Experimenting in picoliter microvials

The use of nanotechnology to produce nanoliter vials for mass spectrometry led to the creation of these even smaller vials for investigating single cells.

Rose Ann Clark
Andrew G. Ewing

Researchers are discovering that combining several disciplines allows their work to move in unexpected directions and extend to new levels; cross-disciplinary research is no longer a rarity in the scientific community. Although the idea of combining chemistry, nanotechnology, and biology might seem unusual, chemists and biochemists willing to move from normal-sized beakers to beakers in the picoliter realm are finding this technology extremely beneficial. Thus, we are seeing the birth of nanoliter, picoliter, and even femtoliter beakers!

Our interest in nanotechnology stems from our current research in neurochemistry. We are analytical chemists investigating single cells to better understand the chemical communication occurring between cells. Nerve and hormonal cells communicate by releasing chemicals that the neighboring cells use as cues to continue or discontinue specific functions (1). Previously, research in this area focused on studying the cells electrochemically in petri dishes with volumes of 3-4mL (2-4). However, the use of nanotechnology to produce nanoliter vials for mass spectrometry sparked our interest in the possible use of even smaller picoliter vials for investigating single cells (5).

By using picoliter beakers, we can work on the cellular or subcellular level without the dilution observed in large volumes. Most of the work we present here focuses on developing picoliter microstructures and further characterization of the electrochemical properties that could affect the results of our research in these tiny beakers. Combining this nanotechnology with electrochemical detection for the analysis of single cells will provide substantial improvements in small-sample handling. Thus, opportunities will be created immediately for biological and chemical analyses, previously unexplored in the solution phase.

Nanotechnology in analytical chemistry
Achieving small-volume sample-handling techniques is of interest in many areas of analytical chemistry. Sample handling limits were pushed to smaller volumes for mass spectrometry; electrochemistry; separation techniques, such as liquid chromatography and capillary electrophoresis (5); and other areas. Vials capable of handling nanoliter volumes were developed for analysis of single cells using microcolumn liquid chromatography and for DNA sample handling in a gel-based separation system (6-8). In all of these applications, the sample size was decreased substantially by the microvials; however, only a small fraction of the total volume is actually analyzed. In efforts to further minimize the sample size to picoliter volumes, ultra-small vials were fabricated photolithographically on silicon wafers for capillary electrophoresis applications and for matrix-assisted laser desorption-ionization mass spectrometry to isolate analytes, such as Bradykinin, cytochrome c, and neurotransmitters to picoliter volumes (9, 10). Another small-volume technique of note uses picoliter-to-femtoliter aqueous drops isolated in organic media where a microscopic diffusional titration can be performed (11-13).

One analytical method that is emerging as an important technique for selective and sensitive detection in ultra-small environments is electrochemistry. Electrodes can be photolithographically fabricated on array structures and used to investigate cellular systems. Bioelectrochemical interfaces were constructed by culturing neurons directly on top of microelectrode arrays (14). Other groups are working on interfacing microelectrodes with biological systems. Connely and co-authors have used nanotechnology to perform cell-electrode interfacing to produce in vitro biosensors (15). Another electrode array system developed for biological environments combines a microscale electrochemical device with small-scale separations-based methods. Nanofabricated electrode arrays were placed at the end of a rectangular channel used for separation based on electrophoresis (16).

Extremely small, individual microelectrodes (1-5 µm) can be made; this has also facilitated the growth in using electrochemistry in the past decade. Cellular investigations in vitro and in vivo requiring detection of minute amounts of material are being conducted with carbon fiber microelectrodes (2-4, 17-19). However, analyzing a sample of a single cell is difficult because of the small volume of material released on stimulation of a cell and the femtomole-to-zeptomole levels of analyte present. To realize the full potential of electrochemical techniques, further developments in small-volume sample handling are needed that allow measurements without diluting the contents of the cell; this can only be done by using containers with extremely small volumes.

Fabrication of microvials
As we have noted, microvials were made for several applications in a variety of sizes. They were constructed in silicon wafers in 1-nL to 9-pL sizes solving small-volume, sample-handling, and dilution problems for injection into separation systems (8, 9). These microvials are effective for applications where it is not necessary to see the analyte. For work involving cells, however, the silicon microvials are not appropriate. To extend the use of the small-volume containers, we have fabricated a polystyrene microvial (20). These vials are transparent, which is advantageous for use with single cells because transmission-optical microscopy can be used to view the cell.

Microvial templates were fabricated at Cornell University's Nanofabrication Facility using our photolithographic and wet chemical-etching techniques (see sidebar, Creating the microvial, below). The five vial sizes (from 310 down to 0.4 pL) are easily manipulated by changing the photomask or controlling the KOH-etch depth. In addition, these transparent polystyrene microvials are easily viewed with transmission-optical microscopy, required for manipulation of extremely small volumes and single cells. The microvials are characterized structurally using standard microscopy and scanning electron microscopy (SEM) to image the microvials (Figure 1).

