A Universal Approach to Prepare Reagents for DNA-Assisted Protein Analysis

The quality of DNA-labeled affinity probes is critical in DNA-assisted protein analyses, such as proximity ligation and extension assays, immuno-PCR, and immuno-rolling circle amplification reactions. Efficient, high-performance methods are therefore required for isolation of pure conjugates from reactions where DNA strands have been coupled to antibodies or recombinant affinity reagents. Here we describe a universal, scalable approach for preparing high-quality oligonucleotide-protein conjugates by sequentially removing any unconjugated affinity reagents and remaining free oligonucleotides from conjugation reactions. We applied the approach to generate high-quality probes using either antibodies or recombinant affinity reagents. The purified high-grade probes were used in proximity ligation assays in solution and in situ, demonstrating both augmented assay sensitivity and improved signal-to-noise ratios.


Introduction
DNA-assisted protein analysis technologies play an increasing role in analyses of proteomes [1], and for disease diagnosis, prognosis [2] and personalized medicine [3]. For optimal performance of such analyses, high-quality and pure DNAcoupled affinity binders are required. A broad variety of methods are available for protein-DNA conjugation, in order to functionalize antibodies for DNA-assisted protein analysis. Commonly, biotinylated antibodies are coupled with oligonucleotide-modified streptavidin molecules. Alternatively, bifunctional cross-linkers can be used to covalently link antibodies to oligonucleotides with suitable chemical modifications. For instance, primary amines of antibodies can be reacted with dibenzylcyclooctyne-NHS esters to allow copper-free click chemistry to attach azide-modified oligonucleotides to the alkyne-functionalized antibodies [4]. Recently, the growing availability of new classes of affinity binders prepared by recombinant techniques, such as recombinant antibody fragments [5], affibodies [6] and designed ankyrin repeat proteins (DARPins) [7], greatly increases opportunities for DNAassisted protein analysis. By expressing recombinant binders in fusion with SNAP protein domains, O 6 -benzylguanine (BG)modified oligonucleotide can be covalently coupled to SNAP domains of the affinity reagents [8]. The use of suitably modified recombinant affinity reagents has the virtue that oligonucleotides can be attached to the recombinant binder at a specific site, remotely from the target binding area, and at a one-to-one (or any other desired) stoichiometry.
In the various conjugation methods, the efficiency of functionalization of affinity binders is usually maximized by employing a large excess of suitably modified oligonucleotides, thereby generating a mixture of the desired conjugates and an excess of free oligonucleotides, along with any remaining unconjugated affinity binders. In DNA-assisted protein analysis, remaining free oligonucleotides are liable to cause target-independent detection signals, increasing assay background. This is particularly deleterious in homogeneous assays where no washing steps are applied, such as in solution phase proximity ligation assays (PLA) [9,10] or proximity extension assays (PEA) [11]. Any unconjugated binders in the detection reactions, be it antibodies or alternative scaffolds, also tend to compromise assay performance by competing with properly functionalized affinity reagents for binding to the respective targets, resulting in reduced detection signals. Present protocols for purifying conjugates may remove either free oligonucleotides or antibodies, but there is a need for a general and convenient procedure to avoid both contaminating species in one simple procedure [12].
We present in this study a general and efficient approach for purification of high-quality DNA-coupled affinity binders from free oligonucleotides and remaining unconjugated natural antibodies or alternative binding scaffolds. Alkyne-modified antibodies and azide-modified oligonucleotides were conjugated by click chemistry, and recombinant DARPin affinity reagents were coupled with BG-activated oligonucleotides via SNAP domains. The conjugation reaction mixtures were subjected to solid-phase purification, where unconjugated binders and excess oligonucleotides were removed in two sequential steps. The purified DNAlabeled affinity reagents were then used in solution phase and in situ PLA to validate the effect of the conjugation and purification protocol on sensitivity and signal-to-noise for the two assays.

Affinity binders, recombinant proteins and oligonucleotides
Goat anti-human IL8 (AF-208-NA) and recombinant human IL8 (208-IL-010) were purchased from RnD Systems. Anti-HER2 DARPins 9.01 and G3, in fusion with both an N-terminal RGS His-tag and a C-terminal SNAP domains, referred to as 9.01-SNAP and G3-SNAP, respectively, were expressed and prepared as described previously [13]. Oligonucleotides with suitable modifications (Table S1) were purchased from Integrated DNA Technology.

