Laterally Aggregated Polyacrylamide Gels for Immunoprobed Isoelectric Focusing

Abstract

Immunoprobed isoelectric focusing (IEF) resolves proteins based on differences in isoelectric point (pI) and then identifies protein targets through immunoprobing of IEF-separated proteins that have been immobilized onto a gel scaffold. During the IEF stage, the gel functions as an anti-convective medium and not as a molecular sieving matrix. During the immunoprobing stage, the gel acts as an immobilization scaffold for IEF-focused proteins via photoactive moieties. Here, we characterized the effect of gel pore size on IEF separation and in-gel immunoassay performance. We modulated polyacrylamide (PA) gel pore size via lateral chain aggregation initiated by PEG monomers. During IEF, the 2% PEG highly porous PA gel formulation offered higher resolution (minimum pI difference ~0.07 ± 0.02) than unmodified 6%T, 3.3%C (benchmark) and 6%T, 8%C (negative control) PA gels. The highly porous gels supported a pH gradient with slope and linearity comparable to benchmark gels. The partition coefficient for antibodies into the highly porous gels (K = 0.35 ± 0.02) was greater than the benchmark (3×) and negative control (1.75×) gels. The highly porous gels also had lower immunoassay background signal than the benchmark (2×) and negative control (3×) gels. Taken together, lateral aggregation creates PA gels that are suitable for both IEF and subsequent in-gel immunoprobing by mitigating immunoprobe exclusion from the gels while facilitating removal of unbound immunoprobe.

Graphical Abstract

Polyacrylamide (PA) gels are widely used for electrophoretic analysis and characterization of proteins.^1–3 Used in important protein electrophoresis modalities,^4,5 PA gels offer exceptional separation resolution because (1) the pore size of the gel approximates the average diameter of globular proteins, making the gel a high-performance sieving matrix and (2) the porous gel retards molecular diffusion, as compared to electrophoresis in free solution (i.e., capillary zone electrophoresis, CZE; capillary isoelectric focusing, cIEF).

In addition to acting as a separation medium for electrophoresis, PA gels offer auxiliary functions. For example, after slab-gel 2-dimensional electrophoresis, the gel matrix is used to transfer specific focused proteins to mass spectrometry (MALDI) for definitive target identification, with the gel region of interest physically cut out of the slab and transferred to the mass spectrometer.⁶ Another example is Western blotting,^7–9 where proteins are separated by size and then probed with immunoreagents, thus facilitating target identification using an immunoassay. Western blotting confers high selectivity but may not afford detection resolution suitable for post-translational modifications and/or splice variants. Such proteoforms may not have detectable molecular mass differences.

Appending an immunoassay to isoelectric focusing (IEF) can provide the resolving power necessary to measure a broader swath of proteoforms. IEF separates protein targets based on differences in isoelectric point (pI), not size.⁵ During IEF, species electromigrate through a pH gradient, stopping at the position in the gradient where the local pH and the pI of the target are equal. Protein targets then “focus” at this position, with a sustained applied electric field and stable pH gradient. Unlike the PA gel of Western blotting (protein sizing), the PA gel of IEF need not function as a molecular sieving matrix. When developed in a miniaturized format, IEF can operate with enhanced heat dissipation (i.e., Joule heating), reduced reagent and sample consumption, and improved separation performance (rapid separation times).

While immunoblotting formats often rely on physical transfer of separated protein targets from the separation lane to a large-pore-size blotting membrane,^10,11 immunoprobed IEF formats have, instead, covalently immobilized separated proteins to the functionalized walls of the IEF separation capillary^12,13 or, in our case, have covalently immobilized protein targets into a PA gel separation matrix^14–17 but only as an anti-convective media (during IEF) and an immobilization scaffold for subsequent immunoprobing (immunoassay).

Consequently, we sought to assess the role of PA gel porosity and microarchitecture in immunoprobed IEF with in-gel target detection. We considered both PA gel formulations similar to slab-gel and published formats (i.e., 6%T 3.3%C PA gel; mean pore radius ~5–92 nm; and fiber radius ~0.5 nm^18,19) as well as PA gels designed to offer larger mean pore radius by inducing lateral aggregation by including a preformed hydrophilic polymer (i.e., PEG) during PA gel polymerization.²⁰ Incorporation of PEG creates a bimodal pore size distribution in the resultant hydrogel, with open fluidic spaces formerly occupied by the porogen and small pores created between bundles of polyacrylamide^18,21,22 For these “highly porous” PA gel formulations, the mean pore radius can be ~250 nm (quantified for a 5%T, 4%C PA gel incorporating 2% PEG with a PA bundle radius at 150 nm), a 100-fold increase.¹⁸ After polymerization, the unfunctionalized PEG freely diffuses out of the PA gel network; however, a small fraction of the PEG is thought to remain within the PA gel as a semi-interpenetrating network, which may inhibit in-gel diffusion of soluble species.²⁰

Specifically, for immunoprobed IEF, we address questions regarding the anti-convective characteristics of highly porous PA gels during IEF and, during immunoprobing, target immobilization to the PA gel matrix, introduction of antibody probe into the PA gels, and removal of unbound antibody probe (background) which impact protein target detection.

