Intervene effect on Streptococcus suis biolm formation of emodin extracted from Rheum ocinale Baill


 Background

 Streptococcus suis is a zoonotic pathogen that causes serious systemic infections in pigs and humans. It is a great threat to pig breeding and public health safety. The formation of biofilm was one of the main reasons that it’s difficult to cure. Rhubarb water extract can inhibit the formation of biofilm of Streptococcus suis, but the main functional ingredient was not clear. And what’s the potential mechanism of emodin intervene the biofilm formation? Importantly, the gene expression of luxS of Streptococcus suis was significantly decreased with emodin treatment, if there is a possibility that emodin could combine with LuxS protein which in turn inhibit the formation of biofilm?
Methods

Optimization of green ultrasonic extraction of emodin from R. officinale by Response Surface Methodology. The activity of antibacterial was evaluated by MIC assay. And the ability that intervene biofilm formation was evaluated through Crystal violet staining and SEM. The combination mode between emodin and LuxS protein was identified through molecular docking.
Results

The optimum extraction conditions (ethanol concentration of 80%, extraction time of 21 min, and liquid to solid ratio of 12.5:1 mL/g) were determined through the interaction analysis of different influencing factors and model verification experiment. The minimal inhibitory concentration (MIC) of RUEP against S.suis was 15.625 µg·mL-1, and sub-inhibitory concentrations of RUEP (1/2, 1/4, 1/8 and 1/16 MIC) significantly inhibited the biofilm formation of S. suis. Moreover, scanning electron microscopy analysis confirmed that RUEP can damage the layered colony community structure in biofilm and effectively intervene in the biofilm formation of Streptococcus suis. Through molecular docking analyzed, the inhibition of biofilm formation could be realized through the formation of hydrogen bond with residues His14 and Glu 60, π-π staking with His61、Phe80 and hydrophobic interaction with Cys127.
Conclusion

Through our study, we found emodin have good antibacterial and antibiofilm formation properties and the potential mechanism of emodin intervention on biofilm formation of Streptococcus suis was inferred.


Results
The optimum extraction conditions (ethanol concentration of 80%, extraction time of 21 min, and liquid to solid ratio of 12.5:1 mL/g) were determined through the interaction analysis of different in uencing factors and model veri cation experiment. The minimal inhibitory concentration (MIC) of RUEP against S.suis was 15.625 µg·mL-1, and sub-inhibitory concentrations of RUEP (1/2, 1/4, 1/8 and 1/16 MIC) signi cantly inhibited the bio lm formation of S. suis. Moreover, scanning electron microscopy analysis con rmed that RUEP can damage the layered colony community structure in bio lm and effectively intervene in the bio lm formation of Streptococcus suis. Through molecular docking analyzed, the inhibition of bio lm formation could be realized through the formation of hydrogen bond with residues His14 and Glu 60, π-π staking with His61 Phe80 and hydrophobic interaction with Cys127.

Conclusion
Through our study, we found emodin have good antibacterial and antibio lm formation properties and the potential mechanism of emodin intervention on bio lm formation of Streptococcus suis was inferred.

