Association of Rahnella victoriana, Enterobacter hormaechei subsp. hoffmannii and Citrobacter braakii with walnut decline

Persian walnut (Juglans regia) has a considerable economic importance worldwide. However, the vigor and vitality of walnut trees were heavily affected by bark canker during the last few years. Irregular longitudinal cankers in the outer bark, stem tissue necrosis, and bleeding with black-colored exudates walnut trees were observed in Kermanshah, Hamedan, Markazi, Alborz, Isfahan, Qom, Semnan, and Razavi Khorasan provinces in western, central and eastern Iran during 2018 and 2019. A total of 150 symptomatic samples were collected from affected walnut trees in order to identify bacteria associated with walnut decline. Two-hundred sixty strains with a metallic green sheen were isolated on EMB-agar medium. The pathogenicity of all strains was proved by inoculating a suspension of the bacterial strains under the bark of immature walnut fruits cv. ‘Hartley’. Ninety-five strains caused necrosis and a dark-colored region in the mesocarp around the inoculation site 14 days post-inoculation. Moreover, 12 representative strains induced necrotic and black-colored tissues in the bark of young green twigs of two-year old walnut seedling cv. ‘Chandler’. The strains were classified into four categories based on conventional phenotypic characters confirmed with the 16S rRNA gene sequences. A phylogenetic tree based on the concatenated sequences of two housekeeping gene fragments, gyrB and infB, indicated that strains including I1, Q6, and S6 were grouped in a cluster with Gibbsiella quercinecans FBR97T as well as strains I2, I5, and KE6 were clustered with Rahnella victoriana FRB 225T. Moreover, strains MR1, MR3, and MR5 were grouped with the Enterobacter hormaechei subsp. hoffmannii DSM 14563T. The phylogenetic analyses based on the partial sequencing of housekeeping genes including fusA, pyrG, and leuS revealed that strains KH1, KH3, and KH7 belong to Citrobacter braakii species. To the best of our knowledge, this is the first report of C. braakii and E. hormaechei as plant pathogens and R. victoriana associated with walnut decline.


Results
Symptoms and bacterial isolation. A total of 150 symptomatic samples were collected from affected walnut trees of Kermanshah, Hamedan, Markazi, Alborz, Isfahan, Qom, Semnan, and Razavi Khorasan, provinces in western, central and eastern Iran (Fig. 1). The frequent symptoms were irregular longitudinal cankers in the outer bark, stem tissue necrosis, bleeding with black-colored exudates, and dark lesions in the inner bark (Fig. 2). Walnut trees seriously affected by cankers showed a progressive loss in vigor and death after few years. The severity of the disease symptoms was higher in some areas, especially in Kermanshah, Hamedan, and www.nature.com/scientificreports/ Qom provinces. Two-hundred sixty strains with metallic green sheen were isolated on eosin methylene blue agar (EMB-agar) medium. The colony of bacterial strains on nutrient agar (NA) was cream, round, convex and smooth with entire margins.
Pathogenicity assay. The pathogenicity of 95 strains was proved on immature walnut fruits cv. 'Hartley' .
Necrosis and a dark-colored region in the mesocarp around the inoculation site were observed 14 days postinoculation ( Table 1). The pathogenicity of 12 representative strains was confirmed on young green twigs of two-year-old walnut seedling cv. 'Chandler' . Necrotic and black-colored tissues appeared in the bark around the inoculation points, and such symptoms were detected in positive control, B. nigrifluens M2, 30 days post-inoculation (Fig. 3). G. quercinecans and Enterobacter hormaechei subsp. hoffmannii strains induced large necrotic areas beyond inoculated sites on immature fruits and branches (Fig. 3). No symptoms were detected in the fruits and young branches infiltrated with sterile distilled water (SDW). Bacteria resembled those used in inoculations were recovered in high populations from the inoculated immature fruits and twigs 14 and 30 days after inoculation, respectively. Re-isolated strains were identical with inoculated ones in all phenotypic characters.
