First apyrase splice variants have different enzymatic properties.

LALP70 is a novel lysosomal membrane protein belonging to the apyrase protein family. The apyrase protein family comprises enzymes capable of cleaving nucleotide tri- and diphosphates in a calcium- or magnesium-dependent manner, not being altered by P-type, F-type, or V-type NTPase inhibitors. In this study we have cloned and sequenced the human LALP70 gene to determine the genomic structure. The gene is organized in 11 introns and 12 exons covering a genomic region of approximately 16 kilobase pairs. By fluorescence in situ hybridization analysis, the hLALP70 gene was mapped to the human chromosome 8p21.1-p21.3. We further show that there is at least one alternatively spliced variant, hLALP70v, which can be generated via an alternative splice side at the 3'-end of exon 7, leading to a protein variant differing in 8 amino acids (VSFASSQQ). This is the first splice variant that has been described in the apyrase protein family. Reverse transcriptase polymerase chain reaction analysis showed an ubiquitous expression of both variants, with different relative mRNA expression levels in different tissues. Comparison of the enzymatic properties of the splice variants revealed a broader substrate specificity for hLALP70v with CTP, UDP, CDP, GTP, and GDP as preferred substrates, while hLALP70 utilized UTP and TTP preferentially. Furthermore, enzyme activity of hLALP70v was equally dependent on Ca(2+) and Mg(2+), being saturated already at 1 mm concentration. In contrast, hLALP70 enzymatic activity were unsaturated up to 10 mm Ca(2+), while Mg(2+) showed a saturation at already 1 mm concentration with 2-3-fold lower enzymatic activity as observed with Ca(2+). Our data suggest that the presence or absence of the 8-amino acid motif VSFASSQQ provoke differences in substrate specificity and divalent cation dependence of hLALP70/hLALP70v.

The apyrase protein family comprises enzymes capable of cleaving nucleotide tri-and diphosphates in a calcium-or magnesiumdependent manner, not being altered by P-type, F-type, or V-type NTPase inhibitors (1). Members of the apyrase family contain up to four homolog conserved sequence stretches, which have been described as apyrase conserved regions (ACRs 1 ; Ref. 2). The substrate specificity differs between different apyrases and seems to be sensitive to small changes in the amino acid sequence outside the ACRs. This was shown for the 97% identical apyrase isoforms (NTP1 and NTP3) of Toxoplasma gondii (3,4) and in human brain E-type apyrase after side-directed mutagenesis of two conserved tryptophan residues (5). Apyrases with a substrate specificity restricted to nucleotide triphosphates are classified as E-type ATPases (1,6). Originally, apyrases have been described as ectoenzymes, like CD 39 (7), or other ecto NTPases (1,6,8). However, meanwhile several apyrases have been localized intracellularly (2, 3, 9 -11).
The human lysomal apyrase-like protein (hLALP70) belongs to the apyrase gene family, based on four ACRs in the Nterminal half of the sequence (11). The hLALP70 cDNA has an open reading frame of 1848 bp, encoding a protein with a molecular mass of 71 kDa (11). Based on sequence analysis, hydrophobicity plot and in vitro transcription/translation studies, hLALP70 is thought to be inserted in the membrane as a type III membrane protein (11,12), with one transmembrane domain at the C terminus and two transmembrane domains at the N terminus of the protein. The amino acid sequence of hLALP70 is nearly identical with the sequence of the human UDPase (hUDPase; Ref. 10), except for an additional stretch of 8 amino acids, located 7 amino acids downstream from the ACR4 in the hLALP70 protein. We have isolated and analyzed the hLALP70 gene and provide evidence that the hUDPase is a splice variant of hLALP70. This is the first splice variant described so far for the apyrase family. RT-PCR analysis revealed a differential expression of hLALP70 and its splice variant. Finally we show that the two proteins differing in the 8 amino acids exhibit differences in their enzymatic properties.

