Modulation of Staphylococcus aureus spreading by water

Staphylococcus aureus is known to spread rapidly and form giant colonies on the surface of soft agar and animal tissues by a process called colony spreading. So far, the mechanisms underlying spreading remain poorly understood. This study investigated the spreading phenomenon by culturing S. aureus and its mutant derivatives on Tryptic Soy Agarose (TSA) medium. We found that S. aureus extracts water from the medium and floats on water at 2.5 h after inoculation, which could be observed using phase contrast microscopy. The floating of the bacteria on water could be verified by confocal microscopy using an S. aureus strain that constitutively expresses green fluorescence protein. This study also found that as the density of bacterial colony increases, a quorum sensing response is triggered, resulting in the synthesis of the biosurfactants, phenolic-soluble modulins (PSMs), which weakens water surface tension, causing water to flood the medium surface to allow the bacteria to spread rapidly. This study reveals a mechanism that explains how an organism lacking a flagellar motor is capable of spreading rapidly on a medium surface, which is important to the understanding of how S. aureus spreads in human tissues to cause infections.


Results
Spreading of S. aureus HG001. S. aureus is known to spread on the surface of TSA medium containing 0.24% agar (TSA-0.24) 1 . Since S. aureus does not have a flagellum or any other surface apparatus that would allow the bacterium to power its movement, we posit that S. aureus uses water and moves passively across the medium surface during spreading. S. aureus spreading is commonly studied on TSA-0.24 agar 1,6,7 . A medium with agar at this concentration contains large amounts of water, therefore, studying water extraction by S. aureus on a plate surface that is already saturated with water can be difficult. Additionally, types of agar used to prepare TSA were found to influence not only the ability but also the speed of S. aureus HG001 spreading. To determine if a plate with a relatively dry surface could be utilized for assaying water extraction, we prepared TSA plates with agarose (SeaKem) to determine how the concentration of agarose and volume of the medium affected the spreading of S. aureus HG001. We found that when the strain was cultured on a 15-ml TSA plate that contained 0.25% agarose (TSA-0.25) in a 9-cm petri plate, S. aureus HG001 spread and its colony covered almost the entire plate with numerous tendrils developing at the colony's edge after 24 h (Fig. 1a). When the agarose concentration was increased to 0.3% (TSA-0.3), the ability of the bacterial spreading was reduced; the size of the colony on this plate was approximately 1.5 cm in diameter (Fig. 1b). When the percentage of agarose was increased further to 0.35% or 0.4% (TSA-0.35 or TSA-0.4, respectively), the strain did not spread (Fig. 1c,d). We also found that increasing the volume of TSA medium in the plate could compensate for the inability of S. aureus to spread on plates containing a high percentage of agarose. When the volume of the medium was increased from 15 ml to 20 ml, the bacteria formed a colony that was larger and thicker than that formed on a 15-ml TSA-0.25 plate (Fig. 1a,e), and the bacteria were able to spread on TSA-0.3 and TSA-0.35 plates (Fig. 1f,g). We also found that the bacteria were unable to spread on a TSA-0.4 plate with 20 ml of TSA medium (Fig. 1h). When the volume of TSA was increased to 25 ml, the bacteria spread on TSA plates that contained 0.25-0.4% agarose ( Fig. 1i-l). Therefore, in this study, we used 25-ml TSA-0.4 plates for studies on water extraction by S. aureus.
Phase contrast microscopic observation of S. aureus HG001 colonies. We observed S. aureus HG001 colonies formed on 25-ml TSA-0.4 plates under a phase contrast microscope at a magnification of 400x. We found that 20 min after inoculation, the colony was flat and bacteria were distributed sparsely within the colony. Many bacteria were found oscillating, typical of Brownian movement (Supplementary Video S1), suggesting that a small amount of water was present in the colony causing the movement. At 2.5 h after inoculation, the morphology of the colony changed significantly. Under the microscope, bacteria at the center of the colony were found in different depths. In the colony, a few bacteria at the bottom appeared to be stationary; those on the top were moving (Supplementary Video S2), showing the presence of water in the colony. Most importantly, although the bacteria were moving in water, they moved passively as the bacteria were all moving in a general direction at the same speed (Supplementary Video S2), suggesting that bacteria in the colony were floating and drifting in flowing water. After the bacteria had been cultured for 4 h, we found that a colony was densely packed by bacteria but were still moving in the same general direction (Supplementary Video S3).
