- Split View
-
Views
-
Cite
Cite
Elizabeth O'Connor, Birgit Eisenhaber, Jane Dalley, Tao Wang, Caroline Missen, Neil Bulleid, Paul N. Bishop, Dorothy Trump, Species specific membrane anchoring of nyctalopin, a small leucine-rich repeat protein, Human Molecular Genetics, Volume 14, Issue 13, 1 July 2005, Pages 1877–1887, https://doi.org/10.1093/hmg/ddi194
- Share Icon Share
Abstract
Mutations in the gene NYX, which encodes nyctalopin, lead to the retinal disorder congenital stationary night blindness which is characterized by defective night vision (nyctalopia) from birth. Nyctalopin is of unknown function but is predicted to be a secreted glycoprotein of the extracellular small leucine-rich repeat (SLRP) proteoglycan and protein family attached to the cell membrane in humans via a glycosylphosphatidylinositol (GPI) anchor but in mouse via a transmembrane domain. We investigated membrane association and attachment for human and mouse nyctalopin and show, conclusively, that human nyctalopin is a GPI anchored protein. In addition, the orthologous mouse protein, although it localizes to the cell surface, is not GPI anchored. We also confirm both mouse and human nyctalopins are glycosylated. Further sequence analysis suggests that chimp, dog and frog nyctalopins are likely to be GPI anchored but that rat nyctalopin is not. This is the first reported example of orthologous proteins which have different mechanisms of cell membrane attachment. Notably, the disease-causing mutations that have been identified to date in the human NYX gene are all distributed throughout the core LRR region and not in the C-terminal GPI anchor signal sequence. We propose that the presence of nyctalopin on the surface of the cell rather than the mechanism of anchoring is crucial to its function.
INTRODUCTION
Congenital stationary night blindness (CSNB) is a non-progressive retinal disorder characterized by defective night vision (nyctalopia) from birth (1) that is clinically and genetically heterogeneous. It is most commonly inherited as an X-linked disorder, but autosomal dominant (Nougaret) (OMIM no. 163500) and recessive (Oguchi) (OMIM no. 258100) forms have been described (2,3). In addition to nyctalopia, patients with the X-linked forms of CSNB have variably high myopia, nystagmus and decreased visual acuity. Mutations in either of the two X-linked genes can cause CSNB. CSNB1 (OMIM no. 310500) is caused by mutations in the NYX gene which maps to Xp11.4 and encodes the leucine-rich glycoprotein, nyctalopin (4,5) and CSNB2 (OMIM no. 300071) is associated with loss of function mutations in the α1-subunit of the L-type calcium channel gene, CACNA1F (6,7). Patients with either form of CSNB have a negative electroretinogram with a b-wave to a-wave ratio of <1 (Schubert–Bornschein waveform) (8), but detailed electrophysiological testing indicates that patients with CSNB1 (also called ‘complete’ CSNB) have very little, if any, rod photoreceptor activity and normal or only mildly abnormal cone activity (9), whereas patients with CSNB2 (‘incomplete’ CSNB) have residual rod activity and abnormal cone activity. The no b-wave (nob) mouse is a naturally occurring animal model for CSNB1 (10–12) with an 85 bp deletion in the mouse Nyx gene (11).
NYX expression is predominantly in the retina and kidney (4), although the expression has been observed in other tissues (5). The functions of nyctalopin are unknown but bioinformatic analysis predicts it to be a secreted glycoprotein, that in humans is attached to the cell membrane via a glycosylphosphatidylinositol (GPI) anchor (4,5). Phylogenetic analysis predicts nyctalopin to be a member of the extracellular small leucine-rich repeat proteoglycan and protein (SLRP) family (4,5). The SLRPs are a subset of the leucine-rich repeat (LRR) superfamily which contains tandem repeats of variable length with the consensus sequence LXXLXLXXNXL. These repeats align to form a curved structure with an inner concave surface forming a β-sheet and an outer convex surface of variable topology (13). SLRP family members are characterized by the presence of disulphide-bonded capping motifs at either end of the LRR domain.
To date, 35 disease-causing mutations in the NYX gene have been identified (4,5,14,15). These are thought to be inactivating and occur throughout most of the coding sequences of the gene. Non-sense mutations are reported throughout the gene and would be likely to result in the loss of part of the LRR region together with the anchor region or result in unstable mRNA. However, to date, no disease-causing mutations have been reported in the 3′ region of the gene encoding the predicted GPI anchor signal sequence and its cleavage site.
GPI anchorage occurs in ∼0.5% of cellular proteins in eukaryotes (16) and is the mechanism of membrane attachment for many cell surface glycoproteins (17). GPI anchor addition occurs in the endoplasmic reticulum and requires a C-terminal GPI signal sequence which is cleaved and immediately replaced with the lipid moiety (18). The site of cleavage of the protein and attachment of the lipid (ω-site) has been characterized (19) and the amino acids surrounding the ω-site (cleavage/attachment region) are critical to successful GPI anchor addition. Proteins can be cleaved from their GPI lipid anchor by the action of phosphatidylinositol phospholipase C and D (PI-PLC and PI-PLD) (20) or, alternatively, by angiotensin-converting enzyme (21).