EXPERIMENTAL PROCEDURES
Cloning and Analyzing of the Human LALP70 Gene-For isolation of genomic clones containing the human LALP70 gene, high density PAC hybridization filters (RLDB2, Berlin, Germany) were used. As probe a 321-bp fragment of the hLALP70 gene, representing the 3Ј-portion of the cDNA sequence (GenBank TM accession number AJ131358), was amplified by PCR using primers (LALP1f, 5Ј-CGG GGC GTT TCC TTT GTCTAC-3Ј and LALP1r, 5Ј-CTC GGC CTG CAT TTT GTT TTT-3Ј, annealing temperature: 58°C). After hybridization a strong signal for PAC131C11, and weaker signals for several additional clones, were * This work was partially supported by Deutsche Forschungsgemeinschaft Grant EL125/1,2. 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  observed. PAC131C11 was used for the determination of the genomic structure by sequencing.
Sequencing templates were generated by means of PCR using cDNA-derived primers. PCR products were separated by gel electrophoresis and purified using the QIAquick PCR Purification Kit (Qiagen). Approximately 75 ng of purified DNA was used for cycle sequencing. The cycle sequencing was performed according to manufacture's recommendations using ABI Prism BigDye terminator kit (PE Biosystems). Products from cycle sequencing were concentrated using standard ethanol precipitation procedure and analyzed on a ABI 373 automated sequencer (Applied Biosystems). The resulting sequences were aligned by the software package Sequencher (Gene Codes, Ann Arbor, MI).
Chromosomal Localization-The chromosomal localization of human LALP70 gene by fluorescence in situ hybridization (FISH) analysis was carried out on spreads of human lymphocyte metaphase chromosomes, with biotinylated PAC131C11 DNA as probe as described previously (13). In brief, 30 ng of total PAC DNA was labeled with biotin-7-dATP by nick translation (Nick Translation Kit, Life Technologies, Inc.) and detected after hybridization procedure with FITC streptavidin, biotinylated goat anti-streptavidin, and FITC streptavidin (Vector Laboratories, Burlingame, CA). Chromosome counterstaining was performed using DAPI, and slides were embedded in Vectashield (Vector). The preparations were analyzed with a Zeiss epifluorescence microscope, equipped with a CCD camera (Photometrics, Mü nchen, Germany) controlled by the software package Smartcapture (Vysis, Bergisch-Gladbach, Germany).
Radiation Hybrid Panel Screen-Additional to FISH localization, hLALP70 was mapped with the Genebridge 4 radiation hybrid panel consisting 93 radiation hybrid clones of the human genome (Research Genetics). Screen were performed in duplicate, and the results of the PCR analysis were submitted to the server at the Whitehead Institute/ MIT Center for Genome Research's radiation hybrid map of the human genome.
RT-PCR Analysis-For the isolation of total RNA samples from different organs indicated in Fig. 3 were snap-frozen in liquid nitrogen and stored at Ϫ80°C. Tissue samples were extracted with RNAclean (AGS) according to the manufacturer's instructions. 6 g were used for the RT reaction in a total volume of 20 l. As primer the lower primer indicated in Fig. 4A was used (20 pmol). Reverse transcription was performed for 2 h at 42°C with 10 units of a reverse transcriptase from avian myeloblastosis virus (U. S. Biochemical Corp.). 5 l from the RT reaction mix was used for a subsequent PCR in a total volume of 50 l using 20 pmol of each primer indicated in Fig. 3 and 10 units of Taq polymerase (Promega). The annealing temperature was set to 56°C and the PCR was run for 35 cycles. 20 l of the PCR reaction mix were analyzed on a 3% Metaphor-agarose gel (FMC Bioproducts) in TBE buffer (90 mM Tris base, 10 mM boric acid, 2 mM EDTA).
