Environmental fungi from cool and warm neighborhoods in the heat island of Baltimore City show differences in thermal susceptibility and pigmentation

Abstract

A major barrier for most fungal species to infect humans is their inability to grow at body temperature (37°C). Global warming and more frequent extreme heat events may impose selection pressures that allow fungal adaptation to higher temperatures. As fungi adapt to warmer environments, they may overcome the thermal barrier that limits infection of warm-blooded hosts, including humans. Cities are heat islands that are up to 8°C warmer than their suburban counterparts and may thus be an important reservoir of thermotolerant fungi that inhabit environments near humans. Here, we describe a novel and inexpensive technique to collect fungal samples from various sites in Baltimore, MD using commercially available taffy candy. Our results show that fungal isolates from warmer neighborhoods show greater thermotolerance and lighter pigmentation relative to isolates of the same species from cooler neighborhoods, suggesting local adaptation. Lighter pigmentation in fungal isolates from warmer areas is consistent with known mechanisms of pigment regulation that modulate fungal temperature. The opportunistic pathogen Rhodotorula mucilaginosa from warmer neighborhoods had a higher resistance to gradual exposure to extreme heat than those from cooler neighborhoods. Our results imply fungal adaptation to increased temperatures in warmer areas of cities. The acquisition of thermotolerance poses a potential risk for humans, as it is necessary for fungal survival within humans.

Introduction

For humans, endothermy provides natural protection against many fungal species with pathogenic potential [1]. Most fungal species are unable to survive at or above human body temperature, and analyses of yeast thermal tolerance have shown that there is a 6% reduction in the number of fungal species able to survive at each degree above 30°C [1, 2].

Natural and human-made disasters are associated with adaptive changes in environmental microbes that can lead to the emergence of new diseases in humans, plants, and animals [3]. Changes in the environment have been hypothesized to enable adaptive changes in fungi and drive the emergence of new fungal pathogens [4], a phenomenon that may have already occurred with Candida auris [4, 5]. Global warming and extreme heat (including prolonged and more frequent heatwaves) may enable environmental fungi with pathogenic potential to acquire higher thermotolerance and thus overcome the mammalian thermal barrier. In recent decades, a sharp emergence of new fungal pathogens such as C. auris and Sporothrix brasiliensis has been proposed to result from environmental pressure to acquire thermotolerance [4, 6]. An analysis of the fungal collection at the Westerdijk Fungal Diversity Institute showed that fungal isolates collected globally from the environment in recent decades have a higher maximum growth temperature than those collected in prior decades, consistent with a possible gain of thermotolerance due to global warming [2]. However, that study is limited as the isolates in their collection are not randomly collected from the environment and may not represent the true diversity of fungi present in the environment over the course of a century.

Environments with high thermal selective pressure include urban centers (cities), which are subjected to the “heat island effect” and tend to be 4–8°C warmer than rural or suburban neighboring regions [7–10]. The warmer environment in cities is predominantly due to reduced green spaces, blocking of natural airflow by buildings, higher heat absorbance by building materials, and anthropomorphic heat production by machines and cars. However, even within cities, there can be a large temperature variation, with some neighborhoods experiencing more severe heat than others, and the warmer neighborhoods are often the home to marginalized communities [10–12]. A study published by the National Oceanographic and Atmospheric Administration (NOAA) reported that in Baltimore City in late August 2018, parks and suburban neighborhoods are ~30.5°C, which is about 8–9°C cooler than the warmest neighborhoods that reach 39.4°C [13].

Currently, there are few studies investigating the mycobiome of urban centers, including areas where there are ample bacterial microbiome analyses, partly due to technical difficulties in performing metagenomic analysis of fungal DNA without adequate reference genomes for comparison and sequence alignment [14, 15]. Few studies have investigated the heat-resistance and phenotypic profiles of urban fungi. An analysis of filamentous environmental fungi from urban sites in Louisville, KY showed faster growth rates and higher enzymatic activities at higher temperatures than the fungi of the same species isolated from rural sites [16].

