Modular fluorescent nanoparticle DNA probes for detection of peptides and proteins

Fluorescently labeled antibody and aptamer probes are used in biological studies to characterize binding interactions, measure concentrations of analytes, and sort cells. Fluorescent nanoparticle labels offer an excellent alternative to standard fluorescent labeling strategies due to their enhanced brightness, stability and multivalency; however, challenges in functionalization and characterization have impeded their use. This work introduces a straightforward approach for preparation of fluorescent nanoparticle probes using commercially available reagents and common laboratory equipment. Fluorescent polystyrene nanoparticles, Thermo Fisher Scientific FluoSpheres, were used in these proof-of-principle studies. Particle passivation was achieved by covalent attachment of amine-PEG-azide to carboxylated particles, neutralizing the surface charge from − 43 to − 15 mV. A conjugation-annealing handle and DNA aptamer probe were attached to the azide-PEG nanoparticle surface either through reaction of pre-annealed handle and probe or through a stepwise reaction of the nanoparticles with the handle followed by aptamer annealing. Nanoparticles functionalized with DNA aptamers targeting histidine tags and VEGF protein had high affinity (EC50s ranging from 3 to 12 nM) and specificity, and were more stable than conventional labels. This protocol for preparation of nanoparticle probes relies solely on commercially available reagents and common equipment, breaking down the barriers to use nanoparticles in biological experiments.


Determination of PEG density for 40-nm particles
To evaluate how PEG density impacted passivation of 40-nm particles, we varied the PEG:nanoparticle ratio and measured the resulting surface charge neutralization using zeta potential measurements. Based on theoretical models, we reduced the PEG:nanoparticle ratio to approximately 10 5 PEG molecules per nanoparticle and tested several concentrations within that order of magnitude. These studies indicated that there was similar surface charge of -10 mV for each experimental group ( Figure S1). From this, we concluded that 1.4x10 5 PEG:nanoparticle was sufficient for passivating 40-nm nanoparticles.

Binding assessment of varied functional PEG ratios
To evaluate the ratio of functionalized to non-functionalized PEG required for probe attachment and successful detection, we varied the ratio of amine-PEG-azide to mPEG-amine during nanoparticle fabrication. Fluorescent nanoparticle probes with ratios of mPEG-amine:amine-PEG-azide of 75:25, 95:5, 99.5:0.5, 99.95:0.05, and 99.995:0.005 were fabricated, conjugated to B1 probes, and evaluated in plate-based binding studies. The ratio of 75:25 and 95:5 showed similar binding, while the lower ratios showed a decline in on-target binding that correlated with amine-PEG-azide density ( Figure S2). As the amine-PEG-azide contains the reactive group for probe attachment, it is expected that if the DNA aptamer probe-complex is in excess of the amine-PEG-azide, the number of probes per particle will increase with the amine-PEG-azide fraction and thus increase on-target binding. This trend was observed for the ratios of 95:5 and lower. However, it is possible that at higher amine-PEG-azide densities, the aptamer-probe complex is no longer in excess or steric hinderance between adjacent probes impacts probe-target interactions. We hypothesize that one of these two mechanisms was responsible for the only marginal increase in binding between the 95:5 and 75:25 groups. For these proof-ofprinciple studies, we utilized the 95:5 mPEG-amine:amine-PEG-azide ratio for PEGylation of the fluorescent nanoparticle probes to maximize binding efficiency while minimizing cost.

Characterization of the conjugation-annealing handle conjugation
To validate the conjugation of the conjugation-annealing handle to the PEG-azide layer of the fluorescent nanoparticles, we quantified the attachment of a fluorescently-labeled conjugationannealing handle. Using our standard nanoparticle fabrication process, we reacted particles with 125 µM (standard), 12.5 µM, 1.25 µM, and 0 µM of a conjugation-annealing handle that contained a Cy5 dye ("DBCO-Oligo-Cy5"). As a control, nanoparticles were also reacted with 12.5 µM of a conjugation-annealing handle that contained a Cy5 dye but no DBCO functional group ("Oligo-Cy5"). A standard plate reader was used to correlate nanoparticle number with fluorescently labeled oligonucleotide number. Nanoparticles reacted with 0 µM DBCO-Oligo-Cy5 and 12.5 µM Oligo-Cy5 resulted in less than 1 oligonucleotide/nanoparticle, while nanoparticles reacted with 1.25, 12.5, and 125 µM DBCO-oligo-Cy5 resulted in about 1.5, 2.7, and 21oligonucleotides/particle, respectively (Supplementary Figure S3). These results indicate that the attachment of the conjugation-annealing handle to the particles is specific and can be modulated through input concentration. In addition, the ratio of ~20 oligos per 40 nm particle is similar to the ~400 oligos per 200 nm particle observed in qPCR ( Figure 3B) if surface area differences are accounted for. Overall, these results indicate that the conjugation-annealing handle attaches to the particle specifically and the density can be controlled.

