Maelstrom is required to position the MTOC in stage 2-6 Drosophila oocytes.

The factors that determine intracellular polarity are largely unknown. In Drosophila oocytes one of the earliest polar events is the positioning of the microtubule-organizing center (MTOC). Here we present data that are consistent with the hypothesis that maelstrom is required for posterior positioning of the MTOC.

Assembly of the Drosophila germ plasm.

The Drosophila melanogaster germ plasm has become the paradigm for understanding both the assembly of a specific cytoplasmic localization during oogenesis and its function. The posterior ooplasm is necessary and sufficient for the induction of germ cells. For its assembly, localization of gurken mRNA and its translation at the posterior pole of early oogenic stages is essential for establishing the posterior pole of the oocyte. Subsequently, oskar mRNA becomes localized to the posterior pole where its translation leads to the assembly of a functional germ plasm. Many gene products are required for producing the posterior polar plasm, but only oskar, tudor, valois, germcell-less and some noncoding RNAs are required for germ cell formation. A key feature of germ cell formation is the precocious segregation of germ cells, which isolates the primordial germ cells from mRNA turnover, new transcription, and continued cell division. nanos is critical for maintaining the transcription quiescent state and it is required to prevent transcription of Sex-lethal in pole cells. In spite of the large body of information about the formation and function of the Drosophila germ plasm, we still do not know what specifically is required to cause the pole cells to be germ cells. A series of unanswered problems is discussed in this chapter.

Regulation of the vitellogenin receptor during Drosophila melanogaster oogenesis.

In many insects, development of the oocyte arrests temporarily just before vitellogenesis, the period when vitellogenins (yolk proteins) accumulate in the oocyte. Following hormonal and environmental cues, development of the oocyte resumes, and endocytosis of vitellogenins begins. An essential component of yolk uptake is the vitellogenin receptor. In this report, we describe the ovarian expression pattern and subcellular localization of the mRNA and protein encoded by the Drosophila melanogaster vitellogenin receptor gene yolkless (yl). yl RNA and protein are both expressed very early during the development of the oocyte, long before vitellogenesis begins. RNA in situ hybridization and lacZ reporter analyses show that yl RNA is synthesized by the germ line nurse cells and then transported to the oocyte. Yl protein is evenly distributed throughout the oocyte during the previtellogenic stages of oogenesis, demonstrating that the failure to take up yolk in these early stage oocyte is not due to the absence of the receptor. The transition to the vitellogenic stages is marked by the accumulation of yolk via clathrin-coated vesicles. After this transition, yolk protein receptor levels increase markedly at the cortex of the egg. Consistent with its role in yolk uptake, immunogold labeling of the receptor reveals Yl in endocytic structures at the cortex of wild-type vitellogenic oocytes. In addition, shortly after the inception of yolk uptake, we find multivesicular bodies where the yolk and receptor are distinctly partitioned. By the end of vitellogenesis, the receptor localizes predominantly to the cortex of the oocyte. However, during oogenesis in yl mutants that express full-length protein yet fail to incorporate yolk proteins, the receptor remains evenly distributed throughout the oocyte.

The gastrulation defective gene of Drosophila melanogaster is a member of the serine protease superfamily.

The establishment of dorsal-ventral polarity in the oocyte involves two sets of genes. One set belongs to the gurken-torpedo signaling pathway and affects the development of the egg chorion as well as the polarity of the embryo. The second set of genes affects only the dorsal-ventral polarity of the embryo but not the eggshell. gastrulation defective is one of the earliest acting of this second set of maternally required genes. We have cloned and characterized the gastrulation defective gene and determined that it encodes a protein structurally related to the serine protease superfamily, which also includes the Snake, Easter, and Nudel proteins. These data provide additional support for the involvement of a protease cascade in generating an asymmetric signal (i.e., asymmetric Spätzle activity) during establishment of dorsal-ventral polarity in the Drosophila embryo.

Eggshell assembly in Drosophila: processing and localization of vitelline membrane and chorion proteins.

