identifi-cation of bioactive molecules. However, they can be misleading, especially while testing low affinity compounds, which might lead to false-positive ef-fect mediated by protein denaturation, colloidal aggregation, and interference with the detection methods. 64

Finally, measuring interaction in cells can elucidate the behavior of the lig-and in living system, providing information on other complex dynamic mech-anisms as receptor dimerization and internalization. 65

20

SC

RE

EN

ING

Experimental contribution SPR

Change of the refractive index, measured as a change in resonance angle or reso-nance wavelength upon binding of mole-cules on a sensor surface

Direct time-resolved determination of interaction, allows kinetic measurement and identification of on- and off-rate

Ligand based NMR Nuclear magnetic resonance chemical shift, given by NMR active nuclei, e.g.H1

Confirmation and affinity determination of ligand binding to unlabeled protein of any size, with the possibility to map the surface pocket for binding, and the bind-ing epitope for the ligand. A specific setup allows calculation of residence time.

DSF Indirect monitor of thermal folding with a fluorescent reporter, using qPCR

Identification of low-molecular-weight ligands that bind and de-/stabilize puri-fied proteins (∆Tm)

X-ray crystallography Structure determination by X-ray diffrac-tion of target crystals

Direct visualization of target-ligand complex, and consequent localization of ligand binding site

FP Perturbation of molecular movement of fluorophores in solution upon binding

Quantitative versatile analysis of di-verse molecular interactions and en-zyme activities at equilibrium

21

Pros Cons Ref

• High-sensitivity method – allows FBLD • Generally low false positive and nega-tive rates • Low amount of target required (~15 nmol protein for assay development and screening of 2000 compounds)

• Difficult discrimination between specific and non-specific binding • Difficult data analysis • Requires stable protein target over time • Expensive material • Immobilization procedure might af-fect the target • Measurements are affected by sol-vent (e.g. DMSO) and buffer mis-match

57,66,67

• Small molecules detection – allows FBLD • Protein and ligand unlabeled • KD determination of weak binders, till 10 mM

• Ligand concentration 10xKD or 100 times more than protein • Limited structural information • Difficult discrimination of promis-cuous binding due to compound ag-gregation • Precipitation/unfolding of the target leads to both false positive and false negative

68–71

• Fast (1000 runs per day) • Homogeneous • Functional knowledge of target not nec-essary • Low protein consuming (80 pm protein per analysis) • Easy data interpretation

• Artifacts easily occurs because of fluorescence quenching or aggrega-tion • Non suitable for disordered protein • ∆Tm values are not always a reflec-tion of ligands relative affinities

72–74

• Low false positive and negative rates • High range of affinity range, < 1M

• Limited to targets for which a robust crystallographic system is available • High protein consumption (~5 nmol for 96-wells crystallization plate)

75–77

• Inexpensive • Homogeneous • Measurement do not destroy samples, so it can be re-measured • Low background • Applicable to disordered proteins

• Fluorescent label required • Auto-fluorescent and light scattering can interfere with the measurement • Difficulties to detect weak binders

78,79

Table I. Methods and technologies used for compounds screening. SPR Surface Plasmon Res-onance; NMR Nuclear Magnetic Resonance; DSF Differential Scanning Fluorimetry; FP Fluorescence Polarization.

22

LE

AD

DIS

CO

VE

RY

AN

D O

PT

IMIZ

AT

ION

Experimental contribution Protein observed NMR

Nuclear magnetic resonance chemical shift, given by NMR active nuclei la-belled protein

Residue-specific, structural information about the binding interface, and quantita-tive affinity determination

AlphaScreen Singlet oxygen bead-based transfer upon interaction, which results in a non-radioactive amplified luminescent signal

Quantification of EC50 value, through competition with labelled/unlabeled mole-cule

ITC Heat released/absorbed upon molecu-lar interaction at constant temperature and pressure

Direct determination of binding thermody-namic, and consequent quantification of affinity. Allows kinetic determination (ki-nITC)

MST Motions of fluorescent molecules through a microscopic temperature gradient

Estimation of KD value from dose-re-sponse curve at different ligand concentra-tion

23

Pros Cons Ref

• Reliable determination of KD • Low false positive and negative rates

• Required a high amount of isotop-ically labelled protein • Large library screening only in multiplexing mode • Suitable for protein < 40kDa be-cause of lack of sensitivity

80,81

• Homogeneous • Easy to perform • Low background • Sensitive • Automatable • Versatile

• Expensive materials • Sensitive to light exposure • Singlet oxygen can be sequestered by compounds in screening libraries that can scavenge radical oxygen • Donor bead photo bleaching • Require a high-energy laser exci-tation source • Difficulties to distinguish specific interaction to co-precipitation/mis-folded protein

82–84

• Easy to use • High sensitivity

• The interaction needs an enthalpic component • Very high protein consumptions • Required high solubility of titrated component • Complex binding interactions are hard to evaluate

85–87

• Applicable to membrane proteins • Immobilization-free system • Big range of KD • Low compound consuming • High-throughput • Fast • No limitations on molecular weight ratio of the binding partners

• Requires fluorescent labelling or strong intrinsic fluorescent proteins • Aggregates can affect the meas-urement 88–91

Table IIa. Methods and technologies used for lead discovery and optimization. NMR Nuclear Magnetic Resonance; ITC Isothermal Titration Calorimetry; MST MicroScale Thermophore-sis.

24

LE

AD

DIS

CO

VE

RY

AN

D O

PT

IMIZ

AT

ION

Experimental contribution

Rotating cell-based ligand binding assay Real time radioactivity/ fluorescence intensity changes on rotating immobi-lized living cells

Direct time-resolved determination of in-teraction in living system. Allows kinetic measurement, and identification of on- and off-rate closer to real environment.

GCI Phase shift induced by changes in the refractive index within an evanescent field near a sensor surface

Direct time-resolved determination of in-teraction, that allows kinetic measurement

SPR* X-ray crystallography*

25

Pros Cons Ref

• Not costly consumables • High KD range measurable

• Compounds needs to be labelled • Low sensitivity • Highly compound consuming • Low throughput

65,92

• High sensitivity method – allows FBLD • Assess drug performance in human blood sample and plasma serum • High KD range measurable • No-clogging risk with either large mole-cules (e.g. liposomes) nor viscous sol-vents and additive

• Highly sensitive for wavelength in-stability • Measurement are affected by sol-vents and buffer mismatch • Expensive material

93–96

Table IIb. Methods and technologies used for lead discovery and optimization. GCI Grating-Coupled Interferometry; SPR Surface Plasmon Resonance. * SPR and X-ray crystallography are described in Table I.

26

Aim

The aim of this PhD project was to develop methods and strategies that can accelerate the discovery of drugs for neglected tropical diseases. It has primar-ily involved the development of biochemical and biophysical assays and using them for fragment library screening and characterization of hits and leads, and in-depth characterization of interactions.

This was divided into following sub-goals: Validating targets in NTDs as suitable for fragment based lead dis-

covery approaches Developing novel SPR-biosensor based assay strategy for initiating a

FBLD campaign, based on library size and information needed (e.g. identification of selective fragments for different target species)

Characterizing interaction kinetic of protein-protein interactions us-ing orthogonal methods, both in vitro and in cell

Understanding the biological complexity upon binding of cargo pro-teins

Exploring the possibility to depict molecular binding intracellularly using Real Time Cell Binding Assay (RT-CBA), combined with in vitro biophysical methods

In brief, this project aimed to provide methods and data that can support the identification of new lead scaffolds for selected NTDs targets, and the optimi-zation of lead compounds.

27

Present Investigation

The development of a new drug is a process that usually takes more than a decade, and comes with a very costly price tag with inherent risks of failure. New methods for accelerating hits discovery (e.g. fragment based) started to change the way of how we approach and perform drug discovery in the early stages. However, only “speeding up” the process cannot account for false pos-itives that end up in pre-clinical studies, causing a massive loss of time and resources. Resources which we do not often have for studying tropical dis-eases, as trypanosomiasis and shigellosis, for which the R&D funding cover an infinitesimal portion of the total amount designated to NTDs research.

From this perspective, implementation of multidisciplinary approaches, to-gether with more rational optimization of assays, plays a crucial role already at the early stages.

The work presented in this thesis considers different aspects of accelerating the discovery of new drugs, while reducing the risk of wasting time on false positive hits through the application of complementary methods.

Paper I Establishing Trypanosoma cruzi farnesyl pyrophosphate synthase as a viable target for biosensor driven fragment based lead discovery. When applying biophysical screening methods for identifying new hit scaf-folds, it is essential to have an extensive understanding of the physicochemical properties of the target and a valid biochemical method for studying both the protein behavior and the effects of the compounds that may be evolved into drugs. This study established methods for the discovery of new drugs for American Trypanosomiasis treatment, targeting farnesyl pyrophosphate syn-thase from Trypanosoma cruzi.

Background Farnesyl Pyrophosphate Synthase (FPPS) is a regulatory enzyme with a key role in the isoprenoid biosynthetic pathway, one of the most targeted routes for discovery of new drugs against trypanosomiasis. 97 Specifically, tcFPPS

28

has been validated as a promising target for drugs development due to its role in the synthesis of essential components for the viability of Trypanosoma cruzi. 98 Nitrogen containing bisphosphonates – clinically used inhibitors of human FPPS – showed antiproliferative and cytocidal activities against the parasite. 99 Current efforts are mainly focused on a structure-activity relation-ship (SAR) driven approach to improve the selectivity of the bisphosphonates for the T. cruzi enzyme. 100 Bisphosphonates are known to have specificity for bone tissue because of their high affinity for calcium. This has led to the hy-pothesis that they could be selective also for trypanosomal calcium-rich orga-nelles and thereby accumulate in the acidocalcisome of the parasite. 101

This paper presents a new strategy to tackle tcFPPS, aiming for a broader chemical exploration by developing methods for identification of new inhibi-tors scaffolds, with high potential for efficient and safe drugs. The chosen ap-proach was fragment based lead discovery (FBLD). It is a relatively fast and cost-efficient strategy, but requires methods with high sensitivity to detect and discriminate fragment hits, which typically have weak affinities. 63 The possi-bility to use FBLD for this target was explored through the application of dif-ferent methods. Two fragment screening assays were developed and tested using a small library.

