The study of reproduction and its artificial manipulation is important in many fields. For example, in sea urchins, an animal's testes can be dissected and sperm is activated by exposure to seawater. Eggs can be released by injecting KCl into the perivisceral cavity, and mixing eggs and sperm in vitro produces fertilization, as evidenced by the appearance of the fertilization membrane and subsequent development of embryos 1. These simple techniques have been the basis for such dissimilar studies as those of Berdishev 1, dealing with the role of fatty acids and cannabinoids in fertilization, to investigation of the gene expression patterns of hybrids by Nielsen and coworkers 2.

Artificial reproduction has also been well-studied in mammals, and cloning of eutherians from somatic cells is now common 345678. Harvested eggs can be enucleated and merged with a somatic cell and the reconstructed embryos cultured in vitro before being implanted into surrogate mothers 8. These methods have opened up new possibilities in both basic and applied science [e.g. 9]. Importantly, artificial fertilization has been utilized as a means of assisting with the conservation effort of declining species 1011.

Frogs have been favorite model organisms in reproductive and developmental biology for many years, mainly because of the ease with which they can be kept in captivity; their external fertilization; easily visible development in large, transparent eggs; and large numbers and ease of manipulation of their eggs. Consequently, research on frogs has often been in the vanguard of advancement in artificial reproduction techniques, and much is known about a few model species such as the African clawed frog, Xenopus laevis and the North American leopard frog, Rana pipiens [e.g. 12131415]. Indeed, the first vertebrate cloned from a somatic nucleus was a frog 16. Briggs and King injected female R. pipiens with male pituitary glands to induce ovulation and deposition of unfertilized eggs. The eggs were mechanically activated by pricking with a needle, a process which brings the pronucleus immediately under the surface of the animal pole. Taking advantage of this situation, the pronuclei were extruded, along with a small amount of cytoplasm, using a glass needle. In other species, such as the Xenopus or the axolotl, UV radiation can be used to destroy the female pronucleus instead 1718. Development was then directed by a somatic nucleus microinjected into the cytoplasm of the enucleated egg.

In another group of experiments, Kroll and Amaya 19 developed an effective and reliable method for creating transgenic Xenopus: testes were macerated in solution and the sperm membranes partially dissolved, allowing access to the condensed chromosomes. Linearized bacterial plasmids containing genes of interest were mixed in with the sperm solution and recombinant ligase was used to covalently insert the bacterial plasmids into the sperm genomic DNA, resulting in the insertion of many copies of the plasmid construct into each genome. These nuclei were then microinjected into mature eggs, generating, under appropriate conditions, hundreds of nonmosaic, transgenic embryos. Such techniques allow the investigation of gene function in these species [e.g. 20].

Clearly, there are enormous advantages to being able to manipulate a species' reproduction in the laboratory. However, despite the multiplicity of studies concentrating on anurans, to date all model species are aquatic breeders. Yet amphibians have the largest diversity of breeding strategies among terrestrial vertebrates, and it is to be expected that species with different reproductive strategies will require different methods for their manipulation in the laboratory. Therefore, many species remain experimentally intractable. Notably, terrestrial-breeding frogs, a very large and diverse group of organisms, are largely inaccessible to reproductive investigations.

The neotropical frog genus Eleutherodactylus is characterized by terrestrial breeding and direct development without an aquatic larval stage. With more than seven hundred described species, this is the largest vertebrate genus 2122. There has been considerable experimental attention focused on Eleutherodactylus frogs, ranging from basic developmental biology 23242526; to ecology [e.g. 27282930]; to the evolution of development 2331. However, there are as yet no available techniques for performing in vitro fertilization in these frogs. The development of such techniques would allow additional investigations into the genetic regulation of direct development in these species and would also assist with conservation of declining populations, an important goal considering the fact that many species of Eleutherodactylus are declining, and several are already extinct 3233.

Eleutherodactylus coqui are small tree frogs with internal fertilization and direct development 34. This species is extremely common in the forests of Puerto Rico, and it has been found that their population size is limited by the availability of retreat sites, as opposed to food resources 35. As with all other studied Eleutherodactylus species, E. coqui embryos develop directly into tiny froglets in terrestrial eggs, without a tadpole stage 36. Protocols for the husbandry of these frogs have been reported, and it is possible to maintain them in the laboratory for multiple generations 3738. A method has also been developed to induce ovulation using an artificial form of luteinizing hormone-releasing hormone (LHRH) 39. It is know in this species that sperm entry occurs at a small disc at the animal pole of the egg, and that polyspermy is apparently common but does not interfere with development 40. Cortical granules and their exocitosis have also been described using electron microscopy 40, but the large (5 mm diameter), opaque and featureless eggs make it difficult to observe the rising of a fertilization membrane. As E. coqui is arguably the best-studied terrestrial-breeding frog, we have focused on this particular species as a model for the development of reproductive techniques.