The Drosophila eggshell consists of three major proteinaceous layers: the vitelline membrane, the inner chorionic layer, and the outer endochorion. During the latter stages of oogenesis, the proteins that comprise these layers are synthesized and secreted by epithelial follicle cells which surround the maturing oocyte. While there is considerable knowledge of the structural units which comprise the eggshell layers, there is little knowledge of how individual proteins function or interact with one another to form the structure. Immunoelectron microscopy was used to follow the distribution of four different eggshell proteins in the assembling and mature eggshell. sV23 and sV17, follicle cell proteins synthesized during the early stages of eggshell formation (stages 8-10), were distributed within the vitelline membrane layer at all stages. Despite marked temporal differences in their accumulation profiles, s36 and s18, putative chorion proteins, were similarly distributed throughout the floor, pillars, and roof of the endochorion. Although the vitelline membrane appears to be morphologically complete by stage 11, developmental Western blots and immunolocalization data indicate that molecular dynamism persists within the layer throughout the subsequent choriogenic stages. During early chorion formation the vitelline membrane appears to act as a reservoir for chorion proteins since s36 was found predominantly in the vitelline membrane layer of stage 12 egg chambers. During the late choriogenic stages (13-14), both sV17 and sV23 are processed to smaller derivatives. Interactions between the eggshell layers were suggested by ultrastructural analysis of a sV23 protein null mutant which showed that the structural integrity of the outer chorion is dependent upon the presence of a vitelline membrane component.

Brainiac encodes a novel, putative secreted protein that cooperates with Grk TGF alpha in the genesis of the follicular epithelium.

brainiac (brn) is involved in a number of developmental events. In addition to being required zygotically for segregation of neuroblasts from epidermoblasts, it is essential for a series of critical steps during oogenesis which also depend upon gurken (grk), a TGF alpha homolog. Animals harboring strong mutations of either grk or EGF receptor tyrosine kinase (Egfr) or doubly mutant for brn and weak grk or Egfr mutations produce ovarian follicles with multiple sets of nurse cell-oocyte complexes. These follicles frequently have discontinuities in the follicular epithelium that uncover nurse cells but not the oocyte. Gaps first appear in the germarium, suggesting that some nurse cells lack affinity for invading prefollicular cells. This is the first evidence that grk, in addition to its involvement in the genesis of anterior-posterior and dorsal-ventral polarity, is also required for Egfr-dependent development of the follicular epithelium that surrounds each nurse cell/oocyte cluster to form an egg chamber. We have used restriction fragment length polymorphisms to localize brn to a 10-kb region within a 300-kb stretch of DNA on the X-chromosome, and we have identified the brn gene by means of RNA rescue. brn codes for a putative secreted protein. brn is expressed in germ cells at the time follicle cells first surround the nurse cell-oocyte complex. Our genetic data suggest that brn acts in a parallel, but partially overlapping pathway to the Grk-Egfr signaling pathway. The brn pathway may help to provide specificity to TGF alpha -Egfr function during oogenesis.

Multiple signaling pathways establish both the individuation and the polarity of the oocyte follicle in Drosophila.

The development of the Drosophila oocyte depends upon a sequential series of interactions between the germline cells and the somatically derived follicle cells to produce individual follicles with appropriate polarities. In the germarium the control of germline cell division depends upon a proper interaction with somatic cells adjacent to the germline stem cells. Both gurken and brainiac are required in the germline, and the Egfr, daughterless, Notch, and Delta genes are required in the somatic cells to produce individual egg chambers with a continuous follicular epithelium. After a follicle forms, components in these same signaling pathways, plus additional genes, are then required for the establishment of the anterior-posterior polarity, followed by the dorsal-ventral polarity of the developing follicle. Initially, gurken mRNA is localized to the posterior edge of the oocyte, where it signals the posterior polar follicle cells to differentiate as posterior. The anterior-posterior assymmetry of the oocyte is then established by a reorganization of the microtubule network, which require a Notch-Delta-dependent signal sent from the posterior polar follicle cells to the oocyte and the activity of protein kinase A in the germ line. This reorganization leads to the localization of the maternal anterior-posterior determinants bicoid and oskar to opposite poles of the oocyte and the repositioning of the oocyte nucleus to the anterior-dorsal surface of the oocyte, gurken mRNA and protein are now concentrated between the oocyte nucleus and the adjacent anterior-dorsal follicle cells, where, in combination with Rhomboid, it locally activates the EGF receptor and its downstream cascade to direct the adjoining cells to adopt a dorsal fate. This process is thought to restrict the action of three follicle cell gene functions, encoded by windbeutel, nudal, and, pipe, to the ventral follicle cells, where they lead to the localized activation of a serine protease cascade required to produce the active Spätzle ligand to activate the Toll receptor. Finally, the termini of the embryo are dependent upon the activation of the Torso receptor, and this requires the localized expression of torso-like in a subset of follicle cells at the anterior and posterior poles of the follicle, which leads to the activation of Trunk, the putative ligand for Torso. In summary, the normal development of the oocyte requires a continuous sequence of germline-follicle cell interactions to provide the polarities responsible for normal development.

