Single-walled carbon nanotube probes for characterization of biofilm degrading enzymes demonstrated against Pseudomonas aeruginosa extracellular matrices

Abstract

Hydrolase co-therapies that degrade biofilm extracellular polymeric substances (EPS) allow for better diffusion of antibiotics and more effective treatment; current methods for quantitatively measuring enzymatic degradation of EPS are not amendable to high-throughput screening. Herein, we present biofilm EPS-functionalized single-walled carbon nanotube (SWCNT) probes for rapid screening of hydrolytic enzyme selectivity and activity on EPS. The extent of biofilm EPS degradation is quantified by monitoring quenching of the SWCNT fluorescence. We used this platform to screen sixteen hydrolases with varying bond breaking selectivity against a panel of wild type Pseudomonas aeruginosa and mutants deficient or altered in one or more EPS. Next, we performed concentration-dependent studies of six enzymes on two common strains found in Cystic Fibrosis (CF) environments and, for each enzyme, extracted three first-order rate constants and their relative contributions by fitting a parallel, multi-site degradation model, with good model fit (R² from 0.65 to 0.97). Reaction rates (turnover rates) are dependent on enzyme concentration and range from 6.67e-11 to 2.80e-3 *s⁻¹ per mg/ml of enzyme. Lastly, we confirmed findings from this new assay using an established crystal-violet staining assay for a subset of the hydrolase panel. In summary, our work shows that this modular sensor is amendable to high-throughput screening of EPS degradation, thereby improving the rate of discovery and development of novel hydrolases.

Graphical Abstract

Introduction

Biofilms are self-aggregating microbial communities enmeshed in self-generated extracellular matrices predominantly composed of polysaccharides, proteins, lipids, and extracellular DNA (eDNA)^1,2. Biofilms frequently form on biological and synthetic surfaces, and contribute significantly towards persistent bacterial infections (65 to 80% of bacterial infections³). Unfortunately, the recalcitrant adhesion and reduced diffusion of external antimicrobial agents also make biofilm infections difficult to treat. The dense extracellular polymeric substances (EPS) of the biofilm confers the bacteria consortium considerable antibiotic resistance^4,5 as well as protection against native immune defenses⁶. Among the many biofilm-producing bacterial species, Pseudomonas aeruginosa is of significant clinical relevance and has served as a model for biofilm development and eradication^7,8

Pseudomonas aeruginosa is a gram-negative, rod-shaped, opportunistic bacterium that causes nosocomial infections in immunocompromised patients⁹. These bacteria thrive inside of patients with burn wounds¹⁰, cystic fibrosis (CF)¹¹, acute leukemia^12,13, and organ transplants¹⁴ causing severe infections. P. aeruginosa has the genetic capacity to produce three different exopolysaccharides namely- Psl, Pel, and alginate¹⁵. Psl is a neutral pentasaccharide subunit that contains D-mannose, L-rhamnose, and D-glucose in 3:1:1 ratio interlinked via β-(1,3), α-(1,3), and α-(1,2) glycosidic linkages¹⁵. This polysaccharide interacts with many proteins to stabilize the biofilm structures¹⁶. Pel is a cationic polysaccharide that is composed of β-(1,4) partially acetylated N-acetyl glucosamine and N-acetyl galactosamine sugars¹⁷. This polysaccharide cross-links with eDNA in the biofilm stalk via ionic interactions¹⁷. Similarly, alginate is an acidic polysaccharide composed of non-repeating subunit of D-mannuronic acid residues interspersed with L-guluronic residues that are O-acetylated at the C5’ position joint by β-(1,4) glycosidic linkages¹⁵.

Recently, the World Health Organization (WHO) identified P. aeruginosa as a critical priority pathogen due to its innate ability to tolerate antibiotics by forming biofilms¹⁸. Major mechanisms present in P. aeruginosa to evade antibiotic treatment can be classified into intrinsic, acquired, and adaptive resistance¹⁹. Intrinsic resistance comprises of efflux pumps and low outer membrane permeability that expels antibiotics out the cell and produces enzymes that deactivates antibiotics¹⁹. Acquired resistance in P. aeruginosa is achieved by horizontal gene transfer of resistance genes or mutational changes^19,20. Finally, adaptive resistance involves the formation of biofilms that limits the diffusion of antibiotics^19,21. Recent studies have shown that P.aeruginosa evades antibiotic treatment by up-regulating expression of a symbiotic filamentous inoviral prophage (Pf4) by forming separated liquid crystalline compartments around bacterial cells²². This protection against antibiotics as well as immune defense mechanisms necessitates novel approaches to eradicate P. aeruginosa biofilms. Because biofilm formation is a critical attribute of P. aeruginosa, several anti-biofilm agents have been developed and studied to target disruption of the biofilm matrix^23–25. Enzymatic degradation of one or more EPS components is one such strategy with growing attention since it would make bacteria more susceptible to the native immune system and allow for better penetration of antibiotics. Several studies have tested the ability of hydrolytic enzymes to disrupt EPS^26–30. For example, β-glucosidase and lyticase have shown to increase susceptibility to ceftazidime by disrupting the β-(1,3) and β-(1,4) glycosidic linkages in P. aeruginosa strains isolated from CF patients³¹. Similarly, alginate lyase and DNase have also shown promise towards treating infections caused by mucoid strains of this bacterium^32,33.

