Glycobiology Advance Access originally published online on January 26, 2005
Glycobiology 2005 15(6):615-624; doi:10.1093/glycob/cwi045
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Glycobiology vol. 15 no. 6 © Oxford University Press 2005; all rights reserved.
Intact 
-1,2-endomannosidase is a typical type II membrane protein
2 GlycoFi, Inc., 21 Lafayette Street, Lebanon, NH 03766; 3 Thayer School of Engineering, Dartmouth College, Hanover, NH 03755; and 4 Department of Pharmacology and Toxicology, Dartmouth Medical School, Hanover, NH 03755
1 To whom correspondence should be addressed; e-mail: tillman.gerngross{at}dartmouth.edu
Received on September 20, 2004; revised on January 10, 2005; accepted on January 19, 2005
 |
Abstract |
|---|
 Top Abstract IntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
Rat endomannosidase is a glycosidic enzyme that catalyzes the cleavage of di-, tri-, or tetrasaccharides (Glc1–3Man), from N-glycosylation intermediates with terminal glucose residues. To date it is the only characterized member of this class of endomannosidic enzymes. Although this protein has been demonstrated to localize to the Golgi lumenal membrane, the mechanism by which this occurs has not yet been determined. Using the rat endomannosidase sequence, we identified three homologs, one each in the human, mouse, and rat genomes. Alignment of the four encoded protein sequences demonstrated that the newly identified sequences are highly conserved but differed significantly at the N-terminus from the previously reported protein. In this study we have cloned two novel endomannosidase sequences from rat and human cDNA libraries, but were unable to amplify the open reading frame of the previously reported rat sequence. Analysis of the rat genome confirmed that the 59- and 39-termini of the previously reported sequence were in fact located on different chromosomes. This, in combination with our inability to amplify the previously reported sequence, indicated that the N-terminus of the rat endomannosidase sequence previously published was likely in error (a cloning artifact), and that the sequences reported in the current study encode the intact proteins. Furthermore, unlike the previous sequence, the three ORFs identified in this study encode proteins containing a single N-terminal transmembrane domain. Here we demonstrate that this region is responsible for Golgi localization and in doing so confirm that endomannosidase is a type II membrane protein, like the majority of other secretory pathway glycosylation enzymes.
Key words: glycosidase / glycosylation / Pichia pastoris / recombinant expression / transmembrane domain
|
Introduction |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
N-linked glycosylation plays a major role in the processing of many cellular and secreted proteins. In eukaryotes the preassembled oligosaccharide glucose3mannose9N-acetylglucosamine2 (Glc3Man9GlcNAc2) is transferred from dolichyl pyrophosphate onto the acceptor site of the protein by an oligosaccharyltransferase complex in the endoplasmic reticulum (ER) (Dempski and Imperiali, 2002
). Subsequently, the terminal 
-1,2-glucose is removed by glucosidase I followed by cleavage of the remaining two -1,3-glucose residues by glucosidase II (Herscovics, 1999). The resulting high-mannose glycan (Man9GlcNAc2) is then processed by the ER-residing -1,2-mannosidase to yield Man8GlcNAc2 and transported to the Golgi, where further modifications are performed (Herscovics, 1999).
Incorrect processing of the glycan structure in the ER can block subsequent modifications, which is the molecular basis of various diseases in humans. For example the absence of glucosidase I results in a congenital disorder of glycosylation type IIb, which is extremely rare, with only one reported human case, leading to early death (Marquardt and Denecke, 2003). Isolation of a Chinese hamster ovary (CHO) cell line (Lec23), deficient in glucosidase I, demonstrated that the predominant glycoform produced by this cell line is Glc3Man9GlcNAc2 (Ray et al., 1991). A mouse lymphoma cell line, PHAR2.7, lacking glucosidase II activity, produced mainly the glycoforms Glc2Man9GlcNAc2 and Glc2Man8GlcNAc2 (Reitman et al., 1982). Further analysis of this latter cell line demonstrated that despite the absence of glucosidase II, high-mannose structures lacking glucose were present, indicating the existence of an alternative pathway for the removal of glucose from glucosylated N-glycan structures (Moore and Spiro, 1992).
The enzyme responsible for this glucosidase-independent reaction was found to be an endomannosidase (Endo; E.C. 3.2.1.130
[EC]
), which catalyzes the release of mono-, di-, and tri-glucosyl-mannose oligosaccharides by cleaving the -1,2-mannosidic bond that links them to high-mannose N-glycans (Bause and Burbach, 1996; Hiraizumi et al., 1993; Lubas and Spiro, 1988). The only Endo gene cloned to date was obtained from a rat liver cDNA library (rEndo), encodes a protein of 451 amino acids with a molecular mass of 52 kDa (Spiro et al., 1997). To distinguish this sequence from those identified in this report it will subsequently be referred to as rEndoSpiro. This enzyme has a neutral pH optimum and does not appear to have any specific cation requirements (Bause and Burbach, 1996). Unlike the glucosidase enzymes, which are localized in the ER, rEndoSpiro is primarily localized in the Golgi (Zuber et al., 2000). This suggests that the enzyme may serve as a backup system by processing glucosylated glycoforms that bypass glucosidase I and II, and have leaked into the Golgi. Furthermore, in contrast to the majority of glycosylation enzymes, which are type II membrane proteins, the cloned rEndoSpiro does not possess a transmembrane domain and is postulated to be membrane associated through myristoylation of the penultimate glycine residue (Spiro et al., 1997).
