Sampling was conducted in the aquatic environments of East Azerbaijan Province, Iran (Table 1). Algae and roots of the grass species Cynodon dactylon were collected from a river and irrigation canal in 50 mL Falcon tubes and stored at 4°C prior to processing in the laboratory. Following surface sterilization for 1 min, tissue samples were cultured on NARF medium (Morita and Tojo, 2007) and incubated at 15°C for 5 days. Upon hyphal growth, a portion of the culture was transferred to WA medium (20.0 g/L agar) for purification using the hyphal tip method (Goh, 1999). Purified isolates were preserved on CMA medium in McCarthy vials at 10°C.

Growth patterns of the isolates were observed 2 weeks after inoculation on various agar media, including corn meal agar (CMA) (MIRMEDIA, Iran), potato dextrose agar (PDA) (Sigma Aldrich, Germany), malt extract agar (MEA) (DIFCO, USA), potato carrot agar (PCA) (Sigma Aldrich, Germany), and V8-juice agar (SIGMA, Germany) at 25°C. Morphological evaluations were performed on sexual and asexual structures produced on autoclaved hemp seeds and ryegrass pieces floating in sterile water from different sources (pond water, distilled water, tap water) (Martin, 1992). Twenty measurements were taken for each structure. Microscopic structures were photographed using a Nikon Eclipse Ti2 microscope with a digital camera system (Nikon, Japan). All purified cultures were deposited in the fungal culture collection of Azarbaijan Shahid Madani University, Tabriz, Iran (AZFC) and Iranian Research Institute of Plant Protection, Tehran, Iran (IRAN). Type specimens were also deposited in the herbarium of the Iranian Research Institute of Plant Protection.

The critical temperatures for growth were determined by incubating the strains on potato carrot agar (PCA) at 0, 2, 5, 10, 15, 20, 25, 30, 35, and 40°C. Descriptions were provided based on ex-type strains, and additional data for strains showing distinct morphological differences were included.

Genomic DNA was extracted from 5-day-old CMA cultures using a modified manual procedure (Möller et al., 1992). Three genomic regions were targeted for amplification: the ITS-rDNA region, which contains the 5.8S nuclear ribosomal RNA gene and the internal transcribed spacers ITS1 and ITS2, and partial sequences of the mitochondrial cox1 and cox2 genes. These regions were amplified using specific primer pairs: ITS5 (GGAAGTAAAAGTCGTAACAAGG) and ITS4 (TCCTCCGCTTATTGATATGC) (White et al., 1990) for the ITS-rDNA region, FM55 (GGCATACCAGCTAAACCTAA) and FM52R (TTAGAATGGAATTAGCACAAC) (Martin, 2000) for the cox1 gene, and FM58 (CCACAAATTTCACTACATTGA) and FM66 (TAGGATTTCAAGATCCTGC) (Villa et al., 2006) for the cox2 gene. All reactions were conducted in a total volume of 50 μL, consisting of 25 μL of ready-to-use PCR Master mix (SinaClon, Iran), 1.2 μM of each primer, 18.6 μL of DNase-free water, and 10 ng of DNA. Amplifications were carried out using a PeqStar 96X universal thermal cycler with the following conditions: 95°C for 5 min, followed by 30 cycles of denaturation at 95°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 1 min, with a final extension step at 72°C for 7 min for ITS-rDNA (White et al., 1990), and 94°C for 5 min, followed by 40 cycles of denaturation at 94°C for 30 s, annealing at 54°C for 30 s, and extension at 72°C for 1 min, with a final extension step at 72°C for 7 min for cox1 and cox2 genes (Martin, 2000; Chenari Bouket et al., 2015).

