i 
 

Synthesis and evaluation of flexible 

pyrimethamine analogues as antifolates 

against drug-resistant malaria 

 

 

 

Thesis by Matthew Maree 

 

Submitted to the Faculty of Science, University of the Witwatersrand, 

Johannesburg, in fulfilment of the requirements for the degree of Doctor of 

Philosophy 

Supervisor: Professor Amanda Rousseau 

Co-Supervisor: Dr Kennedy Ngwira  

2025 



ii 
 

Declaration  

 

I declare that the work presented in this thesis was carried out exclusively by myself under the 

supervision of Professor Amanda Rousseau and Dr Kennedy Ngwira. Biological assessments 

were conducted by the BIOTEC research group in Thailand, the results of which are included 

in this thesis. This thesis is submitted for the Doctor of Philosophy Degree in Science at the 

University of the Witwatersrand, Johannesburg and has not been submitted before for any 

other degree or qualification.  

 

 

 

 

 

 

 

 

 

 

 

Matthew Maree 

March 2025 



iii 
 

Abstract  

In many parts of the world today, malaria still represents a major health crisis, with millions 

of cases and hundreds of thousands of deaths reported each year. The major malaria causing 

parasite is Plasmodium falciparum, which accounts for the majority of cases and deaths 

worldwide. This parasite has shown a remarkable ability to rapidly mutate and develop 

resistance against initially effective drug molecules. The continued threat of malaria today 

motivates the development of new molecules which can remain active despite point 

mutations that arise in the active site of the targeted enzymes.  

Previous work done by A. Rousseau and colleagues showed that modifying existing class II 

antifolates, which target the bifunctional parasite enzyme dihydrofolate reductase – 

thymidylate synthase (DHFR-TS), by adding a flexible four-atom linker allowed for high levels 

of antiplasmodial activity to be maintained against drug-resistant forms of the parasite.  

In the present work, molecular modelling techniques were used to optimize the structural 

elements of flexible pyrimethamine analogues for binding ability against mutant enzymes. The 

effects of various functional groups around a conserved structure were evaluated, and a 

library of compounds was selected for synthesis. Numerous flexible pyrimethamine analogues 

were successfully synthesized making use of partially established experimental procedures, 

and were submitted for various antiplasmodial assessments. The synthesized compounds 

showed excellent activity in single enzyme assays, inhibiting drug-resistant PfDHFR enzymes 

at low nanomolar concentrations. Whole cell assays were also conducted, where a significant 

decrease in activity was observed, with the most active compounds inhibiting the parasite 

cells at low micromolar concentrations. These results suggested that while the compounds 

were effective binders of the target enzyme, they had some pharmacokinetic limitations 

which prevented them from effectively exhibiting their mode of action inside the cell.  

A second-generation of analogues was then envisaged, taking inspiration from existing 

antifolate compounds which are known to have favourable pharmacokinetic properties. 

Methods for the synthesis of the devised second-generation analogues were developed as 

part of this work, and the binding ability of the compounds was validated with further 

molecular modelling studies. We are currently awaiting the results for the biological 

assessments of the second-generation analogues.  



iv 
 

Acknowledgments  

 

First and foremost, my deepest thanks go to my devoted supervisor, Professor Amanda 

Rousseau. Your diligent feedback and tireless support made the completion of this project 

possible. I feel incredibly fortunate to have worked with you over the past four years. 

I am also deeply grateful to Dr Kennedy Ngwira for providing additional support and invaluable 

feedback during the writing process. Special thanks also to Dr Kamogelo Butsi, whose 

expertise was crucial in guiding the early stages of this project and my research. 

Many aspects of this work would not have been possible without the biological assessments 

provided by the BIOTEC research group in Thailand. My heartfelt thanks to all its members for 

their hard work in delivering the biological results. 

I also acknowledge the CSIR for hosting the CHPC, where all the molecular dynamics 

simulations for this project were conducted. 

To the members of the organic lab at Wits, especially my lab mates in 318 — Bianca, Aneesa, 

and Khanya—thank you for your camaraderie and for keeping me sane during the trials and 

tribulations of organic synthesis. 

To my close friends, Matthew Bracken and Alex Plakas, whom I met during honours in 2020 

and continued this journey with, thank you for the countless lunches, coffee breaks, and 

opportunities to vent about failed reactions and endless columns. 

I owe an immeasurable debt to my parents, who supported me throughout my academic 

journey and provided an environment that allowed me to focus on my work. 

Finally, my deepest gratitude to Agi for your unwavering love and support during the most 

challenging times. 



v 
 

Table of contents  

Contents 
Declaration .............................................................................................................................................. ii 

Abstract .................................................................................................................................................. iii 

Acknowledgments .................................................................................................................................. iv 

Table of contents .................................................................................................................................... v 

Chapter 1 – Introduction ......................................................................................................................... 1 

1.1 Malaria .......................................................................................................................................... 1 

1.1.1 Brief history and introduction ................................................................................................ 1 

1.1.2 Summary of Plasmodium falciparum life cycle ...................................................................... 2 

1.1.3 History of malaria prevention and treatment........................................................................ 4 

1.1.4 Social and economic burden of malaria ............................................................................... 14 

1.2 Details of folate metabolism and review of antifolate drugs ..................................................... 17 

1.2.1 Overview of folate metabolism ........................................................................................... 17 

1.2.2 Antifolates in the treatment of various diseases ................................................................. 21 

1.2.3 Resistance to antifolates in the treatment of malaria. ........................................................ 28 

1.2.4 Overcoming resistance to class II antifolates in the treatment of malaria .......................... 31 

1.3 Molecular modelling in drug discovery ....................................................................................... 33 

1.4 Project introduction and aims .................................................................................................... 34 

1.4.1 Background of the current project ...................................................................................... 34 

1.4.2 Aims...................................................................................................................................... 37 

1.4.3 Proposed methodology ........................................................................................................ 38 

Chapter 2 – Molecular Modelling ......................................................................................................... 42 

2.1 Introduction ................................................................................................................................ 42 

2.2 Aims and objectives .................................................................................................................... 45 

2.3 Validation of PfDHFR-TS crystal structures and ligand binding mode ........................................ 47 

2.3.1 Assessment of protein crystal structures............................................................................. 47 

2.3.2 Identification of ligand binding mode and self-docking experiments ................................. 51 

2.3.3 Enrichment docking ............................................................................................................. 57 

2.4 Optimizing interactions in the hydrophobic pocket ................................................................... 59 

2.5 Assessment of X-substituents around the terminal ring ............................................................ 69 

2.6 Molecular dynamics simulations ................................................................................................ 80 

2.7 In silico estimation of physiochemical parameters ..................................................................... 85 

2.8 Concluding remarks and recommendations for synthesis ......................................................... 86 



vi 
 

2.9 Additional molecular modelling experiments ............................................................................. 88 

Chapter 3 – Results and discussion part A: Synthesis and biological evaluations of first-generation 

analogues .............................................................................................................................................. 90 

3.1 Synthesis of flexible pyrimethamine analogues ......................................................................... 90 

3.1.1 Alkylation of phenols with propargyl bromide .................................................................... 92 

3.1.2 Esterification of aliphatic carboxylic acids ........................................................................... 95 

3.1.3 Formation of beta-ketonitriles ............................................................................................. 97 

3.1.4 Microwave assisted formation of 2,4-diaminopyrimidines ............................................... 100 

3.1.5 Iodination of 2,4-diaminopyrimidines ............................................................................... 105 

3.1.6 Sonogashira coupling reaction ........................................................................................... 108 

3.1.7 Hydrogenation of Sonogashira reaction products ............................................................. 115 

3.2 Biological assessments of synthesized flexible pyrimethamine analogues .............................. 119 

3.2.1 Single enzyme inhibition assays ......................................................................................... 123 

3.2.2 Whole P. falciparum cell assays ......................................................................................... 133 

3.2.3 Conclusions and path forwards.......................................................................................... 139 

Chapter 4 – Results and discussion part B: Selection, synthesis, and modelling studies of second-

generation analogues ......................................................................................................................... 142 

4.1: Design of second-generation analogues .................................................................................. 142 

4.2: Proposed synthesis of second-generation analogues ............................................................. 146 

4.2.1: Formation of propargyl substituted phenols and anilines ................................................ 147 

4.2.2 Sonogashira coupling reaction ........................................................................................... 162 

4.2.3 Hydrogenation of Sonogashira reaction products ............................................................. 171 

4.2.4 Removal of Boc protecting group ...................................................................................... 175 

4.2.5 Hydrolysis of esters after Sonogashira and hydrogenation ............................................... 176 

4.2.6 Addition of glutamic acid ................................................................................................... 178 

4.2.7 Deprotection of glutamic acid ........................................................................................... 196 

4.3 Molecular modelling studies of second-generation analogues ................................................ 198 

4.3.1 Introduction ....................................................................................................................... 198 

4.3.2 Induced fit docking ............................................................................................................. 200 

4.3.3 Molecular dynamics simulations ....................................................................................... 208 

4.3.4 In silico estimation of physiochemical parameters ............................................................ 211 

Chapter 5 – Conclusions and future work .......................................................................................... 213 

Chapter 6 – Experimental methods and procedures .......................................................................... 217 

6.1 Synthesis of first-generation analogues.................................................................................... 219 

6.1.1 Alkylation of phenols with propargyl bromide .................................................................. 219 

6.1.2 Synthesis of aliphatic Esters ............................................................................................... 226 



vii 
 

6.1.3 Synthesis of Beta-ketonitriles (3-oxopropanenitriles) ....................................................... 227 

6.1.4 Synthesis of 2,4–Diaminopyrimidines ................................................................................ 229 

6.1.5 Iodination of 2,4–Diaminopyrimidines .............................................................................. 231 

6.1.6 Sonogashira coupling reaction ........................................................................................... 233 

6.1.7 Hydrogenation of alkynes .................................................................................................. 250 

6.2 Synthesis of second-generation analogues .............................................................................. 264 

6.2.1 Esterification of carboxylic acid containing phenols and anilines ..................................... 264 

6.2.2 Alkylation of phenols for second-generation analogues ................................................... 267 

6.2.3 Boc protection of anilines .................................................................................................. 270 

