C. albicans SC5314 (kindly donated by Professor Ruoyu Li from Peking university, China), C. tropicalis ATCC1369 and S. aureus ATCC25923 (both purchased from ATCC) were used in this research. YPD and BHI plus 1% glucose (OXOID, Basingstoke, England) medium were used for the growth of C. albicans (or C. tropicalis) and S. aureus, respectively. RPMI1640 medium (Thermo Fisher Scientific, USA) buffered to a pH of 7.0 with 0.165 M MOPS (Sangon Biotech, Shanghai, China) was used for growing the dual-species biofilms between Candida species and S. aureus.

The chemical components of lemongrass (extracted from C. flexuosus) essential oil (doTERRA, USA) were analyzed by gas chromatography-mass spectra (GC-MS) system (Agilent gas chromatograph 6,890 coupled with Agilent 5973 mass selective detector, USA). The gas chromatographic separation was performed on a HP-INNOWAX capillary column (30 m × 0.25 mm i.d. with 0.25 μm film thickness, Agilent Technology, USA). The injector temperature was set at 250°C and high-purity helium was used as the carrier gas with flow rate of 1 ml/min. The sample was injected with a 1 μl volume and performed in a split mode (split ratio 30:1). The oven temperature was programmed as follows: initial temperature was maintained at 40°C for 2 min and then the temperature was increased to 250°C at 6°C/min and subsequently held for 10 min. The MS was performed in electron ionization mode (EI) with 70 eV ionization energy. The ion source temperature was set at 230°C with scan range from 30 to 300 m/z. The compounds were identified according to the NIST 2014 mass spectral library which is a standard mass spectrometry reference database released by National Institute of Standards and Technology (NIST) in 2014.

Minimal inhibitory concentration (MIC) was performed by serial microdilutions based on previously described method with minor modifications (Miao et al., 2019; Niu et al., 2020). Specifically, Candida and staphylococcal cell suspensions were diluted to 1 × 10⁵ cells/ml in YPD and BHI respectively. Then aliquots of 100 µl of each species were separately transferred into each 96-well microtiter plates. Next, lemongrass essential oil was added in YPD or BHI to obtain final concentrations in the range of 0.0097–10 μg/ml; Similarly, citral was added in YPD or BHI to obtain final concentrations in the range of 0.0078–8 μg/ml. Following 24 h incubation at 37°C, the MIC value was determined as the lowest concentration (volume percent, v/v %) of lemongrass essential oil or citral which inhibited the visible growth of each species after overnight incubation. For each treatment condition, three replicates were performed and the experiment was repeated 3 times.

All strains were stored in 10% glycerol at -80°C. C. albicans or C. tropicalis strain was inoculated on the Sabouraud’s Dextrose Agar (SDA) and S. aureus was grown on the Trypticase Soy Agar (TSA, OXOID, Basingstoke, England) overnight. Then single colony of Candida strains was inoculated in YPD and S. aureus was grown in BHI at 37°C overnight. The cultures were diluted to OD₆₀₀ = 0.01 with RPMI1640, and seeded into 96-well microplates for 24 h at 37°C. The medium was refreshed with lemongrass essential oil or citral (mixture of cis-citral and trans-citral, MedChemExpress, USA) and co-incubated for another 24 h. Then, antimicrobial efficacy against dual-biofilms was evaluated in three aspects. Firstly, biofilm biomass reduction was quantified by the crystal violet (CV) staining. Briefly, dried biofilms were stained with 0.2% CV solution for 45 min, washed with water, and destained with 95% ethanol for 45 min. OD₅₇₀ were measured in a microplate reader (BioTek, USA). The biomass reduction was calculated by the following formula. Reduction ratio (%) = (OD_untreated-OD_treated)/OD_untreated × 100%. Furthermore, reduction ratio of viable cells within biofilm was assessed by Cell Counting Kit-8 (CCK-8) method as previously described (Tan et al., 2019).The depth of color is directly proportional to cellular viability and OD₄₅₀ was measured by microplate reader. The viability reduction ratio was calculated by the same formula as above. Besides, the reduction in viable counts CFU (Colony Forming Units) was detected by plate assay. For this, after treatment, biofilms were dissociated from the surface into PBS and then added onto YPD + vancomycin (2 mg/L) plates (select C. albicans) and TSA+ amphotericin B (2.5 mg/L) plates (select S. aureus) by serial-dilutions plating (Qu et al., 2016). The effectiveness to kill each species of biofilm was determined by calculating the direct reduction of viable cells [Reduction ratio % = (untreated CFU - treated CFU)/untreated CFU ×100%]. For each treatment condition, three replicates were performed and the experiment was repeated 3 times.

