A LEGO®-inspired pipette-free approach for a fully-integrated molecular diagnostic kit isothermally operated at near body temperature for the detection of antibacterial resistance

Abstract

Herein, we introduce a LEGO®-inspired molecular diagnostic microdevice fully integrated with DNA extraction, loop-mediated isothermal amplification (LAMP), and colorimetric detection functionalities for rapid detection of antibacterial resistance in a pipette-free manner. The microdevice system is composed of a ready-to-use microdevice containing all necessary reagents and stamps that offer sample-to-answer diagnosis in a pipette-free manner. In particular, antimicrobial resistance was analyzed through LAMP at a significantly reduced temperature of approximately 40 °C, combined with DNA extraction and detection, which were performed at room temperature (RT). Semi-solid colorimetric detection of LAMP amplicons was introduced using polyethylenimine (PEI)-coated paper and AuNP solutions to enable stable storage and color discrimination between negative and positive samples. To use the introduced microdevice system, a commonly encountered antibiotic-resistant bacterium – Enterococcus faecium – was successfully identified in 93 min, and a limit of detection as low as 10¹ CFU mL⁻¹ was obtained by naked-eye colorimetric detection of LAMP amplicons. The introduced device offers point-of-care applications that accurately and rapidly detect antimicrobial-resistant bacteria. The simple operation and colorimetric detection strategy can help implement point-of-care testing (POCT) in resource-limited areas.

---

Introduction

Nowadays, antimicrobial resistance is a serious public health problem causing serious diseases that are more challenging and costly to treat.^1,2 According to the World Health Organization (WHO), there are some important examples of antimicrobial-resistant strains of bacteria such as methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant Enterococcus (VRE), multi-drug-resistant Mycobacterium tuberculosis (MDR-TB), which are known to result in severe illnesses and increase mortality rates, threatening public health.³ Although polymerase chain rection (PCR) is considered a gold standard of nucleic acid analysis for identifying these bacteria in hospitals or central labs, isothermal nucleic acid amplification tests have also become more popular and emerged as important tools for clinical applications due to their isothermal temperature control, high sensitivity, and low cost, dispensing with the use of a thermal cycler, which is inevitable for PCR.^4–6 Among isothermal approaches, loop-mediated isothermal amplification (LAMP) is simple to perform, enabling rapid end-point detection with up to 100-fold higher sensitivity than PCR. For this reason, LAMP is widely used for on-site diagnostic purposes for the detection of bacteria and viruses, especially for detecting antibiotic-resistant pathogenic bacteria.^7,8 For these reasons, LAMP has been an ideal technique for developing a rapid and sensitive point-of-care testing (POCT) platform for the detection of antibiotic-resistant bacteria in resource-limited areas.^9,10

Various studies have focused on integrating LAMP with microdevices for antimicrobial resistance diagnostics, which offer portability, sensitivity, specificity, and simplified instrumentation.^11–13 In detail, microfluidic platforms enable the integration of DNA extraction, amplification, and detection processes in a single system, which is ideal for POCT applications.^14,15 For example, Papadakis et al. introduced a portable real-time colorimetric LAMP device to detect COVID-19 and cancer mutations within 30 min from patients’ nasopharyngeal samples.¹⁶ Sun et al. fabricated a miniaturized device for SARS-CoV-2 diagnostics integrating extraction and amplification modules, and the total analysis time was about 28 min.¹⁷ In another study, Trinh et al. presented an integrated 3D foldable microdevice (DNA extraction, amplification, and detection) for vancomycin-resistant Enterococcus identification in 60 min.¹⁸ Most of the LAMP previously reported have been performed at a single temperature of 60–65 °C, simplifying the temperature control and operation compared to PCR.^4,19 However, they still require a high temperature which hinders their POCT application for on-site diagnosis in resource-limited areas.^20–22 To date, only a few studies have been conducted to perform LAMP at low temperatures, but this concept has not yet been applied to fabricating a microdevice platform integrated with LAMP. For example, Cai et al. developed a LAMP reaction in which the incubation temperature was around 40 °C using special phosphorothioated primers and additional factors (urea and single-stranded DNA binding protein) to decrease the melting temperature.²³ Although LAMP enabled highly sensitive amplification, the introduced technique required expensive primer modifications. In a recent study, Nam et al. reported a low temperature LAMP for the detection of miR-21 by optimizing the concentrations of MgSO₄ and deoxyribonucleoside triphosphates (dNTPs) and the lengths of DNA probes.²⁴ These results suggest that the low temperature LAMP could enable a great advancement of POCT diagnostic platforms for rapid bacterial and viral detection, especially for antibiotic-resistant pathogens in resource-limited areas.

