The Opportunistic Pathogen Sphaeropsis Sapinea is Found to be one of the Most Abundant Fungi in Symptomless and Diseased Scots Pine in Central- Europe


 Background: The opportunistic and latent pathogen Sphaeropsis sapinea is one of the most important forest pathogens on pine. The fungus may cause Diplodia tip blight on several coniferous trees and disease symptoms come visible when trees are weakened by stress, usually related to injuries and drought. This project compares the mycobiome of healthy and diseased Scots pines. Twigs were sampled in June and September 2018 in a German forest stand with varying health status of the sampled Scots pines. Growth of 2017 and 2018 were sampled and cultivable, filamentous fungi isolated and the mycobiome was analysed by high-throughput sequencing (HTS) of the internal transcribed spacer 2 (ITS2) region. Results: A PERMANOVA analysis confirmed that the mycobiome community composition significantly differed between growth years (p < 0.001) and sampling time (p < 0.001) but not between healthy and diseased trees. Higher amount of S. sapinea was observed in June and the growth year 2017. Besides yeasts which were neglected, 23 ascomyceteous fungal endophytes were isolated. S. sapinea was the most common endophyte isolated and the second common in HTS data. It was highly abundant in symptomless (healthy) trees. Conclusion: Results highlights the ability of S. sapinea to accumulate unnoticed before disease outbreaks, implementing the sudden threat for Scots pines in the future.

Taxa isolated with the culture-based method.  [77,78]. 26 OTUS were assigned to the Exobasidiaceae, which usually form colonies with single-celled conidia but without hyphae. Members of this basidiomycetous Family are commonly none-pine host-speci c plant pathogens [79]. In total 312 (25%) of all detected OTUs with HTS may represent species with yeast or yeast-like stages. 35 OTUs could be classi ed as lamentous Basidiomycota including three ectomycorrhizal fungi, whereas for 48 basidiomycetous OTUS where no further assignment to trophic stage or lifestyle was possible. The ectomycorrhizal fungi (OTU177, OTU896, and OTU1088: Laccaria spp.) can be assumed to have an epiphytic source as symbiotic, root associated species. 274 OTUs were assigned to Ascomycota growing with mycelia, excluding species with probably epiphytic source such as lichens or lichenicolous fungi (38 OTUs), fungicolous or obligate non-pine parasitic fungi (6 OTUs) or ascomycetous sooty molds (2 OTUs). For 125 ascomycetous OTUs where no further assignment to trophic stage was possible, as well as for 372 OTUs which represent Fungi with no signi cant similarity to sequences in database. Usually species of Chytridiomycetes (10 OTUs) are inhabiting soil, fresh water, and saline estuaries or are parasitic on e.g. amphibians. Therefore it is assumed that chytrid OTUs had an epiphytic source as well as the three Glomeromycota (OTU359, OTU437, and OTU42) which are arbuscular mycorrhizal fungi.