Pre-anneal versus post-anneal probe attachment
We evaluated two different approaches for probe attachment to the particle: the "pre-anneal" and "post-anneal" approaches. In pre-annealing, the probe is attached to the fluorescent nanoparticle B1 probe by pre-annealing of conjugation-annealing handle to DNA aptamer probe prior to attachment to the particle. In post-annealing, the conjugation-annealing handle is conjugated to the particle and then the DNA aptamer probe is annealed. The pre-annealing and post-annealing methods were compared through assessment of B1 binding to his-tagged Her2 protein (on-target) and myoglobin (off-target). Pre-annealed probes showed higher binding than the post-annealing approach to on-target proteins (5.5-fold and 2.5-fold binding over background, respectively, Figure S5). Little non-specific binding was observed irrespective of the attachment strategy. It may be possible to optimize the post-annealing approach in studies beyond the scope of this proof-of-principle analysis. First, the concentration of probe could be increased to ensure attachment to all available conjugation-annealing handles. Second, the annealing handle could include a spacer region to reduce steric hinderance due to the PEG layer. Finally, the annealing region could be extended to increase the likelihood of achieving annealing. For the proof-ofprinciple studies described in this work, the pre-annealing method was used.

Binding assessment of alternative conjugation strategies
As our probe annealing attachment technique is unique, we compared on-target binding by probe prepared using our pre-annealing attachment approach to a more standard direct probe conjugation attachment. In the direct conjugation strategy, the 5'-DBCO functional group on the DNA probe was conjugated to the PEG-azide. The original design, using the annealing approach for probe attachment, showed higher binding to its target than the direct DBCO conjugation strategy (4.3-fold versus 2.4-fold binding over background, respectively; Figure S6). This suggest that our annealing approach works as well or better than more standard approaches.

Fluorescent stability of commercially available probes
Our fluorescent nanoparticle probes bound with lower EC50 values than commercially available labels (Figure 6). To investigate if this was due to changes in binding affinity or reduction in fluorescent signal, the brightness of probes was assessed immediately after preparation and at 1 and 2 weeks. Minimal change in fluorescent intensities was observed for the commercially available labels over the course of two weeks (Supplementary Figure S7). This study suggests that the reduction in binding signal observed in the commercially available labels is due to loss of binding affinity, not loss in fluorescent intensity. This provides additional motivation to use the fluorescent nanoparticle probes, as they are more stable.

Relationship between fluorescent intensity and label concentration
The linear range for yellow-green nanoparticles, dark-red nanoparticles, Streptavidin-647 and Streptavidin-APC was determined for the fluorescent plate reader used in binding assay. The linear range was determined to be ~50-18,000 AU for 488/535 and ~100-10,000 for 635/680 em/ex (Supplementary Figure S12). Unless indicated in the figure caption, all data generated in these studies fell within the linear range of the plate reader.

Experimental data replicates
Supplementary Figures S8-11 show additional replicates of the main text figures.

PEG density determination for 40-nm particles
PEGylation was carried out as described in the main text Materials and Methods section "Nanoparticle activation and PEGylation". Particles were resuspended in a 100 mg/mL of a 95:5 mPEG-amine:azide-PEG-amine (MW of PEGs was 2000 g/mol) in PBS at PEG:nanoparticle molar ratios of 1.4x10 5 , 2.1x10 5 , and 2.8x10 5 . Additional sets of 40-nm particles were resuspended in 100 mg/mL of a 75:25, 99.5:0.5, 99.95:0.05, or 99.995:0.005 mPEGamine:azide-PEG-amine ratios. The reactions were incubated at 24 °C with shaking at 800 RPM on a ThermoMixer dry block. After 1 hour, 250 µl PBS was added, samples were washed twice, and resuspended in 500 µl PBS. Zeta potential measurements were obtained as described in the main text Materials and Methods section "Fluorescent nanoparticle probe characterization".

