Disruption of protein rhamnosylation affects the Sporothrix schenckii-host interaction

Highlights

• •Sporothrix schenckii RmlD gene is involved in rhamnose synthesis.
• •The RmlD silencing led to changes in the cell wall composition and organization.
• •Cells with low rhamnose levels are differentially recognized by innate immne cells.
• •The RmlD silenced mutants showed virulence attenuation in the Galleria mellonella.

Abstract

Sporotrichosis is a fungal disease caused by the members of the Sporothrix pathogenic clade, and one of the etiological agents is Sporothrix schenckii. The cell wall of this organism has been previously analyzed and thus far is known to contain an inner layer composed of chitin and β -glucans, and an outer layer of glycoproteins, which are decorated with mannose and rhamnose-containing oligosaccharides. The L-rhamnose biosynthesis pathway is common in bacteria but rare in members of the Fungi kingdom. Therefore, in this study, we aimed to disrupt this metabolic route to assess the contribution of rhamnose during the S. schenckii-host interaction. We identified and silenced in S. schenckii a functional ortholog of the bacterial rmlD gene, which encodes for an essential reductase for the synthesis of nucleotide-activated L-rhamnose. RmlD silencing did not affect fungal growth or morphology but decreased cell wall rhamnose content. Compensatory, the β-1,3-glucan levels increased and were more exposed at the cell surface. Moreover, when incubated with human peripheral blood mononuclear cells, the RmlD silenced mutants differentially stimulated cytokine production when compared with the wild-type strain, reducing TNFα and IL-6 levels and increasing IL-1 β and IL-10 production. Upon incubation with human monocyte-derived macrophages, the silenced strains were more efficiently phagocytosed than the wild-type strain. In both cases, our data suggest that rhamnose-based oligosaccharides are ligands that interact with TLR4. Finally, our findings showed that cell wall rhamnose is required for the S. schenckii virulence in the G. mellonella model of infection.

1. Introduction

Sporotrichosis is an acute or chronic granulomatous subcutaneous mycosis that occurs in humans and other mammals. These infections are caused by members of the pathogenic clade of the Sporothrix genus, which contains the species Sporothrix schenckii, Sporothrix brasiliensis, Sporothrix globosa, and Sporothrix luriei (de Beer et al., 2016). The disease has a worldwide distribution and several clinical manifestations, ranging from the most common acute and chronic cutaneous and subcutaneous lesions to more serious infections involving the lymph nodules, fascia, cartilage, muscle, and bones (Chakrabarti et al., 2015, Lopes-Bezerra et al., 2018a). The etiological agents are widely distributed in the environment, although regionalisms can be observed for some of the species, and thus far, it is considered to be a unique infection caused by a dimorphic fungus with a substantial zoonotic transmission (Chakrabarti et al., 2015).

Similar to the other members of the pathogenic clade, S. schenckii is a dimorphic organism, and the mycelial form is considered the saprophytic morphology found when growing in the environment. In contrast, the yeast-like morphotype is frequently found when invading the host tissues (Mora-Montes et al., 2015, Lopes-Bezerra et al., 2018a).

The fungal cell wall is a dynamic structure that confers protection to stress, controls permeability, is essential to cell viability, morphogenesis, and pathogenesis (Díaz-Jiménez et al., 2012). Moreover, the cell wall contains molecules that induce and modulate innate and adaptive immune responses (Netea et al., 2006, Netea et al., 2008, Latgé et al., 2017, Martinez-Alvarez et al., 2017, Lopes-Bezerra et al., 2018a). The S. schenckii cell wall is composed of chitin, β-1,6-, β-1,4-, and β-1,3-glucans that form the inner layer; whilst the cell surface or outer layer is composed mostly of peptidorhamnomannan, which is a complex of molecules with a wide molecular weight range, containing 12.2% of proteins, 57% of mannose and 33.5% of rhamnose (Lloyd and Bitoon, 1971, Martinez-Alvarez et al., 2017, Lopes-Bezerra et al., 2018b). Similar to S. schenckii, the fungal pathogen Pseudallescheria boydii contains rhamnomannan in its cell wall (Pinto et al., 2001), which activates macrophages in a TLR4-dependent manner (Figueiredo et al., 2010) and is crucial for adhesion to mammalian cells (Pinto et al., 2004). These rhamnomannans from both species are immunogenic structures that are recognized by antibodies (Lopes-Bezerra, 2011). In bacteria, the rhamnose-containing cell wall polysaccharides have been described to have an important role in virulence, increasing adherence and immune evasion, changes in the structure of these polysaccharides caused attenuation of virulence in a mouse model of peritonitis, this reduction in the virulence correlated with enhanced phagocytic uptake and clearance by neutrophils (Teng et al., 2002, Mistou et al., 2016).

Rhamnose is incorporated in many glycoproteins, exopolysaccharides, and other components of the fungal cell wall in Rhynchosporium secales, Cephalotheca purpurea, Cephalotheca reniformis, Penicillium chrysogenum, and Ophiostoma ulmi (Martinez et al., 2012). In both Magnaporthe grisea and Botryotinia fuckeliana, rhamnose synthesis is carried out using UDP-glucose as a precursor, using two enzymes, one of wich has a double enzymatic activity (Fig. 1) (Martinez et al., 2012). The functional orthologs for the protein-encoding genes have also been reported in Botrytis cinerea (Ma et al., 2017). In bacteria, the rhamnose synthesis involves four enzymes with unique enzymatic activities that convert the precursor glucose to dTDP-L-rhamnose, and the main difference with the fungal system is the epimerization and reduction steps that are carried out by individual enzymes, RmlC and RmlD, respectively (Fig. 1) (Graninger et al., 1999, Giraud and Naismith, 2000).

Fig. 1.

Synthesis of nucleotide-bound rhamnose in bacteria and fungi. In bacteria the biosyntethic pathway involves four different enzymes (RmlA, RmlB, RmlC and RmlD) and starts with glucose as precursor. The main differences in fungi are the precursor, UDP-glucose in this case, and that the last two steps in the synthesis are carried out by a single enzyme with bifunctional activity (epimerase and reductase).

To gain insights into the role of rhamnose-containing molecules during the S. schenckii-host interaction, we silenced RmlD and performed a phenotypical characterization of the mutant cells. In particular, we analyzed the changes in the cell wall composition, the interaction with human peripheral blood mononuclear cells (PBMCs), with human monocyte-derived macrophages, and assessed the virulence in Galleria mellonella larvae.

2. Material and methods

2.1. Strains and culture media

Strains used and generated here are listed in Table 1. Fungal cells were maintained at 28 °C in YPD medium, pH 4.5 (1% [w/v] yeast extract, 2% [w/v] gelatin peptone, and 3% [w/v] dextrose). Agrobacterium tumefaciens AGL-1 was maintained at 28 °C in Luria-Bertani medium (0.5 [w/v] yeast extract, 1% [w/v] gelatin peptone, and 1% [w/v] NaCl). When selection required, the medium was supplemented with 100 μg mL⁻¹ ampicillin and 100 μg mL⁻¹ kanamycin. For conidia, cells were grown on solid YPD medium, pH 4.5, at 28 °C for 7 days, and then harvested by surface scratching, as described (Martinez-Alvarez et al., 2017). For hyphal growth, conidia were incubated in YPD broth, pH 4.5, at 28 °C and 120 rpm for 48 h, and harvested by filtering, using a vacuum system and a 5-µm nylon membrane (Monodur®). For yeast-like cells, conidia were incubated in YPD broth, pH 7.8 for 7 days at 37 °C and 120 rpm, as described, and harvested by centrifuging at 5000 × g for 5 min at 4 °C (Martinez-Alvarez et al., 2017). In both cases, cells were washed three times with deionized water and kept at −20 °C until used. Heat inactivation of cells was carried out at 60 °C for 2 h, (Martinez-Alvarez et al., 2017) and confirmation of the viability loss was performed on YPD plates, pH 4.5, incubated at 28 °C for 5 days. Upon Agrobacterium-mediated transformation, S. schenckii cells were selected on YPD plates, pH 4.5 added with 400 mg mL⁻¹ hygromycin B, and incubated at 28 °C for 5 days (Lozoya-Pérez et al., 2018, Lozoya-Perez et al., 2019). Streptococcus mutans Xc was cultured in Todd-Hewitt Broth (THB; Oxoid) or on THB agar at 37 °C with 5% CO₂. When required, THB was supplemented with 10 μg mL⁻¹ erythromycin or 3 μg mL⁻¹ chloramphenicol.

Table 1.

Strains used in this work.

