An Ultrasensitive Bacterial Motor Revealed by Monitoring Signaling Proteins in Single Cells
10.1126/science.287.5458.1652
Crossref journal-article
American Association for the Advancement of Science (AAAS)
Science (221)
Abstract

Understanding biology at the single-cell level requires simultaneous measurements of biochemical parameters and behavioral characteristics in individual cells. Here, the output of individual flagellar motors in Escherichia coli was measured as a function of the intracellular concentration of the chemotactic signaling protein. The concentration of this molecule, fused to green fluorescent protein, was monitored with fluorescence correlation spectroscopy. Motors from different bacteria exhibited an identical steep input-output relation, suggesting that they actively contribute to signal amplification in chemotaxis. This experimental approach can be extended to quantitative in vivo studies of other biochemical networks.

Bibliography

Cluzel, P., Surette, M., & Leibler, S. (2000). An Ultrasensitive Bacterial Motor Revealed by Monitoring Signaling Proteins in Single Cells. Science, 287(5458), 1652–1655.

Authors 3
  1. Philippe Cluzel (first)
  2. Michael Surette (additional)
  3. Stanislas Leibler (additional)
References 48 Referenced 523
  1. A rare example of single-cell study of biochemical networks is provided by a recent analysis of Xenopus oocytes which revealed the mitogen-activated protein kinase cascade as a sharp epigenetic switch [
  2. 10.1126/science.280.5365.895
  3. 10.1038/262467a0
  4. Spiro P. A., Parkinson J. S, Othmer H. G., Proc. Natl. Acad. Sci. U.S.A. 94, 7263 (1997). (10.1073/pnas.94.14.7263) / Proc. Natl. Acad. Sci. U.S.A. by Spiro P. A. (1997)
  5. Alon U., et al., EMBO J. 17, 4238 (1998). (10.1093/emboj/17.15.4238) / EMBO J. by Alon U. (1998)
  6. Scharf B. E., Fahrner K. A., Turner L., Berg H. C., Proc. Natl. Acad. Sci. U.S.A. 95, 201 (1998). (10.1073/pnas.95.1.201) / Proc. Natl. Acad. Sci. U.S.A. by Scharf B. E. (1998)
  7. Kuo S. C., Koshland D. E., J. Bacteriol. 171, 6279 (1989). (10.1128/jb.171.11.6279-6287.1989) / J. Bacteriol. by Kuo S. C. (1989)
  8. Bray D., Levin M. D., Morton-Firth C. J., Nature 393, 6685 (1998). (10.1038/30018) / Nature by Bray D. (1998)
  9. R. M. Macnab in Two Component Signal Transduction J. A. Hoch and T. J. Silhavy Eds. (American Society for Microbiology Washington DC 1995) pp. 181–199.
  10. Larsen S. H., Reader R. W., Kort E. N., Tso W.-W., Adler J., Nature 249, 74 (1974). (10.1038/249074a0) / Nature by Larsen S. H. (1974)
  11. Berg H. C., Brown D. A., Nature 239, 500 (1972). (10.1038/239500a0) / Nature by Berg H. C. (1972)
  12. Welch M., Oosawa K., Aizawa S., Eisenbach M., Proc. Natl. Acad. Sci. U.S.A. 90, 8787 (1993); (10.1073/pnas.90.19.8787) / Proc. Natl. Acad. Sci. U.S.A. by Welch M. (1993)
  13. Barak R., Eisenbach M., Biochemistry 31, 1821 (1992). (10.1021/bi00121a034) / Biochemistry by Barak R. (1992)
  14. In addition to cheY this strain also lacks two other genes coding for CheB which demethylates the receptors and CheZ which dephosphorylates CheY-P. Several lines of evidence from in vitro data and (4) indicate that in this strain essentially all of the CheY present in the cytoplasm is phosphorylated. To verify if the presence of the GFP modified the phosphorylation of CheY-GFP we performed in vitro kinase assays (27). There is no noticeable difference between phosphorylation rates for CheY and CheY-GFP.
  15. CheY-GFP was expressed under control of lacOP with the low-copy plasmid pMGS98 (CmR). This plasmid was engineered by replacing the cheY gene in pLC576 (4) with an in-frame cheY-gfp fusion constructed by polymerase chain reaction amplification of the cheY gene (including the ribosome binding site) and gfp#2 [
  16. 10.1016/0378-1119(95)00685-0
  17. The glass slide was coated with amino-silanes and rabbit antibodies to flagellin that were covalently cross-linked to amino-silanes with glutaraldehyde. Transitions from CW to CCW and from CCW to CW states were manually scored from slowed-down (to 1/3 speed) video recordings and recorded on charts with LabView software (National Instrument Austin TX). Errors on the timing of transitions were less than 0.2 s. The incident excitation light was provided by the 488-nm line of an argon laser (air-cooled ion laser 160 Spectra Physics Mountain View CA). The laser beam was expanded to a diameter of 4 mm attenuated with a neutral density filter down to 0.1 mW and fed to an oil immersion objective ×100 (numerical aperture = 1.3 Olympus Uplan FI Melville NY). We used a dichroic mirror (Q505LP Chroma Brattleboro VT) to set up an epifluorescence illumination. The green fluorescence emission was rejected from the side of the microscope with a cold mirror (CP-SM-550 CVI Livermore CA). The residual blue component was filtered out from a long-pass green filter (06-515 CVI). A convergent lens focused the outcoming beam onto the core (50 μm) of a multimode optical fiber (Spectran Sturbridge MA). The optical fiber was connected to an avalanche photodiode (SPCM EG&G Quebec Canada) that delivered a TTL pulse for each detected photon.
