Isoelectric Focusing: Trapping Proteins at Their Zero-Charge pH

The zero-charge trap: what isoelectric focusing actually does

Isoelectric focusing (IEF) is an electrophoretic technique that separates amphoteric molecules — proteins, peptides, some nucleotides — according to their isoelectric point (pI), the pH at which the molecule carries no net charge. The trick is to run electrophoresis not in a uniform buffer but inside a stable pH gradient that increases monotonically from the anode (low pH) to the cathode (high pH). Every protein placed anywhere in this gradient experiences a force and drifts, but the direction of that force depends on where it sits.

Consider a protein whose pI is 6.8. If it finds itself at pH 5 (below its pI), it is net-positively charged and migrates toward the cathode — up the gradient. As it climbs, its net positive charge shrinks. When it crosses pH 6.8, its net charge inverts to negative and the field now pushes it back toward the anode — down the gradient. The only stable point is pH = pI, where the net charge is zero, the electrophoretic velocity is zero, and any thermal excursion is corrected by a restoring force pointing back toward the pI. This is the defining feature that distinguishes IEF from ordinary zone electrophoresis: it is a steady-state, focusing method, not a kinetic one. Molecules do not merely separate; they are actively compressed against diffusion into razor-sharp bands, and — crucially — the separation is independent of how the sample was loaded or of small variations in run time.

Because the endpoint is a thermodynamic-style steady state rather than a race, two identical proteins loaded at opposite ends of the gel converge on the same final band. This built-in error correction is why IEF resolves single-charge isoforms — deamidation of one asparagine, one phosphorylation, one sialic acid difference on a glycoprotein — that no size- or hydrophobicity-based method can touch.

The physics of focusing: why the band sharpens instead of spreading

The mechanism follows directly from the electrophoretic mobility. A molecule's steady-state velocity is v = μ(x)·E, where E is the field strength and μ(x) is the effective mobility, which is proportional to the net charge Z(pH) and therefore a function of position x through the local pH(x). At the pI, Z = 0 so μ = 0 and v = 0. Svensson (later Svensson-Rilbe) formalized this in 1961 by writing the focusing condition: the band is stable only where the slope of mobility with position is negative,

• dμ/dx < 0 at x = x(pI) — a molecule pushed cathodically (to higher x) must acquire a restoring mobility pointing back anodically, and vice versa.

At steady state, electrophoretic transport toward the pI is exactly balanced by Fickian back-diffusion away from the concentrated band. Solving this transport balance gives a Gaussian concentration profile whose standard deviation is

σ = √( D / ( E · (dμ/d(pH)) · (d(pH)/dx) ) )

Every term in the denominator tells you how to sharpen a band: a steeper local pH gradient d(pH)/dx (use a narrow-range gel), a higher field E, and a protein whose mobility changes steeply with pH near its pI (a large −dμ/dpH, i.e. a good titration slope). The numerator, the diffusion coefficient D, works against you but only as a square root. This is why IEF bands are so much narrower than zone-electrophoresis zones: σ scales as E^(−1/2), so doubling the field narrows the band by ~30% while quadrupling the field halves it.

The resolving power follows from combining band width with gradient slope. In its practical form the minimum resolvable pI difference is

ΔpI(min) ≈ 3·√( D·(d(pH)/dx) / ( E·(−dμ/dpH) ) )

which under routine analytical conditions lands near 0.01–0.02 pH units for carrier-ampholyte gels and can reach ~0.001 in ultra-narrow immobilized gradients spanning only 0.1 pH unit over several centimeters.

Building the gradient: carrier ampholytes versus Immobilines

The entire method hinges on one hard problem: how do you build a pH gradient that is stable under an electric field? A field drives ions, so any gradient made of mobile buffer ions would simply be swept away. Two solutions dominate the field.

