Interspecies and temporal dynamics of bacterial and fungal microbiomes of pistil stigmas in flowers in holoparasitic plants of the Orobanche series Alsaticae (Orobanchaceae)

Little is known about the microbiomes of flower parts, and even less information is available regarding these microorganisms’ colonization of specific niches in parasitic plants. We investigate the temporal interspecies dynamics of the parasitic plants microbiome of flower stigmas in two stages of development: immature stigmas in flower buds and mature stigmas in opened flowers. We compared two related holoparasitic Orobanche species from localities approximately 90 km apart and characterize their bacterial and fungal communities using 16S rRNA gene and ITS sequences, respectively. We identified from 127 to over 228 OTUs per sample for fungi, sequences belonging to genera: Aureobasidium, Cladosporium, Malassezia, Mycosphaerella, and Pleosporales, constituting approximately 53% of the community in total. In the bacterial profile, we recorded 40 to over 68 OTUs per sample consisting of Enterobacteriaceae, and genera Cellulosimicrobium, Pantoea, and Pseudomonas spp., with an approximately 75% frequency. In microbial communities, higher numbers of OTUs colonizing mature stigmas were recorded than in immature. This implies that the dynamics and concurrence of microbial communities were different between O. alsatica and O. bartlingii and underwent significant changes during flower development. To the best of our knowledge, is the first study of the interspecies and temporal dynamics of the bacterial and fungal microbiomes of pistil stigmas in flowers.

Sequencing (NGS). Additionally, we examine and compare how dynamics of microorganisms were shaped between the two species separated around 90 km (the interspecies dynamics), as well as in the development of stigma in the case of immature stigmas from closed flower buds without access to the external environment and from mature stigmas from opened flowers that had access to the external environment (the temporal dynamics).

Results
After quality and length filtering, 297,043 bacterial 16S rRNA reads (from 26,105 to 48,759 per sample) and 506,384 fungal ITS reads (from 46,827 to 81,669 per sample) were obtained from the samples. The number of unfiltered reads that were aligned to these OTUs ranged from 40 to over 68 per sample for bacteria and from 127 to over 228 per sample for fungi. In both microbial communities, higher numbers of OTUs colonizing mature stigmas were recorded compared with immature stigmas in Orobanche alsatica and O. bartlingii (Table 1).
Bacterial and fungal abundance and dominance. The microbiome was composed mainly of the phyla Actinobacteria and Proteobacteria, which constituted approximately 98% of the community (Fig. 1). The analysis of the examined stigma samples allowed us to identify bacteria representing the 15 most numerous OTUs (Table 2) The dominant class of fungi (approximately 57%) in the analysed stigmas was Dothideomycetes, followed by Ascomycota cls Incertae sedis (9.8%), Tremellomycetes (8.6%) and Malasseziomycetes (8.3%) (Fig. 1). Analysis of the composition of fungi present on the stigmas showed the presence of 36 of the most numerous OTUs (Table 3). We noted high abundance-70.45% of the community in total of Mycosphaerella tassiana, Cladosporium delicatulum, Aureobasidium pullulans, Malassezia restricta, Pleosporales, Tetracladium spp., Chalastospora ellipsoidea, Vishniacozyma victoriae and Volucrispora graminea. The eudominant and dominant were represented by Mycosphaerella tassiana (11.01-25.51%), Cladosporium delicatulum (10.39-19.49%), Aureobasidium pullulans (12.11-16. Ecological indices. Analysis of ecological indicators of bacterial communities in stigmas of pistil at different stages of development revealed similar relationships, as did the analysis of fungal communities. In the fungal community of the stigma, the values of Simpson's dominance, Shannon diversity and Pielou's evenness indices for the communities were more similar than those for bacteria. Statistical analysis showed significant differences between bacteria in immature and mature stigmas in the case of the analysed indicators, as well as between fungi significant differences were recorded only for OTUs (Table 1). A lower value for the Simpson index in bacteria was found in the case of mature stigmas in both Orobanche alsatica (average λ = 0.27 vs. 0.39) and O. bartlingii (λ = 0.18 vs. 0.63), which demonstrates a greater species diversity. In the case of fungi, the situation was similar in O. alsatica (λ = 0.07 vs. 0.10), while in O. bartlingii (λ = 0.17 vs. 0.08) in mature stigma, the index was slightly higher. In the case of the Shannon diversity index, there was an inverse relationship. The index values were higher in the case of bacterial communities comparing immature to mature stigmas (H′ = 1.37 vs.  Table 2). The comparison of the composition of microorganisms in immature and mature stigmas of both species showed the existence of similarities and differences in the tested samples, supported by the conducted analyses (Table 2). At the phylum level, we observed a higher frequency of Actinobacteria in immature stigmas than in Proteobacteria, while in mature stigmas, the relationship was reversed (Fig. 1). The most abundant groups of  (1,2,5,6) and mature stigmas from opened flowers (3,4,7,8    ). In addition, OTUs such as Tetracladium spp. were found only in mature stigmas in a single sample (40.82%), as opposed to Ceratobasidiaceae, which was recorded at a high frequency (10.34%) only in immature stigmas (Table 3). In immature stigmas in both analysed species, the samples had a higher frequency of Malassezia restricta (14.53 vs. 2.00%, p = 0.001) than mature stigmas, and most recorded fungi had a similar frequency in both stages.
