isolated in ice-cold Ringer's solution and brought to 20'C pri- or to use. Microinjection and cryomicrodissection proce- dures were as given by Horowitz et al. (1).
Proc. Nati. Acad. Sci. USA Vol. 81, pp. 1426-1430, March 1984 Cell Biology
Artifacts caused by cell microinjection (oocyte/intracellular solute/membrane damage/sodium, potassium)
DAVID S. MILLER, YING-TUNG LAU, AND SAMUEL B. HOROWITZ Department of Physiology and Biophysics, Michigan Cancer Foundation, Detroit, MI 48201
Communicated by Keith R. Porter, November 15, 1983
er's solution were injected in the vegetal hemisphere with 80-100 nl of a 14% gelatin (Difco; gel point, 24-24.50C) solution containing [3H]- or [14C]sucrose (New England Nuclear; 1-7 p.Ci/,gl of gelatin; 1 Ci = 37 GBq) using glass pipets with 10-gm o.d. tips. Pipets were pulled from 860-pum o.d. capillary glass tubing on a David Kopf model 700 vertical pipet puller, then bevelled with an alumina abrasive film (0.3 pum) mounted on a turntable. Volumes injected were O5% of oocyte volumes. A similar bolus/oocyte ratio was used for nongelatin injections. Injected cells were transferred to 100 A.l of fresh Ringer's solution and cooled on ice for 5 min before incubation at 200C. (Loss of all oocyte K into this volume would increase extracellular K by about 1 mM.) After incubation for 4, 8, or 18 hr, cells were mounted on brass blocks and quenched to -1600C in dichlorodifluoromethane/ liquid N2. Medium was sampled for determination of radioactivity. Cells awaiting cryomicrodissection were stored in
ABSTRACT The effects of microi'jection on Ranapipiens oocytes were determined using cryomicrodissection to measure Na, K, water, and injected radiolabeled sucrose (in gelatin) in the nucleus, animal, and vegetal ooplasm and injected bolus (reference phase, RP). The results point to potential problems in the interpretation of microinjection experiments. When oocytes were injected and incubated in Ringer's solution, nucleus, ooplasm, and RP lost K and sucrose and gained Na. Patterns of loss and gain were complex but were consistent with continuous solute leakage at the injection site causing artifactual intracellular diffusion gradients. In spite of leakage, oocytes completed scheduled meiotic maturation when exposed to progesterone. When oocytes were microin'ected and incubated in paraffin oil (a medium in which polar solutes cannot exchange), nuclear and ooplasmic Na, K, and water concentrations remained identical to those in uninjected cells. Neither microinjection per se nor the injected bolus affected intraoocytic solute distributions. These findings imply that, after microinjection in aqueous media, metabolites are lost from and redistribute in cells, and that these artifactual changes are inadequately reflected in the ability of the cell to carry out a complex process. They also show that injection artifacts can be avoided by injecting and incubating cells under paraffin oil.
liquid N2. Oocytes for oil incubation (2) were rinsed in ion-free isotonic sucrose solution, blotted on filter paper, and transferred to paraffin oil (Fisher), previously washed with Ringer's solution in a separatory funnel. Still in oil, the oocytes were injected as described above. Cells were then transferred to 100 p.l of fresh oil, cooled on ice for 5 min, incubated at 20'C, and frozen. Paraffin oil incubation for up to 2 days does not affect oocyte viability as determined by (i) an unchanged appearance; (ii) the ability to maintain constant cell ATP concentrations (about 3 mM in whole cells); (iii) the ability to maintain normal intracellular concentrations and distributions of water, K, Na, 3-O-methylglucose, and aaminoisobutyric acid (Table 2 and unpublished data); and (iv) the ability to complete meiotic maturation when exposed to progesterone (3). Cryomicrodissection was carried out at -450C. Ooplasmic samples were taken at separate animal (lateral to the nucleus) and vegetal hemisphere locations after removal of superficial (cortical) material; nuclear and bolus (reference phase, RP) isolation followed. Samples were weighed (at -100C), dried over P205 at 60'C, and reweighed to determine water content. Dried samples were extracted in boiling water and extracts were analyzed for [14C]- or [3H]sucrose by liquid scintillation counting and for Na and K by atomic absorption spectroscopy (1, 4, 5). Na and K data are presented as concentrations (mM) and sucrose data, as RP-normalized concentration ratios-e.g., N/RP equals nuclear dpm per ml of water divided by RP dpm per ml of water. Sucrose loss was determined for each cell by using the final counts in the incubation medium and reconstructed total counts in the cell. Since extracellular sucrose concentrations in long-term incubations approached cellular concentrations, calculated loss rates are underestimates and should be viewed as lower limits. The error is conservative, strengthening our conclusions. Data are presented as mean + SEM and the significance of differences between sample means was determined by using Student's t test. Equilibrium dialysis experiments
Because of their large size, amphibian oocytes are often used as cellular "test tubes" into which precursors and effectors are introduced by microinjection. During microinjection, the oocyte membrane is punctured and a bolus of foreign material is introduced. Few studies have addressed the problem of injection damage and resultant artifacts, and those that have only provide estimates of loss of injected material. The fates of endogenous substances seem never to be considered, even though normal cell function clearly depends on maintenance of an optimal intracellular environment. We describe here experiments that distinguish and evaluate separately loss of injected material, changes in endogenous solute concentrations, and alterations in intracellular solute distributions after microinjection. The data show that; after microinjection in aqueous media, constitutive solutes (e.g., metabolites) suffer loss and intracellular redistribution while normally excluded (extracellular) solutes enter the cell. However, our findings also suggest a method to avoid microinjection artifacts: injection and incubation in paraffin oil.
MATERIALS AND METHODS Female Rana pipiens were maintained without feeding at 24°C in dechlorinated tap water (containing neomycin sulfate, 1 mg/liter; Upjohn), which was circulated and bubbled in a hibernation enclosure. Folliculate stage Y5 oocytes were isolated in ice-cold Ringer's solution and brought to 20'C prior to use. Microinjection and cryomicrodissection procedures were as given by Horowitz et al. (1). Oocytes in RingThe publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Abbreviation: RP, reference phase.
1426
Cell
Biology: Miller et aL
Proc. Natl. Acad. Sci. USA 81 (1984)
1427
Table 1. Sucrose concentration ratios in Rana oocytes after microinjection and incubation in paraffin oil for various time periods Ratio 1 hr 4 hr 18 hr Nucleus/RP 0.73 ± 0.08 1.09 ± 0.03 1.07 ± 0.03 Animal ooplasm/RP 0.35 ± 0.04 0.52 ± 0.01 0.50 ± 0.00 Vegetal ooplasm/RP 0.32 ± 0.02 0.36 ± 0.01 0.39 ± 0.02 Data are mean ± SEM for four-six cells.
50
C z
w
age, because cells immersed in oil are closed systems that do not lose sucrose to the medium. Diffusional equilibrium was established among intraoocytic compartments by 4 hr, since sucrose distribution did not change during the subsequent 14
w
Ir 10
cr C) 5
4
8
18
HOURS AFTER MICROINJECTION FIG. 1. Loss of microinjected ["4C]sucrose from sults represent mean SEM for 6-16 cells.
oocytes. Re-
show that Na, K, and sucrose distribute uniformly between the water of RP gelatin and the water of saline solutions over wide concentration ranges. Hence, binding or exclusion of these solutes by RP is negligible. Several lines of evidence show that equilibrium RP concentrations expressed on a water basis provide a quantitative measure of nuclear and ooplasmic free solute concentrations (1, 5).
