Chromatin signatures of the Drosophila replication program

DNA replication initiates from thousands of start sites throughout the Drosophila genome and must be coordinated with other ongoing nuclear processes such as transcription to ensure genetic and epigenetic inheritance. Considerable progress has been made toward understanding how chromatin modifications regulate the transcription program; in contrast, relatively little is known about the role of the chromatin landscape in defining how start sites of DNA replication are selected and regulated. This study describes the Drosophila replication program in the context of the chromatin and transcription landscape for multiple cell lines using data generated by the modENCODE consortium. While the cell lines exhibit similar replication programs, there are numerous cell line-specific differences that correlate with changes in the chromatin architecture. Chromatin features were identified that are associated with replication timing, early origin usage, and ORC binding. Primary sequence, activating chromatin marks, and DNA-binding proteins (including chromatin remodelers) contribute in an additive manner to specify ORC-binding sites. Accurate and predictive models were generated from the chromatin data to describe origin usage and strength between cell lines. Multiple activating chromatin modifications contribute to the function and relative strength of replication origins, suggesting that the chromatin environment does not regulate origins of replication as a simple binary switch, but rather acts as a tunable rheostat to regulate replication initiation events (Eaton, 2011).

The chromatin landscape clearly impacts both the expression and the replication of the genome. For example, the transcriptionally active euchromatin typically replicates prior to the repressed heterochromatic sequences. Studies in yeast, Drosophila, and mammalian systems have shown that changes in histone acetylation are associated with changes in the replication program. However, a comprehensive view of the replication program in the context of chromatin modifications and DNA-binding proteins is lacking (Eaton, 2011).

The different modENCODE data types across multiple cell lines (The modENCODE Consortium 2010) allowed the definition of the chromatin and transcription landscape associated with features of the DNA replication program. For each replication data type (replication timing, early origins, and ORC binding), a 43 × 3 matrix was generated, with each column representing a specific cell line and each row representing the enrichment or correlations with chromatin marks, DNA-binding proteins, nucleosome density, histone variants, nucleosome turnover (CATCH-IT), and gene expression (RNA-seq). For the replication timing profiles where there are no discrete peak calls, the Spearman's correlation was calculated between each factor with the whole-genome replication timing profile. For early origins of replication and ORC-binding sites, the median log2 enrichment was calculated of each factor within all BrdU peaks and within 500 bp of ORC ChIP-seq peak centers, respectively (Eaton, 2011).

The selection and regulation of DNA replication origins was found to be associated with distinct sets of chromatin marks and DNA-binding proteins. Prior studies have associated early replication with active transcription and the presence of 'activating' chromatin modifications such as histone acetylation, whereas late replication is associated with 'repressive' chromatin marks such as those found in the heterochromatin. Indeed, this study found that gene expression is positively correlated with replication timing, as are generally euchromatic marks such as H3K4me1 and H3K18ac. In contrast, heterochromatic marks such as H3K27me3 and H3K9me2 are negatively correlated with replication timing. The sequences surrounding early origins were also enriched for activating chromatin marks as well as specific DNA-binding proteins, including chromatin remodeling factors (Eaton, 2011).

Because many of the ORC-binding sites colocalized with promoters of active genes, the ORC-binding sites were separated into those that are TSS proximal (within 1 kb of a TSS) and those that were not at a TSS (distal). Particular interest was placed on chromatin features that are shared between ORC-binding sites both proximal and distal to promoters. Additionally, marks that are specific to ORC sites distal from a promoter will be of interest, as these marks may be required for ORC binding or function in the absence of a promoter (Eaton, 2011).

ORC-binding sites proximal to TSSs were enriched for chromatin remodelers such as the NURF complex (NURF301 [also known as E(BX)], ISWI) as well as other DNA-binding proteins such as GAF, RNA Pol II, and CHRO. These TSS-associated ORC sites were also enriched for H3K9ac, H3K27ac, H3K4me2, and H3K4me3 -- marks frequently found at promoters. Interestingly, those ORC sites that did not overlap with a TSS (distal) were also enriched for chromatin remodelers ISWI and NURF301, as well as GAF, which has also been implicated in chromatin remodeling. Consistent with the idea of ORC localizing to dynamic and active chromatin, an enrichment was found for CATCH-IT and H3.3 at ORC sites both proximal and distal to TSSs, as well as a reduction in bulk nucleosome occupancy. ORC sites not located at promoters were enriched for many of the same histone marks as those at promoters, with a few notable exceptions. A decrease in H3K4me3 was found at ORC sites distal from a promoter, as well as an increase in H3K18ac and H3K4me1 (Eaton, 2011).

Chromatin features specific to transcription start sites such as RNA Pol II and H2Av were decreased at ORC-binding sites distal to promoter elements. A small amount of RNA Pol II signal remained in the TSS distal ORC-binding sites; however, in comparison to the local enrichment of ISWI and GAF, there was a clear reduction in local signal. The remaining signal may be due to unannotated transcription start sites (Eaton, 2011).

Chromatin marks that are associated with active transcription through gene bodies (e.g., H3K79me1, H3K36me1, and H3K36me3) were not found above background levels at any ORC-binding sites. However, H3K36me1 was found specifically flanking those ORC-binding sites that did not coincide with a TSS. ORC has been shown to facilitate the formation of heterochromatin and HP1 binding; however, ORC sites were depleted for heterochromatic histone modifications such as H3K27me3 and H3K9me2/3 and were only slightly enriched for HP1. This may be due, in part, to the inability to map distinct ORC-binding sites in repetitive sequences, a current limitation of high-throughput sequencing approaches (Eaton, 2011).

