This is the submitted version of the article: Maldonado J., Estévez M.-C., Fernández-Gavela A., González-López J.J., González-Guerrero A.B., Lechuga L.M.. Label-free detection of nosocomial bacteria using a nanophotonic interferometric biosensor. Analyst, (2020). 145. : 497 - . 10.1039/c9an01485c. Available at: https://dx.doi.org/10.1039/c9an01485c Analyst Label-free detection of a few number of nosocomial bacteria using a nanophotonic interferometric biosensor Journal: Analyst Manuscript ID Draft Article Type: Paper Date Submitted by the n/a Author: Complete List of Authors: Maldonado, Jesus; Catalan Institute of Nanoscience and Nanotechnology Estevez, M.-Carmen; Catalan Institute of Nanoscience and Nanotechnology Fernández Gavela, Adrián ; Catalan Institute of Nanoscience and Nanotechnology Gonzalez-Lopez, Juan Jose; Hospital Universitari Vall d'Hebron Gonzalez-Guerrero, Ana Belén; Catalan Institute of Nanoscience and Nanotechnology Lechuga, Laura; Catalan Institute of Nanoscience and Nanotechnology (ICN2), Please do not adjust margins Analyst Page 2 of 19 1 2 3 4 5 6 7 ARTICLE 8 9 10 11 12 Label-free detection of a few number of nosocomial bacteria 13 using a nanophotonic interferometric biosensor 14 15 Jesús Maldonadoa, M.-Carmen Estevez*a, Adrián Fernández-Gavelaa, Juan José González-Lópezb, 16 Ana Belén González-Guerreroa and Laura M. Lechugaa 17 Nosocomial infections are a major concern at worldwide level. Early and accurate identification of nosocomial pathogens 18 is crucial to provide a timely and adequate treatment. A prompt response also prevents the progression of the infection to 19 life-threatening conditions, such as septicemia or generalized bloodstream infection. We have implemented two highly 20 sensitive methodologies using an ultrasensitive photonic biosensor based on bimodal waveguide interferometer (BiMW) 21 for the fast detection of Pseudomonas aeruginosa and methicillin-resistant Staphylococcus aureus (MRSA), two of the 22 most prevalent bacteria associated to nosocomial infections. For that, we have developed a biofunctionalization strategy 23 based on the use of a PEGylated silane (silane-PEG-COOH) which provides a highly resistant and bacteria-repelling surface, 24 which is crucial to specifically detect each bacterium. Two differentiated biosensor assays have been settled: one based on 25 a specific direct immunoassay employing polyclonal antibodies for the detection of P. aeruginosa and another one 26 employing aptamers for the direct detection of MRSA. The biosensor immunoassay for P. aeruginosa is fast (it only takes 27 12 min) and specific, detecting concentrations down to 49 cfu mL-1 (cfu: colony forming unit). The second one relies on the 28 use of an aptamer that specifically detects penicillin-binding protein 2a (PBP2a), a protein only expressed in the MRSA 29 mutant, providing the photonic biosensor with the ability of identifying the resistant pathogen MRSA and differentiating it 30 from methicillin-susceptible S. aureus (MSSA). Direct, label-free, and selective detection of whole MRSA bacteria was 31 achieved with a limit of detection around 29 cfu mL-1. Both results can fulfill the clinical requirements of sensitivity for the 32 Received 00th January 20xx, diagnosis of infections, demonstrating the great potential of this interferometric biosensor device as a versatile, sensitive 33 Accepted 00th January 20xx and specific tool for bacterial detection and quantification providing a rapid method for the identification of nosocomial 34 DOI: 10.1039/x0xx00000x pathogens at the point-of-care. 35 36 37 38 39 40 Introduction with kidney, urinary tract, and upper respiratory tract as 41 preferred sites of colonization.3-5 Methicillin-resistant 42 Nosocomial infections (infections acquired at the hospital or Staphylococcus aureus (MRSA) is a leading pathogen 43 primary care centers) and the emergence of antibiotic associated with serious hospital diseases.6-9 MRSA is especially 44 resistant bacteria are major health care issues. Habitually, 1 alarming given its multidrug resistance pattern to β-lactam 45 out of 25 patients at hospital becomes infected and as many as antibiotics. It causes a range of illnesses, from skin and wound 46 1 in 9 dies as a consequence of the acquired infection.1,2 infections, to pneumonia and bloodstream infections that can 47 Pseudomonas aeruginosa contributes to 11% of all nosocomial result in sepsis and death. MRSA has recently been found as 48 infections resulting in high mortality and morbidity rates. It is a the main cause of healthcare associated infections (HAI) which 49 main causative pathogen of surgical and wound infections, in 2014 caused 72,444 invasive infections leading to a 13% of 50 urinary tract infections (UTI), cystic fibrosis, and bacteremia, mortality 10, 11. 51 Moreover, an early detection of the causative pathogens of 52 a Nanobiosensors and Bioanalytical Applications Group, Catalan Institute of nosocomial infections is crucial for prescribing an effective 53 Nanoscience and Nanotechnology (ICN2), CSIC, BIST and CIBER-BBN, Campus UAB, treatment before the disease progresses.12 A rapid 54 Bellaterra, 08193 Barcelona, Spain. identification of the infective pathogen will greatly help in the E-mail: mcarme.estevez@icn2.cat 55 b Department of Clinical Microbiology, Hospital Universitari Vall d'Hebron, Vall selection of the most adequate antimicrobial treatment, 56 d'Hebron Institut de Recerca (VHIR), Universitat Autònoma de Barcelona, 08035 limiting broad-spectrum antibiotic therapy, and reducing 57 Barcelona, Spain morbidity and mortality. Current methods for bacteria ‡Electronic Supplementary Information (ESI) available: [details of any 58 supplementary information available should be included here]. See detection in the laboratory involve culture incubation, which 59 DOI: 10.1039/x0xx00000x takes 2-4 days to be completed. Molecular techniques based 60 Please do not adjust margins Page 3 of 19 Please do not adjust margins Analyst 1 ARTICLE Journal Name 2 3 on PCR detect different nosocomial pathogens in an effective the first order mode. Both modes keep travelling through the 4 and specific way.13 However, these techniques also require sensing region of the bimodal device until the output of the 5 long