Characterization of functional amyloid curli in biofilm formation of an environmental isolate Enterobacter cloacae SBP-8

The biofilm formation by bacteria is a complex process that is strongly mediated by various genetic and environmental factors. Biofilms contribute to disease infestation, especially in chronic infections. It is, therefore important to understand the factors affecting biofilm formation. This study reports the role of a functional amyloid curli in biofilm formation at various abiotic surfaces, including medical devices, by an environmental isolate of Enterobacter cloacae (SBP-8) which has been known for its pathogenic potential. A knockout mutant of csgA, the gene encoding the major structural unit of curli, was created to study the effect of curli on biofilm formation by E. cloacae SBP-8. Our findings confirm the production of curli at 25 °C and 37 °C in the wild-type strain. We further investigated the role of curli in the attachment of E. cloacae SBP-8 to glass, enteral feeding tube, and foley latex catheter. Contrary to the previous studies reporting the curli production below 30 °C in the majority of biofilm-forming bacterial species, we observed its production in E. cloacae SBP-8 at 37 °C. The formation of more intense biofilm in wild-type strain on various surfaces compared to curli-deficient strain (ΔcsgA) at both 25 °C and 37 °C suggested a prominent role of curli in biofilm formation. Further, electron and confocal microscopy studies demonstrated the formation of diffused monolayers of microbial cells on the abiotic surfaces by ΔcsgA strain as compared to the thick biofilm by respective wild-type strain, indicating the involvement of curli in biofilm formation by E. cloacae SBP-8. Overall, our findings provide insight into biofilm formation mediated by curli in E. cloacae SBP-8. Further, we show that it can be expressed at a physiological temperature on all surfaces, thereby indicating the potential role of curli in pathogenesis.


