Lasers play an important role in many parts of modern life, with applications ranging from materials process ing over optical communication to medicine. In order to tailor the properties of lasers to the specific needs of these applications, a wide range of different device configurations and materials has been tested. Today, materials in all of the three aggregate forms—gaseous, liquid, and solid—are widely used as optical gain materials. Until recently, however, little work has been done on using biological materials as the active parts of lasers, although biocompatible and bioderived lasers might enable previously unthinkable approaches for biosensing, light-based diagnosis and therapies as lasers based on biologically grown components can be more naturally integrated into biological systems. Several reports have shown that biological tissue can support random lasing when infiltrated with concentrated laser dye solutions [1
1. M. Siddique, L. Yang, Q. Z. Wang, and R. R. Alfano, Opt. Commun. 117, 475 (1995). [CrossRef]
2. L. Wang, D. Liu, N. He, S. L. Jacques, and S. L. Thomsen, Appl. Opt. 35, 1775 (1996). [CrossRef] [PubMed]
3. R. C. Polson and Z. V. Vardeny, Appl. Phys. Lett. 85, 1289 (2004). [CrossRef]
4. Q. Song, S. Xiao, Z. Xu, J. Liu, X. Sun, V. Drachev, V. M. Shalaev, O. Akkus, and Y. L. Kim, Opt. Lett. 35, 1425 (2010). [CrossRef] [PubMed]
]. Recently, DNA strands were used as nanoscale scaffolds to adjust the separation between two different fluorescent dye molecules in bioinspired microfluidic lasers based on Förster resonance energy transfer (FRET) [5
5. Y. Sun, S. I. Shopova, C.-S. Wu, S. Arnold, and X. Fan, Proc. Natl. Acad. Sci. USA 107, 16039 (2010). [CrossRef] [PubMed]
]. However, all of these structures relied on synthetic optical gain materials.
We have recently reported the generation of laser light from a fluorescent protein and have shown that single mammalian cells that are transiently transfected to express the green fluorescent protein (GFP) can form the active component of a microlaser [6
6. M. C. Gather and S. H. Yun, Nat. Photonics 5, 406 (2011). [CrossRef]
]. GFP is a highly fluorescent macromolecule that is synthesized by nu merous genetically engineered organisms [7
]. In this Letter, we show lasing from colonies of GFP-expressing Escherichia coli
are rod-shaped bacteria that are commonly found in the intestine of humans and animals; most strains are nonharmful. E. coli
also plays an important role in modern biotechnology and synthetic biology [8
8. S. Y. Lee, ed., Systems Biology and Biotechnology of Escherichia Coli (Springer, 2009). [CrossRef]
]. The bacteria can be genetically manipulated to produce valuable substances, such as human insulin [9
9. D. V. Goeddel, D. G. Kleid, F. Bolivar, H. L. Heyneker, D. G. Yansura, R. Crea, T. Hirose, A. Kraszewski, K. Itakura, and A. D. Riggs, Proc. Natl. Acad. Sci. USA 76, 106 (1979). [CrossRef] [PubMed]
] and biofuels [10
10. E. J. Steen, Y. Kang, G. Bokinsky, Z. Hu, A. Schirmer, A. McClure, S. B. d. Cardayre, and J. D. Keasling, Nature 463, 559 (2010). [CrossRef] [PubMed]
], so that these compounds can be harvested in large quantities at low cost for use in medicine and energy. In general, genetic engineering of bacteria tends to be more robust and more efficient than for mammalian cells. In contrast to our previous work on mammalian cells, the genetic transformation of the bacteria used here with a plasmid encoding for GFP expression was stable. This means that the capability to synthesize fluorescent protein is maintained during cell replication. Stable transformation allows for long-term use of bio logical lasers, for instance, by fully harnessing their self-healing nature, the ability to replenish the optical gain-material in response to photo-bleaching of GFP.
To estimate the average amount of GFP per bacteria cell, we measured the brightness of individual cells with a fluorescence microscope ( objective) equipped with a CCD camera (Olympus IX51, SPOT RT3). The measurement was calibrated against the brightness of a recombinant GFP solution held between two microscope slides with a defined separation. We found an average GFP concentration of around , with a variation between bacteria cells. This corresponds to GFP molecules within each bacteria cell (cell volume is approximately one femtoliter).
