University of Groningen
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.
Chapter 6
Conclusions and outlook
Tom E.P. Kimkes
Molecular Systems Biology, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen, Nijenborgh 4, 9747 AG Groningen, the Netherlands
Chapter 6
Summary of this thesis
In this thesis, we first provided a detailed overview of the current knowledge of surface sensing and biofilm initiation in the gram-negative model organism Escherichia coli (Chapter 1). We motivated the selection of eight signal transduction systems for being putatively involved in perception of surface contact. We initially tested two of these systems, CpxAR and RcsCDB, as they were previously implicated in surface sensing (Chapter 2). A reporter gene controlled by RcsB was strongly expressed upon surface contact in our microfluidic device, while CpxR-regulated genes were not. As two completely different experimental approaches failed to show contact-dependent activation of CpxR, we concluded that there is no indication that this system is responsive to surface contact.
Screening of all eight candidate surface sensing systems revealed that the Rcs system is the only one activated by surface contact under our experimental conditions (Chapter 4). Further, surface-mediated activation of Rcs resulted in the wakeup of persister cells, which we conjecture was caused by an adapting metabolic machinery in preparation of nutrient limitations that accompany life in a biofilm.
In Chapter 3, we investigated whether optical tweezers could be a useful tool for long-term handling of single bacteria, which could be used for, amongst other things, dynamic control over surface contact. We found that E. coli could be stably optically trapped with little damage by an oscillating optical trap. We established that trapping for at least thirty minutes is possible, thereby showing the potential of optical manipulation for single-cell studies.
Finally, in Chapter 5, we further characterised the Cpx system and specifically its response to copper ions, which had not yet been thoroughly studied. Here, we found activation of CpxR in absence of its cognate histidine kinase, CpxA. Furthermore, we found medium conditioning by orotate secretion to provide protection against copper.
Outlook
In the last twenty years, large parts of microbiology have become a science of single-cell biology. The experimental setup used here has been one of the key methods to study single cells and their dynamics. Therefore, the finding in this thesis that immobilisation of E. coli between a cover glass and gel pad rapidly induces a surface sensing response, also has implications beyond E. coli and beyond the study of surface sensing. Generally, the possibility that the immobilisation between surfaces may invoke a contact-dependent response is not considered. Therefore, results obtained with this type of experimental setup may have been impacted by surface sensing-induced responses and, for future studies, it is essential that the possibility of such responses are accounted for.
In Chapter 1, the gaps in our knowledge of surface sensing were highlighted. Specifically, while we did know what a bacterium might sense at a solid-liquid interface,
Conclusions and outlook
we knew very little about how they could sense it. The crucial first step towards an improved understanding was therefore to identify which signaling systems are involved in the process of surface sensing. With that knowledge, we can hope to attain a mechanistic understanding of the sensing itself and a systems-level understanding of the initiation of the switch from planktonic to sessile phenotype. Improved knowledge of the mechanisms underlying surface sensing would then facilitate the design of novel materials, coatings and (medical) treatments to prevent surface recognition and subsequent biofilm formation.
In this thesis, we performed a screening of candidate surface sensing systems and we found only a single system to be surface-responsive. Testing a few experimental setups, with different surface properties and contact area, we found that specific setups were required for the activation. This finding implies that specific conditions must be met before surface sensing can occur, which in the case of the Rcs system are either a large contact area or exertion of mechanical forces by two surfaces on either side of the cell. Therefore, we hypothesise that Rcs plays a role at the somewhat later stages of biofilm formation after the initial adhesion, e.g. at the microcolony stage, where adjacent bacteria may act as the surface. Sensing of contact area would resemble a new mechanism of perceiving the local cell density, next to the traditional quorum sensing mechanism based around secreted signalling molecules [1] and cell-cell contact sensing via specific receptor interactions [2]. We speculate that sensing of contact area might have an advantage over the other sensing mechanisms mainly in the case of multispecies biofilms, as it would allow bacteria to respond to the local cell density without having to rely on chemical signals or specific receptors that might differ between species. It would be interesting to see whether this novel sensing mechanism is more widespread.
Since the only surface sensing system that was identified in our screening likely plays a role at a somewhat later stage of biofilm formation, we should also consider the possibility that E. coli K12 does not have any systems in place to perceive initial surface contact. As we know so little about early surface-attached E. coli, possibly the first response occurs only after a few cell divisions, when the contact area increases by contact between cells. While there are indications that E. coli does respond to adhesion within minutes (i.e. by a decrease in respiration [3]), there is no concrete evidence that significant transcriptional changes are taking place immediately following contact with a surface. Indeed, the earliest reported transcriptional responses in E. coli were after four hours of attachment [4]. Therefore, omics data for cells that have been attached for significantly shorter than four hours would provide great insight. Overall, the possibility that adhesion does not elicit an immediate response is a conjecture that should be considered when studying surface sensing.
Although all but one of the candidate surface sensing systems were found to be unresponsive in our microfluidic setup, there might be conditions under which these systems do perceive contact with a surface. Just like the Rcs system requires a large contact area, other systems might also have specific requirements. Thus, while we have now performed a first screening, future work should aspire to include multiple conditions, such as surface properties, bacterial strains, growth phase, nutrient
Chapter 6
availability, etc. Here, for dynamically controlling surface contact, we propose the use of optical tweezers, which we have shown to enable the stable handling of single E. coli cells, while being relatively harmless. Additionally, optical tweezers should allow for establishing not just cell-surface interactions, but also cell-cell interactions by assembling multiple layers of bacteria, to study when the first contact-dependent responses are initiated.
Experimental conditions might play a large role in surface sensing. With regard to the surface, our choice for glass and polyacrylamide pads was based on existing knowledge about persister wakeup, which we expected (and confirmed) to result from surface sensing-induced adaptations. However, these materials likely do not resemble surfaces that E. coli encounter in nature. While the study of responses to abiotic surfaces are definitely valuable, more relevant results might be obtained when E. coli are investigated upon attachment to eukaryotic cells, although this will be harder to do experimentally. Along the same line, as biofilm formation is a virulence-related process, clinical isolates will surely yield results that are more directly applicable to treatment of actual infections, as compared to laboratory-adapted strains [5]. For instance, the most commonly used lab strains lack several types of pili [6, 7], have low expression levels of curli [8] and have an altered lipopolysaccharide structure [9], which all play a role in surface-interactions. However, even laboratory-adapted strains are a good starting point for growing our understanding of biofilm initiation and have practical advantages, such as having well-annotated genome sequences and being non-pathogenic. Growth phase might also affect surface sensing, as it could control expression of, for instance, cell appendages [10]. Furthermore, biofilm formation can be affected by the nutrient availability [11, 12], which could therefore also be related to the ability to sense surface contact. Thus, different experimental conditions might lead to identification of different surface sensing systems.
Knowing which systems are responsible for sensing contact with surfaces opens the way for more detailed investigations. Questions that should be answered are which properties of the surface are sensed, how the sensing is achieved and, perhaps even more importantly, how it can be prevented. Establishing the sensing mechanism could initially involve deletion strains to identify the most upstream sensor, like we did for the Rcs system. Later steps should identify protein-protein interactions upon surface contact, where optical tweezers could be employed to control the attachment and microscopic techniques such as FRET could provide evidence of interactions, as well as the dynamics. Furthermore, a systems-level understanding of biofilm initiation could be developed by omics studies on surface-attached cells at a number of time points after surface contact. However, as already mentioned, there are practical problems when working with early surface-attached cells. If we precisely know which systems are activated by surface contact, it might be possible to circumvent the inherent issues of early sessile cells by activating these systems in planktonic bacteria, as most of the candidate surface sensing systems are also responsive to chemical signals. With transcriptomics, proteomics and metabolomics data of surface-attached cells, a systems-level understanding of the earliest steps of biofilm formation might be attained.
Conclusions and outlook
References
[1] Beloin C, et al. (2004) Global impact of mature biofilm lifestyle on Escherichia coli K-12 gene expression. Molecular Microbiology 51(3):659–674.
[2] Blango MG, Mulvey MA (2009) Bacterial landlines: contact-dependent signaling in bacterial populations. Current Opinion in Microbiology 12:177–181.
[3] Geng J, Beloin C, Ghigo JM, Henry N (2014) Bacteria Hold Their Breath upon Surface Contact as Shown in a Strain of Escherichia coli, Using Dispersed Surfaces and Flow Cytometry Analysis. PloS one 9(7):e102049.
[4] Domka J, Lee J, Bansal T, Wood TK (2007) Temporal gene-expression in Escherichia coli K-12 biofilms. Environmental Microbiology 9(2):332–346. [5] Rossi E, et al. (2018) ”It’s a gut feeling” - Escherichia coli biofilm formation in the
gastrointestinal tract environment. Critical Reviews in Microbiology 44(1):1–30. [6] Xicohtencatl-Cortes J, et al. (2009) The type 4 pili of enterohemorrhagic
Escherichia coli O157:H7 are multipurpose structures with pathogenic attributes. Journal of Bacteriology 191(1):411–421.
[7] Kuehn MJ, Heuser J, Normark S, Hultgren SJ (1992) P pili in uropathogenic E. coli are composite fibres with distinct fibrillar adhesive tips. Nature 356:252–255. [8] Vidal O, et al. (1998) Isolation of an Escherichia coli K-12 Mutant Strain Able To Form Biofilms on Inert Surfaces: Involvement of a New ompR Allele That Increases Curli Expression. Journal of bacteriology 180(9):2442–2449.
[9] Liu D, Reeves PR (1994) Escherichia coli K12 regains its O antigen. Microbiology 140:49–57.
[10] Pesavento C, et al. (2008) Inverse regulatory coordination of motility and curli-mediated adhesion in Escherichia coli. Genes & development 22:2434–2446. [11] Costerton JW, Lewandowski Z, Caldwell DE, Korber DR, Lappin-scott HM
(1995) Microbial biofilms. Annual review of microbiology 49:711–745.
[12] 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.
Summary
Bacterial biofilms are involved in the majority of bacterial infections and complicate their treatment. Moreover, a wide range of surfaces can be colonized by bacteria and high costs are associated with reduced effectiveness of bio-fouled materials and their regular cleaning. Therefore, it is imperative to understand how bacterial biofilm formation is initiated, so that it might be counteracted.
In Chapter 1 of this thesis, we described the current knowledge in the field of surface sensing. As a bacterium approaches a surface, it might sense changes in its local physicochemistry, as the osmolarity, ionic strength, pH and nutrient availability differ near a solid-liquid interface compared to the bulk liquid. The gram-negative model organism Escherichia coli has several mechanisms in place to sense these properties. Furthermore, cell appendages (flagella and pili) can attach to the surface and they are known to be sensitive to adhesion. Finally, if the cell body itself comes into contact with the surface, this could give rise to cell envelope deformations, which could trigger the activation of envelope stress response systems. Several of the signal transduction systems that can be activated by these surface-related signals, regulate the expression of genes that are involved in biofilm formation. However, for most systems it has never been investigated whether they are really activated by surface contact and subsequently initiate biofilm formation. The reason that research in the field of surface sensing has been so limited is probably that investigation of single cells as they transition from the suspended to the attached state is inherently difficult. Therefore, we ended Chapter 1 with an overview of experimental developments that may facilitate further research of surface sensing.
In Chapter 2, we investigated the two previously reported surface sensing systems in E. coli with single-cell resolution, as opposed to the population-level measurements that were employed in the original publications. We found that, indeed, the Rcs system is highly responsive to surface contact, but that the widely-recognised contact-dependent activation of the Cpx system does not occur. Despite many attempts, including replication of the originally reported experiments, we could not confirm a role of CpxAR in surface sensing. Instead, we identified two technical reasons, cell lysis and contamination of the coating material with copper, that might have impacted the previous report. We conclude that, while E. coli ’s Rcs system is surface-responsive, the role of the Cpx system needs to be reconsidered.
In the third chapter, we worked on improving optical trapping with the goal of prolonged single-cell manipulation in a microfluidic setup. We established the upper limit of IR exposure to be 8 J if a conventional, stationary optical trap was used, while all cells tolerated up to 20 J in an oscillating trap. With regard to stability, the oscillating trap was optimal when the effective trap length was about 20% larger than the cell length and when it was oriented perpendicular to the flow direction of the medium. Using the optimal settings and a laser intensity that was just sufficient for stable trapping, we could hold single E. coli cells for at least 30 min without affecting their viability. The ability to hold cells for this long without adverse effects, will
facilitate a wide range of microscopic experiments in which bacteria can be studied while they are not in contact with a surface and where surface attachment can be induced in a controlled manner.
In Chapter 4, we performed a screening, using transcriptional reporters for the eight candidate surface sensing systems in E. coli. Only the Rcs system was found to respond to surface contact in our microfluidic setup. Using a few experimental setups, we established that activation of Rcs required simultaneous contact with two surfaces, a cover glass and a polyacrylamide gel pad, indicating that either a large contact area is required or that the pressure exerted by the two surfaces activates the system. This requirement of extensive surface contact suggests a role for Rcs at a later stage of biofilm development, where contact with surrounding bacteria might trigger the system. The downstream targets of Rcs are consistent with a role in a developing biofilm, as Rcs induces several metabolic genes and transporters, ensuring sufficient nutrient uptake and utilisation under conditions of scarcity as encountered in the centre of a biofilm, and genes for the biosynthesis of the exopolymeric matrix that encapsulates the growing biofilm. Thus, in our investigation of surface sensing systems, we identified only the Rcs system and our findings suggest that it controls adaptation to growth in a developing biofilm.
Finally, in Chapter 5, the activation of CpxR by trace amounts of copper ions was studied. Activation can go via its cognate histidine kinase CpxA, but also proceeds in the absence of that protein, in which case the histidine kinase CusS is required. This cross-talk between pathways might be due to the homology between them, but might also indicate the importance of CpxR in protection against copper, as illustrated by bacteriostatic effects of trace amounts of copper on ∆cpxR cells. Furthermore, our results show that medium conditioning during growth, specifically by excretion of orotic acid, can shield cells from adverse conditions, in this case the presence of copper.
Samenvatting
Bij het merendeel van de bacteri¨ele infecties zijn biofilms betrokken, wat de behandeling kan compliceren. Bovendien kunnen bacteri¨en een grote verscheidenheid aan oppervlakken koloniseren en er zijn hoge kosten gemoeid met de verminderde effectiviteit van biologisch verontreinigde materialen en met het regelmatig reinigen ervan. Het is daarom van groot belang dat we een beter beeld krijgen van de initiatie van biofilmformatie, zodat het in de toekomst beter kan worden bestreden.
In hoofdstuk 1 van dit proefschrift hebben we de huidige kennis over bacteri¨ele oppervlaktewaarneming beschreven. Wanneer een bacterie een oppervlak nadert, zullen de lokale fysicochemische eigenschappen voor de cel veranderen, aangezien de osmolariteit, ionsterkte, pH en beschikbaarheid van voedingsstoffen nabij de grens tussen een vloeistof en een vast materiaal over het algemeen afwijken van de condities verder weg van het oppervlak. Het gram-negatieve modelorganisme Escherichia coli bezit meerdere mechanismen die deze eigenschappen kunnen waarnemen. Verder bezitten bacteri¨en meestal flagella en pili die aan oppervlakken kunnen binden en waarvan bekend is dat ze gevoelig zijn voor dergelijke adhesie. Als ten slotte de cel zelf in contact komt met het oppervlak kan de cel-envelop - het geheel van binnenmembraan, buitenmembraan en periplasma - licht vervormen, wat vervolgens signaaltransductiesystemen (de ”envelope stress responses”) kan activeren. Enkele van de systemen die op deze oppervlaktegerelateerde signalen kunnen reageren, spelen bovendien een rol in de expressie van genen waarvan een deel betrokken is bij biofilmformatie. Echter, het is in de meeste gevallen nooit onderzocht of deze systemen daadwerkelijk geactiveerd worden door contact met een oppervlak en of ze verantwoordelijk zijn voor de initiatie van biofilmformatie. De reden hiervoor is waarschijnlijk dat er beperkte experimentele methoden beschikbaar zijn voor het volgen van individuele bacteri¨en tijdens de overgang van een vrij bewegende naar aangehechte toestand. Daarom hebben we het eerste hoofdstuk afgesloten met een overzicht van experimentele ontwikkelingen die verder onderzoek van opperlaktewaarneming mogelijk kunnen gaan maken.
In hoofdstuk 2 hebben we de twee eerder gerapporteerde oppervlaktewaarnemings-systemen in E. coli onderzocht op enkel-cel niveau, in tegenstelling tot de ensemblemetingen die in de originele publicaties waren toegepast. Wij konden bevestigen dat het Rcs systeem sterk reageert op oppervlaktecontact, maar we vonden geen enkele aanwijzing dat de alom erkende contactafhankelijke activatie van het Cpx systeem plaatsvindt, ondanks vele pogingen, waaronder replicatie van de originele experimenten. We hebben echter twee technische verklaringen gevonden, cellysis en verontreiniging van het gebruikte coatingmateriaal met koper, die mogelijk de experimenten in de originele publicatie be¨ınvloed hebben. Onze conclusie is dat E. coli ’s Rcs systeem oppervlaktecontact waarneemt, maar dat daarentegen de rol van CpxAR moet worden herzien.
In het derde hoofdstuk werkten we aan het optimaliseren van het vangen van E. coli met optical tweezers, met als doel het langdurig manipuleren van individuele cellen in
een microflu¨ıdische chip. Wanneer een conventionele stilstaande ”optical trap” werd gebruikt, tolereerden cellen maximaal 8 J aan infrarood licht, terwijl in het geval van oscillerende tweezers een dosis tot 20 J veilig bleek. Wat betreft de stabiliteit, was de oscillerende optical trap optimaal als de effectieve lengte van de trap ongeveer 20% groter was dan de cellengte en de ori¨entatie loodrecht was ten opzichte van de stroomrichting in het kanaal. Met de optimale configuratie en een laserintensiteit die net voldoende was voor het stabiel vasthouden van een cel, konden we individuele E. coli cellen gedurende ten minste 30 minuten vasthouden zonder hun overleving te be¨ınvloeden. De mogelijkheid om cellen zo lang optisch vast te houden zonder nadelige effecten zal nieuwe microscopische experimenten mogelijk maken waarin bacteri¨en bestudeerd kunnen worden terwijl ze niet in contact zijn met enig oppervlak en waarbij het contact met een oppervlak op een gecontroleerde manier geforceerd kan worden. In hoofdstuk 4 hebben we getest welke van de acht mogelijke oppervlakte-waarnemingssystemen in E. coli, zoals beschreven in hoofdstuk 1, daadwerkelijk geactiveerd worden door contact met een oppervlak. Enigszins verassend reageerde in onze experimenten enkel het Rcs systeem op contact met een oppervlak. Verdere karakterisatie van het Rcs systeem toonde aan dat ofwel een grote contactoppervlakte benodigd is, ofwel dat er enige druk op de cel uitgeoefend moet worden, wil activatie van Rcs optreden. Deze vereiste van uitgebreid oppervlaktecontact suggereert een rol voor Rcs in de latere stappen van biofilmformatie, wanneer contact met aangrenzende bacteri¨en het systeem kan activeren. De cellulaire processen die door Rcs gereguleerd worden zijn consistent met een rol in een ontwikkelende biofilm. Rcs reguleert namelijk enkele metabole genen en transporters en kan op die manier bijdragen aan voldoende activiteit van het metabolisme onder condities van voedselschaarste, die kunnen optreden in een biofilm. Bovendien zorgt het Rcs systeem ervoor dat de exopolymere matrix gevormd wordt die een groeiende biofilm structuur geeft en bescherming biedt. Kortom, alleen Rcs kwam naar voren in onze screening van mogelijke oppervlaktewaarnemende systemen en onze bevindingen suggereren dat dit systeem de aanpassing van de cel aan groei in een ontwikkelende biofilm reguleert. Ten slotte bestudeerden we in hoofdstuk 5 de activatie van CpxR door zeer lage concentraties van koperionen. Activatie kan verlopen via de bijbehorende histidine kinase CpxA, maar gebeurt ook in afwezigheid van dit eiwit zolang de histidine kinase CusS aanwezig is. Deze signaaloverdracht tussen de CusSR en CpxAR systemen zou veroorzaakt kunnen worden door de hoge mate van homologie, maar kan ook zijn ontstaan vanwege de essenti¨ele rol van CpxR in bescherming tegen koper, die blijkt uit de bacteriostatische effecten van zeer lage koperconcentraties in ∆cpxR cellen. Verder laten onze resultaten zien dat conditionering van het medium, specifiek door excretie van orootzuur, cellen kan helpen bij het omgaan met nadelige omstandigheden, in dit geval de aanwezigheid van koper.
Dankwoord
Het begon zes jaar geleden toen Matthias mij tijdens een kop koffie vroeg een masterproject in zijn groep te komen doen. Na het afronden van dit project en de rest van mijn master was het min of meer vanzelfsprekend dat ik mijn promotieonderzoek in de groep van Matthias zou gaan doen. Dat is een goede beslissing geweest, waar ik absoluut geen spijt van heb gehad.
Matthias, bedankt dat je me de mogelijkheid hebt geboden mijn promotieonderzoek onder jouw begeleiding te doen. Bedankt voor vier jaar lang meedenken, al je adviezen en je vertrouwen in mij, waardoor ik uiteindelijk dit proefschrift heb kunnen schrijven. De beoordelingscommissie, Wilbert Bitter, Jan-Maarten van Dijl en Oscar Kuipers, hartelijk dank voor het lezen en beoordelen van dit proefschrift.
Yonathan en Alexine, bedankt dat jullie mijn paranimfen willen zijn. Yonathan, voor ons begon het hier ongeveer tegelijkertijd. Wij deden hier beiden ons eerste masteronderzoek en beiden besloten we terug te komen voor onze PhD. Bedankt voor je fijne gesprekken. Jammer dat onze goede bedoeling een bandje op te richten nooit echt van de grond is gekomen. Succes met het afronden van je PhD. Alexine, we kennen elkaar nog niet zo lang, maar we hebben veel tijd samen doorgebracht in het lab. Bedankt voor alle gezellige momenten.
Zheng, we spent so much time together struggling with the optical tweezers. Although these long experiments in the cold and noisy microscope room could be quite frustrating, especially when at the end of the experiment all cells were lost when a bubble passed through the microfluidic channel, I’m glad that we could work on this project together. I value our conversations in which I learned a lot about China. Kuba, Hannah, Zheng and Ying, thank you for all your critical questions and helpful suggestions during the weekly E. coli meetings and for your help in the lab.
Thanks to all the people I shared an office with in the last four years, and in particular Serdar and Tamara, for the great atmosphere. I also want to thank Alex, Silke, Thanasis, Vakil and all other current and former members of our group for making the lab such an enjoyable place to work.
I would also like to thank some people for my stay in Sweden. Johan Elf, Michael Lawson, Prune Leroy, David Fange, Petter Hammar, ¨Ozden Baltekin och Alexis Boucharin, tack vare er min tid i Uppsala var j¨attetrevligt. Ni l¨arde mig om mikroskopi, mikrofluidik och bildanalys, som jag kunde anv¨anda p˚a forskningen f¨or min PhD. Tack s˚a mycket!
Ook wil ik nog mijn familie noemen. Marian en Dick, bedankt voor jullie hulp in een moeilijke tijd, waardoor ik op sommige momenten door kon gaan met mijn experimenten.
Als laatste wil ik mijn moeder bedanken. Lieve mam, bedankt dat je er altijd voor me bent, voor het aanhoren van al mijn verhalen en je steun wanneer mijn onderzoek weer eens tegenzat en voor je onvoorwaardelijke vertrouwen en geloof in mij. Bedankt dat je me de mogelijkheid hebt geboden zo ver te komen.