Disease class
Even the composition of disease class 0, in isolate data, is dispersed more alone (Fig. 5A) no statistical difference with PERMANOVA analysis (p = 0.062) between disease classes was observed. Based on the "indispecies" analysis M. olivacea was statistically different (p = 0.0001) in disease class 0 (isolate data). The diversity indexes, Simpson (p = 0.0625) and Shannon (p = 0.135), were not statistically different in cultivable data. PERMANOVA analysis (p = 0.141), permutation test and visualization of HTS data did not show any statistical differences in grouping of OTUs (Fig. 6B). Similarly, the diversity indexes (Shannon p = 0.871, Simpson p = 0.826) were not statistically different between disease classes. This indicates that species diversity in a disease class is similar (abundance and evenness of the species present) in HTS data. The PERMANOVA comparison was made also for HTS reads observed only in sampling time June (p = 0.25) or September (p = 0.367) con rming no statistical difference were observed between disease classes. Similarly, the composition of plant pathogens was not statistically different between disease classes (p = 0.699). The comparison of these groups con rmed the null hypothesis that the centroids and dispersion of the OTUs and diversity indexes are equivalent for all groups.
Year of the growth The species composition was statistically different in isolate data (p = 0.001) and the composition of OTUs were statistically different (p = 0.001) in HTS data between year of the growth (2017 vs 2018) (Fig. 7). Diversity indices (Shannon, p = 3.76e-05; Simpson p = 5.67e-07) for isolate data were signi cantly different, indicating higher diversity in 2018. Similarly, diversity indices were signi cantly different for HTS data (Shannon, p = 2e-16; Simpson's, p = 1.44e-13), indicating higher diversity in 2018.
1. polyspora (p= 0.0001) and S. sapinea (p= 0.0008) were found to be statistically different between growth years from HTS data. Similarly, in isolate data S. polyspora (p = 0.0001) was statistically different between growth years. In HTS data set S. polyspora abundance was higher in the growth year 2017 (both sampling times) and in isolate data year 2018. PERMANOVA analysis with HTS data showed that S. sapinea reads were statistically different between sampling times (p = 0.001), growth years (p = 0.001), and disease classes (p = 0.043). The number of S. sapinea reads (HTS data) were higher in June and in growth year 2017. Kruskall-Wallis test for HTS data showed that disease class 0 was different from disease classes 2, 4 and 5 (Fig. 9a). Similarly, disease class 3 was different from 2 and 4 ( Fig. 9a). Four outliers, all from June 2017 samples, were detected from HTS data (disease class 1 = 132497 reads, disease class 2 = 86340 and 804269 reads; disease class 4 = 730005 reads). After removing these outliers PERMANOVA analysis showed differences between sampling times (p = 0.001), growth years (p=0.012) and disease classes (p = 0.019). The numbers of reads were higher in June and in growth year 2017. Kruskall-Wallis comparison showed that disease classes 2 and 4 (the number of reads) were statistically higher from disease classes 0 and 3 (Fig. 9b). After removing outliers, the average number of reads did not differ statistically between disease classes 0 and 5.

Discussion
Culture-based isolation method and HTS accomplish each other Selection of methods is crucial for studies aiming at providing the full range of an organism's microbiome or as in this study, the mycobiome of pines' twigs. The large number of detected OTUs with HTS was to be expected, as high fungal community quantities have been described before, outstanding in their diversity of morphologies and trophic strategies [74].
The assemblage of fungi detected with the cultivable method only detects fungi with the ability to metabolize the provided nutrient medium, which are fast enough to grow out of their wood piece in the given time of the experiment and which are not antagonized in their tissue by surrounding lamentous fungi "on their way" growing out to the surface. Experience, thorough design of the method and careful handling not to oversee outgrowing fungi, as it was guaranteed in this study to our best knowledge, certainly increases the likelihood to detect the fungi growing in the observed samples. Except Diaporthe sp., P. funiculata, and Th. fuckelii, all other 20 fungi isolated within this study were also found in a previous study by Bußkamp et al. [30], where 103 fungal species from 25800 Scots pine twigs segments (in comparison to 1358 segments from this study) were identi ed. Four other Diaporthe and two other Preussia species were identi ed by Bußkamp et al. [30]. Th. fuckelli is a typical endophyte of Scots pine and occurred in the natural distribution area of its host [80]. Usually it is fruiting on dead branches and it is assumed to be member of the fungal self-pruning community of pine [81].
HTS method on the other hand includes not only endophytes, but also epiphytes and non-cultivable species including yeasts. Nevertheless, the methods proved to accomplish each other: surprisingly, only 65 % of the identi ed species from the cultivation method were found in the HTS data. Eight species B. mediterranea, B. nummularia, H. fragiforme, and Rosellinia sp. (all Xylariales, Sordariomycetes), J. rotula (Sordariales, Sordariomycetes), D. acicola and Py. domesticum (both Pezizales, Pezizomycetes), and Pezizomycetes sp. were not detected by HTS (Fig. 5). Species of Sordariomycetes (39 % of all isolated species in this study by classical method) are common endophytic fungi in plant tissues and comprised 31% of all isolations of Scots pine twigs in the study by Bußkamp et al. [30] or 32% of all isolations from pine branches in studies by Sanz-Ros et al. [82]. D. acicola is a typical saprophyte of Scots pine needles and endophyte of Scots pine twigs [30]. Py. domesticum is a pyrophilous cup fungus, occurring on burnt or sterilized soils. It typically fruits within a few weeks after a burn [83]. This indicates that both methods are necessary to provide the mycobiome, especially endophytes of an organism. In contrast to our theory that HTS would detect all fungi found in the sequence data from the amplicon method, roughly half of the species were not detected. Indeed, still the choice of primers and databases proves to be the bottleneck for the discovery of all species [73,84]. The advantage of the culture-based method is the gain of living cultures, which could be tested regard, virulence, antagonism and ecological relevance and function.

