CHARACTERIZATION OF SEA URCHIN TRANSGLUTAMINASE, A PROTEIN REGULATED BY GUANINE/ADENINE NUCLEOTIDES

Transglutaminases (TGs) are calcium-dependent enzymes that catalyze transamidation of glutamine residues to form intermolecular isopeptide bonds. Nine distinct TGs have been identified in mammals, and three of them (types 2, 3 and 5) are regulated by GTP/ATP and are able to hydrolyze GTP, working as bifunctional enzymes. We have isolated a cDNA clone encoding a TG from a cDNA library prepared from the blastula stage of sea urchin Paracentrotus lividus ( PlTG ). The cDNA sequence has an open reading frame coding for a protein of 738 amino acid, including a Cys active-site, and the other two residues critical for catalytic activity, His and Asp. We have studied its expression pattern by in situ hybridization and also demonstrated that the in vitro expressed PlTG had GTP- and ATP-hydrolyzing activity; moreover, GTP inhibited the transamidating activity of this enzyme as that of human TG2, TG3, and TG5.


Introduction
Transglutaminases (TGs) are a diverse family of Ca 2+ -dependent enzymes with distinct genes, structures and biological functions which catalyze post-translational modification of proteins referred as the R-glutaminyl-peptide, amine--glutamyl transferase reaction which leads either to the formation of an isopeptide bond within or between polypeptide chains, or to the covalent incorporation of polyamines into protein substrates (1,2).
TGs are widely distributed in various organisms. In mammals, nine distinct but closely related transglutaminases, have been identified: TGs 1 to 7, coagulation factor XIIIa and the catalytically inactive band 4.2 (1,2). Homologues have been found in invertebrates, in slime molds, plants and bacteria. A phylogenetic analysis of all the TGs indicated that an early gene-duplication event might have given rise to two lineages: one that comprises TG2, TG3, TG5, TG6 and TG7, and erythrocyte band 4.2; and the other to fXIIIa, TG1, TG4 and invertebrate TGs (2).
Among the TG family members, TG2, TG3, TG5 and slime mold TG (3) have been shown to be regulated by guanine nucleotides. TG2 is involved in several biological processes such as cell adhesion, apoptosis, and wound healing (4,5). This enzyme is multifunctional with both proteincross-linking and GTP-hydrolyzing activities (1); it has also been shown to be capable of functioning as a signal-transducing GTP-binding protein, coupled to activated receptors (6, 7), though the physiological significance of this function is yet to be elucidated. The X-ray structure of human TG2 bound with guanosine 5'-diphosphate (GDP) is known, and the main GDP-binding residues have been identified (8) and binding constants for nucleotides have been measured (9). 4 TG cDNAs have been isolated from lower verterbrates, such as fish, and the genes have been found to have structural similarity with those of mammalian genes (12,13). TG cDNAs of a few invertebrates; ascidians (14), Drosophila (15), grasshopper (annulin) (16), limulus (17), crayfish (18) tiger shrimp (Penaeus monodon) (19), starfish (20) have also been cloned but the physiological roles of these invertebrate TGs remain unclear.
Previous results obtained by Cariello et al. (21)(22)(23) strongly suggested that TG or TGs might be involved at various stages in embryonic development of sea urchin. Although these reports described the identification and purification of sea urchin TG activity, no structural information has been presented. To find out more about invertebrate TGs, their physiological roles, and evolutionary relationship to other TGs, we attempted the molecular cloning of a Paracentrotus lividus TG (PlTG). In this study a cDNA clone encoding PlTG was isolated and its expression pattern was analysed by in situ hybridization. By biochemical characterization of in vitro expressed PlTG, we demonstrated that this enzyme showed a calcium-dependent transamidating activity, that was negatively regulated by guanine nucleotides. Furthermore, we demonstrated that PlTG was able to hydrolyze GTP and ATP although there is a low sequence homology with the GTP-binding domain of TG2. PlTG is the first transglutaminase characterized from an invertebrate organism that displays ATP-and GTP-hydrolyzing activity.

Animals and embryos
Adult specimens of Paracentrotus lividus were collected in the Bay of Naples. Specimens were shed by KCl (0,5M) injection and gametes were collected in filtered sea water. Eggs were fertilized with suspensions of sperm and cultured at 18 °C. When the embryos reached the appropriate stage, they were packed by hand centrifugation, frozen at -20°C for RNA preparation or fixed for in situ hybridization.

