How Are Isotopes Used In Biology

Isotope labeling in the lab is mostly a matter of patience and contamination control

I spent three years running metabolic flux experiments with carbon-13 labeled glucose before I stopped being annoyed by the occasional phantom peak in my mass spectra. The issue wasn't the isotope itself. It was the lab water, which apparently had a constant background of dissolved organic carbon with a C-13 signature that looked suspiciously like incomplete metabolic turnover. You learn quickly to treat every reagent as a potential source of contamination, not just the sample. The fundamental trick is simple. You feed cells, animals, or organisms a molecule where one or more atoms are replaced with a heavier or radioactive isotope, then track where that atom ends up. A normal carbon atom weighs 12 units. Carbon-13 weighs 13. Carbon-14 weighs 14 and decays. When you feed something glucose labeled at position 1 with C-13, you can watch that specific carbon atom move through glycolysis, the TCA cycle, and into amino acids or fatty acids simply by measuring the mass shift in extracted metabolites using a mass spectrometer. The instrument detects the mass difference. Everything else is chemistry and careful sampling.

How Are Isotopes Used In Biology

Tracer studies are the most common application. You introduce a labeled compound and follow its path through a biological system. This is how we mapped out most of the intermediates in central metabolism. Before isotope tracers, you were guessing at pathway connections based on enzyme presence. Now you can literally watch a labeled carbon appear in lactate, then alanine, then glutamate, with time-resolved samples taken every few minutes. The pattern of labeling tells you which pathways are active and at what relative rates. Radiometric dating uses the decay of isotopes like C-14 to determine the age of organic material. This works reliably up to about 50,000 years. Beyond that, there isn't enough C-14 left to measure accurately, and you switch to other systems like uranium-lead or potassium-argon for geological timescales. The half-life of C-14 is roughly 5,730 years, which makes it useful for archaeology and recent evolutionary questions but useless for anything older than the last glacial period. Imaging with isotopes has gotten much better in the last decade. Secondary ion mass spectrometry, or SIMS, can map isotope distributions at subcellular resolution. NanoSIMS pushes this further, resolving individual bacterial cells within a biofilm and showing which ones are actively incorporating labeled nutrients. This matters because not every cell in a population is doing the same thing. Bulk measurements hide that variation entirely.

MEDICAL AND DIAGNOSTIC USES are arguably where isotopes touch the most lives. The urea breath test for Helicobacter pylori uses C-13 or C-14 labeled urea. The patient swallows it, and if H. pylori urease is present in the stomach, it breaks the urea down, releasing labeled CO2 that shows up in the breath. Simple, non-invasive, and accurate. The C-14 version uses a tiny amount of radioactivity—less than a chest X-ray—but many clinics have switched to C-13 because it requires no radiation safety protocols and the instrument cost has come down significantly.

Practical considerations that textbooks don't always emphasize

The biggest problem people encounter isn't understanding the concept. It's that isotope experiments are unforgiving of sloppy technique. I once wasted an entire week of culture time because I used glass fiber filters for sample collection that had been stored near a C-13 enriched compound in another lab's freezer. The filters absorbed enough vapor to shift the baseline of every sample processed on that day. We caught it when the unlabeled controls showed anomalous enrichment. That's when you remember to ask where everyone in your building has been storing their labeled stocks. Another issue is isotope fractionation. Biological systems prefer lighter isotopes. Enzymes that use C-12 react slightly faster than those using C-13, simply because the lighter bond is easier to break. This means that when you measure the isotopic composition of a product, part of the signal reflects actual metabolic routing and part reflects kinetic isotope effects that have nothing to do with pathway activity. For most flux analysis work, this effect is small—maybe one to three per mil—but if you're doing high-precision stable isotope probing or studying slow-turnover systems, it can look like a real biological signal. You account for it by running unlabeled controls alongside every experiment and measuring the natural fractionation baseline for your system. Radiation safety is another practical concern that gets short-changed in introductions. P-32 is widely used in molecular biology for labeling nucleic acids, and it's a pure beta emitter with a maximum energy of 1.7 MeV. That means it requires proper shielding—acrylic or Plexiglas, not lead, because high-energy betas produce bremsstrahlung X-rays when they hit dense materials. I've seen people use lead shields for P-32 work, which is worse than using nothing at all. The bremsstrahlung from a lead shield can expose you to more radiation than the beta particles themselves. Use acrylic, keep it thick, and monitor properly.

When working with stable isotopes like C-13 and N-15, the main cost is the labeled compounds themselves. Uniformly labeled C-13 glucose runs about $80 to $150 per gram depending on the supplier and enrichment level. You typically need 10 to 50 millimolar final concentration in culture media, which means you're burning through money fast if you're doing time courses with multiple time points. Some labs recover and reuse labeled media through filtration, but that introduces dilution and cross-contamination risks that are hard to quantify. Most people just buy fresh labeled media and budget accordingly.