The chromatin signatures were examined of promoter elements with and without ORC associated to determine whether there were unique chromatin signatures specific for ORC associated promoters. Since those promoters with proximal ORC binding tend to be far more actively transcribed than those without ORC, the comparison was limited to active promoter elements only. It was found that ORC-associated promoters had modestly increased chromatin remodeling activities, decreased nucleosome occupancy, and greater evidence of nucleosome turn-over relative to other active promoters not associated with ORC. In summary, these results indicate that dynamic chromatin environments may contribute to ORC localization and the subsequent activation of replication origins (Eaton, 2011).

Integrative analysis of gene amplification in Drosophila follicle cells: parameters of origin activation and repression

In metazoans, how replication origins are specified and subsequently activated is not well understood. Drosophila amplicons in follicle cells (DAFCs) are genomic regions that undergo rereplication to increase DNA copy number. All DAFCs were identified by comparative genomic hybridization, uncovering two new amplicons in addition to four known previously. The complete identification of all DAFCs enabled investigation of these in vivo replicons with respect to parameters of transcription, localization of the origin recognition complex (ORC), and histone acetylation, yielding important insights into gene amplification as a metazoan replication model. Significantly, ORC is bound across domains spanning 10 or more kilobases at the DAFC rather than at a specific site. Additionally, ORC is bound at many regions that do not undergo amplification, and, in contrast to cell culture, these regions do not correlate with high gene expression. As a developmental strategy, gene amplification is not the predominant means of achieving high expression levels, even in cells capable of amplification. Intriguingly, it was found that, in some strains, a new amplicon, DAFC-22B, does not amplify, a consequence of distant repression of ORC binding and origin activation. This repression is alleviated when a fragment containing the origin is placed in different genomic contexts (Kim, 2011; full text of article).

Protein Interactions

In Drosophila, four unknown proteins from embryonic extracts copurify (and cosediment) with DmORC2 and DmORC5, suggesting that these form DmORC (Gossen, 1995). One 79 kDa protein component is similar to the molecular weight of Lat. Immunoprecipitation of DmORC2 from Schneider cells results in specific coimmunoprecipitation of Lat. Immunoprecipitation of Lat with anti-Lat antibodies also coimmunoprecipitates DmORC2. The sequence similarity with ScORC3, the association of Lat with DmORC2, and the cell proliferation defects of lat null mutants strongly argue that Lat functions as a subunit of Drosophila ORC (Pinto, 1999).

The origin recognition complex (ORC) is the DNA replication initiator protein in eukaryotes. A functional recombinant Drosophila ORC has been reconstituted and activities of the wild-type and several mutant ORC variants have been compared. Drosophila ORC is an ATPase, and the ORC1 subunit is essential for ATP hydrolysis and for ATP-dependent DNA binding. Moreover, DNA binding by ORC reduces its ATP hydrolysis activity. In vitro, ORC binds to chromatin in an ATP-dependent manner, and this process depends on the functional AAA+ nucleotide-binding domain of ORC1. Mutations in the ATP-binding domain of ORC1 are unable to support cell-free DNA replication. However, mutations in the putative ATP-binding domain of either the ORC4 or ORC5 subunits do not affect either of these functions. Evidence is provided that the Drosophila ORC6 subunit is directly required for all of these activities and that a large pool of ORC6 is present in the cytoplasm, cytologically proximal to the cell membrane. Studies reported here provide the first functional dissection of a metazoan initiator and highlight the basic conserved and divergent features between Drosophila and budding yeast ORC complexes (Chesnokov, 2001).

Six different mutant complexes and wild-type recombinant ORC were prepared. For each case, simultaneous expression of the wild-type or mutant genes in a baculovirus expression system resulted in complexes that could be purified to homogeneity through four chromatographic steps, and the mutant complexes assembled and exhibited no chromatographic differences during the purification. In a final step, the pooled peak fractions were subjected to glycerol-gradient sedimentation (Chesnokov, 2001).

The best understood functions of the yeast ORC are its DNA-binding and ATP hydrolysis functions. The bulk of recombinant (or purified embryonic) Drosophila ORC DNA binding activity is nonspecific and ATP-independent. However, this ATP-independent DNA binding activity can be titrated away with sufficient amount of carrier DNA when the carrier DNA is in a range 50-100 molar excess to the probe DNA. At physiologically relevant ATP concentrations (10 microM to 1 mM), the wild-type ORC binds to DNA 10-50-fold better than either the ORC1A or ORC1B mutant complex. Mutations in either the Walker A or B motif of ORC4 or the Walker A motif of ORC5 have no effect on the formation of ATP-dependent DNA-protein complex. These experiments supports the idea that the recombinant Drosophila ORC, like the recombinant S. cerevisiae homolog, requires only the ORC1 component of the complex to bind ATP for tight DNA interactions. However, the complex missing the ORC6 subunit does not form an ATP-dependent DNA-protein complex (Chesnokov, 2001).

