香京julia种子在线播放

Organically produced apple fruits (Malus domestica) of the cultivar “Topaz” were obtained from the organic storage company Rosenbaum Franz GmbH & Co KG (Pöllau, Austria). Apple samples were taken directly after harvest and after a 6-months storage period. Freshly harvested apples were immediately taken to the laboratory and processed under sterile conditions (in the following named “before storage”). For analyzing impact of HWT on the apple microbiota, 100 apples were stored untreated and 100 apples were subjected to HWT by immersing apples in a 53°C water bath for 3 min. The two groups were stored separately but under the same controlled conditions in the company’s storage chamber for 6 months. Directly after opening storage chambers, fungal infection rate by any fungal pathogen on apples was evaluated. HWT was found to be highly efficient as no disease patterns were observed. Among the 100 apples that were untreated 10% were infected, exhibiting rot diameters of 2.5–4 cm in diameter. A subset of each group, consisting of 10 randomly selected apples, was subjected to amplicon analyses; untreated apples were defined into “untreated healthy” and “untreated diseased.” The apples were transported to the laboratory and processed under sterile conditions. One whole apple for each category and sample (“before storage,” “HWT,” “untreated healthy,” and “untreated diseased”) was cut into smaller pieces and homogenized in a Stomacher laboratory blender (BagMixer, Interscience, St. Nom, France) with 40 ml sterile NaCl (0.85%) solution for 3 min. A total of 4 ml of the solution was centrifuged at 16,000 g for 20 min and the pellet stored at −70°C for further DNA extraction. This way 10 biological replicates for each category were produced.

The resulting pellets from the previous step were subjected to total microbial DNA extraction using the FastDNA SPIN Kit for Soil (MP Biomedicals, Solon, USA) and a FastPrep Instrument (MP Biomedicals, Illkirch, France) for 30 s at 5.0 m/s. Amplicons were prepared in three technical replicates using the primer pair 515f-926r, specific for bacteria and ITS1f-ITS2r specific for fungi. Sequences of primers are listed in Supplementary Table S1. Peptide nucleic acid (PNA) clamps were added to the PCR mix to block amplification of host plastid and mitochondrial 16S DNA (Lundberg et al., 2013). Amplification of the 16S rRNA gene was performed in a total volume of 20 μl [5 × Taq&Go (MP Biomedicals, Illkirch, France), 1.5 μM PNA mix, 10 μM of each primer, PCR-grade water and 1 μl template DNA] under the following cycling conditions: 95°C for 5 min, 35 cycles of 78°C for 5 s, 55°C for 45 s, 72°C for 90 s, and a final elongation at 72°C for 5 min. PCR for amplifying the fungal ITS region was conducted in 20 μl (5 × Taq&Go, 10 μM of each primer, 25 μM MgCl₂, PCR-grade water and 2 μl template DNA) using the cycling conditions: 94°C for 5 min, 30 cycles of 94°C for 30 s, 58°C for 35 s, 72°C for 40 s and a final elongation at 72°C for 10 min. A nested PCR step was performed to add barcoded primers (10 μM) in a total volume of 30 μl for both 16S rRNA gene and ITS region: 95°C for 5 min, 15 cycles of 95°C for 30 s, 53°C for 30 s, 72°C for 30 s, and a final elongation at 72°C for 5 min. Three technical replicates, conducted for each sample, were combined and purified by Wizard SV Gel and PCR Clean-Up System (Promega, Madison, WI, USA). DNA concentrations were measured with Nanodrop 2000 (Thermo Scientific, Wilmington, DE, USA) and samples were combined in equimolar concentration. The amplicons were sequenced on a Illumina MiSeq v2 (2 × 250 bp) machine.

