[4 Microinjection as a Tool to Explore Small GTPase Function

By Brian Storrie


Microinjection overcomes the plasma membrane barrier to the introduction of charged or large nonlipid soluble molecules into cells by the direct insertion of a hollow capillary micropipette into the cell. With the application of pressure, aqueous solution is then directly transferred into either the cytosol or the nucleus. I give specific examples of the application of this approach to the functional study of small GTPases of the Sar1, ARF, and rab family in membrane trafficking between the Golgi apparatus and endoplasmic reticulum (ER). The principles illustrated by these examples should be generally applicable to other small GTPases. Detailed protocols for capillary microinjection using semiautomated equipment are given.


Microinjection is a tool to overcome the plasma membrane permeability barrier to the introduction of charged molecules, polypeptides, or DNA plasmids into cells. In essence, microinjection treats the cell as the test tube and uses a microinjection capillary needle as the pipette to add small volumes of solution to the cytoplasm or nucleus. The approach has major advantages: (i) the technique is highly synchronous with a few hundred cells being injected over 10-20 minutes, (ii) intracellular environment and cell morphology is preserved, (iii) the reaction vessel is small, that is, the size of a cell, and correspondingly reagent dilution is confined to the volume of the cell, and (iv) reagent consumption is limited. No more than 0.5 to 1.0 ^l of solution is required to load the microcapillary. Injection volumes are on the order of 1 pl, ~12 liters or ^5-10% of cell volume. Both flat substratum-attached and rounded-to-free floating cells can be injected. The technique is popular with an early March 2005 PubMed search against ''microinjection'' yielding 15,762 hits and perhaps underappreciated in the small GTPase field with the corresponding search against ''microinjection small GTPase'' giving 134 hits.

Microinjection is a light microscope based technique. The microinjection capillary and its positioning relative to cells are tracked by phase contrast. Moreover, phase contrast is used to identify cell nucleus versus cytoplasm. Microinjection requires in addition to an inverted phase contrast microscope at least two committed pieces of equipment (Fig. 1). These are a micromanipulator to position the injection capillary and a pressure regulator to increase pressure to the capillary selectively as it

Microinjection Brain Tissu

Fig. 1. Year 2004 state-of-the-art setup for microinjection of substratum attached cells. Left to right: Box of Geloader® pipette tips for backloading capillary micropipettes, glass Petri dish bottom with modeling clay bridge for storing newly fabricated capillary micropipettes, Eppendorf FemtoJet pressure regulator (arrowhead), research grade inverted microscope, Eppendorf XYZ motor drive attached to right side of microscope stage (arrow points to capillary micropipette in holder), Eppendorf InjectMan® joystick controller for micromanipulation/injection (double arrows), and black and white video monitor showing substratum attached HeLa cells.

Fig. 1. Year 2004 state-of-the-art setup for microinjection of substratum attached cells. Left to right: Box of Geloader® pipette tips for backloading capillary micropipettes, glass Petri dish bottom with modeling clay bridge for storing newly fabricated capillary micropipettes, Eppendorf FemtoJet pressure regulator (arrowhead), research grade inverted microscope, Eppendorf XYZ motor drive attached to right side of microscope stage (arrow points to capillary micropipette in holder), Eppendorf InjectMan® joystick controller for micromanipulation/injection (double arrows), and black and white video monitor showing substratum attached HeLa cells.

enters a cell. The increased pressure is what expels fluid from the capillary within cells. The micromanipulator is mounted onto the stage of the inverted microscope. Ideally the micromanipulator and pressure regulator work together under microprocessor control. In fact, the microprocessor and the possibility of programmed injection is what places microinjection as a practical technique in the hands of many investigators. In programmed injection, the microprocessor controls the actual dipping of the capillary into the cell and coordinates rapid changes in capillary pressure with capillary position. Hence, the skill level required is reduced and the success rate of injections greatly increased. Microinjection was introduced as a technique in the 1940s (e.g., de Fonbrune, 1949). Microprocessor control was introduced in the 1980s through instrument development efforts at the European Molecular Biology Laboratory (Ansorge, 1982; Pepperkok et al., 1988) and commercialized by Carl Zeiss and Eppendorf. The microprocessor controlled Eppendorf injection system used by my laboratory may be mounted on any of a number of different inverted microscopes. Micro-capillaries suitable for injection into either the cytoplasm or nucleus may be purchased from Eppendorf. Microcapillaries may also be pulled to individual requirements using commercially available equipment such as the Flaming-Brown puller from Sutter Instruments.

