HMG Coenzyme A reductase


hmgcr is expressed in the embryo in a dynamic and tissue-specific pattern. Although hmgcr is not expressed in PGCs, it is expressed in their target tissue, the gonadal mesoderm. In the mesoderm, hmgcr is initially broadly expressed and then becomes restricted to a segmental pattern; this coincides with the stage when PGCs migrate from the endoderm into the mesoderm. Mesodermal hmgcr expression is then further restricted to a cluster of cells in each of parasegments 10, 11 and 12, corresponding to where the developing gonadal mesoderm is found. During this time, the PGCs complete their association with the gonadal mesoderm. The three clusters of hmgcr-expressing cells join to form a continuous band of cells, which is tightly associated with the PGCs in wild-type embryos, but not in hmgcr/clb mutant embryos. Thus, hmgcr is expressed at high levels in the developing gonadal mesodermat the time the PGCs specifically associate with this tissue (Van Doren, 1998).

Since hmgcr is specifically expressed in the gonadal mesoderm and is required for PGCs to associate with this tissue properly, whether hmgcr guides PGCs to the gonadal mesoderm was examined. This was done by ectopically expressing hmgcr in specific tissues using the Gal4/UAS system. When hmgcr is expressed in stripes within the mesoderm and ectoderm, many PGCs become lost and are preferentially found in places of highest ectopic hmgcr expression. These embryos are viable and the gonadal mesoderm develops normally. Thus the correct pattern of hmgcr expression within the mesoderm is essential for the PGCs to associate with the gonadal mesoderm. To test whether hmgcr expression is sufficient to attract PGCs to new locations, hmgcr was expressed in the epidermis and in the nervous system. In both cases, PGCs now associated with the tissue ectopically expressing hmgcr. Therefore, ectopic expression of hmgcr is sufficient to attract germ cells to tissues to which they would never normally migrate. In contrast, when hmgcr was expressed in PGCs, their migration was unaffected and the embryos grew up to be fertile adults. It is concluded that high-level expression of hmgcr in the gonadal mesoderm provides spatial information to guide migrating PGCs (Van Doren, 1998).

Isoprenoids control germ cell migration downstream of HMGCoA reductase

3-hydroxy-3-methylglutaryl coenzyme A reductase (Hmgcr) provides attractive cues to Drosophila germ cells, guiding them toward the embryonic gonad. However, it remains unclear how Hmgcr mediates this attraction. In a genomic analysis of the Hmgcr pathway, it was found that the fly genome lacks several enzymes required for cholesterol biosynthesis, ruling out cholesterol and cholesterol-derived proteins as mediators of PGC migration. Genetic analysis of the pathway revealed that two enzymes, farnesyl-diphosphate synthase and geranylgeranyl-diphosphate synthase, required for the production of isoprenoids, act downstream of Hmgcr in germ cell migration. Consistent with a role in geranylgeranylation, embryos deficient in geranylgeranyl transferase type I show germ cell migration defects. These data, together with similar findings in zebrafish, implicate an isoprenylated protein in germ cell attraction. The specificity and evolutionary conservation of the Hmgcr pathway for germ cells suggest that an attractant common to invertebrates and vertebrates guides germ cells in early embryos (Santos, 2004).

In many species, primordial germ cells (PGCs) originate at a location separate from that of the somatic gonad. Thus, in order to form functional gametes, germ cells often migrate through and along embryonic tissues to reach the somatic part of the gonad. Drosophila germ cell migration provides an excellent system to study the influence of multiple guidance cues on the migratory behavior of a single group of cells. Drosophila germ cells form at the posterior pole of the embryo in close apposition to the primordium of the posterior midgut (PMG). At gastrulation, as the germ band extends dorsally and the PMG invaginates, germ cells are carried to the inside of the embryo. Once in the lumen of the newly formed PMG, germ cells start to actively migrate. Initially, at early stage 10 of embryogenesis, they migrate across the PMG. They then migrate dorsally along the basal side of the PMG at late stage 10 and subsequently migrate away from the midgut toward the adjacent mesoderm at stage 11. In the mesoderm, germ cells associate with somatic gonadal precursor cells (SGPs), three bilateral clusters of mesodermal cells in parasegments ten to twelve. During germ band retraction, starting at stage 12, germ cells and the associated SGPs migrate anteriorly until the gonadal cells round up to coalesce into the embryonic gonad (Santos, 2004).

Genetic analysis of Drosophila germ cell migration has shown that lipid metabolism plays an important role during multiple stages of germ cell migration. wunen and wunen 2, encoding phospholipid phosphatases, have been shown to repel germ cells away from the most ventral region of the PMG into more dorsal regions. Drosophila Hmgcr, the enzyme responsible for the biosynthesis of mevalonate, provides attractive cues to germ cells, guiding them toward the SGPs (Van Doren, 1998). Hmgcr (the gene encoding Hmgcr is initially broadly expressed in the mesoderm, and as development proceeds its expression becomes restricted to the SGPs. Analysis of Hmgcr/columbus mutant embryos suggests that this tissue-specific expression pattern is required to attract germ cells to the mesoderm and later for the association of germ cells with the SGPs. Indeed, expression of Hmgcr is sufficient to attract germ cells as shown by the movement of germ cells toward places of ectopic expression of Hmgcr (Van Doren, 1998). In Hmgcr mutants, most germ cells fail to migrate into the mesoderm and as a result remain associated with the dorsal surface of the PMG. Besides germ cell migration defects, these mutant embryos die at the end of embryogenesis with no obvious patterning defects. Indeed, it is quite striking that mutations in Hmgcr, an enzyme required for the biosynthesis of products such as ubiquinones, carotenoids, isoprenoids, and cholesterol has a phenotype so specific for germ cell migration. Also, it remains unclear what molecule downstream of Hmgcr functions to attract germ cells. One possibility is that Hmgcr is limiting for the biosynthesis of a product that directly attracts germ cells. Indeed, Hmgcr has been described as the rate-limiting step in the mevalonate pathway. Another possibility (Deshpande, 2001) is that Hmgcr regulates protein modifications that are more indirectly required to regulate germ cell attraction (Santos, 2004).

To determine how Hmgcr may mediate germ cell attraction, the biosynthetic pathway downstream of Hmgcr was analyzed. Several enzymes required for cholesterol biosynthesis are not encoded in the fly genome, ruling out cholesterol and cholesterol modified proteins as mediators of PGC migration downstream of Hmgcr. Farnesyl-diphosphate synthase and geranylgeranyl-diphosphate synthase, the enzymes required for the biosynthesis of isoprenoids downstream of Hmgcr, are expressed in the mesoderm and are required for germ cell migration. Mutant embryos for either of these proteins show germ cell migration defects that are similar to those observed in Hmgcr mutant embryos. Furthermore, overexpression of either of these proteins in ectopic locations is sufficient to attract germ cells. Also, it is shown that mutants for geranylgeranyl transferase type I, an enzyme required for transferring geranylgeranyl pyrophosphate to target proteins, have germ cell migration defects similar to Hmgcr. In parallel to this study, Thorpe (2004), using inhibitors of the Hmgcr pathway, also identified geranylgeranylation as a critical step in zebrafish germ cell migration. These data strongly suggest that a geranylgeranylated protein common to vertebrates and invertebrates mediates germ cell attraction downstream of Hmgcr. The striking conservation of a requirement for isoprenoids in Drosophila and zebrafish germ cell migration suggests that evolutionary conserved signals may guide migrating germ cells (Santos, 2004).