The Drosophila yolkless gene encodes a vitellogenin receptor belonging to the low density lipoprotein receptor superfamily.

Sequence comparisons of vitellogenins from a wide range of organisms have identified regions of similarity not only to each other but also to vertebrate apolipoproteins (e.g. apoB-100 and apoE). Furthermore, the chicken vitellogenin receptor, which also binds apolipoproteins receptor (LDLR) superfamily [Bujo, H., Hermann, M., Kaderli, M. O., Jacobsen, L., Sugawara, S., Nimpf, J., Yamamoto, T. & Schneider, W. J. (1994) EMBO J. 13, 5165-5175]. The yolk proteins of higher dipterans are exceptional, however, and instead show similarity to lipoprotein lipases. The molecular characterization of the putative Drosophila melanogaster vitellogenin receptor gene, yolkless (yl), described in this report reveals that the protein it encodes (Yl), is also a member of the LDLR superfamily. The ovary-specific 6.5-kb yl RNA codes for a protein of approximately 210 kDa which contains all three motifs common to the LDLR class of proteins. Within this superfamily, Yl may be related more to the LDLR-related proteins (LRPs), which bind both apolipoproteins and lipoprotein lipases. The similarity of Yl to the other LDLR proteins is restricted to the putative extracellular domain. Most noticeably, the cytoplasmic domain of Yl lacks the typical NPXY sequence which is involved in receptor internalization.

Identification of regions interacting with ovoD mutations: potential new genes involved in germline sex determination or differentiation in Drosophila melanogaster.

Only a few Drosophila melanogaster germline sex determination genes are known, and there have been no systematic screens to identify new genes involved in this important biological process. The ovarian phenotypes produced by females mutant for dominant alleles of the ovo gene are modified in flies with altered doses of other loci involved in germline sex determination in Drosophila (Sex-lethal+, sans fille+ and ovarian tumor+). This observation constitutes the basis for a screen to identify additional genes required for proper establishment of germline sexual identity. We tested 300 deletions, which together cover approximately 58% of the euchromatic portion of the genome, for genetic interactions with ovoD. Hemizygosity for more than a dozen small regions show interactions that either partially suppress or enhance the ovarian phenotypes of females mutant for one or more of the three dominant ovo mutations. These regions probably contain genes whose products act in developmental hierarchies that include ovo+ protein.

Multiple products from the shavenbaby-ovo gene region of Drosophila melanogaster: relationship to genetic complexity.

The Drosophila melanogaster shavenbaby (svb)-ovo gene region is a complex locus, containing two distinct but comutable genetic functions. ovo is required for survival and differentiation of female germ line cells and plays a role in germ line sex determination. In contrast, svb is required in both male and female embryos for the production of epidermal locomotor and sensory structures. Sequences required for the two genetic functions are partially overlapping. ovo corresponds to a previously described germ line-dependent 5.0-kb poly(A)+ mRNA that first appears in the germarium and accumulates in nurse cells during oogenesis. The 5.0-kb mRNA is stored in the egg, but it is rapidly lost in the embryos except for its continued presence in the germ line precursor pole cells. The ovo mRNA predicts a 1,028-amino-acid 110.6-kDa protein homologous with transcription factors. We have identified an embryonic mRNA, 7.1 kb in length, that contains exons partially overlapping those of the 5.0-kb poly(A)+ mRNA. The spatial distribution of this newly discovered transcript during midembryogenesis suggests that it corresponds to the svb function. The arrangement of exons common to the 5.0- and 7.1-kb mRNAs suggests that the Ovo and Svb proteins share DNA-binding specificity conferred by four Cys2-His2 zinc finger motifs but differ functionally in their capacity to interact with other components of the transcription machinery.

Evidence for sex transformation of germline cells in ovarian tumor mutants of Drosophila.

Mutations at a few genetic loci in Drosophila cause ovarian tumors with hundreds of poorly differentiated germ cells. We examined several of these mutants to test the hypothesis that such ovarian tumors contain sex-transformed cells. By testing for expression of male germline traits, we determined that partial germline sex transformation occurs in otu, snf, Sxlfs, and bam ovarian tumors. Thus these genes are likely to be required for proper establishment of germline sexual identity.

Glycoconjugate expression during Drosophila embryogenesis.

