Biology subjects

Butler, G.

Publications and source records attributed to Butler, G..

At least 19 recordsLinked to original sources

bioRxiv2026

Motor signals modulate cortical but not subcortical processing of self-initiated sounds

When we produce sounds ourselves, the brain modulates the auditory neural response through an efference copy mechanism, allowing us to distinguish between self-initiated and externally generated auditory inputs. However, the precise level of the auditory pathway at which this attenuation occurs remains unclear. While evidence from animal models suggests that early auditory processing of self-generated sounds may be modulated by corticofugal signaling, localized cortical modulation would preserve the high-fidelity subcortical sound encoding while allowing flexible, context-dependent processing at higher levels. To probe potential motor influences in the early auditory system, we collected scalp-recorded frequency following responses (FFRs) from 33 normal-hearing adults during active (self-initiated) and passive (externally presented) listening conditions using a 170 ms speech stimulus. Data were collected with a vertical montage that emphasizes subcortical generators of the FFR. We observed no significant differences in the FFR between active and passive conditions in spectral power, response amplitude, pitch tracking, onset latency, or phase consistency. In contrast, cortical event-related potentials showed motor-induced suppression (MIS): reduced early peak amplitudes in the active condition after correcting for motor signals, increased phase consistency prior to auditory feedback, and more precise phase consistency at sound offset. In addition to indicating FFRs can be collected during a wider range of behavioral tasks without substantial motor contamination, our observation of the canonical MIS in cortical signals but not in FFRs suggests that MIS of self-initiated sounds primarily affects later stages of auditory processing rather than the early encoding reflected in the FFR.

neuroscience↗

bioRxiv2026

Indirect genomic effects shape cancer risk across species

Tumour prevalence varies dramatically throughout the animal kingdom despite broadly conserved cellular and developmental processes, raising the question of how evolution has shaped susceptibility 1,2. Here, we link macroevolutionary variation in tumour prevalence to gene-level selection by integrating comparative genomics data from 109 species of birds and mammals using a Bayesian phylogenetic framework to estimate pangenome-wide rates of genetic evolution across >150 million years of evolutionary change. We identify 3,206 genes in which natural selection is associated with shifts in tumour prevalence, with more than 80% of which are linked to reduced prevalence, suggesting pervasive selection for cancer suppression. Using causal phylogenetic inference, we show that genes associated with reduced tumour prevalence act predominantly through indirect effects on body size, revealing growth as a key mediator of cancer risk across species. In contrast, genes associated with increased tumour prevalence exert direct effects independent of body size. Finally, at the species-level, we demonstrate that exceptionally low rates of benign tumours do not necessarily coincide with reduced malignancy, revealing that benign and malignant tumour processes are evolutionarily decoupled. Together, these results reveal how natural selection has fine-tuned the link between genotype, phenotype, and cancer risk across species.

evolutionary biology↗

bioRxiv2026

The Vertebrate Genomes Project Phase I: A global reference genome resource

The Vertebrate Genomes Project (VGP) aims to produce complete and near-error-free reference genomes for all [~]70,000 extant vertebrate species1. Organized in four phases, it progressively targets all vertebrate orders, families, genera, and eventually all species. Here we present the completion of VGP Phase I, delivering reference genomes for [~]95% of vertebrate orders, along with additional lineages within those orders, totaling 816 species and 1.6 trillion base pairs of main haplotype sequence. These genomes were assembled and annotated over an 8-year period (2018-2026) of rapid advances in genome sequencing, assembly, and annotation methods2-4, alongside the growth of associated consortium initiatives and international collaborations5-9. They represent some of the highest-quality vertebrate genomes currently available, and most have become the primary reference for their respective species in public databases. Comparative analyses across a subset of 579 species when we reached a threshold of 85% of orders allowed us to reconstruct the genome of the last common ancestor of all vertebrates 500 million years ago, identify diverse modes of sex chromosome evolution, reveal clade-specific three-dimensional genome architecture, discover methylated epigenetic landscapes across vertebrates, and provide a framework for studying gene and pseudogene evolution, immune loci, cancer-associated genes, and other trait-associated loci. Approximately a quarter of this subset are listed as Vulnerable to Critically Endangered by the IUCN Red List of Threatened Species, and have enabled more advanced genomic investigations of extinction risk. VGP Phase I delivers a reference backbone for vertebrate genomics, enabling discoveries that would otherwise remain out of reach across evolution, conservation, and medicine. Talking pointsO_LIThe flagship paper of VGP Phase I C_LIO_LIThe highest quality collection of genomes within the eukaryotic domain of life C_LIO_LIEvolution of genome sequencing technology quality throughout VGP Phase I C_LIO_LIA driver project that has been a model for multiple large-scale, high-quality reference genome projects C_LIO_LIReleases all currently unpublished genomes in Phase I from scientific study embargoes C_LIO_LIMultiple biological discoveries across the vertebrate tree of life C_LI