TO SIDEBAR: Creating the microvial

Figure 1. It is possible to fabricate microvials...

Microvial volumes were estimated from the measured dimensions (Table 1). Inner and outer diameters were measured with optical microscopy and SEM. Depths were determined using a surface profiler. Silicon microvials are perfectly symmetrical, inverted, square pyramids; the polystyrene microvials show some nonuniformities (Figure 1) (5, 8, 9). These nonuniformities make the size calculation of microvial volumes more difficult.

Table 1. Physical characteristics...

Microvial sizes reported in Table 1 are consistent over a single silicon wafer template. However, vial sizes can vary from wafer to wafer because the depth and, therefore, the outer diameter and volume are controlled by the KOH-etch rate. In addition, some rounding of the edges of the microvial templates is observed. The rounding of the outer edges occurs on the silicon template during the KOH wet chemical-etching procedure and is more pronounced for smaller structures. It was shown that the rounding can be reduced by creating structures on the mask to protect the corners during the etch procedure (21). Similar structures were used on the photomask, implemented for creation of the vials depicted in Figure 1; however, complete compensation has not yet been achieved. Very little rounding is observed when the larger, 310-pL vials (not shown) are produced; however, the protective structures may not have been completely etched away, thus increasing the size of the vial (Table 1). The rounding is more pronounced for 16-pL vials or smaller because of the overall small dimensions at these incredibly small volumes. It is still possible, even given these deviations from ideal, to calculate the volume of these microvials using square pyramids for the larger vials (310 and 85 pL) and conical shapes for the smaller vials (15, 10, 1, and 0.4 pL). Surface roughness contributions were not considered in these calculations; however, calibration of the volume by injection was performed for comparison.

Microvial volumes can also be determined using a microinjector calibrated so the delivery rate is known. Solutions containing 58% glycerol are used during this extended calibration process to ensure minimal evaporation. The vial sizes reported are determined by timing the filling process and calculating volume from a calibration curve. The range of microvial volumes is shown with the standard deviation for six trials shown in Table 1. The values for the smaller microvials compare well with those predicted from calculations based on the actual microvial dimensions. Because the calibrated volumes are close to the true volume, we will now only refer to the calibrated volumes. The 1- and 0.4-pL vial volumes were not determined because it is difficult to calibrate the microinjector used in these experiments for such small volumes.

Research in picoliter volumes
Although small polystyrene vial sizes are well-suited for investigations requiring extremely low sample volumes, rapid evaporation of liquid samples in the picoliter regime is a problem. The rate of evaporation is proportional to the surface area/volume ratio of the sample, which is extremely large in a picoliter vial. Under normal laboratory conditions, an aqueous sample evaporates in seconds from a 16-pL vial. Various methods were used to prevent evaporation in microvials: coating the vial with a membrane lid, saturating the headspace with water, and adding glycerol to lower the vapor pressure (5, 9, 22). In our work, we use glycerol, increased humidity of the headspace, and covering the microvial with a layer of mineral oil in experiments to minimize evaporation. Using these methods, we did not see evaporation during the 1-5 min it took to do the electrochemical experiments.

All of our electrochemical experiments were performed using the two-electrode mode normally used for microelectrodes (20). The working electrode was made from a 5-µm carbon fiber, and the reference is a miniature Ag/AgCl with a 1-µm tip diameter (23). Experiments are conducted on the stage of an inverted microscope equipped with three micromanipulators for positioning the electrodes and an injector (10 to 2 µm). High-precision micromanipulators are necessary to move the electrodes, considering the micrometer dimensions of the vials being used. Experiments also need to be conducted on a vibration-free table to eliminate interferences. Considering the dimensions for the 10-pL microvial (Table 1) and the size of the working and reference electrodes, it is not surprising that micromanipulators and vibration suppression are necessary.

When performing experiments in picoliter vials, we initially fill the vials with solution by pressure-injection with a microinjector. Shown in Figure 2 are 81-pL (top) and 16-pL (bottom) vials before and after a 35% glycerol-dye solution is injected. The injector (shown in the upper right) is manipulated into the vial and then retracted after delivering solution so it does not interfere with data collection. The filling process is easy to follow visually because the angled walls of the vial initially appear dark and become transparent as liquid is added. Once a microvial is filled, the miniature Ag/AgCl reference and the working carbon fiber electrodes are manipulated into it for subsequent electrochemistry (Figure 3).

Figure 2. It is possible to place samples...

Figure 3. The experimental setup...

