香京julia种子在线播放

Human lungs are estimated to inhale 500 to 100,000 fungal spores per day depending on local environmental conditions (Hasleton, 1972; Hamm et al., 2020b). Substantial attention has been paid to mechanisms fungi use to avoid the immune response that lungs can mount against pathogens (e.g. Kwon-Chung and Sugui, 2013; Wiesner and Klein, 2017; Ward et al., 2021). Only recently, however, has it been suggested that healthy lungs may harbor as many fungi as lungs diagnosed with overt infections (Richardson et al., 2019), but it remains unclear which fungi in healthy lung tissues present transiently as a result of spore inhalations and which fungi survive as commensals, pathogens, or mutualists.

Once thought to represent a sterile tissue, the lung is now known to possess a microbiome, including a fungal component, the mycobiome (Hamm et al., 2020b). Studies of the human lung mycobiome have been limited, with samples typically derived from sputum or bronchoalveolar lavage rather than from actual tissues. Previous studies typically focused on diseased lungs (Nguyen et al., 2015; Weaver et al., 2019), so knowledge of the human lung mycobiome and more broadly the mammalian lung mycobiome remains limited.

Museum collections provide opportunities to study animal-microbe interactions across broad spatial scales and temporal archives (Cheng et al., 2011; Carvalho et al., 2017; Dunnum et al., 2017; Schindel and Cook, 2018). The goal of the current study was to explore the value of ultra-frozen lung tissues from museum collections to characterize the lung mycobiome of wild small mammals using molecular and cultivation methods. Lung samples were chosen to represent the species diversity from arid environments across the southwestern United States, in part to determine whether such frozen tissues would help define the range and host species of Coccidioides, the causative agent of coccidioidomycosis (Valley fever).

Long before the advent of modern molecular methods there were indications in the 1940s and 1950s that apparently healthy mammals frequently harbor living fungi in their lung tissues. The evidence for this arose in part from attempts by Dr. C. W. Emmons (Emmons and Ashburn, 1942; Emmons, 1943) to determine whether rodents serve as natural reservoirs for coccidioidomycosis. Those studies demonstrated a relatively high incidence of Coccidioides among specific rodent hosts at certain localities (15% of pocket mice and 17% kangaroo rats at San Carlos Indian Reservation, San Carlos, Arizona; Emmons and Ashburn, 1942). Moreover, these and other studies demonstrated that a related fungus, now generally known as Blastomyces parvus (previously Haplosporangium parvum or Emmonsia parva), was detected even more frequently in mammalian lung tissues of a wide variety of mammalian hosts (Emmons and Ashburn, 1942; Jellison, 1950; Bakerspigel, 1956).

Species of Pneumocystis represent a special case in the context of the lung mycobiome, because they are obligately symbiotic lung fungi. They are broadly distributed across mammals, and it has been suggested that each mammalian species has at least one host-specific species of Pneumocystis (Danesi et al., 2020). Molecular surveys for Pneumocystis in domesticated animals and wildlife have used targeted PCR approaches, and success rates have been mixed depending on the animal species targeted (Demanche et al., 2001; Akbar et al., 2012; Danesi et al., 2020). We report that high-throughput sequencing targeting fungal ribosomal internal transcribed spacer (ITS) sequences identifies this group at high frequency (83%) among diverse and widely distributed small mammals. This approach also provides preliminary information regarding specific fungal-host associations and coevolutionary histories for this group.

Our molecular study also expands the results of early lung studies with respect to Coccidioides and Blastomyces and demonstrates that mammalian lung tissues host a diverse mycobiome, with many species either known or now appearing to be adapted to the lung environment. Our results further indicate that, along with Pneumocystidales, members of the Onygenales, Malasseziales, and Saccharomycetales are common constituents of the natural lung mycobiome. We hypothesize that members of all four orders are typically benign commensals rather than aggressive opportunistic pathogens and that the lung mycobiome provides an excellent opportunity to better understand the evolution of commensalism or pathogenicity.