Strain	Organism	Genotype	Reference
Xc	Streptococcus mutans	Serotype C wild-type strain	(Koga et al., 1989)
SMU ΔrmlD	Streptococcus mutans	Xc ΔrmlD::erm	(van der Beek et al., 2015)
SMU DrmlD + pSsRmlD	Streptococcus mutans	Xc ΔrmlD::erm transformed with pSsrmlD	This work
1099–18 ATCC MYA 4821	Sporothrix schenckii	Wild-type	(Castro et al., 2013)
HSS29	Sporothrix schenckii	1099–18 ATCC MYA 4821 transformed with pBGgHg	This work
HSS30	Sporothrix schenckii	1099–18 ATCC MYA 4821 transformed with pBGgHg	This work
HSS20	Sporothrix schenckii	1099–18 ATCC MYA 4821 transformed with pBGgHg-RmlD	This work
HSS21	Sporothrix schenckii	1099–18 ATCC MYA 4821 transformed with pBGgHg-RmlD	This work
HSS22	Sporothrix schenckii	1099–18 ATCC MYA 4821 transformed with pBGgHg-RmlD	This work
HSS23	Sporothrix schenckii	1099–18 ATCC MYA 4821 transformed with pBGgHg-RmlD	This work
HSS24	Sporothrix schenckii	1099–18 ATCC MYA 4821 transformed with pBGgHg-RmlD	This work
HSS25	Sporothrix schenckii	1099–18 ATCC MYA 4821 transformed with pBGgHg-RmlD	This work

2.2. Heterologous complementation of Streptococcus mutans ΔrmlD

The heterologous expression was performed following a methodology previously described by van der Beek et al. (2015). Briefly, the S. schenckii RmlD open reading frame was cloned into XbaI and BamHI sites of the vector pDC123 (Chaffin and Rubens, 1998), generating pSsRmlD, which was used to transform S. mutans Xc wild-type and selected by the conferred resistance to chloramphenicol. The presence of the plasmid inside the bacteria was confirmed by PCR. The bacterial rmlD gene was knocked out by precise in-frame allelic replacement of rmlD with an erythromycin resistance gene in S. mutans + pSsrmlD (van der Beek et al., 2015). For this, the disruption cassette was constructed as follows. A fragment of 700 bp immediately upstream of rmlD was amplified with the primers 5′-CGCAGCAAGCAGTTACGTGATTTGTTGAAG and 5′-GTTTTGAGAATATTTTATATTTTTGTTCATTATTTTTTCTCCTTTAAAAAGCTTTATTACTATTACC, and a fragment of 674 bp immediately downstream of rmlD was amplified with the primers 5′-AGTTATCTATTATTTAACGGGAGGAAATAATATTTTAGCAAAGAAGGACAGGTTTAAACC-3′ and 5′-CTGAAGGTGATAAATCCGTGCCATA-3′ (underlined sequences correspond to 30 bp to the 5′ and 3′ ends of the erm gene, respectively). The upstream and downstream PCR fragments were fused by PCR with the 738 bp amplicon of the erm gene, amplified from the pDCerm plasmid (Jeng et al., 2003). The resultant rmlD disruption cassette was used to transform S. mutans + pSsRmlD and transformants were selected by the double resistance to erythromycin and chloramphenicol. PCR analysis and sequencing were used to confirm the native rmlD disruption. As a result, S. mutans ΔrmlD + pSsRmlD was generated.

For electron microscopy, overnight cultures were diluted and grown to the mid-log phase, cells harvested by centrifuging, and samples were washed, fixed, and dehydrated as described previously (Garufi et al., 2012). Then, cells were mounted onto 12.5 mm specimen stubs (Agar Scientific) and coated with gold to 1 nm using a Quorum Q150R S sputter coater at 20 mA. Visual examination was performed with a Phenom PRO desktop scanning electron microscope (Phenom-World BV), operated with an acceleration voltage of 10 kV.

2.3. Sporothrix schenckii RmlD silencing

Fungal genomic DNA was isolated as described elsewhere (Robledo-Ortiz et al., 2012) and used, along with the primer pair 5′-CTCGAGATGTCCGACAAACACGTACTCG and 5′-AAGCTTAATGTTGCGGGCAGCACA (underlined sequences are added recognition sites for XhoI and HindIII, respectively) to amplify 315 bp fragment of the 5′ region of the RmlD open reading frame. The amplicon was cloned into the pSilent-1 XhoI and HindIII sites (Nakayashiki et al., 2005) generating the construction pSilent-1-RmlD-sense. The antisense fragment was amplified by PCR using a primer pair with the same sequence as the one used for amplification of the sense region, but with recognitions sites in the 5′ ends for StuI and BglII for forward and reverse primers, respectively. The resulting amplicon was cloned into the StuI and BglII sites of pSilent-1-RmlD-sense, generating pSilent-1-RmlD-sense-antisense. This construction, along with the primer pair 5′-CTGCAGATGCCAGTTGTTCCCAGTGATC and 5′-GAGCTCCCTCTAAACAAGTGTACCTGTGCATT (underlined sequences are added recognition sites for PstI and SacI, respectively), were used to amplify by PCR a 2814 bp amplicon that spans from the Aspergillus nidulans TrpC promoter to the A. nidulans trpC terminator, including the sense-intron-antisense region for RmlD (Lozoya-Perez et al., 2019). The amplicon was cloned into the PstI and SacI sites of the binary vector pBGgHg (Chen et al., 2000), generating pBGgHg-RmlD. This construction was used to transform A. tumefaciens AGL-1, and then cells were incubated in minimal medium (0.34 M K₂HPO₄, 0.16 M NaH₂PO₄, 0.37 M NH₄Cl, 0.24 MgSO₄, 0.04 M KCl, 1.8 mM CaCl₂, FeSO₄ 0.18 mM, pH 7.0 adjusted with 1 N HCl) supplemented with 200 μM acetosyringone (Sigma) for 4.5 h at 28 °C and 250 rpm (Lozoya-Pérez et al., 2018). Aliquots of 100 μL bacterial suspension and 100 μL of conidia at 1 × 10⁶ conidia mL⁻¹ were mixed and incubated for 72 h at 28 °C on a cellophane disk on top of YPD agar. Then, the cellophane was placed on YPD, pH 4.5 added with 400 mg mL⁻¹ hygromycin B and 200 μM cefotaxime, and incubated for 72 h at 28 °C. The transformant cells were subjected to five monoconidial passages in YPD, pH 4.5, and three events of dimorphism induction in YPD, pH 7.8. In all cases, the media were added with 400 mg mL⁻¹ hygromycin B for transformants selection.

2.4. Analysis of gene expression and insertional events in mutant strains

The insertion of the binary plasmid within the fungal genome was confirmed by PCR, using the primer pair 5′-GGCGACCTCGTATTGGGAATC and 5′-CTATTCCTTTGCCCTCGGACGAG-3′, as previously described (Lozoya-Perez et al., 2019). The forward primer aligns in the pdg promoter and the reverse primer inside the hph marker. Total RNA was extracted as described elsewhere (Robledo-Ortiz et al., 2012), cDNA synthesized and purified as reported (Trujillo-Esquivel et al., 2016, Trujillo-Esquivel et al., 2017), quantified in a NanoDrop 2000 (Thermo Scientific), and used in qPCR reactions in a thermocycler StepOne Plus (Life Technologies). The PCR reactions contained the SYBR Green PCR Master Mix (Life Technologies) and the primer pair 5′-GTCCGACAAACACGTACTCG and 5′-GGGTCCTTGTCGACCTTGT that generated an amplicon of 243 bp from the 5′ end of the RmlD open reading frame. To estimate the number of the binary plasmid insertional events the same strategy was used, but genomic DNA was used instead of cDNA. In both RT-qPCR and qPCR data were analyzed with the StepOne software V 2.2 (Life Technologies) and the 2^-ΔΔCt method (Livak and Schmittgen, 2001). The encoding gene for the ribosomal protein L6 was used for data normalization and was amplified with the primer pair 5′-ATTGCGACATCAGAGAAGG and 5′-TCGACCTTCTTGATGTTGG, as reported (Trujillo-Esquivel et al., 2017); while the parental strain 1099–18 ATCC MYA 4821 was considered as the reference condition.

2.5. Cell wall analysis

Cells homogenates from yeast-like cells were mechanically prepared in a Braun homogenizer as described elsewhere (Mora-Montes et al., 2010b). Cell walls were resuspended in deionized water and centrifuged for 10 min at 20 000 × g and 4 °C, the supernatant discarded and the pellet washed additionally five times with deionized water. Then, cell walls were subjected to removal of intracellular contaminant components by serial incubations with hot SDS, β-mercaptoethanol, and NaCl, and then hydrolyzed with 2 M trifluoroacetic acid (Sigma-Aldrich), as described. (Mora-Montes et al., 2007) Acid-hydrolyzed samples were analyzed by High-Performance Anion-Exchange Chromatography with Pulsed Amperometric Detection (HPAEC-PAD) with a Dionex system (Thermo Fisher Scientific), using separation conditions reported elsewhere (Estrada-Mata et al., 2015). For cell wall protein concentration, cleansed walls were alkali-hydrolyzed as reported (Mora-Montes et al., 2007), before quantification using the Pierce BCA Protein Assay (Thermo Fisher Scientific).

To estimate the ability to bind Alcian blue, yeast-like cells were grown to mid-log phase, then pelleted by centrifuging, washed twice with deionized water, and the cell concentration adjusted at an OD₆₀₀ of 0.2 in deionized water. Aliquots of 1 mL were used to interact with 30 μg mL⁻¹ Alcian blue (Sigma-Aldrich) and analyzed as described (Hobson et al., 2004). The cell wall porosity to polycations was calculated as described previously (De Nobel et al., 1990), incubating cells with either 30 µg mL⁻¹ poly-L-lysine (MW 30–70 kDa, Sigma-Aldrich) or 30 µg mL⁻¹ diethylaminoethyl-dextran (MW 500 kDa, Sigma-Aldrich).

