A three-dimensional microfluidic device embedded within a thermal cycler tube for electrokinetic DNA extraction

Abstract

Microfluidic devices have been widely used in modern chemical and biological analyses as stand-alone units, typically in series with other equipment such as extraction columns, manual or robotic pipetting, and even advanced next-generation sequencing systems. While microfluidic devices have enhanced various aspects of laboratory workflows, their integration with established commercial assay platforms remains limited. To this end, we developed a three-dimensional microfluidic insert embedded directly into a commercially available polymerase chain reaction (PCR) tube. This integration creates a microfluidic device compatible with conventional thermal cyclers, which support complex temperature cycling and multiplexed fluorescence detection. The integrated system facilitates key bioassay functions like nucleic acid purification through a selective ionic focusing method known as isotachophoresis (ITP), PCR amplification, and real-time fluorescence detection. We validated the performance of the integrated system by purifying nucleic acids from raw human serum samples and detecting exogenous SARS-CoV-2 N gene using FAM-labeled TaqMan probes, with both the DNA extraction and detection carried out within the same PCR tube. We achieved a detection sensitivity of 100 cp μL⁻¹ within a total process time of 60 min in these experiments. Human serum samples processed without purification show no PCR amplification results. This integrated system demonstrates the powerful concept of integrating microfluidic structures into form factors compatible with the highly complex and sensitive operation of current off-the-shelf systems.

---

Introduction

Modern chemical and biological analyses rely heavily on commercially available analytical systems that offer a wide range of mixing, separation, reaction, incubation, and detection modalities. These leverage manual and robotic pipetting, centrifuges, chemical mixers, physicochemical separations, thermal incubation, and highly automated and complex devices combining multiple steps such as next-generation sequencing systems.^1,2 Microfluidic devices offer the ability to integrate multiple functions—such as purification, mixing, reaction, and detection—on a single platform. An integrated microfluidic platform that combines sample preparation and detection could eliminate unnecessary transfer steps (to avoid contamination), streamline workflows, and improve assay efficiency. To date, microfluidic devices have mostly been used as stand-alone components in series with traditional equipment.^3,4 The integration is often limited to transfer between a planar microfluidic device and a traditional assay system. A typical example is the use of standard, off-the-shelf nucleic acid extraction kits based on surface adsorption which are often used immediately prior to transferring purified analytes (e.g., DNA) into a microfluidic chip for further analysis.^5,6 A recent review⁷ of microfluidic devices integrating the relatively new tools of clustered regularly interspaced palindromic repeats-associated proteins (CRISPR-Cas) reported that only one in twenty studies integrated the difficult steps of sample preparation and enzymatic reactions into their microfluidic chips.

There has been a dearth of studies and demonstrations wherein microfluidic devices are integrated within and fully compatible with commercially available assay systems. One notable integration is so-called microfluidic well plate inserts. Siloam Biosciences⁸ developed a custom 96-well microplate wherein each well also included a connected, curved, in-plane microchannel on the detection side of the plate. The device form factor was compatible with microplate reader and showed advantages to enzyme-linked immunosorbent assays (ELISA), including significant increase in sensitivity due to the high surface area-to-volume ratio. Several reported systems have integrated microfluidic structures in well plate systems for cell culture and microscopy. Gheibi et al.⁹ demonstrated the use of inserts designed to fit standard 12-well plates, creating confined culture conditions that allowed primary hepatocytes to maintain a differentiated phenotype in the absence of convection. Berry et al.¹⁰ created custom microplates with rail channel microfluidic structures designed to achieve capillary-driven flows and support interfaces between hydrogel zones for custom cell signaling assays. The latter was compatible with plate reader equipment with multi-wavelength detection capabilities. To our knowledge, these systems have been limited to modifications of plate-and-well formats where the microfluidic features have in-plane form factor. Further, we know of no such embedded systems (compatible with traditional assay equipment) which have integrated the challenging function of sample preparation for raw samples.

In this work, we explore a new approach wherein a three-dimensional (3D) microfluidic structure is integrated within a form factor compatible with a highly versatile conventional assay system: a thermal cycler with integrated real-time fluorescence detection. Specifically, we created a 3D device that drops into a standard thermal cycler tube compatible with ThermoFisher Scientific Applied Biosystems 7500 Fast cycler. This cycler is a widely used instrument that enables precise thermal control and multi-color fluorescence detection. A key advantage of our approach is the integration of sample preparation using isotachophoresis (ITP) for nucleic acid extraction. Compared to solid-phase extraction methods, ITP has several potential benefits, including eliminating the need for manual buffer exchanges (as with solid-phase extraction); enabling purification from complex samples like blood, urine, and plasma¹¹ without moving parts or centrifugation; and offering an extraction process that is minimal bias with respect to nucleic acid length or sequence.^12,13 Additionally, ITP allows for integration with downstream processes such as enzymatic reaction and amplification.¹³ Using this microfluidics-embedded tube system, we demonstrate integration of the following three important functions of bioassays: sample preparation and nucleic acids purification via ITP, enzymatic amplification via PCR, and real-time fluorescence detection. Each function is achieved within this convenient and compatible form factor, and the latter two functions are achieved in completely sealed tubes to avoid cross-contamination. We demonstrate the detection of target DNA sequences in raw human serum samples. Overall, the system demonstrates the powerful concept of integrating 3D microfluidic structures into form factors compatible with the highly refined and sensitive operation offered by current off-the-shelf systems.

