Effect of Aureobasidium Pullulans S-2 on the Postharvest Microbiome of Tomato During Storage

Background: Biological control of fruit postharvest diseases by antagonistic microorganisms has been considered an effective alternative to chemical fungicides. The inuence of microbial antagonists on fruit-associated microbiome will provide a new perspective for in-depth study of the antagonistic mechanism. In this study, the biocontrol ecacy of A. pullulans S-2 against postharvest diseases of tomatoes was investigated. Meanwhile, the fungal and bacterial microbiota on tomato surfaces were examined by high-throughput sequencing. Results: A. pullulans S-2 can signicantly inhibit the decay rate, maintain fruit rmness and reduce weight loss of tomatoes. In addition, the treatment group can maintain higher titratable acid, ascorbic acid and lycopene than the control group. After using A. pullulans S-2, more dramatic changes were observed in fungal diversity than bacterial in the microbiota. Aureobasidium was signicantly enriched in the treatment group, while Cladosporium, Mycosphaerella, Alternaria and Penicillium were decient compared with the control group. Pantoea, Brevibacterium, Brachybacterium, Serratia, Glutamicibacter and Pseudomonas also had signicant differences between the two groups. Conclusions: This study demonstrated that the application of A. pullulans S-2 resulted in alterations in the bacterial and fungal community and that could inhibit pathogens and decrease fruit disease incidence. It provides new insights into the dynamics of the tomato's surface microbiome after microbial antagonist treatment.


Background
Tomato (Solanum Lycopersicon L.) is an important fruit that is widely grown globally, whose annual production is more than 180 million tons in 2019 (http://www.fao.org/faostat/en/#data/QC). However, during harvesting and transportation, the tomato will be easily damaged due to its fragile skin and soft esh, the disease development results in a substantial economic loss [1]. The primary diseases of postharvest tomatoes are gray mold caused by Botrytis cinerea [2], black rot caused by Alternaria alternata [3] and so on.
Chemical fungicides are still the most common method to control postharvest decay. Still, the long-time dependence on chemical fungicides has resulted in fungicide residues, environmental hazards, and pathogen resistance. Recently, the use of microbial antagonists and their active compounds to control postharvest decay of fruits has been a research hotspot, and this eco-friendly method may replace chemical fungicides in the future [4].
The microbial antagonists are mostly isolated from the plant and soil, and their ability of rapid growth and colonization in wounds is the main criterion. Yeast and yeast-like fungi occupy a very important position in the research of microbial antagonists. Their biocontrol mechanisms mainly consist of competition for nutrients and space, induction of resistance mechanisms and secreting antimicrobial substances [5]. Aureobasidium pullulans, a yeast-like fungus, has been used in postharvest disease prevention and control of several fruits such as avocado, citrus, peach, apple [6-8], It can antagonize several pathogens, such as Penicillium spp., Botrytis cinerea, Phytophthora cactorum [9][10][11].
Previous research mainly focused on the aspects of microbial antagonists, pathogens and fruits to elucidate the mechanism. However, an important part, where the microorganisms act on the surface, is ignored [12,13]. Microorganisms play an important role in the growth of plants, including bene cial, pathogenic and other microorganisms. After harvest, the microbial community dynamics could change due to the factors related to fruit physiology and other abiotic effects, including packaging, storage, and other postharvest treatments [14]. The changes of microorganisms on the fruit surface are closely related to fruit rot during storage [15]. It is generally believed that high microbial diversity and resilient microbiome structure are bene cial for fruits and vegetables. The occurrence of postharvest diseases of fruits is due to the proliferation of pathogenic fungi disturbing the microbial balance [16].
In recent years, several reports were demonstrating changes in microbial communities on fruit surfaces following different physical or chemical treatments. For example, the combination of ozone treatment and polyethylene packaging [17], hot water [18], waxing [19]. These researches have shown that a deeper understanding of changes taking place on fruit surfaces after harvest could provide new ideas for applying post-harvest control methods. However, there is limited information on the effect of microbial antagonists on microbial communities on the surface.
In this study, we screened a strain, A. pullulans S-2, from healthy tomatoes and used it on tomatoes to observe whether it could reduce the incidence of diseases during storage and examine its in uence on the quality of tomatoes. The changes in the bacterial and fungal communities on tomato fruit surface were also investigated using amplicon sequencing of the 16S and ITS conserved regions in the rDNA.

