Transport of microorganisms into cellulose nanofiber mats

Abstract

Nanofiber mats hold potential in numerous applications that interface with microorganisms. However, a fundamental study that quantifies the transport of microorganisms into three-dimensional microenvironments, such as nanofiber mats, has not yet been conducted. Here, we evaluate the microbial uptake capacity of three hydrophilic cellulose sorbents, a high surface area electrospun nanofiber mat, as well as two commercial products, a macrofibrous Fisherbrand fabric and an adsorptive Sartorius membrane. The small average fiber diameter (∼1.0 μm) and large porosity of the nanofiber mats enabled a 21 times greater collection of Escherichia coli K12 per milligram of material than the macrofibrous Fisherbrand controls and 220 times more than the Sartorius controls. In most cases, the exposure time of the nanofiber mats to the microorganisms was sufficient to reach a quasi-equilibrium state of microbial uptake, allowing the calculation of an adsorption coefficient (K_eq) that relates the concentration of cells in the sorbent to the concentration of cells remaining in solution. The K_eq of the nanofiber mats was 420, compared to 9.2 and 0.67 for the Fisherbrand and Sartorius controls, respectively. In addition to E. coli, we studied the cellulose nanofiber mat uptake of two additional medically relevant and distinct microorganisms, Gram-negative Pseudomonas aeruginosa PA01 and Gram-positive Staphylococcus aureus MW2, to probe whether microorganism removal is bacteria-specific. The high uptake capacity of all three bacteria by the nanofiber mats indicates that microbial uptake is independent of the microorganism's adhesion mechanism. This work suggests that cellulose nanofiber mat “sponges” are a green platform technology that has the potential to remove detrimental microorganisms from wounds, trap bacteria within a protective military textile, or remediate contaminated water.

---

Introduction

Electrospinning is a scalable technology capable of producing non-woven mats composed of continuous nano- to microscale diameter fibers from a variety of natural and synthetic polymers.^1–3 Due to their interconnectivity, microscale interstitial spaces, large surface-to-volume ratio, and porosity values greater than 80%,^4,5 electrospun nanofiber mats have been investigated for use in biomedical applications. The intrinsic properties of nanofiber mats have been reported to promote cell respiration, skin regeneration, moisture retention, and removal of exudates.⁶ By controlling the delivery of antibacterial agents from nanofiber mats the bacteria that cause infections can be combatted.^7,8 We suggest that there is an additional untapped opportunity to employ nanofiber mats as “sponges” that can remove bacteria from wounds; this novel application of nanofiber mats would take advantage of their intrinsic and engineered properties.⁹

For example, 25.8 million U.S. patients suffer from diabetes, while another 79 million individuals are at risk for developing the disease.¹⁰ When the disease causes bacterial biofilms¹¹ and/or metabolic dysregulation in patients,¹² their wounds remain perpetually arrested in the inflammatory phase of the healing process.¹³ New bandages that could “suck-out” the bacteria from chronic wounds to effectively remove the bacteria from the infection site and/or kill the bacteria trapped within the nanofiber mat would be beneficial. However, this relies on understanding the bacteria–nanofiber interface and to date, a fundamental study that quantifies the uptake of microorganisms into a nanofiber matrix has not yet been conducted.

For the first time, we systematically quantify the ability of cellulose nanofiber mats to remove both Gram-negative and Gram-positive bacteria from solution. Cellulose was chosen because it is the most abundant natural polymer, hydrophilic,^14–16 and an almost inexhaustible source for environmentally friendly and biocompatible products that interface with microorganisms, including, wearable electronics, food packaging, water remediation technologies, and biomedical devices.^17–20 The capacity of electrospun cellulose nanofiber mats to adsorb Escherichia coli K12 (E. coli K12) was compared to two commercial cellulose controls. Also, by experimentally varying the initial E. coli K12 concentration and the nanofiber mat diameter, we developed models to quantify the quasi-equilibrium and dynamic behavior of bacteria uptake using a three-dimensional nanofiber environment. Whereas previously, modeling the transport of bacteria into porous media has been limited to soil components.^21–23 Our results highlight the promise of highly porous lightweight cellulose nanofiber mats to serve as microorganism “sponges”.

