From May to July 2021, 13 sugarcane leaves with red stripe symptoms were collected in sugarcane fields of the provinces of Guangdong and Guangxi in China ( Table S1 ). The symptoms of the diseased samples were photographed on site and taken to the laboratory for isolation and identification of the causal agent(s) ( Figure S1 ).

The isolation and purification of bacteria was performed using the agar-streak method. Briefly, symptomatic leaves were first disinfected for 1 min with 75% alcohol and then rinsed three times with sterile water. At the junction of the diseased and healthy leaf tissue, two or three fragments (2-3 mm × 4-5 cm) were cut into small pieces that were introduced in a 2 mL-microtube. This microtube contained one 4 mm-steel ball, two 2 mm-steel balls, and 100 µL of sterile water. Leaf tissue was homogenized for 60 s with a MM400 beater (Verder Shanghai Instruments and Equipment Co., Ltd, Shanghai, China). Subsequently, 400 µL of sterile water were added to each tube. After a 30 min incubation, the supernatant of each tube was spread on nutrient agar (NA) medium using the quadrant streak method. Agar plates were then incubated at 30°C for 12-16 h. Bacterial colonies with a yellow or a white color were isolated, purified, and tested by PCR using Aaa-specific primers RS-ITS-F1 and RS-ITS-R1, as described by Li et al. (2018). The PCR program was 3 min at 94°C for initial denaturation; 35 cycles of 45 s at 94°C for denaturation, 30 s at 58°C for annealing, and 45 s at 72°C for extension; and 10 min at 72°C for final extension. All PCR-positive colonies were stored on NA plates at 4°C and in 20% glycerol at -80°C until further investigations. Five colonies of 24 hour-old bacteria growing on NA medium were randomly selected for each strain to measure their colony size.

Freshly isolated and 24-hour-old single colonies of Aaa CNGX08 (white) and CNGD05 (yellow) were suspended in 50 μL sterile H₂O. A loop of bacterial suspension was placed on a copper mesh for 1 min and excess of liquid was removed with a filter paper. Bacteria were stained with a solution of 2.5% phosphor-tungstic acid (pH 7.0) for 3 s. Morphology of bacterial cells was observed with a HT-7700 transmission electron microscope (Hitachi High-Tech, Shanghai, China) and digital images were acquired with an AMT CCD camera (Hitachi High-Tech).

Bacterial genomic DNA was extracted using the Bacterial Genome DNA Extraction Kit (Tiangen, Beijing, China). DNA quality and quantity tests were performed using 1% agar gel electrophoresis and the SynergyTM H1 multifunctional microplate reader (BioTek, Vermont, USA). The DNA samples were stored at -80°C until further use.

Five housekeeping genes (ugpB, pilT, lepA, trpB, and gltA) were amplified by PCR using primer pairs reported by Feng et al. (2009). The 50 μL reaction mix included 1.0 μL bacterial genomic DNA (100 ng/μL), 20 μL Green Taq Mix (Vazyme Biotech, Nanjing, China), 1.0 μL each of forward and reverse primers (10 μmol/L), and 27 μL sterile water. The PCR program was 5 min at 95°C for initial denaturation; 35 cycles of 30 s at 95°C for denaturation, 30 s at 60°C for annealing, and 30 s at 72°C for extension; and 5 min at 72°C for final extension. The expected amplification product had a size of 444, 398, 489, 434, and 481 bp for ugpB, pilT, lepA, trpB, and gltA, respectively. Bacterial genomic DNA of Aaa strain ATCC 19860 (Institute of Microbiology, Chinese Academy of Sciences) was used as the positive control. Sterile distilled water was used as the blank control. All primer information is given in Table S2 .

All the PCR products (sections 2.2 and 2.5) were purified and then cloned into the pMD19-T vector. Three positive clones per PCR product were randomly selected for sequencing by Sangon Biotech (Shanghai, China). Sequences were deposited at the NCBI GenBank database under accession numbers OP738811-OP738827 for the 16S-23S rDNA ITS region, and under accession numbers ON707276-ON707358 and OP747511-OP747512 for the five house-keeping genes.

A restriction fragment length polymorphism (RFLP) analysis of the Aaa sequences obtained with primers RS-ITS-F1/RS-ITS-R1 was performed in silico using restriction enzymes HindIII and EcoRI with DNAMAN software (Lynnon Biosoft, San Ramon, USA). The restriction patterns were identified and labelled as described by Li et al. (2018).