Tomato
The tomato fruits of the cultivar 'Provence' (introduced from the Netherlands) were harvested at the redripe stage from Hanya Organic Farm (32°11'N, 119°27'E), Zhenjiang City, Jiangsu Province, China. After harvest, fruits were packed in foam boxes and transported to our laboratory within one hour time.
Medium size tomatoes with similar maturity indices were selected for experiments.

Biocontrol tests on tomato fruit
A. pullulans S-2 was cultivated in nutrient yeast dextrose broth (NYDB: 8 g nutrient broth, 5 g yeast extract and 10 g dextrose in 1 L of distilled water) at 180 rpm, 28°C for 24 h (two consecutive culturing). Then, the yeast was centrifuged at 5000 × g for 5 min, and washed with sterile water twice, nally suspended in sterile physiological saline (0.85% NaCl). The concentration of cells was adjusted to 1 × 10 8 cells/mL using a binocular microscope and a hemocytometer.
The tomatoes were randomly divided into two groups. The rst group of tomatoes was dipped in A. pullulans S-2 suspension of 1 × 10 8 cells/mL for 1 min designated as the treatment group. The control group was dipped in sterile physiological saline. Treated tomatoes were then kept in a plastic box (48 × 38 × 15 mm) to dry. After drying, the boxes were wrapped with plastic lm to maintain high humidity, and incubated at 10℃ for 8 d and then transferred to shelf life for 4 d at 20℃. Decay incidence and severity, and fruit quality indices ( rmness, weight loss rate, titratable acid, ascorbic acid and lycopene level) were determined.

Firmness
TA-XT2i Texture Analyser (Stable Micro Systems, UK) was used to test the rmness of tomatoes [20]. TA settings were: P5; the test speed was 1 mm/s, the pre-test and after-test speed were 5 mm/s, and displacement was 5 mm. Each tomato was tested at three sites in the equatorial area, and nally, the maximum force (N) was recorded.

Weight loss
The weight of each tomato during storage was recorded, and the percentage of fruit weight after storage relative to the initial fruit weight was calculated.

Titratable acid
Ten grams of tissues from the sarcocarp of 6 tomatoes were ground into a homogenate with 20 mL distilled water. After 1 h, the homogenate was centrifuged at 7000 × g for 10 min to get the supernatant. The supernatant was transferred into a 100 mL volumetric ask and used as the nal extracting solution. 1% phenolphthalein (indicator) was dripped into 10 mL extracting solution and titrated against 0.1 M NaOH [20].

Ascorbic acid
Twenty grams of tissues from the sarcocarp of 6 tomatoes were crushed with 20 g/L oxalic acid solution and diluted to 100 mL. After which, it was centrifuged at 4°C, 7000 × g for 10 min, and the ltrate was collected. Ten milliliters of the ltrate were titrated with 2,6-dichlorophenol indophenol dye [21] and the volume of the titrant was recorded. Ten milliliters of 20 g/L oxalic acid solution were used as a blank.

Lycopene
Lycopene can be extracted with petroleum ether solvent without light [22]. Ten grams of tomatoes were homogenized, then 25 mL water and petroleum ether were added respectively into the homogenate. After 30 min, the supernatant was extracted and 25 mL petroleum ether was added to the homogenate again for extraction. The two supernatants were mixed and the absorbances at 472 nm were recorded.

DNA extraction and amplicon sequencing
Microorganisms on the tomato surface were isolated at 0, 4, 8, and 12 d during storage. The tomatoes were washed with a sterile PBS solution (pH 7.2), and the washed-out solution was collected. Then the solution was ltered through a microporous membrane (diameter 50 mm, pore diameter 0.22 µm) by vacuum ltration [13].

Sequence data analysis
To obtain more clean tags, the adaptor sequences, low-quality sequences, and poly N-containing reads were ltered by QIIME. Then, the effective tags were clustered into operational taxonomic units (OTUs) with ≥ 97% similarity using the UPARSE pipeline. The representative sequence (the highest abundant tag sequence in each OTU) was classi ed into organisms by a naive Bayesian model using an RDP classi er based on the SILVA database [26] and ITS2 database [27], with the con dence threshold value of 0.8.
Alpha diversity indexes were calculated by Welch's t-test and Wilcoxon rank test in R project Vegan package. Considering the presence or absence of species and changes in species abundance, the principal component analysis diagram of β diversity was obtained based on the Weighted UniFrac algorithm. An online platform of OmicShare tools was used to analyze the sequence data (http://www.omicshare.com/tools).

