Emerging salt marshes as a source of Trichoderma arenarium sp. nov. and other fungal bioeffectors for biosaline agriculture

Abstract Aims Sustainable agriculture requires effective and safe biofertilizers and biofungicides with low environmental impact. Natural ecosystems that closely resemble the conditions of biosaline agriculture may present a reservoir for fungal strains that can be used as novel bioeffectors. Methods and Results We isolated a library of fungi from the rhizosphere of three natural halotolerant plants grown in the emerging tidal salt marshes on the south‐east coast of China. DNA barcoding of 116 isolates based on the rRNA ITS1 and 2 and other markers (tef1 or rpb2) revealed 38 fungal species, including plant pathogenic (41%), saprotrophic (24%) and mycoparasitic (28%) taxa. The mycoparasitic fungi were mainly species from the hypocrealean genus Trichoderma, including at least four novel phylotypes. Two of them, representing the taxa Trichoderma arenarium sp. nov. (described here) and T. asperelloides, showed antagonistic activity against five phytopathogenic fungi, and significant growth promotion on tomato seedlings under the conditions of saline agriculture. Conclusions Trichoderma spp. of salt marshes play the role of natural biological control in young soil ecosystems with a putatively premature microbiome. Significance and Impact of the Study The saline soil microbiome is a rich source of halotolerant bioeffectors that can be used in biosaline agriculture.


Introduction
Sustainable agriculture requires high yields of crops, which can be achieved if chemical pesticides and synthetic fertilizers are replaced or combined with environmentally friendly biofungicides and biofertilizers (Altomare and Tringovska 2011). In such products, plant-beneficial micro-organisms positively influence the microbial community in the rhizosphere and, therefore, protect the plants as biological control agents (BCAs) and stimulate their growth as plant growth-promoting microbes (PGPMs) (Vessey 2003). Fungi are the essential members of every soil ecosystem, not only as decomposers of organic (mainly plant) matter but also as biotrophic associates of plants or other organisms (Trillas and Segarra 2009). Although most fungal-plant interactions are mutualistic (those involving mycorrhizal and endophytic fungi), numerous soil-borne diseases of plants are also caused by fungi (Redman et al. 2001). On the other hand, beneficial interactions between plants and fungi are sensitive to disturbances and require extended period to establish. To date, our understanding of these processes in native soil ecosystems remains incomplete.
Some environmental opportunistic fungi that are capable of efficiently colonizing a variety of substrates can interact with a broad range of organisms without becoming pathogenic to plants or to humans. These fungi can be particularly useful for crop protection (Harman et al. 2004). They can rapidly establish in the rhizosphere, compete with plant pathogenic fungi for the resources, and stimulate plant growth (Trillas and Segarra 2009;Harman et al. 2019). Several species of the two hypocrealean genera Clonostachys (Nygren et al. 2018) and Trichoderma (Ascomycota, Druzhinina et al. 2018) are particularly suitable for such purposes because of their versatile mycoparasitism coupled with plant-beneficial properties, including production of phytohormone-like components (Vinale et al. 2009;Cai et al. 2013) and stimulation of plant systemic resistance (Harman et al. 2004;Cai et al. 2013). The diversity of these genera is high, but so far, only a handful of species have been used as bioeffectors in biocontrol formulations Kubicek et al. 2019). However, some of these species also have potentially adverse effects like as mushroom pests (Komo n-Zelazowska et al. 2007;Innocenti et al. 2019) or even as pathogens for immunocompromised humans (Sandoval-Denis et al. 2014;Hatvani et al. 2019). Therefore, new and safe bioeffectors are required.