Results tcFPPS produced essentially according to published protocols revealed a cer-tain instability and loss of function during the time required. Initial work was therefore focused on optimizing procedures and conditions for storage and handling high-quality protein for FBLD. Through Differential Scanning Flu-orimetry (DSF) analysis it was possible to identify the optimal buffer for stor-age. The combination of DSF and other biophysical methods – nanoDSF, Cir-cular Dichroism (CD) and Dynamic Light Scattering (DLS) – confirmed that the revised protocols resulted in the production of a non-aggregated dimeric protein (84kDa), stable over time, and with a correctly folded secondary struc-ture. DSF and nanoDSF were also applied to evaluate the stability of the pro-tein in conditions selected for screening analysis, confirming that the protein was structurally intact. (Fig 4)

29

Fig 4 Orthogonal characterization of produced tcFPPS. (a) Differential scanning fluorimetry (DSF) analysis of effect of storage buffer composition: 25 mM TRIS (blue) and 25 mM HEPES (orange) at pH 6.5, 7.0, 7.5, and 8.0 (from dark to light shade). (b) nanoDSF analysis of sta-bility of tcFPPS stored in 25 mM TRIS–HCl buffer at pH 6.5 and 75 mM NaCl, supplemented-with 1 mM TCEP at 4° C. Freshly thawed sample (solid line) and after 1 week from thawing (dashed line). (c) DSF analysis of effect of additives in the running buffer used for surface plasmon resonance analysis (10 mM HEPES, 150 mM NaCl, 3 mM Mg2+-acetate) on the sta-bility of tcFPPS. Inclusion of 1 mM TCEP alone (bottom, medium green), 0.05% Tween20 alone (top, dark green), or both 1 mM TCEP and 0.05% Tween20 (middle, lightgreen). (d) nanoDSF analysis of stability of tcFPPS in sodium acetate pH 5.0 (used as preconcentration solution for immobilization on sensor surface). Freshly protein sample (solid line) and sample thermally inactivated through incubation at 92° C for 5 min (red dotted line). (e) Circular dichroism spectrum of tcFPPS diluted in water to 0.2 mg/mL. (f) Dynamic light scattering analysis of stock solution

A real time bioluminescence-based assay was developed for evaluating the catalytic activity of produced tcFPPS. It consisted in a three-enzyme coupled assay, which monitors the amount of pyrophosphate (PPi) produced in the tcFPPS reaction, via two steps which convert formed PPi into a light signal. All the three enzymes were catalytically active in the buffer selected. How-ever, conditions were optimized prioritizing the two coupled enzymes (PPDK and luciferase) so that the reaction catalyzed by tcFPPS represented the rate limiting step, as confirmed by control experiments. (Fig 5)

30

Fig 5 Scheme of coupled enzymatic assay for analysis of tcFPPS activity. tcFPPS reaction: Synthesis of farnesyl pyrophosphate (FPP) and pyrophosphate (PPi) from geranyl pyrophos-phate (GPP) and isopentenyl diphosphate (IPP). Pyruvate phosphate dikinase (PPDK) reac-tion: Conversion of PPi into ATP using AMP, and with the associated conversion of phosphoe-nolpyruvate (PEP) into pyruvate (Pyr) and phosphate (PO42−). Luciferase reaction: ATP drives the conversion of luciferin into oxyluciferin using O2 and Mg2+. This is followed by the spontaneous reaction that generates light. AMP and pyrophosphate are recycled into the PPDK reaction.

To confirm that the enzyme was suitable for FBLD, two methods for screening a small fragment library were used: SPR biosensor technology and DSF. Screening conditions were selected with respect to physiological relevance, in terms of both ionic strength and reducing environment.

An SPR-biosensor based assay suitable for fragment screening against tcFPPS was developed and used in independent screening experiments with preparation of different surfaces. An attempt to immobilize thermally un-folded protein confirmed that the target needed to be structurally intact for immobilization. Moreover, a catalytic assay was performed on a manually im-mobilized enzyme on the chip surface outside the instrument – using the same procedure as in the instrument – confirming the target activity even after im-mobilization.

SPR-biosensor experiments run with bisphosphonates resulted in a high de-gree of secondary effects, which indicated that the immobilized enzyme was functional but also that the inhibitors were not suitable as reference. Therefore, the threshold for definition of hits was set to 30% of a theoretical Rmax (RNORM) for each fragment, calculated as following.

⁄ 2.1

31

where 2.2

Of the 90 fragments screened against tcFPPS, 16 hits (18%) were identified in the affinity screen, of which five (5%) were then confirmed through analy-sis of a concentration series. Compound ranking was determined using both Ligand Efficiency (LE) and Binding Efficiency (BE).

LE assesses the apparent binding affinity (KDapp) in relation to the number

of heavy atoms in a molecule, providing a way to compare the affinity of mol-ecules corrected for their size.

/1000 / ∗ 3

where R is the gas constant 8.31451 J/mol K, T is the temperature in Kelvin and nHA is the number of non-hydrogen atoms in the sample molecules. BE was calculated from saturation curves, as the slope of the linear relation-ship between complex concentration and ligand concentrations at very low ligand concentrations; resulting in an estimation of how much compound is bound as solely function of concentration. (Fig 6)

32

Fig 6 Steady-state analysis of interactions between Fragment Hits 1–5 and tcFPPS. Solid curves are theoretical saturation curves based on nonlinear regression analysis of reference subtracted and molecular weight normalized (RNORM= RU/Rmax) data and a simple 1:1 inter-action. The dashed lines represent the slopes of the graphs at low ligand concentrations, from which the binding efficiency (BE) was estimated. LE values calculated from KD

app and BE es-timated are reported in the table.

The same screening was performed in the absence of the cofactor (Mg2+), re-sulting in 25 fragments above the threshold, suggesting a higher degree of interactions.

The same conditions were applied when screening by DSF. The threshold for hit selection was set to ∆Tm ≥ ±1 °C from a DMSO reference, which lead to the identification of 35 hits both in the presence and absence of Mg2+, with the overlap of 26 fragments between the two conditions. (Fig 7)

33

Discussion This paper describes a rational approach to FBLD and hit validation when reference compounds are not available – or not suitable.

High-quality protein and conditions ensuring its functionality in the screen-ing experiments appeared to be crucial. Application of orthogonal biophysical methods for target handling allow a better control of the experimental condi-tions, increasing the reliability of the results while reducing the risk of arti-facts. As a complementary tool, the newly developed catalytic assay is useful for both confirming the functionality of the enzyme, and for discriminating binders from inhibitors.

Two screening methods were here compared – SPR biosensor analysis and DSF – which resulted in a different hit rates, with a limited overlap. The dif-ferences in the results can be attributed to the different principles for two meth-ods for selecting hits. DSF only detects hits affecting the unfolding process of the protein, while SPR identifies any small changes in the density proximal to the biosensor surface. Nevertheless, both methods exhibited differences in hit rates and/or signals while run in presence or absence of the cofactor, support-ing that the structural integrity of the protein under the two conditions is very different.

Another issue addressed in this paper was the ranking of low affinity inter-action binders. Although it was technically possible to estimate KD values and calculate LE, the actual number would not be relevant since the conditions were far from steady state. BE was chosen as a better parameter to rank hit compounds, as a measure of the ligand binding without assuming a mechanis-tic model neither stoichiometry.

Fig 7 Data from primary screening of fragment library against tcFPPS using SPR and DSF, in (a) presence of Mg2+ and (b) absence of Mg2+. Data represent normalised SPR signals (RNORM= RU/Rmax). Fragments with signals above the 0.3 RNORM threshold in a) were defined as hits. Final hits (compounds 1-5) are shown as dark green filled circles, eleven additional initial hits are shown as light green circles. These hits have the same coloring in b).

34

This first study therefore established tcFPPS as a suitable target for FBLD, using an unconventional referencing method and BE for ranking weak bind-ers. These findings will lead to the possibility of discovery new classes of inhibitor against the parasitic enzyme.

Paper II Experimental strategies for initiating fragment-based drug discovery using SPR biosensors

Besides target stability and characterization, different practical issues have also to be taken in consideration before initiating a fragment-based screening campaign. Among all, the choice of a fragment library has a significant im-pact on the experimental strategy, determining which technology to consider, but also which control experiments and orthogonal validation would be more suitable. This study addressed different experimental strategies for initiating a FBLD campaign with two different sized libraries, across five challenging target classes to answer different biological questions – understanding ion channel mechanisms, finding species specificity, comparing surfaces in disor-dered regions.

Background FBLD is currently widely practiced in pharmaceutical companies, comple-menting the well-known and explored high throughput screening (HTS) meth-ods. Due to their small size (up to 300Da) fragments tend to “fit” into adaptive regions of a protein, becoming an essential tool while looking at challenging targets for HTS – such as protein-protein interfaces, and allosteric sites. 102 Moreover, compared to the HTS of millions of drug-like compounds, FBLD reduces the chemical space tremendously (from 1060 to 109 molecules) allow-ing an efficient exploration of the latter within smaller compound libraries. 63

Library sizes can vary from 100 to 100,000 fragments. Fragment libraries containing a few hundreds of molecules are usually the best choice for low throughput methods such as crystallography. 103 Such libraries are therefore usually characterized by compounds with high solubility. Larger libraries can have a better structural diversity, with an optimal coverage of the pharmaco-logically relevant chemical space estimated to be around 2,000 fragments. 42

In this work, two different sized libraries were screened against challenging targets using SPR biosensor technology. 66,104 A library of 1056 fragments (FL1056) was used to tackle an ion channel homologue (Acetylcholine-bind-ing Protein, AChBP), protein-protein interaction (epigenetic target LSD1) and semi-/disordered targets (PTP1B and Tau, respectively). A small library of 90

35

compounds (FL90) was screened against three species of the same metalloen-zyme (tcFPPS, tbFPPS, hFPPS), looking for unique allosteric sites. Different orthogonal methods (NMR, crystallography) were explored to cross-validate hit molecules.

Results Target stability in buffers used for immobilization was initially evaluated via nanoDSF – except for the intrinsically disordered protein Tau. Different im-mobilization methods – amine coupling, His-capture, cross-linked His-cap-ture, thiol coupling and streptavidin coupling – were tested in order to account for possible interferences with the binding sites, and two different strategies were applied based on the library of choice. (Fig 8)

The workflow for screening of FL1056 included an initial Clean Screen, which removed 1% of extremely promiscuous binders from the library. The Clean Screen was then followed by a Binding level screen, where 10% of fragments for each target was selected and an Affinity screen that reduced the final SPR hit of more 80%. Due to the low number of fragments and their intrinsically high solubility, a Clean Screen was not performed for FL90, reducing the compound consump-tion and time required for the experiments. The Binding level screen for FL90 identified ~15% of binders for each of the three enzymes (14 for tcFPPS, 11 for tbFPPS and 13 for hFPPS), and 5% after Affinity screen were confirmed as hit. Final hit molecules previously identified for tcFPPS (Paper I) were re-confirmed also in this study. Four fragments were found to be selective for a specific species of FPPS, and two final hits were selective for the trypanosome enzymes.

