Characterization of the fungal genus Sphaerellopsis associated with rust fungi: species diversity, host-specificity, biogeography, and in-vitro mycoparasitic events of S. macroconidialis on the southern corn rust, Puccinia polysora

Abstract

Sphaerellopsis species are putative hyperparasites of rust fungi and may be promising biological control agents (BCA) of rust diseases. However, few detailed studies limit potential BCA development in Sphaerellopsis. Here, we explored the biogeography, host-specificity, and species diversity of Sphaerellopsis and examined the early infection stage of one species, S. macroconidialis, to infer its trophic status. We randomly screened 5,621 rust specimens spanning 99 genera at the Arthur Fungarium for the presence of Sphaerellopsis. We identified 199 rust specimens infected with Sphaerellopsis species on which we conducted morphological and multi-locus phylogenetic analyses. Five Sphaerellopsis species were recovered, infecting a total of 122 rust species in 18 genera from 34 countries. Sphaerellopsis melampsorinearum sp. nov. is described as a new species based on molecular phylogenetic data and morphological features of the sexual and asexual morphs. Sphaerellopsis paraphysata was the most commonly encountered species, found on 77 rust specimens, followed by Sphaerellopsis macroconidialis on 56 and S. melampsorinearum on 55 examined specimens. The type species, Sphaerellopsis filum, was found on 12 rust specimens and Sphaerellopsis hakeae on a single specimen. We also recovered and documented for the first time, the sexual morph of S. macroconidialis, from a specimen collected in Brazil. Our data indicate that Sphaerellopsis species are not host specific and furthermore that most species are cosmopolitan in distribution. However, S. paraphysata is more abundant in the tropics, and S. hakeae may be restricted to Australia. Finally, we confirm the mycoparasitic strategy of S. macroconidialis through in-vitro interaction tests with the urediniospores of Puccinia polysora. Shortly after germination, hyphae of S. macroconidialis began growing along the germ tubes of P. polysora and coiling around them. After 12 days of co-cultivation, turgor loss was evident in the germ tubes of P. polysora, and appressorium-like structures had formed on urediniospores. The interaction studies indicate that Sphaerellopsis species may be more effective as a BCA during the initial stages of rust establishment.

Introduction

Sphaerellopsis Cooke (Leptosphaeriaceae, Ascomycota) is the most commonly reported fungal genus associated with rust fungi (Pucciniales, Basidiomycota). Sphaerellopsis species have been reported on 369 rust species and 30 genera in more than 50 countries across the globe (Kranz and Brandenburger 1981). The fungus is usually described as solitary to gregarious spherical black pycnidia that develop on sori and, thus, presumably infect rust spores and prevent their dispersion (Eriksson 1966). These black pycnidia are typically found on uredinia, the spore stage most frequently associated with severe rust disease epidemics and long-distance dispersal across continents. However, pycnidia have also been found in association with other rust spore stages. Due to the intimate association of Sphaerellopsis with rust fungi, this fungal genus is tentatively considered a potential biological control agent (BCA) of rust fungi, many of which cause devastating disease epidemics and yield losses worldwide (Chen et al. 2002; Kolmer et al. 2009; Lidwell-Durnin and Lapthorn 2020). Nevertheless, characterization studies of Sphaerellopsis are scarce, which limits its use in applied biological control research.

In 1815, the type species Sphaerellopsis filum was initially described as Sphaeria filum by Bivona-Bernadi on rusts infecting Convolvulus sepium and Populus nigra in Sicily (Bivona-Bernardi 1815). Fries transferred the species to Phoma as Phoma filum in 1823 (Fries 1823). Later, Castagne erected the genus Darluca and treated Sphaeria filum as a synonym of Darluca vagans (Castagne 1851). However, in 1966 Eriksson considered the epithet “vagans” superfluous and prioritized “filum” over “vagans” (Eriksson 1966). In 1908, Spegazzini considered Eudarluca caricis the teleomorph of Darluca filum (Spegazzini 1908). Later in 1951, Keener proved the connection between these two genera experimentally (Keener 1951). Yuan et al. 1998 confirmed this connection by obtaining the asexual morph from the teleomorph in culture studies (Yuan et al. 1998). Lastly, in 1977, Sutton transferred Darluca filum to the genus Sphaerellopsis as S. filum (Sutton 1977). Species of Sphaerellopsis are commonly found in its asexual state, and the sexual morph is rarely observed. Although Sphaerellopsis and Eudarluca are now known to be congeneric, it is still uncertain which Sphaerellopsis species is conspecific with Eudarluca caricis.

Although most scientific publications posit Sphaerellopsis as a mycoparasite of rust fungi, its relationship with these plant pathogens is still poorly understood. While there is some evidence of direct interaction between S. filum and several rust species, the nature of the interaction has not been consistently described and may vary among S. filum strains. For instance, some researchers argue that S. filum can colonize rust spores by penetrating nonspecialized hyphae and disrupting cytoplasm (Carling, D.E. Brown, M.F. Millikan, 1976; Płachecka 2005; Sappin-Trouffy 1896; Whelan et al. 1997). However, other studies report no evident cytoplasmic disruptions of rust spores when S. filum is present (D’Oliveira, 1941; Hulea 1939). In vitro assays demonstrated hyphal growth and conidioma development of S. filum when cultured with intact or ruptured rust spores (Rambo and Bean 1970). However, changes in fungal growth rate do not necessarily demonstrate that S. filum can infect rust fungi. Similarly, lab and field experiments have shown a significant reduction in rust infection when S. filum is present (Black 2012; Gordon and Pfender 2012; Yuan and Han 2000), but these conclusions are contradicted by other studies (Yuan et al. 1999). In recent years, through phylogenetic analyses, several isolates determined as  S. filum have turned out to be incorrectly placed in the genus Sphaerellopsis (Trakunyingcharoen et al. 2014). Thus, new genera were created, and new species within Sphaerellopsis were introduced. Hence, the previous interaction tests of S. filum with rust fungi remain unanswered, as the Sphaerellopsis specimens used in those studies may represent different species, or even belong to other genera.