Cell culture
The HER2-expressing human ovarian cancer cell line SK-OV-3 ((HTB-77; ATCC)) was grown in RPMI1640 medium, and the human fibroblast cell line BJhTERT was grown in Minimum Essential Medium (MEM). Both media were supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine and 1% penicillin-streptomycin (all reagents from Sigma-Aldrich). The cells were cultured at 37uC in a humidified 5% CO 2 environment.
Preparation of antibody-DNA conjugates  activated antibodies were purified over a Zeba spin desalting column (7K MWCO, Thermo Scientific) to remove unreacted DBCO-NHS ester. After purification, the activated antibodies were mixed with a 4-fold molar excess of the azide modified oligonucleotides (Arm1_long, Arm2_long, Arm1, or Arm2, Table  S1) and incubated overnight at 4uC.

Preparation of DARPin-DNA conjugates
Thirty ml of 200 mM 59-aminohexyl modified oligonucleotides (S3primer_long, S3block_long, S3primer and S3block, Table S1) were incubated for 2 h at 37uC with a 30-fold molar excess of an amino-reactive bifunctional cross-linker to functionalize the oligonucleotides with a benzylguanine (BG) moiety (BG-GLA-NHS (S9151S, NEB); 18 ml of a 10 mM solution BG-GLA-NHS freshly dissolved in DMSO). Excess BG-GLA-NHS was removed by separation through two 7K MWCO Zeba spin desalting columns, and the purified BG-modified oligonucleotides were collected. The DARPins 9.01-SNAP and G3-SNAP were reduced to activate the cysteine residue within the SNAP domain in 20 mM DTT for 1 h at 37uC, separately. The reduced DARPins were then mixed with a 3-fold molar excess of the purified BGmodified oligonucleotides, and incubated for 1 h at 37uC.

Two-step purification of conjugates
To remove unconjugated protein, biotinylated capture oligonucleotides were added at a 2-fold molar excess over the partially complementary conjugation oligonucleotides (Arm1Capture for Arm1_long, Arm2Capture for Arm2_long or S3Capture for both S3primer_long and S3block_long) to the respective antibody-or DARPin-oligonucleotide conjugates, followed by incubation for 30 min at RT. The solution was then incubated for 30 min at RT with Streptavidin Sepharose High Performance (17-5113-01, GE Healthcare) using 150 nmol biotinylated capture oligonucleotides per milligram beads. After two washes with PBST (PBS containing 0.05% Tween 20) to remove unconjugated protein, 50 ml of enzymatic elution solution (16 NEB buffer 4, 1 mg/ml BSA, and 1 U/ml of the restriction enzyme MlyI (R0610S, NEB)) was added to the streptavidin Sepharose and incubated overnight at 37uC.
Next, for antibody-DNA conjugates, the supernatant from the MlyI cleavage was incubated with Dynabeads Protein G (10004D, Life Technologies) for 1 h at RT, followed by two washes with PBST to remove unconjugated oligonucleotides. The antibody-DNA conjugates were then eluted by incubating with 50 mM glycine-HCl (pH 2.5) for 5 min at RT and the eluate was immediately adjusted to pH 7 with Tris buffer (pH 8.2).
For DARPin-DNA conjugates, the supernatant from the MlyI cleavage was incubated in phosphate buffer (0.1 M phosphate, 0.5 M NaCl and 0.05% Tween 20, pH 8.3) with DynabeadsH His-Tag Isolation and Pulldown (10103D, Life Technologies) for 1 h at RT. The DARPin-DNA conjugates were washed three times in PBS to remove unconjugated oligonucleotides, and then eluted with 0.1 M imidazole in PBS for 5 min at RT.

