Skip to main content
PMC home page
eLife logoLink to eLife
. 2014 Mar 4;3:e01579. doi: 10.7554/eLife.01579

A model symbiosis reveals a role for sheathed-flagellum rotation in the release of immunogenic lipopolysaccharide

Caitlin A Brennan 1, Jason R Hunt 2,3, Natacha Kremer 1, Benjamin C Krasity 1, Michael A Apicella 2,3, Margaret J McFall-Ngai 1, Edward G Ruby 1,*
Editor: Peter Greenberg4
PMCID: PMC3941163  PMID: 24596150

Abstract

Bacterial flagella mediate host–microbe interactions through tissue tropism during colonization, as well as by activating immune responses. The flagellar shaft of some bacteria, including several human pathogens, is encased in a membranous sheath of unknown function. While it has been hypothesized that the sheath may allow these bacteria to evade host responses to the immunogenic flagellin subunit, this unusual structural feature has remained an enigma. Here we demonstrate that the rotation of the sheathed flagellum in both the mutualist Vibrio fischeri and the pathogen Vibrio cholerae promotes release of a potent bacteria-derived immunogen, lipopolysaccharide, found in the flagellar sheath. We further present a new role for the flagellar sheath in triggering, rather than circumventing, host immune responses in the model squid-vibrio symbiosis. Such an observation not only has implications for the study of bacterial pathogens with sheathed flagella, but also raises important biophysical questions of sheathed-flagellum function.

DOI: http://dx.doi.org/10.7554/eLife.01579.001

Research organism: other

eLife digest

While a few of the bacteria that live in and on the bodies of humans and other animals are harmful and can cause disease, most others can offer benefits to their hosts. Many bacteria—including some important human pathogens—have tails called flagella that rotate to move the bacteria inside its host. However, the immune system can detect parts of these flagella and eliminate the pathogen.

Bacterial flagella are made from filaments of proteins, and some flagella are also enclosed by a sheath that is similar to the outer membrane that encloses certain bacteria. The function of this sheath is unclear, although some researchers have suggested that it might prevent the immune system from detecting the proteins in the flagellum. Now, by studying the interactions between the Hawaiian bobtail squid and a marine bacterium, Brennan et al. show that the sheath can actually alert the host that the bacteria are around.

The Hawaiian bobtail squid collects bioluminescent bacteria within a so-called ‘light organ’. This organ undergoes a number of developmental changes to house the bacteria, and the squid then uses the light from the bacteria to mask its own shadow, which helps it to avoid being detected by predators. Brennan et al. compared how wild-type bacteria and mutant bacteria that either had no flagella, or had flagella that did not rotate, interacted with young squid. Only bacteria with working flagella were able to trigger the normal development of the squid’s light organ, which suggests that the rotating flagella are releasing the signal that tells the squid that the beneficial bacteria are present.

Brennan et al. demonstrated that the rotation of sheathed flagella led to the release of a molecule called lipopolysaccharide. This molecule is known to activate the immune system in animals, and it is one of the bacterial signals that the squid responds to. Moreover, when the flagella of other bacteria with sheaths—such as those that cause cholera—are rotating, there is also an increase in the release of lipopolysaccharide. However, rotation of the flagella of bacteria without sheaths has no such effect. The next challenge will be to test the importance of this release of lipopolysaccharide from rotating flagella on the outcome of bacterial diseases of humans and other animals.

DOI: http://dx.doi.org/10.7554/eLife.01579.002

Introduction

Bacterial flagella are important virulence factors that facilitate tissue tropism in host–microbe interactions. Besides mediating such migration, flagella serve additional functions, such as adherence to or invasion of host cells (Young et al., 2000; Hayashi et al., 2001; Giron et al., 2002). However, in most vertebrate models of pathogenesis, it has been difficult to separate the distinct contribution of flagellar motility in bacterial migration from its other roles, such as immune-system activation (Feuillet et al., 2006; Andersen-Nissen et al., 2007). Bacterial flagella are highly conserved on a molecular level, but display diversity in morphology, number and subcellular localization between species. One such variation is the presence of an outer membrane-derived sheath (Fuerst and Perry, 1988; Geis et al., 1993; Ferooz and Letesson, 2010). While the flagella of most bacteria are unsheathed, such a structure surrounds the filament of several polarly flagellated bacteria, including several with a host-associated lifestyle. The biological function(s) of the sheath remains unknown, but it has been posited to allow these bacteria to evade innate immune recognition of the enclosed flagellins (Gewirtz et al., 2004; Yoon and Mekalanos, 2008). This hypothesized role for the flagellar sheath presents a conundrum: what is the effect of coating a rapidly spinning structure with lipopolysaccharide (LPS), another potent immunostimulatory molecule? As the tools with which to study the flagellar sheath are limited, we sought to use the model symbiosis between Vibrio fischeri and the Hawaiian bobtail squid, in which the events of initial colonization can be observed in real-time, to probe whether the sheathed flagellum plays a role in inducing a symbiont-specific host response.

V. fischeri is a gram-negative, bioluminescent bacterium that is found both in seawater and as a beneficial symbiont colonizing the light-emitting organs of certain fishes and squids. In the mutualism with the Hawaiian bobtail squid, Euprymna scolopes, symbiosis begins when a newly hatched squid specifically recruits V. fischeri cells from ambient seawater. The bacteria first aggregate on ciliated appendages of the light organ, and then migrate through pores and down ducts, ultimately reaching internal crypts (Figure 1A) where the symbiont population is housed (Nyholm and McFall-Ngai, 2004). Successful colonization of the crypts ultimately requires flagellar motility, as amotile mutants are unable to colonize the light-organ to luminous levels (Graf et al., 1994; Brennan et al., 2013). However, the symbionts are sensed by host cells even before they reach the crypts. Indeed, as few as 3–5 bacteria interacting with the ciliated cells of the surface appendages signal both organ-wide changes in host transcription and immune-cell (hemocyte) trafficking (Koropatnick et al., 2007; Kremer et al., 2013). Colonizing V. fischeri cells also induce an extensive morphogenesis of the light organ, beginning with the programmed cell death (apoptosis) of the ciliated appendages that begins within 6 hr of initial exposure to the symbiont, and continues until their complete regression (Montgomery and McFall-Ngai, 1994; Foster et al., 2000).

Figure 1. Motility mutants of V. fischeri do not induce early-stage apoptosis.

(A) Morphology of the juvenile light organ, highlighting the external ciliated epithelium (CE; false-colored blue) in a scanning electron micrograph (left), and the internal features with which the bacteria interact during initiation of symbiotic colonization in a schematic diagram (right). DC, deep crypts; AC, antechamber; D, duct; P, pore; and MF, mucus field. (B–E) Representative laser-scanning confocal microscopy (LSCM) images of acridine orange (AO)-stained juvenile light organs, after exposure to either: no V. fischeri (B, apo); wild type (C); flrA (D); or motB1 (E), as described in the ‘Materials and methods’. White arrowhead in (C) indicates one of the numerous AO-positive nuclei (yellow). Scale bar in (B) represents 100 μm for all the images. (F) Counts of AO-positive nuclei in the light-organ ciliated epithelium of squid exposed to either the indicated V. fischeri strains, or no V. fischeri (apo). (***), p<0.001 by Kruskal–Wallis Analysis of Variance (ANOVA), followed by Dunn’s Multiple Comparison test. The absence of apoptosis in the aposymbiotic squid confirms that the experimental addition of peptidoglycan (PGN) (‘Materials and methods’) is insufficient to induce AO staining. (G) Bacterial localization under conditions used for measurement of early-stage apoptosis (panel F). Localization was determined by examining squid exposed to the indicated strains, genetically labeled to express GFP, using LSCM. Each point represents the bacterial localization in a single light-organ lobe. (H) Level of colonization (colony-forming units, CFU) under conditions used for measurement of early-stage apoptosis (panel F). Light organs were dissected from anesthetized and pithed squid, and then homogenized and plated to determine symbiont number.

DOI: http://dx.doi.org/10.7554/eLife.01579.003

Figure 1.

Figure 1—figure supplement 1. Soft-agar motility of wild-type, flrA and motB1 strains.

Figure 1—figure supplement 1.

Wild-type, motB1 and flrA strains were grown to mid-log phase and inoculated into SWT medium supplemented with 0.3% agar. Plates were visualized after incubating for 11 hr at 28°C.
DOI: http://dx.doi.org/10.7554/eLife.01579.004

Figure 1—figure supplement 2. Negative-stain transmission electron micrograph of a representative flagellated motB1 cell.

Figure 1—figure supplement 2.

The motB1 mutant was grown to mid-log phase in SWT broth, and visualized by transmission electron microscopy (TEM) after negative-staining. Scale bar indicates 1 µm.
DOI: http://dx.doi.org/10.7554/eLife.01579.005

Figure 1—figure supplement 3. Negative-stain transmission electron micrograph of a representative aflagellate flrA cell.

Figure 1—figure supplement 3.

The flrA mutant was grown to mid-log phase in SWT broth, and visualized by negative-stain TEM. Scale bar indicates 1 µm.
DOI: http://dx.doi.org/10.7554/eLife.01579.006

Figure 1—figure supplement 4. Representative LSCM image of an AO-stained light-organ lobe isolated from a squid treated with exogenous V. fischeri lipid A.

Figure 1—figure supplement 4.

As described in the ‘Materials and methods’, squid were exposed for 10 hr to purified V. fischeri lipid A, then stained with acridine orange. White arrowheads indicate two of the AO-positive nuclei (yellow). Scale bar represents 100 μm.
DOI: http://dx.doi.org/10.7554/eLife.01579.007

Figure 1—figure supplement 5. Induction of early-stage apoptosis in response to exogenous lipid A.

Figure 1—figure supplement 5.

