Rapid detection of single bacteria in unprocessed blood using Integrated Comprehensive Droplet Digital Detection

Blood stream infection or sepsis is a major health problem worldwide, with extremely high mortality, which is partly due to the inability to rapidly detect and identify bacteria in the early stages of infection. Here we present a new technology termed ‘Integrated Comprehensive Droplet Digital Detection’ (IC 3D) that can selectively detect bacteria directly from milliliters of diluted blood at single-cell sensitivity in a one-step, culture- and amplification-free process within 1.5–4 h. The IC 3D integrates real-time, DNAzyme-based sensors, droplet microencapsulation and a high-throughput 3D particle counter system. Using Escherichia coli as a target, we demonstrate that the IC 3D can provide absolute quantification of both stock and clinical isolates of E. coli in spiked blood within a broad range of extremely low concentration from 1 to 10,000 bacteria per ml with exceptional robustness and limit of detection in the single digit regime.


Supplementary Figure 3. High throughput droplet generator.
A two layer droplet microfluidic device was designed to increase the throughput of droplet generation. Blood samples were introduced through two inlets on the top layer and split into 8 flow streams. The DNAzyme solution was injected through an inlet located on the other side of the top layer and also split into 8 flow streams. The blood and DNAzyme solution streams were merged one-by-one and passed down through the interlayer connecting holes. Mixtures were then encapsulated within 25 μm droplets by the oil phase introduced from the bottom layer inlets.

Supplementary Figure 4. DNAzyme sensors rapidly detect target bacteria E. coli in HEPES buffer in droplets. a)
Representative fluorescence images show the colocalization of a single Syto17 stained bacterium and DNAzyme sensor signal after 900 s incubation time in the droplet. b) Real-time fluorescence monitoring of a single droplet that contains DNAzyme sensors and a single bacterium. c) Quantification of fluorescence intensity of droplets in experiments described in b). Data are represented as mean ± s.d., n = 3. Threshold was obtained by 'Three sigma rule (three times s.d. of the baseline (relative intensity was stably 1)'. Fluorescent intensity goes above threshold at 8 min. 5 μm droplets are used in this Figure. Supplementary Figure 5. SYTO 17 (red color) stained control Bacillus (a) or target E. coli K12 (b) were spiked in blood at a concentration of 10 7 cells ml -1 which was encapsulated in a single cell manner in droplets with DNAzyme sensor (final blood content is 10% in this data). After a 3-hour reaction, droplets are counted on-chip using our confocal detection system. Red spikes represent droplets that contain SYTO 17 stained cells which are observed on both control (a) and target (b) cells. However, only the target E. coli K12 (b) produced a green color DNAzyme signal that is above the background (i.e., droplets that do not contain cells). At such a high initial cell concentration (10 7 cells ml -1 ), there are occasionally 2 bacteria (i.e., 2 red spikes) observed in one droplet. In these cases, the DNAzyme signal directly correlates to the number of bacteria in the droplet. a) and b) were performed in triplicate and a total of ~70,000 droplets were counted. c) Maximum photon counts of representative droplets that contain 0 or 1 E. coli. Black dot represents the photon count from each droplet. Box plot with an overlay of actual data is shown. Mean value is shown as red dot. n = 200, ****P < 0.0001, Two-tailed Student's t-test. A count is considered as a "positive hit" if it is higher than the threshold (dash lines) that is set to be the maximum photon count of empty droplets. This set of experiments reveals that our encapsulated DNAzyme sensor system possesses zero false positive rate and minimal false negative rate (~0.5%) using the 1D on-chip droplet counting. Figure 6. Schematic diagram of 3D particle counter. 4 Excitation beam from the diode lasers (Laser1 and Laser2) are guided through the excitation filter (F ex ) and the dichroic mirrors (D1 and D2). The beam is then focused into the cuvette through the objective lens (L1) (see photo inset). Emission spectra are collected from the same objective lens and transmitted through the dichroic filters, which is focused via a lens (L2) into the pinhole (PH). The light beam is further collimated by another lens (L3) toward the detection unit. A dichroic filter (D3) splits the emission beam before reaching the emission filters (F em ) placed in front of the two photomultiplier tubes (PMT1 and PMT2). The analog signals from the PMTs are converted through AD converter acquisition card for data analysis. Figure 7. Fluorescent signal kinetics of the DNAzyme reaction in bacterium-containing blood droplets at different reaction time. a) Time course of fluorescence signal in droplets was monitored every 15 minutes using a 3D particle counter (only some representative time points were shown here). Fluorescence signal of droplets containing single bacterium starts to rise and become measureable at 60 min DNAzyme reaction time. A representative positive signal (red box) is zoomed in and shown in b) where it is analyzed with a pattern recognition algorithm (red dashed line). (LB) and incubated with shaking at 37 °C for 14 h. Ten (10) μL of this culture was further inoculated in 5 ml fresh LB for 4 h then 1 ml LB was transferred to a microcentrifuge tube. The mixture was then centrifuged at 10,000 rpm for 4 min at RT and decanted.