Existing methods to characterize changes in biofilm EPS morphology for assessment of biofilm degradation involve propagation of living cells (Table 1). These assays are limited by time, labor, and cost, which limits throughput when screening novel enzymes that degrade the biofilm EPS. To overcome these limitations, fluorescent single-walled carbon nanotubes (SWCNT), are proposed as abiotic proxies to colonies for screening of biofilm-degrading enzymes.

Table 1:

Summary of current methods used to quantitatively characterize biofilms

Type of Assay	Description	Advantages	Disadvantages
Colony Counting34–36	Separating individual cells on agar plate to differentiate between living and dead cells	Inexpensive	Time and labor intensive; Vulnerable to counting error; low throughput
Flow-based Cell Counting34,37,38	Use of lasers to understand surface morphology of the cells	Simultaneous information about multiple cell parameters	Expensive, labor intensive, and low throughput
Microscopic techniques (such as confocal scanning laser microscopy)34	Imaging of the biofilms to understand the morphology	High spatial and temporal resolution of the biofilm; Analysis of the sample in the natural state	Expensive, time consuming and low-throughput; Results may vary depending on the user.
Crystal Violet Staining34,36,39	Stains bacterial peptidoglycan to quantify biofilm biomass	High-throughput, inexpensive, easy to perform	Does not distinguish between live and dead cells; batch-batch variability; lacks sensitivity; Overestimates or underestimates the biofilm biomass

Semiconducting SWCNT are one-dimensional carbon allotropes which are fluorescent in the near infrared (nIR) region. The emission of SWCNT in the nIR leads to minimal absorption and scattering in biological tissues, allowing for measurements in complex samples⁴⁰. SWCNT have high surface sensitivity which enables detection of changes in the local dielectric environment by modulation of exciton dynamics⁴¹. Several fluorescent, biosensors have been developed using SWCNT to detect a diverse range of analytes (sugars⁴², proteins^43–45, DNA⁴⁶, and glycans^47,48); selectivity is achieved by non-covalently functionalizing the SWCNT surface with different amphiphilic polymers^49,50. This functionalization forms a corona phase around the SWCNT which renders the nanoprobes soluble in aqueous solution⁵⁰. Previous works have utilized the SWCNT platform to detect protease activity⁵¹ as well different pathogens^52,53, however, none have used the platform to understand biofilm degradation. Recently, our group has capitalized on the degradation of the SWCNT corona phase for developing robust sensors of hydrolytic enzyme activity^54,55. Herein, we extend this platform to screen for enzymes that degrade biofilm extracellular matrices.

In this study, SWCNT surfaces are non-covalently functionalized with biofilm EPS from a suite of P. aeruginosa strains that differed in the amount and type of EPS components. These probes were then screened with 16 enzymes (glycanases, proteinases, DNases) to determine relative EPS degradation rates and substrate selectivity. EPS degradation was reported by monitoring the modulation of fluorescent signal; this transduction method is hypothesized to result from increased solvent access to the SWCNT surface, thus changing local permittivity and recombination of excitons as proposed in other works^41,54–56. The findings from the initial screen were used to select six enzymes for use in concentration dependent studies. The findings were fitted with a multi-site, first-order kinetic model to understand kinetics governing biofilm EPS degradation. Lastly, the results of the biofilm EPS-SWCNT probe were benchmarked against a standard crystal violet assay to determine limitations and advantages of this approach in screening hydrolase candidates for co-therapies against biofilms.

