OUP user menu

Removal of phytotoxic compounds from torrefied grass fibres by plant-beneficial microorganisms

Radoslava Trifonova, Joeke Postma, Francel W. A. Verstappen, Harro J. Bouwmeester, Jan J. M. H. Ketelaars, Jan-Dirk van Elsas
DOI: http://dx.doi.org/10.1111/j.1574-6941.2008.00508.x 158-166 First published online: 1 October 2008


We aimed to select microorganisms colonizing torrefied grass fibres (TGF) and simultaneously reducing the phytotoxicity which appeared after heat treatment of the fibres. Eighty-eight bacterial strains and one fungus, previously isolated from a sequential enrichment experiment on torrefied fibres and extracts, were tested separately for their capacity to decrease phytotoxicity. Eleven of the bacterial strains and the fungus significantly reduced phytotoxicity. These organisms were checked for their ability to grow on agar containing phenol, 2-methoxyphenol, 2,6-dimethoxyphenol, 2-furalaldehyde, pyrrole-2-carboxaldehyde and furan-2-methanol as sole carbon sources. The fungus F/TGF15 and the bacterial strain 66/TGF15 were able to grow on all six compounds. Strains 15/TGE5, 23/TGE5, 43/TGE20, 56/TGF10 and 95/TGF15 grew on two to four compounds, and strain 72/TGF15 only on one compound. Strains 31/TGE5, 34/TGE5, 48/TGE20 and 70/TGF15 did not grow on any of the single toxic compounds. GC analyses of torrefied grass extracts (TGE) determined which compounds were removed by the microorganisms. F/TGF15 was the only isolate depleting phenol, 2-methoxyphenol, 2-dihydrofuranone and pyrrole-2,5-dione-3-ethyl-4-methyl. Strains 15/TGE5, 23/TGE5, 31/TGE5 and 56/TGF10, and the fungus depleted 2-furalaldehyde, 2-furan-carboxaldehyde-5-methyl, pyrrole-2-carboxaldehyde, 5-acetoxymethyl-2-furaldehyde and benzaldehyde-3-hydroxy-4-methoxy. These promising candidates for colonizing and simultaneously reducing the phytotoxicity of TGF were affiliated with Pseudomonas putida, Serratia plymuthica, Pseudomonas corrugata, Methylobacterium radiotolerans and Coniochaeta ligniaria.

  • Pseudomonas
  • Serratia
  • Methylobacterium
  • Coniochaeta ligniaria
  • 2-furalaldehyde (furfural)
  • phenol


Since the late 1970s, there has been a worldwide search for novel peat substitutes (Raviv et al., 1986; Robertson, 1993). Peat substitutes are required because of a general lack of peat moss resources in some countries and because of limitations to its use in the near future as a result of environmental constraints (Abad et al., 2001). Different compounds have been investigated as potential peat substitutes. Among them are cotton gin trash (Papafotiou et al., 2007), vermicompost (Zaller, 2006), wood fibres (Gruda & Schnitzler, 2006) and pine bark (Yu & Zinati, 2006). Recently, a new candidate for peat replacement within potting soil has also been suggested, namely torrefied grass fibres (TGF) (Trifonova et al., 2008). TGF have good water-holding capacity and they are a renewable substrate.

By exposing grass fibres to 240 °C for 1 h, stabilization of the material and protection against decay are achieved. However, phytotoxic compounds such as phenol and phenolic compounds (2-methoxyphenol, 2,6-dimethoxyphenol), 2-furalaldehyde (furfural), pyrrole-2-carboxaldehyde and furan-2-methanol are produced in the process (Trifonova et al., 2008). Furans and dehydration products thereof (2-furalaldehyde and hydroxymethyl furfural) are produced by dehydration of solubilized sugars. The release of phenolic material can arise from lignin depolymerization. Similar compounds have been found in steam-treated wheat straw exposed for different times at 210 °C, namely: extractable phenolics, tannin, 2-furalaldehyde and hydroxy-methyl-2-furalaldehyde (Castro et al., 1994). Such chemical changes incurred during torrefaction may have been responsible for the phytotoxic effect. Phenol and phenolic compounds have been found to be toxic to germination of maize (Zea mays L.) seeds (Quaratino et al., 2007) and tomato plants (Aranda et al., 2006), and some microorganisms are described to be inhibited by some of those toxic compounds (Zaldivar et al., 2000; Nilsson et al., 2005).