6.2.4 Alkylation of protected anilines ......................................................................................... 272 

6.2.5 Deprotection of alkylated anilines ..................................................................................... 274 

6.2.6 Methylation of alkylated anilines....................................................................................... 276 

6.2.7 Sonogashira coupling reaction for second-generation analogues .................................... 278 

6.2.8 Hydrogenation of second-generation alkynes ................................................................... 284 

6.2.9 Removal of Boc group after Sonogashira and reduction ................................................... 290 

6.2.10 Hydrolysis of esters after Sonogashira and reduction ..................................................... 291 

6.2.11 Protection of glutamic acid .............................................................................................. 294 

6.2.12 Addition of glutamic acid ................................................................................................. 295 

6.2.13 Deprotection of glutamic acid ......................................................................................... 297 

References .......................................................................................................................................... 299 

Appendices .......................................................................................................................................... 305 

Appendix A – Parameters for molecular modelling experiments ................................................... 305 

Appendix B – Predicted physiochemical properties ....................................................................... 314 

Appendix C – Full set of data obtained from biological assays ....................................................... 321 

Appendix D – Predicted physiochemical properties for second-generation analogues ................. 327 

Appendix E – Characterization data for synthesized compounds .................................................. 328 



1 
 

Chapter 1 – Introduction 

 

1.1 Malaria 

 

1.1.1 Brief history and introduction 

Malaria is an infectious disease caused by single-celled parasites of the genus Plasmodium. Its 

effects have been known for thousands of years1. In ancient times it was thought to arise from 

the inhalation of contaminated air2, however, in 1880, the French doctor Alphonse Laveran 

discovered that it was in fact caused by blood-borne microorganisms1. He was awarded the 

1907 Nobel Prize in medicine for this discovery. By the end of the nineteenth century several 

other major breakthroughs had been made in the study of this parasite. Three distinct malaria 

causing species were uncovered by Camillo Golgi (Plasmodium vivax, Plasmodium malariae), 

and Marchiafava and Celli (Plasmodium falciparum) in 1886 and 1890, respectively1. The long-

standing puzzle of malaria transmission was solved by Ronald Ross in 1897, where it was 

proven that the parasite was transmitted by Anopheles mosquitos1. For his work in studying 

the life cycle of the malaria parasite in the mosquito host, Ross received the 1902 Nobel Prize 

in medicine. The next fifty years after this ground-breaking discovery saw many insights gained 

into the details of the parasite life cycle. It was revealed that the parasite undergoes different 

stages of development in its two animal hosts – humans and mosquitos – and it was 

demonstrated that humans could be infected via transmission of sporozoites identified in 

infected mosquitos3. In 1948, sporozoites were identified for the first time in a human by 

studying liver samples from an infected individual. This aided the understanding of the full 

parasite life cycle in humans, from injection of sporozoites into the human host during a 

mosquito blood meal, subsequent proliferation in the liver and eventual infection of red blood 

cells by merozoites, to the uptake of differentiated gametocytes upon feeding by another 

mosquito3. In addition to P. vivax, P. malariae, and P. falciparum, two other Plasmodium 

species capable of causing malaria in humans were classified in the early twentieth century. 

Plasmodium ovale was named by John Stephens in 19224, and Plasmodium knowlesi, which 



2 
 

was first identified in simian hosts in 1932 by Campbell and Napier, was later shown to be able 

to infect humans via zoonosis5. Of the five Plasmodium species known to cause malaria in 

humans, P. falciparum accounts for the vast majority of cases and deaths worldwide6.   

 

1.1.2 Summary of Plasmodium falciparum life cycle 

The life cycle of P. falciparum involves two different host organisms, namely humans and 

female Anopheles mosquitos, and is shown in brief in figure 1.1. Parasite sporozoite cells are 

first transferred to humans through the bites of infected mosquitos [Fig 1, A]. They migrate 

through the epithelium into the blood stream and within a few hours reach the liver where 

they infect hepatocyte cells3,7 [Fig 1, B]. In the initial seven days after being infected, 

individuals exhibit no symptoms, and the disease is not transmissible. During this period, the 

parasite cells undergo rapid asexual proliferation, and become differentiated into 

multinucleate schizont cells. These rupture the infected liver cells, releasing numerous 

merozoites into the systemic circulation [Fig 1, C]. It is estimated that a single sporozoite can 

produce up to 10 000 further differentiated merozoite cells in the liver stage of the infection7,8. 

Once in the blood stream, merozoites quickly infect red blood cells and continue rapidly 

dividing [Fig 1, D]. The infected cells rupture, releasing the newly produced parasite cells into 

the blood stream. The cycle of infection continues and is known as the blood stage [Fig 1, E]. 

The red blood cell population in the body becomes severely diminished, which along with the 

effects of the human immune response, results in symptoms of the disease, namely fatigue, 

fever, and anaemia3,9. A small number of merozoites become differentiated into gametocytes 

during the blood stages and remain inside infected red blood cells [Fig 1, F]. When a mosquito 

takes a blood meal, infected cells are delivered to the mid gut of the insect [Fig 1, G]. The 

gametocytes differentiate into male and female gametes, which fuse to form a zygote [Fig 1, 

H]. This develops further into a motile ookinete, which exits the midgut, attaching itself to the 

outer gut membrane. There it develops into an oocyst, a multinucleate cell which produces 

thousands of new sporozoites over several days8 [Fig 1, I]. Once released, the sporozoites 

travel to the salivary glands of the mosquito, where they can subsequently be injected along 

with the saliva into a new human host when the mosquito feeds3,8 [Fig 1, J].  



3 
 

 

Figure 1.1: Summary of P. falciparum parasite life cycle in human and mosquito hosts. 

[Figure created with BioRender.com] 

 

Symptoms of malaria typically show 10 - 15 days after the initial infection, and if not treated 

rapidly, can lead to severe illness and death. Children under five years of age are the most 

vulnerable, and account for two thirds of malaria deaths worldwide6. Malaria is treatable and 

preventable, however, over half a million people die and hundreds of millions more are 

infected each year. 

 

 



4 
 

1.1.3 History of malaria prevention and treatment 

 

1.1.3.1 Small molecule drugs  

There is a long history of treatments and preventative methods against malaria. Some 

progress was made in these efforts before the nature of the Plasmodium parasite and the 

disease itself was understood. The earliest known treatment for malaria comes from the bark 

of the Cinchona tree, which is indigenous to parts of South America9,10. The effect of the bark 

on fevers was long known by local populations, and in 1820 the active ingredient, quinine 

[Fig 1.2], was isolated by French chemists. An effective extraction procedure was developed, 

and it quickly became mass produced and used across the world10. It was the predominant 

malaria treatment until other, synthetic drugs were created in the 1920s11. It is still 

occasionally used in combination with other drugs for mild malaria cases today. Quinine 

combats malaria by preventing parasite cells from digesting haemoglobin during the blood 

stages of the infection, although the full mode of action remains unresolved11.   

Chloroquine, shown in figure 1.2, was the next major drug to be utilized. Developed in 

Germany in the 1930s as an alternative to quinine, this compound rapidly entered mainstream 

use and like quinine, also targets the blood stages of the Plasmodial infection12. Although an 

extremely effective antimalarial in the first few years of its use, parasites quickly developed 

resistance to chloroquine-based treatments10. The ability of the Plasmodium parasite to 

mutate so readily is one of the single most challenging factors in the eradication of malaria.  

Time and time again, new promising antimalarial agents have been developed and have 

shown excellent activity in the initial years of use, before the parasite invariably develops key 

mutations to give it resistance against these new drugs. It has long been a challenge for 

chemists to create molecules which are able to target the parasite over many years and remain 

active despite any mutations in the active site of the targeted enzyme. By the second half of 

the twentieth century, chloroquine resistance was widespread10, so new drug regimens were 

required.  

 



5 
 

 

Figure 1.2: Structures of early antimalarials quinine and chloroquine 

 

Two new compounds, proguanil and pyrimethamine, included in figure 1.3, were developed 

in 1945 and 1951, respectively, and quickly became widely used. Proguanil, which rearranges 

in the body into the active form of the drug cycloguanil, and pyrimethamine are both 

antifolates which target the bifunctional parasite enzyme dihydrofolate reductase - 

thymidylate synthase (DHFR-TS). This enzyme is a very attractive drug target as it is essential 

for folate metabolism in the parasite life cycle. In both the liver stage and the blood stages of 

the parasite infection in humans, the parasite undergoes large rounds of cellular proliferation, 

requiring a multitude of folate cofactors produced by the parasite’s folate metabolism. By 

interrupting this process, the growth of the parasite can be halted13. Because they target both 

the liver stage and the blood stages, antifolate drugs can be used to prevent the symptoms of 

malaria from developing in the first place, eliminating the parasite in the liver before it enters 

the blood stream, or can be used to treat the symptoms of malaria after they manifest, halting 

the progress of the parasite in the blood stages14.  

 

 

Figure 1.3: Structures of antimalarial class II antifolates cycloguanil and pyrimethamine, and 
the prodrugs proguanil and chlorproguanil 

 



6 
 

After the great initial success of these drugs, the issue of parasite resistance due to specific 

mutations however soon became apparent, and new combination treatment strategies were 

adopted3. Sulfones and sulfonamides, shown in figure 1.4, are another class of compounds 

which interrupt the folate metabolism in parasite cells. These molecules target the enzyme 

dihydropteroate synthase (DHPS) and are thus known as class I antifolates. Cycloguanil, 

pyrimethamine, and other compounds which target the DHFR-TS enzyme are called class II 

antifolates15. By combining both class I and class II antifolates into a single treatment regime, 

the advent of drug-resistance can be delayed, as mutations in two separate parasite enzymes 

are required for complete resistance10. Combination treatments became the norm, with 

pyrimethamine and sulfadoxine being sold under the brand name Fansidar, and 

chlorproguanil (an analogue of proguanil) [Fig 1.3] and dapsone marketed as Lapdap16. 

Despite early success, these treatments soon also lost their effectiveness against mutant 

parasite strains.  