First, dual-species biofilms were grown on coverslips for 24 h, and supernatant was aspirated, and replaced with fresh RPMI-1640 containing lemongrass essential oil (0.0708% and 0.3125%, v/v) or citral (0.125% and 0.5%, v/v) for biofilm compositional analysis and lemongrass essential oil (0.3125%, v/v) or citral (0.5%, v/v) for viability analysis, and then grown for 24 h. Afterwards, biofilm components including carbohydrates, proteins and nucleic acids were visualized as previously described with modification (Shin and Eom, 2019). Briefly, the dual-species biofilms on coverslips were immersed with 4% (v/v) paraformaldehyde for 1 h and dried for 30 min. The amino groups of dual-species biofilm were stained 10 μg/ml fluorescein isothiocyanate isomer I (FITC) for 1 h. To visualize carbohydrates, 1 μg/ml of concanavalin A-Alex Fluor 594 conjugate (Con-A), which reacts with mannose and glucose of biofilm, was incubated for 30 min. Finally, nucleic acids were stained with 5 μg/ml of 4,6-diamidino-2-phenylindoldihydrochloride (DAPI) which binds to double-stranded DNA for 45 min. At the end of each staining, the biofilms were washed with PBS to remove dye residues. Besides, the biofilms were stained with the LIVE/DEAD BacLight Viability Kit (Thermo Fisher Scientific, USA) for 15 min according to the manufacturer’s protocol. Cellular fluorescence was evaluated with CLSM (Leica, SP8, Germany) and the image data was analyzed by Leica LAS AF Lite and ImageJ software (NIH, USA). For the image quantification, the fluorescence data were taken from four randomly selected fields with same size per sample and one of the four fields was chosen to represent the phenotype. In addition, the average of biofilm thickness was quantified by collecting the data from four z-stacks for each experiment condition. The CLSM experiment was independently repeated 3 times.

The C. albicans/S. aureus biofilms were first fixed with 2.5% glutaraldehyde and next dehydrated in 30%, 50%, 70%, 80%, 90%, and 95% dilution series of ethanol for 15 min, respectively. Next the samples were immersed in isoamyl acetate for 15 min and then attached to metallic stubs by carbon stickers and sputter-coated with gold. Finally, the morphology of prepared samples was observed by Hitachi SU8100 SEM microscopy (HITACHI, Japan). The representative SEM images were chosen from three biological replicates.

1 ml overnight culture of bacteria and fungi were diluted to OD₆₀₀ = 0.01, seeded into a 12-well polystyrene microplate, and grown for 24 h. After 24 h, the supernatant was refreshed with RPMI-1640 containing 0.5% citral, wells without treatment were set as control. After 24 h, fungal and bacterial mRNA were extracted according to Omega Yeast RNA kit and the Bacteria kit (Omega BioTek, USA) respectively. cDNA was obtained by FastKing RT Kit (Tiangen, China). The gene expression of C. albicans and S. aureus, were detected by SYBR Green quantitative Real-Time PCR assay (SuperReal PreMix Plus, Tiangen, China) and calculated using the formula 2^-ΔΔCt. The reaction was run on ABI Step One qPCR system (Applied Biosystem, USA) as follows: initial step at 95°C for 15 min, followed by 40 cycles at 95°C for 10 s, 60°C for 30 s. Eighteen seconds rRNA was chosen as internal reference gene for fungi and 16s rRNA for bacteria to normalize the data respectively. The primers ( Table 1 ) were synthesized by Shangya biotech company in China. qPCR analyses of each gene were performed in three biological replicates, each with three technical replicates.

All experiments were performed at least 3 times with statistical significance of different treatment groups calculated by one-way analysis of variance (ANOVA, Bonferroni multiple-comparison test) or by students’ t test. Difference were considered significant if P value was less than 0.05. Statistical analyses and graphs were made using Microsoft Excel and GraphPad Prism 7.

The biologically active component of lemongrass essential oil is citral, which is mixture of two isomeric acyclic monoterpene aldehydes (neral and geranial). Compositional analysis by GC-MS identified a total number of 19 compounds with content exceeding 0.05%, making up ~99.6% of the lemongrass (C. flexuosus) essential oil ( Table 2 ). According to the results, the primary components identified included ~29.4% geranial (trans-citral, α-citral) and ~30.4% neral (cis-citral, β-citral) ( Figure 1 ). This is consistent with a previous study reporting citral as the dominant component of lemongrass essential oil, accounting for more than 60% of total compounds (Hadjilouka et al., 2017). Other identified high-content ingredients included caryophyllene (~25.4%) and indan-1,3-diol monoacetate (~7%). These results therefore validated the composition of lemongrass essential oil used in this study with citral confirmed as the most abundant compound.