2.6. Analysis of cell wall N-linked and O-linked glycans

To trim O-linked glycans 1 × 10⁹ yeast-like cells were resuspended in 1 N NaOH and gently shook for 18 h at room temperature. For N-linked glycans removal, aliquots containing 1 × 10⁹ yeast-like cells were incubated for 20 h at 37 °C with 25 U endoglycosidase H (New England Biolabs) (Lozoya-Perez et al., 2019). In both cases, cells were pellet by centrifuging, and the supernatants were collected, lyophilized, and stored at −20 °C until used. Total sugar content was quantified with the phenol–sulfuric-acid method (Dubois et al., 1956), while acid hydrolysis and monosaccharide separation was performed by HPAEC-PAD as described. (Mora-Montes et al., 2012).

2.7. Chitin and β1,3-glucan staining

For chitin labeling, yeast-like cells were incubated with 500 µg mL⁻¹ fluorescein isothiocyanate conjugated-wheat germ agglutinin (WGA-FITC; Sigma-Aldrich) for 60 min at room temperature, as reported. (Mora-Montes et al., 2011); whilst β1,3-glucan was labeled by incubating cells with 5 µg mL⁻¹ IgG Fc-Dectin-1 chimera (Graham et al., 2006) for 40 min at room temperature, followed by an additional incubation with 1 µg mL⁻¹ donkey anti-Fc IgG-FITC (Sigma-Aldrich) for 40 min at room temperature (Marakalala et al., 2013). Samples were examined under fluorescence microscopy in a Zeiss Axioscope-40 microscope and an Axiocam MRc camera. The fluorescence quantification of 300 cells was collected using the software Adobe Photoshop™ CS6 and the formula: [(total of green pixels-background green pixels) × 100]/total pixels (Pérez-García et al., 2016).

2.8. Ethics statement

The Ethics Committee from Universidad de Guanajuato approved the inclusion of primary human cells in this study (Ref. 17082011). Only healthy adult volunteers that received information about the study and signed the informed consent were enrolled. The inclusion of insects in this project was approved by the same institutional Committee (Ref. CIBIUG-P12-2018).

2.9. Stimulation of cytokine production by human peripheral blood mononuclear cells

PBMCs were isolated from human EDTA-treated venous blood, which upon withdrawn was mixed with Histopaque-1077 (Sigma-Aldrich) and subjected to differential centrifugation, as previously reported (Endres et al., 1988). The fungus-immune cell interactions were performed in round-bottom 96-well microplates, using live yeast-like cells (Martinez-Alvarez et al., 2017). Each well contained 100 μL 1 × 10⁵ yeast-like cells and 100 μL 5 × 10⁵ PBMCs in RPMI 1640 Dutch modification (added with 2 mM glutamine, 0.1 mM pyruvate and 0.05 mg mL⁻¹ gentamycin; all reagents from Sigma-Aldrich). The microplates were incubated for 24 h at 37 °C with 5% (v/v) CO₂, centrifuged for 10 min at 3000 × g at 4 °C, the supernatants saved, and kept at −20 °C until used. When indicated, the PBMCs suspensions were pre-incubated for 1 h at 37 °C and 5% (v/v) CO₂ with 200 μg mL⁻¹ laminarin (Sigma-Aldrich), 10 μg mL⁻¹ anti-TLR2 (eBioscience), or 10 μg mL⁻¹ anti-TLR4 (Santa Cruz Biotechnology), before interaction with yeast-like cells. Preincubations with Isotype matched, 10 μg mL⁻¹ irrelevant IgG₁ antibody (Santa Cruz Biotechnology) were used as a control for TLR4 blocking; while preincubations with 10 μg mL⁻¹ IgG₂aκ (eBioscience) were included to control TLR2 blocking assays (Martinez-Alvarez et al., 2017). To avoid any bias for the presence of bacterial lipopolysaccharide, all the fungus-PBMCs interactions were added with 5 μg mL⁻¹ polymyxin B (Sigma-Aldrich) (Martinez-Alvarez et al., 2017), despite were negative for this bacterial wall component, assessed with a Limulus amebocyte lysate (Sigma-Aldrich). The interleukin 1β (IL-1β) concentration was determined in the collected supernatants by ELISA, using a DuoSet ELISA Development kit (R&D Systems); whereas the tumor necrosis factor-alpha (TNFα), interleukin 6 (IL-6), and interleukin 10 (IL-10) were quantified by ELISA with Standard ABTS ELISA Development kits (Peprotech). In all cases, mock interactions with only human PBMCs were included as controls.

2.10. Analysis of phagocytosis by human monocyte-derived macrophages

Upon isolation as described in the previous section, human PMBCs were differentiated to macrophages using recombinant human granulocyte–macrophage colony-stimulating factor (Sigma-Aldrich) (Pérez-García et al., 2016). Before interactions, yeast-like cells were washed twice with PBS and incubated with 1 mg mL⁻¹ Acridine Orange (Sigma-Aldrich) as described (Hernández-Chávez et al., 2019), and then washed twice with PBS and cell concentration adjusted at 3 × 10⁷ yeast-like cells mL⁻¹. The interactions were performed with a macrophage-yeast ratio of 1:6, in 6-well plates, and a final volume of 800 µL DMEM medium (Sigma-Aldrich) per well (Lozoya-Perez et al., 2019). After 2 h of incubation at 37 °C and 5% (v/v) CO₂, macrophages were washed twice with chilled PBS and stained with 1.25 mg mL⁻¹ Trypan Blue, as an external fluorescence quencher, as described (González-Hernández et al., 2017). The cell–cell interactions were analyzed by flow cytometry in a FACSCanto II equipped with a FACSDiva acquisition system (Becton Dickinson, Franklin Lakes, NJ, USA). A total of 50,000 events were collected per sample or condition, gating for immune cells (González-Hernández et al., 2017, Hernández-Chávez et al., 2018, Lozoya-Perez et al., 2019). Fluorescent signals were obtained using the FL1 (green) and FL2 (red) channels, previously compensated with non-stained macrophages. All the interactions were performed in presence of polymyxin B 5 µg mL⁻¹, and when required, preincubations of macrophages with 200 μg mL⁻¹ laminarin (Sigma-Aldrich), 200 μg mL⁻¹ L-rhamnose (Sigma-Aldrich), 10 μg mL⁻¹ anti-TLR2 (eBioscience), 10 μg mL⁻¹ anti-TLR4 (Santa Cruz Biotechnology), 10 μg mL⁻¹ irrelevant IgG₁ antibody (Santa Cruz Biotechnology), or 10 μg mL⁻¹ irrelevant IgG₂aκ (eBioscience) were performed for 1 h at 37 °C and 5% (v/v) CO₂. Cells emitting fluorescence in the green channel were considered in the early stage of phagocytosis, cells emitting in both the green and red channels were considered to be in the intermediate stage of this immune event; while those only emitting signals in the red channel were considered to be in the late stage of the phagocytic process (Hernández-Chávez et al., 2018, Lozoya-Perez et al., 2019).

2.11. Analysis of fungal virulence in Galleria mellonella

The G. mellonella larvae were from an in-house colony previously established (Clavijo-Giraldo et al., 2016) and were raised and fed ad libitum on corn bran and honey diet (1 kg corn bran, 150 g rice meal, 250 mL bee honey, and 70 mL glycerin) (García-Carnero et al., 2020). Only larvae with irritability, absence of body melanization, and a length of 1.2–1.5 cm were included in the host-fungus interaction assays (Clavijo-Giraldo et al., 2016, García-Carnero et al., 2020). Fungal inocula were injected in the last left pro-leg, previously sanitized with 70% (v/v) ethanol, with the assistance of a Hamilton syringe and a 26-gauge needle (Clavijo-Giraldo et al., 2016). The fungal inoculum was of 1 × 10⁵ yeast-like cells 10 µL ^-1. Larvae were kept in Petri dishes at 37 °C and survival monitored daily for 15 days. To maintain animal hydration, chopped apple was included in the animal housing during the observation period (García-Carnero et al., 2020). The animal silk was removed when evident to delay the transition to the pupa. Loss of irritability and extensive body melanization were taken as signs of insect death. Larvae were analyzed in groups that included 30 animals per experimental condition. As a control, one animal group was injected only with PBS, to assess animal mortality by animal manipulation or mechanical injuries To calculate the colony-forming units (CFUs), alive and dead animals were decapitated with a sterile scalpel, the hemolymph collected, serially diluted, and incubated on YPD plates, pH 4.5, at 28 °C for 72 h.

Hemocytes were quantified in anticoagulated hemolymph as previously reported (García-Carnero et al., 2020); whilst cytotoxicity and phenoloxidase activities were measured in cell-free hemolymph using 20 mM 3,4-dihydroxyDL-phenylalanine (Sigma-Aldrich) and the Pierce LDH Cytotoxicity Assay (Thermo Fisher Scientific), respectively (Martínez-Álvarez et al., 2019, Gómez-Gaviria et al., 2020).

