Enzymes in the Extracellular Matrix of Volvox: an Inducible, Calcium-dependent Phosphatase with a Modular Composition*

The volvocine algae provide the unique opportunity for exploring development of an extracellular matrix.Volvox is the most advanced member of this family and represents the simplest multicellular organism, with differentiated cells, a complete division of labor, and a complex extracellular matrix, which serves structural and enzymatic functions. InVolvox carteri a glycosylated extracellular phosphatase was identified, which is partially released from the extracellular matrix into the growth medium. The phosphatase is synthesized in response to inorganic phosphate starvation and is strictly calcium-dependent. The metalloenzyme has been purified to homogeneity and characterized. Its gene and cDNA have been cloned. Comparisons of genomic and cDNA sequences revealed an extremely intron-rich gene (32 introns). With an apparent molecular mass of 160 kDa the Volvox extracellular phosphatase is the largest phosphatase cloned, with no sequence similarity to any other phosphatase. This enzyme exhibits a modular composition. There are two large domains and a small one. The large domains are highly homologous to each other and therefore most likely originated from gene duplication and fusion. At least one EF-hand motif for calcium binding was identified in this extracellular protein. Volvoxextracellular phosphatase is the first calcium-dependent extracellular phosphatase to be cloned.

One of the prerequisites to promote the transition from unicellularity to multicellularity during evolution was the development of a complex extracellular matrix (ECM) 1 from a simple cell wall. The ECM of a multicellular organism is a complex organelle that serves structural and enzymatic functions (reviewed in Ref. 1). It mediates many developmental responses including regulation of growth and differentiation, wound repair, and pathogen defense. The ECM also plays a key role when adaptations to environmental changes are required.
The evolutionary transition from unicellularity to multicellularity can be analyzed among the volvocaceae, a family of photoautotrophic green algae. The complexity within this family progresses gradually from unicellular Chlamydomonas via Gonium and Pandorina to Eudorina, forming colonies with an increasing cooperation of individual cells. The most advanced member of the volvocaceae, Volvox, has developed a multicellular form of organization with a complete division of labor between two cell types, somatic and reproductive. In Volvox carteri, about 2000 cells are somatic and only 16 cells are reproductive (2,3).
The ECM of Volvox exhibits anatomically distinct structures; based on light and electron microscopic observations the Volvox ECM was divided into four main zones and several subzones (reviewed in Ref. 4). The Volvox ECM lacks cellulose, pectins, or lignin but mainly consists of glycoproteins, and the structures of several ECM glycoproteins are known in molecular detail (reviewed in Ref. 5).
In the cell walls of higher plants several enzymes such as peroxidases, invertases, mannosidases, glucanases, arabinosidases, galactosidases, polygalacturonase, pectin methylesterases, malate dehydrogenase, proteases, and phosphatases have been identified (reviewed in Ref. 1). However, research on these individual proteins has occurred in different plant species and all sorts of plant organs and cell types. Since higher plant ECMs are extremely complex amalgams, relatively little is known about the precise functions and intermolecular interactions. In these complex systems, it is also difficult to investigate in vivo enzymatic reactions to environmental changes. Therefore, a simple model organism as Volvox might be advantageous for ECM studies. So far, the only ECM glycoproteins of Volvox with proven enzymatic activity are a lysozyme/chitinase (6) and an arylsulfatase (7). I am interested in adaptations within the Volvox ECM in response to environmental stimuli. Since phosphorus is an important nutrient and often limiting in many ecosystems, the role of the ECM during phosphorus deprivation was investigated. In this report an atypical calcium-dependent extracellular phosphatase was identified in Volvox, which is synthesized in response to phosphate deprivation. The enzyme has been isolated and characterized. Likewise, the cloning and characterization of its gene is described.
Standard Volvox medium contains 0.16 mM glycerophosphate. Due to an orthophosphate (P i ) contamination in the glycerophosphate (Sigma) used, standard Volvox medium also contains ϳ25 M orthophosphate. In (glycero)phosphate-free Volvox medium also no orthophosphate was detectable.
Assay of Activity-Routinely, phosphatase activity was measured using p-nitrophenyl phosphate (pNPP) as a substrate. The assay contained 40 mM pNPP, 200 mM Tris-HCl, pH 9.5, and enzyme in a final volume of 1 ml. When the metalloenzyme nature of the phosphatase * This work was supported by the Deutsche Forschungsgemeinschaft (SFB 521). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM 1 The abbreviations used are: ECM, extracellular matrix; HPLC, high performance liquid chromatography; PCR, polymerase chain reaction; pNPP, p-nitrophenyl phosphate; RACE, rapid amplification of cDNA ends; RT, reverse transcription; kb, kilobase pairs; bp, base pairs. was discovered, the assay was supplemented with 5 mM CaCl 2 . After incubation at 37°C for 15 min, the reaction was stopped by adding 100 l of 4 M NaOH. Enzyme activity was quantified by measuring the absorbance at 405 nm.
