Staphylococcus aureus SA113 (ATCC35556) was used for the study of hitchhiking motility (Herbert et al., 2010; Tsompanidou et al., 2011). A ΔtagO mutant of S. aureus SA113 contains a deletion in tagO and does not produce WTA (Weidenmaier et al., 2004). For complementing the ΔtagO mutation, the strain was transformed with a TagO-expressing plasmid, pHY-tagO. Pseudomonas aeruginosa PAO1 and its isogenic mutant ΔfliA, a flagellum-defective mutant (Lo et al., 2016) were kindly provided by Professor Hwan-You Chang. Escherichia coli EPI300 (Epicentre Technologies, Wisconsin, United States) and S. aureus RN4220, a restriction-deficient strain (Nair et al., 2011), were used as hosts for cloning. Staphylococcus aureus strains were transformed with pCtuf-ppmch (Mauthe et al., 2012), which encodes mCherry and exhibits red fluorescence. Staphylococcus aureus CGL1190, a bioluminescent strain that constitutively expresses a luciferase reporter. Plasmid pHY-tagO was constructed by inserting a PCR-amplified DNA fragment containing the tagO from S. aureus SA113. The fragment was amplified with primers 5′-CGGTCTAGATAGCACTTGTTACTGCAGCA and 5′-TTACCCGGGATCCCATACAGCTATGCTTT and digested with XbaI and SmaI and finally inserted into the same sites in pHY300pLK (TaKaRa Bio) to generate pHY-tagO. Bacteria were cultured in tryptic soy broth (TSB) and tryptic soy agar (TSA; Oxoid). Mannitol salt agar (BD Difco) was used for differentiating and enumerating S. aureus colonies. Antibiotic-resistant colonies were selected on media that contained ampicillin (100 μg/ml), chloramphenicol (10 μg/ml), kanamycin (50 μg/ml), and tetracycline (5 μg/ml).

To generate a bioluminescent S. aureus strain, a temperature-sensitive plasmid pRP1190 (Plaut et al., 2013), which contains a modified luxBADCE operon from Photorhabdus luminescens, was transformed into S. aureus SA113. Integration of plasmids into bacterial chromosomes by homologous recombination was performed according to a method described previously (Plaut et al., 2013). The plates were imaged using an in vivo Imaging System (IVIS Lumina III, PerkinElmer, Waltham, Massachusetts, United States). Luminescent colonies were selected, and the integration site was confirmed by PCR.

Hitchhiking motility was assayed according to a method described elsewhere (Samad et al., 2017) but with modifications. Staphylococcus aureus and Pseudomonas aeruginosa strains were cultured to the mid-log phase and then mixed at a ratio of 1:100 in TSB. Bacterial mixture (200 μl) that contained 10⁶ colony forming unit (CFU) of S. aureus and 10⁸ CFU of P. aeruginosa was transferred into the wells in a Calgary Biofilm Device (CBD; Innovotech, Edmonton, Alberta, Canada). The plate was placed at room temperature for 1 h. Subsequently, the pegs on a CBD lid was immersed in the wells for 30 s to allow attachment of the bacteria. The pegs on the lid were then washed by immersing the pegs in 200 μl PBS in each well of a 96-well microtiter plate with gentle shaking using a Thermomixer (Eppendorf, Hamburg, Germany). After washing, the lid was transferred and immersed in a new 96-well plate containing 200 μl PBS and 0.1 mm-diameter glass beads (Biospec Products, Bartlesville, Oklahoma, United States). The plate was vortexed gently using a Thermomixer (Eppendorf, Hamburg, Germany) to remove the bacteria from the pegs. The bacterial suspension was then serially diluted and plated on mannitol salt agar. The number of S. aureus SA113 on the pegs was enumerated by viable cell count.

The pegs on the lid of CBD were removed and fixed overnight in 2% glutaraldehyde. The bacteria were examined under a field emission scanning electron microscope (FE-SEM, Hitachi SU8220, Hitachi High-Tech, Tokyo, Japan) according to a method described elsewhere (Samad et al., 2017). Images were colored using Photoshop CS3 (Adobe Systems, San Jose, California, United States).