Main species of Pinus twigs
The main species/OTUs were detected with both methods, but rarer ones only with HTS methods. As con rmed from studies by Bußkamp et al. [30], some of the isolated endophytes of pine twigs may play a role as decomposer or weak pathogens of the host twigs and needles, e.g. S. sapinea, T. conorum-piceae, D. acicola, Ph. lacerum, and Peniophora pini (Schleich. ex DC.) Boidin. For the most of the isolated endophytes no signi cant function is known. Endophytic fungi in twigs of P. sylvestris with culture based methods were analyzed by several authors in the past, e.g. [44,82,[85][86][87][88]  In this study, the most abundant fungus in all disease classes identi ed with HTS was the common foliar endophyte of Scots pine, Sy. polyspora [44,89]. However, it has been noted to cause current season needle necrosis (CSNN) in true r (Abies spp.) across Europe and North America [90][91][92], necrosis on shoots of Pinus pinea L. [93] as well as necrosis on stems and needles on Pinus yunnanensis Franch. [94]. Cleary et al. [24] suggested that this endophyte is opportunistic pathogen, that due changes in climate can potentially increase its pathogenicity. Gremmeniella abietina (Lagerberg) Morelet was found only from 5 samples (115 reads). The pathogen is native to Europe and produces cankers on stems and severe damages leading to the death of its main host tree species Pinus and Picea [66,95]. Like S. sapinea, G. abietina causes crown defoliation and distortion of terminal twigs [96], leading to the assumption that it would occupies the same niches in the host tree. Due to the minor abundance of G. abietina in this study, it indicates to appear as an endophyte.

Role of yet unnoticed isolated yeasts
Previous studies found many yeast species in living or decaying plant parts and are often associated with other organisms like insects [71]. They are adapted to short-term fluctuations in abiotic conditions and to cyclic seasonal changes. Therefore they have physiological adaptations, like pigmentation and extracellular polysaccharides to survive [69]. Phylloplane yeasts also may influence the behavior, fitness, and growth of their hosts, because they produce plant hormone-like metabolites [69]. But there are only few taxa known to be endophytic and there are only few studies on conifer species, e.g. Sequoia sempervirens (D.Don) Endl. [97] P. sylvestris [98], and Pinus tabuliformis Carrière [99]. Beside lamentous wood decaying fungi, yeasts play an important role during the fungal transformation of wood, e.g. producing a partially de-ligni ed material. The efficiency in degrading of plant material differs among wood-decaying, litter-decomposing and plant-pathogenic fungi and yeasts [71]. The different decomposer groups differs in the degradation of cellulose and hemicellulose. Yeasts often form an association with basidiomycetes during the wood decay process and are able to consume lignocellulose-related sugars, usually found in tree bark, leaf litter, and rotting wood.

Mycobiome between sampling factors
Disease class The diversity indexes and species/OTU composition were same in each disease class indicating there is no differences between the main mycobiome between different health status of pine trees (healthy versus diseased) in Diplodia tip blight-diseased forest site. Based on our data, we can hypothesize that there is no health tness effect of mycobiome on symptomless trees. Similarly, the number of pathogens did not increase throughout the disease class. However, in a study by Martín et al [64], a correlation was found between host plant (elm) resistance to pathogens and to the structure of their microbial communities. This observation suggest that the pathogen is restricted by other mycobiome due to niche competition [2]. In our study the genotype of the trees was not de ned.
Instead of mycobiome, maybe the more resistant trees (de ned as healthy) differ from diseased due variation in genetics of the trees.