To understand the mechanisms for the emergence of fungal pathogens, there is a need for more functional characterization of the urban mycobiome and a more detailed characterization of how heat affects urban fungal populations. We hypothesize that cities can serve as sites where thermotolerance and potentially pathogenicity for mammals are selected for, and may emerge, due to the heat island effect, and we hypothesize that we can see evidence for these changes by comparing thermal properties of fungi from warmer neighborhoods versus those from cooler neighborhoods. Consequently, we explored the urban fungal census from sidewalks across thermally diverse neighborhoods in Baltimore, MD using a new technique for collection, and show examples of genotypic and phenotypic characterization of the fungi present, including pigmentation variation and the isolation and identification of clinically relevant polyextremophilic yeasts.

Materials and methods

Sample site and date selection

To select thermally diverse neighborhoods in Baltimore, MD, we used publicly available data on heat disparities across Baltimore City. Based on the most recently available high-resolution data from 2018 published by the NOAA [13], we collected samples from four sites across Baltimore City. For the two warm neighborhood sites, we chose the corner of N. Wolfe St and E. Fayette St. (39°17′38.4″N 76°35′26.5″W, “Fayette St.,” “Site 2”) and the corner of N. Charles St. and Washington Park East (39°17′49.4″N 76°36′55.5″W, “Mt. Vernon,” “Site 3”), representative of the warmest areas in Baltimore. For the average temperature neighborhood site, we chose the corner of S. Wolfe St., and Fell St (39°16′51.8″N 76°35′22.5″W, “Fells Point,” “Site 1”), and for a cool neighborhood site, we chose the corner of N. Charles St. and St. Paul St (39°20′27.7″N 76°37′06.3″W, “Guilford,” “Site 4”). On 20 August 2023, we collected samples from all four sites during sunny weather, between 12 p.m. and 2 p.m. Additional collections were done on 4 September 2023, 21 June 2024, and 28 August 2024, between 12 p.m. and 3 p.m., for only the warmer (Site 2) and cooler (Site 4) sites. All dates were chosen due to sunny weather and were at least 3 days after the most recent precipitation. The two collection dates in June and August 2024 were for the purpose of performing further mold pigmentation measurements. Thermal images to record the temperature of each site were taken using an FLIR C2 IR camera, and images were processed using FIJI (ImageJ) IRimage plugin [17]. Additional temperature readings were done using a GP-300 infrared thermometer (Duraline Systems, West Nyack, NY) to confirm infrared camera measurements. Measurements for the thermometer and FLIR camera were comparable (difference <1°C) (Supplementary Table S1). All temperature recordings (IR and thermal images) were done prior to sample collection.

Sidewalk sample collection

To collect fungal samples from sidewalks, we developed a collection protocol using Starburst taffy candy (Mars Corporation, VA, USA) (Fig. 1A). Starburst taffy is a soft, sticky candy that easily adapts its contour to that of environmental surfaces when pressed against those surfaces with hand pressure and has standardized dimensions and ingredients. The stickiness of the candy picks up the small particles and microbes, which are then released upon dissolving. We used this method instead of conventional swab collection methods to ensure sample collection from the crevices and contours of the sidewalk. To collect environmental fungi found in the gritty and textured surface of a sidewalk, we rolled one yellow Starburst taffy per collection date over an ~100 cm² area on the sidewalk for 30 s, making sure to press the taffy into the sidewalk. The sidewalk sample site was a similar location for each collection date. The taffy candy was then dissolved in 10 ml PBS containing 60 μl of 25% KOH solution to neutralize the citric acid in the candy. After the taffy was dissolved, the suspension was passed through a 100 μm filter to remove debris, and 200 μl aliquots of sidewalk samples were spread on each of three to five 10 cm Sabouraud dextrose agar (SDA) plates with penicillin–streptomycin to limit bacterial growth. SDA was chosen as the isolation agar because it is acidic and inhibits many bacterial species while promoting fungal growth. Samples were incubated at room temperature (22°C) for up to 7 days. To verify, there was little to no contamination from the taffy itself, the taffy was processed as described above except without rolling the taffy on the sidewalk. Three technical replicates from three Starburst taffy (nine cultures in total) alone did not grow fungi when processed with these conditions, except for a single-mold colony that grew after 7 days of incubation, and was most likely an airborne-derived contaminant (Supplementary Fig. S1). This indicates that by culture methods, the Starburst candy itself did not contribute significant fungal contamination.

Figure 1.

Sample collection methods. (A) Fungal sample collection from the sidewalk via Starburst and (B) fungal sample collection from dirt.

