A diverse family of inositol 5-phosphatases playing a role in growth and development in Dictyostelium discoideum

Inositol phosphate-containing molecules play an important role in a broad range of cellular processes. Ino-sitol 5-phosphatases participate in the regulation of these signaling molecules. We have identified four inositol 5-phosphatases in Dictyostelium discoideum , Dd5P1–4, showing a high diversity in domain composition. Dd5P1 possesses only a inositol 5-phosphatase catalytic domain. An unique domain composition is present in Dd5P2 containing a RCC1-like domain. RCC1 has a seven-bladed propeller structure and interacts with G-proteins. Dd5P3 and Dd5P4 have a domain composition similar to human Synaptojanin with a Sac I domain and OCRL with a RhoGAP domain, respectively. We have expressed the catalytic domains and show that these inositol 5-phosphatases have different substrate preferences. Single and double gene inactivation suggest a functional redundancy for Dd5P1, Dd5P2, and Dd5P3. Inactivation of the gene coding for Dd5P4 leads to defects in growth and development. These defects are restored by the expression of the complete protein but not by the 5-phosphatase catalytic domain. Inositol

Inositol phosphates play a role in a variety of eukaryotic cellular processes, including chemotaxis and membrane trafficking.They are regulated by a number of enzymes.The group of phosphatidylinositol 3-kinases (PI3K) 1 phosphorylates the lipid substrates PI, PI(4)P, and PI(4,5)P 2 at the 3-position of the inositol ring (1).The lipid product PI(3,4,5)P 3 has been strongly implicated to be important in chemotaxis in neutrophils and fibroblasts (2,3).PTEN, identified as a tumor suppressor gene (4), reverses the action of PI3K by dephosphorylation of PI(3,4,5)P 3 and PI (3,4)P 2 at the 3-position (5).Another group of enzymes, the inositol 5-phosphatases, can remove the phosphate group at the 5-position of the inositol ring (6 -8).The importance of inositol 5-phosphatase activity in PI (3,4,5)P 3 regulation is demonstrated by SHIP1.In stimulated B-cells, SHIP1 accounts for the major phosphatase activity toward PI(3,4,5)P 3 , and inactivation of SHIP1 leads to an increased and prolonged PI(3,4,5)P 3 production (9).Other inositol 5-phosphatases have been shown to play important roles in a number of cellular processes.Mutations in the inositol 5-phosphatase OCRL are responsible for Lowe syndrome in human (10), and deletion of the presynaptic inositol 5-phosphatase Synaptojanin leads to neurological abnormalities and early death of mice (11).
In the social amoeba, Dictyostelium discoideum chemotaxis toward folic acid and cAMP is an essential strategy for survival (12).Several observations suggest that phosphatidylinositol phosphates mediate chemotaxis and, in particular, the localization of the signal inside D. discoideum cells.The PH domains of a number of proteins involved in chemotaxis, including CRAC, Akt/PKB, and PhdA, have been shown to transiently localize at the leading edge of cells moving in a chemotactic gradient (13)(14)(15).As these PH domains bind to PI(3,4,5)P 3 and PI(3,4)P 2, an asymmetric lipid distribution is implicated by these observations.In pi3k1/2-null cells, a strain with two putative PI3Ks inactivated (16), the transient localization of PhdA can no longer be observed and cells show reduced chemotaxis (15).On the other hand, in PTEN-null cells, a strain in which a putative 3-phosphatase is inactivated, the localization of PH-domains is prolonged and broadened and chemotaxis is also reduced (17,18).Inositol 5-phosphatases may play an important role in the regulation of the phosphoinositides.As this group of enzymes leads to the degradation of PI(3,4,5)P 3 and at the same time formation of PI(3,4)P 2 , another PH-binding molecule, they can be central players in the metabolic route of these signaling molecules.
To investigate the role of inositol 5-phosphatases in chemotaxis and endocytosis, we cloned and characterized four D. discoideum inositol 5-phosphatases.Catalytic activity was determined, indicating that they act as inositol 5-phosphatases.Single and double gene disruptants were obtained and growth, chemotaxis, and development were studied in these knock-out strains.