The end purpose of microinjection is to produce phenotype. I consider in detail here the situation of substratum attached interphase cells and experiments designed to elucidate small GTPase function in secretion. The principles developed are applicable to small GTPases in other metabolic pathways. I will first illustrate approaches taking examples primarily from the work of my laboratory, second use the strengths and weaknesses of these examples to propose a set of good practice standards for microinjection experiments, and finally present a set of detailed protocols for injection experiments. In general, microinjection experiments involve the introduction of either increased amounts of a normal or mutant protein or a potential inhibitor into cells. For the consideration of alternate approaches such as vaccinia virus for rapid, synchronous, high-level transient transfections or siRNA for slow, but effective knockdowns of wild-type protein expression, the reader is directed to other articles within the Methods in Enzymology series.

Example Uses

GTPases act as conformation-dependent molecular switches. In the GTP-bound state, the switch is on. In the GDP-bound state, the switch is off. When switched on, small GTPase act to recruit effectors. GTP7S as a non-hydrolyzable analogue of GTP is an inhibitor of all small GTPases. Activity of individual GTPases may be selectively altered by amino acid substitutions. For example, the activity state of small GTPases may be altered by amino acid substitutions in their primary sequence that affect GTP hydrolysis, guanine nucleotide exchange, nucleotide binding, and potentially specific effector interactions. In addition to altering the amount or activity state of small GTPases, alterations in the amount or activity state of effectors is also an attractive avenue for probing function. Dominant negative phenotypes may be produced by amino acid substitutions that either activate the GTPase switch, for example, mutations that slow GTP hydrolysis, so-called GTP-restricted mutations, or those that inactivate the GTPase switch, for example, mutations that stabilize the association of GDP with the protein, so-called GDP-restricted mutations. Increased amounts of GTP-restricted mutations act as dominant negative inhibitors because they slow the reutilization of effectors. GDP-restricted mutant proteins act not by competing for effectors, but by competing with wildtype protein for binding to guanine nucleotide exchange factors (GEFs). In fact, GDP-restricted small GTPases in general bind more tightly to GEFs than wild type proteins and hence form "dead-end" complexes with the GEF (Feig, 1999). Of course, how specific the effects of a GDP-restricted rab6a, for example, is dependent on how specific the GEF is.

With this background in mind, I now consider a number of specific examples relating to membrane trafficking between the ER and Golgi apparatus. These examples concern Sar1a, a small GTPase required for COPII coat protein recruitment to ER membranes; ARF1, a small GTPase required for COPI coat protein and clathrin recruitment to Golgi apparatus membranes; rab6a/a', small GTPases involved in membrane trafficking at the trans Golgi apparatus/trans Golgi network (TGN); and rab33b, a small GTPase of the medial Golgi apparatus. In all cases, phenotype is assayed on a single cell basis by assessing effects on protein distribution. Lastly antibody microinjection to inhibit cell intoxication by Pseudomonas exotoxin or Shiga-like toxin will be considered as an example of outcome assessment by incorporation of radiolabeled amino acid.

1. GTP7S as a general inhibitor of small GTPase dependent processes - We used microinjected GTP7S to probe the dependence of Golgi apparatus scattering in response to microtubule depolymerizaton to small GTPases (Yang and Storrie, 1998). Vero cells were microinjected with GTP7S from a 500 ^M stock solution. As the microinjection volume is about 5-10% of that of the cell, the intracellular concentrations were between 25 and 50 ^M. These concentrations had no effect on Golgi scattering in a 4.5 h endpoint assay. Higher concentrations led to cell damage. As a positive control, we found that these concentrations of GTP7S were sufficient to disperse the juxtanuclear, Golgi apparatus associated concentration of the COPI coat protein complex. Protein distributions were assessed by immunofluorescence.

Do negative results in such experiments truly indicate that GTPases play no role in a process? The answer may well be no in any given situation. As Pepperkok et al. (1998) showed, GTP7S microinjection fails to block Sar1a function and hence to affect ER exit of vesicular stomatitis virus G protein. Contrary to the expectations of our GTP7S microinjection experiments, we later showed that overexpression of GDP-rab6a delays but does not block microtubule-depolymerization-induced Golgi scattering (Jiang and Storrie, 2005), that is, a small GTPase does have some role in Golgi scattering. The bottom line is that such experiments require careful analysis of protein distribution over time, that is, kinetic analysis of phenotype, and the acknowledgement that some GTPases may be comparatively insensitive to GTP7S.

2. Overexpression of wild-type protein - The levels of small GTPases and effectors must be exquisitely balanced to produce normal resident protein distributions within the secretory pathway. Nevertheless, the finding that overexpression of wild type rab6a (Martinez et al., 1997, transfection) or rab33b (Valsdottir et al., 2001, microinjection) induced the redistribution of Golgi glycosyltransferases to the ER was remarkable. In such experiments, the advantage of microinjection versus transfection was the high degree of synchrony that comes from the narrow time window over which the plasmids are injected into cells. Plasmid injection into the nucleus is often used because it is much simpler to purify a plasmid than to purify from an animal cell source a protein expressed in small amounts or to purify the same protein in an active, properly modified protein from bacteria. Small GTPases are post-translationally modified by fatty acid addition.