Dirt sample collection

Dirt samples were collected by scraping the top layer of dirt along the tree wells, garden patches between sidewalks, and grass-tufted crevices adjacent to the curb. One sample containing at least 5 ml of non-compacted dirt was collected at each location. The 5 ml of dirt sample was added to 45 ml PBS, solubilized at room temperature for 15–30 min, passed through a 100 μm strainer to remove large debris, plant material, and larger insoluble objects, diluted 1:100 in PBS, and 50 μl aliquots of diluted dirt samples were spread on 3–5 10 cm SDA plates containing penicillin–streptomycin and incubated at room temperature (22°C) for up to 7 days. Isolation scheme summarized in Fig. 1B.

Yeast colony isolation

Fungal colonies were isolated from original SDA plates containing material from sidewalk and dirt samples by restreaking on SDA plates. Plates were incubated at room temperature for up to 7 days. Plates were examined daily for fungal growth, both filamentous and yeast like. Yeast-like colonies (small, circular, opaque, and non-filamentous) were picked with a pipette tip and streaked to isolation on fresh yeast peptone dextrose (YPD) agar plates, and incubated at 20°C. Newly appearing colonies derived from the dirt and sidewalk collection plates were picked from the YPD plates daily for up to 7 days. Cells from each collected isolate were observed microscopically to confirm fungal (nonbacterial) identity by cell size, morphology, and appearance of intracellular organelles. Individual fungal colonies were isolated as they appeared, transferred to 96-well plates with freezing media (25% glycerol, 75% YPD broth), and stored at −80°C. Replicate 96-well plates with sterile PBS were inoculated and stored at 4°C. All original plates were wrapped in parafilm and stored at 4°C.

Growth area measurements

All the original samples on SDA agar plates from Sites 2 and 4 were scanned using a Canon CanoScan9000F scanner at 600 dpi after 7 days of growth. Using the FIJI (ImageJ) selection and measure tools, Regions of Interest (ROIs) were manually identified and the total area of fungal growth per plate was determined for each site.

Pigmentation analysis

Pigmentation of fungi was determined as previously described [18]. Briefly, pigmentation of molds growing on the original SDA sample plates was assessed by imaging with a Canon CanoScan9000F scanner at 600 dpi. To measure the pigmentation of yeast isolates, frozen stocks of isolated colonies stored at −80°C were stamped onto YPD agar on an OmniWell agar plate using a flame-sterilized microplate replicator (Boekel Scientific) and grown at room temperature for 7 days. Yeast colonies (which were in spots of ~5 mm in diameter) were scanned using the CanoScan9000F at 600 dpi. Images were processed and the mean gray values of mold and yeast colonies were measured using the FIJI (ImageJ) measure tool.

Thermal absorbance analysis

Agar plates with mold cultures were acclimated to room temperature for 1 h (22°C). The Petri dishes were unlidded and exposed to a prewarmed 19 W 4000 K white light bulb at 50 cm for 10 min. Following the 10 min exposure, culture plates were imaged immediately for heat absorption using an FLIR E96 infrared camera. Images were processed using FIJI (ImageJ) IRimage plugin [17], and temperature measurements were recorded for each mold colony. Mean gray value measurements were taken using a visible light spectrum image taken by the FLIR E96 camera at the time of the experiment.

Heat-ramp assay

Heat-ramp experiments were performed as previously described, with some changes [19]. Frozen stocks of Rhodotorula mucilaginosa and Cystobasidium minutum collected from Sites 2 to 4 were streaked on SDA plates and incubated for 48 h at room temperature. Single colonies of R. mucilaginosa and C. minutum were scraped from SDA plates, suspended in liquid SDA, and adjusted to a density of 0.1 absorbance at 600 nm, and 200 μl of these cultures were grown overnight at 30°C in stationary 96-well plates. Samples were diluted 1:5 in fresh SDA liquid medium, and 100 μl of these cultures were directly treated with a linear heat ramp from 30°C to 55, 56, or 60°C over 8–10 min in a water bath with agitation (LAUDA Scientific, Germany). These temperatures were chosen based on previous literature demonstrating this method as a way to differentiate between heat-ramp-resistant yeast mutants and wildtype [19]. Untreated and heat-ramp–treated strains were immediately spotted (5 μl in SDA) in 5-fold serial dilutions on SDA agar and incubated at 30°C for 48 h. Plates were imaged and Colony Forming Units (CFUs) were counted for each isolate under each treatment.