In the current report we have cloned two novel open reading frames (ORFs) encoding human liver and rat liver Endos from cDNA libraries, and identified a homologous ORF in the mouse (hEndo, rEndo, and mEndo, respectively). Sequence analysis of these three ORFs identified a conserved N-terminal hydrophobic region that we have demonstrated to be responsible for Golgi localization. This indicates that the reported proteins are typical type II membrane proteins and distinguishes them from the previously reported protein (Spiro et al., 1997). Furthermore, evidence presented in the current study demonstrates that the rat Endo sequence cloned by Spiro et al. (1997) is incomplete and possesses a 5'-terminus that is likely to be an artifact of cloning.
|
Results |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
Cloning the intact hEndo and rEndo sequences
By performing a protein BLAST search using the rEndoSpiro sequence (GenBank gi:2642187) we identified a human ORF (GenBank gi:20547442) of 290 amino acids, which showed 87% identity and 94% similarity to amino acids 162 to 451 of the rat ORF. This former ORF was annotated in the human genome as hypothetical protein FLJ12838 though the tissue source was not disclosed. When the DNA 5'-terminus of this human sequence was analyzed using translated BLAST, another human ORF (GenBank gi:18031878) was identified that possessed 95% identity over a 22-amino-acid stretch. This ORF was annotated in the human genome as the protein mandaselin short form from the liver, though the function of this protein was not elucidated. Subsequent reading-frame analysis of this second sequence indicated that 172 amino acids were in-frame upstream of the homologous region. Combining both of these overlapping reading frames produced a putative sequence with an ORF of 462 amino acids (Figure 1A) and a predicted molecular mass of 54 kDa. To determine if both of these overlapping reading frames were indeed a single ORF primers were designed to the 5' and 3' termini of the combined ORF. Subsequent polymerase chain reaction (PCR) amplification produced a fragment of 1389 base pairs encoding the single predicted ORF. Examination of this ORF by BLAST analysis against the human genome (NCBI) maps it to chromosome 6 (6q16). Within this region the ORF is separated into four exons, containing bases 1–544 (exon 1), 545–654 (exon 2), 655–731 (exon 3), and 732–1389 (exon 4), separated by 9.8 kb, 8.0 kb, and 0.8 kb, respectively. Analysis of the protein sequence encoded by this ORF indicated a hydrophobic region from amino acids 10 to 26 (Figure 1B), suggesting the presence of a potential type II transmembrane domain, like the majority of mammalian glycosylation enzymes. This is in contrast to the reported sequence of rEndoSpiro, which does not appear to contain a transmembrane domain but possesses a glycine residue at position 2 in the protein that is speculated to be myristoylated, and thus the mechanism of membrane localization (Spiro et al., 1997
).
|
In addition to the hEndo, we identified a third member of this family of mannosidases by BLAST analysis. This third member was isolated as a cDNA clone (GenBank AK030141 [GenBank] ) from mouse testis as part of the Mouse Genome Encyclopedia Project and annotated based on its homology to the rat protein. This sequence encodes a protein of 462 amino acids with 82% and 85% identity to the rEndoSpiro and hEndo enzymes, respectively. Alignment of the hEndo, mEndo, and rEndoSpiro protein sequences indicated that the C-termini of these proteins are highly conserved but that the N-termini of the hEndo and mEndo vary from that of rEndoSpiro (Figure 2). Like the human sequence, the mouse sequence has a hydrophobic region at the N-terminus suggesting that it also has a potential type II transmembrane domain. The high degree of conservation between all three sequences, from the motif DFQ(K/R)SDRI to the C-terminus, suggests that this region encodes the catalytic domain in each case. Furthermore, owing to the high degree of similarity between the N-termini of the human and mouse proteins, we speculated that a similar form may be present in rat. Using the primer synthesized to amplify the 5'-terminus of the human gi:18031877 ORF (primer hE1) in combination with an rEndo 3'-terminal primer (primer rE2), we were able to amplify a single fragment from rat liver cDNA. This corresponded in size to that of the hEndo ORF. Subsequently, the hEndo ORF was used to perform a BLAST analysis against the rat genome and a potential rEndo ORF was identified that also possessed the typical type II membrane protein architecture. The sequence encoding the ORF mapped to chromosome 5 (5q21) of the rat genome and has similar architecture to the human sequence, being separated into four exons, containing bases 1–544 (exon 1), 545–654 (exon 2), 655–731 (exon 3), and 732–1389 (exon 4), separated by 3.7 kb, 7.0 kb, and 0.8 kb, respectively.
|
Interestingly, analysis of the mouse genome demonstrated that the putative mEndo sequence was also divided into four exons and that splicing sites were located after bases 544, 654, and 731, identical to both hEndo and rEndo sequences. To clone the newly identified rEndo sequence a 5' primer was designed to anneal 40 bases upstream of the putative ATG (primer rE1). Using this primer and an rEndo 3'-terminal primer (primer rE2) we were able to amplify a specific product from rat liver cDNA encoding rEndo with an ORF of 462 amino acids that is 85% identical to hEndo and 92% identical to the putative mouse ORF, mEndo.