2.12. Statistical analysis

Statistical analysis was performed using GraphPad Prism 6 software. The Mann-Whitney U test was used to analyze cytokine stimulation and phagocytosis by human cells. These experiments were carried out in duplicate with samples from eight healthy donors. Animal survival experiments were performed in groups, each containing 30 larvae. Data were analyzed using the Log-rank test and are reported in Kaplan-Meier survival curves. Other experiments were performed at least three times in duplicate and the unpaired t-test was used to establish statistical significance. All data are represented as the mean and standard deviation. In all cases, the significance level was set at P < 0.05.

3. Results

3.1. Identification of the S. schenckii RmlD and functional complementation in bacteria

To disrupt the synthesis of rhamnose-containing molecules in S. schenckii, we first predicted the putative genes involved in L-rhamnose biosynthesis, using the genome sequence already reported for this organism and the biosynthetic pathway described in Streptococcus suis (Martinez et al., 2012, Teixeira et al., 2014). Our analysis indicated that the sequence SPSK_06220 (accession code to GenBank) from S. schenckii showed 55% similarity to dTDP-glucose 4,6-dehydratase encoded by rmlB (accession code WP_024531966), while SPSK_06451 showed a 48% similarity to dTDP-4-dehydrorhamnose reductase encoded by rmlD (WP_174845793). No putative homolog was found for the dTDP-4-dehydrorhamnose 3,5-epimerase encoded for rmlC. In M. grisea, this biosynthetic pathway has also been described, but differently from prokaryotes, this pathway only involves two enzymes, UG4,6-Dh and U4k6dG-ER, which have UDP-glucose 4,6-dehydratase and UDP-4-dehydrorhamnose 3,5-epimerase/UDP-4-dehydrorhamnose reductase activities, respectively. UG4,6-Dh represents the functional ortholog of RmlB and U4k6dG-ER combines the activity of RmlC and RmlD (Martinez et al., 2012). This bifunctional activity of epimerase and reductase for UG4,6-Dh has also been described for Botrytis cinerea bcer and Botryotinia fuckeliana U4k6dG-ER (Martinez et al., 2012, Ma et al., 2017). The S. schenckii putative proteins encoded by SPSK_06220 and SPSK_06451 showed a similarity of 92% and 91% to UG4,6-Dh and U4k6dG-ER, respectively. SPSK_06220 and SPSK_0645 were named here as RmlB and RmlD, respectively, and were found overexpressed in the yeast morphology, according to the recent transcriptomic analysis of S. schenckii dimorphism (Giosa et al., 2020), showing log2FC values of 1.12 and 2.98, respectively. Since the protein encoded by RmlD is expected to have a bifunctional role as both epimerase and reductase and the gene showed increased overexpression when compared to RmlB, we focused our work on this gene. Bioinformatic analyses showed that RmlD contains a Wierenga motif ¹⁰GASGLLGRQ¹⁸, an Mg⁺² binding site ¹³⁷YGESK^141, and a catalytic triad ¹¹²T-¹³⁷Y-¹⁴¹K found in other RmlD homologs (Blankenfeldt et al., 2002). Moreover, it contains a NAD(P)H binding domain, characteristic of the dehydrogenase/reductase superfamily (Blankenfeldt et al., 2002).

To get insights on the role of RmlD in the biosynthetic pathway of L-rhamnose in S. schenckii, heterologous expression and complementation in an S. mutans ΔrmlD mutant strain were performed (van der Beek et al., 2015). It was previously reported that reduction in the rhamnose content significantly attenuated S. mutans growth and affected cell morphology, generating swollen cells and aggregated chains (van der Beek et al., 2015). Heterologous expression with S. schenckii RmlD, under the control of a constitutive promoter, restored the S. mutans ΔrmlD cell morphology in the majority of cells, which formed short chains and grew up unidirectionally. A minority of cells retained the mutant phenotype, i.e., were swollen and aggregated cells (Fig. 2). The growth rate of S. mutans ΔrmlD expressing S. schenckii RmlD was slightly slower than the parental strain, and the buoyancy of S. mutans ΔrmlD + pSsRmlD did not resemble the parental strain, as the cells remained at the bottom of the culture while the WT strain remains in suspension (data not shown). Therefore, these results suggested that S. schenckii RmlD partially complemented the truncated biosynthetic pathway found in S. mutans ΔrmlD.

Fig. 2.

Sporothrix schenckii RmlD partly complements a Streptococcus mutans ΔrmlD mutant strain. Representative scanning electron microscopy images of S. mutans wild type (A), S. mutans ΔrmlD (B), and S. mutans ΔrmlD + pSsRmlD (C) at different growth times. The cell morphology of the wild-type strain appears as short chains with unidirectional growth, the ΔrmlD mutant forms aggregates and long chains with swollen cells. The expression of S. schenckii RmlD restored cell morphology, the cells appear as short chains with unidirectional growth, only a few cells are swollen and forming aggregates.

3.2. Sporothrix schenckii RmlD silencing

To assess the RmlD relevance in S. schenckii fitness and interaction with the host, we generated silenced strains using the previously standardized A. tumefaciens mediated-transformation strategy (Lozoya-Pérez et al., 2018, Lozoya-Perez et al., 2019) and the binary vector pBGgHg-RmlD constructed here, which contains the first 315 bp of the RmlD open reading frame and was cloned in sense and antisense orientations in pBGgHg (Chen et al., 2000). Hygromycin B-resistant cells were selected after transformation, subjected to monoconidial passages, as described in the Materials and methods section, and the insertion of pBGgHg-RmlD within the fungal genome was confirmed by PCR (data not shown). Following this strategy, we selected six colonies for further characterization (HSS20-HSS25). S. schenckii cells were also transformed with the empty pBGgHg vector and two colonies (HSS29 and HSS30) were selected and used as controls to assess any specific effects of the vector on the fungal phenotypes.

Changes in RmlD expression in the selected mutants were evaluated by RT-qPCR assays, using the gene encoding for the ribosomal protein L6 for data normalization (Trujillo-Esquivel et al., 2017), and the parental strain 1099–18 ATCC MYA 4821, referred here as wild-type (WT), as a reference strain. Both strains HSS29 and HSS30 showed RmlD expression levels similar to that observed in the WT strain, confirming that the empty vector did not affect the expression of the target gene (Fig. 3A). However, strains HSS20-HSS25 could be classified into two groups, depending on the RmlD silencing degree. HSS20, HSS21, and HSS22 showed a gene silencing of 52.9 ± 6.9%, 50.1 ± 4.1%, and 55.4 ± 5.6%, respectively, while the second group, composed of HSS23, HSS24, and HSS25 showed a gene silencing higher than 98% (99.7 ± 1.1%, 97.6 ± 1.2%, and 98.5 ± 0.9%, for HSS23, HSS24, and HSS25, respectively; Fig. 3A). The number of insertional events of the binary plasmid was also evaluated in these selected mutant strains by qPCR, amplifying a 243 bp fragment of the RmlD open reading frame. From all the silenced mutants, i.e., HSS20-HSS25 three copies of this RmlD fragment were amplified, while from WT and the control strains HSS29 and HSS30 only one RmlD amplicon was observed (Fig. 3B). The amplified region is part of the sequence cloned in sense and antisense to generate pBGgHg-RmlD, thus the amplification of three copies in the silenced strains can be interpreted as the presence of only one copy of the binary vector per strain and that this occurred in an insertional event outside the RmlD locus. Thus, we continued with the phenotypical characterization of the six silenced strains, to assess any potential contribution of the off-target insertional event on the phenotype, along with the strains transformed with the empty vector. When compared to the WT strain, both control and silenced strains showed the same ability to undergo dimorphism, and similar cell and colony morphology (data not shown). Additionally, no apparent growth defect was detected in these strains, as the doubling times for both mycelia and yeast-like cells of the silenced mutants and control strains were similar to those observed in the WT strain (Fig. 4). Therefore, RmlD silencing did not affect fungal growth or morphology.

Fig. 3.

Analysis of RmlD expression and binary vector insertional events. In A, RT-qPCR reactions amplifying a 243 bp fragment of RmlD open reading frame were used to assess gene expression. In B, analysis of the binary vector insertional events by qPCR, amplifying the same fragment described in A. In both cases, the amplification of the gene encoding for the ribosomal protein L6 data was used for data normalization. Data are means ± SD of three independent experiments performed in duplicates. *P < 0.05 when compared to the WT strain. WT, strain 1099–18 ATCC MYA 4821. Strains HSS29 and HSS30 were transformed with pBGgHg; while HSS20-HSS25 with pBGgHg-RmlD.

Fig. 4.

Doubling times of Sporothrix schenckii wild-type, control, and RmlD-silenced strains. For hyphae, cells were grown in YPD, pH 4.5, and the biomass dry weight determined every 2 h. For yeast-like cells, these were grown in YPD, pH 7.8, and cells quantified in a hemocytometer. From the generated growth curves, the doubling time was calculated for each strain. WT, strain 1099–18 ATCC MYA 4821. Strains HSS29 and HSS30 were transformed with pBGgHg; while HSS20-HSS25 with pBGgHg-RmlD.

