Implications of Below-Ground Allelopathic Interactions of Camelina sativa and Microorganisms for Phosphate Availability and Habitat Maintenance

Toxic breakdown products of young Camelina sativa (L.) Crantz, glucosinolates can eliminate microorganisms in the soil. Since microorganisms are essential for phosphate cycling, only insensitive microorganisms with phosphate-solubilizing activity can improve C. sativa’s phosphate supply. In this study, 33P-labeled phosphate, inductively coupled plasma mass spectrometry and pot experiments unveiled that not only Trichoderma viride and Pseudomonas laurentiana used as phosphate-solubilizing inoculants, but also intrinsic soil microorganisms, including Penicillium aurantiogriseum, and the assemblies of root-colonizing microorganisms solubilized as well phosphate from apatite, trigger off competitive behavior between the organisms. Driving factors in the competitiveness are plant and microbial secondary metabolites, while glucosinolates of Camelina and their breakdown products are regarded as key compounds that inhibit the pathogen P. aurantiogriseum, but also seem to impede root colonization of T. viride. On the other hand, fungal diketopiperazine combined with glucosinolates is fatal to Camelina. The results may contribute to explain the contradictory effects of phosphate-solubilizing microorganisms when used as biofertilizers. Further studies will elucidate impacts of released secondary metabolites on coexisting microorganisms and plants under different environmental conditions.


Introduction
Brassica mulches are used for biofumigation to reduce weed growth and crop diseases [1,2]. Plant species commonly used for biofumigation are cabbage, Indian mustard, rapeseed, white mustard, and oil radish. The biocidal activity of Brassica plants originates from the hydrolysis of their characteristic secondary metabolites with high allelopathic potential, the glucosinolates, resulting in the release of several breakdown products such as sidered in agriculture. This study addresses mutual interferences between Camelina and microorganisms under P-deficient conditions. In the first part, we direct microbial capacities of phosphate solubilization from apatite and uptake of the solubilized phosphate by C. sativa when grown in a P-depleted soil. In the second part, the glucosinolate break down by microorganisms or myrosinase and the influence of resulting products on fungal growth are studied. Focusing on a dominant soil fungus existing in the P-depleted soil, we addressed features of the entangled allelochemical-and phosphate-driven interferences between Camelina, the fungus, root-associated and added microorganisms. We also show in the third part that shoot material from mature Camelina plants almost lacking glucosinolates has no apparent fungicidal properties, yet nevertheless modulates the soil microbial diversity in a soil rich in organic matter. We hypothesize an important role of allelopathic secondary metabolites in the phosphate competitiveness of plants and microorganisms, commonly underestimated in agriculture.

Rhizotron Experiments-Uptake of Phosphate Solubilized from 33 P-Apatite
The uptake of solubilized 33 P-labeled phosphate [24,25] by C. sativa was studied in the presence and absence of the inoculants Trichoderma viride and Pseudomonas laurentiana, both applied as suspensions to the root tips. 33 P-imaging revealed first radioactivity in the leaves at 8 to 9 days after planting. Radioactive 33 P incorporation increased continuously until the shoots were harvested after 27 days, indicating microbial solubilization of 33 P from the apatite ( Figure 1A). The development of many Camelina seedlings was retarded in the presence of T. viride and P. laurentiana during early growth, and other seedlings died a few days after planting. However, most of the surviving Camelina plants caught up with the early growth retardation until harvest. Quantification of absorbed 33 P indicated that inoculation with T. viride and P. laurentiana did not enhance phosphate uptake compared with equally developed control plants ( Figure 1B). competitors. On the other hand, breakdown products of glucosinolates are inhibitory for many fungal species. Fungicidal isothiocyanates may be differentially abundant under phosphate deprivation, which changes the allelopathic potential, their interactions with microorganisms and biofumigant properties of Brassica species. The loss of microbial fitness due to nutrient deficiencies combined with plant-or microbial-derived allelochemicals and the resulting impacts on plants are seldom considered in agriculture. This study addresses mutual interferences between Camelina and microorganisms under P-deficient conditions. In the first part, we direct microbial capacities of phosphate solubilization from apatite and uptake of the solubilized phosphate by C. sativa when grown in a P-depleted soil. In the second part, the glucosinolate break down by microorganisms or myrosinase and the influence of resulting products on fungal growth are studied. Focusing on a dominant soil fungus existing in the P-depleted soil, we addressed features of the entangled allelochemical-and phosphate-driven interferences between Camelina, the fungus, root-associated and added microorganisms. We also show in the third part that shoot material from mature Camelina plants almost lacking glucosinolates has no apparent fungicidal properties, yet nevertheless modulates the soil microbial diversity in a soil rich in organic matter. We hypothesize an important role of allelopathic secondary metabolites in the phosphate competitiveness of plants and microorganisms, commonly underestimated in agriculture.

Rhizotron Experiments-Uptake of Phosphate Solubilized from 33 P-Apatite
The uptake of solubilized 33 P-labeled phosphate [24,25] by C. sativa was studied in the presence and absence of the inoculants Trichoderma viride and Pseudomonas laurentiana, both applied as suspensions to the root tips. 33 P-imaging revealed first radioactivity in the leaves at 8 to 9 days after planting. Radioactive 33 P incorporation increased continuously until the shoots were harvested after 27 days, indicating microbial solubilization of 33 P from the apatite ( Figure 1A). The development of many Camelina seedlings was retarded in the presence of T. viride and P. laurentiana during early growth, and other seedlings died a few days after planting. However, most of the surviving Camelina plants caught up with the early growth retardation until harvest. Quantification of absorbed 33 P indicated that inoculation with T. viride and P. laurentiana did not enhance phosphate uptake compared with equally developed control plants ( Figure 1B). apatite by different microorganisms and microbial consortia. Free phosphate was measured by ICP-MS analyses of PVK medium and the media after culturing the microorganisms. n = 3 with three technical replicates; significance compared to the control ***: p ≤ 0.001 (t-test). See also Figure S1. P. au + Cex: P. aurantiogriseum + Camelina extract added; P. ols + Cex: P. olsonii + Camelina extract added; P.ols: P. olsonii from Camelina seeds; P. pol R: Paenibacillus polymyxa from Camelina roots; Pant: Pantoea ananatis isolate from A. elegans; Pseudomonas spec. (BS4-DS); Rhodococcus spec. (BS2-DS); Arthrobacter spec. (BS5-DS); Paenibacillus spec. (BS1-DS); Ac. eleg. con: Actinomucor elegans consortium; T. viride: liq. T. viride grown in liquid culture; T. viride con. T. viride consortium; Pseud: P. laurentiana; CamRM: Camelina root surface microorganisms; P. aurant: P. aurantiogriseum; Control PVK: Control PVK medium.
Any attempt to reisolate T. viride from the roots, collected from the rhizotron soil after the radioactivity has subsided, failed. It is therefore likely that the fungus did not establish an association with the plants in the rhizotron experiment. The most viable and dominant bacteria from inoculated roots, which could be cultured on Sabouraud agar, were Pseudomonas species, among them probably P. putida, while Acinetobacter spec. were identified from the non-inoculated roots. Presently, further studies are necessary to verify if P. laurentiana is among the colonizing Pseudomonas species.

