Role of Protein Translocation Pathways across the Endoplasmic Reticulum in Trypanosoma brucei*

The translocation of secretory and membrane proteins across the endoplasmic reticulum (ER) membrane is mediated by co-translational (via the signal recognition particle (SRP)) and post-translational mechanisms. In this study, we investigated the relative contributions of these two pathways in trypanosomes. A homologue of SEC71, which functions in the post-translocation chaperone pathway in yeast, was identified and silenced by RNA interference. This factor is essential for parasite viability. In SEC71-silenced cells, signal peptide (SP)-containing proteins traversed the ER, but several were mislocalized, whereas polytopic membrane protein biogenesis was unaffected. Surprisingly trypanosomes can interchangeably utilize two of the pathways to translocate SP-containing proteins except for glycosylphosphatidylinositol-anchored proteins, whose level was reduced in SEC71-silenced cells but not in cells depleted for SRP68, an SRP-binding protein. Entry of SP-containing proteins to the ER was significantly blocked only in cells co-silenced for the two translocation pathways (SEC71 and SRP68). SEC63, a factor essential for both translocation pathways in yeast, was identified and silenced by RNA interference. SEC63 silencing affected entry to the ER of both SP-containing proteins and polytopic membrane proteins, suggesting that, as in yeast, this factor is essential for both translocation pathways in vivo. This study suggests that, unlike bacteria or other eukaryotes, trypanosomes are generally promiscuous in their choice of mechanism for translocating SP-containing proteins to the ER, although the SRP-independent pathway is favored for glycosylphosphatidylinositol-anchored proteins, which are the most abundant surface proteins in these parasites.

Translocation of secretory protein precursors across the endoplasmic reticulum (ER) 4 membrane in eukaryotes and the plasma membrane in prokaryotes utilizes two routes. In higher eukaryotes, most proteins are believed to be targeted co-translationally to the ER by the signal recognition particle (SRP) pathway (1). In the co-translational translocation mechanism, the signal peptide or the transmembrane domain is recognized by the SRP; the ribosome-nascent chain-SRP complex then binds to the membrane via the SRP receptor, and after SRP release the translating ribosome interacts with the translocon (2). The polypeptide chain then moves from the tunnel of the ribosome into the membrane channel using peptide elongation as the driving force for translocation (3).
Eukaryotes can also translocate proteins across the ER membrane post-translationally (4 -6). It is not known how often this pathway is used in higher eukaryotes, but it is very active in fast growing cells such as yeast (2,7). The chaperones DNAJ/K bind to the proteins during their translation in the cytoplasm. Translocation via this pathway begins with binding of the translocation substrate to the SEC complex in the ER membrane. During this step, the cytosolic chaperones are released from the substrate (8). Translocation depends on the luminal protein Kar2p (BiP in higher eukaryotes) (9,10). The binding of the protein to BiP inside the lumen of the ER prevents movement back into the cytosol, resulting in net forward translocation. The process of binding BiP is repeated until the protein completely traverses the ER (11). Both translocation pathways utilize the same channel, which is composed of a conserved heterotrimeric membrane protein complex called the SEC61 complex in eukaryotes and SECY complex in bacteria and archaea (12).
In the yeast Saccharomyces cerevisiae, the complex that mediates post-translational translocation was named the SEC complex (10,13). The SEC complex is composed of Sec63p, Sec62p, Sec71p, and Sec72p. The first two genes, SEC62 and SEC63, are essential for growth, whereas SEC71 and SEC72 are not (14). Mutations in SEC62, SEC71, and SEC72 predominantly affect post-translationally translocated precursors, suggesting that these gene products may be specific for this pathway (10). In contrast, SEC63 is required for both translocation pathways in vivo (15).
Sec63p is a polytopic membrane protein that spans the ER membrane three times. Its luminal J-domain interacts with the ER chaperone Kar2p (16). In addition to the luminal J-domain, the protein also contains two distinct regions that are conserved in all Sec63p identified so far (17). Part of this region was shown to interact with Sec62p and is necessary only for the post-translational pathway, but another repetitive domain also present in the spliceosomal Brr2p is essential for the function of this protein in both the SRP and the post-translational translocation pathways (17).
Mammalian complexes that mediate post-translational translocation across the ER seem to differ from the complexes found in yeast. Mammals possess only SEC62 and SEC63 but not SEC71 and SEC72 (2). Sec62/63 proteins were shown to interact with Sec61p, the translocation channel, but no specific function has been ascribed to these proteins.
Interestingly in vitro studies in yeast demonstrate that Kar2p and Sec63p are required for both co-and post-translational translocation (15,16,18). In mammals, these two factors were found not to be essential for co-translational translocation because the Sec61 complex and the SRP receptor are sufficient for co-translational translocation into proteoliposomes (19). This discrepancy can be explained assuming that Sec63p plays a regulatory role in translocon gating, which is essential in intact membranes but not in artificial proteoliposomes (17).
Unlike eukaryotes, in bacteria such as Escherichia coli, translocation of most, if not all, signal peptide (SP)-containing proteins is mediated by the post-translational route (20). The SRP pathway is used exclusively for polytopic membrane proteins (21,22).
Very little is known about the mechanisms of protein translocation across the ER in trypanosomes. Trypanosomes possess an unusual SRP complex that contains two RNA molecules, 7SL RNA and a tRNA-like molecule (23,24). The Trypanosoma brucei SRP complex is also composed of the four S-domainbinding proteins but lacks the Alu domain-binding proteins (25). Depletion of the SRP pathway by RNAi of SRP54, SRP68, SRP19, or SRP72 does not affect the translocation of several SP-containing proteins to the ER, suggesting the existence of an active SRP-independent pathway for protein translocation (25,26). We suggested that in the absence of the SRP pathway, the SP-containing proteins traverse the ER via an alternative pathway and that mislocalization of the SP-containing proteins results from a defect in intracellular trafficking because of defects in translocation of polytopic membrane proteins (26). Indeed we have recently provided evidence that in trypanosomes the SRP pathway is essential for the biogenesis of polytopic membrane proteins (27). In elucidating the role of SRP and its receptor in protein translocation, we found that silencing of the trypanosome ␣ subunit of the SRP receptor resulted in a remarkable phenotype, which differs from the phenotype observed following SRP protein silencing. The accumulation of SRP-bound ribosomes on the ER membrane during ␣ subunit of the SRP receptor silencing elicited a signal that led to complete shutoff of spliced leader (SL) RNA transcription. Because in trypanosomes, SL RNA is required for the production of each mRNA by trans-splicing, shutoff of SL RNA transcription stopped mRNA and protein production and led to cell death (28).