Human and mouse nyctalopins are 84% identical at the amino acid level, with the region of least homology being at the C-terminus (11). Bioinformatic analysis suggests that mouse nyctalopin is attached to the cell membrane via one or two transmembrane (TM) domains and not a GPI anchor (11). In view of the surprising prediction of human and mouse nyctalopin having different mechanisms of membrane attachment, together with the lack of disease-causing mutations in the 3′ region of the human gene we investigated membrane association and attachment for both human and mouse nyctalopin. First, we show that wild-type (WT) human and mouse nyctalopins are both present on the cell surface and that removal of the anchor sequence of either protein results in secretion into the cell medium, indicating that both human and mouse WT nyctalopin have functional membrane anchors. We also demonstrate that the different anchor predictions are correct as human nyctalopin has a functional GPI anchor but mouse does not. Further sequence analysis suggests chimp, dog and frog nyctalopins are likely to be GPI anchored but that mouse and rat nyctalopins are not. Moreover, nyctalopin is absent in non-vertebrates and appears to represent a new acquirement during the late evolution of chordates. To our knowledge, this is the first example of a protein being GPI anchored in some species but not in others.
RESULTS
Sequence architecture of the nyctalopin protein family and analysis of the C-terminal sequence properties
Sequence similarity searches with human nyctalopin (GenBank accession no. Q9GZU5) in the non-redundant protein database of the National Center of Biological Information (www.ncbi.nih.gov) finds orthologues in chimp (XP_528943), dog (XP_548956), mouse (P83503), rat (XP_228750), chicken (XP_416770), frog (AAH81163) and in the fish Tetraodon nigroviridis (CAG02639) (with E-values below 10−135 in a BLAST search (22). This sequence group is clearly distinguished from other proteins with LRRs by a large gap in E-values (best other hit with E>10−32). Further nyctalopins have been found in the proteomes of the fishes Fugu rubripes (SINFRUP00000147617) and Danio rerio (zebrafish, GENSCAN00000037001) as best, outstanding hits with E-values below 10−80 with BLAST searches (www.ensembl.org). Clear indications for nyctalopin orthologues in non-vertebrates have not been detected.
Several nyctalopin proteins that are derived from genome sequencing efforts without additional experimental information appear incomplete. The N-terminal sequence appears to be missing for the zebrafish and Tetraodon fish proteins, and the Fugu fish and Tetraodon fish proteins are probably C-terminally truncated. Other representatives (from rat and dog) appear too long in the N-terminus and are artificially fused with non-nyctalopin artefacts due to incorrect genome assembly or gene annotation.
As sequence-analytic results for the individual nyctalopins indicate, this protein family is characterized by a typical domain architecture (Fig. 1). Apparently, all nyctalopins carry a cleavable signal peptide, which is confidently predicted with TargetP (23) for the human, mouse, rat, frog, chicken and Fugu fish cases. Analysis of the nyctalopin sequence and comparison with known structures of LRR proteins predicts the presence of 13 LRRs flanked by cysteine-rich capping motifs (Fig. 1). LRR1 and LRR13 contain residues that are involved in the N- and C-terminal capping, respectively. The N-terminal capping motif resembles that of the Nogo receptor (24) and the SLRP decorin (13), whereas the C-terminal capping motif resembles that of the SLRP chondroadherin and platelet glycoprotein 1bα (25).
C-terminal to the LRRs is a region rich in Ser/Thr residues predicted to be O-glycosylated which links to the C-terminal ‘anchor’ regions. These hydrophobic segments could represent C-terminal TM regions or hydrophobic tails of GPI lipid anchor attachment signals, although these tend to be longer (≥20 residues) and more hydrophobic in the first case. TM prediction tools do tend to generate false-positive predictions in protein segments coding for helices buried inside of globular structures (as is the case of LRR proteins). If restricted to the C-terminal segment, the method HMMTOP (26) predicts TM hits only for the mouse sequence (region 455–474) and the homologous region of the rat. Both, DAS-TMfilter (27) and TMHMM (28) search for hydrophobic cores of candidate TM regions and do not report hits for the C-termini of all observed nyctalopin proteins.
The prediction of GPI lipid anchoring with the big-Π tool (29) is very sensitive to the correctness of the sequence and even a single amino acid exchange can modify the prediction. All nyctalopins differ from the consensus derived from known, efficiently GPI lipid anchored metazoan proteins (19). If one allows for a polar window with three residues (QMD) after the hydrophobic C-terminal tail and for a bulky spacer region, the human and chimp nyctalopins are predicted as targets for GPI lipid anchoring (with potential ω-site Gly453 for the human sequence). In addition, the dog and frog nyctalopins are predicted by the plant version of big-Π (30) as GPI lipid anchoring targets (with ω-site Ser509 in the context …LP-S-SGL… and ω-site Asn441, respectively). The mouse and rat sequences are clearly rejected by all versions of big-Π as GPI lipid anchoring targets mainly because of the absence of a potential ω-site. The chicken and the zebrafish sequences have an appropriate ω-site region but the sequence detail of the spacer and tail regions is untypical for a metazoan GPI lipid anchored protein.