In mammals, Hmgcr is the enzyme required for the conversion of 3-hydroxy-3-methylglutaryl coenzyme A into mevalonate. Mevalonate is required for the biosynthesis of many different compounds such as ubiquinones, carotenoids, and isoprenoids and cholesterol. In order to identify the components of the Hmgcr biosynthetic pathway (also called mevalonate pathway) in the fly, sequences of the human proteins involved in this pathway were used to identify the corresponding fly homologs. A single fly gene was found encoding each enzyme required for the biosynthesis of isoprenoids, from mevalonate kinase to farnesyl-diphosphate synthase and geranylgeranyl-diphosphate synthase. Isoprenoids are a family of lipids well known as posttranslational modifiers of proteins. This posttranslational modification of proteins consists of the covalent attachment of farnesyl-pyrophosphate (FPP) or geranylgeranyl-pyrophosphate (GGPP) to specific motifs on the C terminus of target proteins by their respective transferases -- farnesyl transferase (FNT) and geranylgeranyl transferase type I and type II (GGTI and GGTII). Both FNT, GGTI, and GGTII are heterodimeric proteins composed of one α and one β subunit. All of these prenyltransferases are encoded in the fly genome by a single α and β subunit. However, quite strikingly homologs of several enzymes required for the biosynthesis of cholesterol from FPP were not found. These include squalene synthase, squalene monooxygenase, and lanosterol synthase, which together catalyze the first three steps in the biosynthesis of cholesterol as well as 3β-hydroxysteroid-delta(8)-delta(7)-isomerase, sterol-C5-desaturase, 3β-hydroxy-sterol-delta(24)-reductase, and 7-dehydrocholesterol-reductase. While some fly proteins contained small regions of homology with these human proteins, they did not contain any of the conserved signature domains used to identify these specific classes of proteins, and were therefore not considered to be homologs. The lack of homology to enzymes of the cholesterol branch was confirmed at the nucleotide level. Clear fly homologs were identified for three of the enzymes required for the biosynthesis of cholesterol: lanosteroldemethylase, delta(14)-sterol-reductase, and C-4 methylsterol oxidase. The data provide the molecular basis for earlier findings that indicated that insects lack the enzymes to synthesize sterols from mevalonate. To extend this analysis to other insects, the genome of Anopheles gambiae was analyzed for the presence of cholesterol biosynthetic enzymes. As in the case of Drosophila melanogaster, several of the enzymes critical for cholesterol biosynthesis are also not encoded in this species of insects. Thus, the observed widespread inability of insects to survive in sterol free medium could be based on genomic deletion of the enzymes needed to synthesize sterols from mevalonate. Most importantly, with regard to germ cell attraction, these results unequivocally rule out cholesterol or any cholesterol-derived proteins, such as Hedgehog as mediators downstream of Hmgcr (Santos, 2004).

Since genomic analysis eliminated cholesterol as a candidate effector in germ cell migration downstream of Hmgcr, focus was placed on other branches of the mevalonate biosynthetic pathway to determine how Hmgcr attracts germ cells. It is possible that a product of the pathway, such as mevalonate, directly attracts germ cells. Alternatively, products of the pathway, such as the isoprenoids farnesyl-pyrophosphate or geranylgeranyl-pyrophosphate, could be needed for the posttranslational modification of a protein that only upon modification can attract germ cells. In either case, the enzymes required for the biosynthesis of the attractant are expected to colocalize with Hmgcr in the mesoderm, the attractive tissue, at stage 11, when germ cells are migrating away from the PMG toward the mesoderm. Therefore the RNA expression profile of each fly gene identified in the mevalonate pathway was systematically analyzed. Expression pattern analysis showed that most genes involved in this pathway are broadly expressed in the embryo. While expression of several genes was slightly increased in tissues such as the pharynx, the esophagus, and the ring gland, their predominant pattern of expression was uniform in all cells of the developing embryo. This broad expression pattern points to a general need of the mevalonate pathway in all cells of the fly embryo. Only two genes, Farnesyl pyrophosphate synthase (Fpps), the fly gene encoding farnesyl-diphosphate synthase, and quemao (qm), the fly gene encoding geranylgeranyl-diphosphate synthase, were expressed in a pattern consistent with a more specific role in germ cell attraction. fpps RNA expression in the mesoderm resembles Hmgcr expression: both genes are expressed initially uniformly in the mesoderm and subsequently develop a segmental pattern. However, unlike Hmgcr, fpps expression does not become enriched in the SGPs. Besides being expressed in the mesoderm, fpps is also expressed in the nervous system, foregut, and midgut, as well as in the ring gland later during embryogenesis. qm RNA is also highly expressed in the mesoderm, with elevated expression in the visceral mesoderm. Besides being enriched in the mesoderm, qm is also expressed in the PMG and in the nervous system at high levels in late embryonic development. Both fpps and qm are required for the biosynthesis of isoprenoids, a special group of lipids involved in the posttranslational modification of many proteins, suggesting that isoprenoids may regulate germ cell migration (Santos, 2004).

These data establish (1) that squalene and its products such as cholesterol are not produced by the Hmgcr pathway because the enzymes necessary to synthesize sterols are not present in the Drosophila genome, (2) that two enzymes in the Hmgcr pathway, farnesyl-diphosphate synthase (fpps) and geranylgeranyl-diphosphate synthase (qm), are coexpressed with Hmgcr in the mesoderm, (3) that mutations in fpps and qm exhibit similar phenotypes to Hmgcr and enhance a weak Hmgcr phenotype, (4) that misexpression of fpps and qm can attract germ cells to new sites, and (5) that mutants in geranylgeranyl transferase type I exhibit germ cell migration defects similar to those of fpps, qm, and Hmgcr. Taken together, these data suggest that germ cell attraction is dependent upon geranylgeranyl protein modifications. Similar conclusions have been reached in a study on zebrafish germ cell migration using inhibitors of specific enzymatic steps in the Hmgcr pathway (Thorpe, 2004). It is concluded that the Hmgcr pathway plays a specific and evolutionarily conserved role in germ cell attraction by limiting the production of a geranylgeranylated product (Santos, 2004).

Several enzymes required for the synthesis of cholesterol from mevalonate are missing from the fly and Anopheles genomes. This finding explains earlier observations that showed that insects require sterols in their diet, are unable to convert mevalonate to squalene or cholesterol, and lack the enzymatic activity to produce squalene or cholesterol. While these experiments suggested that neither squalene nor cholesterol is synthesized by insects, the basis of the defect remained unclear until this study. Genomic analysis in both D. melanogaster and A. gambiae shows that many of the enzymes required for the biosynthesis of cholesterol are not present in the genome of these species and provides an explanation for the need of insects to obtain cholesterol from their diet. It is unclear when during evolution and how this deletion occurred as the enzymes encoding for the synthesis of cholesterol are not clustered in one region of the genome. Further systematic evolutionary experiments will be required to clarify this aspect (Santos, 2004).

In mammals, cholesterol regulates the transcription of Hmgcr via negative feedback regulation mediated by Sterol Regulatory Element Binding Protein (SREBP). Interestingly, while D. melanogaster Hmgcr is insensitive to sterol regulation, Theopold (1996) found a functional homolog of the Sterol Regulatory Element Binding Protein in Drosophila (DSREBP or Helix loop helix protein 106). However, Drosophila SREBPs regulate the transcription of genes involved in fatty acid biosynthesis and not in cholesterol or isoprenoid biosynthesis. Also, DSREBP cleavage, which activates the protein, is inhibited by palmitate in Drosophila and not by sterols, as in mammals (Rawson 2003; Seegmiller 2002). These findings, together with the observation that enzymes required for the production of squalene are missing from both the fly and the mosquito genome, suggest that the synthesis and regulation of sterols by the mevalonate pathway may have evolved separately in the vertebrate lineage. This evolutionary diversification is in striking contrast to the apparent conservation of the pathway that leads to production of a germ cell attractant produced by the mevalonate pathway (Santos, 2004).

These results demonstrate that two genes, fpps and qm, required for the biosynthesis of isoprenoids are coexpressed with Hmgcr in the mesoderm and that mutations in these genes have a germ cell migration phenotype very similar to that of Hmgcr mutants. Enhancement of a weak Hmgcr migration phenotype by fpps and qm mutants further supports the notion that fpps and qm act downstream of Hmgcr in germ cell migration. However, overexpression of either gene in the nervous system was less effective in attracting germ cells to the new site than overexpression of Hmgcr. This result may be expected if one considers that Hmgcr is normally only weakly expressed in the nervous system; thus the effect of overexpression of enzymes that rely on Hmgcr activity for their substrate would be limited by this basal level. The notion that Hmgcr activity is limiting in the nervous system for production of the germ cell attractant is further supported by the observation that the Hmgcr overexpression phenotype is gene dosage dependent: elav-Gal4 driving one copy of UAS-Hmgcr was sufficient to attract on average 12 germ cells to the nervous system whereas two copies of UAS-Hmgcr were sufficient to attract up to 20 cells. In the mesoderm, in contrast, Hmgcr seems only limiting for germ cell attraction when its activity is reduced beyond 50%; a germ cell migration phenotype is observed in the weak homozygous mutant but not in heterozygous animals (Santos, 2004).