Pot Experiment-Growth of Camelina Seedlings on Dikopshof Soil
These results pointed to other active phosphate-solubilizing microorganisms in the Dikopshof soil that render inoculations of the root tips with T. viride and P. laurentiana suspensions ineffective and perhaps make microorganisms able to kill Camelina seedlings. Also, phosphate-solubilizing microorganisms that have already colonized the roots may be in competitive interaction with the inoculants and soil microorganisms. Therefore, pot experiments using nonsterilized and sterilized Dikopshof soil supplemented with unlabeled apatite were performed. The experiment was started with 3-day-old seedlings of Camelina containing still high amounts of glucosinolates.
The growth of surviving Camelina seedlings in nonsterilized soil was first significantly stimulated in the presence of T. viride and P. laurentiana, evidencing a contribution of the inoculants to plant growth promotion. Later, the difference between inoculated and non-inoculated seedlings was abolished ( Figure 2IA). With sterilized soil, inoculation led again to a stimulatory but not significant effect between day 3 and 7 after planting ( Figure 2IB). The stimulation turned significant between day 10 and 15. Shoot growth of the non-inoculated Camelina plants was retarded in the sterilized soil ( Figure 2IB). The results provide clues that soil microorganisms may contribute to phosphate solubilization. Their elimination by sterilization reduces the amount of solubilized phosphate. To address the question whether an Actinomucor elegans consortium possessing plant protecting properties can also promote Camelina, this consortium was applied to the plants [26] 11. Indeed, the A. elegans consortium significantly stimulated the growth of Camelina after day 13 ( Figure 2IC) in sterilized soil. Taken together, growth monitoring disclosed the existence of interfering organisms, which were eliminated by sterilization.

Abundant Microorganisms in the Dikopshof Soil, on Roots Surfaces of Agar-Grown Camelina and on the Seed Coat
Microbial plating experiments revealed that the Dikopshof soil harbored only a few microorganisms cultivable on Sabouraud agar, among them a fungus that overgrew other soil microorganisms during progressing cultures ( Figure S4). The fungus was identified as Penicillium aurantiogriseum, a plant pathogen known to release numerous phytotoxic and antimicrobial secondary metabolites [27,28]. P. aurantiogriseum developed less mycelium when grown in PVK liquid medium, compared to Sabouraud medium ( Figure S2 and Figure 1A,C). Some of the soil bacteria were identified as belonging to the genera Paenibacillus, Rhodococcus, Arthrobacter and Pseudomonas. Also, some of the microorganisms colonizing the Camelina root surface and seed coat, when seedlings were cultured on Pikovskaya (PVK) agar, were analyzed. Penicillium olsoni and Paenibacillus polymyxa were found on Camelina seed coats and seedlings ( Figure S2). On PVK, most roots were heavily covered with yellow colonies composed of several species (see below), among them a Cytobacillus firmus strain and a yet unidentified Pseudomonas species. The yellow bacterial consortium was not further investigated because it degraded over time. The results underline the variability of root-colonizing microorganisms, depending on the growth conditions. Scheme 1 gives an overview of the microorganisms present in the Dikopshof soil, and further microorganisms used for experiments in this study, here and later in the text. T-test; * p < 0.05; ** p < 0.001; *** p < 0.001; ns not significant. (II) Pot experiments with sterilized 0soil supplemented with mycelia of P. aurantiogriseum (a1,b1); P. aurantiogriseum + the T. viride consortium (a2,b2); T. viride consortium alone (a3,b3); controls without fungi (a4,b4); inoculated with A. elegans consortium (a5,b5). a: Camelina in 0-soil + apatite; b, Camelina + P containing fertilizer; Arrows: dying plants.

Abundant Microorganisms in the Dikopshof Soil, on Roots Surfaces of Agar-Grown Camelina and on the Seed Coat
Microbial plating experiments revealed that the Dikopshof soil harbored only a few microorganisms cultivable on Sabouraud agar, among them a fungus that overgrew other soil microorganisms during progressing cultures ( Figure S4). The fungus was identified as Penicillium aurantiogriseum, a plant pathogen known to release numerous phytotoxic and antimicrobial secondary metabolites [27,28]. P. aurantiogriseum developed less mycelium when grown in PVK liquid medium, compared to Sabouraud medium ( Figures  S2 and 1A,C). Some of the soil bacteria were identified as belonging to the genera Paenibacillus, Rhodococcus, Arthrobacter and Pseudomonas. Also, some of the microorganisms colonizing the Camelina root surface and seed coat, when seedlings were cultured on Pikovskaya (PVK) agar, were analyzed. Penicillium olsoni and Paenibacillus polymyxa were found on Camelina seed coats and seedlings ( Figure S2). On PVK, most roots were heavily covered with yellow colonies composed of several species (see below), among them a Cytobacillus firmus strain and a yet unidentified Pseudomonas species. The yellow bacterial consortium was not further investigated because it degraded over time.

Coculture of Camelina Seedlings with Fungi in 0-Soil Pot Experiments
To study the interaction of the two fungal consortia (T. viride, A. elegans) for supporting plant growth, older seedlings of Camelina were grown in sterilized 0-soil, a nutrient-deficient artificially mixed substrate supplemented with apatite. Seeds were directly germinated in the pots and after 7 days, when the Camelina seedlings had lost about 75% of their glucosinolates, agar plugs covered with mycelia of P. aurantiogriseum, the T. viride or the A. elegans consortium, respectively, were placed into the soil. In the arrangements with only P. aurantiogriseum, 60-70% of Camelina plants died ( Figure 2II). When plugs of the T. viride consortium were added to P. aurantiogriseum, the development of Camelina was comparable to the controls (no fungus) and to plants in pots containing only plugs with the T. viride consortium. These results indicate that P. aurantiogriseum is toxic to Camelina in sterilized 0-soil, and this toxicity can be overcome by addition of the T. viride consortium. Addition of plugs with the A. elegans consortium was stimulatory to the growth of Camelina seedlings, as already observed in the first pot experiment. Interestingly, the A. elegans consortium overgrew P. aurantiogriseum when cocultured on agar ( Figure S2(4d)). The growth behavior of Camelina in the presence of the fungi was compared to growth on sterilized 0-soil supplied with phosphate containing fertilizer. Phosphate supply resulted in a different growth behavior of Camelina in the presence of P. aurantiogriseum, as the fungus did not inhibit Camelina, indicating that the negative effect of P. aurantiogriseum on Camelina seedlings occurs only under phosphate deficiency conditions.

Coculture of Camelina Seedlings with Fungi in 0-Soil Pot Experiments
To study the interaction of the two fungal consortia (T. viride, A. elegans) for supporting plant growth, older seedlings of Camelina were grown in sterilized 0-soil, a nutrientdeficient artificially mixed substrate supplemented with apatite. Seeds were directly germinated in the pots and after 7 days, when the Camelina seedlings had lost about 75% of their glucosinolates, agar plugs covered with mycelia of P. aurantiogriseum, the T. viride or the A. elegans consortium, respectively, were placed into the soil. In the arrangements with only P. aurantiogriseum, 60-70% of Camelina plants died ( Figure 2II). When plugs of the T. viride consortium were added to P. aurantiogriseum, the development of Camelina was comparable to the controls (no fungus) and to plants in pots containing only plugs with the T. viride consortium. These results indicate that P. aurantiogriseum is toxic to Camelina in sterilized 0-soil, and this toxicity can be overcome by addition of the T. viride consortium. Addition of plugs with the A. elegans consortium was stimulatory to the growth of Camelina seedlings, as already observed in the first pot experiment. Interestingly, the A. elegans consortium overgrew P. aurantiogriseum when cocultured on agar ( Figures S2, 4d). The growth behavior of Camelina in the presence of the fungi was compared to growth on sterilized 0-soil supplied with phosphate containing fertilizer. Phosphate supply resulted in a different growth behavior of Camelina in the presence of P. aurantiogriseum, as the fungus did not inhibit Camelina, indicating that the negative effect of P. aurantiogriseum on Camelina seedlings occurs only under phosphate deficiency conditions.

