(Received for publication, November 20, 1996, and in revised form, January 22, 1997)
From the Michigan State University-DOE Plant Research Laboratory, Michigan State University, East Lansing, Michigan 48823-1312
We used complexes of avidin and biotinylated
precursors to generate translocation intermediates that occupy
functional transport sites and thereby block the transport of other
precursor proteins into pea chloroplasts. Cysteine residues of purified
precursor to the small subunit of rubisco (prSS) were modified with the biotinylation reagent
biotin-1-biotinamido-4-[-4-(maleimidomethyl)-cyclohexane-carboxamido]butane. Chemically biotinylated prSS was readily imported into chloroplasts. The addition of avidin, however, resulted in the formation of an
avidin-biotinylated precursor complex that could not be imported into
chloroplasts even when precursors had already engaged the transport
apparatus before avidin was added. On fractionation, the
avidin-biotinylated precursor complex associated with envelope membranes. Titration of transport sites with avidin-biotinylated precursor complexes revealed that saturation was reached at 2,000 molecules/chloroplast. Even with less than saturating levels of complexes, a sufficient number of translocation sites could be occupied
with avidin-precursor complexes so that the import rate of freshly
added radiolabeled prSS was reduced by 35%. From these observations we
conclude that the trapped intermediates were blocking functional
translocation sites. These biotinylated translocation intermediates should be useful in future efforts to purify
and characterize the chloroplastic protein import machinery.
Most chloroplastic proteins are encoded in the nucleus and synthesized on cytoplasmic ribosomes as precursor proteins containing an N-terminal transit peptide (1-3). After translation, precursors are transported across the two envelope membranes into the stroma, where the transit peptide is removed by a stromal processing peptidase (4). It is now widely accepted that the transport process involves a proteinaceous translocation apparatus located in both the outer and inner envelope membranes of the chloroplast. Several approaches have been used to identify components of the chloroplastic translocation machinery (5-10). These diverse approaches have identified several membrane proteins of 34, 70, 75, 86, and 110 kDa in molecular size (10-15). The role of these proteins in the import process remains to be elucidated. Due to the complexity of the translocation machinery, more components are probably required and remain to be identified.
Identification and characterization of components of the mitochondrial import apparatus has been aided greatly by the availability of translocation intermediates stuck in the transport machinery. This strategy has been used rather successfully by several different laboratories. For instance, a translocation intermediate was generated by Eilers and Schatz (16) using a chimeric precursor protein consisting of a mitochondrial targeting signal fused to dihydrofolate reductase (DHFR).1 This fusion protein can be imported into mitochondria. However, on addition of methotrexate, a folate anologue that binds with high affinity to DHFR, the fusion protein becomes stably folded and trapped in the import apparatus (16, 17). This blocked intermediate could then be cross-linked to various components of the mitochondrial import apparatus (18). In addition to identifying components of the mitochondrial import apparatus, translocation intermediates have also provided a means to investigate the energetics of protein import into mitochondria and have been useful in identifying stromal factors involved in unidirectional import into mitochondria (19).
Generating translocation intermediates using intact chloroplasts, however, has been more difficult. Limiting the levels of ATP available in an import reaction allows precursors to associate with chloroplasts but not to be fully translocated (20-22). This interaction was originally called binding, but more recently it has been recognized that it is not a reversible association, and the precursors more probably represent early translocation intermediates (8, 22). Although the topology of this early translocation intermediate is not clearly resolved (10, 40), efforts to trap precursors at a later stage of import have been less successful. For instance, the import of the precursor for 5-enolpyruvyl-shikimate-3-phosphate synthase was reduced but not arrested by the addition of its competitive inhibitor, glyphosate (23). In an approach similar to that used successfully with mitochondria, methotrexate did not block the import into chloroplasts of a chimeric fusion protein involving DHFR (24, 25). Even the introduction of a stop transfer domain from endoplasmic reticulum proteins into precursor proteins did not prevent import into chloroplasts (26). More recently, Wu et al. (15), using the chimeric protein Oee1-DHFR, were able to generate translocation intermediates by preincubating the fusion protein with antibodies against DHFR. Schnell and Blobel (9) also generated translocation intermediates using import conditions. They devised a unique hybrid precursor consisting of prSS fused to the IgG binding domain of staphylococcal protein A. This hybrid yielded two types of translocation intermediate, depending on the conditions. Under low ATP levels the precursor formed the same type of early translocation intermediate described above. With adequate levels of ATP, the chimeric precursor imported slowly, allowing the identification of some intermediates that spanned both the outer and inner envelope membranes. The difficulty in generating translocation intermediates from chloroplast may suggest that its import machinery has different import characteristics than other transport systems.