As described in the ‘Materials and methods’, squid were exposed for 10 hr to either no V. fischeri (apo), wild-type V. fischeri, or purified V. fischeri lipid A alone. Early-stage apoptosis was determined by counting the number of acridine orange (AO)-positive nuclei in the ciliated field of one lobe of the light organ. (***), p<0.001 by Kruskal–Wallis ANOVA, followed by Dunn's Multiple Comparison test.
DOI: http://dx.doi.org/10.7554/eLife.01579.008
Open in a new tab

Results

To address whether the sheathed flagellum mediates symbiont recognition, we first compared the appearance of apoptosis within the host’s ciliated epithelium (Figure 1A) in response to inoculation with either wild-type V. fischeri, or one of two amotile mutants: motB1 (Brennan et al., 2013), which bears normal flagella that do not rotate (Figure 1—figure supplements 1 and 2); and flrA, which is both amotile and aflagellate (Millikan and Ruby, 2003; Brennan et al., 2013; Figure 1—figure supplements 1 and 3). Such amotile mutants do not successfully colonize the light-organ deep crypts (Graf et al., 1994). However, as this apoptosis proceeds over the first few days of symbiotic colonization, we assessed this phenotype an early time-point (10 hr post-inoculation) to optimize apoptopic induction by V. fischeri while selecting conditions under which amotile mutants would not yet exhibit a colonization defect. As previously reported (Montgomery and McFall-Ngai, 1994), squid that were not exposed to V. fischeri (i.e., aposymbiotic, or ‘apo’) exhibited much lower levels of apoptosis (Figure 1B,F) than squid exposed to wild-type V. fischeri (Figure 1C,F). In contrast, exposure to either the motB1 or the flrA strain did not induce early-stage apoptosis above the background levels (Figure 1D–F). Because the actual location at which amotile mutants are arrested in symbiotic initiation has not been defined, we confirmed that both the location and number (Figure 1G,H) of bacteria at this early stage were indistinguishable in squid exposed to wild-type, flrA or motB1 cells. These data show that flagellar motility plays a role in inducing a specific host immune response that is independent of its later role in migration into the light-organ crypts (Brennan et al., 2013), suggesting that biosynthesis of the flagellum, and its subsequent functions, may mediate multiple aspects of symbiotic initiation.

Reports of flagellar-mediated activation of immunity have thus far been limited to responses to the filament protein, flagellin (Hayashi et al., 2001; Franchi et al., 2006). Despite the flagellar sheath, flagellin monomers are a prevalent component of the secretome of both V. cholerae (Xicohtencatl-Cortes et al., 2006) and V. fischeri (unpublished data). However, the failure of the fully flagellated motB1 mutant to induce normal apoptosis demonstrated that the expression of immunogenic flagellin is insufficient to activate apoptosis in the light organ and, instead, suggests that apoptosis is driven by flagellar rotation itself. Previous work (Foster et al., 2000) has shown that apoptosis of the light-organ ciliated epithelium is induced specifically by the lipid-A component of LPS (Figure 1—figure supplements 4 and 5); indeed, LPS is a common activator of innate-immune defenses across the animal kingdom (Choi et al., 1998; Manna and Aggarwal, 1999). However, no effect of flagellar motility on LPS presentation has been reported previously. We posited that, as sheath production and flagellar biogenesis may not be tightly coordinated (Richardson et al., 1990; Josenhans et al., 1995; Ferooz and Letesson, 2010), a sheathed flagellum might shed LPS due to a rotation-weakened association of the sheath with the spinning flagellar filament. This hypothesis is supported by negative-stain transmission electron microscopy (TEM) of sheathed flagella, which has revealed vesicle-like structures (Figure 2A), often at the distal tip of the flagellum, and ∼10–80 nm in diameter (Geis et al., 1993; Millikan and Ruby, 2004; Ferooz and Letesson, 2010). Such images are indicative of weaknesses between the flagellar sheath and shaft, which may be analogous to the foci of disassociation between the outer membrane and peptidoglycan required for the formation of cell-derived outer-membrane vesicles (OMVs) (McBroom and Kuehn, 2007). Thus, rotation of the flagellum would further destabilize these interactions, likely releasing LPS and activating an LPS-dependent immune response, rather than simply shielding against a host response.

Figure 2. Motility mutants of V. fischeri release less LPS into culture supernatants.

(A) Negative-stain TEM of a wild-type V. fischeri cell. Dashed box, shown larger within the solid box, highlights the distal tip of two flagella, one of which displays dissociation of the sheath from the filament in the form of a membrane vesicle-like structure. Scale bar indicates 1 μm. (B) Reactogenic LPS in cell-free supernatants of V. fischeri mid-log (OD600 ≈ 0.5) cultures grown in seawater tryptone (SWT) medium at 28°C was measured by LAL assay. (*), p<0.05; (**), p<0.01; and (***), p<0.001, as analyzed by one-way repeated measures ANOVA, with a posthoc Bonferroni correction. (C) Total LPS levels in cell-free supernatants from mid-log (OD600 ≈ 0.5) cultures grown in SWT at 28°C for indicated strains were determined by quantitative SDS-PAGE analysis (detailed in Figure 2—figure supplement 1).

DOI: http://dx.doi.org/10.7554/eLife.01579.009

Figure 2.

Figure 2—figure supplement 1. Growth of V. fischeri wild-type, motB1 and flrA strains.

Figure 2—figure supplement 1.

Indicated strains were inoculated into SWT and grown at 28°C under microaerobic conditions.
DOI: http://dx.doi.org/10.7554/eLife.01579.010

Figure 2—figure supplement 2. Quantification of total LPS in cell-free supernatants by SDS-PAGE analysis.

Figure 2—figure supplement 2.

(A) Silver-stained gel image of purified V. fischeri LPS separated by SDS-PAGE used to generate a standard curve by densitometry. (B) Total dry weight following dialysis of 50 ml of cell-free supernatants of the indicated strains, and band intensity of 125 μg samples of the purified lipid fractions, visualized by silver staining after SDS-PAGE analysis. Image has been cropped to remove irrelevant lanes, but all samples were separated on the same gel.
DOI: http://dx.doi.org/10.7554/eLife.01579.011
Open in a new tab

We determined whether the motB1 and flrA mutants released less of this signal molecule than wild-type V. fischeri using the Limulus amoebocyte lysate (LAL) assay to quantify the amount of reactogenic LPS in cell-free supernatants prepared from exponentially growing cultures (Figure 2B). While all strains had similar growth characteristics (Figure 2—figure supplement 1), we observed a significant reduction in shed LPS by the motB1 mutant, whose flagella do not rotate, and an even greater reduction by the aflagellate flrA mutant. To confirm that these strains actually released less (and not just a less reactive) LPS, we directly measured the total LPS shed from these strains by purifying the lipid fraction from the cell-free culture supernatants, and visualizing the LPS by quantitative SDS-PAGE (Figure 2—figure supplement 2). This analysis revealed a substantial reduction in total LPS released by both the motB1 and flrA strains (Figure 2C), consistent with the data from the LAL assay (Figure 2B).

To further test our hypothesis, we next examined whether (i) genetic complementation of the motB1 mutant would restore normal levels of LPS shedding; (ii) disruption of flagellar rotation by a motB1-independent method would similarly lessen the LPS observed in culture supernatant; and, (iii) reduced LPS shedding would be observed when flagellar rotation is disrupted in other sheathed bacteria as well. As anticipated, expression of motB1 in trans restored both soft-agar motility and reactogenic LPS levels in cell-free culture supernatants (Figure 3, Figure 3—figure supplement 1). In addition, cell-free culture supernatants from a V. fischeri motX mutant, which lacks a different component of the flagellar motor than motB1, yet is also flagellated but amotile (Figure 4—figure supplements 1 and 2), had similarly reduced levels of reactogenic LPS (Figure 4). Taken together, these data support the notion that flagellar rotation promotes LPS release by V. fischeri cells.

Figure 3. Genetic complementation restores supernatant LPS levels to the motB1 mutant strain.

Levels of reactogenic LPS in mid-log culture supernatants (OD600 ≈ 0.5) of indicated strains, as measured using the LAL assay. (*), p<0.05, and (**), p<0.01, as analyzed by one-way repeated measures ANOVA with a posthoc Bonferroni correction.

DOI: http://dx.doi.org/10.7554/eLife.01579.012

Figure 3.

Figure 3—figure supplement 1. Soft-agar motility of genetically complemented motB1 strains.

Figure 3—figure supplement 1.

Indicated strains were grown to mid-log phase and inoculated into SWT medium containing 0.3% agar. Plates were visualized after incubation for 11 hr at 28°C.
DOI: http://dx.doi.org/10.7554/eLife.01579.013
Open in a new tab

Figure 4. Disruption of motX in V. fischeri reduces supernatant LPS levels.

Levels of reactogenic LPS, as measured using the LAL assay, in cell-free supernatants of mid-log cultures (OD600 ≈ 0.5) of V. fischeri wild-type and motX strains. (**), p<0.01 as determined by paired Student’s t test.

DOI: http://dx.doi.org/10.7554/eLife.01579.014

Figure 4.

Figure 4—figure supplement 1. Soft-agar motility of V. fischeri motX mutant.

Figure 4—figure supplement 1.

Wild-type and motX strains were grown to mid-log phase and inoculated into SWT medium containing 0.3% agar. Plates were visualized after incubating for 8 hr at 28°C.
DOI: http://dx.doi.org/10.7554/eLife.01579.015

Figure 4—figure supplement 2. Negative-stain transmission electron micrograph of a representative motX cell.

Figure 4—figure supplement 2.

The motX mutant was grown to mid-log phase in SWT broth, and visualized by negative-stain TEM, revealing normal sheathed flagella. Scale bar indicates 1 µm.
DOI: http://dx.doi.org/10.7554/eLife.01579.016
Open in a new tab

Finally, we asked whether flagellar rotation-mediated LPS release is characteristic of other bacteria with sheathed flagella by comparing the amount of LPS shed by wild-type Vibrio cholerae O395N1 and its amotile motAB-deletion derivative. As in V. fischeri, disruption of flagellar rotation in V. cholerae significantly reduced LPS release (Figure 5A), supporting the conclusion that this mechanism of LPS release is not unique to V. fischeri, but is conserved among bacteria with sheathed flagella. In contrast, a similar analysis of an Escherichia coli mutant unable to rotate its (unsheathed) flagella revealed no alteration in levels of released LPS (Figure 5B), further indicating that this behavior may be specific to bacteria that sheath their flagella. Nevertheless, it remains possible that rotation-mediated LPS release still occurs in other non-sheathed bacteria, or, perhaps, under other growth conditions.

