This document is the Accepted Manuscript version of a Published Work that appeared in final form in Analytical Chemistry, copyright © American Chemical Society after peer review and technical editing by the publisher. To access the final edited and published work see: https://dx.doi.org/10.1021/acs.analchem.9b04369. Lab in a Tube: Point-of-Care detection of Escherichia coli Niloufar Amina,b,c,‡, Amadeo Sena Torralbaa,‡, Ruslan Álvarez-Diduka, Abbas Afkhamib and Arben Merkoçia,d* a Nanobioelectronics & Biosensors Group, Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Campus UAB, Bellaterra, 08193 Barcelona, Spain. b Faculty of Chemistry, Bu-Ali Sina University, Hamedan, Iran c Food and Drug Laboratory Research Center, Food and Drug Organization, MOH&ME, Tehran, Iran d ICREA, Institució Catalana de Recerca i Estudis Avançats, Pg. Lluís Companys 23, 08010 Barcelona, Spain. ‡ These authors contributed equally to the work. *arben.merkoci@icn2.cat ABSTRACT: Significant levels of infectious diseases caused by instruments, manpower, expense and time that already are not suit- pathogenic bacteria are nowadays a worldwide concern implying able for emergency cases especially in developing countries where considerable public healthcare challenges and huge economic det- medical facilities are unobtainable 7. riment. Due to the fast spread of these bio-threat agents and the To overcome the aforementioned problems, point-of-care (PoC) di- outbreak of diseases, a rapid detection of pathogens at early stages agnostic can be a potential solution for improving the global health is crucial, particularly in low-resource settings. To this aim, we de- managements 8. Toward this important goal, different PoC devices veloped for the first time a new sensing approach carried out in a for detection of E. coli have been implemented including colori- single-step for Escherichia coli O157:H7 detection. The detection metric-based lateral flow immunoassay (LFIA) 9-11 and other non- principle is based on Förster Resonance Energy Transfer using gold paper-based microfluidic devices based on electrochemical detec- nanoclusters as signal reporter and gold nanoparticles conjugated tion 12, colorimetry13, 14, fluorimetry15 and magneto-fluorimetry16. with antibodies as a quencher. The sensing platform includes and LFIA are limited by their poor sensitivity that usually requires sig- UV-LED to provide the proper excitation and consists in a micro- nal amplification, while the other microfluidic systems are rela- tube containing two pieces of fiber glass; one of them is embedded tively expensive and require trained personnel17, 18. Therefore, there with label-free gold nanoclusters and the other one with gold nano- is a lack of a suitable PoC device to detect of E. coli for in field particles conjugated with antibodies. Upon the addition of the sam- assays. ple containing bacteria the florescence of gold nanoclusters is re- Our research group have been involved in developing photolumi- covered. The assay was evaluated by naked eye (on/off) and quan- nescence paper-based platforms for detecting E. coli 19, 20. The re- titatively using a smartphone camera. The biosensor proved to be ported strategies are related to nanoswitch “Off-On” methodology highly specific and sensitive, achieving a limit of detection as low using photoluminescent quantum dots conjugated with antibodies. as 4.0 CFU ml-1. Additionally, recoveries of 110% and 95% were In these platforms, despite of being sensitive and specific, there are obtained when evaluating the platform in spiked river and tap water some drawbacks such as multistep procedures, exclusive readers respectively. platforms and particularly the easy to use for end-user that reduce the applicability of these biosensors for in-field measurements. Previous studies have reported the quenching of gold nanoclusters Keywords: Point-of-Care; Biosensor; Optical detection; Bacteria (AuNCs) fluorescence in the presence of gold nanoparticles detection; FRET; Smartphone-based detection (AuNPs) 21. This phenomenon occurs due to the complementary overlapping between the surface plasmon resonance absorption wavelength of AuNp and the fluorescence emission wavelength of AuNCs. Some biosensors have been developed by exploiting this phenomenon such as the detection of cholesterol22 and uric acid 23. Nowadays, Escherichia coli O157:H7 (E. coli O157:H7) is one of In these cases, the detection is carried out through the production the major threats to public healthcare