PCR amplicons were sequenced by Macrogen (Amsterdam, the Netherlands). Raw sequences were manually edited using SeqManII^® (DNA STAR) and MEGA v. 6 (Tamura et al., 2013). Sequence data from ex-type and reference strains of known Globisporangium species were obtained from NCBI GenBank (www.ncbi.nlm.nih.gov/genbank/) (Table 2). The retrieved sequences were assembled using Geneious (version 5.6) and aligned using the Q-INS-I algorithm in MAFFT (latest version), available on the MAFFT web server (Katoh and Standley, 2013; Katoh et al., 2019), separately for each of the genomic regions. Subsequently, after the removal of gaps, phylogenetic analyses were performed on the TrEase webserver (Mishra et al., 2023) for the individual genes using FastTree2 (Price et al., 2010) for Minimum Evolution, RAxML (Stamatakis, 2014) for Maximum Likelihood, and MrBayes (Ronquist et al., 2012) for Bayesian inference. For the Bayesian analysis, a GTR model was selected, and the analyses were run on random trees for 1,000,000 generations, discarding 30% of the first trees as burn-in steps of the analysis to determine posterior probabilities from the remaining trees. RAxML and FastTree2 trees were generated using the GTRGAMMA and GTR algorithms, respectively, and the reliability of the inferred trees was assessed through bootstrap analysis with 1,000 replications. After ensuring that there were no conflicting topologies in the phylogeny of the individual loci, they were concatenated, with the borders marked to ensure independent modeling of substitution rates for each partition. Multigene phylogenies (ITS, cox1, and cox2) with support values were calculated in the same manner as mentioned above, using three different approaches to assess the robustness of the inferred phylogenies. The sequences obtained in this study were deposited in GenBank, and their accession numbers have been provided in Table 2.

Pathogenicity tests were conducted using a single isolate of each species (IRAN 5217C and IRAN 5173C, respectively). Cucumber (Cucumis sativus L.), a known host for a wide range of oomycetes (Ben-Yephet and Nelson, 1999; Lévesque and De Cock, 2004; Roberts et al., 2016; Sigillo et al., 2020; Zhang et al., 2023), was selected as the test plant. The inoculum was prepared according to the methods of Ahadi et al. (2024) with minor modifications. A sterile substrate containing sandy loam soil, wheat seed, and distilled water was inoculated with five mycelial plugs (5 mm³) from three-day-old PDA cultures of the target isolates. After 9 days of incubation at 25°C, cucumber seeds were sown and grown in a greenhouse at 25 ± 2°C with a 16-h photoperiod for 14 days. Positive controls consisted of plants grown in soil inoculated with each of the test isolates, while negative controls were grown in non-inoculated soil. Symptoms including wilting, crown and root rot, and stem discoloration and deterioration were assessed daily for 14 days. Symptomatic plants were further analyzed to isolate and identify potential pathogens using previously described isolation methods. A randomized complete block design with six replicates per treatment was employed.

In the analysis of multi-locus alignment (Globisporangium species), which included the gene boundaries of ITS (1–1,066), cox1 (1,067–1,560), and cox2 (1,561–2,069), a total of 89 isolates belonging to Globisporangium, along with an outgroup, were examined. The combined dataset (ITS + cox1 + cox2) comprised 2,069 characters, including alignment gaps, with 958 variable characters (636 for ITS-rDNA, 169 for cox1, and 153 for cox2) and 1,002 constant characters (the sequence lengths were 333 bp for ITS-rDNA, 320 bp for cox1, and 349 bp for cox2).

Bayesian analysis confirmed the tree topology obtained from minimum evolution (ME) and maximum likelihood (ML) methods. While most Bayesian posterior probability values were consistent with bootstrap supports, the Bayesian tree was selected to represent the phylogeny, with bootstrap values from ME and ML methods incorporated for comparison (Figure 1). In all three-locus phylogenetic trees (Bayesian, ML, and ME), the four Globisporangium isolates were placed into two distinct, well-supported clades, each representing a separate species. Specifically, the isolates IRAN 5217C and IRAN 5344C formed a robust monophyletic group, with maximum support from Bayesian posterior probability (1) and bootstrap values (100% for both ME and ML). Similarly, isolates IRAN 5173C and IRAN 5345C were clustered together in a separate, strongly supported clade [support values: 1.00 (Bayesian posterior probability), 83% (ME), and 67% (ML)]. In the multi-locus analysis, the topologies of the single-locus phylogenies did not conflict with the respective three-locus phylogenies, confirming the reliability and accuracy of the inferred relationships among these isolates.

