Instant biofumigation using natural products from papaya seed waste for sustainable management of soil-borne plant pathogens

This project started in summer 2020. During this first report period, due to the COVID restrictions and financial challenges facing our farmer collaborators as a result of COVID, we are unable to initiate the proposed field trials on participating farmers’ farms. However, we are able to work with one of our collaborating farmers, Mr. Owen Kaneshiro of the Owen Kaneshiro Farm, Waianae, HI. Using soil collected from various fields on the Kaneshiro farm (picture 1), and through discussions with our collaborating farmer partner, we designed and conducted greenhouse studies to determine the efficacy of the papaya seed biofumigant against Fusarium oxysporum f. sp. letucae and root-knot nematode, Meloidogyne incognita, under different biofumigant formulations and application regimes, using lettuce and kai choi as test crops. Laboratory studies have also been conducted to optimize the preparation of the PGS as biofumigants, and to investigate the impact of PGS application on soil bacterial/archaeal communities.

Papaya seeds contain a high level of BG that can be enzymatically hydrolyzed via myrosinase in aqueous solution to form BITC which has potent pesticide activities against a range of soil-borne phytopathogenic nematodes, insect pests, and fungi. The binary glucosinolate/myrosinase system requires activation by disrupting the cellular compartments that naturally separate the glucosinote from the myrosinase, and by the inclusion of water to trigger the enzymatic hydrolysis. In our previous study (Han et al., 2018) we noted a significantly elevated level of BITC in the oil extracted from papaya seeds that were previously frozen compared with that found in oil extracted from non-frozen seeds. Since BITC is the active ingredient of the papaya seed biofumigant, we tested whether it was necessary to pre-freeze the seeds to obtain high BITC in the PGS. To that end, prior to the greenhouse pot trials, we investigated processing conditions for preparing the PGS and measured BG and BITC contents in the PGS. The factors examined included (i) freezing seeds prior to drying and grinding; (ii) seed drying temperatures; and (iii) seed drying duration.  One group of fresh seeds were frozen at -20°C and then dried at 50°C for 48 h, before being ground in an electric grain mill grinder (S1). A second group of papaya seeds were dried at 50C for 48 h without a prior frozen treatment (S2), and a third group of seeds were oven dried at 40°C for 24 h without a prior frozen treatment (S3).  The resulting PGS samples were reacted with an equal weight of water at room temperature for 6h to synthesize BITC, and seeds oil containing BITC was finally extracted with hexane and analyzed using the cyclocondensation assay and HPLC quantification as described in the Materials/Methods section. The results are shown in Fig. 1. Samples S1 and S2 gave essentially the same amount of BITC after activated with water, indicating freezing of papaya seeds prior to drying and grinding into PGS is unnecessary. BITC is lower in sample S3 than the other two treatment groups (p< 0.05). Drying the seeds for 24 hours at 40°C did not remove sufficient moisture from the seeds, and the resulting PGS formed small clumps which may have reduced the enzymatic conversion of BG to BITC.  Overall, our results indicate that directly drying the papaya seeds at 50°C for 48 hrs without prior freezing is adequate to generate PGS with reproducible BITC content. Further, activation of PGS with equal volume of water for 6 hours at room temperature is sufficient to convert most of the BG to BITC.

We also measured the BG and BITC contents in seeds of two common papaya varieties Rainbow and Sunrise, and tested whether seed BITC synthesis is affected by dry storage. The BG content was averaged about 20-25 µmole/g dry seeds regardless the papaya varieties. As shown in Figure 2, the BITC generation did not vary considerably between the two varieties.  These results also demonstrate that BITC generation from seeds freshly harvested from fruits is similar to that from seeds dried and stored in an air-tight Ziploc bag for 1 month. 

Based upon these results, PGS for subsequent greenhouse trials was prepared by drying papaya seeds at 50°C for 2 days, milled into powders, and stored in Ziploc bags at room temperature prior to use in the trials. We also analyzed the nutrient profile of the defatted PGS. The proximate analysis result is presented in Table 1.

Table 1. Nutrient profile of defatted PGS.

