Rhamnose Binding Protein as an Anti-Bacterial Agent—Targeting Biofilm of Pseudomonas aeruginosa

More than 80% of infectious bacteria form biofilm, which is a bacterial cell community surrounded by secreted polysaccharides, proteins and glycolipids. Such bacterial superstructure increases resistance to antimicrobials and host defenses. Thus, to control these biofilm-forming pathogenic bacteria requires antimicrobial agents with novel mechanisms or properties. Pseudomonas aeruginosa, a Gram-negative opportunistic nosocomial pathogen, is a model strain to study biofilm development and correlation between biofilm formation and infection. In this study, a recombinant hemolymph plasma lectin (rHPLOE) cloned from Taiwanese Tachypleus tridentatus was expressed in an Escherichia coli system. This rHPLOE was shown to have the following properties: (1) Binding to P. aeruginosa PA14 biofilm through a unique molecular interaction with rhamnose-containing moieties on bacteria, leading to reduction of extracellular di-rhamnolipid (a biofilm regulator); (2) decreasing downstream quorum sensing factors, and inhibiting biofilm formation; (3) dispersing the mature biofilm of P. aeruginosa PA14 to improve the efficacies of antibiotics; (4) reducing P. aeruginosa PA14 cytotoxicity to human lung epithelial cells in vitro and (5) inhibiting P. aeruginosa PA14 infection of zebrafish embryos in vivo. Taken together, rHPLOE serves as an anti-biofilm agent with a novel mechanism of recognizing rhamnose moieties in lipopolysaccharides, di-rhamnolipid and structural polysaccharides (Psl) in biofilms. Thus rHPLOE links glycan-recognition to novel anti-biofilm strategies against pathogenic bacteria.


Introduction
Part of the reasons for bacterial resistance to antibiotics are due to the formation of biofilm, composed of secreted polysaccharides, proteins, glycolipids and small molecules in the bacterial microenvironment. Recent development of antimicrobial agents with novel molecular mechanisms to control bacterial infectious diseases has drawn much attention [1]. The inhibition of bacterial biofilm An RBP possessing specific binding of l-rhamnose was discovered from horseshoe crab plasma and recently engineered for expression in the Escherichia coli system [29]. This recombinant horseshoe crab plasma lectin (rHPL) possesses a very low sequence identity with known RBPs and does not have conserved domains. Interestingly, rHPL binds to bacteria or pathogen-associated molecular patterns (PAMPs) by recognizing rhamnose moieties and inhibits the growth of P. aeruginosa [29]. Unlike other RBLs, rHPL only binds to l-rhamnose and rhamnobiose but not to galactose or mannose [30]. This high substrate specificity makes rHPL a prospective candidate to bind rhamnose-containing components in biofilm of P. aeruginosa and examine the biological functions of these bindings.

rHPL OE Was Expressed in E. coli and Purified by Affinity Chromatography
rHPL was successfully expressed in E. coli in 2014. The yield of rHPL purified using a nickel-affinity column was~8 mg/L, and the purity was 93% [29]. To improve the productivity and solubility of rHPL, the codon usage of synthetic hpl (hploe) was optimized for E. coli, and recombinant HPL OE was co-expressed with chaperones. The expression product, rHPL OE , was purified by fast protein liquid chromatography (FPLC) equipped with a HisTrap™ affinity column following purification scheme in supplementary data showed in Table S1. rHPL DM , representing rHPL with two artificial mutations Y88A and F145A, which showed no binding activities to a bacterial cell-( Figure S1A) or pathogen-associated molecular pattern (PAMP; Figure S1B), was applied as a negative control in this study. Figure S2 illustrated the purification chromatogram and SDS polyacrylamide gel electrophoresis analysis data. The purified rHPL OE was desalted and concentrated using an Amicon protein concentrator (10 kDa cut-off) and subjected to further analysis. Figure S3 showed the secondary structure of rHPL OE determined by circular dichroism (CD) spectroscopy in comparison to those of rHPL and rHPL DM , indicating that both rHPL OE and rHPL DM possessed similar structures to rHPL [29]. A yield of 11.33 mg/L rHPL OE with a comparable purity. The yield of rHPL OE was 1.4-fold higher than that of rHPL, suggesting that the chaperone improved the solubility and stability of rHPL in the E. coli system.

rHPL OE Bound to Cell-Free Biofilm Matrix from P. aeruginosa PA14 via Recognizing Rhamnose
Our hypothesis is that rHPL OE might bind to rhamnose-containing components in the biofilm and further interrupt bacterial biofilm development. First, binding capacity of rHPL OE to the cell-free biofilm matrix from P. aeruginosa PA14 was tested. Here, a mature P. aeruginosa PA14 biofilm was extracted using a NaCl solution to give a cell-free biofilm matrix. In the extracted PA14 biofilm, the total protein was 0.19 mg/mL and polysaccharide, 0.32 mg/mL. Di-rhamnolipid, an important QS-factor and putative binding target of rHPL OE , was extracted using chloroform, and was 0.66 µg/mL by methylene blue method. The interaction between rHPL OE and the cell-free biofilm matrix was measured by ELISA, and the inhibitory effect of monosaccharides and alginate on rHPL OE -biofilm interaction was determined using competitive ELISA. rHPL OE at a final concentration of 0.1 µM was mixed with l-rhamnose and loaded into microplate wells coated separately with extracted P. aeruginosa PA14 biofilm. rHPL OE mixed with buffer only was used as a positive control. rHPL DM without rhamnose binding activity served as a negative control. As shown in Figure 1, addition of 25, 50, and 100 mM of l-rhamnose effectively reduced binding of rHPL OE to the extracted biofilm, in comparison with the positive control. The inhibitory constant (K i ) of l-rhamnose to rHPL OE interaction with the biofilm of P. aeruginosa PA14 was 98 mM. The addition of 100 Mm of d-mannose, d-glucose, d-fructose, d-galactose and alginate could not reduce the binding between rHPL OE and the extracted biofilm. These results indicated that rHPL OE specifically bound to rhamnose containing components in P. aeruginosa PA14 biofilm.