Evaluation of the Microbial Capacity for Phosphate Solubilization
Microorganisms mentioned in scheme 1 were evaluated for their ability of phosphate solubilization by measuring the phosphate released from apatite using inductively coupled plasma mass spectrometry (ICP-MS). The Dikopshof soil microorganisms Pseudomonas spec., Rhodococcus spec., Arthobacter spec., Paenibacillus spec., the total washable microbial collection from root surfaces of Camelina grown on cheesecloth, single organisms from Camelina seed coat/root (P. olsonii, P. polymyxa), P. aurantiogriseum, the T. viride suspension, the T. viride consortium, the P. laurentiana suspension, the A. elegans consortium and its Scheme 1. Aforementioned and named after microorganisms addressed in this study. Liquid culture of T. viride does not conserve the consortium.

Evaluation of the Microbial Capacity for Phosphate Solubilization
Microorganisms mentioned in Scheme 1 were evaluated for their ability of phosphate solubilization by measuring the phosphate released from apatite using inductively coupled plasma mass spectrometry (ICP-MS). The Dikopshof soil microorganisms Pseudomonas spec., Rhodococcus spec., Arthobacter spec., Paenibacillus spec., the total washable microbial collection from root surfaces of Camelina grown on cheesecloth, single organisms from Camelina seed coat/root (P. olsonii, P. polymyxa), P. aurantiogriseum, the T. viride suspension, the T. viride consortium, the P. laurentiana suspension, the A. elegans consortium and its consortium member P. ananatis were incubated with apatite in PVK medium, and their ability to release phosphate was quantified by ICP-MS.
P. aurantiogriseum displayed a high capacity to solubilize phosphate ( Figure 1C). The microbial assembly collected from Camelina root surfaces had a similar capacity. The T. viride consortium and T. viride from liquid culture reached only 9-11%, P. laurentiana about 20% of the amount of phosphate solubilized by P. aurantiogriseum. With the A. elegans consortium, almost no free phosphate was measured, but phosphate could be trapped within the nexus of the cotton-like mycelium. This explanation is supported by the relatively high phosphate solubilization found with the P. ananatis isolate from the A. elegans consortium ( Figure 1C). The evaluation of phosphate solubilization confirmed the presence of phosphate-solubilizing microorganisms in the Dikopshof soil, the presence of P-solubilizing microorganisms associated with the root surfaces and with the seeds of Camelina. The phosphate solubilization activity of the root-colonizing microbial assemblies, the fungi P. aurantiogriseum and P. olsonii and further microorganisms could be visualized by culturing on PVK agar plates ( Figure S1). The unstable, yellow bacterial consortium-colonizing Camelina roots could not be used for the measurement of phosphate solubilization by ICP-MS, as culturing of the intact consortium was not possible in liquid PVK for three days. This consortium exhibited the strongest P-solubilization on PVK agar plates (see below).

Camelina s Antifungal Weapon: Glucosinolate Breakdown by Soil Bacteria
The results of the Camelina growth studies under phosphate deficiency and of the phosphate-solubilizing microorganisms prompted us to elucidate the involvement of glucosinolates in the multifaceted interactions. Glucosinolate containing extracts from 3-dayold Camelina seedlings were tested for their effects on the growth of P. aurantiogriseum. As determined by UHPLC-MS, the sterilized aqueous Camelina extract, prepared by redissolving of the dried methanolic extract in water, contained large amounts of the glucosinolates glucoarabin, glucocamelinin and 11-(methylsulfinyl) undecyl-glucosinolates ( Figure 3 and Figure S3). The glucosinolates are by far the dominant secondary metabolites in these Camelina extracts, but we could not identify glucosinolate degradation products. The extracts also contain phenolic compounds as minor constituents. The identified phenolic compounds are listed in Table S1. the first 24 hours to about 50% by Camelina glucosinolates in the presence of myrosinase ( Figure S5). Thus, both strains used for inoculation did not meet our expectation of being glucosinolate tolerant. Triplicates of the different culture supernatants were combined to one sample, technical replicates: n = 3. The amounts of the three glucosinolates left after three days of culturing is given in % of the control (no microorganisms added; 100%, calculated from peak areas). The characteristic chromatogram shows glucoarabin (GS9), glucocamelinin (GS10) and undecyl-glucosinolate (GS11). The picture below the chromatogram illustrates the strongly inhibited growth of P. aurantiogriseum when incubated in liquid PKV medium with Sinapis alba myrosinase added to the methanolic Camelina extract redissolved in water (middle) and incubation in presence of myrosinase containing aqueous extract of Camelina seedlings (right) in comparison to the control (methanolic Camelina extract redissolved in water without myrosinase, left), (n = 6).
The microorganisms collected from the Camelina root surface by bathing were also examined for glucosinolate degradation under phosphate deficiency, but these microorganisms did not degrade gluosinolates, and the addition of P. aurantiogriseum mycelium did not elicit the degradation. Thus, the root-colonizing microorganisms are not affected by glucosinolates released by the seedlings. Triplicates of the different culture supernatants were combined to one sample, technical replicates: n = 3. The amounts of the three glucosinolates left after three days of culturing is given in % of the control (no microorganisms added; 100%, calculated from peak areas). The characteristic chromatogram shows glucoarabin (GS9), glucocamelinin (GS10) and undecyl-glucosinolate (GS11). The picture below the chromatogram illustrates the strongly inhibited growth of P. aurantiogriseum when incubated in liquid PKV medium with Sinapis alba myrosinase added to the methanolic Camelina extract redissolved in water (middle) and incubation in presence of myrosinase containing aqueous extract of Camelina seedlings (right) in comparison to the control (methanolic Camelina extract redissolved in water without myrosinase, left), (n = 6).
When exposed to the above described Camelina extract in PVK medium, P. aurantiogriseum did not degrade the glucosinolates and grew like the controls, while P. olsonii grew better in presence of the extract ( Figure S2). When P. aurantiogriseum was treated with myrosinase-supplemented Camelina extracts to release isothiocyanates, the fungus did not grow at all ( Figure 3). Therefore, P. aurantiogriseum is not inhibited by intact glucosinolates, but it is sensitive to isothiocyanates or other breakdown products.
The finding that P. aurantiogriseum did not degrade glucosinolates from Camelina or other plants, except for sinigrin ( Figure S4), prompted us to examine soil-and rootcolonizing bacteria for their ability to degrade the glucosinolates. The four soil bacteria belonging to the genera Paenibacillus spec.  (Figure 3), were exposed to glucosinolate containing Camelina extract for 3 days in PKV medium. UHPLC-MS analyses of the culture media disclosed different capacities of glucosinolate degradation. The Pseudomonas, Arthrobacter and the Rhodococcus species showed highest degradation capacity, while the Paenibacillus species degraded less than 60% of the glucosinolates ( Figure 3). When Camelina seedlings were directly extracted with water instead of methanol, and the sterilized extract was immediately applied to P. aurantiogriseum cultures, fungal growth was also inhibited ( Figure S2), due to the active myrosinase in this extract. Although the P-solubilization capacity of the inoculants T. viride and P. laurentiana were unexpectedly low, the growth behavior in the presence of Camelina glucosinolates and myrosinase was studied. The T. viride strain did not degrade the glucosinolates, but growth was inhibited when myrosinase was added ( Figure S5). These findings correspond to the behavior of P. aurantiogriseum. The growth of P. laurentiana was inhibited during the first 24 h to about 50% by Camelina glucosinolates in the presence of myrosinase ( Figure S5). Thus, both strains used for inoculation did not meet our expectation of being glucosinolate tolerant.
The microorganisms collected from the Camelina root surface by bathing were also examined for glucosinolate degradation under phosphate deficiency, but these microorganisms did not degrade gluosinolates, and the addition of P. aurantiogriseum mycelium did not elicit the degradation. Thus, the root-colonizing microorganisms are not affected by glucosinolates released by the seedlings.