In greenhouse Trial I, phytotoxicity of lettuce seedlings occurred in all treatments receiving PGS and resulted in plant growth similar to the NA (Picture 2; Fig. 3). However, BM was able to suppress M. incognita and resulted in lettuce growth similar to the autoclaved treatment (without M. incognita inoculation and no fungal infestation) (Fig. 3).

In Trial II, PGS was amended to the soil one week prior to the seedling transplanting. While the canopy width of kai choi was increased by BM compared to the NA control, the shoot and root weights of kai choi in BM amended soil were not different from NA control (Fig 4 B, C). All PGS treatments (PGS 0.5%, PGC 1%, PGC+CE) did not pose phytotoxicity on kai choi as the canopy width, shoot and root weight were all not different from the NA control (Fig. 4 A, B, C). Unfortunately, all PGS treatments and brown mustard (BM) amendment did not reduce the plant disease index using a scale of 1-4 (Fig. 4 D). However, all biofumigation (PGS and BM) significantly (P ≤ 0.05) reduced the number of root pieces infected by F. oxysporum (Fig. 4 E) and the number of M. incognita penetrated into kai choi roots (Fig. 4 F). No Rhizoctonia hyphae was observed on the Komada medium. Initial attempts to monitor the fate of BITC in the PGS-amended soil during the trial produced highly varied results and no clear conclusion can yet be drawn. Uniformity in soil sampling is identified as a potential factor that needs further investigation.  It is also noted that PGS has a relatively high oil content (about 25% per dry seed weight), while BITC is hydrophobic, and hence BITC is likely partitioning to the oil phase, and its release may be affected by gradual seed oil biodegradation in the soil over time. These issues will be further investigated to achieve a more effective PGS application regime as a soil biofumigant.

Trial III was collected from another lettuce field with a history of Fusarium wilt infestation. Same treatments as Trial II were imposed and 3 Manoa lettuce seedlings were planted into each pot for bioassay over 1 month. Lettuce plants responded with positive growth from the NA control when planted in soil amended with BM or 1% PGS (P ≤ 0.05, Fig. 5 A). Only BM resulted in wider canopy width and total shoot biomass higher than NA (Fig. 5 B, C). It is also promising in this trial that biofumigation treatments resulted in a lower plant disease index compared to the NA control (P ≤ 0.05; Fig. 5D). Deviate from the results in Trial II, PGS 0.5% and PGS 1% were the only treatments that suppressed the recovery of kai choi roots with F. oxysporum on the Komada medium (P ≤ 0.05, Fig 5E) much like the autoclaved soil. Whereas, root gall index (1-5 scale) on lettuce was significantly suppressed by all biofumigation treatment including PGS+CE compared to the NA control (P ≤ 0.05, Fig 5F). From this trial, PGS showed promise as a potential sustainable biofumigant. 

To study the effect of PGS treatment on soil microbial diversity, twelve soil samples, triplicate samples from four different treatments (Table 2), were collected from greenhouse Trial 1 and then subjected to 16S rRNA gene amplicon sequencing. Bioinformatics analysis identified a total of 510 bacterial/archaeal genera across the twelve samples. The 40 most abundant genera, which were those operational taxonomic units (OTUs) with relative abundance no less than 1% in at least one sample, were summarized in Figure 6.

Table 2. Sample ID and treatment conditions for twelve samples subjected to sequencing

Analysis of the alpha diversity, including the observed OTUs and Shannon index, were shown in Figure 7A and 7B, respectively. No significant difference among the four treatments were observed with all p-value of Kruskal-Wallis tests higher than 0.1.

Analysis of beta diversity, including the Jaccard distance and unweighted UniFrac distance, were shown in Figure 8A and 8B, respectively. Different treatments showed significant microbial community dissimilarity, with p-value of 0.021 in PERMANOVA test of unweighted UniFrac distance. Additionally, all the three treatment (0.5 CE, 0.5 PGS and PGS1) showed significant microbial community dissimilarity than the control treatment (CK), with p-value of 0.10, 0.09 and 0.10 in pair-wise PERMANOVA test, respectively.

