Abstract The human fungal pathogen Cryptococcus neoformans poses significant health risks, particularly to immunocompromised individuals, such as those with HIV/AIDS. In this study, we investigate the role of an uncharacterized protein, Taf1, in regulating ATP levels and virulence in C. neoformans. Our previous proteomic analyses confirmed the expression of Taf1, encoded by the gene CNAG_04232. We found that the deletion of the TAF1 gene resulted in the upregulation of 204 genes and the downregulation of 908 genes. Gene Ontology analysis indicated that these regulated genes are associated with metabolic and cellular processes, as well as ATP-dependent activities. Notably, the TAF1-deficient mutant exhibited impaired growth at elevated temperatures (39°C). Furthermore, in a murine model of infection, mice inoculated with the taf1Δ mutant demonstrated significantly improved survival compared to those infected with the wild-type strain, suggesting a critical role for Taf1 in virulence. Additionally, KEGG pathway analysis of RNA-Seq and metabolomics data revealed significant alterations in fatty acid biosynthesis and degradation pathways following TAF1 deletion. Collectively, these findings underscore the essential role of Taf1 in modulating cellular energy and its implications for the virulence of C. neoformans, thereby paving the way for potential therapeutic strategies targeting this pathogen. Supplementary Information The online version contains supplementary material available at 10.1186/s12866-025-04207-w. Keywords: Cryptococcus neoformans, Taf1, Virulence, ATP Introduction The opportunistic fungal pathogen Cryptococcus neoformans poses a significant threat to individuals with compromised immune systems, particularly those living with HIV/AIDS [[36]1, [37]2]. Annually, approximately 180,000 cases of cryptococcal meningoencephalitis are reported, with this fungus responsible for 15% of all AIDS-related fatalities [[38]3, [39]4]. The World Health Organization’s recent recognition of C. neoformans as a critically important fungal pathogen underscores the urgent need for further investigation into the serious diseases it causes, characterized by a lack of effective treatment options and the absence of a vaccine [[40]5, [41]6]. This substantial disease burden highlights the pressing necessity to elucidate the mechanisms that facilitate fungal proliferation within vertebrate hosts, which is crucial for the discovery of new therapeutic agents and vaccine targets [[42]7]. A key factor influencing the virulence of C. neoformans is adenosine triphosphate (ATP), which serves as the primary energy currency for cells [[43]8, [44]9]. ATP is crucial not only for cellular metabolism but also for various processes that enable C. neoformans to thrive and cause disease within the host [[45]10, [46]11]. Research has shown that disruption of ATP biosynthesis results in a complete loss of virulence in C. neoformans. For instance, Blundell et al. demonstrated that the loss of adenylosuccinate synthetase (AdSS) leads to adenine auxotrophy and a consequent reduction in virulence in murine models [[47]9]. This study highlighted distinct structural differences between fungal AdSS and its human counterpart, positioning AdSS as a promising target for antifungal drug development [[48]9]. The relationship between ATP levels and virulence in C. neoformans is particularly significant, as the C. neoformans ability to adapt to diverse environmental stresses encountered within the host is central to its pathogenicity. Elevated ATP levels are associated with enhanced growth and survival under hostile conditions, enabling the pathogen to effectively counteract immune challenges [[49]8, [50]12]. Conversely, disruptions in ATP production can lead to metabolic dysregulation, thereby impairing the C. neoformans ability to sustain its virulence [[51]9, [52]11]. The regulation of ATP levels is intricately linked to various metabolic pathways, particularly those involved in the synthesis and degradation of fatty acids [[53]13]. Fatty acids are essential components of cellular membranes, playing a critical role in maintaining membrane integrity, facilitating signaling processes, and serving as energy reserves. Moreover, they significantly impact the modulation of virulence traits in fungal pathogens, including C. neoformans, by influencing energy production and cellular responses to environmental stressors [[54]14–[55]16]. In this study, we characterized the role of the uncharacterized protein Taf1 in regulating ATP levels and its implications for the virulence of C. neoformans. Our findings indicate that the TAF1-deficient mutant (taf1Δ) exhibited growth deficiencies at 39°C, probably resulting in diminished virulence in a murine model