On the other hand, to identify LAMP amplicons, colorimetric detection methods based on naked-eye observation have been widely developed for end-point detection which can also be used in resource-limited areas.²⁵ Nanomaterials (gold nanoparticles (AuNPs) and silver nanoparticles (AgNPs)) are usually employed to visualize the presence of amplicons by color change.^26,27 Among nanomaterials, AuNPs have been widely used for visualization after amplification, especially LAMP amplicon detection, owing to their color shift depending on their aggregation or dispersion state.^28,29 For example, Djisalov et al. used AuNPs for LAMP amplicon detection based on the salt-induced aggregation principle with stabilized and aggregated AuNPs appearing red (positive samples) and violet (negative samples), respectively.³⁰ Moreover, naked-eye detection of LAMP amplicons was performed using DNA-functionalized AuNPs based on target-induced stabilization, which resulted in a red color for positive samples, contrasting with the occurrence of gray aggregates in negative samples.³¹ However, these approaches are not only costly and complicated to modify, but also hard to incorporate into POCT because of the multiple pipetting steps during the detection process.³² Therefore, there is a need for a new colorimetric platform that operates in a pipette-free manner and can be applied to POCT in low-resource settings by realizing inexpensive, portable, rapid, and instrument-free analyses.^33–35

In this study, we introduced a LEGO®-inspired molecular diagnostic microdevice integrating DNA extraction, amplification, and detection for rapid antibacterial resistance diagnosis at a low temperature within 93 min. The device consists of three main parts: (i) the bottom part – a case holding two LEGO® brick-like structures and a microdevice; (ii) the middle part – an enclosed microdevice with four pairs of chambers filled with FTA washing buffer, deionized (DI) water, LAMP reagents, and AuNP solution; (iii) the top part – two LEGO® brick-like structures where each LEGO® brick is adhered with a 3 mm FTA card disc for DNA extraction and polyethylenimine (PEI)-coated filter paper for DNA detection. The introduced LAMP-based microdevice enabled DNA amplification at a low and constant temperature of approximately 40 °C by optimizing LAMP components. Guiding the two LEGO® bricks along the case allows for an easy and smooth uptake and transfer of DNA samples from chamber to chamber without pipetting steps. For the demonstration of this concept, two common antibiotic-resistant pathogenic bacteria including Enterococcus faecium (E. faecium) and Acinetobacter baumannii (A. baumannii) were analyzed as practical samples using the integrated microdevice system.

Principles of colorimetric detection

Fig. 1 shows the overall schematic of the introduced colorimetric method for naked-eye visualization of LAMP amplicons using a combination of PEI-coated filter paper and citratecapped AuNP solution. On one hand, PEI – a high cationic charge polymer – has been widely used for capturing DNA in various forms such as genomic DNA (gDNA) and plasmid DNA, or RNA due to its ability to condense nucleic acids.^36,37 Also, PEI has been widely used for surface modification on nanoparticles or immobilization of nanoparticles onto commercial filter paper for colorimetric sensors or catalytic applications.^38–40 In this study, a positively-charged PEI was coated onto Whatman filter paper by charge interaction and subsequently used as a matrix for colorimetric detection of LAMP amplicons in combination with AuNPs. When the negatively-charged DNAs are soaked into the PEI-coated papers, DNAs directly interact with the PEI layer, forming PEI–DNA complexes. The presence of the PEI–DNA complex can be visually determined by adding citrate-capped AuNP solution. That is, in the presence of LAMP amplicons, the PEI layer is blocked by DNAs, leaving AuNPs well dispersed in solution and maintaining their red color. However, in the absence of LAMP amplicons, the positively-charged amine groups of PEI layer captures the negatively-charged, citrate-capped AuNPs, resulting in the aggregation of the AuNPs and changing the color to gray.

Fig. 1 Schematic showing the mechanism of colorimetric detection of LAMP amplicons using PEI-coated paper discs and citrate-capped AuNP solution.

Materials and methods

Materials

FTA classic cards, FTA washing buffer, and PEI were obtained from Sigma-Aldrich (St Louis, MO, USA). Whatman Grade 2 qualitative filter paper was obtained from GE Healthcare Life Sciences (Chicago, IL, USA). Bst 2.0 polymerase, dNTP mixes (each 10 mM), 100 mM magnesium sulfate (MgSO₄), and 10× isothermal amplification buffer for LAMP reaction were purchased from New England Biolabs (Ipswich, MA, USA). For performing gel electrophoresis, DNA ladders and loading dye were purchased from Genes Laboratories (Seongnam, Korea) and Dyne Bio (Seongnam, Korea), respectively. Agarose powder and sealing films were obtained from BioFact (Daejeon, Korea). Poly(methyl methacrylate) (PMMA) was purchased from Goodfellow (Coraopolis, USA). Primers for amplifying E. faecium and A. baumannii were synthesized from Cosmogenetech (Seoul, Korea).

Sample preparation and DNA extraction

In this study, E. faecium bacteria were grown in both liquid culture media and agar plates. E. faecium was incubated with 5 mL of Luria–Bertani broth (10 g of tryptone, 5 g of yeast extract, and 5 g of NaCl in 1 L of distilled water (DI)) at 37 °C and agitated at 200 rpm for 16 h in an incubator. Viable counts were determined by performing serial dilutions and plating these dilutions on solid LB agar media for E. faecium and incubating at 37 °C for 16 h. For DNA extraction, a 3 mm FTA card was prepared for performing DNA extraction as described in our previous study.⁴¹ In brief, 3 μL of bacterial solution was applied on the FTA card using a pipette and incubated at RT for 10 min for DNA extraction. Then, 20 μL of the FTA purification reagents and DI water were sequentially applied to wash and rinse the FTA card for downstream application, respectively.