and laborious protocols, which can take several hours, waveguide. Both modes have different penetration of their 6 and involve trained personnel and specialized equipment. evanescent fields into the medium and therefore they have 7 Thus, there is a need of new analytical technologies capable of different sensitivities to any refractive index change at the 8 performing pathogen identification, detection and sensor surface.28 Both modes produce an interference pattern 9 quantification with high sensitivity, specificity and speed, at the exit of the waveguide, which is related to the changes 10 which circumvent the use of any additional labels and that induced by any interaction event happening on the sensor 11 require minimal sample volume and processing. surface. This is due to the fact that both modes are differently 12 Several biosensors have been previously employed to detect affected through the evanescent fields which induce a phase 13 bacteria, such as surface plasmon resonance (SPR), variation (ΔΦ) in the light output. In order to perform the 14 demonstrating in some cases adequate limits of detection. experiments, a sensing window is opened along the bimodal 15 However, most of these SPR assays do not offer a direct section of the waveguide by removing the top-cladding, 16 detection of the whole microorganism and rather rely on DNA allowing the evanescent wave of both modes to be in contact 17 bacterial extraction and identification,14 or require the use of with the biomolecules. This sensing configuration provides an 18 nanoparticles and enzymes to amplify the sensor signal.15 excellent sensitivity with a resolution (minimum detectable 19 Electrochemical,16-19 and impedimetric biosensors20 have been change of refractive index) of only 10-8 RIU. 29, 30 20 also employed for nosocomial pathogen detection, showing The BiMW biosensor device has been successfully employed 21 good performance and limit of detection (LOD). However, for the rapid detection of Escherichia coli and Bacillus cereus.31 22 these approaches also need to amplify the signal with enzymes We have now modified the silanization protocol to improve 23 and/or nanoparticles. even more the surface-repelling properties of the biosensor 24 The development of biosensors for whole bacteria detection is surface and its usefulness has been demonstrated with the 25 challenging because bacteria tend to adhere to surfaces. The capture and detection of two highly prevalent nosocomial 26 properties of the sensor surface (i.e. charge and bacteria in a label-free way. Specifically, we have: (i) employed 27 hydrophobicity) and the presence of many different antigens a silanization protocol, which involves the formation of a 28 on the bacteria surface can easily lead to non-specific silane-PEG-COOH layer that avoids non-specific bacteria 29 interactions with the sensor surface.21 These interactions can adsorptions on the sensor surface; (ii) developed a BiMW- 30 affect the overall signal (i.e. resulting in false positive or based immunoassay for P. aeruginosa detection and; (iii) 31 altered quantifications) and must therefore be minimized as developed a detection assay for the specific detection of MRSA 32 much as possible. Introducing polyethylene glycol (PEG) based on the use of an aptamer which solely recognizes PBP2a 33 molecules to cover the surface is a good option as their protein, not present in methicillin-susceptible S. aureus 34 chemical structure confers high flexibility, hydrophilicity and (MSSA). 35 biocompatibility. The surface modification with PEG to The capability of the BiMW biosensor allowed the rapid 36 generate bacteria-repelling surfaces has attracted increasing detection (<20 min) of whole bacteria at low concentrations 37 attention22-25 and grafted PEG coatings have been employed to (8×102 cfu mL-1), reaching LODs between 29-50 cfu mL-1, 38 prevent biofouling of bacteria and cells on different sensor without the need of additional amplification steps or sample 39 surfaces. However, coating protocols involved complex pretreatment. Our approach exemplifies a promising 40 procedures, such as inert conditions, toxic solvents (e.g. alternative for simple, direct, rapid, and label-free bacteria 41 benzene and toluene), intense shaking, and involve long-time detection, which can further be employed in decentralized 42 (>24 h).26, 27 Thus, the generation of a biointerface with the settings, such as emergency units, where prompt infection 43 capability of preventing non-specific adhesion by a detection and bacteria identification become crucial. 44 straightforward method compatible with a direct 45 immobilization would be extremely useful in the biosensor Experimental 46 field. 47 Based on this, we have combined a highly sensitive optical Chemical and biological reagents 48 biosensor with a biosurface functionalization for the rapid, 49 label-free and specific detection of nosocomial bacteria. We 50 have employed a photonic biosensor based on bimodal 51 interferometric waveguides (BiMW), whose sensitivity stands 52 as one of the highest reported in the literature for label-free 53 detection. The working principle of a BiMW biosensor is based 54 on the interferometric design using a single straight waveguide 55 (see Figure 1). The BiMW sensor has a first waveguide zone 56 exhibiting a single-mode behavior, where only the 57 fundamental mode is propagating. After some distance, the 58 core thickness of the waveguide is increased, acting as a modal 59 splitter and supporting two light modes: the fundamental and 60 2 | J. Name., 2012, 00, 1-3 This journal is © The Royal Society of Chemistry 20xx Please do not adjust margins Please do not adjust margins Analyst Page 4 of 19 1 Journal Name ARTICLE 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Fig. 1. a) Scheme of the BiMW device. General view of the main components and the sensing principle. Inset image shows a picture of a BiMW chip 37 containing 20 sensors. b) Schematic representation of the biofunctionalization strategy: the Si3N4 sensor surface is modified with silane-PEG-COOH, and 38 specific bacteria receptors (antibody or aptamers) are covalently immobilized through the COOH group. 