(en)Methods of assessing the effect of a drug on long term memory and for screening a pharmaceutical agent for its ability to modulate long term memory are disclosed.

1.ApplicationNumber: US-6612505-A
1.PublishNumber: US-7572434-B2
2.Date Publish: 20090811
3.Inventor: TULLY TIMOTHY P.
YIN JERRY CHI-PING

4.Inventor Harmonized: TULLY TIMOTHY P(US)
YIN JERRY CHI-PING(US)

5.Country: US
6.Claims:
(en)Methods of assessing the effect of a drug on long term memory and for screening a pharmaceutical agent for its ability to modulate long term memory are disclosed.

7.Description:
(en)RELATED APPLICATIONS
This application is a Continuation of U.S. application Ser. No. 09/419,371 filed Oct. 14, 1999, now U.S. Pat. No. 6,890,516 which is a Divisional of U.S. application Ser. No. 08/809,917 filed Jul. 7, 1997, now U.S. Pat. No. 6,689,557, which is the U.S. National Phase of International Application No. PCT/US95/13198, filed Oct. 6, 1995, which is a Continuation-in-Part of U.S. application Ser. No. 08/361,063, filed Dec. 21, 1994, now U.S. Pat. No. 6,051,559, which is a Continuation-In-Part of U.S. application Ser. No. 08/319,866, filed Oct. 7, 1994, now U.S. Pat. No. 5,929,223. The entire teachings of the above applications are incorporated herein by reference.
This application contains an electronically submitted substitute sequence listing text file which has been filed through the EFS-Web. The materials submitted in the text file and substitute paper copy of the substitute sequence listing are incorporated herein by reference in thei entirety. The text file contains a single file named “17VV7191.txt”(70KB, created on June 17, 2009). The text file was created on Jun. 17, 2009.


GOVERNMENT SUPPORT
This invention was made with Government support under Grant No. 1R01 HD32245-01 awarded by the National Institutes of Health. The United States government has certain rights in this invention.


BACKGROUND OF THE INVENTION
Activation of the cyclic 3′,5′-adenosine monophosphate (cAMP) signal transduction pathway can have long-lasting global consequences through its influence on the expression of specific genes. This is true for simple organisms as well as mammals, where many of the known cAMP-responsive genes can have important neural and endocrine roles. Additional information regarding activation of this pathway would be useful, particularly as this activation pertains to the ability of animals to remember activities or events.
SUMMARY OF THE INVENTION
The present invention is based on Applicants' discovery of the dCREB1 and dCREB2 genes. The present invention is further based on Applicants' discovery that the Drosophila CREB2 gene codes for proteins of opposite functions. One isoform (e.g., dCREB2-a) encodes a cyclic 3′,5′-adenosine monophosphate (cAMP)-responsive transcriptional activator. Another isoform (e.g., dCREB2-b) codes for an antagonist which blocks the activity of the activator.
When the blocking form is placed under the control of the heat-shock promoter, and transgenic flies are made, a brief shift in temperature induces the synthesis of the blocker in the transgenic fly. This induction of the blocker (also referred to herein as the repressor) specifically disrupts long-term, protein synthesis dependent memory of an odor-avoidance behavioral paradigm.
As a result of Applicants' discovery, a method is herein provided to regulate long term memory in an animal. The method of regulating long term memory described herein comprises inducing expression of a dCREB2 gene or a fragment thereof in the animal.
The dCREB2 gene encodes several isoforms. Examples of an isoform encoded by the dCREB2 gene are dCREB2-a, dCREB2-b, dCREB2-c, dCREB2-d, dCREB2-q, dCREB2-r and dCREB2-s.
The isoforms encoded by the dCREB2 gene include cAMP-responsive activator isoforms and antagonistic blocker (or repressor) isoforms of the activator isoforms. Cyclic AMP responsive activator isoforms can function as a cAMP-responsive activator of transcription. Antagonistic repressors can act as a blocker of activators. An example of a cAMP-responsive activator isoform is dCREB2-a. An example of an antagonistic repressor (or blocker) isoform is dCREB2-b. The terms blocker and repressor are used interchangeably herein.
In one embodiment of the invention, the dCREB-2 gene encodes a cAMP-responsive activator isoform and inducing said gene results in the potentiation of long term memory.
Alternatively, inducing the dCREB2 gene encoding a cAMP-responsive activator isoform activates the production of a protein which is necessary for the formation of long term memory.
In another embodiment of the invention, the dCREB2 gene encodes a repressor isoform and inducing said gene results in the blocking of long term memory.
A further embodiment of the invention relates to a method of regulating long term memory in an animal comprising inducing repressor and activator isoforms of dCREB2 wherein long term memory is potentiated in the animal when the net amount of functional activator (ΔC) is greater than zero.
The invention also relates to a method of identifying a substance capable of affecting long term memory in an animal comprising the determination that said substance alters the induction or activity of repressor and activator isoforms of dCREB2 from normal in the animal.
As referred to herein, an activator isoform includes dCREB2-a and functional fragments thereof and a repressor isoform includes dCREB2-b and functional fragments thereof.
Other embodiments of the invention relate to a method of enhancing long term memory formation in an animal comprising increasing the level of activator homodimer from normal, decreasing the level of activator-repressor heterodimer from normal, or decreasing the level of repressor homodimer from normal in the animal.
Still another embodiment of the invention relates to a method of identifying a substance capable of affecting long term memory in an animal comprising the determination that said substance alters activator homodimer, activator-repressor heterodimer and/or repressor homodimer formation from normal in the animal.
As referred to herein, an activator homodimer includes the dCREB2a homodimer, an activator-repressor heterodimer includes the dCREB2a-dCREB2b heterodimer, and a repressor homodimer includes the dCREB2b homodimer.
A further embodiment of the invention relates to isolated DNA encoding a cAMP responsive transcriptional activator. Such a cAMP responsive transcriptional activator can be encoded by a Drosophila dCREB2 gene or by homologues or functional fragments thereof. For example, a cAMP responsive transcriptional activator can be encoded by the dCREB2 gene which codes for dCREB2-a or by a gene encoded by the sequences presented herein.
Still another embodiment of the invention relates to isolated DNA encoding an antagonist of cAMP-inducible transcription. Such an antagonist of cAMP-inducible transcription can be encoded by a Drosophila dCREB2 gene or by homologues or functional fragments thereof. For example, an antagonist of cAMP-inducible transcription can be encoded by the dCREB2 gene which codes for dCREB2-b.
Another embodiment of the invention relates to isolated DNA (SEQ ID NO.: 25) which encodes a Drosophila dCREB2 gene or functional fragments thereof.
A further embodiment of the invention relates to isolated DNA encoding an enhancer-specific activator. Such an enhancer-specific activator can be encoded by a Drosophila dCREB1 gene or by homologues or functional fragments thereof.
Another embodiment of the invention relates to isolated DNA encoding a nitric oxide synthase of Drosophila (DNOS). Such DNA can encode a DNOS of neuronal locus. The DNOS encoded can contain, for example, putative heme, calmodulin, flavin mononucleotide (FMN), flavin adenine dinucleotide (FAD) and nicotinamide adenine dinucleotide phosphate, in its reduced form, (NADPH) binding site domains.
A further embodiment of the invention relates to a method for assessing the effect of a drug on long term memory formation comprising administering the drug to Drosophila , subjecting the Drosophila to classical conditioning to at least one odorant and electrical shock, and assessing the performance index of the classical conditioning, wherein the effect of the drug occurs when it alters the performance index from normal. The drug can affect long term memory formation by, for example, altering the induction or activity of repressor and activator isoforms of dCREB2.
A still further embodiment of the invention relates to the assessment that an animal will have an enhanced or, alternatively, a diminished capability of possessing long term memory. This assessment can be performed by determining the amount of cAMP-responsive activator isoforms, cAMP-responsive repressor or blocker isoforms, or dimers of these isoforms that are present in the animal, where these isoforms are encoded by the CREB2 or a homologous gene. Enhanced capability of possessing long term memory will be more likely as the amount of activator exceeds the amount of repressor, i.e. in direct proportion to the size of the net amount of functional activator (ΔC) when this quantity is greater than zero. Conversely, diminished capability of processing long term memory will be more likely as the amount of repressor exceeds the amount of activator, i.e. in direct proportion to the size of the net amount of functional activator (ΔC) when this quantity is less than zero.
Another embodiment of the invention relates to a screening assay of pharmaceutical agents as enhancers of long term memory or as obstructors of long term memory in animals. The screening assay is performed by determining the change in the amount of cAMP-responsive activator isoforms, cAMP-responsive repressor or blocker isoforms, or dimers of these isoforms that is present in an animal or, more preferably, in a cell culture system or in Drosophila when the pharmaceutical agent is present, in comparison to when the pharmaceutical agent is not present, where these isoforms are encoded by the CREB2 or a homologous gene. Enhancers of long term memory cause a net increase in the amount of activator isoforms relative to the amount of repressor isoforms, i.e. an increase in the net amount of functional activator (ΔC). Obstructors of long term memory cause a net decrease in the amount of activator isoforms relative to the amount of repressor isoforms, i.e. a decrease in the net amount of functional activator (ΔC). The pharmaceutical agent can cause these changes by acting, for example, to alter the expression (transcription or translation) of the respective activator and/or repressor isoforms from the CREB2 or a homologous gene, to alter the formation of activator homodimers, activator-repressor heterodimers and/or repressor homodimers from the expressed isoforms, or to alter the interaction of one or more of these isoform or dimer types at their molecular targets. The long term memory activator isoform/repressor isoform system herein disclosed provides a unique platform for conducting such screening assays.
A further embodiment of the invention relates to an assay of pharmaceutical agents for their property as facilitators or hinderers of long term memory in animals. The assay is performed by administering the pharmaceutical agent to Drosophila prior to subjecting the Drosophila to a Pavlovian olfactory learning regimen. This regimen assesses the long term memory capabilities of the Drosophila by subjecting the flies to a massed and/or a spaced training schedule. Transgenic lines of these flies containing altered dCREB2 genes can be used to further elucidate the long term memory facilitation or hindering property of the pharmaceutical agent. The assay provides data regarding the acquisition of long term memory by the Drosophila after exposure to the pharmaceutical agent. These data are compared to long term memory acquisition data from Drosophila that have not been exposed to the pharmaceutical agent. If the exposed flies display faster or better retained long term memory acquisition than the unexposed flies, the pharmaceutical agent can be considered to be a facilitator of long term memory. Conversely, if the exposed flies display slower or less retained long term memory acquisition than the unexposed flies, the pharmaceutical agent can be considered to be a hinderer of long term memory. Since the genetic locus for this long term memory assay in Drosophila resides in the dCREB2 gene, the results from this assay can be directly applied to other animals that have homologous genetic loci (CREB2 or CREM genes).



BRIEF DESCRIPTION OF THE DRAWING
FIG. 1A depicts the DNA sequence (SEQ ID NO.: 1) and predicted amino acid sequence (SEQ ID NO.: 2) of the dCREB2-a coding region. The basic region and leucine zipper domains are indicated by solid and broken bold underlining, respectively; positively-charged residues in the basic region are circled; periodic leucines in the zipper motif are boxed; glutamines in the activation domain are underlined; the short amino acid motif with target sites for kinases, starting at residue 227, is indicated by a bold outline; and sequences specified by alternatively-spliced exons 2, 4 and 6 are shaded.
FIG. 1B depicts the amino acid sequences of the bZIP domains of dCREB2 (SEQ ID NO.: 3), mammalian CREB (SEQ ID NO.: 4), CREM (SEQ ID NO.: 5) and ATF-1 (SEQ ID NO.: 6). Differences between dCREB2 and CREB are boxed.
FIG. 2 is a schematic diagram of dCREB2 isoforms with the exon boundaries defined with respect to dCREB2-a. Diagram is not drawn to scale.
FIG. 3 is a bar graph representation of results showing pKA-responsive transcriptional activation by dCREB2-a.
FIG. 4 is a bar graph representation of results showing the transcriptional effect of dCREB2-b and a mutant variant on pKA-responsive activation by dCREB2-a.
FIG. 5 depicts the DNA sequence (SEQ ID NO.: 7) and predicted amino acid sequence (SEQ ID NO.: 8) of the dCREB1 coding region. The basic region and leucine zipper domains are indicated by solid and broken bold underlining, respectively; positively-charged residues in the basic region are circled; periodic leucines of the zipper motif are boxed; and in the acid-rich region of the activation domain, negatively-charged amino acids are underlined and proline residues are indicated by diamonds.
FIG. 6 is a bar graph representation of results showing transcriptional activation of a CRE reporter gene by dCREB1 in Drosophila Schneider L2 cell culture.
FIG. 7A is a photomicrograph of a Northern blot depicting the effect of heat shock induction on dCREB2-b expression: wt=wildtype flies; CREB=17-2 transgenic flies; lanes 1-2: no heat shock; lanes 2-3: immediately after heat shock; lanes 5-6: three hours after heat shock.
FIG. 7B is a photograph of a Western blot depicting the effect of heat shock induction on dCREB2-b protein production: wt=wildtype flies; CREB=17-2 transgenic flies; lanes 1-2: no heat shock; lanes 2-3: immediately after heat shock; lanes 5-6: one hour after heat shock; lanes 7-8: three hours after heat shock; lanes 9-10: 9 hours after heat shock; lanes 11-12: 24 hours after heat shock.
FIG. 7C is a photograph of a Western blot depicting the effect of heat shock induction on dCREB2 and dCREB2-mLZ (a mutated dCREB2-b) protein production: wt=17-2 transgenic flies (expressing wildtype blocker, dCREB2-b); m=A2-2 transgenic flies (expressing mutant blocker, dCREB2-mLZ); lanes 1-2: no heat shock; lanes 3-4: immediately after heat shock; lanes 5-6: three hours after heat shock; lanes 7-8: six hours after heat shock.
FIG. 8 is a bar graph representation of results showing the effect of cycloheximide (CXM) feeding, before or after spaced or massed training, on one-day memory retention: stripped bars=+CXM; hatched bars=−CXM.
FIG. 9A is a bar graph representation of results showing the effect of heat shock induction on one-day memory retention in wildtype (Can-S) flies and hs-dCREB2-b transgenic (17-2) flies given spaced or massed training: hatched bars=wildtype (Can-S) flies; stripped bars: hs-dCREB2-b transgenic (17-2) flies; hs=heat shock.
FIG. 9B is a bar graph representation of results showing the effect of heat shock induction on one-day memory retention in wildtype (Can-S) flies or hs-dCREB2-b transgenic (M11-1) flies given spaced or massed training: hatched bars=wildtype (Can-S) flies; stripped bars: hs-dCREB2-b transgenic (M11-1) flies; hs=heat shock.
FIG. 9C is a bar graph representation of results showing the effect of heat shock induction on learning in wildtype (Can-S) flies and hs-dCREB2-b transgenic (17-2) flies given spaced or massed training: hatched bars=wildtype (Can-S) flies; stripped bars: hs-dCREB2-b transgenic (17-2) flies; hs=heat shock.
FIG. 10 is a bar graph representation of results showing the effect of heat shock induction on one-day memory retention in wildtype [w(isoCJ1)] flies, hs-dCREB2-b transgenic (17-2) flies, and mutant hs-dCREB2-mLZ transgenic (A2-2) flies given spaced training: hatched bars=wildtype [w(isoCJ1)] flies; stripped bars=hs-dCREB2-b transgenic (17-2) flies; white bars=mutant hs-dCREB2-mLZ transgenic (A2-2) flies; hs=heat shock.
FIG. 11 is a bar graph representation of results showing the effect of heat shock induction on seven-day memory retention (long term memory) in wildtype (Can-S) flies and hs-dCREB2-b transgenic (17-2) flies given spaced training: hatched bars=wildtype (Can-S) flies; stripped bars=hs-dCREB2-b transgenic (17-2) flies; hs=heat shock.
FIG. 12 is a bar graph representation of results showing the effect of heat shock induction on one-day memory retention in hs-dCREB2-b transgenic (17-2) flies, radish mutant flies, and radish hs-dCREB2-b double mutant (rsh; 17-2) flies given spaced training: hs=heat shock; hatched bars=−hs; stripped bars=+hs.
FIG. 13A is a graphic representation of results showing the effect of repeated training sessions on seven-day memory retention (long term memory) in wildtype (Can-S) flies with long term memory as a function of the number of training sessions indicated by open circles and a negative accelerating exponential Gompertz (growth) function fit to the individual performance indexes (PIs) using a nonlinear iterative least squares method indicated by the solid line.
FIG. 13B is a graphic representation of results showing the effect of the rest interval between each training session on seven-day memory retention (long term memory) in wildtype (Can-S) flies with long term memory as a function of the rest interval indicated by open circles and a negative accelerating exponential Gompertz (growth) function fit to the individual performance indexes (PIs) using a nonlinear iterative least squares method indicated by the solid line.
FIG. 14 depicts a conceptual model of a molecular switch for the formation of long term memory based on differential regulation of CREB isoforms with opposing functions with ΔC indicating the net effect of CREB activators.
FIG. 15A is a bar graph representation of results showing the effect of 48 massed training sessions (48× massed) or 10 spaced training sessions with a 15-minute rest interval (10× spaced) on seven-day memory retention in wildtype (Can-S) flies.
FIG. 15B is a bar graph representation of results showing the effect of one (1×), two (2×) or ten (10×) massed training sessions, three hours after heat-shock induction of the transgene (induced) or in the absence of heat-shock (uninduced), on seven-day memory retention in wildtype (Can-S) flies, hsp-dCREB2-a transgenic (C28) flies, and hsp-dCREB2-a transgenic (C30) flies: black bars=wildtype (Can-S) flies; stripped bars=hsp-dCREB2-a transgenic (C28) flies; and white bars=hsp-dCREB2-a transgenic (C30) flies.
FIG. 15C is a bar graph representation of results showing responses three hours after heat shock in wildtype (Can-S) flies and hsp-dCREB2-a transgenic (C28) flies to odors, either octanol (OCT) or methylcyclohexanol (MCH), or to shock (60 V DC): black bars=wildtype (Can-S) flies; stripped bars=hsp-dCREB2-a transgenic (C28) flies.
FIG. 16A-16C depict the deduced amino acid sequences of DNOS and mammalian NOSs with amino acid numbering starting at the first methionine in each open reading frame (ORF), putative binding domains for cofactors (overlined) demarcated as in previously pubhished reports on mammalian NOSs, and amino acids which have been proposed as contacts with FAD and NADPH based on crystal structure of the ferrodoxin NADP reductase (Karplus, P.A., Science, 251: 60-66 (1991)) conserved at equivalent positions (bullet points): DNOS, Drosophil NOS (SEQ ID NO. 9); RNNOS, rat neuronal NOS (SEQ ID NO. 11); BENOS, bovine endothelial NOS (SEQ ID NO. 10); MMNOS, mouse macrophage NOS (SEQ ID NO. 12). Sequence alignment and secondary structure predictions were performed by Genework 2.3 (IntelliGenetics).
FIG. 16D is a schematic diagram of the domain structure of Drosophila and mammalian NOS proteins with the proposed cofactor-binding sites for heme (H), calmodulin (CaM), flavin mononucleotide (FMN), flavin adenine dinucleotide (FAD), nicotinamide adenine dinucleotide phosphate (NADPH) and the glutamine-rich domain (Q) in DNOS shown.
FIG. 17A is a photograph of a Western blot showing DNOS expression in 293 human embryonic kidney cells.
FIG. 17B is a bar graph representation of results showing DNOS enzyme activity measured in 293 human embryonic kidney cell extracts by conversion of 3 H-L-arginine to 3 H-L-citrulline: in the presence of exogenous Ca 2+ or calmodulin (group B); in the presence of 1 mM EGTA without exogenous Ca 2+ or calmodulin (group C); in the presence of 100 mM L-NAME with exogenous Ca 2+ or calmodulin (group D).
FIG. 18A is a photomicrograph of a Northern blot showing a 5.0 kb dNOS transcipt present in Drosophila heads: H=head; B=body.
FIG. 18B is a photograph of an agarose gel stained with ethidium bromide showing the expression by the dNOS gene of two alternatively spliced mRNA species with the arrows indicating the positions of the DNA fragments of the expected sizes: the 444 bp long-form fragment and the 129 bp short-form fragment. The other bands present in the lane are artifacts from heteroduplexes that failed to denature. KB=size markers.
FIG. 18C depicts the alignment of the deduced amino acid sequence of two protein isoforms of DNOS and mouse neuronal NOS: top part shows the relation between two conceptual Drosophila NOS proteins, DNOS-1 (amino acid residues 408-427 and 513-532 of SEQ ID NO.: 9) and DNOS-2 (SEQ ID NO.: 14), corresponding to the longer and shorter RT-PCR products, respectively; the bottom part shows the relationship between the relevant regions of two protein isoforms of the mouse neuronal NOS, n-NOS-1 (amino acid residues 494-513 and 599-618; SEQ ID NO.: 13 and SEQ ID NO.: 15, respectively) and n-NOS-2 (SEQ ID NO.: 16); and the numbers indicate the positions of the amino acid residues relative to the first methionine in the respective OFRs.
FIG. 19A-19B depicts the nucleotide sequence (SEQ ID NO.: 25) of a dNOS cDNA encoding the DNOS protein. The open reading frame of 4050 bp starts at nucleotide 189 and ends at nucleotide 4248.



