Battery-free radio frequency wireless sensor for bacteria based on their degradation of gelatin-fatty acid composite films

(1)

Contents lists available at ScienceDirect

Electrochimica

Acta

journal homepage: www.elsevier.com/locate/electacta

Battery-free

radio

frequency

wireless

sensor

for

bacteria

based

on

their

degradation

of

gelatin-fatty

acid

composite

ﬁlms

Palraj

Kalimuthu

a, b, 1, ∗

_,

_Juan

_F.

_{Gonzalez-Martinez}

a, b

_,

_Dainius

_Jakubauskas

a, b

_,

Marité Cárdenas

a, b

_,

_Tautgirdas

_Ruzgas

a, b

_,

_Javier

_Sotres

a, b, ∗

a Department of Biomedical Science, Faculty of Health and Society, Malmö University, Malmö 20506, Sweden b Bioﬁlms-Research Center for Biointerfaces, Malmö University, Malmö 20506, Sweden

a

r

t

i

c

l

e

i

n

f

o

Article history:

Received 10 January 2021 Revised 1 March 2021 Accepted 26 March 2021 Available online 1 April 2021 Keywords:

Bacteria detection Composite ﬁlm

Radio frequency identiﬁcation Passive wireless sensor Scanning electron microscopy

a

b

s

t

r

a

c

t

Continuousandautomatedbacteriadetectionispivotalforamyriadofbiomedical,foodsafetyand envi-ronmentalapplications.Thisworkpresentsthefabricationofaprototypeofapassive(battery-free)radio frequency sensorforwirelessdetectionofbacteria. Thesensing mechanismis basedonthe bacterial-induced(proteasesandpeptidases)degradationofglutaraldehyde(GTA)cross-linkedgelatin-caprylicacid (CA)compositefilm.Proteolyticdegradationofthefilmresultedinadecreaseofitsresistivity,a quan-titythatcouldbewirelesslymonitoredbycouplingthefilmtoaradio-frequencyantenna(an inductor-capacitorresonator)andmonitoringthefrequencyforwhichthetransferredpowerbetweenthisantenna and anotherantenna connectedto aVector NetworkAnalyzer(VNA) wasmaximized. We experimen-tallyprovedthisconceptbymonitoringE.colibacteriainaqueousmediumanddetectedat18.0± 2.8h, 23.5± 0.7h,27.0± 2.8h,40.5 ± 3.5h,45.5± 0.7hfortheinitialE.coliconcentrationof3.2× 108_,

6.8× 107_,_2.3_{× 10}6_,_4.3_{× 10}5_,_and_3.6_{× 10}4_CFU/mL,_{respectively.}_Further,_the_E.coli_induced

degrada-tionofthecompositeﬁlmwasinvestigatedbyevaluatingthethicknessoftheﬁlmbyopticalmicroscopy aswellasmorphologybyscanningelectronmicroscopytechniques.

1. Introduction

Bacterial outbreaks are the origin of major diseases and corresponding mortality that arise throughout the world every year [1-3]. Therefore, the detection and quantification of pathogenic bacteria have become key points in various sectors, including biomedical, food safety and environmental applications [4–6]. Hitherto, bacteria detection relies on several techniques such as standard plate counting [7], enzyme-linked immunosorbent as- say [8], polymerase chain reaction (PCR) [9], adenosine triphos- phate (ATP) estimations [10], direct epifluorescent filter technique [11], magnetic [12], electrochemical [13] and immuno-biosensors [14]. However, most of these techniques require numerous time- consuming handling steps, periodic manual sampling, expensive instrumentations and trained personnel for operation. Moreover,

∗_{Corresponding authors at: Department of Biomedical Science, Faculty of Health} and Society, Malmö University, Malmö 20506, Sweden.

E-mail addresses: p.kalimuthu@uq.edu.au (P. Kalimuthu), javier.sotres@mau.se (J. Sotres).

1 Present address: School of Chemistry and Molecular Biosciences, The University of Queensland, St Lucia-4072, Brisbane, Australia.

excluding those based on electrochemical sensing, all the mentioned techniques lack the ability to detect bacteria in real-time or in-ﬁeld as they need to be performed in laboratory environments. A fast, low-cost and easy to use a detection system for bacteria in real-time would help to reduce the time limitation of these methods, allowing to work directly in-ﬁeld without the necessity to bring samples to laboratories.

In this regard, implementing wireless communication has enor- mous potential. Wireless communication is not only beneficial in applications where remote monitoring is of interest but also in those that benefit from minimal human intervention, e.g., monitoring infection with sensors embedded in care products like wound dressings and diapers. Almost any sensor can be rendered wireless by means of appropriate circuitry, (e.g., Wi-Fi, Bluetooth, etc.). However, most of these approaches require batteries. This increases costs and limits implementation in many areas. In this work, we make use instead of the passive (battery-less) chip-less Radio Fre- quency Identification (RFID) technology, where sensors do not require batteries as they are powered by the reader itself [15–18]. While different im plementations of this technology are available, we make use of that where the antennas of reader and sensor are magnetically coupled when close to each other, and the sen-

https://doi.org/10.1016/j.electacta.2021.138275

(2)

insitu growth of three different bacteria, namely Bacillussubtilis,E. coli JM109, and Pseudomonasputida in food samples [24]. In addition, Potyrailo and co-workers developed a passive RFID sensor for monitoring bacterial growth in different food samples using E.coli as a model system [31]. Moreover, several nanomaterials have been incorporated in the RFID sensor/transponders to improve sensitivity towards bacteria. For instance, dextrin capped gold nanoparti- cles (d-AuNP) were used as markers for E.coli C30 0 0 in milk samples, allowing their wireless detection as the presence of nanopar- ticles could shift the resonance frequency of different transponder designs [29, 30]. Mannoor etal. used an LC resonator/transponder integrated onto graphene printed on silk ﬁbroin and showed the potential of this design for detection of bacteria in the oral cavity when integrated on tooth surfaces [27].