Undisturbed ecosystems can be natural sources of lowinput, multifunctional and renewable microbial bioeffectors. In nature, when plants germinate from their seed teguments, they associate with the microbes that exist in the surrounding environment. However, only a select subset of this community becomes associated with roots or established in the rhizosphere (Chaparro et al. 2014;Santhanam et al. 2015). In agriculture, the soil microbial communities are severely disturbed by tilling, culture, weathering and the introduction of various xenobiotics (such as pesticides and fertilizers); thus the soil microbial communities in these ecosystems frequently get reformed (Santhanam et al. 2015;Szoboszlay et al. 2017;Zhang et al. 2017;Hartman et al. 2018). For example, a welldocumented agricultural phenomenon is the high frequency of soil-born disease outbreaks in monocultured crops, which happens due to the unbalanced microbiomes rich in plant pathogenic invertebrates, fungi or bacteria (Santhanam et al. 2015;Hartman et al. 2018;Wang et al. 2019). Some newly formed natural ecosystems may resemble such affected agricultural lands in that they are young and frequently offer similar adverse conditions for microbial communities and plants. Among such ecosystems, the emerging tidal salt marshes in particular may resemble the conditions of biosaline agriculture, where saline (sea) water is used for irrigation in arid or coastal areas (Masters et al. 2007;Ayyam et al. 2019). Native plants in these conditions may be prone to diseases because of the extremely limited vegetation diversity (equivalent to monoculture), the disturbance from seawater intrusion, and the salinization of the soil surface (Li et al. 2018;Ayyam et al. 2019). Interestingly, in most of such seemingly simple natural ecosystems, even single pioneer species of plants stay healthy (Li et al. 2018;Ayyam et al. 2019).
Hence, we hypothesize that the wild plants growing in emerging tidal salt marshes may have queried the soil microbial community to assist them, namely they may have recruited some native bioeffectors as root associations in response to challenges, such as biotic (pests) and abiotic (salinity, oligotrophy and climate) challenges. In this study, we investigate the possibility of beneficial interactions between wild plants and their associated fungi in an emerging tidal salt marsh screening for native bioeffectors potentially suitable for agricultural use.

Materials and Methods
The study area and sample collection The coastal tidal flat (33°15 0 N, 120°45 0 E) in the Jiangsu province of China, spread over 6Á53 9 10 5 ha, represents the largest tidal wetland in eastern Asia (Long et al. 2016;Li et al. 2018). The costal mud flat in Dafeng Nature Reserve is the central part of this area, which keeps growing by 50-200 m per year towards the Yellow Sea. The area is under the influence of the northern subtropical monsoon climate, with a mean annual temperature of 15°C and a mean annual rainfall of 1058 mm (Long et al. 2016;Li et al. 2018;Jiang et al. 2019). Halophytic vegetation like Arundo donax (Poaceae) and Suaeda salsa (Chenopodiaceae) are the pioneer plants on this saline soil, followed by the common reed Phragmites australis (Poaceae) mixed with cogongrass Imperata cylindrica (Poaceae), which are the dominant species after the salinity drops (Li et al. 2018). Therefore, for our study, we selected three plants from three sites to sample their rhizosphere soil: P. australis (site A), S. salsa (site B) and A. donax (site C). The sampling sites are shown in Fig. 1. Nine rhizosphere soil samples located 200 m apart were collected for each plant in June 2019, as described by Cai et al. (2015). Briefly, the whole plant was carefully removed from the soil, and the bulk of the soil was removed from the roots by shaking the plant vigorously. The soil still adhering on the roots was considered as the rhizosphere soil. The rhizosphere soil samples were then stored separately in sterilized bags and transported to the laboratory on ice. Soil chemical properties, including organic matter (OM) content and available phosphorus (AP), were measured as described in our previous study (Jiang et al. 2019). Soil pH and electrical conductivity (EC) were measured in a 1 : 5 (w/v) suspension at 25°C. Soil nitrate nitrogen (NN) and ammoniacal nitrogen (AN) content was analysed with a continuous-flow analyser (AutoAnalyzer 3, Bran + Luebbe GmbH, Germany) as described previously (Cai et al. 2015;Jiang et al. 2019).

