Profiling Proteins in Individual Bacteria
Label-free LC–MS pushes proteomics below the picogram scale, detecting stress-induced protein changes in individual bacteria.
A label-free single-cell proteomics workflow has pushed mass spectrometry to the scale of individual bacteria, detecting stress-induced protein changes in single Bacillus subtilis cells and tentative differences in how individual cells respond.
Bacterial single-cell proteomics presents an unusually demanding sensitivity problem. A bacterium contains less than 1 pg of protein – roughly three orders of magnitude less by volume than a typical eukaryotic cell – while its cell wall complicates lysis and proteins introduced during sample preparation can overwhelm the endogenous signal.
The researchers developed their bacterial single-cell proteomics workflow, termed bacSCP, by adapting a low-volume, one-pot preparation method for bacterial cells. Fluorescent staining helped reliably sort individual cells into 384-well plates, while repeated freeze-thaw cycles lysed intact bacteria before label-free LC–MS analysis.
Applied to individual Escherichia coli and B. subtilis cells, bacSCP typically quantified more than 50 bacterial proteins, predominantly abundant metabolic enzymes and ribosomal proteins. To test whether that coverage could capture a biological response, the team analyzed individual B. subtilis ΔmcsB cells before and after heat shock, detecting significant increases in the protein-quality-control factors GroEL, GroES, and ClpC.
The single-cell measurements also suggested that the response was uneven: four cells showed particularly strong increases in heat-shock proteins, forming a distinct cluster. However, technical variability is substantial at these protein amounts, and the authors caution that some apparent cell-to-cell differences could therefore be analytical rather than biological.
Greater sensitivity and larger cell populations will be needed to separate those sources of variation and reach less abundant proteins.
PFNet Pushes HX-MS to Residue Level
A machine-learning framework infers residue-level opening free energies from peptide-resolved HX-MS data in seconds.
A machine learning framework has extracted residue-level protein stability from conventional hydrogen exchange mass spectrometry (HX-MS) data in seconds, turning peptide-level measurements into quantitative maps of the energetic changes associated with protein interactions and mutations.
HX-MS tracks hydrogen isotope exchange along the protein backbone, but its spatial resolution is usually limited by the peptides generated during digestion. PFNet was developed to overcome that limitation by inferring the free energy of local opening (ΔGop) for individual residues directly from HX-MS isotopic envelopes.
Trained on synthetic datasets designed to reproduce variations in peptide coverage, experimental noise, timepoints, and back exchange, the model also assigns confidence scores to its predictions. Against a synthetic benchmark, PFNet outperformed an existing Bayesian approach while reducing analysis times from tens of hours to seconds.
Tests against hydrogen-exchange NMR provided an experimental benchmark. Across two designed mini-proteins and T4 lysozyme, PFNet-derived ΔGop values showed correlations of 0.63–0.91 with the NMR measurements. Unlike NMR, which is difficult to apply to large proteins and complexes, PFNet requires neither a protein structure nor additional biophysical information.
The researchers then applied PFNet to a much larger system, revisiting existing HX-MS data for the interaction between the SARS-CoV-2 spike protein and ACE2. Conventional peptide-level analysis showed broad protection around the receptor-binding domain; PFNet localized the strongest stabilization to individual residues including Gln498 and Gln493, while also resolving coupled destabilization elsewhere in the domain.
The authors acknowledge that more realistic experimental noise models should improve PFNet’s robustness across HX-MS platforms and protein systems. The next test will be how consistently it can recover residue-level energetics from the messier experimental data encountered across different proteins and laboratories.
N-Glycoproteomics in a Single Tube
A single-pot magnetic-bead method streamlines N-glycopeptide preparation across plasma, serum, and tissue samples.
A magnetic bead-based workflow has simplified the preparation of N-glycopeptides for mass spectrometry, combining protein cleanup, digestion, and glycopeptide enrichment in a single tube.
Glycopeptides ionize relatively poorly, so abundant non-glycosylated peptides and contaminants such as salts and detergents can interfere with their detection by mass spectrometry. For large clinical cohorts, sample preparation therefore needs to combine effective enrichment with a workflow suited to automation.
The researchers combined single-pot solid-phase-enhanced sample preparation (SP3) with hydrophilic interaction liquid chromatography (HILIC), using the same magnetic particles for protein cleanup and digestion followed by N-glycopeptide enrichment. Cellulose-based particles recovered glycopeptides more efficiently and removed more non-glycosylated peptides than the carboxylated polymer beads conventionally used for SP3.