However, most of these RFID-based sensors have inherent drawbacks for monitoring bacteria in aqueous media. Principally, the sensing and antenna sections of the transponder were both exposed to the analyzed liquids, which affects the read signal in different ways. First, not only the presence but also the volume of liquid surrounding the transponder antenna drastically reduces the reading distance [33]. Second, changes in e.g., saline content, temperature and pH of the transponder antenna surrounding media would also inﬂuence the communication with the reader [22– 24,34]. These drawbacks can be overcome by means of a transponder design where the sensor section is connected but well differentiated from the antenna section in the transponder and where only the sensor section is exposed to the investigated aqueous medium [3].

Recently, we reported a radio frequency allied wireless sensor for proteases based on this implementation [33]. The wireless transponder/sensor comprised of two main components: an LC (inductor-capacitor) resonator, i.e., an antenna, connected to two electrodes bridged by a cost-effective chemically cross-linked gelatin-fatty acid composite film. Proteolytic digestion of the composite led to a change in its resistivity and, subsequently, to a decrease of the (characteristic) frequency for which the power transmitted by the reader, a Vector Network Analyzer (VNA) equipped with an antenna, to the transponder was maximized. This characteristic frequency could then be used as a quantity for wirelessly monitoring the presence of proteases. Interestingly, the developed composite film showed high stability in aqueous media (phosphate-buffered saline (PBS) and lysogeny broth (LB) medium) as well as sensitivity towards various proteases. The goal of this work was to investigate the applicability of this sensor concept to monitor the presence of bacteria in aqueous solutions. Our hypothesis was that bacteria, by means of their secreted proteases/peptidases, could as well digest the cross-linked gelatin- fatty acid composite film. The possibility to monitor the presence of bacteria in aqueous environments using a low-cost wireless technology would be of interest for several applications. An example is that of wearable sensors for early detection of infection

and, thus, their early detection would beneﬁt from a sensor based on the degradation of a gelatin-based material. Here, in order to prove the practical applicability of the developed sensor concept, we have employed it to wirelessly detect and quantify E.coli bacteria in aqueous media. The choice of E.coli as a model sample for our study was based not only on it being one of the most ubiqui- tous pathogens but also, from a speciﬁc application point of view, on E.coli being the most common cause of UTI [40].

2. Materialandmethods

2.1. Chemicals

Gelatin (type A, Prod. No. G2500),), glycerol (Prod. No. G5516), caprylic acid (Prod. No. O3907), aqueous solutions of glutaralde- hyde (50 %, w/v) (Prod. No. G7651), tryptone (Prod. No. 16922), yeast (Prod. No. 51475), and sodium chloride (Prod. No. S5886) were purchased from Sigma Aldrich (St. Louis, MO) and used as received. Unless otherwise speciﬁed, all other reagents used were of analytical grade purity and used as supplied. The E.coli utilized in this work was TOP10 chemically competent bacterial strain (Prod. No. C404003, Invitrogen). All the experiments were performed at room temperature (~25 °C). All solutions were prepared with ultra- high quality water (UHQ, resistivity 18.2 M

•cm) processed in El- gastat UHQ II apparatus (Elga Ltd, High Wycombe, Bucks, England). Phosphate-buffered saline (PBS) solution was prepared from the tablets (Prod. No. P4417, Sigma Aldrich) resulting in 137 mM NaCl, 2.7 mM KCl and 10 mM PBS (pH 7.4 at 25 °C).

2.2. Bacteriacultivation,celllysatepreparationandmeasurement Sterile lysogeny broth (LB) medium without antibiotics was used to cultivate bacteria. The LB medium and glassware were au- toclaved at 121 °C for 15 min prior to experiments. The E. coli glycerol stocks of a single bacterial colony were ﬁrst resuspended into 7 mL of LB and incubated at 37 °C (shaking at 200 rpm) overnight, and then subsequently recultivated in 200 mL of sterile LB (same growth conditions) for 5 h. The growth rate of bacteria was monitored at regular time intervals with a UV spectropho- tometer (Perkin Elmer Lambda 35, Perkin Elmer, Waltham, Mas- sachusetts) using the optical density measured at 600 nm wave- length (OD 600). The growing bacteria were removed from the in-

cubator when OD 600 0.6 (4.8 × 108CFU/ml) which is the mid-log

phase in the growth curve [41]and then diluted in the LB medium (preserving the pH) to obtain the desired concentration. The bacterial cultures were stored at 4 °C when not in use to prevent growth and ensure an accurate dilution factor. Samples were brought to room temperature prior to the use through dilution in LB medium. It was found the E.coli cells continuously grew up to OD 600 1.7

(~1.36 × 10 9 _{CFU/ml) and then reached stationary phase (Support-}

(3)

Fig. 1. ( a) Setup used for bacteria wireless detection. ( b) Close view of the reader and transponder antennas.

of stock bacteria (4.8 _{× 10}8 _{CFU/ml) was transferred into 2.5 mL}

of LB medium and used for all wireless and impedance measurements. Speciﬁcally, experiments were carried out using bacteria concentrations range of 3.6 × 104 _{– 3.2}_{× 10}8 _CFU/ml_(OD

600

: 0.0 0 0 045 to 0.4). We also performed experiments with bacteria suspended in phosphate buffer saline (PBS). In this case, bacteria from LB stock solutions were sedimented by centrifugation (40 0 0 rpm for 10 min) and washed twice with PBS solution to remove any nutrients from the LB medium present in the sediment. The samples were then further diluted in PBS buffer down to the desired concentration. Low E. coli concentration values (OD 600 <

0.01) were conﬁrmed by pour plate counting method as reported elsewhere [42].

To obtain E. coli cell lysate, 200 mL of bacterial culture (OD 600 = 0.6) was centrifuged (50 0 0 g, 15 min in 10 °C, Sorvall

RC5B Plus, USA), the cell pellet was washed and resuspended in 30 mL ice-cold PBS buffer. Resuspended E. coli cells were pulse- ultrasonicated on ice (Branson 250 Digital Soniﬁer, Branson Ultra- sonics Corp., USA), until the liquid became transparent. The liquid was subsequently centrifuged (25,0 0 0 g, 20 min in 10 °C, Sorvall RC5B Plus, USA), collected supernatant (cleared cell lysate) was ﬁl- tered through sterile 0.22

μ

m ﬁlters and stored for further use in 4 ◦C.