Cloning of Expression Vectors-The full-length hLALP70 cDNA plus 105 bp 5Ј-and approximately 350 bp 3Ј-untranslated regions was subcloned into the pMCS-5 vector (MoBiTec, Göttingen, Germany) using FseI and XhoI restriction sites after a complete digest with FseI and a partial digest with XhoI. The hLALP70v cDNA was generated from the hLALP70-pMCS plasmid by deletion of the 24 bp between nucleotides 1028 -1051 by PCR amplification using the Quick Change Site-Directed Mutagenesis kit from Stratagene. The sense deletion primer 5Ј-GCG TAC GAA GTC CCC AAA ACT GAA GAA GTA GCT-3Ј was complementary to the antisense deletion primer 5Ј-AGC TAC TTC TTC AGT TTT GGG GAC TTC GTA CGC-3Ј. The deletion mutagenesis was performed according to the manufacturer's instruction. The amplified cDNA was provided in Escherichia coli XL1-blue supercompetent cells (Stratagene). The correctness of the 24-bp deletion was confirmed by DNA sequencing. Both hLALP70 and hLALP70v were cloned into the mammalian expression vector pCL-Neo (Promega) using MluI and XbaI restriction sites.
Cell Culture and Transfection-COS-7 cells were cultivated in Dul-becco's modified Eagle's medium supplemented with 2 g/liter Hepes, 5% fetal calf serum, 5% adult calf serum, 50 g/ml gentamycin, and incubated at 37°C in a humidified chamber equilibrated with 5% CO 2 . COS-7 cells were transfected with the mammalian pCL-Neo expression vector alone or with this vector containing either the cDNA for hLALP70 or for hLALP70v, respectively, using a 25-kDa polyethyleneimine (Aldrich) as transfection reagent. Transfection procedure was performed as described elsewhere (Biederbick et al. (11)). Transfected cells were harvested for the apyrase enzymatic assay 24 h after transfection.
Preparation of COS Cell Crude Membranes-Transfected COS-7 cells were homogenized with a Dounce homogenizer, and nuclei were separated as described elsewhere (14). To separate the crude membrane fraction from the cytosol the postnuclear supernatant were centrifuged with 105,000 ϫ g. The pellets were resuspended in 400 l of 20 mM Hepes, pH 7.4, 120 mM NaCl, 5 mM KCl, 0.2 mM EDTA containing 0.1% Triton X-100. The protein concentration in each sample was determined with bicinchoninic acid (Sigma) according to the manufacture's instructions.
Measurement of Nucleotide Phosphatase Activity-To measure apyrase activity COS-7 membrane suspension containing 6 g total protein were adjusted to 45 l with reaction buffer containing 20 mM Hepes, pH 7.4, 120 mM NaCl, 5 mM KCl, 0.2 mM EDTA, 1 mM NaN 3 , and 0.5 mM Na 3 VO 4 , with or without 5 mM CaCl 2 . After preincubation of 5 min at 37°C nucleotide phosphatase reactions were initiated by the addition of 5 l of the same buffer containing 10 mM nucleotide phosphate substrates to give a final concentration of 1 mM. Samples were incubated for 20 min at 37°C. NTP/NDP hydrolysis under these conditions were linear up to 30 min. Apyrase activity was determined by measuring the inorganic phosphate released as described previously (15,16). Values obtained from samples without CaCl 2 or MgCl 2 were subtracted from those obtained with CaCl 2 or MgCl 2 . All measurements were done in triplicate.

Cloning, Sequencing, and Mapping of the hLALP70 Gene-
Primers specific for the human LALP70 gene identified a PAC clone (131C11) with a 80-kb insert (data not shown). This PAC clone was subsequently used for primer-directed sequence walk to analyze the genomic structure of the human LALP70 gene. Our sequence analysis of the hLALP70 gene revealed that it contains 12 exons and 11 introns, covering a genomic region of approximately 16 kb ( Fig. 1A; Table I). The sizes of the exons range from 8 nucleotides (exon 1) to 325 bp (exon 9). Sequence analysis of the intron-exon junctions revealed that all splice sites obeyed the GT-AG paradigm ( Table I). The hLALP70 gene containing PAC131C11 clone was found to hybridize specifically at a single locus on the short arm of chromosome 8 ( Fig.  2A), and no specific signals were observed on any other chromosome. DAPI banding analysis indicated the localization of hLALP70 to the chromosomal region 8p21.1-p21.3 (Fig. 2B). To map the gene more precisely, we used for radiation hybrid mapping the Genbridge G4 panel (Research Genetics, Huntsville, AL) and PCR. Primers for mapping were described above. We determined that the gene is located 0.4 centiRay distal to WI-961 (LOD Ͼ3.1), which is located in the Genethon map between the markers D8S1734 and D8S1820 (44.9 and 54.2 centimorgans from the top of chromosome 8, respectively). In this area the EST sequence KIAA0392, which represents the 3Ј-end of the hLALP70v cDNA (GenBank TM accession number AF016032), has been located previously (17).