□ S The on-line version of this article (available at http://www.jbc.org)contains appendices A-F.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AY184992, AY184993, AY184994, and AY184995.

EXPERIMENTAL PROCEDURES
Identification and Sequence Analysis-The first putative inositol 5-phosphatase sequence was obtained using degenerated primers complementary to the conserved motifs I and II present in inositol 5-phosphatases (see "Results"), and the PCR product was used to screen a cDNA library kindly provided by Dr. R. A. Firtel.The D. discoideum genomic (www.sdsc.edu/mpr/dicty)and cDNA databases (www.csm.biol.tsukuba.ac.jp) were screened for other putative inositol 5-phosphatases using either the conserved motif I or II.Using Seqman from Lasergene (DNAstar), contigs were formed.Additional sequences were obtained by screening a cDNA library, kindly provided by Dr. R. H. Gomer, with a PCR fragment containing part of the catalytic domain (used primers: Dd5P2: 5P2S1 ϩ 5P2R1; Dd5P3: 5P5S1 ϩ 5P5R1; and Dd5P4: 5P4S1 ϩ 5P4R1, see "Appendix A").The longest clones obtained for Dd5P3 (4212 bp) and Dd5P4 (2795 bp) encode the complete open reading frame of 1377 and 787 amino acids, respectively.For Dd5P2, the longest clone only encoded for the amino acids 1-599.In combination with data from the genomic data base, a complete open reading frame of 1794 amino acids was constructed.The longest clone obtained for Dd5P1 encodes for amino acids 118 -678; the missing part of the 5Ј-open reading frame was obtained from the data base.A comparison of the sequence obtained from cDNA clones with the genomic data base sequences revealed the presence of one intron in each inositol 5-phosphatase gene (Dd5P1: nucleotides 397-499; Dd5P2: nucleotides 53-187; Dd5P3: nucleotides 241-384; and Dd5P4: nucleotides 1357-1622) (GenBank TM accession numbers AY184992, AY184993, AY184994, and AY184995, respectively).BLAST, Smart, Pfam, and Expasy programs were used to analyze the obtained sequences.Alignments are made using Megalign (Lasergene) and Genedoc followed by optimizing by eye.
Northern Blot Analysis-RNA was isolated from cells grown on plates or starved on non-nutrient plates for 3 or 6 h or from tight aggregates, slugs, or culminants using RNeasy Mini kit from Qiagen.Equal amounts of RNA (ϳ40 g) were loaded on the gel.The probes used to screen the cDNA library were used for hybridization of inositol 5-phosphatase RNA.For Dd5P1, the PCR product of primer 6BIr ϩ 6BIf was used.DNA was labeled with [␣-32 P]dCTP using the random primer method (High Prime, Roche Molecular Biochemicals).Blots were incubated overnight at 65 °C in hybridization solution (0.5 M NaH 2 PO 4 / Na 2 HPO 4 , pH 7.0, 7% SDS, 0.2 mg/ml bovine serum albumin).Stringency of the final washes was 0.1 M NaH 2 PO 4 /Na 2 HPO 4 , pH 7.0, 1% SDS at 65 °C.Bands were visualized using a PhosphorImager.
Gene Disruption-Gene inactivation was obtained by replacing part of the gene by the bsr or pyr5/6 cassette.The cassettes were inserted into NdeI/NsiI site of Dd5P1, NdeI/EcoRI (nucleotide 1210) site of Dd5P2, AccI (nucleotide 1648)/BglII site of Dd5P3, and NcoI (nucleotide 887)/NdeI site of Dd5P4.Orientation of the cassette was opposite to the orientation of Dd5P1(bsr and pyr), Dd5P2(bsr), and Dd5P3(bsr) and similar to the orientation of Dd5P2(pyr) and Dd5P4(bsr and pyr).The place of insertion of the selection marker is indicated by an asterisk in Fig. 1.Single gene inactivation was performed with AX3 cells, whereas double gene inactivation was performed with DH1 cells.Potential knock-outs were screened by PCR and confirmed by Southern and/or Northern blots.As initiation from an internal ATG does not occur, at the most the N-terminal part of the protein will be expressed.A truncated RNA transcript could be observed for strains in which Dd5P2, Dd5P3, or Dd5P4 was inactivated.The size corresponded with the expected size for transcription of the first part of the gene.Therefore, the catalytic inositol 5-phosphatase domain is not transcribed and not expressed.