Sampling time
As mentioned in several studies [100][101][102], the composition of fungal communities differs in their temporal variation due to changing weather conditions, the normal cycling of the seasons or the characteristics of the host plant. In this study, we observed statistical differences between the two sampling times (June versus September), indicating the change of mycobiome between seasons. In isolate data diversity indices were indicating higher diversity in September.
However, for HTS data no differences in diversity indices were observed. M. olivacea, T. conorum-piceae and E. nigrum amounts increased in September based years and disease classes (Fig. 9). The number of S. sapinea reads (HTS data) were higher in June and in growth year 2017. As S. sapinea was more common in June, this re ects well with the production of conidial spores of this fungus. The late spring and summer months are de ned as the period of proliferation [17]. However, the moisture play here a role and rainfalls favor S. sapinea spore dispersal regardless the time of the year [17,34]. The abundance of S. sapinea was higher in older tissues, leading to the conclusion that this fungus can initiate the niche in the Scots pine woody tissue.
HTS data showed that abundance of S. sapinea was different as disease class 0 varied from disease classes 2, 4 and 5 (Fig. 9a). Similarly, disease class 3 was different from 2 and 4 (Fig. 9a). After removing outliers differences we detected that disease classes 2 and 4 (the number of reads) were statistically higher from disease classes 0 and 3 (Fig. 9b). Notable is, that after removing outliers, the average number of reads did not differ statistically between disease classes 0 and 5. Isolate data and HTS data (after removing outliers) both showed that the mean numbers of S. sapinea are similar in totally healthy trees (disease class 0, defoliation: 0-5 %) and seriously affected (disease class 5, defoliation: 81-99 %). Similarly, the isolate data could not detect differences between other disease classes. Similarly, HTS showed that disease class 2 (21-40 %) had higher abundance of S. sapinea compared to disease class 3 (defoliation 41-60 %). Indeed, perhaps the health status of the tree is not due the abundance of S. sapinea or its competition with the mycobiome. Rather the susceptibility of the tree is de ned by several abiotic and biotic factors that at the moment remain still unknown. However, in an epidemiological sense, it can be assumed that the S. sapinea accumulate symptomless in the healthy trees. In a study by Brodde et al [34], it is discussed that the fungus can culminate for 10 years in the tree before disease outbreak might happen.

Conclusions
Environmental change is altering the disturbance regimes in many regions already [104], and the risk of disturbances for forest management is increasing under the projected climate change [48,104,105]. Indeed, in the current study eight trees died during the sampling period (June-September 2018) due the extreme drought conditions observed on site. We found S. sapinea as symptomless endophytic fungi with high abundance in healthy looking Scots pine trees. This con rms the hypothesis that S. sapinea is accumulating as an endophyte in the healthy trees. Similarly, this observation can explain the sudden and rapid development of disease epidemics in several areas [34,35] and highlight the ability of the fungus to spread unnoticed [6,23,24]. To mitigate the possible future impacts of climate change, we need to understand all the factors that trigger the development of Diplodia tip blight disease epidemics. It is not known which abiotic or biotic factors actually activate the lifestyle switch of S. sapinea. This highlights the urgent need to take actions so the negative in uence of S. sapinea is restricted and the health of pine-dominated forests is secured in changing world. The research over epidemiology of S. sapinea -P. sylvestris pathosystem is urgently needed. This information should be used to improve forest health (e.g. via resistance breeding) to limit the spread of pathogenic S. sapinea.
In September 2018, eight trees sampled in June were dead, and they were replaced with new trees in the respective disease class. However, three trees could not be replaced, and thus only 27 trees were sampled in September. Altogether we collected samples from 35 different trees, which were cut from stem height of 6-8 metres (114 branch samples, Table 1).
Three randomly selected 2-year-old twigs (growth years 2017 and 2018) per tree were collected. Two twigs were stored in 8 °C and analysed with a culturebased method within the following 48 hours. One twig per tree was immediately placed in liquid nitrogen at the site and stored in -80 °C before processed further for DNA extraction and metabarcoding. From each sampling time (June or September) samples were collected for two growth years 2017 and 2018 from each disease class (Table 1) summing up total 114 samples for HTS study and cultivation study.