Embryos mutant for the β subunit of geranylgeranyl transferase type I display germ cell migration defects that are similar to the ones observed for Hmgcr, fpps, or qm mutants. This suggests that geranylgeranylation of proteins rather than an intermediate of the isoprenoid pathway is necessary for proper germ cell migration in the fly. This conclusion is further supported by the findings of Thorpe (2004) that show that zebrafish embryos treated with geranylgeranyl transferase inhibitors but not with farnesyl transferase inhibitors display strong germ cell migration defects that are similar to the defects observed in fish treated with statins, a family of potent Hmgcr inhibitors. Taken together, these results favor the idea that geranylgeranyl-PP protein modification plays a specific and conserved role in the production of a germ cell attractant. In Drosophila, restriction of expression and consequently activity of the protein modification pathway governed by Hmgcr, fpps, and qm provides the spatial instruction for migrating germ cells. In contrast, enzymes of the isoprenoid pathway have not been found to be expressed specifically in cells that attract germ cells in the zebrafish embryo and uniformly provided mevalonate can rescue the germ cell migration defect after statin treatment. If indeed geranylgeranyl modification of the same protein is critical for germ cell attraction in fly and fish, this difference suggests that in Drosophila the protein substrate is expressed in most cells of the embryo and it is the tissue-specific protein modification that provides spatial information to the attractant, while in zebrafish the attractant may be restricted in its expression but remains inactive without geranylgeranyl modification (Santos, 2004).

There are at least two possible scenarios as to how Hmgcr is attracting germ cells. In one, geranylgeranylation of a protein would indirectly result in the production of the attractant; in the other, a geranylgeranylated protein is directly attracting germ cells (Santos, 2004).

In the indirect model, geranylgeranylated proteins would promote the expression and/or secretion of the attractant molecule. Small G proteins such as Ras, Rac, and Rab are well-known prenylated signaling molecules that affect gene expression and vesicle trafficking as well as cell migration and are good candidates to play such an indirect role in germ cell migration. To test whether Hmgcr may affect gene expression during imaginal disc development, clonal analysis of Hmgcr was performed using the flp-FRT system. It was observed that whereas flies carrying small clones homozygous mutant for Hmgcr in the eye (ey-flp) or randomly in the body (hs-flp) were normal -- flies carrying large mutant clones died as pupae and displayed major body patterning defects. The expression of markers such as engrailed (en) and wingless (wg) was examined in the large Hmgcr mutant clones in wing discs and it was observed that the expression of these genes is altered. en expression appears to be downregulated whereas wg is upregulated and patchy. These results suggest that Hmgcr can indeed affect patterning, either by directly modulating the expression of target genes or perhaps by affecting the intracellular localization of factors required for gene expression. However, in Hmgcr mutant embryos, which display strong germ cell migration defects, no evident changes in the expression profile of either en or wg were seen. The same is true in embryos overexpressing Hmgcr in the nervous system. In fact, zygotic Hmgcr seems not to be required for general gene expression or secretion because Hmgcr zygotic mutant embryos display no major patterning defects, while showing strong germ cell migration defects. This suggests that germ cell migration is more sensitive to Hmgcr levels than is gene expression and secretion and makes the indirect model less likely (Santos, 2004).

In the direct model, a geranylgeranyl-PP modified protein emanating from the mesoderm would be recognized by germ cells and attract their migration toward the source. This model seems at first rather unconventional as geranylgeranylation in general fosters membrane association of proteins rather than promoting modification of a secreted protein that acts at a distance, as described for other lipid modifications including cholesterol. However, the fission yeast M-factor as well as the budding yeast a-mating factor provide well-characterized examples of prenylated proteins working directly as secreted attractants. Mating in yeast requires the reciprocal interaction of secreted pheromones with membrane bound receptors. The S. cerevisiae a-type pheromone is a farnesylated protein that is secreted by a-cells via a nonclassical export mechanism and binds the STE3 (seven-transmembrane G protein coupled receptor) on the surface of α cells. The binding of the a-factor to its receptor in the α cell promotes cytoskeletal rearrangements on this cell and the formation of a polarized center, known as: the 'shmoo'. Farnesyl modification of the a-mating factor is essential for the presentation of the ligand to its receptor and replacement of the farnesyl group by a geranylgeranyl group does not alter significantly the functionality of a-factor. In an analogous manner, one may hypothesize that a geranylgeranylated germ cell attractant may be secreted by the mesoderm and recognized by a G protein coupled receptor family member expressed in germ cells. Indeed, G protein coupled receptors have been shown to be required for germ cell migration in zebrafish, mouse, and flies. However, at this point genetic evidence in flies and fish suggests that the isoprenoid pathway is not involved in the production or modification of the respective ligands, SDF1 the CXCR4 ligand in zebrafish, or the as yet unknown ligand for the fly GPCR Tre1. Another pathway involved in Drosophila germ cell migration is controlled by the Wunen proteins, which are homologs of mammalian phosphatidic acid phosphohydrolases, LPPs. Biochemical and genetic arguments make a direct functional connection between the Wunen/LPP3 and Hmgcr pathways unlikely, because the known phospholipid substrates for Wunen/LPP3 are not among the known products of the Hmgcr pathway and because mutants in Hmgcr affect a different migratory step than wunen mutants (Santos, 2004).

While the nature of the germ cell attractant downstream of HMCCoAr remains elusive, these findings have considerably narrowed the search and provide a working model that should ultimately lead to the identification of a germ cell attractant for Drosophila and possibly beyond (Santos, 2004).

Live imaging of Drosophila gonad formation reveals roles for Six4 in regulating germline and somatic cell migration

Movement of cells, either as amoeboid individuals or in organised groups, is a key feature of organ formation. Both modes of migration occur during Drosophila embryonic gonad development, which therefore provides a paradigm for understanding the contribution of these processes to organ morphogenesis. Gonads of Drosophila are formed from three distinct cell types: primordial germ cells (PGCs), somatic gonadal precursors (SGPs), and in males, male-specific somatic gonadal precursors (msSGPs). These originate in distinct locations and migrate to associate in two intermingled clusters which then compact to form the spherical primitive gonads. PGC movements are well studied, but much less is known of the migratory events and other interactions undergone by their somatic partners. These appear to move in organised groups like, for example, lateral line cells in zebra fish or Drosophila ovarian border cells. This study used time-lapse fluorescence imaging to characterise gonadal cell behaviour in wild type and mutant embryos. The homeodomain transcription factor Six4 is required for the migration of the PGCs and the msSGPs towards the SGPs. A likely cause of this was identified in the case of PGCs; Six4 is required for expression of Hmgcr which codes for HMGCoA reductase and is necessary for attraction of PGCs by SGPs. Six4 affects msSGP migration by a different pathway, since these move normally in Hmgcr mutant embryos. Additionally, embryos lacking fully functional Six4 show a novel phenotype in which the SGPs, which originate in distinct clusters, fail to coalesce to form unified gonads. This work establishes the Drosophila gonad as a model system for the analysis of coordinated cell migrations and morphogenesis using live imaging and demonstrates that Six4 is a key regulator of somatic cell function during gonadogenesis. The data suggest that the initial association of SGP clusters is under distinct control from the movements that drive gonad compaction (Clark, 2007; full text of article).

The hedgehog Pathway Gene shifted Functions together with the hmgcr-Dependent Isoprenoid Biosynthetic Pathway to Orchestrate Germ Cell Migration

The Drosophila embryonic gonad is assembled from two distinct cell types, the Primordial Germ Cells (PGCs) and the Somatic Gonadal Precursor cells (SGPs). The PGCs form at the posterior of blastoderm stage embryos and are subsequently carried inside the embryo during gastrulation. This study has investigated the role of the hedgehog (hh) pathway gene shifted (shf) in directing PGC migration. shf encodes a secreted protein that facilitates the long distance transmission of Hh through the proteoglycan matrix after it is released from basolateral membranes of Hh expressing cells in the wing imaginal disc. shf is expressed in the gonadal mesoderm, and loss- and gain-of-function experiments demonstrate that it is required for PGC migration. Previous studies have established that the hmgcr-dependent isoprenoid biosynthetic pathway plays a pivotal role in generating the PGC attractant both by the SGPs and by other tissues when hmgcr is ectopically expressed. Production of this PGC attractant depends upon shf as well as a second hh pathway gene gγ1. Further linking the PGC attractant to Hh, evidence is presented indicating that ectopic expression of hmgcr in the nervous system (via the elav Gal4) promotes the release/transmission of the Hh ligand from these cells into and through the underlying mesodermal cell layer, where Hh can contact migrating PGCs. Finally, potentiation of Hh by hmgcr appears to depend upon cholesterol modification (Deshpande, 2013).

The synthesis of mevalonic acid by the enzyme Hmgcr is the rate-controlling step in the biosynthesis of isoprenoids and steroids. In mammals, one end-product of the mevalonate pathway, cholesterol, is used to modify the C-terminus of the processed Hh ligand, and this modification plays an important role in controlling the activity of this signaling molecule. Flies lack the enzymes needed for de novo cholesterol biosynthesis and depend instead upon exogenous cholesterol for this modification of the Hh ligand. Nevertheless, the mevalonate biosynthetic pathway is still used to potentiate the release/transmission of the Hh ligand, in this case through (at least in part) the geranylation of the G protein Gγ1 (Deshpande, 2009). Hmgcr as well as the downstream components in the isoprenoid biosynthetic pathway also play a pivotal role in generating the attractant that guides PGC migration both from its native source, the SGPs, and from a variety of different embryonic tissues when ectopically expressed. However, how hmgcr or the other isoprenoid pathway enzymes function in generating the PGC attractant either in the SGPs or at ectopic sites has remained unresolved and contentious. To address this problem this study has focused on the connection between the mevalonate→isoprenoid biosynthetic pathway and two proteins that have been implicated in the long distance basolateral transmission of the Hh-Np ligand, the G protein Gγ1 and the extracellular hh signaling factor Shf (Deshpande, 2013).

