Surface sensing in Escherichia coli
Kimkes, Tom Eric Pieter
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2019
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Kimkes, T. E. P. (2019). Surface sensing in Escherichia coli. Rijksuniversiteit Groningen.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Surface sensing by the Rcs system initiates
adaptation of Escherichia coli to nutrient
limitation
Tom E.P. Kimkes and Matthias Heinemann
Molecular Systems Biology, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen, Nijenborgh 4, 9747 AG Groningen, the Netherlands
Abstract 85
Introduction 86
Results 87
Only the Rcs system is activated by surface contact 87 Induction of RcsB reporters by surface contact requires RcsF 89 Surface sensing via Rcs is involved in the persister phenotype 90 Activation of Rcs requires a large surface contact area 93
Discussion 94
Materials and methods 96
References 99
Abstract
Escherichia coli biofilms are involved in various illnesses, yet we still do not fully understand what triggers biofilm formation. To characterise the mechanism by which biofilm formation is initiated, the transcriptional response to surface contact was investigated. Specifically, the expression of target genes of eight candidate surface sensing systems was quantified in E. coli upon surface attachment in a microfluidic setup. Using this approach, it was established that only the Rcs system is activated on a surface. Further characterisation revealed that the auxiliary pathway component RcsF is required for surface sensing. Activation of Rcs was only observed when E. coli cells were immobilised between two materials, suggesting that a large contact area or mechanical pressure is required for induction, which would suggest a role for Rcs at the microcolony stage of biofilm formation with physical contact with other bacteria. Activation of the Rcs system via surface sensing was found to abolish the state of persistence induced by a previously reported carbon source shift, presumably by increasing the influx of gluconeogenic substrates via upregulation of the DctA transporter. Contact-dependent expression of transporters could be an effective strategy to ensure sufficient nutrient uptake when growing in a developing biofilm where nutrients are often scarce. Our results imply that the Rcs system could be a new target for biofilm prevention.
Introduction
Biofilm formation poses significant technological and medical problems. Biofilms are formed in the majority of bacterial infections [1] and complicate treatment, because of an increased tolerance to antibiotics [2–4]. Also, long-term prevention of biofilm formation on man-made surfaces without adverse side effects, such as toxicity, is challenging and high costs are associated with the regular cleaning of biofouled surfaces, such as the hulls of ships [5, 6].
Most microbiological research is concerned with either planktonic (suspended) cells or the later stages of biofilm formation, whereas research into the initial steps of sessile growth, i.e. the events occurring in the first several hours after adhesion, has been limited. Although we do know what bacteria may sense when adhering to a surface, we only have some indications concerning how they sense it. There are three signals that a surface may present to adhering bacteria: (i) the microenvironment at a solid-liquid interface is often different from the bulk liquid in terms of ionic strength, osmolarity, pH and nutrient availability [7–9]; (ii) attachment of bacteria via flagella, pili or curli exerts forces on these adhesive appendages, which may be perceived via mechanosensitive mechanisms [10–16]; (iii) adhesion may also result in envelope stress, although this has not been fully established [17]. While the model organism Escherichia coli has several signal transduction pathways that are able to perceive such signals, it is in most cases not known if and how they are involved in surface sensing and consequent biofilm initiation.
As reviewed in Chapter 1, the BaeSR, BarA-UvrY, BasSR, CpxAR, EnvZ-OmpR, PSP, RcsCDB and σEsystems might all play a role in surface sensing, as these systems
are sensitive to one or more of the signals conveyed by surface contact and they regulate target genes with a link to biofilm formation. A response to adhesion has only been reported for CpxAR [18] and RcsCDB [19], meaning that involvement of the other systems is only speculative. Considering the magnitude of biofilm-related problems, the knowledge about perception of surface contact is therefore disproportionately limited. With regard to the next steps in biofilm development, i.e. formation of a microcolony from single attached bacteria, a challenge faced by sessile bacteria is that they become immobilised in the growing biofilm and therefore cannot move towards more favourable environments when nutrients get limited, which is bound to happen within a growing biofilm [20]. Eventually, bacteria in the centre of the biofilm may not be able to cope with these low nutrient concentrations and will assume a low growth rate [21]. In order to prepare for the expected nutrient limitation, it would thus be an effective strategy for sessile cells to adapt their metabolism at an early stage. Indeed, a transcriptomics study comparing planktonic and sessile cells, found many upregulated genes with functions in metabolism and transport after four hours on a surface [22]. Also consistent with surface-induced changes in metabolism is the previously observed growth resumption of persister cells upon surface contact [23], because bacterial persistence can be the result of low metabolic activity [24]. Hence, it is plausible that E. coli adapts its metabolic machinery already at the stage of surface
sensing, but it is unclear how this is achieved.
Employing microfluidics and fluorescence microscopy, we investigated the activity of the above mentioned candidate surface sensing systems upon surface attachment. We found that the CpxAR system is not responsive to surface contact, which contradicts previous findings of other labs (see also Chapter 2). Only the Rcs system was found to respond to attachment. Further characterisation of the Rcs response indicated that surface sensing requires the outer membrane lipoprotein RcsF, as well as a large contact area with the surface. Our findings further suggest that activation of Rcs on a surface induced expression of the DctA transporter, increasing the ability to take up carbon sources other than glucose. Taken together, our results indicate that surface sensing by the Rcs system affects metabolism and that this likely happens at the stage of microcolony formation, where the surface contact area is larger due to contact with other bacteria.
Results
Only the Rcs system is activated by surface contact
To induce surface contact and to study the resulting initial responses of E. coli, bacteria carrying the transcriptional fluorescent reporter plasmids in Table 1 were grown exponentially in liquid medium and transferred to the microfluidic device shown in Figure 1. To keep the environmental conditions as constant as possible, the cells
Figure 1: The experimental setup in which surface-mediated induction of the reporters was tested. PAA: polyacrylamide; PDMS: polydimethylsiloxane.
were subjected to spent medium prepared from a culture with the same optical density, such that the absence or presence of a surface was the only difference between growth in the flask and in the microfluidic device. Following the introduction of the bacteria into the microfluidic setup, acquisition of phase contrast and GFP fluorescence images was started within ten minutes. In Supplementary Figure S1 the expression profiles of the reporter genes are shown in the first hours following the introduction of the bacteria into the microfluidic device. Most of the reporters for the potential surface sensing systems showed changes in expression level upon contact with a surface. Why the fluorescence intensities increased for most reporters is yet unclear. Overall, our experimental setup enabled the observation of the transcriptional response of E. coli to surface contact.