causing serious infectious dis- of H2O2 in an enzymatic reaction, AuNPs seeds start growing, in- eases mainly through contaminated water and food 1, 2. Since the creasing the absorbance of AuNPs with the consequently decrease causative pathogen agent can spread rapidly, it is imperative to in the fluorescence emission of AuNCs. Although this strategy has achieve fast diagnosis. In addition, providing a quick decision in been exploited for biosensing applications, so far, to the best of our the field to test the safety of water and food will have a great effect knowledge none of these nanomaterials have being used together on controlling and minimizing mortality rate and financial afflic- as an immunosensor. Specifically, if we are starting from the OFF tion. state, fluorescence could be recovered if the distance between the Traditional methods for the identification and detection of patho- donor/acceptor pair becomes big enough genic bacteria are mainly based on plate cultivation3, enzyme- Certainly E. coli bacteria is big enough to promote this effect. For linked immunosorbent assay (ELISA) and polymerase chain reac- this purpose, we developed a simple, fast and cost-effective “lab on tion (PCR)4-6. Culturing method requires long analysis time (∼ a tube” platform for E. coli detection. The proposed approach relies days) and special labor operation. PCR and ELISA need significant on Förster Resonance Energy Transfer (FRET) mechanism based on gold nanoclusters (AuNCs) as energy donor and antibody-func- conjugated AuNPs followed by an incubation for 1 h under the pre- tionalized AuNPs as energy acceptor. This enables an ON/OFF flu- vious conditions. The excess of antibodies and BSA were removed orescent signal depending on the presence or absence of E. coli in by centrifuging the solution at 14 000 rpm and 4 °C for 30 min, the sample. The assay is carried out inside a microtube that contains followed by re-suspension with 200 µL of Milli-Q water. Finally the reagents embedded in glass fiber membrane. In this regard, the the conjugated AuNPs where stored in 4°C for further use. biosensor can be easily transported and stored up to use. The assay Platform fabrication. The glass fiber membranes were cut into a starts by just introducing the sample into the micro-tube. After 20 15 cm × 8 mm strips. 800 μL of the conjugated AuNPs and AuNCs minutes of incubation, the result can be evaluated by naked eye, in where drop-casted on separate glass fiber strips and dried in a vac- which the presence or absence of fluorescence indicates a yes/no uum chamber for 2.5 hours. Finally they were cut in circles (diam- response to the bacteria. Conveniently, it can be evaluated quanti- eter ≈ 0.6 cm) using a hole puncher and introduced in the micro- tatively by using a 3D printed device with an UV-LED, a tubes. smartphone camera and an image-analysis app. Bacteria preparation. Stock solution (1010 E. coli/ml) of bacteria was prepared by dissolving 10 mg E. coli in 1 ml filtered Milli Q EXPERIMENTAL SECTION water followed by shaking with vortex. In order to determine the Materials and Instruments. All commercial reagents are analyti- concentration of the prepared stock solution, the optical density at cal grade and handled according to the material safety data sheets 600 nm (OD600) was measured using a spectrophotometer and cor- suggested by the suppliers. Hydrogen tetrachloroaurate (III) dehy- roborated with Agilent geometry software drate (HAuCl4 3H2O, 99.9%), Trisodium Citrate, Sodium hydrox- (https://www.chem.agilent.com/store/biocalculators/calcODBacte ide, Phosphate buffered saline tablet (P4417), bovine serum albu- rial.jsp). Standard serial solutions of E. coli (0 - 108 CFU/ml) were min (BSA), were purchased from Sigma-Aldrich (Madrid, Spain). prepared by diluting the suspension stock solution (1010 CFU/ml Anti-E.coli antibody (Ab, ab68451) was obtained from Abcam ≈ 1010 E. coli/ml) in PBS. (Cambridge, UK). Escherichia coli O157:H7 (CECT 4783, E. coli) and Salmonella Typhimurium (CECT 722T, S. typhimurium) Assay performance. Serial dilutions of E. coli (0 - 108 CFU/ml) strains were obtained from Sigma-Aldrich (Madrid, Spain). Ana- where prepared in PBS, 200 μL where drop-casted in the microtube lytical grade nitric acid and hydrochloric acid were purchased from and incubated for 20 min in room temperature. The assay was eval- Fisher Scientific Fiber glass membrane (GFCP00080000) was pur- uated with a spectrophotometer by transferring the solution into a chased