N-acetylation and phosphorylation of Sec complex subunits in the ER membrane
© Soromani et al.; licensee BioMed Central Ltd. 2012
Received: 24 September 2012
Accepted: 11 December 2012
Published: 13 December 2012
Skip to main content
© Soromani et al.; licensee BioMed Central Ltd. 2012
Received: 24 September 2012
Accepted: 11 December 2012
Published: 13 December 2012
Covalent modifications of proteins provide a mechanism to control protein function. Here, we have investigated modifications of the heptameric Sec complex which is responsible for post-translational protein import into the endoplasmic reticulum (ER). It consists of the Sec61 complex (Sec61p, Sbh1p, Sss1p) which on its own mediates cotranslational protein import into the ER and the Sec63 complex (Sec63p, Sec62p, Sec71p, Sec72p). Little is known about the biogenesis and regulation of individual Sec complex subunits.
We show that Sbh1p when it is part of the Sec61 complex is phosphorylated on T5 which is flanked by proline residues. The phosphorylation site is conserved in mammalian Sec61ß, but only partially in birds, and not in other vertebrates or unicellular eukaryotes, suggesting convergent evolution. Mutation of T5 to A did not affect the ability of mutant Sbh1p to complement the growth defect in a Δsbh1Δsbh2 strain, and did not result in a hypophosphorylated protein which shows that alternate sites can be used by the T5 kinase. A survey of yeast phosphoproteome data shows that Sbh1p can be phosphorylated on multiple sites which are organized in two patches, one at the N-terminus of its cytosolic domain, the other proximal to the transmembrane domain. Surprisingly, although N-acetylation has been shown to interfere with ER targeting, we found that both Sbh1p and Sec62p are cotranslationally N-acetylated by NatA, and N-acetyl-proteome data indicate that Sec61p is modified by the same enzyme. Mutation of the N-acetylation site, however, did not affect Sec62p function in posttranslational protein import into the ER. Disabling NatA resulted in growth retardation, but not in co- or posttranslational translocation defects or instability of Sec62p or Sbh1p.
We conclude that N-acetylation of transmembrane and tail-anchored proteins does not interfere with their ER-targeting, and that Sbh1p phosphorylation on T5, which is not present in Sbh2p, plays a non-essential role specific to the Sec61 complex.
Protein phosphorylation is a reversible mechanism used in all kingdoms of life to regulate protein activity, location and stability . Protein N-acetylation which is irreversible can regulate protein stability and protein-protein interactions [2, 3]. Many proteins (50% in yeast) are N-acetylated, the enzymes responsible for N-acetylation have been identified, and their substrate specificities characterized . The role of N-acetylation, however, remains unclear for the majority of substrates to date. Strikingly, proteins bearing N-terminal signal sequences are usually not N-acetylated . If the modification is introduced by mutation, N-acetylation leads to missorting of secretory proteins to the cytosol. These observations led to the conclusion that N-acetylation interferes with ER targeting .
Although protein flux across the ER membrane can be extremely variable, nothing is known about the regulation of the activity of the protein translocation channel in the ER membrane. In yeast the channel is composed of 3 subunits, Sec61p, Sbh1p and Sss1p, which form the Sec61 complex responsible for cotranslational protein import into the ER . The channel subunits are highly conserved with mammalian proteins Sec61α, Sec61ß and Sec61γ. In yeast, posttranslational import into the ER of proteins with less hydrophobic signal sequences is mediated by a heptameric complex which in addition to the Sec61 complex contains the Sec63 complex (Sec63p, Sec62p, Sec71p, Sec72p) . Yeast also express a homologue of Sec61p, Ssh1p, which together with Sss1p and a homologue of Sbh1p, Sbh2p, forms the Ssh1 complex responsible exclusively for co-translational import into the ER . Protein translocation into the ER and the SEC61, SSS1, SEC63 and SEC62 genes are essential. Deletion of either SBH1 or SBH2 does not affect yeast viability, but deletion of both genes leads to temperature-sensitivity .