The availability of discrete translocation intermediates blocked at
later stages in import may help with the identification and
characterization of additional translocation components from the inner
envelope membrane and with investigating the role, if any, of molecular
chaperones and other stromal factors. This article describes a strategy
whereby the import of a biotinylated chloroplastic precursor is blocked
when avidin is present. Available cysteine residues on prSS were
modified with the biotinylation reagent biotin-BMCC (see Fig. 1). When
the chemically biotinylated precursor prSS-BMCC is incubated in the
presence of avidin, a complex forms. This avidin-precursor complex
interacted strongly with chloroplasts but was not able to be imported
when incubated with high levels of ATP. The avidin-precursor complex
fractionated with the membrane fraction. This complex occupied a
sufficient number of import sites to alter the import rate of freshly
added precursor. The potential usefulness of these translocation
intermediates for isolating components of the translocation apparatus
is discussed.
Pea seeds (Pisum sativum var.
little marvel) were obtained from Olds Seed Co. (Madison,
WI). [35S]EXPRE35S35S
(cysteine/methionine) and 35S-methionine translation grade
were obtained from DuPont NEN. Percoll silica gel and Mg-ATP were
purchased from Sigma.
Isopropyl-1-thio--D-galactopyranoside was obtained from
Life Technologies, Inc. Biotin-BMCC, purified avidin, and immobilized
avidin resin were obtained from Pierce. Plasmid pet11D-prSS (27) was
provided by Dr. R. Klein (University of Kentucky, Lexington, KY).
Intact chloroplasts were isolated from 8-12-day-old pea seedlings by homogenization and differential centrifugation followed by sedimentation through a Percoll gradient as described previously (28). Chloroplasts were washed twice in 50 mM HEPES/KOH (pH 7.7), 0.33 M sorbitol (import buffer) and finally resuspended to a concentration of 1 mg of chlorophyll/ml of import buffer.
Overexpression and Purification of prSSThe pet11D-prSS was
introduced into Escherichia coli BL21(DE3) (29, 30). For
35S labeling, prSS was prepared as described by Schnell and
Blobel (9). prSS was sequestered into inclusion bodies on induction with isopropyl-1-thio--D-galactopyranoside. The
inclusion bodies were isolated from E. coli essentially by
the method of Lin and Cheng (31). Purified 35S-prSS was
stored at
80 °C at a concentration of approximately 1 mg/ml, with
a specific activity of 1 × 106 dpm/µg protein.
An SP6 vector containing the full-length precursor of a small subunit of rubisco from pea was used for in vitro transcription and translation (32). In vitro transcription was performed using SP6 RNA polymerase to generate mRNA (26), whereas translation in the presence of [35S]methionine was performed using a wheat germ system, incubating for 90 min at 25 °C as described by Bruce et al. (28).
Biotinylation of prSSApproximately 100 µg of E. coli-expressed 35S-prSS in 6.0 M
guanidine-HCl, 50 mM Tris-HCl (pH 6.8), 10 mM
dithiothreitol was applied to a Sephadex G-25 (Pharmacia Biotech Inc.)
column equilibrated in 6.0 M guanidine-HCl, 50 mM Tris-HCl (pH 6.8) to remove dithiothreitol. The eluant
was collected and added to a biotinylation reaction containing
Biotin-BMCC (Pierce) to a final concentration of 3 mM and
50 mM Tris-HCl (pH 6.8) buffer to a final volume of 100 µl. The reaction was incubated for 1-3 h at room temperature in the
dark. The biotinylation reaction was quenched by the addition of
dithiothreitol to 5 mM and incubated for an additional 20 min. The reaction was diluted 10-fold with 6.0 M
guanidine-HCl, 50 mM Tris-HCl (pH 6.8) and immediately
applied to a Sephadex G-25 column to remove excess biotin-BMCC. The
column was centrifuged at 200 × g for 5 min. The
eluant was collected and concentrated by trichloroacetic acid
precipitation. The insoluble material was collected by centrifugation
at 10,000 × g for 10 min. The pellet was washed two
times with cold acetone, finally solubilized in 100 µl of buffer
containing 6.0 M guanidine-HCl, 50 mM Tris-HCl (pH 6.8), 10 mM dithiothreitol, and stored at 80 °C.
