To investigate post-dormant regulation of trehalose metabolism in the brine shrimp, we cloned and characterized two trehalase cDNAs from embryos of Artemia franciscana using a PCR probe corresponding to a highly conserved region among trehalases. The cDNAs consisted of 2496 and 2485 nucleotides, and had almost the same open reading frame encoding 703 amino acids which showed 46.6–42.6% similarities to trehalases of animals. The calculated molecular mass of the trehalase was 79,995 Da. The deduced sequence had a cleavable signal peptide, a cell adhesion motif, four potential N-glycosylation sites, trehalase signatures and a unique, long carboxyl terminal polypeptide containing a predicted transmembrane region and a potential cAMP-dependent phosphorylation site. Phylogenetic analyses showed a large divergence among trehalases of arthropods. Northern blots revealed the presence of two mRNAs. One of them, a 2.6 kb mRNA, was abundant in the dormant cysts and prenauplii. The other 5.0 kb mRNA was newly synthesized during post-dormant development. Possible mechanisms of trehalase regulation are presented on the basis of the results shown by the Northern blots and developmental changes of trehalase activity.
Various organisms accumulate α, α -trehalose, a non-reducing disaccharide consisting of two glucoses, against physiological and environmental changes. The accumulated trehalose has a role as an anti-stress molecule. For instance, in yeast, such anti-stress effects of trehalose against dehydration, heat shock, freezing and ethanol have been reported (reviewed by Nwaka and Holzer, 1998). These anti-stress effects of trehalose are thought to be due to its protective function on the cell membrane or proteins (Crowe et al., 1987, 1988).
Embryos of Artemia entering dormancy also accumulate α, α -trehalose. This accumulation is not observed in nondormant embryos (Clegg, 1962, 1965). The amount of trehalose in the encysted dormant gastrula represents 15% of its dry weight (Clegg, 1962, 1965). Under appropriate circumstances, the encysted embryos resume their subsequent development and the trehalose is hydrolyzed by a trehalase (EC 188.8.131.52).
In our previous study (Nambu et al., 1997a), a soluble, alkaline trehalase with a molecular mass of 70 kDa on SDS-polyacrylamide gel electrophoresis was purified from developing embryos of Artemia. The Artemia trehalase showed a significant increase in its activity at the prenauplius stage and the activity decreased after hatching (Nambu et al., 1997a). Hydrolysis of trehalose provides glucose to produce glycogen and glycerol, which are respiratory substrates in the Artemia cysts (Clegg, 1964; Ewing and Clegg, 1969). It is also suggested that accumulated glycerol increases the osmotic pressure in the cysts, resulting in a rupture of the shell followed by the emergence of prenauplii surrounded by hatching membrane (Clegg, 1964). Therefore, the temporal regulation of the trehalase activity is significant in post-dormant development. To investigate the mechanisms of trehalase regulation, we cloned a 228 bp cDNA fragment of the Artemia trehalase using a polymerase chain reaction (PCR) (Nambu et al., 1997b). In the present report, we describe the cloning and characterization of Artemia trehalase cDNAs encoding a full open reading frame (ORF) and show the results of Northern blots and phylogenetic analyses.
MATERIALS AND METHODS
Dried cysts of Artemia franciscana from the Great Salt Lake in Utah, USA were obtained from Japan Pet Drugs Co. (Tokyo and Los Angeles, CA, USA). The cysts were hydrated and cultured as described by Nambu et al. (1997a).
Total RNA was extracted from 0.5 g of the wet weight of the embryos and nauplii according to the method described by Maniatis et al. (1982) with some modifications. Each specimen was homogenized in 2.5 ml of a guanidine thiocyanate solution. The homogenate was centrifuged at 4,000 rpm for 5 min and total RNA was extracted from the supernatant. Poly (A)+ RNA was purified using Oligotex dT30 <Super> (Roche Japan, Tokyo, Japan).