Figure 5. Loss of flagellar rotation reduces LPS release by sheathed V. cholerae, but not unsheathed E. coli.

Figure 5.

Open in a new tab

(A) Reactogenic LPS released into mid-log (OD600 ≈ 0.5) culture supernatants by wild-type V. cholerae O395N1 and its motAB (i.e., pomAB) derivative, grown in SWT at 37°C, was measured by the LAL assay. (*), p<0.05 by paired Student’s t test. (B) Reactogenic LPS in mid-log (OD600 ≈ 0.5) culture supernatants of wild-type E. coli and its motAB derivative grown in tryptone medium at 30°C, as measured using the LAL assay.

DOI: http://dx.doi.org/10.7554/eLife.01579.017

Discussion

Using the model squid-vibrio symbiosis as a tool to probe the effect of the enigmatic flagellar sheath on host recognition, we uncover a previously unrecognized role for the rotation of sheathed flagella in mediating the release of the immunogenic molecule LPS in both V. fischeri and its pathogenic congener V. cholerae. The flagellar sheath has proved problematic to study; indeed, to our knowledge, there are no mutations that result in the loss of a sheath in otherwise sheathed flagella. Intriguingly, a few Vibrio species, including the human pathogen Vibrio parahaemolyticus, encode two independent flagellar structures: one set producing sheathed, the other unsheathed flagella (McCarter, 1999). However, pleiotropic effects of mutations in these distinct pathways as well as other technical limitations make a direct comparison of unsheathed and sheathed flagella within this species difficult to interpret (L McCarter, personal communication). Thus, the functional role of the sheath, even as it relates to bacterial motility, remains poorly characterized, and observation of this new phenomenon for the flagellar sheath furthers our understanding of this enigmatic structure and its contribution to bacterial virulence. As this behavior is difficult to probe experimentally, novel approaches, such as the modeling of the biophysical properties underlying lipid-membrane behavior at a sub-micron scale, are needed to inform our understanding of the interactions between the sheath and filament along the length of the flagellum, as well as the physical forces at play between a rotating flagellar sheath and the contiguous, but static, cellular outer membrane. While vesicle-like structures have only been reported at the flagellar tip of V. fischeri (Millikan and Ruby, 2004), another possible region from which LPS may be shed during rotation is the base of the flagellum where the basal body perturbs the outer membrane. Interestingly, cryotomographic imaging has revealed unusual features in the Vibrio flagellar basal body that may help stabilize this interaction (Chen et al., 2011).

Additionally, our results show that flagellar rotation in V. fischeri promotes release of the immunogenic molecule LPS that, during initiation of symbiosis, is required for normal induction of apoptotic tissue development (Foster et al., 2000). Further, by activating the squid’s developmental program in response to symbiont exposure, the flagellar sheath may serve to promote recognition by the host immune system, suggesting the context-dependent nature (mutualism or pathogenesis) of this behavior. Perhaps masking one immunostimulatory molecule (e.g., flagellin) with another (e.g., LPS) can serve to modulate, rather than simply avoid (Gewirtz et al., 2004; Yoon and Mekalanos, 2008), detection by the host immune system.

Surprisingly, although both the flrA and motB1 mutants still release LPS in culture, albeit at a reduced level, we observed complete attenuation of apoptotic induction, suggesting that (i) the threshold of LPS sensed by host cells is exquisitely calibrated to the amount of LPS shed by V. fischeri; (ii) because V. fischeri cells are not dividing during symbiotic initiation, flagellar-mediated LPS release is the predominant source of LPS to which host cells are exposed; and/or (iii) the LPS released by wild-type V. fischeri is somehow presented in a more immunogenically active form than that released by either the flrA or motB1 mutant. This last point emphasizes the microheterogeneous chemical structures of LPS made by bacteria like V. fischeri (Phillips et al., 2011). This heterogeneity, as well as LPS’s insolubility in aqueous solution and the toxicity of spent medium to juvenile squid, makes directly comparing the chemical concentration or specific activity of LPS between culture and symbiotic settings problematic. In this symbiosis, the LPS is produced by a few non-growing bacterial cells at a tissue interface, in high flow-field conditions that further confound quantification.

Finally, if LPS is shed from the flagellar tip as vesicles, other molecules such as flagellin might be enclosed, and these components may act synergistically to induce host responses (although flagellin has not yet been characterized as a morphogen in this association). Alternatively, LPS released from either the tip or base of the flagellum might form vesicles with a smaller diameter than the cell-derived OMVs, thereby enabling tighter interaction with the host-cell surfaces. Further work will build upon these discoveries and clarify the mechanisms underlying signaling during initiation of the squid-vibrio symbiosis; such work includes characterizing the composition of LPS released by V. fischeri during flagellar rotation, identifying the squid receptors that recognize V. fischeri LPS (Goodson et al., 2005; Krasity et al., 2011), and determining any role flagellin may have in activating host responses.

While thus far demonstrated only in V. fischeri and V. cholerae, the phenomenon of flagellar-mediated LPS release may be conserved across other animal-associated bacteria with sheathed flagella. The human pathogens Brucella melitensis and Helicobacter pylori, whose sheathed flagellum and/or flagellar motility serve as important virulence factors (Ottemann and Lowenthal, 2002; Petersen et al., 2011), may similarly release the immunogenic molecule LPS during flagellar rotation and, thereby, regulate immune recognition during infection. However, the complexity of mammalian models of host–microbe interactions can limit efforts to separate the various contributions of flagellar motility, and highlights the role of invertebrate model systems in revealing the underlying principles that govern shared mechanisms of host–microbe interactions.

Materials and methods

Bacterial strains and growth conditions

V. fischeri strains used in this study include: the wild-type E. scolopes light-organ isolate, ES114 (Boettcher and Ruby, 1990); flrA::Tnerm, MB21407 (Brennan et al., 2013); motB1::Tnerm, MB06357 (Brennan et al., 2013); and motX::Tnerm, MB12561 (Brennan et al., 2013). Strains were grown at 28°C in either Luria–Bertani salt (LBS) medium (per liter, 10 g Bacto-tryptone, 5 g yeast extract, 20 g NaCl, and 50 ml of 1 M Tris–HCl buffer, pH 7.5, in deionized water) for overnight growth, or seawater tryptone (SWT) medium (per liter, 5 g Bacto-tryptone, 3 g yeast extract, 3 ml glycerol, 700 ml Instant Ocean [Aquarium Systems, Inc., Mentor, OH] at a salinity of 33–35 ppt, and 300 ml distilled water) for experimental use. V. cholerae strains O395N1 and its ΔpomAB derivative (Gosink and Hase, 2000) (referred to as motAB within this work, for consistency with the V. fischeri and E. coli nomenclature) were grown at 37°C in either SWT for LPS measurements, or Luria-Bertani medium (per liter, 10 g Bacto-tryptone, 5 g yeast extract, and 10 g NaCl) for overnight growth. E. coli strains RP437 and its ΔmotAB derivative, RP6894 (Zhou et al., 1998), were grown at 30°C in tryptone medium (per liter, 10 g Bacto-tryptone and 5 g NaCl). When appropriate, antibiotics were added to media at the following concentrations: erythromycin, 5 μg/ml; and chloramphenicol, 2.5 μg/ml. Growth media were solidified with 1.5% agar unless otherwise indicated.

For growth curves, overnight cultures were used to inoculate 100 µl volumes of SWT, arrayed in a 96-well plate, and a plate-reader was used to measure optical density (OD) at 600 nm over 9 hr of growth at 28°C with shaking. Data are presented as the mean, and standard error of the mean, for three biological replicates.

Assessment of motility and flagellar structure

For soft-agar motility assays, cells were grown to OD600 ≈ 0.3–0.4. Cultures were then normalized to an OD600 of 0.3, and 2 μl of each strain were inoculated into plates containing SWT supplemented with 0.3% agar. Plates were grown at 28°C for 8–11 hr. The diameter of migration by the cells is a measure of their level of flagellar motility.

For examination of flagellar structures, cells were grown in SWT broth with shaking at 28°C to an OD600 of ∼0.3, applied to Pioloform-coated copper grids (Ted Pella Co., Tustin, CA) for 1 min, washed with sterile water for 30 s, and negatively stained for 1 min with NanoW (Nanoprobes, Yaphank, NY). Grids were immediately examined using a Philips CM120 transmission electron microscope (University of Wisconsin Medical School Electron Microscope Facility, Madison, WI).

Construction of the motB1 complementation strain

The motB1 open reading frame was amplified by PCR using primer pair motB1_compF (5′-GCTCTAGACTAACACACAGGAAACAGCTATGGAAGATGAAAACGACTGCA-3′), which introduces an XbaI restriction enzyme site and a ribosomal binding site, and motB1_compR (5′ GCGGTACCGACCTAATCTAAGGCGCA-3′), which introduces a KpnI restriction enzyme site. The resulting product was cloned into the XbaI/KpnI-digested fragment of pVSV105 (Dunn et al., 2006). Both the complementation construct (pCAB98) and vector control were conjugated into wild-type V. fischeri and into the motB1 mutant using standard techniques (Stabb and Ruby, 2002).

Measurement of reactogenic LPS using a chromogenic Limulus amoebocyte lysate (LAL) assay

V. fischeri cultures were prepared by growth in SWT broth with shaking to mid-log phase (OD600 ≈ 0.5). Cells were removed by two successive rounds of centrifugation, and the resultant supernatant was passed through a 0.22-μm pore-sized filter, yielding a purified, cell-free supernatant. Cell-free supernatants were serially diluted in pyrogen-free water, and reactogenic LPS was detected from these samples using the ToxinSensor Chromogenic LAL Endotoxin Assay Kit (Genscript, Piscataway, NJ), per manufacturer’s instruction, except in half–volume reactions. Reactogenic LPS units for each sample were normalized to the OD600 of the culture. The LPS in uninoculated media, serially diluted in pyrogen-free water, was determined to be fewer than 50 reactogenic LPS units per ml. Data are presented as the mean, and standard error of the mean, for at least three sets of biological replicates. Statistical analyses were performed with a repeated-measures ANOVA, or paired Student’s t test when appropriate, to normalize for variability in the basal levels of LPS released by wild-type cells across different experiments.