Sbh1p and Sbh2p interact with multiple partners, and it is not known how these interactions are regulated. Sbh1p and Sbh2p are small tail-anchored proteins in the ER membrane with largely unstructured cytosolic domains and single α-helical transmembrane domains which on their own can complement the temperature sensitivity of a Δsbh1Δsbh2 strain [8, 9]. The cytosolic domain of Sbh2p, however, modulates interactions of its transmembrane domain , and this is likely also the case for the Sbh1p cytosolic domain. Sbh2p is required for efficient transfer of the nascent polypeptide from signal recognition particle (SRP) into the Ssh1 channel, and may signal to the SRP receptor whether the Ssh1 channel is already occupied by a ribosome . In mammalian cells, Sec61ß mediates association of signal peptidase with the Sec61 channel . The cytosolic domain of Sec61ß can bind to the ribosome, and can be crosslinked to nascent secretory proteins within the ribosomal exit tunnel [12, 13], but the principal ribosome receptor in the ER membranen is Sec61p [14, 15]. The cytosolic domain of Sec61ß also serves as GDP-exchange factor for the ß subunit of SRP receptor in the ER membrane . In addition to its function in translocation, SBH1 also interacts both genetically and physically with the exocyst, a protein complex required for fusion of secretory transport vesicles with the plasma membrane . This function is specific to Sbh1p; overexpression of Sbh2p, which is 50% identical to Sbh1p at the amino acid level, does not suppress exocyst mutations . Mammalian Sec61ß also binds to the exocyst . Sbh1p interaction with the exocyst requires its cytosolic domain, but the function of the interaction remains unknown .
Mammalian Sec61ß has been shown to be phosphorylated . In isolated ER membranes phosphorylation is mediated by Protein Kinase C, and phosphorylation of ER-derived microsomes enhances cotranslational protein translocation into these membranes. It was not shown, however, that Sec61ß was the protein whose phosphorylation was responsible for enhanced translocation. Sec61ß was also found to be phosphorylated in intact cells, but the kinase and the site(s) of phosphorylation were not identified. The only other subunit of the Sec complex which is known to be phosphorylated is Sec63p . Phosphorylation enhances association of Sec63p with Sec62p by increasing acidity of the carboxy-terminal region of Sec63p, and is mediated by casein kinase 2 (CK2). Whether Sec63p phosphorylation has a regulatory function and whether mammalian Sec63p is also phosphorylated is unknown so far.
Here we set out to investigate the role of phosphorylation in Sbh1p function. In purified Sec complexes we identified the threonine at position 5 as a phosphorylated residue. Mutation of T5 to alanine, however, did not affect the ability of Sbh1p to complement the temperature-sensitivity of a Δsbh1Δsbh2 strain, and did not result in detectable hypophosphorylation of the cellular pool of the protein. Comparison to multiple phosphoproteome data revealed additional phosphorylation sites in Sbh1p which were modified in several combinations. Analyzing the entire Sec complex by mass spectrometry we were surprised to find that both Sbh1p and Sec62p were N-acetylated although this modification has been shown to inhibit targeting to the ER . Mutation of the N-acetyl acceptor site in Sec62p did not affect post-translational protein import into the ER. A strain deleted for the enzymatically active subunit of the NatA complex, Ard1p (also known as Naa10p), which is responsible for N-acetylation of Sec62p and Sbh1p, showed growth defects that were exacerbated at low and high temperature. The strain, however, displayed no measurable protein translocation defects, and both Sec62p and Sbh1p were stably expressed in the Δard1 strain, suggesting that N-acetylation of these proteins by Nat A is not essential for protein import into the ER.
We then combined the mutations of the individual potential phosphorylation sites in pairs or in a triple alanine mutant. As shown in Figure 2C even the triple alanine mutant still complemented growth of Δsbh1Δsbh2 cells at high temperature. The mutant protein could still be phosphorylated at the ER membrane in vitro to levels comparable with wildtype Sbh1p (Figure 2D, left panel). [32P]Phosphate-labelling of intact yeast cells also revealed that tripleA-Sbh1p was phosphorylated like wildtype Sbh1p, and expressed at the same level (Figure 2D, middle panel). Sec63p, a known phosphoprotein subunit of the Sec complex, served as a control in this experiment (Figure 2D, right panel). We conclude that S21, S38, and S44 are not the phosphate acceptor sites, or not the only phosphate acceptor sites in Sbh1p.
When we used a different prediction programme, Scan ProSite, we were able to identify an additional phosphorylation site in the cytosolic domain of Sbh1p at T33 (Figure 1B, yellow). Mutation of T33 to alanine on its own or in combination with S21A, S38A and S44A, however, did not affect complementation activity of the mutant Sbh1p, nor its phosphorylation level (not shown).
