Introduction

The genus Paraphoma Morgan-Jones & J. F. White (Phaeosphaeriaceae) was established in 1983 with Paraphoma radicina (McAlpine) Morgan-Jones & J. F. White (≡ Pyrenochaeta radicina McAlpine) as the type species (Morgan-Jones and White 1983). Initially, it was suggested that the most informative taxonomic feature for members of this genus was setose pycnidia. However, the presence of such pycnidia in the fungal life cycle is specific for both Paraphoma and Pyrenochaeta De Not. species. Thus, it has been proposed to use ultrastructural features of conidiogenesis to distinguish the species of these two genera. According to those data, Paraphoma species have been placed in the genus Phoma Sacc. as a part of the appropriate section Paraphoma (Boerema et al. 2004).

Substantial difficulties in morphological identification have resulted in poor understanding of the generic and species boundaries generally in coelomycetes and particularly in the genus Phoma and its sections (sensu Boerema et al. 2004). Additionally, classification systems based on morphological features are highly artificial and do not represent evolutionary relationships. Molecular phylogenetic studies based on DNA sequencing have shown that Paraphoma is not a sister clade to phoma-like fungi transferred to the family Didymellaceae, but it is closely related to other genera affiliated with the families Phaeosphaeriaceae (de Gruyter et al. 2010), Cucurbitariaceae, and Coniothyriaceae (Chen et al. 2015). Currently, MycoBank categorizes eight species in the genus Paraphoma, and this group of organisms is actively being investigated. In the last 6 years, at least four new Paraphoma species have been described (Quaedvlieg et al. 2013; Crous et al. 2017; Moslemi et al. 2017).

Convolvulus arvensis and Calystegia sepium are perennial, soboliferous plants and two of the most harmful weeds. Controlling these weeds requires intense tillage and the use of a considerable amount of herbicides (Stetsov and Sadovnikova 2012; Nadtochiy 2008). Consequently, the potential application of biological weed control alternatives, particularly phytopathogenic fungi, has been studied more intensively in recent years. Several members of Phaeosphaeriaceae, including Paraphoma species, can be applied as living mycoherbicides and produce bioactive compounds with herbicide characteristics (Guntli et al. 1998; Poluektova et al. 2018).

The taxonomical diversity of phytopathogenic fungi, which can infect plants of the family Convolvulaceae, was examined in Russia and other countries. Several phoma-like species were revealed on these plants and are listed in Table 1. Three of those species produce several compounds with mycoherbicide activity against C. arvensis: Phoma proboscis (Heiny and Templeton 1991, 1995), Phomopsis convolvuli (Ormeno-Nuñez et al. 1988a, b; Watson et al. 1993; Vogelsang et al. 1998), and Stagonospora convolvuli (Pfirter and Defago 1998; Pfirter et al. 1999; Defago et al. 2001).

Table 1 Phytopathogenic phoma-like fungi that can infect plants of the family Convolvulaceae
Full size table

Due to extensive analysis of fungal biodiversity on weeds in Russia and neighboring countries in 1990–2010, pure cultures of fungi isolated from Convolvulaceae were collected. This collection is stored in the laboratory of Mycology and Phytopathology of the All-Russian Institute of Plant Protection. It includes 70 isolates of phoma-like fungi obtained from C. arvensis and C. sepium. All isolates in this collection were identified based on morphological features. A considerable number of isolates in this collection were not identified at the species level and have unclear definitions, such as Ascochyta sp., Mycosphaerella sp., and Phoma sp. According to preliminary molecular phylogenetic data (Gomzhina et al., unpublished), the collection consists of at least ten genera of phoma-like fungi. Among them, eleven isolates were identified as species of the genus Paraphoma.

The aim of this study was to correctly reidentify Russian Paraphoma isolates collected from Convolvulaceae and to taxonomically revise the fungal isolates, based primarily on a molecular phylogenetic approach and traditional morphological analysis.

Materials and methods

Isolates

As a result of the extensive studies of fungal biodiversity on Convolvulaceae weeds carried out in 1990–2010 in different geographical locations in Russia and territories of neighboring countries, eleven Paraphoma isolates (Table 2) were collected by authors from the leaves of C. arvensis and C. sepium that exhibited typical leaf spot symptoms. To isolate a pure culture of fungus from the leaves, fragments of infected material were surface sterilized with 20 ml of 5% sodium hypochlorite (NaClO) solution. Firstly washed for 2 min with 0.1% sodium dodecyl sulfate (SDS), washed with 5% sodium hypochlorite, and then washed three times with 20 ml of sterile water. After surface sterilization, the samples were placed on potato-sucrose agar (PSA) (Samson et al. 2000) containing antibiotics (100 μg/ml ampicillin, streptomycin, penicillin, HyClone, GE Healthcare Life Science, Austria) and 0.4 μl/l Triton X-100 (PanReac, Spain) to restrict the growth of fungi. The Petri dishes were incubated at 24 °C in the dark and were analyzed on days 7–10 of cultivation. Samples of infected leaves were deposited in the Mycological Herbarium (LEP) of All-Russian Institute of Plant Protection (VIZR). All Paraphoma isolates were stored in plastic microtubes on PSA at + 4 °C in the VIZR pure culture collection.