Statistical analysis
Data were analyzed by Statistical Program SPSS (PC) ver. II. X (SPSS Inc. Chicago, Illinois, USA), and P < 0.05 was considered to be statistically signi cant. All experiments were performed in three technical replicates and three biological replicates.

The effect of A. pullulans S-2 on disease development and fruit quality
Tomatoes in the control group began to rot on the 8th day, with a rot rate of 5.56%, and the rot rate increased to 59.46% on the 12 d (Fig. 1A). While the tomatoes in the A. pullulans S-2 treatment group did not rot on 8 d, the rot rate on the 12 d was 35.56%, which was signi cantly lower than the control group. Although the tomatoes were separated from the plant, they also undergo various physiological activities, and the weight loss rate gradually increases. At the end of storage, the weight loss rate of tomatoes in the control group reached 1.79% and that in the treatment group was less than 1% (Fig. 1B). The rmness of tomatoes changes due to the loss of moisture. The decrease rate of rmness increased (Fig. 1C), but the rmness of the treatment group was always higher than that of the control group. At the same time, the titratable acid (Fig. 1D), ascorbic acid (Fig. 1E), and lycopene (Fig. 1F) of the treatment group were signi cantly higher than those of the control group during the entire storage period.

Alpha and beta diversity on tomato surfaces
The Ace index was used to re ect the alpha diversity, and the species abundance and changes in the two groups during the entire storage period were observed by the box diagram. From Fig. 2A, the Ace index of the treatment group was signi cantly lower than the control group due to the intervention of A. pullulans S-2 on 0, 4, 8 d. However, as the storage time prolonged, various pathogenic fungi in the control group became active, disrupting the balance of tomato surface species, resulting in a decrease in the species diversity of the control group and a decrease in the Ace index. At the same time, due to the activity of pathogenic fungi in the late storage period, the surface diversity of the treatment group increased, and the value of Ace index became larger.
According to the alpha diversity of the bacterium (Fig. 2B), the Ace index of the control group was signi cantly greater than that of the treatment group (P 0.05) on 0, 4 d. However, there is no signi cant difference in the Ace index between the two groups (P 0.05), which indicates that the usage of A. pullulans S-2 caused a decrease in the species abundance of bacteria in the early stage, but the species diversity increased with the extension of storage time in the later stage.
For beta diversity, from the PCoA chart of the fungi (Fig. 3A), the explanation degree of PCo1 was 90.47%, and that of PCo2 was 8.95%. The sum of the two was greater than 50%, and which makes the explanation degree enough. From the whole point of view, the control group is far away from the treatment group during the entire storage period, which means that the ora structure between the two groups is not similar. At the same time, within 0-8 d, the structure of the respective ora of the two groups was close and similar. On the 12 d, due to the decay, the ora of the treatment group and control group altered, which was far away from the point in the previous 8 d, and the ora structure was different.
According to the PCoA diagram of bacterium (Fig. 3B), the explanation degree of PCo1 was 48.16%, and that of PCo2 was 33.84%. The explanation degree of the two was relatively similar, and the sum was greater than 50%. Analysis from PCo1, PY4 is far away from the other groups, and the other groups are close. Focusing on PCo2, PCK0, PY0; PCK8, PY8; PCK12 and PY12 are very close. From the above results, we witnessed that the structures of the bacterial colony between the two groups did not differ signi cantly during the entire storage period.

The in uence on fungal community
According to the analysis of fungal phylum (Fig. 4A), the Ascomycota accounted for more than 99%, and the Basidiomycota constituted less than 1%. Then focus on the fungal genus category (Fig. 4B). Aureobasidium, Cladosporium, Mycosphaerella, Alternaria, and Penicillium were important genera during storage.
From Fig. 5A, in the treatment group, the change curve of Aureobasidium was stable within 8 d, and the proportion was always higher than 90%, then decreased to 79.11% on 12 d. The trend of Aureobasidium was maintained at a low level for 8 d, and increased slightly in the later storage period of the control group. Throughout the storage process, the Aureobasidium of the treatment group was signi cantly higher than the control group.
The proportions of Cladosporium in the two groups were stable for the rst 8 d (Fig. 5B); the control group was always higher than 60%, and in the treatment group, it was less than 5%. In the last four days, the control group rapidly dropped to 29.4%, and the treatment group increased to 12.92%. Cladosporium of the control group was always more than the treatment group during the whole storage period.
The trend curve of Mycosphaerella (Fig. 5C) and Alternaria (Fig. 5D) in the control group were increased initially and then decreased. In contrast, Penicillium (Fig. 5E) was decreased initially and then increased. The proportions of these three genera in the treatment group were very low, even Alternaria and Penicillium were always less than 1%. The contents of these three genera in the control group were signi cantly higher than the treatment group during the entire storage period.