Where the method fails completely

Isotope tracing doesn't work well when the pool you're trying to label is enormous and turns over slowly. If you're studying lipid storage in adipose tissue, feeding C-13 glucose won't tell you much about newly synthesized fat because the adipose lipid pool is largely static and the conversion from glucose to fatty acid is inefficient in vivo. You'd be better off using C-13 acetate, which enters lipid synthesis more directly, or switching to a different approach entirely, like measuring deuterium incorporation from heavy water, which labels all newly synthesized macromolecules regardless of the precursor pathway. S-35 methionine is commonly used for pulse-chase experiments to study protein turnover, but the specific activity of commercially available S-35 methionine has dropped significantly in recent years. Supply chain issues and declining demand for some applications have made it harder to get high-specific-activity material. When I last ordered it, the best I could find was about 1,000 Ci/mmol instead of the 1,000+ range that was standard a decade ago. This means you need to use more radioactivity to get the same signal, which creates its own set of handling and disposal problems. D-lysine or D-leucine from heavy water experiments has become a reasonable alternative for many protein turnover studies, though the math is more complex. The assumption that all carbon in your sample comes from the labeled substrate is another trap. If you're growing bacteria in minimal media with C-13 glucose as the sole carbon source, you might think everything gets labeled. But if there's any atmospheric CO2 dissolution or contamination from unlabeled carbon sources in your water or salts, you'll get a mixed label pattern that looks like partial turnover when it's actually just dirty reagents. Always run a no-glucose control to check for background labeling, and analyze your water and salt stocks for organic carbon content before starting a serious experiment.

Mass spectrometry considerations

The detector matters more than most people expect. A typical GC-MS can resolve mass differences of 1 Dalton easily, which is fine for tracking individual labeled atoms in small metabolites. But when you're looking at a peptide with 10 carbons and you want to know how many of those 10 are labeled, you need resolution that can distinguish M+0 from M+1 from M+2, and so on, across a mass range that might be several hundred Daltons. An orbitrap or a high-resolution Q-TOF handles this much better than a quadrupole. With a quadrupole, the isotope envelope of a heavily labeled peptide can merge into an unreadable blob. You'll know the sample is labeled, but you won't be able to parse the distribution accurately. Ionization method also affects what you can measure. Electrospray ionization preserves the intact molecule and works well for polar metabolites and peptides. But if you're trying to measure isotopic enrichment in structural lipids or hydrophobic compounds, ESI can suppress ionization significantly. You might need to switch to atmospheric pressure chemical ionization or derivatize the compounds before injection. I've seen people spend weeks trying to get clean isotope ratio data on lipid extracts with ESI alone, only to realize the issue was ion suppression from co-eluting matrix compounds. A simple solid-phase extraction cleanup step before analysis solved the problem in two hours. There's also the question of how you process the raw data. Isotope enrichment isn't just reading a peak height. You need to deconvolute the natural isotope distribution from the labeled one. Every element has a natural isotope pattern—C-13 occurs at about 1.1% naturally, N-15 at 0.37%. If you're doing C-13 labeling studies, you're working on top of that existing background. Most software packages handle this correction automatically, but the algorithms assume a certain level of mass resolution and signal-to-noise. If your data is noisy, the correction can introduce more error than it removes. Check the correction output manually for a few samples before trusting the automated pipeline with your full dataset.

When to use radioactivity versus stable isotopes

Radioactive isotopes are more sensitive. You can detect femtomole quantities with a scintillation counter. Stable isotope labeling with MS typically needs picomole to nanomole ranges depending on the instrument and sample type. If you're working with rare samples—biopsy material, single cells, environmental samples with low biomass—the sensitivity advantage of radioactivity can be decisive. The tradeoff is regulation, disposal cost, and safety training. A C-13 experiment requires a mass spectrometer and some consumables. A P-32 experiment requires a licensed radiation safety officer, dedicated workspace, waste disposal contracts, and annual training updates. For teaching labs and routine undergraduate experiments, the classic S-35 methionine pulse-chase or the H-3 thymidine incorporation assay for cell proliferation remain popular because the detection equipment is inexpensive and the results are dramatic. A single autoradiograph shows clear bands or spots that students can interpret without needing access to a mass spectrometer. But for research-grade work where you need quantitative flux data, stable isotope labeling with isotope ratio mass spectrometry or LC-MS/MS is the standard. The data is richer, there's no regulatory overhead, and the samples can be archived and reanalyzed indefinitely without decay concerns. I still keep a small stock of C-14 tagged compounds in the freezer for the occasional experiment where the sensitivity justifies the paperwork. Most of the time, C-13 does the job and the extra compliance burden isn't worth it. The labeling patterns are just as informative if your instrument is calibrated and your samples are clean.