of cryptococcosis. Importantly, the TAF1-deficient mutant demonstrated reduced ATP levels compared to wild-type (WT) cells. Insights from transcriptional profiling using RNA sequencing (RNA-Seq) and metabolomics further support the hypothesis that Taf1 regulates ATP levels, likely through pathways involved in fatty acid biosynthesis and degradation. Overall, these data highlighted Taf1 as a probable contributor to ATP levels and the virulence of C. neoformans, establishing its significance in the context of fungal pathogenesis and energy metabolism. Materials and methods Strain and media This investigation utilized two strains: the Cryptococcus neoformans var. grubii KN99α WT strain and the mutant strain taf1Δ which carries the gene deletion of CNAG_04232 sourced from the Fungal Genetics Stock Center (FGSC). To explore the differences in mRNA and metabolomic profiles between the WT and the mutant strain taf1Δ, cells were cultured in YPD medium (2% dextrose, 1% yeast extract, 2% peptone) for around 16 h at 30°C. Following this incubation period, the cultures were harvested and washed twice with distilled water. The resulting cell pellets were subsequently frozen in liquid nitrogen and stored at −80°C for RNA-seq and metabolomics studies. Three biological replicates were performed. Melanin production To investigate melanin production, L-DOPA agar plates containing 0.7 mM L-3,4-dihydroxyphenylalanine (L-DOPA), 0.1% L-asparagine, 0.1% glucose, 0.025% MgSO4 · 7H[2]O, 0.3% KH[2]PO[4] and 2% agar (pH 5.6) were utilized. C. neoformans cells were spotted on L-DOPA agar plates and incubated for 2 days at temperatures of 30°C, 37°C, or 39°C before capturing images of the results. Serial spot dilution assays Fungal cultures grown overnight were washed twice with PBS, followed by adjusting the cell concentration to 2 × 10^5 cells per ml. After this adjustment, 10-fold serial dilutions were created, and 5 µl from each dilution—spanning from 10^5 to 10^0 cells—was spotted on YPD agar plates (2% dextrose, 1% yeast extract, 2% peptone, 2% agar) or L-DOPA agar plates. The agar plates were subsequently incubated at temperatures of 30°C, 37°C, or 39°C for a duration of 2 days before capturing images of the results. RNA-seq To extract RNA utilizing the Omega Plant RNA Kit (Omega Bio-Tek, Catalog No. R6827), approximately 100 mg of the fungal cell sample was ground in a mortar with liquid nitrogen, and the resulting powder was transferred into a 1.5 mL tube. Then, 500 µl of Buffer RCL, including β-mercaptoethanol, was added, and the mixture was vortexed thoroughly. The sample was incubated in a water bath at 55°C for 3 min, followed by centrifugation at 14,000 g for 5 min to separate the phases. The supernatant was carefully collected and transferred to a gDNA Filter Column in a collection tube, and centrifugation was performed again at 14,000 g for 2 min. Next, the supernatant was mixed with an equal volume of Buffer RCB by pipetting up and down 5 to 10 times. This mixture was divided in half; one portion was transferred to the Hibind RNAMini column and centrifuged at 10,000 g for 1 min, discarding the flow-through, then the column was returned to the collection tube. This process was repeated with the second half of the solution. Subsequently, 400 µl of RWC was added to the column and centrifuged at 10,000 g for 1 min, discarding the flow-through. Then, 500 µl of RNA Wash Buffer II was added, and centrifugation was repeated at 10,000 g for another minute, with the waste discarded once more. After washing, the column was centrifuged at 10,000 g for 2 min to remove residual buffers. The Hibind RNAMini column was transferred to a new 1.5 mL RNase-Free tube, and 50 µl of DEPC-treated water was added to the center of the membrane. It was allowed to stand at room temperature for 2 min before centrifuging at 10,000 g for 1 min. Finally, the eluate was returned into the Hibind RNAMini column for another round of incubation at room temperature for 2 min, and centrifugation at 10,000 g for 1 min was performed to obtain the purified RNA solution. To enrich the extracted mRNA, mRNA Capture Beads were employed. Following purification with these beads, the mRNA underwent fragmentation at elevated temperatures. The resulting fragmented mRNA served as a template for synthesizing the first strand of complementary DNA (cDNA) within a reverse transcription enzyme mixture. During the synthesis of the second strand of cDNA, processes such as end repair and A-tailing were also completed. Subsequently, adapters were ligated to the cDNA, and Hieff NGS^® DNA Selection Beads facilitated the purification of target fragments. The PCR amplification of the library followed, culminating in sequencing with the Illumina Novaseq X Plus. The sequencing