39 40 Triethoxysilane polyethylene glycol carboxylic acid (600 Da) C. Purification of bacteria was carried out by centrifugation at 41 (Silane-PEG-COOH) was purchased from Nanocs (USA). 3000 rpm for 15 min at 4º C. Bacteria were then suspended in 42 Solvents for the sensor cleaning process: dry toluene, acetone, PBST (P. aeruginosa and E. coli) or PBS (MRSA and MSSA). 43 ethanol, hydrochloric acid (HCl, 35-38%), and methanol were Bacteria concentrations were confirmed by a nanophotometer 44 supplied by Panreac (Spain). 1-ethyl-3(3-dimethylaminopropyl) from bioNovacientífica, S.L. (Spain). MRSA aptamer with an 45 carbodiimide hydrochloride (EDC), N-hydroxysulfosuccinimide amino modification at the 3’ end (5’-Aptamer-NH2-3’) and 46 (sulfo-NHS), the reagents for buffers preparation recombinant S. aereus PBP2a protein were purchased from 47 glutaraldehyde, and MgCl2 (1 mM) were purchased from Base Pair technologies (USA). The dry aptamer was suspended 48 Sigma-Aldrich (Spain). The buffers employed are the following: in PBS buffer at a final concentration of 100 µM. Once the 49 PBS (10 mM phosphates, 2.9 mM KCl, 137 mM NaCl pH 7.4); aptamer is dissolved in the buffer, it must be heated to 90 °C 50 PBST (PBS with 0.005% Tween 20); MES (0.1 M, 0.5 M NaCl, pH for 2 min and then cooled down to room temperature before 51 5.5) and HEPES (10 mM pH 7.4), SU-8 2025 and use. 52 polydimethylsiloxane (PDMS) polymers used for the flow cell 53 fabrication were purchased from Microchem (USA) and BiMW sensor and experimental set-up 54 Sylgard® (USA), respectively. The BiMW sensors are designed and fabricated at wafer level 55 Polyclonal anti-Pseudomonas antibody was purchased from using standard microelectronic technology in Clean Room 56 Abcam® (Spain). All bacteria strains were provided by Vall facilities as previously described 32. Each wafer contains 12 57 d’Hebron Hospital (Barcelona, Spain). The bacteria were sensing chips, and each chip included an array of 20 bimodal 58 streaked onto Luria broth (LB) and incubated overnight at 37º waveguide interferometric sensors of 30 mm length with a 59 60 This journal is © The Royal Society of Chemistry 20xx J. Name., 2013, 00, 1-3 | 3 Please do not adjust margins Page 5 of 19 Please do not adjust margins Analyst 1 ARTICLE Journal Name 2 3 pitch of 250 µm between them. A scheme of the employed introduced in the sensor surface during the silanization 4 experimental set-up is shown in the Supporting Information process. Carboxyl groups were first activated by flowing a 5 (SI) (Figure S1). The experimental set-up includes the following solution of 0.2 M EDC/0.05 M sulfo-NHS in MES buffer. Next, a 6 modules: laser, optical components, microfluidic delivery solution of the anti-Pseudomonas (50 µg mL-1 in PBS) was 7 system, temperature controller, photodetector, and injected over the freshly activated sensor surface at a flow rate 8 electronics and data acquisition. A He-Ne laser (λ=632.8 nm, of 10 μL min-1. After immobilization, milliQ water was switched 9 TE polarization) is employed as light source, a 40X microscope to PBST as running buffer, which was circulated over the 10 objective and a beam expander are employed for light in- sensor surface at a constant flow of 20 μL min-1. Bacterial 11 coupling and a two-sectional Si photodiode is employed for samples (between 8x102 to 1x107 cfu mL-1 in PBST) were 12 out-coupling detection. The flow delivery system includes a loaded into a 150 µL loop and flowed over the sensor surface 13 syringe pump for the continuous delivery of the buffer and an with a flow rate of 20 μL min-1. A 100 mM HCl solution 14 injection valve connected to a loop of 150 μL for injecting the (injected at 30 μL min-1) was used as regeneration buffer to 15 samples. The sensor signal (%) obtained for each measurement break the bacteria/antibody interaction. Calibration curves 16 is expressed as ΔΦ considering that a complete interferometric were obtained for each bacterium after performing triplicate 17 oscillation corresponds to 2π rad phase variation. The data measurements of different samples of known concentration. 18 acquisition and analysis included a computer with a data The estimated LOD was calculated as the concentration 19 acquisition card, home-made LabVIEW software, and Origin corresponding to the minimum measurable signal, set as three 20 8.0 software (Originlab, Northampton, MA). times the standard deviation (SD) of the baseline of the sensor 21 For biosensing evaluation in continuous flow, and in order signal. 22 to perform the analysis of aqueous samples, it was necessary 23 to incorporate a microfluidic cell and a flow delivery system. Aptamer biofunctionalization and assay development. 24 The microfluidic cell was fabricated in PDMS polymer, building Aptamer immobilization was carried out using an ex-situ 25 a microfluidic cell with eight channels with a pitch of 250 μm protocol (i.e. outside the optical setup following successive 26 between them. Each microchannel has a length of 18 mm, a incubation/washing steps). In this case, the sensor surface was 27 height of 50 μm and width of 100 μm, which covers the activated with a solution of 0.4 M EDC/0.1 M sulfo-NHS in MES 28 sensing window of each sensor. The alignment of the buffer for 3 h. After rinsing with water, an aptamer solution (2 29 microfluidic system over the bimodal waveguide sensor was μM in PBS containing 1mM MgCl2) was incubated over the 30 performed using a home-made aligner, which comprises three sensor surface overnight at room temperature. Finally, the 31 3-axis micromanipulators and a vacuum pump. In order to sensor chip was washed with PBS and water, dried with a N2 32 obtain a continuous flow, the microfluidic cell was connected stream and mounted on the experimental setup. PBS buffer 33 to the fluidic system described in the previous section. The was set as running buffer for MRSA detection. All the bacteria 34 volume flow rate was modified according to the bioassay samples (ranging from 8×102 to 1.6×107 cfu mL-1) were loaded 35 requirements, oscillating between 5 and 30 μL min-1. in a 150 µL loop and flowed over the sensor surface at a flow 36 rate of 20 μL min-1. After each measurement, HCl (25 mM 37 Sensor chips cleaning and chemical activation injected for 60 s at 20 μL min-1) was flowed to dissociate the 38 The BiMW sensor chips were sonicated in acetone, ethanol, interaction and regenerate the aptamer and reuse the 39 milli-Q water and a 1:1 methanol/HCl solution in order to aptasensor. 