Adult Eleutherodactylus coqui frogs were collected in Puerto Rico near El Verde Field Station in the Luquillo mountains and transported to Tulane University where they were housed and fed as previously described 37. All animals were handled and experiments performed in accordance with the standards outlined in the NIH Guide for the Care and Use of Laboratory Animals.

Natural matings were performed by placing a gravid female and a calling male together as described 37. Mature, unfertilized eggs were obtained by injection of gravid females with 20 μg of des-Gly, D-Ala LHRH ethylamide (Sigma, St. Louis, MO. Catalog Number: L4513), as reported 39. Hormonally induced females were placed in plastic bags and allowed to deposit unfertilized eggs. Eggs were experimentally manipulated without moving them from the surface of the plastic bag where they were deposited. Sperm was obtained from adult male frogs that were anesthetized by immersion in 5% benzocaine solution, decapitated and double-pithed. Testes were removed by dissection and macerated with fine forceps.

In vitro fertilization (IVF) experiments using sperm in solution were carried out by macerating the testes from a single frog in 500 μL of sperm dilution buffer (SDB: 10 mM NaCl, 0.2 mM KCl, 0.1 mM CaCl₂, 0.1 mM MgCl₂, 0.5 mM Hepes pH 7.5), and adding this dropwise over the tops of the eggs or injecting it under the jelly coat using a tuberculin syringe (28.5 gauge, 13 mm length). Alternatively, small pieces of macerated testes were placed directly on top of each egg without the use of buffer solution. Incubation of sperm with egg jelly was accomplished by vigorously vortexing the jelly from one egg in 100 μl of SDB. Sperm was then incubated for 10 min in the jelly/buffer supernatant.

Sperm preparations were checked for morphology, movement and membrane integrity by fluorescent microscopy with Live/Dead sperm stain (propidium iodide and SYBR 14) (Molecular Probes, Eugene, OR) using an Olympus BH-2 microscope 41. Additionally, sperm were stained with 1 μM Lysosensor green fluorescent dye (Molecular Probes, Eugene, OR) and imaged with a Zeiss LSM 510 META laser-scanning confocal microscope 42.

For artificial activation experiments, mature eggs were either mechanically stimulated by gentle poking with fine forceps, taking care to penetrate the jelly coat but not the plasma membrane, or immersed in a solution containing 10 mM CaCl₂ with 0.1 mM A23187 calcium ionophore. Scoring of activation was done by noting the appearance of the first cleavage furrow (see figure 1). Artificial activation experiments were also performed on oocytes dissected directly from the ovisac of gravid females. These oocytes were similarly treated by poking or exposure to 10 mM CaCl₂ with 0.1 mM A23187 calcium ionophore, with or without pretreatment for 12 hrs. with 3 μM progesterone 43.

Eggs were allowed to develop at room temperature in parafilm-sealed 60 or 100 mm polystyrene petri dishes and were moistened with an antibiotic solution consisting of 25 μg/ml amphotericin B, 10 U/ml penicillin and 10 ug/ml streptomycin. Eggs were scored as successfully fertilized following neurulation (stage 2 sensu Townsend and Stewart 36; stage 14 sensu Gosner 44).

In an effort to conserve declining populations of animals, the development of protocols for the artificial manipulation of reproduction is of great interest. In the case of the neotropical frog E. coqui, we have observed that a large proportion of eggs that are laid unfertilized auto-activate. We showed that E. coqui eggs are easily activated by mechanical stimuli, leading to a need for careful manipulation of unfertilized eggs in all reproduction studies. Further, the cleavage pattern seen in mechanically activated eggs is similar to that of both auto-activated and chemically activated eggs, suggesting that mechanical stress, probably incurred in during oviposition, is responsible for the auto-activation mentioned above (table 1). Facile auto-activation of eggs has been reported in other species, complicating reproductive manipulation 48. In E. coqui, this phenomenon may relate to internal fertilization, and it would be interesting to investigate auto-activation of unfertilized eggs in closely related, externally fertilizing species such as E. antillensis 38. Some E. coqui eggs, however, remain intact and can be manipulated, showing signs of activation both at the morphological level, through the initiation of development, as well as the functional, through the slow block to polyspermy. By careful handling, we are now able to fertilize over half of the remaining, functionally viable eggs using a simple procedure. Finally, we have shown that E. coqui eggs do not degenerate rapidly and are capable of undergoing activation up to ten hours after deposition, thus creating a window of time for carrying out experimental procedures. Our results demonstrate efficient IVF in an internally fertilizing, terrestrial-breeding frog and help lay the foundation for future research and conservation possibilities in this unusually large genus of amphibians.

ET carried out the experiments and prepared the manuscript. SFM carried out preliminary experiments and supervised the manuscript. The study was conceived jointly. Both authors approved the final manuscript.