Following torrefaction, the grass fibres represent a microbiological vacuum, offering specific colonizable niches. This would allow a microbial community with beneficial properties, for example, improving plant growth or suppressing plant disease, to be established. However, the phytotoxicity and, by inference, toxicity to microorganisms themselves might pose limitations to this colonization process. We therefore investigated whether microorganisms can transform the aforementioned toxic compounds and eventually grow on them, thus reducing their toxic effect. In previous work, we showed that specific bacterial communities could colonize TGF, resulting in a stable microbial population (Trifonova et al., 2008). Thus far, only a few bacteria, i.e. Escherichia coli (strains LYO1 and KO11), Klebsiella oxytoca (strain P2), have been shown to be able to reductively detoxify furfural (5 and 10 mM) into furfuryl alcohol (Martinez et al., 2000; El Asli et al., 2005; Gutierrez et al., 2006). Bacterial strains from the genera Klebsiella, Enterobacter, Citrobacter, Edwardsiella and Proteus and the ascomycete Coniochaeta ligniaria can detoxify furfural (20 mM) and 5-hydroxymethyl furfural (5-HMF, 15 mM) into furfuryl alcolohol by cometabolism (Castro et al., 1994; López et al., 2004; Nichols et al., 2005). Furfural and 5-HMF are released during acid pretreatment of lignocellulose biomasses and they are toxic for microorganisms used for subsequent fermentation of lignocellulose biomasses. According to Boopathy et al. (1993), furfural and 5-HMF were not used as sole sources of carbon and energy, but they were transformed in the presence of glucose and peptone. The fungi Pleurotus spp. (Tsioulpas et al., 2002), Pycnoporus cinnabarinus and Coriolopsis rigida (Aranda et al., 2006) were found to decrease or totally remove phenol from olive mill residues.

The aim of the current study was to investigate removal of phytotoxic compounds from TGF. For this purpose, reduction of phytotoxicity was checked among 88 bacterial strains and one fungus in a seed germination assay. A selection of 11 bacteria and the fungus was further tested for their ability to grow on six potential phytotoxic compounds, as well as to deplete phytotoxic compounds present in torrefied grass extract (TGE).

Materials and methods

Microbial strains

Eighty-eight bacterial isolates and one fungus of different origin were assessed for their ability to remove phytotoxic compounds. Organisms were obtained from enrichment cultures in TGE and TGF after an initial inoculation with a suspension from fireside soil, from TGE without a soil inoculum (AR), from TGE initially inoculated with an arable soil (WK3) and from TGE initially inoculated with forest soil (FR3) (Trifonova et al., 2008). All strains were maintained as pure cultures on R2A (Difco, Detroit, MI) medium at 25 °C. Long-term storage was in glycerol stocks at −80 °C.

TGF extraction

Extracts of grass fibres were prepared according to the following protocol: 9.5 g of TGF was extracted at room temperature with 250 mL distilled water for 2 h using a magnetic stirrer at 700 r.p.m. Following extraction, the suspension was paper-filtered to remove grass fibre particles, yielding TGE. As a pH 6.5–7 is the optimum for most strains tested, pH was adjusted to 6.8 with KOH. The TGE was then filter-sterilized (Millex GV bacterial filter unit 0.22 μm; Millipore) and stored at 4 °C.

Diminishment of phytotoxicity

The 88 bacterial isolates and the fungus were grown in 20 mL liquid medium containing (1 : 1) TGE and minimal salt medium [MSM: K2HPO4 1.4 g L−1, KH2PO4 1.4 g L−1, MgSO4·7H2O 0.1 g L−1, (NH4)2SO4 1 g L−1, MnSO4·7H2O 0.001 g L−1, FeSO4·7H2O 0.01 g L−1]. The isolates were grown as single strains over 5 days at 25 °C and 180 r.p.m. using a horizontal shaker. Thereafter, the cells were removed from the TGE : MSM culture by centrifugation (9400 g, 10 min) and filter-sterilization (Millex GV bacterial filter unit 0.22 μm; Millipore). The filtrate was used in a germination assay with lettuce seeds (Lactica sativa, cultivar ‘Erika’, Enza Zaden, Enkhuizen, the Netherlands) to monitor the reduction of phytotoxic compounds; 250 μL was used per well of 24-well plates (Greiner Bio-one, CELLSTAR®, Germany). The test was performed in six wells in a randomized block design, each with five seeds. After 48 h, the number of seeds that had germinated was counted. As controls, distilled H2O, TGE : MSM (1 : 1) and TGE were used.

