An oligosaccharide analog exhibited antifungal activity by misleading cell wall organization via targeting PHR transglucosidases

The fungal cell wall is an ideal target for the design of antifungal drugs. In this study we used an analog of cell wall polymer, a highly deacetylated long-chain chitosan oligosaccharide (HCOS), to test its effect against pathogenic Candida strains. Results showed that HCOS was successfully incorporated into the dynamic cell wall organization process and exhibited an apparent antifungal activity against both plankton and mature fungal biofilm, by impairing the cell wall integrity. Unexpectedly, mechanistic studies suggested that HCOS exerts its activity by interfering with family members of PHR β-(1,3)-glucanosyl transferases and affecting the connection and assembly of cell wall polysaccharides. Furthermore, HCOS showed great synergistic activity with different fungicides against Candida cells, especially those in biofilm. These findings indicated HCOS has a great potential as an antifungal drug or drug synergist and proposed a novel antifungal strategy with structure-specific oligosaccharides mimicking cell wall polysaccharide fragments. IMPORTANCE Fungal infections have always been a puzzle in clinical medicine. Only a few antifungal drugs are available for medical usage and the widespread use of antifungal drugs increased the incidence of drug resistance. It is an urgent need for the development of novel treatment strategies against fungal infections. In this study, we proposed a novel strategy targeting to fungal cell wall against C. albicans. To our knowledge, it is the first study to show a cell wall polysaccharide fragment analog integrate into and interfere with the fungal cell wall, indicating a novel antifungal strategy using structure-specific oligosaccharides mimicking cell wall polysaccharide fragments.


INTRODUCTION 35
Worldwide, more than 400,000 life-threatening invasive fungal infections occur yearly (1-3). Since 36 fungi are eukaryotes like their human host, the number of targets available for antifungal drug 37 development remains limited. The fungal cell wall is an ideal target for the design of antifungal drugs 38 because it is composed almost exclusively of molecules that are not presented in the human body yet 39 are essential for fungal growth virulence and viability (4). The new class antifungal compound 40 targeting cell wall, such as echinocandin, has been widely used in medical practice (5). glucanosyltransferase Gas/Gel/Phr/Epd proteins (9). Chitin, a linear polysaccharide consisted of β-1,4-46 linked N-acetylglucosamine residues (10), takes up 1% to 15% of the fungal cell wall mass (11). Chitin 47 is essential for cell survival because of its significant contribution to the rigidity and strength of the 48 cell wall (6,8,12). Studies showed that cell wall chitin possesses diversity in structure and function 49 through variations in the length and degree of deacetylation of the microfibrils (13). Chitin polymers 50 can be deacetylated by chitin deacetylases to generate chitosan, a polymer of glucosamine residues (8). 51 The degree of deacetylation can reach up to 70-80% in some fungi, such as Cryptococcus neoformans 52 (14, 15). Chitosan is structurally similar to chitin and plays an indispensable role in maintaining cell 53 barrier functions and integrity during cell growth (16). Aberrant chitin deacetylation causes a 54 decreased rate of growth and reduced pathogenicity (17,18). 55 The cell wall is a dynamic structure that undergoes constant remodeling, including modulation of the 56 distribution and crosslinking, essential for cell integrity and survival (19,20). Monosaccharide analogs 57 had been used to interfere with the cell wall glycan biogenesis or explore the glycan biosynthetic 58 machinery by "hijacking" glycan biosynthesis pathways (21-23). It gave us a hint of a novel antifungal 59 strategy: an analog of cell wall fragment/fiber incorporated into the process of fungal cell wall 60 organization to impair the cell integrity. In this study, a chitosan fragment with defined structural 61 characteristics (MW: 5 kDa; the degree of deacetylation: above 90%) was used to test this hypothesis 62 in Candida albicans, the most predominant fungal species causing superficial to life-threatening 63 systemic infections (24-26). The chitosan fragment was structurally similar to the cell wall chitosan 64 polymers, but likely has a different degree and pattern of deacetylation. Our studies showed the 65 fragment integrated into the cell wall and influenced the assembly process of cell wall remodeling, 66 exhibiting excellent potential as a synergist used in antifungal therapy, especially for mature Candida 67 biofilm, a resistance form of C. albicans. 68