Interestingly, comparing the newly isolated rEndo sequence to that reported by Spiro et al. (1997) demonstrated that bases 162 to 1389 of the former sequence were identical to bases 129 to 1356 of the original sequence, with the exception of three bases that encoded for silent mutations within the amino acid sequence (data not shown). Conversely, bases 1 to 128 of rEndoSpiro demonstrated no significant homology to the 5' of the sequence isolated in this study. Furthermore, when the rEndoSpiro sequence was mapped to the rat genome, bases 1 to 128 localized to chromosome 9 (9q32), whereas bases 129 to 1356 mapped to chromosome 5 (5q21), suggesting that this Endo sequence was an artifact of cloning. To test this theory we designed primers to PCR amplify the complete ORF sequence reported by Spiro et al. (1997). However, despite numerous attempts, using several primer sets and rat liver cDNA from different commercial sources, we were not able to obtain the reported PCR product (data not shown). Subsequent genome analysis has confirmed that only one copy of Endo is present in each the human, mouse, and rat genomes, and these sequences are confined to a single chromosome in each—thus eliminating the possibility of the presence of a second copy of Endo corresponding to rEndoSpiro in the rat genome. Together the data presented demonstrate that (1) the reported rEndoSpiro sequence is divided onto two chromosomes; (2) the reported rEndoSpiro cannot be PCR amplified (unlike the sequences reported in the current study); and (3) the occurrence of single highly homologous copy of Endo in each of the rat, human, and mouse genomes. In summary, we conclude that the original Endo sequence reported by Spiro et al. (1997) was an artifact.
Functional expression of human and rat Endo in yeast
To confirm that the identified human ORF encodes an Endo, we expressed a recombinant form of the enzyme in the yeast Pichia pastoris. Based on the sequence conservation between the motif DFQ(K/R)SDRI to the C-termini of the four sequences in Figure 2, we postulated that this comprised the catalytic domain. Because the protein sequence of the catalytic domain of rEndo is identical to the previously characterized rEndoSpiro catalytic domain, the former can be used as a control to demonstrate Endo activity. Thus we generated yeast expression vectors encoding rEndo and hEndo lacking the N-terminal 59-amino-acid residues. The expression vector pPICZA was used to create an N-terminal fusion of the P. pastoris alpha-mating factor secretion signal to the Endo catalytic domains. These fusions also resulted in the generation of C-terminal 6x His tags to each catalytic domain, as represented in Figure 3B. To assay Endo activity in the medium, the 2-aminobenzamide (2-AB)-labeled GlcMan5GlcNAc2-2AB glycan (Figure 4A, structure i) was incubated with supernatant from a P. pastoris strain (YSH13) expressing rEndo60–462 6xHis.
|
|
As illustrated in Figure 4A, this substrate used does not possess a terminal -1,2-mannose residue and is thus resistant to -1,2-mannosidase. Following incubation of this substrate with supernatant from YSH13, high-performance liquid chromatography analysis demonstrated a reduction in the retention time, corresponding to the removal of two hexose sugars (Figure 4B). These data are consistent with the removal of a glucose--1,3-mannose disaccharide and the formation of Man4GlcNAc2-2AB (Figure 4A, structure ii). Subsequent incubation of Man4GlcNAc2-2AB with -1,2-mannosidase resulted in the removal of one additional terminal -1,2-mannose residue (data not shown), producing Man3GlcNAc2-2AB glycan (Figure 4A, structure iii). Incubation of the supernatant from P. pastoris strain YSH16, secreting hEndo60–462 6xHis, with GlcMan5GlcNAc2-2AB glycan revealed the same activity as the control rEndo sample (Figure 4B). Time-course analysis of this reaction demonstrated that the initial removal of the two hexose residues was a single event (data not shown). Thus the removal of the glucose--1,3-mannose disaccharide confirms that the isolated ORF encodes a functional human Endo. To further characterize the enzyme, we performed assays at different pHs and temperatures, which revealed a pH optimum of 6.2 (Figure 3C), and 37°C being the most efficient temperature tested (data not shown).
Western blot analysis of the supernatant from P. pastoris strains secreting recombinant rEndo60–462 6xHis and hEndo60–462 6xHis (Figure 3B) with an anti-6xHis antibody could only detect the rat enzyme (data not shown). However, because the activity of both enzymes was detectable in the supernatant, we postulated that proteolytic cleavage of the human isoform had removed the 6xHis tag, resulting in our inability to detect this isoform on the western blot. To investigate this theory, double FLAG tags were engineered into both the rat and human constructs between the N-terminal Kex2 proteolytic cleavage site and the Endo domain (Figure 3B). These constructs showed identical activity to the untagged enzymes but now allowed detection of the human and the rat enzymes prior to Ni-resin purification (Figure 4D). This confirmed that hEndo was being secreted but that the C-terminus was not intact based on the absence of recognition by the anti-6xHis antibody. Furthermore, western blot analyses also indicated that the secretion level of recombinant rEndo was several-fold higher than that of the human protein. The supernatant of the P. pastoris strain (YSH89) expressing rEndo60–462 FLAG/His appeared to contain two discrete forms of the protein, a 59-kDa and a 55-kDa polypeptide (Figure 4D), whereas the hEndo60–462 FLAG/His sample contained only one band of 55 kDa. Interestingly, when these samples were passed over a Ni-affinity resin, only the higher-molecular-weight rat protein (59 kDa) (Figure 4D) was retained and recovered by elution with imidazole.
