Overlapping expression patterns for bif and actin are observed during embryogenesis. During the syncytial blastoderm stage, Bif is localized in defined cytoplasmic domains or caps above somatic nuclei and is closely associated with the plasma membrane. These Bif caps undergo cycle specific reorganization during each mitosis identical to that of actin. Double-labelling experiments show that the expression pattern of Bif just prior to the cellularization stage (nuclear division cycle 14) corresponds to that of F actin and appears in a roughly hexagonal pattern (Bahri, 1997).

The colocalization of both proteins remains throughout cellularization, and they form rings around the remaining intracellular gaps at the leading edges of the invaginating membrane (Bahri, 1997).

Late in embryogenesis, anti-Bif antibody specifically labels the nerve tracts of the central nervous system (CNS). Staining is first seen in stage 13 embryos, where the protein seems to be localized primarily on extending neuronal processes and axons. During stage 14, the staining becomes associated with both the anterior and posterior commissures and later in stage 16 it becomes associated with both commissures and longitudinal connectives of the CNS. Phalloidin stained actin also labels embryonic CNS axons and shows colocalization with bif (Bahri, 1997).


Bif and actin appear to be present at similar subcellular locations at various developmental stages. Given the effect that bif mutations have on rhabdomeres and the axial actin cytoskeleton, the spatial distribution of bif relative to that of actin was examined by carrying out double-labelling experiments. In the wt, anti-Bif antibody stains all photoreceptor cells as soon as they appear posterior to the morphogenetic furrow in the third instar larval eye disc. Longitudinal optical sections of the discs show that bif is apically localized. At this stage, bif localization shows a striking resemblance to actin in the developing photoreceptors. Actin is also found in all R cells as visualized by phalloidin staining and completely overlaps bif in the apical region. Both proteins occupy the region where the future rhabdomeres would normally form. Additional actin staining is detected in the morphogenetic furrow. At ~55 h of pupation, bif becomes localized mainly to the cytoplasmic side of the extending rhabdomeres. Partial colocalization of the bif protein with actin can still be seen at this stage, although most of the actin staining is localized to the tip of the rhabdomere (Bahri, 1997).

Bif is present in R cell bodies (Bahri, 1997). To determine if Bif is also localized to R cell growth cones, third-instar larval eye-brain complexes were stained with an affinity-purified anti-Bif antibody. Consistent with the report by Chia (Bahri, 1997), Bif is detected in wild-type R cell bodies but not in bif mutants. However, it was difficult to tell if Bif is expressed in R cell growth cones in whole-mount tissues due to relatively high background staining in the developing optic lobe. To further address this issue, dissociated third-instar R cells were cultured and stained with anti-Bif antibody. Positive staining was detected in wild-type R cell bodies, axons, and growth cones, but not in bif mutants. The strong staining at the periphery of the cell body indicates that Bif is predominantly associated with plasma membrane. It is concluded that Bif, like Dock and Msn, is present in R cell growth cones (Ruan, 2002).

Effects of Mutation or Deletion

Photoreceptor cells of the Drosophila compound eye begin to develop specialized membrane foldings at the apical surface in midpupation. The microvillar structure ultimately forms the rhabdomere, an actin-rich light-gathering organelle with a characteristic shape and morphology. In a P-element transposition screen, mutations were isolated in a gene, bifocal (bif), which is required for the development of normal rhabdomeres. The morphological defects seen in bif mutant animals, in which the distinct contact domains established by the newly formed rhabdomeres are abnormal, first become apparent during midpupal development. The later defects seen in the mutant adult R cells are more dramatic, with the rhabdomeres enlarged, elongated, and frequently split. bif encodes a novel putative protein of 1063 amino acids that is expressed in the embryo and the larval eye imaginal disc in a pattern identical to that of F actin. During pupal development, Bif localizes to the base of the filamentous actin associated with the forming rhabdomeres along one side of the differentiating R cells. On the basis of its subcellular localization and loss-of-function phenotype, possible roles of Bif in photoreceptor morphogenesis are discussed (Bahri, 1997).

