Klebsiella pneumoniae prevents spore germination and hyphal development of Aspergillus species

Different bacteria and fungi live as commensal organisms as part of the human microbiota, but shifts to a pathogenic state potentially leading to septic infections commonly occur in immunocompromised individuals. Several studies have reported synergistic or antagonistic interactions between individual bacteria and fungi which might be of clinical relevance. Here, we present first evidence for the interaction between Klebsiella pneumoniae and several Aspergillus species including A. fumigatus, A. terreus, A. niger and A. flavus which cohabit in the lungs and the intestines. Microbiological and molecular methods were employed to investigate the interaction in vitro, and the results indicate that Klebsiella pneumoniae is able to prevent Aspergillus spp. spore germination and hyphal development. The inhibitory effect is reversible, as demonstrated by growth recovery of Aspergillus spp. upon inhibition or elimination of the bacteria, and is apparently dependent on the physical interaction with metabolically active bacteria. Molecular analysis of Klebsiella-Aspergillus interaction has shown upregulation of Aspergillus cell wall-related genes and downregulation of hyphae-related genes, suggesting that Klebsiella induces cell wall stress response mechanisms and suppresses filamentous growth. Characterization of polymicrobial interactions may provide the basis for improved clinical management of mixed infections by setting the stage for appropriate diagnostics and ultimately for optimized treatment strategies.

In our studies, we aimed to investigate the interactions between select opportunistic fungal pathogens including different Aspergillus species and the clinically important bacterium Klebsiella pneumoniae. Both species cohabit in various regions of the body including particularly the gut and the lungs 1,2,52,53 . Development of Aspergillus species encompasses the formation of spores or conidia and hyphal filaments. Under favorable environmental conditions, spores germinate into long hyphae which are responsible for tissue invasion and escape from the immune system [54][55][56][57] . Aspergillus species have the capacity to form biofilms in which hyphae play a leading role. The architecture of the lungs in conjunction with hyphal development of Aspergilli makes it difficult for the immune system to clear the fungal pathogens [57][58][59][60] . Next to C. albicans, A. fumigatus is the most prevalent cause of fungal infection in immunocompromised patients 56,61,62 , and is responsible for approximately 90% of invasive aspergilloses, based on clinical reports 63,64 . More recently, however, shifts from A. fumigatus to non-fumigatus Aspergillus species have been observed, involving particularly A. terreus, A. niger and A. flavus [65][66][67][68][69] . On the bacterial side, K. pneumoniae is an emerging pathogen displaying resistance to antibiotic treatment which has been associated with several nosocomial outbreaks 53,[70][71][72][73][74][75][76][77][78][79][80][81][82][83] . The antimicrobial resistance and virulence of numerous clinical strains of K. pneumoniae have been associated with the presence of plasmids carrying resistance genes, the hypermucoviscosity phenotype, capsular polysaccharides and the capacity to form biofilms 75,[84][85][86][87][88][89][90] . Bacteria and fungi are constantly exposed to stress conferred particularly by the host microenvironment and other pathogens. The cell wall is the first point of contact between microorganisms and the host or other pathogens. The cell wall plays an important role in mediating interactions with the external environment relevant for nutrient diffusion and molecule-based signalling. Simultaneously, it protects the cells from oxidative or osmotic stresses, and modulates the response to antimicrobial drugs [91][92][93][94][95] . Fungi are capable of adapting their cell walls in response to stress by activating multiple mechanisms directed towards repair or compensation for cell wall damage. In response to stress, C. albicans was shown to activate the MAPK and Ca 2+ /Calcineurin pathways, leading to upregulation of genes involved in the cell wall assembly, and Aspergillus species respond to stress in a similar manner [96][97][98][99][100][101][102][103][104] . It is important to point out that fungal spores and hyphae display different structures and compositions of the cell wall which can induce differential immune responses by the host 105,106 .
In the present study, we sought to investigate the interactions between four Aspergillus species including A. fumigatus, A. terreus, A. niger and A. flavus, and different strains of K. pneumoniae with low or high capacity of biofilm formation. Our studies provide new insights into the biological behavior of Aspergillus and K. pneumoniae in co-culture, by unravelling the type of interaction and response to stress. Our observations highlight the importance of identifying the presence of polymicrobial infections and potential interactions between the pathogens with regard to optimized diagnostic approaches and appropriate antimicrobial treatment.