Introduction
Biofilm is regarded as a complex, sessile community of microbes found attached to a surface or buried firmly in an extracellular matrix (ECM) as aggregates (Roy et al. 2018). The biofilm lifestyle provides the bacteria strong ability to withstand adverse environmental conditions like starvation and desiccation. The biofilm also leads to a broad range of chronic diseases 1 3 Vol:. (1234567890) and contributes to a major cause of persistent nosocomial infections in immune-compromised patients (Singh et al. 2000;Davies 2003). Bacterial colonization of abiotic materials and biofilm formation have serious detrimental consequences in hospital settings (Le Thi et al. 2001). Approximately 50% of nosocomial infections result from indwelling devices used for the purpose of medical treatments, such as catheters, cardiac pacemakers, joint prostheses, dentures, prosthetic heart valves, and contact lenses (Piozzi et al. 2004;Wu et al. 2015). These implants provide an ideal surface for the attachment of bacterial cells. The biofilm-associated infections are mainly caused by opportunistic bacteria which prominently belong to the "ESKAPE" group and include Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumanii, Pseudomonas aeruginosa, and Enterobacter species (ESKAPE). These microbes represent a global threat in Hospital settings due to their ever-increasing antimicrobial resistance (Santajit and Indrawattana 2016;Panda et al. 2022).
E. cloacae is the most commonly isolated species of the genus Enterobacter, which has been accepted as the etiologic agent of many infections in hospitalized and immune-compromised patients. It has been regarded as a significant bacterial pathogen in recent years (Zhang et al. 2016). E. cloacae has been reported to cause several s infections, including pneumonia, bacteremia, septicemia, urinary tract infections, and bloodstream infections. It is responsible for 3 to 6% of bloodstream infections, with approximate mortality rates ranging from 27 to 61% (Nyenje et al. 2013). Nosocomial infections related to the use of medical devices are associated with a high risk of mortality and increased economic costs (Costerton et al. 1987). In order to determine ways in which medical, industrial, and ecological biofilm contamination may be prevented, it is important to understand the factors that promote bacterial adhesion and the formation of multicellular communities on abiotic surfaces. The formation of biofilms in both pathogenic and non-pathogenic strains requires time-dependent differential expression of ECM components (Flemming and Wingender 2010). These key components facilitate a transition from an independent planktonic state to an organized multicellular community deeply engrained in the matrix (Costerton et al. 1995). In order to persist as a successful biofilm on any surface, the first step is the attachment to a surface. Enteric pathogens produce an array of adhesive structures and proteins for colonization on various biotic and abiotic surfaces (McWilliams and Torres 2015). Various adhesive factors include curli fimbriae, flagella, cellulose, LPS, colanic acid, and several outer membrane proteins (Carter et al. 2016).
Amongst all the above-mentioned factors, curli fibres are regarded as one of the key factors which play a pivotal role in mediating biofilm formation and pathogenesis (Barnhart and Chapman 2006). Curli are defined as functional amyloid which are secreted into the ECM of biofilm by both commensal and pathogenic members of Enterobacteriaceae (Zogaj et al. 2003;Kikuchi et al. 2005). Curli fibres are produced by a dedicated secretion pathway known as the nucleation-precipitation mechanism or the type VIII secretion system (Barnhart and Chapman 2006). Seven curli-specific genes (csg) make up the structural components and assembly apparatus of the curli fibers, which are encoded by two divergently transcribed operons (csgBAC and csgDEFG) (Evans and Chapman 2014). The curli subunit csgA is responsible for interaction with the biotic as well as the abiotic components. Curli are majorly involved in bacterial adherence to surfaces, cell accumulation and are a significant part of the ECM essential for the establishment of developed biofilms (Kikuchi et al. 2005). Curli display direct interaction with the substratum and form inter-bacterial bundles, allowing a cohesive and stable association of cells in biofilm (Prigent-Combaret et al., 2000). They are also regarded as a virulent element as they interact with a wide range of host proteins, such as matrix proteins and contact-phase proteins which are suggested to help bacterial spreading in the host (Gophna et al. 2001). The expression of curli is responsive to many environmental factors, such as temperature below 30 °C, which strongly portrays the ecological role of this unique functional amyloid in non-host environments. Indeed, curli are implicated in the attachment to, or biofilm formation of, enteric pathogens on abiotic surfaces (Ryu et al. 2004;Goulter et al. 2010). To date, curli have primarily been studied in the context of Salmonella and Escherichia coli biofilms (Van Gerven et al. 2018), where they serve as an adhesive and structuring support of the biofilm ECM, together with cellulose and extracellular DNA. However, the role of curli in the biofilm formation of E. cloacae is poorly explored.
The present work aimed to investigate the role of curli in biofilm formation at different abiotic surfaces, including commonly used medical devices, namely catheters and feeding tubes, which are directly attached to patients for treatment. We also investigated curli-mediated biofilm formation by E. cloacae at 25 °C (room temperature) and 37 °C (physiological temperature). As the environmental isolates can reach hospital set-up via personnel and visitors contributing to hospital-associated infections, we used an environmental isolate, E. cloacae SBP-8, known to have pathogenic potential ) as a test organism instead of a clinical isolate in the present study.

Bacterial growth and conditions
We used E. cloacae SBP-8 (Accession No. NAIMCC-B-02025), an environmental isolate obtained from rhizosphere soil (Singh et al. 2017). The culture was grown in Luria Bertani (LB) broth media (Hi-Media) at 37 °C with agitation (150 rpm) when required. The mutant strain (ΔcsgA) was grown in LB broth supplemented with chloramphenicol (30 µg/ml), as stated above. To determine growth kinetics, the overnight grown culture (OD 600 of 1) of wild type and mutant strain (ΔcsgA) were diluted 1:100 in YESCA media in a flask, and the bacterial growth was monitored every hours at OD 600 using a multiscan reader (Thermo-scientific). The glycerol stocks of both strains were made with 20% glycerol (SRL) and stored at − 80 °C for further use.

Generation of the major curli subunit csgA knockout
We generated an in-frame deletion of csgA encoding the curli major sub-subunit of the curli operon of the E. cloacae SBP-8 genome, as described in (Sawitzke et al. 2013) with minor modifications. This method involves replacing the desired gene with an antibiotic (chloramphenicol) resistance cassette using the bacteriophage lambda red recombinase system. Electrocompetent cells were prepared by the standard protocol where bacterial cells were grown with the addition of 1 M Arabinose till OD 600nm of 0.3-0.4, followed by washes with 10% glycerol in ice-cold conditions (Huang et al. 2014). The chloramphenicol (Cm R ) cassette was amplified from the pKD3 plasmid using specific primers. The amplicon contained chloramphenicol cassette flanked with sequences upstream (38 bp) and downstream (21 bp) of csgA. The purified Cm R cassette (850 ng) was electro-transformed into E. cloacae SBP-8, having pACBSR (a plasmid consisting of λ-Red recombination system), using a 2 mm diameter Bio-Rad cuvette using gene pulsar II (Bio-Rad, USA) at 2.5 kV, 25 W for 5 ms. The transformants of E. cloacae SBP-8 with csgA deletion were grown on an LB plate supplemented with chloramphenicol and ampicillin at 37 °C, followed by their screening by PCR using (Cm R ) cassette-specific confirmatory primers. The deletion of the gene was confirmed by using PCR amplification of the genes and the inserted antibiotic cassette. The primers used for the knockout generation and confirmation are listed in (Supplementary material: Table 1).