Experimental Section

Isolation of the EPS

P. aeruginosa wild type PAO1 and various derivatives altered in EPS properties (PAO1Δpsl Δpel, PAO1 Δpel, PAO1 Δpsl, PAO1 Δalg, PD0300-alginate overproducer, MJK8-small colony variant) were used in the study since they are well characterized^57–60. Each strain was grown in tryptic soy broth (TSB) liquid media at 30°C for 24 h. The liquid culture was diluted to ~1.0 (OD₆₆₀) and then 100 μL aliquots of the diluted culture were transferred to tryptic soy agar (TSA) plates to generate a lawn of cells for sufficient cell biomass for EPS isolation. These TSA plates were incubated at 30°C for 48 h. After prolonged incubation, biomass was extracted using a cell scraper (Corning-3010) and transferred into 25 ml 0.85 % saline solution (w/v). The resulting solution was vortexed and then the cell suspension was centrifuged at 10,000 rpm for 10 min and the supernatant carefully decanted: the supernatant was then filtered (0.45 μm filter) and the resulting filtrated contained the EPS.

Ethanol Precipitation- Purification of the Polysaccharides in the Biofilm EPS

To purify the polysaccharides in the biofilm EPS, the 25 ml isolated biofilm EPS was first treated with DNase I (2.5 μL, 1 mg/ml) (Sigma-11284932001) at 37 °C for 4 h to degrade extracellular DNA. Subsequently, the EPS was treated with Proteinase K (2.5 ml, 10 mg/ml) (Sigma- 70663) at 37 °C overnight prior to incubating at 80 °C for 4 h to thermally denature the DNase I and Proteinase K. This denaturation step was performed to avoid interference with the fluorescent assay.

After removal of DNA and protein from the biofilm EPS matrix, 3x volume of ice-cold 100 % ethanol was added, and the solution was stored at −20 °C for 24 h. The resulting solution was then centrifuged at 16,300 × g for 30 min, and purified polysaccharides were recovered as the precipitate. The precipitate was air dried and was resuspended in 5 ml water for functionalizing polysaccharide-SWCNT probes. This protocol was only performed while testing the selectivity of the biofilm EPS-SWCNT probe towards polysaccharides.

Development of the Biofilm EPS-SWCNT Probe

To fabricate the biofilm EPS-SWCNT probe, 4 ml of isolated EPS was transferred to a 5 ml Eppendorf tube. 2 mg of (6,5)-chiral SWCNT (Chasm Advanced Materials, SG65i) was added to the isolated EPS and the resulting mixture was tip sonicated (Qsonica Q125) at 4 W for 40 min. After sonication, the suspension was centrifuged for 5 min at 10,000 × g and the probes were extracted as supernatants. These probes were used as stocks to study the enzymatic degradation of the EPS. This procedure was repeated with EPS from various P. aeruginosa strains altered in EPS production, to develop unique biofilm EPS-SWCNT probes.

Fluorescent Assay for Characterization of the Biofilm EPS-SWCNT Probe

Several enzymes with varying bond breaking abilities were used in the study, details of which are presented in Supporting Information 1 and Supporting Information 2.

To understand the response of the biofilm EPS-SWCNT probes to different hydrolytic enzymes, 80 μL of biofilm EPS-SWCNT probe was added to 20 μL of the enzyme in a black walled, 96-well plate. Addition of the biofilm EPS-SWCNT probe to the hydrolytic enzymes initiated the fluorescence test. The resulting fluorescence response generated from the biofilm EPS-SWCNT probe was recorded using a previously described motorized plate-reader⁵⁵. In brief, the probes were excited at 565 nm by an LED and the resulting emission was recorded at 975 nm using an InGaAs detector. All tests were run in quadruplet to confirm the occurrence of an event and for assessing variation in the response.

Enzymatic Degradation of the Biofilm EPS in vitro

To measure the extent to which hydrolytic enzyme treatment degraded EPS components, we treated pre-formed biofilms with the enzymes to assess their ability to break-down a biofilm. The amount of biofilm biomass was determined using a Crystal Violet (CV) biofilm assay that was adapted from previous work^61,62. P. aeruginosa mutants were grown in Lysogeny broth (LB) at 30 °C for 24 h. This liquid culture was diluted to OD₆₆₀ of ~0.1 and then 100 μL aliquots of the diluted cultures was transferred to a clear U-shaped, 96-well plate. The well-plate was incubated at 30 °C for 48 h to allow for biofilm formation and then the plates were rinsed three times by gently immersing the plate in a tub of water to remove unattached, planktonic bacteria. After rinsing, 100 μL of the enzyme was added to the wells, and the plate was again incubated at 30°C for 24 h to provide sufficient time for the enzymes to degrade biofilm EPS components. The contents of the wells were discarded, and the wells rinsed three times by gently immersing the plate in a tub of water prior to adding 125 μL of 0.1% CV to each well and incubating at room temperature for 30 min. Following incubation, the wells were rinsed again to remove excess CV and then 125 μL of 33% acetic acid was added to each well. The plate was again incubated at room temperature for 30 min to dissolve the CV staining the biofilm cells. By measuring the absorbance (540 nm) of CV we quantified the extent to which enzymatic degradation of the biofilm EPS components disrupted biofilm adherence to the plate well.