The first bifocal mutation resulted from a screen for local transpositions starting with a P transposon, Is(2)P842, located in the 10D cytogenetic interval of the Drosophila genome. Is(2)P842 homozygotes are phenotypically wild-type. Southern analysis of 300 putative local transposition events revealed that one of these, D59, was in fact an ~60-kb deletion which removes the entire coding region of the receptor tyrosine phosphatase gene, DPTP10D, as well as the genomic region located between the P-element insertion site and DPTP10D. Animals homozygous for D59 are viable and exhibit a weak rough-eye phenotype (Bahri, 1997).

Close examination of the bif rough-eye phenotype revealed defects in ommatidial and bristle organization. In the wt, the external surface of the adult eye consists of a hexagonal array of ~800 facets with bristles projecting from alternate facet vertices. In bif mutant eyes, adjacent ommatidia are often fused and bristles are short, missing, or duplicated. Bristle precursors are largely normal in the mutant, as seen in anti-Cut antibody-stained pupal discs, but the fused ommatidia may be caused by occasional loss of pigment cells. Internally, wild-type adult ommatidia contain eight photoreceptor cells (R1 to R8), each of which projects into the center a light-gathering organelle called the rhabdomere. The eight rhabdomeres have round cross sections and are organized in an asymmetric trapezoidal pattern. In bif mutant ommatidia, the majority of the rhabdomeres have abnormal morphology; they become enlarged, elongated, or frequently split as if the rhabdomere were duplicated . Electron micrographs show that what appears to be rhabdomere duplication under an optical microscope is actually a physical split of the mutant rhabdomeres. To determine whether additional R cells are recruited in bif ommatidia, bif pupal eye discs were stained with an anti-Elav antibody. A normal complement of eight photoreceptor cells occupying their usual positions were found in the mutant ommatidia (Bahri, 1997).

The bif rough-eye phenotype is enhanced by Fas1 loss of function and by loss of two regions of the Drosophila genome. A genetic interaction screen was carried out in search of dominant enhancers of the bif rough-eye phenotype. In this screen, bif female virgins were crossed to males carrying mutations or deficiencies on the second or third chromosome. Male progeny hemizygous for bif and heterozygous for an autosomal deficiency or mutation were examined. Two deficiencies, Df(2L)Prl [32F1-3; 33F1-2] and Df(3R)p1 4 [90C02-D01; 91A01-02], were found to dominantly enhance the bif rough-eye phenotype. Closer examination of the mutant eyes revealed that they also had elongated, malformed rhabdomeres, similar to the phenotype observed in bif mutants. However, the enhanced roughness was found to be due to additional loss or irregular arrangement of ommatidia. The region uncovered by Df(2L)Prl, which is responsible for the observed dominant enhancement of the phenotype, was further mapped to the interval 32F1-3; 33B2-3 (Bahri, 1997).

The strong rough-eye phenotype of flies homozygous for bif and heterozygous for Df(2L)Prl is clearly distinguishable from those of animals homozygous for a bif null alone. Based on the interaction with Df(2L)Prl, additional bif alleles were isolated in screens for eye roughness. R38 and R47 are two P-element-induced deletion alleles isolated in this screen. They are internal deletions in the bif gene that remove 2 and 3 kb, respectively, from the coding exon 3 regions of bif. The same strategy was also used to isolate four EMS alleles which failed to complement the D59 mutant allele (Bahri, 1997).

In addition, enhancement of the bif mutant rough-eye phenotype was observed when generating bif males, which were also homozygous for the fas1 mutation. fas1 encodes a neural cell adhesion molecule (9), and a mutant allele exists that was caused by insertion of a large, transposable element at position 89D. Revertant alleles from this insertion that are fas1+ do not enhance the roughness of bif mutant eyes. It is interesting that anti-Fas1 antibody stains all photoreceptor cells of the third instar larval eye disc in a pattern similar to that of bif. No obvious defects are seen in homozygous fas1 eyes. Enhancement of the rough-eye phenotype of bif mutant adults was not observed when another neural cell adhesion molecule, FasIII, was used in the genetic interaction experiment (Bahri, 1997).