Kinetic analysis of ATP hydrolysis with multiple independent wild-type (wt) ORC preparations shows a Km of 1.92 µM and a Vmax of 0.4 mol ATP hydrolyzed per min per mol of complex. Binding to DNA has a small (2-fold) but measurable effect on slowing the rate of ATP hydrolysis by ORC. In these experiments, ATP was not limiting, the mutant ORI complexed to DNA was titrated to its maximal effect. In the absence of any carrier DNA, the saturation is reached at an approximate 2.5-fold molar excess of DNA to ORC. Complexes harboring similar mutations in either ORC4 or ORC5 hydrolyzes ATP with equivalent kinetics to wild type, all displaying Km values and Vmax within the experimental error range of wild type. Consistent with the DNA-binding experiments, the ATP-hydrolysis rate for these mutant complexes is slowed by DNA similar to the effect observed for the wild-type ORC. In contrast, ORC1A or ORC1B mutants have severely crippled enzymatic activity, too close to background to measure any kinetic parameters. The ORC-6 complex is able to hydrolyze ATP at reduced levels, but this activity is unaffected by DNA, consistent with the finding that ORC6 is critical for formation of an ATP-dependent ternary complex (Chesnokov, 2001).

Chromatin binding assays were performed by using both mutant and wt ORC in extracts depleted of membranes. For these experiments Drosophila preblastula embryo extracts were immunodepleted of ORC by using antibody raised against ORC2 and ORC6. The effectiveness of immunodepletion was verified by immunoblotting. Demembranated sperm chromatin was added to the depleted extracts, and the binding activities of mutant and wild-type recombinant DmORC were compared with the endogenous Drosophila ORC. Treatment of the extracts with Apyrase abolishes ORC-chromatin binding, thus it is inferred that the binding process requires ATP. Endogenous ATP levels (which are estimated to be at 30-50 µM) were relied upon to mediate tight chromatin binding. Proteins associated with the chromosomes are separated from the unbound proteins by sedimentation. The results obtained via this assay parallel those obtained by the gel-shift experiments. Recombinant wt ORC, ORC4A, ORC4B, and ORC5A complexes associate with the chromatin with apparently the same efficiency as does endogenous protein, whereas the ORC1A, ORC1B, and ORC 6 complexes are severely crippled (Chesnokov, 2001).

Two independent measures of DNA replication competence were used for accessing the abilities of the mutant complexes to restore activity to depleted extracts. In the first assay, labeled precursor incorporation into high molecular DNA was detected by autoradiography of gels after electrophoresis or in a second assay after CsCl density gradient separation of DNA that was replicated in extracts with the density label precursor BrdUrd. As anticipated from the DNA and chromatin binding results, the ORC1A, ORC1B, and ORCdelta6 complexes were essentially inactive by at least 10-20-fold below the activity of wt recombinant ORC in restoring replication to the extracts. The ORC4A, ORC4B, and ORC5A mutants were effective in reconstitution but were in multiple experiments between 50% and 90% of wild-type complex (Chesnokov, 2001).

It has been concluded that the bulk of the subunits of the Drosophila ORC biochemically behave as a complex. ORC2 antibodies were used to track ORC in fractions from 0-12-h embryo extracts after gel-filtration chromatography. Two broad zones containing ORC were found. The highest apparent molecular weight fractions containing all ORC subunits were pooled and purified. A smaller complex was also detected that was apparently without ORC-1. However, when following ORC6 using ORC6-specific antibodies, a pool of ORC6 devoid of other ORC subunits is detected. No other ORC subunits were found in a form unassociated with other ORC proteins. It is estimated that this free pool is at least one-half of the total ORC6 protein present in these extracts. Given the important role that Drosophila ORC6 plays in cell-free replication and the other activities of ORC, it was of interest to ask whether this separate pool of ORC6 is localized with the other ORC subunits in the cell (Chesnokov, 2001).

Transient ectopic expression of ORC1 or ORC2 GFP-fusion proteins in cultured cells shows a distinct nuclear localization; in unexpected contrast, the GFP-ORC6 fusion protein was found both in the nucleus and cytoplasm. The ORC6 cytoplasmic signal seems to be closely associated, in various focal planes, with the cytoplasmic membranes. These experiments rely on overexpression: this issue was probed further by direct immunofluorescence of endogenous levels of the ORC proteins in Drosophila embryos. Before the onset of cellularization, ORC6 protein localizes only with ORC2 in the nuclear space of both interphase and mitotic cells. However, after cellularization, ORC6 seems to localize in the cytoplasm and nucleus. The signals for ORC6 can be blocked by preincubating the affinity-purified antibodies with recombinant ORC6 proteins and are clearly distinct from the ORC2 pattern. Further work will be required to judge whether the cytoplasmic pool of ORC6 is truly membrane associated, but it is worth noting that the carboxyl terminus of Drosophila ORC6 contains a predicted leucine-zipper region that could be involved in mediating multiple heterologous protein-protein interactions (Chesnokov, 2001).

