1
Integrated Metabolomics and Morphogenesis Reveals 1
Volatile Signaling of the Nematode-Trapping Fungus 2
Arthrobotrys oligospora 3
4
Bai-Le Wang,1 Yong-Hong Chen,
1 Jia-Ning He,
1 Hua-Xi Xue,
1 Ni Yan,
1 Zhi-Jun Zeng,
1 5
Joan W. Bennett,2 Ke-Qin Zhang,
1,* Xue-Mei Niu
1,* 6
7
1State Key Laboratory for Conservation and Utilization of Bio-Resources & Key Laboratory for 8
Microbial Resources of the Ministry of Education, School of life Sciences, Yunnan University, 9
Kunming, 650091, People’s Republic of China 10
2Department of Plant Biology, Rutgers University, New Jersey 08901, United States of America 11
12
*Corresponding author (Tel: 86-871-65032538; Fax: 86-871-65034838: E-mail: [email protected] or 13
15
Running title: Flexible tactics of A. oligospora to trap nematodes 16
17
AEM Accepted Manuscript Posted Online 16 February 2018Appl. Environ. Microbiol. doi:10.1128/AEM.02749-17Copyright © 2018 Wang et al.This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license.
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
2
ABSTRACT 18
The adjustment of metabolic patterns is fundamental to fungal biology and plays vital roles in adaption 19
to diverse ecological challenges. Nematode trapping fungi can switch lifestyles from saprophytic to 20
pathogenic by developing specific trapping devices induced by nematodes to infect their prey as a 21
response to nutrient depletion in nature. However, the chemical identity of the specific fungal 22
metabolites used during the switch remains poorly understood. We hypothesized that these important 23
signal molecules might be volatile in nature. GC-MS was used to carry out comparative analysis of 24
fungal metabolomics during saprophytic and pathogenic lifestyles of the model species Arthrobotrys 25
oligospora. Two media commonly used in research on this species, corn meal agar (CMA) and potato 26
dextrose agar (PDA), were chosen in this study. The fungus produced a small group of volatile furanone 27
and pyrone metabolites that were associated with the switch from saprophytic to pathogenic stages. A. 28
oligospora grown on CMA tended to produce more traps and employ attractive furanones to improve 29
utilization of traps, while fungus grown on PDA developed fewer traps and used nematodetoxic 30
furanone metabolites to compensate for insufficient traps. Another volatile pyrone metabolite, maltol, 31
was identified as a morphological regulator for enhancing trap formation. Deletion of gene 32
AOL_s00079g496 in A. oligospora led to increased furanone attractant (2 folds) in mutants and 33
enhanced attractive activity (1.5 fold) of the fungus, while resulted in decreased trap formation. This 34
investigation provides new insights regarding the comprehensive tactics of fungal adaptation to 35
environmental stress, integrating both morphological and metabolomic mechanisms. 36
KEYWORDS: Nematode-trapping fungi; Arthrobotrys oligospora; Metabolic adaptation; 37
Pathogenicity; Volatile Organic Compounds (VOCs) 38
39
40
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
3
Importance 41
Nematode-trapping fungi are a unique group of soil-living fungi that can switch from saprophytic to 42
pathogenic lifestyle once in contact with nematodes as a response to nutrient depletion. In this study, we 43
investigated the metabolic response during the switch and the key types of metabolites involved in the 44
interaction between fungi and nematodes. Our findings indicated that A. oligospora develop multiple 45
and flexible metabolic tactics corresponding to different morphological responses to nematodes. A. 46
oligospora can use similar volatile furanone and pyrone metabolites with different ecological functions 47
to help capture nematodes in the fungal switch from saprophytic to pathogenic lifestyles. Furthermore, 48
A. oligospora mutants with increased furanone and pyrone metabolites confirmed the results. This 49
investigation reveals the importance of volatile signaling in the comprehensive tactics used by nematode 50
trapping fungi, integrating both morphological and metabolomic mechanisms. 51
52
53
INTRODUCTION 54
Nematode-trapping fungi (NTF) can detect the presence of nematodes and develop specialized mycelial 55
trap devices to infect and consume prey as a response to nutrient depletion (1-4). These fungi are 56
broadly distributed in terrestrial and aquatic ecosystems, and more than 200 species from the phyla 57
Ascomycota, Basidiomycota, and Zygomycota have been described. Their role as natural enemies of 58
parasitic nematodes makes them attractive as biocontrol agents; moreover their unique ability to switch 59
between saprophytic and parasitic lifestyles are of great interest in basic ecological research (5, 6). 60
The direct physical contact with living nematodes has been assumed as the crucial biotic factor 61
necessary to induce the trap formation of NTF (7, 8). Traps are regarded as the key morphological 62
indication of the switch from the saprophytic to the pathogenic lifestyle for NTF (9-15). The nematodes 63
not only induce the formation of fungal traps, but once trapped, they also serve as a food source (3). 64
Considerable progress has been made in our understanding of the evolution and molecular mechanisms 65
of fungal trap formation at genomic, proteomic and transcriptomic levels (14, 16). When nematodes 66
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
4
induce the formation of trapping devices, multiple fungal signal transduction pathways are activated and 67
the downstream genes associated with energy metabolism, biosynthesis of the cell wall and adhesive 68
proteins involved in trap formation are regulated (7, 17). 69
Interestingly, NTF need an organic energy source other than nematodes in order to remain in an 70
active nematophagous state (18, 19). Previous studies have found that corn meal agar (CMA) and potato 71
dextrose agar (PDA) are among the best media for keeping the nematophagous activities of NTF and the 72
most extensively used by experimentalists to observe trap formation induced by nematodes (8, 12, 14, 73
16, 17, 18). The composition of the growth medium is important because fungi produce different 74
numbers of traps and have different nematocidal activities when grown on CMA or PDA. However, the 75
chemical identity of the signaling molecules responsible for these differential responses has remained 76
unclear. 77
It has long been assumed that traps are not the only weapons that NTF use to infect nematodes. In 78
1955, Duddington and Shepherd suggested that NTF could yield an unknown metabolite, 79
“nematotoxin”, to paralyze or kill nematodes because they found that the infected nematodes became 80
inactive before the infection bulb had completely developed (19, 20). Later in 1963, Olthof reported that 81
the filtrates from NTF parasitized nematodes contained the unstable nematode-inactivating substance 82
(21). In 1994, linoleic acid was reported as a putative nematicidal compound from several 83
nematophagous fungi (22). At the same time, a unique class of hybrid oligosporon metabolites found as 84
chemotaxonomic markers were reported from different strains of a model species Arthrobotrys 85
oligospora isolated from The Netherlands (23), Australia (24) and later from China (25, 26). These 86
known metabolites from A. oligospora are non-volatile compounds and exhibit several biological 87
activities, including moderate antibacterial properties, significant autoregulatory effects on the 88
formation of conidiophores and hyphal fusions in A. oligospora (26, 27). However, the chemical identity 89
of the metabolites involved in the lifestyle switch from saprotrophic to pathogenic phases has remained 90
cryptic. We hypothesized that these signals might be volatile in nature and used gas chromatography-91
mass spectrometry (GC-MS) in our analyses. 92
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
5
A. oligospora is commonly found in soils from diverse ecological habits and has emerged as the 93
model species for nematode-trapping fungi (5, 27). Under limited conditions, A. oligospora can form 94
three dimensional (3D) traps in direct contact with nematodes. In order to obtain information on 95
morphological and metabolic changes in A. oligospora, two common media, CMA and PDA, were used 96
in our analyses to investigate medium-specific metabolic features, as well to illuminate general aspects 97
of A. oligospora metabolism. There has been no previous report about the response of the fungus just 98
before the fungus starts to form predatory traps via the direct physical contact with nematodes, so a non-99
direct contact bioassay also was performed between the fungus and living/dead nematodes. Dead 100
nematodes were included in the non-direct contact bioassay in order to further evaluate if it could make 101
different responses to the approaching living and dead nematodes. The time-course designs over short-102
term intervals have provided successive snapshots of the morphological and metabolic status of A. 103
oligospora (model strain 1.1883) in response to a shift from the absence of the nematodes to the 104
presence of nematodes. GC-MS analysis was performed for metabolite profiling to determine 105
similarities and differences in temporal metabolite responses, and we have identified several volatile 106
compounds that exhibit medium-specific responses during the induction of traps in response to the 107
presence of nematodes. 108
109
Materials and Methods 110
Fungal and Nematode Strains, Media, and Treatment Conditions. 111
A. oligospora model species strain YMF1.01883 (ATCC 24927) was used in the fungus-nematode 112
interaction bioassays and cultured on CMA (corn (Kunming, China) 20 g L-1
, agar (Biofroxx, 113
Einhausen, Germany) 15 g L-1
) or PDA (potato (Kunming, China) 200 g L-1
, glucose (Solarbio, Beijing, 114
China) 10 g L-1
, agar 15 g L-1
). All the bioassays were conducted in 9 cm diameter glass Petri dishes. 115
Inocula of A. oligospora YMF1.01883 were cultured at 28°C on PDA plates for one week. Then one 5 116