cDNA cloning and nucleotide sequencing
Double stranded cDNA was synthesized from purified poly (A)+ RNA extracted from 12 hr cultured prenauplii of A. franciscana using oligo(dT)12–18 primer. The double stranded cDNA was ligated with EcoRI-NotI adaptor and a cDNA library was constructed in λ gt10 vector (Stratagene) using gigapack®III Gold (Stratagene). The 228 bp cDNA fragment of the Artemia trehalase was obtained by the degenerated PCR and digoxigenin (DIG)-labelled as described by Nambu et al. (1997b). It was used to screen 2.0 × 105 clones. Phage plaques were lifted up and immobilized on nylon membranes (Boehringer Mannheim, Mannheim, Germany). Hybridization and detection was performed as described by Engler-Blum et al. (1993). Positive clones were isolated and two longer inserts termed TreE2 and TreE3 were fully sequenced using Amersham thermo sequenase fluorescent labelled primer cycle sequence kit (Amersham, Buckinghamshire, UK) and an ABI 373 sequencer (Perkin Elmer, Norwalk, CT, USA).
Northern blot analysis
The full length of TreE3 was DIG-labelled according to the random priming method using a DIG DNA Labelling Kit (Boehringer Mannheim).
Total RNA was prepared as described in the section of RNA extraction, and 20 μg of each total RNA was electrophoresed in a 1.0% agarose gel containing formaldehyde (Sambrook et al., 1989) and blotted on to Hybond N+ membrane (Amersham). Hybridization was performed at 68°C as described by Engler-Blum et al. (1993) followed by washing at 65°C.
Isolation of trehalase cDNAs from A. franciscana
Screening of the cDNA library derived from the 12 hr cultured embryos of A. franciscana with the PCR product resulted in isolation of 12 positive clones. Two longer inserts, TreE2 and TreE3, were used for full-length nucleotide sequencing and further investigation.
Comparison of TreE2 with TreE3
TreE2 consisted of 2496 nucleotides while TreE3 contained 2485 nucleotides. These had a common region consisting of 2476 nucleotides (Fig. 1). Identity between the two cDNAs was 99.7%. Seven replacements were found at the nucleotide level (Fig. 1). Both cDNAs contained the same ORF followed by a termination codon and an “AGTAAA” sequence which was similar to the polyadenylation signal, “AATAAA”. An oligo (A) sequence was found 14 nucleotides downstream from the “AGTAAA” sequence in TreE2 (Fig. 1, 2). The proteins deduced from TreE2 and TreE3 comprised 703 amino acids and showed the same molecular mass of 79,995 Da and the same isoelectric point of 5.22. Proteins encoded by these cDNAs were identical except for replacements of two amino acids (Fig. 2, TreE2: Thr246 →TreE3: Ser246 and TreE2: Val504→TreE3: Leu504).
Characterization of Artemia trehalase cDNA
In a computer-assisted search of databases for SWISS-PROT and PIR, Artemia trehalase showed no significant similarity to other proteins than trehalases. The protein was highly similar to animal trehalases: 46.6% to human (Ishihara et al., 1997), 46.1% to rabbit (Ruf et al., 1990), 43.4% to mealworm beetle (Takiguchi et al., 1992), and 42.6% to silkworm (Su et al., 1993, 1994). The identities of the Artemia trehalase to trehalases of E. coli (Gutierrez et al., 1989; Horlacher et al., 1996) and neutral trehalases of fungi (Kopp et al., 1993, 1994; Feldmann et al., 1994; Wolfe and Lohan, 1994; Nwaka et al., 1995; Amaral et al., 1997; Eck et al., 1997) were relatively low (24.6–32.0%). No significant similarity was observed between the Artemia trehalase and acid trehalases of fungi, such as ATH1 gene product of S. cerevisiae (Destruelle et al., 1995) and treA gene product of Aspergillus nidulans (d'Enfert and Fontaine, 1997).
The weight matrix method described by von Heijne (1986) showed the presence of a cleavable signal sequence consisting of 19 amino acids at the amino(N)-terminal end of the Artemia trehalase (Fig. 2, 3). A cell adhesion motif, Arg-Gly-Asp (reviewed by Ruoslahti and Pierschbacher, 1986, and D'Souza et al., 1991) was found next to the N-terminal cleavable site as shown in Fig. 2. The protein had four potential N-glycosylation sites (Fig. 2). Alignment of animal trehalases showed that the location of the third potential N-glycosylation site was conserved among arthropods (Fig. 3). Trehalase signatures submitted to the PROSITE database (Henrissat and Bairoch 1993) were found on the deduced amino acid sequences (Fig. 2, 3). A unique, long polypeptide was found at the carboxyl terminus (Fig. 2, 3, 4). The PSORT program (Nakai and Kanehisa, 1992) strongly suggested that the carboxyl(C)-terminal polypeptide contained a transmembrane region (Gly620-Glu645, Fig. 2, 4). The predicted transmembrane region was highly hydrophobic as was the N-terminal signal sequence (Fig. 4). A potential cAMP-dependent phosphorylation site was included in a cytoplasmic region of the C-terminal polypeptide (Fig. 2).