Cyclo (Leu-Pro) Impairs a Camelina Root Colonizing Bacterial Consortium and Represses P. olsonii
The deadly impact of P. aurantiogriseum on most of the older Camelina seedlings under the phosphate-deficient condition challenged us to identify fungal allelochemicals that may interact with the low amounts of residual glucosinolates released by these plants. Further hints for phytotoxic allelochemicals released by P. aurantiogriseum were disclosed by incubations of Camelina seedlings with fractions of the P. aurantiogriseum culture medium and by coculturing of Camelina seedlings with the fungus on Phytoagar ( Figure S2). P. aurantiogriseum produces a variety of bioactive metabolites, which are released into the surrounding environment. Their characterization was achieved by extraction of the P. aurantiogriseum culture medium with ethyl acetate followed by UHPLC-HR-MS analysis ( Figure 4, Table S2). Data processing with Metabos-cape ® allowed the identification of nine compounds of which five have been described to be produced by P. aurantiogriseum [27,28] 12: Patuline and penicillic acid, both known for their phytotoxicity, as well as the mycotoxines verrrucosidin, aurantiamine, and auranthine. Furthermore, the bicyclofarnesol derivative fuegin, a sesquiterpene was identified. In addition, we found the antibiotic diketopiperazines (cyclo-dipeptides) cyclo(L-Phe-L-Pro), cyclo(L-Leu-L-Pro), which was most abundant, and maculosin (cyclo(L-Pro-L-Tyr). As diketopiperzines were not yet identified in P. aurantiogriseum, cyclo (Leu-Pro) was tested for effects on the yellow-colored bacterial consortium that colonizes Camelina roots and which was highly efficient in phosphate solubilization. toxines verrrucosidin, aurantiamine, and auranthine. Furthermore, the bicyclofarnesol derivative fuegin, a sesquiterpene was identified. In addition, we found the antibiotic diketopiperazines (cyclo-dipeptides) cyclo(L-Phe-L-Pro), cyclo(L-Leu-L-Pro), which was most abundant, and maculosin (cyclo(L-Pro-L-Tyr). As diketopiperzines were not yet identified in P. aurantiogriseum, cyclo (Leu-Pro) was tested for effects on the yellow-colored bacterial consortium that colonizes Camelina roots and which was highly efficient in phosphate solubilization. When germinating Camelina seeds were placed on PVK agar plates and immediately treated with cyclo(L-Leu-L-Pro), the yellow-colored bacterial colonies, which were dominant on control plates without cyclo(L-Leu-L-Pro), were eliminated ( Figure 5). Also, the plant growth promoting and phosphate-solubilizing fungus P. olsonii was repressed. In contrast, P. polymyxa with a lower capacity of phosphate solubilization was not affected. Growth experiments in LB medium revealed that the yellow bacterial consortium was not inhibited by cyclo(L-Leu-L-Pro) and Camelina extracts when applied separately and incubated for 1 day. However, a combination of cyclo(L-Leu-L-Pro) and low concentrated (0.1% w/v) Camelina extract inhibited the growth significantly (p < 0.001), down to about 50% ( Figure 5). When germinating Camelina seeds were placed on PVK agar plates and immediately treated with cyclo(L-Leu-L-Pro), the yellow-colored bacterial colonies, which were dominant on control plates without cyclo(L-Leu-L-Pro), were eliminated ( Figure 5). Also, the plant growth promoting and phosphate-solubilizing fungus P. olsonii was repressed. In contrast, P. polymyxa with a lower capacity of phosphate solubilization was not affected. Growth experiments in LB medium revealed that the yellow bacterial consortium was not inhibited by cyclo(L-Leu-L-Pro) and Camelina extracts when applied separately and incubated for 1 day. However, a combination of cyclo(L-Leu-L-Pro) and low concentrated (0.1% w/v) Camelina extract inhibited the growth significantly (p < 0.001), down to about 50% ( Figure 5). toxines verrrucosidin, aurantiamine, and auranthine. Furthermore, the bicyclofarnesol derivative fuegin, a sesquiterpene was identified. In addition, we found the antibiotic diketopiperazines (cyclo-dipeptides) cyclo(L-Phe-L-Pro), cyclo(L-Leu-L-Pro), which was most abundant, and maculosin (cyclo(L-Pro-L-Tyr). As diketopiperzines were not yet identified in P. aurantiogriseum, cyclo (Leu-Pro) was tested for effects on the yellow-colored bacterial consortium that colonizes Camelina roots and which was highly efficient in phosphate solubilization. When germinating Camelina seeds were placed on PVK agar plates and immediately treated with cyclo(L-Leu-L-Pro), the yellow-colored bacterial colonies, which were dominant on control plates without cyclo(L-Leu-L-Pro), were eliminated ( Figure 5). Also, the plant growth promoting and phosphate-solubilizing fungus P. olsonii was repressed. In contrast, P. polymyxa with a lower capacity of phosphate solubilization was not affected. Growth experiments in LB medium revealed that the yellow bacterial consortium was not inhibited by cyclo(L-Leu-L-Pro) and Camelina extracts when applied separately and incubated for 1 day. However, a combination of cyclo(L-Leu-L-Pro) and low concentrated (0.1% w/v) Camelina extract inhibited the growth significantly (p < 0.001), down to about 50% ( Figure 5). These experiments point to interactions of Camelina together with its associated microorganisms on the one side, and the soil fungus P. aurantiogriseum on the other side, are accompanied by the release of allelopathic secondary metabolites by each partner for the suppression of competitors. Thus, a released allelochemical can gain new inhibitory potential, when combined with other compounds. Suppression of the yellow bacterial consortium should considerably reduce phosphate solubilization within the rhizosphere, as the consortium forms biofilms at the surface of Camelina roots ( Figure 5 and Figure S2). The results suggest that Camelina glucosinolates and cyclo(L-Leu-L-Pro) from P. aurantiogriseum have the potential to reduce the phosphate availability in the rhizophere, thereby changing the microbial biodiversity on Camelina roots and damaging the plant.

