A Step-by-Step Cell Fractionation Protocol: Complete Bench Workflow With Timings

A step-by-step cell fractionation protocol is a differential centrifugation workflow in which cells are lysed in an isotonic buffer, nuclei are pelleted at roughly 600–1,000 × g, mitochondria at 7,000–10,000 × g, and microsomes at 100,000 × g, yielding four to five fractions whose identity is confirmed by marker enzymes or immunoblotting. The entire procedure, from harvest to validated fractions, takes 4–6 hours for cultured cells or liver tissue and 1.5–3 hours for a subcellular re-fractionation. All steps run at 0–4 °C in the presence of protease inhibitors, and each pellet is washed once to reduce cross-contamination.

How Differential Centrifugation Separates Organelles

Centrifugation separates particles by sedimentation velocity, which depends on particle diameter squared, the density difference between particle and medium, and the applied centrifugal field. A nucleus has roughly 300 times the diameter squared of a 0.1 µm vesicle, so it sediments at a field more than two orders of magnitude lower.

The governing relationship is the sedimentation coefficient, expressed in Svedberg units (10⁻¹³ seconds). Intact nuclei sediment near 10⁵–10⁶ S, mitochondria near 10⁴–10⁵ S, and microsomal vesicles near 10³ S. Because sedimentation time scales with the reciprocal of the squared rotor speed, doubling the g-force cuts pelleting time to roughly a quarter.

In plain terms: heavier and bigger structures fall out first, and each spin leaves the lighter material in the supernatant for the next, harder spin.

Rotor geometry matters as much as speed. A fixed-angle rotor pellets faster and forms a tighter pellet than a swinging-bucket rotor at the same g-force. For a protocol that must be reproducible across labs, record the rotor type, the k-factor, and the g-force rather than the rpm, since rpm has no meaning without the rotor radius. Most published fractionation methods express centrifugal force in × g precisely for this reason.

Quantified Fraction Map by Centrifugal Force

The table below reconciles the g-force and time combinations used across the standard literature, including the classic de Duve scheme for rat liver and the protocol from Graham, 2001, used in most teaching labs. It is the reference point for the rest of this protocol.

Reconciling the thresholds: anything below 1,000 × g is a nuclear spin, 7,000–25,000 × g pellets heavy and light mitochondria, and 100,000 × g is the point at which membrane vesicles and ribosomes sediment, so the four numbers expand into five operational fractions.

The classic de Duve nomenclature labels the 100,000 × g pellet as the "microsomal" fraction and the final supernatant as the cytosol. Do not treat these labels as pure: the "mitochondrial" pellet at 10,000 × g reliably co-pellets lysosomes and peroxisomes, which is why a light mitochondrial spin is added when those organelles matter.

Complete Working Protocol From Harvest to Validated Fractions

Start with 5 × 10⁷ cultured cells or 1–2 g of fresh tissue. Pre-chill the homogenizer, rotor, and all buffers to 0–4 °C before starting; temperature drift above 8 °C is one of the most common causes of failed fractions, because it activates proteases and promotes organelle aggregation.

Cell harvest and swelling. Pellet cells at 300 × g for 5 minutes, wash once in ice-cold PBS, and resuspend in hypotonic buffer (10 mM HEPES, 1.5 mM MgCl₂, 10 mM KCl, pH 7.4). Swell on ice for 10–15 minutes; the hypotonic shock weakens the plasma membrane before mechanical lysis.

Homogenization. Dounce-homogenize with a tight pestle, 20–30 strokes for cells or 30–40 for tissue. Check lysis under a phase-contrast microscope: you want 80–90% broken cells with nuclei still intact.

Nuclear spin. Centrifuge at 600–1,000 × g for 10 minutes at 4 °C. The pellet is the nuclear fraction; wash it once in the same buffer and re-spin to remove trapped mitochondria, which reduces nuclear-mitochondrial contamination from about 20% to under 10%.

Mitochondrial spin. Transfer the supernatant and centrifuge at 7,000–10,000 × g for 10–15 minutes. The M pellet is resuspended in 250 mM sucrose buffer for enzyme assays; the wash supernatant is discarded separately.

Microsomal spin. Centrifuge the post-mitochondrial supernatant at 100,000 × g for 60 minutes in an ultracentrifuge. The P pellet contains ER and plasma membrane vesicles; the S supernatant is cytosol.

Validation. Run 10–20 µg of each fraction on SDS-PAGE and immunoblot for the markers in the table above. A clean fractionation shows each marker enriched in exactly one fraction with less than 15% cross-signal elsewhere.