 

 

Figure 1.4: Common antimalarial class I antifolates sulfadoxine and dapsone 

 

Mefloquine, included in figure 1.5, is a compound that was developed as a result of a large-

scale screening campaign conducted in the 1960s and 1970s with the aim of finding new 

antimalarials17. Like quinine and chloroquine, this molecule contains a quinoline functional 

group, and similarly targets parasite development in the blood stages of the infection12. 

Mefloquine was first used in parts of Asia in the 1970s and become licenced for use in the 

United States in 1989. Today, parasite resistance to mefloquine is widespread10.  

In the modern era the most widely used antimalaria treatment is artemisinin combination 

therapy (ACT). The key component of this regimen is artemisinin, which is shown in figure 1.5. 

This natural product was first isolated from the Artemisia annua plant by Chinese researchers 



7 
 

in 19729. The medicinal value of this plant has been long recognized by indigenous Chinese 

populations, and the isolation of the active ingredient led to much further research. 

Structurally, artemisinin differed drastically from other antimalarial compounds known at the 

time. It is comprised of a complex fused ring system, with the sesquiterpene lactone and 

endoperoxide rings proving essential for activity10. Artemisinin acts through a complicated 

multistep mechanism, involving the generation of free radicals from the peroxide bridge which 

disable key parasite proteins18. The exact steps involved in this process are still unclear. During 

the blood stages of a malaria infection, artemisinin inhibits both the asexual reproduction of 

parasite merozoites, and the formation of gametocytes. The disruption of gametogenesis 

prevents further transmission of the disease from infected individuals treated with 

artemisinin19. This compound has shown excellent activity against many mutant strains of 

Plasmodium and is used today in combination for the first line treatment of drug-resistant 

malaria infections10. In an effort to delay the advent of resistance against artemisinin,  

combination strategies are used. Despite this, partial artemisinin resistance was first observed 

in the early 2000s20. The development of resistance against artemisinin-based therapies poses 

a major threat to the ability to treat malaria, as no other drugs are poised to take their place 

as a first line defence against new infections10.     

 

 

Figure 1.5: Structures of mefloquine and artemisinin antimalarials  



8 
 

Table 1.1 includes a brief summary of several small molecule drugs, highlighting details about 

their usage, mode of action, and information about ongoing clinical trials if relevant.  

 

Table 1.1: Summary of small molecule antimalarial drugs  

Compound Target/Mechanism Comments Clinical Approval 

Status 

Quinine Inhibits haem 

detoxification 

 

One of the oldest 

antimalarial agents; still 

used for severe and 

resistant cases 

Approved – Historical 

use since the 17th 

century; now largely 

replaced by newer 

treatments 

Chloroquine Inhibits haem 

polymerization 

Former first-line therapy; 

effectiveness severely 

limited by resistance 

Approved – Introduced 

in the 1940s; limited 

current use due to 

widespread resistance 

Cycloguanil Class II antifolate; 

Inhibits dihydrofolate 

reductase (DHFR), 

blocking folate synthesis 

Active metabolite of 

proguanil; sometimes 

included in combination 

therapies and as a model 

for research into new 

drugs 

Research/Combination 

Use – Not used as a 

standalone agent 

Pyrimethamine Class II antifolate; 

Inhibits dihydrofolate 

reductase (DHFR), 

blocking folate synthesis 

Commonly used in 

combination with 

sulfadoxine  

Approved – Widely 

used in combination 

therapies despite 

resistance challenges 

Sulfadoxine Class I antifolate; 

Inhibits dihydropteroate 

synthase (DHPS), 

interfering with folate 

synthesis 

Used in synergy with 

pyrimethamine; 

standalone efficiency 

reduced due to resistance 

Approved – Employed 

in combination with 

noted resistance issues 

Mefloquine Effective against resistant 

strains; some negative 

Approved – In clinical 

use since the 1980s 



9 
 

Compound Target/Mechanism Comments Clinical Approval 

Status 

Inhibits haem 

detoxification, similar to 

quinine 

side effects and reduced 

efficacy due to resistance  

Artemisinin Generates reactive free 

radicals that damage 

parasite proteins upon 

activation 

Cornerstone of modern 

malaria treatment; 

rapidly acting 

Approved – Integral 

component of 

Artemisinin-Based 

Combination Therapies 

(ACTs) 

Artesunate Artemisinin derivative 

with improved solubility 

and bioavailability 

Preferred for severe 

malaria due to rapid 

action 

Approved – WHO-

recommended for 

severe malaria 

treatment 

Lumefantrine Inhibits haem 

detoxification  

Typically combined with 

other antimalarials for 

enhanced efficacy 

Approved – Key 

component of ACTs 

since the early 2000s 

WR99210 Class II antifolate; 

Potent inhibitor of 

dihydrofolate reductase 

(DHFR), blocking folate 

synthesis 

Known for its high 

potency as an antifolate; 

primarily used as a 

research tool and 

currently undergoing 

early clinical trials 

Under Early-

Phase/Exploratory 

Clinical Evaluation 

 

1.1.3.2 Vaccines  

It has been a goal of scientists for many decades to produce a vaccine that is effective in 

preventing malaria, but this has proved extremely challenging. By the end of the 20th century, 

successful vaccines for many common diseases had already been developed and implemented 

across the globe. The earliest crude vaccine was the smallpox vaccine, created by Edward 

Jenner in 1796. Almost one hundred years later, Louis Pasteur introduced attenuated vaccines, 

and effective vaccines against rabies and cholera in humans were later developed. The 1920s 

saw the creation of vaccines for diphtheria and tetanus, and the first polio vaccine was 



10 
 

produced in the 1950s by Polish virologist Hilary Koprowski21. Many other key vaccines were 

also produced in the second half of the 20th century, including inoculations against 

Haemophilus influenzae, Hepatitis A, and Rotavirus, to name a few.  

While many other diseases have all but been eradicated by the implementation of vaccines, 

the threat of malaria is still ever-present today. Only recently have vaccines for this disease 

become available, and they appear to show only moderate efficacy. This raises the question 

as to why a malaria vaccine has been out of reach for so long while huge progress has been 

made in the development of vaccines for other diseases. The answer to this is many-fold, with 

the single biggest factor being the greater biological complexity of the malaria parasite. While 

bacterial and viral infections are characterized by a limited number of antigens that can result 

in an immune response, parasite infections often involve several different stages of 

development in their life cycle, each with their own unique suite of associated proteins and 

other potential antigens22. This means that there is no single antigen which would be able to 

elicit a broad human immune response against all the stages of parasite development. In 

addition, the Plasmodium parasite has proven to be highly adept at evading the human 

immune system, and even in individuals who have been previously infected with malaria, a 

complete immunity is not developed22. The vast number of potential antigens displayed by 

the malaria parasite, and the inadequacies of the natural human immune system against this 

disease have made selecting antigens for a potential malaria vaccine very challenging2,22.  

So far two different broad types of vaccine have been investigated. Single target and 

combination vaccines. Single target vaccines contain a single parasite antigen against which 

an immune response is developed, while combination vaccines contain a cocktail of different 

antigens that can elicit a broader immune response. Combination vaccines are currently held 

to a higher standard than their single component counterparts as they may be able to 

generate an immune response against the parasite throughout the course of its life cycle. For 

a single target vaccine to be viable, it needs to be completely effective at eliminating all 

parasite cells before they can progress to the next stage of development. For example, a 

hypothetical single target vaccine which creates immunity against sporozoites will lose much 

of its effectiveness if even a single sporozoite is able to enter the liver and become 

differentiated into a hepatic schizont, which will go onto produce numerous merozoites and 

initiate the blood stages of the disease22. Similarly, for a single target vaccine which creates 



11 
 

immunity against the blood stages of the parasite, some merozoites may escape and become 

further differentiated into gametocytes, which would allow for the subsequent transmission 

of the disease to a new feeding mosquito. An additional challenge faced by single target 

vaccines is the high degree of protein polymorphism exhibited by the parasite, particularly in 

the blood stages of the infection2. This results in a best-case scenario of only partial immunity 

against this stage of the disease. For these reasons, it is unlikely that any single target vaccine 

will ever be highly effective. The development of this type of  vaccine is still essential, however, 

as several different single target vaccines which are effective against specific stages of parasite 

development, or effective against specific polymorphic proteins, may be combined into a 

combination vaccine which is capable of generating robust immunity against each stage in the 

parasite’s life cycle.  

One last factor which cannot be overlooked in contributing to the difficulty in producing a 

malaria vaccine is the lack of financial incentives for large pharmaceutical companies. Malaria 

is a disease that largely affects impoverished tropical areas, and pharmaceutical companies 

are hesitant to spend millions of dollars in the research and development of a vaccine that will 

never generate large profits. The funding for malaria vaccine research has come mostly from 

humanitarian initiatives, such as the Bill and Melinda Gates Foundation. The rapid success in 

the development of the Covid 19 vaccine serves to demonstrate that if a collective, global 

effort is enacted for the development of a new vaccine, then progress can be made rapidly. 

Apart from the biological challenges in developing a malaria vaccine discussed earlier, this lack 

of broad international interest in vaccine development for a disease that largely infects 

impoverished communities is the largest factor influencing the slow progress in reaching this 

goal22.  

While several vaccine candidates are currently undergoing clinical trials, only two have been 

approved by the World Health Organization (WHO) for widespread use. The first of which is 

the RTS,S vaccine, developed by GlaxoSmithKline (GSK) pharmaceuticals2,23. Conceptualized 

in the 1970s, this vaccine underwent more than 40 years of development and testing before 

being recommended for the prevention of malaria in children by the WHO in 2021. Today, it 

is distributed mainly in sub-Saharan Africa, and as of the end of 2023, has been administered 

to more than two million children6. This vaccine makes use of the P. falciparum 

circumsporozoite protein (CSP), which is expressed on the surface of parasite cells during early 



12 
 

liver phase development21. This protein is a major antigen recognized by the immune system 

in the early stages of parasite infection and leads to a strong immune response. The vaccine 

is made up of a truncated portion of the CSP antigen, lacking the N-terminus, which is required 

for entry into host cells21,23, fused with hepatitis B surface antigen (HBsAg), which acts as a 

carrier and immunogenic adjuvant22. In field trials, the RTS,S vaccine exhibited moderate 

efficacy of 30 – 50 %, and was shown to decrease the severity of malaria cases in children by 

lowering the parasite load which enters the blood stages22. Despite this, the immunity 

developed after administration of the RTS,S vaccine diminishes quickly, and 18 months after 

receiving the final dose no appreciable immunity can be observed24.  