3.3. Cell wall composition and protein glycosylation in the silenced Sporothrix schenckii mutants

Since RmlD is likely to participate in the synthesis of nucleotide-activated rhamnose, which is used for protein glycosylation (Ma et al., 2017), and this monosaccharide is found as part of the Sporothrix cell wall (Lopes-Bezerra, 2011, Martinez-Alvarez et al., 2017, Lopes-Bezerra et al., 2018b), we next analyzed the effect of RmlD silencing on the S. schenckii cell wall composition by HPAEC-PAD. Walls from yeast-like cells were isolated, cleansed, and acid hydrolyzed, before the chromatographic separation of monosaccharides (Mora-Montes et al., 2007, Martinez-Alvarez et al., 2017). Both control strains, HSS29 and HSS30 showed a similar sugar composition to the one observed in the WT strain (Table 2). Since these monosaccharides are the building blocks for chitin (glucosamine), glucans (glucose), and protein-modifying glycans (mannose and rhamnose), it is possible to infer the levels of these polysaccharides in the cell walls under analysis (Mora-Montes et al., 2007, Martinez-Alvarez et al., 2017, Lopes-Bezerra et al., 2018b). The strains with intermediate RmlD silencing, HSS20-HSS22, showed no differences when compared among them, but there was a significant increment in glucose content and a reduction of about 60% in the rhamnose levels (Table 2). For the rest of the silencing strains, HSS23-HSS25, that showed gene silencing values close to 100%, the cell wall monosaccharide content was similar among the different clones. However, when compared to the WT strain, there was an increment in glucose content and a reduction in rhamnose levels (Table 2). The changes in glucose levels were not significantly different when compared with the mutant group with intermediate silencing, but rhamnose levels were significantly reduced in strains HSS23-HSSS25 with only 2.5% of the total rhamnose content remaining compared to the WT control strain (Table 2). Glucosamine and mannose levels in all the silenced strains were unaffected by RmlD silencing (Table 2). We also analyzed cell wall porosity to polycations and the total cell wall protein content and found no significant changes in both parameters in any of the silenced mutant or control strains (80.8 ± 13.4% vs. 76.7 ± 11.7% porosity to DEAE-dextran, WT vs average values of silenced and control strains, respectively, P = 0.76; 197.5 ± 23.4 µg of protein mg of cell wall⁻¹ vs. 205.4 ± 19.7 µg of protein mg of cell wall⁻¹, WT vs average values of silenced and control strains, respectively, P = 0.81). The cell wall ability to bind the cationic dye Alcian blue was also assessed. Although there was a trend to bind more dye by the silenced strains, this was not statistically significant (93.7 ± 12.7, 99.5 ± 13.4, and 115.6 ± 15.7 µg dye bound OD^600nm = 1.0^-1 for the WT, average values of control strains, and average values of silenced strains, respectively, P = 0.24).

Table 2.

Cell wall analysis of Sporothrix schenckii wild-type, control, and RmlD-

Cell wall abundance
Organism	Glucosamine (%)	Mannose (%)	Glucose (%)	Rhamnose (%)
Wild type	4.3 ± 2.5	38.3 ± 3.7	42.2 ± 3.0	15.2 ± 1.7
HSS29	4.5 ± 1.8	39.4 ± 5.4	41.4 ± 5.4	14.7 ± 2.1
HSS30	4.0 ± 1.9	40.5 ± 2.5	39.1 ± 4.4	16.4 ± 2.2
HSS20	5.2 ± 2.6	35.8 ± 3.7	52.4 ± 2.3*	6.6 ± 3.8*
HSS21	5.1 ± 2.2	36.1 ± 4.7	53.1 ± 4.9*	5.7 ± 2.9*
HSS22	5.5 ± 3.4	37.1 ± 2.2	51.7 ± 3.3*	5.7 ± 3.4*
HSS23	4.9 ± 2.0	39.4 ± 2.2	55.0 ± 5.4*	0.7 ± 0.2*
HSS24	5.1 ± 2.4	38.8 ± 3.5	55.9 ± 4.8*	0.2 ± 0.1*
HSS25	4.8 ± 3.1	36.5 ± 4.2	58.4 ± 4.9*	0.3 ± 0.2*

silenced strains

* P < 0.05 when compared with the values obtained with the WT strain

Since the reduction in rhamnose content suggested a disruption in the proper protein glycosylation pathway, we next enzymatically removed N-linked glycans by incubating cells with endoglycosidase H followed by acid hydrolysis of isolated glycans to analyze the monosaccharide content HPAEC-PAD. Results showed that all samples yielded similar amounts of total carbohydrate, and those from the WT and the two control strains contained the same proportion of mannose and rhamnose composing the N-linked glycans (Fig. 5A). For strains HSS20-HSS22, there was a significant reduction in rhamnose content of 60% on average, whereas N-linked glycans from strains HSS23-HSS25 contained traces amounts of this monosaccharide (Fig. 5A). Even though the mannose content did not change in any of the samples from the silenced mutants, the ratio mannose:total sugar content varied from 0.7 ± 0.04 for WT or control strains to 0.85 ± 0.03 and 0.95 ± 0.01 for strains HSS20-HSS22 and HSS23-HSS25, respectively. In all cases, trace amounts of glucosamine were detected with no significant changes among the different strains analyzed (Fig. 5A). A similar analysis was carried out with O-linked glycans trimmed from the cell wall by β-elimination. WT or control strains contained similar total sugar content and mannose:rhamnose proportion (Fig. 5B). However, for the HSS20-HSS22 intermediate silenced strains a reduction in total sugar and rhamnose was observed, which was even more pronounced in highly silenced strains HSS23-HSS25 (Fig. 5B). Mannose levels did not show significant changes in samples from the silenced strains (Fig. 5B). Collectively, these data indicated that rhamnose bound to both N-linked and O-linked glycan was reduced in the cell wall of the RmlD-silenced strains, and as consequence, cell wall proteins were decorated with mannose-enriched glycans.

Fig. 5.

The RmlD silencing in Sporothrix schenckii affects the rhamnose content in both N-linked and O-linked glycans from the cell wall. The N-linked (A) or O-linked (B) glycans, were trimmed from the cell wall by incubating yeast-like cells with either endoglycosidase H or sodium hydroxide, respectively, the oligosaccharides acid hydrolyzed and monosaccharide content analyzed by High-Performance Anion-Exchange Chromatography with Pulsed Amperometric Detection. Data are means ± SD of three independent experiments performed in duplicates. *P < 0.05 when compared to WT or control cells. ^†P < 0.05 when compared to WT or control cells, or HSS20-HSS22 strains. WT, strain 1099–18 ATCC MYA 4821. Strains HSS29 and HSS30 were transformed with pBGgHg; while HSS20-HSS25 with pBGgHg-RmlD.

Finally, we analyzed whether RmlD silencing disrupted the organization of the wall structural polysaccharides chitin and β-1,3-glucan. We focused on both polysaccharides because they are localized closer to the plasma membrane (Martinez-Alvarez et al., 2017, Lopes-Bezerra et al., 2018b). Consequently, detection of large proportions of these polysaccharides at the cell surface is indicative of disruption of the S. schenckii cell wall organization (Lozoya-Perez et al., 2019). Polysaccharides were labeled with bulky lectins that are not capable of passing through the cell wall network to bind their targets (Mora-Montes et al., 2011, Marakalala et al., 2013). Therefore, labeling is only observed in areas of the cell wall where these polysaccharides are accessible (Mora-Montes et al., 2011, Martinez-Alvarez et al., 2017, Lozoya-Perez et al., 2019). We included heat-killed (HK) cells as a positive control, since cell inactivation by heat artifactually exposes inner wall components on the cell surface (Gow et al., 2007, Martinez-Alvarez et al., 2017), Results displayed in Fig. 6A indicated that chitin labeling was similar in all analyzed strains, depending on live or HK status, suggesting chitin distribution is not affected by RmlD silencing. In contrast, β-1,3-glucan labeling was low and similar in live cells from WT and control strains but was significantly higher in all the RmlD silenced strains (HSS20-HSS25; Fig. 6B). As expected, heat inactivation exposed β-1,3-glucan, increasing the labeling in WT and control strains (Fig. 6B). HK cells from all RmlD silenced strains showed similar labeling levels that were higher than those observed in WT cells, confirming our earlier observation of higher glucan content in the silenced strains cell wall. Thus, these data indicated that β-1,3-glucan but no chitin is more exposed on the cell surface of strains with RmlD silenced.

Fig. 6.

Cell wall chitin and β-1,3-glucan labeling in Sporothrix schenckii wild-type, control, and RmlD-silenced strains. Yeast-like cells were labeled with fluorescein isothiocyanate conjugated-wheat germ agglutinin for chitin staining (A) or IgG Fc-Dectin-1 chimera for β-1,3-glucan staining (B) as described in Materials and methods, inspected under fluorescence microscopy, and the fluorescence of 300 cells randomly selected was calculated. Data are means ± SD of three independent experiments performed in duplicates. *P < 0.05 when compared to the WT strain. WT, strain 1099–18 ATCC MYA 4821. Strains HSS29 and HSS30 were transformed with pBGgHg; while HSS20-HSS25 with pBGgHg-RmlD.