To determine which sequences deleted in D59 might be responsible for the eye defect, three additional viable P-element-induced deletions (LH114, LH93, and LH44) were isolated. LH114 deletes ~6 kb of genomic DNA starting from the P-element insertion site and extending toward, but not into, DPTP10D. Flies homozygous for this deletion have wt eyes. However, the overlapping deletion LH93, which removes an additional ~0.5 kb of genomic DNA further towards DPTP10D produces a rough-eye phenotype. Furthermore, another overlapping deletion, LH44, that removes ~9.5 kb of genomic DNA starting from the P-element insertion site and extending toward, but not into, DPTP10D, also produces the abnormal eye phenotype. LH93 and LH44 do not complement each other, nor do they complement D59 for the eye phenotype. Analysis of these deletions suggested the possibility that DPTP10D is not involved; rather, an undefined gene (bifocal) located between the DPTP10D transcription unit and the Is(2)P842 insertion site might, in fact, be responsible for the rough-eye phenotype. Lending further support to this view, the other two rough-eye mutations, R47 and R38, which were isolated by taking advantage of a dominant genetic interaction with Df(2L)Prl, also remove DNA sequences located between the DPTP10D transcription unit and the Is(2)P842 insertion site (Bahri, 1997).

To determine whether the region under question harbors an unidentified transcript, an 8.5-kb EcoRI genomic DNA fragment that overlaps part or all of the regions deleted in the LH44, R38, and R47 mutant alleles was used as a probe in a Northern hybridization experiment. The results indicate that this probe hybridizes to several transcripts, the most predominant form of which is ~4.5 kb long (Bahri, 1997).

Examination of the locations of molecular lesions associated with the various deletions generated by mobilizing Is(2)842 and the phenotypes they produce strongly supports the view that the open reading frame encoded by the bif cDNAs is required for the bif+ phenotype. LH114 homozygotes exhibit wild-type eyes; the deletion associated with LH114 removes sequences from the 39-flanking and 3' transcribed nontranslated region of the bif cDNAs but leaves the open reading frame intact. In contrast, D59, LH93, LH44, R47, and R38, all of which produce the mutant eye phenotype, are associated with lesions that remove all or part of the open reading frame. Moreover, with the exception of D59, none of the mutations appear to affect the expression of the DPTP10D protein. These data indicate that it is the new transcription unit associated with the bif cDNAs, and not DPTP10D, that is responsible for the observed mutant eye phenotype (Bahri, 1997).

Defects in bif rhabdomeres are accompanied by abnormal actin localization. Given the apparent structural cell deformities associated with bif, the eye phenotype was analyzed at the subcellular level by using antibodies that recognize cytoskeletal markers such as actin, spectrin, and phosphotyrosine. Actin intensely stains the rhabdomeres, whereas phosphotyrosine localizes to the cytoplasmic region adjacent to the rhabdomeral domain of R cells. In contrast, Spectrin usually delineates the boundary of the R-cell body. In these studies, the bif alleles D59 and LH44 showed results similar to those obtained with R47 (Bahri, 1997).

In wild-type eyes, actin marks the rhabdomeres very early in their development. At 55 h of pupation actin staining is intense in the apical microvillar tips of R cells facing the future interretinal space and is also visible in the smooth stalks extending to the adherens junctions. Proper contacts between these actin-stained apical domains usually ensure normal rhabdomere development. In bif ommatidia, there is no clear distinction between microvillar tips and stalks at this stage. Instead, the early star-like actin staining becomes abnormal, merging together in the center to form an equator-like pattern. The other two markers used in this study, spectrin and phosphotyrosine, are normally localized in bif eye discs (similar to the wild-type) and do not show any additional morphological disruptions in mutant R cells (Bahri, 1997).

The early expression pattern of bif in larval R cells remains intriguing. bif is expressed in the extreme apical region of the cell bodies as soon as they form immediately behind the furrow and long before the initial events associated with rhabdomere development occur, yet there are no obvious deformities in bif mutant R cells at this stage. All R cells form normally. This indicates that bif is not required for the formation of R cells. It is conceivable that a critical level of bif must accumulate in the R cell prior to microvillar extension to ensure proper formation of rhabdomeres (Bahri, 1997).

bif rhabdomeres start to show a variety of morphological defects during early rhabdomere formation. These deformities were found to be accompanied by aberrations in the actin staining pattern. bif and actin colocalize in the apical region of R cells (in the third instar larval eye disc), where future rhabdomeres would normally form (in midpupation). In midpupation, most of bif occupies the cytoplasmic base of the axial actin associated with the forming rhabdomeres along one side of the differentiating R cells. The localization data and the effect on actin distribution suggest that bif may serve as a cytoplasmic support for the microvillar actin filaments and its presence immediately adjacent to apical membrane folding may confer rigidity on the rapidly expanding cytoplasmic surface. Thus, it appears that the defects of bif rhabdomeres reflect a developmental function of the bif gene in rhabdomere growth possibly mediated by involvement of the bif protein in the R-cell cytoskeleton (Bahri, 1997).