3. Overexpression of mutant protein - Considering that either GTP- or GDP-restricted rab6/rab33b, Sar1, or Arf1 all can produce a dominant negative phenotype when introduced into cells, does it really matter which allele is used? Whether these experiments are done as plasmid expression or purified protein microinjection, the answer is very much "Yes." GTP-restricted rab6a or rab33b both induce the redistribution of Golgi resident proteins to the ER while respective GDP-restricted isoform induces, if anything, a more compact, juxtanuclear Golgi apparatus (Girod et al., 1999; Jiang and Storrie, 2005; Martinez et al., 1997; Valsdottir, 2001; Young et al., 2005). Correspondingly, expression of GTP-restricted Arf1 results in the stabilization of COPI association with Golgi membranes while expression of GDP-restricted Arf1 results in the failure to recruit new rounds of COPI coat proteins to Golgi membranes and hence produces decoated Golgi membranes that are absorbed into the ER in a brefeldin A like phenotype (Dascher and Balch, 1994).

In the case of Sar1p, the phenotypic differences between alternate mutations are subtle with respect to Golgi apparatus organization and have led to controversy and confusion in the literature. When microin-jected into HeLa cells in the same plasmid background, both isoforms are effective in inhibiting the transport of VSV-G protein from the ER to the Golgi apparatus (Stroud et al., 2003). At moderate inhibitory protein levels, GTP-restricted Sar1 stabilizes the association of COPII coat proteins with ER membranes and ER exit sites. In some cell types, ER exit sites are juxtanuclearly clustered as is the Golgi apparatus. At high GTP-restricted Sar1 levels, this association is destabilized (Stroud et al., 2003). While on the other hand, GDP-restricted Sar1 prevents COPII coat protein recruitment to ER membranes at any inhibitory concentration. The net pheno-type is in either case a block in protein transport from the ER to the Golgi apparatus. However, in the case of the GTP-restricted mutant, the juxta-nuclear localization of Golgi matrix proteins appears more stable than that of recycling Golgi cisterna l enzyme s (Miles et al., 2001; Seemann et al., 2000; Ward et al., 2001). In the case of the GDP-restricted mutant, no such difference is apparent (Stroud et al., 2003; Ward et al., 2001). All Golgi proteins disperse in a manner consistent with protein cycling between the Golgi apparatus and ER.

In sum, the experimenter must know the small GTPase and the ins and outs of how phenotype might be affected.

4. Direct microinjection of mutant protein - In many ways, this is conceptually the most straightforward experiment. The purified protein is microinjected into the cytoplasm, often in the presence of cycloheximide to inhibit new proteins synthesis, and phenotype is assessed with no need to wait for plasmid-encoded protein expression. As one example, I cite our use of microinjected GTP-Sar1H79N protein to establish that accumulation of Golgi resident proteins in the ER in response to an ER exit block was due to protein recycling, not accumulation of newly synthesized proteins (Girod et al., 1999; Miles et al., 2001; Storrie et al., 1998). In these experiments, HeLa cells were microinjected with purified histidine (his)-tagged Sar1H79N protein at stock concentrations of ^0.75 mg/ml or final end cellular concentration of ^35-75 yg/ml in the presence of either cycloheximide or emetine as strong inhibitors of protein synthesis. (The his-tag is included in the sequence of the engineered protein because it simplifies the isolation of the recombinant protein from E. coli by Ni-affinity chromatography.) The kinetics of Golgi protein redistribution to the ER was then scored over a few hours. The comparative strengths and weaknesses of the approach are apparent from the outcome of the experiments. The onset of the ER exit block is quick and there is no need for protein synthesis. However, the actual level of Sar1 mutant protein is likely low in comparison to what can be achieved by plasmid expression. This point is important in the case of the GTP-restricted protein because it is only at high concentrations that the GTP-restricted protein destabilizes

COPII association with ER membranes (Stroud et al., 2003). Furthermore, it should be noted that in the case of Sar1 the GTP-restricted protein has been the microinjected protein of choice not because it is the right mutant protein but because it is easier to isolate than the GDP-restricted mutant in stable, high activity form. As a final buyer beware note, the reader is reminded that small GTPase are in general lipid modified and that these modifications typically are not produced when isolating the recombinant, his-tagged protein from E. coli. The experimenter hopes that the injected cell will do what E. coli did not.

In sum, direct mutant protein microinjection is an attractive choice for which the major hurdle is isolation of the mutant protein as an active, stable protein.

