Scientists Find Bacterial Molecules in the Brain That Rise and Fall With Sleep, Hinting Microbes Help Control When We Rest

You share your body with trillions of microbes, and scientists are increasingly asking how those microbes and their molecules talk to your brain. One candidate messenger is peptidoglycan, a structural molecule from bacterial cell walls that can be chopped into smaller, biologically active pieces known as muramyl peptides. Decades ago, muramyl peptides were shown to promote sleep when given experimentally, but whether naturally produced peptidoglycan fluctuates inside a healthy brain, and whether those fluctuations matter for sleep, remained unknown. A careful, time-of-day aware study in mice now shows that peptidoglycan is not a static contaminant, but a dynamic signal whose levels depend on brain area, circadian time, and recent sleep history.

The problem: missing pieces in the microbe-brain conversation

Researchers have long observed two related facts, which together create a puzzle. First, muramyl peptides and other bacterial fragments can promote sleep and alter inflammatory signaling. Second, sleep and the gut microbiome interact: sleep loss alters gut microbes, and infection alters sleep. What has been missing is a description of endogenous peptidoglycan levels inside different brain regions over normal daily cycles, and how those endogenous levels respond to sleep loss. Without those measurements, models that place microbial products at the center of sleep regulation remain incomplete. The new study addresses that measurement gap directly, with quantitative assays of peptidoglycan across brain areas and time points, and with gene expression profiling to reveal how the brain responds to sleep loss at the molecular level.

The approach: careful timing, region sampling, and molecular readouts

To capture daily rhythms and sleep-loss effects, the researchers used adult male C57BL/6J mice kept on a 12-hour light, 12-hour dark cycle and monitored sleep noninvasively with piezoelectric sensor cages to avoid surgical implants that might induce inflammation. They sampled brain tissues at five time points chosen to map rest-activity transitions and steady states: zeitgeber time ZT0, ZT3, ZT6, ZT12, and ZT15. The brain regions selected were brainstem, somatosensory cortex, hypothalamus, and olfactory bulb, chosen because of their roles in sleep regulation or direct contact with environmental microbes. For sleep-loss experiments, separate groups underwent gentle-handling sleep deprivation for either 3 hours (ZT0–3, mild) or 6 hours (ZT0–6, moderate), followed by tissue collection at the corresponding ZT.

Peptidoglycan levels in homogenized brain tissue and in serum were quantified using a commercial mouse peptidoglycan ELISA, with all values normalized to tissue wet weight and reported as ng PG per mg tissue, or ng/ml for serum. For gene expression, somatosensory cortex samples from control and sleep-deprived mice were submitted to RNA sequencing. Libraries were prepared and sequenced on Illumina platforms, producing roughly 40 million reads per sample, and differential expression was assessed with established pipelines (HISAT2 alignment, HTSeq counts, DESeq2 statistical testing). The RNA-seq comparison captured cortical transcriptional responses after an 8-hour sleep loss window in a separate set of animals.

The breakthrough discovery

Peptidoglycan levels in the brain follow time-of-day rhythms

Across all measured brain areas, endogenous peptidoglycan showed significant time-of-day variation. The lowest peptidoglycan concentrations occurred at ZT12, the transition from the animals’ main rest period to their active period. Levels rose at ZT15 and at ZT6 relative to ZT12, a pattern that, for many brain areas, mirrored the mice’s hourly sleep pattern. This establishes that peptidoglycan levels in the healthy brain are rhythmic, in step with the animal’s rest-activity cycle.

Brain-region specificity: the brainstem leads the field

Not all brain regions contained the same amounts. The brainstem consistently had the highest peptidoglycan levels compared with olfactory bulb, hypothalamus, and somatosensory cortex, which had lower but broadly similar levels. Importantly, the pattern of daily change was not identical across regions: hypothalamic peptidoglycan showed a distinct pattern, decreasing from ZT0 to ZT12 and then rising from ZT12 to ZT15, whereas other regions tracked sleep amount more closely. These results indicate that peptidoglycan entry, clearance, or local handling differs across brain areas.

Sleep loss shifts peptidoglycan in region- and duration-dependent ways

Short sleep loss produced region-specific and time-dependent changes in brain peptidoglycan. After a 3-hour sleep disruption (ZT0–3), peptidoglycan decreased in brainstem and hypothalamus but increased in somatosensory cortex. Serum peptidoglycan trended lower after 3 hours of sleep loss. By contrast, after 6 hours of sleep loss (ZT0–6) the direction of changes had shifted: brainstem and olfactory bulb showed significant increases in peptidoglycan relative to controls, while serum levels were not significantly altered. These patterns show that the brain’s peptidoglycan landscape is acutely responsive to sleep history, and that the response depends on both brain region and the length of sleep loss.

