Protein Aggregation is Associated with Acinetobacter baumannii Desiccation Tolerance

Abstract

Desiccation tolerance has been implicated as an important characteristic that potentiates the spread of the bacterial pathogen Acinetobacter baumannii on dry surfaces. Here we explore several factors influencing desiccation survival of A. baumannii. At the macroscale level, we find that desiccation tolerance is influenced by cell density and growth phase. A transcriptome analysis indicates that desiccation represents a unique state for A. baumannii compared to commonly studied growth phases and strongly influences pathways responsible for proteostasis. Remarkably, we find that an increase in total cellular protein aggregates, which is often considered deleterious, correlates positively with the ability of A. baumannii to survive desiccation. We show that inducing protein aggregate formation prior to desiccation increases survival and, importantly, that proteins incorporated into cellular aggregates can retain activity. Our results suggest that protein aggregates may promote desiccation tolerance in A. baumannii through preserving and protecting proteins from damage during desiccation until rehydration occurs.

Keywords: Acinetobacter baumannii; desiccation tolerance; protein aggregation.

Figure 1

Growth condition and cell density influence desiccation survival of A. baumannii strain 17978. (A) Percent survival of A. baumannii strain 17978 desiccated over 15 days. The strain grown on an LB agar plate was harvested and adjusted to a cell density of 10⁷, 10⁶ or 10⁵ CFU in 10 μL of water to desiccate on polystyrene at 25 °C and 40% relative humidity before rehydration. Cells before desiccation and after 1, 2, 3, 5, 10, or 15 days of desiccation were serially diluted and spotted for CFU to determine the percent survivals. (B) Percent survivals of 17978 from desiccation when the strain, prior to desiccation, was grown on LB agar plates or in liquid LB to stationary phase or log phase. Bacteria from all three growth phases were adjusted to a cell density of 10⁷ or 10⁶ in 10 μL of water to desiccate for 48 h before rehydration. Statistical analyses were performed on the log10 transformation of the data. Percent survival comparisons were performed by two-way ANOVA with Bonferroni post-hoc test, and the p-values were presented as the stars in the graphs. **** p < 0.0001. ND marks when the percent survival was not detectable. Mean survival with error bars (SEM) was obtained from 3 replicates for (A) and 4 replicates for (B).

Figure 2

Transcriptomic analysis of desiccated and growing bacteria. (A) PCA analysis of the 17978 transcriptomes after desiccation, or after growing in liquid LB to stationary phase or log phase or on an LB agar plate. K-means clustering at n = 4 with Pearson distance was performed on the 8 samples. Samples from each condition clustered together as shown by colors of the dots. Desiccated samples and growth samples were all separated. (B) The number of genes significantly changed in expression level after desiccation compared to each growth phase at a cutoff of p < 0.01. In each comparison, the purple column represents the number of genes down-regulated in desiccation, whereas the yellow one represents the up-regulated genes. (C) The overlaps of upregulated genes in desiccation among the three comparisons. (D) The overlaps of down-regulated genes in desiccation among the three comparisons.

Figure 3

Protein aggregation levels in desiccated and growing bacteria. (A) Protein aggregates isolated from 17978 grown on LB agar plates or in liquid LB to stationary phase or log phase, or desiccated. The desiccated samples were grown either on LB agar plates or in liquid LB to log phase prior to desiccation. Aggregates were purified and normalized by total cellular protein concentration [46]. Samples in each panel were processed on the same gel. (B) Quantification of (A) normalized to the log phase growth sample. Mean area density with error bars (SEM) was obtained from 3 biological replicates. Area density comparisons were performed with Repeated Measures ANOVA followed by Tukey post hoc test. * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 4

Inducing protein aggregate production is associated with increased desiccation tolerance. (A) Protein aggregates isolated from 17978 growing in liquid LB to log phase with no treatment (Log phase), treated with 10 µg/mL spectinomycin (Spc10), or with 4 µg/mL streptomcin (Str4). (B) Quantification of (A) normalized to Log phase samples. (C) Desiccation percent survival of 17978. Bacteria were initially grown in liquid LB to log phase without treatment, treated with 10 µg/mL spectinomycin, or with 4 µg/mL streptomycin prior to desiccation. (D) Protein aggregates isolated from the wild type 17978 carrying an empty vector pABBR (WT::pABBR), a lon deletion mutant with pABBR (Δlon::pABBR), and a lon deletion mutant with lon expressed from the vector pABBR (Δlon::pABBR_lon). (E) Quantification of (D) normalized to wild type 17978. (F) Desiccation percent survivals of the three strains in (D). In B, C, E, F, means with error bars (SEM) were obtained from 3 biological replicates. Comparisons were performed by one-way ANOVA followed by Tukey post-hoc test. * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 5

LacZ function detected from A. baumannii protein aggregates. (A) Western blot for the presence of LacZ. All the samples were prepared from 17978 grown on agar plates. LacZ was detected in the whole cell lysates of 17978 constitutively expressing lacZ from the plasmid pXW1 (+lacZ). LacZ was also detectable from protein aggregates in the +lacZ strain, collected from agar plate growth or after desiccation. LacZ was not detected in whole cell lysates or protein aggregates from the A. baumannii parental strain that did not encode lacZ. Equal numbers of cells were processed for whole cell lysate samples. Aggregate loading was normalized by total cellular protein concentration. The whole cell lysate samples were processed on the same gel, while the aggregate samples were processed together on a separate gel. (B) Protein aggregates were assayed for LacZ activity measuring ONPG cleavage at OD₄₁₀. The signal from buffer was used as baseline and subtracted from all samples. LacZ activity from protein aggregates is reported relative to activity from aggregates isolated from desiccated wild type 17978 (WT, desiccated), which was set at 1. One-way ANOVA with Tukey post-hoc test showed that LacZ activity was higher (** p < 0.01) from protein aggregates extracted from A. baumannii 17978 ectopically expressing lacZ.