Hyperinduction of pectate lyase in Dickeya chrysanthemi EC16 by plant-derived sugars

ABSTRACT Pectate lyase (Pel) synthesis in Dickeya chrysanthemi has been reported to be hyperinduced in planta and also in the medium containing plant extract in addition to polygalacturonate. In this study, the major components of Pel-hyperinducing fractions were found to be glucose, fructose, and sucrose by TLC and NMR. From the analysis of the sugars and their derivatives, it was found that acyclic d-hexoses with the trans relationship between C-3 and C-5 hydroxyl groups were found to be basic structure required for hyperinducing the expression of a major isozyme in infected plants (i.e. pelE). From the fact that some non-metabolizable sugars such as 2-deoxy-d-glucose and d-fucose could lead to hyperinduction and that the hyperinduction was observed only in the medium containing low concentration (<0.25%) but not higher of the sugars was added, these sugars may be considered to participate in hyperinduction as the signal rather than through their metabolism.

Pathogenicity of Dickeya spp. is mainly determined by the production of a large amount of plant cell wall-degrading enzymes and, in particular, pectate lyases (Barras et al. 1994;Toth et al. 2003) that cleave randomly α-1,4 galacturonosyl linkages of polygalacturonate by a β-elimination mechanism (Collmer and Keen 1986). Among the several pectate lyases (Pel), PelE appears to be the most important isozyme in plant maceration (Payne et al. 1987), followed by PelB, PelC, and PelA (Hugouvieux-Cotte-Pattat et al. 1996). Apart from these important virulence factors, Dickeya spp. also respond to small molecules in their environment such as those released by plants, after plant cell wall degradation and pectate metabolism (Charkowski et al. 2012).
Plants employ abundant small molecules in signaling and regulatory processes (Tholl & Aharoni 2014) and plant signals have been thought to play the key role in pathogenplant interaction. For example, Agrobacterium tumefaciens responds to plant-derived molecules such as sugars and phe-nolic compounds as a signal to regulate its virulence genes (Stachel et al. 1985;Bolton et al. 1986;Subramoni et al. 2014). Likewise, D. chrysanthemi recognizes the plant signals for the synthesis of Pel at maximum level (>200-fold) higher than the basal level by adding plant extract together with their substrate (polygalacturonate) compared to only a 9-fold induction occurred with polygalacturonate alone, a process called hyperinduction (Bourson et al. 1993). A novel regulatory protein, called Pir (plant inducible regulator), was found to be responsible for Pel-hyperinduction in responding to the plant-derived signals (Nomura et al. 1998;1999). However, it has not been elucidated what is (are) the signal molecule(s) responsible for Pel-hyperinduction. Here we report the characterization of plant components required for pelE-hyperinduction in D. chrysanthemi EC16 as the acyclic D-hexoses with the trans relationship between C-3 and C-5 hydroxyl groups.

Enzyme assays
Plate assays for Pel were done as previously described (Joko et al. 2007b). The plate assay medium contained 1% PGA, 1% yeast extract, 0.38 μM CaCl 2 in 100 mM Tris-HCl (pH 8.5). The medium was supplemented with 100 μg/ml ampicillin and solidified with 0.8% (w/v) agarose. Each petri plate (φ100 mm) contained 25 ml of the Pel assay medium. Wells were made in the agarose medium with a cork borer (φ5 mm), and the bottoms were sealed with 10 μl of 0.8% (w/v) agarose containing 100 μg/ml ampicillin.
To determine the total activity of Pel, bacterial cells were grown at 27°C until early stationary phase by measuring optical density at OD 660 with BACT-500 (Intertech. Inc., Japan). One ml of the culture was sonicated two times for 20 s at 70 pulse/min (Ultrasonic Disrupter UD-200, Tomy Inc., Tokyo) on the ice and then centrifuged at 12,000 rpm for 5 min to remove cell debris. Ten μl of this supernatant was applied to each well in the plates and incubated at 27°C for 16-18 h. The clear zone around the wells was observed due to the degradation after flooding with 5N of H 2 SO 4 .
Spectrophotometric assay of Pel activity was performed as done by Collmer et al. (1982) with minor modification. After a thorough mixing, the optical density at 230 nm was measured every min (Ultrospect 3000, Pharmacia Biotech, Cambridge, England). One unit of Pel activity was defined as the enzyme required for an increase of 1 × 10 −3 optical density at 230 nm in 1 min.

