Delineation of the Plasma Membrane Targeting Domain of the X-Linked Retinitis Pigmentosa Protein RP2

Abstract

purpose. The X-linked retinitis pigmentosa protein RP2 is predominantly targeted to the plasma membrane. This study delineates the exact amino acid sequence requirements for targeting of RP2 through dual N-terminal acyl modification.

methods. Inhibition of acyl modification with a palmitate analogue was used to confirm the mechanism of intracellular targeting. Mutagenesis of the first 15 residues in a synthetic RP2-green fluorescent protein (GFP) chimera was used to probe the precise requirements for plasma membrane targeting in Chinese hamster ovary (CHO) cells by confocal microscopy and subcellular fractionation.

results. The N-terminal Met-Gly-Cys-X-Phe-Ser-Lys motif of human RP2 is necessary and sufficient for the protein’s plasma membrane localization. This motif includes the accepted consensus sequence for N-myristoyl transferase (NMT) and a site for attachment of a palmitoyl moiety. An interesting feature of the motif is an essential phenylalanine at position 5. This is the first report of the requirement of a specific residue at position 5 within the N-terminal acyl modification motif for normal intracellular targeting. Arginine at position 8 is not essential for plasma membrane localization of the protein, but it improves targeting. The motif is highly conserved and is found in all vertebrate orthologues of human RP2, except mouse. In mouse, however, the Ser6Thr change is concordant with the accepted NMT consensus sequence.

conclusions. Conserved residues mediate the intracellular targeting of RP2, further highlighting the potential significance of the protein’s plasma membrane localization. The delineation of this motif identifies residues in which mutations disrupt the dual acylation of RP2 and almost certainly result in disease.

Retinitis pigmentosa (RP) is the most common form of inherited blindness. The disorder is characterized by night blindness and a loss of peripheral vision, which is followed by a loss of central vision. X-linked RP (XLRP) is a severe form of this retinal degeneration that accounts for 3% to 20% of all familial cases. ^{1 2 3} Mutations in the RP2 gene have been shown to account for between 15% and 20% of XLRP. ⁴ The ubiquitously expressed 350-amino-acid RP2 protein has similarity (30.4% identity over 151 amino acids) to the tubulin-specific chaperone cofactor C. ^{4 5}  

Cofactor C functions in assembling native tubulin heterodimers with other cofactors (A–E). The release of tubulin from cofactor complexes occurs with the hydrolysis of guanosine triphosphate (GTP) by the bound tubulin. ⁶ Cofactors C, D, and E act together as a guanosine triphosphatase (GTPase)–activating protein (GAP) for tubulin, stimulating the hydrolysis of tubulin-GTP to tubulin-guanosine diphosphate (GDP). ^{7 8} Pathogenic amino-acid substitutions in RP2, at residues conserved with cofactor C, suggest a functional homology between the proteins ^{4 9 10} ; however, the function of RP2 is currently unknown. 

We have demonstrated that the RP2 protein has a predominantly plasma membrane localization and have identified sites for myristoylation and palmitoylation. ⁵ Lipid modifications are an important mechanism for targeting proteins to cellular membranes and can act as signals that sort proteins to domains within the plasma membranes of some cells. ¹¹ Myristate (myr) is cotranslationally attached through an amide bond to the N-terminal glycine residue of the protein by an N-myristoyl transferase (NMT) after cleavage of the initiating methionine. ^{12 13 14} The consensus sequence for NMT protein substrates is Met-Gly-X-X-X-Ser/Thr-. The glycine at position 2 is essential, serine or threonine is preferred at position 6, and lysine or arginine is preferred at position 7 and/or 8. ^{12 13 14} Myristoylation of the N-terminal glycine residue of a protein facilitates the attachment of the fatty acid palmitate through a thioester linkage to adjacent cysteine residues, normally at position 3. It has been suggested that proteins containing myr-Gly-Cys transiently interact with diverse intracellular membranes until they are retained in the plasma membrane after palmitoylation by a plasma membrane bound palmitoyl acyl transferase (PAT). ¹⁵ For example, myristoylation of N-acylated green fluorescent protein (GFP) is sufficient to exclude it from the nucleus and induce association with intracellular membranes, but plasma membrane localization requires a second signal: either palmitoylation or a polybasic domain. ¹⁶ The targeting of RP2 to the plasma membrane appears to be dependent on dual acylation, because glycine 2 is necessary for membrane association, and cysteine 3 is necessary for plasma membrane localization. ⁵  