Methods and materials

Design and fabrication of 3D-printed microfluidic structure

A schematic of our 3D-printed device is shown in Fig. 1A and B. This device is designed to drop into and mate with a standard 0.1 mL PCR tube. The structure is asymmetric and is bounded by a flat plane on one side in order to provide space for liquid filling the PCR tube. The structure itself includes three interconnected regions: a trailing electrolyte (TE) reservoir located at the top (one side of which is open to the local atmosphere), an internal sample reservoir, and a 3D spiral microchannel terminating at an exit port near the bottom of the flat side of the structure. This port opens into the rest of the internal space within the PCR tube. The multiple interacting reservoirs and cone-shaped transition region between reservoirs and microchannels are used to aid in dispensing of multiple liquid chemistries and a phase-transformable gel into the device. The interconnected upper reservoirs consist of the TE reservoir and the sample reservoir. The TE reservoir, colored in green in the schematic of Fig. 1A and B, holds a volume of 22 μL. The sample reservoir, colored in red, with a volume of 20 μL, holds the serum sample. The U-shaped inlet of the spiral microchannel and gel are designed to eliminate pressure-driven flow from the sample reservoir to the spiral channel during loading. Fig. 1C shows the actual device 3D printed using an off-the-shelf 3D printer (CADworks3D ProFluidics 285D printer). The image shows the spiral microfluidic channel and reservoirs within the system filled with blue dye.

Fig. 1 3D-printed microfluidic structure and ITP-mediated assay for nucleic acid extraction and detection. (A and B) Isometric view schematics of the asymmetric 3D printed microfluidic part. The spiral channel is first filled with LE buffer, followed by dispensing of sample and TE buffer. 20 μL of LE buffer is also dispensed into the 0.1 mL PCR tube. A biological sample such as serum or plasma is injected into the central sample reservoir (red). Trailing electrolyte buffer is dispensed in the top TE reservoir (green). The spiral channel (purple) is used for nucleic acid extraction via ITP. (C) Actual device 3D printed using a CADworks3D ProFluidics 285D 3D printer. Here, blue food dye was used to visualize internal structure. (D) Insertion of the custom microfluidic device into MicroAmp 0.1 mL PCR tube. The insert is positioned such that its flat surface is viewed edge on. (E) Working principle of nucleic acid extraction. After loading, the 3D structure is inserted into the PCR tube. Positive and negative electrodes are connected to LE and TE reservoirs, respectively. During ITP extraction, negatively charged nucleic acids are focused within an ITP zone and travel downward through the spiral channel. Purified DNA is delivered to the LE reservoir. (F) Schematic of one-batch nucleic acid detection workflow from sample to result. After serum preparation, the tube and insert are loaded, and nucleic acid extraction is performed. This is followed by real-time PCR amplification and detection of target genes.

The outer contour of the device is designed for integration with a separate, free-standing PCR tube. The insertion of our custom microfluidic device in the commercially available tube of MicroAmp 0.1 mL PCR tube (Applied Biosystems) is shown in the image of Fig. 1D. The 0.1 mL MicroAmp, cone-shaped polypropylene tube has a top inner diameter of 5.5 mm, and a vertical height of 15.5 mm. As shown in Fig. 1D, this arrangement allows the outer material and shape of the assembly to fit within and interact with the standard, commercially available PCR tube. The space within the PCR tube other than the inserted 3D part is defined by the inner wetted wall of the PCR tube on one side and by the wetted flat wall of the 3D-printed asymmetric structure on the other side. This space forms a third reservoir, which is termed here the leading electrolyte (LE) reservoir shown in Fig. 1E. The bottom of the LE reservoir is connected to the outlet of the spiral microchannel of the 3D insert, and the top side of the liquid within the LE reservoir is exposed to the atmosphere.