Our Illumina sequencing results produced an average of just over 40 fungal OTUs per lung sample and more than 700 OTUs across the 199 samples examined ( Supplementary Data Sheet 2 ). The following discussion assumes minimal contamination of samples from natural and laboratory environments. As described in Materials and Methods, all PCR-based experiments included negative controls. In addition, while the lung samples employed were collected over a period of twenty-five years by multiple scientists, they were obtained using sterile protocols designed to minimize contamination (Yates, 1996; Yates et al., 1996; Galbreath et al., 2019). The problem of potential environmental and laboratory contamination has received special interest in the context of species of Malassezia, once thought to be restricted to skin but now recognized as common in diverse microbiome and environmental samples unrelated to contamination (Spatz and Richard, 2020). While the potential for contamination by species of Malassezia and other environmental fungi provide a caution for interpreting results of microbiome studies in general, we note that the fungi we have focused on here, especially the Onygenales and Pneumocystidales, are not common in environmental samples.

We recognize four groups of fungi that occur in lung tissues: 1) those present because of incidental inhalation of spores but not truly colonizing lung surfaces or tissues; 2) fungi that are capable of colonizing lung tissue as commensals or pathogens transiently or long-term but are not specifically adapted to the lung; 3) fungi that are common members of the mammalian mycobiome that can exist in lung tissues, but which can also be associated with other tissues such as skin or gut, and 4) fungi adapted to the lung, either obligately (Pneumocystis species) or facultatively (Coccidioides species, Aspergillus fumigatus). These categories will likely be fluid depending on the fungi and mammals in question, for example because of potential differential susceptibility of mammalian species to pathogenicity.

The two most difficult mycobiome groups to distinguish between are those representing incidental inhalation (group 1) and transient colonization (group 2). Fungi present from incidental inhalation might be expected to occur in fewer samples and/or exhibit low read numbers, and their sequences would likely be common in environmental surveys. Such fungi could include sporulating members of the Sordariales, Helotiales, Pezizales, Dothidiales, Eurotiales, Hypocreales, Agaricales, and Pleosporales (Klich, 2002; Fröhlich-Nowoisky et al., 2009; Porras-Alfaro et al., 2011; Hamm et al., 2020a). A study by Kramer et al. of the airway mycobiome of cystic fibrosis patients found that sequences from many fungal species had high fluctuations both among different patients and over time when patients were sampled multiple times, suggesting the pulmonary mycobiome was dominated by species present temporarily because of inhalation of environmental spores (Kramer et al., 2015). Similarly, Rubio-Portillo et al. reported substantial variation among the mycobiomes of human bronchoalveolar lavage samples from different individuals, while in contrast they observed correlations between fungal OTUs from lavage samples and air samples from the home environments of test subjects (Rubio-Portillo et al., 2020). This further supports the potential relevance to lung mycobiome studies of incidental inhalation of spores from the local environment.

The importance of transient colonization (group 2), though difficult to separate from simple incidental inhalation in any specific instance, can be inferred from the fact that common environmental fungi not generally thought of as human pathogens occasionally cause disease, especially in immunocompromised individuals. Such fungi include species of Fusarium, Penicillium, Alternaria, and many others including members of the Mucoromycota (Skiada et al., 2017). And although the immune system of healthy individuals efficiently works against such infections, the triggering of defensive responses may in fact depend on transient colonization (Hernández-Chávez et al., 2017; Gago et al., 2019). With respect to our results, the strong correlation between read abundance and frequency ( Figure 1 ) might support the argument that certain fungi, for example members of the Eurotiales not generally recognized as pathogens, are capable of colonization of healthy lung tissue beyond simply being present passively because of inhalation. This could include those species of Aspergillus that only rarely cause serious infections (Tsang et al., 2020).