An important finding of this study is that the Drosophila ORC complex likely uses mechanisms for binding DNA that are similar to those reported for the budding yeast homolog. Of the three potential ATP binding proteins in ORC, only ORC1 seems to be critical for establishing a tight ternary complex with DNA and for binding to chromatin. Similarly only mutations in the ATP binding domains of ORC1 critically affect a single round of DNA replication in cell-free extracts. Additional experimentation needs to be done to test the roles of the conserved domains in ORC4 and ORC5. Particularly intriguing is the wide conservation of the GKT (Walker A motif) and D (D/EE) (Walker B motif) in the ORC4 subunit. Such domains may be critical for recycling ORC for subsequent rounds of replication or for other activities of the complex in heterochromatin formation or putative check-point control. Drosophila ORC is an ATPase, and again ORC1 seems to play the critical role for ATP hydrolysis, since mutants in the putative ATPase domains of ORC4 and ORC5 do not affect the kinetic parameters of the mutant complex. Nevertheless, it is possible that in the presence of other bound factors, ATP binding or hydrolysis by the other subunits plays some critical role. ATP hydrolysis by any subunit does not seem important for DNA-binding activity. ADP could not mediate such a DNA-protein complex, and ATPgammaS is better at forming a ternary complex than ATP. X-ray crystallographic structure models for several AAA+ proteins have been solved, and a common fold has been observed. The crystal structure model of an archael Cdc6 ortholog was used as a guide for the ATP-binding structures of ORC1. In the nucleotide-binding domain of this protein family, both the GKT and the DE motifs contribute to nucleotide affinity. In fact, similar mutants in the amino-part of the Walker B motif of the S. cerevisiae ORC1 are defective for ATP binding, in contrast to mutations at the carboxyl end of the B motif that are competent for such activity. Moreover, the solvent-exposed surfaces present in these parts of the ORC1 protein may influence interactions with other partners, yielding a mutant complex with altered functions. These studies of the ATPase activity of DmORC indicate that turnover is slower when ORC is bound to DNA, but the effect is significantly less than that observed for the budding yeast complex. Divergence in the way in which these proteins interact with DNA is also highlighted by the critical role that the Drosophila ORC6 protein plays in ATP-dependent DNA binding. Perhaps, given the lack of amino acid homologies found between the ScORC6 and DmORC6 proteins, it is dangerous to consider each to be homologs (Chesnokov, 2001).

Overexpression of ORC1 trans-genes in Drosophila can alter DNA replication patterns. This overexpression leads to detectable levels of BrdUrd incorporation in normally quiescent cells or increased levels of replication in follicle cells normally amplifying the chorion genes. Similar ectopic expression of an ORC1A mutant (ORC1K604E) has no phenotype. The biochemical results with the ORC1A mutant K604A predict that their mutation might have a dominant negative effect on DNA replication in vivo. It is possible that the mutant gene would not be antimorphic by virtue of its not being able to compete with a wild-type ORC1 protein for assembly into complex. Leaving this point aside, one idea favored is that ORC1 is limiting for replication in some cellular environments and, for example, complexes containing solely ORC2-6 wait for ORC1 for activation. These pools may or may not be bound to chromosomal DNA. Recent work in mammalian systems indicates that ORC1 may be more loosely associated with chromatin than is ORC2. ORC2, presumably with some of the subunits, can be pelleted with the chromosomes. The results reveal that intact ORC needs ATP and functional ORC1 to bind tightly to chromatin. Are all of these data compatible, assuming a conservation in basic binding properties for ORC between mammals and Drosophila? Perhaps, in the absence of ORC1 other subunits mediate another sort of chromatin association. More complex notions are possible, including the interaction of unknown chromatin binding proteins that serve to tether a complex lacking ORC1 to the ori sites (Chesnokov, 2001).

It is suggested that ORC6 is another subunit that may play important and perhaps dynamic roles in regulating replication activity. The data show that ORC6 is an essential component of the complex per se and may be directly involved in DNA binding and other replication functions or needed for proper ORC assembly. In H. sapiens extracts, ORC6 is not found associated with other ORC subunits, but when expressed in the baculovirus system with the other ORC genes, the protein does join a six-subunit complex. The high levels of free ORC6 in embryonic and cultured cell extracts is intriguing. A considerable fraction of this pool as judged by cytological methods is cytoplasmic, and the protein is perhaps associated with or proximal to the cytoplasmic membranes. It is possible that this localization enables ORC6 to participate in functions unrelated to DNA replication per se, as has been suggested for the 'latheo' gene product, which is ORC3. Latheo seems to be involved in ion transport at neuromuscular junctions. Data now exist for both the budding yeast and for the Drosophila ORC, which directly indicate that all of the subunits are critical for DNA replication function, and complex models involving traffic of subsets of ORC subunits can be the subject of future work (Chesnokov, 2001).



A polyclonal antiserum was used to assay Lat protein expression in the larval peripheral nervous system and body muscle. The antiserum is specific for the LAT protein in Western blots of third instar larval CNS extracts, recognizing a single 79 kDa protein from wild-type larvae that is not detectable in lat null mutant larvae. In wild-type third instar larvae, the Lat protein is immunologically detected at synaptic boutons at most NMJs on a wide variety of muscle fibers in the ventral abdomen. LAT immunostaining is often only weakly detectable at NMJs containing predominantly large type I boutons, which utilize glutamate as the primary neurotransmitter. Staining is nevertheless clearly evident at type I boutons, including both subtypes Ib and Is at various NMJs, including muscles 4 and 6/7. At many NMJs, LAT immunostaining is most distinct at morphologically smaller boutons, including those resembling type II and type III boutons. These boutons are believed to contain amines and neuropeptides, including octopamine (Monastirioti, 1995) and neuropeptide proctolin (Arg-Tyr-Leu-Pro-Thr) found in type II synapses (Anderson, 1988) and insulin-like peptide, found in type III synapses (Gorczyca, 1993). The amines and neuropeptides may serve as modulatory transmitters. In particular, the NMJ at muscle 12, which receives types I, II, and III innervation, consistently exhibits positive staining, commonly at multiple bouton types at the same NMJ. LAT thus appears not to be segregated to particular body segments, subsets of muscle fibers, or particular bouton subtypes distinguishable by either morphological or physiological criteria. LAT immunoreactivity is also detectable at the NMJ in second instar larvae, indicating the protein is present throughout most of the period of dramatic morphological and functional synaptic maturation (Rohrbough, 1999).