Background
Streptococcus suis (S.suis) is a zoonotic pathogen that causes serious systemic infections in pigs and humans [1]. It can cause severe diseases such as meningitis, pneumonia, septicaemia, sepsis and arthritis even death in human [2]. And human can be infected by contacting with pigs or pork products through skin wounds or consumption of raw pork [3]. Since 1968 the rst case of human infection with S.suis has appeared in the Denmark, growing numbers of human cases were reported in many countries. During the 1998 and 2005 Chinese epidemics, a total of 240 people were infected with S.suis, 53 of whom died [4]. At Thailand in 2010, the incidence proportions of the disease were 6.4/10000 persons, and the fatality rate is as high as 16.5% [5]. At Hong Kong, a great number of housewives presumably infected due to contact with (contaminated) pork product [6]. Besides, S.suis is predominantly an occupational hazard in North America and Europe [7,8]. From here we see that S.suis and the diseases it causes do great harm to public health.
Bio lm formation is a unique life phenomenon for bacteria to adapt to the living environment. More than 1/2 all human microbial infections can be associated with bio lms, emphasizing the impact of bio lm structure on public health [9]. Most clinically isolated S. suis strains can form a three-dimensional layered structure (Bio lm), which is surrounded by self-secreted extracellular matrix to ght against the phagocytosis and clearance of the host and the sterilization of antibiotics [10,11]. Bio lm formation is the main reason why S.suis is di cult to cure and recurrent. Bio lm helps S.suis evade the host's defense and resist the intervention of antibiotics. Bio lm formation was described as a process whereby bacteria adhere to the surfaces of materials (living and non-living), secrete polysaccharide matrix, brin, lipoprotein, etc and surround a large number of bacterial aggregate membranes [12]. The protective cover of bio lms prevents antibiotics getting access to target bacteria. In addition, the lack of oxygen supply and the accumulation of metabolites in the bio lm make the bacteria grow slowly and be insensitive to antibiotics, which are the main reasons for the bacterial resistance. The extracellular matrix of bio lm acts as a physical barrier and isolates bacteria from the body's immune system, thus deprives natural killer cells, phagocytes, speci c antibodies, lysozymes, sensitized T cells and other immune effects on the bacteria [13]. At present, erythromycin, azithromycin, tylosin and aspirin are often applied to stop the bio lm formation of S.suis. However, the fact is that antibiotics and nonsteroidal antibacterial that can effectively kill planktonic bacteria are unable to kill bacteria in extracellular membranes of bio lms.
Furthermore, drug residues in pig and human, multiple resistance [14], allergy and gastrointestinal bleeding and other problems would arise after long-term application of common antibiotics. Thus, the discovery and development of bioactive compounds from traditional Chinese medicine has become a hot spot in the eld of against bacteria [15,16].
Typically, the process of bio lms formation is mainly mediated by bacterial communication through the Quorum-sensing (QS) signaling system [17]. The QS system was responsible for the bio lm formation antibiotic resistance group behavior etc [18]. Importantly, QS system could be activated in the condition of chemical signal autoinducer-2 (AI-2) combined with the related acceptor [19]. The LuxS protein was a key enzyme in the process of AI-2 synthesis. And the LuxS/AI-2 mediated QS system play a vital role in bio lm formation of S.suis [20]. Research found that the gene expression of luxS of S.suis was signi cantly decreased with emodin treatment [21],the LuxS protein may be a target for emodin.
R. o cinale, as one of the most traditional medicinal materials, has been used for thousand years. Moreover, R. o cinale has been included in China Pharmacopoeia, British Pharmacopoeia, European Pharmacopoeia, Korean Pharmacopoeia, and Japan Pharmacopoeia due to its extensive antiin ammatory and bacteriostatic effects [22]. Emodin (1,3,8-trihydroxy-6-methylanthraquinone) is the main bioactive component of R. o cinale. Through many investigations, researchers have con rmed the wide spectrum pharmacological effects of emodin, including anti-cancer [23], anti-metastasis [24], reversion of multidrug resistance [25,26], anti-in ammation [27], anti-virus [28], anti-bacteria [29,30], and so on. Furthermore, derivatives obtained by structural modi cation of emodin have also been observed to have bene cial bioactive effects [31,32]. Currently, natural emodin is mainly obtained from Rheum palmatum L., Rheum tanguticum Maxim ex Balf, and R. o cinale. Among them, R. o cinale is the most abundant source with the lowest price than the others. Therefore, it is vital to nd an effective method for extraction of emodin from R. o cinale with high yield to realize and expand its potential applications in the food, nutraceutical and pharmaceutical industry.
Currently, the major extraction methods for bioactive compounds are mainly the conventional ones including decoction, re ux and percolation [33]. The disadvantages of traditional extraction methods are obvious, such as time wasting, energy sapping and low yield [34]. Over the past years, ultrasoundassisted extraction (UAE) has developed rapidly and is gradually maturing. With the advantages of higher product yield, shorter working time and lower cost, UAE is gradually replacing the traditional methods. Besides, UAE is environmentally friendly while achieving high yield, which is in line with the trend of green extraction [35,36]. More importantly, its low energy consumption and cost, satisfy the needs of industrial production of natural products.
A previous study from our laboratory observed that the water extract of R. o cinale can intervene with the bio lm formation of S.suis [37]. However, it is not clear which ingredients from R. o cinale played the major role in this process. In this study, UAE was applied to extract the active ingredient from R. o cinale. And the optimum conditions were investigated by RSM. Besides, high performance liquid chromatographtandem mass spectrometry (HPLC-MS/MS) was carried out to identify the main active ingredient extracted from R. o cinale. Finally, methods of crystal violet staining and scanning electron microscopy (SEM) were used to detect the inhibitory effect of ethanol extract of R. o cinale on the bio lm formation of S.suis. The whole experiment ow is shown in Fig. 1.