Based on the taxonomic classification of 16S rRNA gene, eight previously-reported plant growth promoting bacteria (Genus/Species) were identified in the soil samples, shown in Figure 9. Different treatments resulted in significant difference in the abundance of plant growth promoting bacteria (ANOVA test, p=0.022), shown in Figure 10. All the PGS treatments (0.5 CE, 0.5 PGS and PGS1) showed significant higher total relative abundances of the eight plant growth promoting bacteria than that in the control treatment (CK), with p-value of 0.002, 0.054 and 0.013 in pair-wise ANOVA test, respectively.

References cited:

Han, Z., Park, A., Su, W.W. 2018. Valorization of papaya fruit waste through low-cost fractionation and microbial conversion of both juice and seed lipids. RSC Advances, 8(49), 27963-27972.

**********

During our second year we collected data pertaining to (1) a field trial of PGS on kai choi in Fusarium-infested soil on a collaborating farmer’s farm (Owen Kaneshiro Farm, Waianae, HI), (2) in-vitro toxicity tests of PGS towards various fungal pathogens, and (3) laboratory studies of biofumigant (BITC) emission from PGS to develop better practices for more efficient use of PGS in the field implementation of PGS-based biofumigation.

Field Trial

A. Biomass and Marketable Yield

Total biomass and marketable yield were measured to determine overall effectiveness of the PGS treatments at preserving the crop for market and maintaining healthy plant growth (Picture 4). All treatments were significantly less in mass for both the total biomass and marketable yield than the conventional Vapam-treated plots within the experiment. However, this included the mustard treatment, which showed no significant difference from the PGS treatments as well (Fig. 11). Considering mustard was an effective biofumigation crop (Waisen et al., 2019) previously tested in the same farm and the positive effects of PGS in our greenhouse pot trials, the current field results suggested that trench application was not an effective approach to deliver PGS or macerated brown mustard.

B. Disease Incidence

Diseased plants were counted to determine the overall effectiveness of PGS and other tested amendments in reducing Fusarium wilt disease incidence in a field infested with F. oxysporum (Picture 5). Of the six treatments, Vapam had the least amount of disease, significantly lower than all other treatments. Both PGS 0.5% and PGS 1% were significantly higher in disease incidence, but comparable to the brown mustard treated plots (Fig. 12). The Sorghum planted plots (not till into the soil) had less diseased than the PGS-treated plots, but not different from the control and mustard. These results suggested that both trench application of PGS and macerated brown mustard and no-till cover cropping of sorghum were not effective biofumigation approaches against the targeted soil-borne pathogens. Future study will examine more effective and feasible biofumigation methods using PGS.

C. Fusarium colonization of kai choi roots

To observe whether or not PGS was able to suppress F. oxysporum and prevent root colonization, kai choi roots collected from the field experiment were sterilized and plated on Fusarium-selective media, then observed for any Fusarium-like mycelial growth protruding from the plated roots (Picture 6).

Control, mustard, Vapam, and 1% PGS treatment groups were all significantly lower in Fusarium growth than the 0.5% PGS and Sorghum treatment groups (Fig. 13). Most surprisingly, there is very little difference between the control (negative) treatment group and the Vapam (positive) treatment group. This is likely due to certain strains of F. oxysporum being known to colonize plant roots as avirulent endophytes in various plants, so we can’t be sure if what emerged from roots on the Komada media are pathogenic or avirulent, symbiotic F. oxysporum strains. In order to verify findings or produce more accurate results, we will develop a qPCR assay specific to the F. oxysporum forma specialis that is pathogenic in kai choi.

Picture 6. Surface-sterilized kai choi root pieces plated on Fusarium-selective Komada media (left) and same kai choi root pieces on Komada media with suspected Fusarium growth (right).

D. Nematode community as soil health indicators

A primary reason for assessing alternative biofumigant methods is due to the harm that Vapam inflicts on the natural soil ecosystem when used. To make this comparison and measure the soil health of PGS-treated plots versus Vapam-treated plots, free-living nematodes were extracted and counted. Total free-living nematode counts were significantly higher in the Sorghum-treated plots compared to all other treated plots. The control, mustard, 1% PGS, 0.5% PGS, and Vapam treated plots were all comparable in overall nematode counts (Fig. 14).