Congo Red assay
The curli production by E. cloacae SBP-8 was determined by Congo Red assay as previously described (Zhou et al. 2013). Briefly, Congo Red agar plates was made by preparing yeast extract and Casamino acid agar (YESCA: 1 g L −1 yeast extract, 10 g L −1 casamino acids, 20 g L −1 agar) and after autoclaving, filter sterilized Congo Red (25 µg ml −1 final concentration; Sigma) and filter sterilized Brilliant Blue G (10 µg ml −1 final concentration; Sigma) were added. E. cloacae SBP-8 strains (wild type and ΔcsgA mutant) were grown in LB broth and incubated at 37 °C overnight. Five microliters of the overnight culture of each strain was spotted on the centre of a thick Congo Red agar plate. The plates were incubated at 25 °C and 37 °C for a time duration of 24 to 96 h. Chloramphenicol (30 g ml −1 ) was added to the growth media for the mutant strain. The images of CR assay were captured using Oneplus 7 phone camera.

Evaluation and quantification of biofilm formation on various surfaces
To evaluate the role of curli in biofilm formation by E. cloacae SBP-8, we compared the extent of biofilm formation in wild-type of E. cloacae SBP-8 with its mutant (ΔcsgA) counterpart at 25 °C and 37 °C on different abiotic surfaces using the standard crystal violet assay as described in (Djeribi et al. 2012). We used an Enteral feeding tube (Polyvinylchloride) (Romolene batch no: 171122361, India) and latex catheter (Romolene batch no: G20082297, India) as the medical devices. Culture tubes and glass slides ( borosilicate) (Borosil, India) were used as glass surfaces.

Biofilm formation on medical devices
Sterile latex catheters (CT) and enteral feeding tubes (EFT) were cut into 0.5 mm thick discs and were aseptically introduced into 5 ml YESCA broth inoculated with E. cloacae SBP-8 and the mutant strain (diluted to 1:100). The tubes were incubated for different periods (24, 48, 72 and 96 h) under static conditions. After respective incubation periods, unattached cells were removed by rinsing the discs of the enteral feeding tube and latex catheter with PBS. The discs were further stained with 1 ml of 0.1% crystal violet. After 30 min of incubation, the crystal violet solution was removed, and the excess stain was rinsed off with a mild wash by PBS buffer. Finally, the biofilm was extracted with 1 ml of 33% glacial acetic acid. After 30 min of incubation, the absorbance of the extracted solution was measured at 570 nm. The resulting absorbance i indicates the formed biofilm (Philips et al. 2017). The extracted solution was diluted 2 to 3 times before measurement. The optical densities were measured using Thermo scientific, multiscan Go spectrophotometer. All biofilm quantification experiments were done in triplicate.

Biofilm formation on glass surface
Glass tubes (1 cm diameter) were used to evaluate the biofilm formation. The tube containing 5 ml YESCA broth was inoculated with the 1% volume of the overnight grown culture of E. cloacae SBP-8 and the mutant strain. As described above, the tubes were incubated at 25 °C and 37 °C for different time intervals under static conditions. The crystal violet assay to evaluate biofilm formation was performed as described above.