Green Manure Experiment-Preservation of Fungi and Dynamic Shifts in Microbial Biodiversity due to Loss of Camelina Glucosinolates
To emphasize the significance of Camelina glucosinolates in inhibiting fungal growth, homogenized Camelina shoots from plants in the anthesis stage were incorporated into an organic farming soil, and the effect on the soil microorganisms was assessed. We anticipated a reduction in negative impacts on fungi when glucosinolate-free shoot material was incorporated into the soil. Due to the low microbial content of the Dikopshof soil, the incorporation was carried out with an organic farming soil abundant in organic matter (Wiesengut soil), which had not been utilized for Brassica cultivation for four years [9].
Lipid fingerprinting was employed to estimate the dynamic microbial biodiversity changes. As illustrated in Figure 6, phospholipid-derived fatty acid (PLFA) profiles underwent significant changes upon application of the plant material. A dramatic increase in eukaryotic and microbial fatty acid markers was observed in comparison to the untreated soil shortly after adding the shoot material, which was further intensified during the next day (t1) This indicates (1) the incorporated shoot material contained active associated fungi and bacteria, and (2) the nutrient input may have triggered the activation and proliferation of previously inactive soil microorganisms. In the untreated soil, 15:0 anteiso, a marker for Gram (+) bacteria, was abundant [9]. However, a considerably different profile of PLFAs was established after the application of shoot material at d0, with the highest increase in branched fatty acid 16:0 14-Me. The stress marker 17:0 cyclo of Gram (−) bacteria increased [9], 18:3 showed a strong increase, and the fungal marker 18:2ω 6,9c also increased. The 15:0 anteiso marker disappeared and did not recover at any point during the experiment. Over the following weeks, the PLFA profile experienced less dramatic changes, characterized by a continuous increase of 16:0 (d0-d63) and a continuous decrease of 18:3 from d1 to d63. An increase of the fungal marker 18:2ω6,9c occurred from d1 to d28, followed by a decrease to d63. Even after 63 days, the original fatty acid composition of the untreated soil characterized by Gram (+) bacterial markers, was not reestablished. The arbuscular mycorrhizal fungi PLFA marker 16:1ω5c was low and was not significantly affected by the shoot material. The increase in the fungal marker revealed a positive effect of the almost glucosinolate-free Camelina shoot material on the abundance of fungi.
According to the clustering shown in the heat map ( Figure 6), a total of five behavioral patterns could be identified in soils enriched with Camelina shoot material regarding the changes in PLFA profiles. The first cluster includes only the marker 16:0, which shows a consistent increase in soil with time after soil enrichment with Camelina shoot material. The second cluster includes the marker 18:3 as well as the fungal marker 18:2ω 6.9c, which increase rapidly on the day Camelina shoot material is applied and maintain this trend for a further three weeks before decreasing. The third cluster includes the Gram (+) bacterial marker 15:0 anteiso, abundant only in untreated soil, decreasing as soon as the soil is enriched with Camelina shoot material. The fourth cluster includes the branched fatty acid 16:0 14-Me, which is high only when Camelina shoot material is applied but diminishes the next day. The remaining PLFA markers all belong to the last cluster and show no clear response to the application of Camelina shoot material.
increase rapidly on the day Camelina shoot material is applied and maintain this trend for a further three weeks before decreasing. The third cluster includes the Gram (+) bacterial marker 15:0 anteiso, abundant only in untreated soil, decreasing as soon as the soil is enriched with Camelina shoot material. The fourth cluster includes the branched fatty acid 16:0 14-Me, which is high only when Camelina shoot material is applied but diminishes the next day. The remaining PLFA markers all belong to the last cluster and show no clear response to the application of Camelina shoot material. Pearson correlations among examined variables showed that, except for the t0 and untreated soil PLFA fatty acid profiles, the remaining variables are highly correlated to each other, showing Pearson correlation values higher than 0.67 to 0.99 ( Figure S6). Apart from the above clustering, we can identify two main groups of fatty acids as shown in Figure 7. Pearson correlations among examined variables showed that, except for the t0 and untreated soil PLFA fatty acid profiles, the remaining variables are highly correlated to each other, showing Pearson correlation values higher than 0.67 to 0.99 ( Figure S6). Apart from the above clustering, we can identify two main groups of fatty acids as shown in Figure 7.   Table S3).

Discussion
The results provide insights on how Camelina root-colonizing microorganisms compete with soil-inhabiting ones for phosphate resources by which allelopathic secondary metabolites are of crucial importance. Glucosinolates and their breakdown products are key allelochemicals involved in the interactions as they preferentially inhibit the growth of soil fungi. The green manure experiment indicates the loss of fungicidal effects in the Wiesengut soil when plant material almost free of glucosinolates was incorporated. Subsequently, an enhanced species diversity of microorganisms including fungi was stated, most of them thought to be originated from the shoot material. This result is considerably different to impacts occurring after applying glucosinolate breakdown products to the Wiesengut soil which has deleterious impacts on fungi, formerly demonstrated by Siebers et al. [9].
Since the Dikopshof soil had no P fertilization for almost 100 years [29], P. aurantiogriseum and bacteria of the Pseudomonas, Rhodococcus, Paenibacillus, and Arthrobacter genera are adapted to low P conditions. All of them solubilize phosphate from apatite. According to the pot experiment with Dikopshof soil, the microorganisms contribute to the phosphate supply of the young Camelina seedlings, explaining why their elimination by soil sterilization led to growth stagnation of Camelina. However, the high mortality of the seedlings soon after planting in the rhizotron and pot experiments also points to toxic interactions. Bacterial conversions of Camelina glucosinolates to less toxic degradation products would weaken the plant s defense against sensitive competitors and favor, for instance, the growth of P. aurantiogriseum. In contrast, soil bacteria contribute to the suppression of fungi, when they degrade the glucosinolates to fungicidal products. On the other hand, Camelina glucosinolates have an unexpected harmful effect by acting in concert with cyclo (L-Leu-L-Pro), one of the compounds released by P. aurantiogriseum. The combination suppresses the yellow bacterial consortium, the most powerful phosphate solubilizer we could identify in this study. Thus, combinations of secondary metabolites can have other effects than the single compounds. In addition, the arsenal of P. aurantiogriseum contains many bioactive secondary metabolites directly toxic to plants, for instance penicillic acid or patulin. It must be expected that all the interfering organisms release secondary metabolites for defense, manipulation, and elimination of competitors, or contrarily, for the promotion of partners, whether as symbionts or as members of microbial assemblies. As mentioned, the quality and quantity of the compounds is influenceable by P availability. Most of the relevant research dealing with molecular effects on a plant consider one microorganism without considering interactions with others. Only more recently, interferences, such as antagonisms, are focused, while secondary metabolites are seldom included [30]. Therefore, we would like to give some examples for several, mainly nonvolatile, secondary metabolites hitherto found to be synthesized by the here addressed microbial taxa in order to open the view for the manifold possibilities of chemical interactions that can occur between the organisms.
Dikopshof soil bacteria: Diketopiperazine cyclo(L-Leu-L-Pro), released by P. aurantiogriseum, is also produced by several bacteria, for instance Streptomyces fungicidicus [31], and Pseudomonas putida [32]. Interestingly, the compound suppresses biofilm formation of some Gram (+) bacteria and has nematocidal activities [33,34]. Amino acid based bioactive secondary metabolites are synthesized by many microorganisms. Arthrobacter spec. produces the cyclic depsipeptide arthroamide, which inhibits quorum sensing [35]. Species of the genus Rhodococcus synthesize multiple bioactive secondary metabolites belonging to different chemical classes, among them nonribosomal peptides, polyketides, and the auxinrelated molecules indole-3-acetaldehyde and indole-3-acetic acid. Rhodococcus species have also a great capacity to detoxify and to degrade natural and synthetic xenobiotics [36,37]. Precursor compounds for biosynthesis are sometimes from other organisms or compounds released by one species can be converted by another one. As found with the two Dikospshof soil habitants Pseudomonas spec. and Arthrobacter spec., also Rhodococcus spec. efficiently degrades Camelina glucosinolates, while Paenibacillus spec. was less successful, almost comparable to P. aurantiogriseum.
The failed root colonization of the T. viride strain could be caused by glucosinolate degradation products generated by soil bacteria. ITCs, the best studied degradation products, are known to inhibit the root colonization of several Trichoderma strains. The isothiocyanate sensitivity of T. viride strains was first reported in 1967 [39,40]. Strainspecific positive or negative effects on plants are known from many Trichoderma species [41]. Presently, it is unclear, which degradation products (ITCs, nitriles, desulfoglucosinolates or their downstream degradation products) are generated by the Dikopshof subsoil bacteria. Some nitriles have fungicidal activity lower than ITCs but the effects of these compounds are much less investigated. Meanwhile, many further degradation products deriving from isothiocyanates, thiocyanates and nitriles have been identified with unknown bioactivities [42].
The T. viride strain used in this study has plant protecting properties but is sensitive to the glucosinolate degradation products, which explains in part the results of the pot experiments with differently aged Camelina seedlings. The positive effects of T. viride on older Camelina seedlings could be due to the release of antifungal compounds for suppressing fungal pathogens such as P. aurantiogriseum, of phytohormones and other plant growth stimulatory compounds [43]. The release of plant growth promoting compounds is also suggested for the A. elegans consortium, which is highly robust against toxins and seems not to produce toxins in detectable amounts, at least when grown in liquid culture [26].
The inhibitory effect of Camelina ITCs on P. laurentiana during short-term incubation was not known, and additional studies are necessary to elucidate long-term effects on the growth behavior. Dominant secondary metabolites of the Pseudomonas species are nonribosomal peptides (see above), bacteriocins, N-acetylglutaminylglutamine amides and ß-lactones. Inoculating young seedlings with T. viride and P. laurentiana seems to complicate the situation as root microbiomes become imbalanced because they are confronted with two further competitive species, in addition to the soil microorganisms. It is known that inoculations with phosphate-solubilizing bacteria alter or even decrease the biodiversity of rhizophere bacteria [44], which supports our assumption.
Microorganisms associated with Camelina: As described, we isolated a yellow bacterial consortium harboring a Cytobacillus firmus strain, Paenibacillus polymyxa and P. olsonii from Camelina root or seed coats. The bacteria C. firmus and P. polymyxa have plant growth promoting properties. Cytobacillus firmus, formerly Bacillus firmus [45], releases indole-3acetic acid (IAA), produces cyclic surfactant and linear lipopeptides, and nonribosomal peptides, all with antibiotic properties. Other secondary metabolites are polyketide-derived compounds and dihydroisocoumarins. P. polymyxa, well known for its plant growthpromoting properties, belongs to the core microbiome of many crops and can also occur as an endophyte [46]. On the other hand, the species can have also inhibitory effects on plants. Plant hormones (indole-3-acetic acid), the peptide antibiotics polymyxins and fusaricidins, antibiotic nonribosomal peptides, and, according to a genome study, probably polyketides, are important secondary metabolites of P. polymyxa. Polymyxins disrupt the plasma membrane integrity and causes leakage and are therefore thought to be useful to control plant pathogens. Polymyxins are inhibitory for Pseudomonas species [47,48].
P. olsonii is a plant growth promoting fungi that produces IAA and ameliorates phosphate supply by phosphate solubilization. Some strains have an endophytic lifestyle, for instance P. olsonii ML37, an endophyte in wheat leaves [49]. The secondary metabolites asperphenamate and N-benzoylphenylalanine seem to be produced only in planta [50]. The polyketide xanthoepocin, is widespread in the genus Penicillium.
Many of the known compounds are not specific for one species and, most important, strains of the same species can differ considerably in their secondary metabolite patterns. The possibilities of chemical interactions between microorganisms and with plants are exceedingly large. Moreover, parental microbial and plant secondary metabolites can be modified by exposed organisms, resulting in detoxification or degradation products, or otherwise modified parental compounds, all leading to changed bioactivities.
Only few compounds have been investigated for their functions and molecular targets in plants, for instance: gliotoxin promotes defense reactions in tomato plants [51], cyclo (L-Leu-L-Pro induces disease resistance in Arabidopsis thaliana [52] and volatiles and nonribosomal peptides of P. polymyxa are reported to be responsible for induced systemic resistance [46].
As stated by Pang et al. [53], the importance of secondary metabolites in molecular communication between interacting organisms is still underestimated, as well as functions of changed secondary metabolite accumulations under nutrient deficiency, which can worsen the negative impacts [54,55]. It is likely that all the microorganisms addressed in this study contribute to the allelochemical interactions. We believe that unbalancing of microbial communities triggers the release of toxins by alarmed species for defense of nutrient resources and habitat. Due to the complexity, dissections of further allelochemical interactions could not be addressed in the present study and disentangling the extraordinarily complex interactions needs further intensive studies. Further studies are directed to the identification of the bacterial strains found in the Dikopshof soil and microbial and plant secondary metabolite exudation when faced with released compounds by other organisms.

