The Argi system: one-step purification of proteins tagged with arginine-rich cell-penetrating peptides

The discovery of cell penetrating peptides (CPPs) opened new perspectives for the delivery of proteins into human cells. It is considered that in the future CPP-mediated transport of therapeutic proteins may find applications in the treatment of human diseases. Despite this fact a fast and simple method for the purification of CPP-tagged proteins, free of additional tags, was not available to date. To fill this gap we developed the Argi system for one-step purification of proteins tagged with arginine rich CPPs.

CPP-tagged therapeutics and to reduce the costs of CPP-tagged protein production in the future, we present here the development of the Argi system -a new affinity chromatography tool for one-step purification of arginine-rich CPP-tagged proteins. The presented system is based on the interaction between a DNA aptamer and arginine-rich CPPs.

Results
Selection of an arginine-rich CPP-binding aptamer. To create the Argi system we employed systematic evolution of ligands by exponential enrichment (SELEX 15,16 ). A DNA aptamer characterized by specific and reversible binding to poly-arginine and the Tat 49-57 peptide, fused either to the N-or C-terminus of the tested protein, under mild buffer conditions, was identified. The selection strategy was similar to the one we employed previously 17 . We used a peptide containing eight consecutive arginines (H 6 G 4 R 8 ) as a selection target during the first and second round of SELEX. Next, from the third to the eighth selection round R 8 -PCNA was used as a target. The progress of selection was evaluated using qPCR with R 8 -PCNA as a target (Fig. 1). The aptamer pool after the eighth selection round was cloned and sequenced. Among the 50 analyzed clones we found 12 different aptamer sequences. Next, using again qPCR the binding of single aptamers to R 8 -GST was evaluated and compared. The highest value of the enrichment parameter was observed for the aptamer sequence number 2 called 24-10 ( Fig. 2) which was selected for further studies. The specificity of this molecule was verified and confirmed (Fig. 3).
To test whether the 24-10 aptamer can be used to capture proteins tagged with either the R 6 -tag, R 8 -tag or Tat 49-57 -tag we analyzed and compared the binding of the following proteins, GFP, GFP-R 6 , GFP-R 8 , GFP-Tat 49-57 , GST, R 6 -GST, R 8 -GST, Tat 49-57 -GST, PCNA, R 6 -PCNA, R 8 -PCNA and Tat 49-57 -PCNA, to the 24-10 or reference  . qPCR analysis of selected aptamer binding to R 8 -GST. The enrichment of single aptamers was expressed as the ratio: binding of tested aptamer/binding of reference aptamer to R 8 -GST. Results are the mean of three measurements. Error bars represent the standard deviation. aptamer immobilized to streptavidin-agarose resin (High Capacity Streptavidin-agarose, HCSA). In line with our expectations, this experiment showed that R 6 -, R 8 -and Tat 49-57 -tagged proteins could be successfully captured only when the 24-10 aptamer was used (Fig. 4, Supplementary Fig. 1).
Identification of the 24-10 aptamer region necessary for arginine-rich CPP binding. To determine the region of the 24-10 aptamer crucial for binding of the studied arginine-rich CPP tags, R 8 -GST was used as a model. With the help of the pull-down assay the binding of R 8 -GST to variants of the 24-10 aptamer shortened from either the 5′ or 3′ end, as well as the full length aptamer, was evaluated based on densitometric analysis (Fig. 5A,B). This experiment showed that the sequence from nucleotides 31 to 70, referred to as the AR aptamer (5'-CTTTGTAATTGGTTCTGAGTTCCGTTGTGGGAGGAACATG-3'), could bind the R 8 -tag the most efficiently. The modeling of the AR aptamer with mfold 18 software predicted the formation of four possible structures (Fig. 6). The calculated folding ΔG for structures A, B, C and D was −10.34, −9.86, −9.76 and −9.43 kcal/mol, respectively.

Analysis of elution conditions for arginine-rich CPP-tagged proteins.
Among the key features which should characterize affinity chromatography systems, used for recombinant protein purification, the efficient binding of the analyte and its simple elution under mild buffer conditions should be considered. The demonstrated specific interaction between the AR aptamer and tested arginine-rich tags, although crucial, was insufficient to fully assess the possible application of the developed ssDNA for chromatography purposes. This is why in the following experiment the conditions of analyte elution from the AR aptamer were analyzed. Testing buffers supplemented with increasing amounts of guanidine hydrochloride (GuHCl) we showed that to elute more than 90% of R 6 -, R 8 -and Tat 49-57 -GST bound to the AR aptamer, 500, 600 and 200 mM GuHCl, respectively was necessary ( Fig. 7A-C). Despite the fact that GuHCl is widely used as a protein denaturant it is well known that at mM concentrations it is usually not harmful but can stabilize proteins 19 .
