Isoelectric Focusing: pI Separation, Ampholytes and IPG Strips

Isoelectric focusing (IEF) separates proteins by the pH at which their net charge is zero. The principle is simple enough to state in a sentence, which is why so many first attempts fail for reasons that have nothing to do with the principle. In practice three things decide whether a run works: how much non-protein charge is in the sample, whether you stopped the run at the right endpoint, and whether your reducing agent survives the part of the pH gradient you care about. Everything else — strip length, pH range, stain choice — is optimisation on top of those three.

This guide covers the version most people run: immobilized pH gradient (IPG) strips as the first dimension of a two-dimensional gel, plus the capillary formats now standard in biopharmaceutical quality control. Specific numbers below come from the Cytiva Immobiline DryStrip instructions for use, the GE Healthcare/Cytiva 2-D Electrophoresis: Principles and Methods handbook, and USP general chapter ⟨1054⟩. Volumes, loads and volt-hour targets are system- and vendor-specific — verify against the instructions for the strips and instrument you actually own before pipetting.

What focusing actually does, and why it is self-sharpening

A protein carries net positive charge below its isoelectric point (pI) and net negative charge above it. Placed in a pH gradient under an electric field, it migrates until it reaches the position where pH equals its pI, at which point its net charge is zero and electrophoretic migration stops. Diffusion still moves it out of that band, but the moment it does it picks up charge again and is pulled back. USP ⟨1054⟩ describes this restoring behaviour as the concentration effect that gives the technique its name: the gradient does not merely separate proteins, it actively concentrates each one into a narrow zone.

That is why IEF behaves unlike the size-based separation in SDS-PAGE, where bands only ever broaden with time. It also means the run has a genuine steady state. USP ⟨1054⟩ gives the classical resolution expression, in which the minimum resolvable pI difference depends on the protein’s diffusion coefficient, the steepness of the pH gradient (dpH/dx), the field strength (E), and how fast the protein’s mobility changes with pH near its pI. Two of those four are yours to control: narrow the pH range and raise the voltage, and resolution improves. The ceiling on voltage is heat — which is why every IEF apparatus is built around a cooled bed.

A computed pI is not a strip-selection tool

The most common planning error is choosing a narrow-range strip from a pI calculated off the amino acid sequence. Sequence-based prediction is far less accurate for whole proteins than people assume. In the benchmark accompanying IPC 2.0 (Kozlowski, Nucleic Acids Research, 2021), a deep-learning predictor reached a root-mean-square deviation of 0.848 pH units against experimental protein pI values on a 581-protein test set; a classical Henderson–Hasselbalch implementation of the Bjellqvist type reached 0.937. Peptides predict far better — RMSD 0.222 for IPC 2.0 — because they lack the folded structure and modifications that shift real pKa values.

An error near one full pH unit is larger than the entire span of a narrow-range strip. Two practical consequences:

• For an unfamiliar protein, scout on a broad range first (pH 3–10 or 3–11 NL), find where it lands, then commit to a narrow range.
• Never treat an observed pI shift as evidence of a specific modification without orthogonal confirmation. Deamidation, sialylation, phosphorylation, C-terminal lysine processing and N-terminal pyroglutamate formation all shift pI in the same direction as one another or as simple carbamylation artefacts. Assigning the cause needs mass spectrometry, not a gel position.

Carrier ampholytes versus immobilized pH gradients

Two chemistries create the gradient, and the difference between them is the single biggest determinant of how reproducible your pI readings are.

Carrier ampholytes (CAs) are mixtures of small amphoteric molecules with a spread of pI values. Under field they sort themselves by pI and buffer their local environment, generating the gradient in situ. The gradient is therefore a dynamic, self-organising structure — and it decays. USP ⟨1054⟩ names the phenomenon cathodic drift: if a gel is focused too long, the gradient decays and focused protein migrates off the cathodic end of the gel entirely. The chapter attributes it to electroendosmosis and absorption of carbon dioxide, and notes candidly that the mechanism is not well understood.

