Electrospray, MALDI, and the choice that shapes every spectrum after it

Ask a laboratory which ionisation source produced a spectrum and you learn a great deal about the spectrum before you have looked at it. Electrospray on a modern instrument will give a peptide of four thousand daltons a family of multiply charged ions and a mass accuracy in the low parts per million. A linear MALDI time-of-flight will give a single predominantly singly charged ion and an accuracy measured in hundreds of parts per million. Both can answer the question, is this the right compound. Only one of them can answer the question, is this the right compound and nothing very close to it.

The instrument does not weigh anything

It is worth being exact about what a mass spectrometer does, because the imprecision propagates. The instrument generates ions from a sample, separates them according to the ratio of their mass to their charge, and counts them at a detector. The horizontal axis of every spectrum is mass-to-charge, conventionally written m/z and expressed in thomsons or in dimensionless units depending on the vendor’s software. Nothing is weighed. Nothing is measured against a reference mass in the sense that a balance measures against a calibration weight.

What follows from this is that every molecular weight on every certificate of analysis in this market is a calculated quantity, derived from a measured m/z by assigning a charge and subtracting the mass contribution of whatever adducted to the molecule to give it that charge — usually protons, sometimes sodium, occasionally potassium or ammonium. The assignment is normally straightforward and normally correct. It is nonetheless an assignment, and when it goes wrong it goes wrong by an integer factor, which is the kind of error that produces confident nonsense rather than a plausible discrepancy.

The practical consequence for a reader is a habit: when a mass figure appears, ask what was observed and what was inferred. A report that gives both — the m/z, the charge, and the derived neutral mass — has answered the question before it was asked.

Electrospray, and why it produces a family of ions

Electrospray ionisation works by pumping a solution of the analyte through a fine capillary held at a potential of a few kilovolts relative to the instrument’s entrance. The liquid emerging from the tip forms a cone and then a jet of charged droplets. As solvent evaporates the droplets shrink, the charge density on their surfaces rises, and at the point where electrostatic repulsion exceeds surface tension they fission into smaller droplets. Repeat this enough times and what is left is a bare, charged analyte ion in the gas phase.

Because the charge is acquired in solution and retained through desolvation, a peptide with several basic residues will carry several protons, and the population of ions reaching the analyser is distributed across charge states. This is the defining characteristic of electrospray spectra and the reason they look bewildering to a first-time reader: a single pure compound produces four or five prominent peaks, none of them at the molecular weight.

The distribution is not noise. It carries information about the number of accessible basic sites and about the conformational state of the molecule in solution, and it shifts predictably with mobile-phase composition and pH. It also has a practical advantage that matters for identity work: dividing the mass by three or four brings a large peptide into the range where instruments achieve their best resolving power and accuracy.¹

MALDI, the matrix, and the singly charged ion

Matrix-assisted laser desorption ionisation takes a different route. The peptide is mixed with a large molar excess of a small organic compound that absorbs strongly at the laser wavelength — α-cyano-4-hydroxycinnamic acid and sinapinic acid are the usual choices for peptides and proteins respectively — and the mixture is dried on a metal target. A pulsed ultraviolet laser strikes the crystal, the matrix absorbs the energy, and a plume of matrix and analyte is ejected into the vacuum with the analyte largely intact and mostly singly protonated.

Two consequences follow. First, MALDI spectra are simple: one predominant ion per compound, at the molecular weight plus one proton, which makes them easy to read and easy to print on a certificate. Second, MALDI is markedly more tolerant of salts, buffers and heterogeneous samples than electrospray, which is why it survives in routine synthesis monitoring where electrospray would require a chromatographic clean-up first.

The trade-offs are equally real. The matrix produces intense chemical background below roughly a thousand daltons, which obscures small fragments. Ion yield varies between compounds and between spots on the same target, making MALDI a poor quantitative technique. And the achievable mass accuracy on a linear instrument at peptide molecular weights is measured in hundreds of parts per million unless a reflectron and delayed extraction are in use.²

Which source a laboratory chooses, and why it should say so

The choice between the two techniques is not a matter of quality but of question. A synthesis chemist watching a coupling proceed wants a fast, salt-tolerant check that the chain has grown by the expected residue, and MALDI on a bench instrument answers that in minutes. An analytical laboratory asked whether a submitted vial contains the labelled compound and nothing closely related to it needs the resolving power and the accuracy that electrospray into a high-field analyser provides, coupled to a chromatographic separation so that species which co-elute can at least be assigned to retention times.