Internal transcribed spacer sequencing

Colonies for each isolate were picked from newly streaked SDA plates of fungal isolates touched lightly with a pipette tip and transferred and deposited along the inside of an empty Polymerase Chain Reaction (PCR) tube in a thin layer. For isolates that were unable to grow following storage at −80°C, we scraped some of the frozen stock and deposited it inside of a PCR tube. The tubes were microwaved on high (1100 W) for 2 min, as previously described as a method for preparing DNA samples for PCR in fungi [20]. We used this method to simplify large throughput isolation of DNA from a high number of samples in a reliable manner. DreamTaq Green PCR Master Mix (ThermoFisher) was added to each tube with 0.2 μM internal transcribed spacer 4 (ITS4) and ITS5 primers to detect species-specific fungal internal transcribed sequences. The sequences of these primers are as follows: ITS4–5′-TCCTCCGCTTATTGATATGC-3′ and ITS5–5′-GGAAGTAAAAGTCGTAACAAGG-3′. PCR products were amplified according to the manufacturer’s protocol; 3 min initial denaturation at 95°C, 35 cycles with 30 s denaturation at 95°C, 30 s annealing at 55°C, and 1 min extension at 72°C, with a final 5 min extension at 72°C. PCR products were run on 1% agarose gel to confirm product amplification, and Sanger sequenced at the Johns Hopkins School of Medicine Sequencing & Synthesis Core facility. Resulting sequences were analyzed using 4Peaks software and NCBI Blast nucleotide feature. ITS sequences can be found in Supplementary Table S2 and on NCBI Genbank with the accession numbers PX120324:PX120404.

Statistical analysis

All statistical tests were performed using Prism GraphPad Version 10.5.0 for MacOS.

Results

To investigate the potential effects of urban heat islands on fungi, we collected samples from four sites (Fig. 2A and B) across Baltimore City based on 2018 air temperatures from the NOAA [13]. Fungal samples were collected between 12 and 3 p.m. in Baltimore City on four dates with reported air temperatures of 27–37°C during the summers of 2023 and 2024. Collections were done during the summer to ensure the fungi collected were exposed to the extreme heat experienced due to the heat island effect and the lack of greenspace. Sample sites represented both warmer and cooler neighborhoods and each site was representative of the environment in its neighborhood based on the most recently available NOAA heat island maps. The four sites chosen for the study were: Site 1, (Fells Point) representing an average air temperature neighborhood, Site 2 (Fayette St.) representing the warmest area within Baltimore City, Site 3 (Mt. Vernon) representing an area with above average temperature, and Site 4 (Guilford) representing a neighborhood with cooler temperatures.

Figure 2.

Thermal properties of sampling sites in Baltimore, MD. (A) Map of the four sites sampled in Baltimore, MD (Site 1—Fells Point, Site 2—Fayette St., Site 3—Mt. Vernon, Site 4—Guilford). (B) Google StreetView images from the sampling sites taken during summer months in recent years. Infrared images of the sidewalks (C) and dirt (D) on 20 August 2023. The visible light photo from the FLIR 96E camera overlaid on the generated thermal image (top rows of C and D), and the raw grayscale temperature image standardized to 27°C (min) and 60°C (max) with the mean temperature from the center (approximate area shown with a red circle) of the individual image measured overlaid (bottom rows of C and D).

At each collection site, we measured sidewalk and dirt surface temperatures (Supplementary Table S1). The surface temperatures we observed were consistent with patterns we expected based on heat island effect air temperatures published by NOAA. On the first and second collection days (20 August 2023 and 4 September 2023), we collected temperature data using an FLIR E96 camera to verify these temperature patterns. On the second through fourth collection dates, surface temperatures were recorded using a GP-300 infrared thermometer. The raw images from the infrared camera to measure surface temperatures on 20 August 2023 are displayed in the upper panels of Fig. 2C and D, with an automatically scaled color range based on the temperature range in the field of view, which differs in each frame. To more directly compare average temperatures between the dirt and sidewalk sample collection sites, the raw image files were converted to a gray scale (Fig. 2C and D, lower rows) to reflect the absolute temperature range. The grayscale image allows for direct comparisons and quantifications, as it is scaled so that the gray value images have 27°C as the darkest pixel value and the whitest pixel values to represent 60°C. The measured average temperature from the approximate area of the red circle at the center of the image is overlaid.

