Fig 1.
General concept of turning a 3D Printer into a molecular device.
We explored the possibility of retrofitting low-cost 3D printers to perform rapid, automated nucleic acid isolation (< 15 min) and amplification (< 25 min). (A) The potential applications of the modified 3D printer in the context of low-cost molecular diagnosis of infectious diseases. Step 1 includes processing varied specimens through magnetic particle based nucleic acid extraction and purification. In step 2, the eluted DNA or RNA can be used in downstream analysis by various molecular detection approaches. Examples are: a) real-time thermal cycling using a commercial thermal cycler (this work), b) probe-based amplification operated by the 3D printer with post-PCR imaging (this work)., c) other molecular detection methods such as a lateral flow assay after PCR or isothermal amplification (underway). (B) Schematic showing the advantages of using the 3D printer for automated extraction. It is faster and requires minimal user-intervention once the process begins. It reduces performance variations induced by users. (C-E) Photos that show the major components that were added to the 3D printer to enable our current work. (C) The Printrbot Play 3D printer (base footprint: 5” × 11”) is the most compact 3D printer we have successfully converted into a nucleic acid extraction device. An 8-sample MPPA (magnetic particle processor attachment) was attached to the luer-lock syringe to carry out magnetic particle based nucleic acid extraction. The magnets were stored inside disposable 0.1-mL PCR tubes to eliminate direct contact between the magnets and samples. (D) The Printrbot Simple 3D printer (base footprint of 11” × 13”) is small enough to fit inside a biosafety cabinet. The 8-sample and 12-sample (shown) MPPA function properly when attached to either the Printrbot Play or Printrbot Simple 3D printer. (E) An adaptor for holding the PCR capillary tubes is attached to the Printrbot Simple 3D printer. The PCR capillary tubes were shuttled by the 3D printer between the 95°C denaturation and the 60°C annealing/extension baths (not shown).
Table 1.
Sequence of primers and probes used in this study.
Fig 2.
The 3D printer can process samples with a wide range of bacterial concentrations.
Different concentrations of B. subtilis in LB broth medium were extracted using our 3D printer protocol (duplicate wells of 3×105 CFU, 3×104 CFU, 3×103 CFU, 3×102 CFU, 30 CFU, and LB broth as a negative control). DNA was eluted in 50 μL of elution buffer. (A) Real-time PCR curves from the samples with a 5-log difference in concentration. (B) The Cq values of real-time PCR in panel A was plotted against the input concentrations (R2 = 0.997, slope = -3.23). The Cq values shown here are slightly better (lower Cq) than those in S4 Fig at the same concentration because we used a smaller elution buffer volume (50 μL instead of 100 μL).
Fig 3.
Sensitivity curve for the detection of C. trachomatis in clinically collected urine.
DNA was extracted from undiluted urine, 10×, 100×, and 1,000× diluted urine samples in duplicate, using Qiagen QIAamp spin columns and the 3D printer with MP based DNA extraction. We note that the urine sample input volume, elution output volume, and template volume used in this run were kept identical to allow for proper performance comparison. The R2 values (0.972 from the 3D printer vs. 0.996 from the Qiagen column) and the slopes (-3.23 vs. -3.24, respectively) of the two methods listed in the plot indicate that the two extraction approaches performed similarly. With similar Cq values obtained for both methods, a paired t-test P-value of 0.49 further confirms that the 3D printer’s performance in extraction is comparable to one of the gold standards used in nucleic acid isolations.
Fig 4.
No discernible specimen contamination during 3D printer automated NA extraction.
The real-time PCR plot shows that processing a high concentration of cells (107 CFU/extraction) in samples does not lead to cross-contamination in the adjacent wells (only 10 CFU/extraction or negative samples). The NTC is the no template control. In addition, the results suggest that the extraction consistency between wells was very high since the Cq values of the six identical samples at 107 CFU/extraction are almost identical (19.24 ± 0.21, n = 6).
Fig 5.
Using the 3D printer to perform RNA extraction.
(A) Real-time RT-PCR plot of dengue virus-spiked samples (undiluted sample, 10× and 100× dilutions). Non-infectious cultured human cells suspended in PBS were used as the diluent as well as the negative control. The NTC is the no template control. (B) Plot of Cq vs concentration of viral particles. The slope of the plot (duplicate at each concentration) shows that extraction and PCR were both efficient (slope of -3.44 cycles/log of concentration).
Fig 6.
Conducting PCR in aluminum water baths heated by a 3D printer’s heated bed.
(A) Top view of the two mini-water baths’ setup. A single heated bed provided the appropriate temperatures (95.6°C and 60.7°C as shown in the data logger) for the NA denaturation and annealing/extension steps needed for PCR. (B) Side view of the setup for capturing fluorescent signal change. Four blue LEDs were placed under the PCR capillary tubes to excite the green fluorescent (FAM) dye-labeled hydrolysis probes inside the tubes. The DNA template used in tube 1 was from undiluted urine sample, tube 2 was 10× diluted urine sample, tube 3 was 100× diluted urine sample, and tube 4 was the NTC. The photo was taken after 35 cycles, by a smartphone camera with an orange plastic filter placed over the lens. The fluorescent signal can also be monitored when the tubes are moved between the water baths by the 3D printer. By viewing the fluorescent intensity of the glass capillary tubes, the user can determine which tubes have positive results and semi-quantitatively determine the template’s concentration if the reactions have not reached the plateau phase of PCR amplification. (C) Gel image from the four tubes shown in panel B. Expected amplicon size is 165 bp. Lane M is DNA ladder (from bottom to top: 50/100/150/200/300/500/800/1500 bp). The intensity of the gel bands corresponds well with the template concentration and the fluorescence from the glass tubes.