Sequence-analytic evidence, therefore, suggests that the rodent proteins are membrane-attached via a C-terminal TM region and do not represent targets for GPI lipid anchoring, however, human and possibly, dog and frog nyctalopins do.
Human and mouse nyctalopins are anchored to the cell surface
Human (HWT) and mouse (MWT) nyctalopin constructs encoding flag tags at the N-terminus of the mature protein were cloned into the pcDNA3.1 expression vector. Flag-tagged constructs lacking the C-terminal (predicted anchor) regions were also generated (HNA, MNA) (Fig. 2A). ‘Anchor swap’ constructs were also produced by an exchange of the C-terminal regions of human and mouse nyctalopin (human+mouse anchor (HAS) and mouse+human anchor (MAS). In other constructs, mutations were introduced into the region around the predicted GPI anchor cleavage/attachment site (residues 452–455) in human (HMUT) and the corresponding amino acids in mouse (MMUT) (residues 447–450). The mutations changed the region around the cleavage recognition site in human from GGAG to TTTT and, correspondingly, in mouse MVFC was changed to TTTT (Fig. 2A).
All constructs were expressed in COS-7 cells and immunoblot analysis of whole cell lysate and concentrated cell medium samples was performed using anti-flag antibodies (Fig. 2B). Major bands were present in the whole cell lysate samples of ∼60 kDa for HWT and 62 kDa for MWT. Neither HWT nor MWT were observed in the concentrated medium samples indicating that these proteins are not secreted. The HNA and MNA proteins are present in the cell lysate, with approximate molecular weights of 58 and 60 kDa, respectively. This 2 kDa difference in apparent molecular weight between the WT and the NA forms of the protein in both human and mouse can be accounted for by the loss of the GPI-anchor in human and the C-terminal TM region in mouse. Both HNA and MNA are secreted into the cell medium, indicating that functional membrane anchors have been lost. All of the secreted HNA and most of the secreted MNA had a slower electrophoretic mobility in the cell lysate (Fig. 2B). The anchor swap (HAS and MAS) constructs, and HMUT and MMUT, were not observed in the concentrated medium samples but were present in the cell lysates with similar apparent molecular weights to the WT constructs (Fig. 2B).
The cell surface localization of nyctalopin was investigated by analysis of cell surface proteins following biotinylation of live cells. Live transfected cells were treated with biotin, thus labelling all cell surface proteins, then lysed. Incubation with streptavidin coated sepharose beads followed by centrifugation resulted in the isolation of the biotinylated cell surface proteins. Western blots probed with the anti-flag antibody of the biotinylated (cell surface) and non-biotinylated (intracellular) fractions indicated that all the nyctalopin constructs expressed in COS-7 cells were present on the cell surfaces as well as within the cells (Fig. 3). Negative controls were carried out in the absence of biotin (shown for HWT only, Fig. 3A). In addition, the membranes were subsequently stripped and reprobed with antibodies against the intracellular control calnexin, which, as expected, was not present in the cell surface fractions (shown for mouse WT and mouse NA) (Fig. 3) confirming the specificity of the fractionation.
Human nyctalopin is GPI anchored, whereas mouse is not
It has been demonstrated previously that GPI anchor addition can be studied in vitro when a GPI-anchored protein is translated in the presence of semi-permeabilized cells (31). In addition, the presence of hydrazine, a specific inhibitor of GPI anchor addition, results in a protein–hydrazine complex which is distinguishable from a GPI-anchored protein product because of the lower apparent molecular weight on SDS–PAGE. This system does not affect the normal protein translation and processing of non-GPI-anchored proteins. HWT and MWT nyctalopin RNA were translated in this system and subjected to SDS–PAGE (Fig. 4A). WT protein translated in K562 cells in the presence of hydrazine has a lower molecular weight than protein translated in the absence of hydrazine, indicating that GPI anchor addition occurs unless inhibited by hydrazine. There was no visible size change of MWT protein products under the same conditions. MWT translation products were then treated with endoglycosidase H, to remove glycosylation, which reduced the apparent molecular weight of the proteins by ∼10 kDa. Again, the major bands are the same in the presence or absence of hydrazine (Fig. 4C). This indicates that mouse protein is not GPI anchored. HWT and MWT translations were also carried out using the K class mutant cell line as a control. The catalytic subunit of the transamidase, a critical enzyme in the GPI anchor addition pathway is defective in this cell line (32). The addition of hydrazine during translation in these cells has no effect on the size of the protein product as GPI anchor addition does not occur; however, the presence of hydrazine reduces the efficiency of the translation and, therefore, results in indistinct bands on SDS–PAGE. This was seen in both control and experimental reactions (Fig. 4).