“The SP3-based N-glycopeptide preparation method using commercially available cellulose magnetic beads is a robust and automate-friendly approach for N-glycoproteomics,” said Kazuki Nakajima of Gifu University in a press release.
Tested on plasma, serum, and tissue, the workflow was also applied to clinical samples from patients with gastric cancer. Serum analysis detected disease-associated changes including increased fucosylated glycopeptides from several acute-phase proteins, while tumor tissue showed broader changes in site-specific glycosylation. The small cohorts – 10 patients for serum and paired tissues from eight – make these proof-of-principle applications rather than biomarker validation.
Gastric tissue also exposed a limitation. Unlike rat brain samples, gastric extracts impaired nano-LC performance, an effect reproduced with porcine gastric mucin. An additional C18 solid-phase extraction step prevented the deterioration, although that cleanup is not yet compatible with automated preparation.
The team is now developing a robotic protocol for large-cohort glycoproteomics as part of Japan’s Human Glycome Atlas Project. “The system will be applied for plasma/serum and tissue sample preparation, and further in-depth glycoproteomics,” said Nakajima.
Turning Up the Heat for LC–IRMS
A high-temperature, high-pressure combustion interface enables accurate carbon isotope measurements of oxidation-resistant chlorinated and brominated compounds.
A high-temperature, high-pressure combustion interface has expanded liquid chromatography–isotope ratio mass spectrometry (LC–IRMS) to oxidation-resistant halogenated compounds, enabling accurate carbon isotope measurements of chlorinated and brominated pollutants.
Compound-specific carbon isotope ratios can help distinguish chemical sources and trace environmental transformations, but LC–IRMS depends on converting analytes quantitatively to carbon dioxide before isotope measurement. Conventional interfaces operate at relatively low temperatures, making compounds with strong carbon–halogen bonds difficult to oxidize without introducing isotope fractionation.
In response, the researchers built an LC–IRMS interface that combines persulfate-assisted oxidation with temperatures up to 600 °C and a pressure of 5.2 MPa. “This unique system enables stable carbon isotope analysis of a wide range of halogenated compounds, providing a new tool for source identification and fate analysis,” said Hiroto Kawashima in a press release.
At an optimized 500 °C, the method produced δ¹³C values within ±1‰ of reference measurements for chlorinated and brominated compounds, including trichloroacetic acid, which has proved difficult to analyze reliably by LC–IRMS. Fluorinated compounds remained more resistant: only monofluoroacetic acid was measured accurately, while compounds with multiple C–F bonds showed incomplete oxidation and large isotope errors.
Sensitivity also remains a limitation. Reliable measurements required roughly 50–60 nmol of carbon, leaving environmentally relevant trichloroacetic acid concentrations three to four orders of magnitude below the method’s current detection capability.
The authors now identify preconcentration and larger-volume injection as priorities for reaching lower concentrations, alongside modified oxidation chemistry to improve C–F bond cleavage. “Because the combustion interface was custom-built in-house, it has a significantly lower cost than commercial options, making it more accessible for other laboratories seeking to expand their analytical capabilities for halogenated compounds,” said Kawashima.
(Mass) Spectacular and Strange
Jewelry With Bite
At first glance, a dozen wallaby teeth from a Queensland rockshelter might not look much like jewelry. But the resin on their roots tells a different story.
The incisors were recovered from the Windmill Way rockshelter in southeast Cape York Peninsula, a site with unusually good preservation of organic material. The wider collection of animal remains was initially thought to be largely non-artefactual, possibly linked to use of the shelter as a dingo den. However, the discovery of a resin ball attached to one macropod tooth prompted researchers to take a closer look.
The study examined 12 macropod mandibular incisors using microscopy, FTIR, liquid chromatography–mass spectrometry, and accelerator mass spectrometry radiocarbon dating. Eleven teeth were identified to species, including agile wallaby, spectacled hare-wallaby, and northern nail-tail wallaby. None were perforated, and there was no convincing use-wear or residue evidence suggesting they had been used as scrapers, cutters, or drills.
Instead, the evidence pointed toward adornment. Ten teeth retained obvious resin at their roots, while two more had microscopic resin traces. Several also preserved plant fibres associated with the resin, and one contained a small fragment of cut hide or skin passing through the adhesive ball. AMS dating of two resin samples placed the objects at around 1,100–1,300 years old.
FTIR and LC-MS linked the resin to Xanthorrhoea, or grass tree. The authors argue that the resin probably helped attach plant fibre string to the teeth, making the finds rare evidence of personal ornamentation outside a burial context.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