After joining forward and reversed paired end reads in QIIME 1.9.1, sequencing data was imported into QIIME 22019.1 and demultiplexed following the QIIME 2 tutorials. The DADA2 algorithm was applied for quality filtering, discarding chimeric sequences and to obtain a feature table [containing sequence variants (SVs)] and representative sequences. Feature classification was performed using a Naïve-Bayes feature classifier trained on the Silva132 release (16S rRNA gene) (Quast et al., 2013) or the UNITE v7.2 release (ITS) (Kõljalg et al., 2013). Sequences of features of interest were further identified on species level using NCBI blast alignment tool. Mitochondria and chloroplast reads were discarded from 16S data. Alpha and Beta diversity was investigated running the core diversity script in QIIME 2 rarefying feature tables to the lowest value of reads present in one sample. Core microbiomes (features present in 50% of the samples) were defined for each sample group and core tables were rejoined to obtain barplots and evaluate taxonomic differences. A taxonomy network was constructed on core genera using Cytoscape version 3.5. (Shannon et al., 2003).

Statistical analysis of metabarcoding data was performed using scripts in QIIME 1.9 as well as QIIME2 2019.1. Alpha diversity was tested using the Kruskal-Wallis test and Beta diversity using Analysis of Similarity (ANOSIM) test. Significant differences (alpha ≤ 0.05) in taxa abundance on genus level were calculated using non-parametric Kruskal-Wallis test and False Discovery Rate (FDR) multiple test correction.

A quantitative real-time PCR (qPCR) was conducted to quantify overall bacterial 16S rRNA and fungal ITS gene copy numbers, as well as postharvest pathogens P. expansum and Neofabraea spp. For specific quantification of bull’s eye rot-causing Neofabraea strains, a primer pair (NeoF, NeoR) was selected that specifically targets the highly conserved β-tubulin gene which was found to amplify the three major pathogens associated with bull’s eye rot (N. alba, N. malicorticis, N. perennans), but no other related fungi (Cao et al., 2013). The primer pair (Pexp_patF_F, Pexp_patF_R) for P. expansum targeted the patF gene (involved in the patulin biosynthesis) and was previously tested for its specificity (Tannous et al., 2015). Primer pairs were used each in 5 pmol/μl concentration and are listed in Supplementary Table S1. All reaction mixes contained 5 μl KAPA CYBR Green, 0.5 μl of each primer, 1 μl template DNA (diluted 1:10 in PCR-grade water), adjusted with PCR-grade water to a final volume of 10 μl. Reaction mix for bacterial amplification was supplemented with 0.15 μl PNA mix to block amplification of host-derived 16S rRNA gene copies. Fluorescence intensities were detected using a Rotor-Gene 6000 real-time rotary analyzer (Corbett Research, Sydney, Australia) with the following cycling conditions: Bacteria: 95°C for 5 min, 45 cycles of 95°C for 20 s, 54°C for 30 s, 72°C for 30 s and a final melt curve of 72 to 95°C. Fungi: 95°C for 5 min, 45 cycles of 95°C for 30 s, 58°C for 35 s, 72°C for 40 s and a final melt curve of 72–95°C. P. expansum: 95°C for 5 min, 45 cycles of 95°C for 20 s, 65°C for 15 s, 72°C for 15 s and a final melt curve of 72–95°C. Neofabraea sp.: 95°C for 5 min, 45 cycles of 95°C for 20 s, 57°C for 15 s, 72°C for 40 s followed by melt curve of 72–95°C. Amplification efficiency of the target was analyzed with the melting curve (Supplementary Figure S1). Three individual qPCR runs were conducted for each replicate. Intermittently occurring gene copy numbers that were found in negative controls were subtracted from the respective sample. Significant differences (p ≤ 0.05) of bacterial and fungal gene copy numbers per apple between the different apple groups were calculated using a pairwise Wilcoxon test (Bonferroni correction) and visualized using ggplot2 in R version 3.5.1.