Phosphatidylinositol-phospholipase D (PI-PLD) specifically cleaves GPI anchors. Incubation of live cells with PI-PLD releases intact proteins from their GPI anchors into the cell medium. We investigated the susceptibility of nyctalopin to PI-PLD by treating transfected cells. Cell lysate and medium samples before and after treatment were subjected to western blot analysis (Fig. 5). Human WT nyctalopin (HWT) was observed in the cell medium following PI-PLD treatment (Fig. 5A). This protein migrated faster on SDS–PAGE (equivalent to 2 kDa) than protein in the whole cell lysate equivalent to the size of the GPI anchor. As a negative control, cells were incubated in PI-PLD buffer only under the same conditions and this did not result in release of the protein into the cell medium. MWT transfected cells were also incubated with PI-PLD (Fig. 5A), the results from this experiment indicate that MWT is not sensitive to PI-PLD.
Cells transfected with the anchor swap constructs were also subjected to PI-PLD treatment (Fig. 5B). The HAS protein was not released into the cell medium after treatment with PI-PLD and, therefore, is not GPI anchored. However, the MAS protein was released by PI-PLD. This confirms that the anchor region of the human WT protein confers GPI anchoring and that the GPI anchoring ability of the human protein is lost when the anchor region is replaced by the C-terminal portion of the mouse protein.
The results of PI-PLD treatment of the mutated constructs is shown in Figure 5C. These constructs contain mutations in the C-terminal region in mouse (MMUT) and around the predicted GPI anchor cleavage/attachment site in human (HMUT). PI-PLD does not release either HMUT or MMUT protein into the cell medium and, therefore, neither of these proteins has a functional GPI anchor, indicating the importance of these residues for GPI anchor addition in the WT human protein.
Nyctalopin is heavily glycosylated
Nyctalopin is predicted to have a molecular weight of ∼50 kDa. However, HWT protein has an apparent molecular weight of ∼60 kDa by SDS–PAGE and the secreted HNA form had an apparent molecular weight of ∼70 kDa. We hypothesized that the observed increase in molecular weight compared with the predicted value was caused by glycosylation of the protein. Computer-based predictions have shown human nyctalopin to contain six putative N sites (33) and four putative O-glycosylation sites (34). HWT and MWT cell lysate samples were incubated with N-glycosidase F, which removes N-glycan side chains, and with enzymes which remove O-glycan chains. Treated and untreated samples were analyzed by western blotting and probing with the anti-flag antibody (Fig. 6A). Treatment with N-glycosidase F reduced the apparent molecular weight of the HWT protein by ∼10 kDa confirming the prediction that nyctalopin is substituted with N-glycan side chains. Treatment with O-glycosidase in isolation had no effect on the size of the protein. However, treatment of nyctalopin with neuraminidase in addition to O-glycosidase reduced the size of the protein (Fig. 6A). This treatment was repeated on concentrated HNA medium samples (Fig. 6B). There is no difference in bands between the untreated and O-glycosidase only treated samples, however, treatment with both neuraminidase and O-glycosidase together slightly decreased the electrophoretic migration suggesting substitution with sialylated O-glycan side chains. These results, therefore, confirm the prediction that nyctalopin is glycosylated.
Cell lysate and medium samples were treated with neuraminidase, O-glycosidase and N-glycosidase and compared with untreated samples. It can be seen from Figure 6B that treatment with glycosidases significantly reduces the size of both the cell lysate and medium samples. However, the secreted protein still has a higher molecular weight which may be due to residual glycosylation.
Mouse nyctalopin appears to have a higher molecular weight than the human protein, running slower on SDS–PAGE. Both HWT and MWT cell lysates were treated with glycosidases which increased the electrophoretic migration, although the mouse protein still migrated slower than the human protein (Fig. 6A). Bovine fetuin was used as a control in all glycosylation reactions (Fig. 6C) as this protein is heavily glycosylated.
DISCUSSION
Mutations in the NYX gene, encoding nyctalopin, lead to CSNB1 in humans and the nob phenotype in mice. Nyctalopin is predicted to be a secreted extracellular leucine-rich glycosylated protein attached to the cell membrane via a GPI anchor in humans (4,5) but surprisingly via a TM domain in mouse (35). Previous work has indicated that nyctalopin is present on the cell surface (36) and it has been assumed, therefore, that the predicted GPI anchoring is functional but there has been no experimental verification of the anchoring mechanism.
In order to experimentally test the bioinformatic predictions, we expressed flag tagged constructs of human and mouse nyctalopin, constructs lacking the C-terminal anchor regions, constructs with swapped C-terminal anchor regions and constructs with the mutations of the predicted anchor cleavage site. For both human and mouse, the proteins lacking the C-terminal anchor sequence were secreted into the cell medium, whereas the WT proteins were present only in the cell lysate indicating the importance of this region in membrane attachment for each protein. Moreover, both HWT and MWT nyctalopin were present on the cell surface consistent with the predictions that nyctalopin is membrane anchored and confirming previous immunofluoresence data (36). Interestingly, the proteins lacking the C-terminal anchor region were also present in the cell surface fraction suggesting that the secreted protein is interacting with a cell surface component and is thus pulled down by the biotin assay. Some SLRP proteins are known to dimerize such as opticin (37) and decorin (38) and it may be that nyctalopin is interacting with endogenous nyctalopin on the cell surface of COS-7 cells as they do express nyctalopin at a low level (data not shown).