Glycoproteins and other glycoconjugates present on the surface of many cell types have been identified and assigned various functions. The extent of variation possible in complex glycan structures and the heterogeneity of glycoconjugate expression between two apparently similar cells has been demonstrated previously by using plant lectins to survey developmental biological models. To examine the array and extent of glycoconjugate roles in embryonic Drosophila neurogenesis, we have used plant lectins to characterize lectin receptor molecules present on the neuronal and non-neuronal cell surfaces during critical stages in axonogenesis and axon fascicle development. A collection of lectins representing a variety of hapten monocarbohydrate specificities uncovered a complex expression pattern in many glycan structures. D-Galactose-specific lectins, Bauhina purpura agglutinin (BPA) and Arachis hypogea agglutinin (peanut agglutinin, PNA), and a D-galactose/N-acetylgalactosamine-specific lectin, Glycine max agglutinin (soybean agglutinin, SBA), all recognized the surface of most cultured neurons and their axons. In the intact embryo, only the PNA and BPA receptors were found on neurons of the central and peripheral nervous systems, while SBA recognized cells of structures other than the nervous system. All three lectins recognize a high molecular weight glycoprotein when used to precipitate lectin receptor from culture homogenates. Results suggest the presence of lectin receptor glycoproteins at temporally and spatially important positions within the embryo and in culture. These glycoproteins may provide functions critical in establishing the final phenotypes of specific cells through either axon guidance/target acquisition or morphogenic adhesive events.

The germline: familiar and newly uncovered properties.

Transplantations of both Drosophila pole cells and mouse primordial germ cells suggest that the pregonadal germline of these two organisms is not pluripotent. In mouse PGCs, however, the potential for activating pluripotency is clearly present, as seen in the relative case of deriving EG cells from migratory-stage PGCs. EG cells are derived from altering the in vitro growth conditions of PGCs by the addition of one factor. Conceivably, straying of PGCs in vivo could also lead to novel fates if they migrated into a suitable environment. Aberrant PGC migration underlies models for the origin of several tumor types. How are migratory-stage PGCs prevented from adopting alternative fates in wild-type development? One solution is to link accurate migration into the gonads with germ cell survival. One mechanism of accomplishing this is through the action of pleiotrophic factors such as Steel. Steel is both a proliferation and migration factor; sterility in Steel mutants is due to poor germ cell survival and improper migration. Networking of migratory PGCs is another mechanism to reduce the chances of individual germ cells straying from the migration path. Postmigratory germ cells apparently undergo another restriction on potency--novel imprinting. Assuming that EG methylation profiles accurately reflect modifications made in their founder PGCs, then "erased" pluripotent EG lines suggest that PGCs can be diverted from the germline lineage. Direct assays on PGCs show that the unmethylated phenotype at Igf2r region 2 is characteristic of late germ cells. The maintenance of germline-like methylation with a switch in PGC fate is somewhat analogous to the situation in lag mutants from Volvox, which retain large gonidial-sized cells that can nevertheless initiate somatic differentiation. Just as it is not possible to readjust cell size with a subsequent change in gonidial differentiation, it may be difficult, without gametogenesis, to reimpose methylation modifications once they have been erased. In addition, the methylation status of some sites/loci may be inconsequential to developmental gene activity in germ cells or EG cells. Recent studies indicate that the methylation imprint at Igf2r is not interpreted at functional level until embryogenesis (51). In summary, we have reviewed some distinctive properties of the development of germ cells, from their early segregation, mitotic and meiotic profile, sex determination, to their migration and sequestration into gonads. Some constraints on development of the germline likely serve to maintain its integrity until maturation, when the gametic genomes must be capable of initiating embryonic development. To accomplish this ultimate goal, germ cells in some organisms may retain a bias towards pluripotency throughout development.(ABSTRACT TRUNCATED AT 400 WORDS)

Sex determination of germ cells in Drosophila.

Many lines of evidence indicate that in Drosophila the mechanism for establishing the sex of the female germline is different from that acting in somatic cells. In the soma Sxl has an embryonic determinative role and is required throughout the life of female flies; in germ cells its expression begins only in the larval ovary. Both the mechanism for activating Sxl and the genes controlled by Sxl are different in the germline. A number of genes have been identified that are essential either for survival (e.g. ovo, otu) or differentiation (snf, Sxl, fl(2)d, bgcn) of female germ cells. ovo is required during embryogenesis for survival of pole cells. Genetic interactions with dominant alleles of ovo and/or Sxl indicate that otu, Sxl, snf and fl(2)d act in the same pathway as does ovo. bgcn differs in that neither ovo nor SxlD mutations affect the bgcn phenotype even though XX bgcn germ cells enter the male pathway. bgcn causes sterility in both sexes. Although the germline defect is cell autonomous in mosaic gonads, bgcn is also required in the somatic tissue for maintaining oogenesis of wild-type germ cells. Several dominant suppressors of bgcn have been identified and some have properties similar to Suppressors of variegation, suggesting that chromatin structure is critical for proper germ cell sex determination.