Figure 1.
Inhibitory effects of monosaccharides on the recombinant hemolymph plasma lectin (rHPLOE)-biofilm interaction. The binding activity of rHPLOE to the PA14 biofilm and inhibitory effects of monosaccharides or alginate on this binding were determined by competitive ELISA. rHPLDM was applied as a negative control. *** p < 0.001 versus the rHPLOE only group (positive control).

Synthetic Rhamnobiosides and a Pentasaccharide Inhibited rHPLOE-Biofilm Interaction
To confirm that binding of rHPLOE to the P. aeruginosa PA14 biofilm was via molecular recognition of rhamnose moiety, synthetic rhamno-oligosaccharides (Figure 2A,B) were used as competitors in a competitive ELISA. The chemical synthesis of phenylthio-rhamnobiosides and the Psl-pentasaccharide will be published elsewhere. The characterization of the new oligosaccharides used in this study is given in the Supplementary Information ( Figure S4-S6).

Figure 1.
Inhibitory effects of monosaccharides on the recombinant hemolymph plasma lectin (rHPL OE )-biofilm interaction. The binding activity of rHPL OE to the PA14 biofilm and inhibitory effects of monosaccharides or alginate on this binding were determined by competitive ELISA. rHPL DM was applied as a negative control. *** p < 0.001 versus the rHPL OE only group (positive control).

Synthetic Rhamnobiosides and a Pentasaccharide Inhibited rHPL OE -Biofilm Interaction
To confirm that binding of rHPL OE to the P. aeruginosa PA14 biofilm was via molecular recognition of rhamnose moiety, synthetic rhamno-oligosaccharides (Figure 2A,B) were used as competitors in a competitive ELISA. The chemical synthesis of phenylthio-rhamnobiosides and the Psl-pentasaccharide will be published elsewhere. The characterization of the new oligosaccharides used in this study is given in the Supplementary Information (Figures S4-S6). Figure 2 indicated that all synthetic compounds competed with the rHPL OE -biofilm interaction. The disaccharides are better inhibitor than monosaccharides, and (1-3)-linked rhamnosides were better than its (1-2)-linked counterparts. Psl-pentasaccharide showed the strongest competition presumably due to longer glycan chain length. The competitive effects of rhamnobiosides were stronger than those of rhamnobioses, indicating either a fixed anomeric configuration or hydrophobicity of aglycones slightly favors binding.

rHPL OE Bound to Di-Rhamnolipid via Recognizing Rhamnose Portion
After confirming the binding activity of rHPL OE to synthetic rhamnobiosides, direct binding activity of rHPL OE to di-rhamnolipid from P. aeruginosa PA14 biofilm was measured by ELISA. Further competitive ELISA indicated that rHPL OE binding to di-rhamnolipid could be competed more effectively by l-rhamnose compared to other monosaccharides ( Figure 3).

rHPL OE Inhibited Biofilm Formation and Dispersed the Preformed Biofilm of P. aeruginosa PA14
This experiment was conducted in M63 broth, a minimal low-osmolarity medium to slow down the growth rate of bacteria and mimic nutrient-depleted conditions for biofilm growth [31]. rHPL DM was applied as a negative control. As shown in Figure 4A, 0.31 µM and higher concentrations up to 5 µM of rHPL OE significantly inhibited the biofilm formation of P. aeruginosa PA14. Treatment with 0.31, 0.63, 1.25, 2.5 and 5 µM rHPL OE significantly reduced biofilm formation to 49%, 51%, 39%, 18% and 13%, respectively. As expected, rHPL DM did not inhibit biofilm formation. rHPL OE was also evaluated for its biofilm dispersion activity with preformed biofilm. The mature biofilm of P. aeruginosa PA14 was dispersed by 16% and 24% upon treatment with 2.5 and 5 µM rHPL OE for 24 h, respectively ( Figure 4B). However, rHPL DM did not disperse the mature biofilm. These results indicated that rHPL OE inhibited biofilm formation and dispersed the mature biofilm of P. aeruginosa PA14 and that both activities perhaps are correlated with the rhamnose binding activity.

rHPLOE Bound to Di-Rhamnolipid via Recognizing Rhamnose Portion
After confirming the binding activity of rHPLOE to synthetic rhamnobiosides, direct binding activity of rHPLOE to di-rhamnolipid from P. aeruginosa PA14 biofilm was measured by ELISA. Further competitive ELISA indicated that rHPLOE binding to di-rhamnolipid could be competed more effectively by L-rhamnose compared to other monosaccharides ( Figure 3).