The reduced signal in the Sbh1p Western Blot of the T5A mutant suggested that the T5A mutation destabilizes Sbh1p. This would not necessarily manifest itself in a functional defect, as SBH2 expressed from its own promoter can complement the growth defect of Δsbh1Δsbh2 yeast at 37°C although it is only expressed at about 10% of Sbh1p (Figure 1A, upper right). We performed a cycloheximide chase to investigate T5A Sbh1p stability. Cells were incubated with cycloheximide to prevent new protein synthesis, aliquots were taken at 0, 45 and 90 min, cells lysed and Sbh1p detected by immunoblotting. As shown in Figure 4C both wildtype and T5A Sbh1p were stable for the entire chase period.
Another potential explanation for the reduced signal in T5A Sbh1p Western Blots is that our antibody does not recognize the mutant protein as well as the wildtype. We raised the antibody against the first 18 amino acids of Sbh1p (Figure 4D, bracket). The antibody recognizes both Sbh1p and Sbh2p on blots (Figure 1A, upper right), but we were unable to precipitate phosphate-labelled Sbh2p (Figure 1A, lower right) or unlabelled Sbh2p (not shown) using the antibody, suggesting that it recognizes Sbh1p better than Sbh2p. Within the first 18 amino acids the only region that differs between the two proteins is amino acids 2–5 (Figure 4D). If this region is decisive for interaction with the antibody, mutation of T5 to A may result in the reduced signal that we see in immunoprecipitations of the phosphorylated mutant protein and since the blots shown in Figure 4B are on the immunoprecipitated material, this would also explain the reduced signal for T5A Sbh1p.
We found that Sbh1p was phosphorylated on T5 (Figure 3). While SBH1 homologues are present in all eukaryotes, the cytosolic domain of the protein is not highly conserved and there has been speculation that it might serve different functions in different organisms . Upon sequence comparison, we found that T5 is conserved in mammals and S. cerevisiae, but not strictly in birds, and not at all in other vertebrates, flies, nematodes, plants, or other yeasts (Figure 5B). In all these organisms, however, there are phosphorylatable serine residues in close proximity, so a position close to the N-terminus could potentially be phosphorylated in all Sbh1p orthologues. Strikingly, the proline in position 4 is conserved yeast, all vertebrates, and even in Drosophila, whereas the proline in position 6, which is part of the recognition sequence for proline-directed kinases, is only present in the species that also have a phosphorylatable residues in position 5 (Figure 5B) .
N-acetylation frequently enhances protein stability . We asked whether Sec62p stability was compromised in the Δard1 strain performing a cycloheximide chase experiment. Samples were taken up to 135 min, cells lysed, and Sec62p detected by immunoblotting. We found that Sec62p was stable in the Δard1 and the wildtype strains for up to 135 min (Figure 7C). In contrast, mutant Sec62p in the sec62-1 strain was unstable and rapidly turned over (Figure 7C). Expression levels of Sbh1p in the Δard1 strain were also similar to wildtype (not shown). We conclude that the growth defect in the Δard1 strain is not due to protein instability of Sec62p or Sbh1p or a protein translocation defect into the ER.
Although N-acetylation is a frequent modification (50% of proteins in yeast) its function is not well understood so far . For soluble secretory proteins it had been shown that N-acetylation interferes with their targeting to the ER . In Sec complexes purified using an HA-tagged Sss1p, however, we found that in both Sec62p and Sbh1p the N-terminal methionine had been cleaved and the following serine N-acetylated (Figure 6A). In a genome-wide identification of N-acetylated proteins, Sec61p was also found to be modified in the same fashion (processing of the methionine, acetylation of S2) . We likely missed the N-acetylated N-terminus of Sec61p because the trypsin digest prior to mass spectromety resulted in a very short N-terminal peptide. Effects of N-acetylation on protein stability, and on protein-protein interactions have been reported [2, 3]. Mutation of the N-acetyl acceptor serine in position 2 in Sec62p had no effect on its function in protein translocation into the ER, however, and deletion of the enzymatically active subunit of the NatA complex, Δard1, did not affect Sec62p stability over a 135 min chase period (Figure 6B, C; Figure 7C). Mutation of S2 to tyrosine in Sec61p had no effect on protein stability at 30°C, but resulted in tunicamycin-sensitivity and temperature sensitivity at 37°C, without growth defects at lower temperatures (not shown). Sbh1p was also stably expressed in the Δard1 strain (not shown), and the Δard1 strain showed no gross defects in protein translocation into the ER either co- or posttranslationally (Figure 7B). Our data suggest that N-acetylation of Sec62p and Sbh1p is not important for protein stability, and not essential for their function in protein import into the ER. N-acetylation of Sec61p affects its function at high temperature (not shown), and will be investigated in more detail in the future. But since our sec61-S2Y mutant was temperature-sensitive only at 37°C, whereas the Δard1 strain had growth defects at all temperatures tested (Figure 7A), it is likely that defects in other NatA substrates contribute to this Δard1 phenotype.
