In Vivo Structure/Function Analysis of the Drosophila fat facets Deubiquitinating Enzyme Gene

Corresponding author: Janice A. Fischer, The University of Texas at Austin, Moffett Molecular Biology Bldg., 2500 Speedway, Austin, TX 78712., jaf{at}mail.utexas.edu (E-mail)

Communicating editor: K. GOLIC

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The Drosophila Fat facets protein is a deubiquitinating enzyme required for patterning the developing compound eye. Ubiquitin, a 76-amino-acid polypeptide, serves as a tag to direct proteins to the proteasome, a protein degradation complex. Deubiquitinating enzymes are a large group of proteins that cleave ubiquitin-protein bonds. Fat facets belongs to a class of deubiquitinating enzymes called Ubps that share a conserved catalytic domain. Fat facets is unique among them in its large size and also because Fat facets is thought to deubiquitinate a specific substrate thereby preventing its proteolysis. Here we asked which portions of the Fat facets protein are essential for its function. P-element constructs that express partial Fat facets proteins were tested for function. In addition, the DNA sequences of 12 mutant fat facets alleles were determined. Finally, regions of amino acid sequence similarity in 18 Drosophila Ubps revealed by the Genome Project were identified. The results indicate functions for specific conserved amino acids in the catalytic region of Fat facets and also indicate that regions of the protein both N- and C-terminal to the catalytic region are required for Fat facets function.

DEUBIQUITINATING enzymes (DUBs) are a large group of proteins that cleave ubiquitin-protein bonds and whose physiological roles and mechanisms of function are poorly understood. Ubiquitin (Ub) is a 76-amino-acid polypeptide that can be linked covalently to other proteins via an isopeptide bond between an internal lysine on the substrate protein and the terminal glycine residue (G76) of Ub (PICKART 1998 ). Ub chains form through isopeptide linkages most often between an internal lysine residue (K48) of the first Ub and G76 of an incoming Ub monomer (PICKART 1998 ). Monoubiquitination can serve as a signal for endocytosis of a membrane protein (HICKE 1999 ; STROUS and GOVERS 1999 ) or it can modulate protein activity (CHEN et al. 1996 ). In contrast, Ub chains mark proteins for degradation by the proteasome, a multi-subunit proteolytic complex (LUPAS and BAUMEISTER 1998 ; RECHSTEINER 1998 ). Once thought to be a mechanism only for disposing of damaged proteins, it is now well established that Ub-mediated proteolysis is widely used to modulate the levels of critical regulatory proteins (KOEPP et al. 1999 ; MANIATIS 1999 ).

There are two classes of DUBs (WILKINSON and HOCHSTRASSER 1998 ): the Uch enzymes (ubiquitin C-terminal hydrolases) and the Ubp enzymes (ubiquitin processing proteases). The functional distinction between Uchs and Ubps is ambiguous but the two enzyme families have structurally distinct catalytic domains (WILKINSON and HOCHSTRASSER 1998 ). Ubps are the larger class of DUBs; yeast have only 1 Uch but 16 Ubps (WILKINSON and HOCHSTRASSER 1998 ) and the Drosophila Genome Project has identified 4 Uchs and 19 Ubps (RUBIN et al. 2000 ). Ubps are distinguished by their two conserved catalytic domains: the Cys domain, centered around the catalytic cysteine residue and the His domain, containing two catalytically important histidine residues (BAKER et al. 1992 ; PAPA and HOCHSTRASSER 1993 ; HUANG et al. 1995 ; WILKINSON and HOCHSTRASSER 1998 ).

One likely function for DUBs is to generate Ub monomers from precursor proteins; Ub is synthesized in the cell as peptide-linked Ub polymers or Ub-protein fusions (PICKART 1998 ). Other general roles in the Ub pathway have been proposed for two yeast Ubps: Ubp4 (Doa4) and Ubp14 are thought to cleave isopeptide-linked Ub chains, either linked to remnants of degraded proteins (Ubp4) or free (Ubp14), thus preventing them from clogging the proteasome (PAPA and HOCHSTRASSER 1993 ; AMERIK et al. 1997 ). An editing function has been proposed for one vertebrate DUB that may cleave specifically the terminal Ub from short isopeptide-linked chains, thereby deubiquitinating and preventing the degradation of proteins with short Ub chains (LAM et al. 1997 ).

The Drosophila fat facets (faf) gene encodes a Ubp that is essential specificially for patterning the developing eye and also for viability of the early embryo (FISCHER-VIZE et al. 1992 ). Flies with null mutations in the faf gene are viable and have only two obvious defects: their eyes are malformed and females lay eggs that undergo several rounds of nuclear cleavage after fertilization but never cellularize (FISCHER-VIZE et al. 1992 ). In the eye, faf is required in a cell communication pathway that prevents particular precursor cells from misdetermination as neurons; in faf mutant eyes, each facet has more than the normal complement of eight photoreceptors (FISCHER-VIZE et al. 1992 ).

Faf is unique among Ubps in that it has been shown genetically to antagonize the Ub-mediated proteolysis pathway (HUANG et al. 1995 ; WU et al. 1999 ). Thus it has been proposed that Faf deubiquitinates a particular substrate or set of substrates thereby preventing their degradation. Genetic experiments have identified a single gene, called liquid facets, that is likely to encode the critical substrate for Faf in the eye (CADAVID et al. 2000 ). However, Faf may also have other substrates in the Drosophila eye (LI et al. 1997 ) and ovary.