Previous studies have established that a rate limiting step in generating the PGC attractant either by the SGPs or by other tissues and cell types is the biosynthesis of geranylgeranyl-pyrophosphate by geranylgeranyl diphosphate synthetase (qm). The control point in the geranylgeranyl-pyrophosphate biosynthetic pathway is the production of mevalonic acid by the enzyme Hmgcr. While hmgcr seems to play a rather similar role in the release/transmission of Hh-Np from hh sending cells, in this case through the geranylation of Gγ1, an important and controversial question is whether the functioning of the hmgcr-->qm biosynthetic pathway in hh signaling has any connection to the generation of the PGC attractant. This question was addressed by determining if the PGC migration defects induced by hmgcr expression in the nervous systems depend upon Gγ1 and Shf. It was found that mutations in both gγ1 and shf dominantly suppress the migration defects induced by ectopic hmgcr. In contrast, reducing the dose of the hmgcr gene dominantly enhances the migration defects induced by hmgcr expression in the nervous system. This later finding is expected since reducing hmgcr activity in the SGPs should further compromise the ability of the attractant generated by the SGPs to compete with the attractant generated in the nervous system. The former findings show that the production/activity of the attractant generated in the nervous system by ectopic hmgcr depends on both gγ1 and shf. By themselves, these results do not exclude the possibility that gγ1 and shf only collaborate with hmgcr when it is ectopically expressed in the nervous system while they are not actually needed for the hmgcr-dependent production of the attractant by the SGPs. However, this scenario seems unlikely. For one, there are PGC migration defects in gγ1 and shf mutant embryos. For another, the Gγ1 protein must be geranylated to function in PGC migration (Deshpande, 2009). Finally, like hmgcr, ectopic expression of gγ1 and shf in the mesoderm and ectoderm perturbs PGC migration (Deshpande, 2013).

Even though Gγ1 and Shf are known to function in the release and transmission of the Hh ligand, it could be argued that these two proteins could also mediate the release/transmission of others molecules, including the 'actual' PGC attractant. Indeed, Gγ1 is likely involved in secretion of other molecules, while the fact that Shf homologs in mammals function in Wnt but not Hh signaling raises the possibility that Shf could promote signaling by an as yet unknown ligand (though not Wg). However, there is evidence that like Gγ1 and Shf, Hh itself depends upon hmgcr and the isoprenoid biosynthetic pathway not only in hh signaling but also in generating an ectopic PGC attractant in the nervous system. This comes from the differences in the effects of ectopically expressed Hh-Np (internally autoproteolytic cleavage product coupled with cholesterol addition) and Hh-N (lacking the cholesterol modification) that would be predicted based on the mechanisms proposed for their transmission. First, the apically transmitted Hh-N ligand would be expected to have a smaller effect on PGC migration when ectopically expressed in the nervous system than Hh-Np. With the caveat that expression of different UAS transgene inserts will not be precisely the same, this prediction holds. Second, the geranylation of Gγ1 in response to ectopic Hmgcr would be expected to promote the basolateral release and subsequent spreading of Hh-Np into the mesoderm. By contrast, ectopic Hmgcr should have less influence on Hh-N, which isn't readily internalized by hh sending cells and spreads mostly along the apical surface. With the same caveat, this predicted distinction is also observed. When co-expressed, Hh-Np and Hmgcr collaborate to strongly potentiate PGC migration defects, while there is a more modest collaboration between Hh-N and Hmgcr (Deshpande, 2013).

Though an imperfect mimic of Hh-Np, advantage was taken of a chimeric Hh-GFP fusion protein to analyze the effects of Hmgcr on the transmission of Hh from cells in the embryonic nervous system. Hh-GFP was found to be less effective than Hh-Np (and even Hh-N) in competing with the PGC attractant produced by the SGPs when it is ectopically expressed using the twi or elav GAL4 drivers. Since Hh-GFP appears to have near but not quite wild type activity in morphogenesis, it is surprising that it is relatively ineffective in altering PGC migration. However, a plausible reason for this discrepancy is that the demands imposed by the assays used to test Hh-GFP activity in each experimental context are quite different. The morphogenesis assay requires that Hh-GFP substitute for Hh-Np. Since animals can readily tolerate heterozygosity for hh, small deficits in the functioning of the chimeric protein might only have minimal effects on morphogenesis. In contrast, in the PGC migration assay the ectopically expressed Hh-GFP must be able to compete with the attractant(s) produced by the SGPs. If Hh-Np is the relevant endogenous PGC attractant, then even subtle deficiencies in the activity of the chimeric Hh-GFP ligand would be expected to compromise its ability to compete with the wild type protein. It would also follow that it should be possible to 'rescue' ectopic Hh-GFP by enhancing its activity. This is the case. While hh-GFP is not very active on its own, it is able to collaborate with hmgcr when co-expressed in the nervous system (Deshpande, 2013).

Previous studies have shown that expressing hmgcr in hh producing cells in the ectoderm increases the overall level of Hh protein and enhances its transmission to adjacent cells. Precisely the same sorts of effects on Hh-GFP are evident when it is 'rescued' by co-expression with hmgcr in the nervous system - Hh-GFP levels are elevated, while its transmission into and through the underlying mesodermal cell layer is appreciably enhanced. These hmgcr dependent effects, particularly on the movement of Hh-GFP from the neuroectoderm into the underlying mesoderm, would also provide a plausible explanation for why this biosynthetic enzyme plays such a pivotal role in PGC migration even though it is not directly responsible for the synthesis of the PGC attractant. In the period when PGCs are migrating through the mesoderm, the SPGs are the only cells in the embryo simultaneously expressing both hmgcr and hh. Consequently the accumulation, release and transmission of Hh-Np will be specifically potentiated in SGPs, but not in other hh expressing cells elsewhere in the mesoderm or in the ectoderm. This would provide a mechanism for ensuring that SGP derived Hh-Np out-competes Hh-Np produced elsewhere. Taken together, these findings support the idea that Hh-Np expressed in the SGPs functions as a PGC attractant. With the caveat that the activities of Hh-GFP are not identical to Hh-Np, the fact that Hh-GFP accumulates on the surface and around the PGCs further bolsters this suggestion. Moreover, in a subset of the PGCs Hh-GFP is closely associated with bulges or protrusions that could potentially be of relevance to the process of migration (Deshpande, 2013).

A number of critical questions remain. For one, it is not clear how reception of the hh signal could actually translate into directed movement. The endpoint of the signaling cascade in the canonical pathway is the transcriptional activation of target genes, including the hh receptor ptc. However, transcription is likely not involved in this instance, as ptc reporters are not activated in PGCs. Moreover, in mammals hh dependent axonal guidance and fibroblast migration are independent of transcription and involve instead the coupling of Smo activation to pathways that mediate the reorganization of the cytoskeleton. Further studies will clearly be required to establish a connection between hh signaling to the PGCs, changes in the cytoskeleton and directed movement. Another unresolved question is whether SGPs produce any other PGC attractants. Although no other plausible candidates have been identified, the current experiments do not exclude the possibility that there are other PGC attractants, even including an attractant(s) whose activity, like Hh-Np, is potentiated by the hmgcr isoprenoid biosynthetic pathway (Deshpande, 2013).

Effects of Mutation or Deletion

HMG-CoA reductase guides migrating primordial germ cells

The enzyme 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase is best known for catalysing a rate-limiting step in cholesterol biosynthesis, but it also participates in the production of a wide variety of other compounds. Some clinical benefits attributed to inhibitors of HMG-CoA reductase are now thought to be independent of any serum cholesterol-lowering effect. This study describes a new cholesterol-independent role for HMG-CoA reductase, in regulating a developmental process: primordial germ cell migration. In Drosophila this enzyme is highly expressed in the somatic gonad and it is necessary for primordial germ cells to migrate to this tissue. Misexpression of HMG-CoA reductase is sufficient to attract primordial germ cells to tissues other than the gonadal mesoderm. It is concluded that the regulated expression of HMG-CoA reductase has a critical developmental function in providing spatial information to guide migrating primordial germ cells (Van Doren, 1998).

