Polyester synthesis genes associated with stress resistance are involved in an insect–bacterium symbiosis

Significance

This study reports a previously unrecognized involvement of polyhydroxyalkanoate (PHA), known as a bacterial endocellular storage polymer, in an insect–bacterium symbiosis. Many bacteria in the environment accumulate PHA granules within their cells, which provide resistance to nutritional depletion and other environmental stresses. Here we demonstrate that synthesis and accumulation of PHA in the symbiont cells are required for normal symbiotic association with, and, consequently, positive fitness effects for the host insect. The requirement of PHA for symbiosis suggests that, contrary to the general expectation, the within-host environment may be, at least in some aspects, stressful for the symbiotic bacteria.

Abstract

Many bacteria accumulate granules of polyhydroxyalkanoate (PHA) within their cells, which confer resistance to nutritional depletion and other environmental stresses. Here, we report an unexpected involvement of the bacterial endocellular storage polymer, PHA, in an insect–bacterium symbiotic association. The bean bug Riptortus pedestris harbors a beneficial and specific gut symbiont of the β-proteobacterial genus Burkholderia, which is orally acquired by host nymphs from the environment every generation and easily cultivable and genetically manipulatable. Biochemical and cytological comparisons between symbiotic and cultured Burkholderia detected more PHA granules consisting of poly-3-hydroxybutyrate and associated phasin (PhaP) protein in the symbiotic Burkholderia. Among major PHA synthesis genes, phaB and phaC were disrupted by homologous recombination together with the phaP gene, whereby ΔphaB, ΔphaC, and ΔphaP mutants were generated. Both in culture and in symbiosis, accumulation of PHA granules was strongly suppressed in ΔphaB and ΔphaC, but only moderately in ΔphaP. In symbiosis, the host insects infected with ΔphaB and ΔphaC exhibited significantly lower symbiont densities and smaller body sizes. These deficient phenotypes associated with ΔphaB and ΔphaC were restored by complementation of the mutants with plasmids encoding a functional phaB/phaC gene. Retention analysis of the plasmids revealed positive selection acting on the functional phaB/phaC in symbiosis. These results indicate that the PHA synthesis genes of the Burkholderia symbiont are required for normal symbiotic association with the Riptortus host. In vitro culturing analyses confirmed vulnerability of the PHA gene mutants to environmental stresses, suggesting that PHA may play a role in resisting stress under symbiotic conditions.

Insects account for the majority of biodiversity described thus far (1), and many of them possess symbiotic bacteria within their gut, tissues, and cells (2). These microbial associates are usually transmitted from mothers to offspring vertically and exert various biological effects on their host insects: some play essential roles like provisioning of essential nutrients (3); others have conditionally beneficial roles such as defense against natural enemies and adaptation to specific ecological conditions (4); and others induce parasitic or pathogenic consequences including negative fitness effects and reproductive aberrations (5). The majority of these symbionts are difficult to culture in vitro, probably because they are highly adapted to the unique environment within their host insects (6, 7). Consequently, biological roles of the symbionts were investigated mainly using experimental and physiological approaches to the whole host–symbiont systems (8). Recently, symbiont genomics has provided powerful culture-independent approaches to understanding the whole picture of candidate genes and mechanisms involved in host–symbiont interactions (3, 9). However, the fastidious nature and the consequent genetic intractability of the symbionts are still the major obstacle to understanding of the mechanisms underlying most insect–microbe symbiotic associations.

On the other hand, among aquatic invertebrates and land plants, it is commonly observed that beneficial symbiotic bacteria are acquired by their hosts from the environment every generation, as in squid–Vibrio and legume–Rhizobium symbiotic associations (10, 11). By making use of these model systems wherein the symbionts are easily cultivable and genetically manipulatable, a variety of symbiont genes involved in initiation, accommodation, and persistence in their hosts, and also symbiont genes involved in their beneficial functions, have been identified (12–15). Such model symbiotic systems also have been desired for insect symbiosis studies.

Recently, the bean bug Riptortus pedestris (Fig. 1A) has emerged as an exceptional insect–bacterium symbiotic system. This insect harbors a beneficial and specific symbiont of the betaproteobacterial genus Burkholderia in a specialized region of the posterior midgut (Fig. 1 B–D) (16). The symbiont is orally acquired by host nymphs from the environment every generation and easily cultivable and genetically manipulatable (17, 18). In this study, by using this model system, we report a previously unknown involvement of a bacterial endocellular storage polymer, polyhydroxyalkanoate (PHA) in the Riptortus–Burkholderia symbiotic association. We found that the PHA synthetic capability of the symbiont is important for sustaining infection levels of the symbiont and consequent fitness effects to the host insect, unveiling an unexpected relevance of bacterial stress resistance to the symbiotic association.

Fig. 1.

Riptortus–Burkholderia gut symbiosis. (A) Adult insect of R. pedestris. (B) Dissected midgut of R. pedestris with the morphologically distinct regions M1, M2, M3, M4B, and M4. (C) Closer look of M4 with numerous crypts aligned in two rows. (D) Transmission electron microscopic image of the cross-sectioned crypts. “C” indicates crypt epithelia, whereas “B” shows crypt cavities filled with the Burkholderia cells. Yellow lines highlight boundaries of the crypt epithelia and the crypt cavities.

Results and Discussion

Accumulation of PHA Granules Within Symbiont Cells.

In an attempt to identify symbiosis-related molecules in the Riptortus–Burkholderia association, symbiotic Burkholderia cells and cultured Burkholderia cells, which represented the same wild-type Burkholderia strain RPE75 (19), were subjected to SDS/PAGE. The overall protein profiles were similar between symbiotic Burkholderia and cultured Burkholderia, except for a 19-kDa protein remarkably more abundant in symbiotic Burkholderia than in cultured Burkholderia (Fig. 2A). The 19-kDa protein was identified as phasin (PhaP)—which is known as a PHA granule-associated protein—by in-gel trypsin digestion followed by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) (Fig. 3).

