Biofumigation on Post-Harvest Diseases of Fruits Using a New Volatile-Producing Fungus of Ceratocystis fimbriata

A variety of volatile organic compounds (VOCs) produced by Ceratocystis fimbriata have strong bioactivity against a wide range of fungi, bacteria and oomycetes. Mycelial growth, conidial production, and spore germination of fungi and oomycetes were significantly inhibited after exposure to cultures of C. fimbriata, and colony formation of bacteria was also inhibited. Two post-harvest diseases, peach brown rot caused by Monilinia fructicola and citrus green mold caused by Penicillium digitatum, were controlled during a 4-day storage by enclosing wound-inoculated fruits with 10 standard diameter Petri plate cultures of C. fimbriata in a 15 L box. The fruits were freshly inoculated at onset of storage and the cultures of C. fimbriata were 6 days old. Percentage of control was 92 and 97%, respectively. After exposure to C. fimbriata VOCs, severely misshapen hyphae and conidia of these two post-harvest pathogens were observed by scanning electron microscopy, and their pathogenicity was lost or greatly reduced.


Introduction
The development and utilization of volatile organic compounds (VOCs) from microorganisms have been of increasing interest since they are naturally produced without chemical synthesis. VOCs are mixtures of carbon-based compounds that are either vapors or highly volatiles [1]. The identified VOCs generated by fungi and bacteria are mainly primary (from the synthesis of DNA, amino and fatty acids, etc.) and secondary metabolites (from intermediates of primary metabolism) [2]. Microbial VOCs may be signaling substances for regulating and controlling certain physiological actions of some plants. The growth of Arabidopsis thaliana was promoted after the exposure to the volatiles from bacteria [3], and the systemic resistance in A. thaliana against Erwinia carotovora was induced [4]. Root length, shoot length and fresh weight were increased by treating lettuce seedlings with VOCs produced by Fusarium oxysporum MSA 35 were covered with paraffin oil at 4°C, and transferred to fresh PDA or CA plate and cultured in darkness at 25°C for 5 to 7 days when needed. To make conidial suspensions of test fungi, the culture of these organisms in Petri dishes were flooded with sterile water containing 0.01% Tween 80 and scraped gently with a sterilized glass rod. The suspension was then filtered through two layers of autoclaved medical gauze. A hemacytometer was used to determine the concentration of the stock suspension of conidia and based on that count the suspensions were diluted to 10 5 conidia/mL. Bacterial strains were recovered from suspensions stored in 15% glycerol at -80°C, transferred to fresh Luria-Bertani (LB) liquid medium and cultured in darkness on a rotating shaker (120 rpm) at 28°C for 2 days. The bacterial suspensions were diluted with sterile water to a final concentration at 10 6 CFU/mL.

Bioassay for C. fimbriata VOCs against plant pathogens
The bioactivity of C. fimbriata VOCs on test pathogens was determined by measuring the growth of test organisms with double Petri dish dual-culture ( Fig 1A). The bioassay system was set up with the bottoms of two 9-cm lidless Petri dishes, which were laid in opposition and then sealed together with two layers of Parafilm (treatment, T). One plate containing PDA or CA, which was on the top of the dual-culture, was inoculated with a plug (diameter = 5 mm) of test fungi or oomycetes, or spread with 100 μL of test bacterial suspension at 10 6 CFU/mL on LB plate. The other one, placed at the bottom, was a 6-day-old culture of C. fimbriata (diameter of colony of C. fimbriata was about 4 cm) inoculated with a plug (diameter of plug = 5 mm) that produced a certain amount of VOCs. Three replicates were conducted, and a same dualculture system lacking C. fimbriata was used as the negative control for each pathogen. The cultures were incubated in darkness at 25°C for fungi or oomycetes, or 28°C for bacteria. The growth of the test organisms, including morphology of colony, growth of mycelia, inhibition of mycelial growth, and conidial production, was measured 5 to 7 days post inoculation (DPI) for the fungi and oomycetes as the diameter of the controls reached about 8 cm and 2 DPI for the The effect of the treatment on conidial germination of test fungi by the volatiles produced by C. fimbriata was also observed. 100 μL of a conidial suspension of test fungi at 10 5 conidia/mL was plated on PDA, and sealed with a 6-day-old culture of C. fimbriata as described above. An untreated by the volatiles produced by C. fimbriata was used as the negative control, with three fructicola growing on a PDA plate for 4 days. R: M. fructicola after 6-day exposure to the VOCs from C. fimbriata transferred to a fresh PDA plate and grown for another 4 days. (d) M. fructicola, exposed to 0, 1, 3, 5 and 7-day old cultures of C. fimbriata for 5 days, respectively. (e) CK: Penicillium digitatum growing on a PDA plate for 6 days. T: P. digitatum growing on a PDA plate with 6-day exposure to the VOCs from C. fimbriata. (f) CK: P. digitatum growing on a PDA plate for 4 days. R: P. digitatum after 6-day exposure to the VOCs from C. fimbriata transferred to a fresh PDA plate and grown for another 4 days. (g) CK: P. italicum growing on a PDA plate for 5 days. T: P. italicum growing on a PDA plate with 5-day exposure to the VOCs from C. fimbriata. (h) CK: Phytophthora capsici growing on a CA plate for 6 days. T: P. capsici growing on a CA plate with 6-day exposure to the VOCs from C. fimbriata. (i) CK: Acidovorax avena growing on a LB plate for 2 days. T: A. avena growing on a LB plate with 2-day exposure to the VOCs from C. fimbriata. replicates for each pathogen. The incubation conditions for these dual-cultures were the same as above, and the germination per 100 conidia was counted for each plate when most of the conidia in the control were germinated.