Statistical Analysis for the Crystal Violet Assay

In the assay to assess hydrolytic enzyme treatment to disrupt biofilms, all experiments were performed in triplicate, each with three technical repeats. This was done to account for the biological variation within the samples. We performed a one-way ANOVA using JMP 15 (SAS, Cary NC) for each strain treated with hydrolytic enzymes to assess their ability to degrade biofilm EPS components. Post-hoc tests were performed using Tukey’s HSD test at p < 0.05.

Results and Discussion

Probe Preparation

We tested seven P. aeruginosa strains that differed either in the type or the amount of EPS produced. Strains PAO1 and derivatives PAO1Δpsl Δpel, PAO1 Δpel, PAO1 Δpsl, PAO1 Δalg were included, as each is unable to make psl, pel, and/or alginate polysaccharides, strain PDO300 is an alginate overproducer, and MJK8 is a small colony variant with altered EPS composition. Small colony variants and alginate overproducers are frequently isolated from cystic fibrosis (CF) patients. From each, biofilm EPS matrix was isolated and used to non-covalently functionalize SWCNT via tip sonication, forming an EPS corona phase (Figure 1(a)). The corona phase around SWCNT is predominantly composed of a blend of different macromolecules namely- polysaccharides, proteins, and extracellular DNAs (quantified in Supporting Information 3).

Figure 1.

Fabrication steps for biofilm EPS functionalized single-walled carbon nanotube (SWCNT) probe (a) (i) Pseudomonas aeruginosa biofilm is grown on a petri dish. (ii, iii) The biofilm is transferred from a petri dish into the saline solution for matrix isolation (iv, v) The biofilm-saline solution is centrifuged and the biofilm extracellular polymer substances (EPS, including DNA, polysaccharide, protein, and lipids) are recovered as supernatant (vi) SWCNT are added to the harvested EPS (vii) The SWCNT-biofilm EPS solution is tip-sonicated (viii) The resulting suspension is centrifuged, and supernatant is extracted, resulting in the solubilized probe (b) Interrogation of the biofilm EPS-SWCNT probe with a custom nIR fluorescent plate reader (c) Sample probe response plotted as % relative signal change vs. time, shown for the PDO300 EPS- SWCNT against the amyloglucosidase enzyme at 300 AGU/ml (n = 4, one standard deviation at each measurement is shown as shaded region). ΔF is the % fluorescence signal change relative to buffer control.

The conformational arrangement of the EPS on the SWCNT is unknown and difficult to model due to the heterogenous composition of the EPS; however, it was sufficiently amphiphilic to make stable dark biofilm EPS-SWCNT probes that could then be tested against hydrolytic enzymes (Supporting Information 4). In general, darker SWCNT suspensions imply that the wrapped polymer around SWCNT is highly amphiphilic and results in higher concentration of SWCNT probe. Next, we performed UV-Vis-nIR absorption spectroscopy to validate two-month stability of the biofilm-EPS SWCNT probes (Supporting Information 5). The peaks at wavelengths of 570 and 990 nm display the excitonic absorption bands E₂₂ and E₁₁ respectively, indicating the presence of (6,5) chiral SWCNT (as expected from vendor specifications). As evident from the spectral graphs, SWCNT probes display dominant excitonic absorption bands even after storing the probes for two months, indicating that SWCNT probes can be stored as a stable library of different biofilm types.