Generating mutants that are defective in quorum sensing and spreading. We have mutagenized S. aureus HG001 using the bursa aurealis transposon 17 . Among the 2000 mutants screened, four mutants did not expand their colony on 25 ml TSA-0.25 plates after 24 h of culturing. Sequencing of these mutants revealed that mutants CGL005 and CGL006 had a transposon insertion in agrD ( Supplementary Fig. S1) and agrC, respectively, which are genes involved in activating quorum sensing in S. aureus 18 . We found that CGL005 did not spread on TSA-0.25 and TSA-0.4 plates (Fig. 2Ab,i). After CGL005 was transformed with pGH-agr, which contains agrD, the spreading ability of CGL005 was restored (Fig. 2Ad,k). Furthermore, CGL005 expanded its colony on the plate that contained PSMα 3, which was shown to be an important PSM for staphylococcal spreading 6 ( Fig. 2Ba-d). Since transcription of psm genes is activated by the Agr quorum-sensing system 9,19 , the addition of synthetic PSMα 3 necessary for spreading showed that the lack of PSM synthesis by CGL005 was responsible for the mutant phenotype. This study also generated a mutant, CGL007, which has its psmα operon deleted, to demonstrate the importance of PSMα to spreading. The spreading ability of CGL007 on TSA-0.25 and TSA-0.4 plates was impaired (Fig. 2Ae,l), as compared with the spreading activity exhibited by the wild-type strain HG001 (Fig. 2Aa,h), due to the lack of PSMα synthesis. The spreading ability of CGL007 was restored when CGL007 was transformed with pGH-psm, which contains the whole psmα operon (Fig. 2Ag,n), but was not restored after the cells were transformed with an empty vector, pGHL7 (Fig. 2Af,m). Moreover, the spreading of CGL007 was promoted by adding PSMα 3 onto TSA-0.25 and TSA-0.4 ( Fig. 2Be-h). These results demonstrated that the Agr quorum-sensing system, which controls PSMs synthesis, is important to spreading by S. aureus HG001. The results are consistent with that reported previously 5-8 . Presence of water in colonies. We cultured S. aureus strains HG001 and CGL005 on 25 ml TSA-0.4 plates and tilted the plates at a 30° angle during incubation. We found that at Hour 3, water started to flow and carry bacteria out of the HG001 colony (Fig. 3b); flowing of water was not observed for the CGL005 strain (Fig. 3e). This was not attributed to the lack of water in the mutant colony, since under a confocal laser-scanning microscope, the present of water in the colony was evident (data not shown). At Hour 6, a significant amount of water flowed out of the HG001 colony and accumulated at the bottom of the plate (Fig. 3c). However, only relatively small amounts of bacteria flowed out of CGL005 colony with no water accumulating at the bottom of the plate (Fig. 3f). These results showed that without an activated quorum sensing system resulting in the inability to synthesize PSMs, water surface tension holds water and prevents water from flowing out of the CGL005 colony.

Confocal laser-scanning microscopy of colonies formed by S. aureus.