Determining the dissociation constant of AR aptamer/arginine-rich CPP complexes. To evaluate the affinity between the developed ssDNA and the studied tags the dissociation constant (Kd) of AR aptamer/arginine-rich tag complexes was determined. The Kd, determined using isothermal titration calorimetry, for the AR aptamer and R 6 -GST, R 8 -GST and Tat 49-57 -GST was 943 ± 37 × 10 −9 M, 532 ± 55 × 10 −9 M and 893 ± 44 × 10 −9 M (Fig. 8), respectively. All reactions were exothermic and enthalpy-entropy driven ( Table 1). The analysis of stoichiometry revealed that one molecule of AR-aptamer can bind two molecules of the tested tags.
Evaluation of AR aptamer application for purification of CPP-tagged proteins. Finally, when the biochemical studies of the Argi system were completed, to test the developed tool in real conditions the purification of nine different protein variants (GFP-R 6 , GFP-R 8 , GFP-Tat 49-57 , R 6 -GST, R 8 -GST, Tat 49-57 -GST, R 6 -PCNA, R 8 -PCNA and Tat 49-57 -PCNA) from E. coli total protein extract was performed and analyzed. These experiments clearly confirmed the applicability of the developed AR aptamer for the purification of recombinant proteins tagged with arginine-rich peptides (Fig. 9). Moreover, we found that using 175 μg of AR aptamer we could purify from 145 to 594 μg of protein (Supplementary Table 1). Additionally, analyzing the reproducibility of the Argi system we found that during nine purification cycles the quality of isolated proteins was not significantly affected ( Fig. 10).

Discussion
In this study we presented the development of the Argi system which is a powerful tool for one-step purification of recombinant proteins tagged with arginine-rich cell-penetrating peptides overexpressed in E. coli cells. Besides the application of poly-arginine peptides for cargo transport, in 1984 the utilization of this tag for the purification of recombinant proteins was proposed 20 , but has not found a wide application in laboratories. The purification of arginine-rich peptide-tagged proteins was based on the use of non-specific ion exchange chromatography. However, the pH of the buffers used during purification was far from the pH range which is usually optimal for protein stability and activity. The applied buffer conditions could possibly be harmful for the biological activity of many proteins. On the other hand other studies demonstrated that the poly-arginine tag could improve protein solubility 21 . Despite previous reports, an affinity chromatography system for the purification of poly-arginine-tagged proteins was never developed. Therefore, the presented new chromatography tool offers a much easier method of arginine rich CPP-tagged protein purification, which was not available up to date.  5), protein with R 6 tag (lanes 2, 6), protein with R 8 tag (lanes 3, 7), protein with Tat 49-57 tag (lanes 4,8). After washing, the protein sample was denatured and half of the protein sample volume of the bound protein was separated on 12% SDS-PAGE gel followed by Coomassie brilliant blue staining. This is one of three independent experiments which is representative.

Materials and Methods
Experimental. Chemicals and plasmids. The chemicals used in this work were obtained from Sigma-Aldrich (USA) and Merck (Germany) unless indicated otherwise. Synthetic aptamers, the ssDNA library and reference aptamer (5′-CATGCTTCCCCAGGGAGATGACTGACTGACTGACTGACTGACTGACT GACTGACTGACTGGAGGAACATGCGTCGCAAAC-3′) were obtained from IBA GmbH (Germany). The template plasmids pDNR-LIB (coding for human PCNA), pGEX4T-1 (coding for GST) and pK7WGF2 (coding for GFP) were purchased from ImaGenes GmbH, GE HealthCare and VIB Department of Plant Systems Biology, University of Ghent, respectively.
Protein immobilization. The H 6 G 4 R 8 peptide (LifeTein, USA) was dissolved in DMSO to a concentration of 1 mg/mL. Next, 10 μg of the peptide was mixed with 0.5 μL 50% cobalt-coated agarose beads (TALON, Clontech) and the mixture was incubated in AS1 buffer (136 mM NaCl, 12 mM KCl, 10 mM Na 2 HPO 4 , 1.7 mM KH 2 PO 4 , 4.9 mM MgCl 2 , 0.01% (v/v) Tween 20; pH 7.5) for 1 h at RT with continuous mixing. To remove the unbound peptide, after immobilization the beads were washed 3x with AS1 buffer. Immobilization of R 8 -GST, GST or The bound protein was denatured. Next, one fortieth of the volume of each protein sample was separated on 12% SDS-PAGE followed by Coomassie staining. This is one of three independent experiments which is representative. (B) Densitometric analysis of data from three independent pull down assays was done using Multispectral Imaging System IMAGER with Launch VisionWorksLS. Results are the mean of three measurements. The error bars represent standard deviation. The results were normalized relative to signal from the protein sample bound by the full length aptamer (100%). R 8 -PCNA to CNBr-activated Sepharose 4B (Pharmacia) was performed according to the protocol supplied by the manufacturer.
