Controlling transgene expression to study Xenopus laevis metamorphosis

Abstract

Sperm-mediated transgenesis of Xenopus laevis is the first application of genetic methodology to an amphibian. However, some transgenes are lethal when they are expressed constitutively. To study the influence of these genes on amphibian metamorphosis and to generate F₁ progeny from mature transgenic adults, these transgenes must be placed under the control of an inducible system so that they can be activated at specific times in development. We show that two well known binary inducible gene expression systems supplement transgenesis for the study of X. laevis metamorphosis, one system controlled by the progesterone analogue RU-486 and the other controlled by the tetracycline derivative doxycycline. By inducing a dominant negative form of the thyroid hormone receptor under the control of doxycycline specifically in the nervous system we have delimited the developmental periods within which thyroid hormone controls innervation of the developing limb from the spinal cord.

Xenopus laevis have been used as a model system for cell and developmental biology for decades. However, the “restriction enzyme-mediated integration” transgenesis method (1) that introduces genes into X. laevis embryos before first cleavage is the first genetic tool applied successfully to amphibians. Homologous and heterologous ubiquitous and tissue-specific constitutive promoters are expressed accurately in the F₀ embryos and tadpoles, and their expression patterns are transmitted faithfully to the next generations (2). We are interested in the genetic pathways that are controlled by thyroid hormone (TH) during amphibian metamorphosis. Two methods that are used to assess gene function during embryogenesis, mRNA injection into the fertilized egg and the injection of morpholino oligonucleotides, have limited use in the study of metamorphosis or any late developmental event. Tadpoles are not even competent to respond to TH until the second week after fertilization (3). A valuable adaptation of a genetic system is one where a gene of interest can be kept silent but then induced by a small molecule at an appropriate time and place in a specific cell type. For the study of metamorphosis and to generate F₁ progeny, this procedure has the advantage that embryogenesis can take place while the transgene is silent. Breeding sexually mature individuals with silent but inducible transgenes can be continued for generations.

In this study, we have demonstrated the value of two binary-inducible systems in transgenic X. laevis, the RU-486/mifepristone-inducible system (4, 5) and the tetracycline (Tet)-inducible system (6–8). The use of these inducible systems allows us to express the transgenes by using tissue-specific promoters. The progesterone analogue RU-486 is the ligand that binds a modified progesterone receptor ligand-binding domain fused to a GAL4 DNA-binding domain and a VP16 activation domain (GLVP) (4). A tissue-specific promoter drives this modified receptor that activates a transgene located on a second plasmid. This transgene of interest is cloned downstream of UAS containing four 17-bp GAL4-binding sites and a minimal promoter from the E1B major late promoter (4). We established a line of X. laevis that is transgenic for both plasmids, and we describe some of the features of this inducible system.

The Tet-inducible system was first described in 1992 (6). The Tet-off version requires the presence of the ligand doxycycline (Dox) to keep the transgene silent, a strategy that is less convenient for long-term rearing of tadpoles. Therefore, we have concentrated on the Tet-on system (7) in which gene expression is induced by addition of Dox. We used an improved Tet-on system (8) that has a very low baseline and a robust induction by doxycycline. We established a transgenic frog line that, when induced with Dox, expresses a dominant negative form of the TH receptor (TRDN) fused to GFP only in neural tissues. When this transgene is expressed constitutively in the nervous system the tadpoles never convert to leg swimming and die at the climax of metamorphosis as quadriplegics (9). We have induced the transgene expression during different windows of tadpole development, and the results suggest that TH is involved in both the formation of motor neurons and the peripheral innervation of the limb muscles.

Materials and Methods

Plasmids. The plasmids CMV/GLVP (4) and UAS-E1b-LUC (5) were gifts from B. W. O'Malley (Baylor College of Medicine, Houston). pUHDrtTA2S-M2 was a gift from W. Hillen (University of Erlangen, Erlangen, Germany) (8). A 700-bp EcoRI+BamHI fragment containing rtTA2S-M2 was isolated and the BamHI site was blunted. This fragment was subcloned into the EcoRI and StuI sites of pCS2+ (10). The resulting construct was cleaved with HindIII and NotI, yielding a 1-kb fragment containing rtTA2S-M2 with an simian virus 40 poly(A) end. This DNA was subcloned into NβT2+plasmid (2, 11) to obtain NβT/rtTA2S-M2. Similarly the HindIII+NotI fragment of pCarGFP2 (12) was replaced with the HindIII+NotI fragment from pCS2+rtTA2S-M2 to obtain pCar/rtTA2S-M2. The plasmid pUHC13-3 was a gift from H. Bujard (University of Heidelberg, Heidelberg) (6). The tetO promoter fragment from pUHC13–3 was PCR-amplified with appropriate primers and subcloned into the SalI+HindIII sites of pCS2+GFP to obtain pCS2+(tetO)GFP. The HindIII+NotI fragment of pCS2+(tetO)GFP was replaced by the HindIII+NotI fragment from pCS TRDN/GFP (12) to obtain pCS2+(tetO)TRDN/GFP.