rHPLOE Inhibited Biofilm Formation and Dispersed the Preformed Biofilm of P. aeruginosa PA14
This experiment was conducted in M63 broth, a minimal low-osmolarity medium to slow down the growth rate of bacteria and mimic nutrient-depleted conditions for biofilm growth [31]. rHPLDM was applied as a negative control. As shown in Figure 4A, 0.31 μM and higher concentrations up to 5 μM of rHPLOE significantly inhibited the biofilm formation of P. aeruginosa PA14. Treatment with 0.31, 0.63, 1.25, 2.5 and 5 μM rHPLOE significantly reduced biofilm formation to 49%, 51%, 39%, 18% and 13%, respectively. As expected, rHPLDM did not inhibit biofilm formation. rHPLOE was also evaluated for its biofilm dispersion activity with preformed biofilm. The mature biofilm of P. aeruginosa PA14 was dispersed by 16% and 24% upon treatment with 2.5 and 5 μM rHPLOE for 24 h, respectively ( Figure 4B). However, rHPLDM did not disperse the mature biofilm. These results indicated that rHPLOE inhibited biofilm formation and dispersed the mature biofilm of P. aeruginosa PA14 and that both activities perhaps are correlated with the rhamnose binding activity.

rHPLOE Inhibited Swarming Activity and Decreased Secreted Rhamnolipids of P. aeruginosa PA14
Swarming motility is positively correlated with the amount of extracellular rhamnolipids, leading to rapid bacterial translocation that promotes efficient colonization of bacterial cells on a surface [32]. Under optimal growth conditions, P. aeruginosa cell population would spread along a solid surface and cover a large area. If swarming were inhibited, the bacterial cell population would gather together to form a colony. The swarming area of P. aeruginosa PA14 was significantly reduced upon treatment with rHPLOE but not rHPLDM ( Figure S7A). The swarming area of PA14 was reduced by 33%, 43%, 48% and 46%, respectively, upon treatment with 0.63, 1.25, 2.5 and 5 μM rHPLOE for 72 h ( Figure S7B). Figure 5 showed that the extracellular rhamnolipid levels of PA14 (11.7 ± 0.4 μg/mL in the control treated with buffer) reduced by 44%, 50%, 54%, 64% and 59% upon treatment with 0.31, 0.63, 1.25, 2.5 and 5 μM rHPLOE, respectively. However, the 5 μM rHPLDM treatment did not affect the extracellular rhamnolipid levels of PA14. This result was consistent with the swarming assay, indicating that the binding of rHPLOE to biofilm inhibited PA14 swarming activity by down-regulating the levels of extracellular di-rhamnolipid. . Inhibitory and dispersion effect on the biofilm of P. aeruginosa PA14 by rHPL OE . Quantification of crystal violet staining associated with (A) the biofilm of P. aeruginosa PA14 after treatment with rHPL OE at the indicated concentration for 24 h and (B) after biofilm formation for 24 h and treatment with rHPL OE at the indicated concentration for a further 24 h. rHPL DM was applied as a negative control. The buffer-treated group was set as 100% (mock). * p < 0.05 and *** p < 0.001 versus the buffer-treated group.

rHPLOE Inhibited Swarming Activity and Decreased Secreted Rhamnolipids of P. aeruginosa PA14
Swarming motility is positively correlated with the amount of extracellular rhamnolipids, leading to rapid bacterial translocation that promotes efficient colonization of bacterial cells on a surface [32]. Under optimal growth conditions, P. aeruginosa cell population would spread along a solid surface and cover a large area. If swarming were inhibited, the bacterial cell population would gather together to form a colony. The swarming area of P. aeruginosa PA14 was significantly reduced upon treatment with rHPL OE but not rHPL DM ( Figure S7A). The swarming area of PA14 was reduced by 33%, 43%, 48% and 46%, respectively, upon treatment with 0.63, 1.25, 2.5 and 5 µM rHPL OE for 72 h ( Figure S7B). Figure 5 showed that the extracellular rhamnolipid levels of PA14 (11.7 ± 0.4 µg/mL in the control treated with buffer) reduced by 44%, 50%, 54%, 64% and 59% upon treatment with 0.31, 0.63, 1.25, 2.5 and 5 µM rHPL OE , respectively. However, the 5 µM rHPL DM treatment did not affect the extracellular rhamnolipid levels of PA14. This result was consistent with the swarming assay, indicating that the binding of rHPL OE to biofilm inhibited PA14 swarming activity by down-regulating the levels of extracellular di-rhamnolipid.
leading to rapid bacterial translocation that promotes efficient colonization of bacterial cells on a surface [32]. Under optimal growth conditions, P. aeruginosa cell population would spread along a solid surface and cover a large area. If swarming were inhibited, the bacterial cell population would gather together to form a colony. The swarming area of P. aeruginosa PA14 was significantly reduced upon treatment with rHPLOE but not rHPLDM ( Figure S7A). The swarming area of PA14 was reduced by 33%, 43%, 48% and 46%, respectively, upon treatment with 0.63, 1.25, 2.5 and 5 μM rHPLOE for 72 h ( Figure S7B). Figure 5 showed that the extracellular rhamnolipid levels of PA14 (11.7 ± 0.4 μg/mL in the control treated with buffer) reduced by 44%, 50%, 54%, 64% and 59% upon treatment with 0.31, 0.63, 1.25, 2.5 and 5 μM rHPLOE, respectively. However, the 5 μM rHPLDM treatment did not affect the extracellular rhamnolipid levels of PA14. This result was consistent with the swarming assay, indicating that the binding of rHPLOE to biofilm inhibited PA14 swarming activity by down-regulating the levels of extracellular di-rhamnolipid.  Down-regulation effect on extracellular rhamnolipids of P. aeruginosa PA14 by rHPL OE or rHPL DM . Extracellular rhamnolipids of P. aeruginosa PA14 treated by rHPL OE or rHPL DM were measured by the chloroform-methyl blue method. *** p < 0.001 versus the buffer-treated group.