genomics↗

bioRxiv2025

TooTranslator: Zero-Shot Classification of Specific Substrates for Transport Proteins by Language Embedding Alignment of Proteins and Chemicals

Transmembrane transport proteins mediate selective movement of ions and metabolites across membranes. Experimental characterization of their substrate specificity is limited. For novel class discovery with limited data, zero-shot learning aims to assign labels to test samples whose label has not been seen previously during training of the model. We introduce TooTranslator, a regression-based model that aligns embeddings from pre-trained protein (ProtBERT), chemical (ChemBERTa), and text (SciB-ERT) language models into a shared latent space, enabling substrate prediction by minimizing distances between protein and substrate embeddings. TooTranslator tackles zero-shot learning for the task of predicting the specific substrate of transmembrane transport proteins. Models using four loss functions are evaluated on protein test sets with seen and unseen substrates. We find no statistically significant difference in performance of the four loss functions. For tests with seen substrates, the models compare with the state-of-the art in performance. For tests with unseen (but known) substrate a top-k approach is needed as only 4.3% of test cases predict the correct label as the nearest label. For top-10 that rises to 17%; for top-50 rises to 54%, and top-100 reaches 80%. TooTranslator demonstrates the potential of multimodal embedding alignment for open-world protein function inference.

bioinformatics↗

bioRxiv2025

Mitotic bypass and endocycling promote cancer cell survival after genotoxic chemotherapy

Genotoxic chemotherapies are central components of the treatment regimen for most cancers but are rarely curative. Drug-tolerant persister cells (DTPs) evade cell death during these treatments by accessing transient adaptive states and can contribute to cancer progression after treatment. Here, we demonstrate that cancer cells can survive genotoxic chemotherapy-induced stress by accessing a DTP state wherein stress-induced bypass of mitosis precipitates continued endocycling which promotes survival by allowing cells to evade mitotic catastrophe and cell death. Mechanistic studies indicate that persistent DNA damage signaling in endocycling persister cells triggers sustained p53-independent CDK1 inhibition by WEE1 and Myt1. Continued survival in endocycling persisters is dependent on activation of this G2 checkpoint, and disrupting WEE1 or Myt1 activity using clinical-stage small molecule inhibitors is sufficient to drive CDK1 reactivation, forcing mitotic entry, catastrophe, and cell death. Our results define endocycling DTPs as targetable mediators of cancer cell persistence after genotoxic therapy.

cell biology↗

bioRxiv2025

Intracellular acidification by bacteria-derived valeric acid is a mechanism of trans-kingdom ecology against Candida parapsilosis colonization

In hematopoietic cell transplant patients, intestinal Candida parapsilosis expansion and translocation cause life-threatening candidemia, yet how commensal intestinal bacteria prevent Candida expansion remains incompletely defined. Here, we trained a machine learning model on supernatant metabolomic profiles of commensal bacteria to identify bacteria-derived inhibitors of fungal growth, with valeric and butyric acid as top hits. Experimental validation confirmed in silico predictions in three systems. First, in patient fecal samples, valeric and butyric acid levels inversely correlated with C. parapsilosis growth. Second, in vitro, valeric acid potently inhibited C. parapsilosis growth by causing intracellular acidification. Third, administration of glycerol valerate, and free or microencapsulated valeric acid blunted C. parapsilosis growth at murine intestinal sites where valeric acid could be detected. Thus, machine learning could identify a mechanistic driver of trans-kingdom ecology limiting C. parapsilosis intestinal expansion and may inform strategies to reduce patient risk of developing candidiasis during high-risk periods.

microbiology↗

bioRxiv2025

Centromeres in budding yeasts are conserved in chromosomal location but not in structure.