EXPERIMENTAL SECTION

Reagents/Chemicals.

Borofloat wafers (University Wafer 516), SU8 3050 photoresist (MicroChem), titanium diisopropoxide bis(acetylacetonate) (Sigma 325252), anhydrous isopropanol (Sigma 278475), a custom in-house-designed mask (CAD/ART Services), GelSlick (Lonza 50640), standard glass slides (VWR), dichlorodimethyl silane (Sigma 440272), 3-(trimethoxysilyl)propyl methacrylate (Sigma 440159), methanol (Sigma 179337), glacial acetic acid (Sigma 8817–46), 30%T 29:1 acrylamide/bis-acrylamide solution (Sigma A3574), 40%T 29:1 acrylamide/bis-acrylamide solution (Sigma A7802), bis-acrylamide powder (Sigma 146072), 1.5 M pH 8.8 TrisHCl (TekNova T1588), N-[3-[(3-benzoylphenyl)formamido]propyl] methacrylamide (BPMAC, custom synthesized by PharmAgra Laboratories), ammonium persulfate (APS, Sigma A3678), N,N,N′,N′-tetramethylethylenediamine (TEMED, Sigma T9281), UV photoinitiator VA086 (Wako Chemicals 61551), 10 kDa poly(ethylene glycol) (PEG, Sigma 92897), borosilicate glass sheets (McMaster-Carr 8476K62), no. 1.5H glass coverslips (Ibidi 0107999097), μ-Slide 4 well glass-bottom chambered coverslips (Ibidi 80427), Array-It Hybridization Cassettes (AHC1×16, Array-It), and permanent lab markers (VWR 52877–310) were used to fabricate materials in this study.

IEF was conducted using the immobilines pK_a 3.6 and pK_a 9.3 acrylamido buffers (Sigma 01716, 01738), SinuLyte pH 4–7 carrier ampholytes (Sigma 05087), an ABS electrophoresis device designed and printed in-house, graphite electrodes (Bio-Rad 1702980), 0.5 mm gel spacers (CBS Scientific MVS0510-R), and TritonX-100 detergent (Sigma X100). Tris-buffered saline with Tween-20 (TBS-T, CST 9997S) was used for gel incubation and wash steps.

The proteins and molecules used in this study were fluorescent pI markers 4.5, 5.5, and 6.6 (Sigma 89149, 77866, 73376), henceforth termed “pI markers,” bovine serum albumin (BSA, Sigma A7030), purified recombinant turboGFP (Evrogen FP552, lot 55201240718) termed “tGFP” (a variant of the GFP with increased fluorescence, MW 27 kDa), primary rabbit-anti-turboGFP antibody (Pierce PA5–22688, lots UA2694351 and UA2718271), secondary polyclonal antibody AlexaFluor-647-labeled donkey-anti-rabbit (Invitrogen A-31573) termed “IgG*.” The IgG* degree of labeling is reported by the manufacturer as five fluorophores per molecule for lot 1964354 (Invitrogen).

Wafer and Gel Fabrication.

SU8 fabrication on a glass wafer was conducted following a standard protocol.²³ Briefly, a custom mask with rails spaced 22 mm apart was used to fabricate features of 40 μm in height (confirmed by optical profilometry) in SU8 on a borofloat wafer. After wafer treatment with GelSlick, polyacrylamide (abbreviated “PA”) gels were fabricated on the wafer and polymerized onto silanized half glass slides. Supplementary Table T1 lists the critical components of each gel condition, using 10 kDa PEG as the preformed hydrophilic polymer and leveraging porogen gel fabrication conditions developed by Righetti and colleagues.¹⁸ After fabrication, gels were incubated for 16–24 h in 1× TBS-T to allow for diffusion of soluble PEG porogen out of the PA gel network, thus forming pores, then rinsed in DI water for 30 min before use.

For this study, we use the canonical notation of %T as the total acrylamide monomer concentration (w/v) in solution and %C as the ratio of bis-acrylamide cross-linker concentration to the total acrylamide monomer concentration.²⁴ To distinguish the six PA gel conditions, as outlined in Supplemental Table T1, we define the 6%T 3.3%C APS/TEMED gels as “benchmark gels,” the 6%T 8%C APS/TEMED + 0% PEG gels as “negative control gels,” the 6%T 8%C + 0.5% PEG gels as “0.5% PEG highly porous PA gels,” the 6%T 8%C + 1% PEG gels as “1% PEG highly porous PA gels,” the 6%T 8%C + 1.5% PEG gels as “1.5% PEG highly porous PA gels,” and the 6%T 8%C APS/TEMED + 2% PEG gels as “2% PEG highly porous PA gels.” The PA gel matrix is formed from incorporation of acrylamide monomers and bis-acrylamide cross-linkers into a randomly organized hydrogel network, inducing heterogeneity in pore size.^25–27

Thermodynamic Partitioning.