rHPL OE Attenuated QS-Associated Factors
The development of biofilms appears to be regulated by the QS system. Most reagents that attenuate biofilm development work through inhibiting production of QS-factors or disrupting function of QS-factors by direct binding to QS-factors [33][34][35]. Among these agents, pyoverdine and pyocyanin are unique toxic molecules related to the virulence of P. aeruginosa PA14. To understand the attenuation mechanism of P. aeruginosa biofilm by rHPL OE , the effect of rHPL OE on QS-mediated azocasein-degrading protease activity and the secretion of two virulent factors, pyoverdine and pyocyanin were evaluated. Bacterial extracellular proteases degrade proteins in host cells (infected tissue), thereby facilitating bacterial invasion and growth [36]. Interestingly, bacterial protease secretion by P. aeruginosa PA14 was down-regulated by rHPL OE treatment ( Figure S8A). The azocasein-degrading assay showed that 1.25 and 2.5 µM rHPL OE respectively decreased 16% and 12% of total proteolytic activities. Treatment with 0.31, 0.63, 1.25 and 2.5 µM rHPL OE decreased the pyoverdine levels by 12%, 8%, 17% and 21%, respectively ( Figure S8B). Treatment with the same concentrations of rHPL OE decreased the pyocyanin levels by 41%, 56%, 45% and 40%, respectively ( Figure S8C).
These results indicated that the binding of rHPL OE to the biofilm down-regulated the protease activities and QS-factors (Table 1). Consistent with previous observations, 2.5 µM rHPL DM did not show an inhibitory effect. Relatively weak inhibition of rHPL OE on protease activity and pyoverdine expression indicated that these effects might be indirect effects. The effect of rHPL OE on pyocyanin production was detectable but not concentration-dependent. A possible alternate reason for this may be that the inhibition effect of rHPL OE on pyocyanin production was compensated by other regulatory pathways.

Combination Treatment with rHPL OE Improved Efficacies of Antibiotics on P. aeruginosa with Preformed Biofilms
Since rHPL OE exhibited dispersion of P. aeruginosa PA14 biofilm, we hypothesized that rHPL OE might increase the activities of antibiotics against P. aeruginosa PA14 in mature biofilms. Total protein assay showed that the combination of rHPL OE with either azithromycin (hydrophobic) or cephalexin (hydrophilic) significantly attenuated the total protein levels in the biofilm. The IC50 of azithromycin and cephalexin to P. aeruginosa PA14 with pre-formed biofilm was 27.3 ± 1.4 and 27.7 ± 0.9 µg/mL, respectively. In the presence of 25 µg/mL azithromycin, in combination with 0.31, 0.63, 1.25 and 2.5 µM rHPL OE reduced the total protein in the biofilm by 19%, 21%, 39% and 43%, respectively, compared to the antibiotic-only control ( Figure 6).  In the presence of 25 μg/mL cephalexin, combination treatment with the same concentrations of rHPLOE reduced the total protein in biofilms by 33%, 33%, 38% and 50%, respectively, compared to the antibiotic-only control ( Figure 6). This observation suggested that rHPLOE might facilitate antibiotics to kill P. aeruginosa PA14 by partially interfering with biofilm regulation and thus destroying the structure of the mature biofilm.

rHPLOE Inhibited Infection of P. aeruginosa in Mammalian Cells and Zebrafish Embryos
During P. aeruginosa infection, rhamnolipids and pyocyanin are the important virulence factors causing cytotoxicity. Since rHPLOE inhibited rhamnolipid and pyocyanin production in P. aeruginosa PA14, we predicted that rHPLOE could also reduce P. aeruginosa PA14 infection inhuman lung cells. To evaluate this hypothesis, cell death of human lung cell line A549 by P. aeruginosa PA14 infection in the presence of rHPLOE was examined. As shown in Figure 7, the PA14-infected A549 cell death was reduced by 14%, 21%, 57% and 73% in the presence of 0.31, 0.63, 1.25 and 2.5 μM rHPLOE, respectively. In the presence of 25 µg/mL cephalexin, combination treatment with the same concentrations of rHPL OE reduced the total protein in biofilms by 33%, 33%, 38% and 50%, respectively, compared to the antibiotic-only control ( Figure 6). This observation suggested that rHPL OE might facilitate antibiotics to kill P. aeruginosa PA14 by partially interfering with biofilm regulation and thus destroying the structure of the mature biofilm.