2.3. Preparationandﬁlmfabricationofglutaraldehyde(GTA) cross-linkedgelatin-fattyacidcomposite

The preparation of the composite material and subsequent film fabrication on screen-printed gold electrodes (SPGE, ref: C223AT, DropSens, Spain) is described elsewhere [33]. Briefly, pristine gelatin solutions were prepared by dissolving (10% w/v) of gelatin powder in UHQ water under heating (50 °C) and stirring conditions. After 30 min, glycerol was added into the hot gelatin solution with a final concentration of 1 % w/v and kept under heating and stirring conditions for additional 15 min. After achieving a homogeneous gelatin-glycerol mixture, the fatty acid, caprylic acid (CA) was slowly added into the complex mixture to reach a final concentration of 64% w/v. Stirring and heating continued for another 30 min until a homogeneous milky white solution was obtained.

The resulting hot gelatin-glycerol-CA composite solution was cast on SPGEs used for impedance and wireless measurements. Prior to the ﬁlm casting, the SPGEs were cleaned in 0.1 M H 2SO 4

by cycling the electrode potential from −0.1 to +1 V as described [43]. Then, 60 μL of the hot composite solution was drop-coated on the cleaned SPGE. After coating, the ﬁlm was cooled at room temperature for 10 min and then incubated in a 0.5 wt % 5 ml glu-

taraldehyde (GTA) in water for 1h. After this step, the films were rinsed thoroughly with UHQ water to remove any non-crosslinked GTA. Finally, the films were cured at room humidity and temperature for 12 h. The thickness and morphology of these films before and after exposure to E.coli was determined by differential-contrast optical microscopy (Nikon Optiphot, Japan) and scanning electron microscopy (see below).

2.4. Vectornetworkanalyzermeasurements

A commercial DG8-SAQ Vector Network Analyzer (VNA) (SDR- Kits, Melksham, UK) was used both to characterize the impedance of the composite ﬁlm during bacterial degradation and as a reader for wireless bacterial monitoring. Both types of experiments were performed with SPGEs. These are comprised of three electrodes: gold working and counter electrodes and a silver reference electrode. For directly measuring the impedance of the composite ﬁlm, the working and counter electrodes of the SPGEs were connected to a 50

coaxial cable (matching the characteristic impedance of the VNA), which was connected through the other end to the transmitter port (TX) of the VNA. The actual experimental setup used in the present study is shown in Fig.1. For wireless measurements, the composite coated SPGE was connected to a rectangu- lar RF antenna (6.8 x 5.4 cm, 2 loops) through the working and counter electrodes with a homemade 3D printed holder ( Fig.1a) whereas the silver reference electrode of the SPEG was unattended. As a reader, we used a homemade copper circular RF antenna (diameter 5.5 cm, 5 loops) connected to the TX port of a VNA. For all reported wireless measurements, reader and sensor antenna were kept at a ﬁxed distance of 1.2 cm.

During the wireless measurements, the magnitude of the reﬂec- tion scattering parameter

|

S11

|

=

1 − P_T

Pmax (where P Tis the trans-

mitted power and P max is the maximum achievable transmitted

power) was monitored continuously in the 3-30 MHz range at a sweep rate of 135 kHz/s.

Both direct and wireless VNA-based measurements were performed by ﬁrst exposing the composite-coated SPGEs to sterilized LB medium or PBS for at least 24 h in order to ensure stability. The sterilization was performed by autoclave the medium for 20 min at 120 °C. Then the freshly prepared composite ﬁlm coated electrodes were transferred to the sterilized LB or PBS solution under UV light and then tightly closed the container. After ascertaining the stability, the coated SPGEs were exposed to bacteria-containing LB medium or PBS solution (5 mL), and the corresponding VNA signal monitored continuously. Further, it was noted that the freshly prepared LB medium exhibited a pH value of 7.2, and it was gradu-

(4)

ally decreased when the E.coli growth increases. On the other hand, no signiﬁcant pH shift observed in E.coli containing PBS solution. 2.5. Scanningelectronmicroscopy

An environmental scanning electron microscope (SEM) (EVO LS10, Zeiss, Germany) was used to visualize the composite ﬁlms before and after bacterial degradation. For this, the SEM was oper- ated in the variable pressure mode (10 Pa, EHT voltage 25 kV) and the images acquired using the backscattered detector.

3. Resultsanddiscussion

3.1. Sensorconcept

The concept of the developed wireless bacteria sensor was detailed in [33]. Briefly, a reader (a Vector Network Analyzer equipped with an antenna) is used to interrogate a transponder with well-differentiated antenna and sensor sections, only the lat- ter being exposed to the investigated aqueous media. A magnetic field is generated when the current flows through the reader antenna. When both reader and transponder antennas come into proximity, this magnetic field induces a current on the transponder. The equivalent circuit for this concept is shown in Fig.2.

The power transmitted from the reader to transponder is maximized at a certain (characteristic) current frequency for which the imaginary part of the equivalent impedance for the whole setup, Z eq(Eq. (1)), is zero [44].

Zeq=Zreader_antenna+

ω

2_M2

Ztransponder

(1) where Z reader-antenna is the impedance for the reader antenna,

Z _transponder is the overall impedance of the transponder i.e., the composite-bridged electrodes (sensor component) coupled to the circular RF antenna, and M is the mutual inductance between reader and transponder antennas. The analytical expression that relates the characteristic frequency, i.e., that for which the imaginary part of Z eq becomes zero, to the resistance and capacitance

of the sensor component, R ts and C ts, is fairly complex. However,

as discussed in [33], provided that the sensor component can be modeled as a resistor (R ts) and a capacitor (C ts) connected in par-

allel, it is reasonable to assume that the transmitted power will be maximized for a frequency close to that for which the imaginary part of the equivalent impedance of the transponder is zero. Speciﬁcally, this condition is fulﬁlled for a frequency,

ω

0, provided

by [33]:

ω

0=

1 Lta

(

Cta+Cts

)

− 1 [

(

Cta+Cts

)

Rts]2 (2)

where L ta and C ta are the inductance and capacitance of the

transponder antenna, respectively. From Eq. (2), it follows that a decrease in the resistance of the sensor section of the transponder, R ts, would lower the frequency for which the transmitted power is

maximized. The proposed bacteria sensor concept is based on this dependence.