Dox Treatment. A50mg/ml stock solution of doxycycline hyclate (Dox; Sigma) was stored in the dark at –20°C. From 5 to 20 tadpoles were raised in 4 liters of 0.1× MMR (10 mM NaCl/0.2 mM KCl/0.1 mM MgCl₂/0.2 mM CaCl₂/0.5 mM Hepes, pH 7.5) with either 5 μg/ml or 50 μg/ml Dox. The medium was changed twice a week.

Immunohistochemistry. Immunohistochemistry with an α-phosphohistone 3 antibody has been described (13).

Luciferase Activity Assay and Protein Content. Two-millimeter pieces of tadpole tail were homogenized with a Tissuemizer in luciferase extraction buffer, the extracts were centrifuged at 22,400 × g for 5 min, and 50 μl of the supernatant was used to determine luciferase activity (14). The luciferase activity for each tail was normalized to total protein in the extract by using the Pierce protein measurement kit.

Results

The RU-486-Inducible System. The RU-486 system was tested in transgenic tadpoles in both F₀ animals and F₁ lines to examine the response to the inducer and its germ-line transmissibility. The two plasmids, CMV/GLVP (4) and UAS-E1b-LUC (5), were cotransformed into X. laevis by the restriction enzyme-mediated integration method (1), the tadpoles were raised to Nieuwkoop and Faber (NF) stage 59 and then injected with 500 ng of RU-486 per g of body weight. The up-regulation of the luciferase reporter in the tails of seven individual F₀ tadpoles (Fig. 1a) was at least one order of magnitude. In one case no detectable uninduced baseline level occurred. A sexually mature transgenic male was mated with a wild-type female frog. As determined by PCR, 52% of the F₁ progeny (14 of 27) were transgenic for both transgenes. As expected all the transgenic progeny had similar baseline levels, and the luciferase reporter was induced to the same extent (Fig. 1b). Even though RU-486 added to the rearing water does induce the transgene, its insolubility makes it difficult to control the amount that is being delivered and to determine the penetration of RU-486 to all tissues. In addition, it is not desirable to repeatedly inject the same animal. For these reasons, we decided to test the Tet method.

Fig. 1.

The extent of inducibility by RU-486 of transgenes under the control of the modified progesterone system in X. laevis F₀ and F₁ tadpoles. (a) Luciferase reporter activity in the tail extracts of seven different F₀ tadpoles at stage 59 before (solid column) and 2 days after (open column) injection with 500 ng of RU-486/g of body weight. (b) Luciferase reporter activity in 1-week-old transgenic F₁ tadpoles. Column A, injected with 70% ethanol alone; column B, 24 h after a single injection of 5 ng of RU-486; column C, 24 h after a single injection of 50 ng of RU-486 in 70% EtOH; column D, uninduced; column E, induced with 10 nM RU-486 added to the tadpole rearing water for 3 days; and column F, induced with 100 nM RU-486 added to the tadpole rearing water for 3 days. In b, values are ± SEM and bars with * are significantly different (P > 0.05) relative to the uninduced transgenic and wild-type animals.

The Tet-Inducible System. To prepare Tet transgenic tadpoles, two plasmids were cotransformed by the restriction enzyme-mediated integration method. One has the ubiquitous simian cytomegalovirus (sCMV) promoter driving the modified Tet-binding protein (8). The second plasmid has Tet operator (TetO) elements upstream of GFP. One-week-old transgenic animals were induced with 5 μg/ml Dox added to the rearing water, and GFP expression was monitored at varying times (Fig. 2 a and b). The GFP was detected within 4 h and reached its peak expression 12 h after addition of Dox. The distribution of GFP expression is identical with that seen by using the same promoter driving GFP under constitutive control (2). To quantify the fold induction and measure the baseline expression a luciferase reporter was used instead of GFP. A plasmid containing the sCMV promoter regulating rtTA2S-M2 was cotransfected with one in which tetO drives luciferase. The F₀ tadpoles were grown to stage 56. About 2 mm of tail was removed from each tadpole before and after 24 h of 5 μg/ml Dox treatment, and extracts were assayed for luciferase activity and normalized for total protein content. The levels of reporter inducibility varied between individuals. However, in all but one animal the baseline of reporter gene activity was low or undetectable (Fig. 2c). Inducible levels varied from two to four orders of magnitude. To test the tissue specificity of this system, Dox-inducible GFP expression was placed under the control of two tissue-specific X. laevis promoters, neural-specific β-tubulin (NβT) (2) and the muscle-specific promoter (pCar) (1, 12). The Dox-induced expression patterns could not be distinguished from those of the constitutively expressed promoters (Fig. 2 d and e).