Table 2 Information of the geographical location and dates of sample collections, from which the investigated isolates were obtained
Full size table

DNA isolation, PCR, and sequencing

Mycelium was scraped from 20-day-old cultures on oatmeal agar (OA, Boerema et al. 2004) and macerated with 0.3-mm glass sand on a Retsch MM400 mixer mill (Retsch, Germany). Genomic DNA was then extracted according to the standard CTAB/chloroform method (Doyle and Doyle 1990).

The ITS and LSU regions of rDNA were amplified and sequenced for all 11 isolates. The RNA polymerase II (RPB2) and β-tubulin (TUB) genes were sequenced for 10 and 9 isolates, respectively. All obtained sequences were deposited into GenBank, and accession numbers are listed in Table 3.

Table 3 Collection details and GenBank accession numbers of isolates
Full size table

The primers ITS1F (Gardes and Bruns 1993) and ITS4 (White et al. 1990) and the primers LR0R (Rehner and Samuels 1994) and LR5 (White et al. 1990) were used to amplify the ITS and LSU regions of the ribosome genes, respectively. The primer T1/T2 (O’Donnell and Cigelnik 1997; Saleh and Leslie 2004) was used to amplify part of the TUB gene, and fRPB2-7cR/fRPB2-5f2 (Liu et al. 1999) was used to amplify part of the RPB2 gene.

The amplification reactions had a total reaction volume of 25 μl, which was composed of dNTPs (200 μМ), each of the forward and reverse primers (ITS1F/ITS4, LR0R/LR5, T1/T2, fRPB2-7cR/fRPB2-5f2) (0.5 μМ), Taq DNA-polymerase (5 U/μl), 10× PCR buffer with Mg2+ and NH4+ ions, and 1–10 ng of total genomic DNA.

The PCR conditions were as follows: predenaturation of DNA at 95 °С for 5 min; 35 cycles of denaturation at 92 °С for 50 s, annealing at 55 °C for 40 s (ITS1F/ITS4 and Т1/Т2) or at 55 °C for 50 s (LR0R/LR5), and elongation at 72 °С for 75 s, followed by a final elongation step for 5 min at 72 °С. The RPB2 gene was amplified according to the touchdown program. All steps were the same as described below, but the annealing temperature consequently declined from 5 cycles of 60 °С for 40 s and 5 cycles of 58 °С for 40 s to 30 cycles of 54 °С for 40 s.

After PCR, amplicons were purified according to a standard method with a DNA-binding silica matrix (Boyle and Lew 1995). Visualization and concentration measurements of the purified PCR products were implemented by electrophoresis in 1% agarose gel stained with ethidium bromide and MassRuler 100 bp as a marker of concentration.

Amplicons were sequenced by Sanger’s method (1977) on ABI Prism 3500 (Applied Biosystems, Hitachi, Japan), with the Big Dye Terminator v3.1 Cycle Sequencing Kit (ABI, Foster City, USA), according to the manufacturer’s instructions.

Phylogenetic analysis

Sequences were assembled using Vector NTI advance v. 11.0 (Invitrogen, Thermo Fischer Scientific, Waltham, USA) and aligned with ClustalX 1.8 (Thompson et al. 1997). All known representative Paraphoma strains and type species of all genera in the family Phaeosphaeriaceae, with Phaeosphaeriopsis glaucopunctata as an outgroup, were obtained from GenBank (Table 4) and included in the analysis. The phylogenetic trees were inferred with RAxML (randomized accelerated maximum likelihood) software (v. 7.2.8, Stamatakis 2006) by the maximum-likelihood (ML) method. Bootstrap support values with 1000 replications were calculated for tree branches.

Table 4 GenBank accession numbers of reference isolates
Full size table

Morphological analysis

Pure cultures were incubated on PSA and OA amended with 200 mg/ml streptomycin sulfate for up to 2 weeks under standard conditions (Boerema et al. 2004). Petri dishes were placed for 1 week in darkness and then for a week under 12-h near-ultraviolet light/12-h dark to stimulate sporulation. Colony diameter was measured after 7 days, and colony morphology was examined after 14 days of incubation. Colony colors on the surface and underside of the inoculated Petri dishes were assessed according to the color charts of Bondartsev (1953). Isolates that did not produce pycnidia on agar medium were cultivated on autoclaved grain (millet, oat, and pearl barley) under near-ultraviolet conditions for 14 days. Observations and measurements of 50 replicates of conidia and conidiomata were conducted with a stereomicroscope Olympus SZX16 (Olympus, Japan) and microscope Olympus BX53.