rHPL OE Inhibited Infection of P. aeruginosa in Mammalian Cells and Zebrafish Embryos
During P. aeruginosa infection, rhamnolipids and pyocyanin are the important virulence factors causing cytotoxicity. Since rHPL OE inhibited rhamnolipid and pyocyanin production in P. aeruginosa PA14, we predicted that rHPL OE could also reduce P. aeruginosa PA14 infection inhuman lung cells. To evaluate this hypothesis, cell death of human lung cell line A549 by P. aeruginosa PA14 infection in the presence of rHPL OE was examined. As shown in Figure 7, the PA14-infected A549 cell death was reduced by 14%, 21%, 57% and 73% in the presence of 0.31, 0.63, 1.25 and 2.5 µM rHPL OE , respectively.
Regarding the negative control, 2.5 µM BSA or 2.5 µM rHPL DM did not influence A549 cell death. These assays clearly demonstrated that the mortality rate of A549 cells killed by P. aeruginosa PA14 decreased due to the presence of rHPL OE . Therefore, rHPL OE attenuated P. aeruginosa PA14 infection in human lung A549 cells through direct interaction between rHPL OE and rhamnose-containing bacterial compounds. causing cytotoxicity. Since rHPLOE inhibited rhamnolipid and pyocyanin production in P. aeruginosa PA14, we predicted that rHPLOE could also reduce P. aeruginosa PA14 infection inhuman lung cells. To evaluate this hypothesis, cell death of human lung cell line A549 by P. aeruginosa PA14 infection in the presence of rHPLOE was examined. As shown in Figure 7, the PA14-infected A549 cell death was reduced by 14%, 21%, 57% and 73% in the presence of 0.31, 0.63, 1.25 and 2.5 μM rHPLOE, respectively.  Zebrafish (Danio rerio) has become an important vertebrate animal model for many disease studies including pathogen infection [37] in recent years. Here fertilized eggs from zebrafish were used as an animal model to evaluate whether rHPL OE attenuated P. aeruginosa PA14 infection in vivo. Zebrafish embryos showed normal development in the absence of P. aeruginosa PA14, while the development of most embryos was significantly delayed or even died at 48 h following infection with P. aeruginosa PA14 (Figure 8). Regarding the negative control, 2.5 μM BSA or 2.5 μM rHPLDM did not influence A549 cell death. These assays clearly demonstrated that the mortality rate of A549 cells killed by P. aeruginosa PA14 decreased due to the presence of rHPLOE. Therefore, rHPLOE attenuated P. aeruginosa PA14 infection in human lung A549 cells through direct interaction between rHPLOE and rhamnose-containing bacterial compounds.
Zebrafish (Danio rerio) has become an important vertebrate animal model for many disease studies including pathogen infection [37] in recent years. Here fertilized eggs from zebrafish were used as an animal model to evaluate whether rHPLOE attenuated P. aeruginosa PA14 infection in vivo. Zebrafish embryos showed normal development in the absence of P. aeruginosa PA14, while the development of most embryos was significantly delayed or even died at 48 h following infection with P. aeruginosa PA14 (Figure 8). However, when P. aeruginosa PA14 was pre-incubated with 5 μM rHPLOE, 82.2% of embryos developed normally as compared to control embryos showing a development rate of 90.9%. Zebrafish embryos treated with 2.5 μM rHPLOE still developed slowly, but almost all of the embryos survived. These results indicated that rHPLOE could protect zebrafish embryos from P. aeruginosa However, when P. aeruginosa PA14 was pre-incubated with 5 µM rHPL OE , 82.2% of embryos developed normally as compared to control embryos showing a development rate of 90.9%. Zebrafish embryos treated with 2.5 µM rHPL OE still developed slowly, but almost all of the embryos survived. These results indicated that rHPL OE could protect zebrafish embryos from P. aeruginosa PA14 infection at concentrations higher than 2.5 µM.

Discussion
To study the carbohydrate components of the biofilm, carbohydrate-binding proteins, such as plant lectins, have been widely applied. Previously many different neutral carbohydrates including N-acetyl-d-glucosamine, N-acetyl-d-galactosamine, d-glucose and d-mannose were identified in the exopolymeric matrix of the biofilm [38]. Lectins with various specificities show an interaction with the carbohydrate components of biofilms. For example, concanavalin A from Canavalia ensiformis seeds is specific for d-glucose and d-mannose [39], and lectin from Triticum vulgaris germs is specific for N-acetyl-d-glucosamine and sialic acid [40]. l-Rhamnose as a unique sugar in bacteria and plants is a common component of the cell wall and capsule of many pathogenic bacteria including Gram-negative P. aeruginosa [41], Salmonella typhimurium [42], Vibrio cholera [43,44] as well as Mycobacterium tuberculosis [45]. P. aeruginosa produces at least three distinct exopolysaccharides that contribute to biofilm development and architecture: Alginate, Pel and Psl [46]. Alginate consists of only uronic acids, Pel is a glucose-rich polysaccharide and Psl consists of a repeating pentasaccharide containing d-mannose, d-glucose and l-rhamnose [47]. Psl serves as a primary structural scaffold for biofilm development. In addition, Psl is involved in early stages of biofilm formation and promoting cell-to-cell interactions [48,49]. Another rhamnose-containing component in the biofilm was the QS-factor di-rhamnolipid. A mutant strain (P. aeruginosa PAO1C-rhlAB) that could not produce rhamnolipids lost its swarming activity, which agreed with our observation in rHPL OE -treated PA14 [50]. rHPL OE bound to both Psl and di-rhamnolipid through targeting rhamnose moiety and synthetic rhamnobiosides-competitive ELISA showed that rHPL OE preferred to bind with α(1-3)-rhamnobiose rather than α(1-2)-rhamnobiose. The linkage of the rhamnose of Psl was reported to be α(1-3) [47] while that of di-rhamnolipid was α(1-2) [51], perhaps favoring that rHPL OE binding to rhamnose component in biofilm was governed by recognition of unique structure feature of rhamnosyl moiety. Although rHPL OE binding to Psl still required more abundant glycan for detail characterization, rHPL OE binding to di-rhamnolipid down-regulated the expression of di-rhamnolipid and QS-factors of P. aeruginosa and further inhibited biofilm formation.
The most important finding of this study was that the binding of rHPL OE to the P. aeruginosa PA14 biofilm inhibited biofilm development and disrupted the mature biofilm. Our data clearly showed that rHPL OE caused these bio-effects by interacting with components in the biofilm including structural polysaccharides or di-rhamnolipid, largely due to the down-regulation of QS-factors including di-rhamnolipid, pyocyanin, pyoverdine and extracellular proteases by interrupting the functions of di-rhamnolipid.
Based on the biofilm dispersion activity of rHPL OE , we proposed that rHPL OE possessed synergistic effects with antibiotics on P. aeruginosa. Two antibiotics commonly applied for treating P. aeruginosa infections, azithromycin, a hydrophobic azalide that kills bacteria by decreasing protein production and cephalexin, a hydrophilic beta-lactam that kills bacteria by inhibiting cell wall synthesis, were used in this study in combination with rHPL OE . The results showed that rHPL OE significantly improved the bactericidal activity of both antibiotics, strongly suggesting that rHPL OE was useful as a biofilm dispersion reagent for deconstructing the biofilm and improving the activities of antibiotics.
Many studies reported that P. aeruginosa infection could be inhibited by down-regulating QS-factors. Importantly, these anti-QS reagents (allicin, triazolyldihydrofuranone and baicalin hydrate) are effective against multidrug-resistant P. aeruginosa [33][34][35]. Studies in the past decade revealed that these anti-QS reagents also inhibited many human infections caused by biofilm-producing bacteria [52,53]. This fact is important for fighting human pathogenic bacteria, as biofilms are found to be involved in over 80% of microbial infections in humans [54]. Since rHPL OE could reduce the levels of pyoverdine and pyocyanin, we speculate that rHPL OE might also inhibit P. aeruginosa infection. It was found that rHPL OE significantly reduced the cytotoxicity towards A549 cells and neutralized toxicity (leading to development retardation or death) of the zebrafish embryo caused by P. aeruginosa PA14.