A concatenated sequence of the five housekeeping genes mentioned in section 2.5 was produced for 54 bacterial strains: 42 Aaa strains (17 obtained in this study and 25 retrieved from the NCBI library), 10 Aac strains, one Aaca strain, and one A. facilis strain (used as outgroup strain). The 54 concatenated sequences were aligned using the ClustalW algorithm in the software MEGA 11 (Tamura et al., 2021). A phylogenetic tree was constructed using the neighbor-joining (NJ) method with 1000 bootstrap replicates. Sequence identity analysis was performed with BioEdit V7.0.9.0 (Borland, Scotts Valley, USA). A heat map was produced with SDTv1.2 (The University of Cape Town, South Africa). The characteristics of all tested strains of Acidovorax are given in Table S1 .

Stalk cuttings with a single bud each were prepared from healthy sugarcane varieties LC07-150 (source of Aaa strain CNGX08) and ZZ8 (source of Aaa strain CNGD05). Cuttings were planted in pots (5.5 x 4 x 6 cm) filled with nutrient soil (Pindstrup Mosebrug A/S, PINDSTRUP, Denmark) and grown in a climate chamber at 30°C with 16 h light/8 h night and 65% humidity. When plants had 1-2 fully expanded leaves, they were inoculated using the leaf puncture method. Briefly, the upper side of the leaf was punctured but not pierced with a disposable syringe and the wounded area was wrapped with sterile cotton. One mL of bacterial suspension (10⁸ CFU/mL) was placed on the cotton for inoculation with a single strain of the pathogen whereas 0.5 mL of each strain was used for mixed inoculations. Only sterile water was used for the control plants. Inoculated plants were grown using the same conditions as described above but humidity was increased to 85%. These plants were also covered with a transparent plastic bag for the first three days post inoculation (dpi).

The pathogenicity assay included eight different treatments, namely two sugarcane varieties (LC07-150 and ZZ8) each inoculated with four different inocula: two single strains of Aaa (CNGX08 and CNGD05), a mix (1:1) of these two strains, and sterile water (negative control). Thirty plants were inoculated for each treatment and the pathogenicity assay was conducted independently two times. Disease incidence (%) was calculated as follows: (Number of plants with red stripes/Total number of inoculated plants) x 100. Area under disease progress curve (AUDPC) was calculated as follows for each treatment of each of the two inoculation experiments (Shaner, 1977):

X_i and X_{i +1} represent the disease incidence at time points T_i and T_{i +1}, respectively; n is the total number of time points (0, 3, 6, 9, 12, and 15 dpi). At 15 dpi, three inoculated leaves with red stripe symptoms were used for bacterial isolation and characterization. Bacterial colonies isolated from symptomatic leaves were identified by PCR with Aaa-specific primers RS-ITS-F1 and RS-ITS-R1 (Li et al., 2018).

Six leaf samples (three from each inoculation experiment) of varieties LC07-150 and ZZ8 inoculated with Aaa strains CNGX08 and CNGD05 were collected at 0, 12, 24, 48, and 72 hours post inoculation (hpi). Endogenous SA content of the leaf tissue was determined following the instructions of a plant SA ELISA kit (Beijing Solarbio Science & Technology Co., Ltd. China). Each leaf subsample (0.1 g each representing the six collected samples) was homogenized in 1 mL of 10 mM PBS buffer (pH = 7.4) using a freeze grinder (JXFSTPRP-CL, Shanghai, China). Homogenized leaf tissue was centrifuged at 2500 rpm and 25°C for 20 min and the supernatant was used for determination of SA content by sandwich ELISA and a 450 nm optical density. Three leaf subsamples were analyzed at each time point for each treatment, and three technical replicates were performed per subsample.

The six leaf samples collected for determination of SA content (section 2.9) were also used to determine relative expression of gene PR-1. The transcript expression of gene PR-1 involved in the SA signaling pathway was determined by RT-qPCR assay using the QuantStudio 3 qPCR System (Applied Biosystems, USA). Total RNA was extracted from the leaf tissue using the TRIzol reagent kit (Invitrogen, USA). The cDNA was synthesized using HiScript II Q RT SuperMix with a qPCR (+gDNA wiper) reverse transcription kit (Vazyme Biotech). The qPCR mix was composed of 10.0 μL of 2× ChamQ Universal SYBR qPCR Master Mix (Vazyme Biotech), 0.4 μL of forward and reverse primers (10 μM), 1.0 μL of cDNA, and sterile high-purity water to obtain a final volume of 20 μL. The qPCR program included denaturation at 95°C for 30 s followed by 40 cycles at 95°C for 10 s, and 60°C for 30 s. The glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene was used as a reference gene (Zheng et al., 2017). The relative expression of gene PR-1 was calculated with the 2^-△△CT method. Characteristics of the primers is given in Table S2 . Three leaf subsamples were analyzed at each time point and for each treatment, and three technical replicates were performed per subsample.

