Bacterial Cell Wall Synthesis Gene uppP Is Required for Burkholderia Colonization of the Stinkbug Gut

Abstract

To establish a host-bacterium symbiotic association, a number of factors involved in symbiosis must operate in a coordinated manner. In insects, bacterial factors for symbiosis have been poorly characterized at the molecular and biochemical levels, since many symbionts have not yet been cultured or are as yet genetically intractable. Recently, the symbiotic association between a stinkbug, Riptortus pedestris, and its beneficial gut bacterium, Burkholderia sp., has emerged as a promising experimental model system, providing opportunities to study insect symbiosis using genetically manipulated symbiotic bacteria. Here, in search of bacterial symbiotic factors, we targeted cell wall components of the Burkholderia symbiont by disruption of uppP gene, which encodes undecaprenyl pyrophosphate phosphatase involved in biosynthesis of various bacterial cell wall components. Under culture conditions, the ΔuppP mutant showed higher susceptibility to lysozyme than the wild-type strain, indicating impaired integrity of peptidoglycan of the mutant. When administered to the host insect, the ΔuppP mutant failed to establish normal symbiotic association: the bacterial cells reached to the symbiotic midgut but neither proliferated nor persisted there. Transformation of the ΔuppP mutant with uppP-encoding plasmid complemented these phenotypic defects: lysozyme susceptibility in vitro was restored, and normal infection and proliferation in the midgut symbiotic organ were observed in vivo. The ΔuppP mutant also exhibited susceptibility to hypotonic, hypertonic, and centrifugal stresses. These results suggest that peptidoglycan cell wall integrity is a stress resistance factor relevant to the successful colonization of the stinkbug midgut by Burkholderia symbiont.

INTRODUCTION

Many insects are in intimate symbiotic associations with bacteria. Such symbiotic bacteria exist in the gut lumen, body cavity, or inside cells. To establish a successful host-symbiont association, a number of molecular factors from the symbiont side, and also from the host side, must work in a coordinated manner. To understand the mechanisms of these intricate host-symbiont interactions, several model symbiotic systems have been used to identify novel symbiotic factors and to determine their molecular functions (1). For example, the legume-Rhizobium nitrogen-fixing symbiosis and the squid-Vibrio luminescent symbiosis have been studied in depth. In both systems, the symbiotic bacteria are easily cultivable and genetically manipulatable and are thus suitable for elucidating the molecular properties of their symbiotic factors (2–8).

However, among insect-microbe symbiotic systems, molecular factors relevant to symbiosis have been poorly characterized except for inferences from genomic information (9–11). The paucity of molecular and biochemical studies is attributed to the difficulty in isolating and culturing symbiotic bacteria outside insect hosts. Consequently, powerful mutant-based molecular genetic approaches have not been effectively applied to insect-microbe symbiotic systems in general. Obligate insect symbionts, such as Buchnera in aphids and Wigglesworthia in tsetse flies, have been associated with their hosts over evolutionary time and are incapable of independent living and thus are uncultivable (9, 12). As for facultative insect symbionts, such as Wolbachia in various insects and Sodalis in tsetse flies, which are transmitted through host generations not only vertically but also horizontally, at least some of them are cultivable outside their host insects and thus potentially genetically manipulable (13–15). However, culturing these symbionts is generally not easy because it requires complex culture media containing either mammalian sera or live insect cells, and the symbionts grow very slowly, are prone to contamination, and are reluctant to form colonies on agar plates (16). Therefore, previous studies on bacterial symbiotic factors using genetically manipulated symbionts have been limited (16–21).

The bean bug Riptortus pedestris belongs to the stinkbug family Alydidae in the insect order Hemiptera. In contrast to previously known insect-bacterium symbiotic systems, nymphal R. pedestris acquires a betaproteobacterial symbiont of the genus Burkholderia not vertically but from the soil environment every generation (22). A posterior region of the insect midgut bears numerous crypts whose lumens are filled with bacterial cells of the symbiotic Burkholderia (23). Reflecting its free-living origin in the environment, the symbiotic Burkholderia is easily cultivable on standard microbiological media and can be experimentally reinfected into the host insect by oral administration (24, 25). Comparisons between symbiotic and asymbiotic insects showed beneficial fitness consequences of Burkholderia infection to the host insect (22, 26). These features of the Riptortus-Burkholderia gut symbiotic system provide unprecedented opportunities to study insect symbiosis at molecular and biochemical levels.