Inhibited percent ð%Þ ¼
Germination rate of control ð%Þ À Germination rate of treated fungi ð%Þ Germination rate of control ð%Þ Â 100% Next, a plug from the edge of each treated pathogen was transferred to a fresh plate (recovered treatment, R) and the growth, including colony morphology, mycelial growth, conidial production and spore germination, was compared to the untreated treatment at 5 DPI. Three replicates were carried out for each pathogen, and the incubation conditions were the same as described above. The experiment was carried out a total of twice.
In order to detect the relationship between C. fimbriata VOCs concentration and the inhibition effect, the test pathogens were inoculated with a fungus or oomycete plug (diameter = 5 mm) on PDA or CA and 100 μL of test bacteria suspension at 10 6 CFU/mL on LB, and treated with the same bioassay system by C. fimbriata cultures of 0, 1, 3, 5 or 7 days old, which had different concentration of VOCs. The treatment with no C. fimbriata inoculated was taken as the negative control in three replicates for each concentration. These cultures were incubated in darkness at 25°C for fungi or oomycetes or 28°C for bacteria. The mycelial or colony growth was measured 5 to 7 DPI for the fungi and oomycetes as the diameter of the controls reached about 8 cm, and 2 DPI for the bacteria. The work was done twice.

Control of peach brown rot and citrus green mold by VOCs from C. fimbriata
Experiments were conducted with commercially grown peaches and citrus bought from Xiyuan market in Beijing. Disease-and injury-free fruits were selected and surface sanitized in 3% NaClO solution for 5 min. The fruits were rinsed with sterile water three times then placed on a clean bench for aseptic drying. The inoculation method was modified as described [21]. For each fruit, three locations were wounded at the equator of equidistance with a sterile inoculating needle. The depth of wounds was about 5 mm. Then, 10 μL of conidial suspension of M. fructicola (10 5 conidia/mL on peach) or P. digitatum (10 5 conidia/mL on citrus) and sterile water were pipetted into two wound locations, leaving the third wound without inoculation as a control. The inoculated fruits were placed in 15 L plastic boxes with two layers of autoclaved medical gauze, soaked with 100 mL of sterile water, placed at the bottom to moisturize and prevent rolling (Fig 2A and 2B). The effect of exposure to C. fimbriata volatiles was detected by placing a specific number of 6-day-old (inoculated with a plus, diameter = 5 mm) C. fimbriata cultures in the boxes. The boxes were placed in an illuminated incubator at 25°C after being closed with fitting plastic lids. The control consisted of inoculated fruits in boxes without inoculated C. fimbriata cultures. The effect of application time for the treatment, VOCs dose and treatment period was determined with the treatment by various amounts of inoculated C. fimbriata cultures per box or different treatment periods: immediately, 24 h, or 48 h, respectively, after inoculation on the fruits. The infected number, lesion size, and control effect were measured and calculated after 4-day treatment. Each experiment was conducted in a completely randomized design with three replicate boxes with 10 fruits in each. For lesion size measurement, 10 fruits were randomly selected from the three boxes per each treatment. Lesion size mm 2 ð Þ¼p Â a 2 Â b 2 , a: the length of disease spot (mm), b: the width of diseases spot (mm).