The budding yeast Saccharomyces cerevisiae has point centromeres, which are much smaller and simpler than centromeres of most other eukaryotes and have a defined DNA sequence. Other yeast taxa have different and highly diverse centromere structures, but a clear picture of how yeast centromeres have evolved is lacking. Here, we investigated nine yeast species in two taxonomic orders that are close outgroups to S. cerevisiae. We find that they have a wide diversity of centromere structures, indicating that multiple transitions of structure have occurred within the last 200 Myr. Some species have centromeres with defined sequence motifs (17 - 200 bp), others consist of Inverted Repeats (IRs), and others have Ty5-like retroelement clusters. Strikingly, the chromosomal locations of centromeres have largely been conserved across taxonomic orders, even as their structures have changed, which suggests that structure replacement occurs in situ. In some Barnettozyma species we find that a single genome can contain chromosomes with different centromere structures - some with IRs and some without - which suggests that a structural transition is underway in this genus. We identified only one example of a centromere moving by a long distance: a new centromere formed recently at the MAT locus of Barnettozyma californica, 250 kb from the previous centromere on that chromosome. Author summaryCentromeres are an evolutionary paradox. Their molecular function is highly conserved, but their structures vary tremendously among eukaryotes. The "point" centromeres of the yeast Saccharomyces cerevisiae are among the most unusual: they are tiny (< 200 bp) and non-repetitive, and unlike other centromeres they contain a single copy of a well-defined sequence motif. However, we have little knowledge about where these point centromeres came from, or more generally about how changes of centromere structure occur during evolution. Here, we characterized centromere structure and location in nine species of budding yeasts, spanning an evolutionary depth of approximately 200 million years. We find that the chromosomal locations of centromeres are extraordinarily well conserved, whereas their structures vary greatly. We show that sequence-defined centromeres are older and more widely distributed than previously realised. We identify some species in which the centromeres of different chromosomes have different structures, which suggests that they are in transition from one structure to another. The centromere variation observed makes it difficult to infer the ancestral structure.

evolutionary biology↗

bioRxiv2025

Small pangenome of Candida parapsilosis reflects overall low intraspecific diversity

Candida parapsilosis is an opportunistic yeast pathogen that can cause life-threatening infections in immunocompromised humans. Whole genome sequencing (WGS) studies of the species have demonstrated remarkably low diversity, with strains typically differing by about 1.5 single nucleotide polymorphisms (SNPs) per 10 kb. However, SNP calling alone does not capture the full extent of genetic variation. Here, we define the pangenome of 372 C. parapsilosis isolates to determine variation in gene content. The pangenome consists of 5,859 genes, of which 48 are not found in the genome of the reference strain. This includes 5,791 core genes (present in [≥] 99.5% of isolates). Four genes, including the allantoin permease gene DAL4, were present in all isolates but were truncated in some strains. The truncated DAL4 was classified as a pseudogene in the reference strain CDC317. CRISPR-Cas9 gene editing showed that removing the early stop codon (producing the full-length Dal4 protein) is associated with improved use of allantoin as a sole nitrogen source. We find that the accessory genome of C. parapsilosis consists of 68 homologous clusters. This includes 38 previously annotated genes, 27 novel paralogs of previously annotated genes and 3 uncharacterised ORFs. Approximately one-third of the accessory genome (24/68 genes) is associated with gene fusions between tandem genes in the major facilitator superfamily (MFS). Additionally, we identified two highly divergent C. parapsilosis strains and find that, despite their increased phylogenetic distance ([~]30 SNPs per 10 kb), both strains have similar gene content to the other 372. ImportanceCandida parapsilosis is a human fungal pathogen, listed in the high priority group by the World Health Organisation. It is an increasing cause of hospital-acquired and drug-resistant infection. Here, we studied the genetic diversity of 372 C. parapsilosis isolates, the largest genomic surveillance of this species to date. We show that there is relatively little genetic variation. However, we identified two more distantly-related isolates from Germany, suggesting that even more sampling may yield more diversity. We find that the pangenome (the cumulative gene content of all isolates) is surprisingly small, compared to other fungal species. Many of the non-core genes are involved in transport. We also find that variations in gene content are associated with nitrogen metabolism, which may contribute to the virulence characteristics of this species.