The cell wall of Gram-negative bacteria is the front-line of interacting with the surrounding environment. It consists of an inner membrane, an outer membrane in which lipopolysaccharide (LPS) forms the outer leaflet, and a periplasmic region where the peptidoglycan layer resides (27). Bacterial cell wall components such as LPS and peptidoglycan are essential for maintaining the structural integrity of bacterial cells and generally required for viability (27, 28). In addition, these cell wall components most likely play a role in bacterial association with host and hence, may function as symbiotic factors. Biosynthesis of bacterial cell wall components, such as LPS and peptidoglycan, requires a key lipid carrier, undecaprenyl phosphate (C₅₅-P), which is generated from dephosphorylation of undecaprenyl pyrophosphate (C₅₅-PP) (29–34). C₅₅-P is a precursor of various cell wall components that are synthesized in the cytoplasm and transported to the periplasm, where further polymerization occurs. After release from the cell wall component precursors, the lipid carrier is in a pyrophosphate form (C₅₅-PP) and requires another dephosphorylation step before being reused as a lipid carrier (35). This dephosphorylation step is catalyzed by C₅₅-PP phosphatase enzymes. Four C₅₅-PP phosphatases have been identified in Escherichia coli: UppP (also called BacA), YbjG, YeiU and PgpB, of which UppP is regarded as the major phosphatase (36, 37).

To identify bacterial symbiotic factors in the Riptortus-Burkholderia symbiosis, we targeted the bacterial cell wall-related uppP gene. We generated an uppP-deficient mutant (ΔuppP) of the Burkholderia symbiont by allelic exchange and homologous recombination. Because the ΔuppP mutant shows 75% reduction of C₅₅-PP phosphatase activity in E. coli (36), we hypothesized that the decrease of C₅₅-PP phosphatase activity affects the cell wall component synthesis, resulting in defected cell wall. Since the actual effects on the cell wall by the uppP mutation are not well characterized, we first examined cell wall components of a ΔuppP Burkholderia strain. Furthermore, the growth phenotypes in vitro and symbiotic phenotypes in vivo of the ΔuppP mutant were compared to those of the wild-type Burkholderia symbiont and an ΔuppP/uppP-complemented mutant transfected with a plasmid encoding a functional uppP gene.

MATERIALS AND METHODS

Bacteria, plasmids, and culture media.

Bacterial strains and plasmids used in the present study are listed in Table 1. E. 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). Cells of Burkholderia symbiont strain RPE161, a spontaneous chloramphenicol-resistant mutant derived from RPE64 (24), were cultured at 30°C in YG medium (0.5% [wt/vol] yeast extract, 0.4% [wt/vol] glucose, and 0.1% [wt/vol] NaCl). Antibiotics were used at the following concentrations unless otherwise described: kanamycin at 30 μg/ml and chloramphenicol at 10 μg/ml.

Table 1.

Bacterial strains and plasmids used this study

Strain or plasmid	Relevant characteristicsa	Source or reference
Strains
Burkholderia symbionts
RPE161	Burkholderia symbiont (RPE64); Cmr	24
BBL005	RPE161 ΔuppP; Cmr	This study
BBL105	BBL005/pBL5, complementation of uppP; Cmr Kmr	This study
Escherichia coli
DH5α	deoR endA1 gyrA96 hsdR17(rK− mK+) phoA recA1 relA1 supE44 thi-1 Δ(lacZYA-argF)U169 ϕ80dlacZΔM15 F− λ−	Toyobo
CC118λpir	Carrying helper plasmid pEVS104; Rifr Kmr	51
Plasmids
pEVS104	oriR6K helper plasmid containing conjugal tra and trb; Kmr	51
pK18mobsacB	pMB1ori allelic exchange vector containing oriT; Kmr	52
pBBR122	Broad host vector; Cmr Kmr	53
pBL5	pBBR122 derivative containing uppP; Kmr	This study

^aCm^r, chloramphenicol resistance; Rif^r, rifampin resistance; Km^r, kanamycin resistance.

Generation of ΔuppP mutant.