hLALP70 is almost identical to a Golgi apyrase (10) except for an additional 24-base pair insert in the hLALP70 sequence. We located these 24 bp at the end of exon 7. The 5Ј-end of the 24-bp stretch provided an alternative splice site (Fig. 1B). Based on this we concluded that the human Golgi apyrase might be a splice variant of hLALP70, further designated as hLALP70v. In the paper published by Wang and Guidotti (10) the authors used a genomic Southern blot approach with EcoRI-or NdeI-digested genomic DNA and a 400-bp fragment from the hUDPase cDNA as hybridization probe, which did not contain a EcoRI or NdeI restriction site. Since they obtained two bands they speculated that there might be two similar genes. Correlation of the 400-bp fragment to the LALP70 gene structure showed that the sequence contained the entire exon 8, as well as parts of exon 7 and exon 9 (Fig. 3A). To confirm that hLALP70 and hLALP70v are transcripts from the same coding sequence, we used PCR primers located in the 400-bp fragment used by Wang and Guidotti (Ref. 10; Fig. 3A, LALP3f/ r). As template we used genomic DNA from six unrelated individuals, all of which gave a single band of approximately 2.9 kb (Fig. 3B, lanes 4 -9, and 3C, lane 4), as expected from the LALP gene structure (Table I and Fig. 3A). When the LALP cDNA was used as template, a 320-bp fragment occurred (Fig.  3B, lane 3). We performed the PCR also with the PAC131C11 DNA as template and obtained the same 2.9-kb fragment (Fig.  3B, lane 2, and 3C, lane 1). Digestion of this fragment, or a fragment generated from genomic DNA (Fig. 3C, lanes 1 and 4) with either EcoRI (Fig. 3C, lanes 2 and 5) or NdeI (Fig. 3C,  lanes 3 and 6), resulted in two bands indicating one restriction site for each enzyme within the 2.9-kb fragment. This was confirmed by sequence analysis of the intron between exons 8 and 9, which contains one EcoRI and one NdeI restriction site. Finally, we digested the PAC131C11 DNA with EcoRI or NdeI, respectively, and performed a Southern blot analysis using the 400-bp fragment as hybridization probe as described in Fig. 3. As shown in Fig. 3D we obtained the two expected bands for each enzyme with an identical hybridization pattern as shown by Wang and Guidotti (10). From these results we have to conclude that hLALP70 and hLALP70v are transcripts from the same gene, generated by alternative splicing.
Identification and Expression of a mLALP Splice Variant-In subsequent experiments we analyzed the expression of mLALP and mLALPv in 14 different mouse tissues by a RT-PCR approach (Fig. 4A). Primers were designed according to a mouse EST (GenBank TM accession number AA497420) representing the hLALP70 mouse homolog (mLALP) containing the 24 bp (Fig. 4A). PCR primer were flanking the 24 bp. The 5Ј-sequence of the upper primer was complemented with a sequence from the human LALP70, since only 11 bp of the 5Ј-upstream mLALP sequence were available from the data base (Fig. 4A). Results from the RT-PCR approach revealed that both mLALP and mLALPv were expressed in all tissues (Fig. 4B). However, the relative level of mLALPv was always much lower than that of mLALP. The highest expression of mLALPv were found in liver and kidney (Fig. 4B). We conclude that mLALP and mLALPv are both ubiquitously expressed, but that the levels of expression might be tissue-specific.