Together these data indicate that although the 55-kDa mass corresponds to the theoretical mass of the expressed ORF, only the upper species of 59 kDa in the rat sample appears to be the complete protein, and both rat and human isoforms are being proteolytically processed at the C-terminus resulting in the removal of the 6xHis tags. In the case of the human isoform, this cleavage is complete and explains the lack of signal in the postpurification sample. In contrast, a small proportion of the expressed rEndo remained intact allowing its detection postpurification and on the initial anti-6xHis western blot. Furthermore, the higher expression levels of rEndo indicate that there are several internal proteolytic cleavage sites, represented by the laddering effect observed in the prepurification sample (Figure 4D).
Tissue distribution of hEndo
hEndo mRNA was detected in all tissues examined, albeit at varying concentrations (Figure 5). In the majority of tissues two mRNA species were observed corresponding to 3.1 kb and 5.1 kb. The size of the latter is substantially larger than that expected from the size of the protein. However, it has been reported previously that rEndoSpiro message is also about 4.9 kb (Spiro et al., 1997). Though we concluded that the cloned rEndoSpiro sequence was incomplete, missing the first 128 base pairs of the ORF, the probe used to detect message in that study was 3' to the deletion. As a consequence the probe would have been hybridizing to the full-length message of the rEndo cloned in the current study. Therefore, the prominent message for both rEndo and hEndo is of similar size, migrating at 
5 kb. Transcript for hEndo is most intense in the liver, from which the ORF was isolated. The apparent order of expression from highest to lowest is liver, kidney, muscle/pancreas, heart, placenta/lung, and brain. The faint expression of Endo in brain tissue (Figure 5) can be intensified with prolonged exposure.
|
Immunolocalization of Endo
To investigate the subcellular localization of hEndo, we constructed and expressed a series of green fluorescent protein (GFP) fusions, see Figure 3C. Expression of the full-length hEndo fused to GFP (hEndo1–462 GFP) demonstrates colocalization with the wheat germ agglutinin (WGA) Golgi marker (Figure 6L), agreeing with previous data that the rEndoSpiro localizes to this compartment (Zuber et al., 2000). However, to date, the mechanism by which Endo localizes to the Golgi has not been identified. To investigate the roles of the putative transmembrane and the catalytic domains in localization, individual GFP fusion constructs of these regions were expressed; see Figure 3C for illustrated fusions. As demonstrated in Figure 6F, fusion of GFP to the N-terminal 59 residues of hEndo (hEndo1–59 GFP) results in Golgi colocalization with WGA in a similar fashion to the full-length protein. In contrast the C-terminal catalytic domain (residues 60–462) appears to play a minimal role in localization because distribution of this fusion is comparable to cells transformed with the control vector (Figure 6I and O). Though the GFP domain expressed in study has no apparent secretion signal, a small proportion of the GFP expressed enters the secretory pathway and as a result passes through the Golgi. Therefore, although there is limited colocalization with WGA, in both hEndo60–462 GFP and control samples, the majority of the GFP signal does not colocalize with WGA. Thus, in hEndo, it is apparent that the N-terminus is responsible for cellular localization of the protein to the Golgi, and like the majority of other glycosylation enzymes, this data confirms that hEndo is a typical type II membrane protein. Likewise, the N-terminus of the rEndo isolated in this study is capable of facilitating Golgi localization confirming that this protein is also a type II membrane protein (Figure 6C).
|
|
Discussion |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
In the present study we isolated cDNA clones encoding hEndo and rEndo while identifying a mouse homolog, mEndo. Comparison of these three ORFs with the previously reported rEndoSpiro sequence (Spiro et al., 1997
) demonstrated high homology from the motif DFQ(K/R)SDRI to the C-termini of the sequences, suggesting that this region encodes an essential fragment of the protein, possibly the catalytic domain. In contrast to the high homology within the putative catalytic domain, the N-termini of the proteins are distinctly different. Hydropathy analysis of the human, rat, and mouse sequences suggested the presence of an N-terminal transmembrane domain, similar to the majority of other glycosylation enzymes. In contrast, rEndoSpiro does not contain such a hydrophobic domain but does contain a glycine residue at position 2 (Spiro et al., 1997). This glycine has been suggested to be a putative myristoylation site that has been implicated as an alternative mechanism for membrane localization (reviewed in Boutin, 1997). In the absence of a typical type II membrane domain, the observed Golgi localization using antibodies raised against the catalytic domain was attributed to myristoylation (Zuber et al., 2000). To confirm the rEndoSpiro sequence and investigate the role of this glycine residue in cellular localization, we attempted to amplify a cDNA corresponding to the reported sequence (Spiro et al., 1997). Though we could amplify DNA encoding the catalytic domain we were unable to isolate the 5'-terminus, even after using several primers located proximal to the reported 5'-terminus and several different sources of rat liver cDNA.
