Fast and sensitive multiplexed diagnostic system enabled by real-time solid-phase PCR assay

Multiplexed pathogenic detection using polymerase chain reaction (PCR) is a powerful technique to detect several pathogens in one assay^1,2,3. For multiplexing, two or more primer sets are designed to amplify and detect different targets in the same PCR reaction, saving time and effort in clinical practice. Multiplex detection is also enabled in quantitative real-time PCR (qPCR), in which DNA products are amplified and quantified via cycle threshold (Ct) values, eliminating the post analysis steps and providing instant and quantified results^4,5. However, multiplexed qPCR instruments require the same number of filters as the number of pathogens, hence limiting the multiplexing capacity⁶. In a typical qPCR instrument, less than five emission channels are usually used, limiting the degree of multiplexing to about 5 due to spectral overlap of fluorophores⁷. Biofire FilmArray (BFR) multiplex instrument is a notable PCR system that runs multiplexed assays for a higher number of pathogens (>15) using a single reading system through compartmentalization of each target in a separate space to target each pathogen separately. However, the BFR system does not offer accurate quantification, and it requires a relatively huge volume for massive compartmentalization.

Solid-phase PCR (SP-PCR), wherein several pathogen-specific probes or primers are immobilized on a solid surface as an array, has been suggested to enable multiplex amplification and minimize the signal interference in multiplex detection^8,9. SP-PCR dramatically enhances multiplex capabilities due to a spatially encoding array that enables targeting multiple specific analytes using a single fluorophore^10,11. However, as SP-PCR is performed in a stationary/static condition, interaction and binding between the generated amplicons of the target genes (in the liquid phase) and the attached pathogen-specific probe (on the solid support) is limited¹², which reduces PCR efficiency. Moreover, the signal reading of the SP-PCR process is usually at the end of the assay, and real-time detection is unattainable unless complicated and advanced optical setups are used, such as confocal scanning¹³ and micro-ring optical resonators¹⁴. Thus, there is an unmet need for a high-level multiplexing of quantitative PCR assays.

Here, we present for the first time an integrated microfluidic platform for real-time, quantitative, and multiplexed SP-PCR assay. The platform incorporates a dual-chamber PCR design, wherein pathogen-specific probes are immobilized on an array in an annealing chamber, and PCR reagent passes forward and backward between the annealing chamber and a denaturing chamber. A new mechanical valve is proposed for precise, bidirectional flow of the PCR reagents. The valve offers robust fluidic control through a structurally simple, single-part design. The approach allows for the temporary eviction of one chamber, so that the signals of the amplified products on the array can be recorded after each cycle. Moreover, active liquid transportation greatly enhances the efficiency of SP-PCR. In addition, the two chambers are maintained at constant temperatures, eliminating the need for conventional thermal cycling and significantly reducing transition and ramping time. Furthermore, sample preparation steps, including nucleic acid purification, are seamlessly integrated into the fluidic workflow, enabling a complete sample-to-answer diagnostic process. This platform allows for scalable multiplexing, accurate real-time quantification, high sensitivity (limit of detection of 10 copies/reaction), and rapid turnaround time (<21 min), while maintaining low-cost consumables and minimal instrumentation requirements. The simplicity and precision of the fluidic architecture make this platform a compelling solution for point-of-care and high-throughput molecular diagnostics.

System layout of real-time multiplex PCR assay

The conventional clinic practice for molecular-based pathogen detection is a lengthy, laborious, and multi-step process that requires skillful staff and expensive instrumentation (Fig. 1A). Increasing the number of targets per assay directly increases the sample preparation and instrumentation complexities¹⁵. In quantitative PCR (qPCR), the signal measurement instrument is unable to meet the multiplexing capabilities of reading multiple fluorescent signals due to spectral overlaps of the existing fluorophores (Fig. 1A). Here, we design an automated sample-to-answer SP-PCR system to perform multiplexed assays and enable real-time monitoring of fluorescent signals from the solid-phase array (Fig. 1B and Fig. S1). The total system consists of sample preparation, amplification sections, heating, and the optical system. Immobilizing multiple specific probes for several targets on a solid surface allows for the signal reading of multiple targets using a single fluorescence wavelength.

A, B Workflow illustration of conventional multiplex assay (top) and our sample-to-answer multiplexed PCR on a solid array (bottom). The total system contains the nucleic acid extraction and amplification sections on a solid array as well as two heaters and the optical system. In step 1, the user loads the sample patient into the chip. in step 2, the system purifies the RNA within 3 min in the extraction section. In steps 3 and 4, RT-PCR is performed through solid phase process on an array of different pathogen targets with real-time monitoring.