DETAILED DESCRIPTION OF THE INVENTION
Applicants have cloned and characterized two genes, designated dCREB2 and dCREB1, isolated through a DNA-binding expression screen of a Drosophila head cDNA library in which a probe containing three cAMP-responsive element (CRE) sites was used.
The dCREB2 gene codes for the first known cAMP-dependent protein kinase (PKA) responsive CREB/ATF transcriptional activator in Drosophila . A protein data base search showed mammalian CREB, CREM and ATF-1 gene products as homologous to dCREB2. For these reasons, dCREB2 is considered to be a member, not only of the CREB/ATF family, but of the specific cAMP-responsive CREB/CREM/ATF-1 subfamily. It is reasonable to expect that dCREB2 is involved in Drosophila processes which are analogous to those which are thought to depend on cAMP-responsive transcriptional activation in other animal systems.
Applicants have shown that the dCREB2 transcript undergoes alternative splicing. Splice products of dCREB2 were found to fall into two broad categories: one class of transcripts (dCREB2-a, -b, -c, -d) which employs alternative splicing of exons 2, 4 and 6 to produce isoforms whose protein products all have the bZIP domains attached to different versions of the activation domain and a second class of transcripts (dCREB2-q, -r, -s) which have splice sites which result in in-frame stop codons at various positions upstream of the bZIP domain. These all predict truncated activation domains without dimerization or DNA binary activity.
dCREB2-a, -b, -c and -d are splice forms that predict variants of the activation domain attached to a common basic region-leucine zipper. These alternative splice forms result in seemingly minor changes in the size and spacing of parts of the activation domain. Nevertheless, alternative splicing of the activation domain has profound effects on the functional properties of dCREB2 products. Isoform dCREB2-a produces a PKA-responsive transcriptional activator in cell culture, whereas dCREB2-b, lacking exons 2 and 6, produces a specific antagonist. This dCREB2 splicing pattern (and its functional consequences) is virtually identical to that seen in the CREM gene. Similarly located, alternatively-spliced exons in the CREM gene determine whether a particular isoform is an activator or an antagonist (deGroot, R. P. and P. Sassone-Corsi, Mol. Endocrinol., 7: 145-153 (1993); Foulkes, N. S. et al., Nature, 355: 80-84 (1992)).
The ability of the phosphorylation domain (KID domain) to activate in trans other constitutive transcription factors which are bound nearby could potentially transform a CREM antagonist (which contains the KID domain but is lacking an exon needed for activation) into a cAMP-responsive activator. Since the modular organization of these molecules has been conserved, dCREB2-d could have this property.
In contrast to the dCREB2 splicing variants that encode isoforms with a basic region-leucine zipper domain, the dCREB2-q, -r and -s splice forms incorporate in-frame stop codons whose predicted protein products are truncated before the bZIP region. Isoforms of this type have been identified among the products of the CREB gene (deGroot, R. P. and P. Sassone-Corsi, Mol. Endocrinol., 7: 145-153 (1993); Ruppert, S. et al., EMBO J, 11: 1503-1512 (1992)) but not the CREM gene. The function of these truncated CREB molecules is not known, but at least one such CREB mRNA is cyclically regulated in rat spermatogenesis (Waeber, G. et al., Mol. Endocrinol., 5: 1418-1430 (1991)).
So far, dCREB2 is the only cAMP-responsive CREB transcription factor isolated from Drosophila . Other Drosophila CREB molecules, BBF-2/dCREB-A (Abel, T. et al., Genes Dev., 6: 466-488 (1992); Smolik, S. M. et al., Mol. Cell Biol., 12: 4123-4131 (1992)), dCREB-B (Usui, T. et al., DNA and Cell Biology, 12(7): 589-595 (1993)) and dCREB1, have less homology to mammalian CREB and CREM. It may be that dCREB2 subsumes functions of both the CREB and CREM genes in Drosophila . The mammalian CREB and CREM genes are remarkably similar to one another in several respects. It has been suggested that CREB and CREM are the product of a gene duplication event (Liu, F. et al., J. Biol. Chem., 268: 6714-6720 (1993); Riabowol, K. T. et al., Cold Spring Harbor Symp. Quant. Biol., 1: 85-90 (1988)). dCREB2 has a striking degree of amino acid sequence similarity to the CREB and CREM genes in the bZIP domain. Moreover, comparison of alternative splicing patterns among CREB, CREM and dCREB2 indicates that dCREB2 generates mRNA splicing isoforms similar to exclusive products of both CREB and CREM. Taken together, the sequence information and the splicing organization suggest that dCREB2 is an ancestor of both the mammalian CREB and CREM genes.
As discussed further herein, one phenomenon in which dCREB2 might act with enduring consequences is in long-term memory. This possibility is a particularly tempting one because recent work in Aplysia indicates that a CREB factor is likely to function in long-term facilitation by inducing an “immediate early” gene (Alberini, C. M. et al., Cell, 76: 1099-1114 (1994); Dash, P. K., Nature, 345: 718-721 (1990)). Recent experiments with a conditionally-expressed dCREB2-b transgene indicate that it has specific effects on long-term memory in Drosophila.
The product of the second gene described herein, dCREB1, also appears to be a member of the CREB/ATF family. Gel-retardation assays indicate that it binds specifically to CREs. It has a basic region and an adjacent leucine zipper at its carboxyl end, but this domain shows limited amino acid sequence similarity to other CREB/ATF genes. The presumed transcriptional activation domain of dCREB1 is of the acid-rich variety. Furthermore, it has no consensus phosphorylation site for PKA. dCREB1 can mediate transcriptional activation from CRE-containing reporters in the Drosophila L2 cell line, but this activation is not dependent on PKA.
A recurrent finding from work on the biology of learning and memory is the central involvement of the cAMP signal transduction pathway. In Aplysia, the cAMP second-messenger system is critically involved in neural events underlying both associative and non-associative modulation of a behavioral reflex (Kandel, E. R. and J. H. Schwartz, Science, 218: 433-443 (1982); Kandel, E. R., et al., In Synaptic Function , Edelmann, G. M., et al. (Eds.), John Wiley and Sons, New York (1987); Byrne, J. H., et al., In Advances in Second Messenger and Phosphoprotein Research , Shenolikar, S. and A. C. Nairn (Eds.), Raven Press, New York, pp. 47-107 (1993)). In Drosophila , two mutants, dunce and rutabaga, were isolated in a behavioral screen for defects in associative learning and are lesioned in genes directly involved in cAMP metabolism (Quinn, W. G., et al., Proc. Natl. Acad. Sci. USA, 71: 708-712 (1974); Dudai, Y., et al., Proc. Natl. Acad. Sci., USA 73: 1684-1688 (1976); Byers, D. et al., Nature, 289: 79-81 (1981); Livingstone, M. S., et al., Cell, 37: 205-215 (1984); Chen, C. N. et al., Proc. Natl. Acad. Sci. USA, 83: 9313-9317 (1986); Levin, L. R., et al., Cell, 68: 479-489 (1992)). These latter observations were extended with a reverse-genetic approach using inducible transgenes expressing peptide inhibitors of cAMP-dependent protein kinase (PKA) and with analyses of mutants in the PKA catalytic subunit (Drain, P. et al., Neuron, 6: 71-82 (1991); Skoulakis, E. M., et al., Neuron, 11: 197-208 (1993)). Recent work on mammalian long-term potentiation (LTP) also has indicated a role for cAMP in synaptic plasticity (Frey, U., et al., Science, 260: 1661-1664 (1993); Huang, Y. Y. and E. R. Kandel, In Learning and Memory , vol. 1, pp. 74-82, Cold Spring Harbor Press, Cold Spring Harbor, N.Y. (1994)).
The formation of long-lasting memory in animals and of long-term facilitation in Aplysia can be disrupted by drugs that interfere with transcription or translation (Agranoff, B. W. et al., Brain Res., 1: 303-309 (1966); Barondes, S. H. and H. D. Cohen, Nature, 218: 271-273 (1968); Davis, H. P. and L. R. Squire, Psychol. Bull., 96: 518-559 (1984); Rosenzweig, M. R. and E. L. Bennett, In Neurobiology of Learning and Memory , Lynch, G., et al. (Eds.), The Guilford Press, New York, pp. 263-288, (1984); Montarolo, P. G., et al., Science, 234: 1249-1254 (1986)). This suggests that memory consolidation requires de novo gene expression. Considered along with the involvement of the cAMP second-messenger pathway, this requirement for newly synthesized gene products suggests a role for cAMP-dependent gene expression in long-term memory (LTM) formation.
In mammals, a subset of genes from the CREB/ATF family are known to mediate cAMP-responsive transcription (Habener, J. F., Mol. Endocrinol., 4: 1087-1094 (1990); deGroot, R. P. and P. Sassone-Corsi, Mol. Endocrinol., 7: 145-153 (1993)). CREBs are members of the basic region-leucine zipper transcription factor superfamily; (Landschulz, W. H. et al., Science, 240: 1759-1764 (1988)). The leucine zipper domain mediates selective homo- and hetero-dimer formation among family members (Hai, T. Y. et al., Genes & Dev., 3: 2083-2090 (1989); Hai, T. and T. Curran, Proc. Natl. Acad. Sci. USA, 88: 3720-3724 (1991)). CREB dimers bind to a conserved enhancer element (CRE) found in the upstream control region of many cAMP-responsive mammalian genes (Yamamoto, K. K., et al., Nature, 334: 494-498 (1988)). Some CREBs become transcriptional activators when specifically phosphorylated by PKA (Gonzalez, G. A. and M. R. Montminy, Cell, 59: 675-680 (1989); Foulkes, N. S. et al., Nature, 355: 80-84 (1992)), while others, isoforms from the CREM gene, are functional antagonists of these PKA-responsive activators (Foulkes, N. S. et al., Cell, 64: 739-749 (1991); Foulkes, N. and P. Sassone-Corsi, Cell, 68: 411-414 (1992)).
Work in Aplysia has shown that cAMP-responsive transcription is involved in long-term synaptic plasticity (Schacher, S. et al., Science, 240: 1667-1669 (1988); Dash, P. K., Nature, 345: 718-721 (1990)). A primary neuronal co-culture system has been used to study facilitation of synaptic transmission between sensory and motor neurons comprising the monosynaptic component of the Aplysia gill-withdrawal reflex. Injection of oligonucleotides containing CRE sites into the nucleus of the sensory neuron specifically blocked long-term facilitation (Dash, P. K., Nature, 345: 718-721 (1990)). This result suggests that titration of CREB activity might disrupt long-term synaptic plasticity.
Described herein is the cloning and characterization of a Drosophila CREB gene, dCREB2. This gene produces several isoforms that share overall structural homology and nearly complete amino acid identity in the basic region-leucine zipper with mammalian CREBs. The dCREB2-a isoform is a PKA-responsive transcriptional activator whereas the dCREB2-b product blocks PKA-responsive transcription by dCREB2-a in cell culture. These molecules with opposing activities are similar in function to isoforms of the mammalian CREM gene (Foulkes, N. S. et al., Cell, 64: 739-749 (1991); Foulkes, N. and P. Sassone-Corsi, Cell, 68: 411-414 (1992); Foulkes, N. S. et al., Nature, 355: 80-84 (1992)). The numerous similarities in sequence and function between dCREB2 and mammalian CREBs suggest that cAMP-responsive transcription is evolutionarily conserved.
Genetic studies of memory formation in Drosophila have revealed that the formation of a protein synthesis-dependent long-term memory (LTM) requires multiple training sessions with a rest interval between them. As discussed further herein, this LTM is blocked specifically by induced expression of a repressor isoform of the cAMP-responsive transcription factor CREB. Also as discussed further herein, LTM information is enhanced after induced expression of an activator form of CREB. Maximum LTM is achieved after just one training session.
To investigate the role of CREBs in long-term memory (LTM) formation in Drosophila , dominant-negative transgenic lines which express dCREB2-b under the control of a heat-shock promoter (hs-dCREB2-b) were generated. Groups of flies, which had been heat-shock induced or left uninduced, were tested for memory retention after Pavlovian olfactory learning. This acute induction regimen minimized potential complications from inappropriate expression of dCREB2-b during development and allowed a clear assessment of the effect of hs-dCREB2-b induction on memory formation.
In Drosophila , consolidated memory after olfactory learning is composed of two genetically distinct components: anesthesia-resistant memory (ARM) and long-term memory (LTM). ARM decays to zero within four days after training, and formation of ARM is insensitive to the protein synthesis inhibitor cycloheximide (CXM) but is disrupted by the radish mutation (Folkers, E., et al., Proc. Natl. Acad. Sci. USA, 90: 8123-8127 (1993)). In contrast, LTM shows essentially no decay over at least seven days, its formation is cycloheximide-sensitive and it is not disrupted by radish. Two different training protocols involving massed and spaced sessions were employed (Ebbinghaus, H., Uber das Gedachtnis, Dover, N.Y.(1885); Baddeley, A. D., The Psychology of Memory , Basic Books, N.Y. (1976)) to dissect memory formation. The massed training procedure consists of ten consecutive training cycles with no rest interval between them, while the spaced training protocol consists of the same number of sessions but with a 15-minute rest between each. Their genetic dissection revealed that the massed protocol produced only ARM, while the spaced protocol produced memory retention composed of both ARM and LTM.
The behavioral results show that formation of LTM is completely blocked by induced expression of hs-dCREB2-b. This effect is remarkable in its behavioral specificity. ARM, a form of consolidated memory genetically distinguishable from LTM, but co-existing with it one-day after spaced training, was not affected. Learning and peripheral behaviors likewise were normal. Thus, the effect of the induced hs-dCREB2-b transgene is specific to LTM.
Induction of the mutant blocker did not affect LTM. This result, together with the molecular data which showed that induction of the wild-type blocker did not have widespread effects on transcription, suggests that the blocker is reasonably specific at the molecular level when it specifically blocks LTM. The wild-type blocker may disrupt cAMP-dependent transcription in vivo, since it can block cAMP-responsive transcription in cell culture. It is reasonable to infer that dimerization is necessary for blocker function and that the wild-type blocker could interfere with cAMP-responsive transcription either by forming heterodimers with dCREB2-a, the activator, or by forming homodimers and competing for DNA binding with homodimers of dCREB2-a. Thus, activators and repressors may form homodimers or heterodimers. It is reasonable to infer that long term memory is enhanced when the level of activator homodimer is increased from normal and/or when the level of activator-repressor heterodimer is decreased from normal and/or when the level of repressor homodimer is decreased from normal. In any case, the molecular target(s) of dCREB2-b are likely to be interesting because of the behavioral specificity of the block of LTM.
In Drosophila , consolidation of memory into long-lasting forms is subject to disruption by various agents. A single-gene mutation radish and the pharmacological agent CXM were used to show that long-lasting memory in flies is dissectable into two components, a CXM-insensitive ARM, which is disrupted by radish, and a CXM-sensitive LTM, which is normal in radish mutants. As described herein, CREB-family members are likely to be involved in the CXM-sensitve, LTM branch of memory consolidation. The results described herein, taken together with the showing that long-term memory is dissectable into a CXM-insentive ARM and a CXM-sensitive LTM, show that only one functional component of consolidated memory after olfactory learning lasts longer than four days, requires de novo protein synthesis and involves CREB-family members.
Based on work in Aplysia, a model has been proposed to describe the molecular mechanism(s) underlying the transition from short-term, protein synthesis-independent to long-term, protein synthesis-dependent synaptic plasticity (Alberini, C. M. et al., Cell, 76: 1099-1114 (1994)). The present work in Drosophila on long-term memory extends this model to the whole organism. Important molecular aspects of this transition seem to involve migration of the catalytic subunit of PKA into the nucleus (Backsai, B. J. et al., Science, 260: 222-226 (1993)) and subsequent phosphorylation and activation of CREB-family members (Dash, P. K., Nature, 345: 718-721 (1990); Kaang, B. K., et al., Neuron, 10: 427-435 (1993)). In flies, it is likely that the endogenous dCREB2-a isoform is one of these nuclear targets. Activated dCREB2-a molecules then might transcribe other target genes, including the immediate early genes—as is apparently the case in Aplysia. (Alberini, C. M. et al., Cell, 76: 1099-1114 (1994)).
It is remarkable that the cAMP signal transduction pathway, including its nuclear components, seem to be required for memory-related functions in each of these species and behavioral tasks. Taken together with cellular analyses of a long-lasting form of LTP in hippocampal slices (Frey, U., et al., Science, 260: 1661-1664 (1993); Huang, Y. Y. and E. R. Kandel, In Learning and Memory, vol. 1, pp. 74-82, Cold Spring Harbor Press, Cold Spring Harbor, N.Y. (1994)), the emerging picture is that cAMP-responsive transcription is a conserved molecular switch involved in the consolidation of short-term memory to long-term memory. Thus, it is reasonable to infer that differential regulation of CREB isoforms serves as a molecular switch for the formation of long term memory.
A universal property of memory formation is that spaced training (repeated training sessions with a rest interval between them) produces stronger, longer-lasting memory than massed training (the same number of training sessions with no rest interval) (Ebbinghaus, H., Uber das Gedachtnis , Dover, N.Y. (1885); Hintzman, D. L., In Theories in Cognitive Psychology: The Loyola Symposium , R. L. Solso (Ed.), pp. 77-99, Lawrence Erlbaum Assoc., Hillsdale, N.J. (1974); Carew, T. J., et al., Science, 175: 451-454 (1972); Frost, W. N., et al., Proc. Natl. Acad. Sci. USA, 82: 8266-8269 (1985)). This phenomenon also exists in fruit flies for a conditioned odor avoidance response (Tully, T. and W. G. Quinn, J. Comp. Physiol. 157: 263-277 (1985)). Genetic dissection of this long-lasting memory has revealed, however, an important difference between massed and spaced training. Spaced training produces two functionally independent forms of consolidated memory, ARM and LTM, while massed training produces only ARM.
As described herein, ARM and LTM differ primarily in their requirement for protein synthesis during induction. ARM is not affected when flies are fed the protein synthesis inhibitor cycloheximide (CXM) immediately before or after training, while LTM is completely blocked under the same feeding conditions. ARM in normal flies also decays away within four days after training, while LTM shows no decay for at least seven days. Thus, protein synthesis is required to induce LTM, but LTM is maintained indefinitely once formed. These latter properties of LTM have been observed throughout the animal kingdom (Davis, H. P. and L. R. Squire, Psychol. Bull., 96: 518-559 (1984); Castellucci, V. F., et al., J. Neurobiol., 20: 1-9 (1989); Erber, J., J. Comp.Physiol.Psychol., 90: 41-46 (1976); Jaffe, K., Physiol.Behav., 25: 367-371 (1980)). The emerging neurobiological interpretation is that formation of LTM involves protein synthesis-dependent structural changes at relevant synapses (Greenough, W. T., TINS, 7: 229-283 (1984); Buonomano, D. V. and J. H. Byrne, Science, 249: 420-423 (1990); Nazif, F. A., et al., Brain Res., 539: 324-327 (1991); Stewart, M. G., In Neural and Behavioural Plasticity: The Use of the Domestic Chick As A Model , R. J. Andrew (Ed.), pp. 305-328, Oxford, Oxford (1991); Bailey, C. H. and E. R. Kandel, Sem. Neurosci., 6:35-44 (1994)). The modern molecular view is that regulation of gene expression underlies this protein synthesis-dependent effect (Goelet, P. et al., Nature, 322: 419-422 (1986); Gall, C. M. and J. C. Lauterborn, In Memory:Organization and Locus of Change , L. R. Squire, et al., (Eds.) pp. 301-329 (1991); Armstrong, R. C. and M. R. Montminy, Annu.Rev.Neurosci., 16: 17-29 (1993)).