Fig. 2.

Accumulation of PHA granules in symbiotic Burkholderia. (A) Protein profiles of symbiotic Burkholderia and cultured Burkholderia in YG medium analyzed by SDS/PAGE. Arrowhead indicates the band of 19-kDa PhaP, a PHA granule-associated protein. (B and C) Transmission electron microscopic images of symbiotic Burkholderia cells (B) and cultured Burkholderia cells (C). Red arrowheads indicate PHA granules. (D and E) Phase-contrast images (Left) and fluorescent images (Right) of symbiotic Burkholderia cells (D) and cultured Burkholderia cells (E), wherein PHA granules are stained with Nile Blue. (F) Flow cytometric histograms of PHA-derived fluorescence from symbiotic Burkholderia cells (gray) and cultured Burkholderia cells (white). Values on the graph are mean fluorescence intensity ± SD.

Fig. 3.

Identification of PhaP protein of the Burkholderia symbiont by LC-MS/MS. (A) Amino acid sequence of PhaP protein deduced from the draft Burkholderia genome sequence. Peptide sequences matched to the LC-MS/MS results are highlighted in red. (B) Identified peptides with ion scores. (C) MS/MS spectrum of the peptide located at amino acid positions 61–79.

PHAs are linear polyesters produced by many bacteria via fermentation of sugars and lipids (Fig. 4A). PHAs are accumulated as discrete granules within bacterial cells, which may sometimes reach levels as high as 90% of the cell dry weight and function as carbon and energy storage for the bacteria (20). Once PHAs are extracted from the bacterial cells, these macromolecules show material properties similar to some common plastics like polypropylene, and hence they have drawn much attention as a biodegradable substitute for synthetic plastics in biotechnology fields (20–23). On the ground that the major part of PHA granule surface is covered with PhaP (24, 25), we suspected that PHA granules may be preferentially accumulated in symbiotic Burkholderia cells. Transmission electron microscopy detected electron-translucent presumable PHA granules in both symbiotic and cultured Burkholderia cells. The size of the granules was apparently larger in the symbiotic Burkholderia cells than in the cultured Burkholderia cells (Fig. 2 B and C). PHAs can be stained with Nile Blue fluorescent dye (26). Fluorescence microscopy detected presumable PHA-derived fluorescence from both symbiotic and cultured Burkholderia cells (Fig. 2 D and E). Flow cytometry demonstrated that symbiotic Burkholderia cells emitted significantly higher PHA-derived fluorescence than cultured Burkholderia cells (P < 0.05, z-test) (Fig. 2F). We extracted PHA fraction from cultured Burkholderia cells and hydrolyzed it, and the resultant monomer fraction was analyzed by gas chromatography coupled with mass spectrometry, thereby identifying the Burkholderia-derived PHA as poly-3-hydroxybutyrate (Fig. S1 A–C). Using different carbon sources, the cultured Burkholderia cells consistently produced 100% poly-3-hydroxybutyrate as the sole PHA component (Fig. S1D).

Fig. 4.

PHA and its synthetic genes. (A) Structure of poly-3-hydroxybutyrate, the sole PHA produced by the Burkholderia symbiont strain RPE75. (B) Main steps of PHA synthesis from acetyl-CoA. PhaA, β-ketothiolase; PhaB, NADPH-dependent acetoacetyl-CoA reductase; PhaC, PHA synthase. (C) Gene arrangements of PHA-related genes on the Burkholderia symbiont genome. Note that phaP is located on a contig distinct from the contig encoding phaC and phaB. Deleted regions in each of the mutants are indicated on the scheme as spaces. phaP and phaR indicate the PHA-associated protein gene and the PHA-regulatory protein gene, respectively. The nucleotide sequences have been deposited in the DNA Data Base in Japan/European Molecular Biology Laboratory/GenBank nucleotide sequence databases (accession nos. AB787502 for phaP, AB787503 for phaB, and AB787504 for phaC).

Generation of Symbiont Mutants Deficient in phaB, phaC, and phaP Genes Involved in PHA Granule Formation.