Growth of isolates on phytotoxic compounds

Eleven bacterial strains and the fungal strain, showing best results in the seed germination assay, were evaluated for their potential growth on phytotoxic compounds in TGF. Six model compounds, phenol (QBiogene), pyrrole-2-carboxaldehyde, 2-methoxyphenol, 2,6-dimethoxyphenol, 2-furaldehyde and furan-2-methanol (Alfa Aesar GmbH & Co. KG, Germany), were used as sole carbon sources in MSM agar medium. Agar plates with 0.1% and 0.01% of each compound were prepared in duplicate. Bacterial strains were inoculated in strikes on the plates and their growth on the agar was evaluated. MSM agar medium without any carbon was used as a negative control and MSM agar medium containing 1% glucose was the positive control.

Removal of phytotoxic compounds

Eleven bacterial strains and the fungal strain, showing best results in the seed germination assay, were cultured on TGE : MSM liquid medium at 25 °C, 160 r.p.m. Bacterial strains were cultured for 5 days and the fungus for 7 days. To determine growth, dilution plating was done on R2A media. Thereafter, the bacterial and fungal cells were removed from TGE : MSM suspension by centrifugation (9400 g, 10 min) and the supernatants were filter-sterilized (Millex GV bacterial filter unit 0.22 μm; Millipore). The resulting filtrate was assessed in the lettuce seed germination assay as described previously and analysed by GC-MS.

GC-MS analyses

Two millilitres of TGE filtrate and 2 mL of dichloromethane were mixed together and vortexed for about 10 s, until a homogeneous solution was obtained. After centrifugation (1200 g, 5 min), the upper inorganic layer was discarded and the organic layer was transferred to a clean vial and passed over a short column (Pasteur capillary pipette filled with siliconized glass wool and anhydrous Na2SO4). The clean filtrate was then collected and stored at −20 °C.

Two microlitres of the filtered organic phase was injected in splitless mode into the injection port of a gas chromatograph (5888 series II, Hewlett-Packard GMI) coupled to a MS (model 5972A, Hewlett-Packard GMI) with a Zebron ZB-5ms column (30 m × 0.25 mm I.D. × 0.25 μm film thickness) (Phenomenex). The carrier gas was He (1 mL min−1). The oven was programmed at an initial temperature of 45 °C for 1 min, with a ramp of 10 °C per min to 310 °C, and final time of 8.5 min. The injection temperature was 250 °C, and the detection temperature was 290 °C. Peaks were tentatively identified by comparison of the spectra with commercial databases as well as with reference compounds: phenol (QBiogene), pyrrole-2-carboxaldehyde, 2-methoxyphenol, 2,6-dimethoxyphenol, 2-furaldehyde and furan-2-methanol (Alfa Aesar GmbH & Co. KG).

Data from bacteria or fungus incubated in TGE : MSM were expressed as percentage of the concentration of compounds present in the TGE : MSM control without microbial growth.


To evaluate the data from the seed germination assay and the compounds from GC-MS, an anova was applied with Genstat release 9.2 (Rothamsted Experimental station, Harpenden, UK). Least significant differences were calculated for P=0.05.

The concentration of the compounds which had been measured with GC-MS in the TGE samples incubated with the different bacteria and the fungus were correlated with the germination of lettuce seeds and with growth of the bacteria in TGE : MSM by multivariate analyses with the statistical program canoco release 4.5 (Ter Braak, 1995). The concentration of the compounds was log transformed and analysed with redundancy analysis (RDA) given that the structure of the data was linear. Scaling of the figures focused on interspecies correlations. Significance of the environmental factors (percentage germination and CFU) was analysed with Monte Carlo permutation based on 499 random permutations.