Extraction of RNA from biofilm
For RNA isolation, strains were grown in YESCA media in a static condition at 25 °C and 37 °C. RNA was isolated from biofilm formed as a ring in the sides of culture tubes using the conventional method (Ares 2012) with minor modifications. The rings were scrapped from 5 tubes using a sterile cell scraper. The pooled rings were immediately mixed with 70% of ammonium sulfate (Merck) and incubated at RT for 5 min, followed by centrifugation at 5000 g for 5 min at 4 °C (Eppendorf, Germany). The supernatant was poured off, and the cells were digested with freshly prepared lysozyme (0.5 mg/ml) and incubated at RT for 15 min. Post digestion of the cells, 15 µl of sodium acetate (CDH, India) and 45 µl of 10% SDS (Sigma, USA) were added to the cells and vortexed briefly for 10 s. After this, 400 µl of P:C:I (Phenol: Chloroform: Isopropanol) solution was added, vortexed to emulsify, and incubated for 10 min at 65 °C. The tube was incubated on ice for 5 min and centrifuged at 14,000 g for 20 min. The resulting aqueous phase was extracted with an equal volume of chloroform (SRL), and the process was repeated. Finally, the aqueous phase was collected in a fresh tube, and RNA was precipitated using 50 µl of sodium acetate and 100% ethanol (Merck) at − 80 °C for overnight. After overnight incubation, the samples were centrifuged at 14,000 g for 30 min. The pellet was washed twice with 70% of ethanol and resuspended in 40 µl of water. For both medical devices, firstly, the culture was removed, and the entire surface (0.5 cm piece) was washed with 1X PBS in order to remove the unbound cells as adapted from (Mandakhalikar et al. 2018). The RNA from the biofilm was extracted as above. RNA samples were checked by agarose gel electrophoresis to assess their integrity and quantified spectrophotometrically. Reverse transcription was performed with 800 ng of total RNA using Verso cDNA Synthesis Kit (Catalog number: AB/1453-A) as per the instruction from the manufacturer. cDNA synthesis efficiency was verified by electrophoresis on agarose gel using gene specific primer for curli.

Gene expression analysis
We investigated the expression of csgA, one of the curli proteins, in E. cloacae SBP-8 at both 25 and 37 °C. Quantitative real-time PCR (qPCR) was performed using Bio-Rad SYBR green dye in Bio-Rad thermocycler by using gene-specific primer for curli. Relative fold change in target gene expression was calculated using 2 − ΔΔ Ct method by normalizing against 16S rRNA. The expression of csgA was determined by using the delta Ct method (2ΔCt), a variation of the Livak method, where ΔCt = Ct (reference gene)-Ct (target gene) (França et al. 2011). All reactions were performed in duplicates, including negative control samples, which never showed significant threshold cycles. The primers used in the study are mentioned in the (Supplementary material: Table 1).

Confocal laser scanning microscopy (CLSM) of biofilm
Intact biofilms of E. cloacae were visualized using an inverted Zeiss (LSM) 880 Airyscan (Carl Zeiss, Jena, Germany) equipped with a C-APOCHRO-MAT 40x/1.2 Water Corr-UVVIS-IR objective. After incubation for 48 to 72 h, the glass slides were removed from the culture tube and washed twice with 1X PBS in order to remove the unbound cells. To visualize the formed biofilm, the wildtype, and mutant strain were stained with 5 µM SYTO9 (Thermoscientific) at room temperature for 30 min, followed by two 10-min washes with 1X PBS. The laser was used at 488 nm for excitation, and the emission was observed at 528 nm (SYTO9). Obtained images and z-stack projections were visualized using Zeiss ZEN System imaging software (Zeiss).

Observation of biofilm by field emission scanning electron microscopy (FESEM)
We were further interested in visualizing the morphology of biofilm formed on the surfaces of latex catheters, glass slides, and the enteral feeding tube employing FESEM using the protocol of (Djeribi et al. 2012) with a minor modification. 0.5 mm thick discs of the medical devices and glass slides were aseptically introduced into tubes containing LB broth inoculated with E. cloacae SBP-8 with a dilution of (1:100) at both 25°and 37 °C under static conditions. The samples were dehydrated and sputter-coated with chromium metal using the Quorum Q150T ES system, and the morphology of the biofilm was observed in a Thermoscientific FEI FE-SEM APREO S SEM (Netherland) system at 20 k V.