Deletion of the chromosomal uppP gene from the Burkholderia symbiont was accomplished by allelic exchange and homologous recombination using a suicide vector pK18mobsacB containing the 5′ (uppP-L) and 3′ (uppP-R) regions of uppP gene. The wild-type Burkholderia symbiont strain RPE161 was subjected to PCR using the primers uppP-L-P1 (5′-TTT AAG CTT GAG TTC GAC TTC GAG CGT GT-3′) and uppP-L-P2 (5′-TTT GGA TCC AAG ACT GCT GAC CGG AAA AA-3′) for the uppP-L region, and the primers uppP-R-P1 (5′-TTT GGA TCC TTC TTC TTC GGC TGG TTC AT-3′) and uppP-R-P2 (5′-TTT GAA TTC GCA CTG GAA AAC CTC AGC A-3′) for the uppP-R region. PCR products and the pK18mobsacB vector were digested with proper restriction enzymes, ligated, and transformed into E. coli DH5α cells. The transformed E. coli cells were selected on LB-agar plates containing 100 μg of kanamycin/ml. Positive colonies carrying a vector with the correct insert were further selected by colony PCR using the primer uppP-L-P1 and the vector primer aphII (5′-ATC CAT CTT GTT CAA TCA TGC G-3′). Donor E. coli cells carrying the pK18mobsacB containing uppP-L and uppP-R were mixed with recipient Burkholderia RPE161 cells and also E. coli CC118λpir cells carrying a helper plasmid pEVS104 to transfer the cloned vector to the RPE161 cells. After allowing a single crossover by culturing cell mixtures of triparental conjugation on YG-agar, RPE161 cells with the first crossover were selected on YG-agar containing chloramphenicol (30 μg/ml) and kanamycin. Positive colonies with the genomic integration of vector DNA were confirmed by PCR using the chromosomal primer uppP-up (5′-GAG GCA ATG AAA CGT ATC GAC-3′) and the vector primer aphII. The second crossover was allowed by culturing cells with the single crossover in YG media and Burkholderia cells with a double crossover were selected on YG-agar containing chloramphenicol and sucrose (200 μg/ml). The mutant strain with deletion of the uppP gene (BBL005) was identified by PCR using the primers uppP-up and uppP-down (5′-CCA GCA TCT GCT CTT TGT CA-3′) and sequencing of the PCR product.

Generation of ΔuppP/uppP-complemented mutant.

A DNA fragment containing the open reading frame of uppP gene was amplified from RPE161 using the primers uppP-com-P1 (5′-GCA CGG CAA TTT TTC TCT TC-3′) and uppP-com-P2 (5′-CGA CTC GAA CGT GTG ACC TA-3′). The amplified DNA fragment was cloned into the DraI site of pBBR122 to generate the plasmid pBL5. The cloned plasmid was introduced into E. coli DH5α cells to generate donor cells. By triparental conjugation with the BBL005 recipient cells and E. coli CC118λpir helper cells, the pBL5 plasmid carried by the donor E. coli DH5α cells was transferred to the recipient Burkholderia BBL005 cells, yielding the complemented Burkholderia BBL105 cells. The complemented mutant strain was selected on YG-agar with chloramphenicol (30 μg/ml) and kanamycin.

Measurement of bacterial growth in liquid media.

Growth curves of the Burkholderia symbiont strains were examined either in YG medium or in minimal medium (0.6% Na₂HPO₄·2H₂O, 0.3% KH₂PO₄, 0.1% NH₄Cl, 0.05% NaCl, 0.0003% CaCl₂, 1 mM MgSO_4, 0.2% glucose). The starting cell solutions were prepared by adjusting the optical density at 600 nm [OD₆₀₀] to 0.05 in either YG medium or minimal medium using primary culture grown in YG medium at 30°C for 18 h. The cell solutions were incubated on a rotator shaker at 180 rpm at 30°C for 36 h, whose OD₆₀₀ was monitored every 2 h using a spectrophotometer (Mecasys, South Korea).

Protein analysis of bacterial lysates.

Burkholderia symbiont cells were harvested at an OD₆₀₀ of 1 after culturing in YG medium. The cells were washed with PBS (137 mM NaCl, 2.7 mM KCl, 8 mM NaH₂PO₄, and 3 mM KH₂PO₄ at pH 7) and resuspended at 2 × 10⁷ cells/μl in PBS. An aliquot of this solution was saved for the whole-lysate fraction (WL). The cell suspension was then sonicated and further diluted to 10⁷ cells/μl equivalent in PBS containing 10 mM EDTA, 100 μg of egg white lysozyme (BioShop Canada, Inc., Canada)/ml, and protease inhibitors (Roche, Germany). After adding one-fourth volume of 10% Triton X-114 (final, 2%), the cell solution was agitated for 1 h at 4°C. The sample was centrifuged (15,000 × g for 20 min at 4°C), and the pellet was saved for the insoluble fraction (IS), while the supernatant was transferred to a new tube. The IS fraction was washed with PBS and resuspended in 1× Laemmli sample buffer. The liquid was incubated at 37°C for 10 min and centrifuged (10,000 × g for 10 min at 25°C). After a 10-min incubation at room temperature, the separated aqueous fraction (AQ) was transferred to a fresh tube and supplemented with Triton X-114 solution to a final concentration of 2% for additional phase partitioning before collecting the final AQ fraction. The Triton X-114 fractions (TX) from both partitionings were combined, and an equal volume of TBSE solution (20 mM Tris-HCl [pH 8], 130 mM NaCl, and 5 mM EDTA) was added. After agitation for 10 min at 4°C, the samples were centrifuged (10,000 × g for 10 min at 25°C) and separated into upper and lower phases at room temperature. The upper layer was then discarded, TBSE was added, and the procedure was repeated. Final TX fractions were precipitated with cold ethanol, and dried precipitates were resuspended in 1× Laemmli sample buffer. Proteins from different phase fractions were separated by SDS–15% PAGE and visualized by staining with Coomassie brilliant blue R250. The loading quantity for each fraction was 7 × 10⁷ cells equivalent for WL, 7 × 10⁷ cells equivalent for AQ, 6 × 10⁸ cells equivalent for TX, and 3 × 10⁸ cells equivalent for IS.