Purification of Phosphatase-Volvox spheroids (HK10) from three 20-liter cultures were collected on a 100-m mesh nylon screen. After extensive washing with Volvox medium lacking phosphate, the algae were incubated in phosphate-free medium under standard conditions for further 24 h. The Volvox algae were collected as described and disrupted ultrasonically (Sonifier B15; Branson, Danbury, CT), and the lysate was cleared by centrifugation (24,000 ϫ g, 20 min). The crude enzyme solution was fractionated and precipitated with ammonium sulfate (30 -80% saturation; centrifugation at 24,000 ϫ g, 20 min). The precipitated protein was dissolved, desalted by dialysis (20 mM Tris-HCl, pH 8.0) for 15 h, and passed through a Sep-Pak Plus C 18 reversed phase cartridge (Waters, Milford, MA). The flow-through was loaded on a QAE-Sephadex A-25 anion exchange column (Pharmacia, Uppsala, Sweden), and the column was developed by applying a NaCl gradient (up to 1 M) in 20 mM Tris-HCl, pH 8.8, at 1 ml/min. Fractions containing enzyme activity were concentrated by ammonium sulfate precipitation as described. The dissolved protein was further fractionated by molecular-sieve chromatography on a HiLoad 16/60 Superdex 200 preparative grade fast protein liquid chromatography column (Pharmacia, Uppsala, Sweden) in 20 mM Tris-HCl, pH 8.8, 200 mM NaCl at a flow rate of 1 ml/min. Fractions containing the enzyme solution were concentrated using a Centriplus-30 concentrator (Amicon, Beverly, MA). After a short ultracentrifugation step (60,000 ϫ g, 10 min) phosphatase was still in solution, but after a long run (90,000 ϫ g, 18 h) the soluble phosphatase was sedimented. Both centrifugation steps were repeated twice. Finally the enzyme pellet was dissolved in 20 mM Tris-HCl, pH 8.8.
Proteolytic Digestion and Separation of Peptides-50 g of phosphatase were applied to a SDS-polyacrylamide gel electrophoresis (6%). The gel was stained with 0.25% Coomassie Blue, 45% methanol, 19% acetic acid, and destained with 30% methanol, 7% acetic acid. The gel slice containing the extracellular phosphatase was excised, cut into small pieces, and crushed. The gel pieces were treated as follows: 2 times for 1 h 30% methanol, 7% acetic acid; 3 times for 1 h 50% methanol, 10% acetic acid; 2 times for 1 h 90% ethanol. Then the gel pieces were dried in vacuo, subsequently soaked in 4% trypsin, 0.2 M NH 4 HCO 3 and incubated at 37°C for 14 h. The suspension was centrifuged, and the supernatant was kept. After three subsequent washing steps (2 times 0.2 M NH 4 HCO 3 ; 1 time 0.2 M NH 4 HCO 3 , 50% acetonitrile), the supernatant and the wash eluates were pooled, filtered (0.22-m Millex-GV4 filter; Millipore Waters, Bedford, MA), brought to 0.1% trifluoroacetic acid, and dried by lyophilization. The peptides were dissolved in 6 M guanidine HCl, 0.1% CF 3 CO 2 H and fractionated by reversed phase HPLC (SMART system; Pharmacia, Uppsala, Sweden) on a RPC C 2 /C 18 (3 m) column (Pharmacia, Uppsala, Sweden). Peptides were eluted by a 30-min linear gradient from 5 to 40% acetonitrile in 0.1% CF 3 CO 2 H with a flow rate of 200 l/min. Peptides were rechromatographed on a RPC C 18 (3 m) column (Pharmacia, Uppsala, Sweden) under the same conditions. Peptides were sequenced by Edman degradation using an automated gas phase peptide sequencer (Applied Biosystems, Foster City, CA).
Cloning of the Gene-The V. carteri genomic library (13) in EMBL 3 (14) was used to clone the phosphatase gene, named phoX. The screening and cloning procedures followed standard techniques (15). DNA sequencing was performed by the chain termination method (16) using T7 DNA polymerase (Pharmacia, Uppsala, Sweden). Synthetic oligonucleotides were used throughout to sequence the phoX gene.