Experimental

Materials and methods

All compounds were used as received. Cellulose acetate (M_w = 30 000 Da), acetone, analytical reagent grade acetic acid (AA), sodium chloride (NaCl), and calcofluor white stain were obtained from Sigma-Aldrich (St. Louis, MO). Sodium hydroxide (NaOH) and Fisherbrand Cellulose Paper (09-801C, Fisherbrand control) was purchased from Fisher Scientific (Fair Lawn, NJ). A commercially available regenerated cellulose adsorptive membrane (1401213, Sartorius control) was purchased from Sartorius Stedim Biotech, Germany. Difco Luria–Bertani (LB) broth was purchased from BD Life Sciences (Franklin Lakes, NJ). Deionized (DI) water was obtained from a Barnstead Nanopure Infinity water purification system (Thermo Fisher Scientific, Waltham, MA).

Cellulose nanofiber mat fabrication

A 15% w/v solution of cellulose acetate in acetone²⁴ was mixed for 24 h at 20 rpm using an Arma-Rotator A-1 (Bethesda, MA). The solution was loaded into a 5 mL Luer-Lock tip syringe capped with a Precision Glide 18 gauge needle (Becton, Dickinson & Co. Franklin Lakes, NJ), which was secured to a PHD Ultra syringe pump (Harvard Apparatus, Plymouth Meeting, PA). Alligator clips were used to connect the positive anode of a high-voltage supply (Gamma High Voltage Research Inc., Ormond Beach, FL) to the needle and the negative anode to a copper plate wrapped in aluminum foil. A constant feed rate of 3 mL h⁻¹, an applied voltage of 25 kV, and a separation distance of 10 cm were used to spin cellulose acetate nanofiber mats. The assembled electrospinning apparatus was housed in an environmental chamber (CleaTech, Santa Ana, CA) with a desiccant unit (Drierite, Xenia, OH) to maintain a temperature of 22 ± 1 °C and a relative humidity of 55%. In this study, all cellulose acetate nanofiber mats were electrospun for 1 h before being converted to cellulose nanofiber mats. As-spun mats sandwiched between Teflon sheets were thermally treated at 208 °C for 1 h and then submerged in a 4 : 1 v/v solution of H₂O/ethanol containing 0.1 M NaOH for 24 h.^14,25 The mats were then washed using DI water and placed in a desiccator for 24 h at room temperature (23 °C) to dry.

Characterization of nanofiber mat, Fisherbrand control, and Sartorius control

Micrographs were acquired using a FEI-Magellan 400 scanning electron microscope (SEM). A sputter machine (Gatan high resolution ion beam coater model 681) was used to coat samples with ∼5 nm of platinum. Fiber diameter distribution was determined using Image J 1.45 software (National Institutes of Health, Bethesda, MD) by measuring 50 random fibers from 5 micrographs. A PerkinElmer Spectrum 100 Fourier transform infrared spectrometer (FTIR) confirmed that cellulose was regenerated after alkaline treatment of the as-spun cellulose acetate nanofiber mats.

A Zeiss Axiovert 4-laser spinning disc confocal microscope (Zeiss confocal, 20× magnification) was used to collect z-stack composite images of cellulose nanofiber mats fluorescently dyed with calcofluor white stain (1 μL mL⁻¹). The 3D composite images from Zen software were imported to Image J 1.45 software from which the average thickness of the nanofiber mats was determined by averaging 50 thickness measurements taken from 5 different nanofiber mats (Fig. S1 of the ESI†). The thicknesses of the Fisherbrand and Sartorius controls were measured using a Mitutoyo micrometer (Aurora, IL). Average thickness was multiplied by surface area to calculate nanofiber mat volume, V_mat. The total internal surface area of the nanofiber mats was estimated using an Autosorb®-iQ system (Quantachrome) using 50 mg of nanofiber mat that were degassed for 2 h at 150 °C. The total surface area was calculated for the nanofiber mat using the Brunauer–Emmett–Teller (BET) method.²⁶ The surface area of the Fisherbrand and Sartorius controls was too low to be estimated using BET.

Quantification of bacteria uptake by the nanofiber mat, Fisherbrand control, and Sartorius control

Escherichia coli K12 (E. coli K12) and Pseudomonas aeruginosa PA01 (P. aeruginosa PA01) were purchased from Leibniz Institute DSMZ (Germany). Staphylococcus aureus MW2 (S. aureus MW2) was a kind gift from Prof. Neil Forbes at the University of Massachusetts Amherst. Bacteria were grown in Luria Broth at 37 °C and re-suspended in a phosphate buffered saline solution (PBS, pH 7.2) to remove residual macromolecules and other growth medium constituents. Throughout the experiment, no external forces were applied.