Carbohydrate analysis.

The WL samples prepared for protein analysis by SDS-PAGE were used for the analysis of bacterial carbohydrates. The WL sample was boiled in 1× Laemmli sample buffer, deproteinated by incubation with 400 μg of proteinase K/ml at 60°C for 1 h, and reboiled prior to SDS-PAGE. Loading amount was 1 × 10⁸ cells equivalent per lane for 12% Laemmli SDS-PAGE gels and 2 × 10⁸ cells equivalent per lane for 12% Tris-Tricine SDS-PAGE gels. Bacterial carbohydrates separated in the gels were visualized using the Pro-Q Emerald 300 lipopolysaccharide gel stain kit (Invitrogen). Briefly, the gels were fixed with 5% acetic acid and 50% methanol, washed three times with 3% acetic acid, incubated with oxidizing solution containing periodic acid for 30 min, washed three times again with 3% acetic acid, and stained with Pro-Q Emerald 300 staining solution for 2 h. After two washes with 3% acetic acid, the gels were observed with the gel documentation system GDS-200.

Lysozyme susceptibility assay.

Frozen mid-log phase Burkholderia cells were thawed and resuspended in PB (10 mM sodium phosphate, pH 7). After washing with PB, 0.9 ml of the Burkholderia cell suspension was prepared at an OD₆₀₀ of 0.77 to 0.78 in PB and transferred to a cuvette for spectrophotometry. After an addition of 0.08 ml of lysis solution (500 μg of egg white lysozyme/ml in PB with 100 mM EDTA), the OD₆₀₀ of the cell suspension was measured every 2 min up to 28 min and then every 5 min until 73 min. As a control, 0.08 ml of PB containing 100 mM EDTA was added to the cell suspension.

Insect rearing and symbiont inoculation.

R. pedestris bean bugs were reared in our insect laboratory at 28°C under a long day regime of 16 h light and 8 h dark as described previously (38). Nymphal insects were reared in clean plastic containers (34 cm by 19.5 cm wide by 27.5 cm high) supplied with soybean seeds and DWA (distilled water containing 0.05% ascorbic acid). Upon reaching adulthood, the insects were transferred to a bigger container (35 cm by 35 cm wide by 40 cm high) in which soybean plant pots were placed for food and cotton pads were attached to the walls as a substrate for egg laying. Eggs were collected daily and transferred to new cages for hatching. Newly molted second instar nymphs were provided with wet cotton balls soaked with a symbiont inoculum solution consisting of mid-log-phase Burkholderia cells suspended in DWA at a concentration of 10⁷ cells/ml.

The care and treatment of Burkholderia cells and insects in all procedures strictly followed the guidelines of the Pusan National University (PNU) Institutional Animal Care and Use Committee (IACUC) and the Living Modified Organ (LMO) Committee.

Diagnostic PCR.

Insects were surface sterilized briefly with 70% ethanol and dissected in PBS in a glass petri dish using fine scissors and forceps under a dissection microscope. Dissected samples of the posterior midgut M4 region were individually subjected to DNA extraction using the QIAamp DNA minikit (Qiagen). Diagnostic PCR was conducted using GoTaq Green Master Mix (Promega) with the supplied buffer system under a temperature profile of 95°C for 10 min, followed by 30 cycles of 95°C for 30 s, 58°C for 30 s, and 72°C for 1 min, and finally 72°C for 2 min using the primers Burk16SF (5′-TTT TGG ACA ATG GGG GCA AC-3′) and Burk16SR (5′-GCT CTT GCG TAG CAA CTA AG-3′), which specifically target 16S rRNA gene of the Burkholderia symbiont (38). PCR products were analyzed by 1% agarose gel electrophoresis and a 100-bp DNA ladder was used to estimate product size.