3.2. Impedancemeasurementsofthecompositeﬁlmduringits bacterialdegradation

The present wireless sensor concept relies on the resistivity changes of the composite used to coat SPGEs resulting from its degradation by the proteases/peptidases secreted from the E.coli bacteria present in the tested media. We monitored, as detailed in the material and methods section, the impedance in the RF spectrum of the composite-bridged electrodes while exposed to an E. coli solution. Experimental data for the real and imaginary components of the impedance of this system during exposure to the mid-log growth phase of E.coli cells (OD 600 = 0.67, equivalent to

5.36 × 108 _{CFU/mL) in LB medium are provided in Fig. S2a and}

S2b, respectively (Supporting Information).

Assuming that the composite-bridged electrodes can be modeled by a capacitor, C, and resistor, R, in parallel, the relationship between the equivalent impedance of the system, Zequivalent, and the current frequency, f, is provided by:

Zequivalent= R 1+j2

π

fCR= R 1+

(

2

π

fCR

)

2− j 2

π

fCR2 1+

(

2

π

fCR

)

2 (3) As shown in Fig. S2a and S2b, Eq. (3) provided a reasonable ﬁt (dashed lines) for the impedance of the composite when exposed to bacteria (solid lines). This allowed quantifying the capacitance and resistance (Fig. S2c and S2d, respectively, Supporting In- formation) over time while exposed to the bacterial solution. It can be observed that the capacitance remained fairly stable during the process (values stayed within a ~2 pF window during the experiment). However, the resistance decreased continuously during exposure of the composite to bacteria from a value of ~650

to ~150

(a value similar to that measured for non-coated electrodes, data not shown). This suggests that the composite film on SPGE was effectively degraded by E.coli. It is well known that E.coli bacteria secrete ~70 proteases/peptidases, including serine, cysteine, as- partic, and metalloproteases [45]. These enzymes fulfil numerous vital functions in cells, including digestive, protective, and regula- tory metabolic processes [46]. During extracellular digestion, peptide bonds in the protein matrix are hydrolyzed, and consequently, the released free amino acids are imported into the bacterial cell and metabolized. This digestion process is utilized in the present study to detect the E.coli bacteria with the aid of biodegradable GTA cross-linked gelatin-CA composite film coated on SPGE. As a result of composite film digestion, in the wireless setup resistivity

(5)

Fig. 3. A schematic illustration for the wireless sensor setup and E.coli detection process.

changes will lead (according to Eq.(2)) to a decrease of the characteristic frequency that will allow to monitor wirelessly the presence of bacteria in the tested solution. The fabricated wireless sensor setup and E.coli detection process are schematically illustrated in Fig.3.

The antimicrobial properties of the fatty acid, CA, incorporated within the gelatin ﬁlm to control its permeability to water, need to be discussed. CA is less sensitive to the gram-negative bacteria due to the difference in the cell wall structure that composed of an additional plasma membrane while compared to the gram-positive bacteria [47]. In addition, the pK a value of CA is 4.9 [48] and,

therefore, it shows bactericidal effect in acidic pH ( < 4.9) and no significant effect at neutral pH [49]. Hence, it is envisaged that CA does not show significant antibacterial effect towards the gram- negative E.coli bacteria at the neutral pH conditions employed in the present study. This is supported by the impedance measurements, which indicate that the resistivity change occurred because of bacterial-induced degradation of the CA incorporated composite film. Moreover, GTA specifically cross-links the free amino groups present in the gelatin polymer network through Schiff base and Michael-type reactions and, thus, the peptide bonds are expected to be unaffected and freely available for proteolytic degradation through exothermic hydrolysis reactions [50, 51].

3.3. Modificationofthethicknessofthecompositefilmsbybacteria The thickness of the composite film on SPGEs was evaluated by optical microscopy before and after degradation by E.coli ( Fig.4). Before conducting the measurement, the freshly prepared composite films were incubated in sterile LB medium for 24 h to attain complete water saturation, and the thickness of these films was estimated to be ~350 μm ( Fig.4a). In contrast, as shown in Fig.4b the thickness of the film was significantly decreased to ~75–100 μm upon exposure to E.coli bacteria as a result of enzymatic degradation. The image is shown in Fig.4b is representative of films for which a stable and constant resistance value was obtained after exposure to 5.36 × 108 _CFU/mL_E.coli_{for 28 h in the impedance}

measurements (Fig. S2). The composite ﬁlm was degraded entirely when exposed to bacteria solution for another few hours. Thus,

Fig. 4. Optical microscopy images of composite ﬁlm modiﬁed SPGEs (a) before and (b) after exposure to E. coli (5.36 × 10 8 CFU/mL) for 28 h.

(6)

Fig. 5. SEM images of SPEGs coated with GTA cross-linked gelatin-CA composites

a) before and b) after exposure for 24 h to E. coli (5.2 × 10 8 CFU/mL) in PBS.

optical microscopy conﬁrmed the susceptibility of the developed composite towards bacterial induce degradation.

3.4. Modiﬁcationofthemorphologyofthecompositeﬁlmsby bacteria

Scanning electron microscopy (SEM) was used to monitor changes in the topology of the GTA cross-linked gelatin-CA composite ﬁlms on SPGEs when exposed to bacterial solutions. SEM images for such a ﬁlm before and after exposure for 24 h to E.coli containing PBS solution (5.2 × 108 _{CFU/mL) are shown in}_Fig.₅_a

and b, respectively. It can be observed that initially the films exhibited a smooth, compact and defect-free surface. However, pits and grooves with diameter values between 10 0 and 40 0 μm were visible on the composite film after being exposed to the bacteria solution. The hollow structure of the films, as inferred from SEM, combined with the overall decrease in thickness, as inferred from optical microscopy, after exposure to E.coli confirmed that this bacteria catabolized the developed composite film with their enzymes and subsequently decreased its resistivity i.e., the proposed bacteria detection mechanism on which the proposed sensing concept relies.