In bacterial cells, PHAs are synthesized mainly from acetyl-CoA via three steps of enzymatic reactions catalyzed by ketothiolase or PhaA, acetyl-CoA reductase or PhaB, and PHA synthase or PhaC (Fig. 4B) (20, 23). The draft genome sequence of the symbiotic Burkholderia strain RPE75 contained two copies of phaA genes plus 10 similar genes with 40–50% amino acid sequence identity, a copy of the phaB gene plus two similar genes with 40–55% amino acid sequence identity and a single copy of the phaC gene without related genes (TBLASTN E-value < 0.0001). The genome data contained six copies of phaP genes, and one of them agreed with the amino acid sequence of PhaP obtained by SDS/PAGE and LC-MS/MS (Fig. 3). Considering the redundancy in phaA genes, we decided to generate Burkholderia mutants deficient in phaB, phaC, and phaP genes. Using allelic exchange by homologous recombination, we established ΔphaB, ΔphaC, and ΔphaP mutants from the wild-type Burkholderia strain RPE75 (Fig. 4C). In general, PHA production is facilitated when nutrient supplies are imbalanced because storing excess nutrients endocellularly is advantageous for the bacteria under such conditions. By polymerizing soluble intermediates into insoluble molecules, the bacterial cells are able to prevent leakage of valuable compounds without osmotic problems (20, 23). Accordingly, flow cytometric analysis of the wild-type Burkholderia cells revealed that bacterial cells cultured in nutritionally poor PB-G medium (phosphate buffer with 1% glucose) emitted drastically higher levels of PHA-derived fluorescence than bacterial cells cultured in nutritionally rich yeast glucose (YG) medium (Fig. 5A). By contrast, the Burkholderia mutants ΔphaP, ΔphaB, and ΔphaC exhibited relatively lower levels of PHA-derived fluorescence even when cultured in nutritionally poor PB-G medium (Fig. 5 B–D). The estimated levels of PHA accumulation in the bacterial cells were on the order of wild type > ΔphaP > ΔphaB > ΔphaC under the nutritionally poor condition (Fig. 5 A–D). The mutant ΔphaP seemed to show some heterogeneity among the bacterial cells, judging from the lower peak and the broader variance of the flow cytometric histogram (Fig. 5B). Fluorescence microscopy of the wild-type and mutant Burkholderia cells confirmed these patterns observed with flow cytometry, including heterogeneous PHA accumulation among ΔphaP mutant cells (Fig. 5 E–H). We also generated the complemented Burkholderia mutant strains ΔphaP/phaP, ΔphaB/phaB, and ΔphaC/phaC by transforming the mutant bacteria with plasmids containing functional phaP, phaB, and phaC, respectively. Fluorescence microscopy of the complemented Burkholderia mutants confirmed restoration of PHA accumulation within the bacterial cells (Fig. 5 I–L).

Fig. 5.

PHA production by wild-type and PHA-related gene mutants of the Burkholderia symbiont in vitro. (A–D) Flow cytometric histograms of PHA-derived fluorescence from the Burkholderia cells cultured in PB-G (gray) and the Burkholderia cells cultured in YG medium (white): (A) wild type, (B) ΔphaP, (C) ΔphaB, and (D) ΔphaC. Values on the histograms are mean fluorescence intensity ± SD. (E–H) Phase-contrast images (Left) and fluorescent images (Right) of the Burkholderia cells cultured in PB-G medium wherein PHA granules are stained with Nile Blue: (E) wild type, (F) ΔphaP, (G) ΔphaB, and (H) ΔphaC. (I–L) Phase-contrast images (Left) and fluorescent images (Right) of the plasmid-transfected Burkholderia cells cultured in PB-G medium wherein PHA granules are stained with Nile Blue: (I) wild type transfected with blank plasmid, (J) ΔphaP complemented with phaP-containing plasmid, (K) ΔphaB complemented with phaB-containing plasmid, and (L) ΔphaC complemented with phaC-containing plasmid.

Symbiont Mutants Deficient in PHA Synthetic Genes Show Symbiosis-Deficient Phenotypes.

Cultured Burkholderia cells of the mutants ΔphaP, ΔphaB, and ΔphaC as well as the wild-type strain were suspended in water and orally administrated to newly molted Riptortus nymphs at the second instar, the competent stage for natural establishment of the gut symbiotic association (18). These insects were reared to the fifth instar and dissected, and their symbiotic organ, the midgut fourth region (see M4 in Fig. 1B), was inspected. The wild-type strain established normal colonization in the symbiotic organ: numerous crypts arranged in two rows were well developed, whitish in color, and full of symbiotic bacteria (Fig. 6A). The ΔphaP mutant also established normal colonization to the midgut crypts, although development of the crypts was less conspicuous in comparison with the wild-type strain (Fig. 6B). On the other hand, the midgut crypts of fifth-instar nymphs infected with the ΔphaB and ΔphaC mutants looked different: the crypts were thin and translucent (Fig. 6 C and D), which is somewhat reminiscent of the atrophied midgut crypts of uninfected insects (Fig. 6E). When the dissected midgut samples were homogenized and cultured on YG agar plates, the colony count data were in accordance with the histological observations: the wild-type strain and the ΔphaP mutant yielded bacterial titers around 10⁷ colony formation units (cfu) per insect, whereas the ΔphaB and ΔphaC mutants exhibited about 5 × 10⁶ cfu per insect (Fig. 6I). Transmission electron microscopy of the midgut crypts revealed striking differences not only in bacterial densities but also in PHA granules in the bacterial cells. The wild-type strain densely populated the crypt cavity, whose cells contained a number of PHA granules (Fig. 6J). The ΔphaP mutant similarly exhibited dense infection in the crypt cavity and bacterial cells with intracellular PHA granules (Fig. 6K). By contrast, the ΔphaB and ΔphaC mutants did not fill up the crypt cavity, and PHA granules were lacking in their cells (Fig. 6 L and M). Cultured Burkholderia cells of the complemented mutants ΔphaP/phaP, ΔphaB/phaB, and ΔphaC/phaC were similarly administrated to second-instar Riptortus nymphs. When the infected insects were inspected at the fifth instar, the deficient phenotypes associated with the ΔphaB and ΔphaC mutants were restored: the midgut crypts were well developed (Fig. 6 G and H), and the symbiont titers were recovered to the levels closer to the wild-type level (Fig. 6I).

Fig. 6.

Infection density and morphology of the wild-type and PHA-related gene mutants of the Burkholderia symbiont in the midgut symbiotic organ of the Riptortus host. (A–H) Midgut M4 region with crypts dissected from fifth-instar nymphs that had been inoculated with the following Burkholderia symbiont strains at the second instar: (A) wild type, (B) ΔphaP, (C) ΔphaB, (D) ΔphaC, (E) uninfected, (F) ΔphaP/phaP, (G) ΔphaB/phaB, and (H) ΔphaC/phaC. (Scale bars, 0.2 mm.) (I) Symbiont titers in the midgut M4 region of fifth-instar nymphs. Means and SE of cfu per insect are shown. Different letters (a, b) on the top of the columns indicate statistically significant differences (P < 0.05; Student t test with Bonferroni correction). (J–M) Transmission electron microscopic images of the Burkholderia symbiont strains in the midgut M4 region of fifth-instar nymphs: (J) wild type, (K) ΔphaP, (L) ΔphaB, and (M) ΔphaC. (Insets) Enlarged images of the symbiont cells.

