The code for directing proteins for translocation across ER membrane: SRP cotranslationally recognizes specific features of a signal sequence

Experimental design

A site-specific photocrosslinking approach was used to investigate the role of different parts of the signal sequence for interaction with SRP. The technique is based on the tRNA-mediated incorporation of a modified amino acid carrying a crosslinking reagent into ribosome-nascent chain complexes (RNCs) [23, 26, 47]. In the present study, RNCs of both wild-type and mutated preprolactin (pPL) were synthesized using an in vitro translation system. All mRNAs used contained a unique amber-stop codon (UAG) in the codon corresponding to the position 18 of the preprolactin signal sequence (Fig. 1b; Table 1). Premature termination of translation occurs at the position of the amber codon in the absence of an amber-suppressor, however termination is suppressed in the presence of an amber-suppressor tRNA (Fig. 1b). Photoreactive N^ε-(5-azido-2-nitrobenzoyl)-Lys-tRNA^amb (εANB-Lys-tRNA^amb) amber-suppressor Lys-tRNA was used for site-specific labeling of pPL nascent chains.

Table 1.

pPL mutants

Mutant	Sequence (N-terminal part of the preprotein)
	Signal sequence	Mature protein
pPL (WT)	MDSKGSSQKGSRLLLLL#VSNLLLCQGVVS	T-----PVCPNGPGN
n-4L	MLSLGSSQLGSLLLLLL#VSNLLLCQGVVS	T-----PVCPNGPGN
n-3LR	MLSLGSSQLGSRLLLLL#VSNLLLCQGVVS	T-----PVCPNGPGN
n-Δ10,R	MRLLLLL#VSNLLLCQGVVS	T-----PVCPNGPGN
Δ1L	MDSKGSSQKGSRLLLLL#VSN-LLCQGVVS	T-----PVCPNGPGN
Δ2L	MDSKGSSQKGSRLLLLL#VSN--LCQGVVS	T-----PVCPNGPGN
Δ3L	MDSKGSSQKGSRLLL-L#VSN--LCQGVVS	T-----PVCPNGPGN
Δ4L	MDSKGSSQKGSRLL--L#VSN--LCQGVVS	T-----PVCPNGPGN
1K-c	MDSKGSSQKGSRLLLLL#VSNLLLCQGVVS	TK----PVCPNGPGN
2K-c	MDSKGSSQKGSRLLLLL#VSNLLLCQGVVS	TKK---PVCPNGPGN
3K-c	MDSKGSSQKGSRLLLLL#VSNLLLCQGVVS	TKKK--PVCPNGPGN
4K-c	MDSKGSSQKGSRLLLLL#VSNLLLCQGVVS	TKKKK-PVCPNGPGN
5K-c	MDSKGSSQKGSRLLLLL#VSNLLLCQGVVS	TKKKKKPVCPNGPGN
n-3KR	MDSKGSSQKKKRLLLLL#VSNLLLCQGVVS	T-----PVCPNGPGN
n-3KR-4K-c	MDSKGSSQKKKRLLLLL#VSNLLLCQGVVS	TKKKK-PVCPNGPGN
5D-c	MDSKGSSQKGSRLLLLL#VSNLLLCQGVVS	TDDDDDPVCPNGPGN
n-2LKR	MLSLGSSQKGSRLLLLLVVSNLLLCQGVVS	T-----PVCPNGPGN
Δ1L*	MDSKGSSQKGSRLLLL-VVSNLLLCQGVVS	T-----PVCPNGPGN
Δ2L*	MDSKGSSQKGSRLLL--VVSNLLLCQGVVS	T-----PVCPNGPGN
Δ3L*	MDSKGSSQKGSRLL---VVSNLLLCQGVVS	T-----PVCPNGPGN

Open in a new tab

Altered amino acids are underlined, positively charged amino acid residues are marked in red, and negatively charged are in cyan. #, location of εANB-Lys probe in the constructs for photocrosslinking. Mutants Δ1L – Δ4L were constructed earlier [23].

Homogeneous populations of [³⁵S] Met-labeled pPL RNCs were prepared in vitro by translation of mRNAs truncated at codon 86. In the absence of a natural stop codon at the end of the mRNA, ribosomes do not dissociate and the nascent chains remain bound. RNCs of wild-type and mutated preprolactin were prepared in the presence of purified canine SRP. Photocrossliking was induced by UV-irradiation, and SRP-photoadducts were visualized by SDS-PAGE. As it is shown in Fig. 1c, a strong photocrosslinking band, corresponding to the expected molecular weight of a nascent pPL and SRP54 photoadduct (NC-SRP54), was observed when SRP was present in the reaction mix, but not in the control reaction without SRP. The result is in good agreement with earlier studies on pPL-SRP interaction [23, 26, 38]. NC-SRP54 photocrosslinking was strictly probe-dependent, as no adduct was found in the control reaction without εANB-Lys-tRNA^amb (Fig. 1c).