3.5. Wirelessbacteriasensinginaqueousmedia

The bacterial degradation of the composite-coated electrodes was wirelessly monitored by means of the setup detailed in Section 2.4. Speciﬁcally, we monitored the magnitude of the for- ward scattering parameter, |S 11|, between the transponder and

Fig. 6. (a) Evolution of | S 11| overtime for a representative experiment where one of the developed bacteria sensors was exposed to E. coli cells (OD 600 = 0.4, equivalent to 3.2 × 10 8 CFU/mL) in LB medium. ( b) Evolution of the characteristic frequency over time measured for individual sensor electrode exposed to three different concentrations of E. coli in LB medium.

reader antennas in the 3–30 MHz range and characterized this spectrum in terms of the characteristic frequency for which |S 11|

attained a minimum value, i.e. the transmitted power was maximized.

At first, we monitored the characteristic frequency with the sensor section of the transponder immersed in bacteria-free aqueous media (we performed experiments in both sterile LB medium and PBS, characteristic frequency vs time plots in these media are shown in Figs.6, 7and8). The sensor modified with freshly prepared composite films exhibited a characteristic frequency slightly higher than 12.0 MHz both in PBS buffer and LB medium. This value shifted marginally to lower values during the first ~ 2 h of in- cubation, but always by a shift value below 0.5 MHz. This process suggests that some absorption of water molecules by the gelatin polymer took place initially. Indeed, in a recent work [33] we showed that the characteristic frequency remained fairly stable for 41 days (~10 3 _h)_in_aqueous_solution._{Nevertheless,}_for_all_our

experiments, the composite-coated electrodes were immersed in bacteria-free aqueous media (sterile LB medium or PBS) at least 24 h prior to exposure to bacteria-containing media. Below, we re-

(7)

Fig. 7. Evolution of the characteristic frequency over time measured for composite ﬁlm modiﬁed sensors exposed to an E.coli concentration of 6.4 × 10 8 CFU/mL in LB medium and in PBS solution.

Fig. 8. Evolution of the characteristic frequency over time measured for compos- ite ﬁlm modiﬁed sensors exposed to E.coli cells lysate and non-lysed E.coli cells (5.2 × 10 8 CFU/mL) in PBS solution.

port and discuss our wireless measurements for E.coli detection in both LB medium and PBS buffer.

Fig.6a illustrates the wireless monitoring over time of the for- ward scattering parameter,

|

S11

|

, for a composite ﬁlm incorporated

transponder when exposed to 3.2 × 10 8_{CFU/mL (0.4 OD}

600in the

exponential growth phase) of E. coli in LB medium at 25 °C. The initial characteristic frequency, 11.8 MHz, continuously decreased until a stable value, ~6.5 MHz, was reached. This proves that the proposed sensor concept can be used to wireless monitor the presence of bacteria by means of their degradation of the composite ﬁlm incorporated in the sensor section of the transponder.

The time evolution of the characteristic frequency for three different experiments where transponders were exposed to 3.6 × 104_,

2.3 _{× 10}6 _{and 3.2}_{× 10}8 _CFU/mL_E.coli_{in LB medium (25 °C) is}

shown in Fig.6b along with a representative control experiment where the sensor was exposed to bacteria-free LB medium. Like in the experiment shown in Fig.6a, the starting characteristic frequency exhibited a similar value of ~11.8 MHz for all experiments and ended up in value within the 6.5–7 MHz range. In all ex-

periments, characteristic frequency vs. time plots for the experiments where the sensors were exposed to bacteria exhibited three regimes: (i) an initial one characterized by a low negative gradient followed by (ii) a high negative gradient regime and (iii) a final stationary regime where the characteristic frequency remained stable. The existence of the initial low negative gradient regime suggests that the initial degradation process does not respond to the frequency change due to the higher thickness of the film, as the similar phenomenon observed in our previous work with different proteases [33]. It can be noted that the time period over which this regime takes place is inversely dependent on bacteria concentration. The high dependence of the characteristic frequency with time characteristic of the second regime indicates a rapid decrease of the composite resistivity, which can be reasonable to associate with significant degradation. In contrast with the first regime, the time period over which this second regime takes place has a mi- nor dependence on the initial bacteria concentration. This regime finished when the E.coli cells secreted enzymes digested the composite film down to an extent where conducting ions were able to freely diffuse between the electrodes of the sensor section of the transponder, and a new plateau is found. The characteristic frequency measured in this regime (~6.5–7 MHz) matched that measured in LB medium for transponders where the sensor section was formed by bare non-coated SPGEs.

It is well known that the LB medium is composed of high nutrients such as tryptone and yeast extract, which are expected to facilitate bacterial growth. Therefore, to investigate the impacts of these nutrients towards the performance of the sensor, the experiments were also performed in PBS solutions. As mentioned in Section2.2, the cultivated E.coli cells were completely washed with PBS solution to remove the LB medium ingredients present in the bacterial sediment. Fig.7shows a comparison between representative experiments where the sensor was exposed to a 6.4 _{× 10}8

CFU/mL concentration of E.coli in both LB medium and PBS buffer. It can be seen that a similar response was obtained in both media. According to the bacteria detection criterion mentioned above (i.e., 0.8 MHz shift of f0) the detection time was found to be ~ 15

and 17 h for LB medium and PBS solution, respectively. The degradation process of the composite film was marginally faster in LB medium when compared to the PBS solution. There are two spec- ulations that bring up these results. First, the nutrients present in the LB medium could lead to faster growth of bacteria than in PBS solution at room temperature and, subsequently, in faster degradation of the composite film. It might also be that bacteria adjusted faster in LB medium as they were grown in the same medium, whereas adapting to PBS buffer might take additional time before starting the degradation process. In Fig.7it can also be observed that the characteristic frequency measured at the beginning of an experiment was slightly lower (~ 0.5 MHz) when investigating media containing bacteria with respect to bacteria-free media (control experiments). This was expected as the presence of bacteria (and that of their secreted and peptidases) is known to affect the di- electric constant of the media [52]. This will in turn, affect the capacitance values of our sensor and, therefore, the measured characteristic frequency. Fig. 7 also shows that the characteristic frequency measured in PBS buffer after complete degradation of the composite was ~ 2 MHz higher than for experiments performed in LB medium. This reflects a difference between the resistivity of LB medium and that of PBS buffer. Nevertheless, these differences do not mask the drastic change in characteristic frequency observed upon degradation of the composite by bacteria. Overall, these results confirmed that the developed sensor could be used to monitor E.coli bacteria regardless of the nutrient content of the media.