PCR Amplification of cDNA Fragments-RNA from phosphatestarved (5 h) Volvox was used to construct a cDNA library covalently linked to magnetic beads according to the instructions of the manufacturer of the beads (Deutsche Dynal, Hamburg, Germany). The cDNA fragments of phosphatase were amplified by PCR, using this library. Rapid amplification of cDNA ends (RACE)-PCR technique was performed as described (17). PCR fragments were ligated into the SmaI site of pUC18 or pBluescript II SK vector (Stratagene, La Jolla, CA). Synthetic oligonucleotides were used to sequence the phosphatase cDNA fragments in both directions.
Construction of the Chimeric ␤-Tubulin Phosphatase Gene with Epitope Tag-To achieve high and constitutive phosphatase production, the phosphatase phoX gene was placed under the control of the Volvox ␤-tubulin promoter (18). For construction of the chimeric gene, genomic clones of Volvox ␤-tubulin (nucleotides 40 -501 of tubulin sequence as deposited in the GenBank TM data base, accession number L24547) and phosphatase were used. An additional EcoRV site was generated by PCR directly upstream of the start codons both of the phoX gene and the Volvox ␤-tubulin gene as described (19), to facilitate ligation of the parent DNAs. To introduce a short marker peptide (DYKDDDDK, FLAG epitope tag) at the C terminus of the phosphatase, the gene splicing by overlap extension, or recombinant PCR technique, was performed as described (20). The first PCR was performed on a genomic phosphatase plasmid with a 19-mer 5Ј sense primer (5Ј-CTGTTAACC-TCTTGCTGAT) and a 44-mer 3Ј antisense primer (5Ј-CTACTTGTCA-TCGTCGTCCTTGTAGTCATCCCAGGACGATACGC) with only half of the primer matching; the other half carries the sequence coding for the epitope tag (tag sequence in bold and stop codon in italics). The second PCR was performed on the same template with a 42-mer 5Ј sense primer (5Ј-GACTACAAGGACGACGATGACAAGTAGATGCG-CGCACGTGAC, in which the underlined nucleotides are complementary to the corresponding region of the antisense primer of the first PCR), and a 20-mer 3Ј antisense primer (5Ј-AGTGATGCAATGATCT-TACA). The third PCR was carried out with both of these PCR products and only the flanking primers. PCRs were performed as described (11). All fragments were joined together by standard techniques (15). The complete construct was confirmed by restriction analysis and sequencing.
Stable Transformation of Volvox-All transformation experiments were performed with the Volvox nitrate reductase gene as a selectable marker on a separate plasmid (21). Plasmid DNAs for transformation of Volvox were purified using anion exchange columns (Qiagen, Hilden, Germany). Gold microprojectiles (1 m in diameter; Bio-Rad) were coated with the required targeting plasmids, and stable transformation of Volvox strain 153-48 was performed as described (21), but using a Biolistic PDS-1000/helium particle gun (Bio-Rad). Strain 153-48 carries a stable loss-of-function mutation within the nitrate reductase gene (nitA) (22). Out of 30 NitA-positive transformants, two transgenic Volvox clones were recovered which expressed the phoX gene constitutively.
Preparation of Anti-phosphatase Antiserum-The peptide AIYFP-NIPPAVTEEEKC was synthesized on a SP4000 peptide synthesizer (Labortec, Bubendorf, Switzerland) by using Fmoc (N-(9-fluorenyl)methyloxycarbonyl) amino acid derivatives. The peptide corresponds to the N terminus of the mature phosphatase with an artificial cysteine at the C-terminal end. The predicted molecular mass and the sequence of the peptide were confirmed by electrospray mass spectrometry using a Single-Stage Quadrupole 7000 mass spectrometer (Finnigan, San Jose, CA) and by Edman degradation using an automated gas phase peptide sequencer (Applied Biosystems, Foster City, CA). The peptide was covalently linked to a maleimide-activated carrier protein (keyhole limpet hemocyanin; Pierce) via the SH group of the artificial cysteine. A New Zealand rabbit was immunized (intradermally) with 100 g of phosphatase peptide coupled to 100 g of carrier protein in 250 l of Freund's complete adjuvant (Sigma). The rabbit was boostered three times at the 20th day (subcutaneously), 30th day (intradermally), and 40th day (subcutaneously) from the first immunization on with 100 g of phosphatase peptide coupled to 100 g of carrier protein in 250 l of Freund's incomplete adjuvant (Sigma). At the 61st day 20 ml of blood were withdrawn from the rabbit's ear vein. After blood clotting had taken place, the blood was centrifuged for 10 min at 2,000 ϫ g and the supernatant was kept. 0.02% sodium ethylmercurithiosalicylate were added to the supernatant and stored at 4°C.