LAMP assays

In this study, we adopted a low temperature LAMP to make the technique more convenient and energy-efficient. To find the lowest possible temperature for the modified LAMP assay, we conducted amplification reactions at various temperatures ranging from 35 to 65 °C. Primer sets for amplifying E. faecium and A. baumannii including F3 and B3 (outer primers), FIP and BIP (inner primers), and LB and LF (loop primers) were designed using the PrimerExplorer V5 program, and primer sequences are shown in Table S1. 0.35 mM dNTPs (0.7 μL), 1× isothermal amplification buffer (20 mM Tris–HCl, 10 mM (NH₄)₂SO₄, 50 mM KCl, 2 mM MgSO₄, and 0.1% Tween 20; pH 8.8 at 25 °C) (2.0 μL), 0.16 U μL⁻¹ of Bst 2.0 DNA polymerase (0.4 μL), a primer mixture (2.0 μL) including 0.2 μM F3 and B3, 1.6 μM FIP and BIP, 0.8 μM LF and LB, and DI water were included in each LAMP mixture (20 μL). Besides, the typical LAMP was also performed at 65 °C for a mixture (20 μL) that contained 1× isothermal amplification buffer (20 mM Tris–HCl, 10 mM (NH₄)₂SO₄, 50 mM KCl, 2 mM MgSO₄, and 0.1% Tween 20; pH 8.8 at 25 °C) (2.5 μL), 1.4 mM dNTPs, 6 mM MgSO₄ (1.5 μL), 0.16 U μL⁻¹ of Bst 2.0 DNA polymerase (0.4 μL), the primer mixture (2.0 μL) including 0.2 μM F3 and B3, 1.6 μM FIP and BIP, 0.8 μM LF and LB, and DI water. For a positive control, E. faecium and A. baumannii were applied onto the FTA card and dipped into the LAMP mixture for amplification of the target gene. A negative control, not containing any target gene, was tested simultaneously to enhance the reliability of the test. Positive and negative controls were incubated at various temperatures ranging from 35 to 65 °C for 60 min, and the results were evaluated by gel electrophoresis and the colorimetric detection method introduced in this study. For on-chip LAMP, a commercial hot plate was used as a heater for controlling the amplification temperature. Also, LAMP amplicons that were used for optimizing the colorimetric detection were kept at 4 °C or −20 °C for short- or long-term storage, respectively.

Specificity and sensitivity tests

E. faecium was amplified with E. faecium primer sets as well as A. baumannii primer sets to perform a specificity test. In the same manner, A. baumannii was tested with A. baumannii primer sets as well as E. faecium primer sets. The amplification results were analyzed by gel electrophoresis and the colorimetric detection method introduced in this study. For the sensitivity test, E. faecium was 10-fold serially diluted down to 10⁰ CFU mL⁻¹ and applied to the FTA card for DNA purification. Subsequently, a LAMP reaction, gel electrophoresis, and colorimetric detection were conducted to investigate the limit of detection (LOD).

Colorimetric detection

For the preparation of PEI-coated paper discs, the PEI solution was diluted with DI water to make various concentrations (0.001, 0.005, 0.01, 0.05, and 0.1% (v/v)) of the PEI solution. The nitrocellulose, regenerated cellulose, glass microfiber, nylon, and Whatman filter papers were cut to have a circular shape (2 mm in diameter) and dipped into the as-prepared PEI solution overnight at room temperature (RT). Afterward, the PEI-coated paper disc was thoroughly washed with DI water to completely remove unreacted PEI and then dried further at RT before use. AuNPs were synthesized following the Turkevich approach that was introduced by Kimling et al.⁴² Briefly, an aqueous solution of gold chloride trihydrate (0.167 mM) was heated with magnetic stirring, followed by the addition of sodium citrate (2.2 mM). After the color of the solution turned red, it was cooled to RT. For the colorimetric detection of the LAMP amplicons, the LAMP amplicons were first reacted with the PEI-coated paper for 10 min, and the PEI-coated paper disc on which DNA is attached was further transferred to a chamber containing citrate-capped AuNPs and reacted for 3 min to visualize the color change. The color intensity of the AuNP solutions was analyzed using ImageJ 1.52a software. Images were captured using a smartphone and processed using ImageJ software for analysis. The captured images were first converted to an 8-bit grayscale format. Then, a fixed-sized circular region of interest (ROI) was manually selected on each reaction chamber to ensure consistent measurements across samples, and the mean gray values were measured. In this way, min/mean/max gray intensity values were quantified and mean gray intensity values were employed to confirm the difference between the positive and negative samples.