Western Blot Analysis-The purified inositol 5-phosphatase catalytic domains were analyzed by 10% SDS-PAGE followed by immunostaining with Penta-His Antibody (Qiagen) and Anti-mouse IgG peroxidase conjugate (Sigma).Bands were visualized using chemiluminescence blotting substrate POD (Roche Molecular Biochemicals) and 1-min exposure to film.Prestained Protein Marker, Broad Range (New England Biolabs) was used to determine the size.
Expression of Proteins in D. discoideum-To express proteins in D. discoideum, the desired DNA fragment was cloned into an extrachromosomal vector containing a neomycin cassette (pMB74 or pAH2).This results in expressing the protein from actin 15 promoter.To express the full-length Dd5P4 in Dd5P4 Ϫ cells, the obtained full-length cDNA clone was inserted into the BglII site of pMB74.To clone the inositol 5-phosphatase catalytic domain, the domain was excised from the pRSETB vector (see above) with BamHI and BglII and cloned into the BglII site of pAH2.The RhoGAP domain was obtained by PCR using primer OE4GAPs1 (see "Appendix A") and the universal T7 primer with cDNA as template.The PCR product was digested with BamHI/XhoI and cloned into the BamHI/XhoI site of pRSETB.The insert was sequenced and cloned into the BglII site of pAH2.
Activity toward PI(4,5)P 2 and PI(3,4,5)P 3 was determined as described previously (24).50 g of phosphatidyl serine and 10 g of [ 32 P]PI(3,4,5)P 3 (see below for preparation method) or 20 g of phosphatidyl serine and 10 g of PI(4,5)P 2 /[ 3 H]PI(4,5)P 2 were resuspended in 50 l of 50 mM Tris-HCl, pH 7.4.Vesicles were formed by sonication, and 1 l of 1 M MgCl 2 and 60 l of (diluted) enzyme were added.After 30 min, the reaction was stopped and the lipids were extracted.Lipids were separated by TLC and visualized by exposure to film or Phosphor-Imager.In the case of the PI(4,5)P 2 assay, the spots were visualized using EN3Hance spray (PerkinElmer Life Sciences) and scraped from the plates (Silica Gel 60, Merck) and radioactivity was measured.

Identification of Four Inositol 5-Phosphatase in D. discoideum-
The first potential inositol 5-phosphatase in D. discoideum was identified by performing a PCR with degenerated primers.The obtained PCR product was used as probe for cDNA library screening.The gene found using this method was called Dd5P1 (D. discoideum 5-phosphatase 1) and codes for a protein of 678 amino acids (see "Appendix B").To identify other putative inositol 5-phosphatases in D. discoideum, we searched the D. discoideum data base for sequences showing homology to the conserved motifs I and II, WXGDXN(Y/F)R and P(A/S)W(C/ T)DR(I/V)L respectively, which are characteristic for inositol 5-phosphatases (6).Using the partial sequence obtained from the data base, complete sequences were obtained from cDNA library screens.Three putative inositol 5-phosphatases were identified (see "Appendix B"), coding for proteins consisting of 1800, 1377, and 787 amino acids, respectively.A comparison of the sequence obtained from cDNA clones with the genomic data base sequences revealed the presence of one intron in each inositol 5-phosphatase gene.The position of the introns is indicated by a triangle in Fig. 1.
High Diversity in Inositol 5-Phosphatase Domain Composition Present in D. discoideum-Alignment of the catalytic domain of the four D. discoideum inositol 5-phosphatases with other inositol 5-phosphatase catalytic domains shows a high degree of identity between the amino acid sequence of the four proteins and type II inositol 5-phosphatases (Fig. 2) (see "Appendix C").All of the four proteins contain the conserved amino acids present in motifs I and II that are important for inositol 5-phosphatase activity (26 -28).Also, the amino acids strongly conserved in other regions of the inositol 5-phosphatase domains are present in all four D. discoideum inositol 5-phosphatases, suggesting that they are active inositol 5-phosphatases.