According to the results of the agglomeration hierarchical grouping analysis, a clear grouping of stigma samples into two clades for bacterial and fungal microorganisms was observed (Fig. 2). By analysing the arrangement of the AHC dendrogram for the distribution of variants, it was observed that two groups were formed in the clade consisting of bacteria. The first one consists of two subgroups that correspond to immature and mature stigmas To determine the correlations between bacterial and fungal microorganisms colonizing stigmas, the Mantel test (linear correlation for Pearson) was performed (Fig. 3). A correlation coefficient of r = 0.783 was obtained at α = 0.05. However, the calculated value of p < 0.0001 was lower than the significance level, which indicates the presence of correlations between the matrices under investigation.
Additionally, the biplot PCA results made it possible to ungroup the composition of microorganisms between samples of O. alsatica and O. bartlingii immature and mature stigmas. The biplot PCA for bacteria and fungi is shown separately in Fig. 4. In each case, the first two factors (F1 and F2) allow us to represent high values of the initial variability of the data, i.e., 79.51% for bacteria and 53.  (Fig. 4)

Discussion
The distances between the first and second principal components identified by PCA as well as AHC analysis (Figs. 3, 4) (Table 1). In the case of the bacterial profile, there were several OTUs that were classified as eudominant, dominant or subdominant, and the rest were much less frequent. On the other hand, fungal microorganisms were more diverse in the case of OTUs being subdominant or rare. However, the similarity of the occurrence and general frequency of especially eudominants and dominants of bacteria and fungi (e.g., Cellulosimicrobium sp., Mycosphaerella tassiana, Cladosporium delicatulum) between O. alsatica and O. bartlingii is extremely interesting. It should be emphasized that the populations of both species are separated by approximately 90 km. Interestingly, substantial abundance of Actinobacteria, Proteobacteria (e.g., genera Pantoea, Pseudomonas, Stenotrophomonas and Moraxellaceae) and Firmicutes (Paenibacillus spp.) in the stigmas of apple trees [27][28][29][30][31] were also confirmed to have been detected in the stigmas of Orobanche tested (Fig. 1). The temporal dynamics in apple stigma microbiome revealed a diverse bacterial community evolving into a community dominated by two families within the phylum Proteobacteria, the Pseudomonadaceae and Enterobacteriaceae 28 , which was also confirmed in our results in mature stigmas. Pantoea agglomerans strain E325 also had biocontrol activity against Erwinia amylovora on apple flower stigmas 39 . In addition, petals of Saponaria officinalis L. and Lotus corniculatus L. were dominated by members of the family Enterobacteriaceae (higher frequency of Serratia sp.) 40 . Serratia marcescens was also identified in mature stigmas of O. bartlingii. We observed also that the phyla and OTUs identified similar to soil microbiomes, especially genera Pseudomonas, Paenibacillus. These and other taxa (genera Bacillus, Methylobacterium, Rhizobium) were also identified in papers about microbiome flowers 24 . The higher prevalence of this group of microorganisms may be due to the fact that parasitic plants have a specific life cycle in which they spend most of their time in the rhizosphere as seeds which remain viable in the soil for many years. The communities of fungi, especially yeasts found in nectar, contain similar groups identified in stigmas in particular, e.g., Aureobasidium pullulans, Cryptococcus spp., Filobasidium wieringae, and Vishniacozyma victoriae. This may be due to the fact that stigmas, like nectaries, are habitats extremely rich in various types of nutrient substances 15,16 . Cellulosimicrobium sp. which was a eudominant or dominant in study stigmas was also found in floral nectar 41 . Some plant pathogens are able to inhibit the development of stigmas and thus, more dangerously, prevent the correct pollination process (e.g., Salmacisia buchloëana). The pistil smut fungus shifts sex ratios to be nearly 100% phenotypically hermaphroditic 28,42,43 . However, this study did not confirm the presence of these pathogens colonizing stigma. Before the flower opens, the sesquiterpene products are emitted in the bud headspace and are then absorbed and accumulated by the stigma and anthers, which supports the proper development of these structures of petunia and provides protection against microorganisms 44 . Additionally, the stigma of the pistil provides a specific microhabitat and may be an attractive source of nutrients, supporting the growth of the population of microorganisms, which may also affect the pollination process itself, stimulating the production of specific metabolites and phytohormones as well as affecting pollinators. Moreover, an extremely interesting issue is to provide information on which insects support the colonization of bacteria and fungi on stigmas. As other studies show, not only typical pollinators can be responsible for this but also the fauna that inhabit flowers on a daily basis 45 .