RESULTS Loss of Microinjected Sucrose. The half-time of sucrose permeation of uninjected Rana oocytes is several weeks, the intact oocyte membrane being almost impermeable (ref. 6 and unpublished data). Permeation of microinjected oocytes was much faster. In a typical experiment, loss half-times were about 6 hr, with kinetics that approximated a single exponential function (Fig. 1). These data indicate rapid onset of loss of injected material after membrane puncture and failure of repair processes to limit loss. Other experiments (data not shown), in which efflux curves were constructed for individual cells, gave loss half-times ranging from 3 to 100 hr for cells injected with sucrose in aqueous gelatin and from 9 to 90 hr for cells injected with sucrose in water. Intracellular Sucrose Distribution. The results of an experiment in which sucrose was microinjected into oocytes immersed in paraffin oil are shown in Table 1. The experiment provides control data required to assess microinjection dam-
hr. At equilibrium, the nucleus/RP sucrose concentration ratio was close to unity, showing that nucleoplasmic and gelatin water were equally accessible as a solvent for sucrose. The animal ooplasm/RP ratio was -0.51 and the vegetal ooplasm/RP ratio -0.38. These ratios are similar to those measured in Desmognathus oocytes (5, 7). Ooplasm/RP ratios are smaller than unity, in part because of the presence in ooplasm of yolk platelets, which contain water of crystal hydration from which sucrose is excluded (7, 8). The animalvegetal difference is explained by the greater yolk platelet density in vegetal than animal ooplasm. Having established 4 hr as a practical intracellular diffusional equilibration time and determined sucrose equilibrium distribution by using oocytes in paraffin oil, we compared these equilibrium distributions to the distribution ratios of oocytes in Ringer's solution, a system in which leakage to the medium occurs. After 4 hr of oil incubation, the intracellular distribution of sucrose in this experiment (Table 2) was the same as the equilibrium distribution found in the previous experiment (Table 1). This was not the case for Ringer's solution-incubated oocytes. In these cells, which were undergoing sucrose loss, nuclear/RP, animal ooplasm/RP, and vegetal ooplasm/RP concentration ratios exceeded equilibrium (oil) values, with the difference increasing with incubation time. At 18 hr, nuclear/RP, animal ooplasm/RP, and vegetal ooplasm/RP concentration ratios were 2.6, 1.9, and 1.7 times their respective equilibrium ratios. Apparently, as sucrose leaked out through the membrane puncture site, an intracellular diffusion gradient formed. The RP, located closest to the puncture, was at a downstream location on the gradient. Vegetal and animal ooplasm (in that order) were midway. The nucleus was furthest upstream. If sucrose distribution ratios in oocytes incubated in Ringer's solution reflect leak-induced intraoocytic diffusional gradients, they should change to equilibrium values when leakage is stopped. To test this prediction, oocytes were microinjected with sucrose and incubated in Ringer's solution for 14 hr. A third of the cells were frozen for analysis; the remainder were incubated for an additional 4 hr, half in Ringer's solution and half in paraffin oil, and then frozen. The
Table 2. Sucrose concentration ratios in oocytes incubated in oil or Ringer's solution for various time periods Ratio Oil Ringer's solution 4 hr 4 hr 8 hr 18 hr 1.00 ± 0.02 2.45 ± 0.240 Nucleus/RP 1.29 ± 0.05* 2.64 ± 0.27t Animal ooplasm/RP 0.53 ± 0.02 0.56 ± 0.02 0.66 ± 0.08 1.01 ± 0.13t Vegetal ooplasm/RP 0.36 ± 0.04 0.37 ± 0.02 0.45 ± 0.05 0.60 ± 0.06t Data are mean ± SEM for 4 cells incubated in oil and 16-20 cells incubated in Ringer's solution. *P < 0.05; Ringer's solution incubation vs. oil incubation. tp < 0.01; Ringer's solution incubation vs. oil incubation.
1428
Cell Biology: Miller et A
Table 3. Sucrose concentration ratios in long-term Ringer's solution-incubated oocytes with and without equilibration in oil Ringer's solution only Ringer's 14 hr 18 hr solution/oil Nucleus/RP 1.69 ± 0.18 1.77 ± 0.09 1.09 ± 0.05* Animal ooplasm/RP 0.54 ± 0.04 0.82 ± 0.17 0.37 ± 0.02t Vegetal ooplasm/RP 0.35 ± 0.03 0.47 ± 0.06 0.39 ± 0.08 Cells in Ringer's solution were microinjected, incubated for 14 hr, and divided into three groups: one was frozen, a second was incubated for an additional 4 hr in Ringer's solution, and the third was incubated in paraffin oil for 4 hr and then frozen. Data are mean SEM for five or six cells. Significances reported are for ratios for Ringer's solution incubated vs. Ringer's solution/oil-incubated oocytes. *P < 0.01. tp < 0.05.