As noted in the Introduction, signal sequences have three distinct structural regions: a positively charged n-region, a central hydrophobic h-region and a more polar c-region that contains the signal peptidase cleavage site. In order to delineate the parts of the signal sequence involved in interaction with SRP we designed mutated pPL proteins with (i) alterations in the charge content of the n-region (addition/removal of positively charged residues and/or decrease in the length of the n-region; mutations of this kind are denoted with “n-” in their names and a mnemonic for the kind of mutations made), (ii) alterations in the length of the h-region (deletions are denoted by Δ), (iii) alterations in the charge content of the early part of the mature protein immediately after the signal peptidase cleavage site (mutations of this kind are denoted with “-c” in their names), and (iv) combinations of i and iii. All mutants are listed in Table 1.

Positively charged residues in the n-region have only weak effects on the SRP-signal sequence interaction

The n-region of the wild-type pPL signal sequence contains one negatively charged residue (Asp at position 2) and three positively charged residues (Lys at positions 4 and 9, Arg at position 12), Table 1. Replacement of the Lys and Arg residues by Leu led to a mild decrease of SRP photocrosslinking (mutants n-3LR and n-4L, Fig. 2a). Deletion of the first 10 amino acid residues within the n-region (mutant n-Δ10,R, leading to decrease of the charge) gave similarly weak effect on the photocrosslinking efficiency. The introduction of two additional positively charged amino acid residues in the n-region did not significantly influence SRP photocrosslinking (mutant n-3KR, Fig. 2a).

Introduction of charged residues into the early part of the mature protein have only weak effects on the SRP-signal sequence interaction

Statistical studies have shown that there is a strong selection against Arg and Lys in the early part of mature portion of bacterial secretory proteins [48, 49]. Although such strong discrimination against positively charged amino acids was not observed in eukaryotic signal sequences, long strings of positively charged residues placed downstream of the signal sequence can block translocation of eukaryotic secretory proteins [50-52]. What is the nature of this inhibition of translocation? Does it influence interaction with SRP? Using the photocrosslinking approach, we found that introduction of one or two positively charged Lys residues in the early portion of mature pPL did not influence the SRP-nascent chain interaction (Fig. 2b). Further increasing the number of Lys residues to 3-5 led to a small decrease in the quantity of the SRP-photoadduct. The introduction of negatively charged amino acids (five Asp residues, Fig. 2b) had almost the same, weak inhibitory effect on the SRP-nascent chain photocrosslinking as had five Lys residues. The efficiency of SRP crosslinking to mutant signal peptides carrying extra positively charged residues both in the n-region and in the early part of mature protein (mutant n-3KR-4K-c) was equal to the crosslinking efficiency seen for the 4K-c mutant alone (Fig. 2a and 2b).

Deletions in the h-region have a strong effect on the SRP-signal sequence interaction

Recently, we have found that deletions of one or more Leu residues in the h-region of the pPL signal peptide compromise its interaction with SRP and lead to the activation of a novel, preemptive quality control mechanism called RAPP [23]. Here we conducted a quantitative analysis of the SRP interactions with the same mutant pPL signal sequences (Table 1). Deletion of only one Leu decreased SRP photocrosslinking efficiency almost by one third, while deletion of the second Leu had little effect, but deletion of three and four Leu led to an almost complete inhibition of SRP photocrosslinking (Fig. 3). These results show that the hydrophobicity of the h-region is critical for the SRP-nascent chain interaction. We hypothesized that it is more likely overall hydrophobicity of the hydrophobic core than its length, that is consistent with earlier findings [43]. The data on the hybrid protein LamB-prolactin suggested that a Leu-rich signal sequence is optimal for interaction with SRP [43]. Similar studies also suggested that the h-region is of major importance for the recognition by SRP [32, 33, 35]. Our results are in a good agreement with these observations. Thus, the signal sequence h-region is the key element for SRP-nascent chain interactions in mammalian cells.

Targeting of pPL with mutated signal peptide to the Sec61 translocon

In order to compare the effects on SRP biding with effects on the full translocation reaction elicited by the different signal sequence mutations, in vitro translocation experiments with dog pancreas rough ER microsomes (RMs) were conducted. Full-length wild-type and mutant pPLs were translated in vitro in the presence of purified RMs, and the efficiency of translocation was determined as the fraction of molecules in which the signal sequence was cleaved by the lumenally disposed signal peptidase. We did not see significant inhibition of cleavage for the proteins with mutations in the signal sequence n-region. All of them were efficiently processed and very little precursor with uncleaved signal sequence was detected (Fig. 4a). These data are in a good agreement with previous findings [53]. However, proteins with charged amino acid residues in the early portion of mature protein were not efficiently cleaved, and in general much more precursor was found for all mutated proteins of that group in comparison with the wild-type protein (Fig. 4b). The uncleaved precursor was not transported across the membrane since it could be digested by proteinase K, as shown for the 4K-c mutant (Fig. 4d). However, a mutant with extra positively charged amino acids both in the n-region and in the mature part, n-3KR-4K-c, was more efficiently translocated than the 4K-c mutant (Fig. 4a and 4b).