At 2711 and 2778 amino acids, the two similar Faf proteins, generated by altenative splicing of the final exon, are the largest known Ubps (FISCHER-VIZE et al. 1992 ; WILKINSON and HOCHSTRASSER 1998 ). One hypothesis to explain the large size of Faf is that it has several substrates and there may be distinct domains along the length of the protein for recognizing each. Yet, the results of genetic and biochemical experiments suggest that the catalytic region of Faf alone may be largely sufficient for its function (TAYA et al. 1998 , TAYA et al. 1999 ; WU et al. 1999 ).

Faf has mouse and human homologs, called Fam and DFFRX/Y, respectively (JONES et al. 1996 ; WOOD et al. 1997 ) and Fam can substitute for Faf in Drosophila (CHEN et al. 2000 ). The predicted amino acid sequences of these genes are highly conserved along most of their lengths, which precludes their comparison as a means of finding potential functional domains.

As a first step toward gaining insight into how the structure of Faf relates to its function, we generated six different deletion mutants of the faf gene and tested each for function in the developing eyes of P-element-transformed flies. In addition, the DNA sequences of 12 faf alleles with point mutations were determined. Finally, we identified conserved regions among the amino acid sequences of the Ubps in the Drosophila genome. There are two main conclusions from this work. First, distinct amino acid residues within the catalytic region of Faf, other than the catalytic residues themselves, have been defined as essential for Faf function. Second, we found unexpectedly that for its essential function in the eye alone, as well as for its ovary function, protein domains spanning nearly the entire Faf protein, both N-terminal and C-terminal to the catalytic region, are required.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila genetics:
All flies were grown on standard food at 25°. The alleles faf^FO8 and faf^BX4 are described in FISCHER-VIZE et al. 1992 and the P{w+, ro-faf+} transformants are described in HUANG and FISCHER-VIZE 1996 . P-element transformation was performed as described previously (SPRADLING 1986 ; FISCHER-VIZE et al. 1992 ). P{w+} insertions were introduced into a faf^FO8/faf^BX4 background using standard crosses.

Plasmid constructions:
Standard procedures (SAMBROOK et al. 1989 ) were used for all subcloning manipulations. Enzymes used for plasmid construction were obtained from New England Biolabs (Beverly, MA), Promega Biotech (Madison, WI), and Boehringer Mannheim (Indianapolis). All of the deletion constructs derive from a plasmid called pBA-Myc1-cDNA, which is a derivative of Bluescript (Stratagene, La Jolla, CA) with its SmaI site changed to AscI (pBAscI) and that has an 8.5-kb AscI fragment (described in HUANG et al. 1995 ) containing a faf cDNA with a Myc epitope tag between amino acids 53 and 54 cloned into the AscI site. For each of the faf constructs, the DNA sequence at the deletion breakpoint was determined to check that the reading frame was restored. Subsequently, an AscI fragment containing each deletion construct (in pB-faf) was ligated into the AscI site of the pRO transformation vector (HUANG and FISCHER-VIZE 1996 ) and a plasmid with the fragment in the correct orientation was identified. faf1: An 0.55-kb 5'-end fragment of faf was isolated from pBA-Myc1-cDNA restricted with HindIII, treated with mung bean nuclease, and then restricted with AscI. An 7.2-kb 3'-end fragment of faf was isolated from pBA-Myc1-cDNA restricted with ScaI and AscI. To generate pB-faf1, the 5'-end and 3'-end fragments were ligated into pBAscI. faf2: pBA-Myc1-cDNA was restricted with EcoNI and StuI, deleting an 0.5-kb fragment, and then treated with Klenow. To generate pB-faf2, the larger fragment was isolated and religated. faf3: An 3.1-kb 5'-end fragment of faf was isolated from pBA-Myc1-cDNA restricted with AscI and DraI. An 3.8-kb 3'-end fragment of faf was isolated from pBA-Myc1-cDNA restricted with AhdI, treated with T4 DNA polymerase, and then restricted with AscI. To generate pB-faf3, the 5'-end and 3'-end fragments were ligated into pBAscI. faf4: pBA-Myc1-cDNA was restricted with SnaBI and HpaI, deleting an 1.6-kb fragment. To generate pB-faf4, the larger fragment was isolated and religated. faf5: pBA-Myc1-cDNA was restricted with SphI and NcoI, deleting an 1.5-kb fragment, and then treated with T4 DNA polymerase. To generate pB-faf5, the larger fragment was religated. faf6: An 2.3-kb fragment of faf was isolated from pB-faf5 restricted with HpaI and AscI and ligated into pBAscI restricted with AscI and EcoRV. An 2.3-kb 3'-end fragment of faf was isolated from the resulting plasmid restricted with HindIII and AscI. An 0.6-kb 5'-end faf fragment was isolated from pB-faf5 restricted with AscI and HindIII. To generate pB-faf6, the 5'-end and 3'-end fragments were ligated into pBAscI.

Western blot analysis:
Protein extracts were prepared from adult transformants as follows. Twenty adult flies of each genotype were heat-shocked at 37° for 1 hr, allowed to recover at 25° for 1 hr, and then frozen in a dry ice/ethanol bath. The flies were thawed, homogenized in 200 ml of 2x Laemmli buffer, boiled for 5 min, and then spun in a microfuge at 4° for 5 min. The supernatants were boiled for another 5 min, size-separated by SDS-PAGE on a 5% gel, and transferred to nitrocellulose. SDS-PAGE, Western transfer, and hybridization of the blot were according to standard procedures (SAMBROOK et al. 1989 ). The primary antibody was mouse monoclonal anti-Myc (Santa Cruz Biochemical) used at 1:200 dilution and the secondary antibody was HRP-conjugated anti-mouse IgG (Vector Laboratories, Burlingame, CA) used at 1:1000. Blots were developed with ECL Renaissance reagents (Amersham, Arlington Heights, IL) used according to the manufacturer's instructions.