5. Direct microinjection of inhibitory antibodies or protein fragments -In the secretory pathway, small GTPases act to recruit effector proteins to membranes. One class of effectors is coat proteins such as COPI and COPII. COPI coat proteins are one example of ARF1 effectors. Micro-injected antibodies to B-COPI, one of seven different subunits of COPI, provide a specific tool to probe the role of COPI in a given process, particularly as ARF1 has been implicated in many processes. As with other purified proteins, the microinjection is into the cytoplasm. Antibodies directed against a B-COPI peptide containing the sequence EAGE react with an exposed region of COPI in vivo and inhibit protein transport from the ER to the plasma membrane (Pepperkok et al., 1993). Importantly they do this by stabilizing the association of COPI coat proteins with Golgi apparatus membranes. Monovalent Fab fragments of these divalent IgG antibodies are effective, indicating that inhibition is due directly to antibody binding to a specific site on B-COPI rather than to divalent antibody-induced crosslinking of COPI components. In our experiments, we used affinity purified EAGE Fab fragments at a stock concentration of ^1.4 mg/ml and found that these had no effect on recycling of resident Golgi glycosyltransferases to the ER (Girod et al., 1999). Importantly, microin-jected EAGE Fab fragments strongly inhibited the recycling of KDEL receptor from the cis Golgi apparatus to the ER. The negative result on resident glycosyltransferases recycling to the ER is only meaningful because of the successful positive control. We concluded from the Fab fragment experiments together with supporting data with GTP-restricted ARF1 expression that resident Golgi protein recycling to the ER is a COPI independent process.

6. Co- and sequential microinjection experiments - Coinjection or sequential microinjection experiments provide an excellent approach to probe for the role of multiple gene products in a process. Here I give one example. Over expresssion of rab6a and rab6a', two rab6 isoforms differing in only three amino acids, both cause redistribution of Golgi resident proteins to the ER (Jiang and Storrie, 2005; Young et al., 2005). If the two rab6 isoforms are redundant, then coexpression of the GDP-restricted form of one should inhibit the redistribution induced by the GTP-restricted form of the other isoform (Jiang and Storrie, 2005). This is exactly what is observed indicating that indeed the two are redundant in the pathway. A simple biochemical explanation for the cross-competition between rab6a and rab6a' is that the two use the same GEF or other machinery in their cycling between GTP- and GDP-bound states. Coinjection-coexpression experiments are technically fairly simple. Sequential microinjection experiments are more technically demanding. If each round of injection has a 40% success rate, that is, successful injection per capillary dip into cell, then the overall successful rate of two sequential injections is the product of each injection or 16%. In practice, I note that the success rate for cytoplasmic injections is higher than that of nuclear injections.

7. Single cell versus biochemical assays for phenotype - The outcome of most microinjection experiments is assayed on a single cell basis. For example, Cascade blue dextran might be used as a coinjection marker to identify microinjected cells and antibody staining might be used to determine the distribution of the protein of interest. Alternatively, an antibody might be used to identify the microinjected or expressed protein. Antibody identification of the injected cells has the major advantage of allowing for assessment of protein-expression levels through staining intensity.

With skill, biochemical assays are possible. For example, Girod et al. (1999) assayed for the ability of anti-EAGE or anti-KDEL receptor antibodies to inhibit cell intoxication by Pseudomonas exotoxin or Shiga-like toxin by assessing the scintillation counting of [35S] methionine incorporation by the injected cells. These experiments used cells cultured on cover-glass fragments and the coverslips were processed for trichloroacetic acid precipitation in situ. Such experiments require mock injection controls for injection-damaged cells act like dead cells with respect to amino acid incorporation.

Some Practical Comments and Suggested Good Practice Standards for Microinjection

Let me be straightforward: microinjection, even with present state-of-the-art semi-automatic injection equipment is a skill that does require a time investment. Based on my experience of having taught microinjection to ~15 people in a research laboratory setting over the last several years, almost anyone can do acceptable microinjection for research purposes with a week of learning. However, true skill, in being able to handle a range of samples, coping with pipette clogging, and being fast, takes more experience than that. These words are not meant to be discouraging. Rather they are meant to be realistic. To date, I have met only one person who did not have the patience or persistence to learn microinjection. The skill once learned is like riding a bicycle, something where proficiency is rapidly regained.

Microinjection can produce remarkably nice quantitative concentration and time responses (see, Jiang and Storrie, 2005; Stroud et al., 2003; Young et al., 2005 for example data). However, doing so requires treating the microinjection capillary (needle) with respect and being quick. In the real world even after centrifugation, samples are prone to aggregation within the narrow confines of the capillary, typical tip diameter 0.5 to 1.0 ^m. Getting the capillary once sample loaded quickly into the cell culture media and down to the level of the cells is important. Injection is done with continuous holding pressure to the capillary (Pc). Pc is set to prevent backflow into the capillary and hence high enough to give a slow continuous flow to the sample. One does not want drying at the tip by being slow to get the capillary into the media. Often there is a bubble at the tip after loading. This bubble should be cleared quickly after bringing the capillary down to cell level. Sufficient Pc, typically 50 hPa, gives a slow positive flow important to limit clogging. If the capillary tends to clog, reload sample and start with a new capillary or recentrifuge the sample and reload. If clogging is a persistent problem, increase capillary diameter by using less heat when pulling the capillary pipette. Make a better pipette for the sample. Do not arbitrarily enlarge the capillary, thus breaking the tip. A broken tip is irregular in shape and hence does not seal well as it enters the cell and is irreproducible as one does a concentration or time series.