We used confocal imaging to measure the thermodynamic partitioning coefficient of IgG* into the PA gels.¹⁶ After silanization of the μ-Slide chambered coverslips, gels were fabricated using wafer molds with 40 μm feature heights within these containers and incubated in TBS-T for 24 h. Gels were then exposed to 1:20 dilution of IgG* solution (from 2 mg/mL stock solution from manufacturer, spun down to remove aggregates) in 2% BSA/TBS-T for >2 h to equilibrium. Confocal imaging experiments were conducted on an inverted Zeiss LSM 710 AxioObserver at the CRL Molecular Imaging Center. Images were acquired at room temperature using a 40× water immersion objective (LD C-Apochromat 40×/1.1 NA W Corr M27, Zeiss) with the correction collar manually assessed to optimal calibration at 0.150 mm. IgG* within the chambered coverslips was imaged using a HeNe633 laser at 17% power, using the MBS488/561/633 beam splitter and the Zen 2010 software (Zeiss). We collected fluorescence image stacks (field of view: 212.55 μm × 212.55 μm; cubic voxels: 0.71 μm × 0.71 μm × 0.70 μm) and analyzed using an in-house Fiji (1.52i, NIH) script.

Equilibrium Swelling Ratio.

The equilibrium swelling ratio was conducted using the Flory–Rehner theory.^24,28–30 After fabrication of PA gels on glass slides with 500 μm spacers to define gel height, gels were weighed immediately on an Ohaus Adventurer Pro weigh station to determine the “fabrication” weight, then incubated in 1× TBS-T for 24 h for PEG diffusion out of the gel. After 12 h of DI water incubation, the equilibrated gel was weighed again for the “hydration” weight, dehydrated fully with a nitrogen gas stream, and weighed a third time for the “dehydration” weight.

IEF and Photocapture Efficiency.

The thin-film IEF was conducted as previously described³¹ with minor modifications to accommodate purified protein solutions. First, gels with a height of 40 μm were rinsed in DI water and then the fluid layer was wicked off the top of the gel. Next, the gel was incubated in 40 μL of a solution of 1% each of the pI markers 4.5, 5.5, and 6.6 (from 1 to 3 mg/mL stock solutions from the manufacturer), and a 3.7 μM GFP (from 37 μM stock from manufacturer) for 30 min at room temperature, protected from light. The three-component IEF lid was fabricated using a 1:100 dilution of the stock 40% Sinulyte carrier ampholytes for a final concentration of 0.4%. Supplemental Table T2 lists the components of the lid gel, which was polymerized for 4 min each at 20 mW/cm² light intensity using a 390 nm UV long-pass filter (Edmund Optics) on an OAI model 30 collimated UV light source.

The PA gel and lid gel were assembled in the ABS electrophoresis device as previously described³¹ and set on an Olympus IX-71 inverted microscope with an Olympus UPlanFi 4× (NA 0.13) objective and a EMCCD Camera iXon2 (Andor), with imaging settings loaded into MetaMorph software (7.10.1.161, Molecular Devices). After a 30 s delay for the soluble reagents in the focusing lid gel to diffuse into the PA gel, IEF was conducted by applying 690 V for 12 min using a Power-Pac high-voltage power supply device (HVPS, BioRad 1645056). During this focusing period, pI markers were imaged using a UV-long pass filter cube (XF100–1, Omega Optical) at 2.5 min intervals, and tGFP was imaged using a GFP filter cube (XF100–3, Omega Optical) 1 min subsequent to the pI markers image. After 12 min of focusing, protein photoimmobilization was induced by application of UV at 100% intensity for 45 s with the Hamamatsu LC8 (Hamamatsu Photonics K.K.), sweeping across the gel assembly at 1.6 cm above the gels. After protein photoimmobilization, gels were rinsed in TBS-T for 30 min to remove uncaptured species and imaged with the same settings used previously. The lab markers used to denote the gel edges along the separation axis are fluorescent in the UV-long pass channel and visible in brightfield imaging.

Micrographs were analyzed using an in-house Matlab (R2015b, MathWorks) script^32,33 adapted to this microwell-free variant of the IEF. Briefly, the micrographs were segmented into regions of interest (ROIs), converted into line plots averaged across the width of the ROI (maintaining the separation axis), and background-subtracted using the average background intensity across the ROI. Gaussian curve fitting to the line plots enables extraction of the peak height, peak location, peak width, area under the curve, signal-to-noise ratio (SNR), and other assay-specific parameters from each ROI. Validation of the Gaussian curve fits is conducted analytically (R² ≥ 0.7 and signal-to-noise ratio SNR ≥ 3) and confirmed manually. For the images taken on the Olympus microscope, 1 pixel is 4 μm × 4 μm.