The 3D model of the tube device was designed in SolidWorks 2024, featuring separate reservoirs and internal spiral microchannels with cross-sectional dimensions of 200 × 180 μm. The model was then 3D-printed with clear microfluidics resin (CADworks3D). More details of the 3D printing material properties are provided in the ESI† document. We also show that the insert fits within tubes of various manufacturers and models, such as low profile tubes by Thermo Scientific, PurePlus 0.2 mL PCR tubes by Labcon, and TempAssure 0.1 mL PCR tubes by USA Scientific (see Fig. S1†). To avoid clogging, after printing, we injected isopropanol into microchannels and used an air gun to clean out resin residue, before 60 s curation in the ultraviolet light chamber.

Integrated microfluidic workflow for nucleic acid extraction and detection

We developed a complete workflow to leverage the PCR tube-embedded spiked-in serum purification, amplification, and detection in less than 60 min, as shown in Fig. 1F. After serum preparation, we insert the 3D-printed microfluidic part into the commercially available PCR tube and effect ITP within this insert device to extract total nucleic acids from the serum and deliver them to the LE reservoir space in the PCR tube in about 15 min. Upon completion of the extraction process, we remove the external electrodes and microfluidic insert, leaving the extracted nucleic acids and the LE buffer in the PCR tube. We then add the PCR master mix, detection reagents (either SYBR Green dye or a fluorescently labeled TaqMan probe), and forward and reverse primers specific to the target DNA sequence to the PCR tube. The PCR tube is sealed with its integrated optical cap. After sealing, the tube is never opened again (e.g., after amplification) in order to minimize contamination. We then insert the tube within the real-time PCR (qPCR) thermal cycler. The programmed thermocycler then performs a fast-mode or standard qPCR. As normal, the presence of target genes in the serum-extracted DNA is indicated by the detection of a fluorescent amplification signal through the real-time PCR system. The system provides quantitative data regarding the specific DNA concentration and estimated copy number within the sample. Additionally, this setup and methodology can be adapted for conducting spike-in control experiments using an exogenous DNA template.

Serum and exogenous spike-in DNA preparation

Red-top (clot) tubes were obtained from Stanford blood center. After the blood collection, the tubes were left undisturbed for half an hour, then the clot was removed by centrifuging at 2000 rcf for 10 min in a centrifuge (Eppendorf centrifuge 5417C). After that, the supernatant liquid component (serum) was transferred into microcentrifuge tubes with a pipette and kept at 2–8 °C. If not analyzed promptly, the serum was stored at −20 °C. We included 1% Triton-X as a mild lysis agent and to promote disassociation of DNA from proteins. We also added 1 μL of Proteinase K (Invitrogen) at 20 mg mL⁻¹ to the 10 μL serum sample. The SARS-CoV-2 N gene fragment (New England Biolabs) was selected as an exogenous oligonucleotide. The serial dilution was prepared by dissolving the initial stock solution in DNase-free, biology-grade water. Controlled conditions were established by spiking a known amount of the target into healthy human serum samples.

All blood samples used in this study were obtained as de-identified specimens from the Stanford Blood Center (SBC) under an agreement for in vitro investigational use. All samples were collected from fully consented donors under the Guidelines of IRB #13942 approved by Stanford University's Administrative Panel on Human Subjects in Medical Research. Donors were screened using SBC's standard donor history questionnaire and underwent testing for transfusion-transmissible infections including ABO/Rh, HIV, hepatitis B/C, HTLV, West Nile and syphilis.

Buffer preparation and dispensation

The LE buffer consisted of 200 mM Tris, 100 mM HCl and 0.5% Polyvinylpyrrolidone (PVP). The TE buffer consisted of 800 mM Tris, 200 mM HEPES and 0.5% PVP. To prevent pressure-driven flow and enhance flow stability for smoother voltage curve, the TE was mixed with phase-changing Pluronic F-127 (Sigma-Aldrich), with a mass percentage of 30%. The LE was first dispensed into the spiral channel with a pipette, followed by the dispensation of TE mixture and treated serum sample. The TE mixture was kept on ice to ensure it remained liquid form until dispensed into the TE reservoir with a syringe, where the TE-Pluronic mixture formed into a gel to seal the top. The Pluronic mixture suppresses pressure- and gravity-driven flow, thereby preventing unwanted mixing of solutions during ITP extraction.

Isotachophoresis operation

To prevent cross-contamination during DNA extraction using ITP, all experiments were conducted within a UV cabinet/PCR Workstation (Azer Scientific) to maintain a sterile environment. We sprayed all the surfaces with 10% bleach (Clorox germicidal bleach). Before each experiment, the 3D-printed part was thoroughly rinsed with 10% bleach and then with deionized water for 2 min to remove residual nucleic acids in the channel. After the rinsing step, 20 μl of LE was first loaded in the PCR tube and the spiral microchannel. The serum sample was dispensed into the central sample reservoir, and high-concentration TE with Pluronic F-127 was then loaded in the top TE reservoir. Note that all components, including pipette tips and tube device, were kept in the ice box before dispensing, since concentrated Pluronic solutions gelate at room temperature. In this way, liquid Pluronic was dispensed (with cold pipette) into the device. Once the Pluronic solution approached room temperature and gelated, we applied a constant current of 80 μA through electrodes (positive in LE reservoir and negative in TE reservoir) using a Keithley 2410 sourcemeter. Voltage and current data were acquired real-time through a custom MATLAB code, and fluorescence images were obtained via HCImageLive.