Full size image

Fluid control and valving system

Figure 2A shows the design of the platform. To enable timely and precise transfer of reagents for nucleic acid purification and SP-PCR assays, the system integrates a new type of built-in mechanical valving system to control the fluid motion (Fig. 2B). The valve is a 3D printed single-part rigid cube inserted in a predesigned empty space of a flexible layer. The rigid cube has a channel (0.3 × 0.3 mm²) in the middle of its core. At a normally closed status of the valve, the channel inside the rigid cube is not aligned with the channel networks of the system, thus buffers are kept trapped in the reservoirs. To open the valve, i.e., enabling the buffer to pass from a buffer reservoir to the system, the valve body is externally pressed. Once the valve is pressed, the valve body is displaced vertically (z-axis), enabling an alignment for the valve channel with the system channel, thus permitting fluid flow (driven by negative pressure) (Movie S1). Vertical displacement of the valve is also possible due to the elastic property in the layers of the platform. To close the valve, the pressing force is removed, hence bringing the valve to the normal closed status. To prevent the undesired fluid flow through or beside the valve wall during the normally closed status of the valve, the size of the rigid valve must be bigger than the space in the system where the valve is designed to fit in. We tested several rigid valve sizes with a fixed size (2.3 mm³) of the host space in the top flexible layer. through the fluid pressure (P) inside the microchannel that the valve system can withstand prior to fluid leakage or burst (Fig. 2D). For example, for a 2.3 mm-side cubic space in the flexible layer, a 2.53 mm-side cube valve, which a 10% bigger size than the host space, shows that the fluid bursts between the two sides of the valve around 40 kPa (Fig. 2D). Such pressure is almost 20 times bigger than the operational pressure of the fluidic system (hydrodynamic pressure drop of the system). This size difference allows the valve to fit tightly between the flexible layers, forming a sealing mechanism for the valve and preventing fluid leakage of the undesired reagent from other chambers. The mechanical motion provides an instant and precise response for fluid control. For automated pressing of multiple valves (valve 1 to valve 7), a preprogrammed motion of 7 pressing pins is used with the aid of a CAM motion system (Fig. S1A). In comparison to several valving systems in the literature that involve a complex level of system integration, such as piezoelectric¹⁶, electrostatic¹⁷, and pneumatic valves¹⁸, the suggested press-to-activate mechanical valve is simple, precise, and reliable.

A Total assay on the chip. B The built-in valve system. The rigid body of the valve was inserted into an empty space of the flexible top layer. The fluid was not allowed to pass at a normally closed status of the valve, wherein the channel in the valve core was not aligned with the main channel of the system. To allow fluid motion, the valve was pressed externally, creating a vertical displacement of the valve, hence aligning the valve channel with the system channel network. The rigid valve returns to its original position and closed status after removing the pressing force. C Structure of the integrated silica matrix. A top inlet enabled the extraction buffers and the RNA to pass through the silica matrix to capture, purify, and release the RNA. D The valve performance. Pressure burst (maximum pressure that the valve can hold before leaking) vs the chip operation pressure for different sizes of the rigid valve. E Evaluation of on-chip extracted RNA using a benchtop real-time PCR machine. The error bars represent the standard deviation calculated from 3 independent experiments.

Full size image

Sample preparation

Nucleic acid (RNA) is extracted in the purification section prior to the SP-PCR. In brief, the viral spiked sample was added to the lysis buffer. The lysed RNA was permitted (controlled by opening V1 and V7) to pass through the silica matrix (SM) (Silica Y-SM-BC-1, Biocomma), wherein the RNA was captured (Fig. 2A and C). Next, a washing buffer 1 and 2 flow through the SM (controlled by V2 and 7 and V3 and V7, respectively). Then, an elution buffer was flown through the SM to elute the RNA from the SM and transfer it into chamber 2 (C2) (controlled by V4 and V6). The eluted RNA was ready for SP-PCR in the amplification section. We evaluated the performance of on-chip RNA extraction by a benchtop real-time PCR machine to estimate the PCR efficiency (via standard curve) (Fig. 2E). Serial dilutions of RNA samples (10 K,1 k, 100 and 10 copies) were spiked on the lysis and the purified RNA was collected from C2 then RT-PCR was run for original RNA dilation and the collected RNA, simultaneously (Fig. 2E). Figure 2E shows a delay in the PCR cycles for the on-chip extracted RNA compared to original RNA stock indicating an RNA copy loss during the extraction. With an average PCR cycle delay of 0.72, the average purification efficiency was found to be 61%.

Real-time, quantitative SP-PCR strategy

Prior to SP-PCR, the RT-PCR reaction mixture is brought from the PCR reagent reservoir and mixed with the pure RNA in C2. A reverse transcription step (RT) is performed in C2 at 50 °C for 4 min to get complementary DNA (cDNA). Afterwards, PCR is carried out by alternating the mixture between chamber 1 (C1) and C2 during each cycle (Fig. 3A–C). The two chambers (C1 and C2) sit above two temperature-heating zones, 60 °C and 95 °C, respectively (Fig. S1A). To move the PCR mixture forward and backward among the C1 and C2, an external pin presses the top side of C2, which is made from flexible material (Fig. 3A). The flexible top side of C2 enables elastic deformation, thus forcing the fluid to pass to C1 via a connecting channel (Fig. 3A and Movie S1). To bring the fluid from C1 to C2, the pressing force is removed from the top of C2, bringing the deformed C2 to its original volume, thus creating a negative pressure in C2 that forces the fluid to refill C2. The pressing instantly (<0.5 s) displaces the fluid from C2 to C1. Upon the PCR mixture transportation, the temperature of the PCR mixture is adjusted to the new chamber temperature in 3.5 s, 5 s, and 10 s for the PCR volumes of 10 μl, 15 μl and 30 μl, respectively (Fig. S2A). Usually, a relatively big PCR volume (>15 μL) is necessary for a reliable PCR assay since a small volume reduces the existence probability of target copies and increases fluid handling issues, such as evaporation or complexity in handling small volumes. However, with large PCR volume, the heating up and cooling down of the PCR solution is lengthy (>1 h). Besides, a PCR cycler requires a bulky thermal cycling system that consists of a heater, a fan, a heatsink, and a precise temperature controller. Here, the suggested dual-chamber PCR system enables a fast thermal response since the PCR solution moves to a preset temperature compartment; hence, there is no need to heat up and cool down the whole system. The PCR solution adjusts to the wall temperature of the compartment instantly. The system exploits the transfer of PCR mixture among the chambers to quickly heat up and cool down the mixture, thus eliminating the need for a bulky thermal cycler (Fig. S2B).