Based on morphology and DNA sequence data, there are currently seven accepted Sphaerellopsis species. Five species are reported as mycoparasites of rust fungi: Sphaerellopsis anomala, S. filum, S. hakeae, S. macroconidialis, and S. paraphysata (Crous et al. 2016; Nag Raj 1993; Trakunyingcharoen et al. 2014); and two are considered saprobic: S. artemisiae and S. isthmospora (Doilom et al. 2021; Phookamsak et al. 2019). Although S. hakeae and S. paraphysata were reported to be associated with rust sori and plant tissue (Crous et al. 2016, 2018), it is unclear if the association with the host plant is parasitic or saprobic. Furthermore, it has not been proven that all Sphaerellopsis species associated with rust fungi are mycoparasites. Therefore, while significant improvements have been made to the taxonomy of Sphaerellopsis (Trakunyingcharoen et al. 2014), the parasitic relationship between its members with rust fungi remain undetermined.

Among the five Sphaerellopsis species known to associate with Pucciniales, S. paraphysata is the only one confirmed to have a mycoparasitic strategy. Secondary metabolites obtained from S. paraphysata disrupted the urediniospores cell wall of Puccinia substriata, leading to cellular component leakage (Ashmitha Sri et al. 2020). Inoculation of the conidia of S. paraphysata on the uredinia of P. substriata reduced rust spore germination by up to 76% (Anandakumar et al. 2019). In addition, the rust disease severity of the rust was 13% when S. paraphysata was present compared to the control of 86% (Anandakumar et al. 2019). However, because the species was also found in plant tissue (Crous et al. 2018), further studies are needed to discard a plant pathogenic strategy for S. paraphysata which would likely negate its application as a potential BCA for rust fungi.

In addition to interaction studies between members of Sphaerellopsis and rust fungi, other ecological studies are essential to characterize the genus and determine if any species could be suitable as a BCA of rusts. For example, knowledge of a natural enemy’s host range and geographic distribution is crucial for environmental risk assessments to prevent releasing new diseases. Furthermore, host-specificity studies can help clarify whether/which Sphaerellopsis species are generalists or host-specific on rust species or genera. Nevertheless, biogeography, species diversity, and the host-specificity of Sphaerellopsis are unknown due to the few records of the currently accepted species. Most of these records are primarily from temperate regions (Ashmitha Sri et al. 2020; Crous et al. 2016; Trakunyingcharoen et al. 2014) with limited records from the tropics.

Because most Sphaerellopsis species are associated with rust fungi, we adopted a strategy of screening vouchered rust specimens for the presence of incidentally co-collected Sphaerellopsis species. The Arthur Fungarium (PUR), housed at Purdue University, is one of the world’s largest collections of rust fungi. It holds approximately 160,000 specimens of 5,000 species collected across a broad geographic distribution and timeline, and it is one of the most diverse collections, with 132 rust genera in 14 families in the world (Purdue Herbaria 2022) with especially rich holdings (ca. 50%) of specimens from the Americas. Therefore, the present study had two aims: (1) to augment distribution and host data on the fungal genus Sphaerellopsis by screening PUR collections, with an emphasis on those originating from the Americas, and evaluating these for signals of host-specificity, and (2) to elucidate the strategy of S. macroconidialis when interacting with rust fungi, using the urediniospores of the southern corn rust, caused by Puccinia polysora, as a model system.

Methods

Collected samples

We collected black fruiting bodies of Sphaerellopsis from preserved rust specimens in the Arthur Fungarium (PUR) supplemented with newly collected material in Peru in 2019 and Puerto Rico in 2018. Rust specimens at PUR are stored in folders sorted by rust species in host plant families and geographic regions. When collecting Sphaerellopsis samples at PUR, we randomly screened these rust specimens by selecting the top, middle, and bottom specimens from the Americas shelf in each rust species folder. When collecting Sphaerellopsis samples from other geographic regions, we randomly selected one rust specimen per folder. Although we screened rust specimens collected across the globe, the Americas was our preferred geographic region in an effort to close this information gap. We screened each rust-infected leaf of every rust specimen under a stereoscope Olympus Model SZ2-ILST (Tokyo, Japan) and screened for visible signs of Sphaerellopsis-type fruiting bodies developing on the sori. Only specimens fruiting exclusively on rust sori but not on surrounding host tissue were removed for further analyses, as the ability to also fruit on host plant tissues would indicate a non-rust-specific pathogen. Then, we removed one Sphaerellopsis-infected sorus with a sterile razor blade per rust specimen. A new blade was used per each specimen to prevent cross-contamination. Each infected sorus was placed in a microcentrifuge tube labeled with the PUR barcode of the rust specimen and a serial number.

When collecting Sphaerellopsis specimens in the field, we first collected rust-infected plant leaves. Then, we looked for black fruiting bodies developing on sori under a stereoscope. If Sphaerellopsis was present, we isolated it by cutting a piece of the plant tissue containing both the sorus and Sphaerellopsis with a razor blade. Then, the plant tissue was sterilized with 1/10 dilution chlorine bleach for one minute and washed three times with sterile water. The piece of plant tissue was inoculated onto Petri dishes containing potato dextrose agar (PDA) and 50 mg/mL chloramphenicol. Petri dishes were shipped to the Aime Lab at Purdue University for further processing. Once the Petri dishes arrived at the Aime lab, we subcultured them until axenic cultures were achieved on PDA and 2% malt extract agar (MEA) with 50 mg/mL chloramphenicol. Isolates were stored long-term on PDA slants at 4 °C and in 15% (v/v) glycerol at -80 °C. Finally, we pressed, dried, and vouchered the collected rust specimens at PUR.

In total, we screened 5,621 Pucciniales collections for the presence of Sphaerellopsis (Supplementary Table 1). The following data were recorded for each Pucciniales specimen that was found to be co-infected with Sphaerellopsis: PUR accession number, rust species name, country of origin, year of collection, host plant family, genus, and species, and geolocation (Supplementary Table 2). Finally, we took macro- and microphotographs of some of the collected Sphaerellopsis samples with an Olympus SC30 camera and image software Olympus cellSens entry version 1.14 under a stereoscope Olympus Model SZ2-ILST and a compound microscope Olympus BH2-RFCA at PUR. Measurements of fungal structures were made using cellSens Standard 1.18 Imaging Software (Olympus).

Identification and species concept

The collected Sphaerellopsis samples were identified using an integrated species concept, based on morphological characters and phylogenetic analyses (Aime et al. 2021). Original descriptions of the currently accepted Sphaerellopsis species were used as references for morphological comparison (Cooke M.C., 1883; Crous et al. 2016; Doilom et al. 2021; Nag Raj 1993; Phookamsak et al. 2019; Trakunyingcharoen et al. 2014).