The pathogenic mutation ΔS6 in RP2 ^{4 17} disrupts the acyl-mediated targeting of RP2 to the plasma membrane, suggesting that membrane localization is essential for the protein’s function in the retina. ⁵ In this study, we delineated the precise plasma membrane targeting requirements for RP2. These data demonstrate that the cellular targeting of the protein is conserved, supporting the hypothesis that RP2 plasma membrane localization is essential to its function in vertebrates. 

Methods

Treatment of Cells with 2-Bromopalmitate

The neuroblastoma cell line SH-SY5Y was maintained in Dulbecco’s modified Eagle’s medium-F12 with 10% fetal calf serum. Cells were seeded at 1 × 10⁵ cells/mL onto chamber slides. After 24 hours, cells were transferred to media containing 2.5% fetal calf serum, with or without 50 μM 2-bromopalmitate (Aldrich, Poole, UK). After a further 16 hours, cells were fixed in ice-cold methanol and processed for immunofluorescence with affinity-purified antisera S974, as we have described previously. ⁵ Fluorescence was detected with a laser scanning confocal microscope (Carl Zeiss, Oberkochen, Germany). 

Database Searching and Sequence Alignment

Using a combination of bioinformatic applications for data mining, the predicted amino acid sequences corresponding to RP2 in a variety of species were identified. NIX and PIX analysis identified RP2 orthologues (provided in the public domain by the Human Genome Mapping Project Resource Center, Cambridge UK and available at http://www.hgmp.mrc.ac.uk/) and Seqman II (DNAStar, Inc, Madison, WI) was used to construct a consensus cDNA sequence. Protein sequences were then aligned using the Clustal algorithm with Megalign software (DNAStar). 

Construction of Plasmids

Full length RP2 cDNA was amplified by PCR from a human brain cDNA library (Clontech, Palo Alto, CA) and cloned into the pGEM-T vector, using oligonucleotide primers based on the untranslated region of the mRNA. Restriction endonuclease recognition sites appropriate for subcloning were introduced by using modified primers for PCR from pGEMT RP2 and were subsequently used to clone RP2 into the BamHI-AgeI site of pEGFP-N1. Cloning into the BamHI-AgeI site of pEGFP-N1 also generated the N-terminal amino acid RP2-GFP chimeras. Oligonucleotides with a BamHI site in the 5′ end and an AgeI site at the 3′ end were annealed, phosphorylated, and cloned into the GFP vector. This strategy was also used to introduce mutations at the N terminus of RP2. 

Confocal Microscopy and Subcellular Fractionation of Cells Expressing RP2-GFP Chimeras

Chinese hamster ovary (CHO) cells were maintained under the same conditions as the SH-SY5Y cells. Transfections were performed 24 hours after seeding using a eukaryotic cell transformation agent (Lipofectamine Plus; Life Technologies, Paisley, UK) according to the manufacturer’s instructions. After a further 48 hours, cells were fixed with 3.7% (vol/vol) formaldehyde and fluorescence imaged as described for the SH-SY5Y cells. For subcellular fractionation, cells were seeded into T25 flasks before transfection. After 48 hours, cells were washed twice and scraped into phosphate-buffered saline (PBS) containing a protease inhibitor cocktail (Aldrich). Cells were then broken by passing them 16 times through a ball-bearing cell homogenizer with a clearance of 12 μm (HGM, Heidelberg, Germany). After cell breakage, homogenates were centrifuged at 12,000g for 20 minutes. Pellets were then resuspended in PBS, and aliquots of pellet and supernatant prepared for Western blot analysis. This procedure efficiently fractionates most of the membranes into the pellet and cytosol into the supernatant, as has been confirmed by the fractionation of a range of marker proteins. Affinity-purified anti-RP2 was used at 1:500 and the anti-GFP (Clontech) was used at the manufacturer’s recommended titer. Immune complexes were visualized by enhanced chemiluminescence detection. 