When suitable tool compounds were available, they were used at saturating concentrations to determine the sensitivity and stability of the surface over time. They also enabled the calculation of a theoretical Rmax (Eq.2.2) for hits.

Fig.8 Targets and workflow applied to respectively FL1056 (a) and FL90 (b).

36

In the absence of suitable reference compounds (FPPS, Tau), other strategies were employed to verify the surface integrity. For example, Binding level screening with the FPPS enzyme variants was performed both in presence and absence of the cofactor. The two conditions resulted in a different hit rates, indicating different functionality of the immobilized targets. A different strat-egy was instead to Tau, where a biotinylated version of the protein highlighted the importance of an oriented immobilization approach.

Different hit selection and ranking methods were proposed. For FL1056, 10% of the compounds were selected after Binding level screen, ranked on the basis of response levels and that exhibited a characteristic square-shape curve. In contrast, fragments with signals between 30% and 100% of their theoretical Rmax were selected after screening of FL90. Sensorgrams quality and a clear concentration dependence were required features for selection of hit in the Affinity screen for both libraries. Hits were ranked on the basis of their affinity (KD

app) and BE, common parameters for ranking of fragments. Orthogonal validation of hits was carried through using NMR (Tau), and

crystallization studies are still ongoing (AChBP, LSD1, FPPS).

Discussion This study described a rational strategy for fragments screening against chal-lenging targets using SPR, considering practical issues – library of choice, experimental strategy, hit ranking, orthogonal validation – and highlighting its applicability to both academic and industrial endeavors.

The best relevant chemical space coverage is obtained within ~2,000 frag-ment library. Libraries of that size – as FL1056 – are considered “relatively small” and are quite common in small pharma and biotech companies. Even if they are more manageable than conventional compound libraries in big pharma, they still require proper handling and instrumentation for a through-put required for efficient screen the entire library against an immobilized tar-get protein. Placing reference compounds along the way is also a good strategy to verify the stability of the surface over time.

Commercially available fragment libraries can be seen as an alternative for academic researchers, facilitating compound handling and affordability. Here a fragment library of ca. a hundred fragments for crystallographic screening was explored. Since the library was designed for crystallography, the com-pounds have a high solubility, favorable characteristic also for SPR and other biophysical technologies. Even in a library of this small size, the chemical diversity was sufficient for identification of hits for both trypanosome and hu-man FPPS.

The size of the library is important also from a methodological perspective, as it influences the hit selection process. When screening FL1056, fragments were initially ranked on the basis of signal intensity and binding behavior. The top 10% responses were defined as hits. When screening a smaller library – as FL90 – it is possible to analyze the interactions more in details already

37

directly after the initial Binding level screen. Signals were normalized for each fragment based on their theoretical Rmax, and hits were defined as compounds having good quality sensorgram and a RNORM between 0.3 and 1, excluding both weaker interactions and super-stoichiometric binders.

All SPR based fragment screening campaigns initiated here resulted in the selection of 4 to 11 (19 for Tau) hits per target, with no significant differences between library A and B. This is a clear indication that both libraries were adequately designed for the identification of few specific scaffolds for drug development. In addition, challenging targets are not inherently problematic for a fragment screening campaign and lack of suitable references does not have to be a problem, even though additional assay development and creative experimental design may be required.

In conclusion, this study demonstrated that a rational approach to the ex-perimental design can lead to the identification of fragment hits also for chal-lenging targets, regardless the size of the library employed. These findings suggest different approaches when setting up an SPR-biosensor based FBLD campaign based on library sizes and information needed (e.g. identify selec-tive fragments for different target species).

Paper III Identification of a Shiga toxin A fragment with an ability to be internalized into Gb3 receptor carrying cells

If the target of interest is a structurally characterized protein-protein interac-tion (PPI), an efficient approach for drug discovery is structure-based. Through the structural analysis of the interface between the two binding part-ners it is actually possible to determine the key residues for the interaction, and therefore design molecules that can disrupt it. This work is the first ex-plorative kinetic study of a PPI target for shigellosis, where the key sequence for the binding is identified and characterized both in vitro and in cell.

Background Shiga (STx) or Shiga-Like toxins are bacterially produced proteins that be-longs to the AB5 family, a multi-protein complex composed by a subunit with enzymatic activity (A) and a homopentamer carrier (B). 105 The release of the toxin by intestinal pathogenic bacteria, such as Shigella dysenteriae, triggers an acute infection called shigellosis, an NTD that mostly affects young chil-dren in less developed countries. 106 STxB acts as a carrier that delivers the STxA into the cell by binding with Gb3, and subsequently becoming, inter-nalized together with its cargo. Gb3 is a glycosphingolipid with restricted tis-sue expression (e.g. intestinal epithelium), but is also over-expressed in certain

38

tumor cell lines. For this reason, the use of STxB for drug delivery has been suggested for cancer therapy. 107,108

Different studies have resolved the mechanism for intracellular trafficking of the toxin, revealing a complex journey through early endosomes, Golgi, cytosol and endoplasmatic reticulum, where STxA can finally be released from its B-partner by a furin-cleavage and reduction of a disulphide bond. Although the protein-protein interface of the complex was described already in 1994, 109 there is few (or no) information about the interaction mechanism for the two proteins.

This study explored the interaction between STxB and six STxA derived peptides of different length, rationally designed for the mechanistic analysis of the StxB-StxA complex interaction. It was characterized analyzed in vitro using different biophysical methods (SPR, Grating-Coupled Interferometry), and complemented by experiments in cells (confocal microscopy) in order to confirm the internalization of the complex.

Results An SPR biosensor-based assay for characterization of the binding mechanism between immobilized STxB and short peptides derived from STxA was de-veloped. (Fig 9) Interaction studies with peptides if two different lenghts – STAS and STAL – initially revealed that it was not possible to detect binding of the shortest peptide (STAS) to STxB. Based on the STAL sequence, six new peptides were designed, where the structure of the C-terminus (native, or amide-modified) and the N-terminus (native, or acyl-modified) were consid-ered

Fig 9 Graphical representation of the longest STxA derived peptide analyzed. PDB 1DM0

Analysis of the secondary structures of these peptides by via CD revealed that peptide 3 to 6 present a defined secondary structure. (Fig 10) However, only

39

peptide 5 and six – which contains both beta-strands and not only the alpha-helix –were the ones which interacted with immobilized STxB at all the dif-ferent physiological pHs, with similar affinities. Furthermore, a remarkable difference in the binding between the two peptides was found in the singular kon and koff values, in both SPR (Fig 11) and interferometry experiments (Fig 12). While peptide 5 was clearly bound to STxB, it had a relatively fast asso-ciation and dissociation rates, peptide 6 had a slower dissociation rate. So alt-hough the additional four C-terminal residues in peptide 6 were not seen in the crystal structure of the STxB-STxA complex due to high flexibility, they nevertheless seem to stabilize the interaction significantly.

Fig 10 Circular dichroism spectra of STxA-derived peptides, diluted in water to 100 µM. Spec-tra for peptides 1 and 3 are not included since their sequences are identical to those for peptides 2 and 4, respectively

40

Fig 11 Analysis of STxA derived peptides 5 and 6 using SPR biosensor assay. a) Sensorgrams of peptides injected in a 3-fold concentration series, up to 10 μM, represented as Signals (RU) over Time (s). Overlays (black lines) represent theoretical curves from fitting data with 1:1 binding (top) and 1:1 Two State (bottom) models. b) Interaction parameters and assay data estimated from the fitting in a)

41

Fig 12 Analysis of STxA derived peptides 5 and 6 using GCI biosensor assay. a) Sensorgrams of peptides injected in a 5-fold concentration series up to 1 µM, represented as Surface Mass (pg/mm2) over Time (s). Overlays (black lines) represent theoretical curves from fitting data with 1:1 binding (top) and Conformational Change (bottom) models. b) Interaction parameters and assay data estimated from the fitting in a)

42

Peptide 6 was consequently selected for in cell studies using Ramos cells, a cancer cell line that overexpresses the Gb3 receptor. Confocal experiments confirmed that the peptide is internalized in presence of its binding partner STxB, whereas no penetration was observed for the peptide alone (Fig 13).

Fig 13 Live imaging of Ramos cells incubated with STxBATTO590-Peptide6FITC 1h. ATTO590 fluorescence by binding of STxB-ATTO590 to the Ramos cells (A), FITC fluores-cence from Peptide 6 (B), and overlap of the two fluorescence images (C). FITC fluorescence from Peptide 6 (D), bright field caption of Ramos cells (E), and overlap of the two fluores-cence images (F).

Discussion This work described a rational approach to the characterization of a PPI target based on the structure of a complex, with the identification of the critical re-gion for the binding and the effect of pH.

Peptides derived from the C-terminal of STxA were designed with respect to the X-ray structure to both define the minimal length for the binding and to understand if it should have free or modified end groups – important for the formation of the alpha-helix and maybe influencing the interactions with STxB. It was revealed that the presence of the two beta-strands is fundamental upon the binding, and that the final flexible residues not seen in the crystal structure play a key role in stabilizing the interaction. The small effects on kon and koff values when varying the pH did not show obvious trends. More gen-erally, the peptides showed very similar affinity for STxB at every pH, con-firming the complex stability at the different physiological conditions encoun-tered by the toxin.

43

Confocal experiments suggested that peptide 6 can be internalized and re-leased intracellularly only in the presence of STxB. This indicates that the sequence identified with peptide 6 is crucial for both binding and internaliza-tion. Molecules/peptides capable of disrupting the interaction between this re-gion and STxB would be able to inhibit the uptake of the toxin into the host.

This study represents the first kinetic characterization of the Shiga toxin complex assembly, identifying the key region of STxA for both binding and internalization process, and putting a spotlight on its complex interaction mechanism.

Paper IV Revealing the dynamic features of STxB-Gb3 co-internalization mechanism of molecular cargo into cancer cells.

For complex biological systems – such as STxB carrier protein – an in vitro characterization might not be enough to grasp the key event for the target role. Using in cell assays can be a possibility to overcome the limitations of in vitro studies and get a better understanding of the system. Internalization into the host cells is one of the strategies that allow this pathogen to start an infection while escaping defense mechanisms. An event such as internalization cannot clearly be studied only in vitro. This study describes the complexity of the Shiga toxin B internalization process, highlighting the importance of not only in-cell studies, but also the fourth dimension: time.