Further, as aforementioned, we have attributed the observed frequency changes to the degradation of the composite ﬁlm by E.coli secreted enzymes. To conﬁrm our conjecture, we tested the

(8)

Fig. 9. Bacteria detection time, corresponding to that for which the characteristic frequency shifted by 0.8 MHz, for different concentrations of E. coli in LB medium. Each data point and corresponding error bar corresponds to the mean and standard deviation value calculated from three different experiments performed at room temperature.

developed sensor in bacterial cell lysate in PBS solution and compared the result with that obtained on non-lysed bacteria cells in PBS solution under identical experimental conditions at room temperature. In both experiments, the initial bacteria concentration was 5.2 × 108 _CFU/mL._Fig. ₈ _{portrays a}_detection_time_of

~17 h (based on the detection criterion of 0.8 MHz) for non-lysed E.coli cells in PBS solution. In contrast, the E.coli cells lysate in PBS solution were not detected until ~40 h. Recently, we reported the degradation of the developed composite film by various proteases and found that the degradation rate varied depends upon protease concentrations [33]. Non-lysed bacteria degraded the composite film faster than bacterial cells lysate. This indicates that E. coli lysate contains a constant concentration of E.coli secreted enzymes [53]. In this environment, it is expected that the composite film will be degraded at a slower rate than in the non-lysed cells solution, where proteases and peptidases will be continuously secreted as supported by our data.

Moreover, we have also explored the quantiﬁcation possibilities of the proposed bacterial sensor concept based on the experiments performed in LB medium. As seen in Fig. 6b, for all experiments, the characteristic frequency versus time plot entered the second regime for a frequency shift of ~0.8 MHz. In contrast, there was no such a frequency decrease observed neither in LB medium nor in PBS buffer, even when immersed for long periods of time (~10 3

h) [33]. Subsequently, we used this characteristic frequency shift, i.e., 0.8 MHz, as a bacteria detection criterion. According to this criterion, E.coli bacteria detection times exhibited an exponential dependence with the initial bulk bacteria concentration, as shown in Fig.9. Speciﬁcally, E.coli bacteria were detected at 18.0 _{± 2.8,} 23.5 ± 0.7, 27.0 ± 2.8, 40.5± 3.5, and 45.5± 0.7 h for the initial concentration of 3.2 × 108_{, 6.8}_{× 10}7_{, 2.3}_{× 10}6_{, 4.3}_{× 10}5 _and

3.6 _{× 10}4 _{CFU/mL, respectively.}_{Thus, the}_{proposed sensor}_does

not only allow monitoring the presence of E.coli cells in a tested aqueous solution but also quantifying their concentration range of 3.2 _{× 10}8_{to 2.3}_{× 10}4_CFU/mL.

On average, the initial characteristic frequency measured in our experiments was ~11.8 MHz, and after complete degradation of the composite, this quantity decreased to a value of ~6.5 MHz, which corresponds to a ~45% change. Characteristic frequency changes around 1% are usually considered as an acceptable wireless sen-

tations, the transponder antenna and, therefore, its coupling with the reader antenna will be affected by parameters like the sam- pled volume as well as changes in its physico-chemical properties of e.g., ionic strength, as shown by Ma and co-workers [22]. They found that the monitored inductance of the sensor was extremely interfered by physicochemical parameters such as sample volume and buffer ion concentration. The implementation proposed in this work, where the transponder is composed of well-separated sensor and antenna sections, only the former exposed to the tested media, avoids this drawback. Thus, changes in the monitored characteristic frequency can be unequivocally associated with the degradation of the composite ﬁlm by E.coli bacteria.

It can be noted that the proposed sensor concept, while still in a prototype stage, is a cost-effective strategy as it basically relies on bridging two electrodes with the (low-cost) developed gelatin-based composite when compared to the expensive lithog- raphy techniques used in other reported sensors [24, 29, 30, 32, 34]. The thickness of the composite film on SPGE is directly correlat- ing with the detection time. So far, our attempts to lower the film thickness led to not enough reproducibility. Currently, to enhance the sensor performances related to the detection time and stability, we are investigating different methods for depositing the composite film e.g., spin coating, as well as the incorporation of other components. Finally, it is also worth to mention that the degradation of this composite could be used as well in any other type of bacterial sensor strategy independently of the reading mechanism.

4. Conclusions

Here, we proposed a novel concept for wireless monitoring of bacteria in aqueous media and validated it by monitoring the presence of E.coli cells in both LB medium and PBS buffer. The sensing mechanism is based on the degradation of a GTA cross-linked gelatin-CA composite film covering two electrodes exposed to the tested media. The composite film was degraded in the presence of E.coli cells as confirmed by both SEM and optical microscopy images. This resulted in a decrease of its resistivity as indicated by impedance measurements. In the proposed sensor concept, the film-coated electrodes were connected to an antenna (an LC resonator), only the former being exposed to the tested aqueous sample. Keeping the transponder antenna outside from the tested sample avoids the drawbacks associated with the use of RF communication through aqueous media. When the transponder with the sensor section exposed to bacterial solutions was monitored wirelessly by means of a VNA equipped with an RF antenna, the frequency for which maximum power was transferred from reader to transponder antenna shifted to lower values. We did not only prove in this way that the proposed sensor could be used to monitor E. coli bacteria in aqueous solutions (both PBS buffer and LB medium were investigated). We also showed that the detection time followed a relationship with the bacteria concentrations.