Western Blot Analysis-About 4000 Volvox spheroids of wild-type or transformed algae were harvested by filtration on a 100-m mesh nylon screen and disrupted ultrasonically. The lysate was cleared by centrifugation (20,000 ϫ g, 3 min) and used for Western blot analysis. Deep zone extracts were as described previously (23). Samples were separated on a 5% standard SDS-polyacrylamide gel, electroblotted to a polyvinylidene fluoride membrane (0.45 m; Millipore Waters, Bedford, MA), and probed using the polyclonal rabbit anti-phosphatase peptide antibody or a mouse anti-FLAG monoclonal antibody M2 (Eastman Kodak Co.). Signals were visualized using the enhanced chemiluminescence Western blotting analysis system (Pharmacia, Uppsala, Sweden).
Assay of Substrate Specificity-Different possible substrates (p-nitro-phenyl phosphate, glycerophosphate, phosphoenolpyruvate, ribose 5-phosphate, phosphorylcholine, O-phospho-L-tyrosine, AMP, cAMP, NADPH, NADH, 1,2-diacyl-glycero-3-phosphate, L-␣-phosphatidylcholine) were assayed by the P i release during incubation in the presence of the phosphatase (15 min at 37°C). For that purpose 100 l of enzyme/substrate solution containing liberated phosphate were mixed with 1 ml of a solution containing 3 mM ammonium molybdate, 80 mM ascorbic acid, and 0.7 M sulfuric acid (24), and the solution was incubated for 20 min at 37°C. Negative controls were performed without the enzyme. The absorbance at 820 nm was measured, and the liberated phosphate was determined diagrammatically by comparison with a standard curve of phosphate. P i concentrations of the growth medium were determined in the same way. Radioactive labeling of ECM extracts, SSG 185, or pherophorin-S with [ 33 P]phosphate was performed in vivo as described before (23). The different ECM glycoproteins were assayed for remaining [ 33 P]phosphate by comparing the incorporated radioactivity prior to and after incubation with the phosphatase on a SDS-polyacrylamide gel followed by fluorography.

Identification of an Inducible Extracellular Phosphatase in
Volvox-Phosphorus is an important and often limiting nutrient (25). Analyzing the kinetics of the P i decrease in the medium during growth of Volvox, I observed contrary to my expectations an increasing orthophosphate concentration. When Volvox colonies were grown for 6 days (or 3 generations) in standard Volvox medium, the orthophosphate concentration increased from ϳ25 to ϳ100 M, whereas the number of colonies rose by 2 orders of magnitude (up to six colonies/ml). Glycerophosphate supplied with the growth medium is the only possible source for orthophosphate production, indicating the action of a potent (extracellular) phosphatase. Thus, the growth medium and different extracts from Volvox spheroids were assayed for this activity using pNPP as a chromogenic substrate. A high level of phosphatase activity was identified in phosphate-deprived Volvox, independent of the developmental stage used. This activity correlates with the appearance of a ϳ160-kDa protein band on SDS-polyacrylamide gels (Fig. 1A). The 160-kDa protein band is not detectable in Volvox grown in the presence of 1 mM orthophosphate. Phosphate-starved wildtype algae secrete about 50% of the activity into the growth medium, and about 5% are in the deep zone of the Volvox ECM (for nomenclature see Ref. 4). The remaining ϳ45% of activity are bound within the cellular zone of the ECM. In a mutant Volvox strain called "dissociator" (26), which is unable to pro-duce a structurally intact ECM and consequently dissociates into single cells shortly after the end of embryogenesis, about 90% of phosphatase activity and most of the ϳ160-kDa protein are secreted into the growth medium.
These results suggested the existence of an extracellular 160-kDa phosphatase in Volvox, which is inducible by phosphate deprivation.
Purification of the Enzyme-In my experiments, purification of phosphatase from large volumes of growth medium caused problems. Therefore, the phosphatase bound within the cellular zone of the ECM served as the starting material. Volvox spheroids were cultivated in phosphate-free medium for 24 h and then harvested. The Volvox algae were disrupted ultrasonically, and the lysate was cleared by centrifugation. The crude extract was fractionated and precipitated with ammonium sulfate (30 -80% saturation), desalted by dialysis, and passed through a reversed phase C 18 cartridge that did not adsorb the enzyme. The flow-through was loaded on an anion exchange column (QAE-Sephadex), and the column was developed with a NaCl gradient. Fractions containing enzyme activity were concentrated and then further fractionated by molecular sieve chromatography (Superdex 200). Due to self-aggregation, the soluble extracellular phosphatase could be sedimented by ultracentrifugation. This final centrifugation step was repeated twice, leading to a homogenous phosphatase preparation (Fig.  1B).
