Injections of Predatory Bacteria Work Alongside Host Immune Cells to Treat Shigella Infection in Zebrafish Larvae

Abstract

Bdellovibrio bacteriovorus are predatory bacteria that invade and kill a range of Gram-negative bacterial pathogens in natural environments and in vitro [1, 2]. In this study, we investigated Bdellovibrio as an injected, antibacterial treatment in vivo, using zebrafish (Danio rerio) larvae infected with an antibiotic-resistant strain of the human pathogen Shigella flexneri. When injected alone, Bdellovibrio can persist for more than 24 hr in vivo yet exert no pathogenic effects on zebrafish larvae. Bdellovibrio injection of zebrafish containing a lethal dose of Shigella promotes pathogen killing, leading to increased zebrafish survival. Live-cell imaging of infected zebrafish reveals that Shigella undergo rounding induced by the invasive predation from Bdellovibrio in vivo. Furthermore, Shigella-dependent replication of Bdellovibrio was captured inside the zebrafish larvae, indicating active predation in vivo. Bdellovibrio can be engulfed and ultimately eliminated by host neutrophils and macrophages, yet have a sufficient dwell time to prey on pathogens. Experiments in immune-compromised zebrafish reveal that maximal therapeutic benefits of Bdellovibrio result from the synergy of both bacterial predation and host immunity, but that in vivo predation contributes significantly to the survival outcome. Our results demonstrate that successful antibacterial therapy can be achieved via the host immune system working together with bacterial predation by Bdellovibrio. Such cooperation may be important to consider in the fight against antibiotic-resistant infections in vivo.

Keywords: Bdellovibrio; Shigella; antibacterial; antibiotic; innate immunity; predation; zebrafish.

Figure 1

Injected Predatory Bdellovibrio Persist in Zebrafish Larvae without Ill Effects and Protect against Shigella Infection In Vivo (A) Cartoon of Bdellovibrio life cycle. (I–III) Motile predatory Bdellovibrio attach to and invade the periplasm of Gram-negative bacteria such as Shigella. (III) Prey bacteria are rounded by DD-endopeptidase action on the cell wall. (IV) Prey bacteria are killed in ∼30 min and kept intact as Bdellovibrio consume their contents and grow. (V and VI) Following replication, Bdellovibrio lyse prey 180–240 min after invasion, releasing further predators. These Bdellovibrio progeny can repeat the predatory cycle. OM, outer membrane; CW, cell wall peptidoglycan; CM, cytoplasmic membrane. (B) Wild-type (WT) AB larvae were injected at 3 dpf in the hindbrain ventricle with 1–10 × 10⁴ PFUs of mCherry-Bdellovibrio (red). The same larvae were imaged over time to observe distribution. Representative images from a single larva are shown here. Scale bar, 100 μm. (C) Enumeration of live Bdellovibrio in PBS-homogenates from larvae injected with mCherry-Bdellovibrio as in (B) over time. Each circle represents a count from an individual larva. Data are pooled from two independent experiments (n = 8 larvae per experiment). Mean ± SEM (horizontal bars) is shown. The p values (versus the 0 hpi time point) were determined by multiple t test. Significance with Bonferroni correction was defined as p < 0.0125. See also Figures S1B–S1D for comparative evaluations of Bdellovibrio persistence from different doses in larvae at different developmental stages. (D) Survival curve of WT AB larvae injected with mCherry-Bdellovibrio as in (B) and incubated at 28°C for 48 hpi. Data are pooled from three independent experiments (n = 22–37 larvae per experiment). (E) WT AB zebrafish larvae were injected in the hindbrain ventricle at 3 dpf with >5 × 10³ CFUs of GFP-S. flexneri (green), followed by a hindbrain injection of either PBS or 1–2 × 10⁵ PFUs of mCherry-Bdellovibrio (red), 30–90 min after the initial Shigella infection. Representative images of the hindbrain ventricle in PBS- or Bdellovibrio-treated zebrafish larvae infected with Shigella are shown. Dotted square shows region of interaction between fluorescent Shigella and Bdellovibrio. For each treatment, the same larva was imaged over time. Scale bar, 100 μm. See also Movie S1. (F) Enumeration of live Shigella in homogenates of larvae injected with S. flexneri and treated with injections of either PBS or Bdellovibrio as in (E) over time. Each circle represents a count from an individual larva. Half-filled circles represent enumerations from larvae at time 0 and are representative of inocula for both conditions. Only viable larvae were included in the analysis. Data are pooled from four independent experiments (up to n = 3 larvae per time point per experiment). Mean ± SEM (horizontal bars) is shown. The p values (between conditions at cognate time points) were determined by unpaired one-tailed Student’s t test. Significance was defined as p < 0.05. (G) Survival curve of larvae injected with S. flexneri and treated with either PBS or Bdellovibrio as in (E). Larvae were incubated at 28°C for 72 hpi. Data are pooled from three independent experiments (n = 22–48 larvae per condition per experiment). Up to three larvae per condition were taken for CFUs at 2, 24, 48, and 72 hr time points. The p value between conditions was determined by log-rank Mantel-Cox test. Significance was defined as p < 0.05. See also Figure S1.