Background STxB is the carrier pentameric subunit of the Shiga toxin (STx). It binds spe-cifically to Gb3 receptors on mammalian cells and is subsequently internal-ized by endocytosis, starting the shigellosis infection. Considering the stability of STx in different physiological environments and its ability to penetrate tis-sue, together with the overexpression of Gb3 in tumors like lymphomas and colorectal carcinomas, STxB has been identified as a promising carrier for specific delivery of small peptides and molecules to cancer cells. 110

Different aspects of the STxB-Gb3 interaction have therefore previously been investigated, from the in vitro characterization by SPR 111 – even in a lipid bilayer environment – to the ability of the toxin to form membrane in-vagination. 112 In vitro apparent affinities were however found to be quite dif-ferent from those observed in cell studies, the latter also differing for the cell lines used. 113 This is expected since the complexity of internalization in living cells cannot be described by a simple, reversible, one-to-one interaction, as is the assumption when estimating affinities in the form of KD values. One pro-cess that explains the complexity is the clustering of plasma membrane Gb3

44

receptors upon STxB binding. Clustered STxB-receptor complexes create a negative curvature, thus increasing the surface area, which facilitates binding of additional STxB pentamers. 114,115

This study evaluated the carrier capacity of STxB as a function of temper-ature, concentration and time. A rotating cell-based ligand binding assay was employed using STxB tagged with a fluorescent label (FITC) or a green fluo-rescent protein (eGFP). The data enable the formulation of a complex model for the interaction mechanism.

Results LigandTracer technology was used as RT-CBA, in order to detect intracellular events with FITC-labelled STxB and STxB-eGFP fusion protein. Daudi, Ra-mos (lymphoma) and HT-29 (colon cancer) cells were chosen as positive con-trol due to their high level of expressed Gb3, while K562 cells were employed as negative control. An initial comparison of STxB binding to Daudi and K562 confirmed that the labelled toxin binds specifically to the receptor as binding was detected to Daudi cells but not K562 cells.

Temperature effects on the binding of STxB-eGFP to its receptor was char-acterized with Daudi and HT-29 cell lines. At room temperature and 37 °C a second event in the interaction processed was observed. (Fig 14) Confocal microscopy experiments confirmed that this event was internalization (Fig 15).

Fig 14 Kinetic studies of interactions between STxB-eGFP and cells at different temperatures. RT-CBA with Daudi (A, B, C) and HT-29 cells (D, E, F) at 8 ºC (A, D), room temperature (B, E) and 37 °C (C, F). Incubation for 3 h (black) and > 3 h (grey). The dotted line represented the time point were dissociation was performed.

45

The concentration dependency of binding was also explored by incubating HT-29 cells with STxB-eGFP concentrations from 3 nM to 270 nM. When normalizing the data in respect to the signal intensity, it appeared that the curves had similar shape for the first 30 minutes. (Fig 16) At lower equilib-rium was reached within one hour, whereas at higher concentrations there was a linear increase in the signal. The latter demonstrated that higher concentra-tions of STxB promote its internalization (Fig 17). The concentration dependency was believed to potentially be a result of re-ceptor clustering (see above). For a better understanding of this process, the binding of STxB to Gb3 expressing cells was monitored after the addition of unlabeled protein and STxB labelled with a fluorescence quencher (ATTO540Q) (Fig). The STxB-quencher decreased the fluorescence signal for both STxB conjugates. The unlabeled STxB increased the signal from the STxB conjugates pre-incubated with the cells, and was consequently not a displacement as initially expected.

Fig 15. Live imaging of Daudi cells incubated with STxB-eGFP and STxB-FITC for 3h. A) Detection of deep red dye on the plasma mem-brane B) Detection of eGFP fluores-cence in the middle layer of the Z scan. C) Over-lap of the fluorescence im-ages. D) transmitted light imaging on Daudi cells. E) Detection of FITC fluores-cence in the middle layer of the Z scan. F) Overlap of the fluorescence and transmitted light imag-es.

Fig 16 Normalization of concentration series at 30 min (left). Percentage of STxB-eGFP in-ternalized per hour related with the surface bound STxB-eGFP (right)

46

Fig 17 Quenching assays on STxB-FITC (A) and STxB-eGFP (B) on Ramos cells, black curves represent STxB-ATTO540Q added after incubation with STxB constructs, grey curves represent STxB-eGFP and STxB-FITC incubations, where in the case of STxB-FITC, the same unlabeled concentration of STxB was added (A grey). (C-D) Displacement assays on Ramos (C) and HT29 cells (D). 30nM (C) and 90nM (D) of STxB-eGFP was incubated for 3 h, fol-lowing by a 3h incubation of 90nM(C) and 270nM(D) of unlabeled STxB. Dissociation was monitored afterwards.

Finally, the influence of time on the interaction between STxB and cells was considered. Experiments with Ramos cells showed that when the time for in-cubation of STxB constructs with cells was increased, the interaction became more complex and could no longer be described by a simple reversible 1:1 interaction mechanism (Fig 18).

Fig 18 Time dependency of StxB-FITC (A,C) and StxBeGFP (B,D). StxB-FITC was incubated with concentrations of 30 and 90 nM at (A) short times of 30 min and 1h or at longer times C) 3h plus 3h. StxB-eGFP was incubated with concentrations of 30 and 90 nM at (B) short times of 30 min and 1h or at longer times D) 3h plus 3h

47

Discussion In this paper, a rational strategy for characterization in cell was applied to un-derstand the complexity of the system of interest, with respect to temperature, concentration and time.

The use of RT-CBA provided detailed information on how STxB is inter-nalized into living cells under different conditions. Investigating the features of STxB conjugated to a small molecule (FITC) and fused to a protein (eGFP) gave insights into how the binding capacity of STxB can be changed by spe-cific coupling of molecules and peptides/protein to the protein.

The effect of internalization clearly discernible at temperatures close to 37 °C. An additional process that can also come into play is receptor recycling, which can change target availability. The concentration dependency of STxB interactions with cells clearly indicates that the dynamics of STxB-Gb3 bind-ing is far more complex than presented in the literature. This was also high-lighted by exploring the time dimension. When analyzing data from longer incubation times, the binding deviates from a simple reversible 1:1 mecha-nism. A possible explanation can be found in the intrinsic internalization mechanism of STxB. It has been observed that the binding of the toxin to the cellular surface forms a tubular membrane invagination. STxB cluster and tub-ular membrane therefore promotes more protein to bind by receptors recruit-ment. This is reflected in the linear increase of the signal and the high amount of STxB detected by confocal microscopy.

More complex models are needed to estimate or compensate for these ad-ditional biological processes, and conventional affinity values assuming a re-versible 1:1 binding may be misleading.

This work therefore highlights the importance of time in the understanding of complex dynamic systems, which involves both binding and internalization processes.

Paper V Bioengineering of living cells for measuring intracellular interactions of small molecules in real-time.

In cell assays are vital for establishing an understanding of target behavior in living system. A further step in this direction is to measure how molecules are exposed to their intracellular targets and to characterize their binding in a living environment, already clearing the ground for preclinical trials. In this perspective, implementing RT-CBA in the early stages of drug discovery is a strategy that might improve the success rate of a drug. This study explored the possibility of developing a time resolved intracellular assay for analysis of

48

interactions between small molecules and a target, by using an engineered target fused to a fluorescent protein.

Background Carbonic anhydrases (CA) are zinc enzymes ubiquitously expressed in pro-karyotes and eukaryotes. They are encoded by four distinct evolutionarily un-related gene families. An interesting family for drug discovery is represented by the α-CA family, found predominantly in mammals. 116 However, repre-sentatives of the α- family are also found in other organisms, including various pathogens, such as Trypanosoma cruzi. 117 The structure of tcCA is unknown, but the catalytic features appear to be very similar to those for human CAII. The parasitic tcCA has recently been validated as a target against Chagas Dis-ease, since inhibitors (e.g. sulfonamides) developed for other forms of CA also leads to a trypanocidal activity. 118,119

hCAII is one of the most studied variants of the enzyme, not only for its importance as a drug target but also as a prototypical model system for bio-physical studies and medicinal chemistry applications. 120

In this study hCAII and protein engineering was used to create a fluorescent target. By fusing it to a fluorescent protein protein (eGFP), it could be used to monitor the binding of sulfonamides/-quencher molecules internalized into living cells. An RT-CBA was employed to detect effects on fluorescence caused by the quenching of hCAII-eGFP by quencher labeled sulfonamides, clear indication of intracellular binding.

Results hCAII-eGFP constructs were developed and used to transform insect and bac-terial cells. The expression in insect cells was monitored for 70h using Ligand-Tracer to confirm the capability of the RT-CBA to detect the fluorescent pro-tein inside the cells, while confocal microscopy indicated that the engineered protein was highly expressed and localized in the cytoplasm. (Fig 19)

A B

Fig 19 Live imaging of Sf9 cells incubated with Bacmid virus for 48 h. A) Detection of deep red dye on the plasma membrane (in blue) B) Overlap of A with the detection of the eGFP fluorescence in the middle layer of the Z scan.

49

An SPR biosensor assay was developed to determine the potential effect of eGFP on the interaction between hCAII and the sulfonamides inhibitors. The kinetic of the interaction between the sulfonamide used as a reference (E67) was the same for immobilized hCAII and hCAII-eGFP, demonstrating that the fusion of eGFP did not affect active site of the enzyme. Moreover, analysis of the interaction between the sulfonamide-quencher AEBSA-Q540 and CAII resulted in a clear 1:1 binding curve. This indicated that labeling of the inhib-itor with quencher did not affect its ability to bind the target protein. (Fig 20)

Fig 20 SPR-biosensor based analysis of interactions between hCAII and hCAII-eGFP, with chosen sulfonamides. Reported KD values were calculated from fitting the sensorgrams with 1:1 binding model (in black)

Based on the SPR results, RT-CBA were performed in insect cells, when the expression level plateau was reached. AEBSA-Q540 was incubated at 25 µM, resulting in a decrease in the fluorescent signal i.e. giving an “inverted” bind-ing curve. By replacement of the media with fresh media without the inhibitor resulted in a return of the fluorescence in the system to that at the start of the experiment. (Fig 21A)

A competition experiment with the reference sulfonamide E67 was also performed. This resulted in a small change in the signal, indicating that the rapidly associating and dissociating reference sulfonamide competes for the binding to AEBSA-Q540. (Fig 21B)

50

Fig 21 A) Real-time expression of hCAII-eGFP in insect cells and binding of AEBSA-Q540. B) Normalization and inversion of the quenched signal upon incubation with AEBSA-Q540 at concentration 90uM and after addition of 80 nM E67 sulfonamide (square).