Microdevice design and fabrication

Fig. 2 shows the overall structure of the LEGO®-inspired molecular diagnostic microdevice composed of three parts including a cassette, a microdevice, and LEGO® bricks. The detailed fabrication process of each component, such as the laser engraved area, microdevice, LEGO® bricks, and cassette is precisely illustrated in Fig. S1. In brief, AutoCAD was used to design the sealing film, microdevice, and LEGO® bricks. Then a laser cutting machine and a computer numerical control (CNC) machine were used for engraving these patterns accordingly. LEGO® bricks 1 and 2 have two solid pillars where the top of a pillar is adhered to an FTA card or PEI-coated paper for performing DNA extraction or detection, respectively. Use of the LEGO® bricks eliminated the necessity for pipettes in transferring DNAs from chamber to chamber for more convenient operation. On the microdevice, microchambers were engraved in parallel where FTA washing buffer, DI water, LAMP reagents, and AuNP solutions were pre-loaded for use during DNA analysis (extraction, amplification, and detection). Afterward, all chambers were completely enclosed using a PCR sealing film which was partially engraved using a laser cutter. The engraved spots were easily ruptured by a gentle push, allowing effortless insertion of the stamps into the chambers without the need to peel off the entire sealing film, minimizing cross-contamination. Moreover, neodymium magnets were inserted into the middle portion of LEGO® brick 1 and the microdevice near the LAMP reagent chamber to prevent evaporation during the LAMP reaction by tightly holding the brick and the LAMP chamber. Moreover, the assembled device served as a cassette for retaining a sealed microdevice and two LEGO® bricks.

Fig. 2 Illustration showing the LEGO®-inspired molecular diagnostic microdevice fabricated for the investigation of antibacterial resistance.

Microdevice operation

As shown in Fig. 3, the LEGO®-inspired molecular diagnostic microdevice was employed to rapidly analyze antibacterial resistance from extraction to detection in a pipette-free manner via a series of stamping motions. Besides, a detailed cross-sectional view of the LEGO®-inspired molecular diagnostic microdevice that contains all the necessary components and reagents for DNA analysis in the detection of antibacterial resistance is also presented. For DNA extraction, 3 μL of bacterial solution was first applied on the FTA card adhered to LEGO® brick 1. Then, LEGO® brick 1 was gently pushed against the engraved sealing film to make punctures to submerge the bacterial DNA-containing FTA cards into the washing buffer in the chambers. LEGO® brick 1 was then introduced into the DI water chambers via the same procedure to remove the remaining washing buffer on the FTA cards. DNA extraction using an FTA card took approximately 20 min. For DNA amplification, LEGO® brick 1 containing the extracted DNA from the FTA card was introduced into the LAMP chamber for conducting LAMP using a commercial hot plate. After the LAMP, the LEGO® brick 1 was removed from the LAMP chamber, and LEGO® brick 2 was introduced into the LAMP chamber to absorb and take up LAMP amplicons for the subsequent detection process. Finally, LEGO® brick 2 was transferred into the AuNP chamber to visualize the color change of the AuNP solution for realizing naked eye-based detection of LAMP amplicons.

Fig. 3 Schematic workflow representing the step-by-step procedures of a naked-eye visual detection method using a LEGO®-inspired molecular diagnostic microdevice.

Statistical analysis

All statistical analyses and data visualization were conducted using GraphPad Prism 9.0 (GraphPad Software Inc., San Diego, CA, USA) or Excel software. Comparisons between groups were performed using unpaired two-tailed t-tests. Statistical significance was defined as p < 0.05. Significance levels were indicated as follows: ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, and ns (not significant) for p ≥ 0.05. Error bars were added after the experiments were performed in triplicate and averaged.

Results and discussion

Optimization of the colorimetric detection

Fig. 4 shows the results of the optimization for the introduced colorimetric detection method using PEI-coated papers. At first, to select the best performing paper matrix, various types of membranes such as nitrocellulose, regenerated cellulose, glass microfiber, nylon, and Whatman filter paper were tested by pre-coating with PEI (0.05%). Among the five filter papers tested, the PEI-coated Whatman filter paper displayed the highest contrast in the color display between the negative and positive samples. However, there was almost no difference in the target band intensities when LAMP reactions were performed under normal (65 °C) and low temperature (40 °C) conditions, as demonstrated in Fig. S2. Based on the UV-vis spectrometry results, after the removal of PEI-coated paper from the AuNP solution, the negative sample displayed a gray color caused by the loss of the plasmonic peak of AuNPs around 520 nm. This phenomenon could be explained by the aggregation of AuNPs due to the electrostatic interaction between the positively-charged PEI and negatively-charged AuNPs. However, the positive sample maintained the red color of AuNPs because the negatively-charged LAMP amplicons competed with the negatively-charged AuNPs and eventually reacted with the positively-charged PEI, leaving negatively-charged AuNPs unreacted. Based on these results, we selected the Whatman paper as a color-inducing matrix to detect LAMP amplicons. Furthermore, the PEI concentration, which was used for coating a filter paper, also plays a critical role in enhancing the colorimetry performance. Fig. 4b shows the effect of PEI concentration on the aggregation of AuNPs in the negative and positive LAMP samples for amplicon detection. Among the five different PEI concentrations tested (0.001, 0.005, 0.01, 0.05, and 0.1%), the negative and positive samples were only clearly distinguished when the concentrations of PEI were 0.05 and 0.1%. Therefore, we selected 0.05% for further testing.