Phenotypic characterization of curli
In order to understand the role of curli in biofilm formation, we generated a csgA mutant. Firstly we used a standard qualitative method to distinguish curli-producing bacteria from non-curliated bacteria, where CR-binding is known to be curli-dependent (Reichhardt et al. 2015). As E. cloacae SBP-8 lacks cellulose, binding of CR to cellulose is ruled out. The wild-type strain produced (left panel) deep red colouration in comparison to the knockout strain (right panel), which did not develop red phenotype on Congo red media at any incubation temperature used in the study, indicating the absence of curli (Fig. 1A). We also tested for the curli production at 37 °C which is represented in (Fig. 1B) which showed light red (left panel) and light pink-white colour (right panel) by wild type and the mutant respectively. In wild-type strain, the curli production was observed at 25 °C for a time interval of 24-96 h. Although we found lesser colouration for 37 °C on the CR-YESCA plate, we further examined the role of curli at this temperature in biofilm formation at varying surfaces.

Contribution of curli to E. cloacae SBP-8 biofilm formation on abiotic surfaces
We first evaluated the biofilm formation of E. cloacae SBP-8 wild-type strain and csgA mutant on different abiotic surfaces (glass, enteric feeding tubes, and latex catheter) where cells were grown in YESCA broth at 25 °C, a condition known to induce the biogenesis of curli. It was also tested at 37 °C to test if the curli are produced at this temperature and potentially contribute to the biofilm formation on abiotic surfaces. To ascertain that csgA deletion does not affect growth of the bacteria, growth kinetics was performed, which did not show a significant growth difference in wild-type and ∆csg knockout (Fig. S1).
We used a crystal violet (CV) assay to evaluate the biofilm formed on various surfaces, which is the most suitable method to quantify the biofilms (Philips et al. 2017). The initial examination of the crystal violet tubes at both 25 and 37 °C revealed differences in the adhered biomass (ring), which clearly indicated that the wild-type strain adhered more than the mutant strain mentioned in the supplementary material (Fig.  S2). In general, the biofilm formation on the glass surface was less intense than the other two surfaces at 25 °C during the initial time period, i.e., up to 72 h ( Fig. 2 A-C). Deletion of csgA negatively affected the biofilm formation on all three surfaces used. However, the percentage decrease upon csgA deletion observed was different at different time intervals and types of surfaces. CV staining showed a 48% and 21.5% decrease in biofilm formation by ∆csgA strain on the glass surface at 72 and 48 h, respectively. In the case of the medical devices, we found a 49% reduction in the biofilm at 48 h on an enteral feeding tube (Fig. 2B). Whereas, for foley latex catheter, we found significant differences from 48 to 96 h (Fig. 2C). The catheters showed 29% and 58% reduction respectively at 48 and 72 h in biofilm formed by ∆csgA strain as compared to its wild-type counterpart (Fig. 2C). Similar results were obtained at 37 °C where deletion of csgA caused decreased biofilm formation by E. cloacae SBP-8 ( Fig. 2D-F). However, the effect of csgA deletion on biofilm formation was less severe at 37 °C than at 25 °C. The ∆csgA strain showed a 23-27%, 11 −21%, and 13 to 29% decrease in biofilm on glass, enteral feeding tube, and latex catheter, respectively, at different time intervals.

Morphological analysis of the biofilms through scanning electron microscopy
To visualize the fine architecture and cell aggregation of the formed biofilms on the glass surface and medical devices (enteral feeding tube and latex Foley catheter), FESEM (Field Emission Scanning Electron Microscopy) analysis was performed. As seen in Figs. 3, 4, and 5, the surfaces were readily and rapidly colonized by a wild-type strain of E. cloacae SBP-8. In comparison to the glass surface, the medical devices formed multi-layers of cells encased in   hrs, the wild-type strain adhered more on to the surface rather than the mutant strain dense uniform EPS, thereby indicating more biofilm formation on medical devices rather than the glass surface. On the contrary, the mutant strain (ΔcsgA) exhibited lesser colonization with sparsely distributed cells on the surface (Figs. 3, 4, and 5). The mutant strains exhibited diffuse layers at all time intervals, along with lesser EPS production. The mutant strain also showed elongated and flat morphology, which has also been reported in E. coli (Azam et al. 2020). The flat and elongated morphology of the cells in the mutant strain has been mentioned in the Supplementary data (Fig S3). Thus, our result indicated that curli is critical for the firm attachment of cells onto the surface and for the three-dimensional development of biofilms in E. cloacae SBP-8.
Further, we performed confocal scanning laser microscopy analysis on the biofilms grown on glass slides at 25 °C and 37 °C to validate the architecture. The biofilms were grown on glass slides. Both wildtype and mutant strains were incubated at static 25 °C and 37 °C temperatures. The samples were stained with Syto-9 (nucleic acid binding dye) and analysed after 48 to 72 h incubation. The wild-type strain showed a greater density of microbial cells widely distributed on the surface at both temperatures, as shown in Fig. 6. The wild-type strain exhibited strong colonization leading to micro-colony formation at one site and biofilms at other sites. This observation represents the phase switch from micro-colonies to a sessile state, as seen in Fig. 6 (C) and (G) with the encircled area. On the contrary, the mutant strain showed weaker adherence, lesser density, and colonization to the surface with diffused layers of cells with the least EPS production at both temperatures Fig. 6 (B) (G) (G) (H). Since curli plays an integral part in shaping the architecture of the formed biofilm, we also performed the z-stack analysis of the formed biofilm as depicted in the supplementary (Fig. S4). We found significant differences in the thickness between the wild-type and mutant strain.