Outlined in the malaria vaccine road map, the WHO has set a goal for the development of a 

vaccine which shows both lasting protective efficacy of at least 75 %, and transmission 

blocking ability by the year 20306. While the RTS,S vaccine falls short of this target, it is still an 

important first step in realizing this goal and can nevertheless be used in the meantime as a 

tool for easing the burden of malaria until a superior vaccine candidate is developed.  

The second malaria vaccine to date which has been approved for widespread use is the R21 

vaccine. This vaccine was developed by scientists at the University of Oxford in collaboration 

with Novavax, a biotechnology company. The R21 vaccine has many similarities to the RTS,S 

vaccine. It also makes use of a truncated portion of the CSP to incite an immune response and 

is supported by hepatitis B derived adjuvants. The main difference between the two vaccines 

is the formulation of the virus-like-particles (VLPs) which are responsible for presenting the 

chosen antigen to the immune system. The RTS,S vaccine contains both fused CSP-HBsAg 

proteins as well as monomeric HBsAg, while the R21 vaccine is made up of only the 

recombinant CSP-HBsAg protein. This allows for the R21 VLPs to display a higher surface 

density of the CSP antigen, eliciting a stronger immune response23. The early-stage clinical 

trials of the R21 vaccine yielded very promising results. Efficacy of up to 75 % was observed in 

children 12 months after the initial three dose regimen, with a fourth dose extending the 

duration of immunity6. This vaccine was recommended by the WHO in October 2023, and is 

currently being distributed alongside the RTS,S vaccine in areas of high risk.   



13 
 

1.1.3.3 Vector control methods 

Apart from medicines and vaccines, vector control remains one of the most promising 

strategies for combating malaria. Hypothetically, if every single Anopheles mosquito was 

eradicated today, then malaria would soon disappear completely without any medical 

interventions. This, however, poses an even greater challenge than the development of new 

antimalarial drugs or a universal vaccine. Despite this, vector control methods have been 

widely implemented, and continue to play a crucial role in limiting the number of cases.  

The two main methods of mosquito vector control used today are the indoor residual spraying 

(IRS) of insecticides, and the use of insecticide treated nets (ITNs). The former involves coating 

the interior surfaces of dwellings with potent insecticides which are able to eliminate any 

mosquitos that come into contact with it. The latter combines the use of nets, which serve to 

create a physical barrier between humans and mosquitos, blocking any potential 

transmission, with insecticides that enhance this effect25. 

The most widely used insecticide for IRS to date is dichlorodiphenyltrichloroethane (DDT), the 

structure of which is included in figure 1.6. This compound was first synthesized in the late 

1800s, but it was not until 1939 that its remarkable insecticidal properties were identified10. 

In low concentrations, this compound is able to disrupt the nervous system of arthropods by 

interfering with the normal functioning of sodium ion channels, leading rapidly to death26. 

The protection offered by DDT IRS lasts up to six months27, and it has been a very effective 

method for reducing the incidence of malaria in highly affected areas28,29. This compound is 

however not without its drawbacks. After more than 60 years of widespread use, mosquito 

populations have developed resistance to DDT27,29. In addition, there are numerous concerns 

about its toxicity, with some evidence suggesting a link between DDT and the development of 

certain types of cancer in humans30.  Additionally, DDT is classified as a persistent organic 

pollutant for its ability to remain intact in ecosystems long after its introduction29. For these 

reasons, the widespread use of DDT was banned in 1972 by the US environmental protection 

agency. Despite the ban, several countries received exemptions and continued the use of DDT 

as a vector control agent in the face of malaria outbreaks. In 2006 the WHO declared DDT 

acceptable for use in areas of high malaria risk and it is still often used today25.  



14 
 

For use in ITNs, pyrethrums, shown in figure 1.6, are the insecticide of choice25. This class of 

compound was first extracted from the flowers of the Tanacetum cinerariifolium plant, the 

insecticidal properties of which have been known for thousands of years4. Numerous synthetic 

analogues have also been created, which are named pyrethroids [Fig 1.6]. These compounds 

exhibit a similar mode of action to DDT, involving the disruption of sodium ion channels inside 

the cell membranes of arthropod cells2. Although widely successful for a number of years, 

many cases of mosquito resistance to pyrethroid based insecticides have been reported. Due 

to this, various combination approaches are currently recommended for ITNs in areas of 

pyrethroid resistance6.  

 

 

Figure 1.6: Structures of the insecticides DDT, the natural pyrethrin 1 and synthetic 
allethrin 1 

 

1.1.4 Social and economic burden of malaria 

Although malaria has been around for many thousands of years, this disease still represents a 

massive health crisis today. In 2022, the estimated number of malaria cases reached 249 

million, with more than 600 000 deaths reported6. While malaria is endemic to more than 80 

countries across the world, the greatest burden of this disease is felt in Africa, which 

accounted for more than 90 % of the total cases and deaths in 20226.  

Due to continued, directed efforts at combating malaria over the last 20 years, substantial 

progress has been made. In the year 2000, an estimated number of 262 million infections, and 

864 000 deaths were reported6. These figures have dropped each year reaching a minimum 

of 229 million cases and 576 000 deaths in 2019. However, due to the impacts of the global 

Covid-19 pandemic which started in early 2020, the number of cases and deaths saw a slight 



15 
 

increase. Humanity is still recovering from these setbacks, and the number of deaths and cases 

reported in 2022 are marginally above the pre-pandemic figures6.  

The global malaria mortality rate has also fallen consistently since 2000. This is defined by the 

number of deaths reported per 100 000 individuals at risk. Improvements in this statistic are 

made both by limiting the number of at-risk individuals, mainly through vector control 

methods, and also improvements in the treatment of malaria cases through medical 

interventions, lowering the proportion of deaths after infection. In 2000 the mortality rate 

was 29 deaths per 100 000 individuals at risk. This reached its lowest value of 14 in 2019. An 

increase in mortality rate was observed in 2021, reaching 15.2. Since then however some 

improvement has been made and the mortality rate reported in 2022 was 14.3, which is close 

to the pre-pandemic levels6.  

One key reason for the success achieved in the last 20 years is the increase in the coverage of 

vector control methods. These are relatively cheap, and highly effective at reducing 

transmission rates. In 2022, 282 million ITNs were distributed in malaria endemic countries. 

More than half of these were pyrethroid-PBO nets, which are treated with a combination of 

pyrethroid and piperonyl butoxide (PBO) insecticides. In many sub-Saharan African nations 

ITNs are the main form of vector control, and it is estimated that in 2022 around 70 % of 

households in sub-Saharan Africa had at least one ITN, a substantial increase from 5 % 

reported in 2000. This increase has however stagnated since 20156. In 2022, 47 countries 

implemented IRS campaigns to prevent the spread of malaria, although the coverage of this 

is even less than that of ITNs, with an estimated 1.8 % of at-risk populations being protected 

by IRS6. 

Children under the age of five years and pregnant women are the most vulnerable to malaria, 

and so special attention needs to be paid to these demographics. Young children accounted 

for 76 % of the total number of deaths from malaria in 2022. Women who have been infected 

with malaria during pregnancy are more likely to die during childbirth, and children born 

under these conditions have a higher risk of child mortality and experiencing various 

developmental issues. In some heavily afflicted areas, up to 40 % of pregnant women had 

malaria infections in 2022. This is very alarming given the potential effects of malaria during 

pregnancy and makes combating malaria during pregnancy an issue of vital importance. It is 



16 
 

therefore recommended that pregnant women who may be exposed to malaria regularly take 

sulfadoxine-pyrimethamine formulations as a preventative measure6. 

In 2016, the WHO and partners introduced a global technical strategy (GTS) for the eventual 

eradication of malaria. This proposal outlined specific goals and provided recommendations 

for reducing the malaria mortality rate and lowering the number of malaria endemic countries 

by the year 2030. Compared to the baseline figures obtained in 2015, a decrease in the 

mortality rate of 40 % by 2020, 75 % by 2025, and 90 % by 2030 were targeted. Furthermore, 

it is hoped that by 2030, at least 35 new countries will be declared malaria free. As of the year 

2022 these targets are not on track to be met.   

As efforts to combat malaria continue to scale up, it is evident that an increase in funding is 

required. In 2022, an estimated $ 4.1 billion was invested in the fight against malaria. This 

represents an increase from previous years, although still far short of the $ 7.8 billion target 

defined by the GTS for 20226. Most of this money is required for the manufacturing and 

distribution of pharmaceuticals and ITNs, as well as the rollout of the recently developed 

vaccines. Aside from this, it is also vital to allocate funds for the research and development of 

new strategies for fighting malaria. ACTs continue to be the pharmaceutical front line defence 

against malaria, but their long-term effectiveness isn’t guaranteed. It is crucial that the drug 

development pipeline continues to produce new molecules to take the place of ACTs, which 

have been extensively used over the last two decades.  

The WHO estimates that due to human intervention against malaria, namely the widespread 

distribution of antimalarial drugs and the implementation of vector control methods, a total 

of 2.1 billion cases and 11.7 million deaths have been prevented between 2000 and 20226. 

This is a remarkable achievement, and it is important to consider this when looking at the vast 

number of cases and deaths that are still reported each year. Even though the threat of malaria 

is still ever-present in many parts of the world today, the situation could be significantly worse, 

and there is still continued hope that the complete eradication of this disease is a feasible 

undertaking.    



17 
 

1.2 Details of folate metabolism and review of antifolate drugs 

 

1.2.1 Overview of folate metabolism  

Folate metabolism is an ancient metabolic pathway that exists in some form in all organisms. 

Folate derivatives acquired from this process are essential cellular cofactors required for 

nucleic acid biosynthesis and DNA replication. Rapidly dividing cells such as cancer, bacteria, 

and parasite cells are heavily dependent on a large supply of folate cofactors for their 

continued proliferation16. The use of compounds which can inhibit the folate metabolism in 

these cells are highly effective at halting their progress, and as such antifolate drugs have been 

widely used in the treatment of various diseases.  