An analysis of variance (one-way ANOVA) was conducted with AUDPC data obtained for the two sugarcane varieties inoculated with the two strains of Aaa (single and mixed inocula). The same type of analysis was performed with each data set of the SA content and of the transcript level of gene PR-1 in plants at each time point (0, 12, 24, 48, and 72 h). Duncan’s test was used to identify mean differences at p ≤ 0.05 between the treatments. All analyses were conducted with the software SAS version 8.01 (SAS Institute Inc., North Carolina, USA).

Seventeen bacterial strains were isolated on NA medium from the 13 sugarcane leaves (A-M) with typical red stripe symptoms collected from the provinces of Guangxi and Guangdong ( Figure S1 ). The color of the colonies varied among these strains ( Figure 1A ). Five leaf samples yielded only white-cream, circular, and translucent colonies whereas another four leaf samples yielded only yellow, circular, and translucent colonies. Colonies of both morphological types were obtained from the remaining four leaf samples. The white-cream colonies grew slower than the yellow colonies as mean size of white-cream colonies was 1.14 mm (± 0.06) vs. 0.91 mm (± 0.06) for yellow colonies after 24 h of growth on NA medium ( Figures 1B, C ). A colony of strain CNGX08 (white-cream) and a colony of CNGD05 (yellow) were taken for observation of bacterial cell morphology by electron microscopy. No bacterial spores were observed in the microscopic preparations and cells of both strains were straight rods with a single polar flagellum ( Figures 1D, E ).

A 454 bp fragment was amplified by PCR from total genomic DNA of the 17 bacterial strains from China using the Aaa-specific primers (RS-ITS-F1/RS-ITS-R1) ( Table S1 ). Based on the nucleotide sequence of this amplicon, the strains originating from two geographical regions (Guangxi vs. Guangdong) or with different morphological types (white-cream vs. yellow) were 98.4-100% identical ( Figure S2 ). These strains had also 98.2-99.5%, 97.7-98.8%, and 96.2-97.3% sequence identity with Aaa strain ATCC 19860 Aac strain W6, and Aaca strain 30134, respectively. The specific PCR amplification and sequence data suggested that the 17 bacterial strains from sugarcane in China belong to the subspecies Aaa.

The 17 Aaa strains from sugarcane in China were divided into three groups based on their in silico restriction patterns (I, II, and IIIb) of the 16S-23S rDNA ITS region (454 bp) cut with enzymes HindIII and EcoRI ( Table 1 ). Pattern RFLP I consisted of two bands (330 and 124 bp) whereas pattern II had three bands (330, 68, and 56 bp). Two bands, one of 386 bp and the other of 68 bp, formed pattern RFLP IIIb. The RFLP pattern I group included seven Aaa strains from the Guangxi province (four with white-cream colonies and three with yellow colonies). RFLP pattern II group contained nine strains from the provinces of Guangxi and Guangdong (four with white-cream colonies and five with yellow colonies). Only strain CNGD01 with white-cream colonies from the province of Guangdong showed RFLP pattern IIIb ( Table 1 ).

To further investigate the genetic variability among the 17 strains of Aaa from sugarcane in China, pairwise sequence comparisons were carried out with the single sequence of five housekeeping genes (ugpB, pilT, lepA, trpB, and gltA) and with the concatenated sequence (2,246 nucleotides) of these five genes. When the housekeeping genes were analyzed separately, the lowest sequence identities were found for gene lepA (95.0-100%), followed by ugpB (95.7-100%) and gltA (96.8-100%). Highest sequence identities were observed for genes trpB (99.5-100%) and pilT (97.9-100%) ( Table S3 ). The 17 strains of Aaa had 97.8-100% sequence identity among each other and 96.7-97.5% with seven strains of Aaa from Argentina when comparisons were based on the concatenated sequence of all five genes. This latter identity was 97.0-97.7%, 94.6-95.4%, and 95.7-96.3% with strain ATCC 19860 of Aaa, strain AacW6 of Aac, and strain 30134 of Aaca, respectively ( Table S3 ).