CFU assay.

Each of the M4 midgut regions dissected from second instar Riptortus nymphs was collected in 50 μl of PB and homogenized by a pestle mortar. The homogenized sample was diluted if necessary and spread on YG-agar plates containing chloramphenicol. After 2 days of incubation at 30°C, colonies on the plates were counted, and the number of symbiont cells in the sample was calculated as the CFU × the dilution factor.

Quantitative PCR.

Quantitative PCR for estimating titers of the Burkholderia symbiont was performed as described previously (38). Dissected midgut samples (either M3 or M4) were individually subjected to DNA extraction by QIAamp DNA minikit (Qiagen). DNA samples were mixed with a master PCR solution containing 2× qPCR premix of QuantiMix SYBR kit (PhileKorea) and the primers BSdnaA-F (5′-AGC GCG AGA TCA GAC GGT CGT CGA T-3′) and BSdnaA-R (5′-TCC GGC AAG TCG CGC ACG CA-3′), which target a 0.15-kb region of dnaA gene of the Burkholderia symbiont. The PCR temperature profile was 40 cycles of 95°C for 10 s, 60°C for 15 s, and 72°C for 15 s. A standard curve for dnaA gene copies was generated using a series of extracted DNA samples containing known numbers of Burkholderia cells.

CFU assay after stress treatments.

For each test, the starting CFU were compared to CFU after the following treatments. (i) For the M4 lysate treatment, midgut M4 regions dissected from fifth-instar Riptortus nymphs were homogenized and heat treated at 55°C for 5 min to kill intrinsic symbiont cells prior to the assay. Different concentrations of the M4 lysate, ranging from 0.0 to 0.4 mg/ml, were incubated with cultured Burkholderia cells at mid-log phase for 1 h at room temperature. After incubation, the samples were diluted, spread on YG-agar plates, cultured for 2 days, and subjected to colony counting. (ii) For the hypotonic test, mid-log-phase Burkholderia cells in YG medium were washed with 10 mM PB and adjusted to ∼10⁷ cells/ml in PB. The cell suspensions were incubated at 30°C for 24 h and subjected to CFU assay. (iii) For the hypertonic test, mid-log-phase Burkholderia cells were adjusted to an OD₆₀₀ of 0.5 to 0.7 in YG medium. The cell suspension was combined with an equal volume of 2 M glucose solution, incubated at 30°C for 24 h, and subjected to CFU assay. (iv) For the centrifugal pressure test, mid-log-phase Burkholderia cells cultured in YG medium were adjusted to 10⁴ cells/ml, placed in 1.5-ml microcentrifuge tubes, centrifuged at 15,000 rpm (20,000 × g) for 30 min, and subjected to CFU assay.

RESULTS

Growth rates of wild-type and mutant Burkholderia symbiont strains.

We disrupted the uppP gene of the wild-type Burkholderia symbiont strain RPE161, thereby establishing a ΔuppP mutant Burkholderia symbiont strain BBL005. By transforming the ΔuppP mutant strain with a plasmid encoding a functional uppP gene, we also generated a ΔuppP/uppP-complemented mutant Burkholderia symbiont strain BBL105. Growth curves of these Burkholderia strains in nutritionally rich yeast-glucose (YG) medium revealed that the wild-type strain and the ΔuppP mutant exhibited similar growth rates, while the ΔuppP/uppP-complemented mutant grew a little slower (Fig. 1A). Growth curves in nutritionally limited minimal medium exhibited similar patterns, although growth rates overall were much slower in minimal medium than in YG medium (Fig. 1B). These results indicate that deletion of the uppP gene does not affect growth of the Burkholderia symbiont under in vitro culture conditions. The slower growth of the ΔuppP/uppP-complemented mutant may be due to a cost of harboring the plasmid.

Fig 1.

Growth curves of the wild-type Burkholderia symbiont strain (RPE161), the ΔuppP mutant strain (BBL005), and the ΔuppP/uppP-complemented mutant strain (BBL105) in YG medium (A) and in minimal medium (B).

Susceptibility of the ΔuppP mutant to lysozyme.