To distinguish between actual induction of the reporter genes and unspecific global effects on
Reporter Regulated by
ais BasSR [25]
bamD σE [26]
cadB EnvZ-OmpR [27]
crl BarA-UvrY [28]
csrB BarA-UvrY [29, 30], BasSR [31], CpxAR [32] fkpA σE [26, 33, 34]
frr
-luxS BarA-UvrY [35] mdtA BaeSR [32, 36, 37]
ompC EnvZ-OmpR [38], BasSR [31], CpxAR [39], σE (-) [33, 34]
osmB RcsCDB [19, 40] apFAB70 -plsB σE [33, 34] pspA PspF [41] pspF PspF (-) [42] rcsA RcsCDB [19, 43] rhsA RcsB [44], EnvZ-OmpR [27] rpoE σE [26, 33], RcsCDB [45] rpsM
-spy BaeSR [32, 46], CpxAR [32, 47, 48], RcsCDB [19, 32, 40] tolB σE [34], CpxAR [32], PspF [32]
tolC BaeSR [32], CpxAR [49] ugd RcsCDB [19], BasSR [31]
ybaY RcsCDB [32]
yccA CpxAR [32, 47, 48, 50] yebE CpxAR [32, 47, 48, 50] yfbE BasSR [51]
Table 1: Reporters used in this work. The known regulation by the eight candidate surface sensing systems is shown with corresponding references. Most reporter plasmids were obtained from [52], with the exception of the csrB, mdtA, apFAB70 and pspA reporters, which were constructed in this work. The regulations shown here are positive, except for auto-repression of pspF and repression of ompC by σE.
gene expression, reporter constructs for two housekeeping genes and one synthetic unregulated promoter were tested in the same experimental setup. The two housekeeping genes were chosen based on their reported relatively constant expression levels throughout different conditions. The tested reporters are frr [53, 54] and rpsM [55], and the synthetic apFAB70 promoter, which comes from a library of constitutive promoters [56]. The expression profiles of these genes upon surface contact are shown in Supplementary Figure S2. In the three hours after introduction into the microfluidic device, the fluorescence of the frr, rpsM and apFAB70 reporter strains increased by respectively 90%, 50% and 110%. Hence, the expression of the control genes is
Figure 2: Fluorescence traces of all reporters. Shown here are the mean fluorescence intensities of E. coli in the microfluidic setup carrying each of the 27 reporters (24 controlled by candidate surface sensing systems (black) and 3 control reporters (blue)). The intensities were normalised by setting the initial fluorescence of each cell to a value of 100. All individual fluorescence traces are shown in Supplementary Figure S1.
comparable to most reporters of the candidate surface sensing systems (Figure 2). To identify those reporters that were specifically induced in the microfluidic device, the initial changes in expression upon surface contact were compared to the frr, rpsM and apFAB70 reporters. We quantified the slopes of the fluorescence signals in the first hour after surface attachment and found that, out of the 24 reporters, six were significantly more induced than the control genes in a Wilcoxon Rank Sum test (p<0.05) and five were significantly less induced (Figure 3). A hypergeometric analysis of the reporters and their known regulators (Table 1) revealed that only reporters controlled by the Rcs system were significantly enriched in the induced subgroup (p = 0.0027). Thus, our screening indicated that only the Rcs system is activated by surface contact in the microfluidic device.
Induction of RcsB reporters by surface contact requires RcsF
Towards confirming that indeed the Rcs system was required for induction of its reporters on a surface, we compared the expression of the rcsA reporter in wild-type bacteria and the ∆rcsB mutant, lacking the response regulator of the Rcs system (Figure 4a). As expected, the rcsA promoter induced much less by the loss of rcsB, to a level comparable to the control genes (Figure 4b). Thus, the response regulator RcsB is required for induction of its reporters on a surface.
To further characterise surface sensing by the Rcs system, the role of the outer membrane component RcsF was investigated. RcsF is known to sense some of the signals that activate the Rcs system, but it is not required for a response to the full range of signals [57, 58]. Using an rcsF knockout strain, we found that loss of this gene had the same effect on expression of the reporter as the loss of response regulator RcsB (Figure 4b). Our results confirm that the Rcs system is activated upon surface contact, in an RcsF-dependent manner.
Surface sensing via Rcs is involved in the persister phenotype
Previous work of our lab found that E. coli cultures display growth bistability, with growing and dormant (persistent) populations, when the carbon source is switched from a glycolytic to a gluconeogenic substrate [59]. When shifting cells from glucose to fumarate medium, 99.9% of the cells enter the persistent population [60]. This method of generating persisters has allowed omics-based characterization of the persister phenotype [60]. However, studying these persister cells under the microscope has not been possible, as they switch from the dormant to the growing state as soon as they are immobilised in the microfluidic device [23]. We hypothesised that cells under the microscope might sense the contact with the surface and subsequently activate a transcriptional response that facilitates ’wakeup’ from the dormant state. A potential mechanism involves the Rcs system, which, via a complex regulation (Supplementary Figure S3) [61–63], could be involved in the expression of the transporter DctA, responsible for the import of several gluconeogenic carbon sources such as fumarate [64] and known to affect the bifurcation into growing and persistent populations after
Figure 3: Comparison of the fluorescence of all reporters following surface contact. The slopes are shown of linear fits to the normalised fluorescence intensities in the first hour. The blue bars are the control reporters considered to be expressed constitutively. Reporters with red bars have slopes significantly different from the controls (Wilcoxon rank sum test, p < 0.05), while reporters with black bars were comparable to the control reporters. Dashed bars are reporters regulated by the Rcs system. The inset shows a histogram of the fluorescence intensity slopes with the same colour coding. Due to the linear scale and the much stronger induction than other reporters, the rcsA reporter was excluded from the histogram.
a nutrient shift [59].