Both appear in this market, and reports rarely distinguish them. That matters because the two techniques have different blind spots. MALDI can induce loss of labile modifications during desorption, so a phosphorylated or otherwise fragile species may be under-represented. Electrospray suppresses ionisation of some analytes in the presence of others, so a minor component of a mixture may be absent from a spectrum in which it is genuinely present.

A certificate stating the source therefore tells a reader which class of error to consider. The Journal has stopped asking suppliers for more testing and started asking them for this line instead, on the grounds that it costs nothing and changes what the existing test can be said to support.

The isotope pattern, and how it declares the charge

Because carbon-13 is present at roughly 1.1% natural abundance, a peptide containing one hundred and ninety carbon atoms will exist substantially as molecules containing one, two or three carbon-13 atoms. In a spectrum this appears as a series of peaks above the monoisotopic peak, separated in mass by approximately 1.00336 daltons and distributed in intensity according to the binomial statistics of the composition.

Two things follow, and both are practically useful. First, the spacing between adjacent isotope peaks in a charge-state cluster is one over the charge: a spacing of 0.5 on the m/z axis means the ion is doubly charged, 0.333 means triply, 0.25 means quadruply. This is the simplest charge assignment available and it requires no assumptions about the sample at all. Second, the relative intensities of the isotope peaks are predictable from the elemental composition, so a cluster whose shape departs markedly from the calculated envelope is evidence that two species are overlapping.

Both observations require an instrument capable of resolving the isotope peaks at the relevant m/z, which is where resolving power stops being a specification-sheet number and becomes the thing that determines whether a spectrum can be interpreted at all. Below roughly ten thousand resolving power, a multiply charged peptide envelope collapses into a single broad hump that carries neither the spacing nor the shape information.

Resolving power, defined and then converted

Resolving power is conventionally defined as m divided by Δm, where Δm is the width of the peak at half its maximum height. An instrument quoted at 30,000 resolving power at m/z 1000 produces peaks roughly 0.033 wide at that position, which is sufficient to separate the isotope peaks of a triply charged ion. The same instrument at m/z 4000 may deliver rather less, because resolving power is not constant across the mass range and the figure on a specification sheet is quoted at whichever mass flatters it.

The number that matters for identity work is whether the instrument can separate two species whose masses differ by the amount you care about. To distinguish a deamidated peptide from its parent at four thousand daltons requires separating peaks 0.98 apart, which is a resolving power of roughly four thousand — modest. To distinguish a glutamine-for-lysine substitution requires separating peaks 0.036 apart at the same mass, which is a resolving power above one hundred thousand. Those two requirements differ by a factor of twenty-five and both are described in the trade by the same phrase, high resolution.

The Journal’s practice, adopted after an exchange with a laboratory that pointed out we had been sloppy about it, is to state the discrimination rather than the specification: not “high-resolution MS”, but “sufficient to resolve a 0.98-dalton shift at the parent mass”. It is longer and it says something.³

The instrument classes, and what each can support

A single quadrupole mass filter provides unit resolution and mass accuracy of a few tenths of a dalton. It is entirely adequate to confirm that a sample is broadly the compound expected and to detect large modifications, and it is the analyser in most low-cost LC-MS systems. It cannot resolve an isotopic envelope at peptide molecular weights and therefore cannot assign charge from spacing.

Time-of-flight analysers separate ions by the time they take to traverse a flight tube. A linear tube gives modest resolving power; adding a reflectron and delayed extraction raises it into the tens of thousands, and modern quadrupole time-of-flight hybrids achieve low single-figure parts-per-million accuracy with routine calibration. Orbital trapping instruments measure the frequency of ion oscillation in an electrostatic field and convert it by Fourier transform, delivering resolving powers from sixty thousand to several hundred thousand and sub-part-per-million accuracy with internal calibration. Fourier-transform ion cyclotron resonance remains the highest-performing class and the least common outside academic facilities.

What this hierarchy means for a reader of certificates is that the instrument named on the document sets a ceiling on what the document can claim, independent of the laboratory’s competence. An unnamed instrument leaves that ceiling unknown, which is why the Journal now treats the absence of an instrument name as a material omission rather than a stylistic one.

Oxidation: sixteen daltons, and the one that gets caught

Methionine oxidises to the sulphoxide with a mass increase of 15.995 daltons, and on to the sulphone at a further 15.995. Tryptophan and histidine oxidise by related routes. The chemistry is driven by dissolved oxygen, by trace peroxides in excipients and in some grades of polysorbate, by light, and by transition-metal contamination, and it proceeds in lyophilised material as well as in solution, though more slowly.