Results

Effects of 2-Bromopalmitate on RP2 Plasma Membrane Localization

The palmitate analogue 2-bromopalmitate has been reported to block both the myristoylation and palmitoylation of proteins, inhibiting membrane binding and correct cellular localization. ¹⁸ Because our previous data had indicated that the mechanism of RP2 membrane targeting is through dual acylation, we investigated the effects of 2-bromopalmitate on the plasma membrane localization of RP2. SH-SY5Y cells were cultured with 2-bromoplamitate and the localization of RP2 observed by immunofluorescent staining with S974, followed by confocal microscopy (Fig. 1) . Cells treated with 2-bromopalmitate showed redistribution of RP2 away from the plasma membrane to intracellular organelles and the cytoplasm. 

Sequence Alignment of RP2 Protein Orthologues

Because human RP2 is targeted to the plasma membrane through dual acylation, we examined orthologues of the protein for dual acylation motifs to assess the likely functional importance of RP2 subcellular localization in a range of animals. Several vertebrate RP2 orthologues were identified by using a combination of bioinformatic applications. Sequence alignments of the N-terminal domains of these proteins showed that they all contain a putative dual N-acylation domain (Fig. 2) . All the RP2 orthologues contain a glycine residue at amino acid position 2, which is the site of myristic acid attachment and a cysteine at amino acid position 3, the site of palmitic acid attachment. The preferred consensus sequence for NMT protein substrates is a serine or threonine at position 6 and a lysine or arginine at position 7 and/or 8. Human RP2 has a serine at position 6, a lysine at position 7, and an arginine at position 8. It is interesting that the mouse RP2 sequence has threonine at position 6, whereas all the other proteins have a serine. The mouse protein also differs at residue 4 where it has a cysteine, whereas the other RP2 orthologues have a phenylalanine. The phenylalanine residue at position 5 occurs in all the RP2 proteins, suggesting this may also represent an important residue for the plasma membrane localization of RP2. 

Subcellular Localization of RP2 N-Acylation Domain Mutants

To delineate the precise amino acid sequence requirements for RP2 plasma membrane targeting we examined the localization of human RP2-GFP chimeric proteins containing full-length RP2 and the N-terminal domain of the protein. We have demonstrated that the N-terminal 15 amino acids of RP2 are sufficient for plasma membrane localization. ⁵ In this study, we mutated all the residues within this region, as shown in Figure 3 , to elucidate the primary sequence requirements for RP2 targeting. 

Wild-type and mutant RP2 constructs were cloned into the GFP expression vector pEGFP-N1, so that peptides were tagged at the C terminus with GFP. CHO cells were transfected with these constructs and analyzed by confocal fluorescence microscopy. Consistent with our previous results using untagged protein, ⁵ full-length, wild-type RP2-GFP localized predominantly to the plasma membrane (Fig. 4A) . The N-terminal 15 amino acids of wild-type protein also localized mainly to the plasma membrane (Fig. 4A) , demonstrating that this region of the protein is sufficient for efficient targeting to the plasma membrane. As expected, N-terminal RP2-GFP constructs containing mutations at residues where fatty acids are attached in myristoylated and palmitoylated proteins (G2 and C3, respectively) did not localize to the plasma membrane (Fig. 4B) . The G2A and G2A/C3S mutant proteins localized throughout the cell with strong staining in the nucleus. This staining pattern is similar to that of GFP alone. The C3S mutant protein showed a different localization pattern and appeared to be associated with intracellular membranes. A cytoplasmic staining pattern with a strong nuclear component was seen in cells expressing N-terminal RP2-GFP proteins with F5A, S6A substitutions and the clinically occurring ΔS6 mutations (Fig. 4B) , suggesting that these residues are essential for dual acylation of RP2. Within the first six residues, F4 was the only amino acid at which mutagenesis (F4A) appeared to have no effect on the localization of RP2, suggesting that this is not an essential part of the consensus sequence for plasma membrane targeting. 