Small scale experiments were conducted to test the efficacy of potential biocontrol agents with and without combined HWT against infection of the fungal pathogens P. exopansum ATCC 7861 (Origin: CBS 325.48) and N. malicorticis (Jacks) Nannfeld (Origin: DSMZ 62715), selected as representative for bull’s eye rot-causing fungal pathogens. More than 800 bacterial strains, isolates from apples, were tested for antagonistic properties toward the two pathogens by dual-culture in vitro assay on Waksman agar (Berg et al., 2002). Bacterial isolates showing highest antagonistic properties toward both fungi were identified by 16S rRNA gene Sanger sequencing (LGC Genomics, Berlin, Germany) and using the NCBI BLAST alignment tool: Pantoea vagans 14E4, Bacillus amyloliquefaciens 14C9, and Pseudomonas paralactis 6F3. Preliminary tests showed Pantoea vagans 14E4 as well as a consortium of the three different bacterial strains on equal proportions to have the best control effect. For in vivo tests, 30 apples from the cultivar “Topaz” per treatment and pathogen were rinsed with water and four artificial wounds (1 cm in diameter and depth) were cut with a sterile knife around the radius of the fruits. Each apple was artificially infected with N. malicorticis (submerged in a 1.6 × 10⁵ conidia/ml solution) or P. expansum (10 μl of a 5 × 10⁴ spores/ml solution) and incubated for 24 h at 20°C. For both, the wound pathogen P. expansum, and the lenticel rot causing N. malicorticis this way an infection could be induced at the wanted locations. Overnight cultures in nutrient broth (Sifin, Berlin, Germany) of bacterial biocontrol strains were cultivated at 30°C and centrifuged at 5,000 rpm for 15 min. The supernatant was discarded and bacterial pellets were resuspended in sterile sodium chloride solution (0.85%). A consortium on equal proportions of all three biocontrol strains was prepared. Suspensions were diluted to an OD₆₀₀ of 0.2 (approximately, 10⁶ cells/ml). Apples, 24 h after inoculation with the fungal pathogens, were treated either with P. vagans 14E4 or the consortium by submerging the apples in the prepared solution for 5 s. HWT groups were previously submerged in 53°C hot water for 3 min and allowed to dry. Negative control samples were stored directly after wounding without pathogen infection and positive control samples were stored after infection with N. malicorticis and P. expansum without further treatment. Results were evaluated after 3 weeks (P. expansum) and 5 weeks (N. malicorticis) storage period under controlled conditions at 4°C. Supplementary Figure S2 exemplifies the temporally disease progression of P. expansum infection, directly, 1 and 3 weeks after wounding. The diameter of infected areas as well as the length of the cuts was measured and statistical significance tested using a pairwise Wilcoxon test (Bonferroni correction) and visualized using ggplot2 in R version 3.5.1.

After quality filtering and removing of chimeric sequences using the DADA2 algorithm and excluding mitochondrial and chloroplast sequences from the 16S rRNA gene fragments, the 16S rRNA and ITS datasets contained 1,071,751 and 880,909 paired reads, respectively. Sequences were assigned to 2,297 bacterial and 613 fungal features and the datasets were rarefied to 1,638 bacterial and 1,319 fungal sequences, according to the sample with the lowest amount of sequences. Core microbiota were defined for each sample group (“before storage,” “HWT,” “untreated healthy,” and “untreated diseased”), by keeping only the features present in 50% of the replicates of the respective group. A 50% cutoff was used to keep enough features for the evaluation of the main features in all groups and to discard rarely occurring features. In total, 205 core bacterial and 89 core fungal features remained that were condensed to 60 and 44 genera, respectively. From those taxa, a network of co-occurring OTUs was constructed to visualize shared taxa and taxa being unique for a specific group (Figure 1). Among 104 bacterial and fungal genera, 23 were shared by all apples, while 22 genera were present in “HWT” and “untreated healthy” apples but absent in all other samples, probably indicating a health-related postharvest microbiome. Additionally, “HWT,” “untreated healthy,” and “before storage” samples hosted 13, 16, and 10 unique taxa, respectively, while no unique taxa were found for “untreated diseased” apples. N. alba was present in all apples, including “before storage” samples, whereas P. expansum only occurred in stored apples.