Pel-hyperinduction by various plant extracts
Extracts of several vegetables (potato tuber, radish root, turnip root, carrot root, Chinese cabbage, cauliflower, celery, green pak choy, asparagus, and broccoli) and fruits (tomato, apple, banana, cantaloupe melon, avocado, strawberry, pineapple, kiwifruit, oranges, and mango) were tested for hyperinducing ability of Pel. Extracts were obtained by cutting 1 g of vegetables and fruits into small pieces, placed into a test tube containing 10 ml of DDW, then autoclaved for 1 min at 121°C. One-tenth volume of this extract was added into the growth medium. After growth in M63 + 0.2% glycerol and 0.2% PGA with or without plant extract, total activity of Pel was assayed spectrophotometrically at OD 230 nm as described above. When the ratio of Pel-specific activity in PGA + plant extract to that in PGA was over fivefold, we considered it as hyperinduced level.

Fractionation of pel-hyperinducing ability in potato extract
The hot-water extract was obtained after the filtration of the autoclaved chopped potato tuber in DDW through filter paper (φ90 mm, Toyo 2500) and the filtrate was fractionated by molecular weight using sequential centrifugal filter devices with the cutoffs of 100, 50, 30, 10, and 3 kDa (Centricon, Millipore, Bedford, USA). To test the possibility of the involvement of macromolecules, the active fraction was treated with either one of lipase, proteinase K, and DNase I for 60 min at 37°C followed by boiling for 15 min to inactivate those enzymes. The filtrate was concentrated by freeze-drying (FDU-830, Eyela, Tokyo) and the debris was dissolved in 100% methanol. After centrifugation of the suspension at 10,000 × g for 10 min, the supernatant and pellet were used as a methanol-soluble fraction and insoluble fraction, respectively. The centrifugation was repeated once more. Both fractions were then evaporated using a rotary evaporator at 30°C; the pellet was dissolved in sterile DDW for testing Pel-hyperinducing ability. The hyperinducing activity in each fraction was assayed by adding each fraction to an equal volume of 2× concentrated M63 minimal medium containing 0.4% glycerol and 0.4% PGA.

Elution from Toyopearl HW-40C column
The methanol-soluble fraction, in which the hyperinducing activity was found, was loaded into the vertical column (φ2 cm × 30 cm) of Toyopearl HW-40C (Tosoh) silica gel and eluted with 100% methanol. The fractions after elution were evaporated using rotary evaporator at 30°C. The pellet was dissolved in 5 ml of sterile DDW and 0.5 ml of this solution was added into M63 + 0.2% glycerol + 0.2% PGA.

Thin-layer chromatography
The fractions showing Pel-hyperinducing ability after the elution from HW-40C silica gel column were spotted onto TLC plate. Then, it was developed in the mixtures of solvents (methanol/DDW, acetic-acid/ethyl-acetate/methanol/DDW, or ethyl-acetate/hexane). TLC plate was allowed to thoroughly air dry and the spots were observed under UV light and after spraying with 5% sulfuric acid in ethanol or with 0.2% naphthoresorcinol in ethanol (Mahfut et al. 2016).

