Life Cycle Dominates the Volatilome Character of Dimorphic Fungus Coccidioides spp.

ABSTRACT Valley fever (coccidioidomycosis) is an endemic fungal pneumonia of the North and South American deserts. The causative agents of Valley fever are the dimorphic fungi Coccidioides immitis and C. posadasii, which grow as mycelia in the environment and as spherules within the lungs of vulnerable hosts. Current diagnostics for Valley fever are severely lacking due to poor sensitivity and invasiveness, contributing to a 23-day median time to diagnosis, and therefore, new diagnostic tools are needed. We are working toward the development of a breath-based diagnostic for coccidioidomycosis, and in this initial study, we characterized the volatile metabolomes (or volatilomes) of in vitro cultures of Coccidioides. Using solid-phase microextraction (SPME) and comprehensive two-dimensional gas chromatography coupled to time of flight mass spectrometry (GC×GC-TOFMS), we characterized the volatile organic compounds (VOCs) produced by six strains of each species during mycelial or spherule growth. We detected a total of 353 VOCs that were at least 2-fold more abundant in a Coccidioides culture than in medium controls and found that the volatile metabolome of Coccidioides is more dependent on the growth phase (spherules versus mycelia) than on the species. The volatile profiles of C. immitis and C. posadasii have strong similarities, indicating that a single suite of Valley fever breath biomarkers can be developed to detect both species. IMPORTANCE Coccidioidomycosis, or Valley fever, causes up to 30% of community-acquired pneumonias in highly populated areas of the U.S. desert southwest where the disease is endemic. The infection is difficult to diagnose by standard serological and histopathological methods, which delays appropriate treatment. Therefore, we are working toward the development of breath-based diagnostics for Valley fever. In this study, we characterized the volatile metabolomes (or volatilomes) of six strains each of Coccidioides immitis and C. posadasii, the dimorphic fungal species that cause Valley fever. By analyzing the volatilomes during the two modes of growth of the fungus—mycelia and spherules—we observed that the life cycle plays a significant role in the volatiles produced by Coccidioides. In contrast, we observed no significant differences in the C. immitis versus C. posadasii volatilomes. These data suggest that life cycle, rather than species, should guide the selection of putative biomarkers for a Valley fever breath test.

and between life cycles but also posited that there would be a subset of VOCs that are common to Coccidioides spp. during spherule growth, which will serve as our candidate biomarkers for breath test development.

RESULTS
Coccidioides isolates in this study. The Coccidioides genus is widely distributed throughout North and South America. The genus is composed of two geographically (19) and genetically (19,20) distinct species, C. posadasii and C. immitis, which are further divided into five well-defined populations (21)(22)(23). C. posadasii has a more extensive biogeographic distribution and is divided into three populations, occurring in Texas/Venezuela, Mexico/South America, and Arizona. The species C. immitis is divided into two populations found in central and southern California and was recently found in Washington State, representing a newly discovered third distinct population (24,25). Representative members of all populations cause disease in humans and other mammals (21)(22)(23). Due to the magnitude of the observed genetic variation and geographic distribution between species, we hypothesized that metabolomic variation exists. To capture diversity across the volatilome between and within species, we included a total of 12 isolates from both species, representing four populations, all of which were isolated from humans ( Table 1).
Characteristics of the Coccidioides volatilome by species and life cycle. We cultured each Coccidioides isolate in biological triplicate using temperatures and oxygen concentrations that induce mycelial or spherule growth but using the same culture medium, facilitating direct comparisons between the VOCs produced during the two life cycles. Using SPME to adsorb VOCs from the headspace of fungal culture filtrates and GCÂGC-TOFMS to analyze the volatilomes, we detected a total of 353 chromatographic peaks that were at least 2-fold more abundant in a Coccidioides culture (n = 72) than in the medium controls (n = 6) (see Table S1 in the supplemental material). Of the 353 volatiles detected in Coccidioides cultures, 28 were identified at level 1 or 2 and therefore were assigned putative names based on mass spectral and chromatographic data (see "Processing and analysis of chromatographic data" in Materials and Methods, below), which included 13 compounds previously associated with human and environmental fungal pathogens (26) (Table 2). For the unnamed compounds, we assigned chemical classifications to 45 of them, which were assigned a level 3 identification based on a combination of mass spectral and chromatographic characteristics (Table S1).