Injections are done under phase contrast microscopy. Good cytoplas-mic injection produces a gentle wave of movement in the cytosol as the injection occurs. Good nuclear injection produces a slight wave of movement in the nucleus. The cell does not bleb during injection. The cell looks normal by phase contrast after the injection. Cell health is good both during, immediately after, and several hours later by phase contrast microscopy. Decrease injection pressures or capillary diameter if these things are not true.

Ideally, in all microinjection experiments, whether plasmid expression or direct protein injection, the injected cells can be identified by antibody staining against the expressed or injected polypeptide. This allows for quantification of expression levels or polypeptide delivery.

Ideally, the concentration of injected polypeptide or plasmid needed to produce phenotype is known from an actual titration experiment, and actual kinetics for the onset of phenotype are determined in preliminary experiments. Ideally, experimental design includes both positive and negative controls.

Finally, as cited in these examples, experience suggests that the experimenter must understand the consequences of using the GDP- versus GTP-restricted isoform as dominant negative mutant.

A Suggested Equipment Setup and Materials List

1. A limited personnel access room for injections. Injections are done into cells in open media tissue culture containers. Contamination is least if people are not continuously coming in and out of the room. There is no need for sterile room conditions or restriction of the room to a single purpose.

2. A stable research-grade inverted microscope such as the Zeiss Axiovert 200. An intermediate lens optovar is useful for fine-tuning microsope magnification to injection situation. The microscope may need to be mounted on a vibration table or not. Frequently this is required in modern forced draft air handling buildings to eliminate building transmitted vibration to the injection capillary.

3. Phase contrast optics: 10x and 32x long-working distance, phase 1 objectives.

4. Capillary pipette puller such as the Sutter Model 97 Flaming/ Brown micropipette puller.

5. Thin wall borosilicate glass capillaries for fabrication of micropipettes. We use a thin wall borosilicate glass capillary with filament from Sutter Instruments (catalog number, BF120-94-10, 1.20 mm outside diameter, 0.94 mm inside diameter, 10 cm overall length). We use Geloader2 tips from Eppendorf (catalog number 22 35 165-6, various distributors) to backload solutions into the fabricated micropipettes.

6. Integrated microprocessor controlled micromanipulator and pressure regulator system. We use the items as a system (InjectMan® NI 2 and FemtoJet®) from Eppendorf AG (Hamburg, Germany) that operates together under microprocessor control. This is the optimal situation. This system is designed for the injection of substratum-attached cells.

7. Video rate CCD camera with high quality monitor. It is much more pleasant to view the injection process on a monitor than through the microscope oculars. Video rate, 25 or 30 frames per second, is important to give a "real-time" sense of micropipette position. We have used at times a C-mounted Cohu 9010 black and white CCD camera paired with a Panasonic monitor or a Dage-MTI black and white CCD camera paired to a Sony monitor. The Dage-MTI camera is the flexible camera but an exercise in overkill for the purpose.

8. CO2 independent culture media for maintaining the cells during the microinjection process. One source is In Vitrogen. Other suppliers are possible. Phenol red free media is best. Remember to prewarm the media to at least room temperature before use. Animal cells round in the cold.

9. Substratum attached cells. My experience is entirely with either substratum attached HeLa or Vero cells. The more spread the cell the better. Depending on purpose, the cells may be grown either on tissue plastic, glass coverglass onto which cells are seeded in a tissue culture dish, or cover glass bottomed tissue culture dishes (MatTek). Cells should ideally be 50-70% confluent for short-term, 6-10 h post injection experiments. Microinjection can be done with suspended cells. Successful suspended cell injection requires a two micropipette-system in which one pipette is used to hold the cell and another to inject the cell.

10. It is convenient to have a CO2 incubator close to the injection setup for culture incubation.

11. Purified plasmid DNA, polypeptides, or membrane impermeant solutions. We routinely use Qiagen Maxi- or Mini-prep prepared plasmids. For microinjection, we prepare plasmid inserts in pCMUIV, a plasmid designed for high-level expression in human cells (Nilsson et al., 1989). We directly resuspend purified DNA in molecular biology grade water. Cells do not react well to EDTA and Tris. Obviously, micoinjection into live cells mandates the avoidance of azide or other preservative in protein solutions.