output contained raw reads, which may have included adapters and low-quality bases that could have compromised downstream assembly and analysis. To ensure high-quality clean reads, fastp (version 0.18.0) [[56]17] was applied for further filtering; a filtering process was implemented that involved discarding any reads with adapter sequences, excluding those that contained over 10% unknown nucleotides (N), and removing low-quality reads characterized by more than 50% of their bases having a Q-value of 20 or lower. For mapping reads to the ribosomal RNA (rRNA) database, the short reads alignment tool Bowtie2 (version 2.2.8) [[57]18] was utilized, enabling the removal of rRNA-mapped reads. The resultant clean reads were then employed for assembly and gene abundance calculations. An index of the reference genome was constructed, and paired-end clean reads were mapped against this reference using HISAT2 (version 2.1.0) with default parameters [[58]19]. The mapped reads for each sample were assembled employing StringTie v1.3.1 through a reference-based approach [[59]20]. For each transcription region, the expression abundance and variation were quantified by calculating the FPKM (fragments per kilobase of transcript per million mapped reads) value using RSEM software. Additionally, the mapped reads were utilized to compute a TPM (Transcripts Per Kilobase of exon model per Million mapped reads) value for further assessment of expression abundance and variability, also through RSEM. Differential expression analysis of RNAs was conducted using DESeq2 software to compare two distinct groups, with edgeR applied for sample comparison [[60]21]. Genes or transcripts meeting the criteria of a false discovery rate (FDR) below 0.05 and an absolute fold change ≥ 2 were classified as differentially expressed. Gene Ontology (GO) enrichment analysis identified significantly enriched GO terms among differentially expressed genes (DEGs) relative to the background of the genome. This mapping process involved associating all DEGs with GO terms from the GO database ([61]http://www.geneontology.org/), followed by a hypergeometric test to determine significant enrichment [[62]22]. Furthermore, genes typically engaged in interactions that facilitated specific biological functions. Pathway-based analysis enhanced the understanding of these functions. The KEGG database served as a primary public resource for pathway-related information [[63]23]. Pathway enrichment analysis discovered significantly enriched metabolic and signaling pathways within the DEGs when compared to the entire genome background. Metabolomics studies The cell samples were ground using a pestle and mortar after being frozen in liquid nitrogen. A 100 mg aliquot was mixed with 1 mL of a cold solvent system consisting of methanol, acetonitrile, and water in a 2:2:1 volumetric ratio, followed by sonication at low temperature for 30 min, performed twice. After centrifugation at 14,000 g for 20 min at 4°C, the supernatant was isolated and subsequently dried using a vacuum centrifuge. A total volume of 100 µL of acetonitrile/water (1:1, v/v) was used to re-dissolve the samples for LC-MS analysis. Utilizing a UHPLC system (1290 Infinity LC, Agilent Technologies) paired with a quadrupole time-of-flight mass spectrometer (AB Sciex TripleTOF 6600, Shanghai), we conducted our analysis. The HILIC separation was executed on an ACQUITY UPLC BEH Amide column (2.1 mm × 100 mm, 1.7 μm particle size, Waters, Ireland). The mobile phase featured two components: component A was a blend of 25 mM ammonium acetate and an equal concentration of ammonium hydroxide in water, while component B comprised acetonitrile. Our gradient program began with 95% B and sustained this level for 0.5 min before gradually declining to 65% over a span of 6.5 min. Following this, the ratio was further decreased to 40% for 1 min, maintaining stability for an additional minute before returning to 95% B in only 0.1 min, followed by a re-equilibration phase that lasted 3 min. We optimized the electrospray ionization (ESI) source parameters, adjusting Ion Source Gas 1 and Gas 2 to 60, the curtain gas to 30, and setting the source temperature at 600°C, with the IonSpray Voltage Floating (ISVF) configured to ± 5500 V. For the MS-only acquisition, we scanned the m/z range of 60 to 1000 Da, allowing for a TOF MS accumulation time of 0.20 s for each spectrum. In the auto MS/MS acquisition, the m/z range was modified to 25-1000 Da, with a product ion scan accumulation time set to 0.05 s per spectrum. We retained features that displayed over 50% nonzero measurements in any group, and compound identification was reliant on precise m/z values (within 10 ppm) as well as MS/MS spectra compared to an internal