Morphological change of the treated pathogens
The conidial and mycelial morphology of treated and recovered M. fructicola and P. digitatum was observed by a scanning electron microscope (SEM, HITACHI S-3400N) equipped with the control software HITACHI S-3400N Scanning Electron Microscope, provided by Functional Genomics Centre of platform 985, CAU, with normal cultures as controls. The PDA medium, with treated mycelia and conidia on its surface, was cut into small pieces (5 mm×5 mm) and fixed with 2.5% glutaraldehyde solution at 4°C overnight. The fixed treated mycelia and conidia with PDA medium was rinsed with phosphate buffer solution (0.1M PBS, pH 7.2) for 20 min, three times. The samples were then dehydrated in an ethanol gradient (30%, 50%, 70%, 80%, 90%, and 100% ethanol), three times for 20 min at each concentration. After the substitution treated with isoamyl acetate three times for 20 min each time, the samples were critical-point dried with CO 2 in an HCP-2 critical-point dryer (Hitachi, Japan) for 8 hours. Then, each sample was placed onto a metal stub with double-sided carbon tape and sprayed with gold ion sputter (EIKO IB-3) for 8 min (3 mA), so that a thin layer of gold was deposited on the surface of the sample. The morphology of mycelia and conidia in three kinds of growth states (treated, recovered, and untreated) were observed by SEM and compared by measuring 10 mycelial diameter, branch number of 100 mycelial tips and 10 conidial dimensions in random. The experiment was carried out a total of twice.

Pathogenicity of the treated pathogens
A method was designed to detect the pathogenicity of the treated and recovered M. fructicola and P. digitatum compared to normally growing cultures. The peach and citrus fruits were surface sterilized and wounded as described above. Since the treated and recovered fungi produced almost no conidia, a 2 mm plug from the edge of the untreated, treated or recovered colony was inoculated on one location of the three wounds on each fruit; the other two wounds were inoculated with 10 μL sterile water and non-inoculated controls. The infection number and lesion size of the two post-harvest diseases was measured 4 DPI. The pathogenicity test was conducted in a completely randomized design with three replicate boxes for each kind of inoculum with 10 fruits. This work was repeated.

Statistical analysis
Data analysis was performed using Excel 2010 (Microsoft) and SPSS statistical program (Version 17.0, International Business Machines Corp., Armonk, New York). Mean values were compared using Student's t test, at significance level α = 0.05.

Effect of C. fimbriata volatiles on various fungi, oomycetes and bacteria
All test fungi, oomycetes and bacteria were greatly inhibited but not killed after exposure to volatiles from C. fimbriata in vitro without direct contact or diffusion through the culture media ( Fig 1A). After 5 to 7 days treatment, the test fungi and oomycetes grew much more slowly than did untreated cultures. Inhibition of mycelial growth from 50% to 85% was observed (Table 2), which indicates that the sensitivity of each target microorganism to these volatiles varies.
The colony morphology of treated fungi and oomycetes appeared abnormal, with thinner hyphae and lighter pigmentation (Fig 1B, 1E and 1H). In addition, some fungi, such as P. italicum, lost some pigmentation after the exposure to C. fimbriata volatiles (Fig 1G). For most of the test fungi, no conidia were observed after 5-day treatment by C. fimbriata VOCs, except P. digitatum, which produced 5×10 4 conidia/mL ( Table 2). The inhibition of mycelial growth and conidial production for each colony of P. italicum could not be calculated, since more than one colony usually formed due to conidial release (Fig 1G). Spore germination of test fungi was also affected by treatment by C. fimbriata VOCs, with observed germination inhibition in the range of 66% to 87% (Table 2). Rhizoctonia sp. belongs to mycelia sterile and does not produce any conidia. V. mali hardly produces conidia under normal conditions and only zoospores are produced by Phytophthora sp., so no data was collected on the effect of the treatment on conidial production of these species. The inhibition of mycelial growth on test fungi and oomycetes, except P. italicum, showed a linear relationship (correlated coefficient (R 2 ) >0.9) with increasing incubation time and thus longer exposure to VOCs (Fig 1D). After exposure to C. fimbriata VOCs, all of the test bacteria were still alive, but none of them was able to produce normal discrete and round colony morphology; instead the colonies were irregularly shaped, and smeared to a bigger patchy zone (Fig 1I).
After the treatment by C. fimbriata volatiles, all test pathogens, including fungi, oomycetes and bacteria, grew after being transferred to a fresh plate. However, their growth speed was slow and their colony morphology abnormal after the 10 th generation (Fig 1C and 1F). The colony diameter of recovered pathogens growing on fresh plates was only 1.0 to 2.5 cm, significantly smaller than that of normally growing cultures (7.5 to 8.9 cm). No conidia were observed in most recovered cultures except in P. digitatum, whose conidial production and spore germination (2 × 10 4 conidia/mL and 12 ± 1%, respectively) were much lower than that of the control (Table 3). This suggests that the treated pathogens cannot recuperate on fresh media, even when free from the exposure to the VOCs from C. fimbriata.