DNA isolation and PCR amplification

The genomic DNA of each potential Sphaerellopsis, collected during the screening, was extracted using the EZNA HP Fungal DNA kit (Omega Bio-Tek, Norcross, Georgia), following the manufacturer’s instructions and modifying only the incubation time in the third step. Instead of 30 min, we incubated the samples overnight to ensure complete lysis of cells in the suspension. We selected the following loci for amplification: the internal transcribed spacer (ITS) and the large subunit (LSU) of the ribosomal DNA repeat, the translation elongation factor 1-α (tef1) and the RNA polymerase II second largest subunit (rpb2). Because most of our Sphaerellopsis specimens were derived from fungarium collections and thus culturing was not possible, we designed specific ITS and LSU primers for amplification of these loci (Table 1). For this, we downloaded all ITS and LSU sequences of verified Sphaerellopsis species (Trakunyingcharoen et al. 2014) from GenBank. We also downloaded sequences of several rust species and of ubiquitous fungal species usually found in dead plant material. Multiple alignments were conducted using MUSCLE version 3.7 (Edgar 2004) in MEGA7 (Kumar et al. 2016). Conserved regions were searched for both loci in Sphaerellopsis sequences, excluding rusts and other fungal sequences. We selected primers that amplify approximately 250 bp in length for ITS amplification and between 600 and 700 bp for LSU amplification. Finally, we performed a BLASTn database search using our selected primers as the query to confirm that the greatest matching hits were Sphaerellopsis sequences. Amplification for each locus was conducted with these new, and previously published (Table 1) primers in 25-µl PCR reactions on a Mastercycler ep gradient Thermal Cycler (Eppendorf model #5341, Hauppauge, New York) that consisted of 12.5 µl of 2× MyTaq Mix (Bioline, Swedesboro, New Jersey), 1.25 µl of each 10 µM primer, and 10 µl of either 1/10 or 1/5 diluted DNA extract. Amplifications of rDNA, tef1 and rpb2 loci were run under the following conditions: initial denaturation at 94 °C for 5 min (95 °C for rpb2/96°C for 2 min for tef1); followed by 40 cycles of denaturation (45 cycles for ITS) at 94 °C for 30 Sect. (95 °C for rpb2), annealing at 51.8 °C for 45 s for ITS/54°C for 45 s for LSU/56°C for 30 s for tef1/55°C for 45 s for rpb2, and elongation at 72 °C for 45 s (1 min for LSU and 30 s for tef1); and final extension at 72 °C for 7 min.

Table 1 Primers for PCR amplification and sequencing used in this study

Electrophoresis and sequencing

We ran the PCR products in 1% agarose and stained them with GelRed (RGB4102, Phoenix Research Products) for 35 min at 110 V in a Bio-Rad electrophoresis tank to visualize PCR products. PCR products of samples that showed bands were sent to Genewiz (South Plainfield, New Jersey) for purification and subsequent sequencing in both directions with the amplification primers (Table 1). Raw sequence reads were edited manually and assembled using Sequencher version 5.2.3 (Gene Codes Co., Ann Arbor, Michigan).

Sequence alignment and phylogenetic trees

The edited sequences were blasted against the NCBI GenBank nucleotide database (http://ncbi.nlm.nih.gov/blast/Blast.cgi) to confirm placement in Sphaerellopsis. To construct datasets, we downloaded publicly available DNA sequences of Sphaerellopsis species as references for our phylogenetic analyses; Alternaria consortialis was chosen as the outgroup (Table 2). Sequences were aligned using MUSCLE version 3.7 (Edgar 2004) in MEGA7 (Kumar et al. 2016). Then, the aligned sequences were trimmed using trimAl version 1.2 (Capella-Gutiérrez et al. 2009) with a minimum percentage of positions to conserve [0-100]: 50; and gap threshold, the fraction of positions without gaps in a column [0–1]: 0.6. We performed maximum likelihood (ML) inference using IQ-TREE (Minh et al. 2020) under partitioned models (Chernomor et al. 2016) and selected the best nucleotide substitution model under Akaike’s information criterion corrected for small sample size (AICc) using ModelFinder (Kalyaanamoorthy et al. 2017). An ultrafast bootstrap analysis was implemented with 1,000 replicates (Hoang et al. 2018). The “-bnni” option was used to reduce the risk of overestimating branch supports with UFBoot due to severe model violations. Finally, phylogenetic reconstructions with bootstrap values were visualized in FigTree version 1.4.3 (http://tree.bio.ed.ac.uk/software/figtree/) and colored in Inkscape (https://inkscape.org).

Table 2 Reference sequences used in phylogenetic analyses

Geographical distribution

The localities of Sphaerellopsis specimens with successfully amplified gene regions were used to build a geographic map. We used the GPS coordinates of each of these specimens when present. Otherwise, we generated approximated coordinates according to the locality description following a geocoding Python Script in the GitHub repository (Lynn 2017). We plotted the geographic data of each specimen on a map and colored each point by the clades formed in the multi-locus phylogenetic tree using the package Geopandas in Python (Jordahl 2014).

Interaction experiments between conidia of Sphaerellopsis macroconidialis and urediniospores of Puccinia polysora

Puccinia polysora was the host from which the strain SP28 of S. macroconidialis, used in this study, was collected. Puccinia polysora is an agriculturally important fungus that causes the destructive disease Southern rust of corn (Sun et al. 2021).

Collection and identification of urediniospores of Puccinia polysora from maize crops

In the summer of 2021, maize leaves infected with P. polysora were harvested from field-grown maize plants at the Southwest Purdue Agricultural Center, Indiana, USA, and brought to the Aime Lab. Urediniospores were collected using a mini cyclone spore collector (Tallgrass Solutions, INC; Manhattan, KS) and stored in gelatin capsules at -80 °C until further use. To confirm the identification of the rust, we amplified the LSU region using the primers of Aime (2006) and the methodologies of Aime et al. (2018) and Aime and McTaggart (2020). We amplified the LSU as it has been shown to be the most informative gene for rust species identification (Aime et al. 2017). The resulting DNA fragment was blasted against the NCBI and the Rust HUBB (Kaishain et al. 2024) databases to confirm identity.