  18. Magde D., Elson E., Webb W. W., Biopolymers 13, 29 (1974); (10.1002/bip.1974.360130103) / Biopolymers by Magde D. (1974)
  19. 10.1002/bip.1974.360130102
  20. ; R. Rigler in Fluorescence Correlation Spectroscopy O. S. Wolfbeis Ed. (Springer-Verlag Berlin 1992) pp. 13–24. (10.1007/978-3-642-77372-3_2)
  21. 10.1073/pnas.96.18.10123
  22. This method allowed us fast and consecutive concentration measurements of cells of interest without bleaching irreversibly (less 5%) the GFP. The lower data point (0 μM) gives the amplitude of the background noise. The limit of the detection ranged from ∼0.15 to 50 μM corresponding to a background noise of three molecules and an upper limit of ∼1000 molecules in the detection volume respectively. Because of the errors in focusing and alignment of the laser spot with respect to the cell body concentration could not be determined with precision better than 15%.
  23. Eigen M., Rigler R., Proc. Natl. Acad. Sci. U.S.A. 91, 5740 (1994); (10.1073/pnas.91.13.5740) / Proc. Natl. Acad. Sci. U.S.A. by Eigen M. (1994)
  24. Maiti S., Haupts U., Webb W. W., Proc. Natl. Acad. Sci. U.S.A. 94, 11753 (1997). (10.1073/pnas.94.22.11753) / Proc. Natl. Acad. Sci. U.S.A. by Maiti S. (1997)
  25. To check whether the presence of a GFP molecule in the CheY-GFP fusion could modify the cooperativity of the binding with the motors we analyzed on video recordings the swimming behavior of individual cells. Two populations of cells carrying either wild-type CheY or CheY-GFP (both expressed from the same low-copy plasmid vector) were grown under the same conditions (see Fig. 2A) and preinduced with 5 μM of IPTG. From these two populations we reported on two histograms the number of cells (out of 60) having either a “smooth ” “intermediate ” or “tumbly” behavior. We defined these three types of behavior as “smooth” for cells having a tumbling frequency (TF) inferior to 0.1 “intermediate” when 0.1 < TF < 1 and “tumbly” when TF > 1. We followed individual cells and manually scored the number of tumbles. We computed TF for single cells as the ratio of the number of tumbles to the duration of the recording. The two histograms were remarkably similar (27). This rather arduous analysis of the cell behavior provides a strong indication that the presence of a GFP molecule in the CheY-GFP fusion should not affect substantially the cooperativity of the binding to the motors and CheY phosphorylation.
  26. The reported times for folding and oxidation of the chromophore of the GFP in vitro were on the order of 10 min and 1 hour respectively [
  27. Reid B. G., Flynn G. C., Biochemistry 36, 6786 (1997); (10.1021/bi970281w) / Biochemistry by Reid B. G. (1997)
  28. ]. We performed experiments in which CheY-GFP was preinduced to different levels for more than 1 hour and thus had time for GFP chromophore formation. For the experiments in which we followed the induction process in the single cell (Fig. 2A) the data points fell on the same curve. This suggests that chromophore oxidation and the folding of the GFP might be faster in vivo than in vitro (28). Our observations were also in agreement with a recent study on a series of NH 2 -terminal fusions of proteins to GFP where the folding of the chromophore was observed to be tightly coupled to the folding of the fused protein (28).
  29. Ishahara A., Segall J. S., Block S. M., Berg H. C., J. Bacteriol. 155, 228 (1983); (10.1128/jb.155.1.228-237.1983) / J. Bacteriol. by Ishahara A. (1983)
  30. . The top curve of figure 5 in the Ishihara et al. article shows a high degree of correlation in the bias between two neighboring motors within the same cell. This result strongly suggests that motors are responding identically to the same signal implying a uniformity of their functioning.
  31. We calculated the Hill coefficient from the Hill plot in a standard fashion by retaining all data points with the CW bias between 0.1 and 0.9 (these limits do not necessarily correspond here to the physiologically relevant range of bacterial behavior).