The role of the ovarian tumor locus in Drosophila melanogaster germ line sex determination.

The locus ovarian tumor (otu) is involved in several aspects of oogenesis in Drosophila melanogaster. The possible role of otu in the determination of the sexual identity of germ cells has not been extensively explored. Some otu alleles produce a phenotype known as ovarian tumors: ovarioles are filled with numerous poorly differentiated germ cells. We show that these mutant germ cells have a morphology similar to primary spermatocytes and that they express male germ line-specific reporter genes. This indicates that they are engaged along the male pathway of germ line differentiation. Consistent with this conclusion, we found that the splicing of Sex-lethal (Sxl) pre-mRNAs occurs in the male-specific mode in otu-transformed germ cells. The position of the otu locus in the regulatory cascade of germ line sex determination has been studied by using mutations that constitutively express the feminizing activity of the Sxl gene. The sexual transformation of the germ cells observed with several combinations of otu alleles can be reversed by constitutive expression of Sxl. This shows that otu acts upstream of Sxl in the process of germ line sex determination. Other phenotypes of otu mutations were not rescued by constitutive expression of Sxl, suggesting that several functions of otu are likely to be independent of sex determination. Finally, we show that the gene dosage of otu modifies the phenotype of ovaries heterozygous for the dominant alleles of ovo, another gene involved in germ line sex determination. One dose of otu+ enhances the ovoD ovarian phenotypes, while three doses partially suppress these phenotypes. Synergistic interaction between ovoD1 and otu alleles leads to the occasional transformation of chromosomally female germ cells into early spermatocytes. These interactions are similar to those observed between ovoD and one allele of the sans fille (snf) locus. Altogether, our results imply that the otu locus acts, along with ovo, snf, and Sxl, in a pathway (or parallel pathways) required for proper sex determination of the female germ line.

The Drosophila clathrin heavy chain gene: clathrin function is essential in a multicellular organism.

The clathrin heavy chain (HC) is the major structural polypeptide of the cytoplasmic surface lattice of clathrin-coated pits and vesicles. As a genetic approach to understanding the role of clathrin in cellular morphogenesis and developmental signal transduction, a clathrin heavy chain (Chc) gene of Drosophila melanogaster has been identified by a combination of molecular and classical genetic approaches. Using degenerate primers based on mammalian and yeast clathrin HC sequences, a small fragment of the HC gene was amplified from genomic Drosophila DNA by the polymerase chain reaction. Genomic and cDNA clones from phage libraries were isolated and analyzed using this fragment as a probe. The amino acid sequence of the Drosophila clathrin HC deduced from cDNA sequences is 80%, 57% and 49% identical, respectively, with the mammalian, Dictyostelium and yeast HCs. Hybridization in situ to larval polytene chromosomes revealed a single Chc locus at position 13F2 on the X chromosome. A 13-kb genomic Drosophila fragment including the Chc transcription unit was reintroduced into the Drosophila genome via P element-mediated germline transformation. This DNA complemented a group of EMS-induced lethal mutations mapping to the same region of the X chromosome, thus identifying the Chc complementation group. Mutant individuals homozygous or hemizygous for the Chc1, Chc2 or Chc3 alleles developed to a late stage of embryogenesis, but failed to hatch to the first larval stage. A fourth allele, Chc4, exhibited polyphasic lethality, with a significant number of homozygous and hemizygous offspring surviving to adulthood. Germline clonal analysis of Chc mutant alleles indicated that the three tight lethal alleles were autonomous cell-lethal mutations in the female germline. In contrast, Chc4 germline clones were viable at a rate comparable to wild type, giving rise to viable adult progeny. However, hemizygous Chc4 males were invariably sterile. The sterility was efficiently rescued by an autosomal copy of the wild-type Chc gene reintroduced on a P element. These findings suggest a specialized role for clathrin in spermatogenesis.

Approaches to identify genes involved in Drosophila embryonic CNS development.