Host Fitness Is Negatively Affected by Infection with Symbiont Mutants Deficient in PHA Synthesis Genes.

We measured fitness parameters of host insects infected with the wild-type strain, ΔphaP mutant, ΔphaB mutant, ΔphaC mutant, or no symbiont. Whereas survival to adulthood was not different among the wild-type–infected, mutant-infected, and uninfected insects (Fig. 7A), time to adulthood was different among them on the order of wild type > ΔphaP > ΔphaB > ΔphaC > no symbiont (Fig. 7B). Here note that a previous study also showed that survival to adulthood is not different among the insects infected with the wild-type symbiont and the insects with no symbiont (17). Body length of adult males was on the order of wild type > ΔphaP > ΔphaB > ΔphaC > no symbiont, wherein the difference between wild type and ΔphaC was statistically significant (Fig. 7C). Body weight of adult males exhibited no remarkable differences among the symbiont genotypes (Fig. 7D). Body length of adult females was on the order of wild type > ΔphaP > ΔphaB > ΔphaC > no symbiont, wherein the difference between wild type and ΔphaC was statistically significant (Fig. 7E). Body weight of adult females was on the order of wild type > ΔphaP > ΔphaB = ΔphaC > no symbiont, wherein the differences between wild type and ΔphaB or ΔphaC were statistically significant (Fig. 7F). In summary, these results showed that the host insects infected with the ΔphaB or ΔphaC mutants tended to exhibit lower fitness values than the host insects infected with the wild-type strain. It was also shown that the extent of fitness decline tended to be more conspicuous in female insects than in male insects, which may be relevant to—although this is speculative—the fact that female insects require more energy resources for egg production. We also measured fitness parameters of host insects infected with the complemented Burkholderia mutants ΔphaP/phaP, ΔphaB/phaB, or ΔphaC/phaC. The declined fitness values observed with the host insects infected with the ΔphaB and ΔphaC mutants were mostly recovered in the host insects infected with the ΔphaB/phaB and ΔphaC/phaC complemented mutants (Fig. 7 B–F).

Fig. 7.

Effects of infection with the wild-type and PHA-related gene mutants of the Burkholderia symbiont on fitness parameters of the Riptortus host. (A) Survival rate. (B) Adult emergence. Red (ΔphaC-infected group) and black (uninfected group) asterisks (*) indicate significant difference in comparison with the wild-type–infected group (P < 0.05; z-test with Bonferroni correction). (C) Body length of adult males. (D) Body weight of adult males. (E) Body length of adult females. (F) Body weight of adult females. In C–F, different letters (a, b, c) on the top of the columns indicate statistically significant differences (P < 0.05; one-way ANOVA with Tukey correction). Means and SEs are shown, and sample sizes are indicated on the columns.

Plasmid Retention Rates.

We examined plasmid retention rates of the complemented Burkholderia strains ΔphaP/phaP, ΔphaB/phaB and ΔphaC/phaC, as well as wild-type/vector, which had been administrated to the host insects at the second instar and were inspected at the fifth instar. Although the ΔphaP/phaP and wild-type/vector strains had lost their plasmids, strikingly, the ΔphaB/phaB and ΔphaC/phaC strains exhibited around 90% plasmid retention rates (Table 1), indicating the importance of phaB and phaC genes for persistent infection and/or proliferation of the symbiont within the host midgut.

Table 1.

Plasmid retention rates in the complemented Burkholderia strains during infection of the Riptortus host

Symbiont strain	Plasmid detection*	Retention rate, %†
Wild type/vector‡	0/20, 0/20, 0/20	0
ΔphaP/phaP	0/20, 0/20, 0/20	0
ΔphaB/phaB	17/20, 18/20, 18/20	88.3 ± 2.9
ΔphaC/phaC	18/20, 18/20, 20/20	93.3 ± 5.8

^*Number of colonies with plasmid/number of colonies examined per insect.

^†Mean ± SD.

^‡Wild-type Burkholderia strain transfected with blank vector.

PHA Synthesis Genes phaB and phaC Are Required for Normal Riptortus–Burkholderia Gut Symbiosis.

On the basis of these results, we conclude that the PHA synthesis genes phaB and phaC of the Burkholderia symbiont are required for normal symbiotic association with the Riptortus host. This finding unveils a previously unrecognized involvement of the endocellular bacterial biopolymer production in the insect–bacterium symbiotic association.

PHA Granule-Associated Protein Gene phaP in Riptortus–Burkholderia Gut Symbiosis.

On the other hand, the results with the ΔphaP mutant were somewhat puzzling. Compared with the other mutants, ΔphaP mutant exhibited less impact on symbiosis in vivo. We speculate that this might be due to the redundancy of phasin proteins. The draft genome sequence of the Burkholderia strain RPE75 contains six phaP genes, and the ΔphaP mutant represents only one of them. Another interesting aspect of the ΔphaP mutant is that only some of the ΔphaP mutant cells highly accumulated PHA granules in vitro (Fig. 5F), whereas all ΔphaP mutant cells looked like they had accumulated PHA granules in vivo (Fig. 6K). The differences between PHA granule accumulation in phaP mutant in vitro (Fig. 5F) and in vivo (Fig. 6K) correlate with the reduced stress resistance in vitro (Fig. 8), but better bacterial survival (Fig. 6I) and host insect fitness in vivo (Fig. 7). It seems that the phaP gene product may behave differently under the symbiotic condition and the cultured condition. Further studies are needed to address the biological roles of the phaP gene(s) in the Riptortus–Burkholderia symbiosis.