Discussion The work described is a rational strategy towards time-resolved evaluation of small molecules binding to intracellular targets, a field of great interest which is still pretty unexplored.

Engineering of proteins and small molecules tagging them with fluorescent probes, has been used widely to enable detection of interactions. However, the modification might affect the behavior of the protein/molecule and therefore their binding. Complementary in vitro experiments – e.g. using SPR biosen-sors – are important not only as controls for the binding, but also to give an indication upon the affinity and the binding mode of the molecule.

The expression of hCAII-eGFP in insect cells could be followed for more than 70h, indicating the robustness of the assay. A limit to this strategy could be the modification of the sulfonamide molecule: ATTO-Q540 is a big label that might greatly influence the internalization capacity. In contrast, a flexible linker between the target and a well-known inert tag like eGFP minimizes the possible effects upon the binding, as demonstrated with the SPR experiments.

In conclusion, the concept of time-resolved proximity-based displacement developed in this study might improve in cell exploration for lead optimiza-tion, providing more information about biological events occurring or concen-trations required for observing intracellular effects with small molecules.

51

Conclusion

Combining improved methods and rational strategies is the key for finding new drugs against challenging but promising targets, which are often ne-glected due to the high risk associated with the conventional drug develop-ment process. The overall goal of this project was therefore to develop a panel of orthogonal biophysical and biochemical assays to boost the early stage drug discovery in neglected tropical diseases.

The major findings in this thesis are as follows: tcFPPS enzyme is a suitable target for FBLD, and an extensive under-

standing of its properties and behavior in different conditions is the key for developing and validating reliable screening assays

The choice of a screening strategy and validation methods are strongly correlated to the target properties and the kind of information needed, but are also influenced by the chosen compound library intrinsic char-acteristics (size, solubility, chemical space coverage)

Implementing structured-based in vitro strategy and in cell assays al-lows a proper characterization of molecular interactions in biological processes, while looking at complex cargo systems as STxA-STxB

The introduction of the time dimension in a cell binding assay is a crucial point to truly understand the complexity of the mechanism and cargo capacity of STxB

Intracellular drug-target engagement in cell culture medium can be monitored with RT-CBA.

As part of the AEGIS consortium, this project provides new and optimized assays for studying the different targets of interest, and overall addresses ra-tional strategies for tackling challenging situations (e.g. lack of suitable refer-ence compounds, target instability, cargo mechanisms) to speed up the early stage drug discovery process.

52

Future at a glance

There are still some fundamental barriers in the development of novel treat-ments for NTDs. Conventional drug discovery approaches meet incredible challenges, because of both lack of funding, and lack of screening platforms with suitable assays and validation. This project combined new and optimized assays into a rational multidisciplinary approach, providing tools for acceler-ating the discovery process and decreasing the cost of drug design.

Implementing an in-depth target characterization to the screening of well-designed libraries is a crucial step for a better assay development, and conse-quently for a lower risk to encounter false positive hits. In this perspective, rational strategies and appropriate selection of methods are the key to succeed in finding new chemical scaffolds addressing NTDs targets, bringing that needed innovation for efficient and safe treatment of those infectious diseases.

Moreover, the combination of in vitro and in cell studies is a powerful tool for tackling complex interaction mechanisms and give a better hint about mol-ecules intracellular behavior. Since the need for a rapid discovery in NTDs requires a rapid validation of possible hits, implementing real time binding characterization in living cells at the early stage represents in this sense a small step towards the future of drug discovery. Orthogonal biophysical in vitro and in cell binding assays can actually compensate each other’s weaknesses, mak-ing the selection process more effective before the pre-clinical studies.

53

Popular summary/Sammanfattning

Covid-19 outbreak in the beginning of this year has already changed our perspective, shaping our lives for indefinite time. The drug discovery process suddenly became a reality show, while we receive constant updates on new vaccines and how well they are proceeding through the pre-clinical and clinical phase, or new information of this tiny in-sidious enemy and its targetable features for finding new drugs. Everything is happening right now, at record time.

But how long does it take to go from nothing to a treatment?

Pandemic aside, the complex process of developing a new drug from the original idea takes usually 12–15 years and cost up to $ 2.6 bil-lion. The main reason for this high cost is due to the failure rate during the clinical trials, when the drug is actually tested in humans, because either the candidates don’t work or they are not safe. The risk of such huge investment is taken into ac-count when it comes to find a new drug for a relatively common con-dition – as anticancer, antidiabetic, analgesic. And even a sudden out-break brings focus and funding for research and development, as hap-pened with Ebola in 2015, and is happening right now with Covid-

Utbrottet av covid-19 i början av det här året har redan förändrat våra perspektiv och format våra liv på obestämd tid. Läkemedels-branschens forskningsprocess har plötsligt blivit en dokusåpa, samti-digt som vi konstant tar emot upp-dateringar om nya vaccin och hur bra de utvecklas genom den för-kliniska och kliniska fasen, eller ny information om denna lilla förrä-diska fiende och dess egenskaper som vi kan rikta in oss på för att hitta nya läkemedel. Allting händer nu, på rekordtid.

Men hur lång tid tar det att gå från ingenting till en behandling?

Även utan en pandemi är proces-sen för att utveckla ett nytt läkeme-del från originalidé till färdig pro-dukt mycket komplex, det kan ta mellan 12-15 år och kosta uppemot 2,6 miljarder dollar. Den största or-saken till denna höga kostnad är graden av misslyckanden under de kliniska försöken, då läkemedlet faktiskt ska testas på människor, ef-tersom det antingen inte fungerar eller inte är säkert. Risken för en så-dan stor investering räknas in när det gäller att hitta ett nytt läkemedel för relativt vanliga tillstånd - som anticancer, antidiabetes, analgetika. Även stora utbrott av sjukdomar

54

19. However, when there is no in-centive to research treatment, dis-eases are neglected. In the Tropics there are a number of infectious dis-eases that First World countries do not often hear about, since they do not make headlines in our principle news channels. Hidden among the neglected population of the world, those diseases silently affect over a billion people worldwide, and dis-figure and kill millions each year. Caused by a variety of different mi-crobes, they go under the name of Neglected Tropical Diseases, NTDs. Compared to other diseases, there is a very small number of new therapeutics for NTDs. For exam-ple, of the 850 new products regis-tered between 2000 and 2011, only 5 were for NTDs and they were all placed as a new indication or a new composition of existing drugs; none were new medicines. Because of lack of interest and consequen-tially funding, the development of new drugs for these diseases has to go through a non-conventional way. In this sense, multi-sector en-gagement through public-private partnerships is a way that has al-ready proved its potential. This the-sis itself is only a piece in a bigger puzzle for accelerating the early stage – hence reducing the cost – of the drug discovery process in ne-glected diseases. A puzzle called AEGIS. First step of the drug dis-covery pipeline, the early stages begin with basic research, when scientists find a key target for influ-encing a biological response corre-

ger dem uppmärksamhet och finan-siering för forskning och utveckl-ing, som ebola under 2015, och som just nu med covid-19.

Men sjukdomar försummas där forskning inte uppmuntras. I Tropi-kerna finns det ett antal infektions-sjukdomar som I-länder inte ofta hör talas som eftersom de inte skapar rubriker i våra huvudsakliga nyhetskanaler. Gömda bland värl-dens försummade populationer på-verkar dessa sjukdomar i tysthet över en miljard människor i hela världen, och vanställer och dödar miljoner varje år. De orsakas av många olika mikrober och går un-der namnet NTDs (Neglected Tro-pical Diseases); Försummade tro-piska sjukdomar. Jämfört med andra sjukdomar finns det väldigt få nya läkemedel för NTDs. Till ex-empel var bara fem av 850 registre-rade nya produkter mellan 2000 och 2011 tillverkade för NTDS, och alla fem grupperades som nya be-teckningar eller nya kompositioner av existerande läkemedel, inga av dem var nya mediciner. Utveckl-ingen av nya läkemedel för dessa sjukdomar måste ske på ett icke-konventionellt sätt eftersom det saknas ett intresse och därmed också finansiering. Sektorsövergri-pande engagemang genom offent-lig-privat samverkan har redan vi-sat sig ha potential. Denna avhand-ling är endast en del av ett större pussel för att påskynda det tidiga stadiet - och följaktligen minska kostnaderna - av upptäcktsproces-sen för läkemedel för försummade sjukdomar. Detta pussel är AEGIS.

55

lated to a disease. Different strate-gies are then designed for finding molecules that “hit” the target of choice, or in other words that are able to interact with the target. The hit molecule is then modified and improved to a “lead” candidate, which is ready to be tested in ani-mal models and finally in human.

This project described different approaches and strategies devel-oped to boost the early stage drug discovery for neglected tropical diseases. Since everything start with a target, the key for develop-ing reliable processes for finding new hits is an extensive under-standing of the properties and be-havior of the target itself in differ-ent conditions. How can we mimic the physiological environment in a test tube, and what is the best con-dition to run the experiment? At this stage of the process we have to compromise a little, but also al-ready consider those parameters for a more reliable result. Using frag-ment of molecules that can be merged in a second time to form the perfect lead candidate, can be a way to speed up the search of a new drug. However, this bring to other practical consideration, such as available resources. A more “real-ity-like” approach described in this thesis is the observation of the tar-get in the cell environment. Time, in this sense, is a precious ally for understanding the complexity of in-tracellular traffic. Looking at the cell system, and even at the intra-cellular interaction between our tar-get and the new selected molecule,

Det första steget i att upptäcka nya läkemedel börjar med grundläg-gande forskning, där man hittar ett huvudmål som kan påverka ett bio-logiskt gensvar som korrelerar till en sjukdom. Sedan utformas olika strategier för att hitta molekyler som kan träffa det valda målet, eller som med andra ord kan interagera med målet. Träffmolekylen modi-fieras och förbättras sedan till en huvudkandidat som är klar att testas på djur och tillslut på människor.

