Infected plant parts of different legumes were collected between 2014 and 2019 ( Figure 1 ; Table 1 ). Small tissue parts of infected leaves, pods and flowers or whole seeds were surface sterilized for 1 min with 1% sodium hypochlorite, rinsed three times in sterile distilled water and placed onto potato-dextrose agar (PDA) with streptomycin (100 mg L^-1) and penicillin (100 mg L^-1). Cultures were incubated at 20°C under 12 h light and 12 h darkness for one week. Single-hyphal isolates were obtained from each colony and grown on PDA (Leyronas et al., 2012).

Pure cultures were evaluated for morphological characteristics and subjected to preliminary molecular genetic analyses of the ITS region, using ITS1-F, ITS4 (reverse) primers. The final collection comprised 278 Botrytis putative isolates showing traits typical of the genus. After the identification to species level by sequencing three nuclear genes RPB2, HSP60 and G3PDH (described in the following section), a set of 21 isolate was selected for further investigations. Isolates of all determined species originated from different hosts were choosen.

For each selected isolate, 10 working cultures were established by transferring a mycelial plug (5 mm ø) from the margins of the one-week-old culture: nine for pathogenicity tests and one for molecular genetic analyses.

About 10 mg of fungal material was collected from petri dishes using a sterile cell scraper with a 20 mm lifter blade (Bellco Glass, USA) (in our experience this amount covers ~ 1/3 of scraper blade) and suspended in 500 µl NucleoMag^® 96 Plant kit (Macherey-Nagel, Germany) Lysis buffer. This was followed by homogenization for 2×60 sec using FastPrep^®-24 instrument and Lysing Matrix D (MP Biomedicals, USA), phenol (Alfa Aesar, Germany) - chloroform (Merck, USA) treatment and subsequent purification with the NucleoMag^® 96 Plant kit according to the manufacturer’s instructions.

The identification to species level was carried out by sequencing three nuclear genes: RNA polymerase II (RPB2), Heat shock protein 60 (HSP60) and Glyceraldehyde 3-phosphate dehydrogenase (G3PDH). Primer combinations ( Table 2 ) were used according to Staats et al. (2005), and Raja et al. (2017) who showed that these three gene sequences were diagnostic for discriminating Botrytis species from each other.

The amplification of RPB2, HSP60 and G3PDH by PCR was carried out for 40 cycles (98°C-5 sec, 59°C-5 sec, 72°C-20 sec, GeneAmp PCR System 9700 (Applied Biosystems, USA)) in a total volume of 20 µl. The reaction mixture was comprised of following components: 10 µl of H2X Phire Plant Direct PCR Master Mix (Thermo Fisher Scientific, USA), 0.3 µM forward primer and 0.3 µM reverse primer ( Table 3 ) (Staats et al., 2005) and eluted 1 µl of fungal DNA solution. The success of the amplification was verified through the inspection of PCR products by agarose gel electrophoresis. Samples were considered positive when they contained a >700bp PCR fragment.

Positive reaction mixtures were cleaned up from excess of dNTPs and primers through employment of Exonuclease I (0.5 µl) (Thermo Fisher Scientific, USA) and Shrimp Alkaline Phosphatase (2 µl) (Thermo Fisher Scientific, USA) (incubated for 40 min at 37°C and inactivated at 95°C for 20 min). Further, 1 µl of cleaned fragment solution was transferred to BigDye^® Terminator v3.1 Cycle Sequencing reaction mixture which was prepared according to the manufacturer’s instructions (Applied Biosystems, USA). Both DNA strands of every PCR product were sequenced using primers ITS1-F, ITS4 (reverse), G3PDHfor1, G3PDHrev1, HSP60for1, HSP60rev1, RPB2for1 and RPB2rev1 and sequencing products were analyzed on 3130xl Genetic Analyzer (Applied Biosystems, USA). The ITS region sequences from each isolate were aligned using MEGA11 software (v. 11.0.13) and the created contig was BLAST^® searched against the NCBI nucleotide database to identify taxonomic sources of records with the highest sequence identity (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