The four inositol 5-phosphatases differ with respect to their domain composition (Fig. 1).Dd5P1 shows the least complex composition containing only the inositol 5-phosphatase domain.A BLAST analysis with the catalytic domain shows the highest score (Expect value ϭ 2e Ϫ54 ) with the catalytic domain of human SHIP2 (30% identity, 44% similarity) (29).This score is clearly higher than scores for non-SHIP-like proteins with the highest score for inositol polyphosphate 5-phosphatase II of Mus musculus (Expect value ϭ 7e Ϫ37 ).The homology with SHIP2 does not expand to the N-terminal part of the protein, because the SH2 domain present in SHIP2 is not present in Dd5P1.
Besides the inositol 5-phosphatase domain, Dd5P2 contains a region homologous to RCC1 ("Appendix D") (30).The structure of RCC1 has been solved, revealing a seven-bladed propeller with each blade comprising by four ␤-sheets (31).RCC1 is known to bind Ran, a small G-protein present at high concentrations in the nucleus.The binding of RCC1 to Ran leads to an increased rate of exchange of GDP for GTP, thus acting as a guanine exchange factor for Ran (32) The other two inositol 5-phosphatases identified in D. discoideum are homologous to human inositol 5-phosphatases.Dd5P3 resembles the synaptojanin-like proteins found in both human and yeast with the highest BLAST score of the catalytic domain with that of INP52p from fission yeast (37% identity, 59% similarity) (34).Similar to human Synaptojanin and yeast INP52p, Dd5P3 has a SacI-like domain including the conserved RXNCXDCLDRTN motif (35) in front of the inositol 5-phosphatase domain (see "Appendix E").The SacI domains of Synaptojanin and INP52p have been shown to remove the phosphate group of PI(4)P, PI(3)P, and at a low rate, both phosphates of PI(3,5)P 2 (36).The long C-terminal part of Dd5P3 does not have any homology with known domains and consists of poly(Q) and poly(N) repeats.
Dd5P4 is homologous to human OCRL (10) and INPP5b (37,38), consisting of a inositol 5-phosphatase domain followed by a RhoGAP domain (see "Appendix F").The catalytic domain has the highest BLAST score with the catalytic domain of human INPP5b (44% identity, 60% similarity).RhoGAP domains are known to catalyze the GTPase activity of Rho proteins.The crystal structure of human RhoGAP has been solved indicating a role for two conserved amino acids, Arg-85 p50RhoGAP and Asn-194 p50RhoGAP , in GAP-activated GTP hydrolysis (39).The role of these amino acids is supported by mutational analysis (39,40).Mutational analysis has also shown that the conserved Arg is not predominantly involved in the binding of Rho proteins (40).In Dd5P4, the Arg and the Asn are substituted by an Ile and Gln, respectively.This observation may suggest that Dd5P4 does not exhibit high RhoGAP activity, but may still bind Rho proteins.
Different Transcription Levels during Development-To determine the transcription levels of the D. discoideum inositol 5-phosphatases, Northern blot analysis was performed.Very low transcription levels were observed for the four inositol 5-phosphatases in all stages of the D. discoideum life cycle (Fig. 3).Dd5P1 and Dd5P3 were equally transcribed in all stages with the exception of the even lower transcription of Dd5P3 in the vegetative stage.The levels of transcription of Dd5P2 were higher during aggregation than during growth and multicellular development.The transcription of Dd5P4 was relatively high during growth, decreased during aggregation, and returned to almost vegetative levels in the multicellular stages.Furthermore, a smaller transcript of 2.5 kilobases was observed in the multicellular stages.
Activity toward Soluble and Lipid Substrates of D. discoideum Inositol 5-Phosphatases-To determine whether the four putative inositol 5-phosphatases identified in D. discoideum can function as inositol 5-phosphatases, the catalytic activity and specificity of the inositol 5-phosphatase domains were studied in vitro.The catalytic domains were expressed as His tag fusion proteins in E. coli and purified.The protein of the expected size could be detected by Western blot analysis for Dd5P1-3 (Fig. 4).The Western blot analysis performed for Dd5P4 showed a band at a higher position than expected (62 instead of 55 kDa), but the purified protein did show inositol 5-phosphatase activity.Unfortunately, because of the very low expression levels obtained for the catalytic domain of Dd5P1, no activity could be determined for this protein.