Protein translocation across the ER in trypanosomes is of special interest because the most important proteins for trypanosome survival within its hosts are GPI-anchored proteins. These include the variant surface glycoproteins that undergo antigenic variation in the bloodstream form of the parasite (29,30) and the EP and GPEET procyclins expressed in the procyclic (insect) form (31). These proteins acquire the GPI anchor in the ER and are further modified in the Golgi and at the plasma membrane (32). It was therefore of great interest to examine whether the translocation and processing of these proteins in trypanosomes differs from that of other SP-containing proteins.
In this study, we examined the role of the SRP-independent post-translational translocation to the ER. Our results suggest that trypanosomes, like yeast and unlike higher eukaryotes, possess the SEC complex protein SEC71. RNAi silencing of this factor suggests that SEC71 is essential for parasite survival. Despite the continued translocation of SP-containing proteins to the ER under SEC71 depletion, several proteins were mislocalized. The SP-containing proteins examined in this study either carry SP only, SP and a single transmembrane domain, or SP and a GPI anchor. No effect on polytopic membrane protein biogenesis was observed in these cells. Unlike bacteria and other eukaryotes, trypanosomes seem to be flexible in utilizing either translocation route for most SP-containing proteins. Only the GPI-anchored proteins were severely affected by depleting the SRP-independent translocation route. The promiscuous entry of SP-containing proteins to the ER was blocked by co-silencing of both SEC71 and SRP68 or SEC63, a factor essential for both translocation pathways in yeast. Silencing of SEC63 blocked entry to the ER of both SP and polytopic membrane proteins as well as GPI-anchored proteins, suggesting that in trypanosomes this factor is essential for both translocation routes. Only after depleting the two translocation routes was accumulation of protein precursors observed. This study highlights the role of the SRP-independent pathway in protein sorting in trypanosomes and its role in the biogenesis of GPI-anchored proteins. The trypanosome co-translational machinery seems to be somewhat different from that identified in yeast and mammals. However, both translocation routes are essential for parasite survival.

EXPERIMENTAL PROCEDURES
Cell Growth and Transfection-Procyclic forms of T. brucei strain 29-13, which carries integrated genes for T7 polymerase and the tetracycline repressor (33), were grown in SDM-79 (34) supplemented with 10% fetal calf serum in the presence of 50 g/ml hygromycin and 15 g/ml G418. Generation of the stem-loop silencing constructs for SEC71 and SEC63 using the primers described in supplemental Table 1 were performed as described previously (26).
Northern Blot Analysis-Total RNA was prepared with TRIzol reagent, and 20 g/lane was separated on a 1.2% agarose gel containing 2.2 M formaldehyde. Specific mRNA was detected with ␣-32 P-random labeled probes (Random Primer DNA Labeling Mix, Biological Industries, Co., Kibutz Beit-Haemek, Israel).
In Vivo Labeling and Immunoprecipitation-Procyclic cells were labeled with 200 Ci/ml L-[ 35 S]methionine/cysteine (Amersham Biosciences) as specifically indicated in the different experiments.
Immunofluorescence and Confocal Microscopy-Cells were washed with phosphate-buffered saline, mounted on poly-L-lysine-coated slides, and fixed in 4% formaldehyde, and immunofluorescence was performed as described in Liu et al. (26). Finally the cells were visualized with a Zeiss LSM 510 META inverted microscope. The details of the source and dilu-tion of all antibodies used are presented in supplemental Table 2.
Scanning Electron Microscopy-Cells were washed with phosphate-buffered saline, mounted on glass coverslips, and subsequently fixed in glutaraldehyde/paraformaldehyde solution, 2% OsO 4 , 2% tannic acid/guanidine hydrochloride (35). After dehydration, the cells were air-dried and gold-coated using a Polaron sputter coater (Fisons). The cells were examined by a Jeol 840 scanning electron microscope (Tokyo, Japan).
Western Blot Analysis-Whole cell lysates (10 6 cell equivalents per lane) were fractionated by SDS-PAGE, transferred to Protran membranes (Whatman), and probed with the indicated antibodies. The details of the antibodies used are presented in supplemental Table 2.
Construction of SEC71-silencing Construct Carrying Blasticidin Resistance-To generate the stem-loop silencing construct for SEC71 carrying a blasticidin resistance gene, the pLew100 construct was digested with the restriction enzymes RsrII and BssHII and ligated with the blasticidin resistance gene fragment amplified with primers 154 and 155 (supplemental Table 1). Cells carrying the silencing construct for SRP68 (25) were transfected with the SEC71-silencing construct and selected with 10 g/ml blasticidin.