Materials And Methods
Bacteria strain and cultural condition S. suis (ATCC 700794) strain was purchased from American Type Culture Collection (ATCC) and maintained in 50% glycerin at -40 °C. The bacteria were cultured at 37 °C in

Single-factor experiments of UAE
In order to explore the best conditions of UAE for emodin, three variables including ultrasonic time, ethanol concentration and liquid to solid ratio were investigated. The ultrasonic time varied from 10 to 30 min; the variation of ethanol concentration was from 50% to 90%; and the liquid to solid ratio was from 5 to 15 (mL/g), respectively. When one of the variables was investigated, the remaining variables were fixed at a certain level.

Experimental design
Based on the results obtained from the single-factor experiments, ethanol concentration (X 1 ), ultrasonic time (X 2 ) and liquid to solid ratio (X 3 ) were the three variables selected and the effect of their interactions on the yield of emodin (Y) was investigated by using RSM.
The three-factor-three-levels Box-Behnken design (BBD) of RSM was carried out by using Design-expert software (version 8.0), and all the independent variables that to be encoded were varied over three levels (-1, 0, 1). As a result, X 1 was varied over a range of 70%-90%. X 2 was varied from 15 to 25 min. Likewise, X 3 was changed from 10 to 15 mL/g. Besides, the codes and its representative values were shown in Table 1 Table 2. The relationship between dependent and independent variables was expressed as a secondorder polynomial equation: All experiments were tested in triplicate and data were analyzed by Design-expert software Under the optimum UAE conditions, the extraction solution was collected, concentrated, lyophilized, and the crude extract was obtained. Hereafter, the crude extract was purified by using a macroporous resin and then stored at -20 °C prior to further analysis. The purified sample was termed "RUEP" and its purity (P) was calculated according to the formula: P (%) = C R V R *100/m R , where in C R and V R were the concentration of emodin (mg/mL) and the volume (mL) of sample solution, respectively, and m R was the mass of frozen sample (mg).
Put Table 2 here Identification of emodin in RUEP by HPLC-MS/MS HPLC-MS/MS was applied to analyze the active ingredient in RUEP. During the process of separation of sample, an ACQUITY UPLC BEH C18 column (100 mm × 2.1 mm, 1.7 mm) Page 9/33 was connected to a HPLC system (SCIEX ExionLC™ AD) and kept at a temperature of 40°C . Importantly, gradient elution was carried out in this experiment. Details were as follows: 0 min, 5% solvent B; 5 min, 95% solvent B; 11 min, 95% solvent B; 12 min, 5% solvent B and 15 min, 5% solvent B. Besides, the mobile phase A and B were 0.1% formic acid and acetonitrile, respectively. The injection volume and flow rate of mobile phase were set at 5 μL and 0.4 mL/min, respectively. The detection system was mass spectrometer Biofilm assay was performed in line with the procedure described previously [40].
Briefly, based on the MIC procedures, the culture time was prolonged to 72 h. The medium was discarded, and then phosphate buffer saline (PBS, pH 7.2) was added into each well to wash away the floating bacteria. After pouring out PBS, 200 μL of 99% methanol was added and the mixture was incubated for 30 min to fix the biofilm. Subsequently, the methanol was poured out, and every well was dyed with 200 μL of crystal violet solution for 30 min. In the end, crystal violet dye solution was discarded and wells were washed with distilled water, and then 200 μL of glacial acetic acid (33%) was added. The optical density value of this mixed solution was measured by using a microtiter plate reader (DG5033A, Huadong, Ltd., Jiangsu, China)