Results

Phylogeny

The adjusted and aligned sequences in the phylogenetic analysis had the following lengths: ITS region, 483 bp; LSU, 845 bp; TUB, 490 bp; and RPB2, 736 bp; the number of polymorphic sites per genome locus was 44 (9.1%), 2 (0.2%), 28 (5.7%), and 32 (4.3%), respectively.

Nucleotide sequences of 11 isolates were used for 2-gene phylogenetic analysis (sequences of ITS and LSU regions). Sequences of 10 isolates in the 3-gene phylogenetic analysis (ITS, LSU, and RPB2) and sequences of only nine isolates in the 4-gene phylogenetic analysis (ITS, LSU, TUB, and RPB2) were suitable for analysis. Three phylogenetic trees were developed: an individual tree of the ITS region, a combined tree for the ITS and TUB genes, and a combined phylogram for all studied loci.

All currently known species of Paraphoma and all our isolates formed a common monophyletic group. Within this group, all our isolates clustered as three distinct clades in all phylogenetic trees with a maximum value of bootstrap support (100%). The composition of these clades was identical in all trees, and these clades did not cluster with any Paraphoma species (Figs. 1, 2, 3). The first clade consisted of four isolates (MF-9.301, MF- 9.298.1, MF-9.265, MF-9.300.1). The second clade included six isolates (MF-9.296.1, MF-9.88, MF-9.294.1, MF-9.240, MF-9.95, MF-9.182.1) (Figs. 2, 3). The third clade included only the isolate MF-9.222 (Fig. 1).

Fig. 1
figure 1

Maximum-likelihood phylogenetic tree inferred from ITS, representing all Paraphoma species and reference of Stagonospora convolvuli

Full size image
Fig. 2
figure 2

Maximum-likelihood phylogenetic tree inferred from ITS and TUB

Full size image
Fig. 3
figure 3

Maximum-likelihood phylogenetic tree inferred from ITS, TUB, and RPB2

Full size image

In the tree of the ITS region (Fig. 1), the isolate MF-9.222 clustered within the same clade with the two reference strains of S. convolvuli (12–039; 01–634), which are represented in GenBank only by these two sequences. Thus, according to the obtained molecular phylogenetic data, isolate MF-9.222 was identified as S. convolvuli. However, due to its position in the Paraphoma cluster, the new taxonomical combination Paraphoma convolvuli has been proposed for S. convolvuli.

The other isolates were divided into two subclades in all phylogenetic trees. The topologies of the combined phylogenetic trees were more detailed, including the topology of clade 2, which was more explicit and allowed the division of isolates MF-9.182, MF-9.299, and MF-9.240 into two subclades (Figs. 2, 3). This clade did not include any type or representative isolate. It is monophyletic and, in all trees, has a maximum value of bootstrap support. Therefore, the isolates of this clade are considered a new species, Paraphoma melnikiae.

Morphology and Taxonomy

Paraphoma convolvuli (Dearn. & House) Gomzhina M. M. & Gasich E. L. comb. nov.

MycoBank MB823867

Basionym Stagonospora convolvuli Dearn. & House

Paraphoma melnikiae Gomzhina M. M. & Gasich E. L., sp. nov.

MycoBank MB823800

Type specimen LEP 131845. The type specimen represents dried leaves of C. arvensis with leaf spots collected in Saint Petersburg on February 17, 2002.

Etymology: Named after Dr. V. A. Melnik (1937–2017), an outstanding Russian mycologist and taxonomist who dedicated his work to different fungi, including phoma-like species.

This fungus causes leaf spots on C. arvensis. Spots are incorrectly rounded with concentric zones and some pycnidia (Fig. 4). Pycnidia are diffuse, semisubmerged, rounded, dark brown, and 100–250 μm (Figs. 5, 6). Conidiophores are reduced to phialidic conidiogenous cells formed from the inner cells of the pycnidial wall, hyaline, discrete, flask-shaped, and 7.29–10.25 × 3.6–4.24 μm (Fig. 7). Conidia are cylindrical with rounded tips, straight or slightly curved, with 0–2 transverse septa, 9–22.5 × 1.5–3.8 μm (Fig. 8). On OA (Fig. 9), the colony diameter is 25–34 mm after 7 days and 42–54 mm after 14 days. Felty-velvet or flocculose felty-velvet aerial mycelia are not abundant, pale-olivaceous. The colors of the colonies on the upper and lower parts are from pale to dark brown, sometimes with dark brown, reddish, and fallow sectors. The color of the colonies could also be from pale to dark vinaceous shades, sometimes with pale-olivaceous sectors. Margins are regular and slightly curved. Pycnidia are sparse, scattered, immersed and semi-immersed, rare in aerial mycelium, dark brown, rounded, hairy, 40–420 μm, with 1–3 ostioles. Conidia are hyaline, multiguttulate, cylindrical with rounded tips, straight or slightly curved, with 0–2 transverse septa 7–16 (10.4 ± 0.5) × 1.5–2.5 (2.0 ± 0.1) μm.