Immunoprobing.

Assessment of immunoprobing efficiency in hydrated gels was conducted as previously described.^34,35 Briefly, gels on full glass slides were rinsed in DI water, dried with a nitrogen stream, and assembled into a 16-well Array-It Hybridization Cassette (2 columns of 8 wells each). The gel region in each well was rehydrated in 100 μL of 2% BSA/TBS-T for 30 min. Each well in the left column was exposed to 50 μL of 200 nM tGFP protein diluted in 2% BSA/TBS-T for 30 min, while each well in the right column was exposed to 50 μL of 2% BSA/TBS-T as blanks. Protein was immobilized to the photoactive gel by application of 18 mW/cm² UV for 300 s using the OAI model 30 collimated UV light source.

After gels were rinsed in TBS-T for 30 min to remove uncaptured molecules, the gels were immunoprobed for tGFP using a standard immunoprobing protocol.³² Briefly, gels were exposed to 80 μL of 1:10 dilution of primary rabbit-anti-tGFP antibody in 2% BSA/TBS-T (667 nM in solution and 1:10 dilution of secondary donkey-anti-rabbit-647 antibody in 2% BSA/TBS-T (1333 nM in solution).

Gels were imaged on the GenePix 4300A microarray scanner (Molecular Devices) for native GFP fluorescence with the 488-filter set and immunoprobed fluorescence signal with the 647-filter set. Images were analyzed using an in-house Fiji script. Briefly, each well was segmented into a region of interest (ROI) avoiding edges. Each ROI was background-subtracted using the average background intensity in an adjacent blank ROI. For each ROI, we measured average fluorescence intensity, SNR, and immunoprobing efficiency. For the images acquired by the GenePix scanner, 1 pixel is 5 μm × 5 μm.

Statistical Analysis.

One-way ANOVA’s with Kruskal–Wallis test and posthoc Dunn’s multiple comparison test with p < 0.05 (*) were performed, after assessment for normality (Q-Q test), using GraphPad Prism version 8.1.1. Linear regression fit was performed using an in-house Matlab script.

RESULTS AND DISCUSSION

We assessed three key functions of the laterally aggregated, photoactive PA gels in immunoprobed IEF: (i) acting as an anti-convective medium during IEF, (ii) acting as a scaffold for UV-induced immobilization of IEF bands, and (iii) facilitating heterogeneous immunoassays of the immobilized protein targets. The following sections investigate each aspect, starting with a physical characterization of the gel formulations.

Physical Characterization of Highly Porous PA Gels.

We designed and fabricated a panel of PA gel formulations composed of highly porous PA gels polymerized in the presence of hydrophilic polymer (0.5–2.0% 10 kDa PEG, 6% T, 8%C), negative control gels (0% PEG, 6%T, 8%C), and benchmark gels (0% PEG, 6%T, 3.3%C). Three notes on rationale: first, benchmark gels were included to facilitate comparison of this study to other immunoprobed IEF studies.³¹ Second, as compared to the benchmark gel formulations, the highly porous PA gel formulations contain an increased bis-acrylamide cross-linker concentration for mechanical robustness, with APS/TEMED polymerization conditions consistent among all gel formulations (Supplemental Table T1). Third, PA gel porosity increases both via incorporation of PEG at increasing molecular weights and increasing concentration.¹⁸ For this study, we chose to selectively vary the concentration of 10 kDa PEG with the PA gel precursor solution, because lower molecular weight PEG requires higher PEG concentrations to induce lateral aggregation (e.g., 25% for 1 kDa PEG).¹⁸

After gel fabrication, we sought to characterize the dependence of matrix porosity on PEG concentration. Using the equilibrium swelling ratio assay, we first sought to validate the relevance of the assay to the gel formulations under study (Figure 1). To address this question, we observed a slight gel swelling as PEG diffuses out of the PA gels during equilibration of a newly fabricated gel with a surrounding solution (Supplemental Figure S1). The slight swelling suggests extensive water retention. In comparing the dry dehydrated and hydrated gel states, the gel swelling confirms that the equilibrium swelling ratio assay is applicable to scrutinize the highly porous PA gels.³⁰

Figure 1.

Equilibrium swelling ratio reports larger average pore size in 2% PEG-containing PA gel formulations. Swelling ratio (Q) for benchmark gel formulation (0% PEG, 6%T, 3.3%C) is 16.7 ± 1.89, for negative control gel formulation (0% PEG, 6%T, 8%C) is 14.8 ± 1.19 and for highly porous PA gels (0.5% to 2.0% PEG, 6%T, 8%C) is maximum at 19.7 ± 3.54 for the 2% PEG formulation. Mean and standard deviation; n = 4–9 gels; significance determined by one-way ANOVA with Kruskal–Wallis test and post-hoc Dunn’s multiple comparison test with p < 0.05 (*).