from Millipore. All glassware were washed with aqua regia, microplate. The fluorescence intensity was measured at 665 nm and rinsed three times with Milli-Q water. Transmission electron (Ext. 365 nm) for each dilution of E. coli and normalized with the microscopy (TEM) images were performed with a FEI Tecnai G2 black (F-F0), where F and F0 corresponded to sample with and with- 20 TWIN electron microscope. Fluorescence emission and Absorb- out E. coli, respectively. For a point-of-care application, the assay ance spectra were collected on a Microplate reader (Molecular de- was evaluated by taking a picture using a mobile phone camera and vice). TS-100 Thermo-Shaker (Biosan, Riga, Latvia) was used as analyzing it with ImageJ software (National Institutes of Health, the stirrer for functionalization of AuNPs with antibodies. Maryland, USA) following the protocol already reported by our Synthesis of BSA-AuNCs. BSA-AuNCs were synthesized accord- group27. All the experiments were replicated three times. ing to a previous protocol24. Briefly, aqueous solutions of HAuCl4 (10 mM, 5 ml) and BSA (50 mg/ml, 5 ml) were mixed and sub- jected to vigorous stirring at 37°C. After 2 min, NaOH solution (1.0 RESULTS AND DISCUSSION M, 0.5 mL) was added into the reaction mixture. The reaction con- Assay principle. The assay was performed in a polypropylene mi- tainer was sealed and allowed to proceed in the dark for 12 h. The crotube, which was used as a reservoir and detection platform. solution became deep brown and showed strong red emission under Firstly, the fluorescent AuNCs were synthesized, purified and a 365-nm UV lamp, pointing out the formation of the AuNCs. The drop-casted on a fiberglass membrane. The fiber glass was then prepared AuNCs were purified through dialysis membrane dried using a vacuum desiccator and cut into circular pieces (Ø 0.6 (MWCO: 1 kDa) against Milli-Q water for one day to remove un- cm) using a hole puncher. The pieces were introduced into the mi- reacted HAuCl4 and sodium hydroxide. Purified AuNCs were kept crotube and stored until running the assay. The same procedure was at 4°C for further use. The concentration of BSA-AuNCs was esti- done for the AuNPs, which were previously conjugated with anti- mated at around 20 mg/ml. bodies against E. coli O157:H7 following the procedure reported in Synthesis and conjugation of AuNPs. The synthesis of AuNPs the experimental section (Figure 1A). In this way, the micro-tube was performed following the protocol reported by Bastús and co- stocked both conjugate pads, one containing the sensing PL probe workers 25. In short, the initial AuNPs seeds were prepared by boil- (AuNCs) and the other containing the quencher label (AuNPs). The ing 150 mL of sodium citrate (2.2 mM) in a sealed condenser, fol- fiber glass was used as a porous material with high absorption and lowed by the injection of 1 ml of HAuCl4 (25 mM). After 10 release capacity of solutions, appropriate long-term storage of sam- minutes, the color of the solution turned from light yellow to pink ples, high stability and good flexibility 28, 29. In order to evaluate and it was cooled up to 90 °C. Next, 1 ml of sodium citrate (60mM) the presence of E. coli in a real environment, the assay starts by was added and after 2 min, 1 ml of HAuCl4 (25mM) was sequen- introducing the sample solution into the microtube followed by a tially added. The solution was kept under stirring at 90 °C for 30 20 minutes incubation step (Figure 1B). The sample solution ena- min and the color of solution finally became wine-red. The final bles the rehydration of the conjugate pad and the eventual release concentration of the AuNPs solution was 3.1 × 1011 NPs/ml. of the stored nanomaterials to the solution. The biosensing mecha- nism is based on Förster Resonance Energy Transfer (FRET), being The AuNPs were bioconjugated with antibodies against E. coli fol- the AuNPs and AuNCs the energy acceptors and donors, respec- lowing the procedure previously reported by our group 26 . First the tively. In this regard, the AuNPs were able to quench the fluores- pH of the AuNPs was adjusted to 8.9 using Borate buffer (10 mM). cence emission of the AuNCs when they were close together. This Next, 100 μL of anti-E. coli (200 μg/mL) were incubated with 1.5 is known as the “turn-off” mode, which is linked to the absence of ml AuNPs for 24 h at 350 rpm and 4 °C. Then 100 μL BSA solution E. coli in the sample. Conversely, in the presence of E. coli, the in Milli-Q water (final