In many animals, including mammals and Drosophila, primordial germ cells (PGCs) form in a region of the embryo separate from the somatic portion of the gonad. Consequently, PGCs migrate through the embryo and make specific contacts with somatic cells to form the gonad. In a genetic screen for mutations that disrupt this process in Drosophila, 15 alleles of a gene called columbus(clb) were identified. clb was cloned and found to encode a Drosophila homologue of HMG-CoA reductase (hmgcr). Twelve ethyl methane sulphonate (EMS)-induced alleles and one transposon-induced allele of clb were examined and all showed molecular defects in the hmgcr transcript: it is therefore concluded that the clb alleles correspond to hmgcr/clb. All missense mutations found in hmgcr alter amino acids in the catalytic domain that are conserved from humans to archaebacteria, suggesting that it is the catalytic activity of HMG-CoA reductase that is important for PGC migration (Van Doren, 1998).

Mutations in hmgcr severely disrupt primordial germ cell migration. In wild-type Drosophila embryos, PGCs first migrate through the posterior endoderm, move to the dorsal surface of the endoderm, and then migrate into the adjacent mesoderm. Once in the mesoderm, the PGCs associate with a subset of mesodermal cells that will give rise to the somatic portion of the gonad (gonadal mesoderm). Subsequently, the PGCs and gonadal mesoderm coalesce to form the embryonic gonad. In hmgcr/clb mutants, some PGCs fail to migrate from the endoderm to the mesoderm, and instead remain associated with the endoderm. In addition, many PGCs that do migrate to the mesoderm fail to associate with the gonadal mesoderm, and instead scatter widely in the embryo. Thus, HMG-CoA reductase is required for PGCs to migrate to the mesoderm and to find their target tissue within the mesoderm (Van Doren, 1998).

Several mutants that show defects in PGC migration affect development of the gonadal mesoderm. Therefore gonad formation was examined in hmgcr/clb mutant embryos. Early specification of gonadal mesoderm precursors occurs normally in hmgcr/clb mutants, as judged by the expression of Zinc-finger homeodomain protein-1 and Clift. At later stages, hmgcr/clb mutant embryos continue to express gonadal mesoderm-specific markers such as Clift, DWnt-2, and the 412 retrotransposon. In addition, the gonadal mesoderm coalesces normally into a correctly patterned embryonic gonad that contains few or no PGCs. Thus, in contrast to previously described mutants, the PGC migration defect observed in hmgcr/clb mutant embryos cannot be attributed to a failure in gonadal mesoderm development. The development of several other tissues occurs normally in hmgcr/clb mutant embryos, including the nervous system and tracheal branches, which require directed cell movement for their formation. It is still possible that these processes are influenced by maternal HMG-CoA reductase which is required for early embryonic development. Although hmgcr/clb mutant embryos do not have any gross morphological defects, they do die at the end of embryogenesis. This is consistent with HMG-CoA reductase being required for some more general cellular function(s), in addition to being specifically required for PGC migration (Van Doren, 1998).

HMGCoA reductase potentiates hedgehog signaling in Drosophila

hh is expressed exclusively in the posterior compartment of the wing disc and orchestrates wing development by signaling the expression of downstream target genes such as decapentaplegic (dpp) and ptc in the anterior compartment. In the absence of hh signaling, these target genes are not properly activated, resulting in defects in growth and patterning along the anterior/posterior axis. Conversely, when hh is inappropriately expressed in the anterior compartment, it activates dpp in a pattern that leads to overgrowth of anterior tissues and the partial duplication of distal wing structures. These gain-of-function phenotypes are associated with a dominant hh mutation, hhMoonrat (hhMrt), that causes a partial transformation of anterior wing to posterior (Felsenfeld, 1995). The anterior-to-posterior transformations induced by the Mrt allele can be dominantly suppressed by mutations in hh signaling pathway genes that are required to promote hh signaling in either the sending or responding cell. To assess if hmgcr influences hh signaling in the wing, interactions with Mrt were tested. As positive controls mutations in the hh signaling pathway gene dispatched (disp) were used; disp is thought to function in the sending cell (Deshpande, 2005).

The Mrt wing blades were assigned to five different classes (classes I-V based on the severity of the wing phenotype (Felsenfeld, 1995). Roughly 75% of the control Mrt wing blades (hhMrt/TM3) fall into classes III and IV, which represent moderate to relatively severe wing deformations. The phenotypic effects of Mrt can be dominantly suppressed by the hmgcr mutation, and 75% of the wing blades in hhMrt/hmgcr1 trans-heterozygotes belong to either classes I or II, which represent nearly normal wing morphology. Moreover the extent of suppression of the Mrt wing phenotypes by hmgcr is equivalent to that observed when a disp mutation is trans to hhMrt (Deshpande, 2005).

One model that could explain the suppression of the Mrt wing phenotypes is that hmgcr potentiates hh signaling. If this is correct, then hmgcr mutants might be expected to exhibit segmentation defects similar to those of known hh pathway genes. To explore this possibility, cuticles of hmgcr embryos were examined. Nearly 30% (14/49) of the hmgcr embryos showed fusions of one or more segments and/or the deletion of pattern elements characteristic of mutations in segment polarity genes. The most prevalent defects were the fusion of abdominal segments 7 and 8 (9/14); however, more severe disruptions in patterning were also evident. The same types and range of patterning defects were observed for another hmgcr allele. The frequency of such defects in control embryos was never more than 3%-5% (Deshpande, 2005).

Although segment polarity defects are clearly evident in hmgcr embryos, the cuticle phenotypes are much less severe than those seen for genes like hh and wg (which give a lawn of denticles). One explanation for the relatively weak segment polarity defects is that maternally derived Hmgcr compensates for the lack of the zygotic gene product. To test this possibility, hmgcr germline clones were generated. While fertile females were not obtained for the strong hmgcr11.57 allele, fertile females were obtained for the hypomorphic allele hmgcr11522. These females were mated to either hmgcr11.57/TM3 Ubx-LacZ or wild-type males. The cuticle phenotypes observed when the germline clone females were mated to heterozygous hmgcr males were examined. The embryos could be divided into roughly four groups. Group I (15%) embryos arrested development without forming cuticle. In a subset of these embryos, abnormal mouth parts and/or filzkorper could be detected. Group II (27%) embryos formed at least some cuticle, but embryos had severe segmentation defects. Many of the embryos in this group had deletions/fusions of cuticle pattern elements. In others, cuticle structures like the denticle belts were incompletely formed. Much less pronounced developmental defects were observed in embryos in groups III and IV. Embryos in group III (22% embryos) had fusions of abdominal segments 7 and 8, but were otherwise normal. Embryos in group IV, which represents about 37% of the embryos, resembled wild-type; however, less than half of these animals hatched, suggesting that they may have other vital defects. Since group III or IV embryos were observed only when the hmgcr germline clone females were mated to wild-type males, it is presumed that embryos in groups I and II were fertilized by hmgcr mutant sperm. Three conclusions can be drawn from these data: (1) there is a substantial hmgcr maternal contribution; (2) the loss of this maternal product can be partially compensated by zygotic expression from the paternal gene; (3) while hmgcr seems to function in the wg-hh regulatory circuit, it must have additional roles that are critical for normal development that may be unrelated to the segment polarity pathway (Deshpande, 2005).

To provide additional evidence that hmgcr functions in segment polarity, the pattern of wg expression was examined in hmgcr mutant embryos. Up until stage 9/10, no defects were discerned in the pattern or level of wg stripe expression in the ectoderm of hmgcrz− embryos. However, beginning around stage 11, wg expression in hmgcrz− embyros is downregulated, and the level of Wg accumulation is reduced compared to wild-type. Further reductions in Wg protein accumulation are evident in older hmgcrz− embryos, though even in these older embryos, some residual Wg protein can still be seen in the ectoderm. These findings indicate that hmgcr resembles hh in that it is not required in the initial activation of wg stripe expression in the ectoderm, but is required to sustain wg expression. In contrast, the effects of reduced hmgcr activity on wg are considerably less severe than those seen in hh null mutant embryos. In the absence of hh, wg stripe expression in the ectoderm disappears almost completely by the end of stage 9, whereas small amounts of Wg protein are still clearly evident in stage 12 and older hmgcrz− embryos. This difference could indicate that hmgcr activity is not essential for maintaining wg expression. Another factor that could contribute to the difference is the substantial maternal contribution of hmgcr. To confirm this possibility, Wg expression was examined in progeny from hmgcr11522 germline clone females mated to hmgcr 11.57/TM3 Ubx:LacZ males. As expected, the effects on Wg expression were more pronounced when maternal hmgcr activity was compromised (Deshpande, 2005).