Next, using the equilibrium swelling ratio assay, we can determine the equilibrium swelling ratio (Q) defined as

$$Q=\frac{{\text{mass}}_{\text{hydrated gel}}}{{\text{mass}}_{\text{dehydrated gel}}}$$ (1)

where mass_{hydrated gel} and mass_{dehydrated gel} are the masses of each gel when measured in the hydrated and dehydrated states, respectively.^24,27,28

In first comparing benchmark gels to negative control gels, we observed a decrease in Q with the increase in bis-acrylamide cross-linker concentration (Figure 1). Comparison of Q values in the PA gel formulations studied here fall into the range of previously reported values for PA gels with 0% PEG (Q = 17.63 ± 0.83; PA gel 10%T, 3%C gel formulation; different chemical polymerization³⁶). In comparison to the benchmark gel formulation (3.3%C), we observed slightly increased opacity in the negative control gel formulation and highly porous PA gel formulations, all of which contained 8%C bis-acrylamide.

In next comparing the negative control gels (0% PEG) to the highly porous PA gels, we observed a statistically significant increase in Q for the 2% PEG highly porous gel formulation (p < 0.05, Figure 1). Similarly, comparison to the benchmark gel formulation showed a significant increase in Q for the 2% PEG gel formulation (p < 0.05, one-way ANOVA performed with Kruskal–Wallis test and post-hoc Dunn’s multiple comparison test). Swelling behavior in the highly porous PA gel formulations is corroborated by swelling behavior of highly porous PA matrices observed by other groups.²²

Taken together, the inclusion of 2% PEG in the PA gel formulations significantly increases Q, thus indicating an increase in gel porosity as desired. Consequently, our study focuses on the 2% PEG gel formulation (2% PEG, 6%T, 8%C) as the matrix through which to understand the three key functions of the gel in immunoprobed IEF: as an anti-convective medium, a scaffold for protein immobilization, and an immunoassay substrate.

IEF Performance in Highly Porous PA Gels.

To understand the different PA gels as anti-convective media for IEF, we investigated IEF behavior during the focusing and equilibrium stages for a carrier ampholyte system. During IEF, molecules electrophorese along a pH gradient to the point in the gradient that matches the isoelectric point (pI) of that respective species (i.e., the pH at which a molecule is net neutrally charged). Once at the pI position, IEF balances diffusive band broadening with a restorative electrophoretic force, thus forming a pseudoequilibrium electrophoretic separation.^5,37–39 The pH gradient follows the law of pH monotony.⁵ Performing IEF in a matrix reduces convection-associated dispersion along the separation axis, yielding sharply focused protein bands.⁴⁰ In addition to affecting convection, the PA gel porosity should affect the electrophoretic mobility and diffusivity of target species (and ampholytes) under certain conditions.

We first characterized the formation and stability of IEF pH gradients formed in the benchmark, negative control, and 2% PEG PA gel formulations. To monitor IEF, we included three fluorescently labeled pI markers (4.5, 5.5, and 6.6 pH units) as indicators of pH gradient formation over time (Supplemental Figure S2). As a proxy for the duration required to establish a stable pH gradient, we monitored IEF until the stable formation of a Gaussian concentration distribution for each pI marker. In the negative control and 2% PEG PA gel formulation (8%C), the focusing time was ~5 min. In the benchmark gel formulation (3.3%C), the focusing time was ~7.5 min. In all cases, the electrical current decreased during the same periods of time, as expected as ampholytes form a pH gradient (Supplemental Figure S3).^37,41

PA gels are known to support cathodic drift during and after IEF pH gradient formation. The drift of the stable pH gradient is generated by electroosmotic flow arising from applying an electric field to a fluid in a PA gel matrix, which includes slightly negatively charged polyacrylamide chains.⁴² To estimate the cathodic drift velocity, we measured the velocity of the 6.6 pI marker (from the 7.5 to 10 min time point of elapsed IEF time) in IEF performed in the benchmark, negative control, and 2% PEG PA gel formulations.

We observed a statistically significant difference in the drift velocity of the pI marker between the benchmark (6%T, 3.3% C) and negative control (6%T, 8%C) gel formulations (benchmark gel, 0.06 ± 0.02 mm/min; negative control, 0.17 ± 0.01 mm/min; mean ± standard deviation; n = 3 gels; one-way ANOVA with Kruskal–Wallis test and post-hoc Dunn’s multiple comparison test). Reported values of cathodic drift during IEF in PA gels are similar (i.e., 0.021 ± 0.003 mm/min in thin-film IEF,⁴³ ~0.07 mm/min in microchannel IEF,⁴⁴ and ~0.1 mm/min in slab IEF⁴²). Finally, in considering the 2% PEG gels, we observed a cathodic drift velocity of the pI marker (0.08 ± 0.02 mm/min) that was not statistically different from the benchmark or negative control gel formulations, thus indicating that the 2% PEG PA gel formulation reduces convection during IEF, as desired.