Why is spaced training required to induce LTM? The massed and spaced procedures both entail ten training sessions; consequently, flies receive equivalent exposure to the relevant stimuli (one odor temporally paired with electric shock and a second odor presented without shock). The only procedural difference between massed and spaced training is the rest interval between each training session. The absence of a rest interval between sessions during massed training does not appear to disrupt the memory formation process. The level of initial learning assayed immediately after massed training is similar to that after spaced training. In addition, ARM levels are similar after both training procedures. Thus, the presence of a rest interval during spaced training seems crucial to the induction of LTM.
To investigate the temporal kinetics of this rest interval in relation to the formation of LTM ( FIGS. 13A and 13B ), it was first established that the usual ten sessions of spaced training produced maximal 7-day memory retention (7-day retention is composed solely of LTM, since ARM decays to zero within four days.
FIG. 13A shows that 15 or 20 training sessions did not improve memory retention. Thus, ten spaced training sessions produces maximal, asymptotic levels of LTM.
LTM as a function of the length of the rest interval between 10 spaced training sessions was then assessed. FIG. 13B reveals a continuous increase in LTM from a O-min rest interval (massed training) to a 10-minute rest interval, at which time LTM levels reach maximum. Longer rest intervals yielded similar memory scores. These observations of LTM formation suggest an underlying biological process, which changes quantitatively during the rest interval between sessions and which accumulates over repeated training sessions.
In transgenic flies, the formation of LTM, but not ARM or any other aspect of learning or memory, is disrupted by induced expression of a repressor form of the cAMP-responsive transcription factor CREB (Example 4). Mutating two amino acids in the “leucine zipper” dimerization domain of this CREB repressor was sufficient to prevent the dominant-negative effect on LTM. Thus, indication of LTM is not only protein synthesis-dependent but also is CREB-dependent. Stated more generally, CREB function is involved specifically in a form of a memory that is induced only by spaced training. This observation was particularly intriguing in light of the molecular nature of CREB.
In Drosophila , transcriptional and/or post-translational regulation of dCREB2 yields several mRNA isoforms. Transient transfection assays in mammalian F9 cells have demonstrated that one of these isoforms (CREB2-a) functions as a cAMP-responsive activator of transcription, while a second isoform (CREB2-b) acts as an antagonistic repressor of the activator (Example 1; cf. Habener, J. F., Mol. Endocrinol., 4: 1087-1094 (1990); Foulkes, N. and P. Sassone-Corsi, Cell, 68: 411-414 (1992)). (This repressor isoform was used previously to generate the inducible transgene mentioned above.) The existence of different CREB isoforms with opposing functions suggested an explanation for the requirement of multiple training sessions with a rest interval between them for the formation of LTM.
In its simplest form, this model (Example 7; FIG. 14 ) supposes that cAMP-dependent protein kinase (PKA), activated during training, induces the synthesis and/or function of both CREB activator and repressor isoforms (cf. Yamamoto, K. K., et al., Nature, 334: 494-498 (1988); Backsai, B. J. et al., Science, 260: 222-226 (1993)). Immediately after training, enough CREB repressor exists to block the ability of CREB activator to induce downstream events. Then, CREB repressor isoforms are inactivated faster than CREB activator isoforms. In this manner, the net amount of functional activator (ΔC=CREB2a−CREB2b) increases during a rest interval and then accumulates over repeated training sessions (with a rest interval) to induce further the downstream targets involved with the formation of LTM (Montarolo, P. G., et al., Science, 234: 1249-1254 (1986); Kaang, B. K., et al., Neuron, 10: 427-435 (1993)).
This model leads to three predictions, which have been confirmed. First, if the functional difference between CREB activator and repressor isoforms is zero (ΔC=0) immediately after one training session, then additional massed training sessions should never yield LTM. FIG. 15A shows that 48 massed training sessions, rather than the usual 10, still does not produce any 7-day memory retention. Second, if the amount of CREB repressor is increased experimentally, ΔC will be negative immediately after training (ΔC<0). Then, enough CREB repressor may not decay during a rest interval to free enough CREB activator for induction of LTM. This has been shown to be the case for spaced training (15-min rest interval) after inducing the expression of a hsp-dCREB2-b (repressor) transgene three hours before training (Example 4). Third, if the amount of CREB activator is increased experimentally, ΔC will be positive immediately after training (ΔC>0). This effect, then, should eliminate or reduce the rest interval required to induce LTM. FIG. 15B shows the results from recent experiments in which the expression of a hsp-dCREB2-a (activator) transgene was induced three hours before training. In these transgenic flies, massed training produced maximal LTM. This effect appeared not to arise trivially, since olfactory acuity, shock reactivity ( FIG. 15C ) and initial learning were normal in transgenic flies after heat shock-induction. Thus, the requirement for a rest interval between training sessions to induce LTM specifically was eliminated.
FIG. 15B also shows that maximal LTM occurred in induced hsp-dCREB2-a transgenic flies after just one training session. The usual requirement for additional training to form a strong, long-lasting memory was no longer necessary. Thus, induced overexpression of a CREB activator has produced in otherwise normal flies, the functional equivalent of a “photographic” memory. This result indicates that the amount of CREB activator present during training—rather than the amount of activated PKA that reaches CREB in the nucleus, for instance (cf. Backsai, B. J. et al., Science, 260: 222-226 (1993); Kaang, B. K., et al., Neuron, 10: 427-435 (1993); Frank, D. A. and M. E. Greenberg, Cell, 79: 5-8 (1994))—is the rate-limiting step of LTM formation. Taken together, results from these experiments support the notion that the opposing functions of CREB activators and repressors act as a “molecular switch” (cf. Foulkes, N. S. et al., Nature, 355: 80-84 (1992)) to determine the parameters of extended training (number of training sessions and rest interval between them) required to form maximum LTM.
To date, seven different dCREB2 RNA isoforms have been identified, and more are hypothesized to exist. Each may be regulated differentially at transcriptional (Meyer, T. E., et al., Endocrinology, 132: 770-780 (1993)) and/or translation levels before or during LTM formation. In addition, different combinations of CREB isoforms may exist in different (neuronal) cell types. Consequently, many different combinations of activator and repressor molecules are possible. From this perspective, the notions that all activators and repressors are induced during a training session or that all repressors inactivate faster than activators (see above) need not be true. Instead, the model requires only that ΔC (the net function of activators and repressors) is less than or equal to zero immediately after training and increases with time (rest interval).
Theoretically, particular combinations of activator and repressor molecules in the relevant neuron(s) should determine the rest interval and/or number of training sessions necessary to produce maximum LTM for any particular task or species. Thus, the molecular identification and biochemical characterization of each CREB activator and repressor isoform used during LTM formation in fruit flies is the next major step toward establishing the validity of our proposed model. Similar experiments in other species may establish its generality.
CREB certainly is not involved exclusively with LTM. The dCREB2 gene, for instance, is expressed in all fruit fly cells and probably acts to regulate several cellular events (Foulkes, N. S. et al., Nature, 355: 80-84 (1992)).
So, what defines the specificity of its effects on LTM? Specificity most likely resides with the neuronal circuitry involved with a particular learning task. For olfactory learning in fruit flies, for instance, CREB probably is modulated via the cAMP second messenger pathway. Genetic disruptions of other components of this pathway are known to affect olfactory learning and memory (Livingstone, M. S., et al., Cell, 37: 205-215 (1984); Drain, P. et al., Neuron, 6: 71-82 (1991); Levin, L. R., et al., Cell, 68: 479-489 (1992); Skoulakis, E. M., et al., Neuron 11: 197-208 (1993); Qiu, Y. and R. L. Davis, Genes Develop. 7: 1447-1458 (1993)). Presumably, the stimuli used during conditioning (training) stimulate the underlying neuronal circuits. The cAMP pathway is activated in (some) neurons participating in the circuit, and CREB-dependent regulation of gene expression ensues in the “memory cells”. This neurobiological perspective potentially will be established in Drosophila by identifying the neurons in which LTM-specific CREB function resides. Experiments using a neuronal co-culture system in Aplysia already have contributed significantly to this issue (Alberini, C. M. et al., Cell, 76: 1099-1114 (1994) and references therein).
The involvement of CREB in memory, or in the structural changes of neurons which underlie memory in vivo, also has been implicated in mollusks (Dash, P. K., Nature, 345: 718-721 (1990); Alberini, C. M. et al., Cell, 76: 1099-1114 (1994)) and in mice (Bourtchuladze, R., et al., Cell, 79: 59-68 (1994)). Ample evidence also exists for the involvement of the cAMP second messenger pathway in associative learning in Aplysia (Kandel, E. R., et al., In Synaptic Function , Edelmann, G. M., et al. (Eds.), John Wiley and Sons, New York (1987); Byrne, J. H., et al., In Advances in Second Messenger and Phosphoprotein Research , Shenolikar, S. and A. C. Nairn (Eds.), Raven Press, New York, pp. 47-107 (1993)) and in rat hippocampal long-term potentiation (LTP), a cellular model of associative learning in vertebrates (Frey, U., et al., Science, 260: 1661-1664 (1993); Huang, Y. Y. and E. R. Kandel, In Learning and Memory , vol. 1, pp. 74-82, Cold Spring Harbor Press, Cold Spring Harbor, N.Y. (1994)). Finally, cellular and biochemical experiments have suggested that CREB function may be modulated by other second messenger pathways (Dash, P. K., et al., Proc. Natl. Acad. Sci. USA 88: 5061-5065 (1991); Ginty, D. D. et al., Science, 260: 238-241 (1993); deGroot, R. P. and P. Sassone-Corsi, Mol. Endocrinol., 7: 145-153 (1993)). These observation suggest that CREB might act as a molecular switch for LTM in many species and tasks.
Finally, why might the formation of LTM require a molecular switch? Many associative events occur only once in an animal's lifetime. Forming long-term memories of such events would be unnecessary and if not counterproductive. Instead, discrete events experienced repeatedly are worth remembering. In essence, a recurring event comprises a relevant signal above the noise of one-time events. Teleologically, then, the molecular switch may act as an information filter to ensure that only discrete but recurring events are remembered. Such a mechanism would serve efficiently to tailor an individual's behavioral repertoire to its unique environment.
The present invention also relates to isolated DNA having sequences which encode (1) a cyclic 3′,5′-adenosine monophosphate (cAMP) responsive transcriptional activator, or a functional fragment thereof, or (2) an antagonist of a cAMP responsive transcriptional activator, or a functional fragment thereof, or (3) both an activator and an antagonist, or functional fragments thereof of both.
The invention relates to isolated DNA having sequences which encode Drosophila dCREB2 isoforms, or functional analogues of a dCREB2 isoform. As referred to herein, a functional analogue of a dCREB2 isoform comprises at least one function characteristic of a Drosophila dCREB2 isoform, such as a cAMP-responsive transcriptional activator function and/or an antagonistic repressor of the cAMP activator function. These functions (i.e., the capacity to mediate PKA-responsive transcription) may be detected by standard assays (e.g., assays which monitor for CREB-dependent activation). For example, assays in F9 cells have been used extensively to study CREB-dependent activation because their endogenous cAMP-responsive system is inactive; (Gonzalez, G. A. et al., Nature, 337: 749-752 (1989); Masson, N. et al., Mol. Cell Biol., 12: 1096-1106 (1992); Masson, N. et al., Nucleic Acids Res., 21: 1163-1169 (1993)).
The invention further relates to isolated DNA having sequences which encode a Drosophila dCREB2 gene or a functional fragment thereof. Isolated DNA meeting these criteria comprise nucleic acids having sequences identical to sequences of naturally occurring Drosophila dCREB2 and portions thereof, or variants of the naturally occurring sequences. Such variants include mutants differing by the addition, deletion or substitution of one or more nucleic acids.
The invention relates to isolated DNA that are characterized by (1) their ability to hybridize to a nucleic acid having the DNA sequence in FIG. 1A (SEQ ID NO.: 1) or its complement, or (2) by their ability to encode a polypeptide of the amino acid sequence in FIG. 1A (SEQ ID NO.: 2) or functional equivalents thereof (i.e., a polypeptide which functions as a cAMP responsive transcriptional activator), or (3) by both characteristics. Isolated nucleic acids meeting these criteria comprise nucleic acids having sequences homologous to sequences of mammalian CREB, CREM and ATF-1 gene products. Isolated nucleic acids meeting these criteria also comprise nucleic acids having sequences identical to sequences of naturally occurring dCREB2 or portions thereof, or variants of the naturally occurring sequences. Such variants include mutants differing by the addition, deletion or substitution of one or more residues, modified nucleic acids in which one or more residues is modified (e.g., DNA or RNA analogs), and mutants comprising one or more modified residues.
Such nucleic acids can be detected and isolated under high stringency conditions or moderate stringency conditions, for example. “High stringency conditions” and “moderate stringency conditions” for nucleic acid hybridizations are explained on pages 2.10.1-2.10.16 (see particularly 2.10.8-11) and pages 6.3.1-6 in Current Protocols in Molecular Biology (Ausubel, F. M. et al., eds, Vol. 1, Suppl. 26, 1991), the teachings of which are incorporated herein by reference. Factors such as probe length, base composition, percent mismatch between the hybridizing sequences, temperature and ionic strength influence the stability of nucleic acid hybrids. Thus, high or moderate stringency conditions can be determined empirically, depending in part upon the characteristics of the known DNA to which other unknown nucleic acids are being compared for homology.
Isolated nucleic acids that are characterized by their ability to hybridize to a nucleic acid having the sequence in FIG. 1A or its complement (e.g., under high or moderate stringency conditions) may further encode a protein or polypeptide which functions as a cAMP responsive transcriptional activator.
The present invention also relates to isolated DNA having sequences which encode an enhancer-specific activator, or a functional fragment thereof.
The invention further relates to isolated DNA having sequences which encode a Drosophila dCREB 1 gene or a functional fragment thereof. Isolated DNA meeting these criteria comprise nucleic acids having sequences identical to sequences of naturally occurring Drosophila dCREB1 and portions thereof, or variants of the naturally occurring sequences. Such variants include mutants differing by the addition, deletion or substitution of one or more nucleic acids.
The invention further relates to isolated DNA that are characterized by (1) their ability to hybridize to a nucleic acid having the DNA sequence in FIG. 5 (SEQ ID NO.: 7) or its complement, or (2) by their ability to encode a polypeptide of the amino acid sequence in FIG. 5 (SEQ ID NO.: 8), or by both characteristics. Isolated DNA meeting these criteria also comprise nucleic acids having sequences identical to sequences of naturally occurring dCREB1 or portions thereof, or variants of the naturally occurring sequences. Such variants include mutants differing by the addition, deletion or substitution of one or more residues, modified nucleic acids in which one or more residues is modified (e.g., DNA or RNA analogs), and mutants comprising one or more modified residues.
Such nucleic acids can be detected and isolated under high stringency conditions or moderate stringency conditions as described above, for example.
Fragments of the isolated DNA which code for polypeptides having a certain function can be identified and isolated by, for example, the method of Jasin, M., et al., U.S. Pat. No. 4,952,501.
Nitric Oxide in Invertebrates: Drosophila dNOS Gene Codes for a Ca 2+ /Calmodulin-Dependent Nitric Oxide Synthase.
Nitric oxide (NO) is a gaseous mediator of a wide variety of biological processes in mammalian organisms. Applicants have cloned the Drosophila gene, dNOS, coding for a Ca 2+ /calmodulin-dependent nitric oxide synthase (NOS). Presence of a functional NOS gene in Drosophila provides conclusive evidence that invertebrates synthesize NO and presumably use it as a messenger molecule. Furthermore, conservation of an alternative RNA splicing pattern between dNOS and vertebrate neuronal NOS, suggests broader functional homology in biochemistry and/or regulation of NOS.
NO is synthesized by nitric oxide synthases (NOSs) during conversion of L-arginine to L-citrulline (Knowels, R. G., et al., Biochem. J, 298: 249 (1994); Nathan, C., et al., J. Biol. Chem., 269: 13725 (1994); Marletta, M. A., J. Biol. Chem., 268: 12231 (1993)). Biochemical characterization of NOSs has distinguished two general classes: (i) constitutive, dependent on exogenous Ca 2+ and calmodulin and (ii) inducible, independent of exogenous Ca 2+ and calmodulin. Analyses of cDNA clones have identified at least three distinct NOS genes in mammals (Bredt, D. S., et al., Nature, 351: 714-718 (1991); Lamas, S., et al., Proc. Natl. Acad. Sci. USA, 89: 6348-6352 (1992); Lyons, C. R., et al., J. Biol. Chem., 267: 6370 (1992); Lowenstein, C. J., et al., Proc. Natl. Acad. Sci. USA, 89: 6711 (1992); Sessa, W. C., et al., J. Biol. Chem., 267:15274 (1992); Geller, D. A., et al., Proc. Natl. Acad. Sci. USA, 90: 3491 (1993); Xie, Q. et al., Science, 256: 225-228 (1992)) neuronal, endothelial and macrophage, the former two of which are constitutive and the latter of which is inducible. The nomenclature for these different isoforms used here is historical, as it is clear now that one or more isoforms can be present in the same tissues (Dinerman, J. L., et al., Proc. Natl. Acad. Sci. USA, 91: 4214-4218 (1994)).