During the P. falciparum life cycle in humans, there are several stages which involve large 

rounds of asexual cellular proliferation, namely the initial formation of merozoites during the 

liver stage and the continued cycle of infection – proliferation – lysis which repeats many times 

during the blood stages. The use of antifolates in the treatment of malaria disrupts these 

processes, and can serve as an effective prophylaxis, halting the development in the liver 

stage, as well as a symptomatic treatment, combating the continued proliferation of parasite 

cells during the blood stages.  

Cells obtain folates in one of two ways, either through de novo synthesis or via a folate salvage 

pathway31. The former involves numerous steps and enzymes, and results in the formation of 

the key cofactor dihydrofolate (DHF), which is subsequently converted to tetrahydrofolate 

(THF) by the enzyme dihydrofolate reductase (DHFR). Animals lack serval key enzymes 

involved in de novo synthesis and must therefore obtain their folate cofactors through a 

different pathway. Typically, the required folate precursors are obtained through diet in the 

form of folic acid. This is converted by the animal’s own DHFR enzymes to DHF, which is then 

further metabolised to THF by the same enzyme. As humans are unable to synthesize their 

own folates from scratch, folic acid, also called vitamin B9, is an essential nutrient32. 

Plasmodium cells have access to both metabolic pathways for producing their folate cofactors 

and can also scavenge folates from the infected host cells.   

 



18 
 

Folate derivatives are structurally conserved and are made up of three key components. The 

first is a pterin ring, which is attached to para-aminobenzoic acid (PABA), and finally a glutamic 

acid residue. These elements are highlighted on the structure of dihydrofolate in figure 1.7.  

 

 

Figure 1.7: Key structural elements of DHF 

  

An overview of de novo folate metabolism in Plasmodium cells is shown in scheme 1.1. The 

first step entails the rearrangement of gaunosine-5’-triphosphate (GTP) to 7,8-

dihydroneopterin triphosphate (DHNTP) and is achieved by the enzyme GTP cyclohydrolase. 

This forms the core pterin ring system present in all folate derivatives. While the GTP 

cyclohydrolase enzyme is also found in animal cells, it does not play the same role in folate 

synthesis as several of the enzymes required for the later steps are absent33. The triphosphate 

tail of the DHNTP is then cleaved by phosphatase enzymes, resulting in the formation of 7,8-

dihydroneopterin (DHN), which is subsequently converted to hydroxymethyldihydropterin 

(HMDP) by the enzyme dihydroneopterin aldolase (DHNA), with the liberation of 

glycolaldehyde. A bifunctional parasitic enzyme in Plasmodium cells catalyses the next two 

transformations. A diphosphate group is first added to the terminal hydroxy of HMDP by the 

hydroxymethyldihydropterin pyrophosphokinase (HPPK) domain. The resulting dihydropterin 

pyrophosphate (DHPP) then undergoes a condensation reaction with p-aminobenzoic acid 

(PABA), catalysed by the dihydropteroate synthase (DHPS) domain of the bifunctional enzyme, 

to form dihydropteroate (DHP)33. Finally, DHP is modified by the addition of a glutamic acid 

residue, facilitated by the enzyme dihydrofolate synthase (DHFS). The resulting product is 

dihydrofolate (DHF), which is reduced to tetrahydrofolate (THF) by dihydrofolate reductase 

(DHFR)14. 



19 
 

 

Scheme 1.1: Overview of de novo folate metabolism in Plasmodium cells, steps also involved 
in the folate salvage route indicated by dashed blue outline 

 

The folate salvage pathway is a convergent route that bypasses many of the early steps of THF 

synthesis found in de novo folate metabolism. While de novo synthesis begins with GTP and 

proceeds through several intermediates, the salvage pathway relies on later folate precursors 

that are typically scavenged from the host’s bloodstream. The points at which the salvage 

pathway converge with the de novo synthesis route are included in scheme 1.1 above in the 



20 
 

dashed blue area. Class I antifolates inhibit the enzyme DHPS, effectively blocking the de novo 

folate synthesis pathway. However, this can be bypassed via the salvage pathway if later folate 

precursors, such as DHP or DHF, are scavenged from the host. In contrast, Class II antifolates 

target DHFR, inhibiting the conversion of DHF to THF, effectively blocking both the de novo 

and salvage routes of folate metabolism. 

Inside the cell, the later folate derivatives usually exist in polyglutamated forms, containing up 

to ten additional glutamic acid residues. While monoglutamate forms are readily transported, 

the polyglutamate forms are not, and build up to high concentrations within the cell. The 

addition and removal of extra glutamate residues serve to regulate the concentration of folate 

cofactors, ensuring that cells requiring large amounts of folates have sufficient access to them 

in the polyglutamated form. In addition, the polyglutamated forms have enhanced abilities as 

enzyme substrates34.  Two main enzymes facilitate this process, extra glutamic acid residues 

are sequentially added to the C-terminus of the glutamic acid present on the folate derivatives 

by the enzyme folylpolyglutamate synthase (FPGS). A second enzyme, gamma-glutamyl 

hydrolase, is able to cleave the extra glutamate residues resulting in the monoglutamate form 

of the folate cofactors for subsequent cellular transport34. These compounds are highly polar 

and require active transport mechanisms to cross cell membranes. In human cells there are 

three main pathways for facilitating the entry of the mono-glutamated folate cofactors into 

cells. These include the transmembrane proteins reduced folate carrier (RFC), the proton-

coupled folate transporter (PCFT), and various folate receptors34.  

THF is the final product of both the de novo biosynthesis pathway, and the folate salvage route. 

This molecule is able to facilitate numerous key transformations required for cell survival. In 

particular it acts as an essential cofactor in one-carbon-transfer reactions33. Single carbon 

units are added to THF in various forms by enzymes, including formyl, methylene, and methyl 

C1 units33. These are then transferred to other substrates in several key biosynthetic pathways, 

including purine and thymidylate synthesis, which are required for DNA and RNA production, 

as well as the synthesis of various amino acids needed for protein formation. Once these 

transformations are complete, the C1 THF derivatives become oxidized back to DHF, which can 

subsequently be reduced to THF, continuing the metabolic cycle.    



21 
 

1.2.2 Antifolates in the treatment of various diseases  

 

1.2.2.1 Treatment of malaria with antifolates 

Many different compounds have been investigated as antifolates for their role in disrupting 

folate metabolism. While there are dozens of enzymes involved in this process, the most 

success has come from targeting two in particular. These are dihydropteroate synthase (DHPS) 

and dihydrofolate reductase (DHFR). Drugs which inhibit the former enzyme are known as 

class I antifolates, while drugs inhibiting the latter enzyme are class II antifolates. Antifolate 

drugs have shown to be very effective against P. falciparum strains but show significantly lower 

activity against other malaria causing Plasmodium species13.  

In the treatment of P. falciparum infections, both class I and class II antifolates have been 

widely used, commonly in combination to enhance the effects. The class I antifolates largely 

consist of sulfadrugs, such as dapsone and sulfadoxine [Fig 1.4]. These resemble one of the 

natural substrates for the DHPS enzyme, PABA. The structural similarities are presented in 

figure 1.8. Dapsone is the most potent known inhibitor of P. falciparum DHPS. It was originally 

synthesized in an attempt to create new dye molecules in the early 1900s and was shown to 

have a significant antimicrobial effect in the 1930s. Further testing also demonstrated the 

effects of this compound as an antimalarial, and today it is used to treat a number of diseases, 

including bacterial and fungal infections, as well as malaria14. While this drug is somewhat 

ineffective when administered in isolation, by combining it with class II antifolates a significant 

increase in the antimalarial effect is observed14. While the target enzyme of class I antifolates, 

DHPS, may be blocked by class I antifolates, the parasite can still obtain folate cofactors 

through the salvage pathway. In order for a complete reduction in folates available for parasite 

cells, the second pathway also needs to be targeted. This can be achieved by combining both 

class I and class II antifolates into a single treatment. While combination treatments have the 

advantage of delaying the development of parasite resistance, resistance still invariably 

emerges after prolonged use of the same active compounds.  

 



22 
 

 

Figure 1.8: Conserved structural elements of PABA and the class I antifolate dapsone 

 

Widely used class II antifolates in the treatment of malaria include compounds such as 

proguanil, chlorproguanil, and pyrimethamine [Fig 1.3]. These inhibit the parasite enzyme 

dihydrofolate reductase by mimicking the natural enzyme substrate DHF. Typically, class II 

antifolates are competitive reversible inhibitors, and work by outcompeting DHF at the active 

site of the DHFR enzyme35. Cycloguanil and its analogues contain a 2,4-diamino-

dihydrotriazine core, while pyrimethamine and its analogues contain a 2,4-diaminopyrimidine 

core. These functional groups resemble the pterin moiety found in DHF, as shown in figure 

1.9, and are essential for binding to the active site of the target enzyme. The binding mode of 

the class II antifolates to the active site of P. falciparum DHFR is highly conserved, and a 

minimum required pharmacophore for these compounds has been identified. This is shown 

in figure 1.10, and includes hydrogen bonds to residues Ile14, Ile164, Asp54, and an additional 

charge mediated interaction with Asp54. Some of these binding elements are not present for 

the natural enzyme substrate DHF, and so compounds which contain all of them can 

outcompete DHF in the active site of the target DHFR enzyme.  

 



23 
 

 

Figure 1.9: Conserved structural elements of DHF and class II antifolates 

 

 

 

Figure 1.10: Binding mode of class II antifolates in the PfDHFR enzyme 

 

1.2.2.2 Treatment of cancer with antifolates 

Cancerous cells are marked by rapid, uncontrolled proliferation. In this process, cells require 

large amounts of folate cofactors, and thus antifolate drugs which limit their supply are 

effective agents in the treatment of cancer. One of the first antifolate compounds to be used 

in this manner is aminopterin35, shown in figure 1.11. This molecule closely resembles the 

natural substrate of the DHFR enzyme, DHF. Both contain a pteridine core moiety, but 

aminopterin has an amino substituent in the 4-position, while DHF has a carbonyl. This 



24 
 

enables additional interactions in the active site, with the primary amine of aminopterin able 

to both accept and donate hydrogen bonds to nearby residues, allowing it to outcompete the 

natural enzyme substrate. Aminopterin was first used to treat certain types of leukaemia in 

children. It showed great promise but was phased out because it had an undesirable toxicity 

profile. Methotrexate (MTX), included in figure 1.11, was its replacement for widespread use. 