A–C Mechanism of fluid transportation during PCR cycling. An external pressure of the C2 forces the fluid to pass into C1. Releasing the force creates negative pressure in C2, hence quickly retreats the fluid into C2. B The RT-PCR in a two-chamber system. RT-PCR starts with the reverse transcriptase (RT) process in C2 to obtain cDNA from the RNA, and then the PCR process takes place in C1 and C2. C The side view for the liquid transport mechanism via the external pressing pin. D Mechanism of the solid-phase PCR. The double-stranded cDNA is heated to 95 °C to separate it into single strands in C1. The single-stranded DNA pieces (amplicons), along with other PCR ingredients, are in the liquid layer above specific DNA probes that are immobilized onto a solid surface. As the PCR mixture is transferred to C2 (60 °C), new DNA is made (via annealing) in the liquid while some of it binds (via hybridization) to the matching probes on the solid surface. As the PCR cycles continue, specific FAM-labeled forward primers extend to the appropriate sites on the template. More DNA is produced, which also acts as a new template for further amplification on the solid surface.

Full size image

For each PCR cycle, the cDNA is denatured at C2, then the reaction mixture is transferred to C1 for the annealing and extension step (Fig. 3B and C). The annealing and extension occur in both the liquid phase and the solid phase. The forward primer is fluorescently labeled with FAM. During the annealing step, new amplicons are formed in the liquid phase with forward and revers primers annealed to the denatured strands and got extended, while portions of the denatured strands (with FAM labeled forward primer) are hybridized to the immobilized probes on the solid-phase array (Fig. 3D). The probes are elongated and the FAM labeled strands stay on the solid-phase assay to provide the fluorescence signal. Such a primer labeling sequence strategy enables detection using a single spectral range (single emission wavelength).

Traditionally, strong fluorescence in PCR solution, originated by the freely labeled primers, prevents a real-time reading for solid array unless a special background suppression or high signal-to-noise ratio mechanism is used, such as confocal scanning¹³, surface plasmon resonance spectroscopy (SPR)¹⁵, opto-fluidic equipment⁸, wavelength shift of a silicon micro-ring optical resonator¹⁴, and microfluidic spatially defined droplet arrays¹⁹. Those alternative techniques involve complicated and excessive optical components, moving objectives for fluorescence scanners or opto-fluidic equipment. Here, we propose the simple two-chamber solution that allows the PCR reagent to be removed from the solid-phase array in C1, thereby enabling clear fluorescence reading. After each PCR cycle, the PCR mixture that contains FAM-labeled forward primer and reverse primer is transferred to C2 to eliminate any source of signal interfering with the array (Fig. 3B), allowing for real-time signal reading of the solid array. Thus, the fluid reciprocates among the two chambers to (i) change the temperature of the PCR mixture from 60 °C (in C1) to ~95 °C (in C2) or vice versa, (ii) temporarily empty C1 for real-time fluorescence signal reading of the solid array after each PCR cycle and iii) provide a continuous and dynamic motion for the PCR solution.

Performance of two-chamber SP-PCR

Initially, to evaluate the proposed dual-chamber PCR system, a traditional liquid-PCR test was performed for a single target (Covid-19 virus) with SYBR green dye. Fluorescent signals were intensified as the PCR number increased, proving the functionality of the system (Fig. S2C). Then, we evaluated our dual-chamber SP-PCR concept via immobilizing 3 spots of a specific probe of Covid-19 virus, 3 spots for negative control (oligonucleotides having no homology to the target sequence), and 3 spots for the FAM probe on the surface of C1 (Fig. 4a). RNA and RT-PCR reagents were added in C2. Then, an RT-PCR was performed through oscillation of the assay mixture between C1 and C2. Fluorescence signal of the solid panel was plotted in real-time after each annealing step in C1 to determine the assay performance (Fig. 4B). Fluorescence measured on the microarray spot increased linearly with the concentration of amplicons in solution and on the solid spots. Since all the PCR reagents are dragged away from the reading chamber (C1), no wash steps are required to reduce the fluorescent background. Figure 4B indicates that the signals of the FAM probe remained relatively stable over the 40 PCR cycles, indicating the stability of the chemical bonding through a 40 times fluid alteration between C1 and C2. Meanwhile, the specific spot for Covid-19 displayed an increase in fluorescent signal as the PCR cycle increased due to the solid-phase PCR reaction. The sensitivity (limit of detection (LOD) of the assay was 10 copies/reaction with PCR efficiency of 83.4% for a single target of covid-19 virus (Fig. 4C).

