Different Cell Wall-Degradation Ability Leads to Tissue-Specificity between Xanthomonas oryzae pv. oryzae and Xanthomonas oryzae pv. oryzicola.

Xanthomonas oryzae pv. oryzae (Xoo) and Xanthomonas oryzae pv. oryzicola (Xoc) lead to the devastating rice bacterial diseases and have a very close genetic relationship. There are tissue-specificity differences between Xoo and Xoc, i.e., Xoo only proliferating in xylem vessels and Xoc spreading in intercellular space of mesophyll cell. But there is little known about the determinants of tissue-specificity between Xoo and Xoc. Here we show that Xoc can spread in the intercellular spaces of mesophyll cells to form streak lesions. But Xoo is restricted to growth in the intercellular spaces of mesophyll cells on the inoculation sites. In vivo, Xoc largely breaks the surface and inner structures of cell wall in mesophyll cells in comparison with Xoo. In vitro, Xoc strongly damages the cellulose filter paper in comparison with Xoo. These results suggest that the stronger cell wall-degradation ability of Xoc than that of Xoo may be directly determining the tissue-specificity.


Introduction
Xanthomonas oryzae pv. oryzae (Xoo) and Xanthomonas oryzae pv. oryzicola (Xoc) lead to bacterial blight and bacterial leaf streak in rice (Oryza sativa L.), respectively [1]. Bacterial blight and bacterial leaf streak are devastating bacterial disease of rice worldwide [1,2]. Xoo and Xoc both belong to Xanthomonas species in the Gammaproteobacteria and have a very close genetic relationship [1,2]. However, Xoo enters rice leaf through hydathodes or wounds then multiplies in xylem vessels to cause leaf blight [1]. In contrast, Xoc gains access to leaf by stomata or wounds and multiplies in the sub-stomatal cavity [1]. Xoc can spread in the intercellular spaces of mesophyll cells to form streak lesions between the veins [1]. But the determinants of Xoo and Xoc parasitic tissue-specificity are still largely unknown [3].
Many factors possibly determine the tissue-specificity of Xanthomonas species among vascular and non-vascular pathogens. The minimal distinguishing differences, such as the subtler difference of amino acid and non-coding nucleotide polymorphisms, maybe determine the tissue-specificity between vascular and non-vascular pathogens [4]. The four positions of amino acid residue from HpaA and XpsD proteins correlate with tissue-specificity among Xoc strain BLS256, Xoo strain KACC100331 and MAFF311018, and other vascular or non-vascular Xanthomonas strains [5]. HpaA is a secreted and translocated protein as the type III secretion system (T3SS) components that affect the secretion of effectors and translocators [5,6]. XpsD is an outer membrane protein that serves as a gatekeeper for the

Normal and Cryo Field-Emission Scanning Electron Microscopy
On the corresponding day, the rice leaves at the inoculation sites were longitudinally cut for exposing cell wall and were then cut into 2 mm × 2 mm blocks with new razors and the filter paper strips were cut into 2 mm × 2 mm blocks, then the leaf blocks and the small strips were immediately immersed in 2.5% (v/v) glutaraldehyde in sodium phosphate buffer (0.1 M, pH 7.2) at 4°C for 12 h. The sample blocks were dehydrated in a gradient series of ethanol, then dried in the critical point equipment (HCP-2, Hitachi, Tokyo, Japan), sputter-coated with platinum in sputtering apparatus (MCIOO, Hitachi, Tokyo, Japan), and observed with a field-emission scanning electron microscope (SU8010, Hitachi, Tokyo, Japan). To quantify the images of cell wall with cellulose microfibrils at high magnification and the images with interstitial cavity between fibers or cellulose microfibrils on filter paper, 8-14 images were obtained from three plants in two independent inoculations and 4-15 images were obtained from three filter papers in one treatment.
For in vivo analyzing the growth of bacteria on rice leaf, the fresh rice leaves with inoculation sites and non-inoculation sites were rapidly frozen in solid liquid-nitrogen, sublimated, and sputter-coated with platinum in cryo-scanning electron microscopy transfer system (PP3010T, Quorum, London, England). The coated leaves were observed under a field-emission scanning electron microscope (SU8010, Hitachi, Tokyo, Japan) at 1 kV and a working distance of 8.3 mm.

