Short Ligand, Cysteine-Modified Warnericin RK Antimicrobial Peptides Favor Highly Sensitive Detection of Legionella pneumophila.

Culture-based methods for the detection of Legionella pneumophila are prohibitively slow and frequently inadequate. The problem has been addressed with biosensing technology that employs a variety of ligands for the specific capture of bacteria. However, the limited success of the application of mammalian antibodies, aptamers, and nucleic acid-based probes for sensitive biosensing has generated growing interest in exploring alternative biosensing architectures, such as those based on antimicrobial peptides (AMP) that are known for their attractive therapeutic applications. We report on the successful employment of cysteine-modified warnericin RK AMP for the operation of a highly sensitive biosensor of L. pneumophila based on digital photocorrosion of GaAs/AlGaAs nanoheterostructures. The replacement of the relatively cumbersome procedure commonly applied for the attachment of antibodies to COOH-terminated mercaptohexadecanoic acid self-assembled monolayers has allowed for a significant reduction in the distance at which bacteria are immobilized above the biosensor surface. An important consequence of this approach is the attractive limit of detection of L. pneumophila estimated at 2 × 102 CFU/mL. The target bacteria were captured four times more efficiently than P. fluorescens, B. subtilis, and E. coli, which is highly promising for environmental monitoring.

Introduction

Rapid detection of pathogenic bacteria in a water environment remains a challenging issue.[1,2] Of particular interest are rapid, portable, non-labor intensive, yet cost-attractive tools for detection of pathogens.[3] The culture-based, colony counting methods have been widely used to detect bacteria,[4] but they are labor and time intensive.[4,5] For instance, Legionella pneumophila may require up to 10 days of incubation for visible detection of colonies.[6] Alternatively, polymerase chain reaction (PCR)-based detection[7] or matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) spectroscopy[8] could both provide relatively fast and accurate detection. However, the need for highly trained personnel and sophisticated lab requirements are the main constraints of these techniques.[9,10] The biosensor-based detection methods of pathogenic bacteria have gained attention due to their potential to offer relatively fast, portable, and easy-to-handle solutions.[11] In that context, various types of L. pneumophila biosensors have been investigated ranging from optical[12,13] and piezoelectric[14,15] to electrochemical.[16,17] However, to the best of our knowledge, an economically attractive method for automated monitoring of water reservoirs for the presence of pathogenic bacteria has yet to be developed. Recently, photoluminescence (PL)-based detection of E. coli has proven to be rapid and relatively sensitive.[18] In this technique, the sensitivity of PL varies with the presence of electrically charged molecules (i.e., proteins, viruses, and bacteria) on the surface of semiconductor nanoheterostructures that have the potential for the realization of a regenerable system designed for automated data collection.[19−22]

L. pneumophila is a pathogenic waterborne bacterium, predominantly found in man-made artificial water reservoirs, i.e., spas and cooling towers.[23,24] Humans who are accidentally exposed may develop a pulmonary infection known as Legionnaire’s disease.[24,25] In 2006, more than 6000 cases were reported in Europe, 400 of them are fatal.[25] Of the 60 reported Legionella species, 85–90% are associated with Legionnaire’s disease.[26] Therefore, the detection of L. pneumophila in water reservoirs has emerged as a public health priority. Culture-based methods have been commonly applied for the detection of L. pneumophila;[27,28] however, they are inefficient. In addition to a multi-day delayed detection, some viable but non-culturable bacteria could not be detected with these techniques. The PCR-based method can address most of the aforementioned problems, but the requirement of trained personnel and sophisticated laboratory facilities limit the application of this technique.[4] Biosensor-based detection of L. pneumophila has the potential to alleviate these deficiencies.