A Real image of the solid-phase PCR for a single target (Covid-19 virus) in C1. The immobilized FAM shows a stable and strong signal through the PCR cycle increment, while the signals of the target spot (Covid-19) gradually increase with the increment of the PCR cycles. B Real-time signal reading for the array spots for Covid-19 virus with copy numbers of 10k, 1k, and non-template target (NTC). The line represents the average of the 3 measurements of the 3 spots of the array. Annealing and denaturing times are 20 s and 5 s, respectively. C The standard curve for PCR efficiency calculation for the SP-PCR with a single target (Covid-19 virus). D Influence of annealing time (hybridization) in C1 on PCR performance. The inset of graph (D) shows the real-time fluorescence with PCR cycles. The error bars represent the standard deviation calculated from 3 independent experiments in (B–D).

Full size image

Conventional SP-PCR runs in a static reaction above a two-dimensional solid array, wherein the diffusion process between the PCR solution and the solid array is the dominant mechanism that enables interaction between the immobilized probes and the PCR components. Pierik et. al. (13) only obtained 58% PCR efficiency after hybridizing amplicons in a static SP-PCR. While in our system, moving the PCR components above the solid array enhances interaction, hence the PCR efficiency. The suggested dual-chamber platform creates an additional fluid motion, i.e., advection, due to the transporting of PCR solution. The displacement of the PCR reagents from C2 to C1 occurs in less than 0.5 s, creating a 2000 μL/min fluid velocity. With such velocity, a turbulent flow occurs inside the C1 (Figs. S6 and S7), and Movie S2). The high SP-PCR efficiency proves the role of the dynamic motion of PCR components in overall performance. To reduce the assay turnaround time, we also optimized the annealing time (Fig. 4D). Apparently, a minimum of 12 s of annealing time is required for the assay without compromising the assay efficiency. Unless mentioned otherwise, we use 12 s annealing for the SP-PCR assay. Such annealing time is important to reduce the essay turnaround time.

Sample-to-answer multiplexed real-time detection

To evaluate the integrated system for multiplexed real-time assay, we designed an array of specific probes for 5 viral pathogen targets (Covid-19, Influenza A (Inf. A), Influenza B (Inf. B), Rhinovirus and Parainfluenza (Para-inf.) viruses) (Fig. 5A). RT-PCR is performed in the two-chamber PCR section for the 5 viral targets after the RNA purification in the extraction section. Figure 5B depicts the fluorescence signals of the solid array in real time after the annealing/hybridization process. The assay successfully distinguished the existence of 5 viruses for the 5 positive cases (spiked viral RNA in lysis buffer). Figure 5C shows 3 positive tests in 3 lanes for the array for the three spiked viral RNA (Covid-19, Inf. A, Inf. B, Rhino, and Para-inf. viruses), while two lanes show no signal, since no viral RNA was spiked (no template control (NTC). The average LOD for the 5 viruses is 10 copies/reaction. The proportion of negative test results out of truly negative samples, i.e., specificity, is 100% for the cutoff at cycle 38 (Fig. 5E–I). Slightly progressive signals start to build up on the solid spot after the 40 cycles, which limit accuracy to distinguish between the positive and negative samples. Hence, we strictly consider a signal after 38 cycles as a negative sample, as running the PCR above 40 cycles shows a slightly stronger signal for samples with low copy numbers (<10 copies) than the negative control (NTC) (Fig. S8).

A Arrangement of the immobilized probe in C1. B Multiplexed PCR for 5 viral pathogens run simultaneously with 10,000 RNA copies and 0 copies. C Multiplexed PCR for 5 viral pathogens with 3 positive lanes (3 viral RNA was spiked), with 5 lanes containing probes immobilized for 5 viral RNAs. D Assay total time. E–I LOD for Covid-19, Inf. A, Inf. B, Rhinovirus and Para-inf., respectively. The different colors in E-I mean each point represents a completely an independent test.

Full size image

With instant transporting and the rapid heating/cooling of the PCR mixture, a short turnaround time of the multiplex assay is achieved. The sample-to-answer test takes ~20 min (Fig. 5D) to perform a multiplex assay, including 3 min for RNA extraction, 4 min RT process, and 14 min for 40-cycle PCR of 15 μL volume. Compared to commercialized multiplex assay platforms, such as Filmarray 2.0 and BioFire® SpotFire® from Biomerieux Inc., that require 45 min and 15 min, respectively, for the sample-to-answer detection, our multiplex system provides a competitive assay time with a simplified system instrumentation thanks to the dual-chamber system that eliminates the thermal cycler and multiple emission filters (Figs. S1 and S5).