(9)

Thus, the proposed sensor concept offers the avenue for bacteria quantiﬁcation. Further, it was found that the degradation of the composite ﬁlm in bacterial cells lysate was slower than in non- lysed bacteria solutions. This indicated that the continuous pro- duction of proteases and peptidases by the E.coli played a critical role in the degradation of the composite leading to earlier detection. The advantages of the proposed sensor concept include real- time, accurate and high sensitive (3.6 × 10 4 _{CFU/ml) bacteria de-}

tection. Moreover, it also has possibilities for low-cost scalable pro- duction. A variety of applications (biomedical, food safety and environmental) would beneﬁt from these characteristics, which out- perform those from many of the currently available techniques for bacteria detection. Future work will focus on improving the detection time by optimizing the thickness and composition of the composite ﬁlm as well as on testing the sensor in more biomedical rel- evant samples like wound exudates and urine.

Creditauthorstatement

Palraj Kalimuthu: Conceptualization, Investigation; Formal analysis; Data curation, Methodology, Roles/Writing - original draft.

JuanF.Gonzalez-Martinez: Methodology, Writing - review & editing. DainiusJakubauskas: Methodology, Writing - review & editing. Marité Cárdenas: Resources; Data curation, Writing - review & editing. Tautgirdas Ruzgas: Resources; Validation; Writing - review & editing. JavierStores: Methodology, Conceptualization, Su- pervision; Validation; Data curation, Funding acquisition; Project administration; Investigation; Writing - review & editing.

Authorcontributions

The manuscript was written through the contributions of all authors.

DeclarationofCompetingInterest

The authors declare that they have no known competing ﬁnan- cial interests or personal relationships that could have appeared to inﬂuence the work reported in this paper.

Acknowledgments

The Knowledge Foundation (Grants No. 20150207 and 20190010), the Swedish Research Council (Grant No. 2 018-04320), the Mats Paulsson’s foundation for research, innovation and development of society, Malmö University and the Gustav Th. Ohlsson Foundation are gratefully acknowledged for ﬁnancial support. M.C. and D.J. thank the Swedish Research Council (Grant No. 2014-3981).

Supplementarymaterials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.electacta.2021.138275.

References

[1] C.D. Iwu , A.I. Okoh , Preharvest transmission routes of fresh produce associated bacterial pathogens with outbreak potentials: a review, Int. J. Environ. Res. Public Health 16 (2019) 4407 .

[2] F.J. Gormley , C.L. Little , N. Rawal , I.A. Gillespie , S. Lebaigue , G.K. Adak , A 17-year review of foodborne outbreaks: describing the continuing decline in England and Wales (1992–2008), Epidemiol. Infect. 139 (5) (2011) 688–699 . [3] S. Wang , H. Duan , W. Zhang , J.W. Li , Analysis of bacterial foodborne disease

outbreaks in China between 1994 and 2005, FEMS Immunol. Med. Microbiol. 51 (1) (2007) 8–13 .

[4] C.R. Kuske , Current and emerging technologies for the study of bacteria in the outdoor air, Curr. Opin. Biotechnol. 17 (2006) 291–296 .

[5] T. Bintsis , Foodborne pathogens, AIMS Microbiol. 3 (2017) 529–563 . [6] G. Skretas , A.K. Meligova , C. Villalonga-Barber , D.J. Mitsiou , M.N. Alexis ,

M. Micha-Screttas , et al. , Engineered chimeric enzymes as tools for drug discovery: generating reliable bacterial screens for the detection, discovery, and assessment of estrogen receptor modulators, J. Am. Chem. Soc. 129 (2007) 8443–8457 .

[7] O. Lazcka , F.J. Del Campo , F.X. Munoz , Pathogen detection: a perspective of tra- ditional methods and biosensors, Biosens. Bioelectron. 22 (2007) 1205–1217 . [8] E. Galikowska , D. Kunikowska , E. Tokarska-Pietrzak , H. Dziadziuszko , J.M. Los ,

P. Golec , et al. , Speciﬁc detection of Salmonella enterica and Escherichia coli strains by using ELISA with bacteriophages as recognition agents, Eur. J. Clin. Microbiol. Infect. Dis. 30 (2011) 1067–1073 .

[9] J.A. Jordan , M.B. Durso , Real-time polymerase chain reaction for detecting bacterial DNA directly from blood of neonates being evaluated for sepsis, J. Mol. Diagn. 7 (2005) 575–581 .

[10] K.Y. Yoon , C.W. Park , J.H. Byeon , J. Hwang , Design and application of an inertial impactor in combination with an atp bioluminescence detector for in situ rapid estimation of the eﬃcacies of air controlling devices on removal of bioaerosols, Environ. Sci. Technol. 44 (2010) 1742–1746 .

[11] B. Seddon , T. McKay , L. MacKenzie , J.S. Kemp , D.R. Fenlon , L. Farr , the use of via blue and DEFT for the detection of yeasts, molds and bacteria, Int. Biodeterior. Biodegrad. 32 (1993) 3–18 .

[12] C.M. Duarte , C. Carneiro , S. Cardoso , P.P. Freitas , R. Bexiga , Semi-quantitative method for Staphylococci magnetic detection in raw milk, J. Dairy Res. 84 (2017) 80–88 .

[13] L. Lu , G. Chee , K. Yamada , S. Jun , Electrochemical impedance spectroscopic technique with a functionalized microwire sensor for rapid detection of foodborne pathogens, Biosens. Bioelectron. 42 (2013) 4 92–4 95 .

[14] Y. Zhu , M. Jovic , A. Lesch , L. Tissieres Lovey , M. Prudent , H. Pick , et al. , Im- muno-aﬃnity amperometric detection of bacterial infections, Angew. Chem. Int. Ed. 57 (2018) 14942–14946 .

[15] G.K. Mc , D. Collins , P. Anandarajah ,A review of chipless remote sensing solutions based on RFID technology, Sensors 19 (22) (2019) 4829 .

[16] Z. Meng, Z. Li, RFID tag as a sensor - a review on the innovative designs and applications, 16(2016) 305.