Although this compound still displayed some toxicity towards healthy cells, it was a marked 

improvement over aminopterin.  

 

 

Figure 1.11: Structure of anticancer antifolates aminopterin and methotrexate (MTX) 

 

The structures of aminopterin and methotrexate are very similar. The nitrogen present in the 

linker of aminopterin is a secondary amine, while in methotrexate it is methylated, becoming 

a tertiary amine. This small change improved the toxicity profile of the drug and allowed for a 

similar level of anti-cancer activity. MTX became the antifolate of choice in the treatment of 

certain cancers in the 1950s and is still in widespread use today35. While successful in practice, 

the detailed mechanism of how MTX is able to inhibit tumour cells was unknown for many 

years. This lack of a resolved mode of action led to challenges in developing other antifolate 

compounds in subsequent years. New anticancer antifolates only become available in the 

early 21st century. Pemetrexed and pralatrexate, shown in figure 1.12, are two such 



25 
 

compounds. Both are structurally related to MTX, with pemetrexed containing a modified 

guanine core and lacking the nitrogen atom in the linker. Pralatrexate also lacks the nitrogen 

atom in the linker, instead containing a carbon with a propargyl substituent. Both of these 

compounds are also effective anticancer antifolates and remain in use today.  

 

 

 

Figure 1.12: Structures of recently developed anticancer antifolates pemetrexed and 
pralatrexate 

 

One important aspect of treating cancer with antifolate drugs is the cellular transport of these 

compounds. Anticancer antifolates can be broken down into two general categories. The first 

are classical antifolates; these are defined as containing a glutamic acid residue36. Like the 

natural folates, these are able to become polyglutamated, which has significant effects on the 

distribution of these drugs throughout the body. Upon administration, they exist in the 

monoglutamated form. As they are extremely polar, and exist as anions under physiological 

conditions, various transmembrane transport proteins are required for entry into cells. Once 

inside the cell, they undergo polyglutamation, preventing them from exiting and resulting in 

bioaccumulation. This is a favourable process as it ensures that the drugs become 

concentrated inside tumour cells, where they can exhibit their desired mode of action. 



26 
 

Resistance of cancer cells to this type of antifolate is largely due to changes in the structure of 

the various transport proteins required for entry into the cancerous cells34. If mutations in 

these enzymes prevent the classical antifolates from entering the targeted cells, then they will 

have no means of exerting their desired action. To overcome this, another class of anticancer 

antifolates have been developed. These are named non-classical, or lipophilic antifolates34. 

Lacking the glutamic acid portion of their classical counterparts, they are not dependent on 

active cellular transport, and their increased lipophilicity allows them to enter cells though 

passive diffusion. One such compound is trimetrexate, shown in figure 1.13. 

 

 

Figure 1.13: Structure of lipophilic antifolate trimetrexate 

 

Trimetrexate is a structural analogue of MTX, containing a 5-methyl quinazoline ring instead 

of the pterin ring found in MTX. In addition, it lacks the glutamic acid substituent on its 

terminal ring, and instead contains a trimethoxy functionality. This compound has shown good 

efficacy against MTX resistant cancer strains as it can operate independently of mutated 

transport proteins which confer resistance to the classical antifolates. Despite this, some level 

of resistance can still develop against the non-classical antifolates, although in this case the 

resistance is due to mutations in the active site of the target enzyme itself35.  

 

1.2.2.3 Treatment of bacterial infections with antifolates 

Folate metabolism is an essential pathway for all bacterial species. Compounds inhibiting this 

process have been widely used as antibiotics. The earliest known being prontosil [Fig 1.14], 

which was first synthesized as a dye molecule and later shown to have potent anti-microbial 

activity37. It is a prodrug for the active compound sulfanilamide [Fig 1.14]. Containing similar 



27 
 

functional groups to PABA, this molecule binds to the bacterial enzyme DHPS, blocking the de 

novo folate synthesis pathway in bacteria cells. Many other antibiotic compounds exist which 

target this enzyme, largely consisting of similar sulfur containing PABA mimics38.  

 

 

Figure 1.14: Structures of antibacterial antifolates sulfanilamide and sulfamethoxazole, and 
the prodrug of sulfanilamide, prontosil 

 

In addition to DHPS, DHFR is the other key enzyme target for antibiotics inhibiting folate 

metabolism. Trimethoprim, shown in figure 1.15, was introduced in the late 1960s and has 

been used in combination with another DHPS inhibitor, sulfamethoxazole [Fig 1.14] very 

successfully for a number of years37. Unfortunately, bacterial resistance has developed against 

antifolate medications. Point mutations in the active site of the target enzymes render the 

currently used drugs unable to bind effectively37.  

 

 

Figure 1.15: Structures of antibacterial antifolates targeting DHFR, trimethoprim, iclaprim, 
and a propargyl linked antifolate 

 

Recent efforts are directed at developing new compounds which can overcome this resistance. 

Iclaprim, included in figure 1.15, is a potential drug candidate which has shown good levels of 

activity against trimethoprim resistant strains and progressed through various clinical trials 



28 
 

until some negative side effects were identified37. Other compounds in development include 

the propargyl-linked antifolates [Fig 1.15], which contain an unsaturated three-atom linker 

between its ring systems39. Various analogues of this nature have been produced and show 

some promise for development into new generation antibiotics.   

 

1.2.3 Resistance to antifolates in the treatment of malaria.  

Widely used in the treatment of malaria for many decades, both class I and class II antifolates 

have lost much of their activity due to the development of resistance in parasite populations. 

Point mutations in the active site of the target enzymes prevent drug molecules from binding, 

while maintaining high affinity for the enzymes’ natural substrates.  

A single A437G point mutation in the active site of the parasite DHPS enzyme leads to 

significant resistance against sulfadoxine and other related class I anti-folates33. This single 

amino acid substitution, alone or combined with other point mutations including A581G, 

S436A, and K540E, results in a many-fold loss of activity against sulfadoxine13. Many P. 

falciparum strains have been identified in various parts of the world containing one or more 

of the above listed drug-resistant mutations. It is thought that many of these mutations 

initially developed in parts of South-East Asia and South America but have become 

widespread today due to human migration patterns33. 

Although initially highly successful, class II antifolates have also been rendered ineffective due 

to mutations developing in the target enzyme DHFR. In P. falciparum, a single bifunctional 

DHFR-TS enzyme is encoded for, containing a DHFR domain and a thymidylate synthase (TS) 

domain. While the DHFR domain converts DHF to THF in the presence of NADPH, the TS 

domain is responsible for thymidine synthesis and makes use of a THF derivative as a cofactor 

in a reductive methylation reaction31. Mutations in the active site of the DHFR domain arise 

as a result of drug pressure after years of continued use of the same active compounds. 

Resistance to pyrimethamine was first observed in the 1950s, shortly after this compound was 

deployed for widespread use13. Initial strategies to combat this newfound resistance included 

the use of combination therapies, but soon after their implementation they too were found 

to be ineffective. P. falciparum resistance against pyrimethamine arises primarily through a 



29 
 

S108N point mutation in the active site of the parasite DHFR enzyme. This single amino acid 

substitution results in a 25-fold decrease in pyrimethamine activity when compared to the 

wild-type enzyme13. Subsequent mutations further decrease the activity of pyrimethamine, 

with quadruple mutant strains containing the above mutation in combination with C59R, 

N51I, and I164L substitutions causing a loss in activity of three orders of magnitude14. These 

mutations cause the shape of the active site of the parasite DHFR enzyme to change, adopting 

a new configuration for which pyrimethamine has a very low binding affinity40. While these 

changes also confer high levels of resistance to cycloguanil, distinct mutations are observed 

as a result of cycloguanil drug pressure. These include A16V and S108T substitutions, which 

are also able to drastically lower the effectiveness of pyrimethamine14. Cycloguanil and 

pyrimethamine are somewhat rigid molecules, both containing inflexible bi-aryl axes between 

their two rings. Because of this, they have a high energy barrier for rotation around this bond, 

meaning they are effectively locked into a single conformation. As such, they are unable to be 

accommodated in the active site of DHFR enzymes containing mutant residues which sterically 

hinder the binding pocket or alter its shape, preventing optimal alignment and interactions. 

Development of new class II antifolates is focused mainly on increasing the flexibility of active 

compounds, with the aim of mimicking the ability of the natural enzyme substrate, DHF, to fit 

optimally within the enzyme's active site, ensuring effective interactions and inhibition, even 

when mutations alter the shape or steric properties of the binding pocket. 

Table 1.2 summarizes the key single amino acid substitutions discussed above and their roles 

in contributing to drug-resistance. It should be noted, however, that these mutations rarely 

occur in isolation. Instead, resistant P. falciparum strains contain multiple DHFR mutations that 

collectively enhance resistance. For instance, the quadruple mutant S108N/N51I/C59R/I164L 

is strongly associated with pyrimethamine resistance, while the A16V/S108T combination 

confers high levels of resistance to cycloguanil.  

 

 

 

 



30 
 

Table 1.2: Single amino acid mutations in drug-resistant PfDHFR and their impact on inhibitor 

activity 

Mutation: Associated drug-

resistance: 

Inhibitor activity: 

S108N Pyrimethamine  Initial mutation conferring moderate resistance 

to pyrimethamine. 

N51I Pyrimethamine  When combined with S108N, increases resistance 

to pyrimethamine. 

C59R Pyrimethamine  Enhances resistance in combination with S108N 

and N51I mutations. 

I164L Pyrimethamine and 

cycloguanil  

High-level resistance to pyrimethamine and 

cycloguanil, especially when combined with 

other mutations. 

A16V Cycloguanil Confers resistance to cycloguanil but not 

pyrimethamine.  

S108T Cycloguanil  When combined with A16V, significantly 

increases resistance to cycloguanil.  