genomics↗

bioRxiv2024

TooT-PLM-P2S: Incorporating secondary structure information into protein language models

In bioinformatics, modeling the protein space to better predict function and structure has benefitted from Protein Language Models (PLMs). Their basis is the proteins amino acid sequence and self-supervised learning. Ankh is a prime example of such a PLM. While there has been some recent work on integrating structure with a PLM to enhance predictive performance, to date there has been no work on integrating secondary structure rather than three-dimensional structure. Here we present TooT-PLM-P2S that begins with the Ankh model pre-trained on 45 million proteins using self-supervised learning. TooT-PLM-P2S builds upon the Ankh model by initially using its pre-trained encoder and decoder. It then undergoes an additional training phase with approximately 10,000 proteins and their corresponding secondary structures. This retraining process modifies the encoder and decoder, resulting in the creation of TooT-PLM-P2S. We then assess the impact of integrating secondary structure information into the Ankh model by comparing Ankh and TooT-PLM-P2S on eight downstream tasks including fluorescence and solubility prediction, sub-cellular localization, and membrane protein classification. For both Ankh and TooT-PLM-P2S the downstream tasks required task-specific training. Few of the results showed statistically significant differences. Ankh outperformed on three of the eight tasks, TooT-PLM-P2S did not outperform on any task for the primary metric. TooT-PLM-P2S did outperform for the precision metric for the task of discriminating membrane proteins from non-membrane proteins. This study requires future work with expanded datasets and refined integration methods.

bioinformatics↗

bioRxiv2024

Oestrogen influences B cell class-switching in individuals with an XX sex chromosome complement

Sex differences in humoral immunity are well-documented, though the mechanisms underpinning these differences remain ill-defined. Here, we demonstrate that post-pubertal cisgender females have higher levels of class-switched B cells compared to age-matched cisgender males. However, whilst sex chromosome-encoded genes characterise most of the differences in total B cell transcriptomes between cisgender-females and -males, sex differences in class-switched B cells are only observed post-pubertally. Accordingly, B cells express high levels of oestrogen receptor 2 (ESR2) and genes known to regulate B cell class-switching are enriched for ESR2-binding sites. Using a gender-diverse cohort of young people, we show that in transgender males (XX chromosomal background), blockade of natal oestrogen reduced the frequency of class-switched B cells, whilst gender-affirming oestradiol treatment in transgender females (XY chromosomal background), did not increase the frequency of class-switched B cells. These data demonstrate that sex hormones and chromosomes work in tandem to impact immune responses, with oestrogen only supporting B cell class-switching on an XX chromosomal background. eTOC summarySex hormones and chromosomes work in tandem to impact immune responses, with oestrogen influencing B cell class-switching exclusively on an XX chromosomal background.

immunology↗

bioRxiv2024

Punctuational evolution is pervasive in distal site metastatic colonization

The evolution of metastasis represents a lethal stage of cancer progression. Yet, the evolutionary kinetics of metastatic disease remain unresolved. Here, using single cell CRISPR-Cas9 lineage tracing data, we show that in metastatic disease, gradual molecular evolution is punctuated by episodes of rapid evolutionary change associated with lineage divergence. By measuring punctuational effects across the metastatic cascade, we show that punctuational effects contribute more to the molecular diversity at distal site metastases compared to the paired primary tumor, suggesting qualitatively different modes of evolution may drive primary and metastatic tumor progression. This is the first empirical evidence for distinct patterns of molecular evolution at early and late stages of metastasis and demonstrates the complex interplay of cell intrinsic and extrinsic factors that shape lethal cancer.