40 remove organic contamination. Then they were rinsed with 41 milli-Q water and dried with a N2 stream. An oxidation step to SEM analysis 42 reveal enough oxide groups on the sensor surface was carried The images of bacterial adhesion to the sensor surface were 43 out by placing the sensor chips in an UV/O3 cleaner (BioForce examined by a 400L field-emission scanning microscope (FE- 44 Nanosciences, USA) for 1 h, followed by their exposure to a SEM; FEI Magellan) operated at an acceleration voltage of 2 kV 45 10% nitric acid solution at 75° C for 25 min. After rinsing with and with a working distance of 4.5 mm. A Quanta 650F 46 milli-Q water, the sensor chips were incubated with a silane- Scanning Electron Microscope (Field Emission Inc., USA), 47 PEG-COOH solution (25 mg mL-1 in ethanol/water 95:5 (v/v)) operated in 2 kV, was used to assess the selectivity of the 48 for 2 h at 4 ºC. Then the sensor chips were rinsed with ethanol aptamer for the detection of MRSA, confirming the binding of 49 and water and dried with a N2 stream. A final curing step was whole MRSA on the aptamer-modified sensor surface. All 50 performed by introducing the sensor chips into a glass bottle bacteria were first suspended in PBST buffer. Si3N4 surfaces 51 and inside a conventional autoclave for 90 min at 120 ºC and without and with silane-PEG-COOH, and biofunctionalized with 52 at a pressure of 1.5 bars. the aptamer were placed in a solution of 107 cfu mL-1 of each 53 bacterium for 1 h, and then PBST buffer was replaced with a 54 Antibody biofunctionalization and assay development fixation buffer (glutaraldehyde 5% in HEPES buffer) and 55 Antibody immobilization was performed in-situ (i.e. by incubated for another hour. The Si3N4 surfaces were washed 56 mounting the silanized sensor chip with the flow cell, placed it with PBST and dried with a nitrogen stream before 57 in the optical setup and flowing the different solutions). Anti- examination by SEM. 58 Pseudomonas antibodies were immobilized on the sensor 59 surface through a covalent binding to the carboxyl groups 60 4 | J. Name., 2012, 00, 1-3 This journal is © The Royal Society of Chemistry 20xx Please do not adjust margins Please do not adjust margins Analyst Page 6 of 19 1 Journal Name ARTICLE 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Fig. 2. SEM images of the bacterial adhesion in untreated Si3N4 sensor surfaces (top) and in sensor surface treated with PEG-silane-COOH (bottom). Scale bars 26 of inset images (top) are 2, 5, and 1 µm, respectively. 27 detailed information). The non-specific adsorptions of 28 29 Results and discussion different bacteria on Si3N4 surfaces coated with the silane- PEG-COOH were first evaluated using S. aureus, as a 30 Analysis of bacterial adhesion onto silicon nitride sensor representative model of Gram-positive bacteria and E. coli and 31 surface P. aeruginosa as representative of Gram-negative bacteria. 32 A key point in the development of a surface refractive index They were selected since they are three of the most common 33 based biosensor is to completely avoid unspecific adsorptions. pathogens identified in nosocomial infections. As can be seen 34 In general, any adsorbed mass on the surface produces a in Figure 2 (top), significant non-specific adsorptions of 35 refractive index change and, therefore, renders in altered bacteria over untreated Si3N4 surfaces were clearly observed, 36 sensor values that affect an accurate quantification. In this being particularly high for S. aureus (Gram-positive) in 37 context, bacteria are especially complex targets since they are comparison with E. coli and mainly with P. aeruginosa (Gram- 38 prone to adsorb on any type of surfaces, giving rise to negative), which showed a much lower adhesion. These results 39 unspecific responses in biosensors, which cannot be exemplify also how different bacteria can behave in terms of 40 distinguishable from the specific ones. The design of a surface adsorption depending on their own surface properties. Even in 41 that minimizes other bacteria adhesion, that only allows the the case of P. aeruginosa, with an apparently low adhesion, 42 interaction of the target bacteria with their specific the resultant signal in the biosensor could be comparatively 43 bioreceptor but not with the sensor surface, and that ideally high, considering the size of bacteria and the extremely 44 avoids additional blocking steps, is mandatory for the sensitivity level of our device.30 On the other hand, the SEM 45 achievement of competitive and high-performance bacteria images of the silanized surfaces demonstrates that the surface 46 biosensor. PEG is known to decrease the interaction between treatment with PEGylated compounds such as the silane-PEG- 47 silicon surfaces and biomolecules because of its steric COOH helps prevent the bacterial adhesion (see Fig. 2, 48 stabilization forces and chain mobility.33 Therefore, we used a bottom) with all the three bacteria tested. Finally, the ability of 49 silane containing a short chain of PEG (MW=600 Da) with a the silane-PEG-COOH modified surface to attach biomolecules 50 carboxylic group at the other end, in such a way that it was then assessed with the immobilization of the antibodies 51 provided a reactive group to be further coupled to (through the free Lys in their structure) and confirmed using 52 bioreceptors. A simple silanization protocol was performed, XPS. For this experiment, silicon nitride surface was cleaned 53 which involved the incubation of the oxidized Si3N4 surface and modified with silane-PEG-COOH following the same 54 with the silane to form an assembled layer, followed by a high- procedure described. Then the carboxyl groups were activated 55 pressure and high-temperature curing step (achieved by with a solution of EDC/sulfo-NHS and an antibody solution (50 56 means of an autoclave) to stabilize the bonds. Employing this µg mL-1) was incubated over silicon nitride surface during 30 57 strategy, a stable and dense PEGylated layer was obtained as min at room temperature. The details of XPS experiments and 58 assessed