Supplementary
After resuspension of the cell pellet in 500 μL of DNAzyme reaction buffer (RxN buffer, 50 mM HEPES (pH 7.5), 150 mM NaCl, 15 mM MaCl 2 , 0.01 % (v/v) Tween 20), the mixture was heated at 90 °C for 5 min to break the cell wall. Cell lysates were then centrifuged and the supernatant was collected as a sample.

Colony forming unit (CFU) assay
The concentration of bacterial cells was determined by the CFU assay using freshly cultured bacterial cells. E. coli K12 or clinical isolates were freshly cultured as described above. Two ml of cultured cells were then serially diluted in 10-fold intervals with LB and divided into two groups; one for stock with 20% glycerol and the other for the CFU assay. Cell concentrations ranging from the 4 th to the 9 th dilution were then plated onto the agar-LB plates using 100 μL of diluted cells. The bacterial suspension was then spread evenly over the entire surface of the plate. After drying, the plates were taped to prevent contamination and incubated at 37 °C overnight or until colonies formed. After the colonies reached a suitable size, they were counted to determine the cell concentration.

DNAzyme activity assay in bulk
The fluorescence-based analysis of the cleavage reaction of the construct, as shown in

Specificity test with different bacterial cells or mammalian cells in bulk
Denaturing polyacrylamide gel electrophoresis (dPAGE)-based specificity assays, as shown in Fig. 2C

Monitoring DNAzyme sensor signals with clinical E. coli isolates in blood in bulk
Eleven bacterial strains (see above) and fresh blood was obtained from UC Irvine Institute for Clinical and Translational Science (ICTS) through the blood donor program (IRB # 2012-9023). Bacterial cells were freshly cultured in LB and limiting dilutions were performed to quantify the cell concentration as described above. 1,000 CFU of each bacterial strain were then spiked in 25 μl of 20% blood and mixed with 24 μl of 2✕ RxN buffer. After gentle mixing, 1 μl of 5 μM DNAzyme was added to the mixture before incubating for 30 min at room temperature. The fluorescence intensity was quantified using a fluorescence microplate reader.

Bacterial lysis
Note that lysozyme (1 mg ml -1 ), a bacteria lysis agent, is pre-mixed in the DNAzyme sensor solution. The use of lysis agents allows the target molecules to be rapidly released from bacteria, which will further decrease the assay time. We have systemically optimized bacteria lysis conditions using various agents including Triton X-100, IGEPAL, SDS and lysozyme alone or in combination, and identified that lysozyme most efficiently lyses bacteria without interfering with droplet formation or DNAzyme sensor function.

1D on-chip detection system
For real-time droplet monitoring on-chip, the fabricated devices were placed onto a translation stage and appropriately aligned with a custom-built confocal spectrometer.
The spectrometer excitation source consisted of a 488-nm CW air-cooled argon ion laser. A dichroic mirror (505DRLP02; Omega Optical, Brattleboro, VT) was used to reflect the 488-nm radiation and to define a vertical axis normal to the surface of the optical table. A 40Χ microscope objective was employed to bring the light to a tight focus