3.4. The Sporothrix schenckii RmlD silencing affected the cytokine stimulation by human peripheral blood mononuclear cells

We have previously demonstrated that defects in the cell wall composition and protein glycosylation pathways affect the S. schenckii ability to stimulate cytokine production by human PBMCs (Lozoya-Perez et al., 2019). Since RmlD silencing affected both the cell wall and protein glycosylation, we next analyzed the outcome of the interaction between these mutant cells with human PBMCs, measuring the production of soluble TNFα, IL-1β, IL-6, and IL-10. For these four cytokines, the WT and the two control strains, HSS29 and HSS30, stimulated similar cytokine levels (Fig. 7). The strains with intermediate levels of RmlD silencing (HSS20, HSS21, and HSS22) induced lower levels of TNFα and IL-6 compared to PBMCs incubated with the WT strain, but this reduction was not significant (P greater than 0.05 in all cases; Fig. 7). In contrast, these three strains stimulated significantly higher levels of both IL-1β and IL-10 (Fig. 7). No differences were observed in the ability to stimulate any of the four cytokines when compared among the three intermediate clones (Fig. 7). Strains HSS23, HSS24, and HSS25, which have nearly abrogated expression of RmlD, stimulated similar levels when compared among them, and for the case of TNFα and IL-6, these were significantly lower to those obtained with the WT, the control strains, or HSS20-HSS22 silenced strains (Fig. 7). The IL-1β and IL-10 stimulation by any of these three silenced strains was higher than that obtained with WT cells but similar when compared to strains HSS20-HSS22 (Fig. 7). Therefore, the RmlD silencing affected the S. schenckii ability to stimulate cytokine production by human PBMCs.

Fig. 7.

Cytokine stimulation by human peripheral blood mononuclear cells. Yeast-like cells and human peripheral blood mononuclear cells (PBMCs) were coincubated, the supernatants were collected and the concentration of TNFα, IL-6, IL1β, and IL-10 was measured by ELISA. Data are means ± SD obtained with samples from eight donors, each assayed in duplicate wells. *P < 0.05 when compared to WT cells. ^†P < 0.05 when compared to cells under no treatment from the same strain. No treatment, PBMCs preincubated with 5 μg mL⁻¹ polymyxin B; Control anti-TLR2, PBMCs preincubated with 5 μg mL⁻¹ polymyxin B and 10 μg mL⁻¹ IgG₂aκ; Control anti-TLR4, PBMCs preincubated with 5 μg mL⁻¹ polymyxin B and 10 μg mL⁻¹ IgG₁; + anti-TLR2, PBMCs preincubated with 5 μg mL⁻¹ polymyxin B and 10 μg mL⁻¹ anti-TLR2; + anti-TLR4, preincubated with 5 μg mL⁻¹ polymyxin B and 10 μg mL⁻¹ anti-TLR4; + laminarin, preincubated with 5 μg mL⁻¹ polymyxin B and 200 μg mL⁻¹ laminarin. Control, mock reactions were no fungal cells were included. WT, strain 1099–18 ATCC MYA 4821. Strains HSS29 and HSS30 were transformed with pBGgHg; while HSS20-HSS25 with pBGgHg-RmlD.

We next assessed the contribution of some immune receptors during the Sporothrix-PMBC interaction, by blocking with specific antagonists. We particularly focused on TLR2, TLR4, and dectin-1, because these three receptors have been demonstrated to be major players in the S. schenckii immune sensing and for the control of experimental sporotrichosis (Sassá et al., 2012, Negrini et al., 2013, Jellmayer et al., 2017, Martinez-Alvarez et al., 2017). PBMCs preincubation with monoclonal antibodies against TLR2 or TLR4 or laminarin, blocking agents for TLR2, TLR4, or dectin-1, respectively, significantly reduced the TNFα stimulation with WT cells, indicating that these three receptors participate in the signaling pathways to stimulate PBMC cytokine production (Fig. 7). A similar trend was observed with the control strains HSS29 or HSS30, and the silenced strains HSS20-HSS22 (Fig. 7). For TNFα stimulation with the silenced strains HSS23-HSS25, there was a TLR2 and dectin-1 participation to a similar extent than in the systems where the WT or any other of the strains was used (Fig. 7). However, preincubation of PBMCs with anti-TLR4 antibodies did not modify TNFα production stimulated by HSS23, HSS24, or HSS25 (Fig. 7). Control interactions with irrelevant isotype-matched antibodies for anti-TLR2 and anti-TLR4 antibodies stimulated similar cytokine levels than PBMCs without any treatment (Fig. 7). A similar trend was observed for IL-6 stimulation, which was blocked by anti-TLR2, anti-TLR4, or laminarin when PBMCs were incubated with WT, HSS29, HSS30, HSS20-HSS22 (Fig. 7), whereas TLR4 blocking did not affect the ability of strains HSS23-HSS25 to stimulate IL-6 production (Fig. 7). IL1β and IL-10 production were completely dependent on Dectin-1 activation, whereas TLR2 and TLR4 activation were dispensable (Fig. 7).

3.5. Phagocytosis of RmlD-silenced Sporothrix schenckii by human monocyte-derived macrophages

Monocyte-derived macrophages were generated from human PBMCs by differentiation with human granulocyte–macrophage colony-stimulating factor, as reported (Pérez-García et al., 2016). The immune cell-yeast-like cell interactions were analyzed by flow cytometry, as this strategy has been reported to be useful to separate cells in the early, intermediate, or late stage of the phagocytic process, depending on the fluorescence signals emitted (Hernández-Chávez et al., 2018, Lozoya-Pérez et al., 2018, Lozoya-Perez et al., 2019). The WT and the control strains HSS29 and HSS30 showed similar abilities to interact with the human immune cells, and no differences in any of the three defined stages were observed among these strains (Fig. 8A). The silenced strains HSS20-HSS25 showed similar phagocytosis levels and the number of immune cells in the early, intermediate, or late stage of the phagocytic event was significantly higher than those observed in the WT or control strains (Fig. 8A). Mock reactions, where human monocyte-derived cells were analyzed in the absence of S. schenckii showed only threshold cell numbers for any of the three defined stages (Fig. 8A). Therefore, the S. schenckii-human-derived macrophage interaction was significantly affected by RmlD silencing.

Fig. 8.

Phagocytosis of RmlD-silenced S. schenckii by human monocyte-derived macrophages. In A, Acridine Orange-labeled yeast-like cells were coincubated with human monocyte-derived macrophages at a macrophage-yeast ratio 1:6, for 2 h at 37 °C and 5% (v/v) CO₂. The macrophages were gated by FACS and 50,000 cells were counted per sample. Macrophages interacting with at least one fluorescent yeast cell were included as counted events. Control, macrophages interacting with no yeast-like cells. *P < 0.05 when compared to WT strain. In B, similar experiments to those described in panel A were performed but with human cells previously preincubated with 5 μg mL⁻¹ polymyxin B and 10 μg mL⁻¹ anti-TLR2 (+anti-TLR2), with 5 μg mL⁻¹ polymyxin B and 200 μg mL⁻¹ L-rhamnose (+Rhamnose), 5 μg mL⁻¹ polymyxin B and 10 μg mL⁻¹ anti-TLR4 (+anti-TLR4), or with 5 μg mL⁻¹ polymyxin B and 200 μg mL⁻¹ laminarin (+Laminarin). No treatment refers to cells preincubated only with 5 μg mL⁻¹ polymyxin B. Results correspond to cells in the late stage of phagocytosis. For all cases, 100% corresponds to the system with no treatment and the absolute values were similar to those shown in panel A. Control, macrophages interacting with no yeast-like cells. *P < 0.05 when compared to the no treatment condition of the same strain. † P < 0.05 when compared to the no treatment condition of the same strains and other strains under the same condition. ‡ P < 0.05 when compared to WT strain under the same condition. For both panels, data represent means ± SD from six donors assayed by duplicate. WT, strain 1099–18 ATCC MYA 4821. Strains HSS29 and HSS30 were transformed with pBGgHg; while HSS20-HSS25 with pBGgHg-RmlD. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

To gain insight into the receptors involved in the phagocytic process, we next preincubated the human cells with agents that block specific receptors and assessed the impact on the number of immune cells interacting with fungal cells. We particularly focused on cells in the late stage of the phagocytic process because this provided differences well defined between the WT and silenced strains (Fig. 8A). Results in Fig. 8B showed that preincubation of human monocyte-derived macrophages with anti-TLR2 antibodies decreased the cell number interacting with yeast-like cells by 18.1 ± 5.7% in all strains analyzed, indicating a minor contribution for this receptor during the phagocytic process and in a similar extent in WT, control, and silenced strains (Fig. 8B). Control reactions with an irrelevant isotype-matched antibody showed no effects on the ability of immune cells to phagocyte fungal cells (data not shown). When similar experiments were conducted with immune cells preincubated with anti-TLR4 antibodies, we found about a 25.5 ± 5.1% reduction in the phagocytosis of WT or control strains (Fig. 8B). This effect was partially lost when the silenced strains HSS20-HSS22 were interacting with immune cells (reduction of about 15.1 ± 4.2%) and was lost in the systems where HSS23-HSS25 strains were included (Fig. 8B). These data suggested that the blocking effect of anti-TLR4 antibodies was lost in the silenced strains HSS23-HSS25, most likely because the TLR4 ligand was absent, i.e., rhamnose. To confirm this observation, we preincubated the human cells with rhamnose before challenged with the fungal cells. In pilot experiments, we found that the minimal rhamnose concentration to block fungal phagocytosis was 200 µg mL⁻¹, and increased concentrations did not significantly change this effect (data not shown). After preincubation with 200 µg mL⁻¹ rhamnose, the monocyte-derived macrophages showed a reduction in the ability to phagocyte the WT or control cells similar to that observed when TLR4 was blocked, and this effect was partially lost in the silenced strains HSS20-HSS22 and absent in the strains HSS23-HSS25 (Fig. 8B). The blocking of dectin-1 by laminarin reduced fungal phagocytosis by about 50.3 ± 8.9% and this observation was similar for all the strains analyzed (Fig. 8B). Therefore, different receptors contribute to S. schenckii yeast-like cells phagocytosis, and TLR4 is required to mediate S. schenckii phagocytosis by a rhamnose-dependent mechanism.