We demonstrated the basic utility of microvials for electrochemical analysis by performing cyclic voltammetric analysis on an electrochemical standard, ferrocene carboxylic acid (20, 24). Voltammograms were collected in several microvial sizes. A representative cyclic voltammogram collected in a 16-pL microvial and one collected in bulk solution at a scan rate of 1 V/s are shown in Figure 4. The cyclic voltammograms show oxidation onset of the ferrocene carboxylic acid at 0.3 V and reach maximum oxidation quickly (notice the current levels off). Once the potential reaches 0.8 V, it is reversed, and ferrocene continues to be oxidized until scanning past the formal potential, and the current returns to zero. The cyclic voltammograms in the 16-pL vial at scan rates of 1 V/s display the normal sigmoidal response expected for steady-state diffusion at a microelectrode (23). The half-wave potential obtained in the microvials, 0.33 ± 0.01 V (n = 20), is identical to that for ferrocene carboxylic acid in bulk solution. The current values for 1-mM ferrocene carboxylic acid in the vials are very close to those of the bulk solution.

Figure 4. Cyclic voltammetry results in microvials...

To push the limits for electrochemical measurements to the smallest volumes, experiments were conducted in 1-pL vials with even smaller carbon fiber (1-µm tip diameter) electrodes (25). For these experiments, measurements must be carried out rapidly (~30 s) to avoid evaporation problems. At the appropriate scan rates, sigmoidal voltammograms are again observed for 1-mM ferrocene carboxylic acid solutions in only 1 pL (20).

Additional studies extended the electrochemistry investigations to a much wider range of parameters (24). Variations in the cyclic voltammetric response are observed at slower scan rates where the diffusion layer is interrupted by the walls of the microvials. The microvials become essentially 3-D thin-layer cells, and all of the material is oxidized or reduced depending on the direction of the cyclic voltammetric scan. The full details of the voltammetric response over a wide range of scan rates are beyond the scope of this paper and can be found elsewhere (24). Electrochemical measurements obtained for ferrocene carboxylic acid as a model system provided critical information about the electrochemical response in these small volumes. With this information in hand, we are proceeding with the investigation of biological samples.

Isolating cells
Successful development of voltammetric measurements in picoliter microvials gives us a means to investigate single-cell systems and reactions in extremely small and restricted volumes. The goal of these experiments is to place a single cell in a small (7- or 16-pL) microvial where the extracellular environment will be restricted, and released materials can be monitored with minimum dilution. To accomplish this goal, we have developed two methods to transfer single cells into the microvial and isolate them. One method involves using a microinjector to transfer of a cell into a microvial, and the second method uses culturing procedures to isolate a single cell. Photomicrographs showing cells transferred by microinjection are displayed in Figure 5. Cells were isolated in 81- and 16-pL vials. The cell in the 81-pL vial still has a substantial volume around it; however, the cell in the 16-pL vial is very close to the vial dimensions. In this case, the cell diameter is ~15 µm, and the bottom of the vial is 20 µm.

Figure 5. Single bovine adrenal cells...

As we mentioned earlier, we used three different methods to prevent evaporation. Using the cell system, the evaporation is controlled by placing a layer of mineral oil above the microvial to prevent the buffer solution from contacting air. This is critically important because it should allow us to carry out measurements on systems of viable cells. Glycerol, on the other hand, permeates the cell membrane and causes the membrane to weaken, which makes transfer difficult.

We were able to transfer single cells and groups of cells and are beginning to look at their electrochemistry in the picoliter volumes. In preliminary experiments, we have placed groups of cells (6-10) into 390-pL vials and carried out voltammetry in the microvial solution surrounding the cells. Because the bovine adrenal cells contain large amounts of norepinephrine and epinephrine (easily oxidized catecholamines), we expected to obtain voltammetry resembling catecholamines in standard solution. In fact, this is the case (Figure 6). The measured catecholamine corresponds to ~60 µM, calculated from the voltammogram, and appears to represent a background amount of these substances released from these cells. Future work will include experiments monitoring stimulated release, but the data shown here demonstrate that these experiments are very feasible. Thus, this technology should have an impact on microanalysis for studying single-cell physiology.

Figure 6. Cyclic voltammograms...

Commercial potential
In addition to being a valuable research tool, microvial technology should have broad applicability in other areas, including sensor technology, and single-cell and DNA analyses. Pantano and Walt are developing optical sensors with nanovials in the ends of fiber optics, and these will be useful for immuno- and DNA sensors (26). Anywhere that minute samples are involved or huge arrays of samples are needed, the microvial technology will be useful.

The analysis of single cells requires incredibly small samples. It is easy to imagine a system that uses microvial technology to create an array of analysis chambers to electrochemically or optically screen large populations of single cells for genotype or phenotype. The same array concept will likely be used someday to carry out polymerase chain reaction chemistry on many DNA samples followed by sequencing or genotyping via one of many methods under development as we write. The nanotechnology leading to experiments in picoliter and even femtoliter volumes indeed promises to make the future an exciting one.

ACKNOWLEDGMENTS

REFERENCES

SEE OTHER HOT ARTICLE: The art of creating a flexible R&D organization