The actin-rich cytoskeleton mediates several cellular processes that lead to changes in cell shape, membrane stability, movement, and generation of force. This may involve multiple actin-binding proteins which drive the changes in actin assembly and distribution. However, in the Drosophila eye the stabilizing effect that bif might have on actin organization may be indirect, since at the molecular level Bif does not contain any obvious actin-binding domain (Bahri, 1997).

The precise nature is not known of the mutations uncovered by the deficiencies that phenotypically interact with bif. However, genetic screens can be carried out to isolate interacting genes and therefore allow the identification of their gene products that influence the bif mutant phenotype. The genetic interaction of bif with Fas1 differs from its interaction with the Prl and p14 deficiencies in that the mutation in fas1 acts synergistically with bif only when both are homozygous. Since Fas1 promotes cell-cell interactions via homophilic binding and Bif seems to be associated with the R-cell cytoskeleton, it is possible that bif and Fas1 are involved in maintaining the photoreceptor cytoarchitecture and the overall ommatidial structure (Bahri, 1997).

The fas1 mutation alone is viable and has no obvious adult eye phenotype, which indicates that this gene may encode a redundant function in the eye. In the embryo, a role for fas1 in the developing CNS axons was established through its genetic interaction with Abelson. Similarly, the genetic interaction with bif provides a basis for the analysis of the functional role that fas1 may play in eye development (Bahri, 1997).

Both dock and msn are required for the proper termination of R1–R6 growth cones in the lamina. The function of Bif in the same pathway, as suggested by the dosage-sensitive interaction between msn and bif, should lead to altered R1–R6 targeting pattern in bif mutants. To address this possibility, the R cell projection pattern was examined in bif mutants using cell-type-specific markers. The projection pattern of all R cell axons in late third-instar larvae was visualized using the R cell-specific marker MAb 24B10. In bif mutants, like in msn mutants, gaps were frequently seen in the R1–R6 growth cone termination site. The R7 and R8 terminal field in the medulla is also mildly disorganized in bif mutants (Ruan, 2002).

To determine if the above defects are due to a failure of R1–R6 growth cones to stop at lamina, the Ro-Tau-lacZ marker was used to assess the targeting pattern of a subset of R1–R6 growth cones at third-instar larval stage. In wild-type, Ro-Tau-lacZ labels R2–R5 axons within the lamina. In bifR47 mutants, like in msn mutants, many R2–R5 axons project through the lamina into the medulla layer. This phenotype was also observed in three other bif alleles examined, including bifR38, bifLH114, and bifEP(x)395 (Ruan, 2002).

To examine if bif mutations also affect R7 projections an R7-specific marker PM181-GAL4-UAS-lacZ was used to label R7 axons in bif mutants. In wild-type, R7 axons migrate from the eye disc through the optic stalk and lamina into the medulla. Their growth cones expand significantly in size after reaching the medulla region. In all bif mutants examined, the R7 projection pattern was indistinguishable from that in wild-type (Ruan, 2002).

R1–R6 targeting defect was unlikely due to abnormal R cell differentiation or fate determination (e.g., transformation of an R1–R6 cell into an R7 or R8 fate) as assessed by using several R cell-specific markers. To further determine if R1–R6 targeting defect in bif mutants reflects a direct role for Bif in R1–R6 growth cones, the development of the target region was examined by using the lamina neuronal differentiation marker Dachshund and the glial marker Repo. The differentiation and organization of lamina neurons appear normal in bif mutants. Lamina glia in bif mutants appear less organized compared to those in wild-type. However, the number of glial cells in the R1–R6 target region in bif mutants is similar to that in wild-type, indicating that the differentiation and migration of lamina glial cells occurs normally in bif mutants. That the arrangement of glia in the target region was mildly disorganized was likely caused by abnormal R cell innervation. Mutations in several other genes (e.g., dock and brakeless) that are required in R cells for targeting also affect the organization of glial cells in the target region. Furthermore, eye-specific expression of a bif transgene under control of the GMR promoter in bif mutants largely rescues R1–R6 targeting defects. These results, together with the data showing the presence of Bif in R cell axons and growth cones, indicate that Bif is required in R cell growth cones for specifying targeting decisions (Ruan, 2002).