This study identified two Globisporangium species from various aquatic environments in Iran. Four isolates, including two from the roots of the grass species Cynodon dactylon growing in an irrigation canal (IRAN 5217C and IRAN 5344C), and two from water surface algae in the Sarab River in Sarab City (IRAN 5173C and IRAN 5345C) were reported as new species: G. parvizense sp. nov. and G. sarabense sp. nov.

Globisporangium parvizense Ahadi, Alizadeh, Chenari Bouket sp. nov. (Figure 2).

MycoBank: 857765

Typification: IRAN. East Azarbaijan: Tabriz (Varanaq), from roots of the grass species Cynodon dactylon, growing within the agricultural irrigation canals, Oct 2022, R. Ahadi (holotype IRAN 18632F). Ex-holotype culture IRAN 5217C = AZFC-RA178-2-1. GenBank: ITS = PV444307; cox1 = PV454623; cox2 = PV454627.

Etymology: The species is named in honor of the first author's father (Parviz), whose assistance was invaluable in the collection of the species.

Morphology: The colony pattern on MEA and PCA is a semi-rosette pattern, pattern on CMA is entirely within the growth medium and pattern on PDA and V8A appears submerged to stellate, with a daily growth rate of 9 mm at 25°C on PCA. The cardinal temperatures are at a minimum of 5°C, an optimum of 25°C, and a maximum of 30°C on PCA. The main hyphae are hyaline, aseptate, and range from 2.4 to 6 μm in width. The globose sporangia appear intercalary or terminal, measuring 11–32 μm in diameter (mean 23 μm). Globose hyphal swelling, terminal or intercalary, 8.5–16.5 μm (mean 10 μm) in diameter. Discharge tube arising from sporangia, short to long, one and rarely two per sporangia, and thick, 13–41 μm (mean 19.5 μm) in length and 7–8.5 μm (mean 8 μm) in width. Zoospores are not observed. Oogonia are globose and smooth, appearing intercalary or terminal, with a diameter range of 14–23 μm (mean 20.5 μm) and produced in single cultures. Antheridia are typically one, occasionally two per oogonium, diclinous. Oospores are plerotic, with one, rarely two per oogonium, measuring 15–23 μm in diameter (mean 20 μm). The wall thickness ranged from 1.8 to 3.6 μm.

Additional specimen examined: IRAN. East Azarbaijan: Tabriz (Varanaq), from roots of the grass species Cynodon dactylon, growing within the agricultural irrigation canals, Oct 2022, R. Ahadi. culture IRAN 5344C = AZFC-RA178-2-2. GenBank: ITS = PV444308; cox1 = PV454624; cox2 = PV454628.

Notes: Phylogenetic analysis revealed a close relationship between Globisporangium parvizense and G. hypogynum, followed by a slightly less close relationship with G. acrogynum and G. schmitthenneri. These species, as well as all other Globisporangium species, can be differentiated from G. parvizense based on ITS, cox1, and cox2 sequence data. Blastn searches on NCBI GenBank revealed that the ITS sequence of the G. parvizense ex-type strain (IRAN 5217C = AZFC-RA178-2-1) exhibited 99% identity with the G. hypogynum type strain CBS23494 (HQ643565), G. schmitthenneri type strain Darke1611 (JF836869), and G. acrogynum isolate CBS54988 (HQ643414) (with two, three, and three nucleotide differences, respectively). Similarly, the cox1 sequence of the ex-type strain shared 99% identity, respectively (with four nucleotide differences for both), with the G. hypogynum type strain CBS23494 (HQ708609), G. acrogynum isolate CBS54988 (HQ708461), and G. schmitthenneri type strain Darke1611 (JF895534). Comparative analysis of G. parvizense with the cox2 sequence showed 98% identity, with seven nucleotide differences compared to G. hypogynum type strain CBS23494 (AB362325), G. acrogynum isolate CBS54988 (AB362324), and G. schmitthenneri type strain Darke1611 (JF895530). The phylogenetic data robustly establish G. parvizense as a phylogenetically isolated species within the genus Globisporangium.

