Diurnal oscillations in gut bacterial load and composition eclipse seasonal and lifetime dynamics in wild meerkats

Abstract

Circadian rhythms in gut microbiota composition are crucial for metabolic function, yet the extent to which they govern microbial dynamics compared to seasonal and lifetime processes remains unknown. Here, we investigate gut bacterial dynamics in wild meerkats (Suricata suricatta) over a 20-year period to compare diurnal, seasonal, and lifetime processes in concert, applying ratios of absolute abundance. We found that diurnal oscillations in bacterial load and composition eclipsed seasonal and lifetime dynamics. Diurnal oscillations were characterised by a peak in Clostridium abundance at dawn, were associated with temperature-constrained foraging schedules, and did not decay with age. Some genera exhibited seasonal fluctuations, whilst others developed with age, although we found little support for microbial senescence in very old meerkats. Strong microbial circadian rhythms in this species may reflect the extreme daily temperature fluctuations typical of arid-zone climates. Our findings demonstrate that accounting for circadian rhythms is essential for future gut microbiome research.

Fig. 1. Study system and sampling distribution.

a Timeline of samples analysed in this study (1997–2019; n = 1109), with lines connecting samples collected from the same meerkat individual (y-axis), and coloured by hours after sunrise. Yellow represents samples collected close to sunrise, purple represents samples collected closer to sunset. Periods of intensive sampling (~2007 and 2015) enable us to account for environmental and social effects at certain periods. b Proportion of time meerkats spend foraging during wet summer (green dashed line) and dry winter (yellow dashed line). Figure modified from Doolan and MacDonald with permission. Solid lines represent sunrise and sunset in summer (green) and winter (yellow). c Seasonal climate across the year measured at the Kalahari Research Station, South Africa, averaged from data between 2009 and 2019. Bars represent total rainfall per month, and red and blue lines represent mean maximum and minimum temperatures, respectively. d Trend in residual body mass across life showing senescence at approximately 5.5 years and 95% credible intervals, modified from Thorley et al.. e Sampling distribution for diurnal, seasonal, and lifetime scales. Source data are provided in the source data file.

Fig. 2. Temporal trends in gut bacterial load (top), alpha diversity (middle), and beta diversity (bottom) across the day (left), year (centre), and life (right).

a–f Smoothed estimates and partial residuals from two GAMMs predicting (a–c) bacterial load and (d–f) observed ASV richness across the day, year, and life. Shaded area represents 95% CIs. g–i Beta diversity MDS ordination coloured by (g) hours after sunrise (orange = morning; afternoon = purple), (h) season (green = wet season; yellow = dry season), and (i) age (blue = young; grey = adult; red = old). For clarity, samples were grouped into discrete categories to generate 95% CI ellipses and group centroids (large circles), which represent samples collected in the morning field session (< 7 h after sunrise) and afternoon (> 7 h after sunrise), wet (October–April) and dry (May–September) season, and age categories (< 1 years, 1–5 years, and >5 years). Source data are provided in the source data file.

Fig. 3. Genera driving temporal trends in beta diversity.

a Summary of taxonomic shifts in relative abundance at the genus-level per 30-min interval from sunrise. The number of samples that were summarised per 30-min slot are indicated by the histogram. b, c Weighted Unifrac ordination plots of (b) axes one and two and (c) three and four, coloured and grouped by the most abundant genus in each sample. Large circles represent group centroids for samples sharing the same most abundant genus. Arrows indicate the direction and influence of significant temporal variables when categorised into morning/afternoon, wet/dry seasons, and young/adult/old. Statistics for temporal variables (arrows) are shown. Source data are provided in the source data file.

Fig. 4. Temporal dynamics of 16 focus genera across scales.

Comparison of temporal dynamics of 16 focus genera across diurnal (a, b), seasonal (c, d), and lifetime (e, f) scales. Top panel shows GAMM abundance estimates across the three temporal scales compared to the mean, where zero (indicated by the dashed line) represents the mean (log10) abundance of each genus. Estimates have been back-transformed in the bottom panel to represent absolute abundance, and bacterial load (grey) is shown for comparison. Only genera that significantly shift across both frozen and freeze-dried samples are shown (see Supplementary Figs. 6–8 for 95% confidence intervals and trends split by storage). Source data are provided in the source data file.

Fig. 5. Summary of temporal (blue) and mechanistic (red) effects on the microbiome (n = 1109).

a Effect sizes for non-linear temporal variables on bacterial load, diversity measures, and abundances of 16 focus genera. b Effect sizes and 95% confidence intervals for eight environmental, biological, and behavioural variables. Dark blue/red denotes significant associations that are consistent across storage treatments, whilst light blue/red denotes significant yet inconsistent across storage. c Partitioning of model R² into variance explained by temporal, mechanistic, and methodological variables. Source data are provided in the source data file.

Fig. 6. Diurnal oscillations do not decay with age.

a–c Diurnal oscillations of 16 focus genera in a) young meerkats (<1 year old; n = 385). b adult meerkats (1–5 years old; n = 627); and (c) particularly old meerkats (>5 years old; n = 97). Zero represents the taxa mean and sample distributions are indicated by the histograms. d–h Model estimates and 95% confidence intervals for (d) Clostridium (sensu stricto 1); (e) Bacteroides; (f) Paeniclostridium, (g) Cellulomonas, and (h) Raoultibacter, split by age category (blue = young; grey = adult; red = old). Source data are provided in the source data file.