Generation of EP1-GFP Construct-The blasticidin S resistance cassette was excised from construct pHD887 (courtesy of C. Clayton, Heidelberg, Germany) using restriction enzymes SmaI and StuI. The DNA fragment was ligated to the FIGURE 1. Multiple sequence alignment of SEC71. Sequence alignment of T. brucei SEC71 with its homologues is shown. T. brucei Tb03.1J15.650 was identified as the SEC71 homolog using the following criteria. The T. brucei data base, GeneDB, was searched using a yeast and fungi SEC71 PROSITE profile (EXXDXXXXXXXLXPXX-XXXXXXXEI) (indicated in pink letters) obtained from Swiss-Prot: Scerevisiae, S. cerevisiae, accession number SEC66_YEAST; Spombe, Saccharomyces pombe, accession number Q9UUA4_SCHPO; and Ncrassa, Neurospora crassa, accession number Q7SH48_NEUCR. The similarity of T. brucei to these sequences is 35, 34.5, and 38%, and identity is 16.9, 10.8, and 11.3%, respectively. The alignment was performed using the ClustalW multiple sequence alignment program. Predicted N-glycosylation sites are indicated by green squares. The predicted transmembrane domain is boxed in pink. Residues marked in color represent identity (red), similarity (green), and weak similarity (blue). pG-⌬164.EG derivative pG-M1.5 (38) 5 that had been cut with XbaI, treated with Klenow enzyme, and cut with StuI. The resulting plasmid, pGbsr-M1.5, was processed with HindIII and StuI, and the insert was ligated to pTSARib (39) cut with HindIII and EheI, yielding construct pTSA-Ribsr. The EP1-GFP cassette was released from pG-⌬LII.EY (38) by restriction with NdeI, treatment with Klenow enzyme, and restriction with HindIII. The insert was ligated to pTSA-Ribsr cut with BamHI, treated with Klenow enzyme, and cut with HindIII. The resulting construct pTSA-Ribsr.EG was linearized with BglII prior to electroporation of trypanosomes. In the EP1-GFP cassette the GFP inserts at position 150 of the open reading frame between the small N-terminal domain and the EP repeats.
Generation of the EP1-GFP⌬GPI Polypeptide-The EP1-GFP⌬GPI was amplified from pKD4.EP1:GFP#C19, deleting the GPI-anchoring sequence by introducing a stop codon directly following the last EP1 repeat. The PCR product lacked nucleotides 1099 -1167 (GAATLKSVALPFAIAAAALVAAF) of the EP1 gene. The PCR was amplified using the primers EP.HindIII and EP⌬BglII (supplemental Table 1). The PCR product was cut with HindIII and BglII and cloned into pTsrib-(blast). For integration, the resulting construct, pTSA-Ribsr.EG, was linearized with SphI.

RESULTS
Trypanosomes Possess a SEC71 Homologue-To study the role of the protein translocation SRP-independent pathway in T. brucei, the genome was first searched for a homologue to SEC71, which participates in protein translocation across the ER in yeast (10,14). A SEC71 homologue was identified only after searching for this protein with a profile that was generated based on multiple alignments of this protein in yeast and other fungi (see Fig. 1). When the T. brucei data base was searched using part of the profile (EXXDXXXXXXXLXPXWXX-XXXXXXXEI), it resulted in a single hit. This protein (Tb03.1J15.650) was shown to carry the entire profile. The predicted open reading frame has the same length (206 amino acids) as the S. cerevisiae Sec71p. Alignment of the T. brucei SEC71 predicted amino acid sequence against its homologues from different yeast and fungi is presented in Fig. 1. A single stretch of amino acids that potentially represents a transmembrane domain is predicted between amino acids 33 and 51 (boxed in pink). No N-terminal signal sequence exists in this protein, suggesting that the first transmembrane domain may function in its targeting. The predicted topology of the T. brucei protein based on the Kyte-Doolittle method (40) agrees well with that of its yeast homologue (14). The yeast proteins possess two potential glycosylation sites in the luminal region upstream to the predicted transmembrane domain. The T. brucei protein homologue also contains such an N-linked glycosylation site (NX(S/T)) upstream to the predicted transmembrane domain. Homologues of SEC71 exist in both Trypanosoma cruzi (Tc00.1047053509453. 20) and Leishmania major (LmjF25.1310).
To investigate the role of SEC71 in protein translocation, we down-regulated its expression by RNAi. We used a construct that expresses the double-stranded RNA in the form of a stemloop structure transcribed from the tetracycline-inducible EP promoter (33). Growth of cells depleted of SEC71 mRNA is presented in Fig. 2A. Growth of the silenced cells (ϩTet) was impaired, and the cells stopped growing 4 days after induction. As expected, a major reduction in SEC71 mRNA was observed that was accompanied by an increase in double-stranded RNA production (Fig. 2B). To examine whether SEC71 silencing induced morphological changes, cells were examined by confocal and scanning electron microscopy. The results shown in Fig.  2C reveal no obvious changes in the size or shape of cells in contrast to the major perturbations observed under SRP depletion (25)(26)(27). It should be noted, however, that silencing of essential genes in T. brucei does not necessarily affect cell morphology (41,42).
Silencing of SEC71 mRNA Does Not Prevent Translocation of SP-containing Proteins-To examine whether SP-containing proteins utilize the SRP-independent pathway, the SEC71 mRNA-depleted cells were analyzed for translocation defects. The results are presented in Fig. 3. In A and B, the effect of SEC71 silencing on p67, a lysosomal protein that is heavily glycosylated, was examined. In T. brucei procyclic forms, mature p67 is present as a 100-kDa glycoprotein (gp100) (43). Cells on 5 M. Gü nzel and M. Engstler, unpublished data. the 3rd day of silencing were pulse-chased, and upon immunoprecipitation, the 100-kDa protein was observed (Fig. 3A). The amount of mature p67 was similar in uninduced and induced cells (lanes 3 and 4, respectively), suggesting that p67 is correctly translocated to the ER under SEC71 depletion. To further investigate the processing of p67, the depleted cells were stained with anti-p67 antibodies. The results, presented in Fig.  3B1, show that in uninduced cells the staining was confined to the single lysosome, whereas in induced cells several regions of staining were observed. The aberrant staining of p67 may either result from aggregation of mislocalized proteins or from fragmentation of the lysosome. Staining of the lysosome using LysoTracker TM (Molecular Probes) (Fig. 3B2) indicated no fragmentation of the lysosome. These results therefore suggest that despite proper translocation to the ER the vesicular transport of p67 was severely affected, leading to its mislocalization.