RESULTS 69
The long-chain chitosan oligosaccharide (HCOS) exhibited apparent antifungal activity by 70 impairing the cell wall integrity 71 Two chitosan oligosaccharide products, a long-chain chitosan oligosaccharide (HCOS, average 72 molecular weight 5 kDa, degree of deacetylation > 90%) and a shorter one (LCOS, average molecular 73 weight 0.8 kDa, degree of deacetylation > 90%), were used to assess the antifungal activity against 74 planktonic C. albicans cells or its biofilms. HCOS but not LCOS showed evident inhibitory activities 75 against fungal cells in both two lifestyles (Fig. 1A and B). It should be noted that the fungicidal activity 76 of HCOS became more potent on the late-stage biofilm, which usually showed much greater resistance 77 to antifungal drugs, such as caspofungin (Fig. 1C). HCOS treated cells were also more sensitive to cell 78 wall stressors (Fig. 1D). Electron microscopic studies showed that after exposure to a fungicidal 79 concentration of HCOS, C. albicans cells exhibited a porous cell surface appearance (Fig. 1E) and an 80 ambiguous and sparse cell wall layer (Fig. 1F). These results demonstrated that HCOS affected the 81 integrity of the cell wall structure.

The long-chain chitosan oligosaccharide (HCOS) was incorporated into the dynamic cell wall 97
To determine the binding site of HCOS on fungal cells, HCOS was labeled with the fluorescein FITC. 98 The HCOS-FITC showed an unaffected antifungal activity as unlabeled HCOS (supplementary

HCOS exerted its activity by interfering the β-(1,3)-glucanosyl transferases activity of Phr1p 152
To assess the influence of HCOS on the transglucosidase activity of Phr1p, laminarin was used as the 153 donor substrate and an SR-labeled laminaripentaose (L5-SR) as the acceptor molecule in an enzyme 154 assay in vitro. The transglycosylation efficiency of Phr1p was significantly inhibited by HCOS but not 155 LCOS (Fig. 4A). However, HCOS could not act as a substrate for Phr1p (Fig. 4A). It suggested that 156 albicans Phr1 null mutant showed less sensitive to HCOS than the wild type (Fig. 4B). Fluorescent 158 microscopic study also indicated that Phr1p was co-localized with HCOS-FITC (Fig. 4C). In the 159 absence of Phr1p, HCOS-FITC still targeted the cell wall but did not showed the accumulation at cell 160 apexes (Fig.4D). These results suggested that Phr1p was the main target of HCOS, while it is not fully 161 responsible for the integration of HCOS into the cell wall.

Synergistic effects of HCOS with antifungal drugs against Candida species 173
HCOS greatly affected the integrity of the cell wall, which provides a critical protective barrier to the 174 fungi, and HCOS acted well on established biofilm (Fig.1). These findings indicated its great potential 175 as an antifungal drug synergist, especially against fungal biofilm. We assessed the synergetic activity 176 of HCOS with different antifungal drugs against C. albicans biofilm. After exposure to 500 μg/mL 177 HCOS together with 1000 μg/mL fluconazole or 1 μg/mL caspofungin, the fungal biofilm mass was 178 decreased by 57% (FLC+HCOS) or 44% (caspofungin+HCOS), respectively (Fig. 5A and C).  (Table 1 and supplementary Table 2).

237
Candida hyphae were stained with purple red color. The scale bar of kidney is 2mm, the scale bar of magnified images of 238 areas inside the blue boxes is 20 µm. * * P< 0.01, * * * P < 0.001 experiments were repeated three times.