We next characterized both the linearity and slope (dpH/dx) of the IEF pH gradient at equilibrium (10 min of elapsed IEF separation time, Figure 2). The three fluorescently labeled pI markers were here, again, used as indicators of a pH gradient (Figure 2A, Supplemental Figure S2). To assess the linearity of each pH gradient, we compared the mean R² statistic for linear fits of pI marker value versus location along the gradient. We hypothesized no notable impact on linearity and slope among the three gel formulations, as carrier ampholyte focusing is dependent on the separation axis length, anolyte and catholyte conditions, and the ampholyte composition, not the hydrogel network.⁵ We observed no significant difference in the slope or the linearity of the pH gradients, among the three gel formulations (Figure 2B,C; one-way ANOVA performed with the Kruskal–Wallis test and post-hoc Dunn’s multiple comparison test). The linearity and pH gradient slope are similar to previously reported values in anti-convective media.^31,43

Figure 2.

Highly porous PA gels are suitable for IEF. (A) Representative inverted fluorescence micrographs and intensity plots of IEF-focused pI markers (4.5, 5.5, and 6.6 markers, black arrows) at 10 min elapsed separation time. (B) Slope of pH gradient is not significantly different between the gel conditions. (C) R² statistic for linear regression fit of the 3 pI markers is not significantly different between the gel conditions. (D) The minimum resolvable pI difference (ΔpI_min) is significantly different between highly porous PA gel formulations with 0% PEG (negative control) and 2% PEG. For all graphs, mean and standard deviation marked by horizontal lines for n = 3 gels per condition. One-way ANOVA with Kruskal–Wallis test and post-hoc Dunn’s multiple comparison test was conducted with p < 0.05 (*).

To conservatively estimate the minimum pI difference for which two neighboring protein peaks are fully resolved⁵ (ΔpI_min), we assessed the triad of fluorescently labeled pI markers through the relationship:

$$\Delta \text{p}{I}_{\text{min}}=3\times \frac{\text{dpH}}{\text{d}x}\times {\sigma }_{\text{protein}}$$ (2)

where dpH/dx is the slope of the pH gradient and σ_protein is a measure of the peak quarter-width (i.e., standard deviation of the Gaussian fit to the fluorescence intensity profile when focused). The estimated ΔpI_min is considered conservative, as the diffusivity of the pI markers is expected to be substantially larger than a moderate-sized protein. The pI markers studied have molecular masses that are 2 orders of magnitude smaller than a moderate-sized protein (i.e., 285 Da for a pI marker). Nevertheless, estimates of ΔpI_min using the pI markers give an indication of the expected suitability of IEF resolution in the 2% PEG highly porous PA gels.

For the 2% PEG highly porous PA gel formulation, we observed a ΔpI_min that was statistically larger than that of the negative control gel formulation (Figure 2D). The ΔpI_min values estimated for the 2% PEG PA gel and negative control PA gel were 0.20 ± 0.03 and 0.12 ± 0.01 (mean ± standard deviation; n = 3 gels), respectively. Differences in the ΔpI_min between the negative control and benchmark gels were not statistically significant (one-way ANOVA performed with Kruskal–Wallis test and post-hoc Dunn’s multiple comparison test). The benchmark PA gel formulation had ΔpI_min = 0.16 ± 0.02.

Next, to estimate a less conservative ΔpI_min, IEF analysis was performed on the well-characterized, fluorescent model protein tGFP, which focuses into 2–3 isoform peaks¹⁵ depending on IEF performance and design. In tGFP analysis via IEF, the 2% PEG highly porous PA gel supported a ΔpI_min = 0.07 ± 0.02 (mean ± standard deviation; n = 3 gels). The benchmark PA gel formulation had a ΔpI_min = 0.11 ± 0.03, with a similar ΔpI_min measured for the negative control (ΔpI_min = 0.12 ± 0.02). The tGFP protein focused into two isoform peaks¹⁵ for all replicate IEF analyses in the 2% PEG PA gels, and in one of three replicates in the negative control gels (Supplemental Figure S4). This IEF analysis of the tGFP isoforms occurred concurrently with that of the 3 fluorescent pI markers (Supplemental Figure S5).

Comparison to other published IEF performance benchmarks suggests that the 2% PEG highly porous PA gels offer ΔpI_min performance on par with or exceeding that of commonly used formats, including capillary IEF⁴⁵ and others (benchmark gels, ΔpI_min = 0.13 ± 0.02 using Polybuffer ampholytes over a pH 4–7 gradient;³¹ IEF in free solution, ΔpI_min = 0.11 for adherent-cell IEF platforms using Zoom ampholytes over a pH 4–7 gradient⁴⁶). Concurrent IEF analyses of pI markers and tGFP isoforms allow estimates of IEF peak capacity, n_c, which reports the number of well-separated protein bands in a single IEF separation lane (n_c = L/4σ, where L = separation lane length, 4σ = peak width).⁵ The n_c for IEF was 12.0 ± 5.9 in the benchmark gels, 16.9 ± 7.9 in the negative control gels, and 19.0 ± 10.2 in the 2% PEG highly porous gels (across 3 replicates each, Supplemental Figure S5). Given the success of the benchmark gels for immunoprobed IEF of complex biospecimens,^31,46 the n_c suggests that the 2% PEG highly porous gels are suitable for high-performance IEF, which is important for resolving protein isoforms and post-translational modifications.