Detailed Protocols

1. Pulling a capillary pipette—Prepulled and mounted capillary pipettes for either cytoplasmic or nuclear injection may be purchased commercially from Eppendorf. Many experimenters pull their own pipettes. A popular pipette puller is the Model 97, Flaming/Brown micropipette puller from Sutter Instrument Company, Novato, CA (Fig. 2A). Steps—The capillary (arrow, Fig. 2B) is inserted through a box heating filament (arrowhead, Fig. 2B) of 2.5 x 2.5 mm box size and 2.5 mm width in the puller, centered, and clamped into place by tightening the thumb wheel of each of the capillary carriers. The capillary is then placed under tension and heat applied to the box filament. A starting heat value is found by the ramp test. In the ramp test, the box filament is gradually heated to a value at which the heated glass just begins to flow. Tension is applied to the glass

Fig. 2. Flaming/Brown micropipette puller (Flaming/Brown). A. full frame view with lid open to show top interior of puller. B. Capillary carriers with inserted capillary (arrow). Arrowhead points to box filament for heating glass. C. Two newly fabricated capillary micropipettes (arrows) after heating the glass to the point of rupture under tension. The safety cover over the box filament has been removed to give better visual access in the photographs.

Fig. 2. Flaming/Brown micropipette puller (Flaming/Brown). A. full frame view with lid open to show top interior of puller. B. Capillary carriers with inserted capillary (arrow). Arrowhead points to box filament for heating glass. C. Two newly fabricated capillary micropipettes (arrows) after heating the glass to the point of rupture under tension. The safety cover over the box filament has been removed to give better visual access in the photographs.

during the ramp test. We then take the value from the ramp test, reduce it by 10-20 and try this first for production micropipettes. Production pipettes have a narrower opening at higher heating and a larger opening at lower heating. What works best is determined by effectiveness in injections into cells. The taper length of the microinjection pipette should be in the range of 5-7 mm. As shown in Fig. 2C, two essentially identical micropipettes are produced each heating cycle.

Newly fabricated pipettes are best used within days of being pulled. Other Sutter specific settings—Pull, 150; Velocity, 100; Time, 135; Pressure, 300 for thin-walled glass

2. Backloading a capillary micropipette—We use Geloader2 plastic pipette tips to introduce precentrifuged solutions for microinjection into capillary micropipettes. The solution is centrifuged for 15-20 min in the cold at full speed in a refrigerated microcentrifuge. We typically prepare final solutions in an end volume of 12 ^l. The final solution consists of diluted plasmid (common end stock concentrations of 10 to 100 ng/^l for pCMUIV inserts) or protein (common end stock concentrations of mg/ ml), any co-injection marker (e.g., Cascade blue dextran [Molecular Probes, Eugene, OR] at an end stock concentration of 3 mg/ml or so), and molecular biology grade water to volume. An example of backloading is shown in Fig. 3A. A left-handed person holds the capillary micropipette and the arrow points to the thin extended tip of the Geloader2. We load 1.5 to 2.5 ^l of solution into the capillary micropipette. This is sufficient for the injection of many thousands of cells. Air pressure applied to the capillary micropipette will force solution down to the end of the capillary. Obviously, the fabricated microcapillary has too wide a tip when its resistance is not sufficient to retain within the micropipette solution at a pressure of 5000 hPa, a typical pressure for clearing clogged particulates from the capillary micropipette.

3. Bringing the capillary micropipette down to cell level—The overall Eppendorf InJectMan manipulator system consists of two parts: a microscope stage-mounted XYZ motorized assembly and a joystick control box (see Fig. 1 for overview picture and Fig. 3B for closeup of stage-mounted assembly). The loaded capillary micropipette is mounted into the holder on the motorized microscope unit, the micropipette pressurized to the continuous or holding pressure (Pc), and the capillary brought down into the cell culture media as fast as possible. Using phase contrast illumination, a bright phase halo forms where the micropipette makes contact with the culture media (Fig. 3C). As shown in Fig. 3C for a right mounted XYZ assembly, the capillary micropipette tip will be to the left of the phase bright spot (arrow pointing towards tip, blurry capillary image, Fig. 3C). The phase spot is useful for first getting the capillary micropipette into the


Fig. 3. Steps in capillary micropipette manipulation. A. Backloading of capillary micropipette with solution (arrow points to Geloader® tip inserted into back opening of micropipette). B. Capillary micropipette in place over open cell culture media (arrow points to capillary micropipette). C. Low power appearance of capillary micropipette in culture media (arrow points towards slightly out of focus capillary tip and phase bright spot is where capillary enters the culture media). 10x objective, phase contrast. D. Appearance of capillary micropipette as it enters the cytoplasm of a HeLa cell. Cytoplasmic microinjections are near the nucleus as the cell is thickest in this area and hence the target depth is greatest.