High-throughput Screening of Biofilm Degrading Enzymes

The biofilm EPS-SWCNT probes were tested against a panel of 16 hydrolytic enzymes to assess their ability to degrade EPS components. A subset of enzymes such as lyticase and β-glucosidase were chosen based on their ability to selectively hydrolyze α-(1,4), β-(1,3) and β-(1,4) glycosidic linkages present between the glucose monomers of the exopolysaccharides (pel, psl, and alginate), whereas enzymes such as endo-levanase and xylanase were chosen as sensor controls primarily because the linkages targeted by these enzymes are absent in the biofilm EPS (Supporting Information 1 for more details on the linkages targeted by various enzymes). The extent of biofilm EPS degradation was quantified by monitoring quenching of the SWCNT fluorescence caused by influx of water molecules due to decomposition of the polymer wrapping; sufficient degradation leads to bare SWCNT which agglomerate and self-quench (see Supporting Information 6 for AFM evidence supporting this mechanism). The fluorescence response of the SWCNT was recorded by a custom nIR fluorescent plate reader⁵⁵ (Figure 1 (b)). The response is a quenching curve (Figure 1 (c), data from PDO300 biofilm EPS-SWCNT probe treated with amyloglucosidase). Lack of signal from the heat inactivated control confirmed that the degradation signal was the result of enzyme itself and not from other interaction of enzyme buffer with biofilm EPS constituents. Furthermore, a buffer control was included to account for solvent effects: this was done by subtracting the buffer response from the enzyme-treatments signals and reported as % fluorescence signal change (ΔF). It is important to note that the throughout the course of the study, buffer was used as the control. An attempt to thermally denature the enzymes was initially done, but there was difficulty denaturing some of the enzymes (such as termamyl and subtilisin A).This could be due to engineered thermal resistance or proprietary buffers that protect the enzymes from denaturation; hence for the entire study blank buffer additions were used for the controls.

A high-concentration enzyme screening panel was first conducted to determine which enzymes had clear activity against the biofilm EPS. A complete panel describing the ΔF response of the seven P. aeruginosa strains against the 16 hydrolytic enzyme types is presented in form of a heat map (Figure 2, see Supporting Information 7 for the raw fluorescence plots). For each combination, four experimental runs were conducted, and standard deviations were calculated to roughly estimate the signal to noise ratio for the probe (Supporting Information 8). If the signal change was greater than 3x standard deviation of replicates (measurement noise), the signal was deemed to be acceptable. The enzymes in this heat map were not normalized to a single concentration because of the lack of standardized enzyme activity units. The enzymes that were suspended in a solvent (xylanase, pullulanase, endo-levanase and α-rhamnosidase) were tested at their maximal stock concentrations, whereas the enzymes that were lyophilized (β-glucosidase, lyticase, and alginate lyase) were tested at 10 mg/ml. A complete list of enzymes at their stock concentrations is presented in Supporting Information 2.

Figure 2.

A heat map showing degradation response of biofilm EPS materials obtained from seven different P. aeruginosa strains by a panel of 16 hydrolases with varying bond breaking biases. The extent of biofilm EPS degradation is expressed as ΔF (% fluorescence signal change relative to background control) (Figure 1c).Buffer additions were used as the controls for all the experiments conducted in the above heatmap.

In assessing this screening data, a lower ΔF indicates higher degradation of the biofilm EPS matrix. A positive ΔF value indicates binding of the enzyme on the nanotube (shielding the SWCNT from further quenching effects). Out of the panel, β- glucosidase, alginate lyase, lyticase, hyaluronidase, amyloglucosidase, and subtilisin showed the most degradation. Our results for β- glucosidase, lyticase, alginate lyase align with prior works^31,63 that examined the efficacy of these enzymes in degrading P. aeruginosa biofilms. Furthermore, there was minimal degradation observed by α-rhamnosidase and DNase despite the presence of the rhamnose (in psl) and eDNA in the biofilm EPS, respectively. This apparent lack of degradation is attributed to strong binding of rhamnose and eDNA to the SWCNT surface which prevents the depolymerization or inaccessibility of these macromolecules by enzymes due to steric hindrance. In contrast, enzymes such as xylanase, pullulanase, and endo-levanase showed an increase in fluorescence signal, indicating binding to the SWCNT rather than degradation of biofilm EPS. Lack of a degradation signal from these enzymes indicate that the glycosidic linkages targeted by these enzymes were absent or inaccessible.

Further inspection of fluorescence responses from PAO1 and its derivatives (PAO1Δpsl Δpel, PAO1 Δpel, PAO1 Δpsl, PAO1 Δalg) reveals some indication of selectivity of enzymes towards pel and psl polysaccharides. For example, the response of lyticase to PAO1 EPS is −10 %, however, when testing this enzyme with the double knockout PAO1 Δpsl Δpel EPS, the response increases to −4.1 % (reduced enzyme activity), and then returns to −8.5% with PAO1 Δalg. This change in fluorescence intensity implies that lyticase targets the majority of linkages present in pel and psl. The residual response of PAO1 Δpsl Δpel EPS (−4.1%) could be attributed to the dual functionality of lyticase towards other glycosidic linkages present in the biofilm EPS. These similar trends are also observed in other enzymes like β-glucosidase and alginate lyase.

