poly-U-binding splicing factor/half pint: Biological Overview | Evolutionary Homologs | Regulation | Developmental Biology | Effects of Mutation | References
Gene name - poly-U-binding splicing factor

Synonyms - half pint (hfp)

Cytological map position - 63A3

Function - mRNA splicing factor

Keywords - alternative splicing, oogenesis

Symbol - pUbsf

FlyBase ID: FBgn0028577

Genetic map position -

Classification - RNA-binding region RNP-1 (RNA recognition motif)

Cellular location - nuclear

NCBI links: Precomputed BLAST | Entrez Gene | UniGene |

Alternative splicing is used by metazoans to increase protein diversity and to alter gene expression during development. However, few factors that control splice site choice in vivo have been identified. Half pint (Hfp; FlyBase designation Poly-U-binding splicing factor) regulates RNA splicing in Drosophila. Females harboring hypomorphic mutations in hfp lay short eggs and show defects in germline mitosis, nuclear morphology, and RNA localization during oogenesis. hfp encodes the Drosophila ortholog of human PUF60 and functions in both constitutive and alternative splicing in vivo. In particular, hfp mutants display striking defects in the developmentally regulated splicing of ovarian tumor (otu). Furthermore, transgenic expression of the missing otu splice form can rescue the ovarian phenotypes of hfp. hfp is also required for efficient splicing of gurken mRNA and in the splicing of eukaryotic initiation factor 4E (eIF4E), which binds to the seven-methylguanosine cap at the 5' end of messenger RNAs and is a limiting factor in translation initiation (Van Buskirk, 2002).

Alternative exon selection is a means of producing multiple transcripts, often encoding distinct polypeptides, from a single gene. At least 35% of human genes are alternatively spliced, leading in some cases to a remarkable diversity of transcripts. Changes in splicing have been shown to alter the activities of transcription factors and cell death regulators, the ligand binding specificities of growth factor receptors and cell adhesion molecules, and the subcellular distribution of proteins. Despite the prevalence of alternative splicing, however, little is known about its regulation in vivo. One notable exception comes from the sex determination pathway of Drosophila, in which the Sex-lethal (Sxl) protein catalyzes a series of sex-specific alternative splicing events that lead to a mode of sexual differentiation and dosage compensation in keeping with the X:A ratio of the animal. Interestingly, the male-specific lethal-2 (msl-2) gene is regulated by Sxl at two levels: Sxl binds within and suppresses the removal of an intron in the 5'UTR of msl-2 and also represses translation of the unspliced, exported message. From this and other examples, it is becoming clear that several aspects of RNA metabolism, from RNA export to translation, are connected with splicing regulation (Van Buskirk, 2002 and references therein).

The gurken (grk) gene of Drosophila, a mammalian TGFalpha homolog, has been used to study many aspects of RNA regulation, including nuclear export, RNA localization, and translational control. During midoogenesis, grk RNA becomes restricted to one side of the oocyte, where it produces a localized source of protein that induces EGF receptor activation in the overlying follicle cells, establishing the dorsal side of the egg shell and embryo. The importance of grk RNA localization is evidenced by the defects in dorsoventral (DV) polarity that result when the grk transcript is not correctly localized. In egg chambers mutant for squid (sqd), grk RNA is not restricted to the dorsal-anterior corner of the stage 9 oocyte and instead forms an anterior ring, resulting in activation of the EGF receptor around the circumference of the oocyte and leading to the production of dorsalized eggs. The sqd gene encodes three isoforms of hrp40, a Drosophila hnRNP protein found within the nuclei and cytoplasm of all cells in the egg chamber, and has been proposed to function in the regulated nuclear export and translational repression of grk mRNA (Van Buskirk, 2002 and references therein).

The encore (enc) gene is also required for proper grk mRNA localization (Hawkins, 1997). However, in enc mutant egg chambers, Grk protein levels are reduced, and these females lay ventralized, not dorsalized, eggs. Thus it appears that enc is not only required for proper grk mRNA localization but for its translation as well. enc encodes a large, cytoplasmic protein of unknown function that becomes concentrated at the dorsal side of the oocyte in stage 9 egg chambers, where it colocalizes with grk mRNA (Van Buskirk, 2000). While the DV defects of enc are cold sensitive, enc females raised at higher temperatures show a highly penetrant extra mitosis in the germline, producing oocytes that are associated with 31 nurse cells instead of the normal 15 (Hawkins, 1996). However, the effectors that mediate this extra mitosis are not known (Van Buskirk, 2002 and references).