Analysis of Substrate Specificity-In our next experiment we analyzed the enzyme substrate specificity of the human LALP70 and its splice variant hLALP70v to determine the influence of the 8 amino acids VSFASSQQ on the enzyme activity. Data from other apyrases have shown that only slight differences in the sequence can have a marked influence on the substrate specificity (4, 5, 18). Moreover, our own experiments revealed that a green fluorescence protein tag also changed the  Table I. The corresponding GenBank TM accession number is AJ246165. B, part of exon 7 illustrating the alternative splice sites used to generate either hLALP70 or hLALP70v, respectively. substrate specificity (data not shown). For this reason we cloned hLALP70 and hLALP70v into an expression vector without any tag. Using lysates from cells that had been transfected with these constructs we measured the apyrase activity for a variety of NTPs and NDPs as indicated in Fig. 5. While hLALP70 exhibited the highest enzyme activity on UTP and TTP, hLALP70v cleaved CTP most efficiently followed by CDP, UDP, and GTP. Thus, hLALP70v has a broad substrate specificity indicative for a typical apyrase based on a definition by enzymatic properties (1). According to these results, hLALP70v does not represent a typical UDPase as has been published on the basis of enzyme data obtained with a LALP70v/myc fusion protein (10). Analysis of Divalent Ion Dependence-Since apyrase activity usually depends on the presence of divalent cations, we compared hLALP70 and hLALP70v activity under a concentration range of Ca 2ϩ and Mg 2ϩ , respectively. As substrate we used UTP for hLALP70 and CTP for hLALP70v. hLALP70 enzymatic activity showed a linear increase up to 10 mM concentrations of Ca 2ϩ (Fig. 6A). In contrast, hLALP70 apyrase activity was already saturated at 1 mM Mg 2ϩ (Fig. 6A). At 5 mM concentrations the enzyme activity was three to four times higher with Ca 2ϩ than with Mg 2ϩ . With hLALP70v, a rapid increase of enzyme activity was observed up to 1 mM Ca 2ϩ and Mg 2ϩ , comparable with the kinetic observed for hLALP70v (Fig. 6B). However, total enzyme activity of hLALP70v was 2-3-fold higher than of hLALP70 for both Ca 2ϩ and Mg 2ϩ . Enzyme activity was saturated for Ca 2ϩ and Mg 2ϩ concentrations higher than 1 mM (Fig. 6B). Thus, hLALP70v can use either Ca 2ϩ or Mg 2ϩ equally, while LALP70 shows a preference for Ca 2ϩ . We conclude, that hLALP70 and hLALP70v have a different dependence on divalent cations based on structural changes induced by the presence or absence of the 8 amino acids VSFASSQQ. DISCUSSION The human apyrase LALP70 is identical to the human Golgi UDPase (10), with the exception of additional 24 bp, which are located in the central region of the LALP70 cDNA sequence (11). Comparison with other apyrases showed that these 24 bp encoding the amino acids VSFASSQQ are unique in the apyrase protein family (Fig. 7). A mouse EST (GenBank TM accession number AA497420) highly homologous to the human LALP70 also contained these 24 bp (Figs. 4A and 7) with only 2 conservative nucleotide exchanges. This indicates that the 8-amino acid stretch has been conserved at least between mouse and man. Analysis of the hLALP70 gene revealed that the 24 bp are located at the end of exon 7. The 5Ј-junction of the 24 bp contains an alternative splice site. However, Wang and Guidotti (10) performed a Southern blot analysis from which they concluded that there might be another similar gene copy in the genome in addition to their hUDPase gene. They used EcoRI or NdeI for the digestion of genomic DNA and a 400-bp fragment as a hybridization probe. They assumed that this fragment, which did not contain a EcoRI or NdeI site, was entirely located in one exon. Correlation of this 400-bp fragment to our estimated gene structure revealed that the fragment covers the entire exon 8 and parts of exon 7 and exon 9.