To measure the height of colonies, strains expressing GFP constitutively were inoculated on 25 ml TSA-0.4 plates and cultured at 37 °C. At each time point, the edge in the colony on the plate was observed (Fig. 4Aa). The Z stacks in which the planes were separated by 10 μ m were acquired from the top to the bottom of the colony using an upright laser-scanning confocal microscope (Leica, TCS-SP2) (Fig. 4Aa). Green fluorescence in the micrographs indicated the distribution of bacteria in the spreading colony. The 0.5 μ m carboxylate-modified FluoSphere beads (Ex580/Em605) (Invitrogen) were also incorporated into the medium to demarcate the plate surface. An illustration showed images of top and side views from the Z stacks (Fig. 4Ab,c), which were generated with Imaris software (Bitplane).  We cultured S. aureus HG001(pRPO-gfp) on 25 ml TSA-0.4 plates. Z stacks of the colonies were acquired under a confocal laser-scanning microscope (Fig. 4Ba). We found that 20 min after inoculation, the colony was thin and approximately 10 μ m high (Fig. 4Ba). At Hours 1 and 2, growth of the bacteria was evident as the intensity of green fluorescence in the colony increased substantially. Meanwhile, the height of the colony increased to 30 μ m (Fig. 4Ba). At Hours 3 and 4, the height of the colony increased to 40 μ m (Fig. 4Ba); at Hour 5, the height of the colony increased to 50 μ m (Fig. 4Ba). We found that at Hour 5 after inoculation, the colony formed by CGL005(pRPO-gfp) was 70 μ m high with an intensity of green fluorescence substantially stronger than that of HG001(pRPO-gfp) (Fig. 4Bb). Macroscopic observation also showed that the HG001(pRPO-gfp) colony expanded but CGL005(pRPO-gfp) expanded very little during the five-hour period. The results indicated that the cell density and the height of colony formed by CGL005(pRPO-gfp) are higher than that formed by HG001(pRPO-gfp), which is due to a lack of PSMs synthesis to reduce the water surface tension of the colony and to restrict the colony expansion. We also examined the colonies formed by CGL007(pRPO-gfp), which has a defect in synthesis of PSMα . At Hour 5, the height of the colony was 60 μ m (Fig. 4Bc), which was lower than the colony formed by CGL005(pRPO-gfp) but higher than HG001(pRPO-gfp).

3-D analysis of the S. aureus HG001 colonies.
We generated 3-D images of S. aureus HG001 colonies using the confocal 2-D images shown in Fig. 4Ba and Imaris software (Bitplane). We found that 20 minutes after inoculation, bacteria were distributed sparsely inside the colony with a ring of accumulation at the colony edge (Fig. 5a). At Hour 1, bacterial growth was evident as more green fluorescence was found within the colony (Fig. 5d). At Hour 2, most of the areas in the colony were covered by bacteria (Fig. 5g). A thick layer of bacteria developed after Hour 3 (Fig. 5j,m,p). Images generated from the side of the colony revealed that bacteria were attached to the plate surface at 20 min after inoculation (Fig. 5b,c). At Hour 1, a space of several μ m separated the green bacteria and the red medium (Fig. 5f), showing that bacteria were floating. During Hours 2-5, the green layer thickened and the distance between bacteria and the plate surface increased (Fig. 5g-r). These results showed that bacteria both grow and float in colonies.

Quorum sensing and S. aureus spreading.
Earlier studies demonstrated that mutants with mutations in agr genes do not spread on soft agar [6][7][8] , suggesting that quorum sensing is required for colony spreading of S. aureus. It is well known that expression of PSMs is activated by Agr quorum-sensing system 9,19 . However, it is currently unknown whether the quorum sensing response occurs in a spreading colony to modulate the spreading behavior. Therefore, a reporter plasmid pPSM-gfp, which contains the promoter of the psmα operon, which requires quorum-sensing activator (AgrA) for transcription, fused with a gfp reporter gene, was used in this study. We applied 2 μ l log phase HG001(pPSM-gfp) cells (1 × 10 7 CFU) to TSA-0.4 plates. We found that cells did not exhibit much green fluorescence before Hour 2 under a confocal laser-scanning microscope (Fig. 6Aa-c). However, strong green fluorescent signals were detected after Hour 3 (Fig. 6Ad-f). We also analyzed GFP expression by immunoblotting. To obtain enough proteins for the analysis, we had to inoculate 100 μ l of the cells, which contained 5 × 10 8 CFU, on one plate. The cells were harvested and homogenized. The lanes in a gel was loaded with 5 μ g of the cell lysate. The results showed that 20 min after inoculation, the amount of GFP was relatively high (Fig. 6B, lane 1). This could be due to the long half-life of GFP and GFP being present in the cells prior to inoculation. However, the levels of GFP decreased at Hours 1 and 2 (Fig. 6B, lanes 2,3) but increased after Hour 3 (Fig. 6B, lanes 4-6). These results suggest that quorum sensing occurred before or at Hour 3, resulting in the activation of the psm promoters.

Discussion
S. aureus is known to spread on a soft agar surface with an astonishing speed of approximately 100 μ m per minute 1 , which parallels or exceeds the swarming speeds of many bacteria 20 . However, the mechanism of spreading by S. aureus differs from that of swarming because S. aureus lacks a flagellar motor to drive itself 1 . Therefore, the purpose of this study was to elucidate the underlying mechanism by which S. aureus spreads on a soft agar surface.