Cultivation of corn plants in the greenhouse and installation of humidity chamber for inoculations

Healthy corn plants (P0574AM™) were cultivated in the greenhouse facility at Lily Hall of Life Sciences, Purdue University. We planted seven 3-gallon pots with two corn seeds per pot. Following germination, we removed the weaker seedling leaving one plant per pot. Plants were maintained at a temperature range between 24 and 30 °C and watered and fertilized as needed. Next, we installed a humidity chamber for rust inoculation in the same room where plants were growing. This chamber consisted of a simple cubic structure (30 cm³) made of PVC pipes and covered with a white four mil plastic sheeting. A door was installed on the chamber for easy access and manipulation of the corn plants once these were inside. A PVC pipe (2 cm diam. and 20 cm length) was also inserted in one side of the chamber to connect a 2.2 L humidifier (AquaOasis™) placed outside the chamber. Finally, we placed a hygrometer inside the chamber to track temperature and humidity.

Rehydration of urediniospores of P. polysora for inoculation

Before inoculating healthy corn plants with P. polysora, we took the urediniospores stored at -80 °C in gel capsules and rehydrated them in two steps. First, the spores contained in gel capsules were thawed at 4 °C for 16 h. Then, the urediniospores contained in gel capsules were placed inside a humidified chamber. This chamber consisted of a sterile plastic container with a 23.5% KOH beaker as a source of water vapor. This concentration of KOH gives approx. 80% of relative humidity inside the container while avoiding water condensation (Rowell 1984). Then, we sealed the chamber with a lid and let the urediniospores rehydrate for 12 h at room temperature. Once urediniospores were rehydrated, we added them into a sterile glass vial containing 0.1% tween 20. We gently mixed the spores with the solution to resuspend them and ensure no clumps were formed.

Inoculation of corn plants with urediniospores of P. polysora in the greenhouse

We used a spore inoculator (Tallgrass Solutions, Manhattan, Kansas) attached to an air compressor (California Air Tools CAT-1P060S) operating in the 2–5 psi range to inoculate healthy corn plants with urediniospores immersed in 0.1% tween 20. Each healthy corn leaf was sprayed with the spore solution at 2 cm from the leaf. Once each plant was covered entirely with the spore solution, we placed them in the humidity chamber and did not close the chamber completely to ensure air circulation. Inoculations were done in the late afternoon when temperatures were lower, which helped moisture stay longer on the leaf surface and facilitated spore germination for successful infection (Borlaug Global Rust 2017). During the infection period, temperatures were held at between 23 and 30 °C and the humidifier was continuously filled with sterile distilled water to keep relative humidity between 50 and 80%. We used 16 daylight hours and eight night hours. Under optimal conditions, we observed rust symptoms on corn leaves between 7 and 15 days after inoculation.

Harvesting of fresh conidia of S. macroconidialis and urediniospores of P. polysora for the in-vitro interaction test

Conidia of S. macroconidialis SP28 from a two-week-old PDA culture were harvested for the interaction test. We added 1 mL of sterile water to the medium, then slightly agitated the petri dish to let the water mix with the conidia for about a minute. Once the water turned milky from presence of suspended conidia, we collected the conidia solution with a micropipette and transferred it to a 2mL tube. Fresh urediniospores of P. polysora infecting corn plants in the greenhouse were also harvested for the interaction test. We collected urediniospores from open and pulverulent sori to ensure the urediniospores were mature and ready to germinate. We gently tapped the rust-infected leaf against a 2mL tube containing 0.1% tween 20 to allow the spores to fall into it. Once the tween 20 solution turned light brown, we closed the lid. The concentration of conidia and urediniospores was measured with a hemocytometer to reach a dilution of 10⁴ spores per mL. The viability of the conidia and urediniospores was checked with Trypan Blue. We used ≥ 80% viable spores as the threshold for the interaction test.

In-vitro interaction between spores of S. macroconidialis and P. polysora

We poured 1 mL of 1% water agar with 50 mg/mL Chloramphenicol into small Petri dishes (50 mm diam.) to set up the interaction test. Then, we added 40 uL of the urediniospore suspension to five Petri dishes. To locate the urediniospores during the interaction test, we drew two points on each side of the bottom of each petri dish with a marker. Each Petri dish was sealed with parafilm and incubated in the dark at 25°C overnight to facilitate urediniospore germination. We observed each petri dish under a compound microscope Olympus BH2-RFCA using a 20X objective the day following inoculation. Petri dishes in which > 70% of the urediniospores germinated were kept for the next step. A minimum of three Petri dishes with > 70% urediniospore germination were used as replicates for the interaction test. Then, we added the conidia of Sphaerellopsis, suspended in water, to the same Petri dishes containing germinated urediniospores at a 1 mm distance from the urediniospores. Petri dishes were sealed again with parafilm and incubated for 24 h at room temperature. After 24 h of co-inoculation, we conducted daily screenings of the plates over the next 12 days. Interactions were observed under a microscope using the 20X objective without opening the lid to avoid contamination. Lids were only removed on the last day of observation to use a 40x objective and to take final pictures. Two negative controls were also used in this interaction test. The first one consisted of three Petri dishes containing urediniospores only. The second consisted of three Petri dishes containing only conidia of S. macroconidialis. The experiment was repeated three times.

Results

Sphaerellopsis recovery from PUR collections

We randomly screened 5,621 rust specimens in 99 rust genera, representing 5% of the total collections at PUR and 58% of the accepted rust genera (Berndt & Aime, n.d. unpublished), for the presence of Sphaerellopsis species that were incidentally co-collected with rust specimens (Supplementary Table 1). Of these 5,621 specimens, we collected 523 black fruiting bodies resembling the fungal genus Sphaerellopsis (Supplementary Table 2). Of these 523 collections, 199 were confirmed as Sphaerellopsis members through phylogenetic analyses and morphology (Table 3). Five Sphaerellopsis species were recovered, infecting 122 rust species in 18 genera from 34 countries.

Table 3 Sphaerellopsis members associated with rust fungi from PUR and identified through molecular and morphological analyses. Reference sequences in bold; NA: data not available

We successfully amplified 195 DNA sequences of Sphaerellopsis from ITS rDNA, 58 sequences from LSU rDNA, 48 from tef1, and eight from rpb2. Although we amplified the four loci for some Sphaerellopsis specimens, degradation of DNA in older specimens limited the ability to obtain complete locus datasets for many specimens. Nevertheless, we successfully amplified the ITS region of 163 specimens collected between 1883 and 1998. The oldest Sphaerellopsis specimen whose ITS region was successfully amplified was collected in 1883 on Melampsora medusae from the United States (voucher number: PUR2041, GenBank accession number: OQ418220). Lastly, we isolated a strain of S. macroconidialis (SP28) from freshly collected material, which was the basis of the interaction experiments between the conidia of S. macroconidialis and germinated urediniospores of Puccinia polysora, the host from which it was collected.