Probe preparation
Poly(A) + RNA from gastrula stage of sea urchin P. lividus was reverse transcribed using the SuperScript Preamplification System for First Strand cDNA Synthesis (GIBCO/BRL) and then amplified directly using PCR according to manufacturer's instructions. The following degenerate primers were used to amplify the coding region of the active-site of a putative sea urchin TG: 5'-GTSMMVTAYGGMCAGTGCTGGGT and 5'-ARRTCHGGYCKSKYCATCCA, where H = A or C or T; K = G or T; M = A or C; R = A or G; S = C or G; V = A or C or G; Y = C or T. The 232-bp reaction product, containing the coding sequence of the TG active-site, was used as sea urchin TG probe.
cDNA library construction and screening cDNA was synthesized with 4 µg of blastula stage poly(A) + RNA. After the addition of EcoRI adapters, the cDNA was inserted into the vector arms of λZAP II (24). This ligated DNA was encapsidated using Gigapack II Gold packaging extract according to the manufacturer's instructions (Stratagene), and was used to infect the E. coli strain XL-1 blue, thus constructing a cDNA library. Approximately 9 x 10 5 recombinant λ phages were screened by plaque hybridization using the sea urchin TG cDNA probe labelled by random priming (Multiprime DNA Labelling System, Amersham). Hybridization was carried out at 60°C for 16 h in Church buffer (1 mM 6 EDTA/0.5 M NaHPO 4 / 7% SDS). Subsequently, the filters were washed 3 times for 10 min each at room temperature in 2x standard saline citrate (SSC)/0.1% SDS and twice for 20 min at 60°C in 2x The cDNA inserts from positive clones were rescued as pBluescript SK(-) by helper phagemediated in vivo excision as described by the manufacturer (Stratagene).

DNA sequencing and analyses
Ten positive cDNA clones were isolated and sequenced. All sequences were carried out with a CEQ 2000XL DNA Analysis System apparatus (Beckman) by Molecular Biology Service of Stazione Zoologica "Anton Dohrn", Naples. The DNA and amino acid sequences were analysed using the GCG computer program (Wisconsin Sequence Analysis Package).
Three of the positive clones contained the entire sequence of the sea urchin TG cDNA.

Plasmid construction
Polymerase chain reaction was used to amplify the PlTG cDNA for subcloning as a GST fusion protein into E. coli expression vector pGEX-2T (Amersham). Full-length PlTG cDNA was amplified using Pfu DNA polymerase and two primers containing an EcoRI restriction site. The amplification product was digested with EcoRI and cloned in-frame with GST into the EcoRI site of pGEX-2T. Restriction digests identified inserts in the correct orientation and DNA sequence analysis verified that no errors had been introduced during the PCR reaction.
A Cys-324 Ala mutant expression vector of PlTG cDNA (GST-PlTG C324A ) was generated using QuickChange TM Site-directed Mutagenesis Kit (Promega) and GST-PlTG as the template, following manufacturer instructions. Mutation was verified by DNA sequence analysis. GST-fusion proteins were eluted with buffer A containing 10 mM reduced glutathione and stored at -80°C. Protein concentration was determined by Protein Assay (Bio-Rad) and bovine serum albumin as the standard. The GST tag was not removed because previous work from many groups has indicated this is unnecessary (26,27). Reaction blanks contained no added calcium.

Expression and purification of GST-fusion proteins
Inhibition assays contained 0-1.5 mM GTP-γ-S in the presence of 0,5 mM CaCl 2 . 9

GTPase and ATPase Activity
GTPase and ATPase activities were determined by the charcoal method as described (29) with some modifications. A 50 µl reaction containing 10 pmol of purified GST or GST-fusion proteins or guinea pig liver TG2, 50 mM Tris-HCl, pH 7.

Cloning of PlTG cDNA
In order to characterize sea urchin TG, we screened a P. lividus blastula stage cDNA library using a partial cDNA fragment of sea urchin transglutaminase as a probe. The probe cDNA was obtained by a PCR on gastrula stage RNA, using two degenerate oligonucleotides encoding the conserved amino acid sequence around the active-site of TGs.
Ten positive clones were isolated out of the 9. amino acid sequence included the Cys324, the other two critical residues for transamidating activity, His383 and Asp406 and the tryptophan, that is a general feature of catalysis in all eukaryotic TGs, present in position 289 (30) (Fig. 1). A putative Ca 2+ binding region (Val481-Arg503), which has been identified in human TG2, was also found ( Fig. 1) (31).
We aligned the PlTG amino acid sequence with other TGs with respect to the middle region around the active-site and Ca 2+ -binding region, which are highly homologous among many TGs. The amino acid sequence of PlTG, in this region, was around 40% identical to those of human TG2, TG3, and TG5 and TGs of crayfish, ascidians, grasshopper, fruit fly, limulus, slime mold, and shrimp, while for starfish TG the identity was around 60% (Fig. 2). PlTG showed the presence of a long amino-terminal region, common to all invertebrate TGs, that is missing in the human enzymes.
In the phylogenetic tree of transglutaminase presented by Lorand and Graham, PlTG was situated closer to other invertebrate TGs than to human and fish TGs. Among human TGs, however, TG4 was placed significantly close to invertebrate TGs (2).
To check whether PlTG showed the presence of a putative GTP-binding region, we performed a sequence alignment of transglutaminases in the region where the key residues for TG2, TG3, and TG5 interaction with GTP have been described following a mutagenesis approach (6) and X-ray studies (8,10). Among the key residues interacting with GTP in TG2, Ser171, Arg476, and Arg478 are conservatively replaced by Thr226, Lys530, and Lys633 respectively in PlTG; but Phe174 and Arg478 are non-conservatively replaced by Asn229 and Val532 (Fig. 2).
PlTG spatial expression. By prism stage the PlTG transcript appeared in the oral ectoderm, with the highest level detectable in the ciliary band ectoderm associated with the growing arm buds. This signal remained constant during pluteus stage (Fig. 3).