Results
K. pneumoniae inhibits spore germination and hyphal development of Aspergillus species. The in vitro interaction between K. pneumoniae and several Aspergillus species, including A. fumigatus, A. terreus, A. niger and A. flavus was characterized. K. pneumoniae and Aspergillus were grown alone and in co-culture for 24 and 48 h, as outlined in the Methods section. The level of inhibition of fungal spore germination was assessed by qPCR at 24 h ( Fig. 1) and imaged by confocal microscopy (Fig. 2). The suppression of Aspergillus growth by K. pneumoniae strains is shown in Fig. 1. Our data have shown that, upon contact with K. pneumoniae, Aspergillus spores are not developing into hyphae (Fig. 1A). This is reflected by the unchanged DNA content of Aspergillus in the co-culture which corresponds to the initial fungal cell loads. In addition, this inhibition is independent of the biofilm forming capacity of K. pneumoniae strains, since the same effect was observed when using low-biofilm (ATCC 13883) and high-biofilm (ATCC 700603) forming strains. As shown in Fig. 1A, the inhibitory effect was observed for all Aspergillus species tested, indicating that this effect is Aspergillus species-independent.
Due to the functional importance of hyphae for tissue invasion and escape from the immune system, we have addressed the effect of K. pneumoniae on hyphal development of the indicated Aspergillus species. We sought to understand whether K. pneumoniae also has an effect on hyphal development and if the inhibitory effect still occurs at the hyphal stage. The results have shown that, in the presence of K. pneumoniae, Aspergillus species were not able to progress with hyphal development by demonstrating significantly decreased DNA contents compared to Aspergillus in single cultures (Fig. 1B), revealing that the inhibitory effect of K. pneumoniae on Aspergillus species is independent of the fungal growth stage.

Biofilm formation of Aspergillus species is affected by the presence of K. pneumoniae. Both
K. pneumoniae and Aspergillus species used in this study are capable of forming biofilms. Based on the observation that K. pneumoniae strains are able to suppress Aspergillus growth and hyphenation, we sought to investigate whether low-and high-biofilm forming strains of K. pneumoniae might also affect the biofilm formation of Aspergillus. A confocal microscopy approach was used to measure the biofilm thickness of bacteria and fungi growing alone and in co-culture. The results have shown that, upon contact with K. pneumoniae strains, the biofilm thickness of A. fumigatus was significantly decreased compared to A. fumigatus growing alone. These observations are supported by the confocal images showing that A. fumigatus in co-culture remained in the form of spores, independently of the K. pneumoniae strain tested ( Fig. 2A). In addition, the high-biofilm forming strain of K. pneumoniae showed thicker biofilms when grown in the presence of A. fumigatus ( Fig. 2A, right plot). We further investigated whether K. pneumoniae strains could also influence the biofilm formation of pre-formed A. fumigatus biofilms. To address this question, A. fumigatus spores were pre-germinated for 12 h, followed by  Inhibition of Aspergillus growth by K. pneumoniae is dependent on direct contact. Based on the documented antagonistic effect of K. pneumoniae strains against the development and growth of Aspergillus species, we have addressed the contact-dependence of the interaction. Transwell plates permitting the growth of bacteria and fungi in physically separated compartments but allowing exchange of diffusible molecules through a porous membrane were employed for the analysis. In the wells (lower compartments), K. pneumoniae strains were grown alone or in co-culture with A. fumigatus, whereas in the inserts with the porous membrane (upper compartments), A. fumigatus was grown alone. This strategy allowed us to investigate whether the inhibitory effect on A. fumigatus was mediated via direct contact with K. pneumoniae or by the secretion of molecules conferring independence from physical contact. The results have shown that A. fumigatus was able to grow in the upper To further assess the putative secretion of inhibitory molecules, we investigated the effect of the bacterial supernatant (SN) on A. fumigatus growth. Supernatants of K. pneumoniae grown alone or in co-culture in biofilm mode were obtained at 12 h and 24 h. Collection of supernatants at different time points was performed to determine if the inhibitory effect is growth phase-dependent, because the secretion of molecules might vary during the bacterial growth cycle. The supernatants were added to cultures of A. fumigatus. After 48 h of incubation, the results have shown that, regardless of the culture conditions or growth phase of the bacteria, various preparations of bacterial SN were unable to inhibit growth of any of the fungal Aspergillus species studied (Fig. S1). By contrast, in the presence of bacteria, either in exponential or stationary growth phase, the inhibitory effect on Aspergillus growth was observed (Fig. S2). By contrast, co-cultures with heat-or UV-killed