Record the total protein yield per fraction with a BCA assay and calculate the percentage recovery against the input lysate; a good preparation recovers 70–90% of total protein across the fractions.

Step-by-Step Timing Table

The timeline assumes four samples processed in parallel, which is typical for a comparison experiment. Single samples run faster, but the fixed spin times dominate.

For cultured cells the competent end-to-end time is 4–6 hours; tissue takes up to 7 hours because of the longer homogenization and a second low-speed spin to clear connective debris.

Subcellular Fractionation Variations and Edge Cases

Not every sample tolerates the same protocol. Three variations cover the majority of edge cases.

Nuclear isolation in hypotonic buffer. When the nuclear fraction is the goal, a 1.5 mM MgCl₂ hypotonic buffer plus NP-40 at 0.1–0.5% selectively lyses the plasma membrane while leaving nuclei intact; this yields nuclei with under 5% cytoplasmic contamination.

OptiPrep density gradient. When subfractionation of the ER or Golgi is required, a discontinuous iodixanol (OptiPrep) gradient at 10–30% resolves ER and Golgi in about 2 hours, adding substantial time but improving purity for downstream glycomics.

Mitochondria from mouse liver. Use 250 mM sucrose, 10 mM Tris, and 1 mM EDTA and keep the 7,000–10,000 × g range; lower speeds lose 40–50% of mitochondria, and higher speeds introduce ER contamination.

The critical number to respect: at 3,000 × g, more than half of mitochondria remain in the supernatant, so the 600–1,000 × g spin is deliberately short to avoid losing them. Keep total centrifugation time below 3 hours from homogenization to final pellet, because the k-factor translates spin hours into a separation efficiency that degrades with delay.

Common Misconceptions and Red Flags

Three misconceptions dominate student and review coverage of cell fractionation, and each one has a measurable cost.

The first is that a higher g-force gives a cleaner fraction. In practice, exceeding 15,000 × g in the "mitochondrial" spin co-pellets the ER and lysosomes, and the fraction loses its identity; a western blot then shows calnexin in a "mitochondrial" lane.

The second is that fractions are pure. In fact, the nuclear pellet typically carries 5–15% mitochondrial signal, and the mitochondrial pellet carries 10–20% ER. A fractionation is only as good as its contamination profile, and the profile must be reported.

The third is that the marker enzyme is definitive. Individual markers overlap between organelles; COX IV is mitochondrial, but if the blot shows it in both N and M lanes, that is a contamination signal, not a case of two mitochondrial pools.

Red flags: a lysate where fewer than 60% of cells are broken, a nuclear pellet that is loose and hard to resuspend, a microsomal spin completed at less than 100,000 × g, and a fractionation with no marker validation at all.

What to Actually Do Next

Run a small pilot with one sample and validate every fraction by marker immunoblot before committing cells to the larger experiment. This costs 3–4 hours and preempts a failed batch of 8–10 preparations.

Split buffers into single-use aliquots and freeze them once; repeated freeze-thaw of buffers with dithiothreitol loses 20–30% reducing capacity. Add protease inhibitor cocktail fresh each day.

Calibrate the centrifuge against a tachometer or the rotor k-factor at least every 6 months, because an uncalibrated rotor can drift the nominal 10,000 × g by 10–20%.

For protein work, resuspend pellets in 1× Laemmli buffer immediately and freeze at −80 °C; avoid repeated centrifugation of the same sample.

Limitations of the Fractionation Approach

Cell fractionation is an approximation, not a physical dissection. The pellet is a population enriched for a structure, and its purity depends on cell type, buffer, and rotor. Two labs running identical protocols can produce fractions with 10–20% differing composition, which is why relative enrichment across fractions is the interpretable metric rather than absolute abundance.

The method also requires about 5 × 10⁶–5 × 10⁷ cells per fraction to detect low-abundance markers, which scales poorly for rare primary tissues. For single cells and sparse samples, modern imaging and proximity-labeling approaches capture spatial information fractionation cannot, and the two methods are complementary rather than interchangeable.

Finally, the g-force thresholds are operational conventions, not biological boundaries; organelle sizes overlap, and a "pure" 100,000 × g microsomal pellet still contains extracellular vesicles and cytoskeletal fragments. Use the protocol as a reproducible first-pass enrichment, and confirm organelle identity with at least two orthogonal markers before drawing mechanistic conclusions.