[17] E. Ghafar-Zadeh , Wireless integrated biosensors for point-of-care diagnostic applications, Sensors 15 (2015) 3236–3261 .

[18] L. Cui , Z. Zhang , N. Gao , Z. Meng , Z. Li , Radio frequency identiﬁcation and sensing techniques and their applications-a review of the state-of-the-art, Sensors 19 (2019) 4012 .

[19] J. Zhang , G.Y. Tian , A.M.J. Marindra , A.I. Sunny , A.B. Zhao , J. Zhang , et al. , A review of passive RFID Tag antenna-based sensors and systems for structural health monitoring applications, Sensors 17 (2) (2017) 265 .

[20] Z. Zou , Q. Chen , Q. Chen , I. Uysal , L. Zheng , Radio frequency identiﬁcation en- abled wireless sensing for intelligent food logistics, Philos. Trans. A Math. Phys. Eng. Sci. 372 (2014) 20130313 .

[21] L.K. Fiddes , N. Yan ,RFID tags for wireless electrochemical detection of volatile chemicals, Sens. Actuators B 186 (2013) 817–823 .

[22] L. Ma , Z. Nie , Y. Huang , S.Z. Yao , Inductance-based sensing technique for wireless, remote-query measurement in liquid media, Sci. China Chem. 53 (2010) 1391–1397 .

[23] S. Bhadra , C. Narvaez , D.J. Thomson , G.E. Bridges ,Non-destructive detection of ﬁsh spoilage using a wireless basic volatile sensor, Talanta 134 (2015) 718–723 . [24] K.G. Ong , J.S. Bitler , C.A. Grimes , L.G. Puckett , L.G. Bachas , Remote query resonant-circuit sensors for monitoring of bacteria growth: application to food quality control, Sensors 2 (2002) 219–232 .

[25] K.G. Ong , J. Wang , R.S. Singh , L.G. Bachas , C.A. Grimes , Monitoring of bacteria growth using a wireless, remote query resonant-circuit sensor: application to environmental sensing, Biosens. Bioelectron. 16 (2001) 305–312 .

[26] K.G. Ong , L.G. Puckett , B.V. Sharma , M. Loiselle , C.A. Grimes , L.G. Bachas , Wire- less passive resonant-circuit sensors for monitoring food quality, Proc. SPIE Int. Soc. Opt. Eng. 4575 (2002) 150–159 .

[27] M.S. Mannoor , H. Tao , J.D. Clayton , A. Sengupta , D.L. Kaplan , R.R. Naik , et al. , Graphene-based wireless bacteria detection on tooth enamel, Nat. Commun. 3 (2012) 17671–17678 .

[28] M.S. Saravanan , J.K. Singh , N. Thirumoorthy , RFID sensors for food safety centre by identifying the physical factors that affecting the food, in: Proceedings of the International Conference on Information Communication and Embedded Systems (ICICES2014), 2014, pp. 1–6 .

[29] S. Karuppuswami , L.L. Matta , E.C. Alocilja , P. Chahal , A wireless RFID Compat- ible sensor tag using gold nanoparticle markers for pathogen detection in the liquid food supply chain, IEEE Sens. Lett. 2 (2018) 1–4 .

[30] L.L. Matta , S. Karuppuswami , P. Chahal , E.C. Alocilja , AuNP-RF sensor: an innovative application of RF technology for sensing pathogens electrically in liquids (SPEL) within the food supply chain, Biosens. Bioelectron. 111 (2018) 152–158 . [31] R.A. Potyrailo , N. Nagraj , Z. Tang , F.J. Mondello , C. Surman , W. Morris , Bat- tery-free radio frequency identiﬁcation (RFID) sensors for food quality and safety, J. Agric. Food Chem. 60 (2012) 8535–8543 .

[32] K.S. Sultan , T.A. Ali , N.A. Fahmy , A. El-Shibiny , Using millimeter-waves for rapid detection of pathogenic bacteria in food based on bacteriophage, Eng. Rep. 1 (2019) e12026 .

[33] P. Kalimuthu , J.F. Gonzalez-Martinez , T. Ruzgas , J. Sotres , Highly stable passive wireless sensor for protease activity based on fatty acid-coupled gelatin composite ﬁlms, Anal. Chem. 92 (2020) 13110–13117 .

[34] R. Narang , S. Mohammadi , M.M. Ashani , H. Sadabadi , H. Hejazi , M.H. Zariﬁ, et al. , Sensitive, real-time and non-intrusive detection of concentration and

(10)

Commun. Biol. 3 (2020) 512 .

[42] E.R. Au - Sanders , Aseptic laboratory techniques: plating methods, JoVE (2012) e3064 .

[43] A. Butterworth , E. Blues , P. Williamson , M. Cardona , L. Gray , D.K. Corrigan , SAM composition and electrode roughness affect performance of a DNA biosensor for antibiotic resistance, Biosensors 9 (2019) 22 .

[44] C.K. Alexander , M.N.O. Sadiku , Fundamentals of Electric Circuits, 6th ed., Mc- Graw-Hill, New York, 2017 .

[45] M. Ehrmann , The Periplasm, ASM Press, American Society for Microbiology, 2006 .

[46] A. Williams , Mechanism of action and speciﬁcity of proteolytic enzymes, Quart. Rev. 23 (1969) 1–17 (London) .

(2019) 159–168 .

[54] A.M.J. Marindra , G.Y. Tian , Chipless RFID sensor tag for metal crack detection and characterization, IEEE Trans. Microwave Theory Tech. 66 (2018) 2452–2462 .

[55] P.G. Bowler , The 10(5) bacterial growth guideline: reassessing its clinical rele- vance in wound healing, Ostomy Wound Manage 49 (2003) 44–53 . [56] A.C. Browne , M. Vearncombe , R.G. Sibbald , High bacterial load in asymptomatic

diabetic patients with neurotrophic ulcers retards wound healing after application of Dermagraft, Ostomy Wound Manage 47 (2001) 44–49 .

[57] M.L. Wilson , L. Gaido , Laboratory diagnosis of urinary tract infections in adult patients, Clin. Infect. Dis. 38 (2004) 1150–1158 .