Phytotoxicity diminishment

The 88 bacterial strains and one fungal strain of different origin (TGE5, TGE20, TGF10, TGF15, AR, WK3 and FR3) were tested for their ability to break down the phytotoxic compounds present in TGE. Five of 15 isolates from TGE5, namely 15/TGE5, 23/TGE5, 26/TGE5, 31/TGE5 and 34/TGE5, showed significant enhancement of seed germination (Table 1), improving germination rates by about 25% (Fig. 1). As the positive result for 26/TGE5 was obtained just once, this strain was not used for further tests. Two of 15 strains obtained from TGE20, namely isolates 43/TGE20 and 48/TGE20, one of six isolates from TGF10, namely isolate 56/TGF10, and five of 10 from TGF15, namely isolates 66/TGF15, 70/TGF15, 72/TGF15, 95/TGF15 and the fungus F/TGF15, significantly enhanced seed germination as well (Table 1). Two of 22 isolates from WK3 (12/WK3 and 19/WK3) and one of 12 from FR3 (4/FR3) diminished phytotoxicity, giving significant increases of germination rates (data not shown). No significant improvements were found for the eight AR isolates. In general, seed germination was improved about 25–35% by single strains. When the four most efficient single isolates from TGE5 were mixed and tested, this yielded a similar enhancement of seed germination as any of the single strains, i.e. 25–35% germination (Fig. 1).

View this table:

Isolates decreasing phytotoxicity in TGE assessed in a lettuce seed germination assay

StrainClosest affiliateGenBank accession numberNo. of tests with germination improvementTotal no. of tests
15/TGE5Pseudomonas putidaEU29338444
23/TGE5Serratia plymuticaEU29338034
26/TGE5Flavobacterium denitrificansEU45083512
31/TGE5Pseudomonas corrugataEU29338335
34/TGE5Stenotrophomonas maltophiliaEU29336622
43/TGE20Rhizobium radiobacterEU29337922
48/TGE20Flavobacterium denitrificansEU29337822
56/TGF10Methylobacterium radiotoleransEU29338933
66/TGF15Leifsonia xyli ssp. xyliEU29337322
70/TGF15Mycobacterium anthracenicumEU29338222
72/TGF15Agrococcus caseiEU29337633
95/TGF15Agromyces aurantiacusEU29338144
F/TGF15Coniochaeta ligniariaEU45083622
  • * Number of tests with a significant improvement of lettuce seed germination (P=0.05).


Lettuce seed germination assay with 15 isolates from TGE5 after incubation in TGE : MSM. ‘Mix’ is the combination of isolates 15/TGE5, 23/TGE5, 26/TGE5 and 31/TGE5. Three controls were applied: H2O, TGE and TGE : MSM. The least significant difference is 18 (P=0.05).

The water control consistently yielded 100% germination, while TGE yielded 0% germination, clearly showing the consistency of toxicity. The control TGE : MSM yielded a slightly higher germination percentage than the TGE control owing to dilution of TGE.

Growth of isolates on phytotoxic compounds

Strains 15/TGE5, 23/TGE5, 31/TGE5, 34/TGE5, 43/TGE20, 48/TGE20, 56/TGF10, 66/TGF15, 70/TGF15, 72/TGF15, 95/TGF15 and F/TGF15 were used for growth tests on potentially phytotoxic compounds. The fungus F/TGF15 and bacterial strain 66/TGF15 were able to grow on all six compounds when these were added at 0.01% to the agar. Strains 15/TGE5, 23/TGE5, 43/TGE20, 56/TGF10 and 95/TGF15 grew on two to four of the compounds (Table 2). Strain 72/TGF15 grew on only one compound. However, strains 31/TGE5, 34/TGE5, 48/TGE20 and 70/TGF15 did not grow on any of the single toxic compounds. All strains grew on the positive control with glucose as carbon source and did not grow on the negative control without any carbon source (data not shown).

View this table:

Growth of selected isolates on agar with potential phytotoxic compounds used as sole carbon source (concentration 0.01%)

StrainClosest affiliatePhenol2-methoxyphenol2-furaldehyde2,6-dimethoxyphenolPyrrole -2-carboxaldehydeFuran- 2-methanol
15/TGE5Pseudomonas putida101001
23/TGE5Serratia plymutica101011
31/TGE5Pseudomonas corrugata000000
34/TGE5Stenotrophomonas maltophilia000000
43/TGE20Rhizobium radiobacter101000
48/TGE20Flavobacterium denitrificans000000
56/TGF10Methylobacterium radiotolerans002001
66/TGF15Leifsonia xyli ssp. xyli221111
70/TGF15Mycobacterium anthracenicum000000
72/TGF15Agrococcus casei000001
95/TGF15Agromyces aurantiacus101001
F/TGF15Coniochaeta ligniaria221212
  • 0, no visual growth; 1, tiny visual colonies of 0.1–0.3 mm; 2, colonies of >0.3 mm.