cancer biology↗

bioRxiv2023

Analysis of clinical Candida parapsilosis isolates reveals copy number variation in key fluconazole resistance genes

We used whole-genome sequencing to analyse a collection of 35 fluconazole resistant and 7 susceptible Candida parapsilosis isolates together with coverage analysis and GWAS techniques to identify new mechanisms of fluconazole resistance. Phylogenetic analysis shows that although the collection is diverse, two probable outbreak groups were identified. We identified copy number variation of two genes, ERG11 and CDR1B, in resistant isolates. Two strains have a CNV at the ERG11 locus; the entire ORF is amplified in one, and only the promoter region is amplified in the other. We show the annotated telomeric gene CDR1B is actually an artefactual in silico fusion of two highly similar neighbouring CDR genes due to an assembly error in the C. parapsilosis CDC317 reference genome. We report highly variable copy numbers of the CDR1B region across the collection. Several strains have increased expansion of the two genes into a tandem array of new chimeric genes. Other strains have experienced a deletion between the two genes creating a single gene with a reciprocal chimerism. We find translocations, duplications, and gene conversion across the CDR gene family in the C. parapsilosis species complex, showing that it is a highly dynamic family.

genomics↗

bioRxiv2023

A general prediction model for substrates of transport proteins

Transport proteins play a crucial role in cellular metabolism and are central to many aspects of molecular biology and medicine. Determining the function of transport proteins experimentally is challenging, as they become unstable when isolated from cell membranes. Machine learning-based predictions could provide an efficient alternative. However, existing methods are limited to predicting a small number of specific substrates or broad transporter classes. These limitations stem partly from using small datasets for model training and a choice of input features that lack sufficient information about the prediction problem. Here, we present SPOT, the first general machine learning model that can successfully predict specific substrates for arbitrary transport proteins, achieving an accuracy above 92% on independent and diverse test data covering widely different transporters and a broad range of metabolites. SPOT uses Transformer Networks to represent transporters and substrates numerically. To overcome the problem of missing negative data for training, it augments a large dataset of known transporter-substrate pairs with carefully sampled random molecules as non-substrates. SPOT not only predicts specific transporter-substrate pairs, but also outperforms previously published models designed to predict broad substrate classes for individual transport proteins. We provide a web server and Python function that allows users to explore the substrate scope of arbitrary transporters.

bioinformatics↗

bioRxiv2023

Exploiting protein language models for the precise classification of ion channels and ion transporters

This study presents TooT-PLM-ionCT, a composite framework consisting of three distinct systems, each with different architectures and trained on unique datasets. Each system within TooT-PLM-ionCT is dedicated to a specific task: segregating ion channels (ICs) and ion transporters (ITs) from other membrane proteins and differentiating ICs from ITs. These systems exploit the capabilities of six diverse Protein Language Models (PLMs) - ProtBERT, ProtBERT-BFD, ESM-1b, ESM-2 (650M parameters), and ESM-2 (15B parameters). As these proteins play a pivotal role in the regulation of ion movement across cellular membranes, they are integral to numerous biological processes and overall cellular vitality. To circumvent the costly and time-consuming nature of wet lab experiments, we harness the predictive prowess of PLMs, drawing parallels with techniques in natural language processing. Our strategy engages six classifiers, embracing both conventional methodologies and a deep learning model, for each of our defined tasks. Furthermore, we delve into critical factors influencing our tasks, including the implications of dataset balancing, the effect of frozen versus fine-tuned PLM representations, and the potential variance between half and full precision floating-point computations. Our empirical results showcase superior performance in distinguishing ITs from other membrane proteins and differentiating ICs from ITs, while the task of discriminating ICs from other membrane proteins exhibits results commensurate with the current state-of-the-art. Author summaryIn our research, we have designed TooT-PLM-ionCT, a composite framework composed of three unique systems, each tailored to a specific protein classification task and trained on different datasets. This framework is our tool for categorizing integral membrane proteins, specifically ion channels and ion transporters. These proteins are essential to the health of cells, as they manage ion movement across cell membranes. To bypass the high costs and long timelines of conventional lab experiments, we have turned to advanced computation methods akin to how computers process human language. Our three-pronged approach harnesses six top-tier Protein Language Models and a range of classifiers to discern between these key proteins. In doing so, we also evaluated the effects of various conditions, like dataset balance, representation methods, and levels of computation precision, on the accuracy of our classification tasks. The outcomes show our framework effectively identifies ion transporters, sets them apart from ion channels, and distinguishes ion channels on par with existing top-notch techniques. The performance, however, can vary based on the task, suggesting that customizing the approach for each task could be beneficial. In the future, we plan to expand the depth and breadth of our protein study by incorporating additional knowledge sources, utilizing more refined representation methods, and testing our framework on larger and diverse protein datasets. This progress sets us on a path to better understand proteins and their roles in cellular health.