Although the work by Kaito and Sekimizu 1 showed that S. aureus spreads only on TSA-0.24 plates, we found that S. aureus HG001 also spreads on TSA-0.4 plates (Fig. 1). We found that on this type of plate, S. aureus HG001 formed a colony filled by water (Supplementary Video S2) with a height of approximately 30-40 μ m as measured by confocal microscopy (Fig. 4Ba). Additionally, water flowed out of the colony formed on a TSA-0.4 plate after the plate was tilted during incubation (Fig. 3a-c), verifying the presence of water in the colony. The results suggest that S. aureus HG001 is capable of extracting water from the medium. Earlier studies showed that the synthesis of PSMs is required for S. aureus to spread [5][6][7] . Therefore, we generated an agrD mutant, CGL005, which is unable to synthesize PSMs. We found that the colony formed by CGL005(pRPO-gfp) at Hour 5 was at least 40% or 20 μ m higher than that formed by HG001(pRPO-gfp) (Fig. 4Ba,b). Due to a lack of colony expansion, the green fluorescence in the CGL005(pRPO-gfp) colony was more intense with a higher bacterial density than the green fluorescence in HG001(pRPO-gfp) (Fig. 4Ba,b). The defect in spreading of CGL005 can be complemented with a plasmid that expressed AgrD ( Fig. 2A), or accomplished by adding synthetic PSMα 3 to the plate (Fig. 2B). Furthermore, less water flowed out of the CGL005 colony than that from a HG001 colony after 3 and 6 h of incubation in tilted plates (Fig. 3d-f), showing that PSMs are required to weaken the water surface tension to allow the water to flow. The impact of PSMs in reducing the water surface tension and colony spreading were also verified using the psmα deletion mutant CGL007 (Figs 2A and 4Bc). These phenomena are strikingly similar to colonies formed by a swarming bacterium, B. subtilis F29-3. When this bacterium is cultured on LB agar containing 0.7% agar, the bacterium extracts water from the medium and uses surfactin to cause water to flood agar surface to facilitate swarming 16 . This study suggests that S. aureus HG001 uses a similar mechanism to spread on TSA.
Since S. aureus lacks a flagellum and an apparatus that would allow the bacteria to move actively as the swarming bacteria do, S. aureus likely moves passively on TSA medium. If bacteria move actively, they should move in different directions with different speeds in a fluid filled environment. However, in a colony that was formed on TSA-0.4 medium, all the S. aureus HG001 cells were moving in one general direction with the same speed (Supplementary Video S2, S3). These results suggest that S. aureus, rather than moving actively, was carried by flowing water across the water surface. Meanwhile, we found that S. aureus was attached to the surface of TSA-0.4 medium. S. aureus gradually floated in a colony after being cultured for more than 1 h (Fig. 5). This finding also suggested the bacteria lack active movement, resulting in floating and drifting in water.
The force that creates water flow likely comes from a continuous water extraction by the bacteria and a weakening of the water surface tension of a spreading colony by PSMs. The transcription of PSM genes is activated by quorum sensing 9,19 and is required for spreading [5][6][7] , suggesting that quorum sensing occurs in a spreading colony. Our confocal laser-scanning microscopic work showed that GFP is not expressed much by the colony of HG001(pPSM-gfp) at Hours 1 and 2 but is strongly expressed after Hour 3 (Fig. 6A). Immunoblot analysis also verified that the expression of GFP started to increase at Hour 3 (Fig. 6B), suggesting that quorum sensing and transcription of the PSM genes start before or at Hour 3. It is likely that cells in the entire colony are transcribing psmα as an earlier study showed that agrP3 activity was similar among the cells between the center and edge in a spreading colony 21 , thus the expression of PSMs in different areas of the spreading colony may be similar. The edge of the colony showing stronger green signal (Fig. 5) may be due to the accumulation of cells by the coffee ring effect 22,23 .