The concentration of the biotinylated precursor protein was
approximately 1 mg/ml.
The import of guanidine-HCl-denatured 35S-prSS or 35S-prSS-BMCC into isolated chloroplast was performed as described by Perry and Keegstra (8). After import, intact chloroplasts were reisolated by centrifugation through 40% Percoll. The recovered chloroplasts were lysed and fractionated (28) to yield a crude membrane and soluble fraction. Both membrane and soluble fractions were analyzed by SDS-PAGE and fluorography.
Import with wheat germ-translated 35S-prSS was performed as described by Tranel et al. (14). The imported proteins were analyzed by SDS-PAGE and fluorography.
Generation of Avidin-Biotinylated Precursor ComplexesTwo strategies were used to generate avidin-biotinylated precursor complexes. In the first, avidin-biotinylated precursor complexes were formed prior to incubation with chloroplasts. To accomplish this, purified 35S-prSS-BMCC expressed by E. coli was incubated in the presence of avidin in import buffer (final volume, 10-40 µl) for 5, 15, or 30 min at 25 °C in the dark. The molar ratio of biotinylated precursor:avidin in the preincubation reaction was 1:2. This ratio should favor formation of complexes with a single biotinylated precursor bound to a single avidin protein. To investigate the import capacity of precursor-avidin complexes, we added aliquots of the reaction directly to an import assay and incubated for 30 min at 25 °C. Chloroplasts were repurified through 40% (v/v) Percoll and analyzed by SDS-PAGE and fluorography.
The second strategy was first to allow biotinylated prSS to engage the translocation apparatus before avidin was added. To accomplish this, we bound 35S-prSS-BMCC to chloroplasts at 4 °C, in the dark, for 20 min in the presence of 75 µM ATP. Intact chloroplasts were repurified through a 40% (v/v) Percoll cushion, washed twice with import buffer, and resuspended in import buffer containing avidin (at a ratio of 2 mol of avidin:1 mol of biotinylated precursor). Complex formation continued at 4 °C, in the dark, with continuous rocking for 30 min. Import for 30 min was initiated by adjusting the ATP concentration to 4 mM and the temperature to 25 °C. Chloroplasts were repurified through 40% (v/v) Percoll and analyzed by SDS-PAGE and fluorography.
Titration of Import Sites with Avidin-Precursor ComplexesIncreasing amounts of E. coli -expressed precursor (15-245 nM) were incubated with chloroplasts at 75 µM ATP in the dark at 4 °C for 10 min. Chloroplasts were repurified through a 40% Percoll cushion, washed with import buffer, and then incubated with import buffer containing avidin. Complex formation continued for an additional 30 min, in the dark, on ice. Import was then initiated by adjusting ATP levels to 4 mM and allowed to continue for 20 min at room temperature in room light. The entire reaction was overlaid onto a 40% Percoll cushion followed by centrifugation. Intact chloroplasts recovered from the pellet were fractionated into a crude membrane and supernatant fraction. The membrane fraction containing the blocked avidin complexes were analyzed by SDS-PAGE and fluorography. Gels were further quantitated directly by PhosphorImager (Molecular Dynamics).
Competition Experiment using Avidin-Precursor ComplexesChloroplasts containing avidin-biotinylated precursor complexes trapped in the import apparatus were prepared as described in the previous section and repurified through a 40% Percoll cushion. Chloroplasts were resuspended in 400 µl of import buffer containing 4 mM Mg-ATP and allowed to incubate at room temperature for 15 min. Wheat germ-translated prSS (~500,000 dpm/100-µl reaction) was added, and import proceeded for 0.5, 2, 4, and 20 min. Import reactions were immediately terminated at the times indicated by diluting 10-fold with cold import buffer and immediately centrifuging through a 40% Percoll cushion. Pellets were resuspended in sample buffer and analyzed by SDS-PAGE and fluorography. Gels were quantitated directly by PhosphorImager.