As a diffusible free-radical gas, NO is a multifunctional messenger affecting many aspects of mammalian physiology [for reviews, see Dawson, T. M., et al., Ann. Neurol., 32: 297 (1992); Nathan, C., FASEB J, 6: 3051 (1992); Moncada, S., et al., N. Eng. J. Med., 329: 2002-2012 (1993); Michel, T., et al., Amer. J. Cardiol., 72: 33C (1993); Schuman, E. M., et al., Annu. Rev. Neurosci., 17: 153-183 (1994)]. NO originally was identified as an endothelium-derived relaxing factor responsible for regulation of vascular tone (Palmer, R. M. J., Nature, 327: 524 (1987); Palmer, R. M. J., et al., Nature, 333: 664 (1988); Ignarro, L. J., et al., Proc. Natl. Acad. Sci. USA, 84: 9265 (1987)) and as a factor involved with macrophage-mediated cytotoxicity (Marletta, M. A., et al., Biochemistry, 21: 8706 (1988); Hibbs, J. B., et al., Biochem. Biophys. Res. Comm., 157: 87 (1989); Steuhr, D. J., et al., J. Exp. Med., 169: 1543 (1989)). Since NO has been implicated in several physiological processes including inhibition of platelet aggregation, promotion of inflammation, inhibition of lymphocyte proliferation and regulation of microcirculation in kidney (Radomski, M., et al., Proc. Natl. Acad. Sci. USA, 87: 5193 (1990); Albina, J. E., J. Immunol., 147: 144 (1991); Katz, R., Am. J. Physiol., 261: F360 (1992); lalenti, A., et al., Eur. J. Pharmacol., 211: 177 (1992)). More recently, NO also has been shown to play a role in cell-cell interactions in mammalian central and peripheral nervous systems—in regulating neurotransmitter release, modulation of NMDA receptor-channel functions, neurotoxicity, nonadrenergic noncholonergic intestinal relaxation (Uemura, Y., et al., Ann. Neurol., 27: 620-625 (1990)) and activity-dependent regulation of neuronal gene expression (Uemura, Y., et al., Ann. Neurol., 27: 620 (1990); Dawson, V. L., et al., Proc. Natl. Acad. Sci. USA, 88: 6368 (1991); Lei, S. Z., et al., Neuron, 8: 1087 (1992); Prast, H., et al., Eur. J. Pharmacol., 216: 139 (1992); Peunova, N., Nature, 364: 450 (1993)). Recent reports of NO function in synaptogenesis and in apoptosis during development of the rat CNS (Bredt, D. S., Neuron, 13: 301 (1994); Roskams, A. J., Neuron, 13: 289 (1994)) suggest that NO regulates activity-dependent mechanism(s) underlying the organization of fine-structure in the cortex (Edelman, G. M., et al., Proc. Natl. Acad. Sci. USA, 89: 11651-11652 (1992)). NO also appears to be involved with long-term potentiation in hippocampus and long-term depression in cerebellum, two forms of synaptic plasticity that may underlie behavioral plasticity (Bohme, G. A., Eur. J. Pharmacol., 199: 379 (1991); Schuman, E. M., Science 254: 1503 (1991); O'Dell, T. J., et al., Proc. Natl. Acad. Sci. USA, 88: 11285 (1991); Shibuki, K., Nature, 349: 326 (1991); Haley, J. E., et al., Neuron, 8: 211 (1992); Zhuo, M., Science, 260: 1946 (1993); Zhuo, M., et al., Neuro Report, 5: 1033 (1994)). Consistent with these cellular studies, inhibition of NOS activity has been shown to disrupt learning and memory (Chapman, P. F., et al., Neuro Report, 3: 567 (1992); Holscher, C., Neurosci. Lett., 145: 165 (1992); Bohme, G. A., et al., Proc. Natl. Acad. Sci. USA, 90: 9191 (1993); Rickard, N. S., Behav. Neurosci., 108: 640-644 (1994)).
Many of the above conclusions are based on pharmacological studies using inhibitors of nitric oxide synthases or donors of NO. Interpretations of such studies usually are limited because the drugs interact with more than one target and they cannot be delivered to specific sites. A molecular genetic approach can overcome these problems, however, by disrupting a specific gene, the product of which may be one of the drug's targets. Recently, such an approach has been attempted in mice via generation of a knock-out mutation of the neuronal NOS (nNOS) (Huang, P. L., et al., Cell, 75: 1273-1286 (1993)). While nNOS mutants appeared fully viable and fertile, minor defects in stomach morphology and hippocampal long-term potentiation were detected (Huang, P. L., et al., Cell, 75: 1273-1286 (1993); O'Dell, T. J., et al., Science, 265: 542-546 (1994)). Moreover, some NOS enzymatic activity still was present in certain regions of the brain, suggesting a role for other NOS genes in the CNS. While. yielding some relevant information about one specific component of NO function, this NNOS disruption existed throughout development. Consequently, functional defects of NOS disruption in adults could not be resolved adequately from structural defects arising during development. Genetic tools exist in Drosophila , in contrast, to limit disruptions of gene functions temporally or spatially.
To identify candidate Drosophila NOS homologs, a fragment of the rat neuronal NOS cDNA (Bredt, D. S., et al., Nature, 351: 714-718 (1991)) was hybridized at low stringency to a phage library of the Drosophila genome as described in Example 11. The rat cDNA fragment encoded the binding domains of FAD and NADPH (amino acids 979-1408 of SEQ ID NO.: 11), which are cofactors required for NOS activity, and therefore were expected to be conserved in fruit flies. Several Drosophila genomic clones were identified with the rat probe and classified into eight contigs. Sequence analysis of three restriction fragments from these genomic clones revealed one (2.4R) with high homology to mammalian NOSs. The deduced amino acid sequence of the ORF encoded within the 2.4R fragment indicated 40% identity to the rat neuronal NOS and binding sites for FAD and NADPH.
The 2.4R DNA fragment then was used to probe a Drosophila adult head cDNA library as described in Example 11, and eight clones were isolated. Restriction analysis indicated that all contained identical inserts and thus, defined a predominant transcript expressed by this Drosophila gene. One clone (c5.3) was sequenced in both directions. The 4491 bp cDNA contained one long ORF of 4350 bp. The methionine initiating this ORF was preceded by ACAAG which is a good match to the translation start consensus (A/CAAA/C) for Drosophila genes (Cavener, D. R., Nucleic Acids Res., 15: 1353-1361 (1987)). Conceptual translation of this ORF yielded a protein of 1350 amino acids with a molecular weight of 151,842 Da.
Comparison of the amino acid sequence of this deduced Drosophila protein (DNOS) (SEQ ID NO.: 9) to sequences of mammalian NOSs revealed that DNOS is 43% identical to neuronal NOS (SEQ ID NO.: 11), 40% identical to endothelial NOS (SEQ ID NO.: 10) and 39% identical to macrophage NOS (SEQ ID NO.: 12). It also revealed similar structural motifs in DNOS ( FIG. 16A-16C ). The C-terminal half of the DNOS protein contains regions of high homology corresponding to the presumptive FMN-, FAD- and NADPH-binding sites. Amino acids thought to be important for making contacts with FAD and NADPH in mammalian NOSs (Bredt, D. S., et al., Nature, 351: 714-718 (1991); Lamas, S., et al., Proc. Natl. Acad. Sci. USA, 89: 6348-6352 (1992); Lyons, C. R., et al., J. Biol. Chem., 267: 6370 (1992); Lowenstein, C. J., et al., Proc. Natl. Acad. Sci. USA, 89: 6711 (1992); Sessa, W. C., et al., J. Biol. Chem., 267: 15274 (1992); Geller, D. A., et al., Proc. Natl. Acad. Sci. USA, 90: 3491 (1993); Xie, Q. et al., Science, 256: 225-228 (1992)) are conserved in DNOS. The middle section of DNOS, between residues 215 and 746 of SEQ ID NO.: 9, showed the highest similarity to mammalian NOSs: it is 61% identical to the neuronal isoform and 53% identical to endothelial and macrophage isoforms. Sequences corresponding to the proposed heme- and calmodulin-binding sites in mammalian enzymes are well-conserved in DNOS. The region located between residues 643-671 of SEQ ID NO.: 9 has the characteristics of a calmodulin-binding domain (basic, amphiphilic α-helix) (O'Neil, K. T., et al., Trends Biochem. Sci., 15: 59-64 (1990)). The amino acid sequence between these two sites is very well conserved among all four NOS proteins, suggesting the location of functionally important domains such as the arginine-binding site (Lamas, S., et al., Proc. Natl. Acad. Sci. USA, 89: 6348-6352 (1992)), tetrahydrobiopterine cofactor binding site or a dimerization domain. DNOS also has a PKA consensus site (Pearson, R. B., Meth. Enzymol., 200: 62-81 (1991)) (at Ser-287 of SEQ ID NO.: 9) in a position similar to neuronal and endothelial NOSs.
The 214 amino acid N-terminal domain of DNOS shows no obvious homology to its equivalent portion of neuronal NOS or to the much shorter N-terminal domains of endothelial and macrophage NOSs. This region of DNOS contains an almost uninterrupted homopolymeric stretch of 24 glutamine residues. Such glutamine-rich domains, found in many Drosophila and vertebrate proteins, have been implicated in protein-protein interactions regulating the activation of transcription (Franks, R. G., Mech. Dev., 45: 269 (1994); Gerber, H.-P., et al., Science, 263: 808 (1994); Regulski, M., et al., EMBO J, 6: 767 (1987)). Thus, this domain of DNOS could be involved with protein-protein interactions necessary for localization and/or regulation of DNOS activity.
The above sequence comparisons suggest that a Drosophila structural homolog of a vertebrate NOS gene was identified. The order of the putative functional domains in the DNOS protein is identical to that of mammalian enzymes ( FIG. 15B ). Structural predictions based on several protein algorithms also indicate that general aspects of DNOS protein secondary structure (hydrophobicity plot, distribution of α-helixes and β-strands) from the putative heme-binding domain to the C-terminus are similar to those of mammalian NOSs. DNOS also does not contain a transmembrane domain, as is the case for vertebrate NOSs. In addition to these general characteristics, several aspects of DNOS structure actually render it most like neuronal NOS: (i) the overall sequence similarity, (ii) the similarity of the putative calmodulin-binding site (55% identical to the neuronal NOS vs. 45% identical to endothelial NOS or vs. 27% identical to macrophage NOS) and (iii) the large N-terminal domain. Neuronal NOS and DNOS also do not contain sites for N-terminal myristoylation, which is the case for endothelial NOS (Lamas, S., et al., Proc. Natl. Acad. Sci. USA, 89: 6348-6352 (1992)), nor do they have a deletion in the middle of the protein, which is the case for macrophage NOS (Xie, Q. et al., Science, 256: 225-228 (1992)).
To establish that Applicants putative DNOS protein had nitric oxide synthase activity, the dNOS cDNA was expressed in 293 human embryonic kidney cells as described in Example 12, which have been used routinely in studies of mammalian NOSs (Bredt, D. S., et al., Nature, 351: 714-718 (1991)). Protein extracts prepared from dNOS-transfected 293 cells as described in Example 12, contained a 150 kD polypeptide, which was recognized by a polyclonal antibody raised against the N-terminal domain of DNOS ( FIG. 17A , lane 293+dNOS). This immunoreactive polypeptide was of a size expected for DNOS and was absent from cells transfected with just the pCGN vector alone ( FIG. 17A , lane 293+vector).
Extracts made from dNOS-transfected 293 cells showed significant NO synthase activity, as measured by the L-arginine to L-citrulline conversion assay as described in Example 12 (0.1276±0.002 pmol/mg/min; FIG. 17B , group B). [In a parallel experiment, the specific activity of rat neuronal NOS expressed from the same vector in 293 cells was 3.0±0.02 pmol/mg/min, N=4]. DNOS activity was dependent on exogenous Ca 2+ /calmodulin and on NADPH, two cofactors necessary for activity of constitutive mammalian NOSs (Iyengar, R., Proc. Natl. Acad. Sci. USA, 84: 6369-6373 (1987); Bredt, D. S., Proc. Natl. Acad. Sci. USA, 87: 682-685 (1990)). DNOS activity was reduced 90% by the Ca 2+ chelator EGTA ( FIG. 17B , group C). Also, 500 μM N-(6-aminohexyl)-1-naphthalene-sulfonamide (W5), a calmodulin antagonist which inhibits activity of neuronal NOS (Bredt, D. S., Proc. Natl. Acad. Sci. USA, 87: 682-685 (1990)), diminished DNOS activity to 18% (0.0222±0.001 pmol/mg/min, N=2). In the absence of exogenous NADPH, DNOS (or nNOS) activity was reduced 20% (0.1061±0.011 pmol/mg/min, N=4 for DNOS; 2.7935±0.033 pmol/mg/min, N=2 for NNOS). DNOS activity also was blocked by inhibitors of mammalian NOSs (Rees, D. D., Br. J. Pharmacol., 101: 746-752 (1990)). N G -nitro-L-arginine methyl ester (L-NAME) reduced DNOS activity 84% ( FIG. 17B , group D), and 100 μM N G -monomethyl-L-arginine acetate produced a complete block (0.0001±0.0002 pmol/mg/min, N=2). These enzymatic data demonstrate that DNOS is a Ca 2+ /calmodulin-dependent nitric oxide synthase.
Northern blot analysis indicated a 5.0 kb dNOS transcript which was expressed predominantly in adult fly heads but not bodies ( FIG. 18A ). More sensitive RT-PCR experiments as described in Example 13, however, detected dNOS message in poly(A) + RNA from fly bodies. Neuronal NOS genes from mice and humans produce two alternatively spliced transcripts, the shorter one of which yields a protein containing a 105 amino acid in-frame deletion (residues 504-608 in mouse or rat neuronal NOS) (Ogura, T., Biochem. Biophys. Res. Commun., 193: 1014-1022 (1993)). RT-PCR amplification of Drosophila head mRNA produced two DNA fragments: the 444 bp fragment corresponded to vertebrate long form and the 129 bp fragment corresponded to vertebrate short form ( FIG. 18B ). Conceptual translation of the 129 bp sequence confirmed a splicing pattern identical to that for the NNOS gene ( FIG. 18C ). Presence of the short NOS isoform in Drosophila strengthens the notion that it may play an important role in NOS biochemistry.
The discovery of a NOS homolog in Drosophila provides definitive proof that invertebrates produce NO and, as suggested by recent reports, most likely use it for intercellular signaling. These data also suggest that a NOS gene was present in an ancestor common to vertebrates and arthropods, implying that NOS has existed for at least 600 million years. Thus, it is expected that NOS genes are prevalent throughout the animal kingdom.
Consistent with this view are existing histochemical data. NOS activity has been detected in several invertebrate tissue extracts: in Lymulus polyphemus Radomski, M. W., Philos. Trans. R. Soc. Lond. B. Biol. Sci., 334: 129-133 (1992)), in the locust brain (Elphick, M. R., et al., Brain Res., 619: 344-346 (1993)), in the salivary gland of Rhodniusprolixus (Ribeiro, J. M. C., et al., FEBS Lett., 330: 165-168 (1993)(34)) and in various tissues of Lymnaea stagnalis (Elofsson, R., et al., Neuro Report, 4: 279-282 (1993)). Applications of NOS inhibitors or NO-generating substances have been shown to modulate the activity of buccal motoneurones in Lymnaea stagnalis (Elofsson, R., et al., Neuro Report, 4: 279-282 (1993)) and the oscillatory dynamics of olfactory neurons in procerebral lobe of Limax maximus (Gelperin, A., Nature, 369: 61-63 (1994)). NADPH-diaphorase staining, a relatively specific indicator of NOS protein in fixed vertebrate tissue samples (Dawson, T. M., et al., Proc. — Natl. Acad. Sci . USA, 88: 7797 (1991); Hope, B. T., et al., Proc. Natl. Acad. Sci. USA, 88: 2811 (1991)), also has suggested the presence of NOS in Drosophila heads (Muller, U., Naturwissenschaft, 80: 524-526 (1993)). The present molecular cloning of dNOS considerably strengthens the validity of these observations.
Sophisticated genetic analyses of NOS function are available in Drosophila . Classical genetics will allow the creation of point mutations and deletions in dNOS, resulting in full or partial loss of dNOS function. Such mutations will permit detailed studies of the role of NOS during development.
The invention further relates to isolated DNA that are characterized by their ability to encode a polypeptide of the amino acid sequence in FIG. 16A-16C (SEQ ID NO.: 9) or functional equivalents thereof (i.e., a polypeptide which synthesizes nitric oxide). Isolated DNA meeting this criteria comprise amino acids having sequences homologous to sequences of mammalian NOS gene products (i.e., neuronal, endothelial and macrophage NOSs). The DNA sequence represented in SEQ ID NO.: 25 is an example of such an isolated DNA. Isolated DNA meeting these criteria also comprise amino acids having sequences identical to sequences of naturally occurring dNOS or portions thereof, or variants of the naturally occurring sequences. Such variants include mutants differing by the addition, deletion or substitution of one or more residues, modified nucleic acids in which one or more residues is modified (e.g., DNA or RNA analogs), and mutants comprising one or more modified residues.
Such nucleic acids can be detected and isolated under high stringency conditions or moderate stringency conditions, for example. “High stringency conditions” and “moderate stringency conditions” for nucleic acid hybridizations are explained on pages 2.10.1-2.10.16 (see particularly 2.10.8-11) and pages 6.3.1-6 in Current Protocols in Molecular Biology (Ausubel, F. M. et al., eds, Vol. 1, Suppl. 26, 1991), the teachings of which are incorporated herein by reference. Factors such as probe length, base composition, percent mismatch between the hybridizing sequences, temperature and ionic strength influence the stability of nucleic acid hybrids. Thus, high or moderate stringency conditions can be determined empirically, depending in part upon the characteristics of the known DNA to which other unknown nucleic acids are being compared for homology.
Isolated DNA that are characterized by their ability to encode a polypeptide of the amino acid sequence in FIG. 16A-16C , encode a protein or polypeptide having at least one function of a Drosophila NOS, such as a catalytic activity (e.g., synthesis of nitric oxide) and/or binding function (e.g., putative heme, calmodulin, FMN, FAD and NADPH binding). The catalytic or binding function of a protein or polypeptide encoded by hybridizing nucleic acid may be detected by standard enzymatic assays for activity or binding (e.g., assays which monitor conversion of L-arginine to L-citrulline). Functions characteristic of dNOS may also be assessed by in vivo complementation activity or other suitable methods. Enzymatic assays, complementation tests, or other suitable methods can also be used in procedures for the identification and/or isolation of nucleic acids which encode a polypeptide having the amino acid sequence in FIG. 16A-16C or functional equivalents thereof.
The present invention will now be illustrated by the following examples, which are not intended to be limiting in any way.
EXAMPLES
The following materials and methods were used in the work described in Examples 1 and 2.
Expression Cloning of dCREB1 and dCREB2
Standard protocols for expression cloning by DNA-binding (Ausubel, F., Current Protocols in Molecular Biology, John Wiley and Sons, New York, 1994; Singh, H. et al., Cell, 52: 415-423 (1988)) were followed except as noted. A double-stranded, 3×CRE oligonucleotide was synthesized and cloned between the XbaI and KpnI sites of pGEM7Zf+ (Promega). The sequence of one strand of the oligonucleotide was 5′CGTCTAGATCTA TGACTGAA TA TGACGTAA TA TGACGTAA TGGTACCA GATCTGGCC 3′ (SEQ ID NO.: 17), with the CRE sites underlined. The oligonucleotide was excised as a BglII/HindIII fragment and labeled by filling-in the overhanging ends with Klenow fragment in the presence of [α 32 P]dGTP, [α 32 P]dCTP and unlabeled dATP and dTTP (Ausubel, F., Current Protocols in Molecular Biology , John Wiley and Sons, New York, 1994)). Just prior to use, the labeled fragment was pre-absorbed to blank nitrocellulose filters to reduce background binding. All other steps were as described (Ausubel, F., Current — Protocols in Molecular Biology , John Wiley and Sons, New York, 1994)). After secondary and tertiary lifts, positive clones were subcloned into pKS+ (Stratagene) and sequences.
Gel Shift Analysis
Gel-mobility shift assays were performed as in Ausubel, F., Current Protocols in Molecular Biology , John Wiley and Sons, New York, 1994, with the following modifications. The 4% polyacrylamide gel (crosslinking ratio 80:1) was cast and run using 5× Tris-glycine buffer (Ausubel, F., Current Protocols in Molecular Biology , John Wiley and Sons, New York, 1994)) supplemented with 3 mM MgCl 2 . The oligonucleotides used as the DNA probes were boiled and slowly cooled to room temperature at a concentration of 50 μg/ml in 0.1 M NaCl. 50 ng of double-stranded probe was end-labeled using polynucleotide kinase in the presence of 100 μCi of [γ 32 P]ATP. The double-stranded oligonucleotides were purified on a native polyacrylamide gel and used in a mobility shift assay at about 0.5 ng/reaction.
For dCREB2, the original dCREB2-b cDNA was subcloned and subjected to site-directed mutagenesis to introduce restriction sites immediately 5′ and 3′ of the open reading frame. This open reading frame was subcloned into the pET11A expression vector (Novagen) and used to induce expression of the protein in bacteria. The cells containing this vector were grown at 30° C. to an approximate density of 2×10 8 /ml and heat-induced at 37° C. for 2 hours. The cells were collected by centrifugation and lysed according to Buratowski, S. et al., Proc. Natl. Acad. Sci. USA, 88: 7509-7513 (1991). The crude extract was clarified by centrifugation and loaded onto a DEAE colum previously equilibrated with 50 mM Tris HCl, pH 8.0, 10% sucrose, 100 mM KCl. Step elutions with increasing amounts of KCl in the same buffer were used to elute the dCREB2-b protein, which was assayed using the gel mobility-shift assay. The peak fraction was dialyzed against the loading buffer and used in the binding experiment. The specific competitor that was used was the wild-type CRE oligonucleotide. The sequence of one strand of the double-stranded oligonucleotides used in the gel shift analysis are listed. For the first two oligonucleotides, wild-type and mutant CREs are underlined.