Detta projekt beskriver olika till-vägagångssätt och strategier som är utvecklade för att förbättra det ti-diga stadiet av läkemedelsproces-sen för NTDs. Eftersom allting bör-jar med ett mål, är en omfattande förståelse för målets egenskaper och beteende i olika förhållanden nyckeln till att utveckla en pålitlig process. Hur kan vi imitera den psykologiska miljön i ett provrör, och vilka är de bästa förhållandena för att utföra ett experiment? I detta steg av processen måste man kom-promissa, men också överväga dessa parametrar för ett mer pålit-ligt resultat. Att sammanfoga mole-kylfragment en andra gång för att utforma den perfekta huvudkandi-daten kan vara ett sätt att påskynda sökandet efter ett nytt läkemedel. I och med detta måste man dock ta hänsyn till andra praktikaliteter, såsom tillgängliga resurser. Ett mer realistiskt tillvägagångssätt, vilket beskrivs i denna avhandling, är ob-servation av målet i en cellmiljö. På det sättet blir också tiden en värde-full allierad i förståelsen av den komplexa intracellulära trafiken.

56

might improve the success rate of the drug by making the selection process more effective before the pre-clinical studies. Since the need for a rapid discovery in NTDs re-quires a rapid validation of possible hits, implementing this studies in living cells at the early stage repre-sents in this sense a small step to-wards the future of drug discovery.

Genom att studera cellsystemet, och den intracellulära interaktionen mellan målet och den nyvalda mo-lekylen, är det möjligt att förbättra läkemedlets framgångsgrad genom att göra urvalsprocessen mer effek-tiv före de förkliniska studierna. Ef-tersom behovet av en snabb upp-täckt inom NTDs kräver ett snabbt godkännande av möjliga träffar re-presenterar genomförandet av dessa studier i ett tidigt stadium ett litet steg mot framtiden av läkeme-delsupptäckter.

57

Acknowledgements

A deep breath. Even though it is common knowledge that this would be the most read part

of the thesis, it is likewise understood that this is also the last part to be written. And I am no better, doing this the night before the submission deadline, sur-rounded by books, scribbled notes, other doctoral thesis from people I am more or less aware of, and an impressive amount of USB sticks that I have no idea where they came from. This is a moment of calm before the storm, and one can clearly see all the signs. When the time stretches, as a melted clock from a weird painting. When thoughts about past and future are mixing each other up. When the eyes threaten to close, but the mind is restless.

I left Italy four years ago, and a strange coincidence made me find a small “To do” list I wrote as a teenager, kind of prophetic in a sense: “Move to Sweden, and stay there for a while”. It did not say what I would have found here, or what to do (hiking? Not a fan), but this part of my four years in the beautiful, cold Scandinavia is now filled with great memories.

And for this, I have many people to thanks.

First of all, I thank my supervisor, Helena Danielson. It took me a while to not call you “professor Danielson”, and even more to understand that sometimes it is ok to stop talking and leave some silence filling the room. Your passion for research is more than inspiring, and it is what made me literally jumping out of the bed when I received your first e-mail after the interview in Munich. You have been by my side in the definitely not easy path that followed that freaky Monday, and it is mostly thank to you that I have learned a lot in these four years, both at professional and at personal level.

Thanks to my co-supervisor, Doreen Dobritzsch. Even if you said I should have bothered you more, you gave me great insights from a crystallographic point of you, and also in listening my teaching troubles with Biokemi I. I am positive that it really required a lot of patient, sometimes!

A very big THANK YOU! to my semi- supervisor and great colleague, Helena Nordström. Thanks for being as nerd as I am, and share knowledge, and frustration, and confusion, and WhatOnEarthIsThat in the magic kingdom of SPR. You made my experimental days definitely more fun! Thanks for be-ing always there, even in the uh-oh times. May the Force be with you.

Thanks to Gunnar Johansson, for the insightful discussions and for intro-ducing me to the bioluminescent world. Grazie mille!

58

Grazie Ylva Ivarsson, for actually being responsible of this entire PhD, which otherwise I would not have been known about.

Thanks to Micke Widersten, for lighting up my thoughts and interest in the basic of Biochemistry. And for not laughing in my face, when in my first months I assured you I would have completed Biochemistry III during my first year. And well…

Thanks to The Danielson’s, always in motion and everyday more. Thanks to Ted, for sharing the ups and downs of the PhD life, but anyway BENE. Thanks to Mostafa, for being an incredible office mate, who is never surprised by my sudden speeches out of the blue. Thank you, Gun, for your tips, your passion, and for sharing my frustration with tcCA – I am not done with that, yet. Thanks to Khyati, who always has a good word and a minute to chat. Thank you, Sandra, for your efficient way of taking care of every little prob-lem. Thanks Emil, for the past and the present talks. Thanks to the mystic half-Danielson/half-Dobritzsch Daniela, for the small chats, I wish you all the best with your PhD!

Thanks to Gigi, because you are a Danielson’s ghost, but definitely alive. The lab would be a way more boring if it wasn’t for its inhabitants. Espe-

cially the weekend ones. A huge THANK YOU to Caroline and Johanna, the early birds. Thank you both for reminding me that sometimes it is ok to open up, a little… too much. Thank you, Caroline, for breakfast with gossips, for the private talks, and for the infinite borrowings. Thank you, Johanna, for the tea-times with all what it means, for trying not to plan too much just because of my Italian nature, and not judging my zombie face when I had to come too early-for-me for teaching. You both are precious, and the best lab buddies – even though I cannot follow your time-schedule.

Thanks Susanne, for the MMORPG experience, and for the great baking skills that I definitely took advantage of. Thanks Ali, a constant presence in my extra-hour work. Seriously, sometimes it seems you live there!

Gina, thanks for being that kind of neighbor you can go have a tea with. I wish you all the best with your PhD! Thank you, Lianne, for the waterpolo experience and all the rest.

Mille grazie a Leandro, il mio ex-compagno di scrivania che mi manca tanto. Thank you, for being always cheerful, and sharing funny jokes.

Thanks to the computational guys, who migrated one floor up. Thank you, Erik, for the lunch stories; to Maxim, I am sorry I betrayed the idea of getting a dog, and got two cats instead! Grazie Emiliano perché mi hai permesso di romperti un po’ le scatole e, niente, alle volte c’è solo bisogno di ciacolare. And to Joana and Malin, who are cited a little bit below as well!

A huge THANKS to my ex-ex-office mate Dirk, who helped me in many (im)possible ways. Thank you, Claire, for being always a presence in my life, even though you are so damn far.

Thanks Anzhelika (or better Prof. Vorobyeva), Piero and Claudia, for all the fun.

59

Thanks to my student for the great job and for keeping up with my perks as a supervisor. Thank you, Leo, Tom and Charlotte. And thanks to Lilla, es-pecially for “the drops”. I wish you all a great future.

This project would be only an idea, if it wasn’t for all the people who par-ticipated in it. So I would like to thanks all my collaborators, in good times and in bad.

Thanks to BioThema’s in Handen, for accepting me as one of them during one, very cold, December. Thanks to Arne, Leo, Demet and Marie (former colleague).

Thanks to the Klebe’s group in Marburg for hosting me in the best way possible, and introducing me to the Marburg experience. Thanks to prof. Klebe, Steffen, Stefan, Stefania and Maria Giovanna.

Thanks to the AEGIS people for being a crazy company of my semesters. You made everything better. Grazie Laurita, Atilio, Patrick e Roberto. Grazie Francesca, perché sia Uppsala che Marburg sono state un’avventura! And a big thanks to Eva for all the help!

Grazie, Vale, perché dal primo momento in cui ci siamo incontrate nella hall dell’albergo, mi sei sempre stata vicina (seppure lontana). Un abbraccio enorme che spero di darti presto, perché manchi un sacco. #GENio

Grazie ai Modellisti Anonimi, perché anche se siamo ognuno in uno Stato diverso, siamo sempre i Modellisti Anonimi. Grazie, Alberto e Alessandro!

A special thanks to Joana and Fabian, to Malin and Francesco because we definitely have to “Get together” again, when this freaking pandemic calms down.

Grazie agli Italiani di Uppsala, perché mi avete fatta sentire a casa, e deci-samente meno sola. Grazie ad Alberto, Armin, Marco Cap, Marco Rep (forse più italiano di me), Fabio, Claudio, and a big thanks to Hedda! Un ringrazia-mento speciale a Olivia, per i cataloghi che sto ancora guardando (anche se mannaggia-il-covid se ne riparla per il 2021), e per aver tradotto il mio “inte-ressantissimo” Popular Summary.

Il più grande grazie va alla mia sorellina, Marta. La parte più difficile di questi quattro anni è stata non poterti vedere tutti i giorni! E un grosso grazie va anche al mio fratello acquisito, Aure, perché semplicemente ti voglio bene. Grazie ai miei genitori, che mi hanno supportata in ogni mia scelta. So che non è stato facile vedermi andare via, e soprattutto sapermi qui, lontana in quest’ultimo periodo. Non vedo l’ora di potervi riabbracciare tutti.

E João... per essere una che parla sempre tantissimo, non so esattamente cosa dire. Volevo scrivere in Portoghese, ma ormai è notte fonda, e faccio addirittura fatica a scrivere in Italiano. La cosa incredibile è che tu sei ancora qui, accanto a me. Ci sei sempre stato. Grazie perché hai saputo darmi forza, quando ne avevo più bisogno. Perché mi hai supportata e sopportata in ogni modo, ed è grazie a te che sono riuscita a superare tutto – anche il brutto – di questi quattro anni. Amo-te, noivo.

60

Finally, thanks to you, who are reading my words. Hope there are not too many mistakes, I have never written a book before – I know I will have one week to check everything, so I am quite optimistic about it.

And if there is a take-home message to my story, well:

It might be hard, but nobody’s alone.

61

References

1. Flamel, N., Orandus, E. & Westcott, W. W. Nicholas Flammel: His exposition of the hieroglifical figures which he caused to be painted upon an arch in St. Innocents church yard in Paris : concerning both the theory and practice of the philosophers stone. (1890).

2. Dong, X., Milholland, B. & Vijg, J. Evidence for a limit to human lifespan. Nature 538, 257–259 (2016).

3. Rankin, S. A., Bradley, R. L., Miller, G. & Mildenhall, K. B. A 100-Year Review: A century of dairy processing advancements—Pasteurization, cleaning and sanitation, and sanitary equipment design. J. Dairy Sci. 100, 9903–9915 (2017).

4. Smelt, J. P. P. M. & Brul, S. Thermal Inactivation of Microorganisms. Crit. Rev. Food Sci. Nutr. 54, 1371–1385 (2014).

5. Scott, E., Bloomfield, S. F. & Barlow, C. G. Evaluation of disinfectants in the domestic environment under ‘in use’ conditions. J. Hyg. (Lond). 92, 193–203 (1984).

6. Berche, P. Louis Pasteur, from crystals of life to vaccination. Clin. Microbiol. Infect. 18, 1–6 (2012).

7. Durand, G. A., Raoult, D. & Dubourg, G. Antibiotic discovery: history, methods and perspectives. Int. J. Antimicrob. Agents 53, 371–382 (2019).