Using 6-well plates, each porous sorbent (nanofiber mat, Fisherbrand control, and Sartorius control) was punched into a circle with a diameter of 2.54 cm and incubated in a bacterial solution with an initial concentration of 1.52 × 10⁸ cells per mL (5 mL per well). A control sample (no sorbent, bacteria solution only) was run in parallel to each experiment and six trials for each sorbent were performed. Additional experiments were conducted as a function of nanofiber mat diameter (2.54, 2.22, 1.91, and 1.27 cm) and initial E. coli K12 concentration (1.52, 4.46, 6.32 × 10⁸ cells per mL). For all experiments, the sorbents were incubated for 2 h at 37 °C at 150 rpm; over this period of time a portion of the bacteria transferred from the surrounding solution to the sorbent. The optical density of the solution in both the sample and control well were monitored using the McFarland 0.5 standard, which is equal to 1 to 2 × 10⁸ cells per mL.²⁷ Concentrations were measured using a BioTek EL×800 Absorbance Microplate Reader at an absorbance of 600 nm. A calibration curve (Fig. S2 of the ESI†) was developed to convert the microplate reading to optical density, and then to cell concentration. These concentrations were confirmed using plate counting. To calculate the total number of cells removed by the sorbent at time, t, we calculated the difference between the concentration of bacteria in the sample well containing a sorbent and the concentration of bacteria in the control well.

A Zeiss confocal (20× magnification) was used to (1) collect micrographs that qualitatively confirmed E. coli K12 collection by the nanofiber mat, Fisherbrand control, and Sartorius control, and (2) to examine where within the sorbents the bacteria collected. Cellulose was dyed fluorescently with calcofluor white stain (1 μL mL⁻¹) and E. coli K12 was tagged using a plasmid green fluorescent protein. Zen software was used to generate 3D composite images.

Modeling of bacteria uptake by the nanofiber mat

The quantity of E. coli K12 collected at the longest time (t = 120 min), for the largest mat (2.54 cm), as a function of initial cell concentration was used to determine a quasi-equilibrium model. The adsorption coefficient, K_eq, from the equilibrium model was utilized to build a dynamic model to capture the transient behavior^28,29 of the collection of bacteria into nanofiber mats of various mat diameters. A first-order kinetic model, with external transport mechanisms assumed to be controlling, was postulated based on observations of the data. The model was fit to time-dependent uptake data for the four different mat diameters, at one initial concentration. The dependence of the fitted rate constant k_m on mat diameter was examined. Full details of the modeling procedures are provided in the ESI.†

Results and discussion

Characteristics of the nanofiber mat, Fisherbrand control, and Sartorius control

Cellulose acetate was successfully electrospun and converted via alkaline treatments to regenerated cellulose nanofiber mats (Fig. 1). Consistent with literature,^14,24 the as-spun cellulose acetate nanofibers, as well as the cellulose nanofibers had a ribbon morphology (Fig. S3 of the ESI†). The materials' properties of the nanofiber mats are summarized in Table 1. From SEM micrographs, the average fiber diameter of the cellulose nanofiber mats was determined to be 1.08 ± 0.46 μm. Fig. S3E of the ESI† displays the fiber diameter distribution of the electrospun nanofiber mats. FTIR spectra of the as-spun cellulose acetate and the cellulose nanofiber mats is displayed on Fig. S4 of the ESI.† Notably, the disappearance of the 1750 cm⁻¹ peak indicates that the acetate groups have been replaced with hydroxyl groups.²⁵ The total surface area of the cellulose nanofiber mats was estimated to be 4.5 m² g⁻¹, which is consistent with literature, see Fig. S4 of the ESI.†³⁰ All nanofiber mats were electrospun for 1 h; the bulk thickness of the nanofiber mats was determined to be 42.4 ± 12 μm using confocal microscopy.

Fig. 1 SEM micrographs display the morphology and average fiber diameter of cellulose sorbents: nanofiber mat, Fisherbrand control, and Sartorius control.

Table 1 Summary of the materials properties of cellulose sorbents: electrospun nanofiber mat, Fisherbrand control, and Sartorius control

Material property	Nanofiber mat				Fisherbrand	Sartorius
Fiber diameter (μm)	1.08 ± 0.46				16.5 ± 9.6	0.33 ± 0.2
Total surface area (m^2 g^−1)	4.5				N/A	N/A
Mat thickness (μm)	42.4 ± 12				151.6 ± 4.9	154.7 ± 5.5
Mat diameter (cm)	2.54	2.22	1.91	1.27	2.54	2.54
Mat volume (cm^3)	0.0215	0.0164	0.0121	0.0054	0.0765	0.0780
Mat weight (mg)	9.3 ± 1.8	6.18 ± 1.0	4.0 ± 0.8	1.8 ± 0.3	35.6 ± 1.0	37.4 ± 0.7
Density (mg cm^−3)	432.6	376.8	330.6	333.3	465.4	479.5