To test the involvement of the Rcs system in persister wakeup on a surface, we used the nutrient-shift method to generate persisters in wild-type, ∆rcsF and ∆rcsB strains. Eighteen hours after the nutrient shift, we transferred the cells to the microfluidic setup under the microscope and quantified the percentage of cells in the persister state (i.e. cells that were not growing in volume). Whereas the ratios of persisters and growing cells were comparable for all three strains in liquid culture (Supplementary Figure S4), we found the ratios to be markedly different already after 15 minutes in the microfluidic device (Figure 5a). Specifically, in all strains the percentage of cells in the persister state was lower in the microfluidic device than in liquid culture, suggesting that the surface contact had induced growth in part of the persister population. Moreover, while deletions of rcsB and rcsF had no effect on persistence in liquid culture, we found that the strain lacking the rcsF gene retained more cells in the persister state than the wild-type and ∆rcsB strains when surface-attached. During the next hours in the microfluidic device, further growth resumption of persisters occurred and we found the dynamics of persister wakeup to be affected by mutations in the Rcs system (Figure 5b). Thus, our results indicate that the Rcs system plays a role in persister wakeup in the microfluidic device.
Our results indicate that the rcsF deletion caused fewer persisters to resume growth
Figure 4: Expression of rcsA in wild-type and Rcs mutants on a surface. (a) Schematic representation of the Rcs pathway. OM: outer membrane, IM: inner membrane. (b) Expression of the rcsA transcriptional reporter on a surface in the wild-type (black), rcsB (red) and rcsF (blue) deletion strains. The graphs show the mean and 95% confidence intervals of at least 40 bacteria from two independent experiments. All values are normalised to the initial intensity of each cell.
Figure 5: Growth resumption of persisters of wild-type, rcsF, rcsB and dctA mutants on a surface. Bacteria were switched from glucose to fumarate minimal medium to generate persisters. After 18 h, the cells were introduced in the microfluidic setup of Figure 1. (a) Shown here is the percentage of bacteria that are in the persister state at the first time point of the time-lapse microscopy (i.e. after∼15 min in the microfluidic device). Persister cells were defined as those cells that did not grow in volume or divide. The fraction of persisters in liquid culture, 18 h after the nutrient shift, is shown in Supplementary Figure S4. (b) The percentage of persisters throughout the first 6 h of time-lapse microscopy is shown, including only the cells that resume growth (i.e. start to increase in volume) between 15 minutes and 6 hours after introduction into the microfluidic device. dctA∗: BW25113∆dctA complemented with pNTR-SD-dctA and induced with 50 µM IPTG. Each experiment is based on at least 59 cells (average 83).
upon surface contact, while the loss of rcsB had the opposite effect. This finding seemingly contradicts our results in Figure 4b, where we showed that the ∆rcsF and ∆rcsB strains are equally defective in controlling the expression of the reporter gene. This apparent disagreement exemplifies the complexity of transcriptional regulation by the Rcs system, involving both phosphorylated and unphosphorylated RcsB and multiple auxiliary factors [57, 65, 66]. While the ∆rcsF strain loses the ability to phosphorylate RcsB in response to surface attachment, transcriptional regulation that involves unphosphorylated RcsB remains functional, while that is not the case in the ∆rcsB mutant. Hence, our results are not contradictory, but rather illustrate the complexity of the Rcs system.
To further elucidate the mechanism by which Rcs affects growth resumption of persisters generated by the nutrient-shift method, we wanted to test whether the fumarate transporter DctA was involved, as the transcriptional regulation of dctA may involve RcsB at multiple regulatory steps (Supplementary Figure S3). To this end, we decoupled the expression of the dctA gene from the Rcs system, such that the
Figure 6: Activity of Rcs on several surfaces. Bacteria carrying the rcsA reporter were subjected to different surfaces to identify the requirements for induction. (a) Bacteria were trapped with optical tweezers (i.e. there was no surface contact) for 30 min with a laser power of 8 mW, n = 6. (b) Bacteria that spontaneously attached in a flow channel with silanised cover glass, n = 18. (c) Same as b, but with an untreated cover glass, n = 20. (d) Bacteria that were immobilised between a silanised cover glass and polyacrylamide pad, n = 20. The fluorescence increase in d is the same as in Figure 4b where a polyacrylamide setup with an untreated glass was used. Shown here in grey are the individual cells and in black are the mean ± 95% confidence intervals. All values are normalised to the initial intensity.
hypothesised induction of dctA on a surface would not happen. We achieved this decoupling by deleting the dctA gene and complementing the cells with plasmid-borne IPTG-inducible dctA. Because the expression level of DctA affects persister formation [59], we first determined that supplementing both the glucose and fumarate minimal media with 50 µM IPTG resulted in a comparable persister fraction as the wild-type after a nutrient shift, indicating similar expression levels. Eighteen hours after the nutrient shift, we introduced the cells into the microfluidic device and quantified the fraction of persister cells. The percentage of persisters fifteen minutes after introduction in the microfluidic device (Figure 5a), as well as the wakeup dynamics in the next several hours (Figure 5b), were the same as for the ∆rcsF strain. Hence, the Rcs system acts via dctA to cause growth resumption of persister cells. Overall, we have shown that persister wakeup on a surface is dependent on the Rcs system and DctA, which likely allows increased metabolic activity due to a higher fumarate uptake rate.
Activation of Rcs requires a large surface contact area
After finding that Rcs is activated by surface contact and consequently causes dormant cells to resume growth, we endeavoured to more closely investigate the dynamics of surface sensing by inducing the surface contact in a controlled manner. To this end, we introduced exponentially growing bacteria into a flow channel, trapped them with optical tweezers and attached them via electrostatic interactions to a silanised cover glass. The silane layer was applied to the cover glass to ensure sufficiently strong attachment, as bacteria could more easily detach from untreated glass. In Chapter 3,
we have shown that E. coli remains viable in an oscillating optical trap for at least 30 min under carefully optimized conditions. Here, we first found that the expression of the reporter remained constant during 30 min of optical trapping (Figure 6a), indicating that the trapping itself does not activate this envelope-stress response. Second, when trapped bacteria were brought into contact with the silanised cover glass, the Rcs system still did not respond for at least two hours.