Overexpression of Msn causes pretarget termination of many R cell growth cones. To determine if overexpression of Bif also causes a similar dominant phenotype, the R cell projection pattern was examined in wild-type larvae carrying an UAS-bif transgene under control of the eye-specific GMR-GAL4 driver. Indeed, overexpression of Bif induces an axonal projection phenotype similar, if not identical, to that caused by Msn overexpression. In all larvae carrying one copy of GMR-GAL4 and UAS-bif, the R1–R6 termination layer displays a mild defect compared to that in wild-type. The lamina plexus was often split due to the uneven termination of R1–R6 growth cones, while the R7 and R8 terminal field in the medulla remains largely normal (Ruan, 2002).

When two copies of UAS-bif were present, the phenotype became much more severe in all hemispheres examined. The number of prematurely terminated R cell axons increased dramatically. In addition, the R7 and R8 terminal field was also severely disrupted. Interestingly, the increase in the dosage of the bif transgene also severely disrupted the organization of R cell clusters in all eye discs examined. Many R cell bodies migrated abnormally toward the most posterior end and accumulated at the entry point of the optic stalk. This R cell migration phenotype was never observed in larvae overexpressing Msn, which may reflect a more general role for Bif in cytoskeletal regulation (Ruan, 2002).

To examine if the bif gain-of-function phenotype is sensitive to the dosage of msn, the strong msn loss-of-function allele msn102 was crossed into the bif overexpression background. No modification of the bif overexpression phenotype was observed, which is in marked contrast to the observation that msn overexpression phenotype can be largely suppressed by reducing the dosage of bif. However, while this data is consistent with data that indicate bif functions downstream of msn in the pathway, the possibility cannot be excluded that the reduced amount of Msn is sufficient for the action of excess Bif since the analysis was not performed in msn null mutants due to technical reasons (Ruan, 2002).

The effect of Bif overexpression was also assessed by placing UAS-bif under control of the R7-specific driver PM181-GAL4. In the absence of UAS-bif, all R7 axons projected into the medulla. While R7 differentiation and the organization of R7 cell bodies remain normal in larvae overexpressing either msn or bif in R7 cells, a small number of R7 axons were frequently observed that terminated within the lamina soon after exiting the optic stalk. This phenotype, however, was less severe and less penetrant than that caused by overexpression of msn or bif under control of the GMR-GAL4 driver, which might be due to the difference in either the level or the timing of GAL4 expression (Ruan, 2002).

In summary, like overexpression of msn, overexpression of bif also causes pretarget growth cone termination. Unlike overexpression of msn that only affects R cell growth cone targeting, overexpression of bif also causes abnormal migration of R cell clusters. Since the organization of R clusters is not affected in third-instar bif loss-of-function mutants, it is suspected that excess bif induces this R cell migration phenotype by interfering with cytoskeletal events in R cell bodies that are normally mediated by other proteins (Ruan, 2002).

Each photoreceptor rhabdomere is an interconnected stack of membranes that emerges from the apical neuronal membrane. The process of rhabdomere site selection, initiation, and elaboration is poorly understood. The results indicate a role for Amphiphysin (Amph) in organizing the intricately folded rhabdomere membrane. (1) Accumulation of Amph on the apical surface occurs together with an enrichment of F-actin at this site. (2) Overexpression of Amph (by targeted misexpression or in bifocal mutants) results in the delocalization of endogenous Amph and the mislocalization of both F-actin, leading to either loss or ectopic rhabdomeres. bifocal codes for a novel protein that colocalizes with actin and is necessary for photoreceptor morphogenesis (Bahri, 1997). Loss of rhabdomeres may be due to excess Amph titrating out factors necessary for the proper development of rhabdomeres; ectopic rhabdomeres may be due to ectopic Amph recruiting sufficient F-actin or other proteins to a second site in the cell and thus triggering formation of an extra rhabdomere (Zelhof, 2001).