---

We compared our nanofiber mats to two model commercial controls, Fisherbrand and Sartorius. All three sorbents had the same cellulose chemistry – a promising polymer for protective clothing,³¹ wound bandages,³² filtration, and adsorption,¹⁵ as well as a similar fiber morphology (on different length scales). Additionally, the Fisherbrand and Sartorius materials have previously served as controls to electrospun cellulose nanofiber mats in studies that quantified the adsorption of proteins.^33,34 The materials properties of the Fisherbrand and Sartorius controls are summarized in Table 1. The Fisherbrand control contains cellulose macrofibers with an average fiber diameter of 16.5 ± 9.6 μm, an order of magnitude larger than our electrospun nanofiber diameter. The fiber diameter distribution of the Fisherbrand controls is provided on Fig. S3 of the ESI.† While individual Fisherbrand control fibers are generally cylindrical and continuous, they appeared less smooth than the electrospun nanofibers. The average thickness of the Fisherbrand control was 151.6 ± 4.9 μm. Analogous to the nanofiber mat, the Sartorius control is regenerated cellulose; micrographs display that it had a webbed-fiber morphology, Fig. 1. The average diameter of the webs were 0.33 ± 0.2 μm, the pores had an average diameter of 1.70 ± 0.6 μm, and the overall thickness of the Sartorius control was determined to be 154.7 ± 5.5 μm. The surface area was too low to be estimated for the Fisherbrand and Sartorius controls.

E. coli K12 uptake by the nanofiber mat, Fisherbrand control, and Sartorius control

While the use of nanofiber mats as effective size selective sieves has been extensively studied,^35–38 the demonstration that nanofiber mats can serve as a simple, inexpensive way to uptake bacteria from a wound is an emerging concept that has not yet been quantified.^9,39 Ideally, a lightweight nanofiber mat could simply be placed within the contaminated area and removed after bacteria have transported into the mat.

The total number of E. coli K12 cells collected over 120 min by the electrospun nanofiber mat, Fisherbrand control, and Sartorius control is shown in Fig. 2A. After 120 min, the nanofiber mat collected 4.2 and 55.3 times more E. coli K12 cells than the Fisherbrand and Sartorius controls, respectively. Per milligram, the nanofiber mat collected 21 and 220 times more E. coli K12 than the Fisherbrand and Sartorius controls, respectively. After 15 min, the electrospun nanofiber mats statistically, collected significantly more cells than both controls. After 120 min, all materials have collected a statistically different quantity of E. coli K12. When we compare our nanofiber mat to data previously published on activated carbon, our nanofiber mat uptakes approximately 16 times more E. coli K12 cells per milligram of material.⁴⁰ Digital images taken before and after the nanofiber mats were used as sponges show that the water visually changed from opaque to clear indicating that there was a significant amount of microorganisms removed (Fig. 2B). The number of E. coli K12 cells adsorbed per mass, volume, and surface area are summarized in Table 2.

Fig. 2 (A) The total number of E. coli K12 collected by the nanofiber mat, Fisherbrand control, and Sartorius control over 120 min. Error bars represent the standard deviation of six trials. Statistical significance at 120 min is denoted by an asterisk. (B) Digital pictures of the initial cloudy E. coli K12 solution (time = 0 min) and the clear solution after 120 min treatment with a nanofiber mat. (C) Fluorescent micrographs of nanofiber mat, Fisherbrand control, and Sartorius control acquired after 120 min of incubation with E. coli K12. All cellulose sorbents fluoresce blue, and E. coli K12 fluoresce green. The scale bar displayed is 20 μm. For (A–C) the initial concentration of E. coli K12 was 1.52 × 10⁸ cells per mL and a constant sorbent diameter of 2.54 cm was used.