Analysis of Drosophila eyes:
Scanning electron micrographs and 1-µm plastic sections of adult Drosophila eyes were prepared as described previously (HUANG et al. 1995 ). The fraction of wild-type facets was calculated by scoring 100–250 facets per eye in one sectioned eye from at least three different flies of each genotype.

DNA sequence analysis:
Mutant faf alleles were amplified by the polymerase chain reaction (PCR) using total genomic DNA prepared from a single fly homozygous or hemizygous [in trans to Df(3L)faf^BP; FISCHER-VIZE et al. 1992] for each mutant allele. Genomic DNA was prepared by adding the fly to a microfuge tube containing 50 µl of buffer (10 mM Tris pH 8.2, 1 mM EDTA, 25 mM NaCl) and 1 µl of proteinase K (20 mg/ml). The fly was homogenized with a pipet tip and then incubated at 37° for 1 hr and then at 100° for 2 min to inactivate the proteinase K. A 4-µl aliquot of this homogenate was used in a single PCR reaction. Fourteen primer pairs, each of which generated a PCR product ranging in size from 350 to 1000 bp, were used to amplify each faf allele in 14 pieces, each of which was sequenced directly by automated fluorimetric methods. The DNA sequences of each of the seventeen faf gene exons, all intron splice consensus sequences, and all introns except 1, 3, 4, and 16 were determined. To distinguish bona fide allele mutations from PCR-induced mutations, PCR products with non-wild-type DNA sequences were reamplified from genomic DNA a second time and their sequences were determined again. If the same mutation was found a second time, then it was considered to be amplified from the endogenous gene and not a PCR-induced mutation. In all cases, when more than one mutation was found in a single allele, only one reappeared in the second amplification. Details concerning the primer sequences and the PCR reaction conditions will be furnished on request.

Drosophila Ubp amino acid sequence analysis:
The Ubp amino acid sequences were obtained by using Query GadFly (FLYBASE 1999 ) for proteins with UCH motifs. The amino acid sequences obtained were then subjected to BLOCKS analysis (www.blocks.fhcrc.org/blockmkr/make_blocks.html).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Construction of six ro-faf transgenes:
We have shown previously that single amino acid substitutions in the key Cys or His residues of the catalytic domain severely attenuated or abolished the function of Faf in the eye (HUANG et al. 1995 ). In addition, a fragment of Faf containing mainly the catalytic domain can cleave synthetic peptide-linked Ub-protein substrates in bacteria efficiently (HUANG et al. 1995 ). To test whether domains of Faf outside the catalytic region are also important for Faf function in the eye, five different deleted forms of the faf cDNA were constructed (faf1–faf5), each of which encodes a Faf protein with a block of 163 to 514 amino acids removed (Fig 1). In addition, a sixth construct (faf6) containing mainly the catalytic domain was generated. Each of the deletion constructs was generated in the context of a faf cDNA that encodes the smaller of the two forms of Faf protein (2711 amino acids) as this cDNA was shown previously to complement completely the function of the endogenous faf gene in the eye (HUANG et al. 1995 ).

View larger version (18K): In this window In a new window Download PPT slide   Figure 1. Faf proteins. The structures of the Faf protein derivatives expressed by the six different ro-faf genes are shown. Each Faf protein has a Myc-epitope tag inserted between residues 53 and 54 and the catalytic domain is approximately between residues 1670 and 2060 of the 2711 amino acid wild-type Faf protein. The locations of the amino acid residues deleted in each protein are shown and enumerated at right. The relative ability of each construct to substitute for the endogenous faf gene in the eye is also shown at right (see Fig 3A).

Each deleted cDNA construct was cloned into a P-element transformation vector called pRO (HUANG and FISCHER-VIZE 1996 ) that activates transcription using a rough gene enhancer and a heat shock protein 70 promoter. The pRO vector drives expression in a band of undifferentiated cells surrounding the facet preclusters early in eye development and also later in a subset of four of the eight photoreceptor cells (KIMMEL et al. 1990 ; HEBERLEIN et al. 1994 ; DOKUCU et al. 1996 ). We have shown previously that due to their early expression in cells surrounding facet preclusters, transgenes in which the pRO vector drives expression of the wild-type faf cDNA (P{ro-faf+}) complement completely the eye defects in faf null mutants (HUANG and FISCHER-VIZE 1996 ).

Complementation of the faf mutant eye phenotype by the ro-faf transgenes:
P-element transformant lines were generated with each of the six ro-faf constructs, and to test each for function, each P element was introduced into a faf^- background. The particular faf^- background used, faf^BX4/faf^FO8, lacks all or nearly all endogenous faf activity; faf^BX4 is a null mutation and faf^FO8 is a strong mutant allele (FISCHER-VIZE et al. 1992 ). The eyes of flies with two copies of each P-element construct in several independent transformant lines were analyzed. Tangential sections of compound eyes revealing the anatomy of each facet (Fig 2) were scored for the fraction of normally developed as opposed to aberrantly assembled facets (Fig 3A).

View larger version (105K): In this window In a new window Download PPT slide   Figure 2. Eyes of faf mutants expressing ro-faf transgenes. Scanning electron micrographs (A, C, E, and G) and apical tangential sections (B, D, F, and H) of adult Drosophila compound eyes are shown. (A and B) Wild type; the external eye surface is regular and in each facet seven of the eight photoreceptors (1–7) are visible. (C and D) fafBX4/fafFO8; the external eye surface is irregular and in the majority of facets there are more than the wild-type number of photoreceptors. At least some of the ectopic photoreceptor cells arise from the "mystery cells" that are associated with photoreceptor precursors early in eye development (TOMLINSON and READY 1987 ); in faf mutants, the mystery cells are often misspecified as photoreceptors (FISCHER-VIZE et al. 1992 ). (E and F) P{w+, ro-faf1}, fafBX4/fafFO8; one copy of the P element complements the faf mutant eye phenotype well. Most of the facets appear wild type. (G and H) P{w+, ro-faf6}, fafBX4/fafFO8; one copy of the P element complements the faf mutant eye phenotype weakly. Most of the facets are malformed.

