DNA topology influences molecular machine lifetime in human serum

Lifetimes and operational performance were investigated for a DNA nanomachine and linear probe in human serum and blood to elucidate design principles for future biomedical applications of DNA-based devices.

The absorption spectrum of a 1% solution of whole human blood in 1×PBS was measured using an Agilent Technologies Cary 300 and is shown in Fig. S1. The excitation and emission wavelengths are indicated for the TET and Cy5.5 dyes. The bar widths correspond to the instrument bandwidths used for data acquisition.  Prior to incubation in 70% serum for native polyacrylamide gel electrophoresis (PAGE) experiments (described in Supplementary Information S4), the nanomachine and linear probe were purified via native PAGE to filter out excess DNA and malformed structures. The Cy5.5-IBRQ linear probe appeared to agglomerate in solution and could not be purified. Neither the Cy5.5-IBRQ nanomachine nor the TET-IBFQ devices exhibited this behavior. Fig. S2 shows the results of a typical native PAGE purification attempt with multiple variations of the dyequencher combinations on the linear probe. As seen in the gel image, both Cy5.5-IBRQ and Cy5.5-IBFQ linear probe exhibited the agglomeration and slow migration behavior. Moreira, et al. found various dye-quencher modifications to alter the thermodynamic stability of duplex DNA. 1 Though they did not specifically study the Cy5.5-IBRQ pair, they did find that similar cyanine dyes and IBRQ increased duplex stability in cacodylate buffer. It is possible that this phenomenon is partially to blame for the strange gel migration behavior of our probe, though the nature of the Cy5.5-quencher interaction requires further study.

Supplementary Information S4: Incubation in Serum followed by Native PAGE Analysis
Samples for degradation experiments were prepared at 250 nM in a 100 µL 7:3 serum:PBS mixture and incubated at 37 °C for varying times. Diluting human serum can shorten DNA device lifetimes, as endogenous levels of actin and saline in human serum greatly inhibit the activity of some nucleases. 2 Note that dilution with PBS, as done here, should not reverse the inhibition of nuclease activity in human serum. After incubation, the DNA from the samples was extracted using 50 µL chloroform and 50 µL 25:24:1 phenol:chloroform:isoamyl alcohol (Sigma-Aldrich), followed by centrifugation at 13,000 rcf for 10 minutes. PAGE analysis of the degraded samples was performed using 1.5 mm thick native 10% polyacrylamide gels made with 1×TAE, 6.25 mM magnesium acetate, pH 8.0 gel and running buffer. Samples were loaded using 19:1 loading buffer of ficoll and 0.04 wt% bromophenol blue and run at 150 V for 120 minutes at 20 °C. Completed gels were stained with SYBR Gold (Life Technologies) and imaged using a multiplexed fluorescence detection and gel documentation system (FluorChemQ, ProteinSimple). Representative native PAGE gel sections for human serum samples are shown in  To quantify the degree of degradation, intensity profiles were acquired from each incubation time stamp lane. The background intensity of each lane was subtracted from the intensity profile to obtain the net peak height prior to normalization, as illustrated in Fig. S4. The net peak heights of all samples for a particular gel were then normalized to the net peak height of the 1 min incubation sample from that gel. This normalized intensity reflects an approximate fraction of operational devices remaining at a particular time. Fig. S5 shows the normalized intensity-time profiles. After baseline subtraction, but prior to normalization, band intensities for the first two time stamps were frequently approximately equal, with variations primarily due to pipetting and Goltry, et al.

DNA Machines in Serum and Blood Supplementary Information
Electronic Supplementary Information -6 experimental error, indicating that very little degradation occurred in the first few minutes of incubation and validating the choice of the 1 min incubation sample as the control for normalization.  As in previous studies, 3,4 the data were fit to a first-order exponential decay model: where D is the DNA device concentration, D ! is the initial concentration, t is the incubation time, and λ is the decay rate.

DNA Machines in Serum and Blood Supplementary Information
Electronic Supplementary Information -7 The decay rate is related to the mean lifetime τ by The half-life t !/! can then be calculated as Substituting Eq. (S2) into (S1) and dividing by D ! to get an expression for the normalized band intensity I yields The data were fit to Eq. (S4) using OriginPro software to obtain values for τ, which were then used to calculate t !/! and λ. The results of this analysis are shown graphically as dashed curves in Fig. S5 and numerically in Table S2. Mean lifetimes in human serum are also shown in Fig.  2f.
Nearly all of the degradation data were well described by the first-order exponential decay model (R 2 > ~0.93), but it was noted that the TET-modified Bright probe appeared to settle out to a nonzero final fluorescence, indicating possible failure of this model (Fig. S5c, orange data set, R 2 = ~0.86). Similar behavior was observed, though not discussed, in data from previous publications (Conway, et al., Supplementary Figure S12, T80 HEG data set) and may warrant modification of the mathematical model used for such degradation studies. 4   TABLE S2. Results of first-order exponential decay analysis for the DNA nanomachine and linear probe incubated in 70% human serum and 70% non-heat-inactivated FBS at 37 °C (fitted curves shown in Fig. S5). No data is given for the Cy5.5-IBRQ Dark probe because of purification issues, as described in Supplementary Information S3.