A likely explanation for the translocation defect seen for the mutants carrying extra positively charged residues in the early part of the mature chain is that such mutations may affect the orientation of the signal peptide in the ER membrane, effectively imparting an N_lum-C_cyt orientation on the signal sequence [48, 54, 55]. In this orientation, the signal sequence will not become a substrate for the lumenally disposed signal peptidase, and the protein will remain accessible to proteinase K.

Mutated pPLs containing different deletions in the h-region of the signal sequence were also tested for translocation across the ER membrane (Fig. 4c; see also Karamyshev et al., 2014 [23]). Earlier, we found that some deletions in the hydrophobic core led to inhibition of pPL translocation into RMs [23]. Here, we tested additional deletions in the signal sequence (Δ1L*, Δ2L*, Δ3L*; Table 1) and assayed translocation of these proteins into RMs (Fig. 4c). While only the unprocessed form of pPL was observed in the absence of RMs, the majority of the wild-type protein was processed in the presence of RMs. However, deletions of 1, 2 or 3 Leu residues led to a gradual inhibition of signal sequence cleavage, with practically complete absence of mature form for a mutant with a 3 Leu deletion. The uncleaved precursor protein but not the mature form was susceptible to proteinase K digestion, as expected (Fig. 4d).

We conclude that deletions in the h-region strongly inhibit translocation and/or targeting to the membrane. These results generally correlate with the photocrosslinking results (Fig. 3).

Plasmids, DNA techniques, mRNA, and tRNA

Plasmids containing bovine pPL constructs were made from the previously described pSPBP4 plasmid, a 950 bp NcoI-PstI fragment encodes preprolactin inserted into pSP64T vector [8]. Plasmids containing pPLs with an amber-stop codon and deletions in the hydrophobic core were constructed earlier [23]. Standard DNA techniques were performed according to the protocols of Sambrook and coworkers [72]. Site-directed mutagenesis was performed by a PCR technique using Pfu turbo DNA polymerase (Stratagene, La Jolla, CA). All mutations were confirmed by DNA sequencing. DNA fragments for in vitro transcription were generated by PCR under standard conditions using Ex Taq DNA polymerase (TaKaRa, Shiga, Japan). N^ε-(5-azido-2-nitrobenzoyl)-Lys-tRNA^amb (εANB-Lys-tRNA^amb) and truncated mRNAs were prepared as before [26, 47].

In vitro translation and site-specific photocrosslinking

In vitro translation for photocrosslinking experiments was conducted in wheat germ cell-free translation system as before [47, 73] with some modifications. Standard reaction (25-50 μl) contained 50 mM HEPES (pH 7.5), 100 mM KOAc, 3.5 mM Mg(OAc)₂, 1.0 μCi/μl [³⁵S]methionine, 1-1.4 pmoles/μl εANB-Lys-tRNA^amb (or unmodified Lys-tRNA^amb), and other components as described [47, 73, 74]. Where indicated, 0.06 μM canine SRP was added to the reaction mix. Translation was conducted in the dark for 30-40 min at 26^oC. Samples were photolyzed on ice for 15 min using a 500 W mercury arc lamp as described [73]. After photolysis, samples were sedimented through a 130 μl sucrose cushion (0.5 M sucrose, 50 mM HEPES (pH 7.5), 40 mM KOAc, 5 mM Mg(OAc)₂) in TLA 100 rotor (Beckman Optima TL centrifuge) at 100,000 rpm for 5 min at 4^oC. The pellets were resuspended in SDS sample buffer and analyzed by 10-15% gradient SDS-electrophoresis. [¹⁴C] molecular weight protein markers (Gibco or Amersham) were loaded on the gel and used for calculation of photoadducts molecular weight. Gels were visualized on a Bio-Rad Molecular Imager FX. Quantity One image software (Bio-Rad) was used for analysis. Quantity of each photoadduct was normalized to the amount of the corresponding nascent chain, and then calculated relatively to wild-type pPL crosslink.

Translations of full-length pPL mRNAs were performed at 30°C for 60 min as described [75, 76] in a total volume of 14 μl containing nuclease treated reticulocyte lysate, 40 units of RNase inhibitor, [³⁵S]Met (0.7 mCi/ml), 70 mM of each amino acid except Met, and 4 equivalents of RMs. Samples were analyzed by 12% SDS-PAGE. Gels were visualized on a Fuji FLA-3000 phosphorimager using the Image Reader V1.8J/Image Gauge V 3.45 software.

Proteinase K treatment

Translocation of polypeptides to the luminal side of the microsomes was assayed by resistance to exogenously added proteinase K and by analysis of constructs for protection. Proteinase K treatment was performed as before [77, 78] by adding CaCl₂ (13 mM final concentration) together with proteinase K (67 μg/ml final concentration) to the translation reaction. After incubating for 30 min on ice, the proteinase K was inactivated by adding PMSF (670 μg/ml final concentration) and the samples were further incubated for 5 min on ice before SDS-PAGE analysis.

Canine SRP, ER microsomes

Canine SRP was purified as described previously [79], except that the sucrose density sedimentation was omitted. SRP concentration was determined using ∈_{280 nm} = 1.0 × 10⁶ M⁻¹ cm⁻¹. Column-washed dog rough microsomes were prepared as before [80].