Half pint was isolated in a yeast two-hybrid screen with Enc; mutations in enc affect grk RNA localization and germline mitosis. Half pint is so named based on the mutant phenotype, which includes the production of short eggs and the formation of egg chambers that contain eight instead of sixteen germline cells. Surprisingly, half pint encodes the Drosophila homolog of human PUF60, an RNA binding protein characterized as a general splicing factor. Drosophila Hfp does indeed affect splicing: it specifically regulates the alternative splicing of a subset of genes within the ovary (Van Buskirk, 2002).

The encore (enc) gene is required for the regulation of germline mitosis, karyosome formation, and establishment of dorsoventral (DV) polarity of the Drosophila egg and embryo (Hawkins, 1996; Hawkins, 1997; Van Buskirk, 2000). enc encodes a large cytoplasmic protein with a single conserved region that contains an R3H motif (Van Buskirk, 2000), which is implicated in RNA binding. A portion of the Enc protein containing this conserved domain was used as bait in a two-hybrid screen for interacting proteins. A single factor that shows a specific interaction with Enc was recovered, and this interaction has been confirmed through coimmunoprecipitation. The gene encoding this interacting protein has been named half pint (Van Buskirk, 2002).

Multiple transcripts produced from alternative splicing can encode proteins with different or even opposing functions, and, hence, the regulation of alternative splicing can have profound biological consequences. However, few factors that regulate splice site selection in vivo have been identified, and much of the knowledge of splicing regulation comes from in vitro studies. The difficulty in isolating factors with effects on alternative splicing in vivo has, in part, led to the hypothesis that splice site choice is influenced by many factors that act combinatorially, each exerting a small influence on any one splicing event. Hfp, however, represents a factor with a significant role in alternative splicing regulation. Reductions in Hfp activity disrupt the developmentally regulated splicing of ovarian tumor, splice site usage within eIF4E and the efficiency of grk splicing. However, several other splicing events appear to be unaffected. The specificity of Hfp function is highlighted by the observation that, though both the otu and sqd splicing patterns shift between early and late oogenesis, otu, but not sqd, splicing is regulated by Hfp (Van Buskirk, 2002).

In vivo analysis of Hfp function along with in vitro studies on the human homolog, PUF60, demonstrate a role for this factor in splicing. However, the human protein has been implicated in a number of other processes. It has been identified as Ro-BP1, a component of RNP complexes of unknown function found in both nuclear and cytoplasmic compartments (Peek, 1993). It has also been isolated as FBP-interacting repressor (FIR), shown to block activated transcription by interacting with the TFIIH complex (Liu, 2000). PUF60/Ro-BP1/FIR therefore appears to represent an unusually multifunctional protein, participating directly in transcriptional repression and pre-mRNA splicing and also associating with RNA-protein complexes in the cytoplasm (Van Buskirk, 2002).

The observation that all of the ovarian phenotypes of hfp can be rescued by expression of the missing 104 kDa Otu splice form underscores the importance of otu splicing regulation and reveals previously undetected aspects of Otu function. While some otu egg chambers that have too few or too many germline cells have been observed, and the tumorous phenotype of strong otu mutations indicates a role in cell proliferation, none of the otu mutations specifically prevent the fourth mitosis of the germline cluster. The observations that Otu104 can rescue this defect in hfp and that the otu11 mutation can suppress the enc mitotic defect point to a role for Otu in the precise regulation of the four rounds of germline mitosis. Similarly, while otu egg chambers have been shown to display defects in the localization of oskar RNA to the posterior pole of the oocyte, a defect also observed in hfp, a role for otu in the localization of grk mRNA has not been detected. These results suggest that otu does indeed play a role in grk RNA localization and that it is the 104 kDa isoform that performs this function. How does Otu participate in these diverse processes, from the regulation of germline mitosis to the dispersal of polytene chromosomes to the localization of RNAs? Mutations in otu are associated with aberrations in actin distribution, including ectopic accumulation of actin filaments within the fusome, a germline organelle implicated in the control of cystocyte division. Thus, otu may affect germline mitosis by regulating the interaction between fusome components and actin microfilaments. Likewise, otu may participate in mRNA localization by regulating the interaction of localized RNAs with the cortical actin cytoskeleton, which may act to anchor transcripts at their subcellular destinations. A direct role for otu in RNA localization is supported by its cofractionation with RNP complexes (Van Buskirk, 2002).

