Anal. Chem. 2001, 73, 4625-4631
Spatial Control of Cellular Measurements with the Laser Micropipet Huaina Li, Christopher E. Sims, Hayley Y. Wu, and Nancy L. Allbritton*
Department of Physiology and Biophysics and Center for Biomedical Engineering, University of California, Irvine, California 92697
Continued progress in understanding cellular physiology requires new strategies for biochemical measurements in solitary cells, multiple cells, and subcompartments of cells. Large spatial gradients in the concentrations of molecules and presumably the activities of enzymes can occur in cells. Consequently, there is a critical need for measurement techniques for mammalian cells with control over the numbers or regions of cells interrogated. In the present work, we developed a strategy to rapidly load the cytoplasmic contents of either multiple cells or a subregion of a single cell into a capillary. A single, focused pulse from a laser created a mechanical shock wave which disrupted a group of cells or a portion of a cell in the path of the shock wave. Simultaneously, the cytoplasm was loaded into a capillary for electrophoretic separation. The size of the region of cellular disruption (and therefore the volume of cytoplasm collected) was controlled by the amount of energy in the laser pulse. Higher energies could be used to sample groups of cells while much lower energies could be utilized to selectively sample the tip of a neuronal process. The feasibility of performing measurements on subcellular compartments was also demonstrated by targeting reporter molecules to these compartments. A reporter localized to the nucleus was detected on the electropherogram following laser-mediated disruption of the cell and the nucleus. Finally, we demonstrate that this method terminated cellular reactions with sufficient rapidity that cellular membrane repair mechanisms were not activated during cytoplasmic collection. The combined ability to preselect a spatial region of a cell or cells and to rapidly load that region into a capillary will greatly enhance the utility of CE in the biochemical analysis of cells. Progress in biology is fueled by the growth of new analytical techniques that can be applied to biologic systems. An example is the development of new strategies to analyze cells at the singlecell or subcellular level. Prominent among these techniques are a variety of microscopy-based and flow cytometric methods. Fluorescent indicators and green fluorescent protein (GFP) fusions with other proteins are arguably the most successful of * Corresponding author: (phone) 949-824-6493; (fax) 949-824-8540; (e-mail)
[email protected]. 10.1021/ac0105235 CCC: $20.00 Published on Web 08/17/2001
© 2001 American Chemical Society
these techniques.1,2 When combined with microscopy, both tools allow cellular processes to be monitored at high spatial resolution over time. The fluorescent indicators typically function by responding to the binding of a ligand with a shift in their fluorescent properties (quantum yield, emission wavelength, or excitation wavelength). They have been most successful for the measurement of intracellular ion concentrations (particularly calcium) but have not been generalizable to other ligands. With flow cytometry, the same indicators can be used to measure the properties of large numbers of cells in a short time but without spatial resolution and without multiple time points from the same cell. Both the indicator and GFP-tagging methods have opened new doors in biologic research; however, they continue to suffer from a number of weaknesses. To quantitate the fluorescence, the indicator or GFP-protein must frequently be present at concentrations that are substantially greater than similar molecules within the cells.3 For example, the GFP-protein may be expressed at concentrations that are orders of magnitude greater than that of the native protein in the cell. This overexpression can lead to artifactual alterations in cellular behavior. In addition, the number of cellular properties that can be simultaneously measured from the same cell is typically very low due to spectral overlap between the fluorophores. For living cells, two to three fluorescent probes are generally the maximum number that can be quantitated by microscopy except in highly idealized situations. Thus, only two to three properties of the cell are measured simultaneously. It is now clear that cellular signaling involves the combined interactions of large numbers of interconnected pathways with each pathway composed of many different proteins.4 How the proteins function as elements in the signaling network remains to be understood. A major reason is a lack of analytic tools that enable the measurement of large numbers of cellular properties simultaneously from the same single cell. Analysis of single cells by capillary electrophoresis (CE) holds the promise of simultaneously measuring multiple attributes of a single cell. This is primarily due to the high separation efficiency of CE and the ease with which picoliter samples are handled. The analysis of single cells by CE was pioneered by a number of groups including those of Yeung, Sweedler, Ewing, Jorgensen, (1) Zacharias, D. A.; Baird, G. S.; Tsien, R. Y. Curr. Opin. Neurobiol.. 2000, 10, 416-2. (2) Tsien, R. Y. Annu. Rev. Biochem. 1998, 67, 509-44. (3) Yokoe, H.; Meyer, T. Nat. Biotechnol. 1996, 14, 1252-6. (4) Jordan, J.; Landau, E. M.; Lyengar, R. Cell 2000, 103, 193-200.
