Characterization of Acinetobacter baumannii Copper Resistance Reveals a Role in Virulence

Abstract

Acinetobacter baumannii is often highly drug-resistant and causes severe infections in compromised patients. These infections can be life threatening due to limited treatment options. Copper is inherently antimicrobial and increasing evidence indicates that copper containing formulations may serve as non-traditional therapeutics against multidrug-resistant bacteria. We previously reported that A. baumannii is sensitive to high concentrations of copper. To understand A. baumannii copper resistance at the molecular level, herein we identified putative copper resistance components and characterized 21 strains bearing mutations in these genes. Eight of the strains displayed a copper sensitive phenotype (pcoA, pcoB, copB, copA/cueO, copR/cusR, copS/cusS, copC, copD); the putative functions of these proteins include copper transport, oxidation, sequestration, and regulation. Importantly, many of these mutant strains still showed increased sensitivity to copper while in a biofilm. Inductively coupled plasma mass spectrometry revealed that many of these strains had defects in copper mobilization, as the mutant strains accumulated more intracellular copper than the wild-type strain. Given the crucial antimicrobial role of copper-mediated killing employed by the immune system, virulence of these mutant strains was investigated in Galleria mellonella; many of the mutant strains were attenuated. Finally, the cusR and copD strains were also investigated in the murine pneumonia model; both were found to be important for full virulence. Thus, copper possesses antimicrobial activity against multidrug-resistant A. baumannii, and copper sensitivity is further increased when copper homeostasis mechanisms are interrupted. Importantly, these proteins are crucial for full virulence of A. baumannii and may represent novel drug targets.

Keywords: Acinetobacter baumannii; Galleria; copper; metal; pathogenesis.

FIGURE 1

Organization of regions A–D in the AB5075 chromosome. We previously identified 22 putative copper-related genes in the genome of AB5075 (Williams et al., 2016). These genes are located in four chromosomal regions named A–D. The relative location of each region on the AB5075 chromosome is shown on the left (not to scale), and the genes and regions are depicted to scale relative to one another. Each operon is colored uniquely, and the same colors are maintained in Figure 9. Gene annotations are shown when available, otherwise ORF numbers are used.

FIGURE 2

Growth of highly copper sensitive AB5075 mutant and complementation strains. Strains were grown for 6 h at 37°C in M9 medium supplemented with the concentration of CuSO₄ indicated in the legend. Growth was measured both by turbidity (OD₆₀₀, data not shown) and enumeration of viable colonies (CFU/mL). Black lines are used to group the mutant strain with the corresponding complemented derivative. The data are presented as geometric mean and SEM of three biologically independent experiments. Two-way ANOVA with Tukey’s adjustment for multiple comparisons was used to compare growth in each copper concentration at each timepoint. The symbols indicate which copper treatments were statistically different (P < 0.05) from the 0 mM control, as follows: ∼, growth in 1.5 mM CuSO₄ was different from the control; +, growth in 1 and 1.5 mM CuSO₄ was different from the control; *, growth in 0.5, 1, and 1.5 mM CuSO₄ was different from the control; and #, growth in 0.25, 0.5, 1, and 1.5 mM CuSO₄ was different from the control.

FIGURE 3

Growth of moderately copper sensitive AB5075 mutant and complementation strains. Strains were grown for 6 h at 37°C in M9 medium supplemented with the concentration of CuSO₄ indicated in the legend. Growth was measured both by turbidity (OD₆₀₀, data not shown) and enumeration of viable colonies (CFU/mL). Black lines are used to group the mutant strain with the corresponding complemented derivative. The data are presented as geometric mean and SEM of three biologically independent experiments. Two-way ANOVA with Tukey’s adjustment for multiple comparisons was used to compare growth in each copper concentration at each timepoint. The symbols indicate which copper treatments were statistically different (P < 0.05) from the 0 mM control, as follows: ∼, growth in 1.5 mM CuSO₄ was different from the control; +, growth in 1 and 1.5 mM CuSO₄ was different from the control.

FIGURE 4

Growth of AB4857 and AB5711 carrying Region D. Region D from AB5075 was cloned into pAJ100 and transformed into AB4857 and AB5711, which do not naturally encode the region D genes. Strains carrying the empty vector, pAJ100, were used as controls. The strain background is indicated in the graph title (AB4857, left; AB5711, right); strains carrying the control vector pAJ100 are indicated with dotted lines, and strains carrying pAJ100-Region D are indicated with solid lines. Strains were grown for 6 h at 37°C in M9 medium supplemented with the indicated concentration of CuSO₄. Growth was measured both by turbidity (OD₆₀₀, data not shown) and enumeration of viable colonies (CFU/mL). The data are presented as geometric mean and SEM of three biologically independent experiments. Two-way ANOVA with Tukey’s adjustment for multiple comparisons was used to compare growth of the strain carrying the pAJ100 vector to the strain carrying pAJ100-Region D in each copper concentration at each timepoint. The symbols indicate how many comparisons were statistically significant (P < 0.05) at each timepoint, as follows: ∼, growth in 1.5 mM CuSO₄ was different between the two strains; +, growth in 1 and 1.5 mM CuSO₄ was different between the two strains *, growth in 0.5, 1, and 1.5 mM CuSO₄ was different between the two strains.