Transmission Electron Microscopy
The leaf tissues at the inoculation site and non-inoculation site (adjacent to inoculation site) were cut into 2 mm × 1 mm blocks which were then fixed in 2.5% (v/v) glutaraldehyde in sodium phosphate buffer (0.1 M, pH 7.2) at 4°C overnight, and washed for 30 min with the same buffer three times. The leaf blocks were post-fixed in 1% (w/v) osmium tetroxide for 2 h, washed 30 min with the same buffer three times, and dehydrated in a series of acetone concentration. Dehydrated blocks were progressively infiltrated and embedded in Spurr's resin (SPI, SPI Chem, West Chester, United States), then polymerized at 65°C for 48 h. The samples were cut into ultrathin sections (60-70 nm thick) with diamond knife, stained with 2% uranyl acetate, and observed with a Hitachi transmission electron microscope (H-7650, Hitachi, Japan) at 80 kV.

Cellulose Synthase Gene Expression Analysis
Total RNA of rice leaves isolated from the 3-cm leaf fragments including inoculations sites, was extracted by using Trizol reagent (Invitrogen, Carlsbad, CA, USA). An aliquot (5 µg) of total RNA was firstly treated with RNase-free DNase I (Invitrogen, Carlsbad, CA, USA) to remove potentially contaminating DNA, and then first-strand cDNA was reverse transcribed with oligo(dT) 18 primer using M-MLV reverse transcriptase (Promega, Madison, WI, USA) according to the manufacturer's protocols. Quantitative reverse transcription polymerase chain reaction (qRT-PCR) was carried out by using cellulose synthase gene (CESA) specific primers (Supplementary Table S1), as described previously [20]. The transcript level of rice actin gene was used to standardize the RNA sample of each RT-PCR, and the expression level relative to that of IR24 inoculated by PXO99 A at 0 h was calculated. Each qRT-PCR assay was biologically carried out twice with similar results, within each repetition having three technical replicates and only one biological replicate was presented.

Statistical Analysis
The significant differences among lesion length, percentage of images with cellulose microfibril, percentage of image with interstitial cavity between fibers or cellulose microfibrils on filter paper and relative expression level of CESA genes were assessed using pairwise Student's t-test in Excel (Microsoft, http://www.microsoftstore.com).

The Leaf Tissue Morphology of Susceptible Rice Infected with Xoo and Xoc
Leaf is a major infection site of Xoo and Xoc [1]. To study whether Xoo can grow and multiply in the extracellular spaces of mesophyll cells, we firstly analyzed the leaf phenotypes of susceptible rice IR24 inoculated with Xoo strain PXO99 A and Xoc strain RH3. On the inoculation sites, the leaves inoculated with Xoo represented water-soaked symptom with little brown color around inoculation sites and the leaves inoculated with Xoc were yellow colored at 5 days after infection (DAI) ( Figure 1A). There were many bacteria around mesophyll cells in leaf inoculated by Xoo and Xoc on inoculation sites ( Figure 1B-1,C-3). In the non-inoculation sites (adjacent to the inoculation site), only rice leaves inoculated with Xoc represented yellow streak lesions with yellow exudates on leaf surface ( Figure 1A) and contained many bacteria around the mesophyll cells ( Figure 1C-4); but there was no bacterium around the mesophyll cells in rice leaf inoculated by Xoo ( Figure 1B-2). At 5 DAI, some bacterial exudates showed on stoma of inoculation site for both leaves inoculated with Xoo and Xoc (Supplementary Figure S1); but there were many bacterial exudates on stoma of non-inoculation site (adjacent to the inoculation site) leaf inoculated with Xoc and no bacterial exudates on stoma of non-inoculation site leaf inoculated with Xoo (Supplementary Figure S1). At 14 DAI, the lesion length of the rice leaf inoculated by Xoo strain PXO99 A was only 0.5 cm (the diameter of inoculation syringe was 0.5 cm); the lesion length of rice leaf inoculated by Xoc strain RH3 exceeded 2 cm ( Figure 1D).

The Cell Wall Surface Structures of Mesophyll Cells in Rice Infected with Xoo and Xoc
Plant cell wall is the first physical barrier against pathogens in plant-pathogen interaction [21]. Therefore, we analyzed the cell wall structures of susceptible rice IR24 leaves infected with Xoo and Xoc. In leaves infected with Xoo strain PXO99 A , the cell wall surface of the mesophyll cells was still flat (low magnification images) and represented many cellulose microfibrils (black arrow) (high magnification images) at 3 and 5 DAI which were almost similar to a lot of cellulose microfibrils on the flat surface of cell wall at 0 DAI ( Figure 2A). In rice leaves infected with Xoc strain RH3, the cell wall surface of mesophyll cells had more pits (low magnification images), especially on the sites of cell wall in contact with bacterium, and fewer cellulose microfibrils (high magnificent images) at 3 and 5 DAI in comparison with the flat cell wall surface and many cellulose microfibrils at 0 DAI ( Figure 2B). At 5 DAI, the number of images with cellulose microfibrils (high magnification image) from leaves inoculated with Xoo strain PXO99 A were significantly (p < 0.01) more than the number of images with cellulose microfibrils from leaves inoculated with Xoc strain RH3 ( Figure 2C).