concentration 1%(w/w)) was added to the 2 antibody-functionalized AuNPs is attached to the surface of the as shutter speed, ISO and focus, obtaining reproducible and accu- bacteria avoiding the approach of the AuNCs. In this way, due to rate images. In this regard, a 3D printed device was developed to the fact that FRET mechanism is possible up to 30 nm and E. coli introduce the micro-tube using the cap as the sample holder. The where 1µm sized 30, the fluorescence signal was restored, switching device contains a 365 nm UV-LED to excite the AuNCs and a mi- to a “turn-on” mode (Figure 1C). It should be noticed the specific- cro lens to adjust the focal length of the camera to the minimum ity of the assay. In this regard, AuNCs were mixed with AuNPs possible (Figure 1D). The pictures were analyzed using ImageJ with and without bacteria. It can be observed in Figure S1 that app (smartphone version) to transform the brightness levels into a there is no recovery in the fluorescence of AuNCs when AuNPs are numerical value, which allows quantification. Since most of the not conjugated with antibodies. people across the world has a smartphone as a common equipment Finally, a quantitative outcome was obtained using a smartphone in our day-to-day life, this proof of concept makes a highly prom- camera. Nowadays, almost any smartphone has a good quality ising device for point-of-care assay and field applications without CMOS sensor and the possibility to fix the picture parameters such the need to a conventional bulky analyzer such as fluorimeter. Figure 1. Schematic representation of the biosensing platform. (A) Two fiber glass strips embedded with AuNCs and AuNPs conjugated with anti-E. coli antibody respectively. (B) Operation principle; Sample is added directly to the micro-tube containing the two strips. (C) Naked eye evaluation of the ON/OFF switch mechanism. (D) Smartphone based device for quantitative evaluation of the E. coli concentra- tion. Characterization of AuNCs and conjugated-AuNPs. The syn- has a spherical and uniform size of around 40 nm. Besides, its con- thesized AuNCs were characterized by transmission electron mi- jugation to antibodies against E. coli was corroborated by the pres- croscopy (TEM). As observed in Figure 2A, they have a uniform ence of red shift from 525 to 535 nm in the UV-vis spectrum, indi- morphology and a narrow size distribution with an average diame- cating a modification on the AuNPs surface (Figure S2) 31. Finally, ter of about 1.0 nm. Furthermore, a fluorescence emission peak at in order to achieve FRET, the absorbance spectrum of energy ac- 665 nm when excited at 365 nm and a strong red emission under ceptor must be overlapped with the emission spectrum of the donor UV light (Figure 2B). The synthesized AuNPs were also charac- probe. This phenomenon can be observed in Figure 2B, where the terized with TEM, as shown in Figure 2C and Figure 2D. AuNPs emission spectrum of the AuNCs was clearly overlapped with the absorbance spectrum of the conjugated-AuNPs. 3 Figure 2. (A) TEM image of BSA-AuNCs. Inset: the size distribtion histogram of the as prepared AuNCs. (B) Fluorescence spectra of BSA- AuNCs (red line) and absorption spectrum of conjugated AuNPs with Anti-E. coli (black line). Inset: Digital photo of AuNCs under UV light. (C) TEM image of Citrate stabilized- AuNPs. Inset: graph corresponds to size distribution diagram. (D) High magnified of as-prepared AuNPs. Assay optimization. The highest concentration with the higher flu- reduce the development cost of the biosensor). This was done by orescence emission of AuNCs was tested to evaluate the size de- performing the well-stablished gold aggregation test (GAT) 32, pendence on the quenching efficiency of AuNPs. It is known that, which determines the maximum concentration of antibody required the bigger the size of nanoparticles, the highest the red shift in the to cover the AuNPs surface. Based on the results shown in Figure wavelength, causing more overlap with the emission of AuNCs 21. 