Nucleotide alignment dataset and phylogenetic inferences

Our multi-locus phylogenetic analysis consisted of a four-locus-concatenated dataset of 1996 characters, of which 352 were parsimony-informative. The percentage of parsimony-informative characters per gene region was 3.4% for ITS, 1.3% for LSU, 7.1% for tef1, and 5.8% for rpb2. We analyzed 219 individuals, of which 16 were sequences from previously identified Sphaerellopsis taxa, and Alternaria consortialis served as an outgroup taxon (Table 3). The following models were selected by ModelFinder (AICc): JC for ITS, GTR + F + R2 for LSU, TIM + F + G4 for tef1, and TIM2e + I for rpb2. Our maximum likelihood analysis revealed eight supported clades (Fig. 1, supplementary Figure S1), all of which have bootstrap support ≥ 70%.

Fig. 1

The top-scoring ML phylogenetic tree of the genus Sphaerellopsis reconstructed from the four-locus-concatenated dataset (ITS, LSU, tef1, and rpb2). The ML bootstrap value is presented above each branch. Colors delimit clades, each labeled with the corresponding Sphaerellopsis species. Taxa labels are written on the tree as “PUR voucher,” “the host rust where Sphaerellopsis was found,” and “the origin/locality of each specimen.” Reference sequences and outgroup taxa are written in bold. The tree was rooted to Alternaria consortialis CBS 104.31. Refer to supplementary figure S1. to see the fully resolved phylogram showing branch lengths and support values

Species Diversity of Sphaerellopsis associated with rust fungi

Five species of Sphaerellopsis were recovered from our sampling: four of the seven previously accepted species, and one undescribed species (Fig. 1). Sphaerellopsis paraphysata was the most common species within our screened collections, found on 77 rust specimens, followed by S. macroconidialis on 56 and S. filum on 12. Sphaerellopsis hakeae was found in one rust specimen, and S. artemisiae and S. isthmospora were not found in this study. One species, Sphaerellopsis anomala is not represented in our analyses, due to lack of sequence data. Two other well-supported clades were found in the phylogeny that do not represent previously published Sphaerellopsis species. One of these consisted of a single specimen found on sori of Puccinia montanensis from the United States collected in 1896 (voucher number: PUR23925). We amplified the ITS region of this specimen and took macro photographs. However, due to the scarce and dry material, the specimen’s morphology and amplification of other gene regions were impossible. Thus, it is uncertain whether this represents an undescribed species. The other clade comprised 55 specimens from Pucciniales collected between 1883 and 2016, including a specimen that contained both the asexual and sexual morphs. This new species is described as Sphaerellopsis melampsorinearum sp. nov. below.

Taxonomy

Sphaerellopsis melampsorinearum Gomez-Zap. & Aime, sp. nov.

Figure 2.

Fig. 2

Sphaerellopsis melampsorinearum sp. nov. (PUL F29362, PUL 29,360, PUL F29361). A, B Conidiomata developed on sori. C, D Ascomata developed on sori. E Outer layers of conidioma, textura angularis. F, G Conidia. H Vertical section through ascomata. I Conidiogenous cells. J, K Asci and pseudoparaphyses. K, L Ascospores. Bars: a–d = 100 μm, e–g = 10 μm, h = 20 μm, i–l = 10 μm

Mycobank No: MB847464.

Etymology: Named after the large number of rust hosts that belong to the suborder Melampsorineae.

Diagnosis: Similar to S. filum but differs in conidiomata size (up to 107 μm diam.), and conidia length [(9.1–)10.3–14.3(–16.6) µm].

Type: Holotype: United States, Indiana, Tippecanoe County, West Lafayette, on urediniospores of Melampsora medusae infecting Populus deltoides, 19 September 2015, M. Catherine Aime, s.n. (PUL F29362 (ex-PURN15307); GenBank accessions ITS–OQ418354, LSU–OQ418183, tef1–OQ743726, rpb2–OQ587607).

Description: Asexual morph–conidiomata associated with rust sori, pycnidial, erumpent, aggregated, globose, 48–107 μm, with central ostiole, outer layers dark brown cells textura angularis, 3.8–6.92 μm diam. Paraphyses not observed. Conidiophores reduced to conidiogenous cells. Conidiogenous cells line the inner cavity and are smooth, hyaline, globose to ampulliform. Conidia fusoid, hyaline, smooth, guttulate, 1-septate, slightly constricted at the septum, apex subobtuse, tapering to truncate hilum, (9.1–)10.3–14.3(–16.6) × (3–5) µm. Sexual morph–ascomata associated with rust sori, 76–162 μm diam., solitary or gregarious; loci immersed, brown in outer zone consisting of two to three rows of dark cells, hyaline in inner part, subglobose to ampulliform, with protruding papillate neck and ostiole. Pseudoparaphyses are filiform, septate, hyaline. Asci numerous, 8-spored, bitunicate, cylindrical-clavate, short stipitate, 59.6–101.3 × 8.5–10.4 μm. Ascospores irregularly biseriate, fusiform, hyaline to pale yellow, 15.2–21.3 × 3.8–6.0, 1-septate, slightly constricted at the septum, surrounded by a mucous sheath not easily perceived.

Substrate/Host: on rust sori of several rust species, principally species of the genus Melampsora, but also known to infect Coleosporium spp., Puccinia spp., Uromyces spp., and Gerwasia holwayi.

Distribution: Argentina, Bolivia, Brazil, Colombia, Ecuador, Germany, Japan, Peru, Trinidad, the continental United States of America, Venezuela.