A single colony of GFP-expressing bacteria was picked from a plate and transferred with an inoculating loop onto the surface of a dielectric mirror with high- reflectivity (
) in the
range of the spectrum and good transmission for blue-light (Y2 coating, CVI). Figure 1a
shows a scanning electron microscopy (SEM) image of the bacteria colony and reveals that cells are closely packed. Individual E.coli
have an elongated cylindrical shape with an average length and diameter of
, respectively. For the laser experiment, the colony was covered with a drop of glycerol to reduce scattering of light at the cell wall interfaces and then covered with a second mirror to form a Fabry–Perot type laser cavity. Spacer beads (diameter,
) were inserted to protect the bacteria from rupturing due to the weight of the top-mirror and to control the cavity length.
We used the pulsed output of an optical parametric oscillator (OPO) that was tuned to
to optically pump the GFP-expressing bacteria (Quanta Ray MOPO- 700, Spectra-Physics, pulse duration
). The energy of the excitation pulses was adjusted with neutral density filters and monitored with an energy meter. The pulses were reflected into the laser cavity described above by a dichroic mirror (
long-pass) and focused with a
achromatic lens. The pulse energies quoted in the following are the energy of single pulses behind the focusing lens, but before the front mirror of the cavity. The emission from the bacteria culture was collected with the same lens, separated from back-reflected pump light by the dichroic mirror and fed to a camera and a
spectrograph connected to a cooled CCD camera (Andor). See Fig. 1b
for a schematic illustration of the setup.
shows the integrated output from the bacteria filled resonator as a function of the energy of the pump pulses. A distinct threshold can be observed at a pump energy of
. This value is higher than what we reported previously for lasers based on recombinant protein solution and single mammalian cells. We attribute the increased threshold to a combination of two effects: (1) The presence of multiple small E.coli
cells in the cavity results in higher scattering loss during each round-trip than for a situation where only a single mammalian cell (typical diameter,
) is placed inside the resonator. (2) The plane–plane resonator configuration used here is only marginally stable and we do not expect that the closely packed bacteria cells between the mirrors provide an equally efficient lens effect for stabilizing the resonator as a single large mammalian cell [6
6. M. C. Gather and S. H. Yun, Nat. Photonics 5, 406 (2011). [CrossRef]
]. Thus, the gain provided by the GFP-expressing bacteria will have to compensate for substantial diffraction losses.
For pump energies just above the lasing threshold, the output spectrum showed a well-defined mode structure (Fig. 3
, top). The wavelength separation between the peaks (
) is consistent with the expected longitudinal mode separation (
is the refractive index of the glycerol filling the cavity). At higher pump energies, however, these lines expanded into an ensemble of closely spaced peaks (Fig. 3
, bottom). Such a spectral profile is indicative of simultaneous lasing action in multiple nondegenerate transverse laser modes [12
12. O. Svelto, Principles of Lasers (Springer–Verlag, New York, 2009).
shows the spatial output pattern of an E.coli
cell laser pumped with an extended excitation beam. For this experiment, the
focusing lens was replaced by a
microscope objective and the excitation beam was expanded to a diameter of
. Figures 4a
show the output when the structure is pumped close to the lasing threshold (
above threshold, respectively. Figure 4c
shows a coregistered bright field image of the bacteria laser. As expected, lasing is observed in regions where bacteria are densely packed and the optical gain provided by GFP molecules is high. The complex and seemingly random output pattern of the laser presumably results from the disordered ensembles of bacteria clusters that cause the laser to operate at higher order transverse laser modes. As the energy of the pump pulses was increased, additional parts of the sample began to lase [compare inset to Fig. 4b
] while the previously lasing regions substantially gained in brightness. We conclude that, at high pump energies the observed spatial patterns represent a superposition of several higher-order transverse laser modes. This is consistent with the observation that ensembles of peaks form in the output spectrum of the laser as the pump energy is increased (Fig. 3
, bottom). The lobe-size of the modes is in some cases consistent with the size of individual bacteria [inset to Fig. 4b
]. However, considering the dense and multilayered packing of E.coli
, we think the lobes were formed by coherent optical interference rather than representing individual bacteria.
In conclusion, we have shown that lasing can be generated from colonies of E.coli bacteria genetically transformed to synthesize fluorescent protein. Lasing was evidenced by clear threshold behavior and discrete peaks in the emission spectrum. Demonstration of lasing from bacteria that inherit the capability to synthesize GFP upon cell division is an important step towards large-scale self-sustained biological lasers. We note that the cavity mirrors are critical to the lasing action. In the absence of either of the mirrors, no lasing was observed up to a pump flux of . In the future, it will be interesting to see whether bacteria colonies can be configured to form random or ordered structures providing sufficient optical feedback to produce laser light without external mirrors.