mm diameter disk of the fungus was cut with a sterile cork borer and was inoculated onto either CMA 117
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
6
or PDA. The cultures were incubated at 28°C until the fungal lawn occupied half of the Petri plate to 118
obtain strong fresh mycelia. 119
Two treatment bioassays were performed to evaluate the fungal responses during the A. 120
oligospora-nematodes interaction: direct physical contact and non-direct contact. Caenorhabditis 121
elegans (strain N2) was cultured in oatmeal medium at 22°C for 6-7 days. The fungal strains treated 122
without nematodes on PDA or CMA, were used as controls for both treatment bioassays. 123
For the non-direct contact bioassay, the bottom of the Petri plate containing the fungal lawn was 124
inverted over a second Petri plate bottom of identical size containing 1 mL solution of mixed stage 125
living nematodes or dead nematodes (29). In this treatment, the fungi and the nematodes shared the 126
same atmosphere but had no direct physical contact. The two half Petri plates were sealed together with 127
Parafilm and then incubated in a dark chamber at 28 °C. For the control group, 1 mL sterile H2O was 128
used in place of the nematode suspension. Half of the live nematodes were submerged for 20 min in 129
45C water to prepare dead nematodes. In each treatment the fungus and nematodes were harvested 130
separately at a 6 hour interval for the first 48 hours and at a 24 hour interval for the subsequent 4 days, 131
respectively. In total, four biological replicates were performed for metabolomic analysis. 132
For the direct contact bioassay, 1 mL mixed-stage living nematode solution (about 3000 nematodes) 133
was directly added in the center of the fungal lawn; 1 mL sterile H2O was used as control. The Petri 134
dishes were sealed with Parafilm and incubated in a dark chamber at 28°C. The fungal lawn was 135
observed and harvested at a 6 hour interval for the first 48 hours and at a 24 hour interval for the 136
subsequent 4 days, respectively. The fungal lawns with nematodes from each time point were extracted 137
with methanol for metabolomic analysis. 138
139
Morphological Analysis. 140
For characterization of fungal growth, development and morphological transitions, microscopy was 141
performed according to the protocols outlined previously (16). Observations of the morphological 142
transitions of hyphal fusion to two-dimensional (2D) nets and morphological transitions to three-143
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
7
dimensional (3D) traps were assessed at a 6 hour interval for first 48 hours and at a 24 hour interval for 144
subsequent 4 days, respectively. The hyphal fusions and 3D traps were evaluated with a binocular 145
microscope (10 x magnification, Olympus, Japan). Seven fields in each fungal culture were picked at 146
random for observation; microscopic counting was repeated three times; and the data obtained were 147
analyzed statistically. Image stacks were processed using Imaris 6.3.1 (Bitplane) to generate images for 148
publication. The mean corrected data for the fungal strains treated with nematodes were obtained from 149
the outcome of the data in the test minus the data in the control group. 150
151
Metabolomic Profiling. 152
The metabolomic profiling analysis involved sample extraction, metabolite detection, metabolomic data 153
preprocessing (e.g., metabolite feature extraction, chromatographic peak alignment, data reduction), and 154
statistical analysis. The metabolic profiles were obtained from direct contact and non-direct contact 155
bioassays conducted on two media. Each treatment group consisted of 4 replicates and a corresponding 156
control group with same number replicates. The fungal mycelial lawn were harvested and extracted 157
twice with 30 mL methanol under ultrasonic conditions for 30 min in an ice-cooled bath-type sonicator. 158
Each methanol-soluble extract was centrifuged for 3 min at 10 000 × g and 4°C and the supernatant was 159
concentrated to dryness under vacuum. Each dried extract was resuspended in 1 mL methanol under 160
ultrasonic conditions for 20 min in an ice-cooled bath-type sonicator, and then filtered through 0.22 μm 161
membranes. The filtrates were stored at -80°C prior to GC-MS analyses. 162
GC-EI-MS analyses were performed as described (30) using a Hewlett-Packard gas chromatograph 163
5890 series II Plus linked to a Hewlett-Packard 5972 mass spectrometer system (Hewlett-Packard, San 164
Diego, CA, USA) equipped with a 30 m long, 0.25 mm i.d., and 0.5 μm film thickness HP5-MS 165
capillary column. The temperatures were programmed from 100 to 300 C at a rate of 5C/min. Helium 166
was used as a carrier gas at a flow rate of 0.7 mL/min. The split ratio was 1:20, the injector temperature 167
280 C, the interface temperature 300 C, and the ionization voltage 70 eV. 168
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
8
Identification of peaks was performed through retention time index and mass spectrum. 169
Compounds from the strains were designated as metabolites if they were identified with a match 900 on 170
a scale of 0 to 1,000 and retention index (RI) deviation of 3.0 (31, 32). The semiquantitative analysis of 171
the main compounds was performed through internal normalization with the area of each compound. 172
The addition of each area of the compounds corresponds to 100% area (33). 173
174 Data Analysis. 175
The data matrix was analyzed by Principal Component Analysis (PCA (34). The principle component 176
calculations were performed using TIGR MultiExperiment Viewer (MeV) software with a Centering 177
Mode, based on means, and visualized by using the Eigenvalues of the first principal component (x-axis) 178
and the second principal component (y-axis) or second principal component (x-axis) and the third 179
principal component (y-axis) (34). Each point on the plot represents an individual sample, and each 180
point on the loading plot represents a contribution of an individual metabolite to the score plot. 181
Accordingly, chemical components responsible for the differences between samples detected in the 182
scores plot can be extracted from the corresponding loadings. 183
Samples were clustered using unsupervised hierarchical cluster analysis (HCA) that provides 184
organization of primary data sets without predefined classification. Data were visualized by 185
dendrograms. Logarithmized values of metabolite relative concentrations were implemented in TIGR 186
Mev software in an unsupervised hierarchical cluster analysis (HCA) using Pearson correlation (34, 35). 187
188
Chemotaxis and Nematodetoxic Assays. 189
In order to evaluate if these volatile metabolites and mutants have nematode-attracting ability, 190
chemotaxis assays were preformed in 9 cm plates containing assay medium (20% agar, 5 mM potassium 191
phosphate pH 6.0, 1 mM CaCl2, 1 mM MgSO4) according to published protocols (36, 37). Two marks, 192
at opposite ends, were made on the back of the Petri plate, about 1 cm from the edge of the plate. 193
Between 100 and 200 washed adult nematodes were placed near the center of a 9 cm assay plate with 194
the putative attractant at one end of the plate and an aliquot of 1µL solvent ethanol was placed over the 195
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
9
other mark as control. An aliquot containing each respective test compound was suspended in 1µL of 196
ethanol and placed on the agar over one mark. Test compounds, including propanoic acid (1), 3-ethoxy-197
1,2-propanediol (2), 2(5H)-furanone (5), furan-2-ylmethanol (6), furan-2-carbaldehyde (7), 5-198
methylfuran-2-carbaldehyde (8), and n-Hexadecanoic acid (13), (Z,Z)-9,12-methyl octadecadienoate (14) 199
were obtained from Sigma-Aldrich USA, and D-(+)-Talose (12) was obtained from TCI Tokyo 200
Chemical Industry Co., Ltd. Japan. To evaluate the mutant strains, a 6 mm diameter disk of mycelium 201
grown for 2 days on CMA medium was used as test sample. For negative controls, a 6 mm diameter 202
disk of the wildtype strain on CMA medium was used. About 100 washed C. elegans adult nematodes 203
in M9 buffer were placed near the center of the plate, equidistant from the two marks. After 1hr, the 204
number of C. elegans at the putative attractant area and at the control area was counted. A chemotaxis 205
index was calculated based on the enrichment of animals at the attractant as following formula: 206
chemotaxis index = (the number of nematodes at the attractant area − the number of nematodes at the 207
control)/the total number of the Nematodes. The chemotaxis index varied from +1.0 to -1.0. In this 208
assay, a chemotaxis index of 1.0 represents complete preference for the test sample, and an index of 0 209
represents an equal distribution. 210
The nematode toxicity test was performed according to a previously published protocol (38). 211
About 300 C. elegans were dispensed into 3.5 cm plates containing 1 mL of M9 buffer with variable 212
amounts of pure metabolites (dissolved in DMSO) per plate. The same volume solvent of DMSO (0.5% 213
DMSO, v/v) was used as a negative control group and 1 g/mL ivermectin (Sigma-Aldrich USA) was 214
used as a positive control. Worms were exposed for 24 h at 20 °C, and the number of dead or living 215
worms was determined by the absence/presence of touch-provoked movement when probed with a 216
platinum wire. The median lethal concentration (LC50) value was calculated using the probit method 217
(38). All treatments were conducted in triplicate. 218
219
Mutant Construction. 220
The annotation of the genome of A. oligospora revealed five putative PKS genes including 221
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
10
AOL_s00043g287, AOL_s00043g828, AOL_s00079g496, AOL_s00215g283, and AOL_s00215g926 222
(16). The gene AOL_s00043g287 encodes a type III PKS and is designated PKS III-1. The genes 223