Analysis with nuclear magnetic resonance
The samples for NMR were prepared by dissolving an analyte in a deuterium solvent [methyl-d 3 alcohol-d (CD 3 OD) and deuterium oxide (D 2 O)]. Sample dissolved in 100% methanol was evaporated using rotary evaporator for 30 min. A deuterium solvent was added into the flask, mixed carefully, and evaporated for 30 min. Several drops of deuterium solvent were added to the flask, mixed thoroughly, and poured into NMR microtube (φ5 mm) for the analysis with FT-NMR (JNM-EX270, Jeol). (Shen et al. 1992) was fused into the genes of Pel  The cell density was checked every 2 h by reading its optical density (OD 660 nm) with Bact-Monitor 500 (Intertech.Inc., Japan), and the expression of pelE was assayed by reading the light production from the fused lux genes using chemiluminescence reader (Hamamatsu Photonics, Hamamatsu, Japan). The mean of 10 times of 1 s readings in 500 μl samples at every 2 h was expressed as the specific activity in terms of cell density (cpm/OD 660 ).

Effects of plant extract on pel-hyperinduction
When D. chrysanthemi EC16 was grown in M63 minimal medium containing 0.2% glycerol and 0.2% PGA together with 1% (w/v) potato extract, the hyperinduction of Pel was confirmed ( Figure 1). Besides potato extract, the extract of various vegetables such as radish, turnip, carrot, Chinese cabbage, cauliflower, celery, green pak choy, asparagus, and broccoli and those from many fruits except avocado and mango were confirmed to hyperinduce the synthesis of Pel too (Figure 2(A,B)).

Identification of pel-hyperinducing compound(s) in potato
When a hot-water extract of potato was treated with lipase, proteinase K, and Dnase I, the hyperinducing ability of Pel was not affected. Thus, the component(s) responsible for Pel-hyperinduction seem not to be these macromolecular substrates or their breakdown products. Also, as the extract could be autoclaved at 121°C for 10 min without losing the Pel-hyperinducing ability, the compound(s) may be heat stable (Figure 3(A)). In the treatment with centrifugationdependent filtration (Centricon) with several sizes of molecular sieves, the Pel-hyperinducing ability was found in the fraction smaller than 3 kDa (Figure 3(B)). When potato extract was dried by rotary evaporator and redissolved in 100% methanol, the Pel-hyperinducing ability was found in methanol soluble fraction (Figure 3(C)).

Silica gel column and TLC
When the above methanol-soluble fraction was loaded onto Toyopearl HW-40C silica gel (Tosoh Corp., Tokyo) vertical column (φ2 cm × 30 cm) and eluted with 100% methanol, Pel-hyperinducing activity was mainly found in 5 ml fractions with 4, 5, 6, and 7; these fractions were analyzed by ascending TLC on silica gel and were developed with chloroform/methanol (1:2), then were sprayed with H 2 SO 4 /EtOH (5:95). The small tailing spots appeared on the silica TLC plate suggested the presence of sugars in these fractions. Fractions 5 and 6, which showed similar Rf values at ca. 0.35 on the TLC plate, were collected and processed again with a silica gel column. The re-collected fractions were then analyzed by Figure 1. Synthesis of Pel in M63 minimal medium containing 0.2% glycerol, M63 + 0.2% glycerol + 0.2% PGA, or M63 + 02% glycerol + 0.2% PGA + 1% potato extract. Total Pel activity was visualized by plate assay method as described in the Section 'Materials and methods.' Figure 2. The effect of various plant extracts on Pel-specific activity. D. chrysanthemi EC16 was grown in M63-0.2% glycerol supplemented with 0.2% PGA and the extracts of vegetables (A) and fruits (B). Pel activity was assayed spectrophotometrically and expressed as its specific activity (U/OD 660 ). Bacterial strain was grown in indicated medium at 27°C until early stationary. The data were shown as the means and standard errors of three replicates (shown in bars).
spraying with naphthoresorcinol reagent onto TLC after the development with the solvents: ethyl acetate/2-propanol/ DDW (6:6:1). Few clear spots with small tailing were observed. When the samples were treated with acetate anhydride and pyridine (1:2), they showed the similar spots at known Rf values as those of authentic D-glucose, D-fructose, and sucrose on TLC plate (Figure 4).