  

While different amino acid substitutions can hinder the effectiveness of drug molecules, the 

mutational possibilities are limited by the fact that the natural substrate still needs to bind 

and undergo the required reaction13. This is fortunate as it means the enzyme cannot 

indefinitely mutate to avoid inhibition without compromising its essential function, thereby 

preserving the potential for the development of new effective class II antifolates. 

There are numerous crystal structures of the PfDHFR enzyme available in the Protein Data 

Bank (PDB), containing a variety of bound small molecule ligands and bearing several different 

mutations in the enzyme active site, which are responsible for the enzyme’s resistance to 

inhibition. Understanding the structure of the enzyme is key for designing effective class II 

antifolates which can remain active in the presence of resistance conferring mutations.  

 

 



31 
 

1.2.4 Overcoming resistance to class II antifolates in the treatment of 

malaria 

Much insight has been gained into the mechanisms of parasite resistance thanks to the 

resolution of crystal structures of mutant enzymes. These detailed structures have revealed 

how specific mutations alter the active site and affect drug binding, providing valuable 

information for designing more effective inhibitors. As mentioned above, one attractive 

strategy for overcoming resistance is to introduce increased flexibility in drug molecules, 

thereby allowing them to arrange themselves favourably in the active site maintaining the 

same desired interactions but avoiding steric clashes from mutant residues40. In the 

development of new cycloguanil and pyrimethamine analogues, additional flexibility is 

imparted by adding a multi-atom linker between the two ring systems. In the case of 

cycloguanil and pyrimethamine, the two rings are directly joined, greatly limiting the 

rotational possibilities of these molecules due to the rigid bi-aryl axis. With the addition of a 

flexible linker, the rotational freedom between the rings is increased, allowing the molecules 

to adopt conformations that can better fit into the active site of mutant enzymes. This 

enhanced flexibility can improve binding affinity and effectiveness, potentially overcoming 

parasite resistance mechanisms13. 

Many flexible analogues of cycloguanil and pyrimethamine have been developed in recent 

years and have shown improved binding ability against mutant DHFR parasite enzymes. One 

of the first compounds of this nature to be synthesized is WR99210, shown in figure 1.16.  Like 

cycloguanil, this molecule contains a 2,4-diamino-dihydrotriazine core, with a dimethyl 

substituent in the 6-position. It however has a five-atom flexible linker in the 5-position 

connected to the halogenated terminal ring. Although highly potent against drug-resistant 

strains of P. falciparum, this compound has never been developed into a successful drug 

candidate for various pharmacokinetic and toxicity related factors13.  Among these is low oral 

bioavailability. It is suggested that under the acidic conditions of the gastrointestinal tract, the 

triazine core of WR99210 becomes fully protonated, and as such, is unable to diffuse across 

cell membranes for absorption14. Cycloguanil is able to overcome this by being administered 

in pro-drug form. Rearrangement into the active compound only occurs in the liver, after it 

has already entered systemic circulation. PS-15 [Fig 1.16] is a pro-drug form of WR99210, 



32 
 

acting in the same manner as proguanil, and has shown greatly improved oral bioavailability13. 

Pyrimidines are slightly less basic than triazines and exist in an equilibrium between 

protonated and unprotonated species in the gastrointestinal tract. This allows for increased 

levels of absorption and may make them more suitable for the development of new drugs41. 

P218, shown in figure 1.16, is a flexible pyrimethamine analogue reported in 2012. It retains 

the conserved 2,4-diamino-6-ethyl-pyrimidine core of pyrimethamine41. Similarly to 

WR99210, it features a five-atom linker connected to a substituted benzene ring. However, 

instead of halogen substituents, P218 contains a 2-carboxyethyl group attached to its terminal 

ring. This allows for additional charge mediated interactions to asparagine 122 in the active 

site of the target enzyme, analogous to the binding of the natural substrate DHF41. This 

compound displayed excellent activity in vitro, inhibiting quadruple mutant drug-resistant P. 

falciparum strains in the low nanomolar range in whole cell assays. P218 is currently 

undergoing clinical trials for administration in humans. Early results show promising efficacy 

for this compound, including favourable toxicity and pharmacokinetic profiles42.  

 

 

Figure 1.16: Structures of flexible class II antifolates, WR99210 and P218. PS-15, a pro-drug 
of WR99210, is also included  



33 
 

Both WR99210 and P218 were developed using rational drug design principles. Insights into 

the binding mode of the natural substrate and the mechanisms of resistance to existing drugs 

were obtained by studying the crystal structures of mutated enzymes. This information guided 

the design of effective inhibitors of drug-resistant P. falciparum DHFR. In the case of WR99210, 

a homology model was constructed, and it was predicted that increased flexibility between 

the rings of cycloguanil would enhance binding against mutant enzymes13. In the development 

of P218, various protein crystal structures containing synthesized molecules bound in the 

active site of the target enzyme were obtained. This allowed researchers to identify additional 

potential binding interactions for new drug molecules and led to the selection of the 2-

carboxyethyl substituent to form charge mediated bonds to asparagine 122.  

The development of new therapeutic compounds is often guided by Target Candidate Profiles 

(TCP) and Target Product Profiles (TPP), which outline the essential characteristics a drug must 

possess to be a viable therapeutic option. For new antimalarials, these criteria have been well 

defined by the Medicines for Malaria Venture (MMV), which states that novel compounds 

must exhibit potent activity against P. falciparum, favourable pharmacokinetics suitable for 

single dose or short course treatments, and low risk for the development of parasite 

resistance42. Several specific TCP goals have also been established, ranging from molecules 

that block parasite transmission, by targeting gametocytes, to compounds that act on the 

insect vector rather than the parasite itself. Class II antifolates align most strongly with TCP-

142, which includes compounds capable of interrupting the asexual blood-stage of the parasite 

infection, either as monotherapies or in combination with other suitable antimalarials.  

 

1.3 Molecular modelling in drug discovery  

Molecular modelling is a powerful tool which can accelerate many of the most resource 

intensive and time-consuming stages in the drug development process. High-throughput 

screening can be used to identify hit compounds from large libraries of potential candidates, 

and further molecular docking experiments are useful in the early-stages of lead development 

to assess the specific effects of functional group changes. Molecular dynamics (MD) 

simulations offer insights into the dynamic nature of protein-ligand interactions by accounting 



34 
 

for the inherent flexibility of amino acid residues. Unlike static crystal structures or simple 

docking experiments, MD simulations capture the full range of conformational changes and 

interactions that occur in realistic biological systems. Furthermore, various physicochemical 

parameters of drug candidates can be evaluated and their pharmacokinetic characteristics 

predicted using ADMET (absorption, distribution, metabolism, excretion, toxicity) prediction 

tools.  

Until recently, the identification of new hit compounds required screening hundreds or 

thousands of potential candidates in various in vitro assays. This is an immense undertaking, 

with no guarantee of success. Computational high-throughput screening makes it possible for 

this process to be done entirely in silico and allows for a greater range of structural diversity 

to be assessed as evaluated compounds do not actually have to be synthesized or isolated. In 

lead development, chemists often synthesize vast libraries of analogues, making slight 

functional group changes to elucidate the nature of substrate binding to the target enzyme 

and to generate a required minimum pharmacophore through trial and error. This process can 

be expedited using molecular modelling techniques, which allow for specific interactions to 

be predicted and optimized in silico, thereby reducing the need for extensive laboratory 

synthesis and in vitro testing. The prediction of physicochemical parameters also greatly 

assists in the drug development process. Optimization of pharmacokinetics by means of  

various ADMET properties can be done before any synthesis takes place, allowing for a more 

efficient and streamlined process, and increasing the likelihood of developing a successful 

drug candidate. 

 

1.4 Project introduction and aims 

 

1.4.1 Background of the current project  

In 2016, A. Rousseau and colleagues reported a series of flexible cycloguanil analogues which 

showed excellent activity against drug-resistant P. falciparum strains in both single enzyme 

and whole cell assays40. Based on the rationale that a flexible linker would allow for improved 



35 
 

binding against mutant enzymes, a series of compounds was synthesized containing linkers of 

different lengths. The compounds featured a conserved 2,4-diamino-dihydrotriazine core, 

with a phenyl group in the 6-position, and the flexible linker attached to the 5-position of the 

dihydrotriazine ring, given by the general structure shown in figure 1.17. 

 

 

Figure 1.17: General structure of flexible cycloguanil analogues reported in 2016 

 

Various linker lengths were assessed, ranging from two to six atoms. It was found that 

analogues with a four-atom linker performed the best, with the most potent compounds 

inhibiting drug-resistant P. falciparum in low nanomolar concentrations in whole cell assays. 

The effect of various substituents around the two phenyl rings were also investigated, and it 

was shown that while various groups were tolerated around the terminal ring, an 

unsubstituted phenyl ring in the 6-position of the dihydrotriazine core was optimal. The most 

active compound synthesized in this series is shown in figure 1.18. 

 

 

Figure 1.18: Structure of the most active compound of the series synthesized in 2016 

 

Although these compounds displayed excellent anti-plasmodial activity, their synthesis was 

not well optimized, and they contained a chiral centre. Without the possibility of 

enantioselective synthesis in their preparation, racemic mixtures were evaluated for activity. 



36 
 

For these reasons a second-generation of compounds was prepared containing a pyrimidine 

core instead of the dihydrotriazine core44. The pyrimidine core is fully aromatised, and the 

carbon in the 6-position of the ring sp2 hybridized. This eliminated the chiral centre and 

allowed for a more effective synthetic route to be employed. It was hoped that these new 

compounds would display comparable activity to their dihydrotriazine predecessors, this 

however was unfortunately not the case. The most active compounds in this series only 

inhibited drug-resistant P. falciparum in the low micromolar range in whole cell assays, 

representing a significant loss in activity. The general structure of these second-generation 

analogues is shown in figure 1.19. 

 

 

Figure 1.19: General structure of second-generation flexible analogues 

 

It was postulated that the observed loss in activity was due to the inclusion of a new rigid 

bond in the second-generation analogues. As the pyrimidine core is fully aromatized, a high 

energy barrier exists for the rotation of the phenyl ring at the 6-position of the pyrimidine core 

around the newly formed bi-aryl axis. This rigidity effectively locks molecules into a single 

conformation, preventing them from avoiding mutant residues in the active site of the target 

enzyme, drastically lowering their binding ability. While disappointing, these results informed 

the next stages of lead development.  