AOL_s00043g828, AOL_s00079g496, AOL_s00215g283, and AOL_s00215g926 encode type I PKSs 224
and are designated PKS I-1, PKS I-2, PKS I-3, and PKS I-4, respectively. A modified protoplast 225
transformation method (30) for genetic disruption of these PKS genes was applied using double-226
crossover recombination with the hygromycin-resistance gene (hyg) as a selection marker, followed by 227
identification of desired mutants using diagnostic PCR. The two homologous regions were amplified 228
from A. oligospora genomic DNA using primers containing overlapping regions with the vector pAg1-229
H3 and the hyg-resistance cassette. 230
Genomic DNA of A. oligospora was extracted as previously described (16). Restriction 231
endonucleases and DNA modifying enzymes were purchased from New England Biolabs (Beverly, 232
MA). In-Fusion® HD Cloning Kits were purchased from Clontech Laboratories (Mountain View, CA). 233
The left and right DNA fragments flanking the hygromycin resistant gene (hygR) in pAg1-H3 vector 234
were amplified from the genomic DNA of A. oligospora by PCR (GXL high-fidelity DNA Polymerase 235
TaKaRa Biotechnology Co. Ltd, Dalian, China) using primer sets as following. The disruption vector 236
for PKS III-1 gene AOL_s00043g287 was constructed with primer sets: 287-5f 237
(TCGAGCTCGGTACCAAGGCCCGGGTAAGACGGTGTAGAGGGCTGC), 287-5r 238
(GAGGCCTGATCATCGATGGGCCCGGACTTAGACTGGGCACT), 287-3f 239
(GCGATCGCGGCCGGCCGGCGCGCCGCCGAGGTCTTCTGGAAA) and 287-3r 240
(GAGTCACGAAGCTTGCATGCCTGCAGGTGTGCCGTTGCTTGGTAA). The disruption vector for 241
PKS I-1 gene AOL_s00043g828 was constructed with primer sets: 828-5f 242
(GAGCTCGGTACCAAGGCCCGGGTGCGTCACTTTGTTCATC), 828-5r 243
(CGAGGCCTGATCATCGATGGGCCCTAAATCTATCGTCGGGTAC), 828-3f 244
(GCGATCGCGGCCGGCCGGCGCGCCTCACGGAACAGGCACTAC) and 828-3r 245
(TCACGAAGCTTGCATGCCTGCAGGCAGACGATCTATCCCACC). The disruption vector for PKS 246
I-2 gene AOL_s00079g496 was constructed with primer sets: 496-5f: 247
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
11
AGCTCGGTACCAAGGCCCGGGTTTGTTATAGAAATGCCTCC and 496-5r: 248
GAGGCCTGATCATCGATGGGCCCGTCTTACCCAACTTAGCG, and 496-3f: 249
GCGATCGCGGCCGGCCGGCGCGCCAGATAGTAAGGATGGGCAG and 496-3r: 250
TCACGAAGCTTGCATGCCTGCAGGTGAAACGCAGACGGGTAA. The disruption vector for PKS 251
I-3 gene AOL_s00215g283 was constructed with primer sets: primer sets 283-5f, 283-5r, 283-3f and 252
283-3r (30). The disruption vector for PKS I-4 gene AOL_s00215g926 was constructed with primer sets: 253
926-5f (GAGCTCGGTACCAAGGCCCGGGGCCGTAAGTAAATTGTCTG), 926-5r 254
(AGGCCTGATCATCGATGGGCCCCAAGTGCGTGGTAGGAGC), 926-3f 255
(TCTAGAGGATCCCCCGACTAGTGTGGCGTTCGTAGTGATG) and 926-3r 256
(CACGAAGCTTGCATGCCTGCAGGTTCCAGTAGGACCGTGTA). 257
The DNA fragments (5' flanks and 3' flanks) were purified using PCR Clean-up Kit (Macherey-258
Nagel Inc, Düren, Germany) and NucleoSpin Gel, and were inserted into the specific sites of pAg1-H3 259
vector, respectively, by In-Fusion method to generate the completed disruption pAg1-H3-5′-3′ vector. 260
The homologous fragment amplifications were carried out as follows. Twenty-five L PCR 261
amplification system, using GXL high-fidelity DNA polymerase following the manufacturer’s 262
instructions (Takara) was applied. Half of one microliter of the prepared genomic DNA from A. 263
oligospora was added as template. All PCRs were performed in a Veriti 96-well thermal cycler 264
(Applied Biosystems, Foster city, CA). The amplification program contained predenaturation at 98°C 265
for 4 min followed by 30 cycles of denaturation at 98°C for 10 s, annealing at 57°C for 15 s, and 266
elongation at 68°C for 2 min, with a final extension step at 68°C for 10 min. 267
Medium PDASS (PDA supplemented with 0.6 M sucrose, 0.3 g/L yeast extract, 0.3 g/L tryptone, 268
0.3 g/L peptone, and 200 μg/mL hygromycin B (Roche Applied Science, Mannheim, Germany) for 269
selecting transformants) was applied to carry out protoplast regeneration. Four 1-1.2 cm diameter 270
mycelia plugs from 7 d fungal strain on YMA medium (2 g/L yeast extract, 10 g/L malt extract, and 18 271
g/L agar) were inoculated into 100 mL of TG medium [1% tryptone (Oxoid, Basingstoke, U.K.), and 272
1% glucose] and cultured at 30 °C at 180 rpm for 36 h. The mycelia were harvested and resuspended in 273
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
12
20 mL of a filter-sterilized enzyme solution that contained 120 mg of lysing enzymes (Sigma, St. Louis, 274
MO), 0.4 mL of cellulase (Sigma, St. Louis, MO), and 100 mg of snailase (Solarbio, Beijing, China) in 275
0.6 M MgSO4 at pH 6.0. The suspension was incubated for 4 h at 28 °C on a rotary shaker at 180 rpm. 276
Protoplasts were collected by filtering through six layers of sterile lens-cleaning tissue and centrifuged 277
at 1000 g. The protoplasts were washed twice with KTC (1.2 M KCl, 10 mM Tris-HCl, 50 mM CaCl2) 278
solution and finally resuspended in the same solution. 279
The protoplast-based protocol for the disruption of the targeted genes in A. oligospora was performed as 280
described previously (16). About 150 μL protoplasts (circa 8.0×107 /mL) were mixed with 10 μg linear 281
DNA in a 1.5 mL centrifuge tube. After 30 min of incubation on ice, 600 μL of PTC (50 mM CaCl2, 20 282
mM Tris-HCl, 50% polyethylene glycol 6000, pH 7.5) was added into the mixture and mixed gently. 283
After incubation at 28°C for 1 h, regeneration for 12 h, the putatively transformed protoplasts were 284
plated onto PDAS medium (PDA supplemented with 5 g/L molasses, 0.6 M
saccharose, 0.3 g/L
yeast 285
extract, 0.3 g/L tryptone, and 0.3 g/L
casein peptone) containing 200 μg/mL of hygromycin B. 286
Transformation colonies were selected after incubation at 28°C for 6-8 d, and every single colony was 287
transferred to a new plate containing TYGA medium (10 g/L tryptone, 10 g/L
glucose, 5 g/L
yeast 288
extract, 5 g/L molasses, 18 g/L agar) containing 200 μg/mL of hygromycin B. After incubation for 5 d at 289
28 °C, genomic DNA of putative transformants were extracted and were verified by PCR to check for 290
the integration of genes in the genome. Five mutants deficient in these PKS genes, respectively, were 291
screened out and confirmed by PCR. Knockout of the PKS I-2 gene AOL_s00079g496 was further 292
confirmed by southern blot analysis. Southern analysis was carried out according to the instructions 293
provided by the Chemiluminescent Nucleic Acid Detection Module (Thermo, Rockford, USA). The 294
primer pair KS-5f (TGTATTCCGTTTCGGTCTGC) and KS-3r (TTGAACCAACACGATTCTGC) 295
were used as Southern hybridization probes, and restriction enzyme Age I was used to digest the 296
genomic DNA of the wild-type A. oligospora and the mutant ∆AOL_s00079g496 for Southern analysis. 297
All the mutants were maintained and cultured on the same media in the same way as the wildtype strain. 298
The metabolites from cultures of these mutants and the wildtype strain were extracted and analyzed by 299
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
13
HPLC and GC-MS methods. GC-MS analysis was performed as described above in Metabolomic 300
Profiling. 301
302
HPLC Analysis. 303
HPLC analysis was carried out using a HP 1200 unit (Agilent, Waldbronn, Germany), employing the 304
following instrumental conditions: column, CAPCELL PAK C18, 5 m; 4.6 ×250 mm (Shiseido, Tokyo, 305
Japan); mobile phase A, 0.1% formic acid in water; mobile phase B, 0.1% formic acid in acetonitrile. 306
The LC conditions were performed as described previously (30) and were manually optimized on the 307
basis of separation patterns as gradient program of B: 0 min, 10% B; 2 min, 10% B; 10 min, 25% B; 30 308
min, 35% B; 35 min, 50% B; 45 min, 90% B; 47 min, 10% B; 49 min, 10% B. UV spectra were 309
recorded at 220-400 nm. 310
311
RESULTS 312
Differences in hyphal morphogenesis of A. oligospora on CMA and PDA in response to nematodes 313
In the non-direct contact bioassay, living and dead nematodes were used to evaluate if the fungus had 314
different morphological and metabolic responses. The morphological responses of A. oligospora grown 315
on CMA and PDA to the presence of nematodes under two modes of contact in 144 h were evaluated 316
and found to be significantly different (Fig. 1). In direct contact with nematodes, the fungal strains 317
grown on both PDA and CMA developed 3D traps. In our study, within 6 h on CMA, the formation of 318
3D traps was observed, while on PDA the formation of 3D traps was not observed until 12 h. Not only 319
did the fungus on CMA produce traps in a shorter time, after 24 h, fungal strains grown on CMA had 320
more traps than those on PDA. After 30 h, when exposed to nematodes, the fungus cultivated on CMA 321
produced the 3D traps at a level of 100 cm-2
, while the fungal strains grown on PDA formed fewer than 322
half this number of traps. The fungus grown on PDA took 12 hours longer to develop 3D traps at near 323
80 cm-2
which is 20% fewer than that of the fungal strains on CMA (Fig. 1). 324
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
14
The non-direct contact bioassay was performed according to published protocols (29). The bottom 325
portion of two glass Petri plates of identical size were used, one containing the fungal culture, the other 326
containing C. elegans nematodes. The fungal plate was inverted over the plate containing the nematodes, 327
but there was no direct contact between the worms and the fungal mycelium. Under this condition of 328
non-direct exposure to nematodes, at 24 h, no obvious morphological transition was observed in the 329
fungi cultured on either CMA or PDA (Fig. 1). However, hyphal fusions were observed. A 330
morphological transition of hypha fusions was observed at 30 h for the fungi grown on CMA and at 42 331
h on PDA. On CMA, the fungal strains developed 30% more hyphal fusions than on PDA (Fig. 1). 332
Interestingly, the numbers of the hyphal fusions produced by the fungus grown on CMA reached the 333
maximum at 80 cm-2
within 96 h, and then quickly decreased to 30 cm-2
. However, no such change in 334