The Cell Wall Inner Structure of Mesophyll Cells in Rice Infected with Xoo and Xoc
Under transmission electron microscopy (TEM) analyzing rice mesophyll cells, the normal cell wall represented low electron-density (transparent) layer structure; the cell wall broken by pathogens represented high electron-density (opaque) layer structure because of osmiophilic material deposition in cell wall [22,23]. To further investigate the cell wall inner structure, we observed the cell wall of the mesophyll cell in susceptible rice IR24 inoculated by Xoo strain PXO99 A and Xoc strain RH3 under TEM. There was slightly higher electron density of cell wall (shown by black arrow head) and distinct layers of cell wall in mesophyll cells of leaves inoculated with Xoo at 3, 5 DAI in comparison with the electron-density of cell wall at 0 DAI ( Figure 3A). However, there was dramatically higher electron-density of cell wall (black arrow head) and non-layers of cell wall in mesophyll cells of leaves inoculated with Xoc at 3, 5 DAI in comparison with the electron density of cell wall at 0 DAI ( Figure 3B). At the same time, the out layer of the cell wall contacting with the bacterium consisted of many electron-dense (osmiophilic) particles in mesophyll cell of leaves inoculated with Xoc at 3 DAI ( Figure 3B). Cellulose as the major component of plant cell wall is synthesized by cellulose synthase genes (CESA) [21]. To further analyze the cell wall inner structure, we found that the relative expression levels of CESA4/7/9 significantly reduced in IR24 rice leaves inoculated both with Xoo strain PXO99 A Pathogens 2020, 9, 187 5 of 11 and with Xoc strain RH3 (Supplementary Figure S2). However, the expression of CESA4/7/9 in rice leaves inoculated by Xoc were significantly (p < 0.01 and p < 0.05) induced to lower levels than in rice leaves inoculated by Xoo at 2, 4, 12, and 24 hours after inoculation (Supplementary Figure S2). 1A). There were many bacteria around mesophyll cells in leaf inoculated by Xoo and Xoc on inoculation sites ( Figure 1B-1,C-3). In the non-inoculation sites (adjacent to the inoculation site), only rice leaves inoculated with Xoc represented yellow streak lesions with yellow exudates on leaf surface ( Figure 1A) and contained many bacteria around the mesophyll cells ( Figure 1C-4); but there was no bacterium around the mesophyll cells in rice leaf inoculated by Xoo ( Figure 1B-2). At 5 DAI, some bacterial exudates showed on stoma of inoculation site for both leaves inoculated with Xoo and Xoc (Supplementary Figure S1); but there were many bacterial exudates on stoma of non-inoculation site (adjacent to the inoculation site) leaf inoculated with Xoc and no bacterial exudates on stoma of non-inoculation site leaf inoculated with Xoo (Supplementary Figure S1). At 14 DAI, the lesion length of the rice leaf inoculated by Xoo strain PXO99 A was only 0.5 cm (the diameter of inoculation syringe was 0.5 cm); the lesion length of rice leaf inoculated by Xoc strain RH3 exceeded 2 cm ( Figure 1D).

The Cellulose Filter Papers Less Damaged by Xoo than Xoc
The qualitive filter paper (Grade 1, GE Whatman, Hangzhou, China), which mostly consisted of cellulose, hemicellulose, and lignin (www.gelifesciences.com), were treated with bacteria to in vitro analyze the cell wall-degradation ability of Xoo strain PXO99 A and Xoc strain RH3. After 5 days, the surface of filter paper treated by Xoo showed few interstitial cavities between big fibers in low magnification image and no interstitial cavities between the densely packing microfibrils in high magnification image ( Figure 4A); however, the surface of filter paper treated by Xoc showed many interstitial cavities between big fibers in low magnification image and many interstitial cavities between microfibrils in high magnification image ( Figure 4B). Meanwhile, the surface of filter paper treated by H 2 O (negative control) also showed few interstitial cavities between big fibers or microfibrils after 5 days treatment ( Figure 4C). The number of images with interstitial cavities between fibers or microfibrils in filter paper treated by Xoc strain RH3 was 3-fold higher than the number of images with interstitial cavities between fibers or microfibrils in filter paper treated by Xoo strain PXO99 A or H 2 O ( Figure 4D). Pathogens 2020, 9, 187 6 of 11 magnification images) at 3 and 5 DAI which were almost similar to a lot of cellulose microfibrils on the flat surface of cell wall at 0 DAI (Figure 2A). In rice leaves infected with Xoc strain RH3, the cell wall surface of mesophyll cells had more pits (low magnification images), especially on the sites of cell wall in contact with bacterium, and fewer cellulose microfibrils (high magnificent images) at 3 and 5 DAI in comparison with the flat cell wall surface and many cellulose microfibrils at 0 DAI ( Figure 2B). At 5 DAI, the number of images with cellulose microfibrils (high magnification image) from leaves inoculated with Xoo strain PXO99 A were significantly (p < 0.01) more than the number of images with cellulose microfibrils from leaves inoculated with Xoc strain RH3 ( Figure 2C).