Phenotypic characteristics. All strains were negative in Gram reaction, oxidase test and levan sucrase activity. All strains were fermentative aerobes and catalase positive. None of strains produced pink or fluorescent pigments on YDC and King's B media, respectively. The strains were classified into four categories based on conventional phenotypic characters. Sixty-two strains belonging to G. quercinecans were negative for gelatin and starch hydrolysis, whereas positive for indole production and esculin hydrolysis. Fourteen strains characterized as Rahnella victoriana were able to hydrolyze Tween 20, Tween 80, gelatin and esculin. R. victoriana strains did not produce indole or urease. The third phenotypic group consists of 11 strains was identified as Enterobacter hormaechei subsp. hoffmannii, which hydrolyzed gelatin, however, did not utilize starch, esculin, Tween 20 and Tween 80. These strains produced indole and urease ( Table 2). The last phenotypic group, including eight strains was recognized as Citrobacter braakii, which utilize starch. Strains of this group did not utilized esculin, Tween 20, Tween 80 and gelatin (Table 2).

Phylogenetic analysis.
Amplification and partial sequencing of 16S rRNA gene of one representative strain from each phenotypic group were performed. Moreover, the gyrB and infB of nine representative strains of three phenotypic groups and fusA, pyrG, and leuS of those of fourth group were amplified and partially sequenced. The accession numbers of sequences obtained in this study were deposited in the NCBI GenBank database are listed in Table 3. Pairwise comparison the 16S rRNA gene nucleotide sequences of strains I1 and I2 were 100% identical with those of G. quercinecans FBR97 T and R. victoriana FRB 225 T , respectively. Moreover, a phylogenetic tree reconstructed based on the concatenated sequences of two housekeeping gene fragments, gyrB (601 bp) and infB (615 bp), revealed that the strains I1, Q6 and S6 were grouped in a cluster with G. quercinecans FBR97 T . Additionally, the strains I2, I5, and KE6 were clustered with R. victoriana FRB 225 T , not only in the phylogenetic tree based on the concatenated sequences (Fig. 4), but also in each of the single gene-based trees (Data not shown).
The partial 16S rRNA gene sequence of strain MR1 was identical to that of E. hormaechei subsp. hoffmannii DSM 14563 T . The maximum-likelihood phylogenic tree, including strains identified as Enterobacter were  www.nature.com/scientificreports/ collected from Markazi province and type strains of Enterobacter species was created using concatenated sequences of gyrB (684 bp) and infB (906 bp) genes. The strains MR1, MR3, and MR5 were grouped with E. hormaechei subsp. hoffmannii DSM 14563 T . The resulting cluster presented a robust bootstrap value (Fig. 5).
The partial 16S rRNA gene sequence of strain KH7 collected from Razavi Khorasan province indicated that this pathogenic strain on walnut belongs to Citrobacter genus with 99.63 and 99.44% similarity with Citrbacter freundii ATCC 8090 T and Citrobacter braakii ATCC5113 T , respectively. The phylogenetic analyses based on the partial sequencing of three housekeeping genes, fusA (633 bp), pyrG (305), and leuS (640 bp) demonstrated that strains KH1, KH3, and KH7 belong to C. braakii species in a monophyletic clade with high bootstrap support (Fig. 6).
As Fig. 1 shows, G. quercinecans was identified associated with walnut decline in all sampled provinces except Markazi and Razavi Khorasan whereas the agents of disease in theses provinces were E. hormaechei subsp. hoffmannii and C. braakii, respectively. Furthermore, R. victoriana strains were in association with walnut decline symptoms from Isfahan and Kermanshah provinces. www.nature.com/scientificreports/

Discussion
Iran is the third producer of Persian walnut worldwide. However, its walnut trees have been under threat from decline during last decade. The causal agents of disease were identified in the southern, northern, and northwestern provinces 2,12-14 . Nevertheless, despite high severity and prevalence of walnut decline, the etiology of the disease has not yet been studied in the most western, central and eastern provinces. In this study, the bacterial strains in association with the walnut decline samples originating from eight provinces of Iran were identified using classical and molecular methods. The pathogenic bacterial strains were categorized into four species including G. quercinecans, R. victoriana, C. braakii, and E. hormaechei subsp. hoffmannii. G. quercinecans was recognized as the causal agent of acute oak decline (AOD) in the United Kingdom 22 , Spain 23,24 , Switzerland 25 , Iran 26 and Poland 27 . In addition, it was known as the causal agents of the necrotic tissues of apple and pear trees in Germany and Austria, as well as exudates from a pear tree in China 28 . Furthermore, G. quercinecans was reported as the causal agents of the bacterial canker disease of Russian olive 29 , and the walnut canker disease in Iran 3 . R. victoriana was isolated from oak trees suffering AOD in the UK 18 and Iran 30 . Moreover, www.nature.com/scientificreports/  Table 3. Phenotypic characters of walnut pathogenic strains isolated from decline symptoms from different provinces of Iran. Gibbsiella quercinecans FBR97 T and Rahnella victoriana FBR225 T strains used as control.