Many of the steps involved in formation of the Drosophila embryonic central nervous system (CNS) have been identified by both descriptive and experimental studies. In this review we will describe the various approaches that have been used to identify molecules involved in CNS development and the advantages and disadvantages of each of them. Our discussion will by no means be exhaustive; but rather we will discuss our experiences with each approach and provide an overview of what has been learned by using these methodologies. Finally, we will discuss methods that have been recently developed and how they are likely to provide further insight into CNS development.

Characterization of a putative Drosophila GTP-binding protein.

We generated a set of monoclonal antibodies raised against Drosophila antigen eluted from a lectin affinity column. One antibody, mAb 13D5 recognizes an antigen found in the most dorsal regions of the ventral midline of Drosophila embryos at stages prior to and during axonogenesis. 13D5 recognizes cells dorsal to the ectoderm in the extended germ band beginning approximately 7 hours after fertilization and in the peripheral nervous system (PNS) at about 13 hours. In addition, 24-hour-old cultures of isolated embryonic neuroblasts possess a number of cells that express the 13D5 antigen and are not recognized by horseradish peroxidase antisera. These cells extend ramified processes with multiple growth cone-like structures or large individual processes with a broad growth cone structure. 13D5 immunoprecipitates a single band with an apparent molecular mass of 58 kDa and isolated a 1.9 kb EcoRI fragment from the lambda gt11 expression libraries. In situ hybridization to staged embryos using the digoxygenin-labeled probe reveals a pattern of expression in cells just lateral to the dorsal-most regions (mesectoderm) of the ventral midline in 8- to 9-hour embryos. In situ hybridization to cultured cells derived from whole embryos reveals several cell types with differing morphologies that express transcript recognized by the digoxygenin-labeled probe. These cells may possess either broad processes containing detectable transcript, or long thin processes with no detectable transcript. Northern analysis reveals a 2.1 kb RNA transcript detectable in all embryonic stages. Nucleotide sequence obtained from the 1.9 kb insert reveals homology with the GTP-binding regions of two signal recognition particle receptors (SNRP) isolated from canine and human tissues.

Contributions of cooling and warming rate and developmental stage to the survival of Drosophila embryos cooled to -205 degrees C.

Because of their high susceptibility to chilling injury, permeabilized Drosophila embryos can not be cryobiologically preserved by slow freezing at rates low enough to prevent the formation of intraembryonic ice. Calculations indicated that to outrun the chilling injury they must be cooled and warmed rapidly at an estimated 20,000 degrees C/min or faster. Ordinarily, such cooling rates would inevitably produce lethal intracellular ice. To prevent this, embryos must contain and be surrounded by sufficiently high concentrations of glass-promoting solutes to induce vitrification on cooling and prevent devitrification on warming. Like Steponkus et al. (Nature 345, 170, 1990) we have used ethylene glycol as the solute and have exposed permeabilized 12-h embryos to it in two steps. (Permeabilization was effected by exposing dechorionated embryos to a mixture of 0.3% 1-butanol in n-heptane for 90 or 110 s.) The two steps were (i) a 30-min exposure to 2 M ethylene glycol at 23 degrees C and (ii) a 5-min exposure to 8.5 M ethylene glycol [+/- 10% polyvinylpyrrolidone (PVP)] at 5 degrees C. The volumetric response to the first step indicates that full permeation of the 2 M glycol has been approached by 30 min. The point of the second step is to raise the intraembryonic concentration of ethylene glycol to near 8.5 M ethylene glycol by osmotic dehydration. Survival based on hatching is some 45% at this point. When 12-h embryos in 8.5 M glycol containing 10% PVP are then cooled to -205 degrees C at approximately 100,000 degrees C/min and warmed at about that rate, an average of about 12% survive (hatch), although in about half the runs 15-29% survive. Survivals in the absence of PVP are usually poorer but have been as high as 40%. Currently, 5% of the surviving larvae develop to adult flies (Steponkus et al. reported 18% hatching and 3% development to adult). Embryos that develop but do not hatch show readily detectable abnormalities in mouth parts and dorsal closure. Very high warming rates are much more critical to survival than are very high cooling rates; for example, none survive when warming is 2000 degrees C/min. The deleterious effect of slow warming is exerted between -80 and -40 degrees C. The lack of reciprocity between the effects of time spent cooling and time spent warming argues against ascribing death to chilling injury. Rather, it and other data argue for ascribing death to the devitrification during warming of cytoplasm that vitrified during cooling.(ABSTRACT TRUNCATED AT 250 WORDS)