Measurement of total LPS by quantitative, silver-stained, SDS-PAGE analysis

To remove salts from 50 ml of purified cell-free supernatants of V. fischeri (described above in ‘Measurement of reactogenic LPS using a chromogenic Limulus amoebocyte lysate (LAL) assay’), samples were dialyzed against four liters of distilled, deionized H2O using 3500 MWCO SnakeSkin dialysis tubing (ThermoScientific, Rockford, IL) and five water changes. The dialyzed samples were lyophilized and resuspended in 10 ml DNaseI buffer (10 mM Tris–HCl, pH 7.6, 2.5 mM MgCl2, 0.5 mM CaCl2) and treated with 1 mg/ml DNaseI/RNaseA (Roche, Indianapolis, IN) overnight at 37°C. Following nuclease treatment, an equal volume of 95% phenol at 65°C was added. Samples were vortexed and incubated at 65°C for 30 min, then cooled on ice and centrifuged at 3000×g for 10 min at 4°C. The aqueous layer was collected and the phenol layer was back extracted with an equal volume of deionized water pre-warmed to 65°C. The aqueous layers were combined and residual phenol was removed by addition of one-tenth volume of 0.3 M sodium acetate, and precipitation three times with 3 volumes of absolute ethanol. The resulting pellets were resuspended in HPLC-grade water, frozen, and again lyophilized.

Total purified LPS samples from each 50 ml of culture were measured for dry weight, resuspended to a final concentration of 25 μg/μl, and analyzed by SDS-polyacrylamide gel electrophoresis (PAGE) using NuPAGE pre-cast 4–12% Bis-Tris polyacrylamide gels (Novex, Grand Island, NY). The gels were loaded with 5 μl from each preparation, as well as known masses of purified V. fischeri LPS to produce a standard curve. Following electrophoresis, gels were analyzed by silver staining and densitometry using ImageJ. A standard curve was generated based on the band intensities of the purified V. fischeri LPS fractions, and the concentrations of LPS in the culture supernatants were estimated from this standard curve.

Squid colonization experiments

Freshly hatched juvenile squid were exposed to 100 μg of lysozyme-treated peptidoglycan per ml of filter-sterilized Instant Ocean at 32–36 ppt (PGN-FSIO) for 2 hr to induce mucus shedding (Nyholm et al., 2002). In addition to enhancing the solubility of the peptidoglycan, this lysozyme treatment can result in smaller, more immunogenic breakdown products (like TCT). As indicated in individual experiments, either V. fischeri, at a final concentration of ∼5000 colony-forming units (CFU) per ml, or 1 μg/ml V. fischeri lipid A were added to the PGN-FSIO.

After 10 hr, squid were exposed to 0.001% acridine orange (AO) for 5 min, washed in filter-sterilized Instant Ocean (FSIO), and anesthetized in 2% ethanol in FSIO. Squid were then ventrally dissected to expose the light organ, and examined by laser-scanning confocal microscopy (LSCM). AO-positive nuclei in the ciliated epithelial field were counted for one light-organ lobe per squid as a measure of cell-death induction (Delic et al., 1991; Foster et al., 2000). Data are presented as the mean, and standard error of the mean, for 32–38 squid per condition.

For experiments in which symbiont number was determined, squid were colonized as above and, at 10 hr post-inoculation, rinsed three times in FSIO, anesthetized in 2% ethanol in FSIO, and ventrally dissected to remove the light organ. Individual light organs were then homogenized, and each homogenate was diluted and spread onto LBS agar for CFU. Data are presented as the mean, and standard error of the mean, for 36–37 squid per condition.

To visualize bacterial localization, squid were similarly colonized with bacterial strains carrying pVSV102, which harbors a green-fluorescent protein (GFP)-encoding gene under constitutive expression (Dunn et al., 2006), using the conditions described for assaying early-stage apoptosis. After 9.5 hr in PGN-FSIO containing the V. fischeri inoculum, animals were exposed to Cell Tracker Orange CMRA (Invitrogen Molecular Probes, Carlsbad, CA) for 30 min to label host tissue. Squid were then anesthetized in 2% ethanol in FSIO, pithed, and examined by LSCM as described previously (Nyholm et al., 2000; Sycuro et al., 2006). Bacterial localization was determined by the presence of GFP-fluorescing cells, relative to notable features of the juvenile light organ (Figure 1A). Data indicate the location of the bacterial aggregate that has progressed the furthest in a single light-organ lobe, for 13–15 squid per condition. Thus, the decreased induction of apoptosis by these mutants does not result simply from a differential arrest earlier in migration into the light organ crypts.

Scanning electron microscopy was performed on freshly hatched squid, as previously described (Koch et al., 2013).

Acknowledgements

We thank E Koch (University of Wisconsin–Madison) for graciously providing the scanning electron micrograph of a juvenile squid, C Häse (Oregon State University) for supplying V. cholerae strains, JS Parkinson (University of Utah) for providing E. coli strains, and RA Welch (University of Wisconsin–Madison) for sharing research space. We thank Morgan Beeby (Imperial College London) and Rob Phillips (Caltech) for helpful discussions. This work was supported by NIH grants RR12294/OD11024 (to EGR and MJM-N) and AI50661 (to MJM-N). CAB was supported by an NIH Molecular Biosciences Training Grant (T32 GM07215), and an NIH Microbes in Health and Disease Training Grant (T32 AI055397), to the University of Wisconsin–Madison.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Funding Information

This paper was supported by the following grants:

  • National Institutes of Health RR12294, OD11024 to Margaret J McFall-Ngai, Edward G Ruby.

  • National Institutes of Health AI50661 to Margaret J McFall-Ngai.

  • National Institutes of Health Molecular Biosciences Training Grant to University of Wisconsin–Madison to Caitlin A Brennan.

  • National Institutes of Health Microbes in Health and Disease Training Grant to Caitlin A Brennan.

Additional information

Competing interests

The authors declare that no competing interests exist.

Author contributions

CAB, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article.

JRH, Acquisition of data, Analysis and interpretation of data.

NK, Analysis and interpretation of data, Drafting or revising the article.

BCK, Analysis and interpretation of data, Drafting or revising the article.

MAA, Analysis and interpretation of data, Drafting or revising the article.

MJM-N, Conception and design, Analysis and interpretation of data, Drafting or revising the article.

EGR, Conception and design, Analysis and interpretation of data, Drafting or revising the article.