We investigated membrane anchorage by specifically testing for the occurrence of GPI-anchor attachment. The results of the in vitro transcription/translation assays of HWT and MWT in semi-permeabilized cells prove, conclusively, that GPI anchor addition occurs post-translation in human nyctalopin. The results of the mouse translations indicated that the protein was not subject to GPI modification and further data from PI-PLD assays confirm these results. The residues of the cleavage/attachment (ω-site) site region are critical for GPI anchor addition. Mutating the predicted ω-site region in human nyctalopin (amino acids 452–455, GGAG) to amino acids that would prevent GPI anchorage (TTTT) resulted in nyctalopin being present on the cell surface but not released by PI-PLD, suggesting a conversion to a TM domain. In addition, the exchange of the C-terminal regions of human and mouse nyctalopin resulted in the reversal of susceptibility to PI-PLD providing further evidence that the C-terminal region of human nyctalopin, but not mouse nyctalopin, is a signal sequence for GPI anchor addition.
All nyctalopin constructs expressed in this study had a higher apparent molecular weight than predicted from the amino acid sequence due to modification of the protein by glycosylation. Mouse nyctalopin, despite a smaller transcript length, appears ∼2 kDa higher than human after separation by SDS–PAGE, in WT and the mutated constructs and is, therefore, not accounted for by GPI cleavage. Exchange of the anchor regions reverses this size difference and we proposed that it was because of an additional glycosylation event occurring in the C-terminal region of the mouse protein. However, treatment of the human and mouse WT proteins with glycosidases did not reduce the protein sizes to the same extent. HNA and MNA proteins secreted into the cell medium but consistently have an apparent higher molecular weight than cellular proteins and in the mouse is observed as a duplet, which may be due to differences in glycosylation. The size of these glycoproteins reduces dramatically after treatment with glycosidases but does not reduce to the same extent as cellular protein and requires further investigation.
The family of vertebrate nyctalopins described in this work is a novel group in the SLRP superfamily. Apparently, nyctalopin is absent in non-vertebrates and represents a new acquirement during the late evolution of chordates. The varying mechanism of membrane attachment within the nyctalopin family is novel. Although this is the first described example of alternative membrane binding with a GPI lipid anchor or a TM region within a protein family, the variation between different types of acyl- or prenyl-anchors and other membrane attachment factors has already been reported for groups of closely related proteins (39). Within the family of folate receptors, three human isoforms α, β and γ differ in their proportions of secreted and membrane anchored protein: the α isoform is GPI lipid anchored, γ is secreted and β is ∼50% secreted and 50% GPI anchored (40). These examples emphasize that even a high degree of similarity in the globular domain part of the protein sequence does not necessarily imply the conservation of post-translational modification sites or localization signals in the non-globular segments.
In general, nyctalopin proteins seem to have a very weak C-terminal hydrophobic core. As the prediction results show, this does not depend on what type of signal (GPI or C-terminal TM region) they carry. In some cases, neither a GPI signal nor a TM region could be predicted. It should be noted that TM regions in TM aggregates are often not very hydrophobic. Alternatively, nyctalopins might fulfil their biological function being a member of a membrane anchored protein complex without the strong need to have their own membrane anchor. Such a situation has been shown for the PIG-K/GPI8 protein family (41) where some members of the family carry an own C-terminal TM region and others do not.
The results presented here show, conclusively, that human nyctalopin is a GPI anchored protein but that mouse nyctalopin, although it localizes to the cell surface, is not. This is the first reported example of orthologous proteins having different mechanisms of cell membrane attachment. Notably, the disease-causing mutations that have been identified to date in the human NYX gene are all distributed throughout the core LRR region and not in the C-terminal GPI anchor signal sequence. The function of oligodendrocyte myelin glycoprotein (OMgp), another GPI anchored LRR protein has been shown to be dependent on the LRR region only as deletion of the GPI region does not result in the loss of function (42). We propose a similar mechanism for nyctalopin function and suggest that its presence on the surface of the cell rather than its mechanism of anchoring is crucial to its function.
The function of nyctalopin in the retina remains unknown. Recent data suggest a role in ON bipolar cell signalling as the CSNB1 ERG findings can be mimicked in non-human primates by pharmacological blocking of the ON bipolar cell pathway with 2-amino-4-phosphonobutyrate (43). In addition, mutations have recently been described in CSNB patients in the gene GRM6 which encodes the glutamate receptor MgluR6, present only in the synapses of ON bipolar cell dendrites (44). These intriguing data suggest that nyctalopin and MgluR6 have closely related roles in synapse formation and signal processing in the ON bipolar cell pathway.
