Multiplexed Protein-Interaction Studies Using the HaloLink™ Protein Array System

Introduction

Functional protein arrays allow researchers to probe thousands of proteins immobilized on a surface to detect proteome-scale protein-interaction networks^{(1) (2)} . Researchers also are increasingly using these arrays in clinical diagnostic applications to look for antibody response to cancer antigens and infectious agent^{(3) (4)} . The main advantage of using protein arrays is the miniaturization and multiplexing capabilities they provide, enabling large amounts of information to be generated rapidly from small amounts of sample. However, before protein arrays become a mainstream tool for biological applications, several technological hurdles remain to be solved.

Scientists interested in using functional protein arrays for their specific biological applications have limited options. The first option is to use preprinted arrays with specific protein content already immobilized on the surface. However, the utility of these arrays is limited because most researchers want to use their own specific content. A second option for researchers is to create custom protein arrays containing content of their interest, which is a daunting task because of the need to:

• clone, express, purify and validate the function of the protein content.
• select a capture surface and attachment chemistry that will maintain the three dimensional conformation of proteins.
• optimize complex robotics to create reproducible arrays.

Creating Custom Arrays Using the HaloLink™ Protein Array Systems

The HaloLink™ Protein Array Systems are designed to address the needs of scientists interested in using protein arrays for medium-throughput, multiplexed protein interaction studies. Innovations in cell-free protein expression systems, HaloTag® capture technology and surface engineering are combined to simplify the process of making custom protein arrays. First, cell-free transcription and translation systems—TnT® T7 Quick Coupled Transcription/Translation System and TnT® SP6 High-Yield Wheat Germ Protein Expression System—are used to express proteins directly from plasmid DNA. These systems are easy to use, rapid, and ideal for expressing proteins toxic to the cell line of choice. In addition, they allow protein modification with nonnatural amino acids containing fluorescent or other reactive labels to facilitate downstream detection.

Second, proteins of interest are expressed as HaloTag® fusion proteins in cell-free systems. The HaloTag® protein is a 33kDa tag engineered to form a covalent bond with its HaloTag® Ligand. This tag allows covalent and oriented capture of proteins of interest on HaloLink™ Slides directly from the cell-free expression system, hence eliminating the need for any prior purification step.

Third, HaloLink™ glass slides are coated with a polyethylene glycol-based hydrogel that is known to resist nonspecific protein adsorption and to minimize surface-induced denaturation of specific proteins. Finally, we used a silicone gasket to create 50 wells on the glass slides, allowing multiple assays to be performed manually on the same slide without any specialized equipment.

Figure 1 provides an overview of the workflow for custom array fabrication and downstream application using HaloLink™ Protein Array Systems. To create a custom array, proteins of interest are expressed in a cell-free system as HaloTag® fusion proteins and added directly to wells on the HaloLink™ Slides to capture fusion proteins at the surface. The custom array can then be used for downstream applications such as protein:protein or protein:DNA interaction studies. The combination of cell-free expression systems and HaloTag® attachment chemistry allows creation of a medium-throughput custom array and performance of protein interaction studies in less than 8 hours.

Figure 1. Overview of the HaloLink™ Protein Array Systems experimental workflow.

These systems allow the user to create a medium-throughput custom array and perform downstream applications in less than eight hours.

Identifying Protein:Protein Interactions Using the HaloLink™ Protein Array Systems

To demonstrate the utility of HaloLink™ Protein Array Systems for multiplexed protein:protein interaction experiments, we studied interactions among eight HaloTag® bait fusion proteins and three prey proteins. The protein interaction study was divided into two steps: First, we expressed eight HaloTag® bait fusion proteins using the TnT® T7 Quick Coupled Transcription/Translation System and captured them to produce the custom HaloLink™ protein array. The HaloTag® bait fusion proteins used in the study were: HaloTag®-cJun, HaloTag®-cFos, HaloTag®-Protein Kinase A (PKA), HaloTag®-R1α, HaloTag®-p53, HaloTag®-p65, HaloTag®-β-galactosidase (βgal) and HaloTag® protein only (control). The second step was to probe the custom HaloLink™ array with three prey proteins for bait:prey interactions. The prey proteins selected for the study were c-Fos, c-Jun and p50.