Figure 2

Bdellovibrio Prey on Shigella In Vitro and In Vivo inside Living Zebrafish (A) GFP-S. flexneri (green) were incubated in vitro, in the presence or absence of mCherry-Bdellovibrio (red), and visualized by wide-field fluorescent microscopy. Representative images, including rod-shaped Shigella and rounded Shigella invaded by smaller comma-shaped Bdellovibrio, were taken at 1 hr post-mixing. Scale bar, 1 μm. (B) 5–12 × 10⁸ CFUs of GFP-S. flexneri were incubated, in vitro, in 10 mL CaHEPES buffer for 21 hr in the presence or absence of ∼6.2 × 10¹⁰ PFUs of mCherry-Bdellovibrio. Live Shigella were enumerated over time. Data are pooled from three independent experiments. Mean ± SEM (horizontal bars) is shown. The p value between conditions was determined by paired one-tailed Student’s t test. Significance was defined as p < 0.05. (C and D) 2–7 × 10⁷ CFUs of GFP-S. flexneri, 8.4–10.4 × 10⁹ PFUs of mCherry-Bdellovibrio, or both GFP-S. flexneri and mCherry-Bdellovibrio were incubated in vitro in CaHEPES buffer at 37°C. (C) Optical density 600 (OD₆₀₀) representing Shigella numbers (Bdellovibrio are too small to contribute to OD₆₀₀) or (D) mCherry fluorescence intensity representing Bdellovibrio numbers was measured every 30 min for 6 hr using a microplate reader (results plotted every 1 hr). Mean ± SEM from three biological replicates with three technical replicates each is shown. The p value between conditions was determined by paired one-tailed Student’s t test. Significance was defined as p < 0.05. (E) WT AB zebrafish larvae were injected at 3 dpf in the tail muscle with 10³ CFUs of GFP-S. flexneri (green) followed by a tail muscle injection of 1–2 × 10⁵ PFUs of mCherry-Bdellovibrio (red) 30–90 min after the initial Shigella infection. Larvae were imaged by confocal microscopy at 20× magnification. Representative images show the different morphologies of Shigella in vivo, including the typical rod-shaped Shigella (arrow) and also a high proportion of rounded Shigella (arrowheads) at regions of interaction with Bdellovibrio. Scale bar, 10 μm. (F) Representative images of predation of Shigella by Bdellovibrio in vivo, inside a larva injected as in (E) and imaged by high-resolution confocal microscopy at 63× magnification. Frames captured over time show stages of Bdellovibrio (red) invasive predation and rounding of Shigella (green) in vivo. Scale bar, 2.5 μm. mpi, minutes post-infection. See also Movie S2. (G) WT AB zebrafish larvae were injected in the hindbrain ventricle at 3 dpf with 2–6 × 10⁵ CFUs of GFP-S. flexneri (green) alone or followed by a hindbrain injection of 1–30 × 10² PFUs of mCherry-Bdellovibrio (red) 30–90 min after the initial Shigella infection. Bdellovibrio were diluted 100-fold from usual injections to facilitate enumeration of any replicated predators. Enumeration of live Bdellovibrio in PBS-treated homogenates of larvae over time is shown. Each circle represents a count from an individual larva. Half-filled circles represent enumerations from larvae at time 0 and are representative of inocula for both conditions. Only viable larvae were included in the analysis. Data are pooled from two independent experiments (up to n = 3 larvae per time point per experiment). Mean ± SEM (horizontal bars) is shown. The p values (versus the 0 hpi time point) were determined by multiple t test. Significance with Bonferroni correction was defined as p < 0.0125. Of note, p values (not displayed on figure) between conditions at cognate time points were determined by unpaired one-tailed Student’s t test with significance defined as p < 0.05. These are as follows: p < 0.001 between conditions at 5 hr and p < 0.0852 at 24 hr.