The total numbers of volatiles produced by C. posadasii and C. immitis were similar, at 291 and 309, respectively ( Fig. 1; Table S2). Within each species, the two life cycles are quite divergent in their volatilomes, sharing less than one-third of the total volatilome ( Fig. 1A and B). However, within each life cycle, the two species share many similarities, with more than one-half of the VOCs being produced by both C. posadasii and C. immitis ( Fig. 1C and D). In aggregate across the genus, or subdivided by species, the spherule volatilome is larger than the mycelial volatilome, but this appears to be driven by considerable differences in volatilome sizes observed in a few isolates (C. posadasii B3221 and GT-166 and C. immitis RS) (Table S2).
Analyzing the relative abundances of individual analytes, we observed some statistically significant differences in VOCs produced during the two life cycles but not between the two species. We performed a Mann-Whitney U test using Benjamini-Hochberg false discovery rate (FDR) correction comparing all spherule and mycelial VOCs and identified 43 that were significantly different in abundance (P , 0.05) (Fig. 2). Of the 43 volatiles, 32 of them had a higher abundance in spherules, compared to 11 in mycelia, which mirrors the overall difference in the sizes of the volatilomes (Fig. 1). Among the VOCs detected in higher abundances in one life cycle than in another were the aromatic hydrocarbon 1-methyl-3-(1-methylethyl)-benzene and the aldehyde hexanal, which are associated with mycelial cultures, while the alcohols 2-(2-butoxyethoxy)ethanol and 2-hexen-1-ol, the ketone cyclohexanone, and the heteroaromatic 2,4,6-trimethylpyridine were associated with spherule cultures. There were no VOCs that were significantly different in relative abundance between C. posadasii and C. immitis.
Most (.70%) of the spherule and mycelial VOCs detected in this study were produced by four or fewer Coccidioides strains in each life cycle. Of the 272 spherule volatiles that were detected, only 35 (13%) were detected in at least two-thirds of all spherule cultures (Fig. 3A), 12 of which were also significantly more abundant in spherule than in mycelial cultures. Nineteen of 224 mycelial volatiles (8%) were detected in at least two-thirds of all mycelial samples (Fig. 3B), and only 2 were significantly different between life cycles (VOCs 199 and 428). However, both volatiles were produced in higher relative abundances in spherules. Only two VOCs, VOCs 39 and 748, were detected in two-thirds of both spherule and mycelial samples. Interestingly, the relative abundance of a VOC was usually consistent across all strains. For example, in every spherule culture in which limonene (VOC 383) was detected, it was present at a relative concentration that was between 2-and 10-fold higher than that in the medium (Fig. 3A).
Multivariate analyses of the Coccidioides volatilome. Principal-component analysis (PCA), using the 78 cultures and medium blanks as observations and 353 Coccidioides VOCs as variables, shows that the volatilomes of individual strains and culture conditions are highly reproducible (Fig. S1A) and that they cluster by life cycle and not by species (Fig. 4). This observation is reinforced when we include only mycelia or spherules in the PCA, where no separation between species is observed ( Fig. S1B and C), or include only C. immitis or C. posadasii, where we still observe a clear separation by life cycle (Fig. S1D and E). Evaluating these subsets of the data, some interesting variations between strains emerge. The type strain for C. immitis is RS; however, in both mycelial and spherule life cycles, we observe that its volatile metabolome is quite unique from those of the other C. immitis strains (Fig. S1). C. posadasii also has a metabolic outlier, RMSCC3506, but only during mycelial growth ( Fig. S1B and E). These outliers cannot be fully explained by the size of the volatilomes as being abnormally large or small (Table S2).