Optical setup and image analysis

The fluorescence images of ITP zone were acquired through an sCMOS camera (Hamamatsu C13440) with an epifluorescence microscope (Nikon Eclipse TE300). The microscope objective (Nikon) with 2× magnification was connected with a 45-degree mirror to enable imaging of a vertical plane. The tube device was vertically positioned in a semi-micro cuvette filled with index-matching glycerol oil to perform non-invasive imaging through the sidewall. A 3D printed holder was utilized to fix the cuvette and electrodes in the experimental set-up, see Fig. S2.†

Real-time PCR amplification and probe-based detection

qPCR was performed using the Applied Biosystems 7500 Fast thermocycling system with two different detection methods. The detailed probe-based plate setup is provided in this section, while the SYBR green detection method is described in the ESI† (see SYBR green-based qPCR detection). Spike-in SARS-CoV-2 N gene fragments in serum were detected via amplification of a 72-base pair internal segment using sequence-specific primers (forward: 5′GACCCCAAAATCAGCGAAAT, reverse: 5′TCTGGTTACTGCCAGTTGAATCTG). PCR was performed in a total reaction volume of 40 μL, consisting of 20 μL master mix (TaqMan Universal PCR Master Mix, ThermoFisher Scientific), 1 μL of 20 μM forward primer, 1 μL of 20 μM reverse primer, 1 μL of 10 μM FAM-labelled TaqMan Probe (IDT), and sample. The probe sequence is /56-FAM/ACCCCGCAT/ZEN/TACGTTTGGT GGACC/3IABkFQ/. Raw amplification plots were processed in QuantStudio (ThermoFisher) to generate quantification cycle (C_q) values. All reactions were performed in quadruplicate for standard curve and for tube-embedded extracted nucleic acids detection. For SYBR Green detection method, post-amplification melting-curve analysis was used to verify the presence of a single PCR product (see ESI†).

Results and discussion

Nucleic acids purification via isotachophoresis and visualization

The schematic shown in Fig. 1E is our custom-integrated system, including the tube-embedded monolithic 3D printed structure, a commercially available PCR tube, and external electrodes. The system uses ITP to purify and transport DNA. ITP is a versatile process applicable to a wide variety of sample purification, including DNA purification from complex samples like human serum.¹² Nucleic acids are highly soluble with a high electrophoretic mobility magnitude and a low pK_a of around 3, and this makes them suitable for purification via anionic ITP. ITP employs an electric field to selectively focus DNA between the TE and LE, while impurities (including PCR inhibitors) remain in the TE zone.¹² Positive and negative platinum electrodes are connected to the LE and TE reservoirs, respectively (the overall setup is shown in Fig. S2†). Current flows from the LE to the TE. The electric field migrates DNA out of the sample reservoir and into the three-dimensional channel in the shape of a narrowing helix within the inset device. DNA purification occurs within this uniform-cross-section spiral channel. Negatively charged nucleic acids are selectively bracketed by the TE and LE co-ions as shown in Fig. 1E, and ITP delivers purified nucleic acids from the serum to the peripheral LE chamber. In this way, purified DNA is separated from strong PCR inhibitors in serum, enabling efficient amplification.^12,13

We developed a custom visualization setup (not required for assay, see Fig. S2†) to obtain epifluorescence videos of the ITP migration of analytes through the spiral channel. As an example, we include a movie showing the migration of fluorescein in the channel as ESI† file Movie S1. Fig. 2A presents a composite image of eight sequential snapshots obtained from this video, illustrating the progression of the fluorescent ITP zone. Note the increase in intensity of the ITP-focused material as the ITP zone migrates down the channel (see Fig. 2A and S3†). This increase of fluorescence in ITP zone is attributed to the semi-infinite injection, which continuously supplies the analyte to the ITP zone. Images from this video were extracted at a constant time interval of 30 s and then superposed to create the composite shown in Fig. 2A. Time points t₁, t₃, t₅, t₇ represent positions where ITP zone is closer to the microscope, while t₀, t₂, t₄, t₆ correspond to positions where the ITP zone is farther from the microscope. We also developed a custom image processing code to identify and track the ITP zone. Fig. 2B shows the measured trajectory of the detected ITP zone, and this aligns with the projected microchannel outline. Here, the abscissa is the horizontal direction, and the ordinate is the vertical direction within the imaging plane. The latter measured trajectory data is color-coded to show the temporal progression of the ITP zone. In Fig. 2C, we show a plot of the measured in-plane ITP velocity vectors as a function of time (the time axis here increases downward, consistent with the downward progression). The in-plane velocity magnitude is also plotted as a grey, filled curve. The velocity vectors clearly show the periodicity of the observed in-plane (x–y plane) velocity as the ITP zone travels through the 3D trajectory of the helical channel.