Our results further show that there is a EcoRI and a NdeI site in the intron between exon 8 and exon 9 ( Fig. 3C and sequence data from the intron), suggesting that there should be two bands Southern blot analysis after digestion of genomic DNA with one of these enzymes and subsequent hybridization with the 400-bp fragment. This is the case in the paper of Wang and Guidotti (10) and also in the experiment shown in Fig. 3D where the PAC clone DNA was used. The band size and pattern seen in Fig. 3D are identical with those described by Wang and Guidotti (10), where the small EcoRI band results from the two EcoRI sites indicated in Fig. 3A. Finally, PCR analysis of six genomic human DNAs from unrelated individuals (Fig. 3B) further confirmed that there is only one gene copy in the human genome for LALP70 and the hUDPase described by Wang and Guidotti (10). Taken together, with the lack of any indication for more than one gene for hLALP70 and the hUD-Pase, we conclude that the human Golgi UDPase is a splice variant of hLALP70 (hLALP70v). hLALP70 and hLALP70v are the first splice variants that have been described so far in the apyrase protein family.
In preliminary RT-PCR experiments using primer flanking the 24-bp sequence we observed a differential expression of hLALP70 and hLALP70v in different human cell lines (data not shown). To investigate the expression pattern of mLALP and mLALPv in different tissues, we designed PCR primer flanking the 24 bp stretch in the mouse EST (Fig. 4A). Since only a few bases 5Ј-upstream of the 24-bp sequence were given by the EST, we completed the upper primer sequence from the adequate human sequence (Fig. 4A). The RT-PCR results show that in all tissues analyzed, mLALP expression was superior to the expression of mLALPv, but with variations in their ratio (Fig. 4B). However, RNAs used for RT-PCR were from whole tissue samples containing a variety of different cell types. Thus, it cannot be excluded that in some cell types, which contribute only to a small percentage to a given tissue, the mLALP/mLALPv ratio is inverted. Moreover, the turnover rates of the two proteins might be different adjusting the protein amount of mLALP and mLALPv in the cell to other levels as reflected by their relative amount of RNA. While our data clearly show that both mLALP and mLALPv are expressed ubiquitously, the detailed and cell-specific expression levels of these two proteins have to be elucidated in further experiments on RNA and protein level.
In initial experiments to measure the hLALP70 enzyme activity we used hLALP70 and a hLALP70/green fluorescent protein (GFP) fusion protein. We observed that the GFP had an influence on the substrate specificity (data not shown). Thus, in RT-PCR analysis of mLALP70 and mLALP70v expression. A, alignment of the mouse EST homolog to hLALP70. Underlined sequence in small letters shows the 24 bp that are removed in the splice variant hLALP70v. Short bold sequences represent the upper and lower primer, respectively, which were used for RT-PCR. Note that the first 8 nucleotides of the upper primer were derived from the human sequence. B, RT-PCR results obtained with total RNA from mouse organs as indicated. While the 155-bp band represents mLALP, the 131-bp band represents mLALPv.
FIG. 5. Substrate specificity of hLALP70 and hLALP70v. A crude membrane preparation from cells transfected either with a hLALP70 cDNA, or a hLALP70v cDNA, or the vector alone was used to measure the nucleoside phosphatase activity in the presence of NaN 3 and Na 3 VO 4 as described under "Experimental Procedures." Different NTP and NDP substrates were used as indicated. The assay was performed in the presence or absence of 5 mM Ca 2ϩ , and and the data obtained with Ca 2ϩ were corrected with the values obtained without Ca 2ϩ . Mean values from three separate measurements are given. the present study we used hLALP70 and hLALP70v without any tag for the measurement of the enzyme activity. Our results indicate that hLALP70v has a broad substrate specificity with the ability of cleaving all nucleotide di-and triphosphates with the exception of adenosine di-and triphosphate. In contrast to the data published by Wang and Guidotti (10), hLALP70v did not prefer UDP as optimal substrate. Rather, CTP was the preferred substrate, followed by UDP and CDP. This difference might be due to a myc tag, which was present in the construct used by Wang and Guidotti (10). This might have implications in the substrate specificity, comparable with what we observed for the GFP tag. Actually, Wang and Guidotti (10) did not include CTP in their substrate list, and it is not evident whether they used CTP as a substrate at all. Our results depending on hLALP70v without any tag does not confirm that this enzyme is a UDPase.