We also acknowledge that A. fumigatus, sequences of which were common in our dataset (based on BLAST hits), is special in that it causes the most frequently diagnosed lung mycosis and is at the same time a common environmental saprotroph. Aspergillus fumigatus has multiple adaptations that permit colonization of lung tissues, including small spore size, thermotolerance, and molecular mechanisms that allow avoidance of immune responses (Robert and Casadevall, 2009; Casadevall, 2012; Kwon-Chung and Sugui, 2013; Raffa and Keller, 2019); and in that sense it might fit into either group 4 or group 2. Currently, it is difficult to know for certain whether the apparent frequent occurrence of A. fumigatus in the lungs of small mammals reflects the large number of spores in the environment or colonization of lung tissues. However, the fact that in NMDS analyses samples with high read counts for OTU3 (presumed A. fumigatus) formed a cluster relative to samples with low and no OTU3 reads ( Figure 3D ) may provide support for colonization in such samples.

Taken as a whole, our results have similarities and differences with respect to results obtained for the lung mycobiome of humans. Fungi commonly detected in healthy human lungs have included Candida (Saccharomycetales) Malassezia (Malasseziales), Cladosporium (Cladosporiales), Penicillium and Aspergillus (Eurotiales), and Pneumocystis (Tipton et al., 2017). Similarly, OTUs representing the Malasseziales and Saccharomycetales (including Candida), commonly associated with the human mycobiome (group 3), were well represented in our data (83% and 49% of the samples, respectively; Figure 1 ; Table 1 ). The relatively high diversity of Malasseziales fungi (eight OTUs) in our samples is similar to results reported previously for both healthy and diseased human lungs (van Woerden et al., 2013; Willger et al., 2014; Fraczek et al., 2018).

In our results, many OTUs were both abundant in sequence reads and frequency across samples despite being from different small mammals and various locations ( Figure 1 ). This contrasts with results reported for humans that indicated high inter-individual variability for many OTUs (Kramer et al., 2015; Rubio-Portillo et al., 2020). In addition, sequences from members of the Onygenales were frequent in our samples but have been rare or non-existent in human samples. One potential contributing factor to these differences between our study and human studies is that we sampled lung tissues as opposed to bronchoalveolar lavage or sputum, and as a result our samples have the potential to detect fungi more deeply in tissues than would be obtained with lavage or sputum sampling. Another potential difference could reflect the fact that small, fossorial mammals spend substantial amounts of time in close contact with the soil environment, which could increase exposure to fungi in the soil environment, thereby reflecting more intense sampling of fungi in soil.

The most compelling results from our study relate to the Pneumocystidales and Onygenales, offering new insights into the biology of these organisms. Species of Pneumocystis have long been reported to be common obligate lung fungi with substantial host specificity, and most species are yet to be formally described (Mazars et al., 1997; Demanche et al., 2001; Laakkonen et al., 2001; Akbar et al., 2012). Members of this group possess genomic and physiological adaptations to be obligate biotrophs of mammals (Ma et al., 2016), and they clearly belong to the group of fungi adapted to the lung (group 4). Surveys for species of Pneumocystis in the lungs of non-human primates and bats using specific Pneumocystis-directed PCR produced positive results in 30-40% of the lung tissues examined (Demanche et al., 2001; Akbar et al., 2012), although one study reported a high incidence in shrews (80% of 83 samples; Danesi et al., 2020). Our Illumina approach resulted in 83% of the samples (165 of 199) containing sequences from members of this group. This percentage is almost certainly an underestimate, given that we examined only small amounts of lung tissue from each animal. Our results suggest that in some instances a single lung can harbor multiple Pneumocystis lineages, similar to results reported for P. carinii and P. wakefieldiae (Cushion et al., 2004; Latinne et al., 2021). For example, several different individuals of Dipodomys heermanni had more than 1000 reads assigned to both OTUs 27 and 1292. Moreover, in some instances our data suggest the possibility of lineage spillover across rodent families ( Figure 2B ). These are congruent with the findings of Latinne et al. (2021) obtained for Southeast Asian rodents indicating that certain Pneumocystis lineages may not be restricted to a single host. Nevertheless, host specificity of Pneumocystis species and their animal hosts is apparent for multiple genera of Heteromyidae especially Dipodomys ( Figure 2B ). The evolution of Pneumocystis host niche is complex and not well understood (Babb-Biernacki et al., 2020). Our study suggests that high-throughput non-targeted sequencing can be valuable to gain information toward understanding Pneumocystis diversity and host specificity.