To confirm these findings, the expression of the Engrailed (En) protein was examined in hmgcr mutant embryos. hh signaling is required to maintain a high level of En expression in the stripes, and, in hh mutants, en expression begins to decay around stages 10-11. hmgcr is also required to maintain a high level of En expression, and, in embryos lacking zygotic hmgcr activity, En expression is reduced compared to wild-type by stage 11 (Deshpande, 2005).

The failure to maintain high levels of wg expression in older embryos would be consistent with the idea that Hmgcr is required for sending and/or receiving the Hh ligand. To test this hypothesis, the distribution of the Smo protein was compared in wild-type, hmgcrz− (hmgcr11.57), and hh embryos. Previous studies have shown that reception of the Hh signal stabilizes Smo protein and induces it to relocalize from intracellular membrane vesicles to membranes on the cell surface. In wild-type embryos, the effects of Hh signaling on Smo stability and localization can be visualized as a series of stripes that are about five cells wide. In these stripes, Smo is concentrated predominantly at the surface of the cell, giving a ring around the edge of each cell in the stripe in confocal cross-sections. The stripes are separated by a band of about five cells that have a lower level of localized Smo. In hmgcrz− embryos, the stripe pattern is much less well defined. Moreover, unlike wild-type, the Smo protein is not tightly localized to the cell surface in many of the cells in the stripe, but, instead, it is distributed in the cytoplasm. Though the Smo localization pattern across each segment in hmgcrz− embryos is disrupted, the effects on Smo are not as severe as those seen in hh null embryos (Deshpande, 2005).

The defect in Smo relocalization in hmgcrz− embryos supports the idea that Hmgcr activity is required for the production and/or activity of the Hh ligand. To test this possibility further, the pattern of Hh accumulation was compared in wild-type and hmgcrz− (hmgcr11.57) embryos. In wild-type embryos, Hh is expressed in each parasegment in a two cell wide stripe, and the protein in these cells is distributed around the membrane in a punctate pattern. Extending outward in either direction from the stripe is a relatively sharp gradient of Hh protein. Like the cells in the stripe, the Hh protein associated with the interstripe cells is generally distributed in a punctate pattern around the membrane. No defects in Hh protein expression are apparent in hmgcrz− embryos, and, as seen in wild-type, there is a two cell wide stripe of Hh-expressing cells in each parasegment. Moreover, like wild-type, the protein is concentrated in a punctate pattern around the cell membrane. In contrast, the amount of Hh protein in the hmgcrz− stripes is considerably higher than wild-type. Concomitant with the increase in the level of Hh in cells in each stripe, the amount of protein in interstripes is greatly reduced in hmgcrz− embryos relative to that seen in wild-type. Similar results were obtained for hmgcrm−z− (Deshpande, 2005).

The abnormal pattern of accumulation of Hh seen in hmgcr mutant embryos suggests that hmgcr is required for the efficient release of Hh from the two cells that express this ligand and/or in the transport of Hh from these cells to the adjacent receiving cells. To test this idea further, the effects of ectopically produced Hmgcr were examined on expression of the Hh target gene wg. It was reasoned that if Hmgcr functions primarily in Hh-producing cells to promote the efficient release or dispersal of the Hh ligand, then overexpression of Hmgcr in these cells might be expected to have a more pronounced effect on wg than overexpression in the neighboring Hh-receiving cells. To direct Hmgcr expression in cells that normally produce the Hh ligand, an hh-Gal4 driver was used, while, for the control, either a ptc or a wg driver was used to direct Hmgcr expression in cells that normally respond to the Hh ligand (Deshpande, 2005).

These expectations were met. In embryos in which Hmgcr is expressed in Hh-receiving cells by using the ptc or wg driver, the pattern of Wg accumulation resembles that of wild-type. Wg is expressed in a single cell wide stripe in each parasegment, and it localizes in these cells in a punctate pattern near the cell membrane. A low level of Wg associated with the membranes of cells in the interstripe region can also be detected. In contrast, when Hmgcr is expressed in Hh-producing cells, the level of Wg accumulation cells in the wg stripe is substantially upregulated. Moreover, an expansion of the stripe from a single cell to a two cell wide stripe was sometimes observed. In addition, high amounts of Wg could be seen extending through much of the interstripe region (Deshpande, 2005).

The driver-dependent effects of Hmgcr on Wg accumulation would be consistent with the idea that Hmgcr is most effective in enhancing Hh signaling when it is expressed in Hh-producing cells. To test this idea further, the distribution of Hh protein was examined in embryos in which Hmgcr was expressed under the direction of the hh, wg, and ptc drivers. When Hmgcr expression is directed by the wg or ptc drivers, the distribution of Hh protein resembles that seen in wild-type embryos. Hh accumulates in a two cell wide stripe in each parasegment, while there is only a relatively low level of Hh protein to either side of this stripe. A different result is obtained with the hh driver. Though the Hh parasegmental stripes are still discernable, the stripes are much broader than in wild-type (or when Hmgcr expression is controlled by wg or ptc-Gal4), and there are high levels of Hh extending to almost the middle of the interstripe region. The effect of ectopic hmgcr was also examined by using a paired-Gal4 driver that drives expression in alternate segments. As expected, it was found that the breadth of the Hh stripe was increased in alternate segments (Deshpande, 2005).

In the embryo, hh and wg establish an autoregulatory circuit in which signaling by one ligand potentiates signaling by the other. Thus, it is formally possible that the upregulation of hh signaling evident when hmgcr is overexpressed by using the hh driver is the indirect consequence of augmenting the reception of the wg signal in hh-expressing cells. To exclude this possibility, whether hh signaling can be potentiated by ectopic expression of hmgcr in hh-sending cells was tested in the wing disc, in which there is no autoregulatory circuit between hh and wg. In the wing disc, the Hh ligand is expressed in the posterior compartment, and it promotes Ptc protein accumulation in the anterior compartment along the compartment boundary. When Hmgcr is expressed in the receiving cells by using the ptc-Gal4 driver, there is little effect on Ptc accumulation, and it resembles that in wild-type. By contrast, when Hmgcr is ectopically expressed in hh-sending cells by using the hh-Gal4 driver, Ptc accumulation is upregulated. These findings indicate that Hmgcr can function in hh sending cells in the wing disc to potentiate hh signaling (Deshpande, 2005).

Although neither hh nor disp is haplo-insufficient with respect to germ cell migration, synergistic genetic interactions are observed when mutations in these two genes are combined in trans. Germ cell migration is essentially indistinguishable from wild-type in embryos heterozygous for an hh mutation, and fewer than 20% of the stage 13 to stage 16 embryos have four or more mispositioned germ cells. This is also true for embryos heterozygous for a mutation in disp. In contrast, germ cell migration defects are readily apparent in the trans combination, and nearly 90% of the stage 13 to stage 16 embryos have ten or more mispositioned germ cells (Deshpande, 2005).

If the requirement for hmgcr function in germ cell migration is related to its role in promoting the transmission or movement of the Hh ligand, then equivalent synergistic genetic interactions between hmgcr and either hh or disp should be found. This is the case. There are minor germ cell migration defects in hmgcr/+ embryos, and about 35% of the stage 13 to stage 16 embryos have four or more mispositioned germ cells. These minor defects are substantially exacerbated when hmgcr is combined with mutations in hh or disp. In hh/hmgcr trans-heterozygotes, more than 95% of the stage 13 to stage 16 embryos have ten or more mispositioned germ cells. Similarly, like the disp/hh combination, a high frequency of germ cell migration defects are evident in the disp/hmgcr trans combination (Deshpande, 2005).

In embryos heterozygous for the hh gain-of-function allele hhMrt, there are defects in germ cell migration, and about 60% of the stage 13 to stage 16 embryos have four or more mispositioned germ cells. Like the wing abnormalities in hhMrt flies, this weak germ cell migration phenotype is presumed to arise from the misexpression of Hh protein. However, the mechanism is likely to be different from that involved in the mispecification of anterior compartment cells by ectopic Hh. In this case, the ectopic Hh expressed by the Mrt allele probably competes with the protein produced by the somatic gonadal precursor cells as an attractant and misdirects the migrating germ cells. If the effects of Mrt on migration are due to competition, it was reasoned that it should be possible to enhance the germ cell migration phenotype of hhMrt by reducing the potency of the Hh signal emanating from the somatic gonadal precursor cells. Consistent with this expectation, mutations in both hh and disp significantly increase the severity of the migration defects seen in hhMrt, and, in each case, almost all of the embryos had ten or more mispositioned germ cells. An hmgcr mutation also substantially enhances the hhMrt migration defects, and its effects are equivalent to mutations in either hh or disp (Deshpande, 2005).

slow as molasses, acting in a parallel pathway to Hmgcr, is required for polarized membrane growth and germ cell migration in Drosophila

slow as molasses (slam) is required for germ cell migration. In slam zygotic mutants, germ cells fail to transit off the midgut into the mesoderm. slam is required in parallel to HMG Coenzyme A reductase, another germ cell migration gene. Because slam RNA and protein are expressed earlier than the time when defects are observed in germ cell migration, it is proposed that Slam is required for the localization of a signal to the basal side of blastoderm cells that is needed later in the posterior midgut to guide germ cells (Stein, 2002).