Orobancheae species are poorly known in terms of the bacterial and fungal microbiota that inhabit them [32][33][34][35][36][37][38] . These microorganisms can have the potential to mitigate the impacts of unfavourable environmental conditions, as well as the negative effects of plant pathogen infections 19,35,36 . Due to the dynamic and brief development of species from the Orobanchaceae family, microorganisms can respond to and potentially aid the current needs of these heterotrophic plants; for example, microorganisms in the sunflower rhizosphere affect parasitic seed germination and growth 46 as well as endophytic bacteria in seeds of Cistanche armena are able to improve the tolerance of parasitic plants under stress conditions in their natural habitat 37 . On the other hand, for the few parasitic plants that are dangerous weeds of economic importance, microbial communities may play an important role in mitigating the negative effects of infections caused by these plants. Thus, the production of various metabolites by microorganisms can support their host plant in different ways 32,47 . Notably, interactions among these species are significant determinants of the overall composition and function of plant microbiota 34 .There is no data on microorganisms colonizing the stigma of the pistil, which may be of key importance in their adaptation to the environment and reproductive biology, especially in the pollination phase. The development of the stigmas of the studied species of holoparasitic plants takes only a few days. In this short period of time, crucial for the plant, a number of mechanisms are generated that help them communicate with the environment in response to specific stimuli. In addition, a variety of substances produced within the stigmas are important in relationships with other organisms that use the same habitat island. The possibility of colonization of microorganisms from the external environment in mature stigmas occurs both day and night because the flowers of the tested Orobanchaceae species remain constantly open. This abundant and specialized ecological community of the stigma consists of commensal as well as symbiotic and pathogenic microorganisms. These microorganisms can be transferred horizontally through environmental transfer via pollinator, atmospheric or soil contamination and vertically during the parasitic plant life cycle 48 . Seed-fungal communities have been observed to be transmitted horizontally by the environment and soil versus seed-bacterial communities, which had mostly vertical transmission 49 . However, insect pollination is an ecological process involved in the transmission of bacteria from flowers to seeds; thus, the seed microbiota consists of microorganisms inhabiting not only the plant vascular tissues but also the flowers 50 . As microorganisms can interact with parasitic plants, notably during the early stages, they may have played a role in specialization. This connection allows for the exchange of various substances and microorganisms that inhabit the internal tissues of plants. There is a mutual transfer of microorganisms and homogenization between the host and the parasite during the interaction 32,34 . The presence of rhizosphere-related bacteria in mature stigmas of O. bartlingii may also be associated with soil contamination or/and the presence of a parasitic plant in the anthropogenic area. This is because the seeds of parasitic plants are disseminated in the soil where they can lie for decades. This can increase competition between soil microbiota and seed-born pioneer endophytes and modify the microbial profile 35 . Enterobacteriaceae species, detected in seeds, suggest a possible bacterial www.nature.com/scientificreports/ transmission to the seed through insect pollinators 51 , also recorded in our study in mature stigmas at a higher frequency. In mature stigmas, horizontal transmission could be due to environmental microbial deposition on flowers and pollinators. It is noteworthy that using bumblebees as vectors of various biocontrol agents is becoming increasingly popular. The potential of the yeast-like biocontrol fungus Aureobasidium pullulans vectored by bumblebees (Bombus terrestris) has been investigated, which significantly reduced the fungal pathogens. The performance and activity of the bees were not negatively affected by A. pullulans 52 .