The efficiency of bio-recognition elements is crucial in order that biosensor technology be able to offer a selective, sensitive, and accurate measurement of the target.[29,30] Numerous bio-recognition elements, such as antibodies (Ab), carbohydrates, aptamers, peptides, as well as combinations of these, have been widely explored in different biosensing platforms.[11] Among them, Ab have been considered as an attractive option since they can be highly specific toward the antigenic target.[31,32] However, Ab suffer from lack of stability, especially under extreme environmental conditions, such as high/low pH and elevated temperatures,[31] and they often require additional conjugating compounds, such as neutravidin, biotin, or avidin.[33] These conjugations increase the number of interfaces, which could affect the reproducible performance of a biosensor. Recently, some studies have shown that antimicrobial peptides (AMPs) could be reasonable candidates for bio-recognition in biosensing platforms.[32,34,35] The multiple molecular niches of an AMP seem partially responsible for the strong interaction with bacteria and fungi surface moieties.[31] Furthermore, it has been reported that some cationic AMPs could maintain their activity in harsh environments, even after boiling and autoclaving.[36,37] The considerably superior stabilities of AMPs over those of typical globular proteins[31,37] justify the AMP research expected to lead to the replacement of typical Ab-based bio-recognition elements.

In the past few years, several AMPs such as magainin I,[38] clavanin A,[39] and polymyxin B[40] have been investigated as bio-recognition probes, with some AMPs demonstrating a highly specific recognition capacity.[31] Mannoor et al. reported that a gold electrode functionalized with magainin I AMP showed differential binding affinity to the pathogenic bacterial strains of E. coli and Salmonella at 107 CFU/mL.[31] Another study reported that a gold microelectrode functionalized with Leucocin A (Leu A) exhibited high binding affinity to Listeria monocytogenes.[41] It has been reported that warnericin RK is a membrane active peptide, which shows high specificity to L. pneumophila.(42) However, the exact interaction between a peptide and a bacterium is not clearly understood. It has been proposed that initially, the peptide attaches to the target cell surface due to a general electrostatic interaction with negatively charged bacteria. This initial association is followed by a specific interaction of the peptide with a specific, yet unidentified, L. pneumophila membrane moiety.[43] It has also been reported that the selectivity of warnericin RK to L. pneumophila might be related to the fatty acid composition of the cell membrane.[43]

In our recent study, we employed a digital photocorrosion (DIP) biosensor biofunctionalized with a 16-mercaptohexadecanoic acid (MHDA) linker[54] for interfacing warnericin RK AMP, which allowed detection of L. pneumophila at 103 CFU/mL.[44] Given that a DIP biosensor is sensitive to the flow of electric charge between the biosensor and immobilized bacteria, we hypothesized that a short-linker biosensor, consisting also of a reduced number of interfaces, could exhibit a significantly enhanced sensitivity. Thus, we report here on the operation of an innovative DIP biosensor comprising cysteine-modified RK AMP (Cys-AMP) designed for rapid detection of L. pneumophila. A successful investigation was also carried out by demonstrating a negligible specificity of the biosensor toward P. fluorescens, B. subtilis, and E. coli.

Results and Discussion

Functionalization of GaAs/AlGaAs Biosensors

The immobilization of peptides on the surface of GaAs was evaluated by FTIR analysis as presented in Figure (for a detailed list of peak positions see Table S1). The FTIR absorbance spectra were obtained by subtracting from the spectrum of freshly etched GaAs (001) sample. The absorbance band at 1235 cm–1 was assigned to amide III, while the band at 1519 cm–1 could be assigned to amide II.[50,51] The absorbance at 1540 and 1655 cm–1 is a characteristic for C=O stretching of amide I.[52,53]

Figure 1

Representative FTIR absorbance spectra of thiol and peptide related peaks (a), and amide A absorbance spectra collected for different peptide concentrations (b).