The phenotype of slamwaldo mutant embryos is strikingly similar to that of Hmgcr mutants. In both mutants, germ cells fail to move off the midgut and do not associate with SGPs in the mesoderm at stage 11-12. Some germ cells are seen correctly migrating in both single mutants, even in null alleles of Hmgcr. An instructive role as a germ cell attractant was demonstrated for Hmgcr by the finding that mis-expression of Hmgcr leads to attraction of germ cells to the ectopic site independent of SGP differentiation (Van Doren, 1998; Stein, 2002).

To analyze a possible interaction between the two genes, Hmgcr expression and function was analyzed in slam mutant embryos. Hmgcr RNA is properly expressed in slamwaldo mutants, indicating that slam is not required upstream of Hmgcr for its RNA expression. To test whether slam acts downstream of Hmgcr, the levels of Slam activity were reduced after ectopic expression of Hmgcr. It was found that ectopic Hmgcr expression is still capable of attracting germ cells in a slamwaldo heterozygous mutant background, suggesting that Slam is not required for Hmgcr-mediated germ cell attraction to the mesoderm. Together, these data make it less likely that Slam and Hmgcr act within the same pathway and favor the hypothesis that Hmgcr and slam act independently and provide separate guidance cues. This conclusion is further supported by the analysis of slamwaldo; Hmgcr double mutants. The germ cell migration phenotype in double mutant embryos was stronger than either single mutant, as no germ cells move off the midgut. However, the specificity of this defect is unclear as the double mutant embryos were poorly differentiated, a phenotype that is not found in collections from either single mutant. The fact that the double mutant has a novel phenotype suggests that, while the activity of Hmgcr and slam may not directly rely on each other’s function, the two genes may act in parallel and may regulate common downstream pathways (Stein, 2002).

The mevalonate pathway controls heart formation in Drosophila by isoprenylation of Gγ1

The early morphogenetic mechanisms involved in heart formation are evolutionarily conserved. A screen for genes that control Drosophila heart development revealed a cardiac defect in which pericardial and cardial cells dissociate, which causes loss of cardiac function and embryonic lethality. This phenotype resulted from mutations in the genes encoding HMG-CoA reductase, downstream enzymes in the mevalonate pathway, and G protein Gγ1, which is geranylgeranylated, thus representing an end point of isoprenoid biosynthesis. These findings reveal a cardial cell-autonomous requirement of Gγ1 geranylgeranylation for heart formation and suggest the involvement of the mevalonate pathway in congenital heart disease (Yi, 2006).

Mutations in genes controlling heart development frequently cause fatal cardiac malformations, the most common type of birth defect in humans. Because many of the mechanisms involved in heart development are evolutionarily conserved, the fruit fly Drosophila represents a powerful model for genetically dissecting this complex developmental process. The Drosophila heart, or dorsal vessel, which pumps bloodlike cells through an open circulatory system, is composed of parallel rows of contractile cardial cells (cardioblasts) tightly attached to pericardial cells; the latter perform supportive and secretory functions (Yi, 2006).

A P-element genetic screen was performed for Drosophila mutants with heart defecusing transgenic flies harboring a green fluorescent protein (GFP) transgene under control of the Hand enhancer, which is specific for cardial cells, pericardial cells, and the lymph gland—a hematopoietic organ in fruit flies. The Hand-GFP transgene allows visualization of the developing heart at single-cell resolution. Among a collection of mutants with cardiac abnormalities, a heart defect was observed in which pericardial cells dissociated from cardioblasts in the dorsal vessel at the end of embryogenesis. This phenotype was termed 'broken hearted' (bro). Five such mutants of different genetic loci are described in this study. In contrast to the wild-type dorsal vessel in which the pericardial cells are intimately associated with cardioblasts, in each of these mutants, the relative positions of pericardial cells and cardioblasts changed with each heartbeat (Yi, 2006).

The P element in the bro1 locus [l(3)01152] is located in the first exon of the hydroxymethyl-glutaryl (HMG)coenzyme A (CoA) reductase gene (HMGCR), which is expressed in the dorsal vessel and the gonadal mesoderm, where it is required for migration of primordial germ cells (Van Doren, 1998). Mutants trans-heterozygous for HMGCR01152 and a deficiency line Df(3R)Exel9013, in which the HMGCR gene is deleted, or two EMS mutants, HMGCRclb26.31 and HMGCRclb11.54, showed similar, but more severe, cardiac defects than homozygous HMGCR01152 mutants. Expression of HMGCR in the heart, with the use of a Hand-GAL4 driver and a UAS-HMGCR transgene, rescued the cardiac defects in the HMGCR01152 mutant (Yi, 2006).

HMGCR controls a rate-limiting step in the conversion of HMG-CoA into mevalonate, a precursor for the synthesis of cholesterol and isoprene derivatives that modify the C termini of proteins containing a CAAX motif (C, cysteine; A, aliphatic amino acid; X, any amino acid). In contrast to mammalian cells, Drosophila does not use the mevalonate pathway to synthesize cholesterol. Injection of embryos at the syncytial blastoderm stage with 0.1 µM mevinolin, a statin drug that lowers cholesterol level by inhibiting HMGCR activity, caused cardiac defects at stage 17 similar to those of the HMGCR mutants (Yi, 2006).

To investigate whether either of the two major isoprenoids, farnesyl pyrophosphate (farnesyl-PP) and geranylgeranyl pyrophosphate (geranylgeranyl-PP), might be required for heart formation, mutants were examined in the genes encoding geranylgeranyl pyrophosphate synthase (GGPPS) and geranylgeranyl transferase type I ß subunit (ßGGT-I), which act downstream of HMGCR and are required for the biosynthesis of geranylgeranyl-PP or transfer of geranylgeranyl-PP to protein, to find out whether they also cause cardiac defects. Indeed, GGPPS (also called qm) mutant embryos showed 100% penetrance for the bro phenotype, just as HMGCR mutants did, and at least 30% of the ßGGT-I mutants displayed the same phenotype. In contrast, two deficiency lines [Df(2L)Exel6010 or Df(3R)Exel6269] deleting either the farnesyl transferase α (CG2976) or ß (CG17565) subunit did not display similar cardiac defects. These findings suggested that the cardiac defects of HMGCR mutant embryos resulted from a failure of geranylgeranylation of a target substrate protein required for the adhesion between cardioblasts and pericardial cells (Yi, 2006).

Analysis of another bro mutant (bro4) suggested that the G protein γ subunit 1 (Gγ1), which contains a C-terminal CAAX motif, is the substrate of this geranylgeranylation modification required for heart formation. The P element in the bro4 locus l(2)k08017 is inserted into the splice donor site after the first exon of the Gγ1 gene. Gγ1 expression level was reduced by more than 50% in homozygous l(2)k08017 embryos, which suggested that l(2)k08017 is a hypomorphic mutant allele of the Gγ1 gene. Mutants trans-heterozygous for the l(2)k08017 insertion and a deficiency that deletes the Gγ1 gene [Df(2R)H3E1] or for Df(2R)H3E1 and a Gγ1 null allele showed the same cardiac defects as the homozygous l(2)k08017 embryos. Double mutants of the hypomorphic HMGCR and Gγ1 alleles showed a more severe cardiac defect than either single mutant. A fifth bro mutation was mapped to the Sar1 gene, which encodes a guanosine triphosphatase that controls budding of vesicles overlaid with coat protein complex II (COPII) from the endoplasmic reticulum (ER) to the Golgi network (Yi, 2006).

The developmental onset of cardiac defects was identical in the HMGCR, Gγ1, GGPPS/qm, ßGGT-I, and Sar1 mutants. Cardioblasts and pericardial cells were properly specified and aligned until stage 16. However, at stage 17, pericardial cells began to dissociate from the dorsal vessel. These observations suggest that these genes are required to maintain cardiac integrity. The phenotypes of the different mutants were also comparable, except for the two HMGCR EMS mutants or the HMGCR01152/Df(3R)Exel9013 mutant, which was more severe and showed distortion of the shape of the dorsal vessel (Yi, 2006).