Protein Chemical Studies-Automated Edman degradation of the purified Volvox phosphatase showed that its N-terminal amino acid sequence is AIYFPNI. In order to get additional sequence information from internal peptides, Volvox phosphatase, localized in a SDS-polyacrylamide gel by Coomassie stain, was digested by incubating the corresponding gel slice with trypsin. The resulting peptide mixture was eluted from the polyacrylamide gel and chromatographed by reversed phase C 2 /C 18 HPLC. All peaks were rechromatographed by C 18 HPLC. Samples of well separated peaks were subjected to amino acid sequence analysis on an automated gas-phase sequencer. The amino acid sequence data obtained are summarized in Table I. Peptide 10 includes a N-glycosylation motif, and amino acid 10 of this peptide gave no signal in automated Edman degradation, which indicates N-glycosylation at this site.
Amplification of a cDNA Fragment by PCR-The amino acid sequence of the tryptic peptide 6 ( Table I) was used to synthesize degenerated oligonucleotide primers. The sense primer 5Ј-GARGAYATHGGNGG was designed from amino acid positions 3-7. The antisense primer 5Ј-TCRTANGTRAAYTG was designed from amino acid positions 13-17 of the same peptide. The latter primer was used to reverse transcribe total RNA of phosphate-starved V. carteri. The resulting cDNA was ampli-  VSVEAALVMPDK fied by PCR using both the sense and antisense primers. Forty cycles of amplification yielded a cDNA fragment of 44 bp in length, which was cloned and sequenced. The deduced amino acid sequence matched the sequence of peptide 6, with glycine at position 8. Genomic and cDNA Clones-A genomic library of V. carteri constructed in the replacement vector EMBL 3 was screened using the 44-bp cDNA fragment as a probe. Eighty-five positive clones were identified out of ϳ60,000 phages screened. The ϳ16-kb insert of one of these phage clones was subcloned and sequenced (phoX-1 in Fig. 2A). It was noticed that the ϳ16-kb insert did not contain the 3Ј-terminal end of the phosphatase gene. Therefore, a genomic probe, derived from the outermost 3Ј-end of the sequenced phage insert, was used for an additional screening of the EMBL 3 library. This produced ϳ90 positive clones, and 5 were characterized. One of these phage clones contained the missing 3Ј-terminal end of the phosphatase gene and, therefore, was subcloned and sequenced (phoX-2 in Fig. 2A). The partially overlapping inserts of the phages phoX-1 and -2 allowed the determination of the complete sequence of the V. carteri phosphatase gene, named phoX, including the 5Ј-and 3Ј-untranslated regions (Fig. 2A).
The complete 5.1-kb cDNA-sequence of the Volvox phosphatase was obtained by RT-PCR and RACE-PCR (17) technology. The 3Ј-RACE-PCRs revealed two different ends of the extracellular phosphatase mRNA, one 45 bp longer than the other. The applied strategy to collect the complete nucleotide sequence of the phosphatase cDNA is summarized schematically in Fig. 2B. The sequence was submitted to the GenBank TM EBI data bank with accession number AJ012458.
The exon-intron organization of the phoX gene was deduced by comparison of genomic and cDNA sequences. Fig. 2A summarizes the relative locations of the 33 exons and 32 introns within ϳ15 kb of phosphatase sequence. The smallest exon encodes for only 11 amino acids. 73% of the genomic DNA sequence between start and stop codon is derived from introns.
The Phosphatase Polypeptide-The deduced amino acid sequence of the Volvox phosphatase is shown in Fig. 3. The PSORT protein localization program (27) predicts a cleavable hydrophobic leader sequence (amino acids 1-23) which is in accordance with the sequenced N terminus of the mature phosphatase. A total of eight possible N-glycosylation sites (NXT and NXS) were found. As mentioned above, glycosylation of one of those sites has been confirmed by protein chemistry (peptide 10, Table I). Based on the cDNA data the molecular mass of the mature phosphatase protein should be 144 kDa, but SDS-gel analysis revealed a somewhat higher molecular mass (ϳ160 kDa). The difference may result from post-translational modification, e.g. glycosylation.
There is an internal sequence similarity within the phosphatase sequence. Amino acids 30 -596 show 46% identity to amino acids 610 -1185, indicating a modular composition of phosphatase; there are two quite similar large domains, I and II, each about 570 amino acids in length, and a small C-terminal domain III, which is about 180 amino acids in length (Fig. 3).