Figure 3

Bdellovibrio Is Recognized and Engulfed by Zebrafish Leukocytes In Vivo (A) 1–2 × 10⁵ PFUs of mCherry-Bdellovibrio were injected into the hindbrain ventricle of Tg(mpx:GFP)ⁱ¹¹⁴ zebrafish larvae at 3 dpf, and interactions between neutrophils (green) and Bdellovibrio (red) were visualized by fluorescent stereomicroscopy. Representative images from a single larva over time are shown. Scale bar, 100 μm. See also Movie S3. (B) 1–2 × 10⁵ PFUs of mTeal-Bdellovibrio were injected into the hindbrain ventricle of Tg(mpeg1:Gal4-FF)^gl25/Tg(UAS-E1b:nfsB.mCherry)^c264 zebrafish larvae at 3 dpf, and interactions between macrophages (green) and Bdellovibrio (red) were visualized by fluorescent stereomicroscopy. Representative images from a single larva over time are shown. Scale bar, 100 μm. See also Movie S3. (C) Tg(mpx:GFP)ⁱ¹¹⁴ zebrafish larvae were injected with PBS or Bdellovibrio as in (A), and GFP-expressing neutrophils present in the head region were quantified at 6 hpi. Each circle represents a count from an individual larva. Data are pooled from two independent experiments. The p value between conditions was determined by unpaired one-tailed Student’s t test. Significance was defined as p < 0.05. (D) Tg(mpeg1:Gal4-FF)^gl25/Tg(UAS-E1b:nfsB.mCherry)^c264 zebrafish larvae were injected with PBS or Bdellovibrio as in (B), and mCherry-expressing macrophages present in the head region were quantified at 6 hpi. Each circle represents a count from an individual larva. Data are pooled from two independent experiments. The p value between conditions was determined by unpaired one-tailed Student’s t test. Significance was defined as p < 0.05. (E) Tg(mpx:GFP)ⁱ¹¹⁴ zebrafish larvae were pre-treated using control (CTRL) or Pu.1-targeting morpholino (MO) to deplete leukocytes. Morphants were injected in the hindbrain ventricle at 3 dpf with either PBS or 3–5 × 10⁵ PFU mCherry-Bdellovibrio. Live Bdellovibrio were enumerated from PBS homogenates of larvae. Each circle represents a count from an individual larva. Half-filled circles represent enumerations from larvae at time 0 and are representative of inocula for both conditions. Mean ± SEM (horizontal bars) is shown. The p value (between conditions at cognate time points) was determined by unpaired one-tailed Student’s t test. Significance was defined as p < 0.05. As inoculums from independent experiments were variable up to a log-fold, a representative of three independent experiments performed is shown. See also Figure S2E. See also Figure S2.

Figure 4

Bdellovibrio Work alongside Innate Immune Cells to Protect against Shigella Infection In Vivo (A) Tg(mpx:GFP)ⁱ¹¹⁴ zebrafish larvae were pre-treated using control (CTRL) or Pu.1-targeting morpholinos (MO) to deplete leukocytes. Morphants were injected in the hindbrain ventricle at 3 dpf with >5 × 10³ CFUs of GFP-S. flexneri followed by a hindbrain injection of PBS or 1–2 × 10⁵ PFUs of mCherry-Bdellovibrio 30–90 min after the initial Shigella infection. Live Shigella were enumerated from larval homogenates. Each circle represents a count from an individual larva. Half-filled circles represent enumerations from larvae at time 0 and are representative of inocula for both conditions. Only viable larvae were included in the analysis. Data are pooled from three independent experiments (up to n = 3 larvae per time point per experiment). Mean ± SEM (horizontal bars) is shown. Top graph represents collated data. Bottom graph represents only Bdellovibrio-treated larvae, a subset of the above data. The p value (between conditions at cognate time points) was determined by unpaired one-tailed Student’s t test. Significance was defined as p < 0.05. ND, not determined at 72 hpi due to high morphant mortality reducing the samples available. (B) Survival curve of control (CTRL) or Pu.1 morphant larvae, injected with S. flexneri and treated with Bdellovibrio as in (A). Larvae were incubated at 28°C for 72 hpi. Data are pooled from three independent experiments (n = 12–40 larvae per condition per experiment). Up to three larvae per condition were taken for CFU at 6, 24, and 48 hr time points. Top graph represents collated data. Bottom graph represents only Bdellovibrio-treated larvae, a subset of the above data. The p value between conditions was determined by log-rank Mantel-Cox test. Significance was defined as p < 0.05. (C) Model for the therapeutic benefit of Bdellovibrio as an antibacterial agent against S. flexneri in vivo. The zebrafish immune system alone is unable to control high doses of Shigella (green) injected into the hindbrain; without treatment, bacterial replication results in death of the larva. Injection of live predatory Bdellovibrio (red) 30–90 min after Shigella infection is therapeutically beneficial to the host. Here, live invasive predation of Shigella by Bdellovibrio rounds and then kills the Shigella, significantly reducing host bacterial burden. Remaining Shigella and Bdellovibrio themselves are ultimately cleared by host processes, including leukocyte action. Together, the immune system cooperates with predation to clear bacterial infection and promote survival. See also Figure S3.