K. pneumoniae did not show any reduction of fungal growth, demonstrating the requirement of live bacteria for the inhibitory effect (Fig. 5A). Aspergillus species remain viable upon interaction with K. pneumoniae. In order to assess whether the observed inhibitory effect of K. pneumoniae on the growth and development of various Aspergillus species is sustainable or reversible, the fungi were grown alone or in co-culture with K. pneumoniae. After 24 h growth in biofilm mode, aliquots of the single cultures and the co-cultures were streaked onto YPD agar plates containing the antibiotic kanamycin to prevent bacterial growth. The cultures were evaluated after 48 h of incubation, and confirmed efficient elimination of the bacteria. The co-cultures of K. pneumoniae strains with Aspergillus species revealed recovery of fungal growth upon exposure to kanamycin (Fig. 4), demonstrating that the fungi maintain their viability and ability to re-initiate growth once the inhibitory effect of the bacteria is eliminated. The same effect was observed when using other antibiotics including colistin or tetracycline which target different structures of the bacteria (Fig. S3). Confocal microscopy was used to assess quantitatively the growth recovery of A. fumigatus upon interaction with K. pneumoniae and subsequent treatment with an antibiotic. Co-cultures of A. fumigatus and K. pneumoniae were initiated at time zero, followed by antibiotic treatment with colistin at 6 h for elimination of the bacteria. After 24 h growth in co-culture or alone, measurements of the biofilm thickness of A. fumigatus revealed that, in the antibiotic-treated cultures, A. fumigatus biofilms were similar to A. fumigatus in single culture. The respective confocal microscopy images also showed hyphal development, contributing to the biofilm thickness. In addition, the Petri dishes also showed growth recovery of A. fumigatus in co-culture after antibiotic treatment, similar to cultures of the fungus growing alone (Fig. 5B).
Growth inhibition of Aspergillus species by K. pneumoniae is not affected by nutrient shortage. Nutrient competition for the carbon source is a potential limiting factor for fungal growth during the interaction of K. pneumoniae with Aspergillus spp. To address this notion, the interaction between A. fumigatus and K. pneumoniae (ATCC strain 700603), YPD containing a standard glucose concentration of 2% was compared with YPD containing 4% glucose. The microorganisms were grown alone and in co-culture for 48 h in YPD containing the indicated glucose concentrations. No major differences in fungal growth were observed during the interaction with bacteria (Fig. S4), suggesting that the availability or lack of carbon was not a relevant factor for the antagonistic effect exerted by the bacteria on fungal growth.
K. pneumoniae renders A. fumigatus sensitive to cell wall stress and induces upregulation of cell wall-related genes. We investigated whether K. pneumoniae could induce cell wall stress in A. fumigatus and render the fungus more or less susceptible to stressors including antifungal drug treatment. After 24 h of growth in single culture or co-culture, A. fumigatus and K. pneumoniae were isolated by flow sorting. Equal numbers of sorted cells were seeded into 12-well agar plates containing a combination of the antibiotics kanamycin and colistin to prevent bacterial growth. Moreover, the plates contained different concentrations of oxidative (H 2 O 2 ), osmotic (sorbitol, sodium dodecyl sulfate (SDS)) and cell wall stressors (Calcofluor white (CFW)) as well as antifungal drugs (Fig. 6A). The sensitivity to these stresses was compared between A. fumigatus grown alone or upon exposure to K. pneumoniae, and the latter constellation revealed at least two-fold greater sensitivity (Fig. 6B). Additionally, three different classes of antifungal drugs including Amphotericin B, Voriconazole and Micafungin were tested to assess the sensitivity of A. fumigatus grown under the conditions indicated above. However, no differential sensitivity has been observed (Fig. 6C). To address the genetic basis of stress response of A. fumigatus, expression analysis of cell wall-related genes was performed at 24 hours after adding K. pneumoniae to A. fumigatus cultures pre-germinated for 12 h. Transcript levels of three genes were quantified by qPCR including CHS3, a regulator of chitin synthase expression, EXG1, a conveyer of glucan β-glucosidase activity, and MPKC, a putative mitogen-activated protein kinase (MAPK) involved in oxidative stress response. Normalization was performed against the reference gene TUBA 107-109 and A. fumigatus grown alone. Significant upregulation of the A. fumigatus genes CHS3, EXG1 and MPKC was documented upon exposure to K. pneumoniae (Fig. 6D). Conversely, significant downregulation of genes associated with hyphal development such as CRZA and PKAR was identified (Fig. 6E). CRZA is a transcription factor in the calcineurin pathway, regulating conidial germination, hyphal growth and virulence [110][111][112][113] , and PKAR is a protein kinase in the cyclic adenosine monophosphate (cAMP) pathway also regulating hyphal formation 114,115 .