Conclusions
Binding and down-regulation of di-rhamnolipid has not attracted much attention so far. In this study, we optimized the production of rHPL in E. coli using chaperone co-expression. rHPL OE , a highly specific rhamnose binding protein, bound to not only bacterial cells and PAMPs [29] but also extracted cell-free biofilm from P. aeruginosa. In addition, such interaction inhibited biofilm formation and dispersed mature biofilm through down-regulating secretion of di-rhamnolipid in biofilm and further down-regulating other QS-factors including extracellular proteases, pyoverdine and pyocyanin. As a biofilm dispersion reagent, rHPL OE increased the antibiotic activity against P. aeruginosa PA14 with pre-formed biofilm. Therefore, rHPL OE promised to be an effective anti-biofilm agent for combination therapy. At cellular and animal levels, rHPL OE inhibited the infection and toxicity of P. aeruginosa PA14 towards human lung epithelial cells and zebrafish embryos. These results indicated that rHPL OE served as a novel anti-biofilm agent by targeting rhamnose-containing components in biofilm, which in turn linked glycan-recognition to novel anti-biofilm strategies against pathogenic bacteria.

Bacteria Strains, Growth Medium and Plasmid
Escherichia coli TOP10F' (Invitrogen, Waltham, MA, USA) was used as a host for vector construction and DNA manipulation. E. coli BL21(DE3) (Invitrogen, Waltham, MA, USA) was used as a host for protein expression. Pseudomonas aeruginosa PA14 (serotype O19) was kindly provided by Dr. Hwan-You Chang (Institute of Molecular Medicine, National Tsing Hua University, Hsinchu, Taiwan). The vector pET-23a (+) (Novagen, Burlington, MA, USA) with a T7 promoter was used for recombinant protein expression in E. coli cells and sequence analysis. Takara's Chaperone Plasmid Set (#3340, TaKaRa, Shiga, Japan) was used for chaperone co-expression. All other buffers and reagents were of the highest commercial purity.

Biofilm Formation and Extraction
Bacterial biofilm extraction was as described by Chibaet et al. [55]. P. aeruginosa PA14 grown on LB plates was inoculated in LB medium and incubated overnight at 37 • C with 250 rpm circular shaking. A portion of the overnight culture was 1000-fold diluted in 10 mL LB medium and incubated at 37 • C to an A600nm of 1. Bacterial cells with biofilm were harvested from incubated solution by centrifugation at 8000× g for 10 min at 25 • C (KUBOTA, Osaka, Japan). The harvested pellet was re-suspended with 1 mL of 1.5 M NaCl to extract a cell-free biofilm component. The suspensions were centrifuged at 5000× g for 10 min at 25 • C (KUBOTA, Osaka, Japan), and the supernatants containing the biofilm fraction were collected.

Polysaccharide Quantification in the Extracted Biofilm
Absolute ethanol (≥99.8%, #32221), phenol (#P1037) and sulfuric acid (95%-98%, #435589) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Polysaccharide in biofilm was measured as described by Tribedi and Sil [56]. Extracted biofilm solution was mixed with 2.2 volumes of chilled absolute ethanol, incubated at −20 • C for 1 h and centrifuged at 3500× g for 20 min at 4 • C (KUBOTA, Osaka, Japan). The pellet containing exopolysaccharides was dissolved in sterile water and measured by the phenol-sulfuric acid method. Fifty microliters of re-suspended sample was mixed with 200 µL phenol, and then, 750 µL sulfuric acid was added. The solution was left standing for 40 min, and vigorously shaken. After shaking, the A 490nm was measured using a spectrophotometer (U3310, HITACHI, Tokyo, Japan), and the amount of total sugar (excluding amino sugars) was calculated with a standard curve established using glucose.