As three transmembrane subunits of the Sec complex in the ER are N-acetylated, this argues against a general interference of this modification with protein targeting to the ER. In fact, Forte and colleagues  had only demonstrated an interference of N-acetylation with posttranslational targeting of soluble proteins to the yeast ER. For cotranslational targeting, the authors had found that these proteins, even if they contained an N-terminal sequence compatible with N-acetylation, were usually not modified . The authors argued that competition of NatA and SRP for their ribosomal binding site would lead to presence of only one or the other near the exit tunnel of the large ribosomal subunit and thus in the presence of SRP NatA-mediated N-acetylation could not occur . Here we find that two cotranslationally ER-targeted transmembrane proteins, Sec61p and Sec62p, are N-acetylated on a NatA consensus sequence (Figure 6 and ), suggesting that the binding of NatA and SRP to the nascent protein are limiting, not their binding to the ribosome. In soluble secretory proteins with N-terminal signal sequences SRP and NatA compete with each other for binding to the the nascent protein N-terminus. In transmembrane proteins like the Sec complex subunits the binding sites for NatA and SRP in the nascent protein are physically separated from each other, so both can bind. We also found the tail-anchored protein Sbh1p N-acetylated which shows that N-acetylation does not interfere with posttranslational ER-targeting by the GET machinery. Taken together our data and those in Forte et al.  suggest that N-acetylation only interferes with posttranslational ER targeting via the Sec63 complex.
Sbh1p is a small tail-anchored protein which interacts with multiple partners (see Introduction). It is the only subunit of the Sec61 channel that is non-essential, so it likely enhances speed of translocation or efficiency of targeting rather than forming a part of the channel proper. This view is supported by the crystal structure of the archaeal SecYEG complex, where SecG is associated with the periphery of the channel . Sbh1p also is the only subunit of the Sec61 complex that has functions outside the protein translocation channel . Protein-protein interactions can be regulated by covalent modifications such as N-acetylation and phosphorylation [1, 3]. Phosphorylation in contrast to N-acetylation is reversible and therefore allows flexible regulation of specific interactions depending on specific circumstances.
The cytosolic domain of Sbh1p contains several predicted phosphorylation sites, and its mammalian orthologue, Sec61ß, had been shown to be phosphorylated, but the modified sites had not been identified, nor had the function of Sbh1p phosphorylation been investigated in detail . The authors had demonstrated that in vitro phosphorylation of microsomal membranes enhances protein translocation into the ER, but since they had found three proteins involved in protein translocation phosphorylated (docking protein α, TRAP, and Sec61ß), the contribution of Sec61ß phosphorylation to the translocation enhancement remained unclear. When we mutated the predicted phosphorylation sites in Sbh1p individually or up to a quadruple combination (S21, T33, S38, S44) to alanine, we did not observe hypophosphorylation of Sbh1p, and even the quadruple mutant was still able to complement the temperature-sensitivity of a Δsbh1Δsbh2 strain (Figure 2, and not shown). We therefore purified Sbh1p with Sec61 complexes using membranes from an SSS1-HA strain and subjected the purified protein to mass spectrometry analysis. We found that Sbh1p when it is part of the Sec61 complex is phosphorylated on a single site, T5 (Figure 3). Mutation of T5 to alanine, either alone or in combination with the only phosphorylatable residue not covered in our mass spectrometry analysis, S27, did again not result in detectable hypophosphorylation of Sbh1p in yeast cells, or in loss of function in translocation (Figure 4A, B). Mutation of T5 did result in a reduced signal on Western Blots, but because T5 is part of the peptide against which our antibody was raised, it remains unclear whether this reduction in signal is due to decreased interaction with the antibody, or to lower expression of T5A Sbh1p (Figure 4B, D). In a cycloheximide chase experiment, however, T5A Sbh1p was stable (Figure 4C).