E. coli expression of wild type and mutant PlTG
In order to study biochemical activity of PlTG we carried out expression in E.coli of fulllength sea urchin transglutaminase. The coding region of sea urchin TG cDNA was amplified by PCR and cloned in-frame with GST into a plasmid vector, pGEX2T to construct pGEX2T-PlTG

TG activity of GST-PlTG fusion protein
TG activity of GST-PlTG wild-type and mutated fusion proteins was investigated using the incorporation of 14 C-putrescine into N,N'-dimethylcasein as detailed in experimental procedures.
GST-PlTG was a functionally active TG, showing an increase of transglutaminase activity dependent on proteins concentration (data not shown). GST-PlTG activity was comparable to that of commercially available guinea pig liver TG2 (Fig. 5).
As expected, mutation of the active Cys resulted in no TG activity for the GST-PlTG C324A mutant (Fig. 5).
The activity of GST-PlTG was completely dependent on the presence of calcium, increasing with increasing Ca 2+ concentrations, giving 50% activity with 1mM Ca 2+ and reaching the maximum activity with 4mM Ca 2+ under the assay conditions used (data not shown). Purified GST, used as control, had no detectable transglutaminase activity (fig. 5).
The nucleotide GTP has been shown to inhibit type 2 (like guinea pig liver TG), type 3 and type 5 transglutaminase activity. In contrast, the transglutaminase activity of other TGs, like that from Limulus haemocyte, was not affected by GTP even at millimolar concentrations. To study the effect of GTP on GST-PlTG transglutaminase activity we carried out an inhibitory assay using a non-by guest on March 24, 2020 http://www.jbc.org/ Downloaded from 13 hydrolyzable analog of GTP, GTP-γ-S, at different concentrations. Concentrations of GTP-γ-S, in the millimolar range had a detrimental effect on GST-PlTG activity and the degree of inhibition was different depending on calcium concentration. Using 1 mM GTP-γ-S, 60% inhibition was obtained in the presence of 0,5 mM Ca 2+ concentration (Fig. 6). Complete inhibition of GST-PlTG occurred with 1,5 mM GTP-γ-S in the presence of 0,5 mM Ca 2+ concentration. The inhibitory effect caused by GTP-γ-S could be reversed by increasing the concentration of calcium up to 2mM (data not shown).

GTPase and ATPase activity of GST-PlTG fusion proteins.
Not all residues implicated in GTP-binding in TG2 are conserved in PlTG. However, the recent finding of a GTPase activity in TG3, that was believed to lack such an activity on the basis of the sequence similarity, encouraged us to test if the PlTG is able to bind and hydrolyze ATP and GTP.
The GTPase activity was dependent on the protein concentration (data not shown) and on substrate concentration. The effect of increasing GTP concentration on rates of GTP hydrolysis revealed for the GST-PlTG fusion protein an apparent K m of 10 µM in the range of K m of human erythrocyte TG2 (14 µM) (32), using the same assay conditions (Fig. 8). The GTPase activity of the GST-PlTG and GST-PlTG C324A fusion proteins were specifically inhibited in a dose-dependent manner by GTP-γ-S. The IC 50 values for GTP-γ-S of the GTPase activity were comparable for the two tested proteins (10 µM and 13 µM respectively) (Fig. 9).