In contrast, the rat isoform identified in this study (rEndo), encoding a transmembrane region, was readily amplified. Interestingly, when the rEndoSpiro sequence was BLAST against the recently published rat genome (Gibbs et al., 2004), bases 129–1356 of the ORF were located on chromosome 5, whereas bases 8–129 were located on chromosome 9. Conversely, the 5' sequence of rEndo is contiguous with the other three exons on chromosome 5, encoding the catalytic domain. Likewise, the complete hEndo and mEndo ORFs are located on single chromosomes, as expected. These observations, as enumerated in Results, indicated that we had isolated the intact sequences of human and rat endomannosidases, while proving that the previously published rat sequence (Spiro et al., 1997) was incomplete.
Analysis of the tissue distribution of hEndo transcript shows widespread expression, similar to the rat isoform (Spiro et al., 1997). Both hEndo and rEndo are highly expressed in the liver and kidney, but other tissues have significantly different expression patterns. Interestingly, in contrast to the hEndo, the rat isoform exhibited high expression levels in both the brain and lung (Spiro et al., 1997). The widespread expression of both isoforms of this enzyme in rat and human suggests that Endo may play a broad house-keeping role in the processing of N-glycans.
The expression of active recombinant hEndo in P. pastoris confirms the function of the isolated ORF. Interestingly, expression levels of the rat isoform, although highly homologous at the nucleotide and protein level, is several-fold higher than for the human protein. This may be due to the rat enzyme being inherently more stable during expression or in the culture medium. Both enzymes were processed at the C-termini, with the human enzyme completely processed with virtually no full-length form (59 kDa) of the enzyme detectable by western blot. In contrast, incomplete processing of the recombinant rEndo demonstrated the presence of some of the intact 59-kDa band. Similarly, when the rEndo was expressed in Escherichia coli the protein was proteolytically cleaved at the C-terminus (Spiro et al., 1997). Furthermore, affinity chromatographic purification of the rat isoform from rat liver demonstrates the presence of two forms, 56 and 60 kDa (Hiraizumi et al., 1994). Together, these data indicate that both hEndo and rEndo proteins are susceptible to proteolytic processing. The similar sizes of these two enzymes following proteolysis suggests that the cleavage site is identical, though the precise cleavage site in the bacterial, yeast, and mammalian expressed enzyme remains to be determined.
Traditionally, the removal of glucose from N-glycans was thought to occur exclusively in the ER by glucosidases I and II. However, the identification of the rEndo and its localization to the cis and medial Golgi demonstrated that glucose trimming may also occur subsequent to the ER (Roth et al., 2003; Zuber et al., 2000). The specific role of Endo is currently not understood. Affinity-purification of rEndo demonstrated copurification with calreticulin, suggesting a possible role in the quality control of N-glycosylation (Spiro et al., 1996). Alternatively, Endo may provide the cell with the ability to recover and properly mature glucosylated structures that have bypassed glucosidase trimming in the ER. Removing the glucose--1,3-mannose disaccharide from a glucosylated high-mannose structure creates a substrate for Golgi-residing mannosidases and glycosyltransferase enzymes, which allows the further maturation to complex N-glycans.
A phylogenetic survey of Endo indicates that this protein has emerged recently during evolution and is restricted to members of the chordate phylum, including mammals, birds, reptiles, amphibians, and bony fish, with the only exception being that it has also been identified in Mollusca (Dairaku and Spiro, 1997). In this study we identify three members of this family of glycosidic enzymes from mammalian sources that displayed distinct type II membrane protein architecture. Furthermore, we confirm that the transmembrane region is responsible for Golgi localization, elucidating for the first time the mechanism by which Endo is retained and membrane associated in this compartment.
|
Materials and methods |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
Cloning the hEndo and novel rEndo
To confirm that the two human sequences (GenBank: 18031877 and 20547441), described in Results, encode a single ORF we designed primers specific to the 5'-terminus of the gi:18031877 ORF and the 3'-terminus of the gi: 20547441 ORF (5'- ATGGCAAAGTTTCGGAGAAGGACTTGC-3' and 5'- TTAAGAAACAGGCAGCTGGCGATCTAATGC-3', respectively; named hE1 and hE2, respectively). These primers were used to amplify a 1389-bp fragment from human liver cDNA (BD Biosciences, Palo Alto, CA) using Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA) as recommended by the manufacturers, under the cycling conditions: 95°C for 1 min, 1 cycle: 95°C for 30 s, 60°C for 1 min, 72°C for 2.5 min, 30 cycles; 72°C for 5 min, 1 cycle. The DNA fragment produced was incubated with Taq DNA polymerase for 10 min at 68°C and cloned into pCR2.1 (Invitrogen, Carlsbad, CA). DNA sequencing confirmed that both of the human sequences identified by BLAST, produced one complete ORF, encoding hEndo. This confirmed construct was named pSH131.
The ORF encoding rEndo was amplified from rat liver cDNA (Clontech, Palo Alto, CA), as described, using the primers rE1 and rE2 (5'-GCAAAGCGTTTACTAGATTGGAGTG-3' and 5'-CCTTAATTAATTAAGAAACAGGCAGCTGGCGATCTAATGC-3', respectively). Subsequently the amplified DNA fragment containing the 1389-bp ORF was cloned into pCR2.1 (Invitrogen), sequenced, and designated pSH354.