Potential & used phosphorylation sites in the cytosolic domain of Sbh1p
Mutant complements Δsbh1Δsbh2
Position phosphorylated by mass spec
✓ ✓ ✓
Phosphorylation of S2 of Sbh1p was detected in one study . This is unusual because S2 at the N-terminus of Sbh1p is also N-acetylated (Figure 3 and Figure 6), and there are no reports that we are aware of of N-termini that have both modifications. One possible explanation is that there are two populations of Sbh1p, one which is phosphorylated at S2, the other N-acetylated at S2. But mutation of S2 to alanine destabilized all Sbh1p dramatically and made it close to undetectable on Western Blots, although the low amount of residual protein was still able to complement the growth defect of Δsbh1Δsbh2 yeast (not shown and Table 1). Alanine is less efficiently N-acetylated than serine, but the fact that Sbh1p is stable in a Δard1 strain (not shown) suggests strongly that it is the lack of phosphorylation that is critical in the S2A mutant . Sbh1p is a tail-anchored protein which is inserted into the ER membrane posttranslationally after dissociation from the ribosome, and the NatA complex is ribosome-associated [31, 32]. Perhaps S2 is N-acetylated during biosynthesis and phosphorylation of the same residue early during biogenesis stabilizes the protein in a specific conformation that improves its interaction with chaperones or the insertion machinery in the ER membrane. If phosphorylation of S2 were important during biogenesis only, the phosphate might be removed once the protein is inserted into the ER membrane which would explain why in most studies S2 was not found phosphorylated (Table 1).
Phosphorylation of T5 of Sbh1p was detected in two phosphoproteome studies by mass spectrometry, but in both cases other sites were also found to be modified [28, 29]. The principal difference between these studies and ours is that in the phosphoproteome analyses total cellular Sbh1p was detected, so when multiple sites were found phosphorylated it is not clear whether these were in the same molecule or in different populations of Sbh1p. In our work, we identified T5P Sbh1p in Sec61 complexes that were either free or associated with the Sec63 complex (Figure 3). We have no information on the phosphorylation status of Sbh1p in ribosome-associated Sec61 complexes, of exocyst-associated and of free Sbh1p. Comparison of our own and the phosphoproteome data suggest that the phosphorylation state of and sites used in Sbh1p may differ depending on its interaction partners, and/or on growth conditions.
The kinases modifying Sbh1p may include proline directed kinases like Cdc28 for T5, and perhaps casein kinase 2 (CK2) for the membrane proximal residues S35 and S38 . The latter do not fulfil the consensus sequence for CK2 modification, but CK2 has been shown to phosphorylate other residues in non-consensus contexts and CK2 modifies Sec63p, which would bring it into close proximity with Sbh1p in the Sec complex [1, 20]. In our mutagenesis studies, we never saw hypophosphorylation attributable to mutation of a specific residue, even when we mutated all available serines and threonines in the cytosolic domain of Sbh1p individually, or when we mutated several sites in combination (Figure 2, Figure 4; Table 1). Kinases in yeast tend to be promiscuous, however, and if their actual target residue is missing they can phosphorylate other serine or threonine residues in the vicinity . Promiscuous phosphorylation may therefore be part of the explanation for the lack of hypophosphorylation in the T5A mutant.
Mammalian Sec61ß (mouse and human) has also been found in several phosphoproteome analyses to be modified on multiple sites, but the pattern of phosphorylation of Sec61ß was different from Sbh1p (all residues highlighted in the first line of Figure 1B, with the exception of S43 & T48; ProSite). Only some these these sites are conserved in yeast including T5 (yeast numbering: T5, S21, T33), and of these only T5 was found phosphorylated in yeast (Table 1). When Gruss and colleagues investigated the phosphorylation of mammalian Sec61ß, they did not identify the modified residues, but found after tryptic digest and phosphopeptide mapping that a single peptide was predominantly phosphorylated (O. Gruss, personal communication). From the distribution of trypsin cleavage sites and phosphorylated residues this peptide is almost certainly identical to the N-terminal 15 amino acids of Sec61ß which includes T5 and two serine residues which were found to be phosphorylated in phospho-proteome studies (Figure 1B; ProSite).