Discussion
We have demonstrated that the interaction between different strains of K. pneumoniae and various Aspergillus species in vitro mediates inhibition of fungal spore germination and hyphal growth. In addition, K. pneumoniae was also able to impair biofilm formation of Aspergillus species. This effect was shown to require the presence of live, actively growing bacteria and to be dependent on physical contact between the microorganisms. The inhibitory effect was reversible, as revealed by the resumption of fungal growth upon elimination of the bacteria by antibiotic treatment. During the interaction with K. pneumoniae, Aspergillus adopted a dormant or standby mode, during which the genes regulating hyphal development were downregulated, conferring low energy supply requirement for survival. This notion is in accordance with the observation that the effects of the interaction were apparently independent of more or less abundant carbon source, suggesting that competition for nutrients may not be a major factor in the inhibition of fungal growth.
Moreover, our data suggest that K. pneumoniae confers stress to the fungal cells resulting in the upregulation of protective mechanisms which involve remodelling and reinforcement of the fungal cell wall. This effect was demonstrated by increased sensitivity to stress induced by agents conferring osmotic, oxidative or cell wall-targeted challenges. In response to the stress effects, Aspergillus showed elevated expression of the CHS3, EXG1 and MPKC genes, which are involved in cell wall remodelling and oxidative stress response. This observation is in line with earlier studies indicating that, upon exposure to stress, fungi initiate a response based on activating cell wall salvage pathways that compensate for cell wall damage as a rescue mechanism 96-100 including, for example, an increase in transcript abundance of MPKC in response to oxidative stress [102][103][104][116][117][118] . Since the formation of biofilms can per se lead to resistance against antimicrobial treatment owing to the decreased physical accessibility of the pathogens by drugs 89,119-121 , additional effects of microbial interactions within mixed biofilms affecting the protective mechanisms of individual pathogens may further enhance the resistance to therapy. The understanding of processes occurring during bacterial-fungal interactions (BFI) is therefore essential for the development of appropriate diagnostic and therapeutic approaches. The inhibitory or stimulating effect of BFI mediated by secreted molecules has been demonstrated for different constellations involving molecular cross-talk via production of quorum sensing molecules [12][13][14][15][16][18][19][20]25,26,122 . Earlier observations of BFI involving K. pneumoniae and C. albicans revealed a similar antagonistic effect mediated by the bacteria, but the mechanisms of the interaction have remained unclear 123 . By contrast, an example for synergistic effects is a report on the interaction between Klebsiella aerogenes and Cryptococcus neoformans, which revealed the induction of melanin production in the fungus mediated by bacterial secretion of dopamine, leading to enhanced protection of Cryptococcus from macrophages 12 . K. pneumoniae was also found to increase spore germination and hyphal growth of Glomus deserticola, a vesicular-arbuscular mycorrhizal fungus, and the authors suggested that K. pneumoniae may produce a diffusible compound resulting in hyphal extension 124 .