TSA plates prepared with 0.4% agarose (TSA-0.4; Seakem, Lonza Rockland, Maine, United States) were used for swarming assay. Bacterial culture (2 μl) that contained 1×10⁴ CFU of S. aureus and 1 × 10⁶ CFU of P. aeruginosa was spotted at the center of a TSA-0.4 plate. The plate was subsequently incubated at 37°C for 4 h to allow P. aeruginosa to swarm. The bacteria at the center and the edge of a swarm were picked with a toothpick, which was inserted vertically into the bottom of the plate. The bacteria adhered to the toothpick were dissolved in 1 ml PBS. After serial dilution, bacteria were plated on mannitol salt agar. The number of S. aureus in the bacterial suspension was enumerated by viable cell count.

To detect the binding of LPS from P. aeruginosa to S. aureus, FITC-LPS (Sigma-Aldrich, United States) and free LPS (Sigma-Aldrich, United States) were mixed with 10⁵ CFU of S. aureus. The fluorescence intensity was detected, and the number of S. aureus SA113 bound to FITC-LPS was enumerated with a Guava easyCyte flow cytometer (Merck Millipore). To assess the binding of WTA of S. aureus or free lipoteichoic acids (LTA; Sigma-Aldrich, United States) to P. aeruginosa, mid-log phase S. aureus and P. aeruginosa were stained with hexidium iodide (Thermo Fisher Scientific, Waltham, MA, United States) and SYTO 9 (Invitrogen; Thermo Fisher Scientific) according to the instruction provided by the vendors. Labeled S. aureus and P. aeruginosa were mixed at a ratio of 1:1 and 10⁵ CFU/ml of the bacteria was used for flow cytometry analysis. Five thousand fluorescent-stained cells were counted in a gate set in a hexidium iodide versus SYTO 9 dot plot.

Caenorhabditis elegans N2 was cultured and maintained according to a method described previously (Brenner, 1974). Staphylococcus aureus strains harboring pCtuf-ppmch and SYTO 9-labeled P. aeruginosa were used in this study. Staphylococcus aureus and P. aeruginosa were subcultured for 4 h. Then, bacterial pellets were collected, mixed, and suspended in 200 μl PBS. The bacterial suspension was inoculated on TSA plates, and the inoculum was allowed to dry at room temperature. Caenorhabditis elegans at the L4 stage were placed onto bacteria-containing TSA plates and incubated at 20°C for 30 min. The worms were then transferred and mounted onto 2% agarose pads on a slide covered with 1% levamisole (Sigma-Aldrich, United States) to paralyze the worms. Bacteria in the digestive tracts of C. elegans were observed and imaged under a confocal laser-scanning microscope at 0 and 60 min after feeding, worms were imaged under a Olympus FV10i microscope. Meanwhile, the images were captured at 30 s interval for 5 min and converted into a time-lapse video. The distance of bacterial migration was measured and analyzed using FV10-ASW 4.2 software (Olympus).

Six- to eight-week-old BALB/c mice were used in this study. The animal experiments were approved by the Chang Gung University Animal Care Committee (approval no. CGU108-191) and performed in accordance with the approved animal care guidelines and protocols. To improve the imaging quality, the backs of the mice were shaved. Mice were anesthetized by isoflurane inhalation and 50 μl of bacterial suspension, which contained 1 × 10⁷ CFU of luminescent S. aureus CGL1190 and 5 × 10⁷ CFU of P. aeruginosa PAO1, was injected subcutaneously into the dorsal region of the animals. The mice were then imaged with an IVIS instrument at 0 and 6 h post-injection. After observation, the mice were sacrificed, and the skin tissues at the injection site and the areas 2 cm away from the injection site were collected using a single hole puncher. The skin tissues were homogenized in 1 ml sterile PBS followed by centrifugation at 500 ×g for 5 min. The supernatant was serially diluted and plated on mannitol salt agar plates. The number of luminescent S. aureus was determined by viable cell count.