To test whether the exposure to the IR beam had abolished surface sensing by Rcs, we observed cells that had spontaneously attached to the silanised surface without having been exposed to the optical tweezers. Again, we found no induction of the transcriptional reporter on a silanised glass surface (Figure 6b), indicating that the optical trapping did not abolish surface sensing and that there must be another reason why the silanised flow channel did not activate the Rcs system.
We hypothesised that the silane layer obscured some property of the cover glass that is necessary for sensing by Rcs, as the previous experiments where we did see activation of the system involved uncoated glass. Therefore, we tested a flow channel with uncoated glass. While many cells detached from the surface, we again found no increase in reporter fluorescence in the cells that remained attached (Figure 6c). This result indicates that the silane coating of the cover glass was not the cause of the lack of Rcs activation. As we have excluded the optical trapping and surface coating as explanations for lack of Rcs response, we conclude that the other difference with our previous experiments, the absence of a polyacrylamide pad, must explain why we observe no Rcs activation. Indeed, we found that the reporter fluorescence increases rapidly when cells were immobilised between a cover glass and polyacrylamide pad, regardless of surface coating (Figure 6d). Thus, activation of the surface sensing system RcsCDB is dependent on the presence of a polyacrylamide pad, possibly because it increases the surface contact area or because of pressure exerted by the two surfaces on either side of the cell.
Discussion
In this study, we performed a screening to determine which candidate surface sensing systems are activated when E. coli cells are brought into contact with a cover glass and polyacrylamide pad. Using a microfluidic device and 24 fluorescent transcriptional reporters for eight candidate surface sensing systems, we found that only the RcsB regulon is significantly upregulated by such surface contact (Figure 3). We confirmed that RcsB was required for the increased expression of its reporters. Also, the outer membrane lipoprotein RcsF was essential for surface-mediated activation of the Rcs system (Figure 4). Furthermore, we have shown that contact-dependent activation of Rcs does not occur when cells are immobilised only on a cover glass, but that it additionally requires a polyacrylamide pad.
Connecting the result of our screening to bacterial persistence, we established a role of surface sensing by the Rcs system in growth resumption of dormant cells generated by a previously described carbon source shift method [59] (Figure 5). Persistence
can result from very low metabolic activity, either occurring stochastically in a population of growing cells or induced in the laboratory, e.g. by shifting bacteria to a carbon source to which they are not adapted [24, 59]. We therefore hypothesised that the growth resumption of persisters on a surface might be caused by an RcsB-controlled increase in metabolic activity. DctA, a transporter for fumarate, malate and succinate [64], could be the link between Rcs and dormancy, as a complex regulatory network (Supplementary Figure S3) connects RcsB to dctA and this transporter would allow increased uptake of the new carbon source. We rendered the expression of the dctA gene independent of RcsB and found that surface-induced growth resumption of persisters was greatly reduced. We conclude that regulation of dctA by the Rcs system is involved in persister wakeup, likely by increasing the potential for uptake of the new carbon source.
Upregulation of transporters in surface-grown bacteria can be seen as a strategy to increase the range of nutrients that can be utilised, in order to cope with future adverse conditions. Generally, when running out of nutrients, bacteria would search by chemotaxis for a new environment that can sustain growth [67], but this is not possible when cells are attached to a surface. Therefore, especially in adherent cells, adapting the metabolism to utilise multiple carbon sources can be essential to ensure growth in suboptimal conditions where nutrients are scarce and need to be shared with the surrounding cells in the developing biofilm. Our results suggest that Rcs-dependent surface sensing upregulates DctA, allowing for uptake of several gluconeogenic substrates [64]. Consistently, increasing this uptake capability may sustain E. coli biofilms growing in their natural habitat, as succinate, another substrate for DctA, is produced by the gut microbiota in relatively high amounts [68]. While the Rcs regulon is not well-characterised, it may include several more transporters: galP (galactose uptake), the lac operon (lactose utilization) and yfdC (putative formate transporter) [19, 32]. Thus, multiple transporters may be upregulated on a surface, ensuring sufficient nutrient uptake capability.
We have found that the Rcs system cannot be activated merely by attachment to a flat glass surface, but that it requires another surface on top of the cell, in our case a polyacrylamide pad (Figure 6). Consistently, in a previous report [19] of contact-dependent activation of the Rcs system, cells were immobilised by filtering them onto a nitrocellulose membrane with a pore size that E. coli might be able to enter [69], and therefore these bacteria might also have experienced a large surface contact area. In addition, it was found that persisters would switch to the growing phenotype, suggesting Rcs activation, in a different microfluidic device [70], in which cells were in contact with both the cover glass and a PDMS pad (Johan Elf, personal communication). Thus, while surface contact on multiple sides of the cell is needed for a response of Rcs, it seems that the properties of the surface (i.e. hydrophilic or hydrophobic) do not matter. We envision that in a developing biofilm, adjacent bacteria would fulfil the function of surface, suggesting a role for Rcs after the initial adhesion, e.g. at the microcolony stage. Based on the downstream targets of the Rcs system, a role at the somewhat later stages of biofilm formation has previously been suggested [58].
The question now is why contact with multiple surfaces is needed for induction of the Rcs system. As the RcsCD dimer can act both as a kinase and a phosphatase on RcsB [58], one possibility would be that a small surface contact area would not sufficiently shift the balance between the two enzymatic activities: While RcsCD that is localised at the cell-surface interface would be active as a kinase, RcsCD that is further away from the site of surface contact may not change its activity and may therefore still act as a phosphatase, which would prevent the level of phosphorylated RcsB from increasing. Alternatively, the pressure exerted by the two surfaces on either side on the cell may be the signal activating the system. RcsF activates the system by interactions with an inner membrane protein, likely IgaA [71], as RcsF that is localised to the periplasm or bound to the inner membrane constitutively activates the system [72]. Recently, it was shown that the length of the unstructured region of RcsF is crucial for its function and that a longer linker region was needed when the inner and outer membranes were further apart [73]. Possibly, the pressure exerted by the two surfaces in our experiments slightly deformed the cell envelope [17, 69], decreasing the distance between the membranes and facilitating direct contact between RcsF and IgaA.