The intense bands at 2922 and 2850 cm–1 observed in Figure a are typical of CH2 asymmetric and symmetric vibrations and are related to the thiol groups of peptides reported in the literature.[1,32] The absorbance bands at 1653 and 1587/1734 cm–1 were assigned to amide I and amide II, respectively.[55,56] Similarly, the peptide immobilized at the C-terminal with a free N-terminal region shows a characteristic peak at 1653 cm–1.[52,57] Furthermore, the band observed at 1734 cm–1 corresponds to C=O stretching of lateral chain functions and some hydrolyzed ester functions.[55,58] The intense peaks at 1653 and 1734 cm–1 suggest a helical conformation.[52,57,58] Furthermore, the band at 1587 cm–1 suggests the presence of a N–H bond for amide II,[59] while the band at 3324 cm–1 could be assigned to amide A.[60,61] Therefore, the amide related peaks in the FTIR spectra (1235, 1519, 1587, 1653, 1734, and 3324 cm–1) confirm the successful immobilization of peptides on the surface of GaAs via the cysteine linker of peptides. It is noticeable that the peaks of a similar intensity were observed in Figure a for amide I, II, and III of different concentration peptides, while the amide A intensity varied with the increasing concentrations of peptides as presented in Figure b. The peak intensity increased with the increasing concentrations of peptides until 50 μg/mL, thereafter, the peak intensity did not correlate with higher peptide concentrations. Hence, the 50 μg/mL could be considered as the optimum peptide concentration for GaAs functionalization.

The AFM evaluation of Cys-AMP functionalized GaAs surface topography is illustrated in Figure , while the corresponding σRMS values are plotted in Figure S1. The micrograph in Figure a presents a freshly etched GaAs surface, and Figure b–h demonstrates the roughness of the GaAs surface functionalized with peptides of different concentrations. As shown in Figure a, the freshly etched GaAs surface is characterized by σRMS = 0.54 nm, which is consistent with the previously published results.[22] Following exposure to peptides at concentrations of 2, 5, 10, 25, 50, 75, and 100 μg/mL, the GaAs surface was characterized by σRMS of 0.73, 0.78, 1.02, 1.16, 1.45, 1.49, and 1.49 nm, respectively (Figure b–h). A comparable surface roughness is observed for GaAs exposed to 50 and 100 μg/mL of peptides, which is consistent with the saturation effect, also recorded with the FTIR measurements (Figure b). Clearly, the concentration of peptides at 50 μg/mL appears optimal for the functionalization of the GaAs surface.

Figure 2

Representative AFM micrographs of GaAs reference surface (a), and peptide-coated GaAs at 2 (b), 5 (c), 10 (d), 25 (e), 50 (f), 75 (g), and 100 μg/mL (h).

Figure a illustrates the dependence of the water contact angle of the GaAs surface on the concentration of peptides employed for functionalization. The contact angle values decreasing from 83 to 66° were observed for surfaces functionalized with peptides at 2, 5, 10, 25, 50, 75, and 100 μg/mL. Note that the oxidized surface of GaAs is characterized by contact angles exceeding 90°.[62] The increased hydrophilicity of GaAs following the deposition of peptides is consistent with the results of Date et al.[63] who showed that the air bubble angles decreased substantially with increasing concentrations of peptides on the gold surface.

Figure 3

Water contact angle of the GaAs surface exposed to different concentrations of cysteine-modified peptides (a), and representative XPS spectra in the C 1s absorption region for the uncoated GaAs (b), and exposed to 50 μg/mL of a peptide solution (c). The error bars in (a) represent standard deviations of three repetitions.

XPS data for bare and peptide functionalized (50 μg/mL) GaAs surfaces are presented in Figure b,c. The C 1s spectra for both functionalized and non-functionalized samples were observed at 284.8 eV, ascribed to C–H and C–C bonds.[64] The peak at 286.3 eV could be assigned to the carbon atoms of the C=O or C–N[65,66] or to the O=C—N.[64,67] The amide related peak at 288.08 eV,[67,68] observed only for the functionalized sample, confirms the presence of peptides.