NMR analysis
13 C NMR (nuclear magnetic resonance) analysis of the abovementioned Pel-hyperinducing fractions also indicated the fraction contained mainly a mixture of sugars ( Figure 5 (A)). These NMR peaks were compared with those of several authentic sugars ( Figure 5(B)). The peak pattern and comparative analysis of spectra of NMR data indicated that this active fraction contained D-glucose, D-fructose, and sucrose as the major components.

Hyperinducibility of Pel by various sugars
Based on the results of TLC and NMR analyses, we tested the effects of several commercially available sugars on Pel production of D. chrysanthemi EC16. At their concentration, less than 0.25% (w/v), D-glucose, D-fructose, and sucrose in addition to 0.2% PGA and 0.2% glycerol could hyperinduce the synthesis of Pel. However, at higher concentration of   these sugars (>0.25%), they seemed to have repressed it ( Figure 6(A-C)). When the expression of pelE isozyme was examined using gene fusion construct with promoter-less luxA-E cassette of Vibrio fischerii, low concentration (0.1%) (w/v) of the monosaccharides such as L-arabinose, D-fructose, D-glucose, Dgalactose, D-mannose, 2-deoxy-D-glucose and D-fucose, Dribose, D-xylose, the disaccharides such as melibiose, sucrose, and the trisaccharides such as raffinose, and sugar alcohol such as D-mannitol were found to hyperinduce the expression of pelE (Table 2).
In this study, low concentration of many of the tested monosaccharides, disaccharides, and trisaccharides was shown to hyperinduce the expression of pelE in the simultaneous presence of PGA at a similar level as that of Pel in the medium containing PGA and plant extract. However, the other pel isozymes did not show hyperinduction in the same medium (Table 3). We found that pelA::lux expression was increased in acidic conditions (data not shown), and thus the different isozymes of Pel may recognize different signals for its hyperinduction.
The inducible monosaccharides commonly exist both as cyclic and acyclic form (D-form) and share common trans relationship between C-3 and C-5 hydroxyl group. There were some exceptions, for example, D-ribose, D-xylose, and L-arabinose which possess only five carbons (pentose) and did not share the configuration at C-3 and C-5 ( Figure 7) were able to hyperinduce pelE expression. Methylated monosaccharide glycosides such as methyl β-L-arabinopyranoside, methyl α-D-galactopyranoside, methyl α-D-glucopyranoside, methyl α-D-mannopyranoside, methyl α-Lrhamnopyranoside, methyl β-D-xylopyranoside which cannot perform acyclic (Figure 8) form failed to hyperinduce the expression of pelE. When these sugars are esterified with a methyl group, which cannot form the acyclic anymore, lost also the ability to Pel-hyperinduction. So, the minor acyclic form of the above hexose may be effective for its hyperinduction. 3-O-Methyl-D-glucose which OH was methylated at C-3 still exists as acyclic form but non-inducible, indicating that C-3 hydroxyl group would be essential for the hyperinduction. Other monosaccharides such as D-arabinose, L-glucose, L-ribose, L-xylose, L-rhamnose, and rare sugars (Dpsicose, D-sorbose, L-fructose, L-psicose, L-sorbose, L-tagatose) except d-tagatose and d-sorbitol which is able to form acyclic but having different a configuration of C-3 and C-5 hydroxyl groups from the above failed to hyperinduce the expression of pelE (Figure 9).
The oligosaccharides such as sucrose and raffinose, which exist only as cyclic (Figure 10), were effective for the hyperinduction. This may be due to their cleavage which leads to the above acyclic hexose with the same configuration at C-3 and C-5 hydroxyl groups. In the case of melibiose, it also possesses same C-3 and C-5 hydroxyl configuration and can perform an acyclic form of its monosaccharide chain, but it could hyperinduce pelE. This may be due to the effect of the linkage to other monosaccharide at C-6. However, non-effective oligosaccharides such as D-cellobiose, lactose, maltose, and maltotriose which share common configuration at C-3 and C-5 having linkage to other monosaccharide at position C-4 ( Figure 11). This may be the reason why they were found Figure 7. Chemical structure of inducible monosaccharides. These monosaccharides commonly exist both as cyclic and acyclic form (D-form) and share common trans relationship between C-3 and C-5 hydroxyl group. D-ribose, D-xylose, and L-arabinose which possess only five carbons (pentose) and did not share the configuration at C-3 and C-5 are exception. Lux-specific activity was expressed as the (cpm/OD 660 ) of the bacterial culture. The data were expressed as the mean of peak Lux-specific activities from at least three independent experiments.