Phosphorylation of the cellular pool of Sbh1p was dynamic (Figure 5A) in contrast to phosphorylation of Sec63p (Figure 5A), thus phosphorylation of Sbh1p has the potential to regulate its interactions with its various partners. Phosphorylation at specific sites may either enhance a specific interaction or prevent it . The phosphate that we found on T5 may prevent binding of the Sbh1p cytosolic domain to the ribosome when the Sec61 complex is associated with the Sec63 complex. Based on crosslinking experiments to nascent chains the N-terminus of Sbh1p can reach into the peptide exit tunnel of the large ribosomal subunit . Phosporylation of T5, which adds bulk and two negative charges to a position fixed by two flanking prolines, will almost certainly reduce access of the Sbh1p N-terminus to the peptide exit tunnel and may at the same time reduce affinity of Sbh1p and the Sec61 complex for the ribosome. Kinases frequently are dependent on each other , so phosphorylation of T5 may enhance phosphorylation of additional sites in Sbh1p when it has dissociated from the Sec63 complex which may signal to the SRP receptor that this translocon is unoccupied and available to receive a new nascent chain (similar to Sbh2p in ). Interaction with the SRP receptor may then trigger dephosphorylation of T5 and allow Sbh1p to bind to the ribosome again. Other scenarios in which T5 phosphorylation regulates interaction of Sbh1p with the Sec63 complex, promotes interaction of Sbh1p with the Sec61 complex, or prevents interaction with the exocyst are also plausible. In Sbh2p, which forms a strictly cotranslational translocon with the Sec61p homologue Ssh1p and does not interact with the exocyst, the threonine in position 5 is not conserved (Figure 4D). T5 phosphorylation is therefore likely to have an effect on interactions that are specific to Sbh1p, i.e. interaction with Sec61p, with the Sec63 complex or with the exocyst.
The N-terminal cytosolic domain of Sbh1p is predicted to be largely unstructured and indeed in the crystal structure of the archaeal channel the SecG cytosolic domain is not visible . The Sbh1p cytosolic domain also contains a relatively large number of basic amino acids most of which are conserved between yeast and dog (Figure 1B). One conceivable effect of T5 phosphorylation is a folding back of the Sbh1p N-terminus and interaction of the negatively charged phospho-T5 with one of the basic patches more proximal to the membrane (amino acids 15–17 KRK, or 30/31 KK). This phosphorylation-induced structure may then promote or prevent interaction with known Sbh1p binding partners, and/or with additional kinases. The prolines flanking T5 may also be subject to cis-trans isomerization by phosphate-directed prolyl isomerases like Pin1p which may help the phosphorylated Sbh1p N-terminus to acquire its functionally relevant structure .
The phosphorylation site at T5 almost certainly has an interesting role, as both the site and its flanking prolines are conserved between S. cerevisiae and mammals (Figure 1B), but not Xenopus, fish, invertebrate metazoa, or other yeasts (Figure 5B). Kinases evolved largely after the development of metazoa, so phosphorylation sites, kinases, and the actual use of the phosphorylation sites are generally poorly conserved between yeast and mammals [1, 34, 35]. In two birds, chicken (Gallus gallus) and zebrafinch (Taeriopygia guttata), the residues surrounding the T5 phospho-acceptor site are conserved (Figure 5B) and in chicken, T5 is replaced by a phosphorylatable serine, whereas in zebrafinch T5 is replaced by asparagine (Figure 5B). Altogether the data suggest that the N-terminus of Sec61ß proteins is evolving rapidly, and the fact that the PTP motif occurs in yeast and again in some birds and all mammals suggests a case of convergent evolution.
As three subunits of the ER-resident Sec complex are N-acetylated, the modification per se is unlikely to interfere with targeting to the ER. That N-acetylation can occur on two cotranslationally targeted transmembrane proteins, Sec61p and Sec62p, but not on cotranslationally targeted soluble proteins , suggests that in signal-sequence bearing proteins binding of NatA and SRP to the nascent protein are limiting, not their binding to the ribosome. Since we also found the tail-anchored protein Sbh1p N-acetylated, our data and those in Forte et al.  demonstrate that N-acetylation exclusively interferes with posttranslational ER targeting via the Sec63 complex.