3B, we concluded that 175 μg/ml of anti-E. coli were enough to Among the tested sizes, 40 nm was selected as the most suitable for cover the surface of the AuNPs, however we round it up to 200 the assay (Figure S3). Besides bigger nanoparticles (>40nm) pro- μg/ml in the subsequent assays. duce more quenching, the antibodies conjugation protocols are The required concentration of AuNCs to achieve an effective FRET fully optimized for 40 nm. Moreover, if bigger nanoparticles are and recovery was another parameter to be optimized. In this con- used the required amount of antibodies to cover the surface will be text, (AuNP at fixed concentration of 1.16x1012 NP/ml) the fluores- higher, causing a more expensive device. To evaluate how the cence recovery in presence of E. coli was evaluated using different AuNP concentration affects the quenching of AuNCs, the stock so- AuNCs concentrations (from 0.5 to 8.0 mg/ml). As expected, an lution (3.1 × 1011 NPs/ml) of AuNP was concentrated (by centrifu- excess of AuNCs brought to signal saturation while a low concen- gation) and diluted. It can be observed in Figure 3A that indeed tration of AuNCs let to negligible fluorescence intensity (Figure higher concentrations of AuNP produces more quenching on the 3C). Therefore, the fluorescence enhancement efficiency (F-F0) light emission of AuNCs. Nevertheless, we have selected 1.16x1012 was employed as a criterion to evaluate the best signal where F and NP/ml as a compromise between efficiency and affordability since F0 correspond to the intensities of fluorescence of AuNCs/AuNPs there is no noteworthy quenching at higher concentrations. system in the presence and absence of E. coli, respectively. Hereby, Furthermore, the antibody concentration on the AuNPs was also 6.0 mg/ml was determined as the optimal concentration of AuNCs. optimized as it dramatically affects the sensitivity and the cost-ef- fectiveness of the platform (using less antibody can considerably 4 Figure 3. Optimization of experimental conditions. (A) Pre-optimization of the quenching efficiency due to AuNPs concentration (B) The effect of the concentration of the immobilized antibody on the surface of AuNPs (GAT) at NaCl 1 % and 0.58× 10 12/ml AuNPs in various concentration of antibody (C) Effect of the concentration of AuNCs as a fluorophore agent on fluorescence recovery efficiency (F-F0) where F and F0 correspond to the fluorescence intensities of AuNCs/AuNPs system in the presence and absence of E. coli, respectively at 105 CFU/mL E. coli (D) The effect of the concentration of anti-E. coli- labeled AuNPs on quenching efficiency (F0-F)/F0 where F and F0 corre- spond to the fluorescence intensities of AuNCs in the presence and absence of Ab-AuNPs, respectively. The concentration of AuNCs was 6.0 mg/ml. Although AuNPs concentration was previously optimized (pre op- exponentially and reached equilibrium after 20 minutes. This re- timization strategy) those nanoparticles were not conjugated with veals the assay time required before performing the final evalua- anti-E. coli antibodies. It is important to double check the concen- tion. tration of conjugated AuNPs required to achieve an effective FRET Smartphone-based detection of E. coli. The developed platform taking in consideration the previously optimized AuNCs concen- was firstly evaluated under optimal conditions by spiking known tration. In this regard, conjugated AuNPs ranging from 0.31 to 4.46 concentrations of E. coli (0 to 107 CFU/mL) in PBS buffer (10 mM (x1012 NPs/mL) were evaluated. As observed in Figure 3D, 2.32 pH 7.4) and using a spectrophotometer. As observed in Figure 4A, (x1012 AuNPs/mL) allowed a 68% quenching efficiency, while the fluorescence signal gradually increases with higher concentra- higher and lower concentrations ended up with signal saturation tion of E. coli, with a good correlation (exponential behavior log[E. and low quenching, respectively. It is also essential to optimize this coli]) in the range of 0 to 105 CFU/mL (Figure 4B), being F and F0 parameter for fluorescence restoration efficiency (Figure S4). the fluorescence intensities at 665 nm in the presence and absence Therefore, this was the concentration of conjugated AuNPs se- of E. coli, respectively. The limit of detection (LOD) was calcu- lected for the next experiments. lated as low as 1.37 CFU/mL based on 3 times “s/m” ratio, where Finally, in order to better understand the response rate of our plat- “s” is the standard deviation of the blank fluorescence signal (three form to E. coli detection, a comprehensive kinetic study was per- replicates) and “m” is the slope of the related calibration curve. Fur- formed (Figure S5) showing that the fluorescence signal increased thermore, the assay was evaluated with a smartphone camera, in order to get as close as possible to point-of-care conditions. In this 5 context, a Samsung galaxy S7 smartphone was used to take a pic- sponding to absence of E. coli (F0). The calibration