Table 2 Summary of the E. coli K12 removal capacity of cellulose sorbents: electrospun nanofiber mat, Fisherbrand control, and Sartorius control

	Total E. coli K12 at t = 120 min
	Nanofiber mat				Fisherbrand	Sartorius
Mat diameter (cm)	2.54	2.22	1.91	1.27	2.54	2.54
Adsorption (cells)	4.61 × 10^8	4.32 × 10^8	3.59 × 10^8	1.16 × 10^8	1.11 × 10^8	8.34 × 10^6
Adsorption (%)	72	52	44	14	12	1
Adsorption per weight (cells per mg)	4.98 × 10^7	6.99 × 10^7	8.97 × 10^7	6.43 × 10^7	3.11 × 10^6	2.23 × 10^5
Adsorption per volume (cells per cm^3)	2.15 × 10^10	2.63 × 10^10	2.97 × 10^10	2.16 × 10^10	1.45 × 10^9	1.07 × 10^8
Adsorption per surface area (cells per m^2)	1.11 × 10^10	1.55 × 10^10	1.99 × 10^10	1.43 × 10^10	N/A	N/A

---

Fluorescent micrographs in Fig. 2C were used to (1) confirm qualitatively that the nanofiber mats collected more bacteria than the controls and to (2) gain insight into where the bacteria physically collected on the cellulose sorbents. Visually, it is evident that many more green fluorescent E. coli K12 cells collected on the nanofiber mats than on the Fisherbrand or Sartorius controls. On the nanofiber mat, the bacteria appear to be both attaching to the nanofibers and filling the void space between the nanofibers. Confocal z-stack images were acquired at three points during the bacteria collection experiments (t = 15, 60, and 120 min), which confirmed that bacteria were always present throughout the entire nanofiber mat; after only 15 min of E. coli K12 collection no gradient within the nanofiber mat was observed.

The Fisherbrand control exhibited poor cell adsorption. Fewer bacteria adsorbed onto the Fisherbrand control than the nanofiber mat and a majority of the E. coli K12 appeared to be attached to the macrofibers. The Fisherbrand control had a statistically higher average fiber diameter than our nanofiber mats. This suggests that increasing the fiber diameter of a sorbent does not increase adsorption. The lowest amount of bacteria uptake was achieved by the Sartorius control, where cells did not appear to be preferentially adsorbing onto the webbed-fibers or within the pore voids. This suggests that simply continuing to decrease the fiber diameter alone is not sufficient to increase bacteria uptake into fiber mats. Previous literature reports that higher adhesion of bacteria is possible by using surfaces that conform to their size, offer surface roughness, are porous, or have a higher surface area.⁴¹ Our findings demonstrate that the E. coli K12, which are ∼0.5 μm in width by 2 μm in length, prefer the high porosity nanofiber mats with a ∼1 μm diameter over the Fisherbrand or Sartorius controls.

Equilibrium and diffusion kinetics of bacteria uptake by the nanofiber mat

Adsorption isotherm models, predominantly Langmuir isotherms, have been used to characterize microbial behavior in porous media.²¹ To explore the applicability of an isotherm model to our nanofiber mats, we experimentally varied the initial E. coli K12 concentration (1.52 × 10⁸, 4.46 × 10⁸, 6.32 × 10⁸ cells per mL) and measured uptake over time for the nanofiber mats with a diameter of 2.54 cm. The experimental data collected (Fig. 3B) suggest that an equilibrium state with respect to the removal of cells from the solution is reached by 120 min for each of the initial E. coli concentrations.

Fig. 3 (A) Summary and color representation of the different conditions tested, including, three initial E. coli K12 concentrations (1.52 × 10⁸, 4.46 × 10⁸, 6.32 × 10⁸ cells per mL) and four nanofiber mat diameters (1.27, 1.91, 2.22, 2.54 cm). (B) The total number of E. coli K12 cells collected as a function of initial cell concentration and time are displayed. The diameter of the nanofiber mat (2.54 cm) was held constant. (C) The concentration of E. coli K12 collected by the nanofiber mat vs. the concentration remaining in the bulk solution at t = 120 min. The subsequent quasi-equilibrium model (line) is derived from experimental data collected by varying the initial cell concentration (B). The experimental data collected by varying nanofiber mat diameter (from D), as well as the Fisherbrand control (black) and Sartorius control (red) are plotted to benchmark the model. (D) The total number of E. coli K12 collected as a function of nanofiber mat diameter and time. The initial concentration of cells was held constant at 1.52 × 10⁸ cells per mL. The dynamic model (solid lines) predicts the uptake of bacteria as a function of nanofiber mat diameter and time. (B and D) Error bars represent the standard deviation of six trials. (E) The rate constants determined from the dynamic model increase with increasing nanofiber mat diameter.