G3PDH, HSP60, and RPB2 gene sequences were aligned with other Botrytis sequences, downloaded from NCBI nucleotide database, including the sequences of Sclerotinia sclerotiorum and Monilinia fructigena as outgroup taxa (Staats et al., 2005). The alignment of sequences was carried out by Multiple Alignment using Fast Fourier Transform (MAFFT) (Madeira et al., 2022), alignments were trimmed by trimAl software employing the strict method (Capella-Gutiérrez et al., 2009) and clustered by MEGA5 software employing the ‘Neighbor Joining’ algorithm to construct the phylogenetic tree (Tamura et al., 2011).

The test of pathogenicity was performed in vitro on leaves of two cultivars of each of faba bean, field pea, narrow-leafed lupin and soybean, and one of yellow lupin ( Table 2 ).

Two layers of sterile filter paper sheets were used to line metal trays (40 × 30 × 5 cm). Four 28 cm long glass sticks were arranged into each tray to support the leaves. An equal amount of water was poured into each tray, so that it was moist but water did not accumulate in puddles. Leaves were obtained from the middle level of plants at BBCH growth stage 35 (5 visibly extended internodes) and rinsed twice with sterile water. Whole leaves of faba bean and field pea were put onto the glass sticks in the metal trays so that only the petiole touched the wet filter paper. Soybean leaves were divided into leaflets and placed in metal trays so that only the cut surface touched the water. For lupin leaves, three glass sticks were tied together and three middle leaflets were detached and placed on them so as not to touch the filter paper.

Each tray contained 30 plant leaflets, one for each of the isolates and some controls. Agar plugs with Botrytis isolates (5 mm ø) were placed on leaflets with the mycelium side down. Two non-inoculated agar plugs were used in each tray as a control. After inoculation, the trays were covered with plastic film to maintain high humidity and incubated at 20°C in the dark for 72 h. The diameter of the formed lesions under the inoculated plug was measured crosswise once a day until some of lesions reached the edge of the leaf. Four replicates were prepared and the experiment was repeated, giving 8 replicates. A further repeat was conducted for narrow-leafed lupin ‘Probor’, giving 12 replicates. Mean lesion diameter was considered as the measure of pathogenicity.

Leaflet datasets were subjected to analysis of variance using SPSS v. 27 (IBM Inc., Chicago, IL, USA) within each host genus (yellow and narrow-leafed lupins were analysed together). A simple split-plot model was applied to the data from each day, where the culture box was the main plot and the leaflet was the subplot. The repeated-measures model was then applied to 4 days of data from faba bean and soybean as well as to 3 days of data from lupins and pea. The output was corrected to take into account the split-plot design of the experiment.

The 26 isolates collected from legumes in this study with the Botrytis spp. isolates from GenBank formed a clade with 99% bootstrap support, distinct from the closely related M. fructigena and S. sclerotiorum outgroups. Branches with under 50% bootstrap support were not included in the phylogenetic tree ( Figure 2 ). Phylogenetic analysis of the combined gene datasets confirmed that all isolates belonged to the genus Botrytis. Six isolates clustered with B. cinerea references, three with B. fabae, seven with B. fabiopsis and three with B. pseudocinerea. One isolate was grouped with B. medusae and one isolate with B. euroamericana ( Table 1 ). Five isolates did not clearly group with identifiable clades, so they were not considered further in this study.

The first symptoms of infection were observed 24 h after inoculation of all host cultivars. The main effects of Botrytis species and isolates were significant for all measures on all host species ( Table 4 ). The effects of replicate and of the day × replicate interaction were significant for pea, because lesion diameters were 1 mm greater in replicates 1 to 4 than those in replicates 5 to 8 on day 1, which continued until day 3, when it was no longer significant.

