BSc Immunology and Pharmacology (Bioinformatics): The Structural Analysis of Protein Kinases from Trypanosoma and Leishmania Parasites

Computational structural analysis of three Leishmania mexicana protein kinases using AlphaFold 3.0, NCBI BLAST, PyMOL, and JupyterLab revealed conserved functional motifs and residue-level interaction patterns that may inform future drug discovery efforts against leishmaniasis, a neglected tropical disease affecting millions worldwide. The project was completed as a BSc (Hons) dissertation in Immunology and Pharmacology at the Strathclyde Institute of Pharmacy and Biomedical Sciences, University of Strathclyde, under the supervision of Dr Martin Wiese.

Leishmaniasis is a devastating parasitic disease caused by Leishmania protozoan parasites, responsible for over 30,000 deaths annually and placing an ongoing burden on healthcare systems across endemic regions. Despite this, treatment options remain limited, and resistance to existing drugs such as amphotericin B, miltefosine, and pentavalent antimonials continues to grow. Identifying novel therapeutic targets is therefore a pressing priority. This dissertation investigated three mitogen-activated protein kinases (MAPKs) from Leishmania mexicana — LmxM.25.1990, LmxM.17.0490, and LmxM.14.1300 — using a computational bioinformatics pipeline to characterise their three-dimensional structures, predict protein-protein interaction sites, and evaluate their potential as drug targets.

The Burden of Leishmaniasis and the Case for Kinase Targeting

Leishmaniasis presents in three main clinical forms. Cutaneous leishmaniasis, the most common, causes skin sores and scarring at the site of infection. Visceral leishmaniasis is the most severe, targeting the spleen, liver, and bone marrow and proving fatal if left untreated. Mucocutaneous leishmaniasis, the least common, destroys mucous membranes of the nose, throat, and mouth. Approximately 1.7 billion people live in areas at risk of contracting the disease, and control remains challenging due to environmental factors, drug resistance, and the absence of effective vaccines.

Current treatments are hampered by significant side effects and increasing resistance. Amphotericin B resistance arises through mutations in sterol biosynthesis pathways that reduce drug binding, while miltefosine resistance is driven by alterations in transporter genes that impair drug uptake. These mechanisms highlight the urgent need for alternative therapeutic strategies targeting essential parasite biology.

Protein kinases represent a compelling avenue for intervention. The three kinases selected for this study — LmxM.25.1990, LmxM.17.0490, and LmxM.14.1300 — are serine/threonine kinases within the MAPK cascade, critical regulators of stress response, differentiation, and survival in Leishmania mexicana. Their selection was based on characterised transferase and catalytic activities, their role in ATP binding, and their positioning within the parasite kinome.

Computational Methodology

The analytical pipeline combined several complementary bioinformatics tools to build a comprehensive structural picture of the three target kinases.

Kinase Identification and Sequence Analysis: Candidate kinases were retrieved from GenBank and TriTrypDB through a stepwise screening process. An initial pool of 335 genes was filtered for serine/threonine specificity, then reduced to 91 by protein kinase domain presence, and further refined to 34 by removing structurally similar kinases. NCBI BLAST was then used to align amino acid sequences from L. mexicana, identifying conserved motifs and interaction points critical for kinase functionality.

AlphaFold Structure Prediction: High-resolution three-dimensional models of all three kinases were generated using AlphaFold 3.0, which applies deep learning to predict protein conformations from amino acid sequences. Structural confidence was assessed using pLDDT (predicted local Distance Difference Test) scores at the residue level, with segments scoring below 50 excluded from functional interpretation. Global interaction confidence was evaluated using iPTM (interface predicted Template Modelling) and pTM scores to assess the reliability of predicted protein-protein interfaces.

Structural Visualisation and Refinement: Initial examination of predicted models was performed in FirstGlance in Jmol, which enabled qualitative assessment of secondary structure composition, ATP-binding pockets, and catalytic residues. Structures were then imported into PyMOL for detailed visualisation and quantitative measurement of residue distances, salt bridge interactions, and hydrogen bond networks. Side-chain conformations were optimised through energy minimisation to reduce steric clashes.

Quantitative Analysis in JupyterLab: Refined structures were exported in Protein Data Bank format and imported into a JupyterLab environment, where custom Python scripts using NumPy, pandas, SciPy, and Matplotlib calculated interatomic distances among functionally relevant residues. A two-tailed t-test was used to assess the statistical significance of differences in structural confidence scores across the three kinase models.

Implications for Drug Discovery

LmxM.14.1300 emerges as a particularly strong drug target candidate, given its robust structural integrity, higher pLDDT confidence, and narrower range of salt-bridge distances, which may enable more precise inhibitor design. LmxM.25.1990 and LmxM.17.0490, while displaying greater structural flexibility, also present opportunities for allosteric targeting if their extended or transient contacts prove critical for kinase cascade function.

The findings situate kinase inhibition as a promising strategy for addressing drug resistance in leishmaniasis treatment. By targeting essential regulators of the MAPK cascade, it may be possible to disrupt the parasite’s stress response and survival mechanisms in ways that existing drugs targeting membrane sterols or metabolic pathways do not.

Limitations and Future Directions

As with all computational studies, important limitations apply. In silico methods cannot fully replicate dynamic protein conformations or the complexity of the parasite’s intracellular environment. Flexible loop regions in LmxM.25.1990 and LmxM.17.0490 may assume ordered structures only under specific stress conditions not captured by static AlphaFold models. Experimental validation through X-ray crystallography, cryo-electron microscopy, or site-directed mutagenesis will be essential to confirm the predicted salt bridge interactions and their functional relevance.

Future research could also extend comparative analysis to Trypanosoma brucei and T. cruzi to identify structural features conserved across kinetoplastid species, potentially identifying broad-spectrum targets. Molecular dynamics simulations would complement static AlphaFold predictions by refining estimates of loop mobility and allosteric site accessibility, while integration of transcriptomic and proteomic data during parasite differentiation could reveal additional windows of vulnerability for therapeutic intervention.

Conclusion

This dissertation demonstrates how a multi-tool computational pipeline can generate meaningful structural and functional insights into Leishmania mexicana protein kinases with relevance to drug discovery. By combining sequence alignment, AlphaFold structure prediction, residue-level visualisation, and quantitative distance analysis, the study identified LmxM.14.1300 as a structurally stable and well-defined target, while also highlighting the potential allosteric significance of the more flexible loop regions in LmxM.25.1990 and LmxM.17.0490.

As drug resistance continues to undermine existing leishmaniasis treatments, computational approaches like this provide a cost-effective and scalable foundation for identifying candidate targets and prioritising candidates for experimental validation. Translating these structural insights into drug development will ultimately require in vitro and in vivo assays, but this work establishes a clear and evidenced rationale for pursuing Leishmania mexicana kinases as a therapeutic avenue.

The dissertation was awarded a mark of 76, placing it within the A band at distinction level, reflecting the quality and rigour of the computational analysis undertaken. This result validated the depth of the structural investigation and the strength of the bioinformatics pipeline developed throughout the project.

Acknowledgement

I would like to thank my supervisor Dr Martin Wiese for his guidance, expertise, and unwavering support throughout this project. I am also grateful to the lecturers and colleagues at the Strathclyde Institute of Pharmacy and Biomedical Sciences for their helpful discussions on bioinformatics and structural analysis.