Carrier ampholytes (CAs). Harry Svensson's insight was to use the analytes' own trick against them. If you fill the gel with a heterogeneous mixture of hundreds of small amphoteric buffer molecules, each with its own pI spread finely across (say) pH 3–10, then when the field is applied each ampholyte focuses at its own pI — exactly like a protein. Because they are present in bulk and have good buffering capacity right at their pI, they collectively lay down a smooth, self-organizing, stable pH staircase. Olof Vesterberg's 1969 synthesis — reacting polyethylene-polyamines with acrylic acid to make a statistical mixture of oligoamino-oligocarboxylic acids — turned this into a commercial reagent (Ampholine, later Servalyte, Pharmalyte). The weakness is that CAs are only quasi-stationary: over long runs the whole plateau slowly migrates toward the cathode (cathodic drift), flattening the basic end and losing basic proteins.

Immobilized pH gradients (IPG). In 1982 Pier Giorgio Righetti, Bjellqvist and co-workers eliminated the drift by grafting the buffering groups directly onto the gel backbone. Their Immobiline reagents are acrylamido derivatives — CH₂=CH–CO–NH–R, where R carries a single weak acid (–COOH) or weak base (tertiary amino) group of precisely known pKₐ. During gel casting, a gradient mixer blends two Immobiline solutions so the concentration of each buffering species varies linearly along the gel; the Henderson–Hasselbalch balance of grafted acids and bases then defines a fixed pH at every position. Because the buffering groups are copolymerized into the polyacrylamide, they cannot migrate: the gradient is welded in place. IPG strips (marketed as ReadyStrip / Immobiline DryStrip) give reproducible, drift-free gradients, high loading capacity, and the extreme narrow-range resolution that carrier ampholytes cannot sustain.

A worked example: predicting a protein's pI and its band width

Take β-lactoglobulin A, a well-studied whey protein with a measured pI near 5.1–5.2. Its net charge as a function of pH is the sum over all ionizable groups of their fractional charges, each given by Henderson–Hasselbalch. For a cationic group (Lys, Arg, His, N-terminus) the fractional positive charge is 1/(1 + 10^(pH−pKₐ)); for an anionic group (Asp, Glu, Cys, Tyr, C-terminus) the fractional negative charge is −1/(1 + 10^(pKₐ−pH)). The pI is the pH at which Σ(charges) = 0 — found by solving Z(pH) = 0 numerically (there is no closed form for a real protein with dozens of groups). Modern tools (ExPASy Compute pI/Mw, using Bjellqvist's empirically corrected pKₐ set derived from IEF data itself) do exactly this and reproduce measured pIs to within a few hundredths of a unit for most soluble proteins.

Now estimate the focused band. Near its pI a protein's charge changes roughly linearly with pH; for a typical small globular protein like β-lactoglobulin −dμ/dpH is of order 3×10⁻⁹ m²·V⁻¹·s⁻¹ per pH unit (an order-of-magnitude estimate, not a tabulated value). On a 7-cm IPG strip spanning pH 4–7, the gradient slope d(pH)/dx ≈ 3/0.07 = 43 pH units per meter. With D ≈ 7×10⁻¹¹ m²·s⁻¹ and a field E ≈ 200 V/cm = 2×10⁴ V/m, the steady-state band standard deviation is σ = √(D / (E·(dμ/dpH)·(d(pH)/dx))) ≈ √(7×10⁻¹¹ / (2×10⁴·3×10⁻⁹·43)) ≈ 1.6×10⁻⁴ m, about 0.16 mm. That figure — a sub-millimeter band on a 70-mm strip — is exactly what makes IEF the workhorse first dimension: hundreds of resolvable positions across one strip.

The practical stopping rule is volt-hours. Focusing is complete when the integral ∫E·dt has driven every protein to its steady state; typical analytical IPG runs accumulate 10,000–60,000 V·h. Under-focusing leaves streaks (proteins still migrating); vast over-focusing risks water splitting and precipitation of proteins at their pI, where their solubility is minimal — a subtlety we return to below.

Limits, artifacts, and the awkward truth about pI at zero charge

IEF is powerful but riddled with subtleties that reward a careful operator.