On 20 August 2023, Site 1 had a warm sidewalk with an average of 40.7°C, while a cooler dirt sample ambient temperature (27.2°C) (Fig. 2C and D). As expected, the dirt and sidewalks from Sites 2 and 3 collection sites had the warmest surfaces tested. Sidewalks in these locations were ~38.4–39.8°C (Fig. 2C), with the dirt temperature at either site of sample collection reaching up to ~53–57.7°C on average, with some areas on the surface reaching up to 60°C (Fig. 2D). Both sidewalk and dirt samples from Site 4 were ~27°C, which was approximately the same as the reported air temperature at the time of collection (Supplementary Table S1), and shows a reduced thermal pressure compared to the high surface temperatures at other areas within the city.

In the subsequent sample collections on 4 September 2023, 21 June 2024, and 28 August 2024, we collected only from Sites 2 and 4 to simplify the collection and comparison process on the basis that they are on two ends of the heat-island landscape within Baltimore City, with Site 2 being among the warmest areas both on the day of our study and as reported by NOAA and Site 4 being the coolest test area. In the subsequent fungal sample collection dates at Sites 2 and 4, surface temperatures recorded via infrared thermometers remained consistent with what we observed on 20 August 2023, and 4 September 2023, using the infrared camera (Supplementary Table S1).

We successfully isolated yeast and mold from all four sites on 20 August 2023, and from Sites 4 and 2 on the remaining collection dates, including the sidewalk and dirt that reached exceedingly high surface temperatures. While fungi were isolated from samples across all neighborhoods, the abundance of culturable fungi varied by location. Notably, for the samples plated from sidewalk and dirt from four collection dates, less area of the petri dish cultures was covered in mold and yeast in plates plated with material from the warmest (Site 2) samples than the plates from the coolest (Site 4) site (Fig. 3). While we were able to recover viable fungi in these extreme temperature environments in cities, there was a reduced amount of viable and culturable fungi present in the hotter areas. This culture-based method indicates that there are fewer culturable fungi present in the warmer neighborhoods, but this method does not account for the presence of species that may have a faster growth rate or species that may have a naturally larger growth area, causing increased relative coverage of the media plate.

Figure 3.

Fungal growth is sparser from Site 2 (warm) samples than Site 4 (cool) samples. Less plate surface area is covered by fungal growth from samples collected from Site 2 compared to those collected at Site 4. Error bars represent 95% CI from the mean. Samples were collected on four collection dates. Each dot represents the percentage of agar area covered in fungal growth from one plate. Statistical significance was calculated with an unpaired t-test.

One fungal property that indicates selection towards certain thermal environments is pigmentation. Pigments produced by fungi can absorb and retain heat, which impacts the cell temperature and supports fungal growth in a heat-deprived environment through a process described as “thermal melanism” [21]. Overall, darker colors tend to absorb more heat energy, and these pigments are often composed of melanins, which have a remarkable capacity to absorb electromagnetic radiation and convert it to heat [18, 21]. We compared mold pigmentation across the urban landscape from sidewalk and dirt on all four collection dates and found that molds from Site 2 (warm—40°C–57°C) were significantly less pigmented (lighter in color) than those from Site 4 (cool—27°C–33°C), with mean gray values differing by ~10-pixel values (Fig. 4A). Similarly, yeast from the first two collection dates were about 12-pixel values lighter from Site 2 (warm) compared to Site 4 (cool) (Fig. 4A and B). As an alternative approach, we directly assessed fungal heat absorbance by mold colonies as measured by an infrared camera. Consistent with having darker pigmentation, the mold found in Site 4 (cool) absorbed more heat during a 10-min exposure to white light than the mold from Site 2 (warm) (Fig. 4C and D). We would expect similar effects for yeast, though their small size precluded analyses by this method. Three mold plates each from samples collected from Sites 2 and 4 collected on 20 August 2023, and 4 September 2023, were chosen at random. There was a correlation between the thermal absorbance (temperature) following exposure to white light (Fig. 4C) and the mean gray value of the colonies (Fig. 4E). This correlation is suggestive of thermal melanism/thermal albinism, as these phenomena have been called, to regulate thermal properties of the fungi [18, 21]. These pigmentation patterns have been previously described for yeast collected at different latitudes, where fungi from arctic regions display enhanced pigmentation that promotes heat absorbance, while equatorial fungi have less pigmentation to avoid excessive heat absorbance [18].