Although life cycle and strain differences in the Coccidioides volatilome are observable in the PCA, less than one-quarter of the total variance is captured by the first two principal components. Thus, we also used hierarchical clustering analysis (HCA) to compare the Coccidioides volatilomes (Fig. 5). The HCA reinforces our finding that the life cycle is the most significant contributor to the Coccidioides volatilome, and there are not global similarities in the volatilomes within each species or populations of the species. There are two main clusters of VOCs driving the separation of the mycelial and spherule life cycles in the HCA. Cluster 1 includes 56 VOCs, which are generally present Life Cycle Dominates the Volatilome of Coccidioides in a higher abundance in mycelial samples than in spherule samples, 5 of which are significantly different between life cycles (P , 0.05 with Benjamini-Hochberg FDR correction) (Table S1). In cluster 1, seven compounds are identified: the hydrocarbon 3undecene, the aromatic hydrocarbons 1-ethyl-4-methyl-benzene and 1-methyl-3-(1methylethyl)-benzene, the alcohol 2-butyl-1-octanol, the aldehyde decanal, and the monoterpenes limonene and b-pinene. As in the PCA, we find that C. immitis RS and C. posadasii RMSCC3506 cluster separately when grown as mycelia, partially driven by the low abundances of VOCs from cluster 1. Cluster 2 includes 16 VOCs that are present in spherule samples while generally being absent or found at a lower abundance in mycelial samples, all of which are significantly different between life cycles (P , 0.05 with Benjamini-Hochberg FDR correction). Cyclohexanone (a ketone) and 2,4,6-trimethylpyridine (a heteroaromatic) are two VOCs in this cluster that we were able to identify.

DISCUSSION
The genus Coccidioides is divided into two closely related and putatively allopatric species, with a strong phylogeographic signal and population substructure within each species (20). However, we find that the character of the Coccidioides volatilome is determined by life cycle rather than species. Within each species, the two life cycles have a relatively small overlap in their shared volatilome, with 27% and 31% being shared for C. posadasii and C. immitis, respectively (Fig. 1). While there were no volatiles that were significantly different in relative abundance between C. posadasii and C. immitis, we observed 43 statistically significant differences in the VOCs produced during the two life cycles (Fig. 2). This global difference between life cycles is the primary driver of separation that we observed in multivariate analyses using PCA (Fig. 4) and HCA (Fig. 5). These findings are in line with the large-scale transcriptional changes that occur when Coccidioides species transitions from mycelial to spherule growth: between 20% and 40% of the Coccidioides genes are differentially regulated during the shift between the environmental and parasitic life cycles (27)(28)(29)(30). It has also been shown that the proteomes of the two life cycles are distinct, with statistically significant differences between mycelial and spherule life cycles (31). In contrast to the significant differences in the volatilomes between life cycles, we observed no uniform differences between the volatilomes of C. posadasii and C. immitis. There is a high degree of overlap (56 to 57%) between the species' volatilomes when comparing the same life cycle (Fig. 1). This similarity is also visible in the PCA (Fig. 4; see also Fig. S1 in the supplemental material) and HCA (Fig. 5), as the volatilomes do not cluster by species even when we Open shapes indicate VOCs with at least a 2-fold-higher relative abundance than in medium blanks, and solid shapes indicate VOCs with at least a 10-fold-higher relative abundance. Strains are color-coded as blue for C. immitis and red for C. posadasii. The percentages of listed VOCs detected in each strain and the percentages of strains producing each VOC are noted at the bottom and right sides of each chart, respectively. Compound identifications: 97, acetic acid; 239, 1-hexanol; 255, 2-hexen-1-ol; 383, limonene; 428, 2,4,6,-trimethylpyridine; 820, 2-butyl-1-octanol.
Life Cycle Dominates the Volatilome of Coccidioides control for the life cycle. Global similarities between C. posadasii and C. immitis have also been demonstrated at the protein level (31,32).