Cloning and plasmid construction. Nucleotide sequences coding for GFP-R 6 /R 8 /Tat 49-57 , R 8 /Tat 49-57 -GST and R 8 /Tat 49-57 -PCNA were amplified using sets of specific primers (Supplementary Table 2) and cloned into the pET29a vector as described previously 17 with minor modifications. The following PCR conditions were used: preliminary denaturation at 94 °C for 2 min, 30 cycles of amplification (94 °C for 40 s, 58 °C for 30 s, 72 °C for 3 min), final incubation at 72 °C for 7 min. The PCR products were purified, digested with BamHI and NdeI (FastDigest, Thermo Scientific) restriction enzymes and cloned into the pET29a expression vector using T4 DNA ligase (Thermo Scientific), followed by sequencing. The construction of E. coli expression vectors carrying PCNAand GFP-coding sequences was previously described 17 . Open reading frames coding for R 6 -GST and R 6 -PCNA were synthetized by Genscript.
Protein purification. All the purification steps were performed at 4 °C similarly as described previously 17 . The E. coli cells containing all variants of GFP, GST and PCNA were resuspended in GFP buffer (50 mM Tris-HCl, 3 M NaCl, pH 8.0), PBS(140 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 ,1.8 mM KH 2 PO 4 , pH 7.3) buffer and PCNA buffer (50 mM Tris-HCl, pH 7.6), respectively. Phenylmethylsulfonyl fluoride (PMSF) was then added to 1 mM Figure 7. Determination of optimal GuHCl concentration for analyte elution. 5′-biotinylated AR aptamer bound to streptavidin agarose resin was incubated with GST fused with the indicated tag. After washing, the bound protein was eluted using Tris-HCl buffer supplemented with the appropriate GuHCl concentration. Next, the protein samples were denatured and one fortieth of the volume of each sample was separated on 12% SDS-PAGE followed by Coomassie staining and analyzed using Multispectral Imaging System IMAGER with Launch VisionWorksLS. The results were normalized relative to signal from the protein sample eluted with buffer supplemented with 0 mM GuHCl (0% eluted protein). Recombinant GST protein tagged with: (A) R 6 , (B) R 8 and (C) Tat 49-57 peptide. Each panel presents the results of one of three independent experiments which is representative.
Scientific RepoRts | 7: 2619 | DOI:10.1038/s41598-017-02432-6 final concentration. Next, the samples were sonicated (10 min, 5 s pulses, 10 s break) and centrifuged (30,000x g for 30 min) to obtain E. coli total protein extracts. Next, the proteins were purified as described below. At the final step of protein purification the fractions containing the protein of interest were pooled, dialyzed against a S buffer (50 mM Tris-HCl, 150 mM NaCl, 15% (v/v) glycerol, pH 7.5), frozen in liquid nitrogen and stored at −80 °C until use.
GFP and GFP-R 6 /R 8 / Tat 49-57 purification. The protein extract was loaded onto a 5 mL HiTrap Butyl HP column (GE Healthcare) equilibrated with GFP buffer. The unbound proteins were washed out with GFP buffer. Next, the bound proteins were eluted using 200 mL of linear gradient from 3-0 M NaCl. To obtain the satisfactory protein purity, the fractions containing the protein of interest were dialysed against PCNA buffer and loaded onto a 5 mL High Q (Bio-Rad) column. After washing with PCNA buffer the bound proteins were eluted using 200 mL of linear gradient from 0-0.8 M NaCl in PCNA buffer.
GST and R 6 /R 8 / Tat 49-57 -GST purification. The protein extract was loaded onto a 2 mL Glutathione Sepharose 4B column (GE Healthcare) equilibrated with PBS buffer. The unbound proteins were washed out with PBS buffer. Next, the bound protein was eluted with GST elution buffer (50 mM Tris-HCl, 50 mM reduced glutathione, pH 8.0).
PCNA and R 6 /R 8 / Tat 49-57 -PCNA purification. The proteins were purified according to the protocol described previously for PCNA and His 3 -PCNA 17 . SELEX procedure. The SELEX procedure was performed as described previously 17 with some modifications. AS1 buffer was used for aptamer selection. The H 6 G 4 R 8 peptide bound to TALON cobalt-coated agarose beads (Clontech), which were shown previously to provide good selection power 22 , was used for selection rounds I and II. For selection cycles III to VIII R 8 -PCNA, immobilized to to CNBr-activated Sepharose 4B (Pharmacia), was used as a target molecule.