Fig. 2 Experimental visualization of a moving fluorescent ITP interface within the custom 3D microfluidic structure. (A) Overlaid false-color, microscopic epifluorescence images of the ITP zone obtained at equal time intervals. As time progresses, the ITP zone spirals downward, with increasing intensity indicating rising sample concentration. The spiral microchannel outline is overlaid for reference. (B) Measured fluorescent ITP zone trajectory overlaid on an inverted-grey-scale, time-averaged image of the structure. The trajectory of ITP zone is detected and represented with a false color scale showing the measured time progression. The trajectory closely matches the spiral channel outline. Scale bars in (A) and (B) are 1 mm. (C) In-plane velocity vector plots of the ITP zone as a function of time. U_x represents the horizontal velocity, U_y is the vertical velocity, and |U_s| is the total velocity magnitude within the image plane, calculated as . The velocity vector magnitudes and directions exhibit the expected periodicity as the ITP zone travels at a constant velocity through the 3D channel structure. The triangle tick marks in (B) and (C) correspond to the annotated time points in (A).

The endpoint of each extraction process was discerned by a levelling off of the measured voltage versus time curves (see Fig. S5†); at which point, the current was terminated. At this point also, the ITP zone was just outside of and near exit of the spiral channel. The TE liquid has a lower mass density than the LE liquid. Hence, by terminating the channel near the bottom of the tube, we purposely set up a condition wherein the lower density TE (and the DNA solute) was pushed up by buoyancy forces into the LE reservoir. The buoyancy effect is present because the mass density of LE is greater than that of the adjusted TE (ATE) buffer. A visualization of the buoyant plume of the ITP zone exiting into the LE reservoir is presented in Fig. S4.† The composition and concentration of the ATE zone were simulated using the client-based application for fast electrophoresis simulation (CAFES) developed by Santiago lab.¹⁴ ATE contains 164 mM Tris and 64 mM HEPES, with a measured density of 0.94 g mL⁻¹, compared to 0.99 g mL⁻¹ for LE.

In this way, we leveraged buoyancy-driven mixing to distribute the extracted DNA within the contents of the LE chamber. The LE can also include reaction chemistries to effect reactions with the extracted DNA. After the ITP process, we remove the external electrodes and the microfluidic insert, add reaction chemistries, seal the tube, and place it into the thermal cycler wherein qPCR and subsequently post-amplification analyses are performed.

System performance for SARS-CoV-2 detection

We demonstrated the full functionality of our system by performing qPCR on DNA purified from raw human serum samples. The experimental data are shown in Fig. 3. To demonstrate our system, we chose non-human DNA fragment SARS-CoV-2 N gene as an exogenous oligonucleotide for detection. Exogenous genes are typically used as spike-in control to eliminate high inter-specimen variability¹⁵ and can be employed as biomarkers for the diagnosis of pathogenic infections. Spiking known quantities of DNA into human serum enables precise quantification of system performance. Serum specimens were spiked with various amount of SARS-CoV-2 N gene with final concentrations of 10² to 10⁵ cp μL⁻¹ in serum prior to DNA extraction. For each DNA concentration, combined ITP extraction and amplification experiments were performed in quadruplicate. We performed qPCR using Applied Biosystem 7500 Fast Real-Time PCR System. SARS-CoV-2 N genes extracted from serum were detected via amplification of a 72-base pair internal fragment using sequence-specific primers, and the products were detected via two different ways, probe-based detection (see Fig. 3) or SYBR Green-based detection (see ESI†).

Fig. 3 Real-time PCR results at various nucleic acids concentrations in serum and associated controls. Standard amplification curves of pure DNA in buffer without serum are shown in gray (polygon data symbols). (A) qPCR amplification curves for pure sample in the absence of serum. (B) qPCR amplification curves from serum samples and with and without ITP purification. TaqMan-probed qPCR for SARS-CoV-2 N gene from 10x serial dilutions (curves without data symbols) show successful amplification enabled by ITP extraction. The presence of 8 μL serum in a standard 40 μL reaction system (sans ITP) inhibits qPCR at all DNA concentrations (circle data symbols). A normalized dye fluorescence (ΔR_n) threshold of 0.07 was used to determine quantification cycle (C_q) values. Note the legend for the circular data symbols (blue) is provided in part (A) of the figure and labelled “without ITP”. The inset of (B) shows measured C_q values (as a function of copies per reaction, cp rxn⁻¹) for nucleic acid dilution series for both pure DNA (no serum) and for DNA purified from serum via ITP in our device. A standard curve is shown for clear buffer (no serum, bottom) together with DNA extracted with our system (from serum, top). The assay's detection limit extends below about 100 cp μL⁻¹ in serum.

