Human activity, not environmental factors, drives Scedosporium and Lomentospora distribution in Taiwan

Abstract

Scedosporium and Lomentospora species are emerging fungal pathogens capable of causing severe infections in both immunocompetent and immunocompromised individuals. Previous environmental surveys have suggested potential associations between these fungi and various soil chemical parameters, though the relative influence of human activity versus environmental factors has not been systematically evaluated. Here, we conducted a comprehensive survey of 406 soil samples from 132 locations across Taiwan, analyzing fungal abundance alongside soil physicochemical parameters and the Human Footprint Index (HFI). We recovered 236 fungal isolates comprising 10 species, with S. boydii (32.2%), S. apiospermum (30.9%), and S. dehoogii (14.4%) being the most prevalent. The highest fungal burdens were observed in urban environments (up to 1293 CFU/g), particularly in public spaces and healthcare facilities. Statistical analysis revealed a significant positive correlation between fungal abundance and HFI (r = 0.143, P = .005), while soil chemical parameters including nitrogen, carbon, pH, electrical conductivity, and various base cations showed no significant associations despite their wide ranges. These findings indicate that anthropogenic disturbance of environments, rather than soil chemistry, is the primary driver of Scedosporium and Lomentospora distribution in Taiwan. This understanding holds important implications for predicting infection risks and developing targeted public health strategies, particularly in rapidly urbanizing regions. Future studies incorporating more specific indicators of human impact may further elucidate the mechanisms underlying these distribution patterns.

Introduction

Scedosporium and Lomentospora species are clinically significant fungal pathogens capable of causing a wide spectrum of human infections. These range from localized infections like mycetoma following traumatic inoculation in immunocompetent hosts to severe disseminated infections in immunocompromised patients, particularly those with hematologic malignancies or transplant recipients.^1,2 In cystic fibrosis patients, these fungi rank second among filamentous fungi colonizing the airways,³ and they are particularly concerning due to their ability to cause allergic bronchopulmonary disease and potential progression to invasive infection following lung transplantation.⁴ The mortality rates are notably high, reaching up to 80% in cases of disseminated infection and 75% in neutropenic patients with invasive disease.² Due to their intrinsic resistance to most available antifungal drugs and increasing clinical significance,^5–7Scedosporium/Lomentospora species have been recently included in the World Health Organization’s first-ever Fungal Priority Pathogens List (FPPL),⁸ highlighting their global importance as emerging pathogens requiring urgent attention and research focus.

Environmental surveys conducted worldwide have provided important insights into the ecological preferences of these fungi. Studies in Austria and the Netherlands found no Scedosporium species in natural habitats such as forests and sand dunes, but reported high recovery rates from industrial areas, parks and playgrounds, and agricultural lands.⁹ In France, the highest densities were found in human-impacted areas, with S. dehoogii predominantly recovered from industrial areas and S. aurantiacum from agricultural regions.¹⁰ Australian surveys revealed S. aurantiacum (54.6% of isolates) and L. prolificans (43%) as predominant species in urban environments, with significantly higher fungal burdens in areas within 3 km of city centers compared to rural sites.¹⁰ In Thailand, environmental surveys in human-impacted public parks found high abundance and diversity of Scedosporium species, with S. apiospermum as the predominant species, accounting for 86.6% of isolates in Bangkok parks. Broader surveys across regions with dense human activity and tourism further confirmed S. apiospermum’s dominance, revealing significant genetic diversity among isolates, including potential novel species.^11,12 Mexican surveys across 25 states found these fungi primarily in areas of high human activity, including urban gardens (49%), industrial parks (24%), and sports parks (14%).¹³ Notably, these studies consistently demonstrated a strong association between human activities and the presence of Scedosporium/Lomentospora species, while also revealing geographical variations in species distribution.