rate of trap formation was observed for the fungal strains grown on PDA. The numbers of hyphal 335
fusions produced by the strains grown on PDA increased steadily and reached 80 cm-2
at the end of the 336
observation period. Exposure to living or dead nematodes made no obvious difference in the formation 337
of hyphal fusions when the fungus was grown on CMA. However, on PDA about 20% more hyphal 338
fusions were observed with exposure to live nematodes than with exposure to dead ones (Fig. 1). The 339
formation of 3D traps was not observed for the fungi grown on CMA until 72 h; 3D traps were observed 340
on PDA after 96 h. Then, while the numbers of 3D traps on both media increased slowly, they remained 341
at a low level (Fig. 1). In summary, the fungi grown on CMA developed more traps, and did so at a 342
faster rate, than those grown on PDA. In the absence of direct contact with nematodes, the fungi grown 343
on either CMA or PDA medium developed more hyphal fusions than 3D traps. 344
345
Metabolites from A. oligospora grown on CMA or PDA during the time course between the 346
saprophytic and pathogenic stages 347
. Time course metabolite profiles of A. oligospora YMF1.01883 grown on CMA and PDA treated with 348
nematodes, including direct contact and non-direct contacts with live and dead nematodes, and treated 349
without nematodes were analyzed with GC-MS analysis (Tables S1-S2). In non-direct contact bioassay, 350
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
15
nematodes were also collected at regular intervals for GC/MS analyses in order to remove the effect of 351
the nematodes on A. oligospora in direct contact bioassay. Four replicates for each treatment on one 352
medium at one time point led to in total 382 fungal samples and 60 nematode samples for metabolite 353
analysis. At one time point, the metabolite profiles of the fungal strains treated with nematodes were 354
compared with that of the fungal strains treated without nematodes to evaluate the metabolites varying 355
in contents. In order to get more information about the potential metabolites, the peaks were designated 356
as metabolites if they were identified with a match 700 on a scale of 0 to 1,000 to that data in the inborn 357
library. All the metabolites which showed significant changes in concentrations during the time course 358
metabolite profiles were considered (Tables S3-S4). 359
The metabolite profiles of A. oligospora on CMA and PDA, respectively, under four treatments at 360
24h were analyzed (Fig. S1). It was obvious to note that despite four types of treatments, the fungal 361
strains on the same medium shared quite similar metabolite patterns. It seemed that direct contact 362
between nematodes and fungi did not make an obvious difference in the fungal metabolite profiles. It is 363
also clear that the metabolite profiles of the fungi grown on CMA were different from those grown on 364
PDA. Comparison with the corresponding control groups without nematodes revealed that 34 out of 70 365
metabolites from the strains on CMA medium and 16 out of 80 metabolites from those on PDA medium 366
were significantly up- or down-regulated during the time course profiles of the fungal contacts with 367
nematodes. Among 34 varying metabolites detected from the fungal strains on CMA, 18 metabolites 368
were found from all the three groups treated with nematodes, 13 metabolites from the group treated with 369
nematodes in direct contact, 2 metabolites from the group treated with live nematodes in non-direct 370
contact, and 1 metabolite from the group treated with dead nematodes in non-direct contact. Among 16 371
varying metabolites detected from the fungal strains on PDA, 9 metabolites were found from all the 372
three groups treated with nematodes, 5 metabolites from the group treated with nematodes in direct 373
contact, 1 metabolite from the group treated with live nematodes in non-direct contact, and 1 metabolite 374
from the group treated with dead nematodes in non-direct contact. These metabolites included short 375
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
16
chained acids and esters, furanones and pyranones, linoleic acid derivatives, purines, and phenolic 376
compounds (Tables S5-S6). 377
378
Analysis of metabolic patterns of A. oligospora grown on CMA and PDA between the saprophytic 379
and the pathogenic stages 380
The profiled metabolite data were analyzed using principal component analysis (PCA). PCA is a useful 381
clustering method for exploratory data analysis and requires no previous knowledge of data structures. 382
The PCA score trajectories of logarithmically transformed metabolite concentrations from A. oligospora 383
during saprophytic and the pathogenic stages, growing on CMA and PDA, are depicted (Fig. 2). These 384
data points are clustered into two distinct groups in the plot maps (Fig. 2A and 2B), indicating clear 385
differences in the fungal extract metabolome between fungi growing on the two different media. The 386
first two principal components account together for 90.2% of the variance. Overall, the first principal 387
component mainly reflected differences in media. 388
The PCA plots of dendrograms from experiments and metabolite data in CMA and in PDA, 389
respectively, are depicted in Fig. 2C and 2D. The metabolic profiles of the CMA groups displayed more 390
extensive responses to nematodes than the PDA groups, compared with their corresponding time-series 391
controls. A notable transformation of metabolic changes was observed in the CMA groups treated with 392
nematodes under the two different modes of contact. During the time courses of fungal strains 393
cohabitating with nematodes under two different modes of contact, the PCA plot of the metabolites of 394
fungi grown on CMA shows that the experimental groups separate into four main branches: 1) the 395
control branch of A. oligospora cohabiting without nematodes (C); 2) the branch of A. oligospora 396
cohabiting under direct contact with nematodes (DC); 3) the branch of A. oligospora cohabiting under 397
non-direct contact with live nematodes (NDC-L); and 4) the branch of A. oligospora cohabiting under 398
non-direct contact with dead nematodes (NDC-D) (Fig. 2C). The fungal strains grown on CMA medium 399
had different metabolic responses not only to the approach and the access of nematodes, but also to the 400
presence of living or dead nematodes. In contrast, on PDA, no obvious distribution of metabolite data in 401
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
17
the PCA plot was observed with either the modes of contact or the viability status of nematodes (Fig. 402
2D). 403
Hierarchical clustering was applied to organize the metabolites based on their relative levels across 404
samples and to discern linkages between these metabolites (Fig. 3). A subset of small molecular 405
metabolite categories, including 11 metabolites in the CMA group and 9 metabolites in the PDA group, 406
were significantly changed while the fungi cohabited with nematodes from 6 h to 96h under both media. 407
Among these metabolites, 6 metabolites in the CMA group consistently changed patterns during the 408
time course. These were propanoic acid (1), 3-ethoxy-1,2-propanediol (2), 6-methoxy-9H-purin-2-409
amine (3), Hexahydro-2,6-epoxyfuro[3,2]-3-ol (4), 2(5H)-furanone (5), and furan-2-ylmethanol (6) (Fig. 410
4). On PDA medium, 7 metabolites showed changing patterns during the time course metabolite profiles. 411
These were furan-2-carbaldehyde (7), 5-methylfuran-2-carbaldehyde (8), 2H-pyran-2,6(3H)-dione (9), 412
3-hydroxy-2-methyl-4H-pyran-4-one (10), (R)-1-phenyl-1,2-ethanediol (11), D-(+)-Talose (12), n-413
hexadecanoic acid (13), and (9Z,12 Z)-methyl octadeca-9,12-dienoate (14, methyl ester of linoleic acid) 414
(Fig. 4). These compounds may have potential functional roles in the interaction between the fungal 415
strains and nematodes. 416
417
Characterization of the target metabolites during the fungus-nematode interaction. 418
The roles of 12 of the 14 individual metabolites were evaluated using 12 commercially available 419
compounds, including propanoic acid (1), 3-ethoxy-1,2-propanediol (2), 6-methoxy-9H-purin-2-amine 420
(3), 2(5H)-furanone (5), furan-2-ylmethanol (6), furan-2-carbaldehyde (7), 5-methylfuran-2-421
carbaldehyde (8), 3-hydroxy-2-methyl-4H-pyran-4-one (10), (R)-1-phenyl-1,2-ethanediol (11), D-(+)-422
Talose (12), n-hexadecanoic acid (13), and (9Z, 12Z)-methyl octadeca-9,12-dienoate (14). The use of 423
chemical standards allowed us to test the ability of individual compounds to attract or poison nematodes, 424
as well as to observe their effects on fungal development and morphology. 425
During the preliminary chemotaxis bioassay, among the metabolites tested at a concentration of 426
1mg/mL, 0.1 mg/mL, and 0.01mg/mL C. elegans worms were attracted toward only three metabolites: 427
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
18
2(5H)-furanone (5), furan-2-yl methanol (6) and furan-2-carbaldehyde (7). Interestingly, these three 428
metabolites all share a furan ring and have similar molecular weights. Compounds 5 and 6 were 429
characterized from the fungus on CMA, while compound 7 was characterized from the fungus on PDA, 430
based on the time course metabolic profiles of the fungal strains (Fig. 4). To characterize the chemotaxis 431
responses to these three volatile attractants further, worms were tested at concentrations of 1000, 500, 432
250, 100, 50, 25, 10, 5, and 1 µg/mL, and a chemotaxis index was calculated based on the enrichment of 433
animals at the attractant. The chemotaxis index could vary from 1.0 (perfect attraction) to -1.0 (perfect 434
repulsion). Weakly attractive ethanol was used as the control. The metabolite 2(5H)-furanone (5) 435
functioned as an attractant through a broad range of concentrations, displaying the strongest nematode-436
attracting ability at a concentration of 250 g/mL (Fig. 5A). Furan-2-yl methanol (6) showed a more 437