• Proteins are least soluble at their pI. Zero net charge means minimal electrostatic repulsion between molecules, so aggregation and precipitation peak precisely where you are trying to focus them. This is why IEF is run with high concentrations of a neutral chaotrope (7–9 M urea, often plus thiourea) and a zwitterionic detergent such as CHAPS to keep hydrophobic and membrane proteins in solution.
• The measured pI is operationally defined. A protein's pI depends on ionic strength, temperature, and bound ligands, and the value reported by IEF is the pH read in the gel under those conditions — not a platonic constant. Urea, for example, perturbs pKₐ values and shifts apparent pIs upward by several tenths of a unit, which is why one always calibrates with pI marker proteins run alongside.
• Cathodic drift and plateau flattening plague carrier-ampholyte gels: electroendosmosis and the slow migration of the ampholyte plateau erode the basic end, so proteins with pI > 9 (histones, ribosomal proteins) are notoriously hard to see. IPGs largely cure this but very basic ranges (pH 9–12) still demand special Immobiline chemistries.
• Carrier ampholytes interfere downstream. They absorb in the UV, bind some proteins, and suppress signal in mass spectrometry, so they must be washed out before MS — one more reason the empty IPG matrix is preferred for proteomics.

There is also a conceptual caveat worth stating plainly: pI is not the pH of zero absolute charge but of zero net charge. At its pI a protein is a bristling collection of positive and negative groups (a zwitterion writ large) whose charges happen to cancel. Two proteins can share a pI while differing enormously in charge magnitude and titration behavior — which is exactly why IEF's resolution depends on −dμ/dpH (the steepness of the titration curve) and not merely on the pI value itself.

From clinical bands to the first dimension of proteomics

The reason IEF earned a permanent place in the analytical canon is its coupling to a second, orthogonal separation. In 1975 Patrick O'Farrell published two-dimensional gel electrophoresis (2D-PAGE): run IEF first (separating by pI), then run SDS-PAGE perpendicular to it (separating by mass). Because pI and molecular weight are essentially uncorrelated, the two dimensions multiply to give a map on which a single mammalian cell lysate resolves into thousands of distinct spots. O'Farrell's original resolution of ~1100 Escherichia coli proteins on one gel was, at the time, an astonishing leap and effectively founded quantitative proteomics. The switch to IPG strips as the first dimension (Bjellqvist and Righetti's technology, adopted through the late 1980s–90s) made 2D-PAGE reproducible enough for inter-laboratory databases and, later, robust coupling to MALDI and ESI mass spectrometry for spot identification.

Beyond proteomics, IEF is a clinical and industrial standard. Clinical labs use it to detect hemoglobinopathies — HbS, HbC and HbE separate cleanly from HbA by their pI shifts — and to resolve oligoclonal bands in cerebrospinal fluid, a cornerstone diagnostic for multiple sclerosis. In biopharmaceutical quality control, capillary IEF (cIEF) and imaged cIEF have become the reference methods for the charge-heterogeneity profile of monoclonal-antibody therapeutics: deamidation, C-terminal lysine clipping, and sialylation each shift the pI, and regulators expect a fingerprint of these charge variants for every batch. cIEF replaces the gel with a fused-silica capillary, focuses in minutes at fields up to ~500 V/cm, and then mobilizes the focused zones past a UV detector by chemical or hydrodynamic means — marrying the resolving power of Svensson's 1961 steady state to the speed and quantitation of modern instrumentation.

Carrier-ampholyte IEF versus immobilized pH gradient (IPG) focusing

Property	Carrier ampholytes (CA)	Immobilized pH gradient (IPG)
pH-forming species	Soluble amphoteric buffers that migrate and self-arrange	Acrylamido buffers (Immobilines) covalently grafted to the gel
Gradient stability	Drifts (cathodic drift) over long runs; plateau flattens	Fixed and reproducible; no drift, run indefinitely to steady state
Maximum resolution	~ΔpI 0.02	~ΔpI 0.001 in ultra-narrow (0.1 unit) ranges
Loading capacity	Modest; ampholytes co-elute and can complex analytes	High; empty matrix, good for prep-scale and MS
Inventors / year	Svensson-Rilbe theory 1961; Vesterberg synthesis 1969	Righetti, Bjellqvist et al., 1982