The lack of significant differences between the C. posadasii and C. immitis volatilomes is due, in part, to large differences observed between strains of the same species; the volatilome highlights some unique aspects of a few of the individual Coccidioides strains that we analyzed. For example, RS is the type strain of C. immitis; however, it appears to be a metabolic outlier during both life cycles (Fig. S1) but especially during mycelial growth (Fig. 5). Additionally, the mycelial life cycle of C. posadasii RMSCC3506 appears to be a metabolic outlier compared to the other strains grown as mycelia ( Fig. 5; Fig. S1). These mycelial outliers appear to be driven, at least in part, by the lack of volatiles that make up cluster 1 in the HCA (Fig. 5). None of the 56 VOCs found in cluster 1, which are more abundant in mycelial than in spherule cultures, are detected in RMSCC3506 cultures, and only 2 are detected in RS mycelial cultures. These two strains also lack VOCs 39 and 748, which were detected in the majority of strains for both life cycles but absent from both life cycles of RS and mycelial cultures of RMSCC3506 (Fig. 3). Additionally, none of the 11 VOCs that were significantly more abundant in mycelial than in spherule cultures ( Fig. 2 and Table S1) are detected in RMSCC3506, and only 2 are detected in RS. Beyond contributing to high levels of volatilome variance in Coccidioides, the significance of the strain-level differences relative to fungal biology or infection is not known. Variation in virulence among strains in murine models of coccidioidomycosis has been observed but has not been tested in the context of species or with a mechanistic hypothesis (33). In future studies of mouse models and human infections, we plan to investigate correlations between the Coccidioides volatilome and virulence and infection outcomes.
It should be noted that although the spherule and mycelial media had the same composition, we detected different VOCs in the headspace of the blanks (Fig. 4 and 5; Table S1). We interpret these differences as being the influence of the cultures on the blanks rather than the influence of the blank media on the cultures. Because the medium blanks were incubated and processed in parallel with their respective samples, we posit that the blank media absorbed volatiles from the culture environment (e.g., from fungal spherule or mycelial cultures in shared incubators), which introduced differences in their volatilomes. An analysis of fresh blank media (not incubated with the cultures) would be required to test this hypothesis. Our data indicate that the spherule media absorbed a higher abundance of compounds than the mycelial media (Table S1), which may be due to the spherule life cycle producing higher numbers and/or abundances of individual compounds. Therefore, it is possible that there are additional significant spherule volatiles that were not identified in this study since a VOC was defined as "detected" based on its relative abundance versus medium blanks. The results from this study highlight two crucial factors that need to be considered in the development and translation of volatilomics (and other functional genomics data) into biomarkers. First, there can be a significant amount of intraspecies variation such that the volatilome of a single strain is not representative of the species as a whole (17,(34)(35)(36). Therefore, an analysis of multiple strains should be performed to identify a core volatilome, and the more genetically and phenotypically diverse the species, the more strains will need to be analyzed to determine the consensus set of volatiles produced by the majority of them. Second, our data show that the growth phase of the organism has a strong influence on the volatilome, and therefore, using in vitro culture conditions that can most closely replicate the in vivo environment and microbial growth phases may increase the likelihood of putative biomarker translation from bench to bedside. In addition to controlling the growth phase of dimorphic organisms, macronutrient, micronutrient, and oxygen availability should all be considered in the development of in vitro model systems that replicate the in vivo physiology of the pathogen (14,17,(36)(37)(38)(39)(40)(41). In this study, we observed 35 VOCs present in at least 67% of all spherule strains (Fig. 3); as this is the parasitic life cycle found within the host, we hypothesize that the most highly conserved spherule VOCs have the highest probability of translating to animal models and to human Valley fever breath biomarkers.