Quantitative real-time PCR. qPCR analysis of aptamer binding, enrichment and specificity was performed as described previously 17 . R 8 -PCNA was used as a target when the enrichment parameter was analyzed for the aptamer pool after each SELEX cycle. R 8 -GST was used to test the enrichment of particular aptamer sequences. R 8 -GST and GST were used to evaluate the specificity of the 24-10 aptamer.  Table 1. Thermodynamic parameters of AR aptamer interaction with R 6 -GST, R 8 -GST and Tat 49-57 -GST determined according to the one set of sites model with assumption that the ligand is in the measurement cell.
Pull-down-based assays. Assays were performed at 4 °C using a universal spin column (MoBiTec) and High Capacity Streptavidin Agarose (HCSA) resin (ThermoScientific). After each washing or elution step, the resin was incubated for 5 min before centrifugation (800xg for 30 s).  The purification of the tested protein was repeated nine times using the same resin. Next, 4 μg of protein sample eluted after the first and ninth purification cycle was denatured and separated on 12% SDS-PAGE followed by Coomassie staining. This is one of three independent experiments which is representative.
Determination of protein elution conditions. The experiment was performed as described in section (Determination of the 24-10 aptamer fragment essential for arginine-rich peptide tag binding) with minor modifications. 20 μL of 50% (v/v) HCSA resin coupled with 1 nmole of the 5′-biotinylated AR aptamer (5′-CTTTGTAATTGGTTCTGAGTTCCGTTGTGGGAGGAACATG-3′) was incubated with the R 6 /R 8 or Tat 49-57 -GST protein (final concentration 4 μM) in 500 μl of ARGI buffer. After washing out the unbound protein the bound protein was eluted with ARGI buffer (5 × 500 μL) supplemented with different concentrations of guanidine hydrochloride (GuHCl), in the following ranges: 0-600 mM for R 6 -GST, 0-700 mM for R 8 -GST and 0-300 mM for Tat 49-57 -GST. The aptamer-bound protein was denatured with 1 × GLB at 95 °C for 5 min. Next, one fortieth of the volume of each protein sample was separated on 12% SDS-PAGE followed by Coomassie brilliant blue staining.
Purification of R 6 /R 8 / Tat 49-57 -tagged proteins using the AR aptamer. The experiment was performed as described in section (Pull-down assay) with minor modifications. AS1 buffer was replaced with ARGI buffer. 14 nmoles of the 5′-biotinylated AR aptamer were immobilized using 150 μL of 50% (v/v) HCSA resin. After washing, the resin was incubated for 1 h with gentle agitation with 1 ml of E. coli total protein extract (final concentration 10 mg/mL) prepared similarly as described in section (Protein purification), in ARGI buffer. Unbound proteins were washed out with ARGI buffer, and the bound protein was eluted by rinsing the resin 5 × with ARGI buffer supplemented with appropriate concentrations of GuHCl. The eluted protein was concentrated using Amicon Ultra-0.5(Merck Millipore Ltd.) and dialyzed against S buffer without glycerol. The Bradford protein assay (Bio-Rad) was used to measure protein concentration in the samples which were next denaturated at 95 °C for 5 min and separated on 12% SDS-PAGE, followed by Coomassie brilliant blue staining.
Reproducibility of AR aptamer-based chromatography. The experiment was performed as described in section (Purification of R6/R8/Tat49-57-tagged proteins using the AR aptamer) with minor modifications. After each purification procedure the HCSA-AR aptamer resin was regenerated with R buffer (50 mM Tris-HCl, 1 M NaCl, 1 M GuHCl, 0.01% (v/v) Tween 20, pH 7.5). When not used the HCSA-AR aptamer resin was stored in S buffer at −80 °C.
Dissociation constant determination. The isothermal titration calorimetry experiments were carried out in duplicates at 25 °C using a VP-ITC instrument (MicroCal, Northampton, MA, USA). Typically, 30 injections of 4 µL aliquots of 200 µM AR aptamer were added into a 1.4355 ml calorimeter cell containing 20 µM tagged protein in 50 mM phosphate buffer (pH 7.6) containing 207 mM NaCl. The injection speed was 0.5 µL/s with 4 min intervals between injections. All solutions were degassed under vacuum prior to use in ITC experiments. In order to ensure proper mixing after each injection, a constant stirring speed of 300 rpm was maintained during the experiment. The heat of the AR aptamer dilution was used to correct the total heat of binding prior to data analysis. The nonlinear analysis was performed according to the one set of sites model using Origin7 software.
Computational Analysis. The models of AR aptamer secondary structure and ΔG were calculated using the mfold web server 18 . The calculation was performed for the following conditions: temperature 25 °C and 300 mM NaCl.