Analytical Chemistry, Vol. 73, No. 19, October 1, 2001 4625
and Kennedy.5-16 Recently Dovichi and collaborators have begun to develop strategies employing CE to measure the proteome of single cells.17,18 Allbritton and colleagues have devised methods to map enzyme activation in single cells.19 A variety of other technologies utilizing CE for the analysis of cells are under development.20-22 The major strengths of these methods are their potential to measure large numbers of cellular attributes simultaneously and to quantify many cellular parameters not currently measurable at the single-cell level. However, since the above methods destroy the integrity of the cell, the ability to measure cellular attributes over time is sacrificed as is all spatial information. A number of investigators have developed strategies to sample cellular subcompartments for analysis by CE. Ewing and colleagues inserted the etched tip of a capillary into the large neurons of Planorbus corneus and electrophoresed a portion of the cytoplasm into a capillary.8,9 In a related strategy, Allbritton and cohorts used a vacuum to aspirate cytoplasm from small regions of the giant oocytes of Xenopus laevis.23 Sweedler and colleagues loaded cells onto plates containing a MALDI matrix and then used a focused laser to selectively vaporize portions of cells for analysis by mass spectroscopy.24 While these methods are suitable for the measurement of many cellular parameters, they are not appropriate for others. The techniques are either limited to giant cells (>1 nL) much greater in size then the typical mammalian cell (e1 pL) or involve significant perturbations to the cell prior to sampling. Even when these perturbations occur only seconds before sampling, substantial changes in the physiology of the cell can occur, for example, changes in ion concentrations or enzyme activation. Additional measurement strategies with spatial control are needed for CE to reach its full potential as a tool to analyze cellular contents. The laser micropipet system (LMS) is a CE-based method that uses a laser-generated shock wave to lyse small mammalian cells and terminate cellular reactions on very rapid time scales ( 5). To demonstrate that the contents of a process could be collected, separated, and detected, cells were loaded with carboxyfluorescein and Oregon green carboxylic acid and the lumen of a capillary was positioned above a neuronal process (Figure 4A). The process was collected into the capillary with the LMS and electrophoresis initiated (Figure 4B). The capillary was then manually moved away from the cell during electrophoresis (Figure (39) Ghosh, A.; Greenberg, M. E. Science 1995, 268, 239-47. (40) Haugland, R. P., Handbook of Fluorescent Probes and Research Chemicals. 6th ed.; Molecular Probes: Eugene, OR, 1996.
Figure 3. Electropherograms from a single and multiple cells containing carboxyfluorescein and Oregon green carboxylic acid. (A) Electropherogram of Oregon green carboxylic acid (6-isomer) and 5/6-carboxyfluorescein (mixed isomers) standards (10 nL of 5 and 50 nM, respectively) in a buffer. The mixed isomers of carboxyfluorescein migrated as a bifurcated peak (2) and the Oregon green carboxylic acid (6-isomer) as a single peak (1). (B-D) Electrophoretic traces from 1 (B), 4 (C), and 10 (D) cells loaded with Oregon green carboxylic acid and 5/6-carboxyfluorescein. The cells were incubated with Oregon green 488 carboxylic acid diacetate (6-isomer) and carboxyfluorescein diacetate (5- and 6-isomers), washed, and then lysed and loaded into the capillary using the LMS. The energy of the single laser pulse utilized to lyse the cells was 40 (B), 70 (C), and 80 (D) µJ. Electrophoresis was performed in a 75-µm-i.d. capillary with a potential of 9 kV.
Analytical Chemistry, Vol. 73, No. 19, October 1, 2001
4629
Figure 4. Collection of the contents of a neuronal process. An NGFdifferentiated PC12 cell grown on a cover slip was placed on the stage of a microscope and a process of the cell was centered beneath the lumen of a fused-silica capillary (30-µm internal diameter). Shown are transmitted light images. The lumen of the capillary was placed 25 µm above the neuronal tip and consequently is out of focus in the image. Panel A is the video frame coincident with the delivery of a single laser pulse (2 µJ) localized to the very end of the process. The 5-ns pulse occurred randomly during the 33-ms frame. Panel B is the next video frame, i.e. 33 ms later in time. In panel C, the capillary was removed so that the remaining portion of the neuronal process is seen more clearly.
4C). The neuronal cell and remaining neurite remained intact and attached to the cover slip. Figure 5 shows a representative electropherogram of the contents of a PC12 growth cone. The migration times of the Oregon green carboxylic acid and mixed isomers of carboxyfluorescein were similar to that of the standards. The ability to selectively and rapidly collect a portion of a cell for analysis for CE should be of great utility in studying the spatial control of signaling events in cells. This is particularly true in neurobiology where the biochemistry of the large cell body frequently overshadows that of the much smaller axonal and dendritic tips. Subcellular Measurements by Probe Localization. Many chemical processes are localized to discrete cellular compartments within the main cell body, for example, the nucleus or endoplasmic reticulum. This enables a cell to activate a signaling process in one subregion but not in another. Alternatively, constitutively active enzymes can be sequestered in compartments away from their substrates and then enzyme and substrate mixed in a 4630 Analytical Chemistry, Vol. 73, No. 19, October 1, 2001
Figure 5. Electropherograms of a single neuronal process containing the fluorophores carboxyfluorescein and Oregon green carboxylic acid. (A) Electropherogram of Oregon green 488 carboxylic acid (6isomer) and 5/6-carboxyfluorescein standards (∼100 pL of 10 and 50 nM, respectively) in buffer. The peaks are labeled as in Figure 3. (B) An NGF-differentiated PC12 cell was loaded with Oregon green carboxylic acid and 5/6-carboxyfluorescein by incubation with the diacetate forms. A process of the cell was centered beneath the lumen of a fused-silica capillary (30-µm internal diameter), and the contents of the process were selectively collected as described in Figure 4. To separate the contents of the cellular process, 21 kV was applied across the capillary (∼36 µA).