Characterization of conjugation-annealing handle conjugation
Nanoparticle functionalization was carried out according to the method described through PEGylation. After PEGylation, nanoparticles were reacted with 125 µM (standard), 12.5 µM, 1.25 µM, and 0 µM of a conjugation-annealing handle that contained a Cy5 dye ("DBCO-Oligo-Cy5"; /5DBCOTEG/TGTGGAGAGGAAGATGGTA/3Cy5Sp/). As a control, nanoparticles were also reacted with 12.5 µM of a conjugation-annealing handle that contained a Cy5 dye but no DBCO functional group ("Oligo-Cy5"; TGTGGAGAGGAAGATGGTA/3Cy5Sp/). The remainder of the standard protocol was followed. The fluorescent intensities of nanoparticles were measured on a Tecan Spark at em/ex of 488/535 (nanoparticles) and 610/670 (Cy5) in a black 96-well plate. A standard curve of nanoparticles and the fluorescent oligonucleotides was used to determine the number of oligos/nanoparticle.

Alternative probe conjugation strategies
Oligonucleotides were purchased from Integrated DNA Technologies (IDT) for DNA attachment through a conjugation-annealing handle or from Eurofins Scientific for the direct conjugation of the DBCO-B1 probe ( Table 1). The attachment of DNA was achieved by pre-or post-annealing the probe to the particle, as described in the main text Material and Methods section "DNA aptamer probe attachment" (Figure 1A). Direct conjugation of the probe to the PEG layer was performed as a comparison to more standard approaches. For direct conjugation, 40-nm nanoparticles were PEGylated as described in the main text. The DBCO-B1 probe was reacted with azide-PEGylated particles at a 300:1 ratio in PBS at 24 °C overnight with shaking at 800 RPM. Following incubation, 180 µl PBS was added, and samples were washed twice and resuspended in 500 µl PBS. The final sample was stored at 4 °C.

In-house peptide synthesis
Peptides were synthesized on an Intavis Multipep RSi synthesizer. All

Limit of detection for fluorescent plate reader
For each label used in these studies, a titration with two-fold dilutions into PBS was performed in a 96-well black plate. The fluorescent intensities were measured as described in the main text and graphed in Excel. The linear region for each label was determined evaluating the quality of fit after fitting a regression line using the method of least squares in Excel Software.

Fluorescent Nanoparticle Probe Protocol
Buffers: Reaction buffer: 20 mM MES, 500 mM NaCl, pH 6 Wash buffer: PBS; 10 mM Na2HPO4, 1.8 mM KH2PO4, 137 mM NaCl, 2.7 mM KCl, pH 7.4 Nanoparticle Wash: Nanoparticle washes are carried out throughout this protocol and consist of the following steps: 1. Centrifuge nanoparticles at 31,000 xg for 30 minutes prior to PEGylation or 60 minutes post PEGylation in a 1.5-mL tube to form a pellet. 2. Remove supernatant, taking care not to disturb the pellet. 3. Add appropriate buffer and volume to the pellet, as noted in procedure.
4. Redisperse the pellet by pipetting up and down while the nanoparticle tube is partially immersed in a standard laboratory sonication bath (Branson Bransonic Ultrasonic Cleaner 8510R-DTH) until no large aggregates of nanoparticles are visible (usually 10-30 seconds).