We next examined the expression of the flagellar pocket protein CRAM. CRAM is a type I membrane protein with a predicted molecular mass of 130 kDa that is concentrated in the flagellar pocket, an invagination of the cell surface of trypanosomes where endocytosis and exocytosis take place. The glycosylated form of the protein migrates as a 200-kDa protein (44). Western blotting showed that CRAM in induced cells was synthesized at similar levels compared with uninduced cells and was properly glycosylated (Fig. 3C). In addition, we examined the level and processing of the ERresident chaperone BiP and compared it with its level in cells depleted for SRP68. The cell line silenced for SRP68 that carries a stem-loop construct was described previously by us (25).
The results presented in Fig. 3D reveal that in either SEC71-or SRP68-depleted cells the level of BiP was slightly increased, and the protein was properly processed, suggesting that the protein is translocated to the ER. The accumulation of BiP may result from the accumulation of misfolded proteins in the ER as a consequence of depleting either of the translocation pathways. These data suggest that in the absence of the SRP-independent pathway the SP-containing proteins are normally translocated to the ER, most probably via the SRP pathway, but several are mislocalized.
The Biogenesis of Polytopic Membrane Proteins Is Dependent on the SRP Pathway but Not on the SRP-independent Pathway-We recently demonstrated that biogenesis of polytopic membrane proteins is dependent on the SRP pathway (27). Here we examined whether the SRP-independent pathway is also required for the translocation of polytopic membrane proteins. In this study the translocation of three polytopic membrane proteins was examined: TbNT8.1 (45), Tb29-2 (46), and vacuolar H ϩ -pyrophosphatase (47). The vacuolar H ϩ -pyrophosphatase functions in the acidification of the acidocalcisomes, which are cytoplasmic vesicles rich in calcium, phosphorous, and magnesium. The protein possesses a signal peptide and 16 transmembrane (TM) domains (47). The level of vacuolar H ϩ -pyrophosphatase was markedly reduced in SRP68-depleted cells but not in SEC71-silenced cells (Fig. 4A).
We next examined the level of two additional membrane proteins, Tb29-1 and Tb29-2. These proteins have a common large domain carrying octapeptide repeats. However, Tb29-2 contains in addition a large C-terminal domain with eight large hydrophobic TM domains, whereas Tb29-1 does not contain any obvious TM domains (46). Antibodies raised against Tb29 recognize both Tb29-1 and Tb29-2. The antibodies were used to probe a Western blot lysates of SEC71-silenced cells with SRP68-silenced cells as a control. The results (Fig. 4B) suggest that whereas the polytopic protein Tb29-2 was severely affected in SRP68-depleted cells no effect on this protein was observed in SEC71-silenced cells. The level of the SP-containing protein (Tb29-1) was unaffected in both SRP68-and SEC71-depleted cells. Finally we examined the biogenesis of TbNT8.1 fused at its N terminus to GFP (45). This protein is a permease that is  NOVEMBER 14, 2008 • VOLUME 283 • NUMBER 46 composed of 11 TM domains. Transgenic parasites carrying the SEC71-silencing construct were transfected with the plasmid carrying the GFP-fused TbNT8.1 transporter. The expression of TbNT8.1 was examined in uninduced and silenced cells (3 days after induction). The results indicate no effect on the level of membrane-bound TbNT8.1 as a result of SEC71 silencing (Fig. 4C1). We then examined the localization of the protein by confocal microscopy. No change in its localization was observed in the SEC71-silenced cells (Fig. 4C2). These results indicate that polytopic membrane protein biogenesis is not affected in cells depleted of the SRP-independent pathway and hence that these proteins are exclusively translocated via the SRP pathway (27).

Major Reduction of EP and gp63 during SEC71 Silencing-
The most abundant surface proteins of the T. brucei procyclic form are the GPI-anchored proteins EP and GPEET (48), which are present in several million copies per cell (49). The mature forms of EP migrate as heterogeneous populations of 45-50 kDa (36). The heterogeneity is a result of the addition of branched poly-N-acetyllactosamine structures to the GPI anchor (49,50). It was therefore of great interest to examine the production and processing of EP proteins under SEC71 silencing. The results presented in Fig. 5A1 demonstrate that in the uninduced cells the EP coat was uniform and heavy, whereas upon silencing the coat became uneven and much thinner. In the non-silenced cells, intracellular EP was localized around the flagellar pocket, whereas in the induced cells EP accumulated in vesicles, suggesting that the protein is trapped and not completely processed (Fig. 5A2). A, whole cell lysates from SRP68 or SEC71 mRNA-silenced cells using the extraction method described previously (26). The proteins were separated on 10% SDS-polyacrylamide gels and subjected to Western blot analysis using anti-vacuolar H ϩ -pyrophosphatase (VHϩppase) antiserum. B, the same as in A but the proteins were separated on an 8% SDS-polyacrylamide gel, and the blot was incubated with anti-Tb-29 antiserum. C, effect of SEC71 silencing on TbNT8.1-GFP expression. C1, membrane and soluble fractions were prepared as described under "Experimental Procedures" from uninduced (ϪTet) SEC71-silenced cells and cells after 3 days of induction (ϩTet) and were subjected to Western blot analysis with anti-GFP antibodies. Anti-hnRNPD 0 antibodies were used to control for the quality of fractionation. C2, localization of TbNT8.1-GFP. Cells were fixed with 4% formaldehyde for 25 min and visualized by confocal microscopy. Scale bar, 10 m. on the 3rd day after induction were subjected to immunofluorescence assay. Cells were fixed with 4% formaldehyde and incubated with anti-EP mAb 247 before (1) or after (2) treatment with Nonidet P-40 to enable labeling of surface or intracellular expression of the protein, respectively. Scale bar, 10 m. B, whole cell lysates from SEC71 or SRP68 mRNA-silenced cells were subjected to Western blot analysis using anti-EP mAb 247. C, whole cell lysates from SEC71 mRNA-silenced cells were subjected to Western blot analysis using anti-gp63 antiserum. Anti-hnRNPD 0 antibodies were used to control loading. D, Northern blot analysis of EP mRNA upon SEC71 depletion. RNA was prepared from uninduced cells (ϪTet) and cells after 3 days of induction (ϩTet). Total RNA (20 g/lane) was subjected to Northern blot analysis with random labeled probes. The transcripts examined are indicated by arrows.