Estimation of bacterial and fungal abundance and isolation of fungi
The standard 10-fold dilution plating method was adopted for screening and isolation of bacteria and fungi from the collected soil samples. Specifically, 5 g of each soil sample was suspended in 45 ml of sterilized distilled water and was serially diluted for another 1000 folds. From the last two dilutions, 100 ll of the soil suspension was spread over the surface of LB plates (Thermo Fisher Scientific) for bacteria and PDA (BD Difco, Becton, Dickinson and Company, Franklin Lakes, NJ, USA) supplemented with 400 lg ml À1 chloramphenicol plates for fungi. The colony-forming units (CFU) on each plate were recorded separately for bacteria and fungi. Distinct fungal colonies were purified with the single-cell separation method (Zeb et al. 2019).

DNA barcoding and phylogenetic analysis
All isolated fungi were DNA barcoded using the primary (ITS1 and 2 rRNA; White et al. 1990), and the secondary (the RNA polymerase II subunit B gene, rpb2; Liu et al. 1999 (White et al. 1990). The Trichoderma strains were further sequenced for the rpb2 and tef1 with the primer pairs of fRPB2-5f and fRPB2-7cr (Liu et al. 1999) and EF1 and EF2 (O'Donnell et al. 1998) respectively. All sequences were aligned with MUSCLE that integrated in the MEGA 5 software for each locus separately and were grouped to phylotypes (Tamura et al. 2011). Unique phylotypes were subjected to the sequence similarity search tool BLASTN against the NCBI GenBank database (http://www.ncbi.nlm.nih.gov, February 2020). A species was assigned to the query strain when sequences of ITS1 and 2 rRNA were found to be identical to the type or published reference strains. Strains with the possibility of being putatively new species and ambiguous cases were assigned at the genus level. Fungi identified as Trichoderma by means of ITS1 and 2 rRNA were then further investigated by the analysis of the diagnostic fragment of tef1 and of rpb2 using a sequence similarity search against the NCBI GenBank and TrichoBLAST (www.isth.info; Kopchinskiy et al. 2005) databases. The closely related sequences found in the GenBank database were retrieved.
For phylogenetic analysis, all the obtained sequences were aligned using MUSCLE 3.8.31 integrated in ALIVIEW 1.23 (Larsson 2014). Isolates from the same soil sample sharing identical sequences of the three DNA barcode markers were treated as one fungal haplotype (genet). The sequence similarity search using NCBI BLASTN with the ITS1 and 2 and the rpb2 and tef1 sequences was performed to retrieve the vouchered sequences of the closely related strains and the identified species in the public database. The corresponding sequences of the type or published reference strains of the most closely related species were also downloaded based on the best BLAST hits. Alignment files were then generated for each marker, and the flanking areas were manually trimmed. The Bayesian information criterion was used to select the best fit model with ModelFinder (Kalyaanamoorthy et al. 2017) implemented in IQ-TREE 1.6.12 (Nguyen et al. 2015). Maximum likelihood (ML) analysis was computed with IQ-TREE. ML bootstrap proportions were computed for 1000 replicates. The obtained phylogenetic trees were viewed in FigTree v1.4.4 and edited in Corel Draw 2018.

Phenotypic assays
For the assessment of macro-morphology, fungi were inoculated on three different media-PDA, SNA (synthetic low nutrient agar, Nirenberg 1976) and CMD (4% cornmeal + 2% dextrose; Jaklitsch 2009)-and incubated at 25°C with 12 h of illumination and 12 h of darkness for 7 days. The macro-morphology of the strains was recorded with a Canon EOS 70D (equipped with a Canon 100 lm macro lens) under white light. The micro-morphology was investigated using a Leica DMi8 microscope (Leica, Wetzlar, Germany) and a cryo-scanning electron microscope (cryo-SEM, Quorum PP3010T integrated onto a Hitachi SU8010 FE-SEM, Japan). In the cryo-SEM, the fungal culture was rapidly frozen in liquid nitrogen slush, fractured at À140°C and coated with 5 nm of platinum.