Surface Coverage with Bacteria

Representative optical micrographs of the GaAs surface functionalized with different concentrations of Cys-AMP and exposed to the L. pneumophila suspension at 106 CFU/mL are shown in Figure a–h, while the capture efficiencies are summarized in Figure i. The uncoated (reference) surface of GaAs was able to capture ∼72 bacteria/mm2, whereas peptides at 2, 5, 10, 25, 50, 75, and 100 μg/mL yielded 217, 301, 394, 1009, 1868, 1920, and 2011 average bacteria/mm2, respectively. The captured bacteria steadily increased with the concentration of peptide to 50 μg/mL, thereafter demonstrating a tendency toward saturation.

Figure 4

Representative optical micrographs of L. pneumophila captured on the GaAs surface using several concentrations of peptides: (a) reference, (b) 2, (c) 5, (d) 10, (e) 25, (f) 50, (g) 75, and (h) 100 μg/mL, and (i) density of captured L. pneumophila on the reference (R) and peptide functionalized GaAs surfaces. The asterisks indicate significantly different values compared to the reference as determined by the Student’s t test (n = 3, p < 0.05).

To evaluate the specificity of peptide toward L. pneumophila, a series of tests were carried out against the non-target P. fluorescens, B. subtilis, and E. coli bacteria at 106 CFU/mL with the GaAs chips functionalized with either peptides cysteine-modified warnericin RK (Cys-WRK) AMP or anti-L. pneumophila Ab. The representative micrographs of the biochip surfaces are shown in Figure , and resulting bacterial capture efficiencies are illustrated in the Figure . The Cys-WRK AMP functionalized biochips captured L. pneumophila, B. subtilis, P. fluorescens, and E. coli, on average, at 2018, 477, 331, and 216 bacteria/mm2, respectively. This is compared with the ability of anti-L. pneumophila Ab functionalized biochips to capture the same bacteria, on average, at 742, 217, 186, and 165 bacteria/mm2, respectively. These results clearly demonstrate that RK AMP peptide, and as expected, anti-L. pneumophila Ab-coated GaAs surfaces captured L. pneumophila more efficiently compared to other bacteria. Notably, the Cys-WRK AMP functionalized GaAs biochips captured L. pneumophila at least four times more efficiently than the other investigated bacteria.

Figure 5

Representative optical micrographs of different bacteria on uncoated GaAs surface (a–d), and GaAs surfaces functionalized with AMP at 50 μg/mL (e–h) and Ab at 100 μg/mL (i–l).

Figure 6

Summary of the results indicating that a cysteine-modified warnericin AMP biosensor captured L. pneumophila four times more efficiently than the other investigated bacteria. The asterisks indicate significantly different values compared to the reference (p < 0.05) as determined by the Students t test (n = 3).

A number of studies have evaluated binding affinity as well as interaction between peptide and bacteria on the surface of biosensor substrates. For instance, Etayash et al.[32] observed that the 24AA LeuA-conjugated gold substrate exhibited high binding specificity toward Gram-positive bacteria, while lower specificity was observed for short length peptide 14AA LeuA. In another study,[1] gold substrates functionalized with 37AA LeuA were incubated with four different bacteria (i.e., E. coli, Listeria innocua, Coronabacterium divergens, and Listeria monocytogenes) and it was found that the L. monocytogenes demonstrated the highest binding efficiency compared to other bacteria. Recent studies have reported that L. pneumophila sensitivity toward warnericin could be due to the lipid composition of the bacterial membrane. Verdon et al.[43] investigated the sensitivity of L. pneumophila to warnericin RK and found that the presence of branched-chain fatty acids on the surface of the bacteria play a crucial role in the sensitivity of the bacteria to the peptide. Legionella contains unusually high amounts of phosphatidylcholines (30%), which are predominantly present in eukaryotic cells only.[69] Furthermore, phosphatidylcholine is not typically present in other bacterial cell membranes.[24,69] A number of Legionella-specific peptides have been reported.[24,43] Of these, only three peptides, warnericin RK being one of them, were found to be specific toward L. pneumophila serogroups 1, 3, 5, and 6.[24] Therefore, the results obtained here, in agreement with previous reports, suggest that the innovative warnericin RK-conjugated GaAs-based biosensor could be an attractive system for specific detection of L. pneumophila.