To investigate the antimicrobial action of lemongrass essential oil and citral against planktonic phase of Candida species and S. aureus, the MIC values (v/v%) of each species and dual-species were determined ( Table 3 ). The MIC of lemongrass essential oil against planktonic C. albicans, C. tropicalis and S. aureus were 0.0781%, 0.039%, and 0.0781%, respectively. The MIC of citral against C. albicans, C. tropicalis, and S. aureus were 0.0313%, 0.0156%, and 0.0313%, respectively. Regarding the dual-species planktonic Candida species and S. aureus, The MIC of lemongrass essential oil against C. albicans/S. aureus and C. tropicalis/S. aureus were 0.1563% and 0.0781%, respectively. The MIC of citral against C. albicans/S. aureus and C. tropicalis/S. aureus were 0.125% and 0.0313%, respectively. Therefore, both lemongrass essential oil and citral at low concentrations were enough to inhibit growth of planktonic S. aureus and Candida species.

To evaluate the efficacy of lemongrass essential oil and citral against dual-species biofilms formed by Candida species and S. aureus, crystal violet (CV) staining and CCK-8 test were performed to quantify biofilm biomass and viability respectively. The volume percent (v/v%) of lemongrass essential oil at the range from 0.097% to 10% were used to assess its antibiofilm activity. As present in Figures 2A, C , treatment by 0.3125% lemongrass essential oil significantly reduced the biofilm biomass by ~80% and cell viability by ~85%, showing the highest activity to eliminate biofilm. Interestingly, treatment with concentrations more than 0.3125% actually diminished the antibiofilm effects although the effects on cell viability remained maximal beyond 0.3125%. Similarly, experiments conducted with citral showed 0.5% treatment strongly reduced the biofilm biomass by ~74% but higher concentrations led to reduced activity ( Figure 2B ). At this concertation citral reduced cell viability by ~87%, but like the results with lemongrass essential oil, higher concentrations were similarly effective against biofilm viability ( Figure 2D ).

The next series of experiments turned to examine the effects of the two agents on dual-species biofilms formed by C. tropicalis and S. aureus. Following the same approach, treatment with lemongrass essential oil significantly reduced the biofilm biomass, and in this instance the inhibition was greater with increasing concentrations with a maximal inhibition of ~80% at 10% ( Figure 3A ). Intriguingly, cell viability measurements showed again that 0.3125% was a threshold concentration for affecting dual-species biofilms of C. tropicalis and S. aureus ( Figure 3C ). Comparative assays with citral showed treatment with 8% citral maximally reduced biofilm biomass by ~82%, and with 0.25% or higher concentrations of citral evidently led to maximal cell viability reduction by ~71% ( Figures 3B, D ).

The preceding data provided evidence that lemongrass essential oil and citral could reduce both biofilm biomass and viability of C. albicans/S. aureus dual-species biofilms. However, these assays did not differentiate the effectiveness of these agents against the individual species inside the biofilm. To ascertain this, CFU measurements were undertaken after treating the dual-species biofilms with either lemongrass essential oil or citral ( Figures 4A, B ). Treating the biofilms with sub-optimal and optimal concentrations of lemongrass essential oil, 0.0781% and 0.3125%, respectively, resulted in killing rates of ~67% and ~97% for C. albicans in comparison to ~62% and ~95% for S. aureus ( Figures 4C, E ). Comparable experiments with citral at 0.125% killed ~70% C. albicans and ~60% S. aureus, while treatment with 0.5% citral killed ~96% C. albicans and ~94% S. aureus ( Figures 4D, F ).

To investigate whether lemongrass essential oil and citral impact the architecture of C. albicans/S. aureus dual-species biofilms, three major components of biofilm matrix, namely nucleic acids, proteins and carbohydrates, were examined by CLSM. As shown in Figure 5A , treatment of either lemongrass essential oil or citral resulted in obvious changes in the biofilm matrix. In particular, the dense network of filaments in untreated biofilms became more sparsely distributed after treatment. Treatment with optimal dose of lemongrass essential oil (0.3125%) exerted strong effect, substantially impairing the coadhesion between C. albicans and S. aureus compared to the sub-optimal treatment (white square versus red square in Figure 5A ). However, compared to the sub-optimal treatment, application of optimal dose of citral (0.5%) also reduced the coadhesion while the effect was not significant as the lemongrass essential oil did (white square versus yellow square in Figure 5A ). These observations were therefore largely consistent with the treatment outcomes, with lemongrass essential oil being marginally more effective than citral treatment, although both agents reduced matrix density.