The cysteine residues of chemically purified prSS
were modified by reaction with biotin-BMCC (see Fig. 1).
Biotinylation of prSS was confirmed by SDS-PAGE followed by transfer to
Immobilon P and detection of the biotinylated proteins with avidin
conjugated to alkaline phosphatase (Fig. 2A,
lane 2). Biotinylation of prSS resulted in a mobility shift
of approximately 2 kDa when analyzed by SDS-PAGE and fluorography (Fig.
2B, compare lanes 1 and 2). The
molecular weight of biotin-BMCC is 533 and, if all four cysteine residues of prSS were modified with biotin-BMCC, this would produce a
2-kDa change. Alternatively, biotinylation of the precursor may alter
its affinity for SDS, thereby contributing to the observed mobility
shift of prSS-BMCC. Several other biotinylated bands also appear (Fig.
2A, lane 2). These additional bands most likely represent
contaminates from E. coli, because they are not radiolabeled (Fig. 2, compare A, lane 2, with B, lane 2).
Finally, all the radiolabeled prSS shifted to a higher molecular weight
form, indicating that biotinylation was complete (Fig. 2B,
lane 2). Biotinylation of prSS was also confirmed by
demonstrating that approximately 95% of prSS-BMCC bound to a column
containing immobilized avidin, whereas prSS did not (data not
shown).
Biotinylated prSS can be imported into intact pea chloroplasts (Fig.
3). Protease protection assays using thermolysin
confirmed that both mature-sized SS and SS-BMCC were no longer
sensitive to digestion and therefore had entered chloroplasts (Fig. 3,
lanes 3 and 6, respectively). Both SS and SS-BMCC
were present in the soluble fraction after import (Fig. 3, lanes
3 and 6, respectively). Moreover, the mobility of
imported SS-BMCC was similar to that of SS. However, the modified SS
still contained biotin, because on analysis by SDS-PAGE, followed by
transfer to Immobilon P, the biotinylated SS could be detected with
avidin conjugated to alkaline phosphatase (data not shown).
Avidin Blocks the Import of Biotinylated prSS
We sought to determine whether the addition of avidin would halt translocation of biotinylated prSS, thereby generating intermediates stuck in the translocation apparatus. Two different strategies were used during these efforts. In the first, avidin and biotinylated precursor were incubated together and allowed formation of a complex before adding them to chloroplasts in the presence of sufficient ATP to support import. In the second, biotinylated precursors were bound to the surface of chloroplasts in the presence of low ATP levels. Avidin was then added before the ATP levels were raised.
In the experiment shown in Fig. 4, avidin was first
incubated with biotinylated precursors for 5, 15, or 30 min before the complexes were added to chloroplasts. The data in Fig. 4 show that
these complexes bound to chloroplasts; however, they were not imported
(Fig. 4, compare lane 8 with lanes 9, 11, and
13). In control experiments, biotinylated prSS was imported
in the absence of avidin, (Fig. 4, lane 8), and the import
of unmodified prSS was not blocked by avidin (Fig. 4, compare
lanes 3 and 5). Complexes of avidin-biotinylated
prSS fractionated with membranes (Fig. 4, lanes 9, 11, and
13). In control experiments, the products derived from
biotinylated prSS imported in the absence of avidin (Fig. 4, lane
8) and SS imported with or without avidin (Fig. 4, lanes
3 and 5) were present in the soluble fraction. The
simplest explanation is that biotinylated prSS is trapped in the
translocation apparatus when avidin is present.
However, biotinylated prSS was not processed by the stromal peptidase when avidin was present. From this result, we conclude that complexes of avidin and biotinylated precursors do not extend through the import apparatus sufficiently far to gain access to the stroma processing protease. Indeed, it is technically possible, although unlikely, that the complexes of avidin and biotinylated precursors were not specifically associating with the import apparatus but, rather, were simply aggregating at the surface of chloroplasts. To address this concern, biotinylated precursors were bound to chloroplasts before the addition of avidin.