The Cell Wall Inner Structure of Mesophyll Cells in Rice Infected with Xoo and Xoc
Under transmission electron microscopy (TEM) analyzing rice mesophyll cells, the normal cell wall represented low electron-density (transparent) layer structure; the cell wall broken by pathogens represented high electron-density (opaque) layer structure because of osmiophilic material deposition in cell wall [22,23]. To further investigate the cell wall inner structure, we observed the cell wall of the mesophyll cell in susceptible rice IR24 inoculated by Xoo strain PXO99 A and Xoc strain RH3 under TEM. There was slightly higher electron density of cell wall (shown by black arrow head) and distinct layers of cell wall in mesophyll cells of leaves inoculated with Xoo at 3, 5 DAI in comparison with the electron-density of cell wall at 0 DAI ( Figure 3A). However, there was dramatically higher electron-density of cell wall (black arrow head) and non-layers of cell wall in mesophyll cells of leaves inoculated with Xoc at 3, 5 DAI in comparison with the electron density of cell wall at 0 DAI ( Figure 3B). At the same time, the out layer of the cell wall contacting with the bacterium consisted of many electron-dense (osmiophilic) particles in mesophyll cell of leaves inoculated with Xoc at 3 DAI ( Figure 3B). Cellulose as the major component of plant cell wall is synthesized by cellulose synthase genes (CESA) [21]. To further analyze the cell wall inner structure, we found that the relative expression levels of CESA4/7/9 significantly reduced in IR24 rice leaves inoculated both with Xoo strain PXO99 A and with Xoc strain RH3 (Supplementary Figure S2). However, the expression of CESA4/7/9 in rice leaves inoculated by Xoc were significantly (p < 0.01 and p < 0.05) induced to lower levels than in rice leaves inoculated by Xoo at 2, 4, 12, and 24 hours after inoculation (Supplementary Figure S2).  between microfibrils in high magnification image ( Figure 4B). Meanwhile, the surface of filter paper treated by H2O (negative control) also showed few interstitial cavities between big fibers or microfibrils after 5 days treatment ( Figure 4C). The number of images with interstitial cavities between fibers or microfibrils in filter paper treated by Xoc strain RH3 was 3-fold higher than the number of images with interstitial cavities between fibers or microfibrils in filter paper treated by Xoo strain PXO99 A or H2O ( Figure 4D).  The percentage of image with interstitial cavity between fibers in total observed images from filter papers treated by Xoo strain PXO99 A , Xoc strain RH3 and H 2 O at the 5 th day. Data represent mean (at least three images were observed from three different filter papers) ± standard deviation (SD). Double asterisk (**) stand for the significant differences between percentage of images with interstitial cavities among fibers or microfibrils between infilter paper treated by PXO99 A and infilter paper treated by RH3, between infilter paper treated by RH3 and infilter paper treated by H 2 O (negative control) at p < 0.01. n, the total number of observed images.