We then examined changes in the level of this protein in the SEC71-silenced cells compared with SRP68-depleted cells. The results (Fig. 5B) demonstrate a difference between these two silenced cells lines. Whereas a major reduction of the mature form of EP was observed in the level of EP in the SEC71-silenced cells, no effect on the level of EP was observed in SRP68-depleted cells. These data suggest that EP production and processing depend on the SRP-independent pathway.
We then examined whether this effect is specific to EP or is shared with other GPI-anchored proteins. We therefore examined the level of a less abundant GPI-anchored protein, gp63, in the SEC71-silenced cells. Results shown in Fig. 5C suggest that gp63 levels were also markedly reduced in these silenced cells, suggesting that the effect is characteristic of GPI-anchored proteins. The reduction in the level of EP protein might conceivably result from regulatory effects on EP production either affecting transcriptional activation or mRNA stability. The level of EP mRNA was therefore examined before and after SEC71 silencing; the results ( Fig. 5D) show no effect of SEC71 depletion on the level of EP mRNA, suggesting that reduction in the level of EP stems from changes in its biosynthesis during translation/translocation or maturation of the protein.
To further investigate whether the reduction in EP is the result of defects in biogenesis of EP and GPEET, procyclic trypanosomes were labeled with [ 3 H]ethanolamine in uninduced and silenced cells. Ethanolamine is a component of the GPI anchor, which is attached to the C-terminal amino acid of the protein precursor in the ER (51). It was previously shown that after uptake into the cells, ethanolamine is rapidly converted into phosphatidylethanolamine (52). After labeling, GPI-anchored proteins from procyclic trypanosomes can be sequentially extracted by various solvents (36). We found that the uptake of [ 3 H]ethanolamine into uninduced and induced cells (48 h in the presence of tetracycline) during a 24-h labeling period was comparable (results not shown). However, the butanol extract of uninduced cells that contains most of the mature forms of EP and GPEET (36) was found to contain more than 10 times the amount of radioactivity relative to the corresponding extract from induced cells. Accordingly analysis of the butanol extracts by SDS-PAGE followed by fluorography showed a clear reduction in labeled EP and GPEET after SEC71 depletion (Fig.  6, left two lanes). The Triton extracts, which contain less hydrophilic (presumably immature) forms of EP and GPEET (53), also showed a reduction in incorporated radioactivity in the silenced cells, but to a lesser extent, compared with the butanol extracts (Fig. 6, middle two lanes). The decreased levels of labeled EP and GPEET in the induced cells were not because of decreased GPI anchor biosynthesis because induced and uninduced cells showed similar levels of labeled free GPIs (Fig. 6, two right lanes), which are extracted into the aqueous phase of the chloroform:methanol:water extract (37). The reduction in the level of these GPI-anchored proteins could be explained by reduction in entry into the ER, further suggesting that although small amounts of EP can traverse the ER in the absence of the SRP-independent pathway, most probably by the SRP pathway, this entry is less efficient. We cannot exclude the possibility that in the SEC71 knockdown cells small amounts of proteins still translocate via the SRP-independent pathway.

The Reduction in the Level of GPI-anchored Protein under SEC71 Depletion Is Not the Result of Defects in GPI Anchoring-
The decrease in GPI-anchored proteins under SEC71 silencing may also result from defects affecting C-terminal peptide cleavage and/or GPI transamidation. We therefore examined the level of EP and of EP lacking the GPI anchor in uninduced and in cells silenced for SEC71. To this end, an EP protein containing an HA tag and carrying multiple methionines to enable efficient labeling of the protein (EPMH) was used (Fig. 7, A and  B) (54). Pulse-chase labeling and immunoprecipitation was performed with anti-HA antibodies. The results presented in Fig.  7A demonstrate two major immunoprecipitated products, which were observed previously using the same construct; the first product was a ϳ40-kDa precursor that represents a protein that was modified by N-glycosylation and GPI anchoring and hence was able to translocate into the ER, and the second band represented fully modified GPI-anchored protein. The same experiment was performed using cells carrying the EPMH construct lacking the C-terminal domain with the GPI anchoring site (EPMH⌬GPI) (54). The level of immunoprecipitated EP was markedly reduced upon SEC71 silencing for both EPMH and EPMH⌬GPI, suggesting that it is not the ability to undergo GPI anchoring that causes the reduction in EP or its derivative upon SEC71 silencing. Interestingly no preprotein was observed following silencing most probably because of the rapid degradation of the preprotein, which failed to translocate into the ER. Even upon treatment of the cells with either lactacystin (55) or FMK-024 (morpholinourea-phenylalanine-homophenylalanine-fluoromethyl ketone from Enzyme Systems Products, Livermore, CA) (56), which inhibit degradation by the proteasome or lysosome, respectively, no precursor repre-

Protein Translocation in T. brucei
We next examined the translocation of another reporter, an EP1-GFP fusion that was also engineered to lack the GPI anchor, by deleting the C terminus domain of the protein (see "Experimental Procedures"). The expression of this fusion was examined under SEC71 silencing in pulse-chase experiments (Fig. 7B). The results suggest that the low level of EP under SEC71 silencing is not caused by defects in GPI anchoring but most likely by the inability to traverse the ER under this depletion. Indeed inspecting the SP sequence of the GPI-anchored proteins suggests that the SP of GPI-anchored proteins is sig-nificantly less hydrophobic compared with the SP of other translocated proteins (see "Discussion").