Fig. 8.

Growth and survival of the wild-type and PHA-related gene mutants of the Burkholderia symbionts under stressful culturing conditions. (A) Growth curves in nutrition-rich YG medium. (B) Growth curves in nutrition-poor minimal medium. (C) Survival after nutritional depletion in phosphate buffer for 2 d. (D) Survival after heat stress in 45 °C water bath for 10 min. (E) Survival after osmotic stress in 1 M glucose for 24 h. (F) Survival after oxidative stress in 0.5% or 2% hydrogen peroxide for 2 d. In C–F, different letters (a, b, c, d) on the top of the columns indicate statistically significant differences (P < 0.05; one-way ANOVA with Tukey correction). Means and SDs are shown.

Relevance of PHA Synthesis to Bacterial Ecology and Legume–Rhizobium Symbiosis.

Previous studies of other symbiotic systems, such as legume–Rhizobium nitrogen-fixing symbiosis and squid–Vibrio luminescent symbiosis, have highlighted the important biological roles of extracellular bacterial biopolymers, or biofilms, in the establishment and maintenance of the symbiotic associations (27, 28). However, PHAs are generally accumulated within the bacterial cells as polymer granules and are normally not excreted outside unless the cells are lysed (20, 23). How, then, can endocellular bacterial biopolymers like PHAs contribute to an insect–bacterium symbiotic association?

In diverse bacteria, PHAs play important ecological roles as internal storage polymers, thereby improving their survival during starvation (20). Probably relevant to the storage function, it has been shown that PHA accumulation improves bacterial tolerance to high temperature, exposure to reactive oxygen species, osmotic stress, UV irradiation, desiccation, and other environmental stresses (29). In the symbiotic context, PHA granules have been identified in rhizobial cells. Some species belonging to Rhizobium, Bradyrhizobium, and Azorhizobium accumulate PHA granules both in free life and in symbiosis, whereas in other species, such as Sinorhizobium meliloti, accumulation of PHA granules is observed only in free life but not in nitrogen-fixing bacteroids (30). Although advantages of PHA accumulation in free-living rhizobia, such as better reproduction and survival under low-nutrient conditions, have been clearly demonstrated (31), biological roles of PHAs in symbiotic rhizobia have been elusive and controversial (32, 33). Some studies on rhizobial mutants deficient in PHA granule formation revealed that ability to synthesize and use PHA is not essential for symbiosis (30, 34), whereas other studies reported that rhizobial mutants unable to synthesize PHA are less competitive for growth and nodule occupancy (35, 36). Several studies demonstrated that, interestingly, rhizobial mutants deficient in PHA synthesis tend to exhibit elevated nitrogen-fixing activities and improved host fitness, probably because disrupted PHA synthesis allocates metabolic resources toward nitrogen fixation (30, 32). However, the metabolic antagonism between PHA synthesis and nitrogen fixation seems contradictory to the observations that some rhizobial bacteroids accumulate a large amount of PHAs (37). In these contexts, how these ecological aspects of PHAs in rhizobia and other bacteria are relevant to PHA synthesis in the Burkholderia symbiont is of great interest.

PHA Synthesis Genes Are Involved in Symbiont Resistance to Environmental Stresses Under Cultured Conditions.

To gain insights into the mechanisms of how PHA synthesis is involved in Riptortus–Burkholderia symbiosis, the ΔphaP, ΔphaB, and ΔphaC mutants together with the wild-type strain of the Burkholderia symbiont were examined for their growth characteristics in vitro under various environmental conditions. In nutritionally rich YG medium, all of the symbiont strains showed similar growth patterns, except that the ΔphaC mutant attained lower density at stationary phase (Fig. 8A). By contrast, in nutritionally minimal medium with 0.2% glucose as the sole carbon source, the ΔphaP, ΔphaB, and ΔphaC mutants exhibited remarkably slower growth than the wild-type strain (Fig. 8B). When the symbiont strains were cultured in phosphate buffer containing no carbon source, survival after 2 d was significantly lower in the ΔphaP, ΔphaB, and ΔphaC mutants than in the wild-type strain (Fig. 8C). When the symbiont strains were heat-treated at 45 °C for 10 min, the mutant strains exhibited significantly lower survival rates than the wild-type strain (Fig. 8D). When the symbiont strains were incubated with 1 M glucose, survival after 24 h was lower in the ΔphaP than in the wild-type strain, and, strikingly, the ΔphaB and ΔphaC mutants suffered nearly 80% mortality (Fig. 8E). These deficiencies associated with the ΔphaP, ΔphaB, and ΔphaC mutants were mostly restored in the ΔphaP/phaP, ΔphaB/phaB, and ΔphaC/phaC complemented mutants (Fig. 8 B–E). Meanwhile, when the symbiont strains were cultured in the presence of hydrogen peroxide, their cell proliferation was certainly inhibited in a dose-dependent manner, but the levels of inhibition showed no significant differences between the mutant and wild-type strains (Fig. 8F). These results indicated that the PHA synthesis genes may be important for resistance of the Burkholderia symbiont against such environmental stress factors as nutritional depletion, high osmotic pressure, and high temperature.

Possible Relationship of PHA Production to Environmental Stress in Riptortus–Burkholderia Symbiosis.