bioinformatics↗

bioRxiv2022

CRISPR-Cas9 editing induces Loss of Heterozygosity in the pathogenic yeast Candida parapsilosis

Genetic manipulation is often used to study gene function. However, unplanned genome changes (including Single Nucleotide Polymorphisms (SNPs), aneuploidy and Loss of Heterozygosity (LOH)) can affect the phenotypic traits of the engineered strains. Here, we show that CRISPR-Cas9 editing can induce LOH in the diploid human pathogenic yeast Candida parapsilosis. We sequenced the genomes of ten isolates that were edited with CRISPR-Cas9 and found that the designed changes were present in nine. However, we also observed LOH in all isolates, and aneuploidy in two isolates. LOH occurred most commonly downstream of the Cas9 cut site and extended to the telomere in three isolates. In two isolates we observed LOH on chromosomes that were not targeted by CRISPR-Cas9. Two different isolates exhibited cysteine and methionine auxotrophy caused by LOH at a heterozygous site in MET10, approximately 11 and 157 kb downstream from the Cas9 target site, respectively. C. parapsilosis isolates have relatively low levels of heterozygosity. However, our results show that mutation complementation to confirm observed phenotypes is important even when using CRISPR-Cas9, which is now the gold standard of genetic engineering. IMPORTANCECRISPR-Cas9 has greatly streamlined gene editing, and is now the gold standard and first choice for genetic engineering. However, we show that in diploid species extra care should be taken in confirming the cause of any phenotypic changes observed. We show that the Cas9-induced double strand break is often associated with loss of heterozygosity in the asexual diploid human fungal pathogen Candida parapsilosis. This can result in deleterious heterozygous variants (e.g. stop gain in one allele) becoming homozygous resulting in unplanned phenotypic changes. Our results stress the importance of mutation complementation even when using CRISPR-Cas9.

microbiology↗

bioRxiv2022

Predicting the specific substrate for transmembrane transport proteins using BERT language model

Transmembrane transport proteins play a vital role in cells metabolism by the selective passage of substrates through the cell membrane. Metabolic network reconstruction requires transport reactions that describe the specific substrate transported as well as the metabolic reactions of enzyme catalysis. In this paper, we apply BERT (Bidirectional Encoder Representations from Transformers) language model for protein sequences to predict one of 12 specific substrates. Our UniProt-ICAT-100 dataset is automatically constructed from UniProt using the ChEBI and GO ontologies to identify 4,112 proteins transporting 12 inorganic anion or cation substrates. We classified this dataset using three different models including Logistic Regression with an MCC of 0.81 and accuracy of 97.5%; Feed-forward Neural Networks classifier with an MCC of 0.88 and accuracy of 98.5%. Our third model utilizes a Fine-tuned BERT language model to predict the specific substrate with an MCC of 0.95 and accuracy of 99.3% on an independent test set.