3.6. RmlD silencing affected the Sporothrix schenckii virulence in Galleria mellonella larvae

To assess whether the RmlD silencing affected the ability of S. schenckii to kill the host, we infected Galleria mellonella larvae with the different strains used in this study and compared the survival. We selected this alternative model to evaluate virulence because it provides basic lethality parameters of fungal strains similar to that generated with mice (Clavijo-Giraldo et al., 2016, Lozoya-Pérez et al., 2018, Lozoya-Perez et al., 2019, Lozoya-Pérez et al., 2020, Macêdo-Sales et al., 2020). Results showed that groups inoculated with the WT or the control strains HSS29 and HSS30 have similar median survival times of 6.0 ± 0.5 days and killed about 74.5 ± 5.1% of the insect population after the two-week observation period (Fig. 9). No significant differences were observed among these three fungal strains (Fig. 9). Incubation of Galleria with strains HSS20-HSS22 resulted in a median survival time of 12.0 ± 0.2 days killing 40.0 ± 3.3% of insects (Fig. 9). The mortality curves generated with strains HSS23-HSS25 were similar among them, but in contrast to the rest of the strains under analysis, most of the insect population survived at the end of the observation period, with median survivals of more than 15 days and killing only 3.0 ± 0.5% of the infected population (Fig. 9). The defect in the ability to kill insects was unlikely to be associated with defects in fungal cells to grow inside larvae, as similar CFUs were recovered from insects inoculated with the WT, control, or silenced strains (Table 3). The Galleria control group inoculated only with PBS showed 100% survival during the observation period (data not shown). Next, we analyzed whether this increased survival in insects infected with RmlD-silenced strains could be associated with decreased effects on cell damage inside the larva body, measured as the cell-free lactate dehydrogenase found in the hemocele (García-Carnero et al., 2020, Gómez-Gaviria et al., 2020, Lozoya-Pérez et al., 2020), and referred here as cytotoxicity. As expected, high and similar cytotoxicity levels were observed in the insect groups inoculated with the WT or the HSS29 and HSS30 control strains, and this parameter was significantly reduced in the groups inoculated with strains HSS20, HSS21, or HSS22 (Table 3). Interestingly, cytotoxicity in the larvae inoculated with strains HSS23, HSS24, or HSS25 was even lower and significantly different from the levels observed in the other silenced strains, the WT or the control strains (Table 3). The insect immune effectors, hemocytes, and phenol oxidase have been recently reported to be responsive defense elicitors that are modulated depending on the Sporothrix virulence (Lozoya-Pérez et al., 2020). Therefore, we quantified these two parameters in the insect groups inoculated with the silences strains.

Fig. 9.

Mortality of Galleria mellonella larvae inoculated with RmlD-silenced Sporothrix schenckii. Groups containing 30 larvae were inoculated with 1 × 10⁵ yeast-like cells in 10 μL of PBS and survival recorded daily for 15 days. Data are shown in Kaplan–Meier plots. The statistical analysis showed no differences among the WT, HSS29, and HSS30 strains (P = 0.68), but the silenced strains generated survival curves with significantly increased median survival times (P < 0.05). When compared among them, no differences were observed in the curves generated with strains HSS20, HSS21, and HSS22 (P = 0.76), nor with those generated with strains HSS23, HSS24, and HSS25 (P = 0.88). WT, strain 1099–18 ATCC MYA 4821. Strains HSS29 and HSS30 were transformed with pBGgHg; while HSS20-HSS25 with pBGgHg-RmlD.

Table 3.

Fungal burden, cytotoxicity, hemocyte, and phenoloxidase levels in Galleria mellonella inoculated with RmlD-silenced Sporothrix schenckii.

Strain	Colony-Forming Units (×105) a	Cytotoxicity (%) b	Hemocytes (×106) mL−1c	Phenoloxidase d
PBS e	0.0 ± 0.0	15.5 ± 4.8	3.2 ± 0.5	0.6 ± 0.3
WT f	2.5 ± 0.6	81.4 ± 11.4	8.3 ± 0.7	3.0 ± 0.4
HSS29	2.6 ± 0.5	87.5 ± 13.0	7.7 ± 0.8	3.1 ± 0.2
HSS30	2.9 ± 0.7	79.8 ± 9.7	8.1 ± 0.7	3.3 ± 0.4
HSS20	2.6 ± 0.4	56.7 ± 10.5 * †	6.0 ± 0.4 * †	2.2 ± 0.5 * †
HSS21	2.5 ± 0.7	50.4 ± 11.4 * †	5.7 ± 0.3 * †	2.0 ± 0.5 * †
HSS22	2.5 ± 0.6	50.1 ± 10.0 * †	5.6 ± 0.6 * †	2.1 ± 0.3 * †
HSS23	2.8 ± 0.5	28.5 ± 8.8 *	4.2 ± 0.5 *	1.2 ± 0.3 *
HSS24	2.9 ± 0.8	24.1 ± 6.6 *	4.2 ± 0.6 *	1.0 ± 0.3 *
HSS25	2.8 ± 0.9	28.9 ± 11.8 *	4.5 ± 0.4 *	0.9 ± 0.4 *

^aSurviving and dead animals were decapitated, and the hemolymph from dead and survival animals was collected and used to calculate the colony-forming units by incubating on YPD plates.

^bRefers to the free lactate dehydrogenase activity quantified in the cell-free hemolymph from inoculated insects. Data were normalized to the enzyme activity determined in lysed hemocytes, which were considered as the 100%.

^cCalculated from the infected animals' hemolymph.

^dCalculated in the cell-free hemolymph and defined as the Δ_490nm min⁻¹ μg protein ^-1.

^eControl group inoculated only with PBS.

^fWT, strain 1099–18 ATCC MYA 4821. Strains HSS29 and HSS30 were transformed with pBGgHg; while HSS20-HSS25 with pBGgHg-RmlD. * P < 0.05 when compared with the values obtained in animals infected with the WT strain.

^†P < 0.05 when compared with the measurements in animals infected with the WT or HSS23-HSS25 strains.

Results indicated that the highest hemocyte countings and phenoloxidase activity were found in the larvae inoculated with the WT, or the control strains HSS29 or HSS30, whereas groups infected with strains HSS20-HSS22 showed significantly lower levels of both parameters (Table 3). The groups infected with strains HSS23-HSS25 showed similar values of both hemocytes and phenoloxidase activity and these were lower when compared to those generated by the WT, control, or other silenced strains (Table 3). Collectively, these data indicated that RmlD silencing negatively affected the S. schenckii ability to kill G. mellonella larvae and the ability to stimulate insect immune effectors.

4. Discussion

S. schenckii is one of the causative agents of sporotrichosis, an acute or chronic granulomatous subcutaneous mycosis (Mora-Montes et al., 2015). As with other pathogenic fungi, the cell wall modulates the interaction with the host immune system, and contains species-specific components (Netea et al., 2008, Arana et al., 2009, Díaz-Jiménez et al., 2012, Martínez-Álvarez et al., 2014, Martinez-Alvarez et al., 2017, Lopes-Bezerra et al., 2018b, Garcia-Rubio et al., 2020). One of the most distinctive S. schenckii cell wall components is the presence of glycoproteins modified with glycans consisting of mannose and rhamnose, which is a complex of different macromolecules that collectively are named peptidorhamnomannan (Lloyd and Bitoon, 1971, Lopes-Bezerra, 2011, Mora-Montes et al., 2015, Lopes-Bezerra et al., 2018b).