Sleep loss alters cortical genes that detect or respond to peptidoglycan

RNA sequencing of somatosensory cortex after an 8-hour sleep loss window revealed widespread transcriptional changes. Of the 5,501 genes that changed significantly, roughly half were upregulated and half were downregulated. The researchers specifically examined genes known to be involved in bacterial fragment detection and signaling. Notable results included strong upregulation of the peptidoglycan recognition protein Pglyrp1 and increased expression of immune and signaling-related genes such as Nfil3, Myd88, Il18, and certain MHC component genes. In contrast, some bacterial cell wall binding or response genes including Siglec1, Trem2, and Lbp were downregulated. The pattern suggests that sleep loss both changes peptidoglycan amounts and alters the brain’s molecular machinery for detecting and responding to microbial fragments.

Why these results matter

They strengthen the microbial contribution to sleep physiology

The idea that microbial products can influence sleep is not new, but this study provides quantitative evidence that peptidoglycan, an abundant microbial cell wall component, is dynamically present in the healthy brain in synchrony with daily rest-activity cycles and responsive to sleep loss. That temporal and spatial structure is exactly what one would expect from a physiologically relevant signaling molecule rather than a noise-level contaminant. The findings therefore lend weight to models in which microbe-derived molecules contribute to the homeostatic regulation of sleep.

They connect circadian microbial activity and brain function

Microbial cell wall synthesis in the gut and microbial community composition are themselves rhythmic, linked to feeding and circadian host behavior. If peptidoglycan release into circulation is rhythmic, and if brain entry or clearance is modulated by daily changes in brain vasculature or glymphatic clearance, then peptidoglycan may be a molecular bridge that conveys time-of-day information from the microbiome to brain networks that gate sleep and arousal. This bridges two active fields—circadian microbiome biology and sleep neurobiology—and proposes a concrete molecular link.

They suggest candidate mechanisms and targets

The strong, sleep-loss associated upregulation of Pglyrp1 and changes in other immune signaling genes provide mechanistic clues. Pglyrp1 is known to induce cytokines such as IL1 and TNF, which are established sleep-regulatory substances. Thus, peptidoglycan detection in brain tissue could act through Pglyrp1 and downstream cytokines to promote localized sleep-related processes. If validated, that pathway could become a target for interventions designed to correct sleep problems driven by dysbiosis or to modulate inflammation-linked sleep disturbance.

Caveats and what still needs to be done

The scientists explicitly note limitations that temper the conclusions and guide next steps. First, many of the reported effects are subtle and sample sizes were modest. For the RNA-seq experiment the group size was n = 3 per condition at the single collection timepoint, a design chosen to capture a focused physiological response but one that lacks temporal depth and power to find small changes. Second, correlation does not prove causation: changes in brain peptidoglycan may reflect altered entry, local production or clearance, or redistribution rather than direct signaling events. Third, how peptidoglycan travels from gut to brain, and whether the same patterns occur in females or in other species including humans, remain open questions. The scientists call for time-resolved gene expression studies, manipulations of peptidoglycan or its receptors, and expanded sampling across sexes and ages to build a causal chain.

Practical takeaways and future horizons

For clinicians and the public, the study is not yet a call to action, but it does reshape how we should think about sleep. Sleep biology may be a property of the holobiont, the host plus its microbes, rather than of the host alone. This reframing opens possibilities: therapies that alter the microbiome or block specific peptidoglycan sensing pathways could one day complement behavioral sleep treatments, especially where inflammation and altered gut permeability are involved. For researchers, the paper highlights new experimental directions: manipulating Pglyrp1 in specific brain regions, tracing labeled peptidoglycan from gut to brain, and mapping peptidoglycan rhythms in disease models.

Conclusion: small fragments, large implications

The study demonstrates that bacterial peptidoglycan is a dynamic component of the healthy mouse brain, with levels that depend on the region sampled, the time of day, and recent sleep history. The observation that sleep loss changes both peptidoglycan levels and the expression of peptidoglycan-linked genes in cortex supports a model in which microbial fragments participate in the network of signals that regulate sleep. While the data are descriptive and modest in scope, they provide necessary quantitative grounding for the hypothesis that microbe-host interactions are integral to sleep biology. Follow-up experiments that increase temporal resolution, test causality, and extend findings to other sexes and species will be needed before these molecular rhythms can be translated into human sleep medicine. For now, the paper invites a fundamental shift in perspective: sleep may be, in part, a property of our shared biology with microbes.

Key methods and numbers at a glance

• Species and strain: Male C57BL/6J mice, 2–5 months old.
• Brain areas measured: brainstem, somatosensory cortex, hypothalamus, olfactory bulb.
• Time points: ZT0, ZT3, ZT6, ZT12, ZT15.
• Sleep deprivation: gentle handling for 3 h (ZT0–3, n = 8) or 6 h (ZT0–6, n = 9).
• PG quantitation: commercial mouse PG ELISA, normalized to tissue wet weight (ng/mg).
• RNA-seq: somatosensory cortex, n = 3 control, n = 3 SD, ~40 million reads/sample, differential expression by DESeq2.

The study was published in Frontiers in Neuroscience on July 16, 2025.