Globisporangium parvizense is morphologically distinct from its phylogenetically close species based on several morphological characteristics. A detailed comparison of G. parvizense, G. hypogynum, G. acrogynum, and G. schmitthenneri reveals several distinct features. G. parvizense is characterized by narrower hyphae, up to 6 μm in width, compared to 7.7, 8, and 8.3 μm in G. acrogynum, G. schmitthenneri, and G. hypogynum, respectively. While G. parvizense exclusively forms globose sporangia, G. schmitthenneri also produces lemon-shaped sporangia. Globisporangium parvizense also has smaller oogonia (up to 23 μm) compared to G. hypogynum (up to 35 μm), and consistently smooth oogonia, unlike G. acrogynum. Globisporangium parvizense typically has one, occasionally two, antheridia per oogonium, while G. schmitthenneri can have up to three, and G. acrogynum has only one antheridium per oogonium. Globisporangium parvizense is uniquely characterized by the consistent formation of two spherical oospores per oogonium, while G. acrogynum and G. hypogynum each produce a single oospore. These morphological characteristics, including narrower hyphae, simpler sporangial shapes, consistent oospore formation, and unique antheridial structures, clearly differentiate G. parvizense from its closely related species (Table 3).

Globisporangium sarabense Ahadi, Alizadeh, Chenari Bouket sp. nov. (Figure 3).

MycoBank: 857791

Typification: IRAN. East Azarbaijan: Sarab, from water surface algae in Sarab River, Oct 2022, R. Ahadi (holotype IRAN 18544F). Ex-holotype culture IRAN 5173C = AZFC-RA98-1-1. GenBank: ITS = PV444306; cox1 = PV454622; cox2 = PV454626.

Etymology: Referring to the type location, Sarab City.

Morphology: The colony pattern on MEA is a rosette pattern, whereas the pattern on CMA, PDA, PCA, and V8A are submerged to stellate. Daily growth at 25°C on PCA is 20 mm. Cardinal temperatures are a minimum of 5°C, an optimum of 25°C, and a maximum of 35°C on Potato Carrot Agar. Main hyphae are hyaline, aseptate, and 2.4–6 μm (mean 3.1 μm) wide. Sporangia are globose or subglobose, intercalary or terminal, 17–26.8 μm (mean 21.6 μm) in diameter, and lemon-shaped sporangia are intercalary or terminal, 23.1–31.7 μm (mean 27.4 μm) in length and 17–26.8 μm (mean 21.7 μm) in width. The discharge tube is 12.8–10.3 μm (mean 11.5 μm) in length and 3–3.6 μm (mean 3.3 μm) in width. Zoospores are not observed. Oogonia are rarely formed, mostly terminal, smooth, globose, 18.3–23.1 μm (mean 20 μm) in diameter, and produced in a single culture. Antheridia are one per oogonium, mostly diclinous. Oospores are globose, aplerotic, one per oogonium, measuring 12.2–23.3 μm (mean 15.8 μm) in diameter, with wall thickness ranging from 0.6 to 2.4 μm.

Additional specimen examined: IRAN. East Azarbaijan, Sarab, from water surface algae in Sarab River, Oct 2022, R. Ahadi. culture AZFC-RA98-1-2. GenBank: ITS = PV444305; cox1 = PV454621; cox2 = PV454625.