Analytically this is the easy case, and it deserves to be described as such in an article otherwise concerned with what cannot be seen. A sixteen-dalton shift is resolvable on essentially any instrument, and the oxidised species is usually chromatographically distinct enough to appear as a separate peak under a reasonably shallow gradient. A spectrum showing a plus-sixteen satellite at a few per cent of the parent intensity is unambiguous evidence of oxidation, and its absence is meaningful evidence of the opposite.

Which is why the Journal’s standing request to laboratories in this market is for the spectrum rather than the verdict. A reproduced spectrum, even at the modest resolution of a routine instrument, allows a reader to look for the plus-sixteen satellite themselves. A conformance statement does not, and cannot be made to. This is the cheapest available improvement to identity reporting in the trade and it consists of printing a picture the laboratory has already produced.

How an identity result travels once it leaves the laboratory

Follow a mass spectrum through the market and its meaning changes at every step. A laboratory issues a report to whoever submitted the sample, stating what was observed on a named instrument on a named date. The submitter — a vendor, in most cases — extracts a figure and a verdict onto a certificate of analysis for the lot. A reseller reproduces the certificate, or a portion of it. A listing page distils the whole chain into a phrase: identity verified.

Nothing dishonest need happen at any step for the final phrase to support far more than the original report does. The instrument’s resolving power is lost at step two. The convention behind the theoretical mass is lost at step two or three. The date, the batch and the submitter’s identity survive unevenly. By the time the claim reaches a buyer it has become a property of the product rather than a record of a measurement on one vial from one lot on one day.

This is the structural reason the Journal reports identity claims by asking for the underlying laboratory report rather than the certificate. When a supplier supplies it, the claim usually holds up and often turns out to be stronger than the certificate suggested. When a supplier cannot locate it, that is itself information about how far back the documentary chain reaches, and we report that too, without inferring anything about the material.

The standing rule in this department

This publication applies one rule to every identity claim it reports, and it is worth stating in isolation because it governs the rest. A mass measurement supports a statement about composition. Only a fragmentation or mapping experiment supports a statement about sequence. Where a source says identity was confirmed, we report that a mass was measured, unless we have seen evidence of the second kind.

The rule has consequences we accept. It makes our coverage read as more sceptical than the underlying documents, because the documents claim more than they establish. It occasionally irritates laboratories which have in fact done sequence-level work and have simply not printed it, and the remedy there is a two-line email which we are glad to receive. And it means we cannot describe any research-grade vial in this market as sequence-confirmed, because on the evidence available to us almost none are.

What the rule is not is an accusation. Nothing in this article suggests that vendors are selling material other than what they label, and the Journal has no evidence of that in respect of any company it covers. The claim is narrower and, we think, harder to argue with: the documentation in general circulation does not have the discriminating power that the language on it implies, and the gap between the two is where every avoidable dispute in this market begins.

If this market spent one more pound on identity, where should it go

A fair question, and the Journal’s answer has changed. Our first instinct was to argue for sequence confirmation on every lot, and the arithmetic does not support it: peptide mapping on every batch would raise the analytical cost per vial by a multiple, and the failure mode it protects against — a wholly substituted or permuted sequence — is not the one we see evidence of.

The better allocation, on our present assessment, is orthogonal. Identity by high-resolution intact mass on every lot, at a resolving power sufficient to resolve a one-dalton shift at the parent mass, with the spectrum reproduced. Sequence confirmation once per synthesis campaign rather than once per lot, on the reasoning that the sequence is a property of the process and the lot-to-lot risk is degradation rather than misconstruction. And a chromatographic method shallow enough to separate the deamidated form, because that is the change most likely to have occurred between the certificate and the buyer.

That package is not expensive. Two of the twenty companies in our dossier programme already do something close to the first item, and one has told us it is costed for the second. Whether any of it happens depends on whether buyers ever ask, which is a market question rather than a scientific one and is therefore the harder of the two.

Readers who take one thing from this piece should take the arithmetic. Isotope spacing is one over the charge. Deamidation is one dalton, oxidation is sixteen, and stereochemical inversion is nothing at all. A tolerance of ±1 dalton on a four-thousand-dalton peptide is two hundred and forty parts per million and excludes almost nothing worth excluding. Those four facts are sufficient to read most of the identity claims in circulation, and they fit on the back of an envelope.