Readout and data analysis of solution phase PLA
The results of solution phase PLA with real-time PCR readout were recorded using the MxPro software (Stratagene) and the cycle of threshold (Ct) was exported to Excel 2007 (Microsoft). The plots of data were either prepared in Excel 2007 (Microsoft) or using the statistical computing environment ''R'' (Bell Laboratories). The LOD was defined as the concentration of protein corresponding to a Ct value two standard deviations above the average assay background. The amplification template numbers N were calculated from Ct values using N = 2 402Ct , assuming that the Ct value for one molecule of template was 40. Coefficients of variation (CV) were calculated as CV% = StDev (N)/mean (N)6100%, where StDev and mean is the standard deviation and mean of quadruplicates of each sample.
In situ PLA and immuno-RCA Ten thousand SK-OV-3 cells were seeded per well on Lab-Tek Chamber Slides (154534, Thermo Scientific), and cultivated overnight. Cells were fixed with 3.7% paraformaldehyde for 30 min and permeabilized in PBS containing 0.1% Triton X-100 for 15 min. To block unspecific binding of the probes 16blocking buffer (Olink Bioscience) was added to the cells and incubated for 1 h at 37uC. The blocking buffer was replaced by 40 ml of the purified and unpurified probes (20 nM of each purified 9.01-SNAP-S3primer_long and G3-SNAP-S3block_long conjugates for PLA; 5 nM of purified G3-SNAP-S3block_long conjugates and unpurified G3-SNAP-S3block conjugates for immuno rolling circle amplification (RCA)), diluted in the diluent buffer (Olink Bioscience). The probes solution was incubated with the cells for 1.5 h at 37uC, followed by two washes with washing buffer (DUO82049, Olink Bioscience). In situ PLA detection was then completed using the Duolink in situ detection kit Orange (excitation 554 nm/emission 576 nm, DUO92007, Olink Bioscience). For immuno-RCA the incubation with G3-SNAP conjugates was followed, after two washes, by a step where a 40 ml ligation mix (40 mM Tris-HCl, 10 mM MgCl 2 , 0.1 mM ATP, 100 nM S3 padlock oligonucleotide and 0.02 U/ml T4 DNA ligase (Thermo Scientific)) was added for a 30 min incubation at 37uC. After ligation and two washes, 40 ml amplification and detection mix (Duolink in situ detection kit Orange, DUO92007, Olink Bioscience) was applied to complete the immuno-RCA detection. Next, all slides were incubated with 1.25 mg/ml Hoechst 33342 (Life Technologies) for 20 min at RT to counterstain cell nuclei. The slides were then washed twice for 10 min in 16 buffer B (DUO82049, Olink Bioscience), followed by a 1 min wash step in 0.016 buffer B and mounting with Vectashield (H-1000, Vector Labs). In parallel, the cells on slides were analyzed with in situ PLA using 20 nM of pure G3-SNAP and 9.01-SNAP probes spiked with 20 nM of unconjugated G3-SNAP and 9.01-SNAP. The cells were also detected with immuno-RCA using 5 nM of G3-SNAP probe spiked with unconjugated G3-SNAP at different concentrations (40, 80, 160 or 320 nM).

Imaging and data analysis
Cell images were acquired using an Axioplan 2 epifluorescence microscope (Zeiss), equipped with a 100 W mercury lamp, a cooled CCD camera (AxioCam HRm, Zeiss), a computercontrolled filter wheel with excitation and emission filters for visualization of DAPI (excitation 350 nm/emission 470 nm) and Cy3 (excitation 550 nm/emission 570 nm), and a 206 objective (Plan-Apochromat 440640-9903, Zeiss). Images were collected as compilations of Z-stacks and processed with AxioVision software (version 4.8, Zeiss). Image analysis and quantification of RCA products were conducted with the Duolink Image Tool (DUO90806, Olink Biosciences).