We have demonstrated a real-time sample-to-answer assay based on a solid-phase PCR array for multiplexed detection. The proposed system allows real-time signal reading, hence quantification, for solid-phase PCR without using external or special optical equipment. The real-time quantification was enabled through moving the PCR solution away from the solid-phase array after each annealing step, hence reading the signals of the array in an empty space. To allow the fluid movement, a two-chamber PCR system was integrated, where DNA annealing, extension, and signal reading occur in one chamber (C1), and DNA denaturation occurs in another chamber (C2). To achieve the sample-to-answer assay, which includes several reagent additions, we have also introduced a new built-in valve system enabling precise and simple reagent manipulation for both nucleic acid (RNA/DNA) purification and amplification steps. To assess the utility of the total platform in clinical scenarios, we performed multiplexed viral detection for 5 viruses simultaneously, analyzing the impact of PCR dynamic reagent motion on solid-phase PCR assay sensitivity, and characterizing the valve system properties of manipulating several assay reagents and buffers.

For manipulating the assay reagent and performing nucleic acid purification, the valving system is an essential component for sophisticated assay performance. However, to ensure fluid control robustness, the valve includes several parts that increase the system integration complexity. The passive valving, which usually depends on the surface tension of the chip material and storage conditions, lacks performance robustness. Hence, we have developed a built-in single-part valve for precise fluid control. The valve that consists of a rigid cube with a hole inserted into a predesigned flexible space traps buffers in reservoirs as long as the central hole remains misaligned with the chip channels. When externally pressed, the valve aligns with system channels, enabling fluid flow. The elastic platform layers facilitate vertical displacement, ensuring an instant response and preventing inner fluid leakage. The prevention of leakage mechanism is guaranteed by designing the rigid cube slightly larger than its housing space, creating a tight seal. In contrast to several suggested valving systems, our press-to-activate valve is extremely simple (single part), precise (instant response), and reliable (mechanical actuation), while the valving system doesn’t require additional layers or oil to prevent fluid leakage. With the integration of a valve system (7 valves) and a silica matrix, we designed an auxiliary unit for nucleic acid purification on the system for the important role of providing a sample-to-answer test without middle-involved steps to avoid sample contamination (Fig. 3). The nucleic acid purification takes 3 min on the disposable chip with a fully automated platform.

With the nucleic acid (RNA) being purified and ready for RT-PCR, two chambers (C1 and C2) communicate with each other through a narrow channel, allowing the RT-PCR reaction mixture to be reciprocated back and forth between the two chambers, with two different heating zones, for PCR thermal cycling. In C1, real-time signal reading of the amplified PCR products on the immobilized probes was enabled by performing hybridization for labeled amplicons to the complementary probes on the array. Because the PCR mixture that contains the unattached FAM-labeled primers is driven away to C2 during each PCR cycle, the signal of the amplified amplicon on the solid array is distinguishable without the need for extra optical tools. Hence, cumulative data enables a sensitive single reading for a multiplexed quantitative analysis of the PCR-based assay in real time, with an average sensitivity of 10 copies/per reaction of the viral loads. Furthermore, the two-chamber PCR system provides rapid temperature changes (Fig. S2B) and enhanced mixing (Figs. S6 and S7), thereby significantly reducing assay processing time as compared to known solid array systems^1,13. Such assay performance enhancement, in terms of sensitivity and shortening the annealing process time, can be attributed to the dynamic motion of the PCR reagent above the solid array, where an extra active mixing tool displays performance superiority compared to the traditional stationary solid-phase PCR, which experiences slow diffusion mixing. Enhancing the diffusion of PCR products near the solid array surface plays an important role in improving the amplification efficiency of the solid-phase PCR, while transporting the PCR mixture away from the solid array enables real-time and sensitive single reading unlike the non-quantitative end-point solid-phase PCR⁸. Additionally, since the immobilized probes/primers experience mild temperature (60 °C) at the surface of C1 during the PCR process, a thermal degradation or denaturation, which may result in a drop-off in primer efficiency, is avoided. Hence, a dual-chamber system is ideal for real time multiplexed assay.

In practical terms, the proposed set-up allows a practitioner to inject the patient sample (nasal swab or saliva) directly into the microfluidic device and obtain the multiplex assay result in ~20 min without compromising any purification or amplification steps. Although several platforms proposed a rapid PCR system^20,21, usually a significant step is eliminated or compromised for purification or amplification steps to claim a rapid assay. In Fig. 4D, we showed that, for example, reducing the annealing time beyond (12 s) can dramatically postpone the Ct-value of the PCR process. For low-copy number viral sample, such a delay can provide a false negative (beyond 38 Ct-Value). The dual-chamber PCR system presents a scalable multiplexed assay for several pathogen targets. The assay is fast and sensitive with high real-time quantitative detection capabilities. The average LOD of several tested viral pathogens was 10 copies/reaction. Below such a copy number, the possibility of false positive results increases due to the interaction between the non-specific products and the immobilized probe. In brief, such quantitative amplification detection with scalable multiplexing capabilities is achieved by the combination of liquid and solid phase amplifications and real-time assessment of hybridization to an array of multiple targets in a two-chamber PCR system. A limitation of the present study is that assay performance was evaluated using purified RNA spiked into lysis buffer; therefore, future studies will be conducted to validate sensitivity, specificity, and robustness using clinically relevant nasal swab and saliva samples.