View larger version (40K): In this window In a new window Download PPT slide   Figure 3. Complementation of the faf mutant phenotype by ro-faf transgenes. Histograms indicate the fraction of wild-type facets in faf mutant (fafBX4/fafFO8) flies transformed with ro-faf transgenes. (A) The fraction of wild-type facets in faf mutants and in faf mutants containing two copies of a P element expressing ro-faf+ or each of the ro-faf constructs (1–6) is shown. The fraction of wild-type facets is shown for each individual transformant line analyzed. Standard deviations represent variability between flies within a single transformant line. (B) The fraction of wild-type facets in flies as in A is shown. Flies contain one copy (1) or two copies (2) of the particular P element shown at bottom.

Functional analysis of each ro-faf construct in the eye leads to three main observations. First, none of the six deletion constructs retains wild-type activity (Fig 3A). Second, faf2 retains only slightly more activity than ro-faf6 (Fig 3A), suggesting that the 164 amino acids deleted in ro-faf2 (Fig 1) may be highly significant functionally. And finally, ro-faf6, which expresses the smallest Faf protein derivative consisting mainly of the catalytic domain (Fig 1), retains only slight ability to complement faf mutations (Fig 3A).

We wanted to determine whether a small difference in expression level (for example, twofold) could have an effect on the ability of each ro-faf construct to substitute for the endogenous faf gene in the eye. Thus, we analyzed the eyes of faf^- flies bearing single copies of each ro-faf P element and compared their phenotypes with faf^- flies containing two copies of the same P element (Fig 3B). Significant differences in the abilities of ro-faf genes to complement the faf mutant eye phenotype were observed only for ro-faf3 and ro-faf4 lines; for these two constructs, two copies complement the faf mutant eye phenotype significantly more effectively than one copy (Fig 3B).

Relative levels of wild-type and Faf proteins in transformed flies:
To determine whether or not the six Faf proteins and wild-type Faf accumulate similarly when expressed as ro-faf transgenes, Faf protein in extracts from heat-shocked transformant flies were visualized on protein blots. The Faf proteins expressed by the transgenes are Myc-epitope tagged (Fig 1) so that by using anti-Myc antibodies to visualize the Faf proteins only Faf protein expressed by the transgenes was detected. We found that wild-type Faf and the Faf proteins accumulated to similar levels except for Faf6 (Fig 4); in extracts from each of the two ro-faf6 lines, significantly more Faf protein was detected than in any of the other extracts (Fig 4 and data not shown).

View larger version (73K): In this window In a new window Download PPT slide   Figure 4. Faf protein accumulation in transformed flies. Faf proteins expressed from ro-faf transgenes in protein extracts from heat-shocked adult flies were visualized on protein blots using antibodies to the Myc epitope. Wild-type (wt) Faf expressed from ro-faf+ and the six deletion derivatives (Faf1–6) expressed from ro-faf1-6 are shown. No Myc-tagged Faf protein is detected in protein extracts from w- flies containing no P-element construct. Each lane contains protein from one fly equivalent. Arrows indicate Faf and Faf protein bands.

Thus we conclude that the differences in the ability of the ro-faf constructs to complement the faf mutant eye phenotype are not due primarily to an effect on Faf protein accumulation, but rather to differences in the activities of the Faf proteins. Although Faf6 protein accumulates to much higher levels than wild-type Faf and all of the other Faf proteins, ro-faf6 retains nearly no ability to complement faf mutations (Fig 3A). This observation demonstrates clearly that the catalytic domain is insufficient in vivo, at least in the Drosophila eye, for Faf function.

In addition, the particular sensitivity to copy number of the ro-faf3 and ro-faf4 constructs is unlikely to be due to lower Faf3 and Faf4 protein accumulation than the other Fafs or wild-type Faf. Instead, the particular nature of the effect that the deletions have on the activity of Faf could render Faf3 and Faf4 more sensitive to twofold concentration differences than the other Faf proteins with low activity.

DNA sequences of faf alleles with point mutations:
The DNA sequences of 14 point mutant faf alleles were determined and the results are shown in Fig 5. All the mutants have a similar maternal effect lethal phenotype: females homozygous for each mutant allele produce embryos that never cellularize (FISCHER-VIZE et al. 1992 ). By contrast, the alleles fall into two groups on the basis of their mutant eye phenotypes: five of them have extremely weak mutant eye phenotypes (>90% wild-type facets) when homozygous or in trans to strong alleles and seven have strong mutant eye phenotypes (<5% wild-type facets; FISCHER-VIZE et al. 1992 ; Fig 5). Mostly likely, cellularization of the embryo simply requires a higher level of Faf activity than does eye patterning.