Comparison of the Volvox phosphatase sequence with other phosphatases or all the sequences of the Swiss-Prot protein sequence data base (BLASTP search; Ref. 28) did not reveal any significant sequence similarities.
Immunological Detection-Rabbit polyclonal antibodies were raised against the synthetic N-terminal phosphatase peptide AIYFPNIPPAVTEEEKC. The artificial cysteine at the C terminus of this peptide was used for covalent attachment to a carrier protein (keyhole limpet hemocyanin). The peptide antibody specifically recognizes the 160-kDa phosphatase band in lysates of phosphate-deprived Volvox (Fig. 4A). The quantitative distribution of the immunopositive material in different extracts from phosphate-starved Volvox (growth medium, whole cell lysate, deep zone extract) was just as found for the phosphatase activity described above (data not shown). Identical amounts of phosphatase activity from these extracts or from the purified enzyme were loaded on a SDS-polyacrylamide gel and analyzed by Western blotting (Fig. 4B). There is no significant difference in intensity between the immunopositive band derived from the purified phosphatase and the immunopositive bands derived from different extracts, indicating that the purified extracellular phosphatase is the predominant or sole phosphatase enzyme in these extracts.
Constitutive Overexpression of Phosphatase in Volvox-Transgenic Volvox were generated which express the phoX gene with a C-terminal epitope tag (FLAG tag). To allow an enzymatic differentiation between endogenous and recombinant phosphatase, the recombinant phosphatase was constitu- tively expressed using the strong ␤-tubulin promoter of Volvox (for details see "Experimental Procedures"). Transgenic Volvox algae were produced as described previously (21), and the presence of the chimeric mRNA in transformants was confirmed by RT-PCR (data not shown). The total phosphatase activity in transgenics was about 50 -80-fold higher than in the parent strain (153-48) when both were grown in phosphate-sufficient medium. These results confirm the identity of the phoX gene as the phosphatase gene. In Western blot analysis a monoclonal antibody raised against the FLAG epitope tag stains a 160-kDa band in lysates of the transgenics (Fig. 4C).
Phosphatase release from the ECM into the growth medium exhibited a striking behavior; the higher the density of the Volvox spheroids the more phosphatase was released (the cell membranes remained undamaged). Up to 80% total phosphatase activity could be found in the growth medium if Volvox colonies were incubated in close contact to each other. The concentrated growth medium originating from the same number of transgenics incubated at different cell densities is shown in a Western blot analysis in Fig. 4D. Later I recognized that this density-dependent effect is not only a property of transgenics grown in phosphate-sufficient medium but also of wildtype algae grown in phosphate-deficient medium (data not shown). Possibly, the Volvox ECM is partially degraded by the action of enzymes, if the colonies are in close contact and more and more phosphatase can be released.
The Metalloenzyme Nature-To determine whether divalent cations were required for the activity of the extracellular phosphatase of Volvox, the purified enzyme was treated with the metal ion chelators EDTA or EGTA. Concentrations as low as 5 M EDTA or EGTA resulted in complete inactivation of the enzyme. To investigate whether the phosphatase activity could be recovered after inactivation, the enzyme was treated with 2 mM EGTA, followed by the addition of a variety of metal ion species at 10 mM, namely Cu 2ϩ , Mg 2ϩ , Ca 2ϩ , Fe 2ϩ , Mn 2ϩ , Zn 2ϩ , and Co 2ϩ . Only calcium permitted 70% reactivation of the EGTA-inactivated enzyme (Fig. 5A). Reconstitution of activity by calcium was investigated in detail by adding varying concentrations of CaCl 2 to the EGTA-inactivated enzyme and measuring phosphatase activity (Fig. 5B). All of the metals used were in the form of chloride salts and chloride ions had no detectable effect on phosphatase activity at the concentrations used (data not shown). Accordingly, the extracellular phosphatase of Volvox is a calcium-dependent metalloenzyme. The amino acid sequence of extracellular phosphatase reveals at least one typical "EF hand" motif ( Fig. 5C). This motif is localized at amino acid positions 210 -239 (Fig. 3) within domain I of phosphatase.
Substrate Specificity and Inhibitors-The extracellular phosphatase of Volvox has a broad substrate specificity (Table II). The best substrates are p-nitrophenyl phosphate, glycerophosphate, and phosphoenolpyruvate, but all the other phosphomonoesters tested were also accepted. Since the phosphomonoester within the amino acid O-phospho-L-tyrosine is a substrate, the extracellular phosphatase exhibits a protein phosphatase activity as well. Contrary to these broad phosphomonoesterase activities, the phosphatase is unable to hydrolyze the phosphodiester in cAMP or the phosphoanhydride bond in NADH. Therefore, Volvox extracellular phosphatase is a phosphomonoesterase.