Our phosphosite analysis of Sbh1p suggests that several Sbh1p populations exist in the cell which are phosphorylated at different sites in different combinations. The phosphate on T5 of Sbh1p must play a particularly important Sbh1p-specific role since the site evolved twice independently, in S. cerevisiae and in mammals, and is not present in Sbh2p. Phosphorylation is a reversible modification known to affect protein-protein interactions, thus it is likely that the phosphorylation status of an individual Sbh1p molecule determines its interaction partners, and vice versa. In the future we will therefore investigate the phosphorylation status of exocyst- and ribosome-associated Sbh1p by mass spectrometry, and subsequently make the readout for functional analysis of specific phospho-site mutations specific to the complex in which they were found to be modified.
NY179 (SBH1 SBH2 MAT a leu2-3,112 ura3-52), H3223 (MAT a KanMx::sbh1 leu2-3,112 ura3-52 GAL + ), H3203 (MAT a HphMx::sbh2 leu2-3,112 ura3-52 GAL + ), H3231 (MAT a KanMx::sbh1 HphMx::sbh2 leu2-3,112 ura3-52 GAL + ) were gifts from Jussi Jäntti and used to characterize Sbh1p phosphorylation sites [7, 8]. KRY739 (SSS1-HA::TRP1 ssh1Δ::ADE2 + ) was a gift from Kai-Uwe Kalies and used to purify Sec complexes. BY4742 (MAT a his3-1 leu2 lys2 ura3 can1-100), BY17299 (Δdoa10::KanMX6 in BY4742), and BY10976 (Δard1::KanMX6 in BY4742) were from the Euroscarf collection and a gift from Manfred Schmitt. RSY1132 (MAT a leu2,3-112 trp1-1Δ ura3-52 sec61-3) , RSY1294 (MAT a leu2,3-112 can1-100 leu2-3,112 his3-11,15 trp1-1 ura3-1 ade2-1 psec61-32) , RSY529 (MAT a leu2,3-112 his4-619 ura3-52 sec62-1)  were gifts from Randy Schekman. KRY712 (BMA38a MAT a his3-Δ200 leu2-3.112 ura3-1 trp1-Δ1 ade2-1 can1-100 [pCEN-LEU2 sec61-302) was a control for the translocation reporter constructs .
SBH1 cloned into pCR2.1-TOPO plasmid was kindly provided by Jussi Jäntti (Helsinki University, Finland). The QuickChange Site-Directed Mutagenesis Kit (Stratagene, UK) was used to introduce single or multiple base mutations in SBH1. Substitutions were verified by sequencing. Mutated sbh1 was excised from pCR2.1-TOPO plasmids using EcoRV/BamHI, and subcloned into pRS305, pRS415 and pRS424. The second amino acid of Sec62p was changed from serine to tyrosine as above in SEC62 in pUC19. After verifying the sequence mutant sec62-S2Y was subcloned using SacI/HindIII into pRS315. Plasmids with PHO8-URA3 and PRC1-URA3 are described in .
Microsomes were prepared from SBH1/2 wildtype or mutant cells as in . Labelling reactions contained 5eq membranes in B88 with 0.1 μM okadaic acid and calyculin A (Calbiochem, UK), 5 μM GTP, 2 mM CaCl2, 40 μCi γ-32P]ATP (Amersham Biosciences, UK). Reactions were incubated at 30°C for 30 min and okadaic acid and calyculin A (500 nM) in B88 were added prior to sedimentation at 14,100 x g at 4°C for 5 min. Membranes were lysed in 50 μl 2% SDS, incubated at 65°C for 10 min, Sbh proteins immunoprecipitated with anti-Sbh1p serum raised by us against the first 18 amino acids of Sbh1p. Samples were analyzed by SDS-PAGE on 15% gels and autoradiography.