View larger version (27K): In this window In a new window Download PPT slide   Figure 5. Molecular mutations in faf alleles. At top are shown the 17 exons (black boxes) and introns (lines connecting the boxes) of the faf gene. The start codon and two alternate stop codons are indicated, as well as the locations of the catalytic domains containing the key cysteine (Cys) and histidine (His) residues. The positions of the DNA lesions in 13 different faf mutant alleles are indicated beneath the exons. The boldface allele names indicate strong mutants and the others are weak mutant alleles. At bottom, the amino acids within each exon are indicated by number. Five of the mutations are due to single base changes that result in altered codon identities: fafFO8 (CACTAC), fafB3 (CAGTAG), fafB6 (TGGTGA), fafB7 (TGGTGA), fafB8 (GAAAAA). One mutation involves two base changes: fafBX5 (GTA just downstream of G2452). Four mutations are small deletions: fafFBB12 (66-bp deletion including the 3' splice acceptor site of intron 4 and part of exon 5), fafBX1 (12-bp in-frame deletion), fafBX3 (15-bp deletion in the middle of intron 11), fafBX15 [deletion of 4 bp (GGGT)]. Two mutations are insertions: fafB4 [deletion of TTT codon and insertion of 18 bp: (AATCCCAACAATCTACTG)], fafB5 [deletion of AG and insertion of 15 bp (TAATTTTTTTTTTAA)]. The fafB6 allele sequence is surprising; it has a stop codon in exon 3 but imparts a weak mutant eye phenotype. The most likely explanation is that there is an alternative splice within intron 2 and exon 3, such that the part of exon 3 containing this lesion is not always used. Two alleles, fafBX8 and fafBX10, have no lesions within any exon or any of the introns sequenced (see MATERIALS AND METHODS). Most likely, their lesions lie within transcriptional control sequences. The fafBX4 allele is an inversion (FISCHER-VIZE et al. 1992 ). In the process of sequencing the mutations, we found two errors in the sequence of exon 17, which, when corrected, resulted in the larger form of Faf protein having a slightly longer open reading frame in exon 17 than reported previously (FISCHER-VIZE et al. 1992 ). These corrections agree with the Genome Project sequence data and have been sent to GenBank. Also, the numbers of the catalytic His residues were reported in error previously (HUANG et al. 1995 ) due to a mistake in sequence numbering (FISCHER-VIZE et al. 1992 ). The correct numbers (His1978 and His1986) are shown here. The faf alleles shown in this figure are the only ones that remain of the point mutations reported in FISCHER-VIZE et al. 1992 , except for fafBX13, which was not sequenced because it appears to have a cytological rearrangement.

Eight of the mutant faf alleles have revealing molecular lesions that allowed them to be sorted into four groups. The first group contains only faf^FO8, in which the second catalytic His residue is changed to Tyr (Fig 5). As faf^FO8 is a strong allele, this confirms the results of previous experiments that indicated the importance of this His residue for Faf function (HUANG et al. 1995 ).

The second group consists of three strong alleles, faf^B8, faf^B7, and faf^BX1, whose molecular lesions define amino acids within the catalytic region that may be essential for function. In faf^B8, a single Glu residue between the Cys and His domains, conserved in Fam (WOOD et al. 1997 ), is changed to Lys. This result could indicate a specific function for the Glu residue in catalysis, as the Glu residue is within a motif conserved among all Drosophila Ubps (BLOCK 3, consensus "D", see below). The faf^BX1 allele contains a deletion of the amino acids MLFY, which are just C-terminal to the His domain and conserved among yeast Ubps (WILKINSON and HOCHSTRASSER 1998 ), Drosophila Ubps (BLOCK 8, see below), and Fam (WOOD et al. 1997 ). This result indicates a requirement for these conserved residues for Ubp function. Similarly, the mutation in faf^B7 introduces a stop codon just prior to the MLFY residues. The faf^BX1 lesion is within the deletion in faf5, thus confirming the importance of this region for Faf function.

The third group consists solely of faf^B4, a weak mutant allele in which a Phe residue near the beginning of Faf protein is replaced by an insertion of six other amino acids (Fig 5). The Phe residue is within the region deleted in faf1, thus confirming a role for the region upstream of the catalytic domain in Faf function.

The molecular lesions within mutations of the fourth group, consisting of three weak mutant alleles, faf^B5, faf^BX5, and faf^BX3, show that Faf protein regions well C-terminal to the His domain are also important for Faf function. The faf^B5 and faf^BX5 alleles have frameshift mutations near the C terminus of Faf; the mutation in faf^BX5 destroys only the terminal 250 amino acids. The faf^BX3 allele has a 15-bp deletion in the middle of intron 11, which is normally only 60 bp in length. This intron may not be spliced at all, in which case a stop codon would be encountered within the intron and a truncated protein, similar to that produced by faf^B5, would be generated. Alternatively, the mutation may render the splicing of intron 11 less efficient, in which case the result would be that some wild-type Faf protein would be present, but less of it.

Comparison of Drosophila Ubp amino acid sequences:
All of the protein sequences classified as Ubps in the GadFly database in Flybase were submitted to BLOCKS analysis (MATERIALS AND METHODS). The results indicated that Drosophila have at least 18, and probably 19, Ubps (Fig 6 legend). Eight blocks of sequence conservation within the catalytic region, including the Cys domain (Block 1) and the His domain (Block 6), were identified (Fig 6). The conserved Blocks are similiar in position and sequence to those found for the yeast Ubps (WILKINSON and HOCHSTRASSER 1998 ) but the motifs have diverged. In addition, Faf is larger than any Ubp previously reported in any organism and is by far the largest Ubp in Drosophila; it is more than twice as large as the next largest Ubp (Fig 6). The size of Faf's catalytic region, however, is moderate. Thus, the bulk of Faf is unique sequence outside the catalytic region.