In order to compare taxonomic composition of the four groups, Figure 2 was constructed for the bacterial (Figure 2A) and fungal (Figure 2B) core microbiota of each group on genus level, where genera represented with less than 1% of the number of reads are clustered as “Other.” The microbiota within the four different groups showed great taxonomic variability, especially when apples before storage were compared to stored apples. The bacterial microbiota within all samples was highly dominated by Proteobacteria, ranging from 65% in “before storage” samples up to 80% in “untreated healthy” apples. Apples “before storage” had additionally a high abundance of Bacteroidetes (32%) compared to the other groups (3–8%), whereas all stored apple samples prevailed in Actinobacteria abundance (9–20%) over “before storage” samples (1%). Sphingomonas was the most abundant genus in all groups (35–46%). Hymenobacter (31%) and Massilia (13%) were furthermore highly abundant in apples before storage. Pseudomonas (7–11%) and Methylobacterium (7%) were abundant in healthy apples after storage, whereas diseased apples after storage showed high abundances of Methylobacterium (12%) and Frondihabitans (11%) (Figure 2A). In total, the core microbiota of the four groups “before storage”, “HWT”, “untreated healthy” and “untreated diseased” contained 15, 50, 49, and 18 bacterial genera, respectively.

The fungal microbiota was dominated by Ascomycota, ranging from 72% in “untreated healthy” samples up to 97% in “untreated diseased” apples. Basidiomycota were more abundant in healthy apples before (19%) and after (11–26%) storage, compared to “untreated diseased” apples (3.5%). On genus level, Mycosphaerella dominated “before storage” samples (30%), followed by Alternaria (19%), Vishniacozyma (12%), Cladosporium (8%), and Aureobasidium (7%). Stored “HWT” samples were dominated by a not further assigned taxon of Hypocreales (20%), followed by Cladosporium (15%), P. expansum (11%), Acremonium, and Didymellaceae sp. (each 10%) and Vishniacozyma (9%). Almost the same fungal genera were highly abundant in stored “untreated healthy” samples, with Vishniacozyma (21%) being the main representative, except P. expansum featuring only 1% abundance. Stored “untreated diseased” apples were almost exclusively composed of the two postharvest pathogens P. expansum (45%) and N. alba (42%) (Figure 2B). Both fungi were present in “HWT” and “untreated healthy” apples, although with less relative abundance. “Before storage” apples contained 0.1% N. alba, while P. expansum was absent. The samples “before storage”, “HWT”, “untreated healthy” and “untreated diseased” contained 28, 27, 33, and 18 fungal core genera, respectively.

The bacterial and fungal diversity within the apple samples was assessed by Shannon diversity index. Apples from the category “before storage” showed significantly the lowest bacterial diversity (H′ = 5.19 ± 0.8), followed by stored apples from the category “untreated diseased” (H′ = 5.72 ± 0.3). Both were significantly less diverse than stored “untreated healthy” (H′ = 6.46 ± 0.6) and “HWT” samples featuring highest bacterial diversity (H′ = 6.68 ± 0.4) (Figure 3A). Fungal diversity was highly decreased in stored “untreated diseased” apples (H′ = 1.93 ± 0.8), being significantly lower compared to all healthy apples: “before storage”: H′ = 3.77 ± 0.5, “HWT”: H′ = 3.87 ± 0.6 and “untreated healthy”: H′ = 4.31 ± 0.1 (Figure 3B).

Beta diversity analyses, applied on the whole bacterial and fungal dataset and based on Bray Curtis distance matrix, indicated clear clustering between apples before and after storage in all cases (Figures 3C,D). Statistical significance in bacterial composition, assessed via pairwise ANOSIM (Table 1), revealed significant differences between all groups, except for the comparison of “HWT” and “untreated healthy” samples. Highest variability was found when “before storage” samples were compared to the remaining groups. The fungal composition was significantly different between all four groups, while difference between “HWT” and “untreated healthy” samples was lowest.