Our results support the hypothesis that certain Onygenalean fungi, including species of Coccidioides and members of the Blastomyces parvus group, are adapted to infect and persist in lung tissues. Taylor and Barker have argued that species of Coccidioides are adapted to be endozoan inhabitants of small-mammal lungs rather than soil fungi with the ability to cause opportunistic infections (Taylor and Barker, 2019). This argument is based in part on genomic studies of Coccidioides species and other members of the Onygenales that showed a reduction of genes involved in plant decomposition and an increase in genes involved in the degradation of animal proteins (Sharpton et al., 2009). Moreover, sequences from species of Coccidioides and Blastomyces are rare in environmental surveys including studies that sample sites where species of Coccidioides are endemic (Porras-Alfaro et al., 2011; Bates et al., 2012), and the detection of Coccidioides species in soils is sporadic even with robust methods (Greene et al., 2000; Barker et al., 2012). Further, OTUs from the Onygenales typically do not have close relatives among sequences from environmental samples (uncultured fungi) in GenBank. This paucity of representation in environmental surveys contrasts with the diversity (19 OTUs) and prevalence (92%) of Onygenalean OTUs in the lung tissues examined here. An illustration of this disparity is found in a comparison of BLAST hits at GenBank using the eight most frequent Eurotiales OTUs in our dataset with those obtained with the eight most frequent Onygenales OTUs ( Supplementary Table 5 ). The Eurotiales had equally good hits to GenBank entries from environmental samples and sequences obtained from known organisms. In contrast, the Onygenales sequences produced close matches to sequences from known organisms but not to environmental sequences.

Although coccidioidomycosis has been reported in diverse animal species (del Rocío Reyes-Montes et al., 2016), there are only two studies reporting the direct isolation of Coccidioides species from wild small mammals. One reported cultured Coccidioides from pocket mice (Chaetodipus intermedius, C. penicillatus, and C. baileyi, reported as Perognathus in the literature but later reassigned to Chaetodipus; Hafner and Hafner, 1983), the southern grasshopper mouse (Onychomys torridus), and a Merriam’s kangaroo rat (Dipodomys merriami) trapped in Arizona (Emmons and Ashburn, 1942). In Baja Mexico, a serology assay detected coccidioidal antibodies in a deer mouse (Peromyscus maniculatus) and desert woodrat (Neotoma lepida; Catalán-Dibene et al., 2014). Our ITS2 data represent the first high-throughput sequencing approach to attempt detection of Coccidioides species in wild small-mammal lungs. Here, we report 14 species of small mammals with sequences from Coccidioides, nine of which (Ammospermophilus harrisii, Dipodomys heermanni, Neotoma albigula, Neotoma stephensi, Otospermophilus variegatus, Perognathus ampulus, Peromyscus boylii, Sylvilagus audubonii, and Thomomys bottae) are newly documented host species ( Figure 4A ).