Control of peach brown rot and citrus green mold
In the treatment without C. fimbriata volatiles (negative control), the fruits were 100% infected by M. fructicola or P. digitatum and visible rot or mold was observed on wound locations inoculated with conidial suspension (Fig 2C). In the treatments with C. fimbriata volatiles, no visible symptoms were observed on the fruits whether they were inoculated with conidial y and z = mean value ± standard deviation (SD). * indicates a significant difference between the control and the treatment by C. fimbriata volatiles (P < 0.05). doi:10.1371/journal.pone.0132009.t002 suspension or sterile water (Fig 2D and 2E). The disease control effect on peach brown rot and citrus green mold was 94% and 91% respectively when the VOCs treatment was performed immediately after inoculation with the conidia suspension, which was the optimum application time for the treatment; when treated 24 hours post inoculation (HPI) disease control sharply decreased to 35% and 34%, respectively. Disease control was only 15% and 9% respectively when treated 48 HPI for the two diseases (Fig 3).
The effect of VOCs dose on the two diseases was determined by fumigating the inoculated fruits with different amounts of C. fimbriata cultures in boxes for 4 days immediately after the inoculation. Untreated peaches and citrus were 100% infected, with lesion sizes of 909 mm 2 and 987 mm 2 respectively. Disease control for both of the plant diseases improved as the number of C. fimbriata cultures per box increased, with control as high as 90% and 98% with 9 plates and as low as 44% and 57% with three plates (Fig 4).
Control effect treated by various treatment periods was tested with 10 C. fimbriata cultures in each box immediately after the inoculation. The control effect of peach brown rot and citrus green mold increased as the time of the treatment lasted longer, with control at 92% and 97% at 96 HPI (Fig 5), while the control fruits were completely infected.

Morphological change of the treated pathogens
To evaluate the morphological change of the test pathogens treated by VOCs from C. fimbriata, the morphology of the treated mycelia and conidia was observed by SEM after 5 days treatment. In the control samples, M. fructicola grew healthily on PDA medium, presented regular and homogeneous mycelia with typical tapered apices (Fig 6A), and had normal barrelshaped conidia (Fig 6J). Both hyphae and conidia had a smooth appearing cell wall surface. The treated samples exhibited abnormal growth with irregular distortions. The severely degenerating hyphae were collapsed, curling, distorted or twisted (Fig 6B and 6C), with swollen Table 3. Mycelial growth, conidial production and spore germination of the recovered pathogens growing on a fresh plate transferred from treated cultures.

Species
Diameter of colony (cm) x Conidial production (conidia/mL) apexes with extensive branches (Fig 6D-6F), and the conidia were enlarged and distorted ( Fig  6K). After recovering on a fresh PDA plate, M. fructicola still had a similarly misshapen morphology with that of the treated fungus; the hyphae and conidia were swollen, branched and twisted (Fig 6G-6I and 6L). Measurement results showed that the diameter of M. fructicola and P. digitatum control hyphae was significantly wider than that of treated and recovered treatment. The conidia size of M. fructicola in treated and recovered cultures was significantly larger than that in the control. However, there was no significant difference among the treated, recovered or normal P. digitatum cultures. One hundred mycelial tips were observed in each treatment for M. fructicola and P. digitatum, and almost all of the observed hyphae in the treated or recovered cultures were branched, compared to that with regular morphology in the control ( Table 4).