Detection of L. pneumophila

Detection of L. pneumophila was carried out with the PL effect employed for monitoring DIP of GaAs/AlGaAs nanoheterostructures. The PL scans of the Cys-AMP functionalized biochips exposed to different concentrations of L. pneumophila are shown in Figure . The PL maxima were observed at 15, 20, 27, 34, 46, 57, and 76 min for 0, 102, 5 × 102, 103, 104, 105, and 106 CFU/mL of L. pneumophila, respectively. The details of this experiment are summarized in Table . Under optimized conditions, the PL maximum at 20 min obtained for a bacterial suspension at 102 CFU/mL is delayed from the 15 min maximum observed for the reference sample. At the same time, the 21 min PL maximum observed for the mixed suspension of B. subtilis at 105 CFU/mL and L. pneumophila at 102 CFU/mL suggests that the limit of detection of the biosensor is at ∼2 × 102 CFU/mL. The delayed positions of PL maxima revealed for the growing concentrations of L. pneumophila are consistent with the sensitivity of DIP GaAs/AlGaAs nanoheterostructures to the presence of bacteria immobilized on the biosensor surface. In this system, the rate of photocorrosion of GaAs/AlGaAs nanoheterostructures is delayed due to the charge transfer between bacteria and the semiconductor, as suggested previously.[21,70] A mixed suspension of B. subtilis at 105 CFU/mL with L. pneumophila at 102 CFU/mL was used as a control to demonstrate the specificity of a proposed biosensor. When mixed together, the bacteria showed a PL maximum at 21 min, whereas L. pneumophila alone at 102 CFU/mL yielded a PL maximum at 20 min, which illustrates that the PL maxima were not affected in a measurable manner by the presence of non-target bacteria. We also observed that inter-experimental (different biochips) errors for determining PL maxima varied less than 13%, which indicated a relatively highly reproducible detection. Furthermore, the reproducible response of the biosensor was demonstrated for GaAs/AlGaAs chips functionalized with peptide solutions stored at room temperature for 30 days. The related PL scan, collected for L. pneumophila at 5 × 102 CFU/mL (sample S), revealed the PL maximum position at 27 min, which is similar to that obtained for the fresh peptide solution. A summary of several recent studies reporting on biosensing of L. pneumophila is provided in Table S2. Aziziyan et al.[33] detected 104 CFU/mL of L. pneumophila using an Ab functionalized GaAs/AlGaAs DIP biosensor. In their subsequent study,[21] they improved the detection limit to 103 CFU/mL by decorating bacteria with sodium dodecyl sulfate (SDS). However, a decoration step of bacteria with SDS increases the complexity of a detection protocol, and thus, it may not be entirely advantageous in comparison to the simple process of detecting L. pneumophila with a Cys-AMP-based biosensor.

Figure 7

(a) Normalized PL intensity of AMP functionalized GaAs/AlGaAs DIP biochips (wafer D3422) exposed to different concentrations of L. pneumophila in 0.1 × PBS, (b) PL peak positions vs different concentrations of L. pneumophila. The PL peak positions obtained for L. pneumophila are statistically different compared to either 0.1× PBS (reference), 102 CFU/mL of L. pneumophila, or the negative control test for B. subtilis + L. pneumophila exposed surfaces, as determined by the Student’s t test (n = 3, p < 0.05). The dashed line highlights the biosensing resolution of the device against peak positions of the negative test and the results obtained for L. pneumophila suspension at 102 CFU/mL.