Step 1: Creating the Custom Array

The eight bait proteins were cloned into pFN19K (N-terminal HaloTag® fusion) Flexi® vectors and expressed in the TnT® T7 Quick Coupled Transcription/Translation System (see supplementary information for details). After expression, 5.0μl/well of samples were added to the individual wells on HaloLink™ slides (Table 1), incubated for 1.0 hour, washed and dried to obtain a custom array. Column 1 contained a dilution series of HaloTag® standard protein. Column 2 had HaloTag® bait fusion proteins diluted 1:20-fold in PBSB (PBS with 10 mg/ml BSA). Both of these columns were probed in subsequent steps using Anti-HaloTag® Polyclonal Antibody (pAb) to confirm and estimate the binding of HaloTag® fusion protein. Columns 3–5 contain eight HaloTag® fusion bait proteins in triplicate for protein:protein interaction studies using the three prey proteins and two additional negative controls (cell-free lysate without DNA, and PBSB-only).

Step 2: Probing the Custom Array

The custom HaloLink™ array fabricated above was interrogated with prey (probe) proteins to identify key protein:protein interactions. Prey proteins (c-Jun, c-Fos and p50) were cloned into TS1K Flexi® vectors (non-HaloTag® analog of pFN19K) and expressed in TnT T7 Quick  Coupled Transcription/Translation System (see supplementary information for details). Prey proteins were biotinylated during expression using Transcend™ Biotinylated tRNA. Biotinylated prey proteins, c-Jun, c-Fos and p50, were added to columns 3, 4 and 5, respectively, of the custom array and incubated for 1.0 hour. After washing and drying the probed slide, positive interactions were detected by incubating with Alexa Fluor® 647 labeled streptavidin. In parallel, columns 1 and 2 of custom array were probed using Anti-HaloTag® pAb followed by secondary antibody labeled with Alexa Fluor® 633. The slide was scanned to detect the positive interactions (Figure 2, Panel A).

Figure 2. Detecting capture of HaloTag® bait fusion proteins on HaloLink™ Slides.

Panel A. Scanned image of the HaloLink™ custom array probed using prey proteins and the Anti-HaloTag® pAb. Panel B. Plot of the signal intensity obtained from HaloTag® Standard Protein (column 1 on the slide) and HaloTag® fusion bait proteins after probing with Anti-HaloTag® pAb (column 2 on the slide). Median fluorecent intensity is calculated from a 2.0mm diameter area.

Figure 2, Panel B, is the plot of median fluorescent intensity obtained from columns 1 and 2 of the scanned slide image and shows the relative amount of HaloTag® fusion proteins captured on the HaloLink™ slide. The binding of HaloTag® fusion protein to the HaloLink™ slide was specific because no signal was seen when the active site of HaloTag® protein is blocked by incubation with excess amount of HaloTag® Ligand (data not shown). The data shown with the HaloTag® Standard Protein in Figure 2, Panel B, were useful for normalizing the experimental variations resulting from: a) differences in expression levels of HaloTag® fusion proteins; b) differences in capture efficiency of individual HaloTag® fusion proteins; and c) intra- and inter-lot variability of HaloLink™ slides.

Figure 3 is the plot of fluorescent signal obtained when the HaloLink™ custom array (Figure 2, Panel A, columns 3–5) was probed with prey proteins. Positive interaction of c-Jun prey protein was seen with c-Jun,
c-Fos, p53 and p65 bait protein. c-Fos interacted with c-Jun, p53 and p65 bait proteins, and p5o interacted with p53, p65 and β-galactosidase. These interactions have been previously reported in the literature^{(5) (6)} , except for interaction between p50 and β-galactosidase.

Figure 3. Quantitating signal from HaloLink™ Arrays to identify protein:protein interactions.

Plot showing the fluorescent signal obtained from interaction of HaloTag® bait fusion proteins with prey proteins. Background (Bkg.) signal is the average of six wells that contained only cell-free lysate with no HaloTag® protein or PBSB. The standard deviation (δ) was calculated from these six wells. Bkg. + 3σ was used as the cutoff for positive protein:protein interaction.

Studying Small-Molecule-Modulated Protein:Protein Interaction using the HaloLink™ Protein Array Systems

We also tested the use of the HaloLink™ Protein Array Systems to study the modulation of protein:protein interactions by small molecules. We chose to study interaction between FK506 binding protein (FKBP) and the FKBP rapamycin binding domain of mTOR (FRB) proteins⁽⁷⁾ . Rapamycin is needed for FKBP and FRB to interact, and the protein:protein interaction can be modulated by changing rapamycin concentrations.

FRB was expressed as HaloTag® bait fusion protein in a TnT® SP6 High-Yield Wheat Germ Expression System and captured on HaloLink™ slides. Purified FKBP-GST (20μg/ml in PBSB) was added to the slides in the presence of varying amounts of rapamycin. The amount of FKBP-GST bound to FRB was detected using an anti-GST antibody and secondary antibody labeled with Alexa Fluor® 633. Figure 4 shows that FKBP binding is dependent on the concentration of rapamycin. No signal was measured in the absence of rapamycin.