Co-silencing of the SRP and the SRP-independent Pathways Severely Affects Entry of SP-containing Proteins to the ER-Si-
lencing of each of the translocation routes still permitted the translocation of SP-containing proteins to the ER, suggesting that trypanosomes are promiscuous in choosing the translocation routes for these proteins. Even the GPI-anchored proteins that seem to be a preferential substrate of the SRP-independent pathway still traverse the ER in its absence, most probably via the SRP pathway. This promiscuity would prevent accumulation of preprotein because in the absence of one pathway the protein translocates using the alternative route. As already mentioned, we cannot rule out the possibility that in SEC71 knockdown cells translocation via the SRP-independent pathway was not completely blocked.
However, co-silencing of the two pathways should completely block entry of SP-containing proteins to the ER. To this end, we generated a cell line carrying the constructs to silence SEC71 and SRP68. The growth of these parasites was severely affected upon silencing (results not shown). We confirmed the expression of both silencing constructs by examining the production of double-stranded RNA to either SEC71 or SRP68 genes using gene-specific probes (Fig. 8A). Next the effect of co-silencing on protein translocation was examined. Cells were induced for silencing for 3 days, and proteins were subjected to Western analysis. The results demonstrate major reduction in the level of EP (Fig. 8B) and CRAM (Fig. 8C). Accumulation of CRAM precursor at the expected size (130 kDa) (44) was also observed. These results support the notion that because of the flexibility in choosing the translocation pathway even if only a single pathway for entry to the ER is active, the proteins traverse the ER, and no precursors accumulate.  However, blocking both pathways prevents entry to the ER, and precursors start to accumulate.
SEC63 Is an Essential Gene in T. brucei-To further investigate the factors that participate in entry of proteins to the ER, we searched for factors that govern both translocation pathways. In yeast, the Sec63p is required for both translocation pathways (15,16,18).
Bioinformatics analysis identified the SEC63 homologue in T. brucei. The deduced amino acid sequence of this protein was compared with its homologues in other eukaryotes (Fig. 9). The T. brucei protein, like its yeast homologue, is composed of thee transmembrane domains with a luminal DNAJ domain situated between the second and the third transmembrane domains. Like its homologues, the trypanosome protein contains the Brr2like domain (18,57). This domain is situated between the third transmembrane domain and the C-terminal part of the protein. However, the trypanosome protein is smaller than its homologues (490 residues compared with 663 and 760 in yeast and mammals, respectively). The truncation is situated in the C terminus. In particular, the trypanosome protein is missing two conserved regions, 52 predominantly acidic residues and an extreme C-terminal 14-residue sequence that was shown in yeast to interact with Sec62p (18, 57) (Fig. 9). The lack of Sec62p binding domain on the trypanosome Sec63p protein is not surprising given that there is no recognizable SEC62 homologue in the trypanosome genome. This suggests that trypanosomes may have a more restricted set of proteins that function in the SRP-independent translocation route.
To investigate the role of Sec63p in protein translocation and to examine whether this factor participates in both translocation pathways in trypanosomes, the expression of SEC63 was silenced by RNAi using a stem-loop construct as described above. SEC63 mRNA was completely eliminated in the silenced cells (Fig. 10A). SEC63 is an essential gene because its depletion resulted in growth arrest (Fig. 10B). The morphology of the parasites was examined upon silencing and exhibited major changes such as rounding of the cells as well as kinetoplast and nuclear multiplication without cytokinesis (Fig. 10C). This phe- FIGURE 9. Multiple sequence alignment of SEC63. The alignment was performed using the ClustalW multiple sequence alignment program. Residues depicted in red, green, and blue represent identity, similarity, and weak similarity, respectively. The sequences were obtained from Swiss-Prot: Hsapiens, Homo sapiens, accession number NM_007214; Mmusculus, Mus musculus, accession number NM_153055; Scerevisiae, S. cerevisiae, accession number NP_014897; and Tbrucei, T. brucei brucei, GeneDB accession number Tb09.211.1550. The T. brucei sequence shares identity of 30, 30, and 40% with human, mouse, and yeast, respectively. The predicted three transmembrane domains are boxed in pink. The DNAJ domain is indicated in green. The Brr2-like (Brl) domain is boxed in yellow, and the SEC62 interaction domain is boxed in blue. NOVEMBER 14, 2008 • VOLUME 283 • NUMBER 46 notype is reminiscent of the one observed previously for SRP54-and SRP receptor subunit ␣-silenced cells (26,28). Remarkably after 3 days, the cells induced the spliced leader silencing pathway (28), and the level of all mRNAs was reduced because of shutoff of SL RNA transcription. 6 Thus, to investigate the role of SEC63 in protein translocation, we examined the protein translocation defects 2 days after induction before spliced leader silencing was elicited.

Protein Translocation in T. brucei
SEC63-depleted Cells Fail to Translocate SP-containing Proteins to the ER-We next examined the levels of EP, p67, BiP, and CRAM before and after SEC63 silencing. The results (Fig.  11) indicate major reduction of all the proteins tested. Fig. 11A1 demonstrates a major reduction in EP levels. This reduction was also visualized by immunofluorescence (Fig. 11A2). Reduction in p67 by immunoprecipitation and by immunofluorescence was also observed (Fig. 11B, 1 and 2). Although the levels of EP, p67, and BiP were reduced, no precursors of these proteins were detected. In contrast, the CRAM level was not only reduced but the 130-kDa preprotein (44) was accumulated (Fig.  11C) as observed following the co-silencing of SEC71 and SEC68 (Fig. 8). Silencing of SEC63 blocked the biogenesis of polytopic membrane proteins as shown by the level of vacuolar H ϩ -pyrophosphatase (Fig. 11D). The severe effect of SEC63 depletion on all membrane protein biogenesis suggests that SEC63 is essential not only for the SRP-independent but also for the SRP translocation pathway.