The final C-terminal residues of all G protein γ subunits contain a CAAX motif in which the variable amino acid X determines the type of lipid modification: If X is serine, methionine, alanine, or glutamine, the cysteine is modified by farnesylation, whereas if X is leucine or valine, it is modified by geranylgeranylation. Using an in vitro prenylation assay, it was found that Drosophila Gγ1 protein, which contains a CAAX motif of Cys-Thr-Val-Leu (CTVL), was modified by geranylgeranylation, but not by farnesylation, in agreement with the requirement of GGPPS/qm and ßGGT-I for cardiac development (Yi, 2006).

To determine directly if geranylgeranylation of Gγ1 is essential for heart development, whether wild-type and mutant forms of Gγ1 protein could rescue the cardiac defect of the Gγ1 mutant was tested. Targeted expression of wild-type Gγ1 in the heart was sufficient to rescue the cardiac defects of Gγ1 mutants, whereas mutant forms of Gγ1, in which geranylgeranylation was abolished by either a substitution of Ser for Cys67 (Gγ1-C67S) in the CAAX box or a deletion of the CAAX box (Gγ1-δCAAX), failed to rescue the cardiac defects in Gγ1 mutants. It is concluded that geranylgeranylation of the CAAX motif of Gγ1 is required for its normal activity during Drosophila heart formation (Yi, 2006).

Lipid modification of the CAAX motif facilitates the association of proteins with membranes. To further explore how geranylgeranylation of Gγ1 affects its biological function, the subcellular localization of the Gγ1 protein was examined in Drosophila S2R+ cells. Wild-type Gγ1 protein was always excluded from the nucleus in S2R+ cells, whereas the two mutant forms of Gγ1, which were not geranylgeranylated, were located throughout the cytoplasm and nucleus. Because Gγ1 is a small protein and can enter the nucleus freely, the specific localization of wild-type Gγ1 protein to the cytoplasm likely reflects its interaction with membranous structures, which requires modification by geranylgeranylation (Yi, 2006).

In S2R+ cells treated with three HMGCR inhibitors (atorvastatin, mevinolin, and simvastatin), as well as HMGCR double-stranded RNA, the wild-type Gγ1 protein displayed the same abnormal subcellular distribution as the two mutant forms of Gγ1. These findings suggest that abnormal subcellular localization of Gγ1 accounts for the cardiac defects in the mevalonate pathway mutants and Gγ1 mutants. Gα has also been shown to be required at an earlier stage of heart development for proper alignment of cardioblasts (Fremion, 1999), which is distinct from the function of Gγ1 revealed here (Yi, 2006).

Cardiac defects of HMGCR or Gγ1 mutants could be completely rescued by targeted expression of UAS-HMGCR and UAS-Gγ1 transgenes, respectively, using a Hand-GAL4 driver, which directs expression in both cardioblasts and pericardial cells, or a Mef2-GAL4 driver, which is expressed in cardioblasts but not in pericardial cells. In contrast, targeted expression of HMGCR or Gγ1 using Dot-GAL4, which drives expression only in pericardial cells, failed to rescue the cardiac defects in either mutant. These results demonstrate that HMGCR and Gγ1 function specifically in cardioblasts to adhere with pericardial cells and exclude the possibility that the bro cardiac phenotype arises secondarily from general metabolic abnormalities (Yi, 2006).

HMGCR and downstream enzymes in the biochemical pathway leading to the synthesis of geranylgeranyl-PP are specifically required in cardioblasts to modify Gγ1. It is proposed that geranylgeranylation, which is required for the proper intracellular localization of Gγ1, is in turn required for generating a signal for pericardial cells to adhere to cardioblasts throughout heart formation. Indeed, Gßγ has been shown to control Golgi apparatus organization and vesicle formation during exocytosis in mammalian cells. The finding that a mutation in Sar1 causes the same cardiac phenotype as the Gγ1 mutation further supports the possibility that this collection of mutations perturbs the secretion of a factor required for maintenance of cardiac integrity. Inhibition of this pathway with statins results in cardiac defects similar to those resulting from mutations in HMGCR and downstream genes required for isoprenoid biosynthesis, which raises the possibility that congenital heart defects reportedly associated with the use of statins, which are contraindicated during pregnancy, may reflect perturbation in a similar developmental pathway (Yi, 2006).

HMGCR has also been shown to be required for recruitment of primordial germ cells (PGCs) to the gonad in Drosophila, but the protein target(s) of the mevalonate pathway that mediate this process have not been identified. Perhaps Gγ1 functions in the gonad mesoderm to guide PGC migration. It is speculated that lipid modifications mediated by the mevalonate pathway contribute to directed cell migration and subsequent cell-cell adhesion in diverse cell types. Given the conservation of cardiac developmental control mechanisms, it will be of interest to investigate the potential involvement of the mevalonate pathway in mammalian heart development and congenital heart disease (Yi, 2006).


Reference names in red indicate recommended papers.

Clark, I. B., Jarman, A. P. and Finnegan, D. J. (2007). Live imaging of Drosophila gonad formation reveals roles for Six4 in regulating germline and somatic cell migration. BMC Dev. Biol. 7: 52. PubMed citation; Online text

Deshpande, G., Swanhart, L., Chiang, P. and Schedl, P. D. (2001). Hedgehog signaling in germ cell migration. Cell 106: 759-769. 11572781

Deshpande, G. and Schedl, P. (2005). HMGCoA reductase potentiates hedgehog signaling in Drosophila melanogaster. Dev. Cell 9(5): 629-38. 16256738

Deshpande, G., Godishala, A. and Schedl, P. (2009). Gγ1, a downstream target for the hmgcr-isoprenoid biosynthetic pathway, is required for releasing the Hedgehog ligand and directing germ cell migration. PLoS Genet 5: e1000333. PubMed ID: 19132091

Deshpande, G., Zhou, K., Wan, J. Y., Friedrich, J., Jourjine, N., Smith, D. and Schedl, P. (2013). The hedgehog Pathway Gene shifted Functions together with the hmgcr-Dependent Isoprenoid Biosynthetic Pathway to Orchestrate Germ Cell Migration. PLoS Genet 9: e1003720. PubMed ID: 24068944gγ1

Felsenfeld, A. L. and Kennison, J. A. (1995). Positional signaling by hedgehog in Drosophila imaginal disc development. Development 121: 1-10. 7867491

Flury, I., et al. (2005). INSIG: a broadly conserved transmembrane chaperone for sterol-sensing domain proteins. EMBO J. 24(22): 3917-26. 16270032

Gertler, F. B., Chiu, C. Y., Richter-Mann, L. and Chin, D. J. (1988). Developmental and metabolic regulation of the Drosophila melanogaster 3-hydroxy-3-methylglutaryl coenzyme A reductase. Mol. Cell. Biol. 8(7): 2713-21. 3136321

Rawson, R. B. (2003). The SREBP pathway-insights from Insigs and insects. Nat. Rev. Mol. Cell Biol. 4: 631-640. 12923525

Santos, A. C. and Lehmann, R. (2004). Isoprenoids control germ cell migration downstream of HMGCoA reductase. Dev. Cell. 6(2): 283-93. 14960281

Seegmiller, A. C., Dobrosotskaya, I., Goldstein, J. L., Ho, Y. K., Brown, M. S. and Rawson, R. B. (2002). The SREBP pathway in Drosophila: regulation by palmitate, not sterols. Dev. Cell 2: 229-238. 12923525

Stein, J. A., Broihier, H. T., Moore, L. A. and Lehmann, R. (2002). Slow as Molasses is required for polarized membrane growth and germ cell migration in Drosophila. Development 129: 3925-3934. 12135929

Theopold, U., Ekengren, S., Hultmark, D. (1996). HLH106, a Drosophila transcription factor with similarity to the vertebrate sterol responsive element binding protein. Proc. Natl. Acad. Sci. 93(3): 1195-1199. 8577739

Thorpe, J. L., Doitsidou, M., Ho, S. Y., Raz, E., Farber, S. A. (2004). Germ cell migration in zebrafish is dependent on HMGCoA reductase activity and prenylation. Dev. Cell 6(2): 295-302. 14960282

Van Doren, M., Broihier, H. T., Moore, L. A. and Lehmann, R. (1998). HMG-CoA reductase guides migrating primordial germ cells. Nature 396: 466-469. 9853754

Yi, P., Han, Z., Li, X. and Olson, E. N. (2006). The mevalonate pathway controls heart formation in Drosophila by isoprenylation of Gγ1. Science 313: 1301-1303. PubMed Citation: 16857902

HMG Coenzyme A reductase : Biological Overview | Evolutionary Homologs | Developmental Biology | Effects of Mutation | References

date revised: 10 June 2009

Home page: The Interactive Fly © 2006 Thomas Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.