complex response, being attractive when undiluted but somewhat repulsive at low concentrations. 438
Among the 12 metabolites tested for their toxicity towards C. elegans, 5-methylfuran-2-439
carbaldehyde (8) showed toxic activity against nematodes with a LC50 value of 369 μg/mL in 12 h. The 440
other compounds tested did not display obvious toxic effects at the concentrations tested in these 441
experiments (Fig. 5B). 442
The same 12 compounds were applied to the fungal cultivation media. In comparison with the 443
solvent control, fungal strains treated with 2.5 g/mL of 3-hydroxy-2-methyl-4H-pyran-4-one (10), also 444
known as maltol (10), displayed a significant increase in the formation of 3D traps induced by 445
nematodes. Over 12 h, the number of adhesive 3D traps formed by the fungus grown on the media 446
treated with maltol (10) was 189 cm−2
(Fig. 5C), or 30% more than control untreated media (142 cm−2
). 447
448
Functional validation of the furanone and pyrone metabolites during the fungus-nematode 449
interaction. 450
Bioinformatics analysis of the A. oligospora genome revealed five putative polyketide synthase (PKS) 451
genes including a type III PKS synthase gene AOL_s00043g287 (PKS III-1), and four type I PKS 452
synthase genes AOL_s00043g828 (PKS I-1), AOL_s00079g496 (PKS I-2), AOL_s00215g283 (PKS I-3), 453
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
19
and the other AOL_s00215g926 (PKS I-4) (30). Disruption of these five PKS genes were performed and 454
five mutants ΔAOL_s00043g287 (PKS III-1), ΔAOL_s00043g828 (PKS I-1), ΔAOL_s00079g496 (PKS 455
I-2), ΔAOL_s00215g283 (PKS I-3), ΔAOL_s00215g926 (PKS I-4) were screened from 20, 7, 8, 22, and 456
23 transformants, respectively, by genomic DNA isolation and diagnostic PCR (see Fig. S2 and Fig. S3). 457
The metabolites from cultures of these mutants and the wildtype strain were extracted and analyzed by 458
HPLC and GC-MS methods according to the standard protocols (30). In the HPLC profiles, the mutants 459
AOL_s00215g283 (PKS I-3) and AOL_s00215g926 (PKS I-4) lacked most of the peaks with 460
retention times ranging between 21 and 40 min (see Table S8). Mutant AOL_s00043g287 (PKS III-1) 461
displayed the same HPLC and GC-MS profiles as the wildtype strain. Mutant AOL_s00043g828 (PKS 462
I-1) showed three peaks that were not observed in wild type at retention times at 11.68, 463
17.87 and18.82 min in the HPLC profile, and then characterized as non-furanone and non-pyrone 464
metabolites by comparison with the standard samples, and further GC-MS analysis. Only the HPLC 465
profile of mutant AOL_s00079g496 (PKS I-2) displayed an obvious difference in the peak for the 466
attractant compound 2(5H)-furanone (5), while most of other peaks were similar to the wild type profile 467
(Fig. 6). It was interesting to note that even grown on PDA, the mutant AOL_s00079g496 (PKS I-2) 468
yielded 200% higher 2(5H)-furanone (5) than the wildtype strain (Fig. 6). Further chemotaxis bioassays 469
performed on CMA also revealed that 150% more worms were attracted to AOL_s00079g496 (PKS I-470
2) than to the wildtype strain (Fig. 7), strongly confirming the nematode-attracting function of 2(5H)-471
furanone (5). 472
Mutant AOL_s00079g496 (PKS I-2) showed the same growth rates and conditions as the wild-type 473
strain both on CMA and on PDA within 6 days. However, from 9 days on, AOL_s00079g496 grown 474
on PDA displayed much fluffier aerial mycelia than the wildtype (Fig. 8). In addition, 475
AOL_s00079g496 showed increased spore formations but decreased germination rates than the wild 476
type strain (Fig. 7). When nematodes were added to 4 day PDA cultures, it was surprising to note that 477
AOL_s00079g496 formed fewer traps than the wild-type strain (Fig. 7). After 24 h, the number of 478
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
20
adhesive traps formed by AOL_s00079g496 was 69 cm−2
with 10% fewer than the number formed by 479
the wildtype strain (77 cm−2
) (Fig. 7). Accordingly, the number of nematodes infected by the traps 480
produced by the mutant AOL_s00079g496 was 35% fewer than that of the wildtype strain (Fig. 7). 481
Nevertheless, in both the wild type and the mutant strains, all the nematodes were dead within 36 h, 482
despite the fact that AOL_s00079g496 made fewer traps. Further GC-MS analysis revealed that 483
AOL_s00079g496 (PKS I-2) produced more furanone and pyrone metabolites including furan-2-484
ylmethanol (6), furan-2-carbaldehyde (7), 1-(furan-2-yl)propan-1-one (an ethyl derivative of 7), 5-485
(hydroxymethyl)furan-2-carbaldehyde (a hydroxy derivative of nematicidal 8), and 3,5-dihydroxy-6-486
methyl-2H-pyran-4(3H)-one (a hydroxy derivative of 10) than the wild type strain (Fig. 9). 487
488
DISCUSSION 489
Nematode trapping fungi have fascinated scientists for decades and many earlier workers have observed 490
the way in which the presence of nematodes alters the morphology and metabolism of trap forming 491
species. Although earlier studies detected attractant and nematocidal metabolites by their activities, the 492
compounds were never chemically identified (21-28). Therefore, we hypothesized that these signaling 493
molecules might be volatile in nature. In our analyses, we used GC-MS and were able to separate and 494
chemically characterize the metabolites, as well as elucidate their biological activities in attracting 495
nematodes, in inducing trap formation, or in killing nematodes. 496
Under direct physical contact with nematodes, fungi grown on CMA produced more 3D traps than 497
those grown on PDA and did so at a faster rate. Similar results were obtained in the non-direct contact 498
bioassay, however, instead of 3D trap formation, fungal hypha fusions were observed in the non-direct 499
assay. The fungi grown on CMA developed more hyphal fusions and 3D traps than those on PDA. Only 500
at a late stage after 72 h on CMA and 96 h on PDA were a few 3D traps observed. Previous studies 501
suggested that trap formation also requires a hyphal fusion event during initial stages (39), and hyphal 502
fusions were regarded as defensive structures of nematode-trapping fungi (5, 26). This might indicate 503
that in the face of the approaching nematodes, the fungus first moved into a defensive posture through 504
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
21
hyphal fusion before trap formation induced by direct contact with nematodes. Both bioassays revealed 505
that A. oligospora grown on CMA displayed greater morphological transitions in response to the 506
presence of nematodes than when it was grown on PDA. 507
The time course of metabolite profiling indicated that the growth medium influenced metabolism 508
more profoundly than the mode of contact with nematodes. When grown on CMA, almost half of (48%) 509
the total detected metabolites changed in response to physical or indirect contact with nematodes; when 510
grown on PDA only 11% of metabolites displayed significant abundance changes in response to direct 511
or indirect contact with nematodes. The fungal strains grown on CMA medium had more extensive 512
metabolic responses to nematodes compared with those grown on PDA medium. The PCA plot of 513
metabolite data from fungal strains on CMA clustered into four main branches, corresponding with the 514
experimental treatments according to the modes of contact (direct or indirect) and the status of 515
nematodes (living or dead). In other words, the fungal strains grown on CMA had different metabolic 516
patterns in response not only to the approach of nematodes, but also to the presence of living or dead 517
nematodes. In summary, our morphological and metabolic analyses indicate complex relationships 518
between media, fungal sensitivity, and morphological transitions. When A. oligospora was grown on 519
CMA, it made quicker and stronger responses in both morphology and metabolism even before having 520
direct contact with the nematodes. 521
The metabolomics analyses suggest a role of particular volatile metabolites in initiating the 522
morphological transition of the nematode-trapping fungus. Volatile compounds are emitted by many 523
species of fungi and serve many ecological functions in nature, as well as having been exploited for 524
their role in food flavor and as indirect indicators of the presence of fungal growth (40, 41, 42). 525
However, the role of individual fungal volatile substances in fungal-nematode ecological interactions is 526
poorly understood. 527
Metabolites 1-14 were screened out from the metabolic profiles of the fungal strains grown on 528
CMA and PDA during the switch from the saprophytic to pathogenic stages. Among them, two 529
furanone metabolites (5 and 6) emitted by the fungus grown on CMA were found to attract the 530
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
22
nematodes to the fungal colony. An early study on comparison of the interaction of the free-living 531
nematode Panagrellus redivivus with nematophagous fungi and non-nematophagous fungi showed that 532
nematophagous fungi preferred to attract nematodes while non-nematophagous fungi repelled them (28). 533
Our data not only confirm that the nematode-trapping fungi can chemically lure the prey to traps but 534
also provides the identification of specific attractive compounds as volatile in nature. 535
Among the changing metabolites produced by the fungus grown on the PDA between saprotrophic 536
and pathogenic phases, one volatile furanone metabolite, 5-methylfuran-2-carbaldehyde (8), was found 537
to significantly paralyze and kill the nematodes. There is a long standing assumption that an unstable 538
nematode-inactivating chemical compound produced by the fungus might be a volatile metabolite (21) 539
and our work confirms this assumption. We also found that the amount of the long chain metabolite, (9Z, 540
12Z)-methyl octadeca-9,12-dienoate (14, methyl ester of linoleic acid) increased significantly in the 541
fungus grown on PDA under contact with nematodes. Other research had identified linoleic acid as a 542