The in uence on bacterial community
From the perspective of bacterial phylum taxonomy (Fig. 6A), Proteobacteria, Actinobacteria and Firmicutes were the most important phyla during the entire storage, accounting for 72%, 19% and 2%, respectively. Among them, the dynamic trend of Proteobacteria (Fig. 7A) was increased and then decreased during the 12 d of storage. The treatment of A. pullulans S-2 advanced the decline time point from 8 d to 4 d, and the content of the treatment group in the late storage period was signi cantly lower than the control group.
During the 12 d storage, Actinobacteria (Fig. 7B) was rst declined and then increased. The increasing time point was advanced from 8 d to 4 d in the treatment group, and the proportion of Actinobacteria was greater than the control group during the later storage period. The Firmicutes (Fig. 7C) trend was always decreasing, and Firmicutes in the control group were always higher than the treatment group.
From the bacterial genus category (Fig. 6B), the six genera, namely Pantoea, Brevibacterium, Brachybacterium, Serratia, Glutamicibacter, and Pseudomonas were showed their importance. The panoramic view of the situation (Fig. 8) showed that the tendency of Pantoea and Pseudomonas was raised at the beginning and declined later. But, Brevibacterium, Brachybacterium, and Glutamicibacter were showed the opposite trend. The proportion of Serratia was stable during 0-8 d, and rapidly increased in 8-12 d.
After using A. pullulans S-2, the proliferation rate of Pantoea (Fig. 8A) was slowed down, and the decay rate became faster; the decay rate of Brevibacterium (Fig. 8C) and Brachybacterium (Fig. 8F) became faster in the early stage of storage, and the proliferation rate was slower in the later stage.
The proportion of Serratia (Fig. 8D) and Glutamicibacter (Fig. 8E) was increased in the late storage period, and they were the dominant bacterium in the treatment group. Pseudomonas (Fig. 8B) was the dominant bacteria in the early stage of the treatment group. Still, due to the rapid growth in the early stage, the total amount of Pseudomonas had been decreasing since the fourth day, but it was still signi cantly higher than the control group in the end.