MATERIALS AND METHODS
Cloning of NYX and Nyx
PCR of the NYX gene was carried out using the Advantage GC-2 PCR kit (Clontech) on a Peltier thermal cycler 200. Primers were synthesized by MWG-biotech. Owing to the high GC content and regular occurrence of mutations following PCR on long templates, several fragments of the gene were amplified from genomic DNA and cloned using the pGEM®-T Easy Vector System II (Promega). Unique restriction sites throughout NYX were used to assemble the whole gene in the pcDNA3.1 vector. Mouse Nyx was cloned into pcDNA3.1 from an IMAGE cDNA clone (IMAGE:5362608). The 24 bp flag tag sequence (GACTACAAGGATGACGATGACAAG) was incorporated into the constructs by PCR using forward primers covering the sfiI restriction site and including the additional 24 bp in frame (human—TGGGCCGTGGGGGCCGACTACAAGGATGACGATGACAAGTGCGCCCGCGCTTGTCCCGCCG, mouse—AAAGGCCACCGAGGCCGACTACCCGGATGACGATGACAAGTGTCTGCGGGCCTGCCCTGCGGC). The amplified fragments were cloned into NYX and Nyx in pcDNA3.1 using unique restriction sites. Flag-tagged constructs will be subsequently referred to as HWT (human) and MWT (mouse).
Constructs lacking the predicted anchor region were produced in two stages. PCR of 200 bp fragments was carried out using reverse primers containing a stop codon (TGA) at codon 455 (human) and 446 (mouse). These fragments were cloned back into HWT and MWT to produce flag-tagged HNA (human) and MNA (mouse) constructs. The ‘anchor swap’ constructs were created by an exchange of the sequence 3′ of the unique FseI restriction sites at 1107 (human) and 1092 bp (mouse). These constructs are referred to as HAS (human with mouse anchor region) and MAS (mouse with human anchor).
Mutation of amino acids 452–455 in human (GGAG-TTTT) and 447–450 in mouse (MVFC-TTTT) was carried out using the QuikChange® Site-Directed Mutagenesis kit according to manufacturers instructions. Complimentary sets of primers were used (human—CTCCTCCCGTGGGGTGACAACCACGACCCGGCAGCCCTGG, mouse—CGGTCCCATGCAGTGACGACCACCACCTACAAGGCCACG). HWT and MWT were used as templates and the mutated constructs are referred to as HMUT and MMUT.
Sequences of all constructs were confirmed by DNA sequencing. The Endofree Plasmid Maxi Kit (Qiagen) was used to produce endotoxin free plasmid DNA suitable for transfections.
Cell culture and transfections
COS-7 cells were cultured in Dulbeccos modified Eagle's medium, supplemented with 2 mm L-glutamine, 10% fetal bovine serum and 100 U/ml streptomycin/penicillin, in 75 cm2 flasks, at 37°C, 5% CO2. Approximately 1×106 cells per well were seeded in six well plates 24 h prior to transfection. Transfections were carried out using lipofect AMINE (Invitrogen). Cells were transfected with 1 µg of endotoxin-free DNA in a 1% solution of lipofectAMINE in Optimem-1 (Gibco) medium. After incubation for 5 h at 37°C, the transfection medium was removed and replaced with Optimem-1 medium. Transfected cells were incubated for 72 h.
Western blot analysis
Samples in SDS-loading buffer (62.5 mm Tris–HCl, pH 6.8, 25% glycerol, 2% SDS, 5% β-mercaptoethanol and Bromophenol blue) were denatured at 95°C prior to SDS–PAGE. Denatured proteins were run through a 4% acrylamide stacking gel at 100 V and then through a 10% separating gel at 150 V. Proteins were blotted onto Hybond-ECL Nitrocellulose membrane (Amersham) using a Trans-Blot SD Semi-Dry Transfer Cell (Bio-Rad) at 15 V for 1 h. Membranes were blocked with 5% marvel in TBS. Blots were incubated with primary antibody [Anti-flag® M2 monoclonal antibody (Sigma)—1:2000 dilution or anti-calnexin goat polyclonal (Santa Cruz Biotechnology)—1:500] in 5% marvel. Membranes were then washed 4×5 min with 1×TBS-T and incubated with secondary antibody [goat anti-mouse HRP (for flag), rabbit anti-goat HRP (for calnexin), both Dakocytomation], 1:2000 dilution in 5% marvel, for 90 min at room temperature. After washing, membranes were incubated with ECL Western Blotting Detection Reagents (Amersham Biosciences) for 1 min, then exposed onto Hyperfilm ECL (Amersham Biosciences). Coomassie staining of acrylamide gels was carried out using the Bio-safe Coomassie (Bio-Rad) system according to the manufacturer's instructions.
Cell surface biotinylation
All steps were carried out at 4°C. Medium samples from transfected cells were concentrated 20× using Vivaspin concentrator columns as described earlier. Cells were washed 3× with 1× PBS and incubated with 0.5 mg/ml EZ-Link Sulfo-NHS-LC-biotinat (Pierce) in PBS for 30 min. The free biotin was quenched by washing three times with cold PBS containing 100 mm glycine. Cells were lysed in RIPA buffer (1% IGEPAL, 0.5% Deoxycholate and 0.1% SDS in PBS) containing protease inhibitors for 30 min. Lysates were centrifuged at 15 000×g, 15 min at 4°C and the supernatant (whole cell lysate) recovered. Half of this fraction was incubated with equal volumes of streptavidin–sepharose beads overnight with rotation. The intracellular fraction (supernatant) was recovered after centrifuging at 15 000×g for 15 min. The beads were washed 4× in RIPA buffer, centrifuging after each wash. The beads were then resuspended in SDS-loading buffer and denatured at 100°C for 5 min prior to centrifuging at 15 000×g for 15 min. The supernatant (biotinylated cell surface fraction) was recovered. Medium and cell fractions were subjected to western blot analysis.