Phylogenetic trees were drawn by the neighbor-joining method (Saitou and Nei, 1987) (Fig. 5). The results revealed that all the trehalases could be classified into three groups as follows: animal trehalases including the Artemia trehalase, trehalases of E. coli, and neutral trehalases of fungi (Fig. 5A). Genetic distances between the Artemia trehalase and mammalian trehalases were shorter than those between the Artemia trehalase and trehalases of insects (Fig. 5B).
Developmental expression of Artemia trehalase
Since the nucleotide sequences of TreE2 and TreE3 were almost identical to each other, the labelled TreE3 was used as a probe to detect both types of mRNA.
A 2.6-kb mRNA was detected in all the samples of A. franciscana (Fig. 6). The 2.6-kb transcript was abundant in the dormant (0 hr) cysts and embryos at 3 hr and 6 hr of the culture (Fig. 6). The transcript then decreased and was weakly detected in the 30 hr sample. A 5.0-kb transcript was observed after 3 hr, which increased by 6–12 hr and decreased thereafter. The amount of the 5.0-kb transcript was lower than that of the 2.6-kb transcript in all the samples.
In the present report, we described cloning and characterization of the trehalase cDNAs from crustacean Artemia. Most parts these two cDNAs, termed TreE2 and TreE3, were the same and both contained the same ORF of 2109 nucleotides. Seven replacements were found at the nucleotide level between TreE2 and TreE3, resulting in two replacements on coded 703 amino acids. The rate of change at the nucleotide level between TreE2 and TreE3 was 0.3%, much lower than those among alleles of Na/K-ATPase α1 subunit of A. franciscana reported by García-Sáez et al. (1997). Therefore, TreE2 and TreE3 appear to be derived from point-mutated alleles of one trehalase gene rather than from two different trehalase genes. Sequencing analyses of the PCR products showed that the mRNAs were in the ratio of 1:1 (data not shown). As the dried cysts which we used in the present report were from a wild population, rather than from an established strain, variants of trehalase gene might be observed in individuals or in sub-populations.
The presence of potential N-glycosylation sites on the predicted peptide is consistent with our previous conclusion that the Artemia trehalase is a glycoprotein (Nambu et al., 1997a).
The most significant feature of the Artemia trehalase is the unique, long C-terminal polypeptide. The region consisted of 107 amino acids (Ile597-Glu703) and had no significant similarity to any other proteins. The predicted transmembrane region (Gly620-Glu645) was present on the C-terminal polypeptide (Fig. 2, 4) and about 60 amino acids from the C-terminal region were possibly located in the cytoplasm. Membranebound trehalases have been reported in silkworm (Azuma and Yamashita, 1985; Takesue et al., 1989) and mammals (Takesue et al., 1986; Ruf et al., 1990; Sasai-Takedatsu et al., 1996; Oesterreicher et al., 1998). The presence of the cell adhesion motif also suggests the association of Artemia trehalase to the cell membrane.
The predicted molecular mass of the Artemia trehalase without the N-terminal signal sequence was 77,936 Da and substantially larger than that of the purified soluble protein, 70 kDa (Nambu et al., 1997a). The molecular mass of the soluble trehalase decreased to 66 kDa after endoglycosidase H treatment in our previous study (Nambu et al., 1997a). The solubility of the Artemia trehalase is incompatible with the presence of the predicted transmembrane region. As similar sequences to the C-terminal polypeptide of the Artemia trehalase were not found among other trehalases (Fig. 3), it is probable that the C-terminal polypeptide consisting of 107 amino acids is not necessary for trehalase activity and might be removed. The molecular mass of the deduced Artemia trehalase without the N-terminal signal sequence and the C-terminal polypeptide was calculated to be 66,272 Da, showing a good agreement with the molecular mass of the purified Artemia trehalase after the endoglycosidase H treatment.