Pathogenicity of the treated pathogens
The loss of pathogenicity determinants was another possible reason for the bioactivity of C. fimbriata VOCs on the test pathogens. In the control, all of the fruits inoculated with normal M. fructicola or P. digitatum were 100% infected with typical symptoms of peach brown rot and citrus green mold. However, no visible symptoms were observed on the fruits inoculated with treated or recovered pathogens (Fig 2F). A plug from the colony edge of M. fructicola treated with the C. fimbriata volatiles was transferred to the 10 th generation. The colony grew, but the growth rate was slow, and had the similar colony morphology compared to the first generation of treated culture. It is possible that the pathogenicity of M. fructicola and P. digitatum was lost or weakened after being treated by C. fimbriata VOCs due to damage to cell wall or organelles, which cannot recuperate when the treated pathogens were transferred to fresh culture media.

Discussion
VOCs from C. fimbriata have significant inhibition on the test fungi, oomycetes and bacteria. Misshapen mycelia and conidia were observed by SEM after treated with VOCs. These prominent morphological changes in treated samples were possibly due to the damage of the cell wall or the lack of cytoplasm [28,29]. And it also might be from the destruction of organelles in the endomembrane system [30]. Morphological alterations and damage on the hyphae or conidia were the possible reasons for the slow mycelial growth speed and low conidial germination. VOCs also inhibited pigment production, such as in Fusarium sp., which was also observed in other treated fungi by VOCs from bacteria [31]. Interestingly, the changed characteristics affected by VOCs remained even after 10 generations of transfer. If the morphological changes are carried over the generations, it is possible that genetic mutations happened. However, as we did not conduct further studies, it is too early to make such a conclusive statement, but is definitely worth to examine in future work. Exposure to VOCs from C. fimbriata showed significant inhibition of test fungi, oomycetes and bacteria in vitro and excellent control of peach brown rot and citrus green mold. Unlike some traditional biological control agents, which must colonize wounds or some other susceptible sites to be effective [32], the VOCs from C. fimbriata, acting as a biofumigant, does not require contact. Such treatment may be more compatible with integrated disease or pest management systems than current biological or chemical fungicides that require spraying or drenching applications, causing problems such as fungicide residue, air pollution, and water purification [21]. The VOCs from wild-type strain F. oxysporum MSA 35 repressed gene expression of two virulence genes in F. oxysporum f. sp. lactucae [33]. It can be deduced that there also might be some genes associated with pathogenicity affected by VOCs from C. fimbriata.
We have qualitatively and quantitatively determined the VOCs from C. fimbriata by headspace GC-MS (unpublished data). Seven compounds were identified, and the three most abundant were butyl acetate, ethyl acetate and ethanol, which accounted for 97% of the total VOCs yield. Mixtures of one or more of these pure chemicals were used to simulate the treatment on the test pathogens at the approximate concentration of VOCs naturally produced by C. fimbriata, but there was no obvious inhibition effect observed. Inhibition may be due to the synergistic effect of all VOCs from C. fimbriata, including those that cannot be detected by current identification methods. This is similar to results from other VOCs producing fungi, such as Muscodor sp. [21,34]. Since C. fimbriata is also a soil-borne pathogen, the authors proposed that the practical application could be to screen a mutant without pathogenicity or to incubate this fungus in a container and transfer the VOCs produced into a separate airtight container by pump, which should avoid the risk of possible infection by C. fimbriata by direct contact with agricultural products or through the soil.
There are few studies on antimicrobial volatiles from C. fimbriata, and to the best of our knowledge, this is the first report on the potential of C. fimbriata against other plant pathogens and plant diseases control. The bioactivity of VOCs from C. fimbriata suggested that this biofumigation could be applied for diseases control under airtight conditions, not only on fresh fruits but also in cereals, vegetables, seeds, and seedlings.