Fig. 2.

Inducibility and tissue-specific expression of the Tet system in transgenic Xenopus F₀ and F₁ tadpoles. Expression of the GFP reporter gene in a 1-week-old F₀ transgenic animal before (a) and after (b) 12 h of treatment with 5 μg/ml Dox. (c) Luciferase activity in the tail extracts of 11 different F₀ transgenic animals before (solid bar) and after (open bar) induction with 5 μg/ml Dox for 48 h. Tissue-specific expression in 1-week-old tadpoles induced with 5 μg/ml Dox using the NβT (d) and cardiac actin (e) promoters driving inducible GFP. [Bars = 500 μm(a, b, and e) and 200 μm(d).]

Inducing a Transgene Within Developmental Windows. When a dominant negative form of TH receptor α (TRDN) is expressed specifically in the developing brain and spinal cord, the tadpoles cannot convert from tail to leg swimming at metamorphic climax (NF stage 60) (9). The most severe phenotype is complete paralysis of the fully developed limbs and death at metamorphic climax (9). Less severely affected animals have a delayed conversion to leg swimming at the climax of metamorphosis. This phenotype has its origins in the spinal cord where TH controls the formation of limb motor neurons (9). In previous experiments the constitutive NβT promoter expressed the TRDN transgene throughout embryonic and tadpole development in the entire nervous system. Because the severely affected animals die at the climax of metamorphosis it is impossible to obtain a breeding line. A Dox-inducible NβT/TRDN (fused to GFP) transgenic male frog was grown to sexual maturity and bred with a wild-type female, producing hundreds of F₁ tadpoles. This transgenic male has both transgenes inserted together into two different chromosomes because 75% of the progeny can be induced to express the TRDN-GFP reporter. The GFP-fused TRDN transgene was observed in F₁ progeny 4 h after addition of 50 μg/ml Dox to the rearing water (Fig. 3 a and b). The GFP/TRDN transgene product remained visible at a low level 24 h after removing Dox from the water (Fig. 3c) but could no longer be detected 48 h after Dox withdrawal (data not shown). The effectiveness of the NβT/TRDN transgene in F₁ animals was demonstrated by treating 1-week-old tadpoles with 50 μg/ml Dox and 10 nM 3,5,3′-triiodothyronine for 4 days. The TH-induced cell division that normally occurs in the cells lining the brain ventricle (13) is blocked by the transgene (Fig. 4). More than 50 uninduced transgenic tadpoles were grown to metamorphosis. None of them had even a mild leg-swimming phenotype. Tadpoles were sorted into control and transgenic groups during embryogenesis by their GFP expression after a 24-h induction with 50 μg/ml Dox. This early and brief 24-h induction of the transgene had no effect on development and metamorphosis. Transgenic tadpoles that were treated throughout tadpole life (from stage 46 to climax) all developed a strong phenotype (Table 1). The limb paralysis phenotype occurred in tadpoles treated with high levels of Dox from stage 46 to 52 when limb motor neuron progenitor cells are proliferating in the spinal cord. None of the tadpoles grown in the lower concentration of Dox (5 μg/ml) developed a severe phenotype. Transgenic tadpoles that were induced to express the transgenes as late as stage 56 developed the uncoordinated limb phenotype (Table 1). Tadpoles induced with Dox after stage 58 were normal.

Fig. 3.

Kinetics of induction in the brain of the TRDN/GFP fusion protein transgene driven by the NβT promoter. Dox (50 μg/ml) was added to the rearing water. The same animal before induction (a), after 12 h of Dox induction (b), and 24 h after withdrawal of Dox (c). (The left is anterior in all images.) (Bars = 500 μm.)

Fig. 4.

TH-induced proliferation of the cells that line the brain ventricle was monitored by anti-PH3 staining (bright spots) in 1-week-old transgenic tadpoles after 4 days of treatment with 10 nM 3,5,3′-triiodothyronine (a), 10 nM 3,5,3′-triiodothyronine plus 50 μg/ml Dox (b), or untreated tadpoles (c). (The left is anterior in all images.) (Bars = 500 μm.)