Based on the results from this study, we propose a mechanism to explain how S. aureus spreads. This mechanism involves a continuous water extraction from the medium. As the bacterial density of the colony increases, quorum sensing occurs, which triggers the transcription of psm genes and the synthesis of PSMs. These biosurfactants weaken the water surface tension of the colony allowing water to flow. The water flow then carries the bacteria and causes a rapid spreading of the bacteria across the agarose surface. In nature, an ability to move rapidly gives bacteria advantages to obtain food 24 and is an important virulence factor for many pathogens [25][26][27][28] . An earlier study showed that S. aureus is capable of spreading on fresh pork meat 6 , showing that this organism is able to spread on soft tissues. Our study reveals how S. aureus spreads, which is important in the understanding of how this pathogen moves in human tissues to cause diseases.

Materials and Methods
Strains and culture conditions. S. aureus HG001 29 , a derivative of S. aureus NCTC8325, is able to spread on soft agar. S. aureus CGL005 is a mutant strain derived from strain HG001, which has a transposon inserted in agrD ( Supplementary Fig. S1). S. aureus CGL007 is a mutant strain with a deletion in psmα operon. Tryptic soy broth (TSB) and tryptic soy agar (TSA) were used to culture S. aureus. TSA plates were prepared 20 min before inoculation according to the method of Kaito and Sekimizu 1 , except that the plates were prepared with agarose from Seakem (Lonza Rockland, Inc.). Bacteria were cultured in TSB overnight, and washed in PBS. The plates were inoculated with 2 μ l (1 × 10 7 CFU) of the bacteria and then cultured at 37 °C. Strains containing pRPO-gfp or pPSM-gfp were cultured on TSA plates supplemented with 10 μ g/ml chloramphenicol. PSMα 3 was synthesized chemically by Kelowna International Scientific Inc. (Taiwan) and was applied (2 μ l, 1 mM) to the center of 25 ml TSA-0.4 plates. After drying for 30 min in a hood, the plate was inoculated with CGL005 or CGL007 at the same location to demonstrate that PSMα 3 is required for spreading by the mutant strains.
Transposon mutagenesis. To identify the genes that are involved in colony spreading, S. aureus HG001 was mutagenized with a mariner-based transposon, bursa aurealis, according to a method described elsewhere 17 . The transposon contains the short inverted repeats of the hornfly transposon and the ermB resistance marker. Insertion of bursa aurealis into the genome resulted in resistance to erythromycin. S. aureus HG001 were sequentially transformed with pBursa and pFA545, which contains genes encoding mariner transposase and confers resistance to tetracycline and ampicillin. The resulting transformants were spread on TSA containing erythromycin (TSAerm) and incubated at 30 °C for 24 hr. After induction of transposase expression and curing the plasmid by incubating the transformants at 42 °C for 2 days, cells containing transposon insertion in the chromosome were selected on TSAerm. Approximately 2000 mutants were collected. Mutants that were defective in spreading were selected on TSA medium that contained 0.25% agarose (TSA-0.25). The site of transposon insertion was determined by inverse PCR using primers that were complementary to the border sequences of the transposon according to a method described elsewhere 17 . The amplified DNA fragment was then sequenced to identify the location of transposon insertion, using the genome sequence of S. aureus NCTC8325 (accession number: NC_007795.1) 30 , which is the parental strain of S. aureus HG001.

Construction of PSM mutant strain of S. aureus.