Results and Discussion
The general approach to prepare high-grade DNA-labeled affinity binders is schematically illustrated in Figure 1, exemplified with the preparation of antibody-oligonucleotide conjugates. First, the conjugation was performed in solution by mixing DBCOmodified antibodies with a 4-fold molar excess of azide-modified oligonucleotides. After incubation, the resulting mix containing antibody-DNA conjugates and remaining free antibodies and oligonucleotides, was incubated with biotinylated capture oligonucleotides at 2-fold molar excess over conjugation oligonucleotides, and the products were immobilized on streptavidin-coated Sepharose beads. After washes to remove unconjugated antibodies, the captured antibody-DNA conjugates along with unconjugated oligonucleotides were both released by enzymatic cleavage using the restriction enzyme MlyI that cleaves at the border of the complementary region between the capture and conjugate oligonucleotides (Fig. 1b) to release single-stranded DNA-coupled conjugates. The eluate was then captured on protein G beads and washes were conducted to remove any free oligonucleotides having no antibodies attached. Finally, purified conjugates were eluted from the protein G beads by lowering the pH.
In order to monitor the purification and MlyI cleavage steps, the antibodies were also conjugated separately to two shorter oligonucleotides (Arm1 and Arm2) corresponding precisely to the cleaved form. When these conjugates with shorter oligonucleotides were electrophoresed, they resulted in bands of identical molecular weight as the conjugates released by restriction digestion.
The conjugation reaction mixtures, the eluates after enzymatic cleavage, and the purified conjugates were all analyzed by polyacrylamide gel electrophoresis and visualized with silver staining (Figure 2a). The conjugation reaction mixtures (lanes 6 and 7) resulted in a smear of bands similar to and larger than unconjugated antibodies (around 150 kDa, lane 5) and bands at the bottom of the gel corresponding to sizes of unconjugated DNA oligonucleotides (lane 3 and 4). After MlyI cleavage and elution, the cleaved excess oligonucleotides in the eluate from antibodies conjugated to Arm1_long (lane 8) migrated similarly to Arm1 (lane 4), which corresponds to the length of the cleaved products. This indicates that the MlyI cleavage works efficiently on the solid support. Under the native conditions that the samples were electrophoresed, the migration of conjugated antibodies in lanes 8 and 9 (around 170 kDa or larger) was only slightly slower than that of unconjugated antibodies (around 150 kDa, lane 5). After the whole purification protocol (Fig. 1) the oligonucleotides observed in the reaction mixture at the bottom of lane 8 were completely removed, and pure antibody-DNA conjugates were obtained and visualized (lane 9). The yield for antibody conjugates after the two-step purification was approximately 20% of the starting antibody input.
To investigate if a variation of the purification method would also be suitable for recombinant affinity reagents, a procedure similar to the one described above was performed for DARPin-SNAP constructs (HER2-specific G3-SNAP and 9.01-SNAP) [14,15]. DARPins were expressed in fusion with a His-tag and a SNAP protein domain, allowing the convenient conjugation and purification of the DARPins. After reduction, the reactive cysteine residue within the SNAP domain readily reacts with benzylguanine (BG) to generate a covalent bond while releasing a guanine group [8]. The DARPin-SNAP constructs were first reduced with DTT and then reacted with BG-modified oligonucleotides to form DARPin-SNAP-DNA conjugates in solution [13]. To remove any remaining unconjugated protein, DARPin-SNAP-DNA conjugates were immobilized by hybridization to biotinylated oligonucleotides bound to streptavidin-coated Sepharose beads, in a manner similar to that used for the purification of antibody-DNA conjugates, and unconjugated DARPins were washed away. After restriction digestion with MlyI, the conjugated protein was bound to cobalt-chelating Dynabeads for his-tag isolation, and excess free oligonucleotides were then removed by washing. The pure DARPin-SNAP conjugates were finally eluted in a yield of around 60% out of the starting material.
To monitor the purification, the unpurified G3-SNAP conjugation reaction mixtures, eluates after enzymatic digestion, and the pure G3-SNAP conjugates were subjected to electrophoresis (as described above) and visualized with protein silver staining ( Figure 2b). G3-SNAP conjugates with S3 block_long oligonucleotides (approximately 55 kDa, lane 6) and S3 block oligonucleotides (approximately 50 kDa, lane 7) showed intense bands at positions expected for the conjugates, in comparison to the unconjugated G3-SNAP (35 kDa, lane 5). Despite the use of a three-fold molar excess of oligonucleotides in the conjugation reaction, some remaining unconjugated G3-SNAP was nonetheless seen in the conjugation reaction mixture of G3-SNAP and S3 block_long (Fig. 2b, lane 6) and of G3-SNAP and S3 block (Fig. 2b, lane 7). The excess of oligonucleotides was also observed at the bottom of the gel (lane 6 and lane 7). After MlyI digestion, the eluate (lane 8) was visualized as two bands, corresponding to the conjugate and excess free S3 block oligonucleotides (which had been cleaved from excess S3 block_long). This demonstrated reliable cleavage by the MlyI restriction enzyme and efficient removal of unconjugated G3-SNAP. Finally, as a result of the twostep purification, a single, clean band of pure G3-SNAP-DNA conjugates was seen (lane 9), confirming that both free oligonucleotides and unconjugated G3-SNAP had been removed.