data show that, when sucrose leakage was terminated by immersion in oil, nucleus/RP, animal ooplasm/RP, and vegetal ooplasm/RP concentration ratios returned to equilibrium values (Table 3). Intracellular Cation Concentrations and Distributions. When transmembrane K, Na, and water fluxes are at steady state, these substances are at intraoocytic diffusional equilibrium. If no membrane leakage occurred, the injection of a RP would require a transient adjustment as RP, nuclear, and ooplasmic solutes and water interdiffuse but no change in steady-state distribution would occur (1). This was the case when oocytes were microinjected and incubated in oil (Table 4). Nuclear and ooplasmic water, K, and Na concentrations were unchanged 1 hr after injection (compare with uninjected controls) and remained constant for the next 17 hr. RP K and Na concentrations measure the intracellular free fraction of these cations (4). Comparison of these values with those in Ringer's solution demonstrates the role of plasma membrane active transport in maintaining an oocyteRinger's solution activity gradient. Thus, RP K is 138 mM in oil-incubated oocytes (Table 5), compared with 2.5 mM in Ringer's solution, while RP Na is 14 mM (Table 6), compared with 118 mM in Ringer's solution. The behavior of K and Na in oocytes suffering microinjection-induced leakage can be expected to differ. Potassium should behave somewhat like sucrose, since it will also leak through the injection site from a higher intracellular activity to a lower Ringer's solution activity. Na leakage will be in the opposite direction. However, K and Na differ from sucrose in several respects. First, the oocyte membrane is normally permeable to K and Na. Second, about half of ooplasmic K and more than half of ooplasmic Na undergoes very slow isotopic exchange (9-13), primarily because of fractions "bound" to yolk (4, 8). Because of bound K and Na,
Proc. Natl. Acad Sci. USA 81 (1984) Table 4. Potassium, sodium, and water concentrations in uninjected oocytes and in microinjected oocytes incubated in oil for various time periods Microinjected, oil incubated 1 hr 4 hr 18 hr Uninjected K (mM) Nucleus 148 + 13 139 + 4 147 2 131 ± 9 Animal ooplasm 137 ± 18 143 ± 5 141 ± 4 125 ± 7 Vegetal ooplasm 119 ± 8 112 ± 3 111 + 3 129 ± 8 RP 138 ± 3 140 2 145± 3 Na (mM) Nucleus 22 ± 6 15 ± 4 23 3 27 ± 3 Animal ooplasm 51 ± 4 46 ± 2 46 2 44 ± 2 Vegetal ooplasm 42 ± 3 52 ± 2 55 3 55 ± 3 RP 17 ± 3 17 3 14 ± 1 Water (% wet weight) Nucleus 86 ± 1 88 ± 4 89 ± 1 88 ± 9 Animal ooplasm 46 ± 18 47 ± 5 46 ± 4 45 ± 7 Vegetal ooplasm 42 ± 8 42 ± 3 43 ± 3 43 ± 8 RP 82 ± 3 86 ± 2 85 ± 3 Data are mean ± SEM for four-six cells.
and the additional complication that "free" cations are partially excluded by yolk water, equilibrium relationships between ooplasmic and RP Na and K concentrations are nonlinear. Consequently, interpretation of ooplasm/RP and nucleus/RP Na and K concentration ratios from Ringer's solution-incubated microinjected oocytes is complicated and must be approached in a guarded manner. Potassium. As one might expect from the sucrose data, incubation in Ringer's solution reduced K concentrations in all cellular regions (Table 5). In the first 4 hr, reductions were modest, suggesting that, at early times, the active transport capacity of the damaged cell membrane was adequate to largely compensate for K loss through the injection site. By 8 hr, almost one-half of nuclear and more than two-thirds of RP K was lost. Apparently a significant failure of active transport capacity occurred between 4 and 8 hr. The internal distribution of K in microinjected, Ringer's solution-incubated oocytes appears consistent with leakage from a damaged microinjection site. Nuclear/RP, animal ooplasm/RP, and vegetal ooplasm/RP K concentration ratios were all higher than the equilibrium values determined in oilincubated cells (Table 5). As in the case of sucrose, this pattern suggests the presence of a gradient in which K activity is lowest adjacent to the puncture site and higher at internal regions distant from the site. Sodium. Despite the membrane damage that can be inferred from the sucrose data, only a modest increase oc-
Table 5. Potassium concentrations and concentration ratios in oocytes that were microinjected and incubated in oil, microinjected and incubated in Ringer's solution, or not injected Injected Oil Ringer's solution Uninjected 4 hr 4 hr 8 hr 18 hr Concentrations (mM) Nucleus 147 3 150 ±3 141 ±4 83 ± 5t 44 ± 7* Animal ooplasm 145 ± 7 148 ±5 128 ±6 74 ± 4t 45 ± 9* 111 ± 9 113 ± 4 117 ± 5 55 ± 3P Vegetal ooplasm 35 ± 9* RP 138 ±7 109 +6t 40 +5t 16 +5* Ratios 1.1 ± 0.04 1.4 O.lt 2.4 ± 0.2t 3.2 ± 0.2* Nucleus/RP Animal ooplasm/RP 1.1 ± 0.05 1.3 ± 0.1 2.2 ± 0.2t 2.4 ± 0.3t 0.9 ± 0.06 1.1 ± 0.1 1.7 ± 0.2t Vegetal ooplasm/RP 1.8 ± 0.2t Data are mean ± SEM for four (uninjected and oil) or 15-20 (Ringer's solution) oocytes. Significances reported are for comparison with oil-incubation oocyte values. *P < 0.001. tp < 0.05. tP < 0.01.