Scanning electron microscopy (SEM) observation
Two pieces of sterilized frosted glass were firstly put into a 6-well microplate, and then the biofilm culture was conducted according to the method mentioned above. One well without adding RUEP was used as a control group, the other well adding RUEP was used as an experimental group. After the biofilm formation of S.suis on frosted glass, the culture medium was discarded and 2.5% of glutaraldehyde was added, then, the mixture was placed in a refrigerator (4 °C) for 1.5 h. At the rinsing stage, 0.1 mol/L of PBS (pH 6.8) was applied three times at an interval of 10 min each. In order to achieve the purpose of dehydration, 70%, 80%, and 90% ethanol was respectively added in experimental units for one time, and then 100% ethanol was added three times at an interval of 10 min each. Next, a mixture of tert-butyl alcohol solution and ethanol (1:1) and a pure tert-butyl alcohol solution were added respectively at an interval of 15 min. The glass was stored in a refrigerator at -20 °C, and then it was pre-frozen for 30 min and dried in a Freeze dryer (ES-2030, HITACHI, Japan) for 4 h. The glass was fixed on the sample table with conductive tape and coated the gold film with a thickness of 100-150 Å using an ion sputtering coater (E1010, HITACHI, Japan). Finally, the observation of S.suis biofilm was carried out under an electron microscopy (S-3400N, HITACHI, Japan).

Molecular Docking
Molecular docking was one of the main methods in Computer-Aided Drug Design. The method was widely applied on discovery of drug target and study of potential mechanism of drug action [41]. The three-dimension structure of LuxS of S.suis was obtained from the Protein Data Bank (https://www.rcsb.org) (PDB ID: 4XCH) [42]. The emodin structure was obtained from Pubchem (https://pubchem.ncbi.nlm.nih.gov) [43]. CDOCKER was selected for docking, which was a docking method performed based on CHARMm. Besides, semi flexible docking is adopted in the docking process, namely, all the residues of protein was fixed and the chemical compound was movable. The best conformation of compound with the lowest energy was selected for further analyzed. The binding model was analyzed and visualized by PyMoL v2.0.6. software [44].

Statistical analysis
All experiments were repeated in triplicate. Statistical analysis was carried out using SPSS 19.0. And p < 0.05 was considered as statistically significant.

Effect of variables on extraction of emodin
The influence of different single-factors on emodin yield was investigated. At singlefactors experiments, only the studied factor was varied at certain range, while other variables were fixed at certain level. The details were as follows: ethanol concentration was 70%, liquid to solid ratio was 10:1 mL/g and the ultrasonic time was 20 min. Besides, the ultrasonic power was kept at 500 W in all experiments. The result from Fig. 2A indicated that emodin yield increased gradually with time prolonged, and reached peaked at 20 min in which the yield was 2.46±0.07 mg/g. But the yield began to decrease as the time prolonged beyond 20 min. As presented in Fig. 2B, the yield was found to increase progressively with increasing ethanol concentration and the high yield of 2.44±0.18 mg/g was obtained at ethanol concentration of 80%. When ethanol concentration was above 80%, the yield began to decrease slowly. As could been seen from Fig. 2C, the yield of emodin was found to increase as liquid to solid ratio increased from 5 to 12.5:1 mL/g. And the yield of emodin reached the highest point (2.28±0.15 mg/g) when the ratio was 12.5:1 mL/g. The amount of emodin happened to decrease as liquid to solid ratio increased beyond 12.5:1 mL/g. Consequently, ultrasonic time of 20 min, ethanol concentration of 80% and liquid to solid ratio of 12.5:1 mL/g were confirmed for subsequent investigation.
Put Figure 2 [45,46]. The matrix designed by BBD and corresponding results of RSM experiments were given in Table 2. Observed from Table 2, the yields of emodin were in the range of 1.79-2.94 mg/g with the change of extraction conditions. Besides, the predicted data and observed data had a close consistence, and there was no statistically significant difference (p>0.05) between them, indicating that the model built by RSM can accurately predicted the yield of emodin. The details of the predicted ANOVA model were shown in Table 3. From the data in Table 3, a second-order polynomial equation about predicted response for emodin yield was obtained by applying regression analysis on the experimental data, which was as follows (Eq. (4)). Importantly, the model adequacy was verified to ensure the accuracy of the results (Fig. 3). After the evaluation of the model, response surface analysis of the results was carried out, response surface graphics and contours plots of different variables was obtained (Fig. 4).
Put Table 3  Besides, three parallel experiments were carried out under these amended conditions. The yield of emodin was 2.77±0.06 mg/g, which was agreed closely to the predicted value of 2.82 mg/g.