Fig. 3C is an adsorption isotherm created from the longest-time data points from Fig. 3B. Here, the concentration of bacteria taken up by the 2.54 cm diameter nanofiber mat is plotted as a function of the concentration of bacteria remaining in the bulk solution, at t = 120 min. The three data points (darkest blue, yellow, green) lie on a straight line, suggesting that we are in the Henry's law or linear regime of adsorption. The slope of a straight line fit of the data is the equilibrium adsorption coefficient, K_eq, which indicates the uptake capacity of the nanofiber mats, per the equation:

c_m = K_eqc_b

where c_m is the concentration of bacteria inside the mat and c_b is the concentration of bacteria in the bulk solution. K_eq was determined to be 420 ± 82, with 95% confidence bounds; this means that a nanofiber mat will contain ∼420 times as many E. coli K12 cells as an equal volume of solution at equilibrium. The K_eq of the Fisherbrand and Sartorius controls, were determined to be much lower 9.24 and 0.67, respectively. These values are plotted on Fig. 3C as black and red points that lie well below the isotherm for the nanofiber mats, thus reiterating the impressive removal capability of the nanofiber mats. The data represented by the lighter-blue symbols in Fig. 3C are described next.

To further understand the dynamics of the E. coli K12 uptake into the mats, we measured bacterial uptake over time for nanofiber mats of various diameters (1.27, 1.91, 2.22, and 2.54 cm) starting from the same initial concentration of 1.52 × 10⁸ cells per mL, Fig. 3D. After 120 min, it appeared that all of the nanofiber mats have (at least nearly) reached quasi-equilibrium except for the smallest diameter nanofiber mat (lightest blue circle), for which the data have still not begun to plateau. The overlay of the t = 120 min data for the three smaller-diameter mats onto the quasi-equilibrium model (Fig. 3C) confirms that all mats have reach quasi-equilibrium except the smallest nanofiber mat (lightest blue circle), which lies below the isotherm. The number of E. coli K12 removed per mass, volume, and surface area for all nanofiber mats at t = 120 min are provided in Table 2. The total removal capacity increased with increasing nanofiber mat diameter while the mass-, volume-, and area-normalized capacities fluctuate with no clear pattern, as expected.

A dynamic model was developed to describe the behavior seen in the experimental uptake data, under an assumption that transport processes external to the mat are controlling uptake. The data in Fig. 3D clearly show that the rate of bacteria removal is greater for physically larger mats, which indicates that bacterial transport within the mat is not the rate-limiting process. Furthermore, z-stack confocal micrographs indicated that bacteria were uniformly distributed throughout the void space of the nanofiber mats at all-time points during adsorption experiments, which further confirmed that transport inside the nanofiber mat was not the rate limiting factor. Under the further assumption that the concentration difference between the mat and the bulk solution is driving the transport, the following equation was derived,

where k_m is the first-order rate constant and the other symbols have been defined previously. Details of the derivation can be found in the ESI.†

The solid curves in Fig. 3D are the fits of the model to the experimental data. The variable k_m was the only fitted parameter in each case, as the values of all other variables are assumed known (including K_eq from the fit in Fig. 3C). The overall quality of the fits support the apparent first-order kinetics for bacterial removal by the nanofiber mat, consistent with Liu and Ford's report that bacteria may aggregate in a thin layer outside of a porous medium (50 to 500 μm diameter spherical particles) before diffusing in from there.²³

The values of the rate constant, k_m, as estimated from the dynamic model fits are shown as a function of nanofiber mat diameter in Fig. 3E. The geometry of the experiment is such that the mat is a two-dimensional target floating on the surface of the bacterial solution covering only a fraction of that surface. If the E. coli K12 removal were controlled by external diffusion or some other process that is limited by the rate at which the cells reach the surface of the mat, then one would expect k_m to scale with the surface area of the mat, i.e., with the square of the mat diameter. Fig. 3E displays that the dependence of k_m on the nanofiber mat diameter does indeed appear to have nonlinear, possibly a quadratic character. Overall the results confirm that some external process, which is dependent on the geometry of the mat within the test well, is controlling bacterial uptake in our experiments.

Nanofiber mat “sponges” uptake additional microorganisms

High surface-to-volume area nanofiber mats featuring 1 μm diameter fibers hold potential as sponges that can efficiently remove E. coli K12 from a contaminated area such as a wound. As a further proof-of-concept we challenged the nanofiber mat with an additional model Gram-negative microbe, P. aeruginosa PA01, as well as with the model Gram-positive microbe, S. aureus MW2, Fig. 4. At t = 120 min, the nanofiber mat collected statistically more S. aureus MW2 (6.28 × 10⁸ cells) and P. aeruginosa PA01 (5.96 × 10⁸ cells) than E. coli K12 (4.61 × 10⁸ cells). Since these microbes have different transport and adhesion mechanisms, we suggest that the ability of the nanofiber mats to remove microbes is independent of their adhesion mechanisms. E. coli have receptor specific binding through organelle, specifically type 1 frimbraie, auto transporter proteins, and aggregative frimbraie, that can permanently bind the bacteria to surfaces.⁴² S. aureus lack these extracellular organelle and instead rely on protein adhesions through microbial surface components recognizing adhesive matrix molecules (MSCRAMM).^43,44 P. aeruginosa PA01 contain two distinct lipopolysaccharide (LPS) O-polysaccharide species, A- and B-band LPS, which control the surface hydrophobicity and surface charge on the bacteria thus dictating its ability to adhere to a material.⁴⁵ All of the bacteria strains tested have different adhesion mechanisms and thus, we suggest that the nanofiber morphology and void spaces within the electrospun mat predominantly dictate the removal capacity.