On the basis of CLSM images, alteration of nucleic acid, protein and carbohydrate contents were quantified after treatment by these two agents. As shown in Figure 5B , compared to the sub-optimal treatment (0.0781% lemongrass essential oil and 0.125% citral), application of 0.3125% lemongrass essential oil and 0.5% citral significantly reduced the biovolume of nucleic acids by ~86% and ~63% respectively. Furthermore, these optimal treatments also caused sharp reductions in proteins by ~89% and ~77% ( Figure 5C ), as well as carbohydrates by ~68% and ~44% ( Figure 5D ), respectively. Overall, these treatments reduced the total biovolume of three main biofilm components by ~85% and ~65% ( Figure 5E ), respectively. Again, based on the measurable changes in biofilm architecture, these results implied that lemongrass essential oil showed slight advantages over citral in inhibiting the compositional biovolume of dual-species biofilms.

Next, live/dead cell staining was performed by CLSM to further investigate effects of lemongrass essential oil and citral on the biofilm structure. As shown in Figure 6A , exposure of biofilms to 0.3125% lemongrass essential oil and 0.5% citral resulted in striking reductions in viable cells present in biofilm. As with the biovolume measures, the CLSM images could also be analyzed to estimate the cell viability rates within the treated biofilms. Image analysis revealed that 0.3125% lemongrass essential oil and 0.5% citral reduced the percent of viable cells from ~ 92% to ~14% and ~13%, respectively ( Figure 6B ). Moreover, the values calculated by CLSM were fully consistent with the activities of lemongrass essential oil and citral measured in both CCK-8 and CFU assays (refer Figures 2 and 4 ). In addition, 0.3125% lemongrass essential oil and 0.5% citral treatments also dramatically reduced the biofilm thickness from ~22 μm to ~7 μm and ~6 μm, respectively ( Figure 6C ), consistent with the reduced content of three major matrix components after treatment.

To visualize the alterations of micro-architecture of dual-species biofilms after exposure to lemongrass essential oil and citral, SEM microscopy was applied. As shown in Figure 7 , following treatment of 0.5% citral, the number of adherent bacterial cells was obviously decreased. Additionally, compared with the hyphae without treatment, treated hyphae underwent clearly changes with thinner filaments, wrinkled surface and fragile hyphae segmentations (indicated by yellow arrows in the figure). Treatment by 0.3125% lemongrass essential oil resulted in even further reduction of adherent bacterial cells, as well as shrunk surface compared to that treated with 0.5% citral (indicated by red arrows in the figure). These results were in line with the CLSM observations shown in Figure 5 , indicating both lemongrass essential oil and citral could impose inhibitory effects on the coadhesion between C. albicans and S. aureus, as well as changing the hyphal morphology of C. albicans in dual-species biofilms.

To dissect molecular responses of each species to citral, qPCR assay was performed to examine the transcriptional level of biofilm-related genes. The relative fold change in gene expressions of C. albicans and S. aureus were normalized to each housekeeping gene 18s rRNA and 16s rRNA respectively and calculated by the DeltaDeltaCt method. Comparing gene expression changes in dual-species biofilms between untreated and 0.125% citral treated samples showed remarkable downregulated expression of Candida adhesion related genes als3 and hwp1 by ~ 95% and ~ 90%, respectively ( Figure 8A ). Furthermore, SAP genes including sap1, sap2, and sap3 which encode secreted aspartyl proteinases were significantly repressed by ~94.1%, ~96.2%, and ~96.8%, respectively. These data indicated citral counteracted C. albicans in dual-species biofilms through exerting effects on Candida adhesins and virulent factors.

As for the transcriptional levels of S. aureus in dual-species biofilms, treatment by citral critically influenced several biosynthetic pathways and signaling system of S. aureus. As shown in Figure 8B , treatment with 0.5% citral led to dramatically downregulation of central toxin regulator AgrA and α-toxin (encoded by hla gene) by ~ 69.8% and ~ 87.3%, respectively. Moreover, expression of acpP, accA and fapR, genes encoding crucial components involved in fatty acids biosynthesis, were also prominently repressed (~70.5%, ~84.3%, and ~76.2% reductions, respectively). In addition, expressions of MurE and Pbp1, which are essential for the peptidoglycan biosynthesis, were downregulated by ~33.2% and ~54%, respectively. Thus, these data indicated that citral inhibited S. aureus in dual-species biofilms by affecting quorum sensing system, fatty acids and peptidoglycan biosynthesis.