None of the isolates grew on the compounds at 0.1%, with the exception of F/TGF15 which was able to grow on 0.1% 2,6-dimethoxyphenol. Concentrations of the model compounds were between 6.8 and 28 mg kg−1 TGF, corresponding to between 0.0007% and 0.003%. Thus, 0.1% represents a much higher concentration than present in TGF.

Removal of phytotoxic compounds

Figure 2 provides an example of the compounds present in TGE, as well as in TGE inoculated with strain 15/TGE5. Several peaks disappeared after growth of this isolate.


Chromatographic peaks present in TGE : MSM without microorganisms (control) (upper panel) and TGE : MSM incubated with isolate 15/TGE5 over 5 days (lower panel). Compounds [retention time (in min)]: (A) 2-furalaldehyde (retention time 4.57), (B) 2-dihydrofuranone (5.70), (C) 2-furan-carboxaldehyde-5-methyl (6.46), (D) phenol (6.63), (E) pyrrole-2-carboxaldehyde (7.19), (F) 2-methoxyphenol (8.41), (G) pyrrole-2,5-dione-3-ethyl-4-methyl (10.56), (H) 5-acetoxymethyl-2-furaldehyde (11.62), (I) benzaldehyde-3 hydroxy-4-methoxy (12.90), (J) benzopyran-2-one (13.51), (K) ethanone-1-(4-hydroxy-3-methoxyphenyl) (14.01), (L) 5-acetylaminomethyl-4-amino-2-methylpyrimidine (14.52).

Most of the bacteria and the fungus multiplied in TGE : MSM and CFU were detected from log 5 up to log 9.4. Data for the few isolates that, for unknown reasons, did not grow during incubation in TGE : MSM were excluded from the multivariate statistical analysis presented in Fig. 3. This figure illustrates the correlation between the germination rate of lettuce seeds, CFU microbial numbers and the 14 most relevant compounds that were detected in TGE. Most duplicate samples showed very similar values and both axes as well as the two explanatory factors (CFU, percentage germination) were significant (P=0.002). Significant reductions of several of the potentially phytotoxic compounds in TGE were measured. The ordination plot in Fig. 3 shows that the fungal isolate F/TGF15 and the bacterial isolate 56/TGF10 had the highest increase in seed germination (isolates are in the same direction as the vector percentage germination). Most compounds correlated negatively with percentage germination, indicating that a decrease of these compounds enhanced germination of lettuce seeds. Vectors of some of the compounds had an almost identical direction, i.e. compounds with retention times of 4.57, 6.46, 11.62 and 12.88 min, showing that their depletion in the different samples was similar. The most active strains in removing toxic compounds from TGE were F/TGF15, and 56/TGF10. Strains 23/TGE5, 34/TGE5, 15/TGE5 and 31/TGE5 were able to remove or deplete several compounds as well. Strains 43/TGE20, 72/TGF15 and 95/TGF15 were the least effective in removing compounds.


Ordination plot of duplicate TGE : MSM samples incubated with the different strains (points with strain numbers 15/TGE5, 23/TGE5, 31/TGE5, 34/TGE5, 43/TGE20, 56/TGF10, 72/TGF15, 95/TGF15, F/TGF15) and the control, the presence of distinct compounds (dotted vectors) and germination of lettuce seeds and growth of the bacteria in TGE : MSM as explanatory factors (filled vector). The plots were generated by RDA. Values on the axes indicate the percentage of total variation of the compounds explained by each axis. Vectors pointing in the same direction are positively correlated and those pointing in opposite directions are negatively correlated. Compounds [retention time (in min)]: (A) 2-furalaldehyde (4.57), (B) 2-dihydrofuranone (5.70), (C) 2-furan-carboxaldehyde-5-methyl (6.46), (D) phenol (6.63), (E) pyrrole-2-carboxaldehyde (7.19), (F) 2-methoxyphenol (8.41), (G) pyrrole-2,5-dione-3-ethyl-4-methyl (10.56), (H) 5-acetoxymethyl-2-furaldehyde (11.62), (I) benzaldehyde-3 hydroxy-4-methoxy (12.90), (J) benzopyran-2-one (13.51), (K) ethanone-1-(4-hydroxy-3-methoxyphenyl) (14.01), (L) 5-acetylaminomethyl-4-amino-2-methylpyrimidine (14.52), and (N) 2-furanmethanol (4.80).

Remarkably, 2-furanmethanol with a retention time of 4.80 min was absent in the TGE : MSM control (Fig. 2) and increased in some samples with bacterial growth (Fig. 3), indicating that this compound arises due to bacterial growth. This compound is likely to be a breakdown product of one of the potentially phytotoxic compounds that were degraded.