Plant Material
Seeds of Camelina sativa (L.) Crantz were harvested from plants grown as described by Hölzl and Dörmann [56]. Camelina plants used for soil incorporation were grown in a phytotron (160 µmol m−2 s −1 light, 25 • C and 65% humidity). Seeds were placed in soilfilled pots, watered three times a week and fertilized once a week. Shoots were harvested when plants reached BBCH scale 64-73 (late flowering).
For preparing extracts, Camelina seedlings were grown hydroponically from seeds on cheesecloth under natural conditions. Three-day-old Camelina seedlings were harvested, dried between filter paper, weighed, and extracted with methanol (1:1 w/v). The homogenate was centrifuged at 20,000 g for 5 min at 4 • C and the supernatant evaporated to dryness. The residue was dissolved in water, the resulting suspension centrifuged at 20,000× g for 5min. The supernatant was aliquoted and stored at −20 • C. Before use, thawed portions were sterile filtrated (syringe filter Excalibur, pore size 0.22 µm, Labomedic, Germany). For direct application of myrosinase containing Camelina extracts, the extraction was performed with water, the homogenate centrifuged as mentioned, the supernatant sterile filtrated and portions of the filtrate directly added to microbial cultures. The extracts were checked for the presence of intact glucosinolates using the UHPLC/MS method described.

Microorganisms Used for Rhizotron and Pot Experiments
The T. viride F-00612 consortium (collection of D.K. Zabolotny, Institute of Microbiology and Virology, National Academy of Sciences, Ukraine) was grown in liquid YEP medium for 48 h and the diluted suspension (OD600 = 0.1) used for inoculations in the rhizotron and pot experiments with Dikopshof soil (DI; 50 • 48 29 N, 6 • 57 11 E). In contrast to culturing on agar, the T. viride F-00612 consortium loses associated microorganisms, such as Enterobacter ludwigii (accession No MH915584), E. cloacae (accession No MH915583), Acinetobacter calcoaceticus (accession No MH915582), Bacillus pumilus (accession No MH915587), B. subtilis (accession No MH915585), and B. safensis (accession No MH915586), when grown in liquid medium [57]. When used as a consortium, culturing was performed on Czapek agar.
Seed coat-colonizing Actinomucor elegans (accession KM404167, registered at the DMSZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig) as strain AbRoF1 used for morphological characterization) established stable associations with Abutilon theophrasti root-colonizing microorganisms after reinoculation [58]. The resulting consortium contains among others Pantoea ananatis (DSM ID 14-714C) and the yeast Papilliotrema baii (DSM 100638) as two of the most viable microorganisms on culture plates.

Synthesis of 33 P Labeled
Apatite, Rhizotron Growth Studies and 33 P-Imaging 33 P-apatite synthesis was performed according to Wolff et al. [59]. The rhizotron experiments followed roughly the method described by Bauke et al. [24]. 33 P-apatite was mixed with P-depleted, decalcified Dikopshof soil [24,25], placed into rhizotrons and one 4-day-old Camelina seedling per rhizotron was planted. Roots of the seedlings were inoculated with 20 µL P. laurentiana and T. viride suspensions, then placed into 12 rhizotrons. Non-inoculated plants were used as controls. Rhizotrons were then placed at an angle of 45 • in a climate chamber with a day/night-length of 12 h under a slow transition and a light intensity of 320 µmol m−2 s −1 PAR. Temperatures were set to 22 • C and 18 • C, respectively, at a relative air humidity of 50%.
After 3 weeks, the experiment was finished by harvesting the plants. Plants were separated from soil, dried, and digested (6 h, 180 • C) with 4 mL 65% HNO 3 in a Loftfield apparatus (Loftfields Analytische Lösungen, Neu Eichenberg, Germany). After dilution and filtration, aliquots were mixed with 10 mL scintillation cocktail (ULTIMA Gold XR, PerkinElmer, Solingen, Germany) and the incorporated radioactivity subsequently measured by a Tri-Carb ® 3110TR Liquid scintillation counter (PerkinElmer, Solingen, Germany).

Reisolation of Inoculants from Roots after Rhizotron Experiments
Camelina roots from suspension-inoculated plants were removed from the rhizotron soil after decay of radioactivity and soil particles were carefully picked off prior to placing the roots on Czapek, LB or Sabouraud agar plates. After culturing for 14 days at 25 • C in the dark, the plates were examined for T. viride and P. laurentiana and other colonies.