DISCUSSION 240
The fungal cell wall is a protective barrier that is critical for cell viability, making it an attractive target 241 for the development of antifungal therapies, especially because mammalian cells have no equivalent 242 structures (6, 27). Compounds inhibiting cell wall polysaccharide biosynthesis, such as echinocandins 243 and nikkomycins, have been developed, and some of them became crucial first-line drugs for 244 antifungal treatment in clinical practice (28, 29). In this study, a highly deacetylated long-chain 245 chitosan oligosaccharide HCOS mimicking cell wall chitin/chitosan fragments were used to test 246 whether it can sneak into the cell wall remodeling process and interfere its integrity. Results showed 247 that HCOS was promptly incorporated into the Candida cell wall, disturbed the wall morphology, and 248 presented apparent antifungal activity on both planktons and biofilm ( Fig. 1 and Fig. 2). HCOS exerted 249 its activity most probably by interfering with family members of PHR β-(1,3)-glucanosyl transferases 250 wall of the heat-killed C. albicans (supplementary Fig. 2). It suggested that cell wall proteins were 260 critical for the binding of HCOS to the cell wall. Also, HCOS was mainly in the alkali-insoluble chitin 261 potion of cell wall exacts (Fig. 2D). The alkali-insoluble portion from HCOS-treated cell wall exacts 262 showed an increased ratio of glucosamine compared to untreated cells (supplementary Fig. 4). These 263 observations suggested that HCOS sneaked into the cell wall chitin/chitosan organization process. The 264 oligosaccharide size is critical for this action because of a short chain chitosan oligosaccharide LCOS 265 accumulated into the cell instead of the cell surface ( Fig. 2A and B). Also, LCOS did not exert 266 noticeable antifungal effects ( Fig. 1A and B). Previous studies suggested that the chitosan 267 polysaccharide exerted antifungal activity by targeting plasma membrane (32-35). Therefore, 268 incorporation with the cell wall was the key to the antifungal activity of the chitosan oligosaccharide, 269 which highly depends on the oligomer's size, not too short, not too long. 270 Using affinity chromatography, we tried to isolate and identify C. albicans cell wall proteins binding 271 HCOS. We initially use Zymolyase-treated cell samples to isolate HCOS-binding proteins, but did not 272 find any proteins bound to the affinity column (data not shown). We then added the GPI phospholipase 273 together with Zymolyase to treat cells. Multiple bands were eluted from the column (supplementary 274 (36), glucan elongation and branching (37). Enzymatic analysis showed that HCOS acts as an inhibitor 300 rather than a substrate for Phr1p transglucosidases (Fig. 4A). Fluorescence microscope results showed 301 that HCOS could still bind to the cell wall of Phr1 null mutant (Fig. 4D). However, HCOS was not 302 enriched at the tip of the Phr1 null cell, unlike its localization on the wild type of cell surface. We also 303 hydrolyzed the cell wall of HCOS treated cells and analyzed the component of degraded cell walls, 304 including chitin/chitosan fragments, glucan fragments, and chitin-glucan fragments. We did not see 305 obvious difference in the composition of the degradation products between HCOS treated and 306 untreated samples (data not shown). Therefore, these results suggested that HCOS accumulated in the 307 cell wall via non-covalent interactions. PHR1 was important for the apex distribution and antifungal 308 The RNA expression level of the PHR1 gene was upregulated through the maturation of the Candida 323 biofilm (Fig. 3G-I). It explains the enhanced activity of HCOS against the mature biofilm (Fig. 1C). 324 Usually, the maturation of biofilm increases the antifungal resistance of fungal cells (42). Surprisingly, 325 the fungicidal activity of HCOS became more robust following the maturation of the biofilm (Fig. 1C). 326 Moreover, a combination of first-line antifungal drugs and HCOS effectively killed planktons and 327 eliminated fungal biofilms of several common pathogenic Candida species (Table 1 and  328   Supplementary Table 2), suggesting a synergistic activity of HCOS and fungicides. Antifungal 329 treatment on biofilm-associated fungal infections is often ineffective, leading to recurrent and chronic 330 infections and biofilm-specific drugs are not available currently (42). A combination of HCOS and 331 fluconazole significantly increased the survival rate and reduced the renal fungal burden of C. albicans 332 infected mice (Fig. 6). These results strongly suggested that HCOS had excellent potential for 333 developing biofilm-specific antifungal drugs or drug synergists. Although it required a relative high 334 concentration of HCOS to exhibit antifungal activity, high concentration HCOS did not show obvious 335 cell toxicity to mammalian cells ( Fig.6A and B) (43). Several studies showed that as high as 2000 336 mg/kg/day orally administered chitooligosaccharide showed no toxicity in the rat (44). Thus, 337 chitooligosaccharide is much safer than commonly used antifungal drugs. To be noted, our animal 338 studies showed that 50mg/kg HCOS alone already showed noticeable improvement in the survival rate. 339 Natural glycan-based drugs might need relatively high concentration to treat the pathogenic infection. 340 However, they usually do not have high toxicity and strong side effect. Also, HCOS showed a 341 significant synergist effect with other fungicides ( Table 1, Table S2). In this case, the effective 342 concentration of both HCOS (around 100 μg/ml) and the drug is lower. In the antibiofilm assay, HCOS 343 alone even showed a lower MIC value compared to fluconazole. In fact, chitosan and chitosan 344 oligosaccharide have been widely used as foods and biomedical materials (45, 46). It has been proved 345 with good biocompatibility and safety to human body. Moreover, HCOS might be used as a carrier of 346 other cell wall acting compounds or further modified to improve the antifungal activity. This study 347 opens new therapeutic perspectives for treating human candidiasis. 348