Identification of emodin in RUEP
After treatment with macroporous resin, RUEP with the purity of 84.47 ± 0.07% was obtained. The main active substance in RUEP was identified by comparing retention time and mass spectral data with standard based on HPLC-MS/MS analysis (Fig. 5). The compound could be identified as emodin.
Put Figure 5 here

Effect of RUEP on biofilm formation of S.suis
The inhibition ability of biofilm formation was evaluated by MIC assay (Fig. 6A). In The structure of biofilm of S.suis was observed with assistant of SEM. In total two kinds of groups, one is positive control group, one is treated with 1/2 MIC emodin. The experimental result was shown in Fig. 6B. A thick biofilm consisting of aggregates and microcolonies of S.suis almost completely covered the surface of the rough glass slide. A three-dimensional biofilm with multiple layers and aggregates was formed among the communicating S.suis cells. However, after treatment with 1/2 MIC of RUEP (7.8125 μg·mL -1 ), the complete biofilm of S.suis was not existed.

Molecular docking
The molecular docking result processed by PyMOL software and presented in Fig. 7.
The LuxS protein structure of S.suis was shown in Fig. 7A. The blue part was alpha helix and the purple was β-folding, while the pink represented irregular curl. As can be seen in  Table 4 here

Discussion
At investigation of single-factors in uence on emodin yield, the overall trend is similar: With the increase of variables, it rst increases to a certain vertex and then decreases. For in uence of ultrasonic time, on ultrasonic extraction of rosmarinic acid from Orthosiphon stamineus Benth leaves, the similar tendency was found [47]. In general, the ultrasonic wave needs time to crush the cell walls of R. o cinale and extract the cell contents. More importantly, with time increasing, the process through which solvent enters inside the cell is promoted by cavitation and heat, and thus this enables the solvent to penetrate inside the cell and dissolve the target compound [48]. However, the extracted emodin may be degraded and its structure may be destroyed when the time is too long in the ultrasonic surrounding [49]. On effect of ethanol concentration, the effect tendency was consistent with the reports on the extraction of orientin and avonoids [50,51]. On one hand, dielectric constant of extraction reduces with the amount of ethanol increasing, and then solvent solubility and diffusion of target components are improved, which in turn result in the high yield of emodin [52]. On the other hand, the polarity of solvent changes with the adjustment of ethanol concentration. When ethanol concentration was around 80%, the polarity of solvent was similar with emodin and thus maximized the solubility of emodin in solvent and resulted in a higher yield. However, when the concentration of organic solvent is too high in the extracting medium, some unfavorable situation would happen such as dehydration of cells and denaturation on of cell wall proteins, which may inhibit the diffusion of the target compounds in solvent and cause a reduction in the yield of target compounds. In addition, in the presence of small amounts of water, plant materials expand more effectively, increasing the surface area of contact between the plant substrate and the solvent [53].
Finally, the effect of liquid to solid ratio is discussed. At a very low liquid to solid ratio, the volume of the extracting solvent is so small that it is unable to completely extract the target compounds from the plant [54]. With increase in the concentration of the extraction solvent, there will be a high concentration difference between the solvent and plant materials and further lead to the release of the target compound. But as the ratio was beyond 12.5:1 mL/g, emodin yield decreased, which may be due to that the transfer of ultrasonic energy was compromised in the excessive extraction solvent. Meanwhile, the dissolved quantity of other soluble matter from plant cell would increase [55].