FIGURE 5

Copper-dependent killing of AB5075 mutant strain biofilms and planktonic cells. Cultures were statically incubated in a 24-well tissue culture-treated plate at 37°C to establish biofilms. After 24 h, the broth was replaced with M9 medium containing 1.5 mM CuSO₄. After an additional 0, 6, or 24 h of incubation, bacteria from both the supernatant (planktonic) and the biofilm were plated for enumeration of CFU. Percent survival was calculated as the percentage of bacteria present at time zero under each condition. Data are presented as mean and range for three biologically independent experiments. One-way ANOVA with Dunnett’s correction was used to compare each mutant strain to the wild-type at each indicated condition. Statistically significant comparisons are indicated with the asterisks, as follows: **P < 0.01; ***P < 0.001; ****P < 0.0001.

FIGURE 6

Accumulation of intracellular copper. Log phase cultures were treated with 0.25 mM CuSO₄ and samples were collected at the indicated timepoints. During the wash phase, the remaining culture volume was washed and resuspended in fresh M9 medium without supplemented copper. Cell samples that were collected at each timepoint were washed and dried, and copper content was measured by ICP-MS. Data are presented as the mean and range for three biologically independent experiments. Two-way ANOVA with Sidak’s correction was used to compare the cusR mutant strain to the wild-type at all timepoints (top); two-way ANOVA with Dunnett’s correction was used to compare all the mutant strains to wild-type at all timepoints (bottom). Statistically significant comparisons are indicated with an asterisk (P < 0.01). At 0 min post-wash, five strains were significantly different from wild-type: copB, cueO, cusR, cusS, and copD (bottom).

FIGURE 7

Survival of Galleria mellonella infected with AB5075 mutant strains. G. mellonella were injected with approximately 5.0 × 10⁴ CFU of the indicated strain and survival was monitored for 5 days. The gene interrupted in each mutant strain is indicated in the graph title. Experiments were repeated using two or three different orders of larvae with 12–15 larvae per experimental group; the total n ranged from 25 to 45 larvae per strain. Kaplan–Meier survival curves were compared (excluding Untouched and PBS controls) using the Mantel–Cox test with Holm’s correction for multiple comparisons. Statistically significant comparisons of the mutant strain to the wild-type strain are indicated by bars with asterisks signifying the P-value as follows: *P < 0.05; **P < 0.01; ***P < 0.001.

FIGURE 8

Assessment of virulence of AB5075 mutant and complemented strains in a murine pneumonia model. Mice were infected intranasally with approximately 5 × 10⁶ CFU of the indicated strain and survival was monitored for 5 days. Experiments with the cusR mutant strain and complemented derivative were performed in four biological replicates with five mice per group (n = 20 per strain), spaced two and two over time to ensure reproducibility; two biological replicates were completed with the copD mutant strain and complemented derivative (n = 10 mice per strain). Kaplan–Meier survival curves were compared using the Mantel–Cox test with Holm’s correction for multiple comparisons. Statistically significant comparisons are indicated by bars with asterisks signifying the P-value as follows: *P < 0.05; **P < 0.01; ***P < 0.001. The only significant comparison in the top panel is wild-type AB5075 vs. the cusR mutant strain.

FIGURE 9

Model of putative copper resistance mechanisms in AB5075. Predicted protein identifications were determined by amino acid sequence homology to known proteins from other bacteria, and proteins with putative copper resistance functions are indicated. In the mutational analysis described herein, proteins putatively involved in transport (PcoB, CopB, CopD), oxidation (PcoA, CueO), chaperoning (CopC), and regulation (CusR, CusS) individually contributed to copper resistance in AB5075. Copper ATPases are highly conserved inner-membrane proteins that efflux copper ions out of the cytoplasm; AB5075 appears to have two copper ATPases: CopA1 and CopA2. The RND family efflux pump CzcCBA may efflux copper from the periplasm to the extracellular milieu. Three other putative membrane proteins in AB5075 may also have roles in copper transport, though the function of these proteins is not yet well-defined in any bacteria (CopB, PcoB, and CopD). Periplasmic copper oxidases reduce damage by converting Cu⁺ to the less toxic Cu²⁺; CueO and PcoA are both predicted copper oxidases. CopC is a highly conserved copper chaperone that is found in the periplasm; it prevents damage by binding and sequestering copper ions and may facilitate movement of copper ions between other copper-related proteins, i.e., delivering copper ions to an efflux pump. Two well-conserved putative regulatory systems are also present in AB5075: CueR senses cytoplasmic copper levels and then upregulates copper resistance genes in its copper-bound form, and CusRS is a two-component system that is thought to be activated by periplasmic copper. The regulons of CueR and CusR have not yet been defined in AB5075, though putative binding sites have been identified in regions B and D, respectively, and these putative target genes are indicated in the figure.