Wild-type 3xCRE (SEQ ID NO.: 18):


5′ AAA TGACGTAA CGGAAA TGACGTAA CGGAAA TGACGTAA CG 3′;



Mutant 3xmCRE (SEQ ID NO.: 19):

5′ AAA TGAATTAA CGGAAA TGAATTA ACGGAAA TGAATTAA CGG 3′;



Nonspecific competitor #1 (SEQ ID NO.: 20):

5′ TGCACGGGTTTTCGACGTTCACTGGTAGTGTCTGATGAGGCCGAAAG

GCCGAAACGCGATGCCCATAACCACCACGCTCAG 3′;



Nonspecific competitor #2 (SEQ ID NO.: 21):

5′ TCGACCCACAGTTTCGGGTTTTCGAGCAAGTCTGCTAGTGTCTGATG

AGGCCGAAAGGCCGAAACGCGAAGCCGTATTGCACCACGCTCATCGAGAA

GGC 3′;



Nonspecific competitor #3 (SEQ ID NO.: 22):

5′ CTAGAGCTTGCAAGCATGCTTGCAAGCAAGCATGCTTGCAAGCATGC

TTGCAAGC 3′;



Nonspecific competitor #4 (SEQ ID NO.: 23):

5′ CTCTAGAGCGTACGCAAGCGTACGCAAGCGTACG 3′




























For dCREB1, heat-induced bacterial extracts (Ausubel, F., Current Protocols in Molecular Biology , John Wiley and Sons, New York, 1994) were made from the original phage clone integrated by lysogeny. Extract from a bacteria lysogenized with another plaque (which did not bind to CRE sites) from the screen was used as a negative control. Competition experiments were done using a 4-100 fold molar excess (relative to the probe) of unlabeled, wild-type CRE oligonucleotides or unlabeled, mutant CRE oligonucleotides.
Northern Blots
Total head and body RNA was isolated from flies according to the protocol of Drain, P. et al., Neuron, 6:71-82 (1991). Total RNA from all other developmental stages was a gift from Eric Schaeffer. All RNA samples were selected twice on oligo-dT columns (5 Prime-3 Prime) to isolate poly A+ RNA. Two μg of poly A+ RNA was fractionated on 1.2% formaldehyde-formamide agarose gels, transferred to nitrocellulose and probed using an uniformly labeled, strand-specific, antisense RNA (aRNA) probe. The template for the synthesis of aRNA was one of the partial cDNA clones isolated from the library screen (pJY199). This cDNA contained the carboxyl-terminal 86 amino acids of the dCREB2-b protein and about 585 bp of 3′ untranslated mRNA. All Northern blots were washed at high stringency (0.1% SDS, 0.1×SSC, 65° C.).
In situ Hybridization to Tissue Sections
Frozen frontal sections were cut and processed under RNAse-free conditions, essentially as described in Nighorn, A. et al., Neuron, 6:455-467 (1991), with modifications for riboprobes as noted here. Digoxigenin-labeled riboprobes were made from pJY199 using the Genius kit (Boehringer-Mannheim). One μg of Xba-linearized template and T3 RNA polymerase was used to make the antisense probe, while one μg of EcoRI-linearized template together with T7 RNA polymerase was used for the control sense probe. Alkaline hydrolysis (30 minutes at 60° C.) was used to reduce the average probe size to about 200 bases. The hydrolyzed probe was diluted 1:250 in hybridization solution (Nighorn, A. et al., Neuron, 6:455-467 (1991)), boiled, quickly cooled on ice, added to the slides and hybridized at 42° C. overnight. The slides were then treated with RNAse A (20 μg/ml RNAse A in 0.5 M NaCl/10 mM Tris pH8 for 1 hour at 37° C.) prior to two 50° C. washes. Digoxigenin detection was as described.
Reverse Transcription Coupled With the Polymerase Chain Reaction (RT-PCR) Analysis of dCREB2 and Identification of Alternatively Spliced Exons
The template for reverse transcription coupled with the polymerase chain reaction (RT-PCR) was total RNA or poly A+ RNA isolated from Drosophila heads as in Drain, P. et al., Neuron, 6: 71-82 (1991). Total RNA used was exhaustively digested with RNase-free DNase 1 (50 μg of RNA digested with 50 units of DNase I for 60-90′ at 37° C. followed by phenol, phenol/chloroform extraction, and ethanol precipitation) prior to use. Results from separate experiments indicate that this DNase-treatment effectively eliminates the possibility of PCR products derived from any contaminating genomic DNA. Two rounds of selection using commercial oligo-dT columns (5 Prime-3 Prime) were used to isolate poly A+ RNA from total RNA. The template for an individual reaction was either 100-200 ng of total RNA, or 10-20 ng of poly A+ RNA.
The RT-PCR reactions were performed following the specifications of the supplier (Perkin-Elmer) with a “Hot Start” modification (Perkin-Elmer RT-PCR kit instructions). All components of the RT reaction, except the rTth enzyme, were assembled at 75° C., and the reaction was initiated by adding the enzyme and lowering the temperature to 70° C. At the end of 15 minutes, the preheated (to 75° C.) PCR components (including trace amounts of [α 32 P]dCTP) were added quickly, the reaction tubes were put into a pre-heated thermocycler, and the PCR amplification begun. Cycling parameters for reactions (100 μl total volume) in a Perkin-Elmer 480 thermocycler were 94° C. for 60 seconds, followed by 70° C. for 90 seconds. For reactions (50 μl) in an MJ Minicycler the parameters were 94° C. for 45 seconds and 70° C. for 90 seconds.
All primers used in these procedures were designed to have 26 nucleotides complementary to their target sequence. Some primers had additional nucleotides for restriction sites at their 5′ ends to facilitate subsequent cloning of the products. Primers were designed to have about 50% GC content, with a G or C nucleotide at their 3′ most end and with no G/C runs longer than 3. For RT-PCR reactions with a given pair of primers, the Mg +2 concentration was optimized by running a series of pilot reactions, at Mg +2 concentrations ranging from 0.6 mM to 3.0 mM. Reaction products were analyzed on denaturing urea-polyacrylamide gels by autoradiography. Any product that appeared larger than the band predicted from the cDNA sequence was purified from a preparative native gel, re-amplified using the same primers, gel-purified, subcloned and sequenced.
To verify that a given RT-PCR product was truly derived from RNA, control reactions were run to show that the appearance of the product was eliminated by RNase A treatment of the template RNA. Also, products generated from reactions using total RNA as the template were re-isolated from reactions using twice-selected polyA+ RNA as template.
Plasmids
Expression constructs for transient transfection experiments in Drosophila were made in the expression vector pAct5CPPA (Han, K. et al., Cell, 56: 573-583 (1989)) or in pAcQ. pAcQ is a close derivative of pAct5CPPA in which the XbaI site at the 5′ end of the 2.5 kb actin promoter fragment was destroyed and additional sites were inserted in the polylinker. pAc-dCREB1 was made by subcloning a KpnI-SacI fragment containing the complete dCREB1 open reading frame (from a cDNA subcloned into pKS+) into pAct5CPPA. pAc-PKA was constructed by subcloning an EcoRV fragment encoding the Drosophila PKA catalytic subunit (Foster, J. L. et al., J. Biol. Chem., 263: 1676-1681 (1988)) from a modified pHSREM1 construct (Drain, P. et al., Neuron, 6: 71-82 (1991)) into pAct5CPPA. To make the 3×CRE-lacZ reporter construct for Drosophila cell culture, the double-stranded, wild-type 3×CRE oligonucleotide used in the gel shift analysis was cloned into the KpnI-XbaI backbone of HZ50PL (Hiromi, Y. and W. J. Gehring, Cell, 50: 963-974 (1987)), a reporter construct made for enhancer testing which has cloning sites in front of a minimal hsp70 promoter-lacZ fusion gene.
RSV-dCREB2-a was constructed in a long series of cloning steps. Essentially, the activator-encoding open reading frame was first reconstructed on the plasmid pKS+ by sequentially adding each of the three exons (exons 2, 4 and 6) into the original cDNA of dCREB2-b, which had been subcloned from phage DNA into pKS+. Site-directed mutagenesis was used to introduce unique restriction enzyme sites both 5′ and 3′ of the dCREB2-b open reading frame, and these sites facilitated the subcloning process and allowed removal of 5′ and 3′ untranslated sequences. Once the activator was assembled, the resulting open reading frame was sequenced to confirm the cloning steps and moved into a modified RSV vector which contained a polylinker located between the RSV promoter and the SV40 polyadenylation sequences (RSV-0). RSV-dCREB2-b was made by moving the original dCREB2-b cDNA (which had been subcloned into pKS+) into RSV-0.
Other constructs used in experiments were: pCaE (pMtC) (Mellon, P. L. et al., Proc. Natl. Acad. Sci. USA, 86: 4887-4891 (1989)), which contains the cDNA for mouse PKA catalytic subunit cloned under the mouse metallothionein 1 promoter; RSV-βgal (Edlund, T. et al., Science, 230: 912-916 (1985)), which expresses the lacZ gene under control of the Rous sarcoma long terminal repeat promoter (Gorman, C. M. et al., Proc. Natl. Acad. Sci. USA, 79: 6777-6781 (1982)). RSV-CREB (Gonzalez, G. A. et al., Nature, 337: 749-752 (1989)) is a CREB cDNA fragment containing the 341-amino acid open reading frame under the RSV LTR-promoter in RSV-SG, and the D(−71) CAT reporter (Montminy, M. R. et al., Proc. Natl. Acad. Sci. USA, 83: 6682-6686 (1986)) which is a fusion of a CRE-containing fragment of the rat somatostatin promoter and the bacterial CAT coding region.
F9 Cell Culture and Transfection
Undifferentiated F9 cells were maintained and transfected using the calcium phosphate method as described in Darrow, A. L. et al., “Maintenance and Use of F9 Teratocarcinoma Cells”, In Meth. Enzymol., v. 190 (1990), except that chloroquine was added to 100 mM immediately before transfection and precipitates were washed off ten hours after transfection, at which time the dishes received fresh chloroquine-free medium. Amounts of DNA in transfections were made equivalent by adding RSV-0 where required. Cells were harvested 30 hours after transfection. Extracts were made by three cycles of freeze/thawing, with brief vortexing between cycles. Particulate matter was cleared from extracts by ten minutes of centrifugation in the cold. β-galactosidase assays were performed as described in Miller, J. H., Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., 1972. CAT assays were performed as described in Sheen, J. Y. and B. Seed, Gene, 67: 271-277 (1988) using aliquots of extract heat-treated at 65° C. for ten minutes and centrifuged for ten minutes to remove debris. Results reported are from three experiments run on different days with at least three dishes per condition within each experiment. Error bars represent standard error of the mean, with error propagation taken into account (Grossman, M. and H. W. Norton, J. Hered., 71: 295-297 (1980)).
Drosophila Cell Culture and Transient Transfection
Schneider L2 cells in Schneider's medium (Sigma) supplemented with 10% fetal bovine serum (FBS) or Kc167 cells in D-22 medium (Sigma) supplemented with 10% FBS, were transfected by the calcium phosphate method essentially as described in Krasnow, M. A. et al., Cell, 57: 1031-1043 (1989), with the following differences. Kc167 cells were plated at 2×10 6 cells/ml and chloroquine was added to a final concentration of 100 mM immediately prior to transfection. A total of 10 μg of plasmid DNA per dish was used for L2 transfections and 25 μg per dish for Kc167 transfections. DNA masses in transfections were made equivalent with pGEM7Zf+ where required. Precipitates were left undisturbed on L2 cells until harvest, but for Kc167 cells the original medium was replaced with fresh, chloroquine-free medium after twelve hours. Cells were harvested thirty-six to forty-eight hours after transfection. Extracts were made and enzymatic assays performed as described above for F9 cells. Results reported for transfections are averages of at least three experiments run on different days, with at least duplicate dishes for each condition within experiments. Error bars represent standard error of the mean, with error propagation taken into account (Grossman, M. and H. W. Norton, J. Hered., 71: 295-297 (1980)).
β-galactosidase (βgal) and Chloramphenicol Acetyl Transferase (CAT) Assays
β-galactosidase assays were run and activity calculated as described in Miller, J. H., Experiments in Molecular Genetics , Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., 1972. CAT assays were performed essentially according to Sheen, J. Y. and B. Seed, Gene, 67: 271-277 (1988)) using supernatants from heat-treated aliquots of extracts (65° C. for 10 minutes and then centrifuged for 10 minutes). Relative activity was calculated according to Sheen, J. Y. and B. Seed, Gene, 67: 271-277 (1988)).
PKA-Responsive Transcriptional Activation by dCREB2-a
F9 cells were transiently transfected with 10 μg of D(−71) CAT plasmid as a CRE-directed reporter. 5 μg of RSV-βgal reporter was included in each dish as a normalization control for transfection efficiency. Different groups received 8 μg of dCREB2-a expression vector and 4 μg of PKA expression vector, separately or in combination. Results are expressed as CAT/Pgal enzyme activity ratios, standardized to values obtained with PKA-transfected dishes.
Transcriptional Effect of dCREB2-b and a Mutant Variant on PKA-Responsive Activation by dCREB2-a
F9 cells were transiently cotransfected with 10 μg of D(−71) CAT along with the indicated combinations of the following expression constructs: RSV-dCREB2-a (5 μg); pMtC (2 μg); RSV-dCREB2-b (5 μg); and RSV-mLZ-dCREB2-b, which expresses a leucine-zipper mutant of dCREB2-b (5 μg). The DNA mass for each dish was made up to 27 μg with RSV-0. Other experimental conditions are as described above under “PKA-Responsive Transcriptional Activation by dCREB2-a”.
Transcriptional Activation of a CRE Reporter Gene by dCREB1 in Drosophila Schneider L2 Cell Culture
The cells were transiently transfected with a dCREB1 expression construct (1 μg), with or without a construct which expresses Drosophila PKA. 3×CRE-βgal reporter (1 μg) and the normalization Ac-CAT reporter (1 μg) were included in each dish. Expression vectors not present in particular dishes were replaced by pACQ.
Example 1
Isolation and Characterization of dCREB2
Two different genes were isolated in a DNA-binding expression screen of a Drosophila head cDNA library using a probe containing three CRE sites (3×CRE). Many clones were obtained for the dCREB2 gene, while only one clone was obtained for dCREB1. The dCREB2 clones had two alternatively-spliced open reading frames, dCREB2-b and dCREB2-c (see FIG. 2 ). These differed only in the presence or absence of exon 4 and in their 5′ and 3′ untranslated regions. The inferred translation product of dCREB2-b showed very high sequence similarity to the amino acid sequences of the basic region/leucine zipper (bZIP) domains of mammalian CREB (SEQ ID NO.: 4), CREM (SEQ ID NO.: 5) and ATF-1 (SEQ ID NO.: 6) (see FIG. 1B ).
Chromosomal in situ hybridization using a dCREB2 probe localized the gene to a diffuse band at 17A2 on the X chromosome, a region which contains several lethal complementation groups (Eberl, D. F. et al., Genetics, 30: 569-583 (1992)).
To determine the DNA binding properties of dCREB2-b, the DNA binding activity of dCREB2-b was assayed using a gel mobility-shift assay. Bacterial extracts expressing the dCREB2-b protein retarded the migration of a triplicated CRE probe (3×CRE). The protein had lower, but detectable, affinity for a mutated 3×CRE oligonucleotide. Competition experiments using unlabeled competitor oligonucleotides showed that the binding of dCREB2-b to 3×CRE was specific with higher affinity for CRE sites than to nonspecific DNA. Together with the conserved amino acid sequence, this functional similarity suggested that dCREB2 was a CREB family member.
The expression pattern of dCREB2 was determined by Northern blot analysis of poly A+RNA from various developmental stages. A complex pattern with at least 12 different transcript sizes was found. Two bands of approximately 0.8 and 3.5 kb were common to all of the stages. The adult head contained transcripts of at least six sizes (0.8, 1.2, 1.6, 1.9, 2.3 and 3.5 kb). In situ hybridization to RNA in Drosophila head tissue sections showed staining in all cells. In the brain, cell bodies but not neuropil were stained.
dCREB2 has alternatively-spliced forms. Initial transfection experiments showed that the dCREB2-c isoform was not a PKA-responsive transcriptional activator. This information, together with the complex developmental expression pattern and the known use of alternative splicing of the CREM gene to generate PKA-responsive activators (Foulkes, N. and P. Sassone-Corsi, Cell, 68: 411-414 (1992); Foulkes, N. S. et al., Nature, 355: 80-84 (1992)) suggested that additional domains might be required to code for an activator.
Reverse transcription coupled with the polymerase chain reaction (RT-PCR) was used to identify new exons. Comparison of the genomic DNA sequence with that of cDNAs indicated the general exon/intron organization and assisted in the search for additional exons. Sense and antisense primers spaced across the dCREB2-b cDNA were synthesized and used pairwise in RT-PCR reactions primed with Drosophila head RNA. Reactions with primers in exons 5 and 7 (see FIG. 2 ) generated two products, one with the predicted size (compared with the cDNA clones) and one that was larger. The larger fragment was cloned and its sequence suggested the presence of exon 6 (see FIG. 1A ; SEQ ID NO.: 1). A primer within exon 6 was synthesized, end-labeled and used to screen a Drosophila head cDNA library. Two clones were isolated, sequenced and found to be identical. This splicing isoform, dCREB2-d, confirmed the splice junctions and exon sequence inferred from the RT-PCR products.
Initial attempts to isolate exon 2 proved difficult. The genomic sequence that separated exons 1 and 3 (see FIG. 2 ) was examined and one relatively extensive open reading frame (ORF) was identified. Three antisense primers, only one of which lay inside this ORF, were synthesized based on the intron sequence. Three sets of RT-PCR reactions were run in parallel, each using one of the three antisense primers and a sense primer in exon 1. Only the reaction that used the antisense primer in the ORF produced a PCR product. The sequence of this product matched a continuous stretch of nucleotides from the genomic sequence, extending 3′ from exon 1 past the splice junction in the dCREB2-b cDNA to the location of the antisense primer in the ORF. This fragment suggested that exon 1 might be extended in some mRNAs by use of an alternative 5′ splice site located 3′ to the site used to make dCREB2-b. Based on the newly-identified exon sequences, a sense primer was made. This primer was used with an antisense primer in exon 3 to generate a new product whose sequence established the location of the new 5′ splice site. The sequence added to exon 1 by alternative 5′ splice site selection is denoted exon 2. The exon 2 sequence also showed that the same 3′ splice site was used both for the original cDNA and for exon 2. To independently verify this alternative splicing pattern, RT-PCR was carried out with a primer that spanned the 3′ splice junction of exon 2 and a primer in exon 1. The sequence of the product corroborated the splice junctions of exon 2 shown in FIG. 1A (SEQ ID NO.: 1).
To determine if exons 2 and 6 could be spliced into the same molecule, an RT-PCR reaction was carried out with primers in exons 2 and 6. The reaction produced a product of the size predicted by coordinate splicing of these two exons and the identity of this product was confirmed by extensive restriction analysis.
dCREB2 is a Drosophila CREB/ATF gene. FIG. 1A shows the DNA sequence (SEQ ID NO.: 1) and inferred amino acid sequence (SEQ ID NO.: 2) of dCREB2-a, the longest ORF that can result from the identified alternative splicing events. The indicated translation start site for this ORF is probably authentic because: i) stop codons occur upstream from this ATG in all reading frames in our dCREB2 cDNAs (sequences not shown) ii) this ATG was selected by computer (Sheen, J. Y. and B. Seed, Gene, 67: 271-277 (1988)) as the best ribosome binding site in the DNA sequence that contains the ORF; and iii) use of the next ATG in the open reading frame 480 nucleotides downstream would not produce an inferred product that is a PKA-dependent activator (see below). This information does not exclude the possibility that internal translation initiation sites may be used in this transcript, as happens in the CREM gene's S-CREM isoform (Delmas, V. et al., Proc. Natl. Acad. Sci. USA, 89: 4226-4230 (1992)).
The dCREB2-a open reading frame predicts a protein of 361 amino acids with a carboxyl-terminal bZIP domain (SEQ ID NO.: 3) highly homologous to those of mammalian CREB (SEQ ID NO.: 4) and CREM (SEQ ID NO.: 5) (see FIG. 1B ). The inferred dCREB2-a product has a small region of amino acids containing consensus phosphorylation sites for PKA, calcium/calmodulin-dependent kinase II (CaM kinase II) and protein kinase C (PKC) at a position similar to that of the P-box in CREB, CREM and ATF-1. The amino-terminal third of the predicted dCREB2-a is rich in glutamines (including runs of four and five residues). Glutamine-rich activation domains occur in CREB, CREM, and other eukaryotic transcription factors, including some from Drosophila (Courey, A. J. and R. Tijan, “Mechanisms of Transcriptional Control as Revealed by Studies of the Human Transcription Factor Sp 1”, In Transcriptional Regulation , vol. 2, McKnight, S. L. and K. R. Yamamoto (eds.), Cold Spring Harbor Press, Cold Spring Harbor, N.Y., 1992; Mitchell, P. J. and R. Tijan, Science, 245: 371-378 (1989)).
A computerized amino acid sequence homology search with the predicted dCREB2-a protein sequence (SEQ ID NO.: 2) identifies CREB, CREM and ATF-1 gene products as the closest matches to dCREB2-a. The homology is particularly striking in the carboxyl-terminal bZIP domain where 49 of 54 amino acids are identical with their counterparts in mammalian CREB ( FIG. 1B ). The homology is less striking, albeit substantial, in the activation domain. Lower conservation in this domain is also characteristic of the mammalian CREB and CREM genes (Masquilier, D. et al., Cell Growth Differ., 4: 931-937 (1993)).
FIG. 2 shows the exon organization of all of the dCREB2 alternative splice forms that we have detected, both as cDNAs and by RT-PCR. Splice products of dCREB2 fall into two broad categories. One class of transcripts (dCREB2-a, -b, -c, -d) employs alternative splicing of exons 2, 4 and 6 to produce isoforms whose protein products all have the bZIP domains attached to different versions of the activation domain. The second class of transcripts (dCREB2-q, -r, -s) all have splice sites which result in in-frame stop codons at various positions upstream of the bZIP domain. These all predict truncated activation domains without dimeriation or DNA binding activity.
Two different dCREB2 isoforms, dCREB2-a and dCREB2-b, have opposite roles in PKA-responsive transcription. The capacity of isoforms of the dCREB2 gene to mediate PKA-responsive transcription was tested in F9 cells. These cells have been used extensively to study CREB-dependent activation because their endogenous cAMP-responsive transcription system is inactive (Gonzalez, G. A. et al., Nature, 337: 749-752 (1989); Masson, N. et al., Mol. Cell Biol., 12: 1096-1106 (1992); Masson, N. et al., Nucleic Acids Res., 21: 1163-1169 (1993)). Candidate cAMP-responsive transcription factors, synthesized from expression vectors, were transiently transfected with and without a construct expressing the PKA catalytic subunit. CREB-dependent changes in gene expression were measured using a cotransfected construct that has a CRE-containing promoter fused to a reporter gene.
The product of the dCREB2-a isoform is a PKA-dependent activator of transcription ( FIG. 3 ). Transfection of PKA or dCREB2-a alone gave only modest activation above baseline values. Cotransfection of dCREB2-a and PKA together, however, gave levels of activation 5.4-fold greater than the activation seen with PKA alone.
dCREB2-b did not act as a PKA-dependent transactivator. When transfected together with the reporter and PKA, it failed to stimulate reporter activity. Instead, it functioned as a direct antagonist of PKA-dependent activation by dCREB2-a ( FIG. 4 ). Cotranfection of equimolar amounts of the dCREB2-a and dCREB2-b expression constructs, along with PKA and the reporter, resulted in a nearly complete block of PKA-dependent activation from the CRE-containing reporter.
The strong homology between the leucine zippers of dCREB2 (SEQ ID NO.: 3), CREB (SEQ ID NO.: 4) and CREM (SEQ ID NO.: 5) (see FIG. 1B ) suggested that mutations which abolish CREB dimerization (Dwarki, V. J. et al., EMBO J, 9: 225-232 (1990)) should also affect dCREB2 dimerization. The mutant Drosophila molecule mLZ-dCREB2-b was made by introducing two single-base changes that convert the middle two leucines of the leucine zipper to valines. An identical mutation in CREB abolishes homodimerization in vitro (Dwarki, V. J. et al., EMBO J., 9: 225-232 (1990)). Cotransfected mLZ-dCREB2-b failed to block PKA-dependent activation by dCREB2-a ( FIG. 4 ).
Example 2
Isolation and Characterization of dCREB1
A single cDNA representing the dCREB1 gene was isolated in the same screen of a Drosophila lambda gt11 expression library that yielded the dCREB2 cDNAs. The sequence of the dCREB1 cDNA contained a complete open reading frame specifying a 266 amino acid protein with a carboxyl-terminal leucine zipper four repeats long and an adjacent basic region ( FIG. 5 ; SEQ ID NO.: 7). The amino-terminal half of the inferred protein contains an acid-rich activation domain, with glutamate, asparate and proline residues spaced throughout. dCREB1 has consensus phosphorylation sites for CaM kinase II and PKC throughout its length, but has no predicted phosphoacceptor site for PKA.
Gel shift analysis showed higher-affinity binding of the dCREB1 protein to 3×CRE than to 3×mCRE. Transcriptional activation by dCREB1 was assayed with transient cotransfection experiments using the Drosophila L2 and Kc 167 cell culture lines. In L2 cells, dCREB1 activates transcription from CREs, but this effect is not enhanced by cotransfection of PKA ( FIG. 6 ). In Kc167 cells, dCREB1 fails to activate reporter expression either alone or with cotransfected PKA expression constructs.
Genomic Southern blot analysis indicates that dCREB1 is a single copy gene, and chromosomal in situ hybridization shows that it is located at 54A on the right arm of chromosome 2.
These results show that dCREB1 is a non-PKA responsive CREB family member from Drosophila.
The following materials and methods were used in the work described in Examples 3 and 4.
Isolating Transgenic Flies
EcoRI restriction sites were added (using PCR) just 5′ to the putative translation initiation site and just 3′ to the translation termination site in the dCREB2-b cDNA. This fragment was sequenced and subcloned into CaSpeR hs43, a mini-white transformation vector which contains the hsp70 promoter, in the orientation so that the dCREB2-b open reading frame is regulated by the hsp70 promoter. Germ-line transformation was accomplished using standard techniques (Spradling, A. C. and G. M. Rubin, Science, 218: 341-347 (1982); Rubin, G. M. and A. Spradling, Science, 218: 348-353 (1982)). Two transgenic lines, 17-2 and M11-1, each with one independent P-element insertion were generated and characterized. They appeared normal in general appearance, fertility and viability. These transgenic lines were outcrossed for at least five generations to w(CS10) (Dura, J-M., et al., J. Neurogenet., 9: 1-14 (1993)), which itself had been outcrossed for ten generations to a wild-type (Can-S) stock. This extensive series of outcrossing is necessary to equilibrate the genetic background to that of Canton-S. Flies homozygous for the 17-2 transgene were bred and used for all experiments.
The mutant blocker has been described previously (see Example 1). The mutations were substituted into an otherwise wild-type blocker construct and germ-line transformants were made by injecting into w(isoCJ1) embryoes. Flies homozygous for the A2-2 transgene insertion were bred and used for all experiments. w(isoCJ1) is a subline of w(CS10) (see above) carrying isogenic X, 2 nd and 3 rd chromosomes and was constructed by Dr. C. Jones in our laboratory. Originally 40 such sublines were w(CS10) using standard chromosome balancer stocks. Olfactory acuity, shock reactivity, learning and three-hr memory after one-cycle training then were assayed in each isogenic subline. As expected, a range of scores among the sublines was obtained. w(isoCJ1) yielded scores that were most like those of w(CS10) on each of these assays. By injecting DNA into the relatively homogeneous genetic background of w(isoCJ1), outcrossing of the resulting germ-line transformants to equilibrate heterogeneous) genetic backgrounds was not necessary.
Cycloheximide Feeding and Heat-Shock Regimen
For experiments on memory retention after one-cycle training and on retrograde amnesia, flies were fed 35 mM cycloheximide (+CMX; Sigma) in 4% sucrose (w/v) or 4% sucrose alone (−CXM) at 25° C. Groups of 100 flies were placed in feeding tubes (Falcon 2017) containing two 1.0×2.5 cm Whatmann 3MM filter paper strips that were soaked with a total of 250 μl of solution.
For experiments on one day retention after massed or spaced training, flies were fed 35 mM CXM and (w/v) 5% glucose dissolved in 3% ethanol. Groups of 100 flies were placed in feeding tubes (Falcon 2017) containing one 1.0×2.5 cm Whatmann 3MM filter paper strips that was soaked with a total of 126 μl of solution.
For experiments on learning after one-cycle training, olfactory acuity, and shock reactivity, flies were fed a 5% glucose, 3% ethanol solution alone or 35 mM CXM in the glucose/ethanol solution.
The feeding period was limited to 12-14 hrs prior to training, or to the 24-hr retention interval after training. Flies which were fed prior to training were transferred directly to the training apparatus after feeding, subjected to massed or spaced training, then transferred to test tubes containing filter paper strips soaked with 5% glucose during the 24-hr interval. Flies which were fed after training were trained, then transferred immediately to test tubes containing filter paper strips soaked with 5% glucose solution which was laced with 35 mM CXM. Flies remained in the test tubes for the duration of the 24-hr retention interval.