Fig. 4 The total number of S. aureus MW2, P. aeruginosa PA01, and E. coli K12 collected as a function of time is displayed. An initial concentration of 1.52 × 10⁸ cells per mL was used and the diameter of the nanofiber mat (2.54 cm) was held constant. Error bars represent the standard deviation of six trials.

Conclusions

Overall, the high adsorption of multiple microorganisms achieved by the electrospun cellulose nanofiber mats demonstrates their potential application as sponges. On a per milligram basis, our cellulose nanofiber mat collected 21 and 220 times more E. coli K12 than the Fisherbrand control and Sartorius control, respectively. While all sorbents had the same surface chemistry, the nanofiber mat demonstrated the best removal efficiency due to its ideal diameter fibers and larger available pore structures. By pairing the experimental data with the quasi-equilibrium model and diffusion kinetics of E. coli K12 removal, we provide insight into the properties and parameters that result in high microorganism removal using nanofiber mats.

Acknowledgements

This work was kindly supported by the NSF-sponsored Institute for Cellular Engineering IGERT program (DGE-0654128), the James M. Douglas Career Development Faculty Award, and the UMass Center for Hierarchical Manufacturing (CHM), a NSF Nanoscale Science and Engineering Center (CMMI-1025020). Thanks to Dr Shelly Peyton, Lauren Barney, and Jesus Alvelo for help with nanofiber mat staining and confocal imaging. Thanks to Dr Wei Fan and Paul Dornath for help conducting nitrogen adsorption experiments. We acknowledge the use of facilities at the W. M. Keck Center for Electron Microscopy.