The first priority for diagnosing pneumonia etiology is determining whether the underlying cause is bacterial, fungal, or viral, as this will guide the selection of antibiotics, antifungals, or neither for treating the infection. We have initiated a study to directly compare coccidioidomycosis, bacterial pneumonia, and viral pneumonia in Arizona residents to determine the specificity of the Coccidioides spherule volatiles for diagnosing the fungal pneumonia that is endemic in this region of the United States. There have not yet been any human breath studies designed to compare multiple fungal and bacterial pneumonias to determine if there might also be a universal set of fungal breath volatiles, but given the limited overlap in the regions of the United States where fungal pneumonias are endemic (42), a set of universal fungal breath volatiles will not be required to substantially improve pneumonia diagnostics. It may be of clinical or epidemiological interest to differentiate coccidioidomycosis from other fungal pneumonias such as aspergillosis, blastomycosis, histoplasmosis, or cryptococcosis. Previous studies have shown that bacterial lung pathogens have unique and reproducible volatilomes that can be used to detect and identify organisms at the genus, species, or strain level (11), and therefore, differentiating fungal pneumonias is feasible.
Conclusion. The Coccidioides volatilome is strongly influenced by the saprobic versus parasitic life cycles, while the two species, C. posadasii and C. immitis, are indistinguishable by their volatile profiles, in part due to high strain-to-strain variability within the species. We identified 35 VOCs produced by the majority of the Coccidioides strains when grown as spherules, the parasitic form of the fungus found in Valley fever lung infections and in disseminated fungal disease. Our results show that as future efforts are made to develop biomarkers for Coccidioides and other mycoses caused by dimorphic fungi, it is vitally important to control for the fungal life cycle to represent parasitic growth in the host and to sample multiple strains of the pathogen to identify volatiles that are more highly conserved within the infectious genus or species.

MATERIALS AND METHODS
Fungal isolates and growth conditions. The Coccidioides isolates used in the study were obtained from human patients and are listed in Table 1. All Coccidioides isolates were grown under biosafety level 3 (BSL-3) containment, under conditions that induce mycelial or spherule growth.
Fungal growth conditions were performed as described previously (18,43) and are summarized here in brief. For mycelial growth, a 50-ml vented falcon tube containing 10 ml of RPMI medium (filter-sterilized RPMI 1640, 10% fetal bovine serum) was inoculated with a 1-cm by 1-cm 2Â glucose yeast extract (GYE) agar plug for each strain. These plates were inoculated using 100 ml of a glycerol stock, spread across the plate, and cultured at 30°C for 2 weeks. Control RPMI medium was inoculated with a plug from sterile 2Â GYE agar medium. Each sample, including the medium control, was prepared in triplicate. Cultures were grown on a shaking incubator at 150 rpm at 30°C for 96 h. For spherule cultures, a 50-ml vented falcon tube containing 10 ml of RPMI medium was inoculated to a final concentration of 1.0 Â 10 5 arthroconidia/ml in 1Â phosphate-buffered saline (PBS). Arthroconidia were grown and harvested as previously described (18,43). Strains RMSCC2343 and RMSCC3505 did not produce enough conidia to achieve 1.0 Â 10 5 arthroconidia/ml and were inoculated at 7.0 Â 10 4 and 4.0 Â 10 4 arthroconidia/ml, respectively. Control medium was inoculated with 1 ml of sterile 1Â PBS. Cultures were grown on a shaking incubator at 150 rpm at 39°C in 10% CO 2 for 96 h. Mycelial and spherule cultures were spun at 12,000 Â g at 4°C for 10 min to pellet the cells. The supernatant was removed, placed in a Nanosep MF centrifugal device with a Bio-Inert membrane 0.2-mm spin filter, and centrifuged at 3,200 Â g for 4 min. The filtrate was stored at 280°C until volatile metabolomics analysis.
To ensure the sterility of the filtrates for metabolomics analyses outside a BSL-3 containment facility, 10% of all sample filtrates were plated on 2Â GYE and incubated at 30°C for 96 h to ensure the complete removal of viable pathogen particles. No growth was observed in any replicate.