Immunoblotting IEF in Highly Porous PA Gels.

After IEF analysis, the focused proteins were immobilized and probed using an immunoassay to complete the immunoprobed IEF assay. IEF and the PA gel formulations studied here are designed to support in situ immunoprobing, a transfer-free immunoblotting design. Transfer-free blotting reduces losses during resolubilization of IEF-focused protein targets, which is advantageous to the analytical sensitivity of the IEF assay. In situ immunoprobing uses the PA gel matrix as a scaffold for protein target immobilization, followed by an in-gel immunoassay on each immobilized protein target. Light-based blotting (immobilization) toggles the IEF matrix from an anti-convective medium to the immobilization scaffold. One common approach to light-based blotting uses photoactivation (UV) of a light-sensitive monomer polymerized into each PA gel (i.e., N-[3-[(3-benzoylphenyl)formamido]propyl] metha-crylamide, BPMAC). Upon UV activation, BPMAC abstracts hydrogen from neighboring C–H bonds, thus irreversibly cross-linking protein targets to the gel matrix.^23,47

For the transfer-free immunoblotting design, we first assessed each PA gel formulation as an immobilization scaffold. We define photocapture efficiency, η, as

$$\eta =\frac{{\text{tGFP}}_{\text{post-immobilization}}}{{\text{tGFP}}_{\text{pre-immobilization}}}$$ (3)

where tGFP_{pre-immobilization} and tGFP_{post-immobilization} are the areas-under-the-curve (AUCs) of focused protein (tGFP) before exposure of the protein peak to UV light and after exposure to UV light, respectively (Figure 3A). We further considered the minimum resolvable pI difference for each formulation (Figure 3B). Across all three gel formulations, no difference was measured in the η values (benchmark gels, η = 27 ± 15%; negative control gels, η = 25 ± 10%; 2% PEG gels, η = 19 ± 11%; mean ± standard deviation; n = 3 gels, Figure 3C). Using the benchmark gel formulation, immunoprobed IEF on tGFP from single mammalian cell lysate (a more complex sample matrix) performed with η = 17 ± 2%.³¹ Note that the protein loading was not designed to control the total volume (and concomitantly, mass) of sample introduced to the gel (Supplemental Figure S6). Analyses of η for the 2% PEG highly porous PA gel formulation suggests the material is a suitable scaffold for UV-induced protein immobilization.

Figure 3.

Characterization of photocapture efficiency in highly porous PA gels. (A) Representative inverted fluorescence micrographs and intensity plots of IEF-focused tGFP (arrows) at 11 min of IEF (left) and after tGFP immobilization and wash (right). (B) Estimated minimum resolvable pI difference, using the most abundant isoforms, is not statistically different between gel formulations. (C) tGFP photocapture efficiency (η) over all tGFP isoforms indicates no significant difference. For all graphs, mean and standard deviation marked by horizontal lines for n = 3 gels per condition. One-way ANOVA with Kruskal–Wallis test and post-hoc Dunn’s multiple comparison test was conducted with p < 0.05 (*).

For the transfer-free immunoblotting design, we next assessed the performance of the in situ immunoassay (immunoprobing) step. Performing an immunoassay on target immobilized in a hydrogel makes consideration of the thermodynamic partitioning (K) characteristics of the hydrogel important. Gel microarchitecture affects the size-exclusion phenomena observed with small-pore gels and large-molecular-weight soluble species.^17,48,49 In PA gels typically dominated by size-exclusion mechanisms, thermodynamic partitioning reduces the concentration of immunoprobe in the gel, as compared to in free solution, and impacts binding kinetics of the immunoassay. In turn, the concentration of immunoprobe in the gel impacts the analytical sensitivity of the immunoprobed IEF assay, as well as reagent consumption.

First, given the relationship between PEG concentration and resultant porosity of each PA gel formulation, we sought to measure the partition coefficient (K) of a canonical immunoprobe (IgG*, a fluorescently labeled primary antibody), using the following relationship:^17,27,48

$$K=\frac{{\text{IgG}}^{*}{\text{fluorescence}}_{\text{gel}}-{\text{autofluorescence}}_{\text{gel},\text{blank}}}{{\text{IgG}}^{*}{\text{fluorescence}}_{\text{solution}}-{\text{autofluorescence}}_{\text{solution},\text{blank}}}$$ (4)

Here, IgG* fluorescence gel is the fluorescent signal from the labeled immunoprobe in the gel, IgG* fluorescence _solution is the fluorescence signal of the labeled immunoprobe in the solution, autofluorescence _{gel, blank} is the background autofluorescence in the gel without IgG*, and autofluorescence _{solution, blank} is the autofluorescence in the solution without IgG*.