It is important to point out that the observations of BFI in vitro may not necessarily be reflected by identical implications in vivo, as demonstrated for the interaction between C. albicans and P. aeruginosa which were shown to display inhibitory effects in vitro 13,32,125,126 but synergistic effects in mouse and zebrafish in vivo models 127,128 . Co-infections affecting the lungs of patients with cystic fibrosis resulted in worse clinical outcome in comparison to infections with each of the pathogens individually 14,21,[35][36][37] .
Although this observation may not necessarily indicate discrepant in vitro and in vivo effects of the BFI, it highlights the need to verify observations made in vitro by exploiting models reflecting the complexity encountered in the human host organism. Such analyses may involve the use of organoids generated from primary biopsy materials or animal models. We are currently generating intestinal organoids derived from colon and jejunum/ ileum biopsies from patients undergoing diagnostic endoscopy in order to facilitate studies of BFI in a system mimicking the human bowel. Moreover, we are planning to study the interactions in a zebrafish model which permits analyses in the presence of an immune system closely resembling the situation in the human host 129,130 .
Once confirmed in additional model systems, the observations presented may have important clinical implications. If, in the presence of co-localized infections by K. pneumoniae and various Aspergillus species, treatment with antibiotics could indeed unleash fungal growth, it may lead to rapid expansion of Aspergillus requiring timely therapeutic intervention to prevent severe and potentially life-threatening disease in immunocompromised hosts. The identification of metabolites, such as specific quorum sensing molecules released during BFI, could be exploited for clinical diagnosis of these co-infections, and provide a basis for appropriate treatment measures. Proteomic, metabolomic and genetic analyses of the interaction between K. pneumoniae and Aspergilli addition of bacteria for a total of 24 h. Transcript levels of the A. fumigatus single culture were set to one, and the transcript levels in the co-culture were normalized to the single culture. Additionally, normalization was also performed against the reference gene TUBA. Data are presented as mean + SE of four independent experiments. are currently ongoing, and will expectedly reveal candidate molecules for future diagnostic exploitation, with the ultimate aim to improve the control and outcome of polymicrobial infections, particularly in the immunocompromised patient setting.

Materials and Methods
Strains and growth conditions. Aspergillus and Klebsiella pneumoniae strains used in this study are listed in Table 1. Two different strains of K. pneumoniae were used, a low-biofilm (ATCC 13883) and a high-biofilm (ATCC 700603) forming strain 75 . Aspergillus strains were maintained in Malt Extract agar (MEA, Sigma) plates for 3-4 days at 37 °C. Spores were collected into 1x PBS + 0.1% Tween 20, following filtration by using a 40 μm cell strainer. Spores were stored at 4 °C for up to one week for subsequent use in the experiments. K. pneumoniae strains were maintained in Luria-Bertani (LB) agar plates and grown overnight in liquid LB medium at 37 °C with agitation at 180 rpm. Equal cell numbers of Aspergillus spp. and K. pneumoniae (10 6 cells/mL) were used in co-culture experiments. For qualitative assessment of bacterial-fungal interaction, Aspergillus spp. and K. pneumoniae were grown alone and in co-culture in sterile-filtered yeast extract peptone dextrose (YPD) medium (Formedium, Norfolk, UK) at 37 °C. The microorganisms were grown as biofilms, with static incubation in 35 × 10 mm tissue culture dishes (CytoOne, Starlab GmbH, Ahrensburg, Germany). For co-culture experiments, fungi and bacteria were either mixed at time zero or K. pneumoniae strains were added after pre-germination of Aspergillus spores, permitting the formation of hyphae for up to 12 hours (h). For analysis by confocal microscopy, the same growth procedure was performed using IBIDI plates (µ-dish, 35  are listed in Table 1. Primers were pre-tested for correct amplification prior to use in qPCR analysis. The thermocycler used for qPCR reactions was the Realplex Mastercycler epgradient S, Eppendorf (Hamburg, Germany). Testing for recovery of fungal viability. Aspergillus spp. and K. pneumoniae were grown alone and in co-culture in sterile-filtered YPD medium (Formedium, Norfolk, UK) at 37 °C. These were grown as biofilms with static incubation in 35 × 10 mm tissue culture dishes (CytoOne, Starlab GmbH, Ahrensburg, Germany). Co-cultures of fungi and bacteria were mixed together at time zero and incubated for 24 h. An aliquot of these cultures (100 µL) was then streaked on YPD agar plates containing the antibiotic kanamycin 1000 µg/mL to select for fungal growth. Imaging was performed (Canon Macro Lens EF-S 60 mm) after 48 h incubation at 37 °C. As a negative control, to check for absence of bacterial growth, K. pneumoniae strains were streaked alone onto YPD/ Kanamycin agar plates. Positive controls were implemented by growing Aspergillus spp. alone in YPD/Kanamycin agar plates to check for unaffected fungal growth in the presence of this antibiotic.