To gain more specific information on LAT synaptic localization, double-staining experiments were carried out at the wild-type NMJ with antibodies against LAT and other known pre- or postsynaptic proteins. Confocal microscopy of immunofluorescence reveals colocalization of LAT with the presynaptic vesicle-associated cysteine string protein (CSP) (Zinsmaier, 1994) at multiple bouton types, including both types I and II, suggesting LAT is located predominantly presynaptically and colocalizes with synaptic vesicles. LAT and CSP immunostaining often show nearly complete overlap at smaller boutons resembling type II and type III boutons, while at type I boutons, LAT staining usually appears less distinct and contained within a subarea of CSP expression. LAT localization was also examined in preparations double stained for postsynaptic glutamate receptors (DGluR2a) and the postsynaptic Discs-large (Dlg) and position-specific ß (ßPS) integrin proteins, which are localized in the subsynaptic reticulum of type I boutons. LAT staining typically occupies smaller areas contained within the broader staining pattern of the respective postsynaptic proteins, including the central area of boutons. These results confirm that LAT is localized at type I boutons and strongly suggest that the protein is expressed predominantly, if not exclusively, in the presynaptic compartment (Rohrbough, 1999).

Effects of Mutation or Deletion

Genetic dissection of learning and memory in Drosophila has been limited by the existence of ethyl methanesulfonate (EMS)-induced mutations in only a small number of X-linked genes. To remedy this shortcoming, a P element mutagenesis has been carried out to screen for autosomal mutations that disrupt associative learning and/or memory. The generation of 'P-tagged' mutant alleles will expedite molecular cloning of these new genes. A behavior-genetic characterization of latheoP1, a recessive, hypomorphic mutation of an essential gene, has been carried out. latheoP1 flies perform poorly in olfactory avoidance conditioning experiments. This performance deficit could not be attributed to abnormal olfactory acuity or shock reactivity, two task-relevant 'peripheral' behaviors used during classical conditioning. Thus, the latheoP1 mutation appears to affect learning/memory specifically. Consistent with chromosomal in situ localization of the P element insertion, deficiencies of the 49F region of the second chromosome fail to complement the behavioral effect of the latheoP1 mutation. Further complementation analyses between latheoP1 and lethal alleles, produced by excision of the latheoP1 insert or by EMS or gamma-rays, in the 49F region maps the latheo mutation to one vital complementation group. Flies heterozygous for latheoP1 and one of two EMS lethal alleles or one lethal excision allele also show the behavioral deficits, thereby demonstrating that the behavioral and lethal phenotypes co-map to the same locus (Boynton, 1992).

Histological analyses of late third instar larvae reveal the absence of all imaginal discs and an undersized CNS in lat mutants. In contrast, the gross morphology of imaginal discs and CNS appear normal in second instar mutants. During the pupal stage in Drosophila, larval structures outside of the CNS are histolyzed, and adult structures are generated from proliferating clusters of cells in imaginal discs. Hence, degeneration of imaginal discs in late third instar larvae likely produces the pupal lethality associated with these lat mutants. These morphological defects are accompanied by an apparent defect in DNA replication. Proliferating cells in normal and mutant lat larvae CNSs were examined by labeling their chromosomes with bromo deoxyuridine (BrdU) during DNA replication. In late third instar larvae (96 hr posthatching), cell proliferation in a normal CNS is maximal. Lateral portions of the brain hemispheres, where cells of the optic lobes are proliferating, show particularly high levels of BrdU incorporation. In contrast to this high level of cell proliferation occurring in normal larvae, virtually no BrdU incorporation can be seen in the CNS of homozygous lethal lat mutants. Consistent with this observation, the mutant CNS appears smaller at this late stage of larval development (Pinto, 1999).

Given the defect in neuroblast (Nb) proliferation in the CNS of pupal lethal lat larvae, the gross morphology of several anatomical regions of the adult brain in homozygous viable lat mutants was evaluated. Homozygous latP1 mutants (which display defective olfactory associative learning but normal sensorimotor responses) show a 20% reduction only in mushroom body neuropillar volume, while flies hemizygous for latP1 and a chromosomal deficiency (Df) (which display performance deficits in olfactory associative learning and in sensorimotor response) show an 80% reduction in mushroom bodies and a 20% reduction in central complex. These observations suggest a development etiology for the learning defect of adult viable latP1 mutants (aberrant mushroom bodies) and for the more severe performance deficits of latP1/Df hemizygotes (aberrant mushroom bodies and central complex (Pinto, 1999).