Culture-based Isolation, Morphological And Molecular Identi cation
The twigs were divided based on growth years, 2017 and 2018. The shoots were defoliated, washed and surface sterilized as described in Bußkamp et al. [30].
Thereafter, shoots were cut into 5 mm pieces and plated on malt yeast peptone agar (MYP) modi ed after Langer [106]. The Petri dishes were incubated for up to three weeks at room temperature (ca. 22 °C) at natural day/night cycle. They were visually checked for developing colonies on weekly basis. Emerging mycelia were sub-cultured separately on MYP. Isolated strains were assigned to mycelial morphotypes and identi ed by micromorphological characters. For identifying fungi, a ZEISS Axiostar plus microscope was used and standard procedures for fungi described in Lee and Langer [107] were followed. In addition to standard literature recommended by Oertel [108] for determination of fungi and forest diseases, the following literature was used e.g. [83,[109][110][111][112][113][114][115][116][117]. One representative strain of each morphotype was used for molecular identi cation.
Fungal DNA was extracted for molecular identi cation following the protocol of Keriö et al. [118]. Taq DNA polymerase (Microsynth) was used for PCR ampli cation of ITS regions with primer pair, ITS1-F [119] and ITS4 [120]. Brie y, the PCR protocol was as follows: 1X HR PCR Buffer, 200 µM dNTP, 0.5 µM primer 1, 0.5 µM primer 2, 100 ng template DNA, 0.2 U/µl DNA polymerase; the reaction was adjusted to 25 µl with autoclaved MQ H2O. The PCR conditions used for ITS region were 94 °C for 3 min; 30 cycles of 94 °C for 30 s, 55 °C for 1 min, 72 °C for 1 min, and 72 °C for 10 min. Possible contaminations were determined with a negative control using sterile water as template in PCR protocols. RedStain was used to con rm DNA amplicons on a 1.5% agarose gel and the visual detection was made by ultraviolet transillumination. PCR products were puri ed and sequenced using the ITS4 primer at Microsynth SEQLAB (Göttingen, Germany). The ITS sequences were extracted with an open source software utility (https://microbiology.se/software/itsx/) to extract the ITS2 subregion from the fungal nuclear ITS sequences [121]. The ITS1 and ITS2 sequences were used for BLASTN [122]) searches against GenBank/NCBI [123] to provide taxonomic identi cation. Intraspeci c ITS similarity for the sequenced fungi of 98-100% was used at species level and further con rmed the morphological identi cation.