Immobilized pH gradients (IPGs) solve this by covalently building the buffering groups into the matrix. Immobiline reagents are acrylamide monomers each carrying a single weak acidic or basic group; two solutions of differing buffer composition are mixed during casting, and the acrylamide portion copolymerises with acrylamide and bisacrylamide so the buffering groups become part of the gel. The gradient cannot migrate because it is chemically bolted down.

The resolution difference is large and quantified in USP ⟨1054⟩: proteins differing in pI by as little as 0.02 pH units may be resolved on a carrier-ampholyte gel, whereas an immobilized pH gradient can resolve differences of approximately 0.001 pH units. Independent measurement supports the stability claim — Lomeli and Herr (Analytical Chemistry, 2024) reported a 43-fold reduction in cathodic drift for a microscale mixed-bed IPG gel relative to a carrier-ampholyte separation in the same device.

Note that this is not an either/or in practice. Standard IPG rehydration solutions still contain carrier ampholytes at 0.5–2%, because they improve protein solubility and even out conductivity across the gradient without displacing the immobilized gradient.

Choosing strip length, pH range and load

Cytiva Immobiline DryStrip gels are 3 mm wide and 0.5 mm thick, supplied dried on a plastic backing in 7, 11, 13, 18 and 24 cm lengths and stored at −20 °C. Longer strips and narrower pH ranges both buy resolution, and both cost run time and sample.

Loads scale with both length and range. From the Immobiline DryStrip instructions for use, for pH 3–10:

• 7 cm: 3–6 µg for silver stain, 25–60 µg preparative (Coomassie)
• 18 cm: 20–40 µg silver, 100–500 µg preparative
• 24 cm: 30–60 µg silver, 200–600 µg preparative

Narrower ranges take more protein because the same load is spread over fewer species per centimetre — a 24 cm pH 4–7 strip is rated 45–90 µg for silver and 200–1300 µg preparative, and basic ranges higher still. These figures were established with total protein from E. coli extracts; a sample dominated by a few abundant species will overload well below them.

Sample application method is dictated by pH range, not convenience. Rehydration loading (sample included in the rehydration solution) is the default for broad ranges. For basic intervals such as pH 6–9 and 6–11, Cytiva recommends anodic cup application, with a maximum of about 100 µg protein per 100 µL and no more than 150 µL in the cup — higher concentrations risk precipitation at the cup. Preparative loads on basic ranges go on by anodic paper-bridge loading instead. Albumin-rich samples such as plasma are an exception: cathodic cup loading, with DTT in both sample and rehydration solution.

Salt is the failure mode that mimics every other failure mode

If a run streaks, blanks out at the ends, or simply never reaches voltage, check conductivity before you change anything else. The Cytiva handbook is unambiguous about the numbers:

• Rehydration loading: keep the salt concentration in the rehydration solution below 10 mM.
• Cup loading: up to 50 mM may be tolerated — though proteins may precipitate at the application point as they move abruptly into a lower-salt environment.

The mechanism explains the symptom set. Salt raises strip conductivity, so the power supply holds a low voltage; no protein focuses until the ions have migrated to the strip ends, which eats the run time. Water movement follows the ions, so one end of the strip swells while the other dries. The visible result is a strip with large unfocused regions at either end, read on the second-dimension gel as horizontal streaking or empty zones. USP ⟨1054⟩ gives the same advice from the pharmacopoeial side: prepare the sample in deionized water or 2% ampholytes, desalting by dialysis or gel filtration if necessary.

Ionic detergent is the other conductivity trap, and it is easy to introduce accidentally. SDS is incompatible with IEF because it confers charge on proteins — the exact property IEF is trying to read. If a lysis protocol used SDS, the handbook gives two routes: dilute into a large excess of nonionic or zwitterionic detergent, or separate the SDS by acetone precipitation. Cleanup pays for itself: in the handbook’s own comparison, rat liver protein extracted with Triton X-100 and precipitated with acetone gave 758 silver-stained spots, versus 801 after a dedicated 2-D cleanup, versus 726 for a plain urea buffer extract.