by X-ray photoelectron spectroscopy (XPS) (see SI for the comparison of the survey spectra for untreated Si3N4, 59 60 This journal is © The Royal Society of Chemistry 20xx J. Name., 2013, 00, 1-3 | 5 Please do not adjust margins Page 7 of 19 Please do not adjust margins Analyst 1 ARTICLE Journal Name 2 3 silane-PEG-COOH and sensor surfaces with immobilized specific approach to identify these bacteria. Besides antibodies 4 antibody are discussed in SI and Figure S1-S4 therein. commercially available, a specific aptamer was recently 5 developed that specifically recognizes PBP2b. In order to take 6 Immunoassay for the detection of Pseudomonas aeruginosa 7 The BiMW sensor modified with silane-PEG-COOH was placed 8 in the experimental set-up. Then, the carboxyl groups of the 9 silane-PEG-COOH surface were activated with a solution of 10 EDC/sulfo-NHS. An antibody solution was flowed over the 11 sensor surface and the immobilization was monitored in real- 12 time. The mean immobilization signal of the antibody was 13 7.5±2 x2π rad, which represents a high and reproducible 14 antibody coverage due to the covalent attachment (Figure S5). 15 Suspensions of P. aeruginosa at different concentrations were 16 sequentially flowed through the sensor surface. 17 Representative signals of specific P. aeruginosa detection (1.6 18 x 105 cfu mL-1) and non-specific E. coli (1 x 105 cfu mL-1) used as 19 control are shown in Fig.3a. These results show a ΔΦ of 0.4x2π 20 rad for the detection of P. aeruginosa and a negligible signal 21 for the E. coli sample, respectively, which confirms the 22 specificity of the immunoassay provided by the antibody. 23 Additional E. coli samples at different concentrations (from 105 24 to 107 cfu mL-1) were also flowed and no signal was observed 25 for any of them, confirming that the signal contribution was 26 coming only from the specific detection of P. aeruginosa (see 27 Figure 3b). The interaction between the antibodies and the 28 captured bacteria can be further interrupted by the injection 29 of 100 mM HCl for 1 min (see sensorgram in Fig.3c as a 30 representative binding and regeneration cycle) without 31 altering the sensor surface density of antibodies, as can be 32 seen by the complete recovery of the baseline. Under these 33 conditions, the biosensor was reusable up to 5 cycles using the 34 same bioreceptor layer before degradation of the antibody 35 layer (i.e. the signal corresponding to bacteria binding 36 decreased significantly). A calibration curve was generated 37 after analyzing a wide range of bacteria concentrations. The 38 lowest concentration analyzed was 8x102 cfu mL-1 resulting in 39 a ΔΦ of 0.074x2π rad. (see Fig. 3b). Considering the sample 40 volume injected (i.e. 150 µL) this corresponds to 120 bacteria 41 detected. Moreover, considering the minimum measurable 42 signal, which was set as three times the standard deviation of 43 the baseline (8.5×10-5×2π rad), an estimated LOD of 49 cfu mL- 1 was achieved (i.e. around 8 bacteria in a 150 µL volume Fig. 3. (a) Real time sensorgrams obtained for P. aeruginosa and E. coli 44 (negative control). (b) Calibration curve for P. aeruginosa (black squares) 45 sample). These concentration values fall comfortably within and for non-specific bacteria E. coli (red circles). Inset graph: linear 46 the cut-off values in infections such as UTIs (i.e. between 104- representation with exponential fitting (y= y0 + A 𝑒 ―𝑥/1 ) of the P. 47 105 cfu mL-1)34 upper and low respiratory tract infections (cut- aeruginosa detection. Linear part was fitted to a linear regression off at 103 cfu mL-1)35, 36 and could even be suitable for more (y=1.7484 × 10-6x, R2=0.9846). All data show mean + SD of triplicate 48 measurements. (c) Real time sensorgram of P. aeruginosa detection (1.6 x 49 restrictive cut-offs such as for blood stream infections.37 103 cfu mL-1), followed by a regeneration step (dash line indicates the 50 baseline). 51 Aptamer-based assay for the detection of MRSA 52 Detection of PBP2a protein advantage of the silane-PEG-COOH chemistry, we decided to 53 Methicillin-resistant S. aureus emerges after acquiring the modify the aptamer with an amino (-NH2) terminal group and 54 mecA resistance gene, which encodes PBP2a protein.38, 39 This use the same surface chemistry employed with the antibody 55 protein is a unique transpetidase that catalyzes cell-wall previously. We first attempted the detection of pure PBP2a 56 crosslinking and it is not inhibited by β-lactam antibiotics protein in order to evaluate the affinity and the integrity of the 57 (contrary to what happens with the four S. aureus native aptamer after immobilization and its specificity against other 58 Penicillin-binding proteins (PBPs)). Therefore, the detection of interfering proteins. As negative control, we selected a protein 59 this distinctive protein, only present in the MRSA strain, is a located in the cell wall of the Mycobacterium tuberculosis (also 60 6 | J. Name., 2012, 00, 1-3 This journal is © The Royal Society of Chemistry 20xx Please do not adjust margins Please do not adjust margins Analyst Page 8 of 19 1 Journal Name ARTICLE 2 3 a gram-positive bacterium, such as MRSA) as PBP2b in MRSA. 4 The capability of the aptamer to detect PBP2a protein was 5 evaluated at different concentrations (between 50 and 500 ng 6 mL-1); the negative control protein was tested at two 7 concentrations: 500 ng mL-1 and 5 μg mL-1 (ten times higher 8 than the highest concentration of PBP2b evaluated). 9 Representative signals of two concentrations of PBP2a and the 10 negative control are shown in Fig. 4a. These signals show a ΔΦ 11 of 0.1×2π rad for the detection of 100 ng mL-1, and 0.3×2π rad 12 for the detection of 500 ng mL-1 of PBP2a protein, respectively, 13 while the signals for the negative controls are negligible, 14 indicating the excellent specificity of the aptamer. The range of 15 PBP2a concentrations evaluated showed a linear 16 concentration-dependent profile as can be seen in Figure 4b,c. 17 A calibration curve for the direct detection of PBP2a was 18 generated, reaching an estimated LOD of 3.27 ng mL-1 (Fig. 4c). 