nematicidal metabolite in the mycelial extracts of several pathogenic fungi of the genus Arthrobotrys 543
(22). However, in our study, linoleic acid and its ethyl ester did not show obvious inhibitory effects on 544
nematodes. 545
In addition to the furanone metabolites involved in the interaction between the fungus and 546
nematodes, a volatile pyrone metabolite, 3-hydroxy-2-methyl-4-pyrone (maltol, 10) was identified as a 547
morphological regulator. Maltol (10) is found widely in various beans and other plant sources such as 548
larch tree bark, pine needles, and roasted malt (from which it gets its name) (43), but it has rarely been 549
described as a microbial metabolite (44, 45). Maltol is responsible for much of the characteristic smell 550
of red ginseng (46), and has been used to impart a sweet aroma to commercial fragrances. Maltol has 551
been marketed as a safe and reliable food flavor-enhancing agent for freshly baked breads and cakes and 552
also as food preservative and natural antioxidant (47). Maltol also is used as a bidentate metal ligand for 553
administered drugs (45). A recent study revealed that maltol found in the root exudates from crabgrass 554
can affect the growth of maize shoots and reduce the soil microbial biomass carbon by acting as an 555
allelochemical that interferes with plant growth and the microbial community of soils (48). However, 556
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
23
our study indicates that maltol acts as a morphological regulator for fungi. In our study, maltol was 557
found to be involved in regulating the formation of 3D traps of nematode-trapping fungus. Since maltol 558
is widely distributed in plants, it is interesting to speculate whether there is a co-evolutionary 559
relationship between maltol from plants attacked by the nematodes, and the induction of trap formation 560
by nematode-trapping fungi. 561
Furanone and pyrone metabolites are known to be important plant fruit constituents (49, 50). For 562
example, the 4-hydroxy-3(2H)-furanones associated with fruit aromas act to attract animals to fruits in 563
order to ensure seed dispersal. Furanones may function as inter-organism signal molecules in various 564
plant ecosystems (51). In plants, furanone and pyrone metabolites originate directly from carbohydrates 565
hexoses and pentoses as Maillard reaction products (52). Previous studies have suggested that furanone 566
might be derived from phosphorylated carbohydrates in tomato and yeast, and furaneol was from D-567
fructose-1,6-diphosphate. Hexose diphosphate was also assumed as biogenetic precursor to 4-hydroxy-568
5-methyl-2-methylene-3(2H)-furanone likely converted by an as yet unknown enzyme in tomato 569
(Solanum lycopersicum) and strawberry (Fragaria ananassa) (43, 45). However, the biogenetic 570
pathways of furanones and those of pyrones such as maltol (10) still remain unknown. 571
Polyketides are the most abundant class of fungal secondary metabolites (16). Because polyketides, 572
furanones, and pyrone metabolites all derive from the same precursors that are obtained from hexose 573
utilization, we studied the effects of all the PKS genes on the production of furanones and pyrone 574
metabolites in A. oligospora. Our previous report revealed that the knockout of the PKS I-3 gene 575
AOL_s00215g283 led to the abolishment of the morphological regulatory arthrosporols and high trap 576
formations (30). To elucidate the effects of genes in the production of furanone and pyrone metabolites, 577
mutants deficient in each of all the five PKS genes of A. oligospoara were constructed. The mutant with 578
loss of the PKS I-2 gene AOL_s00079g496 showed increased production of the furanone and pyrone 579
metabolites on both CMA and PDA medium, with no obvious changes in other metabolites or on 580
morphology. The content of nematode-attracting furanone and pyrone metabolites in the mutant strain 581
was double that of the wildtype strain. The fact that mutant AOL_s00079g496 (PKS I-2) displayed 582
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
24
150% stronger nematode-attracting activity was in good agreement with 200% higher of attractants in 583
the mutant strain. Although there were fewer traps induced by the strain in which the PKS I-2 gene 584
AOL_s00079g496 had been knocked out, the overall predatory ability remained the same as wild type. 585
We hypothesize that the extra production of nematodetoxic furanone metabolites compensated for the 586
lower number of traps. 587
In conclusion, we found that A. oligospora on CMA and PDA can sense the approaching 588
nematodes and develop hyphal fusions (Fig. 10). A. oligospora grown on both CMA and PDA produced 589
small volatile furanone and pyrone metabolites in response to the presence of nematodes. The fungus 590
cultivated on CMA medium made furanone metabolites that attracted nematodes, while the fungus 591
grown on PDA medium produced nematodetoxic furanone metabolites (Fig. 10). Mutation resulting in 592
the increase of furanone and pyrone metabolites led to increased attractive activity and decreased trap 593
formations of A. oligospora mutant, confirming the above results from integrated morphological and 594
metabolic analysis. These results show that the fungus flexibly adjusts its metabolic activity to 595
complement morphological changes, thereby potentially affecting fungal nematode-trapping ability and 596
differential trap formation. 597
598
ACKNOWLEDGMENT 599
This work was sponsored by projects from U1502262 and 31470169, and Yunnan University Program 600
for Excellent Young Talents awarded to X.N. (XT412003). 601
602 SUPPORTING INFORMATION 603
Additional Supporting Information may be found in the online version of this article at the publisher's 604
web-site. 605
REFERENCES 606
1. Pramer D. 1964. Nematode-trapping fungi. Science 144:382-388. 607
2. Barron GL. 1977. The nematode-destroying fungi. 608
Canadian Biological Publications Ltd,Guelph, Ontario, Canada. 609
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
25
3. Nordbring-Hertz B, Jansson HB, Tunlid A. 2011. Nematophagous fungi. eLS. John Wiley & 610
Sons, Ltd, Chichester, UK. 611
4. Moosavi MR, Zare R. 2012. Fungi as biological control agents of plant-parasitic nematodes, p 612
67-107. In Mérillon, J. M., Ramawat, K. G. (ed), Plant defence: biological control,Vol. 12. 613
Springer Netherlands, Dordrecht, Netherlands. 614
5. Nordbring-Hertz B. 1988. Nematophagous fungi: strategies for nematode exploitation and for 615
survival. Microbiol Sci 5:108-116. 616
6. Li J, Zou C, Xu J, Ji X, Niu X, Yang J, Huang X, Zhang K. 2015. Molecular Mechanisms of 617
Nematode-Nematophagous Microbe Interactions: Basis for Biological Control of Plant-Parasitic 618
Nematodes. Annu Rev Phytopathol 53:67-95. 619
7. Nordbring-Hertz B. 2004. Morphogenesis in the nematode-trapping fungus Arthrobotrys 620
oligospora - an extensive plasticity of infection structures. Mycologist 18:125-133. 621
8. Yang Y, Yang EC, An ZQ, Liu XZ. 2007. Evolution of nematode-trapping cells of predatory 622
fungi of the Orbiliaceae based on evidence from rRNA-encoding DNA and multiprotein 623
sequences. Proc Natl Acad Sci USA. 104:8379-8384. 624
9. Pramer D, Stoll NR. 1959. Nemin: a morphogenic substance causing trap formation by 625
predaceous fungi. Science 129:966-967. 626
10. Dijksterhuis J, Sjollema KA, Veenhuis M, Harder W. 1994. Competitive interactions 627
between two nematophagous fungi during infection and digestion of the nematode Panagrellus 628
redivivus. Mycol Res 98:1458-1462. 629
11. Xie H, Aminuzzaman F, Xu L, Lai Y, Li F, Liu X. 2010. Trap induction and trapping in eight 630
nematode-trapping fungi (Orbiliaceae) as affected by juvenile stage of Caenorhabditis elegans. 631
Mycopathologia 169:467-473. 632
12. Hsueh Y-P, Mahanti P, Schroeder FC, Sternberg PW. 2013. Nematode-trapping fungi 633
eavesdrop on nematode pheromones. Curr Biol 23:83-86. 634
13. Wang X, Li G-H, Zou C-G, Ji X-L, Liu T, Zhao P-J, Liang L-M, Xu J-P, An Z-Q, Zheng X. 635
2014. Bacteria can mobilize nematode-trapping fungi to kill nematodes. Nat Commun 5:5776-636
5785. 637
14. Su H, Zhao Y, Zhou J, Feng H, Jiang D, Zhang KQ, Yang J. 2015. Trapping devices of 638
nematode-trapping fungi: formation, evolution, and genomic perspectives. Biol Rev 31:371-378. 639
15. Li L, Yang M, Luo J, Qu Q, Chen Y, Liang L, Zhang K. 2016. Nematode-trapping fungi and 640
fungus-associated bacteria interactions: the role of bacterial diketopiperazines and biofilms on 641
Arthrobotrys oligospora surface in hyphal morphogenesis. Environ Microbiol 18:3827-3839. 642
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
26
16. Yang J, Wang L, Ji X, Feng Y, Li X, Zou C, Xu J, Ren Y, Mi Q, Wu J. 2011. Genomic and 643
Proteomic Analyses of the Fungus Arthrobotrys oligospora Provide Insights into Nematode-644
Trap Formation. PLoS Pathog 7:e1002179. 645
17. Singh UB, Sahu A, Singh R, Singh DP, Meena KK, Srivastava J, Manna M. 2012. 646
Evaluation of biocontrol potential of Arthrobotrys oligospora against Meloidogyne graminicola 647
and Rhizoctonia solani in Rice (Oryza sativa L.). Biol Control 60:262-270. 648
18. Zlitni S, Ferruccio LF, Brown ED. 2013. Metabolic suppression identifies new antibacterial 649
inhibitors under nutrient limitation. Nat Chem Biol 9:796-804. 650
19. Duddington C. 1955. Fungi that attack microscopic animals. Bot Rev 21:377-439. 651
20. Shepherd AM. 1955. Formation of the infection bulb in Arthrobotrys oligospora Fresenius. 652
Nature 175:475 653
21. Olthof TH, Estey R. 1963. A nematotoxin produced by the nematophagous fungus Arthrobotrys 654
oligospora Fresenius. Nature 197:514-515. 655
22. Stadler M, Anke H, Sterner O. 1993. Linoleic acid — The nematicidal principle of several 656
nematophagous fungi and its production in trap-forming submerged cultures. Arch Microbiol 657
160:401-405. 658
23. Stadler M, Sterner O, Anke H. 1993. New biologically active compounds from the nematode-659