Both hfp and enc affect the regulation of germline mitosis. While hfp mutant egg chambers often contain eight germline cells, enc mutations result in an extra round of division, resulting in egg chambers with 32 germline cells. Overexpression of Otu104 can rescue the division defect of hfp and reduction of Otu104 activity can suppress the enc extra division. However, though Otu104 levels are strongly decreased in hfp, no converse overproduction of this isoform is detected in enc mutants, and, thus, enc does not appear to antagonize Hfp's splicing activity. It is possible that enc regulates germline mitosis through a pathway independent of hfp. However, the observed interaction between Hfp and Enc proteins raises the possibility that these genes function in a common pathway. If Enc does not antagonize Hfp's splicing activity, though, under what circumstances do these proteins interact? It may be useful to consider the Hfp-Enc interaction in terms of the roles these genes play in DV patterning. hfp and enc mutants both show defects in grk RNA localization, and enc is required for the accumulation of Grk protein (Hawkins, 1997). hfp plays a role in grk splicing. As has been observed for other splicing factors, it is possible that Hfp remains associated with its targets during nuclear export and perhaps delivers the spliced grk transcript to Enc in the cytoplasm. Enc may in turn be required for both anchoring and translation of the grk message. It is thus proposed that Hfp may have two roles: one in splicing regulation and one in the subcellular targeting of RNAs (Van Buskirk, 2002).

Since Otu104 overexpression can rescue the RNA localization defect of the hypomorphic allele hfp9, the delivery of grk RNA to the anchoring machinery must also be dependent on otu. Otu's tudor domain and observed cofractionation in RNP complexes suggest that it might associate with mRNAs during transport or anchoring. It is therefore proposed that Hfp may also be present within the RNP complexes with which Otu fractionates. In support of this, the vertebrate homolog of Hfp has been isolated as a component of nucleo-cytoplasmic RNP complexes (Peek, 1993), and Hfp has also been found to copurify with proteins found in cytoplasmic RNP particles (J. Wilhelm, personal communication to Van Buskirk, 2002). This putative cytoplasmic function of Hfp may involve an interaction with Enc. For instance, Enc may be required to release grk RNA from an intermediate Hfp-/Otu-dependent stage of transport to an anchored, translationally competent state (Van Buskirk, 2002).

Much of the knowledge of splicing regulation in vivo comes from the study of the Sex-lethal (Sxl) protein of Drosophila. This factor has been shown to control splice site usage by blocking U2snRNP auxiliary factor (U2AF) from binding to the polypyrimidine tract upstream of the 3' splice acceptor site. Thus, Sxl binding represses usage of the proximal 3' splice site, and a competing splice acceptor was used. How does Hfp function to regulate splice site choice? In the case of the otu transcript, does hfp promote splicing to exon 6a or repress selection of exon 7? In vitro studies with the human homolog of hfp, PUF60, point to a stimulatory role (Page-McCaw, 1999). In these studies, PUF60 [60 kDa poly(U) binding factor] was isolated as part of a complex that could stimulate efficient splicing of several introns in Hela nuclear extracts. This complex, along with U2AF, was found to facilitate the association of U2snRNP with the pre-mRNA, thus promoting spliceosome assembly (Van Buskirk, 2002).

The rat homolog of Hfp, Siah-BP, was isolated in a three-hybrid screen for factors that could bind to a splicing enhancer, a short RNA sequence required for neuron-specific alternative splicing of the amyloid precursor protein (APP) gene (Poleev, 2000). Siah-BP was also found to directly bind the large subunit of U2AF, consistent with studies on PUF60 suggesting that it acts in conjunction with U2AF in the early steps of splicing (Page-McCaw, 1999). The ability of the human PUF complex to stimulate splicing of all introns assayed suggests that it acts constitutively, facilitating spliceosome assembly at all polypyrimidine tracts. Half pint, however, appears to possess sequence specificity, since it is capable of influencing splice site choice and affects only a subset of genes analyzed. This apparent discrepancy between the human and Drosophila homologs may come from the artificial nature of the in vitro splicing assay, in which addition of perhaps nonphysiological levels of splicing factors may stimulate the splicing of targets that would not normally be affected in vivo. Alternatively, the orthologs may indeed differ in their target range, a point that could be addressed through in vitro analysis of Hfp activity (Van Buskirk, 2002).

Little is known about the cis-acting regulatory elements that control alternative splicing, posing a challenge for genomic protein prediction models. The recognition element for Hfp is likely to be somewhat degenerate, since the known Hfp targets do not share an obvious binding site. RNA-protein crosslinking studies and/or the identification of several more Hfp target genes may define this binding site, providing insight into the elusive cis-acting sequences that control alternate splice site selection during development (Van Buskirk, 2002).