Notes: Phylogenetic analysis revealed a close phylogenetic relationship between G. sarabense, G. lucens, and G. tabrizense, followed by a slightly less close relationship with G. viniferum and G. debaryanum. The ex-type strain of G. sarabense (IRAN 5173C = AZFC-RA98-1-1) exhibited 94% ITS identity with the G. lucens strain CBS113342, 96% identity in the cox1 gene, and <96% identity in the cox2 gene, as well as <94% ITS identity with the G. tabrizense strain IRAN 18502F. It showed 96.62% identity in the cox1 gene and <97% identity in the cox2 gene. However, it also displayed 21 and 15 nucleotide differences in ITS, 18 and 17 in cox1, and 10 and 12 in cox2, respectively, compared to this reference strain. Blastn searches on NCBI GenBank indicated that the ITS sequence of G. sarabense shared the highest similarity (97% identity, with 18 nucleotide differences and 5 gaps) with the G. viniferum isolate Tr-Sm01 (MT251151). Its cox1 sequence was most similar (97%) to Pythium sp. isolate Kb003 (ON202818), and the cox2 sequence was most similar (98%) to G. intermedium (MG256947). The ex-type strain of G. sarabense shared 96% identity in ITS, <95% in cox1 and cox2 genes with the G. viniferum voucher CBS119168, differing by 20 nucleotides in ITS, 16 in cox1, and 12 in cox2. Additionally, it shared 97% ITS identity, 96% cox1 identity, and 97% cox2 identity with G. debaryanum CBS75296, but exhibited 19 nucleotide differences in ITS, 16 in cox1, and 12 in cox2. The examined loci consistently placed G. sarabense in a separate phylogenetic clade, confirming its status as a distinct species within Globisporangium.

Globisporangium sarabense shares some morphological similarities with G. lucens, G. tabrizense, G. viniferum, and G. debaryanum. However, distinct morphological features differentiate G. sarabense from its congeners. Compared to G. lucens, G. sarabense lacks zoospores, while G. lucens produces zoospores. Furthermore, G. sarabense has single oospores, whereas G. lucens occasionally has two oospores per oogonium. Oogonia chains were not identified in G. sarabense, but G. lucens exhibits the formation of oogonia chains. In terms of sporangium morphology, G. sarabense exhibits greater diversity, including lemon-shaped, globose, or subglobose forms, while G. lucens primarily has globose or subglobose sporangia. Globisporangium sarabense possesses a single antheridium per oogonium, whereas G. lucens may have as many as five antheridia.

In contrast to G. tabrizense, G. sarabense is characterized by the absence of filamentous sporangia and the possession of a discharge tube, which further distinguishes it from G. tabrizense. Furthermore, in G. sarabense, the formation of oogonia is infrequent, and in contrast to G. tabrizense, the development of oogonia chains does not occur. In G. tabrizense, as many as two antheridia are produced, whereas in G. sarabense, only a single antheridium is formed. Oospores in G. tabrizense produce both plerotic and aplerotic types, with two oospores per oogonium. In contrast, G. sarabense only produces a single aplerotic oospore. Moreover, G. viniferum has up to three plerotic oospores per oogonium, but G. sarabense has a single, aplerotic oospore.

Differentiating G. sarabense from G. debaryanum is primarily based on the absence of zoospores and the production of only aplerotic oospores in G. sarabense. At the same time, G. debaryanum produces mostly plerotic and rarely aplerotic oospores. Additionally, in G. sarabense, the wall thickness of the oospore is 1.5 μm wider than in G. debaryanum.

In summary, G. sarabense is morphologically distinct from its congeners due to its unique combination of sporangia types, absence of zoospores, oospore characteristics, and number of antheridia. These morphological differences highlight the taxonomic distinctiveness of G. sarabense within the genus Globisporangium (Table 4 and Figure 3).