Most treatment groups were comparable in species richness (number of different species present) measures, however the Vapam plots were significantly less species rich (Fig. 15). This is likely due to the that Vapam indiscriminately kills everything in the soil that it is used in (nematodes, fungi, beneficial microbes, etc.), thus creating a uniformity in the species present in the soil as the species that are able to bounce back quickly. It’s unlikely in these conditions to observe and record and long generational/large nematodes in the soil samples. So, despite not seeing a large difference among the PGS, sorghum, mustard, and control treated plots, there is already an observable increase in species richness in the soil not treated with Vapam. This is an indicator that in one crop rotation without the synthetic fumigant, the soil is becoming healthier. 

PGS biofumigation plate assay

In our initial plate assay design, fungal cultures were inoculated onto the PGS-amended agar plates (PGS was blended in and encapsulated in the agar) and incubated for subsequent colony count. We encountered microbial contamination problems with this assay format. Some of the papaya seed indigenous microbes apparently survived the seed drying and milling processes and grew in the PGS-amended agar plates. Subsequently, we developed an alternative plate assay, in which the agar plate was hollowed out in its center and water-amended PGS was used to fill the hole. As a PGS-free control, the hole was either not filled in with the PGS, or without creating the center hole. This assay format worked well, showing little signs of contamination from the indigenous papaya seed microbes, and was adopted for subsequent in-vitro bioassays. Of the three fungi tested, the F. oxysporum plate assay showed greatest overall significant difference in growth, reducing the mycelial growth to almost zero compared to the control group (Picture 7, top left; Fig.16). Thus, it appears that PGS biofumigation is highly effective against F. oxysporum in vitro. However, both F. solani and Setophoma sp. had no significant difference in growth between the control and two treatments groups, with the PGS biofumigation appearing to have no effect on the mycelial growth (Picture 7, top right and bottom; Fig. 17). For Setophoma sp., there was a slight difference between the control group and the group treated with 0.5g PGS + 0.5g water, but this was not easily observed (Picture 7, bottom; Fig. 17). 

Emission of BITC from PGS

In intact papaya seeds, cellular membranes separate glucosinolate (the substrate) from myrosinase (the enzyme), and hence efficient breakdown of the membrane barriers is essential to promote the enzymatic hydrolysis upon water addition. One of the simplest approach to break down the membranes is to mill the dried seeds into powders, called “PGS” herein. This method is inexpensive, though it may not completely degrade all membrane barriers in the seeds. In our prior study (Han et al. 2018), we could detect ample amounts of BITC after extracting water-activated PGS using organic solvents. The result indicated that high levels of BITC could indeed be produced by PGS upon water activation. However, since organic solvents were used to extract the BITC, we do not know the extent BITC can be released by itself from the PGS without extraction. In year 2, we designed and evaluated a series of laboratory experimental tests, to gain a better understanding about the temporal pattern and extent of BITC emission from water-activated PGS, and to investigate how this process is affected by PGS water content, PGS particle size, freezing pretreatment, and inclusion of soil. These fundamental characterizations are essential to develop better practices for more efficient implementation of PGS-based biofumigation in the field. Because currently we do not have a gas chromatography system for measure BITC in the vapor phase, we had to develop an alternative assay platform based on measurement using liquid chromatography. In this regard, since BITC is a hydrophobic compound that shows good solubility in plant oil (triacylglycerol) and mineral oil (alkanes), we tested commercial soybean oil and mineral oil as a liquid trap for BITC emitted from PGS, and the oil was then analyzed for BITC using HPLC. The transport of BITC from water-activated PGS to the vapor phase surrounding the PGS involves multiple steps. First, the addition of water to the PGS activates the myrosinase enzyme which catalyzes the hydrolysis of benzyl glucosinolate (BG) in an aqueous phase. Second, the BITC thus generated is hydrophobic, and hence it tends to partition into the seed fat, or emit to the vapor phase (interstitial air space) followed by eventual release into the headspace above the PGS bed. It may also be absorbed onto PGS particle surfaces as it diffuses through the bed of PGS.