As noted, due to the simplicity of the platform, we expect that clinical application of the suggested multiplex real-time detection platform with sample-to-answer capabilities will be both straightforward and cost-effective, with huge potential adoption in clinics or point-of-care areas. The application of the proposed system is applicable for disease screening for poor-resource clinics or national ports for its structure simplicity and rapid turnaround assay time during an epidemic or pandemic pathogen outbreak, with the ability to design and implant certain group of targets on a solid-phase array. Additionally, the simple thermal system drastically reduces the size and weight of the system PCR, making it a suitable system for drive-throw tests with significant ramping up and cooling down time reduction. We tested only 5 pathogen targets on an array as proof of concept for the multiplexing capabilities. However, printing a larger array is obtainable, as we have shown in Fig. S4. A yet further aspect of providing a solid array on a separate substrate, which can be slotted in the second chamber during the chip production, may improve manufacturing flexibility for targeting the total analysis system to different analytes. This allows the system’s user to utilize the same PCR platform for producing different batches of analysis chips directed at different analytes.

Platform structure and fabrication

The platform consists of a disposable device and permanent auxiliary parts. (Fig. S1) The disposable device (30 × 50 mm²), wherein the assay takes place, consists of a bottom layer of a 3-dimensional (3D) printed biocompatible material (Biomed, Formlabs) and a top layer of flexible polydimethylsiloxane (PDMS) material. The 3D-printed bottom layer has an engraved space (35 × 4.5 × 2 mm³) to host a flexible PDMS cuboid that allows structure deformability during the valve operation. Chambers and the connecting microchannels lie in the bottom layer (Fig. S2A). The top PDMS layer contains only the cubic grooves for the rigid valve. Flexible material is used for the top layer since the operation of the valves and C2 requires material deformation. Because the PDMS is an air-permeable material, which increases the evaporation of the PCR solution, the bottom and side walls of the disposable device were rigid 3D-printed material. Details of fabrication and the platform components are listed in Table S1. The valve cube is made of a 3D-printed rigid material (Biomid resin, Formlabs). The body of the valve was inserted inside the grooves of the top layer (Fig. S1). The permanent parts of the platform are two constant-temperature heaters fixed at the bottom of the device, excitation and emission filters with a camera fixed at the top of the device, and a CAM system for fluid control. The CAM system consists of 7 CAMs, 7 followers, 7 pins, and 7 springs to mechanically control the valve and the fluid motion in C2. A stepper motor is used to control the motion of the CAM (Fig. S2A and Movie S3).

PCR chamber design

The two chambers (C1 and C2) for the PCR process possess different shapes and features since they perform different steps of annealing and denture, respectively. C1 was designed shallow (250μl-depth) and long (3 mm-length) to increase the surface-to-volume ratio, hence increasing the heat dissipation of the hot PCR mixture (95 °C) coming from the C2. C2 is designed as a hemisphere shape to enable a complete fluid transfer once the pressing-pin is engaged, since the pressing-pin tip has a hemisphere shape. A microchannel (130 μm × 130 μm × 2700 μm) connects C1 and C2 for the fluid alternating among the chambers. The microchannel cross-section (A) was small to increase the fluid velocity (v) upon the fluid transporting among the chambers, given that the volumetric fluid flow rate (Q) is proportional to v and A (Q = v × A). The microchannel was long (2.7 mm) to prevent a thermal gradient between C1 and C2, although this length creates a microchannel total liquid volume of ~0.4 μL (2% of the total assay volume). The microchannel was on the top of C2 hemisphere to allow any generated bubble to pass forward to C1 during the pin pressing, while the microchannel has smooth-rounded edges at the C1 side to enable smooth fluid motion from C1 to C2.

Surface functionalization and probe immobilization

Attachment of the probe to the surface of the array compartment is performed via functionalizing the surface of the microfluidic space. The surface of the C1 was functionalized with an aldehyde group (-CHO) before the probe immobilization (Fig. S3). Initially, we tried to functionalize the 3D-printed material at the bottom of C1. However, unrepeatable probe density and uniformity were detected in specific spots (Fig. S4A), thus we coated the 3D-printed surface of C1 with PDMS since PDMS has a hydroxyl group (OH-) upon oxygen plasma treatment, then proceed for PDMS surface functionalization for the solid array. Immobilization of probes harboring a C6-amino linker to PDMS surfaces was performed as reported²² with some modifications. Briefly, the PDMS layer above the 3-D printed surface was thoroughly cleaned with acetone, RNase-free water, and dried with N₂. Next, the surface was subjected to oxygen plasma treatment for 5 min and immersed in 5% (3-Aminopropyl) triethoxysilane solution. After overnight incubation at room temperature, the surface was washed with RNase-free water, dried with N₂, and cured for 1 h at 110 °C. Afterward, the surface was immersed in a 5% glutaraldehyde solution for 2 h at room temperature, washed with RNase-free water, and dried with N₂. Finally, 3 μL of each probe (100 µM) were spotted on their respective lane by employing the column array dispenser developed by our team (Fig. S4 and Movie S3). Finally, 1% bovine serum albumin (BSA) in phosphate-buffered saline was added in C1 and C2 to reduce non-specific binding, then washed with RNase-free water. BSA also prevents the PCR products from adsorption on the surface of the PCR chamber. Probe attachment and uniformity on the spots were ensured by allowing the dispensing columns that contain the probe solution to be in continuous contact with the treated PDMS surface for 12 h at room temperature under controlled humidity conditions (humidity index ~98%). The stability of the immobilized probe against the continuous motion of the PCR mixture is shown in the unchanged intensity of the FAM probe after 40 cycles of fluid transporting between C1 and C2, thanks to the strong chemical immobilization.