to be not inducible. Trehalose which cannot form acyclic form was also found to be non-effective. These findings suggested the importance of acyclic form existence and OH configuration of sugars for hyperinduction of pelE in D. chrysanthemi EC16. The non-metabolizable sugars for D. chrysanthemi EC16 such as 2-deoxy-D-glucose and D-fucose (6-deoxy-D-galactose) which share common configuration at C-3 and C-5 hydroxyl groups were also shown to be equally effective for hyperinduction of the synthesis of Pel. Thus, these sugars may act as signals probably in a similar manner as the interaction with periplasmic binding protein and a transmembrane signal protein as shown by ChvE of A. tumefaciens or galactose/glucose-binding protein (GBP) of E. coli (Shimoda et al. 1993). Likewise, Shimoda and associates (1990) investigated the effect of 2-deoxy-D-glucose on the vir genes expression. Although A. tumefaciens did not use this sugar analog for growth, it was shown that this sugar was required for a marked increase in the expression of vir genes. Noneffective sugars for the induction of vir genes in A. tumefaciens have a different C-3 stereochemical structure from that of the effective sugars.
In A. tumefaciens, the monosaccharides that induce the expression of vir genes are clustered in aldoses group. Periplasmic loop of VirA protein, a member of two-component regulatory system virA-virG (Stachel & Zambrysky 1986;Figure 9. Chemical structure of non-inducible monosaccharides. D-arabinose, L-glucose, L-ribose, L-xylose, L-rhamnose, and rare sugars (D-psicose, D-sorbose, L-fructose, L-psicose, L-sorbose, L-tagatose) except D-tagatose and D-sorbitol which is able to form acyclic but having different configuration of C-3 and C-5 hydroxyl groups was non-effective.  Leroux et al. 1987), and the periplasmic sugar-binding protein ChvE (Huang et al. 1990) are required to specifically recognize the stereostructures of aldoses ).
The effective sugars for Pel-hyperinduction in D. chrysanthemi seemed to share similarity with those for vir induction in A. tumefaciens, though there were some exceptions. L-arabinose, D-fucose, D-galactose, D-glucose, D-xylose, and 2-deoxy-D-glucose were shown to be effective for both plant pathogenic bacteria; however, D-fructose, D-ribose, D-mannitol, and sucrose were shown to be effective only for D. chrysanthemi EC16 but not for A. tumefaciens as described by Cangelosi et al. (1990). This variability might be due either to the difference in their degradation or to the difference in the structure of some sensor protein(s) in D. chrysanthemi and that in A. tumefaciens. The elucidation of this specific recognition of sugars by such sensor (or receptor) proteins will lead to a better understanding of the mechanism of in planta hyperinduction of Pel. Figure 10. Chemical structure of effective oligosaccharides. Sucrose and raffinose, which exist only as cyclic were effective for the hyperinduction. Their cleavage may lead to the above acyclic hexose with the same configuration at C-3 and C-5 hydroxyl groups. In the case of melibiose, though it possesses same C-3 and C-5 hydroxyl configuration and can perform an acyclic form of its monosaccharide chain, it linked to other monosaccharide at C-6. Figure 11. Chemical structure of non-effective oligosaccharides. D-cellobiose, lactose, maltose, and maltotriose which share common configuration at C-3 and C-5 linked to other monosaccharide at position C-4. Trehalose was also found to be only in cyclic form.