To generate psmα deletion mutant, gene replacement using the temperature-sensitive plasmid pMAD was performed according to the method described elsewhere 31 . A 1.3-kb DNA fragment containing the entire spectinomycin-resistance gene spc 31 was amplified using the primer Spc-F (5′ -CCGCTCGAGAGTAGTTCACCACCTTTTC) and Spc-R (5′ -CGCCCCGGGTTTATTGTTTTCTAAAATC). The PCR product was digested with XhoI and SmaI and inserted into the same sites in pGEM-7Z to generate p7Z-spc. The upstream and downstream regions of psmα operon were amplified by PCR using the primer pairs Up-F1 and Up-R1 (5′ -GATGAGCTCATAATGTAATACCCCAGCAG and 5′ -CTACCCGGGATAGTTATCTTGTGCGTAAT), Scientific RepoRts | 6:25233 | DOI: 10.1038/srep25233 Dn-F1 and Dn-R1 (5′ GAACTCGAGTTCTCAGGCCACTATACCAA and 5′ GCGGCATGCACAATAC AATCACGTAGCAT), respectively. The PCR products were cut using SacI, SmaI, XhoI and SphI, and inserted into the SacI-SmaI and XhoI-SphI sites in p7Z-spc to generate p7Z-007. The fragment containing spc flanked with the upstream and downstream regions of psmα operon was amplified from p7Z-007 using primer Up-F2 (5′ -GATGTCGACATAATGTAATACCCCAGCAG) and Dn-R2 (5′ -GCGAGATCTAC AATACAATCACGTAGCAT). The PCR product was digested with SalI and BglII and inserted into the same sites in pMAD to generate pMAD-Δ psmα . Allelic replacement of the psmα operon by a spectinomycin cassette (spc) was performed by introducing the pMAD-Δ psmα into S. aureus strain HG001 to generate the psmα mutant strain CGL007 according to the method described previously 31 . Plasmid construction. The promoter of rpoD gene of B. subtilis was amplified by PCR, using primers 5′ -CGCGTCTAGAATCAACAGAATCAAAGAGGA and 5′ -CGCGGGATCCATGTATATGAATTTGTCGAA. The amplified fragment was cut with BamHI and XbaI and inserted at the BamHI-XbaI sites in pRU-gfp 16 to generate pRU-RPOD. A DNA fragment that contained the rpoD promoter along with the gfp sequence in pRU-RPOD was isolated by EcoRI and XbaI digestions, and the fragment was inserted into the same sites in a shuttle vector pGHL6 32 to generate pRPO-gfp. A DNA fragment that contained the psmα promoter was amplified using primers 5′ -GGTCTAGAATGAGCTTAACCTCTATTAA and 5′ -CAAGGATCCGCTTATGAGTTAACTTCAT, and S. aureus HG001 DNA was used as a template. The DNA fragment was cut by BamHI and XbaI and then inserted into the BamHI-XbaI sites in pRPO-gfp to generate pPSM-gfp. To complement the mutants CGL005 and CGL007, pGH-agr and pGH-psm were constructed. The vector pGHL6 was cut with EcoRI and self-ligated to generate pGHL7, which was used to construct pGH-agr and pGH-psm. A fragment containing agrB, agrD, agrC and their promoter was amplified by PCR, using primers Agr-F (5′ -AGCTTCTAGACTGAGTCCAAGGAAACTAAC) and Agr-R (5′ -AGCTGAATTCGGATCGTCTTCGCAAATGAA) and S. aureus HG001 chromosome as a template. The PCR product was then cut with XbaI and EcoRI, and inserted into the XbaI-EcoRI sites in a vector, pGHL7 to generate pGH-agr. The psmα operon with its promoter was amplified by PCR, using primers PSM-F (5′ -GGTCTAGAATGAGCTTAACCTCTATTAA) and PSM-R (5′ -GGAGATCTTT AAGTATTCAATTCAATTCGCTTA) and the S. aureus HG001 chromosome as a template. The PCR product was cut with XbaI and BglII and inserted into the XbaI-BglII sites in pGHL7 to generate pGH-psm.
Microscopy. Bacterial colonies were observed under a Zeiss Axioskop phase contrast microscope at a magnification of 400x. Bacterial movement was recorded using a Sony NEX-5N camera. To measure the height of colonies, S. aureus HG001(pRPO-gfp), S. aureus CGL005(pRPO-gfp) and S. aureus CGL007(pRPO-gfp) were inoculated on 25 ml TSA-0.4 plates that contained 1/100th of its volume of 0.5 μ m carboxylate-modified FluoSphere beads (Invitrogen, Ex580/Em605) and cultured at 37 °C for indicated time. At each time point, one plate was moved out from the incubator, cover removed, and placed on the stage of an upright microscope for observation. Different plates were used for the indicated time point. A Z stack in which the planes were separated by 10 μ m was acquired using an upright laser-scanning confocal microscope (Leica, TCS-SP2). Green fluorescence indicated the distribution of bacteria in the spreading colony. The last plane showing red fluorescence was used to indicate the plate surface. The height of the colony was determined by green fluorescence using the method described elsewhere 16 . 3-D images of colonies were generated from the Z stacks using Imaris software.
Gravity experiments. To demonstrate that colonies contained water, plates were tilted at 30° during incubation to allow water and bacteria to flow out of the colonies according to the method described earlier 16 . Quorum sensing. TSA-0.4 plates were inoculated with 2 μ l PBS-washed log-phase bacteria (approximately