References

  1. Andersen-Nissen E, Hawn TR, Smith KD, Nachman A, Lampano AE, Uematsu S, Akira S, Aderem A. 2007. Cutting edge: Tlr5-/- mice are more susceptible to Escherichia coli urinary tract infection. Journal of Immunology 178:4717–4720 [DOI] [PubMed] [Google Scholar]
  2. Boettcher KJ, Ruby EG. 1990. Depressed light emission by symbiotic Vibrio fischeri of the sepiolid squid Euprymna scolopes. Journal of Bacteriology 172:3701–3706 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Brennan CA, Mandel MJ, Gyllborg MC, Thomasgard KA, Ruby EG. 2013. Genetic determinants of swimming motility in the squid light-organ symbiont Vibrio fischeri. Microbiology Open 2:576–594. 10.1002/mbo3.96 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Chen S, Beeby M, Murphy GE, Leadbetter JR, Hendrixson DR, Briegel A, Li Z, Shi J, Tocheva EI, Müller A, Dobro MJ, Jensen GJ. 2011. Structural diversity of bacterial flagellar motors. European Molecular Biology Organization 30:2972–2981. 10.1038/emboj.2011.186 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Choi KB, Wong F, Harlan JM, Chaudhary PM, Hood L, Karsan A. 1998. Lipopolysaccharide mediates endothelial apoptosis by a FADD-dependent pathway. The Journal of Biological Chemistry 273:20185–20188. 10.1074/jbc.273.32.20185 [DOI] [PubMed] [Google Scholar]
  6. Delic J, Coppey J, Magdelenat H, Coppey-Moisan M. 1991. Impossibility of acridine orange intercalation in nuclear DNA of the living cell. Experimental Cell Research 194:147–153. 10.1016/0014-4827(91)90144-J [DOI] [PubMed] [Google Scholar]
  7. Dunn AK, Millikan DS, Adin DM, Bose JL, Stabb EV. 2006. New rfp- and pES213-derived tools for analyzing symbiotic Vibrio fischeri reveal patterns of infection and lux expression in situ. Applied and Environmental Microbiology 72:802–810. 10.1128/AEM.72.1.802-810.2006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Ferooz J, Letesson JJ. 2010. Morphological analysis of the sheathed flagellum of Brucella melitensis. BMC Research Notes 3:333. 10.1186/1756-0500-3-333 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Feuillet V, Medjane S, Mondor I, Demaria O, Pagni PP, Galan JE, Flavell RA, Alexopoulou L. 2006. Involvement of Toll-like receptor 5 in the recognition of flagellated bacteria. Proceedings of the National Academy of Sciences of the United States of America 103:12487–12492. 10.1073/pnas.0605200103 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Foster JS, Apicella MA, McFall-Ngai MJ. 2000. Vibrio fischeri lipopolysaccharide induces developmental apoptosis, but not complete morphogenesis, of the Euprymna scolopes symbiotic light organ. Developmental Biology 226:242–254. 10.1006/dbio.2000.9868 [DOI] [PubMed] [Google Scholar]
  11. Franchi L, Amer A, Body-Malapel M, Kanneganti TD, Ozoren N, Jagirdar R, Inohara N, Vandenabeele P, Bertin J, Coyle A, Grant EP, Núñez G. 2006. Cytosolic flagellin requires Ipaf for activation of caspase-1 and interleukin 1beta in salmonella-infected macrophages. Nature Immunology 7:576–582. 10.1038/ni1346 [DOI] [PubMed] [Google Scholar]
  12. Fuerst JA, Perry JW. 1988. Demonstration of lipopolysaccharide on sheathed flagella of Vibrio cholerae O:1 by protein A-gold immunoelectron microscopy. Journal of Bacteriology 170:1488–1494 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Geis G, Suerbaum S, Forsthoff B, Leying H, Opferkuch W. 1993. Ultrastructure and biochemical studies of the flagellar sheath of Helicobacter pylori. Journal of Medical Microbiology 38:371–377. 10.1099/00222615-38-5-371 [DOI] [PubMed] [Google Scholar]
  14. Gewirtz AT, Yu Y, Krishna US, Israel DA, Lyons SL, Peek RM., Jnr 2004. Helicobacter pylori flagellin evades toll-like receptor 5-mediated innate immunity. The Journal of Infectious Diseases 189:1914–1920. 10.1086/386289 [DOI] [PubMed] [Google Scholar]
  15. Giron JA, Torres AG, Freer E, Kaper JB. 2002. The flagella of enteropathogenic Escherichia coli mediate adherence to epithelial cells. Molecular Microbiology 44:361–379. 10.1046/j.1365-2958.2002.02899.x [DOI] [PubMed] [Google Scholar]
  16. Goodson MS, Kojadinovic M, Troll JV, Scheetz TE, Casavant TL, Soares MB, McFall-Ngai MJ. 2005. Identifying components of the NF-kappaB pathway in the beneficial Euprymna scolopes-Vibrio fischeri light organ symbiosis. Applied and Environmental Microbiology 71:6934–6946. 10.1128/AEM.71.11.6934-6946.2005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Gosink KK, Hase CC. 2000. Requirements for conversion of the Na(+)-driven flagellar motor of Vibrio cholerae to the H(+)-driven motor of Escherichia coli. Journal of Bacteriology 182:4234–4240. 10.1128/JB.182.15.4234-4240.2000 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Graf J, Dunlap PV, Ruby EG. 1994. Effect of transposon-induced motility mutations on colonization of the host light organ by Vibrio fischeri. Journal of Bacteriology 176:6986–6991 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hayashi F, Smith KD, Ozinsky A, Hawn TR, Yi EC, Goodlett DR, Eng JK, Akira S, Underhill DM, Aderem A. 2001. The innate immune response to bacterial flagellin is mediated by Toll-like receptor 5. Nature 410:1099–1103. 10.10398/35074106 [DOI] [PubMed] [Google Scholar]
  20. Josenhans C, Labigne A, Suerbaum S. 1995. Comparative ultrastructural and functional studies of Helicobacter pylori and Helicobacter mustelae flagellin mutants: both flagellin subunits, FlaA and FlaB, are necessary for full motility in Helicobacter species. Journal of Bacteriology 177:3010–3020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Koch EJ, Miyashiro T, McFall-Ngai ML, Ruby EG. 2013. Features governing symbiont persistence in the squid-vibrio association. Molecular Ecology 10.1111/mec.12474 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Koropatnick TA, Kimbell JR, McFall-Ngai MJ. 2007. Responses of host hemocytes during the initiation of the squid-Vibrio symbiosis. The Biological Bulletin 212:29–39. 10.2307/25066578 [DOI] [PubMed] [Google Scholar]
  23. Krasity BC, Troll JV, Weiss JP, McFall-Ngai MJ. 2011. LBP/BPI proteins and their relatives: conservation over evolution and roles in mutualism. Biochemical Society Transactions 39:1039–1044. 10.1042/BST0391039 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kremer NK, Philipp E, Carpentier MC, Brennan CA, Kraemer L, Altura MA, Augustin R, Häsler R, Heath-Heckman EA, Peyer SM, Schwartzman J, Rader BA, Ruby EG, Rosenstiel P, McFall-Ngai MJ. 2013. Initial symbiont contact orchestrates host-organ-wide transcriptional changes that prime tissue colonization. Cell Host & Microbe 14:183–194. 10.1016/j.chom.2013.07.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Manna SK, Aggarwal BB. 1999. Lipopolysaccharide inhibits TNF-induced apoptosis: role of nuclear factor-kappaB activation and reactive oxygen intermediates. Journal of Immunology 162:1510–1518 [PubMed] [Google Scholar]
  26. McBroom AJ, Kuehn MJ. 2007. Release of outer membrane vesicles by Gram-negative bacteria is a novel envelope stress response. Molecular Microbiology 63:545–558. 10.1111/j.1365-2958.2006.05522.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. McCarter L. 1999. The multiple identities of Vibrio parahaemolyticus. Journal of Molecular Microbiology Biotechnology 1:51–57 [PubMed] [Google Scholar]
  28. Millikan DS, Ruby EG. 2003. FlrA, a σ54-dependent transcriptional activator in Vibrio fischeri, is required for motility and symbiotic light-organ colonization. Journal of Bacteriology 185:3547–3557. 10.1128/JB.185.12.3547-3557.2003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Millikan DS, Ruby EG. 2004. Vibrio fischeri flagellin A is essential for normal motility and for symbiotic competence during initial squid light organ colonization. Journal of Bacteriology 186:4315–4325. 10.1128/JB.186.13.4315-4325.2004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Montgomery MK, McFall-Ngai M. 1994. Bacterial symbionts induce host organ morphogenesis during early postembryonic development of the squid Euprymna scolopes. Development 120:1719–1729 [DOI] [PubMed] [Google Scholar]
  31. Nyholm SV, Deplancke B, Gaskins HR, Apicella MA, McFall-Ngai MJ. 2002. Roles of Vibrio fischeri and nonsymbiotic bacteria in the dynamics of mucus secretion during symbiont colonization of the Euprymna scolopes light organ. Applied and Environmental Microbiology 68:5113–5122. 10.1128/AEM.68.10.5113-5122.2002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Nyholm SV, McFall-Ngai MJ. 2004. The winnowing: establishing the squid-vibrio symbiosis. Nature Reviews Microbiology 2:632–642. 10.1038/nrmicro957 [DOI] [PubMed] [Google Scholar]
  33. Nyholm SV, Stabb EV, Ruby EG, McFall-Ngai MJ. 2000. Establishment of an animal-bacterial association: recruiting symbiotic vibrios from the environment. Proceedings of the National Academy of Sciences of the United States of America 97:10231–10235. 10.1073/pnas.97.18.10231 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Ottemann KM, Lowenthal AC. 2002. Helicobacter pylori uses motility for initial colonization and to attain robust infection. Infection and Immunity 70:1984–1990. 10.1128/IAI.70.4.1984-1990.2002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Petersen E, Chaudhuri P, Gourley C, Harms J, Splitter G. 2011. Brucella melitensis cyclic di-GMP phosphodiesterase BpdA controls expression of flagellar genes. Journal of Bacteriology 193:5683–5691. 10.1128/JB.00428-11 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Phillips NJ, Adin DM, Stabb EV, McFall-Ngai MJ, Apicella MA, Gibson BW. 2011. The lipid A from Vibrio fischeri lipopolysaccharide: a unique structure bearing a phosphoglycerol moiety. The Journal of Biological Chemistry 286:21203–21219. 10.1074/jbc.M111.239475 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Richardson K, Nixon L, Mostow P, Kaper JB, Michalski J. 1990. Transposon-induced non-motile mutants of Vibrio cholerae. Journal of General Microbiology 136:717–725. 10.1099/00221287-136-4-717 [DOI] [PubMed] [Google Scholar]
  38. Stabb EV, Ruby EG. 2002. RP4-based plasmids for conjugation between Escherichia coli and members of the Vibrionaceae. Methods in Enzymology 358:413–426. 10.1016/S0076-6879(02)58106-4 [DOI] [PubMed] [Google Scholar]
  39. Sycuro LK, Ruby EG, McFall-Ngai M. 2006. Confocal microscopy of the light organ crypts in juvenile Euprymna scolopes reveals their morphological complexity and dynamic function in symbiosis. Journal of Morphology 267:555–568. 10.1002/jmor.10422 [DOI] [PubMed] [Google Scholar]
  40. Xicohtencatl-Cortes J, Lyons S, Chaparro AP, Hernandez DR, Saldana Z, Ledesma MA, Rendón MA, Gewirtz AT, Klose KE, Girón JA. 2006. Identification of proinflammatory flagellin proteins in supernatants of Vibrio cholerae O1 by proteomics analysis. Molecular & Cellular Proteomics 5:2374–2383. 10.1074/mcp.M600228-MCP200 [DOI] [PubMed] [Google Scholar]
  41. Yoon SS, Mekalanos JJ. 2008. Decreased potency of the Vibrio cholerae sheathed flagellum to trigger host innate immunity. Infection and Immunity 76:1282–1288. 10.1128/IAI.00736-07 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Young GM, Badger JL, Miller VL. 2000. Motility is required to initiate host cell invasion by Yersinia enterocolitica. Infection and Immunity 68:4323–4326. 10.1128/IAI.68.7.4323-4326.2000 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Zhou J, Sharp LL, Tang HL, Lloyd SA, Billings S, Braun TF, Blair DF. 1998. Function of protonatable residues in the flagellar motor of Escherichia coli: a critical role for Asp 32 of MotB. Journal of Bacteriology 180:2729–2735 [DOI] [PMC free article] [PubMed] [Google Scholar]
eLife. 2014 Mar 4;3:e01579. doi: 10.7554/eLife.01579.018

Decision letter

Editor: Peter Greenberg1

eLife posts the editorial decision letter and author response on a selection of the published articles (subject to the approval of the authors). An edited version of the letter sent to the authors after peer review is shown, indicating the substantive concerns or comments; minor concerns are not usually shown. Reviewers have the opportunity to discuss the decision before the letter is sent (see review process). Similarly, the author response typically shows only responses to the major concerns raised by the reviewers.

Thank you for sending your work entitled “Rotation of sheathed bacterial flagella promotes an immune response during host colonization in a model symbiosis” for consideration at eLife. Your article has been favorably evaluated by a Senior editor and 3 reviewers, one of whom is a member of our Board of Reviewing Editors.

The Reviewing editor and the other reviewers discussed their comments before we reached this decision, and the Reviewing editor has assembled the following comments to help you prepare a revised submission.

The three reviewers all believe your manuscript is very interesting and appropriate for eLife, but some modifications are needed for it to gain acceptance. This will require addition of some experimental data. The suggestions in the individual reviews are pretty clear and consistent:

1) We feel that you need to show that the levels of LPS in wildtype vs mutant supernatants can account for the relative capacities to induce apoptosis.

2) We also think it is important to measure peptidoglycan in the supernatants. Your previous work shows it is involved in the apoptotic response or some other way nailing down the conclusion that it is specifically shed LPS that is responsible for apoptosis.