Salinity and pH tolerance assays for fungi were performed in (Costar TM 96-well microplates, Corning, NY, USA). Two microlitres of spore suspension (10 8 spores ml À1 ) of each strain were inoculated into 198 ll of 30% Murashige Skoog basal salt mixture medium (MS, Sigma-Aldrich, USA) supplemented with 1% glucose (MSG), and incubated at 25°C in darkness. The salinity of the MSG medium was previously adjusted with NaCl to concentrations at 0, 0Á5, 1Á0 and 1Á5 mol l À1 . In another assay, the pH gradient was set up as pH values at 5Á0, 7Á0, 8Á0 and 9Á0. Growth was monitored as O.D.750 nm of each well every 12 or 24 h using a Spectra Max iD3 microplate reader (Molecular Devices, USA).

Fungal dual confrontation assays
The antagonistic activity of the selected Trichoderma isolates was investigated by dual confrontation assays, as described in Zhang et al.  (Derntl et al. 2017). Alternaria cf. alternata TUCIM 10217 and Macrophoma sp. TUCIM 10254 were isolated in this study (see below). Briefly, a plug of fresh culture (6 mm) of an opponent fungus was placed 1 cm from the edge of the PDA plate (9 cm diameter) and incubated at 25°C in darkness for 24 h. Then a similar culture plug of the Trichoderma sp. was placed on the opposite edge of the same plate. The fungi were allowed to grow under the above incubating condition for 14 days, and the fungal combat on each plate was recorded with a Canon EOS 70D camera.

Plant growth promotion experiment
To analyse the growth promotion effect of the selected Trichoderma spp. on plant, a pot experiment was carried out with tomato seedlings (Solanum lycopersicum L. cv. ). Three seedlings, all 3 weeks old, were planted in each pot containing 300 g of a mixture (w/ w = 1 : 1) of vermiculite (1-3 mm) and perlite (1-3 mm) at a pH of 6Á0. The salinity of the growth substrate was adjusted by adding NaCl to 0Á5% and 0Á75%, representing medium and high salinity stress conditions, respectively, and using a 0% NaCl group as the control. Three millilitres of Trichoderma spore suspension (10 8 spores ml À1 ) were inoculated to the roots in each pot. Ten millilitres of 10% MS irrigation was applied every 2 weeks. The plants were allowed to grow at 25°C under cycled illumination conditions (light : darkness = 16 : 8) for 6 weeks. At the end of the experiment, data regarding plant growth and health, including plant height, fresh and dry biomass, and the SPAD value for measuring the leaf chlorophyll content, were recoded for each seedling (N = 12 per each treatment). Root development was measured using a root scanner (Epson Perfection v700 Photo, Seiko Epson, Japan), as described previously (Cai et al. 2013).

Statistical analysis
The means and the standard deviations of the data were calculated using PASW 18.0 (IBM Corporation, Chicago, IL, USA). Multiple comparisons were performed using the analysis of variance (ANOVA) and Duncan's multiple range test (P = 0Á05) integrated in PASW 18.0. The heatmap was plotted in R v3Á2.2.

Study area and sampling sites
The study area, Dafeng Nature Reserve, is located at the east costal region of China, which faces the Yellow Sea. The area consists of the emerging salt marshes (Solonchak, IUSS Working Group WRB, FAO 2015) that formed 50 years ago and is still growing towards the sea due to the large amount of sediment carried by the Yellow River and the Yangtze River (Li et al. 2018). The reserve is a typical coastal mud flat, characterized by a gentle slope formed with successive saline soil. The land offers a unique opportunity to study hydromorphic soil development, vegetation succession and microbiome assembly (Long et al. 2016). The natural vegetation succession in this area starts with the giant cane A. donax (Poaceae) close to the sea shore, followed by the highly halotolerant native red plant S. salsa (Chenopodiaceae), and the cosmopolitan fire-adapted grass I. cylindrica (Poaceae). In the relatively mature ecosystems several kilometres inland, the marshes are colonized by the common reed P. australis (Poaceae). Large colonies of P. australis, S. salsa and A. donax, occupying several square kilometres, undisturbed by human activities, were selected as sampling sites A, B and C respectively (Fig. 1).