Primer and probe design

Primers and probes designed for this study are presented in Table S2. A FAM fluorophore was included in the 5’-end of forward primers to allow the detection of PCR-amplified products. A C6-amino linker was included in the 5’-end of the probes to facilitate attachment to PDMS surfaces. Forward and reverse oligos were purchased from TAG Copenhagen, and probes were purchased from IDT.

Array arrangement

To track and confirm probe immobilization, a probe with a random nucleotide sequence was designed with a C6-amino linker on its 5’-end and FAM fluorophore on its 3’-end (Immo-P, Table S2). The fluorescent signal of this probe was assessed after probe immobilization and PCR amplification (Fig. S3). Once the immobilization protocol was optimized and the probe attachment ensured, this reporter probe was omitted from the final multiplex RT-PCR array.

RNA samples

Synthetic RNA controls for Covid-19 (SARS-CoV-2, MN908947.3), Influenza A H1N1 (NC_026433), Influenza B (NC_002209), Rhinovirus 89 (NC_001617.1), and Parainfluenza 1 (NC_003461.1) were purchased from Twist Bioscience.

Nucleic acid purification and PCR buffers

RNA purification kit (QIAamp Viral TNA Min Kit, QIAGEN) was uploaded in their predesign reservoirs as follows: 50 μl, 50 μl, 50 μl, 10 μl for lysis, wash 1, wash 2 and elution buffers, respectively. For a 20-μL RT-PCR test, the SensiFAST™ Probe No-ROX One-Step kit from Meridian Bioscience was used. 16 µl of RT-PCR master mix, which includes 6 µl of purified RNA, was combined with 4 µl of a solution with all 10 forward and reverse primers mixed in equimolar concentrations (0.4 μM). Amplification in this system is initiated by a forward primer in solution and a surface-immobilized reverse probe. While the immobilized strand participates in surface-based amplification, released strands can contribute to solution-phase amplification, which is predominantly driven by the forward primer. Equimolar concentrations of forward and reverse primers in solution were used to maintain compatibility with standard PCR conditions and to support both solution-phase and surface-associated amplification. As the assay performed reliably under these conditions, further optimization of primer concentrations was not conducted. The volume of the reagents is scaled up or down according to the total volume of the assay. RT was performed in C2 at 45 °C. Next, the C2 temperature increased to 95 °C to allow polymerase activation for 1 min. Unless specified, PCR amplification was performed by promoting cDNA denaturation in C2 for 5 s at 95 °C and then transferring the solution to C1 to perform annealing and extension for 12 s at 60 °C.

Fluorescence measurements

The fluorescence intensities of all spots shown on the array are measured at the end of the annealing step of the liquid-phase PCR (the hybridization step of the solid-phase PCR) using a CMOS camera (CS165MU/M) and ThorCam software. We used the average fluorescent intensity of three spots of the same target, i.e., probe sequence, for each measurement. We normalized the recorded signals to the measured fluorescence intensity of the C6-amino linker-FAM fluorophore. For the traditional liquid phase PCR (Fig. 4B), the fluorescent signal was produced when SYBR Green (excitation at 498nm- wavelength) was activated after binding to the double-stranded DNA (dsDNA) generated. The full image for C2 was taken after each cycle, wherein the fluorescent signal was proportional to the dsDNA present in the reaction. Details of fluorescence signal acquisition are in SI.

1. Dikdan, R. J. et al. Multiplex PCR assays for identifying all major severe acute respiratory syndrome Coronavirus 2 variants. J. Mol. Diagn. 24, 309–319 (2022).

   Article  Google Scholar 

2. Li, J. et al. Simultaneous detection of multiple respiratory pathogens using an integrated microfluidic chip. Anal. Chem. 96, 13768–13776 (2024).

   Google Scholar 

3. Welch, N. L. et al. Multiplexed CRISPR-based microfluidic platform for clinical testing of respiratory viruses and identification of SARS-CoV-2 variants. Nat. Med. 28, 1083–1094 (2022).

   Article  Google Scholar 

4. Kreitmann, L. et al. Next-generation molecular diagnostics: leveraging digital technologies to enhance multiplexing in real-time PCR. TrAC Trends Anal. Chem. 160, 116963 (2023).

   Article  Google Scholar 

5. Artika, I. M., Dewi, Y. P., Nainggolan, I. M., Siregar, J. E. & Antonjaya, U. Real-time polymerase chain reaction: current techniques, applications, and role in COVID-19 diagnosis. Genes 13, 2387 (2022).

   Article  Google Scholar 

6. Clark, T. W. et al. Rapid multiplex PCR for respiratory viruses reduces time to result and improves clinical care: results of a systematic review and meta-analysis. J. Infect. 86, 462–475 (2023).