Anti-Biofilm Assay
KH 2 PO 4 (≥99%, #P5655), K 2 HPO 4 (≥99%, #P3768), (NH 4 ) 2 SO 4 (≥99%, #A4418), MgSO 4 (≥99.5%, #M7506), L-arginine (≥98%, #A5006), polymyxin B (meets USP testing specifications, #P0972) and crystal violet (≥90%, #C0775) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Biofilm formation was described by O'Toole [57]. Briefly, overnight culture of the P. aeruginosa PA14 in LB medium was diluted with fresh M63 medium (22 mM KH 2 PO 4 , 40 mM K 2 HPO 4 , 15 mM (NH 4 ) 2 SO 4 , 1 mM MgSO 4 and 0.4% l-arginine). A total of 1 × 10 6 CFU/well bacterial cells were then cultured with various concentrations of rHPL OE in a final volume of 100 µL M63 medium onto a 96-well clear round bottom polystyrene microplate (#3788, Thermo Fisher, Waltham, MA, USA) and incubated at 37 • C for 24 h without shaking. A standard antibiotic polymyxin B (50 µg/mL) often used to kill P. aeruginosa in other studies was applied as a positive control, and M63 only was used as a negative control. The suspended culture was then discarded, and the plate was washed with ddH 2 O twice to remove any remaining suspended cells in the microtiter wells. The biofilm was then stained with 125 µL of 0.1% crystal violet for 15 min, after which the stained biofilm was washed with ddH 2 O five times to remove any unbound dye. The crystal violet bound to the biofilm was solubilized with 125 µL 30% (v/v) acetic acid and quantified by measuring the A 595nm using the microplate spectrophotometer (iMark Microplate Absorbance Reader, Bio-Rad, Hercules, CA, USA). For the biofilm dispersion test, biofilm cultures were grown statically in 80 µL of M63 medium for 24 h, followed by the addition of 20 µL of the indicated concentration of rHPL OE in M63 medium and incubation for an additional 24 h. The addition of polymyxin B (50 µg/mL) to the culture before biofilm formation was used as a positive control.

Swarming Motility Measurement
The swarming motility of P. aeruginosa PA14 was investigated in plates containing swarming motility media (LB with 0.5% (wt) glucose and 0.6% (wt) agar). Agar was purchased from OXOID (#LP0011, Basingstoke, Hampshire, UK). An aliquot of 2 µL of motility media containing 1 × 10 6 CFU/mL bacterial cells was inoculated in plates with different concentrations of rHPL OE . Subsequently, spots were dired for 20 min at room temperature and incubated at 37 • C for 48 h. The diameter of the circular bacterial growth was measured [50].

Extracellular Rhamnolipid Quantification
Methylene blue (≥82%, #M9140), sodium borate (ACS reagent, ≥99.5%, #S9640) and chloroform (≥99.5%, #C2432) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Rhamnolipid was quantified as described by Pinzon et al. [58]. The supernatant of the anti-biofilm test was collected, and bacterial cells were removed by centrifugation (KUBOTA, Osaka, Japan). The pH of the cell-free supernatant was first adjusted to 2.3 ± 0.2 using 1 N HCl. Two hundred microliters of acidified sample was then extracted with the same volume of chloroform five times. One milliliter of the chloroform extract was carefully removed and mixed with 40 µL of 0.1% methylene blue (freshly prepared and pH adjusted to 8.6 ± 0.2 with 50 mM borate buffer) and 960 µL of distilled water in a 2 mL tube. After being vigorously shaken for 4 min, the samples were left to stand for 15 min. The chloroform phase was transferred into a quartz cuvette and value of A 638 nm was measured by spectrophotometer (U3310, HITACHI, Tokyo, Japan). The values were converted to rhamnolipid concentrations using a calibration curve established by applying the same procedure to standard rhamnolipid solutions of different concentrations.

Azocasein Degradation Assay
The proteolytic activity was determined in the cell-free culture supernatant according to the method by Kessler et al. [59]. The amount of protease released by P. aeruginosa treated by rHPL OE , the supernatant of the anti-biofilm test was collected by centrifugation (KUBOTA, Osaka, Japan).
One hundred and fifty microliters of supernatant were added to 0.3% of 1 mL azocasein (#A2765, Sigma-Aldrich St. Louis, MO, USA) in 50 mM Tris-HCl (pH 7.5) and incubated at 37 • C for 15 min. The reaction was stopped by the addition of 0.5 mL 10% trichloroacetic acid, and the mixture was centrifuged at 8000× g for 5 min (KUBOTA, Osaka, Japan) to obtain a clear supernatant. The absorbance of the clear supernatant was then measured at 400 nm (U3310, HITACHI, Tokyo, Japan) [60].

Pyoverdine Quantification Assay
P. aeruginosa was incubated with different concentrations of rHPL OE at 37 • C for 24 h. Thereafter, the cultured liquid was centrifuged at 8000× g for 15 min (KUBOTA, Osaka, Japan), and the cell-free supernatant was used for the pyoverdine measurement. The relative concentrations of pyoverdine in all of the treated supernatants with respect to the control was measured through a fluorescence microplate spectrophotometer (Victor2, PerkinElmer, Waltham, MA, USA) at an excitation wavelength of 405 nm and an emission wavelength of 465 nm [61].