For heat-shock induction, flies were collected within two days of eclosion, placed in glass bottles in groups of about 600, and incubated overnight at 25° C. and 70% relative humidity. The next day, three hours before training, groups of approximately 100 flies were transferred to foam-stoppered glass shell vials containing a strip of filter paper to absorb excess moisture. The vials then were submerged in a 37° C. water bath until the bottom of the foam stopper (inside the vial) was below the surface of the water, thereby insuring that the flies could not escape heat-shock. The vial remained submerged for 30 min, after which the flies were transferred to a standard food vial for a 3-hr recovery period at 25° C. and 70% relative humidity. Training began immediately after the recovery period.
Pavlovian Learning and Memory and Testing
Flies were trained with an automated version of the learning procedure of Tully, T. and W. G. Quinn, J. Comp. Physiol., 157: 263-277 (1985). In brief, flies were trapped in a training chamber, the inside of which was covered with an electrifiable copper grid. Groups of about 100 flies were exposed sequentially to two odors [either octanol (OCT) or methylcyclohexanol (MCH)], which were carried through the training chamber in a current of air, for 60 seconds with 45 seconds rest intervals after each odor presenation. During exposure to the first odor, flies also were subjected to twelve 1.5-second pulses of 60 V DC with a 5-second interpulse interval. After training, flies were transferred to food vials for a particular retention interval. Conditioned odor-avoidance responses then were tested by transferring flies to the choice point of a T-maze, where they were exposed simultaneously to OCT and MCH carried in the distal ends of the T-maze arms and out the choice point on converging currents of air. Flies were allowed to distribute themselves in the T-maze arms for two minutes, after which they were trapped in their respective arms, anesthetized and counted. The “percent correct” then was calculated as the number of flies avoiding the shock-paired odor (they were in the opposite T-maze arm) divided by the total number of flies in both arms. (The number of flies left at the choice point, which usually was less than 5%, were not included in this calculation). Finally, a performance index (PI) was calculated by averaging the percent corrects of two reciprocal groups of flies—one where OCT and shock were paired, the other where MCH and shock were paired—and then by normalizing the average so that a PI=0 represented a 50:50 distribution in the T-maze and a PI=100 represented 100% avoidance of the shock-paired odor. For these studies, three different training protocols were used: 1. One-cycle training consisted of the training session just described. 2. Massed training consisted of 10 of these training cycles delivered one right after the other. 3. Spaced training consisted of 10 training cycles with a 15-min rest interval between each. One-cycle training was used to assay learning, while massed and spaced was used to assay consolidated memories.
Olfactory Acuity and Shock Reactivity
Odor avoidance responses to OCT or to MCH at two different concentrations—one (10 0 ) usually used in conditioning experiments and a 100-fold (10 −2 ) dilution thereof—were quantified in various groups of flies in the absence of heat shock and 3 hr or 24 hr after heat shock with the method of Boynton, S. and T. Tully, Genetics, 131: 655-672 (1992). Briefly, flies are placed in a T-maze and given a choice between an odor and air. The odors are naturally aversive, and flies usually choose air and avoid the T-maze arm containing the odor. For shock reactivity, flies are given a choice between an electrified grid in one T-maze arm, and an unconnected grid in the other. After the flies have distributed themselves, they are anesthetized, counted and a PI is calculated.
Statistical Analyses of Behavioral Data
Since each PI is an average of two percentages, the Central Limit Theorem predicts that they should be distributed normally (see Sokal, R. R. and F. J. Rohlf, Biometry, 2nd Edition, W.H. Freeman and Company, New York (1981)). This expectation was shown to be true by an empirical determination with data from Tully, T. and W. G. Quinn, J. Comp. Physiol., 157: 263-277 (1985) and Tully, T. and D. Gold, J. Neurogenet., 9: 55-71 (1993). Thus, untransformed (raw) data were analyzed parametrically with JMP2.1 statistical software (SAS Institute Inc., Cary N.C.). Since preliminary experiments preceded all of the experiments summarized herein, all pairwise comparisons were planned. To maintain an experimentwise error rate of alpha=0.05, the critical P values for these individual comparisons were adjusted accordingly (Sokal, R. R. and F. J. Rohlf, Biometry, 2nd Edition, W.H. Freeman and Company, New York (1981)) and are listed below for each experiment.
All experiments were designed in a balanced fashion with N=2 PIs per group collected per day; then replicated days were added to generate final Ns. In each experiment, the experimenter (M.D.) was blind to genotype.
A. One-day memory in wild-type flies fed CXM before or immediately after massed or spaced training ( FIG. 8 ): PIs from these four drug treatments (−CXM before, −CXM after, +CXM before and +CXM after) and two training procedures (massed and spaced) were subjected to a TWO-WAY ANOVA with DRUG (F (3,56) =8.93; P<0.001) and TRAINing (F (1,56) =18.10, P<0.001) as main effects and DRUG×TRAIN (F (3,56) =4.68, P=0.006) as the interaction term. P values from subsequent planned comparisons are summarized in FIG. 8 . The six planned comparisons were judged significant if P≦0.01.
B. One-day memory after massed or spaced training in dCREB2-b transgenic flies ( FIGS. 9A and 9B ): In experiments with the 17-2 transgenic line, PIs from two strains (Can-S and 17-2) and four training-regimens (spaced −hs, spaced +hs, massed −hs and massed +hs) were subjected to a TWO-WAY ANOVA with STRAIN (F (1,40) =1.57; P=0.22) and TRAINing-regimen (F (3,40) =25.81, P<0.001) as main effects and STRAIN×TRAIN (F (3,40) =6.62, P=0.001) as the interaction term. A similar analysis was done with data from the M 11-1 transgenic line, yielding STRAIN (F (1,40) =2.81; P=0.10), TRAINing-regimen (F (3,40) =11.97, P<0.001) and STRAIN×TRAIN (F (3,40) =3.37, P=0.03) effects. P values from subsequent planned comparisons are summarized in FIGS. 9A and 9B . In each experiment, the seven planned comparisons were judged significant if P≦0.01.
C. Learning after one-cycle training in 17-2 transgenic flies ( FIG. 9C ): PIs from two strains (Can-S and 17-2) and three heat-shock regimens [−hs, +hs (3 hr) and +hs (24 hr)] were subjected to a TWO-WAY ANOVA with STRAIN (F (1,30) =0.69; P=0.41) and HEAT-shock regimen (F (2,30) =10.29, P<0.001) as main effects and STRAIN×HEAT (F (2,30) =0.71, P=0.50) as the interaction term. P values from subsequent planned comparisons are summarized in FIG. 9C . The three planned comparisons were judged significant if P≦0.02.
D. One-day memory after spaced training in A2-2 transgenic flies ( FIG. 10 ): PIs from these three strains [w(isoCJ1), 17-2 and A2-2] and two heat-shock regimens [−hs and +hs (3 hr)] were subjected to a TWO-WAY ANOVA with STRAIN (F (2,30) =9.43, P<0.001) and HEAT-shock regimen (F (1,30) =9.84, P=0.004) as main effects and STRAIN×HEAT (F (2,30) =5.71, P=0.008) as the interaction term. P values from subsequent comparisons are summarized in FIG. 10 . The six planned comparisons were judged significant if P≦0.01.
E. Olfactory acuity in 17-2 flies (Table): PIs from these two strains (Can-S and 17-2), four different odor-levels (OCT-10 0 , OCT-10 −2 , MCH-10 0 and MCH-10 −2 ) and three heat-shock regimens [−hs. +hs (3 hr) and +hs (24 hr)] were subjected to a THREE-WAY ANOVA with STRAIN (F (1,184) =0.12, P=0.73), ODOR-level (F (3,184) =126.77, P<0.001) and HEAT-shock regimen (F (2,184) =3.55, P=0.03) as main effects, STRAIN×ODOR (F (3,184) =1.23, P=0.30), STRAIN×HEAT (F (2,184) =0.33, P=0.72) and ODOR×HEAT (F (6,184) =3.14, P=0.006) as two-way interaction terms and STRAIN×ODOR×HEAT (F (6,184) =0.48, P=0.83) as the three-way interaction term. P values from subsequent planned comparisons are summarized in the Table. The twelve planned comparisons were significant if P≦0.005.
F. Shock reactivity in 17-2 flies (Table): PIs from these two strains (Can-S and 17-2), two shock groups (60V and 20V) and three heat−shock regimens [−hs, +hs (3 hr) and +hs (24 hr)] were subjected to a THREE-WAY ANOVA with STRAIN (F (1,84) =0.50, P=0.48), SHOCK (F (1,84) =97.78, P<0.001) and HEAT-shock regimen (F (2,84) =3.36, P=0.04) as main effects, STRAIN×SHOCK (F (1,84) =1.12, P=0.29), STRAIN×HEAT (F (2,84) =1.06, P=0.35) and SHOCK×HEAT (F (2,84) =6.66, P=0.002) as two-way interaction terms and STRAIN×SHOCK×HEAT (F (2,84) =1.75, P=0.18) as the three-way interaction term. P values from subsequent planned comparisons are summarized in the Table. The six planned comparisons were judged significant if P≦0.01.
G. Seven-day memory after spaced training in 17-2 flies ( FIG. 11 ): PIs from two strains (Can-S and 17-2) and two heat-shock regimens [−hs and +hs(3 hr)] were subjected to a TWO-WAY ANOVA with STRAIN (F (1,20) =6.09; P=0.02) and HEAT-shock regimen (F (1,20) =16.30, P=0.001) as main effects and STRAIN×TRAIN (F (1,20) =7.73, P=0.01) as the interaction term. P values from subsequent planned comparisons are summarized in FIG. 11 . The three planned comparisons were judged significant if P≦0.02.
H. One-day memory after spaced training in rsh; 17-2 double mutants ( FIG. 12 ): PIs from three strains (17-2, rsh and rsh; 17-2) and two heat-shock regimens [−hs and +hs (3 hr)] were subjected to a TWO WAY ANOVA with STRAIN (F (2,30) =32.05; P<0.001) and HEAT-shock regimen (F (1,30) =59.68, P<0.001) as main effects and STRAIN×TRAIN (F (2,30) =11.59, P<0.001) as the interaction term. P values from subsequent planned comparisons are summarized in FIG. 12 . The five planned comparisons were judged significant if P≦0.01.
I. Learning after one-cycle training in rsh; 17-2 mutants (see text): PIs from these two strains (Can-S and rsh; 17-2) and two heat-shock regimens [−hs and +hs (3 hr)] were subjected to a TWO-WAY ANOVA with STRAIN (F (1,20) =86.85, P<0.001) and HEAT-shock regimen (F (1,20) =0.02, P<0.89) as main effects and STRAIN×HEAT (F (1,20) =0.86, P=0.37) as the interaction term. P values from subsequent planned comparisons are summarized in the Table. The two planned comparisons were significant if P≦0.03.
J. Olfactory acuity in rsh; 17-2 flies (Table): PIs from these two strains (Can-S and rsh; 17-2), four different odor-levels (OCT-10 0 , OCT-10 −2 , MCH-10 0 and MCH-10 −2 ) and two heat-shock regimens [−hs, and +hs (3 hr)] were subjected to a THREE-WAY ANOVA with STRAIN (F (1,112) =0.02, P=0.88), ODOR-level (F (3,112) 50.03, P<0.001) and HEAT-shock regimen (F (1,112) =29.86, P<0.001) as main effects, STRAIN×ODOR (F 3,112) =2.15, P=0.10), STRAIN×HEAT (F (1,112) =0.34, P=0.56) and ODOR×HEAT (F (3,12) =6.41, P=0.001) as two-way interaction terms and STRAIN×ODOR×HEAT (F (3,112) 1.12, P=0.35) as the three-way interaction term. P values from subsequent planned comparisons are summarized in the Table. The eight planned comparisons were judged significant if P≦0.01.
K. Shock reactivity in rsh; 17-2 flies (Table): PIs from these two strains (Can-S and rsh; 17-2), two shock groups (60V and 20V) and two heat-shock regimens [−hs and +hs (3 hr)] were subjected to THREE-WAY ANOVA with STRAIN (F (1,56) =0.51, P=0.48), SHOCK (F (1,56) =88.14, P<0.001) and HEAT-shock regimen (F (1,56) =0.08, P=0.77) as main effects, STRAIN×SHOCK (F (1,56) =0.12, P=0.73, STRAIN×HEAT (F (1,56) =0.03, P=0.86) and SHOCK×HEAT (F (1,56) =0.39, P=0.53) as two-way interaction terms and STRAIN×SHOCK×HEAT (F (1,84) =1.58, P=0.21) as the three-way interaction term. P values from subsequent planned comparisons are summarized in the Table. The four planned comparisons were judged significant if P≦0.01.
Northern Analysis
For RNA collection, the heat-shock regimen was the same as for behavioral experiments. For any indicated time interval between heat-shock and collection, flies rested in food-containing vials at 25° C. Flies were collected and quickly frozen in liquid nitrogen. All Northern analyses used head RNA. The tube of frozen flies was repeatedly rapped sharply on a hard surface, causing the heads to fall off. The detached frozen heads were recovered by sieving on dry ice. Approximately 1000 heads were pooled for RNA preparation. Wild-type and transgenic flies for each individual time point always were processed in parallel. Flies that were not induced were handled in a similar manner to induced flies, except that the vials were not placed at 37° C. Total head RNA was isolated from each group of flies, and poly A+ RNA was isolated using oligo dT columns according to the instructions of the manufacturer (5′−−−>3′ Inc.). The concentration of poly A+ mRNA was measured spectrophotometrically, and 0.5 mg of mRNA per lane was loaded and run on 1.2% formaldehyde-agarose gels. Northern blots were prepared, probed and washed (0.1×SSC at 65° C.) as described (Ausubel, F., Current Protocols in Molecular Biology , John Wiley and Sons, New York, 1994). For detection of the transgene, an 843 bp dCREB2-b cDNA fragment was subcloned into pKS+ and used to generate a uniformly-labeled antisense riboprobe. This fragment codes for the carboxyl-terminal 86 amino acids of the dCREB2-b protein plus 3′ untranslated mRNA.
Western Blot Analysis and Antiserum
Western blot analysis was performed using a rabbit antiserum raised against a peptide corresponding to 16 amino acids in the basic region of the dCREB2-b cDNA with an additional COOH terminal Cys. The sequence of the peptide was: (SEQ ID NO.: 24) NH2-RKREIRLQKNREAAREC-COOH. The peptide was synthesized and coupled to Sulfo-SMCC (Pierce) activated keyhole lympet hemocyanin. The antigen was injected into rabbits (100 μg) and boosted at two week intervals. Sera was bled and tested for immune reactivity towards bacterially-expressed dCREB2-b protein. The antiserum was passed through a CM Affi-gel Blue column (Biorad), and the flow-through was concentrated by ammonium sulfate precipitation, resuspended and dialyzed against PBS (0.14 M NaCl, 2.7 mM KCl, 4.3 mM Na 2 HPO 4 7H 2 0, 1.4 mM KH 2 P0 4 , pH 7.3). The dialyzed serum was affinity-purified using a peptide column made using an Ag/Ab Immobilization kit (Immunopure from Pierce). After the antiserum was eluted using a 4 M MgCl 2 , 0.1 M HEPES pH 6.0 buffer, it was dialyzed into PBS and frozen.
Each data point represents approximately 5 fly heads. Groups of about 25-50 flies were collected and quickly frozen on liquid nitrogen until all of the time points had been collected. Heads were isolated resuspended in approximately 200 μl of 1× Laemmli sample buffer, allowed to thaw and homogenized with a Dounce type B pestle. Samples were boiled for 5 minutes, and centrifuged for 10 minutes at room temperature in an Eppendoff microcentrifuge. The supernatants were collected and boiled again just prior to loading onto protein gels. Standard procedures were used to separate equal amounts of proteins from each sample on 12% polyacrylamide-SDS gels and to transfer them to PVDF membranes by electroblotting (Ausubel, F., Current Protocols in Molecular Biology , John Wiley and Sons, New York, 1994).
The membranes were blocked for 60 minutes with a 5% BSA solution made up in TBST (10 mM Tris, pH 7.9, 150 mM NaCl, 0.05% Tween 20). The primary antibody was diluted 1:1000 in TBST and incubated with the filter for 30 minutes. The membranes were washed three times with TBST for 5 minutes each time and then incubated for 30 minutes with an alkaline phosphatase-conjugated anti-rabbit IgG second antibody (Promega) diluted 1:7500 in TBST. The membranes were washed three more times as before and developed using a chromogenic alkaline phosphatase reaction according to manufacturers suggestions (Promega).
Example 3
Transgene Expression Increased After Heat-Shock Induction
In order to interpret the effects of transgene induction on behavior, dCREB2-b expression in transgenic flies (17-2) after heat-shock induction was measured. Northern blot analysis revealed elevated levels of hs-dCREB2-b message in the 17-2 flies immediately and three hours after heat-shock ( FIG. 7A ). This induction was also detectable in brain cells using in situ hybridization. Western blot showed increased dCREB2-b protein immediately after induction ( FIG. 7B ). Elevated levels of the dCREB2-b protein were seen nine hours later and were still detectable twenty four hours after induction. These data indicate that increased amounts of dCREB2-b existed in brain cells throughout spaced training, which ended about six hours after heat induction.
The behavioral experiments also used transgenic flies (A2-2) which expressed a mutated dCREB2-b protein (dCREB2-mLZ). These mutations changed the two internal leucine residues of the leucine zipper to valine residues, and these changes have been shown to result in a protein which is unable to form dimers (Dwarki, V. J. et al., EMBO J., 9: 225-232 (1990)). In transient co-transfection assays, the mutant protein was unable to block PKA-dependent transcription mediated by dCREB2-a, while the wild-type protein had blocking function. Western blot analysis showed that the wild-type and mutant blocker are expressed at similar levels beginning immediately after heat-shock induction and lasting for at least 6 hours ( FIG. 7C ). Therefore, it is unlikely that these two proteins have large differences in expression levels or stability in the transgenic flies.
Northern blot analysis of two different housekeeping genes, myosin light chain (Parker, V. P., et al., Mol. Cell Biol., 5: 3058-3068 (1985)) and elongation factor α (Hovemann, B., et al., Nucleic Acids Res., 16: 3175-3194 (1988)), showed that steady-state levels of their RNAs were unaffected after transgene induction for at least 3 hours. Gel shift analysis using two different consensus DNA binding sites showed that there was no large effect on the gel shift species which formed after transgene induction for at least 9 hours. Cotransfection of the blocker did not interfere with the activity of a transcription factor from a different family in cell culture. Considered together, hs-dCREB2-b probably had fairly specific molecular modes of action after induction.
Example 4
Assessment of the Role of CREBs in Long-Term Memory Formation
Flies were fed 35 mM cycloheximide (CXM) for 12-14 hours before, or for the 24-hr retention interval immediately after, massed or spaced training ( FIG. 8 ). Each of these CXM feeding regimens significantly reduced one-day memory after spaced training but had no effect on one-day memory after massed training ( FIG. 8 ). Thus, cyclohexmide feeding immediately before or after spaced training disrupts one-day memory. These results suggest that protein synthesis is required soon after training for the formation of long-lasting memory.
The results in FIG. 8 show that cycloheximide feeding affects one-day retention after spaced training but not massed training. Different groups of wild-type (Can-S) flies were fed 5% glucose solution alone (hatched bars) or laced with 35 mM CXM (striped bars) either for 12-14 hr overnight before massed or spaced training or for the 24-hr retention interval immediately after training. One-day memory retention was significantly lower than normal in flies fed CXM before (P<0.001) or after (P<0.001) spaced training. In both cases, one-day retention in CXM-fed flies was reduced to a level similar to one-day memory after massed training in glucose-fed flies (P=0.88 for CXM before training and P=0.71 for CXM after training). In contrast, no difference was detected between CXM-fed and control flies for one-day memory after massed training (P=0.49 and P=0.46, respectively).
One day retention after spaced training was unaffected in uninduced (−hs) transgenic flies (17-2) but was significantly reduced in induced (+hs) transgenic flies ( FIG. 9A ). In contrast, one-day retention after massed training was normal in both uninduced and induced transgenic flies ( FIG. 9A ). Comparisons of one-day retention after spaced or massed training between wild-type flies with (+hs) or without (−hs) heat-shock indicated that the heat-shock regimen itself did not have a non-specific effect on memory after either training protocol. Thus induction of the dCREB2-b transgene only affected (i.e., disrupted) one-day memory after spaced training.
One day retention after spaced or massed training in M11-1, a second line carrying an independent hs-dCREB2-b insertion, also was tested. Results with M11-1 were similar to those obtained with 17-2 ( FIG. 9B ). These results show that the effect of induced hs-dCREB2-b does not depend on any particular insertion site of the transgene.
The results in FIGS. 9A-9C show that induction of the dCREB2-b transgene disrupts one-day memory after spaced training, while one-day memory after massed training and learning are normal.
In FIG. 9A , different groups of wild-type (Can-S) flies (hatched bars) or hs-dCREB2-b transgenic (17-2) flies (striped bars) were given spaced training or massed training in the absence of heat shock (−hs) or three hours after heat shock (+hs). After training, flies were transferred to standard food vials and stored at 18° C. until one-day memory was assayed. No differences in one-day memory after spaced or massed training were detected between Can-S vs. 17-2 flies in the absence of heat shock (−hs; P=0.83 and 0.63, respectively). When flies were trained three hours after heat shock (+hs), however, one-day memory was significantly different between Can-S vs. 17-2 flies after spaced training (P<0.001) but not after massed training (P=0.23). In fact, the one-day memory after spaced training was no different than that after massed training in induced 17-2 flies (P=0.59). In addition, the heat-shock regimen did not produce a non-specific effect on one-day retention after spaced (P=0.59) or massed (P=1.00) training in Can-S flies. N=6 performance indices (PIs) per group.
The experiment described in FIG. 9A was repeated in FIG. 9B with a second, independently derived dCREB2-b transgenic line, M11-1 (striped bars). Here again, a) no differences in one-day memory after spaced or massed training were detected between Can-S vs. M11-1 flies in the absence of heat-shock (−hs; P=0.83 and 0.86, respectively), b) a significant difference between Can-S v. M11-1 for one-day memory after spaced training (P<0.001) but not after massed training (P=0.85) when trained three hours after heat-shock (+hs), c) one-day memory after spaced training was no different than that after massed training in induced M11-1 flies (P=0.43) and d) the heat-shock regimen did not produce a non-specific effect on one-day retention after spaced (P=0.59) or massed (P=0.94) training in Can-S flies. N=6 PIs per group.
If induction of the transgene specifically affected LTM via disruption of gene expression, then learning should not be affected, since it does not require new protein synthesis. Different groups of flies were trained using one-cycle training either without heat-shock, or three or twenty four hours after heat-shock. These time points after induction were selected to correspond to the times when flies were trained and tested in the previous experiments (see FIGS. 9A and 9B ). Induction of the transgene (dCREB2-b) in the 17-2 line had no effect on learning in either case ( FIG. 9C ).
In FIG. 9C , different groups of Can-S flies (hatched bars) or 17-2 transgenic flies (striped bars) received one-cycle training in the absence of heat shock (−hs) or three (+hs 3 hr) or 24 (+hs 24 hr) hours after heat-shock and then were tested immediately afterwards. In each case, no differences between Can-S vs. 17-2 flies were detected (Ps=0.28, 0.64 and 0.42, respectively), indicating that learning was normal in induced or uninduced transgenic flies. N=6 PIs per group.
Induction of the transgene which contained the mutant blocker (A2-2) did not affect one-day retention after spaced training, while the wild-type blocker (17-2) had a dramatic effect ( FIG. 10 ). The w(iso CJ1) flies, whose one-day retention also was unaffected by heat induction, is the isogenic control for the mutant blocker transgenic flies. Since Western blot analysis showed that wild-type and mutant blockers probably have similar expression levels, this result suggests that the blocker requires an intact leucine zipper to function effectively.
FIG. 10 shows that induction of the hs-dCREB2-mLZ mutant blocker does not affect one-day retention after spaced training. Different groups of wild-type [w(isoCJ1)], hs-dCREB2-b transgenic (17-2) or mutant hs-dCREB2-mLZ transgenic flies (A2-2) received spaced training in the absence of heat-shock (−hs) or three hours after heat-shock (+hs). The flies were then handled and tested as in FIG. 9A . No differences in one-day memory after spaced training were detected between w(isoCJ1) vs. 17-2 flies or between w(isoCJ1) vs. A2-2 flies in the absence of heat shock (−hs; P=0.38 and 0.59, respectively). When flies were trained three hours after heat shock (+hs), however, one-day memory after spaced training was significantly different between w(isoCJ1) vs. 17-2 flies (P<0.001)—as in FIG. 9 A—but was not different between w(isoCJ1) vs. A2-2 flies (P=0.78). In addition, the heat-shock regimen did not produce a non-specific effect on one-day retention after spaced training in w(isoCJ1) or A2-2 flies (P=0.40 and P=0.97, respectively. N=6 performance indices (PIs) per group.
Olfactory acuity and shock reactivity are component behaviors essential for flies to properly learn odor-shock associations. The Table shows the scores for these peripheral behaviors for Can-S versus 17-2 flies. With or without heat-shock, olfactory acuity and shock reactivity were normal in 17-2 transgenic flies. Thus, hs-dCREB2-b induction does not affect olfactory acuity or shock reactivity.
If induction of hs-dCREB2-b blocks long-term memory (LTM), then long-lasting memory also should be blocked. In wild-type flies, seven-day retention after spaced training consists solely of the CXM-sensitive LTM because the CXM insensitive ARM component has decayed away. In uninduced transgenic flies (17-2), seven-day retention after spaced training was similar to retention in uninduced wild-type flies (P=0.83; FIG. 11 ). Seven-day retention was severely disrupted, however, in transgenic flies which were trained three hours after heat-shock (P=0.001) and did not differ from zero (P=0.91). In contrast, the heat-shock protocol had no detectable effect on seven-day memory in wild-type flies (P=0.39). Thus, induction of hs-dCREB2-b disrupts long-term memory (LTM).
FIG. 11 shows that induction of hs-dCREB2-b completely abolishes 7-day memory retention. Previous analyses of radish mutants indicated that memory retention four or more days after spaced training reflects the sole presence of LTM. Thus, the effect of induced hs-dCREB2-b on LTM was verified by comparing 7-day retention after spaced training in Can-S (hatched bars) vs. 17-2 transgenic (striped bars) flies that were trained in the absence of heat-shock (−hs) or three hours after heat shock (+hs). Flies were stored in standard food vials at 18° C. during the retention interval. N=6 PIs per group. Seven-day retention after spaced training did not differ between Can S and 17-2 in the absence of heat-shock (P=0.83) but was significantly lower than normal in 17-2 flies after heat-shock (P=0.002). In fact, 7-day retention after spaced training in induced 17-2 transgenic files was not different from zero (P=0.92). In addition, the heat-shock regimen did not affect 7-day retention after spaced training non-specifically in Can-S flies (P=0.39).
If induction of the hs-dCREB2-b transgene specifically blocks LTM, then it should only affect the CXM-sensitive component of consolidated memory after spaced training. For both transgenic lines, 17-2 and M11-1, the effect of transgene induction looked similar to the effect that CXM had on wild-type flies (compare FIG. 8 with FIGS. 9A and 9B ). This similarity suggested that the induced dCREB2-b protein completely blocked CXM-sensitive memory, leaving ARM intact. The radish mutation disrupts ARM (Folkers, E., et al., Proc. Natl. Acad. Sci. USA, 90: 8123-8127 (1993)), leaving only LTM one day after spaced training. Thus, a radish hs-dCREB2-b “double mutant” (rsh; 17-2) was constructed to allow examination of LTM in the absence of ARM. In the absence of heat-shock, rsh; 17-2 double-mutants and radish single-gene mutants yielded equivalent one-day retention after spaced training ( FIG. 12 ). In contrast, when these flies were heat-shocked three hours before spaced training, one-day retention was undetectable in rsh; 17-2 flies but remained at mutant levels in radish flies. The double mutant also showed normal (radish-like) learning (P=0.59) and normal (wild-type) olfactory acuity and shock reactivity in the absence of heat-shock versus three hours after heat shock (see the Table).
FIG. 12 shows that induction of hs-dCREB2-b completely abolishes one-day memory after spaced training in radish; 17-2 “double mutants.” Since radish is known to disrupt ARM, a clear view of the effect of hs-dCREB2-b on LTM was obtained in radish; 17-2 flies. One-day retention after spaced training was assayed in rsh; 17-2 double mutants and in 17-2 and rsh single-gene mutants as controls. Flies were trained in the absence of heat-shock (hatched bars) or three hours after heat-shock (striped bars) and stored at 18° C. during the retention interval. As usual, induction of hs-dCREB2-b produced significantly lower one-day memory after spaced training in 17-2-flies (P<0.001). The heat-shock regimen, however, had no effect on such memory in radish mutants (P=0.52), which reflects only the presence of LTM. In contrast, heat-shock produced significantly lower scores in rsh; 17-2 double mutants (P<0.001), which were not different from zero (P=0.20). N=6 PIs per group.