Discussion
According to previous research, antagonistic yeast may replace chemical fungicides, prevent the postharvest diseases of fruits and vegetables, and signi cantly maintain the good qualities of fruits and vegetables. In this study (Fig. 1), the application of A. pullulans S-2 signi cantly reduced the incidence of tomato postharvest diseases, also maintained fruit moisture, rmness, TA, ascorbic acid and lycopene.
Our results were highly comparable with the previous study [28].
The in uence of A. pullulans S-2 on the change of microbial community structure on tomato surface is the key point of our study. The high-throughput sequencing technology was used to identify the microbiome of tomato surface during storage. From the analysis of ACE index (Fig. 2), we found that the application of A. pullulans S-2 had a more signi cant impact on fungal diversity. The rapid colonization of A. pullulans S-2 and inhibition of other pathogenic fungi in the early stage of storage resulted in the decrease of fungal diversity, so the diversity of the treatment group was less than the control group. At the later stage of storage, the tomato fruit began to rot, the abundance of pathogenic microorganisms and the microbial diversity of the treatment group was increased. The previous study has shown that rotting could signi cantly reduce fungal diversity [15], so the microbial diversity of the control group decreased due to the severe rot ( Fig. 2A).
Then, we analyzed the effect of A. pullulans S-2 on the microbial community structure of tomato surface from the perspective of colony composition. From the perspective of fungi (Fig. 4B), after A. pullulans S-2 treatment, the ve genera namely Aureobasidium, Cladosporium, Mycosphaerella, Alternaria, and Penicillium were changed signi cantly in the control and treatment groups. In the treatment group, only Aureobasidium was signi cantly higher than the control group, and the other four genera were lower compared to the control group (Fig. 5). Aureobasidium is a ubiquitous strain on the surface of tomatoes, and it also occupies a certain proportion in the naturally grown group (CK).
Cladosporium, Mycosphaerella, Alternaria, Penicillium are several common pathogens. Studies have shown that Cladosporium mainly exists in the leaves of various plants and can cause tomato leaf mold disease [29,30]. Mycosphaerella is the primary source of melon and banana leaf spot [31,32], but it doesn't seem to cause postharvest diseases. Alternaria and Penicillium can infect various fruits and vegetables, leading to black spot and blue mold of postharvest tomato [33][34][35].
It was observed that A. pullulans S-2, showed an outstanding ability to survive and colonize, and compete with pathogenic fungi such as Cladosporium, Mycosphaerella, Alternaria, Penicillium, for nutrition and space, thereby inhibiting the growth of pathogenic fungi and achieving the effect of disease control. Our results are consistent with the previous report that A. pullulans can compete for nutrition and space to prevent pathogenic fungi [36].
Stating from the bacterial genus (Fig. 8), the main genera found in our study was reported to be bene cial to plants on the surface. Among them, Pantoea is quite special. It is widespread in plants, and most of them are screened out from diseased plants. Therefore, it is generally believed that bacteria belonging to this genus were pathogenic in the early stage. Among them, the notorious pathogens, including Pantoea stewartia subsp. Stewartii and Pantoea ananatis, were reported to infect corn and rice [37]. However, with the deepening of research, it was found that some species in the genus have nitrogen xation and are considered to plant growth-promoting bacteria. They play a role in controlling Botrytis cinerea, which is very complex and fascinating [38][39][40].
Most species in the genus Brevibacterium have been proven to promote plant growth, among which Brevibacterium casei MH8a can enhance the ability of white mustard seeds to absorb metals, and Brevibacterium linens in Compost teas have a speci c control effect on tomato diseases [41,42]. In order to enhance the bene cial effects of Brevibacterium strains, it has been used in combination with other strains. For example, the synthetic community of Brevibacterium frigoritolerans HRS1 and the other three bacteria has a stronger immune effect against tomato bacterial wilt [43]. Brevibacterium halotolerans and Trichoderma harzianum have a synergistic interaction in improving the growth and yield of peppermint [44]. There are reports that the bacteria in the genus Brachybacterium also can promote plant growth [45].
Studies proved that Serratia, Glutamicibacter, and Pseudomonas had been recognized as plant growthpromoting bacteria. Serratia proteamaculans can effectively control tomato early blight and promote plant growth [46]. A salt-tolerant PGPR strain Glutamicibacter sp. YD01 can stimulate plant growth and development and alleviate the adverse effects on plants under salt stress conditions [47]. Pseudomonas was often isolated from the rhizosphere of plants, and some of them have been reported as rhizosphere bacteria that promote plant growth [48,49].
The effect of A. pullulans S-2 on bacteria is 'simple and clear' on the fungal community and mainly affects the growth and decay rate of many bacterial species. The proportion of Pantoea, Brevibacterium, and Brachybacterium in the treatment group was signi cantly lower than that of the control group, and the three genera Serratia, Glutamicibacter, and Pseudomonas were signi cantly higher compared to the control group. It can be understood as the 'replacement' of the dominant species.

Conclusions
A. pullulans S-2 can be used as a microbial antagonist to prevent and control postharvest tomato diseases. It affected the composition of the fungal community and the growth and decline rate of various bacteria to achieve the 'replacement' of the dominant ora. In addition, it can maintain the qualities of postharvest tomatoes, thereby reducing the incidence of diseases. This study found several potentially bene cial bacterial taxa from tomato surfaces during storage. Further studies about the role of these bacteria on host and biocontrol e cacy of the combination of potentially bene cial bacteria and A. pullulans S-2 will be needed. This study should provide a better understanding of the mechanisms which microbial antagonists control diseases.

Declarations
Ethics approval and consent to participate Not applicable Consent for publication Not applicable Availability of data and material All raw sequence data have been made available in the NCBI Sequence Read Archive (SRA) database under the BioProject PRJNA730651.

Competing interests
The authors declare that they have no competing interests.

Funding
This work was supported by the National Natural Science Foundation of China (31901743, 31772037, 31772369).   The PCoA (principal co-ordinate analysis) analysis for fungal (A) and bacterial (B) communities based on Weighted UniFrac at different storage times. The change curves of several important genera of fungal during storage, * = P < 0.05.