Fig. 3. Steps in capillary micropipette manipulation. A. Backloading of capillary micropipette with solution (arrow points to Geloader® tip inserted into back opening of micropipette). B. Capillary micropipette in place over open cell culture media (arrow points to capillary micropipette). C. Low power appearance of capillary micropipette in culture media (arrow points towards slightly out of focus capillary tip and phase bright spot is where capillary enters the culture media). 10x objective, phase contrast. D. Appearance of capillary micropipette as it enters the cytoplasm of a HeLa cell. Cytoplasmic microinjections are near the nucleus as the cell is thickest in this area and hence the target depth is greatest.

media. After that, all attention should be on the actual capillary tip, not the spot. Under a 10x objective, bring the capillary tip down to almost in focus with the cells, shift to a 32x objective for final fine adjustments of the tip relative to cells. It is important to shift phase contrast objectives while the capillary tip is out of focus above the cells. If the tip is too near cell level, it may be broken from vibration as the 32x objective is clicked into place.

Most people break several capillary micropipettes in the course of learning the coarse micromanipulation to place the capillary near cell level. Experienced personnel do all coarse manipulation steps with the 32x objective rather than first with a 10x objective.

4. Injecting substratum attached cells - The Eppendorf system uses a microprocessor to dip the capillary micropipette into the cell, increase air pressure within the pipette to injection pressure (Pi), and then withdraw the capillary from the cell. The machine like the autostart system on a Formula 1 race car is better than a person. The operator sets or maintains several parameters. First the operator must set the Z-limit. Using the fine movement setting on the manipulator, the micropipette tip is either brought down to focus at the same height as the cell or actually inserted into the nucleus or cytoplasm (Fig. 3D) and then Z-limit keypad button pressed. The capillary micropipette is then brought up to a "cruising" height above the cells. The tip will be slightly out of focus with respect to the cells. It will be sufficiently high above the cells to avoid collisions as it is moved from cell to cell. The operator sets the injection time. We routinely use 0.3 sec. When the operator presses the button on top of the joy stick, the capillary micropipette tip is automatically moved down into the cells to the Z-limit. Injector pressure automatically increases to Pi, stays there for the injection time, and then the pipette is removed from the cell. All of this is done at a 45-degree pipette entry angle into and out of the cell. Ideally all is set now for repeated nuclear or cytoplasmic injections. Small wave-like movement of cytoplasm or nucleoplasm will be seen during the injection process.

Unfortunately, the surface of the culture dish or glass coverglass is not perfectly flat. This means that as the micropipette is moved from cell to cell the Z-limit position must be adjusted up or down by keypad. Moreover, cell height differs with phase of the cell cycle. Furthermore, in practice, capillary micropipettes clog (see following).

An experienced person can inject a few hundred cells in 10 to 20 min— from starting to load the capillary to bringing the capillary back up out of the culture media after having completed injecting cells. We typically start injecting at a pressure of 300 hPa. We then adjust Pi downwards depending on whether damage is being done to the cells. My experience is that it is better to kill a few cells initially and be sure that transfer is occurring than to be too careful. Being too careful often results in nondetectable transfer as the capillary is dipped into the cell. For a good flowing solution, a Pi of ^100 hPa produces successful transfer with about 50% of the downward capillary dips. Some solutions require a higher Pi than 300 hPa to produce good flow into the cell.

Clearing clogged capillary micropipettes. In the real world, capillary micropipettes clog during microinjection. The 5-7 mm glass taper of the micropipette tends to catalyze particulate formation in the injection solution. The faster the individual injections are done the less the problem. Antibodies at high concentration tend to be a problem. Sar1 mutants tend to be a problem. For a well-behaved solution, a single capillary micropipette is sufficient to inject a few hundred cells with no clogs. More commonly there may be a bubble from loading the capillary that needs to be cleared or an actual particulate clog. Particulate clogs appear as white spots in the taper of the tip. The first approach to either is increasing the pressure to cleaning pressure, 5000 hPa or more, briefly. This almost always suffices for bubbles and often for clogs. When doing this, the micropipette should be over a cell-free region and sufficiently above the substratum to allow for slight pipette movement under the high pressure. If high pressure does not suffice, high pressure combined with tapping on the pipette holder often will suffice. Again, the pipette needs to be somewhat high and over a cell-free region to allow for vibration. There is nothing gained in having the pipette collide with cells. Impaled cells can often be dislodged by the same approach. Recentrifuging the injection solution can help. Clogging is more of a problem at high plasmid or protein concentrations. Increasing capillary pipette diameter can help.

In the real world, I have sometimes used 3-4 capillary micropipettes to successfully inject cells for a single time point or single concentration.

Assaying for phenotype in injected cells. The microinjected cells are identified independently of phenotype by a fluorescent coinjection marker or antibody to expressed or injected protein. Phenotype assay depends on experimental design. Examples of different assays and experiment situations are given in earlier sections of this article. Protein expression levels may be assessed by antibody staining intensities using current technology CDD cameras and software (e.g., Jiang and Storrie, 2005; Young et al., 2005).