Previous studies have shown that the product of the UppP-mediated enzymatic reaction, C₅₅-P, is involved in biosynthesis of various cell wall components including peptidoglycan, LPS, colanic acid, and teichoic acid (30–34, 39). Hence, we compared protein composition, carbohydrate expression and lysozyme susceptibility of the wild-type Burkholderia symbiont strain and the ΔuppP mutant strain. In sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of proteins extracted from the cultured Burkholderia cells, the whole lysates (WL) were partitioned into water-soluble (aqueous [AQ]), Triton X-114-soluble (TX), and water/Triton X-114-insoluble (IS) fractions. No notable differences in protein profiles were detected between the wild-type strain and ΔuppP mutant (Fig. 2A). Carbohydrates of proteinase K-treated bacterial lysates were separated by SDS-PAGE and subjected to periodic acid oxidation and fluorescent staining. Ladder patterns representing repeating units of LPS O-antigen and high-molecular-weight bacterial carbohydrates were commonly detected in the wild-type strain and ΔuppP mutant, and the profiles exhibited no apparent differences between them (Fig. 2B). On the other hand, when lysozyme was added to bacterial cell suspensions, the ΔuppP mutant exhibited a much greater reduction in turbidity than the wild-type strain, and the reduction in turbidity was restored in the ΔuppP/uppP-complemented mutant to the level of the wild-type strain (Fig. 2C). These results indicate that cell wall integrity of the ΔuppP mutant is impaired by disruption of the uppP gene.

Fig 2.

In vitro characterization of the ΔuppP mutant strain BBL005. (A) Protein analysis of the Burkholderia symbiont strains by SDS-PAGE. WL, whole lysate; AQ, aqueous soluble fraction; TX, Triton X-114 soluble fraction; IS, insoluble fraction. (B) Carbohydrate analysis by SDS-PAGE. (C) Lysozyme susceptibility assay of the Burkholderia strains. Error bars indicate standard deviations.

Atrophied host symbiotic organ and symbiosis defect in the Riptortus host infected with the ΔuppP mutant.

To examine the symbiotic properties of the ΔuppP mutant, wild-type, ΔuppP, or ΔuppP/uppP Burkholderia cells were orally administered to early second instar Riptortus nymphs. The insects were reared to the fourth instar, and their midgut symbiotic organs were dissected and inspected morphologically. In wild-type-infected insects, the symbiotic organs were well developed and hazy in color, which was indicative of bacterial cells filling the midgut crypts (Fig. 3A). In ΔuppP-infected insects, by contrast, the symbiotic organs were atrophied and translucent in color (Fig. 3B), which was reminiscent of the symbiotic organs of uninfected control insects (Fig. 3C). In ΔuppP/uppP-infected insects, the well-developed hazy symbiotic organs were restored (Fig. 3D). Diagnostic PCR of the dissected symbiotic organs confirmed the absence of symbiont infection in the ΔuppP-infected insects (Fig. 3E). These results indicate that the ΔuppP mutant strain is deficient in symbiosis and that disruption of the uppP gene is responsible for this phenotype.

Fig 3.

(A to D) Morphology of host symbiotic midgut inoculated with Burkholderia symbiont strains: wild-type strain RPE161 (A), ΔuppP mutant strain BBL005 (B), uninfected control (C), and ΔuppP/uppP-complemented mutant strain BBL105 (D). Insects were orally administered with the Burkholderia cells at the second instar and dissected for inspection of the midgut at the fourth instar. (E) Diagnostic PCR detection of the Burkholderia infection in midgut dissected from third-, fourth-, and fifth-instar nymphs.

Initial infection but no proliferation of the ΔuppP mutant in the host symbiotic organ.

To compare the initial infection processes of the wild-type strain and the ΔuppP mutant, second-instar Riptortus nymphs were orally administered with the cultured symbiont strains and maintained for 10, 15, 20, or 25 h after inoculation. Subsequently, their midguts were dissected, individually subjected to DNA extraction, and analyzed by quantitative PCR targeting dnaA gene of the Burkholderia symbiont strain (Fig. 4A). The wild-type strain was already detectable in both the M3 region and the M4 symbiotic region of the host midgut at 10 h after inoculation, and the symbiont population steadily increased at 15, 20, and 25 h after inoculation. In contrast, the ΔuppP mutant was also detected in both the M3 region and the M4 symbiotic region of the host midgut at 10 h after inoculation, but the symbiont population exhibited no increase at 15, 20, or 25 h after inoculation (Fig. 4A). We also performed a CFU assay for the wild-type strain, ΔuppP mutant, and ΔuppP/uppP-complemented mutant on dissected midgut samples from second instar Riptortus nymphs at 36 and 63 h after inoculation (Fig. 4B). At these later stages, the infection titers of the ΔuppP mutant (∼10² per insect) were drastically lower than those of the wild-type strain (10⁴ to 10⁵ per insect). Notably, infection titers of the ΔuppP/uppP-complemented mutant exhibited significant restoration to 10³ to 10⁴ per insect (Fig. 4B). These results indicate that the ΔuppP mutant is certainly incorporated into the host midgut but cannot proliferate and survive in the symbiotic organ, thereby failing to establish the symbiotic association with the Riptortus host.