Cell
Biology: Miller et aL
Proc. NatL Acad. Sci. USA 81 (1984)
1429
Table 6. Sodium concentrations and concentration ratios in oocytes that were microinjected and incubated in oil, microinjected and incubated in Ringer's solution, or not injected Injected Ringer's solution Oil 18 hr 8 hr 4 hr 4 hr Uninjected Concentrations (mM) 164 ± 19* 91 ± 6* 20 ± 2* 12 ± 4 8±2 Nucleus 177 ± 13* 128 ± 4* 63 ± 3 56 ± 3 56 ± 5 Animal ooplasm 176 ± 16t 140 ± 3* 99 ± 4 95 ± 3 87 ± 4 Vegetal ooplasm 107 ± 8* 102 ± 5* 35 ± 4t 14 ± 5 RP Ratios 1.4 ± 0.1* 0.9 ± 0.04 0.6 ± 0.02 0.8 ± 0.1 Nucleus/RP 2.1 ± 0.5 1.3 ± 0.1t 1.9 ± 0.2t 4.4 ± 0.6 Animal ooplasm/RP 2.3 ± 0.6t 1.4 ± 0.1f 2.9 ± 0.3t 7.5 ± 1.3 Vegetal ooplasm/RP Data are mean ± SEM for 4 (uninjected and oil) or 15-20 (Ringer's) oocytes. Significances are for comparison with oilincubated values. *P < 0.001. tP < 0.05. tP < 0.01.
curred in RP (or free) Na of injected oocytes during the first 4 hr of incubation in Ringer's solution (Table 6). At 4 hr, a 21 mM increase in Na was compensated by an :29 mM decrease in K (Tables 5 and 6). However, in the next 4 hr, 69 mM of RP K were lost and 67 mM of Na were gained. As noted above, this massive stoichiometric reversal appears to mark a failure of the active transport mechanism that compensates for leakage. Indeed, after 18 hr in Ringer's solution, RP Na in microinjected oocytes was nearly equal to Ringer's solution Na. In oil-incubated and uninjected cells, animal and vegetal ooplasmic Na concentrations were several times those of the nucleus or the RP (Table 4). This reflects the fact that Na in yolk is about 100 mM but Na in interyolk cytoplasm and nucleoplasm is 150~~~
100 z
O100~~~~~
0 50
z
1-
150 E
4?
0 100
z
0 0
00
50
Be 0
50
% SUCROSE
100
LOSS
FIG. 2. Whole oocyte Na and K concentrations in microinjected cells that were incubated in progesterone-containing Ringer's solution as a function of loss of injected material. Open circles represent mean values for uninjected cells; variabilities were too small to be shown here (see Table 7). Closed circles represent values for individual injected cells.