Figure 4 shows the removal of several of the compounds by the separate strains. 2-Furalaldehyde and pyrrole-2-carboxaldehyde were depleted by strains 15/TGE5, 23/TGE5, 31/TGE5, 56/TGF10 and F/TGF15 and partially by 34/TGE5 (Fig. 4, upper). Similarly, 2-furan-carboxaldehyde-5-methyl, 5-acetoxymethyl-2-furaldehyde and benzaldehyde-3-hydroxy-4-methoxy were depleted by the same strains (data not shown). The fungus F/TGF15 was the only strain that removed phenol, 2-methoxyphenol (Fig. 4, middle), 2-dihydro-furanone and pyrrole-2,5-dione-3-ethyl-4-methyl (data not shown). Ethanone-1-(4-hydroxy-3-methoxy phenyl) was only removed by strains 56/TGF10 and F/TGF15 and was partially, but significantly, decreased by strains 15/TGE5 and 23/TGE5 (Fig. 4, lower). Benzopyran-2-one was not removed by any of the tested strains, although 15/TGE5, 23/TGE5, 56/TGF10, 95/TGF15 and F/TGF15 reduced the concentration significantly (Fig. 4, lower).


Removal of compounds from TGE by the bacterial and fungal strains 15/TGE5, 23/TGE5, 31/TGE5, 34/TGE5, 43/TGE20, 56/TGF10, 72/TGF15, 95/TGF15 and F/TGF15 as compared with the control (c), which was TGE : MSM without a microbial inoculant. Scale bars with an asterisk are significantly (P=0.05) lower than the control value.

As two major groups of compounds were detected to be present in TGE, namely 2-furaldehyde and its derivatives and phenol and its derivatives, two representative compounds from those two groups were quantified in detail. The actual concentrations present in the TGE : MSM were 0.68 mg L−1 for 2-furaldehyde and 0.07 mg L−1 for 2-methoxyphenol.


We investigated a selection of strains for their ability to remove or transform phytotoxic compounds present in torrefied grass, a renewable substrate that can be used as an ingredient of potting soil. Eighty-eight strains from different origins, which had previously been isolated from enriched TGE and fibres (Trifonova et al., 2008), were tested. A lettuce seed germination assay monitored the reduction of toxicity to lettuce seeds in TGE overall. Several of the isolates, including the fungus C. ligniaria, significantly improved lettuce seed germination. The phytotoxicity-diminishing microorganisms belonged to different taxonomic groups, indicating a broad taxonomic spread of the capacity to reduce phytotoxicity across bacteria: Pseudomonadaceae, Burkholderiaceae, Enterobacteriaceae, Methylobacteriaceae, Flavobacteriaceae, Microbacteriaceae, Rhizobiaceae, Mycobacteriaceae and Xanthomonadaceae. These phytotoxicity-diminishing bacteria and the fungus C. ligniaria are interesting candidates for the recolonization of TGF, to decrease phytotoxicity as well as to create a complex and stable microbial community in the substrate.

In a previous study, the following potentially phytotoxic compounds were measured in the TGF: phenol (11 mg kg−1) and phenolic compounds such as 2-methoxyphenol (10.8 mg kg−1), 2,6-dimethoxyphenol (10.2 mg kg−1), 2-furalaldehyde (27.9 mg kg−1), pyrrole-2-carboxaldehyde (6.8 mg kg−1) and furan-2-methanol (8.6 mg kg−1) (Trifonova et al., 2008). As the above-mentioned seed germination assay is a ‘black-box’ strategy, it was not known which of the compounds were depleted by the microorganism. To obtain data on separate phytotoxic compounds, additional experiments were performed: (1) growth of the selected isolates on toxic compounds used as single carbon sources, and (2) assessment of removal of compounds present in TGE by the growth of selected isolates. For the latter, compounds were measured by GC-MS.