We successfully purified and detected known amounts of DNA spiked into human serum using our custom-designed tube-embedded microfluidic device. Fig. 3A shows the amplification of DNA samples in the absence of serum. No template controls (NTC) refer to qPCR reactions that do not contain the SARS-CoV-2 N gene in the reaction system. The latter contains 20 μL TaqMan master mix, 1 μL TaqMan probe, 2 μL forward and reverse primers, and 17 μL LE buffer in a 40 μL system. Fig. 3B shows amplification of the DNA purified with our system. NTC-ITP controls refer to amplification results of samples without spiked-in DNA templates but otherwise processed by our microfluidic device. The negative signals from these NTC-ITP confirm there was no detectable cross-contamination or residual DNA in the device. NTC, NTC-ITP and PCR from equivalent amount of exogenous DNA concentration in serum without ITP extraction showed no amplification after 40 cycles (highlighting the strong PCR inhibition of serum). We separately validated our system using a SYBR Green-based qPCR method and post-amplification melting curve analysis (see Fig. S7†). The single peak in the melting curve confirmed the specificity of the amplicon and demonstrated the adaptability of our workflow to different qPCR detection chemistries.

The pH of the LE buffer was measured with ROSS Ultra pH Electrodes (Thermo Scientific Orion) as 8.3. This value is close to the typical and efficient PCR pH of 8.4.¹⁶ To further assess the impact of LE on the PCR reaction, we conducted additional control experiments and confirmed that LE buffer does not reduce amplicon yield. No significant differences were observed between reactions with and without LE buffer (see Fig. S8†).

Fig. 3B insets shows the quantification cycle (C_q) value as a function of concentrations of SARS-CoV-2 N gene. We show linear regression fits with each set of data. First, the lower (gray) data is a standard curve obtained from qPCR reactions using DNA templates in pure LE buffer, without serum. The qPCR efficiency for standard curve was calculated to be 94.7%. The high R² value indicates a strong linear relationship between the log of the DNA amount and the C_q value. This is expected given the qPCR instrument, and the spike-in DNA dilution process. The upper (red) line is the detection curve for DNA extracted, purified, and amplified from serum samples using our system. We attribute the large difference between C_q values (ΔC_t ∼ 7) of ITP extraction from serum versus pure DNA control with no serum to the incomplete extraction of DNA from the sample during the ITP process. There may also be some finite amount of DNA adsorption to the 3D printed surfaces.

To quantify the possibility of residual inhibitory effects and protein contamination from serum, we measured the A260/A280 ratio of the samples before ITP purification and after ITP using an Invitrogen Nanodrop One Spectrophotometer (Thermo Scientific). Prior to ITP purification, the sample mixture exhibited an average A260/A280 ratio of 0.54, using molecular biology grade water as a blank. The Nanodrop instrument's Acclaro Contaminant ID feature identified protein contamination with a coefficient of variation (CV) of 8.0%. After ITP extraction from serum, the purified product showed an average A260/A280 ratio of 1.34, using LE buffer as the blank. The improved A260/280 ratio indicates a significant reduction in protein contamination. This value is within the range reported for extracted cell-free DNA from human blood plasma, blood serum, or whole blood using commercially available kits such as QIAamp Circulating Nucleic Acid Kit (Qiagen)^17–19 or AccuPrep Genomic DNA extraction kit (Bioneer)¹⁸ and is not expected to affect the downstream applications.

For the integrated tube-embedded extraction and detection, exogenous DNA concentration of around 100 cp μL⁻¹ in serum was the detection limit for these experiments. We note the lowest two concentrations yielded approximately the same C_q. We attribute this to the effect of random sampling error near the assay's inherent sensitivity threshold.²⁰ The qPCR efficiency was 90.3% for the majority of the data (ranging from 10⁴ to 10⁶ cp rxn⁻¹), excluding the lowest copy number data points (10³ cp rxn⁻¹) from the efficiency calculation. Below 10 cp μL⁻¹, the qPCR fails to detect target amplification. In its current form, our device does not improve the sensitivity of viral nucleic acid detection compared to commercially available kits. As a comparison, the QIAamp DNA Blood Mini kit (Qiagen), a widely used column-based extraction method, has a reported limit of detection of 560 cp mL⁻¹ of Epstein–Barr virus (EBV) from 400 μL of whole blood, corresponding to 220 cp rxn⁻¹.²¹ And the MAGMAX viral/pathogen nucleic acid isolation kit (Thermo Fisher), a magnetic-bead-based system, achieves a detection limit of 50 cp of EBV per extraction from human plasma, blood and urine.²² Our system currently detects approximately 100 cp μL⁻¹, suggesting potential applications in endogenous cell-free DNA detection.²³ The system may therefore find applications in non-invasive diagnostics of hereditary diseases or cancers. We hypothesize that further optimization of the system design could result in an improved limit of detection. For example, increasing the spiral channel length (and its cross-sectional area) should result in a greater amount of DNA extracted from the finite amount of DNA dispensed into the sample reservoir.