TABLE



Olfactory acuity and shock reactivity in Can-S (wild-type), 17-2 (hs-dCREB2-b

transgenic) and rsh; 17-2 (radish, hs-dCREB2-b “double mutant”) flies a






Olfactory Acuity
Shock







OCT
MCH
Reactivity










Heat Shock
Group
10 0 *
10 −2
10 0 *
10 −2
60 V
20 V



−hs
CAN-S
58 ± 3
32 ± 3
80 ± 2
33 ± 7
79 ± 5
52 ± 5


17-2
60 ± 3
34 ± 8
77 ± 3
37 ± 5
87 ± 3
43 ± 2

+hs
CAN-S
69 ± 4
41 ± 4
77 ± 2
25 ± 9
74 ± 5
58 ± 6

(3 hrs)
17-2
71 ± 4
37 ± 3
76 ± 5
26 ± 3
78 ± 3
67 ± 5

+hs
CAN-S
66 ± 2
56 ± 8
79 ± 4
33 ± 2
84 ± 3
63 ± 3

(24 hr)
17-2
65 ± 3
42 ± 6
76 ± 3
41 ± 5
85 ± 2
60 ± 6

−hs
CAN-S
51 ± 4
39 ± 5
72 ± 5
33 ± 7
87 ± 3
52 ± 5


rsh; 17-2
57 ± 3
39 ± 5
74 ± 5
29 ± 4
82 ± 4
53 ± 6

+hs
CAN-S
72 ± 4
48 ± 3
66 ± 2
60 ± 3
80 ± 4
58 ± 6

(3 hr)
rsh; 17-2
68 ± 4
46 ± 6
78 ± 2
49 ± 4
83 ± 1
50 ± 5



a Olfactory acuity and shock reactivity were assayed in untrained flies with the methods of Boynton, S. and T. Tully, Genetics, 131: 655-672 (1992) and Dura, J-M., et al., J. Neurogenet., 9: 1-14 (1993), respectively (see Examples for more details). N = 98 PIs per group. Planned comparisons between Can-S vs. mutant flies failed to detect any significant differences with any heat-shock regimen.