As discussed above, we anticipated that when entry to the ER is completely blocked, preprotein should accumulate. We then examined the processing of EP1-GFP (described above) under SEC63 depletion. The results (Fig. 12A) demonstrate that the pre-EP1-GFP accumulates in silenced cells. Because the low level of the precursor may reflect its rapid degradation, we examined whether it might be degraded by the proteasome or the lysosome (Fig. 12B). We treated the silenced cells with lactacystin and FMK-024. These treatments did not change the level of precursor, suggesting that if such degradation occurs it is not mediated by either the proteasome or the lysosome.

DISCUSSION
In this study, two factors involved in entry of proteins to the ER were investigated: SEC71, a component of the SRPindependent pathway, and SEC63, which is essential for both translocation pathways in yeast (15,16,18). Our results demonstrate that the SRP-independent pathway is an essential pathway in trypanosomes. However, as opposed to E. coli and yeast, we found that trypanosomes do not appear to have a single dedicated pathway for the majority of SP-containing proteins examined in this study. We demonstrated here that upon depletion of the SRP-independent translocation pathway SP-containing proteins maintain their ability to translocate. Surprisingly GPI-anchored proteins were found to be far more reliant on the SRP-independent pathway, suggesting that trypanosomes are not completely promiscuous in choosing the translocation route for each of their proteins. Depletion of SEC63 severely impaired the translocation of all SP-containing proteins to the ER and also blocked polytopic membrane protein biogenesis, suggesting that this factor is also essential for both translocation pathways in trypanosomes. Our results suggest that trypanosomes can utilize the two translocation pathways interchangeably for translocating most SP-containing proteins into the ER. However, despite entry to the ER following specific depletion of either pathway, several proteins were mislocalized, suggesting that both pathways are essential for proper trafficking of proteins within the cell. Only under SEC63 silencing or co-silencing of SRP68 and SEC71 was entry to the ER almost completely blocked. On the other hand, polytopic membrane protein biogenesis was unaffected when the SRP-independent pathway was depleted.
Protein Translocation Machinery in Trypanosomes Versus That in Yeast, Mammals, and Other Parasites-In the yeast S. cerevisiae, a clear division between the two translocation pathways exists. For instance under SEC71 depletion, the translocation of ␣-factor and carboxypeptidase proteins is perturbed, but SRP substrates such as invertase and Kar2p are translocated normally (58). Mammals lack homologues to SEC71 and SEC72, and the role of SEC62/63 in protein translocation is still an open question. Although mammals contain SEC62/63, it appears that, in contrast to yeast, there is a single type of higher order complex for protein translocation (2,17). Recent studies in mammalian cells suggested that signal peptide-containing 6 H. Goldshmidt and S. Michaeli, unpublished data. proteins traverse the ER via the SRP pathway (59), but no study to date has investigated the role of the SRPindependent pathway in translocation into and across the ER in vivo.
Trypanosomes seem to behave differently from either yeast or mammals. Trypanosomes, like yeast, possess SEC71, which functions exclusively in SRP-independent translocation. But in contrast to yeast where the translocated proteins are faithful to and dependent on a single translocation route (58,60,61), trypanosomes are promiscuous in choosing their translocation route to the ER (Fig. 3). Only the GPI-anchored proteins seem to be translocated exclusively by the SRP-independent pathway (Figs. 5-7).
However, unlike yeast, trypanosomes lack SEC72 and SEC62 but possess a SEC63 homologue. The SEC63 protein lacks the domain known to interact with SEC62 (17,18), supporting the notion that this factor does not exist in trypanosomes. Trypanosomes may therefore possess a minimal SRP-independent apparatus that is able to recognize and translocate signal peptide-containing proteins. This complex may have evolved exclusively to translocate the very abundant GPI-anchored proteins that are crucial for parasite survival within the host. Note that we cannot rule out the possibility that there might be trypanosome homologues of the missing factors (SEC62 and SEC72) that deviate so strongly from the yeast sequence that they escaped detection by bioinformatics screening. It is interesting to note that another parasite, Plasmodium falciparum, does not carry SEC71 or SEC72 but possesses both SEC62 and SEC63, suggesting that in this parasite, as in mammals, a single complex for ER translocation might exist (62).
In yeast, SEC63 is part of the post-translational SEC complex, but Sec63p depletion in S. cerevisiae affects the translocation of both the signal peptide-containing proteins and integral membrane proteins, which are SRP substrates. This sug- FIGURE 11. Translocation of GPI-anchored proteins, SP-containing proteins, and polytopic membrane proteins to the ER is blocked in SEC63-silenced cells. A1, Western blot analysis was performed as described under "Experimental Procedures." The blot was incubated with anti-EP mAb 247 and anti-BiP antibodies. Reactivity with anti-hnRNPD 0 was used to control for equal loading. A2, immunofluorescence with anti-EP mAb 247. Uninduced cells (ϪTet) and induced cells (ϩTet) on the 2nd day after induction were fixed with 4% formaldehyde and incubated with anti-EP mAb 247 (surface staining). Nuclei (N) and kinetoplasts (K) were stained with 4Ј,6-diamidino-2-phenylindole. Scale bar, 5 m. B1, immunoprecipitation of p67. Total cell lysates (lanes 1 and 2) and the immunoprecipitated products (lanes 3 and 4) from uninduced cells (lanes 1 and 3) and cells after 2 days of induction (lanes 2 and 4) were analyzed on a 10% SDS-polyacrylamide gel. The immunoprecipitated product (gp100) is indicated by an arrow. B2, immunofluorescence microscopy with anti-p67 antiserum. The yellow arrows indicate the p67-stained lysosome (ϪTet), and the blue arrows indicate the aberrant p67 localization (ϩTet). Scale bar, 5 m. The protocols for B1 and B2 are as in Fig. 3, A and B, respectively. C, Western blot analysis of CRAM protein. Whole cell lysates (10 6 cells/lane) were prepared from uninduced cells (ϪTet) and cells after 2 days of induction (ϩTet), separated on a 8% SDS-polyacrylamide gel, and subjected to Western blot analysis using anti-CRAM antibodies. The mature form (mCRAM) and the unglycosylated form of CRAM (pCRAM) are indicated. Reactivity with anti-hnRNPD 0 was used to control equal loading. D, Western blot analysis of vacuolar H ϩ -pyrophosphatase (VHϩppase). Whole cell lysates were separated on a 10% SDS-polyacrylamide gel and subjected to Western blot analysis using anti-vacuolar H ϩ -pyrophosphatase antiserum.