In vitro transcription/translation
Five micrograms of linearized plasmid DNA (HWT and MWT) were used as templates in RNA synthesis using T7 RNA polymerase. The reactions were carried out using the mMessenger mMachine T7 kit (Ambion) according to the manufacturer's instructions. The reactions were terminated after 90 min incubation at 37°C and subjected to phenol/chloroform extraction followed by ethanol precipitation. RNA pellets were resuspended in 50 µl of sterile water containing DTT and RNasin and stored at −80°C.
The in vitro translation of HWT and MWT was carried out as previously described (31). Translations, in semi-permeabilized K562 and class K mutant K562 cells, were carried out in the presence and absence of hydrazine. Semi-permeabilized cells were prepared as previously described (45). Endoglycosidase H treatment of MWT translation products was also carried out as described previously (31). All the samples were resuspended in SDS-loading buffer, and resolved by SDS–PAGE through 9% gels. Gels were fixed [10% (v/v) methanol, 10% (v/v) acetic acid], dried and visualized using a FujiBas 2000 phosphoimager.
PI-PLD assays
Lyophilized PI-PLD (Calbiochem) was dissolved in PI-PLD buffer (10 mm Tris–HCl pH 8.0, 0.05% BSA, 0.1% Triton X-100) to a stock concentration of 0.05 U/µl. The medium was removed from transfected cells and retained. The cells were washed with Optimem-1 medium prior to incubation with 1 U/ml PI-PLD in Optimem-1 for 1 h at 37°C. Negative controls were incubated without enzyme in Optimem-1 with PI-PLD buffer. Medium samples before and after PI-PLD treatment were centrifuged for 5 min at 16 000 g, 4°C to pellet cell debris, concentrated 20× at 3000 RCF, 4°C in Vivaspin 2 ml Concentrator (Vivascience) columns in a Sorvall Legend RT Centrifuge before addition of 6× SDS–PAGE loading buffer. Cells were washed with Optimem-1 and whole cell lysate samples were collected in 6× SDS–PAGE loading buffer.
Glycosylation
Bovine Fetuin (Sigma) (10 µg) was used as a control in all glycosylation reactions. Fetuin control samples were resolved on SDS–PAGE and results visualized following Coomassie staining.
N-glycosylation
Medium samples from transfected cells were concentrated 100× using Vivaspin concentrator columns as described earlier, then diluted one in five in N-glycosidase buffer (0.5% IGEPAL (Sigma) in PBS, 0.1% SDS, 0.1% β-mercaptoethanol). Cell lysate samples were collected by scraping cells into N-glycosidase buffer. Twenty five microlitres cell lysate and medium samples were denatured at 100°C for 3 min, cooled and 0.05 m EDTA was added prior to incubation with 1 U N-glycosidase F (Roche) overnight at 37°C. Control reactions minus enzyme were also incubated.
O-glycosylation
Transfected cells were homogenized in O-glycosidase buffer [10 mm sodium phosphate pH 7.2 with EDTA-free protease inhibitors (Roche)]. Medium samples (concentrated 100×) were diluted one in five in O-glycosidase buffer. Twenty five microlitres reactions were incubated at 37°C overnight with 1 mU O-glycosidase (Roche). To test for sialyated O-glycosylation reactions containing both 1 mU O-glycosidase and 20 mU neuraminidase (Roche) were also performed. Reactions containing all three enzymes for de-glycosylation were set up with samples in O-glycosidase buffer. Six times SDS–PAGE loading buffer was added to all glycosylation samples prior to SDS–PAGE and western blotting.
ACKNOWLEDGEMENTS
This work was supported by the Medical Research Council UK. E.O'C. is an MRC PhD student. P.N.B. is a Wellcome Trust Senior Research Fellow in Clinical Science.
Conflict of Interest statement. None declared.