Generation of recombinant Endo constructs and expression
To generate a secreted yeast form of the hEndo, a region encoding the putative catalytic domain was expressed in the EasySelect Pichia Expression kit (Invitrogen) as recommended by the manufacturer. Briefly, PCR was used to amplify the ORF fragment containing bases 178 to 1386 from pSH131 using the primers hE3 and hE4 (5'-GAATTCGCCACCATGGATTTCCAAAAGAGTGACAGA ATCAACAG-3' and 5'-GAATTCCCAGAAACAGGC AGCTGGCGATC-3', respectively, with an EcoRI restriction site engineered into each). The conditions used with Pfu Turbo were: 95°C for 1 min, 1 cycle; 95°C for 30 s, 55°C for 30 s, 72°C for 3 min, 25 cycles; 72°C for 3 min, 1 cycle. The product was incubated with Taq DNA polymerase, cloned into pCR2.1, and sequenced as described; the resulting clone was designated pSH178. From this construct the hEndo fragment was excised by digestion with EcoRI and subcloned into pPicZA (Invitrogen) digested with the same enzyme, producing pAW105. This construct was transformed into the P. pastoris yeast strain GS115 supplied with the EasySelect Pichia Expression kit, producing the strain YSH16. Subsequently the strain was grown in BMGY to an OD600 of 2 and induced in BMMY for 48 h at 30°C, as recommended by the kit manufacturers.
To confirm that the isolated ORF was an Endo, the rEndo catalytic domain was amplified and expressed in parallel as a positive control. Briefly, bases 178–1386 of rEndo, corresponding to the putative catalytic domain (amino acids 60–462), were amplified from rat liver cDNA (BD Biosciences) using the same conditions to amplify the cDNA encoding full-length rEndo already described. The primers used were rE3 and rE4 (5'-GAATTCGCCACCATGGACTTCCAAAGGAGTGATCGAATCGACAT-GG-3' and 5'-GAATTCCCTGAAGCAGGCAGCTGTTGATCC-3', respectively, with an EcoRI restriction site engineered into each). The PCR product was cloned into pCR2.1 and sequenced, and the resultant construct named pSH179. Subsequently, the encoded catalytic domain of rEndo was subcloned into pPicZA (Invitrogen) and expressed in GS115 as described, producing the construct pAW106 and the strain YSH13. It must be noted that the rEndo DNA sequence amplified for this pAW106 construct was generated from rat liver cDNA prior to obtaining the clone containing the complete rEndo sequence present in pSH354. However, sequencing has confirmed that rEndo bases 178–1386 were identical in each construct.
To detect recombinant hEndo and rEndo by western blot, a double FLAG tag was engineered 3' to the Kex2 cleavage site of the alpha-mating factor and 5' to the EcoRI restriction, used for Endo cloning, in pPicZA as follows. Briefly, the phosphorylated oligonucleotides FLAG1 and FLAG2 (5'-P-AATTTATGGACTACAAGGATGACG ACGACAAGG-3' and 5'-P-AATTCCTTGTCGTCGTC ATCCTTGTAGTCCATA-3', respectively) were annealed as previously described (Sambrook and Russell, 2001), and ligated into pPicZA digested with EcoRI. A construct containing two tandem FLAG tags in the correct orientation was named pSH241. Subsequently, the encoded catalytic domains (amino acid residues 60–462) of rEndo and hEndo were digested from pSH179 and pSH178 with EcoRI and ligated into pSH241, digested with the same enzyme. The resultant constructs encoding recombinant rEndo60–462 FLAG/His and hEndo60–462 FLAG/His were named pSH245 and pSH246, respectively. Transformation of these constructs into GS115 produced the strains YSH89 and YSH90, respectively. Expression and secretion of Endos from these strains was performed as already described.
In vitro characterization of recombinant Endo
GlcMan5GlcNAc2, substrate for Endo assays, was isolated from the och1 alg3 double knockout mutant strain RDP25 (Davidson et al., 2004). 2-AB-labeled GlcMan5GlcNAc2 was added to 10 µl of culture supernatant and incubated at 37°C for 8 h or overnight. Ten microliters of water was then added, and subsequently the glycans were separated by size and charge using an Econosil NH2 (Altech, Avondale, PA) 4.6 x 250 mm, 5 micron bead, amino-bound silica column as previously described (Choi et al., 2003). The pH optimum analysis was carried out by adjusting the culture supernatant to various pH (4.5–8.5, see Figure 4C) before incubating with the substrate at 37°C for 8 h. The preferred temperature for hEndo was performed by incubating the substrate with culture supernatant at room temperature, 30°C and 37°C.
Northern blot analysis
Tissue distribution of hEndo transcript was determined with a human multiple tissue northern blot (BD Biosciences) normalized to 2 µg purified poly A+ RNA from each of the tissues. A 550-bp fragment of human Endo was amplified from pSH131 using Taq DNA polymerase (Promega, Madison, WI) with the primers hE9 (5'-GTTAACCACCTCAGGGTCTCGGAGTATTCG-3' and hE2 (described previously). Subsequently, the probe was generated with this PCR product by [32P]dCTP-labeling using the RadPrime DNA Labeling System (Invitrogen).