HCOS affinity chromatography of C. albicans cell wall proteins. 443
The HCOS affinity column was prepared as described previously (59). An aliquot of 100 mg HCOS 444 was dissolved in 10 mL 0.1 M NaHCO3-Na2CO3 buffer (pH 8.3), then added into 10 mL NHS-445 activated agarose. After the reaction mixture was shaken (50 rpm) at 4 °C for 14 h, the product was 446 washed with 20% ethyl alcohol then packed into a 1.5 × 5 cm column. To determine the β-1,3-glucanosyltransferase activity of the Phr1 protein, we adopted the previously

Degree of deacetylation of the cell wall chitin by HPLC analysis. 499
The cell wall of C. albicans was extracted as previously described (65). C. albicans were grown in 500 YPD to mid-exponential phase and treated with HCOS, LCOS for 2 h at the concentration of 500 501 μg/mL, respectively. Cells were washed with sterile water and lysed with liquid nitrogen extraction 502 thoroughly. The pellets were washed with 1 M NaCl repeatedly to remove contaminating cytoplasmic 503 proteins and freeze-dried. The cell wall (20 mg) was hydrolyzed with 2 M trifluoroacetic acid for 3 h 504 at 120 °C and freeze-dried after concentrating by rotary evaporation. The samples were resuspended 505 in water to a 10 mg/ml concentration and analyzed by HPLC as described previously (66), and the 506 relative proportions of monosaccharides in the cell wall were calculated. 507

Distribution of COS in the cell wall. 508
The cells were treated with HCOS-FITC and LCOS-FITC for 2 h at the concentration of 500 μg/mL 509 to identify the cell wall components that HCOS incorporates. The cell wall extraction method is as 510 were collected, and the absorbance at 405 nm was measured using a TECAN Infinite M200 PRO 526 multifunction microplate reader (TECAN, Grodig, Austria). Additionally, rabbit red cells incubated 527 with PBS served as a negative control group, and those incubated with distilled water served as a 528 positive control (serving as 100% hemolysis). 529

Murine model of disseminated candidiasis. 530
The sample size was selected based on the preliminary infection trial (n = 5 for mice models). Mice 531 were relocated randomly to treatment or control cages. For the mouse disseminated Candidiasis model, 532 Kunming female mice (n = 5 per group) were infected with a dose of 2.5 × 10 6 C. albicans cells 533 suspension by intravenous injection via the tail vein. Mice were treated 2 h post-infection with a 534 specified intraperitoneal administration of fluconazole (0.1mg/kg, 5mg/kg) and oral gavage HCOS 535 (10mg/kg, 50mg/kg) alone or the combination of fluconazole with HCOS (0.1mg/kg + 50mg/kg, 536 0.1mg/kg +10mg/kg), 0.9% NaCl as control. The animals were weighed every day, and after 7 days of 537 treatment, the mice were euthanized, and one kidney was removed aseptically, placed in PBS, and then 538 homogenized via bead beating. Serial dilutions of the homogenized kidneys were plated on YPD agar 539 for the enumeration of fungal colonies. CFU values in kidneys were expressed as CFU/g of tissue, then 540 transformed into log10 units. The differences between groups were analyzed by analysis of variance. 541 Another kidney was fixed in 4% paraformaldehyde, embedded in paraffin wax, and sectioned 542 longitudinally. Specimens were stained with periodic acid-Schiff (PAS) for the assessment of fungal 543 invasion according to previous publication (68). 544

Statistical analysis. 545
Graphical evaluations were performed with GraphPad Prism v5.0 (GraphPad Software Inc., San Diego, 546 CA). Analysis of variance (ANOVA) was used to evaluate significant differences. Data are presented 547 as means ± SD. A two-tailed Student's t-test was performed to compare two groups and one-way 548 ANOVA for multiple group analysis. The P-value < 0.05 or 0.01 was considered as significant. All 549 data were analyzed using SPSS 19.0 software (SPSS, USA). 550

ACKNOWLEDGMENTS 551
We are grateful to Xixia. Li, and Xueke Tan for helping with sample preparation and taking TEM 552 images at the Center for Biological Imaging (CBI), Institute of Biophysics, Chinese Academy of 553 Science. We thank Prof. Huang for generously providing Phr1 mutant strains and thank Prof. Guan for 554 providing GPI enzyme expression plasmid. 555 HCOS-FITC at concentration of 500 µg/mL for 24 h, the biofilm cell viability was measured 745 by XTT assay and normalized to the control. Data represented as means ± SD (n=6). n.s., 746 non-significant. 747  HCOS (500 µg/mL) or their combination with a flow rate of 5 μL/min for 2 h. The biofilm in 785 microchannels was stained with 1μg/mL DAPI and 1μg/m PI and observed by fluorescence 786 microscope. Scale bar, 100 μm. 787