Concentration dependent Studies and Kinetic modeling of biofilm EPS degradation

Enzymes that showed the most degradation from the initial screen (β-glucosidase, hyaluronidase, subtilisin, amyloglucosidase, lyticase, and alginate lyase) were then used in concentration dependent studies on two P. aeruginosa phenotypes commonly isolated from the CF lung environment, an alginate overproducer (PDO300) and a small colony variant (MJK8). End point data clearly depict anticipated concentration dependence, i.e., the fluorescence signal decreases with increasing enzyme concentration (Figure 3).

Figure 3.

Concentration dependence of hydrolytic enzymes on degrading biofilm EPS obtained from two P. aeruginosa phenotypic variants commonly isolated from the CF lung environment. (a) PDO300-alginate overproducer. (b) MJK8-small colony variant. β-glucosidase, hyaluronidase, lyticase, and alginate lyase were at a stock concentration of 10 mg/ml whereas subtilisin and amyloglucosidase were at 2.60 AU/g and 300 AGU/ml, respectively. The error bars represent one standard deviation (n =4). Buffer was used as control in the above concentration enzyme panel.

The dynamic response data was then used to assess kinetics governing biofilm EPS degradation. A three-site, first-order kinetic model is proposed which effectively segregates the EPS into three parts: fast, medium, and slow degradability component based on substrate accessibility. In reality, there would be a broader spectrum of sites, but this simplification is more tractable mathematically (rather than an infinite set of parallel reactions) and provides physical intuition to substrate accessibility. The three-independent reactions in parallel are as follows

$$\mathit{\text{Slow}}\mathit{\text{Degradation}}\mathit{\text{Rate}}=-\frac{d{S}_1}{dt}=k_{1}E{S}_1$$ (1)

$$\mathit{\text{Medium}}\mathit{\text{Degradation}}\mathit{\text{Rate}}=-\frac{d{S}_2}{dt}=k_{2}E{S}_2$$ (2)

$$\mathit{\text{Fast}}\mathit{\text{Degradation}}\mathit{\text{Rate}}=-\frac{d{S}_3}{dt}=k_{3}E{S}_3$$ (3)

Integrating equation (1), (2), and (3) and combining leads to

$$X_1=\frac{S_{1t}}{S_{10}}=\text{exp}\left(-k_{1}Et\right)$$ (4)

$$X_2=\frac{S_{2t}}{S_{20}}=\text{exp}\left(-k_{2}Et\right)$$ (5)

$$X_3=\frac{S_{3t}}{S_{30}}=\text{exp}\left(-k_{3}Et\right)$$ (6)

$$X_{\mathit{\text{total}}}= a{X}_1+b{X}_2+c{X}_3$$ (7)

Where S_o is the initial substrate concentration, S_t is the substrate concentration at time (t), E is the enzyme concentration (constant) and k₁, k₂, and k₃ are first-order rate constants which are assigned to the slow, medium, and fast parts of the parallel reaction. Note that the ratio of current substrate to start substrate can be used to describe the extent of reaction for the slow, medium, and fast reactions (X₁, X₂ and X₃) (Supporting Information 9 for complete model derivation). These are combined for total extent of reaction, X_total and a, b, and c are constants that describe the relative contributions of the individual reactions (a + b + c = 1). The total extent of reaction is then corelated to the normalized change in fluorescence.

The concentration gradient tests of enzymes were normalized to correlate to extent of signal attenuation (Supporting Information 10), and the combined kinetic model (total extent of reaction) was fit to each enzyme-biofilm EPS data set to extract the first-order rate constants (k₁, k₂, and k₃) and the relative contributions of the slow, medium, and fast reactions (a, b, and c) (Supporting Information 11, MATLAB code that was used to perform the fitting can be found in Supporting Information 12). A sample fitting of the model to the lyticase concentration gradient tests of PDO300 and MJK8 is shown (Figure 4) and the remaining concentration gradient tests with the kinetic model fit can be found in Supporting Information 13. During the course of fitting, enzyme concentrations that did not show any degradation (binding only) were omitted primarily because the proposed model is not well suited for binding kinetics (Figure 4 (b)).

Figure 4.