The Amph overexpression eye phenotype (split/ectopic rhabdomeres) is similar to the loss-of-function bifocal mutant phenotype (Bahri, 1997), suggesting that Amph and bifocal may act in the same genetic pathway. Bifocal is a novel protein that is colocalized with Amph at the apical membrane of newly formed embryonic cells, as well as at the apical membrane of photoreceptor neurons. In photoreceptor neurons, Bifocal expression precedes Amph and lasts into adulthood. Amph mutants show normal Bifocal localization, but bifocal mutants show delocalization of Amph into a broad apical domain matching that of F-actin, rather than its normal tight apical crescent. The bifocal mutant phenotype can be suppressed by reducing amph gene dosage by 50%. Thus, Bifocal-mediate localization of Amph to the future rhabdomere membrane domain is essential for normal eye development (Zelhof, 2001).

To determine the developmental origin of the Amph overexpression phenotype, Amph, Bifocal and F-actin localization were assayed during rhabdomere development. Wild-type photoreceptors show an even distribution of Bifocal, Amph and F-actin at the apical surface of the cell. By contrast, photoreceptors overexpressing Amph have an abnormal punctate 'ball' of F-actin and Bifocal at the apical cortex and Amph is delocalized from the apical membrane. It is concluded that excess Amph protein leads to destabilization of the normal apical Amph localization, a mislocalization of Bifocal protein and F-actin, and the subsequent failure in rhabdomere morphogenesis (Zelhof, 2001).

These data leads to the following model for Amph function. Bifocal is localized to the apical membrane of photoreceptor neurons, where it recruits Amph and other proteins. This protein complex then aids in the morphogenesis of the intricately folded rhabdomere membrane. Loss of Amph results in a mild phenotype, perhaps because Bifocal and additional proteins are still apically localized and can promote rhabdomere morphogenesis. However, when Amph is mislocalized outside of the apical membrane domain (in Amph overexpression experiments or in bifocal mutants), it can induce the formation of ectopic rhabdomere membrane domains or inhibit rhabdomere morphogenesis by relocating or titrating away the necessary protein complex to form a rhabdomere (Zelhof, 2001).


Babu, K., Cai, Y., Bahri, S. Yang, X. and Chia, W. (2004). Roles of Bifocal, Homer, and F-actin in anchoring Oskar to the posterior cortex of Drosophila oocytes. Genes Dev. 18: 138-143. 14752008

Babu, K., Bahri, S., Alphey, L. and Chia, W. (2005). Bifocal and PP1 interaction regulates targeting of the R-cell growth cone in Drosophila. Dev. Biol. 288: 372-386. 16280124

Bahri, S. M., Yang, X. and Chia, W. (1997). The Drosophila bifocal gene encodes a novel protein which colocalizes with actin and is necessary for photoreceptor morphogenesis. Mol. Cell. Biol. 17: 5521-5529. 9271427

Helps, N..R., Cohen, P. T., Bahri, S. M., Chia, W. and Babu, K. (2001). Interaction with protein phosphatase 1 is essential for bifocal function during the morphogenesis of the Drosophila compound eye. Mol. Cell. Biol. 21: 2154-2164. 11238949

Ruan, W., Long, H. Vuong, D. H. and Rao, Y. (2002). Bifocal is a downstream target of the Ste20-like serine/threonine kinase Misshapen in regulating photoreceptor growth cone targeting in Drosophila. Neuron 36: 831-842. 12467587

Sisson, J. C., Field, C., Ventura, R., Royou, A. and Sullivan, W. (2000). Lava Lamp, a novel peripheral golgi protein, is required for Drosophila melanogaster cellularization. J. Cell Biol. 151: 905-918. 11076973

Zelhof, A. C., et al. (2001). Drosophila Amphiphysin is implicated in protein localization and membrane morphogenesis but not in synaptic vesicle endocytosis. Development 128: 5005-5015. 11748137

bifocal: Biological Overview | Regulation | Developmental Biology | Effects of Mutation

date revised: 10 June 2004

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

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