Morphological and functional examinations of three homozygous lethal lat genotypes were examined. Homozygous lat- animals of all three genotypes develop relatively normally through the first and second instars and exhibit coordinated locomotion similar to wild-type larvae. Thereafter, their movement and feeding behavior, though functional, becomes progressively less vigorous. Many mutant larvae grow to normal size but usually remain in the food past the normal wandering stage (5-6 days after egg laying [AEL]) and delay pupation for several days beyond the normal period. Lethality in lat- larvae and early pupae eventually results from a loss of postembryonic cell division, leading to a complete absence of CNS cell proliferation and imaginal development in late third instar (Boynton, 1993; Pinto, 1999). All three lethal lat- alleles appear to be strong protein hypomorphs or nulls, based on the near or complete absence of detectable LAT protein immunostaining at lat- NMJs in immunohistological assays (Rohrbough, 1999).

The neuromuscular morphology of lat- mutant larvae, including the stereotypic muscle and innervation patterns, appears largely normal. Mutant NMJs are present at the normal synaptic locations and exhibit morphological terminal elaborations, similar to wild-type NMJs. Presynaptic boutons at lat- NMJs are normal in size and at the light microscope level appear to possess a normal level of transmitter vesicle proteins, such as CSP. Alterations in synaptic terminal morphology have been described for other Drosophila learning and memory mutants. The number of terminal branches and synaptic boutons at lat- NMJs were examined, using anti-CSP to visualize presynaptic terminals. The terminal branching pattern at the muscle 12 NMJ, which receives multiple innervation and forms boutons of three to four subtypes, is similar to normal for mutant larvae. However, both lat- mutant strains have about 20% fewer terminal branches than do wild-type terminals, due to fewer higher order branch segments at mutant terminals. Consistent with reduced terminal branching, lat- NMJs also have about 20% fewer synaptic boutons than do wild-type NMJs at both muscle 12 and muscle 6/7. Similar but statistically insignificant decreases are observed at lat- muscle 12 and 6/7 NMJs, which show more variability in bouton number. However, parallel morphological changes are not observed at the muscle 4 NMJ, where lat- NMJs have a statistically insignificant increased bouton number compared to normal In summary, lat- mutant NMJs exhibit only mild alterations in synaptic morphology. At the more complex muscle 12 NMJ, mutant terminal complexity appears to be slightly reduced on the basis of terminal branching and bouton number. In contrast, at the simpler muscle 4 NMJ, mutant terminals have normal or slightly increased bouton numbers and terminal complexity. These morphological differences are completely insufficient to account for the observed functional alterations in transmission at mutant synapses (Rohrbough, 1999).

Place memory retention in Drosophila

Some memories last longer than others, with some lasting a lifetime. Using several approaches memory phases have been identified. How are these different phases encoded, and do these different phases have similar temporal properties across learning situations? Place memory in Drosophila using the heat-box provides an excellent opportunity to examine the commonalities of genetically-defined memory phases across learning contexts. This study determines optimal conditions to test place memories that last up to three hours. An aversive temperature of 41°C was identified as critical for establishing a long-lasting place memory. Interestingly, adding an intermittent-training protocol only slightly increased place memory when intermediate aversive temperatures were used, and slightly extended the stability of a memory. Genetic analysis of this memory identified four genes as critical for place memory within minutes of training. The role of the rutabaga type I adenylyl cyclase was confirmed, and the latheo Orc3 origin of recognition complex component, the novel gene encoded by pastrel, and the small GTPase rac were all identified as essential for normal place memory. Examination of the dopamine and ecdysone receptor (DopEcR) did not reveal a function for this gene in place memory. When compared to the role of these genes in other memory types, these results suggest that there are genes that have both common and specific roles in memory formation across learning contexts. Importantly, contrasting the timing for the function of these four genes, plus a previously described role of the radish gene, in place memory with the temporal requirement of these genes in classical olfactory conditioning reveals variability in the timing of genetically-defined memory phases depending on the type of learning (Ostrowski, 2014).

Temperature as an aversive reinforcer interacts with training conditions to induce place memories of different stabilities. Previous work showed that intermittent training for Drosophila in space and place memory increases memory performance up to two hours after training. Shown in this study is that temperatures at or above 41°C are needed for induction of this longer lasting memory. That is, 37°C and below can act as an aversive reinforcer and condition flies to avoid a part of the training chamber, but continued avoidance decays within minutes of training. It is only with a temperature of 41°C that an hours-long memory is induced with massed and intermittent training. This abrupt difference in the length of the memory after training with the higher temperature may reflect a threshold of some sort, the steepness of which is currently unknown. This could arise from a differential input to the reinforcing circuit from separate sensory systems, like the Trp family of receptors, or from altered output from one of these sensory systems. Future studies on different temperature responsive proteins may differentiate between these possibilities (Ostrowski, 2014).

Genetic analysis challenges the use of time as a critical factor in determining a memory phase. Memory phases in the fly were initially examined after classical olfactory conditioning where an odorant is typically paired with an aversive electric shock or a rewarding sugar. Four different memory phases have been classified based roughly on time after training and genetic/pharmacological manipulations. Short-term memory after olfactory learning is measured within minutes of training; long-term memory and anesthesia resistant memory start to be active within hours and are increasingly important for memories at the 24 h range and longer. An intermediate memory is thought to be important in the interval between short-term and long-term memories. That time alone is a critical factor in determining these phases loses support when comparing flies with different mutations in aversive and rewarded olfactory memory. For example, the long-known mutant radish was originally shown to be important in the hours-long range after aversive olfactory training and genetically classified the anesthesia-resistant memory. Interestingly, this gene is important within minutes of training in rewarded olfactory memory (Ostrowski, 2014).