Fig. 4 Results showing various conditions for optimizing the colorimetric-based detection method using PEI-coated paper. (a) Various types of papers were coated with PEI solution among which Whatman filter paper was selected as the optimum paper based on the displayed color and UV-vis spectrometry. (b) Various concentrations of AuNPs were tested and 0.05% was selected as the optimum concentration based on the displayed color and UV-vis spectrometry. NC: negative control. PC: positive control. The experiments were repeated three times.

Optimization of the efficiency of LAMP conducted at near body temperature

The low temperature at which the LAMP was conducted in this study was adapted from a previous study with a slight modification.²⁴ According to the previous reports, the concentrations of dNTPs and MgSO₄ played important roles in LAMP when performed at 40 °C, which was confirmed as shown in Fig. S3. According to the gel image, the LAMP performed at 40 °C was successfully carried out with lower concentrations of MgSO₄ (2 mM) and dNTPs (0.35 mM), which were chosen for further experiments. The results were further confirmed by the introduced colorimetric technique enabling discrimination between the positive and negative samples. Although the exact mechanism is not clear, the previous study suggested that the LAMP performed at low temperature was enabled by lowering the Mg²⁺ concentration to prevent strong hybridization and avoid non-specific binding of DNAs.²⁴ Because Mg²⁺ ions can directly interact with the phosphate groups of DNA to reduce the electrostatic repulsion between the strands, adjusting their concentration can greatly affect the melting temperature of dsDNAs.⁴³ Besides, since dNTPs chelate with Mg²⁺ and reduce free Mg²⁺ in the reaction solution, the concentration of dNTPs was also reduced accordingly.⁴⁴ Moreover, Fig. 5 shows the results of the optimization test on the amplification temperature and time for the LAMP. As shown in Fig. 5a, on-tube LAMPs were performed at 35, 40, 50, and 60 °C for 60 min. As shown in the gel image, the LAMP reaction successfully amplified the genomic DNA of E. faecium bacteria in the temperature ranging from 40 to 60 °C. Also, colorimetric detection was performed to analyze the LAMP amplicons. The color intensities were almost identical regardless of the temperatures applied. Considering power consumption, 40 °C was considered to be the optimal temperature. As shown in Fig. 5b, LAMP was performed for 40, 50, and 60 min at 40 °C. Although the LAMPs were successfully performed for both 50 and 60 min, detected both by gel electrophoresis and the colorimetric technique, 60 min displayed more distinct color contrast as shown in the absorbance measurement. Based on these results, a reaction temperature of 40 °C and a reaction time of 60 min were finally selected for further experiments.

Fig. 5 (a) Results showing the optimized conditions when conducting LAMP at various temperatures and analyzed by gel electrophoresis, colorimetry, and UV-vis spectrometry. (b) Results showing the optimized reaction time by conducting LAMP at 40 °C and analyzed by gel electrophoresis, colorimetry, and UV-vis spectrometry. NC: negative control. PC: positive control. The experiments were repeated three times.

Specificity and sensitivity tests

Specificity test. To test the specificity of the LAMP when performed at low temperature, targeted DNA and non-targeted DNA were amplified with matched and non-matched primer sets. Fig. 6a shows the results of the LAMP and colorimetric detection when amplified with E. faecium primer sets. As shown in the gel electrophoresis image, E. faecium primer sets selectively amplified the target gene from E. faecium, whereas the target gene from A. baumannii was not amplified. Thus, a ladder-like band appeared exclusively when E. faecium DNA was present. Alternatively, with E. faecium primer sets, the DNA from A. baumannii was not amplified. Similarly, the AuNP solution remained red in color in the presence of E. faecium DNA, while the color of the AuNPs turned gray in the absence of E. faecium DNA as well as in a negative control sample. Moreover, the color of the paper discs was further analyzed using ImageJ software and quantified as shown in the mean gray intensity graph. On the other hand, Fig. 6b shows the results of the LAMP assay and colorimetric detection when amplified with A. baumannii primer sets. A ladder-like band only appeared in the presence of A. baumannii DNA, which indicated the high selectivity of LAMP performed for the detection of A. baumannii. Subsequently, detection using PEI-coated paper discs was also conducted. In the absence of A. baumannii, the AuNP solution turned gray, while the AuNP solution remained red in the presence of A. baumannii. From these results, the specificity of LAMP was confirmed to be highly sufficient for discriminating the target of interest.

Fig. 6 Results showing the specificity of the LAMP and colorimetric detection using PEI-coated paper discs. (a) Results of gel electrophoresis, colorimetric detection, and mean gray intensity graph when amplified with E. faecium primer sets. (b) Results of gel electrophoresis, colorimetric detection, and mean gray intensity graph when amplified with A. baumannii primer sets. NC: negative control. The experiments were repeated three times. Significance levels were indicated as follows: ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, and ns (not significant) for p ≥ 0.05.