hLALP70 exhibited a different substrate specificity compared with hLALP70v. The preferred substrates were UTP and TTP, and all other nucleotide di-and triphosphates were not or only to a much lesser extent cleaved. This indicates that the 8 extra amino acids in LALP70 have a strong influence on the substrate specificity of this enzyme. The investigation of other apyrases also revealed a strong dependence of enzyme proper-ties on minor sequence differences. Apyrase isoforms (NTP1 and NTP3) encoded from two different genes in Toxoplasma gondii had an amino acid sequence identity of 97% but differed markedly in their substrate binding and specificity (4). This was attributed to two 12-amino acid-long sequences, FIT-GREMLASID and IVTGGGMLAAIN, which are located at the C terminus of these proteins, where they constitute the region of highest dissimilarity between the two isoforms. However, no similarity between our 8-amino acid sequence and the two 12-amino acid sequences are obvious. Furthermore, a human brain ectoapyrase was used for site-directed mutagenesis of two conserved tryptophan residues (W187A, W459A; Ref. 5), which are also conserved in hLALP70 (Trp 227 , Trp 526 ; Ref. 11). While the W187A mutation abrogated the enzymatic activity, a stimulation of the NTPase activity was observed in case of the W459A mutation. Other single amino acid mutations revealed similar results (6). Finally, in addition to the changes in substrate specificity, the hLALP70-specific motif VSFASSQQ has also an influence on the enzymatic dependence on divalent ions like Ca 2ϩ and Mg 2ϩ .
hLALP70v, which has been described as a Golgi-resident protein (10), is a homolog to the yeast intracellular apyrase GDA1, also located in the Golgi apparatus. GDA1 functions as a GDPase, converting GDP to GMP, which is then transported from the Golgi lumen into the cytosol in exchange with a GDP activated sugar (9). Since GDP-activated sugars are essential for protein glycosylation in the Golgi compartment, GDA1 mutants exhibit less glycosylated proteins and an increased level of GDP in the Golgi cisternae (9). Recently, a mouse UDPase has been cloned and characterized, which is localized in the endoplasmic reticulum (ER-UDPase; Ref. 19). It is thought that the ER-UDPase can unfold a similar function in the ER as GDA1 in the Golgi complex, being critical in the reutilization of UDP during reglycosylation of misfolded glycoproteins (20 -22). ER-UDPase, which sequence is unrelated to that of the hLALP proteins, is a soluble protein. In contrast to hLALP70v and to the ER-UDPase, hLALP70 has been located in autophagic/ lysosomal vacuoles (11). It has to be elucidated by further experiments, whether hLALP70 is also active in the reutilization of NTPs/NDPs from the lysosomal/autophagic compartment, as has been proposed earlier (11). In lysosomes only nucleoside transport systems have been described so far (23), and one can assume that NTPs and NDPs trapped in a lysosomal vacuoles have to be degraded at least to NMPs, in analogy to the Golgi compartment and the rough endoplasmic reticulum (RER), to be transported into the cytosol. However, since the lysosomal compartment does not provide a protein glycosylation machinery, it is unlikely that the outward transport of monophosphate nucleotides or nucleosides is linked to an in-FIG. 6. Dependence of hLALP70 and hLALP70v on divalent ions. Nucleoside phosphatase activity in crude membrane preparations was measured as in the experiment described in the legend to Fig. 4, but with different concentrations of Mg 2ϩ or Ca 2ϩ . A, LALP70 nucleoside phosphatase activity in dependence on Ca 2ϩ and Mg 2ϩ . B, LALP70v nucleoside phosphatase activity in dependence on Ca 2ϩ and Mg 2ϩ .
Although most of a transiently expressed hLALP70/GFP fusion protein was colocalized with lamp1-positive vacuoles, a further association with the Golgi apparatus and the ER could not be ruled out (11). Furthermore, the Golgi localization of hLALP70v has been shown only indirectly (10), and it remains to be shown that this splice variant is exclusively located in this compartment or might also occur in lysosomes. Thus, the question remains open whether the amino acid motif VSFASSQQ also contains sorting information.