The investigations conducted in the present project are a continuation of this work. The same 

core functionalities are conserved, specifically the 2,4-diaminopyrimidine core with a four-

atom flexible linker in the 5-position, attached to a substituted terminal ring. Instead of a 

phenyl substituent at the 6-position, however, various linear or cyclic aliphatic hydrophobic 

groups are included. This approach preserves the desired hydrophobic interactions in the 

active site while allowing for free rotation of the new substituent. This ensures that the 



37 
 

molecules can remain highly flexible and are able to adopt favourable conformations for 

binding to mutant enzymes.  

 

1.4.2 Aims 

Based on the earlier findings that a four-atom linker is optimal, and that a pyrimidine core 

with an aliphatic hydrophobic substituent in the 6-position has some advantages over a 

dihydrotriazine core, compounds with the general structure given in figure 1.20 are 

investigated.  

 

 

Figure 1.20: General structure of compounds investigated in this project 

 

Molecular modelling tools are used to evaluate the effectiveness of different substituents on 

the provided general structure. To investigate the dimensions of the hydrophobic pocket 

within the active site, various cyclic and linear aliphatic hydrophobic groups of differing sizes 

are examined as R-substituents. Additionally, numerous analogues with common aromatic 

substituents at the X-position are assessed to potentially identify new interactions that could 

enhance binding affinity to mutant enzymes. ADMET prediction tools are also employed to 

ensure that the compounds exhibit favourable pharmacokinetic properties which are essential 

for the development of successful drug candidates. Furthermore, molecular dynamics (MD) 

simulations are performed on some of the more promising compounds identified in the 

molecular docking experiments to gain insight into the specific nature of the binding 

interactions to the target enzyme. Based on the results of the combined molecular modelling 

experiments, several candidates are selected for synthesis using partially established 

methods, which are optimized for the efficient synthesis of the selected compounds. The 

prepared compounds are then put forward for in vitro antimalarial assessments, including 



38 
 

both single enzyme PfDHFR and whole cell assays against wild-type and drug-resistant strains 

of P. falciparum. Following these initial evaluations, structural changes may be implemented 

in the design of second-generation compounds for improved activity or pharmacokinetic 

properties.  

 

1.4.3 Proposed methodology  

 

1.4.3.1 Molecular modelling 

For the molecular modelling portion, various protocols and algorithms are implemented, all 

of which are contained within the Schrodinger molecular modelling suite of software. These 

methods are discussed in detail in chapter 2, but a brief outline is provided below.   

Relevant protein crystal structures of drug-resistant PfDHFR enzymes are imported from the 

protein data bank (PDB). They are prepared for subsequent modelling by refinement with 

epik, and energy minimization using the OPLS4 force field, as this is the most advanced force 

field available in the version of the modelling software used. Ligand design and library 

enumeration are carried out, generating numerous structures containing different R and X-

substituents. Docking scores for these compounds are then calculated using the Glide 

algorithm. Additional Induced Fit Docking (IFD) procedures are also implemented, in order to 

account for protein flexibility in the docking experiments. Further molecular dynamics 

simulations are conducted on some of the top scoring compounds in order to accurately 

assess the binding mode of the compounds to the target enzyme. In addition to the methods 

mentioned above, QikProp is used for the estimation of various physiochemical properties in 

order to fully assess the suitability of the compounds of interest as potential drug candidates 

for oral administration.  

 



39 
 

1.4.3.2 Synthesis 

The target compounds are formed from the attachment of the flexible linker and terminal ring 

portions to the 5-position of the 2,4-diaminopyrimidine core. A disconnection approach to the 

synthesis is shown in scheme 1.2  

 

 

Scheme 1.2: Retrosynthesis of flexible pyrimethamine analogues 

 

The general synthetic method for the preparation of the final compounds is shown in scheme 

1.3 and described in the following paragraph.  

 



40 
 

 

Scheme 1.3: Synthetic procedure for the production of flexible pyrimethamine analogues. 

 

Reagents and conditions: i) Propargyl bromide, K2CO3, acetone, RT, 24 hrs; ii) H2SO4, MeOH, 

reflux, 24 hrs; iii) KOtBu, MeCN, IPA, 2-MeTHF, RT, 2 hrs; iv) Guanidine HCl, KOtBu, 2-MeTHF, 

MW, 150 W, 150 °C; v) I2, HIO3, H2SO4, CH3COOH, H2O, reflux, 24 hrs; vi) Pd(PPh3)2Cl2, CuI, DIPA, 

DMF, reflux, 24 hrs; vii) Pd/C, H2(g), MeOH, RT, 24 hrs.  

 

Appropriately substituted phenols (1) are alkylated with propargyl bromide under basic 

conditions, yielding the four-atom linker and terminal ring portion of the desired compounds 

(2). The formation of the 2,4-diaminopyrimidine core occurs over several steps. Starting with 

commercially available carboxylic acids containing appropriate R-substitutions (3), an 

esterification reaction is performed with methanol under acidic conditions to obtain the 

corresponding methyl esters (4). The esters then undergo a base mediated nucleophilic 

substitution reaction with acetonitrile, forming the desired beta-keto nitrile products (5). 

Next, these are combined with guanidine hydrochloride under basic conditions in a microwave 

assisted ring closure reaction for the formation of the 2,4-diaminopyrimidine core (6). For the 

subsequent attachment of the linker and terminal ring portions in a cross-coupling reaction, 



41 
 

the 5-position of the pyrimidine core needs to be halogenated. This is achieved by a reaction 

with iodine under acidic conditions, forming the desired 5-iodo pyrimidines (7). The iodinated 

intermediates are then combined with the previously prepared propargyl substituted phenols 

in a palladium catalysed Sonogashira coupling reaction. This connects the 2,4-

diaminopyrimidine core to the linker and terminal ring portions, affording the alkyne 

containing intermediates 8. Finally, the triple bond is reduced in a hydrogenation reaction to 

yield the fully flexible final compounds (9) over a total of seven steps.  

 

1.4.3.3 In vitro antimalaria testing  

Synthesized compounds are evaluated for anti-plasmodial activity and human toxicity in single 

enzyme and whole cell assays. PfDHFR and HsDHFR enzymes are expressed, and inhibition 

constant (Ki) values reported. Compounds are further tested against TM4/8.2 (wild type 

PfDHFR) and V1/S (QM PfDHFR) strains cultured in human erythrocyte cells, and two 

mammalian derived cell lines for cytotoxicity assessments. The assay details, and discussion 

of the results are presented in chapter 3. All assays are conducted by the BIOTEC research 

group in Thailand. The general method followed for the in vitro experiments is provided in 

chapter 6. 



42 
 

Chapter 2 – Molecular Modelling 

 

2.1 Introduction 

This chapter deals with the extensive molecular modelling studies that were conducted to 

identify compounds for synthesis and testing against drug-resistant strains of Plasmodium 

falciparum. The Schrödinger molecular modelling suite was used, and all simulations were run 

in the Maestro user interface. Schrödinger release version 2023-245 was used for generating 

the molecular dynamics simulations, and version 2024-246 was used for everything else. The 

specific parameters used for each experiment are included in screenshots of the Maestro user 

interface contained in appendix A. 

Protein crystal structures of mutant Plasmodium falciparum DHFR-TS enzymes were imported 

directly from the RCSB Protein Data Bank (PDB). All structures were obtained by X-ray 

crystallography and have various ligands bound in the dihydrofolate reductase active site, 

along with the cofactor NADPH. Protein structures were prepared for subsequent analysis 

using the built-in protein preparation workflow47,48. This is a vital first step in any molecular 

modelling work as protein structures obtained from the PDB are often incomplete. The protein 

preparation workflow fills in missing portions of the protein structure, including absent 

hydrogen atoms and highly disordered, solvent exposed loops which are not able to be 

resolved from crystallography. The hydrogen bonding network within the protein is optimized 

at a given pH, and the final protein structure is minimized using an energy force field. The 

OPLS4 force field49 was used for this portion of the project.   

The virtual screening workflow was used for method validation and to gain insight into the 

binding mode of the compounds assessed. This tool combines ligand preparation50, receptor 

grid generation, and Glide docking51–55 into a single user interface. Ligand preparation 

generates the possible ionization and tautomerization states of the compounds at a specified 

pH and uses Epik to calculate associated energy penalties for each state. Receptor grids are 

required for molecular docking and are generated by including the protein region surrounding 

the bound ligand, along with any specified docking constraints, into a three-dimensional 

structure file.  



43 
 

For Glide docking, three different algorithms are available: High Throughput Virtual Screening 

(HTVS), Standard Precision (SP), and Extra Precision (XP). In each method, numerous ligand 

conformations are generated and sampled for positive interactions against the defined 

receptor grid. The sampled poses are scored by calculating various energy terms related to 

protein-ligand contacts. These include hydrogen bonding, lipophilic interactions, electrostatic 

interactions, and van der Waals dispersion forces. These parameters are combined into the 

overall GlideScore, which is well suited for distinguishing between compounds with high 

binding ability and those with weak affinity for the active site.  

The three Glide algorithms vary in the extent to which they sample ligand conformational 

possibilities, representing a trade-off between speed and accuracy. GlideHTVS is the fastest 

and performs only limited conformational sampling, while GlideSP employs an exhaustive 

sampling approach. GlideXP is the most resource-intensive algorithm, using an anchor-and-

grow method that iteratively refines good initial poses to identify the most favourable ligand 

binding conformations. GlideXP also introduces additional energy parameters to the 

GlideScore, including a hydrophobic enclosure term that rewards ligands for their potential to 

displace water molecules from hydrophobic sites.  

The ligand state penalties calculated during ligand preparation are incorporated into the 

GlideScore to produce the Docking Score. This ensures that scores for unfavourable ligand 

states, which might otherwise rank highly, are adjusted accordingly. GlideScore and Docking 

Score are both reported in units of kcal/mol, however, they are not directly comparable with 

experimentally obtained binding energies.  

Further Induced Fit Docking (IFD)56–58 experiments were conducted to assess the flexible 

nature of the protein active site, and to determine any potential new binding interactions. 

Traditional docking methods are limited by the fa