To test whether residues after the N-terminal six amino acids of RP2 were required for plasma membrane targeting, chimeric proteins were generated in which residues between 7 and 15 were mutated to glycine (Fig. 3) . The first six amino acids of RP2 were not sufficient to localize the fusion protein to the plasma membrane (Fig. 4C) . However, the first seven amino acids of RP2 were sufficient to localize the majority of protein to the plasma membrane (Fig. 4C) , although there was more cytoplasmic staining observed compared with the N-terminal 15-amino-acid construct (Fig. 4A) . Previous reports have suggested that when putatively acylated N-terminal sequences are appended to GFP without a linker, plasma membrane localization can be prevented. ^{19 20} A chimera of the first eight amino acids of RP2 fused directly to GFP unexpectedly also localized predominantly to the plasma membrane (Fig. 4C) . 

Membrane Association of RP2 N-Acylation Domain Mutants

To complement the morphologic data, we analyzed wild-type and mutant RP2 partitioning by subcellular fractionation. CHO cells were transiently transfected with RP2-GFP constructs. After cell breakage and fractionation of cytosolic and membranous fractions by centrifugation, the presence of RP2-GFP proteins in the supernatant and membrane-containing pellet fractions were determined by Western blot analysis with RP2 and GFP antibodies. Consistent with our previous results using untagged protein, ⁵ full-length, wild-type RP2-GFP partitioned to the pellet fraction (Fig. 5A) . The presence of the C-terminal GFP tag did not appear to affect detection of RP2 by affinity-purified sera (S974). The antisera also detected RP2-GFP chimeric proteins containing just the N-terminal 15 amino acids of RP2 (Fig. 5B) . However, chimeric proteins consisting of the first eight residues or less were not detected (Fig. 5B) . This suggests that residues 8 to 15 of RP2 contain an epitope for the polyclonal antisera used in this study. All the RP2-GFP N-terminal mutant proteins were expressed efficiently by transiently transfected CHO cells, although protein levels were variable (Fig. 5B) . The N-terminal 15-amino-acid wild type, C3S, and F4A RP2-GFP proteins were more abundant in pellet than supernatant fractions (Fig. 5C) , suggesting that these proteins are membrane associated. In contrast, the G2A, G2A/C3S, F5A, S6A, and ΔS6 mutant proteins fractionated principally to the supernatant (Fig. 5C) , indicating they are not attached to cellular membranes. Chimeric protein containing the first seven amino acids of RP2 was detected entirely in the pellet fraction by using a GFP antibody, whereas the first six amino acids of RP2 localized in both the pellet and supernatant fraction. This suggests that the first six RP2 amino acids are not sufficient for complete membrane association (Fig. 5D) , when residue 7 is mutated from a lysine to a glycine. The first eight amino acids of RP2 coupled directly to GFP separated in the pellet fraction (Fig. 5D) . GFP alone partitioned predominantly to the supernatant fraction, as would be expected of a cytosolic protein (Fig. 5D) . 

