QUICK SEARCH:   [advanced] Author: Keyword(s):
Home     Help     Feedback     Subscriptions     Archive     Search     Table of Contents

This Article Summary Full Text Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Zelhof, A. C. Articles by Doe, C. Q. Search for Related Content PubMed PubMed Citation Articles by Zelhof, A. C. Articles by Doe, C. Q.

Drosophila Amphiphysin is implicated in protein localization and membrane morphogenesis but not in synaptic vesicle endocytosis

¹ Institute of Neuroscience, HHMI, University of Oregon 1254, Eugene, OR 97403, USA
² Section of Neurobiology, Institute for Cellular and Molecular Biology, University of Texas at Austin. Austin, TX 78712, USA
³ Department of Biology, HHMI, University of California, San Diego. La Jolla, CA 92093, USA
⁴ Department of Genetics, University of Cambridge, Downing Street, Cambridge CB2 3EH, UK
⁵ Department of Biochemistry, University of Cambridge, Tennis Court Road, Cambridge CB2 1QW, UK

View larger version (37K): [in a new window]   Fig. 1. Amphiphysin gene organization and characterization of mutant alleles. (A) Amphiphysin consists of 10 exons extending over 18 kb of DNA, including a large 13 kb second intron that contains no detectable genes. The EP(2)2175 transposon is inserted 412 bp from the AUG encoding the first methionine of the cDNA; imprecise excision of the EP transposon produced the amph26 and amph54 null alleles. In amph26 there is a 3428 bp deletion (breakpoints –1925/1502) that removes the entire first exon; in amph54 there is a 1834 bp deletion (breakpoints –581/1253) that removes the entire first exon but does not remove the first exon of Sin3A. All phenotypes are observed with either mutant. The Sin3A and Gq-3 genes are adjacent to the amph gene, but neither amph mutation affects the coding sequences of Sin3A or Gq-3 genes, and the Sin3A and amph26 mutations complement each other (A. Razzaq, personal communication), consistent with amph26 having no effect on Sin3A function. (B) Western analysis of Drosophila extracts from Amphiphysin mutant and precise excisions (amph+1). In each of the putative null mutants, amph26 and amph54 (data not shown), there is no immunoreactivity detected. (C) Schematic representations of the four Amph isoforms. In each case, the alternative splicing occurs in exon 8, reducing or eliminating the central domain of Amphiphysin (top); sequence analysis of the four isoforms from the end of exon 7 to the start of exon 9 (bottom). (D) Drosophila Amphiphysin does not interact with Drosophila Dynamin: western analysis and Coomassie staining of Dynamin binding to the SH3 domain of Amphiphysin. Drosophila extract was mixed with three different fusion proteins: GST, GST fused to the SH3 domain of Amph (GST-SH3) and GST fused to a mutated form of the SH3 domain (GST-SH3*). In each case, we could not detect any binding of Dynamin to the GST proteins. The asterisks indicate the GST proteins.

View larger version (84K): [in a new window]   Fig. 2. Amphiphysin is postsynaptic at the larval neuromuscular junction. (A-D) Amphiphysin (green) does not colocalize with presynaptic marker Cysteine String Protein (red) at third instar neuromuscular synapses; Amph staining surrounds the CSP staining, as expected for a postsynaptic protein. Low-magnification view of the NMJs at muscle 6 and 7 of abdominal segment 2 showing Amph (A), Csp (B) and the merged image (C); a high-magnification image of the region indicated with an arrow in C is shown in D. (E-H) Amphiphysin (green) is precisely colocalized with the postsynaptic marker Discs large (red) at third instar type I neuromuscular synapses. Low-magnification view showing Amph (E), Dlg (F) and the merged image (G); a high-magnification image of the region indicated with an arrow in G is shown in H.

View larger version (69K): [in a new window]   Fig. 3. Amphiphysin is required for postsynaptic protein localization and synaptic function at larval type I neuromuscular junctions. (A) Miniature excitatory junctional potentials (mEJP) recordings reveal there is a small but significant increase in quantal size in amph mutants. Left panels show representative traces of mEJPs from amph26 mutant and wild-type (amph+1) flies. Histograms and comparative cumulative plots are shown on the right. The average mEJP for the amph+1/amph+1 is 0.81±0.01 mV (n=2804) and for amph26/amph26 is 0.94±0.02 mV (n=2148) (P<0.05). (B) There is no change in the amplitude of the elicited response in amph mutants. Elicited response is measured as excitatory junctional potentials (EJPs) for amph26 mutants and wild type (amph+1). Representative EJP tracings from amph26/amph26 and amph+1/amph+1 larva (top). Histograms representing the average amplitude for each genotype (bottom): amph26/amph26 35.5±1.29 mV (n=16) and amph+1/amph+1 34.9±1.22 mV (n=15) P>0.1. (C) Immunohistochemical examination of the third instar NMJ of wild type (top) and amph26 mutant (bottom) larva. The NMJs of muscles 6 and 7 of abdominal segment 2 are shown. The localization of Dlg, Lgl and GluRIIB are shown. In amph mutants, Lgl is absent from the synapse but still maintained in the M band of the muscle. GluRIIB protein localization at the NMJ is identical in wild type and amph26 mutants. (D) Immunohistochemical examination of the third instar larval NMJ in wild type (top) and amph26 mutants (bottom). The NMJs of muscle 6 and 7 of abdominal segment 2 are shown. The boxed areas represent the magnified views of the accompanying panels. In amph mutants, both Scrib and Dlg still show localization to the synapse but there is protein delocalized throughout the muscle.