cDNA clone length - 2415

Bases in 5' UTR - 120

Exons - 6

Bases in 3' UTR - 381


Amino Acids - 637

Structural Domains

The 2.5 kb hfp cDNA corresponds to the observed size of the transcript, which is highly enriched within the ovary and early embryo but which must also be present at other stages of development, given the lethal hfp phenotype. The predicted Hfp open reading frame encodes a 637-amino acid, 68 kDa protein. The central portion of the protein contains two consensus RNA recognition motifs (RRMs), found in a number of RNA binding proteins, including poly(A) binding proteins and splicing factors. The C terminus contains an RRM-like domain that possesses lower similarity to the consensus RRM motif. Database searches reveal that hfp is identical to Drosophila poly(U) binding splicing factor (pUbsf), a gene identified by homology to human PUF60, which was isolated in a screen for poly(U) binding factors that could stimulate splicing in vitro (Page-McCaw, 1999). The isolation of a predicted splicing factor in the two-hybrid screen came as a surprise, given the cytoplasmic localization of Encore (Van Buskirk, 2002).


PUF60/Ro-BP1/FIR/Siah-BP: a homolog of poly-U-binding splicing factor/half pint

Ro ribonucleoprotein particles (Ro RNPs) are complexes of several proteins with a small RNA polymerase III-transcribed Ro RNA. Despite their relative abundance and evolutionary conservation no function has as yet been ascribed to these complexes. Also their subcellular distribution is still largely unknown since immunofluorescence studies concerning their localization have produced conflicting data. Cell enucleation has been used to fractionate cells into cytoplasmic and nuclear fractions. Analysis of these fractions has revealed an exclusively cytoplasmic localization for the Ro RNPs. The majority of the Ro RNAs have been shown to be stably associated with all three known Ro RNP proteins. Although no Ro RNAs could be detected in the nuclear fraction, the Ro RNP-specific proteins were abundantly present. These nuclear non-Ro RNA-associated proteins have been shown to be capable of binding Ro RNAs (Peek, 1993).

A new pyrimidine-tract binding factor, PUF, has been identified that is required, together with U2AF, for efficient reconstitution of RNA splicing in vitro. The activity has been purified and consists of two proteins, PUF60 and the previously described splicing factor p54. p54 and PUF60 form a stable complex in vitro when cotranslated in a reaction mixture. PUF activity, in conjunction with U2AF, facilitates the association of U2 snRNP with the pre-mRNA. This reaction is dependent upon the presence of the large subunit of U2AF, U2AF65, but not the small subunit U2AF35. PUF60 is homologous to both U2AF65 and the yeast splicing factor Mud2p. The C-terminal domain of PUF60, the PUMP domain, is distantly related to the RNA-recognition motif domain, and is probably important in protein-protein interactions (Page-McCaw, 1999).

FUSE-binding protein (FBP) binds the single-stranded far upstream element of active c-myc genes, possesses potent transcription activation and repression domains, and is necessary for c-myc expression. A novel 60 kDa protein, the FBP interacting repressor (FIR), blocks activator-dependent (but not basal) transcription through TFIIH. Recruited through FBP's nucleic acid-binding domain, FIR forms a ternary complex with FBP and FUSE. FIR repressed a c-myc reporter via the FUSE. The amino terminus of FIR contains an activator-selective repression domain capable of acting in cis or even in trans in vivo and in vitro. The repression domain of FIR targets only TFIIH's p89/XPB helicase, required at several stages in transcription, but not factors required for promoter selection. Thus, FIR locks TFIIH in an activation-resistant configuration that still supports basal transcription (Liu, 2000).

Two clones were isolated in a three-hybrid screen of a rat fetal brain P5 cDNA library with an intronic splicing enhancer of the amyloid precursor protein (APP) gene as RNA bait. These clones represent the rat homologs of the previously described genes CUG-binding protein (CUG-BP) and Siah-binding protein (Siah-BP). Both interact in a sequence-specific manner with the RNA bait used for library screening as well as with the CUG repeat. In contrast, no interactions were observed in the three-hybrid assay with other baits tested. In two-hybrid assays, Siah-BP interacts with U2AF65 as well as with itself. EWS, an RGG-type RNA-binding protein associated with Ewing sarcoma, was identified as an interacting partner for the CUG-BP homolog in a two-hybrid assay for protein-protein interactions performed with various factors involved in RNA metabolism. Splicing assays performed by RT-PCR from cells cotransfected with certain cDNAs and an APP minigene, used as a reporter, indicate exclusion of exon 8 if the CUG-BP homolog is present. It is concluded that clone AF169013 and its counterpart in human CUG-BP could be the trans-acting factors that interact with the splicing enhancer downstream of exon 8, and in this way influence alternative splicing of the APP minigene (Poleev, 2000).

poly-U-binding splicing factor/half pint: Biological Overview | Regulation | Developmental Biology | Effects of Mutation | References

date revised: 15 July 2002

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

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