Table 1

PL Maxima Obtained for the Reference (PBS) Run and Different Concentrations of L. pneumophila (All Experiments Repeated for at Least Three Times)

bacteria	concentrations (CFU/mL)	PL maxima (min)
PBS	0.1×	15 ± 13%
L. pneumophila	102	20 ± 10%
L. pneumophila	5 × 102	27 ± 11%
L. pneumophila	5 × 102(S)	27 ± 9%
L. pneumophila	103	34 ± 9%
L. pneumophila	104	46 ± 9%
L. pneumophila	105	57 ± 7%
L. pneumophila	106	76 ± 7%
B. subtilis + L. pneumophila	105 + 102	21 ± 13%

Experimental Section

Materials and Reagents

Undoped, double-side-polished GaAs (001) chips (Wafer WV 23084, Wafer Technology Ltd., Washington, USA) were used for measuring bacteria capture efficiency. GaAs/Al0.35Ga0.65As nanoheterostructure wafers (Canadian Photonics Fabrication Centre, Ottawa, Ontario, Canada) were used for monitoring the DIP process of biofunctionalized chips. The details for employing GaAs/AlGaAs nanoheterostructures in DIP biochips have been reported elsewhere.[19,33] Semiconductor grade isopropanol, acetone, and OptiClear were purchased, respectively, from Fisher Scientific (Ottawa, Canada), National Diagnostics (Mississauga, Canada), and ACP (Montréal, Canada). Anhydrous ethanol was purchased from Commercial Alcohols Inc. (Brampton, Canada). Ammonium hydroxide (28% of NH4OH) used for removing oxides from the GaAs surface was purchased from Anachemia (Richmond, Canada). Phosphate-buffered saline solution (PBS; 10×, pH 7.4) and 16-mercaptohexadecanoic acid (MHDA) thiol were purchased from Sigma-Aldrich (Oakville, Canada) and ViroStat, Inc. (Portland, ME), respectively. Anti-L. pneumophila polyclonal Ab were purchased from ViroStat, Inc. Green fluorescent L. pneumophila JR32 was obtained from the Faculty of Agricultural and Environmental Sciences, McGill University (Ste-Anne de Bellevue, Québec, Canada). Bacillus subtilis ATCC 60514 and Escherichia coli ATCC 25922 were obtained from the Department of Biology, Université de Sherbrooke (Quebec, Canada), and Pseudomonas fluorescens ATCC 13525 was purchased from Cedarlane (Burlington, Ontario, Canada). Cys-AMPs (GenScript, Piscataway, USA) were employed to achieve robust functionalization of GaAs/AlGaAs chips, thus, taking advantage of the strong affinity of sulfur toward Ga and As.[45]

Biofunctionalization of GaAs-Based Chip Surface

Bulk GaAs (001) chips, 2 mm × 2 mm, were used for carrying out Fourier transform infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM), and bacteria capture efficiency measurements. The samples of bulk GaAs and GaAs/AlGaAs nanoheterostructures were cleaned in ultrasonic baths of acetone, OptiClear, and isopropanol for 5 min each, and then dried with high purity nitrogen gas.[46,47] Thereafter, native oxides present on the surface of the samples were removed with 28% NH4OH (2 min at room temperature) followed by immediate dipping of the samples in degassed ethanol and rinsing with copious amounts of degassed DI water. Different concentrations of peptide solutions (2–100 μg/mL) were prepared for functionalizing the GaAs and GaAs/AlGaAs chips. To investigate the stability of the proposed biosensor, aliquots of peptide solution (50 μg/mL) were stored at room temperature for up to 30 days. Functionalization was achieved by immersing cleaned samples in peptide solution for 1 h. The functionalized chips were sonicated in degassed DI water for 1 min and immediately rinsed with degassed DI water to remove non-immobilized peptides.