Discussion
The presence of a putative transglutaminase in sea urchin was first envisaged (33)

Biochemical and molecular features of PlTG
Although the overall identity with the mammalian TG primary sequence is low in the deduced amino acid sequence of the Paracentrotus TG, the middle region of the sequence is significantly conserved. In fact, the catalytic core domain region was highly homologous to the corresponding regions of the human TG2 and of other invertebrate TGs (38). Compared with human TG2, PlTG showed a long amino-terminal region, that was also found in all invertebrate TGs, suggesting that this was a common characteristic of non-mammalian TGs.
In the PlTG deduced primary sequence, the amino acid residues surrounding the active-site Cys ( Fig. 2) are almost all identical to those of invertebrates TGs, except Drosophila melanogaster and 16 crayfish TG. In addition to this catalytic Cys site, His and Asp, which shape a catalytic triad with Cys, are also conserved. Reactions of TGs is similar to those of papain and proceed by a kinetic pathway of acylation and deacylation But in contrast to papain proteases, TGs need a preequilibration with the second substrate and the Trp289 makes an important contribution to stabilizing the transition-state intermediates; in fact this residue is present in all eukaryotic TGs, from slime mold to mammals (30). Furthermore, a putative Ca 2+ -binding region reported in mammalian TG2 was also found (31). This is consistent with the finding that Ca 2+ was required for the enzymatic activity of PlTG. In the absence of Ca 2+ , TG2 adopts an inactive conformation that prevents the reactivity of Cys277. Structural data show that the activation of TG2 involves a displacement of protein domains, with increased reactivity and substrate accessibility to the activesite. Our results suggest that an acyl-transfer reaction, similar to that of mammalian TG2, is executed in the catalytic reaction of PlTG. In fact, the substitution of the active-site Cys with an alanine residue resulted in the total loss of the PlTG activity. The mechanism of PlTG activation was counteracted by the inhibitor GTP, and this inhibition was sensitive to Ca 2+ concentration. The same mechanism was found for TG2, TG3, and TG5.
A comparison of TG3 and TG5 sequences versus TG2, together with the previous structure of the TG2-GDP complex, could not explicitly predict that TG3 and TG5 would bind and hydrolyze GTP, because the amino acid residues interacting with GTP are different and not always conservatively replaced. The finding that all three enzymes exhibit GTPase activity (10,11) suggested that several other TGs enzymes could use other types of residues to mediate essential interactions with guanine nucleotides. All these data drive us to explore whether PlTG could hydrolyze guanine nucleotides.
We observed that PlTG not only binds GTP and ATP, but is also able to hydrolyze both GTP and ATP, like TG2 (27). Instead, TG3 showed only GTPase activity and no measurable ATPase activity, while TG5 showed GTPase activity and was able to bind ATP without any ATPase activity. The effect of GTP concentration on rates of GTP hydrolysis revealed for PlTG an apparent K m of approximately 10 µM, in the range of K m of human erythrocyte TG 2 (14 µM) and of guinea pig liver TG (4 µM) (28,31), which was about 15-fold higher than K m (0,3 µM) of a typical GTPase (39).
The ATPase activity of PlTG was considerably greater than ATPase activity of guinea pig liver TG, showing a higher affinity of PlTG for ATP. The GTPase and ATPase activity of PlTG does not require the active-site cysteine, in fact the mutate form of PlTG retained GTP and ATP-binding properties that were even increased probably due to the presence of proteolytic fragments. This was demonstrated for human TG2, where the removal of C-terminal fragment increases the NTP hydrolysis activity, probably due to a different folding of the ATPase GTPase catalytic domain (27).
These results confirmed that GTPase activity of transglutaminases is not tightly related to the primary sequence but should be evaluated experimentally.
In its GTP-bound form, TG2 function as a signal transduction molecule by acting as a classic Gprotein. As G-protein, TG2 is involved in the transmembrane transmission of the α1-adrenergic and thromboxane receptors to their effector enzyme phospholipase C-delta1 (PLC-δ1) (7). Coupling of TG2 to PLC-δ1 activates the hydrolysis of membrane-bound inositol phospholipids leading to generation of the second messangers 1,4,5-triphosphate (IP3) and diacylglycerol and subsequent intracellular Ca 2+ mobilization and protein kinase C activation (40). Moreover, GTP acts as noncompetitive inhibitor for transamidating activity. Therefore, under physiological conditions TG activities could be regulated by local concentrations of calcium and nucleotides (GTP and ATP), as previously proposed for TG2 (5). have also been shown to bind GTP/ATP and to undergo a GTPase cycle. In this paper we demonstrate, for the first time, a GTP-hydrolyzing activity in an invertebrate TG. This finding has important implication for the evolution of TGs, and in particular for the evolution of their role in signaling. In fact, as previously reported PlTG together with slime molde TG (2) have evolved from an early duplication event in a lineage different from the TG2, TG3 and TG5 lineage. Hence, the demonstration of GTP hydrolysis in PlTG adds weight to emerging evidence that the bifunctionality of TGs is a general function rather than a derived property acquired by only one lineage after the early gene-duplication event. This feature has been successively lost in some descendents such as fXIIIa.
Whether the GTP-hydrolyzing activity of PlTG is related to certain cellular signalling in sea urchin remains to be determined. However the presence of the PlTG mRNA in regions where important morphogenetic rearrangements occurred during embryogenesis, could suggest additional physiological roles, besides the cross-linking of proteins, for PlTG.