Procedure:
This procedure is written for any scale of nanoparticle preparation, but suggested masses, volumes, and concentrations are listed for 40-nm and 200-nm particle preparations used throughout the protocol. 1. Redisperse the stock tube of carboxylate-modified microspheres (see Table 1 for more information) for 10 seconds following step 4 of "Nanoparticle Wash" to ensure nanoparticles are well distributed in solution. a. Dissolve mPEG-amine and azide-PEG-amine (see Table 1 for details) at 100 mg/mL concentration in wash buffer. b. Resuspend particles in 95:5 volume ratio of mPEG-amine:azide-PEG-amine such that PEG:nanoparticle molecular ratio is 3.5x10 7 and 1.4x10 5 for 200-and 40-nm nanoparticles, respectively. Suggested volumes: i. 40-nm particles: 235 µL 100 mg/mL mPEG-amine and 12.7 µL 100 mg/mL azide-PEG-amine. ii. 200-nm particles: 23.5 µL 100 mg/mL mPEG-amine and 1.3 µL 100 mg/mL azide-PEG-amine. 6. Incubate at 24 °C for 1 hour with shaking at 800 RPM on a ThermoMixer dry block. (Note: This incubation step may be allowed to proceed overnight. a. 40-nm particles: 500 µL wash buffer. b. 200-nm particles: 125 µL wash buffer. 9. Combine particles with the conjugation-annealing handle or probe for the pre-anneal or postanneal probe attachment approach. React 125 µM of the DNA aptamer probe-complex or the conjugation annealing handle with particles at a 30,000:1 or 125:1 DNA:nanoparticle molar ratio for 200-and 40-nm particles, respectively: a. Pre-anneal: i. Pre-anneal the conjugation-annealing handle and probe (see Table 3  i. Post-anneal the conjugation-annealing handle (see Table 3 for details) to the particle by adding conjugation annealing handle to nanoparticle pellet and add appropriate volumes of 1x and 10x wash buffer to achieve a final concentration of 1x wash buffer and a final volume 3.3x the initial nanoparticle volume. Incubate at 24 °C for at least 16 hours with shaking at 800 RPM on a ThermoMixer dry block. Suggested volume: 1. 200-nm particles: For a final volume of 20 µl, combine 8.5 µl conjugation-annealing handle, 0.85 µl 10x wash buffer, and 10.7 µl wash buffer. ii. Wash particles and redisperse in wash buffer equal to 28x the initial particle volume twice. Suggested volume: 1. 200-nm particles: 180 µL. iii. Wash particles. Redisperse in a solution of 125 µM DNA aptamer probe at equal molar ratio to the conjugation-annealing handle. Bring to 3.3x the initial particle volume using water and incubate at 95 °C for 5 minutes. Centrifuge briefly to remove condensation from the tube lid. Incubate on benchtop for 10-15 minutes to allow the two complementary regions of DNA to anneal. Suggested volumes: 1. 200-nm particles: Combine 8.5 µl of 125 µM probe and 11.5 µl DNase/RNase free water. 10. Add 1 µL wash buffer/µL PEG solution for 40-nm particles and 7 µL wash buffer/µL PEG solution for 200-nm particles. Suggested volumes: a. 40-nm particles: 250 µL wash buffer. b. 200-nm particles: 180 µL wash buffer. 11. Wash and redisperse particles in 2 µL wash buffer/µL PEG solution PBS for 40-nm particles and 5 µL wash buffer/µL PEG solution for 200-nm particles. (Note: The pellet will be less defined, and some nanoparticles may remain in solution after this step.) Suggested volumes: a. 40-nm particles: 500 µL wash buffer. b. 200-nm particles: 125 µL wash buffer. 12. Repeat Step 11. Store nanoparticles in wash buffer at 4 °C.

PEG Density Calculations
The density of PEG-36 (MW of 1.6 kDa) required to achieve brush layer conformation was determined utilizing theoretical and experimental models [2][3][4] . The PEG conformation depends upon two parameters: 1) the Flory radius (RF), which is the radius of the PEG coil and is dependent upon molecular weight and 2) the distance between PEG molecule graft sites (D). The relationship between RF and D dictates the PEG conformation: If D > RF, the PEG layer will be a mushroom conformation; if D < RF, it will be a brush layer; and if D < 0.36 RF it will be a dense brush layer [2][3][4] . RF can be calculated using the following equations 2-4 : RF = αN 3/5 where α = the length of one monomer and N = the number of monomers/polymer chain.

Assumptions:
• D = distance between PEG molecules • PEG distance to achieve a dense brush layer: < (0.36) , < 1.08 • PEG distance as a function of PEG spacing: Distance between two PEG molecules:          AlexaFluor 647 Conjugate, and B1 Streptavidin APC Conjugate against HHH targets. These binding curves were used to create the box and whiskers plot shown in Figure 6A, and one representative graph per probe from this panel is shown in Figure 6A.