8. Gaudillière, J. P. How pharmaceuticals became patentable: The production and appropriation of drugs in the twentieth century. Hist. Technol. 24, 99–106 (2008).

9. Moors, E. H. M., Cohen, A. F. & Schellekens, H. Towards a sustainable system of drug development. Drug Discov. Today 19, 1711–1720 (2014).

10. Keating, C. Ken Warren and the Rockefeller Foundation’s Great Neglected Diseases Network, 1978–1988: The Transformation of Tropical and Global Medicine. Mol. Med. 20, S24–S30 (2014).

11. R. Choffnes, E. & Relman, D. A. The Causes and Impacts of Neglected Tropical and Zoonotic Diseases. The National Academies Press (National Academies Press, 2011). doi:10.17226/13087

12. Johnston, K. L., Ford, L. & Taylor, M. J. Overcoming the challenges of drug discovery for neglected tropical diseases: The A·WOL experience. J. Biomol. Screen. 19, 335–343 (2014).

13. Policy Cures Research. 2019 G-Finder neglected disease report: Neglected Disease Research and Development: Uneven Progress. (2020).

14. Weng, H. B., Chen, H. X. & Wang, M. W. Innovation in neglected tropical disease drug discovery and development. Infect. Dis. Poverty 7, 1–9 (2018).

15. Smith, J. & Taylor, E. M. What Is Next for NTDs in the Era of the Sustainable Development Goals? PLoS Negl. Trop. Dis. 10, e0004719 (2016).

16. Kennedy, P. G. E. & Rodgers, J. Clinical and neuropathogenetic aspects of human African trypanosomiasis. Front. Immunol. 10, 1–11 (2019).

62

17. Echeverria, L. E. & Morillo, C. A. American Trypanosomiasis (Chagas Disease). Infect. Dis. Clin. North Am. 33, 119–134 (2019).

18. Barrett, M. P. et al. The trypanosomiases. Lancet 362, 1469–1480 (2003). 19. Savioli, D. D. and L. Accelerating Work to Overcome the Global Impact of

Neglected Tropical Diseases: A Roadmap for Implementation. World Heal. Organ. 1–42 (2012). doi:WHO/HTM/NTD/2012.1

20. World Health Organization. Generic framework for control, elimination and eradication of neglected tropical diseases. (2016).

21. Pollastri, M. P. Fexinidazole: A New Drug for African Sleeping Sickness on the Horizon. Trends Parasitol. 34, 178–179 (2018).

22. Aksoy, S., Buscher, P., Lehane, M., Solano, P. & Van Den Abbeele, J. Human African trypanosomiasis control: Achievements and challenges. PLoS Negl. Trop. Dis. 11, 1–6 (2017).

23. Fairlamb, A. H. & Horn, D. Melarsoprol Resistance in African Trypanosomiasis. Trends Parasitol. 34, 481–492 (2018).

24. Kennedy, P. G. E. Update on human African trypanosomiasis (sleeping sickness). J. Neurol. 266, 2334–2337 (2019).

25. Jenkins, C. A guide to neglected tropical diseases prioritised by the World Health Organisation. Financial Times (2017).

26. World Health Organization. WHO methods and data sources for country‐level causes of death 2000‐2012. Global Health Estimates Technical Paper WHO/HIS/HSI/GHE/2014.7 (2018). doi:ISBN 9789241564854

27. Bump, J. B., Reich, M. R. & Johnson, A. M. Diarrhoeal diseases and the global health agenda: Measuring and changing priority. Health Policy Plan. 28, 799–808 (2013).

28. Kotloff, K. L. The Burden and Etiology of Diarrheal Illness in Developing Countries. Pediatr. Clin. North Am. 64, 799–814 (2017).

29. D. Legros & D. Legros. Guidelines for the control of shigellosis, including epidemics due to Shigella dysenteriae type 1. World Heal. Organ. 1–70 (2005).

30. Ashkenazi, S. Shigella infections in children: New insights. Semin. Pediatr. Infect. Dis. 15, 246–252 (2004).

31. Schnupf, P. & Sansonetti, P. J. Shigella Pathogenesis: New Insights through Advanced Methodologies. Microbiol. Spectr. 7, 1–24 (2019).

32. Lampel, K. A., Formal†, S. B. & Maurelli, A. T. A Brief History of Shigella. EcoSal Plus 8, (2018).

33. Puzari, M., Sharma, M. & Chetia, P. Emergence of antibiotic resistant Shigella species: A matter of concern. J. Infect. Public Health 11, 451–454 (2018).

34. Strovel, J. et al. Early Drug Discovery and Development Guidelines: For Academic Researchers, Collaborators, and Start-up Companies. Assay Guid. Man. 1–35 (2004).

35. Opassi, G., Gesù, A. & Massarotti, A. The hitchhiker’s guide to the chemical-biological galaxy. Drug Discov. Today 23, 565–574 (2018).

36. MacArron, R. et al. Impact of high-throughput screening in biomedical research. Nat. Rev. Drug Discov. 10, 188–195 (2011).

37. Wildey, M. J., Haunso, A., Tudor, M., Webb, M. & Connick, J. H. High-Throughput Screening. in Annual Reports in Medicinal Chemistry 50, 149–195 (Elsevier Inc., 2017).

38. Fox, S. et al. High-throughput screening: Update on practices and success. J. Biomol. Screen. 11, 864–869 (2006).

39. Chessari, G. & Woodhead, A. J. From fragment to clinical candidate—a historical perspective. Drug Discov. Today 14, 668–675 (2009).

63

40. Hubbard, R. E. & Murray, J. B. Experiences in fragment-based lead discovery. in Methods in Enzymology 493, 509–531 (Elsevier Inc., 2011).

41. Erlanson, D. A., Fesik, S. W., Hubbard, R. E., Jahnke, W. & Jhoti, H. Twenty years on: The impact of fragments on drug discovery. Nat. Rev. Drug Discov. 15, 605–619 (2016).

42. Shi, Y. & von Itzstein, M. How size matters: Diversity for fragment library design. Molecules 24, 1–10 (2019).

43. Jencks, W. P. On the attribution and additivity of binding energies. Proc. Natl. Acad. Sci. U. S. A. 78, 4046–4050 (1981).

44. Bottegoni, G., Favia, A. D., Recanatini, M. & Cavalli, A. The role of fragment-based and computational methods in polypharmacology. Drug Discov. Today 17, 23–34 (2012).

45. Baker, M. Fragment-based lead discovery grows up. Nat. Rev. Drug Discov. 12, 5–7 (2013).

46. Singh, M., Tam, B. & Akabayov, B. NMR-fragment based virtual screening: A brief overview. Molecules 23, (2018).

47. Mackey, T. K. & Liang, B. A. Threats from emerging and re-emerging neglected tropical diseases (NTDs). Infect. Ecol. Epidemiol. 2, 18667 (2012).

48. Balakrishnan, V. S. Ending neglected tropical diseases. Lancet. Infect. Dis. 17, 584–585 (2017).

49. Kefalidou, A. A., Nations, U. & Affairs, S. Brief for GSDR – 2016 Update Sustainable drug development for Neglected Tropical Diseases. 1–6 (2016).

50. DNDi. Making Medical History To Meet the Needs of Neglected Patients Drugs for Neglected Diseases Initiative 2018 Annual Report. 1–54 (2019). doi:https://www.dndi.org/annualreport2018/

51. Pierce, R. J., MacDougall, J., Leurs, R. & Costi, M. P. The Future of Drug Development for Neglected Tropical Diseases: How the European Commission Can Continue to Make a Difference. Trends Parasitol. 33, 581–583 (2017).

52. Coaker, H. Neglected tropical disease research: Rethinking the drug discovery model. Future Med. Chem. 7, 693–700 (2015).

53. Laursen, L. A neglected cause. Nature 533, S68–S69 (2016). 54. Mandal, S., Moudgil, M. & Mandal, S. K. Rational drug design. Eur. J.

Pharmacol. 625, 90–100 (2009). 55. Dawidowski, M. et al. Inhibitors of PEX14 disrupt protein import into

glycosomes and kill Trypanosoma parasites. Science (80-. ). 355, 1416–1420 (2017).

56. Joseph-McCarthy, D., Campbell, A. J., Kern, G. & Moustakas, D. Fragment-Based Lead Discovery and Design. J. Chem. Inf. Model. 54, 693–704 (2014).

57. Renaud, J. P. et al. Biophysics in drug discovery: Impact, challenges and opportunities. Nat. Rev. Drug Discov. 15, 679–698 (2016).

58. Genick, C. C. & Wright, S. K. Biophysics: for HTS hit validation, chemical lead optimization, and beyond. Expert Opin. Drug Discov. 12, 897–907 (2017).

59. Hennig, M., Ruf, A. & Huber, W. Combining Biophysical Screening and X-Ray Crystallography for Fragment-Based Drug Discovery. in Peptide-Based Materials 310, 115–143 (2011).

60. Copeland, R. A. Enzymes. A Practical Introduction to Structure, Mechanism, and Data Analysis. (Wiley-VCH, 2000).

61. Copeland, R. A., Pompliano, D. L. & Meek, T. D. Drug–target residence time and its implications for lead optimization. Nat. Rev. Drug Discov. 5, 730–739 (2006).

64

62. Folmer, R. H. A. Drug target residence time: a misleading concept. Drug Discov. Today 23, 12–16 (2018).

63. Davis, B. J. & Erlanson, D. A. Learning from our mistakes: The ‘unknown knowns’ in fragment screening. Bioorg. Med. Chem. Lett. 23, 2844–2852 (2013).

64. Aldrich, C. et al. The Ecstasy and Agony of Assay Interference Compounds. ACS Chem. Biol. 12, 575–578 (2017).

65. Björkelund, H., Gedda, L. & Andersson, K. Comparing the epidermal growth factor interaction with four different cell lines: Intriguing effects imply strong dependency of cellular context. PLoS One 6, 1–7 (2011).

66. Chavanieu, A. & Pugnière, M. Developments in SPR Fragment Screening. Expert Opin. Drug Discov. 11, 489–499 (2016).

67. Shepherd, C. A., Hopkins, A. L. & Navratilova, I. Fragment screening by SPR and advanced application to GPCRs. Prog. Biophys. Mol. Biol. 116, 113–123 (2014).

68. Nitsche, C. & Otting, G. NMR studies of ligand binding. Curr. Opin. Struct. Biol. 48, 16–22 (2018).

69. Hevener, K. E. et al. Hit-to-Lead: Hit Validation and Assessment. in Methods in Enzymology 610, 265–309 (Elsevier Inc., 2018).