View larger version (23K): In this window In a new window Download PPT slide   Figure 6. Regions of similarity among the Drosophila Ubps. The 18 Drosophila Ubps are represented as black lines and the eight blocks of similarity, all within the catalytic region, are depicted as colored rectangles. The Ubps are aligned by Block 1, which contains the catalytic Cys residue. A consensus sequence for each block is shown at the bottom left. The amino acids in boldface are invariant and the catalytic Cys and His residues are underlined. The asterisks indicate that no clear consensus amino acids could be identified. The asterisk in parentheses indicates that an amino acid may or may not be present at that position. A probable nineteenth Ubp, CG8334, contains Blocks 1–3, but none of the others, perhaps due to a sequencing error. One protein, CG8232, listed as a Ubp in GadFly, does not have any of the conserved BLOCKS.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Faf is a Ubp required for patterning the Drosophila eye and for cellularization of the embryo. Faf is the largest known Ubp and the only deubiquitinating enyzme thought to counteract the proteolysis machinery by deubiquitinating particular substrates and thereby preventing their degradation. Six different deleted forms of the faf gene were expressed in Drosophila P-element transformants and tested for their ability to substitute for the endogenous faf gene in the eye. In addition, the DNA sequences of 12 point mutant faf alleles were determined. To aid analysis of these results, conserved sequences in the 18 Drosophila Ubps were identified. Unexpectedly, we found that protein domains along the entire length of the Faf protein are required for full activity of the protein.

In the eye, there is genetic evidence that Faf has one critical pathway, and the substrate in this pathway may be the Liquid facets protein (CADAVID et al. 2000 ). Thus, the faf mutant eye phenotype most likely reflects the ability of Faf to locate and deubiquitinate one substrate. The results presented here indicate that domains along virtually the entire primary structure of Faf are required for this one function. Thus, the unusually large size of Faf cannot be explained simply by the presence of multiple substrate-binding domains arranged in a linear fashion along the protein sequence. The same Faf protein domains are required for the eye and ovary functions of Faf; all of the mutants, whether they are weak or strong in the eye, have the same maternal effect lethal phenotype (FISCHER-VIZE et al. 1992 ). Thus, if there is a different substrate in the ovary, domains of Faf both N- and C-terminal to the catalytic region are also required for its recognition and deubiquitination, and these domains overlap those required in the eye.

The results of previous biochemical experiments suggest that only the catalytic domain of Faf may be required for its function. First, when expressed in bacteria, a fragment of Faf containing only the catalytic region can deubiquitinate artificial substrates in the form of peptide-linked Ub-protein fusions (HUANG et al. 1995 ), suggesting that the catalytic domain has the intrinsic ability to recognize Ub. Other experiments suggest that the catalytic domain of Fam can also recognize a specific substrate, even when that substrate is not ubiquitinated. The catalytic domain of Fam specifically binds to particular domains of two proteins, AF-6 and ß-catenin, in vitro and also in cultured cells (TAYA et al. 1998 , TAYA et al. 1999 ). In addition, when expressed in cultured cells, AF-6 can be ubiquitinated and the catalytic domain of Fam can deubiquitinate it (TAYA et al. 1998 ). The in vivo significance of the Fam/AF-6 and Fam/ß-catenin interactions is not yet clear. Fam and AF-6 co-localize in some mouse tissues (KANAI-AZUMA et al. 2000 ), suggesting that their interaction may be significant. However, genetic evidence (CHEN et al. 2000 ) suggests that neither the Drosophila homolog of Af-6 (Canoe, MIYAMOTO et al. 1995 ) or ß-catenin (Armadillo, WIESCHAUS et al. 1984 ) is an important Faf substrate in the eye. Nevertheless, in in vitro and cell culture assays, the catalytic domains of Faf/Fam alone can recognize Ub and specific substrates.

The results of genetic experiments also have suggested that the catalytic domain of Faf alone might be sufficient for its function. When expressed in the fly eye with the same expression vector used here, either of two different yeast Ubps, Ubp2 and Ubp3, can substitute for the endogenous Faf protein more effectively than any of the Faf proteins except for Faf1 (WU et al. 1999 ). [In these experiments, Ubp2 and Ubp3 accumulate to levels similar to Faf6 and Faf+, respectively (WU et al. 1999 ; Z. WU and J. A. FISCHER, unpublished data).] However, neither Ubp2 nor Ubp3 shows obvious amino acid sequence similarity with Faf outside the catalytic domain. One possible explanation for these results is that the catalytic domain of Faf, which is similar to that of Ubp2 and Ubp3, is the essential part of the protein. However, yeast Ubp4, which shares the conserved catalytic domain, cannot substitute for faf (WU et al. 1999 ). There are several alternative explanations. For example, structural similarities between Faf and the two yeast Ubps may have escaped detection. Alternatively, Ubp2 and Ubp3 may recognize the substrate of Faf in a different way than Faf does; the yeast Ubps might have an enhanced ability, relative to Faf, to bind to Ub chains linked to some substrates. This idea seems plausible especially for Ubp2 as it may be a more promiscuous deubiquitinating enzyme than Faf. When overexpressed in yeast, Ubp2 inhibits proteolysis generally and deubiquitinates a variety of substrates (BAKER et al. 1992 ). In addition, a high level of Ubp2 expression in the Drosophila eye disrupts eye development, presumably because it deubiquitinates proteins inappropriately (WU et al. 1999 ).

We conclude that, in vivo, the large unique regions of Faf are required for full protein activity. The large size of Faf appears to reflect a requirement for many Faf protein domains in the recognition of one substrate, rather than a linear array of domains for recognition of multiple substrates. We speculate that, in vivo, Faf might locate its substrate by binding to several different proteins in a complex. Perhaps when Faf/Fam and the substrate is overexpressed in cell culture assays or in vitro, the need for Faf/Fam interactions with other proteins in the complex is overcome by the nonphysiological increase in concentration of both proteins. Further experiments are required to test this idea.

	ACKNOWLEDGMENTS

We thank the University of Texas, Austin, DNA Analysis Center for performing the DNA sequencing. We thank Maya Zhang and Kerrie-Ann Smyth for helping to determine the DNA sequences of two alleles. We also thank J. Mendenhall for performing the scanning electron microscopy, J. Loera for technical assistance, and G. Gage for help in preparing the figures. This work was supported by a grant from the National Institutes of Health (HD30680) to J.A.F. Support for X.C. also came from the Institute for Cellular and Molecular Biology, the Center for Developmental Biology, and the Department of Zoology at the University of Texas at Austin.