regulated manner as needed by the cell. One of the means by which this site-directed control of chemical processes is made possible is the presence of intrinsic signals in biomolecules that direct the molecules to specific regions of the cell.28,41 Prominent among these signaling sequences are nuclear localization sequences (NLS), short amino acid sequences that when attached to another molecule cause the fusion molecule to be selectively localized in the nucleus.28 To demonstrate the feasibility of detecting a localized probe by CE, two NLS-linked molecules were synthesized. Fluorescein alone or F-PKC was attached to an NLS (F-NLS and F-PKC-NLS, respectively) and microinjected into a PC12 cell. Figure 6A and B show the transmitted light and fluorescent image, respectively, of a cell loaded with F-PKC-NLS. The probe is clearly concentrated in the nucleus. A typical CE trace of a cell microinjected with F-NLS is shown in Figure 6D. A single peak with the same migration time as the F-NLS standard (Figure 6C) is present on the electropherogram. Using localized substrates such as F-PKC-NLS, it should be possible to measure the activation of enzymes in discrete cellular compartments. A large number of signaling sequences have been described,28,41 all of which can be exploited to selectively localize desired probes to specific subcellular locations. CONCLUSIONS We have demonstrated that the LMS terminates cellular reactions with sufficient speed that artifactual substrate metabo(41) Rizzuto, R.; Brini, M.; Pozzan, T. Methods Cell Biol. 1994, 40, 339-58.
Figure 6. Nuclear localization of molecules by attachment of an NLS. Transmitted light (A) and fluorescent (B) images of a PC12 cell microinjected with F-NLS-PKC. The nuclear region (marked by adjacent #) was brightly fluorescent, whereas no fluorescence was detectable in the adjacent cytoplasm (marked by adjacent *). (C) Electropherogram of a standard solution of F-NLS (∼0.6 nL of 5 nM). (D) Electrophoretic trace from a PC12 cell microinjected with F-NLS. The cell was lysed with the LMS (40-µJ laser pulse). Electrophoresis was performed in an uncoated capillary (50-µm i.d.,12 kV).
lism does not occur in response to cell lysis. Following lysis of a cell with the LMS, the substrates for PKC and CamKII remained unphosphorylated. These two enzymes are expected to be the initial enzymes activated in response to cell lysis since they are involved in repair of the plasma membrane and can be rapidly activated by an influx of extracellular calcium through a breach in the membrane.42,43 We also demonstrated that phosphorylation of substrates and potentially proteolytic breakdown occur with slower methods of cell lysis making these methods unsuitable for some types of cellular analysis. Spatial localization of proteins particularly enzymes has been identified as a mechanism to regulate the flow of information (42) Bi, G.; Alderton, J. M.; Steinhardt, R. A. J. Cell Biol. 1995, 131, 1747-58. (43) Steinhardt, R. A.; Bi, G.; Alderton, J. M. Science 1994, 263, 390-3.
through signaling cascades in cells.3,39 For this reason, the ability to perform spatially localized measurements in cells is becoming increasingly important in cell biology. We have demonstrated that the LMS can be utilized to perform spatially localized measurements on cells. Two paradigms are possible. In the first, a selected region of cytoplasm is removed or detached from the cell and loaded into a capillary. The extent of the region collected is controlled by the amount of energy in the laser pulse generating the shock wave. A weakness of this method is that the shock wave must have access to the cellular subregion without interacting with nontargeted regions of the cell (or these will also be lysed). Thus, it may be limited to cellular protrusions and processes. An advantage of the method, however, is that sufficiently large depositions of energy by the laser create shock waves capable of simultaneously lysing many cells enabling population measurements to be performed. The second method of spatial localization does not rely on titration of the laser pulse energy but rather on the actual localization of intracellular probes with signaling sequences. This method has the advantage that virtually any cellular subcompartment can be interrogated as long as a localization sequence is available. Furthermore, measurements will be specific to the targeted compartment even when the entire cell is lysed and loaded into the capillary. These adaptations of CEbased methods to provide spatial information about the biochemistry of cells will enrich the biologists’ and chemists’ tool chest. The ultimate goal is to reveal the intricacies of cellular signaling. ACKNOWLEDGMENT The authors thank Tatiana Krasieva for helpful discussions and assistance, and the Laser Microbeam and Medical Program, an NIH-funded user facility, for use of equipment and facilities. This work was supported by the NIH (RO1NS/MH39310, R24RR/ CA14892, and KO8AI01513). Received for review May 7, 2001. Accepted July 13, 2001. AC0105235
Analytical Chemistry, Vol. 73, No. 19, October 1, 2001
4631