70. Harner, M. J., Frank, A. O. & Fesik, S. W. Fragment-based drug discovery using NMR spectroscopy. J Biomol NMR 56, 65–75 (2013).

71. Moschen, T. et al. Measurement of Ligand-Target Residence Times by 1H Relaxation Dispersion NMR Spectroscopy. J. Med. Chem. 59, 10788–10793 (2016).

72. Genick, C. C. et al. Applications of biophysics in high- Throughput screening hit validation. J. Biomol. Screen. 19, 707–714 (2014).

73. Niesen, F. H., Berglund, H. & Vedadi, M. The use of differential scanning fluorimetry to detect ligand interactions that promote protein stability. Nat. Protoc. 2, 2212–2221 (2007).

74. Bruylants, G., Wouters, J. & Michaux, C. Differential Scanning Calorimetry in Life Science: Thermodynamics, Stability, Molecular Recognition and Application in Drug Design. Curr. Med. Chem. 12, 2011–2020 (2005).

75. Davis, A. M., Teague, S. J. & Kleywegt, G. J. Application and limitations of x-ray crystallographic data in structure-based ligand and drug design. Angew. Chemie - Int. Ed. 42, 2718–2736 (2003).

76. Russo Krauss, I., Merlino, A., Vergara, A. & Sica, F. An overview of biological macromolecule crystallization. Int. J. Mol. Sci. 14, 11643–11691 (2013).

77. Chegkazi, M. S., Mamais, M., Sotiropoulou, A. I. & Chrysina, E. D. Rational drug design using integrative structural biology. Methods Mol. Biol. 1824, 89–112 (2018).

78. Burke, T., Loniello, K., Beebe, J. & Ervin, K. Development and Application of Fluorescence Polarization Assays in Drug Discovery. Comb. Chem. High Throughput Screen. 6, 183–194 (2012).

79. Lea, W. A. & Simeonov, A. Fluorescence polarization assays in small molecule screening. Expert Opin. Drug Discov. 6, 17–32 (2011).

80. Teilum, K., Kunze, M. B. A., Erlendsson, S. & Kragelund, B. B. (S)Pinning down protein interactions by NMR. Protein Sci. 26, 436–451 (2017).

81. Fesik, S. W. NMR structure-based drug design. J. Biomol. NMR 3, 261–269 (1993).

82. Sierecki, E. et al. A cell-free approach to accelerate the study of protein-protein interactions in vitro. Interface Focus 3, (2013).

65

83. Veloria, J. R., Devkota, A. K., Cho, E. J. & Dalby, K. N. Development of a cost effective and robust AlphaScreen® platform for application. Biotechniques 64, 181–183 (2018).

84. Eglen, R. M. et al. The Use of AlphaScreen Technology in HTS: Current Status. Curr. Chem. Genomics 1, 2–10 (2008).

85. Kabiri, M. & Unsworth, L. D. Application of isothermal titration calorimetry for characterizing thermodynamic parameters of biomolecular interactions: Peptide self-assembly and protein adsorption case studies. Biomacromolecules 15, 3463–3473 (2014).

86. Liang, Y. Applications of isothermal titration calorimetry in protein science. Acta Biochim. Biophys. Sin. (Shanghai). 40, 565–576 (2008).

87. Ghai, R., Falconer, R. J. & Collins, B. M. Applications of isothermal titration calorimetry in pure and applied research-survey of the literature from 2010. J. Mol. Recognit. 25, 32–52 (2012).

88. Asmari, M., Ratih, R., Alhazmi, H. A. & El Deeb, S. Thermophoresis for characterizing biomolecular interaction. Methods 146, (2018).

89. Dijkman, P. M., Lea, W. A., Gmbh, N. T., Lazic, A. & Gmbh, N. T. Microscale Thermophoresis Quantifies Biomolecular Interactions under Previously Challenging Conditions. Methods 59, 301–315 (2014).

90. Jerabek-Willemsen, M., Wienken, C. J., Braun, D., Baaske, P. & Duhr, S. Molecular interaction studies using microscale thermophoresis. Assay Drug Dev. Technol. 9, 342–353 (2011).

91. Seidel, S. A. I. et al. Label-Free Microscale Thermophoresis Discriminates Sites and Affinity of Protein–Ligand Binding. Angew. Commun. 51, 10656–10659 (2012).

92. Spiegelberg, D., Stenberg, J., Haylock, A. K. & Nestor, M. A real-time in vitro assay as a potential predictor of in vivo tumor imaging properties. Nucl. Med. Biol. 43, 12–18 (2016).

93. Diéguez, L. et al. Effect of the Refractive Index of Buffer Solutions in Evanescent Optical Biosensors. Sens. Lett. 7, 851–855 (2009).

94. Kozma, P., Hámori, A., Kurunczi, S., Cottier, K. & Horvath, R. Grating coupled optical waveguide interferometer for label-free biosensing. Sensors Actuators, B Chem. 155, 446–450 (2011).

95. Patko, D., Cottier, K., Hamori, A. & Horvath, R. Single beam grating coupled interferometry: high resolution miniaturized label-free sensor for plate based parallel screening. Opt. Express 20, 23162 (2012).

96. Kozma, P., Kehl, F., Ehrentreich-Förster, E., Stamm, C. & Bier, F. F. Integrated planar optical waveguide interferometer biosensors: A comparative review. Biosens. Bioelectron. 58, 287–307 (2014).

97. de Souza, W. & Rodrigues, J. C. F. Sterol Biosynthesis Pathway as Target for Anti-trypanosomatid Drugs. Interdiscip. Perspect. Infect. Dis. 2009, 1–19 (2009).

98. Montalvetti, A. et al. Farnesyl pyrophosphate synthase is an essential enzyme in Trypanosoma brucei: In vitro RNA interference and in vivo inhibition studies. J. Biol. Chem. 278, 17075–17083 (2003).

99. Demoro, B. et al. Ibandronate metal complexes: solution behavior and antiparasitic activity. J. Biol. Inorg. Chem. 23, 303–312 (2018).

100. Recher, M. et al. Design, synthesis and biological evaluation of sulfur-containing 1,1-bisphosphonic acids as antiparasitic agents. Eur. J. Med. Chem. 60, 431–440 (2013).

66

101. Montalvetti, A. et al. Bisphosphonates Are Potent Inhibitors of Trypanosoma cruzi Farnesyl Pyrophosphate Synthase. J. Biol. Chem. 276, 33930–33937 (2001).

102. Price, A. J., Howard, S. & Cons, B. D. Fragment-based drug discovery and its application to challenging drug targets. Essays Biochem. 61, 475–484 (2017).

103. Hartshorn, M. J. et al. Fragment-Based Lead Discovery Using X-ray Crystallography. J. Med. Chem. 48, 403–413 (2005).

104. Danielson, U. Fragment Library Screening and Lead Characterization Using SPR Biosensors. Curr. Top. Med. Chem. 9, 1725–1735 (2009).

105. Fraser, M. E. et al. Structure of Shiga Toxin Type 2 (Stx2) from Escherichia coli O157:H7. J. Biol. Chem. 279, 27511–27517 (2004).

106. Kotloff, K. L., Riddle, M. S., Platts-Mills, J. A., Pavlinac, P. & Zaidi, A. K. M. Shigellosis. Lancet 391, 801–812 (2018).

107. Janssen, K. P. et al. In vivo tumor targeting using a novel intestinal pathogen-based delivery approach. Cancer Res. 66, 7230–7236 (2006).

108. Luginbuehl, V., Meier, N., Kovar, K. & Rohrer, J. Intracellular drug delivery: Potential usefulness of engineered Shiga toxin subunit B for targeted cancer therapy. Biotechnol. Adv. 36, 613–623 (2018).

109. Fraser, M. E., Chernaia, M. M., Kozlov, Y. V. & James, M. N. G. Crystal structure of the holotoxino from Shigella dysenteriae at 2.5 Å resolution. Nat. Struct. Mol. Biol. 1, 59–64 (1994).

110. Müller, S. K. et al. Gb3-binding lectins as potential carriers for transcellular drug delivery. Expert Opin. Drug Deliv. 14, 141–153 (2017).

111. Nakajima, H. et al. Kinetic Analysis of Binding between Shiga Toxin and Receptor Glycolipid Gb3Cer by Surface Plasmon Resonance. J. Biol. Chem. 276, 42915–42922 (2001).

112. Watkins, E. B. et al. Shiga toxin induces lipid compression : a mechanism for generating membrane curvature Shiga toxin induces lipid compression : a mechanism for generating. (2019). doi:10.1021/acs.nanolett.9b03001

113. Gallegos, K. M. et al. Shiga toxin binding to glycolipids and glycans. PLoS One 7, (2012).

114. Rydell, G. E. et al. Rab12 Localizes to Shiga Toxin-Induced Plasma Membrane Invaginations and Controls Toxin Transport. Traffic 15, 772–787 (2014).

115. Pina, D. G., Johannes, L. & Castanho, M. A. R. B. Shiga toxin B-subunit sequential binding to its natural receptor in lipid membranes. Biochim. Biophys. Acta - Biomembr. 1768, 628–636 (2007).

116. Supuran, C. Carbonic Anhydrases An Overview. Curr. Pharm. Des. 14, 603–614 (2008).

117. Vermelho, A. B. et al. Carbonic anhydrases from Trypanosoma and Leishmania as anti-protozoan drug targets. Bioorganic Med. Chem. 25, 1543–1555 (2017).

118. Supuran, C. T. Inhibition of carbonic anhydrase from Trypanosoma cruzi for the management of Chagas disease: An underexplored therapeutic opportunity. Future Med. Chem. 8, 311–324 (2016).

119. Pan, P. et al. Cloning, characterization, and sulfonamide and thiol inhibition studies of an α-carbonic anhydrase from Trypanosoma cruzi, the causative agent of Chagas disease. J. Med. Chem. 56, 1761–1771 (2013).

120. Pastorekova, S., Parkkila, S., Pastorek, J. & Supuran, C. T. Carbonic anhydrases: Current state of the art, therapeutic applications and future prospects. J. Enzyme Inhib. Med. Chem. 19, 199–229 (2004).

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1932

Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science andTechnology, Uppsala University, is usually a summary of anumber of papers. A few copies of the complete dissertationare kept at major Swedish research libraries, while thesummary alone is distributed internationally throughthe series Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Science and Technology.(Prior to January, 2005, the series was published under thetitle “Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-408721

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2020

accelerating the discovery of drugs for neglected...

Documents