Environmental factors, particularly nutrient availability, strongly influence the distribution and abundance of Scedosporium/Lomentospora species.^9,14 These fungi thrive in nitrogen-rich environments with elevated nitrate levels, explaining their prevalence in fertilized agricultural lands and areas of intensive animal activity.⁹ Other significant factors include phosphorus availability, organic matter (OM) content, and temperature, with warmer conditions (around 25°C) enhancing the effects of both nutrient availability and environmental pollutants.^14,15 While these fungi can adapt to various pH (4.9–8.5) and salinity conditions, they show optimal growth in moderately alkaline soils (pH 6.1–7.6).^9,11,14 Notably, hydrocarbon contamination, especially from diesel fuel, correlates with higher fungal populations.^15–17 However, most evidence comes from field observations, with only one experimental study validating the synergistic effects of temperature, nitrate, and diesel contamination.¹⁵ These findings suggest that anthropogenic factors, rather than soil chemistry alone, may primarily drive Scedosporium/Lomentospora distribution in natural ecosystems.

In Taiwan, our previous study revealed high recovery rates of Scedosporium species from soil samples across urban and hospital areas in Taiwan.¹⁸ To build upon these findings and better understand the environmental factors governing the distribution of these medically relevant fungi, we conducted an expanded environmental survey incorporating three key components lacking in our previous study: (1) more comprehensive geographical coverage, including eastern Taiwan, (2) systematic analysis of soil physicochemical parameters, and (3) quantitative assessment of human impact using the Human Footprint Index (HFI). This multifaceted approach aims to delineate the relative importance of anthropogenic influence versus edaphic factors in determining the abundance and distribution of Scedosporium and Lomentospora species across Taiwan’s diverse landscapes.

Materials and methods

Sample sites

A total of 132 distinct locations across Taiwan were investigated in this study, comprising the 79 previously sampled locations in Huang et al. (2023)¹⁸ and an additional 53 new sampling sites, particularly strengthening the coverage in eastern Taiwan (Fig. 1, Supplementary Table 1). Land use type classification for all sampling sites was based on the Land Use Investigation (LUI) Map from The National Land Surveying and Mapping Center (NLSC) (https://www.nlsc.gov.tw/), Taiwan. Across the 132 locations, a total of 406 soil samples were collected, resulting from 1 to 17 replicates per location to account for spatial variability. At each replicate site, 1–4 sampling units were randomly selected within a square meter. For each sampling unit, soil was collected from five points with 10–20 cm intervals from the center using a metal spatula sterilized with 70% ethanol between samples to avoid cross-contamination. Soil was collected at a depth of 7–10 cm beneath the soil surface, a range chosen based on Kaltseis et al. (2009)⁹, which showed comparable Scedosporium colony-forming unit (CFU) counts across depths. This depth minimizes surface variability while ensuring consistent sampling. Samples were pooled in a sterile plastic bag and gently mixed, then transferred to a 50 ml sterile Falcon tube. All soil samples were stored at 4°C until processing.

Isolation and quantification of fungal burden in soils

For each soil sampling unit, 10 g of soil was air-dried at room temperature for 1–2 days, placed in a 100 ml sterile Erlenmeyer flask, and homogenized with 90 ml sterilized water containing 1–2 drops of Tween 20. The mixture was thoroughly vortexed and left to stand for 15 min. A total of 250 μl of suspension was inoculated onto five plates of Dichloran Glycerol (DG18) Agar Base (Himedia, India) supplemented with chloramphenicol at a concentration of 20 µl/ml. Plates were incubated at 37°C for 5–7 days and examined at intervals for the emergence of characteristic colonies of Scedosporium/Lomentospora. All Scedosporium/Lomentospora colonies on five plates of each sampling unit were counted, from which the fungal burdens were determined based on the mean value of CFU/g of soil dry weight.

Molecular identification of fungal isolates

For molecular identification, we amplified the internal transcribed spacer (ITS) region using a two-step barcoding approach.^19–21 The first PCR used a 25 μl reaction mixture containing DNA template, primer sets (NS1B1ngs: CCTNGTTGATYCTGCCAGT and LR5: TCCTGAGGGAAACTTCG), and PCR master mix. Amplification consisted of initial denaturation (95°C, 2 min), followed by 25 cycles of denaturation (95°C, 30 s), annealing (64–66°C, 30 s), and extension (72°C, 1 min), with final extension at 72°C for 7 min. The second PCR used the first-step product with barcoded primers. Products were purified using Ampure Xp Beads (0.4× ratio) and sequenced on an Oxford Nanopore Technologies GridION using the SQK-LSK110 kit.

We reconstructed the phylogeny of the isolated species using the ITS sequences. Sequences of type specimens of Scedosporium and Lomentospora species were retrieved from NCBI GenBank. Sequences of Petriellopsis africana (CBS 311.72), Petriella setifera (CBS 385.87), and Petriellopsis sordida (CBS 144612) were selected as the outgroup. Sequences were aligned using MAFFT.²² Phylogenetically informative regions were selected using ClipKIT²³ with the gappy strategy. Phylogenetic trees were reconstructed using IQ-TREE2²⁴ with recommended partition parameters inferred using ModelTest-NG.²⁵ Trees were visualized using Interactive Tree of Life version 4.²⁶

Soil chemical properties analysis

All soil chemical analyses were performed on air-dried soil samples that passed through a 2-mm sieve. Each analysis was performed in duplicate, and mean values were used for subsequent statistical analyses.

Total nitrogen (N) in soil samples was determined using the semimacro Kjeldahl method with a KB-8 digestion block and VAP 300 steam distillation system (Gerhardt GmbH, Germany). Soil OM content was assessed via wet oxidation following the Walkley–Black procedure. Water content was determined by oven-drying soil samples at 105°C until constant weight was achieved. Soil pH and electrical conductivity (EC) were determined using a pH/EC/TDS/Temperature Portable Meter HI9814 (HANNA Instruments, Woonsocket, RI, USA) in a 1:2.5 soil:water suspension.

Available phosphorus was measured using the Bray-1 method with spectrophotometric determination. Exchangeable cations (Na⁺, K⁺, Mg²⁺, and Ca²⁺) were assessed by atomic absorption spectrophotometry (AAS) using a Z 5300 instrument following the manufacturer’s recommendations (Hitachi—Science & Technology, Tokyo, Japan). For this analysis, soil samples were extracted with 1 m ammonium acetate (pH 7.0), and the extracts were analyzed for cation concentrations.

Statistical analyses

Statistical analyses were performed using R version 4.1.0 with essential packages.^27–33 Samples were classified as culture-positive for Scedosporium and Lomentospora species based on CFU detection, and their prevalence was analyzed across land usage types (public, recreational, agricultural, and forest). Species occurrence frequency was calculated as the proportion of each species among total detections.

Shannon diversity index was calculated to assess species richness and evenness within each sample. Beta diversity analysis using Bray–Curtis dissimilarity matrix assessed species composition differences between samples. Variability among land usage types was examined through permutational analysis of multivariate dispersions (PERMDISP), followed by anova and Dunn’s test with Bonferroni correction for pairwise comparisons.

Geographic distribution of fungal burden was visualized across Taiwan, with point size representing sample count and color intensity indicating mean fungal burden per location. Sites with mean fungal burdens exceeding 200 CFUs were specifically labeled.

To analyze relationships between fungal burden and environmental factors, log-transformed CFU data were correlated with the HFI^34,35 and soil parameters (nitrogen, carbon, organic content, water content, pH, EC, available phosphorus, and cations) using generalized linear models with a quasi-Poisson family.

Results

Isolation and quantification of fungal burden in soils

A total of 236 fungal isolates belonging to 10 different species within the Scedosporium and Lomentospora species were recovered from soil samples collected across Taiwan (Fig. 1). The most frequently isolated species was S. boydii (76 isolates, 32.2%), followed by S. apiospermum (73 isolates, 30.9%), and S. dehoogii (34 isolates, 14.4%). Less commonly isolated species included S. aurantiacum (19 isolates, 8%), S. hainanense (12 isolates, 5%), P. angusta (11 isolates, 4.7%), and L. prolificans (5 isolates, 2.1%). Scedosporium haikouense was rarely isolated (3 isolates, 1.3%), S. minutisporum (2 isolates, 0.8%), and S. ellipsosporium (1 isolate, 0.4%). The newly generated sequences from this study have been deposited in NCBI GenBank under accession numbers PQ569163 to PQ569295, encompassing 133 isolates. For additional sequence accession numbers, refer to Huang et al. (2023).¹⁸

Fungal burdens varied considerably among sampling sites and land usage types (Fig. 1). The highest fungal burdens were observed in urban environments, particularly in public areas and recreational sites. Notable examples included Taipei Veterans General Hospital Fenglin Branch (up to 1293 CFU/g), Tainan High Speed Rail Station (up to 580 CFU/g), Rushan Visitor Center (up to 516 CFU/g), and Heping Village Park (up to 320 CFU/g). In contrast, agricultural areas and forest sites generally showed lower fungal burdens, typically ranging from 0 to 200 CFU/g. Analysis of alpha diversity revealed a significant positive correlation between fungal abundance and the Shannon diversity index (r = 0.417, P = 6.76 × 10⁻⁷) (Fig. 2), suggesting that sites with higher fungal loads also supported more diverse fungal communities.

Multiple species were often isolated from the same sampling site, particularly in urban areas. Analysis of species co-occurrence revealed that recreational and public areas harbored higher average numbers of coexisting species (0.72 and 0.65 species per site, respectively) compared to agricultural areas (0.32 species per site) and natural forests (0.25 species per site). The most diverse communities were found at Chihshang No. 3 Park (four species: S. apiospermum, S. hainanense, S. minutisporum, and S. dehoogii), followed by Hsinchu Park (three species) and Taipei Veterans General Hospital Fenglin Branch (three species). These findings suggest that recreational and public environments might provide more suitable conditions for diverse Scedosporium and Lomentospora communities compared to agricultural lands and natural forests.

Fungal community dispersion and land use correlations

Analysis of multivariate dispersion revealed differences in fungal community heterogeneity across land usage types. Public and recreational areas showed greater community dispersion (distances to median of 0.607 and 0.560, respectively) compared to agricultural and forest areas (0.541 and 0.537, respectively). However, anova results on the homogeneity of multivariate dispersions confirmed no significant differences among land usage types (F = 0.540, df = 3, 204, P = .655), indicating the degree of variation within each land usage type was similar. Pairwise comparisons using Dunn’s test revealed no significant differences in community dispersion between any land usage types (all adjusted P-values = 1.0) (Fig. 3).

Analysis of soil samples revealed wide variation in chemical parameters across sampling sites. Total nitrogen content ranged from 0.038 to 0.613%, carbon content from 0.398 to 11.363%, and organic content from 0.644 to 19.211%. Water content varied between 1.064 and 10.609%, while pH values ranged from 4.665 to 7.735. EC fluctuated between 0.360 and 1.215 mS/cm, and available phosphorus showed substantial variation from 2.477 to 324.624 ppm. Base cations exhibited the following ranges: sodium (0.105%–0.675%), potassium (0.006%–0.157%), magnesium (0.296%–4.907%), and calcium (8.110%–63.475%). The HFI of sampling sites ranged from 8 to 71, indicating a broad spectrum of human impact.

Statistical analysis showed a significant positive correlation between fungal abundance (mean CFU) and the HFI (r = 0.143, P = .005), indicating that human activity influences fungal distribution (Fig. 4). In contrast, soil chemical factors showed weaker and mostly non-significant correlations with fungal abundance. EC (r = −0.084, P = .580), total nitrogen (r = −0.254, P = .189), and organic content (r = −0.201, P = .346) exhibited slight negative but non-significant associations, while water content (r = 0.006, P = .972) and pH (r = 0.004, P = .983) showed positive but negligible effects. Other nutrients, including phosphorus, sodium, magnesium, calcium, and potassium, also had no statistically significant correlations (all P > .05). These findings suggest that human activity has a stronger influence on fungal abundance than individual soil chemical properties.

Discussion

This study offers a comprehensive ecological analysis of Scedosporium and Lomentospora species distribution across Taiwan, emphasizing the influence of human activity on their habitat preferences. By examining the HFI, soil chemical parameters, and mapping species distributions from 406 diverse soil samples, we identified a strong association between these fungi and anthropogenically disturbed environments. Contrary to prior assumptions that soil chemistry is one of the major determinants, our findings highlight human activity as the primary driver of their distribution. This suggests that urbanization and human presence create unique ecological niches conducive to the persistence and proliferation of these species.

Anthropogenic environments as selective niches for Scedosporium/Lomentospora species

Our ecological survey reveals an intriguing pattern of habitat specialization among Scedosporium and Lomentospora species, with a preference for anthropogenically modified environments. The highest fungal burdens were detected in urban centers, particularly in intensively used public spaces (Tainan High Speed Rail Station: 580 CFU/g) and healthcare facilities (Taipei Veterans General Hospital: 1293 CFU/g). This distribution pattern suggests these fungi have successfully adapted to exploit niches created by human disturbance of the environment. Global ecological studies consistently demonstrate this anthropogenic habitat preference —from European surveys showing these fungi thriving in industrial zones and wastewater treatment plants while being absent from natural forests,^9,10 to Australian studies documenting higher populations in urban cores versus peripheral areas,³⁶ to African surveys finding them concentrated in polluted urban soils,¹⁴ to Thailand studies revealing their prevalence in high human-traffic areas.^11,12 This consistency across diverse geographical regions and urban contexts provides strong evidence that these fungi have evolved traits allowing them to exploit conditions characteristic of human-modified environments.

The strong positive correlation between fungal abundance and Shannon diversity (r = 0.417, P < .001) (Fig. 2) further supports this conclusion, indicating that human-modified environments create conditions favorable not only for increased fungal loads but also for greater diversity of Scedosporium/Lomentospora species. Rather than selecting for single-species dominance, anthropogenic disturbance appears to generate heterogeneous microhabitats that facilitate the coexistence of multiple Scedosporium and Lomentospora species. The particularly high fungal burdens we observed (up to 1293 CFU/g) exceed previously reported densities, indicating these organisms may find optimal growth conditions in Taiwan’s intensively developed urban environments.

Biogeographical patterns reveal regional differences in species community

Our survey documents a distinctive species composition in Taiwan’s urban environments, with S. boydii (32.2%) and S. apiospermum (30.9%) as predominant species, followed by S. dehoogii (14.4%) and S. aurantiacum (8%). This pattern differs notably from other regions: Thailand reported overwhelming S. apiospermum dominance (86.6%),¹¹ while Australian studies showed S. aurantiacum (54.6%) and L. prolificans (43%) prevalence.³⁶ European surveys also revealed regional variations, with S. dehoogii dominating in France¹⁰ and S. apiospermum (69.8%) in Austria.⁹ Unlike these regions where single species typically dominate, Taiwan shows a more balanced community structure. These geographical variations likely reflect complex interactions between urbanization patterns, historical biogeographical factors, and local environmental conditions.

Culture-based detection reveals complex community structure

Our culture-based approach using DG18 agar achieved a 51.2% positive detection rate, comparable to previous studies using various selective media.^11,18,36 These results suggest the potential utility of DG18 in environmental surveys for Scedosporium and Lomentospora, though its efficacy compared to established selective media like DRBC, SceSel+, and Scedo-Select III requires systematic evaluation.^37–39 However, beyond simple detection, our methodology revealed complex patterns of species co-occurrence that provide insights into community assembly. The discovery of up to four coexisting species at single sites (e.g., Chihshang No. 3 Park harboring S. apiospermum, S. hainanense, S. minutisporum, and S. dehoogii) suggests these fungi might have evolved mechanisms for resource partitioning that enable coexistence. This level of species co-occurrence exceeds that typically reported in previous studies, which often found only one or two species per site.^10,11,14,36 While different selective media have been employed across studies (SceSel+, Scedo-Select III), the consistent range of positive culture rates (40%–60%) suggests these methods effectively capture the culturable portion of these fungal communities. The higher species co-occurrence we detected may reflect either more sensitive detection methods or genuinely more complex community structures in Taiwan’s urban environments.