The purified catalytic domains of Dd5P2, Dd5P3, and Dd5P4 were incubated with the phospholipids PI(4,5)P 2 and PI(3,4, 5)P 3 and the water-soluble inositol phosphates Ins(1,4,5)P 3 and Ins(1,3,4,5)P 4 .Substrate degradation was quantified using thin-layer chromatography for the phospholipids and ion-exchange chromatography for the water-soluble inositol phosphates.The results are summarized in Fig. 5 demonstrating good degradation of PI(4,5)P 2 by all inositol 5-phosphatases, FIG. 4. Western blot analysis of the purified inositol 5-phosphatase catalytic domains used for activity measurements.The catalytic domains were expressed in E. coli, purified by binding to nickel-nitrilotriacetic acid, and eluted with 3 ml of imidazole-containing binding buffer.The proteins were visualized using Coomassie Brilliant Blue staining (A) or using a His-tag antibody followed by chemiluminescence (B).The amount of elute loaded onto the gel was 10 l of each sample (A) or 7.5, 1.5, 1.5, 0.05, and 7.5 l for Dd5P1, Dd5P2, Dd5P3, Dd5P4, and empty pRSETB vector, respectively (B).Sizes of standards (in kilodaltons) are shown on the left.The arrows indicate a band at the expected size for Dd5P1.comparable to wild-type cells.Slug formation was also normal as well as the formation of fruiting bodies.The spores that were formed were viable (data not shown).Also, Dd5P1/2 Ϫ , Dd5P1/3 Ϫ , or Dd5P2/3 Ϫ cells showed no defect in growth or development.
The single disruption of Dd5P4 did affect growth and development.The growth rate in axenic medium was significantly lower for Dd5P4 cells compared with wild-type AX3 cells (Fig. 7).Whereas the doubling time of wild-type cells is ϳ15 h in shaking culture, the doubling time of Dd5P4-deficient cells is 41 h.The growth rate on bacterial lawns was affected as well for Dd5P4 Ϫ cells.Amoeba were deposited on a bacterial lawn and incubated at 22 °C.Approximately 90% wild-type amoeba formed visible plaques within 4 days, whereas Ͻ1% Dd5P4 Ϫ cells formed visible plaques after 9 days.The development of the Dd5P4 cells when grown on bacterial lawns was affected as well.The inactivation of Dd5P4 lead to the formation of multiple-tipped aggregates (Fig. 8).
Overexpression of Full-length Dd5P4 but Not the Separate Domains Rescues Dd5P4 Ϫ Cells-The overexpression of the full-length Dd5P4 restored the defects of Dd5P4 Ϫ cells in growth and development, confirming that the disruption of Dd5P4 was indeed responsible for the observed defects.Growth rate in axenic medium was comparable to the growth rate of wild-type cells (AX3) with a doubling time of 15 h (Fig. 7).When grown on bacterial plates, ϳ90% of the amoeba formed visible plaques within 4 days, and the multiple-tipped phenotype observed for the Dd5P4 Ϫ cells could not be observed anymore (Fig. 8).Overexpression of the inositol 5-phosphatase catalytic domain or the RhoGAP domain of Dd5P4 in the Dd5P4 knock-out strains (Dd5P4 Ϫ /cat5P4 and Dd5P4 Ϫ / GAP5P4, respectively) did not restore any of the defects.As can be seen in Fig. 7, growth rate in axenic medium was not enhanced by overexpression of either one of the domains.Also, growth rate and development on bacterial lawns were not restored (data not shown).

DISCUSSION
The implicated role of phosphoinositide molecules in signal localization has lead to an increased interest in metabolizing enzymes such as inositol 5-phosphatases.The amount of data on mammalian inositol 5-phosphatases has expanded rapidly over the last few years, showing important functions for inositol 5-phosphatases in several processes (6).