Identification of the Dominant Fungus and Accompanying Bacteria Present in the Dikopshof Soil
Microorganisms present in the Dikopshof soil were isolated by serial dilution as described in Siebers et al. [9], and cultivated on different media (Sabouraud agar, Pikovskaya agar, Czapek agar). Only a few microorganisms were detected after 2 weeks. Discrete colonies of a dominant fungus were further cultured on Sabouraud agar. A representative plate was used for identification of an assumed Penicillium species by the DSMZ. For identification, the ITS rDNA sequences and morphological markers were used. The species was identified as Penicillium aurantiogriseum Dierckx (isolate 20-165) with a sequence identity of 100% to the reference (GeneBank accession No AF033476).
Single bacterial colonies were picked as template for touchdown-PCR with the bacterial 16S primers (27f AGAGTTTGATCMTGGCTCAG, 1492r(s) CGGTTGTTACGACTT) as described in Schütz et al. [60]. The PCR products were purified with the NucleoSpin Gel and PCR Clean-up Kit (Macherey& Nagel, Düren, Germany) and sequenced. The bacterial sequences were searched against the NCBI Nucleotide database using BLASTN, leading to the identification of species belonging to the genera Paenibacillus, Rhodococcus, Arthobacter, and Pseudomonas, (accession numbers: BS1_Paenibacillus_sp ON620168; BS2_Rhodococcus_sp ON620169; BS4_Pseudomonas_sp ON620170; BS5_Arthrobacter_sp ON620171). The PCR methods with the corresponding primers were also used for the identification of most viable bacteria harvested from the roots after the rhizotron experiment and culturing on Sabouraud agar. The Gene-Bank accession number of Pseudomonas sp. P16.1 is: OR123075, and of Acinetobacter sp. P17.2: OR123076 (NCBI Nucleotide database).

Setups of Pot Experiments
For the first pot experiment nonsterilized and sterilized Dikopshof soil was used. 50 g of the soil was thoroughly mixed with 2 g unlabeled, analogously synthesized apatite, filled in adequate pots, and watered until the soil was well moistened. Three Camelina seedlings (3 days old) were planted in each pot. Root tips of the seedlings were inoculated with 20 µL P. laurentiana and T. viride suspensions. The plants were cultured in a phytotron (160 µmol m −2 s −1 light, 25 • C and 65% humidity) until three leaves were developed and cotyledons became yellowish (14-18 days). The plants were watered every second day and fertilized with P fertilizer at days 8-10 when the first two leaves were unfolded. Shoot length was monitored during the entire culture at days shown in Figure 3. Each experiment was performed with 15 pots per plant species and was repeated three times.
The second, 21-day-pot experiment was performed with 25 g 0-soil mixed with 1 g apatite, filled in pots and watered as described above. Two or three Camelina seeds were placed into each pot. After seven days, 0.5 cm 2 agar plugs covered with mycelia either of the P. aurantiogriseum, the T. viride-or the A. elegans consortium were harvested and incorporated to the soil in the following manner: 5 pots-control (no fungus), 5 pots P. aurantiogriseum, 5 pots P. aurantiogriseum + T. viride consortium, 5 pots T. viride consortium, and 5 pots A. elegans, (n = 3). The plants were watered every second day and fertilized with P fertilizer after germination and then every week. The second set up was repeated with P-containing fertilizer.

Determination of Released Phosphate by Inductively Coupled Plasma Mass Spectrometry (ICP-MS)
P. aurantiogriseum, T. viride (grown in liquid culture), the T. viride consortium, the A. elegans consortium and P. olsonii from the Camelina seed coat were precultivated on Sabouraud agar. The collected washable microorganisms from the root surfaces and all other bacteria were precultured in LB liquid medium. Agar plugs (100 mg) of the fungi and 500 µL of the microorganisms grown in LB medium (OD600nm 0.1) were transferred to flasks containing 15 mL PVK (Pikovskaya medium) and cultured for 6 days at 21 • C in the dark. All incubations were performed in triplicate using material from diverse cultures of the given species. Subsequently, the cultures were filtered and the filtrates centrifuged at 20,000 rpm for 15 min at 4 • C to pellet microbial material and apatite particles. The supernatants were transferred to new tubes and again centrifuged at 20,000 rpm for 10 min. The supernatants of the triplicates were combined, aliquoted (2 mL) and stored at −20 • C until analysis. Aliquots of the PVK medium were treated in the same way as the control. Sample preparation for ICP-MS measurements was performed as follows: After centrifugation, the supernatant was transferred into a round-bottomed 15 mL PFA vial (Savillex, Eden Prairie, USA) and placed on a heating plate at 80 • C to be completely dried down in customer-designed laminar flow box in a clean-room. The dried material was redissolved in a mixture of 1 mL 68% ultrapure HNO 3 and 0.5 mL 30% H 2 O 2 . The closed vial was heated up at 120 • C on a heating plate for 1 h to dissolve any organic matter. After digestion, the solution was dried again and then redissolved in 1 mL 0.3 M HNO 3 for further dilution before the determination of P concentration started.
The P concentrations were analyzed by quadrupole inductively coupled plasma mass spectrometry (ICP-QMS, Agilent 7900, Agilent, Bremen, Germany). The measurements were performed after two hours of warm-up time to reduce the P background towards 10 ppb (10 µg/L).

Identification of Glucosinolates and Derived Compounds by UHPLC-MS
For the identification, supernatants of the centrifuged culture media and Camelina seedling extracts were used. Screening and quantification were carried out using an ultrahigh-performance liquid chromatography (UHPLC)-electrospray-mass spectrometer (MS) consisting of an ACQUITY UPLC system equipped with a Xevo TQ-S triple quadrupole mass spectrometer (Waters, Eschborn, Germany) using a Nucleodur C18 Gravity-SB column (150 × 3 mm, 3 µm; Macherey-Nagel, Düren, Germany) thermostated to 25 • C. The 59 min gradient elution was performed at a flow rate of 1 mL/min with Millipore water (Millipore GmbH, Schwalbach, Germany) with 0.1% formic acid (pH 3.0) as solvent A and acetonitril with 0.1% formic acid as solvent B (VWR International GmbH, Darmstadt, Germany, LC-MS grade). The mass spectra were recorded with an ESCi source in positive and negative full scan mode.

Identification of Secondary Metabolites Released from Penicillium Aurantiogriseum by UHPLC/MS-MS
P. aurantiogriseum was cultured for four weeks in Czapek yeast medium. The medium was filtrated and the filtrate centrifuged at 10,000× g for 15 min. The supernatant was extracted with ethyl acetate and the organic and aqueous phases evaporated to dryness. The dried ethyl acetate phase was used for compound identification. The dry residue was reconstituted with CH 3 CN/MeOH (3:1) using vortexing/sonication. After centrifugation, the supernatant was filtered using a Socorex ® borosilicate glass syringe and 0.2 µm MilliQ Millipore ® LCR filters. An aliquot of the filtrate was 10× diluted with MeOH prior to injection (1 µL).
Chromatographic separation was performed on an UltiMate 3000 UHPLC system from Thermo Fisher Scientific (Waltham, MA, USA) equipped with an Acquity BEH C18 column (2.1 × 100 mm, 1.7 µm, Waters) at a flow rate of 450 µL/min at 30 • C. The mobile phase consisted of H 2 O + 0.1% formic acid (A) and acetonitrile + 0.1% formic acid (B). A gradient elution was performed from 5 to 30% B in 8 min, increase to 100% B in 1 min and flushing at 100% B during 2 min. The UHPLC system was connected to a high-resolution QExactive orbitrap mass spectrometer (Thermo Fisher Scientific) equipped with a heated electrospray ion source. Masses were calibrated below <2 ppm accuracy using the Thermo Fischer Pierce calibration solution. Data were acquired in (+)-and (-)-ESI mode, with a mass range from m/z 100 to 800 and 35,000 resolution. MS/MS experiments were performed with a normalized collision energy of 30 eV.