Fig 4.

Quantitative analyses of the Burkholderia symbiont strains in the host symbiotic organs of second instar Riptortus nymphs. (A) Quantitative PCR analysis of infection densities of the wild-type strain RPE161 and the ΔuppP mutant BBL005 at 10, 15, 20, and 25 h after inoculation. (B) CFU quantification of infection densities of the wild-type strain RPE161, the ΔuppP mutant BBL005, and the ΔuppP/uppP-complemented mutant BBL105 at 36 and 63 h after inoculation. Different letters (a and b) indicate statistically significant differences (unpaired Student t test; *, P < 0.05; **, P < 0.01; ****, P < 0.0001).

Effect of symbiotic organ lysate on the ΔuppP mutant.

Considering the lysozyme susceptibility of the ΔuppP mutant (Fig. 2C) and its incapability of survival in the host symbiotic organ (Fig. 3 and 4), we hypothesized that the host symbiotic organ may possess bactericidal activities to which the wild-type strain is resistant but the ΔuppP mutant is susceptible. To explore this possibility, we dissected fifth-instar Riptortus nymphs and collected their midguts. The dissected symbiotic organs were homogenized and heat-treated to inactivate intrinsic Burkholderia cells, and the lysates at different concentrations were applied to cultured wild-type Burkholderia cells and ΔuppP mutant cells. No significant effect of the midgut lysate was observed on either the wild-type strain or the ΔuppP mutant (Fig. 5A).

Fig 5.

(A) Survival of the wild-type Burkholderia symbiont strain RPE161 and the ΔuppP mutant strain BBL005 when symbiotic midgut lysates from fifth instar Riptortus nymphs were added to the cultured bacteria. (B to D) Survival of the wild-type strain RPE161, the ΔuppP mutant BBL005, and the ΔuppP/uppP-complemented mutant BBL105 under environmental stress conditions. (B) Under a hypotonic condition in 10 mM phosphate buffer for 24 h. (C) Under a hypertonic condition in 1 M glucose for 24 h. (D) Under a high gravity condition of centrifugation at 20,000 × g for 30 min. Different letters (a and b) indicate statistically significant differences (unpaired Student t test with a Bonferroni correction; P < 0.05).

Survival of the ΔuppP mutant under environmental stress conditions.

The wild-type strain, the ΔuppP mutant, and the ΔuppP/uppP-complemented mutant of the Burkholderia symbiont were exposed to several stressful conditions in vitro, and their survival was evaluated by CFU assay. Under a hypotonic condition in 10 mM phosphate buffer for 24 h, the ΔuppP mutant exhibited a significantly lower survival rate than the wild-type strain and the ΔuppP/uppP-complemented mutant (Fig. 5B). Under a hypertonic condition in 1 M glucose for 24 h, the ΔuppP mutant also showed a significantly lower survival rate than the wild-type strain and the ΔuppP/uppP-complemented mutant (Fig. 5C). When centrifugal pressure at 20,000 × g for 30 min was applied to the cultured Burkholderia cells, the ΔuppP mutant again showed a significantly lower survival rate than the wild-type strain and the ΔuppP/uppP-complemented mutant (Fig. 5D). These results strongly suggest that the ΔuppP mutant is susceptible to environmental stresses, which is likely attributable to the impaired cell wall integrity caused by the disruption of uppP gene.

DISCUSSION

In this study, we show that the ΔuppP mutant of the Burkholderia symbiont fails to establish symbiosis in the host midgut M4 region, while the ΔuppP/uppP-complemented mutant restores normal association with the host midgut (Fig. 3 and 4). These results indicate that uppP gene of the Burkholderia symbiont is essential for establishing normal gut symbiotic association with the Riptortus host.

In E. coli and other bacteria, uppP gene is involved in the biosynthesis of various cell wall components, including peptidoglycan, LPS, and others (29–34). The ΔuppP mutant of the Burkholderia symbiont exhibits higher susceptibility to lysozyme than the wild-type strain, whereas the ΔuppP/uppP-complemented mutant shows a restored level of lysozyme susceptibility comparable to that of the wild-type strain (Fig. 2C). These results indicate that disruption of the uppP gene impairs integrity of the symbiont cell wall and suggest that the cell wall defect is likely relevant to the symbiosis defect of the ΔuppP mutant.