The biological roles of the Burkholderia symbiont for the Riptortus host are thought to be primarily nutrient provisioning for complementing a nutritionally imbalanced plant sap diet (17) and secondarily detoxification of detrimental chemicals (38). At present, there is no compelling evidence to support that PHA production is directly involved in these mutualistic functions of the Burkholderia symbiont, although the possibility that PHA metabolism may somehow affect the nutritional and detoxification activities of the symbiont cannot be ruled out. It seems more likely that PHA production is important for colonization and proliferation of the Burkholderia symbiont under the symbiotic condition on the grounds that (i) the ΔphaB and ΔphaC mutants exhibited significantly lower infection densities than the wild-type strain (Fig. 6); (ii) the ΔphaB/phaB and ΔphaC/phaC complemented mutants restored the infection densities (Fig. 6); and (iii) the phaB/phaC-encoding plasmids in the complemented mutant bacteria were selectively maintained in the host symbiotic organ, suggesting positive selection for the PHA synthesis genes (Table 1). Living inside the insect body may seem to imply a nutrient-rich, favorable condition for the symbiont. However, contrary to the general expectation, we point out the possibility that the symbiotic gut environment, at least in some aspects, may be restricted and hostile for the symbiont. An example may be the doubling time of the Burkholderia symbiont under different conditions. Second-instar Riptortus nymphs acquire about 10⁴ cells of the Burkholderia symbiont, fifth-instar nymphs harbor around 6 × 10⁷ symbiont cells in the midgut symbiotic organ, and it takes about 14 d for second-instar nymphs to become fifth-instar nymphs (18). Under the symbiotic condition, therefore, doubling time of the symbiont is estimated as about 27 h. Under cultured conditions, meanwhile, doubling time of the symbiont is estimated as about 4.5 h in a rich medium (Fig. 8A) and 11.3 h in a minimal medium (Fig. 8B). Hence, proliferation of the Burkholderia symbiont in the host symbiotic organ is restricted to a much slower rate than proliferation in the minimal medium. In the midgut crypts, the symbiont cells are densely packed (Fig. 1D and Fig. 6J), where they survive and proliferate over 2 mo of the host’s lifetime. Hence, it is conceivable, although speculative, that the symbiont cells within the symbiotic organ may suffer from restriction of some nutrients, oxygen, and/or other factors due to the high cell densities. Recent studies on diverse insect–microbe symbiotic associations have revealed that the host organisms regulate infection densities of their symbionts by means of phagocytosis, antimicrobial peptides, lysozymes, reactive oxygen species, etc. (39–42). These innate immune mechanisms may also comprise host-related stress factors for the symbionts. Considering our experimental data and these circumstances, we hypothesize (i) that PHA production contributes to the Burkholderia symbiont by ensuring better survival and proliferation under the symbiotic condition wherein the symbiont suffers nutritional and other stresses; (ii) that this is the reason why the ΔphaB and ΔphaC mutants exhibit lower infection densities than the wild-type symbiont strain (Fig. 6); and (iii) that the lower symbiont densities are the reason why the host insects infected with the ΔphaB and ΔphaC mutants suffer lower fitness values than the host insects harboring the wild-type strain (Fig. 7).

Conclusion and Perspective.

In this study, we identified an important biological role of the bacterial endocellular storage polymer, PHA, for maintenance of the Burkholderia–Ritortus gut symbiosis, unveiling a previously unrecognized molecular component required for some insect–bacterium symbiotic associations. This finding suggests the possibility that, contrary to the general expectation, the within-host environment may be stressful for the symbiotic bacteria. For identifying the nature of the “symbiotic stress,” direct measurements of microenvironmental parameters within the host midgut such as nutrient levels, osmotic pressure, oxygen concentration, redox status, levels of reactive oxygen species, etc., will be needed. We do not find PHA synthesis genes in the genomes of insect endosymbionts such as Buchnera of aphids and others, which are exclusively transmitted vertically and have lost any free-living ability. It seems plausible that the PHA synthetic pathway has been primarily selected for survival at the free-living stage of the Burkholderia symbiont and secondarily co-opted for survival under the symbiotic condition within the host, which may exemplify different evolutionary trajectories of vertically transmitted and environmentally acquired symbiotic bacteria. Under contemporary energy problems, the possibility of converting biomass into biofuels by microbial consortia, including gut symbiotic microbiota of termites, has been argued, despite its practical difficulties (43, 44). PHAs have been investigated as renewable biological plastic substitutes produced by microorganisms (20, 21). The unexpected finding of a biological connection of PHA synthesis to insect symbiosis highlights another intriguing interface between basic microbiology, symbiosis, and biotechnology.

Materials and Methods

Bacteria, Plasmids, and Media.

The bacteria strains and plasmids used in this study are listed in Table S1. Escherichia coli cells were cultured at 37 °C in LB medium [1% (wt/vol) tryptone, 0.5% (wt/vol) yeast extract, and 0.5% (wt/vol) NaCl], and Burkholderia cells were cultured at 26 °C in YG medium [0.5% (wt/vol) yeast extract, 0.4% (wt/vol) glucose, and 0.1% (wt/vol) NaCl] unless otherwise described. Antibiotic concentrations were as follows: 30 μg/mL for rifampicin, 50 μg/mL for kanamycin, and 30 μg/mL for chloramphenicol.

Generation of Deletion Mutant Strains.