curve is ob- ture of the microtubes after running the assay (20 minutes incuba- served in Figure 4D, which shows a power fluorescence increase . tion). The micro-tube cap was irradiated under a UV LED (Ex. 365 It is worth to notice a change in the mathematical function in light nm) and the smartphone was attached to a holder, the pictures were of the results obtained with the fluorimeter and the smartphone. always taken in the same conditions; at ISO 100 and shooter speed Graphing the values according to the function obtained with the 1/60 s. As illustrated in Figure 4C, the sample containing 0 fluorometer (log[E. coli]), it can be seen an exponential behavior CFU/mL of E. coli, defined as blank sample, is characterized by which means that the results obtained with the smartphone follows low brightness levels due to the “turn-off” FRET effect. However, a power function from 0 to 107 CFU/mL with a LOD of 4.0 with the increase of E. coli concentration there was a gradual in- CFU/mL. (Figure 4D inset) crease in fluorescence (“turn-on” mode). In order to obtain a nu- The developed biosensor performance of E. coli detection was merical value, the pictures were analyzed with Image J software, in compared to some of the reported point of care sensors and sum- which the fluorescence intensity is transformed into a value from 0 marized in Table S1. to 255 (where 0 is dark and 255 is maximum brightness). The flu- orescence intensity values were normalized with the signal corre- Figure 4. Quantification of E. coli O157:H7 by developed biosensor. (A) Fluorescence spectrum of AuNCs/Ab-AuNPs upon addition of different concentrations of E. coli from 0 to 107 CFU/ mL. (B) Calibration curve of the assay by using fluorimeter. (F and F 0 correspond to the fluorescence intensities of AuNCs/AuNPs system in the presence and absence of E. coli, respectively). (C) Typical images of AuNCs/Ab- AuNPs system in the presence of various concentration of E. coli in the range of 0 - 107 CFU/mL (0, 103, 104, 105, 106 CFU/mL from left to right) captured by smartphone camera under 365 nm UV light irradiation; (D) Calibration curve resulted by processing images in the mobile phone. Error bars represent the standard deviation (n=3) inset: linear behavior of a power fuction. 6 Real samples evaluation. To further test the practicability of the 6 3.5x10 biosensor, the assay was conducted by spiking known concentra- tions of E. coli (102 and 105 CFU) to tap water and river water sam- 6 3.0x10 ples collected from river Ter (Vic, Spain). The samples were eval- uated following the procedure previously reported and the recover- 6 2.5x10 ies where calculated by comparing the results obtained in real sam- ples (i) with the ones obtained in PBS (ii), using the following equa- 6 2.0x10 F-F0 tion: 𝑅𝑒𝑐𝑜𝑣𝑒𝑟𝑦 (%) = 100 𝑥 (𝐹 − 𝐹0)𝑖 /(𝐹 − 𝐹0)𝑖𝑖 10 Salmon. 10 E.coli.+ 6 1.5x10 10 E.coli The results are summarized in Table 1. The obtained recoveries 5 6 1.0x10 5 2 prove that there is no drastic interference when evaluating samples 10 Salmon. with a complex matrix, thus the developed platform can be applied 10 E.coli.+ 10 E.coli 5 for the detection of E. coli in real samples at the point-of-care. 5.0x10 5 2 2 5 0.0 0 10 Salmon. Table.1. Detection of E. coli bacteria spiked in water samples (n=3) [E.Coli](CFU mL-1) Sample Spiked #F-F0 in *F-F0 in Recovery E.coli standard real sample (%)a Figure 5. Specificity of the developed fluorescence biosensor to- (CFUml-1) buffer ward E. coli in the presence of Salmonella as interfering bacteria. River 102 1440843 1503611 103.70 The error bars are the standard deviation of the three measurements. water 105 2791463 2648975 105.37 102 1440843 1400975 97.23 CONCLUSIONS Tap 105 2783852 2740975 98.45 The proposed all-in-one device combines outstanding advantage of water both the glass fiber membrane as a cost-effective substrate with a high loading capacity and the microtube platform as simple and Recovery (%) =100 × (F-F0*/ F-F0#) portable set-up. Moreover, the detection method applied in the de- veloped biosensor is fast (20 minutes) and doesn’t require washing Determination of the assays specificity. The specificity of the bi- steps. It has proved to be highly sensitive and specific for detection osensor toward E. coli was assessed by evaluation samples contain- of E. coli, although its application can be extended to other big an- ing E.coli and Salmonella typhimurium, which is a similar family alytes. Finally, the possibility to do smartphone-based detection