Homogenous -washing-independent -proximity ligation and extension reactions in solution have proven useful to analyze proteins in biofluids [9,11]. Solution phase PLA and PEA require 1 ml of sample, and detection depends on dual recognition by oligonucleotide-modified affinity reagents, followed by DNA amplification via quantitative PCR. By suitable design of the DNA strands, it is possible to avoid detection signals from noncognate pairs of antibodies or affinity reagents, thus avoiding increasing problems with cross-reactivity upon multiplexing. Accordingly, assays are available for up to 96 proteins in the same reaction, with performance similar to that of individual assays [16,17]. Since the homogeneous proximity reactions (PLA and PEA) do not involve any washing steps, it is necessary to use high-quality probes where both unconjugated antibodies and free oligonucleotides have been removed, as they would otherwise be expected to compromise the performance of solution phase reactions.
To assess the effect of purifying the conjugates, we compared the performance of pure and unpurified conjugates (probes) in assays for detection of human interleukin 8 (IL8) protein using solution phase PLA. The reactions were performed in PLA buffer alone or with 10% or 50% chicken serum. The purified conjugates resulted in a limit of detection (LOD) of 0.06 pM, which represents an approximately 16-fold increased sensitivity of detection over the unpurified conjugates (0.97 pM) (Figure 3a, Table 1). In complex samples with 10% or 50% chicken serum, solution phase PLA probed with pure conjugates demonstrated robust and reproducible measurement of IL8 protein (Figure 3b, Table 1), and the chicken serum did not impair the LOD of the assay or its linear range compared to the corresponding assay in PLA buffer ( Figure 3b).
Next, we sought to determine to what extent unconjugated antibodies or free oligonucleotides would interfere with the performance of PLA in solution. We assayed samples containing 10 pM IL8 in PLA buffer using purified conjugates to which different amounts of unconjugated antibodies specific for IL8 or unconjugated oligonucleotides (Arm1 and Arm2) had been added. Unconjugated antibodies at a concentration of 480 pM in assays using 60 pM conjugates caused a 2-fold decrease in detection signals (Figure 3c). This is in accordance with the expectation that unconjugated antibodies would compete with the PLA probes for binding to their cognate targets and thus reduce detection efficiency. Free oligonucleotides spiked at a final concentration of 30 pM into detection reactions with 100 pM IL8 increased assay background by approximately two Ct units (Figure 3d), indicating that free oligonucleotides in the reactions cause an increased assay background and lead to a decreased signal-to-noise ratio by around 4-fold at these concentrations.
We also investigated the effects of impure probes on in situ protein detection reactions. The purified DARPin-DNA conjugates were used alongside an unpurified DARPin-DNA conjugate in immuno-RCA to detect HER2 proteins in fixed SK-OV-3 cells abundantly expressing HER2 protein [18] and BJhTERT cells lacking the protein [19]. At a probe concentration of 5 nM the pure G3-SNAP conjugates resulted in a 3-fold lower background than the unpurified G3-SNAP conjugates in BJhTERT cells (Figure 4a and b). The increased background from unpurified probes in BJhTERT cells probably resulted from any free oligonucleotides that may resist washes. The signal-to-noise ratio, the number of RCA products in SK-OV-3 cells divided by those in BJhTERT cells, was lower in assays using unpurified probes compared to those using purified conjugates (Figure 4a and b). This is likely a consequence both of unconjugated affinity reagents (G3-SNAP) blocking target molecules and free oligonucleotides resistant to washes that may contribute to nonspecific signals. In order to assess the effect of free affinity reagents in these assays, G3-SNAP was added back to the purified G3-SNAP conjugates that were used both in immuno-RCA and in situ PLA. As shown in Figure 4c, the presence of 40 nM free G3-SNAP greatly reduced detection signals of HER2 proteins in SK-OV-3 cells. At a concentration 64-fold higher than that of the conjugates, unconjugated DARPins abolished detection of HER2 proteins in SK-OV-3 cells (Figure 4c). The numbers of detection signals for HER2 proteins using in situ PLA decreased by a factor of four when the purified conjugates of either G3-SNAP or 9.01-SNAP were mixed with an equal concentration of the corresponding free affinity reagents (Figure 4d), underlining the importance of purified detection reagents.
In summary, we have demonstrated a generalized and robust approach for preparing high quality oligonucleotide-modified affinity reagents, using either antibodies or recombinant binders. The obtained oligonucleotide-conjugated affinity reagents exhibited greatly improved performance in protein analysis using DNAassisted protein detection reactions. The two-step purification on solid supports is suitable for scalable, high-throughput preparation of high-grade oligonucleotide-conjugated antibody-based or recombinant affinity reagents.