Cyclo(L-Leu-L-Pro)-Tolerant and Intolerant Microorganisms from Camalia sativa Seed Coats
The fungus associated with C. sativa seeds was isolated and cultured on TSB medium. Colonies of the most abundant bacteria growing on Pikovskaya plates untreated/treated with 200 µl 1 mM cyclo(L-Leu-L-Pro) were picked and used as templates for identification as described above. The cyclo(L-Leu-L-Pro)-tolerant species was identified as Paenibacillus polymyxa (accession No. Paenibacillus_sp ON620175). The intolerant yellow colonies were composed of distinct species, thus presenting a bacterial consortium. One of the species which was most abundant in the young colonies matches several Pseudomonas species with identical sequence similarity (accession number: Pseudomonas_sp ON620172), whereby Pseudomonas aeruginosa could be excluded by PCR analysis using specific primers. Another species that proliferated from aging yellow colonies was identified by the DSMZ. A partial 16SrRNA sequence indicates the strain as most likely belonging to Cytobacillus firmus (syn. Bacillus firmus, ID 22-142; identity 99.9%). Due to the impoverishment of species and their change in abundance during culturing, it was not yet possible to comprehensively characterize this complex consortium.
The identification of the fungus associated with Camelina seed coats was performed by the DSMZ (isolate ID: . For identification, the ITS rDNA fragment and the partial βtubulin gene were used as barcodes for sequencing. The species was identified as Penicillium olsonii Bainier & Sartory (strain CBS 232.60 Ex-Typus). The sequences were compared with reference sequences (Genbank, MycoID und INDOOR; Penicillium olsonii btub EF652020, KAS6229).

Growth Behavior of the Bacterial Consortium in the Presence of Cyclo(L-Leu-L-Pro) and Camelina Glucosinolate Containing Extract
Bacterial consortium cultures were inoculated in LB medium to an OD600nm 0.1, grown at 37 • C for 150 min, followed by the addition of (a) cyclo(L-Leu-L-Pro), 1.33 mM in methanol; (b) Camelina extract, 0.1% (v/v) in methanol; and (c) cyclo(L-Leu-L-Pro), 1.33 mM in methanol and Camelina extract, 0.1% (v/v) in methanol. Growth was normalized to OD600nm at which compounds were added to the culture.

Microbial Degradation Capacity of Camelina Glucosinolates
P. aurantiogriseum was cultured on Sabouraud agar until colonies were 5 cm in diameter, after which 2-3 mg mycelium were removed from the agar and placed into flasks containing 15 mL PVK medium with 300 µL methanolic Camelina extract redissolved in water. The addition of Camelina extract was repeated after 3 days of culturing for 6 days at 21 • C in the dark. Cultures were grown in triplicates. After 6 days, the cultures were filtrated, the filtrates centrifuged at 20,000 g for 10 min and the supernatants combined, resulting in one sample. Further cultures using the same designs were supplemented with three units myrosinase (from Sinapis alba, Merck, Germany) every two days or with 500 µL of Camelina root-colonizing microorganisms (RCM), which were prepared as mentioned above. Camelina root-colonizing microorganisms without the fungus were evaluated by inoculation of 15 mL PKV medium with 500 µL of the LB precultured assembly for 6 days under the same conditions as described. All samples were aliquoted and stored at −20 • C until analysis which were performed by UHPLC-MS/MS as described above.

Green Manure Experiment with Camelina Shoot Material
Wiesengut soil (180 g) was filled in a 200 mL glass beaker. Camelina above-ground plant material (phenological growth stage: anthesis, BBCH scale: 64-73) was crushed with a homogenizer (Philips HR 2870/50 Minimixer) after adding water in a ratio 1:1 (plant/water (w/v)). Fifty grams of homogenized plant material were added to the soil resulting in a ratio of 0.28g plant material/g soil) and thoroughly mixed. The beakers were closed with a glass lid and Parafilm and stored in the dark at 21 • C. Samples were drawn directly after application (t0), and then after 1, 7, 14, 21, 28, and 63 days. Treatments were carried out in three biological replicates.

PLFA Analysis
The microbial community structure was described by analysis of the phospholipid fatty acid composition (PLFA) in the soil. Lipid were extracted according to Bligh and Dyer [61] and analyzed as described in Kruse et al. [62]. Lipids were fractionated according to their polarity by solid phase extraction [63]. Soil samples were dried for 24 h at 105 • C in the oven to determine the soil moisture content and dry weight. For PLFA analysis, the acyl groups of the separated lipids were cleaved and converted into their methyl esters by methanolysis (FAMEs), [64]. For quantification, an internal standard (100 µL of tridecanoic acid, 50 µg/mL in methanol) was used. All chemicals were of analytical grade. In total 28 PLFAs were identified in the soil samples. Specific PLFAs (or combinations thereof) were assigned to certain groups of organisms as described in Siebers et al. [9]. For plant material and other eukaryotes such as algae, PLFA 18:3 was used as marker. FAMEs were analyzed using an Agilent 7890 gas chromatograph with Supelco SP-2380 capillary column and a flame ionization detector.

Statistics
Statistical analysis of growth and phosphate data was performed with PRISM 9.0. Significant differences were calculated by use of the student's t-test. Variables were subjected to one-way analysis of variance (ANOVA). Normality was tested according to Anderson-Darling, DÁgostino & Pearson, Shapiro-Wilk and Kolmogorow Smirnov. Student's t-tests were performed in addition and p-values provided in results. Results are presented in the figures as means ± standard deviation.
To evaluate the effects of green manure application on microorganism, we used clustering and a clustered heatmap to reveal hierarchical clusters in data matrices [65]. To this end, we used heatmap and factoextra packages in R. In addition to the clustering and clustered heatmap, we also performed principal component analysis (PCA) and then both the observations and the original variables were illustrated in the principal component space [66]. In a biplot, closely aligned variables are positively correlated with each other where the stronger the correlation is when the larger the arrows are. Negatively correlated variables are aligned in opposite directions and the strength of the correlation is again measured by the magnitude of the arrows. Noncorrelated variables are typically shown by arrows that are aligned in 90 degrees to each other.

Conclusions
In conclusion, the results indicate that microorganisms already associated with the roots considerably contribute to phosphate solubilization, arising conflicts with soil microorganisms and those used for inoculation under phosphate-deficient conditions. Exemplified with P. aurantiogriseum and young C. sativa seedlings, secondary metabolites have a pivotal role in damaging or even killing of competing organisms and of those which release fungicidal compounds. Such functions are presently underestimated for most of the secondary metabolites in agriculture. However, they may contribute to the contradictory effects of phosphate-solubilizing microorganisms when used as biofertilizers in agriculture [67]. The presented study provides an idea of the importance of secondary metabolites in establishing alliances and in manipulation of accompanied organisms, which certainly include alterations of their epigenetics, transcriptomics, and proteomics. In further studies, impacts of released secondary metabolites on coexisting microorganisms and plants under different environmental condition should be considered.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/plants12152815/s1, Figure S1: Microorganisms isolated from Dikopshof soil. Figure S2: Phosphate solubilization of Penicillium aurantiogriseum in the presence of the glucosinolates and additional information. Figure S3: Identification of Camelina sativa glucosinolates. Figure S4: Allelopathic plant-microbe interaction and phosphate solubilzation. Figure S5: Effects of Camelina glucosinolates/myrosinase on inoculants. Figure S6: Pearson correlation values higher than 0.67 to 0.99Nd addition information. Figure S7: Principal Component analysis and additional information. Figure S8: The eigenvalues. Figure S9: The cos2 values. Figure S10: Illustration of the contribution (in percent) of the examined variables (untreat, t0 . . .t63) on PC1 and PC2. Table  S1: Identification of major phenolic secondary metabolites; Table S2: Penicillium aurantiogriseum compound identification. Table S3: Calculated correlations between examined variables and PC1 and PC2; Additional experiment: Test for the presence of viral plant pathogens in the mycelium of P. aurantiogriseum.

Data Availability Statement:
The datasets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.