   Article  Google Scholar 

7. Blumenfeld, N. R. et al. Multiplexed reverse-transcriptase quantitative polymerase chain reaction using plasmonic nanoparticles for point-of-care COVID-19 diagnosis. Nat. Nanotechnol. 17, 984–992 (2022).

   Article  ADS  Google Scholar 

8. Khodakov, D. A. & Ellis, A. V. Recent developments in nucleic acid identification using solid-phase enzymatic assays. Microchim. Acta 181, 1633–1646 (2014).

   Article  Google Scholar 

9. Seo, S. E. et al. Ultrafast molecular diagnosis-based solid-phase photonic PCR for respiratory pathogen variant discrimination. Adv. Funct. Mater. 34, 2402972 (2024).

   Article  Google Scholar 

10. Xiong, M. et al. Space-coded microchip for multiplexed respiratory virus detection via CRISPR-Cas12a and RPA. Talanta 291, 127815 (2025).

    Article  Google Scholar 

11. Xu, Z. et al. Microfluidic space coding for multiplexed nucleic acid detection via CRISPR-Cas12a and recombinase polymerase amplification. Nat. Commun. 13, 6480 (2022).

    Article  ADS  Google Scholar 

12. Gu, Z. et al. Solid-phase PCR based on thermostable, encoded magnetic microspheres for simple, highly sensitive and multiplexed nucleic acid detection. Sens Actuators B Chem. 298, 126953 (2019).

    Article  Google Scholar 

13. Pierik, A. et al. Real time quantitative amplification detection on a microarray: towards high multiplex quantitative PCR. Lab Chip 12, 1897–1902 (2012).

    Article  Google Scholar 

14. Shin, Y., Perera, A. P., Kim, K. W. & Park, M. K. Real-time, label-free isothermal solid-phase amplification/detection (ISAD) device for rapid detection of genetic alteration in cancers. Lab Chip 13, 2106–2114 (2013).

    Article  Google Scholar 

15. Pollet, J., Janssen, K. P. F., Knez, K. & Lammertyn, J. Real-time monitoring of solid-phase PCR using fiber-optic SPR. Small 7, 1003–1006 (2011).

    Article  Google Scholar 

16. Feng, G. H. & Kim, E. S. Piezoelectrically actuated dome-shaped diaphragm micropump. J. Microelectromech. Syst. 14, 192–199 (2005).

    Article  ADS  Google Scholar 

17. MacHauf, A., Nemirovsky, Y. & Dinnar, U. A membrane micropump electrostatically actuated across the working fluid. J. Micromech. Microeng. 15, 2309–2316 (2005).

    Article  Google Scholar 

18. Inman, W. et al. Design, modeling and fabrication of a constant flow pneumatic micropump. J. Micromech. Microeng. 17, 891–899 (2007).

    Article  ADS  Google Scholar 

19. Zhang, R., Gong, H. Q., Zeng, X., Lou, C. & Sze, C. A microfluidic liquid phase nucleic acid purification chip to selectively isolate DNA or RNA from low copy/single bacterial cells in minute sample volume followed by direct on-chip quantitative PCR assay. Anal. Chem. 85, 1484–1491 (2013).

    Article  ADS  Google Scholar 

20. Seder, I., Coronel-Tellez, R., Helalat, S. H. & Sun, Y. Fully integrated sample-in-answer-out platform for viral detection using digital reverse transcription recombinase polymerase amplification (dRT-RPA). Biosens. Bioelectron. 237, 115487 (2023).

    Article  Google Scholar 

21. Lee, S. H. et al. Bubble-free rapid microfluidic PCR. Biosens. Bioelectron. 126, 725–733 (2019).

    Article  Google Scholar 

22. Chou, W. C., Hu, W. P., Yang, Y. S., Chan, H. W. H. & Chen, W. Y. Neutralized chimeric DNA probe for the improvement of GC-rich RNA detection specificity on the nanowire field-effect transistor. Sci. Rep. 9, 11056 (2019).

    Article  ADS  Google Scholar 

Download references

Authors and Affiliations

1. Department of Health Technology, Technical University of Denmark, Lyngby, Denmark

   Islam Seder, Rodrigo Coronel Téllez, Jing Zhang, Tao Zheng & Yi Sun

2. Appolon Bioteck, Chaponnay, France

   Stephen McCleary, Saïd El Mouatassim & Jean-François Brugère

Contributions

I.S., Y.S., S.M., S.E.M. and J.F.B. conceived the idea, with Y.S. supervising the project. I.S. fabricated and characterized the device, designed and performed experiments, and conducted data analysis. R.T. designed and tested primers and probes. J.Z. and T.Z. contributed data analysis and comments on the manuscript. I.S. wrote the manuscript with input from all authors. All authors approved the final version of the manuscript.

Corresponding authors

Correspondence to Islam Seder or Yi Sun.

Competing interests

Y.S., I.S., and R.C.T. have a patent application pending assigned to their institution. A European patent application related to the invention, “Fluid handling system for multiplex pathogen detection: a fast and sensitive real-time PCR assay on a solid array” (Application No. 25189320.2), has been filed. The other authors declare no competing financial interests or non-financial interests.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions