Morphogen-defined patterning of Escherichia coli enabled by an externally tunable band-pass filter
© Sohka et al; licensee BioMed Central Ltd. 2009
Received: 07 May 2009
Accepted: 08 July 2009
Published: 08 July 2009
Gradients of morphogens pattern cell fate – a phenomenon that is especially important during development. A simple model system for studying how morphogens pattern cell behavior would overcome difficulties inherent in the study of natural morphogens in vivo. A synthetic biology approach to building such a system is attractive.
Using an externally-tunable band-pass filter paradigm, we engineered Escherichia coli cells to function as a model system for the study of how multiple morphogens can pattern cell behavior. We demonstrate how our system exhibits behavior such as morphogen crosstalk and how the cells' growth and fluorescence can be patterned in a number of complex patterns. We extend our cell patterning from 2D cultures on the surface of plates to 3D cultures in soft agarose medium.
Our system offers a convenient, well-defined model system for fundamental studies on how multiple morphogen gradients can affect cell fate and lead to pattern formation. Our design principles could be applied to eukaryotic cells to develop other models systems for studying development or for enabling the patterning of cells for applications such as tissue engineering and biomaterials.
Morphogens are substances that induce different cell fates in a concentration dependent manner [1, 2]. Morphogens are especially important in development because emission of a morphogen from a source can lead to the formation of different cell types in a defined spatial relationship to the source and to each other. The study of natural morphogens during development is challenging owing to the complexity of in vivo systems and the extremely low concentrations of most morphogens [3, 4]. The engineering of less complex biological systems with position-specific gene expression would provide convenient model systems for testing fundamental concepts and hypotheses of how morphogens can pattern cell fate.
Synthetic biology seeks to create biological systems with desired properties and behaviors [5–7], including the ability to pattern cell behavior. For example, genetic circuits have been built in E. coli cells such that fluorescent proteins are expressed only when a particular compound falls within a certain concentration range [8, 9]. Such cells function similar to a band-detect filter and the fluorescence pattern that emerges follows the contour lines of the gradient of the compound.
We have recently engineered E. coli cells to function as a band-pass filter for β-lactam antibiotics and cellular β-lactamase (BLA) activity in which the position and width of the band can be tuned externally . When challenged to grow in the presence of the β-lactam antibiotic ampicillin (Amp) and tetracycline (Tet), cells grow and fluoresce only if the cellular β-lactamase activity (which is regulated through an IPTG-inducible promoter) falls within a narrow range. In this system, Amp and IPTG function as "morphogens" and can be used together with Tet to pattern cell growth and fluorescence. Amp and IPTG are not true morphogens because E. coli is prokaryotic. However, this simple model system can be used for fundamental studies on how morphogens can pattern cell behavior . For example, the opposing effects of Amp and IPTG were used to break the radial symmetry imposed by radial concentration gradients and pattern cell growth in linear and other patterns that lacked radial symmetry (i.e. cells were patterned in defined ways that did not follow contour lines of the concentration gradient).
In this report we describe how the inclusion of an additional promoter that responds to a different cell permeable molecule can result in more complex patterns of growth and gene expression. In response to morphogen/chemical gradients, mixtures of cells with different levels of enzyme activity and different fluorescent protein genes achieved position-specific multi-gene expression patterns resulting in a complex multicolor pattern. In addition, we demonstrate how morphogens in our system can exhibit crosstalk, and we expand our pattering methods from two-dimensions on the surface of a plate to a three-dimensional culture.
Results and discussion
Band-pass growth behavior
Since, intracellular Amp concentration is reduced by cellular β-lactamase activity, the enzyme (BLA) and substrate (Amp) pair acts as an attenuator for the entire system (Fig. 1a). Increased enzyme activity shifts the Amp concentration range required for growth to higher levels. Likewise, an increase in Amp concentration requires an increase in enzyme activity for growth. In this manner, the band-pass filter is externally tunable. In liquid culture, growth requires a particular ratio of β-lactamase activity to Amp. The same is true for growth on agar plates. Ring radius is the same for cells with very different β-lactamase activity provided the ratio of Amp to β-lactamase activity is the same (Fig 1c). Our cells function like a sensor and can be used to detect enzyme activity or substrate concentration in a very narrow range and with wide dynamic range. In addition, both Amp and IPTG function as morphogens  that fit Wolpert's French flag model . Thus, the band-pass paradigm transforms E. coli into a simple model system for how morphogens affect cell fate during development.
Crosstalk between different morphogens
In this experiment, both strains' ability to hydrolyze Amp was set to be the same by using the IPTG concentration necessary to achieve a minimum inhibitory concentration (MIC) for Amp of 64 μg/ml. In the case of cells expressing P99-BLA, the effect of CA and Amp appears to be largely additive (Fig. 2), as expected. Fluorescence is not seen in between the CA disk and the closest Amp disc because the total concentration of Amp and CA in this region is too high for growth. CA does not affect the concentration of Amp because it is too poor an inhibitor of P99-BLA. In the case of cells expressing TEM1-BLA, crosstalk is observed presumably because CA distorts the Amp gradient through inhibition of TEM1-BLA. This expands the region of no growth beyond what is seen in cells expressing P99-BLA (Fig. 2). CA and Amp are acting synergistically since CA serves to help Amp prevent cell growth by inhibiting TEM1-BLA's Amp degradation activity. These experiments illustrate that our model system can be used to study crosstalk between multiple morphogens. In addition, our system can be used in a simple visual method for characterizing substrate/inhibitor specificity of β-lactamases.
Patterning of cell growth and fluorescence
Our system can be used to study how simple growth patterns and gene expression profiles can emerge from a homogenous population of cells. However, actual developmental phenomena are much more complicated than this. After cell fate has been patterned by one morphogen, cell behavior might be further regulated by other signals. We sought to capture this in our simple model system by engineering the cells to respond to the signal of another cell permeable molecule. To achieve this, we added a third compatible vector that had a red fluorescent protein gene (RFPec) under the control of an arabinose inducible promoter (Fig. 3b). By the appropriate patterning of Amp, IPTG and arabinose, we achieved localized RFPec expression without any change to the growth patterning (Fig. 3c). Thus, in addition to designating where cells grow, we can further pattern the cells with different phenotypes in a location-dependent manner. This is exemplified by our patterning of straight lines of cells that are precisely segmented in color (Fig 3c).
Multi-cellular separation and patterning of fluorescence
Three-dimensional growth patterning
External tunability and multiple morphogens are important attributes for a useful model developmental system and key for achieving complex patterning of cell behavior. We have demonstrated how a simple band-pass filter genetic circuit in E. coli can enable external tunability and can have multiple morphogens with crosstalk and opposing effects. This enables complex patterning including multi-colored line drawing (Fig 3c) and the separation of different cell types into complex patterns based on enzyme activity (Fig 4). Our cells can be patterned in one-dimension (lines), two dimensions (e.g. rings on plates) and three dimensions (e.g. spheres). Our system thus offers a convenient, well-defined model system for fundamental studies on how morphogens gradients can affect cell fate and lead to pattern formation. Our design principles for engineering an externally tunable band-pass filter could be applied to eukaryotic cells to develop other models systems for studying development or for enabling the patterning of cell fate for applications such as tissue engineering and biomaterials.
Such studies would be simplified if morphogen gradients were permanently established instead of evolving over time. In our experiments, the morphogen gradient was continuously changing over the experimental period due to diffusion and degradation. This results in broad, complex gradients that are more difficult to accurately model, and the resulting patterns have relatively large dimensions (i.e. mm to cm scale). With steeper and more stable morphogen gradients, more precise and smaller scale patterning should be possible. The combination of continuous production  and rapid degradation of the morphogen at defined locations is one possible strategy for achieving this.
Plasmid constructions for band-pass filter have previously described . Plasmid maps and regulatory relationships are summarized in Fig. 1a. The reporter plasmid (pTS1) has streptomycin (Sm)/spectinomycin (Sp) resistance and carries the CloDF13 replication origin. The β-lactamase expression plasmid pDIM-C8 has chloramphenicol (Cm) resistance and a p15a replication origin, which is compatible with the CloDF13 replicon. All genes cloned into pDIM-C8 are under the control of the IPTG inducible tac promoter whose leaky expression was suppressed by LacI from pTS1. Several versions of the pDIM-C8 were constructed and named with an abbreviation for the name of the gene in place of the "X".
pBAD-Fluor (Fig 3b) confers kanamycin (Kan) resistance and has a ColE1 replication origin, which is compatible with the CloDF13 and p15a replicons. The rank of the plasmid copy number is pTS1 > pBAD-Fluor > pDIM-C8 as judged by ethidium bromide staining intensity plasmid DNA isolated from strains harboring all three plasmids. pBAD-Fluor was constructed by replacing the Amp R gene of pBAD/Myc-His A (Invitrogen) with Kan resistant gene with constitutive promoter of pET24d(+) (Novagen) using the BspHI site. The NcoI site (overlaps the ATG start codon) was eliminated and changed to BglII site by PCR based methods. Three version of plasmid pBAD-Fluor were constructed, each of which can express one of three fluorescent proteins (RFPec, eYFP or eCFP). The genes encoding eYFP and eCFP were PCR amplified from commercially available plasmid (Clontech) such that cloning sites BglII and KpnI (for ligation into pBAD-Fluor) were added at the 5' and 3' ends, respectively. The gene encoding RFPec was created by PCR amplification using mutagenic primers and DNA encoding DsRed2 (Clontech) as the template. In addition to introducing the BglII and KpnI sites for cloning, these primers introduced mutations in the 5' region of the DsRed2 gene to created RFPec – a modified version of DsRed2 that expresses very efficiently under the pBAD promoter in E. coli .
Strains and their properties are listed in Additional file 1. E. coli SNO301 (ampD1, ampA1, ampC8, pyrB, recA, rpsL) was used. The ampD1 mutation allows β-lactam hyper-inducible transcription of genes under the control of the ampC promoter . The rpsL mutation confers streptomycin resistance to this host, the same resistance marker used to maintain pTS1. However, we observed that pTS1 was stably maintained in SNO301 cells in the presence of 100 μg/ml Sm. The rpsL mutation does not confer spectinomycin (Sp) resistance. We used Sp instead of Sm in some experiments. Multiple aliquots of frozen cell stocks for plating and growth experiments were prepared from overnight LB cultures at a final optical density at 600 nm of 0.5 and contained 15% glycerol. These stocks were stored at -80°C.
LB-agar growth experiments
Ten-fold dilutions of frozen stocks (150 μl for 86 mm round plates and 206 μl for 91 mm square plates) were spread evenly over the entire LB-agar (1.5%) plates containing Sm (100 μg/ml), Cm (50 μg/ml) and various concentrations of IPTG, Tet and arabinose. For multi-fluorescent color pattering, plates also contained Kan (30 μg/ml). For the plate in Fig. 4, 50 μl of each of the three strains were mixed before plating. Plate volume was 20 ml for 86 mm round plates and 27.5 ml for 91 mm square plates. Ninety minutes after plating, dry one cm diameter paper discs (six mm discs for Fig. 2) were placed at the desired position using a schematic map laid underneath the plate to ensure reproducible placement. The disks were spotted with Amp, potassium clavulanate (CA), IPTG, arabinose and/or Tet as indicated. Some spots had mixtures of two compounds. The detailed conditions for each experiment can be found in Additional file 1. The plates were incubated at 37°C for 21 hours.
After disc removal growth images were acquired by CCD camera under the white light reflection conditions using ChemiDoc XRS (BioRad). Fluorescence images were acquired by Typhoon 9410 imager (GE Lifescience) using 200 μm resolution and agar top focused (+3 mm). Fluorescent images were taken under the following conditions. For imaging GFP (green): 350 V/488 nm laser excitation (ex), 526 nm short pass (SP) filter emission (em) detection; for imaging eCFP (cyan): 450 V/457 nm ex, 526 nm SP em; for imaging eYFP (yellow): 450 V/532 nm ex, 555 nm band pass filter em; and for imaging RFPec (Red): 500 V/633 nm ex, 580 nm SP em. Greyscale images were converted to color and the contrasts were adjusted by using Image Quant TL software (GE Lifescience). Due to the excitation laser wavelength and emission filter restrictions, it was not possible to completely isolate eCFP/GFP or eYFP/GFP signals. Thus, in Fig. 4, there is a small amount of GFP signal in the eCFP (cyan) and eYFP (yellow) signals. Two colors were merged in Fig. 3c (i.e. GFP (green) and RFPec (Red)). Three colors were merged in Fig. 4 (i.e. eCFP (cyan), eYFP (yellow) and RFPec (Red)).
Three-dimensional growth patterning
LB medium containing 0.5% (w/v) of low melting point agarose (LMP agarose; Invitrogen) was heated in a microwave until the agarose was molten and was then incubated at 39°C to keep from solidifying. Frozen stocks of RH01 cells were diluted 333-fold into this LB-agarose medium containing Sp (30 μg/ml), Cm (50 μg/ml) and Tet (30 μg/ml). IPTG (300 μM) was included in the sphere growth experiments. For sphere shaped growth (Fig. 5a), the Amp-containing cylinder was embedded in the LB-agarose-cell using a method involving three layers. First, 5 ml of the LB-agarose-cell solution was poured into a glass vial (size: 28 mm × 58 mm; Fisher) and placed on ice until solidified. Immediately after solidification, 0.5 ml of the LB-agarose-cell solution was layered on top. Next, a 2% (w/v) agarose cylinder gel (3 mm diameter × 4 mm height) containing Amp (2.13 mg/ml) was pushed into the center of the unsolidified thin solution. After this layer had solidified and immobilized the cylinder, a final 5 ml of the LB-agarose-cell solution was layered on top. After the final layer had solidified, the vial was incubated for 21 hr at 37°C. For layered growth (Fig. 5b), first 600 μl of 2% agarose was placed in the bottom of the vial. After solidification, a paper disc (1 cm) was placed on the center and 15 μl of 300 mM IPTG was spotted on it. The vial was incubated at room temperature for 6 hrs. Next, 10 ml of LB-agarose-cell solution was layered on top and the vial was placed on ice until the layer solidified. A second paper disc was placed on the gel, to which 15 μl of 12.5 mg/ml Amp was spotted. The vial was incubated for 21 hr at 37°C. In both experiments, the heights of the gel were about 2 cm. Growth images were acquired by digital camera (Canon) equipped with a macro lens under the white LED light from the bottom reflection conditions using super-macro mode. Contrasts and color balances were adjusted by using Microsoft photo editor to better visualize the growth pattern.
Strain descriptions and detailed disc pattern information can be found in Additional file 1.
This work was supported by NIH/NIGMS and a gift from Asahi Kasei Corporation.
- Wolpert L: Positional information and the spatial pattern of cellular differentiation. J Theor Biol 1969, 25: 1-47. 10.1016/S0022-5193(69)80016-0View ArticleGoogle Scholar
- Wolpert L: One hundred years of positional information. Trends Genet 1996, 12: 359-364. 10.1016/S0168-9525(96)80019-9View ArticleGoogle Scholar
- Ashe HL, Briscoe J: The interpretation of morphogen gradients. Development 2006, 133: 385-394. 10.1242/dev.02238View ArticleGoogle Scholar
- Gurdon JB, Bourillot PY: Morphogen gradient interpretation. Nature 2001, 413: 797-803. 10.1038/35101500View ArticleGoogle Scholar
- Andrianantoandro E, Basu S, Karig DK, Weiss R: Synthetic biology: new engineering rules for an emerging discipline. Mol Syst Biol 2006., 2:Google Scholar
- Drubin DA, Way JC, Silver PA: Designing biological systems. Genes Dev 2007, 21: 242-254. 10.1101/gad.1507207View ArticleGoogle Scholar
- Tan C, Song H, Niemi J, You L: A synthetic biology challenge: making cells compute. Mol Biosyst 2007, 3: 343-353. 10.1039/b618473cView ArticleGoogle Scholar
- Basu S, Gerchman Y, Collins CH, Arnold FH, Weiss R: A synthetic multicellular system for programmed pattern formation. Nature 2005, 434: 1130-1134. 10.1038/nature03461View ArticleGoogle Scholar
- Bernhardt K, Carter EJ, Chand NS, Lee J, Xu Y, Ajioka JW, Goncalves JM, Haseloff J, Micklem G, Rowe D: New tools for self-organised pattern formation. IET Synthetic Biology 2007, 1: 29-31. 10.1049/iet-stb:20070020View ArticleGoogle Scholar
- Sohka T, Heins RA, Phelan RM, Greisler JM, Townsend CA, Ostermeier M: An externally-tunable bacterial band-pass filter. Proc Natl Acad Sci USA 2009, 106: 10135-10140. 10.1073/pnas.0901246106View ArticleGoogle Scholar
- Jacobs C, Frere JM, Normark S: Cytosolic intermediates for cell wall biosynthesis and degradation control inducible beta-lactam resistance in gram-negative bacteria. Cell 1997, 88: 823-832. 10.1016/S0092-8674(00)81928-5View ArticleGoogle Scholar
- Valtonen SJ, Kurittu JS, Karp MT: A luminescent Escherichia coli biosensor for the high throughput detection of beta-lactams. J Biomol Screen 2002, 7: 127-134.Google Scholar
- Baggaley KH, Brown AG, Schofield CJ: Chemistry and biosynthesis of clavulanic acid and other clavams. Nat Prod Rep 1997, 14: 309-333. 10.1039/np9971400309View ArticleGoogle Scholar
- Beharry Z, Chen H, Gadhachanda VR, Buynak JD, Palzkill T: Evaluation of penicillin-based inhibitors of the class A and B beta-lactamases from Bacillus anthracis. Biochem Biophys Res Commun 2004, 313: 541-545. 10.1016/j.bbrc.2003.11.158View ArticleGoogle Scholar
- Payne DJ, Du W, Bateson JH: beta-Lactamase epidemiology and the utility of established and novel beta-lactamase inhibitors. Expert Opin Investig Drugs 2000, 9: 247-261. 10.1517/135437184.108.40.206View ArticleGoogle Scholar
- Tsien RY: The green fluorescent protein. Annu Rev Biochem 1998, 67: 509-544. 10.1146/annurev.biochem.67.1.509View ArticleGoogle Scholar
- Pfleger BF, Fawzi NJ, Keasling JD: Optimization of DsRed production in Escherichia coli: effect of ribosome binding site sequestration on translation efficiency. Biotechnol Bioeng 2005, 92: 553-558. 10.1002/bit.20630View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.