Why the ΔuppP mutant cannot establish infection in the host's symbiotic organ is currently elusive. Considering the lysozyme susceptibility and the impaired cell wall integrity of the ΔuppP mutant (Fig. 2C), a hypothesis is that the symbiotic midgut is producing bactericidal factors such as lysozymes or antimicrobial peptides, to which the wild-type symbiont is resistant but the ΔuppP mutant (and possibly also nonsymbiotic bacteria) is susceptible. Our results that lysates of the midgut M4 region of fifth-instar Riptortus nymphs affected neither the wild-type symbiont nor the ΔuppP mutant (Fig. 5A) do not support this hypothesis, but it should be noted that the lysate was heat treated to kill intrinsic Burkholderia cells and thus heat-sensitive bactericidal factors may have been inactivated by the treatment. Although the midgut lysate was prepared from fifth-instar nymphs due to difficulty in collecting sufficient amount of the sample from younger nymphs, it should also be noted that the Burkholderia infection initially establishes in the host midgut at the second instar, not the fifth. Interestingly, a recent transcriptomic analysis of the midgut regions of Riptortus nymphs revealed that host antimicrobial genes, such as a c-type lysozyme gene and a defensin-like gene, are highly expressed in asymbiotic insects but scarcely expressed in symbiotic insects (40). In the bacteriocytes of the grain weevils, an antimicrobial peptide, coleoptericin A, regulates the population and proliferation of the Sodalis-allied endosymbiont (41). In the bacteriocytes of the pea aphid, two i-type lysozyme genes are specifically expressed and represent the most abundant transcripts in the symbiotic cells, presumably regulating the population and proliferation of the Buchnera endosymbiont (42). Hence, the possibility cannot be ruled out that such bactericidal gene products are preferentially expressed in the Riptortus midgut, act on the symbiont cell wall, and result in the infection failure of the ΔuppP mutant of the Burkholderia symbiont.

Considering the susceptibility of the ΔuppP mutant to environmental stresses, such as low osmolality, high osmolality, and high centrifugal pressure (Fig. 5B to D), an alternative hypothesis is that the symbiotic conditions within the host midgut entail some environmental stresses, to which the wild-type symbiont is resistant but the ΔuppP mutant is susceptible. Although the nature of the “symbiotic stress” is unknown, it may be osmotic, anoxic, nutritional, immunological, or a combination of these. In this context, a recent study demonstrated a crucial involvement of bacterial stress-related genes in the Riptortus-Burkholderia symbiosis: disruption of symbiont genes for synthesizing an endocellular storage polyester, polyhydroxyalkanoate (PHA), which confers bacterial resistance to nutritional depletion and other environmental stresses, resulted in failure of normal symbiotic association, while complementation of the PHA synthesis genes rescued the symbiosis defect (54). It should be noted that the “bactericidal factor hypothesis” and the “symbiotic stress hypothesis” may not necessarily be mutually exclusive, on the ground that the bactericidal factors could be regarded as comprising host-derived immunological stresses.

The cell wall is located on the outer surface of bacterial cells as a front line of host-symbiont interactions. Therefore, considerable attention has been paid to the possible relevance of the symbiont cell wall to symbiosis, particularly to interactions with host's innate immunity. For example, some endosymbiotic bacteria, such as Spiroplasma and Wolbachia, exhibit remarkable degeneration in their cell wall, thereby eliciting no or little innate immune responses of their host insects (43–46). Transcriptomic comparisons between symbiotic and asymbiotic host insects have revealed that a variety of immunity-related genes, including lysozyme genes and antimicrobial peptide genes, are upregulated in symbiosis-associated patterns (40, 42, 47–50). To our knowledge, apart from general studies of bacterial cell wall changes and host immune responses, this study is the first to unequivocally identify that a specific cell wall biosynthesis-related symbiont gene is required for an insect-bacterium symbiotic association.

On the basis of previous studies in squid-Vibrio, nematode-Photorhabdus/Xenorhabdus, and other model symbiotic systems, Ruby in 2008 classified symbiosis-deficient bacterial mutants into (i) initiation mutants, which are unable to establish infection in the host, (ii) accommodation mutants, which can establish infection but fail to reach the usual infection density, and (iii) persistence mutants, which at first establish infection normally but are unable to maintain the normal infection level (1). Under these criteria, the ΔuppP mutant can be regarded as a mutant between an initiation mutant and accommodation mutant, because it is able to infect initially but fails to establish colonization in the Riptortus host. The cell wall deficiency of the ΔuppP mutant most likely affects the initial host-symbiont association, which highlights a previously underexplored aspect of insect-bacterium symbiotic associations.