A novel biosynthetic gene cluster across the Pantoea species complex is important for pathogenicity in onion

Onion center rot is caused by at least four species of Pantoea (P. ananatis, P. agglomerans, P. allii, and P. stewartii subsp. indologenes). Critical onion pathogenicity determinants for P. ananatis were recently described but whether those determinants are common among other onion-pathogenic Pantoea species remains unknown. In this work, we report onion pathogenicity determinants in P. stewartii subsp. indologenes and P. allii. We identified two distinct secondary metabolite biosynthetic gene clusters in different strains of onion pathogenic P. stewartii subsp. indologenes. One cluster is similar to the previously described HiVir phosphonate biosynthetic cluster of P. ananatis and another is a novel putative phosphonate biosynthetic gene cluster, which we name “Halophos”. The Halophos gene cluster was also identified in P. allii strains. Both clusters are predicted to be phosphonate biosynthetic clusters based on the presence of a characteristic phosphenolpyruvate phosphomutase (pepM) gene. The deletion of pepM gene from either the P. stewartii subsp. indologenes HiVir or Halophos clusters caused loss of necrosis on onion leaves and red onion scales, and resulted in significantly lower bacterial populations compared to the corresponding wildtype and complemented strains. Seven (halB-halH) out of eleven genes (halA-halK) in the Halophos gene cluster are required for onion necrosis phenotypes. The onion non-pathogenic strain PNA15-2 gained the capacity to cause foliar necrosis on onion via exogenous expression of a minimal seven gene Halophos cluster (halB -halH). Furthermore, cell-free culture filtrates of PNA14-12 expressing the intact Halophos-gene cluster caused necrosis on onion leaves consistent with the presence of a secreted toxin. Together, these observations indicated that pepM genes in both phosphonate biosynthetic gene clusters (HiVir and Halophos) are important for Pantoea spp. onion pathogenicity and the biosynthetic product of the Halophos cluster causes necrosis on onion leaf tissue. Overall, this is the first report of onion pathogenicity determinants in P. stewartii subsp. indologenes and P. allii. Author summary Onion center rot is caused by multiple Pantoea species including P. stewartii subsp. indologenes and P. allii. We identified two distinct secondary metabolite biosynthetic clusters associated with onion pathogenic strains, the validated HiVir phosphonate cluster and a putative phosphonate biosynthetic cluster that we named as Halophos based on the associated “halo” phenotype on the red onion scales. We found that pepM genes from each cluster (HiVir and Halophos) are required for onion infection by P. stewartii subsp. indologenes and P. allii but not for millet infection by P. stewartii subsp. indologenes. Conversely, the T3SS was important for millet infection by P. stewartii subsp. indologenes but not onion infection. Induction of the intact Halophos cluster was associated with the accumulation of a necrosis-inducing factor in culture, which suggests it might be a secreted phytotoxin. Seven of the eleven Halophos cluster genes are required for onion necrosis phenotypes and expression of this minimal cluster conferred a limited onion necrosis phenotype to an onion-non-pathogenic Pantoea strain. We provide evidence of a Halophos biosynthetic gene cluster to be associated with onion pathogenicity in strains of P. stewartii subsp. indologenes and P. allii.

suggesting that the putative phosphonate compounds may be essential for pathogenicity 124 in onion. The Halophos cluster is adjacent to a phage/plasmid primase gene in P. stewartii 125 subsp. indologenes PNA14-9, PNA14-11, and PNA14-12. The Halophos cluster from 126 PNA14-12 is located within a genomic island predicted by IslandViewer 4. In addition, the 127 Halophos cluster has a lower GC%, higher average effective number of codons (Nc), and 128 lower average codon adaptation index (CAI), compared to their corresponding whole 129 genomes ( The Halophos gene cluster is predicted to contain eleven co-transcribed genes 135 ( binding enzyme (halJ), and pyridoxamine 5'-phosphate oxidase (halK) (Fig 1 and Table   141 2). Notably, the halJ sequence is predicted to possess an adenylation domain and a 142 condensation domain, suggesting that it might be involved in non-ribosomal peptide 143 synthetase activity. to the same length and used for constructing a neighbor-joining tree. The bootstrap 153 support values > 70% were shown at the node. The conserved clusters (not to scale) 154 were colored with the same colors representing homologous proteins. Non-conserved 155 proteins were in white, and fragmented pseudogenes were in stripes. Transposase 156 genes were labeled as Tn.

181
Comparison of the two phosphonate biosynthetic gene clusters, HiVir and 182 Halophos revealed that only two genes were similar based on predicted annotations: 183 pepM and the major facilitator superfamily (MFS) transporter gene (Tables 2 and 3).

471
Similar to PNA14-12, P. allii LMG24248 also showed a pink halo phenotype on the 472 red onion scale, while its pepM mutant did not develop any symptoms (Fig 4). indicates the modes of action between HiVir and Halophos products are different.

490
Halophos was predicted in silico to encode a putative phosphonate biosynthetic 491 gene cluster. In this study, we characterized individual genes in the Halophos gene cluster, 492 and found that halB-H but not halA, halI, halJ, halK genes contribute to onion 493 pathogenicity (Fig 5). We showed that by transferring a minimal cluster of halB to halH to 494 an onion-non-pathogenic strain PNA15-2, the strain was able to cause foliar necrosis on 495 onion (Figs 5D and 5E). However, the mean lesion length of PNA15-2 pBS46::halB-H 496 was relatively small. This may be due to the low expression level of the plasmid carrying 497 a large fragment of halB to halH (8,614 bp) or other components present in PNA14-12 498 but absent in PNA15-2 that might be needed to facilitate the expression of halB to halH.

499
In addition, to our surprise, halI, the MFS transporter gene from the Halophos cluster did 500 not contribute to onion pathogenicity (Fig 5), while hvrI, the MFS transporter gene from 501 the HiVir cluster in P. ananatis was shown to be important for onion leaf virulence [8]. We speculate that PNA14-12 does not rely on halI but other unknown means to transport its 503 product of Halophos out of the cell.

525
Bacterial strains and inoculum preparation. 526 Bacterial strains and plasmids used in the study are listed in S2 Table. Naturally    and foliage inoculation assays, 10 μl of 1 x 10 6 CFU/ml bacterial suspensions were used.

597
The pictures and lesion lengths were recorded at 5 dpi. were streaked on LB agar with trimethoprim and stored. The chromosomal insertion of 615 the plasmid was confirmed by PCR (S3 Table).