Conclusions

We developed a novel microfluidic device paradigm by embedding a 3D microfluidic structure within a commercially available PCR tube. Our workflow integrates ITP-based nucleic acid extraction from raw samples, qPCR amplification, and probe-based target detection. Using this device, we successfully detected the SARS-CoV-2 N gene spiked into raw human serum samples, demonstrating effective extraction and detection capabilities.

One key advantage of our tube-embedded microfluidic device is its design as an all-in-one sample handling and reaction system, enhancing both time efficiency and ease of use. For instance, our device reduces total assay time from sample collection to result readout to less than 60 min. This turnaround is significantly shorter than traditional solid-phase purification and PCR-based detection methods, which often exceed 2 h and involve ten or more manual steps. We include in the ESI† discussion and comparison to other commercial purification kits. Additionally, our device includes reservoirs, microfluidic chambers, and three-dimensional fluid-handling features which add critical functionalities to a standard reaction tube. Integrated features for electric field control of molecular transport, purification, and buoyancy-based mixing enable versatile sample handling. Moreover, the device's outer contour is engineered for compatibility with a range of standard PCR tube types and models, facilitating integration into diverse laboratory workflows and minimizing the need for specialized consumables. Compatibility with thermal cyclers allows for precise temperature control, efficient enzymatic reactions, and sensitive fluorescence detection, collectively enhancing its utility for nucleic acid diagnostics. Beyond the current demonstration with human serum samples, ITP-based extraction could be adapted for a broader range of biological samples, including plasma, whole blood, saliva, and tissue lysates. Expanding its use to these matrices would further enhance its applicability in clinical diagnostics.

One disadvantage of our current device is the variability in calibration and uniformity across 3D printing batches. We attribute this limitation to the constraints of the 3D printer's finest achievable feature dimensions (30 μm as claimed). This variability affects the ITP extraction process, resulting in extraction times ranging from 10 to 20 min with the same constant current applied, see extraction time histogram in Fig. S6.† Further optimization of the 3D printing parameters or exploration of alternative fabrication methods may be necessary to mitigate these inconsistencies. 3D printing was used in this work only as a rapid prototyping method to demonstrate the new form factor and electrokinetic process. Any future commercial implementation of this device should explore large-scale production of plastic microfluidic tubes. We hypothesize these could be fabricated via injection molding, for example, as two separate pieces that are subsequently assembled to accommodate the complex geometry. A second fabrication possibility may be blow molding to create parts with internal cavities. Our approach also supports parallelization, enabling scalable nucleic acid testing through robotic dispensing and electrode arrays for diagnostics. An additional limitation of the current system is the challenging buffer handling process in the introduction of Pluronics F-127 into the TE reservoir. Currently, we use syringes to facilitate injection and prevent trapping bubbles in the channel or reservoirs. Development of more complex dispensing structures and methods are suggested for further development. The limit of detection of our current design is 1000 cp rxn⁻¹, which is approximately 10-fold higher (less sensitive) than that of the state-of-the-art extraction techniques.^21,22 We attribute this mostly to the low sample volume capacity of the structure of 20 μL, and to the limited extraction efficiency. This work serves as a proof of concept for integrating microfluidic extraction within a thermal cycler tube. Future work may focus on increasing sample volume and/or increasing the length of the extraction channel to improve extraction efficiency.

The demand for integrated, user-friendly microfluidic solutions is growing, driven by the need for efficient, point-of-care nucleic acid testing in both clinical and research settings. Traditional PCR-based nucleic acid detection systems are often time-consuming and require multiple instruments and steps, making them less suitable for rapid diagnostics. The current integrated device may address this gap by providing an all-in-one solution that reduces complexity and the time for nucleic acid extraction, amplification, and detection.

Data availability

The data supporting the findings of this article are available within the article as part of the Methods and materials section and its ESI.†

Author contributions

Conceptualization: J. G. S.; methodology: J. G. S., Q. J., X. Z., Y. W., A. S. A., D. A. H.; investigation: Q. J., X. Z., A. S. A., D. A. H.; visualization: Q. J.; writing – original draft: Q. J., J. G. S.; writing – review & editing: all authors.