Figure 4.

Fungi collected in thermally diverse neighborhoods exhibit thermal and pigment differences. The mean gray value of mold (A) and yeast (B) isolated from Site 2 is higher than the mean gray value of the fungi isolated from Site 4, indicating less pigmentation in the Site 2 fungi. Each dot represents values from individual colonies. Following exposure to light, fungi from the Site 4 (cool) samples get warmer than those from the Site 2 samples (Panels C and D). Correlation between colony temperature and the mean gray value, where the darker colonies get warmer (Panel E). Panels A and B show mean values with 95% CI and t-tests, while C shows median temperature and non-parametric Mann–Whitney test. Each dot represents mean gray value (Panels A, B, E) or temperature value (Panels C, E) for an individual mold (Panel A, C, E) or yeast colony (Panel B). Samples in (A) were collected from 20 August 2023, 4 September 2023, 21 June 2024, and 28 August 2024. Samples in (Panels B to E) were collected on 20 August 2023, and 4 September 2023.

We sought to compare the thermal properties of fungi of the same or closely related species collected from warmer versus cooler sites. First, it was necessary to identify the species of the urban isolates by PCR amplifying and sequencing the ITS sequence of the 107 collected fungal isolates. We identified the yeast isolates based on the species that most closely matched the ITS DNA sequence on NCBI GenBank, usually with high sequence identity. Notably, rare human pathogens within the ubiquitous Cystobasidiomycetes class were found in all four neighborhood sampling sites, including Rhodotorula spp., Cystobasidium spp., and Cystofilobasidium macerans, (Supplementary Table S2). The identification of the same fungal species in different sampling sites allowed us to compare the characteristics of the same species of fungi when exposed to different environmental pressures. We found R. mucilaginosa, a ubiquitous environmental yeast and rare human pathogen, in both Site 2 (warmer) and Site 4 (cooler). We tested whether the R. mucilaginosa isolate from Site 2 location (Isolate 200S) had a different susceptibility to heat stress compared to those isolated from Site 4 (414S, 423S, and 432S) in laboratory settings. We found that while none were able to grow at continuous 37°C incubation, the single Site 2 isolate, 200S, had greater resistance to killing following exposure to a 30–55°C and 30–56°C heat-ramp stress, remaining more viable compared to the three isolates from Site 4 (Fig. 5A and B).

Figure 5.

R. mucilaginosa from warmer neighborhoods has greater resistance to heat-ramp stress. R. mucilaginosa from the sidewalk of Site 2 (200S) show increased resistance to heat-ramp cell death stress compared to R. mucilaginosa strains collected from the cooler Site 4 sidewalk (414S, 423S, and 432S) after gradual heat ramp to 55°C (Panels a and B). Significance of two-way Analysis of Variance (ANOVA) with multiple comparisons to the 200S for A and B at each condition with three biological replicates. Each dot represents CFUs from an individual biological replicate. **** represents P < .0001, ** represents P < .01, and ns represents P > .05.

We found fungi belonging to the Cystobasidium genus, which includes C. minutum, a rare opportunistic fungal pathogen [22]. Fungi of the Cystobasidium genus were found only at the two warmest sites, Sites 2 and 3. Although one of the C. minutum isolates from Site 3 sidewalk (305S) and one Cystobasidium lysinophilum isolate from Site 2 dirt (242D) were unable to grow at the higher temperatures (Supplementary Fig. S2A), another C. minutum isolate recovered from the 38.4°C sidewalk in Site 3 (300S) was thermotolerant (able to grow at 37°C). We performed similar experiments with the two C. minutum isolated from Site 3 sidewalk (300S and 305S) and found that the 300S isolate was more resistant to heat ramp at 55°C and 60°C (Supplementary Fig. S2B and C). In fact, this 300S C. minutum isolate was the only one able to grow after brief exposure to 30–60°C heat ramp. Together, these data suggest that even within one geographic location, a subset of yeast belonging to the same species is undergoing adaptation or selection for extreme heat. No Cystobasidium isolates were obtained from the cooler Site 4.

Of the 44 yeast-like isolates collected from Site 2 in our study on 20 August 2023 and 4 September 2023, 13 (29.5%) were identified as Aureobasidium pullulans, and had the closest sequence match to A. pullulans PE_11 strain (Supplementary Table S2, Supplementary Fig. S3). A. pullulans is notable since it is a polyextremotolerant fungus, and is tolerant to temperature, saline, and nutrient stress, and was first isolated in glacial ice [23]. A. pullulans isolates had a microscopically diverse morphology with hyphae, yeast-like structures, and other abnormal shapes (Supplementary Fig. S3). Additionally, 5 of the 44 isolates (11%) were identified as Zalaria obscura, a fungus closely related to A. pullulans with similar resistance to environmental stressors [24]. Conversely, only 2 out of the 51 (4%) of the isolates from Site 4 were identified as A. pullulans (Supplementary Table S2). Although, we were unable to perform phenotypic analysis on our A. pullulans isolates due to the inability to grow them from our cryopreserved stock at −80°C, the presence of these species in a neighborhood experiencing extreme heat in Baltimore City is notable and places further context on this fungus’ niche in the environment and as a potential human pathogen.

Discussion

As areas vulnerable to extreme heat conditions, cities are potential crucibles for the adaptation of fungi to warmer temperatures, and this can include fungal species with pathogenic potential that could lead to the emergence of new human fungal pathogens. Here, we investigated the thermal tolerance of fungi in various neighborhoods of a city using a new method to collect and culture environmental fungi. We found that fungi from warmer neighborhoods manifested different thermal tolerances (as measured through a gradual heat-ramp assay) and pigmentation compared to those of colder neighborhoods and our study suggests that the extreme ranges in heat in neighborhoods within urban centers may be sufficient to drive similar differences. While other variables may factor into these findings (e.g. UV exposure, differential exposure to fungi from trees, birds, rodents, and humans, and differences in air flow), our study supports the notion that heat is a major driving factor in the fungal differences between these neighborhoods.

We collected samples during summer months and following periods of warm weather without precipitation to ensure that the fungal samples present were exposed to the extreme heat faced due to the heat island effect and thus had a selective pressure to adapt to heat in warmer neighborhoods, or less of a pressure in the cooler neighborhoods. The temperatures of the sidewalk and dirt in the warmer Site 2 location were 40 and 55°C, respectively, which are both above the typical temperature ranges of fungi, and above human body temperature (37°C), thus providing a selective pressure for higher heat tolerances that, if it selects for higher thermotolerance, could allow them to survive at mammalian body temperature. Meanwhile, the temperatures at the cooler, more wooded and shaded, Site 4 neighborhood were around 27°C, and thus did not provide a selective pressure for growth at or above mammalian body temperature. We noted lower fungal mass growing in agar from warmer sites, consistent with a reduced mycotic flora at those sites. Given that many fungal species are naturally hypothermic [25], and that the fungal viability of fungal species declines rapidly about 30°C [1], one might anticipate that the reduced fungal census in warmer sites reflects a reduced ability to handle heat stress.

R. mucilaginosa isolated from Site 2 (warm) was better equipped to survive gradual exposure to extreme heat (55°C) compared to isolates of the same species isolated from Site 4 (cool). This suggests adaptation to the greater temperatures of the sidewalk at Site 2, implying selection for R. mucilaginosa that can withstand extreme thermal environments for short periods of time, although all four were unable to grow at 37°C. This potential adaptation in R. mucilaginosa to heat is of interest since this fungus causes rare cases of systemic infection in immunocompromised individuals [26–28], and thus demonstrates some ability to withstand growth at human body temperature and host immune defenses. The addition of enhanced thermotolerant properties due to extreme environmental heat may allow this fungus to be even more equipped to survive within mammalian hosts. A relative of Rhodotorula, Rhodosporidiobolus fluvialis, was recently found in two independent clinical infections in China [29]. Researchers found that exposure of this fungus to 37°C conditions induced genetic and phenotypic changes that caused the fungus to have enhanced thermotolerance, drug resistance, and higher virulence phenotypes. These findings support the idea that fungi related to Rhodotorula spp., are in a position to adapt to a warming environment, which can lead to enhanced frequency or severity of cases of disease with this fungus. Rhodotorula spp., is found in nearly every environmental niche across the globe, and is particularly enriched within the guano of the lesser black-backed gull Larus fuscus [30]. This finding is interesting for several reasons. Firstly, the guano samples were specifically collected from urban sites suggesting that gulls may disemminate Rhodotorula within cities through their guano, as was with pigeons and Cryptococcus neoformans [31]. Secondly, the internal temperature of L. fuscus is 41.2°C [32], which is over human body temperature. This indicates that Rhodotorula may have a natural ability to withstand, if not grow, at higher temperatures within an organism, and be primed for inhabiting warm urban sidewalks upon extraction.