Citrobacter braakii (8 strains)
Host         www.nature.com/scientificreports/ R. victoriana strain JZ-GX1 was identified as a growth-promoting rhizobacteria on poplar roots 31 . To the best of our knowledge, this is the first report of R. victoriana in association with walnut decline. Pathogenicity of G. quercinecans strains is attributed to production of plant cell wall-degrading enzymes (PCWDEs) such as rhamnogalacturonase, cellulase, β-glucosidase, α-galactosidase, galactanase, and polyglactoronase whereas pathogenicity of R. victoriana species is ascribed to α-and β-galactosidase, cellulose and a type III secretion system 32 . R. victoriana and G. quercinecans were isolated from symptomatic walnut trees in Isfahan and Kermanshah provinces. Nevertheless, the role of individual lesion microbiota as components of a complex pathobiome, is unknown.
Human bacterial pathogens which cause diseases in plants and vice versa are dramatically increased during the last decades. For instance, Burkholderia gladioli and Burkholderia glumae, routinely isolated from panicle blight of rice 33 , were identified in pneumonia infections in patients with the chronic granulomatous disease 34,35 . Agrobacterium pusense inciting sepsis in human beings 36 , induced tumors on Lawson cypress crown 37 . Moreover, Erwinia billingiae, causing mango bacterial canker disease 38 , was known as the agent of cutaneous infection and bacteremia in humans 39 . Stenotrophomonas maltophilia, an emerging human pathogen 40 , is also responsible for necrotic lesions and stem bleeding on oak trees in Iran 41 . In addition, C. freundii, a frequent cause of nosocomial infections, was recognized as the agent of canker on white mulberry trunks and scaffold branches of trees 42,43 . Species C. braakii considered an as opportunistic human pathogen, was in association with various infections 44 . Furthermore, E. hormaechei is a causative pathogen for human nosocomial infections 45 . These data reveal that C. braakii and E. hormaechei subsp. hoffmannii are new plant pathogens which cause walnut decline.
Overall, the present study highlights the fact that walnut decline has to be considered a complex disease in which several bacteria are involved. Identification and characterization of all bacteria associated with walnut decline will eventually improve our understanding of the etiology of the disease and may result in improved management techniques for control. Future research is necessary to find beneficial bacteria serves as biological control for a variety of bacterial agent of walnut decline or finding resistant genotypes.

Material and methods
Sampling and bacterial isolation. Symptomatic samples were collected from walnut orchards of different provinces of Iran including, Kermanshah, Hamedan, Markazi, Alborz, Isfahan, Qom, Semnan, and Razavi Khorasan during 2018-2019. The samples were washed under tap water for 5 min and were disinfected using ethanol 70% for 30 s and were rinsed three times with SDW. Then the symptomatic tissues were grounded in SDW and incubated at room temperature for one hour. Additionally, more pieces were incubated in 20 ml SDW in incubator shaker for 24 h, and a loopful of resulting suspensions were streaked on plates of EMB-agar (Merck, Darmstadt, Germany) medium containing 1% glycerol and 0.5% yeast extract and were incubated at 28 °C for 72 h. Single colonies were purified on NA plates and routinely maintained in 15% glycerol stocks at − 80 °C.