Early research from Emmons and colleagues suggested a higher correlation with heteromyid rodents and Coccidioides in the environment (Emmons and Ashburn, 1942; Emmons, 1943). Within the Heteromyidae rodents in our study, the genus Chaetodipus had a 15% positivity rate and Dipodomys had a 10% positivity rate, but they did not greatly differ from our overall average Coccidioides detection frequency (12%) ( Figure 4B ). In terms of future studies, Neotoma (woodrats) may be of particular interest with a Coccidioides positive rate of 24%, two-fold higher than our overall average ( Figure 4B ). An overlap analysis between the distribution models of rodents and species of Coccidioides has suggested that Neotoma lepida is a predominant co-occurring species (Ocampo-Chavira et al., 2020).

We were limited by sample size in terms of the Leporidae, Sciuridae, and Muridae. Notwithstanding, one Sylvilagus audubonii (desert cottontail) sample was positive for Coccidioides ( Figure 4A ). Sylvilagus audubonii DNA was found in association with Coccidioides positive burrows in Arizona (Kollath et al., 2019). Despite having only five samples of sciurids (squirrels), we detected Coccidioides in two genera (Ammospermophilus and Otospermophilus) corresponding to 40% of the samples. Only two murid samples were obtained, and neither were positive for Coccidioides.

With the expected expansion of coccidioidomycosis endemic regions in the United States (Lauer, 2017; Gorris et al., 2019), efforts have increased to model the factors affecting the distribution of the disease and predict its expansion under climate change scenarios. These efforts have, however, focused on environmental factors (climate, soil, elevation, and land cover variables) to predict ecological niches for C. immitis and C. posadasii (Baptista-Rosas et al., 2007; Gorris et al., 2018; Weaver et al., 2020; Dobos et al., 2021). Our study supports the prediction that small mammals play important roles as host reservoirs of Coccidioides species (del Rocío Reyes-Montes et al., 2016; Taylor and Barker, 2019), suggesting that the ecology, distribution, and taxonomy of small mammals should be incorporated into Coccidioides modeling efforts.

The presence of Coccidioides sequences does not appear to disrupt the lung mycobiome ( Figure 3C ). Studies with chytridiomycosis in frogs (Jani and Briggs, 2014) and snake fungal disease (Allender et al., 2018) suggest that these infections cause a disruption of the microbiome and a decrease in both fungal and bacterial diversity with fungal pathogen introduction. In sea star wasting disease, changes in microbial community composition occurred during disease progression, with decreasing species richness only in the late stages of the disease (Lloyd and Pespeni, 2018). The presence of Coccidioides without mycobiome disruption might imply a primarily commensal relationship with its host until a change in host conditions favoring a pathogenic state.

Emmons suggested a correlation between the presence of Blastomyces parvus (previously Haplosporangium parvum) and Coccidioides stating, “The isolation of H. parvum from rodents of a given area is taken to indicate that C. immitis is probably also present in that area” (Emmons, 1943). Our ITS2 sequencing revealed Blastomyces sequences in 85% of the samples suggesting that it is more common in small-mammal lungs than are species of Coccidioides (12%). When abundances were rarified, Blastomyces was present without Coccidioides in 77 samples, but only co-occurred with Coccidioides in 6 samples. A Spearman correlation and a permutation test found no statistical correlation between the two Onygenalean genera ( Supplementary Table 3 ). Even without correlation, it is notable that Blastomyces parvus is in high frequency in small-mammal lungs despite its ability to cause adiaspiromycosis (Emmons and Ashburn, 1942). Emmonsia crescens (OTU15), detected in 16% of the samples, is an additional etiological agent of adiaspiromycosis ( Table 1 , Figure 2A ) (Emmons and Jellison, 1960). Adiaspiromycosis often presents as a pulmonary infection common to fossorial mammals like rodents but is much rarer in humans (Anstead et al., 2012).