Concentration dependent studies of lyticase on (a) PDO300 EPS- SWCNT and (b) MJK8 EPS- SWCNT with kinetic fitting of the multi-site degradation model (dashed lines, single model for all concentrations). The shaded region represents one standard deviation (n =4)

The extracted kinetic parameters (Supporting Information 11) reveal several interesting features about the degradation kinetics of the biofilm EPS. The relative contribution of the slow reaction sites is significantly more than the other two reactions (a > 0.8 or 80% contribution), which implies that most of the biofilm EPS is relatively inaccessible to hydrolysis and slower to cleave. The medium-rate degradation is >2x the fast rate except for lyticase, alginate lyase, subtilisin, and hyaluronidase for the alginate overproducer (PDO300), indicating that these enzymes have greater effect on degrading this EPS. The hyaluronidase effect on the alginate overproducer might appear to be an outlier, with the medium kinetic rate dominating, however inspecting the rates shows that the slow and medium constants are nearly the same, and thus would be lumped together as the same slow reaction rate.

Testing Selectivity of the Biofilm EPS-SWCNT Probe

Ethanol precipitation was performed on PDO300 biofilm EPS to purify polysaccharides. These purified polysaccharides were functionalized on SWCNT to assess selectivity of the biofilm EPS probe, by evaluating DNase, proteinase K, and alginate lyase (Figure 5). The dominant signal change (−13.89 %) was produced by the alginate lyase, which has activity against polysaccharides. A small fluorescence response from the DNase (−3.54 %) and proteinase K (−2.91 %) were also observed. This is likely due to presence of small amounts of DNA and protein in the purified polysaccharides sample. Selectivity could be further improved by optimizing the ethanol precipitation protocol. Notwithstanding, this data indicates that single EPS components can be placed on different SWCNT probes to provide a multiplexed assay. This could be done by exploiting unique SWCNT chiral signatures and new, efficient methods of chiral separation⁶⁴.

Figure 5.

Selectivity test of alginate lyase, DNase, and proteinase K on purified PDO300 polysaccharide-SWCNT probe. Shaded region represents one standard deviation (n = 4) with buffer as control.

An additional observation from the purified polysaccharide probe response to alginate lyase is that the response increased significantly (−13.89% purified PDO300 polysaccharide vs −5.52% crude PDO300 biofilm EPS SWCNT probes treated with alginate lyase, compared to Figure 2). The increase in signal intensity can be attributed to the purification protocol, which removes other macromolecules. Thus, when SWCNTs are functionalized with purified polysaccharides there is likely less steric hindrance and more coverage of the SWCNT surface with a specific polysaccharide. When the enzymes approach the SWCNT surface, the substrate is easily accessible, resulting in more degradation, and higher signal intensity. Other purification protocols could be implemented and optimized to target other EPS components and thus expand the utility of these probes to assess the selectivity of the hydrolytic enzymes towards specific biofilm macromolecules.

Comparison of the Biofilm EPS-SWCNT Probe with the Crystal Violet Staining Assay.

We wanted to assess whether the selected hydrolytic enzymes were able to disrupt P. aeruginosa biofilms, thereby providing confidence that SWCNT probes could be used to effectively screen for hydrolytic enzymes capable of degrading intact biofilms. Our approach was to expose preformed biofilms with the enzymes and to measure the remaining biofilm biomass using a widely used crystal violet (CV) (Figure 6). A lower signal means a reduction in biofilm material caused by enzymatic treatment. For the assay, a total of six enzymes (three that demonstrated strong degradation response and three poor) were chosen from the initial screening (Figure 2). Furthermore, for positive and negative controls untreated biofilms (without the enzyme treatments) and sterile broth respectively were used.

Figure 6.

Enzymatic degradation of (a) PDO300-Alginate Overproducer and (b) MJK8-small colony variant biofilm EPS characterized with the crystal violet assay. Alginate lyase and Carolina cellulase were used at 10 mg/ml whereas subtilisin, amyloglucosidase, pullulanase, and termamyl were used 2.60 AU/g, 300 AGU/ml, 1000 NPUN/g, and 120 KNU-T/g, respectively. The error bars represent one standard deviation of three biological replicates. An * indicates biofilm biomass was significantly lower in the enzyme treatments compared to the untreated control based on Tukey’s HSD test (P < 0.05).

Assessing the results from the biofilm assay, the CV tests partially corroborate the SWCNT probe results. For the PDO300 biofilm, alginate lyase, amyloglucosidase, subtilisin, pullulanase, and termamyl showed significant reduction in biofilm biomass (P < 0.05), whereas the Carolina cellulase (P > 0.05) did not (detailed statistical results in Supporting Information 14). All the results except pullulanase and termamyl corresponded with results from the PDO300 EPS-SWCNT probe. Similarly, for MJK8 biofilm, all the enzymes except pullulanase and Carolina cellulase were congruent with the response of the MJK8 EPS-SWCNT probes.