Preparation of Bacteria

E. coli,P. fluorescens, and B. subtilis were obtained from fresh cultures in a Luria-Bertani (LB) medium. L. pneumophila ssp1 were cultured in a buffered charcoal yeast extract agar (BCYE), supplemented with isopropyl thio-β-galactoside (IPTG) and l-cysteine. Subsequently, after 3-day incubation, several colonies of L. pneumophila were transferred to 0.1× PBS, and concentration of bacteria was determined by OD600nm measurement (0.1 OD 600nm corresponds to 6.4 × 107Legionella/mL). Serial dilutions were carried out in 0.1× PBS to achieve the test concentrations.

Biosensor Architecture

Following the removal of native oxides from the surface of GaAs/AlGaAs chips, the samples were immersed for 1 h in Cys-AMPs suspended in DI water. Subsequently, a 1 min sonication in degassed DI water was applied to remove non-immobilized peptides. However, it was found that light rinsing with DI water was sufficient to remove weakly bonded (physisorbed) peptides, which may be important for the future development of a procedure for automated biofunctionalization. The biofunctionalized chips, typically less than 60 min from their fabrication, were exposed to different suspensions of bacteria. Figure illustrates the process of a biosensor fabrication. Notice that the strong interaction of warnericin RK AMP with L. pneumophila is expected to result in the rapid breaking of the bacterial outer membrane as illustrated by the inset in Figure . It is important to note that the entire detection procedure of this biosensor could be completed within ∼2 h, including the biofunctionalization step, as compared to the more than 20 h required by a biosensor employing an alkanethiol self-assembled monolayer. Furthermore, of potential importance to the operation of charge sensitive sensors (DIP, electrochemical or field-effect devices) is the remarkably short, 2 nm distance between the biochip surface and bacteria immobilized with the Cys-AMP architecture, which could affect the process of charge transfer.

Figure 8

Cysteine-modified warnericin RK antimicrobial peptide-based architecture of the L. pneumophila biosensor employing GaAs/AlGaAs nanoheterostructure chips. The inset illustrates the proximity of interaction between bacterial cell membrane and peptides.

In the case of anti-L. pneumophila Ab functionalization, the etched samples (after being treated with 28% NH4OH) were immersed in 1 mM of MHDA thiol in 10 mL of deoxygenated ethanol for 20 h. After the thiolation, the functionalized chips were sonicated in degassed ethanol for 1 min and immediately rinsed with degassed ethanol to remove non-immobilized thiols. Thereafter, the −COOH terminals of thiolated samples were activated using 0.4 M EDC:0.1 M NHS (1:1) solution for 30 min and immediately rinsed with DI water. Then, the samples were incubated in 100 μg/mL of anti-L. pneumophila polyclonal Ab for 1 h. Finally, both AMP and Ab functionalized samples were incubated with heat-killed L. pneumophila at 106 CFU/mL for 1 h. The GaAs/Al0.35Ga0.65As nanoheterostructures were used for the fabrication of 2 mm × 2 mm DIP biochips functionalized by following the aforementioned protocols. A schematic diagram of a typical sample functionalization and bacteria attachment to GaAs/AlGaAs-MHDA-AMP and GaAs/AlGaAs-MHDA-Ab functionalized biochips is presented in Figure S2.

Fourier Transform Infrared Spectroscopy Analysis

The FTIR absorption spectroscopy measurements were collected using a Bruker Optics Hyperion 2000 FTIR system. The spectra were collected with a resolution at 4 cm–1, and individual spectra were averaged over 1000 scans. All FTIR data were recorded with a liquid N2 chilled HgCdTe (mercury cadmium telluride) IR detector. A reference GaAs sample was obtained by consecutive cleanings in ultrasound baths with OptiClear, acetone, isopropanol, acetone, and ethanol (5 min each), and then etched with a 28% NH4OH solution.

Atomic Force Microscopy Analysis

Topographic images of functionalized GaAs samples were taken with an atomic force microscope (AFM, Shimadzu Instruments, SPM-9700, Japan) operating at room temperature (∼25 ± 2 °C). The root mean square surface roughness (σRMS) was calculated based on scans collected from 5 μm × 5 μm surface areas of the investigated samples. Images were analyzed using AFM Gwyddion software (version 2.53).