Figure 4. Plot of fluorescent signal from HaloTag®-FRB-FKBP-GST interaction as a function of rapamycin concentration.

The plot is shown on a log-log scale for clarity. Each point represents an average and standard deviation from triplicate experiments. Units are relative fluorescence units (RFU).

Performing Enzyme Assays Using the HaloLink™ Protein Array Systems

Using HaloTag®-PKA enzyme, we show that the 50-well format of the HaloLink™ Protein Array Systems also allows researchers to perform fluorescent enzyme assays. HaloTag®-PKA was expressed in TnT® SP6 High-Yield Wheat Germ Protein Expression System and captured on a HaloLink™ Slide in a 1-hour incubation. The slide was washed with PBSI containing 5% glycerol and spin-dried. The 5% glycerol in the wash buffer was needed to maintain enzyme activity during washing and drying⁽⁸⁾ . PKA activity was assayed using the ProFluor® Assay (Figure 5; see supplementary information for details). After the enzyme assay, the amount of PKA at the surface was determined using anti-PKA antibody. Results of the PKA activity assay correlated very well with the amount of PKA at the surface. With the ProFluor® Assay, the fluorescent signal has an inverse relationship with PKA activity, hence a lower amount of PKA at the surface will result in higher fluorescent signal. In addition to PKA, we have recently demonstrated the feasibility of performing other fluorescent enzyme assays on HaloLink™ slides⁽⁸⁾ .

Figure 5. PKA enzyme assay on HaloLink™ Slides.

HaloTag®-PKA fusion protein was expressed using TnT® SP6 High-Yield Wheat Germ Protein Expression Extract and serially diluted in PBSB. Five-microliter aliquots of each dilution were added in triplicate on a HaloLink™ Slide. The slide was incubated for 1 hour in a humidity chamber, washed with PBSB containing 5% glycerol and dried. The captured PKA was assayed using the ProFluor® PKA Assay (left Y axis). The amount of PKA on the HaloLink™ Slide was quantitated using an anti-PKA antibody (right Y axis).

Maintaining Stability of Proteins Captured on a HaloLink™ Slide

A major concern for functional protein arrays is the stability of captured proteins. Our investigation⁽⁸⁾  identified that immobilized proteins may lose their functionality either during slide processing steps of washing and drying or during storage of printed slides. Using three enzymes (PKA, β-lactamase and β-galactosidase) fused with HaloTag® protein, we observed that adding 5% glycerol to wash buffers improved the stability of immobilized enzymes during slide processing steps. Surprisingly in the same study we found that washing and spin-drying had no significant impact on protein:protein interactions. All three enzymes immobilized on the HaloLink™ Slides retained their activity when stored at –20°C in 50% glycerol. However, storing the slide at –20°C occasionally resulted in the 50-well gasket coming off during subsequent spinning and drying steps.

To address the stability issue of printed proteins, we explored the ‘print-on-demand’ concept. All proteins needed for an array experiment may be expressed simultaneously in 96- or 384-well plates, stored at –70°C and used to print protein at the time of experiment. Our study showed no decrease in enzyme activity in the cell-free reaction mixes containing enzymes fused to the HaloTag® protein were exposed to multiple freeze-thaw cycles, captured on HaloLink™ slides and assayed for enzyme activity.

Using a single batch of proteins for printing proteins may eliminate problems with experimental data that result from batch-to-batch variation in cell-free expression systems. However we caution that it is indeed possible our freeze-thaw results may not be universally applicable, and optimization may be required for individual proteins.

Summary

In summary, we have shown that HaloLink™ Protein Array Systems can be used successfully to create custom protein arrays on hydrogel-coated surfaces. HaloTag®-fused proteins can be specifically captured from TnT® T7 Quick and SP6 High-Yield Wheat Germ Protein Expression Systems without any prior purification step, simplifying the protein array fabrication process. We show that HaloLink™ Slides allow multiplexed protein:protein interaction studies, screens for small molecule modulators of protein:protein interactions and fluorescent enzyme activity assays.

Acknowledgments

The authors thank the following HaloLink® Protein Arrays team members for their contributions: Brian Andersen, Carol Lindsay, Laurent Bernad, Mike Scurria, Patricia Bresnahan, Tanya Quint, Tenaya Noce, Paula Phenix, James Altwies, Michael Bayline and Ned Reimer.