Manuscript received June 9, 2000; Accepted for publication August 23, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AMERIK, A. Y., S. SWAMINATHAN, B. A. KRANTZ, K. D. WILKINSON, and M. HOCHSTRASSER, 1997  In vivo disassembly of free polyubiquitin chains by yeast Ubp14 modulates rates of protein degradation by the proteasome. EMBO J. 16:4826-4838[Medline].

BAKER, R. T., J. W. TOBIAS, and A. VARSHAVSKY, 1992  Ubiquitin-specific proteases of Saccharomyces cerevisiae. Cloning of UBP2 and UBP3, and functional analysis of the UBP gene family. J. Biol. Chem. 267:23364-23375[Abstract/Free Full Text].

CADAVID, A. L. M., A. GINZEL, and J. A. FISCHER, 2000  The function of the Drosophila Fat facets deubiquitinating enzyme in limiting photoreceptor cell number is intimately associated with endocytosis. Development 127:1727-1736[Abstract].

CHEN, Z. J., L. PARENT, and T. MANIATIS, 1996  Site-specific phosphorylation to IkappaBalpha by a novel ubiquitination-dependent protein kinase activity. Cell 84:853-862[Medline].

CHEN, X., E. OVERSTREET, S. A. WOOD, and J. A. FISCHER, 2000  On the conservation of function of the Drosophila fat facets deubiquitinating enzyme and Fam, its mouse homolog. Dev. Genes Evol. in press.

DOKUCU, M. E., S. L. ZIPURSKY, and R. L. CAGAN, 1996  Atonal, rough and the resolution of proneural clusters in the developing Drosophila retina. Development 122:4139-4147[Abstract].

FISCHER-VIZE, J. A., G. M. RUBIN, and R. LEHMANN, 1992  The fat facets gene is required for Drosophila eye and embryo development. Development 116:985-1000[Abstract].

FLYBASE,, 1999  The Flybase database of the Drosophila genome projects and community literature. Nucleic Acids Res. 27:85-88. Available from http://flybase.bio.indiana.edu/[Abstract/Free Full Text].

HEBERLEIN, U., A. PENTON, S. FALSAFI, D. HACKETT, and G. M. RUBIN, 1994  The C-terminus of the homeodomain is required for functional specificity of the Drosophila rough gene. Mech. Dev. 48:35-49[Medline].

HICKE, L., 1999  Gettin' down with ubiquitin: turning off cell-surface receptors, transporters and channels. Trends Cell Biol. 9:107-111[Medline].

HUANG, Y. and J. A. FISCHER-VIZE, 1996  Undifferentiated cells in the developing Drosophila eye influence facet assembly and require the Fat facets deubiquitinating enzyme. Development 122:3207-3216[Abstract].

HUANG, Y., R. T. BAKER, and J. A. FISCHER-VIZE, 1995  Control of cell fate by a deubiquitinating enzyme encoded by the fat facets gene. Science 270:1828-1831[Abstract/Free Full Text].

JONES, M. H., R. A. FURLONG, H. BURKIN, I. J. CHALMERS, and G. M. BROWN et al., 1996  The Drosophila developmental gene fat facets has a human homologue in Xp11.4 which escapes X-inactivation and has related sequences on Yq11.2. Hum. Mol. Genet. 5:1695-1701[Abstract/Free Full Text].

KANAI-AZUMA, M., J. S. MATTICK, K. KAIBUCHI, and S. A. WOOD, 2000  Co-localization of Fam and Af-6, the mammalian homologs of Drosophila faf and canoe, in mouse eye development. Mech. Dev. 91:383-386[Medline].

KIMMEL, B. E., U. HEBERLEIN, and G. M. RUBIN, 1990  The homeodomain protein Rough is expressed in a subset of cells in the developing Drosophila eye where it can specify photoreceptor cell subtype. Genes Dev. 4:712-727[Abstract/Free Full Text].

KOEPP, D. M., J. W. HARPER, and S. J. ELLEDGE, 1999  How the cyclin became a cyclin: regulated proteolysis in the cell cycle. Cell 97:431-434[Medline].

LAM, Y. A., W. XU, G. N. DEMARTINO, and R. E. COHEN, 1997  Editing of ubiquitin conjugates by an isopeptidase in the 26S proteasome. Nature 385:737-740[Medline].

LI, Q., I. K. HARIHARAN, Y. HUANG, and J. A. FISCHER, 1997  Genetic interactions with Rap1 and Ras1 reveal a second function for the fat facets gene in Drosophila eye development. Proc. Natl. Acad. Sci. USA 94:12515-12520[Abstract/Free Full Text].

LUPAS, A., and W. BAUMEISTER, 1998 The 20S proteasome, pp. 127–146 in Ubiquitin and the Biology of the Cell, edited by J.-M. PETERS, J. R. HARRIS and D. FINLEY. Plenum Press, New York.

MANIATIS, T., 1999  A ubiquitin ligase complex essential for the NF-kappaB, Wnt/Wingless, and Hedgehog signaling pathways. Genes Dev. 13:505-510[Free Full Text].

MIYAMOTO, H., I. NIHONMATSU, S. KONDO, R. UEDA, and S. TOGASHI et al., 1995  canoe encodes a novel protein containing a GLGF/DHR motif and functions with Notch and scabrous in common developmental pathways in Drosophila. Genes Dev. 9:612-625[Abstract/Free Full Text].

PAPA, F. R. and M. HOCHSTRASSER, 1993  The yeast doa4 gene encodes a deubiquitinating enzyme related to a product of the human tre-2 oncogene. Nature 366:313-319[Medline].