SDS–PAGE and western blotting
Media from the P. pastoris cultures was analyzed for endomannosidase secretion by running samples on a 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) (Laemmli, 1970) using the Bio-Rad Mini-Protean II (Hercules, CA) apparatus. The proteins were then transferred onto a nitrocellulose membrane (Schleicher & Schuell, Keene, NH). Recombinant Endo was detected using the anti-FLAG M2 monoclonal in combination with a goat anti-mouse horseradish peroxidase–conjugated secondary antibody and visualized with the ECL western detection system (Amersham Biosciences, Buckinghamshire, U.K.).
Generation of GFP tagged Endo constructs and immunolocalization
Three C-terminal GFP tagged forms of hEndo were generated corresponding to the full-length protein (hEndo1–462 GFP), the N-terminus encoding the putative transmembrane domain (hEndo1–59 GFP) and the C-terminal catalytic domain (hEndo60–462 GFP). In addition a C-terminal GFP tagged form of the rEndo N-terminus (rEndo1–59 GFP) was generated. The rat N-terminal fusion was generated using pSH354 as template and the primers rE5 and rE6 (5'-GGAATTCGCCACCATGGCAAAATTCCGAAGAAGGACCTGCATC-3' and 5'-CCGAATTCGGCTTTGTTTT CCAAGTGGGTATTCGGTGG-3', respectively). Subsequently this fragment was cloned into pCR2.1, sequenced, and designated pSH355. The DNA encoding the N-terminal fragment of the hEndo (amino acids 1–59) was amplified from pSH131 using the primers hE5 and hE6 (5'-GAATTCGCCACC ATGGCAAAGTTTCGGAGAAGGACTTGCATC-3' and 5'-CCGAATTCAAAATTTTTCCCCAAATGAATAGTTCG-3', respectively). The resulting fragment was cloned into pCR2.1, sequenced, and designated pSH356. The DNA fragment encoding the C-terminal region of hEndo was amplified from pSH131 using the primers hE7 and hE8 (5'-GAATTCGCCACCATGGATTTCCAAAAGAGTGACAGAATCAACAG-3' and 5'-GAATTCAGAAACAGGCAGCTGGCGATC-3'). The subsequent fragment was cloned into pCR2.1, sequenced, and designated pSH357. The DNA encoding full-length hEndo was amplified using the template pSH131 and the primers hE5 and hE8, already described. The resulting fragment was cloned into pCR2.1 and designated pSH358. Subsequently, the constructs pSH355 to pSH358 were digested with EcoRI and the resulting fragments subcloned into pEGFP-N2 (BD Biosciences) producing the constructs pAP10, pAP11, pAP12, and pAP13 encoding the rEndo1–59 GFP, hEndo1–59 GFP, hEndo60–462 GFP, and hEndo1–462 GFP, respectively. See Figure 3C for illustration of the fusions generated.
CHO-K1 cells were transiently transfected with the constructs pAP10 to pAP13 and pEGFP-N2 to localize the GFP fusions as follows. Briefly, CHO-K1 cells were seeded at a density of 2 x 105 cells/well into six-well plates containing 22 x 22mm glass coverslips. After overnight incubation at 37°C, 4 µg of each construct was transformed into three separate wells using GenePORTER 2 Transfection reagent (Gene Therapy Systems, San Diego, CA) as described by the manufacturer. The cells were processed 48 h posttransformation by washing the coverslips three times with phosphate buffered saline (PBS) and fixed for 15 min in 4% paraformaldehyde. Subsequently the coverslips were washed three times with PBS and incubated with 2.5 µg/ml WGA-RITC (Vector Laboratories, Burlingame, CA) in PBS containing 0.2% Triton 100 for 30 min. The coverslips were washed three times with PBS and Vectashield H-1000 (Vector Laboratories) applied prior to mounting. The cells were analyzed using a Leica TCS SP confocal microscope setup for GFP and rhodamine detection. The separate images obtained for each fluorophore were superimposed using Adobe Photoshop.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
We thank Dr. M. Spinella (Department of Pharmacology and Toxicology, Dartmouth Medical School, Hanover) for his help in performing the northern blots. The nucleotide sequences reported in this article have been submitted to the GenBank/EBL Data Bank with accession numbers AY372528 [GenBank] and AY599499 [GenBank] for human and rat endomannosidases, respectively.
|
Abbreviations |
|---|
2-AB, 2-aminobenzamide; CHO, Chinese hamster ovary; Endo, endomannosidase; ER, endoplasmic reticulum; GFP, green flurorescent protein; ORF, open reading frame; PBS, phosphate buffered saline; PCR, polymerase chain reaction; SDS–PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis; WGA, wheat germ agglutinin
|
References |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
Bause, E. and Burbach, M. (1996) Purification and enzymatic properties of endo-alpha 1, 2-mannosidase from pig liver involved in oligosaccharide processing. Biol. Chem., 377, 639–646.[Web of Science][Medline]
Boutin, J.A. (1997) Myristoylation. Cell Signal, 9, 15–35.[CrossRef][Web of Science][Medline]
Choi, B.K., Bobrowicz, P., Davidson, R.C., Hamilton, S.R., Kung, D.H., Li, H., Miele, R.G., Nett, J.H., Wildt, S., and Gerngross, T.U. (2003) Use of combinatorial genetic libraries to humanize N-linked glycosylation in the yeast P. pastoris. Proc. Natl Acad. Sci. USA, 100, 5022–5027.