19 20 Detection of MRSA bacteria 21 In order to evaluate the capability of the aptasensor for whole 22 MRSA detection and establish its overall performance, several 23 solutions of MRSA bacteria were prepared in PBS and injected 24 over the aptamer modified sensor surface. MSSA was also 25 analyzed to assess the specificity provided by the aptamer. 26 Representative signals of MRSA detection (1.6×106 cfu mL-1) 27 and MSSA (2×106 cfu mL-1) are shown in Fig. 4a. These signals 28 show a ΔΦ of 0.41×2π rad for the detection of MRSA, and no 29 signal for the MSSA sample. We established the regeneration 30 conditions that enable the dissociation of the MRSA-aptamer 31 binding, without affecting the aptamer activity or integrity as 32 can be seen in Fig. 4b, the use of 25 mM HCl for 60 s was 33 enough to disrupt the interaction and remove all the bacteria 34 bound to the aptamer, therefore, fully recovering the baseline. 35 Other solutions such as NaOH or NaCl were not able to 36 dissociate the interaction or they affected the conformation of 37 the aptamer, disabling it for further interactions. Under the 38 optimal regeneration conditions, the biosensor could be 39 reused up to 12 cycles before signal decrease (see figure S6 in 40 SI). Figure 4c shows some representative signals of specific 41 MRSA detection at different concentrations. The calibration 42 curve is depicted in Figure 4d. A ΔΦ signal of 0.04×2π rad was 43 Fig. 3. a) Real time sensorgrams obtained for the detection of PBP2a protein observed for the lowest concentration injected (8×102 cfu mL- 1) and increased gradually as the bacteria concentration was 44 and for the control protein over the aptamer biofunctionalized sensor. b) 45 Real-time sensorgrams for several PBP2a concentrations over aptamer- also higher. Considering the baseline signal (8.5×10-5×2π rad), modified sensor chips. c) Calibration curve obtained for PBP2a protein in an estimated LOD around 29 cfu mL-1 was calculated from the 46 buffer. All data show mean + SD of triplicate measurements. 47 linear fitting for the MRSA detection (inset in the Fig. 4d). This 48 level is comparable to the one obtained for P. aeruginosa using 49 antibodies, complying also with the current clinical 50 requirements for the detection of nosocomial infections. 40, 41 51 52 53 54 55 56 57 58 59 60 This journal is © The Royal Society of Chemistry 20xx J. Name., 2013, 00, 1-3 | 7 Please do not adjust margins Page 9 of 19 Please do not adjust margins Analyst 1 ARTICLE Journal Name 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Fig. 4. a) Real time sensorgrams of MRSA (black line) versus MSSA measurement (Red line); b) Real time sensorgram of the MRSA detection, followed by the 30 regeneration step with HCl 25 mM; c) Real time sensorgrams showing the detection of MRSA bacteria at different bacteria concentrations (red dashed line 31 indicates the baseline). d) Calibration curve for whole MRSA detection with the BiMW aptasensor. Signal obtained for MSSA at 106 cfu mL-1 is also shown. Inset graph shows the linear fitting (y=2.967×10-6x, R2=0.9697) for the low bacteria concentration range (i.e. between 800 and 1.6 ×105 cfu mL-1). All data show mean + 32 SD of triplicate measurements 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 Fig. 5. SEM study for the confirmation of the selectivity and specificity of the MRSA aptamer biofunctionalized Si3N4 sensor surfaces. a) MSSA (107 cfu mL-1) on a 58 Si3N4 sensor surface modified with only silane-PEG-COOH (i.e. no aptamer immobilized); b) MRSA (107 cfu mL-1) on a Si3N4 sensor surface modified with only silane-PEG-COOH (i.e. no aptamer immobilized); c) MSSA (107 cfu mL-1) on a sensor surface with MRSA aptamer immobilized and d) MRSA (107 cfu mL-1) on a 59 sensor surface with MRSA aptamer immobilized. 60 8 | J. Name., 2012, 00, 1-3 This journal is © The Royal Society of Chemistry 20xx Please do not adjust margins Please do not adjust margins Analyst Page 10 of 19 1 Journal Name ARTICLE 2 3 Conclusions 1. K.S. Park, C.H. Huang, K. Lee K, Y-E Yoo, C.M. Castro, R. 4 Weissleder, and H. Lee Sci. Adv., 2016; 2: e1600300. We have implemented a rapid, label-free, and highly sensitive 5 BiMW biosensor able to identify and quantify nosocomial 2. S.S. Magill, J.R. Edwards, W. Bamberg, Z.G. Beldavs, G. Dumyati, 6 bacteria. Our strategy is based on the combination of a sensor M.A. Kainer, R. Lynfield, M. Maloney, L. McAllister-Hollod, J. Nadle, 7 S.M. Ray, D.L. Thompson, L.E. Wilson and S.K. Fridkin, N. Engl. J. surface modified with a silane-PEG-COOH, which provides an 8 Med. 2014; 370: 1198-208. excellent bacteria-repelling coating, and the use of appropriate 9 specific receptors (i.e. both antibodies and aptamers). With 3. M. Balasoiu, A.T. Balasoiu, R. Manescu, C. Avramescu and O. 10 this biofunctionalization strategy, non-specific bacteria Ionete, Curr. Health Sci. J. 2014; 40: 85-92. 11 adsorption was successfully eliminated, avoiding extra blocking 4. P. Shukla, R.K. Garg and A.K. Dahiya. Role of technology to 12 steps, which can lengthen the process and affect the final combat nosocomial infections. Apollo Medicine. 2016; 13: 71-3. 13 sensitivity. The silane-PEG-COOH surface was evaluated using 14 5. J.B. Lyczak, C.L. Cannon and G.B. Pier, Microbes. Infect. 2000; 2: SEM to assess its performance and its bacteria-repelling 15 1051-60. capabilities and the biofunctionalized surface was additionally 16 characterized with XPS. The subsequent addition of a suitable 6. S. Stefani, D.R. Chung, J.A. Lindsay, A.W. Friedrich, A.M. Kearns 17 antibody layer led to a BiMW immunosensor that showed a and H. Westh, Int. J. Antimicrob. Agents. 2012; 39: 273-82. 18 selective label-free detection of P. aeruginosa with a valuable 7. X. Lu, D.R. Samuelson, Y. Xu, H. Zhang, S. Wang, B.A. Rasco, J. Xu 19 LOD of only 49 cfu mL-1. Moreover, the versatile chemistry and M.E. Konkel, Anal. Chem. 2013; 85: 2320-7. 20 provided by the COOH groups in the silanized sensor chips, 21 8. S.R. Harris, E.J. Feil, M.T. Holden, M.A. Quail, E.K. Nickerson, N. facilitated also the immobilization of an amino-modified 22 Chantratita, S. Gardete, A. Tavares, N. Day, J.A. Lindsay, J.D. aptamer, providing an efficient and highly specific detection of Edgeworth, H. de Lencastre, J. Parkhill, S.J. Peacock, S.D. Bentley, 23 MRSA bacteria, reaching a LOD of 29 cfu mL-1, avoiding any Science. 2010; 327: 469-74. 