In conclusion, we identified only RcsCDB as a surface sensing system in our experimental setup and found that its activation leads to growth resumption of persisters, via regulation of dctA. Hence, our results implicate E. coli ’s Rcs system in two medically relevant phenotypes, biofilm formation and persistence, which makes it a promising target for prevention of biofilm formation and treatment of persistent infections.
Materials and methods
Bacterial strains and growth conditions
E. coli K12 strains MG1655 and BW25113 were used in this work. MG1655 with the transcriptional reporter plasmids shown in Table 1 were obtained from the Promoter Collection [52], with the exception of the csrB, mdtA, pspA and apFAB70 reporters. The first three of these reporters were constructed by amplifying the corresponding promoter regions from the MG1655 genome (for primers, see Table 2) and inserting them by restriction-ligation with BamHI and XhoI into plasmid pUA139, which is the same backbone as the other reporter plasmids. The apFAB70 reporter was constructed by annealing of two DNA oligos (Table 2) and insertion into the same pUA139 vector. The BW25113 deletion strains of rcsB, rcsF and dctA were taken from the Keio Collection [74]. For transfer of these deletions to the MG1655 strain P1-phage transduction was carried out. The pNTR-SD-dctA plasmid with IPTG-inducible dctA gene, was taken from the Mobile Plasmid Collection [75].
Bacteria were grown at 37◦C in an orbital shaker (300 rpm), in M9 minimal medium [59] supplemented with 0.4% glucose. The medium was additionally supplemented with 25 µg/ml kanamycin or ampicillin where appropriate. Preparation of spent medium
Primer Sequence
BamHI-csrB -fw TTTGGATCCCAACTTAAGCCTCTTCTGTAATCC
XhoI-csrB -rev TTTTCTCGAGGAAGTGTCATCATCCTGATGTTC
BamHI-mdtA-fw TTTGGATCCCGTCGGCTTACCTCCTTTC
XhoI-mdtA-rev TTTTCTCGAGCTTGCCAGAACCAGAATGC
BamHI-pspA-fw TTTGGATCCGAAAGCTGTTCGCCTCAC
XhoI-pspA-rev TTTTCTCGAGCAGAGCGTTGATGTTGGC
BamHI-apFAB70 -XhoI TTTGGATCCTTGACATCGCATCTTTTTGTACCT--ATAATGTGTGGATCTCGAGTTTT
BamHI-apFAB70 -XhoI-rc AAAACTCGAGATCCACACATTATAGGTACAAAA--AGATGCGATGTCAAGGATCCAAA
Table 2: Primers/oligos used in this work.
was done by spinning down bacterial cultures at 1000 g at 4◦C and subsequent filtering of the supernatant through a 0.22 µm pore-size bottle-top filter made of PES (Thermo Scientific Nalgene). Spent medium was always taken from cultures at the same OD600
as the culture used for the surface contact experiments.
Microfluidics
Bacteria were grown to an OD600 of 0.4-0.5 before introduction into the microfluidic
device. We used three microfluidic devices in this work, that were described previously (Chapters 2, 3 and [76]). For the experiments involving the polyacrylamide pad, 5 µL of bacterial culture was placed on the cover glass and a 10% polyacrylamide gel pad, which had been incubated for one hour in spent medium, was placed on top of the cells. The setup was completed by a polydimethylsiloxane (PDMS) slab moulded with a channel, through which there was a flow of 24 µL/min spent medium over the gel pad, controlled by a syringe pump (Pump 11 Elite, Harvard Apparatus). Image acquisition was started within 10 min following the introduction of bacteria into the device. For the flow channel experiments, the microfluidic device was constructed by bonding a cover glass to a PDMS mould, flushed with spent medium for 15 min and loaded with bacteria. When sufficient bacteria were surface-attached, a flow of 1.5 µL/min spent medium was initiated using a air-pressurised flow control system (OB1, Elveflow). The optical tweezer experiment was carried out as described in Chapter 3.
Microscopy
For image acquisition, a Nikon Eclipse Ti-E inverted microscope was used, with Nikon CFI Plan Apo Lambda DM 100X Oil objective, Lumencor Aura illumination system (485 nm LED for excitation of GFP) and Andor iXon 897 EM-CCD camera. The following filters were employed: excitation filter bandpass 470/40 nm, dichroic mirror
495 nm and emission filter 525/50 nm (AHF Analysentechnik F46-470). Focus was maintained by Nikon’s PFS3 system. Every 10 min phase contrast and GFP images (200 ms exposure time) were acquired at multiple positions. The microscope was controlled by NIS Elements v4.51 software.
Image analysis
Microscopy images were segmented manually with in-house software written in Matlab (R2014a, MathWorks Inc.). The detected cells were analysed by applying the identified ROIs to background-corrected GFP images. The background correction was done by first subtracting the signal intensity of images without any bacteria, followed by division of each pixel by a correction factor to correct for uneven illumination. The correction factors were determined by smoothing the intensities on a position without cells with a 3x3 point moving average and then dividing the intensity of each pixel by the mean of all pixels.
Generating persisters
Persisters were generated as described before [59]. A BW25113 culture was grown exponentially in glucose minimal medium until an OD600 around 0.5. Then 2.5·109
cells were pelleted by centrifugation, washed to remove glucose and transferred to two flasks with prewarmed fumarate minimal medium. After 17 h, one flask was used to prepare spent medium. After 18 h, cells were taken from the second flask and used for the microfluidics experiment. For determination of the fractions of growing and persister cells, stainings with the membrane-intercalating dye PKH26 were carried out as described [59].
Statistical analyses
Induction of transcriptional reporters was tested against the three control reporters by Wilcoxon rank sum test in Matlab, with a significance threshold of 0.05. Enrichment of reporters for each candidate surface sensing system in the induced subgroup was tested for significance by a hypergeometric analysis.