Pyocyanin Quantification Assay
The pyocyanin quantification assay was performed according to the method described by Essar et al. [62]. P. aeruginosa was incubated with different concentrations of rHPL OE at 37 • C for 24 h, and the cell-free supernatant was collected by centrifugation (KUBOTA, Osaka, Japan). Five milliliters of the supernatant were extracted with 3 mL of chloroform and the chloroform layer was re-extracted with 1 mL of 0.2 N HCl to produce an orange yellow to pink solution, and the chloroform phase was transferred into a quartz cuvette and value of A 520 nm was measured by a spectrophotometer (U3310, HITACHI, Tokyo, Japan).

Synergistic Effect of rHPL OE with Antibiotics
The experiment design was according to the method described by Das et al. [63]. P. aeruginosa with a mature biofilm was treated with rHPL OE in combination with either azithromycin (≥95%, #75199, Sigma-Aldrich, St. Louis, MO, USA) [63] or cephalexin (≥95%, #C4895, Sigma-Aldrich, St. Louis, MO, USA) [64]. Solutions of rHPL OE in 0.3125 µM, 0.625 µM, 1.25 µM and 2.5 µM in combination with different concentrations of antibiotics were directly added to the cultured bacterial liquid with a mature biofilm and held at 37 • C for 24 h. The total protein and amount of biofilm were measured to validate the synergistic antibacterial effect of rHPL OE and antibiotics.

Total Protein Concentration Measurement in the Biofilm
Concentration of the extractable protein was determined as a measure of the P. aeruginosa biofilm population density. The microbial population density in the biofilm is assumed to be directly proportional to the extractable protein concentration [56]. After incubation, planktonic P. aeruginosa cells were removed, and adhered cells remained in the biofilm were gently washed with sterile PBS and boiled for 30 min in 5 mL of 0.5 N NaOH (ACS reagent, ≥97%, #221465, Sigma-Aldrich, St. Louis, MO, USA) to extract the protein. After that, the suspension was centrifuged at 8000× g for 5 min (KUBOTA, Osaka, Japan), and the resulting clear supernatant was collected. The supernatant protein concentration was then measured by the bicinchoninic acid (BCA) protein assay kit (#23225, Thermo Fisher, Waltham, MA, USA).

Anti-A549 Infection Assay
The Roswell Park Memorial Institute (RPMI)-1640 medium (#10-040 CMS) and Antibiotics-Antimycotic solution (PSA, #30-004-CIS) was purchased from Corning (Corning, NY, USA). Characterized fetal bovine serum (FBS, #SH30071) was purchased from HyClone Laboratory (Logan, UT, USA). The A549 infection model was modified from method described by Chi [65]. A549 cells (ATCC ® number: CCL-185™, BCRC, Hsinchu, Taiwan), adenocarcinomic human lung cells, were seeded in 96-well tissue culture plates at 2 × 10 4 containing 100 µL of RPMI-1640 medium supplemented with 10% (v/v) fetal FBS and 1% (v/v) PSA and allowed to grow at 37 • C for 16 to 18 h. Culture supernatants were removed, the monolayer was washed once with PBS buffer, and then 50 µL of serum-free RPMI-1640 containing two times the indicated concentration of rHPL OE was added to cells and incubated for 30 min, followed by inoculation with P. aeruginosa PA14. RPMI-1640 with 5 µM BSA or rHPL DM was used as negative controls. For inoculation, the fresh bacterial cells cultured in LB broth were washed with PBS, re-suspended and diluted in RPMI-1640 medium to a concentration of 1 × 10 8 CFU/mL. Thereafter, 50 µL of the bacterial dilution was applied to the rHPL OE -treated A549 cells at a multiplicity of infection (MOI) of 50 (i.e., 1 × 10 6 CFU/50 µL/well). The blank group was A549 cells treated by 2.5 µM rHPL OE without bacterial infection. After infection for 16 h at 37 • C, the A549 cell viability was determined by the AlamarBlue cell viability assay (BUF012B, Bio-Rad, Hercules, CA, USA). It should be noted that 50 µg/mL polymyxin B should be added to alamarBlue reagent to prevent a survival signal from residual bacteria. The level of viability, expressed as a percentage, was calculated as follows: % viability = [OD of assay cells/OD of control cells] × 100.

Anti-Zebrafish Infection Assay
Zebrafish embryos were purchased from Gendanio Biotech Inc. (New Taipei City, Taiwan). Healthy, transparent and regular embryos were selected and aliquoted into 96-well plates with four embryos per well, followed by incubation at 28 • C in 100 µL embryo water. Thereafter, 100 µL of P. aeruginosa (1 × 10 8 CFU/mL) pre-incubated with 5 mM rHPL OE for 1 h was added to the embryos at 24 h post-fertilization (hpf). The zebrafish embryos were further incubated at 28 • C, and the development of each embryo in 24 h and 48 h was observed using an inverted microscope (TS100, NIKON, Tokyo, Japan) equipped with a digital camera. The number of larva at 48 h in each group was counted, and the rate of successful development was calculated. All zebrafish related experiments were conducted in accordance with the ethical guidelines of Council of Agriculture, Executive Yuan and Ministry of Science and Technology.

Statistical Analyses
All statistical analyses were carried out using GraphPad Prism version 5.01 for Windows (GraphPad Software, La Jolla, CA, USA). Each value was the average of three measurements, where the presented data is the mean ± SD. All means were compared by one-way ANOVA.