We found a strong presence of the extremotolerant fungus A. pullulans in hotter neighborhoods. This fungus is capable of surviving saline, thermal, and nutrient-poor stresses, and can also be an opportunistic human pathogen [33–35]. The presence of this fungus is thus notable as it indicates the selection for extremotolerant fungal populations, which may correlate to ability to tolerate other stressors including mammalian immune response. Gostinčar et al. [23] has noted the potential importance of A. pullulans as an emerging pathogen due to its ability to resist extreme conditions, hyperplasticity, and its presence within the indoor built environment.

The previous comparative study of urban and rural fungi [16], along with the longitudinal analysis of thermal tolerance in culture collection isolates [2] and our results in this study, are each consistent with the notion that fungi are adapting to our changing climate. We have found evidence that urban fungi on the sidewalk and soil of warmer neighborhoods have adaptations that allow them to mitigate heat and survive exposure to extreme heat. Molds and yeast in warmer neighborhoods had reduced pigmentation to avoid absorbing heat. Melanins play diverse roles in fungal biology, including to increase heat absorption, reduce desiccation, and to protect from UV exposure [21]. The inverse relationship seen between pigmentation and neighborhood temperature indicates heat is a main driver of the fungal pigmentation differences, rather than the direct relationship one would expect if it was related to UV or desiccation protection. Previous studies have found a correlation between lighter yeast pigmentation and warmer climates, and darker pigmentations and cooler climates, on a global scale [18]. The employment of darker pigments to help absorb heat is a phenomenon known as “thermal melanism,” which relies on the natural tendency of darker colors to absorb more heat and is used by certain organisms for thermal regulation [21, 36]. We see these previously observed pigmentation gradients in both mold and yeast on a local scale between two microclimates within Baltimore City.

We found that species like R. mucilaginosa can survive exposure to extreme temperatures (55°C) better than those from cooler sites. Even within warmer sites, fungi like C. minutum appear to show varied adaptation to heat where some isolates can survive up to a 60°C heat ramp while another cannot, even though they are presumably recently related. Within these environments, we also see polyextremotolerant fungi such as A. pullulans. Many of these neighborhoods experiencing extreme heat vulnerability (and the risk of emerging thermotolerance in fungi), are also often inhabited by marginalized communities [10, 12, 37]. This is particularly worrisome, as these communities are also the ones already most vulnerable to fungal infection due to healthcare disparities and other social determinants of health [38].

In summary, our study shows differences in thermal tolerance for fungal species within a city corresponding to the neighborhoods of origin and their average ambient and surface temperatures. Our findings suggest that the heat islands defined by cities could be sites for more rapid fungal adaptation to higher temperatures, which brings the possibility that some fungal species with pathogenic potential may be able to overcome human thermal defenses. Other variables including sun exposure, foot traffic, and wildlife can play additional roles in the differences between fungal populations in these sites. To prepare for the possibility that urban heat islands may be the next source of human fungal pathogens, surveillance studies will be needed to establish which fungal species are present, which species exhibit signs of heat adaptations, and which have other virulence-associated traits. Without this systematic surveillance, we risk playing a game of catch up when new pathogenic fungi emerge in the clinic. While this study used culture-based methods to evaluate phenotypic traits of fungi present in the urban environment, future studies may include metagenomic and culture-independent methods to have a complete picture of the fungal landscape within cities, including unculturable fungi. Given that humans and fungi in city ecosystems are in close proximity to one another and that the number of immunocompromised patients are increasing yearly [39], rapid fungal thermal evolution in city heat islands has the potential to bring us new fungal diseases in the years ahead.

Conflicts of interest

None declared.

Funding

None declared.

Data availability

All data generated or analyzed during this study are included in this published article. ITS sequences can be found in Supplementary Table S2 and on NCBI Genbank with the accession numbers PX120324:PX120404.