standard database. To increase the coverage of metabolites, we performed analyses in both positive and negative ion modes separately. The inclusion of quality control (QC) samples was essential for validating the reliability of our metabolomics study [[64]24]. We carried out principal component analysis (PCA) using the R language package gmodels (v2.18.1) [[65]25], where the denser clustering of QC samples indicated stronger reliability. Metabolites were prioritized based on their variable importance in projection (VIP) scores obtained from the (O)PLS model, establishing a threshold of VIP ≥ 1 for effective differentiation between the two groups. In addition, we executed a T-test for univariate analysis, successfully identifying differential metabolites at a significance threshold of FDR < 0.05. These metabolites were subsequently aligned with the KEGG database for annotation and enrichment analysis [[66]23]. Assessment of intracellular ATP levels and mitochondrial membrane potential To measure intracellular ATP levels and mitochondrial membrane potential, C. neoformans cells were initially cultured overnight in YPD medium at 30°C with shaking at 130 rpm. After this growth phase, the cells were washed twice with PBS. Intracellular ATP levels were measured using the BacTiter-Glo Microbial Cell Viability Assay kit (Promega, USA), and the signal was detected with a TECAN Infinite E Plex microplate reader. For evaluating mitochondrial membrane potential, the cells were treated with 50 nM Mito-Tracker Red CMXRos, which accumulates in mitochondria in a manner dependent on the mitochondrial membrane potential. This property allows it to serve as a probe for detecting changes in mitochondrial membrane potential, which can indicate apoptosis. The cells were incubated at 30°C for 1 h and subsequently washed twice with PBS. The fluorescence intensity of the sample, containing 5 × 10^7 cells, was analyzed using a TECAN Infinite E Plex microplate reader. All experiments were conducted with three biological replicates, and statistical significance was determined using a Student’s t-test. Virulence assays For conducting virulence assays, we obtained 6-week-old female BALB/c mice from the Experimental Animal Center at Nantong University, located in Nantong, China. The WT strain and the taf1Δ mutant strain were cultivated overnight in YPD medium at 30°C. Following cultivation, the strains were washed with PBS and then resuspended in PBS at a concentration of 1 × 10^6 cells/ml. The inoculation process involved intranasal instillation of 50 µl of the cell suspension, resulting in an inoculum of 5 × 10^4 cells for each mouse. Each strain was tested using groups of 10 mice. After inoculation, the health status of the mice was assessed daily. CO[2] asphyxiation was utilized to euthanize mice that had reached the humane endpoint. This research adhered to the ethical guidelines regarding animal use set forth by Nantong University, ensuring that all mice were treated in a manner that minimized suffering. Survival analysis was conducted using Kaplan-Meier (K-M) curves and the log-rank (Mantel-Cox) test. Results TAF1-lacking mutant exhibits reduced virulence In our previous proteomic analyses of C. neoformans, we confirmed the expression of Taf1 [[67]11], a hypothetical protein encoded by the gene CNAG_04232 in the C. neoformans genome. Here we showed that the mutant taf1Δ displayed significantly impaired growth at 39°C (Fig. [68]1A). However, when cultured in media supplemented with L-DOPA (L-3,4-dihydroxyphenylalanine) under the same temperature conditions, the taf1Δ mutant exhibited normal melanin production (Fig. [69]1A). To explore the role of Taf1 in virulence-related phenotypes, we hypothesized that the absence of Taf1 would result in decreased pathogenicity in a murine model. To test this hypothesis, we intranasally inoculated mice with cells from both the WT strain and the taf1Δ mutant. By day 27 post-infection, all mice infected with the WT strain succumbed, while a significant proportion of mice infected with the taf1Δ mutant survived well beyond this timeframe (Fig. [70]1B). These findings suggest that Taf1 is essential for the virulence of C. neoformans. Fig. 1. [71]Fig. 1 [72]Open in a new tab Taf1 is essential for thermotolerance and virulence. A To assess the thermotolerance, spot assays were performed on YPD medium and L-DOPA plates, examining the response of each strain at 30°C, 37°C, or 39°C. Specifically, L-DOPA plates were utilized to evaluate melanin production in the various strains. B A total of ten female BALB/c mice were subjected to intranasal inoculation with 10^5 cells of either the WT strain (KN99) or the taf1Δ mutant strain. To assess survival outcomes, the log-rank Mantel-Cox test was employed