Curli plays a role in initial attachment to abiotic surfaces
We tested the role of curli in the biofilm formation at both temperatures 25°and 37 °C through real-time Fig. 4 FESEM micrographs of the biofilm formed on the enteral feeding tube at 25° and 37°C by E. cloacae and its corresponding csgA mutant (10,000X). At different time points from 24-96 hrs, the wild-type strain adhered more on to the surface rather than the mutant strain qRT-PCR. We determined the curli expression at both temperatures on all the surfaces. For 25 °C (glass), we found curli expression to significantly increase from 72 to 96 h (p > 0.0442) (Fig. 7A). In the case of the enteral feeding tube, we found a significant increase from 48 to72 h (p > 0.0173) (Fig. 7B). For the foley latex catheter we found the curli expression to increase from 72 to 96 h (p > 0.0252) (Fig. 7C). The gene expression studies were also carried out at 37 °C. For the glass surface (Fig. 7D), we found a significant increase of csgA gene expression from 48 to 72 h (p > 0.0048) followed by a decrease from 72 to 96 h (p > 0.0048). For the enteral feeding tube (Fig. 7E), there was higher expression of csgA at 48 h which significantly (p > 0.0048) decreased at 72 to 96 h (p > 0.0442). Similarly, for the foley latex catheter (Fig. 7F), there was a significant increase in csgA expression from 48 to 72 h.