*10 0 is manual concentration and corresponds to 10 −3 for bubbler.







































The following materials and methods were used in the work described in Examples 5 and 6.
Pavlovian Learning and Memory and Testing
During one training session, a group of about 100 flies was exposed sequentially to two odors [either octanol (OCT) or methylcyclohexanol (MCH)] for 60 seconds with 45-second rest intervals after each odor presentation. During exposure to the first odor, flies received twelve 1.5-second pulses of 60 V DC with a 5-second interpulse interval.
After training, flies were transferred to food vials and stored at 18° C. for a seven-day retention interval. Conditioned odor-avoidance responses then were tested by transferring files to the choice point of a T-maze, where they were exposed simultaneously to OCT and MCH carried on converging currents of air in the distal ends of the T-maze arms and out the choice point.
Flies were allowed to distribute themselves in the T-maze arms for 120s, after which they were trapped in their respective arms, anesthetized and counted. The “percent correct” then was calculated as the number of flies avoiding the shock-paired odor (they were in the opposite T-maze arm) divided by the total number of flies in both arms. (The number of flies left at the choice point, which usually was less than 5%, were not included in this calculation.) Finally, a performance index (PI) was calculated by averaging the percent corrects of two reciprocal groups of flies—one where OCT and shock were paired, the other where MCH and shock were paired—and then by normalizing the average so that a PI=0 represented a 50:50 distribution in the T-maze and a PI=100 represented 100% avoidance of the shock-paired odor.
All behavioral experiments were designed in a balanced fashion with N=2 PIs per group collected per day; then replicated across days to generate final Ns. In all experiments, the experimenter was blind to genotype.
Statistical Analyses of Behavior Data
PIs are distributed normally (Tully, T. and D. Gold, J. Neurogenet., 9: 55-71 (1993)). Consequently, untransformed (raw) data were analyzed parametrically with JMP3.01 statistical software (SAS Institute Inc., Cary N.C.). Negative accelerating exponential Gompertz (growth) functions (see Lewis, D., Quantitative Methods in Psychology , McGraw-Hill, New York, N.Y. (1960)) were fit to the data in FIGS. 13A and 13B via nonlinear least squares with iteration.
Example 5
Effect on Long Term Memory of Repeated Training Sessions
Seven-day memory retention (a measure of long term memory) in wild-type (Can-S) flies is induced incrementally by repeated training sessions. An automated version of a discriminative classical conditioning procedure was used to electroshock flies during exposure to one odor (CS+) but not to a second odor (CS−). Seven days after one or more training sessions, memory retention of conditioned odor avoidance responses was quantified in a T-maze, where flies were presented the CS+ and CS-simultaneously for 120 seconds.
In FIG. 13A , long term memory as a function of the number of training sessions is indicated by open circles. One training session produced a mean performance index (PI±SEM) near zero. Additional training sessions with a 15-minute rest interval between each, however, yielded a steady increase in mean PIs with a maximum of 39 after ten training sessions. Ten additional training sessions produced similar performance. A nonlinear “growth” function (solid line) was fit to the individual PIs using an iterative least squares method. N=13, 6, 6, 6, 13, 7, 7, 7, 7, 6, 7 and 7 PIs for groups receiving 1-10, 15 and 20 training sessions, respectively.
Example 6
Effect on Long Term Memory of the Rest Interval between Each Training Session
Seven-day memory retention (a measure of long term memory) in wild-type (Can-S) flies is induced incrementally by the rest interval between each training session. As described in Example 5, an automated version of a discriminative classical conditioning procedure was used to electroshock flies during exposure to one odor (CS+) but not to a second odor (CS−). Seven days after one or more training sessions, memory retention of conditioned odor avoidance responses was quantified in a T-maze, where flies were presented the CS+ and CS− simultaneously for 120 seconds.
In FIG. 13B , long term memory as a function of the rest interval is indicated by open circles. Ten training sessions with no rest interval between each (massed training) produced a mean PI near zero. Increasing the rest interval between each of ten training sessions yielded a steady increase in mean PIs with a maximum of 34 for a 10-minute rest interval. Rest intervals up to ten minutes longer produced similar performance. A nonlinear growth function (solid line) was fit to the data as above. N=12, 6, 6, 6, 6, 13, 7, 7, 7, 7, 7, 7 and 7 PIs for groups receiving 0-10, 15 and 20 minutes of rest between each training session.
The following materials and methods were used in the work described in Examples 7-10.
Isolating Transgenic Flies
EcoRI restriction sites were added (using PCR) just 5′ to the putative translation initiation site and just 3′ to the translation termination site in the dCREB2-a cDNA. This fragment was sequenced and subcloned into CaSpeR hs43, a mini-white transformation vector which contains the hsp70 promoter, in the orientation so that the dCREB2-a open reading frame is regulated by the hsp70 promoter. Germ-line transformation was accomplished by injecting into isogenic w(isoCJ1) embryoes using standard techniques (Spradling, A. C. and G. M. Rubin, Science, 218: 341-347 (1982); Rubin, G. M. and A. Spradling, Science, 218: 348-353 (1982)). By injecting DNA into the relatively homogeneous genetic background of w(isoCJ1), outcrossing of the resulting germ-line transformants to equilibrate (heterogeneous) genetic backgrounds was not necessary. Two transgenic lines, C28 and C30, each with one independent P-element insertion were generated and characterized. They appeared normal in general appearance, fertility and viability. Flies homozygous for the C28 or C30 transgene were bred and used for all experiments.
Heat Shock Regimen
For heat-shock induction, flies were collected within two days of eclosion, placed in glass bottles in groups of about 600, and incubated overnight at 25° C. and 70% relative humidity. The next day, three hours before training, groups of approximately 100 flies were transferred to foam-stoppered glass shell vials containing a strip of filter paper to absorb excess moisture. The vials then were submerged in a 37° C. water bath until the bottom of the foam stopper (inside the vial) was below the surface of the water, thereby insuring that the flies could not escape heat-shock. The vial remained submerged for 30 minutes, after which the flies were transferred to a standard food vial for a 3-hr recovery period at 25° C. and 70% relative humidity. Training began immediately after the recovery period.
Statistical Analyses of Behavior Data
PIs from the three strains (Can-S, C28 and C30) and six training-regiments (1×+hs, 2×massed+hs, 10× massed +hs, 1×−hs, 2× massed −hs and 10× massed −hs) were subjected to a TWO-WAY ANOVA with STRAIN (F (2,102) =48.34; P<0.001) and TRAINing-regimen (F (5,102) =25.47, P<0.001) as main effects and STRAIN×TRAIN (F (10,102) =5.85, P<0.001) as the interaction term. Since preliminary experiments preceded all of the experiments summarized herein, all pairwise comparisons were planned. To maintain an experimentwise error rate of alpha=0.05, the individual comparisons summarized in FIG. 15B were judged significant if P<0.002 (Sokal, R. R. and F. J. Rohlf, Biometry, 2nd Edition, W.H. Freeman and Company, New York (1981)).
Example 7
A Molecular Switch for the Formation of Long Term Memory
FIG. 14 presents a conceptual method of a molecular switch for the formation of LTM, based on differential regulation of CREB isoforms with opposing functions.
Immediately after one training session, the relevant CREB activators and repressors are induced. Their combined functions (rather than molecular concentrations) are equivalent and yielded no net effect of CREB activators. Thus, repeated sessions of massed training (no rest interval) never induce LTM (see FIG. 15A ). CREB repressors functionally inactivate faster than CREB activators, however, yielding an increasing net effect of CREB activators (ΔC) with time (see FIG. 13B ). If ΔC is positive at the end of a particular rest interval during spaced training, then CREB activators are free to initiate downstream events involved with the formation of LTM. Usually, ΔC after one training session is not large enough to yield much LTM. Thus, repeated spaced training sessions serve to increase ΔC incrementally eventually to produce maximal LTM (see FIG. 13A ).
Experimental verification of two predictions from this model involving CREB as a molecular switch for long term memory formation is shown in FIGS. 15A-15C and discussed in Examples 8-10.
Example 8
Effect on Long Term Memory of Having Equal Amounts of CREB Activators and Repressors Immediately After One Training Session
The model of a molecular switch for LTM predicts that the functional effects of all CREB activators and repressors are equal immediately after one training session (ΔC=0). If no rest interval occurred between additional training sessions (massed training), then functional CREB activator would not accumulate, thereby preventing the induction of downstream events required for LTM induction.
To test this notion, wild-type (Can-S) flies were subjected to 48, instead of the usual 10 (see FIG. 15B ), massed training sessions (48× massed) or, as a positive control, to 10 spaced training sessions with a 15-minute rest interval (10× spaced). Seven-day memory after such massed training was near zero ( FIG. 15A ), while that after spaced training was near its usual maximum value (see FIG. 13A ). Thus, nearly five times the usual amount of massed training still did not induce LTM. N=6 PIs for each group.
PIs from two groups (10× spaced or 48× massed) of wild-type (Can-S) files were subjected to a ONE-WAY ANOVA with GROUP (F [10] =51.13; P<0.001) as the main effect. A subsequent planned comparison revealed that the mean PI of the 48× massed group did not differ significantly from zero (t[ 10]=1.66 ; P=0.127).
Example 9
Effect on Long Term Memory of Increasing Amounts of CREB Activator
The model of a molecular switch for LTM predicts that experimentally increasing the amount of CREB activator will eliminate the requirements for at least 10 repeated training sessions with a 10-minute rest interval between each to produce maximal LTM.
To test this idea, two transgenic lines (C28 and C30) carrying an inducible hsp-dCREB2-a activator construct inserted into different cytological locations were generated. Different groups of flies from these two transgenic lines were subjected, along with wild-type (Can-S) files, to 1 (1×), 2 (2×) or (10×) massed training sessions three hours after heat-shock induction of the transgene (induced) or in the absence of heat-shock (uninduced).
Without heat-shock, seven-day memory in all three strains did not differ from zero after one, two or ten massed training sessions (all Ps>0.002). With heat-shock, seven-day memory in wild-type flies remained near zero in each massed training group (all Ps>0.002). In contrast, seven-day memory was significant (near the maximum of 35) after ten massed sessions in both the C28 and C30 transgenic lines (all Ps<0.0001). Moreover, seven-day memory after one training session was similar to that after ten training sessions in both C28 (P=0.89) and C30 (P=0.89) transgenic flies. Thus, maximum LTM was induced after just one training session in transgenic flies expressing abnormally high levels of CREB activator. N=10, 4 and 6 PIs for each group of Can-S, C28 and C30, respectively.
Example 10
Olfactory Acuity and Shock Reactivity
Odor avoidance responses to OCT or to MCH were quantified with the method of Boynton, S. and T. Tully, Genetics, 131: 655-672 (1992), given a choice between an odor and air. The odors are naturally aversive, and flies usually chose air and avoided the T-maze arm containing the odor. After 120 seconds, the flies were trapped in their respective arms of the T-maze, anesthetized and counted. A PI was calculated as a normalized percent correctly avoiding the odor (cf. Example 5). PIs from these two strains and two odor-groups (OCT and MCH) were subjected to a TWO-WAY ANOVA with STRAIN (F (1,12) =1.57, P=0.23) and ODOR (F (1,12) =0.07, P=0.80) as main effects and DRUG×ODOR (F (1,12) =0.15, P=0.71) as the interaction term. The two subsequent planned comparisons were judged significant if P<0.025.
Shock reactivity was quantified with the method of Dura, J-M., et al., J. Neurogenet., 9: 1-14 (1993) in wild-type (Can-S) flies, or in a transgenic line (C28) carrying an inducible hsp-dCREB2-a construct, three hours after a 30-minute heat shock at 37° C. Briefly, flies were placed in a T-maze and given a choice between an electrified grid (60V DC) in one T-maze arm and an unconnected grid in the other. After 120 seconds, the flies were trapped in their respective T-maze arms, anesthetized and counted. A PI was calculated as for olfactory acuity. PIs from these two strains were subjected to a ONE-WAY ANOVA with STRAIN (F (1,6) =13.03, P=0.01) as the main effect.
Naive avoidance responses to odors or to shock three hours after heat-shock did not differ between wild-type (Can-S) versus transgenic (C28) flies for the two odorants (MCH and OCT) used for conditioning experiments (P=0.27, 0.55, respectively). N=4 PIs per group. Naive shock avoidance responses three hours after training for transgenic flies were slightly lower than those for wild-flies (P=0.01). N=4 PIs per group.
Examples 11-13 pertain to the Drosophila nitric oxide synthase work.
Example 11
Low Stringency Hybridization to a Phage Library of the Drosophila
Genome and Screening of Drosophila cDNA Library
6×10 4 plaques of a genomic Drosophila λDASH library with the 1.3 kb Bgl II fragment of rat neuronal NOS cDNA (residues 3282-4573) under low stringency conditions of 40% formamide were screened as described in W. M. McGinnis et al., Nature, 308: 428 (1984). Fifty positive phage were purified and grouped based on inter se hybridization. The contig containing the 2.4R fragment of dNOS was comprised of 15 phage clones. Regions of cross-hybridization to the rat probe were identified, subcloned and three of them were sequenced. The other two did not contain sequences homologous to any protein in the database. A Drosophila head cDNA library (a gift from P. Salvaterra) was screened with the 2.4R fragment isolated from phage clone λ8.11 in standard conditions. All phage purification and cloning steps were done with standard methods (J. Sambrook, E. F. Fritsch, T. Maniatis, Molecular Cloning. A Laboratory Manual (Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., 1989)). cDNA fragments were subcloned into Bluescript (Stratagene) and sequenced on both strands with Sequenase 2.0 (USB).
Example 12
Activity of Drosophila Nitric Oxide Synthase (dNOS)
The expression construct for activity assays contained dNOS cDNA (with an XbaI site engineered immediately upstream of the ATG codon) cloned into the XbaI and SmaI sites of the pCGN expression vector (M. Tanaka and W. Herr, Cell, 60: 375 (1990)). 293 human kidney cells were transfected with 15 μg of the dNOS construct, or vector DNA, and precipitated with calcium phosphate as described in (M. J. Imperiale, L. T. Feldman and J. R. Nevins, Cell, 35: 127 (1983)). Cells were collected 2 days later and protein extracts were prepared as described in (J. Sambrook, E. F. Fritsch, T. Maniatis, Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., 1989)).
The fusion protein for raising anti-DNOS antibodies was made by cloning a 0.29 kb Eam11051-SacI fragment of dNOS cDNA (this fragment codes for 97 N-terminal amino acids of dNOS ORF) into EcoRI site of pGEX-KG (K. Guan and J. E. Dixon, Anal. Biochem., 192: 262 (1991)). The fusion protein was expressed in BL21 E. coli strain and purified over Glutathione-Sepharose columns (Pharmacia) as described in G. J. Hannon, D. Demetrick and D. Beach, Genes & Dev., 7: 2378 (1993). Immunization of rabbits, and serum preparation, was done by Hazleton Research Products, Inc. (Denver). The DNOS protein was detected on Western blots using a 1:500 dilution of rabbit serum, and cross-reacting bands were visualized with anti-rabbit alkaline phosphatase conjugate (Promega) according to the protocol provided.
The enzymatic assay was done essentially as described previously (D. Bredt and S. Snyder, Proc. Natl. Acad. Sci. USA, 87: 682 (1990)). A 100 ml reaction mixture containing 25 μl (50-100 μg) of soluble protein extract, 50 mM Hepes pH 7.4, 3>1M FAD, 3 μM FMN, 10 μM tetrahydrobiopterin (ICN), 1 mM DTT, 0.8 mM CaCl 2 , 1 mM NADPH, 10 μg/ml calmodulin, 2 μl of [ 3 H]L-arginine (35.7 Ci/mmol, NEN) and 50 mM L-valine in was incubated for 60 minutes at 37° C. The reaction was stopped with 0.5 ml 20 mM Hepes pH 5.5, 2 mM EDTA, 2 mM EGTA, loaded on 0.5 ml Dowex AG 50WX-8 (Na + form) column and eluted with 3×0.5 ml of the stop buffer. Radioactivity present in the eluent was quantified in a scintillation counter.
FIGS. 17A-17B show the expression of DNOS enzymatic activity in 293 kidney cells. FIG. 17A shows the results of a Western blot analysis of protein extracts from 293 cells transfected with vector alone (lane 293+vector) or with dNOS cDNA construct (lane 293+dNOS). 25 μg of soluble protein extracts was resolved on 7.5% polyacrylamide gel, transferred to nitrocellulose membrane and treated with anti-DNOS antibody. The arrow indicates the position of the DNOS protein. Positions of molecular weight markers (in kD) are shown on the left.
FIG. 17B shows siginificant DNOS enzyme activity measured in 293 cell extracts by conversion of 3 H-L-arginine to 3 H-L-citrulline. Enzymatic activity was detected only in cells transfected with dNOS cDNA construct (groups B-D) and is presented as specific activity (pmol of citrulline/mg/min.). The DNOS activity also was measured in the presence of 1 mM EGTA without exogenous Ca 2+ or calmodulin (group C), or in the presence of 100 mM L-NAME (group D). N=4 reactions per group.
Example 13
Splicing Pattern of dNOS
Heads and bodies of adult flies were separated on sieves. Total RNA was isolated by the guanidinium isothiocyanate method (P. Chomczynski and N. Sacchi, Anal. Biochem., 162: 156 (1987)). Poly(A) + RNA selection, Northern blot and hybridization were done with standard methods (J. Sambrook, E. F. Fritsch, T. Maniatis, Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., 1989)). The blot was hybridized with random-primed dNOS cDNA (10 6 cpm/ml), washed in 0.1×SSC and 0.1% SDS at 65° C. and exposed to X-ray film for 72 hours. Two 25-mer primers (corresponding to residues 1374-1399 (the top primer) and 1793-1817 (the bottom primer) in the dNOS sequence) were used to amplify fragments of two dNOS splice products. Each RT-PCR reaction contained 30 ng of poly(A) + head RNA. In the first stage (RT), 90 ng of the bottom primer and 5U of rTth polymerase (Perkin-Elmers) were added and the mixture was incubated in the MJ Research Minicycler™ in the following sequence of conditions: 95° C./1 minute, 67° C./45 seconds, 70° C./13 minutes. The second stage (PCR) was carried out as follows: 94° C./45 seconds, 63° C./45 seconds, 70° C./90 seconds and was repeated for 35 cycles. Products of the reaction were analyzed on a denaturing polyacrylamide (8%) gel. Bands of interest were isolated, reamplified, cloned into pCR1000 (InVitrogen) and sequenced with Sequenase kit (USB).
Northern blot analysis of dNOS expression in adult flies shows a 5.0 kb dNOS transcript present in heads ( FIG. 18A ). Each lane contained 10 mg of poly (A) + mRNA isolated from Drosophila heads (H) or bodies (B). The Northern blot was hybridized with the dNOS cDNA as described above. Positions of size markers (in kb) are shown on the left. The blot was overprobed with myosin light chain (MLC) (Parker, V. P., Mol. Cell Biol., 5: 3058-3068 (1985)) as a standard for RNA concentration.
FIG. 18B shows that the dNOS gene expresses two alternatively spliced mRNA species. RT-PCR reactions were performed on poly(A) + mRNA isolated from Drosophila heads and were resolved on 8% polyacrylamide gel. Arrows indicate the positions of DNA fragments of expected sizes: the 444 bp long-form fragment and the 129 bp short-form fragment (lane +RNA). Other bands present in this lane are artifacts from heteroduplexes that failed to denature. Poly(A) + mRNA was omitted from the control reaction (lane −RNA), which otherwise was done in identical conditions. Size markers (kb ladder) are shown in the middle lane (KB).
Equivalents
Those skilled in the art will know, or be able to ascertain, using no more than routine experimentation, many equivalents to the specific embodiments of the invention described herein. These and all other equivalents are intended to be encompassed by the following claims.