Additional materials examined: Brazil, São Paulo, on urediniospores of Puccinia vernoniae-mollis infecting leaves of Vernonia sp., 17 February 1989, Anibal de Carvalho 89 − 7, containing teleomorph, (PUL F29361 (ex-PURN9763); GenBank accessions: ITS–OQ418405; LSU– OQ418210). Colombia, Antioquia, on urediniospores of Melampsora larici-populina infecting Populus nigra, 20 March 1989, V.M Pardo-Cardona s.n. (PUL F29359 (ex-PURN4015); GenBank accessions: ITS–OQ418383; LSU–OQ418199). Peru, Ucayali, on urediniospores of Uromyces yurimaguasensis, 22 October 2016, M. Catherine Aime MCA6471. (PUL F29363 (ex-PURN16392); GenBank accessions: ITS–OQ418361; LSU–OQ418189). United States of America, Illinois, on urediniospores of Melampsora sp. infecting Populus sp., 22 September 2012, M. Catherine Aime MCA5030 (PUL 29,360 (ex-PURN6730); GenBank accessions: ITS– OQ418402, LSU–OQ418209, tef1–OQ743742); Georgia, on urediniospores of Coleosporium helianthi infecting Silphium compositum, 24 August 1977, Yoshitaka Ono, John McCain & Joe F. Hennen 10,185 (PUL F29357 (ex-PUR88233); GenBank accession: ITS–OQ418265).

Notes. The conidiomata and length of conidia of S. melampsorinearum are smaller than for any other described Sphaerellopsis species. However, the width of the conidia of S. melampsorinearum is similar to S. anomala, S. filum, and S. macroconidialis. Sphaerellopsis melampsorinearum is distributed worldwide and infects a range of rust species in the Pucciniaceae, Phragmidiaceae, Melampsoraceae, and Coleosporiaceae. However, 41 out of 55 hosts rust hosts of S. melampsorinearum belong to suborder Melampsorineae.

The sexual morph of S. macroconidialis

Sphaerellopsis macroconidialis is known from the asexual morph, and no sexual morph has been described. However, in this study, we recovered one specimen containing the sexual morph of S. macroconidialis (Fig. 3). This specimen was found in Brazil, Rio de Janeiro, associated with telia of Puccinia wedeliicola infecting the host plant Wedelia trichostephia, collected on 7 May 1922, by E.W.D Holway, #1822, (PUL F29358 (ex-PURF8347)) (Fig. 3). The ITS sequence obtained from this specimen shared 100% identity (239/239 no gaps) with S. macroconidialis CBS 233.51 (GenBank Accession No. MH856836.1). Morphology of the sexual morph is as follows: ascomata developing on rust sori, up to 123 μm diam., brown in outer zone, cells textura parenchymatic, hyaline in inner part, erumpent, gregarious; loci subglobose to ampulliform. Pseudoparaphyses are filiform, septate, hyaline. Asci are numerous, 8-spored, bitunicate, cylindrical-clavate, short stipitate, 68.2–106.7 × 7.3–11.1 μm. Ascospores are irregularly biseriate, fusiform, hyaline to pale yellow, 17.2–23 × 4.8–6.0, 1-septate, slightly constricted at the septum, surrounded by a hyaline mucous sheath not easily perceived.

Fig. 3

Teleomorph of S. macroconidialis (PUL F29358). A Ascomata. B Vertical section through ascomata. C Ascospore. D Asci and pseudoparaphyses. E Asci and ascospores. Bars: a = 200 μm, b, c = 20 μm, d, e = 10 μm

In-vitro interaction test between S. macroconidialis and P. polysora 

The interaction test of this study confirms the mycoparasitic strategy of S. macroconidialis on rust fungi. Five days after the co-cultivation, we observed hyphae of S. macroconidialis growing along the germ tubes of P. polysora and coiling around them (Fig.4).

Fig. 4

Light micrographs of Sphaerellopsis macroconidialis interacting with germinated urediniospores of Puccinia polysora in-vitro. A─B Day one after co-cultivation. A Red arrows point to the urediniospore and its germ tube, and the green arrow points to S. macroconidialisS. macroconidialis hypha. B Black arrows point to the first contact. C Negative control, hyphae of S. macroconidialis alone on day 12. D─E Day five after co-cultivation. Hyphae of S. macroconidialis form coils and tightly encircle germ tubes of P. polysora. Black arrows point to dense coils. F Negative control, urediniospores, and germ tubes alone on day 12. G─J Day 12 after co-cultivation. G Dense coils around a germ tube of P. polysora. H An appressorium (arrow) attached to the urediniospore. I and J Loss of turgor of germ tube of P. polysora. Scale bars: A-B, D-J = 20 μm, C = 50 μm

The germ tubes of P. polysora measured 6.5 μm in diam., while those of S. macroconidialis measured 1.8 μm in diam., making them easy to distinguish. During the first day after co-cultivation, we observed the first contact between germinated conidia of S. macroconidialis and germ tubes of urediniospores of P. polysora. Then, during the next four days, hyphae of S. macroconidialis started to grow over the urediniospores and their germ tubes, but without clear evidence of antagonism. However, on the fifth day of co-cultivation, S. macroconidialis began coiling around rust germ hyphae. Coils tightly encircled the germ tubes. However, the cell wall of the germ tubes was not disrupted. Such coils were not seen on S. macroconidialis hyphae inoculated alone. During the next six days, we did not notice any new sign of mycoparasitic mechanism against P. polysora. Nonetheless, on day 12, we noticed the formation of an appressorium attached to a urediniospore and turgor loss of a few germ tubes already coiled by S. macroconidialis. Loss of turgor was not seen on germinated urediniospores inoculated alone. After 12 days of observations, S. macroconidialis hyphae grew abundantly, and no other antagonistic events could be observed.

Discussion

Characterization studies of fungi with potential as BCAs are essential to the development of applied microbial biocontrol of plant diseases. Although the fungal genus Sphaerellopsis is commonly considered a rust mycoparasite due to its association with several rust species, studies of this genus are scarce, and its biocontrol potential is unknown. To evaluate Sphaerellopsis as a candidate BCA, we screened thousands of rust collections for the presence of Sphaerellopsis (Supp. Table 1). We generated sequence data for nearly 200 Sphaerellopsis specimens found on rust fungi collections at four loci, including the ITS, which has previously been shown as a good barcoding region for Sphaerellopsis species (Trakunyingcharoen et al. 2014) and three other loci—LSU, rpb2, and tef1 for phylogenetic resolution (Fig. 1). We then use these data to characterize various aspects of Sphaerellopsis biology including species diversity, geographic distribution, and host specificity. Finally, we examined the interactions between S. macroconidialis and Puccinia polysora to infer initial infection strategies. These results can help determine the suitability of the application based on the BCA’s location and mode of action.

