Prolonged bacterial lag time results in small colony variants that represent a sub-population of persisters

Abstract

Persisters are a subpopulation of bacteria that are not killed by antibiotics even though they lack genetic resistance. Here we provide evidence that persisters can manifest as small colony variants (SCVs) in clinical infections. We analyze growth kinetics of Staphylococcus aureus sampled from in vivo conditions and in vitro stress conditions that mimic growth in host compartments. We report that SCVs arise as a result of a long lag time, and that this phenotype emerges de novo during the growth phase in various stress conditions including abscesses and acidic media. We further observe that long lag time correlates with antibiotic usage. These observations suggest that treatment strategies should be carefully tailored to address bacterial persisters in clinics.

Fig. 1

S. aureus recovered from human and murine abscesses displayed heterogeneous colony sizes including a fraction of small colony variants. Distribution of colony sizes 24 h after plating. Bacteria that were sampled directly from a patient abscess (a) or mice abscesses (b, see also SI 4), from eukaryotic A549 cells (c) and from liquid cultures grown at different pHs (d). Bacteria grown exponentially in complex LB medium served as control (dashed black lines). Shaded areas depict the standard deviation. The colored areas under each curve mark colonies whose area was at least five times smaller than the area of the most common colony type, the criterion that is generally used to identify SCVs (the figures depict colony radius instead of colony area, and SCVs would thus be defined as colonies whose radius is at least a factor 2.23, i.e., √5, smaller than the radius of the most common colony type. The corresponding frequencies are given above the distributions (labeled as “%nsSCV”, “bdl”: below detection limit). Insets show representative images of colonies 24 h after plating. Small colonies are indicated by an arrow. All histograms use a binning of 100 µm. Groups in d define homogeneous subsets of SCV fractions; treatments with a different group letter. a–d Show significant differences in the proportion of SCVs (at P < 0.05, t-test, Bonferroni correction for multiple testing). ** represents a statistical significance of a two-sided t-test at 0.01 threshold; shades represent s.d.

Fig. 2

SCV formation is a consequence of a late emergence of colonies, and is caused by a delay in the first division at the level of single cells. a Schematic of three different scenarios that can explain why a colony is small when measured at a specific time point (vertical black line). b Colony growth of bacteria sampled from murine abscesses and compared to bacteria from an exponentially growing culture (n = 50 for each condition). Colonies sampled from the murine abscess tended to emerge later than colonies sampled from the exponentially growing culture. c Colony growth of bacteria sampled from a liquid culture at neutral (pH 7.4) or acidic (pH 5.5) medium after 3 days of incubation. The same number of colonies is represented for each condition (n = 50). Colonies sampled from pH 5.5 tended to emerge later than colonies sampled from pH 7.4. d Representative images of colonies acquired at different time points after sampling from a liquid acidic culture medium (pH 4.0). Scale bar, 2 mm. Colonies appeared at different time points (arrows). e Colony growth of the stable SCV strain Cowan hemB::ermB on LB agar or on LB agar supplemented with sheep blood (n = 50 for each condition). The colonies of this stable SCV strain grew at a markedly lower rate on LB plates as compared to blood plates that complement their auxotrophy. f Cumulative distributions of the time points of the first cell division of bacteria sampled from a murine abscess model (n = 5 mice) and from an exponentially growing culture determined by automated time-lapse microscopy. The vertical gray area marks the period at the beginning of the experiment where cell divisions could occur but not be observed (** for P < 0.01). g Cumulative distributions of the time points of the first cell division of bacteria sampled from liquid cultures with different pHs and from an exponentially growing culture at neutral pH (n > 3). Groups mark homogeneous subsets of the proportion of bacteria with a lag time (time to first division) over 6 h; treatments with a different group letter are significantly different (at P < 0.05, t-test, Bonferroni correction for multiple testing)

Fig. 3

SCVs are formed during the growth phase and their proportion increases during stationary phase. a When inoculating a low number of bacteria (2 × 10⁴ ml⁻¹) in acidic media (pH 5.5), the resulting populations continue to grow for about 100 h. Lines represent changes in optical density over time in two independent replicates. b The lines depict the number of SCVs (i.e, colonies with a lag time of more than 6 h; dashed lines, stars) and the number of colonies with larger radius (solid lines, open circles) in these cultures. Seventy-two hours after inoculation, the absolute number of SCVs exceeds the total number of bacteria in the initial inoculum, and SCVs were thus formed during the growth phase. Horizontal black lines are guides to the eye. c When bacteria were inoculated at higher concentration, stationary phase was reached earlier and at a similar optical density as in the experiments depicted in a. We used these growth experiments for analyzing effects of stationary phase on the frequency of SCVs. d After entering stationary phase, the absolute number of SCVs does not vary substantially over time (stars), while the absolute number of colonies with larger radius decreases (empty circles). As a consequence, the proportion of SCVs increases in the population without increase in absolute numbers. Lines are exponential regression for 2 replicates on the indicated time range (dashed: SCV colonies, solid: non-SCV colonies; the estimated rates (for large colonies, L, and small colonies, S) with standard deviations of the slopes are depicted in the panel)

Fig. 4

The proportion of bacteria in lag phase correlates with the proportion of bacteria surviving antibiotics. a Illustration of the persister assay. b, c Depict the fraction of bacteria that were killed as a function of time upon exposure to the antibiotics flucloxacillin (b) and ciprofloxacin (c) at 10-fold the minimum inhibitory concentrations. Bacteria were sampled (n = 3) from cultures grown in neutral (pH 7.4) or acidic (pH 5.5) medium and then exposed to fresh growth medium containing antibiotics. Each data point was calculated as the ratio between the number of colony forming units (CFU) at a specific time point after antibiotic addition and the number of CFU in the inoculum. Continuous lines represent the one-phase association fit with extra sum of squares F-test to compare K values, plateau and Y₀ are different. d Bacteria recovered from mouse abscesses survived 24 h antibiotic exposure (40-fold MIC) better than bacteria sampled from a stationary-phase pH 7.4 culture. n = 3; **** correspond to threshold of P < 0.05 and 0.001, respectively for a two-sided student's test. Error bars depict s.e.m. e The fraction of bacteria in a sample that had a lag time of more than 6 h upon plating correlated with the fraction of bacteria in the same sample that survived a 6 h exposure to antibiotics. Cultures were sampled at day 3 or 5 to obtain different proportion of bacteria with a lag time of more than 6 h. Black line corresponds to linear fit (P < 0.001, rho = 1, Spearman’s Rank-Order Correlation). f Exposure to antibiotics increased the lag time of individual bacterial cells, as measured by time-lapse microscopy (n = 3; n = 2 for pH 7.4 at 10-fold MIC). This increase was small for bacteria sampled from cultures with neutral pH (gray curves), and more pronounced for bacteria sampled from acidic cultures (green curves). Curves show averages of 3 replicates, and shaded areas depict s.d.