A possible cAMP-dependent phosphorylation site was found on the predicted C-terminal cytoplasmic region of the Artemia trehalase, although we do not have any evidences that the Artemia trehalase is regulated by cAMP-dependent phosphorylation. The cAMP-dependent phosphorylation site has not been found in other animal trehalases. In yeasts, cAMP-dependent phosphorylation sites in the N-terminal region and cAMP-dependent regulation of yeast trehalase have been reported (Kopp et al., 1993, 1994; Amaral et al., 1997; Eck et al., 1997). Although the site may be removed by the post-translational processing discussed above, it is possible that the potential cAMP-dependent phosphorylation site plays a role in an immature Artemia trehalase.
Phylogenetic trees of trehalase are shown in Fig. 5. On the phylogenetic tree of animal trehalase, the Artemia trehalase was closely related to the mammalian trehalases (Fig 5B). The distributions and the physiological meanings of the arthropod trehalases were reported to be different from each other (Azuma et al., 1985; Takesue et al., 1989; Takiguchi et al., 1992; Yaginuma et al., 1996; Nambu et al., 1997a). Adaptation of each trehalases to different roles and physiological environments might result in their variety in amino acid sequences.
The results of Northern blot analysis confirmed the presence of the two transcripts previously described (Fig. 6). The 2.2 kb mRNA of the Artemia trehalase found in the previous report (Nambu et al., 1997b) was not detectable when we changed a batch of the cysts.
Three working hypotheses on the two transcripts might be proposed as follows. First, the transcripts of Artemia trehalase might originated from alternative splicing of one gene.
The second possibility is that the 5.0 kb transcript is a precursor and processed into the 2.6 kb transcript. The third possibility is that the two transcripts are derived from two highly homologous genes.
The abundance of the 2.6 kb transcript in the dormant cysts of A. franciscana suggests its importance in early period of the post-dormant development. It may be called as a cryptic RNA. It was observed that 18S rRNA gradually increased during development (Fig. 6). This means that the ratio of rRNA to total RNA is relatively low in the dormant cysts. This observation may be partly due to accumulations of cryptic RNA, such as trehalase mRNA, in the dormant cysts.
The activity of the trehalase was low in the hydrated cysts and began to increase after 6 hr in culture (Nambu et al., 1997a) in spite of the abundant presence of the 2.6 kb transcript in the dormant cysts. The discrepancy between the amount of the mRNA and the trehalase activity suggests a probable presence of post-transcriptional regulation of the enzyme. Similar inconsistency between the amount of mRNA and protein products have been reported in the early period of post-dormant development of Artemia (Fisher et al., 1986; Díaz-Guerra et al., 1989).
The Artemia embryos have mechanisms inactivating protein synthesis during their dormancy. The presence of protein synthesis inhibitor (Warner et al., 1977), a lack of template activity of some mRNAs in the dormant cysts by restricted location of mRNAs into mRNP particles (Grosfeld and Littarue, 1975), and low levels of initiation factors (Sierra et al., 1974) have been reported. Moens and Kondo (1976) reported that the specific template activity of the polysomes in the cysts was low and increased linearly up to 13 hr of incubation. The discrepancy between the amount of mRNA and the trehalase activity in the embryos of A. franciscana might be explained by such inhibitions.
The appearance of the 5.0 kb transcript in 3 hr embryos of A. franciscana is a sign of initiation of post-dormant transcription of the Artemia trehalase. If the 5.0 kb transcript is the precursor of the 2.6 kb transcript, the appearance of the 5.0 kb transcript indicates a supplement of the 2.6 kb transcript. The amount of the 5.0 kb transcript increased in accordance with the increase in the trehalase activity (Nambu et al., 1997a). Therefore, it is still possible that transcriptional regulation of the 5.0 kb mRNA directly controls the trehalase activity.
The structural and functional differences of the two transcripts and the mechanisms of the regulation of the trehalase activity are open for future investigation. Further analysis of transcriptional or translational regulations of trehalase would help us to understand the trehalose metabolism in resuming development of Artemia.
We are deeply indebted to Drs. Shimada, Akasaka and Mitsunaga-Nakatsubo, Hiroshima University, for their helpful discussions and supports. Thanks are also offered to Drs. Satoh and Makabe and their co-workers, Kyoto University, for their advice on phylogenetic analysis of trehalases.