Sensitivity test. Fig. 7 shows the results of the sensitivity test of the LAMP performed at 40 °C, which were obtained using both agarose gel electrophoresis and the colorimetric detection technique. As described above, a 10-fold serially diluted E. faecium sample was prepared and used to carry out the sensitivity test as shown in Fig. 7a. Thus, ladder-like bands appeared from 10⁸ to 10¹ CFU mL⁻¹ through agarose gel electrophoresis, which indicated that the successful amplification of DNA was realized via LAMP conducted at near body temperature (40 °C). On the other hand, when the template was 10⁰ CFU mL⁻¹, amplification was not successful due to the low concentration of DNA. This result corresponded well with a typical LAMP conducted at 65 °C.⁴⁵ Furthermore, colorimetric detection was also performed using PEI-coated paper discs and AuNP solutions, which showed LOD results similar to those for agarose gel electrophoresis as shown in Fig. 7b. The AuNP solution remained red in color from 10⁸ to 10¹ CFU mL⁻¹, while both the 10⁰ CFU mL⁻¹ sample and the negative control turned gray. Moreover, these colorimetric detection results were further analyzed using ImageJ software. As shown in the mean gray intensity graph (Fig. 7c), the sensitivity of the PEI-coated paper disc for DNA detection was successfully determined by the color change of AuNP solution from red to grey. The LOD of the LAMP performed at 40 °C was approximately 10¹ CFU mL⁻¹, and the results were successfully confirmed by both agarose gel electrophoresis and the introduced colorimetric detection technique using the combination of AuNP solution and PEI-coated paper discs.

Fig. 7 Results showing the sensitivity of the colorimetric detection using PEI-coated discs. (a) Gel electrophoresis, (b) colorimetric detection, and (c) mean gray intensity graph obtained for the detection of E. faecium. NC: negative control. The experiments were repeated three times. Significance levels were indicated as follows: ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, and ns (not significant) for p ≥ 0.05.

On-chip detection of antibacterial resistance

Fig. 8a shows the overall images of the LEGO®-inspired molecular diagnostic microdevice (right image) and the detailed description for each chamber of the microdevice. In detail, the microdevice was flexibly assembled to create a device like LEGO® blocks by combining five main components of the microdevice including a frame case, two LEGO® brick-like structures, a microdevice, and a cover film (left image). The device operation was minimized by eliminating the use of pipetting and replacing its function with stamps. The two stamps and the frame case could still be reused by replacing them with an FTA card and PEI-coated paper discs. In particular, the on-chip experiment was carried out under low temperature conditions where RT for both DNA extraction and colorimetric detection and 40 °C for DNA amplification via LAMP were employed. As shown in Fig. 8b and c, E. faecium was successfully detected utilizing the introduced microdevice. For on-chip LAMP, the device was placed on a hot plate and incubated at 40 °C for 60 min to perform DNA amplification, while the average temperature measured from the two chambers using an IR camera was 39.85 ± 0.35 °C (n = 8). Based on the gel electrophoresis, a ladder-like band was observed in the presence of target bacteria, which signifies that DNA could be simultaneously purified and amplified on-chip at near body temperature via LAMP (40 °C) (Fig. 8b). A negative control was tested simultaneously to ensure the reliability of the experiment. PEI-coated paper disc-based colorimetric detection was also successfully performed to achieve naked-eye detection at RT (Fig. 8c). The red color of the AuNP solution changed to gray in the absence of LAMP amplicons, whereas the red color was retained in the presence of LAMP amplicons. The color difference between the negative and positive samples was successfully distinguished on the chip. These results proved that the introduced microdevice is not only suitable for rapid bacterial detection in a sample-to-answer manner, but it is also applicable for detecting different types of bacteria from different sample resources.

Fig. 8 (a) Real images showing the overall components of the LEGO®-inspired molecular diagnostic microdevice. (b) Gel electrophoresis results of E. faecium amplified utilizing the LEGO®-inspired molecular diagnostic microdevice. (c) Schematics and photos showing the results of the colorimetric detection observed at the dotted squares inside the chambers. NC: negative control. PC: positive control. The experiments were repeated three times.

Conclusions

In summary, the present work introduced a LEGO®-inspired molecular diagnostic microdevice integrating DNA extraction, LAMP performed at 40 °C, and AuNP-mediated colorimetric detection in a solid form realized on PEI-coated paper for the detection of two common antibiotic-resistant pathogenic bacteria. Low temperature LAMP was combined with FTA card-based DNA extraction and PEI-coated paper disc-based colorimetric detection on a single device, and all steps were realized and operated at near body temperature (40 °C) in 93 min. The introduced integrated microdevice system developed for visualizing LAMP amplicons achieved an LOD as low as 10¹ CFU mL⁻¹ for E. faecium. The simple operation and effectiveness of the introduced colorimetric LAMP-based technique can help protect hospitalized patients from nosocomial infections. While the current platform can help clinicians make better decisions on antibiotic prescriptions mitigating the spread of multidrug-resistant bacteria by preventing over-prescription of antibiotics, the current platform can be further optimized to be more eligible for clinical applications utilizing real-world samples such as blood, saliva, and urine samples. In addition, the overall reaction time for low temperature LAMP can be shortened by optimizing the amounts or types of reagents and enzymes used. Regarding the long-term storage issue, lyophilized LAMP reagents could be applied which can be stored at 4 °C for over a month.^46,47 In this way, the entire microdevice can be stored at 4 °C for over a month, eliminating the need for cold storage. The LEGO®-inspired molecular diagnostic microdevice combined with a machine learning approach may eventually lead to the remote interpretation of the results which is more suitable in the era of a pandemic or for the future of personalized healthcare technology.