Dairaku, K. and Spiro, R.G. (1997) Phylogenetic survey of endomannosidase indicates late evolutionary appearance of this N-linked oligosaccharide processing enzyme. Glycobiology, 7, 579–586.
Davidson, R.C., Nett, J.H., Renfer, E., Li, H., Stadheim, T.A., Miller, B.J., Miele, R.G., Hamilton, S.R., Choi, B.K., Mitchell, T.I., and Wildt, S. (2004) Functional analysis of the ALG3 gene encoding the Dol-P-Man: Man5GlcNAc2-PP-Dol mannosyltransferase enzyme of P. pastoris. Glycobiology, 14, 399–407.
Dempski, R.E. Jr. and Imperiali, B. (2002) Oligosaccharyl transferase: gatekeeper to the secretory pathway. Curr. Opin. Chem. Biol., 6, 844–850.[CrossRef][Web of Science][Medline]
Gibbs, R.A., Weinstock, G.M., Metzker, M.L., Muzny, D.M., Sodergren, E.J., Scherer, S., Scott, G., Steffen, D., Worley, K.C., Burch, P.E., and others. (2004) Genome sequence of the brown Norway rat yields insights into mammalian evolution. Nature, 428, 493–521.[CrossRef][Medline]
Herscovics, A. (1999) Importance of glycosidases in mammalian glycoprotein biosynthesis. Biochim. Biophys. Acta, 1473, 96–107.[Medline]
Higgins, D.G. and Sharp, P.M. (1989) Fast and sensitive multiple sequence alignments on a microcomputer. Comput. Appl. Biosci., 5, 151–153.
Hiraizumi, S., Spohr, U., and Spiro, R.G. (1993) Characterization of endomannosidase inhibitors and evaluation of their effect on N-linked oligosaccharide processing during glycoprotein biosynthesis. J. Biol. Chem., 268, 9927–9935.
Hiraizumi, S., Spohr, U. and Spiro, R.G. (1994) Ligand affinity chromatographic purification of rat liver Golgi endomannosidase. J. Biol. Chem., 269, 4697–4700.
Kyte, J. and Doolittle, R.F. (1982) A simple method for displaying the hydropathic character of a protein. J. Mol. Biol., 157, 105–132.[CrossRef][Web of Science][Medline]
Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680–685.[CrossRef][Medline]
Lubas, W.A. and Spiro, R.G. (1988) Evaluation of the role of rat liver Golgi endo-alpha-D-mannosidase in processing N-linked oligosaccharides. J. Biol. Chem., 263, 3990–3998.
Marquardt, T. and Denecke, J. (2003) Congenital disorders of glycosylation: review of their molecular bases, clinical presentations and specific therapies. Eur. J. Pediatr., 162, 359–379.[Web of Science][Medline]
Moore, S.E. and Spiro, R.G. (1992) Characterization of the endomannosidase pathway for the processing of N-linked oligosaccharides in glucosidase II-deficient and parent mouse lymphoma cells. J. Biol. Chem., 267, 8443–8451.
Ray, M.K., Yang, J., Sundaram, S., and Stanley, P. (1991) A novel glycosylation phenotype expressed by Lec23, a Chinese hamster ovary mutant deficient in alpha-glucosidase I. J. Biol. Chem., 266, 22818–22825.
Reitman, M.L., Trowbridge, I.S., and Kornfeld, S. (1982) A lectin-resistant mouse lymphoma cell line is deficient in glucosidase II, a glycoprotein-processing enzyme. J. Biol. Chem., 257, 10357–10363.
Roth, J., Ziak, M., and Zuber, C. (2003) The role of glucosidase II and endomannosidase in glucose trimming of asparagine-linked oligosaccharides. Biochimie, 85, 287–294.[Medline]
Sambrook, J. and Russell, D.W. (2001) Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Spiro, M.J., Bhoyroo, V.D., and Spiro, R.G. (1997) Molecular cloning and expression of rat liver endo-alpha-mannosidase, an N-linked oligosaccharide processing enzyme. J. Biol. Chem., 272, 29356–29363.
Spiro, R.G., Zhu, Q., Bhoyroo, V., and Soling, H.D. (1996) Definition of the lectin-like properties of the molecular chaperone, calreticulin, and demonstration of its copurification with endomannosidase from rat liver Golgi. J. Biol. Chem., 271, 11588–11594.
Zuber, C., Spiro, M.J., Guhl, B., Spiro, R.G., and Roth, J. (2000) Golgi apparatus immunolocalization of endomannosidase suggests post-endoplasmic reticulum glucose trimming: implications for quality control. Mol. Biol. Cell, 11, 4227–4240.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Torossi, B. Guhl, J. Roth, and M. Ziak Endomannosidase undergoes phosphorylation in the Golgi apparatus Glycobiology, January 1, 2010; 20(1): 55 - 61. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Farrer, H. R. Kranzler, Y. Yu, R. D. Weiss, K. T. Brady, R. Anton, J. F. Cubells, and J. Gelernter Association of Variants in MANEA With Cocaine-Related Behaviors Arch Gen Psychiatry, March 1, 2009; 66(3): 267 - 274. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||