Volatile metabolomics analysis by SPME-GC3GC-TOFMS. The Coccidioides species culture filtrates and medium blanks were allowed to thaw at 4°C overnight, and 2 ml was then transferred and sealed in sterilized 10-ml GC headspace vials with polytetrafluoroethylene (PTFE)-silicone septum screw caps. All samples were stored for up to 12 days at 4°C until analyzed. Samples were randomized for analysis. Volatile metabolite sampling was performed by solid-phase microextraction (SPME) using a Gerstel multipurpose sampler directed by Maestro software. Sample extraction and injection parameters are provided in Table S3 in the supplemental material (autosampler method). Volatile metabolite analysis was performed by two-dimensional gas chromatography-time of flight mass spectrometry (GCÂGC-TOFMS) using Leco (St. Joseph, MI) Pegasus 4D and Agilent 7890 GC instruments. Chromatographic, mass spectrometric, and peak detection parameters are provided in Table S3 (GCÂGC method and mass spectrometry method). An external alkane standard mixture (C 8 to C 20 ; Sigma-Aldrich, St. Louis, MO) was sampled multiple times for calculating retention indices (RIs). The injection, chromatographic, and mass spectrometric methods for analyzing the alkane standards were the same as the methods for the samples.
Processing and analysis of chromatographic data. Data collection, processing, and alignment were performed using ChromaTOF software version 4.71 with the Statistical Compare package (Leco Corp.), using the parameters listed in Table S3 (data processing method).
Peaks were assigned a putative identification based on mass spectral similarity and RI data, and the confidence of those identifications is indicated by assigning levels 1 to 4 (with 1 being the highest) (44). Peaks with a level 1 identification were identified based on mass spectral and RI matches with external standards. Peaks with a level 2 identification were identified based on $800 mass spectral matches by a forward search of the NIST 2011 library and RIs that are consistent with the midpolar Rxi-624Sil stationary phase, as previously described (34), but using an RI range of 0 to 43% (empirically determined by comparing the Rxi-624Sil RIs for Grob mix standards to published polar and nonpolar values). Level 1, 2, and 3 compounds were assigned to chemical functional groups based upon characteristic mass spectral fragmentation patterns and second-dimension retention times, as previously described (35). Level 4 compounds have mass spectral matches of ,600 or RIs that do not match previously published values and are reported as unknowns.
Data postprocessing and statistical analyses. The data postprocessing steps are depicted in Fig. S2. Before statistical analyses, compounds eluting prior to 358 s (acetone retention time) and siloxanes (i.e., chromatographic artifacts) were removed from the peak table. Peaks that were present in only one of the three biological replicates were imputed to zero for that sample, while missing values for peaks that were present in two out of three biological replicates were imputed to half of the minimum value across all biological replicates. The relative abundances of compounds across chromatograms were normalized using probabilistic quotient normalization (PQN) (45) in R version 3.4.3. The data were log 10 transformed, intraclass correlation coefficients (ICCs) were calculated using R ICC package version 2.3.0, and peaks with an ICC of ,0.75 were not further processed. Analytes were retained for further analysis if they were at least 2-fold more abundant in any Coccidioides culture than in the medium controls and were considered to have been detected in that culture. Principal-component analysis was performed using R factoextra package version 1.0.5 with the biological replicates as observations and the absolute peak intensities (mean centered and scaled to unit variance) as variables. The relatedness of samples based on their volatile metabolomes was assessed using hierarchical clustering analysis on Pearson's correlation between isolates and the Euclidean distance between volatiles using R pheatmap package version 1.0.12. Geometric means of the biological replicates were calculated and used to determine the statistical difference in abundances between life cycles or species using a Mann-Whitney U test and Benjamini-Hochberg false discovery rate correction with an a value of 0.05, using R stats package version 3.5.3. The relative abundance of volatiles significantly different between life cycles was calculated by dividing the mean mycelial peak abundance by the mean spherule peak abundance.
Data availability. Metabolomic data (chemical feature peak areas and retention time information) included in this study are available at the NIH Common Fund's National Metabolomics Data Repository (NMDR) website, the Metabolomics Workbench, at www.metabolomicsworkbench.org, where it has been assigned project identifier PR0001064 and study identifier ST001659 (https://doi.org/10.21228/ M85H6W).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.