24 recognition of the MSSA. The combination of the high 25 sensitivity of the BiMW sensor with the features of the specific 9. K.T. Kavanagh, L.E. Calderon and D.M. Sama, Antimicrob. Resist. 26 aptamer represents an excellent analytical tool to distinguish Infect.Control. 2015; 4: 4. 27 between resistant and susceptible S. aureus in a direct and 10. Prevention CDC. Active bacterial coresurveillance report, 28 rapid way. emerging infections program network, methicillinresistant 29 The methodology used to generate the bacteria-repelling Staphylococcus aureus. 2014. 30 surface, based on a simple process of autoclaving curing of the 11. A. Hassoun, P.K. Linden and B. Friedman, Crit. Care. 2017; 21: 31 reactive silane, can be further employed as a general 211. 32 methodology for developing stable modified sensor surfaces 33 for other silicon photonic biosensors. Overall, the results 12. S. Harbarth, J.Garbino, J. Pugin, J.A. Romand, D. Lew and D. 34 Pittet, Am.J. Med. 2003; 115: 529-35. obtained demonstrate the competitive potential of the BiMW 35 biosensor technology, in combination with a silane-PEG-COOH 13. K.L. Thong, M.Y. Lai, C.S. Teh and K.H. Chua, Trop. Biomed. 36 interface, to provide a label-free, rapid, stable, selective and 2011; 28: 21-31. 37 specific biosensor device for the identification of nosocomial 14. K. Nawattanapaiboon, W. Kiatpathomchai, P. Santanirand A. 38 pathogens. Moreover, our approach could be further Vongsakulyanon, R. Amarit, A. Somboonkaew, B. Sutapun, T. 39 translated to a multiplexed configuration, by simply addressing Srikhirin, Biosens. Bioelectron. 2015; 74: 335-40. 40 the immobilization of different antibodies and aptamers in 41 15. F. Melaine, M. Saad, S. Faucher and M. Tabrizian Anal. Chem. each individual bimodal waveguides and could be easily 42 2017; 89: 7802-7. integrated in a point-of-care instrument to be deployed in 43 clinical settings, in order to facilitate fast bacterial diagnosis, 16. L. Xu, W. Liang, Y. Wen, L. Wang, X. Yang, S. Re, N. Jia, X. Zuo, G. 44 one of the urgent needs in our healthcare system. Liu, Biosens. Bioelectron. 2018; 99: 424-30. 45 17. C.R. Nemr, S.J. Smith, W. Liu, A.H. Mepham, R.M. Mohamadi, 46 M. Labib and S.O. Kelley Analytical Chemistry. 2019, 91, 2847-2853 . 47 Conflicts of interest 48 18. Y. Yang, Y-Y. Yu, Y-Z. Wang, C.L. Zhang, J.X. Wang, Z. Fang, H. Lv, There are no conflicts to declare. J.J. Zhong and Y.C. Yong, Biosens. Bioelectron. 2017; 98: 338-44. 49 50 19. H.J. Sismaet HJ, A.J. Pinto and E.D. Goluch, Biosens. Bioelectron. 51 2017; 97: 65-9. Acknowledgements 52 J. Maldonado acknowledges the Mexican National Council for 20. H. Lee, J.O. Keem, H. Cho, J.M. Choi, W.S. Chung, D.Y. Jeon, D.S. 53 Lee and Y.B. Shin Biosens. Bioelectron. 2018; 118: 153-9. 54 Science and Technology (CONACYT). ICN2 is supported by the 55 Severo Ochoa program from Spanish MINECO (Grant No. SEV- 21. A. Ahmed, J.V. Rushworth, N.A. Hirst and P.A. Millner, Clin. 56 2017-0706). Microbiol. Rev. 2014; 27: 631-46. 57 22. F. Boulmedais, B. Frisch, O. Etienne, P. Lavalle, C. Picart, J. Ogier, 58 J.C. Voegel, P. Schaaf, C. Egles, Biomaterials. 2004; 25: 2003-11. 59 References 60 This journal is © The Royal Society of Chemistry 20xx J. Name., 2013, 00, 1-3 | 9 Please do not adjust margins Page 11 of 19 Please do not adjust margins Analyst 1 ARTICLE Journal Name 2 3 23. J. Jin, Y. Han, C. Zhang, J. Liu, W. Jiang, J.Yin and H. Liang, 4 Coll.Surf. B Biointerfaces. 2015; 136: 838-44. 5 24. K.D. Park, Y.S. Kim, D.K. Han, Y.H. Kim, E.H.B. Lee, H. Suh, K.S. 6 Choi, Biomaterials. 1998; 19: 851-9. 7 25. Q. Wang, E. Uzunoglu, Y. Wu and M. Libera, ACS Appl. 8 Mater.Inter. 2012; 4: 2498-506. 9 10 26. M. Zhang, T. Desai and M. Ferrari, Biomaterials. 1998; 19: 953- 60. 11 12 27. V. Hynninen, L. Vuori, M. Hannula, K. Tapio, K. Lahtonen, T. 13 Isoniemi, E. Lehtonen, M. Hirsimäki, J.J. Toppari, M. Valden, V.P. 14 Hytönen Sci. Rep. 2016; 6: 29324. 15 28. S. Dante, D. Duval, D. Fariña, A.B. González-Guerrero and L.M. 16 Lechuga, Laser Photon. Rev. 2015; 9: 248-55. 17 29. D. Grajales, A. Fernández Gavela, C. Domínguez, J.R. Sendra and 18 L.M. Lechuga, J.Phys. Photonics. 2019; 1: 025002. 19 20 30. A.B. González-Guerrero, J. Maldonado, S. Dante, D. Grajales and 21 L.M. Lechuga LM, J. Biophoton. 2017; 10: 61-7. 22 31. J. Maldonado, A.B. González-Guerrero, C. Domínguez and L.M. 23 Lechuga, Biosens.Bioelectron. 2016; 85: 310-6. 24 32. K.E. Zinoviev, A.B. Gonzalez-Guerrero, C. Dominguez and L.M. 25 Lechuga, J.Lightwave Technol. 2011; 29: 1926-30. 26 27 33. S. Lan, M. Veiseh and M. Zhang, Biosens. Bioelectron. 2005; 20: 28 1697-708. 29 34. K.B. Roberts and E.R. Wald ER, Pediatrics, 2018; 141: 30 e20173239. 31 35. M. Woodhead F. Blasi, S. Ewig, J. Garau, G. Huchon, M. Ieven, A. 32 Ortqvist, T. Schaberg, A. Torres, G. van der Heijden, R. Read, T.J. 33 Verheij, Eur. Resp. J. 2005; 26: 1138-80. 34 35 36. K. Loens, L. Van Heirstraeten, S. Malhotra-Kumar, H. Goossens 36 and M. Ieven, J. Clin. Microbiol. 2009; 47: 21-31. 37 37. O. Opota and A. Croxatto, Clin. Microbiol. Infect. 2015; 21: 313- 38 22. 39 38. B. Blázquez, L.I. Llarrull, J.R. Luque-Ortega, C. Alfonso, B. 40 Boggess and S. Mobashery, Biochemistry. 2014; 53: 1548-50. 41 42 39. T.K. McKinney, V.K. Sharma, W.A. Craig and G.L. Archer, J. 43 Bacteriol. 2001; 183: 6862-8. 44 40. J.H. Kwon, M.K. Fausone, H. Du, A. Robicsek and L.R. Peterson, 45 Am. J. Clin. Pathol. 2012; 137: 778-84. 46 41. E.J.G. Pollitt, P.T. Szkuta, N. Burns and S.J. Foster PLOS 47 Pathogens. 2018; 14: e1007112. 48 49 50 51 52 53 54 55 56 57 58 59 60 10 | J. Name., 2012, 00, 1-3 This journal is © The Royal Society of Chemistry 20xx Please do not adjust margins Analyst Page 12 of 19 1 2 3 4 TOC 5 6 Two methodologies using a BiMW biosensor have been developed for the fast, sensitive 7 detection of P. aeruginosa and MRSA bacteria 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 13 of 19 Analyst 1 2 3 4 5 6 7 8 9 10 11 Supplementary Information 12 13 14 15 16 Label-free detection of few number of nosocomial bacteria 17 18 using a nanophotonic interferometric biosensor 19 20 Jesús Maldonadoa, M.