Sphaerellopsis species frequencies

Sphaerellopsis macroconidialis and S. paraphysata were the most common species associated with rust fungi in this study. S. macroconidialis was found on species in ten rust genera, and S. paraphysata on 12 rust genera (Fig. 1). Contrary to expectations, the type species, S. filum, previously reported from 30 rust genera and 369 rust species (Kranz and Brandenburger 1981), was not frequently collected. We found S. filum associated with only three rust genera: Melampsora, Puccinia, and Tranzschelia. As prior studies have shown, other species of Sphaerellopsis were frequently misidentified as S. filum in the past (Trakunyingcharoen et al. 2014), which could explain the discrepancy.

Host-specificity of  Pucciniales-infecting Sphaerellopsis species

Prior studies have found host-specificity in species of Sphaerellopsis (Liesebach and Zaspel 2004), Nischwitz et al. 2005), Kajamuhan et al. 2015). In contrast, our study does not show any signature of host-specificity for the Sphaerellopsis species analyzed (Fig. 1). For example, S. macroconidialis was found to be associated with species from several genera across the rust tree of life (Aime and McTaggart 2020) including Chaconia, Phakopsora, Phragmidium, Puccinia, Ravenelia, and Uropyxis, among others (Fig. 1). Similarly, S. paraphysata was associated with species from multiple rust genera including Crossopsora, Kweilingia, Melampsora, Mikronegeria, Phakopsora, Puccinia, Sorataea, and Uromyces, among others. Sphaerellopsis filum was recovered infecting hosts from three suborders of Pucciniales; and S. melampsorinearum was found on four families of Pucciniales, with the majority of hosts within the subphylum Melampsorineae.

Differences between our and previous work are likely due to limited sampling in prior studies, which only examined Sphaerellopsis species associated with Puccinia species on grass hosts and Melampsora species on poplars. The dataset of Liesebach and Zaspel (2004) and Nischwitz et al. (2005) did not exceed 20 isolates, and the sampling of Kajamuhan et al. (2015) comprised 82 isolates collected from Puccinia species. In contrast, our dataset covered 19 rust genera and 216 specimens. Although S. paraphysata and S. macroconidialis are predominantly associated with Puccinia specimens and S. melampsorinearum with Melampsora specimens, both are also associated with rust species from other genera. Likewise, our data do not show preference of Sphaerellopsis species for rusts at even the family rank. For example, S. paraphysata included rust hosts in the families Melampsoraceae, Phakopsoraceae, and Pucciniaceae, which span three different rust subphyla.

Interestingly, we found Sphaerellopsis species infecting rusts on many economically important hosts such as maize, wheat, and poplars. However, we did not find any Sphaerellopsis infecting Hemileia vastatrix, the causal agent of coffee leaf rust, despite examination of 42 specimens of this rust collected from throughout its range. Nor does Hemileia vastatrix appear on prior lists of Sphaerellopsis rust hosts (Kranz and Brandenburger 1981). Keener (1934) suggested that a possible limiting factor of Sphaerellopsis infection could be the type of sorus produced. Hemileia species, for example, form suprastomatal sori that protrude through the stoma like a “bouquet” (McCain 1983) and do not tear the epidermis of the host plant. In addition to Hemileia, we also screened other rust specimens of the family Zaghouaniaceae that form suprastomatal sori; all were also free of Sphaerellopsis (Supplementary Table 1). Only one specimen of Mikronegeria fagi was found with associated black fruiting bodies resembling Sphaerellopsis. However, due to the scarcity and age of this particular specimen, we were unable to confirm it as a species of Sphaerellopsis. Thus, it remains inconclusive, but likely, that Sphaerellopsis species are restricted to infecting hosts that do not form suprastomatal sori.

Geographic distribution

This study included Sphaerellopsis specimens associated with Pucciniales collected in 34 countries across the globe. Eight specimens were from Africa, 11 from Asia, three from Europe, 55 from North America, 115 from the Neotropics, and five from Oceania (Fig. 5; Table 3). Our results suggest that S. macroconidialis, S. paraphysata, S. filum, and S. melampsorinearum have a cosmopolitan distribution and are adapted to different environmental conditions in both temperate and tropical regions (Fig. 6). However, S. paraphysata appears to be more abundant in the tropics. Sphaerellopsis hakeae may be an exception to this pattern, as both specimens of this species analyzed were from Australia, from where it was also described (Crous et al. 2016). The small sample size in our study limits any conclusive inferences, but it is worth noting that this species was not recovered even among the other Oceania specimens examined. Similarly, our study did not recover any additional specimens of S. artemisiae or S. isthmospora, both currently only known from China (Doilom et al. 2021; Phookamsak et al. 2019).

Fig. 5

Origin/localities of the confirmed Sphaerellopsis specimens associated with rust fungi. The numerator above the bar indicates the number of Sphaerellopsis specimens collected per country; the denominator indicates the total rust specimens screened at PUR for the presence of Sphaerellopsis per country. Countries are colored by geographic regions

Fig. 6

Geographical distribution of Sphaerellopsis specimens examined in this study. Each circle represents one specimen, and each color represents one Sphaerellopsis species

The dispersal biology of Sphaerellopsis species is not well studied. Kuhlman et al. (1978) hypothesized that conidia of Sphaerellopsis did not disperse over long distances but rather spread locally via water splashing to nearby hosts. Our results would suggest that Sphaerellopsis could also be capable of long-distance dispersal. Rust spores can be dispersed through wind currents and may cross continents, and it is possible that the much smaller conidia of Sphaerellopsis species may be passively dispersed along with their much larger host spores. However, further studies in the dispersion mode of Sphaerellopsis are necessary to support this hypothesis.