Fungal Metabarcoding And Data Analysis
The frozen twigs were divided based on growth years 2017 and 2018. The twigs were defoliated and each sample was ground using the Mixer Mill MM 400 from Retsch GmbH with a set program of 25.0 Hz for 20 s to prevent thawing of the samples. The samples and the corresponding milling equipment were handled with liquid nitrogen throughout the entire milling process. The ground product was then stored in 1.5-ml tubes at -80℃. DNA was extracted from 50 mg of the homogenized wood sample using the "innuPREP Plant DNA Kit" (Analytik Jena AG, Jena, Germany), according to the manufacturer's instructions. DNA products were sent to Microsynth SEQLAB (Switzerland). Illumina MiSeq sequencing of amplicons were successful for 95 samples (83%) (Table 1). To sequence the internal transcribed spacer (ITS2) regions of the fungal 18S rRNA gene, two-step Nextera PCR libraries [124] using the primer pair ITS3 (5′-GCA TCG ATG AAG AAC GCA GC -3′) and ITS4 (5′-TCC TCC GCT TAT TGA TAT GC -3′) were created [125]. Subsequently the Illumina MiSeq platform and a v2 500 cycles kit were used to sequence the PCR libraries. The produced paired-end reads which passed Illumina's chastity lter were subject to demultiplexing and trimming of Illumina adaptor residuals using Illumina's real time analysis software included in the MiSeq reporter software v2.6 (no further re nement or selection). The quality of the reads was checked with the software FastQC version 0.11.8 [125]. The locus speci c ITS2 primers were trimmed from the sequencing reads with the software cutadapt v2.8 [126]. Paired-end reads were discarded if the primer could not be trimmed. Trimmed forward and reverse reads of each paired-end read were merged to in-silico reform the sequenced molecule considering a minimum overlap of 15 bases using the software USEARCH version 11.0.667. Merged sequences were then quality ltered allowing a maximum of one expected error per merged read and discarding those containing ambiguous bases. From the remaining reads the ITS2 subregions were extracted with help of the ITSx software suite v1.1.2 [121] and its included fungi database. The extracted sequences were then denoised using the UNOISE algorithm implemented in USEARCH to form operational taxonomic units (OTUs) discarding singletons and chimeras in the process. The resulting OTU abundance table was ltered for possible bleed-in contaminations using the UNCROSS algorithm. OTUs were compared against the reference sequences of the UNITE database and taxonomies were predicted considering a minimum con dence threshold of 0.5 using the SINTAX algorithm implemented in USEARCH. Rarefaction analysis were performed with the R software packages phyloseq v1.26.1 and vegan v2.5-5. Libraries, sequencing and data analysis described in this section were performed by Microsynth AG (Balgach, Switzerland). Additional BLAST searches against NCBI genebank was done manually.
For the statistical analyses the normalized data of HTS data was used. For isolates the exact number of isolates was used. All data analyses were conducted in R version 3.5.1 [127]. For each HTS sample (95) and 114 isolate samples The Shannon-Wiener index [128] and The Simpson index [129] were calculated.
The Permutational Multivariate Analysis of Variance (PERMANOVA) in VEGAN package version 2.4 [130] was used to test the statistical differences/similarities in community structure between (HTS and isolate data) samples (factors: growth year, disease class, sampling time). Permutation test (permutest.betadisper, method = bray) was used to observe and visualize the differences/similarities in dispersion between OTU composition in HTS and isolate (growth year, disease class, sampling time) in VEGAN package version 2.4 [130]. ONE-WAY-ANOVA was used to test the statistical differences/similarities in diversity indexes. S. sapinea reads were analysed with PERMANOVA and further with Kruskall-Wallis test (TukeyHSD test was used to search the differences between groups) for HTS data. ONE-WAY-ANOVA was used for isolate data (normally distributed) and TukeyHSD test was used to search the differences between groups (in disease classes). Welch Two Sample t-test was used for S. sapinea isolate data to detect differences between groups in growth year and sampling time. The statistically different OTUs (factors: growth year, disease class, sampling time) were detected with R package called 'indicspecies' [131].
The FUNGuild database v1.0 database (https://github.com/UMNFuN/FUNGuild) was used to assess the ecological and functional levels of OTUs identi ed to the species level [132]. Trophic levels included: pathogens (in FUNGuild referred to as pathotrophic fungi), saprotrophs, and mutualists (in FUNGuild referred to as symbiotrophic fungi). However, all fungal taxa were also categorized into trophic levels manually based on authors expertise and literature study. During manual curation trophic modes were assigned to endophytes, epiphytes, plant pathogens and wood-decay fungi. The plant pathogen composition between disease classes were analyzed with PERMANOVA and visualized with permutation test. Availability of data and materials  Dispersion of species of culture-based isolation data (a) and OTUs of HTS data; (b) of each sample (permution test with bray method) in each disease class.

Figure 6
Dispersion of species of isolate data (a) and OTUs of HTS data; (b) in each sample between sampling times June (black) and September (red).   Boxplot of number of Sphaeropsis sapinea HTS reads observed in disease classes (0-5) with (a) and without (b) outliers.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.