-Carmen Estéveza*, Adrián Fernández-Gavelaa, Juan José González- 21 Lópezb, Ana Belén González-Guerreroa and Laura M. Lechugaa, 22 23 a Nanobiosensors and Bioanalytical Applications Group, Catalan Institute of Nanoscience 24 25 and Nanotechnology (ICN2), CSIC, BIST and CIBER-BBN, Campus UAB, Bellaterra, 08193 26 27 Barcelona, Spain. 28 29 b Department of Clinical Microbiology, Hospital Universitari Vall d'Hebron, Vall d'Hebron 30 31 32 Institut de Recerca (VHIR), Universitat Autònoma de Barcelona, 08035 Barcelona, Spain 33 34 35 36 37 *Corresponding author: 38 39 E-mail address: mcarmen.estevez@icn2.cat 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Analyst Page 14 of 19 1 2 3 4 5 1. XPS analysis 6 7 The XPS experiments were performed using a Phoibos 150 analyzer (SPECS GmbH, Berlin, 8 9 10 Germany) in ultra-high vacuum condition (base pressure 1 × 10-10 mbar). A monochromatic 11 12 AlKα X-ray source (1486.74 eV) operating at 400 W was employed. Survey spectra were 13 14 collected from 0 to 1380 eV with a pass energy of 50 eV, and high-resolution spectra were 15 16 collected for each element (e.g. C, N and O) with a pass energy of 20 eV. Survey and high- 17 18 resolution spectra were collected at 0º take off angle, defined as the angle spanned by the 19 20 electron path to the analyser and the sample surface. The spectra were obtained at room 21 22 temperature. 23 24 25 XPS survey spectra illustrated significant changes in carbon, silicon, oxygen and nitrogen 26 27 peaks, due to the addition of the different molecular layers (see Figure S2).1 The spectra 28 29 with the antibody immobilized shows a significant increase of the carbon and decrease of the 30 31 nitrogen and the silicon, both being contributions from the Si3N4 substrate and the SiO2 32 33 34 oxidation layer. This demonstrates how Si3N4 is not exposed but coated with silane-PEG- 35 36 COOH and antibodies. The elemental analysis indicates a remarkable increase of carbon 37 38 content after antibody immobilization. Carbon increased from 6.2% (Si3N4) to 20.7% on the 39 40 antibody coated surface while nitrogen levels decreased on the silanized surface and was 41 42 even lower with the antibody attached (see Table S1). The high-resolution C 1s spectra (see 43 44 Figure S3) showed that the C 1s photoelectron peaks are much broader after the 45 46 immobilization with antibodies. The C 1s peaks were deconvoluted according to binding 47 48 energies of carbons in antibody. The first peak was centred at 284.7 (Figure S3a) and 49 50 related to aliphatic carbon (C-C). The second peak at 286.5 eV corresponds to carbon 51 52 53 bonded to oxygen group (C-O) (Figure S3b) which evidences the PEG-silane-COOH 54 55 presence. The antibody attachment by the peak around 287.5 eV, corresponds to –C (-O)- 56 57 NH2 peptide carbon (Figure S3c).2, 3 The binding energies of N 1s detected on the samples 58 4, 5 59 (close to 400 eV) are typical for organic materials and are related to C-N bonds, at 398.2 60 Page 15 of 19 Analyst 1 2 3 eV, the spectre corresponding to Si3N4 (see Figure S4a). At the same time, the silane-PEG- 4 5 COOH (Figure S4.b) and the antibody onto Si3N4 (Figure S4c) are indicative of complete 6 7 coverage of the silicon nitride surface. The component around 402 eV peak was assigned to 8 9 10 imide-N, found on the amino acid proteins (see Figure S4.c). 6 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 Figure S1. Schematic representation of the experimental set-up for the evaluation of the BiMW 39 biosensors system. The set-up comprised three mainly components: a flow cell and flow delivery 40 system, optical system for in coupling of the light and the read-out and data acquisition system. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Analyst Page 16 of 19 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Figure S2. XPS survey spectra of: a) Si3N4, b) Si3N4 coated with silane-PEG-COOH and c) silanized 28 surface with antibody covalently attached to the COOH groups. 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 Figure S3. XPS high resolution spectra of C 1s, a) untreated Si3N4, b) silane-PEG-COOH and c) 56 antibody immobilized. 57 58 59 60 Page 17 of 19 Analyst 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Figure S4. XPS high resolution spectra of N 1s, a) untreated Si3N4, b) silane-PEG-COOH and c) 29 antibody immobilized. 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 Figure S5. Real-time sensorgram showing the antibody immobilization. Phase variation due to the 56 refractive index bulk change and the immobilization of antibody is 7.7×2π rad (ΔΦIn-ΔΦOut). 57 58 59 60 Analyst Page 18 of 19 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Figure S6. Real-time detection of 1, 7, 8, 10, and 12 measurement-regeneration cycles of MRSA 28 detection regenerated with 25 mM HCl. 29 30 31 32 Table S1. Elemental composition of Si3N4 susbtrates before and after surface modification with 33 silane-PEG-COOH and antibody 34 35 Sample Elemental composition (at %) 36 C N O Si 37 38 SI3N4 6.26 33.62 14.59 45.53 39 Silane-PEG 10.15 26.68 21.86 41.31 40 41 Antibody 23.32 22.37 20.7 33 42 43 44 References 45 46 1. Michel R, Pasche S, Textor M and Castner DG. Influence of PEG architecture on protein 47 adsorption and conformation. Langmuir. 2005; 21: 12327-32. 48 2. Lebugle A, Subirade M and Gueguen J. Structural characteristics of a globular protein investigated 49 50 by X-ray photoelectron spectroscopy: comparison between a legumin film and a powdered 51 legumin. Biochim. Biophys. Acta 1995; 1248: 107-14. 52 3. Sharma, S., R.W. Johnson, and T.A. Desai. XPS and AFM analysis of antifouling PEG interfaces for 53 microfabricated silicon biosensors. Biosens. Bioelectron. 2004; 20(2): 227-39. 54 4. Serro AP, Gispert MP, Martins MC, Brogueira P, Colaco R and Saramago B. Adsorption of albumin 55 on prosthetic materials: implication for tribological behavior. J.Biomed. Mater. A. 2006; 78: 581-9. 56 5. Vanea E, Tămăşan M, Albon C and Simon V. Synthesis and characterisation of a new composite 57 aluminosilicate bioceramic. J. Non-Crystall. Solids 2011; 357: 3791-6. 58 59 60 Page 19 of 19 Analyst 1 2 3 6. Chen R, Willcox MD, Ho KK, Smyth D and Kumar N. Antimicrobial peptide melimine coating for 4 titanium and its in vivo antibacterial activity in rodent subcutaneous infection models. 5 6 Biomaterials. 2016; 85: 142-51. 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60