Notes and references

1. L. Persano, A. Camposeo, C. Tekmen and D. Pisignano, Macromol. Mater. Eng., 2013, 298, 504–520.
2. W. E. Teo and S. Ramakrishna, Nanotechnology, 2006, 17, R89–R106.
3. J. Doshi and D. H. Reneker, J. Electrost., 2012, 35, 1–10.
4. L. Ghasemi-Mobarakeh, D. Semnani and M. Morshed, J. Appl. Polym. Sci., 2007, 106, 2536–2542.
5. S. Ramakrishna, K. Fujihara, W.-E. Teo, T. Yong, Z. Ma and R. Ramaseshan, Mater. Today, 2006, 9, 40–50.
6. Y. Zhang, C. T. Lim, S. Ramakrishna and Z.-M. Huang, J. Mater. Sci.: Mater. Med., 2005, 16, 933–946.
7. J. D. Schiffman and M. Elimelech, ACS Appl. Mater. Interfaces, 2011, 3, 462–468.
8. K. A. Rieger and J. D. Schiffman, Carbohydr. Polym., 2014, 113, 561–568.
9. M. Abrigo, P. Kingshott and S. L. McArthur, ACS Appl. Mater. Interfaces, 2015, 7, 7644–7652.
10. J. X. Zhang, World J. Diabetes, 2011, 2, 105–107.
11. L. Hall-Stoodley, J. W. Costerton and P. Stoodley, Nat. Rev. Microbiol., 2004, 2, 95–108.
12. G. C. Gurtner, S. Werner, Y. Barrandon and M. T. Longaker, Nature, 2008, 453, 314–321.
13. K. A. Rieger, N. P. Birch and J. D. Schiffman, J. Mater. Chem. B, 2013, 1, 4531.
14. M. W. Frey, Polym. Rev., 2008, 48, 378–391.
15. A. W. Carpenter, C.-F. de Lannoy and M. R. Wiesner, Environ. Sci. Technol., 2015, 49, 5277–5287.
16. H. P. S. Abdul Khalil, A. H. Bhat and A. F. Ireana Yusra, Carbohydr. Polym., 2012, 87, 963–979.
17. D. Klemm, B. Heublein, H.-P. Fink and A. Bohn, Angew. Chem., Int. Ed. Engl., 2005, 44, 3358–3393.
18. H. P. S. Abdul Khalil, Y. Davoudpour, M. N. Islam, A. Mustapha, K. Sudesh, R. Dungani and M. Jawaid, Carbohydr. Polym., 2014, 99, 649–665.
19. M. M. Reddy, S. Vivekanandhan, M. Misra, S. K. Bhatia and A. K. Mohanty, Prog. Polym. Sci., 2013, 38, 1653–1689.
20. M. Jorfi and E. J. Foster, J. Appl. Polym. Sci., 2014, 132, 1–19.
21. N. Tufenkji, Adv. Water Resour., 2007, 30, 1455–1469.
22. S. M. Nielsen, I. Nesterov and A. A. Shapiro, Transp. Porous Media, 2014, 102, 227–259.
23. J. Liu and R. M. Ford, Environ. Sci. Technol., 2009, 43, 8874–8880.
24. H. Liu and Y.-L. Hsieh, J. Polym. Sci., Part B: Polym. Phys., 2002, 40, 2119–2129.
25. Z. Ma, M. Kotaki and S. Ramakrishna, J. Membr. Sci., 2005, 265, 115–123.
26. P. Dornath and W. Fan, Microporous Mesoporous Mater., 2014, 191, 10–17.
27. I. Wiegand, K. Hilpert and R. E. W. Hancock, Nat. Protoc., 2008, 3, 163–175.
28. D. M. Ruthven, Principles of Adsorption and Adsorption Processes, John Wiley & Sons, Inc., 1984.
29. J. Crank, The Mathematics of Diffusion, Oxford University Press, New York, 2nd edn, 1975.
30. P. Lu and Y.-L. Hsieh, ACS Appl. Mater. Interfaces, 2010, 2, 2413–2420.
31. H. L. Schreuder-Gibson, Q. Truong, J. E. Walker, J. R. Owens, J. D. Wander and W. E. Jones, MRS Bull., 2003, 28, 574–578.
32. P. Zahedi, I. Rezaeian, S.-O. Ranaei-Siadat, S.-H. Jafari and P. Supaphol, Polym. Adv. Technol., 2009, 21, 77–95.
33. S. R. Dods, O. Hardick, B. Stevens and D. G. Bracewell, J. Chromatogr. A, 2015, 1376, 74–83.
34. L. Zhang, T. Menkhaus and H. Fong, J. Membr. Sci., 2008, 319, 176–184.
35. D. Aussawasathien, C. Teerawattananon and A. Vongachariya, J. Membr. Sci., 2008, 315, 11–19.
36. S. Ramakrishna, R. Jose, P. S. Archana, A. S. Nair, R. Balamurugan, J. Venugopal and W. E. Teo, J. Mater. Sci., 2010, 45, 6283–6312.
37. H. Ma, C. Burger, B. S. Hsiao and B. Chu, J. Mater. Chem., 2011, 21, 7507.
38. H. Ma, B. S. Hsiao and B. Chu, J. Membr. Sci., 2014, 452, 446–452.
39. M. Abrigo, P. Kingshott and S. L. McArthur, Biointerphases, 2015, 10, 04A301.
40. J. Rivera-Utrilla, I. Bautista-Toledo, M. Ferro-Garcia and C. Moreno-Castilla, J. Chem. Technol. Biotechnol., 2001, 76, 1209–1215.
41. M. Katsikogianni and Y. F. Missirlis, Eur. Cells Mater., 2004, 8, 37–57.
42. W. M. Dunne, Clin. Microbiol. Rev., 2002, 15, 155–166.
43. T. J. Foster and M. Höök, Trends Microbiol., 1998, 6, 484–488.
44. S. R. Clarke and S. J. Foster, Adv. Microb. Physiol., 2006, 51, 187–224.
45. S. A. Makin and T. J. Beveridge, Microbiology, 1996, 142, 299–307.

---

Footnote

† Electronic supplementary information (ESI) available: Fig. S1 is a z-stack composite image used to determine the average thickness of the cellulose nanofiber mats. Fig. S2 is a calibration curve for converting microorganism plate readings to CFU mL⁻¹. Fig. S3 provides SEM micrographs of cellulose acetate and cellulose nanofiber mats, along with the fiber diameter distributions of the cellulose nanofiber mat and Fisherbrand control. Fig. S4 displays FTIR spectra and a summary table of the characteristic peaks of cellulose acetate and cellulose nanofiber mats and the nitrogen adsorption plot of cellulose nanofiber mats. Further details on the dynamic model are also provided. See DOI: 10.1039/c6ra01394e

---