I express my sincere appreciation to Rainer Pepperkok (European Molecular Biology Laboratory, EMBL, Heidelberg) for teaching me microinjection on the now-discontinued Zeiss Automated Microinjection System in the early 1990s. I also express appreciation to Rainer Saffrich (EMBL) for help in setting up my first microinjection system in the United States. Work in the author's laboratory has been supported by grants from the National Science Foundation and the National Institutes of Health.


Ansorge, W. (1982). Improved system for capillary microinjection into living cells. Exp. Cell Res. 140, 31-37.

Dascher, C., and Balch, W. E. (1994). Dominant inhibitory mutants of ARF1 block endoplasmic reticulum to Golgi transport and trigger disassembly of the Golgi apparatus. J. Biol. Chem. 269, 1437-1448. de Fonbrune, P. (1949). "Technique de micromanipulation." Monographies de l'Institut

Pasteur. Masson et Cie, Paris. Feig, L. A. (1999). Tools of the trade: Use of dominant-inhibitory mutants of ras-family GTPases. Nat. Cell Biol. 1, E25-E27.

Girod, A., Storrie, B., Simpson, J. C., Johannes, L., Goud, B., Roberts, L. M., Lord, J. M., Nilsson, T., and Pepperkok, R. (1999). Evidence for a COP-I-independent transport route from the Golgi complex to the endoplasmic reticulum. Nat. Cell Biol. 1, 423-430.

Jiang, S., and Storrie, B. (2005). Cisternal rab proteins regulate Golgi apparatus redistribution in response to hypotonic stress. Mol. Biol. Cell 16, 2586-2596.

Martinez, O., Antony, C., Pehau-Arnaudet, G., Berger, E. G., Salamero, J., and Goud, B (1997). GTP-bound forms of rab6 induce the redistribution of Golgi proteins into the endoplasmic reticulum. Proc. Natl. Acad. Sci. USA 94, 1828-1833.

Miles, S., McManus, H., Forsten, K. E., and Storrie, B. (2001). Evidence that the entire Golgi apparatus cycles in interphase HeLa cells: Sensitivity of Golgi matrix proteins to an ER exit block. J. Cell Biol. 155, 543-555.

Nilsson, T., Jackson, M., and Pederson, P. A. (1989). Short cytoplasmic sequences serve as retention signals for transmembrane proteins in the endoplasmic reticulum. Cell 58, 707-718.

Pepperkok, R., Schneider, C., Philipson, L., and Ansorge, W. (1988). Single cell assay with an automated capillary microinjection system. Exp. Cell Res. 178, 369-376.

Pepperkok, R., Scheel, J., Horstmann, H., Hauri, H. P., Griffiths, G., and Kreis, T. E. (1993). B-COPI is essential for biosynthetic membrane transport from the endoplasmic reticulum to the Golgi complex in vivo. Cell 16, 71-82.

Pepperkok, R., Lowe, M., Burke, B., and Kreis, T. E. (1998). Three distinct steps in transport of vesicular stomatitis virus glycoprotein from the ER to the cell surface in vivo with differential sensitivities to GTP_S. J. Cell Sci. 111, 1877-1888.

Seemann, J., Jokitalo, E., Pypaert, M., and Warren, G. (2000). Matrix proteins can generate the higher order architecture of the Golgi apparatus. Nature 407, 1022-1026.

Storrie, B., White, J., Rottger, S., Stelzer, E., Suganuma, T., and Nilsson, T. (1998). Recycling of Golgi-resident glycosyltransferases through the ER reveals a novel pathway and provides an explanation for nocodazole-induced Golgi scattering. J. Cell Biol. 143, 1505-1521.

Stroud, W. J., Jiang, S., Jacks, G., and Storrie, B. (2003). Persistence of Golgi matrix distribution exhibits the same dependence on Sar1p activity as a Golgi glycosyltransferase. Traffic 4, 631-641.

Valsdottir, R., Hashimoto, H., Ashman, K., Koda, T., Storrie, B., and Nilsson, T. (2001). Identification of rabaptin-5, rabex-5, and GM130 as putative effectors of rab33b, a regulator of retrograde traffic between the Golgi apparatus and ER. FEBS Lett. 508, 201-209.

Ward, T. H., Polishchuk, R. S., Caplan, S., Hirschberg, K., and Lippincott-Schwartz, J. (2001). Maintenance of Golgi structure and function depends on the integrity of ER export. J. Cell Biol. 155, 557-570.

Yang, W., and Storrie, B. (1998). Scattered Golgi elements during microtubule disruption are initially enriched in trans-Golgi proteins. Mol. Biol. Cell 9, 191-207.

Young, J., Stauber, T., Vernos, I., Pepperkok, R., and Nilsson, T. (2005). Regulation of microtubule-dependent recycling at the TGN by rab6a and rab6a'. Mol. Biol. Cell 16, 162-177.

Was this article helpful?

0 0

Post a comment