Acknowledgments
We would like to thank Jakub Radzikowski, Hannah Schramke, Ying Liu and Zheng Zhang for discussions and Johan Elf (Uppsala University, Sweden) for testing persisters in his microfluidic device. This work was financed by the Netherlands Organisation for Scientific Research (NWO) through a VIDI grant to MH (project number 864.11.001).
References
[1] Potera C (1999) Forging a Link Between Biofilms and Disease. Science 283(5409):1837–1839.
[2] Mah TFC, O’Toole GA (2001) Mechanisms of biofilm resistance to antimicrobial agents. Trends in microbiology 9(1):34–39.
[3] Stewart PS (1996) Theoretical aspects of antibiotic diffusion into microbial biofilms. Antimicrobial agents and chemotherapy 40(11):2517–2522.
[4] Anderl JN, Franklin MJ, Stewart PS (2000) Role of Antibiotic Penetration Limitation in Klebsiella pneumoniae Biofilm Resistance to Ampicillin and Ciprofloxacin. Antimicrobial agents and chemotherapy 44(7):1818–1824.
[5] Schultz MP, Bendick JA, Holm ER, Hertel WM (2011) Economic impact of biofouling on a naval surface ship. Biofouling 27(1):87–98.
[6] Callow JA, Callow ME (2011) Trends in the development of environmentally friendly fouling-resistant marine coatings. Nature communications 2:244. [7] Goodman AE, Marshall KC (1995) Genetic responses of bacteria at surfaces in
Microbial biofilms, eds. Costerton JW, Lappin-Scott HM. (Cambridge University Press, Cambridge), pp. 80–98.
[8] Hong Y, Brown DG (2010) Alteration of bacterial surface electrostatic potential and pH upon adhesion to a solid surface and impacts to cellular bioenergetics. Biotechnology and Bioengineering 105(5):965–972.
[9] Tuson HH, Weibel DB (2013) Bacteria-surface interactions. Soft Matter 9:4368– 4380.
[10] Pratt LA, Kolter R (1998) Genetic analysis of Escherichia coli biofilm formation: roles of flagella, motility, chemotaxis and type I pili. Molecular microbiology 30(2):285–293.
[11] van Houdt R, Michiels CW (2005) Role of bacterial cell surface structures in Escherichia coli biofilm formation. Research in microbiology 156:626–633. [12] Belas R (2014) Biofilms, flagella, and mechanosensing of surfaces by bacteria.
Trends in microbiology 22(9):517–527.
[13] Nord AL, et al. (2017) Catch bond drives stator mechanosensitivity in the bacterial flagellar motor. Proceedings of the National Academy of Sciences 114(49):12952–12957.
[14] Zhang JP, Normark S (1996) Induction of gene expression in Escherichia coli after pilus-mediated adherence. Science 273:1234–1236.
[15] Otto K, Norbeck J, Larsson T, Karlsson KA, Hermansson M (2001) Adhesion of type 1-fimbriated Escherichia coli to abiotic surfaces leads to altered composition of outer membrane proteins. Journal of Bacteriology 183(8):2445–2453.
[16] Bhomkar P, Materi W, Semenchenko V, Wishart DS (2010) Transcriptional Response Of E. coli Upon FimH-mediated Fimbrial Adhesion. Gene Regulation and Systems Biology 4:1–17.
[17] Chen Y, Harapanahalli AK, Busscher HJ, Norde W, van der Mei HC (2014) Nanoscale Cell Wall Deformation Impacts Long-Range Bacterial Adhesion Forces on Surfaces. Applied and Environmental Microbiology 80(2):637–643.
[18] Otto K, Silhavy TJ (2002) Surface sensing and adhesion of Escherichia coli controlled by the Cpx-signaling pathway. Proceedings of the National Academy of Sciences 99(4):2287–2292.
[19] Ferri`eres L, Clarke DJ (2003) The RcsC sensor kinase is required for normal biofilm formation in Escherichia coli K-12 and controls the expression of a regulon in response to growth on a solid surface. Molecular Microbiology 50(5):1665–1682. [20] Monds RD, O’Toole GA (2009) The developmental model of microbial biofilms:
ten years of a paradigm up for review. Trends in microbiology 17(2):73–87. [21] Wentland EJ, Stewart PS, Huang CT, McFeters GA (1996) Spatial variations
in growth rate within Klebsiella pneumoniae colonies and biofilm. Biotechnology progress 12:316–321.
[22] Domka J, Lee J, Bansal T, Wood TK (2007) Temporal gene-expression in Escherichia coli K-12 biofilms. Environmental Microbiology 9(2):332–346. [23] Radzikowski J, van Kessel S, Heinemann M (2017) Low quorum and/or surface
contact abolish persister formation after a nutrient shift in Systems biology of bacterial persistence, a metabolism-driven strategy for survival. (Rijksuniversiteit Groningen, Groningen), pp. 260–297.
[24] Radzikowski JL, Schramke H, Heinemann M (2017) Bacterial persistence from a system-level perspective. Current Opinion in Biotechnology 46:98–105.
[25] Lee LJ, Barrett JA, Poole RK (2005) Genome-Wide Transcriptional Response of Chemostat-Cultured Escherichia coli to Zinc. Journal of bacteriology 187(3):1124–1134.
[26] Dartigalongue C, Missiakas D, Raina S (2001) Characterization of the Escherichia coli sigma(E) regulon. The Journal of biological chemistry 276(24):20866–20875. [27] Quinn HJ, Cameron ADS, Dorman CJ (2014) Bacterial regulon evolution: distinct responses and roles for the identical OmpR proteins of Salmonella typhimurium and Escherichia coli in the acid stress response. PLoS genetics 10(3):e1004215.
[28] Sahu SN, et al. (2003) The bacterial adaptive response gene, barA, encodes a novel conserved histidine kinase regulatory switch for adaptation and modulation of metabolism in Escherichia coli. Molecular and cellular biochemistry 253:167–177. [29] Suzuki K, et al. (2002) Regulatory Circuitry of the CsrA/CsrB and BarA/UvrY
Systems of Escherichia coli . Journal of bacteriology 184(18):5130–5140.