The comparative analysis of soil properties revealed high pH (ca. 8Á5) at all three sites, and no difference in ammoniacal nitrogen (AN) or available phosphate (AP) between the three sampling sites (P > 0Á05, Table 1). The nitrate nitrogen (NN) and organic matter (OM) slightly increased with increased distance from the sea (P < 0Á05), but remained comparable. However, the soils in the three sites had very different salinization and electrical conductivity (EC, an indicator of the total salinity of soil) values, with the lowest salinity at site A and the highest at site C ( Fig. 1 and Table 1).

Fungal abundance correlates with soil properties
The abundance of cultured bacteria and fungi decreased significantly from site A to site C (Fig. 2a). The most closely related soil properties to microbial abundance were EC, OM and NN; on the other hand, pH values, AN and AP were not clearly related to it (Fig. 2b). Specifically, both bacterial and fungal abundances were positively correlated with OM and NN, and were negatively correlated with soil EC values.
Statistically significantly different values are labelled with different letters (N = 9, ANOVA, P < 0Á05). The bold font highlights the statistically significantly largest values among the sites. We isolated 50 fungal strains from the rhizosphere of P. australis (site A) and 50 from the rhizosphere of S. salsa (site B), and only 16 from the rhizosphere of A. donax (site C) (Fig. 1, Table 2). DNA barcoding based on the internal transcribed spacers (ITS1 and 2) of the ribosomal RNA gene cluster revealed in total 38 fungal phylotypes. Of these, 65 isolates could be reliably identified by the sequence similarity to the vouchered isolates deposited in public databases and confirmed by taxonomic literatures (Table 2), and 20 more isolates were identified after sequencing additional DNA barcoding markers, such as fragments of tef1 and rpb2 genes ( Table 2). In total, 85 isolates were thus identified by species, but the taxonomic position of 31 additional isolates (like four Trichoderma spp., Coniothyrium sp. TUCIM 1024, a new hypocrealean strain TUCIM 10250, and others, Table 2) remained undefined, suggesting the existence of putatively new taxa or species that have no corresponding DNA barcodes in public databases.
Although all the plants sampled appeared healthy, the fungi isolated from rhizosphere of S. salsa (site B) were predominantly species that are known to be plant pathogenic (Macrophoma sp., Alternaria spp., F. equiseti, and others; Table 2). Fungi isolated from the two other sites, site A and site C, were ecologically equally versatile, although the habitats differed in salinity. Thus, the rhizosphere of P. australis (site A) was dominated by a putatively new phylotype of Trichoderma, T. sp. TUCIM 10301, followed by four other putatively new Trichoderma spp., T. asperelloides and T. caerulescens, but also the two other mycoparasitic fungi (Coniothyrium sp. TUCIM 10243 and Paraconiothyrium estuarinum TUCIM 10279), and a variety of common saprotrophic fungi, such as species of Aspergillus, Penicillium (Eurotiales), and some common Mucoromycotina (Mucor spp., Mortierella spp.; Table 2, Fig. 1). Similarly, a mixture of mycoparasitic and saprotrophic fungi was recovered from the samples of site C. As this site is located near the costal line, we also found a few aquatic or marine fungi there (Phaeosphaeria spartinae from Pleosporales and hypocrealean Paracremonium binnewijzendii). Interestingly, the diversity recovered from the invasive environmentally opportunistic plant species, the common reed and the giant cane, was rich in the environmental opportunistic species of fungi, that are, Trichoderma spp., Aspergillus spp. and Mucor spp.