Fig. 4
figure 4

Leaf spots on Convolvulus arvensis caused by Paraphoma melnikiae sp. nov. from the type material

Full size image
Fig. 5
figure 5

Immersed pycnidia of Paraphoma melnikiae sp. nov. on leaves of Convolvulus arvensis from the type material

Full size image
Fig. 6
figure 6

Pycnidia of Paraphoma melnikiae sp. nov. on leaves of Convolvulus arvensis from the type material

Full size image
Fig. 7
figure 7

Conidiogenous cells of Paraphoma melnikiae sp. nov. from the type material

Full size image
Fig. 8
figure 8

Conidia of Paraphoma melnikiae sp. nov. from the type material

Full size image
Fig. 9
figure 9

Morphology of Paraphoma melnikiae sp. nov. colonies on OA (ex-type isolate MF–9.88)

Full size image

Chlamydospores were absent. Perithecia were not observed.

Note: Morphologically, the conidia of P. melnikiae differ from conidia of the closely related species P. convolvuli in shape and size. The conidia of P. convolvuli are longer (15–18 μm) and have more transverse septa (2–3) (Saccardo 1931).

Discussion

Based on the morphological characteristics, all isolates were primarily identified as S. convolvuli by the authors. However, the conidia of the studied isolates were broader and less elongated than the conidia of Stagonospora species. Our isolates did not possess typical morphological features of pycnidia and conidia to identify them as members of the genus Paraphoma. The pycnidia of the studied isolates were not setose, and the conidia were longer than typical Paraphoma conidia. It is known that such morphological characteristics are highly variable and do not represent phylogenetic relationships among fungi in this group.

Unlike the traditional morphological approach, molecular-phylogenetic methods allow the identification of all isolates as members of the genus Paraphoma. Based on molecular data, a new combination, P. convolvuli, was proposed for S. convolvuli. Isolate MF-9.222 should be identified as P. convolvuli, whereas isolates from clade 1 should be identified as Paraphoma cf. convolvuli. All Paraphoma isolates from clade 1 shared similar morphological features with P. convolvuli isolate MF-9.222 but differed from it by a single deletion in the ITS sequence and one insertion in the LSU sequence. Clade 1 was monophyletic and well supported; all isolates were obtained from C. sepium, not from C. arvensis, as was isolate MF-9.222 and the reference P. convolvuli. Apparently, these isolates are new species of the genus Paraphoma, but this requires subsequent validation.

Isolates of the second phylogenetic clade were treated as a new species of Paraphoma, P. melnikiae. This new taxon was proposed according to the polyphasic approach to species recognition (Consolidated Species Concepts) and based on phylogenetic, morphological, and biological characteristics.

To construct phylogenetic hypotheses for closely related phoma-like species, the most informative loci are ITS, TUB, and RPB2. The sequencing of these loci was implemented in this study and resulted in robust, well-supported phylogenetic clades in the phylograms. Thus, to resolve phylogenetic relationships among Paraphoma species, this set of loci is also taxonomically informative. Despite this being used in non-taxonomic studies, the molecular identification of phoma-like fungi is often based only on sequences of the ITS region. Thus, data on sequences of phylogenetic informative loci of particular phoma-like species in GenBank are often presented one-sidedly and scantly. The implementation of phylogenetic studies in such cases becomes difficult. Although it was previously suggested not to identify phoma-like fungi only by morphological features and to take into account molecular traits, now it is not recommended to identify these fungi only by sequencing the ITS loci, as the most popular region for phylogenetic studies, but to include sequences of other informative regions of DNA in the phylogenetic analysis.

A majority of Paraphoma species are widely distributed, occurring as soil-borne fungi causing diseases of aboveground parts of plants. Analysis of the pure culture collection of phoma-like fungi derived from Convolvulaceae showed that species of the genus Paraphoma were detected in Russia, Kazakhstan, and Ukraine. P. convolvuli was found on C. arvensis only in Kazakhstan, whereas closely related isolates of Paraphoma cf. convolvuli (probably a new species) were detected only in one location in Russia, in Saint Petersburg, on C. sepium. The new species P. melnikiae was found on C. arvensis in two locations in Russia (Saint Petersburg and Vladivostok) and in Ukraine.