Heat-, UV-killed and antibiotic-treated bacteria. Heat-killing (HK) of K. pneumoniae cells was performed in a Thermomixer comfort (Eppendorf) at 95 °C for 30 min in PBS, whereas UV-killing (UVK) was performed in a Stratalinker for 3 cycles at 9999 × 100 µjoules with plate shaking between the cycles. Following HK or UVK, K. pneumoniae cells were added to Aspergillus spp in co-culture in 35 × 10 mm tissue culture dishes (CytoOne, Starlab GmbH, Ahrensburg, Germany) as described above (Strains and growth conditions). Imaging was performed after 48 h incubation. As a control, to check for efficiency of killing, an aliquot (100 µL) of HK and UVK cells was streaked onto LB agar plates and incubated at 37 °C. Treatment of K. pneumoniae with the antibiotics colistin, kanamycin or tetracycline was performed at the concentrations of 100, 2000 and 10 µg/mL, respectively. Aspergillus spp. and K. pneumoniae were grown alone and in co-culture in YPD medium at 37 °C in 35 × 10 mm tissue culture dishes (CytoOne, Starlab GmbH, Ahrensburg, Germany). Co-cultures of fungi and bacteria were mixed together at time zero, and antibiotics were added to the culture at 6 h. Imaging was performed after 48 h incubation.
Cell sorting of A. fumigatus and cell wall stress assay. Sorting of A. fumigatus cells growing alone or in co-culture with K. pneumoniae was performed after 24 h growth in 35 × 10 mm tissue culture dishes (CytoOne, Starlab GmbH, Ahrensburg, Germany). Following growth, cultures were collected into 15 mL Falcon tubes, spun down for 5 min at 4.400 g and SN was discarded. A volume of 2 mL 1x PBS was added and samples were vortexed vigorously. An aliquot (1 mL) was added into a FACS tube through a membrane on the lid. These lids were removed and CFW was added to the cells at a final concentration of 10 µg/mL. Samples were then processed using a FACSaria cell sorter (BD Biosciences) by sorting the cells based on CFW fluorescence (BV421 detector). A total of 20 000 cells were sorted into 1 mL 1x PBS. From these, a volume of 10 µL was spotted onto 12-well agar plates.
To perform the cell wall stress assay, 12-well agar plates were prepared using YPD agar containing the antibiotics colistin at 100 µg/mL and kanamycin at 1000 µg/mL to preclude any possibility of bacterial growth. In addition, each plate was complemented with different concentrations of chemical cell wall stressors and antifungals. The cell wall stresses included: H 2 O 2 1 mM, 2 mM and 4 mM; sorbitol 1 M, 2 M and 4 M; sodium dodecyl sulfate (SDS) 50 µg/mL, 100 µg/mL and 200 µg/mL; CFW 5 µg/mL, 10 µg/mL and 20 µg/mL. The antifungal drugs and their corresponding concentrations included amphotericin B (AmB) 0.25 µg/mL, 0.5 µg/mL and 1 µg/mL; voriconazole 0.025 µg/mL, 0.05 µg/mL and 0.1 µg/mL; micafungin 4 µg/mL, 8 µg/mL and 16 µg/mL. Plates were incubated at 37 °C and imaged (Canon Macro Lens EF-S 60 mm) after 2 days. K. pneumoniae growing alone was also spotted onto these plates as a control for complete growth inhibition of bacterial cells.
Statistics. The significance of differences between Aspergillus-K. pneumoniae co-cultures and single cultures was determined by using the T-test with one-tailed distribution for paired samples (Excel software). P values < 0.05 were considered significant. The calculations were based on at least three independent biological replicates.