pathogen. Dilutions of E. coli, Salmonella and a mixture of both empowers the developed biosensor as unique PoC tests with real bacteria where spiked in PBS. The assay was carried out following applicability in resource limited settings. the previous procedure. As shown in Figure 5, the normalized flu- orescence (FL) intensity increased with higher concentrations of E. coli (0 to 105 CFU/mL). Conversely, the normalized FL intensity ASSOCIATED CONTENT obtained for a sample with 105 CFU/mL of Salmonella was similar to the one obtained for 0 CFU of E. coli. This result prove that the Supporting Information developed biosensor was only specific for E. coli due to the AuNPs UV-vis spectra of AuNPs and Ab-AuNPs; Effect of concentration functionalization with highly specific antibodies against E. coli of Ab-AuNPs on the photoluminescence restoration; Kinetic be- O157:H7. This was further corroborated by evaluating solutions havior studies of AuNCs/Ab-AuNPs in the presence of E. coli. containing a mixture of 105 CFU/mL Salmonella + 102 CFU/mL E. coli and 105 CFU/mL E. coli + 102 CFU/mL Salmonella. In the first AUTHOR INFORMATION case, the obtained signal was similar to the blank (0 CFU/mL E. coli) and in the second case the signal was similar to the one ob- * Corresponding author tained for 105 CFU/mL E. coli. Therefore, the results confirmed E-mail address: arben.merkoci@icn2.cat that the platform has a high specificity response and when applied in real samples, the achieved signal could be linked only to E. coli Note: detection. The authors declare no competing financial interest. ACKNOWLEDGMENTS The ICN2 is funded by the CERCA Programme / Generalitat de Catalunya. The ICN2 is supported by the Severo Ochoa program of the Spanish Ministry of Economy, Industry and Competitiveness (MINECO, grant No. SEV-2017-0706). Financial support was obtained under MINECO project MAT2017-87202-P. N.A. acknowledges Food and Drug Laboratory Research Center, Food and Drug Organization of Iran and Ministry of Science, Research and Technology of Iran for scholarship during this research time. A.S.T. acknowledges Universitat Autònoma de Barcelona for the 7 possibility of carrying out this work inside the framework of 16.Banerjee, T.; Sulthana, S.; Shelby, T.; Heckert, B.; Jewell, J.; its PhD Programme in Biotechnology. Woody, K.; Karimnia, V.; McAfee, J.; Santra, S., Multiparametric magneto-fluorescent nanosensors for the ultrasensitive detection of REFERENCES Escherichia coli O157: H7. ACS Infect. Dis. 2016, 2 (10), 667-673. 17.Zarei, M., Infectious pathogens meet point-of-care diagnostics. 1.Painter, J. A.; Hoekstra, R. M.; Ayers, T.; Tauxe, R. V.; Braden, Biosens. Bioelectron. 2018, 106, 193-203. C. R.; Angulo, F. J.; Griffin, P. M., Attribution of foodborne 18.Nasseri, B.; Soleimani, N.; Rabiee, N.; Kalbasi, A.; Karimi, M.; illnesses, hospitalizations, and deaths to food commodities by using Hamblin, M. R., Point-of-care microfluidic devices for pathogen outbreak data, United States, 1998–2008. Emerging Infect. Dis. detection. Biosens. and Bioelectron. 2018, 117, 112-128. 2013, 19 (3), 407. 19.Cheeveewattanagul, N.; Morales‐Narváez, E.; Hassan, A. R. H.; 2.Scallan, E.; Griffin, P. M.; Angulo, F. J.; Tauxe, R. V.; Hoekstra, Bergua, J. F.; Surareungchai, W.; Somasundrum, M.; Merkoçi, A., R. M., Foodborne illness acquired in the United States— Straightforward immunosensing platform based on graphene unspecified agents. Emerging Infect. Dis. 2011, 17 (1), 16. oxide‐decorated nanopaper: a highly sensitive and fast biosensing 3.Opota, O.; Croxatto, A.; Prod'hom, G.; Greub, G., Blood culture- approach. Adv. Funct. Mater. 2017, 27 (38), 1702741. based diagnosis of bacteraemia: state of the art. Clin. Microbiol. 20.Morales-Narváez, E.; Naghdi, T.; Zor, E.; Merkoçi, A., Infect. 2015, 21 (4), 313-322. Photoluminescent lateral-flow immunoassay revealed by graphene 4.Souza, R. B.; Trevisol, D. J.; Schuelter-Trevisol, F., Bacterial oxide: highly sensitive paper-based pathogen detection. Anal. sensitivity to fosfomycin in pregnant women with urinary infection. chem. 2015, 87 (16), 8573-8577. Braz. J. Infect. Dis. 2015, 19 (3), 319-323. 21.Qin, H.; Ma, D.; Du, J., Distance dependent fluorescence 5.Zourob, M.; Elwary, S.; Turner, A. P., Principles of bacterial quenching and enhancement of gold nanoclusters by gold detection: biosensors, recognition receptors and microsystems. nanoparticles. Spectrochim. Acta A 2018, 189, 161-166. Springer Science & Business Media: 2008. 22.Chang, H.-C.; Ho, J.