  32. Standard techniques of immunoblotting ignored the inherent diversity of bacteria which is especially acute when proteins in the cells are expressed from plasmids. Variability of protein distribution within a population of cells depends on precise experimental conditions and the choice of strains plasmids or promoters [see e.g.
  33. 10.1073/pnas.94.3.814
  34. ]. Therefore in all of these methods the true output characteristic of flagellar motors was effectively convoluted with [CheY-P] distributions leading typically to smooth-out characteristics with lower Hill coefficients. To mimic the use of immunoblots we simulated on a computer the convolution of our data (sigmoid motors characteristic curve Fig. 2A) with a Gaussian distribution of proteins (σ = 24% of the mean) [
  35. Levin M. D., Morton-Firth C. J., Abouhamad W. N., Bourret R. B., Bray D., Biophys J. 74, 175 (1998); (10.1016/S0006-3495(98)77777-X) / Biophys J. by Levin M. D. (1998)
  36. ]. In good agreement with our data and (4–6) we found that Hill coefficients ranging from 8 to 20 were respectively smoothed out from 3.9 to 4.2 when convoluted with such distribution. Generally immunoblots may thus be inadequate tools for defining sharp slopes of biological switches. To check the validity of these computer simulations we performed the same experiment as in (4) with wild-type CheY and CheY-GFP fusion. For several concentrations of IPTG (0 1 2.5 and 5 μM) the two populations of cells showed the same swimming behavior [thanks to U. Alon these data were provided with the same setup as in (4)]. In both cases wild-type CheY and CheY-GFP fusion reproduced identical results to those reported in figure 5 of (4) (27).
  37. Welch M., Oosawa K., Aizawa S.-I., Eisenbach M., Biochemistry 33, 10470 (1994); (10.1021/bi00200a031) / Biochemistry by Welch M. (1994)
  38. Zhao R., Amsler C. D., Matsumura P., Khan S., J. Bacteriol. 178, 258 (1996). (10.1128/jb.178.1.258-265.1996) / J. Bacteriol. by Zhao R. (1996)
  39. 10.1126/science.7089556
  40. Goldbeter A., Koshland D. E., Proc. Natl. Acad. Sci. U.S.A. 78, 6840 (1981). (10.1073/pnas.78.11.6840) / Proc. Natl. Acad. Sci. U.S.A. by Goldbeter A. (1981)
  41. CheZ is a specific phosphatase interacting with CheY-P. It has recently been discovered in vitro [
  42. Blat Y., Gillepsie B., Bren A., Dahlquist F. W., Eisenbach M., J. Mol. Biol. 284, 1191 (1998); (10.1006/jmbi.1998.2224) / J. Mol. Biol. by Blat Y. (1998)
  43. ] that the CheZ activity is cooperative with respect to CheY-P and that the phosphatase activity is regulated at the level of the ratio CheY-P/CheZ interaction [
  44. Scharf B. E., Fahrner K. A., Berg H. C., J. Bacteriol. 180, 5123 (1998); (10.1128/JB.180.19.5123-5128.1998) / J. Bacteriol. by Scharf B. E. (1998)
  45. ]. It might be tempting to think that CheZ acts as a CheY-P concentration regulator. CheZ would contribute to the adjustment of the CheY-P concentration in the functioning range of the motors (3 μM). As noted by Scharf et al. “the swarming abilities of both strains RP437 and AW405 suggested that the absolute values of CheY and CheZ are not as crucial for chemotaxis as their ratios” (p. 5125). Conversely one could argue that the natural level of expression of the CheZ gene is sufficient without any further regulation to dephosphorylate the right amount of CheY-P to adjust the system to the operational range of the motors.
  46. P. Cluzel M. Surette S. Leibler data not shown.
  47. Waldo G. S., Standish B. M., Berendzen J., Terwilliger T. C., Nature Biotechnol. 17, 691 (1999). (10.1038/10904) / Nature Biotechnol. by Waldo G. S. (1999)
  48. We thank U. Alon N. Barkai G. Bonnet M. Elowitz T. Griggs C. Guet T. Silhavy J. Stock J. Vilar and E. Winfree for many helpful discussions and comments on the manuscript. This work was partially sponsored by the NIH. P.C. acknowledges support by a fellowship from the Program in Mathematics and Molecular Biology at the Florida State University with funding from the NSF under grant DMS-9406348.
Dates
Type When
Created 23 years ago (July 27, 2002, 5:40 a.m.)
Deposited 1 year, 7 months ago (Jan. 13, 2024, 5:44 a.m.)
Indexed 2 weeks ago (Aug. 12, 2025, 6:18 p.m.)
Issued 25 years, 5 months ago (March 3, 2000)
Published 25 years, 5 months ago (March 3, 2000)
Published Print 25 years, 5 months ago (March 3, 2000)
Funders 0

None

@article{Cluzel_2000, title={An Ultrasensitive Bacterial Motor Revealed by Monitoring Signaling Proteins in Single Cells}, volume={287}, ISSN={1095-9203}, url={http://dx.doi.org/10.1126/science.287.5458.1652}, DOI={10.1126/science.287.5458.1652}, number={5458}, journal={Science}, publisher={American Association for the Advancement of Science (AAAS)}, author={Cluzel, Philippe and Surette, Michael and Leibler, Stanislas}, year={2000}, month=mar, pages={1652–1655} }