Sexual morphs of Sphaerellopsis

Eudarluca has been considered the sexual morph of Sphaerellopsis. Because these are congeneric (Keener 1951; Yuan et al. 1998), Sphaerellopsis, the older name, has priority for these fungi. The genus Eudarluca was erected in 1908 by Spegazzini to place “a new pyrenomycete” associated with the uredosori of an unknown rust, infecting Canna sp. in the Botanical Garden in São Paulo, Brazil (Spegazzini 1908). Spegazzini subsequently named Eudarluca australis as the type species of the genus. However, later in 1966, Eriksson combined several species with E. australis into E. caricis based on an overview of the taxonomy, nomenclature, and ecology of E. caricis (Eriksson 1966). The specific epithet “caricis” was kept based on the basionym Sphaeria caricis described by Fries in 1823 (Fries 1823). The original specimen of Sphaeria caricis was collected from uredinia of a rust species on Carex spp. Since Sphaerellopsis filum was misapplied in the past as the most common species associated with rust fungi, it was thought to be congeneric with E. caricis (Yuan et al. 1998), a position that was not supported by the detailed analyses of Trakunyingcharoen et al. (2014). While we were able to identify the sexual morph of S. macroconidialis on telia of Puccinia wedeliicola infecting the host plant Wedelia trichostephia, and the sexual morph of S. melampsorinearum on uredinia of Puccinia vernoniae-mollis on the leaves of Vernonia sp., we were unsuccessful in locating a sexual morph of a Sphaerellopsis specimen that would be consistent with E. caricis, and thus the asexual morph and correct name for this species remains unknown.

The two sexual specimens of Sphaerellopsis described in this study (Figs. 2 and 3) were collected in the Neotropics. Eriksson (1966), Ramakrishnan and Narasimhalu (1941), and Sebesta (1963) found that high humidity, such as is found in the tropics, favored production of the sexual morph in Sphaerellopsis. Similarly, when Västerbotten found the teleomorphic state of Sphaerellopsis in the Summer of 1962 in northern Sweden, the locality was a hollow in a compost heap, a few meters from a rivulet giving microclimate conditions “similar to the tropics” (Eriksson 1966). The host plant may also play a role in development of the sexual morph of Sphaerellopsis. For example, Eriksson (1966) noted that the sexual morph was most commonly found on plants in Poaceae and Cyperaceae due to their continuous growth and ability to form high-humidity microclimates. Although this hypothesis has not yet been experimentally tested, our findings are consistent with a high humidity requirement for sexual morph development.

Sphaerellopsis infection strategies and antagonism between S. macroconidialis and P. polysora

The two earliest diverging species in our analyses, S. artemisiae and S. isthmospora, were not recovered on any rust samples in our study and are likely not associated with Pucciniales (Fig. 1). It has been posited that several trophic strategies ranging from mycoparasitism to saprotrophism to plant pathogen may have evolved within Sphaerellopsis (e.g., Hulea 1939; Eriksson 1966). Nicolas and Villanueva (1965) posited that the anamorph of Sphaerellopsis species might be able to utilize a large number of carbon compounds; Eriksson (1966) hypothesized that Sphaerellopsis species might feed on plant tissue but that other factors, such as specific compounds secreted from the rust, might be required for Sphaerellopsis to develop its fruiting bodies. Our data would support a hypothesis of an original plant-associated trophic strategy for members of this genus, that later transitioned to a mycoparasitic strategy on plant pathogenic rusts.

This study confirms S. macroconidialis as a mycoparasite of rust fungi. Coiling and appressorium formation by S. macroconidialis and turgor loss of germ tubes of P. polysora are evident signs of antagonistic relationships between these two fungi (Fig. 4). Appressorium formation and coiling are the most common mechanisms of mycoparasites to attack their host pathogens. For example, Trichoderma harzianum and Trichoderma atroviride show the same mechanism, coiling around its host, Rhizoctonia solani, and forming appressoria as an early event before hyphal damage (Benhamou and Chet 1993; Benítez et al. 2004; Chet et al. 1981; Rocha-Ramírez et al. 2002). Similarly, Simplicillium lanosoniveum and Cladosporium tenuissimum form appressoria and helix-shaped hyphae around urediniospores of the soybean rust Phakopsora pachyrhizi (Ward et al. 2011), and aeciospores of the two-needle pine stem rusts Cronartium flaccidum, and Peridermium pini (Moricca et al. 2001), respectively. Sphaerellopsis paraphysata has also been found coiling around urediniospores of Puccinia substriata, but appressorium formation was not seen in this study (Anandakumar et al. 2019).

The formation of helix-shaped hyphae of mycoparasites around the structures of their fungal hosts is a phenomenon usually dependent on lectin recognition. Fungal lectins are carbohydrate-binding proteins located on the fungal surface, which play a role in the recognition and defense of other organisms (Lebreton et al. 2021). Once the mycoparasite recognizes the lectins of the fungal host upon first physical contact, the mycoparasite hyphae start coiling around the fungal host for colonization and further infection (Omann and Zeilinger 2010). Thus, since S. macroconidialis was observed coiling around germ tubes of P. polysora on day five after co-cultivation, genes coding for lectins-binding proteins might be up-regulated during the first four days. Many lectins have been identified in filamentous fungi and yeasts (Lebreton et al. 2021), but information on these proteins in rust fungi is scarce.

Although we observed appressorium formation by S. macroconidialis in the interaction test, these were rare. We only observed one appressorium-like structure attached to a urediniospore on day 12 of co-cultivation (Fig. 4 H). This appressorium was not formed over the germ pore of the urediniospore, and the spore showed no signs of turgor loss. Because we stopped our observations on day 12 due to the overgrowth of S. macroconidialis hyphae, it is impossible to know if the appressorium had any mycoparasitic effect on the rust spore. Appressorium formation was also not observed on Sphaerellopsis paraphysata infecting urediniospores of P. substriata previously (Anandakumar et al. 2019). Given the late appearance of appressoria, Sphaerellopsis likely do not utilize these as the primary mode for penetrating rust fungi. In contrast, Sphaerellopsis species are likely to secrete lytic enzymes (e.g., chitinases, glucanases, and proteases) to infect host rusts once their hyphae coiled around rust structures. We noticed this effect on day 12, where some germ tubes of P. polysora lost turgor (Fig. 4 I, J). Although we did not conduct studies to detect enzymatic secretion, our experimental design may be helpful for future secretome analyses.

Conclusion

In this study we attempted to fill some of the knowledge gaps surrounding Sphaerellopsis, with emphasis on obtaining data that would help to evaluate species as potential biological control agents for diseases caused by rust fungi. We demonstrate that Sphaerellopsis species are widespread and often incidentally co-collected with their rust hosts. Therefore, herbarium specimens may provide a rich source of data about these fungi. Also, Sphaerellopsis species do not appear to be specific to their rust hosts, in general, although there is a signal that some species may be climatically adapted. One new species recovered from herbarium specimens was described, S. melampsorinearum, and  the sexual morph of S. macroconidialis was characterized. Finally, we confirmed that mycoparasitic strategy of S. macroconidialis on P. polysora.