trappmg fungus Arthrobotrys oligospora fresen. Z Naturforsch C 48:843-850. 660
24. Anderson MG, Jarman TB, Rickards RW. 1995. Structures and absolute configurations of 661
antibiotics of the oligosporon group from the nematode-trapping fungus Arthrobotrys 662
oligospora. J Antibiot 48:391-398. 663
25. Wei LX, Zhang HX, Tan JL, Chu YS, Li N, Xue HX, Wang YL, Niu XM, Zhang Y, Zhang 664
KQ. 2011. Arthrobotrisins A-C, Oligosporons from the Nematode-Trapping Fungus 665
Arthrobotrys oligospora. J Nat Prod 74:1526-1530. 666
26. Zhang HX, Tan JL, Wei LX, Wang YL, Zhang CP, Wu DK, Zhu CY, Zhang Y, Zhang KQ, 667
Niu XM. 2012. Morphology regulatory metabolites from Arthrobotrys oligospora. J Nat Prod 668
75:1419-1423. 669
27. Niu X, Zhang K. 2011. Arthrobotrys oligospora: a model organism for understanding the 670
interaction between fungi and nematodes. Mycology 2:59-78. 671
28. Saxena G, Dayal R, Mukerji KG. 1987. Interaction of nematodes with nematophagus fungi: 672
induction of trap formation, attraction and detection of attractants. FEMS Microbiol Lett 45:319-673
327. 674
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
27
29. Niu Q, Huang X, Zhang L, Xu J, Yang D, Wei K, NiuX, An Z, Bennett JW, Zou C, Yang K, 675
Zhang K. 2010. A Trojan horse mechanism of bacterial pathogenesis against nematodes. Proc 676
Natl Acad Sci USA. 107:16631-16636. 677
30. Xu ZF, Wang BL, Sun HK, Yan N, Zeng ZJ, Zhang KQ, Niu XM. 2015. High Trap 678
Formation and Low Metabolite Production by Disruption of the Polyketide Synthase Gene 679
Involved in the Biosynthesis of Arthrosporols from Nematode-Trapping Fungus Arthrobotrys 680
oligospora. J Agric Food Chem 63:9076-9082. 681
31. Li Z, Yao Q, Dearth SP, Entler MR, Castro Gonzalez HF, Uehling JK, Vilgalys RJ, Hurst 682
GB, Campagna SR, Labbé JL. 2017. Integrated proteomics and metabolomics suggests 683
symbiotic metabolism and multimodal regulation in a fungal-endobacterial system. Environ 684
Microbiol 19:1041-1053. 685
32. Meissner S, Steinhauser D, Dittmann E. 2015. Metabolomic analysis indicates a pivotal role 686
of the hepatotoxin microcystin in high light adaptation of Microcystis. Environ Microbiol 687
17:1497-1509. 688
33. Schieberle P, Molyneux RJ. 2012. Quantitation of sensory-active and bioactive constituents of 689
food: A Journal of Agricultural and Food Chemistry perspective. J Agric Food Chem 60:2404-690
2408. 691
34. Wu R, Wu Z, Wang X, Yang P, Yu D, Zhao C, Xu G, Kang L. 2012. Metabolomic analysis 692
reveals that carnitines are key regulatory metabolites in phase transition of the locusts. Proc Natl 693
Acad Sci USA. 109:3259-3263. 694
35. Frimmersdorf E, Horatzek S, Pelnikevich A, Wiehlmann L, Schomburg D. 2010. How 695
Pseudomonas aeruginosa adapts to various environments: a metabolomic approach. Environ 696
Microbiol 12:1734–1747. 697
36. Macosko EZ, Pokala N, Feinberg EH, Chalasani SH, Butcher RA, Clardy J, Bargmann CI. 698
2009. A hub-and-spoke circuit drives pheromone attraction and social behaviour in C. elegans. 699
Nature 458:1171-1175 700
37. Srinivasan J, Reuss SHV, Bose N, Zaslaver A, Mahanti P, Ho MC, O'Doherty OG, Edison 701
AS, Sternberg PW, Schroeder FC. 2012. A Modular Library of Small Molecule Signals 702
Regulates Social Behaviors in Caenorhabditis elegans. PLoS Biol 10:e1001237. 703
38. Wang YL, Li LF, Li DX, Wang B, Zhang K, Niu X. 2015. Yellow Pigment Aurovertins 704
Mediate Interactions between the Pathogenic Fungus Pochonia chlamydosporia and Its 705
Nematode Host. J Agric Food Chem 63:6577-6587. 706
39. Nordbring-Hertz B, Friman E, Veenhuis M. 1989. Hyphal fusion during initial stages of trap 707
formation in Arthrobotrys oligospora. Avan Leeuw J Microb 55:237-244. 708
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
28
40. Bitas V, Kim HS, Bennett JW, Kang S. 2013. Sniffing on microbes: diverse roles of microbial 709
volatile organic compounds in plant health. Mol Plant-Microbe Interact 26:835-843. 710
41. Hung R, Lee S, Bennett JW. 2015. Fungal volatile organic compounds and their role in 711
ecosystems. Appl Microbiol Biotechnol 99:3395-3405. 712
42. Schrader J. 2007. Microbial Flavour Production, p 507-574. In Ralf Günter Berger 713
(ed), Flavours and Fragrances. Springer Berlin Heidelberg, German. 714
43. Guentert M. 2007. The Flavour and Fragrance Industry—Past, Present, and Future, p 1-14. In 715
Ralf Günter Berger (ed), Flavours and Fragrances. Springer Berlin Heidelberg, German. 716
44. Edris AE. 2007. Pharmaceutical and therapeutic potentials of essential oils and their individual 717
volatile constituents: a review. Phytother Res 21:308-323. 718
45. Yi Z, Lu Q, Liu L, Li X, Liu E, Han L, Fang S, Gao X, Tao W. 2014. New maltol glycosides 719
from Flos Sophorae. J Nat Med 69:249-254. 720
46. Sangjun L, Taewha M, Jaehwan L. 2010. Increases of 2-furanmethanol and maltol in Korean 721
red ginseng during explosive puffing process. J Food Sci 75:C147-151. 722
47. Ye H, Qi X, Hu JN, Han XY, Wei L, Zhao LC. 2015. Maltol, a Food Flavoring Agent, 723
Attenuates Acute Alcohol-Induced Oxidative Damage in Mice. Nutrients 7:682-696. 724
48. Zhou B, Kong CH, Li YH, Wang P, Xu XH. 2013. Crabgrass (Digitaria sanguinalis) 725
Allelochemicals That Interfere with Crop Growth and the Soil Microbial Community. J Agric 726
Food Chem 61:5310-5317. 727
49. Pichersky E, Gershenzon J. 2002. The formation and function of plant volatiles: perfumes for 728
pollinator attraction and defense. Curr Opin Plant Biol 5:237-243. 729
50. Slaughter JC. 1999. The naturally occurring furanones: formation and function from 730
pheromone to food. Biol Rev 74:259-276. 731
51. Iason GR, Dicke M, Hartley SE. 2012. The Ecology of Plant Secondary Metabolites: From 732
Genes to Global Processes. Cambridge University Press, New York, USA. 733
52. Schwab W, Davidovichrikanati R, Lewinsohn E. 2008. Biosynthesis of plant-derived flavor 734
compounds. Plant J 54:712-732. 735
736
737
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
29
738
Figure Legends 739
Figure 1. The morphological responses of A. oligospora grown on two different media, CMA (red) and 740
PDA (black), to living nematodes (solid, L) or dead nematodes (hollow, D), under two different modes 741
of contact, direct contact (full line, DC, A) and non-direct contact (dash line, NDC, B) in 144h. 3D traps 742
(triangle) and hyphal fusions (circle). The corrected values are the difference values between the data 743
obtained in fungal strains treated with nematodes and those obtained in fungal strains treated without 744
nematodes. 745
746
Figure 2. A): PCA score-plots between fungal samples during the saprophytic and pathogenic phases 747
grown on two different media, CMA and PDA. (principal component 1 versus principal component 2; 748
component 1, 0.85 and component 2, 0.05). B): PCA score-plots between fungal samples during the 749
saprophytic and pathogenic phases grown on two different media CMA and PDA. (principal component 750
1 versus principal component 3; component 1: 0.85, component 2: 0.05 and component 3: 0.03); C): 751
PCA score-plot between the saprophytic and pathogenic fungal samples grown on CMA (principal 752
component 2 versus principal component 3; component 1: 0.60, component 2:0.12 and component 3: 753
0.07). D): PCA score-plot between the saprophytic and pathogenic fungal samples grown on PDA 754
(principal component 2 versus principal component 3; component 1: 0.65, component 2: 0.11 and 755
component 3: 0.08). A. oligospora growing without nematodes in 144h as control group (C); A. 756
oligospora growing under direct contact with living nematodes in 144h (DC); A. oligospora growing 757
under non-direct contact live nematodes in 144h (NDC-L); A. oligospora growing under non-direct 758
contact with dead nematodes in 144h (NDC-D). 759
760
Figure 3. Unsupervised hierarchical clustering of the logarithmically transformed (log2) relative 761
concentrations of metabolites from the methanol extracts of A. oligospora YMF1.01883 at different 762
growth phases cultivated in CMA and PDA media. Up): Unsupervised hierarchical clustering of the 763
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
30
logarithmically transformed (log2) relative concentrations of 33 metabolites from the methanol extracts 764
of wild type A. oligospora YMF1.01883 cultivated in CMA medium. Down): Unsupervised hierarchical 765
clustering of the logarithmically transformed into (log2) metabolite relative concentrations of 16 766
metabolites from the methanol extracts of wild type A. oligospora cultivated in PDA medium. A. 767
oligospora growing without nematodes in 144 h as control group (C); A. oligospora growing under 768
direct contact with nematodes in 144 h (DC); A. oligospora growing under non-direct contact with 769
living nematodes in 144 h (NDC-L); A. oligospora growing under non-direct contact with dead 770
nematodes in 144 h (NDC-D). 771
772
Figure 4. The structures of 14 metabolites and their abundances within the time courses from the 773
saprophytic to the pathogenic lifestyles of the fungus. (6 metabolites in the CMA group including 1-6, 774
and 8 metabolites in the PDA group including 7-14. Control (green): A. oligospora growing without 775
nematodes in 144h as control group; Direct Contact (blue): A. oligospora growing under direct contact 776
with living nematodes in 144h; Non-Direct Contact Live (rose): A. oligospora growing under non-direct 777
contact with living nematodes in 144h; Non-Direct Contact Dead (red): A. oligospora growing under 778
non-direct contact with dead nematodes in 144h. (n=5). 779
780
Figure 5. A): Attracting activities of three metabolites, 2(5H)-furanone (5), furan-2-ylmethanol (6), and 781
furan-2-carbaldehyde (7) for nematode C. elegans. B): Effect of 5-methylfuran-2-carbaldehyde (8) on 782
the mortality of C. elegans at 12 hours with 1µg/mL Ivermectin used as a positive control group. C): 783
Effect of maltol (10, M) at the concentration of 2.5 g/mL on trap formations of A. oligospora as 784
treated without maltol as control (C). **: P <0.01, *: P <0.05, n=5. 785