PICKART, C. M., 1998 Polyubiquitin chains, pp. 19–63 in Ubiquitin and the Biology of the Cell, edited by J.-M. PETERS, J. R. HARRIS and D. FINLEY. Plenum Press, New York.

RECHSTEINER, M., 1998 The 26S proteasome, pp. 147–189 in Ubiquitin and the Biology of the Cell, edited by J.-M. PETERS, J. R. HARRIS and D. FINLEY. Plenum Press, New York.

RUBIN, G. M., M. D. YANDELL, J. R. WORTMAN, G. L. G. MIKLOS, and C. R. NELSON et al., 2000  Comparative genomics of eukaryotes. Science 287:2204-2215[Abstract/Free Full Text].

SAMBROOK, G., E. F. FRITSCH and T. MANIATIS, 1989 Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SPRADLING, A. C., 1986 P-element mediated transformation, pp. 175–197 in Drosophila: A Practical Approach, edited by D. B. ROBERTS. IRL Press, Oxford.

STROUS, G. J. and R. GOVERS, 1999  The ubiquitin-proteasome system and endocytosis. J. Cell Sci. 112:1417-1423[Abstract].

TAYA, S., T. YAMAMOTO, K. KANO, Y. KAWANO, and A. IWAMATSU et al., 1998  The Ras target AF-6 is a substrate of the Fam deubiquitinating enzyme. J. Cell Biol. 142:1053-1062[Abstract/Free Full Text].

TAYA, S., T. YAMAMOTO, M. KANAI-AZUMA, S. A. WOOD, and K. KAIBUCHI, 1999  The deubiquitinating enzyme Fam interacts with and stabilizes ß-catenin. Genes Cells 4:757-767[Abstract].

TOMLINSON, A. and D. F. READY, 1987  Neuronal differentiation in the Drosophila ommatidium. Dev. Biol. 120:366-376[Medline].

WIESCHAUS, E., C. NUSSLEIN-VOLHARD, and G. JURGENS, 1984  Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster: zygotic loci on the X-chromosome and the fourth chromosome. Roux's Arch. Dev. Biol. 193:296-307.

WILKINSON, K. D., and M. HOCHSTRASSER, 1998 The deubiquitinating enzymes, pp. 99–125 in Ubiquitin and the Biology of the Cell, edited by J.-M. PETERS, J. R. HARRIS and D. FINLEY. Plenum Press, New York.

WOOD, S. A., W. S. PASCOE, K. RU, T. YAMADA, and J. HIRCHENHAIN et al., 1997  Cloning and expression analysis of a novel mouse gene with sequence similarity to the Drosophila fat facets gene. Mech. Dev. 63:29-38[Medline].

WU, Z., Q. LI, M. FORTINI, and J. A. FISCHER, 1999  Genetic analysis of the role of the Drosophila fat facets gene in the ubiquitin pathway. Dev. Genet. 25:312-320[Medline].

This article has been cited by other articles:

				T. Tsichritzis, P. C. Gaentzsch, S. Kosmidis, A. E. Brown, E. M. Skoulakis, P. Ligoxygakis, and G. Mosialos A Drosophila ortholog of the human cylindromatosis tumor suppressor gene regulates triglyceride content and antibacterial defense Development, July 15, 2007; 134(14): 2605 - 2614. [Abstract] [Full Text] [PDF]

				J. A. Fischer, S. Acosta, A. Kenny, C. Cater, C. Robinson, and J. Hook Drosophila Klarsicht Has Distinct Subcellular Localization Domains for Nuclear Envelope and Microtubule Localization in the Eye Genetics, November 1, 2004; 168(3): 1385 - 1393. [Abstract] [Full Text] [PDF]

				E. Overstreet, E. Fitch, and J. A. Fischer Fat facets and Liquid facets promote Delta endocytosis and Delta signaling in the signaling cells Development, November 1, 2004; 131(21): 5355 - 5366. [Abstract] [Full Text] [PDF]

				K. Patterson, A. B. Molofsky, C. Robinson, S. Acosta, C. Cater, and J. A. Fischer The Functions of Klarsicht and Nuclear Lamin in Developmentally Regulated Nuclear Migrations of Photoreceptor Cells in the Drosophila Eye Mol. Biol. Cell, February 1, 2004; 15(2): 600 - 610. [Abstract] [Full Text] [PDF]

				J. H. Kim, K. C. Park, S. S. Chung, O. Bang, and C. H. Chung Deubiquitinating Enzymes as Cellular Regulators J. Biochem., July 1, 2003; 134(1): 9 - 18. [Abstract] [Full Text] [PDF]

				N. A. Boase, R. A. Lockington, J. R. J. Adams, L. Rodbourn, and J. M. Kelly Molecular Characterization and Analysis of the acrB Gene of Aspergillus nidulans: A Gene Identified by Genetic Interaction As a Component of the Regulatory Network That Includes the CreB Deubiquitination Enzyme Genetics, May 1, 2003; 164(1): 95 - 104. [Abstract] [Full Text] [PDF]

				M. T. Pena-Rangel, I. Rodriguez, and J. R. Riesgo-Escovar A Misexpression Study Examining Dorsal Thorax Formation in Drosophila melanogaster Genetics, March 1, 2002; 160(3): 1035 - 1050. [Abstract] [Full Text] [PDF]

				X. Chen, Q. Li, and J. A. Fischer Genetic Analysis of the Drosophila DNAprim Gene: The Function of the 60-kD Primase Subunit of DNA Polymerase Opposes the fat facets Signaling Pathway in the Developing Eye Genetics, December 1, 2000; 156(4): 1787 - 1795. [Abstract] [Full Text]