Edaphic factors show limited influence on distribution

Our analysis of soil chemical properties revealed complex patterns in fungal distribution. Despite examining comprehensive parameters including total nitrogen, carbon, organic content, and various base cations, none showed significant correlations with fungal abundance (Fig. 4). Notably, only the HFI demonstrated a significant positive correlation (r = 0.143, P = .005). While previous studies reported associations with specific soil parameters—such as ammonium concentration in Austria⁹ and optimal pH ranges (6.1–7.5) in multiple regions^9,11—our findings suggest more complex relationships. Although we found these fungi across diverse pH conditions (4.6–7.7), similar to observations in Western France (pH 4.9–8.5),¹⁰ pH showed no significant correlation with abundance. The predominant influence of human activity, rather than specific soil conditions, suggests that anthropogenic disturbance is the primary driver of Scedosporium/Lomentospora distribution. Their ability to thrive across varied chemical conditions reflects their metabolic versatility^17,40 and has important implications for understanding infection risks in urban environments.^17,40

Human activity as primary driver of distribution: quantitative evidence

Integrating the HFI with traditional analytical methods provides quantitative evidence linking human activity to the distribution of Scedosporium and Lomentospora. The HFI is constructed from multiple indicators of human pressure—such as built environments, population density, nighttime lights, roads, and navigable waterways—summed to create a composite measure of human influence at a 1 km² resolution.³⁴ Our analysis shows a significant positive correlation between HFI and these fungi. Whereas modest, this correlation is notable given our large sample size (n = 393) and exceeds associations with measured soil parameters. This advances past studies, which relied on categorical land use classifications [e.g., Kaltseis et al. (2009)⁹; Harun et al. (2010)³⁶]. The patterns in species co-occurrence support human influence, with higher fungal species richness in recreational and public areas (0.72 and 0.65 species per site) compared to agricultural (0.32) and forested areas (0.25). The modest correlation coefficient and weak associations with soil parameters suggest fungal distribution is shaped by multiple interacting factors beyond direct human influence, including habitat heterogeneity in human-modified landscapes. These fungi show a preference for human-impacted environments, particularly in polluted waters and sewage systems characteristic of urbanized areas,^1,41,42 as illustrated by numerous case reports of Scedosporium infections following near-drowning incidents.^43–45 Their ability to flourish under the poor aeration and high osmotic pressures typical of polluted environments helps explain their emergence as opportunistic pathogens in human-modified landscapes.

Conclusion

Our findings emphasize that human-disturbed environments significantly shape the distribution and diversity of Scedosporium and Lomentospora species across Taiwan. While the HFI provided a quantifiable link between human activity and fungal distribution, the modest correlation coefficient suggests a need for more specific and relevant indicators to fully capture these patterns. Targeted indicators—such as urban heat intensity, air quality levels, and waste density—could offer clearer insights into how anthropogenic factors shape fungal communities. Expanding research into their genomic adaptations and resilience mechanisms could further enhance our understanding of their pathogenicity and inform strategies to mitigate their impact, particularly in healthcare and densely populated urban settings. Together, these refined approaches may improve predictive models and guide more effective public health strategies to manage the risks posed by these emerging pathogens.

Author contributions

Hsin-Mao Wu (Formal analysis, Investigation), Yu-Hsuan Fan (Investigation, Methodology), Guan-Jie Phang (Investigation, Methodology), Wen-Ting Zeng (Investigation, Methodology), Khaled Abdrabo El-Sayid Abdrabo (Investigation, Methodology), Yu-Ting Wu (Investigation, Methodology, Supervision), Pei-Lun Sun (Resources, Supervision, Validation, Writing—review & editing), Ying-Hong Lin (Resources, Supervision, Validation, Writing—review & editing), and Yin-Tse Huang (Conceptualization, Data curation, Formal analysis, Methodology, Resources, Supervision, Validation, Writing—original draft, Writing—review & editing).

Funding

Y.H.L. was supported by National Pingtung University of Science and Technology. Y.T.H. was supported by Kaohsiung Medical University (NPUSTKMU-113-P003).

Conflict of interest

The authors report no conflicts of interest.

Notes

^‡Co-first authorship.

References