Conclusions
Holoparasitic plants are a specific group of plants that are obligately host-dependent. In this case, there is also a likelihood that these plants benefit from the presence of bacteria and fungi colonizing the stigmas that cooperate or compete with each other during the development of stigmas, lasting only several days. The variability of microorganisms between immature stigmas from closed flowers and mature stigmas from opened flowers shows that thus far unexplored niches are dynamic, so the whole plant is also able to respond to situations from the external environment. The presence of a greater variety of fungi may be related to the fact that the stigmas provide a convenient microhabitat and an attractive source of nutrients that are more suitable for the growth of fungi, including the symbiotic yeasts that dominate the fungal profile of the stigmas. Thus, the stigmas of parasitic plants, although small, constitute a very rich microenvironment that is extremely diverse in terms of the presence of the described microorganisms. Furthermore, understanding the diversity and role of the stigma microbiome related to their reproductive biology can help to better protect these endangered species. To define the potential function and role of these bacterial and fungal microbiomes for parasitic plants, especially for the development of stigma and pollination, more research is needed. The flowers of both species were placed into sterile plastic tubes. Samples of parasitic plants with fully opened and closed petal flowers (flower bud) were selected separately. The samples were transported the same day to the laboratory and stored at 4 ± 0.5 °C until the analysis, which was carried out within 24 h. Approximately 250 stigmas from fully opened (mature stigmas) and closed petal flowers (immature stigmas) (approximately 1 g of each sample) were immediately dissected, with sterile measures, from the flowers into plastic tubes and stored at − 80 ± 0.5 °C until microbial analysis. A total of eight samples were used for further analyses, four per species, i.e., two for closed (buds) petal flowers (immature stigmas) and two for opened flowers (mature stigmas). For these samples, the pistil was cut as close as possible to the base of the stigma. The Sequencing. Laboratory analyses (DNA extraction and PCR amplification) were performed by A&A Biotechnology (Poland), whereas NGS library preparation and Illumina and Sanger sequencing (the 16S rRNA and ITS products) were conducted by Macrogen (The Netherlands). The microbial communities colonizing the analysed samples were examined by sequencing the V3-V4 region of the 16S rRNA gene and ITS region. The gene fragments were amplified with the PCR primers recommended for the Illumina technique. Primers ITS3F (GCA TCG ATG AAG AAC GCA GC) and ITS4R (TCC TCC GCT TAT TGA TAT GC) for fungal ITS library, 341F (CCT ACG GGNGGC WGC AG) and 805R (GAC TAC HVGGG TAT CTA ATC C) for bacterial 16S rRNA libraries were employed. The primers were developed by adding Illumina adapter overhang nucleotide sequences to the PCR primers used for this sequencing platform. Amplicons were indexed using the Nextera ® XT Index Kit according to the manufacturer's instructions. DNA was sequenced in Illumina MiSeq in 2 × 250 paired-end mode. www.nature.com/scientificreports/ Sequencing results were saved in FASTQ files and uploaded to the MetaGenome Rapid Annotation Subsystems Technology (MG-RAST) server for analysis 63 . Each file underwent quality control (QC), which included quality filtering (removing sequences with ≥ 5 ambiguous base pairs) and length filtering (removing sequences with a length ≥ 2 standard deviations from the mean). Sequences below three reads (singletons) or with abundance less than 0.0005% were removed after generating the ASV Statistical calculations. The OTU compositions of bacteria and fungi were analysed. The taxonomic diversities of the analysed OTUs were determined with the use of Simpson dominance (λ), Shannon diversity index (H'), and Pielou's evenness index (J′) [67][68][69] . Domination classes were determined according to previous work Przemieniecki et al. 70 for bacteria and Kurowski et al. 71 for fungi. The normality of the distribution of the obtained results was tested (Shapiro-Wilk test) and equality of variances was tested Levene's Test. Then, t-test or the Mann-Whitney U test for normally or non-normally distributed datasets was used to evaluate the statistical differences of microorganisms colonizing Orobanche alsatica versus O. bartlingii, as well as immature stigmas from closed flowers versus mature stigmas from opened flowers.