In order to identify bacterial and fungal taxa that potentially contribute to pathogen resistance in “untreated healthy” apples, significant differences in taxa abundance between “untreated healthy” and “untreated diseased” samples were calculated (Supplementary Table S2). A total of 42 bacterial and 28 fungal taxa were found significantly higher in “untreated healthy” apples as well as 2 fungal taxa (P. expansum and N. alba) being significantly increased in “untreated diseased” apples. Higher numbers of taxa in “untreated healthy” apples were found for e.g., Sphingomonas, Pseudomonas, and Methylobacterium as well as Vishniacozyma, Cladosporium, and Acremonium.

Additionally, the impact of HWT on the apple postharvest microbiota was evaluated as well, by calculating significant differences in taxa abundance between “HWT” and “untreated healthy” apples (Supplementary Table S3). A total of 25 bacterial and 22 fungal genera were found to be significantly different abundant between the two groups. Significantly increased in “HWT” were, e.g., Hymenobacter, Rathayibacter as well as Filobasidium; increased in “untreated healthy” were, e.g., Curtobacterium, Rhodococcus as well as Penicillium and Alternaria. However, as previous stated, the overall bacterial microbiome and diversity was not significantly different between the two groups only the fungal microbial composition was slightly changed.

A real time PCR was performed to quantify total bacterial 16S rRNA and fungal ITS gene copy numbers. Bull’s eye rot-causing Neofabraea strains and P. expansum were specifically quantified as well (Figure 4). No significant differences in 16S rRNA gene copy abundance was observed between the four different apple groups; neither between apple “before storage” and all stored apples, nor within the stored groups (Figure 4A). Pathogen infection as well as HWT did accordingly not affect the bacterial abundance in apples. Regarding the total fungal ITS genes we found significantly higher abundances within “untreated diseased” apples compared to all other groups (Figure 4B), due to significant increase of both storage pathogens Neofabraea and P. expansum (Figures 4C,D, respectively). Neofabraea was already present in “before storage” apples in similar abundances as in “HWT” and “untreated healthy” apples while P. expansum was almost absent in apples “before storage.” Overall, fungi were found to proliferate more efficiently compared to bacteria in stored apples, as showed via calculating the prokaryote to eukaryote ratio (Figure 4E). Whereas the ratio was almost balanced in apples before storage (58% bacteria and 42% fungi), fungal genes increased up to the two-fold in stored, healthy apples. A dramatic increase of fungal genes was however observed within stored, diseased apples; 99.4% of total microbial genes detected were fungal.

The efficacy of potential biocontrol strains (P. vagans 14E4, B. amyloliquefaciens 14C9 and P. paralactis 6F3) identified using antagonistic screening methods was tested in small-scale storage experiments with or without combined HWT against N. malicorticis and P. expansum. P. vagans E14 was applied as single agent as well as combined with the other potential biocontrol strains in form of a consortium. Negative control apples that were wounded artificially but not infected with fungal pathogens appeared to be unaffected after two as well as after 5 weeks of storage. Positive control apples that were inoculated with the fungal pathogens and untreated showed 100% infection rate for N. malicorticis and 96% for P. expansum (Figure 5A). Treatment using biocontrol strains slightly decreased infection rates, however, still up to 88% of apples were infected. HWT reduced infection rates of N. malicorticis and P. expansum to 58 and 75%, respectively. Overall, combining HWT and the biocontrol consortium reduced the total infection rates the most (up to 42%). Similar results were shown when the infection diameter was measured (Figure 5B). Here, no significant differences in infection diameter were found between positive control samples and apples treated with biocontrol strains that were not subjected to HWT. In contrast, HWT approved to be efficient in reducing pathogen infection rates, while the combined treatment of HWT and potential biocontrol strains resulted in even less infection.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.