Table 1. Summary of quadriplegic phenotypes in F₁ transgenic animals treated with different concentrations of Dox and stages.

		No. of tadpoles with phenotype†
Developmental stages in Dox treatment	Dox dosage*	No phenotype	Mild phenotype	Severe phenotype
46-52	High	1	3	1
46-59	High	0	0	3
52-59	Low	2	7	0
52/53-59	High	0	5	9
56-end	High	1	3	3
58-end	High	6	1	0
59-end	High	5	0	0

^*High dosage is 50 μg/ml; low dosage is 5 μg/ml.

^†Mild phenotype, leg swimming is delayed at climax; severe phenotype, severe paralysis, animals die without completing metamorphosis.

Discussion

The most commonly used method to study gene function in X. laevis embryos, injection of mRNA into the fertilized egg, is not useful for the study of metamorphosis, which cannot be studied until the second week after fertilization. The first technique to address gene function in metamorphosis was the intramuscular injection of cloned genes directly into the tadpole tail (15). These genes are expressed in a mosaic manner in a limited number of tail muscle cells of living tadpoles. On the other hand, sperm-mediated (restriction enzyme-mediated integration) transgenesis (1) integrates a transgene into the genome before first cleavage. This method coupled with an inducible gene expression system that is controlled by a tissue-specific promoter provides a plausible way to direct transgene expression temporally and spatially and to establish lines of transgenic animals with lethal transgenes expressed in specific tissues. Our laboratory has previously assessed the value of the heat shock (2) and metallothionine (unpublished data) promoters, but both of them result in substantial baselines of reporter activity. The metallothionine promoter also has a tissue-specific expression bias (e.g., stronger expression in pronephros and pharynx).

In this article, we compared two different strategies for inducing transgene expression in transgenic X. laevis. Although RU-486 added to the rearing water up-regulated a ubiquitously expressed transgene, it was not tested for prolonged exposure, as was the Dox inducer for the Tet system. RU-486 is very insoluble in water so that controlling its delivery is a problem. Multiple intraperitoneal injections over a long period are detrimental to the tadpole's development and longevity. This method is suitable for experiments that require a single injection of the inducer. The very low baseline of transgene expression of the RU-486-inducible system (Fig. 1b) will make it useful for tissue ablation studies such as those using diphtheria toxin (16).

The second system we tested, the Tet-inducible system, has great advantages for use in conjunction with Xenopus transgenesis. Dox can be added throughout any window of development.

The extent and specificity of transgene expression depends only on the fidelity of the promoter driving the Tet activator and the amount of the inducer. The ability to vary the level of transgene expression by changing the concentration of the inducer as tadpoles develop is another advantage of the inducible system. The highest dose of Dox that we used was 50 μg/ml in the rearing water. One-week-old tadpoles grew to metamorphosis in the continual presence of 50 μg/ml Dox. The animals eat, grow, and ultimately metamorphose normally under these conditions. The only abnormality observed after long-term exposure to a high concentration of Dox was a twisted tail due to a notochord defect. The lower dose of 5 μg/ml Dox induces the transgene to a lower level, causing the milder leg paralysis phenotype. This lower Dox concentration does not affect notochord morphology. With the Tet-inducible system driving the TRDN transgene, we analyzed the developmental times when TH controls limb innervation. We confirmed that premetamorphic expression of the transgene (up to NF stage 52; Table 1), a time when limb motor neurons are proliferating in the spinal cord (9), causes a mild phenotype. Likewise, application of the inducer beginning at NF stage 52 when the number of limb motor neurons is highest also results in the delayed leg swimming that is characteristic of this phenotype (9). By stage 56 nerves have entered the growing limbs and the number of limb-specific motor neurons has dropped to a fraction of the peak number reached at stage 52 (17, 18). However, the final clustering of the acetylcholine receptors in the limbs does not occur until about stage 58 (9). This experiment strongly indicates that at least two important TH-induced events are involved in the successful innervation of limbs, an early TH-controlled replication of ventricular cells in the spinal cord that give rise to limb motor neurons and then the final maturation of synapses between the nerves that enter the growing limbs and the muscle.

We have shown that transgenic X. laevis carrying a lethal transgene, controlled by an inducible system, can be grown to sexual maturity and will breed and transmit the transgene to F₁ progeny. The Tet system is particularly valuable for keeping a transgene silent and then inducing it within a precise window of development in a specific cell type. This methodology will have benefits for Xenopus research, in general, but it will be essential for genetic modifications that influence metamorphosis.