Discussion
E. cloacae, a member of ESKAPE group, is one of the important nosocomial pathogens which can form biofilms on abiotic surfaces including medical devices. Biofilm formation is a highly regulated, complex, and a dynamic process mediated by multiple genetic, physiological, and environmental factors. Since biofilm formation is one of the key factors, which imparts virulence to most of the organisms, the present study investigated the potential role of curli, important proteinaceous component of the biofilm matrix, in the biofilm formation of an pathogenic environmental isolate (non-clinical) of E. cloacae (SBP-8) . To the best of our knowledge, the role of curli in biofilm formation by E. cloacae is underexplored. To understand the role of curli, we generated a knockout mutant of csgA, the structural protein and monomer of curli fimbriae, employing ʎ red-recombinase system.
The preliminary screening of curli using CR assay confirmed the production of curli by E. cloacae SBP-8 at both 25 and 37 °C. Since CR can also 24-96 hrs, the wild-type strain adhered more on to the surface rather than the mutant strain bind to other bacterial factors like cellulose, we used ΔcsgA mutant of E. cloacae. Since in our preliminary results of calcoflour assay and genome sequence (Genome Accession: PRJNA338268 submitted to NCBI) analysis of E. cloacae SBP-8 indicated lack of cellulose synthesis (unpublished results), the CR binding with cellulose is ruled out. Hence, the loss of colour in mutant bacterial colony confirmed that the colour change on CR-agar plate was due to curli production (Fig. 1). It has also been well-documented in case of E. coli O157:H7 that congo red affinity is strongly dependent on the curli production at both 25 and 37 °C temperature (Sharma and Bearson 2013). Similar findings have been reported in various member from the family of Enterobacteriaceae which includes to enterotoxigenic E. coli, Citrobacter freundii, C. koseri/farmeri, E. aerogenes, E. cloacae, E. sakazakii, Klebsiella oxytoca, K. pneumoniae, Proteus mirabilis, and Raoultella ornithinolytica which showed binding to Congo red 25 and 37 °C (Zogaj et al. 2003;Szabó et al. 2005).
To investigate the role of curli in biofilm production by E. cloacae SBP-8, crystal violet (CV) assay was performed under static conditions at 25 and 37 °C. Our studies have shown that curli deletion had a positive effect on different surfaces with varying temperature. As discussed earlier curli plays a important role in the initial phases of the biofilm development. Decreased biofilm formation in ∆csgA mutant at both temperatures and on all the surfaces indicated that curli does play a role in biofilm formation of E. cloacae. Similar observation was reported in E. coli K-12 and E. coli strain MG1655 where curli deficient strains adhered less to the polystyrene surface (Beloin et al. 2008;Azam et al. 2020). Another report on E. coli also showed that the curli facilitates attachment and provides a scaffold for the maturation of biofilms, where it was shown to adhere to polyurethane sheets at both 25 and 37 °C (Kikuchi et al. 2005). In this case, it was shown that the mutant strain formed lesser biofilm on the above-mentioned surfaces and caused a delay in the maturation of the formed biofilms. However, the effect of csg gene deletion on biofilm formation was highest between 48 and 72 h of biofilm development on different surfaces. It highlights the fact that curli is primarily required for Fig. 6 Confocal laser scanning microscopy images of biofilm formed by the wild-type and the mutant strain (ΔcsgA) incubated at 25°C and 37°C for 48 and 72 hrs under static conditions. All of these CLSM XY, or XZ-scan images were visualized with SYTO9 9 (green). (Scale :100µm). Dense multi-layered biofilms were seen in the wild-type strains as depicted in (A), (C), (E), (G) Lesser cellular density with monolayer and micro-colony were seen in the mutant strain as depicted (B), (D), F), (H) adhesion during the initiation of biofilm formation and the development of multi-layered cell clusters on various surfaces (Prigent-Combaret et al., 2000;Le Thi et al. 2001). Similar studies have shown that curli is important for different strains of E. coli, including O157:H7 and STEC (Shiga toxin-producing E. coli), to attach to biotic and abiotic surfaces such as glass, stainless steel, polystyrene, and stainless steel (Cookson et al. 2002;Jain and Chen 2007;Uhlich et al. 2009). Another study demonstrated that attachment and biofilm formation on glass and polystyrene surfaces by curli-producing Salmonella strains was more efficient than by non-curli-producing strains (Austin et al. 1998).
To correlate the role of curli with biofilm formation at the molecular level, we further quantified the expression of csgA, a structural gene of curli at 25 and 37 °C. In general, the curli fibers are known to be expressed at a lower temperature (25-30 °C), as observed in our study as well, but not at 37 °C in various enteric bacteria (Kikuchi et al. 2005). On the contrary, we observed csgA expression also at 37 °C on all the above-mentioned surfaces. Our observation aligned with the study carried out in different pathogenic strains of E. coli only including six pathogenic isolates of E. coli O:157, E. coli O157:H7 (Ben Nasr et al. 1996;Uhlich et al. 2001;Gualdi et al. 2008). Similar findings have been reported in STEC E. coli, where it has been demonstrated that the maturation of the biofilm cannot take place without the expression of curli at 37 °C (Dewanti and Wong 1995). As mentioned earlier that curli is expressed at 37 °C in the case of E. coli K-12, helping in a strong attachment to the polyurethane sheets (Kikuchi et al. 2005). Our results suggest that expression of curli at 37 °C is favourable for the organism to adhere to the biotic and abiotic surfaces and is thought to be strain dependent owing to the complex regulatory network governing curli biogenesis (Gophna et al. 2001). Natural variation in curli production has been reported for E. coli isolates and STEC O157:H7 populations originating from diverse environments (Maurer et al. 1998;Scheuerman et al. 1998;Carter et al. 2016).