Deletion of the chromosomal phaB, phaC, and phaP genes was accomplished by allelic exchange, following homologous recombination, using the suicide vector pK18mobsacB harboring the 5′ region and 3′ region of the gene of interest. The 5′ and 3′ regions of target genes were first amplified from the Burkholderia symbiont by PCR using the primers listed in Table S2. After digestion of the amplified PCR products and the pK18mobsacB vector with appropriate restriction enzymes, they were ligated and transformed into E. coli DH5α cells. The transformed E. coli cells were selected on LB–agar plates containing kanamycin. The positive colonies carrying vector with correct inserts were further selected by colony PCR using the 5′ region primers [indicated under “Primer name” column by “(gene name)-L-P1” in Table S2] and the vector primer aphII (5′-ATCCATCTTGTTCAATCATGCG-3′). These donor cells carrying pK18mobsacB containing the 5′ and 3′ regions of the gene of interest were then mixed with recipient Burkholderia RPE75 cells along with the helper cell HBL1 to transfer the cloned vector to the Burkholderia RPE75. After allowing the first crossover (single crossover) by culturing cell mixture of triparental conjugation on YG–agar plates, RPE75 cells with the first crossover were selected on YG–agar plates containing rifampicin and kanamycin. The positive colonies with the genomic integration of vector DNA were confirmed by PCR using the upstream 5′ region primers (phaP-up, phaB-up, or phaC-up; Table S3) and the vector primer aphII. The second crossover was allowed by culturing the cells with a single crossover in YG medium and then selecting on YG–agar plates containing rifampicin and sucrose (200 μg/mL). RPE75 with deletion of the gene of interest by the double crossover was identified by PCR using the upstream 5′ region primers (phaP-up, phaB-up, or phaC-up) and the downstream 3′ region primers (phaP-down, phaB-down, or phaC-R-P2; Tables S2 and S3). The scheme of molecular organization of the genes with deletion is shown in Fig. 4C.

Generation of Complemented Strains.

To complement deletion mutants, we used the broad host-range vector pBBR122 to clone phaP, phaB, and phaC genes. The blunt-end gene inserts were prepared by PCR using the primers listed in Table S2 (phaP-com-P1 and phaP-com-P2 for phaP, phaB-com-P1 and phaB-com-P2 for phaB, and phaC-com-P1 and phaC-com-P2 for phaC). The amplified DNA fragments were cloned into the DraI site of pBBR122 and transformed to E. coli DH5α cells. Using triparental conjugation with HBL1 and pBBR122 carrying the genes of interest, phaP, phaB, and phaC were transferred to the recipient Burkholderia mutant strains ΔphaP, ΔphaB, and ΔphaC, respectively. The complemented strains were selected on YG–agar plates containing rifampicin and kanamycin.

SDS/PAGE of Proteins from Symbiotic and Cultured Burkholderia Cells.

Symbiotic Burkholderia cells were freshly collected from the midgut of fifth-instar Riptortus nymphs. The midgut M4 regions were collected by dissection, cut by fine scissors in a 1.5-mL centrifuge tube containing 50 μL of PB (10 mM sodium phosphate, pH 7), suspended by pipetting with 1 mL of PB, and filtered through a 5-μm pore mesh. The bacterial cells in the filtrate were washed with PB by centrifugation and resuspension and adjusted to 10⁸ cells/μL in PB. Cultured Burkholderia cells were prepared by growing RPE75 strain to midlog phase and harvesting by centrifugation (5,000 × g for 10 min). Following washing with PB, the suspension of cultured Burkholderia cells was adjusted to 10⁸ cells/μL in PB. The bacterial cells were lysed in 1× Laemmli sample buffer, boiled at 95 °C for 5 min, and 5 × 10⁷ cells per each lane were subjected to SDS/PAGE in 15% gels. Protein bands in the gel were stained with Coomassie Brilliant Blue R250.

LC-MS/MS of 19-kDa Protein.

The identification of the protein band was performed as described (45). The protein band was excised from the gel, destained with 50% (vol/vol) methanol, and dried by vacuum centrifugation. The dried gel was rehydrated with 2 μL of 50 ng/μL trypsin (Promega) and then incubated in 20 μL of 50 mM Tris⋅HCl (pH 8.5) containing 0.1% (wt/vol) n-decyl-β-d-glucopyranoside (Sigma) at 37 °C for 18 h. The resultant digest was analyzed by a nano-LC (1100 series; Agilent Technologies) with a home-made capillary column packed with Inertsil ODS3 (GL Science) at a flow rate of 500 nL/min using a linear gradient of 0–40% solvent B in 40 min, where the solvent A and the solvent B consisted of 0.075% (vol/vol) formic acid in water and 0.075% (vol/vol) formic acid and 80% (vol/vol) acetonitrile in water, respectively. The eluate was analyzed by a hybrid quadrupole-TOF instrument (Q-Tof2; Waters). The resultant MS/MS data were applied to Mascot Software (Matrix Science) for protein identification.

Fluorescence Staining of PHA Granules.

Midlog phase Burkholderia cells were adjusted to 0.5 OD₆₀₀ in PB containing 1% (wt/vol) glucose and incubated on a rotator shaker at 170 × g at 26 °C for 12 h for PHA granule generation. Burkholderia cells grown in YG media up to midlog phase were also used. The bacterial cells were washed with PB, dropped onto glass slides, and fixed with heat. Symbiotic Burkholderia cells were prepared from dissected midgut of fifth-instar Riptortus nymphs as described above, washed with PB, and heat-fixed on glass slides. The bacterial preparations were stained with 0.05% Nile Blue (Santa Cruz Biotechnology, Inc.) dissolved in ethanol. After 5 min of incubation at room temperature, the preparations were rinsed in tap water, treated with 8% (vol/vol) acetic acid for 1 min, rinsed in tap water, air-dried, covered by coverslip with water, and observed under a phase-contrast and fluorescence microscope (Olympus BX50). PHA-derived fluorescence was observed using a 550-nm dichroic mirror with exciter filter BP450-480 and barrier filter BA515.

PHA Measurement by Flow Cytometry.