Human inositol 5-phosphatases are divided in two groups.Type I inositol 5-phosphatases only convert the water-soluble substrates Ins(1,4,5)P 3 and Ins(1,3,4,5)P 4 .They do not convert lipid substrates.Type II inositol 5-phosphatases do convert phosphoinositides (e.g.PI(4,5)P 2 and PI(3,4,5)P 3 ), and in most cases, they also convert water-soluble substrates.The insertions present in Type I enzymes may prevent the enzymes from binding to the membrane surface, which may explain the differences in substrate specificity between Types I and II inositol   6.Small droplet chemotaxis assay.Chemotaxis toward different concentrations of cAMP (10 Ϫ9 , 10 Ϫ8 , 10 Ϫ7 , and 10 Ϫ6 M cAMP) was determined for wild-type AX3 and knock-out cell lines using the small droplet chemotaxis assay.Asterisks, significant differences from wildtype according to the Student's t test (p Ͻ 0.001) using data of all four cAMP concentrations; NS, no significant difference from wild type.Type II inositol 5-phosphatases can be divided in three subgroups on the basis of domain composition.SHIP1 represents the group of SH2 domain containing inositol 5-phosphatases Synaptojanin, the SacI domain containing enzymes, and OCRL, the RhoGAP domain containing inositol 5-phosphatases (Fig. 1).Mammals contain all three groups of proteins.Six non-mammalian organisms containing putative inositol 5-phosphatases have been sequenced completely so far.The metazoa C. elegans and D. melanogaster lack inositol 5-phosphatases containing a SH2 domain, but both contain one Synaptojanin-like protein ((41) and Q9W296, respectively) and one OCRL-like protein (O17590 and O46049, respectively).The diversity in the plant Arabidopsis thaliana, the yeast strains Saccharomyces cerevisiae and Schizosaccharomyces pombe, and the microsporidia Encephalitozoon cuniculi is even less.The inositol 5-phosphatases of A. thaliana contain only WD40 repeats as additional domains (e.g.Q9SKB7 and O80560), whereas in yeast only Synaptojanin-like proteins are present (28,34).The one inositol 5-phosphatase of E. cuniculi (CAD25856) only contains the inositol 5-phosphatase catalytic domain.In prokaryotes, no inositol 5-phosphatases have been identified.

5-phosphatases (28). The insertions present in
The four inositol 5-phosphatases in D. discoideum show a high diversity in domain composition compared with the evolutionary position of this organism between plants and yeast.A homologue of OCRL (Dd5P4) and a homologue of Synaptojanin (Dd5P3) were identified.Also, a unique combination of a RCC1like domain in front of the inositol 5-phosphatase domain was identified (Dd5P2).This combination of domains has not been found in any other protein reported or present in the Gen-Bank TM so far.
Knock-out strains with one or a combination of two inactive genes for Dd5P1, Dd5P2, and Dd5P3 show no defects in growth and development, suggesting a redundancy among these proteins.Chemotaxis toward cAMP is similar or slightly improved in all of the inositol 5-phosphatase knock-out cell lines compared with wild type.Recently, it has been shown that the 3-phosphatase PTEN is the major PI(3,4,5)P 3 /PI(3,4)P 2 -degrading enzyme.It has been suggested that inositol 5-phosphatases add an addition layer of regulation of these molecules, fine-tuning the chemotactic signal (42).Our results support the role of the group of inositol 5-phosphatases as a minor negative regulator of chemotaxis in D. discoideum.It will be interesting to study the effects of either overexpression or inactivation of the inositol 5-phosphatases in a PTEN-null background.