The between-group differences were determined using two-tailed Student’s t test and one-way analysis of variance (ANOVA) test. According to the Shapiro–Wilk’s test, the results of C. elegans studies did not exhibit a normal distribution; the Kruskal-Wallis test was further used for statistical analysis of the C. elegans studies. All statistical analysis were performed using GraphPad Prism software version 8.0 (GraphPad Software, San Diego, California, United States).

An earlier study demonstrated that S. aureus achieves movement in a liquid environment by interacting with a swimming bacterium (Samad et al., 2017). A hitchhiking motility assay (Samad et al., 2017) was modified and used in this study to measure the ability of P. aeruginosa to carry S. aureus during swimming. We added a mixture of S. aureus and P. aeruginosa into the wells of a CBD. After the bacterial mixture was stratified into two different layers with motile bacteria on the top and nonmotile bacteria at the bottom in the wells, we collected the bacteria on the top using the pegs on the lid of the CBD for CFU enumeration (Figure 1A). The results showed that when S. aureus SA113 was incubated in wells alone, the average CFU of S. aureus SA113 that adhered to a peg was 173. The CFU of bacteria on the pegs increased 17-fold to 2,966 CFU when S. aureus was incubated with P. aeruginosa PAO1 (Figure 1B). We also used a ΔfliA mutant of P. aeruginosa PAO1, which lacks flagellum. Compared with flagellated P. aeruginosa PAO1, the mutant carried 100 times less S. aureus SA113 to the pegs (Figure 1C), indicating that adherence of S. aureus to the pegs depends on the motility of P. aeruginosa. These results were consistent with an earlier finding that S. aureus is carried by P. aeruginosa to migrate (Samad et al., 2017). The attachment of P. aeruginosa PAO1 and S. aureus SA113 to the pegs was also verified by SEM (Figure 1A).

The interaction between S. aureus and P. aeruginosa suggested that surface components of S. aureus were involved in the hitchhiking motility. Our study found that the number of S. aureus that adhered to a peg was reduced after deleting tagO from strain SA113, which is required for the synthesis of wall teichoic acids (WTA; Weidenmaier et al., 2004). The results showed that when cocultured with P. aeruginosa PAO1, the average CFU of S. aureus SA113 that adhered to a peg was 3.1 × 10³ (Figure 2A). However, after tagO was deleted, the average CFU of bacteria adhered to a peg reduced 4.3-fold to 7 × 10² (Figure 2A). The number increased to 3.8 × 10³ CFU per peg when the mutant strain was transformed with pHY-tagO, a plasmid expressing TagO (Figure 2A), showing the importance of WTA in the hitchhiking motility. Similar results were also observed by SEM (Figure 2A, lower panel).

The study further examined whether P. aeruginosa carried S. aureus during swarming. It is known that S. aureus spreads on soft agar surfaces, a movement resembles swarming (Kaito and Sekimizu, 2007). As S. aureus SA113 does not spread due to its agr-quorum sensing system is defective (Herbert et al., 2010; Tsompanidou et al., 2011), the movement of S. aureus SA113 during P. aeruginosa swarming could not be attributed to the spreading of S. aureus. Two microliter bacterial culture that contained 1 × 10⁴ CFU of S. aureus and 1 × 10⁶ CFU of P. aeruginosa was spotted at the center of a TSA-0.4 plate. Four hours after inoculation on agar medium with a mixture containing S. aureus SA113 and P. aeruginosa PAO1, the bacteria at the center and the edge were picked with a toothpick (Figure 2B). The results showed that the average CFU of S. aureus SA113, ΔtagO mutant of SA113, and the ΔtagO mutant of SA113 transformed with pHY-tagO at the center were similar (Figure 2C). However, at the edge of the swarm, average CFU of S. aureus SA113, the ΔtagO mutant, and the mutant harboring pHY-tagO were 251, 20 and 158 CFU, respectively (Figure 2C). The results indicated that S. aureus SA113 does not migrate to the edge of a swarm efficiently after tagO is deleted. According to the results from Figures 1, 2A, P. aeruginosa interacted with WTA of S. aureus and carried S. aureus to move around in the liquid environment, which facilitated S. aureus up to top layer and adhered to pegs. The phenomenon of WTA dependent carriage between P. aeruginosa and S. aureus is also observed on surface. Meanwhile, the whole swarm colony was harvested for CFU counting. The CFU of S. aureus increased from 1 × 10⁴ to 1 × 10⁵, indicating that during co-incubation with P. aeruginosa for 4 h, S. aureus still multiplied.