As could be seen from Table 3, the model had an F-value of 25.57 with a very low probability (p = 0.0002), suggesting that the model was signi cant. There was only a 0.01% chance that an F-value this large could occur due to chance. Parameters such as the determination co-e cient (R 2 ), adjusted determination co-e cient (R 2 adj ) and co-e cient of variance (C.V%) were considered as the indices to evaluate the applicability of the model [56]. The value of R 2 was 0.9705, indicating the powerful correlation among the anticipated and experimental values. Besides, the value of R 2 adj was 0.9325 indicated the experimental values could be expected by using the model considerably. Furthermore, the predicted determination coe cient value (R 2 pre , 0.8188) was in reasonable agreement with the value of R 2 adj . Lower value of C.V% (4.78) suggested experimental values and predicted values had subtle deviations, which also indicated a high degree precision and reproducibility of the model [57]. Moreover, the adequacy of the built model to completely explain the experimental data depends on the value of lack-of-t. The F-value of lack of t was 1.64 (p = 0.3148), proving it was not signi cant compare to the pure error and good tting degree of the regression model [58]. Besides, the coe cients of terms including X 2 , X 1 X 2 , X 2 X 3 , X 1 2 , X 2 2 and X 3 2 were signi cant for the mathematical model with p-value less than 0.05. By contrast, p-value greater than 0.05 means the model term was non-signi cant, such as X 1 , X 3 and X 1 X 3 . As a result, only the terms with signi cant coe cient were applied for model building (shown in Eq. (4)). Moreover, the model had a maxima point for yield, namely, the predicted value, due to the quadratic coe cients of X 1 2 , X 2 2 and X 3 2 were negative. A full analysis was carried out on F-values of the items in the regression model, the order of signi cance for variable that in uenced the yield of emodin was: ultrasonic time (X 2 ) > ethanol concentration (X 1 ) > liquid to solid ratio (X 3 ).
The model adequacy was veri ed by the kinds of diagnostic plots such as predicted versus actual, normal % probability, and internally studentized residuals, which were vital for checking the accuracy and reliability of the model. The plot of normal % probability was shown in Fig. 3A. As observed, the distribution was closed to the straight line, meaning that there was no deviation of the variance and normal distribution was observed. From analysis in Fig. 3B and Fig. 3C (the plot of internally studentized residuals), all data was in a reasonable range. Besides, as could been seen from Fig. 3D, the data was close to the straight line, proving the predicted value from the model was adequate and had a high correlation with the experimental data. Consequently, the model built by RSM was reliable for the estimation of UAE for emodin from R. o cinale.
To investigate the interaction of these three test variables on the yield of emodin, response surface was plotted by varying two of the test variables while keeping the third factor xed [59]. The steeper the response surface slope is and the closer the contour is to the ellipse, the more sensitive the response value is to the variety of factors [60]. The response surface result can be visualized by the two contours plots and the three-dimensional (3D) plots. The effect of the interaction between ethanol concentration (X 1 ) and ultrasonic time (X 2 ) on the yield of emodin was shown in Fig. 4A and Fig. 4D while the liquid to solid ratio (X 3 ) was xed at 12.5:1 mg/L. It could be found that with increasing ethanol concentration from 70-80% and ultrasonic time from 15 to 20 min, there was a concomitant increase in the yield of emodin, while it was observed that the yield gradually decreased when these two test variables increased further. And their interaction had obvious impact on the yield because of the contours plot tended to the ellipse, which was agree with the signi cance analysis result of X 1 X 2 from Table 3. Besides, the ultrasonic time had a signi cant effect on the yield due to the steep slope. Figure 4B and Fig. 4F represented the effect of ethanol concentration (X 1 ) and liquid to solid ratio (X 3 ) on the yield of emodin when ultrasonic time (X 2 ) was xed at 20 min. The optimum yield of emodin was obtained when ethanol concentration and liquid to solid ratio were extremely close to center point, in which ethanol concentration was 80% and liquid to solid ratio was 12.5:1 mL/g. The contours plot was close to circle con rmed their interaction was not signi cant. As ethanol concentration (X 1 ) was xed at 80%, the three-dimension plot and contours of interaction between ultrasonic time (X 2 ) and liquid to solid ratio (X 3 ) was presented in Fig. 4C and Fig. 4G.
A similar change in emodin yield was observed with these two variables. The maximum yield of emodin was obtained when ultrasonic time and liquid to solid ratio were approximate 20 min and 12.5:1 mg/L, respectively. From Table 3, the interaction of X 2 X 3 on yield was signi cant and with the existence of ellipse contours plot, indicating a strong interaction existed between ultrasonic time and liquid to solid ratio. This showed the interaction between these two variables can affect the yield of emodin dramatically. However, the in uence of ultrasonic time was a little higher than liquid to solid ratio because of its steeper slopes.