Several phosphatases can effectively hydrolyze the phosphomonoester bond of isolated phosphoproteins (29 -31). The ECM of Volvox owns several phosphorus-containing glycoproteins (32)(33)(34). Their quality as a substrate for the Volvox phosphatase was investigated. The 33 P-labeled ECM glycoproteins pherophorin-S (23), SSG 185 (13), or whole 33 P-labeled ECM extracts were tested for 33 P liberation by the phosphatase (Table II). No 33 P was liberated with any of these substrates, indicating the absence of phosphomonoesters in the ECM structure of Volvox. Since a phosphodiester bridging two arabinose residues via their 5-positions was isolated from hydrolysates of SSG 185 (35) and pherophorin-S (23), this type of phosphodiester might be the predominant or sole phosphorus FIG. 4. Western blot analysis of wild-type or recombinant extracellular phosphatase after SDS-polyacrylamide gel electrophoresis (5%). A, wild-type Volvox were grown in phosphate-sufficient (ϩ) or -deficient (Ϫ) medium for 24 h and then disintegrated by ultrasonic treatment. The crude cell lysates were analyzed by a Western blot using anti-extracellular phosphatase antibodies for detection. B, identical amounts of phosphatase activity (measured at pH 9.5) from purified phosphatase enzyme (pure) and from different extracts of phosphate-starved wild-type Volvox, namely concentrated growth medium (1), an ultrasonic lysate of whole cells (2), and a deep zone extract (3) were analyzed by a Western blot using anti-extracellular phosphatase antibodies for detection. C, wild-type algae and transformants expressing the recombinant FLAG-tagged phosphatase were grown in phosphate-sufficient medium, and lysates of whole cells were analyzed by a Western blot using anti-FLAG antibodies. D, the same number of algae constitutively expressing the tagged phosphatase were incubated in different volumes (density: 100 colonies/ml or 700 colonies/ml) in phosphate-sufficient medium for 12 h. Then the media of both samples were concentrated and analyzed by a Western blot using anti-FLAG antibodies. compound in the ECM of Volvox.
Some substances, besides EGTA and EDTA, were identified that affect phosphatase activity negatively: phosphate, sulfate, and vanadate are inhibitors of the phosphatase from Volvox. In the presence of 20 mM phosphate, sulfate, or vanadate activity decreases to 3, 27, or 15%, respectively. Vanadate is known to be a protein-phosphotyrosine phosphatase inhibitor (36). Different amino acids at concentrations up to 20 mM, such as L-homoarginine, L-cysteine, L-phenylalanine, that inhibit other phosphatases (37,38) had no significant effect on the activity of the Volvox phosphatase (Table II).
Additional Enzymatic Properties-The Volvox extracellular phosphatase exhibits a broad temperature optimum at 25-45°C and is rather heat-sensitive; after 10 min at 56 or 65°C only 35 or 1% of the initial activity remains.
Phosphatases are routinely categorized by the pH required for optimal activity. The extracellular phosphatase of Volvox showed maximum activity at pH 9.5, with only 5% of the maximum activity at pH 7.0. Therefore, the Volvox phosphatase was classified as an alkaline phosphatase (EC 3.1.3.1).
Extracellular phosphatase from Volvox is insensitive toward non-ionic detergents. With 10% Triton X-100 no decrease of activity was detectable. Contrary to non-ionic detergents, the ionic detergent SDS (0.3%) led to complete inactivation/denaturation. DISCUSSION In nearly all lakes, phosphorus is the limiting element that determines the total biomass of algae (25, 39 -43). Most of the soluble phosphate in a lake has been washed out of the sediments. But in many soils most of this soluble phosphate is esterified to organic compounds, making it impossible for many organisms to absorb it directly (44). Since in lakes overgrown with algae the orthophosphate concentration falls into the submicromolar range (40), the possession of an extracellular phosphatase with a broad substrate specificity is a great advantage for algae like Volvox. Besides the synthesis of an extracellular phosphatase, Volvox shows an additional adaptation to phosphorus deprivation. When phosphorus becomes limiting in a lake, Volvox makes use of its capacity for phototaxis and chemotaxis; Volvox accumulates near the surface during the day and swims downward at night to zones near the ground where phosphate or phosphate ester concentrations are clearly higher (45-48). Due to the coordinated power of 4000 -8000 beating flagellas a multicellular Volvox alga reaches a speed of 5 m/h and comes as deep as 30 -40 m below the surface, in contrast to the much slower unicellular algae (48).