Discussion

To provide further evidence that the plasma membrane targeting of RP2 is dependent on dual N-terminal acylation of the protein, we have demonstrated that RP2 localization can be disrupted by treatment with a palmitate analogue. 2-Bromopalmitate is an inhibitor of protein fatty acylation with some specificity for palmitoylation. ¹⁸ It has been reported to block Fyn fatty acylation in general and palmitoylation in particular. ¹⁸ It has also been shown to inhibit the palmitoylation of the chemokine-HIV receptor CCR5. ²¹  

Sequence alignment of RP2 orthologues from a range of vertebrate species showed that the dual-acyl modification domain is highly conserved. This supports the hypothesis that the plasma membrane localization of the protein is probably necessary for its function in vertebrates. Thus, it is unlikely that RP2 functions exclusively in tubulin folding. It may, however, still interact with tubulin and/or microtubules and could provide a link between membranes and the cytoskeleton, perhaps as part of the cellular protein traffic machinery or a signaling cascade. ²² The recent identification of adenosine diphosphate (ADP) ribosylation factor (ARF)-like proteins and src as interacting partners of RP2 (a function that appears to be conserved with cofactor C and D) supports this hypothesis. ²³  

All the RP2 orthologues identified contained the NMT Met-Gly-X-X-X-Ser/Thr- consensus sequence. Additional residues within the N terminus of the protein were also conserved between species. We have previously shown the signal for plasma membrane localization of RP2 is within the N-terminal 15 amino acids of the protein. ⁵ To pinpoint the amino acid sequence of RP2 responsible for its localization, we mutated all the N-terminal 15-amino-acid residues. These data confirm our previous findings ⁵ that glycine 2, where a myristoyl moiety attaches, and cysteine 3, where a palmitoyl attaches, are essential residues for the plasma membrane localization of RP2. 

The substitution F5A resulted in the unexpected finding that RP2 localized throughout the cell. These data suggest that phenylalanine 5 also represents an essential residue for the plasma membrane targeting of RP2. Because RP2(F5A) is localized throughout the cell and does not appear to be associated with internal membranes, as for the C3S mutation, it would appear that the F5A mutation is disrupting myristoylation of the protein and is not dependent on palmitoylation alone. This is the first report suggesting a residue at position 5 of a myristoylated protein is essential for NMT activity. This finding is particularly surprising, because the F5A substitution represents the exchange of one residue with a nonpolar side chain for another, also alanine at position 5 is found in several other myristoylated proteins. ^{12 13} However, because this residue is conserved in all RP2 protein orthologues, it appears to be a particular requirement of the RP2 intracellular targeting motif. The clinically occurring ΔS6 mutation changes the N-terminal RP2 sequence to Met-Gly-Cys-Phe-Phe-Lys-Arg-Arg-. It has been reported that the majority of (70%), but not all, myristoylated proteins have a serine or threonine at position 6. ¹³ Other residues that have been reported at position 6 do not include lysine; however, myristoylated proteins have been reported with alanine at position 6. ¹³ Our data show that RP2(S6A) does not localize to the plasma membrane, therefore confirming that serine/threonine at position 6 is essential for RP2 myristoylation. The first seven amino acids of RP2 are sufficient to target the majority of the protein to the plasma membrane and contain the reported NMT consensus sequence with the preferred lysine at position 7. ¹² The arginine at position 8, which has also been reported to be a NMT-preferred residue, ¹² was not essential for RP2 plasma membrane localization but increased the amount of RP2 that was targeted to the plasma membrane. 

It has been suggested that RP2 may not be targeted to the plasma membrane in all cell types. A study using full length RP2 tagged at the C terminus with GFP reported the protein was targeted to the plasma membrane in HeLa cells, ²⁴ similar to our previous observations. ⁵ In two human fibroblast cell lines and COS-7 cells, however, the protein was observed to be predominantly in the cytoplasm. ²⁴ It was suggested that such a cell type–dependent plasma membrane targeting of RP2 could be the result of specific trafficking factors in some types of cell. ²⁴ Because the acyl modification machinery is ubiquitously expressed in human tissues, however, we anticipate that RP2 would be acylated in most cell types and that this would lead to its membrane association. Indeed, we have observed the efficient plasma membrane targeting of RP2 in all the diverse cell types we have examined, including COS-7 cells (Chapple JP, Cheetham ME, unpublished observations, 2001). The observed differences in membrane targeting could therefore reflect differences in experimental conditions that may have important implications for the regulation of RP2 localization, as opposed to cell-type differences in NMT or PAT. Future studies of the dynamics of RP2 membrane targeting and the definition of RP2 localization in cells in vivo, particularly in the retina, are essential to better understanding of the role of the protein’s membrane localization. 