Our gene expression analysis also showed that curli is required for attachment on the above-mentioned surfaces at both temperatures. We found Fig. 7 Curli gene expression at various surfaces. Panel A-C shows gene expression level of csgA at 25°C on Glass (A), Enteral feeding tube (B), and Foley latex catheter (C). Panels D-F shows csgA profile at 37 °C on at Glass (D), Enteral feeding tube (E), and Foley latex catheter (C)). Unpaired Student's t-test was performed. Statistical significance: *, P≤ 0.05, **, P≤ 0.01, ***, P≤ 0.001, ****, P<0.0001, ns = not significant increased curli expression on the glass surface, an enteral feeding tube (PVC), and latex foley catheter. Medical implants, including catheters and stents, are vulnerable to colonization by biofilms (Austin et al. 1998). Various findings have shown that curli significantly helps in attachment to hydrophilic surfaces. Due to their roughness and irregularities in the polymeric substances, Biomaterials provide easy bacterial adhesion and attachment for bacterial colonies. Our data also depicted that biofilm formation was more on medical devices than the glass surface (Eginton et al. 1995;Scheuerman et al. 1998). It has been reported that particular surface structures, such as the type I secretion system and curli fimbriae enable the cells to stabilize on the surface (Ryu et al. 2004).
The overall reduction in the biofilm formation by E. cloacae SBP-8 ΔcsgA gives us insight that curli in enteric bacteria could potentially contribute towards the colonization on various surfaces. Although not yet formally demonstrated, our observations support the notion that curli could indeed be expressed by E. cloacae SBP-8 growing in-vivo, especially in biofilm, where environmental conditions and selective pressures are considerably different from in vitro conditions. Thus, once expressed, curli will have the capacity to interact with human cells at physiological temperatures and contribute to the pathophysiology of bacterial infectious disease.
Our CLSM analysis of biofilm formed by E. coli strain W3110 under static conditions at 30 °C revealed that the curli is restricted to the micro-colony formation stages, whereas the mutant ΔcsgA strain showed lesser adhesion to the surface with low cell density (Besharova et al. 2016).
Curli have been identified as a morphological envelope structure of major importance for biofilm formation in E.coli (Vidal et al. 1998). Our findings through FESEM analysis demonstrated that curli is required for biofilm formation on various surfaces. Since the wild-type strain produced substantially thicker layers of cells than the mutant strains, it clearly appears that curli functions to stabilize intercellular contacts allowing aggregation to occur and thereby increasing the biofilm thickness (Austin et al. 1998). Similar to the function of thin aggregative fimbriae, curli produced by diarrheagenic E. coli plays an influential role by forming stiff colony phenotypes and cell clumping in order to facilitate strong adhesion (Collinson et al. 1992). Similar to our observations, several studies using SEM and TEM have shown clear role of curli in mediating cell-to cell contact and forming thick bundles which binds cells together on to the surface in enteropathogenic E. coli (Giron et al. 1991). Auto aggregation was shown to be high in ompR234 E. coli K-12 suggesting that this property results from the formation of interbacterial curli bundles formation (Sohel et al. 1996;Vidal et al. 1998).
Altogether, results obtained in this study allow us to draw up a model that explains biofilm development on abiotic surfaces. Curli plays fundamental role in two steps required for biofilm development: initial bacterial attachment and three-dimensional biofilm formation. Curli helps in adhesion and becomes an important requirement by the bacterial cells to attach to varying surface with different physio-chemical properties. Although curli have been shown to mediate binding of host proteins in the mammalian host, we speculate that the varying conditions encountered outside the host enhance curli expression. Our study shows that curli are more specifically adapted to bacterial attachment on inert surfaces. To cope up successfully with diverse environmental conditions, it seems convincing that E. cloacae SBP-8 may possess adaptive programmes for optimizing growth and survival in each ecological niche as demonstrated through our study.

Conclusions
The present work demonstrates the production of curli and its role in biofilm formation on various abiotic surfaces, including medical devices by E. cloacae SBP-8, an environmental isolate. Despite being from the exogenous origin (not from a clinical setup), the isolate forms a strong biofilm on various surfaces, indicating the potential of environmental E. cloacae to infect immunocompromised patients through medical device-associated microbial contaminations. E. cloacae SBP-8 forms biofilm on the tested medical devices and glass surfaces with varied intensity. With the deletion of csgA gene encoding curli, the biofilm formation is significantly affected, particularly at the initial time period (24-72 h), which suggests that curli fibres are required to initiate biofilm formation during the adherence of microbial cells on the surfaces. Our results suggest that curli production is important in the process of initial attachment and conferring cell aggregation properties to E. cloacae SBP-8 on abiotic materials. It can be speculated that the functional amyloid of E. cloacae SBP-8 is one of the key factors required for successful colonization on different materials, thereby contributing to hospitalassociated infection through contaminated medical devices. Hence curli fibres serve as a potential target for biofilm interventions and infection control.