It was speculated that Volvox has made another adaptation to phosphorus deprivation (49) 2 : the ECM of Volvox owns many phosphorus-containing glycoproteins (32)(33)(34), and these glycoproteins could be an extracellular phosphorus store. However, no 33 P was liberated from whole 33 P-labeled ECM extracts by the extracellular phosphatase. A phosphodiester bridging two arabinose residues was isolated from several ECM glycoproteins of Volvox (23,35). Thus, all phosphate within the ECM of Volvox could be in the form of phosphodiesters, which have structural or functional tasks, but represent no phosphorus store.
Unlike most phosphatases, the Volvox enzyme requires Ca 2ϩ ions for its activity. The calcium dependence, together with the high pH optimum of the enzyme, might be an adaptation to activity in hard water, where the pH and calcium concentrations (CaCO 3 ) are high. A high concentration of calcium results in a very low concentration of soluble orthophosphate, making an extracellular phosphatase advantageous.
Only two other extracellular phosphatases are known that are also calcium-dependent, one from Haloarcula marismortui and one from Chlamydomonas reinhardtii. The extracellular phosphatase from the halophilic archaebacterium H. marismortui (50) has the same molecular weight and a comparable pH optimum as the Volvox enzyme, but this phosphatase requires 4 M salt for activity. The Chlamydomonas extracellular phosphatase (51) is somewhat larger (190 kDa) than the Volvox enzyme but has the same pH optimum and a comparable substrate specificity. Unfortunately, both these enzymes have not yet been cloned, excluding a comparison at the molecular level. Cloning of the Chlamydomonas phosphatase would be of particular interest for evolutionary studies.
Apart from the requirement for calcium the size of the Volvox phosphatase is striking. Subunit molecular masses of most alkaline phosphatases range from 47 to 87 kDa (52,53). Beside the Volvox enzyme (160 kDa) there are only the two large extracellular phosphatases of Chlamydomonas (190 kDa) and Haloarcula (160 kDa) mentioned above and another of the unicellular cyanobacterium Synechococcus (145 kDa) (54, 55) with a high molecular weight. The Synechococcus phosphatase has been cloned (55), but there is no similarity to the Volvox enzyme, and the Synechococcus enzyme has no modular structure.
The size of the Volvox extracellular phosphatase could play an important role within the ECM. Small soluble proteins leave the ECM mesh-work if they are not cross-linked or differently bound, but ECM enzymes need some mobility for activity (56). Due to its size sufficient amounts of the Volvox phosphatase might be kept back within the ECM mesh-work. In this context, the intramolecular domain duplication found in domains I and 2 A. Hallmann, unpublished data.

TABLE II
Substrate specificity and inhibitors of Volvox extracellular phosphatase Substrate specificity is as follows: the P i release caused by the phosphatase after incubation with different substances at 5 mM was determined; the P i release with pNPP was set at 100% (the K m for pNPP was 30 M). The P i release from the 33 P-labeled substances was determined as described under "Experimental Procedures." Inhibitors are as follows: pNPP hydrolysis by the phosphatase without an inhibitor was set at 100%. II of the Volvox phosphatase could be understood as a way to raise the molecular weight during evolution. Simultaneously this duplication could have given the opportunity to obtain two active centers within one polypeptide. During evolution these domains could have developed a different substrate specificity to reach the broad substrate spectrum of today's phosphatase. The positions and sequences of the 32 introns within the phosphatase gene indicate that the intramolecular domain duplication of domains I and II is not a recent event; only three introns are at the same positions in domains I and II (intron 1 in domain I and intron 16 in domain II, 3 in I, and 18 in II, 9 in I, and 23 in II), and the intron sequences are completely different. Interestingly, there are only few genes having more introns than this phosphatase gene, e.g. the human or mouse genes for the major fibrillar collagens contain 51-54 introns within ϳ30 kb (57). There is another explanation for the intramolecular domain duplication. Some phosphatases like the alkaline phosphatase of Escherichia coli exist as dimers of identical subunits (52). Instead of using two identical molecules to create a functional enzyme, two (nearly) identical domains fused within a single polypeptide could fulfill the same purpose.
This extracellular phosphatase can be used as a tool for genetic experiments in Volvox. The phosphatase structural gene with its C-terminal epitope tag could serve as a reporter gene that results in an enzymatically and immunologically detectable gene product.