Conflicts of interest

Juan G. Santiago and Qi Jiang are listed as coinventors on a pending provisional patent application related to this work filed at the U.S. Patent and Trademark Office (no. 63/721153). Other authors declare no competing interests.

Acknowledgements

Qi Jiang acknowledges support from the Stanford Graduate Fellowship. We thank Xinyi Chen from the Stanley Qi Lab for assistance with and access to the Nanodrop measurements.

References

1. P. Dettinger, T. Kull, G. Arekatla, N. Ahmed, Y. Zhang, F. Schneiter, A. Wehling, D. Schirmacher, S. Kawamura, D. Loeffler and T. Schroeder, Nat. Commun., 2022, 13(1), 2999.
2. T. Hu, N. Chitnis, D. Monos and A. Dinh, Hum. Immunol., 2021, 82, 801–811.
3. S. F. Berlanda, M. Breitfeld, C. L. Dietsche and P. S. Dittrich, Anal. Chem., 2021, 93, 311–331.
4. S. Battat, D. A. Weitz and G. M. Whitesides, Lab Chip, 2022, 22, 530–536.
5. Q. Huang, X. Shan, R. Cao, X. Jin, X. Lin, Q. He, Y. Zhu, R. Fu, W. Du, W. Lv, Y. Xia and G. Huang, Micromachines, 2021, 12(12), 1582.
6. J. Fassy, C. Lacoux, S. Leroy, L. Noussair, S. Hubac, A. Degoutte, G. Vassaux, V. Leclercq, D. Rouquié, C. H. Marquette, M. Rottman, P. Touron, A. Lemoine, J. L. Herrmann, P. Barbry, J. L. Nahon, L. E. Zaragosi and B. Mari, PLoS One, 2021, 16(4), e0243333.
7. A. S. Avaro and J. G. Santiago, Lab Chip, 2023, 23, 938.
8. J. Kai, A. Puntambekar, N. Santiago, S. H. Lee, D. W. Sehy, V. Moore, J. Han and C. H. Ahn, Lab Chip, 2012, 12, 4257–4262.
9. P. Gheibi, K. J. Son, G. Stybayeva and A. Revzin, Integr. Biol., 2017, 9, 427–435.
10. S. B. Berry, T. Zhang, J. H. Day, X. Su, I. Z. Wilson, E. Berthier and A. B. Theberge, Lab Chip, 2017, 17, 4253–4264.
11. V. N. Kondratova, O. I. Serdyuk, V. P. Shelepov, G. I. Potapova and A. V. Likhtenshtein, Dokl. Biochem. Biophys., 2005, 402, 200.
12. A. Rogacs, L. A. Marshall and J. G. Santiago, J. Chromatogr. A, 2014, 1335, 105–120.
13. A. Ramachandran and J. G. Santiago, Chem. Rev., 2022, 122, 12904–12976.
14. A. S. Avaro, Y. Sun, K. Jiang, S. S. Bahga and J. G. Santiago, Anal. Chem., 2021, 93, 15768–15774.
15. G. C. O'Connell, P. D. Chantler and T. L. Barr, Lab. Med., 2017, 48, 332–338.
16. J. M. Wages, Encyclopedia of analytical science, 2005, vol. 243.
17. H. Lee, C. Park, W. Na, K. H. Park and S. Shin, npj Precis. Oncol., 2020, 4, 3.
18. F. Karami, M. Reza, N. Daloii, S. Hantooshzadeh and M. H. Modarressi, J. Fam. Reprod. Health, 2017, 11, 3.
19. C. J. Jorgez, D. D. Dang, J. L. Simpson, D. E. Lewis and F. Z. Bischoff, Genet. Med., 2006, 8, 615–619.
20. P. J. Schmidt, N. Acosta, A. H. S. Chik, P. M. D'Aoust, R. Delatolla, H. A. Dhiyebi, M. B. Glier, C. R. J. Hubert, J. Kopetzky, C. S. Mangat, X. L. Pang, S. W. Peterson, N. Prystajecky, Y. Qiu, M. R. Servos and M. B. Emelko, Front. Microbiol., 2023, 14, 1048661.
21. S. Laus, L. A. Kingsley, M. Green and R. M. Wadowsky, J. Mol. Diagn., 2011, 13, 695–700.
22. Applied Biosystems, Sensitive detection of viral nucleic acids using the MagMAX Viral/Pathogen Nucleic Acid Isolation Kit, Thermo Fisher Scientific, 2019.
23. J. Michelson, S. Rausser, A. Peng, T. Yu, G. Sturm, C. Trumpff, B. A. Kaufman, A. J. Rai and M. Picard, Mitochondrion, 2023, 71, 26–39.

---

Footnote

† Electronic supplementary information (ESI) available: Section S.1 to S.7. See DOI: https://doi.org/10.1039/d4lc01062k

---