-a. A., Gold nanocluster-assisted 6.Zhao, Z.; Yan, R.; Yi, X.; Li, J.; Rao, J.; Guo, Z.; Yang, Y.; Li, fluorescent detection for hydrogen peroxide and cholesterol based W.; Li, Y.-Q.; Chen, C., Bacteria-activated theranostic nanoprobes on the inner filter effect of gold nanoparticles. Anal. chem. 2015, against methicillin-resistant Staphylococcus aureus infection. ACS 87 (20), 10362-10367. nano 2017, 11 (5), 4428-4438. 23.Liu, Y.; Li, H.; Guo, B.; Wei, L.; Chen, B.; Zhang, Y., Gold 7.Russo, L.; Leva Bueno, J.; Bergua, J. F.; Costantini, M.; nanoclusters as switch-off fluorescent probe for detection of uric Giannetto, M.; Puntes, V.; de la Escosura-Muñiz, A.; Merkoçi, A., acid based on the inner filter effect of hydrogen peroxide-mediated Low-cost strategy for the development of a rapid electrochemical enlargement of gold nanoparticles. Biosens. Bioelectron. 2017, 91, assay for bacteria detection based on AuAg nanoshells. ACS 734-740. Omega 2018, 3 (12), 18849-18856. 24.Xie, J.; Zheng, Y.; Ying, J. Y., Protein-directed synthesis of 8.Yetisen, A. K.; Akram, M. S.; Lowe, C. R., based microfluidic highly fluorescent gold nanoclusters. J. Am. Chem. Soc. 2009, 131 point-of-care diagnostic devices. Lab Chip 2013, 13 (12), 2210- (3), 888-889. 2251. 25.Bastús, N. G.; Comenge, J.; Puntes, V., Kinetically controlled 9.Huang, H.; Zhao, G.; Dou, W., Portable and quantitative point- seeded growth synthesis of citrate-stabilized gold nanoparticles of of-care monitoring of Escherichia coli O157: H7 using a personal up to 200 nm: size focusing versus Ostwald ripening. Langmuir glucose meter based on immunochromatographic assay. Biosens. 2011, 27 (17), 11098-11105. Bioelectron. 2018, 107, 266-271. 26.Haiss, W.; Thanh, N. T.; Aveyard, J.; Fernig, D. G., 10.Luo, K.; Hu, L.; Guo, Q.; Wu, C.; Wu, S.; Liu, D.; Xiong, Y.; Determination of size and concentration of gold nanoparticles from Lai, W., Comparison of 4 label-based immunochromatographic UV− Vis spectra. Anal. chem. 2007, 79 (11), 4215-4221. assays for the detection of Escherichia coli O157: H7 in milk. J. 27.Álvarez-Diduk, R.; Orozco, J.; Merkoçi, A., Paper strip- Dairy Sci. 2017, 100 (7), 5176-5187. embedded graphene quantum dots: a screening device with a 11.Wu, W.; Zhao, S.; Mao, Y.; Fang, Z.; Lu, X.; Zeng, L., A smartphone readout. Sci. Rep. 2017, 7 (1), 976. sensitive lateral flow biosensor for Escherichia coli O157: H7 28.Oyama, Y.; Osaki, T.; Kamiya, K.; Kawano, R.; Honjoh, T.; detection based on aptamer mediated strand displacement Shibata, H.; Ide, T.; Takeuchi, S., A glass fiber sheet-based amplification. Anal. Chim. Acta 2015, 861, 62-68. electroosmotic lateral flow immunoassay for point-of-care testing. 12.Li, Y.; Xiong, Y.; Fang, L.; Jiang, L.; Huang, H.; Deng, J.; Lab Chip 2012, 12 (24), 5155-5159. Liang, W.; Zheng, J., An electrochemical strategy using 29.O’Farrell, B., Evolution in lateral flow–based immunoassay multifunctional nanoconjugates for efficient simultaneous systems. In Lateral flow immunoassay, Springer: 2009; pp 1-33. detection of Escherichia coli O157: H7 and Vibrio cholerae O1. 30.Morales‐Narváez, E.; Hassan, A. R.; Merkoçi, A., Graphene Theranostics 2017, 7 (4), 935. Oxide as a Pathogen‐Revealing Agent: Sensing with a Digital‐Like 13.Chen, J.; Alcaine, S. D.; Jackson, A. A.; Rotello, V. M.; Nugen, Response. Angew. Chem. 2013, 52 (51), 13779-13783. S. R., Development of engineered bacteriophages for Escherichia 31.Chen, W.; Cao, F.; Zheng, W.; Tian, Y.; Xianyu, Y.; Xu, P.; coli detection and high-throughput antibiotic resistance Zhang, W.; Wang, Z.; Deng, K.; Jiang, X., Detection of the determination. ACS Sens. 2017, 2 (4), 484-489. nanomolar level of total Cr [(III) and (VI)] by functionalized gold 14.Lagally, E.; Scherer, J.; Blazej, R.; Toriello, N.; Diep, B.; nanoparticles and a smartphone with the assistance of theoretical Ramchandani, M.; Sensabaugh, G.; Riley, L.; Mathies, R., calculation models. Nanoscale 2015, 7 (5), 2042-2049. Integrated portable genetic analysis microsystem for 32.Ambrosi, A.; Castañeda, M. T.; Killard, A. J.; Smyth, M. R.; pathogen/infectious disease detection. Anal. chem. 2004, 76 (11), Alegret, S.; Merkoçi, A., Double-codified gold nanolabels for 3162-3170. enhanced immunoanalysis. Anal. chem. 2007, 79 (14), 5232-5240. 15.Sugino, H.; Ozaki, K.; Shirasaki, Y.; Arakawa, T.; Shoji, S.; Funatsu, T., On-chip microfluidic sorting with fluorescence spectrum detection and multiway separation. Lab Chip 2009, 9 (9), 1254-1260. 8 Graphical Table of Contents 9