for statistical evaluation between the two groups of mice. The analysis revealed significant differences in survival rates between those infected with the WT strain and those infected with the taf1Δ mutant strain (*, P < 0.05) Mutant lacking TAF1 has decreased ATP levels and downregulated ATP-dependent activities To further elucidate the molecular mechanisms associated with TAF1 deletion, we performed a comprehensive analysis of the transcriptomic profiles of C. neoformans. RNA-Seq revealed that the absence of the TAF1 gene led to the upregulation of 204 genes and downregulation of 908 genes (Fig. [73]2A and Supplementary Data 1). Subsequent GO analysis indicated that a significant proportion of these differentially expressed genes are involved in various biological processes, particularly in metabolic pathways, where most exhibited downregulation. Notably, a detailed classification of gene functions revealed a marked decrease in ATP-dependent activities, suggesting that Taf1 plays a critical role in maintaining energy homeostasis. Additional analysis of cellular components corroborated these findings, showing substantial downregulation of genes associated with cellular anatomical structures, which may have implications for cellular integrity and functionality (Fig. [74]3). Experimental assessments further confirmed a significant reduction in cellular ATP levels following the knockout of TAF1, as depicted in Fig. [75]4A. Interestingly, when we assessed the mitochondrial membrane potential of C. neoformans post-TAF1 deletion, we observed an unexpected increase in membrane potential in the knockout cells (Fig. [76]4B). This observation may indicate the activation of a compensatory mechanism in response to the energy deficit. Overall, these findings enhance our understanding of the intricate molecular mechanisms by which Taf1 regulates cellular energy dynamics and metabolic processes in C. neoformans. Fig. 2. [77]Fig. 2 [78]Open in a new tab The impact of Taf1 loss on gene and metabolite expression. A Bar charts show the number of differentially expressed genes: up-regulated (red) and down-regulated (blue) with a false discovery rate (FDR) of less than 0.05 and a log 2 fold change greater than 1 or less than − 1. B Bar charts display the number of differentially expressed metabolites: up-regulated (red) and down-regulated (blue), both with FDR < 0.05. The efficiency of metabolite identification is shown for negative ion mode (NEG) and positive ion mode (POS) Fig. 3. [79]Fig. 3 [80]Open in a new tab The Gene Ontology (GO) categories for differentially expressed genes were identified through RNA-seq analysis of the taf1Δ mutant, utilizing three biological replicates Fig. 4. Fig. 4 [81]Open in a new tab Significant changes in intracellular ATP levels and mitochondrial membrane potential were observed following the knockout of the TAF1 gene. Statistical analysis was performed using a Student’s t-test with three biological replicates. Data are presented as mean ± standard deviation. Asterisks indicate statistical significance: *, P < 0.05; ***, P < 0.001. Figures were created using Origin software Taf1 may impact fatty acid biosynthesis and degradation The deletion of TAF1 appears to affect fatty acid biosynthesis and degradation, as indicated by the KEGG pathway analysis of RNA-Seq data (Fig. [82]5). Our untargeted metabolomics approach further clarifies the metabolic changes associated with TAF1 deletion, which are summarized in Fig. [83]2B. In the negative ionization mode, we identified 19 distinct metabolite changes, with 8 metabolites showing increased levels and 11 displaying decreased levels (Fig. [84]2B and Supplementary Data 2). In the positive ionization mode, we detected 20 individual alterations in metabolites, with 14 increasing and 6 decreasing (Fig. [85]2B and Supplementary Data 2). Further KEGG pathway analysis corroborated these findings, revealing significant changes in pathways related to fatty acid biosynthesis and degradation, as well as alterations in oxidative phosphorylation and purine metabolism (Fig. [86]6). Overall, TAF1 may play a role in influencing fatty acid biosynthesis and degradation. Fig. 5. [87]Fig. 5 [88]Open in a new tab The KEGG enrichment analysis reveals the top 30 pathways enriched for differentially expressed genes, based on three biological replicates Fig. 6. [89]Fig. 6 [90]Open in a new tab The KEGG enrichment analysis of metabolites with differential expression, highlighting the top 30 enriched pathways identified through through negative ion mass spectrometry (NEG) and positive ion mass spectrometry (POS), based on three biological replicates Discussion In this study, we elucidate the role of Taf1, an uncharacterized protein, in regulating ATP levels and