We hope you can modify the manuscript in accordance with the reviewer suggestions below. It will then make a very nice contribution to the journal. Please provide a point-by-point disposition with a revised manuscript.

Reviewer #1:

Brennen et al. is a clear easy read that uncovers a very interesting function of the sheath encasing flagella of a handful of bacterial species that exhibit important associations with animals. These associations include those of important human pathogens like Vibrio cholerae. The work uncovered that rotation of the flagella results in release of LPS (endotoxin) from bacterial cells. The LPS serves to stimulate apoptosis in relevant host cells. What I particularly like about the manuscript is that the authors use a special mutualistic association that is very well suited to uncover this novel function of sheathed flagella. They go on to present some limited evidence that this could very well be relevant in the pathogen V. cholerae. I have one single concern of substance and a number of more cosmetic issues.

The substantive concern is what is the role of rotating flagella in establishment of the symbiosis? Maybe I missed it in the paper. Can MotAB and MotX mutants successfully colonize the squid even though they can’t trigger apoptosis? If so, why and what does this say about the importance of flagellar sheath shedding and apoptosis? If the infection by Mot mutants does not proceed normally, where does it stall out?

Minor comments:

“Despite the flagellar sheath, flagellin monomers are a prevalent component of the secretome of V. fischeri (Graber and Ruby, unpublished data)”: Is it acceptable to cite unpublished data in eLife?

“Assuming our hypothesis is accurate” is unfortunate wording. Science should assume hypotheses are incorrect and attempt to disprove them by experimentation, not the other way around. Change wording to something about “To further test the hypothesis...”

“However, pleiotropic effects of mutations in these distinct pathways (L. McCarter, personal communication) prevent direct comparison of unsheathed and sheathed flagella within one species”: Another personal communication.

“As this behavior is difficult to probe experimentally, novel approaches, such as the modeling of the biophysical properties underlying lipid interactions at a sub-micron scale, are needed to inform our understanding of the forces at play between a rotating flagellar sheath and the contiguous, but static, cellular outer membrane”: This text is about how one might study this further. It is a vague statement about biophysical approaches. It is not useful.

Figure 1G: I don't quite get this figure. It is just the presence or absence of bacteria in these locations? If so it should be clear. If something can be said about abundance it would be useful.

Figure 1–figure supplement 2: The images are of full-page blow-ups of single bacterial cells (one with flagella and one without). First they don't need to be this big. Second, and more important, images of single cells are not convincing.

Figure 2–figure supplement 1: Bacterial growth is usually plotted on a log scale over linear time because growth is logarithmic. I recommend revising this figure.

Reviewer #2:

The manuscript by Ruby and colleagues presents an intriguing theory that the rotation of sheathed flagella found on certain bacteria results in release of lipopolysaccharide that can serve to stimulate host innate immune pathways. They present convincing data that Vibrio fischeri strains with no or amotile flagella produce less secreted LPS. They also present data showing that these strains have less capacity to induce epithelial apoptosis in the squid, the first step of symbiont-induced morphogenesis. They do not, however, establish whether the reduction in secreted LPS in these strains is responsible for the failure to induce morphogenesis. They also do not establish whether the alleged outer membrane vesicles at the tips of wild type flagella, observed in electron micrographs, require flagellar motility for their production.

1) A key basis for this paper is the assertion that LPS acts as the signal for the symbiont-induced morphogenesis of the squid host. The data in Figure 1–figure supplement 5 showing comparable apoptosis per ciliated field producing by lipid A exposure and V. fischeri colonization is different from the results reported in a previous paper from this group, by Koropatnick et al. 2004 (Figure 2H), in which LPS has a modest effect inducing apoptosis and interacts synergistically with peptidoglycan to induce this effect. It's possible that lipid A is more potent than crude LPS. However, the real difference in these two sets of results is the extent of apoptosis induced by the symbiont, with significantly lower numbers in this manuscript's figure. The authors should comment on the differences between these two sets of results.

2) Based on the Koropatnick et al. 2004 results showing synergistic induction of apoptosis with combined LPS and PGN, it might be possible that the lack of morphogenesis induced by the flrA and motB mutants is due to the reduction of multiple morphogenic signals, as opposed to just LPS. For example, rotation of the flagellar shaft at the point of connection with the cellular membrane could enhance release of PGN. The authors should test whether levels of peptidoglycan and/or peptidoglycan monomers are reduced in the supernatant of the flrA, motB, and motX mutants relative to wild type cells. Additionally, the flagellin monomers that the authors state are secreted by wild type V. fischeri may act in conjunction with LPS to stimulate morphogenesis and these may well be reduced in the mutants.

3) The set of experiments that would convincingly show that the reduction in secreted LPS in the flrA and motB mutants is responsible for their failure to induce apoptosis would be to 1) show that secreted fractions from the wild type but not the mutant cells can induce apopotosis, and 2) show that depletion of LPS from the wild type cells (for example with the antibody used in Koropatnick et al.) eliminated that apoptosis-inducing activity in the secreted fraction from the wild type cells. These may be technically challenging experiments, but the authors need to provide more data on the connection between what these cells secrete and the functional capacities of these secretions in the symbiosis.

4) The authors discuss the issue of the threshhold concentrations of LPS to induce apoptosis, but they should provide more information on the functional and observed concentrations of these molecules. The experiment shown in Figure 1–figure supplement 5 uses 1 ug/ml lipid A. How does this compare to the concentrations of lipid A present in the LPS released by the flrA, motB, and motX mutants?

5) Based on the characterization of mutants with non-motile but intact flagella, the authors postulate that the biophysics of flagellar rotation result in LPS release. Their one piece of microscopic evidence for this is the electron micrograph in Figure 2A, allegedly showing vesicular release from flagellar tips. The authors should provide more analysis of this phenomenon. First are the alleged vesicles comparable in size to outer membrane vesicles? Second, at what frequency are they observed? Finally, if they are the result of flagellar rotation, then they should not be observed in preparations of cells with amotile flagella. Is this the case?

Reviewer #3:

The manuscript by Brennan et al. investigates the sheathed flagellum of Vibrio fischeri. The reason for the sheath, an extension of the outermembrane, covering the filament has remained mysterious. The authors suggest that it may be provoking an immune response to LPS. They showed that LPS is shed into the environment when the flagellum is rotating, and that bacteria that cannot rotate their flagellum are defective at inducing apoptosis in host squid cells. They additionally showed that a flagellated but non-motile V. cholerae strain also releases less LPS into the environment and suggest that this might be conserved among all bacteria with sheathed flagella. In general the observation is very interesting and may point at a reason for the existence of the sheath surrounding the flagellar filament. Specific comments on the manuscript:

1) One piece was missing from their story: they showed that flagellated non-motile bacteria are defective at inducing apoptosis, and that these bacteria shed less LPS, but they didn’t show that the supernatants from these bacteria (vs the wildtype supe) are defective at inducing apoptosis. Given their argument that the shed LPS is what is stimulating apoptosis, they should show that the supes from the bacteria do/don’t induce apoptosis in the squid cells, rather than using purified LPS. They should do the same with the V. cholerae supes as well, since the source of the LPS shouldn’t make much difference.

2) Another control should be included, which is the addition of phenamil, a specific inhibitor of the sodium-driven motor, to the wildtype cells-this will stop flagellar rotation, and if their hypothesis is correct, there will be less LPS in the supe, which will stimulate less apoptosis.

eLife. 2014 Mar 4;3:e01579. doi: 10.7554/eLife.01579.019

Author response

1) We feel that you need to show that the levels of LPS in wildtype vs mutant supernatants can account for the relative capacities to induce apoptosis.

We understand the concern underlying this comment; however, there are two experimental realities that make this point not as easily resolvable as we, or the reviewers, would like. First, addition of culture medium is toxic to the animal, so this approach is untenable. Second, LPS is not a molecule amenable to simple solution-chemistry assumptions; i.e., it is produced by bacterial cultures in several chemical and physical forms (Phillips et al., 2011), each with a different specific activity, depending on the target/receptor.

For these reasons, which we elaborate upon below in the specific responses, our work simply proposes that because (i) non-rotating flagella release less total LPS, and (ii) LPS is the only molecule known to induce early-stage morphogenesis, then the most parsimonious and reasonable (though not airtight) explanation is that non-rotating flagella induce less early-stage morphogenesis because of their lower level of LPS release. We also feel that the strength and impact of the manuscript is in its use of the squid-vibrio system to discover a new phenomenon of general importance (see Reviewer #1’s first paragraph), rather than in using the phenomenon to describe something new about already extensively studied signals in this specific symbiosis.

To address the reviewers’ concern, we have changed the title, and made modifications throughout the manuscript, as well as others noted in the specific responses, below. These changes do not lower the impact of the work, which uses a model symbiosis to unveil a new biological behavior that occurs in V. fischeri, V. cholerae, and perhaps other bacteria, with consequences of importance in host-microbe interactions.

2) We also think it is important to measure peptidoglycan in the supernatants. Your previous work shows it is involved in the apoptotic response or some other way nailing down the conclusion that it is specifically shed LPS that is responsible for apoptosis.

We thank the reviewers for their careful consideration of the published literature in examining our work. However, there appears to be some confusion with regard to this point. LPS, particularly the lipid-A moiety, specifically induces early-stage apoptosis (as measured by acridine-orange staining) during symbiotic initiation (the appropriate reference for these data is Foster et al., 2000, not Koropatnick et al., 2004). The Koropatnick et al. paper measures apoptosis after 24 h, well into the persistence stage of the symbiosis, and many hours after V. fischeri cells have already colonized the light-organ crypts and undergone substantial proliferation (∼100,000-fold). At such post-colonization time points, LPS/lipid A synergizes with TCT to drive apoptosis into its later stages. In contrast, at the 10-h time point examined in this study, LPS/lipid A is the inducing molecule, and TCT has no known synergistic effect. This difference in experimental timing (10 h vs 24 h) also accounts for the higher numbers of apoptotic nuclei reported in that publication (as noted by Reviewer 2).