786
Figure 6. HPLC analysis of the methanol extracts of the PD cultural broths from the wildtype strain 787
(black line) and the mutant AOL_s00079g496 (Δ79g496 ) (red line). The blue line indicates the peak 788
of 2(5H)-furanone (5) at around 4 min. 789
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
31
Figure 7. A): Comparison of attracting activities of wild type A. oligospora and mutant 790
ΔAOL_s00079g496 (496). B): Comparison of spore formations of A. oligospora and mutant 791
AOL_s00079g496. C): Comparison of spore germination rates of A. oligospora and mutant 792
AOL_s00079g496. D): Comparison of trap formations of wild type A. oligospora and mutant 793
ΔAOL_s00079g496. E): Comparison of nematode capturing abilities of wild type A. oligospora and 794
mutant ΔAOL_s00079g496. ***: P <0.001, **: P <0.01, *: P <0.05, n=4. 795
796
Figure 8. Comparison of mycelial morphology of wild-type strain and the mutant ΔAOL_s00079g496 797
(Δ79g496-KS) on PDA plates (15 d). 798
799
Figure 9. GC-MS analysis of the methanol extracts of the wildtype strain (black line) and the mutant 800
ΔAOL_s00079g496 (Δ79g496) (red line) on CMA. The red arrow refers to the compounds detected in 801
the mutant strain. 802
803
Figure 10. The morphological and metabolic adaptation of wild type A. oligospora on two media CMA 804
and PDA. In direct contact with nematodes, the fungus grown on CMA develops more traps than those 805
grown on PDA; in non-direct contact with nematodes, the fungus grown on PDA develops more hyphal 806
fusion than on CMA. The fungus grown on CMA produced an attractant molecule, 2(5H)-furonone, 807
while the fungus on PDA produced a nematicide metabolite, 5-methyl furan 2-carbaldehyde, as well as 808
a stimulator, maltol, that increased trap numbers. 809
810
811
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
32
812
0 20 40 60 80 100 120 1400
20
40
60
80
100
Traps of DC on CMATraps of DC on PDA
(h)
Corr
ecte
d v
alu
es
of
morp
holo
gic
al
tran
siti
on
s A
0 20 40 60 80 100 120 1400
20
40
60
80
100
Traps of NDC-D on PDA Traps of NDC-D on CMA
Traps of NDC-L on PDA Traps of NDC-L on CMA
Hyphal fusion of NDC-L on PDA Hyphal fusion of NDC-L on CMA
Hyphal fusion of NDC-D on PDA Hyphal fusion of NDC-D on CMA
(h)
Corr
ecte
d v
alu
es
of
morp
holo
gic
al
tran
siti
on
s B
813
Figure 1. The morphological responses of A. oligospora grown on two different media, CMA (red) and 814
PDA (black), to living nematodes (solid, L) or dead nematodes (hollow, D), under two different modes 815
of contact, direct contact (full line, DC, A) and non-direct contact (dash line, NDC, B) in 144h. 3D traps 816
(triangle) and hyphal fusions (circle). The corrected values are the difference values between the data 817
obtained in fungal strains treated with nematodes and those obtained in fungal strains treated without 818
nematodes. 819
820
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
33
821
A
B
C
D
Figure 2. A): PCA score-plots between fungal metabolic samples during the saprophytic and 822
pathogenic phases grown on two different media, CMA and PDA. (principal component 1 versus 823
principal component 2; component 1, 0.85 and component 2, 0.05). B): PCA score-plots between fungal 824
samples during the saprophytic and pathogenic phases grown on two different media CMA and PDA. 825
(principal component 1 versus principal component 3; component 1: 0.85, component 2: 0.05 and 826
component 3: 0.03); C): PCA score-plot between the saprophytic and pathogenic fungal samples grown 827
on CMA (principal component 2 versus principal component 3; component 1: 0.60, component 2:0.12 828
and component 3: 0.07). D): PCA score-plot between the saprophytic and pathogenic fungal samples 829
grown on PDA (principal component 2 versus principal component 3; component 1: 0.65, component 2: 830
0.11 and component 3: 0.08). A. oligospora growing without nematodes in 144h as control group (C); A. 831
oligospora growing under direct contact with living nematodes in 144h (DC); A. oligospora growing 832
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
34
under non-direct contact live nematodes in 144h (NDC-L); A. oligospora growing under non-direct 833
contact with dead nematodes in 144h (NDC-D). 834
835
836
837
Figure 3. Unsupervised hierarchical clustering of the logarithmically transformed (log2) relative 838
concentrations of metabolites from the methanol extracts of A. oligospora YMF1.01883 at different 839
growth phases cultivated in CMA and PDA media. Up): Unsupervised hierarchical clustering of the 840
logarithmically transformed (log2) relative concentrations of 33 metabolites from the methanol extracts 841
of wild type A. oligospora YMF1.01883 cultivated in CMA medium. Down): Unsupervised hierarchical 842
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
35
clustering of the logarithmically transformed into (log2) metabolite relative concentrations of 16 843
metabolites from the methanol extracts of wild type A. oligospora cultivated in PDA medium. A. 844
oligospora growing without nematodes in 144 h as control group (C); A. oligospora growing under 845
direct contact with nematodes in 144 h (DC); A. oligospora growing under non-direct contact with 846
living nematodes in 144 h (NDC-L); A. oligospora growing under non-direct contact with dead 847
nematodes in 144 h (NDC-D). 848
849
850
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
37
853 854 Figure 4. The structures of 14 metabolites and their abundances within the time courses from the 855
saprophytic to the pathogenic lifestyles of the fungus. (6 metabolites in the CMA group including 1-6, 856
and 8 metabolites in the PDA group including 7-14. Control (green): A. oligospora growing without 857
nematodes in 144h as control group; Direct Contact (blue): A. oligospora growing under direct contact 858
with living nematodes in 144h; Non-Direct Contact Live (rose): A. oligospora growing under non-direct 859
contact with living nematodes in 144h; Non-Direct Contact Dead (red): A. oligospora growing under 860
non-direct contact with dead nematodes in 144h. (n=4). 861
862
863
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
38
0 13
827
555
011
00IV
M
0
25
50
75
100
**
**
(g/mL)
(%)
Mort
ali
ty r
ate
s of
nem
ato
des
12 240
100
200
300
400
(h)
(/cm2)
C
M*
Nu
mb
ers
of
trap
s
1000 50
025
010
0 50 25 10 5 1
-0.2
0.0
0.2
0.4
0.6
0.8
2(5H)-Furanone
Furan-2-carbaldehyde
Furan-2-ylmethanol
Concn. (g/mL)
Ch
emota
ctic
in
dex
A
B C
864
Figure 5. A): Attracting activities of three metabolites, 2(5H)-furanone (5), furan-2-ylmethanol (6), and 865
furan-2-carbaldehyde (7) for nematode C. elegans. B): Effect of 5-methylfuran-2-carbaldehyde (8) on 866
the mortality of C. elegans at 12 hours with 1µg/mL Ivermectin used as a positive control group. C): 867
Effect of maltol (10, M) at the concentration of 2.5 g/mL on trap formations of A. oligospora as 868
treated without maltol as control (C). **: P <0.01, *: P <0.05, n=5. 869
870
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
39
871
872
873 874 Figure 6. HPLC analysis of the methanol extracts of the PD cultural broths from the wildtype strain 875
(black line) and the mutant AOL_s00079g496 (Δ79g496 ) (red line). The blue line indicates the peak 876
of 2(5H)-furanone (5) at around 4 min.877
min
0 10 20 30 40
mAU
400
800
1200
1600
WT
79g496
2(5H)-Furanone
4 6
2(5H)-Furanone
WT
79g496
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
40
878
879
24 480
25
50
75
****
(h)
(%)
Att
ract
ed n
emato
des
0.0
1.5
3.0
4.5 **
WT Δ496
(105/mL)
Nu
mb
ers
of
spore
s
2h 4h0
50
100
150
***
**
(%)
Sp
ore
ger
min
ati
on
rate
s
12 24 36 480
60
120
180
(h)
(cm-2)
**
Nu
mb
ers
of
trap
s
24 360
50
100
(h)
**
*
(%)
Cap
ture
d n
emato
des
0.01.53.04.5
WT
Δ496
A B C
D E
880
Figure 7. A): Comparison of attracting activities of wild type A. oligospora and mutant 881
ΔAOL_s00079g496 (496). B): Comparison of spore formations of A. oligospora and mutant 882
AOL_s00079g496. C): Comparison of spore germination rates of A. oligospora and mutant 883
AOL_s00079g496. D): Comparison of trap formations of wild type A. oligospora and mutant 884
ΔAOL_s00079g496. E): Comparison of nematode capturing abilities of wild type A. oligospora and 885
mutant ΔAOL_s00079g496. ***: P <0.001, **: P <0.01, *: P <0.05, n=4. 886
887
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
41
888
WT Δ79g496-KS
889
Figure 8. Comparison of mycelial morphology of wild-type strain and the mutant ΔAOL_s00079g496 890
(Δ79g496-KS) on PDA plates (15 d). 891
892
893
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
42
894 Figure 9. GC-MS analysis of the methanol extracts of the wildtype strain (black line) and the mutant 895
ΔAOL_s00079g496 (Δ79g496) (red line) on CMA. The red arrow refers to the compounds detected in 896
the mutant strain. 897
898
10 20 30 40 0
20
40
60
80
min
(x
10 5 )
2 3 4
5
1
W
T
△ 79g496
O
3 . 1 - ( F u r a n - 2 - y l ) p r o p a n - 1 - o n e
O
O O H H O
4 . 3 , 5 - D i h y d r o x y - 6 - m e t h y l - 2 H - p y r a n - 4 ( 3 H ) - o n e
5 . 5 - ( h y d r o x y m e t h y l ) f u r a n - 2 - c a r b a l d e h y d e
1 . F u r a n - 2 - c a r b a l d e h y d e ( 6 )
2 . F u r a n - 2 - y l m e t h a n o l ( 7 )
O
O O H
O O
O O H O
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from
43
899
Attractant
Stimulator
Nematicide
Fungus
Traps
Hyphal
Fusions
Hyphal
Fusions
Traps
900
Figure 10. The morphological and metabolic adaptation of wild type A. oligospora on two media CMA 901
and PDA. In direct contact with nematodes, the fungus grown on CMA develops more traps than those 902
grown on PDA; in non-direct contact with nematodes, the fungus grown on PDA develops more hyphal 903
fusion than on CMA. The fungus grown on CMA produced an attractant molecule, 2(5H)-furonone, 904
while the fungus on PDA produced a nematicide metabolite, 5-methyl furan 2-carbaldehyde, as well as 905
a stimulator, maltol, that increased trap numbers. 906
907
on May 25, 2018 by guest
http://aem.asm
.org/D
ownloaded from