bioinformatics↗

bioRxiv2022

TooT-SC: Predicting Eleven Substrate Classes of Transmembrane Transport Proteins

BackgroundTransporters form a significant proportion of the proteome and play an important role in mediating the movement of compounds across membranes. Transport proteins are difficult to characterize experimentally, so there is a need for computational tools that predict the substrates transported in order to annotate the large number of genomes being sequenced. Recently we developed a dataset of eleven substrate classes from Swiss-Prot using the ChEBI ontology as the basis for the definition of the classes. ResultsWe extend our earlier work TranCEP, which predicted seven substrate classes, to the new dataset with eleven substrate classes. Like TranCEP, TooT-SC combines pairwise amino acid composition (PAAC) of the protein, with evolutionary information captured in a multiple sequence alignment (MSA) using TM-Coffee, and restriction to important positions of the alignment using TCS. Our experimental results show that TooT-SC significantly outperforms the state-of-the-art predictors, including our earlier work, with an overall MCC of 0.82 and the MCC for the eleven classes ranging from 0.66 to 1.00. ConclusionTooT-SC is a useful tool with high performance covering a broad range of substrate classes. The results quantify the contribution made by each type of information used during the prediction process. We believe the methodology is applicable more generally for protein sequence analysis.

bioinformatics↗

bioRxiv2021

Resistance to miltefosine results from amplification of the RTA3 floppase or inactivation of flippases in Candida parapsilosis

Flippases and floppases are two classes of proteins that have opposing functions in the maintenance of lipid asymmetry of the plasma membrane. Flippases translocate lipids from the exoplasmic leaflet to the cytosolic leaflet, and floppases act in the opposite direction. Phosphatidylcholine (PC) is a major component of the eukaryotic plasma membrane and is asymmetrically distributed, being more abundant in the exoplasmic leaflet. Here we show that gene amplification of a putative PC floppase or double disruption of two PC flippases in the pathogenic yeast Candida parapsilosis results in resistance to miltefosine, an alkylphosphocholine drug that affects PC metabolism that has recently been granted orphan drug designation approval by the US FDA for treatment of invasive candidiasis. We analysed the genomes of 170 C. parapsilosis isolates and found that 107 of them have copy number variations (CNVs) at the RTA3 gene. RTA3 encodes a putative PC floppase whose deletion is known to increase the inward translocation of PC in Candida albicans. RTA3 copy number ranges from 2 to >40 across the C. parapsilosis isolates. Interestingly, 16 distinct CNVs with unique endpoints were identified, and phylogenetic analysis shows that almost all of them have originated only once. We found that increased copy number of RTA3 correlates with miltefosine resistance. Additionally, we conducted an adaptive laboratory evolution experiment in which two C. parapsilosis isolates were cultured in increasing concentrations of miltefosine over 26 days. Two genes, CPAR2_303950 and CPAR2_102700, gained homozygous protein-disrupting mutations in the evolved strains and code for putative PC flippases homologous to S. cerevisiae DNF1. Our results indicate that alteration of lipid asymmetry across the plasma membrane is a key mechanism of miltefosine resistance. We also find that C. parapsilosis is likely to gain resistance to miltefosine rapidly, because many isolates carry loss-of-function alleles in one of the flippase genes. Author summaryMiltefosine was developed as an anticancer drug but is commonly used to treat infections with the protozoan parasites Leishmania and Trypanosoma cruzi. More recently, it has been used to treat fungal infections, and in 2021 it was designated as an orphan drug by the US Food and Drug Administration for treatment of invasive candidiasis. Miltefosine is a derivative of phosphatidylcholine (PC), a major constituent of the cell membrane. PC and other phospholipids are asymmetrically distributed across the cell membrane. The mechanism of action of miltefosine is unknown. Here, we show that either increasing the activity of a putative floppase, which controls outward "flop" movement of phospholipids, or decreasing the activity of flippases, which control inward "flip" movement, results in increased resistance of the fungal pathogen Candida parapsilosis to miltefosine. This result suggests that miltefosine acts by controlling the localisation of PC or other phospholipids in the membrane. Importantly, we find that many C. parapsilosis isolates carry mutations in one flippase gene, which renders them partially resistant to miltefosine, and prone to easily acquiring increased resistance.

genomics↗