The incomplete agreement between the two methods can be attributed to technical shortcomings of the CV assay as well as fundamental differences in the biofilm EPS structure present in each assay. The CV assay relies on two manual washing steps designed to remove unattached bacteria and unbound CV from the walls of the microtiter plate wells³⁹. There can be great variability in the intensity of the rinsing steps which can inadvertently dislodge biofilm cells, leading error in estimation of the biofilm biomass. This error in estimation is evident when we replicate these CV tests and find varying, relative levels of enzyme susceptibility (Supporting Information 15, repeated the assay in LB medium). Furthermore, the assay also lacks an internal standard, which makes it difficult to compare results between experiments or studies and to access reproducibility of the data^39,65. For example, some studies have performed the CV assay in the LB media⁶⁶ while others have used TSB media⁶⁷. The difference in the inoculum media can highly impact the cell-wall interactions in the microtiter plate and can consequently affect the biofilm formation. Also, it is important to note that the biofilm structure on the microtiter wall is fundamentally different than the EPS absorbed on the SWCNT. The SWCNT is more accessible to degradation in solution than a three-dimensional biofilm which limits diffusion of hydrolytic enzymes, such as the cellulase in the PDO300 biofilm. These points highlight limitations to comparison and the need for better standard assays.

It is important to note that biofilm EPS-SWCNTs, though capable of transducing degradation efficiently, are still subject of batch-to-batch variation which currently limit their ability for relative comparison and high-throughput design campaigns. As evident from the raw fluorescent curves (Supporting Information 7), the response due to buffer addition (control) varies with each sensor preparation. Most often the signal decreases indicating susceptibility of SWCNT surface to buffer change. Sometimes there is a small increase in signal, indicating a rearrangement of the EPS coating that protects the surface from solvent effects. Therefore, for each experiment and each newly synthesized probe we conducted the buffer control and compared sensor response relative to that. This batch-to-batch variation is a challenge when working with colloidal probes. Main sources of variability occur during the sample preparation. Despite careful attempts at weighing out materials and controlling process parameters, each batch of probes are prepared with slightly different amounts of SWCNT and subjected to different levels of sonication power caused by different tip sonication positions⁶⁸. This leads to variation in the amount of surface defects and lengths, which in turn affects the fluorescence intensity and sensor response^69–71. These variations among the biofilm EPS-SWCNT probes can be improved by using automated sample preparation (robotic sonication) or other suspension techniques that have less variability (such as suspending a large batch with surfactants and then dialyzing to gently introduce EPS wrappings). These would improve consistency in nanotube preparation and minimize variation among samples, key steps needed for using this technique in high throughput screening

Conclusions

Degradation and permeabilization of biofilm EPS present an effective route to improve susceptibility to antibiotics and allows access of native immune cells. In this study, we have demonstrated that SWCNT probes functionalized with biofilm EPS can be used to screen for hydrolytic enzymes that degrade target biofilm EPS. The unique sensitivity of SWCNT to modifications of the EPS allows for fast response times (<40 min for 96 well plate) and makes the probes suitable for high throughput screening; moreover, SWCNT probes can be stored for months as a library of different biofilm EPS types thereby avoiding the bottleneck of working with living cells. Furthermore, we have also shown that a multi-site degradation model can be fit to the raw data to understand the kinetics governing the EPS degradation. Due to the heterogeneity of the biofilm, it is difficult to thermodynamically model the biofilm EPS on SWCNT; however, a simplified coarse-grained, molecular dynamics model may be used in the future to understand the mechanisms governing EPS degradation. In addition, we have also shown that enzyme substrate preferences can be transduced with the biofilm EPS-SWCNT probes when using purified EPS materials. Future works can exploit the distinct spectral signatures of different SWCNT chiral species to multiplex probes with different EPS materials. These biofilm EPS-SWCNT probes are capable of providing real-time data, which is a distinct advantage over the common end-point assays like biofilm adherence assays. The real time responses can be used to assess the degradation profile of the biofilm EPS and further determine the dosage concentration of the antibiotics. In summary, our work presents biofilm EPS-SWCNT probes that are amendable to high throughput screening of biofilm EPS degrading enzymes. This unique functionality can be used to accelerate the rate of discovery and development of hydrolase-based therapeutics as well as enzyme cocktails used to combat biofilms that foul mechanical or medical device surfaces.