Pathogenicity assays. Pathogenicity of all isolated strains was carried out on walnut immature fruits cv. 'Hartley' . The fruits were disinfected using ethanol 70% for three min 46 . Twenty µl of bacterial suspension (10 6 CFU/ml) were infiltrated in each inoculation site. Inoculated fruits were incubated in Ziploc bags at 26 °C with 90% relative humidity 47 . Each strain was inoculated in three replicates. Symptom progression was observed for 14 days. Moreover pathogenicity of 12 representative strains (namely I1, Q6, S6, I2, I5, KE6, MR1, MR3, MR5, KH1, KH3, and KH7) was performed by inoculating 60 μl of bacterial suspension (10 8 CFU/ml) on green twigs of two-years old walnut clonal seedlings cv. Chandler. The inoculated stems were covered using parafilm and maintained in a greenhouse under a cycle of 16 h light at 28 °C and 8 h dark at 24 °C with 75% relative humidity. Symptoms were monitored one-month post-inoculation. B. nigrifluens M2 13 and SDW were used as positive and negative controls, respectively. Pathogenicity assays were repeated at least twice. To fulfill Koch's postulates, re-isolation from 12 inoculated fruits and branches was performed 14-and 30-days post-inoculation on EMB-agar, respectively. The re-isolated strains were identified by phenotypic characters.
Phenotypic characteristic. All pathogenic strains as well as 12 re-isolated strains from walnut inoculated immature fruits and twigs namely I1, Q6, S6, I2, I5, KE6, MR1, MR3, MR5, KH1, KH3, and KH7 were subjected to physiological, morphological and biochemical tests including, Gram 48 , oxidase reactions 49 , fermentative metabolism 50 , and levan formation from sucrose. Moreover, the production of pink and fluorescent pigments on YDC and King's B was performed 51 . Hydrolysis of starch, esculin, Tween 20, and Tween 80 was performed according to Schaad et al. 51 . Utilization of carbon sources was assayed using the basal medium of Ayers et al. 52 complemented with filter-sterilized carbohydrate sources at 0.25% final concentration. G. quercinecans LMG25500 T and R. victoriana FRB 225 T were also used as reference strains in all tests. Phylogenetic analysis. Genomic DNA was extracted from 24 h culture using the alkaline lysis 53 . The PCR was performed in a volume of 30 µl containing 2 × Taq DNA Polymerase MasterMix (Ampliqon A/S, Odense, Denmark), 10 pmol of each primer, and 1 µl of template DNA. PCR was performed in BioER XP cycler (Zhejiang, China).
The PCR amplification of the 16S rRNA gene of four representative strains including I1, I2, MR1, and KH7 was performed using the primer pair 27F/1492R 54,55 . PCR was performed under the following conditions initial denaturation 5 min at 94 °C followed by 30 cycles of 35 s at 94 °C, 30 s at 55 °C and 2 min at 72 °C, with a final extension of 10 min at 72 °C.
Amplification of gyrB and infB genes was performed using gyrB 01-F/gyrB 02-R and infB 01-F/infB 02-R primer pairs (Table 4 16 ), respectively using the conditions previously published 16  www.nature.com/scientificreports/ The PCR program used for amplification of fusA, leuS, and pyrG genes consisted of an initial denaturation 2 min at 94 °C followed by 10 cycles of 1 min at 94 °C, 1 min at 60 °C and 1 min at 72 °C, in the following, 20 cycles of 1 min at 94 °C, 1 min at 60 °C and 1 min at 72 °C with a final extension of 5 min at 72 °C (Table 4 56 ). The PCR products were electrophoresed in 0.5 × TBE buffer (tris-borate) and visualized by Safe DNA Gel Stain (Pishgam, Tehran, Iran).
The amplified fragments were sequenced by "Cardiogenetic Research Center" at "Shahid Rajaee Heart Hospital" (Tehran, Iran) using Sanger method. Sequences were compared with those deposited in National Center for Biotechnology Information (NCBI) and edited using BioEdit v. 7.0.0 57 . Alignments were performed in CLUSTAL-W 58 . Maximum likelihood phylogenetic trees were constructed using MEGA X 59  Ethical approval. This research does not contain any studies with human participants or animals.
We confirm that all the methods were carried out in accordance with relevant Institutional guidelines and regulations.