Author contributions

Kieu The Loan Trinh: conceptualization, methodology, formal analysis, investigation, resources, and writing – original draft. So Yeon Park: methodology, software, data curation, resources, and writing – original draft. Hyunji Lee: formal analysis and investigation. Nae Yoon Lee: conceptualization, methodology, resources, data curation, project administration, funding acquisition, investigation, writing – review & editing, and supervision.

Conflicts of interest

There are no conflicts to declare.

Data availability

All data generated or analysed during this study are included in this published article. See DOI: https://doi.org/10.1039/d5an00628g.

Acknowledgements

This research was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (RS-2023-00208684) and also by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (RS-2021-NR060117).

References

1. A. Frei, A. D. Verderosa, A. G. Elliott, J. Zuegg and M. A. Blaskovich, Nat. Rev. Chem., 2023, 7, 202–224.
2. M. Xie, M. Gao, Y. Yun, M. Malmsten, V. M. Rotello, R. Zboril, O. Akhavan, A. Kraskouski, J. Amalraj, X. Cai, J. Lu, H. Zheng and R. Li, Angew. Chem., Int. Ed., 2023, 62, e202217345.
3. I. Frost, H. Sati, P. Garcia-Vello, M. Hasso-Agopsowicz, C. Lienhardt, V. Gigante and P. Beyer, Lancet Microbe, 2023, 4, e113–e125.
4. N. Atceken, M. M. Alseed, S. R. Dabbagh, A. K. Yetisen and S. Tasoglu, Adv. Eng. Mater., 2023, 25, 2201174.
5. K. Mao, H. Zhang, F. Ran, H. Cao, R. Feng, W. Du, X. Li and Z. Yang, J. Hazard. Mater., 2024, 462, 132793.
6. M. De Felice, M. De Falco, A. Serra, V. Frisulli, A. Antonacci, R. Isticato, L. de Stefano and V. Scognamiglio, TrAC, Trends Anal. Chem., 2023, 165, 117135.
7. N. Garg, F. J. Ahmad and S. Kar, Curr. Res. Microb. Sci., 2022, 3, 100120.
8. M. K. Nwe, N. Jangpromma and L. Taemaitree, Sci. Rep., 2024, 14, 5916.
9. K. Mao, X. Min, H. Zhang, K. Zhang, H. Cao, Y. Guo and Z. Yang, J. Controlled Release, 2020, 322, 187–199.
10. D. Yigci, N. Atçeken, A. K. Yetisen and S. Tasoglu, ACS Omega, 2023, 8, 43357–43373.
11. A. Sen, M. Masetty, S. Weerakoon, C. Morris, J. S. Yadav, S. Apewokin, J. Trannguyen, M. Broom and A. Priye, Biosens. Bioelectron., 2024, 257, 116292.
12. T. N. D. Trinh and N. Y. Lee, ACS Sens., 2021, 6, 2902–2910.
13. P. Chen, C. Chen, H. Su, M. Zhou, S. Li, W. Du, X. Feng and B. F. Liu, Talanta, 2021, 224, 121844.
14. N. Wang, J. Zhang, B. Xiao and A. Chen, Talanta, 2024, 267, 125150.
15. H. Mashouf, B. Talebjedi, N. Tasnim, M. Tan, S. Alousi, S. Pakpour and M. Hoorfar, Chem. Eng. Process. – Process Intensif., 2023, 189, 109394.
16. G. Papadakis, A. K. Pantazis, N. Fikas, S. Chatziioannidou, V. Tsiakalou, K. Michaelidou, V. Pogka, M. Megariti, M. Vardaki, K. Giarentis, J. Heaney, E. Nastouli, T. Karamitros, A. Mentis, A. Zafiropoulos, G. Sourvinos, S. Agelaki and E. Gizeli, Sci. Rep., 2022, 12, 3775.
17. A. Sun, P. Vopařilová, X. Liu, B. Kou, T. Řezníček, T. Lednický, S. Ni, J. Kudr, O. Ziyka, Z. Fohlerova, P. Pajer, H. Zhang and P. Neužil, Microsyst. Nanoeng., 2024, 10, 66.
18. T. N. D. Trinh, D. A. Thai and N. Y. Lee, Sens. Actuators, B, 2021, 345, 130362.
19. B. Xiao, R. Zhao, N. Wang, J. Zhang, X. Sun and A. Chen, TrAC, Trends Anal. Chem., 2023, 158, 116836.