Through comparing the experiment result with the predicted, proving the accuracy and effectiveness of the model. Moreover, the yield of emodin was improved than other extraction methods. Zhou's et al obtained the emodin yield of 2.53 ± 0.09 mg/g with assistance of microwave [61]. By using ultrasonic nebulization extraction method, researchers obtained the emodin yield of 2.04 mg/g [62]. Besides, Wu's et al extracted emodin from Rheum palmatum L, the yield of 2.21 ± 0.02 mg/g under optimum ultrasonicassisted conditions [63]. Consequently, it could be considered that this optimum condition could extract emodin from R. o cinale effectively.
In the process of identi cation, as could be seen from Fig. 5A, emodin standard was eluted at 5.053 min, with a [M + H] + at m/z 269.0460. Next, the molecular ion with m/z 269.0460 was selected as parent ion to perform a secondary scan, and the secondary mass spectrum of emodin was obtained in Fig. 5B. It mainly produced MS 2 fragment ion at m/z 225.0572 and m/z 241.0505. Under the same condition, the elution time of the constituent in RUEP was 5.051 min, the molecular cation from this active ingredient was detected at m/z 269.0458 (Fig. 5C). This result was consistent with Chen et al' s report [64]. Besides, it mainly produced MS 2 fragment ion at m/z 225.0569 and m/z 241.0503 through the secondary scan (Fig. 3D). Through analysis the primary and secondary mass spectra, the data from emodin standard and the main active ingredient in RUEP was basically same. Importantly, the application of secondary mass spectrometry ensured the accuracy of the results. Therefore, the main active ingredient in RUEP was identi ed as emodin.
Through analyzed molecular docking result, it was indicating that the emodin could combine with the LuxS protein through molecular docking. From analyzed the force between the ligand and residues, the shape of hydrophobic interaction would discharge the water molecules in the binding pocket, increasing the entropy of the system. Binding a nity of emodin to Luxs Protein was increased due to water molecules in hydrophobic environment were released. Besides, the formation of hydrogen bond leads their bond stronger. Among the above residues, His61 and Cys127 was highly conserved residues for binding in LuxS protein no matter in S.suis or other species, including Enterococcus facecium, H. pylori, Vibrio cholerae, Vibrio harveyi, Escherischia coli, Salmonella enterica, Neisseria meningitidis, Actinobacillus pleuropneumoniae etc [65]. In addition, Phe80 was an important substrate binding residue in S.suis LuxS protein that differ from other species. It's responsible for the enzymatic activity, the production of AI-2 and the ability of bio lm formation [66]. Consequently, it could be considered that emodin could be well combine with LuxS protein of S.suis. It's consistent with induced t theory. We speculated the emodin would combine with LuxS protein at the active site, changing the conformation, in uencing the function of the protein. When the ligand-protein complex formed, would decrease the production of AI-2 and inhibited the ability of bio lm formation. However, this is one of potential mechanism that emodin intervene the bio lm formation of S.suis, further exploration about the intervene mechanism needs to carried out.

Conclusion
In this study, ultrasonic-assisted extraction of emodin from R. o cinale was optimized using RSM. The yield of emodin was 2.77 ± 0.06 mg/g under the optimum conditions, which was agreed closely to the predicted value of 2.82 mg/g. Ultrasonic time signi cantly in uenced the yield of emodin, followed by ethanol concentration and liquid to solid ratio. The main active ingredient in RUEP extracted from R. o cinale was identi ed as emodin by HPLC-MS/MS. Moreover, RUEP exhibited potent antibacterial property and inhibitory ability on bio lm formation of S.suis. The inhibition of bio lm formation could be realized through the formation of hydrogen bond with residues His14 and Glu60, π-π stacking with residues His61 Phe80 and hydrophobic interaction with residue Cys127. Therefore, RUEP from R. o cinale may have potential value in the preparation of veterinary drugs and feed additives.   Figure 1 The graphic of speci c experimental procedure.