X-ray Photoelectron Spectroscopy Analysis

XPS spectra were recorded with a Kratos Analytical AXIS (Ultra DLD XPS) spectrometer employing an Al Kα source (1486.6 eV) operating at 150 W. The XPS data were obtained with a 60° take-off angle with respect to the surface normal. The carbon signals were measured and fitted using Casa XPS software for both bulk GaAs and GaAs/AlGaAs biochip samples. The binding energy reference to the adventitious C 1s peak at 284.8 eV positioned the As 3d5/2 peak at 40.8 eV, which was subsequently used as a nominal calibration.

Contact Angle Measurements

The water hydrophilicity of the peptide functionalized GaAs surface was determined at room temperature using commercial static water contact angle measurement equipment (KRÜSS DSA30). The GaAs surface was exposed to a 10 μL droplet of Milli-Q water, and after 5 s, the contact angle of the GaAs-water interface was calculated.

Optical Microscopy Analysis

Optical microscopy (Nikon Instruments, Inc.) was used to determine the density of immobilized bacteria on the biochip surface. The images were taken at 200× magnification in three different regions of each sample surface. All experiments were repeated at least three times. The bacteria surface coverage was calculated using ImageJ software.[48]

Photoluminescence Measurements

The detection of bacteria was carried out at room temperature with a DIP GaAs/AlGaAs biosensor whose PL was measured with a quantum semiconductor photonic biosensing (QSPB) reader described elsewhere.[33,49] Reference measurements and bacteria-coated biochips were irradiated with 5 s pulses delivering 17 mW/cm2 each, in every 20 s period, using a light emitting diode (LED) operating at a wavelength of 660 nm. The PL signal and images of the biochips collected in situ were recorded with a charge-coupled device (CCD) camera. Experiments were carried out in a 0.1× PBS solution, and runs without bacteria were used to obtain the reference signal. All experiments were repeated at least three times.

Conclusions

This study has demonstrated the innovative concept of a cysteine-modified warnericin RK antimicrobial peptide (Cys-AMP) architecture for construction of a biosensor for rapid detection of L. pneumophila in an aqueous environment. The biosensing architecture was employed for functionalization of GaAs/AlGaAs nanoheterostructure biosensors operating on the principle of a digital photocorrosion. The role of peptide concentration on the efficiency of capturing L. pneumophila was investigated with FTIR, AFM, XPS, and water contact angle measurements. The absorbance band peaks related to peptide, observed at 1653 cm–1 (amide I), 1734 cm–1/1538 cm–1 (amide II) and 3324 cm–1 (amide A), confirmed the chemisorption of peptide on the GaAs surface. Our results showed that 50 μg/mL of Cys-AMP was the optimum concentration as determined by maximum capture of L. pneumophila visualized with optical microscopy. The detection sensitivity of the developed biosensor was investigated in the range of 102 to 106 CFU/mL of L. pneumophila, with the limit of detection estimated at 2 × 102 CFU/mL. Thus, the investigated GaAs/AlGaAs nanoheterostructure DIP biosensors demonstrate functionality, which is attractive for the rapid and direct detection of L. pneumophila present in a water environment at a relatively low concentration. The specificity of the biosensor was rated against P. fluorescens, B. subtilis, and E. coli abundantly found in samples of the environmental water. The Cys-AMP functionalized GaAs biochips showed a capture efficiency of over four times greater for L. pneumophila compared to the other investigated bacteria. The important consequence of the proposed Cys-AMP biosensing architecture is that it requires a relatively short time for completion, which may be found attractive for the operation of other biosensors of L. pneumophila compatible with the thiolation procedure. Furthermore, the short length of the employed ligand could potentially result in an enhanced charge transfer between bacteria and the biochip surface, thus leading to an enhanced performance of charge sensing biosensors.