An earlier study showed that Bifidobacterium interacts with the LPS of E. coli to neutralize its toxicity (Park et al., 2007), implying that LPS is a potential candidate involved in associations with Gram-positive bacteria. To investigate the components of P. aeruginosa that interacted with WTA of S. aureus, free LPS from P. aeruginosa was used in the hitchhiking motility assay according to the method described in Figure 1A. The results showed that when free LPS was mixed with S. aureus before coincubation with P. aeruginosa, the number of S. aureus that adhered to a peg decreased in a dose-dependent manner (Figure 3A). Average CFU of S. aureus that adhered to the pegs was 3.6 × 10³. The CFU decreased to 2.4 × 10³, 1.1 × 10³, and 0.4 × 10³ when the bacteria were cultured in TSB containing 0.1, 1, and 10 μg/ml free LPS, respectively (Figure 3A). The results indicated that adding LPS competes for the binding of S. aureus to P. aeruginosa. Flow cytometry analysis also verified the binding of FITC-LPS to S. aureus (Figure 3B). The results demonstrated that the mean fluorescence intensity (MFI) of S. aureus SA113 alone was 19. The MFI increased to 253 when S. aureus SA113 was incubated with FITC-LPS. When unlabeled LPS was added, the MFI of FITC-LPS bound S. aureus decreased to 186 (Figure 3C). To assess whether WTA of S. aureus was required for the binding to LPS of P. aeruginosa, free LPS, hexidium iodide-labeled S. aureus SA113, ΔtagO mutant of SA113, and the mutant transformed with pHY-tagO and SYTO 9-labeled P. aeruginosa were used in flow cytometry analysis. The results demonstrated that when S. aureus SA113 was mixed with P. aeruginosa at a ratio of 1:1, the average percentage of P. aeruginosa associated with S. aureus SA113 was 32 (Figure 3E). When tagO was deleted from strain SA113, the average percentage of P. aeruginosa associated with the mutant reduced to 21.8 (Figure 3E); but increased to 34.7 when the mutant strain was transformed with pHY-tagO (Figure 3E). We also added free LPS to the mixture to determine whether LPS competes for the binding of S. aureus to P. aeruginosa. The results showed that compared with the strains untreated with LPS, the addition of free LPS decreased the binding percentage of P. aeruginosa to S. aureus SA113, the SA113 ΔtagO mutant and the mutant transformed with pHY-tagO from 32, 21.8, and 34.7 to 16, 6.3, and 18, respectively (Figure 3E). Furthermore, we examined whether lipoteichoic acids (LTA), which is a teichoic acid that anchored to the membrane lipid, is involved in the interaction between S. aureus and P. aeruginosa. When free LTA from S. aureus was used in the hitchhiking motility assay, the similar results were obtained as that of LPS (Figure 3D). The results showed that adding free LTA competes for the binding of S. aureus to P. aeruginosa in a dose-dependent manner (Figure 3D). Flow cytometry analysis also confirmed the percentage reductions after adding LTA. The results demonstrated that adding free LTA decreased the average percentage of the binding of P. aeruginosa to S. aureus SA113, the SA113 ΔtagO mutant, and the mutant transformed with pHY-tagO from 32, 21.8, and 34.7 to 10.4, 7.9, and 10.8 (Figure 3E), verifying that P. aeruginosa interacts with S. aureus via the interaction of LPS with WTA or LTA (Figure 3E). However, in comparison with S. aureus SA113, addition of free LTA to the ΔtagO mutant of SA113 or the mutant transformed with pHY-tagO, the differences of the binding of P. aeruginosa to S. aureus is not statistically significant between the groups (Figure 3E). The results suggested that WTA and LTA target different components on the P. aeruginosa surface for interaction.