Yeast grown in YNBD –Leu media to OD600 = 1 were pelleted, washed once with YNBD –Leu –PO4 media, and resuspended in phosphate free medium to OD600 = 1.5. Cells were labelled with [32P]PO4 (75 μCi per sample, Amersham Biosciences UK) for 1 Â½ hr at 30°C with gentle agitation. Equal volumes of TCA (25%) were used to precipitate cells for 20 min on ice; pellets were washed twice in ice-cold acetone before air-drying. Cells were lysed with 100 μl lysis buffer (50 mM Tri-Cl pH 7.5, 1 mM EDTA, 1% SDS, 6 M urea) and an equal volume of glass beads (acid washed, Sigma) by two cycles of 1 min vortexing and 1 and 10 min heating at 65°C respectively. The lysate was diluted in IP buffer containing 100 nM okadaic acid and calyculin A and 1 mM AEBSF (Calbiochem UK). Cell debris were sedimented by centrifugation prior to immunoprecipitation of Sbh1p and Sec63p. For pulse-chase experiments cells were labelled for 10 min as above, the samples (duplicates) were washed in phosphate-free media (YNBD –Leu –PO4) and resuspended to the original OD600 (1.5) with YNBD –Leu media and chased for 5, 15, 30, 60 min.
For Sbh1p cells grown in YNBD -Leu to OD600 = 1.5 were washed in YNB –Leu, resuspended in that medium to the same OD and incubated at 30°C for 30 min in a shaking incubator. Cycloheximide (Roche UK) was added to 50 μg/ml  and the cells incubated for up to 90 min. The reaction was stopped by addition of equal volumes of cold TCA (25%) for 20 min. Cells were lysed as above and Sbh1p detected by immunoblotting. For Sec62p cells were grown in YPD to OD600 = 1.0, cycloheximide added to 200 μg/ml, samples taken at the indicated times, and cells lysed by bead-beating. After gel electrophoresis, Sec62p was detected by immunoblotting.
7,500eq of microsomal membranes from the SSS1-HA strain in B88 were and lysed in 2.5% digitonin buffer [50 mM HEPES-KOH pH: 7.4, 400 mM potassium acetate, 8 mM magnesium acetate, 100 nM okadaic acid & calyculin A (Calbiochem UK),1 x protease inhibitor cocktail (Roche UK), 10% glycerol, 2 mM DTT, 2.5% digitonin (high purity, Calbiochem UK)] for 30 min at 4°C with gentle agitation. Solubilised membranes were sedimented (70,000 rpm, TLA-100.3 Beckman rotor) for 1 h at 4°C; the supernatant was diluted with the same buffer without digitonin, to 1% digitonin. Monoclonal anti-HA agarose conjugate (Sigma UK) was washed 3x with TBS and twice with 10 ml digitonin buffer, then lysate was added for overnight binding at 4°C with rotation. The resin was washed twice with 10 ml of solubilization buffer with 1% digitonin not containing glycerol, 3x with 10 ml TBS at 4°C. Bound proteins were eluted at room temperature with 4 ml of 200 mM Glycine-HCl pH: 2.5, TCA-precipitated and analysed by 12% SDS-PAGE and Coomassie Brilliant Blue (BioRad UK) staining. The band corresponding to Sbh1p was excised and stored in 10% Methanol before trypsin-digest and mass spectrometry analysis. Purification for detection of N-acetylation was done as above, but without phosphatase inhibitors. Bands were excised, digested with trypsin overnight, and resulting peptides analyzed by MALDI-TOF/TOF mass spectrometry on a MALDI 4800 TOF/TOF analyzer (Applied Biosystems). Tryptic peptide mass fingerprints were measured in positive reflector mode with subsequent MSMS fragmentation. Combined MS & MSMS data were identified using the NCBInr protein data base.
We thank Jussi Jäntti, Kai Kalies, and Manfred Schmitt for strains, Randy Schekman for antibodies against Sec63p, Sec62p, and Sss1p, and Manfred Schmitt, Jeff Brodsky, Schiampietro Schiavo, and especially Gert-Wieland Kohring for carefully reading the manuscript. Special thanks to Carol Robinson and Sarah Maslen (Chemistry Department, Cambridge) for identifying phosphorylated T5 in Sbh1p by mass spectrometry, and to Thomas Arnesen, Sabine Rospert, David Owen, Oliver Gruss, Chris Cheng de Vries, and Arend Sidow for helpful discussions. The phosphorylation work was funded by a Wellcome Senior Fellowship to K.R. (042216) and an MRC studentship to C.S. The N-acetylation work was funded by the Saarland University. K.R. thanks the teachers at her son's kindergarden for helping much beyond their call of duty to give her time to write this paper.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.