Inactivation of Dd5P4 resulted in defects in growth and development.The reduced growth rate, either in axenic culture or grown on bacterial plates, suggests a role for Dd5P4 in endocytosis.Development is also affected as the cells form multiple-tipped aggregates.The relative high transcription levels of Dd5P4 at vegetative and multicellular stages support the role of Dd5P4 in growth and development.Knock-out strains of PI3K1 ϩ 2 are defective in chemotaxis, growth, and development (15,16).The cells grow slowly on bacterial lawns and in axenic medium and form multiple-tipped aggregates resembling the phenotype of Dd5P4.No enzyme activity measurements have been reported for the PI3K in D. discoideum, but they have been suggested to catalyze the formation of PI(3)P, PI(3,4)P 2 , and PI(3,4,5)P 3 on basis of sequence homology.Our results support a role for PI(3,4,5)P 3 and/or PI(3,4)P 2 in growth and development; Dd5P4 regulates the levels of these phospholipids by degradation of PI(3,4,5)P 3 and production of PI(3,4)P 2. On the other hand, the effect of inactivation of Dd5P4 could also be assigned to its action on either PI(4,5)P 2 or Ins(1,4,5)P 3 , two signaling molecules also implicated in endocytosis (20,21).In addition, the RhoGAP domain could be responsible for the function of Dd5P4 in endocytosis.Rho proteins have been shown to play a role in both development and growth.The defects in growth and development of Dd5P4 Ϫ cells can be rescued by overexpression of the full-length protein.These defects can not be restored by transfection of D. discoideum cells with an expression vector containing either the inositol 5-phosphatase catalytic domain or the RhoGAP domain.Although we can not exclude that the domains are not properly expressed or folded, the fact that the inositol 5-phosphatase catalytic domain expressed in E. coli is catalytically active would suggest that inositol 5-phosphatase activity is not sufficient to restore the defects of Dd5P4 Ϫ cells.It is possible that the RhoGAP and inositol 5-phosphatase catalytic domain act together to perform its function in growth.The binding of a Rho protein to the RhoGAP domain could affect the inositol 5-phosphatase activity.This would lead to a direct interaction between the Rho and phosphoinositide pathways.It would be interesting to see if and which Rho protein binds to the GAP domain of Dd5P4.
. The highest BLAST score of the catalytic domain of Dd5P2 is shown with the catalytic domain of mouse inositol polyphosphate 5-phosphatase b (INPP5b), a RhoGAP domain containing protein (41% identity, 59% similarity) (33).The long C terminus of Dd5P2 (ϳ1000 amino acids) does not represent a known domain structure.Three poly(N) stretches and a poly(D/E) stretch are present in this C-terminal part of the protein.These kind of stretches are not unusual in D. discoideum proteins.

FIG. 5 .
FIG. 5. Activity of the catalytic domains of inositol 5-phosphatase Dd5P1-4.A, representative TLC analysis of the reaction toward PI(3,4,5)P 3 using 60 l of elute for Dd5P3 and Dd5P4.Results for Dd5P1 were similar to results obtained for the negative control (empty pRSETB vector) (Ϫ).Results obtained for Dd5P2 were similar to results obtained for Dd5P4.PI(3,4,5)P 3 and PI(4,5)P 2 were used as standards.B, representative TLC analysis of the reaction toward PI(4,5)P 2 using 60 l of elute for Dd5P2 and Dd5P3 and 30 l of elute for Dd5P4.Results for Dd5P1 were similar to results obtained for the negative control (Ϫ).C, activity of Dd5P1-4 is expressed in either nanomoles of Ins(1,4,5)P 3 or piccolos of Ins(1,3,4,5)P 4 per min per microgram of protein or in percentage of degraded PI(4,5)P 2 or PI(3,4,5)P 3 per min per microgram of protein.
Type I inositol 5-phosphatases are not present in either one of the D. discoideum inositol 5-phosphatases.As expected by this sequence analysis, the D. discoideum 5-phosphatases Dd5P2, Dd5P3, and Dd5P4 catalyze the dephosphorylation of both water-soluble and lipid substrates, which classify them as Type II inositol 5-phosphatases.The D. discoideum genome has been sequenced to near completion, making it unlikely that more inositol 5-phosphatases containing the motifs WXGDXN(Y/F)R and P(A/S)W(C/T)DR(I/V)L are present.Therefore, probably no Type I and only four Type II inositol 5-phosphatases are present in D. discoideum.

TABLE I
Substrate specificity of Dd5P1-4 towards different substratesCatalytic domains of Dd5P1-4 were expressed in E. coli, purified, and incubated with the four substrates.Every additional "ϩ" indicates an additional doubling in activity.ND, not determined because of low expression levels.