Discussion
In rice leaf tissue, Xoo proliferates in xylem vessel, while Xoc multiplies in intercellular spaces of mesophyll tissue [1]. However, many wild type Xoo strains and Xoc strains, which are infiltrated into mesophyll tissue after 3-7 days, can both cause water-soaked lesions on leaves of susceptible rice varieties [17,24,25]. At the 5 th day after infiltration inoculation, Xoo and Xoc proliferated in the Pathogens 2020, 9, 187 8 of 11 intercellular space between mesophyll cells, which resulted in yellow lesions and bacterial exudate formations on inoculation sites of leaves ( Figure 1B-1,C-3; Supplementary Figure S1). These evidences suggest that Xoo and Xoc can both grow in intercellular spaces between mesophyll cells. But at the non-inoculation sites (adjacent to inoculation site) of leaves inoculated by Xoo, no Xoo bacteria were observed in the intercellular spaces between mesophyll cells following by no lesions and bacterial exudate formations ( Figure 1A,B-2; Supplementary Figure S1A). However, many bacteria of Xoc were growing in the intercellular spaces between mesophyll cells which caused streak lesions and bacterial exudates formations on the leaves of non-inoculation sites ( Figure 1A,C-4; Supplementary Figure S1B). Almost all the wild Xoc strains can infect indica and japonica rice leaves to form streak lesions along the veins [1,2,8,11]. The mesophyll tissue of rice leaf is composed of the lobed mesophyll cells that are joined to one another by the tight fusion cell wall of the lobes [26]. So, Xoc bacteria must break the cell wall junction between the lobes of mesophyll cell then spread to the neighbor intercellular spaces between mesophyll cells in rice leaf. These evidences indicate that Xoo and Xoc can both grow in the intercellular spaces between mesophyll cells. But only Xoc can break the cell wall junction between mesophyll cells to spread in the extracellular spaces and form streak lesions.
The cell wall is consisted of cellulose microfibrils embedded in gel-like matrix of hemicellulose (xylan, glucuronoxylan, xyloglucan, arabinoxylan, mixed linkage glucan, or glucomannan), pectic polysaccharides, and minor amounts of structural proteins [27]. The T2SS of Xoo can secrete CWDEs such as cellulase, polygalacturonase, xylanase, lipase, and endoglucanase as virulence factors facilitating Xoo invading rice [8,16,28]. The T2SS-deficience or the single endoglucanase gene (eglXoB) loss of Xoo are severely virulence-deficient in rice [28,29]. However, the single mutation of cellulose or lipase gene causes only a partial loss of Xoo virulence [16]. The purified celullase, endoglucanase and lipase of Xoo can induce rice defense response which are suppressed by the T3SS of Xoo [16]. These evidences suggest that the Xoo growing in intercellular of mesophyll cells or multiplying in xylem vessels are dependent on the T3SS but independent on the most CWDEs secreted by T2SS [16]. The T2SS-deficient mutant of Xoc is also severely virulence-deficient in rice [8]. The T2SS component-XpsD protein and the extracellular protease A (EcpA) secreted by T2SS are tissue-specific virulence factors between Xoc strain BLS256 and Xoo strains [5,8]. Many genes of cellulase, xylanase and lipase are analyzed in the genome of Xoc strains [4,12]. The gene mutants of CWDEs are not screened in Xoc strain BLS256 Tn5-mutant library [8] which is possibly because of the CWDEs keeping the housekeeping role in Xoc growth [30]. Furthermore, the cell wall of mesophyll cell and the cellulose filter papers were more severely damaged by Xoc than that of Xoo (Figures 2-4). Meanwhile, Xoc can break the cell wall of lobes in mesophyll cells to form streak lesions ( Figure 1). Therefore, in comparison with Xoo, Xoc has the stronger cell wall-degradation ability by breaking cell wall linkage of mesophyll cell lobes to facilitate tissue-specific spread in the intercellular spaces between mesophyll cells. The strong cell wall degradation ability of Xoc, which may directly determine the tissue-specificity of Xoc, should be determined by the CWDEs in T2SS signal pathways.
A lot of cell wall-related genes of rice are involved in rice-Xoo/Xoc interactions. Rice major resistance (MR) genes Xa4 and Xa27 against Xoo are associated with the up-regulated expression of cellulose synthase genes and the secondary cell-wall thickening in vascular bundle elements [31,32]. Rice lines with overexpression of EXPA1/5/10, which encoding cell wall-loosening proteins, show increased susceptibility to Xoo and Xoc [33]. These evidences imply that Xoo and Xoc may both impair the cell wall mediated resistance in rice. However, no MR gene against Xoc is identified in rice until now [2], only some cell wall-related genes are associated with rice resistance to Xoc. The overexpression of polygalacturonase-inhibiting protein 1/4 (OsPGIP1/4) and small heat shock protein (OsHsp18.0-Cl) genes enhance resistance to bacterial leaf streak in rice [34][35][36]. The polygalacturonase-inhibiting protein functions in inhibiting the activity of enzymes which secreted by bacteria to degrade cell wall component-pectic polysaccharides [34]. The expression of nucleotide-binding site-leucine rich repeats (NB-LRR) maize MR gene Rxo1, which mediates rice resistance to Xoc, can induce lignin deposition on the inoculation sites [37]. Furthermore, The T3SS defective strain of Xoc strain GX01 induces the