To further investigate the evolutionary clustering of Aaa, the 17 strains of this subspecies obtained in this study from sugarcane in China were compared to 25 additional strains of Aaa (from bentgrass, maize, millet, rice, sorghum, sugarcane, and Vasey grass), 10 strains of Aac (from melon and watermelon), and one strain of Aaca (from orchid) retrieved from the NCBI library ( Table S1 ). MLST conducted with the concatenated sequence of five housekeeping genes (section 3.4) revealed that all A. avenae strains were divided at subspecies level into three major evolutionary clades ( Figure 2 ). The 10 strains of Aac (from China, Israel, and the USA) and strain 30134 of Aaca (from the USA) were distributed in two separate evolutionary clades. The 42 strains of Aaa forming the third clade were further distributed into three subclades. The 17 strains of Aaa from sugarcane obtained in this study and Aaa strain 30179 from sorghum in Brazil formed subclade Aaa-I, while seven strains from sugarcane in Argentina and 11 additional strains from bentgrass, maize, rice, and Vasey grass from Japan and the USA grouped in subclade Aaa-II. Six strains of Aaa from Creeping bentgrass originating from the USA formed subclade Aaa-III ( Figure 2 ). In subclade Aaa-I, the 17 strains of Aaa from sugarcane obtained in this study were distributed in three different phylogenetic subgroups.

Red stripe symptoms developed in sugarcane varieties LC07-150 and ZZ8 after single and mixed inoculation with strains CNGX08 and CNGD05 of Aaa from China. At 3 dpi, water-green stripes appeared in the middle of the leaf blade and these lesions rapidly turned into red-brown stripes. The sugarcane plants of varieties LC07-150 and ZZ8 inoculated with sterile water remained symptomless ( Figures 3A, B ). The disease incidence of all inoculation treatments with Aaa was greater than 20 and 65% at 3 and 12 dpi, respectively ( Figure 3C ). At 15 dpi, lowest incidence (73.3%) was observed for variety LC07-150 inoculated with strain CNGD05 and highest incidence (88.3%) was found for the same variety inoculated with strain CNGX08. After isolation from symptomatic plants, the strains of Aaa shared 100% sequence identities of ITS of 16S-23S rDNA and same colony color with inoculated strain.

A significant bacterial strain effect (P = 0.0007) but a not significant sugarcane variety effect (P = 0.0680) was found based on AUDPC analysis of all inoculation treatments of combined experiments. The variety x strain interaction effect was also not significant (P = 0.0591). Strain CNGX08 of Aaa was more pathogenic than strain CNGD05 in sugarcane variety LC07-150, while strain CNGD05 was more pathogenic than strain CNGX08 in sugarcane variety ZZ8 ( Table 2 ). Strain CNGX08 was also more pathogenic in LC07-150 when inoculated alone than in mixed inoculation with strain CNGD05. The same result was observed for strain CNGD05 in variety ZZ8.

To investigate whether the SA signal transduction pathway was involved in sugarcane affected by red stripe, the SA content and the expression of the SA-mediated resistance gene (PR-1) were determined in sugarcane varieties LC07-150 and ZZ8 after single or mixed inoculation with strains CNGX08 and CNGD05 of Aaa. Similar changes in SA content were observed in the two sugarcane varieties 0-72 hpi, regardless of the bacterial inoculum ( Figure 4 ). When compared to the control plants inoculated with water, the SA content increased in varieties LC07-150 and ZZ8 after all Aaa treatments. Highest SA increases (40-65%) were observed in these varieties at 24-48 hpi. Significant differences between Aaa treatments were observed at some time points but they were not always consistent throughout the 12-72 hpi. Nevertheless, SA content was greater in varieties LC07-150 and ZZ8 inoculated with strain CNGX08 of Aaa than with strain CNGD05 at 12, 24, and 72 hpi ( Figure 4 ).

Expression of gene PR-1 involved in the SA signal transduction pathway was significantly upregulated in varieties LC07-150 and ZZ8 inoculated with the two strains of Aaa (CNGX08 and CNGD05), as compared to control plants inoculated with water only ( Figure 5 ). Significant increases of PR-1 expression were observed for single and mixed inocula as soon as 12 hpi, but greatest expression was found at 72 hpi. At his time point, PR-1 transcript levels had increased 16-23 times in variety LC07-150 after single or mixed inoculation with strains CNGX08 and CNGD05 of Aaa. A 21-30-time increase was observed for the same strains in variety ZZ8. In each infected sugarcane variety, relative expression of gene PR-1 differed among strains at one time point or another but no consistent pattern was observed, except at 24 and 48 hpi. At these two time points, relative expression of gene PR-1 was higher in plants of the two varieties inoculated simultaneously with strains CNGX08 and CNGD05 than in plants inoculated with the single strains of the pathogen.