As PEG concentration increases, we anticipate K will also increase. Increased K directly increases the concentration of immunoreagent available for the downstream immunoassay, impacting assay sensitivity.¹⁷ The value of K ranges from 0 (no entry of the soluble species into the gel at equilibrium) to 1.0 (near-equivalent concentrations of the soluble species between free solution and the gel at equilibrium) for species with no specific interactions with the gel network.

As reported in Figure 4, we observed a K for the 2% PEG highly porous PA gels that was significantly larger than both the benchmark gels and negative control gels (benchmark gels, K = 0.11 ± 0.01; negative control gels, K = 0.20 ± 0.01; 2% PEG highly porous PA gels, K = 0.35 ± 0.02; mean ± standard deviation; n = 3–6 gels). These observed K values are corroborated by confocal microscopy of PA gel formulations similar to our benchmark gel formulations.¹⁶ Characterization of PA gels created under lateral chain aggregation conditions by scanning electron microscopy (SEM) reports similar porosity.¹⁸

Figure 4.

Partition coefficient, K, is sensitive to PA gel formulation. Partitioning of AlexaFluor-647-labeled donkey-anti-rabbit antibody (IgG*) into the hydrogel network indicates a statistically significant increase in K due to the gel formulation and incorporation of the PEG. Mean and standard deviation marked for n = 3–6 gels per condition.

The 2-fold increase in K between the benchmark gels and negative control gel is hypothesized to arise from bis-acrylamide aggregating into bundles in the higher %C formulation, as has been previously observed (20%C and 60%C PA gels),⁵⁰ and corroborated by freeze-etched transmission electron microscopy (TEM) of similar PA gel formulations⁵¹ and an observed decrease in elastic modulus⁵² (10%C PA gels versus lower %C PA gels).

Lastly, we sought to compare the dependence of background signal on the gel formulation, after in-gel immunoassay completion. Background signal can be attributed to immunoprobe retained in the PA gel after each round of washing. High background signal can degrade detection sensitivity and is an acute challenge in PA gel immunoassay formats. Here, we considered an immunoassay probed by an unlabeled primary polyclonal antibody to the target, followed by detection using a fluorescently labeled secondary polyclonal antibody. We sought a well-characterized protein target, so opted to employ the naturally fluorescent tGFP protein.

The postimmunoassay background signal in the 2% PEG highly porous PA gels was 2× and 3× lower than background signal measured in the benchmark gels and negative control gels, respectively (benchmark gels, 1850 ± 1060 AFU; negative control gels, 1290 ± 380 AFU; 2% PEG highly porous PA gels, 606 ± 72 AFU). We attribute the lower background measured in the highly porous PA gels to reduced entropic trapping of antibody probes in these larger pore size matrixes. Freely diffusing molecules are known to become trapped in regions of hydrogel pore-size inhomogeneity.⁴⁹ In these larger pore size regions, molecules can be trapped in a free energy minimum.

Entropic trapping is particularly relevant for macromolecules of intermediate size.⁴⁹ Immunoreagents such as antibodies (5–8 nm hydrodynamic radius^53,54) would be expected to be entropically trapped in the benchmark PA gels (5–92 nm mean pore radius^18,19). In the 2% PEG highly porous PA gels, however, the estimated mean pore radius is substantially larger (~250 nm).¹⁸

The observed decrease in background signal during immunoprobing in the 2% PEG highly porous PA gels is attributed to reduced entropic trapping of antibody probes in the hydrogel. Reduced entropic trapping would facilitate more effective washout of both the primary and secondary antibodies during the immunoprobing steps. Overcoming the high post-immunoprobing background signal makes the 2% PEG highly porous PA gels an attractive material in which to perform immunoprobed IEF.

CONCLUSIONS

We sought to design a PA gel to enhance performance of the immunoassay stage of immunoprobed IEF, while maintaining the separations performance of the IEF stage. Here, we utilized lateral aggregation during PA gel polymerization to increase the mean pore size of PA gels. To do this, we devised PA gel formulations that included the hydrophilic polymer PEG. We validated that we created significantly larger porous PA gels and observed a significant increase in antibody partitioning into the gel matrix. Comparing the highly porous PA gel to benchmark gels (utilized in previous immunoprobed IEF assays), we observed no significant impact on the formation or stability of the IEF pH gradient and an improvement in separation resolution. Consequently, the highly porous PA gels are suitable for IEF analysis of protein targets. In the subsequent immunoassay stage, the large pore size gels allow for more effective washout (removal) of unbound immunoreagent at immunoassay completion. The highly porous PA matrixes therefore serve to support high-performance IEF separations while providing an immobilization scaffold that is well-suited for heterogeneous immunoassays needed for selective target readout.