References
Jacobi, F.K., Andreasson, S., Langrova, H., Meindl, A., Zrenner, E., Apfelstedt-Sylla, E. and Pusch, C.M. (
Nettleship, E. (
Maw, M.A., John, S., Jablonka, S., Muller, B., Kumaramanickavel, G., Oehlmann, R., Denton, M.J. and Gal, A. (
Bech-Hansen, N.T., Naylor, M.J., Maybaum, T.A., Sparkes, R.L., Koop, B., Birch, D.G., Bergen, A.A., Prinsen, C.F., Polomeno, R.C., Gal, A. et al. (
Pusch, C.M., Zeitz, C., Brandau, O., Pesch, K., Achatz, H., Feil, S., Scharfe, C., Maurer, J., Jacobi, F.K., Pinckers, A. et al. (
Strom, T.M., Nyakatura, G., Apfelstedt-Sylla, E., Hellebrand, H., Lorenz, B., Weber, B.H., Wutz, K., Gutwillinger, N., Ruther, K., Drescher, B. et al. (
Bech-Hansen, N.T., Naylor, M.J., Maybaum, T.A., Pearce, W.G., Koop, B., Fishman, G.A., Mets, M., Musarella, M.A. and Boycott, K.M. (
Scholl, H.P., Langrova, H., Pusch, C.M., Wissinger, B., Zrenner, E. and Apfelstedt-Sylla, E. (
Miyake, Y., Yagasaki, K., Horiguchi, M., Kawase, Y. and Kanda, T. (
Pardue, M.T., McCall, M.A., LaVail, M.M., Gregg, R.G. and Peachey, N.S. (
Gregg, R.G., Mukhopadhyay, S., Candille, S.I., Ball, S.L., Pardue, M.T., McCall, M.A. and Peachey, N.S. (
Candille, S.I., Pardue, M.T., McCall, M.A., Peachey, N.S. and Gregg, R.G. (
Scott, P.G., McEwan, P.A., Dodd, C.M., Bergmann, E.M., Bishop, P.N. and Bella, J. (
Zito, I., Allen, L.E., Patel, R.J., Meindl, A., Bradshaw, K., Yates, J.R., Bird, A.C., Erskine, L., Cheetham, M.E., Webster, A.R. et al. (
Zeitz, C., Minotti, R., Feil, S., Matyas, G., Cremers, F.P., Hoyng, C.B. and Berger, W. (
Eisenhaber, B., Bork, P. and Eisenhaber, F. (
Coyne, K.E., Crisci, A. and Lublin, D.M. (
Caras, I.W., Weddell, G.N. and Williams, S.R. (
Eisenhaber, B., Bork, P. and Eisenhaber, F. (
Udenfriend, S. and Kodukula, K. (
Kondoh, G., Tojo, H., Nakatani, Y., Komazawa, N., Murata, C., Yamagata, K., Maeda, Y., Kinoshita, T., Okabe, M., Taguchi, R. et al. (
Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W. and Lipman, D.J. (
Emanuelsson, O., Nielsen, H., Brunak, S. and von Heijne, G. (
Barton, W.A., Liu, B.P., Tzvetkova, D., Jeffrey, P.D., Fournier, A.E., Sah, D., Cate, R., Strittmatter, S.M. and Nikolov, D.B. (
Uff, S., Clemetson, J.M., Harrison, T., Clemetson, K.J. and Emsley, J. (
Tusnady, G.E. and Simon, I. (
Cserzo, M., Eisenhaber, F., Eisenhaber, B. and Simon, I. (
Krogh, A., Larsson, B., von Heijne, G. and Sonnhammer, E.L. (
Eisenhaber, B., Bork, P. and Eisenhaber, F. (
Eisenhaber, B., Wildpaner, M., Schultz, C.J., Borner, G.H., Dupree, P. and Eisenhaber, F. (
Dalley, J.A. and Bulleid, N.J. (
Yu, J., Nagarajan, S., Knez, J.J., Udenfriend, S., Chen, R. and Medof, M.E. (
Sigrist, C.J., Cerutti, L., Hulo, N., Gattiker, A., Falquet, L., Pagni, M., Bairoch, A. and Bucher, P. (
Julenius, K., Molgaard, A., Gupta, R. and Brunak, S. (
Pesch, K., Zeitz, C., Fries, J.E., Munscher, S., Pusch, C.M., Kohler, K., Berger, W. and Wissinger, B. (
Zeitz, C., Scherthan, H., Freier, S., Feil, S., Suckow, V., Schweiger, S. and Berger, W. (
Le Goff, M.M., Hindson, V.J., Jowitt, T.A., Scott, P.G. and Bishop, P.N. (
Scott, P.G., Grossmann, J.G., Dodd, C.M., Sheehan, J.K. and Bishop, P.N. (
Maurer-Stroh, S., Gouda, M., Novatchkova, M., Schleiffer, A., Schneider, G., Sirota, F.L., Wildpaner, M., Hayashi, N. and Eisenhaber, F. (
Wang, J., Shen, F., Yan, W., Wu, M. and Ratnam, M. (
Eisenhaber, B., Maurer-Stroh, S., Novatchkova, M., Schneider, G. and Eisenhaber, F. (
Vourc'h, P., Moreau, T., Arbion, F., Marouillat-Vedrine, S., Muh, J.P. and Andres, C. (
Khan, N.W., Kondo, M., Hiriyanna, K.T., Jamison, J.A., Bush, R.A., Sieving, P.A. (
Dryja, T.P., McGee, T.L., Berson, E.L., Fishman, G.A., Sandberg, M.A., Alexander, K.R., Derlacki, D.J., Rajagopalan, A.S. (
Wilson, R., Allen, A.J., Oliver, J., Brookman, J.L., High, S. and Bulleid, N.J. (