The L-rhamnose biosynthetic pathway has been described in bacteria, plants, and some fungal species (Giraud and Naismith, 2000, Blankenfeldt et al., 2002, Watt et al., 2004, Martinez et al., 2012, Ma et al., 2017, Wagstaff et al., 2019). Our search within the S. schenckii genome for putative orthologous genes encoding for enzymes participating in L-rhamnose synthesis identified only two possible candidates, similar to other fungal systems (Ma et al., 2017). Both RmlB and RmlD are overexpressed in the yeast-like cells (Giosa et al., 2020), correlating with the observations that in this morphology a higher cell wall rhamnose content is found than in hyphae (Martinez-Alvarez et al., 2017). Heterologous complementation in S. mutants showed that the Sporothrix RmlD gene restored the aberrant growth and morphology phenotype of S. mutans ΔrmlD almost to a level similar to that observed in the parental strain, strongly suggesting its participation in L-rhamnose synthesis. It is worth mentioning that not all cells within the population showed a restored phenotype. This may be explained because S. schenckii RmlD has a slower activity rate than the native enzyme because of its inability to form dimers in the bacterial milieu, which is essential for the activity of many enzymes (Giraud and Naismith, 2000). In addition, enzyme performance could also be affected by the nucleotide bound to the monosaccharide, which differs between bacteria and fungi (Martinez et al., 2012). Indeed, it has been reported that the UDP binding to the nucleotide-diphosphate domain enhances the reactivity of NAD⁺ in short-chain dehydrogenase/reductase enzymes, an enzyme family where RmlD is grouped (Kavanagh et al., 2008). Another explanation is that S. schenckii RmlD competes with S. mutans RmlC for the substrate, leading to a less amount of rhamnose in the cells. Nonetheless, these results, along with the reduction in rhamnose content upon RmlD silencing support the notion that this gene participates in the S. schenckii rhamnose synthesis.

The molecular tools for the S. schenckii genomic manipulation are currently scarce, and thus we included the well-characterized methodology for gene silencing (Rodriguez-Caban et al., 2011, Lozoya-Perez et al., 2019). Even though the insertion of the binary plasmid is random, the inclusion of three independent mutants with similar silencing levels allowed us to discard that the observed phenotypes could be due to insertional events rather than the RmlD silencing.

Rhamnose biosynthesis is essential in bacteria such as Streptococcus pyogenes and Mycobacterium smegmatis (Ma et al., 2002, van der Beek et al., 2015) but dispensable in fungi, having minimal contributions on cell morphology or development (Ma et al., 2017, Santhanam et al., 2017). In agreement with these observations, we found here that RmlD silencing, and therefore rhamnose depletion did not affect the doubling time, cell and colony morphology, and ability to undergo dimorphism. It is thought that binding of L-rhamnose to N-linked and O-linked glycans is a terminal event (Lopes-Bezerra, 2011). This would offer a possible explanation of the lack of a severe phenotype in RmlD-silencing mutant cells. In B. cinerea, upon disruption of the RmlD the accumulation of UDP-4-keto-6-deoxy-glucose, the RmlD substrate, led to defects in cell morphology (Ma et al., 2017). The apparent discrepancy of this observation with our results could be explained by the fact that residual RmlD transcription was found in the HSS23-HSS25, and this could avoid the accumulation of toxic levels of this compound. Alternatively, UDP-4-keto-6-deoxy-glucose may indeed accumulate in the S. schenckii RmlD-silenced strains and this species simply has a different sensitivity degree to UDP-4-keto-6-deoxy-glucose when compared to that observed in B. cinerea.

The cell wall integrity pathway is a signaling network that allows the fungal cell to respond and adapt to stresses that affect the wall, including permanent modifications such as the ones generated by mutations (Bates et al., 2006, Mora-Montes et al., 2007, Valiante, 2017). The increment in β-1,3-glucan content and exposure at the cell surface is likely a consequence of this pathway activation upon reduction of rhamnose content in the silenced mutants. In line with this, Candida spp. mutant strains, where mannan elaboration is disrupted, showed increased β-1,3-glucan content as a compensatory mechanism to the weakness of the cell wall outer layer (Bates et al., 2005, Bates et al., 2006, Mora-Montes et al., 2007, Mora-Montes et al., 2010a, Navarro-Arias et al., 2016, Pérez-García et al., 2016, Hernández-Chávez et al., 2019), and interestingly, the disruption of the N-linked glycan outer chain elaboration in S. schenckii had as a consequence a reduction in rhamnose content and increment in glucan levels (Lozoya-Perez et al., 2019), suggesting that rhamnose levels attached to N-linked glycans could be the sensors that trigger activation of the cell wall integrity pathway. A second possible explanation to this observation, which does not exclude the previous one, is that both β−1,3-glucan and rhamnose biosynthetic pathways share UDP-glucose as the same precursor, and therefore, disruption of the rhamnosylation pathway might increase the precursor levels that are used in the synthesis of this wall structural polysaccharide. We currently do not have a feasible explanation for the fact that only one polysaccharide was modified in these cells and this is an observation that remains to be addressed. Despite rhamnose reduction was observed in both N-linked and O-linked glycans from RmlD-silenced strains, this did not significantly change the total sugar content in N-linked glycans, suggesting rhamnose is not the main component of these structures, and supporting the notion that it is a monosaccharide that modifies the branches attached to the outer chain (Lopes-Bezerra, 2011, Teixeira et al., 2014, Lopes-Bezerra et al., 2018a). On the contrary, rhamnose reduction led to a significantly minor total sugar content in the O-linked glycans from the RmlD-silenced strains, suggesting that rhamnose has a significant contribution to their structure, as previously reported for the case of the peptidorhamnomannan O-linked glycans (Lopes-Bezerra, 2011). Thus, we hypothesize that O-linked glycans in the silenced mutant are shorter than those found in the WT strain, making the cell wall outer layer thinner and less bulky, allowing β-1,3-glucans to be more exposed at the surface

The RmlD-silenced mutants showed a differential ability to stimulate cytokine production by human PBMCs, and as consequence lower TNFα and IL-6 levels but higher IL-1 β and IL-10 production were observed when compared to WT cells. The use of blocking agents suggested that TLR4 is the receptor for rhamnose present in the S. schenckii cell wall, and this observation is in line with a previous report, which indicates that this immune receptor recognizes P. boydii rhamnomannan (Figueiredo et al., 2010). The production of both TNFα and IL-6 by human PBMCs depends mainly on the engagement of dectin-1, TLR2, and TLR4 with their ligands (Martinez-Alvarez et al., 2017), and most likely in a co-stimulatory event (Dennehy et al., 2008, Netea et al., 2008, Reid et al., 2009). Thus, the reduction of the TLR4 ligand in the RmlD-silenced mutants could account for the observed changes in the levels of these proinflammatory cytokines. When interacting with S. schenckii, dectin-1 is the main PBMCs receptor required for IL-1 β and IL-10 production (Martinez-Alvarez et al., 2017), thus, it is likely that the increased β-1,3-glucan content and exposure at the cell surface positively influenced the ability of the mutant strains to stimulate both IL-1 β and IL-10. Dectin-1 is predominantly expressed in macrophages (Taylor et al., 2002), and its engagement with the ligand is sufficient to trigger phagocytosis (Heinsbroek et al., 2008). Therefore, our finding reported here about the Sporothrix-macrophage interaction is likely explained by dectin-1 dependent phagocytosis. However, our observations also indicate that the TLR4-L-rhamnose interplay, and to a lesser extent the TLR2-ligand interaction are also relevant during the sensing of this fungal species by human monocyte-derived macrophages.

The RmlD-silenced mutants showed virulence attenuation in the G. mellonella infection model, suggesting that rhamnose is required to display the ability to kill larvae. This virulence attenuation is unlikely to in vivo growth defect, as similar fungal burdens were found in WT, control, and silenced strains. This observation is in line with those in B. cinerea and Verticillum dahlia, where rhamnose-depleted mutants lost the ability to colonize and cause damage to the host (Ma et al., 2017, Santhanam et al., 2017). It is worthy of mention that these are phytopathogens, and our study here with S. schenckii reports for the first time a link between virulence and rhamnose in a mammalian’s fungal pathogen. The fungal adhesion to host cells is crucial for colonization and dissemination, and in S. schenckii exists a correlation between the virulence and the interaction of adhesins with fibronectin (Teixeira et al., 2009). The loss of rhamnose from the S. schenckii cell wall could affect the cell’s ability to adhere properly to the host cells, as both the peptidorhamnomannan and rhamnose have been described as part of the adhesive mechanism molecules (Pinto et al., 2004, Santhanam et al., 2017), possibly explaining the virulence attenuation upon RmlD silencing.

In conclusion, S. schenckii RmlD is involved in the synthesis of L-rhamnose and the presence of this carbohydrate at the cell wall is important during the interaction of this fungus with human PBMCs and monocyte-derived macrophages, in TLR4-dependent mechanisms. Moreover, RmlD is required for S. schenckii virulence.

CRediT authorship contribution statement

Alma K. Tamez-Castrellón: Data curation, Formal analysis, Writing - original draft. Samantha L. van der Beek: Data curation, Formal analysis, Writing - original draft, Writing - review & editing, Writing - review & editing. Luz A. López-Ramírez: Data curation, Formal analysis, Project administration, Writing - original draft, Writing - review & editing. Iván Martínez-Duncker: Data curation, Formal analysis, Writing - original draft, Writing - review & editing. Nancy E. Lozoya-Pérez: Data curation, Formal analysis, Writing - original draft, Writing - review & editing. Nina M. van Sorge: Data curation, Formal analysis, Funding acquisition, Writing - original draft, Writing - review & editing. Héctor M. Mora-Montes: Data curation, Formal analysis, Funding acquisition, Project administration, Writing - original draft, Writing - review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.