[30] Zere TR, et al. (2015) Genomic targets and features of BarA-UvrY (-SirA) signal transduction systems. PLoS ONE 10(12):e0145035.
[31] Ogasawara H, Shinohara S, Yamamoto K, Ishihama A (2012) Novel regulation targets of the metal-response BasS-BasR two-component system of Escherichia coli. Microbiology 158:1482–1492.
[32] Bury-Mon´e S, et al. (2009) Global analysis of extracytoplasmic stress signaling in Escherichia coli . PLoS Genetics 5(9):e1000651.
[33] Rhodius VA, Suh WC, Nonaka G, West J, Gross CA (2006) Conserved and variable functions of the sigma E stress response in related genomes. PLoS biology 4(1 e2):43–59.
[34] Kabir MS, et al. (2005) Cell lysis directed by sigma E in early stationary phase and effect of induction of the rpoE gene on global gene expression in Escherichia coli . Microbiology 151(8):2721–2735.
[35] Mitra A, Herren CD, Patel IR, Coleman A, Mukhopadhyay S (2016) Integration of AI-2 Based Cell-Cell Signaling with Metabolic Cues in Escherichia coli . PloS one 11(6):e0157532.
[36] Nagakubo S, Nishino K, Hirata T, Yamaguchi A (2002) The Putative Response Regulator BaeR Stimulates Multidrug Resistance of Escherichia coli via a Novel Multidrug Exporter System, MdtABC. Journal of bacteriology 184(15):4161– 4167.
[37] Baranova N, Nikaido H (2002) The BaeSR Two-Component Regulatory System Activates Transcription of the yegMNOB (mdtABCD ) Transporter Gene Cluster in Escherichia coli and Increases Its Resistance to Novobiocin and Deoxycholate. Journal of bacteriology 184(15):4168–4176.
[38] Mizuno T, Mizushima S (1990) Signal transduction and gene regulation through the phosphorylation of two regulatory components: the molecular basis for the osmotic regulation of the porin genes. Molecular microbiology 4(7):1077–1082. [39] Batchelor E, Walthers D, Kenney LJ, Goulian M (2005) The Escherichia coli
CpxA-CpxR Envelope Stress Response System Regulates Expression of the Porins OmpF and OmpC. Journal of bacteriology 187(16):5723–5731.
[40] Hagiwara D, et al. (2003) Genome-Wide Analyses Revealing a Signaling Network of the RcsC-YojN-RcsB Phosphorelay System in Escherichia coli . Journal of bacteriology 185(19):5735–5746.
[41] Jovanovic G, Weiner L, Model P (1996) Identification, Nucleotide Sequence, and Characterization of PspF, the Transcriptional Activator of the Escherichia coli Stress-Induced psp Operon. Journal of bacteriology 178(7):1936–1945.
[42] Jovanovic G, Dworkin J, Model P (1997) Autogenous control of PspF, a constitutively active enhancer-binding protein of Escherichia coli . Journal of bacteriology 179(16):5232–5237.
[43] Wehland M, Bernhard F (2000) The RcsAB box. Characterization of a new operator essential for the regulation of exopolysaccharide biosynthesis in enteric bacteria. Journal of Biological Chemistry 275(10):7013–7020.
[44] Stratmann T, Pul ¨U, Wurm R, Wagner R, Schnetz K (2012) RcsB-BglJ activates the Escherichia coli leuO gene, encoding an H-NS antagonist and pleiotropic regulator of virulence determinants. Molecular Microbiology 83(6):1109–1123. [45] Klein G, et al. (2016) Multiple transcriptional factors regulate transcription of
the rpoE Gene in Escherichia coli under different growth conditions and when the lipopolysaccharide biosynthesis is defective. Journal of Biological Chemistry 291(44):22999–23019.
[46] Raffa RG, Raivio TL (2002) A third envelope stress signal transduction pathway in Escherichia coli. Molecular microbiology 45(6):1599–1611.
[47] Price NL, Raivio TL (2009) Characterization of the Cpx regulon in Escherichia coli strain MC4100. Journal of bacteriology 191(6):1798–1815.
[48] Raivio TL, Leblanc SKD, Price NL (2013) The Escherichia coli Cpx envelope stress response regulates genes of diverse function that impact antibiotic resistance and membrane integrity. Journal of bacteriology 195(12):2755–2767. [49] Weatherspoon-Griffin N, Yang D, Kong W, Hua Z, Shi Y (2014) The CpxR/CpxA
Two-component Regulatory System Up-regulates the Multidrug Resistance Cascade to Facilitate Escherichia coli Resistance to a Model Antimicrobial Peptide. Journal of Biological Chemistry 289(47):32571–32582.
[50] Yamamoto K, Ishihama A (2006) Characterization of copper-inducible promoters regulated by CpxA/CpxR in Escherichia coli. Bioscience, Biotechnology and Biochemistry 70(7):1688–1695.
[51] Hagiwara D, Yamashino T, Mizuno T (2004) A Genome-Wide View of the Escherichia coli BasS-BasR Two-component System Implicated in Iron-responses. Bioscience, Biotechnology and Biochemistry 68(8):1758–1767.
[52] Zaslaver A, et al. (2006) A comprehensive library of fluorescent transcriptional reporters for Escherichia coli . Nature methods 3(8):623–628.
[53] Allen TE, et al. (2003) Genome-Scale Analysis of the Uses of the Escherichia coli Genome: Model-Driven Analysis of Heterogeneous Data Sets. J. Bacteriol. 185(21):6392–6399.
[54] Liu M, et al. (2005) Global transcriptional programs reveal a carbon source foraging strategy by Escherichia coli . Journal of Biological Chemistry 280(16):15921–15927.
[55] Hautefort I, Proen¸ca MJ, Hinton JCD (2003) Single-Copy Green Fluorescent Protein Gene Fusions Allow Accurate Measurement of Salmonella Gene Expression In Vitro and during Infection of Mammalian Cells. Applied and Environmental Microbiology 69(12):7480–7491.