material injected. We can expect, therefore, differences in the degree of damage in experiments carried out in different laboratories and with different protocols. Our kinetic data indicate that spontaneous repair does not follow microinjection in Rana oocytes. Good practice seems to require that the fates of both the injected material and endogenous solutes be monitored in every microinjection experiment. Because these control data are not routinely reported in microinjection studies, the extent of the problem is unknown. However, some measurements of loss of injected material after microinjection in aqueous media are available. For example, Ecker and Smith (15) found that [3H]leucine was lost from Rana oocytes at a maximum rate of 0.5%/min (a halftime of -2 hr) and Tso et al. (16) found that 16-36% of injected progesterone was lost from Xenopus oocytes after 5 hr (half-times of 7.7-19.8 hr). These observations are comparable with those reported here for sucrose, suggesting that our experience is not atypical. However, it must be noted that loss rates for different substances are not really comparable measures of membrane damage because they are importantly influenced (increased and decreased) by factors that may be unrelated to damage-i.e., metabolic incorporation, binding, and the normal permeability properties of the membrane. Oocyte microinjection experiments have provided considerable data on cellular changes occurring during oogenesis and on the molecular mechanisms involved in transcription, translation, and protein synthesis (for reviews, see refs. 14, 17-19). At least one of these processes, transcription by Xenopus oocytes, appears to continue for several days after microinjection. Perhaps Xenopus oocytes have repair capacity not present in Rana. Nevertheless, the present results warn that potential artifacts caused by puncturing the plasma membrane should not be ignored in the interpretation of these data. Oocytes are obviously not delicate cells. Otherwise, these artifacts would result in the disruption of all function and would be easily detected. Maturation experiments show that microinjected oocytes responded to progesterone
normally. In summary, our results show that microinjection in physiological saline may damage oocytes and that this damage should be monitored and its consequences considered. In experiments in which oocytes must be maintained in aqueous media some problems may be ameliorated by determining injection and incubation conditions that minimize membrane leakage and loss of ion regulation-e.g., small pipet tips and low volumes of material injected. However, no improvement is likely to completely prevent the considerable variability inherent in the microinjection process; when possible, leakage and electrolyte concentrations should be monitored for each cell studied. Fortunately, leakage can often be avoided by injecting and incubating cells in paraffin oil. Oocytes in oil retain viability as measured by physiological, biochemical, and developmental parameters. Oil incubation may result in some loss of flexibility in experimental design, but cells can be treated with water-soluble substrates and agents during preincubation in Ringer's solution, exposed directly to oil-soluble agents dissolved in paraffin oil, and microinjected with tracers and intracellular effectors. We thank L. Tluczek, S. Madry, S. Harmon, and J. White for excellent technical assistance and Dr. Philip Paine for valuable advice. This study was supported by National Institutes of Health Grants GM 19548 and HD 12512 and an institutional grant from the United Foundation of Greater Detroit. 1. 2. 3. 4. 5.
Horowitz, S. B., Paine, P. L., Tluczek, L. & Reynhout, J. K. (1979) Biophys. J. 25, 33-44. Merriam, R. W. (1971) Exp. Cell Res. 68, 81-87. Lau, Y. T., Reynhout, J. K. & Horowitz, S. B. Dev. Biol., in press. Horowitz, S. B. & Paine, P. L. (1979) Biophys. J. 25, 45-69. Horowitz, S. B. & Pearson, T. W. (1981). Mol. Cell Biol. 1,
769-784. 6. Horowitz, S. B. (1972) J. Cell Biol. 54, 609-625. 7. Paine, P. L., Pearson, T. W., Tluczek, L. & Horowitz, S. B. (1981) Nature (London) 291, 258-261. 8. Tluczek, L. M., Lau, Y. T. & Horowitz, S. B. Dev. Biol., in press. 9. Ecker, R. E. & Smith, L. D. (1971) J. Cell Physiol. 77, 61-70. 10. Tupper, J. T. & Maloff, B. L. (1973) J. Exp. Zool. 185, 133144. 11. Horowitz, S. B. & Fenichel, I. R. (1970) Cell Biol. 47, 120131. 12. Century, T. J., Fenichel, I. R. & Horowitz, S. B. (1970) J. Cell Sci. 7, 5-13. 13. Frank, M. & Horowitz, S. B. (1980) Am. J. Physiol. 238, C133-138. 14. Smith, L. D. (1975) in The Biochemistry of Animal Development, ed. Weber, R. (Academic, New York), Vol. 3, pp. 1-46. 15. Ecker, R. E. & Smith, L. D. (1968) Dev. Biol. 18, 232-249. 16. Tso, J., Thibier, C., Mulner, 0. & Ozon, R. (1982) Proc. Natl. Acad. Sci. USA 79, 5552-5556. 17. Masui, Y. & Clarke, H. J. (1979) Int. Rev. Cytol. 57, 185-282. 18. Gurdon, J. B. & Melton, D. A. (1981) Annu. Rev. Genet. 15, 189-218. 19. Lane, C. D. & Knowland, J. (1975) in The Biochemistry ofAnimal Development, ed. Weber, R. (Academic, New York), Vol. 3, pp. 145-181.