Two Trichoderma strains tolerate high salinity and alkaline pH Trichoderma spp. are well-recognized plant-beneficial fungi that are used as bioeffectors in biofungicides for controlling fungal diseases in crops (biocontrol) and/or in biofertilizers for plant growth promotion (see review in Druzhinina et al. 2011). The diversity of the isolated Trichoderma strains in this study consisted of seven phylotypes (Table 2), of which two could be reliably identified to the species level (T. asperelloides and T. caerulescens; see below) and five were putatively new taxa. Therefore, in order to select possible bioeffective strains that can be used under the conditions of biosaline agriculture, we first tested the tolerance of the strains to high salinity  and alkaline pH, the parameters that represent or extend the conditions of their native habitat. One strain per each of the seven phylotypes was randomly selected for these tests. Based on the results given in Table 1 (that the salinity of the three sites ranged from 0Á36 to 2Á3%, with pH consistently around 8Á4-8Á5), four gradients of each stress factor were set (Fig. 3). As shown in Fig. 3, strain TUCIM 10320 grew significantly more in the presence of 0Á5 mol l À1 NaCl (2Á9% NaCl, close to the natural salinity of site C) and 1Á0 mol l À1 NaCl, compared to the other strains and to itself when grown under nonsaline conditions (ANOVA, P < 0Á05). Therefore, we assume that this strain is halophilic, while the others are halotolerant. Several strains were sensitive to NaCl (Fig. 3). Furthermore, the growth of strains T. sp. TUCIM 10301 and T. sp. TUCIM 10329 was significantly greater than the growth of other strains T. sp. TUCIM 10328, T. sp. TUCIM 10323, T. sp. 10325 and T. caerulescens TUCIM 10321 under the condition of 0Á5 mol l À1 NaCl. However, the growth of all the strains tested declined dramatically when the NaCl concentration reached 1Á5 mol l À1 (ca. 8%).
The halophilic strain T. asperelloides TUCIM 10320 best adapted to alkaline pH values, followed by strains T. sp. TUCIM 10301 and T. sp. TUCIM 10328. The other Trichoderma spp. strains, TUCIM 10323, TUCIM 10325, TUCIM 10329 and T. caerulescens TUCIM 10321, showed comparatively weaker growth than the above three strains under the test conditions. Based on their adaptability to the two stress factors tested, strains T. asperelloides TUCIM 10320 and T. sp. TUCIM 10301 were selected for subsequent experiments.

Phylogenetic and phenotypic analysis reveals a new Trichoderma species
To reveal the taxonomic position of the bioeffective T. sp. TUCIM 10301, which by far dominated our culture  (Jaklitsch et al. 2013). The most similar sequences were from the strain HZA5 of a recently described species T. dorothopsis (deposited as Trichoderma sp. AA-2019, Tomah et al. 2020), which was also isolated from soil in China, and which shared a 98.77% rpb2 (GenBank: MH647795) and a 97Á52% tef1 (GenBank: MK850827) phylotype with TUCIM 10301 (E-value was equal to zero for both comparisons). The similarity of strain TUCIM 10301 to the most closely related defined species Trichoderma dingleyae and Trichoderma taiwanense was, respectively, 97Á29 and 97Á12% for rpb2, and 85Á53 and 91Á06% for tef1. This indicates that TUCIM 10301 belongs to the Trichoderma Section of this genus. The taxonomy report obtained from this search revealed that besides T. dorothopsis, T. dingleyae and T. taiwanense, the query strain was also related to T. sp. strain IQ 11 (namely TUCIM 4882 from South America) and T. sp. TUCIM 5745 from South-east Asia. The ML phylogram (Fig. 4a) constructed with rpb2 sequences demonstrated that the five isolates, formed a statistically supported clade separate from the most closely related genetic neighbours (T. dorothopsis, T. dingleyae, T. taiwanense, T. sp. TUCIM 5745 and T. sp. TUCIM 4882). Similar tree topology supporting the presence of this clade was also obtained for the tef1 phylogenetic marker (Fig. 4a). Thus, the isolates represented by T. sp. TUCIM 10301 met the criteria of the genealogical concordance phylogenetic species recognition concept (Taylor et al. 2000), as they form distinct clades on the phylograms constructed based on the two unlinked loci (rpb2 and tef1) and also have a unique ITS1 and 2 rRNA phylotype. Therefore, we recognize it as a new species described below as T. arenarium sp. nov. Strain TUCIM 10320 was found to be identical to the type strain of T. asperelloides G.J.S. 04-111 (Samuels et al. 2010) when the rpb2 and tef1 loci were used, as shown in Fig. 4b, and thus it was identified as T. asperelloides.