Several genes that are important for early to late phases of classical olfactory conditioning are critical on a finer time scale in place memory. Mutation of both the rut and lat genes leads to reduced aversive olfactory memory tested immediately after training, as well as longer time points. Although it is currently unclear when during the life-cycle these genes are important for place memory, mutation of rut and lat reduces memory directly after training. Furthermore, both the rut and lat products have been implicated in synaptic plasticity at the neuromuscular junction (NMJ), which suggests a role for these genes in early stages of learning and memory. It is pretty straight-forward that the rut-encoded type I adenylyl cyclase is also acting early on in associative processes in place learning. The lat gene encoding a subunit of the origin of replication (orc3) is also localized to the pre-synaptic specializations at the NMJs). The lat-orc3 also acts early-on in associative processes for place learning. How the lat-orc3 product is related to regulation of cAMP levels is, however, not as clear. The rut and lat results add to our understanding of an apparently common set of short-term changes in memory between olfactory and place memory, which include a common function of the S6 kinase II, an atypical tribbles kinase, and the arouser EPS8L3. And, the recently identified role of the foxp transcription factor specifically in operant learning, as tested in a flight simulator, suggests another set of genes that could be important for operant place memory in the minutes range (Ostrowski, 2014).

Late memory phases in classical olfactory conditioning depend on a set of genes that are important for place memory within minutes. The first challenge to a common timing of a memory phase came from the radish gene. In contrast to a role in the hours range after olfactory learning, radish mutant flies have a deficit in operant place memory within minutes of training. Furthermore, the pst gene (CG8588), encoding a novel product, has been previously shown to have a specific defect in aversive olfactory memory 24 h after spaced training. That is, the pst mutant flies have a normal short-term olfactory memory but a defective memory 1 day later. Interestingly, in the heat-box pst mutant flies already show a significant decrement in place memory immediately after training. This place memory defect seems to get worse within the first hour after training, reduced to ~50% of normal after 60 min. Thus, this 'long-term memory gene' is also involved in a memory within minutes of training in a second learning situation (Ostrowski, 2014).

Using the classical aversive olfactory learning paradigm the rac small GTPase has been identified as a key regulator in memory retention. Inhibition of Rac activity slows early olfactory memory decay, leading to elevated memory levels one hour after training, but becoming increasingly important 2 h after training. There does not appear to be an effect of Rac inhibition in olfactory memory in the minutes range after training. Transgenic flies with inhibited Rac function also have an increase in memory retention after place memory training. However, the first evidence of an increase in memory performance is within 10 min. Impressively, significant place memory was still evident up to 5 h after training, far beyond the range that can be typically measured in wild-type flies. Thus, while rac has a more general role in stabilizing memories, the timing of this function depends again on the type of memory trace that is formed (Ostrowski, 2014).

Not all memory genes first identified in other contexts, however, play a significant role in place memory. The DopEcR gene has been implicated in several behaviors, including a 30 min memory after courtship conditioning. This G-protein linked receptor is responsive to both dopamine and the steroid hormone ecdysone. Remarkably, DopEcR has been shown to interact with the cAMP cascade through double mutant and pharmacological tests. Using conditions that induce a robust and lasting place memory, the DopEcR mutant flies do not show a defect in memory directly after training or at 1 h post-training. This is despite the fact that the rut and cAMP-phosphodiesterase genes (dunce) are critical for place memory. It may be that DopEcR is not required for this type of learning and would be consistent with the independence of place memory from dopamine signaling. Alternatively, other redundant pathways may compensate for the reduction in DopEcR function caused by the DopEcRPB1 allele. One might further speculate that other types of behavioral plasticity, such as reversal learning or memory enhancement after unpredicted high temperature exposures in the heat-box might be more sensitive to DopEcR changes. Future experiments will determine if this is the case (Ostrowski, 2014).

Memory stability across learning contexts in Drosophila has some common genetic mechanisms, but the timing for gene action depends on the type of learning. That this study has added several genes here, including lat, pst, and rac as regulators of memory stability in operant place memory suggests that there are at least some common molecular processes in memory stability across different learning types. However, the timing of these genetically-defined phases depends on what is learnt. It is speculated that an ideal system to regulate memory stability would be one that activates its own decline. That is, a given memory type should activate the process of decreasing memory expression. This might work with the recruitment of a reinforcing pathway, like the dopaminergic signal that is important for both the acquisition of an associative olfactory memory and the active process of forgetting that association. In this case an odor associated with shock gives rise to a memory trace in mushroom body neurons that depends on a set of dopamine neurons that is important for both memory acquisition and decline. Whether this type of aminergic-based system applies to other forms of memory is not yet known. However, if an aminergic-based signal is common in memory decline, as appears to be the case with the Rac-based mechanism, differences in the types of aminergic neurons or innervation targets could give rise to the altered stabilities of behaviorally expressed memories (Ostrowski, 2014).