The fungus, which was one of the most effective strains in improving seed germination, grew on each of the six toxic compounds tested and removed most of the potentially phytotoxic compounds present in TGE, including phenol and 2-methoxyphenol, as shown by GC-MS analyses. Phenol and phenolic compounds are toxic for maize seeds (Quaratino et al., 2007) and tomato plants (Aranda et al., 2006). Thus, depletion of these compounds from TGE can be expected to increase seed germination of lettuce, which is relatively sensitive to toxic compounds. Coniochaeta ligniaria is a soil-born fungus, commonly occurring on decaying wood and wood pulp (Domsch et al., 1993). Nichols et al. (2005) isolated a strain of C. ligniaria, NRRL30616, from soil that metabolizes many inhibitory compounds such as phenol, 2-furaldehyde and 5-HMF, and reduces the concentration of 2-furalaldehyde, HMF (furan dehydration products) and acetate in corn stover hydrolysates essentially to zero.

Six of the bacteria were able to break down or decrease the concentration of several of the compounds present in TGE. These isolates, 15/TGE5, 23/TGE5, 31/TGE5, 34/TGE5, 56/TGF10 and 95/TGF15, were affiliated with Pseudomonas putida, Serratia plymutica, Pseudomonas corrugata, Stenotrophomonas maltophilia, Methylobacterium radiotolerans and Agromyces aurantiacus, respectively. Intriguingly, the first four isolates were all Gammaproteobacteria originating from the same treatment; they were isolated from the fifth enrichment step in TGE. Isolates 15/TGE5, 23/TGE5 and 56/TGF10 and 95/TGF15 were able to grow on several of the potentially phytotoxic compounds as sole carbon source. However, isolates 31/TGE5 and 34/TGE5 did not grow on any of those compounds as sole carbon source. It is possible that the model compounds, admittedly provided in relatively high concentrations compared with the natural substrate, were simply toxic to the strains, and thus limited the ability for growth. Cometabolism is another likely explanation for the discrepancy between the lack of growth on sole carbon sources and the ability to deplete such carbon sources in a mixed substrate such as TGE. In a previous study, it was observed that the strains often did not grow on formic acid and acetic acid as a sole carbon source, but did grow on a combination of the two (Trifonova et al., 2008). For S. maltophilia utilization of one substrate has been reported to have a considerable effect on the utilization of another substrate, when studying consumption of aniline and glucose (Zissi & Lyberatos, 1999).

The most promising bacterium for colonization of TGF is isolate 56/TGF10, affiliated with M. radiotolerans. It increased seed germination by about 35%. This isolate not only removed 2-furaldehyde, 2-furancarboxaldehyde-5-methyl, pyrrole-2-carboxaldehyde and 5-acetoxymethyl-2-furaldehyde, but it was also able to deplete fully ethanone 1-(4-hydroxy-3-methoxy phenyl) and 5-acetylaminomethyl-4-amino-2-methyl pyrimidine, whereas the other four bacterial strains were able only to reduce their concentration. Interestingly, this isolate was from enrichment of fibres rather than of extract.

It is still not fully clear which of the compounds present in TGF and TGE are responsible for the phytotoxicity. Phenolic compounds certainly play a role, and they are depleted by the fungus, but not by the bacteria. Therefore, the effect of the bacteria should be due to depletion of one or more of the other compounds present in TGE. The phytotoxicity of 2-furalaldehyde derivatives is unclear. 2-Furalaldehyde has been reported to be toxic to rumen microorganisms (Kyuma & Takigawa, 1988), although Castro et al. (1994) disproved this. Phytotoxicity of the TGE was found to be concentration-dependent, as phytotoxicity decreased following dilution of the extract. The combination of all the compounds and their derivatives may be causing the phytotoxicity. Thus, testing the separate compounds for their phytotoxic effect on lettuce seeds is of limited value. As germination of lettuce seed was improved at most about 35%, TGF will most likely be applied in future experiments in a combined mix with other potting soil ingredients.

In summary, the fungus C. ligniaria removed almost all potentially phytotoxic compounds present in TGE, including phenolic compounds. Additionally, four interesting bacterial isolates were selected to be capable of removing or decreasing concentrations of several phytotoxic compounds in TGE, but not the phenolic compounds. These four bacteria, i.e. P. putida, Stenotrophomonas plymutica, P. corrugata and M. radiotolerans, are all known for their antagonistic properties against plant pathogens (Zarnowski et al., 2002; Guo et al., 2007; Kai et al., 2007). Thus, the results of our research showed that a microbial community with the capacity to remove phytotoxic compounds from TGE and the potential to promote plant growth or plant health was selected. This is an important step in the development of a renewable substrate containing a beneficial plant microbial community.


This research was financially supported by the Dutch Ministry of Agriculture, Nature and Food Quality.


  • Editor: Christoph Tebbe


View Abstract