Getting the sample into a low-conductivity, fully solubilised state is the same problem addressed at larger scale in protein purification, and the same trade-off applies: every cleanup step that removes salt also loses protein.

The rehydration solution: four decisions in one tube

A standard urea rehydration stock from the Cytiva handbook is 8 M urea, 2% (w/v) CHAPS, 0.5% or 2% (v/v) carrier ampholyte matched to the strip range, and 0.002% bromophenol blue, with DTT added immediately before use at 7 mg per 2.5 mL aliquot (roughly 18 mM). For samples with poorly soluble or membrane proteins, the thiourea variant substitutes 7 M urea plus 2 M thiourea.

Four decisions are embedded in that recipe:

1. Chaotrope. Urea can be pushed to 9 or 9.8 M if solubility demands it. USP ⟨1054⟩ notes that 3 M is often sufficient simply to stop proteins precipitating at their pI, with up to 8 M available.
2. Detergent. It must have zero net charge. Nonionic or zwitterionic only — CHAPS, Triton X-100, NP-40, octyl glucoside, or the alkylamidosulfobetaines ASB-14 and ASB-16, typically 0.5–4%.
3. Ampholyte concentration. This is the non-obvious one. 2% improves solubilisation and raises tolerance to salt, but it also raises conductivity, so the run holds a lower ceiling voltage and takes longer. 0.5% gives higher voltage and shorter, cleaner runs. Cytiva recommends 0.5% for the IPGphor strip holder and for pH 3–11 NL strips specifically, where high voltage and a short run are what keep the result streak-free; 2% when solubility or sample salt forces it.
4. Reducing agent. Covered below — the standard answer is wrong for basic gradients.

Rehydration volumes are fixed by strip length and should be dispensed exactly: 125 µL for 7 cm, 200 µL for 11 cm, 250 µL for 13 cm, 340 µL for 18 cm and 450 µL for 24 cm, rehydrating for 10 to 20 hours. Wear gloves throughout: skin keratin is a classic source of phantom spots.

Focusing is measured in volt-hours, not hours

The endpoint of an IEF run is a cumulative quantity — volts integrated over time (kVh) — not a wall-clock duration. A salty sample spends its first hours at low voltage accumulating almost no volt-hours, so a run timed by the clock will be stopped before anything has focused. Cytiva’s guideline totals for the IPGphor 3 illustrate how steeply the requirement scales:

• 7 cm, pH 3–10: 5.0–6.5 kVh, about 2.5 hours
• 24 cm, pH 3–10: 35–45 kVh
• 24 cm, pH 4–7: 45–60 kVh
• 24 cm, pH 6–9: 60–85 kVh

Narrower ranges need more volt-hours than broad ones on the same strip length, and basic ranges need the most. Cytiva permits 8000 V with the regular or cup-loading strip holder and 10000 V with the cup-loading manifold, running at 20 °C and 50 µA per strip. For crude, salty or paper-bridge-loaded samples the low-voltage first step may be extended by up to four hours to let salt migrate out before high voltage is applied — that extension is the correct response to a conductivity problem you could not remove upstream.

Two endpoint traps worth naming. First, bromophenol blue is not an endpoint indicator. Cytiva flags this explicitly: the dye reaches the end of the strip well before the sample is focused. Its only diagnostic value is that if it does not move at all, no current is flowing. Second, over-focusing is a real failure, not just wasted time — on carrier-ampholyte systems it is precisely how cathodic drift loses your basic proteins off the end of the gel.

Where no vendor guideline exists, USP ⟨1054⟩ describes the empirical way to find the steady state for a slab system: apply a coloured protein such as haemoglobin at several positions on the gel and focus until every application point yields an identical band pattern. The same chapter warns not to apply samples close to either electrode, and suggests trying three positions during method development, since a protein applied at opposite ends may not give identical patterns.