Anderson, M. S., Halpern, M. E. and Keshishian, H. (1988). Identification of the neuropeptide transmitter proctolin in Drosophila larvae: characterization of muscle fiber-specific neuromuscular endings. J. Neurosci. 8: 242-255. PubMed Citation: 2892897

Bell, S. P., et al. (1996). The multidomain structure of Orc1p reveals similarity to regulators of DNA replication and transcriptional silencing. Cell 83(4): 563-8. PubMed Citation: 7585959

Boynton, S. and Tully, T. (1992). latheo, a new gene involved in associative learning and memory in Drosophila melanogaster, identified from P element mutagenesis. Genetics 131(3): 655-72. PubMed Citation: 1321066

Carpenter, P. B. and Dunphy, W.G. (1998). Identification of a novel 81-kDa component of the Xenopus origin recognition complex. J. Biol. Chem. 273: 24891-24897. PubMed Citation: 9733795

Chesnokov, I., Remus, D. and Botchan, M. (2001). Functional analysis of mutant and wild-type Drosophila origin recognition complex. Proc. Natl. Acad. Sci. 98: 11997-12002. 11593009

Davis, G. W., Schuster, C. M. and Goodman, C. S. (1996). Genetic dissection of structural and functional components of synaptic plasticity. III. CREB is necessary for presynaptic functional plasticity. Neuron 17: 669-679. PubMed Citation: 8893024

DiAntonio, A. (1999). Learning something ORIGINal at the Drosophila neuromuscular junction. Neuron 23: 1-2. PubMed Citation: 10402184

Eaton, M. L., et al. (2011). Chromatin signatures of the Drosophila replication program. Genome Res. 21(2): 164-74. PubMed Citation: 21177973

Fischer, S. A., Fischer, T. M. and Carew, T. J. (1997). Multiple overlapping processes underlying short-term synaptic enhancement. Trends Neurosci. 20: 170-177

Fossier, P., Baux, G. and Tauc, L. (1993). Role of different types of Ca2+ channels and a reticulum-like Ca2+ pump in neurotransmitter release. J. Physiol. (Paris) 87: 3-14

Fossier, P., Deibler, M. F., Mothet, J. P., Israel, M., Tauc, L. and Baux, G. (1998). Control of the calcium concentration involved in acetylcholine release and its facilitation: an additional role for synaptic vesicles? Neuroscience 85: 85-91

Gatti, M. and Baker, B.S. (1989). Genes controlling essential cell-cycle functions in Drosophila melanogaster. Genes Dev. 3: 438-453

Gorczyca, M., Augart, C. and Budnik, V. (1993). Insulin-like receptor and insulin-like peptide are localized at neuromuscular junctions in Drosophila. J. Neurosci. 13: 3692-3704

Gossen, M., Pak, D. T. S., Hansen, S. K., Acharya, J. K. and Botchan, M. R. (1995). A Drosophila homolog of the yeast origin recognition complex. Science 270: 1674-1677

Keshishian, H., Chiba, A., Chang, T. N., Halfon, M. S., Harkins, E. W., Jarecki, J., Wang, L., Anderson, M. S., Cash, S., Halpern, M. E. and Johansen, J. (1993). Cellular mechanisms governing synaptic development in Drosophila melanogaster. J. Neurobiol. 24: 757-787

Kim, J. C., et al. (2011). Integrative analysis of gene amplification in Drosophila follicle cells: parameters of origin activation and repression. Genes Dev. 25(13): 1384-98. PubMed Citation: 21724831

Landis, G., Kelly, R., Spradling, A. C. and Tower, J. (1997). The k43 gene, required for chorion gene amplification and diploid cell chromosome replication, encodes the Drosophila homolog of yeast origin recognition complex subunit 2. Proc. Natl. Acad. Sci. 94: 3888-3892

Monastirioti, M., Gorczyca, M., Rapus, J., Eckart, M., White, K. and Budnick, V. (1995). Octopamine immunoreactivity in the fruit fly Drosophila melanogaster. J. Comp. Neurol. 356: 275-287

Ostrowski, D., Kahsai, L., Kramer, E. F., Knutson, P. and Zars, T. (2015). Place memory retention in Drosophila. Neurobiol Learn Mem 123: 217-224. PubMed ID: 26143995

Pinto, S., et al. (1999). latheo encodes a subunit of the origin recognition complex and disrupts neuronal proliferation and adult olfactory memory when mutant. Neuron 23: 45-54. PubMed ID: 10402192

Quintana, D. G., et al. (1997). Identification of HsORC4, a member of the human origin of replication recognition complex. J. Biol. Chem. 272(45): 28247-51

Quintana, D. G., et al. (1998). ORC5L, a new member of the human origin recognition complex, is deleted in uterine leiomyomas and malignant myeloid diseases. J. Biol. Chem. 273(42): 27137-45

Rohrbough, J., Pinto, S., Mihalek, R., Tully, T. and Broadie, K. (1999). latheo, a Drosophila gene involved in learning, regulates functional synaptic plasticity. Neuron 23: 55-70. PubMed ID: 10402193

Zhong, Y., and Wu, C.-F. (1991). Altered synaptic plasticity in Drosophila memory mutants with a defective cyclic AMP cascade. Science 251: 198-201

Zhong, Y., Budnik, V. and Wu, C. F. (1992). Synaptic plasticity in Drosophila memory and hyperexcitable mutants: role of cAMP cascade. J. Neurosci. 12(2): 644-51

Zinsmaier, K. E., Eberle, K. K., Buchner, E., Walter, N. and Benzer, S. (1994). Paralysis and early death in cysteine string protein mutants of Drosophila. Science 263: 977-980

Zucker, R. S. (1993). Calcium and transmitter release. J. Physiol. (Paris) 87: 25-36

latheo: Biological Overview | Evolutionary Homologs | Regulation | Developmental Biology | Effects of Mutation

date revised: 10 February 2012

Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.

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