Lung samples employed in this study were chosen to reflect the geographic range of human coccidioidomycosis from low to high incidence to determine if we could detect sequences from species of Coccidioides. One of the most revealing aspects of our results, however, had to do with the high frequency of sequences from B. parvus and other members of the Onygenales among lung samples (92%). In addition, B. parvus was the most common fungal species cultured from lung tissue. We argue that the endozoan, small mammal reservoir hypothesis (Taylor and Barker, 2019) should be expanded beyond species of Coccidioides to encompass multiple Ajellomycetaceae (Onygenales) fungi including Blastomyces, Emergomyces, Emmonsia, and Emmonsiellopsis, an argument with significant clinical relevance. The phylogenetic diversity within this group is becoming ever more important to the medical field as over the last four decades there has been an increase in reports of novel Emmonsia-like human pathogens (Schwartz et al., 2015). Our study and others (Danesi et al., 2020) demonstrate the commonality of these Emmonsia-like fungi globally in wild animals. Host jumps have allowed and will likely continue to allow these fungi to evolve and diversify followed by radiation, specialization, and speciation which in some cases could increase their virulence (Thines, 2019).

This study was made possible by the frozen tissue collections of the UNM Museum of Southwestern Biology and the UCB Museum of Vertebrate Zoology. While acknowledging the value of the availability of diverse, frozen tissue collections, we also note that molecular barcoding of specimens to confirm host identifications is best practice (Dunnum et al., 2017). About 11% of the samples in this study showed incongruence between host identification based on cyt b sequences and that reported in museum databases. Nine of these frozen samples can be attributed to tissues that were apparently mislabeled at the time of collection, while 13 samples represent misidentification of closely related cryptic species for which morphology alone is inconclusive. Misidentification based on morphological features alone has been reported to be as high as 10% in some pathogen studies (Müller et al., 2013); however, without the ability to both test the host identity with a molecular approach and return to a physical voucher, these kinds of incongruencies would have been impossible to rectify. Directly linking genetic data to a physical specimen should become a necessary component of infectious disease studies (Thompson et al., 2021).

The datasets presented in this study are available in online repositories. Access to the datasets is at https://www.ncbi.nlm.nih.gov/ with the accession number: PRJNA769405. Further inquiries should be directed to the corresponding author.

Ethical review and approval was not required for the animal study because this study involved frozen lung tissues from museum collections deposited over several decades by multiple researchers. No live animals were employed or sacrificed for the study. Although not directly related to our study, typical animal handling and current Animal Welfare Assurance protocols employed by scientists at the University of New Mexico Museum of Southwestern Biology are presented in Materials and Methods under the subheading “Tissue acquisition.”

PS-H, DN, JT, and JC conceptualized the project. PS-H, KM, LM, SL, and JC collected data. PS-H, KC, SL, and DN analyzed the data. PS-H and DN prepared the initial draft of the manuscript, and all authors participated in revisions. PS-H, JT, and DN provided funding. All authors contributed to the article and approved the submitted version.

PS-H was supported in part by a UNM Sevilleta LTER Summer Graduate Student Fellowship (NSF awards DEB 1655499 and DEB 1440478) and the UNM Sevilleta Field Station endowment fund, with additional support from research awards from the UNM Graduate and Professional Student Association and the Department of Biology Graduate Research Allocations Committee. PS-H was also supported by the Mycological Society of America Graduate Research Fellowship and the John W. Rippon Research Award. JT acknowledges the Valley Fever Research Initiative at the University of California, VFR-19-633952 and grant R01 AI148336 from the NIAID. JC acknowledges NSF support (NSF1561342, 2033482). Support was also provided by the UNM Department of Biology’s Molecular Biology Facility, supported by the UNM Center for Evolutionary & Theoretical Immunology (CETI) under National Institutes of Health grant P30GM110907.

This project was made possible by frozen tissue collections preserved at the Museum of Southwestern Biology at the University of New Mexico (UNM) and the Museum of Vertebrate Zoology at University of California Berkeley. The authors would like to thank the field crews that collected these tissues, in particular Dr. James L. Patton. We additionally would like to thank the UNM Center for Advanced Research Computing, supported in part by the National Science Foundation (NSF), for providing high performance computing resources used in this work.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.