One apparent conflict is worth resolving explicitly, because it confuses people moving between formats. USP ⟨1054⟩ recommends cooling the bed to approximately 4 °C for thick slab IEF; Cytiva specifies 20 °C for IPG strips on the IPGphor. Both are correct for their format — urea-containing IPG strips must stay warm enough that the urea does not crystallise. Follow the temperature given for your apparatus, not the one from a different one.

Basic pH ranges break the standard DTT recipe

This is the mistake that produces the most puzzling gels, because the protocol was followed correctly and the result still streaks above pH 7.

DTT is weakly acidic. During focusing it becomes negatively charged and migrates away from the basic end of the strip toward the anode (Frontiers in Chemistry, 2017). Proteins in the basic region are left without reducing protection partway through the run, their thiols reoxidise, and intra- and intermolecular mixed disulfides form. The Cytiva handbook describes the visible outcome: horizontal streaks and extra spots, specifically at pH > 7, plus a spot pattern inflated by the several oxidation states of the same protein. The 2017 review argues the standard post-IEF reduction/alkylation cannot fix this, because it only addresses reoxidation during SDS-PAGE — the damage was already done during focusing.

Two established responses:

• Thiol stabilisation instead of reduction. A disulfide reagent (Cytiva’s DeStreak Reagent, hydroxyethyl disulfide) converts protein thiols to stable mixed disulfides so there is nothing left to reoxidise, reducing streaking particularly across pH 7–11. It is compatible with sample solutions containing up to 20 mM reducing agent, but the strip instructions are explicit that in the rehydration solution you use either DeStreak Reagent or DTT, never both. Excess free reductant defeats the reagent.
• Reduce and alkylate before IEF rather than after. The 2017 review found pre-IEF reduction and alkylation more effective at preventing oxidation-driven horizontal streaking. It is a larger protocol change with its own caveats, and conditions need optimising per sample.

One further basic-range detail from the strip instructions: for strips exceeding pH 9, soak the cathodic electrode pad in DeStreak rehydration solution rather than water.

Never heat a urea-containing sample

Urea in solution degrades to cyanate/isocyanate, which carbamylates protein amino groups. Because carbamylation removes positive charge, it shifts pI — producing artefactual “charge trains”, series of evenly spaced spots that look exactly like a genuine post-translational modification ladder and are not.

The temperature limit depends where you look, and the Cytiva handbook itself gives two figures: its sample-preparation chapter says the solution temperature must not exceed 37 °C once urea has been added, while its handling-precautions and troubleshooting sections say not to heat urea-containing solutions above 30 °C. Treat 30 °C as the working limit — the cost of the conservative number is nothing. USP ⟨1054⟩ adds the complementary rule: where urea is used, only fresh solutions should be prepared, to prevent carbamylation. Standing urea stock accumulates cyanate whether or not anyone heated it.

The practical corollary catches people transferring habits from SDS-PAGE: you cannot boil an IEF sample. Whatever denaturation the protocol needs must come from the chaotrope and detergent, not from heat.

Equilibration: the handover to the second dimension

Before a focused IPG strip can run as the first dimension of a 2-D gel, its proteins must be coated in SDS and their disulfides dealt with. The Cytiva protocol is two 15-minute steps in the same base buffer — 6 M urea, 75 mM Tris-HCl pH 8.8, 29.3% glycerol, 2% SDS, 0.002% bromophenol blue — split into two portions with different additions made just before use:

1. DTT, 100 mg per 10 mL, 15 minutes. Reduces disulfides.
2. Iodoacetamide, 250 mg per 10 mL, 15 minutes. Alkylates the resulting free thiols so they cannot reoxidise during the second dimension.

Volumes scale with strip length: 2.5–5 mL for 7 cm, 5–10 mL for 11 and 13 cm, 10–15 mL for 18 and 24 cm. The handbook stresses consistency in timing across strips — equilibration is a diffusion process, and inconsistent durations shift how much protein is lost from the strip, which shows up as unexplained quantitative variation between gels you intended to compare. Orientation matters too: the “+” end of the strip is acidic and must face the anode, and in flatbed formats the strip goes gel-side down onto the second-dimension gel.