[56] Mutalik VK, et al. (2013) Precise and reliable gene expression via standard transcription and translation initiation elements. Nature methods 10(4):354–360. [57] Majdalani N, Heck M, Stout V, Gottesman S (2005) Role of RcsF in Signaling to the Rcs Phosphorelay Pathway in Escherichia coli . Journal of bacteriology 187(19):6770–6778.
[58] Majdalani N, Gottesman S (2006) The Rcs phosphorelay: a complex signal transduction system. Annual review of microbiology 59:379–405.
[59] Kotte O, Volkmer B, Radzikowski JL, Heinemann M (2014) Phenotypic bistability in Escherichia coli ’s central carbon metabolism. Molecular Systems Biology 10:736.
[60] Radzikowski JL, et al. (2016) Bacterial persistence is an active σSstress response to metabolic flux limitation. Molecular Systems Biology 12(9):882.
[61] Castani´e-Cornet MP, et al. (2010) Acid stress response in Escherichia coli : mechanism of regulation of gadA transcription by RcsB and GadE. Nucleic acids research 38(11):3546–3554.
[62] Tucker DL, et al. (2003) Genes of the GadX-GadW Regulon in Escherichia coli . Journal of bacteriology 185(10):3190–3201.
[63] Boogerd FC, Boe L, Michelsen O, Jensen PR (1998) atp mutants of Escherichia coli fail to grow on succinate due to a transport deficiency. Journal of Bacteriology 180(22):5855–5859.
[64] Baker KE, Ditullio KP, Neuhard J, Kelln RA (1996) Utilization of orotate as a pyrimidine source by Salmonella typhimurium and Escherichia coli requires the dicarboxylate transport protein encoded by dctA. J Bacteriol 178(24):7099–7105. [65] Venkatesh GR, et al. (2010) BglJ-RcsB heterodimers relieve repression of the Escherichia coli bgl operon by H-NS. Journal of Bacteriology 192(24):6456–6464. [66] Pannen D, Fabisch M, Gausling L, Schnetz K (2016) Interaction of the RcsB response regulator with auxiliary transcription regulators in Escherichia coli . Journal of Biological Chemistry 291(5):2357–2370.
[67] Bi S, Sourjik V (2018) Stimulus sensing and signal processing in bacterial chemotaxis. Current Opinion in Microbiology 45:22–29.
[68] De Vadder F, et al. (2016) Microbiota-Produced Succinate Improves Glucose Homeostasis via Intestinal Gluconeogenesis. Cell Metabolism 24:151–157. [69] M¨annik J, Driessen R, Galajda P, Keymer JE, Dekker C (2009) Bacterial growth
and motility in sub-micron constrictions. Proceedings of the National Academy of Sciences 106(35):14861–14866.
[70] Ullman G, et al. (2013) High-throughput gene expression analysis at the level of single proteins using a microfluidic turbidostat and automated cell tracking. Phil. Trans. R. Soc. B 368(20120025):1–8.
[71] Cho SH, et al. (2014) Detecting envelope stress by monitoring beta-barrel assembly. Cell 159(7):1652–1664.
[72] Farris C, Sanowar S, Bader MW, Pfuetzner R, Miller SI (2010) Antimicrobial peptides activate the Rcs regulon through the outer membrane lipoprotein RcsF. Journal of bacteriology 192(19):4894–4903.
[73] Asmar AT, et al. (2017) Communication across the bacterial cell envelope depends on the size of the periplasm. PLoS Biology 15(12):e2004303.
[74] Baba T, et al. (2006) Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Molecular systems biology 4474(2):2006.0008.
[75] Saka K, et al. (2005) A Complete Set of Escherichia coli Open Reading Frames in Mobile Plasmids Facilitating Genetic Studies. DNA Research 12(1):63–68. [76] Huberts DH, et al. (2013) Construction and use of a microfluidic dissection
platform for long-term imaging of cellular processes in budding yeast. Nature Protocols 8(6):1019–1027.
[77] Davies SJ, et al. (1999) Inactivation and Regulation of the Aerobic C4-Dicarboxylate Transport (dctA) Gene of Escherichia coli . Journal of bacteriology 181(18):5624–5635.
[78] Gama-Castro S, et al. (2016) RegulonDB version 9.0: High-level integration of gene regulation, coexpression, motif clustering and beyond. Nucleic Acids Research 44:D133–D143.
Supplement
Figure S1: (Next four pages) Fluorescence traces of all tested reporters on a surface. Bacteria carrying a reporter plasmid were introduced in the microfluidic device and the fluorescence intensities were determined every ten minutes. For each cell, the fluorescence was normalised to the initial intensity. On this page and the next two pages are the reporters that had a similar or lower level of induction on a surface than the three control reporters, in alphabetical order, with identical y-axis scaling. On the fourth page are the reporters that were significantly stronger induced (note that the y-axis scaling is different for those reporters). Each graph shows the mean and 95% confidence intervals of a median number of 20 cells.
Figure S2: Fluorescence traces of three control reporters on a surface. Bacteria carrying a reporter plasmid were introduced in the microfluidic device and the fluorescence intensities were determined every ten minutes. For each cell, the fluorescence was normalised to the initial intensity. Each graph shows the mean and 95% confidence intervals of 50 cells.
Figure S3: Scheme of dctA regulation Only relevant components and regulatory steps are shown. The figure is based on references [61–63, 77], as well as RegulonDB [78]. DctR and RcsB can form a heterodimer, but no target genes are known [66].
Figure S4: Fraction of growing cells in flask. Exponentially growing cultures in glucose minimal medium were stained with the membrane-intercalating dye PKH26 and shifted to fumarate minimal medium. The fluorescence distributions of wild-type and Rcs mutants were measured by flow cytometry at several time points after the carbon source switch and the initial fraction of growing cells was determined (as described in Ref. [59]) to be below 0.1%. These graphs show the fluorescence distributions (10log) after 18 hours, corresponding to the time points that microscopy experiments of Figure 5 were started. The percentages of growing cells 18 h after the switch (the left peaks in these graphs) were determined to be 7.1%, 5.8% and 5.6%, for wild-type, ∆rcsF and ∆rcsB respectively.