Differences between cultivars were significant in soybean, where lesions were larger on ‘Skulptor’ than those on ‘Laulema’ from day 2 onwards, and in pea, where the lesions on ‘Doloresa’ were larger than those on ‘Florida’ on all dates. The cultivar × species interaction was significant in soybean and lupin, while the cultivar × isolate interaction was significant for all species except faba bean. On days 1 and 2, B. cinerea and B. pseudocinerea caused much larger lesions on the yellow lupin ‘Lejaskurzeme’ than on the narrow-leafed cultivars. On day 1, lesions of B. medusae on ‘Probor’ were much larger than those on the other two lupin cultivars and on days 2 and 3, ‘Sonet’ had smaller lesions than the other two cultivars. On days 2 and 3, both B. euroamericana and B. fabae produced larger lesions on ‘Sonet’ than on the other two lupins.

To compare the pathogenicity of Botrytis spp., the optimal assessment day was examined for each host species, when the leaves were still fresh and the lesion diameter was determined. For pea, the optimal assessment day was the third, for lupins the second, for soybean the fourth, and for faba bean the fifth day after infection. Based on the size of the lesions and the characteristics of spot formation in each host species, we divided the pathogenicity of the isolates into three classes: – low (primary lesion only), moderate, and high ( Table 5 ).

In faba bean, B. fabae caused the largest lesions on day 5 ( Figure 3 ), with isolate 17B28 being the strongest ( Table 5 ). The lesions produced by B. pseudocinerea and B. fabiopsis were somewhat smaller than those caused by B. fabae and were similar to each other. B. fabiopsis isolate 17F4 produced lesions second to those of B. fabae 17B28 in size. Isolates 18B11 and 18B2-3 of B. cinerea exhibited high pathogenicity on faba bean leaves; however, as a species it was less pathogenic than B. pseudocinerea. Lesions caused by B. euroamericana were observed only in cv. ‘Lielplatone’. B. medusae did not cause any injury in either faba bean cultivar.

On day 3, all species caused lesions on pea leaves, and the most pathogenic were B. medusae and B. cinerea ( Figure 4 ). B. fabae was significantly less pathogenic. In general, B. fabiopsis caused primary or moderate lesions, except 17F4 and 19B175 ( Table 4 ), which formed large lesions on both the tested pea cultivars. Isolate 17F4 caused lesions that were even larger than those of the most pathogenic B. cinerea isolates. B. pseudocinerea isolate 18B2-3 was also responsible for large lesions.

On lupin leaves, B. medusae was the most pathogenic ( Table 5 ) and caused significantly larger lesions than the other Botrytis spp. ( Figure 5 ). B. pseudocinerea and B. cinerea caused moderate to large lesions, depending on the isolate, and their pathogenicity was significantly higher than that of B. fabae, B. fabiopsis and B. euroamericana, although B. fabiopsis isolate 17F4 produced lesions that were nearly as large as those of the most pathogenic B. cinerea isolates ( Table 5 ).

Lesion sizes were smaller in soybean than on the other host species ( Figure 6 ). B. fabae, B. euroamericana and B. medusae caused only primary lesions on both soybean cultivars. One B. cinerea, two B. fabiopsis and one B. pseudocinerea isolate caused lesions larger than 5.5 mm in diameter ( Table 5 ), that larger than those caused by the agar plug of inoculum. Cultivar ‘Skulptor’ was notably more susceptible to B. pseudocinerea than cultivar ‘Laulema’.

Certain isolates were highly pathogenic to all host species; notably B. fabiopsis 17F4 (highly pathogenic on all), B. cinerea 19B009 and B. pseudocinerea 18B2-3 (highly pathogenic on three and moderately pathogenic on one), and B. fabiopsis 19B175 and B. pseudocinerea 18B11 (highly pathogenic on two and moderately pathogenic on two host species ( Table 5 )). At the other extreme, B. fabiopsis 17F9 and B11, along with B. pseudocinerea 18B6 had, low pathogenicity in all hosts.