Capillary and imaged capillary IEF

Outside 2-D proteomics, most IEF now runs in capillaries. Capillary IEF (cIEF) performs the same separation inside a coated capillary; imaged cIEF (icIEF) images the entire capillary at once with a whole-column detector, so no mobilisation step is needed and the focused pattern is read directly. It has become the standard method for charge-variant analysis of therapeutic proteins, where acidic and basic species arising from deamidation, C-terminal lysine variants, N-terminal pyroglutamate, glycation and sialylated glycans are release and stability attributes.

The practical contrasts with gel IEF are worth knowing before choosing a platform:

• Time scale. Focusing takes minutes rather than the hours-to-overnight of a long IPG strip.
• Gradient chemistry. cIEF is a carrier-ampholyte technique, so ampholyte selection and concentration dominate method development. A 2025 review in Frontiers in Chemistry notes that raising total ampholyte concentration past roughly 3–4% increases background and complicates peak integration.
• Calibration. pI is assigned by interpolation between pI markers. The common practice of using two or three markers assumes the pH gradient is linear; that review reports the gradient is not uniformly resolved across its range, and that using more markers with non-linear regression changes assigned pI values.
• Precision. The same review cites a multi-laboratory NISTmAb study with pI RSD below 0.3% and charge-variant peak-area agreement near 1% across laboratories — far tighter than densitometry of a stained gel.

USP ⟨1054⟩ anticipates this: it permits variations on the detailed slab procedure, including automated instrumentation and immobilized pH gradients, subject to validation, and names the criteria — formation of a stable pH gradient of the desired characteristics, assessed using coloured pH markers of known pI, and comparison against the reference substance electropherogram. On a gel, IEF may serve as an identity test, a limit test, or a quantitative test, but only the last requires densitometry and validation.

Troubleshooting by symptom

• Horizontal streaking across the whole gel, empty regions at the strip ends, run never reached voltage — conductivity. Salt above 10 mM under rehydration loading, residual SDS, or endogenous small ionic molecules. Desalt or precipitate; extend the low-voltage step by up to four hours.
• Streaking confined to the basic half, above pH 7 — thiol reoxidation after DTT has migrated anodally. Switch to a thiol-stabilising disulfide reagent, or reduce and alkylate before IEF.
• Evenly spaced horizontal spot trains where you expected one spot — carbamylation. Urea solution was heated or was not fresh. Remake urea stocks and keep everything below 30 °C after urea is added.
• Basic proteins missing entirely — on carrier-ampholyte systems, cathodic drift from over-focusing; on IPG, either an over-long run or loss during equilibration. Check the volt-hour total against the vendor range rather than the clock.
• Vertical smears from a single spot — overload for the detection method. Compare load against the stain-specific table for that strip length and pH range, remembering the tables assume a complex extract, not a purified protein.
• Spots at unexpected positions plus extra faint spots everywhere — keratin contamination. Gloves for every step, and clean strip holders with the correct cleaning solution.
• No current at all, dye immobile — electrode contact, electrode pad hydration, or strip orientation. The dye moving is the only thing bromophenol blue reliably tells you.

Frequently asked questions

What is the difference between isoelectric focusing and native gel electrophoresis?

Native PAGE separates by a combination of charge, size and shape in a buffer of fixed pH, so a band position confounds several properties. IEF separates on a single property — the pH at which net charge is zero — and reaches a steady state where each protein sits at its pI regardless of how long the field is applied (within the limits of gradient stability). Neither is denaturing in the SDS sense, but IEF is usually run under chaotropic denaturation with urea, which native PAGE avoids.

Why can I not use SDS in an IEF sample buffer?

Because SDS binds proteins at a roughly constant mass ratio and imposes a near-uniform negative charge, which destroys the property IEF measures. The Cytiva handbook requires that any SDS used for lysis be either diluted into a large excess of nonionic or zwitterionic detergent, or removed by acetone precipitation, before IEF. Residual SDS is a documented cause of horizontal streaking.