gests that SEC63 is essential for both translocation pathways (15,18). In trypanosomes, SEC63 also affected both pathways because defects in the translocation of EP as well as polytopic membrane proteins were observed, representing substrates of both translocation pathways (Ref. 27 and this study). Interestingly in vitro translocation using mammalian proteoliposomes demonstrated that SRP-mediated translocation can take place even without Sec63p (19). It will be therefore of great interest to determine the role of SEC63 in ER translocation in mammalian cells in vivo and to determine whether this factor is also needed for both types of ER translocation routes as in trypanosomes and yeast.
The Specific Defects Induced by Perturbing Protein Translocation in Trypanosomes-In trypanosomes, depletion of a single pathway did not affect the level of the translocated SPcontaining proteins (except for GPI-anchored proteins). Surprisingly even under SEC63 depletion, no precursors were detected for p67 and the GPI-anchored proteins. Precursors were detected only under SEC63 depletion using an EP1-GFP fusion, but no effect on the level of this precursor was observed when either the proteasome or lysosome were inhibited; these results suggest that a specific protease(s) may exist to degrade these precursors. In recent years, different types of intramembrane proteases have been described (63) that may degrade misfolded proteins that accumulate in the ER. This protease(s) in trypanosomes most probably requires the function of at least one of the translocation pathways because no EP1-GFP precursor was observed in SEC71-silenced cells, although such precursors accumulated in SEC63 silenced cells.
Despite the entry of the SP-containing proteins to the ER under SEC71 or SRP depletion (26), several proteins were mislocalized. The dependence on the two pathways for proper sorting may stem from the fact that specific proteins involved in post-ER transport are heavily dependent on one or the other of the pathways. Indeed BiP, which only needs to traverse the ER and does not utilize the intracellular trafficking machinery, is correctly localized when either the SRP or the SRP-independent pathway are depleted. Interestingly down-regulation of the SRP pathway also results in selective defects in post-ER membrane trafficking in mammalian cells (59).
The reduction in the level of GPI-anchored proteins observed selectively under depletion of the SRP-independent pathway could have resulted from either destabilization of the translocated proteins because of defects in GPI anchoring or because the protein failed to translocate into the ER. The finding that the level of EPMH⌬GPI was reduced upon SEC71 depletion, like its parental construct EPMH, suggests that it is not the defects in the GPI anchoring process that cause the reduction in the GPI-anchored proteins during this silencing. We can therefore conclude that translocation defects across the ER are the main reason for the low level of these proteins in the silenced cells. However, we cannot completely exclude the possibility that in SEC71-silenced cells there may be defects in the final stages of GPI-anchored protein maturation.
Choice of the Optimal Translocation Pathway and the Dependence of GPI-anchored Proteins on the SRP-independent Pathway-In E. coli, the distinction between SRP substrates and those proteins that utilize the post-translational pathway seems to be simple and to depend mostly on the protein sequence itself (64). Increasing the hydrophobicity of SP shifts the translocation from the post-to the co-translational pathway. In mammals, RNAi of the SRP pathway leads to reduced levels of both SP-containing proteins and polytopic membrane proteins. In fact, the level of polytopic membrane proteins is less affected than that of SP-containing proteins most probably because these have higher affinity to the SRP and therefore win the competition for limited SRP (59). In yeast, each translocated protein is associated with a dedicated pathway, which most probably is also governed by the hydrophobicity of the protein domain presented to the translocation system (SP or TM). The very effective machinery for post-translational translocation in yeast may reflect the need to quickly translocate proteins in these fast growing organisms. On the other hand, SRP is extensively utilized in mammals for translocation of both SP-containing proteins and polytopic membrane proteins. In more slowly dividing mammalian cells, the utilization of the SRP complex may add another level of regulation. Indeed regulation exists in the signal peptide sequence itself that determines the efficiency by which the proteins translocate to the ER. Different signal sequences respond differentially to the accumulation of unfolded proteins in the ER (65).
Indeed in trypanosomes as well the dependence of GPI-anchored proteins on the SRP-independent pathway for translo- cation may stem from the properties of their SP. To test this hypothesis, we examined the hydropathy score of GPI-anchored proteins compared with other proteins carrying SP in the trypanosome genome using the SPScan program (Accelrys Software Inc.). We found that GPI-anchored proteins had a significantly lower score. For example, the scores of the GPI-anchored proteins are as follows: EP1 (Tb10.6k15.0020), Ϫ63.1; GPEET-2 (Tb927.6.510), Ϫ63.1; gp63 (Tb927.8.1610), Ϫ55.1; and variant surface glycoprotein (Tb927.4.5560), Ϫ62.4. On the other hand, non-GPI anchored proteins carrying SP possess more hydrophobic SPs with the following scores: glucose-regulated protein 78 (Tb11.02.5500), Ϫ87.5; p67 (Tb927.5.1810), Ϫ71.4; coated vesicle membrane protein erv25 (Tb10.70.0120), Ϫ100; and acidic phosphatase (Tb927.5.630), Ϫ94.1. The selective reliance of GPIanchored proteins on the SRP-independent pathway may therefore reflect the poor recognition of the SP of these proteins by the SRP, making them a natural substrate for the SRP-independent pathway. The need to meet the great demand for the most abundant surface proteins in procyclic forms of the parasite, EP and GPEET, may explain why these proteins became reliant on the SRP-independent pathway for their swift and effective translocation. This finding may be used in the future to search specifically for drugs to inhibit this protein translocation pathway particularly as this pathway appears to be less important to the mammalian host.