In our tests, we typically use 2 g of dried PGS in each test. We first examined the BITC trapping efficiency by soybean oil vs mineral oil. A water:PGS ratio of 3 was used so that a thin layer of water is viable above the PGS bed which was overlaid with 2 mL of oil. The tube was incubated under room temperature for 40 hours before the oil samples were taken and analyzed using HPLC. As shown in Fig. 18, mineral oil appeared to be a more effective BITC trap than the soybean oil. More importantly, this result provided direct proof that not only BITC can be generated by PGS upon water activation, but also it can be released out of the PGS matrix and migrate into the headspace. Next, we investigated the temporal pattern of BITC release from PGS. Using mineral oil as the trap, and with 2 grams of PGS plus 6 grams of water, the BITC continued to partition into the oil trap over the 32-hr testing period, as shown in Figure 19. We also investigated the effect of soil on BITC release. Sorption of released BITC onto soil surfaces could potentially reduce the efficacy of BITC. In one treatment, 2 g of PGS was mixed into 4 g of soil and placed in a 50 mL centrifuge tube, followed by adding sufficient water until a thin layer of water was viable above the soil/PGS bed. In the other treatment, 2 g of PGS was placed in the bottom of a 50-mL centrifuge tube, overlaid with 4 g of soil, followed by adding water until a thin layer of water was viable above the soil/PGS bed.  In both treatments, 2 mL of mineral oil was overlaid on the water layer.  As shown in Figure 20, by dispersing PGS in soil, much higher BITC emission was detected, which could have implications on how to apply the PGS in soil in the field.  Because in these tests, the interstitial space in the PGS bed was filled with water, BITC needs to migrate through this aqueous phase before it can partition into the oil trap on the top. Typically, PGS-amended soil in the field is not waterlogged, the hydrophobic BITC is likely to diffuse more readily in the air space between soil particles and hence greater and faster BITC release is expected in the field. In summary, our results provide direct proofs of BITC release from water-activated PGS, and that the release process likely lasted for several days. The vapor pressure of BITC is two orders lower than that of allyl-isothiocyanate (AITC, active biofumigant in mustard), and hence slower release than AITC is expected.

Water plays an important role in the conversion of BG to BITC in PGS. On the one hand, water is necessary for the myrosinase to function. However, too much water may rinse out BG (which is water soluble), and may cause BITC degradation. Too much water is also harmful to the crops. We examined BITC release at four different water:PGS ratios (0.5, 1, 2, and 3) using the oil-trap assay. At a 2:1 water:PGS ratio, a higher BITC production was noted (36.05±12.65 mM) compared to our initial treatment (3:1 water:PGS; 22.51±0.43 mM). At further reduced water:PGS ratios of 1 and 0.5, the BITC production was even higher (120.14±1.80 mM and 108.58±36.71 mM, respectively). However, we likely cannot judge the water effect solely on these data. As the water content was decreased, oil on the top seeped into the PGS bed after several hours of incubation. Therefore, some of the oil was in direct contact with the PGS matrix, and BITC likely partitioned into this oil phase once it was produced from BG hydrolysis. This result however suggests addition of some oil (such as vegetable or mineral oil) to the PGS when it is amended to the soil might be beneficial in modulating the release characteristics of BITC. Partition of BITC in an oil phase can help stabilizing the hydrophobic biofumigant and prolong its release from the PGS. This aspect will be investigated further during the next report period.

To investigate two additional potentially important factors, i.e., PGS particle size and seed freezing, we slightly modified the assay. We blended 2 g PGS with 2 g of water and 2 g of mineral oil in a 50 mL centrifuge tube and incubated for 18 hours. The tube was then centrifuged, and the oil sampled for BITC measurement.  As shown in Figure 21, PGS with finer particle size generated more BITC than coarser PGS.  Freezing treatment of papaya seeds prior to drying and grinding into PGS also improved BITC release from PGS (Figure 22). Finer PGS particles likely promoted contact between BG and myrosinase by breaking down tissue/membrane barriers in the papaya seeds and also increased interfacial area for better mass transport. By freezing fresh seeds, further degradation of internal tissues in papaya seeds may occur which can augment breakdown of cellular membrane barriers and promote BITC production.

References cited:

Waisen, P., B. S. Sipes, and K.-H. Wang. 2019. Potential of biofumigant cover crops as open-end trap crops against root-knot and reniform nematodes. Nematropica 49:254-264.