Fig. 5 shows results from such an experimental approach.
Import of native prSS from the prebound state was not affected by the
presence of avidin (Fig. 5, lane 3); however, in the
presence of avidin, import of biotinylated prSS import was blocked
(Fig. 5, lane 7). Complexes of avidin-biotinylated prSS
assembled from the prebound state fractionated with membranes. In the
absence of avidin, prebound biotinylated prSS was imported into
chloroplasts and fractionated to the soluble portion (Fig. 5,
lane 6). From these results we conclude that prebinding of
biotinylated precursor to chloroplasts does not prevent avidin from
interacting with available biotinylated cysteines within the precursor.
However, it is uncertain which biotinylated cysteine is interacting
with avidin.
Titration of Import Sites using Avidin-Biotinylated Precursor Complexes
Import intermediates can serve as valuable tools for
determining the number of import sites present on a chloroplast.
Because binding of biotinylated precursors is specific, as demonstrated by their nearly complete import on addition of ATP (Fig. 5, lane 6), the complexes formed with these precursors should be useful to
titrate chloroplastic import sites. Import sites became saturated with
increasing amounts of avidin-biotinylated precursor complexes (Fig.
6A). The quantities of avidin-biotinylated
prSS complexes that associated with the membrane fraction were
measured, and the resulting data were graphed (Fig. 6B). The
number of complexes reached saturation at approximately 2,000 molecules/chloroplast (Fig. 6B). In a control experiment,
biotinylated prSS at a high concentration (245 nM) in the
absence of avidin was imported into chloroplasts and was present in the
soluble fraction (data not shown). This control confirms that
association of biotinylated precursors with membranes was a result of
avidin complex formation and not due to aggregation.
Avidin-Precursor Complexes Inhibit the Import of Wheat Germ-translated prSS
The results presented above suggest that
under import conditions avidin-precursor complexes can initiate
translocation but become blocked in import sites because of the
presence of a bulky avidin protein. If partially translocated avidin
complexes remain stably associated with the import machinery, the
limited number of import sites of a chloroplast will eventually be
inactivated, leading to a reduction in the capacity of chloroplasts to
import other precursor proteins. To examine this possibility,
chloroplasts containing trapped avidin-biotinylated precursor complexes
were incubated with wheat germ-translated 35S-prSS, and its
rate of import was monitored. Chloroplasts were incubated with 79 nM biotinylated prSS, and avidin complexes formed as
described in Fig. 6. When radiolabeled prSS was examined, its import
was inhibited (Fig. 7A, compare rows
4 and 1). Fig. 7B quantitates data in Fig.
7A and shows that the rate of import of prSS is reduced by
approximately 35% when import sites are blocked with
avidin-biotinylated precursor complexes (Fig. 7B, compare
lines 1 and 4). This inhibition was dependent on
the formation of avidin-precursor complexes, because the rate of import
of prSS was not affected when avidin alone (Fig. 7, A, row
2, and B, line 2) or biotinylated prSS (Fig. 7,
A, row 3, and B, line 3) were incubated
independently. From these results, we conclude that avidin-biotinylated
precursor complexes remain specifically associated with import sites
that are also used by prSS. Blockage of the import apparatus by these
complexes prevents the import of prSS into chloroplasts.
The successful generation of translocation intermediates provides an effective tool in dissecting the process by which proteins traverse the chloroplastic envelope membranes. Translocation intermediates can be used to identify, characterize, and purify components of the import apparatus in addition to delineating individual steps in the import process. In the present investigation, we have used biotinylated precursors to generate translocation intermediates that block protein import into chloroplasts. We have demonstrated that avidin-precursor complexes could be formed by adding avidin to biotinylated precursors prebound to the import machinery. The bulky avidin complex was unable to be translocated through the transport machinery, resulting in obstructed import sites. The effectiveness of this obstruction was shown when freshly translated prSS was added to chloroplasts containing obstructed import sites. Our titration experiment showed that a 79 nM concentration of biotinylated precursor was sufficient to occupy approximately 800 translocation sites per chloroplast once avidin was added (Fig. 6). This concentration should result in occupation by avidin-biotinylated precursor complexes of approximately 40% of the translocation sites on a chloroplast. At this nonsaturating concentration of complex, it was demonstrated that import of freshly added prSS was inhibited by 35%, indicating that blockage of functional import sites had occurred (Fig. 7). The degree of inhibition is consistent with the reduction in the number of functional unobstructed translocation sites. However, the most important point is that avidin-precursor complexes are blocking functional import sites, as evidenced by the reduced rate of import of freshly added precursor proteins.