Supplementary Information S5: Kinetics experiments in Buffer, Human Serum, Blood, and FBS
To assess the effect of enzymatic degradation on dynamic device function, the nanomachine and linear probe were operated with Cy5.5-IBRQ labels in human serum, whole human blood, and PBS at 25 and 37 °C, and their fluorescence emission was measured as a function of time, as shown in Figs. 3 and 4. These DNA reaction kinetics experiments were performed in 100% PBS; 97.5% human serum/2.5% PBS; and 97.5% heparinized whole human blood/2.5% PBS. Fuels were added in constant volume and varying concentration so that each injection was at 50% molar excess relative to the previous injection. The high excess of fuel was added to drive the state transitions rapidly and decouple device degradation from fuel degradation. Teflon coated magnetic stir bars (Fisher Scientific) were used to mix solutions or solutions were mixed by hand using pipettes. Experiments were run in 4 mm specialty optical glass cuvettes (Starna Cells) or 10 mm methacrylate cuvettes (Fisher Scientific).
Kinetics measurements in PBS and serum were acquired using Agilent Technologies Cary Eclipse fluorescence spectrophotometers using the excitation and emission wavelengths provided in Supplementary Information S2. Due to the high absorbance of blood and the 90° orientation of the excitation and emission paths within Cary Eclipse spectrophotometers, kinetics measurements in whole blood were recorded using a custom laser-based fluorescence photometer designed for Cy5.5, as described in Supplementary Information S6.
The nanomachine was also operated in fetal bovine serum (FBS) and the serum from a second volunteer. In FBS, the nanomachine kinetics data exhibited clear evidence for degradation. As seen in Fig. S6a, the fluorescence of the Closed state reached only 50% of the Relaxed state intensity and returned to only 60% in the second Relaxed state. However, the large increase in intensity when transitioned to the Open state indicates that the primary degradation was to the exposed single-stranded toehold region of the F 1 fuel strand. As seen in the schematics (Fig. 1a), the fuel F 2 can Open the Closed nanomachine by binding directly to the available single-stranded portion of the actuator and displace any F 1 fuel strands that were rendered inaccessible to the cF 1 strand by degradation of the toehold. Nonetheless, after 200 minutes, the nanomachine was rendered inoperable and state transitions for the second cycle were greatly diminished. The kinetics data in serum from a second volunteer (Fig. S6c) were virtually indistinguishable from those in the serum of the first volunteer (Fig. S6b, reproduced from Fig. 4b). Goltry, et al.

DNA Machines in Serum and Blood Supplementary Information
Electronic Supplementary Information -10

Supplementary Information S6: Custom Laser-Based Epi-Fluorescence Photometer
To avoid the signal loss inherent in fluorescence measurements in which the excitation and emission are measured at a relative angle of 90° in highly absorbing media, a custom laser-based epi-fluorescence photometer was constructed to measure fluorescence emission from the same surface as the excitation. A schematic of the system is shown in Fig. S7. The system consists of a temperature and current controlled laser diode (Opnext HL6756MG 15 mW continuous wave diode and LXI-GP controller), whose output is modulated by a chopper (Oriel) at 205 Hz. The diode output was tuned to ~670 nm. The drive current and an absorptive neutral density filter were used to adjust the laser intensity, and two mirrors and lenses were used to route and focus the incident beam. A single-edge dichroic mirror (Semrock, long-pass edge at 700 nm, OD2 for wavelengths below 691 nm) reflected the incoming beam toward the sample and allowed the Cy5.5 fluorescence signal to pass through to a photomultiplier module (PMT, Hamamatsu) to amplify the signal. The PMT output was input to a lock-in amplifier (Stanford Research Systems) referenced to the chopper frequency. Because this system was not equipped with a temperature controller for the sample stage, measurements could only be taken at ambient temperature. The baseline detector current of the system was ~6 fA with no solution in the system. A 250 nM sample of the Cy5.5-IBRQ nanomachine in 100% heparinized whole blood yielded a detector current of ~20 pA.

Supplementary Information S7: Kinetics and Native PAGE Results for Nicked Nanomachine in Buffer
The sequences for the nicked nanomachine (Fig. S8) are the same as those listed in Table S1, with the B strand split into two halves of equal length, B 1 and B 2 . The nicked nanomachine was operated in 1×PBS (Fig. S9b), and while the Closed state was still distinct, the Open state fluorescence was not discernable from the Relaxed state. The presence of a nick on the actuator domain allows for a greater range of motion in the Relaxed state, while also decreasing the rigidity of the Open state; hence, the fluorescent distinction between Relaxed and Open states is negligible for the nicked nanomachine. For comparison, operation of the intact nanomachine in 1×PBS is shown in Fig. S9a (same as Fig. 3a). S8. Schematic of the nicked DNA nanomachine. Similar to the original three-state DNA nanomachine (Fig. 1), the nicked DNA nanomachine transitions between Relaxed, Closed, and Open states with the addition of fuel strands, F 1 and F 2 , and their complements, cF 1 and cF 2 .
The intact and nicked nanomachines were also compared via native 10% PAGE. Though the migration rates of the Relaxed and Open states were indistinguishable in the presence of a nick on the actuator domain, the migration rate of the Closed nanomachine was decreased when a nick was present. PAGE results from incubation experiments in human serum (Fig. 2) did not indicate this type of degradation occurring for the intact Closed nanomachine; however, the mean lifetimes for the Relaxed and Open nanomachine states reported in this publication may include minimally degraded structures such as the nicked nanomachine, since such structures migrate at the same rate as the intact versions.   In the Closed state, the nicked nanomachine migrated significantly slower than its intact counterpart.