In conclusion, the N-terminal Met-Gly-Cys-X-Phe-Ser-Lys motif of RP2 was sufficient for efficient plasma membrane targeting of the protein. The identification of this motif as solely responsible for intracellular targeting of RP2 indicates that noncovalent interactions with other membrane proteins are not playing a role in intracellular localization. The motif is conserved in all the RP2 orthologues that have been identified, except for mouse, where there is a threonine at position 6. However, because the previously published consensus sequence for NMT allows a threonine at position 6 it is highly likely that mouse RP2 is also localized to the plasma membrane. Mutation of any of these residues is likely to disrupt the dual acylation of RP2 and result in disease. The next challenge will be to determine the function of RP2 on the plasma membrane of retina-specific cell types and to study the functional relationship with cofactor C. 

 Supported by the Wellcome Trust, the Biotechnology and Biological Sciences Research Council (BBSRC), and Fight for Sight.

 Submitted for publication December 26, 2001; accepted January 25, 2002.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Effects of 2-bromopalmitate on localization of RP2 in SH-SY5Y cells. Confocal immunofluorescence of control cells and cells treated with 50 μM 2-bromopalmitate for 16 hours. Images are 50 μm².

Figure 2.

 

The N terminus of the RP2 protein is conserved and contains sites for dual N-acylation (bold). The full-length predicted protein sequence of RP2 from Homo sapiens (H.sap) and orthologues from Mus musculus (M.mus), Galus galus (G.galus), Xenopus laevis (X.laev), and Danio rerio (D.rerio) were aligned. The N termini of the aligned proteins is shown.

Figure 3.

 

RP2 wild-type and mutated sequences appended to the amino terminus of GFP used in this study. Mutated residues are shown in bold. Residues are numbered from the N-terminal methionine of RP2 1–15.

Figure 4.

 

Confocal localization of N-terminal RP2-GFP chimeras in CHO cells. (A) Full-length, wild-type RP2-GFP and the N-terminal 15 amino acids of RP2 appended to GFP (N-term RP2-GFP). (B) N-terminal 15-amino-acid RP2-GFP chimeras containing G2A, C3S, G2A/C3S, F4A, F5A, S6A, and ΔS6 mutations. The localization of GFP alone is also shown. (C) RP2-GFP chimeras containing the first seven (1–7) and the first six (1–6) amino acids of RP2 separated from GFP by a 10-glycine linker and the first eight amino acids of RP2 fused directly to GFP (1–8). Forty-eight hours after transfection, cells were formaldehyde fixed and processed for confocal microscopy. All images are 95 μm².

Figure 5.

 

Effects of mutagenesis within the N-acylation domain of RP2 on membrane association. CHO cells were transiently transfected with RP2-GFP chimeras. (A) Western blot analysis showing levels of full-length wild-type RP2-GFP in total cell lysate (T), supernatant (S) and pellet fractions (P), as detected by antisera raised against RP2 and GFP. (B) Western blot showing the expression of the N-terminal 15-amino-acid RP2-GFP chimeras. Also shown is the RP2-GFP chimeras containing the first seven and first six amino acids of RP2, the first eight amino acids of RP2 fused directly to GFP, and GFP alone, as detected by antisera raised against RP2 and GFP. (C) Western blot showing comparison of levels of N-terminal 15-amino-acid RP2-GFP chimeras in supernatant and pellet fractions as detected by anti-RP2. (D) Western blot showing comparison of levels of RP2-GFP chimeras containing the first seven and the first six amino acids of RP2, the first eight amino acids of RP2 fused directly to GFP, and GFP alone in supernatant and pellet fractions, as detected by antisera raised against GFP.