Additionally, as indicated in the Methods, all of our experiments were performed in the presence of exogenously added peptidoglycan, which likely contains TCT as a breakdown product, a point that we have added to the text for clarification. Significantly, this cue is always present in natural seawater because of the presence of bacterioplankton, and it causes the hatchling squid to produce mucus, a requirement for efficient colonization (Nyholm et al., 2002). The level of peptidoglycan added to the bacteria-free artificial seawater to induce mucus shedding in this experiment would likely negate the effect of any differences in TCT release between the test strains and the aposymbiotic condition at this early time point (Figure 1, panels B and F). In fact, the absence of significant apoptosis in the aposymbiotic squid (which were exposed to the added peptidoglycan) confirms that this addition is insufficient to induce acridine-orange staining (Figure 1F). We have added this point to the figure legend for clarification. Further, unpublished data from a different study that examined another host response during initiation (hemocyte trafficking into the light organ appendages, which is induced by TCT alone) indicates there is no significant difference in this TCT-dependent response between squid exposed to either wild-type or motB1 strains of V. fischeri.

Nevertheless, we agree with the reviewer that it certainly remains possible that, in addition to LPS, other molecules (peptidoglycan, flagellin, outer membrane proteins, etc.) are differentially released upon flagellar rotation, and we have added this idea to the Discussion.

Reviewer #1:

Brennen et al. is a clear easy read that uncovers a very interesting function of the sheath encasing flagella of a handful of bacterial species that exhibit important associations with animals. These associations include those of important human pathogens like Vibrio cholerae. The work uncovered that rotation of the flagella results in release of LPS (endotoxin) from bacterial cells. The LPS serves to stimulate apoptosis in relevant host cells. What I particularly like about the manuscript is that the authors use a special mutualistic association that is very well suited to uncover this novel function of sheathed flagella. They go on to present some limited evidence that this could very well be relevant in the pathogen V. cholerae. I have one single concern of substance and a number of more cosmetic issues.

The substantive concern is what is the role of rotating flagella in establishment of the symbiosis? Maybe I missed it in the paper. Can MotAB and MotX mutants successfully colonize the squid even though they can’t trigger apoptosis? If so, why and what does this say about the importance of flagellar sheath shedding and apoptosis? If the infection by Mot mutants does not proceed normally, where does it stall out?

We thank the reviewer for alerting us to the need to be explicit about this point. We have added a phrase to clarify that amotile mutants, including the motB1 and motX mutants, do not successfully colonize the juvenile light organ. The hypothesis in the field, as with other systems, is that it is the inability of these mutants to correctly localize to their symbiotic niche – in this case, the deep crypts – that prevents colonization, but the precise stage/location at which initiation is arrested has not yet been thoroughly determined. Therefore, it remains possible that the defects of the strains in colonizing the light organ are ultimately due, in part, to defects in host recognition/response rather than simply in their failure to migrate into the deep crypts. To clarify this latter point, we have added a phrase to the text.

Minor comments:

“Despite the flagellar sheath, flagellin monomers are a prevalent component of the secretome of V. fischeri (Graber and Ruby, unpublished data)”: Is it acceptable to cite unpublished data in eLife?

While the data from V. fischeri are unpublished, the results are not unanticipated from studies of other sheathed bacteria. We have maintained the citation to our unpublished data, but we have also added a reference to work performed in V. cholerae (Xicohtencatl-Cortes et al., 2006) and modified the sentence to strengthen our statement.

“Assuming our hypothesis is accurate” is unfortunate wording. Science should assume hypotheses are incorrect and attempt to disprove them by experimentation, not the other way around. Change wording to something about “To further test the hypothesis...”

We agree with the reviewer, and we have modified this sentence in accord with their comment.

“However, pleiotropic effects of mutations in these distinct pathways (L. McCarter, personal communication) prevent direct comparison of unsheathed and sheathed flagella within one species”: Another personal communication.

While we would like to replace this personal communication with an appropriate citation, the problem we are addressing has not been explicitly stated in the literature. Linda McCarter is the expert in the field of Vibrio parahaemolyticus flagellar motility, and, when asked about the potential for such an experiment, she responded “V[ibrio] P[arahaemolyticus] is complicated by the biology of the two interacting flagellar systems and the not well-matched backgrounds of the existing mutants,” and described the collateral effects upon other cellular behaviors of disrupting either flagellar system. (Dr. McCarter gave us permission to quote her in our response.) We have modified this sentence to more accurately reflect her observation.

“As this behavior is difficult to probe experimentally, novel approaches, such as the modeling of the biophysical properties underlying lipid interactions at a sub-micron scale, are needed to inform our understanding of the forces at play between a rotating flagellar sheath and the contiguous, but static, cellular outer membrane”: This text is about how one might study this further. It is a vague statement about biophysical approaches. It is not useful.

We would like to argue that this text helps place our research into a broader context. In numerous discussions of the flagellar sheath with other scientists, including both biophysicists and microbiologists who are experts on the outer membrane and its stresses, it has remained a point of amazement that no one has considered how such a modification to the flagellar apparatus might affect membrane stability. Therefore, we request permission to retain this single sentence in the Discussion so that readers are made aware of this current lack of understanding about sheathed-flagellum function.

Figure 1G: I don't quite get this figure. It is just the presence or absence of bacteria in these locations? If so it should be clear. If something can be said about abundance it would be useful.

The data in Figure 1G demonstrate that the tissue location of the flrA and motB1 mutants is the same as that of wild-type cells at the time we are comparing them (i.e., the earliest stages of initiation), and that the phenotypes we report for these mutants do not result simply from a differential arrest earlier in migration into the light organ crypts (as mentioned above, the point of their ultimate arrest has not yet been defined). However, we can appreciate the reviewer’s confusion with this figure, and have added a sentence in the Materials and methods to clarify that these data indicate the relative tissue location of mutant and wild-type bacteria at the time of our assays.

Concerning abundance data, while the squid-vibrio system allows much greater resolution of the initiation process than vertebrate models, enumerating bacteria at different locations within live animals still is a difficult task, especially after the bacteria get further into host tissue (e.g., through the pores and into the ducts and antechamber). However, as indicated in the figure legend, we have paired the data in Figure 1G with a determination of the total colony-forming units in the entire organ (Figure 1H), and these values are indistinguishable between the strains.

Figure 1–figure supplement 2: The images are of full-page blow-ups of single bacterial cells (one with flagella and one without). First they don't need to be this big. Second, and more important, images of single cells are not convincing.

We apologize for the confusion. These images are not intended to be full-sized. We believe that’s just the way the upload-process handles the figure supplements. The images are provided to support the statements that the flrA and motB1 mutants have the expected flagellar phenotypes (aflagellate for flrA, and flagellate for motB1) just like in homologous mutants in other bacteria. The images are simply representative cells, which we have clarified in the figure titles (Figure 1–figure supplements 2 and 3, and Figure 4–figure supplement 2). While expected, because of studies of homologous mutants in V. cholerae and other Vibrio spp., the presence or absence of a flagellum had not been described previously for neither any known flagellar motor mutant in V. fischeri nor this particular flrA mutant, although these data have been reported for a previous construction of an flrA mutant (Millikan and Ruby, 2003). As these results are unsurprising, such confirmation seems useful, but not essential, for interpretation of our results, and is placed as supplementary information.

Figure 2–figure supplement 1: Bacterial growth is usually plotted on a log scale over linear time because growth is logarithmic. I recommend revising this figure.

We have modified the figure per the recommendation of the reviewer.

Reviewer #2:

The manuscript by Ruby and colleagues presents an intriguing theory that the rotation of sheathed flagella found on certain bacteria results in release of lipopolysaccharide that can serve to stimulate host innate immune pathways. They present convincing data that Vibrio fischeri strains with no or amotile flagella produce less secreted LPS. They also present data showing that these strains have less capacity to induce epithelial apoptosis in the squid, the first step of symbiont-induced morphogenesis. They do not, however, establish whether the reduction in secreted LPS in these strains is responsible for the failure to induce morphogenesis. They also do not establish whether the alleged outer membrane vesicles at the tips of wild type flagella, observed in electron micrographs, require flagellar motility for their production.

The reviewer’s comment indicates a desire to directly relate the amount of LPS released by colonizing bacteria to an amount of added purified LPS sufficient to induce morphogenesis (in the absence of bacteria). However, as mentioned in our first response, above, as well as in new text, working with LPS is far more complicated than typical solution chemistry. The addition of exogenous LPS assumes no context-dependency in terms of either bacterial proximity (e.g., the binding of V. fischeri to the host cells/receptors) or the form of LPS (micelles, vesicles, fragment size, etc.) presented to the host. As we mention in the Discussion, we expect that the manner in which LPS is presented is an important part of the way it is sensed by the host cells, and that culture supernatants contain LPS released not only by flagellar rotation, but also from other sources like outer membrane vesicles, cell lysis, etc, that have accumulated during growth. In the context of symbiotic initiation, during which there is no V. fischeri proliferation within the timeframe of our experiments, such other sources are less likely to be relevant to the induction of early-stage apoptosis. Taken together, we believe that the proposed experiment would be difficult to interpret, problematic to quantify, and provide little biologically relevant insight into the host responses that are naturally triggered by colonizing V. fischeri cells.

1) A key basis for this paper is the assertion that LPS acts as the signal for the symbiont-induced morphogenesis of the squid host. The data in Figure 1–figure supplement 5 showing comparable apoptosis per ciliated field producing by lipid A exposure and V. fischeri colonization is different from the results reported in a previous paper from this group, by Koropatnick et al. 2004 (Figure 2H), in which LPS has a modest effect inducing apoptosis and interacts synergistically with peptidoglycan to induce this effect. It's possible that lipid A is more potent than crude LPS. However, the real difference in these two sets of results is the extent of apoptosis induced by the symbiont, with significantly lower numbers in this manuscript's figure. The authors should comment on the differences between these two sets of results.

We thank the reviewer for their thorough reading of past literature in considering this manuscript. As described in an earlier response (combined-review comment #2), the different numbers of apoptotic nuclei reported in the two works is because they are examining the tissues at two different time points during symbiotic colonization. In Koropatnick et al., the authors were examining a later time point (24 h post-inoculation), by which time the symbionts have colonized the light organ’s deep crypts and undergone many rounds of doubling. In the work presented here, we are examining the induction of early-stage apoptosis that occurs when non-growing V. fischeri cells are still initiating the symbiosis (10 h post-inoculation), well before they reach the deep crypts, where subsequent proliferation occurs. While inoculated animals have significantly higher apoptosis levels than uninoculated ones at both 10 and 24 h, this developmental (and experimental) difference accounts for the higher absolute numbers of apoptotic nuclei observed in the Koropatnick et al. study.