Trichoderma arenarium sp. nov. and T. asperelloides combat a variety of plant pathogenic fungi In order to investigate whether the isolated Trichoderma strains have potential in biocontrol of plant pathogens, dual confrontation assays were done between the two Trichoderma spp. (TUCIM 10301 and 10320) and five phytopathogenic fungi. We used two fungi isolated in this study (Alternaria cf. alternata TUCIM 10217 and Macrophoma sp. TUCIM 10254) and three other reported pathogenic fungi, F. odoratissimum TUCIM 4848, R. solani TUCIM 3753 and Pestalotiopsis fici TUCIM 5788. The results showed that T. arenarium sp. nov. TUCIM 10301 and T. asperelloides TUCIM 10320 efficiently combated and overgrew the two sympatric fungi as well as R. solani TUCIM 3753 (Fig. 5). However, these two Trichoderma strains both showed weaker antagonism against P. fici TUCIM 5788. T. asperelloides TUCIM 10320 could not combat P. fici and remained in a 'deadlock' stage (where the growth of one fungus is limited by another; see more about fungal 'deadlock' in Zhang et al. 2019  conidia ring surrounding the P. fici colony. As for F. odoratissimum TUCIM 4848, T. arenarium sp. nov. TUCIM 10301 overgrew on it partially, while T. asperelloides TUCIM 10320 completely combatted this fungus and formed dense conidia above it. This response is relatively rare for Trichoderma spp. .
Trichoderma arenarium sp. nov. and T. asperelloides promote plant growth in conditions of high salinity To test whether the obtained T. arenarium sp. nov. and T. asperelloides strains can be used for plant growth promotion in biosaline agriculture, a pot experiment was carried out with a model plant (tomato, S. lycopersicum L.) under three different salinity conditions (0, 0Á5 and 0Á75% NaCl). The evaluation of tomato seedlings (Table 3) showed that the inoculation of T. arenarium sp. nov. TUCIM 10301 and T. asperelloides TUCIM 10320 significantly (ANOVA, P < 0Á05) promoted the biomass and the height of the seedlings compared to the control at both medium (0Á5%) and high (0Á75%) salinity conditions, as well as at the nonsalinity condition. Specifically, the Trichoderma inoculations increased the dry weight of the seedlings by 30-81% under the salty conditions and by 41-107% under the nonsalinity conditions relative to the Trichoderma-free control. Moreover, the effect of the Trichoderma inoculations on SPAD reads (which measure the relative chlorophyll content in leaves) suggested that Trichoderma played a role in eliminating the chlorophyll reduction that normally caused by high salinity. As salinity has a severe negative effect on roots (Ayyam et al. 2019), we also used a root scanner to evaluate root development in a detailed way. The results (Fig. 6) showed that the Trichoderma inoculations significantly (ANOVA, P < 0Á05) promoted the total root length and the number of root tips compared to the control, while correspondingly, the root diameters were Statistically significantly different values are labelled with different letters (N = 12, ANOVA, P < 0Á05). The bold font highlights the statistically significantly largest values among the treatments.