20. M. Shang, J. Guo and J. Guo, Sens. Diagn., 2023, 2, 1123–1144.
21. I. H. Mahardika, S. Naorungroj, W. Khamcharoen, S. Kin, N. Rodthongkum, O. Chailapakul and K. Shin, Adv. NanoBiomed Res., 2023, 3, 2300058.
22. H. Xu, A. Xia, D. Wang, Y. Zhang, S. Deng, W. Lu, J. Lou, Q. Zhong, F. Zhang, L. Zhou, W. Zhang, W. Y. Wang, C. Yang, K. Chang, W. Fu, J. Cui, M. Gan, D. Lou and M. Chen, Sci. Adv., 2020, 6, eaaz7445.
23. S. Cai, C. Jung, S. Bhadra and A. D. Ellington, Anal. Chem., 2018, 90, 8290–8294.
24. D. Nam, S. Kim, J. H. Kim, S. Lee, D. Kim, J. Son, D. Kim, B. S. Cha, E. S. Lee and K. S. Park, Biosensors, 2023, 13, 367.
25. Y. J. Moon, S. Y. Lee, U. Kim and S. W. Oh, Food Sci. Biotechnol., 2023, 33, 203–209.
26. F. Huang, Y. Jiang, Q. Wu, C. Zheng, S. Huang, H. Yang, G. Xiang and L. Zheng, Food Chem., 2024, 460, 140573.
27. H. Ye, C. Nowak, Y. Liu, Y. Li, T. Zhang, L. Bleris and Z. Qin, Small, 2022, 18, 2107832.
28. S. Grammatikos, I. Svoliantopoulos and E. Gizeli, Anal. Chem., 2023, 95, 18514–18521.
29. R. Sivakumar, V. P. Dinh and N. Y. Lee, Lab Chip, 2021, 21, 700–709.
30. M. Djisalov, L. Janjušević, V. Léguillier, L. S. Zorić, C. Farre, J. Anba-Mondoloni, J. Vidic and I. Gadjanski, Sci. Rep., 2024, 14, 15539.
31. X. Zhang, Y. Tian, Y. Shi, J. Liu, C. Zhao, C. C. Chang, T. Takarada, M. Maeda and G. Wang, ACS Infect. Dis., 2024, 10, 2668–2678.
32. D. Liu, Y. Wang, X. Li, M. Li, Q. Wu, Y. Song, Z. Zhu and C. Yang, Aggregate, 2022, 3, e184.
33. M. Bao, S. Zhang, C. T. Pas, S. J. Dollery, R. V. Bushnell, F. N. U. Yuqing, R. Liu, G. Lu, G. J. Tobin and K. Du, Lab Chip, 2022, 22, 4849–4859.
34. Y. Kinoshita, T. Tayama, K. Kitamura, M. Salimullah, H. Uchida, M. Suzuki, Y. Husimi and K. Nishigaki, BMC Biotechnol., 2010, 10, 1–16.
35. S. Y. Park, K. T. L. Trinh, Y. J. Song and N. Y. Lee, Biotechnol. J., 2024, 19, 2300521.
36. A. Baker, M. Saltik, H. Lehrmann, I. Killisch, V. Mautner, G. Lamm, G. Christofori and M. Cotten, Gene Ther., 1997, 4, 773–782.
37. H. Sun, X. Zhu, L. Zhang, Y. Zhang and D. Wang, Mater. Sci. Eng., C, 2010, 30, 311–315.
38. S. Wu, G. Li, W. Liu, D. Yu, G. Li, X. Liu, Z. Song, H. Wang and H. Liu, Nano Energy, 2022, 93, 106859.
39. Y. Zhao, L. Liu, D. Shi, X. Shi and M. Shen, Nanoscale Adv., 2019, 1, 342–346.
40. M. L. M. Budlayan, J. N. Patricio, J. P. B. Lagare, L. B. de la Rosa, S. D. Arco, A. C. Alguno, E. S. Austria, J. P. Manigo and R. Y. Capangpangan, Funct. Compos. Struct., 2021, 3, 035007.
41. K. T. L. Trinh, T. N. D. Trinh and N. Y. Lee, Biosens. Bioelectron., 2019, 135, 120–128.
42. J. Kimling, M. Maier, B. Okenve, V. Kotaidis, H. Ballot and A. Plech, J. Phys. Chem. B, 2006, 110, 15700–15707.
43. R. Owczarzy, B. G. Moreria, Y. You, M. A. Behlke and J. A. Walder, Biochemistry, 2008, 47, 5336–5353.
44. K. H. Roux, Cold Spring Harb. Protoc., 2009, 4, 66.
45. S. Y. Park, R. Sivakumar and N. Y. Lee, Biosensors, 2024, 14, 284.
46. N. Jothinarayanan, F. Karlsen and L. E. Roseng, Appl. Biochem. Biotechnol., 2024, 196, 5249–5264.
47. H. W. Chen and W. M. Ching, Heliyon, 2017, 3, e00415.

---

Footnote

† These authors contributed equally to this work.

---