The observation that the complexes block import of other precursors indicates they must be in close proximity to the translocation apparatus. One likely explanation is that the precursor is physically stuck in the channel of the import apparatus, thereby preventing the import of other precursors. Another possibility is that the precursor has slipped out of the channel but remains sufficiently close that the avidin moiety of the complex is sterically hindering the binding of freshly added precursor to receptor(s) located on the surface of the chloroplast. However, further work will be required to elucidate the topology or configuration of the arrested precursor-avidin complex.
Others have used different strategies to estimate the number of binding sites located on a chloroplast. Using wheat germ-translated prSS, Friedman and Keegstra (33) demonstrated that saturation of binding occurred at 1,500-3,000 precursors/chloroplast. Schnell and Blobel (9), when using the urea-denatured fusion protein prSS/protA, calculated that saturation of binding was obtained at 2,000 precursors/chloroplast. We have attempted to use halted complexes to calculate the number of translocation sites per chloroplast (Fig. 6). After quantitating membranes with blocked avidin complexes, we calculated that saturation occurred at approximately 2,000 precursors/chloroplast (Fig. 6B). This is in agreement with the results of Friedman and Keegstra (33) and Schnell and Blobel (9). However, the number of precursors associated with a chloroplast, as reported by Friedman and Keegstra (33) and Schnell and Blobel (9), represents a combination of binding and early transport intermediates. In binding experiments precursors interact with both surface receptor(s) and components of the import apparatus. Halted avidin-precursor complexes, however, represent precursors that have exclusively engaged the import apparatus at a later stage in import. The level of saturation calculated from our titration experiments therefore represents the number of translocation sites present per chloroplast. Studies with translocation intermediates using mitochondria have estimated that the number of import sites is 102-103/mitochondrion (17, 34). Another study using an avidin fusion protein to generate translocation intermediates calculated approximately 600 import sites/mitochondrion (35).
The translocation intermediates generated cannot be chased into the stroma even under import conditions. This is in contrast to the translocation intermediate generated by Schnell and Blobel (9) using a chimeric precursor protein containing portions of prSS fused to protein A. In their studies a translocation intermediate was observed early during the time course of import, but at later times the fusion protein was chased into the stroma. Thus the approach reported here provides a means for generating permanently arrested translocation intermediates, whereas the approach of Schnell and Blobel (9) produced a slowly imported chimeric precursor that allowed a transitory translocation intermediate to be identified.
Our experiments also revealed that biotinylated precursors can serve as a valuable tool for examining and isolating the chloroplastic import machinery. Modification of precursors with biotin does not dramatically affect their capacity for specific binding to chloroplast. Import likewise does not appear to be greatly impaired. However, processing of newly imported biotinylated prSS appears to be aberrant, since the mobilities of SS-BMCC and SS are similar (Fig. 3, compare lanes 3 and 6). Of the four cysteines within prSS that can be modified with biotin-BMCC, one is located near the cleavage site for the transit peptide (see Fig. 1). This particular modification appears to result in aberrant processing of prSS-BMCC. This result is not surprising; numerous studies on chloroplastic transit peptides in which various amino acid residues have been mutated near the cleavage site have yielded similar reports of aberrant processing (36-38). Robinson and Ellis (39) demonstrated that when cysteines on prSS are modified with iodoacetic acid, import is not affected, but processing is aberrant. However, we cannot exclude other possible explanations for the similar mobility of SS and SS-BMCC. Regardless of this uncertainty, the utility of biotinylated precursors is not diminished.
Biotinylated precursors should be useful in identifying and isolating components of the chloroplastic import machinery. One strategy would entail binding biotinylated precursors to the translocation apparatus, followed by purification of translocation complexes by affinity chromatography using immobilized avidin. This and other possibilities for investigating the import process using biotinylated precursors are underway.