virulence in C. neoformans. Our findings demonstrate that the taf1Δ mutant exhibits impaired growth at 39°C and reduced virulence in a murine model of cryptococcosis, underscoring the essential role of Taf1 in the infection process of C. neoformans. The ability to grow at temperatures between 37°C and 39°C is crucial for microbial infection in human hosts [[91]26]. Our results indicated that Taf1 was necessary for robust growth at 39°C, suggesting its involvement in the thermal adaptation required for virulence. Notably, at 30–37°C, the taf1Δ mutant exhibited growth comparable to that of the WT strain. This discrepancy raises an important question: why does the taf1Δ mutant display a growth defect only at 39°C and not at 30–37°C? We propose that this difference may be attributed to the distinct metabolic demands and stress responses elicited at higher temperatures. At 39°C, the taf1Δ mutant appears to struggle with thermal stress, which could compromise its ability to maintain essential cellular functions, including metabolic flexibility. In contrast, at 30–37°C, the mutant may be able to compensate for the loss of Taf1 through alternative pathways or mechanisms that are insufficient under the more stressful conditions at elevated temperatures. A deeper exploration of these temperature-dependent responses would enhance our interpretation of the results and provide insights into the mechanisms underlying the growth phenotype observed in the taf1Δ mutant. Transcriptomic analyses revealed that deletion of the TAF1 gene lead to the downregulation of genes associated with ATP-dependent activities. Our untargeted metabolomics further indicated alterations in oxidative phosphorylation and purine metabolism. Mitochondria serve as the primary sites for ATP synthesis, and this process is intricately linked to both oxidative phosphorylation and purine metabolism. While oxidative phosphorylation is the principal mechanism for ATP generation, purine metabolism is essential for ATP synthesis and various cellular functions. We hypothesize that Taf1 may regulate intracellular ATP levels by influencing these metabolic pathways, with the observed decline in ATP levels contributing to reduced ATP-dependent activities in the taf1Δ mutant. Additionally, our metabolomics analysis reveals significant alterations in fatty acid metabolism following TAF1 deletion. This finding connects Taf1 role in energy regulation to its impact on virulence, as fatty acids are critical not only for energy production but also as key components of cellular membranes and signaling molecules [[92]13, [93]16, [94]27]. The interplay between ATP dynamics and fatty acid metabolism is particularly noteworthy; the reduction in ATP levels in the taf1Δ mutant correlates with impaired metabolic flexibility, which is vital for the pathogenic potential of C. neoformans. Future investigations should assess whether the taf1Δ mutant exhibits altered sensitivity to cell wall and membrane stressors. This metabolic flexibility is crucial for C. neoformans to adapt to the diverse and often hostile conditions encountered in the host environment, thereby enhancing its survival and virulence. The impaired growth at elevated temperatures observed in the taf1Δ mutant suggests a compromised ability to withstand physiological stresses during infection. Overall, these findings reinforce the notion that Taf1 is integral to the metabolic adaptability of C. neoformans, facilitating its navigation through complex environments and evasion of immune defenses. Furthermore, the differential expression profiles observed in the taf1Δ mutant provide a compelling foundation for future investigations into the specific metabolic pathways regulated by Taf1. Understanding these pathways could unveil new therapeutic targets aimed at disrupting energy regulation and virulence in C. neoformans. For instance, elucidating the precise mechanisms by which Taf1 influences fatty acid biosynthesis and degradation could lead to innovative strategies for enhancing antifungal efficacy and mitigating the impact of this pathogen. Conclusion Our study provides evidence highlighting the role of Taf1 in regulating ATP levels, infection and fatty acid metabolism in C. neoformans. The complex interplay between Taf1, metabolic regulation, and pathogenicity underscores the necessity for further research to elucidate the molecular mechanisms underlying these processes. Gaining insight into these interactions is essential for developing novel therapeutic strategies aimed at effectively combating fungal infections and improving clinical outcomes for affected individuals. Supplementary Information [95]Supplementary Material 1.^ (22.8KB, xlsx) [96]Supplementary Material 2.^ (762.8KB, xlsx) Acknowledgements