View larger version (68K): [in a new window]   Fig. 4. Amphiphysin is localized to the apical membrane domain of epithelial and neural cell types. (A-D) Lateral views of a developmental time series of a stage 5 embryo stained for Amph. At early stage 5, the syncitial nuclei have migrated to the periphery and the membrane has begun to invaginate between the nuclei. Amph is first localized to the apical surface and migrates basally as the membrane extends inwards during cellularization. (D) Upon the completion of cellularization at the end of stage 5, Amph is again localized apically. (E-H) Double labels of Amph and Bifocal protein localization during early cellularization (E,G) and after completion of cellularization (F,H). Amph protein (E,F) is enriched at the invaginating membrane during cellularization, whereas Bifocal protein (G,H) remains localized to the apical membrane domain. (I) Amph is apical in ectoderm and neuroblasts. Lateral view of a stage 9 embryo showing apical Amph in the ventral ectoderm and in neuroblasts (arrows). (J) Lateral view of a stage 16 embryo to emphasize that Amph is not detected in the embryonic CNS (bottom left panel) when compared with the staining in the epidermis and trachea (top panel) and hindgut (bottom right panel). (K,L) Amph is detected at the apical (lumenal) membrane of tracheal tubules and of the esophagus. (K) Stage 15 embryo showing Amph expression at the apical surface in mature tracheal tubules. (L) Stage 15 embryo showing Amph expression in the esophagus. In addition, the non-epithelial secretory garland cells flanking the esophagus also express Amph. (M,N) Amph is localized to apical membrane and vesicles in internal tubular epithelia, such as the hindgut (M) and salivary glands (N). The arrow indicates Amph immunoreactive vesicles.

View larger version (145K): [in a new window]   Fig. 5. Amphiphysin is localized to the apical membrane before rhabdomere formation in photoreceptor neurons. Amph protein (green) and F-actin (red) are shown in pupal photoreceptor neurons (A-F) or in adult eyes (G-I). Each image represents a single confocal section of a z-series. (A-C) Twenty-four hours APF, only actin accumulates at the apical surface of the photoreceptor cells, whereas Amph is found throughout the cell. (D-F) Forty-eight hours APF, Amph accumulates on the apical surface of the photoreceptor cells, where the rhabdomeres will develop. There is some overlap between F-actin and Amph, but F-actin becomes tightly colocalized with Amph only at 55 hours APF (see Fig. 7F). (G-I) In the adult eye, F-actin (red) strongly labels the rhabdomere membrane, whereas Amph (green) is specifically expressed in the lens-secreting cone cells above the rhabdomeres (arrows). Amph expression is also found in cells of the adult head.

View larger version (75K): [in a new window]   Fig. 6. Amphiphysin organizes the rhabdomere membrane domain in photoreceptor neurons. (A,B) Transmission electron microscopy analysis of photoreceptor cells overexpressing Amph. GMR-GAL4 is a transgenic line that contains the glass enhancer driving GAL4. This line provides strong expression of GAL4 in every cell posterior to the morphogenetic furrow. (A) GMR-GAL4 control. One copy of the transgenic line does not result in any detectable defects. (B) The addition of one copy of UAS-Amph results in split and ectopic rhabdomeres (asterisks) or the loss of rhabdomeres (arrowhead). (C,D) Optical sections through the developing retinal epithelium 55 hours APF stained for actin. (C) GMR-GAL4 control. F-Actin localizes in tight crescents on the apical surface of each photoreceptor cell. (D) GMR-GAL4 and UAS-Amph. The overexpression of Amph results in F-Actin accumulating in tight ball like structures versus the characteristic crescents seen in the control. (E,F) Optical sections through the developing retinal epithelium 55 hours APF stained for Bifocal. (E) GMR-GAL4 control. Bifocal localizes to the apical surface of photoreceptor cells. (F) GMR-GAL4 and UAS-Amph. Bifocal is mislocalized in each of the photoreceptor cells overexpressing Amph. (G,H) Optical sections through the developing retinal epithelium 55 hours APF stained for Discs large (red) and Amph (green). (G) GMR-GAL4 control. Amph shows normal apical localization. (H) GMR-GAL4 and UAS-Amph. The overexpression of Amph results in the delocalization of Amph from the apical surface of photoreceptor cells but overexpression does not affect patterning of the ommatidia.

View larger version (118K): [in a new window]   Fig. 7. Amph and Bifocal act oppositely for proper rhabdomere development. (A-C) Transmission electron micrographs of genetic combinations of Bifocal and Amph. (A) amph26/amph26 mutant. The rhabdomeres are closely packed and occasionally fused (arrow). (B) bifocalR38/bifocalR38. As reported previously, bifocal mutants have split and elongated rhabdomeres. (C) bifocalR38/Y; amph26/+. Removal of one copy of amph results in a partial rescue of the bifocal phenotype. First actin localization appears normal (data not shown) and subsequently fewer rhabdomeres are split and elongated. (D-F) Optical sections through the developing retinal epithelium 55 hours APF of a bifocal mutant stained for Amph (green) and F-actin (red) (D) Amph expression in bifocalR38/bifocalR38. Amph localization in a bifocal mutant still demarcates the apical surface but is very diffuse rather than a tight crescent. (E) F-Actin expression in bifocalR38/bifocalR38. Like Amph, Actin still localizes to the apical surface but shows diffuse staining. (F) F-Actin still colocalizes with Amph, as seen in normal development.