Discussion
Soil arguably harbours the world's most diverse microbiome (Jansson and Hofmockel 2020). Plants anchor in the soil by their roots and recruit particular microbial taxa from the soil marketplace as potential partners (Turner et al. 2013;Santhanam et al. 2015). Our understanding on this process and the factors governing behind is very limited for most microbial taxa. As for fungi, besides the interactions of plants with mycorrhizal and phytopathogenic fungi (which have been frequently studied), the mechanisms driving the nonpathogenic fungi in rhizosphere remain unknown (Redman et al. 2001;Harman et al. 2019). In this study, by screening the cultured fungi in the rhizospheres of several pioneer plant species found in the emerging tidal salt marshes, we inadvertently recapitulated a common biological question: why do some Trichoderma species preferentially enrich in rhizosphere, or even colonize on roots? Similarly to what frequently happens in agriculture (Trillas and Segarra 2009;Szoboszlay et al. 2017;Hartman et al. 2018), the perennation of some Poaceae species colonizing the tidal salt marshes results in an accumulation of some specific phytopathogenic fungi (e.g. Macrophoma sp., Alternaria spp., and Fusarium spp.) in their rhizosphere. Consequently, Trichoderma spp., as a mycoparasite (Kubicek et al. 2011;Druzhinina et al. 2018), may trace fungi, including phytopathogenic ones in such ecosystems, thus becoming root associated. Although we are not able to exclude other possible factors attracting Trichoderma spp. to roots, it could be evidence of biocontrol happening in nature. Throughout evolutionary history, native wild plants growing in this ecosystem may have been querying their soil microbial community to assist them in dealing with potential challenges (like the phytopathogen accumulation here). And this may help us empower crops to perform the same by screening the native bioeffectors for the specific plants or for the established ecosystem. The results of this study confirm the initial hypotheses and show that some of the isolated strains can be used as bioeffectors in agriculture, since the Trichoderma spp. found in the sample area significantly promoted plant growth under various salinity conditions and were able to antagonize the sympatric and allopatric plant pathogenic fungi. Besides, five of the seven Trichoderma phylotypes found could be putatively recognized as new species, suggesting that there may be a huge potential source of new microbial taxa hidden in these young extreme ecosystems. Similar observations were noted in several other studies of marine Trichoderma (Gal-Hemed et al. 2011;Vacondio et al. 2015), which also detected putatively new phylotypes.
The new species T. arenarium sp. nov., which is described here having the closest sibling T. dorothopsis (type strain HZA5, not isolated in this study but also found in the soil of the Yangtze River basin, Tomah et al. 2020), may be a local species associated with the coastal soils in this region, as no other strain records were found in other locations so far. However, as the present sampling land is formed from the large amount of sediment carried by the Yellow River and the Yangtze River, the Trichoderma strains may have also been introduced from upstream habitats, like the Gobi Desert or the Loess Plateau, where the massive sediments of the Yellow Sea originate.
Saline soils are widespread all over the world, accounting for 7-8% of the Earth's surface (Artiola et al. 2019). Coastal saline soil, such as that found in salt marshes, represents a subclass of saline soils, and is recognized as an important potential land resource for agricultural development (Long et al. 2016;Ayyam et al. 2019). However, crop growth in such areas is usually very limited, due to the high salinity and low nutrient availability in the soil (Ayyam et al., 2019). Regardless of breeding salttolerant plant cultivars, in this study, we showed that a possible alternative is to identify bioeffectors from local or similar ecosystems for use in saline soil agriculture. The work of Hingole and Pathak (2016) also highlighted the saline soil microbiome as a rich source of halotolerant bioeffectors. In our case, the S. salsa rhizosphere was found to be unsuitable as a source of novel bioeffective strains, as it maintained a very different mycoflora. Compared to the isolates from the P. australis and A. donax samples, the screening of the S. salsa rhizosphere yielded mainly phytopathogenic fungi, suggesting the possibility of plant-specific selection in fungal enrichment. Moreover, among the fungivorous fungal genera, Trichoderma is the largest taxon, with many ubiquitously distributed species (Kubicek et al. 2011;Druzhinina et al. 2018;Kubicek et al. 2019). Most species (80%) Friedl and Druzhinina 2012;Kubicek et al. 2019) have been found to be relatively rare, but a few dozens of species are known to be present in soils all over the world and are considered environmental opportunists with cosmopolitan distribution. In the present work, the most frequent Trichoderma species was T. arenarium sp. nov., followed by several other species within the section Trichoderma, rather than the T. harzianum sensu lato group that frequently found in soil (Druzhinina et al. 2010;Chaverri et al. 2015), indicating that T. arenarium sp. nov. is well adapted to the local niche. Therefore, the study also demonstrates that native bioeffectors may be more effective than the allopatric strains in developing local biocontrol products. As for coastal saline lands, biosaline agriculture offers a solution to the imbalance between the limited arable land and the growing human population by using salt-affected soil and water (Ayyam . This requires the selection of suitable halophytes not only for the plants to be grown but also for the possible associated microorganisms 2020.