2) Based on the Koropatnick et al. 2004 results showing synergistic induction of apoptosis with combined LPS and PGN, it might be possible that the lack of morphogenesis induced by the flrA and motB mutants is due to the reduction of multiple morphogenic signals, as opposed to just LPS. For example, rotation of the flagellar shaft at the point of connection with the cellular membrane could enhance release of PGN. The authors should test whether levels of peptidoglycan and/or peptidoglycan monomers are reduced in the supernatant of the flrA, motB, and motX mutants relative to wild type cells. Additionally, the flagellin monomers that the authors state are secreted by wild type V. fischeri may act in conjunction with LPS to stimulate morphogenesis and these may well be reduced in the mutants.

We have addressed the PGN/TCT issue in the response to the combined review, above. As for the cellular location of LPS shedding, we completely agree that another possible source of the released LPS could be the interface at the base of the flagellum, between the flagellar sheath and the static outer membrane. Currently, though, there exists no evidence to support this hypothesis. It would clearly be an important next step to ask whether there are additional molecules (such as flagellin) that might be differentially shed between these strains, and what their activity might be (synergistically or otherwise). However, in this manuscript we have described a new role for the rotation of the sheathed flagellum, and believe that testing all other candidate surface components for their presence and possible synergy will take considerable effort, and is beyond the scope of this initial announcement and characterization of an important discovery.

3) The set of experiments that would convincingly show that the reduction in secreted LPS in the flrA and motB mutants is responsible for their failure to induce apoptosis would be to 1) show that secreted fractions from the wild type but not the mutant cells can induce apopotosis, and 2) show that depletion of LPS from the wild type cells (for example with the antibody used in Koropatnick et al.) eliminated that apoptosis-inducing activity in the secreted fraction from the wild type cells. These may be technically challenging experiments, but the authors need to provide more data on the connection between what these cells secrete and the functional capacities of these secretions in the symbiosis.

We are unable to perform the first suggested experiment because, as mentioned above, we can’t add culture supernatants to seawater because they are toxic to the animals; too many other metabolic waste products accumulate during the culture’s growth. One could argue that we can simply isolate the LPS from those supernatants, and add it to the animal’s seawater, but that is essentially what we have done already with the addition of purified LPS. Also, as noted earlier, choosing what amount of this material to add would be arbitrary, and impossible to quantitatively relate to the natural process, which is initiated by a few non-growing cells attached to host tissue in a active flow field.

For the same reason we can’t remove the LPS from these supernatants with LPS antibody, as the toxic products will still be present. In addition, our previous published studies show that, while treating the seawater with antibody removes LPS, such an addition can not be made during a colonization experiment, presumably because the antibody binds the V. fischeri cells (including the flagellar sheath) and interferes with the bacterium’s surface association with the host cilia. This is not even a specific effect: as a similar blocking of colonization occurs with the addition of an antibody to OmpU, the bacterium’s major outer membrane porin (Aeckersburg et al., 2001).

4) The authors discuss the issue of the threshhold concentrations of LPS to induce apoptosis, but they should provide more information on the functional and observed concentrations of these molecules. The experiment shown in Figure 1–figure supplement 5 uses 1 ug/ml lipid A. How does this compare to the concentrations of lipid A present in the LPS released by the flrA, motB, and motX mutants?

As we described in response to a previous comment, the concept of a ‘concentration’ of LPS present in a culture is problematic, and implies that all the possible forms LPS comes in have the same specific activities for a given response. Our purpose in mentioning a threshold LPS level is only to suggest that a relatively small difference in LPS release might trigger a significant host response. Because the light organ’s surface epithelium is constantly exposed to seawater, it experiences the low levels of LPS generated by the presence of the naturally occurring bacterioplankton; thus, these host cells must differentiate between an ‘authentic’ indicator of the presence of the symbiont (∼103 cells/ml), and this background ‘noise’ of the normal bacterioplankton (>106 cells/ml). Such a dilemma suggests that the signaling pathways are finely tuned to the additional LPS presented by the rotating flagella of the 5-10 V. fischeri cells attached directly to the host tissue (Altura et al). Thus, it is hard to estimate the amount of LPS released by individual cells during symbiotic initiation. As previously mentioned, the few V. fischeri cells initiating the response are not growing during the early hours of the association examined here, and the dilution effect of the flow field around the cilia to which they are attached is a mystery; thus, even back-of-the envelope calculations based on LPS release in culture require substantial, unsupported, assumptions be made. In any case, we agree that this is an important point to clarify for the reader, and have added to the text to describe it.

The data (Figure 1–figure supplement 4) the reviewer mentions serve merely as a control for our experiments, confirming that, as in Foster et al., 2000, lipid A is sufficient to induce early-stage apoptosis under the conditions used in our work. Given that this experiment is only a methodological control, and not critical to our argument, we are willing to remove these data if the editor feels they distract the reader from the overall message of the manuscript.

5) Based on the characterization of mutants with non-motile but intact flagella, the authors postulate that the biophysics of flagellar rotation result in LPS release. Their one piece of microscopic evidence for this is the electron micrograph in Figure 2A, allegedly showing vesicular release from flagellar tips. The authors should provide more analysis of this phenomenon. First are the alleged vesicles comparable in size to outer membrane vesicles? Second, at what frequency are they observed? Finally, if they are the result of flagellar rotation, then they should not be observed in preparations of cells with amotile flagella. Is this the case?

We can see that our original wording was not clear, and we have clarified our intentions in providing this image. Rather than suggesting that these vesicle-like structures are the definitive source of shed LPS (which, as the reviewer points out, would require answering a number of other descriptive and experimental questions), we intended these data to support the notion that structural weaknesses can appear between the flagellar sheath and shaft, analogous to the weaknesses between the outer membrane and peptidoglycan layer required for cellular OMV formation. Further, their location is consistent with similar structures observed in the sheathed flagella of Helicobacter pylori (Geis et al.; Josenhans et al.).

We have not performed quantitative analyses of vesicle formation from EMs because one could argue that it would be (i) less likely to see them on the non-rotating mutants, as they would not form as rapidly, or (ii) more likely to see them, as, once formed, they would not be as easily shed. Instead, we feel the most rigorous and quantitative indication of LPS shedding is the measurement of LPS in culture supernatants we performed. Thus, we have now made it clear (line 115) that the illustration of vesicle-like structures associated with the flagellum is provided simply to indicate a possible sheath-related source of the LPS appearing in the supernatant.

As discussed later, it is quite possible that the release of LPS more typically occurs elsewhere, such as at the base of the flagellum. Parsing out the location(s) of LPS release, and their chemical form(s), is the subject of another ongoing study requiring the development of new techniques that are beyond the scope of this paper. Here, our goal is to provide new data on a subject that has made very little advancement over the past 30 years, and to share our findings at this stage of our studies so as to encourage a reopening of flagellar-sheath research by investigators using different bacterial species, any one of which may prove to be a key model for future discoveries.

Reviewer #3:

The manuscript by Brennan et al. investigates the sheathed flagellum of Vibrio fischeri. The reason for the sheath, an extension of the outermembrane, covering the filament has remained mysterious. The authors suggest that it may be provoking an immune response to LPS. They showed that LPS is shed into the environment when the flagellum is rotating, and that bacteria that cannot rotate their flagellum are defective at inducing apoptosis in host squid cells. They additionally showed that a flagellated but non-motile V. cholerae strain also releases less LPS into the environment and suggest that this might be conserved among all bacteria with sheathed flagella. In general the observation is very interesting and may point at a reason for the existence of the sheath surrounding the flagellar filament. Specific comments on the manuscript:

1) One piece was missing from their story: they showed that flagellated non-motile bacteria are defective at inducing apoptosis, and that these bacteria shed less LPS, but they didn’t show that the supernatants from these bacteria (vs the wildtype supe) are defective at inducing apoptosis. Given their argument that the shed LPS is what is stimulating apoptosis, they should show that the supes from the bacteria do/don’t induce apoptosis in the squid cells, rather than using purified LPS. They should do the same with the V. cholerae supes as well, since the source of the LPS shouldn’t make much difference.

As described above, addition of bacterial culture supernatants to live animals is not a viable approach.

2) Another control should be included, which is the addition of phenamil, a specific inhibitor of the sodium-driven motor, to the wildtype cells-this will stop flagellar rotation, and if their hypothesis is correct, there will be less LPS in the supe, which will stimulate less apoptosis.

We appreciate and agree with the logic underlying this reviewer’s experiment and, not surprisingly, we had already examined the effect of phenamil on LPS release in these strains, producing the data shown here (see Author response image 1). In comparing cell-free culture supernatants exposed to either DMSO (the solvent control) or phenamil, we observed that, upon addition of phenamil, the motB1 mutant no longer exhibited reduced LPS shedding as compared to wild-type cells similarly exposed to phenamil, supporting the conclusion that both chemical disruption of the sodium-motor pump and flagellar rotation serve the same functions in LPS shedding. However, the overall level of LPS in wild-type supernatants was not reduced after exposure to phenamil, but, rather, it trends toward being slightly higher for both strains, compared to the DMSO controls.

Author response image 1.

Author response image 1.

Open in a new tab

A possible explanation for this difference became apparent when we prepared the cells for this experiment; phenamil addition delayed the growth of V. fischeri, even at concentrations at which cells are still motile when observed by microscopy. Thus, addition of phenamil to the culture supernatants does not specifically interrupt flagellar rotation, as hoped, but rather has collateral effects on V. fischeri physiology and/or stress response. For instance, as a marine organism, V. fischeri uses sodium pumps not only for its flagellar motor, but also for nutrient transport via sodium symporters. With this caveat in mind, these data could be interpreted as supporting our hypothesis. However, we cannot rule out that these indirect effects of phenamil exposure might mask other differences. While we did not include these phenamil data in the final publication, because we were concerned they might be misinterpreted by the reader, we are certainly willing to do so at the request of the editor.

Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd

RESOURCES