Burkholderia cells (3 × 10⁹ cells per sample) were washed with distilled water three times, fixed with 70% ethanol at 4 °C for 30 min, and stained with 100 μL of 0.05% Nile Blue for 10 min at room temperature. After washing with distilled water three times, the cells were incubated with 8% acetic acid for 1 min and again washed with distilled water three times. The stained cells were suspended in 0.5 mL of distilled water and analyzed on a flow cytometer (Beckman Coulter, model FC500) with excitation at 488 nm by argon laser. PHA-derived fluorescence was measured in channel FL2 (585 ± 42-nm bandpass filter).

Insect Rearing and Burkholderia Infection.

Colonies of R. pedestris were maintained in our insect laboratory at 26 °C under a long-day condition of 16 h light and 8 h dark (18). Nymphal insects were reared in clean plastic containers supplied with soybean seeds and distilled water containing 0.05% ascorbic acid (DWA). When they reached adulthood, soybean pots were placed in the container, and cotton pads were attached to the wall of the container. Eggs laid on the cotton pads were collected daily and transferred to new containers for collecting newborn nymphs. When the nymphs molted to the second instar, a symbiont inoculum solution was provided as wet cotton balls in a small petri dish. The inoculum solution consisted of the midlog phase of the cultured Burkholderia (10⁷cells/mL) suspended in DWA.

Transmission Electron Microscopy.

Either dissected midgut M4 regions or cultured symbiont cells were prefixed with 2.5% (wt/vol) glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.4) at 4 °C for 18 h, washed three times with 0.1 M sodium cacodylate buffer (pH 7.4) at room temperature for 15 min each, and postfixed with 1% (wt/vol) osmium tetroxide in 0.1 M sodium cacodylate buffer (pH 7.4) for 1 h at room temperature. After three washings, the samples were dehydrated and cleared through an ethanol and propylene oxide series and embedded in Epon 812 resin. The embedded samples were trimmed and sectioned on an ultramicrotome (Reichert SuperNova, Leica) The sections were stained with uranyl acetate and lead citrate and observed under a transmission electron microscope (HITACHI H-7600).

Colony Formation Unit Assay.

Each midgut M4 region dissected from a fifth-instar Riptortus nymph was homogenized in 100 μL of PB by pestle mortar in a 1.5-mL plastic tube. The homogenate was diluted to 1:10,000 with PB, and 50 μL was spread onto a rifampicin-containing YG–agar plate. After 2 d of incubation at 26 °C, the number of colonies on the plate was counted, and the bacterial titer in the original homogenate was calculated by the number of colonies × the dilution factor.

Plasmid Retention Analysis.

Second-instar Riptortus nymphs were orally administrated with the plasmid-transfected Burkholderia strains. Their midgut M4 region was dissected at the fifth instar, homogenized, and spread onto rifampicin-containing YG–agar plates. Twenty colonies from each insect were subjected to PCR using the primers vector-P1 and vector-P2 targeting the kanamycin-resistant gene of the pBBR122 vector (Table S3).

Measurement of Fitness Parameters of the Host.

Survival rate was measured by observing the mortality of insects from the day of Burkholderia symbiont infection until the middle of the fifth-instar stage. Adult emergence was monitored by closely inspecting late-fifth-instar nymphs and counting the number of newly molted adult insects every day. Statistics analysis of adult emergence data was performed using a z-test for proportion using the equation z = (P₁ − P₂)/[P(1 − P)(1/n₁ + 1/n₂)]^0.5, where P₁ and P₂ are proportions of adult insects in each given group, P is the average proportions of two samples, and n₁ and n₂ are sample sizes. Young adult insects were examined for their body length and dry body weight. For weight measurement, the insects were immersed in acetone for 5 min and then completely dried by incubating in a 70 °C oven.

Growth Characteristics of the Symbiont.

Growth curves of the Burkholderia symbiont strains were measured in a minimal medium (0.2% glucose, 0.6% Na₂HPO₄⋅2H₂O, 0.3% KH₂PO₄, 0.1% NH₄Cl, 0.05% NaCl, 0.0003% CaCl₂, 1 mM MgSO₄) and also in YG medium. The symbiont strains were grown primarily in YG medium at 26 °C for 18 h. Using the primary culture, starting cell solutions were prepared by adjusting OD₆₀₀ to 0.05 in YG medium or minimal medium and incubated on a rotator shaker at 170 × g at 26 °C for 24 h. The bacterial cultures were monitored for OD₆₀₀ every 2 or 3 h using a spectrophotometer (Mecasys).

Survival of the Symbiont Under Stress Conditions.

For each of the stress tests, the starting cfu were compared with the cfu after the stress treatment. For the nutrition depletion test, the symbiont cells at midlog phase grown in YG medium were washed with PB, and the cell concentration was adjusted to about 1.2 × 10⁹ cells/mL in PB. The cell solutions were incubated at 26 °C for 2 d, and the number of living cells were counted by cfu assay. For the heat stress test, the symbiont cells at midlog phase grown in YG medium were adjusted to 0.5 OD₆₀₀ in YG medium. Then the cell solutions were incubated in a 45 °C water bath for 10 min. For the osmotic stress test, the symbiont cells at midlog phase grown in YG medium were adjusted to 0.5 OD₆₀₀ in YG medium. Then the same volume of 2-M glucose solution was added to the cell solution and incubated at 26 °C for 24 h. For the oxidative stress test, the symbiont cells at midlog phase grown in YG medium were adjusted to 4 × 10⁵ cells/mL, and 1 mL of the medium was evenly spread onto each YG agar plate containing rifampicin. Whatman paper discs (7 mm in diameter) were placed on the plates, and 8 μL of either 0.5% or 2% (vol/vol) hydrogen peroxide solution was added onto each of the discs. The plates were incubated at 26 °C for 2 d and checked for the size of the growth inhibition zone.