How much salt can my sample contain?

Under rehydration loading, the Cytiva handbook advises keeping the rehydration solution below 10 mM salt. With cup loading, up to 50 mM may be tolerated, though proteins can precipitate at the application point when they meet the lower-salt gel. These are guidance figures for that vendor’s strips; other systems differ, and the tolerable level also depends on how much sample volume you are loading.

Should I use a broad or narrow pH range strip?

Start broad. A narrow range gives markedly better resolution because it flattens dpH/dx, but you need to know where your proteins actually land, and computed pI is not accurate enough to place them — the published prediction error for whole proteins is close to one pH unit. Run pH 3–10 or 3–11 NL first, then choose a narrow range that brackets what you observed. Budget more volt-hours for the narrow strip.

What are volt-hours and why does my protocol specify them?

Volt-hours (usually kVh) are the integral of applied voltage over time, and they measure how much focusing work has actually been delivered. A run whose voltage was suppressed by salt accumulates volt-hours very slowly, so identical elapsed times can mean completely different degrees of focusing. Instruments integrate this automatically and terminate on the target. Guideline totals range from roughly 5–6.5 kVh for a 7 cm pH 3–10 strip to 60–85 kVh for a 24 cm pH 6–9 strip.

Can I store IPG strips after focusing?

Focused strips are normally equilibrated and run in the second dimension immediately, and the Cytiva protocol assumes the second-dimension gel is ready before equilibration begins. Vendor handbooks give preservation guidance for delayed use, but every delay adds opportunities for thiol reoxidation and diffusion of the focused zones. Treat storage as a fallback, not a plan.

Is IEF quantitative?

It can be, with validation. USP ⟨1054⟩ distinguishes three uses of a gel IEF result: an identity test comparing migration pattern to a standard, a limit test comparing band density subjectively, and a quantitative test using densitometry — the last explicitly subject to validation. Capillary formats are considerably more precise: reported inter-laboratory RSDs for icIEF charge-variant peak areas are around 1%, which is why regulated charge-heterogeneity testing has largely moved off gels.

How does IEF relate to two-dimensional gel electrophoresis?

IEF is the first dimension. Proteins are resolved by pI along the strip, the strip is equilibrated into SDS, and it is then laid across the top of an SDS-PAGE gel so the second separation runs perpendicular, by molecular weight. The combination is what gives 2-D gels their resolving power — see the broader overview of gel electrophoresis methods for where it sits among the alternatives.

References

• Cytiva. Immobiline DryStrip — Instructions for Use (29716430 AA, 2022). Strip dimensions, rehydration volumes, sample loads, and volt-hour running guidelines.
• GE Healthcare (now Cytiva). 2-D Electrophoresis: Principles and Methods, Handbook 80-6429-60AC. Immobiline chemistry, rehydration and equilibration solutions, contaminant limits, troubleshooting.
• United States Pharmacopeia. ⟨1054⟩ Biotechnology-Derived Articles — Isoelectric Focusing, Revision Bulletin, official 1 July 2009. Resolution limits for carrier ampholyte versus immobilized gradients, cathodic drift, validation criteria.
• Lomeli G, Herr AE. Reducing cathodic drift during isoelectric focusing using microscale immobilized pH gradient gels. Analytical Chemistry, 2024. PubMed 38716690.
• Kozlowski LP. IPC 2.0: prediction of isoelectric point and pKa dissociation constants. Nucleic Acids Research 2021;49(W1):W285. Benchmark accuracy of predicted versus experimental pI.
• Reduction and alkylation of proteins in 2D gel electrophoresis: before or after isoelectric focusing? Frontiers in Chemistry, 2017. DTT migration away from the basic end and pre-IEF reduction/alkylation.
• Exploring imaged capillary isoelectric focusing parameters for enhanced charge variants quality control. Frontiers in Chemistry, 2025;13:1536222. icIEF ampholyte selection, pI marker calibration and reported precision.