After demonstrating that P. aeruginosa PAO1 promotes the movement of S. aureus SA113 on surface, this study further investigated whether P. aeruginosa facilitates the movement of S. aureus in a C. elegans model. In this study, S. aureus SA113, the SA113 ΔtagO mutant and the mutant with pHY-tagO were transformed with pCtuf-ppmch to express mCherry (Mauthe et al., 2012). Meanwhile, P. aeruginosa PAO1 was stained with SYTO 9 to exhibit green fluorescence.

Caenorhabditis elegans was fed with bacteria and the movement of the bacteria in the digestive track was followed and imaged. The images revealed that when the worms were fed with S. aureus SA113, little movement of the bacteria was observed; the bacteria at the same location during an hour; the bacteria moved only 0.3 μm within an hour (Figures 4Aa–f, C). When the worms were fed a mixture of S. aureus SA113 and P. aeruginosa PAO1, S. aureus SA113 moved along with P. aeruginosa PAO1 with a distance of 6.0 μm (Figures 4Ba–f, C). When the worms were fed with a bacterial mixture containing the ΔtagO mutant of SA113 and P. aeruginosa PAO1, little movement of the mutant was observed; the mutant moved only 0.9 μm (Figures 4Bg–l, C). However, the mutant was found to co-migrate with P. aeruginosa PAO1 and reached a distance of 4.3 μm at 1 h after feeding if the mutant was transformed with pHY-tagO to complement the mutation (Figures 4Bm–r, C). Time-lapse video showed that S. aureus (red) moved together with P. aeruginosa (green; Supplementary Videos S1, S2). However, due to prolonged exposure to laser damaged the digestive tract and was lethal to the worm, this study was unable to collect the results beyond the 1 h period.

Staphylococcus aureus and P. aeruginosa are commonly associated with wound infection (DeLeon et al., 2014). To clarify whether P. aeruginosa promotes spreading of S. aureus in skin tissues, we used IVIS and observed the movement of S. aureus CGL1190, a strain exhibiting bioluminescence, in mice. The results showed that bioluminescence signals were present in the injection site, indicating that the bioluminescent S. aureus was useful in a mouse skin model (Figures 5Aa,c). Compared to inoculations with S. aureus alone, S. aureus spread farther when co-inoculated with P. aeruginosa (Figures 5Ab,d). To quantify the number of S. aureus moved away from the original injection site, the skin tissues at the inoculated areas and the areas 2 cm away were removed from the mice to enumerate the bacteria in the regions. The results showed that at 6 h after inoculation with S. aureus SA113 alone, the tissue in the area 2 cm away from the inoculation site contained only a few S. aureus, the average CFU was 3 (Figure 5B). In contrast, when inoculated with the bacterial mixture containing S. aureus SA113 and P. aeruginosa PAO1, the average CFU of S. aureus in the skin tissue 2 cm away from the inoculation site was 1.9 × 10⁴ (Figure 5B). However, the average CFU of S. aureus was 1.5 × 10³ CFU when inoculated with a mixture containing the ΔtagO mutant and P. aeruginosa PAO1 (Figure 5B). The results indicated that P. aeruginosa promotes the movement of S. aureus in the mouse skin model. The results also showed that WTA is required for hitchhiking P. aeruginosa by S. aureus SA113 in vivo. Furthermore, we plated 50 μl blood collected from the mice at 6 h post-inoculation, on mannitol salt agar plates and did not detected any S. aureus and P. aeruginosa in the blood. Additionally, our IVIS study revealed that the bioluminescence was largely confined near the tail, rather than the whole body suggesting that bacteria movement does not involve the dispersion of the bacteria via the circulatory system. The results excluded the possibility that the bacterial dissemination was due to the entry of the bacteria into the bloodstream.